drm/amd/display: Introduce DML2
authorQingqing Zhuo <Qingqing.Zhuo@amd.com>
Fri, 28 Jul 2023 21:55:30 +0000 (17:55 -0400)
committerAlex Deucher <alexander.deucher@amd.com>
Mon, 9 Oct 2023 20:48:51 +0000 (16:48 -0400)
DC is transitioning from DML to DML2, and this commit introduces all the
required changes for some of the already available ASICs and adds the
required code infra to support new ASICs under DML2. DML2 is also a
generated code that provides better mode verification and programming
models for software/hardware, and it enables a better way to create
validation tools. This version is more like a middle step to the
complete transition to the DML2 version.

Changes since V1:
- Alex: Fix typos

Changes since V2:
- Update DC includes

Changes since V3:
- Fix 32 bit compilation issues on x86

Changes since V4:
- Avoid compilation of DML2 on some not supported 32-bit architecture
- Update commit message

Co-developed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Co-developed-by: Roman Li <roman.li@amd.com>
Signed-off-by: Roman Li <roman.li@amd.com>
Signed-off-by: Qingqing Zhuo <Qingqing.Zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
36 files changed:
drivers/gpu/drm/amd/display/dc/Makefile
drivers/gpu/drm/amd/display/dc/core/dc.c
drivers/gpu/drm/amd/display/dc/core/dc_resource.c
drivers/gpu/drm/amd/display/dc/dc.h
drivers/gpu/drm/amd/display/dc/dcn32/dcn32_resource.c
drivers/gpu/drm/amd/display/dc/dcn321/dcn321_resource.c
drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c
drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c
drivers/gpu/drm/amd/display/dc/dml2/Makefile [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/cmntypes.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/display_mode_core_structs.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/display_mode_lib_defines.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/display_mode_util.c [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/display_mode_util.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_types.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml2_internal_types.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml2_mall_phantom.c [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml2_mall_phantom.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml2_policy.c [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml2_policy.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml2_utils.c [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml2_utils.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml_assert.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml_depedencies.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml_display_rq_dlg_calc.c [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml_display_rq_dlg_calc.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/dml2/dml_logging.h [new file with mode: 0644]
drivers/gpu/drm/amd/display/dc/inc/core_types.h

index dafa34b..53ec782 100644 (file)
@@ -28,8 +28,8 @@ ifdef CONFIG_DRM_AMD_DC_FP
 
 KCOV_INSTRUMENT := n
 
-DC_LIBS += dcn20
 DC_LIBS += dcn10
+DC_LIBS += dcn20
 DC_LIBS += dcn21
 DC_LIBS += dcn201
 DC_LIBS += dcn30
@@ -44,6 +44,7 @@ DC_LIBS += dcn32
 DC_LIBS += dcn321
 DC_LIBS += dcn35
 DC_LIBS += dml
+DC_LIBS += dml2
 endif
 
 DC_LIBS += dce120
index b3af7b1..2436a29 100644 (file)
@@ -77,6 +77,8 @@
 
 #include "hw_sequencer_private.h"
 
+#include "dml2/dml2_internal_types.h"
+
 #include "dce/dmub_outbox.h"
 
 #define CTX \
@@ -2176,6 +2178,11 @@ struct dc_state *dc_create_state(struct dc *dc)
 
        init_state(dc, context);
 
+#ifdef CONFIG_DRM_AMD_DC_FP
+       if (dc->debug.using_dml2) {
+               dml2_create(dc, &dc->dml2_options, &context->bw_ctx.dml2);
+       }
+#endif
        kref_init(&context->refcount);
 
        return context;
@@ -2185,11 +2192,25 @@ struct dc_state *dc_copy_state(struct dc_state *src_ctx)
 {
        int i, j;
        struct dc_state *new_ctx = kvmalloc(sizeof(struct dc_state), GFP_KERNEL);
+#ifdef CONFIG_DRM_AMD_DC_FP
+       struct dml2_context *dml2 =  NULL;
+#endif
 
        if (!new_ctx)
                return NULL;
        memcpy(new_ctx, src_ctx, sizeof(struct dc_state));
 
+#ifdef CONFIG_DRM_AMD_DC_FP
+       if (new_ctx->bw_ctx.dml2) {
+               dml2 = kzalloc(sizeof(struct dml2_context), GFP_KERNEL);
+               if (!dml2)
+                       return NULL;
+
+               memcpy(dml2, src_ctx->bw_ctx.dml2, sizeof(struct dml2_context));
+               new_ctx->bw_ctx.dml2 = dml2;
+       }
+#endif
+
        for (i = 0; i < MAX_PIPES; i++) {
                        struct pipe_ctx *cur_pipe = &new_ctx->res_ctx.pipe_ctx[i];
 
@@ -2228,6 +2249,12 @@ static void dc_state_free(struct kref *kref)
 {
        struct dc_state *context = container_of(kref, struct dc_state, refcount);
        dc_resource_state_destruct(context);
+
+#ifdef CONFIG_DRM_AMD_DC_FP
+       dml2_destroy(context->bw_ctx.dml2);
+       context->bw_ctx.dml2 = 0;
+#endif
+
        kvfree(context);
 }
 
@@ -4679,6 +4706,9 @@ bool dc_set_power_state(
 {
        struct kref refcount;
        struct display_mode_lib *dml;
+#ifdef CONFIG_DRM_AMD_DC_FP
+       struct dml2_context *dml2 = NULL;
+#endif
 
        if (!dc->current_state)
                return true;
@@ -4698,6 +4728,10 @@ bool dc_set_power_state(
 
                break;
        default:
+#ifdef CONFIG_DRM_AMD_DC_FP
+               if (dc->debug.using_dml2)
+                       dml2 = dc->current_state->bw_ctx.dml2;
+#endif
                ASSERT(dc->current_state->stream_count == 0);
                /* Zero out the current context so that on resume we start with
                 * clean state, and dc hw programming optimizations will not
@@ -4724,6 +4758,11 @@ bool dc_set_power_state(
 
                kfree(dml);
 
+#ifdef CONFIG_DRM_AMD_DC_FP
+               if (dc->debug.using_dml2)
+                       dc->current_state->bw_ctx.dml2 = dml2;
+#endif
+
                break;
        }
 
index aa7b5db..c1e2f0e 100644 (file)
@@ -41,6 +41,7 @@
 #include "dpcd_defs.h"
 #include "link_enc_cfg.h"
 #include "link.h"
+#include "clk_mgr.h"
 #include "virtual/virtual_link_hwss.h"
 #include "link/hwss/link_hwss_dio.h"
 #include "link/hwss/link_hwss_dpia.h"
@@ -86,6 +87,8 @@
        dc->ctx->logger
 #define DC_LOGGER_INIT(logger)
 
+#include "dml2/dml2_wrapper.h"
+
 #define UNABLE_TO_SPLIT -1
 
 enum dce_version resource_parse_asic_id(struct hw_asic_id asic_id)
@@ -318,6 +321,10 @@ struct resource_pool *dc_create_resource_pool(struct dc  *dc,
                                res_pool->ref_clocks.xtalin_clock_inKhz;
                        res_pool->ref_clocks.dchub_ref_clock_inKhz =
                                res_pool->ref_clocks.xtalin_clock_inKhz;
+                       if ((res_pool->hubbub->funcs->get_dchub_ref_freq))
+                               res_pool->hubbub->funcs->get_dchub_ref_freq(res_pool->hubbub,
+                                       res_pool->ref_clocks.dccg_ref_clock_inKhz,
+                                       &res_pool->ref_clocks.dchub_ref_clock_inKhz);
                } else
                        ASSERT_CRITICAL(false);
        }
@@ -4358,9 +4365,22 @@ void dc_resource_state_copy_construct(
 {
        int i, j;
        struct kref refcount = dst_ctx->refcount;
+#ifdef CONFIG_DRM_AMD_DC_FP
+       struct dml2_context *dml2 = NULL;
+
+       // Need to preserve allocated dml2 context
+       if (src_ctx->clk_mgr->ctx->dc->debug.using_dml2)
+               dml2 = dst_ctx->bw_ctx.dml2;
+#endif
 
        *dst_ctx = *src_ctx;
 
+#ifdef CONFIG_DRM_AMD_DC_FP
+       // Preserve allocated dml2 context
+       if (src_ctx->clk_mgr->ctx->dc->debug.using_dml2)
+               dst_ctx->bw_ctx.dml2 = dml2;
+#endif
+
        for (i = 0; i < MAX_PIPES; i++) {
                struct pipe_ctx *cur_pipe = &dst_ctx->res_ctx.pipe_ctx[i];
 
index d38194a..8233cd9 100644 (file)
@@ -40,6 +40,8 @@
 #include "inc/hw/dmcu.h"
 #include "dml/display_mode_lib.h"
 
+#include "dml2/dml2_wrapper.h"
+
 struct abm_save_restore;
 
 /* forward declaration */
@@ -942,6 +944,7 @@ struct dc_debug_options {
        bool dml_disallow_alternate_prefetch_modes;
        bool use_legacy_soc_bb_mechanism;
        bool exit_idle_opt_for_cursor_updates;
+       bool using_dml2;
        bool enable_single_display_2to1_odm_policy;
        bool enable_double_buffered_dsc_pg_support;
        bool enable_dp_dig_pixel_rate_div_policy;
@@ -1049,6 +1052,8 @@ struct dc {
                        struct _vcs_dpi_voltage_scaling_st clock_limits[DC__VOLTAGE_STATES];
                } update_bw_bounding_box;
        } scratch;
+
+       struct dml2_configuration_options dml2_options;
 };
 
 enum frame_buffer_mode {
index a74d4ca..4eda12e 100644 (file)
@@ -89,6 +89,8 @@
 #include "dcn20/dcn20_vmid.h"
 #include "dml/dcn32/dcn32_fpu.h"
 
+#include "dml2/dml2_wrapper.h"
+
 #define DC_LOGGER_INIT(logger)
 
 enum dcn32_clk_src_array_id {
@@ -714,6 +716,7 @@ static const struct dc_debug_options debug_defaults_drv = {
        .use_max_lb = true,
        .force_disable_subvp = false,
        .exit_idle_opt_for_cursor_updates = true,
+       .using_dml2 = false,
        .enable_single_display_2to1_odm_policy = true,
 
        /* Must match enable_single_display_2to1_odm_policy to support dynamic ODM transitions*/
@@ -1805,9 +1808,7 @@ void dcn32_add_phantom_pipes(struct dc *dc, struct dc_state *context,
        }
 }
 
-bool dcn32_validate_bandwidth(struct dc *dc,
-               struct dc_state *context,
-               bool fast_validate)
+static bool dml1_validate(struct dc *dc, struct dc_state *context, bool fast_validate)
 {
        bool out = false;
 
@@ -1885,6 +1886,19 @@ validate_out:
        return out;
 }
 
+bool dcn32_validate_bandwidth(struct dc *dc,
+               struct dc_state *context,
+               bool fast_validate)
+{
+       bool out = false;
+
+       if (dc->debug.using_dml2)
+               out = dml2_validate(dc, context, fast_validate);
+       else
+               out = dml1_validate(dc, context, fast_validate);
+       return out;
+}
+
 int dcn32_populate_dml_pipes_from_context(
        struct dc *dc, struct dc_state *context,
        display_e2e_pipe_params_st *pipes,
@@ -2422,6 +2436,47 @@ static bool dcn32_resource_construct(
                pool->base.oem_device = NULL;
        }
 
+       dc->dml2_options.dcn_pipe_count = pool->base.pipe_count;
+       dc->dml2_options.use_native_pstate_optimization = false;
+       dc->dml2_options.use_native_soc_bb_construction = true;
+       dc->dml2_options.minimize_dispclk_using_odm = true;
+
+       dc->dml2_options.callbacks.dc = dc;
+       dc->dml2_options.callbacks.build_scaling_params = &resource_build_scaling_params;
+       dc->dml2_options.callbacks.can_support_mclk_switch_using_fw_based_vblank_stretch = &dcn30_can_support_mclk_switch_using_fw_based_vblank_stretch;
+       dc->dml2_options.callbacks.acquire_secondary_pipe_for_mpc_odm = &dc_resource_acquire_secondary_pipe_for_mpc_odm_legacy;
+
+       dc->dml2_options.svp_pstate.callbacks.dc = dc;
+       dc->dml2_options.svp_pstate.callbacks.add_plane_to_context = &dc_add_plane_to_context;
+       dc->dml2_options.svp_pstate.callbacks.add_stream_to_ctx = &dc_add_stream_to_ctx;
+       dc->dml2_options.svp_pstate.callbacks.build_scaling_params = &resource_build_scaling_params;
+       dc->dml2_options.svp_pstate.callbacks.create_plane = &dc_create_plane_state;
+       dc->dml2_options.svp_pstate.callbacks.remove_plane_from_context = &dc_remove_plane_from_context;
+       dc->dml2_options.svp_pstate.callbacks.remove_stream_from_ctx = &dc_remove_stream_from_ctx;
+       dc->dml2_options.svp_pstate.callbacks.create_stream_for_sink = &dc_create_stream_for_sink;
+       dc->dml2_options.svp_pstate.callbacks.plane_state_release = &dc_plane_state_release;
+       dc->dml2_options.svp_pstate.callbacks.stream_release = &dc_stream_release;
+       dc->dml2_options.svp_pstate.callbacks.release_dsc = &dcn20_release_dsc;
+
+       dc->dml2_options.svp_pstate.subvp_fw_processing_delay_us = dc->caps.subvp_fw_processing_delay_us;
+       dc->dml2_options.svp_pstate.subvp_prefetch_end_to_mall_start_us = dc->caps.subvp_prefetch_end_to_mall_start_us;
+       dc->dml2_options.svp_pstate.subvp_pstate_allow_width_us = dc->caps.subvp_pstate_allow_width_us;
+       dc->dml2_options.svp_pstate.subvp_swath_height_margin_lines = dc->caps.subvp_swath_height_margin_lines;
+
+       dc->dml2_options.svp_pstate.force_disable_subvp = dc->debug.force_disable_subvp;
+       dc->dml2_options.svp_pstate.force_enable_subvp = dc->debug.force_subvp_mclk_switch;
+
+       dc->dml2_options.mall_cfg.cache_line_size_bytes = dc->caps.cache_line_size;
+       dc->dml2_options.mall_cfg.cache_num_ways = dc->caps.cache_num_ways;
+       dc->dml2_options.mall_cfg.max_cab_allocation_bytes = dc->caps.max_cab_allocation_bytes;
+       dc->dml2_options.mall_cfg.mblk_height_4bpe_pixels = DCN3_2_MBLK_HEIGHT_4BPE;
+       dc->dml2_options.mall_cfg.mblk_height_8bpe_pixels = DCN3_2_MBLK_HEIGHT_8BPE;
+       dc->dml2_options.mall_cfg.mblk_size_bytes = DCN3_2_MALL_MBLK_SIZE_BYTES;
+       dc->dml2_options.mall_cfg.mblk_width_pixels = DCN3_2_MBLK_WIDTH;
+
+       dc->dml2_options.max_segments_per_hubp = 18;
+       dc->dml2_options.det_segment_size = DCN3_2_DET_SEG_SIZE;
+
        if (ASICREV_IS_GC_11_0_3(dc->ctx->asic_id.hw_internal_rev) && (dc->config.sdpif_request_limit_words_per_umc == 0))
                dc->config.sdpif_request_limit_words_per_umc = 16;
 
index f11ff6f..2cb6369 100644 (file)
@@ -1989,6 +1989,47 @@ static bool dcn321_resource_construct(
                pool->base.oem_device = NULL;
        }
 
+       dc->dml2_options.dcn_pipe_count = pool->base.pipe_count;
+       dc->dml2_options.use_native_pstate_optimization = false;
+       dc->dml2_options.use_native_soc_bb_construction = true;
+       dc->dml2_options.minimize_dispclk_using_odm = true;
+
+       dc->dml2_options.callbacks.dc = dc;
+       dc->dml2_options.callbacks.build_scaling_params = &resource_build_scaling_params;
+       dc->dml2_options.callbacks.can_support_mclk_switch_using_fw_based_vblank_stretch = &dcn30_can_support_mclk_switch_using_fw_based_vblank_stretch;
+       dc->dml2_options.callbacks.acquire_secondary_pipe_for_mpc_odm = &dc_resource_acquire_secondary_pipe_for_mpc_odm_legacy;
+
+       dc->dml2_options.svp_pstate.callbacks.dc = dc;
+       dc->dml2_options.svp_pstate.callbacks.add_plane_to_context = &dc_add_plane_to_context;
+       dc->dml2_options.svp_pstate.callbacks.add_stream_to_ctx = &dc_add_stream_to_ctx;
+       dc->dml2_options.svp_pstate.callbacks.build_scaling_params = &resource_build_scaling_params;
+       dc->dml2_options.svp_pstate.callbacks.create_plane = &dc_create_plane_state;
+       dc->dml2_options.svp_pstate.callbacks.remove_plane_from_context = &dc_remove_plane_from_context;
+       dc->dml2_options.svp_pstate.callbacks.remove_stream_from_ctx = &dc_remove_stream_from_ctx;
+       dc->dml2_options.svp_pstate.callbacks.create_stream_for_sink = &dc_create_stream_for_sink;
+       dc->dml2_options.svp_pstate.callbacks.plane_state_release = &dc_plane_state_release;
+       dc->dml2_options.svp_pstate.callbacks.stream_release = &dc_stream_release;
+       dc->dml2_options.svp_pstate.callbacks.release_dsc = &dcn20_release_dsc;
+
+       dc->dml2_options.svp_pstate.subvp_fw_processing_delay_us = dc->caps.subvp_fw_processing_delay_us;
+       dc->dml2_options.svp_pstate.subvp_prefetch_end_to_mall_start_us = dc->caps.subvp_prefetch_end_to_mall_start_us;
+       dc->dml2_options.svp_pstate.subvp_pstate_allow_width_us = dc->caps.subvp_pstate_allow_width_us;
+       dc->dml2_options.svp_pstate.subvp_swath_height_margin_lines = dc->caps.subvp_swath_height_margin_lines;
+
+       dc->dml2_options.svp_pstate.force_disable_subvp = dc->debug.force_disable_subvp;
+       dc->dml2_options.svp_pstate.force_enable_subvp = dc->debug.force_subvp_mclk_switch;
+
+       dc->dml2_options.mall_cfg.cache_line_size_bytes = dc->caps.cache_line_size;
+       dc->dml2_options.mall_cfg.cache_num_ways = dc->caps.cache_num_ways;
+       dc->dml2_options.mall_cfg.max_cab_allocation_bytes = dc->caps.max_cab_allocation_bytes;
+       dc->dml2_options.mall_cfg.mblk_height_4bpe_pixels = DCN3_2_MBLK_HEIGHT_4BPE;
+       dc->dml2_options.mall_cfg.mblk_height_8bpe_pixels = DCN3_2_MBLK_HEIGHT_8BPE;
+       dc->dml2_options.mall_cfg.mblk_size_bytes = DCN3_2_MALL_MBLK_SIZE_BYTES;
+       dc->dml2_options.mall_cfg.mblk_width_pixels = DCN3_2_MBLK_WIDTH;
+
+       dc->dml2_options.max_segments_per_hubp = 18;
+       dc->dml2_options.det_segment_size = DCN3_2_DET_SEG_SIZE;
+
        return true;
 
 create_fail:
index 7179c2b..b011779 100644 (file)
@@ -2957,12 +2957,14 @@ void dcn32_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_pa
        /* Override from passed dc->bb_overrides if available*/
        if ((int)(dcn3_2_soc.sr_exit_time_us * 1000) != dc->bb_overrides.sr_exit_time_ns
                        && dc->bb_overrides.sr_exit_time_ns) {
+               dc->dml2_options.bbox_overrides.sr_exit_latency_us =
                dcn3_2_soc.sr_exit_time_us = dc->bb_overrides.sr_exit_time_ns / 1000.0;
        }
 
        if ((int)(dcn3_2_soc.sr_enter_plus_exit_time_us * 1000)
                        != dc->bb_overrides.sr_enter_plus_exit_time_ns
                        && dc->bb_overrides.sr_enter_plus_exit_time_ns) {
+               dc->dml2_options.bbox_overrides.sr_enter_plus_exit_latency_us =
                dcn3_2_soc.sr_enter_plus_exit_time_us =
                        dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0;
        }
@@ -2970,12 +2972,14 @@ void dcn32_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_pa
        if ((int)(dcn3_2_soc.urgent_latency_us * 1000) != dc->bb_overrides.urgent_latency_ns
                && dc->bb_overrides.urgent_latency_ns) {
                dcn3_2_soc.urgent_latency_us = dc->bb_overrides.urgent_latency_ns / 1000.0;
+               dc->dml2_options.bbox_overrides.urgent_latency_us =
                dcn3_2_soc.urgent_latency_pixel_data_only_us = dc->bb_overrides.urgent_latency_ns / 1000.0;
        }
 
        if ((int)(dcn3_2_soc.dram_clock_change_latency_us * 1000)
                        != dc->bb_overrides.dram_clock_change_latency_ns
                        && dc->bb_overrides.dram_clock_change_latency_ns) {
+               dc->dml2_options.bbox_overrides.dram_clock_change_latency_us =
                dcn3_2_soc.dram_clock_change_latency_us =
                        dc->bb_overrides.dram_clock_change_latency_ns / 1000.0;
        }
@@ -2983,6 +2987,7 @@ void dcn32_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_pa
        if ((int)(dcn3_2_soc.fclk_change_latency_us * 1000)
                        != dc->bb_overrides.fclk_clock_change_latency_ns
                        && dc->bb_overrides.fclk_clock_change_latency_ns) {
+               dc->dml2_options.bbox_overrides.fclk_change_latency_us =
                dcn3_2_soc.fclk_change_latency_us =
                        dc->bb_overrides.fclk_clock_change_latency_ns / 1000;
        }
@@ -3000,14 +3005,17 @@ void dcn32_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_pa
 
                if (dc->ctx->dc_bios->funcs->get_soc_bb_info(dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) {
                        if (bb_info.dram_clock_change_latency_100ns > 0)
+                               dc->dml2_options.bbox_overrides.dram_clock_change_latency_us =
                                dcn3_2_soc.dram_clock_change_latency_us =
                                        bb_info.dram_clock_change_latency_100ns * 10;
 
                        if (bb_info.dram_sr_enter_exit_latency_100ns > 0)
+                               dc->dml2_options.bbox_overrides.sr_enter_plus_exit_latency_us =
                                dcn3_2_soc.sr_enter_plus_exit_time_us =
                                        bb_info.dram_sr_enter_exit_latency_100ns * 10;
 
                        if (bb_info.dram_sr_exit_latency_100ns > 0)
+                               dc->dml2_options.bbox_overrides.sr_exit_latency_us =
                                dcn3_2_soc.sr_exit_time_us =
                                        bb_info.dram_sr_exit_latency_100ns * 10;
                }
@@ -3015,12 +3023,14 @@ void dcn32_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_pa
 
        /* Override from VBIOS for num_chan */
        if (dc->ctx->dc_bios->vram_info.num_chans) {
+               dc->dml2_options.bbox_overrides.dram_num_chan =
                dcn3_2_soc.num_chans = dc->ctx->dc_bios->vram_info.num_chans;
                dcn3_2_soc.mall_allocated_for_dcn_mbytes = (double)(dcn32_calc_num_avail_chans_for_mall(dc,
                        dc->ctx->dc_bios->vram_info.num_chans) * dc->caps.mall_size_per_mem_channel);
        }
 
        if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes)
+               dc->dml2_options.bbox_overrides.dram_chanel_width_bytes =
                dcn3_2_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes;
 
        /* DML DSC delay factor workaround */
@@ -3031,6 +3041,10 @@ void dcn32_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_pa
        /* Override dispclk_dppclk_vco_speed_mhz from Clk Mgr */
        dcn3_2_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
        dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+       dc->dml2_options.bbox_overrides.disp_pll_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+       dc->dml2_options.bbox_overrides.xtalclk_mhz = dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency / 1000.0;
+       dc->dml2_options.bbox_overrides.dchub_refclk_mhz = dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000.0;
+       dc->dml2_options.bbox_overrides.dprefclk_mhz = dc->clk_mgr->dprefclk_khz / 1000.0;
 
        /* Overrides Clock levelsfrom CLK Mgr table entries as reported by PM FW */
        if (bw_params->clk_table.entries[0].memclk_mhz) {
@@ -3186,6 +3200,72 @@ void dcn32_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_pa
                if (dc->current_state)
                        dml_init_instance(&dc->current_state->bw_ctx.dml, &dcn3_2_soc, &dcn3_2_ip, DML_PROJECT_DCN32);
        }
+
+       if (dc->clk_mgr->bw_params->clk_table.num_entries > 1) {
+               unsigned int i = 0;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_states = dc->clk_mgr->bw_params->clk_table.num_entries;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dcfclk_levels =
+                       dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dcfclk_levels;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_fclk_levels =
+                       dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_fclk_levels;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_memclk_levels =
+                       dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_memclk_levels;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_socclk_levels =
+                       dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_socclk_levels;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dtbclk_levels =
+                       dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dtbclk_levels;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dispclk_levels =
+                       dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dispclk_levels;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dppclk_levels =
+                       dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dppclk_levels;
+
+               for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dcfclk_levels; i++) {
+                       if (dc->clk_mgr->bw_params->clk_table.entries[i].dcfclk_mhz)
+                               dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dcfclk_mhz =
+                                       dc->clk_mgr->bw_params->clk_table.entries[i].dcfclk_mhz;
+               }
+
+               for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_fclk_levels; i++) {
+                       if (dc->clk_mgr->bw_params->clk_table.entries[i].fclk_mhz)
+                               dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].fclk_mhz =
+                                       dc->clk_mgr->bw_params->clk_table.entries[i].fclk_mhz;
+               }
+
+               for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_memclk_levels; i++) {
+                       if (dc->clk_mgr->bw_params->clk_table.entries[i].memclk_mhz)
+                               dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].memclk_mhz =
+                                       dc->clk_mgr->bw_params->clk_table.entries[i].memclk_mhz;
+               }
+
+               for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_socclk_levels; i++) {
+                       if (dc->clk_mgr->bw_params->clk_table.entries[i].socclk_mhz)
+                               dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].socclk_mhz =
+                                       dc->clk_mgr->bw_params->clk_table.entries[i].socclk_mhz;
+               }
+
+               for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dtbclk_levels; i++) {
+                       if (dc->clk_mgr->bw_params->clk_table.entries[i].dtbclk_mhz)
+                               dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dtbclk_mhz =
+                                       dc->clk_mgr->bw_params->clk_table.entries[i].dtbclk_mhz;
+               }
+
+               for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dispclk_levels; i++) {
+                       if (dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz) {
+                               dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dispclk_mhz =
+                                       dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz;
+                               dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dppclk_mhz =
+                                       dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz;
+                       }
+               }
+       }
 }
 
 void dcn32_zero_pipe_dcc_fraction(display_e2e_pipe_params_st *pipes,
index b26fcf8..ff4d795 100644 (file)
@@ -616,12 +616,14 @@ void dcn321_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_p
        /* Override from passed dc->bb_overrides if available*/
        if ((int)(dcn3_21_soc.sr_exit_time_us * 1000) != dc->bb_overrides.sr_exit_time_ns
                        && dc->bb_overrides.sr_exit_time_ns) {
+               dc->dml2_options.bbox_overrides.sr_exit_latency_us =
                dcn3_21_soc.sr_exit_time_us = dc->bb_overrides.sr_exit_time_ns / 1000.0;
        }
 
        if ((int)(dcn3_21_soc.sr_enter_plus_exit_time_us * 1000)
                        != dc->bb_overrides.sr_enter_plus_exit_time_ns
                        && dc->bb_overrides.sr_enter_plus_exit_time_ns) {
+               dc->dml2_options.bbox_overrides.sr_enter_plus_exit_latency_us =
                dcn3_21_soc.sr_enter_plus_exit_time_us =
                        dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0;
        }
@@ -629,12 +631,14 @@ void dcn321_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_p
        if ((int)(dcn3_21_soc.urgent_latency_us * 1000) != dc->bb_overrides.urgent_latency_ns
                && dc->bb_overrides.urgent_latency_ns) {
                dcn3_21_soc.urgent_latency_us = dc->bb_overrides.urgent_latency_ns / 1000.0;
+               dc->dml2_options.bbox_overrides.urgent_latency_us =
                dcn3_21_soc.urgent_latency_pixel_data_only_us = dc->bb_overrides.urgent_latency_ns / 1000.0;
        }
 
        if ((int)(dcn3_21_soc.dram_clock_change_latency_us * 1000)
                        != dc->bb_overrides.dram_clock_change_latency_ns
                        && dc->bb_overrides.dram_clock_change_latency_ns) {
+               dc->dml2_options.bbox_overrides.dram_clock_change_latency_us =
                dcn3_21_soc.dram_clock_change_latency_us =
                        dc->bb_overrides.dram_clock_change_latency_ns / 1000.0;
        }
@@ -642,6 +646,7 @@ void dcn321_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_p
        if ((int)(dcn3_21_soc.fclk_change_latency_us * 1000)
                        != dc->bb_overrides.fclk_clock_change_latency_ns
                        && dc->bb_overrides.fclk_clock_change_latency_ns) {
+               dc->dml2_options.bbox_overrides.fclk_change_latency_us =
                dcn3_21_soc.fclk_change_latency_us =
                        dc->bb_overrides.fclk_clock_change_latency_ns / 1000;
        }
@@ -659,14 +664,17 @@ void dcn321_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_p
 
                if (dc->ctx->dc_bios->funcs->get_soc_bb_info(dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) {
                        if (bb_info.dram_clock_change_latency_100ns > 0)
+                               dc->dml2_options.bbox_overrides.dram_clock_change_latency_us =
                                dcn3_21_soc.dram_clock_change_latency_us =
                                        bb_info.dram_clock_change_latency_100ns * 10;
 
                        if (bb_info.dram_sr_enter_exit_latency_100ns > 0)
+                               dc->dml2_options.bbox_overrides.sr_enter_plus_exit_latency_us =
                                dcn3_21_soc.sr_enter_plus_exit_time_us =
                                        bb_info.dram_sr_enter_exit_latency_100ns * 10;
 
                        if (bb_info.dram_sr_exit_latency_100ns > 0)
+                               dc->dml2_options.bbox_overrides.sr_exit_latency_us =
                                dcn3_21_soc.sr_exit_time_us =
                                        bb_info.dram_sr_exit_latency_100ns * 10;
                }
@@ -674,12 +682,14 @@ void dcn321_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_p
 
        /* Override from VBIOS for num_chan */
        if (dc->ctx->dc_bios->vram_info.num_chans) {
+               dc->dml2_options.bbox_overrides.dram_num_chan =
                dcn3_21_soc.num_chans = dc->ctx->dc_bios->vram_info.num_chans;
                dcn3_21_soc.mall_allocated_for_dcn_mbytes = (double)(dcn32_calc_num_avail_chans_for_mall(dc,
                        dc->ctx->dc_bios->vram_info.num_chans) * dc->caps.mall_size_per_mem_channel);
        }
 
        if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes)
+               dc->dml2_options.bbox_overrides.dram_chanel_width_bytes =
                dcn3_21_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes;
 
        /* DML DSC delay factor workaround */
@@ -690,6 +700,10 @@ void dcn321_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_p
        /* Override dispclk_dppclk_vco_speed_mhz from Clk Mgr */
        dcn3_21_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
        dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+       dc->dml2_options.bbox_overrides.disp_pll_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+       dc->dml2_options.bbox_overrides.xtalclk_mhz = dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency / 1000.0;
+       dc->dml2_options.bbox_overrides.dchub_refclk_mhz = dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000.0;
+       dc->dml2_options.bbox_overrides.dprefclk_mhz = dc->clk_mgr->dprefclk_khz / 1000.0;
 
        /* Overrides Clock levelsfrom CLK Mgr table entries as reported by PM FW */
        if (dc->debug.use_legacy_soc_bb_mechanism) {
@@ -836,5 +850,72 @@ void dcn321_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_p
        dml_init_instance(&dc->dml, &dcn3_21_soc, &dcn3_21_ip, DML_PROJECT_DCN32);
        if (dc->current_state)
                dml_init_instance(&dc->current_state->bw_ctx.dml, &dcn3_21_soc, &dcn3_21_ip, DML_PROJECT_DCN32);
+
+       if (dc->clk_mgr->bw_params->clk_table.num_entries > 1) {
+               unsigned int i = 0;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_states = dc->clk_mgr->bw_params->clk_table.num_entries;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dcfclk_levels =
+                       dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dcfclk_levels;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_fclk_levels =
+                       dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_fclk_levels;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_memclk_levels =
+                       dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_memclk_levels;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_socclk_levels =
+                       dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_socclk_levels;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dtbclk_levels =
+                       dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dtbclk_levels;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dispclk_levels =
+                       dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dispclk_levels;
+
+               dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dppclk_levels =
+                       dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dppclk_levels;
+
+
+               for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dcfclk_levels; i++) {
+                       if (dc->clk_mgr->bw_params->clk_table.entries[i].dcfclk_mhz)
+                               dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dcfclk_mhz =
+                                       dc->clk_mgr->bw_params->clk_table.entries[i].dcfclk_mhz;
+               }
+
+               for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_fclk_levels; i++) {
+                       if (dc->clk_mgr->bw_params->clk_table.entries[i].fclk_mhz)
+                               dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].fclk_mhz =
+                                       dc->clk_mgr->bw_params->clk_table.entries[i].fclk_mhz;
+               }
+
+               for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_memclk_levels; i++) {
+                       if (dc->clk_mgr->bw_params->clk_table.entries[i].memclk_mhz)
+                               dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].memclk_mhz =
+                                       dc->clk_mgr->bw_params->clk_table.entries[i].memclk_mhz;
+               }
+
+               for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_socclk_levels; i++) {
+                       if (dc->clk_mgr->bw_params->clk_table.entries[i].socclk_mhz)
+                               dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].socclk_mhz =
+                                       dc->clk_mgr->bw_params->clk_table.entries[i].socclk_mhz;
+               }
+
+               for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dtbclk_levels; i++) {
+                       if (dc->clk_mgr->bw_params->clk_table.entries[i].dtbclk_mhz)
+                               dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dtbclk_mhz =
+                                       dc->clk_mgr->bw_params->clk_table.entries[i].dtbclk_mhz;
+               }
+
+               for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dispclk_levels; i++) {
+                       if (dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz) {
+                               dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dispclk_mhz =
+                                       dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz;
+                               dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dppclk_mhz =
+                                       dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz;
+                       }
+               }
+       }
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/Makefile b/drivers/gpu/drm/amd/display/dc/dml2/Makefile
new file mode 100644 (file)
index 0000000..f35ed8d
--- /dev/null
@@ -0,0 +1,91 @@
+# SPDX-License-Identifier: MIT */
+#
+# Copyright 2023 Advanced Micro Devices, Inc.
+#
+# Permission is hereby granted, free of charge, to any person obtaining a
+# copy of this software and associated documentation files (the "Software"),
+# to deal in the Software without restriction, including without limitation
+# the rights to use, copy, modify, merge, publish, distribute, sublicense,
+# and/or sell copies of the Software, and to permit persons to whom the
+# Software is furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in
+# all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+# THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+# OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+# ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+# OTHER DEALINGS IN THE SOFTWARE.
+#
+# makefile for dml2
+
+ifdef CONFIG_X86
+dml2_ccflags-$(CONFIG_CC_IS_GCC) := -mhard-float
+dml2_ccflags := $(dml2_ccflags-y) -msse
+endif
+
+ifdef CONFIG_PPC64
+dml2_ccflags := -mhard-float -maltivec
+endif
+
+ifdef CONFIG_ARM64
+dml2_rcflags := -mgeneral-regs-only
+endif
+
+ifdef CONFIG_LOONGARCH
+dml2_ccflags := -mfpu=64
+dml2_rcflags := -msoft-float
+endif
+
+ifdef CONFIG_CC_IS_GCC
+ifeq ($(call cc-ifversion, -lt, 0701, y), y)
+IS_OLD_GCC = 1
+endif
+endif
+
+ifdef CONFIG_X86
+ifdef IS_OLD_GCC
+# Stack alignment mismatch, proceed with caution.
+# GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
+# (8B stack alignment).
+dml2_ccflags += -mpreferred-stack-boundary=4
+else
+dml2_ccflags += -msse2
+endif
+endif
+
+ifneq ($(CONFIG_FRAME_WARN),0)
+frame_warn_flag := -Wframe-larger-than=2048
+endif
+
+CFLAGS_$(AMDDALPATH)/dc/dml2/display_mode_core.o := $(dml2_ccflags) -Wframe-larger-than=2048
+CFLAGS_$(AMDDALPATH)/dc/dml2/display_mode_util.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2/dml2_wrapper.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2/dml2_utils.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2/dml2_policy.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2/dml2_translation_helper.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2/dml2_mall_phantom.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2/dml_display_rq_dlg_calc.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2/dml2_dc_resource_mgmt.o := $(dml2_ccflags)
+
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/display_mode_core.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/display_mode_util.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml2_wrapper.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml2_utils.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml2_policy.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml2_translation_helper.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml2_mall_phantom.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml_display_rq_dlg_calc.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml2_dc_resource_mgmt.o := $(dml2_rcflags)
+
+DML2 = display_mode_core.o display_mode_util.o dml2_wrapper.o \
+               dml2_utils.o dml2_policy.o dml2_translation_helper.o dml2_dc_resource_mgmt.o dml2_mall_phantom.o \
+               dml_display_rq_dlg_calc.o
+
+AMD_DAL_DML2 = $(addprefix $(AMDDALPATH)/dc/dml2/,$(DML2))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DML2)
+
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/cmntypes.h b/drivers/gpu/drm/amd/display/dc/dml2/cmntypes.h
new file mode 100644 (file)
index 0000000..5450aa5
--- /dev/null
@@ -0,0 +1,92 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __CMNTYPES_H__
+#define __CMNTYPES_H__
+
+#ifdef __GNUC__
+#if __GNUC__ == 4 && __GNUC_MINOR__ > 7
+typedef unsigned int uint;
+#endif
+#endif
+
+typedef signed char int8, *pint8;
+typedef signed short int16, *pint16;
+typedef signed int int32, *pint32;
+typedef signed int64, *pint64;
+
+typedef unsigned char uint8, *puint8;
+typedef unsigned short uint16, *puint16;
+typedef unsigned int uint32, *puint32;
+typedef unsigned uint64, *puint64;
+
+typedef unsigned long int ulong;
+typedef unsigned char uchar;
+typedef unsigned int uint;
+
+typedef void *pvoid;
+typedef char *pchar;
+typedef const void *const_pvoid;
+typedef const char *const_pchar;
+
+typedef struct rgba_struct {
+    uint8 a;
+    uint8 r;
+    uint8 g;
+    uint8 b;
+} rgba_t;
+
+typedef struct {
+    uint8 blue;
+    uint8 green;
+    uint8 red;
+    uint8 alpha;
+} gen_color_t;
+
+typedef union {
+       uint32          val;
+       gen_color_t f;
+} gen_color_u;
+
+//
+// Types to make it easy to get or set the bits of a float/double.
+// Avoids automatic casting from int to float and back.
+//
+#if 0
+typedef union {
+       uint32 i;
+       float f;
+} uintfloat32;
+
+typedef union {
+       uint64 i;
+       double f;
+} uintfloat64;
+
+#ifndef UNREFERENCED_PARAMETER
+#define UNREFERENCED_PARAMETER(x) x = x
+#endif
+#endif
+
+#endif  //__CMNTYPES_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c
new file mode 100644 (file)
index 0000000..e65e86c
--- /dev/null
@@ -0,0 +1,10275 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "display_mode_core.h"
+#include "display_mode_util.h"
+#include "display_mode_lib_defines.h"
+
+#include "dml_assert.h"
+
+#define DML2_MAX_FMT_420_BUFFER_WIDTH 4096
+// ---------------------------
+//  Declaration Begins
+// ---------------------------
+static void CalculateBytePerPixelAndBlockSizes(
+       enum dml_source_format_class SourcePixelFormat,
+       enum dml_swizzle_mode SurfaceTiling,
+       // Output
+       dml_uint_t *BytePerPixelY,
+       dml_uint_t *BytePerPixelC,
+       dml_float_t *BytePerPixelDETY,
+       dml_float_t *BytePerPixelDETC,
+       dml_uint_t *BlockHeight256BytesY,
+       dml_uint_t *BlockHeight256BytesC,
+       dml_uint_t *BlockWidth256BytesY,
+       dml_uint_t *BlockWidth256BytesC,
+       dml_uint_t *MacroTileHeightY,
+       dml_uint_t *MacroTileHeightC,
+       dml_uint_t *MacroTileWidthY,
+       dml_uint_t *MacroTileWidthC);
+
+static dml_float_t CalculateWriteBackDISPCLK(
+       enum dml_source_format_class WritebackPixelFormat,
+       dml_float_t PixelClock,
+       dml_float_t WritebackHRatio,
+       dml_float_t WritebackVRatio,
+       dml_uint_t WritebackHTaps,
+       dml_uint_t WritebackVTaps,
+       dml_uint_t WritebackSourceWidth,
+       dml_uint_t WritebackDestinationWidth,
+       dml_uint_t HTotal,
+       dml_uint_t WritebackLineBufferSize,
+       dml_float_t DISPCLKDPPCLKVCOSpeed);
+
+static void CalculateVMRowAndSwath(
+       struct display_mode_lib_scratch_st *s,
+       struct CalculateVMRowAndSwath_params_st *p);
+
+static void CalculateOutputLink(
+       dml_float_t PHYCLKPerState,
+       dml_float_t PHYCLKD18PerState,
+       dml_float_t PHYCLKD32PerState,
+       dml_float_t Downspreading,
+       dml_bool_t IsMainSurfaceUsingTheIndicatedTiming,
+       enum dml_output_encoder_class Output,
+       enum dml_output_format_class OutputFormat,
+       dml_uint_t HTotal,
+       dml_uint_t HActive,
+       dml_float_t PixelClockBackEnd,
+       dml_float_t ForcedOutputLinkBPP,
+       dml_uint_t DSCInputBitPerComponent,
+       dml_uint_t NumberOfDSCSlices,
+       dml_float_t AudioSampleRate,
+       dml_uint_t AudioSampleLayout,
+       enum dml_odm_mode ODMModeNoDSC,
+       enum dml_odm_mode ODMModeDSC,
+       enum dml_dsc_enable DSCEnable,
+       dml_uint_t OutputLinkDPLanes,
+       enum dml_output_link_dp_rate OutputLinkDPRate,
+
+       // Output
+       dml_bool_t *RequiresDSC,
+       dml_bool_t *RequiresFEC,
+       dml_float_t *OutBpp,
+       enum dml_output_type_and_rate__type *OutputType,
+       enum dml_output_type_and_rate__rate *OutputRate,
+       dml_uint_t *RequiredSlots);
+
+static void CalculateODMMode(
+       dml_uint_t MaximumPixelsPerLinePerDSCUnit,
+       dml_uint_t HActive,
+       enum dml_output_encoder_class Output,
+       enum dml_output_format_class OutputFormat,
+       enum dml_odm_use_policy ODMUse,
+       dml_float_t StateDispclk,
+       dml_float_t MaxDispclk,
+       dml_bool_t DSCEnable,
+       dml_uint_t TotalNumberOfActiveDPP,
+       dml_uint_t MaxNumDPP,
+       dml_float_t PixelClock,
+       dml_float_t DISPCLKDPPCLKDSCCLKDownSpreading,
+       dml_float_t DISPCLKRampingMargin,
+       dml_float_t DISPCLKDPPCLKVCOSpeed,
+
+       // Output
+       dml_bool_t *TotalAvailablePipesSupport,
+       dml_uint_t *NumberOfDPP,
+       enum dml_odm_mode *ODMMode,
+       dml_float_t *RequiredDISPCLKPerSurface);
+
+static dml_float_t CalculateRequiredDispclk(
+       enum dml_odm_mode ODMMode,
+       dml_float_t PixelClock,
+       dml_float_t DISPCLKDPPCLKDSCCLKDownSpreading,
+       dml_float_t DISPCLKRampingMargin,
+       dml_float_t DISPCLKDPPCLKVCOSpeed,
+       dml_float_t MaxDispclkSingle);
+
+static void CalculateSinglePipeDPPCLKAndSCLThroughput(
+       dml_float_t HRatio,
+       dml_float_t HRatioChroma,
+       dml_float_t VRatio,
+       dml_float_t VRatioChroma,
+       dml_float_t MaxDCHUBToPSCLThroughput,
+       dml_float_t MaxPSCLToLBThroughput,
+       dml_float_t PixelClock,
+       enum dml_source_format_class SourcePixelFormat,
+       dml_uint_t HTaps,
+       dml_uint_t HTapsChroma,
+       dml_uint_t VTaps,
+       dml_uint_t VTapsChroma,
+
+       // Output
+       dml_float_t *PSCL_THROUGHPUT,
+       dml_float_t *PSCL_THROUGHPUT_CHROMA,
+       dml_float_t *DPPCLKUsingSingleDPP);
+
+static void CalculateDPPCLK(
+       dml_uint_t NumberOfActiveSurfaces,
+       dml_float_t DISPCLKDPPCLKDSCCLKDownSpreading,
+       dml_float_t DISPCLKDPPCLKVCOSpeed,
+       dml_float_t DPPCLKUsingSingleDPP[],
+       dml_uint_t DPPPerSurface[],
+
+       // Output
+       dml_float_t *GlobalDPPCLK,
+       dml_float_t Dppclk[]);
+
+static void CalculateMALLUseForStaticScreen(
+       dml_uint_t NumberOfActiveSurfaces,
+       dml_uint_t MALLAllocatedForDCNFinal,
+       enum dml_use_mall_for_static_screen_mode *UseMALLForStaticScreen,
+       dml_uint_t SurfaceSizeInMALL[],
+       dml_bool_t one_row_per_frame_fits_in_buffer[],
+
+       // Output
+       dml_bool_t UsesMALLForStaticScreen[]);
+
+static dml_uint_t dscceComputeDelay(
+       dml_uint_t bpc,
+       dml_float_t BPP,
+       dml_uint_t sliceWidth,
+       dml_uint_t numSlices,
+       enum dml_output_format_class pixelFormat,
+       enum dml_output_encoder_class Output);
+
+static dml_uint_t dscComputeDelay(enum dml_output_format_class pixelFormat,
+       enum dml_output_encoder_class Output);
+
+static dml_bool_t CalculatePrefetchSchedule(struct display_mode_lib_scratch_st *scratch,
+       struct CalculatePrefetchSchedule_params_st *p);
+
+static dml_float_t RoundToDFSGranularity(dml_float_t Clock, dml_bool_t round_up, dml_float_t VCOSpeed);
+
+static void CalculateDCCConfiguration(
+       dml_bool_t DCCEnabled,
+       dml_bool_t DCCProgrammingAssumesScanDirectionUnknown,
+       enum dml_source_format_class SourcePixelFormat,
+       dml_uint_t SurfaceWidthLuma,
+       dml_uint_t SurfaceWidthChroma,
+       dml_uint_t SurfaceHeightLuma,
+       dml_uint_t SurfaceHeightChroma,
+       dml_uint_t nomDETInKByte,
+       dml_uint_t RequestHeight256ByteLuma,
+       dml_uint_t RequestHeight256ByteChroma,
+       enum dml_swizzle_mode TilingFormat,
+       dml_uint_t BytePerPixelY,
+       dml_uint_t BytePerPixelC,
+       dml_float_t BytePerPixelDETY,
+       dml_float_t BytePerPixelDETC,
+       enum dml_rotation_angle SourceScan,
+       // Output
+       dml_uint_t *MaxUncompressedBlockLuma,
+       dml_uint_t *MaxUncompressedBlockChroma,
+       dml_uint_t *MaxCompressedBlockLuma,
+       dml_uint_t *MaxCompressedBlockChroma,
+       dml_uint_t *IndependentBlockLuma,
+       dml_uint_t *IndependentBlockChroma);
+
+static dml_uint_t CalculatePrefetchSourceLines(
+       dml_float_t VRatio,
+       dml_uint_t VTaps,
+       dml_bool_t Interlace,
+       dml_bool_t ProgressiveToInterlaceUnitInOPP,
+       dml_uint_t SwathHeight,
+       enum dml_rotation_angle SourceScan,
+       dml_bool_t ViewportStationary,
+       dml_uint_t SwathWidth,
+       dml_uint_t ViewportHeight,
+       dml_uint_t ViewportXStart,
+       dml_uint_t ViewportYStart,
+
+       // Output
+       dml_uint_t *VInitPreFill,
+       dml_uint_t *MaxNumSwath);
+
+static dml_uint_t CalculateVMAndRowBytes(
+       dml_bool_t ViewportStationary,
+       dml_bool_t DCCEnable,
+       dml_uint_t NumberOfDPPs,
+       dml_uint_t BlockHeight256Bytes,
+       dml_uint_t BlockWidth256Bytes,
+       enum dml_source_format_class SourcePixelFormat,
+       dml_uint_t SurfaceTiling,
+       dml_uint_t BytePerPixel,
+       enum dml_rotation_angle SourceScan,
+       dml_uint_t SwathWidth,
+       dml_uint_t ViewportHeight,
+       dml_uint_t ViewportXStart,
+       dml_uint_t ViewportYStart,
+       dml_bool_t GPUVMEnable,
+       dml_uint_t GPUVMMaxPageTableLevels,
+       dml_uint_t GPUVMMinPageSizeKBytes,
+       dml_uint_t PTEBufferSizeInRequests,
+       dml_uint_t Pitch,
+       dml_uint_t DCCMetaPitch,
+       dml_uint_t MacroTileWidth,
+       dml_uint_t MacroTileHeight,
+
+       // Output
+       dml_uint_t *MetaRowByte,
+       dml_uint_t *PixelPTEBytesPerRow,
+       dml_uint_t *PixelPTEBytesPerRowStorage, // for PTE buffer size check
+       dml_uint_t *dpte_row_width_ub,
+       dml_uint_t *dpte_row_height,
+       dml_uint_t *dpte_row_height_linear,
+       dml_uint_t *PixelPTEBytesPerRow_one_row_per_frame,
+       dml_uint_t *dpte_row_width_ub_one_row_per_frame,
+       dml_uint_t *dpte_row_height_one_row_per_frame,
+       dml_uint_t *MetaRequestWidth,
+       dml_uint_t *MetaRequestHeight,
+       dml_uint_t *meta_row_width,
+       dml_uint_t *meta_row_height,
+       dml_uint_t *PixelPTEReqWidth,
+       dml_uint_t *PixelPTEReqHeight,
+       dml_uint_t *PTERequestSize,
+       dml_uint_t *DPDE0BytesFrame,
+       dml_uint_t *MetaPTEBytesFrame);
+
+static dml_float_t CalculateTWait(
+       dml_uint_t PrefetchMode,
+       enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange,
+       dml_bool_t SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+       dml_bool_t DRRDisplay,
+       dml_float_t DRAMClockChangeLatency,
+       dml_float_t FCLKChangeLatency,
+       dml_float_t UrgentLatency,
+       dml_float_t SREnterPlusExitTime);
+
+static void CalculatePrefetchMode(
+       enum dml_prefetch_modes AllowForPStateChangeOrStutterInVBlank,
+       dml_uint_t *MinPrefetchMode,
+       dml_uint_t *MaxPrefetchMode);
+
+static void CalculateRowBandwidth(
+       dml_bool_t GPUVMEnable,
+       enum dml_source_format_class SourcePixelFormat,
+       dml_float_t VRatio,
+       dml_float_t VRatioChroma,
+       dml_bool_t DCCEnable,
+       dml_float_t LineTime,
+       dml_uint_t MetaRowByteLuma,
+       dml_uint_t MetaRowByteChroma,
+       dml_uint_t meta_row_height_luma,
+       dml_uint_t meta_row_height_chroma,
+       dml_uint_t PixelPTEBytesPerRowLuma,
+       dml_uint_t PixelPTEBytesPerRowChroma,
+       dml_uint_t dpte_row_height_luma,
+       dml_uint_t dpte_row_height_chroma,
+       // Output
+       dml_float_t *meta_row_bw,
+       dml_float_t *dpte_row_bw);
+
+static void CalculateFlipSchedule(
+       dml_float_t HostVMInefficiencyFactor,
+       dml_float_t UrgentExtraLatency,
+       dml_float_t UrgentLatency,
+       dml_uint_t GPUVMMaxPageTableLevels,
+       dml_bool_t HostVMEnable,
+       dml_uint_t HostVMMaxNonCachedPageTableLevels,
+       dml_bool_t GPUVMEnable,
+       dml_uint_t HostVMMinPageSize,
+       dml_float_t PDEAndMetaPTEBytesPerFrame,
+       dml_float_t MetaRowBytes,
+       dml_float_t DPTEBytesPerRow,
+       dml_float_t BandwidthAvailableForImmediateFlip,
+       dml_uint_t TotImmediateFlipBytes,
+       enum dml_source_format_class SourcePixelFormat,
+       dml_float_t LineTime,
+       dml_float_t VRatio,
+       dml_float_t VRatioChroma,
+       dml_float_t Tno_bw,
+       dml_bool_t DCCEnable,
+       dml_uint_t dpte_row_height,
+       dml_uint_t meta_row_height,
+       dml_uint_t dpte_row_height_chroma,
+       dml_uint_t meta_row_height_chroma,
+       dml_bool_t use_one_row_for_frame_flip,
+
+       // Output
+       dml_float_t *DestinationLinesToRequestVMInImmediateFlip,
+       dml_float_t *DestinationLinesToRequestRowInImmediateFlip,
+       dml_float_t *final_flip_bw,
+       dml_bool_t *ImmediateFlipSupportedForPipe);
+
+static dml_float_t CalculateWriteBackDelay(
+       enum dml_source_format_class WritebackPixelFormat,
+       dml_float_t WritebackHRatio,
+       dml_float_t WritebackVRatio,
+       dml_uint_t WritebackVTaps,
+       dml_uint_t WritebackDestinationWidth,
+       dml_uint_t WritebackDestinationHeight,
+       dml_uint_t WritebackSourceHeight,
+       dml_uint_t HTotal);
+
+static void CalculateVUpdateAndDynamicMetadataParameters(
+       dml_uint_t MaxInterDCNTileRepeaters,
+       dml_float_t Dppclk,
+       dml_float_t DISPCLK,
+       dml_float_t DCFClkDeepSleep,
+       dml_float_t PixelClock,
+       dml_uint_t HTotal,
+       dml_uint_t VBlank,
+       dml_uint_t DynamicMetadataTransmittedBytes,
+       dml_uint_t DynamicMetadataLinesBeforeActiveRequired,
+       dml_uint_t InterlaceEnable,
+       dml_bool_t ProgressiveToInterlaceUnitInOPP,
+       dml_float_t *TSetup,
+       dml_float_t *Tdmbf,
+       dml_float_t *Tdmec,
+       dml_float_t *Tdmsks,
+       dml_uint_t *VUpdateOffsetPix,
+       dml_uint_t *VUpdateWidthPix,
+       dml_uint_t *VReadyOffsetPix);
+
+static void PixelClockAdjustmentForProgressiveToInterlaceUnit(struct dml_display_cfg_st *display_cfg, dml_bool_t ptoi_supported);
+
+static dml_float_t TruncToValidBPP(
+       dml_float_t LinkBitRate,
+       dml_uint_t Lanes,
+       dml_uint_t HTotal,
+       dml_uint_t HActive,
+       dml_float_t PixelClock,
+       dml_float_t DesiredBPP,
+       dml_bool_t DSCEnable,
+       enum dml_output_encoder_class Output,
+       enum dml_output_format_class Format,
+       dml_uint_t DSCInputBitPerComponent,
+       dml_uint_t DSCSlices,
+       dml_uint_t AudioRate,
+       dml_uint_t AudioLayout,
+       enum dml_odm_mode ODMModeNoDSC,
+       enum dml_odm_mode ODMModeDSC,
+       // Output
+       dml_uint_t *RequiredSlotsSingle);
+
+static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
+       struct display_mode_lib_scratch_st *s,
+       struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params_st *p);
+
+static void CalculateDCFCLKDeepSleep(
+       dml_uint_t NumberOfActiveSurfaces,
+       dml_uint_t BytePerPixelY[],
+       dml_uint_t BytePerPixelC[],
+       dml_float_t VRatio[],
+       dml_float_t VRatioChroma[],
+       dml_uint_t SwathWidthY[],
+       dml_uint_t SwathWidthC[],
+       dml_uint_t DPPPerSurface[],
+       dml_float_t HRatio[],
+       dml_float_t HRatioChroma[],
+       dml_float_t PixelClock[],
+       dml_float_t PSCL_THROUGHPUT[],
+       dml_float_t PSCL_THROUGHPUT_CHROMA[],
+       dml_float_t Dppclk[],
+       dml_float_t ReadBandwidthLuma[],
+       dml_float_t ReadBandwidthChroma[],
+       dml_uint_t ReturnBusWidth,
+
+       // Output
+       dml_float_t *DCFCLKDeepSleep);
+
+static void CalculateUrgentBurstFactor(
+       enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange,
+       dml_uint_t swath_width_luma_ub,
+       dml_uint_t swath_width_chroma_ub,
+       dml_uint_t SwathHeightY,
+       dml_uint_t SwathHeightC,
+       dml_float_t LineTime,
+       dml_float_t UrgentLatency,
+       dml_float_t CursorBufferSize,
+       dml_uint_t CursorWidth,
+       dml_uint_t CursorBPP,
+       dml_float_t VRatio,
+       dml_float_t VRatioC,
+       dml_float_t BytePerPixelInDETY,
+       dml_float_t BytePerPixelInDETC,
+       dml_uint_t DETBufferSizeY,
+       dml_uint_t DETBufferSizeC,
+       // Output
+       dml_float_t *UrgentBurstFactorCursor,
+       dml_float_t *UrgentBurstFactorLuma,
+       dml_float_t *UrgentBurstFactorChroma,
+       dml_bool_t *NotEnoughUrgentLatencyHiding);
+
+static dml_float_t RequiredDTBCLK(
+       dml_bool_t DSCEnable,
+       dml_float_t PixelClock,
+       enum dml_output_format_class OutputFormat,
+       dml_float_t OutputBpp,
+       dml_uint_t DSCSlices,
+       dml_uint_t HTotal,
+       dml_uint_t HActive,
+       dml_uint_t AudioRate,
+       dml_uint_t AudioLayoutSingle);
+
+static void UseMinimumDCFCLK(
+       struct display_mode_lib_scratch_st *scratch,
+       struct UseMinimumDCFCLK_params_st *p);
+
+static void CalculatePixelDeliveryTimes(
+       dml_uint_t NumberOfActiveSurfaces,
+       dml_float_t VRatio[],
+       dml_float_t VRatioChroma[],
+       dml_float_t VRatioPrefetchY[],
+       dml_float_t VRatioPrefetchC[],
+       dml_uint_t swath_width_luma_ub[],
+       dml_uint_t swath_width_chroma_ub[],
+       dml_uint_t DPPPerSurface[],
+       dml_float_t HRatio[],
+       dml_float_t HRatioChroma[],
+       dml_float_t PixelClock[],
+       dml_float_t PSCL_THROUGHPUT[],
+       dml_float_t PSCL_THROUGHPUT_CHROMA[],
+       dml_float_t Dppclk[],
+       dml_uint_t BytePerPixelC[],
+       enum dml_rotation_angle SourceScan[],
+       dml_uint_t NumberOfCursors[],
+       dml_uint_t CursorWidth[],
+       dml_uint_t CursorBPP[],
+       dml_uint_t BlockWidth256BytesY[],
+       dml_uint_t BlockHeight256BytesY[],
+       dml_uint_t BlockWidth256BytesC[],
+       dml_uint_t BlockHeight256BytesC[],
+
+       // Output
+       dml_float_t DisplayPipeLineDeliveryTimeLuma[],
+       dml_float_t DisplayPipeLineDeliveryTimeChroma[],
+       dml_float_t DisplayPipeLineDeliveryTimeLumaPrefetch[],
+       dml_float_t DisplayPipeLineDeliveryTimeChromaPrefetch[],
+       dml_float_t DisplayPipeRequestDeliveryTimeLuma[],
+       dml_float_t DisplayPipeRequestDeliveryTimeChroma[],
+       dml_float_t DisplayPipeRequestDeliveryTimeLumaPrefetch[],
+       dml_float_t DisplayPipeRequestDeliveryTimeChromaPrefetch[],
+       dml_float_t CursorRequestDeliveryTime[],
+       dml_float_t CursorRequestDeliveryTimePrefetch[]);
+
+static void CalculateMetaAndPTETimes(
+       dml_bool_t use_one_row_for_frame[],
+       dml_uint_t NumberOfActiveSurfaces,
+       dml_bool_t GPUVMEnable,
+       dml_uint_t MetaChunkSize,
+       dml_uint_t MinMetaChunkSizeBytes,
+       dml_uint_t HTotal[],
+       dml_float_t VRatio[],
+       dml_float_t VRatioChroma[],
+       dml_float_t DestinationLinesToRequestRowInVBlank[],
+       dml_float_t DestinationLinesToRequestRowInImmediateFlip[],
+       dml_bool_t DCCEnable[],
+       dml_float_t PixelClock[],
+       dml_uint_t BytePerPixelY[],
+       dml_uint_t BytePerPixelC[],
+       enum dml_rotation_angle SourceScan[],
+       dml_uint_t dpte_row_height[],
+       dml_uint_t dpte_row_height_chroma[],
+       dml_uint_t meta_row_width[],
+       dml_uint_t meta_row_width_chroma[],
+       dml_uint_t meta_row_height[],
+       dml_uint_t meta_row_height_chroma[],
+       dml_uint_t meta_req_width[],
+       dml_uint_t meta_req_width_chroma[],
+       dml_uint_t meta_req_height[],
+       dml_uint_t meta_req_height_chroma[],
+       dml_uint_t dpte_group_bytes[],
+       dml_uint_t PTERequestSizeY[],
+       dml_uint_t PTERequestSizeC[],
+       dml_uint_t PixelPTEReqWidthY[],
+       dml_uint_t PixelPTEReqHeightY[],
+       dml_uint_t PixelPTEReqWidthC[],
+       dml_uint_t PixelPTEReqHeightC[],
+       dml_uint_t dpte_row_width_luma_ub[],
+       dml_uint_t dpte_row_width_chroma_ub[],
+
+       // Output
+       dml_float_t DST_Y_PER_PTE_ROW_NOM_L[],
+       dml_float_t DST_Y_PER_PTE_ROW_NOM_C[],
+       dml_float_t DST_Y_PER_META_ROW_NOM_L[],
+       dml_float_t DST_Y_PER_META_ROW_NOM_C[],
+       dml_float_t TimePerMetaChunkNominal[],
+       dml_float_t TimePerChromaMetaChunkNominal[],
+       dml_float_t TimePerMetaChunkVBlank[],
+       dml_float_t TimePerChromaMetaChunkVBlank[],
+       dml_float_t TimePerMetaChunkFlip[],
+       dml_float_t TimePerChromaMetaChunkFlip[],
+       dml_float_t time_per_pte_group_nom_luma[],
+       dml_float_t time_per_pte_group_vblank_luma[],
+       dml_float_t time_per_pte_group_flip_luma[],
+       dml_float_t time_per_pte_group_nom_chroma[],
+       dml_float_t time_per_pte_group_vblank_chroma[],
+       dml_float_t time_per_pte_group_flip_chroma[]);
+
+static void CalculateVMGroupAndRequestTimes(
+       dml_uint_t NumberOfActiveSurfaces,
+       dml_bool_t GPUVMEnable,
+       dml_uint_t GPUVMMaxPageTableLevels,
+       dml_uint_t HTotal[],
+       dml_uint_t BytePerPixelC[],
+       dml_float_t DestinationLinesToRequestVMInVBlank[],
+       dml_float_t DestinationLinesToRequestVMInImmediateFlip[],
+       dml_bool_t DCCEnable[],
+       dml_float_t PixelClock[],
+       dml_uint_t dpte_row_width_luma_ub[],
+       dml_uint_t dpte_row_width_chroma_ub[],
+       dml_uint_t vm_group_bytes[],
+       dml_uint_t dpde0_bytes_per_frame_ub_l[],
+       dml_uint_t dpde0_bytes_per_frame_ub_c[],
+       dml_uint_t meta_pte_bytes_per_frame_ub_l[],
+       dml_uint_t meta_pte_bytes_per_frame_ub_c[],
+
+       // Output
+       dml_float_t TimePerVMGroupVBlank[],
+       dml_float_t TimePerVMGroupFlip[],
+       dml_float_t TimePerVMRequestVBlank[],
+       dml_float_t TimePerVMRequestFlip[]);
+
+static void CalculateStutterEfficiency(
+       struct display_mode_lib_scratch_st *scratch,
+       struct CalculateStutterEfficiency_params_st *p);
+
+static void CalculateSwathAndDETConfiguration(
+       struct display_mode_lib_scratch_st *scratch,
+       struct CalculateSwathAndDETConfiguration_params_st *p);
+
+static void CalculateSwathWidth(
+       dml_bool_t ForceSingleDPP,
+       dml_uint_t NumberOfActiveSurfaces,
+       enum dml_source_format_class SourcePixelFormat[],
+       enum dml_rotation_angle SourceScan[],
+       dml_bool_t ViewportStationary[],
+       dml_uint_t ViewportWidth[],
+       dml_uint_t ViewportHeight[],
+       dml_uint_t ViewportXStart[],
+       dml_uint_t ViewportYStart[],
+       dml_uint_t ViewportXStartC[],
+       dml_uint_t ViewportYStartC[],
+       dml_uint_t SurfaceWidthY[],
+       dml_uint_t SurfaceWidthC[],
+       dml_uint_t SurfaceHeightY[],
+       dml_uint_t SurfaceHeightC[],
+       enum dml_odm_mode ODMMode[],
+       dml_uint_t BytePerPixY[],
+       dml_uint_t BytePerPixC[],
+       dml_uint_t Read256BytesBlockHeightY[],
+       dml_uint_t Read256BytesBlockHeightC[],
+       dml_uint_t Read256BytesBlockWidthY[],
+       dml_uint_t Read256BytesBlockWidthC[],
+       dml_uint_t BlendingAndTiming[],
+       dml_uint_t HActive[],
+       dml_float_t HRatio[],
+       dml_uint_t DPPPerSurface[],
+
+       // Output
+       dml_uint_t SwathWidthSingleDPPY[],
+       dml_uint_t SwathWidthSingleDPPC[],
+       dml_uint_t SwathWidthY[],
+       dml_uint_t SwathWidthC[],
+       dml_uint_t MaximumSwathHeightY[],
+       dml_uint_t MaximumSwathHeightC[],
+       dml_uint_t swath_width_luma_ub[],
+       dml_uint_t swath_width_chroma_ub[]);
+
+static dml_float_t CalculateExtraLatency(
+       dml_uint_t RoundTripPingLatencyCycles,
+       dml_uint_t ReorderingBytes,
+       dml_float_t DCFCLK,
+       dml_uint_t TotalNumberOfActiveDPP,
+       dml_uint_t PixelChunkSizeInKByte,
+       dml_uint_t TotalNumberOfDCCActiveDPP,
+       dml_uint_t MetaChunkSize,
+       dml_float_t ReturnBW,
+       dml_bool_t GPUVMEnable,
+       dml_bool_t HostVMEnable,
+       dml_uint_t NumberOfActiveSurfaces,
+       dml_uint_t NumberOfDPP[],
+       dml_uint_t dpte_group_bytes[],
+       dml_float_t HostVMInefficiencyFactor,
+       dml_uint_t HostVMMinPageSize,
+       dml_uint_t HostVMMaxNonCachedPageTableLevels);
+
+static dml_uint_t CalculateExtraLatencyBytes(
+       dml_uint_t ReorderingBytes,
+       dml_uint_t TotalNumberOfActiveDPP,
+       dml_uint_t PixelChunkSizeInKByte,
+       dml_uint_t TotalNumberOfDCCActiveDPP,
+       dml_uint_t MetaChunkSize,
+       dml_bool_t GPUVMEnable,
+       dml_bool_t HostVMEnable,
+       dml_uint_t NumberOfActiveSurfaces,
+       dml_uint_t NumberOfDPP[],
+       dml_uint_t dpte_group_bytes[],
+       dml_float_t HostVMInefficiencyFactor,
+       dml_uint_t HostVMMinPageSize,
+       dml_uint_t HostVMMaxNonCachedPageTableLevels);
+
+static dml_float_t CalculateUrgentLatency(
+       dml_float_t UrgentLatencyPixelDataOnly,
+       dml_float_t UrgentLatencyPixelMixedWithVMData,
+       dml_float_t UrgentLatencyVMDataOnly,
+       dml_bool_t DoUrgentLatencyAdjustment,
+       dml_float_t UrgentLatencyAdjustmentFabricClockComponent,
+       dml_float_t UrgentLatencyAdjustmentFabricClockReference,
+       dml_float_t FabricClockSingle);
+
+static dml_bool_t UnboundedRequest(
+       enum dml_unbounded_requesting_policy UseUnboundedRequestingFinal,
+       dml_uint_t TotalNumberOfActiveDPP,
+       dml_bool_t NoChromaOrLinear,
+       enum dml_output_encoder_class Output);
+
+static void CalculateSurfaceSizeInMall(
+       dml_uint_t NumberOfActiveSurfaces,
+       dml_uint_t MALLAllocatedForDCN,
+       enum dml_use_mall_for_static_screen_mode UseMALLForStaticScreen[],
+       dml_bool_t DCCEnable[],
+       dml_bool_t ViewportStationary[],
+       dml_uint_t ViewportXStartY[],
+       dml_uint_t ViewportYStartY[],
+       dml_uint_t ViewportXStartC[],
+       dml_uint_t ViewportYStartC[],
+       dml_uint_t ViewportWidthY[],
+       dml_uint_t ViewportHeightY[],
+       dml_uint_t BytesPerPixelY[],
+       dml_uint_t ViewportWidthC[],
+       dml_uint_t ViewportHeightC[],
+       dml_uint_t BytesPerPixelC[],
+       dml_uint_t SurfaceWidthY[],
+       dml_uint_t SurfaceWidthC[],
+       dml_uint_t SurfaceHeightY[],
+       dml_uint_t SurfaceHeightC[],
+       dml_uint_t Read256BytesBlockWidthY[],
+       dml_uint_t Read256BytesBlockWidthC[],
+       dml_uint_t Read256BytesBlockHeightY[],
+       dml_uint_t Read256BytesBlockHeightC[],
+       dml_uint_t ReadBlockWidthY[],
+       dml_uint_t ReadBlockWidthC[],
+       dml_uint_t ReadBlockHeightY[],
+       dml_uint_t ReadBlockHeightC[],
+
+       // Output
+       dml_uint_t SurfaceSizeInMALL[],
+       dml_bool_t *ExceededMALLSize);
+
+static void CalculateDETBufferSize(
+       dml_uint_t DETSizeOverride[],
+       enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+       dml_bool_t ForceSingleDPP,
+       dml_uint_t NumberOfActiveSurfaces,
+       dml_bool_t UnboundedRequestEnabled,
+       dml_uint_t nomDETInKByte,
+       dml_uint_t MaxTotalDETInKByte,
+       dml_uint_t ConfigReturnBufferSizeInKByte,
+       dml_uint_t MinCompressedBufferSizeInKByte,
+       dml_uint_t ConfigReturnBufferSegmentSizeInkByte,
+       dml_uint_t CompressedBufferSegmentSizeInkByteFinal,
+       enum dml_source_format_class SourcePixelFormat[],
+       dml_float_t ReadBandwidthLuma[],
+       dml_float_t ReadBandwidthChroma[],
+       dml_uint_t RotesY[],
+       dml_uint_t RoundedUpMaxSwathSizeBytesC[],
+       dml_uint_t DPPPerSurface[],
+       // Output
+       dml_uint_t DETBufferSizeInKByte[],
+       dml_uint_t *CompressedBufferSizeInkByte);
+
+static void CalculateMaxDETAndMinCompressedBufferSize(
+       dml_uint_t ConfigReturnBufferSizeInKByte,
+       dml_uint_t ConfigReturnBufferSegmentSizeInKByte,
+       dml_uint_t ROBBufferSizeInKByte,
+       dml_uint_t MaxNumDPP,
+       dml_bool_t nomDETInKByteOverrideEnable,
+       dml_uint_t nomDETInKByteOverrideValue,
+
+       // Output
+       dml_uint_t *MaxTotalDETInKByte,
+       dml_uint_t *nomDETInKByte,
+       dml_uint_t *MinCompressedBufferSizeInKByte);
+
+static dml_uint_t DSCDelayRequirement(
+       dml_bool_t DSCEnabled,
+       enum dml_odm_mode ODMMode,
+       dml_uint_t DSCInputBitPerComponent,
+       dml_float_t OutputBpp,
+       dml_uint_t HActive,
+       dml_uint_t HTotal,
+       dml_uint_t NumberOfDSCSlices,
+       enum dml_output_format_class OutputFormat,
+       enum dml_output_encoder_class Output,
+       dml_float_t PixelClock,
+       dml_float_t PixelClockBackEnd);
+
+static dml_bool_t CalculateVActiveBandwithSupport(
+       dml_uint_t NumberOfActiveSurfaces,
+       dml_float_t ReturnBW,
+       dml_bool_t NotUrgentLatencyHiding[],
+       dml_float_t ReadBandwidthLuma[],
+       dml_float_t ReadBandwidthChroma[],
+       dml_float_t cursor_bw[],
+       dml_float_t meta_row_bandwidth[],
+       dml_float_t dpte_row_bandwidth[],
+       dml_uint_t NumberOfDPP[],
+       dml_float_t UrgentBurstFactorLuma[],
+       dml_float_t UrgentBurstFactorChroma[],
+       dml_float_t UrgentBurstFactorCursor[]);
+
+static void CalculatePrefetchBandwithSupport(
+       dml_uint_t NumberOfActiveSurfaces,
+       dml_float_t ReturnBW,
+       enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+       dml_bool_t NotUrgentLatencyHiding[],
+       dml_float_t ReadBandwidthLuma[],
+       dml_float_t ReadBandwidthChroma[],
+       dml_float_t PrefetchBandwidthLuma[],
+       dml_float_t PrefetchBandwidthChroma[],
+       dml_float_t cursor_bw[],
+       dml_float_t meta_row_bandwidth[],
+       dml_float_t dpte_row_bandwidth[],
+       dml_float_t cursor_bw_pre[],
+       dml_float_t prefetch_vmrow_bw[],
+       dml_uint_t NumberOfDPP[],
+       dml_float_t UrgentBurstFactorLuma[],
+       dml_float_t UrgentBurstFactorChroma[],
+       dml_float_t UrgentBurstFactorCursor[],
+       dml_float_t UrgentBurstFactorLumaPre[],
+       dml_float_t UrgentBurstFactorChromaPre[],
+       dml_float_t UrgentBurstFactorCursorPre[],
+
+       // Output
+       dml_float_t *PrefetchBandwidth,
+       dml_float_t *PrefetchBandwidthNotIncludingMALLPrefetch,
+       dml_float_t *FractionOfUrgentBandwidth,
+       dml_bool_t *PrefetchBandwidthSupport);
+
+static dml_float_t CalculateBandwidthAvailableForImmediateFlip(
+       dml_uint_t NumberOfActiveSurfaces,
+       dml_float_t ReturnBW,
+       dml_float_t ReadBandwidthLuma[],
+       dml_float_t ReadBandwidthChroma[],
+       dml_float_t PrefetchBandwidthLuma[],
+       dml_float_t PrefetchBandwidthChroma[],
+       dml_float_t cursor_bw[],
+       dml_float_t cursor_bw_pre[],
+       dml_uint_t NumberOfDPP[],
+       dml_float_t UrgentBurstFactorLuma[],
+       dml_float_t UrgentBurstFactorChroma[],
+       dml_float_t UrgentBurstFactorCursor[],
+       dml_float_t UrgentBurstFactorLumaPre[],
+       dml_float_t UrgentBurstFactorChromaPre[],
+       dml_float_t UrgentBurstFactorCursorPre[]);
+
+static void CalculateImmediateFlipBandwithSupport(
+       dml_uint_t NumberOfActiveSurfaces,
+       dml_float_t ReturnBW,
+       enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+       enum dml_immediate_flip_requirement ImmediateFlipRequirement[],
+       dml_float_t final_flip_bw[],
+       dml_float_t ReadBandwidthLuma[],
+       dml_float_t ReadBandwidthChroma[],
+       dml_float_t PrefetchBandwidthLuma[],
+       dml_float_t PrefetchBandwidthChroma[],
+       dml_float_t cursor_bw[],
+       dml_float_t meta_row_bandwidth[],
+       dml_float_t dpte_row_bandwidth[],
+       dml_float_t cursor_bw_pre[],
+       dml_float_t prefetch_vmrow_bw[],
+       dml_uint_t NumberOfDPP[],
+       dml_float_t UrgentBurstFactorLuma[],
+       dml_float_t UrgentBurstFactorChroma[],
+       dml_float_t UrgentBurstFactorCursor[],
+       dml_float_t UrgentBurstFactorLumaPre[],
+       dml_float_t UrgentBurstFactorChromaPre[],
+       dml_float_t UrgentBurstFactorCursorPre[],
+
+       // Output
+       dml_float_t *TotalBandwidth,
+       dml_float_t *TotalBandwidthNotIncludingMALLPrefetch,
+       dml_float_t *FractionOfUrgentBandwidth,
+       dml_bool_t *ImmediateFlipBandwidthSupport);
+
+// ---------------------------
+//  Declaration Ends
+// ---------------------------
+
+static dml_uint_t dscceComputeDelay(
+       dml_uint_t bpc,
+       dml_float_t BPP,
+       dml_uint_t sliceWidth,
+       dml_uint_t numSlices,
+       enum dml_output_format_class pixelFormat,
+       enum dml_output_encoder_class Output)
+{
+       // valid bpc = source bits per component in the set of {8, 10, 12}
+       // valid bpp = increments of 1/16 of a bit
+       // min = 6/7/8 in N420/N422/444, respectively
+       // max = such that compression is 1:1
+       //valid sliceWidth = number of pixels per slice line, must be less than or equal to 5184/numSlices (or 4096/numSlices in 420 mode)
+       //valid numSlices = number of slices in the horiziontal direction per DSC engine in the set of {1, 2, 3, 4}
+       //valid pixelFormat = pixel/color format in the set of {:N444_RGB, :S422, :N422, :N420}
+
+       // fixed value
+       dml_uint_t rcModelSize = 8192;
+
+       // N422/N420 operate at 2 pixels per clock
+       dml_uint_t pixelsPerClock, lstall, D, initalXmitDelay, w, s, ix, wx, p, l0, a, ax, L,
+               Delay, pixels;
+
+       if (pixelFormat == dml_420)
+               pixelsPerClock = 2;
+       // #all other modes operate at 1 pixel per clock
+       else if (pixelFormat == dml_444)
+               pixelsPerClock = 1;
+       else if (pixelFormat == dml_n422)
+               pixelsPerClock = 2;
+       else
+               pixelsPerClock = 1;
+
+       //initial transmit delay as per PPS
+       initalXmitDelay = (dml_uint_t)(dml_round(rcModelSize / 2.0 / BPP / pixelsPerClock, 1));
+
+       //compute ssm delay
+       if (bpc == 8)
+               D = 81;
+       else if (bpc == 10)
+               D = 89;
+       else
+               D = 113;
+
+       //divide by pixel per cycle to compute slice width as seen by DSC
+       w = sliceWidth / pixelsPerClock;
+
+       //422 mode has an additional cycle of delay
+       if (pixelFormat == dml_420 || pixelFormat == dml_444 || pixelFormat == dml_n422)
+               s = 0;
+       else
+               s = 1;
+
+       //main calculation for the dscce
+       ix = initalXmitDelay + 45;
+       wx = (w + 2) / 3;
+       p = 3 * wx - w;
+       l0 = ix / w;
+       a = ix + p * l0;
+       ax = (a + 2) / 3 + D + 6 + 1;
+       L = (ax + wx - 1) / wx;
+       if ((ix % w) == 0 && p != 0)
+               lstall = 1;
+       else
+               lstall = 0;
+       Delay = L * wx * (numSlices - 1) + ax + s + lstall + 22;
+
+       //dsc processes 3 pixel containers per cycle and a container can contain 1 or 2 pixels
+       pixels = Delay * 3 * pixelsPerClock;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: bpc: %u\n", __func__, bpc);
+       dml_print("DML::%s: BPP: %f\n", __func__, BPP);
+       dml_print("DML::%s: sliceWidth: %u\n", __func__, sliceWidth);
+       dml_print("DML::%s: numSlices: %u\n", __func__, numSlices);
+       dml_print("DML::%s: pixelFormat: %u\n", __func__, pixelFormat);
+       dml_print("DML::%s: Output: %u\n", __func__, Output);
+       dml_print("DML::%s: pixels: %u\n", __func__, pixels);
+#endif
+       return pixels;
+}
+
+static dml_uint_t dscComputeDelay(enum dml_output_format_class pixelFormat, enum dml_output_encoder_class Output)
+{
+       dml_uint_t Delay = 0;
+
+       if (pixelFormat == dml_420) {
+               // sfr
+               Delay = Delay + 2;
+               // dsccif
+               Delay = Delay + 0;
+               // dscc - input deserializer
+               Delay = Delay + 3;
+               // dscc gets pixels every other cycle
+               Delay = Delay + 2;
+               // dscc - input cdc fifo
+               Delay = Delay + 12;
+               // dscc gets pixels every other cycle
+               Delay = Delay + 13;
+               // dscc - cdc uncertainty
+               Delay = Delay + 2;
+               // dscc - output cdc fifo
+               Delay = Delay + 7;
+               // dscc gets pixels every other cycle
+               Delay = Delay + 3;
+               // dscc - cdc uncertainty
+               Delay = Delay + 2;
+               // dscc - output serializer
+               Delay = Delay + 1;
+               // sft
+               Delay = Delay + 1;
+       } else if (pixelFormat == dml_n422) {
+       // sfr
+       Delay = Delay + 2;
+       // dsccif
+       Delay = Delay + 1;
+       // dscc - input deserializer
+       Delay = Delay + 5;
+       // dscc - input cdc fifo
+       Delay = Delay + 25;
+       // dscc - cdc uncertainty
+       Delay = Delay + 2;
+       // dscc - output cdc fifo
+       Delay = Delay + 10;
+       // dscc - cdc uncertainty
+       Delay = Delay + 2;
+       // dscc - output serializer
+       Delay = Delay + 1;
+       // sft
+       Delay = Delay + 1;
+       } else {
+       // sfr
+       Delay = Delay + 2;
+       // dsccif
+       Delay = Delay + 0;
+       // dscc - input deserializer
+       Delay = Delay + 3;
+       // dscc - input cdc fifo
+       Delay = Delay + 12;
+       // dscc - cdc uncertainty
+       Delay = Delay + 2;
+       // dscc - output cdc fifo
+       Delay = Delay + 7;
+       // dscc - output serializer
+       Delay = Delay + 1;
+       // dscc - cdc uncertainty
+       Delay = Delay + 2;
+       // sft
+       Delay = Delay + 1;
+       }
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: pixelFormat = %u\n", __func__, pixelFormat);
+       dml_print("DML::%s: Delay = %u\n", __func__, Delay);
+#endif
+
+       return Delay;
+}
+
+static dml_bool_t CalculatePrefetchSchedule(struct display_mode_lib_scratch_st *scratch,
+       struct CalculatePrefetchSchedule_params_st *p)
+{
+       struct CalculatePrefetchSchedule_locals_st *s = &scratch->CalculatePrefetchSchedule_locals;
+
+       s->MyError = false;
+       s->DPPCycles = 0;
+       s->DISPCLKCycles = 0;
+       s->DSTTotalPixelsAfterScaler = 0.0;
+       s->LineTime = 0.0;
+       s->dst_y_prefetch_equ = 0.0;
+       s->prefetch_bw_oto = 0.0;
+       s->Tvm_oto = 0.0;
+       s->Tr0_oto = 0.0;
+       s->Tvm_oto_lines = 0.0;
+       s->Tr0_oto_lines = 0.0;
+       s->dst_y_prefetch_oto = 0.0;
+       s->TimeForFetchingMetaPTE = 0.0;
+       s->TimeForFetchingRowInVBlank = 0.0;
+       s->LinesToRequestPrefetchPixelData = 0.0;
+       s->HostVMDynamicLevelsTrips = 0;
+       s->trip_to_mem = 0.0;
+       s->Tvm_trips = 0.0;
+       s->Tr0_trips = 0.0;
+       s->Tvm_trips_rounded = 0.0;
+       s->Tr0_trips_rounded = 0.0;
+       s->max_Tsw = 0.0;
+       s->Lsw_oto = 0.0;
+       s->Tpre_rounded = 0.0;
+       s->prefetch_bw_equ = 0.0;
+       s->Tvm_equ = 0.0;
+       s->Tr0_equ = 0.0;
+       s->Tdmbf = 0.0;
+       s->Tdmec = 0.0;
+       s->Tdmsks = 0.0;
+       s->prefetch_sw_bytes = 0.0;
+       s->prefetch_bw_pr = 0.0;
+       s->bytes_pp = 0.0;
+       s->dep_bytes = 0.0;
+       s->min_Lsw_oto = 0.0;
+       s->Tsw_est1 = 0.0;
+       s->Tsw_est3 = 0.0;
+
+       if (p->GPUVMEnable == true && p->HostVMEnable == true) {
+               s->HostVMDynamicLevelsTrips = p->HostVMMaxNonCachedPageTableLevels;
+       } else {
+               s->HostVMDynamicLevelsTrips = 0;
+       }
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: GPUVMEnable = %u\n", __func__, p->GPUVMEnable);
+       dml_print("DML::%s: GPUVMPageTableLevels = %u\n", __func__, p->GPUVMPageTableLevels);
+       dml_print("DML::%s: DCCEnable = %u\n", __func__, p->myPipe->DCCEnable);
+       dml_print("DML::%s: VStartup = %u\n", __func__, p->VStartup);
+       dml_print("DML::%s: MaxVStartup = %u\n", __func__, p->MaxVStartup);
+       dml_print("DML::%s: HostVMEnable = %u\n", __func__, p->HostVMEnable);
+       dml_print("DML::%s: HostVMInefficiencyFactor= %f\n", __func__, p->HostVMInefficiencyFactor);
+       dml_print("DML::%s: myPipe->Dppclk = %f\n", __func__, p->myPipe->Dppclk);
+#endif
+       CalculateVUpdateAndDynamicMetadataParameters(
+       p->MaxInterDCNTileRepeaters,
+       p->myPipe->Dppclk,
+       p->myPipe->Dispclk,
+       p->myPipe->DCFClkDeepSleep,
+       p->myPipe->PixelClock,
+       p->myPipe->HTotal,
+       p->myPipe->VBlank,
+       p->DynamicMetadataTransmittedBytes,
+       p->DynamicMetadataLinesBeforeActiveRequired,
+       p->myPipe->InterlaceEnable,
+       p->myPipe->ProgressiveToInterlaceUnitInOPP,
+       p->TSetup,
+
+       // Output
+       &s->Tdmbf,
+       &s->Tdmec,
+       &s->Tdmsks,
+       p->VUpdateOffsetPix,
+       p->VUpdateWidthPix,
+       p->VReadyOffsetPix);
+
+       s->LineTime = p->myPipe->HTotal / p->myPipe->PixelClock;
+       s->trip_to_mem = p->UrgentLatency;
+       s->Tvm_trips = p->UrgentExtraLatency + s->trip_to_mem * (p->GPUVMPageTableLevels * (s->HostVMDynamicLevelsTrips + 1) - 1);
+
+       if (p->DynamicMetadataVMEnabled == true) {
+               *p->Tdmdl = p->TWait + s->Tvm_trips + s->trip_to_mem;
+       } else {
+               *p->Tdmdl = p->TWait + p->UrgentExtraLatency;
+       }
+
+#ifdef __DML_VBA_ALLOW_DELTA__
+       if (DynamicMetadataEnable == false) {
+               *Tdmdl = 0.0;
+       }
+#endif
+
+       if (p->DynamicMetadataEnable == true) {
+               if (p->VStartup * s->LineTime < *p->TSetup + *p->Tdmdl + s->Tdmbf + s->Tdmec + s->Tdmsks) {
+                       *p->NotEnoughTimeForDynamicMetadata = true;
+                       dml_print("DML::%s: Not Enough Time for Dynamic Meta!\n", __func__);
+                       dml_print("DML::%s: Tdmbf: %fus - time for dmd transfer from dchub to dio output buffer\n", __func__, s->Tdmbf);
+                       dml_print("DML::%s: Tdmec: %fus - time dio takes to transfer dmd\n", __func__, s->Tdmec);
+                       dml_print("DML::%s: Tdmsks: %fus - time before active dmd must complete transmission at dio\n", __func__, s->Tdmsks);
+                       dml_print("DML::%s: Tdmdl: %fus - time for fabric to become ready and fetch dmd \n", __func__, *p->Tdmdl);
+               } else {
+                       *p->NotEnoughTimeForDynamicMetadata = false;
+               }
+       } else {
+               *p->NotEnoughTimeForDynamicMetadata = false;
+       }
+
+       *p->Tdmdl_vm = (p->DynamicMetadataEnable == true && p->DynamicMetadataVMEnabled == true && p->GPUVMEnable == true ? p->TWait + s->Tvm_trips : 0);
+
+       if (p->myPipe->ScalerEnabled)
+               s->DPPCycles = (dml_uint_t)(p->DPPCLKDelaySubtotalPlusCNVCFormater + p->DPPCLKDelaySCL);
+       else
+               s->DPPCycles = (dml_uint_t)(p->DPPCLKDelaySubtotalPlusCNVCFormater + p->DPPCLKDelaySCLLBOnly);
+
+       s->DPPCycles = (dml_uint_t)(s->DPPCycles + p->myPipe->NumberOfCursors * p->DPPCLKDelayCNVCCursor);
+
+       s->DISPCLKCycles = (dml_uint_t)p->DISPCLKDelaySubtotal;
+
+       if (p->myPipe->Dppclk == 0.0 || p->myPipe->Dispclk == 0.0)
+               return true;
+
+       *p->DSTXAfterScaler = (dml_uint_t) dml_round(s->DPPCycles * p->myPipe->PixelClock / p->myPipe->Dppclk + s->DISPCLKCycles * p->myPipe->PixelClock / p->myPipe->Dispclk + p->DSCDelay, 1.0);
+       *p->DSTXAfterScaler = (dml_uint_t) dml_round(*p->DSTXAfterScaler + (p->myPipe->ODMMode != dml_odm_mode_bypass ? 18 : 0) + (p->myPipe->DPPPerSurface - 1) * p->DPP_RECOUT_WIDTH +
+                                               ((p->myPipe->ODMMode == dml_odm_mode_split_1to2 || p->myPipe->ODMMode == dml_odm_mode_mso_1to2) ? (dml_float_t)p->myPipe->HActive / 2.0 : 0) +
+                                               ((p->myPipe->ODMMode == dml_odm_mode_mso_1to4) ? (dml_float_t)p->myPipe->HActive * 3.0 / 4.0 : 0), 1.0);
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: DPPCycles = %u\n", __func__, s->DPPCycles);
+       dml_print("DML::%s: PixelClock = %f\n", __func__, p->myPipe->PixelClock);
+       dml_print("DML::%s: Dppclk = %f\n", __func__, p->myPipe->Dppclk);
+       dml_print("DML::%s: DISPCLKCycles = %u\n", __func__, s->DISPCLKCycles);
+       dml_print("DML::%s: DISPCLK = %f\n", __func__, p->myPipe->Dispclk);
+       dml_print("DML::%s: DSCDelay = %u\n", __func__, p->DSCDelay);
+       dml_print("DML::%s: ODMMode = %u\n", __func__, p->myPipe->ODMMode);
+       dml_print("DML::%s: DPP_RECOUT_WIDTH = %u\n", __func__, p->DPP_RECOUT_WIDTH);
+       dml_print("DML::%s: DSTXAfterScaler = %u\n", __func__, *p->DSTXAfterScaler);
+#endif
+
+       if (p->OutputFormat == dml_420 || (p->myPipe->InterlaceEnable && p->myPipe->ProgressiveToInterlaceUnitInOPP))
+               *p->DSTYAfterScaler = 1;
+       else
+               *p->DSTYAfterScaler = 0;
+
+       s->DSTTotalPixelsAfterScaler = *p->DSTYAfterScaler * p->myPipe->HTotal + *p->DSTXAfterScaler;
+       *p->DSTYAfterScaler = (dml_uint_t)(dml_floor(s->DSTTotalPixelsAfterScaler / p->myPipe->HTotal, 1));
+       *p->DSTXAfterScaler = (dml_uint_t)(s->DSTTotalPixelsAfterScaler - ((dml_float_t) (*p->DSTYAfterScaler * p->myPipe->HTotal)));
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: DSTXAfterScaler = %u (final)\n", __func__,  *p->DSTXAfterScaler);
+       dml_print("DML::%s: DSTYAfterScaler = %u (final)\n", __func__, *p->DSTYAfterScaler);
+#endif
+
+       s->MyError = false;
+
+       s->Tr0_trips = s->trip_to_mem * (s->HostVMDynamicLevelsTrips + 1);
+
+       if (p->GPUVMEnable == true) {
+               s->Tvm_trips_rounded = dml_ceil(4.0 * s->Tvm_trips / s->LineTime, 1.0) / 4.0 * s->LineTime;
+               s->Tr0_trips_rounded = dml_ceil(4.0 * s->Tr0_trips / s->LineTime, 1.0) / 4.0 * s->LineTime;
+               if (p->GPUVMPageTableLevels >= 3) {
+                       *p->Tno_bw = p->UrgentExtraLatency + s->trip_to_mem * (dml_float_t) ((p->GPUVMPageTableLevels - 2) * (s->HostVMDynamicLevelsTrips + 1) - 1);
+       } else if (p->GPUVMPageTableLevels == 1 && p->myPipe->DCCEnable != true) {
+                       s->Tr0_trips_rounded = dml_ceil(4.0 * p->UrgentExtraLatency / s->LineTime, 1.0) / 4.0 * s->LineTime;
+                       *p->Tno_bw = p->UrgentExtraLatency;
+               } else {
+                       *p->Tno_bw = 0;
+               }
+       } else if (p->myPipe->DCCEnable == true) {
+               s->Tvm_trips_rounded = s->LineTime / 4.0;
+               s->Tr0_trips_rounded = dml_ceil(4.0 * s->Tr0_trips / s->LineTime, 1.0) / 4.0 * s->LineTime;
+               *p->Tno_bw = 0;
+       } else {
+               s->Tvm_trips_rounded = s->LineTime / 4.0;
+               s->Tr0_trips_rounded = s->LineTime / 2.0;
+               *p->Tno_bw = 0;
+       }
+       s->Tvm_trips_rounded = dml_max(s->Tvm_trips_rounded, s->LineTime / 4.0);
+       s->Tr0_trips_rounded = dml_max(s->Tr0_trips_rounded, s->LineTime / 4.0);
+
+       if (p->myPipe->SourcePixelFormat == dml_420_8 || p->myPipe->SourcePixelFormat == dml_420_10 || p->myPipe->SourcePixelFormat == dml_420_12) {
+               s->bytes_pp = p->myPipe->BytePerPixelY + p->myPipe->BytePerPixelC / 4;
+       } else {
+               s->bytes_pp = p->myPipe->BytePerPixelY + p->myPipe->BytePerPixelC;
+       }
+
+       s->prefetch_bw_pr = s->bytes_pp * p->myPipe->PixelClock / (dml_float_t)p->myPipe->DPPPerSurface;
+       if (p->myPipe->VRatio < 1.0)
+               s->prefetch_bw_pr = p->myPipe->VRatio * s->prefetch_bw_pr;
+
+       s->max_Tsw = (dml_max(p->PrefetchSourceLinesY, p->PrefetchSourceLinesC) * s->LineTime);
+
+       s->prefetch_sw_bytes = p->PrefetchSourceLinesY * p->swath_width_luma_ub * p->myPipe->BytePerPixelY + p->PrefetchSourceLinesC * p->swath_width_chroma_ub * p->myPipe->BytePerPixelC;
+       s->prefetch_bw_oto = dml_max(s->prefetch_bw_pr, s->prefetch_sw_bytes / s->max_Tsw);
+
+       s->min_Lsw_oto = dml_max(p->PrefetchSourceLinesY, p->PrefetchSourceLinesC) / __DML_MAX_VRATIO_PRE_OTO__;
+       s->min_Lsw_oto = dml_max(s->min_Lsw_oto, 1.0);
+       s->Lsw_oto = dml_ceil(4.0 * dml_max(s->prefetch_sw_bytes / s->prefetch_bw_oto / s->LineTime, s->min_Lsw_oto), 1.0) / 4.0;
+
+       if (p->GPUVMEnable == true) {
+               s->Tvm_oto = dml_max3(
+                       s->Tvm_trips,
+                       *p->Tno_bw + p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor / s->prefetch_bw_oto,
+                       s->LineTime / 4.0);
+       } else
+               s->Tvm_oto = s->LineTime / 4.0;
+
+       if ((p->GPUVMEnable == true || p->myPipe->DCCEnable == true)) {
+               s->Tr0_oto = dml_max4(
+                       s->Tr0_trips,
+                       (p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / s->prefetch_bw_oto,
+                       (s->LineTime - s->Tvm_oto)/2.0,
+                       s->LineTime / 4.0);
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: Tr0_oto max0 = %f\n", __func__, (p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / s->prefetch_bw_oto);
+       dml_print("DML::%s: Tr0_oto max1 = %f\n", __func__, s->Tr0_trips);
+       dml_print("DML::%s: Tr0_oto max2 = %f\n", __func__, s->LineTime - s->Tvm_oto);
+       dml_print("DML::%s: Tr0_oto max3 = %f\n", __func__, s->LineTime / 4);
+#endif
+       } else
+               s->Tr0_oto = (s->LineTime - s->Tvm_oto) / 2.0;
+
+       s->Tvm_oto_lines = dml_ceil(4.0 * s->Tvm_oto / s->LineTime, 1) / 4.0;
+       s->Tr0_oto_lines = dml_ceil(4.0 * s->Tr0_oto / s->LineTime, 1) / 4.0;
+       s->dst_y_prefetch_oto = s->Tvm_oto_lines + 2 * s->Tr0_oto_lines + s->Lsw_oto;
+
+       s->dst_y_prefetch_equ = p->VStartup - (*p->TSetup + dml_max(p->TWait + p->TCalc, *p->Tdmdl)) / s->LineTime - (*p->DSTYAfterScaler + (dml_float_t) *p->DSTXAfterScaler / (dml_float_t)p->myPipe->HTotal);
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: HTotal = %u\n", __func__, p->myPipe->HTotal);
+       dml_print("DML::%s: min_Lsw_oto = %f\n", __func__, s->min_Lsw_oto);
+       dml_print("DML::%s: *Tno_bw = %f\n", __func__, *p->Tno_bw);
+       dml_print("DML::%s: UrgentExtraLatency = %f\n", __func__, p->UrgentExtraLatency);
+       dml_print("DML::%s: trip_to_mem = %f\n", __func__, s->trip_to_mem);
+       dml_print("DML::%s: BytePerPixelY = %u\n", __func__, p->myPipe->BytePerPixelY);
+       dml_print("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY);
+       dml_print("DML::%s: swath_width_luma_ub = %u\n", __func__, p->swath_width_luma_ub);
+       dml_print("DML::%s: BytePerPixelC = %u\n", __func__, p->myPipe->BytePerPixelC);
+       dml_print("DML::%s: PrefetchSourceLinesC = %f\n", __func__, p->PrefetchSourceLinesC);
+       dml_print("DML::%s: swath_width_chroma_ub = %u\n", __func__, p->swath_width_chroma_ub);
+       dml_print("DML::%s: prefetch_sw_bytes = %f\n", __func__, s->prefetch_sw_bytes);
+       dml_print("DML::%s: bytes_pp = %f\n", __func__, s->bytes_pp);
+       dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %u\n", __func__, p->PDEAndMetaPTEBytesFrame);
+       dml_print("DML::%s: MetaRowByte = %u\n", __func__, p->MetaRowByte);
+       dml_print("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, p->PixelPTEBytesPerRow);
+       dml_print("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor);
+       dml_print("DML::%s: Tvm_trips = %f\n", __func__, s->Tvm_trips);
+       dml_print("DML::%s: Tr0_trips = %f\n", __func__, s->Tr0_trips);
+       dml_print("DML::%s: prefetch_bw_oto = %f\n", __func__, s->prefetch_bw_oto);
+       dml_print("DML::%s: Tr0_oto = %f\n", __func__, s->Tr0_oto);
+       dml_print("DML::%s: Tvm_oto = %f\n", __func__, s->Tvm_oto);
+       dml_print("DML::%s: Tvm_oto_lines = %f\n", __func__, s->Tvm_oto_lines);
+       dml_print("DML::%s: Tr0_oto_lines = %f\n", __func__, s->Tr0_oto_lines);
+       dml_print("DML::%s: Lsw_oto = %f\n", __func__, s->Lsw_oto);
+       dml_print("DML::%s: dst_y_prefetch_oto = %f\n", __func__, s->dst_y_prefetch_oto);
+       dml_print("DML::%s: dst_y_prefetch_equ = %f\n", __func__, s->dst_y_prefetch_equ);
+#endif
+
+       s->dst_y_prefetch_equ = dml_floor(4.0 * (s->dst_y_prefetch_equ + 0.125), 1) / 4.0;
+       s->Tpre_rounded = s->dst_y_prefetch_equ * s->LineTime;
+
+       dml_print("DML::%s: dst_y_prefetch_equ: %f (after round)\n", __func__, s->dst_y_prefetch_equ);
+
+       dml_print("DML::%s: LineTime: %f\n", __func__, s->LineTime);
+       dml_print("DML::%s: VStartup: %u\n", __func__, p->VStartup);
+       dml_print("DML::%s: Tvstartup: %fus - time between vstartup and first pixel of active\n", __func__, p->VStartup * s->LineTime);
+       dml_print("DML::%s: TSetup: %fus - time from vstartup to vready\n", __func__, *p->TSetup);
+       dml_print("DML::%s: TCalc: %fus - time for calculations in dchub starting at vready\n", __func__, p->TCalc);
+       dml_print("DML::%s: TWait: %fus - time for fabric to become ready max(pstate exit,cstate enter/exit, urgent latency) after TCalc\n", __func__, p->TWait);
+       dml_print("DML::%s: Tdmbf: %fus - time for dmd transfer from dchub to dio output buffer\n", __func__, s->Tdmbf);
+       dml_print("DML::%s: Tdmec: %fus - time dio takes to transfer dmd\n", __func__, s->Tdmec);
+       dml_print("DML::%s: Tdmsks: %fus - time before active dmd must complete transmission at dio\n", __func__, s->Tdmsks);
+       dml_print("DML::%s: Tdmdl_vm: %fus - time for vm stages of dmd \n", __func__, *p->Tdmdl_vm);
+       dml_print("DML::%s: Tdmdl: %fus - time for fabric to become ready and fetch dmd \n", __func__, *p->Tdmdl);
+       dml_print("DML::%s: DSTXAfterScaler: %u pixels - number of pixel clocks pipeline and buffer delay after scaler \n", __func__, *p->DSTXAfterScaler);
+       dml_print("DML::%s: DSTYAfterScaler: %u lines - number of lines of pipeline and buffer delay after scaler \n", __func__, *p->DSTYAfterScaler);
+
+       s->dep_bytes = dml_max(p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor, p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor);
+
+       if (s->prefetch_sw_bytes < s->dep_bytes) {
+               s->prefetch_sw_bytes = 2 * s->dep_bytes;
+       }
+
+       *p->DestinationLinesToRequestVMInVBlank = 0;
+       *p->DestinationLinesToRequestRowInVBlank = 0;
+       *p->VRatioPrefetchY = 0;
+       *p->VRatioPrefetchC = 0;
+       *p->RequiredPrefetchPixDataBWLuma = 0;
+       if (s->dst_y_prefetch_equ > 1) {
+
+               if (s->Tpre_rounded - *p->Tno_bw > 0) {
+               s->PrefetchBandwidth1 = (p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor + 2 * p->MetaRowByte
+                                       + 2 * p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor
+                                       + s->prefetch_sw_bytes)
+                                       / (s->Tpre_rounded - *p->Tno_bw);
+                       s->Tsw_est1 = s->prefetch_sw_bytes / s->PrefetchBandwidth1;
+               } else
+                       s->PrefetchBandwidth1 = 0;
+
+               if (p->VStartup == p->MaxVStartup && (s->Tsw_est1 / s->LineTime < s->min_Lsw_oto) && s->Tpre_rounded - s->min_Lsw_oto * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw > 0) {
+                       s->PrefetchBandwidth1 = (p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor + 2 * p->MetaRowByte + 2 * p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) /
+                                                               (s->Tpre_rounded - s->min_Lsw_oto * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw);
+               }
+
+               if (s->Tpre_rounded - *p->Tno_bw - 2 * s->Tr0_trips_rounded > 0)
+                       s->PrefetchBandwidth2 = (p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor + s->prefetch_sw_bytes) /
+                                                                       (s->Tpre_rounded - *p->Tno_bw - 2 * s->Tr0_trips_rounded);
+               else
+                       s->PrefetchBandwidth2 = 0;
+
+               if (s->Tpre_rounded - s->Tvm_trips_rounded > 0) {
+                       s->PrefetchBandwidth3 = (2 * p->MetaRowByte + 2 * p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + s->prefetch_sw_bytes) /
+                                                                       (s->Tpre_rounded - s->Tvm_trips_rounded);
+                       s->Tsw_est3 = s->prefetch_sw_bytes / s->PrefetchBandwidth3;
+       }
+               else
+                       s->PrefetchBandwidth3 = 0;
+
+
+               if (p->VStartup == p->MaxVStartup && (s->Tsw_est3 / s->LineTime < s->min_Lsw_oto) && s->Tpre_rounded - s->min_Lsw_oto * s->LineTime - 0.5 * s->LineTime - s->Tvm_trips_rounded > 0) {
+                       s->PrefetchBandwidth3 = (2 * p->MetaRowByte + 2 * p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / (s->Tpre_rounded - s->min_Lsw_oto * s->LineTime - 0.5 * s->LineTime - s->Tvm_trips_rounded);
+               }
+
+               if (s->Tpre_rounded - s->Tvm_trips_rounded - 2 * s->Tr0_trips_rounded > 0)
+                       s->PrefetchBandwidth4 = s->prefetch_sw_bytes / (s->Tpre_rounded - s->Tvm_trips_rounded - 2 * s->Tr0_trips_rounded);
+               else
+                       s->PrefetchBandwidth4 = 0;
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: Tpre_rounded: %f\n", __func__, s->Tpre_rounded);
+               dml_print("DML::%s: Tno_bw: %f\n", __func__, *p->Tno_bw);
+               dml_print("DML::%s: Tvm_trips_rounded: %f\n", __func__, s->Tvm_trips_rounded);
+               dml_print("DML::%s: Tsw_est1: %f\n", __func__, s->Tsw_est1);
+               dml_print("DML::%s: Tsw_est3: %f\n", __func__, s->Tsw_est3);
+               dml_print("DML::%s: PrefetchBandwidth1: %f\n", __func__, s->PrefetchBandwidth1);
+               dml_print("DML::%s: PrefetchBandwidth2: %f\n", __func__, s->PrefetchBandwidth2);
+               dml_print("DML::%s: PrefetchBandwidth3: %f\n", __func__, s->PrefetchBandwidth3);
+               dml_print("DML::%s: PrefetchBandwidth4: %f\n", __func__, s->PrefetchBandwidth4);
+#endif
+               {
+                       dml_bool_t Case1OK;
+                       dml_bool_t Case2OK;
+                       dml_bool_t Case3OK;
+
+                       if (s->PrefetchBandwidth1 > 0) {
+                               if (*p->Tno_bw + p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor / s->PrefetchBandwidth1 >= s->Tvm_trips_rounded && (p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / s->PrefetchBandwidth1 >= s->Tr0_trips_rounded) {
+                                       Case1OK = true;
+                               } else {
+                                       Case1OK = false;
+                               }
+                       } else {
+                               Case1OK = false;
+                       }
+
+                       if (s->PrefetchBandwidth2 > 0) {
+                               if (*p->Tno_bw + p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor / s->PrefetchBandwidth2 >= s->Tvm_trips_rounded && (p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / s->PrefetchBandwidth2 < s->Tr0_trips_rounded) {
+                                       Case2OK = true;
+                               } else {
+                                       Case2OK = false;
+                               }
+                       } else {
+                               Case2OK = false;
+                       }
+
+                       if (s->PrefetchBandwidth3 > 0) {
+                               if (*p->Tno_bw + p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor / s->PrefetchBandwidth3 < s->Tvm_trips_rounded && (p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / s->PrefetchBandwidth3 >= s->Tr0_trips_rounded) {
+                                       Case3OK = true;
+                               } else {
+                                       Case3OK = false;
+                               }
+                       } else {
+                               Case3OK = false;
+                       }
+
+                       if (Case1OK) {
+                               s->prefetch_bw_equ = s->PrefetchBandwidth1;
+                       } else if (Case2OK) {
+                               s->prefetch_bw_equ = s->PrefetchBandwidth2;
+                       } else if (Case3OK) {
+                               s->prefetch_bw_equ = s->PrefetchBandwidth3;
+                       } else {
+                               s->prefetch_bw_equ = s->PrefetchBandwidth4;
+                       }
+
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: Case1OK: %u\n", __func__, Case1OK);
+                       dml_print("DML::%s: Case2OK: %u\n", __func__, Case2OK);
+                       dml_print("DML::%s: Case3OK: %u\n", __func__, Case3OK);
+                       dml_print("DML::%s: prefetch_bw_equ: %f\n", __func__, s->prefetch_bw_equ);
+#endif
+
+                       if (s->prefetch_bw_equ > 0) {
+                               if (p->GPUVMEnable == true) {
+                                       s->Tvm_equ = dml_max3(*p->Tno_bw + p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor / s->prefetch_bw_equ, s->Tvm_trips, s->LineTime / 4);
+                               } else {
+                                       s->Tvm_equ = s->LineTime / 4;
+                               }
+
+                               if ((p->GPUVMEnable == true || p->myPipe->DCCEnable == true)) {
+                                       s->Tr0_equ = dml_max4((p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / s->prefetch_bw_equ, s->Tr0_trips, (s->LineTime - s->Tvm_equ) / 2, s->LineTime / 4);
+                               } else {
+                                       s->Tr0_equ = (s->LineTime - s->Tvm_equ) / 2;
+                               }
+                       } else {
+                               s->Tvm_equ = 0;
+                               s->Tr0_equ = 0;
+                               dml_print("DML::%s: prefetch_bw_equ equals 0!\n", __func__);
+                       }
+               }
+
+
+               if (s->dst_y_prefetch_oto < s->dst_y_prefetch_equ) {
+                       *p->DestinationLinesForPrefetch = s->dst_y_prefetch_oto;
+                       s->TimeForFetchingMetaPTE = s->Tvm_oto;
+                       s->TimeForFetchingRowInVBlank = s->Tr0_oto;
+
+                       *p->DestinationLinesToRequestVMInVBlank = dml_ceil(4.0 * s->TimeForFetchingMetaPTE / s->LineTime, 1.0) / 4.0;
+                       *p->DestinationLinesToRequestRowInVBlank = dml_ceil(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0;
+               } else {
+                       *p->DestinationLinesForPrefetch = s->dst_y_prefetch_equ;
+                       s->TimeForFetchingMetaPTE = s->Tvm_equ;
+                       s->TimeForFetchingRowInVBlank = s->Tr0_equ;
+
+                       if (p->VStartup == p->MaxVStartup && p->EnhancedPrefetchScheduleAccelerationFinal != 0) {
+                               *p->DestinationLinesToRequestVMInVBlank = dml_floor(4.0 * s->TimeForFetchingMetaPTE / s->LineTime, 1.0) / 4.0;
+                               *p->DestinationLinesToRequestRowInVBlank = dml_floor(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0;
+                       } else {
+                               *p->DestinationLinesToRequestVMInVBlank = dml_ceil(4.0 * s->TimeForFetchingMetaPTE / s->LineTime, 1.0) / 4.0;
+                               *p->DestinationLinesToRequestRowInVBlank = dml_ceil(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0;
+                       }
+               }
+
+               s->LinesToRequestPrefetchPixelData = *p->DestinationLinesForPrefetch - *p->DestinationLinesToRequestVMInVBlank - 2 * *p->DestinationLinesToRequestRowInVBlank;
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: DestinationLinesForPrefetch = %f\n", __func__, *p->DestinationLinesForPrefetch);
+               dml_print("DML::%s: DestinationLinesToRequestVMInVBlank = %f\n", __func__, *p->DestinationLinesToRequestVMInVBlank);
+               dml_print("DML::%s: TimeForFetchingRowInVBlank = %f\n", __func__, s->TimeForFetchingRowInVBlank);
+               dml_print("DML::%s: LineTime = %f\n", __func__, s->LineTime);
+               dml_print("DML::%s: DestinationLinesToRequestRowInVBlank = %f\n", __func__, *p->DestinationLinesToRequestRowInVBlank);
+               dml_print("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY);
+               dml_print("DML::%s: LinesToRequestPrefetchPixelData = %f\n", __func__, s->LinesToRequestPrefetchPixelData);
+#endif
+
+               if (s->LinesToRequestPrefetchPixelData >= 1 && s->prefetch_bw_equ > 0) {
+                       *p->VRatioPrefetchY = (dml_float_t)p->PrefetchSourceLinesY / s->LinesToRequestPrefetchPixelData;
+                       *p->VRatioPrefetchY = dml_max(*p->VRatioPrefetchY, 1.0);
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: VRatioPrefetchY = %f\n", __func__, *p->VRatioPrefetchY);
+                       dml_print("DML::%s: SwathHeightY = %u\n", __func__, p->SwathHeightY);
+                       dml_print("DML::%s: VInitPreFillY = %u\n", __func__, p->VInitPreFillY);
+#endif
+                       if ((p->SwathHeightY > 4) && (p->VInitPreFillY > 3)) {
+                               if (s->LinesToRequestPrefetchPixelData > (p->VInitPreFillY - 3.0) / 2.0) {
+                                       *p->VRatioPrefetchY = dml_max(*p->VRatioPrefetchY,
+                                                                               (dml_float_t)p->MaxNumSwathY * p->SwathHeightY / (s->LinesToRequestPrefetchPixelData - (p->VInitPreFillY - 3.0) / 2.0));
+                       } else {
+                               s->MyError = true;
+                               dml_print("DML::%s: MyErr set. LinesToRequestPrefetchPixelData=%f VinitPreFillY=%u\n", __func__, s->LinesToRequestPrefetchPixelData, p->VInitPreFillY);
+                               *p->VRatioPrefetchY = 0;
+                       }
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: VRatioPrefetchY = %f\n", __func__, *p->VRatioPrefetchY);
+                       dml_print("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY);
+                       dml_print("DML::%s: MaxNumSwathY = %u\n", __func__, p->MaxNumSwathY);
+#endif
+                       }
+
+                       *p->VRatioPrefetchC = (dml_float_t)p->PrefetchSourceLinesC / s->LinesToRequestPrefetchPixelData;
+                       *p->VRatioPrefetchC = dml_max(*p->VRatioPrefetchC, 1.0);
+
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: VRatioPrefetchC = %f\n", __func__, *p->VRatioPrefetchC);
+                       dml_print("DML::%s: SwathHeightC = %u\n", __func__, p->SwathHeightC);
+                       dml_print("DML::%s: VInitPreFillC = %u\n", __func__, p->VInitPreFillC);
+#endif
+                       if ((p->SwathHeightC > 4) && (p->VInitPreFillC > 3)) {
+                               if (s->LinesToRequestPrefetchPixelData > (p->VInitPreFillC - 3.0) / 2.0) {
+                                       *p->VRatioPrefetchC = dml_max(*p->VRatioPrefetchC, (dml_float_t)p->MaxNumSwathC * p->SwathHeightC / (s->LinesToRequestPrefetchPixelData - (p->VInitPreFillC - 3.0) / 2.0));
+                               } else {
+                                       s->MyError = true;
+                                       dml_print("DML::%s: MyErr set. LinesToRequestPrefetchPixelData=%f VInitPreFillC=%u\n", __func__, s->LinesToRequestPrefetchPixelData, p->VInitPreFillC);
+                                       *p->VRatioPrefetchC = 0;
+                               }
+#ifdef __DML_VBA_DEBUG__
+                               dml_print("DML::%s: VRatioPrefetchC = %f\n", __func__, *p->VRatioPrefetchC);
+                               dml_print("DML::%s: PrefetchSourceLinesC = %f\n", __func__, p->PrefetchSourceLinesC);
+                               dml_print("DML::%s: MaxNumSwathC = %u\n", __func__, p->MaxNumSwathC);
+#endif
+                       }
+
+                       *p->RequiredPrefetchPixDataBWLuma = (dml_float_t)p->PrefetchSourceLinesY / s->LinesToRequestPrefetchPixelData
+                               * p->myPipe->BytePerPixelY
+                               * p->swath_width_luma_ub / s->LineTime;
+
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: BytePerPixelY = %u\n", __func__, p->myPipe->BytePerPixelY);
+                       dml_print("DML::%s: swath_width_luma_ub = %u\n", __func__, p->swath_width_luma_ub);
+                       dml_print("DML::%s: LineTime = %f\n", __func__, s->LineTime);
+                       dml_print("DML::%s: RequiredPrefetchPixDataBWLuma = %f\n", __func__, *p->RequiredPrefetchPixDataBWLuma);
+#endif
+                       *p->RequiredPrefetchPixDataBWChroma = (dml_float_t)p->PrefetchSourceLinesC / s->LinesToRequestPrefetchPixelData
+                               *p->myPipe->BytePerPixelC
+                               *p->swath_width_chroma_ub / s->LineTime;
+               } else {
+                       s->MyError = true;
+                       dml_print("DML:%s: MyErr set. LinesToRequestPrefetchPixelData: %f, should be > 0\n", __func__, s->LinesToRequestPrefetchPixelData);
+                       *p->VRatioPrefetchY = 0;
+                       *p->VRatioPrefetchC = 0;
+                       *p->RequiredPrefetchPixDataBWLuma = 0;
+                       *p->RequiredPrefetchPixDataBWChroma = 0;
+               }
+
+               dml_print("DML: Tpre: %fus - sum of time to request meta pte, 2 x data pte + meta data, swaths\n", (dml_float_t)s->LinesToRequestPrefetchPixelData * s->LineTime + 2.0 * s->TimeForFetchingRowInVBlank + s->TimeForFetchingMetaPTE);
+               dml_print("DML: Tvm: %fus - time to fetch page tables for meta surface\n", s->TimeForFetchingMetaPTE);
+               dml_print("DML: Tr0: %fus - time to fetch first row of data pagetables and first row of meta data (done in parallel)\n", s->TimeForFetchingRowInVBlank);
+               dml_print("DML: Tsw: %fus = time to fetch enough pixel data and cursor data to feed the scalers init position and detile\n", (dml_float_t)s->LinesToRequestPrefetchPixelData * s->LineTime);
+               dml_print("DML: To: %fus - time for propogation from scaler to optc\n", (*p->DSTYAfterScaler + ((dml_float_t) (*p->DSTXAfterScaler) / (dml_float_t)p->myPipe->HTotal)) * s->LineTime);
+               dml_print("DML: Tvstartup - TSetup - Tcalc - Twait - Tpre - To > 0\n");
+               dml_print("DML: Tslack(pre): %fus - time left over in schedule\n", p->VStartup * s->LineTime - s->TimeForFetchingMetaPTE - 2 * s->TimeForFetchingRowInVBlank - (*p->DSTYAfterScaler + ((dml_float_t) (*p->DSTXAfterScaler) / (dml_float_t)p->myPipe->HTotal)) * s->LineTime - p->TWait - p->TCalc - *p->TSetup);
+               dml_print("DML: row_bytes = dpte_row_bytes (per_pipe) = PixelPTEBytesPerRow = : %u\n", p->PixelPTEBytesPerRow);
+
+       } else {
+               s->MyError = true;
+               dml_print("DML::%s: MyErr set, dst_y_prefetch_equ = %f (should be > 1)\n", __func__, s->dst_y_prefetch_equ);
+               s->TimeForFetchingMetaPTE = 0;
+               s->TimeForFetchingRowInVBlank = 0;
+               *p->DestinationLinesToRequestVMInVBlank = 0;
+               *p->DestinationLinesToRequestRowInVBlank = 0;
+               s->LinesToRequestPrefetchPixelData = 0;
+               *p->VRatioPrefetchY = 0;
+               *p->VRatioPrefetchC = 0;
+               *p->RequiredPrefetchPixDataBWLuma = 0;
+               *p->RequiredPrefetchPixDataBWChroma = 0;
+       }
+
+       {
+               dml_float_t prefetch_vm_bw;
+               dml_float_t prefetch_row_bw;
+
+               if (p->PDEAndMetaPTEBytesFrame == 0) {
+                       prefetch_vm_bw = 0;
+               } else if (*p->DestinationLinesToRequestVMInVBlank > 0) {
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %u\n", __func__, p->PDEAndMetaPTEBytesFrame);
+                       dml_print("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor);
+                       dml_print("DML::%s: DestinationLinesToRequestVMInVBlank = %f\n", __func__, *p->DestinationLinesToRequestVMInVBlank);
+                       dml_print("DML::%s: LineTime = %f\n", __func__, s->LineTime);
+#endif
+               prefetch_vm_bw = p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor / (*p->DestinationLinesToRequestVMInVBlank * s->LineTime);
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: prefetch_vm_bw = %f\n", __func__, prefetch_vm_bw);
+#endif
+               } else {
+                       prefetch_vm_bw = 0;
+                       s->MyError = true;
+                       dml_print("DML::%s: MyErr set. DestinationLinesToRequestVMInVBlank=%f (should be > 0)\n", __func__, *p->DestinationLinesToRequestVMInVBlank);
+               }
+
+               if (p->MetaRowByte + p->PixelPTEBytesPerRow == 0) {
+                       prefetch_row_bw = 0;
+               } else if (*p->DestinationLinesToRequestRowInVBlank > 0) {
+                       prefetch_row_bw = (p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / (*p->DestinationLinesToRequestRowInVBlank * s->LineTime);
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: MetaRowByte = %u\n", __func__, p->MetaRowByte);
+               dml_print("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, p->PixelPTEBytesPerRow);
+               dml_print("DML::%s: DestinationLinesToRequestRowInVBlank = %f\n", __func__, *p->DestinationLinesToRequestRowInVBlank);
+               dml_print("DML::%s: prefetch_row_bw = %f\n", __func__, prefetch_row_bw);
+#endif
+               } else {
+                       prefetch_row_bw = 0;
+                       s->MyError = true;
+                       dml_print("DML::%s: MyErr set. DestinationLinesToRequestRowInVBlank=%f (should be > 0)\n", __func__, *p->DestinationLinesToRequestRowInVBlank);
+               }
+
+               *p->prefetch_vmrow_bw = dml_max(prefetch_vm_bw, prefetch_row_bw);
+       }
+
+       if (s->MyError) {
+               s->TimeForFetchingMetaPTE = 0;
+               s->TimeForFetchingRowInVBlank = 0;
+               *p->DestinationLinesToRequestVMInVBlank = 0;
+               *p->DestinationLinesToRequestRowInVBlank = 0;
+               *p->DestinationLinesForPrefetch = 0;
+               s->LinesToRequestPrefetchPixelData = 0;
+               *p->VRatioPrefetchY = 0;
+               *p->VRatioPrefetchC = 0;
+               *p->RequiredPrefetchPixDataBWLuma = 0;
+               *p->RequiredPrefetchPixDataBWChroma = 0;
+       }
+
+       return s->MyError;
+} // CalculatePrefetchSchedule
+
+static void CalculateBytePerPixelAndBlockSizes(
+       enum dml_source_format_class SourcePixelFormat,
+       enum dml_swizzle_mode SurfaceTiling,
+
+       // Output
+       dml_uint_t *BytePerPixelY,
+       dml_uint_t *BytePerPixelC,
+       dml_float_t *BytePerPixelDETY,
+       dml_float_t *BytePerPixelDETC,
+       dml_uint_t *BlockHeight256BytesY,
+       dml_uint_t *BlockHeight256BytesC,
+       dml_uint_t *BlockWidth256BytesY,
+       dml_uint_t *BlockWidth256BytesC,
+       dml_uint_t *MacroTileHeightY,
+       dml_uint_t *MacroTileHeightC,
+       dml_uint_t *MacroTileWidthY,
+       dml_uint_t *MacroTileWidthC)
+{
+       if (SourcePixelFormat == dml_444_64) {
+               *BytePerPixelDETY = 8;
+               *BytePerPixelDETC = 0;
+               *BytePerPixelY = 8;
+               *BytePerPixelC = 0;
+       } else if (SourcePixelFormat == dml_444_32 || SourcePixelFormat == dml_rgbe) {
+               *BytePerPixelDETY = 4;
+               *BytePerPixelDETC = 0;
+               *BytePerPixelY = 4;
+               *BytePerPixelC = 0;
+       } else if (SourcePixelFormat == dml_444_16 || SourcePixelFormat == dml_mono_16) {
+               *BytePerPixelDETY = 2;
+               *BytePerPixelDETC = 0;
+               *BytePerPixelY = 2;
+               *BytePerPixelC = 0;
+       } else if (SourcePixelFormat == dml_444_8 || SourcePixelFormat == dml_mono_8) {
+               *BytePerPixelDETY = 1;
+               *BytePerPixelDETC = 0;
+               *BytePerPixelY = 1;
+               *BytePerPixelC = 0;
+       } else if (SourcePixelFormat == dml_rgbe_alpha) {
+               *BytePerPixelDETY = 4;
+               *BytePerPixelDETC = 1;
+               *BytePerPixelY = 4;
+               *BytePerPixelC = 1;
+       } else if (SourcePixelFormat == dml_420_8) {
+               *BytePerPixelDETY = 1;
+               *BytePerPixelDETC = 2;
+               *BytePerPixelY = 1;
+               *BytePerPixelC = 2;
+       } else if (SourcePixelFormat == dml_420_12) {
+               *BytePerPixelDETY = 2;
+               *BytePerPixelDETC = 4;
+               *BytePerPixelY = 2;
+               *BytePerPixelC = 4;
+       } else {
+               *BytePerPixelDETY = (dml_float_t) (4.0 / 3);
+               *BytePerPixelDETC = (dml_float_t) (8.0 / 3);
+               *BytePerPixelY = 2;
+               *BytePerPixelC = 4;
+       }
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: SourcePixelFormat = %u\n", __func__, SourcePixelFormat);
+       dml_print("DML::%s: BytePerPixelDETY = %f\n", __func__, *BytePerPixelDETY);
+       dml_print("DML::%s: BytePerPixelDETC = %f\n", __func__, *BytePerPixelDETC);
+       dml_print("DML::%s: BytePerPixelY = %u\n", __func__, *BytePerPixelY);
+       dml_print("DML::%s: BytePerPixelC = %u\n", __func__, *BytePerPixelC);
+#endif
+       if ((SourcePixelFormat == dml_444_64 || SourcePixelFormat == dml_444_32
+               || SourcePixelFormat == dml_444_16
+               || SourcePixelFormat == dml_444_8
+               || SourcePixelFormat == dml_mono_16
+               || SourcePixelFormat == dml_mono_8
+               || SourcePixelFormat == dml_rgbe)) {
+               if (SurfaceTiling == dml_sw_linear) {
+                       *BlockHeight256BytesY = 1;
+               } else if (SourcePixelFormat == dml_444_64) {
+                       *BlockHeight256BytesY = 4;
+               } else if (SourcePixelFormat == dml_444_8) {
+                       *BlockHeight256BytesY = 16;
+               } else {
+                       *BlockHeight256BytesY = 8;
+               }
+               *BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
+               *BlockHeight256BytesC = 0;
+               *BlockWidth256BytesC = 0;
+       } else {
+               if (SurfaceTiling == dml_sw_linear) {
+                       *BlockHeight256BytesY = 1;
+                       *BlockHeight256BytesC = 1;
+               } else if (SourcePixelFormat == dml_rgbe_alpha) {
+                       *BlockHeight256BytesY = 8;
+                       *BlockHeight256BytesC = 16;
+               } else if (SourcePixelFormat == dml_420_8) {
+                       *BlockHeight256BytesY = 16;
+                       *BlockHeight256BytesC = 8;
+               } else {
+                       *BlockHeight256BytesY = 8;
+                       *BlockHeight256BytesC = 8;
+               }
+               *BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
+               *BlockWidth256BytesC = 256U / *BytePerPixelC / *BlockHeight256BytesC;
+       }
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: BlockWidth256BytesY = %u\n", __func__, *BlockWidth256BytesY);
+       dml_print("DML::%s: BlockHeight256BytesY = %u\n", __func__, *BlockHeight256BytesY);
+       dml_print("DML::%s: BlockWidth256BytesC = %u\n", __func__, *BlockWidth256BytesC);
+       dml_print("DML::%s: BlockHeight256BytesC = %u\n", __func__, *BlockHeight256BytesC);
+#endif
+
+       if (SurfaceTiling == dml_sw_linear) {
+               *MacroTileHeightY = *BlockHeight256BytesY;
+               *MacroTileWidthY = 256 / *BytePerPixelY / *MacroTileHeightY;
+               *MacroTileHeightC = *BlockHeight256BytesC;
+               if (*MacroTileHeightC == 0) {
+                       *MacroTileWidthC = 0;
+               } else {
+                       *MacroTileWidthC = 256 / *BytePerPixelC / *MacroTileHeightC;
+       }
+       } else if (SurfaceTiling == dml_sw_64kb_d || SurfaceTiling == dml_sw_64kb_d_t || SurfaceTiling == dml_sw_64kb_d_x || SurfaceTiling == dml_sw_64kb_r_x) {
+               *MacroTileHeightY = 16 * *BlockHeight256BytesY;
+               *MacroTileWidthY = 65536 / *BytePerPixelY / *MacroTileHeightY;
+               *MacroTileHeightC = 16 * *BlockHeight256BytesC;
+               if (*MacroTileHeightC == 0) {
+                       *MacroTileWidthC = 0;
+               } else {
+                       *MacroTileWidthC = 65536 / *BytePerPixelC / *MacroTileHeightC;
+               }
+       } else {
+               *MacroTileHeightY = 32 * *BlockHeight256BytesY;
+               *MacroTileWidthY = 65536 * 4 / *BytePerPixelY / *MacroTileHeightY;
+               *MacroTileHeightC = 32 * *BlockHeight256BytesC;
+               if (*MacroTileHeightC == 0) {
+                       *MacroTileWidthC = 0;
+               } else {
+                       *MacroTileWidthC = 65536 * 4 / *BytePerPixelC / *MacroTileHeightC;
+               }
+       }
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: MacroTileWidthY = %u\n", __func__, *MacroTileWidthY);
+       dml_print("DML::%s: MacroTileHeightY = %u\n", __func__, *MacroTileHeightY);
+       dml_print("DML::%s: MacroTileWidthC = %u\n", __func__, *MacroTileWidthC);
+       dml_print("DML::%s: MacroTileHeightC = %u\n", __func__, *MacroTileHeightC);
+#endif
+} // CalculateBytePerPixelAndBlockSizes
+
+static dml_float_t CalculateTWait(
+               dml_uint_t PrefetchMode,
+               enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange,
+               dml_bool_t SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+               dml_bool_t DRRDisplay,
+               dml_float_t DRAMClockChangeLatency,
+               dml_float_t FCLKChangeLatency,
+               dml_float_t UrgentLatency,
+               dml_float_t SREnterPlusExitTime)
+{
+       dml_float_t TWait = 0.0;
+
+       if (PrefetchMode == 0 &&
+                       !(UseMALLForPStateChange == dml_use_mall_pstate_change_full_frame) && !(UseMALLForPStateChange == dml_use_mall_pstate_change_sub_viewport) &&
+                       !(UseMALLForPStateChange == dml_use_mall_pstate_change_phantom_pipe) && !(SynchronizeDRRDisplaysForUCLKPStateChangeFinal && DRRDisplay)) {
+               TWait = dml_max3(DRAMClockChangeLatency + UrgentLatency, SREnterPlusExitTime, UrgentLatency);
+       } else if (PrefetchMode <= 1 && !(UseMALLForPStateChange == dml_use_mall_pstate_change_phantom_pipe)) {
+               TWait = dml_max3(FCLKChangeLatency + UrgentLatency, SREnterPlusExitTime, UrgentLatency);
+       } else if (PrefetchMode <= 2 && !(UseMALLForPStateChange == dml_use_mall_pstate_change_phantom_pipe)) {
+               TWait = dml_max(SREnterPlusExitTime, UrgentLatency);
+       } else {
+               TWait = UrgentLatency;
+       }
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: PrefetchMode = %u\n", __func__, PrefetchMode);
+       dml_print("DML::%s: TWait = %f\n", __func__, TWait);
+#endif
+       return TWait;
+} // CalculateTWait
+
+
+/// @brief Calculate the "starting point" for prefetch calculation
+///  if AllowForPStateChangeOrStutterInVBlank is set as a particular requirement, then the mode evalulation
+///  will only be done at the given mode. If no specific requirement (i.e. *_if_possible), then will just go from
+///  try all the prefetch mode in decreasing order of "difficulty" (start from 0 which means all power saving
+///  features).
+static void CalculatePrefetchMode(
+               enum dml_prefetch_modes AllowForPStateChangeOrStutterInVBlank,
+               dml_uint_t *MinPrefetchMode,
+               dml_uint_t *MaxPrefetchMode)
+{
+       if (AllowForPStateChangeOrStutterInVBlank == dml_prefetch_support_uclk_fclk_and_stutter_if_possible) {
+               *MinPrefetchMode = 0;   // consider all pwr saving features
+               *MaxPrefetchMode = 3;   // consider just urgent latency
+       } else {
+               if (AllowForPStateChangeOrStutterInVBlank == dml_prefetch_support_none) {
+                       *MinPrefetchMode = 3;
+               } else if (AllowForPStateChangeOrStutterInVBlank == dml_prefetch_support_stutter) {
+                       *MinPrefetchMode = 2;
+               } else if (AllowForPStateChangeOrStutterInVBlank == dml_prefetch_support_fclk_and_stutter) {
+                       *MinPrefetchMode = 1;
+               } else if (AllowForPStateChangeOrStutterInVBlank == dml_prefetch_support_uclk_fclk_and_stutter) {
+                       *MinPrefetchMode = 0;
+               } else {
+                       dml_print("ERROR: Invalid AllowForPStateChangeOrStutterInVBlank setting! val=%u\n", AllowForPStateChangeOrStutterInVBlank);
+                       ASSERT(0);
+               }
+               *MaxPrefetchMode = *MinPrefetchMode;
+       }
+} // CalculatePrefetchMode
+
+static dml_float_t CalculateWriteBackDISPCLK(
+               enum dml_source_format_class WritebackPixelFormat,
+               dml_float_t PixelClock,
+               dml_float_t WritebackHRatio,
+               dml_float_t WritebackVRatio,
+               dml_uint_t WritebackHTaps,
+               dml_uint_t WritebackVTaps,
+               dml_uint_t WritebackSourceWidth,
+               dml_uint_t WritebackDestinationWidth,
+               dml_uint_t HTotal,
+               dml_uint_t WritebackLineBufferSize,
+               dml_float_t DISPCLKDPPCLKVCOSpeed)
+{
+       dml_float_t DISPCLK_H, DISPCLK_V, DISPCLK_HB;
+
+       DISPCLK_H = PixelClock * dml_ceil(WritebackHTaps / 8.0, 1) / WritebackHRatio;
+       DISPCLK_V = PixelClock * (WritebackVTaps * dml_ceil(WritebackDestinationWidth / 6.0, 1) + 8.0) / (dml_float_t) HTotal;
+       DISPCLK_HB = PixelClock * WritebackVTaps * (WritebackDestinationWidth * WritebackVTaps - WritebackLineBufferSize / 57.0) / 6.0 / (dml_float_t) WritebackSourceWidth;
+       return RoundToDFSGranularity(dml_max3(DISPCLK_H, DISPCLK_V, DISPCLK_HB), 1, DISPCLKDPPCLKVCOSpeed);
+}
+
+static dml_float_t CalculateWriteBackDelay(
+               enum dml_source_format_class WritebackPixelFormat,
+               dml_float_t WritebackHRatio,
+               dml_float_t WritebackVRatio,
+               dml_uint_t WritebackVTaps,
+               dml_uint_t WritebackDestinationWidth,
+               dml_uint_t WritebackDestinationHeight,
+               dml_uint_t WritebackSourceHeight,
+               dml_uint_t HTotal)
+{
+       dml_float_t CalculateWriteBackDelay;
+       dml_float_t Line_length;
+       dml_float_t Output_lines_last_notclamped;
+       dml_float_t WritebackVInit;
+
+       WritebackVInit = (WritebackVRatio + WritebackVTaps + 1) / 2;
+       Line_length = dml_max((dml_float_t) WritebackDestinationWidth, dml_ceil((dml_float_t)WritebackDestinationWidth / 6.0, 1.0) * WritebackVTaps);
+       Output_lines_last_notclamped = WritebackDestinationHeight - 1 - dml_ceil(((dml_float_t)WritebackSourceHeight - (dml_float_t) WritebackVInit) / (dml_float_t)WritebackVRatio, 1.0);
+       if (Output_lines_last_notclamped < 0) {
+               CalculateWriteBackDelay = 0;
+       } else {
+               CalculateWriteBackDelay = Output_lines_last_notclamped * Line_length + (HTotal - WritebackDestinationWidth) + 80;
+       }
+       return CalculateWriteBackDelay;
+}
+
+static void CalculateVUpdateAndDynamicMetadataParameters(
+               dml_uint_t MaxInterDCNTileRepeaters,
+               dml_float_t Dppclk,
+               dml_float_t Dispclk,
+               dml_float_t DCFClkDeepSleep,
+               dml_float_t PixelClock,
+               dml_uint_t HTotal,
+               dml_uint_t VBlank,
+               dml_uint_t DynamicMetadataTransmittedBytes,
+               dml_uint_t DynamicMetadataLinesBeforeActiveRequired,
+               dml_uint_t InterlaceEnable,
+               dml_bool_t ProgressiveToInterlaceUnitInOPP,
+
+               // Output
+               dml_float_t *TSetup,
+               dml_float_t *Tdmbf,
+               dml_float_t *Tdmec,
+               dml_float_t *Tdmsks,
+               dml_uint_t *VUpdateOffsetPix,
+               dml_uint_t *VUpdateWidthPix,
+               dml_uint_t *VReadyOffsetPix)
+{
+       dml_float_t TotalRepeaterDelayTime;
+       TotalRepeaterDelayTime = MaxInterDCNTileRepeaters * (2 / Dppclk + 3 / Dispclk);
+       *VUpdateWidthPix = (dml_uint_t)(dml_ceil((14.0 / DCFClkDeepSleep + 12.0 / Dppclk + TotalRepeaterDelayTime) * PixelClock, 1.0));
+       *VReadyOffsetPix = (dml_uint_t)(dml_ceil(dml_max(150.0 / Dppclk, TotalRepeaterDelayTime + 20.0 / DCFClkDeepSleep + 10.0 / Dppclk) * PixelClock, 1.0));
+       *VUpdateOffsetPix = (dml_uint_t)(dml_ceil(HTotal / 4.0, 1.0));
+       *TSetup = (*VUpdateOffsetPix + *VUpdateWidthPix + *VReadyOffsetPix) / PixelClock;
+       *Tdmbf = DynamicMetadataTransmittedBytes / 4.0 / Dispclk;
+       *Tdmec = HTotal / PixelClock;
+
+       if (DynamicMetadataLinesBeforeActiveRequired == 0) {
+               *Tdmsks = VBlank * HTotal / PixelClock / 2.0;
+       } else {
+               *Tdmsks = DynamicMetadataLinesBeforeActiveRequired * HTotal / PixelClock;
+       }
+       if (InterlaceEnable == 1 && ProgressiveToInterlaceUnitInOPP == false) {
+               *Tdmsks = *Tdmsks / 2;
+       }
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: DynamicMetadataLinesBeforeActiveRequired = %u\n", __func__, DynamicMetadataLinesBeforeActiveRequired);
+       dml_print("DML::%s: VBlank = %u\n", __func__, VBlank);
+       dml_print("DML::%s: HTotal = %u\n", __func__, HTotal);
+       dml_print("DML::%s: PixelClock = %f\n", __func__, PixelClock);
+       dml_print("DML::%s: Dppclk = %f\n", __func__, Dppclk);
+       dml_print("DML::%s: DCFClkDeepSleep = %f\n", __func__, DCFClkDeepSleep);
+       dml_print("DML::%s: MaxInterDCNTileRepeaters = %u\n", __func__, MaxInterDCNTileRepeaters);
+       dml_print("DML::%s: TotalRepeaterDelayTime = %f\n", __func__, TotalRepeaterDelayTime);
+
+       dml_print("DML::%s: VUpdateWidthPix = %u\n", __func__, *VUpdateWidthPix);
+       dml_print("DML::%s: VReadyOffsetPix = %u\n", __func__, *VReadyOffsetPix);
+       dml_print("DML::%s: VUpdateOffsetPix = %u\n", __func__, *VUpdateOffsetPix);
+
+       dml_print("DML::%s: Tdmsks = %f\n", __func__, *Tdmsks);
+#endif
+}
+
+static void CalculateRowBandwidth(
+               dml_bool_t GPUVMEnable,
+               enum dml_source_format_class SourcePixelFormat,
+               dml_float_t VRatio,
+               dml_float_t VRatioChroma,
+               dml_bool_t DCCEnable,
+               dml_float_t LineTime,
+               dml_uint_t MetaRowByteLuma,
+               dml_uint_t MetaRowByteChroma,
+               dml_uint_t meta_row_height_luma,
+               dml_uint_t meta_row_height_chroma,
+               dml_uint_t PixelPTEBytesPerRowLuma,
+               dml_uint_t PixelPTEBytesPerRowChroma,
+               dml_uint_t dpte_row_height_luma,
+               dml_uint_t dpte_row_height_chroma,
+               // Output
+               dml_float_t *meta_row_bw,
+               dml_float_t *dpte_row_bw)
+{
+       if (DCCEnable != true) {
+               *meta_row_bw = 0;
+       } else if (SourcePixelFormat == dml_420_8 || SourcePixelFormat == dml_420_10 || SourcePixelFormat == dml_420_12 || SourcePixelFormat == dml_rgbe_alpha) {
+               *meta_row_bw = VRatio * MetaRowByteLuma / (meta_row_height_luma * LineTime)
+                               + VRatioChroma * MetaRowByteChroma
+                                       / (meta_row_height_chroma * LineTime);
+       } else {
+               *meta_row_bw = VRatio * MetaRowByteLuma / (meta_row_height_luma * LineTime);
+       }
+
+       if (GPUVMEnable != true) {
+               *dpte_row_bw = 0;
+       } else if (SourcePixelFormat == dml_420_8 || SourcePixelFormat == dml_420_10 || SourcePixelFormat == dml_420_12 || SourcePixelFormat == dml_rgbe_alpha) {
+               *dpte_row_bw = VRatio * PixelPTEBytesPerRowLuma / (dpte_row_height_luma * LineTime)
+                               + VRatioChroma * PixelPTEBytesPerRowChroma
+                                       / (dpte_row_height_chroma * LineTime);
+       } else {
+               *dpte_row_bw = VRatio * PixelPTEBytesPerRowLuma / (dpte_row_height_luma * LineTime);
+       }
+}
+
+/// @brief Determine immediate flip schedule given bw remaining after considering the prefetch schedule
+/// @param BandwidthAvailableForImmediateFlip Bandwidth available for iflip for all planes
+static void CalculateFlipSchedule(
+               dml_float_t HostVMInefficiencyFactor,
+               dml_float_t UrgentExtraLatency,
+               dml_float_t UrgentLatency,
+               dml_uint_t GPUVMMaxPageTableLevels,
+               dml_bool_t HostVMEnable,
+               dml_uint_t HostVMMaxNonCachedPageTableLevels,
+               dml_bool_t GPUVMEnable,
+               dml_uint_t HostVMMinPageSize,
+               dml_float_t PDEAndMetaPTEBytesPerFrame,
+               dml_float_t MetaRowBytes,
+               dml_float_t DPTEBytesPerRow,
+               dml_float_t BandwidthAvailableForImmediateFlip,
+               dml_uint_t TotImmediateFlipBytes,
+               enum dml_source_format_class SourcePixelFormat,
+               dml_float_t LineTime,
+               dml_float_t VRatio,
+               dml_float_t VRatioChroma,
+               dml_float_t Tno_bw,
+               dml_bool_t DCCEnable,
+               dml_uint_t dpte_row_height,
+               dml_uint_t meta_row_height,
+               dml_uint_t dpte_row_height_chroma,
+               dml_uint_t meta_row_height_chroma,
+               dml_bool_t use_one_row_for_frame_flip,
+
+               // Output
+               dml_float_t *DestinationLinesToRequestVMInImmediateFlip,
+               dml_float_t *DestinationLinesToRequestRowInImmediateFlip,
+               dml_float_t *final_flip_bw,
+               dml_bool_t *ImmediateFlipSupportedForPipe)
+{
+       dml_float_t min_row_time = 0.0;
+       dml_uint_t HostVMDynamicLevelsTrips = 0;
+       dml_float_t TimeForFetchingMetaPTEImmediateFlip = 0;
+       dml_float_t TimeForFetchingRowInVBlankImmediateFlip = 0;
+       dml_float_t ImmediateFlipBW = 0; // @brief The immediate flip bandwidth for this pipe
+
+       if (GPUVMEnable == true && HostVMEnable == true) {
+               HostVMDynamicLevelsTrips = HostVMMaxNonCachedPageTableLevels;
+       } else {
+               HostVMDynamicLevelsTrips = 0;
+       }
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: TotImmediateFlipBytes = %u\n", __func__, TotImmediateFlipBytes);
+       dml_print("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, HostVMInefficiencyFactor);
+       dml_print("DML::%s: UrgentLatency = %f\n", __func__, UrgentLatency);
+       dml_print("DML::%s: BandwidthAvailableForImmediateFlip = %f\n", __func__, BandwidthAvailableForImmediateFlip);
+#endif
+
+       if (TotImmediateFlipBytes > 0) {
+               if (use_one_row_for_frame_flip) {
+                       ImmediateFlipBW = (PDEAndMetaPTEBytesPerFrame + MetaRowBytes + 2.0 * DPTEBytesPerRow) * BandwidthAvailableForImmediateFlip / (dml_float_t) TotImmediateFlipBytes;
+               } else {
+                       ImmediateFlipBW = (PDEAndMetaPTEBytesPerFrame + MetaRowBytes + DPTEBytesPerRow) * BandwidthAvailableForImmediateFlip / (dml_float_t) TotImmediateFlipBytes;
+               }
+               if (GPUVMEnable == true) {
+                       TimeForFetchingMetaPTEImmediateFlip = dml_max3(Tno_bw + PDEAndMetaPTEBytesPerFrame * HostVMInefficiencyFactor / ImmediateFlipBW,
+                                                                                                               UrgentExtraLatency + UrgentLatency * (GPUVMMaxPageTableLevels * (HostVMDynamicLevelsTrips + 1) - 1),
+                                                                                                               LineTime / 4.0);
+               } else {
+                       TimeForFetchingMetaPTEImmediateFlip = 0;
+               }
+               if ((GPUVMEnable == true || DCCEnable == true)) {
+                       TimeForFetchingRowInVBlankImmediateFlip = dml_max3((MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / ImmediateFlipBW, UrgentLatency * (HostVMDynamicLevelsTrips + 1), LineTime / 4.0);
+               } else {
+                       TimeForFetchingRowInVBlankImmediateFlip = 0;
+               }
+
+               *DestinationLinesToRequestVMInImmediateFlip = dml_ceil(4.0 * (TimeForFetchingMetaPTEImmediateFlip / LineTime), 1.0) / 4.0;
+               *DestinationLinesToRequestRowInImmediateFlip = dml_ceil(4.0 * (TimeForFetchingRowInVBlankImmediateFlip / LineTime), 1.0) / 4.0;
+
+               if (GPUVMEnable == true) {
+                       *final_flip_bw = dml_max(PDEAndMetaPTEBytesPerFrame * HostVMInefficiencyFactor / (*DestinationLinesToRequestVMInImmediateFlip * LineTime),
+                                                               (MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / (*DestinationLinesToRequestRowInImmediateFlip * LineTime));
+               } else if ((GPUVMEnable == true || DCCEnable == true)) {
+                       *final_flip_bw = (MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / (*DestinationLinesToRequestRowInImmediateFlip * LineTime);
+               } else {
+                       *final_flip_bw = 0;
+               }
+       } else {
+               TimeForFetchingMetaPTEImmediateFlip = 0;
+               TimeForFetchingRowInVBlankImmediateFlip = 0;
+               *DestinationLinesToRequestVMInImmediateFlip = 0;
+               *DestinationLinesToRequestRowInImmediateFlip = 0;
+               *final_flip_bw = 0;
+       }
+
+       if (SourcePixelFormat == dml_420_8 || SourcePixelFormat == dml_420_10 || SourcePixelFormat == dml_rgbe_alpha) {
+               if (GPUVMEnable == true && DCCEnable != true) {
+                       min_row_time = dml_min(dpte_row_height * LineTime / VRatio, dpte_row_height_chroma * LineTime / VRatioChroma);
+               } else if (GPUVMEnable != true && DCCEnable == true) {
+                       min_row_time = dml_min(meta_row_height * LineTime / VRatio, meta_row_height_chroma * LineTime / VRatioChroma);
+               } else {
+                       min_row_time = dml_min4(dpte_row_height * LineTime / VRatio, meta_row_height * LineTime / VRatio, dpte_row_height_chroma * LineTime / VRatioChroma, meta_row_height_chroma * LineTime / VRatioChroma);
+               }
+       } else {
+               if (GPUVMEnable == true && DCCEnable != true) {
+                       min_row_time = dpte_row_height * LineTime / VRatio;
+               } else if (GPUVMEnable != true && DCCEnable == true) {
+                       min_row_time = meta_row_height * LineTime / VRatio;
+               } else {
+                       min_row_time = dml_min(dpte_row_height * LineTime / VRatio, meta_row_height * LineTime / VRatio);
+               }
+       }
+
+       if (*DestinationLinesToRequestVMInImmediateFlip >= 32 || *DestinationLinesToRequestRowInImmediateFlip >= 16 || TimeForFetchingMetaPTEImmediateFlip + 2 * TimeForFetchingRowInVBlankImmediateFlip > min_row_time) {
+               *ImmediateFlipSupportedForPipe = false;
+       } else {
+               *ImmediateFlipSupportedForPipe = true;
+       }
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: GPUVMEnable = %u\n", __func__, GPUVMEnable);
+       dml_print("DML::%s: DCCEnable = %u\n", __func__, DCCEnable);
+
+       dml_print("DML::%s: MetaRowBytes = %f\n", __func__, MetaRowBytes);
+       dml_print("DML::%s: DPTEBytesPerRow = %f\n", __func__, DPTEBytesPerRow);
+       dml_print("DML::%s: BandwidthAvailableForImmediateFlip = %f\n", __func__, BandwidthAvailableForImmediateFlip);
+       dml_print("DML::%s: TotImmediateFlipBytes = %u\n", __func__, TotImmediateFlipBytes);
+       dml_print("DML::%s: ImmediateFlipBW = %f\n", __func__, ImmediateFlipBW);
+       dml_print("DML::%s: PDEAndMetaPTEBytesPerFrame = %f\n", __func__, PDEAndMetaPTEBytesPerFrame);
+       dml_print("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, HostVMInefficiencyFactor);
+       dml_print("DML::%s: LineTime = %f\n", __func__, LineTime);
+       dml_print("DML::%s: final_flip_bw = %f\n", __func__, *final_flip_bw);
+
+       dml_print("DML::%s: DestinationLinesToRequestVMInImmediateFlip = %f\n", __func__, *DestinationLinesToRequestVMInImmediateFlip);
+       dml_print("DML::%s: DestinationLinesToRequestRowInImmediateFlip = %f\n", __func__, *DestinationLinesToRequestRowInImmediateFlip);
+       dml_print("DML::%s: TimeForFetchingMetaPTEImmediateFlip = %f\n", __func__, TimeForFetchingMetaPTEImmediateFlip);
+       dml_print("DML::%s: TimeForFetchingRowInVBlankImmediateFlip = %f\n", __func__, TimeForFetchingRowInVBlankImmediateFlip);
+       dml_print("DML::%s: min_row_time = %f\n", __func__, min_row_time);
+       dml_print("DML::%s: ImmediateFlipSupportedForPipe = %u\n", __func__, *ImmediateFlipSupportedForPipe);
+#endif
+} // CalculateFlipSchedule
+
+static dml_float_t RoundToDFSGranularity(dml_float_t Clock, dml_bool_t round_up, dml_float_t VCOSpeed)
+{
+       if (Clock <= 0.0)
+               return 0.0;
+       else {
+               if (round_up)
+                       return VCOSpeed * 4.0 / dml_floor(VCOSpeed * 4.0 / Clock, 1.0);
+               else
+                       return VCOSpeed * 4.0 / dml_ceil(VCOSpeed * 4.0 / Clock, 1.0);
+       }
+}
+
+static void CalculateDCCConfiguration(
+               dml_bool_t DCCEnabled,
+               dml_bool_t DCCProgrammingAssumesScanDirectionUnknown,
+               enum dml_source_format_class SourcePixelFormat,
+               dml_uint_t SurfaceWidthLuma,
+               dml_uint_t SurfaceWidthChroma,
+               dml_uint_t SurfaceHeightLuma,
+               dml_uint_t SurfaceHeightChroma,
+               dml_uint_t nomDETInKByte,
+               dml_uint_t RequestHeight256ByteLuma,
+               dml_uint_t RequestHeight256ByteChroma,
+               enum dml_swizzle_mode TilingFormat,
+               dml_uint_t BytePerPixelY,
+               dml_uint_t BytePerPixelC,
+               dml_float_t BytePerPixelDETY,
+               dml_float_t BytePerPixelDETC,
+               enum dml_rotation_angle SourceScan,
+               // Output
+               dml_uint_t *MaxUncompressedBlockLuma,
+               dml_uint_t *MaxUncompressedBlockChroma,
+               dml_uint_t *MaxCompressedBlockLuma,
+               dml_uint_t *MaxCompressedBlockChroma,
+               dml_uint_t *IndependentBlockLuma,
+               dml_uint_t *IndependentBlockChroma)
+{
+       dml_uint_t DETBufferSizeForDCC = nomDETInKByte * 1024;
+
+       dml_uint_t yuv420;
+       dml_uint_t horz_div_l;
+       dml_uint_t horz_div_c;
+       dml_uint_t vert_div_l;
+       dml_uint_t vert_div_c;
+
+       dml_uint_t swath_buf_size;
+       dml_float_t detile_buf_vp_horz_limit;
+       dml_float_t detile_buf_vp_vert_limit;
+
+       dml_uint_t MAS_vp_horz_limit;
+       dml_uint_t MAS_vp_vert_limit;
+       dml_uint_t max_vp_horz_width;
+       dml_uint_t max_vp_vert_height;
+       dml_uint_t eff_surf_width_l;
+       dml_uint_t eff_surf_width_c;
+       dml_uint_t eff_surf_height_l;
+       dml_uint_t eff_surf_height_c;
+
+       dml_uint_t full_swath_bytes_horz_wc_l;
+       dml_uint_t full_swath_bytes_horz_wc_c;
+       dml_uint_t full_swath_bytes_vert_wc_l;
+       dml_uint_t full_swath_bytes_vert_wc_c;
+
+       dml_uint_t req128_horz_wc_l;
+       dml_uint_t req128_horz_wc_c;
+       dml_uint_t req128_vert_wc_l;
+       dml_uint_t req128_vert_wc_c;
+
+       dml_uint_t   segment_order_horz_contiguous_luma;
+       dml_uint_t   segment_order_horz_contiguous_chroma;
+       dml_uint_t   segment_order_vert_contiguous_luma;
+       dml_uint_t   segment_order_vert_contiguous_chroma;
+
+       typedef enum{
+               REQ_256Bytes,
+               REQ_128BytesNonContiguous,
+               REQ_128BytesContiguous,
+               REQ_NA
+       } RequestType;
+
+       RequestType   RequestLuma;
+       RequestType   RequestChroma;
+
+       yuv420 = ((SourcePixelFormat == dml_420_8 || SourcePixelFormat == dml_420_10 || SourcePixelFormat == dml_420_12) ? 1 : 0);
+       horz_div_l = 1;
+       horz_div_c = 1;
+       vert_div_l = 1;
+       vert_div_c = 1;
+
+       if (BytePerPixelY == 1)
+               vert_div_l = 0;
+       if (BytePerPixelC == 1)
+               vert_div_c = 0;
+
+       if (BytePerPixelC == 0) {
+               swath_buf_size = DETBufferSizeForDCC / 2 - 2 * 256;
+               detile_buf_vp_horz_limit = (dml_float_t) swath_buf_size / ((dml_float_t) RequestHeight256ByteLuma * BytePerPixelY / (1 + horz_div_l));
+               detile_buf_vp_vert_limit = (dml_float_t) swath_buf_size / (256.0 / RequestHeight256ByteLuma / (1 + vert_div_l));
+       } else {
+               swath_buf_size = DETBufferSizeForDCC / 2 - 2 * 2 * 256;
+               detile_buf_vp_horz_limit = (dml_float_t) swath_buf_size / ((dml_float_t) RequestHeight256ByteLuma * BytePerPixelY / (1 + horz_div_l) + (dml_float_t) RequestHeight256ByteChroma * BytePerPixelC / (1 + horz_div_c) / (1 + yuv420));
+               detile_buf_vp_vert_limit = (dml_float_t) swath_buf_size / (256.0 / RequestHeight256ByteLuma / (1 + vert_div_l) + 256.0 / RequestHeight256ByteChroma / (1 + vert_div_c) / (1 + yuv420));
+       }
+
+       if (SourcePixelFormat == dml_420_10) {
+               detile_buf_vp_horz_limit = 1.5 * detile_buf_vp_horz_limit;
+               detile_buf_vp_vert_limit = 1.5 * detile_buf_vp_vert_limit;
+       }
+
+       detile_buf_vp_horz_limit = dml_floor(detile_buf_vp_horz_limit - 1, 16);
+       detile_buf_vp_vert_limit = dml_floor(detile_buf_vp_vert_limit - 1, 16);
+
+       MAS_vp_horz_limit = SourcePixelFormat == dml_rgbe_alpha ? 3840 : 6144;
+       MAS_vp_vert_limit = SourcePixelFormat == dml_rgbe_alpha ? 3840 : (BytePerPixelY == 8 ? 3072 : 6144);
+       max_vp_horz_width = (dml_uint_t)(dml_min((dml_float_t) MAS_vp_horz_limit, detile_buf_vp_horz_limit));
+       max_vp_vert_height = (dml_uint_t)(dml_min((dml_float_t) MAS_vp_vert_limit, detile_buf_vp_vert_limit));
+       eff_surf_width_l = (SurfaceWidthLuma > max_vp_horz_width ? max_vp_horz_width : SurfaceWidthLuma);
+       eff_surf_width_c = eff_surf_width_l / (1 + yuv420);
+       eff_surf_height_l = (SurfaceHeightLuma > max_vp_vert_height ? max_vp_vert_height : SurfaceHeightLuma);
+       eff_surf_height_c = eff_surf_height_l / (1 + yuv420);
+
+       full_swath_bytes_horz_wc_l = eff_surf_width_l * RequestHeight256ByteLuma * BytePerPixelY;
+       full_swath_bytes_vert_wc_l = eff_surf_height_l * 256 / RequestHeight256ByteLuma;
+       if (BytePerPixelC > 0) {
+               full_swath_bytes_horz_wc_c = eff_surf_width_c * RequestHeight256ByteChroma * BytePerPixelC;
+               full_swath_bytes_vert_wc_c = eff_surf_height_c * 256 / RequestHeight256ByteChroma;
+       } else {
+               full_swath_bytes_horz_wc_c = 0;
+               full_swath_bytes_vert_wc_c = 0;
+       }
+
+       if (SourcePixelFormat == dml_420_10) {
+               full_swath_bytes_horz_wc_l = (dml_uint_t)(dml_ceil((dml_float_t) full_swath_bytes_horz_wc_l * 2.0 / 3.0, 256.0));
+               full_swath_bytes_horz_wc_c = (dml_uint_t)(dml_ceil((dml_float_t) full_swath_bytes_horz_wc_c * 2.0 / 3.0, 256.0));
+               full_swath_bytes_vert_wc_l = (dml_uint_t)(dml_ceil((dml_float_t) full_swath_bytes_vert_wc_l * 2.0 / 3.0, 256.0));
+               full_swath_bytes_vert_wc_c = (dml_uint_t)(dml_ceil((dml_float_t) full_swath_bytes_vert_wc_c * 2.0 / 3.0, 256.0));
+       }
+
+       if (2 * full_swath_bytes_horz_wc_l + 2 * full_swath_bytes_horz_wc_c <= DETBufferSizeForDCC) {
+               req128_horz_wc_l = 0;
+               req128_horz_wc_c = 0;
+       } else if (full_swath_bytes_horz_wc_l < 1.5 * full_swath_bytes_horz_wc_c && 2 * full_swath_bytes_horz_wc_l + full_swath_bytes_horz_wc_c <= DETBufferSizeForDCC) {
+               req128_horz_wc_l = 0;
+               req128_horz_wc_c = 1;
+       } else if (full_swath_bytes_horz_wc_l >= 1.5 * full_swath_bytes_horz_wc_c && full_swath_bytes_horz_wc_l + 2 * full_swath_bytes_horz_wc_c <= DETBufferSizeForDCC) {
+               req128_horz_wc_l = 1;
+               req128_horz_wc_c = 0;
+       } else {
+               req128_horz_wc_l = 1;
+               req128_horz_wc_c = 1;
+       }
+
+       if (2 * full_swath_bytes_vert_wc_l + 2 * full_swath_bytes_vert_wc_c <= DETBufferSizeForDCC) {
+               req128_vert_wc_l = 0;
+               req128_vert_wc_c = 0;
+       } else if (full_swath_bytes_vert_wc_l < 1.5 * full_swath_bytes_vert_wc_c && 2 * full_swath_bytes_vert_wc_l + full_swath_bytes_vert_wc_c <= DETBufferSizeForDCC) {
+               req128_vert_wc_l = 0;
+               req128_vert_wc_c = 1;
+       } else if (full_swath_bytes_vert_wc_l >= 1.5 * full_swath_bytes_vert_wc_c && full_swath_bytes_vert_wc_l + 2 * full_swath_bytes_vert_wc_c <= DETBufferSizeForDCC) {
+               req128_vert_wc_l = 1;
+               req128_vert_wc_c = 0;
+       } else {
+               req128_vert_wc_l = 1;
+               req128_vert_wc_c = 1;
+       }
+
+       if (BytePerPixelY == 2) {
+               segment_order_horz_contiguous_luma = 0;
+               segment_order_vert_contiguous_luma = 1;
+       } else {
+               segment_order_horz_contiguous_luma = 1;
+               segment_order_vert_contiguous_luma = 0;
+       }
+
+       if (BytePerPixelC == 2) {
+               segment_order_horz_contiguous_chroma = 0;
+               segment_order_vert_contiguous_chroma = 1;
+       } else {
+               segment_order_horz_contiguous_chroma = 1;
+               segment_order_vert_contiguous_chroma = 0;
+       }
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: DCCEnabled = %u\n", __func__, DCCEnabled);
+       dml_print("DML::%s: nomDETInKByte = %u\n", __func__, nomDETInKByte);
+       dml_print("DML::%s: DETBufferSizeForDCC = %u\n", __func__, DETBufferSizeForDCC);
+       dml_print("DML::%s: req128_horz_wc_l = %u\n", __func__, req128_horz_wc_l);
+       dml_print("DML::%s: req128_horz_wc_c = %u\n", __func__, req128_horz_wc_c);
+       dml_print("DML::%s: full_swath_bytes_horz_wc_l = %u\n", __func__, full_swath_bytes_horz_wc_l);
+       dml_print("DML::%s: full_swath_bytes_vert_wc_c = %u\n", __func__, full_swath_bytes_vert_wc_c);
+       dml_print("DML::%s: segment_order_horz_contiguous_luma = %u\n", __func__, segment_order_horz_contiguous_luma);
+       dml_print("DML::%s: segment_order_horz_contiguous_chroma = %u\n", __func__, segment_order_horz_contiguous_chroma);
+#endif
+
+       if (DCCProgrammingAssumesScanDirectionUnknown == true) {
+               if (req128_horz_wc_l == 0 && req128_vert_wc_l == 0) {
+                       RequestLuma = REQ_256Bytes;
+               } else if ((req128_horz_wc_l == 1 && segment_order_horz_contiguous_luma == 0) || (req128_vert_wc_l == 1 && segment_order_vert_contiguous_luma == 0)) {
+                       RequestLuma = REQ_128BytesNonContiguous;
+               } else {
+                       RequestLuma = REQ_128BytesContiguous;
+               }
+               if (req128_horz_wc_c == 0 && req128_vert_wc_c == 0) {
+                       RequestChroma = REQ_256Bytes;
+               } else if ((req128_horz_wc_c == 1 && segment_order_horz_contiguous_chroma == 0) || (req128_vert_wc_c == 1 && segment_order_vert_contiguous_chroma == 0)) {
+                       RequestChroma = REQ_128BytesNonContiguous;
+               } else {
+                       RequestChroma = REQ_128BytesContiguous;
+               }
+       } else if (!dml_is_vertical_rotation(SourceScan)) {
+               if (req128_horz_wc_l == 0) {
+                       RequestLuma = REQ_256Bytes;
+               } else if (segment_order_horz_contiguous_luma == 0) {
+                       RequestLuma = REQ_128BytesNonContiguous;
+               } else {
+                       RequestLuma = REQ_128BytesContiguous;
+               }
+               if (req128_horz_wc_c == 0) {
+                       RequestChroma = REQ_256Bytes;
+               } else if (segment_order_horz_contiguous_chroma == 0) {
+                       RequestChroma = REQ_128BytesNonContiguous;
+               } else {
+                       RequestChroma = REQ_128BytesContiguous;
+               }
+       } else {
+               if (req128_vert_wc_l == 0) {
+                       RequestLuma = REQ_256Bytes;
+               } else if (segment_order_vert_contiguous_luma == 0) {
+                       RequestLuma = REQ_128BytesNonContiguous;
+               } else {
+                       RequestLuma = REQ_128BytesContiguous;
+               }
+               if (req128_vert_wc_c == 0) {
+                       RequestChroma = REQ_256Bytes;
+               } else if (segment_order_vert_contiguous_chroma == 0) {
+                       RequestChroma = REQ_128BytesNonContiguous;
+               } else {
+                       RequestChroma = REQ_128BytesContiguous;
+               }
+       }
+
+       if (RequestLuma == REQ_256Bytes) {
+               *MaxUncompressedBlockLuma = 256;
+               *MaxCompressedBlockLuma = 256;
+               *IndependentBlockLuma = 0;
+       } else if (RequestLuma == REQ_128BytesContiguous) {
+               *MaxUncompressedBlockLuma = 256;
+               *MaxCompressedBlockLuma = 128;
+               *IndependentBlockLuma = 128;
+       } else {
+               *MaxUncompressedBlockLuma = 256;
+               *MaxCompressedBlockLuma = 64;
+               *IndependentBlockLuma = 64;
+       }
+
+       if (RequestChroma == REQ_256Bytes) {
+               *MaxUncompressedBlockChroma = 256;
+               *MaxCompressedBlockChroma = 256;
+               *IndependentBlockChroma = 0;
+       } else if (RequestChroma == REQ_128BytesContiguous) {
+               *MaxUncompressedBlockChroma = 256;
+               *MaxCompressedBlockChroma = 128;
+               *IndependentBlockChroma = 128;
+       } else {
+               *MaxUncompressedBlockChroma = 256;
+               *MaxCompressedBlockChroma = 64;
+               *IndependentBlockChroma = 64;
+       }
+
+       if (DCCEnabled != true || BytePerPixelC == 0) {
+               *MaxUncompressedBlockChroma = 0;
+               *MaxCompressedBlockChroma = 0;
+               *IndependentBlockChroma = 0;
+       }
+
+       if (DCCEnabled != true) {
+               *MaxUncompressedBlockLuma = 0;
+               *MaxCompressedBlockLuma = 0;
+               *IndependentBlockLuma = 0;
+       }
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: MaxUncompressedBlockLuma = %u\n", __func__, *MaxUncompressedBlockLuma);
+       dml_print("DML::%s: MaxCompressedBlockLuma = %u\n", __func__, *MaxCompressedBlockLuma);
+       dml_print("DML::%s: IndependentBlockLuma = %u\n", __func__, *IndependentBlockLuma);
+       dml_print("DML::%s: MaxUncompressedBlockChroma = %u\n", __func__, *MaxUncompressedBlockChroma);
+       dml_print("DML::%s: MaxCompressedBlockChroma = %u\n", __func__, *MaxCompressedBlockChroma);
+       dml_print("DML::%s: IndependentBlockChroma = %u\n", __func__, *IndependentBlockChroma);
+#endif
+
+} // CalculateDCCConfiguration
+
+static dml_uint_t CalculatePrefetchSourceLines(
+               dml_float_t VRatio,
+               dml_uint_t VTaps,
+               dml_bool_t Interlace,
+               dml_bool_t ProgressiveToInterlaceUnitInOPP,
+               dml_uint_t SwathHeight,
+               enum dml_rotation_angle SourceScan,
+               dml_bool_t ViewportStationary,
+               dml_uint_t SwathWidth,
+               dml_uint_t ViewportHeight,
+               dml_uint_t ViewportXStart,
+               dml_uint_t ViewportYStart,
+
+               // Output
+               dml_uint_t *VInitPreFill,
+               dml_uint_t *MaxNumSwath)
+{
+
+       dml_uint_t vp_start_rot = 0;
+       dml_uint_t sw0_tmp = 0;
+       dml_uint_t MaxPartialSwath = 0;
+       dml_float_t numLines = 0;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: VRatio = %f\n", __func__, VRatio);
+       dml_print("DML::%s: VTaps = %u\n", __func__, VTaps);
+       dml_print("DML::%s: ViewportXStart = %u\n", __func__, ViewportXStart);
+       dml_print("DML::%s: ViewportYStart = %u\n", __func__, ViewportYStart);
+       dml_print("DML::%s: ViewportStationary = %u\n", __func__, ViewportStationary);
+       dml_print("DML::%s: SwathHeight = %u\n", __func__, SwathHeight);
+#endif
+       if (ProgressiveToInterlaceUnitInOPP)
+               *VInitPreFill = (dml_uint_t)(dml_floor((VRatio + (dml_float_t) VTaps + 1) / 2.0, 1));
+       else
+               *VInitPreFill = (dml_uint_t)(dml_floor((VRatio + (dml_float_t) VTaps + 1 + Interlace * 0.5 * VRatio) / 2.0, 1));
+
+       if (ViewportStationary) {
+               if (SourceScan == dml_rotation_180 || SourceScan == dml_rotation_180m) {
+                       vp_start_rot = SwathHeight - (((dml_uint_t) (ViewportYStart + ViewportHeight - 1) % SwathHeight) + 1);
+               } else if (SourceScan == dml_rotation_270 || SourceScan == dml_rotation_90m) {
+                       vp_start_rot = ViewportXStart;
+               } else if (SourceScan == dml_rotation_90 || SourceScan == dml_rotation_270m) {
+                       vp_start_rot = SwathHeight - (((dml_uint_t)(ViewportYStart + SwathWidth - 1) % SwathHeight) + 1);
+               } else {
+                       vp_start_rot = ViewportYStart;
+               }
+               sw0_tmp = SwathHeight - (vp_start_rot % SwathHeight);
+               if (sw0_tmp < *VInitPreFill) {
+                       *MaxNumSwath = (dml_uint_t)(dml_ceil((*VInitPreFill - sw0_tmp) / (dml_float_t) SwathHeight, 1) + 1);
+               } else {
+                       *MaxNumSwath = 1;
+               }
+               MaxPartialSwath = (dml_uint_t)(dml_max(1, (dml_uint_t) (vp_start_rot + *VInitPreFill - 1) % SwathHeight));
+       } else {
+               *MaxNumSwath = (dml_uint_t)(dml_ceil((*VInitPreFill - 1.0) / (dml_float_t) SwathHeight, 1) + 1);
+               if (*VInitPreFill > 1) {
+                       MaxPartialSwath = (dml_uint_t)(dml_max(1, (dml_uint_t) (*VInitPreFill - 2) % SwathHeight));
+               } else {
+                       MaxPartialSwath = (dml_uint_t)(dml_max(1, (dml_uint_t) (*VInitPreFill + SwathHeight - 2) % SwathHeight));
+               }
+       }
+       numLines = *MaxNumSwath * SwathHeight + MaxPartialSwath;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: vp_start_rot = %u\n", __func__, vp_start_rot);
+       dml_print("DML::%s: VInitPreFill = %u\n", __func__, *VInitPreFill);
+       dml_print("DML::%s: MaxPartialSwath = %u\n", __func__, MaxPartialSwath);
+       dml_print("DML::%s: MaxNumSwath = %u\n", __func__, *MaxNumSwath);
+       dml_print("DML::%s: Prefetch source lines = %3.2f\n", __func__, numLines);
+#endif
+       return (dml_uint_t)(numLines);
+
+} // CalculatePrefetchSourceLines
+
+static dml_uint_t CalculateVMAndRowBytes(
+               dml_bool_t ViewportStationary,
+               dml_bool_t DCCEnable,
+               dml_uint_t NumberOfDPPs,
+               dml_uint_t BlockHeight256Bytes,
+               dml_uint_t BlockWidth256Bytes,
+               enum dml_source_format_class SourcePixelFormat,
+               dml_uint_t SurfaceTiling,
+               dml_uint_t BytePerPixel,
+               enum dml_rotation_angle SourceScan,
+               dml_uint_t SwathWidth,
+               dml_uint_t ViewportHeight,
+               dml_uint_t ViewportXStart,
+               dml_uint_t ViewportYStart,
+               dml_bool_t GPUVMEnable,
+               dml_uint_t GPUVMMaxPageTableLevels,
+               dml_uint_t GPUVMMinPageSizeKBytes,
+               dml_uint_t PTEBufferSizeInRequests,
+               dml_uint_t Pitch,
+               dml_uint_t DCCMetaPitch,
+               dml_uint_t MacroTileWidth,
+               dml_uint_t MacroTileHeight,
+
+               // Output
+               dml_uint_t *MetaRowByte,
+               dml_uint_t *PixelPTEBytesPerRow, // for bandwidth calculation
+               dml_uint_t *PixelPTEBytesPerRowStorage, // for PTE buffer size check
+               dml_uint_t *dpte_row_width_ub,
+               dml_uint_t *dpte_row_height,
+               dml_uint_t *dpte_row_height_linear,
+               dml_uint_t *PixelPTEBytesPerRow_one_row_per_frame,
+               dml_uint_t *dpte_row_width_ub_one_row_per_frame,
+               dml_uint_t *dpte_row_height_one_row_per_frame,
+               dml_uint_t *MetaRequestWidth,
+               dml_uint_t *MetaRequestHeight,
+               dml_uint_t *meta_row_width,
+               dml_uint_t *meta_row_height,
+               dml_uint_t *PixelPTEReqWidth,
+               dml_uint_t *PixelPTEReqHeight,
+               dml_uint_t *PTERequestSize,
+               dml_uint_t *DPDE0BytesFrame,
+               dml_uint_t *MetaPTEBytesFrame)
+{
+       dml_uint_t MPDEBytesFrame;
+       dml_uint_t DCCMetaSurfaceBytes;
+       dml_uint_t ExtraDPDEBytesFrame;
+       dml_uint_t PDEAndMetaPTEBytesFrame;
+       dml_uint_t MacroTileSizeBytes;
+       dml_uint_t vp_height_meta_ub;
+       dml_uint_t vp_height_dpte_ub;
+
+       dml_uint_t PixelPTEReqWidth_linear = 0; // VBA_DELTA. VBA doesn't calculate this
+
+       *MetaRequestHeight = 8 * BlockHeight256Bytes;
+       *MetaRequestWidth = 8 * BlockWidth256Bytes;
+       if (SurfaceTiling == dml_sw_linear) {
+               *meta_row_height = 32;
+               *meta_row_width = (dml_uint_t)(dml_floor(ViewportXStart + SwathWidth + *MetaRequestWidth - 1, *MetaRequestWidth) - dml_floor(ViewportXStart, *MetaRequestWidth));
+       } else if (!dml_is_vertical_rotation(SourceScan)) {
+               *meta_row_height = *MetaRequestHeight;
+               if (ViewportStationary && NumberOfDPPs == 1) {
+                       *meta_row_width = (dml_uint_t)(dml_floor(ViewportXStart + SwathWidth + *MetaRequestWidth - 1, *MetaRequestWidth) - dml_floor(ViewportXStart, *MetaRequestWidth));
+               } else {
+                       *meta_row_width = (dml_uint_t)(dml_ceil(SwathWidth - 1, *MetaRequestWidth) + *MetaRequestWidth);
+               }
+               *MetaRowByte = (dml_uint_t)(*meta_row_width * *MetaRequestHeight * BytePerPixel / 256.0);
+       } else {
+               *meta_row_height = *MetaRequestWidth;
+               if (ViewportStationary && NumberOfDPPs == 1) {
+                       *meta_row_width = (dml_uint_t)(dml_floor(ViewportYStart + ViewportHeight + *MetaRequestHeight - 1, *MetaRequestHeight) - dml_floor(ViewportYStart, *MetaRequestHeight));
+               } else {
+                       *meta_row_width = (dml_uint_t)(dml_ceil(SwathWidth - 1, *MetaRequestHeight) + *MetaRequestHeight);
+               }
+               *MetaRowByte = (dml_uint_t)(*meta_row_width * *MetaRequestWidth * BytePerPixel / 256.0);
+       }
+
+       if (ViewportStationary && (NumberOfDPPs == 1 || !dml_is_vertical_rotation(SourceScan))) {
+               vp_height_meta_ub = (dml_uint_t)(dml_floor(ViewportYStart + ViewportHeight + 64 * BlockHeight256Bytes - 1, 64 * BlockHeight256Bytes) - dml_floor(ViewportYStart, 64 * BlockHeight256Bytes));
+       } else if (!dml_is_vertical_rotation(SourceScan)) {
+               vp_height_meta_ub = (dml_uint_t)(dml_ceil(ViewportHeight - 1, 64 * BlockHeight256Bytes) + 64 * BlockHeight256Bytes);
+       } else {
+               vp_height_meta_ub = (dml_uint_t)(dml_ceil(SwathWidth - 1, 64 * BlockHeight256Bytes) + 64 * BlockHeight256Bytes);
+       }
+
+       DCCMetaSurfaceBytes = (dml_uint_t)(DCCMetaPitch * vp_height_meta_ub * BytePerPixel / 256.0);
+
+       if (GPUVMEnable == true) {
+               *MetaPTEBytesFrame = (dml_uint_t)((dml_ceil((dml_float_t) (DCCMetaSurfaceBytes - 4.0 * 1024.0) / (8 * 4.0 * 1024), 1) + 1) * 64);
+               MPDEBytesFrame = 128 * (GPUVMMaxPageTableLevels - 1);
+       } else {
+               *MetaPTEBytesFrame = 0;
+               MPDEBytesFrame = 0;
+       }
+
+       if (DCCEnable != true) {
+               *MetaPTEBytesFrame = 0;
+               MPDEBytesFrame = 0;
+               *MetaRowByte = 0;
+       }
+
+       MacroTileSizeBytes = MacroTileWidth * BytePerPixel * MacroTileHeight;
+
+       if (ViewportStationary && (NumberOfDPPs == 1 || !dml_is_vertical_rotation(SourceScan))) {
+               vp_height_dpte_ub = (dml_uint_t)(dml_floor(ViewportYStart + ViewportHeight + MacroTileHeight - 1, MacroTileHeight) - dml_floor(ViewportYStart, MacroTileHeight));
+       } else if (!dml_is_vertical_rotation(SourceScan)) {
+               vp_height_dpte_ub = (dml_uint_t)(dml_ceil(ViewportHeight - 1, MacroTileHeight) + MacroTileHeight);
+       } else {
+               vp_height_dpte_ub = (dml_uint_t)(dml_ceil(SwathWidth - 1, MacroTileHeight) + MacroTileHeight);
+       }
+
+       if (GPUVMEnable == true && GPUVMMaxPageTableLevels > 1) {
+               *DPDE0BytesFrame = (dml_uint_t)(64 * (dml_ceil((dml_float_t) (Pitch * vp_height_dpte_ub * BytePerPixel - MacroTileSizeBytes) / (dml_float_t) (8 * 2097152), 1) + 1));
+               ExtraDPDEBytesFrame = 128 * (GPUVMMaxPageTableLevels - 2);
+       } else {
+               *DPDE0BytesFrame = 0;
+               ExtraDPDEBytesFrame = 0;
+       }
+
+       PDEAndMetaPTEBytesFrame = *MetaPTEBytesFrame + MPDEBytesFrame + *DPDE0BytesFrame + ExtraDPDEBytesFrame;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: DCCEnable = %u\n", __func__, DCCEnable);
+       dml_print("DML::%s: GPUVMEnable = %u\n", __func__, GPUVMEnable);
+       dml_print("DML::%s: SwModeLinear = %u\n", __func__, SurfaceTiling == dml_sw_linear);
+       dml_print("DML::%s: BytePerPixel = %u\n", __func__, BytePerPixel);
+       dml_print("DML::%s: GPUVMMaxPageTableLevels = %u\n", __func__, GPUVMMaxPageTableLevels);
+       dml_print("DML::%s: BlockHeight256Bytes = %u\n", __func__, BlockHeight256Bytes);
+       dml_print("DML::%s: BlockWidth256Bytes = %u\n", __func__, BlockWidth256Bytes);
+       dml_print("DML::%s: MacroTileHeight = %u\n", __func__, MacroTileHeight);
+       dml_print("DML::%s: MacroTileWidth = %u\n", __func__, MacroTileWidth);
+       dml_print("DML::%s: MetaPTEBytesFrame = %u\n", __func__, *MetaPTEBytesFrame);
+       dml_print("DML::%s: MPDEBytesFrame = %u\n", __func__, MPDEBytesFrame);
+       dml_print("DML::%s: DPDE0BytesFrame = %u\n", __func__, *DPDE0BytesFrame);
+       dml_print("DML::%s: ExtraDPDEBytesFrame= %u\n", __func__, ExtraDPDEBytesFrame);
+       dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %u\n", __func__, PDEAndMetaPTEBytesFrame);
+       dml_print("DML::%s: ViewportHeight = %u\n", __func__, ViewportHeight);
+       dml_print("DML::%s: SwathWidth = %u\n", __func__, SwathWidth);
+       dml_print("DML::%s: vp_height_dpte_ub = %u\n", __func__, vp_height_dpte_ub);
+#endif
+
+       if (SurfaceTiling == dml_sw_linear) {
+               *PixelPTEReqHeight = 1;
+               *PixelPTEReqWidth = GPUVMMinPageSizeKBytes * 1024 * 8 / BytePerPixel;
+               PixelPTEReqWidth_linear = GPUVMMinPageSizeKBytes * 1024 * 8 / BytePerPixel;
+               *PTERequestSize = 64;
+       } else if (GPUVMMinPageSizeKBytes == 4) {
+               *PixelPTEReqHeight = 16 * BlockHeight256Bytes;
+               *PixelPTEReqWidth = 16 * BlockWidth256Bytes;
+               *PTERequestSize = 128;
+       } else {
+               *PixelPTEReqHeight = MacroTileHeight;
+               *PixelPTEReqWidth = 8 *  1024 * GPUVMMinPageSizeKBytes / (MacroTileHeight * BytePerPixel);
+               *PTERequestSize = 64;
+       }
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: GPUVMMinPageSizeKBytes = %u\n", __func__, GPUVMMinPageSizeKBytes);
+       dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %u (after HostVM factor)\n", __func__, PDEAndMetaPTEBytesFrame);
+       dml_print("DML::%s: PixelPTEReqHeight = %u\n", __func__, *PixelPTEReqHeight);
+       dml_print("DML::%s: PixelPTEReqWidth = %u\n", __func__, *PixelPTEReqWidth);
+       dml_print("DML::%s: PixelPTEReqWidth_linear = %u\n", __func__, PixelPTEReqWidth_linear);
+       dml_print("DML::%s: PTERequestSize = %u\n", __func__, *PTERequestSize);
+       dml_print("DML::%s: Pitch = %u\n", __func__, Pitch);
+#endif
+
+       *dpte_row_height_one_row_per_frame = vp_height_dpte_ub;
+       *dpte_row_width_ub_one_row_per_frame = (dml_uint_t)((dml_ceil(((dml_float_t)Pitch * (dml_float_t) *dpte_row_height_one_row_per_frame / (dml_float_t) *PixelPTEReqHeight - 1) / (dml_float_t) *PixelPTEReqWidth, 1) + 1) * (dml_float_t) *PixelPTEReqWidth);
+       *PixelPTEBytesPerRow_one_row_per_frame = (dml_uint_t)((dml_float_t) *dpte_row_width_ub_one_row_per_frame / (dml_float_t) *PixelPTEReqWidth * *PTERequestSize);
+
+       if (SurfaceTiling == dml_sw_linear) {
+               *dpte_row_height = (dml_uint_t)(dml_min(128, 1 << (dml_uint_t) dml_floor(dml_log2(PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch), 1)));
+               dml_print("DML::%s: dpte_row_height term 1 = %u\n", __func__, PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch);
+               dml_print("DML::%s: dpte_row_height term 2 = %f\n", __func__, dml_log2(PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch));
+               dml_print("DML::%s: dpte_row_height term 3 = %f\n", __func__, dml_floor(dml_log2(PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch), 1));
+               dml_print("DML::%s: dpte_row_height term 4 = %u\n", __func__, 1 << (dml_uint_t) dml_floor(dml_log2(PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch), 1));
+               dml_print("DML::%s: dpte_row_height = %u\n", __func__, *dpte_row_height);
+
+               *dpte_row_width_ub = (dml_uint_t)(dml_ceil(((dml_float_t) Pitch * (dml_float_t) *dpte_row_height - 1), (dml_float_t) *PixelPTEReqWidth) + *PixelPTEReqWidth);
+               *PixelPTEBytesPerRow = (dml_uint_t)((dml_float_t) *dpte_row_width_ub / (dml_float_t) *PixelPTEReqWidth * *PTERequestSize);
+
+               // VBA_DELTA, VBA doesn't have programming value for pte row height linear.
+               *dpte_row_height_linear = 1 << (dml_uint_t) dml_floor(dml_log2(PTEBufferSizeInRequests * PixelPTEReqWidth_linear / Pitch), 1);
+               if (*dpte_row_height_linear > 128)
+                       *dpte_row_height_linear = 128;
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: dpte_row_width_ub = %u (linear)\n", __func__, *dpte_row_width_ub);
+#endif
+
+       } else if (!dml_is_vertical_rotation(SourceScan)) {
+               *dpte_row_height = *PixelPTEReqHeight;
+
+               if (GPUVMMinPageSizeKBytes > 64) {
+                       *dpte_row_width_ub = (dml_uint_t)((dml_ceil(((dml_float_t) Pitch * (dml_float_t) *dpte_row_height / (dml_float_t) *PixelPTEReqHeight - 1) / (dml_float_t) *PixelPTEReqWidth, 1) + 1) * *PixelPTEReqWidth);
+               } else if (ViewportStationary && (NumberOfDPPs == 1)) {
+                       *dpte_row_width_ub = (dml_uint_t)(dml_floor(ViewportXStart + SwathWidth + *PixelPTEReqWidth - 1, *PixelPTEReqWidth) - dml_floor(ViewportXStart, *PixelPTEReqWidth));
+               } else {
+                       *dpte_row_width_ub = (dml_uint_t)((dml_ceil((dml_float_t) (SwathWidth - 1) / (dml_float_t)*PixelPTEReqWidth, 1) + 1.0) * *PixelPTEReqWidth);
+               }
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: dpte_row_width_ub = %u (tiled horz)\n", __func__, *dpte_row_width_ub);
+#endif
+
+               ASSERT(*PixelPTEReqWidth);
+               if (*PixelPTEReqWidth != 0)
+                       *PixelPTEBytesPerRow = *dpte_row_width_ub / *PixelPTEReqWidth * *PTERequestSize;
+       } else {
+               *dpte_row_height = (dml_uint_t)(dml_min(*PixelPTEReqWidth, MacroTileWidth));
+
+               if (ViewportStationary && (NumberOfDPPs == 1)) {
+                       *dpte_row_width_ub = (dml_uint_t)(dml_floor(ViewportYStart + ViewportHeight + *PixelPTEReqHeight - 1, *PixelPTEReqHeight) - dml_floor(ViewportYStart, *PixelPTEReqHeight));
+               } else {
+                       *dpte_row_width_ub = (dml_uint_t)((dml_ceil((dml_float_t) (SwathWidth - 1) / (dml_float_t) *PixelPTEReqHeight, 1) + 1) * *PixelPTEReqHeight);
+               }
+
+               *PixelPTEBytesPerRow = (dml_uint_t)((dml_float_t) *dpte_row_width_ub / (dml_float_t) *PixelPTEReqHeight * *PTERequestSize);
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: dpte_row_width_ub = %u (tiled vert)\n", __func__, *dpte_row_width_ub);
+#endif
+       }
+
+       if (GPUVMEnable != true)
+               *PixelPTEBytesPerRow = 0;
+
+       *PixelPTEBytesPerRowStorage = *PixelPTEBytesPerRow;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: GPUVMMinPageSizeKBytes = %u\n", __func__, GPUVMMinPageSizeKBytes);
+       dml_print("DML::%s: GPUVMEnable = %u\n", __func__, GPUVMEnable);
+       dml_print("DML::%s: dpte_row_height = %u\n", __func__, *dpte_row_height);
+       dml_print("DML::%s: dpte_row_height_linear = %u\n", __func__, *dpte_row_height_linear);
+       dml_print("DML::%s: dpte_row_width_ub = %u\n", __func__, *dpte_row_width_ub);
+       dml_print("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, *PixelPTEBytesPerRow);
+       dml_print("DML::%s: PixelPTEBytesPerRowStorage = %u\n", __func__, *PixelPTEBytesPerRowStorage);
+       dml_print("DML::%s: PTEBufferSizeInRequests = %u\n", __func__, PTEBufferSizeInRequests);
+       dml_print("DML::%s: dpte_row_height_one_row_per_frame = %u\n", __func__, *dpte_row_height_one_row_per_frame);
+       dml_print("DML::%s: dpte_row_width_ub_one_row_per_frame = %u\n", __func__, *dpte_row_width_ub_one_row_per_frame);
+       dml_print("DML::%s: PixelPTEBytesPerRow_one_row_per_frame = %u\n", __func__, *PixelPTEBytesPerRow_one_row_per_frame);
+#endif
+
+       dml_print("DML: vm_bytes = meta_pte_bytes_per_frame (per_pipe) = MetaPTEBytesFrame = : %i\n", *MetaPTEBytesFrame);
+
+       return PDEAndMetaPTEBytesFrame;
+} // CalculateVMAndRowBytes
+
+static void PixelClockAdjustmentForProgressiveToInterlaceUnit(struct dml_display_cfg_st *display_cfg, dml_bool_t ptoi_supported)
+{
+       dml_uint_t num_active_planes = dml_get_num_active_planes(display_cfg);
+
+       //Progressive To Interlace Unit Effect
+       for (dml_uint_t k = 0; k < num_active_planes; ++k) {
+               display_cfg->output.PixelClockBackEnd[k] = display_cfg->timing.PixelClock[k];
+               if (display_cfg->timing.Interlace[k] == 1 && ptoi_supported == true) {
+                       display_cfg->timing.PixelClock[k] = 2 * display_cfg->timing.PixelClock[k];
+               }
+       }
+}
+
+static dml_float_t TruncToValidBPP(
+               dml_float_t LinkBitRate,
+               dml_uint_t Lanes,
+               dml_uint_t HTotal,
+               dml_uint_t HActive,
+               dml_float_t PixelClock,
+               dml_float_t DesiredBPP,
+               dml_bool_t DSCEnable,
+               enum dml_output_encoder_class Output,
+               enum dml_output_format_class Format,
+               dml_uint_t DSCInputBitPerComponent,
+               dml_uint_t DSCSlices,
+               dml_uint_t AudioRate,
+               dml_uint_t AudioLayout,
+               enum dml_odm_mode ODMModeNoDSC,
+               enum dml_odm_mode ODMModeDSC,
+
+               // Output
+               dml_uint_t *RequiredSlots)
+{
+       dml_float_t MaxLinkBPP;
+       dml_uint_t MinDSCBPP;
+       dml_float_t MaxDSCBPP;
+       dml_uint_t NonDSCBPP0;
+       dml_uint_t NonDSCBPP1;
+       dml_uint_t NonDSCBPP2;
+
+       if (Format == dml_420) {
+               NonDSCBPP0 = 12;
+               NonDSCBPP1 = 15;
+               NonDSCBPP2 = 18;
+               MinDSCBPP = 6;
+               MaxDSCBPP = 1.5 * DSCInputBitPerComponent - 1 / 16;
+       } else if (Format == dml_444) {
+               NonDSCBPP0 = 24;
+               NonDSCBPP1 = 30;
+               NonDSCBPP2 = 36;
+               MinDSCBPP = 8;
+               MaxDSCBPP = 3 * DSCInputBitPerComponent - 1.0 / 16;
+       } else {
+               if (Output == dml_hdmi) {
+                       NonDSCBPP0 = 24;
+                       NonDSCBPP1 = 24;
+                       NonDSCBPP2 = 24;
+               } else {
+                       NonDSCBPP0 = 16;
+                       NonDSCBPP1 = 20;
+                       NonDSCBPP2 = 24;
+       }
+       if (Format == dml_n422) {
+               MinDSCBPP = 7;
+                       MaxDSCBPP = 2 * DSCInputBitPerComponent - 1.0 / 16.0;
+               } else {
+                       MinDSCBPP = 8;
+                       MaxDSCBPP = 3 * DSCInputBitPerComponent - 1.0 / 16.0;
+               }
+       }
+
+       if (Output == dml_dp2p0) {
+               MaxLinkBPP = LinkBitRate * Lanes / PixelClock * 128.0 / 132.0 * 383.0 / 384.0 * 65536.0 / 65540.0;
+       } else if (DSCEnable && Output == dml_dp) {
+               MaxLinkBPP = LinkBitRate / 10.0 * 8.0 * Lanes / PixelClock * (1 - 2.4 / 100);
+       } else {
+               MaxLinkBPP = LinkBitRate / 10.0 * 8.0 * Lanes / PixelClock;
+       }
+
+       if (DSCEnable) {
+               if (ODMModeDSC == dml_odm_mode_combine_4to1) {
+                       MaxLinkBPP = dml_min(MaxLinkBPP, 16);
+               } else if (ODMModeDSC == dml_odm_mode_combine_2to1) {
+                       MaxLinkBPP = dml_min(MaxLinkBPP, 32);
+               } else if (ODMModeDSC == dml_odm_mode_split_1to2) {
+                       MaxLinkBPP = 2 * MaxLinkBPP;
+               }
+       } else {
+               if (ODMModeNoDSC == dml_odm_mode_combine_4to1) {
+                       MaxLinkBPP = dml_min(MaxLinkBPP, 16);
+               } else if (ODMModeNoDSC == dml_odm_mode_combine_2to1) {
+                       MaxLinkBPP = dml_min(MaxLinkBPP, 32);
+               } else if (ODMModeNoDSC == dml_odm_mode_split_1to2) {
+                       MaxLinkBPP = 2 * MaxLinkBPP;
+               }
+       }
+
+       if (DesiredBPP == 0) {
+               if (DSCEnable) {
+                       if (MaxLinkBPP < MinDSCBPP) {
+                               return __DML_DPP_INVALID__;
+                       } else if (MaxLinkBPP >= MaxDSCBPP) {
+                               return MaxDSCBPP;
+                       } else {
+                               return dml_floor(16.0 * MaxLinkBPP, 1.0) / 16.0;
+                       }
+               } else {
+                       if (MaxLinkBPP >= NonDSCBPP2) {
+                               return NonDSCBPP2;
+                       } else if (MaxLinkBPP >= NonDSCBPP1) {
+                               return NonDSCBPP1;
+                       } else if (MaxLinkBPP >= NonDSCBPP0) {
+                               return NonDSCBPP0;
+                       } else {
+                               return __DML_DPP_INVALID__;
+                       }
+               }
+       } else {
+               if (!((DSCEnable == false && (DesiredBPP == NonDSCBPP2 || DesiredBPP == NonDSCBPP1 || DesiredBPP == NonDSCBPP0)) ||
+                               (DSCEnable && DesiredBPP >= MinDSCBPP && DesiredBPP <= MaxDSCBPP))) {
+                       return __DML_DPP_INVALID__;
+               } else {
+                       return DesiredBPP;
+               }
+       }
+
+       *RequiredSlots = (dml_uint_t)(dml_ceil(DesiredBPP / MaxLinkBPP * 64, 1));
+
+       return __DML_DPP_INVALID__;
+} // TruncToValidBPP
+
+static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
+               struct display_mode_lib_scratch_st *scratch,
+               struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params_st *p)
+{
+       struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals_st *s = &scratch->CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals;
+
+       s->TotalActiveWriteback = 0;
+       p->Watermark->UrgentWatermark = p->mmSOCParameters.UrgentLatency + p->mmSOCParameters.ExtraLatency;
+       p->Watermark->USRRetrainingWatermark = p->mmSOCParameters.UrgentLatency + p->mmSOCParameters.ExtraLatency + p->mmSOCParameters.USRRetrainingLatency + p->mmSOCParameters.SMNLatency;
+       p->Watermark->DRAMClockChangeWatermark = p->mmSOCParameters.DRAMClockChangeLatency + p->Watermark->UrgentWatermark;
+       p->Watermark->FCLKChangeWatermark = p->mmSOCParameters.FCLKChangeLatency + p->Watermark->UrgentWatermark;
+       p->Watermark->StutterExitWatermark = p->mmSOCParameters.SRExitTime + p->mmSOCParameters.ExtraLatency + 10 / p->DCFClkDeepSleep;
+       p->Watermark->StutterEnterPlusExitWatermark = p->mmSOCParameters.SREnterPlusExitTime + p->mmSOCParameters.ExtraLatency + 10 / p->DCFClkDeepSleep;
+       p->Watermark->Z8StutterExitWatermark = p->mmSOCParameters.SRExitZ8Time + p->mmSOCParameters.ExtraLatency + 10 / p->DCFClkDeepSleep;
+       p->Watermark->Z8StutterEnterPlusExitWatermark = p->mmSOCParameters.SREnterPlusExitZ8Time + p->mmSOCParameters.ExtraLatency + 10 / p->DCFClkDeepSleep;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: UrgentLatency = %f\n", __func__, p->mmSOCParameters.UrgentLatency);
+       dml_print("DML::%s: ExtraLatency = %f\n", __func__, p->mmSOCParameters.ExtraLatency);
+       dml_print("DML::%s: DRAMClockChangeLatency = %f\n", __func__, p->mmSOCParameters.DRAMClockChangeLatency);
+       dml_print("DML::%s: UrgentWatermark = %f\n", __func__, p->Watermark->UrgentWatermark);
+       dml_print("DML::%s: USRRetrainingWatermark = %f\n", __func__, p->Watermark->USRRetrainingWatermark);
+       dml_print("DML::%s: DRAMClockChangeWatermark = %f\n", __func__, p->Watermark->DRAMClockChangeWatermark);
+       dml_print("DML::%s: FCLKChangeWatermark = %f\n", __func__, p->Watermark->FCLKChangeWatermark);
+       dml_print("DML::%s: StutterExitWatermark = %f\n", __func__, p->Watermark->StutterExitWatermark);
+       dml_print("DML::%s: StutterEnterPlusExitWatermark = %f\n", __func__, p->Watermark->StutterEnterPlusExitWatermark);
+       dml_print("DML::%s: Z8StutterExitWatermark = %f\n", __func__, p->Watermark->Z8StutterExitWatermark);
+       dml_print("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n", __func__, p->Watermark->Z8StutterEnterPlusExitWatermark);
+#endif
+
+       s->TotalActiveWriteback = 0;
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               if (p->WritebackEnable[k] == true) {
+                       s->TotalActiveWriteback = s->TotalActiveWriteback + 1;
+               }
+       }
+
+       if (s->TotalActiveWriteback <= 1) {
+               p->Watermark->WritebackUrgentWatermark = p->mmSOCParameters.WritebackLatency;
+       } else {
+               p->Watermark->WritebackUrgentWatermark = p->mmSOCParameters.WritebackLatency + p->WritebackChunkSize * 1024.0 / 32.0 / p->SOCCLK;
+       }
+       if (p->USRRetrainingRequiredFinal)
+               p->Watermark->WritebackUrgentWatermark = p->Watermark->WritebackUrgentWatermark + p->mmSOCParameters.USRRetrainingLatency;
+
+       if (s->TotalActiveWriteback <= 1) {
+               p->Watermark->WritebackDRAMClockChangeWatermark = p->mmSOCParameters.DRAMClockChangeLatency + p->mmSOCParameters.WritebackLatency;
+               p->Watermark->WritebackFCLKChangeWatermark = p->mmSOCParameters.FCLKChangeLatency + p->mmSOCParameters.WritebackLatency;
+       } else {
+               p->Watermark->WritebackDRAMClockChangeWatermark = p->mmSOCParameters.DRAMClockChangeLatency + p->mmSOCParameters.WritebackLatency + p->WritebackChunkSize * 1024.0 / 32.0 / p->SOCCLK;
+               p->Watermark->WritebackFCLKChangeWatermark = p->mmSOCParameters.FCLKChangeLatency + p->mmSOCParameters.WritebackLatency + p->WritebackChunkSize * 1024 / 32 / p->SOCCLK;
+       }
+
+       if (p->USRRetrainingRequiredFinal)
+               p->Watermark->WritebackDRAMClockChangeWatermark = p->Watermark->WritebackDRAMClockChangeWatermark + p->mmSOCParameters.USRRetrainingLatency;
+
+       if (p->USRRetrainingRequiredFinal)
+               p->Watermark->WritebackFCLKChangeWatermark = p->Watermark->WritebackFCLKChangeWatermark + p->mmSOCParameters.USRRetrainingLatency;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: WritebackDRAMClockChangeWatermark = %f\n", __func__, p->Watermark->WritebackDRAMClockChangeWatermark);
+       dml_print("DML::%s: WritebackFCLKChangeWatermark = %f\n", __func__, p->Watermark->WritebackFCLKChangeWatermark);
+       dml_print("DML::%s: WritebackUrgentWatermark = %f\n", __func__, p->Watermark->WritebackUrgentWatermark);
+       dml_print("DML::%s: USRRetrainingRequiredFinal = %u\n", __func__, p->USRRetrainingRequiredFinal);
+       dml_print("DML::%s: USRRetrainingLatency = %f\n", __func__, p->mmSOCParameters.USRRetrainingLatency);
+#endif
+
+       s->TotalPixelBW = 0.0;
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               s->TotalPixelBW = s->TotalPixelBW + p->DPPPerSurface[k]
+                                       * (p->SwathWidthY[k] * p->BytePerPixelDETY[k] * p->VRatio[k] + p->SwathWidthC[k] * p->BytePerPixelDETC[k] * p->VRatioChroma[k]) / (p->HTotal[k] / p->PixelClock[k]);
+       }
+
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+
+               s->LBLatencyHidingSourceLinesY[k] = (dml_uint_t)(dml_min((dml_float_t)p->MaxLineBufferLines, dml_floor((dml_float_t)p->LineBufferSize / (dml_float_t)p->LBBitPerPixel[k] / ((dml_float_t)p->SwathWidthY[k] / dml_max(p->HRatio[k], 1.0)), 1)) - (p->VTaps[k] - 1));
+               s->LBLatencyHidingSourceLinesC[k] = (dml_uint_t)(dml_min((dml_float_t)p->MaxLineBufferLines, dml_floor((dml_float_t)p->LineBufferSize / (dml_float_t)p->LBBitPerPixel[k] / ((dml_float_t)p->SwathWidthC[k] / dml_max(p->HRatioChroma[k], 1.0)), 1)) - (p->VTapsChroma[k] - 1));
+
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, MaxLineBufferLines = %u\n", __func__, k, p->MaxLineBufferLines);
+               dml_print("DML::%s: k=%u, LineBufferSize = %u\n", __func__, k, p->LineBufferSize);
+               dml_print("DML::%s: k=%u, LBBitPerPixel = %u\n", __func__, k, p->LBBitPerPixel[k]);
+               dml_print("DML::%s: k=%u, HRatio = %f\n", __func__, k, p->HRatio[k]);
+               dml_print("DML::%s: k=%u, VTaps              = %u\n", __func__, k, p->VTaps[k]);
+#endif
+
+               s->EffectiveLBLatencyHidingY = s->LBLatencyHidingSourceLinesY[k] / p->VRatio[k] * (p->HTotal[k] / p->PixelClock[k]);
+               s->EffectiveLBLatencyHidingC = s->LBLatencyHidingSourceLinesC[k] / p->VRatioChroma[k] * (p->HTotal[k] / p->PixelClock[k]);
+
+               s->EffectiveDETBufferSizeY = p->DETBufferSizeY[k];
+               if (p->UnboundedRequestEnabled) {
+                       s->EffectiveDETBufferSizeY = s->EffectiveDETBufferSizeY + p->CompressedBufferSizeInkByte * 1024 * (p->SwathWidthY[k] * p->BytePerPixelDETY[k] * p->VRatio[k]) / (p->HTotal[k] / p->PixelClock[k]) / s->TotalPixelBW;
+               }
+
+               s->LinesInDETY[k] = (dml_float_t)s->EffectiveDETBufferSizeY / p->BytePerPixelDETY[k] / p->SwathWidthY[k];
+               s->LinesInDETYRoundedDownToSwath[k] = (dml_uint_t)(dml_floor(s->LinesInDETY[k], p->SwathHeightY[k]));
+               s->FullDETBufferingTimeY = s->LinesInDETYRoundedDownToSwath[k] * (p->HTotal[k] / p->PixelClock[k]) / p->VRatio[k];
+
+               s->ActiveClockChangeLatencyHidingY = s->EffectiveLBLatencyHidingY + s->FullDETBufferingTimeY - ((dml_float_t)p->DSTXAfterScaler[k] / (dml_float_t)p->HTotal[k] + (dml_float_t)p->DSTYAfterScaler[k]) * (dml_float_t)p->HTotal[k] / p->PixelClock[k];
+
+               if (p->NumberOfActiveSurfaces > 1) {
+                       s->ActiveClockChangeLatencyHidingY = s->ActiveClockChangeLatencyHidingY - (1.0 - 1.0 / (dml_float_t)p->NumberOfActiveSurfaces) * (dml_float_t)p->SwathHeightY[k] * (dml_float_t)p->HTotal[k] / p->PixelClock[k] / p->VRatio[k];
+       }
+
+               if (p->BytePerPixelDETC[k] > 0) {
+                       s->LinesInDETC[k] = p->DETBufferSizeC[k] / p->BytePerPixelDETC[k] / p->SwathWidthC[k];
+                       s->LinesInDETCRoundedDownToSwath[k] = (dml_uint_t)(dml_floor(s->LinesInDETC[k], p->SwathHeightC[k]));
+                       s->FullDETBufferingTimeC = s->LinesInDETCRoundedDownToSwath[k] * (p->HTotal[k] / p->PixelClock[k]) / p->VRatioChroma[k];
+                       s->ActiveClockChangeLatencyHidingC = s->EffectiveLBLatencyHidingC + s->FullDETBufferingTimeC - ((dml_float_t)p->DSTXAfterScaler[k] / (dml_float_t)p->HTotal[k] + (dml_float_t)p->DSTYAfterScaler[k]) * (dml_float_t)p->HTotal[k] / p->PixelClock[k];
+                       if (p->NumberOfActiveSurfaces > 1) {
+                       s->ActiveClockChangeLatencyHidingC = s->ActiveClockChangeLatencyHidingC - (1.0 - 1.0 / (dml_float_t)p->NumberOfActiveSurfaces) * (dml_float_t)p->SwathHeightC[k] * (dml_float_t)p->HTotal[k] / p->PixelClock[k] / p->VRatioChroma[k];
+                       }
+                       s->ActiveClockChangeLatencyHiding = dml_min(s->ActiveClockChangeLatencyHidingY, s->ActiveClockChangeLatencyHidingC);
+               } else {
+                       s->ActiveClockChangeLatencyHiding = s->ActiveClockChangeLatencyHidingY;
+               }
+
+               s->ActiveDRAMClockChangeLatencyMargin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->UrgentWatermark - p->Watermark->DRAMClockChangeWatermark;
+               s->ActiveFCLKChangeLatencyMargin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->UrgentWatermark - p->Watermark->FCLKChangeWatermark;
+               s->USRRetrainingLatencyMargin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->USRRetrainingWatermark;
+
+               if (p->WritebackEnable[k]) {
+                       s->WritebackLatencyHiding = (dml_float_t)p->WritebackInterfaceBufferSize * 1024.0 / ((dml_float_t)p->WritebackDestinationWidth[k] * (dml_float_t)p->WritebackDestinationHeight[k] / ((dml_float_t)p->WritebackSourceHeight[k] * (dml_float_t)p->HTotal[k] / p->PixelClock[k]) * 4.0);
+                       if (p->WritebackPixelFormat[k] == dml_444_64) {
+                               s->WritebackLatencyHiding = s->WritebackLatencyHiding / 2;
+                       }
+                       s->WritebackDRAMClockChangeLatencyMargin = s->WritebackLatencyHiding - p->Watermark->WritebackDRAMClockChangeWatermark;
+
+                       s->WritebackFCLKChangeLatencyMargin = s->WritebackLatencyHiding - p->Watermark->WritebackFCLKChangeWatermark;
+
+                       s->ActiveDRAMClockChangeLatencyMargin[k] = dml_min(s->ActiveDRAMClockChangeLatencyMargin[k], s->WritebackFCLKChangeLatencyMargin);
+                       s->ActiveFCLKChangeLatencyMargin[k] = dml_min(s->ActiveFCLKChangeLatencyMargin[k], s->WritebackDRAMClockChangeLatencyMargin);
+               }
+               p->MaxActiveDRAMClockChangeLatencySupported[k] =  (p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe) ? 0 : (s->ActiveDRAMClockChangeLatencyMargin[k] + p->mmSOCParameters.DRAMClockChangeLatency);
+               p->ActiveDRAMClockChangeLatencyMargin[k] = s->ActiveDRAMClockChangeLatencyMargin[k];
+       }
+
+       *p->USRRetrainingSupport = true;
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               if ((p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) && (s->USRRetrainingLatencyMargin[k] < 0)) {
+                       *p->USRRetrainingSupport = false;
+               }
+       }
+
+       s->FoundCriticalSurface = false;
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               if ((p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) && ((!s->FoundCriticalSurface)
+                       || ((s->ActiveFCLKChangeLatencyMargin[k] + p->mmSOCParameters.FCLKChangeLatency) < *p->MaxActiveFCLKChangeLatencySupported))) {
+                       s->FoundCriticalSurface = true;
+                       *p->MaxActiveFCLKChangeLatencySupported = s->ActiveFCLKChangeLatencyMargin[k] + p->mmSOCParameters.FCLKChangeLatency;
+               }
+       }
+
+       for (dml_uint_t i = 0; i < p->NumberOfActiveSurfaces; ++i) {
+               for (dml_uint_t j = 0; j < p->NumberOfActiveSurfaces; ++j) {
+                       if (i == j ||
+                               (p->BlendingAndTiming[i] == i && p->BlendingAndTiming[j] == i) ||
+                               (p->BlendingAndTiming[j] == j && p->BlendingAndTiming[i] == j) ||
+                               (p->BlendingAndTiming[i] == p->BlendingAndTiming[j] && p->BlendingAndTiming[i] != i) ||
+                               (p->SynchronizeTimingsFinal && p->PixelClock[i] == p->PixelClock[j] && p->HTotal[i] == p->HTotal[j] && p->VTotal[i] == p->VTotal[j] && p->VActive[i] == p->VActive[j]) ||
+                               (p->SynchronizeDRRDisplaysForUCLKPStateChangeFinal && (p->DRRDisplay[i] || p->DRRDisplay[j]))) {
+                               s->SynchronizedSurfaces[i][j] = true;
+                       } else {
+                               s->SynchronizedSurfaces[i][j] = false;
+                       }
+               }
+       }
+
+       s->FCLKChangeSupportNumber = 0;
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               if ((p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) &&  (s->ActiveFCLKChangeLatencyMargin[k] < 0)) {
+                       if (!(p->PrefetchMode[k] <= 1)) {
+                               s->FCLKChangeSupportNumber = 3;
+                       } else if (s->FCLKChangeSupportNumber == 0) {
+                               s->FCLKChangeSupportNumber = ((p->SynchronizeDRRDisplaysForUCLKPStateChangeFinal && p->DRRDisplay[k]) ? 2 : 1);
+                               s->LastSurfaceWithoutMargin = k;
+                       } else if (((s->FCLKChangeSupportNumber == 1) && (p->DRRDisplay[k] || (!s->SynchronizedSurfaces[s->LastSurfaceWithoutMargin][k]))) || (s->FCLKChangeSupportNumber == 2))
+                               s->FCLKChangeSupportNumber = 3;
+               }
+       }
+
+       if (s->FCLKChangeSupportNumber == 0) {
+               *p->FCLKChangeSupport = dml_fclock_change_vactive;
+       } else if ((s->FCLKChangeSupportNumber == 1) || (s->FCLKChangeSupportNumber == 2)) {
+               *p->FCLKChangeSupport = dml_fclock_change_vblank;
+       } else {
+               *p->FCLKChangeSupport = dml_fclock_change_unsupported;
+       }
+
+       s->DRAMClockChangeMethod = 0;
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               if (p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_full_frame)
+                       s->DRAMClockChangeMethod = 1;
+               else if (p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_sub_viewport)
+                       s->DRAMClockChangeMethod = 2;
+       }
+
+       s->DRAMClockChangeSupportNumber = 0;
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               if (((s->DRAMClockChangeMethod == 0) && (s->ActiveDRAMClockChangeLatencyMargin[k] < 0)) ||
+                       ((s->DRAMClockChangeMethod == 1) && (p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_full_frame)) ||
+                       ((s->DRAMClockChangeMethod == 2) && (p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_sub_viewport) && (p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe))) {
+                       if (p->PrefetchMode[k] != 0) { // Don't need to support DRAM clock change, PrefetchMode 0 means needs DRAM clock change support
+                               s->DRAMClockChangeSupportNumber = 3;
+                       } else if (s->DRAMClockChangeSupportNumber == 0) {
+                               s->DRAMClockChangeSupportNumber =  (p->SynchronizeDRRDisplaysForUCLKPStateChangeFinal && p->DRRDisplay[k]) ? 2 : 1;
+                               s->LastSurfaceWithoutMargin = k;
+                       } else if (((s->DRAMClockChangeSupportNumber == 1) && (p->DRRDisplay[k] || !s->SynchronizedSurfaces[s->LastSurfaceWithoutMargin][k])) || (s->DRAMClockChangeSupportNumber == 2)) {
+                               s->DRAMClockChangeSupportNumber = 3;
+                       }
+               }
+       }
+
+       if (s->DRAMClockChangeMethod == 0) { // No MALL usage
+               if (s->DRAMClockChangeSupportNumber == 0) {
+                       *p->DRAMClockChangeSupport = dml_dram_clock_change_vactive;
+               } else if (s->DRAMClockChangeSupportNumber == 1) {
+                       *p->DRAMClockChangeSupport = dml_dram_clock_change_vblank;
+               } else if (s->DRAMClockChangeSupportNumber == 2) {
+                       *p->DRAMClockChangeSupport = dml_dram_clock_change_vblank_drr;
+               } else {
+                       *p->DRAMClockChangeSupport = dml_dram_clock_change_unsupported;
+               }
+       } else if (s->DRAMClockChangeMethod == 1) { // Any pipe using MALL full frame
+               if (s->DRAMClockChangeSupportNumber == 0) {
+                       *p->DRAMClockChangeSupport = dml_dram_clock_change_vactive_w_mall_full_frame;
+               } else if (s->DRAMClockChangeSupportNumber == 1) {
+                       *p->DRAMClockChangeSupport = dml_dram_clock_change_vblank_w_mall_full_frame;
+               } else if (s->DRAMClockChangeSupportNumber == 2) {
+                       *p->DRAMClockChangeSupport = dml_dram_clock_change_vblank_drr_w_mall_full_frame;
+               } else {
+                       *p->DRAMClockChangeSupport = dml_dram_clock_change_unsupported;
+               }
+       } else { // Any pipe using MALL subviewport
+               if (s->DRAMClockChangeSupportNumber == 0) {
+                       *p->DRAMClockChangeSupport = dml_dram_clock_change_vactive_w_mall_sub_vp;
+               } else if (s->DRAMClockChangeSupportNumber == 1) {
+                       *p->DRAMClockChangeSupport = dml_dram_clock_change_vblank_w_mall_sub_vp;
+               } else if (s->DRAMClockChangeSupportNumber == 2) {
+                       *p->DRAMClockChangeSupport = dml_dram_clock_change_vblank_drr_w_mall_sub_vp;
+               } else {
+                       *p->DRAMClockChangeSupport = dml_dram_clock_change_unsupported;
+               }
+       }
+
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               s->dst_y_pstate = (dml_uint_t)(dml_ceil((p->mmSOCParameters.DRAMClockChangeLatency + p->mmSOCParameters.UrgentLatency) / (p->HTotal[k] / p->PixelClock[k]), 1));
+               s->src_y_pstate_l = (dml_uint_t)(dml_ceil(s->dst_y_pstate * p->VRatio[k], p->SwathHeightY[k]));
+               s->src_y_ahead_l = (dml_uint_t)(dml_floor(p->DETBufferSizeY[k] / p->BytePerPixelDETY[k] / p->SwathWidthY[k], p->SwathHeightY[k]) + s->LBLatencyHidingSourceLinesY[k]);
+               s->sub_vp_lines_l = s->src_y_pstate_l + s->src_y_ahead_l + p->meta_row_height[k];
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, DETBufferSizeY = %u\n", __func__, k, p->DETBufferSizeY[k]);
+               dml_print("DML::%s: k=%u, BytePerPixelDETY = %f\n", __func__, k, p->BytePerPixelDETY[k]);
+               dml_print("DML::%s: k=%u, SwathWidthY = %u\n", __func__, k, p->SwathWidthY[k]);
+               dml_print("DML::%s: k=%u, SwathHeightY = %u\n", __func__, k, p->SwathHeightY[k]);
+               dml_print("DML::%s: k=%u, LBLatencyHidingSourceLinesY = %u\n", __func__, k, s->LBLatencyHidingSourceLinesY[k]);
+               dml_print("DML::%s: k=%u, dst_y_pstate = %u\n", __func__, k, s->dst_y_pstate);
+               dml_print("DML::%s: k=%u, src_y_pstate_l = %u\n", __func__, k, s->src_y_pstate_l);
+               dml_print("DML::%s: k=%u, src_y_ahead_l = %u\n", __func__, k, s->src_y_ahead_l);
+               dml_print("DML::%s: k=%u, meta_row_height = %u\n", __func__, k, p->meta_row_height[k]);
+               dml_print("DML::%s: k=%u, sub_vp_lines_l  = %u\n", __func__, k, s->sub_vp_lines_l);
+#endif
+               p->SubViewportLinesNeededInMALL[k] = s->sub_vp_lines_l;
+
+               if (p->BytePerPixelDETC[k] > 0) {
+               s->src_y_pstate_c = (dml_uint_t)(dml_ceil(s->dst_y_pstate * p->VRatioChroma[k], p->SwathHeightC[k]));
+               s->src_y_ahead_c = (dml_uint_t)(dml_floor(p->DETBufferSizeC[k] / p->BytePerPixelDETC[k] / p->SwathWidthC[k], p->SwathHeightC[k]) + s->LBLatencyHidingSourceLinesC[k]);
+               s->sub_vp_lines_c = s->src_y_pstate_c + s->src_y_ahead_c + p->meta_row_height_chroma[k];
+               p->SubViewportLinesNeededInMALL[k] = (dml_uint_t)(dml_max(s->sub_vp_lines_l, s->sub_vp_lines_c));
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, src_y_pstate_c = %u\n", __func__, k, s->src_y_pstate_c);
+               dml_print("DML::%s: k=%u, src_y_ahead_c = %u\n", __func__, k, s->src_y_ahead_c);
+               dml_print("DML::%s: k=%u, meta_row_height_chroma = %u\n", __func__, k, p->meta_row_height_chroma[k]);
+               dml_print("DML::%s: k=%u, sub_vp_lines_c            = %u\n", __func__, k, s->sub_vp_lines_c);
+#endif
+               }
+       }
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: DRAMClockChangeSupport = %u\n", __func__, *p->DRAMClockChangeSupport);
+       dml_print("DML::%s: FCLKChangeSupport = %u\n", __func__, *p->FCLKChangeSupport);
+       dml_print("DML::%s: MaxActiveFCLKChangeLatencySupported = %f\n", __func__, *p->MaxActiveFCLKChangeLatencySupported);
+       dml_print("DML::%s: USRRetrainingSupport                        = %u\n", __func__, *p->USRRetrainingSupport);
+#endif
+} // CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport
+
+static void CalculateDCFCLKDeepSleep(
+               dml_uint_t NumberOfActiveSurfaces,
+               dml_uint_t BytePerPixelY[],
+               dml_uint_t BytePerPixelC[],
+               dml_float_t VRatio[],
+               dml_float_t VRatioChroma[],
+               dml_uint_t SwathWidthY[],
+               dml_uint_t SwathWidthC[],
+               dml_uint_t DPPPerSurface[],
+               dml_float_t HRatio[],
+               dml_float_t HRatioChroma[],
+               dml_float_t PixelClock[],
+               dml_float_t PSCL_THROUGHPUT[],
+               dml_float_t PSCL_THROUGHPUT_CHROMA[],
+               dml_float_t Dppclk[],
+               dml_float_t ReadBandwidthLuma[],
+               dml_float_t ReadBandwidthChroma[],
+               dml_uint_t ReturnBusWidth,
+
+               // Output
+               dml_float_t *DCFClkDeepSleep)
+{
+       dml_float_t DisplayPipeLineDeliveryTimeLuma;
+       dml_float_t DisplayPipeLineDeliveryTimeChroma;
+       dml_float_t DCFClkDeepSleepPerSurface[__DML_NUM_PLANES__];
+       dml_float_t ReadBandwidth = 0.0;
+
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+               if (VRatio[k] <= 1) {
+                       DisplayPipeLineDeliveryTimeLuma = SwathWidthY[k] * DPPPerSurface[k] / HRatio[k] / PixelClock[k];
+               } else {
+                       DisplayPipeLineDeliveryTimeLuma = SwathWidthY[k] / PSCL_THROUGHPUT[k] / Dppclk[k];
+               }
+               if (BytePerPixelC[k] == 0) {
+                       DisplayPipeLineDeliveryTimeChroma = 0;
+               } else {
+                       if (VRatioChroma[k] <= 1) {
+                               DisplayPipeLineDeliveryTimeChroma = SwathWidthC[k] * DPPPerSurface[k] / HRatioChroma[k] / PixelClock[k];
+                       } else {
+                               DisplayPipeLineDeliveryTimeChroma = SwathWidthC[k] / PSCL_THROUGHPUT_CHROMA[k] / Dppclk[k];
+                       }
+               }
+
+               if (BytePerPixelC[k] > 0) {
+                       DCFClkDeepSleepPerSurface[k] = dml_max(__DML_MIN_DCFCLK_FACTOR__ * SwathWidthY[k] * BytePerPixelY[k] / 32.0 / DisplayPipeLineDeliveryTimeLuma,
+                                                                                                               __DML_MIN_DCFCLK_FACTOR__ * SwathWidthC[k] * BytePerPixelC[k] / 32.0 / DisplayPipeLineDeliveryTimeChroma);
+               } else {
+                       DCFClkDeepSleepPerSurface[k] = __DML_MIN_DCFCLK_FACTOR__ * SwathWidthY[k] * BytePerPixelY[k] / 64.0 / DisplayPipeLineDeliveryTimeLuma;
+               }
+               DCFClkDeepSleepPerSurface[k] = dml_max(DCFClkDeepSleepPerSurface[k], PixelClock[k] / 16);
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, PixelClock = %f\n", __func__, k, PixelClock[k]);
+               dml_print("DML::%s: k=%u, DCFClkDeepSleepPerSurface = %f\n", __func__, k, DCFClkDeepSleepPerSurface[k]);
+#endif
+       }
+
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               ReadBandwidth = ReadBandwidth + ReadBandwidthLuma[k] + ReadBandwidthChroma[k];
+       }
+
+       *DCFClkDeepSleep = dml_max(8.0, __DML_MIN_DCFCLK_FACTOR__ * ReadBandwidth / (dml_float_t) ReturnBusWidth);
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: __DML_MIN_DCFCLK_FACTOR__ = %f\n", __func__, __DML_MIN_DCFCLK_FACTOR__);
+               dml_print("DML::%s: ReadBandwidth = %f\n", __func__, ReadBandwidth);
+               dml_print("DML::%s: ReturnBusWidth = %u\n", __func__, ReturnBusWidth);
+               dml_print("DML::%s: DCFClkDeepSleep = %f\n", __func__, *DCFClkDeepSleep);
+#endif
+
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               *DCFClkDeepSleep = dml_max(*DCFClkDeepSleep, DCFClkDeepSleepPerSurface[k]);
+       }
+       dml_print("DML::%s: DCFClkDeepSleep = %f (final)\n", __func__, *DCFClkDeepSleep);
+} // CalculateDCFCLKDeepSleep
+
+static void CalculateUrgentBurstFactor(
+               enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange,
+               dml_uint_t swath_width_luma_ub,
+               dml_uint_t swath_width_chroma_ub,
+               dml_uint_t SwathHeightY,
+               dml_uint_t SwathHeightC,
+               dml_float_t LineTime,
+               dml_float_t UrgentLatency,
+               dml_float_t CursorBufferSize,
+               dml_uint_t CursorWidth,
+               dml_uint_t CursorBPP,
+               dml_float_t VRatio,
+               dml_float_t VRatioC,
+               dml_float_t BytePerPixelInDETY,
+               dml_float_t BytePerPixelInDETC,
+               dml_uint_t DETBufferSizeY,
+               dml_uint_t DETBufferSizeC,
+               // Output
+               dml_float_t *UrgentBurstFactorCursor,
+               dml_float_t *UrgentBurstFactorLuma,
+               dml_float_t *UrgentBurstFactorChroma,
+               dml_bool_t *NotEnoughUrgentLatencyHiding)
+{
+       dml_float_t LinesInDETLuma;
+       dml_float_t LinesInDETChroma;
+       dml_uint_t LinesInCursorBuffer;
+       dml_float_t CursorBufferSizeInTime;
+       dml_float_t DETBufferSizeInTimeLuma;
+       dml_float_t DETBufferSizeInTimeChroma;
+
+       *NotEnoughUrgentLatencyHiding = 0;
+
+       if (CursorWidth > 0) {
+               LinesInCursorBuffer = 1 << (dml_uint_t) dml_floor(dml_log2(CursorBufferSize * 1024.0 / (CursorWidth * CursorBPP / 8.0)), 1.0);
+               if (VRatio > 0) {
+                       CursorBufferSizeInTime = LinesInCursorBuffer * LineTime / VRatio;
+                       if (CursorBufferSizeInTime - UrgentLatency <= 0) {
+                               *NotEnoughUrgentLatencyHiding = 1;
+                               *UrgentBurstFactorCursor = 0;
+                       } else {
+                               *UrgentBurstFactorCursor = CursorBufferSizeInTime / (CursorBufferSizeInTime - UrgentLatency);
+                       }
+               } else {
+                       *UrgentBurstFactorCursor = 1;
+               }
+       }
+
+       LinesInDETLuma = (UseMALLForPStateChange == dml_use_mall_pstate_change_phantom_pipe ? 1024*1024 : DETBufferSizeY) / BytePerPixelInDETY / swath_width_luma_ub;
+
+       if (VRatio > 0) {
+               DETBufferSizeInTimeLuma = dml_floor(LinesInDETLuma, SwathHeightY) * LineTime / VRatio;
+               if (DETBufferSizeInTimeLuma - UrgentLatency <= 0) {
+                       *NotEnoughUrgentLatencyHiding = 1;
+                       *UrgentBurstFactorLuma = 0;
+               } else {
+                       *UrgentBurstFactorLuma = DETBufferSizeInTimeLuma / (DETBufferSizeInTimeLuma - UrgentLatency);
+               }
+       } else {
+               *UrgentBurstFactorLuma = 1;
+       }
+
+       if (BytePerPixelInDETC > 0) {
+               LinesInDETChroma = (UseMALLForPStateChange == dml_use_mall_pstate_change_phantom_pipe ? 1024*1024 : DETBufferSizeC) / BytePerPixelInDETC / swath_width_chroma_ub;
+
+               if (VRatioC > 0) {
+                       DETBufferSizeInTimeChroma = dml_floor(LinesInDETChroma, SwathHeightC) * LineTime / VRatioC;
+                       if (DETBufferSizeInTimeChroma - UrgentLatency <= 0) {
+                               *NotEnoughUrgentLatencyHiding = 1;
+                               *UrgentBurstFactorChroma = 0;
+                       } else {
+                               *UrgentBurstFactorChroma = DETBufferSizeInTimeChroma / (DETBufferSizeInTimeChroma - UrgentLatency);
+                       }
+               } else {
+                       *UrgentBurstFactorChroma = 1;
+               }
+       }
+} // CalculateUrgentBurstFactor
+
+static void CalculatePixelDeliveryTimes(
+               dml_uint_t NumberOfActiveSurfaces,
+               dml_float_t VRatio[],
+               dml_float_t VRatioChroma[],
+               dml_float_t VRatioPrefetchY[],
+               dml_float_t VRatioPrefetchC[],
+               dml_uint_t swath_width_luma_ub[],
+               dml_uint_t swath_width_chroma_ub[],
+               dml_uint_t DPPPerSurface[],
+               dml_float_t HRatio[],
+               dml_float_t HRatioChroma[],
+               dml_float_t PixelClock[],
+               dml_float_t PSCL_THROUGHPUT[],
+               dml_float_t PSCL_THROUGHPUT_CHROMA[],
+               dml_float_t Dppclk[],
+               dml_uint_t BytePerPixelC[],
+               enum dml_rotation_angle SourceScan[],
+               dml_uint_t NumberOfCursors[],
+               dml_uint_t CursorWidth[],
+               dml_uint_t CursorBPP[],
+               dml_uint_t BlockWidth256BytesY[],
+               dml_uint_t BlockHeight256BytesY[],
+               dml_uint_t BlockWidth256BytesC[],
+               dml_uint_t BlockHeight256BytesC[],
+
+               // Output
+               dml_float_t DisplayPipeLineDeliveryTimeLuma[],
+               dml_float_t DisplayPipeLineDeliveryTimeChroma[],
+               dml_float_t DisplayPipeLineDeliveryTimeLumaPrefetch[],
+               dml_float_t DisplayPipeLineDeliveryTimeChromaPrefetch[],
+               dml_float_t DisplayPipeRequestDeliveryTimeLuma[],
+               dml_float_t DisplayPipeRequestDeliveryTimeChroma[],
+               dml_float_t DisplayPipeRequestDeliveryTimeLumaPrefetch[],
+               dml_float_t DisplayPipeRequestDeliveryTimeChromaPrefetch[],
+               dml_float_t CursorRequestDeliveryTime[],
+               dml_float_t CursorRequestDeliveryTimePrefetch[])
+{
+               dml_float_t req_per_swath_ub;
+
+               for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u : HRatio = %f\n", __func__, k, HRatio[k]);
+               dml_print("DML::%s: k=%u : VRatio = %f\n", __func__, k, VRatio[k]);
+               dml_print("DML::%s: k=%u : HRatioChroma = %f\n", __func__, k, HRatioChroma[k]);
+               dml_print("DML::%s: k=%u : VRatioChroma = %f\n", __func__, k, VRatioChroma[k]);
+               dml_print("DML::%s: k=%u : swath_width_luma_ub = %u\n", __func__, k, swath_width_luma_ub[k]);
+               dml_print("DML::%s: k=%u : swath_width_chroma_ub = %u\n", __func__, k, swath_width_chroma_ub[k]);
+               dml_print("DML::%s: k=%u : PSCL_THROUGHPUT = %f\n", __func__, k, PSCL_THROUGHPUT[k]);
+               dml_print("DML::%s: k=%u : PSCL_THROUGHPUT_CHROMA = %f\n", __func__, k, PSCL_THROUGHPUT_CHROMA[k]);
+               dml_print("DML::%s: k=%u : DPPPerSurface = %u\n", __func__, k, DPPPerSurface[k]);
+               dml_print("DML::%s: k=%u : PixelClock = %f\n", __func__, k, PixelClock[k]);
+               dml_print("DML::%s: k=%u : Dppclk = %f\n", __func__, k, Dppclk[k]);
+#endif
+
+               if (VRatio[k] <= 1) {
+                       DisplayPipeLineDeliveryTimeLuma[k] = swath_width_luma_ub[k] * DPPPerSurface[k] / HRatio[k] / PixelClock[k];
+               } else {
+                       DisplayPipeLineDeliveryTimeLuma[k] = swath_width_luma_ub[k] / PSCL_THROUGHPUT[k] / Dppclk[k];
+               }
+
+               if (BytePerPixelC[k] == 0) {
+                       DisplayPipeLineDeliveryTimeChroma[k] = 0;
+               } else {
+                       if (VRatioChroma[k] <= 1) {
+                               DisplayPipeLineDeliveryTimeChroma[k] = swath_width_chroma_ub[k] * DPPPerSurface[k] / HRatioChroma[k] / PixelClock[k];
+                       } else {
+                               DisplayPipeLineDeliveryTimeChroma[k] = swath_width_chroma_ub[k] / PSCL_THROUGHPUT_CHROMA[k] / Dppclk[k];
+                       }
+               }
+
+               if (VRatioPrefetchY[k] <= 1) {
+                       DisplayPipeLineDeliveryTimeLumaPrefetch[k] = swath_width_luma_ub[k] * DPPPerSurface[k] / HRatio[k] / PixelClock[k];
+               } else {
+                       DisplayPipeLineDeliveryTimeLumaPrefetch[k] = swath_width_luma_ub[k] / PSCL_THROUGHPUT[k] / Dppclk[k];
+               }
+
+               if (BytePerPixelC[k] == 0) {
+                       DisplayPipeLineDeliveryTimeChromaPrefetch[k] = 0;
+               } else {
+                       if (VRatioPrefetchC[k] <= 1) {
+                               DisplayPipeLineDeliveryTimeChromaPrefetch[k] = swath_width_chroma_ub[k] * DPPPerSurface[k] / HRatioChroma[k] / PixelClock[k];
+                       } else {
+                               DisplayPipeLineDeliveryTimeChromaPrefetch[k] = swath_width_chroma_ub[k] / PSCL_THROUGHPUT_CHROMA[k] / Dppclk[k];
+                       }
+               }
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u : DisplayPipeLineDeliveryTimeLuma = %f\n", __func__, k, DisplayPipeLineDeliveryTimeLuma[k]);
+               dml_print("DML::%s: k=%u : DisplayPipeLineDeliveryTimeLumaPrefetch = %f\n", __func__, k, DisplayPipeLineDeliveryTimeLumaPrefetch[k]);
+               dml_print("DML::%s: k=%u : DisplayPipeLineDeliveryTimeChroma = %f\n", __func__, k, DisplayPipeLineDeliveryTimeChroma[k]);
+               dml_print("DML::%s: k=%u : DisplayPipeLineDeliveryTimeChromaPrefetch = %f\n", __func__, k, DisplayPipeLineDeliveryTimeChromaPrefetch[k]);
+#endif
+       }
+
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               if (!dml_is_vertical_rotation(SourceScan[k])) {
+                       req_per_swath_ub = swath_width_luma_ub[k] / BlockWidth256BytesY[k];
+               } else {
+                       req_per_swath_ub = swath_width_luma_ub[k] / BlockHeight256BytesY[k];
+               }
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u : req_per_swath_ub = %f (Luma)\n", __func__, k, req_per_swath_ub);
+#endif
+
+               DisplayPipeRequestDeliveryTimeLuma[k] = DisplayPipeLineDeliveryTimeLuma[k] / req_per_swath_ub;
+               DisplayPipeRequestDeliveryTimeLumaPrefetch[k] = DisplayPipeLineDeliveryTimeLumaPrefetch[k] / req_per_swath_ub;
+               if (BytePerPixelC[k] == 0) {
+                       DisplayPipeRequestDeliveryTimeChroma[k] = 0;
+                       DisplayPipeRequestDeliveryTimeChromaPrefetch[k] = 0;
+               } else {
+                       if (!dml_is_vertical_rotation(SourceScan[k])) {
+                               req_per_swath_ub = swath_width_chroma_ub[k] / BlockWidth256BytesC[k];
+                       } else {
+                               req_per_swath_ub = swath_width_chroma_ub[k] / BlockHeight256BytesC[k];
+                       }
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%u : req_per_swath_ub = %f (Chroma)\n", __func__, k, req_per_swath_ub);
+#endif
+                       DisplayPipeRequestDeliveryTimeChroma[k] = DisplayPipeLineDeliveryTimeChroma[k] / req_per_swath_ub;
+                       DisplayPipeRequestDeliveryTimeChromaPrefetch[k] = DisplayPipeLineDeliveryTimeChromaPrefetch[k] / req_per_swath_ub;
+               }
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeLuma = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeLuma[k]);
+               dml_print("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeLumaPrefetch = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeLumaPrefetch[k]);
+               dml_print("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeChroma = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeChroma[k]);
+               dml_print("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeChromaPrefetch = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeChromaPrefetch[k]);
+#endif
+       }
+
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               dml_uint_t cursor_req_per_width;
+               cursor_req_per_width = (dml_uint_t)(dml_ceil((dml_float_t) CursorWidth[k] * (dml_float_t) CursorBPP[k] / 256.0 / 8.0, 1.0));
+               if (NumberOfCursors[k] > 0) {
+                       if (VRatio[k] <= 1) {
+                               CursorRequestDeliveryTime[k] = (dml_float_t) CursorWidth[k] / HRatio[k] / PixelClock[k] / cursor_req_per_width;
+                       } else {
+                               CursorRequestDeliveryTime[k] = (dml_float_t) CursorWidth[k] / PSCL_THROUGHPUT[k] / Dppclk[k] / cursor_req_per_width;
+                       }
+                       if (VRatioPrefetchY[k] <= 1) {
+                               CursorRequestDeliveryTimePrefetch[k] = (dml_float_t) CursorWidth[k] / HRatio[k] / PixelClock[k] / cursor_req_per_width;
+               } else {
+                       CursorRequestDeliveryTimePrefetch[k] = (dml_float_t) CursorWidth[k] / PSCL_THROUGHPUT[k] / Dppclk[k] / cursor_req_per_width;
+                       }
+               } else {
+                       CursorRequestDeliveryTime[k] = 0;
+                       CursorRequestDeliveryTimePrefetch[k] = 0;
+               }
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u : NumberOfCursors = %u\n", __func__, k, NumberOfCursors[k]);
+               dml_print("DML::%s: k=%u : CursorRequestDeliveryTime = %f\n", __func__, k, CursorRequestDeliveryTime[k]);
+               dml_print("DML::%s: k=%u : CursorRequestDeliveryTimePrefetch = %f\n", __func__, k, CursorRequestDeliveryTimePrefetch[k]);
+#endif
+       }
+} // CalculatePixelDeliveryTimes
+
+static void CalculateMetaAndPTETimes(
+               dml_bool_t use_one_row_for_frame[],
+               dml_uint_t NumberOfActiveSurfaces,
+               dml_bool_t GPUVMEnable,
+               dml_uint_t MetaChunkSize,
+               dml_uint_t MinMetaChunkSizeBytes,
+               dml_uint_t HTotal[],
+               dml_float_t VRatio[],
+               dml_float_t VRatioChroma[],
+               dml_float_t DestinationLinesToRequestRowInVBlank[],
+               dml_float_t DestinationLinesToRequestRowInImmediateFlip[],
+               dml_bool_t DCCEnable[],
+               dml_float_t PixelClock[],
+               dml_uint_t BytePerPixelY[],
+               dml_uint_t BytePerPixelC[],
+               enum dml_rotation_angle SourceScan[],
+               dml_uint_t dpte_row_height[],
+               dml_uint_t dpte_row_height_chroma[],
+               dml_uint_t meta_row_width[],
+               dml_uint_t meta_row_width_chroma[],
+               dml_uint_t meta_row_height[],
+               dml_uint_t meta_row_height_chroma[],
+               dml_uint_t meta_req_width[],
+               dml_uint_t meta_req_width_chroma[],
+               dml_uint_t meta_req_height[],
+               dml_uint_t meta_req_height_chroma[],
+               dml_uint_t dpte_group_bytes[],
+               dml_uint_t PTERequestSizeY[],
+               dml_uint_t PTERequestSizeC[],
+               dml_uint_t PixelPTEReqWidthY[],
+               dml_uint_t PixelPTEReqHeightY[],
+               dml_uint_t PixelPTEReqWidthC[],
+               dml_uint_t PixelPTEReqHeightC[],
+               dml_uint_t dpte_row_width_luma_ub[],
+               dml_uint_t dpte_row_width_chroma_ub[],
+
+               // Output
+               dml_float_t DST_Y_PER_PTE_ROW_NOM_L[],
+               dml_float_t DST_Y_PER_PTE_ROW_NOM_C[],
+               dml_float_t DST_Y_PER_META_ROW_NOM_L[],
+               dml_float_t DST_Y_PER_META_ROW_NOM_C[],
+               dml_float_t TimePerMetaChunkNominal[],
+               dml_float_t TimePerChromaMetaChunkNominal[],
+               dml_float_t TimePerMetaChunkVBlank[],
+               dml_float_t TimePerChromaMetaChunkVBlank[],
+               dml_float_t TimePerMetaChunkFlip[],
+               dml_float_t TimePerChromaMetaChunkFlip[],
+               dml_float_t time_per_pte_group_nom_luma[],
+               dml_float_t time_per_pte_group_vblank_luma[],
+               dml_float_t time_per_pte_group_flip_luma[],
+               dml_float_t time_per_pte_group_nom_chroma[],
+               dml_float_t time_per_pte_group_vblank_chroma[],
+               dml_float_t time_per_pte_group_flip_chroma[])
+{
+       dml_uint_t meta_chunk_width;
+       dml_uint_t min_meta_chunk_width;
+       dml_uint_t meta_chunk_per_row_int;
+       dml_uint_t meta_row_remainder;
+       dml_uint_t meta_chunk_threshold;
+       dml_uint_t meta_chunks_per_row_ub;
+       dml_uint_t meta_chunk_width_chroma;
+       dml_uint_t min_meta_chunk_width_chroma;
+       dml_uint_t meta_chunk_per_row_int_chroma;
+       dml_uint_t meta_row_remainder_chroma;
+       dml_uint_t meta_chunk_threshold_chroma;
+       dml_uint_t meta_chunks_per_row_ub_chroma;
+       dml_uint_t dpte_group_width_luma;
+       dml_uint_t dpte_groups_per_row_luma_ub;
+       dml_uint_t dpte_group_width_chroma;
+       dml_uint_t dpte_groups_per_row_chroma_ub;
+
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               DST_Y_PER_PTE_ROW_NOM_L[k] = dpte_row_height[k] / VRatio[k];
+               if (BytePerPixelC[k] == 0) {
+                       DST_Y_PER_PTE_ROW_NOM_C[k] = 0;
+               } else {
+                       DST_Y_PER_PTE_ROW_NOM_C[k] = dpte_row_height_chroma[k] / VRatioChroma[k];
+               }
+               DST_Y_PER_META_ROW_NOM_L[k] = meta_row_height[k] / VRatio[k];
+               if (BytePerPixelC[k] == 0) {
+                       DST_Y_PER_META_ROW_NOM_C[k] = 0;
+               } else {
+                       DST_Y_PER_META_ROW_NOM_C[k] = meta_row_height_chroma[k] / VRatioChroma[k];
+               }
+       }
+
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               if (DCCEnable[k] == true) {
+                       meta_chunk_width = MetaChunkSize * 1024 * 256 / BytePerPixelY[k] / meta_row_height[k];
+                       min_meta_chunk_width = MinMetaChunkSizeBytes * 256 / BytePerPixelY[k] / meta_row_height[k];
+                       meta_chunk_per_row_int = meta_row_width[k] / meta_chunk_width;
+                       meta_row_remainder = meta_row_width[k] % meta_chunk_width;
+                       if (!dml_is_vertical_rotation(SourceScan[k])) {
+                               meta_chunk_threshold = 2 * min_meta_chunk_width - meta_req_width[k];
+                       } else {
+                               meta_chunk_threshold = 2 * min_meta_chunk_width - meta_req_height[k];
+                       }
+                       if (meta_row_remainder <= meta_chunk_threshold) {
+                               meta_chunks_per_row_ub = meta_chunk_per_row_int + 1;
+                       } else {
+                               meta_chunks_per_row_ub = meta_chunk_per_row_int + 2;
+                       }
+                       TimePerMetaChunkNominal[k] = meta_row_height[k] / VRatio[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub;
+                       TimePerMetaChunkVBlank[k] = DestinationLinesToRequestRowInVBlank[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub;
+                       TimePerMetaChunkFlip[k] = DestinationLinesToRequestRowInImmediateFlip[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub;
+                       if (BytePerPixelC[k] == 0) {
+                               TimePerChromaMetaChunkNominal[k] = 0;
+                               TimePerChromaMetaChunkVBlank[k] = 0;
+                               TimePerChromaMetaChunkFlip[k] = 0;
+                       } else {
+                               meta_chunk_width_chroma = MetaChunkSize * 1024 * 256 / BytePerPixelC[k] / meta_row_height_chroma[k];
+                               min_meta_chunk_width_chroma = MinMetaChunkSizeBytes * 256 / BytePerPixelC[k] / meta_row_height_chroma[k];
+                               meta_chunk_per_row_int_chroma = (dml_uint_t)((dml_float_t) meta_row_width_chroma[k] / meta_chunk_width_chroma);
+                               meta_row_remainder_chroma = meta_row_width_chroma[k] % meta_chunk_width_chroma;
+                               if (!dml_is_vertical_rotation(SourceScan[k])) {
+                                       meta_chunk_threshold_chroma = 2 * min_meta_chunk_width_chroma - meta_req_width_chroma[k];
+                               } else {
+                                       meta_chunk_threshold_chroma = 2 * min_meta_chunk_width_chroma - meta_req_height_chroma[k];
+                               }
+                               if (meta_row_remainder_chroma <= meta_chunk_threshold_chroma) {
+                                       meta_chunks_per_row_ub_chroma = meta_chunk_per_row_int_chroma + 1;
+                               } else {
+                                       meta_chunks_per_row_ub_chroma = meta_chunk_per_row_int_chroma + 2;
+                               }
+                               TimePerChromaMetaChunkNominal[k] = meta_row_height_chroma[k] / VRatioChroma[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub_chroma;
+                               TimePerChromaMetaChunkVBlank[k] = DestinationLinesToRequestRowInVBlank[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub_chroma;
+                               TimePerChromaMetaChunkFlip[k] = DestinationLinesToRequestRowInImmediateFlip[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub_chroma;
+                       }
+               } else {
+                       TimePerMetaChunkNominal[k] = 0;
+                       TimePerMetaChunkVBlank[k] = 0;
+                       TimePerMetaChunkFlip[k] = 0;
+                       TimePerChromaMetaChunkNominal[k] = 0;
+                       TimePerChromaMetaChunkVBlank[k] = 0;
+                       TimePerChromaMetaChunkFlip[k] = 0;
+               }
+       }
+
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               if (GPUVMEnable == true) {
+                       if (!dml_is_vertical_rotation(SourceScan[k])) {
+                               dpte_group_width_luma = (dml_uint_t)((dml_float_t) dpte_group_bytes[k] / (dml_float_t) PTERequestSizeY[k] * PixelPTEReqWidthY[k]);
+                       } else {
+                               dpte_group_width_luma = (dml_uint_t)((dml_float_t) dpte_group_bytes[k] / (dml_float_t) PTERequestSizeY[k] * PixelPTEReqHeightY[k]);
+                       }
+
+                       if (use_one_row_for_frame[k]) {
+                               dpte_groups_per_row_luma_ub = (dml_uint_t)(dml_ceil((dml_float_t) dpte_row_width_luma_ub[k] / (dml_float_t) dpte_group_width_luma / 2.0, 1.0));
+                       } else {
+                               dpte_groups_per_row_luma_ub = (dml_uint_t)(dml_ceil((dml_float_t) dpte_row_width_luma_ub[k] / (dml_float_t) dpte_group_width_luma, 1.0));
+                       }
+
+                       dml_print("DML::%s: k=%u, use_one_row_for_frame = %u\n", __func__, k, use_one_row_for_frame[k]);
+                       dml_print("DML::%s: k=%u, dpte_group_bytes = %u\n", __func__, k, dpte_group_bytes[k]);
+                       dml_print("DML::%s: k=%u, PTERequestSizeY = %u\n", __func__, k, PTERequestSizeY[k]);
+                       dml_print("DML::%s: k=%u, PixelPTEReqWidthY = %u\n", __func__, k, PixelPTEReqWidthY[k]);
+                       dml_print("DML::%s: k=%u, PixelPTEReqHeightY = %u\n", __func__, k, PixelPTEReqHeightY[k]);
+                       dml_print("DML::%s: k=%u, dpte_row_width_luma_ub = %u\n", __func__, k, dpte_row_width_luma_ub[k]);
+                       dml_print("DML::%s: k=%u, dpte_group_width_luma = %u\n", __func__, k, dpte_group_width_luma);
+                       dml_print("DML::%s: k=%u, dpte_groups_per_row_luma_ub = %u\n", __func__, k, dpte_groups_per_row_luma_ub);
+
+                       time_per_pte_group_nom_luma[k] = DST_Y_PER_PTE_ROW_NOM_L[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_luma_ub;
+                       time_per_pte_group_vblank_luma[k] = DestinationLinesToRequestRowInVBlank[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_luma_ub;
+                       time_per_pte_group_flip_luma[k] = DestinationLinesToRequestRowInImmediateFlip[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_luma_ub;
+                       if (BytePerPixelC[k] == 0) {
+                               time_per_pte_group_nom_chroma[k] = 0;
+                               time_per_pte_group_vblank_chroma[k] = 0;
+                               time_per_pte_group_flip_chroma[k] = 0;
+                       } else {
+                               if (!dml_is_vertical_rotation(SourceScan[k])) {
+                                       dpte_group_width_chroma = (dml_uint_t)((dml_float_t) dpte_group_bytes[k] / (dml_float_t) PTERequestSizeC[k] * PixelPTEReqWidthC[k]);
+                               } else {
+                                       dpte_group_width_chroma = (dml_uint_t)((dml_float_t) dpte_group_bytes[k] / (dml_float_t) PTERequestSizeC[k] * PixelPTEReqHeightC[k]);
+                               }
+
+                               if (use_one_row_for_frame[k]) {
+                                       dpte_groups_per_row_chroma_ub = (dml_uint_t)(dml_ceil((dml_float_t) dpte_row_width_chroma_ub[k] / (dml_float_t) dpte_group_width_chroma / 2.0, 1.0));
+                               } else {
+                                       dpte_groups_per_row_chroma_ub = (dml_uint_t)(dml_ceil((dml_float_t) dpte_row_width_chroma_ub[k] / (dml_float_t) dpte_group_width_chroma, 1.0));
+                               }
+                               dml_print("DML::%s: k=%u, dpte_row_width_chroma_ub = %u\n", __func__, k, dpte_row_width_chroma_ub[k]);
+                               dml_print("DML::%s: k=%u, dpte_group_width_chroma = %u\n", __func__, k, dpte_group_width_chroma);
+                               dml_print("DML::%s: k=%u, dpte_groups_per_row_chroma_ub = %u\n", __func__, k, dpte_groups_per_row_chroma_ub);
+
+                               time_per_pte_group_nom_chroma[k] = DST_Y_PER_PTE_ROW_NOM_C[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_chroma_ub;
+                               time_per_pte_group_vblank_chroma[k] = DestinationLinesToRequestRowInVBlank[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_chroma_ub;
+                               time_per_pte_group_flip_chroma[k] = DestinationLinesToRequestRowInImmediateFlip[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_chroma_ub;
+                       }
+               } else {
+                       time_per_pte_group_nom_luma[k] = 0;
+                       time_per_pte_group_vblank_luma[k] = 0;
+                       time_per_pte_group_flip_luma[k] = 0;
+                       time_per_pte_group_nom_chroma[k] = 0;
+                       time_per_pte_group_vblank_chroma[k] = 0;
+                       time_per_pte_group_flip_chroma[k] = 0;
+               }
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, DestinationLinesToRequestRowInVBlank = %f\n", __func__, k, DestinationLinesToRequestRowInVBlank[k]);
+               dml_print("DML::%s: k=%u, DestinationLinesToRequestRowInImmediateFlip = %f\n", __func__, k, DestinationLinesToRequestRowInImmediateFlip[k]);
+
+               dml_print("DML::%s: k=%u, DST_Y_PER_PTE_ROW_NOM_L = %f\n", __func__, k, DST_Y_PER_PTE_ROW_NOM_L[k]);
+               dml_print("DML::%s: k=%u, DST_Y_PER_PTE_ROW_NOM_C = %f\n", __func__, k, DST_Y_PER_PTE_ROW_NOM_C[k]);
+               dml_print("DML::%s: k=%u, DST_Y_PER_META_ROW_NOM_L = %f\n", __func__, k, DST_Y_PER_META_ROW_NOM_L[k]);
+               dml_print("DML::%s: k=%u, DST_Y_PER_META_ROW_NOM_C = %f\n", __func__, k, DST_Y_PER_META_ROW_NOM_C[k]);
+               dml_print("DML::%s: k=%u, TimePerMetaChunkNominal = %f\n", __func__, k, TimePerMetaChunkNominal[k]);
+               dml_print("DML::%s: k=%u, TimePerMetaChunkVBlank = %f\n", __func__, k, TimePerMetaChunkVBlank[k]);
+               dml_print("DML::%s: k=%u, TimePerMetaChunkFlip = %f\n", __func__, k, TimePerMetaChunkFlip[k]);
+               dml_print("DML::%s: k=%u, TimePerChromaMetaChunkNominal = %f\n", __func__, k, TimePerChromaMetaChunkNominal[k]);
+               dml_print("DML::%s: k=%u, TimePerChromaMetaChunkVBlank = %f\n", __func__, k, TimePerChromaMetaChunkVBlank[k]);
+               dml_print("DML::%s: k=%u, TimePerChromaMetaChunkFlip = %f\n", __func__, k, TimePerChromaMetaChunkFlip[k]);
+               dml_print("DML::%s: k=%u, time_per_pte_group_nom_luma = %f\n", __func__, k, time_per_pte_group_nom_luma[k]);
+               dml_print("DML::%s: k=%u, time_per_pte_group_vblank_luma = %f\n", __func__, k, time_per_pte_group_vblank_luma[k]);
+               dml_print("DML::%s: k=%u, time_per_pte_group_flip_luma = %f\n", __func__, k, time_per_pte_group_flip_luma[k]);
+               dml_print("DML::%s: k=%u, time_per_pte_group_nom_chroma = %f\n", __func__, k, time_per_pte_group_nom_chroma[k]);
+               dml_print("DML::%s: k=%u, time_per_pte_group_vblank_chroma = %f\n", __func__, k, time_per_pte_group_vblank_chroma[k]);
+               dml_print("DML::%s: k=%u, time_per_pte_group_flip_chroma   = %f\n", __func__, k, time_per_pte_group_flip_chroma[k]);
+#endif
+       }
+} // CalculateMetaAndPTETimes
+
+static void CalculateVMGroupAndRequestTimes(
+               dml_uint_t NumberOfActiveSurfaces,
+               dml_bool_t GPUVMEnable,
+               dml_uint_t GPUVMMaxPageTableLevels,
+               dml_uint_t HTotal[],
+               dml_uint_t BytePerPixelC[],
+               dml_float_t DestinationLinesToRequestVMInVBlank[],
+               dml_float_t DestinationLinesToRequestVMInImmediateFlip[],
+               dml_bool_t DCCEnable[],
+               dml_float_t PixelClock[],
+               dml_uint_t dpte_row_width_luma_ub[],
+               dml_uint_t dpte_row_width_chroma_ub[],
+               dml_uint_t vm_group_bytes[],
+               dml_uint_t dpde0_bytes_per_frame_ub_l[],
+               dml_uint_t dpde0_bytes_per_frame_ub_c[],
+               dml_uint_t meta_pte_bytes_per_frame_ub_l[],
+               dml_uint_t meta_pte_bytes_per_frame_ub_c[],
+
+               // Output
+               dml_float_t TimePerVMGroupVBlank[],
+               dml_float_t TimePerVMGroupFlip[],
+               dml_float_t TimePerVMRequestVBlank[],
+               dml_float_t TimePerVMRequestFlip[])
+{
+       dml_uint_t num_group_per_lower_vm_stage;
+       dml_uint_t num_req_per_lower_vm_stage;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces);
+       dml_print("DML::%s: GPUVMEnable = %u\n", __func__, GPUVMEnable);
+#endif
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, DCCEnable = %u\n", __func__, k, DCCEnable[k]);
+               dml_print("DML::%s: k=%u, vm_group_bytes = %u\n", __func__, k, vm_group_bytes[k]);
+               dml_print("DML::%s: k=%u, dpde0_bytes_per_frame_ub_l = %u\n", __func__, k, dpde0_bytes_per_frame_ub_l[k]);
+               dml_print("DML::%s: k=%u, dpde0_bytes_per_frame_ub_c = %u\n", __func__, k, dpde0_bytes_per_frame_ub_c[k]);
+               dml_print("DML::%s: k=%u, meta_pte_bytes_per_frame_ub_l = %u\n", __func__, k, meta_pte_bytes_per_frame_ub_l[k]);
+               dml_print("DML::%s: k=%u, meta_pte_bytes_per_frame_ub_c = %u\n", __func__, k, meta_pte_bytes_per_frame_ub_c[k]);
+#endif
+
+               if (GPUVMEnable == true && (DCCEnable[k] == true || GPUVMMaxPageTableLevels > 1)) {
+                       if (DCCEnable[k] == false) {
+                               if (BytePerPixelC[k] > 0) {
+                                       num_group_per_lower_vm_stage = (dml_uint_t) (dml_ceil((dml_float_t) dpde0_bytes_per_frame_ub_l[k] / (dml_float_t) vm_group_bytes[k], 1.0) +
+                                                                                                       dml_ceil((dml_float_t) dpde0_bytes_per_frame_ub_c[k] / (dml_float_t) vm_group_bytes[k], 1.0));
+                               } else {
+                                       num_group_per_lower_vm_stage = (dml_uint_t) (dml_ceil((dml_float_t) dpde0_bytes_per_frame_ub_l[k] / (dml_float_t) vm_group_bytes[k], 1.0));
+                               }
+                       } else {
+                               if (GPUVMMaxPageTableLevels == 1) {
+                                       if (BytePerPixelC[k] > 0) {
+                                               num_group_per_lower_vm_stage = (dml_uint_t)(dml_ceil((dml_float_t) (meta_pte_bytes_per_frame_ub_l[k]) / (dml_float_t) (vm_group_bytes[k]), 1.0) +
+                                                                                                                                       dml_ceil((dml_float_t) (meta_pte_bytes_per_frame_ub_c[k]) / (dml_float_t) (vm_group_bytes[k]), 1.0));
+                                       } else {
+                                               num_group_per_lower_vm_stage = (dml_uint_t)(dml_ceil((dml_float_t) (meta_pte_bytes_per_frame_ub_l[k]) / (dml_float_t) (vm_group_bytes[k]), 1.0));
+                                       }
+                               } else {
+                                       if (BytePerPixelC[k] > 0) {
+                                               num_group_per_lower_vm_stage = (dml_uint_t)(2.0 + dml_ceil((dml_float_t) (dpde0_bytes_per_frame_ub_l[k]) / (dml_float_t) (vm_group_bytes[k]), 1) +
+                                                                                                               dml_ceil((dml_float_t) (dpde0_bytes_per_frame_ub_c[k]) / (dml_float_t) (vm_group_bytes[k]), 1) +
+                                                                                                               dml_ceil((dml_float_t) (meta_pte_bytes_per_frame_ub_l[k]) / (dml_float_t) (vm_group_bytes[k]), 1) +
+                                                                                                               dml_ceil((dml_float_t) (meta_pte_bytes_per_frame_ub_c[k]) / (dml_float_t) (vm_group_bytes[k]), 1));
+                                       } else {
+                                               num_group_per_lower_vm_stage = (dml_uint_t)(1.0 + dml_ceil((dml_float_t) (dpde0_bytes_per_frame_ub_l[k]) / (dml_float_t) (vm_group_bytes[k]), 1) +
+                                                                                                                                       dml_ceil((dml_float_t) (meta_pte_bytes_per_frame_ub_l[k]) / (dml_float_t) (vm_group_bytes[k]), 1));
+                                       }
+                               }
+                       }
+
+                       if (DCCEnable[k] == false) {
+                               if (BytePerPixelC[k] > 0) {
+                                       num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] / 64 + dpde0_bytes_per_frame_ub_c[k] / 64;
+                               } else {
+                                       num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] / 64;
+                               }
+                       } else {
+                               if (GPUVMMaxPageTableLevels == 1) {
+                                       if (BytePerPixelC[k] > 0) {
+                                               num_req_per_lower_vm_stage = meta_pte_bytes_per_frame_ub_l[k] / 64 + meta_pte_bytes_per_frame_ub_c[k] / 64;
+                                       } else {
+                                               num_req_per_lower_vm_stage = meta_pte_bytes_per_frame_ub_l[k] / 64;
+                                       }
+                               } else {
+                                       if (BytePerPixelC[k] > 0) {
+                                               num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] / 64 + dpde0_bytes_per_frame_ub_c[k] / 64 + meta_pte_bytes_per_frame_ub_l[k] / 64 + meta_pte_bytes_per_frame_ub_c[k] / 64;
+                                       } else {
+                                               num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] / 64 + meta_pte_bytes_per_frame_ub_l[k] / 64;
+                                       }
+                               }
+                       }
+
+                       TimePerVMGroupVBlank[k] = DestinationLinesToRequestVMInVBlank[k] * HTotal[k] / PixelClock[k] / num_group_per_lower_vm_stage;
+                       TimePerVMGroupFlip[k] = DestinationLinesToRequestVMInImmediateFlip[k] * HTotal[k] / PixelClock[k] / num_group_per_lower_vm_stage;
+                       TimePerVMRequestVBlank[k] = DestinationLinesToRequestVMInVBlank[k] * HTotal[k] / PixelClock[k] / num_req_per_lower_vm_stage;
+                       TimePerVMRequestFlip[k] = DestinationLinesToRequestVMInImmediateFlip[k] * HTotal[k] / PixelClock[k] / num_req_per_lower_vm_stage;
+
+                       if (GPUVMMaxPageTableLevels > 2) {
+                               TimePerVMGroupVBlank[k] = TimePerVMGroupVBlank[k] / 2;
+                               TimePerVMGroupFlip[k] = TimePerVMGroupFlip[k] / 2;
+                               TimePerVMRequestVBlank[k] = TimePerVMRequestVBlank[k] / 2;
+                               TimePerVMRequestFlip[k]    = TimePerVMRequestFlip[k] / 2;
+                       }
+
+               } else {
+                       TimePerVMGroupVBlank[k] = 0;
+                       TimePerVMGroupFlip[k] = 0;
+                       TimePerVMRequestVBlank[k] = 0;
+                       TimePerVMRequestFlip[k] = 0;
+               }
+
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%u, TimePerVMGroupVBlank = %f\n", __func__, k, TimePerVMGroupVBlank[k]);
+                       dml_print("DML::%s: k=%u, TimePerVMGroupFlip = %f\n", __func__, k, TimePerVMGroupFlip[k]);
+                       dml_print("DML::%s: k=%u, TimePerVMRequestVBlank = %f\n", __func__, k, TimePerVMRequestVBlank[k]);
+                       dml_print("DML::%s: k=%u, TimePerVMRequestFlip = %f\n", __func__, k, TimePerVMRequestFlip[k]);
+#endif
+       }
+} // CalculateVMGroupAndRequestTimes
+
+static void CalculateStutterEfficiency(struct display_mode_lib_scratch_st *scratch,
+               struct CalculateStutterEfficiency_params_st *p)
+{
+       dml_float_t DETBufferingTimeY = 0;
+       dml_float_t SwathWidthYCriticalSurface = 0;
+       dml_float_t SwathHeightYCriticalSurface = 0;
+       dml_float_t VActiveTimeCriticalSurface = 0;
+       dml_float_t FrameTimeCriticalSurface = 0;
+       dml_uint_t BytePerPixelYCriticalSurface = 0;
+       dml_float_t LinesToFinishSwathTransferStutterCriticalSurface = 0;
+       dml_uint_t DETBufferSizeYCriticalSurface = 0;
+       dml_float_t MinTTUVBlankCriticalSurface = 0;
+       dml_uint_t BlockWidth256BytesYCriticalSurface = 0;
+       dml_bool_t SinglePlaneCriticalSurface = 0;
+       dml_bool_t SinglePipeCriticalSurface = 0;
+       dml_float_t TotalCompressedReadBandwidth = 0;
+       dml_float_t TotalRowReadBandwidth = 0;
+       dml_float_t AverageDCCCompressionRate = 0;
+       dml_float_t EffectiveCompressedBufferSize = 0;
+       dml_float_t PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = 0;
+       dml_float_t StutterBurstTime = 0;
+       dml_uint_t TotalActiveWriteback = 0;
+       dml_float_t LinesInDETY = 0;
+       dml_float_t LinesInDETYRoundedDownToSwath = 0;
+       dml_float_t MaximumEffectiveCompressionLuma = 0;
+       dml_float_t MaximumEffectiveCompressionChroma = 0;
+       dml_float_t TotalZeroSizeRequestReadBandwidth = 0;
+       dml_float_t TotalZeroSizeCompressedReadBandwidth = 0;
+       dml_float_t AverageDCCZeroSizeFraction = 0;
+       dml_float_t AverageZeroSizeCompressionRate = 0;
+
+       dml_bool_t FoundCriticalSurface = false;
+
+       dml_uint_t TotalNumberOfActiveOTG = 0;
+       dml_float_t SinglePixelClock;
+       dml_uint_t SingleHTotal;
+       dml_uint_t SingleVTotal;
+       dml_bool_t SameTiming = true;
+
+       dml_float_t LastStutterPeriod = 0.0;
+       dml_float_t LastZ8StutterPeriod = 0.0;
+
+       dml_uint_t SwathSizeCriticalSurface;
+       dml_uint_t LastChunkOfSwathSize;
+       dml_uint_t MissingPartOfLastSwathOfDETSize;
+
+       TotalZeroSizeRequestReadBandwidth = 0;
+       TotalZeroSizeCompressedReadBandwidth = 0;
+       TotalRowReadBandwidth = 0;
+       TotalCompressedReadBandwidth = 0;
+
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+       if (p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) {
+               if (p->DCCEnable[k] == true) {
+                       if ((dml_is_vertical_rotation(p->SourceScan[k]) && p->BlockWidth256BytesY[k] > p->SwathHeightY[k]) || (!dml_is_vertical_rotation(p->SourceScan[k]) && p->BlockHeight256BytesY[k] > p->SwathHeightY[k]) || p->DCCYMaxUncompressedBlock[k] < 256) {
+                               MaximumEffectiveCompressionLuma = 2;
+                       } else {
+                               MaximumEffectiveCompressionLuma = 4;
+                       }
+                       TotalCompressedReadBandwidth = TotalCompressedReadBandwidth + p->ReadBandwidthSurfaceLuma[k] / dml_min(p->NetDCCRateLuma[k], MaximumEffectiveCompressionLuma);
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%u, ReadBandwidthSurfaceLuma = %f\n", __func__, k, p->ReadBandwidthSurfaceLuma[k]);
+                       dml_print("DML::%s: k=%u, NetDCCRateLuma = %f\n", __func__, k, p->NetDCCRateLuma[k]);
+                       dml_print("DML::%s: k=%u, MaximumEffectiveCompressionLuma = %f\n", __func__, k, MaximumEffectiveCompressionLuma);
+#endif
+                       TotalZeroSizeRequestReadBandwidth = TotalZeroSizeRequestReadBandwidth + p->ReadBandwidthSurfaceLuma[k] * p->DCCFractionOfZeroSizeRequestsLuma[k];
+                       TotalZeroSizeCompressedReadBandwidth = TotalZeroSizeCompressedReadBandwidth + p->ReadBandwidthSurfaceLuma[k] * p->DCCFractionOfZeroSizeRequestsLuma[k] / MaximumEffectiveCompressionLuma;
+
+                       if (p->ReadBandwidthSurfaceChroma[k] > 0) {
+                               if ((dml_is_vertical_rotation(p->SourceScan[k]) && p->BlockWidth256BytesC[k] > p->SwathHeightC[k]) || (!dml_is_vertical_rotation(p->SourceScan[k]) && p->BlockHeight256BytesC[k] > p->SwathHeightC[k]) || p->DCCCMaxUncompressedBlock[k] < 256) {
+                                       MaximumEffectiveCompressionChroma = 2;
+                               } else {
+                                       MaximumEffectiveCompressionChroma = 4;
+                               }
+                               TotalCompressedReadBandwidth = TotalCompressedReadBandwidth + p->ReadBandwidthSurfaceChroma[k] / dml_min(p->NetDCCRateChroma[k], MaximumEffectiveCompressionChroma);
+#ifdef __DML_VBA_DEBUG__
+                               dml_print("DML::%s: k=%u, ReadBandwidthSurfaceChroma = %f\n", __func__, k, p->ReadBandwidthSurfaceChroma[k]);
+                               dml_print("DML::%s: k=%u, NetDCCRateChroma = %f\n", __func__, k, p->NetDCCRateChroma[k]);
+                               dml_print("DML::%s: k=%u, MaximumEffectiveCompressionChroma = %f\n", __func__, k, MaximumEffectiveCompressionChroma);
+#endif
+                               TotalZeroSizeRequestReadBandwidth = TotalZeroSizeRequestReadBandwidth + p->ReadBandwidthSurfaceChroma[k] * p->DCCFractionOfZeroSizeRequestsChroma[k];
+                               TotalZeroSizeCompressedReadBandwidth = TotalZeroSizeCompressedReadBandwidth + p->ReadBandwidthSurfaceChroma[k] * p->DCCFractionOfZeroSizeRequestsChroma[k] / MaximumEffectiveCompressionChroma;
+                       }
+               } else {
+                       TotalCompressedReadBandwidth = TotalCompressedReadBandwidth + p->ReadBandwidthSurfaceLuma[k] + p->ReadBandwidthSurfaceChroma[k];
+               }
+               TotalRowReadBandwidth = TotalRowReadBandwidth + p->DPPPerSurface[k] * (p->meta_row_bw[k] + p->dpte_row_bw[k]);
+       }
+       }
+
+       AverageDCCCompressionRate = p->TotalDataReadBandwidth / TotalCompressedReadBandwidth;
+       AverageDCCZeroSizeFraction = TotalZeroSizeRequestReadBandwidth / p->TotalDataReadBandwidth;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: UnboundedRequestEnabled = %u\n", __func__, p->UnboundedRequestEnabled);
+       dml_print("DML::%s: TotalCompressedReadBandwidth = %f\n", __func__, TotalCompressedReadBandwidth);
+       dml_print("DML::%s: TotalZeroSizeRequestReadBandwidth = %f\n", __func__, TotalZeroSizeRequestReadBandwidth);
+       dml_print("DML::%s: TotalZeroSizeCompressedReadBandwidth = %f\n", __func__, TotalZeroSizeCompressedReadBandwidth);
+       dml_print("DML::%s: MaximumEffectiveCompressionLuma = %f\n", __func__, MaximumEffectiveCompressionLuma);
+       dml_print("DML::%s: MaximumEffectiveCompressionChroma = %f\n", __func__, MaximumEffectiveCompressionChroma);
+       dml_print("DML::%s: AverageDCCCompressionRate = %f\n", __func__, AverageDCCCompressionRate);
+       dml_print("DML::%s: AverageDCCZeroSizeFraction = %f\n", __func__, AverageDCCZeroSizeFraction);
+       dml_print("DML::%s: CompbufReservedSpace64B = %u\n", __func__, p->CompbufReservedSpace64B);
+       dml_print("DML::%s: CompbufReservedSpaceZs = %u\n", __func__, p->CompbufReservedSpaceZs);
+       dml_print("DML::%s: CompressedBufferSizeInkByte = %u\n", __func__, p->CompressedBufferSizeInkByte);
+#endif
+       if (AverageDCCZeroSizeFraction == 1) {
+               AverageZeroSizeCompressionRate = TotalZeroSizeRequestReadBandwidth / TotalZeroSizeCompressedReadBandwidth;
+               EffectiveCompressedBufferSize = (dml_float_t)p->MetaFIFOSizeInKEntries * 1024 * 64 * AverageZeroSizeCompressionRate + ((dml_float_t)p->ZeroSizeBufferEntries - p->CompbufReservedSpaceZs) * 64 * AverageZeroSizeCompressionRate;
+       }       else if (AverageDCCZeroSizeFraction > 0) {
+               AverageZeroSizeCompressionRate = TotalZeroSizeRequestReadBandwidth / TotalZeroSizeCompressedReadBandwidth;
+               EffectiveCompressedBufferSize = dml_min((dml_float_t)p->CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate,
+                                                                                       (dml_float_t)p->MetaFIFOSizeInKEntries * 1024 * 64 / (AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate + 1 / AverageDCCCompressionRate)) +
+                                                                               dml_min(((dml_float_t)p->ROBBufferSizeInKByte * 1024 - p->CompbufReservedSpace64B * 64) * AverageDCCCompressionRate,
+                                                                                       ((dml_float_t)p->ZeroSizeBufferEntries - p->CompbufReservedSpaceZs) * 64 / (AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate));
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: min 1 = %f\n", __func__, p->CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate);
+               dml_print("DML::%s: min 2 = %f\n", __func__, p->MetaFIFOSizeInKEntries * 1024 * 64 / (AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate + 1 / AverageDCCCompressionRate));
+               dml_print("DML::%s: min 3 = %f\n", __func__, (p->ROBBufferSizeInKByte * 1024 - p->CompbufReservedSpace64B * 64) * AverageDCCCompressionRate);
+               dml_print("DML::%s: min 4 = %f\n", __func__, (p->ZeroSizeBufferEntries - p->CompbufReservedSpaceZs) * 64 / (AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate));
+#endif
+       } else {
+               EffectiveCompressedBufferSize = dml_min((dml_float_t)p->CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate,
+                                                                                               (dml_float_t)p->MetaFIFOSizeInKEntries * 1024 * 64 * AverageDCCCompressionRate) +
+                                                                                               ((dml_float_t)p->ROBBufferSizeInKByte * 1024 - p->CompbufReservedSpace64B * 64) * AverageDCCCompressionRate;
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: min 1 = %f\n", __func__, p->CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate);
+               dml_print("DML::%s: min 2 = %f\n", __func__, p->MetaFIFOSizeInKEntries * 1024 * 64 * AverageDCCCompressionRate);
+#endif
+       }
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: MetaFIFOSizeInKEntries = %u\n", __func__, p->MetaFIFOSizeInKEntries);
+       dml_print("DML::%s: AverageZeroSizeCompressionRate = %f\n", __func__, AverageZeroSizeCompressionRate);
+       dml_print("DML::%s: EffectiveCompressedBufferSize = %f\n", __func__, EffectiveCompressedBufferSize);
+#endif
+
+       *p->StutterPeriod = 0;
+
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+       if (p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) {
+               LinesInDETY = ((dml_float_t)p->DETBufferSizeY[k] + (p->UnboundedRequestEnabled == true ? EffectiveCompressedBufferSize : 0) * p->ReadBandwidthSurfaceLuma[k] / p->TotalDataReadBandwidth) / p->BytePerPixelDETY[k] / p->SwathWidthY[k];
+               LinesInDETYRoundedDownToSwath = dml_floor(LinesInDETY, p->SwathHeightY[k]);
+               DETBufferingTimeY = LinesInDETYRoundedDownToSwath * ((dml_float_t)p->HTotal[k] / p->PixelClock[k]) / p->VRatio[k];
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, DETBufferSizeY = %u\n", __func__, k, p->DETBufferSizeY[k]);
+               dml_print("DML::%s: k=%u, BytePerPixelDETY = %f\n", __func__, k, p->BytePerPixelDETY[k]);
+               dml_print("DML::%s: k=%u, SwathWidthY = %u\n", __func__, k, p->SwathWidthY[k]);
+               dml_print("DML::%s: k=%u, ReadBandwidthSurfaceLuma = %f\n", __func__, k, p->ReadBandwidthSurfaceLuma[k]);
+               dml_print("DML::%s: k=%u, TotalDataReadBandwidth = %f\n", __func__, k, p->TotalDataReadBandwidth);
+               dml_print("DML::%s: k=%u, LinesInDETY = %f\n", __func__, k, LinesInDETY);
+               dml_print("DML::%s: k=%u, LinesInDETYRoundedDownToSwath = %f\n", __func__, k, LinesInDETYRoundedDownToSwath);
+               dml_print("DML::%s: k=%u, HTotal = %u\n", __func__, k, p->HTotal[k]);
+               dml_print("DML::%s: k=%u, PixelClock = %f\n", __func__, k, p->PixelClock[k]);
+               dml_print("DML::%s: k=%u, VRatio = %f\n", __func__, k, p->VRatio[k]);
+               dml_print("DML::%s: k=%u, DETBufferingTimeY = %f\n", __func__, k, DETBufferingTimeY);
+               dml_print("DML::%s: k=%u,PixelClock = %f\n", __func__, k, p->PixelClock[k]);
+#endif
+
+               if (!FoundCriticalSurface || DETBufferingTimeY < *p->StutterPeriod) {
+                       dml_bool_t isInterlaceTiming = p->Interlace[k] && !p->ProgressiveToInterlaceUnitInOPP;
+
+                       FoundCriticalSurface = true;
+                       *p->StutterPeriod = DETBufferingTimeY;
+                       FrameTimeCriticalSurface = (isInterlaceTiming ? dml_floor((dml_float_t)p->VTotal[k]/2.0, 1.0) : p->VTotal[k]) * (dml_float_t)p->HTotal[k] / p->PixelClock[k];
+                       VActiveTimeCriticalSurface = (isInterlaceTiming ? dml_floor((dml_float_t)p->VActive[k]/2.0, 1.0) : p->VActive[k]) * (dml_float_t)p->HTotal[k] / p->PixelClock[k];
+                       BytePerPixelYCriticalSurface = p->BytePerPixelY[k];
+                       SwathWidthYCriticalSurface = p->SwathWidthY[k];
+                       SwathHeightYCriticalSurface = p->SwathHeightY[k];
+                       BlockWidth256BytesYCriticalSurface = p->BlockWidth256BytesY[k];
+                       LinesToFinishSwathTransferStutterCriticalSurface = p->SwathHeightY[k] - (LinesInDETY - LinesInDETYRoundedDownToSwath);
+                       DETBufferSizeYCriticalSurface = p->DETBufferSizeY[k];
+                       MinTTUVBlankCriticalSurface = p->MinTTUVBlank[k];
+                       SinglePlaneCriticalSurface = (p->ReadBandwidthSurfaceChroma[k] == 0);
+                       SinglePipeCriticalSurface = (p->DPPPerSurface[k] == 1);
+
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%u, FoundCriticalSurface = %u\n", __func__, k, FoundCriticalSurface);
+                       dml_print("DML::%s: k=%u, StutterPeriod = %f\n", __func__, k, *p->StutterPeriod);
+                       dml_print("DML::%s: k=%u, MinTTUVBlankCriticalSurface = %f\n", __func__, k, MinTTUVBlankCriticalSurface);
+                       dml_print("DML::%s: k=%u, FrameTimeCriticalSurface = %f\n", __func__, k, FrameTimeCriticalSurface);
+                       dml_print("DML::%s: k=%u, VActiveTimeCriticalSurface = %f\n", __func__, k, VActiveTimeCriticalSurface);
+                       dml_print("DML::%s: k=%u, BytePerPixelYCriticalSurface = %u\n", __func__, k, BytePerPixelYCriticalSurface);
+                       dml_print("DML::%s: k=%u, SwathWidthYCriticalSurface = %f\n", __func__, k, SwathWidthYCriticalSurface);
+                       dml_print("DML::%s: k=%u, SwathHeightYCriticalSurface = %f\n", __func__, k, SwathHeightYCriticalSurface);
+                       dml_print("DML::%s: k=%u, BlockWidth256BytesYCriticalSurface = %u\n", __func__, k, BlockWidth256BytesYCriticalSurface);
+                       dml_print("DML::%s: k=%u, SinglePlaneCriticalSurface = %u\n", __func__, k, SinglePlaneCriticalSurface);
+                       dml_print("DML::%s: k=%u, SinglePipeCriticalSurface = %u\n", __func__, k, SinglePipeCriticalSurface);
+                       dml_print("DML::%s: k=%u, LinesToFinishSwathTransferStutterCriticalSurface = %f\n", __func__, k, LinesToFinishSwathTransferStutterCriticalSurface);
+#endif
+               }
+       }
+       }
+
+       PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = dml_min(*p->StutterPeriod * p->TotalDataReadBandwidth, EffectiveCompressedBufferSize);
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: ROBBufferSizeInKByte = %u\n", __func__, p->ROBBufferSizeInKByte);
+       dml_print("DML::%s: AverageDCCCompressionRate = %f\n", __func__, AverageDCCCompressionRate);
+       dml_print("DML::%s: StutterPeriod * TotalDataReadBandwidth = %f\n", __func__, *p->StutterPeriod * p->TotalDataReadBandwidth);
+       dml_print("DML::%s: ROBBufferSizeInKByte * 1024 * AverageDCCCompressionRate + EffectiveCompressedBufferSize = %f\n", __func__, p->ROBBufferSizeInKByte * 1024 * AverageDCCCompressionRate + EffectiveCompressedBufferSize);
+       dml_print("DML::%s: EffectiveCompressedBufferSize = %f\n", __func__, EffectiveCompressedBufferSize);
+       dml_print("DML::%s: PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = %f\n", __func__, PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer);
+       dml_print("DML::%s: ReturnBW = %f\n", __func__, p->ReturnBW);
+       dml_print("DML::%s: TotalDataReadBandwidth = %f\n", __func__, p->TotalDataReadBandwidth);
+       dml_print("DML::%s: TotalRowReadBandwidth = %f\n", __func__, TotalRowReadBandwidth);
+       dml_print("DML::%s: DCFCLK = %f\n", __func__, p->DCFCLK);
+#endif
+
+       StutterBurstTime = PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer / AverageDCCCompressionRate / p->ReturnBW + (*p->StutterPeriod * p->TotalDataReadBandwidth - PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer) / (p->DCFCLK * 64) + *p->StutterPeriod * TotalRowReadBandwidth / p->ReturnBW;
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: Part 1 = %f\n", __func__, PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer / AverageDCCCompressionRate / p->ReturnBW);
+       dml_print("DML::%s: StutterPeriod * TotalDataReadBandwidth = %f\n", __func__, (*p->StutterPeriod * p->TotalDataReadBandwidth));
+       dml_print("DML::%s: Part 2 = %f\n", __func__, (*p->StutterPeriod * p->TotalDataReadBandwidth - PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer) / (p->DCFCLK * 64));
+       dml_print("DML::%s: Part 3 = %f\n", __func__, *p->StutterPeriod * TotalRowReadBandwidth / p->ReturnBW);
+       dml_print("DML::%s: StutterBurstTime = %f\n", __func__, StutterBurstTime);
+#endif
+       StutterBurstTime = dml_max(StutterBurstTime, LinesToFinishSwathTransferStutterCriticalSurface * BytePerPixelYCriticalSurface * SwathWidthYCriticalSurface / p->ReturnBW);
+
+       dml_print("DML::%s: Time to finish residue swath=%f\n", __func__, LinesToFinishSwathTransferStutterCriticalSurface * BytePerPixelYCriticalSurface * SwathWidthYCriticalSurface / p->ReturnBW);
+
+       TotalActiveWriteback = 0;
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               if (p->WritebackEnable[k]) {
+                       TotalActiveWriteback = TotalActiveWriteback + 1;
+               }
+       }
+
+       if (TotalActiveWriteback == 0) {
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: SRExitTime = %f\n", __func__, p->SRExitTime);
+               dml_print("DML::%s: SRExitZ8Time = %f\n", __func__, p->SRExitZ8Time);
+               dml_print("DML::%s: StutterBurstTime = %f (final)\n", __func__, StutterBurstTime);
+               dml_print("DML::%s: StutterPeriod = %f\n", __func__, *p->StutterPeriod);
+#endif
+               *p->StutterEfficiencyNotIncludingVBlank = dml_max(0., 1 - (p->SRExitTime + StutterBurstTime) / *p->StutterPeriod) * 100;
+               *p->Z8StutterEfficiencyNotIncludingVBlank = dml_max(0., 1 - (p->SRExitZ8Time + StutterBurstTime) / *p->StutterPeriod) * 100;
+               *p->NumberOfStutterBurstsPerFrame = (*p->StutterEfficiencyNotIncludingVBlank > 0 ? (dml_uint_t)(dml_ceil(VActiveTimeCriticalSurface / *p->StutterPeriod, 1)) : 0);
+               *p->Z8NumberOfStutterBurstsPerFrame = (*p->Z8StutterEfficiencyNotIncludingVBlank > 0 ? (dml_uint_t)(dml_ceil(VActiveTimeCriticalSurface / *p->StutterPeriod, 1)) : 0);
+       } else {
+               *p->StutterEfficiencyNotIncludingVBlank = 0.;
+               *p->Z8StutterEfficiencyNotIncludingVBlank = 0.;
+               *p->NumberOfStutterBurstsPerFrame = 0;
+               *p->Z8NumberOfStutterBurstsPerFrame = 0;
+       }
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: VActiveTimeCriticalSurface = %f\n", __func__, VActiveTimeCriticalSurface);
+       dml_print("DML::%s: StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->StutterEfficiencyNotIncludingVBlank);
+       dml_print("DML::%s: Z8StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->Z8StutterEfficiencyNotIncludingVBlank);
+       dml_print("DML::%s: NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->NumberOfStutterBurstsPerFrame);
+       dml_print("DML::%s: Z8NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->Z8NumberOfStutterBurstsPerFrame);
+#endif
+
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               if (p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) {
+                       if (p->BlendingAndTiming[k] == k) {
+                               if (TotalNumberOfActiveOTG == 0) {
+                                       SinglePixelClock = p->PixelClock[k];
+                                       SingleHTotal = p->HTotal[k];
+                                       SingleVTotal = p->VTotal[k];
+                               } else if (SinglePixelClock != p->PixelClock[k] || SingleHTotal != p->HTotal[k] || SingleVTotal != p->VTotal[k]) {
+                                       SameTiming = false;
+                               }
+                               TotalNumberOfActiveOTG = TotalNumberOfActiveOTG + 1;
+                       }
+               }
+       }
+
+       if (*p->StutterEfficiencyNotIncludingVBlank > 0) {
+               LastStutterPeriod = VActiveTimeCriticalSurface - (*p->NumberOfStutterBurstsPerFrame - 1) * *p->StutterPeriod;
+
+               if ((p->SynchronizeTimingsFinal || TotalNumberOfActiveOTG == 1) && SameTiming &&
+                       LastStutterPeriod + MinTTUVBlankCriticalSurface > p->StutterEnterPlusExitWatermark) {
+                       *p->StutterEfficiency = (1 - (*p->NumberOfStutterBurstsPerFrame * p->SRExitTime + StutterBurstTime * VActiveTimeCriticalSurface / *p->StutterPeriod) / FrameTimeCriticalSurface) * 100;
+               } else {
+                       *p->StutterEfficiency = *p->StutterEfficiencyNotIncludingVBlank;
+               }
+       } else {
+               *p->StutterEfficiency = 0;
+       }
+
+       if (*p->Z8StutterEfficiencyNotIncludingVBlank > 0) {
+               LastZ8StutterPeriod = VActiveTimeCriticalSurface - (*p->NumberOfStutterBurstsPerFrame - 1) * *p->StutterPeriod;
+               if ((p->SynchronizeTimingsFinal || TotalNumberOfActiveOTG == 1) && SameTiming && LastZ8StutterPeriod + MinTTUVBlankCriticalSurface > p->Z8StutterEnterPlusExitWatermark) {
+                       *p->Z8StutterEfficiency = (1 - (*p->NumberOfStutterBurstsPerFrame * p->SRExitZ8Time + StutterBurstTime * VActiveTimeCriticalSurface / *p->StutterPeriod) / FrameTimeCriticalSurface) * 100;
+               } else {
+                       *p->Z8StutterEfficiency = *p->Z8StutterEfficiencyNotIncludingVBlank;
+               }
+       } else {
+               *p->Z8StutterEfficiency = 0.;
+       }
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: LastZ8StutterPeriod = %f\n", __func__, LastZ8StutterPeriod);
+               dml_print("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n", __func__, p->Z8StutterEnterPlusExitWatermark);
+               dml_print("DML::%s: StutterBurstTime = %f\n", __func__, StutterBurstTime);
+               dml_print("DML::%s: StutterPeriod = %f\n", __func__, *p->StutterPeriod);
+               dml_print("DML::%s: StutterEfficiency = %f\n", __func__, *p->StutterEfficiency);
+               dml_print("DML::%s: Z8StutterEfficiency = %f\n", __func__, *p->Z8StutterEfficiency);
+               dml_print("DML::%s: StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->StutterEfficiencyNotIncludingVBlank);
+               dml_print("DML::%s: Z8NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->Z8NumberOfStutterBurstsPerFrame);
+#endif
+
+               SwathSizeCriticalSurface = (dml_uint_t)(BytePerPixelYCriticalSurface * SwathHeightYCriticalSurface * dml_ceil(SwathWidthYCriticalSurface, BlockWidth256BytesYCriticalSurface));
+               LastChunkOfSwathSize = SwathSizeCriticalSurface % (p->PixelChunkSizeInKByte * 1024);
+               MissingPartOfLastSwathOfDETSize = (dml_uint_t)(dml_ceil(DETBufferSizeYCriticalSurface, SwathSizeCriticalSurface) - DETBufferSizeYCriticalSurface);
+
+               *p->DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE = !(!p->UnboundedRequestEnabled && (p->NumberOfActiveSurfaces == 1) && SinglePlaneCriticalSurface && SinglePipeCriticalSurface && (LastChunkOfSwathSize > 0) &&
+                       (LastChunkOfSwathSize <= 4096) && (MissingPartOfLastSwathOfDETSize > 0) && (MissingPartOfLastSwathOfDETSize <= LastChunkOfSwathSize));
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: SwathSizeCriticalSurface = %u\n", __func__, SwathSizeCriticalSurface);
+       dml_print("DML::%s: DETBufferSizeYCriticalSurface = %u\n", __func__, DETBufferSizeYCriticalSurface);
+       dml_print("DML::%s: PixelChunkSizeInKByte = %u\n", __func__, p->PixelChunkSizeInKByte);
+       dml_print("DML::%s: LastChunkOfSwathSize = %u\n", __func__, LastChunkOfSwathSize);
+       dml_print("DML::%s: MissingPartOfLastSwathOfDETSize = %u\n", __func__, MissingPartOfLastSwathOfDETSize);
+       dml_print("DML::%s: DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE = %u\n", __func__, *p->DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE);
+#endif
+} // CalculateStutterEfficiency
+
+/// \CalculateSwathAndDETConfiguration
+/// @brief Calculates swath width and different return buffers sizing (DET, CDB, etc.)
+static void CalculateSwathAndDETConfiguration(struct display_mode_lib_scratch_st *scratch,
+       struct CalculateSwathAndDETConfiguration_params_st *p)
+{
+       dml_uint_t MaximumSwathHeightY[__DML_NUM_PLANES__];
+       dml_uint_t MaximumSwathHeightC[__DML_NUM_PLANES__];
+       dml_uint_t RoundedUpMaxSwathSizeBytesY[__DML_NUM_PLANES__];
+       dml_uint_t RoundedUpMaxSwathSizeBytesC[__DML_NUM_PLANES__];
+       dml_uint_t RoundedUpSwathSizeBytesY[__DML_NUM_PLANES__];
+       dml_uint_t RoundedUpSwathSizeBytesC[__DML_NUM_PLANES__];
+       dml_uint_t SwathWidthSingleDPP[__DML_NUM_PLANES__];
+       dml_uint_t SwathWidthSingleDPPChroma[__DML_NUM_PLANES__];
+
+       dml_uint_t TotalActiveDPP = 0;
+       dml_bool_t NoChromaOrLinearSurfaces = true;
+       dml_uint_t SurfaceDoingUnboundedRequest = 0;
+
+       dml_uint_t DETBufferSizeInKByteForSwathCalculation;
+
+       const long TTUFIFODEPTH = 8;
+       const long MAXIMUMCOMPRESSION = 4;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: ForceSingleDPP = %u\n", __func__, p->ForceSingleDPP);
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               dml_print("DML::%s: DPPPerSurface[%u] = %u\n", __func__, k, p->DPPPerSurface[k]);
+       }
+#endif
+       CalculateSwathWidth(p->ForceSingleDPP,
+               p->NumberOfActiveSurfaces,
+               p->SourcePixelFormat,
+               p->SourceScan,
+               p->ViewportStationary,
+               p->ViewportWidth,
+               p->ViewportHeight,
+               p->ViewportXStart,
+               p->ViewportYStart,
+               p->ViewportXStartC,
+               p->ViewportYStartC,
+               p->SurfaceWidthY,
+               p->SurfaceWidthC,
+               p->SurfaceHeightY,
+               p->SurfaceHeightC,
+               p->ODMMode,
+               p->BytePerPixY,
+               p->BytePerPixC,
+               p->Read256BytesBlockHeightY,
+               p->Read256BytesBlockHeightC,
+               p->Read256BytesBlockWidthY,
+               p->Read256BytesBlockWidthC,
+               p->BlendingAndTiming,
+               p->HActive,
+               p->HRatio,
+               p->DPPPerSurface,
+
+               // Output
+               SwathWidthSingleDPP,
+               SwathWidthSingleDPPChroma,
+               p->SwathWidth,
+               p->SwathWidthChroma,
+               MaximumSwathHeightY,
+               MaximumSwathHeightC,
+               p->swath_width_luma_ub,
+               p->swath_width_chroma_ub);
+
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               RoundedUpMaxSwathSizeBytesY[k] = (dml_uint_t)(p->swath_width_luma_ub[k] * p->BytePerPixDETY[k] * MaximumSwathHeightY[k]);
+               RoundedUpMaxSwathSizeBytesC[k] = (dml_uint_t)(p->swath_width_chroma_ub[k] * p->BytePerPixDETC[k] * MaximumSwathHeightC[k]);
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u DPPPerSurface = %u\n", __func__, k, p->DPPPerSurface[k]);
+               dml_print("DML::%s: k=%u swath_width_luma_ub = %u\n", __func__, k, p->swath_width_luma_ub[k]);
+               dml_print("DML::%s: k=%u BytePerPixDETY = %f\n", __func__, k, p->BytePerPixDETY[k]);
+               dml_print("DML::%s: k=%u MaximumSwathHeightY = %u\n", __func__, k, MaximumSwathHeightY[k]);
+               dml_print("DML::%s: k=%u RoundedUpMaxSwathSizeBytesY = %u\n", __func__, k, RoundedUpMaxSwathSizeBytesY[k]);
+               dml_print("DML::%s: k=%u swath_width_chroma_ub = %u\n", __func__, k, p->swath_width_chroma_ub[k]);
+               dml_print("DML::%s: k=%u BytePerPixDETC = %f\n", __func__, k, p->BytePerPixDETC[k]);
+               dml_print("DML::%s: k=%u MaximumSwathHeightC = %u\n", __func__, k, MaximumSwathHeightC[k]);
+               dml_print("DML::%s: k=%u RoundedUpMaxSwathSizeBytesC = %u\n", __func__, k, RoundedUpMaxSwathSizeBytesC[k]);
+#endif
+               if (p->SourcePixelFormat[k] == dml_420_10) {
+                       RoundedUpMaxSwathSizeBytesY[k] = (dml_uint_t)(dml_ceil((dml_float_t) RoundedUpMaxSwathSizeBytesY[k], 256));
+                       RoundedUpMaxSwathSizeBytesC[k] = (dml_uint_t)(dml_ceil((dml_float_t) RoundedUpMaxSwathSizeBytesC[k], 256));
+               }
+       }
+
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               TotalActiveDPP = TotalActiveDPP + (p->ForceSingleDPP ? 1 : p->DPPPerSurface[k]);
+               if (p->DPPPerSurface[k] > 0)
+                       SurfaceDoingUnboundedRequest = k;
+               if (p->SourcePixelFormat[k] == dml_420_8 || p->SourcePixelFormat[k] == dml_420_10 ||
+                               p->SourcePixelFormat[k] == dml_420_12 || p->SourcePixelFormat[k] == dml_rgbe_alpha
+                               || p->SurfaceTiling[k] == dml_sw_linear) {
+                       NoChromaOrLinearSurfaces = false;
+               }
+       }
+
+       *p->UnboundedRequestEnabled = UnboundedRequest(p->UseUnboundedRequestingFinal, TotalActiveDPP,
+                       NoChromaOrLinearSurfaces, p->Output[0]);
+
+       CalculateDETBufferSize(p->DETSizeOverride,
+               p->UseMALLForPStateChange,
+               p->ForceSingleDPP,
+               p->NumberOfActiveSurfaces,
+               *p->UnboundedRequestEnabled,
+               p->nomDETInKByte,
+               p->MaxTotalDETInKByte,
+               p->ConfigReturnBufferSizeInKByte,
+               p->MinCompressedBufferSizeInKByte,
+               p->ConfigReturnBufferSegmentSizeInkByte,
+               p->CompressedBufferSegmentSizeInkByteFinal,
+               p->SourcePixelFormat,
+               p->ReadBandwidthLuma,
+               p->ReadBandwidthChroma,
+               RoundedUpMaxSwathSizeBytesY,
+               RoundedUpMaxSwathSizeBytesC,
+               p->DPPPerSurface,
+
+               // Output
+               p->DETBufferSizeInKByte, // per hubp pipe
+               p->CompressedBufferSizeInkByte);
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: TotalActiveDPP = %u\n", __func__, TotalActiveDPP);
+               dml_print("DML::%s: nomDETInKByte = %u\n", __func__, p->nomDETInKByte);
+               dml_print("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, p->ConfigReturnBufferSizeInKByte);
+               dml_print("DML::%s: UseUnboundedRequestingFinal = %u\n", __func__, p->UseUnboundedRequestingFinal);
+               dml_print("DML::%s: UnboundedRequestEnabled = %u\n", __func__, *p->UnboundedRequestEnabled);
+               dml_print("DML::%s: CompressedBufferSizeInkByte = %u\n", __func__, *p->CompressedBufferSizeInkByte);
+#endif
+
+       *p->ViewportSizeSupport = true;
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+
+               DETBufferSizeInKByteForSwathCalculation =  (p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe ? 1024 : p->DETBufferSizeInKByte[k]);
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: k=%u DETBufferSizeInKByteForSwathCalculation = %u\n", __func__, k, DETBufferSizeInKByteForSwathCalculation);
+#endif
+
+               if (RoundedUpMaxSwathSizeBytesY[k] + RoundedUpMaxSwathSizeBytesC[k] <= DETBufferSizeInKByteForSwathCalculation * 1024 / 2) {
+                       p->SwathHeightY[k] = MaximumSwathHeightY[k];
+                       p->SwathHeightC[k] = MaximumSwathHeightC[k];
+                       RoundedUpSwathSizeBytesY[k] = RoundedUpMaxSwathSizeBytesY[k];
+                       RoundedUpSwathSizeBytesC[k] = RoundedUpMaxSwathSizeBytesC[k];
+               } else if (RoundedUpMaxSwathSizeBytesY[k] >= 1.5 * RoundedUpMaxSwathSizeBytesC[k] && RoundedUpMaxSwathSizeBytesY[k] / 2 + RoundedUpMaxSwathSizeBytesC[k] <= DETBufferSizeInKByteForSwathCalculation * 1024 / 2) {
+                       p->SwathHeightY[k] = MaximumSwathHeightY[k] / 2;
+                       p->SwathHeightC[k] = MaximumSwathHeightC[k];
+                       RoundedUpSwathSizeBytesY[k] = RoundedUpMaxSwathSizeBytesY[k] / 2;
+                       RoundedUpSwathSizeBytesC[k] = RoundedUpMaxSwathSizeBytesC[k];
+               } else if (RoundedUpMaxSwathSizeBytesY[k] < 1.5 * RoundedUpMaxSwathSizeBytesC[k] && RoundedUpMaxSwathSizeBytesY[k] + RoundedUpMaxSwathSizeBytesC[k] / 2 <= DETBufferSizeInKByteForSwathCalculation * 1024 / 2) {
+                       p->SwathHeightY[k] = MaximumSwathHeightY[k];
+                       p->SwathHeightC[k] = MaximumSwathHeightC[k] / 2;
+                       RoundedUpSwathSizeBytesY[k] = RoundedUpMaxSwathSizeBytesY[k];
+                       RoundedUpSwathSizeBytesC[k] = RoundedUpMaxSwathSizeBytesC[k] / 2;
+               } else {
+                       p->SwathHeightY[k] = MaximumSwathHeightY[k] / 2;
+                       p->SwathHeightC[k] = MaximumSwathHeightC[k] / 2;
+                       RoundedUpSwathSizeBytesY[k] = RoundedUpMaxSwathSizeBytesY[k] / 2;
+                       RoundedUpSwathSizeBytesC[k] = RoundedUpMaxSwathSizeBytesC[k] / 2;
+               }
+
+               if ((RoundedUpMaxSwathSizeBytesY[k] / 2 + RoundedUpMaxSwathSizeBytesC[k] / 2 > DETBufferSizeInKByteForSwathCalculation * 1024 / 2) ||
+                       p->SwathWidth[k] > p->MaximumSwathWidthLuma[k] || (p->SwathHeightC[k] > 0 && p->SwathWidthChroma[k] > p->MaximumSwathWidthChroma[k])) {
+                       *p->ViewportSizeSupport = false;
+                       p->ViewportSizeSupportPerSurface[k] = false;
+               } else {
+                       p->ViewportSizeSupportPerSurface[k] = true;
+               }
+
+       if (p->SwathHeightC[k] == 0) {
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%u All DET for plane0\n", __func__, k);
+#endif
+                       p->DETBufferSizeY[k] = p->DETBufferSizeInKByte[k] * 1024;
+                       p->DETBufferSizeC[k] = 0;
+               } else if (RoundedUpSwathSizeBytesY[k] <= 1.5 * RoundedUpSwathSizeBytesC[k]) {
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%u Half DET for plane0, half for plane1\n", __func__, k);
+#endif
+                       p->DETBufferSizeY[k] = p->DETBufferSizeInKByte[k] * 1024 / 2;
+                       p->DETBufferSizeC[k] = p->DETBufferSizeInKByte[k] * 1024 / 2;
+               } else {
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%u 2/3 DET for plane0, 1/3 for plane1\n", __func__, k);
+#endif
+                       p->DETBufferSizeY[k] = (dml_uint_t)(dml_floor(p->DETBufferSizeInKByte[k] * 1024 * 2 / 3, 1024));
+                       p->DETBufferSizeC[k] = p->DETBufferSizeInKByte[k] * 1024 - p->DETBufferSizeY[k];
+               }
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u SwathHeightY = %u\n", __func__, k, p->SwathHeightY[k]);
+               dml_print("DML::%s: k=%u SwathHeightC = %u\n", __func__, k, p->SwathHeightC[k]);
+               dml_print("DML::%s: k=%u RoundedUpMaxSwathSizeBytesY = %u\n", __func__, k, RoundedUpMaxSwathSizeBytesY[k]);
+               dml_print("DML::%s: k=%u RoundedUpMaxSwathSizeBytesC = %u\n", __func__, k, RoundedUpMaxSwathSizeBytesC[k]);
+               dml_print("DML::%s: k=%u RoundedUpSwathSizeBytesY = %u\n", __func__, k, RoundedUpSwathSizeBytesY[k]);
+               dml_print("DML::%s: k=%u RoundedUpSwathSizeBytesC = %u\n", __func__, k, RoundedUpSwathSizeBytesC[k]);
+               dml_print("DML::%s: k=%u DETBufferSizeInKByte = %u\n", __func__, k, p->DETBufferSizeInKByte[k]);
+               dml_print("DML::%s: k=%u DETBufferSizeY = %u\n", __func__, k, p->DETBufferSizeY[k]);
+               dml_print("DML::%s: k=%u DETBufferSizeC = %u\n", __func__, k, p->DETBufferSizeC[k]);
+               dml_print("DML::%s: k=%u ViewportSizeSupportPerSurface = %u\n", __func__, k, p->ViewportSizeSupportPerSurface[k]);
+#endif
+
+       }
+
+       *p->compbuf_reserved_space_64b = 2 * p->PixelChunkSizeInKByte * 1024 / 64;
+       if (p->UnboundedRequestEnabled) {
+               *p->compbuf_reserved_space_64b = dml_max(*p->compbuf_reserved_space_64b,
+                               (dml_float_t)(p->ROBBufferSizeInKByte * 1024/64)
+                               - (dml_float_t)(RoundedUpSwathSizeBytesY[SurfaceDoingUnboundedRequest] * TTUFIFODEPTH / MAXIMUMCOMPRESSION/64));
+       }
+       *p->compbuf_reserved_space_zs = 2 * p->PixelChunkSizeInKByte * 1024 / 256;
+} // CalculateSwathAndDETConfiguration
+
+static void CalculateSwathWidth(
+               dml_bool_t ForceSingleDPP,
+               dml_uint_t NumberOfActiveSurfaces,
+               enum dml_source_format_class SourcePixelFormat[],
+               enum dml_rotation_angle SourceScan[],
+               dml_bool_t ViewportStationary[],
+               dml_uint_t ViewportWidth[],
+               dml_uint_t ViewportHeight[],
+               dml_uint_t ViewportXStart[],
+               dml_uint_t ViewportYStart[],
+               dml_uint_t ViewportXStartC[],
+               dml_uint_t ViewportYStartC[],
+               dml_uint_t SurfaceWidthY[],
+               dml_uint_t SurfaceWidthC[],
+               dml_uint_t SurfaceHeightY[],
+               dml_uint_t SurfaceHeightC[],
+               enum dml_odm_mode ODMMode[],
+               dml_uint_t BytePerPixY[],
+               dml_uint_t BytePerPixC[],
+               dml_uint_t Read256BytesBlockHeightY[],
+               dml_uint_t Read256BytesBlockHeightC[],
+               dml_uint_t Read256BytesBlockWidthY[],
+               dml_uint_t Read256BytesBlockWidthC[],
+               dml_uint_t BlendingAndTiming[],
+               dml_uint_t HActive[],
+               dml_float_t HRatio[],
+               dml_uint_t DPPPerSurface[],
+
+               // Output
+               dml_uint_t SwathWidthSingleDPPY[],
+               dml_uint_t SwathWidthSingleDPPC[],
+               dml_uint_t SwathWidthY[], // per-pipe
+               dml_uint_t SwathWidthC[], // per-pipe
+               dml_uint_t MaximumSwathHeightY[],
+               dml_uint_t MaximumSwathHeightC[],
+               dml_uint_t swath_width_luma_ub[], // per-pipe
+               dml_uint_t swath_width_chroma_ub[]) // per-pipe
+{
+       enum dml_odm_mode   MainSurfaceODMMode;
+       dml_uint_t surface_width_ub_l;
+       dml_uint_t surface_height_ub_l;
+       dml_uint_t surface_width_ub_c = 0;
+       dml_uint_t surface_height_ub_c = 0;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: ForceSingleDPP = %u\n", __func__, ForceSingleDPP);
+       dml_print("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces);
+#endif
+
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               if (!dml_is_vertical_rotation(SourceScan[k])) {
+                       SwathWidthSingleDPPY[k] = ViewportWidth[k];
+               } else {
+                       SwathWidthSingleDPPY[k] = ViewportHeight[k];
+               }
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u ViewportWidth=%u\n", __func__, k, ViewportWidth[k]);
+               dml_print("DML::%s: k=%u ViewportHeight=%u\n", __func__, k, ViewportHeight[k]);
+               dml_print("DML::%s: k=%u DPPPerSurface=%u\n", __func__, k, DPPPerSurface[k]);
+#endif
+
+               MainSurfaceODMMode = ODMMode[k];
+               for (dml_uint_t j = 0; j < NumberOfActiveSurfaces; ++j) {
+                       if (BlendingAndTiming[k] == j) {
+                               MainSurfaceODMMode = ODMMode[j];
+                       }
+               }
+
+               if (ForceSingleDPP) {
+                       SwathWidthY[k] = SwathWidthSingleDPPY[k];
+               } else {
+                       if (MainSurfaceODMMode == dml_odm_mode_combine_4to1) {
+                               SwathWidthY[k] = (dml_uint_t)(dml_min(SwathWidthSingleDPPY[k], dml_round(HActive[k] / 4.0 * HRatio[k], true)));
+                       } else if (MainSurfaceODMMode == dml_odm_mode_combine_2to1) {
+                               SwathWidthY[k] = (dml_uint_t)(dml_min(SwathWidthSingleDPPY[k], dml_round(HActive[k] / 2.0 * HRatio[k], true)));
+                       } else if (DPPPerSurface[k] == 2) {
+                               SwathWidthY[k] = SwathWidthSingleDPPY[k] / 2;
+                       } else {
+                               SwathWidthY[k] = SwathWidthSingleDPPY[k];
+                       }
+               }
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u HActive=%u\n", __func__, k, HActive[k]);
+               dml_print("DML::%s: k=%u HRatio=%f\n", __func__, k, HRatio[k]);
+               dml_print("DML::%s: k=%u MainSurfaceODMMode=%u\n", __func__, k, MainSurfaceODMMode);
+               dml_print("DML::%s: k=%u SwathWidthSingleDPPY=%u\n", __func__, k, SwathWidthSingleDPPY[k]);
+               dml_print("DML::%s: k=%u SwathWidthY=%u\n", __func__, k, SwathWidthY[k]);
+#endif
+
+               if (SourcePixelFormat[k] == dml_420_8 || SourcePixelFormat[k] == dml_420_10 || SourcePixelFormat[k] == dml_420_12) {
+                       SwathWidthC[k] = SwathWidthY[k] / 2;
+                       SwathWidthSingleDPPC[k] = SwathWidthSingleDPPY[k] / 2;
+               } else {
+                       SwathWidthC[k] = SwathWidthY[k];
+                       SwathWidthSingleDPPC[k] = SwathWidthSingleDPPY[k];
+               }
+
+               if (ForceSingleDPP == true) {
+                       SwathWidthY[k] = SwathWidthSingleDPPY[k];
+                       SwathWidthC[k] = SwathWidthSingleDPPC[k];
+               }
+
+               surface_width_ub_l = (dml_uint_t)dml_ceil(SurfaceWidthY[k], Read256BytesBlockWidthY[k]);
+               surface_height_ub_l = (dml_uint_t)dml_ceil(SurfaceHeightY[k], Read256BytesBlockHeightY[k]);
+
+               if (!dml_is_vertical_rotation(SourceScan[k])) {
+                       MaximumSwathHeightY[k] = Read256BytesBlockHeightY[k];
+                       MaximumSwathHeightC[k] = Read256BytesBlockHeightC[k];
+                       if (ViewportStationary[k] && DPPPerSurface[k] == 1) {
+                               swath_width_luma_ub[k] = (dml_uint_t)(dml_min(surface_width_ub_l, dml_floor(ViewportXStart[k] + SwathWidthY[k] + Read256BytesBlockWidthY[k] - 1, Read256BytesBlockWidthY[k]) - dml_floor(ViewportXStart[k], Read256BytesBlockWidthY[k])));
+                       } else {
+                               swath_width_luma_ub[k] = (dml_uint_t)(dml_min(surface_width_ub_l, dml_ceil(SwathWidthY[k] - 1, Read256BytesBlockWidthY[k]) + Read256BytesBlockWidthY[k]));
+                       }
+                       if (BytePerPixC[k] > 0) {
+                               surface_width_ub_c = (dml_uint_t)dml_ceil(SurfaceWidthC[k], Read256BytesBlockWidthC[k]);
+                               if (ViewportStationary[k] && DPPPerSurface[k] == 1) {
+                                       swath_width_chroma_ub[k] = (dml_uint_t)(dml_min(surface_width_ub_c, dml_floor(ViewportXStartC[k] + SwathWidthC[k] + Read256BytesBlockWidthC[k] - 1, Read256BytesBlockWidthC[k]) - dml_floor(ViewportXStartC[k], Read256BytesBlockWidthC[k])));
+                               } else {
+                                       swath_width_chroma_ub[k] = (dml_uint_t)(dml_min(surface_width_ub_c, dml_ceil(SwathWidthC[k] - 1, Read256BytesBlockWidthC[k]) + Read256BytesBlockWidthC[k]));
+                               }
+                       } else {
+                               swath_width_chroma_ub[k] = 0;
+                       }
+               } else {
+                       MaximumSwathHeightY[k] = Read256BytesBlockWidthY[k];
+                       MaximumSwathHeightC[k] = Read256BytesBlockWidthC[k];
+
+                       if (ViewportStationary[k] && DPPPerSurface[k] == 1) {
+                               swath_width_luma_ub[k] = (dml_uint_t)(dml_min(surface_height_ub_l, dml_floor(ViewportYStart[k] + SwathWidthY[k] + Read256BytesBlockHeightY[k] - 1, Read256BytesBlockHeightY[k]) - dml_floor(ViewportYStart[k], Read256BytesBlockHeightY[k])));
+                       } else {
+                               swath_width_luma_ub[k] = (dml_uint_t)(dml_min(surface_height_ub_l, dml_ceil(SwathWidthY[k] - 1, Read256BytesBlockHeightY[k]) + Read256BytesBlockHeightY[k]));
+                       }
+                       if (BytePerPixC[k] > 0) {
+                               surface_height_ub_c = (dml_uint_t)dml_ceil(SurfaceHeightC[k], Read256BytesBlockHeightC[k]);
+                               if (ViewportStationary[k] && DPPPerSurface[k] == 1) {
+                                       swath_width_chroma_ub[k] = (dml_uint_t)(dml_min(surface_height_ub_c, dml_floor(ViewportYStartC[k] + SwathWidthC[k] + Read256BytesBlockHeightC[k] - 1, Read256BytesBlockHeightC[k]) - dml_floor(ViewportYStartC[k], Read256BytesBlockHeightC[k])));
+                               } else {
+                                       swath_width_chroma_ub[k] = (dml_uint_t)(dml_min(surface_height_ub_c, dml_ceil(SwathWidthC[k] - 1, Read256BytesBlockHeightC[k]) + Read256BytesBlockHeightC[k]));
+                               }
+                       } else {
+                               swath_width_chroma_ub[k] = 0;
+                       }
+               }
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u surface_width_ub_l=%u\n", __func__, k, surface_width_ub_l);
+               dml_print("DML::%s: k=%u surface_height_ub_l=%u\n", __func__, k, surface_height_ub_l);
+               dml_print("DML::%s: k=%u surface_width_ub_c=%u\n", __func__, k, surface_width_ub_c);
+               dml_print("DML::%s: k=%u surface_height_ub_c=%u\n", __func__, k, surface_height_ub_c);
+               dml_print("DML::%s: k=%u Read256BytesBlockWidthY=%u\n", __func__, k, Read256BytesBlockWidthY[k]);
+               dml_print("DML::%s: k=%u Read256BytesBlockHeightY=%u\n", __func__, k, Read256BytesBlockHeightY[k]);
+               dml_print("DML::%s: k=%u Read256BytesBlockWidthC=%u\n", __func__, k, Read256BytesBlockWidthC[k]);
+               dml_print("DML::%s: k=%u Read256BytesBlockHeightC=%u\n", __func__, k, Read256BytesBlockHeightC[k]);
+               dml_print("DML::%s: k=%u ViewportStationary=%u\n", __func__, k, ViewportStationary[k]);
+               dml_print("DML::%s: k=%u DPPPerSurface=%u\n", __func__, k, DPPPerSurface[k]);
+               dml_print("DML::%s: k=%u swath_width_luma_ub=%u\n", __func__, k, swath_width_luma_ub[k]);
+               dml_print("DML::%s: k=%u swath_width_chroma_ub=%u\n", __func__, k, swath_width_chroma_ub[k]);
+               dml_print("DML::%s: k=%u MaximumSwathHeightY=%u\n", __func__, k, MaximumSwathHeightY[k]);
+               dml_print("DML::%s: k=%u MaximumSwathHeightC=%u\n", __func__, k, MaximumSwathHeightC[k]);
+#endif
+
+       }
+} // CalculateSwathWidth
+
+static  dml_float_t CalculateExtraLatency(
+               dml_uint_t RoundTripPingLatencyCycles,
+               dml_uint_t ReorderingBytes,
+               dml_float_t DCFCLK,
+               dml_uint_t TotalNumberOfActiveDPP,
+               dml_uint_t PixelChunkSizeInKByte,
+               dml_uint_t TotalNumberOfDCCActiveDPP,
+               dml_uint_t MetaChunkSize,
+               dml_float_t ReturnBW,
+               dml_bool_t GPUVMEnable,
+               dml_bool_t HostVMEnable,
+               dml_uint_t NumberOfActiveSurfaces,
+               dml_uint_t NumberOfDPP[],
+               dml_uint_t dpte_group_bytes[],
+               dml_float_t HostVMInefficiencyFactor,
+               dml_uint_t HostVMMinPageSize,
+               dml_uint_t HostVMMaxNonCachedPageTableLevels)
+{
+       dml_float_t ExtraLatencyBytes;
+       dml_float_t ExtraLatency;
+
+       ExtraLatencyBytes = CalculateExtraLatencyBytes(
+                       ReorderingBytes,
+                       TotalNumberOfActiveDPP,
+                       PixelChunkSizeInKByte,
+                       TotalNumberOfDCCActiveDPP,
+                       MetaChunkSize,
+                       GPUVMEnable,
+                       HostVMEnable,
+                       NumberOfActiveSurfaces,
+                       NumberOfDPP,
+                       dpte_group_bytes,
+                       HostVMInefficiencyFactor,
+                       HostVMMinPageSize,
+                       HostVMMaxNonCachedPageTableLevels);
+
+       ExtraLatency = (RoundTripPingLatencyCycles + __DML_ARB_TO_RET_DELAY__) / DCFCLK + ExtraLatencyBytes / ReturnBW;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: RoundTripPingLatencyCycles=%u\n", __func__, RoundTripPingLatencyCycles);
+       dml_print("DML::%s: DCFCLK=%f\n", __func__, DCFCLK);
+       dml_print("DML::%s: ExtraLatencyBytes=%f\n", __func__, ExtraLatencyBytes);
+       dml_print("DML::%s: ReturnBW=%f\n", __func__, ReturnBW);
+       dml_print("DML::%s: ExtraLatency=%f\n", __func__, ExtraLatency);
+#endif
+
+       return ExtraLatency;
+} // CalculateExtraLatency
+
+static dml_uint_t CalculateHostVMDynamicLevels(
+                                                                       dml_bool_t GPUVMEnable,
+                                                                       dml_bool_t HostVMEnable,
+                                                                       dml_uint_t HostVMMinPageSize,
+                                                                       dml_uint_t HostVMMaxNonCachedPageTableLevels)
+{
+       dml_uint_t HostVMDynamicLevels = 0;
+
+       if (GPUVMEnable && HostVMEnable) {
+               if (HostVMMinPageSize < 2048)
+                       HostVMDynamicLevels = HostVMMaxNonCachedPageTableLevels;
+               else if (HostVMMinPageSize >= 2048 && HostVMMinPageSize < 1048576)
+                       HostVMDynamicLevels = (dml_uint_t) dml_max(0, (dml_float_t) HostVMMaxNonCachedPageTableLevels - 1);
+               else
+                       HostVMDynamicLevels = (dml_uint_t) dml_max(0, (dml_float_t) HostVMMaxNonCachedPageTableLevels - 2);
+       } else {
+               HostVMDynamicLevels = 0;
+       }
+       return HostVMDynamicLevels;
+}
+
+static dml_uint_t CalculateExtraLatencyBytes(dml_uint_t ReorderingBytes,
+                                                                               dml_uint_t TotalNumberOfActiveDPP,
+                                                                               dml_uint_t PixelChunkSizeInKByte,
+                                                                               dml_uint_t TotalNumberOfDCCActiveDPP,
+                                                                               dml_uint_t MetaChunkSize,
+                                                                               dml_bool_t GPUVMEnable,
+                                                                               dml_bool_t HostVMEnable,
+                                                                               dml_uint_t NumberOfActiveSurfaces,
+                                                                               dml_uint_t NumberOfDPP[],
+                                                                               dml_uint_t dpte_group_bytes[],
+                                                                               dml_float_t HostVMInefficiencyFactor,
+                                                                               dml_uint_t HostVMMinPageSize,
+                                                                               dml_uint_t HostVMMaxNonCachedPageTableLevels)
+{
+       dml_uint_t  HostVMDynamicLevels = CalculateHostVMDynamicLevels(GPUVMEnable, HostVMEnable, HostVMMinPageSize, HostVMMaxNonCachedPageTableLevels);
+       dml_float_t ret                 = ReorderingBytes + (TotalNumberOfActiveDPP * PixelChunkSizeInKByte + TotalNumberOfDCCActiveDPP * MetaChunkSize) * 1024.0;
+
+       if (GPUVMEnable == true) {
+               for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+                       ret = ret + NumberOfDPP[k] * dpte_group_bytes[k] * (1 + 8 * HostVMDynamicLevels) * HostVMInefficiencyFactor;
+               }
+       }
+       return (dml_uint_t)(ret);
+}
+
+static dml_float_t CalculateUrgentLatency(
+               dml_float_t UrgentLatencyPixelDataOnly,
+               dml_float_t UrgentLatencyPixelMixedWithVMData,
+               dml_float_t UrgentLatencyVMDataOnly,
+               dml_bool_t DoUrgentLatencyAdjustment,
+               dml_float_t UrgentLatencyAdjustmentFabricClockComponent,
+               dml_float_t UrgentLatencyAdjustmentFabricClockReference,
+               dml_float_t FabricClock)
+{
+       dml_float_t   ret;
+
+       ret = dml_max3(UrgentLatencyPixelDataOnly, UrgentLatencyPixelMixedWithVMData, UrgentLatencyVMDataOnly);
+       if (DoUrgentLatencyAdjustment == true) {
+               ret = ret + UrgentLatencyAdjustmentFabricClockComponent * (UrgentLatencyAdjustmentFabricClockReference / FabricClock - 1);
+       }
+       return ret;
+}
+
+static dml_float_t RequiredDTBCLK(
+               dml_bool_t DSCEnable,
+               dml_float_t PixelClock,
+               enum dml_output_format_class OutputFormat,
+               dml_float_t OutputBpp,
+               dml_uint_t DSCSlices,
+               dml_uint_t HTotal,
+               dml_uint_t HActive,
+               dml_uint_t AudioRate,
+               dml_uint_t AudioLayout)
+{
+       if (DSCEnable != true) {
+               return dml_max(PixelClock / 4.0 * OutputBpp / 24.0, 25.0);
+       } else {
+               dml_float_t PixelWordRate = PixelClock / (OutputFormat == dml_444 ? 1 : 2);
+               dml_float_t HCActive = dml_ceil(DSCSlices * dml_ceil(OutputBpp * dml_ceil(HActive / DSCSlices, 1) / 8.0, 1) / 3.0, 1);
+               dml_float_t HCBlank = 64 + 32 * dml_ceil(AudioRate * (AudioLayout == 1 ? 1 : 0.25) * HTotal / (PixelClock * 1000), 1);
+               dml_float_t AverageTribyteRate = PixelWordRate * (HCActive + HCBlank) / HTotal;
+               dml_float_t HActiveTribyteRate = PixelWordRate * HCActive / HActive;
+               return dml_max4(PixelWordRate / 4.0, AverageTribyteRate / 4.0, HActiveTribyteRate / 4.0, 25.0) * 1.002;
+       }
+}
+
+static void UseMinimumDCFCLK(struct display_mode_lib_scratch_st *scratch, struct UseMinimumDCFCLK_params_st *p)
+{
+       struct UseMinimumDCFCLK_locals_st *s = &scratch->UseMinimumDCFCLK_locals;
+
+       s->NormalEfficiency = p->PercentOfIdealSDPPortBWReceivedAfterUrgLatency / 100.0;
+       for (dml_uint_t j = 0; j < 2; ++j) {
+
+
+               s->TotalMaxPrefetchFlipDPTERowBandwidth[j] = 0;
+               for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+                       s->TotalMaxPrefetchFlipDPTERowBandwidth[j] = s->TotalMaxPrefetchFlipDPTERowBandwidth[j] + p->NoOfDPP[j][k] * p->DPTEBytesPerRow[j][k] / (15.75 * p->HTotal[k] / p->PixelClock[k]);
+               }
+
+               for (dml_uint_t k = 0; k <= p->NumberOfActiveSurfaces - 1; ++k) {
+                       s->NoOfDPPState[k] = p->NoOfDPP[j][k];
+               }
+
+               s->DPTEBandwidth = s->TotalMaxPrefetchFlipDPTERowBandwidth[j];
+
+               s->DCFCLKRequiredForAverageBandwidth = dml_max(p->ProjectedDCFCLKDeepSleep[j], s->DPTEBandwidth / s->NormalEfficiency / p->ReturnBusWidth);
+
+               s->ExtraLatencyBytes = CalculateExtraLatencyBytes(p->ReorderingBytes, p->TotalNumberOfActiveDPP[j], p->PixelChunkSizeInKByte, p->TotalNumberOfDCCActiveDPP[j],
+                       p->MetaChunkSize, p->GPUVMEnable, p->HostVMEnable, p->NumberOfActiveSurfaces, s->NoOfDPPState, p->dpte_group_bytes,
+                                                                                               1, p->HostVMMinPageSize, p->HostVMMaxNonCachedPageTableLevels);
+               s->ExtraLatencyCycles = p->RoundTripPingLatencyCycles + __DML_ARB_TO_RET_DELAY__ + s->ExtraLatencyBytes / s->NormalEfficiency / p->ReturnBusWidth;
+               for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+                       dml_float_t DCFCLKCyclesRequiredInPrefetch;
+                       dml_float_t PrefetchTime;
+
+                       s->PixelDCFCLKCyclesRequiredInPrefetch[k] = (p->PrefetchLinesY[j][k] * p->swath_width_luma_ub_all_states[j][k] * p->BytePerPixelY[k] + p->PrefetchLinesC[j][k] * p->swath_width_chroma_ub_all_states[j][k] * p->BytePerPixelC[k]) / s->NormalEfficiency / p->ReturnBusWidth;
+                       DCFCLKCyclesRequiredInPrefetch = 2 * s->ExtraLatencyCycles / s->NoOfDPPState[k] + p->PDEAndMetaPTEBytesPerFrame[j][k] / s->NormalEfficiency / s->NormalEfficiency / p->ReturnBusWidth * (p->GPUVMMaxPageTableLevels > 2 ? 1 : 0) + 2 * p->DPTEBytesPerRow[j][k] / s->NormalEfficiency / s->NormalEfficiency / p->ReturnBusWidth + 2 * p->MetaRowBytes[j][k] / s->NormalEfficiency / p->ReturnBusWidth + s->PixelDCFCLKCyclesRequiredInPrefetch[k];
+                       s->PrefetchPixelLinesTime[k] = dml_max(p->PrefetchLinesY[j][k], p->PrefetchLinesC[j][k]) * p->HTotal[k] / p->PixelClock[k];
+                       s->DynamicMetadataVMExtraLatency[k] = (p->GPUVMEnable == true && p->DynamicMetadataEnable[k] == true && p->DynamicMetadataVMEnabled == true) ? p->UrgLatency * p->GPUVMMaxPageTableLevels * (p->HostVMEnable == true ? p->HostVMMaxNonCachedPageTableLevels + 1 : 1) : 0;
+
+                       s->MinimumTWait = CalculateTWait(p->MaxPrefetchMode,
+                               p->UseMALLForPStateChange[k],
+                               p->SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+                               p->DRRDisplay[k],
+                               p->DRAMClockChangeLatencyFinal,
+                               p->FCLKChangeLatency,
+                               p->UrgLatency,
+                               p->SREnterPlusExitTime);
+
+                       PrefetchTime = (p->MaximumVStartup[j][k] - 1) * p->HTotal[k] / p->PixelClock[k] - s->MinimumTWait - p->UrgLatency * ((p->GPUVMMaxPageTableLevels <= 2 ? p->GPUVMMaxPageTableLevels : p->GPUVMMaxPageTableLevels - 2) *  (p->HostVMEnable == true ? p->HostVMMaxNonCachedPageTableLevels + 1 : 1) - 1) - s->DynamicMetadataVMExtraLatency[k];
+
+                       if (PrefetchTime > 0) {
+                               dml_float_t ExpectedVRatioPrefetch;
+                               ExpectedVRatioPrefetch = s->PrefetchPixelLinesTime[k] / (PrefetchTime * s->PixelDCFCLKCyclesRequiredInPrefetch[k] / DCFCLKCyclesRequiredInPrefetch);
+                               s->DCFCLKRequiredForPeakBandwidthPerSurface[k] = s->NoOfDPPState[k] * s->PixelDCFCLKCyclesRequiredInPrefetch[k] / s->PrefetchPixelLinesTime[k] * dml_max(1.0, ExpectedVRatioPrefetch) * dml_max(1.0, ExpectedVRatioPrefetch / 4);
+                               if (p->HostVMEnable == true || p->ImmediateFlipRequirement == true) {
+                                       s->DCFCLKRequiredForPeakBandwidthPerSurface[k] = s->DCFCLKRequiredForPeakBandwidthPerSurface[k] + s->NoOfDPPState[k] * s->DPTEBandwidth / s->NormalEfficiency / s->NormalEfficiency / p->ReturnBusWidth;
+                               }
+                       } else {
+                               s->DCFCLKRequiredForPeakBandwidthPerSurface[k] = p->DCFCLKPerState;
+                       }
+                       if (p->DynamicMetadataEnable[k] == true) {
+                               dml_float_t TSetupPipe;
+                               dml_float_t TdmbfPipe;
+                               dml_float_t TdmsksPipe;
+                               dml_float_t TdmecPipe;
+                               dml_float_t AllowedTimeForUrgentExtraLatency;
+
+                               CalculateVUpdateAndDynamicMetadataParameters(
+                                       p->MaxInterDCNTileRepeaters,
+                                       p->RequiredDPPCLKPerSurface[j][k],
+                                       p->RequiredDISPCLK[j],
+                                       p->ProjectedDCFCLKDeepSleep[j],
+                                       p->PixelClock[k],
+                                       p->HTotal[k],
+                                       p->VTotal[k] - p->VActive[k],
+                                       p->DynamicMetadataTransmittedBytes[k],
+                                       p->DynamicMetadataLinesBeforeActiveRequired[k],
+                                       p->Interlace[k],
+                                       p->ProgressiveToInterlaceUnitInOPP,
+
+                                       // Output
+                                       &TSetupPipe,
+                                       &TdmbfPipe,
+                                       &TdmecPipe,
+                                       &TdmsksPipe,
+                                       &s->dummy1,
+                                       &s->dummy2,
+                                       &s->dummy3);
+
+                               AllowedTimeForUrgentExtraLatency = p->MaximumVStartup[j][k] * p->HTotal[k] / p->PixelClock[k] - s->MinimumTWait - TSetupPipe - TdmbfPipe - TdmecPipe - TdmsksPipe - s->DynamicMetadataVMExtraLatency[k];
+                               if (AllowedTimeForUrgentExtraLatency > 0) {
+                                       s->DCFCLKRequiredForPeakBandwidthPerSurface[k] = dml_max(s->DCFCLKRequiredForPeakBandwidthPerSurface[k], s->ExtraLatencyCycles / AllowedTimeForUrgentExtraLatency);
+                               } else {
+                                       s->DCFCLKRequiredForPeakBandwidthPerSurface[k] = p->DCFCLKPerState;
+                               }
+                       }
+               }
+               s->DCFCLKRequiredForPeakBandwidth = 0;
+               for (dml_uint_t k = 0; k <= p->NumberOfActiveSurfaces - 1; ++k) {
+                       s->DCFCLKRequiredForPeakBandwidth = s->DCFCLKRequiredForPeakBandwidth + s->DCFCLKRequiredForPeakBandwidthPerSurface[k];
+               }
+               s->MinimumTvmPlus2Tr0 = p->UrgLatency * (p->GPUVMEnable == true ? (p->HostVMEnable == true ? (p->GPUVMMaxPageTableLevels + 2) * (p->HostVMMaxNonCachedPageTableLevels + 1) - 1 : p->GPUVMMaxPageTableLevels + 1) : 0);
+               for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+                       dml_float_t MaximumTvmPlus2Tr0PlusTsw;
+                       MaximumTvmPlus2Tr0PlusTsw = (p->MaximumVStartup[j][k] - 2) * p->HTotal[k] / p->PixelClock[k] - s->MinimumTWait - s->DynamicMetadataVMExtraLatency[k];
+                       if (MaximumTvmPlus2Tr0PlusTsw <= s->MinimumTvmPlus2Tr0 + s->PrefetchPixelLinesTime[k] / 4) {
+                               s->DCFCLKRequiredForPeakBandwidth = p->DCFCLKPerState;
+                       } else {
+                               s->DCFCLKRequiredForPeakBandwidth = dml_max3(s->DCFCLKRequiredForPeakBandwidth,
+                                                                                                               2 * s->ExtraLatencyCycles / (MaximumTvmPlus2Tr0PlusTsw - s->MinimumTvmPlus2Tr0 - s->PrefetchPixelLinesTime[k] / 4),
+                                                                                                               (2 * s->ExtraLatencyCycles + s->PixelDCFCLKCyclesRequiredInPrefetch[k]) / (MaximumTvmPlus2Tr0PlusTsw - s->MinimumTvmPlus2Tr0));
+                       }
+               }
+               p->DCFCLKState[j] = dml_min(p->DCFCLKPerState, 1.05 * dml_max(s->DCFCLKRequiredForAverageBandwidth, s->DCFCLKRequiredForPeakBandwidth));
+       }
+}
+
+
+static dml_bool_t UnboundedRequest(enum dml_unbounded_requesting_policy UseUnboundedRequestingFinal,
+                                               dml_uint_t TotalNumberOfActiveDPP,
+                                               dml_bool_t NoChromaOrLinear,
+                                               enum dml_output_encoder_class Output)
+{
+       dml_bool_t ret_val = false;
+
+       ret_val = (UseUnboundedRequestingFinal != dml_unbounded_requesting_disable
+                       && TotalNumberOfActiveDPP == 1 && NoChromaOrLinear);
+       if (UseUnboundedRequestingFinal == dml_unbounded_requesting_edp_only && Output != dml_edp) {
+               ret_val = false;
+       }
+       return (ret_val);
+}
+
+static void CalculateSurfaceSizeInMall(
+               dml_uint_t NumberOfActiveSurfaces,
+               dml_uint_t MALLAllocatedForDCN,
+               enum dml_use_mall_for_static_screen_mode UseMALLForStaticScreen[],
+               dml_bool_t DCCEnable[],
+               dml_bool_t ViewportStationary[],
+               dml_uint_t ViewportXStartY[],
+               dml_uint_t ViewportYStartY[],
+               dml_uint_t ViewportXStartC[],
+               dml_uint_t ViewportYStartC[],
+               dml_uint_t ViewportWidthY[],
+               dml_uint_t ViewportHeightY[],
+               dml_uint_t BytesPerPixelY[],
+               dml_uint_t ViewportWidthC[],
+               dml_uint_t ViewportHeightC[],
+               dml_uint_t BytesPerPixelC[],
+               dml_uint_t SurfaceWidthY[],
+               dml_uint_t SurfaceWidthC[],
+               dml_uint_t SurfaceHeightY[],
+               dml_uint_t SurfaceHeightC[],
+               dml_uint_t Read256BytesBlockWidthY[],
+               dml_uint_t Read256BytesBlockWidthC[],
+               dml_uint_t Read256BytesBlockHeightY[],
+               dml_uint_t Read256BytesBlockHeightC[],
+               dml_uint_t ReadBlockWidthY[],
+               dml_uint_t ReadBlockWidthC[],
+               dml_uint_t ReadBlockHeightY[],
+               dml_uint_t ReadBlockHeightC[],
+
+               // Output
+               dml_uint_t SurfaceSizeInMALL[],
+               dml_bool_t *ExceededMALLSize)
+{
+       dml_uint_t TotalSurfaceSizeInMALL  = 0;
+
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               if (ViewportStationary[k]) {
+                       SurfaceSizeInMALL[k] = (dml_uint_t)(dml_min(dml_ceil(SurfaceWidthY[k], ReadBlockWidthY[k]), dml_floor(ViewportXStartY[k] + ViewportWidthY[k] + ReadBlockWidthY[k] - 1, ReadBlockWidthY[k]) - dml_floor(ViewportXStartY[k], ReadBlockWidthY[k])) *
+                                                                       dml_min(dml_ceil(SurfaceHeightY[k], ReadBlockHeightY[k]), dml_floor(ViewportYStartY[k] + ViewportHeightY[k] + ReadBlockHeightY[k] - 1, ReadBlockHeightY[k]) - dml_floor(ViewportYStartY[k], ReadBlockHeightY[k])) *
+                                                                       BytesPerPixelY[k]);
+
+                       if (ReadBlockWidthC[k] > 0) {
+                               SurfaceSizeInMALL[k] = (dml_uint_t)(SurfaceSizeInMALL[k] +
+                                                                               dml_min(dml_ceil(SurfaceWidthC[k], ReadBlockWidthC[k]), dml_floor(ViewportXStartC[k] + ViewportWidthC[k] + ReadBlockWidthC[k] - 1, ReadBlockWidthC[k]) - dml_floor(ViewportXStartC[k], ReadBlockWidthC[k])) *
+                                                                               dml_min(dml_ceil(SurfaceHeightC[k], ReadBlockHeightC[k]), dml_floor(ViewportYStartC[k] + ViewportHeightC[k] + ReadBlockHeightC[k] - 1, ReadBlockHeightC[k]) - dml_floor(ViewportYStartC[k], ReadBlockHeightC[k])) * BytesPerPixelC[k]);
+                       }
+                       if (DCCEnable[k] == true) {
+                               SurfaceSizeInMALL[k] = (dml_uint_t)(SurfaceSizeInMALL[k] +
+                                       dml_min(dml_ceil(SurfaceWidthY[k], 8 * Read256BytesBlockWidthY[k]), dml_floor(ViewportXStartY[k] + ViewportWidthY[k] + 8 * Read256BytesBlockWidthY[k] - 1, 8 * Read256BytesBlockWidthY[k]) - dml_floor(ViewportXStartY[k], 8 * Read256BytesBlockWidthY[k])) *
+                                       dml_min(dml_ceil(SurfaceHeightY[k], 8 * Read256BytesBlockHeightY[k]), dml_floor(ViewportYStartY[k] + ViewportHeightY[k] + 8 * Read256BytesBlockHeightY[k] - 1, 8 * Read256BytesBlockHeightY[k]) - dml_floor(ViewportYStartY[k], 8 * Read256BytesBlockHeightY[k])) * BytesPerPixelY[k] / 256);
+                               if (Read256BytesBlockWidthC[k] > 0) {
+                                       SurfaceSizeInMALL[k] = (dml_uint_t)(SurfaceSizeInMALL[k] +
+                                               dml_min(dml_ceil(SurfaceWidthC[k], 8 * Read256BytesBlockWidthC[k]), dml_floor(ViewportXStartC[k] + ViewportWidthC[k] + 8 * Read256BytesBlockWidthC[k] - 1, 8 * Read256BytesBlockWidthC[k]) - dml_floor(ViewportXStartC[k], 8 * Read256BytesBlockWidthC[k])) *
+                                               dml_min(dml_ceil(SurfaceHeightC[k], 8 * Read256BytesBlockHeightC[k]), dml_floor(ViewportYStartC[k] + ViewportHeightC[k] + 8 * Read256BytesBlockHeightC[k] - 1, 8 * Read256BytesBlockHeightC[k]) - dml_floor(ViewportYStartC[k], 8 * Read256BytesBlockHeightC[k])) * BytesPerPixelC[k] / 256);
+                               }
+                       }
+               } else {
+                       SurfaceSizeInMALL[k] = (dml_uint_t)(dml_ceil(dml_min(SurfaceWidthY[k], ViewportWidthY[k] + ReadBlockWidthY[k] - 1), ReadBlockWidthY[k]) * dml_ceil(dml_min(SurfaceHeightY[k], ViewportHeightY[k] + ReadBlockHeightY[k] - 1), ReadBlockHeightY[k]) * BytesPerPixelY[k]);
+                       if (ReadBlockWidthC[k] > 0) {
+                               SurfaceSizeInMALL[k] = (dml_uint_t)(SurfaceSizeInMALL[k] +
+                                                                               dml_ceil(dml_min(SurfaceWidthC[k], ViewportWidthC[k] + ReadBlockWidthC[k] - 1), ReadBlockWidthC[k]) *
+                                                                               dml_ceil(dml_min(SurfaceHeightC[k], ViewportHeightC[k] + ReadBlockHeightC[k] - 1), ReadBlockHeightC[k]) * BytesPerPixelC[k]);
+               }
+               if (DCCEnable[k] == true) {
+                       SurfaceSizeInMALL[k] = (dml_uint_t)(SurfaceSizeInMALL[k] +
+                                       dml_ceil(dml_min(SurfaceWidthY[k], ViewportWidthY[k] + 8 * Read256BytesBlockWidthY[k] - 1), 8 * Read256BytesBlockWidthY[k]) *
+                                       dml_ceil(dml_min(SurfaceHeightY[k], ViewportHeightY[k] + 8 * Read256BytesBlockHeightY[k] - 1), 8 * Read256BytesBlockHeightY[k]) * BytesPerPixelY[k] / 256);
+
+                               if (Read256BytesBlockWidthC[k] > 0) {
+                                       SurfaceSizeInMALL[k] = (dml_uint_t)(SurfaceSizeInMALL[k] +
+                                               dml_ceil(dml_min(SurfaceWidthC[k], ViewportWidthC[k] + 8 * Read256BytesBlockWidthC[k] - 1), 8 * Read256BytesBlockWidthC[k]) *
+                                               dml_ceil(dml_min(SurfaceHeightC[k], ViewportHeightC[k] + 8 * Read256BytesBlockHeightC[k] - 1), 8 * Read256BytesBlockHeightC[k]) * BytesPerPixelC[k] / 256);
+                               }
+                       }
+               }
+       }
+
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               if (UseMALLForStaticScreen[k] == dml_use_mall_static_screen_enable)
+                       TotalSurfaceSizeInMALL = TotalSurfaceSizeInMALL + SurfaceSizeInMALL[k];
+       }
+       *ExceededMALLSize = (TotalSurfaceSizeInMALL <= MALLAllocatedForDCN * 1024 * 1024 ? false : true);
+} // CalculateSurfaceSizeInMall
+
+static void CalculateDETBufferSize(
+                                               dml_uint_t DETSizeOverride[],
+                                               enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+                                               dml_bool_t ForceSingleDPP,
+                                               dml_uint_t NumberOfActiveSurfaces,
+                                               dml_bool_t UnboundedRequestEnabled,
+                                               dml_uint_t nomDETInKByte,
+                                               dml_uint_t MaxTotalDETInKByte,
+                                               dml_uint_t ConfigReturnBufferSizeInKByte,
+                                               dml_uint_t MinCompressedBufferSizeInKByte,
+                                               dml_uint_t ConfigReturnBufferSegmentSizeInkByte,
+                                               dml_uint_t CompressedBufferSegmentSizeInkByteFinal,
+                                               enum dml_source_format_class SourcePixelFormat[],
+                                               dml_float_t ReadBandwidthLuma[],
+                                               dml_float_t ReadBandwidthChroma[],
+                                               dml_uint_t RoundedUpMaxSwathSizeBytesY[],
+                                               dml_uint_t RoundedUpMaxSwathSizeBytesC[],
+                                               dml_uint_t DPPPerSurface[],
+                                               // Output
+                                               dml_uint_t DETBufferSizeInKByte[],
+                                               dml_uint_t *CompressedBufferSizeInkByte)
+{
+       dml_uint_t DETBufferSizePoolInKByte;
+       dml_uint_t NextDETBufferPieceInKByte;
+       dml_bool_t DETPieceAssignedToThisSurfaceAlready[__DML_NUM_PLANES__];
+       dml_bool_t NextPotentialSurfaceToAssignDETPieceFound;
+       dml_uint_t NextSurfaceToAssignDETPiece;
+       dml_float_t TotalBandwidth;
+       dml_float_t BandwidthOfSurfacesNotAssignedDETPiece;
+       dml_uint_t max_minDET;
+       dml_uint_t minDET;
+       dml_uint_t minDET_pipe;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: ForceSingleDPP = %u\n", __func__, ForceSingleDPP);
+       dml_print("DML::%s: nomDETInKByte = %u\n", __func__, nomDETInKByte);
+       dml_print("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces);
+       dml_print("DML::%s: UnboundedRequestEnabled = %u\n", __func__, UnboundedRequestEnabled);
+       dml_print("DML::%s: MaxTotalDETInKByte = %u\n", __func__, MaxTotalDETInKByte);
+       dml_print("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, ConfigReturnBufferSizeInKByte);
+       dml_print("DML::%s: MinCompressedBufferSizeInKByte = %u\n", __func__, MinCompressedBufferSizeInKByte);
+       dml_print("DML::%s: CompressedBufferSegmentSizeInkByteFinal = %u\n", __func__, CompressedBufferSegmentSizeInkByteFinal);
+#endif
+
+       // Note: Will use default det size if that fits 2 swaths
+       if (UnboundedRequestEnabled) {
+               if (DETSizeOverride[0] > 0) {
+                       DETBufferSizeInKByte[0] = DETSizeOverride[0];
+               } else {
+                       DETBufferSizeInKByte[0] = (dml_uint_t) dml_max(128.0, dml_ceil(2.0 * ((dml_float_t) RoundedUpMaxSwathSizeBytesY[0] + (dml_float_t) RoundedUpMaxSwathSizeBytesC[0]) / 1024.0, ConfigReturnBufferSegmentSizeInkByte));
+               }
+               *CompressedBufferSizeInkByte = ConfigReturnBufferSizeInKByte - DETBufferSizeInKByte[0];
+       } else {
+               DETBufferSizePoolInKByte = MaxTotalDETInKByte;
+               for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+                       DETBufferSizeInKByte[k] = 0;
+                       if (SourcePixelFormat[k] == dml_420_8 || SourcePixelFormat[k] == dml_420_10 || SourcePixelFormat[k] == dml_420_12) {
+                               max_minDET = nomDETInKByte - ConfigReturnBufferSegmentSizeInkByte;
+                       } else {
+                               max_minDET = nomDETInKByte;
+                       }
+                       minDET = 128;
+                       minDET_pipe = 0;
+
+                       // add DET resource until can hold 2 full swaths
+                       while (minDET <= max_minDET && minDET_pipe == 0) {
+                               if (2.0 * ((dml_float_t) RoundedUpMaxSwathSizeBytesY[k] + (dml_float_t) RoundedUpMaxSwathSizeBytesC[k]) / 1024.0 <= minDET)
+                                       minDET_pipe = minDET;
+                               minDET = minDET + ConfigReturnBufferSegmentSizeInkByte;
+                       }
+
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%u minDET = %u\n", __func__, k, minDET);
+                       dml_print("DML::%s: k=%u max_minDET = %u\n", __func__, k, max_minDET);
+                       dml_print("DML::%s: k=%u minDET_pipe = %u\n", __func__, k, minDET_pipe);
+                       dml_print("DML::%s: k=%u RoundedUpMaxSwathSizeBytesY = %u\n", __func__, k, RoundedUpMaxSwathSizeBytesY[k]);
+                       dml_print("DML::%s: k=%u RoundedUpMaxSwathSizeBytesC = %u\n", __func__, k, RoundedUpMaxSwathSizeBytesC[k]);
+#endif
+
+                       if (minDET_pipe == 0) {
+                               minDET_pipe = (dml_uint_t)(dml_max(128, dml_ceil(((dml_float_t)RoundedUpMaxSwathSizeBytesY[k] + (dml_float_t)RoundedUpMaxSwathSizeBytesC[k]) / 1024.0, ConfigReturnBufferSegmentSizeInkByte)));
+#ifdef __DML_VBA_DEBUG__
+                               dml_print("DML::%s: k=%u minDET_pipe = %u (assume each plane take half DET)\n", __func__, k, minDET_pipe);
+#endif
+                       }
+
+                       if (UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe) {
+                               DETBufferSizeInKByte[k] = 0;
+                       } else if (DETSizeOverride[k] > 0) {
+                               DETBufferSizeInKByte[k] = DETSizeOverride[k];
+                               DETBufferSizePoolInKByte = DETBufferSizePoolInKByte - (ForceSingleDPP ? 1 : DPPPerSurface[k]) * DETSizeOverride[k];
+                       } else if ((ForceSingleDPP ? 1 : DPPPerSurface[k]) * minDET_pipe <= DETBufferSizePoolInKByte) {
+                               DETBufferSizeInKByte[k] = minDET_pipe;
+                               DETBufferSizePoolInKByte = DETBufferSizePoolInKByte - (ForceSingleDPP ? 1 : DPPPerSurface[k]) * minDET_pipe;
+                       }
+
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%u DPPPerSurface = %u\n", __func__, k, DPPPerSurface[k]);
+                       dml_print("DML::%s: k=%u DETSizeOverride = %u\n", __func__, k, DETSizeOverride[k]);
+                       dml_print("DML::%s: k=%u DETBufferSizeInKByte = %u\n", __func__, k, DETBufferSizeInKByte[k]);
+                       dml_print("DML::%s: DETBufferSizePoolInKByte = %u\n", __func__, DETBufferSizePoolInKByte);
+#endif
+               }
+
+               TotalBandwidth = 0;
+               for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+                       if (UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe)
+                               TotalBandwidth = TotalBandwidth + ReadBandwidthLuma[k] + ReadBandwidthChroma[k];
+               }
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: --- Before bandwidth adjustment ---\n", __func__);
+               for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+                       dml_print("DML::%s: k=%u DETBufferSizeInKByte   = %u\n", __func__, k, DETBufferSizeInKByte[k]);
+               }
+               dml_print("DML::%s: --- DET allocation with bandwidth ---\n", __func__);
+#endif
+               dml_print("DML::%s: TotalBandwidth = %f\n", __func__, TotalBandwidth);
+               BandwidthOfSurfacesNotAssignedDETPiece = TotalBandwidth;
+               for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+                       if (UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe) {
+                               DETPieceAssignedToThisSurfaceAlready[k] = true;
+                       } else if (DETSizeOverride[k] > 0 || (((dml_float_t) (ForceSingleDPP ? 1 : DPPPerSurface[k]) * (dml_float_t) DETBufferSizeInKByte[k] / (dml_float_t) MaxTotalDETInKByte) >= ((ReadBandwidthLuma[k] + ReadBandwidthChroma[k]) / TotalBandwidth))) {
+                               DETPieceAssignedToThisSurfaceAlready[k] = true;
+                               BandwidthOfSurfacesNotAssignedDETPiece = BandwidthOfSurfacesNotAssignedDETPiece - ReadBandwidthLuma[k] - ReadBandwidthChroma[k];
+                       } else {
+                               DETPieceAssignedToThisSurfaceAlready[k] = false;
+                       }
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%u DETPieceAssignedToThisSurfaceAlready = %u\n", __func__, k, DETPieceAssignedToThisSurfaceAlready[k]);
+                       dml_print("DML::%s: k=%u BandwidthOfSurfacesNotAssignedDETPiece = %f\n", __func__, k, BandwidthOfSurfacesNotAssignedDETPiece);
+#endif
+               }
+
+               for (dml_uint_t j = 0; j < NumberOfActiveSurfaces; ++j) {
+                       NextPotentialSurfaceToAssignDETPieceFound = false;
+                       NextSurfaceToAssignDETPiece = 0;
+
+                       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+#ifdef __DML_VBA_DEBUG__
+                               dml_print("DML::%s: j=%u k=%u, ReadBandwidthLuma[k] = %f\n", __func__, j, k, ReadBandwidthLuma[k]);
+                               dml_print("DML::%s: j=%u k=%u, ReadBandwidthChroma[k] = %f\n", __func__, j, k, ReadBandwidthChroma[k]);
+                               dml_print("DML::%s: j=%u k=%u, ReadBandwidthLuma[Next] = %f\n", __func__, j, k, ReadBandwidthLuma[NextSurfaceToAssignDETPiece]);
+                               dml_print("DML::%s: j=%u k=%u, ReadBandwidthChroma[Next] = %f\n", __func__, j, k, ReadBandwidthChroma[NextSurfaceToAssignDETPiece]);
+                               dml_print("DML::%s: j=%u k=%u, NextSurfaceToAssignDETPiece = %u\n", __func__, j, k, NextSurfaceToAssignDETPiece);
+#endif
+                               if (!DETPieceAssignedToThisSurfaceAlready[k] && (!NextPotentialSurfaceToAssignDETPieceFound ||
+                                       ReadBandwidthLuma[k] + ReadBandwidthChroma[k] < ReadBandwidthLuma[NextSurfaceToAssignDETPiece] + ReadBandwidthChroma[NextSurfaceToAssignDETPiece])) {
+                                       NextSurfaceToAssignDETPiece = k;
+                                       NextPotentialSurfaceToAssignDETPieceFound = true;
+                               }
+#ifdef __DML_VBA_DEBUG__
+                               dml_print("DML::%s: j=%u k=%u, DETPieceAssignedToThisSurfaceAlready = %u\n", __func__, j, k, DETPieceAssignedToThisSurfaceAlready[k]);
+                               dml_print("DML::%s: j=%u k=%u, NextPotentialSurfaceToAssignDETPieceFound = %u\n", __func__, j, k, NextPotentialSurfaceToAssignDETPieceFound);
+#endif
+                       }
+
+                       if (NextPotentialSurfaceToAssignDETPieceFound) {
+                               // Note: To show the banker's rounding behavior in VBA and also the fact that the DET buffer size varies due to precision issue
+                               //
+                               //dml_float_t tmp1 =  ((dml_float_t) DETBufferSizePoolInKByte * (ReadBandwidthLuma[NextSurfaceToAssignDETPiece] + ReadBandwidthChroma[NextSurfaceToAssignDETPiece]) / BandwidthOfSurfacesNotAssignedDETPiece /
+                               //                         ((ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]) * 64.0));
+                               //dml_float_t tmp2 =  dml_round((dml_float_t) DETBufferSizePoolInKByte * (ReadBandwidthLuma[NextSurfaceToAssignDETPiece] + ReadBandwidthChroma[NextSurfaceToAssignDETPiece]) / BandwidthOfSurfacesNotAssignedDETPiece /
+                               //                         ((ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]) * 64.0));
+                               //
+                               //dml_print("DML::%s: j=%u, tmp1 = %f\n", __func__, j, tmp1);
+                               //dml_print("DML::%s: j=%u, tmp2 = %f\n", __func__, j, tmp2);
+
+                               NextDETBufferPieceInKByte = (dml_uint_t)(dml_min(
+                                                                                       dml_round((dml_float_t) DETBufferSizePoolInKByte * (ReadBandwidthLuma[NextSurfaceToAssignDETPiece] + ReadBandwidthChroma[NextSurfaceToAssignDETPiece]) / BandwidthOfSurfacesNotAssignedDETPiece /
+                                                                                               ((ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]) * ConfigReturnBufferSegmentSizeInkByte), true)
+                                                                                               * (ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]) * ConfigReturnBufferSegmentSizeInkByte,
+                                                                                       dml_floor((dml_float_t) DETBufferSizePoolInKByte, (ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]) * ConfigReturnBufferSegmentSizeInkByte)));
+
+#ifdef __DML_VBA_DEBUG__
+                               dml_print("DML::%s: j=%u, DETBufferSizePoolInKByte = %u\n", __func__, j, DETBufferSizePoolInKByte);
+                               dml_print("DML::%s: j=%u, NextSurfaceToAssignDETPiece = %u\n", __func__, j, NextSurfaceToAssignDETPiece);
+                               dml_print("DML::%s: j=%u, ReadBandwidthLuma[%u] = %f\n", __func__, j, NextSurfaceToAssignDETPiece, ReadBandwidthLuma[NextSurfaceToAssignDETPiece]);
+                               dml_print("DML::%s: j=%u, ReadBandwidthChroma[%u] = %f\n", __func__, j, NextSurfaceToAssignDETPiece, ReadBandwidthChroma[NextSurfaceToAssignDETPiece]);
+                               dml_print("DML::%s: j=%u, BandwidthOfSurfacesNotAssignedDETPiece = %f\n", __func__, j, BandwidthOfSurfacesNotAssignedDETPiece);
+                               dml_print("DML::%s: j=%u, NextDETBufferPieceInKByte = %u\n", __func__, j, NextDETBufferPieceInKByte);
+                               dml_print("DML::%s: j=%u, DETBufferSizeInKByte[%u] increases from %u ", __func__, j, NextSurfaceToAssignDETPiece, DETBufferSizeInKByte[NextSurfaceToAssignDETPiece]);
+#endif
+
+                               DETBufferSizeInKByte[NextSurfaceToAssignDETPiece] = DETBufferSizeInKByte[NextSurfaceToAssignDETPiece] + NextDETBufferPieceInKByte / (ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]);
+#ifdef __DML_VBA_DEBUG__
+                               dml_print("to %u\n", DETBufferSizeInKByte[NextSurfaceToAssignDETPiece]);
+#endif
+
+                               DETBufferSizePoolInKByte = DETBufferSizePoolInKByte - NextDETBufferPieceInKByte;
+                               DETPieceAssignedToThisSurfaceAlready[NextSurfaceToAssignDETPiece] = true;
+                               BandwidthOfSurfacesNotAssignedDETPiece = BandwidthOfSurfacesNotAssignedDETPiece - (ReadBandwidthLuma[NextSurfaceToAssignDETPiece] + ReadBandwidthChroma[NextSurfaceToAssignDETPiece]);
+                       }
+               }
+               *CompressedBufferSizeInkByte = MinCompressedBufferSizeInKByte;
+       }
+       *CompressedBufferSizeInkByte = *CompressedBufferSizeInkByte * CompressedBufferSegmentSizeInkByteFinal / ConfigReturnBufferSegmentSizeInkByte;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: --- After bandwidth adjustment ---\n", __func__);
+       dml_print("DML::%s: CompressedBufferSizeInkByte = %u\n", __func__, *CompressedBufferSizeInkByte);
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               dml_print("DML::%s: k=%u DETBufferSizeInKByte = %u (TotalReadBandWidth=%f)\n", __func__, k, DETBufferSizeInKByte[k], ReadBandwidthLuma[k] + ReadBandwidthChroma[k]);
+       }
+#endif
+} // CalculateDETBufferSize
+
+
+/// @brief Calculate the bound for return buffer sizing
+static void CalculateMaxDETAndMinCompressedBufferSize(
+               dml_uint_t  ConfigReturnBufferSizeInKByte,
+               dml_uint_t  ConfigReturnBufferSegmentSizeInKByte,
+               dml_uint_t  ROBBufferSizeInKByte,
+               dml_uint_t MaxNumDPP,
+               dml_bool_t nomDETInKByteOverrideEnable, // VBA_DELTA, allow DV to override default DET size
+               dml_uint_t nomDETInKByteOverrideValue,  // VBA_DELTA
+
+               // Output
+               dml_uint_t *MaxTotalDETInKByte,
+               dml_uint_t *nomDETInKByte,
+               dml_uint_t *MinCompressedBufferSizeInKByte)
+{
+       *MaxTotalDETInKByte = ConfigReturnBufferSizeInKByte - ConfigReturnBufferSegmentSizeInKByte;
+       *nomDETInKByte = (dml_uint_t)(dml_floor((dml_float_t) *MaxTotalDETInKByte / (dml_float_t) MaxNumDPP, ConfigReturnBufferSegmentSizeInKByte));
+       *MinCompressedBufferSizeInKByte = ConfigReturnBufferSizeInKByte - *MaxTotalDETInKByte;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, ConfigReturnBufferSizeInKByte);
+       dml_print("DML::%s: ROBBufferSizeInKByte = %u\n", __func__, ROBBufferSizeInKByte);
+       dml_print("DML::%s: MaxNumDPP = %u\n", __func__, MaxNumDPP);
+       dml_print("DML::%s: MaxTotalDETInKByte = %u\n", __func__, *MaxTotalDETInKByte);
+       dml_print("DML::%s: nomDETInKByte = %u\n", __func__, *nomDETInKByte);
+       dml_print("DML::%s: MinCompressedBufferSizeInKByte = %u\n", __func__, *MinCompressedBufferSizeInKByte);
+#endif
+
+       if (nomDETInKByteOverrideEnable) {
+               *nomDETInKByte = nomDETInKByteOverrideValue;
+               dml_print("DML::%s: nomDETInKByte = %u (overrided)\n", __func__, *nomDETInKByte);
+       }
+} // CalculateMaxDETAndMinCompressedBufferSize
+
+/// @brief Calculate all the RQ request attributes, like row height and # swath
+static void CalculateVMRowAndSwath(struct display_mode_lib_scratch_st *scratch,
+               struct CalculateVMRowAndSwath_params_st *p)
+{
+       struct CalculateVMRowAndSwath_locals_st *s = &scratch->CalculateVMRowAndSwath_locals;
+
+       s->HostVMDynamicLevels = CalculateHostVMDynamicLevels(p->GPUVMEnable, p->HostVMEnable, p->HostVMMinPageSize, p->HostVMMaxNonCachedPageTableLevels);
+
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               if (p->HostVMEnable == true) {
+                       p->vm_group_bytes[k] = 512;
+                       p->dpte_group_bytes[k] = 512;
+               } else if (p->GPUVMEnable == true) {
+                       p->vm_group_bytes[k] = 2048;
+                       if (p->GPUVMMinPageSizeKBytes[k] >= 64 && dml_is_vertical_rotation(p->myPipe[k].SourceScan)) {
+                               p->dpte_group_bytes[k] = 512;
+                       } else {
+                               p->dpte_group_bytes[k] = 2048;
+                       }
+               } else {
+                       p->vm_group_bytes[k] = 0;
+                       p->dpte_group_bytes[k] = 0;
+               }
+
+               if (p->myPipe[k].SourcePixelFormat == dml_420_8 || p->myPipe[k].SourcePixelFormat == dml_420_10 ||
+                       p->myPipe[k].SourcePixelFormat == dml_420_12 || p->myPipe[k].SourcePixelFormat == dml_rgbe_alpha) {
+                       if ((p->myPipe[k].SourcePixelFormat == dml_420_10 || p->myPipe[k].SourcePixelFormat == dml_420_12) && !dml_is_vertical_rotation(p->myPipe[k].SourceScan)) {
+                               s->PTEBufferSizeInRequestsForLuma[k] = (p->PTEBufferSizeInRequestsLuma + p->PTEBufferSizeInRequestsChroma) / 2;
+                               s->PTEBufferSizeInRequestsForChroma[k] = s->PTEBufferSizeInRequestsForLuma[k];
+                       } else {
+                               s->PTEBufferSizeInRequestsForLuma[k] = p->PTEBufferSizeInRequestsLuma;
+                               s->PTEBufferSizeInRequestsForChroma[k] = p->PTEBufferSizeInRequestsChroma;
+                       }
+
+                       s->PDEAndMetaPTEBytesFrameC = CalculateVMAndRowBytes(
+                               p->myPipe[k].ViewportStationary,
+                               p->myPipe[k].DCCEnable,
+                               p->myPipe[k].DPPPerSurface,
+                               p->myPipe[k].BlockHeight256BytesC,
+                               p->myPipe[k].BlockWidth256BytesC,
+                               p->myPipe[k].SourcePixelFormat,
+                               p->myPipe[k].SurfaceTiling,
+                               p->myPipe[k].BytePerPixelC,
+                               p->myPipe[k].SourceScan,
+                               p->SwathWidthC[k],
+                               p->myPipe[k].ViewportHeightChroma,
+                               p->myPipe[k].ViewportXStartC,
+                               p->myPipe[k].ViewportYStartC,
+                               p->GPUVMEnable,
+                               p->GPUVMMaxPageTableLevels,
+                               p->GPUVMMinPageSizeKBytes[k],
+                               s->PTEBufferSizeInRequestsForChroma[k],
+                               p->myPipe[k].PitchC,
+                               p->myPipe[k].DCCMetaPitchC,
+                               p->myPipe[k].BlockWidthC,
+                               p->myPipe[k].BlockHeightC,
+
+                               // Output
+                               &s->MetaRowByteC[k],
+                               &s->PixelPTEBytesPerRowC[k],
+                               &s->PixelPTEBytesPerRowStorageC[k],
+                               &p->dpte_row_width_chroma_ub[k],
+                               &p->dpte_row_height_chroma[k],
+                               &p->dpte_row_height_linear_chroma[k],
+                               &s->PixelPTEBytesPerRowC_one_row_per_frame[k],
+                               &s->dpte_row_width_chroma_ub_one_row_per_frame[k],
+                               &s->dpte_row_height_chroma_one_row_per_frame[k],
+                               &p->meta_req_width_chroma[k],
+                               &p->meta_req_height_chroma[k],
+                               &p->meta_row_width_chroma[k],
+                               &p->meta_row_height_chroma[k],
+                               &p->PixelPTEReqWidthC[k],
+                               &p->PixelPTEReqHeightC[k],
+                               &p->PTERequestSizeC[k],
+                               &p->dpde0_bytes_per_frame_ub_c[k],
+                               &p->meta_pte_bytes_per_frame_ub_c[k]);
+
+                       p->PrefetchSourceLinesC[k] = CalculatePrefetchSourceLines (
+                               p->myPipe[k].VRatioChroma,
+                               p->myPipe[k].VTapsChroma,
+                               p->myPipe[k].InterlaceEnable,
+                               p->myPipe[k].ProgressiveToInterlaceUnitInOPP,
+                               p->myPipe[k].SwathHeightC,
+                               p->myPipe[k].SourceScan,
+                               p->myPipe[k].ViewportStationary,
+                               p->SwathWidthC[k],
+                               p->myPipe[k].ViewportHeightChroma,
+                               p->myPipe[k].ViewportXStartC,
+                               p->myPipe[k].ViewportYStartC,
+
+                               // Output
+                               &p->VInitPreFillC[k],
+                               &p->MaxNumSwathC[k]);
+               } else {
+                       s->PTEBufferSizeInRequestsForLuma[k] = p->PTEBufferSizeInRequestsLuma + p->PTEBufferSizeInRequestsChroma;
+                       s->PTEBufferSizeInRequestsForChroma[k] = 0;
+                       s->PixelPTEBytesPerRowC[k] = 0;
+                       s->PixelPTEBytesPerRowStorageC[k] = 0;
+                       s->PDEAndMetaPTEBytesFrameC = 0;
+                       s->MetaRowByteC[k] = 0;
+                       p->MaxNumSwathC[k] = 0;
+                       p->PrefetchSourceLinesC[k] = 0;
+                       s->dpte_row_height_chroma_one_row_per_frame[k] = 0;
+                       s->dpte_row_width_chroma_ub_one_row_per_frame[k] = 0;
+                       s->PixelPTEBytesPerRowC_one_row_per_frame[k] = 0;
+               }
+
+               s->PDEAndMetaPTEBytesFrameY = CalculateVMAndRowBytes(
+                       p->myPipe[k].ViewportStationary,
+                       p->myPipe[k].DCCEnable,
+                       p->myPipe[k].DPPPerSurface,
+                       p->myPipe[k].BlockHeight256BytesY,
+                       p->myPipe[k].BlockWidth256BytesY,
+                       p->myPipe[k].SourcePixelFormat,
+                       p->myPipe[k].SurfaceTiling,
+                       p->myPipe[k].BytePerPixelY,
+                       p->myPipe[k].SourceScan,
+                       p->SwathWidthY[k],
+                       p->myPipe[k].ViewportHeight,
+                       p->myPipe[k].ViewportXStart,
+                       p->myPipe[k].ViewportYStart,
+                       p->GPUVMEnable,
+                       p->GPUVMMaxPageTableLevels,
+                       p->GPUVMMinPageSizeKBytes[k],
+                       s->PTEBufferSizeInRequestsForLuma[k],
+                       p->myPipe[k].PitchY,
+                       p->myPipe[k].DCCMetaPitchY,
+                       p->myPipe[k].BlockWidthY,
+                       p->myPipe[k].BlockHeightY,
+
+                       // Output
+                       &s->MetaRowByteY[k],
+                       &s->PixelPTEBytesPerRowY[k],
+                       &s->PixelPTEBytesPerRowStorageY[k],
+                       &p->dpte_row_width_luma_ub[k],
+                       &p->dpte_row_height_luma[k],
+                       &p->dpte_row_height_linear_luma[k],
+                       &s->PixelPTEBytesPerRowY_one_row_per_frame[k],
+                       &s->dpte_row_width_luma_ub_one_row_per_frame[k],
+                       &s->dpte_row_height_luma_one_row_per_frame[k],
+                       &p->meta_req_width[k],
+                       &p->meta_req_height[k],
+                       &p->meta_row_width[k],
+                       &p->meta_row_height[k],
+                       &p->PixelPTEReqWidthY[k],
+                       &p->PixelPTEReqHeightY[k],
+                       &p->PTERequestSizeY[k],
+                       &p->dpde0_bytes_per_frame_ub_l[k],
+                       &p->meta_pte_bytes_per_frame_ub_l[k]);
+
+                       p->PrefetchSourceLinesY[k] = CalculatePrefetchSourceLines(
+                       p->myPipe[k].VRatio,
+                       p->myPipe[k].VTaps,
+                       p->myPipe[k].InterlaceEnable,
+                       p->myPipe[k].ProgressiveToInterlaceUnitInOPP,
+                       p->myPipe[k].SwathHeightY,
+                       p->myPipe[k].SourceScan,
+                       p->myPipe[k].ViewportStationary,
+                       p->SwathWidthY[k],
+                       p->myPipe[k].ViewportHeight,
+                       p->myPipe[k].ViewportXStart,
+                       p->myPipe[k].ViewportYStart,
+
+                       // Output
+                       &p->VInitPreFillY[k],
+                       &p->MaxNumSwathY[k]);
+
+               p->PDEAndMetaPTEBytesFrame[k] = (s->PDEAndMetaPTEBytesFrameY + s->PDEAndMetaPTEBytesFrameC) * (1 + 8 * s->HostVMDynamicLevels);
+               p->MetaRowByte[k] = s->MetaRowByteY[k] + s->MetaRowByteC[k];
+
+               if (s->PixelPTEBytesPerRowStorageY[k] <= 64 * s->PTEBufferSizeInRequestsForLuma[k] && s->PixelPTEBytesPerRowStorageC[k] <= 64 * s->PTEBufferSizeInRequestsForChroma[k]) {
+                       p->PTEBufferSizeNotExceeded[k] = true;
+               } else {
+                       p->PTEBufferSizeNotExceeded[k] = false;
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%u, PixelPTEBytesPerRowY = %u\n", __func__, k, s->PixelPTEBytesPerRowY[k]);
+                       dml_print("DML::%s: k=%u, PixelPTEBytesPerRowC = %u\n", __func__, k, s->PixelPTEBytesPerRowC[k]);
+                       dml_print("DML::%s: k=%u, PixelPTEBytesPerRowStorageY = %u\n", __func__, k, s->PixelPTEBytesPerRowStorageY[k]);
+                       dml_print("DML::%s: k=%u, PixelPTEBytesPerRowStorageC = %u\n", __func__, k, s->PixelPTEBytesPerRowStorageC[k]);
+                       dml_print("DML::%s: k=%u, PTEBufferSizeInRequestsForLuma = %u\n", __func__, k, s->PTEBufferSizeInRequestsForLuma[k]);
+                       dml_print("DML::%s: k=%u, PTEBufferSizeInRequestsForChroma = %u\n", __func__, k, s->PTEBufferSizeInRequestsForChroma[k]);
+                       dml_print("DML::%s: k=%u, PTEBufferSizeNotExceeded          = %u\n",  __func__, k, p->PTEBufferSizeNotExceeded[k]);
+#endif
+               }
+               s->one_row_per_frame_fits_in_buffer[k] = (s->PixelPTEBytesPerRowY_one_row_per_frame[k] <= 64 * 2 * s->PTEBufferSizeInRequestsForLuma[k] &&
+                       s->PixelPTEBytesPerRowC_one_row_per_frame[k] <= 64 * 2 * s->PTEBufferSizeInRequestsForChroma[k]);
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, PDEAndMetaPTEBytesFrame = %u\n", __func__, k, p->PDEAndMetaPTEBytesFrame[k]);
+               dml_print("DML::%s: k=%u, PDEAndMetaPTEBytesFrameY = %u\n", __func__, k, s->PDEAndMetaPTEBytesFrameY);
+               dml_print("DML::%s: k=%u, PDEAndMetaPTEBytesFrameC = %u\n", __func__, k, s->PDEAndMetaPTEBytesFrameC);
+               dml_print("DML::%s: k=%u, HostVMDynamicLevels = %u\n", __func__, k, s->HostVMDynamicLevels);
+               dml_print("DML::%s: k=%u, one_row_per_frame_fits_in_buffer = %u\n", __func__, k, s->one_row_per_frame_fits_in_buffer[k]);
+               dml_print("DML::%s: k=%u, PixelPTEBytesPerRowY_one_row_per_frame = %u\n", __func__, k, s->PixelPTEBytesPerRowY_one_row_per_frame[k]);
+               dml_print("DML::%s: k=%u, PixelPTEBytesPerRowC_one_row_per_frame    = %u\n",  __func__, k, s->PixelPTEBytesPerRowC_one_row_per_frame[k]);
+#endif
+       }
+
+       CalculateMALLUseForStaticScreen(
+               p->NumberOfActiveSurfaces,
+               p->MALLAllocatedForDCN,
+               p->UseMALLForStaticScreen,   // mode
+               p->SurfaceSizeInMALL,
+               s->one_row_per_frame_fits_in_buffer,
+               // Output
+               p->UsesMALLForStaticScreen); // boolen
+
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+               if (p->PTEBufferModeOverrideEn[k] == 1) {
+                       p->PTE_BUFFER_MODE[k] = p->PTEBufferModeOverrideVal[k];
+               }
+               p->PTE_BUFFER_MODE[k] = p->myPipe[k].FORCE_ONE_ROW_FOR_FRAME || p->UsesMALLForStaticScreen[k] || (p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_sub_viewport) ||
+                       (p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe) || (p->GPUVMMinPageSizeKBytes[k] > 64);
+               p->BIGK_FRAGMENT_SIZE[k] = (dml_uint_t)(dml_log2(p->GPUVMMinPageSizeKBytes[k] * 1024) - 12);
+       }
+
+       for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, SurfaceSizeInMALL         = %u\n",  __func__, k, p->SurfaceSizeInMALL[k]);
+               dml_print("DML::%s: k=%u, UsesMALLForStaticScreen   = %u\n",  __func__, k, p->UsesMALLForStaticScreen[k]);
+#endif
+               p->use_one_row_for_frame[k] = p->myPipe[k].FORCE_ONE_ROW_FOR_FRAME || p->UsesMALLForStaticScreen[k] || (p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_sub_viewport) ||
+                                                                       (p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe) || (p->GPUVMMinPageSizeKBytes[k] > 64 && dml_is_vertical_rotation(p->myPipe[k].SourceScan));
+
+               p->use_one_row_for_frame_flip[k] = p->use_one_row_for_frame[k] && !(p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_full_frame);
+
+               if (p->use_one_row_for_frame[k]) {
+                       p->dpte_row_height_luma[k] = s->dpte_row_height_luma_one_row_per_frame[k];
+                       p->dpte_row_width_luma_ub[k] = s->dpte_row_width_luma_ub_one_row_per_frame[k];
+                       s->PixelPTEBytesPerRowY[k] = s->PixelPTEBytesPerRowY_one_row_per_frame[k];
+                       p->dpte_row_height_chroma[k] = s->dpte_row_height_chroma_one_row_per_frame[k];
+                       p->dpte_row_width_chroma_ub[k] = s->dpte_row_width_chroma_ub_one_row_per_frame[k];
+                       s->PixelPTEBytesPerRowC[k] = s->PixelPTEBytesPerRowC_one_row_per_frame[k];
+                       p->PTEBufferSizeNotExceeded[k] = s->one_row_per_frame_fits_in_buffer[k];
+               }
+
+               if (p->MetaRowByte[k] <= p->DCCMetaBufferSizeBytes) {
+                       p->DCCMetaBufferSizeNotExceeded[k] = true;
+               } else {
+                       p->DCCMetaBufferSizeNotExceeded[k] = false;
+
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%u, MetaRowByte                   = %u\n",  __func__, k, p->MetaRowByte[k]);
+                       dml_print("DML::%s: k=%u, DCCMetaBufferSizeBytes        = %u\n",  __func__, k, p->DCCMetaBufferSizeBytes);
+                       dml_print("DML::%s: k=%u, DCCMetaBufferSizeNotExceeded  = %u\n",  __func__, k, p->DCCMetaBufferSizeNotExceeded[k]);
+#endif
+               }
+               s->PixelPTEBytesPerRowY[k] = s->PixelPTEBytesPerRowY[k] * (1 + 8 * s->HostVMDynamicLevels);
+               s->PixelPTEBytesPerRowC[k] = s->PixelPTEBytesPerRowC[k] * (1 + 8 * s->HostVMDynamicLevels);
+               p->PixelPTEBytesPerRow[k] = s->PixelPTEBytesPerRowY[k] + s->PixelPTEBytesPerRowC[k];
+               if (p->use_one_row_for_frame[k])
+                       p->PixelPTEBytesPerRow[k] = p->PixelPTEBytesPerRow[k] / 2;
+
+               CalculateRowBandwidth(
+                       p->GPUVMEnable,
+                       p->myPipe[k].SourcePixelFormat,
+                       p->myPipe[k].VRatio,
+                       p->myPipe[k].VRatioChroma,
+                       p->myPipe[k].DCCEnable,
+                       p->myPipe[k].HTotal / p->myPipe[k].PixelClock,
+                       s->MetaRowByteY[k],
+                       s->MetaRowByteC[k],
+                       p->meta_row_height[k],
+                       p->meta_row_height_chroma[k],
+                       s->PixelPTEBytesPerRowY[k],
+                       s->PixelPTEBytesPerRowC[k],
+                       p->dpte_row_height_luma[k],
+                       p->dpte_row_height_chroma[k],
+
+                       // Output
+                       &p->meta_row_bw[k],
+                       &p->dpte_row_bw[k]);
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, use_one_row_for_frame = %u\n", __func__, k, p->use_one_row_for_frame[k]);
+               dml_print("DML::%s: k=%u, use_one_row_for_frame_flip = %u\n", __func__, k, p->use_one_row_for_frame_flip[k]);
+               dml_print("DML::%s: k=%u, UseMALLForPStateChange = %u\n", __func__, k, p->UseMALLForPStateChange[k]);
+               dml_print("DML::%s: k=%u, dpte_row_height_luma = %u\n", __func__, k, p->dpte_row_height_luma[k]);
+               dml_print("DML::%s: k=%u, dpte_row_width_luma_ub = %u\n", __func__, k, p->dpte_row_width_luma_ub[k]);
+               dml_print("DML::%s: k=%u, PixelPTEBytesPerRowY = %u\n", __func__, k, s->PixelPTEBytesPerRowY[k]);
+               dml_print("DML::%s: k=%u, dpte_row_height_chroma = %u\n", __func__, k, p->dpte_row_height_chroma[k]);
+               dml_print("DML::%s: k=%u, dpte_row_width_chroma_ub = %u\n", __func__, k, p->dpte_row_width_chroma_ub[k]);
+               dml_print("DML::%s: k=%u, PixelPTEBytesPerRowC = %u\n", __func__, k, s->PixelPTEBytesPerRowC[k]);
+               dml_print("DML::%s: k=%u, PixelPTEBytesPerRow = %u\n", __func__, k, p->PixelPTEBytesPerRow[k]);
+               dml_print("DML::%s: k=%u, PTEBufferSizeNotExceeded = %u\n", __func__, k, p->PTEBufferSizeNotExceeded[k]);
+               dml_print("DML::%s: k=%u, PTE_BUFFER_MODE = %u\n", __func__, k, p->PTE_BUFFER_MODE[k]);
+               dml_print("DML::%s: k=%u, BIGK_FRAGMENT_SIZE     = %u\n", __func__, k, p->BIGK_FRAGMENT_SIZE[k]);
+#endif
+       }
+}
+
+static void CalculateOutputLink(
+               dml_float_t PHYCLKPerState,
+               dml_float_t PHYCLKD18PerState,
+               dml_float_t PHYCLKD32PerState,
+               dml_float_t Downspreading,
+               dml_bool_t IsMainSurfaceUsingTheIndicatedTiming,
+               enum dml_output_encoder_class Output,
+               enum dml_output_format_class OutputFormat,
+               dml_uint_t HTotal,
+               dml_uint_t HActive,
+               dml_float_t PixelClockBackEnd,
+               dml_float_t ForcedOutputLinkBPP,
+               dml_uint_t DSCInputBitPerComponent,
+               dml_uint_t NumberOfDSCSlices,
+               dml_float_t AudioSampleRate,
+               dml_uint_t AudioSampleLayout,
+               enum dml_odm_mode ODMModeNoDSC,
+               enum dml_odm_mode ODMModeDSC,
+               enum dml_dsc_enable DSCEnable,
+               dml_uint_t OutputLinkDPLanes,
+               enum dml_output_link_dp_rate OutputLinkDPRate,
+
+               // Output
+               dml_bool_t *RequiresDSC,
+               dml_bool_t *RequiresFEC,
+               dml_float_t *OutBpp,
+               enum dml_output_type_and_rate__type *OutputType,
+               enum dml_output_type_and_rate__rate *OutputRate,
+               dml_uint_t *RequiredSlots)
+{
+       dml_bool_t LinkDSCEnable;
+       dml_uint_t dummy;
+       *RequiresDSC = false;
+       *RequiresFEC = false;
+       *OutBpp = 0;
+
+       *OutputType = dml_output_type_unknown;
+       *OutputRate = dml_output_rate_unknown;
+
+       if (IsMainSurfaceUsingTheIndicatedTiming) {
+               if (Output == dml_hdmi) {
+                       *RequiresDSC = false;
+                       *RequiresFEC = false;
+                       *OutBpp = TruncToValidBPP(dml_min(600, PHYCLKPerState) * 10, 3, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, false, Output,
+                                                                       OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, &dummy);
+                       //OutputTypeAndRate = "HDMI";
+                       *OutputType = dml_output_type_hdmi;
+
+               } else if (Output == dml_dp || Output == dml_dp2p0 || Output == dml_edp) {
+                       if (DSCEnable == dml_dsc_enable) {
+                               *RequiresDSC = true;
+                               LinkDSCEnable = true;
+                               if (Output == dml_dp || Output == dml_dp2p0) {
+                                       *RequiresFEC = true;
+                               } else {
+                                       *RequiresFEC = false;
+                               }
+                       } else {
+                               *RequiresDSC = false;
+                               LinkDSCEnable = false;
+                               if (Output == dml_dp2p0) {
+                                       *RequiresFEC = true;
+                               } else {
+                                       *RequiresFEC = false;
+                               }
+                       }
+                       if (Output == dml_dp2p0) {
+                               *OutBpp = 0;
+                               if ((OutputLinkDPRate == dml_dp_rate_na || OutputLinkDPRate == dml_dp_rate_uhbr10) && PHYCLKD32PerState >= 10000 / 32) {
+                                       *OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 10000, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+                                                                                               OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+                                       if (*OutBpp == 0 && PHYCLKD32PerState < 13500 / 32 && DSCEnable == dml_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+                                               *RequiresDSC = true;
+                                               LinkDSCEnable = true;
+                                               *OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 10000, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+                                                                                                       OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+                                       }
+                                       //OutputTypeAndRate = Output & " UHBR10";
+                                       *OutputType = dml_output_type_dp2p0;
+                                       *OutputRate = dml_output_rate_dp_rate_uhbr10;
+                               }
+                               if ((OutputLinkDPRate == dml_dp_rate_na || OutputLinkDPRate == dml_dp_rate_uhbr13p5) && *OutBpp == 0 && PHYCLKD32PerState >= 13500 / 32) {
+                                       *OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 13500, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+                                                                                               OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+
+                                       if (OutBpp == 0 && PHYCLKD32PerState < 20000 / 32 && DSCEnable == dml_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+                                               *RequiresDSC = true;
+                                               LinkDSCEnable = true;
+                                               *OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 13500, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+                                                                                               OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+                                       }
+                                       //OutputTypeAndRate = Output & " UHBR13p5";
+                                       *OutputType = dml_output_type_dp2p0;
+                                       *OutputRate = dml_output_rate_dp_rate_uhbr13p5;
+                               }
+                               if ((OutputLinkDPRate == dml_dp_rate_na || OutputLinkDPRate == dml_dp_rate_uhbr20) && *OutBpp == 0 && PHYCLKD32PerState >= 20000 / 32) {
+                                       *OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 20000, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+                                                                                       OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+                                       if (*OutBpp == 0 && DSCEnable == dml_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+                                               *RequiresDSC = true;
+                                               LinkDSCEnable = true;
+                                               *OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 20000, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+                                                                                               OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+                                       }
+                                       //OutputTypeAndRate = Output & " UHBR20";
+                                       *OutputType = dml_output_type_dp2p0;
+                                       *OutputRate = dml_output_rate_dp_rate_uhbr20;
+                               }
+                       } else { // output is dp or edp
+                               *OutBpp = 0;
+                               if ((OutputLinkDPRate == dml_dp_rate_na || OutputLinkDPRate == dml_dp_rate_hbr) && PHYCLKPerState >= 270) {
+                                       *OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 2700, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+                                                                                       OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+                                       if (*OutBpp == 0 && PHYCLKPerState < 540 && DSCEnable == dml_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+                                               *RequiresDSC = true;
+                                               LinkDSCEnable = true;
+                                               if (Output == dml_dp) {
+                                                       *RequiresFEC = true;
+                                               }
+                                               *OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 2700, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+                                                                                               OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+                                       }
+                                       //OutputTypeAndRate = Output & " HBR";
+                                       *OutputType = (Output == dml_dp) ? dml_output_type_dp : dml_output_type_edp;
+                                       *OutputRate = dml_output_rate_dp_rate_hbr;
+                               }
+                               if ((OutputLinkDPRate == dml_dp_rate_na || OutputLinkDPRate == dml_dp_rate_hbr2) && *OutBpp == 0 && PHYCLKPerState >= 540) {
+                                       *OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 5400, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+                                                                                       OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+
+                                       if (*OutBpp == 0 && PHYCLKPerState < 810 && DSCEnable == dml_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+                                               *RequiresDSC = true;
+                                               LinkDSCEnable = true;
+                                               if (Output == dml_dp) {
+                                                       *RequiresFEC = true;
+                                               }
+                                               *OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 5400, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+                                                                                               OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+                                       }
+                                       //OutputTypeAndRate = Output & " HBR2";
+                                       *OutputType = (Output == dml_dp) ? dml_output_type_dp : dml_output_type_edp;
+                                       *OutputRate = dml_output_rate_dp_rate_hbr2;
+                               }
+                               if ((OutputLinkDPRate == dml_dp_rate_na || OutputLinkDPRate == dml_dp_rate_hbr3) && *OutBpp == 0 && PHYCLKPerState >= 810) { // VBA_ERROR, vba code doesn't have hbr3 check
+                                       *OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 8100, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+                                                                                       OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+
+                                       if (*OutBpp == 0 && DSCEnable == dml_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+                                               *RequiresDSC = true;
+                                               LinkDSCEnable = true;
+                                               if (Output == dml_dp) {
+                                                       *RequiresFEC = true;
+                                               }
+                                               *OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 8100, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+                                                                                               OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+                                       }
+                                       //OutputTypeAndRate = Output & " HBR3";
+                                       *OutputType = (Output == dml_dp) ? dml_output_type_dp : dml_output_type_edp;
+                                       *OutputRate = dml_output_rate_dp_rate_hbr3;
+                               }
+                       }
+               }
+       }
+}
+
+/// @brief Determine the ODM mode and number of DPP used per plane based on dispclk, dsc usage, odm usage policy
+static void CalculateODMMode(
+               dml_uint_t MaximumPixelsPerLinePerDSCUnit,
+               dml_uint_t HActive,
+               enum dml_output_encoder_class Output,
+               enum dml_output_format_class OutputFormat,
+               enum dml_odm_use_policy ODMUse,
+               dml_float_t StateDispclk,
+               dml_float_t MaxDispclk,
+               dml_bool_t DSCEnable,
+               dml_uint_t TotalNumberOfActiveDPP,
+               dml_uint_t MaxNumDPP,
+               dml_float_t PixelClock,
+               dml_float_t DISPCLKDPPCLKDSCCLKDownSpreading,
+               dml_float_t DISPCLKRampingMargin,
+               dml_float_t DISPCLKDPPCLKVCOSpeed,
+
+               // Output
+               dml_bool_t *TotalAvailablePipesSupport,
+               dml_uint_t *NumberOfDPP,
+               enum dml_odm_mode *ODMMode,
+               dml_float_t *RequiredDISPCLKPerSurface)
+{
+
+       dml_float_t SurfaceRequiredDISPCLKWithoutODMCombine;
+       dml_float_t SurfaceRequiredDISPCLKWithODMCombineTwoToOne;
+       dml_float_t SurfaceRequiredDISPCLKWithODMCombineFourToOne;
+
+       SurfaceRequiredDISPCLKWithoutODMCombine = CalculateRequiredDispclk(dml_odm_mode_bypass, PixelClock, DISPCLKDPPCLKDSCCLKDownSpreading, DISPCLKRampingMargin, DISPCLKDPPCLKVCOSpeed, MaxDispclk);
+       SurfaceRequiredDISPCLKWithODMCombineTwoToOne = CalculateRequiredDispclk(dml_odm_mode_combine_2to1, PixelClock, DISPCLKDPPCLKDSCCLKDownSpreading, DISPCLKRampingMargin, DISPCLKDPPCLKVCOSpeed, MaxDispclk);
+       SurfaceRequiredDISPCLKWithODMCombineFourToOne = CalculateRequiredDispclk(dml_odm_mode_combine_4to1, PixelClock, DISPCLKDPPCLKDSCCLKDownSpreading, DISPCLKRampingMargin, DISPCLKDPPCLKVCOSpeed, MaxDispclk);
+       *TotalAvailablePipesSupport = true;
+
+       if (OutputFormat == dml_420) {
+               if (HActive > 4 * DML2_MAX_FMT_420_BUFFER_WIDTH)
+                       *TotalAvailablePipesSupport = false;
+               else if (HActive > 2 * DML2_MAX_FMT_420_BUFFER_WIDTH)
+                       ODMUse = dml_odm_use_policy_combine_4to1;
+               else if (HActive > DML2_MAX_FMT_420_BUFFER_WIDTH)
+                       ODMUse = dml_odm_use_policy_combine_2to1;
+               if (Output == dml_hdmi && ODMUse == dml_odm_use_policy_combine_2to1)
+                       *TotalAvailablePipesSupport = false;
+               if ((Output == dml_hdmi || Output == dml_dp || Output == dml_edp) && ODMUse == dml_odm_use_policy_combine_4to1)
+                       *TotalAvailablePipesSupport = false;
+       }
+
+       if (ODMUse == dml_odm_use_policy_bypass || ODMUse == dml_odm_use_policy_combine_as_needed)
+               *ODMMode = dml_odm_mode_bypass;
+       else if (ODMUse == dml_odm_use_policy_combine_2to1)
+               *ODMMode = dml_odm_mode_combine_2to1;
+       else if (ODMUse == dml_odm_use_policy_combine_4to1)
+               *ODMMode = dml_odm_mode_combine_4to1;
+       else if (ODMUse == dml_odm_use_policy_split_1to2)
+               *ODMMode = dml_odm_mode_split_1to2;
+       else if (ODMUse == dml_odm_use_policy_mso_1to2)
+               *ODMMode = dml_odm_mode_mso_1to2;
+       else if (ODMUse == dml_odm_use_policy_mso_1to4)
+               *ODMMode = dml_odm_mode_mso_1to4;
+
+       *RequiredDISPCLKPerSurface = SurfaceRequiredDISPCLKWithoutODMCombine;
+       *NumberOfDPP = 0;
+
+       if (!(Output == dml_hdmi || Output == dml_dp || Output == dml_edp) && (ODMUse == dml_odm_use_policy_combine_4to1 || (ODMUse == dml_odm_use_policy_combine_as_needed &&
+               (SurfaceRequiredDISPCLKWithODMCombineTwoToOne > StateDispclk || (DSCEnable && (HActive > 2 * MaximumPixelsPerLinePerDSCUnit)))))) {
+               if (TotalNumberOfActiveDPP + 4 <= MaxNumDPP) {
+                       *ODMMode = dml_odm_mode_combine_4to1;
+                       *RequiredDISPCLKPerSurface = SurfaceRequiredDISPCLKWithODMCombineFourToOne;
+                       *NumberOfDPP = 4;
+               } else {
+                       *TotalAvailablePipesSupport = false;
+               }
+       } else if (Output != dml_hdmi && (ODMUse == dml_odm_use_policy_combine_2to1 || (ODMUse == dml_odm_use_policy_combine_as_needed &&
+                               ((SurfaceRequiredDISPCLKWithoutODMCombine > StateDispclk && SurfaceRequiredDISPCLKWithODMCombineTwoToOne <= StateDispclk) ||
+                               (DSCEnable && (HActive > MaximumPixelsPerLinePerDSCUnit)))))) {
+               if (TotalNumberOfActiveDPP + 2 <= MaxNumDPP) {
+                       *ODMMode = dml_odm_mode_combine_2to1;
+                       *RequiredDISPCLKPerSurface = SurfaceRequiredDISPCLKWithODMCombineTwoToOne;
+                       *NumberOfDPP = 2;
+               } else {
+                       *TotalAvailablePipesSupport = false;
+               }
+       } else {
+               if (TotalNumberOfActiveDPP + 1 <= MaxNumDPP) {
+                       *NumberOfDPP = 1;
+               } else {
+                       *TotalAvailablePipesSupport = false;
+               }
+       }
+}
+
+/// @brief Calculate the required DISPCLK given the odm mode and pixclk
+static dml_float_t CalculateRequiredDispclk(
+               enum dml_odm_mode ODMMode,
+               dml_float_t PixelClock,
+               dml_float_t DISPCLKDPPCLKDSCCLKDownSpreading,
+               dml_float_t DISPCLKRampingMargin,
+               dml_float_t DISPCLKDPPCLKVCOSpeed,
+               dml_float_t MaxDispclk)
+{
+       dml_float_t RequiredDispclk = 0.;
+       dml_float_t PixelClockAfterODM;
+
+       dml_float_t DISPCLKWithRampingRoundedToDFSGranularity;
+       dml_float_t DISPCLKWithoutRampingRoundedToDFSGranularity;
+       dml_float_t MaxDispclkRoundedDownToDFSGranularity;
+
+       if (ODMMode == dml_odm_mode_combine_4to1) {
+               PixelClockAfterODM = PixelClock / 4;
+       } else if (ODMMode == dml_odm_mode_combine_2to1) {
+               PixelClockAfterODM = PixelClock / 2;
+       } else {
+               PixelClockAfterODM = PixelClock;
+       }
+
+       DISPCLKWithRampingRoundedToDFSGranularity = RoundToDFSGranularity(PixelClockAfterODM * (1.0 + DISPCLKDPPCLKDSCCLKDownSpreading / 100.0) * (1 + DISPCLKRampingMargin / 100.0), 1, DISPCLKDPPCLKVCOSpeed);
+       DISPCLKWithoutRampingRoundedToDFSGranularity = RoundToDFSGranularity(PixelClockAfterODM * (1.0 + DISPCLKDPPCLKDSCCLKDownSpreading / 100.0), 1, DISPCLKDPPCLKVCOSpeed);
+       MaxDispclkRoundedDownToDFSGranularity = RoundToDFSGranularity(MaxDispclk, 0, DISPCLKDPPCLKVCOSpeed);
+
+       if (DISPCLKWithoutRampingRoundedToDFSGranularity > MaxDispclkRoundedDownToDFSGranularity) {
+               RequiredDispclk = DISPCLKWithoutRampingRoundedToDFSGranularity;
+       } else if (DISPCLKWithRampingRoundedToDFSGranularity > MaxDispclkRoundedDownToDFSGranularity) {
+               RequiredDispclk = MaxDispclkRoundedDownToDFSGranularity;
+       } else {
+               RequiredDispclk = DISPCLKWithRampingRoundedToDFSGranularity;
+       }
+
+       return RequiredDispclk;
+}
+
+/// @brief Determine DPPCLK if there only one DPP per plane, main factor is the pixel rate and DPP scaling parameter
+static void CalculateSinglePipeDPPCLKAndSCLThroughput(
+               dml_float_t HRatio,
+               dml_float_t HRatioChroma,
+               dml_float_t VRatio,
+               dml_float_t VRatioChroma,
+               dml_float_t MaxDCHUBToPSCLThroughput,
+               dml_float_t MaxPSCLToLBThroughput,
+               dml_float_t PixelClock,
+               enum dml_source_format_class SourcePixelFormat,
+               dml_uint_t HTaps,
+               dml_uint_t HTapsChroma,
+               dml_uint_t VTaps,
+               dml_uint_t VTapsChroma,
+
+               // Output
+               dml_float_t *PSCL_THROUGHPUT,
+               dml_float_t *PSCL_THROUGHPUT_CHROMA,
+               dml_float_t *DPPCLKUsingSingleDPP)
+{
+       dml_float_t DPPCLKUsingSingleDPPLuma;
+       dml_float_t DPPCLKUsingSingleDPPChroma;
+
+       if (HRatio > 1) {
+               *PSCL_THROUGHPUT = dml_min(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput * HRatio / dml_ceil((dml_float_t) HTaps / 6.0, 1.0));
+       } else {
+               *PSCL_THROUGHPUT = dml_min(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput);
+       }
+
+       DPPCLKUsingSingleDPPLuma = PixelClock * dml_max3(VTaps / 6 * dml_min(1, HRatio), HRatio * VRatio / *PSCL_THROUGHPUT, 1);
+
+       if ((HTaps > 6 || VTaps > 6) && DPPCLKUsingSingleDPPLuma < 2 * PixelClock)
+               DPPCLKUsingSingleDPPLuma = 2 * PixelClock;
+
+       if ((SourcePixelFormat != dml_420_8 && SourcePixelFormat != dml_420_10 && SourcePixelFormat != dml_420_12 && SourcePixelFormat != dml_rgbe_alpha)) {
+               *PSCL_THROUGHPUT_CHROMA = 0;
+               *DPPCLKUsingSingleDPP = DPPCLKUsingSingleDPPLuma;
+       } else {
+               if (HRatioChroma > 1) {
+                       *PSCL_THROUGHPUT_CHROMA = dml_min(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput * HRatioChroma / dml_ceil((dml_float_t) HTapsChroma / 6.0, 1.0));
+               } else {
+                       *PSCL_THROUGHPUT_CHROMA = dml_min(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput);
+               }
+               DPPCLKUsingSingleDPPChroma = PixelClock * dml_max3(VTapsChroma / 6 * dml_min(1, HRatioChroma),
+                                                                                                                       HRatioChroma * VRatioChroma / *PSCL_THROUGHPUT_CHROMA, 1);
+               if ((HTapsChroma > 6 || VTapsChroma > 6) && DPPCLKUsingSingleDPPChroma < 2 * PixelClock)
+                       DPPCLKUsingSingleDPPChroma = 2 * PixelClock;
+               *DPPCLKUsingSingleDPP = dml_max(DPPCLKUsingSingleDPPLuma, DPPCLKUsingSingleDPPChroma);
+       }
+}
+
+/// @brief Calculate the actual dppclk freq
+/// @param DPPCLKUsingSingleDPP DppClk freq required if there is only 1 DPP per plane
+/// @param DPPPerSurface Number of DPP for each plane
+static void CalculateDPPCLK(
+               dml_uint_t NumberOfActiveSurfaces,
+               dml_float_t DISPCLKDPPCLKDSCCLKDownSpreading,
+               dml_float_t DISPCLKDPPCLKVCOSpeed,
+               dml_float_t DPPCLKUsingSingleDPP[],
+               dml_uint_t DPPPerSurface[],
+
+               // Output
+               dml_float_t *GlobalDPPCLK,
+               dml_float_t Dppclk[])
+{
+       *GlobalDPPCLK = 0;
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               Dppclk[k] = DPPCLKUsingSingleDPP[k] / DPPPerSurface[k] * (1 + DISPCLKDPPCLKDSCCLKDownSpreading / 100.0);
+               *GlobalDPPCLK = dml_max(*GlobalDPPCLK, Dppclk[k]);
+       }
+       *GlobalDPPCLK = RoundToDFSGranularity(*GlobalDPPCLK, 1, DISPCLKDPPCLKVCOSpeed);
+
+       dml_print("DML::%s: GlobalDPPCLK = %f\n", __func__, *GlobalDPPCLK);
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               Dppclk[k] = *GlobalDPPCLK / 255.0 * dml_ceil(Dppclk[k] * 255.0 / *GlobalDPPCLK, 1.0);
+               dml_print("DML::%s: Dppclk[%0d] = %f\n", __func__, k, Dppclk[k]);
+       }
+}
+
+static void CalculateMALLUseForStaticScreen(
+               dml_uint_t NumberOfActiveSurfaces,
+               dml_uint_t MALLAllocatedForDCNFinal,
+               enum dml_use_mall_for_static_screen_mode *UseMALLForStaticScreen,
+               dml_uint_t SurfaceSizeInMALL[],
+               dml_bool_t one_row_per_frame_fits_in_buffer[],
+
+               // Output
+               dml_bool_t UsesMALLForStaticScreen[])
+{
+
+       dml_uint_t SurfaceToAddToMALL;
+       dml_bool_t CanAddAnotherSurfaceToMALL;
+       dml_uint_t TotalSurfaceSizeInMALL;
+
+       TotalSurfaceSizeInMALL = 0;
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               UsesMALLForStaticScreen[k] = (UseMALLForStaticScreen[k] == dml_use_mall_static_screen_enable);
+               if (UsesMALLForStaticScreen[k])
+                       TotalSurfaceSizeInMALL = TotalSurfaceSizeInMALL + SurfaceSizeInMALL[k];
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, UsesMALLForStaticScreen = %u\n",  __func__, k, UsesMALLForStaticScreen[k]);
+               dml_print("DML::%s: k=%u, TotalSurfaceSizeInMALL = %u\n",  __func__, k, TotalSurfaceSizeInMALL);
+#endif
+       }
+
+       SurfaceToAddToMALL = 0;
+       CanAddAnotherSurfaceToMALL = true;
+       while (CanAddAnotherSurfaceToMALL) {
+               CanAddAnotherSurfaceToMALL = false;
+               for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+                       if (TotalSurfaceSizeInMALL + SurfaceSizeInMALL[k] <= MALLAllocatedForDCNFinal * 1024 * 1024 &&
+                               !UsesMALLForStaticScreen[k] && UseMALLForStaticScreen[k] != dml_use_mall_static_screen_disable && one_row_per_frame_fits_in_buffer[k] &&
+                               (!CanAddAnotherSurfaceToMALL || SurfaceSizeInMALL[k] < SurfaceSizeInMALL[SurfaceToAddToMALL])) {
+                               CanAddAnotherSurfaceToMALL = true;
+                               SurfaceToAddToMALL = k;
+                               dml_print("DML::%s: k=%u, UseMALLForStaticScreen = %u (dis, en, optimize)\n",  __func__, k, UseMALLForStaticScreen[k]);
+                       }
+               }
+               if (CanAddAnotherSurfaceToMALL) {
+                       UsesMALLForStaticScreen[SurfaceToAddToMALL] = true;
+                       TotalSurfaceSizeInMALL = TotalSurfaceSizeInMALL + SurfaceSizeInMALL[SurfaceToAddToMALL];
+
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: SurfaceToAddToMALL       = %u\n",  __func__, SurfaceToAddToMALL);
+                       dml_print("DML::%s: TotalSurfaceSizeInMALL   = %u\n",  __func__, TotalSurfaceSizeInMALL);
+#endif
+               }
+       }
+}
+
+// @brief Calculate return bw for VM only traffic
+dml_float_t dml_get_return_bw_mbps_vm_only(
+                                                               const struct soc_bounding_box_st *soc,
+                                                               dml_bool_t use_ideal_dram_bw_strobe,
+                                                               dml_bool_t HostVMEnable,
+                                                               dml_float_t DCFCLK,
+                                                               dml_float_t FabricClock,
+                                                               dml_float_t DRAMSpeed)
+{
+       dml_float_t VMDataOnlyReturnBW =
+               dml_min3(soc->return_bus_width_bytes * DCFCLK * soc->pct_ideal_sdp_bw_after_urgent / 100.0,
+                               FabricClock * soc->fabric_datapath_to_dcn_data_return_bytes * soc->pct_ideal_sdp_bw_after_urgent / 100.0,
+                               DRAMSpeed * soc->num_chans *  soc->dram_channel_width_bytes *
+                               ((use_ideal_dram_bw_strobe && !HostVMEnable) ? soc->pct_ideal_dram_bw_after_urgent_strobe :  soc->pct_ideal_dram_bw_after_urgent_vm_only) / 100.0);
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: use_ideal_dram_bw_strobe = %u\n", __func__, use_ideal_dram_bw_strobe);
+       dml_print("DML::%s: HostVMEnable = %u\n", __func__, HostVMEnable);
+       dml_print("DML::%s: DCFCLK = %f\n", __func__, DCFCLK);
+       dml_print("DML::%s: FabricClock = %f\n", __func__, FabricClock);
+       dml_print("DML::%s: DRAMSpeed = %f\n", __func__, DRAMSpeed);
+       dml_print("DML::%s: VMDataOnlyReturnBW = %f\n", __func__, VMDataOnlyReturnBW);
+#endif
+       return VMDataOnlyReturnBW;
+}
+
+// Function: dml_get_return_bw_mbps
+// Megabyte per second
+dml_float_t dml_get_return_bw_mbps(
+                                               const struct soc_bounding_box_st *soc,
+                                               dml_bool_t use_ideal_dram_bw_strobe,
+                                               dml_bool_t HostVMEnable,
+                                               dml_float_t DCFCLK,
+                                               dml_float_t FabricClock,
+                                               dml_float_t DRAMSpeed)
+{
+       dml_float_t ReturnBW = 0.;
+       dml_float_t IdealSDPPortBandwidth    = soc->return_bus_width_bytes * DCFCLK;
+       dml_float_t IdealFabricBandwidth     = FabricClock * soc->fabric_datapath_to_dcn_data_return_bytes;
+       dml_float_t IdealDRAMBandwidth       = DRAMSpeed * soc->num_chans * soc->dram_channel_width_bytes;
+       dml_float_t PixelDataOnlyReturnBW    = dml_min3(IdealSDPPortBandwidth * soc->pct_ideal_sdp_bw_after_urgent / 100,
+                                                                                               IdealFabricBandwidth * soc->pct_ideal_fabric_bw_after_urgent / 100,
+                                                                                               IdealDRAMBandwidth * ((use_ideal_dram_bw_strobe && !HostVMEnable) ? soc->pct_ideal_dram_bw_after_urgent_strobe :
+                                                                                                                                                                               soc->pct_ideal_dram_bw_after_urgent_pixel_only) / 100);
+       dml_float_t PixelMixedWithVMDataReturnBW = dml_min3(IdealSDPPortBandwidth * soc->pct_ideal_sdp_bw_after_urgent / 100,
+                                                                                                       IdealFabricBandwidth * soc->pct_ideal_fabric_bw_after_urgent / 100,
+                                                                                                       IdealDRAMBandwidth * ((use_ideal_dram_bw_strobe && !HostVMEnable) ? soc->pct_ideal_dram_bw_after_urgent_strobe :
+                                                                                                                                                                                       soc->pct_ideal_dram_bw_after_urgent_pixel_and_vm) / 100);
+
+       if (HostVMEnable != true) {
+               ReturnBW = PixelDataOnlyReturnBW;
+       } else {
+               ReturnBW = PixelMixedWithVMDataReturnBW;
+       }
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: use_ideal_dram_bw_strobe = %u\n", __func__, use_ideal_dram_bw_strobe);
+       dml_print("DML::%s: HostVMEnable = %u\n", __func__, HostVMEnable);
+       dml_print("DML::%s: DCFCLK = %f\n", __func__, DCFCLK);
+       dml_print("DML::%s: FabricClock = %f\n", __func__, FabricClock);
+       dml_print("DML::%s: DRAMSpeed = %f\n", __func__, DRAMSpeed);
+       dml_print("DML::%s: IdealSDPPortBandwidth = %f\n", __func__, IdealSDPPortBandwidth);
+       dml_print("DML::%s: IdealFabricBandwidth = %f\n", __func__, IdealFabricBandwidth);
+       dml_print("DML::%s: IdealDRAMBandwidth = %f\n", __func__, IdealDRAMBandwidth);
+       dml_print("DML::%s: PixelDataOnlyReturnBW = %f\n", __func__, PixelDataOnlyReturnBW);
+       dml_print("DML::%s: PixelMixedWithVMDataReturnBW = %f\n", __func__, PixelMixedWithVMDataReturnBW);
+       dml_print("DML::%s: ReturnBW                  = %f MBps\n", __func__, ReturnBW);
+#endif
+       return ReturnBW;
+}
+
+// Function: dml_get_return_dram_bw_mbps
+// Megabyte per second
+static dml_float_t dml_get_return_dram_bw_mbps(
+                                               const struct soc_bounding_box_st *soc,
+                                               dml_bool_t use_ideal_dram_bw_strobe,
+                                               dml_bool_t HostVMEnable,
+                                               dml_float_t DRAMSpeed)
+{
+       dml_float_t ReturnDRAMBW = 0.;
+       dml_float_t IdealDRAMBandwidth       = DRAMSpeed * soc->num_chans * soc->dram_channel_width_bytes;
+       dml_float_t PixelDataOnlyReturnBW    = IdealDRAMBandwidth * ((use_ideal_dram_bw_strobe && !HostVMEnable) ? soc->pct_ideal_dram_bw_after_urgent_strobe :
+                                                                                                                                                                               soc->pct_ideal_dram_bw_after_urgent_pixel_only) / 100;
+       dml_float_t PixelMixedWithVMDataReturnBW =  IdealDRAMBandwidth * ((use_ideal_dram_bw_strobe && !HostVMEnable) ? soc->pct_ideal_dram_bw_after_urgent_strobe :
+                                                                                                                                                                                       soc->pct_ideal_dram_bw_after_urgent_pixel_and_vm) / 100;
+
+       if (HostVMEnable != true) {
+               ReturnDRAMBW = PixelDataOnlyReturnBW;
+       } else {
+               ReturnDRAMBW = PixelMixedWithVMDataReturnBW;
+       }
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: use_ideal_dram_bw_strobe = %u\n", __func__, use_ideal_dram_bw_strobe);
+       dml_print("DML::%s: HostVMEnable = %u\n", __func__, HostVMEnable);
+       dml_print("DML::%s: DRAMSpeed = %f\n", __func__, DRAMSpeed);
+       dml_print("DML::%s: IdealDRAMBandwidth = %f\n", __func__, IdealDRAMBandwidth);
+       dml_print("DML::%s: PixelDataOnlyReturnBW = %f\n", __func__, PixelDataOnlyReturnBW);
+       dml_print("DML::%s: PixelMixedWithVMDataReturnBW = %f\n", __func__, PixelMixedWithVMDataReturnBW);
+       dml_print("DML::%s: ReturnDRAMBW                     = %f MBps\n", __func__, ReturnDRAMBW);
+#endif
+       return ReturnDRAMBW;
+}
+
+/// @brief BACKEND
+static dml_uint_t DSCDelayRequirement(
+                                               dml_bool_t DSCEnabled,
+                                               enum dml_odm_mode ODMMode,
+                                               dml_uint_t DSCInputBitPerComponent,
+                                               dml_float_t OutputBpp,
+                                               dml_uint_t HActive,
+                                               dml_uint_t HTotal,
+                                               dml_uint_t NumberOfDSCSlices,
+                                               enum dml_output_format_class OutputFormat,
+                                               enum dml_output_encoder_class Output,
+                                               dml_float_t PixelClock,
+                                               dml_float_t PixelClockBackEnd)
+{
+       dml_uint_t DSCDelayRequirement_val = 0;
+
+       if (DSCEnabled == true && OutputBpp != 0) {
+               if (ODMMode == dml_odm_mode_combine_4to1) {
+                       DSCDelayRequirement_val = 4 * (dscceComputeDelay(DSCInputBitPerComponent, OutputBpp, (dml_uint_t)(dml_ceil((dml_float_t) HActive / (dml_float_t) NumberOfDSCSlices, 1.0)),
+                                                                                               (dml_uint_t) (NumberOfDSCSlices / 4.0), OutputFormat, Output) + dscComputeDelay(OutputFormat, Output));
+               } else if (ODMMode == dml_odm_mode_combine_2to1) {
+                       DSCDelayRequirement_val = 2 * (dscceComputeDelay(DSCInputBitPerComponent, OutputBpp, (dml_uint_t)(dml_ceil((dml_float_t) HActive / (dml_float_t) NumberOfDSCSlices, 1.0)),
+                                                                                               (dml_uint_t) (NumberOfDSCSlices / 2.0), OutputFormat, Output) + dscComputeDelay(OutputFormat, Output));
+               } else {
+                       DSCDelayRequirement_val = dscceComputeDelay(DSCInputBitPerComponent, OutputBpp, (dml_uint_t)((dml_float_t) dml_ceil(HActive / (dml_float_t) NumberOfDSCSlices, 1.0)),
+                                                                               NumberOfDSCSlices, OutputFormat, Output) + dscComputeDelay(OutputFormat, Output);
+               }
+               DSCDelayRequirement_val = (dml_uint_t)(DSCDelayRequirement_val + (HTotal - HActive) * dml_ceil((dml_float_t) DSCDelayRequirement_val / (dml_float_t) HActive, 1.0));
+               DSCDelayRequirement_val = (dml_uint_t)(DSCDelayRequirement_val * PixelClock / PixelClockBackEnd);
+
+       } else {
+               DSCDelayRequirement_val = 0;
+       }
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: DSCEnabled = %u\n", __func__, DSCEnabled);
+       dml_print("DML::%s: ODMMode = %u\n", __func__, ODMMode);
+       dml_print("DML::%s: OutputBpp = %f\n", __func__, OutputBpp);
+       dml_print("DML::%s: HActive = %u\n", __func__, HActive);
+       dml_print("DML::%s: HTotal = %u\n", __func__, HTotal);
+       dml_print("DML::%s: PixelClock = %f\n", __func__, PixelClock);
+       dml_print("DML::%s: PixelClockBackEnd = %f\n", __func__, PixelClockBackEnd);
+       dml_print("DML::%s: OutputFormat = %u\n", __func__, OutputFormat);
+       dml_print("DML::%s: DSCInputBitPerComponent = %u\n", __func__, DSCInputBitPerComponent);
+       dml_print("DML::%s: NumberOfDSCSlices = %u\n", __func__, NumberOfDSCSlices);
+       dml_print("DML::%s: DSCDelayRequirement_val = %u\n", __func__, DSCDelayRequirement_val);
+#endif
+
+       return DSCDelayRequirement_val;
+}
+
+static dml_bool_t CalculateVActiveBandwithSupport(dml_uint_t NumberOfActiveSurfaces,
+                                                                               dml_float_t ReturnBW,
+                                                                               dml_bool_t NotUrgentLatencyHiding[],
+                                                                               dml_float_t ReadBandwidthLuma[],
+                                                                               dml_float_t ReadBandwidthChroma[],
+                                                                               dml_float_t cursor_bw[],
+                                                                               dml_float_t meta_row_bandwidth[],
+                                                                               dml_float_t dpte_row_bandwidth[],
+                                                                               dml_uint_t NumberOfDPP[],
+                                                                               dml_float_t UrgentBurstFactorLuma[],
+                                                                               dml_float_t UrgentBurstFactorChroma[],
+                                                                               dml_float_t UrgentBurstFactorCursor[])
+{
+       dml_bool_t NotEnoughUrgentLatencyHiding = false;
+       dml_bool_t CalculateVActiveBandwithSupport_val = false;
+       dml_float_t VActiveBandwith = 0;
+
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               if (NotUrgentLatencyHiding[k]) {
+                       NotEnoughUrgentLatencyHiding = true;
+               }
+       }
+
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               VActiveBandwith = VActiveBandwith + ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k] + NumberOfDPP[k] * meta_row_bandwidth[k] + NumberOfDPP[k] * dpte_row_bandwidth[k];
+       }
+
+       CalculateVActiveBandwithSupport_val = (VActiveBandwith <= ReturnBW) && !NotEnoughUrgentLatencyHiding;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: NotEnoughUrgentLatencyHiding        = %u\n", __func__, NotEnoughUrgentLatencyHiding);
+       dml_print("DML::%s: VActiveBandwith                     = %f\n", __func__, VActiveBandwith);
+       dml_print("DML::%s: ReturnBW                            = %f\n", __func__, ReturnBW);
+       dml_print("DML::%s: CalculateVActiveBandwithSupport_val = %u\n", __func__, CalculateVActiveBandwithSupport_val);
+#endif
+       return CalculateVActiveBandwithSupport_val;
+}
+
+static void CalculatePrefetchBandwithSupport(
+                                                                               dml_uint_t NumberOfActiveSurfaces,
+                                                                               dml_float_t ReturnBW,
+                                                                               enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+                                                                               dml_bool_t NotUrgentLatencyHiding[],
+                                                                               dml_float_t ReadBandwidthLuma[],
+                                                                               dml_float_t ReadBandwidthChroma[],
+                                                                               dml_float_t PrefetchBandwidthLuma[],
+                                                                               dml_float_t PrefetchBandwidthChroma[],
+                                                                               dml_float_t cursor_bw[],
+                                                                               dml_float_t meta_row_bandwidth[],
+                                                                               dml_float_t dpte_row_bandwidth[],
+                                                                               dml_float_t cursor_bw_pre[],
+                                                                               dml_float_t prefetch_vmrow_bw[],
+                                                                               dml_uint_t NumberOfDPP[],
+                                                                               dml_float_t UrgentBurstFactorLuma[],
+                                                                               dml_float_t UrgentBurstFactorChroma[],
+                                                                               dml_float_t UrgentBurstFactorCursor[],
+                                                                               dml_float_t UrgentBurstFactorLumaPre[],
+                                                                               dml_float_t UrgentBurstFactorChromaPre[],
+                                                                               dml_float_t UrgentBurstFactorCursorPre[],
+
+                                                                               // Output
+                                                                               dml_float_t *PrefetchBandwidth,
+                                                                               dml_float_t *PrefetchBandwidthNotIncludingMALLPrefetch,
+                                                                               dml_float_t *FractionOfUrgentBandwidth,
+                                                                               dml_bool_t *PrefetchBandwidthSupport)
+{
+       dml_bool_t NotEnoughUrgentLatencyHiding = false;
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               if (NotUrgentLatencyHiding[k]) {
+                       NotEnoughUrgentLatencyHiding = true;
+               }
+       }
+
+       *PrefetchBandwidth = 0;
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               *PrefetchBandwidth = *PrefetchBandwidth + dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+                                                                                                               ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k] + NumberOfDPP[k] * (meta_row_bandwidth[k] + dpte_row_bandwidth[k]),
+                                                                                                               NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k]) + cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+       }
+
+       *PrefetchBandwidthNotIncludingMALLPrefetch = 0;
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               if (UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe)
+                       *PrefetchBandwidthNotIncludingMALLPrefetch = *PrefetchBandwidthNotIncludingMALLPrefetch
+                               + dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+                               ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k]
+                               + cursor_bw[k] * UrgentBurstFactorCursor[k]
+                               + NumberOfDPP[k] * (meta_row_bandwidth[k] + dpte_row_bandwidth[k]),
+                               NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k]
+                               + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k])
+                               + cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+       }
+
+       *PrefetchBandwidthSupport = (*PrefetchBandwidth <= ReturnBW) && !NotEnoughUrgentLatencyHiding;
+       *FractionOfUrgentBandwidth = *PrefetchBandwidth / ReturnBW;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: ReturnBW = %f\n", __func__, ReturnBW);
+       dml_print("DML::%s: PrefetchBandwidth = %f\n", __func__, *PrefetchBandwidth);
+       dml_print("DML::%s: FractionOfUrgentBandwidth = %f\n", __func__, *FractionOfUrgentBandwidth);
+   dml_print("DML::%s: PrefetchBandwidthSupport = %u\n", __func__, *PrefetchBandwidthSupport);
+#endif
+}
+
+static dml_float_t CalculateBandwidthAvailableForImmediateFlip(
+                                                                                                       dml_uint_t NumberOfActiveSurfaces,
+                                                                                                       dml_float_t ReturnBW,
+                                                                                                       dml_float_t ReadBandwidthLuma[],
+                                                                                                       dml_float_t ReadBandwidthChroma[],
+                                                                                                       dml_float_t PrefetchBandwidthLuma[],
+                                                                                                       dml_float_t PrefetchBandwidthChroma[],
+                                                                                                       dml_float_t cursor_bw[],
+                                                                                                       dml_float_t cursor_bw_pre[],
+                                                                                                       dml_uint_t NumberOfDPP[],
+                                                                                                       dml_float_t UrgentBurstFactorLuma[],
+                                                                                                       dml_float_t UrgentBurstFactorChroma[],
+                                                                                                       dml_float_t UrgentBurstFactorCursor[],
+                                                                                                       dml_float_t UrgentBurstFactorLumaPre[],
+                                                                                                       dml_float_t UrgentBurstFactorChromaPre[],
+                                                                                                       dml_float_t UrgentBurstFactorCursorPre[])
+{
+       dml_float_t ret_val = ReturnBW;
+
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               ret_val = ret_val - dml_max(ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k],
+                                                                       NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k]) +
+                                                                       cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u\n", __func__, k);
+               dml_print("DML::%s: NumberOfDPP = %u\n", __func__, NumberOfDPP[k]);
+               dml_print("DML::%s: ReadBandwidthLuma = %f\n", __func__, ReadBandwidthLuma[k]);
+               dml_print("DML::%s: UrgentBurstFactorLuma = %f\n", __func__, UrgentBurstFactorLuma[k]);
+               dml_print("DML::%s: ReadBandwidthChroma = %f\n", __func__, ReadBandwidthChroma[k]);
+               dml_print("DML::%s: UrgentBurstFactorChroma = %f\n", __func__, UrgentBurstFactorChroma[k]);
+               dml_print("DML::%s: cursor_bw = %f\n", __func__, cursor_bw[k]);
+               dml_print("DML::%s: UrgentBurstFactorCursor = %f\n", __func__, UrgentBurstFactorCursor[k]);
+
+               dml_print("DML::%s: PrefetchBandwidthLuma = %f\n", __func__, PrefetchBandwidthLuma[k]);
+               dml_print("DML::%s: UrgentBurstFactorLumaPre = %f\n", __func__, UrgentBurstFactorLumaPre[k]);
+               dml_print("DML::%s: PrefetchBandwidthChroma = %f\n", __func__, PrefetchBandwidthChroma[k]);
+               dml_print("DML::%s: UrgentBurstFactorChromaPre = %f\n", __func__, UrgentBurstFactorChromaPre[k]);
+               dml_print("DML::%s: cursor_bw_pre = %f\n", __func__, cursor_bw_pre[k]);
+               dml_print("DML::%s: UrgentBurstFactorCursorPre = %f\n", __func__, UrgentBurstFactorCursorPre[k]);
+               dml_print("DML::%s: ret_val              = %f\n", __func__, ret_val);
+#endif
+       }
+
+       return ret_val;
+}
+
+static void CalculateImmediateFlipBandwithSupport(
+                                                                                       dml_uint_t NumberOfActiveSurfaces,
+                                                                                       dml_float_t ReturnBW,
+                                                                                       enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+                                                                                       enum dml_immediate_flip_requirement ImmediateFlipRequirement[],
+                                                                                       dml_float_t final_flip_bw[],
+                                                                                       dml_float_t ReadBandwidthLuma[],
+                                                                                       dml_float_t ReadBandwidthChroma[],
+                                                                                       dml_float_t PrefetchBandwidthLuma[],
+                                                                                       dml_float_t PrefetchBandwidthChroma[],
+                                                                                       dml_float_t cursor_bw[],
+                                                                                       dml_float_t meta_row_bandwidth[],
+                                                                                       dml_float_t dpte_row_bandwidth[],
+                                                                                       dml_float_t cursor_bw_pre[],
+                                                                                       dml_float_t prefetch_vmrow_bw[],
+                                                                                       dml_uint_t NumberOfDPP[],
+                                                                                       dml_float_t UrgentBurstFactorLuma[],
+                                                                                       dml_float_t UrgentBurstFactorChroma[],
+                                                                                       dml_float_t UrgentBurstFactorCursor[],
+                                                                                       dml_float_t UrgentBurstFactorLumaPre[],
+                                                                                       dml_float_t UrgentBurstFactorChromaPre[],
+                                                                                       dml_float_t UrgentBurstFactorCursorPre[],
+
+                                                                                       // Output
+                                                                                       dml_float_t *TotalBandwidth,
+                                                                                       dml_float_t *TotalBandwidthNotIncludingMALLPrefetch,
+                                                                                       dml_float_t *FractionOfUrgentBandwidth,
+                                                                                       dml_bool_t *ImmediateFlipBandwidthSupport)
+{
+       *TotalBandwidth = 0;
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               if (ImmediateFlipRequirement[k] != dml_immediate_flip_not_required) {
+
+
+
+                       *TotalBandwidth = *TotalBandwidth + dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+                                                                                                               NumberOfDPP[k] * final_flip_bw[k] + ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k],
+                                                                                                               NumberOfDPP[k] * (final_flip_bw[k] + PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k]) + cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+               } else {
+                       *TotalBandwidth = *TotalBandwidth + dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+                                                                                                               NumberOfDPP[k] * (meta_row_bandwidth[k] + dpte_row_bandwidth[k]) + ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k],
+                                                                                                               NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k]) + cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+               }
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k = %u\n", __func__, k);
+               dml_print("DML::%s: ImmediateFlipRequirement = %u\n", __func__, ImmediateFlipRequirement[k]);
+               dml_print("DML::%s: TotalBandwidth = %f\n", __func__, *TotalBandwidth);
+               dml_print("DML::%s: NumberOfDPP = %u\n", __func__, NumberOfDPP[k]);
+               dml_print("DML::%s: prefetch_vmrow_bw = %f\n", __func__, prefetch_vmrow_bw[k]);
+               dml_print("DML::%s: final_flip_bw = %f\n", __func__, final_flip_bw[k]);
+               dml_print("DML::%s: ReadBandwidthLuma = %f\n", __func__, ReadBandwidthLuma[k]);
+               dml_print("DML::%s: UrgentBurstFactorLuma = %f\n", __func__, UrgentBurstFactorLuma[k]);
+               dml_print("DML::%s: ReadBandwidthChroma = %f\n", __func__, ReadBandwidthChroma[k]);
+               dml_print("DML::%s: UrgentBurstFactorChroma = %f\n", __func__, UrgentBurstFactorChroma[k]);
+               dml_print("DML::%s: cursor_bw = %f\n", __func__, cursor_bw[k]);
+               dml_print("DML::%s: UrgentBurstFactorCursor = %f\n", __func__, UrgentBurstFactorCursor[k]);
+               dml_print("DML::%s: PrefetchBandwidthLuma = %f\n", __func__, PrefetchBandwidthLuma[k]);
+               dml_print("DML::%s: UrgentBurstFactorLumaPre = %f\n", __func__, UrgentBurstFactorLumaPre[k]);
+               dml_print("DML::%s: PrefetchBandwidthChroma = %f\n", __func__, PrefetchBandwidthChroma[k]);
+               dml_print("DML::%s: UrgentBurstFactorChromaPre = %f\n", __func__, UrgentBurstFactorChromaPre[k]);
+               dml_print("DML::%s: cursor_bw_pre = %f\n", __func__, cursor_bw_pre[k]);
+               dml_print("DML::%s: UrgentBurstFactorCursorPre = %f\n", __func__, UrgentBurstFactorCursorPre[k]);
+               dml_print("DML::%s: meta_row_bandwidth = %f\n", __func__, meta_row_bandwidth[k]);
+               dml_print("DML::%s: dpte_row_bandwidth          = %f\n", __func__, dpte_row_bandwidth[k]);
+#endif
+       }
+
+       *TotalBandwidthNotIncludingMALLPrefetch = 0;
+       for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+               if (UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) {
+                       if (ImmediateFlipRequirement[k] != dml_immediate_flip_not_required)
+                               *TotalBandwidthNotIncludingMALLPrefetch = *TotalBandwidthNotIncludingMALLPrefetch + dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+                                       NumberOfDPP[k] * final_flip_bw[k] + ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k],
+                                       NumberOfDPP[k] * (final_flip_bw[k] + PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k])
+                                       + cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+                       else
+                               *TotalBandwidthNotIncludingMALLPrefetch = *TotalBandwidthNotIncludingMALLPrefetch + dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+                                       NumberOfDPP[k] * (meta_row_bandwidth[k] + dpte_row_bandwidth[k])
+                                       + ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k],
+                                       NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k])
+                                       + cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+               }
+       }
+
+       *ImmediateFlipBandwidthSupport = (*TotalBandwidth <= ReturnBW);
+       *FractionOfUrgentBandwidth = *TotalBandwidth / ReturnBW;
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: ReturnBW = %f\n", __func__, ReturnBW);
+       dml_print("DML::%s: TotalBandwidth = %f\n", __func__, *TotalBandwidth);
+       dml_print("DML::%s: ImmediateFlipBandwidthSupport = %u\n", __func__, *ImmediateFlipBandwidthSupport);
+#endif
+}
+
+/// @brief Calculate the maximum vstartup for mode support and mode programming consideration
+///         Bounded by min of actual vblank and input vblank_nom, dont want vstartup/ready to start too early if actual vbllank is huge
+static dml_uint_t CalculateMaxVStartup(
+                       dml_uint_t          plane_idx,
+                       dml_bool_t          ptoi_supported,
+                       dml_uint_t          vblank_nom_default_us,
+                       struct dml_timing_cfg_st  *timing,
+                       dml_float_t         write_back_delay_us)
+{
+       dml_uint_t vblank_size = 0;
+       dml_uint_t max_vstartup_lines = 0;
+
+       dml_float_t line_time_us = (dml_float_t) timing->HTotal[plane_idx] / timing->PixelClock[plane_idx];
+       dml_uint_t vblank_actual = timing->VTotal[plane_idx] - timing->VActive[plane_idx];
+       dml_uint_t vblank_nom_default_in_line = (dml_uint_t) dml_floor((dml_float_t) vblank_nom_default_us/line_time_us, 1.0);
+       dml_uint_t vblank_nom_input = (dml_uint_t) dml_min(timing->VBlankNom[plane_idx], vblank_nom_default_in_line);
+       dml_uint_t vblank_avail = (vblank_nom_input == 0) ? vblank_nom_default_in_line : vblank_nom_input;
+
+       vblank_size = (dml_uint_t) dml_min(vblank_actual, vblank_avail);
+
+   if (timing->Interlace[plane_idx] && !ptoi_supported)
+               max_vstartup_lines = (dml_uint_t) (dml_floor(vblank_size/2.0, 1.0));
+       else
+               max_vstartup_lines = vblank_size - (dml_uint_t) dml_max(1.0, dml_ceil(write_back_delay_us/line_time_us, 1.0));
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: plane_idx = %u\n", __func__, plane_idx);
+       dml_print("DML::%s: VBlankNom = %u\n", __func__, timing->VBlankNom[plane_idx]);
+       dml_print("DML::%s: vblank_nom_default_us = %u\n", __func__, vblank_nom_default_us);
+       dml_print("DML::%s: line_time_us = %f\n", __func__, line_time_us);
+       dml_print("DML::%s: vblank_actual = %u\n", __func__, vblank_actual);
+       dml_print("DML::%s: vblank_avail = %u\n", __func__, vblank_avail);
+       dml_print("DML::%s: max_vstartup_lines = %u\n", __func__, max_vstartup_lines);
+#endif
+       return max_vstartup_lines;
+}
+
+/// @brief The Mode Support function.
+dml_bool_t dml_core_mode_support(struct display_mode_lib_st *mode_lib)
+{
+       struct dml_core_mode_support_locals_st *s = &mode_lib->scratch.dml_core_mode_support_locals;
+       struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params_st *CalculateWatermarks_params = &mode_lib->scratch.CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params;
+       struct CalculateVMRowAndSwath_params_st *CalculateVMRowAndSwath_params = &mode_lib->scratch.CalculateVMRowAndSwath_params;
+       struct UseMinimumDCFCLK_params_st *UseMinimumDCFCLK_params = &mode_lib->scratch.UseMinimumDCFCLK_params;
+       struct CalculateSwathAndDETConfiguration_params_st *CalculateSwathAndDETConfiguration_params = &mode_lib->scratch.CalculateSwathAndDETConfiguration_params;
+       struct CalculatePrefetchSchedule_params_st *CalculatePrefetchSchedule_params = &mode_lib->scratch.CalculatePrefetchSchedule_params;
+       struct DmlPipe *myPipe;
+
+       dml_uint_t j, k, m;
+
+       mode_lib->ms.num_active_planes = dml_get_num_active_planes(&mode_lib->ms.cache_display_cfg);
+       dml_print("DML::%s: num_active_planes = %u\n", __func__, mode_lib->ms.num_active_planes);
+
+       CalculateMaxDETAndMinCompressedBufferSize(
+                                                               mode_lib->ms.ip.config_return_buffer_size_in_kbytes,
+                                                               mode_lib->ms.ip.config_return_buffer_segment_size_in_kbytes,
+                                                               mode_lib->ms.ip.rob_buffer_size_kbytes,
+                                                               mode_lib->ms.ip.max_num_dpp,
+                                                               mode_lib->ms.policy.NomDETInKByteOverrideEnable,   // VBA_DELTA
+                                                               mode_lib->ms.policy.NomDETInKByteOverrideValue,    // VBA_DELTA
+
+                                                               /* Output */
+                                                               &mode_lib->ms.MaxTotalDETInKByte,
+                                                               &mode_lib->ms.NomDETInKByte,
+                                                               &mode_lib->ms.MinCompressedBufferSizeInKByte);
+
+       PixelClockAdjustmentForProgressiveToInterlaceUnit(&mode_lib->ms.cache_display_cfg, mode_lib->ms.ip.ptoi_supported);
+
+
+       /*MODE SUPPORT, VOLTAGE STATE AND SOC CONFIGURATION*/
+
+       /*Scale Ratio, taps Support Check*/
+       mode_lib->ms.support.ScaleRatioAndTapsSupport = true;
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               if (mode_lib->ms.cache_display_cfg.plane.ScalerEnabled[k] == false
+                               && ((mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_64
+                                               && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_32
+                                               && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_16
+                                               && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_mono_16
+                                               && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_mono_8
+                                               && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_rgbe
+                                               && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_rgbe_alpha)
+                                               || mode_lib->ms.cache_display_cfg.plane.HRatio[k] != 1.0
+                                               || mode_lib->ms.cache_display_cfg.plane.HTaps[k] != 1.0
+                                               || mode_lib->ms.cache_display_cfg.plane.VRatio[k] != 1.0
+                                               || mode_lib->ms.cache_display_cfg.plane.VTaps[k] != 1.0)) {
+                       mode_lib->ms.support.ScaleRatioAndTapsSupport = false;
+               } else if (mode_lib->ms.cache_display_cfg.plane.VTaps[k] < 1.0 || mode_lib->ms.cache_display_cfg.plane.VTaps[k] > 8.0
+                               || mode_lib->ms.cache_display_cfg.plane.HTaps[k] < 1.0 || mode_lib->ms.cache_display_cfg.plane.HTaps[k] > 8.0
+                               || (mode_lib->ms.cache_display_cfg.plane.HTaps[k] > 1.0 && (mode_lib->ms.cache_display_cfg.plane.HTaps[k] % 2) == 1)
+                               || mode_lib->ms.cache_display_cfg.plane.HRatio[k] > mode_lib->ms.ip.max_hscl_ratio
+                               || mode_lib->ms.cache_display_cfg.plane.VRatio[k] > mode_lib->ms.ip.max_vscl_ratio
+                               || mode_lib->ms.cache_display_cfg.plane.HRatio[k] > mode_lib->ms.cache_display_cfg.plane.HTaps[k]
+                               || mode_lib->ms.cache_display_cfg.plane.VRatio[k] > mode_lib->ms.cache_display_cfg.plane.VTaps[k]
+                               || (mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_64
+                                       && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_32
+                                       && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_16
+                                       && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_mono_16
+                                       && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_mono_8
+                                       && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_rgbe
+                                       && (mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k] < 1 || mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k] > 8 || mode_lib->ms.cache_display_cfg.plane.HTapsChroma[k] < 1 || mode_lib->ms.cache_display_cfg.plane.HTapsChroma[k] > 8 ||
+                                               (mode_lib->ms.cache_display_cfg.plane.HTapsChroma[k] > 1 && mode_lib->ms.cache_display_cfg.plane.HTapsChroma[k] % 2 == 1) ||
+                                       mode_lib->ms.cache_display_cfg.plane.HRatioChroma[k] > mode_lib->ms.ip.max_hscl_ratio ||
+                                       mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k] > mode_lib->ms.ip.max_vscl_ratio ||
+                                       mode_lib->ms.cache_display_cfg.plane.HRatioChroma[k] > mode_lib->ms.cache_display_cfg.plane.HTapsChroma[k] ||
+                                       mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k] > mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k]))) {
+                       mode_lib->ms.support.ScaleRatioAndTapsSupport = false;
+               }
+       }
+
+       /*Source Format, Pixel Format and Scan Support Check*/
+       mode_lib->ms.support.SourceFormatPixelAndScanSupport = true;
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                       if (mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k] == dml_sw_linear && (!(!dml_is_vertical_rotation(mode_lib->ms.cache_display_cfg.plane.SourceScan[k])) || mode_lib->ms.cache_display_cfg.surface.DCCEnable[k] == true)) {
+                       mode_lib->ms.support.SourceFormatPixelAndScanSupport = false;
+               }
+       }
+
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               CalculateBytePerPixelAndBlockSizes(
+                                                               mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k],
+                                                               mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k],
+
+                                                               /* Output */
+                                                               &mode_lib->ms.BytePerPixelY[k],
+                                                               &mode_lib->ms.BytePerPixelC[k],
+                                                               &mode_lib->ms.BytePerPixelInDETY[k],
+                                                               &mode_lib->ms.BytePerPixelInDETC[k],
+                                                               &mode_lib->ms.Read256BlockHeightY[k],
+                                                               &mode_lib->ms.Read256BlockHeightC[k],
+                                                               &mode_lib->ms.Read256BlockWidthY[k],
+                                                               &mode_lib->ms.Read256BlockWidthC[k],
+                                                               &mode_lib->ms.MacroTileHeightY[k],
+                                                               &mode_lib->ms.MacroTileHeightC[k],
+                                                               &mode_lib->ms.MacroTileWidthY[k],
+                                                               &mode_lib->ms.MacroTileWidthC[k]);
+       }
+
+       /*Bandwidth Support Check*/
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               if (!dml_is_vertical_rotation(mode_lib->ms.cache_display_cfg.plane.SourceScan[k])) {
+                       mode_lib->ms.SwathWidthYSingleDPP[k] = mode_lib->ms.cache_display_cfg.plane.ViewportWidth[k];
+                       mode_lib->ms.SwathWidthCSingleDPP[k] = mode_lib->ms.cache_display_cfg.plane.ViewportWidthChroma[k];
+               } else {
+                       mode_lib->ms.SwathWidthYSingleDPP[k] = mode_lib->ms.cache_display_cfg.plane.ViewportHeight[k];
+                       mode_lib->ms.SwathWidthCSingleDPP[k] = mode_lib->ms.cache_display_cfg.plane.ViewportHeightChroma[k];
+               }
+       }
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               mode_lib->ms.ReadBandwidthLuma[k] = mode_lib->ms.SwathWidthYSingleDPP[k] * dml_ceil(mode_lib->ms.BytePerPixelInDETY[k], 1.0) / (mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * mode_lib->ms.cache_display_cfg.plane.VRatio[k];
+               mode_lib->ms.ReadBandwidthChroma[k] = mode_lib->ms.SwathWidthYSingleDPP[k] / 2 * dml_ceil(mode_lib->ms.BytePerPixelInDETC[k], 2.0) / (mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * mode_lib->ms.cache_display_cfg.plane.VRatio[k] / 2.0;
+       }
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true
+                               && mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[k] == dml_444_64) {
+                       mode_lib->ms.WriteBandwidth[k] = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k]
+                                       * mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[k]
+                                       / (mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[k]
+                                                       * mode_lib->ms.cache_display_cfg.timing.HTotal[k]
+                                                       / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * 8.0;
+               } else if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true) {
+                       mode_lib->ms.WriteBandwidth[k] = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k]
+                                       * mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[k]
+                                       / (mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[k]
+                                                       * mode_lib->ms.cache_display_cfg.timing.HTotal[k]
+                                                       / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * 4.0;
+               } else {
+                       mode_lib->ms.WriteBandwidth[k] = 0.0;
+               }
+       }
+
+       /*Writeback Latency support check*/
+       mode_lib->ms.support.WritebackLatencySupport = true;
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true &&
+                       (mode_lib->ms.WriteBandwidth[k] > mode_lib->ms.ip.writeback_interface_buffer_size_kbytes * 1024 / mode_lib->ms.state.writeback_latency_us)) {
+                       mode_lib->ms.support.WritebackLatencySupport = false;
+               }
+       }
+
+       /*Writeback Mode Support Check*/
+       s->TotalNumberOfActiveWriteback = 0;
+       for (k = 0; k <= (dml_uint_t) mode_lib->ms.num_active_planes - 1; k++) {
+               if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true) {
+                       s->TotalNumberOfActiveWriteback = s->TotalNumberOfActiveWriteback + 1;
+               }
+   }
+
+       mode_lib->ms.support.EnoughWritebackUnits = 1;
+       if (s->TotalNumberOfActiveWriteback > (dml_uint_t) mode_lib->ms.ip.max_num_wb) {
+               mode_lib->ms.support.EnoughWritebackUnits = false;
+       }
+
+       /*Writeback Scale Ratio and Taps Support Check*/
+       mode_lib->ms.support.WritebackScaleRatioAndTapsSupport = true;
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true) {
+                       if (mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[k] > mode_lib->ms.ip.writeback_max_hscl_ratio
+                               || mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[k] > mode_lib->ms.ip.writeback_max_vscl_ratio
+                               || mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[k] < mode_lib->ms.ip.writeback_min_hscl_ratio
+                               || mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[k] < mode_lib->ms.ip.writeback_min_vscl_ratio
+                               || mode_lib->ms.cache_display_cfg.writeback.WritebackHTaps[k] > (dml_uint_t) mode_lib->ms.ip.writeback_max_hscl_taps
+                               || mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[k] > (dml_uint_t) mode_lib->ms.ip.writeback_max_vscl_taps
+                               || mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[k] > (dml_uint_t) mode_lib->ms.cache_display_cfg.writeback.WritebackHTaps[k]
+                               || mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[k] > (dml_uint_t) mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[k]
+                               || (mode_lib->ms.cache_display_cfg.writeback.WritebackHTaps[k] > 2.0 && ((mode_lib->ms.cache_display_cfg.writeback.WritebackHTaps[k] % 2) == 1))) {
+                               mode_lib->ms.support.WritebackScaleRatioAndTapsSupport = false;
+                       }
+                       if (2.0 * mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k] * (mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[k] - 1) * 57 > mode_lib->ms.ip.writeback_line_buffer_buffer_size) {
+                               mode_lib->ms.support.WritebackScaleRatioAndTapsSupport = false;
+                       }
+               }
+       }
+
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               CalculateSinglePipeDPPCLKAndSCLThroughput(
+                               mode_lib->ms.cache_display_cfg.plane.HRatio[k],
+                               mode_lib->ms.cache_display_cfg.plane.HRatioChroma[k],
+                               mode_lib->ms.cache_display_cfg.plane.VRatio[k],
+                               mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k],
+                               mode_lib->ms.ip.max_dchub_pscl_bw_pix_per_clk,
+                               mode_lib->ms.ip.max_pscl_lb_bw_pix_per_clk,
+                               mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+                               mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k],
+                               mode_lib->ms.cache_display_cfg.plane.HTaps[k],
+                               mode_lib->ms.cache_display_cfg.plane.HTapsChroma[k],
+                               mode_lib->ms.cache_display_cfg.plane.VTaps[k],
+                               mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k],
+                               /* Output */
+                               &mode_lib->ms.PSCL_FACTOR[k],
+                               &mode_lib->ms.PSCL_FACTOR_CHROMA[k],
+                               &mode_lib->ms.MinDPPCLKUsingSingleDPP[k]);
+       }
+
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               if (mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k] == dml_sw_linear) {
+                       s->MaximumSwathWidthSupportLuma = 8192;
+               } else if (!dml_is_vertical_rotation(mode_lib->ms.cache_display_cfg.plane.SourceScan[k]) && mode_lib->ms.BytePerPixelC[k] > 0 && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_rgbe_alpha) {
+                       s->MaximumSwathWidthSupportLuma = 7680;
+               } else if (dml_is_vertical_rotation(mode_lib->ms.cache_display_cfg.plane.SourceScan[k]) && mode_lib->ms.BytePerPixelC[k] > 0 && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_rgbe_alpha) {
+                       s->MaximumSwathWidthSupportLuma = 4320;
+               } else if (mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_rgbe_alpha) {
+                       s->MaximumSwathWidthSupportLuma = 3840;
+               } else if (dml_is_vertical_rotation(mode_lib->ms.cache_display_cfg.plane.SourceScan[k]) && mode_lib->ms.BytePerPixelY[k] == 8 && mode_lib->ms.cache_display_cfg.surface.DCCEnable[k] == true) {
+                       s->MaximumSwathWidthSupportLuma = 3072;
+               } else {
+                       s->MaximumSwathWidthSupportLuma = 6144;
+               }
+
+       if (mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_420_8 || mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_420_10 || mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_420_12) {
+                       s->MaximumSwathWidthSupportChroma = (dml_uint_t)(s->MaximumSwathWidthSupportLuma / 2.0);
+               } else {
+                       s->MaximumSwathWidthSupportChroma = s->MaximumSwathWidthSupportLuma;
+       }
+               mode_lib->ms.MaximumSwathWidthInLineBufferLuma = mode_lib->ms.ip.line_buffer_size_bits * dml_max(mode_lib->ms.cache_display_cfg.plane.HRatio[k], 1.0) / mode_lib->ms.cache_display_cfg.plane.LBBitPerPixel[k] /
+                                                                                                                       (mode_lib->ms.cache_display_cfg.plane.VTaps[k] + dml_max(dml_ceil(mode_lib->ms.cache_display_cfg.plane.VRatio[k], 1.0) - 2, 0.0));
+               if (mode_lib->ms.BytePerPixelC[k] == 0.0) {
+                       mode_lib->ms.MaximumSwathWidthInLineBufferChroma = 0;
+               } else {
+                       mode_lib->ms.MaximumSwathWidthInLineBufferChroma =
+                                                       mode_lib->ms.ip.line_buffer_size_bits
+                                                                       * dml_max(mode_lib->ms.cache_display_cfg.plane.HRatioChroma[k], 1.0)
+                                                                       / mode_lib->ms.cache_display_cfg.plane.LBBitPerPixel[k]
+                                                                       / (mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k]
+                                                                       + dml_max(dml_ceil(mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k], 1.0) - 2, 0.0));
+               }
+               mode_lib->ms.MaximumSwathWidthLuma[k] = dml_min(s->MaximumSwathWidthSupportLuma, mode_lib->ms.MaximumSwathWidthInLineBufferLuma);
+               mode_lib->ms.MaximumSwathWidthChroma[k] = dml_min(s->MaximumSwathWidthSupportChroma, mode_lib->ms.MaximumSwathWidthInLineBufferChroma);
+       }
+
+       /*Number Of DSC Slices*/
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k) {
+                       if (mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] > 4800) {
+                               mode_lib->ms.support.NumberOfDSCSlices[k] = (dml_uint_t)(dml_ceil(mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] / 600, 4));
+                       } else if (mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] > 2400) {
+                               mode_lib->ms.support.NumberOfDSCSlices[k] = 8;
+                       } else if (mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] > 1200) {
+                               mode_lib->ms.support.NumberOfDSCSlices[k] = 4;
+                       } else if (mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] > 340) {
+                               mode_lib->ms.support.NumberOfDSCSlices[k] = 2;
+                       } else {
+                               mode_lib->ms.support.NumberOfDSCSlices[k] = 1;
+                       }
+               } else {
+                       mode_lib->ms.support.NumberOfDSCSlices[k] = 0;
+               }
+       }
+
+       CalculateSwathAndDETConfiguration_params->DETSizeOverride = mode_lib->ms.cache_display_cfg.plane.DETSizeOverride;
+       CalculateSwathAndDETConfiguration_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+       CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSizeInKByte = mode_lib->ms.ip.config_return_buffer_size_in_kbytes;
+       CalculateSwathAndDETConfiguration_params->ROBBufferSizeInKByte = mode_lib->ms.ip.rob_buffer_size_kbytes;
+       CalculateSwathAndDETConfiguration_params->MaxTotalDETInKByte = mode_lib->ms.MaxTotalDETInKByte;
+       CalculateSwathAndDETConfiguration_params->MinCompressedBufferSizeInKByte = mode_lib->ms.MinCompressedBufferSizeInKByte;
+       CalculateSwathAndDETConfiguration_params->PixelChunkSizeInKByte = mode_lib->ms.ip.pixel_chunk_size_kbytes;
+       CalculateSwathAndDETConfiguration_params->ForceSingleDPP = 1;
+       CalculateSwathAndDETConfiguration_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+       CalculateSwathAndDETConfiguration_params->nomDETInKByte = mode_lib->ms.NomDETInKByte;
+       CalculateSwathAndDETConfiguration_params->UseUnboundedRequestingFinal = mode_lib->ms.policy.UseUnboundedRequesting;
+       CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSegmentSizeInkByte = mode_lib->ms.ip.config_return_buffer_segment_size_in_kbytes;
+       CalculateSwathAndDETConfiguration_params->CompressedBufferSegmentSizeInkByteFinal = mode_lib->ms.ip.compressed_buffer_segment_size_in_kbytes;
+       CalculateSwathAndDETConfiguration_params->Output = mode_lib->ms.cache_display_cfg.output.OutputEncoder;
+       CalculateSwathAndDETConfiguration_params->ReadBandwidthLuma = mode_lib->ms.ReadBandwidthLuma;
+       CalculateSwathAndDETConfiguration_params->ReadBandwidthChroma = mode_lib->ms.ReadBandwidthChroma;
+       CalculateSwathAndDETConfiguration_params->MaximumSwathWidthLuma = mode_lib->ms.MaximumSwathWidthLuma;
+       CalculateSwathAndDETConfiguration_params->MaximumSwathWidthChroma = mode_lib->ms.MaximumSwathWidthChroma;
+       CalculateSwathAndDETConfiguration_params->SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan;
+       CalculateSwathAndDETConfiguration_params->ViewportStationary = mode_lib->ms.cache_display_cfg.plane.ViewportStationary;
+       CalculateSwathAndDETConfiguration_params->SourcePixelFormat = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat;
+       CalculateSwathAndDETConfiguration_params->SurfaceTiling = mode_lib->ms.cache_display_cfg.surface.SurfaceTiling;
+       CalculateSwathAndDETConfiguration_params->ViewportWidth = mode_lib->ms.cache_display_cfg.plane.ViewportWidth;
+       CalculateSwathAndDETConfiguration_params->ViewportHeight = mode_lib->ms.cache_display_cfg.plane.ViewportHeight;
+       CalculateSwathAndDETConfiguration_params->ViewportXStart = mode_lib->ms.cache_display_cfg.plane.ViewportXStart;
+       CalculateSwathAndDETConfiguration_params->ViewportYStart = mode_lib->ms.cache_display_cfg.plane.ViewportYStart;
+       CalculateSwathAndDETConfiguration_params->ViewportXStartC = mode_lib->ms.cache_display_cfg.plane.ViewportXStartC;
+       CalculateSwathAndDETConfiguration_params->ViewportYStartC = mode_lib->ms.cache_display_cfg.plane.ViewportYStartC;
+       CalculateSwathAndDETConfiguration_params->SurfaceWidthY = mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY;
+       CalculateSwathAndDETConfiguration_params->SurfaceWidthC = mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC;
+       CalculateSwathAndDETConfiguration_params->SurfaceHeightY = mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY;
+       CalculateSwathAndDETConfiguration_params->SurfaceHeightC = mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC;
+       CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightY = mode_lib->ms.Read256BlockHeightY;
+       CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightC = mode_lib->ms.Read256BlockHeightC;
+       CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthY = mode_lib->ms.Read256BlockWidthY;
+       CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthC = mode_lib->ms.Read256BlockWidthC;
+       CalculateSwathAndDETConfiguration_params->ODMMode = s->dummy_odm_mode;
+       CalculateSwathAndDETConfiguration_params->BlendingAndTiming = mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming;
+       CalculateSwathAndDETConfiguration_params->BytePerPixY = mode_lib->ms.BytePerPixelY;
+       CalculateSwathAndDETConfiguration_params->BytePerPixC = mode_lib->ms.BytePerPixelC;
+       CalculateSwathAndDETConfiguration_params->BytePerPixDETY = mode_lib->ms.BytePerPixelInDETY;
+       CalculateSwathAndDETConfiguration_params->BytePerPixDETC = mode_lib->ms.BytePerPixelInDETC;
+       CalculateSwathAndDETConfiguration_params->HActive = mode_lib->ms.cache_display_cfg.timing.HActive;
+       CalculateSwathAndDETConfiguration_params->HRatio = mode_lib->ms.cache_display_cfg.plane.HRatio;
+       CalculateSwathAndDETConfiguration_params->HRatioChroma = mode_lib->ms.cache_display_cfg.plane.HRatioChroma;
+       CalculateSwathAndDETConfiguration_params->DPPPerSurface = s->dummy_integer_array[0];
+       CalculateSwathAndDETConfiguration_params->swath_width_luma_ub = s->dummy_integer_array[1];
+       CalculateSwathAndDETConfiguration_params->swath_width_chroma_ub = s->dummy_integer_array[2];
+       CalculateSwathAndDETConfiguration_params->SwathWidth = s->dummy_integer_array[3];
+       CalculateSwathAndDETConfiguration_params->SwathWidthChroma = s->dummy_integer_array[4];
+       CalculateSwathAndDETConfiguration_params->SwathHeightY = s->dummy_integer_array[5];
+       CalculateSwathAndDETConfiguration_params->SwathHeightC = s->dummy_integer_array[6];
+       CalculateSwathAndDETConfiguration_params->DETBufferSizeInKByte = s->dummy_integer_array[7];
+       CalculateSwathAndDETConfiguration_params->DETBufferSizeY = mode_lib->ms.DETBufferSizeY;
+       CalculateSwathAndDETConfiguration_params->DETBufferSizeC = mode_lib->ms.DETBufferSizeC;
+       CalculateSwathAndDETConfiguration_params->UnboundedRequestEnabled = &s->dummy_boolean[0];
+       CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_64b = &s->dummy_integer[2];
+       CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_zs = &s->dummy_integer[1];
+       CalculateSwathAndDETConfiguration_params->CompressedBufferSizeInkByte = &s->dummy_integer[0];
+       CalculateSwathAndDETConfiguration_params->ViewportSizeSupportPerSurface = mode_lib->ms.SingleDPPViewportSizeSupportPerSurface;
+       CalculateSwathAndDETConfiguration_params->ViewportSizeSupport = &s->dummy_boolean[1];
+
+       CalculateSwathAndDETConfiguration(&mode_lib->scratch,
+       CalculateSwathAndDETConfiguration_params); /* dml_bool_t *ViewportSizeSupport */
+
+       s->MPCCombineMethodAsNeededForPStateChangeAndVoltage = false;
+       s->MPCCombineMethodAsPossible = false;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.policy.MPCCombineUse[k] == dml_mpc_as_needed_for_pstate_and_voltage)
+                       s->MPCCombineMethodAsNeededForPStateChangeAndVoltage = true;
+               if (mode_lib->ms.policy.MPCCombineUse[k] == dml_mpc_as_possible)
+                       s->MPCCombineMethodAsPossible = true;
+       }
+       mode_lib->ms.support.MPCCombineMethodIncompatible = s->MPCCombineMethodAsNeededForPStateChangeAndVoltage && s->MPCCombineMethodAsPossible;
+
+       for (j = 0; j < 2; j++) {
+               mode_lib->ms.TotalNumberOfActiveDPP[j] = 0;
+               mode_lib->ms.support.TotalAvailablePipesSupport[j] = true;
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       CalculateODMMode(
+                                       mode_lib->ms.ip.maximum_pixels_per_line_per_dsc_unit,
+                                       mode_lib->ms.cache_display_cfg.timing.HActive[k],
+                                       mode_lib->ms.cache_display_cfg.output.OutputEncoder[k],
+                                       mode_lib->ms.cache_display_cfg.output.OutputFormat[k],
+                                       mode_lib->ms.policy.ODMUse[k],
+                                       mode_lib->ms.state.dispclk_mhz,
+                                       mode_lib->ms.max_state.dispclk_mhz,
+                                       false, // DSCEnable
+                                       mode_lib->ms.TotalNumberOfActiveDPP[j],
+                                       mode_lib->ms.ip.max_num_dpp,
+                                       mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+                                       mode_lib->ms.soc.dcn_downspread_percent,
+                                       mode_lib->ms.ip.dispclk_ramp_margin_percent,
+                                       mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz,
+
+                                       /* Output */
+                                       &s->TotalAvailablePipesSupportNoDSC,
+                                       &s->NumberOfDPPNoDSC,
+                                       &s->ODMModeNoDSC,
+                                       &s->RequiredDISPCLKPerSurfaceNoDSC);
+
+                       CalculateODMMode(
+                                       mode_lib->ms.ip.maximum_pixels_per_line_per_dsc_unit,
+                                       mode_lib->ms.cache_display_cfg.timing.HActive[k],
+                                       mode_lib->ms.cache_display_cfg.output.OutputEncoder[k],
+                                       mode_lib->ms.cache_display_cfg.output.OutputFormat[k],
+                                       mode_lib->ms.policy.ODMUse[k],
+                                       mode_lib->ms.state.dispclk_mhz,
+                                       mode_lib->ms.max_state.dispclk_mhz,
+                                       true, // DSCEnable
+                                       mode_lib->ms.TotalNumberOfActiveDPP[j],
+                                       mode_lib->ms.ip.max_num_dpp,
+                                       mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+                                       mode_lib->ms.soc.dcn_downspread_percent,
+                                       mode_lib->ms.ip.dispclk_ramp_margin_percent,
+                                       mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz,
+
+                                       /* Output */
+                                       &s->TotalAvailablePipesSupportDSC,
+                                       &s->NumberOfDPPDSC,
+                                       &s->ODMModeDSC,
+                                       &s->RequiredDISPCLKPerSurfaceDSC);
+
+                       CalculateOutputLink(
+                                       mode_lib->ms.state.phyclk_mhz,
+                                       mode_lib->ms.state.phyclk_d18_mhz,
+                                       mode_lib->ms.state.phyclk_d32_mhz,
+                                       mode_lib->ms.soc.phy_downspread_percent,
+                                       (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k),
+                                       mode_lib->ms.cache_display_cfg.output.OutputEncoder[k],
+                                       mode_lib->ms.cache_display_cfg.output.OutputFormat[k],
+                                       mode_lib->ms.cache_display_cfg.timing.HTotal[k],
+                                       mode_lib->ms.cache_display_cfg.timing.HActive[k],
+                                       mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k],
+                                       mode_lib->ms.cache_display_cfg.output.ForcedOutputLinkBPP[k],
+                                       mode_lib->ms.cache_display_cfg.output.DSCInputBitPerComponent[k],
+                                       mode_lib->ms.support.NumberOfDSCSlices[k],
+                                       mode_lib->ms.cache_display_cfg.output.AudioSampleRate[k],
+                                       mode_lib->ms.cache_display_cfg.output.AudioSampleLayout[k],
+                                       s->ODMModeNoDSC,
+                                       s->ODMModeDSC,
+                                       mode_lib->ms.cache_display_cfg.output.DSCEnable[k],
+                                       mode_lib->ms.cache_display_cfg.output.OutputLinkDPLanes[k],
+                                       mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k],
+
+                                       /* Output */
+                                       &mode_lib->ms.RequiresDSC[k],
+                                       &mode_lib->ms.RequiresFEC[k],
+                                       &mode_lib->ms.OutputBppPerState[k],
+                                       &mode_lib->ms.OutputTypePerState[k], // VBA_DELTA, VBA uses a string to represent type and rate, but DML uses enum, don't want to rely on strng
+                                       &mode_lib->ms.OutputRatePerState[k],
+                                       &mode_lib->ms.RequiredSlots[k]);
+
+                       if (mode_lib->ms.RequiresDSC[k] == false) {
+                               mode_lib->ms.ODMModePerState[k] = s->ODMModeNoDSC;
+                               mode_lib->ms.RequiredDISPCLKPerSurface[j][k] = s->RequiredDISPCLKPerSurfaceNoDSC;
+                               if (!s->TotalAvailablePipesSupportNoDSC)
+                                       mode_lib->ms.support.TotalAvailablePipesSupport[j] = false;
+                               mode_lib->ms.TotalNumberOfActiveDPP[j] = mode_lib->ms.TotalNumberOfActiveDPP[j] + s->NumberOfDPPNoDSC;
+                       } else {
+                               mode_lib->ms.ODMModePerState[k] = s->ODMModeDSC;
+                               mode_lib->ms.RequiredDISPCLKPerSurface[j][k] = s->RequiredDISPCLKPerSurfaceDSC;
+                               if (!s->TotalAvailablePipesSupportDSC)
+                                       mode_lib->ms.support.TotalAvailablePipesSupport[j] = false;
+                               mode_lib->ms.TotalNumberOfActiveDPP[j] = mode_lib->ms.TotalNumberOfActiveDPP[j] + s->NumberOfDPPDSC;
+                       }
+               }
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       if (mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_4to1) {
+                               mode_lib->ms.MPCCombine[j][k] = false;
+                               mode_lib->ms.NoOfDPP[j][k] = 4;
+                       } else if (mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_2to1) {
+                               mode_lib->ms.MPCCombine[j][k] = false;
+                               mode_lib->ms.NoOfDPP[j][k] = 2;
+                       } else if (mode_lib->ms.policy.MPCCombineUse[k] == dml_mpc_disabled) {
+                               mode_lib->ms.MPCCombine[j][k] = false;
+                               mode_lib->ms.NoOfDPP[j][k] = 1;
+                       } else if (RoundToDFSGranularity(mode_lib->ms.MinDPPCLKUsingSingleDPP[k] * (1 + mode_lib->ms.soc.dcn_downspread_percent / 100),
+                                                                                       1, mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz) <= mode_lib->ms.state.dppclk_mhz &&
+                                                                                       mode_lib->ms.SingleDPPViewportSizeSupportPerSurface[k] == true) {
+                               mode_lib->ms.MPCCombine[j][k] = false;
+                               mode_lib->ms.NoOfDPP[j][k] = 1;
+                       } else if (mode_lib->ms.TotalNumberOfActiveDPP[j] < (dml_uint_t) mode_lib->ms.ip.max_num_dpp) {
+                               mode_lib->ms.MPCCombine[j][k] = true;
+                               mode_lib->ms.NoOfDPP[j][k] = 2;
+                               mode_lib->ms.TotalNumberOfActiveDPP[j] = (dml_uint_t) mode_lib->ms.TotalNumberOfActiveDPP[j] + 1;
+                       } else {
+                               mode_lib->ms.MPCCombine[j][k] = false;
+                               mode_lib->ms.NoOfDPP[j][k] = 1;
+                               mode_lib->ms.support.TotalAvailablePipesSupport[j] = false;
+                       }
+               }
+
+               mode_lib->ms.TotalNumberOfSingleDPPSurfaces[j] = 0;
+               s->NoChromaOrLinear = true;
+               for (k = 0; k < (dml_uint_t) mode_lib->ms.num_active_planes; ++k) {
+                       if (mode_lib->ms.NoOfDPP[j][k] == 1)
+                               mode_lib->ms.TotalNumberOfSingleDPPSurfaces[j] = mode_lib->ms.TotalNumberOfSingleDPPSurfaces[j] + 1;
+                       if (mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_420_8
+                                       || mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_420_10
+                                       || mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_420_12
+                                       || mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_rgbe_alpha
+                                       || mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k] == dml_sw_linear) {
+                               s->NoChromaOrLinear = false;
+                       }
+               }
+
+               if (j == 1 && !UnboundedRequest(mode_lib->ms.policy.UseUnboundedRequesting,
+                               mode_lib->ms.TotalNumberOfActiveDPP[j], s->NoChromaOrLinear,
+                               mode_lib->ms.cache_display_cfg.output.OutputEncoder[0])) {
+                       while (!(mode_lib->ms.TotalNumberOfActiveDPP[j] >= (dml_uint_t) mode_lib->ms.ip.max_num_dpp || mode_lib->ms.TotalNumberOfSingleDPPSurfaces[j] == 0)) {
+                               s->BWOfNonCombinedSurfaceOfMaximumBandwidth = 0;
+                               s->NumberOfNonCombinedSurfaceOfMaximumBandwidth = 0;
+                               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                                       if (mode_lib->ms.policy.MPCCombineUse[k] != dml_mpc_disabled && mode_lib->ms.policy.MPCCombineUse[k] != dml_mpc_as_needed_for_voltage &&
+                                               mode_lib->ms.ReadBandwidthLuma[k] + mode_lib->ms.ReadBandwidthChroma[k] > s->BWOfNonCombinedSurfaceOfMaximumBandwidth &&
+                                               (mode_lib->ms.ODMModePerState[k] != dml_odm_mode_combine_2to1 && mode_lib->ms.ODMModePerState[k] != dml_odm_mode_combine_4to1) &&
+                                               mode_lib->ms.MPCCombine[j][k] == false) {
+                                               s->BWOfNonCombinedSurfaceOfMaximumBandwidth = mode_lib->ms.ReadBandwidthLuma[k] + mode_lib->ms.ReadBandwidthChroma[k];
+                                               s->NumberOfNonCombinedSurfaceOfMaximumBandwidth = k;
+                                       }
+                               }
+                               mode_lib->ms.MPCCombine[j][s->NumberOfNonCombinedSurfaceOfMaximumBandwidth] = true;
+                               mode_lib->ms.NoOfDPP[j][s->NumberOfNonCombinedSurfaceOfMaximumBandwidth] = 2;
+                               mode_lib->ms.TotalNumberOfActiveDPP[j] = mode_lib->ms.TotalNumberOfActiveDPP[j] + 1;
+                               mode_lib->ms.TotalNumberOfSingleDPPSurfaces[j] = mode_lib->ms.TotalNumberOfSingleDPPSurfaces[j] - 1;
+                       }
+               }
+
+               //DISPCLK/DPPCLK
+               mode_lib->ms.WritebackRequiredDISPCLK = 0;
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k]) {
+                               mode_lib->ms.WritebackRequiredDISPCLK = dml_max(mode_lib->ms.WritebackRequiredDISPCLK,
+                                                                                                                                       CalculateWriteBackDISPCLK(mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[k],
+                                                                                                                                                                                       mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+                                                                                                                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[k],
+                                                                                                                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[k],
+                                                                                                                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackHTaps[k],
+                                                                                                                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[k],
+                                                                                                                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackSourceWidth[k],
+                                                                                                                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k],
+                                                                                                                                                                                       mode_lib->ms.cache_display_cfg.timing.HTotal[k],
+                                                                                                                                                                                       mode_lib->ms.ip.writeback_line_buffer_buffer_size,
+                                                                                                                                                                                       mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz));
+                       }
+               }
+
+                mode_lib->ms.RequiredDISPCLK[j] = mode_lib->ms.WritebackRequiredDISPCLK;
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       mode_lib->ms.RequiredDISPCLK[j] = dml_max(mode_lib->ms.RequiredDISPCLK[j], mode_lib->ms.RequiredDISPCLKPerSurface[j][k]);
+               }
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       mode_lib->ms.NoOfDPPThisState[k] = mode_lib->ms.NoOfDPP[j][k];
+               }
+
+               CalculateDPPCLK(mode_lib->ms.num_active_planes,
+                                       mode_lib->ms.soc.dcn_downspread_percent,
+                                       mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz,
+                                       mode_lib->ms.MinDPPCLKUsingSingleDPP,
+                                       mode_lib->ms.NoOfDPPThisState,
+                                       /* Output */
+                                       &mode_lib->ms.GlobalDPPCLK,
+                                       mode_lib->ms.RequiredDPPCLKThisState);
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       mode_lib->ms.RequiredDPPCLKPerSurface[j][k] = mode_lib->ms.RequiredDPPCLKThisState[k];
+               }
+
+               mode_lib->ms.support.DISPCLK_DPPCLK_Support[j] = !((mode_lib->ms.RequiredDISPCLK[j] > mode_lib->ms.state.dispclk_mhz) || (mode_lib->ms.GlobalDPPCLK > mode_lib->ms.state.dppclk_mhz));
+
+               if (mode_lib->ms.TotalNumberOfActiveDPP[j] > (dml_uint_t) mode_lib->ms.ip.max_num_dpp) {
+                       mode_lib->ms.support.TotalAvailablePipesSupport[j] = false;
+               }
+       } // j
+
+       /* Total Available OTG, HDMIFRL, DP Support Check */
+       s->TotalNumberOfActiveOTG = 0;
+       s->TotalNumberOfActiveHDMIFRL = 0;
+       s->TotalNumberOfActiveDP2p0 = 0;
+       s->TotalNumberOfActiveDP2p0Outputs = 0;
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k) {
+                       s->TotalNumberOfActiveOTG = s->TotalNumberOfActiveOTG + 1;
+                       if (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl)
+                               s->TotalNumberOfActiveHDMIFRL = s->TotalNumberOfActiveHDMIFRL + 1;
+                       if (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp2p0) {
+                               s->TotalNumberOfActiveDP2p0 = s->TotalNumberOfActiveDP2p0 + 1;
+                               if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[k] == k || mode_lib->ms.cache_display_cfg.output.OutputMultistreamEn[k] == false) {
+                                       s->TotalNumberOfActiveDP2p0Outputs = s->TotalNumberOfActiveDP2p0Outputs + 1;
+                               }
+                       }
+               }
+       }
+
+       mode_lib->ms.support.NumberOfOTGSupport      = (s->TotalNumberOfActiveOTG <= (dml_uint_t) mode_lib->ms.ip.max_num_otg);
+       mode_lib->ms.support.NumberOfHDMIFRLSupport  = (s->TotalNumberOfActiveHDMIFRL <= (dml_uint_t) mode_lib->ms.ip.max_num_hdmi_frl_outputs);
+       mode_lib->ms.support.NumberOfDP2p0Support    = (s->TotalNumberOfActiveDP2p0 <= (dml_uint_t) mode_lib->ms.ip.max_num_dp2p0_streams && s->TotalNumberOfActiveDP2p0Outputs <= (dml_uint_t) mode_lib->ms.ip.max_num_dp2p0_outputs);
+
+       /* Display IO and DSC Support Check */
+       mode_lib->ms.support.NonsupportedDSCInputBPC = false;
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               if (mode_lib->ms.cache_display_cfg.output.OutputDisabled[k] == false &&
+                       !(mode_lib->ms.cache_display_cfg.output.DSCInputBitPerComponent[k] == 12.0
+                               || mode_lib->ms.cache_display_cfg.output.DSCInputBitPerComponent[k] == 10.0
+                               || mode_lib->ms.cache_display_cfg.output.DSCInputBitPerComponent[k] == 8.0
+                               || mode_lib->ms.cache_display_cfg.output.DSCInputBitPerComponent[k] > (dml_uint_t) mode_lib->ms.ip.maximum_dsc_bits_per_component
+                               )) {
+                       mode_lib->ms.support.NonsupportedDSCInputBPC = true;
+               }
+       }
+
+       mode_lib->ms.support.ExceededMultistreamSlots = false;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[k] == k) {
+                       s->TotalSlots = mode_lib->ms.RequiredSlots[k];
+                       for (j = 0; j < mode_lib->ms.num_active_planes; ++j) {
+                               if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[j] == k)
+                                       s->TotalSlots = s->TotalSlots + mode_lib->ms.RequiredSlots[j];
+                       }
+                       if (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp && s->TotalSlots > 63)
+                               mode_lib->ms.support.ExceededMultistreamSlots = true;
+                       if (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp2p0 && s->TotalSlots > 64)
+                               mode_lib->ms.support.ExceededMultistreamSlots = true;
+               }
+       }
+       mode_lib->ms.support.LinkCapacitySupport = true;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.cache_display_cfg.output.OutputDisabled[k] == false &&
+                       mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k && (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp2p0 || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_edp ||
+                       mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmi || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl) && mode_lib->ms.OutputBppPerState[k] == 0) {
+                       mode_lib->ms.support.LinkCapacitySupport = false;
+               }
+       }
+
+       mode_lib->ms.support.P2IWith420 = false;
+       mode_lib->ms.support.DSCOnlyIfNecessaryWithBPP = false;
+       mode_lib->ms.support.DSC422NativeNotSupported = false;
+       mode_lib->ms.support.LinkRateDoesNotMatchDPVersion = false;
+       mode_lib->ms.support.LinkRateForMultistreamNotIndicated = false;
+       mode_lib->ms.support.BPPForMultistreamNotIndicated = false;
+       mode_lib->ms.support.MultistreamWithHDMIOreDP = false;
+       mode_lib->ms.support.MSOOrODMSplitWithNonDPLink = false;
+       mode_lib->ms.support.NotEnoughLanesForMSO = false;
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k && (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp2p0 || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_edp ||
+                                                                                                               mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmi || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl)) {
+                       if (mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_420 && mode_lib->ms.cache_display_cfg.timing.Interlace[k] == 1 && mode_lib->ms.ip.ptoi_supported == true)
+                               mode_lib->ms.support.P2IWith420 = true;
+
+                       if (mode_lib->ms.cache_display_cfg.output.DSCEnable[k] == dml_dsc_enable_if_necessary && mode_lib->ms.cache_display_cfg.output.ForcedOutputLinkBPP[k] != 0)
+                               mode_lib->ms.support.DSCOnlyIfNecessaryWithBPP = true;
+                       if ((mode_lib->ms.cache_display_cfg.output.DSCEnable[k] == dml_dsc_enable || mode_lib->ms.cache_display_cfg.output.DSCEnable[k] == dml_dsc_enable_if_necessary) && mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_n422 && !mode_lib->ms.ip.dsc422_native_support)
+                               mode_lib->ms.support.DSC422NativeNotSupported = true;
+
+                       if (((mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k] == dml_dp_rate_hbr || mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k] == dml_dp_rate_hbr2 || mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k] == dml_dp_rate_hbr3) &&
+                                       mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] != dml_dp && mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] != dml_edp) ||
+                                       ((mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k] == dml_dp_rate_uhbr10 || mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k] == dml_dp_rate_uhbr13p5 || mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k] == dml_dp_rate_uhbr20) &&
+                                       mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] != dml_dp2p0))
+                               mode_lib->ms.support.LinkRateDoesNotMatchDPVersion = true;
+
+                       if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamEn[k] == 1) {
+                               if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[k] == k && mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k] == dml_dp_rate_na)
+                                       mode_lib->ms.support.LinkRateForMultistreamNotIndicated = true;
+                               if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[k] == k && mode_lib->ms.cache_display_cfg.output.ForcedOutputLinkBPP[k] == 0)
+                                       mode_lib->ms.support.BPPForMultistreamNotIndicated = true;
+                               for (j = 0; j < mode_lib->ms.num_active_planes; ++j) {
+                                       if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[k] == j && mode_lib->ms.cache_display_cfg.output.ForcedOutputLinkBPP[k] == 0)
+                                               mode_lib->ms.support.BPPForMultistreamNotIndicated = true;
+                               }
+                       }
+
+                       if ((mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_edp || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmi || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl)) {
+                               if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamEn[k] == 1 && mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[k] == k)
+                                       mode_lib->ms.support.MultistreamWithHDMIOreDP = true;
+                               for (j = 0; j < mode_lib->ms.num_active_planes; ++j) {
+                                       if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamEn[k] == 1 && mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[k] == j)
+                                               mode_lib->ms.support.MultistreamWithHDMIOreDP = true;
+                               }
+                       }
+                       if (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] != dml_dp && (mode_lib->ms.policy.ODMUse[k] == dml_odm_use_policy_split_1to2 ||
+                               mode_lib->ms.policy.ODMUse[k] == dml_odm_use_policy_mso_1to2 || mode_lib->ms.policy.ODMUse[k] == dml_odm_use_policy_mso_1to4))
+                               mode_lib->ms.support.MSOOrODMSplitWithNonDPLink = true;
+
+                       if ((mode_lib->ms.policy.ODMUse[k] == dml_odm_use_policy_mso_1to2 && mode_lib->ms.cache_display_cfg.output.OutputLinkDPLanes[k] < 2) ||
+                               (mode_lib->ms.policy.ODMUse[k] == dml_odm_use_policy_mso_1to4 && mode_lib->ms.cache_display_cfg.output.OutputLinkDPLanes[k] < 4))
+                               mode_lib->ms.support.NotEnoughLanesForMSO = true;
+               }
+       }
+
+       mode_lib->ms.support.DTBCLKRequiredMoreThanSupported = false;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k &&
+                               mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl &&
+                               RequiredDTBCLK(
+                                                       mode_lib->ms.RequiresDSC[k],
+                                                       mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k],
+                                                       mode_lib->ms.cache_display_cfg.output.OutputFormat[k],
+                                                       mode_lib->ms.OutputBppPerState[k],
+                                                       mode_lib->ms.support.NumberOfDSCSlices[k],
+                                                       mode_lib->ms.cache_display_cfg.timing.HTotal[k],
+                                                       mode_lib->ms.cache_display_cfg.timing.HActive[k],
+                                                       mode_lib->ms.cache_display_cfg.output.AudioSampleRate[k],
+                                                       mode_lib->ms.cache_display_cfg.output.AudioSampleLayout[k]) > mode_lib->ms.state.dtbclk_mhz) {
+                                                               mode_lib->ms.support.DTBCLKRequiredMoreThanSupported = true;
+                                                       }
+       }
+
+       mode_lib->ms.support.ODMCombineTwoToOneSupportCheckOK = true;
+       mode_lib->ms.support.ODMCombineFourToOneSupportCheckOK = true;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k && mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_2to1 && mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmi) {
+                       mode_lib->ms.support.ODMCombineTwoToOneSupportCheckOK = false;
+               }
+               if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k && mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_4to1 && (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp ||
+                       mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_edp || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmi)) {
+                       mode_lib->ms.support.ODMCombineFourToOneSupportCheckOK = false;
+               }
+       }
+
+       mode_lib->ms.support.DSCCLKRequiredMoreThanSupported = false;
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k) {
+                       if (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp ||
+                               mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp2p0 ||
+                               mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_edp ||
+                               mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl) {
+                               if (mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_420) {
+                                       s->DSCFormatFactor = 2;
+                               } else if (mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_444) {
+                                       s->DSCFormatFactor = 1;
+                               } else if (mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_n422 || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl) {
+                                       s->DSCFormatFactor = 2;
+                               } else {
+                                       s->DSCFormatFactor = 1;
+                               }
+#ifdef __DML_VBA_DEBUG__
+                               dml_print("DML::%s: k=%u, RequiresDSC = %u\n",  __func__, k, mode_lib->ms.RequiresDSC[k]);
+#endif
+                               if (mode_lib->ms.RequiresDSC[k] == true) {
+                                       if (mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_4to1) {
+                                               if (mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] / 12.0 / (dml_float_t)s->DSCFormatFactor > (1.0 - mode_lib->ms.soc.dcn_downspread_percent / 100.0) * mode_lib->ms.state.dscclk_mhz) {
+#ifdef __DML_VBA_DEBUG__
+                                                       dml_print("DML::%s: k=%u, PixelClockBackEnd     = %f\n",  __func__, k, mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k]);
+                                                       dml_print("DML::%s: k=%u, DSCCLKPerState        = %f\n",  __func__, k, mode_lib->ms.state.dscclk_mhz);
+                                                       dml_print("DML::%s: k=%u, DSCFormatFactor       = %u\n",  __func__, k, s->DSCFormatFactor);
+#endif
+                                                       mode_lib->ms.support.DSCCLKRequiredMoreThanSupported = true;
+                                               }
+                                       } else if (mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_2to1) {
+                                               if (mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] / 6.0 / (dml_float_t)s->DSCFormatFactor > (1.0 - mode_lib->ms.soc.dcn_downspread_percent / 100.0) * mode_lib->ms.state.dscclk_mhz) {
+                                                       mode_lib->ms.support.DSCCLKRequiredMoreThanSupported = true;
+                                               }
+                                       } else {
+                                               if (mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] / 3.0 / (dml_float_t)s->DSCFormatFactor > (1.0 - mode_lib->ms.soc.dcn_downspread_percent / 100.0) * mode_lib->ms.state.dscclk_mhz) {
+                                                       mode_lib->ms.support.DSCCLKRequiredMoreThanSupported = true;
+                                               }
+                                       }
+                               }
+                       }
+               }
+       }
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: DSCCLKRequiredMoreThanSupported = %u\n",  __func__, mode_lib->ms.support.DSCCLKRequiredMoreThanSupported);
+#endif
+
+       /* Check DSC Unit and Slices Support */
+       mode_lib->ms.support.NotEnoughDSCUnits = false;
+       mode_lib->ms.support.NotEnoughDSCSlices = false;
+       s->TotalDSCUnitsRequired = 0;
+       mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport = true;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.RequiresDSC[k] == true) {
+                       if (mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_4to1) {
+                               if (mode_lib->ms.cache_display_cfg.timing.HActive[k] > 4 * (dml_uint_t) mode_lib->ms.ip.maximum_pixels_per_line_per_dsc_unit)
+                                       mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport = false;
+                               s->TotalDSCUnitsRequired = s->TotalDSCUnitsRequired + 4;
+                               if (mode_lib->ms.support.NumberOfDSCSlices[k] > 16)
+                                       mode_lib->ms.support.NotEnoughDSCSlices = true;
+                       } else if (mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_2to1) {
+                               if (mode_lib->ms.cache_display_cfg.timing.HActive[k] > 2 * (dml_uint_t) mode_lib->ms.ip.maximum_pixels_per_line_per_dsc_unit)
+                                       mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport = false;
+                               s->TotalDSCUnitsRequired = s->TotalDSCUnitsRequired + 2;
+                               if (mode_lib->ms.support.NumberOfDSCSlices[k] > 8)
+                                       mode_lib->ms.support.NotEnoughDSCSlices = true;
+                       } else {
+                               if (mode_lib->ms.cache_display_cfg.timing.HActive[k] > (dml_uint_t) mode_lib->ms.ip.maximum_pixels_per_line_per_dsc_unit)
+                                       mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport = false;
+                               s->TotalDSCUnitsRequired = s->TotalDSCUnitsRequired + 1;
+                               if (mode_lib->ms.support.NumberOfDSCSlices[k] > 4)
+                                       mode_lib->ms.support.NotEnoughDSCSlices = true;
+                       }
+               }
+       }
+   if (s->TotalDSCUnitsRequired > (dml_uint_t) mode_lib->ms.ip.num_dsc) {
+               mode_lib->ms.support.NotEnoughDSCUnits = true;
+       }
+
+       /*DSC Delay per state*/
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               mode_lib->ms.DSCDelayPerState[k] = DSCDelayRequirement(mode_lib->ms.RequiresDSC[k],
+                                                                                                       mode_lib->ms.ODMModePerState[k],
+                                                                                                       mode_lib->ms.cache_display_cfg.output.DSCInputBitPerComponent[k],
+                                                                                                       mode_lib->ms.OutputBppPerState[k],
+                                                                                                       mode_lib->ms.cache_display_cfg.timing.HActive[k],
+                                                                                                       mode_lib->ms.cache_display_cfg.timing.HTotal[k],
+                                                                                                       mode_lib->ms.support.NumberOfDSCSlices[k],
+                                                                                                       mode_lib->ms.cache_display_cfg.output.OutputFormat[k],
+                                                                                                       mode_lib->ms.cache_display_cfg.output.OutputEncoder[k],
+                                                                                                       mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+                                                                                                       mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k]);
+       }
+
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               for (m = 0; m <= mode_lib->ms.num_active_planes - 1; m++) {
+                       for (j = 0; j <= mode_lib->ms.num_active_planes - 1; j++) {
+                               if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == m && mode_lib->ms.RequiresDSC[m] == true) {
+                                       mode_lib->ms.DSCDelayPerState[k] = mode_lib->ms.DSCDelayPerState[m];
+                               }
+                       }
+               }
+       }
+
+       //Calculate Swath, DET Configuration, DCFCLKDeepSleep
+       //
+       for (j = 0; j < 2; ++j) {
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       mode_lib->ms.RequiredDPPCLKThisState[k] = mode_lib->ms.RequiredDPPCLKPerSurface[j][k];
+                       mode_lib->ms.NoOfDPPThisState[k] = mode_lib->ms.NoOfDPP[j][k];
+                       mode_lib->ms.ODMModeThisState[k] = mode_lib->ms.ODMModePerState[k];
+               }
+
+               CalculateSwathAndDETConfiguration_params->DETSizeOverride = mode_lib->ms.cache_display_cfg.plane.DETSizeOverride;
+               CalculateSwathAndDETConfiguration_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+               CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSizeInKByte = mode_lib->ms.ip.config_return_buffer_size_in_kbytes;
+               CalculateSwathAndDETConfiguration_params->ROBBufferSizeInKByte = mode_lib->ms.ip.rob_buffer_size_kbytes;
+               CalculateSwathAndDETConfiguration_params->MaxTotalDETInKByte = mode_lib->ms.MaxTotalDETInKByte;
+               CalculateSwathAndDETConfiguration_params->MinCompressedBufferSizeInKByte = mode_lib->ms.MinCompressedBufferSizeInKByte;
+               CalculateSwathAndDETConfiguration_params->PixelChunkSizeInKByte = mode_lib->ms.ip.pixel_chunk_size_kbytes;
+               CalculateSwathAndDETConfiguration_params->ForceSingleDPP = false;
+               CalculateSwathAndDETConfiguration_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+               CalculateSwathAndDETConfiguration_params->nomDETInKByte = mode_lib->ms.NomDETInKByte;
+               CalculateSwathAndDETConfiguration_params->UseUnboundedRequestingFinal = mode_lib->ms.policy.UseUnboundedRequesting;
+               CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSegmentSizeInkByte = mode_lib->ms.ip.config_return_buffer_segment_size_in_kbytes;
+               CalculateSwathAndDETConfiguration_params->CompressedBufferSegmentSizeInkByteFinal = mode_lib->ms.ip.compressed_buffer_segment_size_in_kbytes;
+               CalculateSwathAndDETConfiguration_params->Output = mode_lib->ms.cache_display_cfg.output.OutputEncoder;
+               CalculateSwathAndDETConfiguration_params->ReadBandwidthLuma = mode_lib->ms.ReadBandwidthLuma;
+               CalculateSwathAndDETConfiguration_params->ReadBandwidthChroma = mode_lib->ms.ReadBandwidthChroma;
+               CalculateSwathAndDETConfiguration_params->MaximumSwathWidthLuma = mode_lib->ms.MaximumSwathWidthLuma;
+               CalculateSwathAndDETConfiguration_params->MaximumSwathWidthChroma = mode_lib->ms.MaximumSwathWidthChroma;
+               CalculateSwathAndDETConfiguration_params->SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan;
+               CalculateSwathAndDETConfiguration_params->ViewportStationary = mode_lib->ms.cache_display_cfg.plane.ViewportStationary;
+               CalculateSwathAndDETConfiguration_params->SourcePixelFormat = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat;
+               CalculateSwathAndDETConfiguration_params->SurfaceTiling = mode_lib->ms.cache_display_cfg.surface.SurfaceTiling;
+               CalculateSwathAndDETConfiguration_params->ViewportWidth = mode_lib->ms.cache_display_cfg.plane.ViewportWidth;
+               CalculateSwathAndDETConfiguration_params->ViewportHeight = mode_lib->ms.cache_display_cfg.plane.ViewportHeight;
+               CalculateSwathAndDETConfiguration_params->ViewportXStart = mode_lib->ms.cache_display_cfg.plane.ViewportXStart;
+               CalculateSwathAndDETConfiguration_params->ViewportYStart = mode_lib->ms.cache_display_cfg.plane.ViewportYStart;
+               CalculateSwathAndDETConfiguration_params->ViewportXStartC = mode_lib->ms.cache_display_cfg.plane.ViewportXStartC;
+               CalculateSwathAndDETConfiguration_params->ViewportYStartC = mode_lib->ms.cache_display_cfg.plane.ViewportYStartC;
+               CalculateSwathAndDETConfiguration_params->SurfaceWidthY = mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY;
+               CalculateSwathAndDETConfiguration_params->SurfaceWidthC = mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC;
+               CalculateSwathAndDETConfiguration_params->SurfaceHeightY = mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY;
+               CalculateSwathAndDETConfiguration_params->SurfaceHeightC = mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC;
+               CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightY = mode_lib->ms.Read256BlockHeightY;
+               CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightC = mode_lib->ms.Read256BlockHeightC;
+               CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthY = mode_lib->ms.Read256BlockWidthY;
+               CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthC = mode_lib->ms.Read256BlockWidthC;
+               CalculateSwathAndDETConfiguration_params->ODMMode = mode_lib->ms.ODMModeThisState;
+               CalculateSwathAndDETConfiguration_params->BlendingAndTiming = mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming;
+               CalculateSwathAndDETConfiguration_params->BytePerPixY = mode_lib->ms.BytePerPixelY;
+               CalculateSwathAndDETConfiguration_params->BytePerPixC = mode_lib->ms.BytePerPixelC;
+               CalculateSwathAndDETConfiguration_params->BytePerPixDETY = mode_lib->ms.BytePerPixelInDETY;
+               CalculateSwathAndDETConfiguration_params->BytePerPixDETC = mode_lib->ms.BytePerPixelInDETC;
+               CalculateSwathAndDETConfiguration_params->HActive = mode_lib->ms.cache_display_cfg.timing.HActive;
+               CalculateSwathAndDETConfiguration_params->HRatio = mode_lib->ms.cache_display_cfg.plane.HRatio;
+               CalculateSwathAndDETConfiguration_params->HRatioChroma = mode_lib->ms.cache_display_cfg.plane.HRatioChroma;
+               CalculateSwathAndDETConfiguration_params->DPPPerSurface = mode_lib->ms.NoOfDPPThisState;
+               CalculateSwathAndDETConfiguration_params->swath_width_luma_ub = mode_lib->ms.swath_width_luma_ub_this_state;
+               CalculateSwathAndDETConfiguration_params->swath_width_chroma_ub = mode_lib->ms.swath_width_chroma_ub_this_state;
+               CalculateSwathAndDETConfiguration_params->SwathWidth = mode_lib->ms.SwathWidthYThisState;
+               CalculateSwathAndDETConfiguration_params->SwathWidthChroma = mode_lib->ms.SwathWidthCThisState;
+               CalculateSwathAndDETConfiguration_params->SwathHeightY = mode_lib->ms.SwathHeightYThisState;
+               CalculateSwathAndDETConfiguration_params->SwathHeightC = mode_lib->ms.SwathHeightCThisState;
+               CalculateSwathAndDETConfiguration_params->DETBufferSizeInKByte = mode_lib->ms.DETBufferSizeInKByteThisState;
+               CalculateSwathAndDETConfiguration_params->DETBufferSizeY = mode_lib->ms.DETBufferSizeYThisState;
+               CalculateSwathAndDETConfiguration_params->DETBufferSizeC = mode_lib->ms.DETBufferSizeCThisState;
+               CalculateSwathAndDETConfiguration_params->UnboundedRequestEnabled = &mode_lib->ms.UnboundedRequestEnabledThisState;
+               CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_64b = &s->dummy_integer[2];
+               CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_zs = &s->dummy_integer[1];
+               CalculateSwathAndDETConfiguration_params->CompressedBufferSizeInkByte = &mode_lib->ms.CompressedBufferSizeInkByteThisState;
+               CalculateSwathAndDETConfiguration_params->ViewportSizeSupportPerSurface = s->dummy_boolean_array[0];
+               CalculateSwathAndDETConfiguration_params->ViewportSizeSupport = &mode_lib->ms.support.ViewportSizeSupport[j];
+
+               CalculateSwathAndDETConfiguration(&mode_lib->scratch,
+               CalculateSwathAndDETConfiguration_params);
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       mode_lib->ms.swath_width_luma_ub_all_states[j][k] = mode_lib->ms.swath_width_luma_ub_this_state[k];
+                       mode_lib->ms.swath_width_chroma_ub_all_states[j][k] = mode_lib->ms.swath_width_chroma_ub_this_state[k];
+                       mode_lib->ms.SwathWidthYAllStates[j][k] = mode_lib->ms.SwathWidthYThisState[k];
+                       mode_lib->ms.SwathWidthCAllStates[j][k] = mode_lib->ms.SwathWidthCThisState[k];
+                       mode_lib->ms.SwathHeightYAllStates[j][k] = mode_lib->ms.SwathHeightYThisState[k];
+                       mode_lib->ms.SwathHeightCAllStates[j][k] = mode_lib->ms.SwathHeightCThisState[k];
+                       mode_lib->ms.UnboundedRequestEnabledAllStates[j] = mode_lib->ms.UnboundedRequestEnabledThisState;
+                       mode_lib->ms.CompressedBufferSizeInkByteAllStates[j] = mode_lib->ms.CompressedBufferSizeInkByteThisState;
+                       mode_lib->ms.DETBufferSizeInKByteAllStates[j][k] = mode_lib->ms.DETBufferSizeInKByteThisState[k];
+                       mode_lib->ms.DETBufferSizeYAllStates[j][k] = mode_lib->ms.DETBufferSizeYThisState[k];
+                       mode_lib->ms.DETBufferSizeCAllStates[j][k] = mode_lib->ms.DETBufferSizeCThisState[k];
+               }
+       }
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               mode_lib->ms.cursor_bw[k] = mode_lib->ms.cache_display_cfg.plane.NumberOfCursors[k] * mode_lib->ms.cache_display_cfg.plane.CursorWidth[k] * mode_lib->ms.cache_display_cfg.plane.CursorBPP[k] / 8.0 / (mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * mode_lib->ms.cache_display_cfg.plane.VRatio[k];
+       }
+
+       CalculateSurfaceSizeInMall(
+                       mode_lib->ms.num_active_planes,
+                       mode_lib->ms.soc.mall_allocated_for_dcn_mbytes,
+                       mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen,
+                       mode_lib->ms.cache_display_cfg.surface.DCCEnable,
+                       mode_lib->ms.cache_display_cfg.plane.ViewportStationary,
+                       mode_lib->ms.cache_display_cfg.plane.ViewportXStart,
+                       mode_lib->ms.cache_display_cfg.plane.ViewportYStart,
+                       mode_lib->ms.cache_display_cfg.plane.ViewportXStartC,
+                       mode_lib->ms.cache_display_cfg.plane.ViewportYStartC,
+                       mode_lib->ms.cache_display_cfg.plane.ViewportWidth,
+                       mode_lib->ms.cache_display_cfg.plane.ViewportHeight,
+                       mode_lib->ms.BytePerPixelY,
+                       mode_lib->ms.cache_display_cfg.plane.ViewportWidthChroma,
+                       mode_lib->ms.cache_display_cfg.plane.ViewportHeightChroma,
+                       mode_lib->ms.BytePerPixelC,
+                       mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY,
+                       mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC,
+                       mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY,
+                       mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC,
+                       mode_lib->ms.Read256BlockWidthY,
+                       mode_lib->ms.Read256BlockWidthC,
+                       mode_lib->ms.Read256BlockHeightY,
+                       mode_lib->ms.Read256BlockHeightC,
+                       mode_lib->ms.MacroTileWidthY,
+                       mode_lib->ms.MacroTileWidthC,
+                       mode_lib->ms.MacroTileHeightY,
+                       mode_lib->ms.MacroTileHeightC,
+
+                       /* Output */
+                       mode_lib->ms.SurfaceSizeInMALL,
+                       &mode_lib->ms.support.ExceededMALLSize);
+
+       for (j = 0; j < 2; j++) {
+               for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                       mode_lib->ms.swath_width_luma_ub_this_state[k] = mode_lib->ms.swath_width_luma_ub_all_states[j][k];
+                       mode_lib->ms.swath_width_chroma_ub_this_state[k] = mode_lib->ms.swath_width_chroma_ub_all_states[j][k];
+                       mode_lib->ms.SwathWidthYThisState[k] = mode_lib->ms.SwathWidthYAllStates[j][k];
+                       mode_lib->ms.SwathWidthCThisState[k] = mode_lib->ms.SwathWidthCAllStates[j][k];
+                       mode_lib->ms.SwathHeightYThisState[k] = mode_lib->ms.SwathHeightYAllStates[j][k];
+                       mode_lib->ms.SwathHeightCThisState[k] = mode_lib->ms.SwathHeightCAllStates[j][k];
+                       mode_lib->ms.DETBufferSizeInKByteThisState[k] = mode_lib->ms.DETBufferSizeInKByteAllStates[j][k];
+                       mode_lib->ms.DETBufferSizeYThisState[k] = mode_lib->ms.DETBufferSizeYAllStates[j][k];
+                       mode_lib->ms.DETBufferSizeCThisState[k] = mode_lib->ms.DETBufferSizeCAllStates[j][k];
+                       mode_lib->ms.RequiredDPPCLKThisState[k] = mode_lib->ms.RequiredDPPCLKPerSurface[j][k];
+                       mode_lib->ms.NoOfDPPThisState[k] = mode_lib->ms.NoOfDPP[j][k];
+               }
+
+               mode_lib->ms.TotalNumberOfDCCActiveDPP[j] = 0;
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       if (mode_lib->ms.cache_display_cfg.surface.DCCEnable[k] == true) {
+                               mode_lib->ms.TotalNumberOfDCCActiveDPP[j] = mode_lib->ms.TotalNumberOfDCCActiveDPP[j] + mode_lib->ms.NoOfDPP[j][k];
+                       }
+               }
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       s->SurfParameters[k].PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock[k];
+                       s->SurfParameters[k].DPPPerSurface = mode_lib->ms.NoOfDPP[j][k];
+                       s->SurfParameters[k].SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan[k];
+                       s->SurfParameters[k].ViewportHeight = mode_lib->ms.cache_display_cfg.plane.ViewportHeight[k];
+                       s->SurfParameters[k].ViewportHeightChroma = mode_lib->ms.cache_display_cfg.plane.ViewportHeightChroma[k];
+                       s->SurfParameters[k].BlockWidth256BytesY = mode_lib->ms.Read256BlockWidthY[k];
+                       s->SurfParameters[k].BlockHeight256BytesY = mode_lib->ms.Read256BlockHeightY[k];
+                       s->SurfParameters[k].BlockWidth256BytesC = mode_lib->ms.Read256BlockWidthC[k];
+                       s->SurfParameters[k].BlockHeight256BytesC = mode_lib->ms.Read256BlockHeightC[k];
+                       s->SurfParameters[k].BlockWidthY = mode_lib->ms.MacroTileWidthY[k];
+                       s->SurfParameters[k].BlockHeightY = mode_lib->ms.MacroTileHeightY[k];
+                       s->SurfParameters[k].BlockWidthC = mode_lib->ms.MacroTileWidthC[k];
+                       s->SurfParameters[k].BlockHeightC = mode_lib->ms.MacroTileHeightC[k];
+                       s->SurfParameters[k].InterlaceEnable = mode_lib->ms.cache_display_cfg.timing.Interlace[k];
+                       s->SurfParameters[k].HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal[k];
+                       s->SurfParameters[k].DCCEnable = mode_lib->ms.cache_display_cfg.surface.DCCEnable[k];
+                       s->SurfParameters[k].SourcePixelFormat = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k];
+                       s->SurfParameters[k].SurfaceTiling = mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k];
+                       s->SurfParameters[k].BytePerPixelY = mode_lib->ms.BytePerPixelY[k];
+                       s->SurfParameters[k].BytePerPixelC = mode_lib->ms.BytePerPixelC[k];
+                       s->SurfParameters[k].ProgressiveToInterlaceUnitInOPP = mode_lib->ms.ip.ptoi_supported;
+                       s->SurfParameters[k].VRatio = mode_lib->ms.cache_display_cfg.plane.VRatio[k];
+                       s->SurfParameters[k].VRatioChroma = mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k];
+                       s->SurfParameters[k].VTaps = mode_lib->ms.cache_display_cfg.plane.VTaps[k];
+                       s->SurfParameters[k].VTapsChroma = mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k];
+                       s->SurfParameters[k].PitchY = mode_lib->ms.cache_display_cfg.surface.PitchY[k];
+                       s->SurfParameters[k].DCCMetaPitchY = mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchY[k];
+                       s->SurfParameters[k].PitchC = mode_lib->ms.cache_display_cfg.surface.PitchC[k];
+                       s->SurfParameters[k].DCCMetaPitchC = mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchC[k];
+                       s->SurfParameters[k].ViewportStationary = mode_lib->ms.cache_display_cfg.plane.ViewportStationary[k];
+                       s->SurfParameters[k].ViewportXStart = mode_lib->ms.cache_display_cfg.plane.ViewportXStart[k];
+                       s->SurfParameters[k].ViewportYStart = mode_lib->ms.cache_display_cfg.plane.ViewportYStart[k];
+                       s->SurfParameters[k].ViewportXStartC = mode_lib->ms.cache_display_cfg.plane.ViewportXStartC[k];
+                       s->SurfParameters[k].ViewportYStartC = mode_lib->ms.cache_display_cfg.plane.ViewportYStartC[k];
+                       s->SurfParameters[k].FORCE_ONE_ROW_FOR_FRAME = mode_lib->ms.cache_display_cfg.plane.ForceOneRowForFrame[k];
+                       s->SurfParameters[k].SwathHeightY = mode_lib->ms.SwathHeightYThisState[k];
+                       s->SurfParameters[k].SwathHeightC = mode_lib->ms.SwathHeightCThisState[k];
+               }
+
+               CalculateVMRowAndSwath_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+               CalculateVMRowAndSwath_params->myPipe = s->SurfParameters;
+               CalculateVMRowAndSwath_params->SurfaceSizeInMALL = mode_lib->ms.SurfaceSizeInMALL;
+               CalculateVMRowAndSwath_params->PTEBufferSizeInRequestsLuma = mode_lib->ms.ip.dpte_buffer_size_in_pte_reqs_luma;
+               CalculateVMRowAndSwath_params->PTEBufferSizeInRequestsChroma = mode_lib->ms.ip.dpte_buffer_size_in_pte_reqs_chroma;
+               CalculateVMRowAndSwath_params->DCCMetaBufferSizeBytes = mode_lib->ms.ip.dcc_meta_buffer_size_bytes;
+               CalculateVMRowAndSwath_params->UseMALLForStaticScreen = mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen;
+               CalculateVMRowAndSwath_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+               CalculateVMRowAndSwath_params->MALLAllocatedForDCN = mode_lib->ms.soc.mall_allocated_for_dcn_mbytes;
+               CalculateVMRowAndSwath_params->SwathWidthY = mode_lib->ms.SwathWidthYThisState;
+               CalculateVMRowAndSwath_params->SwathWidthC = mode_lib->ms.SwathWidthCThisState;
+               CalculateVMRowAndSwath_params->GPUVMEnable = mode_lib->ms.cache_display_cfg.plane.GPUVMEnable;
+               CalculateVMRowAndSwath_params->HostVMEnable = mode_lib->ms.cache_display_cfg.plane.HostVMEnable;
+               CalculateVMRowAndSwath_params->HostVMMaxNonCachedPageTableLevels = mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels;
+               CalculateVMRowAndSwath_params->GPUVMMaxPageTableLevels = mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels;
+               CalculateVMRowAndSwath_params->GPUVMMinPageSizeKBytes = mode_lib->ms.cache_display_cfg.plane.GPUVMMinPageSizeKBytes;
+               CalculateVMRowAndSwath_params->HostVMMinPageSize = mode_lib->ms.soc.hostvm_min_page_size_kbytes;
+               CalculateVMRowAndSwath_params->PTEBufferModeOverrideEn = mode_lib->ms.cache_display_cfg.plane.PTEBufferModeOverrideEn;
+               CalculateVMRowAndSwath_params->PTEBufferModeOverrideVal = mode_lib->ms.cache_display_cfg.plane.PTEBufferMode;
+               CalculateVMRowAndSwath_params->PTEBufferSizeNotExceeded = mode_lib->ms.PTEBufferSizeNotExceededPerState;
+               CalculateVMRowAndSwath_params->DCCMetaBufferSizeNotExceeded = mode_lib->ms.DCCMetaBufferSizeNotExceededPerState;
+               CalculateVMRowAndSwath_params->dpte_row_width_luma_ub = s->dummy_integer_array[0];
+               CalculateVMRowAndSwath_params->dpte_row_width_chroma_ub = s->dummy_integer_array[1];
+               CalculateVMRowAndSwath_params->dpte_row_height_luma = mode_lib->ms.dpte_row_height;
+               CalculateVMRowAndSwath_params->dpte_row_height_chroma = mode_lib->ms.dpte_row_height_chroma;
+               CalculateVMRowAndSwath_params->dpte_row_height_linear_luma = s->dummy_integer_array[2]; // VBA_DELTA
+               CalculateVMRowAndSwath_params->dpte_row_height_linear_chroma = s->dummy_integer_array[3]; // VBA_DELTA
+               CalculateVMRowAndSwath_params->meta_req_width = s->dummy_integer_array[4];
+               CalculateVMRowAndSwath_params->meta_req_width_chroma = s->dummy_integer_array[5];
+               CalculateVMRowAndSwath_params->meta_req_height = s->dummy_integer_array[6];
+               CalculateVMRowAndSwath_params->meta_req_height_chroma = s->dummy_integer_array[7];
+               CalculateVMRowAndSwath_params->meta_row_width = s->dummy_integer_array[8];
+               CalculateVMRowAndSwath_params->meta_row_width_chroma = s->dummy_integer_array[9];
+               CalculateVMRowAndSwath_params->meta_row_height = mode_lib->ms.meta_row_height;
+               CalculateVMRowAndSwath_params->meta_row_height_chroma = mode_lib->ms.meta_row_height_chroma;
+               CalculateVMRowAndSwath_params->vm_group_bytes = s->dummy_integer_array[10];
+               CalculateVMRowAndSwath_params->dpte_group_bytes = mode_lib->ms.dpte_group_bytes;
+               CalculateVMRowAndSwath_params->PixelPTEReqWidthY = s->dummy_integer_array[11];
+               CalculateVMRowAndSwath_params->PixelPTEReqHeightY = s->dummy_integer_array[12];
+               CalculateVMRowAndSwath_params->PTERequestSizeY = s->dummy_integer_array[13];
+               CalculateVMRowAndSwath_params->PixelPTEReqWidthC = s->dummy_integer_array[14];
+               CalculateVMRowAndSwath_params->PixelPTEReqHeightC = s->dummy_integer_array[15];
+               CalculateVMRowAndSwath_params->PTERequestSizeC = s->dummy_integer_array[16];
+               CalculateVMRowAndSwath_params->dpde0_bytes_per_frame_ub_l = s->dummy_integer_array[17];
+               CalculateVMRowAndSwath_params->meta_pte_bytes_per_frame_ub_l = s->dummy_integer_array[18];
+               CalculateVMRowAndSwath_params->dpde0_bytes_per_frame_ub_c = s->dummy_integer_array[19];
+               CalculateVMRowAndSwath_params->meta_pte_bytes_per_frame_ub_c = s->dummy_integer_array[20];
+               CalculateVMRowAndSwath_params->PrefetchSourceLinesY = mode_lib->ms.PrefetchLinesYThisState;
+               CalculateVMRowAndSwath_params->PrefetchSourceLinesC = mode_lib->ms.PrefetchLinesCThisState;
+               CalculateVMRowAndSwath_params->VInitPreFillY = mode_lib->ms.PrefillY;
+               CalculateVMRowAndSwath_params->VInitPreFillC = mode_lib->ms.PrefillC;
+               CalculateVMRowAndSwath_params->MaxNumSwathY = mode_lib->ms.MaxNumSwY;
+               CalculateVMRowAndSwath_params->MaxNumSwathC = mode_lib->ms.MaxNumSwC;
+               CalculateVMRowAndSwath_params->meta_row_bw = mode_lib->ms.meta_row_bandwidth_this_state;
+               CalculateVMRowAndSwath_params->dpte_row_bw = mode_lib->ms.dpte_row_bandwidth_this_state;
+               CalculateVMRowAndSwath_params->PixelPTEBytesPerRow = mode_lib->ms.DPTEBytesPerRowThisState;
+               CalculateVMRowAndSwath_params->PDEAndMetaPTEBytesFrame = mode_lib->ms.PDEAndMetaPTEBytesPerFrameThisState;
+               CalculateVMRowAndSwath_params->MetaRowByte = mode_lib->ms.MetaRowBytesThisState;
+               CalculateVMRowAndSwath_params->use_one_row_for_frame = mode_lib->ms.use_one_row_for_frame_this_state;
+               CalculateVMRowAndSwath_params->use_one_row_for_frame_flip = mode_lib->ms.use_one_row_for_frame_flip_this_state;
+               CalculateVMRowAndSwath_params->UsesMALLForStaticScreen = s->dummy_boolean_array[0];
+               CalculateVMRowAndSwath_params->PTE_BUFFER_MODE = s->dummy_boolean_array[1];
+               CalculateVMRowAndSwath_params->BIGK_FRAGMENT_SIZE = s->dummy_integer_array[21];
+
+               CalculateVMRowAndSwath(&mode_lib->scratch,
+                       CalculateVMRowAndSwath_params);
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       mode_lib->ms.PrefetchLinesY[j][k] = mode_lib->ms.PrefetchLinesYThisState[k];
+                       mode_lib->ms.PrefetchLinesC[j][k] = mode_lib->ms.PrefetchLinesCThisState[k];
+                       mode_lib->ms.meta_row_bandwidth[j][k] = mode_lib->ms.meta_row_bandwidth_this_state[k];
+                       mode_lib->ms.dpte_row_bandwidth[j][k] = mode_lib->ms.dpte_row_bandwidth_this_state[k];
+                       mode_lib->ms.DPTEBytesPerRow[j][k] = mode_lib->ms.DPTEBytesPerRowThisState[k];
+                       mode_lib->ms.PDEAndMetaPTEBytesPerFrame[j][k] = mode_lib->ms.PDEAndMetaPTEBytesPerFrameThisState[k];
+                       mode_lib->ms.MetaRowBytes[j][k] = mode_lib->ms.MetaRowBytesThisState[k];
+                       mode_lib->ms.use_one_row_for_frame[j][k] = mode_lib->ms.use_one_row_for_frame_this_state[k];
+                       mode_lib->ms.use_one_row_for_frame_flip[j][k] = mode_lib->ms.use_one_row_for_frame_flip_this_state[k];
+               }
+
+               mode_lib->ms.support.PTEBufferSizeNotExceeded[j] = true;
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       if (mode_lib->ms.PTEBufferSizeNotExceededPerState[k] == false)
+                               mode_lib->ms.support.PTEBufferSizeNotExceeded[j] = false;
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: j=%u k=%u, PTEBufferSizeNotExceededPerState[%u] = %u\n",  __func__, j, k, k, mode_lib->ms.PTEBufferSizeNotExceededPerState[k]);
+#endif
+               }
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: PTEBufferSizeNotExceeded[%u] = %u\n",  __func__, j, mode_lib->ms.support.PTEBufferSizeNotExceeded[j]);
+#endif
+
+               mode_lib->ms.support.DCCMetaBufferSizeNotExceeded[j] = true;
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       if (mode_lib->ms.DCCMetaBufferSizeNotExceededPerState[k] == false)
+                               mode_lib->ms.support.DCCMetaBufferSizeNotExceeded[j] = false;
+               }
+
+               mode_lib->ms.UrgLatency = CalculateUrgentLatency(mode_lib->ms.state.urgent_latency_pixel_data_only_us,
+                                                                                                       mode_lib->ms.state.urgent_latency_pixel_mixed_with_vm_data_us,
+                                                                                                       mode_lib->ms.state.urgent_latency_vm_data_only_us,
+                                                                                                       mode_lib->ms.soc.do_urgent_latency_adjustment,
+                                                                                                       mode_lib->ms.state.urgent_latency_adjustment_fabric_clock_component_us,
+                                                                                                       mode_lib->ms.state.urgent_latency_adjustment_fabric_clock_reference_mhz,
+                                                                                                       mode_lib->ms.state.fabricclk_mhz);
+
+               /* Getter functions work at mp interface so copy the urgent latency to mp*/
+               mode_lib->mp.UrgentLatency = mode_lib->ms.UrgLatency;
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       CalculateUrgentBurstFactor(
+                               mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k],
+                               mode_lib->ms.swath_width_luma_ub_this_state[k],
+                               mode_lib->ms.swath_width_chroma_ub_this_state[k],
+                               mode_lib->ms.SwathHeightYThisState[k],
+                               mode_lib->ms.SwathHeightCThisState[k],
+                               (dml_float_t) mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+                               mode_lib->ms.UrgLatency,
+                               mode_lib->ms.ip.cursor_buffer_size,
+                               mode_lib->ms.cache_display_cfg.plane.CursorWidth[k],
+                               mode_lib->ms.cache_display_cfg.plane.CursorBPP[k],
+                               mode_lib->ms.cache_display_cfg.plane.VRatio[k],
+                               mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k],
+                               mode_lib->ms.BytePerPixelInDETY[k],
+                               mode_lib->ms.BytePerPixelInDETC[k],
+                               mode_lib->ms.DETBufferSizeYThisState[k],
+                               mode_lib->ms.DETBufferSizeCThisState[k],
+                               /* Output */
+                               &mode_lib->ms.UrgentBurstFactorCursor[k],
+                               &mode_lib->ms.UrgentBurstFactorLuma[k],
+                               &mode_lib->ms.UrgentBurstFactorChroma[k],
+                               &mode_lib->ms.NotUrgentLatencyHiding[k]);
+               }
+
+               CalculateDCFCLKDeepSleep(
+                               mode_lib->ms.num_active_planes,
+                               mode_lib->ms.BytePerPixelY,
+                               mode_lib->ms.BytePerPixelC,
+                               mode_lib->ms.cache_display_cfg.plane.VRatio,
+                               mode_lib->ms.cache_display_cfg.plane.VRatioChroma,
+                               mode_lib->ms.SwathWidthYThisState,
+                               mode_lib->ms.SwathWidthCThisState,
+                               mode_lib->ms.NoOfDPPThisState,
+                               mode_lib->ms.cache_display_cfg.plane.HRatio,
+                               mode_lib->ms.cache_display_cfg.plane.HRatioChroma,
+                               mode_lib->ms.cache_display_cfg.timing.PixelClock,
+                               mode_lib->ms.PSCL_FACTOR,
+                               mode_lib->ms.PSCL_FACTOR_CHROMA,
+                               mode_lib->ms.RequiredDPPCLKThisState,
+                               mode_lib->ms.ReadBandwidthLuma,
+                               mode_lib->ms.ReadBandwidthChroma,
+                               mode_lib->ms.soc.return_bus_width_bytes,
+
+                               /* Output */
+                               &mode_lib->ms.ProjectedDCFCLKDeepSleep[j]);
+       }
+
+       //Calculate Return BW
+       for (j = 0; j < 2; ++j) {
+               for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                       if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k) {
+                               if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true) {
+                                       mode_lib->ms.WritebackDelayTime[k] = mode_lib->ms.state.writeback_latency_us + CalculateWriteBackDelay(
+                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[k],
+                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[k],
+                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[k],
+                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[k],
+                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k],
+                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[k],
+                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[k],
+                                                                       mode_lib->ms.cache_display_cfg.timing.HTotal[k]) / mode_lib->ms.RequiredDISPCLK[j];
+                               } else {
+                                       mode_lib->ms.WritebackDelayTime[k] = 0.0;
+                               }
+                               for (m = 0; m <= mode_lib->ms.num_active_planes - 1; m++) {
+                                       if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[m] == k && mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[m] == true) {
+                                               mode_lib->ms.WritebackDelayTime[k] = dml_max(mode_lib->ms.WritebackDelayTime[k],
+                                                                                       mode_lib->ms.state.writeback_latency_us + CalculateWriteBackDelay(
+                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[m],
+                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[m],
+                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[m],
+                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[m],
+                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[m],
+                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[m],
+                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[m],
+                                                                                       mode_lib->ms.cache_display_cfg.timing.HTotal[m]) / mode_lib->ms.RequiredDISPCLK[j]);
+                                       }
+                               }
+                       }
+               }
+               for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                       for (m = 0; m <= mode_lib->ms.num_active_planes - 1; m++) {
+                               if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == m) {
+                                       mode_lib->ms.WritebackDelayTime[k] = mode_lib->ms.WritebackDelayTime[m];
+                               }
+                       }
+               }
+       s->MaxVStartupAllPlanes[j] = 0;  // max vstartup among all planes
+
+               for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                       s->MaximumVStartup[j][k] = CalculateMaxVStartup(k,
+                                                                                                               mode_lib->ms.ip.ptoi_supported,
+                                                                                                               mode_lib->ms.ip.vblank_nom_default_us,
+                                                                                                               &mode_lib->ms.cache_display_cfg.timing,
+                                                                                                               mode_lib->ms.WritebackDelayTime[k]);
+
+                       s->MaxVStartupAllPlanes[j] = (dml_uint_t)(dml_max(s->MaxVStartupAllPlanes[j], s->MaximumVStartup[j][k]));
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%u, MaxVStartupAllPlanes[%u] = %u\n", __func__, k, j, s->MaxVStartupAllPlanes[j]);
+                       dml_print("DML::%s: k=%u, MaximumVStartup[%u][%u] = %u\n", __func__, k, j, k, s->MaximumVStartup[j][k]);
+#endif
+               }
+       }
+
+       s->ReorderingBytes = (dml_uint_t)(mode_lib->ms.soc.num_chans * dml_max3(mode_lib->ms.soc.urgent_out_of_order_return_per_channel_pixel_only_bytes,
+                                                                                                                               mode_lib->ms.soc.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes,
+                                                                                                                               mode_lib->ms.soc.urgent_out_of_order_return_per_channel_vm_only_bytes));
+
+       for (j = 0; j < 2; ++j) {
+               mode_lib->ms.DCFCLKState[j] = mode_lib->ms.state.dcfclk_mhz;
+       }
+
+       /* Immediate Flip and MALL parameters */
+       s->ImmediateFlipRequiredFinal = false;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               s->ImmediateFlipRequiredFinal = s->ImmediateFlipRequiredFinal || (mode_lib->ms.policy.ImmediateFlipRequirement[k] == dml_immediate_flip_required);
+       }
+
+       mode_lib->ms.support.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified = false;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               mode_lib->ms.support.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified = mode_lib->ms.support.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified ||
+                                                                                                                                                                                       ((mode_lib->ms.policy.ImmediateFlipRequirement[k] != dml_immediate_flip_required) &&
+                                                                                                                                                                                       (mode_lib->ms.policy.ImmediateFlipRequirement[k] != dml_immediate_flip_not_required));
+       }
+       mode_lib->ms.support.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified = mode_lib->ms.support.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified && s->ImmediateFlipRequiredFinal;
+
+       mode_lib->ms.support.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe = false;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               mode_lib->ms.support.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe =
+                                                                               mode_lib->ms.support.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe || ((mode_lib->ms.cache_display_cfg.plane.HostVMEnable == true || mode_lib->ms.policy.ImmediateFlipRequirement[k] != dml_immediate_flip_not_required) &&
+                                                                               (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] == dml_use_mall_pstate_change_full_frame || mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe));
+       }
+
+       mode_lib->ms.support.InvalidCombinationOfMALLUseForPStateAndStaticScreen = false;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               mode_lib->ms.support.InvalidCombinationOfMALLUseForPStateAndStaticScreen = mode_lib->ms.support.InvalidCombinationOfMALLUseForPStateAndStaticScreen ||
+                                                                                                                               ((mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen[k] == dml_use_mall_static_screen_enable || mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen[k] == dml_use_mall_static_screen_optimize) && (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe)) ||
+                                                                                                                               ((mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen[k] == dml_use_mall_static_screen_disable || mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen[k] == dml_use_mall_static_screen_optimize) && (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] == dml_use_mall_pstate_change_full_frame));
+       }
+
+       s->FullFrameMALLPStateMethod = false;
+       s->SubViewportMALLPStateMethod = false;
+       s->PhantomPipeMALLPStateMethod = false;
+       s->SubViewportMALLRefreshGreaterThan120Hz = false;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] == dml_use_mall_pstate_change_full_frame)
+                       s->FullFrameMALLPStateMethod = true;
+               if (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] == dml_use_mall_pstate_change_sub_viewport) {
+                       s->SubViewportMALLPStateMethod = true;
+                       if (mode_lib->ms.cache_display_cfg.timing.RefreshRate[k] > 120)
+                               s->SubViewportMALLRefreshGreaterThan120Hz = true;
+               }
+               if (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe)
+                       s->PhantomPipeMALLPStateMethod = true;
+       }
+       mode_lib->ms.support.InvalidCombinationOfMALLUseForPState = (s->SubViewportMALLPStateMethod != s->PhantomPipeMALLPStateMethod)
+       || (s->SubViewportMALLPStateMethod && s->FullFrameMALLPStateMethod) || s->SubViewportMALLRefreshGreaterThan120Hz;
+
+    if (mode_lib->ms.policy.UseMinimumRequiredDCFCLK == true) {
+               UseMinimumDCFCLK_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+               UseMinimumDCFCLK_params->DRRDisplay = mode_lib->ms.cache_display_cfg.timing.DRRDisplay;
+               UseMinimumDCFCLK_params->SynchronizeDRRDisplaysForUCLKPStateChangeFinal = mode_lib->ms.policy.SynchronizeDRRDisplaysForUCLKPStateChangeFinal;
+               UseMinimumDCFCLK_params->MaxInterDCNTileRepeaters = mode_lib->ms.ip.max_inter_dcn_tile_repeaters;
+               UseMinimumDCFCLK_params->MaxPrefetchMode = dml_prefetch_support_stutter;
+               UseMinimumDCFCLK_params->DRAMClockChangeLatencyFinal = mode_lib->ms.state.dram_clock_change_latency_us;
+               UseMinimumDCFCLK_params->FCLKChangeLatency = mode_lib->ms.state.fclk_change_latency_us;
+               UseMinimumDCFCLK_params->SREnterPlusExitTime = mode_lib->ms.state.sr_enter_plus_exit_time_us;
+               UseMinimumDCFCLK_params->ReturnBusWidth = mode_lib->ms.soc.return_bus_width_bytes;
+               UseMinimumDCFCLK_params->RoundTripPingLatencyCycles = mode_lib->ms.soc.round_trip_ping_latency_dcfclk_cycles;
+               UseMinimumDCFCLK_params->ReorderingBytes = s->ReorderingBytes;
+               UseMinimumDCFCLK_params->PixelChunkSizeInKByte = mode_lib->ms.ip.pixel_chunk_size_kbytes;
+               UseMinimumDCFCLK_params->MetaChunkSize = mode_lib->ms.ip.meta_chunk_size_kbytes;
+               UseMinimumDCFCLK_params->GPUVMEnable = mode_lib->ms.cache_display_cfg.plane.GPUVMEnable;
+               UseMinimumDCFCLK_params->GPUVMMaxPageTableLevels = mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels;
+               UseMinimumDCFCLK_params->HostVMEnable = mode_lib->ms.cache_display_cfg.plane.HostVMEnable;
+               UseMinimumDCFCLK_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+               UseMinimumDCFCLK_params->HostVMMinPageSize = mode_lib->ms.soc.hostvm_min_page_size_kbytes;
+               UseMinimumDCFCLK_params->HostVMMaxNonCachedPageTableLevels = mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels;
+               UseMinimumDCFCLK_params->DynamicMetadataVMEnabled = mode_lib->ms.ip.dynamic_metadata_vm_enabled;
+               UseMinimumDCFCLK_params->ImmediateFlipRequirement = s->ImmediateFlipRequiredFinal;
+               UseMinimumDCFCLK_params->ProgressiveToInterlaceUnitInOPP = mode_lib->ms.ip.ptoi_supported;
+               UseMinimumDCFCLK_params->MaxAveragePercentOfIdealSDPPortBWDisplayCanUseInNormalSystemOperation = mode_lib->ms.soc.max_avg_sdp_bw_use_normal_percent;
+               UseMinimumDCFCLK_params->PercentOfIdealSDPPortBWReceivedAfterUrgLatency = mode_lib->ms.soc.pct_ideal_sdp_bw_after_urgent;
+               UseMinimumDCFCLK_params->VTotal = mode_lib->ms.cache_display_cfg.timing.VTotal;
+               UseMinimumDCFCLK_params->VActive = mode_lib->ms.cache_display_cfg.timing.VActive;
+               UseMinimumDCFCLK_params->DynamicMetadataTransmittedBytes = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataTransmittedBytes;
+               UseMinimumDCFCLK_params->DynamicMetadataLinesBeforeActiveRequired = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataLinesBeforeActiveRequired;
+               UseMinimumDCFCLK_params->Interlace = mode_lib->ms.cache_display_cfg.timing.Interlace;
+               UseMinimumDCFCLK_params->RequiredDPPCLKPerSurface = mode_lib->ms.RequiredDPPCLKPerSurface;
+               UseMinimumDCFCLK_params->RequiredDISPCLK = mode_lib->ms.RequiredDISPCLK;
+               UseMinimumDCFCLK_params->UrgLatency = mode_lib->ms.UrgLatency;
+               UseMinimumDCFCLK_params->NoOfDPP = mode_lib->ms.NoOfDPP;
+               UseMinimumDCFCLK_params->ProjectedDCFCLKDeepSleep = mode_lib->ms.ProjectedDCFCLKDeepSleep;
+               UseMinimumDCFCLK_params->MaximumVStartup = s->MaximumVStartup;
+               UseMinimumDCFCLK_params->TotalNumberOfActiveDPP = mode_lib->ms.TotalNumberOfActiveDPP;
+               UseMinimumDCFCLK_params->TotalNumberOfDCCActiveDPP = mode_lib->ms.TotalNumberOfDCCActiveDPP;
+               UseMinimumDCFCLK_params->dpte_group_bytes = mode_lib->ms.dpte_group_bytes;
+               UseMinimumDCFCLK_params->PrefetchLinesY = mode_lib->ms.PrefetchLinesY;
+               UseMinimumDCFCLK_params->PrefetchLinesC = mode_lib->ms.PrefetchLinesC;
+               UseMinimumDCFCLK_params->swath_width_luma_ub_all_states = mode_lib->ms.swath_width_luma_ub_all_states;
+               UseMinimumDCFCLK_params->swath_width_chroma_ub_all_states = mode_lib->ms.swath_width_chroma_ub_all_states;
+               UseMinimumDCFCLK_params->BytePerPixelY = mode_lib->ms.BytePerPixelY;
+               UseMinimumDCFCLK_params->BytePerPixelC = mode_lib->ms.BytePerPixelC;
+               UseMinimumDCFCLK_params->HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal;
+               UseMinimumDCFCLK_params->PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock;
+               UseMinimumDCFCLK_params->PDEAndMetaPTEBytesPerFrame = mode_lib->ms.PDEAndMetaPTEBytesPerFrame;
+               UseMinimumDCFCLK_params->DPTEBytesPerRow = mode_lib->ms.DPTEBytesPerRow;
+               UseMinimumDCFCLK_params->MetaRowBytes = mode_lib->ms.MetaRowBytes;
+               UseMinimumDCFCLK_params->DynamicMetadataEnable = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataEnable;
+               UseMinimumDCFCLK_params->ReadBandwidthLuma = mode_lib->ms.ReadBandwidthLuma;
+               UseMinimumDCFCLK_params->ReadBandwidthChroma = mode_lib->ms.ReadBandwidthChroma;
+               UseMinimumDCFCLK_params->DCFCLKPerState = mode_lib->ms.state.dcfclk_mhz;
+               UseMinimumDCFCLK_params->DCFCLKState = mode_lib->ms.DCFCLKState;
+
+               UseMinimumDCFCLK(&mode_lib->scratch,
+               UseMinimumDCFCLK_params);
+
+        } // UseMinimumRequiredDCFCLK == true
+
+       for (j = 0; j < 2; ++j) {
+               mode_lib->ms.ReturnBWPerState[j] = dml_get_return_bw_mbps(&mode_lib->ms.soc, mode_lib->ms.state.use_ideal_dram_bw_strobe,
+                                                                                                                               mode_lib->ms.cache_display_cfg.plane.HostVMEnable, mode_lib->ms.DCFCLKState[j], mode_lib->ms.state.fabricclk_mhz,
+                                                                                                                               mode_lib->ms.state.dram_speed_mts);
+               mode_lib->ms.ReturnDRAMBWPerState[j] = dml_get_return_dram_bw_mbps(&mode_lib->ms.soc, mode_lib->ms.state.use_ideal_dram_bw_strobe,
+                                                                                                                               mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+                                                                                                                               mode_lib->ms.state.dram_speed_mts);
+       }
+
+       //Re-ordering Buffer Support Check
+       for (j = 0; j < 2; ++j) {
+               if ((mode_lib->ms.ip.rob_buffer_size_kbytes - mode_lib->ms.ip.pixel_chunk_size_kbytes) * 1024 / mode_lib->ms.ReturnBWPerState[j] >
+                       (mode_lib->ms.soc.round_trip_ping_latency_dcfclk_cycles + 32) / mode_lib->ms.DCFCLKState[j] + s->ReorderingBytes / mode_lib->ms.ReturnBWPerState[j]) {
+                       mode_lib->ms.support.ROBSupport[j] = true;
+               } else {
+                       mode_lib->ms.support.ROBSupport[j] = false;
+               }
+               dml_print("DML::%s: DEBUG ROBSupport[%u] = %u (%u)\n",  __func__, j, mode_lib->ms.support.ROBSupport[j], __LINE__);
+       }
+
+       //Vertical Active BW support check
+       s->MaxTotalVActiveRDBandwidth = 0;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               s->MaxTotalVActiveRDBandwidth = s->MaxTotalVActiveRDBandwidth + mode_lib->ms.ReadBandwidthLuma[k] + mode_lib->ms.ReadBandwidthChroma[k];
+       }
+
+       for (j = 0; j < 2; ++j) {
+               mode_lib->ms.support.MaxTotalVerticalActiveAvailableBandwidth[j] = dml_min3(mode_lib->ms.soc.return_bus_width_bytes * mode_lib->ms.DCFCLKState[j] * mode_lib->ms.soc.max_avg_sdp_bw_use_normal_percent / 100.0,
+                                                                                                                                       mode_lib->ms.state.fabricclk_mhz * mode_lib->ms.soc.fabric_datapath_to_dcn_data_return_bytes * mode_lib->ms.soc.max_avg_fabric_bw_use_normal_percent / 100.0,
+                                                                                                                                       mode_lib->ms.state.dram_speed_mts * mode_lib->ms.soc.num_chans * mode_lib->ms.soc.dram_channel_width_bytes *
+                                                                                                                                       ((mode_lib->ms.state.use_ideal_dram_bw_strobe && !mode_lib->ms.cache_display_cfg.plane.HostVMEnable) ?
+                                                                                                                                       mode_lib->ms.soc.max_avg_dram_bw_use_normal_strobe_percent : mode_lib->ms.soc.max_avg_dram_bw_use_normal_percent) / 100.0);
+
+               if (s->MaxTotalVActiveRDBandwidth <= mode_lib->ms.support.MaxTotalVerticalActiveAvailableBandwidth[j]) {
+                       mode_lib->ms.support.TotalVerticalActiveBandwidthSupport[j] = true;
+               } else {
+                       mode_lib->ms.support.TotalVerticalActiveBandwidthSupport[j] = false;
+               }
+       }
+
+       /* Prefetch Check */
+
+       for (j = 0; j < 2; ++j) {
+               mode_lib->ms.TimeCalc = 24 / mode_lib->ms.ProjectedDCFCLKDeepSleep[j];
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       mode_lib->ms.NoOfDPPThisState[k] = mode_lib->ms.NoOfDPP[j][k];
+                       mode_lib->ms.swath_width_luma_ub_this_state[k] = mode_lib->ms.swath_width_luma_ub_all_states[j][k];
+                       mode_lib->ms.swath_width_chroma_ub_this_state[k] = mode_lib->ms.swath_width_chroma_ub_all_states[j][k];
+                       mode_lib->ms.SwathWidthYThisState[k] = mode_lib->ms.SwathWidthYAllStates[j][k];
+                       mode_lib->ms.SwathWidthCThisState[k] = mode_lib->ms.SwathWidthCAllStates[j][k];
+                       mode_lib->ms.SwathHeightYThisState[k] = mode_lib->ms.SwathHeightYAllStates[j][k];
+                       mode_lib->ms.SwathHeightCThisState[k] = mode_lib->ms.SwathHeightCAllStates[j][k];
+                       mode_lib->ms.UnboundedRequestEnabledThisState = mode_lib->ms.UnboundedRequestEnabledAllStates[j];
+                       mode_lib->ms.CompressedBufferSizeInkByteThisState = mode_lib->ms.CompressedBufferSizeInkByteAllStates[j];
+                       mode_lib->ms.DETBufferSizeInKByteThisState[k] = mode_lib->ms.DETBufferSizeInKByteAllStates[j][k];
+                       mode_lib->ms.DETBufferSizeYThisState[k] = mode_lib->ms.DETBufferSizeYAllStates[j][k];
+                       mode_lib->ms.DETBufferSizeCThisState[k] = mode_lib->ms.DETBufferSizeCAllStates[j][k];
+               }
+
+               mode_lib->ms.support.VActiveBandwithSupport[j] = CalculateVActiveBandwithSupport(
+                       mode_lib->ms.num_active_planes,
+                       mode_lib->ms.ReturnBWPerState[j],
+                       mode_lib->ms.NotUrgentLatencyHiding,
+                       mode_lib->ms.ReadBandwidthLuma,
+                       mode_lib->ms.ReadBandwidthChroma,
+                       mode_lib->ms.cursor_bw,
+                       mode_lib->ms.meta_row_bandwidth_this_state,
+                       mode_lib->ms.dpte_row_bandwidth_this_state,
+                       mode_lib->ms.NoOfDPPThisState,
+                       mode_lib->ms.UrgentBurstFactorLuma,
+                       mode_lib->ms.UrgentBurstFactorChroma,
+                       mode_lib->ms.UrgentBurstFactorCursor);
+
+               s->VMDataOnlyReturnBWPerState = dml_get_return_bw_mbps_vm_only(
+                                                                                                                                       &mode_lib->ms.soc,
+                                                                                                                                       mode_lib->ms.state.use_ideal_dram_bw_strobe,
+                                                                                                                                       mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+                                                                                                                                       mode_lib->ms.DCFCLKState[j],
+                                                                                                                                       mode_lib->ms.state.fabricclk_mhz,
+                                                                                                                                       mode_lib->ms.state.dram_speed_mts);
+
+               s->HostVMInefficiencyFactor = 1;
+               if (mode_lib->ms.cache_display_cfg.plane.GPUVMEnable && mode_lib->ms.cache_display_cfg.plane.HostVMEnable)
+                       s->HostVMInefficiencyFactor = mode_lib->ms.ReturnBWPerState[j] / s->VMDataOnlyReturnBWPerState;
+
+               mode_lib->ms.ExtraLatency = CalculateExtraLatency(
+                               mode_lib->ms.soc.round_trip_ping_latency_dcfclk_cycles,
+                               s->ReorderingBytes,
+                               mode_lib->ms.DCFCLKState[j],
+                               mode_lib->ms.TotalNumberOfActiveDPP[j],
+                               mode_lib->ms.ip.pixel_chunk_size_kbytes,
+                               mode_lib->ms.TotalNumberOfDCCActiveDPP[j],
+                               mode_lib->ms.ip.meta_chunk_size_kbytes,
+                               mode_lib->ms.ReturnBWPerState[j],
+                               mode_lib->ms.cache_display_cfg.plane.GPUVMEnable,
+                               mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+                               mode_lib->ms.num_active_planes,
+                               mode_lib->ms.NoOfDPPThisState,
+                               mode_lib->ms.dpte_group_bytes,
+                               s->HostVMInefficiencyFactor,
+                               mode_lib->ms.soc.hostvm_min_page_size_kbytes,
+                               mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels);
+
+               s->NextMaxVStartup = s->MaxVStartupAllPlanes[j];
+               s->MaxVStartup = 0;
+               s->AllPrefetchModeTested = true;
+               for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                       CalculatePrefetchMode(mode_lib->ms.policy.AllowForPStateChangeOrStutterInVBlank[k], &s->MinPrefetchMode[k], &s->MaxPrefetchMode[k]);
+                       s->NextPrefetchMode[k] = s->MinPrefetchMode[k];
+               }
+
+               do {
+                       s->MaxVStartup = s->NextMaxVStartup;
+                       s->AllPrefetchModeTested = true;
+
+                       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                               mode_lib->ms.PrefetchMode[k] = s->NextPrefetchMode[k];
+                               mode_lib->ms.TWait = CalculateTWait(
+                                                               mode_lib->ms.PrefetchMode[k],
+                                                               mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k],
+                                                               mode_lib->ms.policy.SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+                                                               mode_lib->ms.cache_display_cfg.timing.DRRDisplay[k],
+                                                               mode_lib->ms.state.dram_clock_change_latency_us,
+                                                               mode_lib->ms.state.fclk_change_latency_us,
+                                                               mode_lib->ms.UrgLatency,
+                                                               mode_lib->ms.state.sr_enter_plus_exit_time_us);
+
+                               myPipe = &s->myPipe;
+                               myPipe->Dppclk = mode_lib->ms.RequiredDPPCLKPerSurface[j][k];
+                               myPipe->Dispclk = mode_lib->ms.RequiredDISPCLK[j];
+                               myPipe->PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock[k];
+                               myPipe->DCFClkDeepSleep = mode_lib->ms.ProjectedDCFCLKDeepSleep[j];
+                               myPipe->DPPPerSurface = mode_lib->ms.NoOfDPP[j][k];
+                               myPipe->ScalerEnabled = mode_lib->ms.cache_display_cfg.plane.ScalerEnabled[k];
+                               myPipe->SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan[k];
+                               myPipe->BlockWidth256BytesY = mode_lib->ms.Read256BlockWidthY[k];
+                               myPipe->BlockHeight256BytesY = mode_lib->ms.Read256BlockHeightY[k];
+                               myPipe->BlockWidth256BytesC = mode_lib->ms.Read256BlockWidthC[k];
+                               myPipe->BlockHeight256BytesC = mode_lib->ms.Read256BlockHeightC[k];
+                               myPipe->InterlaceEnable = mode_lib->ms.cache_display_cfg.timing.Interlace[k];
+                               myPipe->NumberOfCursors = mode_lib->ms.cache_display_cfg.plane.NumberOfCursors[k];
+                               myPipe->VBlank = mode_lib->ms.cache_display_cfg.timing.VTotal[k] - mode_lib->ms.cache_display_cfg.timing.VActive[k];
+                               myPipe->HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal[k];
+                               myPipe->HActive = mode_lib->ms.cache_display_cfg.timing.HActive[k];
+                               myPipe->DCCEnable = mode_lib->ms.cache_display_cfg.surface.DCCEnable[k];
+                               myPipe->ODMMode = mode_lib->ms.ODMModePerState[k];
+                               myPipe->SourcePixelFormat = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k];
+                               myPipe->BytePerPixelY = mode_lib->ms.BytePerPixelY[k];
+                               myPipe->BytePerPixelC = mode_lib->ms.BytePerPixelC[k];
+                               myPipe->ProgressiveToInterlaceUnitInOPP = mode_lib->ms.ip.ptoi_supported;
+
+#ifdef __DML_VBA_DEBUG__
+                               dml_print("DML::%s: Calling CalculatePrefetchSchedule for j=%u, k=%u\n", __func__, j, k);
+                               dml_print("DML::%s: MaximumVStartup = %u\n", __func__, s->MaximumVStartup[j][k]);
+                               dml_print("DML::%s: MaxVStartup = %u\n", __func__, s->MaxVStartup);
+                               dml_print("DML::%s: NextPrefetchMode = %u\n", __func__, s->NextPrefetchMode[k]);
+                               dml_print("DML::%s: AllowForPStateChangeOrStutterInVBlank = %u\n", __func__, mode_lib->ms.policy.AllowForPStateChangeOrStutterInVBlank[k]);
+                               dml_print("DML::%s: PrefetchMode = %u\n", __func__, mode_lib->ms.PrefetchMode[k]);
+#endif
+
+                               CalculatePrefetchSchedule_params->EnhancedPrefetchScheduleAccelerationFinal = mode_lib->ms.policy.EnhancedPrefetchScheduleAccelerationFinal;
+                               CalculatePrefetchSchedule_params->HostVMInefficiencyFactor = s->HostVMInefficiencyFactor;
+                               CalculatePrefetchSchedule_params->myPipe = myPipe;
+                               CalculatePrefetchSchedule_params->DSCDelay = mode_lib->ms.DSCDelayPerState[k];
+                               CalculatePrefetchSchedule_params->DPPCLKDelaySubtotalPlusCNVCFormater = mode_lib->ms.ip.dppclk_delay_subtotal + mode_lib->ms.ip.dppclk_delay_cnvc_formatter;
+                               CalculatePrefetchSchedule_params->DPPCLKDelaySCL = mode_lib->ms.ip.dppclk_delay_scl;
+                               CalculatePrefetchSchedule_params->DPPCLKDelaySCLLBOnly = mode_lib->ms.ip.dppclk_delay_scl_lb_only;
+                               CalculatePrefetchSchedule_params->DPPCLKDelayCNVCCursor = mode_lib->ms.ip.dppclk_delay_cnvc_cursor;
+                               CalculatePrefetchSchedule_params->DISPCLKDelaySubtotal = mode_lib->ms.ip.dispclk_delay_subtotal;
+                               CalculatePrefetchSchedule_params->DPP_RECOUT_WIDTH = (dml_uint_t)(mode_lib->ms.SwathWidthYThisState[k] / mode_lib->ms.cache_display_cfg.plane.HRatio[k]);
+                               CalculatePrefetchSchedule_params->OutputFormat = mode_lib->ms.cache_display_cfg.output.OutputFormat[k];
+                               CalculatePrefetchSchedule_params->MaxInterDCNTileRepeaters = mode_lib->ms.ip.max_inter_dcn_tile_repeaters;
+                               CalculatePrefetchSchedule_params->VStartup = (dml_uint_t)(dml_min(s->MaxVStartup, s->MaximumVStartup[j][k]));
+                               CalculatePrefetchSchedule_params->MaxVStartup = s->MaximumVStartup[j][k];
+                               CalculatePrefetchSchedule_params->GPUVMPageTableLevels = mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels;
+                               CalculatePrefetchSchedule_params->GPUVMEnable = mode_lib->ms.cache_display_cfg.plane.GPUVMEnable;
+                               CalculatePrefetchSchedule_params->HostVMEnable = mode_lib->ms.cache_display_cfg.plane.HostVMEnable;
+                               CalculatePrefetchSchedule_params->HostVMMaxNonCachedPageTableLevels = mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels;
+                               CalculatePrefetchSchedule_params->HostVMMinPageSize = mode_lib->ms.soc.hostvm_min_page_size_kbytes;
+                               CalculatePrefetchSchedule_params->DynamicMetadataEnable = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataEnable[k];
+                               CalculatePrefetchSchedule_params->DynamicMetadataVMEnabled = mode_lib->ms.ip.dynamic_metadata_vm_enabled;
+                               CalculatePrefetchSchedule_params->DynamicMetadataLinesBeforeActiveRequired = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataLinesBeforeActiveRequired[k];
+                               CalculatePrefetchSchedule_params->DynamicMetadataTransmittedBytes = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataTransmittedBytes[k];
+                               CalculatePrefetchSchedule_params->UrgentLatency = mode_lib->ms.UrgLatency;
+                               CalculatePrefetchSchedule_params->UrgentExtraLatency = mode_lib->ms.ExtraLatency;
+                               CalculatePrefetchSchedule_params->TCalc = mode_lib->ms.TimeCalc;
+                               CalculatePrefetchSchedule_params->PDEAndMetaPTEBytesFrame = mode_lib->ms.PDEAndMetaPTEBytesPerFrame[j][k];
+                               CalculatePrefetchSchedule_params->MetaRowByte = mode_lib->ms.MetaRowBytes[j][k];
+                               CalculatePrefetchSchedule_params->PixelPTEBytesPerRow = mode_lib->ms.DPTEBytesPerRow[j][k];
+                               CalculatePrefetchSchedule_params->PrefetchSourceLinesY = mode_lib->ms.PrefetchLinesY[j][k];
+                               CalculatePrefetchSchedule_params->VInitPreFillY = mode_lib->ms.PrefillY[k];
+                               CalculatePrefetchSchedule_params->MaxNumSwathY = mode_lib->ms.MaxNumSwY[k];
+                               CalculatePrefetchSchedule_params->PrefetchSourceLinesC = mode_lib->ms.PrefetchLinesC[j][k];
+                               CalculatePrefetchSchedule_params->VInitPreFillC = mode_lib->ms.PrefillC[k];
+                               CalculatePrefetchSchedule_params->MaxNumSwathC = mode_lib->ms.MaxNumSwC[k];
+                               CalculatePrefetchSchedule_params->swath_width_luma_ub = mode_lib->ms.swath_width_luma_ub_this_state[k];
+                               CalculatePrefetchSchedule_params->swath_width_chroma_ub = mode_lib->ms.swath_width_chroma_ub_this_state[k];
+                               CalculatePrefetchSchedule_params->SwathHeightY = mode_lib->ms.SwathHeightYThisState[k];
+                               CalculatePrefetchSchedule_params->SwathHeightC = mode_lib->ms.SwathHeightCThisState[k];
+                               CalculatePrefetchSchedule_params->TWait = mode_lib->ms.TWait;
+                               CalculatePrefetchSchedule_params->DSTXAfterScaler = &s->DSTXAfterScaler[k];
+                               CalculatePrefetchSchedule_params->DSTYAfterScaler = &s->DSTYAfterScaler[k];
+                               CalculatePrefetchSchedule_params->DestinationLinesForPrefetch = &mode_lib->ms.LineTimesForPrefetch[k];
+                               CalculatePrefetchSchedule_params->DestinationLinesToRequestVMInVBlank = &mode_lib->ms.LinesForMetaPTE[k];
+                               CalculatePrefetchSchedule_params->DestinationLinesToRequestRowInVBlank = &mode_lib->ms.LinesForMetaAndDPTERow[k];
+                               CalculatePrefetchSchedule_params->VRatioPrefetchY = &mode_lib->ms.VRatioPreY[j][k];
+                               CalculatePrefetchSchedule_params->VRatioPrefetchC = &mode_lib->ms.VRatioPreC[j][k];
+                               CalculatePrefetchSchedule_params->RequiredPrefetchPixDataBWLuma = &mode_lib->ms.RequiredPrefetchPixelDataBWLuma[k];
+                               CalculatePrefetchSchedule_params->RequiredPrefetchPixDataBWChroma = &mode_lib->ms.RequiredPrefetchPixelDataBWChroma[k];
+                               CalculatePrefetchSchedule_params->NotEnoughTimeForDynamicMetadata = &mode_lib->ms.support.NoTimeForDynamicMetadata[j][k];
+                               CalculatePrefetchSchedule_params->Tno_bw = &mode_lib->ms.Tno_bw[k];
+                               CalculatePrefetchSchedule_params->prefetch_vmrow_bw = &mode_lib->ms.prefetch_vmrow_bw[k];
+                               CalculatePrefetchSchedule_params->Tdmdl_vm = &s->dummy_single[0];
+                               CalculatePrefetchSchedule_params->Tdmdl = &s->dummy_single[1];
+                               CalculatePrefetchSchedule_params->TSetup = &s->dummy_single[2];
+                               CalculatePrefetchSchedule_params->VUpdateOffsetPix = &s->dummy_integer[0];
+                               CalculatePrefetchSchedule_params->VUpdateWidthPix = &s->dummy_integer[1];
+                               CalculatePrefetchSchedule_params->VReadyOffsetPix = &s->dummy_integer[2];
+
+                               mode_lib->ms.support.NoTimeForPrefetch[j][k] =
+                                                               CalculatePrefetchSchedule(&mode_lib->scratch,
+                                                               CalculatePrefetchSchedule_params);
+                       }
+
+                       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                                       CalculateUrgentBurstFactor(
+                                                       mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k],
+                                                       mode_lib->ms.swath_width_luma_ub_this_state[k],
+                                                       mode_lib->ms.swath_width_chroma_ub_this_state[k],
+                                                       mode_lib->ms.SwathHeightYThisState[k],
+                                                       mode_lib->ms.SwathHeightCThisState[k],
+                                                       mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+                                                       mode_lib->ms.UrgLatency,
+                                                       mode_lib->ms.ip.cursor_buffer_size,
+                                                       mode_lib->ms.cache_display_cfg.plane.CursorWidth[k],
+                                                       mode_lib->ms.cache_display_cfg.plane.CursorBPP[k],
+                                                       mode_lib->ms.VRatioPreY[j][k],
+                                                       mode_lib->ms.VRatioPreC[j][k],
+                                                       mode_lib->ms.BytePerPixelInDETY[k],
+                                                       mode_lib->ms.BytePerPixelInDETC[k],
+                                                       mode_lib->ms.DETBufferSizeYThisState[k],
+                                                       mode_lib->ms.DETBufferSizeCThisState[k],
+                                                       /* Output */
+                                                       &mode_lib->ms.UrgentBurstFactorCursorPre[k],
+                                                       &mode_lib->ms.UrgentBurstFactorLumaPre[k],
+                                                       &mode_lib->ms.UrgentBurstFactorChroma[k],
+                                                       &mode_lib->ms.NotUrgentLatencyHidingPre[k]);
+
+                                       mode_lib->ms.cursor_bw_pre[k] = mode_lib->ms.cache_display_cfg.plane.NumberOfCursors[k] * mode_lib->ms.cache_display_cfg.plane.CursorWidth[k] *
+                                                                                                       mode_lib->ms.cache_display_cfg.plane.CursorBPP[k] / 8.0 / (mode_lib->ms.cache_display_cfg.timing.HTotal[k] /
+                                                                                                       mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * mode_lib->ms.VRatioPreY[j][k];
+                       }
+
+                       {
+                       CalculatePrefetchBandwithSupport(
+                               mode_lib->ms.num_active_planes,
+                               mode_lib->ms.ReturnBWPerState[j],
+                               mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange,
+                               mode_lib->ms.NotUrgentLatencyHidingPre,
+                               mode_lib->ms.ReadBandwidthLuma,
+                               mode_lib->ms.ReadBandwidthChroma,
+                               mode_lib->ms.RequiredPrefetchPixelDataBWLuma,
+                               mode_lib->ms.RequiredPrefetchPixelDataBWChroma,
+                               mode_lib->ms.cursor_bw,
+                               mode_lib->ms.meta_row_bandwidth_this_state,
+                               mode_lib->ms.dpte_row_bandwidth_this_state,
+                               mode_lib->ms.cursor_bw_pre,
+                               mode_lib->ms.prefetch_vmrow_bw,
+                               mode_lib->ms.NoOfDPPThisState,
+                               mode_lib->ms.UrgentBurstFactorLuma,
+                               mode_lib->ms.UrgentBurstFactorChroma,
+                               mode_lib->ms.UrgentBurstFactorCursor,
+                               mode_lib->ms.UrgentBurstFactorLumaPre,
+                               mode_lib->ms.UrgentBurstFactorChromaPre,
+                               mode_lib->ms.UrgentBurstFactorCursorPre,
+
+                               /* output */
+                               &s->dummy_single[0], // dml_float_t *PrefetchBandwidth
+                               &s->dummy_single[1], // dml_float_t *PrefetchBandwidthNotIncludingMALLPrefetch
+                               &mode_lib->mp.FractionOfUrgentBandwidth, // dml_float_t *FractionOfUrgentBandwidth
+                               &mode_lib->ms.support.PrefetchSupported[j]);
+                       }
+
+                       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                               if (mode_lib->ms.LineTimesForPrefetch[k] < 2.0
+                                       || mode_lib->ms.LinesForMetaPTE[k] >= 32.0
+                                       || mode_lib->ms.LinesForMetaAndDPTERow[k] >= 16.0
+                                       || mode_lib->ms.support.NoTimeForPrefetch[j][k] == true) {
+                                               mode_lib->ms.support.PrefetchSupported[j] = false;
+                               }
+                       }
+
+                       mode_lib->ms.support.DynamicMetadataSupported[j] = true;
+                       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                               if (mode_lib->ms.support.NoTimeForDynamicMetadata[j][k] == true) {
+                                       mode_lib->ms.support.DynamicMetadataSupported[j] = false;
+                               }
+                       }
+
+                       mode_lib->ms.support.VRatioInPrefetchSupported[j] = true;
+                       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                               if (mode_lib->ms.support.NoTimeForPrefetch[j][k] == true ||
+                                       mode_lib->ms.VRatioPreY[j][k] > __DML_MAX_VRATIO_PRE_ENHANCE_PREFETCH_ACC__ ||
+                                       mode_lib->ms.VRatioPreC[j][k] > __DML_MAX_VRATIO_PRE_ENHANCE_PREFETCH_ACC__ ||
+                                       ((s->MaxVStartup < s->MaximumVStartup[j][k] || mode_lib->ms.policy.EnhancedPrefetchScheduleAccelerationFinal == 0) &&
+                                               (mode_lib->ms.VRatioPreY[j][k] > __DML_MAX_VRATIO_PRE__ || mode_lib->ms.VRatioPreC[j][k] > __DML_MAX_VRATIO_PRE__))) {
+                                                       mode_lib->ms.support.VRatioInPrefetchSupported[j] = false;
+                               }
+                       }
+
+                       s->AnyLinesForVMOrRowTooLarge = false;
+                       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                               if (mode_lib->ms.LinesForMetaAndDPTERow[k] >= 16 || mode_lib->ms.LinesForMetaPTE[k] >= 32) {
+                                       s->AnyLinesForVMOrRowTooLarge = true;
+                               }
+                       }
+
+                       if (mode_lib->ms.support.PrefetchSupported[j] == true && mode_lib->ms.support.VRatioInPrefetchSupported[j] == true) {
+                               mode_lib->ms.BandwidthAvailableForImmediateFlip = CalculateBandwidthAvailableForImmediateFlip(
+                                               mode_lib->ms.num_active_planes,
+                                               mode_lib->ms.ReturnBWPerState[j],
+                                               mode_lib->ms.ReadBandwidthLuma,
+                                               mode_lib->ms.ReadBandwidthChroma,
+                                               mode_lib->ms.RequiredPrefetchPixelDataBWLuma,
+                                               mode_lib->ms.RequiredPrefetchPixelDataBWChroma,
+                                               mode_lib->ms.cursor_bw,
+                                               mode_lib->ms.cursor_bw_pre,
+                                               mode_lib->ms.NoOfDPPThisState,
+                                               mode_lib->ms.UrgentBurstFactorLuma,
+                                               mode_lib->ms.UrgentBurstFactorChroma,
+                                               mode_lib->ms.UrgentBurstFactorCursor,
+                                               mode_lib->ms.UrgentBurstFactorLumaPre,
+                                               mode_lib->ms.UrgentBurstFactorChromaPre,
+                                               mode_lib->ms.UrgentBurstFactorCursorPre);
+
+                               mode_lib->ms.TotImmediateFlipBytes = 0;
+                               for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                                       if (!(mode_lib->ms.policy.ImmediateFlipRequirement[k] == dml_immediate_flip_not_required)) {
+                                               mode_lib->ms.TotImmediateFlipBytes = mode_lib->ms.TotImmediateFlipBytes + mode_lib->ms.NoOfDPP[j][k] * mode_lib->ms.PDEAndMetaPTEBytesPerFrame[j][k] + mode_lib->ms.MetaRowBytes[j][k];
+                                               if (mode_lib->ms.use_one_row_for_frame_flip[j][k]) {
+                                                       mode_lib->ms.TotImmediateFlipBytes = mode_lib->ms.TotImmediateFlipBytes + mode_lib->ms.NoOfDPP[j][k] * (2 * mode_lib->ms.DPTEBytesPerRow[j][k]);
+                                               } else {
+                                                       mode_lib->ms.TotImmediateFlipBytes = mode_lib->ms.TotImmediateFlipBytes + mode_lib->ms.NoOfDPP[j][k] * mode_lib->ms.DPTEBytesPerRow[j][k];
+                                               }
+                                       }
+                               }
+
+                               for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                                       CalculateFlipSchedule(
+                                               s->HostVMInefficiencyFactor,
+                                               mode_lib->ms.ExtraLatency,
+                                               mode_lib->ms.UrgLatency,
+                                               mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels,
+                                               mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+                                               mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels,
+                                               mode_lib->ms.cache_display_cfg.plane.GPUVMEnable,
+                                               mode_lib->ms.soc.hostvm_min_page_size_kbytes,
+                                               mode_lib->ms.PDEAndMetaPTEBytesPerFrame[j][k],
+                                               mode_lib->ms.MetaRowBytes[j][k],
+                                               mode_lib->ms.DPTEBytesPerRow[j][k],
+                                               mode_lib->ms.BandwidthAvailableForImmediateFlip,
+                                               mode_lib->ms.TotImmediateFlipBytes,
+                                               mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k],
+                                               (mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]),
+                                               mode_lib->ms.cache_display_cfg.plane.VRatio[k],
+                                               mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k],
+                                               mode_lib->ms.Tno_bw[k],
+                                               mode_lib->ms.cache_display_cfg.surface.DCCEnable[k],
+                                               mode_lib->ms.dpte_row_height[k],
+                                               mode_lib->ms.meta_row_height[k],
+                                               mode_lib->ms.dpte_row_height_chroma[k],
+                                               mode_lib->ms.meta_row_height_chroma[k],
+                                               mode_lib->ms.use_one_row_for_frame_flip[j][k], // 24
+
+                                               /* Output */
+                                               &mode_lib->ms.DestinationLinesToRequestVMInImmediateFlip[k],
+                                               &mode_lib->ms.DestinationLinesToRequestRowInImmediateFlip[k],
+                                               &mode_lib->ms.final_flip_bw[k],
+                                               &mode_lib->ms.ImmediateFlipSupportedForPipe[k]);
+                               }
+
+                               {
+                               CalculateImmediateFlipBandwithSupport(mode_lib->ms.num_active_planes,
+                                                                                                       mode_lib->ms.ReturnBWPerState[j],
+                                                                                                       mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange,
+                                                                                                       mode_lib->ms.policy.ImmediateFlipRequirement,
+                                                                                                       mode_lib->ms.final_flip_bw,
+                                                                                                       mode_lib->ms.ReadBandwidthLuma,
+                                                                                                       mode_lib->ms.ReadBandwidthChroma,
+                                                                                                       mode_lib->ms.RequiredPrefetchPixelDataBWLuma,
+                                                                                                       mode_lib->ms.RequiredPrefetchPixelDataBWChroma,
+                                                                                                       mode_lib->ms.cursor_bw,
+                                                                                                       mode_lib->ms.meta_row_bandwidth_this_state,
+                                                                                                       mode_lib->ms.dpte_row_bandwidth_this_state,
+                                                                                                       mode_lib->ms.cursor_bw_pre,
+                                                                                                       mode_lib->ms.prefetch_vmrow_bw,
+                                                                                                       mode_lib->ms.NoOfDPP[j], // VBA_ERROR DPPPerSurface is not assigned at this point, should use NoOfDpp here
+                                                                                                       mode_lib->ms.UrgentBurstFactorLuma,
+                                                                                                       mode_lib->ms.UrgentBurstFactorChroma,
+                                                                                                       mode_lib->ms.UrgentBurstFactorCursor,
+                                                                                                       mode_lib->ms.UrgentBurstFactorLumaPre,
+                                                                                                       mode_lib->ms.UrgentBurstFactorChromaPre,
+                                                                                                       mode_lib->ms.UrgentBurstFactorCursorPre,
+
+                                                                                                       /* output */
+                                                                                                       &s->dummy_single[0], // dml_float_t *TotalBandwidth
+                                                                                                       &s->dummy_single[1], // dml_float_t *TotalBandwidthNotIncludingMALLPrefetch
+                                                                                                       &s->dummy_single[2], // dml_float_t *FractionOfUrgentBandwidth
+                                                                                                       &mode_lib->ms.support.ImmediateFlipSupportedForState[j]); // dml_bool_t *ImmediateFlipBandwidthSupport
+                               }
+
+                               for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                                       if (!(mode_lib->ms.policy.ImmediateFlipRequirement[k] == dml_immediate_flip_not_required) && (mode_lib->ms.ImmediateFlipSupportedForPipe[k] == false))
+                                               mode_lib->ms.support.ImmediateFlipSupportedForState[j] = false;
+                               }
+
+                       } else { // if prefetch not support, assume iflip not supported
+                               mode_lib->ms.support.ImmediateFlipSupportedForState[j] = false;
+                       }
+
+                       if (s->MaxVStartup <= __DML_VBA_MIN_VSTARTUP__ || s->AnyLinesForVMOrRowTooLarge == false) {
+                               s->NextMaxVStartup = s->MaxVStartupAllPlanes[j];
+                               for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                                       s->NextPrefetchMode[k] = s->NextPrefetchMode[k] + 1;
+
+                                       if (s->NextPrefetchMode[k] <= s->MaxPrefetchMode[k])
+                                               s->AllPrefetchModeTested = false;
+                               }
+                       } else {
+                               s->NextMaxVStartup = s->NextMaxVStartup - 1;
+                       }
+               } while (!((mode_lib->ms.support.PrefetchSupported[j] == true && mode_lib->ms.support.DynamicMetadataSupported[j] == true &&
+                                       mode_lib->ms.support.VRatioInPrefetchSupported[j] == true &&
+                                       // consider flip support is okay if when there is no hostvm and the user does't require a iflip OR the flip bw is ok
+                                       // If there is hostvm, DCN needs to support iflip for invalidation
+                                       ((s->ImmediateFlipRequiredFinal) || mode_lib->ms.support.ImmediateFlipSupportedForState[j] == true)) ||
+                                       (s->NextMaxVStartup == s->MaxVStartupAllPlanes[j] && s->AllPrefetchModeTested)));
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       mode_lib->ms.use_one_row_for_frame_this_state[k] = mode_lib->ms.use_one_row_for_frame[j][k];
+               }
+
+               s->mSOCParameters.UrgentLatency = mode_lib->ms.UrgLatency;
+               s->mSOCParameters.ExtraLatency = mode_lib->ms.ExtraLatency;
+               s->mSOCParameters.WritebackLatency = mode_lib->ms.state.writeback_latency_us;
+               s->mSOCParameters.DRAMClockChangeLatency = mode_lib->ms.state.dram_clock_change_latency_us;
+               s->mSOCParameters.FCLKChangeLatency = mode_lib->ms.state.fclk_change_latency_us;
+               s->mSOCParameters.SRExitTime = mode_lib->ms.state.sr_exit_time_us;
+               s->mSOCParameters.SREnterPlusExitTime = mode_lib->ms.state.sr_enter_plus_exit_time_us;
+               s->mSOCParameters.SRExitZ8Time = mode_lib->ms.state.sr_exit_z8_time_us;
+               s->mSOCParameters.SREnterPlusExitZ8Time = mode_lib->ms.state.sr_enter_plus_exit_z8_time_us;
+               s->mSOCParameters.USRRetrainingLatency = mode_lib->ms.state.usr_retraining_latency_us;
+               s->mSOCParameters.SMNLatency = mode_lib->ms.soc.smn_latency_us;
+
+               CalculateWatermarks_params->USRRetrainingRequiredFinal = mode_lib->ms.policy.USRRetrainingRequiredFinal;
+               CalculateWatermarks_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+               CalculateWatermarks_params->PrefetchMode = mode_lib->ms.PrefetchMode;
+               CalculateWatermarks_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+               CalculateWatermarks_params->MaxLineBufferLines = mode_lib->ms.ip.max_line_buffer_lines;
+               CalculateWatermarks_params->LineBufferSize = mode_lib->ms.ip.line_buffer_size_bits;
+               CalculateWatermarks_params->WritebackInterfaceBufferSize = mode_lib->ms.ip.writeback_interface_buffer_size_kbytes;
+               CalculateWatermarks_params->DCFCLK = mode_lib->ms.DCFCLKState[j];
+               CalculateWatermarks_params->ReturnBW = mode_lib->ms.ReturnBWPerState[j];
+               CalculateWatermarks_params->SynchronizeTimingsFinal = mode_lib->ms.policy.SynchronizeTimingsFinal;
+               CalculateWatermarks_params->SynchronizeDRRDisplaysForUCLKPStateChangeFinal = mode_lib->ms.policy.SynchronizeDRRDisplaysForUCLKPStateChangeFinal;
+               CalculateWatermarks_params->DRRDisplay = mode_lib->ms.cache_display_cfg.timing.DRRDisplay;
+               CalculateWatermarks_params->dpte_group_bytes = mode_lib->ms.dpte_group_bytes;
+               CalculateWatermarks_params->meta_row_height = mode_lib->ms.meta_row_height;
+               CalculateWatermarks_params->meta_row_height_chroma = mode_lib->ms.meta_row_height_chroma;
+               CalculateWatermarks_params->mmSOCParameters = s->mSOCParameters;
+               CalculateWatermarks_params->WritebackChunkSize = mode_lib->ms.ip.writeback_chunk_size_kbytes;
+               CalculateWatermarks_params->SOCCLK = mode_lib->ms.state.socclk_mhz;
+               CalculateWatermarks_params->DCFClkDeepSleep = mode_lib->ms.ProjectedDCFCLKDeepSleep[j];
+               CalculateWatermarks_params->DETBufferSizeY = mode_lib->ms.DETBufferSizeYThisState;
+               CalculateWatermarks_params->DETBufferSizeC = mode_lib->ms.DETBufferSizeCThisState;
+               CalculateWatermarks_params->SwathHeightY = mode_lib->ms.SwathHeightYThisState;
+               CalculateWatermarks_params->SwathHeightC = mode_lib->ms.SwathHeightCThisState;
+               CalculateWatermarks_params->LBBitPerPixel = mode_lib->ms.cache_display_cfg.plane.LBBitPerPixel;
+               CalculateWatermarks_params->SwathWidthY = mode_lib->ms.SwathWidthYThisState;
+               CalculateWatermarks_params->SwathWidthC = mode_lib->ms.SwathWidthCThisState;
+               CalculateWatermarks_params->HRatio = mode_lib->ms.cache_display_cfg.plane.HRatio;
+               CalculateWatermarks_params->HRatioChroma = mode_lib->ms.cache_display_cfg.plane.HRatioChroma;
+               CalculateWatermarks_params->VTaps = mode_lib->ms.cache_display_cfg.plane.VTaps;
+               CalculateWatermarks_params->VTapsChroma = mode_lib->ms.cache_display_cfg.plane.VTapsChroma;
+               CalculateWatermarks_params->VRatio = mode_lib->ms.cache_display_cfg.plane.VRatio;
+               CalculateWatermarks_params->VRatioChroma = mode_lib->ms.cache_display_cfg.plane.VRatioChroma;
+               CalculateWatermarks_params->HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal;
+               CalculateWatermarks_params->VTotal = mode_lib->ms.cache_display_cfg.timing.VTotal;
+               CalculateWatermarks_params->VActive = mode_lib->ms.cache_display_cfg.timing.VActive;
+               CalculateWatermarks_params->PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock;
+               CalculateWatermarks_params->BlendingAndTiming = mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming;
+               CalculateWatermarks_params->DPPPerSurface = mode_lib->ms.NoOfDPPThisState;
+               CalculateWatermarks_params->BytePerPixelDETY = mode_lib->ms.BytePerPixelInDETY;
+               CalculateWatermarks_params->BytePerPixelDETC = mode_lib->ms.BytePerPixelInDETC;
+               CalculateWatermarks_params->DSTXAfterScaler = s->DSTXAfterScaler;
+               CalculateWatermarks_params->DSTYAfterScaler = s->DSTYAfterScaler;
+               CalculateWatermarks_params->WritebackEnable = mode_lib->ms.cache_display_cfg.writeback.WritebackEnable;
+               CalculateWatermarks_params->WritebackPixelFormat = mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat;
+               CalculateWatermarks_params->WritebackDestinationWidth = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth;
+               CalculateWatermarks_params->WritebackDestinationHeight = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight;
+               CalculateWatermarks_params->WritebackSourceHeight = mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight;
+               CalculateWatermarks_params->UnboundedRequestEnabled = mode_lib->ms.UnboundedRequestEnabledThisState;
+               CalculateWatermarks_params->CompressedBufferSizeInkByte = mode_lib->ms.CompressedBufferSizeInkByteThisState;
+
+               // Output
+               CalculateWatermarks_params->Watermark = &s->dummy_watermark; // Watermarks *Watermark
+               CalculateWatermarks_params->DRAMClockChangeSupport = &mode_lib->ms.support.DRAMClockChangeSupport[j];
+               CalculateWatermarks_params->MaxActiveDRAMClockChangeLatencySupported = &s->dummy_single_array[0]; // dml_float_t *MaxActiveDRAMClockChangeLatencySupported[]
+               CalculateWatermarks_params->SubViewportLinesNeededInMALL = &mode_lib->ms.SubViewportLinesNeededInMALL[j]; // dml_uint_t SubViewportLinesNeededInMALL[]
+               CalculateWatermarks_params->FCLKChangeSupport = &mode_lib->ms.support.FCLKChangeSupport[j];
+               CalculateWatermarks_params->MaxActiveFCLKChangeLatencySupported = &s->dummy_single[0]; // dml_float_t *MaxActiveFCLKChangeLatencySupported
+               CalculateWatermarks_params->USRRetrainingSupport = &mode_lib->ms.support.USRRetrainingSupport[j];
+               CalculateWatermarks_params->ActiveDRAMClockChangeLatencyMargin = mode_lib->ms.support.ActiveDRAMClockChangeLatencyMargin;
+
+               CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(&mode_lib->scratch,
+                       CalculateWatermarks_params);
+
+       } // for j
+       // End of Prefetch Check
+       dml_print("DML::%s: Done prefetch calculation\n", __func__);
+
+       /*Cursor Support Check*/
+       mode_lib->ms.support.CursorSupport = true;
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               if (mode_lib->ms.cache_display_cfg.plane.CursorWidth[k] > 0.0) {
+                       if (mode_lib->ms.cache_display_cfg.plane.CursorBPP[k] == 64 && mode_lib->ms.ip.cursor_64bpp_support == false) {
+                               mode_lib->ms.support.CursorSupport = false;
+                       }
+               }
+       }
+
+       /*Valid Pitch Check*/
+       mode_lib->ms.support.PitchSupport = true;
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               mode_lib->ms.support.AlignedYPitch[k] = dml_ceil(
+                               dml_max(mode_lib->ms.cache_display_cfg.surface.PitchY[k], mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY[k]),
+                               mode_lib->ms.MacroTileWidthY[k]);
+               if (mode_lib->ms.cache_display_cfg.surface.DCCEnable[k] == true) {
+                       mode_lib->ms.support.AlignedDCCMetaPitchY[k] = dml_ceil(dml_max(mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchY[k], mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY[k]), 64.0 * mode_lib->ms.Read256BlockWidthY[k]);
+               } else {
+                       mode_lib->ms.support.AlignedDCCMetaPitchY[k] = mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchY[k];
+               }
+               if (mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_64
+                       && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_32
+                       && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_16
+                       && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_mono_16
+                       && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_rgbe
+                       && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_mono_8) {
+                       mode_lib->ms.support.AlignedCPitch[k] = dml_ceil(dml_max(mode_lib->ms.cache_display_cfg.surface.PitchC[k], mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC[k]), mode_lib->ms.MacroTileWidthC[k]);
+                       if (mode_lib->ms.cache_display_cfg.surface.DCCEnable[k] == true) {
+                               mode_lib->ms.support.AlignedDCCMetaPitchC[k] = dml_ceil(dml_max(mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchC[k], mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC[k]), 64.0 * mode_lib->ms.Read256BlockWidthC[k]);
+                       } else {
+                               mode_lib->ms.support.AlignedDCCMetaPitchC[k] = mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchC[k];
+                       }
+               } else {
+                       mode_lib->ms.support.AlignedCPitch[k] = mode_lib->ms.cache_display_cfg.surface.PitchC[k];
+                       mode_lib->ms.support.AlignedDCCMetaPitchC[k] = mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchC[k];
+               }
+               if (mode_lib->ms.support.AlignedYPitch[k] > mode_lib->ms.cache_display_cfg.surface.PitchY[k] || mode_lib->ms.support.AlignedCPitch[k] > mode_lib->ms.cache_display_cfg.surface.PitchC[k] ||
+                       mode_lib->ms.support.AlignedDCCMetaPitchY[k] > mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchY[k] || mode_lib->ms.support.AlignedDCCMetaPitchC[k] > mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchC[k]) {
+                       mode_lib->ms.support.PitchSupport = false;
+               }
+       }
+
+       mode_lib->ms.support.ViewportExceedsSurface = false;
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               if (mode_lib->ms.cache_display_cfg.plane.ViewportWidth[k] > mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY[k] || mode_lib->ms.cache_display_cfg.plane.ViewportHeight[k] > mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY[k]) {
+                       mode_lib->ms.support.ViewportExceedsSurface = true;
+                       if (mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_64 && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_32 &&
+                               mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_16 && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_8 && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_rgbe) {
+                               if (mode_lib->ms.cache_display_cfg.plane.ViewportWidthChroma[k] > mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC[k] || mode_lib->ms.cache_display_cfg.plane.ViewportHeightChroma[k] > mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC[k]) {
+                                       mode_lib->ms.support.ViewportExceedsSurface = true;
+                               }
+                       }
+               }
+       }
+
+       /*Mode Support, Voltage State and SOC Configuration*/
+       for (j = 0; j < 2; j++) { // j iterator is for the combine mode off or on
+               dml_print("DML::%s: checking support for j=%u\n", __func__, j);
+               dml_print("DML::%s: state_idx=%0d max_state_idx=%0d\n", __func__, mode_lib->ms.state_idx, mode_lib->ms.max_state_idx);
+
+               s->is_max_pwr_state = (mode_lib->ms.max_state_idx == mode_lib->ms.state_idx);
+               s->is_max_dram_pwr_state = (mode_lib->ms.max_state.dram_speed_mts == mode_lib->ms.state.dram_speed_mts);
+
+               s->dram_clock_change_support = (!mode_lib->ms.policy.DRAMClockChangeRequirementFinal ||
+                                                                                       (s->is_max_dram_pwr_state && mode_lib->policy.AssumeModeSupportAtMaxPwrStateEvenDRAMClockChangeNotSupported) ||
+                                                                                       mode_lib->ms.support.DRAMClockChangeSupport[j] != dml_dram_clock_change_unsupported);
+               s->f_clock_change_support = (!mode_lib->ms.policy.FCLKChangeRequirementFinal ||
+                                                                                       (s->is_max_pwr_state && mode_lib->policy.AssumeModeSupportAtMaxPwrStateEvenFClockChangeNotSupported) ||
+                                                                                       mode_lib->ms.support.FCLKChangeSupport[j] != dml_fclock_change_unsupported);
+
+               if (mode_lib->ms.support.ScaleRatioAndTapsSupport == true
+                       && mode_lib->ms.support.SourceFormatPixelAndScanSupport == true
+                       && mode_lib->ms.support.ViewportSizeSupport[j] == true
+                       && !mode_lib->ms.support.LinkRateDoesNotMatchDPVersion
+                       && !mode_lib->ms.support.LinkRateForMultistreamNotIndicated
+                       && !mode_lib->ms.support.BPPForMultistreamNotIndicated
+                       && !mode_lib->ms.support.MultistreamWithHDMIOreDP
+                       && !mode_lib->ms.support.ExceededMultistreamSlots
+                       && !mode_lib->ms.support.MSOOrODMSplitWithNonDPLink
+                       && !mode_lib->ms.support.NotEnoughLanesForMSO
+                       && mode_lib->ms.support.LinkCapacitySupport == true
+                       && !mode_lib->ms.support.P2IWith420
+                       && !mode_lib->ms.support.DSCOnlyIfNecessaryWithBPP
+                       && !mode_lib->ms.support.DSC422NativeNotSupported
+                       && !mode_lib->ms.support.MPCCombineMethodIncompatible
+                       && mode_lib->ms.support.ODMCombineTwoToOneSupportCheckOK == true
+                       && mode_lib->ms.support.ODMCombineFourToOneSupportCheckOK == true
+                       && mode_lib->ms.support.NotEnoughDSCUnits == false
+                       && !mode_lib->ms.support.NotEnoughDSCSlices
+                       && !mode_lib->ms.support.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe
+                       && !mode_lib->ms.support.InvalidCombinationOfMALLUseForPStateAndStaticScreen
+                       && mode_lib->ms.support.DSCCLKRequiredMoreThanSupported == false
+                       && mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport
+                       && mode_lib->ms.support.DTBCLKRequiredMoreThanSupported == false
+                       && !mode_lib->ms.support.InvalidCombinationOfMALLUseForPState
+                       && !mode_lib->ms.support.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified
+                       && mode_lib->ms.support.ROBSupport[j] == true
+                       && mode_lib->ms.support.DISPCLK_DPPCLK_Support[j] == true
+                       && mode_lib->ms.support.TotalAvailablePipesSupport[j] == true
+                       && mode_lib->ms.support.NumberOfOTGSupport == true
+                       && mode_lib->ms.support.NumberOfHDMIFRLSupport == true
+                       && mode_lib->ms.support.NumberOfDP2p0Support == true
+                       && mode_lib->ms.support.EnoughWritebackUnits == true
+                       && mode_lib->ms.support.WritebackLatencySupport == true
+                       && mode_lib->ms.support.WritebackScaleRatioAndTapsSupport == true
+                       && mode_lib->ms.support.CursorSupport == true
+                       && mode_lib->ms.support.PitchSupport == true
+                       && mode_lib->ms.support.ViewportExceedsSurface == false
+                       && mode_lib->ms.support.PrefetchSupported[j] == true
+                       && mode_lib->ms.support.VActiveBandwithSupport[j] == true
+                       && mode_lib->ms.support.DynamicMetadataSupported[j] == true
+                       && mode_lib->ms.support.TotalVerticalActiveBandwidthSupport[j] == true
+                       && mode_lib->ms.support.VRatioInPrefetchSupported[j] == true
+                       && mode_lib->ms.support.PTEBufferSizeNotExceeded[j] == true
+                       && mode_lib->ms.support.DCCMetaBufferSizeNotExceeded[j] == true
+                       && mode_lib->ms.support.NonsupportedDSCInputBPC == false
+                       && !mode_lib->ms.support.ExceededMALLSize
+                       && ((mode_lib->ms.cache_display_cfg.plane.HostVMEnable == false && !s->ImmediateFlipRequiredFinal) || mode_lib->ms.support.ImmediateFlipSupportedForState[j])
+                       && s->dram_clock_change_support == true
+                       && s->f_clock_change_support == true
+                       && (!mode_lib->ms.policy.USRRetrainingRequiredFinal || mode_lib->ms.support.USRRetrainingSupport[j])) {
+                       dml_print("DML::%s: mode is supported\n", __func__);
+                       mode_lib->ms.support.ModeSupport[j] = true;
+               } else {
+                       dml_print("DML::%s: mode is NOT supported\n", __func__);
+                       mode_lib->ms.support.ModeSupport[j] = false;
+                       dml_print_mode_support(mode_lib, j);
+               }
+       }
+
+       mode_lib->ms.support.MaximumMPCCombine = 0;
+       mode_lib->ms.support.ModeIsSupported = 0;
+       if (mode_lib->ms.support.ModeSupport[0] == true || mode_lib->ms.support.ModeSupport[1] == true) {  // if the mode is supported by either no combine or mpccombine
+               mode_lib->ms.support.ModeIsSupported = mode_lib->ms.support.ModeSupport[0] == true || mode_lib->ms.support.ModeSupport[1] == true;
+
+               // Determine if MPC combine is necessary, depends on if using MPC combine will help dram clock change or fclk change, etc.
+               if ((mode_lib->ms.support.ModeSupport[0] == false && mode_lib->ms.support.ModeSupport[1] == true) || s->MPCCombineMethodAsPossible ||
+                       (s->MPCCombineMethodAsNeededForPStateChangeAndVoltage && mode_lib->ms.policy.DRAMClockChangeRequirementFinal &&
+                       (((mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vactive || mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vactive_w_mall_full_frame || mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vactive_w_mall_sub_vp) &&
+                       !(mode_lib->ms.support.DRAMClockChangeSupport[0] == dml_dram_clock_change_vactive || mode_lib->ms.support.DRAMClockChangeSupport[0] == dml_dram_clock_change_vactive_w_mall_full_frame || mode_lib->ms.support.DRAMClockChangeSupport[0] == dml_dram_clock_change_vactive_w_mall_sub_vp)) ||
+                       ((mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vblank || mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vblank_drr
+               || mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vblank_w_mall_full_frame || mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vblank_drr_w_mall_full_frame
+               || mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vblank_w_mall_sub_vp || mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vblank_drr_w_mall_sub_vp
+               ) &&
+                               mode_lib->ms.support.DRAMClockChangeSupport[0] == dml_dram_clock_change_unsupported)))
+                       || (s->MPCCombineMethodAsNeededForPStateChangeAndVoltage && mode_lib->ms.policy.FCLKChangeRequirementFinal &&
+                               ((mode_lib->ms.support.FCLKChangeSupport[1] == dml_fclock_change_vactive && mode_lib->ms.support.FCLKChangeSupport[0] != dml_fclock_change_vactive) ||
+                               (mode_lib->ms.support.FCLKChangeSupport[1] == dml_fclock_change_vblank && mode_lib->ms.support.FCLKChangeSupport[0] == dml_fclock_change_unsupported)))) {
+                       mode_lib->ms.support.MaximumMPCCombine = 1;
+               } else {
+                       mode_lib->ms.support.MaximumMPCCombine = 0;
+               }
+       }
+
+       // Since now the mode_support work on 1 particular power state, so there is only 1 state idx (index 0).
+       mode_lib->ms.support.ImmediateFlipSupport          = mode_lib->ms.support.ImmediateFlipSupportedForState[mode_lib->ms.support.MaximumMPCCombine];   // Consider flip support if max combine support imm flip
+       mode_lib->ms.support.UnboundedRequestEnabled       = mode_lib->ms.UnboundedRequestEnabledAllStates[mode_lib->ms.support.MaximumMPCCombine];         // Not used, informational
+       mode_lib->ms.support.CompressedBufferSizeInkByte   = mode_lib->ms.CompressedBufferSizeInkByteAllStates[mode_lib->ms.support.MaximumMPCCombine];     // Not used, informational
+
+       dml_print("DML::%s: ModeIsSupported                = %u\n", __func__, mode_lib->ms.support.ModeIsSupported);
+       dml_print("DML::%s: MaximumMPCCombine              = %u\n", __func__, mode_lib->ms.support.MaximumMPCCombine);
+       dml_print("DML::%s: ImmediateFlipSupport           = %u\n", __func__, mode_lib->ms.support.ImmediateFlipSupport);
+       dml_print("DML::%s: UnboundedRequestEnabled        = %u\n", __func__, mode_lib->ms.support.UnboundedRequestEnabled);
+       dml_print("DML::%s: CompressedBufferSizeInkByte    = %u\n", __func__, mode_lib->ms.support.CompressedBufferSizeInkByte);
+
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               mode_lib->ms.support.MPCCombineEnable[k]   = mode_lib->ms.MPCCombine[mode_lib->ms.support.MaximumMPCCombine][k];
+               mode_lib->ms.support.DPPPerSurface[k]      = mode_lib->ms.NoOfDPP[mode_lib->ms.support.MaximumMPCCombine][k];
+               mode_lib->ms.SwathHeightY[k]               = mode_lib->ms.SwathHeightYAllStates[mode_lib->ms.support.MaximumMPCCombine][k];
+               mode_lib->ms.SwathHeightC[k]               = mode_lib->ms.SwathHeightCAllStates[mode_lib->ms.support.MaximumMPCCombine][k];
+               mode_lib->ms.DETBufferSizeInKByte[k]       = mode_lib->ms.DETBufferSizeInKByteAllStates[mode_lib->ms.support.MaximumMPCCombine][k];
+               mode_lib->ms.DETBufferSizeY[k]             = mode_lib->ms.DETBufferSizeYAllStates[mode_lib->ms.support.MaximumMPCCombine][k];
+               mode_lib->ms.DETBufferSizeC[k]             = mode_lib->ms.DETBufferSizeCAllStates[mode_lib->ms.support.MaximumMPCCombine][k];
+       }
+
+       mode_lib->ms.DRAMSpeed     = mode_lib->ms.state.dram_speed_mts;
+       mode_lib->ms.FabricClock   = mode_lib->ms.state.fabricclk_mhz;
+       mode_lib->ms.SOCCLK        = mode_lib->ms.state.socclk_mhz;
+       mode_lib->ms.DCFCLK        = mode_lib->ms.DCFCLKState[mode_lib->ms.support.MaximumMPCCombine];
+       mode_lib->ms.ReturnBW      = mode_lib->ms.ReturnBWPerState[mode_lib->ms.support.MaximumMPCCombine];
+       mode_lib->ms.ReturnDRAMBW  = mode_lib->ms.ReturnDRAMBWPerState[mode_lib->ms.support.MaximumMPCCombine];
+
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k) {
+                       mode_lib->ms.support.ODMMode[k] = mode_lib->ms.ODMModePerState[k];
+               } else {
+                       mode_lib->ms.support.ODMMode[k] = dml_odm_mode_bypass;
+               }
+
+               mode_lib->ms.support.DSCEnabled[k] = mode_lib->ms.RequiresDSC[k];
+               mode_lib->ms.support.FECEnabled[k] = mode_lib->ms.RequiresFEC[k];
+               mode_lib->ms.support.OutputBpp[k] = mode_lib->ms.OutputBppPerState[k];
+               mode_lib->ms.support.OutputType[k] = mode_lib->ms.OutputTypePerState[k];
+               mode_lib->ms.support.OutputRate[k] = mode_lib->ms.OutputRatePerState[k];
+               mode_lib->ms.support.SubViewportLinesNeededInMALL[k] = mode_lib->ms.SubViewportLinesNeededInMALL[k];
+       }
+
+       return mode_lib->ms.support.ModeIsSupported;
+} // dml_core_mode_support
+
+/// @brief This function calculates some parameters thats are needed ahead of the mode programming function all
+void dml_core_mode_support_partial(struct display_mode_lib_st *mode_lib)
+{
+       CalculateMaxDETAndMinCompressedBufferSize(
+                                                               mode_lib->ms.ip.config_return_buffer_size_in_kbytes,
+                                                               mode_lib->ms.ip.config_return_buffer_segment_size_in_kbytes,
+                                                               mode_lib->ms.ip.rob_buffer_size_kbytes,
+                                                               mode_lib->ms.ip.max_num_dpp,
+                                                               mode_lib->ms.policy.NomDETInKByteOverrideEnable,
+                                                               mode_lib->ms.policy.NomDETInKByteOverrideValue,
+
+                                                               /* Output */
+                                                               &mode_lib->ms.MaxTotalDETInKByte,
+                                                               &mode_lib->ms.NomDETInKByte,
+                                                               &mode_lib->ms.MinCompressedBufferSizeInKByte);
+
+       PixelClockAdjustmentForProgressiveToInterlaceUnit(&mode_lib->ms.cache_display_cfg, mode_lib->ms.ip.ptoi_supported);
+
+       mode_lib->ms.ReturnBW = dml_get_return_bw_mbps(&mode_lib->ms.soc,
+                                                                                                       mode_lib->ms.state.use_ideal_dram_bw_strobe,
+                                                                                                       mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+                                                                                                       mode_lib->ms.DCFCLK,
+                                                                                                       mode_lib->ms.FabricClock,
+                                                                                                       mode_lib->ms.DRAMSpeed);
+       dml_print("DML::%s: ReturnBW = %f\n", __func__, mode_lib->ms.ReturnBW);
+
+} // dml_core_mode_support_partial
+
+/// @brief This is the mode programming function. It is assumed the display cfg is support at the given power state
+void dml_core_mode_programming(struct display_mode_lib_st *mode_lib, const struct dml_clk_cfg_st *clk_cfg)
+{
+       struct dml_core_mode_programming_locals_st *s = &mode_lib->scratch.dml_core_mode_programming_locals;
+       struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params_st *CalculateWatermarks_params = &mode_lib->scratch.CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params;
+       struct CalculateVMRowAndSwath_params_st *CalculateVMRowAndSwath_params = &mode_lib->scratch.CalculateVMRowAndSwath_params;
+       struct CalculateSwathAndDETConfiguration_params_st *CalculateSwathAndDETConfiguration_params = &mode_lib->scratch.CalculateSwathAndDETConfiguration_params;
+       struct CalculateStutterEfficiency_params_st *CalculateStutterEfficiency_params = &mode_lib->scratch.CalculateStutterEfficiency_params;
+       struct CalculatePrefetchSchedule_params_st *CalculatePrefetchSchedule_params = &mode_lib->scratch.CalculatePrefetchSchedule_params;
+
+       struct mode_program_st   *locals    = &mode_lib->mp;
+       struct DmlPipe *myPipe;
+       dml_uint_t j = 0, k = 0;
+       dml_float_t TWait;
+       dml_bool_t isInterlaceTiming;
+
+       mode_lib->ms.num_active_planes = dml_get_num_active_planes(&mode_lib->ms.cache_display_cfg);
+       mode_lib->mp.num_active_pipes = dml_get_num_active_pipes(&mode_lib->ms.cache_display_cfg);
+       dml_calc_pipe_plane_mapping(&mode_lib->ms.cache_display_cfg.hw, mode_lib->mp.pipe_plane);
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: --- START --- \n",  __func__);
+       dml_print("DML::%s: num_active_planes = %u\n", __func__, mode_lib->ms.num_active_planes);
+       dml_print("DML::%s: num_active_pipes = %u\n", __func__, mode_lib->mp.num_active_pipes);
+#endif
+
+       s->DSCFormatFactor          = 0;
+
+       // Unlike dppclk and dispclk which can be calculated in mode_programming
+       // DCFCLK is calculated in mode_support (which is the state bbox dcfclk or min dcfclk if min dcfclk option is used in mode support calculation)
+       if (clk_cfg->dcfclk_option != dml_use_override_freq)
+               locals->Dcfclk = mode_lib->ms.DCFCLK;
+       else
+               locals->Dcfclk = clk_cfg->dcfclk_freq_mhz;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print_dml_policy(&mode_lib->ms.policy);
+       dml_print_soc_state_bounding_box(&mode_lib->ms.state);
+       dml_print_soc_bounding_box(&mode_lib->ms.soc);
+       dml_print_clk_cfg(clk_cfg);
+
+       dml_print("DML::%s: ImmediateFlipSupport = %u\n", __func__, mode_lib->ms.support.ImmediateFlipSupport);
+       dml_print("DML::%s: Using DCFCLK = %f\n", __func__, locals->Dcfclk);
+       dml_print("DML::%s: Using SOCCLK = %f\n", __func__, mode_lib->ms.SOCCLK);
+#endif
+
+       locals->WritebackDISPCLK = 0.0;
+       locals->GlobalDPPCLK = 0.0;
+
+       // DISPCLK and DPPCLK Calculation
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k]) {
+                       locals->WritebackDISPCLK =
+                                       dml_max(
+                                                       locals->WritebackDISPCLK,
+                                                       CalculateWriteBackDISPCLK(
+                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[k],
+                                                                       mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[k],
+                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[k],
+                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackHTaps[k],
+                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[k],
+                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackSourceWidth[k],
+                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k],
+                                                                       mode_lib->ms.cache_display_cfg.timing.HTotal[k],
+                                                                       mode_lib->ms.ip.writeback_line_buffer_buffer_size,
+                                                                       mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz));
+               }
+       }
+
+       locals->Dispclk_calculated = locals->WritebackDISPCLK;
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k) {
+                               locals->Dispclk_calculated = dml_max(locals->Dispclk_calculated, CalculateRequiredDispclk(
+                                                                                                                                                                       mode_lib->ms.cache_display_cfg.hw.ODMMode[k],
+                                                                                                                                                                       mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+                                                                                                                                                                       mode_lib->ms.soc.dcn_downspread_percent,
+                                                                                                                                                                       mode_lib->ms.ip.dispclk_ramp_margin_percent,
+                                                                                                                                                                       mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz,
+                                                                                                                                                                       mode_lib->ms.max_state.dispclk_mhz));
+               }
+       }
+       if (clk_cfg->dispclk_option == dml_use_required_freq)
+               locals->Dispclk = locals->Dispclk_calculated;
+       else if (clk_cfg->dispclk_option == dml_use_override_freq)
+               locals->Dispclk = clk_cfg->dispclk_freq_mhz;
+       else
+               locals->Dispclk = mode_lib->ms.state.dispclk_mhz;
+#ifdef __DML_VBA_DEBUG__
+        dml_print("DML::%s: Using Dispclk = %f\n", __func__, locals->Dispclk);
+#endif
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               CalculateSinglePipeDPPCLKAndSCLThroughput(
+                               mode_lib->ms.cache_display_cfg.plane.HRatio[k],
+                               mode_lib->ms.cache_display_cfg.plane.HRatioChroma[k],
+                               mode_lib->ms.cache_display_cfg.plane.VRatio[k],
+                               mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k],
+                               mode_lib->ms.ip.max_dchub_pscl_bw_pix_per_clk,
+                               mode_lib->ms.ip.max_pscl_lb_bw_pix_per_clk,
+                               mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+                               mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k],
+                               mode_lib->ms.cache_display_cfg.plane.HTaps[k],
+                               mode_lib->ms.cache_display_cfg.plane.HTapsChroma[k],
+                               mode_lib->ms.cache_display_cfg.plane.VTaps[k],
+                               mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k],
+
+                               /* Output */
+                               &locals->PSCL_THROUGHPUT[k],
+                               &locals->PSCL_THROUGHPUT_CHROMA[k],
+                               &locals->DPPCLKUsingSingleDPP[k]);
+       }
+
+       CalculateDPPCLK(mode_lib->ms.num_active_planes,
+                                       mode_lib->ms.soc.dcn_downspread_percent,
+                                       mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz,
+                                       locals->DPPCLKUsingSingleDPP,
+                                       mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+                                       /* Output */
+                                       &locals->GlobalDPPCLK,
+                                       locals->Dppclk_calculated);
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (clk_cfg->dppclk_option[k] == dml_use_required_freq)
+                       locals->Dppclk[k] = locals->Dppclk_calculated[k];
+               else if (clk_cfg->dppclk_option[k] == dml_use_override_freq)
+                       locals->Dppclk[k] = clk_cfg->dppclk_freq_mhz[k];
+               else
+                       locals->Dppclk[k] = mode_lib->ms.state.dppclk_mhz;
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: Using Dppclk[%0d] = %f\n", __func__, k, locals->Dppclk[k]);
+#endif
+       }
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               CalculateBytePerPixelAndBlockSizes(
+                               mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k],
+                               mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k],
+
+                               /* Output */
+                               &locals->BytePerPixelY[k],
+                               &locals->BytePerPixelC[k],
+                               &locals->BytePerPixelDETY[k],
+                               &locals->BytePerPixelDETC[k],
+                               &locals->BlockHeight256BytesY[k],
+                               &locals->BlockHeight256BytesC[k],
+                               &locals->BlockWidth256BytesY[k],
+                               &locals->BlockWidth256BytesC[k],
+                               &locals->BlockHeightY[k],
+                               &locals->BlockHeightC[k],
+                               &locals->BlockWidthY[k],
+                               &locals->BlockWidthC[k]);
+       }
+
+
+       dml_print("DML::%s: %u\n", __func__, __LINE__);
+       CalculateSwathWidth(
+               false,  // ForceSingleDPP
+               mode_lib->ms.num_active_planes,
+               mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat,
+               mode_lib->ms.cache_display_cfg.plane.SourceScan,
+               mode_lib->ms.cache_display_cfg.plane.ViewportStationary,
+               mode_lib->ms.cache_display_cfg.plane.ViewportWidth,
+               mode_lib->ms.cache_display_cfg.plane.ViewportHeight,
+               mode_lib->ms.cache_display_cfg.plane.ViewportXStart,
+               mode_lib->ms.cache_display_cfg.plane.ViewportYStart,
+               mode_lib->ms.cache_display_cfg.plane.ViewportXStartC,
+               mode_lib->ms.cache_display_cfg.plane.ViewportYStartC,
+               mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY,
+               mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC,
+               mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY,
+               mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC,
+               mode_lib->ms.cache_display_cfg.hw.ODMMode,
+               locals->BytePerPixelY,
+               locals->BytePerPixelC,
+               locals->BlockHeight256BytesY,
+               locals->BlockHeight256BytesC,
+               locals->BlockWidth256BytesY,
+               locals->BlockWidth256BytesC,
+               mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming,
+               mode_lib->ms.cache_display_cfg.timing.HActive,
+               mode_lib->ms.cache_display_cfg.plane.HRatio,
+               mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+
+               /* Output */
+               locals->SwathWidthSingleDPPY,
+               locals->SwathWidthSingleDPPC,
+               locals->SwathWidthY,
+               locals->SwathWidthC,
+               s->dummy_integer_array[0], // dml_uint_t MaximumSwathHeightY[]
+               s->dummy_integer_array[1], // dml_uint_t MaximumSwathHeightC[]
+               locals->swath_width_luma_ub,
+               locals->swath_width_chroma_ub);
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               locals->ReadBandwidthSurfaceLuma[k] = locals->SwathWidthSingleDPPY[k] * locals->BytePerPixelY[k] / (mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * mode_lib->ms.cache_display_cfg.plane.VRatio[k];
+               locals->ReadBandwidthSurfaceChroma[k] = locals->SwathWidthSingleDPPC[k] * locals->BytePerPixelC[k] / (mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k];
+               dml_print("DML::%s: ReadBandwidthSurfaceLuma[%i] = %fBps\n", __func__, k, locals->ReadBandwidthSurfaceLuma[k]);
+               dml_print("DML::%s: ReadBandwidthSurfaceChroma[%i] = %fBps\n", __func__, k, locals->ReadBandwidthSurfaceChroma[k]);
+       }
+
+       CalculateSwathAndDETConfiguration_params->DETSizeOverride = mode_lib->ms.cache_display_cfg.plane.DETSizeOverride;
+       CalculateSwathAndDETConfiguration_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+       CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSizeInKByte = mode_lib->ms.ip.config_return_buffer_size_in_kbytes;
+       CalculateSwathAndDETConfiguration_params->ROBBufferSizeInKByte = mode_lib->ms.ip.rob_buffer_size_kbytes;
+       CalculateSwathAndDETConfiguration_params->MaxTotalDETInKByte = mode_lib->ms.MaxTotalDETInKByte;
+       CalculateSwathAndDETConfiguration_params->MinCompressedBufferSizeInKByte = mode_lib->ms.MinCompressedBufferSizeInKByte;
+       CalculateSwathAndDETConfiguration_params->PixelChunkSizeInKByte = mode_lib->ms.ip.pixel_chunk_size_kbytes;
+       CalculateSwathAndDETConfiguration_params->ForceSingleDPP = false;
+       CalculateSwathAndDETConfiguration_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+       CalculateSwathAndDETConfiguration_params->nomDETInKByte = mode_lib->ms.NomDETInKByte;
+       CalculateSwathAndDETConfiguration_params->UseUnboundedRequestingFinal = mode_lib->ms.policy.UseUnboundedRequesting;
+       CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSegmentSizeInkByte = mode_lib->ms.ip.config_return_buffer_segment_size_in_kbytes;
+       CalculateSwathAndDETConfiguration_params->CompressedBufferSegmentSizeInkByteFinal = mode_lib->ms.ip.compressed_buffer_segment_size_in_kbytes;
+       CalculateSwathAndDETConfiguration_params->Output = s->dummy_output_encoder_array;
+       CalculateSwathAndDETConfiguration_params->ReadBandwidthLuma = locals->ReadBandwidthSurfaceLuma;
+       CalculateSwathAndDETConfiguration_params->ReadBandwidthChroma = locals->ReadBandwidthSurfaceChroma;
+       CalculateSwathAndDETConfiguration_params->MaximumSwathWidthLuma = s->dummy_single_array[0];
+       CalculateSwathAndDETConfiguration_params->MaximumSwathWidthChroma = s->dummy_single_array[1];
+       CalculateSwathAndDETConfiguration_params->SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan;
+       CalculateSwathAndDETConfiguration_params->ViewportStationary = mode_lib->ms.cache_display_cfg.plane.ViewportStationary;
+       CalculateSwathAndDETConfiguration_params->SourcePixelFormat = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat;
+       CalculateSwathAndDETConfiguration_params->SurfaceTiling = mode_lib->ms.cache_display_cfg.surface.SurfaceTiling;
+       CalculateSwathAndDETConfiguration_params->ViewportWidth = mode_lib->ms.cache_display_cfg.plane.ViewportWidth;
+       CalculateSwathAndDETConfiguration_params->ViewportHeight = mode_lib->ms.cache_display_cfg.plane.ViewportHeight;
+       CalculateSwathAndDETConfiguration_params->ViewportXStart = mode_lib->ms.cache_display_cfg.plane.ViewportXStart;
+       CalculateSwathAndDETConfiguration_params->ViewportYStart = mode_lib->ms.cache_display_cfg.plane.ViewportYStart;
+       CalculateSwathAndDETConfiguration_params->ViewportXStartC = mode_lib->ms.cache_display_cfg.plane.ViewportXStartC;
+       CalculateSwathAndDETConfiguration_params->ViewportYStartC = mode_lib->ms.cache_display_cfg.plane.ViewportYStartC;
+       CalculateSwathAndDETConfiguration_params->SurfaceWidthY = mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY;
+       CalculateSwathAndDETConfiguration_params->SurfaceWidthC = mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC;
+       CalculateSwathAndDETConfiguration_params->SurfaceHeightY = mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY;
+       CalculateSwathAndDETConfiguration_params->SurfaceHeightC = mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC;
+       CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightY = locals->BlockHeight256BytesY;
+       CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightC = locals->BlockHeight256BytesC;
+       CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthY = locals->BlockWidth256BytesY;
+       CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthC = locals->BlockWidth256BytesC;
+       CalculateSwathAndDETConfiguration_params->ODMMode = mode_lib->ms.cache_display_cfg.hw.ODMMode;
+       CalculateSwathAndDETConfiguration_params->BlendingAndTiming = mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming;
+       CalculateSwathAndDETConfiguration_params->BytePerPixY = locals->BytePerPixelY;
+       CalculateSwathAndDETConfiguration_params->BytePerPixC = locals->BytePerPixelC;
+       CalculateSwathAndDETConfiguration_params->BytePerPixDETY = locals->BytePerPixelDETY;
+       CalculateSwathAndDETConfiguration_params->BytePerPixDETC = locals->BytePerPixelDETC;
+       CalculateSwathAndDETConfiguration_params->HActive = mode_lib->ms.cache_display_cfg.timing.HActive;
+       CalculateSwathAndDETConfiguration_params->HRatio = mode_lib->ms.cache_display_cfg.plane.HRatio;
+       CalculateSwathAndDETConfiguration_params->HRatioChroma = mode_lib->ms.cache_display_cfg.plane.HRatioChroma;
+       CalculateSwathAndDETConfiguration_params->DPPPerSurface = mode_lib->ms.cache_display_cfg.hw.DPPPerSurface;
+       CalculateSwathAndDETConfiguration_params->swath_width_luma_ub = s->dummy_long_array[0];
+       CalculateSwathAndDETConfiguration_params->swath_width_chroma_ub = s->dummy_long_array[1];
+       CalculateSwathAndDETConfiguration_params->SwathWidth = s->dummy_long_array[2];
+       CalculateSwathAndDETConfiguration_params->SwathWidthChroma = s->dummy_long_array[3];
+       CalculateSwathAndDETConfiguration_params->SwathHeightY = locals->SwathHeightY;
+       CalculateSwathAndDETConfiguration_params->SwathHeightC = locals->SwathHeightC;
+       CalculateSwathAndDETConfiguration_params->DETBufferSizeInKByte = locals->DETBufferSizeInKByte;
+       CalculateSwathAndDETConfiguration_params->DETBufferSizeY = locals->DETBufferSizeY;
+       CalculateSwathAndDETConfiguration_params->DETBufferSizeC = locals->DETBufferSizeC;
+       CalculateSwathAndDETConfiguration_params->UnboundedRequestEnabled = &locals->UnboundedRequestEnabled;
+       CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_64b = &locals->compbuf_reserved_space_64b;
+       CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_zs = &locals->compbuf_reserved_space_zs;
+       CalculateSwathAndDETConfiguration_params->CompressedBufferSizeInkByte = &locals->CompressedBufferSizeInkByte;
+       CalculateSwathAndDETConfiguration_params->ViewportSizeSupportPerSurface = &s->dummy_boolean_array[0][0];
+       CalculateSwathAndDETConfiguration_params->ViewportSizeSupport = &s->dummy_boolean[0];
+
+       // VBA_DELTA
+       // Calculate DET size, swath height here. In VBA, they are calculated in mode check stage
+       CalculateSwathAndDETConfiguration(&mode_lib->scratch,
+               CalculateSwathAndDETConfiguration_params);
+
+       // DCFCLK Deep Sleep
+       CalculateDCFCLKDeepSleep(
+                       mode_lib->ms.num_active_planes,
+                       locals->BytePerPixelY,
+                       locals->BytePerPixelC,
+                       mode_lib->ms.cache_display_cfg.plane.VRatio,
+                       mode_lib->ms.cache_display_cfg.plane.VRatioChroma,
+                       locals->SwathWidthY,
+                       locals->SwathWidthC,
+                       mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+                       mode_lib->ms.cache_display_cfg.plane.HRatio,
+                       mode_lib->ms.cache_display_cfg.plane.HRatioChroma,
+                       mode_lib->ms.cache_display_cfg.timing.PixelClock,
+                       locals->PSCL_THROUGHPUT,
+                       locals->PSCL_THROUGHPUT_CHROMA,
+                       locals->Dppclk,
+                       locals->ReadBandwidthSurfaceLuma,
+                       locals->ReadBandwidthSurfaceChroma,
+                       mode_lib->ms.soc.return_bus_width_bytes,
+
+                       /* Output */
+                       &locals->DCFCLKDeepSleep);
+
+       // DSCCLK
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if ((mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] != k) || !mode_lib->ms.cache_display_cfg.hw.DSCEnabled[k]) {
+                       locals->DSCCLK_calculated[k] = 0.0;
+               } else {
+                       if (mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_420)
+                               s->DSCFormatFactor = 2;
+                       else if (mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_444)
+                               s->DSCFormatFactor = 1;
+                       else if (mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_n422 || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl)
+                               s->DSCFormatFactor = 2;
+                       else
+                               s->DSCFormatFactor = 1;
+                       if (mode_lib->ms.cache_display_cfg.hw.ODMMode[k] == dml_odm_mode_combine_4to1)
+                               locals->DSCCLK_calculated[k] = mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] / 12 / s->DSCFormatFactor / (1 - mode_lib->ms.soc.dcn_downspread_percent / 100);
+                       else if (mode_lib->ms.cache_display_cfg.hw.ODMMode[k] == dml_odm_mode_combine_2to1)
+                               locals->DSCCLK_calculated[k] = mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] / 6 / s->DSCFormatFactor / (1 - mode_lib->ms.soc.dcn_downspread_percent / 100);
+                       else
+                               locals->DSCCLK_calculated[k] = mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] / 3 / s->DSCFormatFactor / (1 - mode_lib->ms.soc.dcn_downspread_percent / 100);
+               }
+       }
+
+       // DSC Delay
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               locals->DSCDelay[k] = DSCDelayRequirement(mode_lib->ms.cache_display_cfg.hw.DSCEnabled[k],
+                                                                                               mode_lib->ms.cache_display_cfg.hw.ODMMode[k],
+                                                                                               mode_lib->ms.cache_display_cfg.output.DSCInputBitPerComponent[k],
+                                                                                               mode_lib->ms.cache_display_cfg.output.OutputBpp[k],
+                                                                                               mode_lib->ms.cache_display_cfg.timing.HActive[k],
+                                                                                               mode_lib->ms.cache_display_cfg.timing.HTotal[k],
+                                                                                               mode_lib->ms.cache_display_cfg.hw.NumberOfDSCSlices[k],
+                                                                                               mode_lib->ms.cache_display_cfg.output.OutputFormat[k],
+                                                                                               mode_lib->ms.cache_display_cfg.output.OutputEncoder[k],
+                                                                                               mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+                                                                                               mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k]);
+       }
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k)
+               for (j = 0; j < mode_lib->ms.num_active_planes; ++j) // NumberOfSurfaces
+                       if (j != k && mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == j && mode_lib->ms.cache_display_cfg.hw.DSCEnabled[j])
+                               locals->DSCDelay[k] = locals->DSCDelay[j];
+
+       // Prefetch
+       CalculateSurfaceSizeInMall(
+               mode_lib->ms.num_active_planes,
+               mode_lib->ms.soc.mall_allocated_for_dcn_mbytes,
+               mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen,
+               mode_lib->ms.cache_display_cfg.surface.DCCEnable,
+               mode_lib->ms.cache_display_cfg.plane.ViewportStationary,
+               mode_lib->ms.cache_display_cfg.plane.ViewportXStart,
+               mode_lib->ms.cache_display_cfg.plane.ViewportYStart,
+               mode_lib->ms.cache_display_cfg.plane.ViewportXStartC,
+               mode_lib->ms.cache_display_cfg.plane.ViewportYStartC,
+               mode_lib->ms.cache_display_cfg.plane.ViewportWidth,
+               mode_lib->ms.cache_display_cfg.plane.ViewportHeight,
+               locals->BytePerPixelY,
+               mode_lib->ms.cache_display_cfg.plane.ViewportWidthChroma,
+               mode_lib->ms.cache_display_cfg.plane.ViewportHeightChroma,
+               locals->BytePerPixelC,
+               mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY,
+               mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC,
+               mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY,
+               mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC,
+               locals->BlockWidth256BytesY,
+               locals->BlockWidth256BytesC,
+               locals->BlockHeight256BytesY,
+               locals->BlockHeight256BytesC,
+               locals->BlockWidthY,
+               locals->BlockWidthC,
+               locals->BlockHeightY,
+               locals->BlockHeightC,
+
+               /* Output */
+               locals->SurfaceSizeInTheMALL,
+               &s->dummy_boolean[0]); /* dml_bool_t *ExceededMALLSize */
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               s->SurfaceParameters[k].PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock[k];
+               s->SurfaceParameters[k].DPPPerSurface = mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k];
+               s->SurfaceParameters[k].SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan[k];
+               s->SurfaceParameters[k].ViewportHeight = mode_lib->ms.cache_display_cfg.plane.ViewportHeight[k];
+               s->SurfaceParameters[k].ViewportHeightChroma = mode_lib->ms.cache_display_cfg.plane.ViewportHeightChroma[k];
+               s->SurfaceParameters[k].BlockWidth256BytesY = locals->BlockWidth256BytesY[k];
+               s->SurfaceParameters[k].BlockHeight256BytesY = locals->BlockHeight256BytesY[k];
+               s->SurfaceParameters[k].BlockWidth256BytesC = locals->BlockWidth256BytesC[k];
+               s->SurfaceParameters[k].BlockHeight256BytesC = locals->BlockHeight256BytesC[k];
+               s->SurfaceParameters[k].BlockWidthY = locals->BlockWidthY[k];
+               s->SurfaceParameters[k].BlockHeightY = locals->BlockHeightY[k];
+               s->SurfaceParameters[k].BlockWidthC = locals->BlockWidthC[k];
+               s->SurfaceParameters[k].BlockHeightC = locals->BlockHeightC[k];
+               s->SurfaceParameters[k].InterlaceEnable = mode_lib->ms.cache_display_cfg.timing.Interlace[k];
+               s->SurfaceParameters[k].HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal[k];
+               s->SurfaceParameters[k].DCCEnable = mode_lib->ms.cache_display_cfg.surface.DCCEnable[k];
+               s->SurfaceParameters[k].SourcePixelFormat = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k];
+               s->SurfaceParameters[k].SurfaceTiling = mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k];
+               s->SurfaceParameters[k].BytePerPixelY = locals->BytePerPixelY[k];
+               s->SurfaceParameters[k].BytePerPixelC = locals->BytePerPixelC[k];
+               s->SurfaceParameters[k].ProgressiveToInterlaceUnitInOPP = mode_lib->ms.ip.ptoi_supported;
+               s->SurfaceParameters[k].VRatio = mode_lib->ms.cache_display_cfg.plane.VRatio[k];
+               s->SurfaceParameters[k].VRatioChroma = mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k];
+               s->SurfaceParameters[k].VTaps = mode_lib->ms.cache_display_cfg.plane.VTaps[k];
+               s->SurfaceParameters[k].VTapsChroma = mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k];
+               s->SurfaceParameters[k].PitchY = mode_lib->ms.cache_display_cfg.surface.PitchY[k];
+               s->SurfaceParameters[k].DCCMetaPitchY = mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchY[k];
+               s->SurfaceParameters[k].PitchC = mode_lib->ms.cache_display_cfg.surface.PitchC[k];
+               s->SurfaceParameters[k].DCCMetaPitchC = mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchC[k];
+               s->SurfaceParameters[k].ViewportStationary = mode_lib->ms.cache_display_cfg.plane.ViewportStationary[k];
+               s->SurfaceParameters[k].ViewportXStart = mode_lib->ms.cache_display_cfg.plane.ViewportXStart[k];
+               s->SurfaceParameters[k].ViewportYStart = mode_lib->ms.cache_display_cfg.plane.ViewportYStart[k];
+               s->SurfaceParameters[k].ViewportXStartC = mode_lib->ms.cache_display_cfg.plane.ViewportXStartC[k];
+               s->SurfaceParameters[k].ViewportYStartC = mode_lib->ms.cache_display_cfg.plane.ViewportYStartC[k];
+               s->SurfaceParameters[k].FORCE_ONE_ROW_FOR_FRAME = mode_lib->ms.cache_display_cfg.plane.ForceOneRowForFrame[k];
+               s->SurfaceParameters[k].SwathHeightY = locals->SwathHeightY[k];
+               s->SurfaceParameters[k].SwathHeightC = locals->SwathHeightC[k];
+       }
+
+       CalculateVMRowAndSwath_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+       CalculateVMRowAndSwath_params->myPipe = s->SurfaceParameters;
+       CalculateVMRowAndSwath_params->SurfaceSizeInMALL = locals->SurfaceSizeInTheMALL;
+       CalculateVMRowAndSwath_params->PTEBufferSizeInRequestsLuma = mode_lib->ms.ip.dpte_buffer_size_in_pte_reqs_luma;
+       CalculateVMRowAndSwath_params->PTEBufferSizeInRequestsChroma = mode_lib->ms.ip.dpte_buffer_size_in_pte_reqs_chroma;
+       CalculateVMRowAndSwath_params->DCCMetaBufferSizeBytes = mode_lib->ms.ip.dcc_meta_buffer_size_bytes;
+       CalculateVMRowAndSwath_params->UseMALLForStaticScreen = mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen;
+       CalculateVMRowAndSwath_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+       CalculateVMRowAndSwath_params->MALLAllocatedForDCN = mode_lib->ms.soc.mall_allocated_for_dcn_mbytes;
+       CalculateVMRowAndSwath_params->SwathWidthY = locals->SwathWidthY;
+       CalculateVMRowAndSwath_params->SwathWidthC = locals->SwathWidthC;
+       CalculateVMRowAndSwath_params->GPUVMEnable = mode_lib->ms.cache_display_cfg.plane.GPUVMEnable;
+       CalculateVMRowAndSwath_params->HostVMEnable = mode_lib->ms.cache_display_cfg.plane.HostVMEnable;
+       CalculateVMRowAndSwath_params->HostVMMaxNonCachedPageTableLevels = mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels;
+       CalculateVMRowAndSwath_params->GPUVMMaxPageTableLevels = mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels;
+       CalculateVMRowAndSwath_params->GPUVMMinPageSizeKBytes = mode_lib->ms.cache_display_cfg.plane.GPUVMMinPageSizeKBytes;
+       CalculateVMRowAndSwath_params->HostVMMinPageSize = mode_lib->ms.soc.hostvm_min_page_size_kbytes;
+       CalculateVMRowAndSwath_params->PTEBufferModeOverrideEn = mode_lib->ms.cache_display_cfg.plane.PTEBufferModeOverrideEn;
+       CalculateVMRowAndSwath_params->PTEBufferModeOverrideVal = mode_lib->ms.cache_display_cfg.plane.PTEBufferMode;
+       CalculateVMRowAndSwath_params->PTEBufferSizeNotExceeded = s->dummy_boolean_array[0];
+       CalculateVMRowAndSwath_params->DCCMetaBufferSizeNotExceeded = s->dummy_boolean_array[1];
+       CalculateVMRowAndSwath_params->dpte_row_width_luma_ub = locals->dpte_row_width_luma_ub;
+       CalculateVMRowAndSwath_params->dpte_row_width_chroma_ub = locals->dpte_row_width_chroma_ub;
+       CalculateVMRowAndSwath_params->dpte_row_height_luma = locals->dpte_row_height;
+       CalculateVMRowAndSwath_params->dpte_row_height_chroma = locals->dpte_row_height_chroma;
+       CalculateVMRowAndSwath_params->dpte_row_height_linear_luma = locals->dpte_row_height_linear;
+       CalculateVMRowAndSwath_params->dpte_row_height_linear_chroma = locals->dpte_row_height_linear_chroma;
+       CalculateVMRowAndSwath_params->meta_req_width = locals->meta_req_width;
+       CalculateVMRowAndSwath_params->meta_req_width_chroma = locals->meta_req_width_chroma;
+       CalculateVMRowAndSwath_params->meta_req_height = locals->meta_req_height;
+       CalculateVMRowAndSwath_params->meta_req_height_chroma = locals->meta_req_height_chroma;
+       CalculateVMRowAndSwath_params->meta_row_width = locals->meta_row_width;
+       CalculateVMRowAndSwath_params->meta_row_width_chroma = locals->meta_row_width_chroma;
+       CalculateVMRowAndSwath_params->meta_row_height = locals->meta_row_height;
+       CalculateVMRowAndSwath_params->meta_row_height_chroma = locals->meta_row_height_chroma;
+       CalculateVMRowAndSwath_params->vm_group_bytes = locals->vm_group_bytes;
+       CalculateVMRowAndSwath_params->dpte_group_bytes = locals->dpte_group_bytes;
+       CalculateVMRowAndSwath_params->PixelPTEReqWidthY = locals->PixelPTEReqWidthY;
+       CalculateVMRowAndSwath_params->PixelPTEReqHeightY = locals->PixelPTEReqHeightY;
+       CalculateVMRowAndSwath_params->PTERequestSizeY = locals->PTERequestSizeY;
+       CalculateVMRowAndSwath_params->PixelPTEReqWidthC = locals->PixelPTEReqWidthC;
+       CalculateVMRowAndSwath_params->PixelPTEReqHeightC = locals->PixelPTEReqHeightC;
+       CalculateVMRowAndSwath_params->PTERequestSizeC = locals->PTERequestSizeC;
+       CalculateVMRowAndSwath_params->dpde0_bytes_per_frame_ub_l = locals->dpde0_bytes_per_frame_ub_l;
+       CalculateVMRowAndSwath_params->meta_pte_bytes_per_frame_ub_l = locals->meta_pte_bytes_per_frame_ub_l;
+       CalculateVMRowAndSwath_params->dpde0_bytes_per_frame_ub_c = locals->dpde0_bytes_per_frame_ub_c;
+       CalculateVMRowAndSwath_params->meta_pte_bytes_per_frame_ub_c = locals->meta_pte_bytes_per_frame_ub_c;
+       CalculateVMRowAndSwath_params->PrefetchSourceLinesY = locals->PrefetchSourceLinesY;
+       CalculateVMRowAndSwath_params->PrefetchSourceLinesC = locals->PrefetchSourceLinesC;
+       CalculateVMRowAndSwath_params->VInitPreFillY = locals->VInitPreFillY;
+       CalculateVMRowAndSwath_params->VInitPreFillC = locals->VInitPreFillC;
+       CalculateVMRowAndSwath_params->MaxNumSwathY = locals->MaxNumSwathY;
+       CalculateVMRowAndSwath_params->MaxNumSwathC = locals->MaxNumSwathC;
+       CalculateVMRowAndSwath_params->meta_row_bw = locals->meta_row_bw;
+       CalculateVMRowAndSwath_params->dpte_row_bw = locals->dpte_row_bw;
+       CalculateVMRowAndSwath_params->PixelPTEBytesPerRow = locals->PixelPTEBytesPerRow;
+       CalculateVMRowAndSwath_params->PDEAndMetaPTEBytesFrame = locals->PDEAndMetaPTEBytesFrame;
+       CalculateVMRowAndSwath_params->MetaRowByte = locals->MetaRowByte;
+       CalculateVMRowAndSwath_params->use_one_row_for_frame = locals->use_one_row_for_frame;
+       CalculateVMRowAndSwath_params->use_one_row_for_frame_flip = locals->use_one_row_for_frame_flip;
+       CalculateVMRowAndSwath_params->UsesMALLForStaticScreen = locals->UsesMALLForStaticScreen;
+       CalculateVMRowAndSwath_params->PTE_BUFFER_MODE = locals->PTE_BUFFER_MODE;
+       CalculateVMRowAndSwath_params->BIGK_FRAGMENT_SIZE = locals->BIGK_FRAGMENT_SIZE;
+
+       CalculateVMRowAndSwath(&mode_lib->scratch,
+       CalculateVMRowAndSwath_params);
+
+       s->ReorderBytes = (dml_uint_t)(mode_lib->ms.soc.num_chans * dml_max3(
+                       mode_lib->ms.soc.urgent_out_of_order_return_per_channel_pixel_only_bytes,
+                       mode_lib->ms.soc.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes,
+                       mode_lib->ms.soc.urgent_out_of_order_return_per_channel_vm_only_bytes));
+
+       s->VMDataOnlyReturnBW = dml_get_return_bw_mbps_vm_only(&mode_lib->ms.soc,
+                                                                                                                                       mode_lib->ms.state.use_ideal_dram_bw_strobe,
+                                                                                                                                       mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+                                                                                                                                       locals->Dcfclk,
+                                                                                                                                       mode_lib->ms.FabricClock,
+                                                                                                                                       mode_lib->ms.DRAMSpeed);
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: locals->Dcfclk = %f\n", __func__, locals->Dcfclk);
+       dml_print("DML::%s: mode_lib->ms.soc.return_bus_width_bytes = %u\n", __func__, mode_lib->ms.soc.return_bus_width_bytes);
+       dml_print("DML::%s: mode_lib->ms.FabricClock = %f\n", __func__, mode_lib->ms.FabricClock);
+       dml_print("DML::%s: mode_lib->ms.soc.fabric_datapath_to_dcn_data_return_bytes = %u\n", __func__, mode_lib->ms.soc.fabric_datapath_to_dcn_data_return_bytes);
+       dml_print("DML::%s: mode_lib->ms.soc.pct_ideal_sdp_bw_after_urgent = %f\n", __func__, mode_lib->ms.soc.pct_ideal_sdp_bw_after_urgent);
+       dml_print("DML::%s: mode_lib->ms.DRAMSpeed = %f\n", __func__, mode_lib->ms.DRAMSpeed);
+       dml_print("DML::%s: mode_lib->ms.soc.num_chans = %u\n", __func__, mode_lib->ms.soc.num_chans);
+       dml_print("DML::%s: mode_lib->ms.soc.dram_channel_width_bytes = %u\n", __func__, mode_lib->ms.soc.dram_channel_width_bytes);
+       dml_print("DML::%s: mode_lib->ms.state_idx = %u\n", __func__, mode_lib->ms.state_idx);
+       dml_print("DML::%s: mode_lib->ms.max_state_idx = %u\n", __func__, mode_lib->ms.max_state_idx);
+       dml_print("DML::%s: mode_lib->ms.state.use_ideal_dram_bw_strobe = %u\n", __func__, mode_lib->ms.state.use_ideal_dram_bw_strobe);
+       dml_print("DML::%s: VMDataOnlyReturnBW = %f\n", __func__, s->VMDataOnlyReturnBW);
+       dml_print("DML::%s: ReturnBW = %f\n",  __func__, mode_lib->ms.ReturnBW);
+#endif
+
+       s->HostVMInefficiencyFactor = 1.0;
+       if (mode_lib->ms.cache_display_cfg.plane.GPUVMEnable && mode_lib->ms.cache_display_cfg.plane.HostVMEnable)
+               s->HostVMInefficiencyFactor = mode_lib->ms.ReturnBW / s->VMDataOnlyReturnBW;
+
+       s->TotalDCCActiveDPP = 0;
+       s->TotalActiveDPP = 0;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               s->TotalActiveDPP = s->TotalActiveDPP + mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k];
+               if (mode_lib->ms.cache_display_cfg.surface.DCCEnable[k])
+                       s->TotalDCCActiveDPP = s->TotalDCCActiveDPP + mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k];
+       }
+
+       locals->UrgentExtraLatency = CalculateExtraLatency(
+                       mode_lib->ms.soc.round_trip_ping_latency_dcfclk_cycles,
+                       s->ReorderBytes,
+                       locals->Dcfclk,
+                       s->TotalActiveDPP,
+                       mode_lib->ms.ip.pixel_chunk_size_kbytes,
+                       s->TotalDCCActiveDPP,
+                       mode_lib->ms.ip.meta_chunk_size_kbytes,
+                       mode_lib->ms.ReturnBW,
+                       mode_lib->ms.cache_display_cfg.plane.GPUVMEnable,
+                       mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+                       mode_lib->ms.num_active_planes,
+                       mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+                       locals->dpte_group_bytes,
+                       s->HostVMInefficiencyFactor,
+                       mode_lib->ms.soc.hostvm_min_page_size_kbytes,
+                       mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels);
+
+       locals->TCalc = 24.0 / locals->DCFCLKDeepSleep;
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k) {
+                       if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true) {
+                               locals->WritebackDelay[k] =
+                                               mode_lib->ms.state.writeback_latency_us
+                                                               + CalculateWriteBackDelay(
+                                                                               mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[k],
+                                                                               mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[k],
+                                                                               mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[k],
+                                                                               mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[k],
+                                                                               mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k],
+                                                                               mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[k],
+                                                                               mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[k],
+                                                                               mode_lib->ms.cache_display_cfg.timing.HTotal[k]) / locals->Dispclk;
+                       } else
+                               locals->WritebackDelay[k] = 0;
+                       for (j = 0; j < mode_lib->ms.num_active_planes; ++j) {
+                               if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[j] == k
+                                               && mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[j] == true) {
+                                       locals->WritebackDelay[k] =
+                                                       dml_max(
+                                                                       locals->WritebackDelay[k],
+                                                                       mode_lib->ms.state.writeback_latency_us
+                                                                                       + CalculateWriteBackDelay(
+                                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[j],
+                                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[j],
+                                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[j],
+                                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[j],
+                                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[j],
+                                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[j],
+                                                                                                       mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[j],
+                                                                                                       mode_lib->ms.cache_display_cfg.timing.HTotal[k]) / locals->Dispclk);
+                               }
+                       }
+               }
+       }
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k)
+               for (j = 0; j < mode_lib->ms.num_active_planes; ++j)
+                       if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == j)
+                               locals->WritebackDelay[k] = locals->WritebackDelay[j];
+
+       locals->UrgentLatency = CalculateUrgentLatency(mode_lib->ms.state.urgent_latency_pixel_data_only_us,
+                                                                                               mode_lib->ms.state.urgent_latency_pixel_mixed_with_vm_data_us,
+                                                                                               mode_lib->ms.state.urgent_latency_vm_data_only_us,
+                                                                                               mode_lib->ms.soc.do_urgent_latency_adjustment,
+                                                                                               mode_lib->ms.state.urgent_latency_adjustment_fabric_clock_component_us,
+                                                                                               mode_lib->ms.state.urgent_latency_adjustment_fabric_clock_reference_mhz,
+                                                                                               mode_lib->ms.FabricClock);
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               CalculateUrgentBurstFactor(mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k],
+                                                                       locals->swath_width_luma_ub[k],
+                                                                       locals->swath_width_chroma_ub[k],
+                                                                       locals->SwathHeightY[k],
+                                                                       locals->SwathHeightC[k],
+                                                                       mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+                                                                       locals->UrgentLatency,
+                                                                       mode_lib->ms.ip.cursor_buffer_size,
+                                                                       mode_lib->ms.cache_display_cfg.plane.CursorWidth[k],
+                                                                       mode_lib->ms.cache_display_cfg.plane.CursorBPP[k],
+                                                                       mode_lib->ms.cache_display_cfg.plane.VRatio[k],
+                                                                       mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k],
+                                                                       locals->BytePerPixelDETY[k],
+                                                                       locals->BytePerPixelDETC[k],
+                                                                       locals->DETBufferSizeY[k],
+                                                                       locals->DETBufferSizeC[k],
+
+                                                                       /* output */
+                                                                       &locals->UrgBurstFactorCursor[k],
+                                                                       &locals->UrgBurstFactorLuma[k],
+                                                                       &locals->UrgBurstFactorChroma[k],
+                                                                       &locals->NoUrgentLatencyHiding[k]);
+
+               locals->cursor_bw[k] = mode_lib->ms.cache_display_cfg.plane.NumberOfCursors[k] * mode_lib->ms.cache_display_cfg.plane.CursorWidth[k] * mode_lib->ms.cache_display_cfg.plane.CursorBPP[k] / 8.0 /
+                                                               ((dml_float_t) mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * mode_lib->ms.cache_display_cfg.plane.VRatio[k];
+       }
+
+       s->VStartupLines = __DML_VBA_MIN_VSTARTUP__;
+       s->MaxVStartupAllPlanes = 0;
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               s->MaxVStartupLines[k] = CalculateMaxVStartup(k,
+                                                                                                       mode_lib->ms.ip.ptoi_supported,
+                                                                                                       mode_lib->ms.ip.vblank_nom_default_us,
+                                                                                                       &mode_lib->ms.cache_display_cfg.timing,
+                                                                                                       locals->WritebackDelay[k]);
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u MaxVStartupLines   = %u\n", __func__, k, s->MaxVStartupLines[k]);
+               dml_print("DML::%s: k=%u WritebackDelay     = %f\n", __func__, k, locals->WritebackDelay[k]);
+#endif
+       }
+
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k)
+               s->MaxVStartupAllPlanes = (dml_uint_t)(dml_max(s->MaxVStartupAllPlanes, s->MaxVStartupLines[k]));
+
+       s->ImmediateFlipRequirementFinal = false;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               s->ImmediateFlipRequirementFinal = s->ImmediateFlipRequirementFinal || (mode_lib->ms.policy.ImmediateFlipRequirement[k] == dml_immediate_flip_required);
+       }
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: ImmediateFlipRequirementFinal = %u\n", __func__, s->ImmediateFlipRequirementFinal);
+#endif
+
+       // The prefetch scheduling should only be calculated once as per AllowForPStateChangeOrStutterInVBlank requirement
+       // If the AllowForPStateChangeOrStutterInVBlank requirement is not strict (i.e. only try those power saving feature
+       // if possible, then will try to program for the best power saving features in order of diffculty (dram, fclk, stutter)
+       s->iteration = 0;
+       s->MaxTotalRDBandwidth = 0;
+       s->AllPrefetchModeTested = false;
+       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+               CalculatePrefetchMode(mode_lib->ms.policy.AllowForPStateChangeOrStutterInVBlank[k], &s->MinPrefetchMode[k], &s->MaxPrefetchMode[k]);
+               s->NextPrefetchMode[k] = s->MinPrefetchMode[k];
+       }
+
+       do {
+               s->MaxTotalRDBandwidthNoUrgentBurst = 0.0;
+               s->DestinationLineTimesForPrefetchLessThan2 = false;
+               s->VRatioPrefetchMoreThanMax = false;
+
+               dml_print("DML::%s: Start one iteration: VStartupLines = %u\n", __func__, s->VStartupLines);
+
+               s->AllPrefetchModeTested = true;
+               s->MaxTotalRDBandwidth = 0;
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       locals->PrefetchMode[k] = s->NextPrefetchMode[k];
+                       TWait = CalculateTWait(
+                                       locals->PrefetchMode[k],
+                                       mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k],
+                                       mode_lib->ms.policy.SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+                                       mode_lib->ms.cache_display_cfg.timing.DRRDisplay[k],
+                                       mode_lib->ms.state.dram_clock_change_latency_us,
+                                       mode_lib->ms.state.fclk_change_latency_us,
+                                       locals->UrgentLatency,
+                                       mode_lib->ms.state.sr_enter_plus_exit_time_us);
+
+                       myPipe = &s->myPipe;
+                       myPipe->Dppclk = locals->Dppclk[k];
+                       myPipe->Dispclk = locals->Dispclk;
+                       myPipe->PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock[k];
+                       myPipe->DCFClkDeepSleep = locals->DCFCLKDeepSleep;
+                       myPipe->DPPPerSurface = mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k];
+                       myPipe->ScalerEnabled = mode_lib->ms.cache_display_cfg.plane.ScalerEnabled[k];
+                       myPipe->SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan[k];
+                       myPipe->BlockWidth256BytesY = locals->BlockWidth256BytesY[k];
+                       myPipe->BlockHeight256BytesY = locals->BlockHeight256BytesY[k];
+                       myPipe->BlockWidth256BytesC = locals->BlockWidth256BytesC[k];
+                       myPipe->BlockHeight256BytesC = locals->BlockHeight256BytesC[k];
+                       myPipe->InterlaceEnable = mode_lib->ms.cache_display_cfg.timing.Interlace[k];
+                       myPipe->NumberOfCursors = mode_lib->ms.cache_display_cfg.plane.NumberOfCursors[k];
+                       myPipe->VBlank = mode_lib->ms.cache_display_cfg.timing.VTotal[k] - mode_lib->ms.cache_display_cfg.timing.VActive[k];
+                       myPipe->HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal[k];
+                       myPipe->HActive = mode_lib->ms.cache_display_cfg.timing.HActive[k];
+                       myPipe->DCCEnable = mode_lib->ms.cache_display_cfg.surface.DCCEnable[k];
+                       myPipe->ODMMode = mode_lib->ms.cache_display_cfg.hw.ODMMode[k];
+                       myPipe->SourcePixelFormat = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k];
+                       myPipe->BytePerPixelY = locals->BytePerPixelY[k];
+                       myPipe->BytePerPixelC = locals->BytePerPixelC[k];
+                       myPipe->ProgressiveToInterlaceUnitInOPP = mode_lib->ms.ip.ptoi_supported;
+
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: Calling CalculatePrefetchSchedule for k=%u\n", __func__, k);
+                       dml_print("DML::%s: AllowForPStateChangeOrStutterInVBlank = %u\n", __func__, mode_lib->ms.policy.AllowForPStateChangeOrStutterInVBlank[k]);
+                       dml_print("DML::%s: PrefetchMode[k] = %u (Min=%u Max=%u)\n", __func__, locals->PrefetchMode[k], s->MinPrefetchMode[k], s->MaxPrefetchMode[k]);
+#endif
+
+                       CalculatePrefetchSchedule_params->EnhancedPrefetchScheduleAccelerationFinal = mode_lib->ms.policy.EnhancedPrefetchScheduleAccelerationFinal;
+                       CalculatePrefetchSchedule_params->HostVMInefficiencyFactor = s->HostVMInefficiencyFactor;
+                       CalculatePrefetchSchedule_params->myPipe = myPipe;
+                       CalculatePrefetchSchedule_params->DSCDelay = locals->DSCDelay[k];
+                       CalculatePrefetchSchedule_params->DPPCLKDelaySubtotalPlusCNVCFormater = mode_lib->ms.ip.dppclk_delay_subtotal + mode_lib->ms.ip.dppclk_delay_cnvc_formatter;
+                       CalculatePrefetchSchedule_params->DPPCLKDelaySCL = mode_lib->ms.ip.dppclk_delay_scl;
+                       CalculatePrefetchSchedule_params->DPPCLKDelaySCLLBOnly = mode_lib->ms.ip.dppclk_delay_scl_lb_only;
+                       CalculatePrefetchSchedule_params->DPPCLKDelayCNVCCursor = mode_lib->ms.ip.dppclk_delay_cnvc_cursor;
+                       CalculatePrefetchSchedule_params->DISPCLKDelaySubtotal = mode_lib->ms.ip.dispclk_delay_subtotal;
+                       CalculatePrefetchSchedule_params->DPP_RECOUT_WIDTH = (dml_uint_t)(locals->SwathWidthY[k] / mode_lib->ms.cache_display_cfg.plane.HRatio[k]);
+                       CalculatePrefetchSchedule_params->OutputFormat = mode_lib->ms.cache_display_cfg.output.OutputFormat[k];
+                       CalculatePrefetchSchedule_params->MaxInterDCNTileRepeaters = mode_lib->ms.ip.max_inter_dcn_tile_repeaters;
+                       CalculatePrefetchSchedule_params->VStartup = (dml_uint_t)(dml_min(s->VStartupLines, s->MaxVStartupLines[k]));
+                       CalculatePrefetchSchedule_params->MaxVStartup = s->MaxVStartupLines[k];
+                       CalculatePrefetchSchedule_params->GPUVMPageTableLevels = mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels;
+                       CalculatePrefetchSchedule_params->GPUVMEnable = mode_lib->ms.cache_display_cfg.plane.GPUVMEnable;
+                       CalculatePrefetchSchedule_params->HostVMEnable = mode_lib->ms.cache_display_cfg.plane.HostVMEnable;
+                       CalculatePrefetchSchedule_params->HostVMMaxNonCachedPageTableLevels = mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels;
+                       CalculatePrefetchSchedule_params->HostVMMinPageSize = mode_lib->ms.soc.hostvm_min_page_size_kbytes;
+                       CalculatePrefetchSchedule_params->DynamicMetadataEnable = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataEnable[k];
+                       CalculatePrefetchSchedule_params->DynamicMetadataVMEnabled = mode_lib->ms.ip.dynamic_metadata_vm_enabled;
+                       CalculatePrefetchSchedule_params->DynamicMetadataLinesBeforeActiveRequired = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataLinesBeforeActiveRequired[k];
+                       CalculatePrefetchSchedule_params->DynamicMetadataTransmittedBytes = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataTransmittedBytes[k];
+                       CalculatePrefetchSchedule_params->UrgentLatency = locals->UrgentLatency;
+                       CalculatePrefetchSchedule_params->UrgentExtraLatency = locals->UrgentExtraLatency;
+                       CalculatePrefetchSchedule_params->TCalc = locals->TCalc;
+                       CalculatePrefetchSchedule_params->PDEAndMetaPTEBytesFrame = locals->PDEAndMetaPTEBytesFrame[k];
+                       CalculatePrefetchSchedule_params->MetaRowByte = locals->MetaRowByte[k];
+                       CalculatePrefetchSchedule_params->PixelPTEBytesPerRow = locals->PixelPTEBytesPerRow[k];
+                       CalculatePrefetchSchedule_params->PrefetchSourceLinesY = locals->PrefetchSourceLinesY[k];
+                       CalculatePrefetchSchedule_params->VInitPreFillY = locals->VInitPreFillY[k];
+                       CalculatePrefetchSchedule_params->MaxNumSwathY = locals->MaxNumSwathY[k];
+                       CalculatePrefetchSchedule_params->PrefetchSourceLinesC = locals->PrefetchSourceLinesC[k];
+                       CalculatePrefetchSchedule_params->VInitPreFillC = locals->VInitPreFillC[k];
+                       CalculatePrefetchSchedule_params->MaxNumSwathC = locals->MaxNumSwathC[k];
+                       CalculatePrefetchSchedule_params->swath_width_luma_ub = locals->swath_width_luma_ub[k];
+                       CalculatePrefetchSchedule_params->swath_width_chroma_ub = locals->swath_width_chroma_ub[k];
+                       CalculatePrefetchSchedule_params->SwathHeightY = locals->SwathHeightY[k];
+                       CalculatePrefetchSchedule_params->SwathHeightC = locals->SwathHeightC[k];
+                       CalculatePrefetchSchedule_params->TWait = TWait;
+                       CalculatePrefetchSchedule_params->DSTXAfterScaler = &locals->DSTXAfterScaler[k];
+                       CalculatePrefetchSchedule_params->DSTYAfterScaler = &locals->DSTYAfterScaler[k];
+                       CalculatePrefetchSchedule_params->DestinationLinesForPrefetch = &locals->DestinationLinesForPrefetch[k];
+                       CalculatePrefetchSchedule_params->DestinationLinesToRequestVMInVBlank = &locals->DestinationLinesToRequestVMInVBlank[k];
+                       CalculatePrefetchSchedule_params->DestinationLinesToRequestRowInVBlank = &locals->DestinationLinesToRequestRowInVBlank[k];
+                       CalculatePrefetchSchedule_params->VRatioPrefetchY = &locals->VRatioPrefetchY[k];
+                       CalculatePrefetchSchedule_params->VRatioPrefetchC = &locals->VRatioPrefetchC[k];
+                       CalculatePrefetchSchedule_params->RequiredPrefetchPixDataBWLuma = &locals->RequiredPrefetchPixDataBWLuma[k];
+                       CalculatePrefetchSchedule_params->RequiredPrefetchPixDataBWChroma = &locals->RequiredPrefetchPixDataBWChroma[k];
+                       CalculatePrefetchSchedule_params->NotEnoughTimeForDynamicMetadata = &locals->NotEnoughTimeForDynamicMetadata[k];
+                       CalculatePrefetchSchedule_params->Tno_bw = &locals->Tno_bw[k];
+                       CalculatePrefetchSchedule_params->prefetch_vmrow_bw = &locals->prefetch_vmrow_bw[k];
+                       CalculatePrefetchSchedule_params->Tdmdl_vm = &locals->Tdmdl_vm[k];
+                       CalculatePrefetchSchedule_params->Tdmdl = &locals->Tdmdl[k];
+                       CalculatePrefetchSchedule_params->TSetup = &locals->TSetup[k];
+                       CalculatePrefetchSchedule_params->VUpdateOffsetPix = &locals->VUpdateOffsetPix[k];
+                       CalculatePrefetchSchedule_params->VUpdateWidthPix = &locals->VUpdateWidthPix[k];
+                       CalculatePrefetchSchedule_params->VReadyOffsetPix = &locals->VReadyOffsetPix[k];
+
+                       locals->NoTimeToPrefetch[k] =
+                               CalculatePrefetchSchedule(&mode_lib->scratch,
+                                       CalculatePrefetchSchedule_params);
+
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%0u NoTimeToPrefetch=%0d\n", __func__, k, locals->NoTimeToPrefetch[k]);
+#endif
+                       locals->VStartup[k] = (dml_uint_t)(dml_min(s->VStartupLines, s->MaxVStartupLines[k]));
+                       locals->VStartupMin[k] = locals->VStartup[k];
+               }
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       CalculateUrgentBurstFactor(
+                               mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k],
+                               locals->swath_width_luma_ub[k],
+                               locals->swath_width_chroma_ub[k],
+                               locals->SwathHeightY[k],
+                               locals->SwathHeightC[k],
+                               mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+                               locals->UrgentLatency,
+                               mode_lib->ms.ip.cursor_buffer_size,
+                               mode_lib->ms.cache_display_cfg.plane.CursorWidth[k],
+                               mode_lib->ms.cache_display_cfg.plane.CursorBPP[k],
+                               locals->VRatioPrefetchY[k],
+                               locals->VRatioPrefetchC[k],
+                               locals->BytePerPixelDETY[k],
+                               locals->BytePerPixelDETC[k],
+                               locals->DETBufferSizeY[k],
+                               locals->DETBufferSizeC[k],
+                               /* Output */
+                               &locals->UrgBurstFactorCursorPre[k],
+                               &locals->UrgBurstFactorLumaPre[k],
+                               &locals->UrgBurstFactorChromaPre[k],
+                               &locals->NoUrgentLatencyHidingPre[k]);
+
+                       locals->cursor_bw_pre[k] = mode_lib->ms.cache_display_cfg.plane.NumberOfCursors[k] * mode_lib->ms.cache_display_cfg.plane.CursorWidth[k] * mode_lib->ms.cache_display_cfg.plane.CursorBPP[k] / 8.0 / (mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * locals->VRatioPrefetchY[k];
+
+#ifdef __DML_VBA_DEBUG__
+                       dml_print("DML::%s: k=%0u DPPPerSurface=%u\n", __func__, k, mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k]);
+                       dml_print("DML::%s: k=%0u UrgBurstFactorLuma=%f\n", __func__, k, locals->UrgBurstFactorLuma[k]);
+                       dml_print("DML::%s: k=%0u UrgBurstFactorChroma=%f\n", __func__, k, locals->UrgBurstFactorChroma[k]);
+                       dml_print("DML::%s: k=%0u UrgBurstFactorLumaPre=%f\n", __func__, k, locals->UrgBurstFactorLumaPre[k]);
+                       dml_print("DML::%s: k=%0u UrgBurstFactorChromaPre=%f\n", __func__, k, locals->UrgBurstFactorChromaPre[k]);
+
+                       dml_print("DML::%s: k=%0u VRatioPrefetchY=%f\n", __func__, k, locals->VRatioPrefetchY[k]);
+                       dml_print("DML::%s: k=%0u VRatioY=%f\n", __func__, k, mode_lib->ms.cache_display_cfg.plane.VRatio[k]);
+
+                       dml_print("DML::%s: k=%0u prefetch_vmrow_bw=%f\n", __func__, k, locals->prefetch_vmrow_bw[k]);
+                       dml_print("DML::%s: k=%0u ReadBandwidthSurfaceLuma=%f\n", __func__, k, locals->ReadBandwidthSurfaceLuma[k]);
+                       dml_print("DML::%s: k=%0u ReadBandwidthSurfaceChroma=%f\n", __func__, k, locals->ReadBandwidthSurfaceChroma[k]);
+                       dml_print("DML::%s: k=%0u cursor_bw=%f\n", __func__, k, locals->cursor_bw[k]);
+                       dml_print("DML::%s: k=%0u meta_row_bw=%f\n", __func__, k, locals->meta_row_bw[k]);
+                       dml_print("DML::%s: k=%0u dpte_row_bw=%f\n", __func__, k, locals->dpte_row_bw[k]);
+                       dml_print("DML::%s: k=%0u RequiredPrefetchPixDataBWLuma=%f\n", __func__, k, locals->RequiredPrefetchPixDataBWLuma[k]);
+                       dml_print("DML::%s: k=%0u RequiredPrefetchPixDataBWChroma=%f\n", __func__, k, locals->RequiredPrefetchPixDataBWChroma[k]);
+                       dml_print("DML::%s: k=%0u cursor_bw_pre=%f\n", __func__, k, locals->cursor_bw_pre[k]);
+                       dml_print("DML::%s: k=%0u MaxTotalRDBandwidthNoUrgentBurst=%f\n", __func__, k, s->MaxTotalRDBandwidthNoUrgentBurst);
+#endif
+                       if (locals->DestinationLinesForPrefetch[k] < 2)
+                               s->DestinationLineTimesForPrefetchLessThan2 = true;
+
+                       if (locals->VRatioPrefetchY[k] > __DML_MAX_VRATIO_PRE_ENHANCE_PREFETCH_ACC__ ||
+                               locals->VRatioPrefetchC[k] > __DML_MAX_VRATIO_PRE_ENHANCE_PREFETCH_ACC__ ||
+                               ((s->VStartupLines < s->MaxVStartupLines[k] || mode_lib->ms.policy.EnhancedPrefetchScheduleAccelerationFinal == 0) &&
+                                       (locals->VRatioPrefetchY[k] > __DML_MAX_VRATIO_PRE__ || locals->VRatioPrefetchC[k] > __DML_MAX_VRATIO_PRE__)))
+                               s->VRatioPrefetchMoreThanMax = true;
+
+                       //dml_bool_t DestinationLinesToRequestVMInVBlankEqualOrMoreThan32 = false;
+                       //dml_bool_t DestinationLinesToRequestRowInVBlankEqualOrMoreThan16 = false;
+                       //if (locals->DestinationLinesToRequestVMInVBlank[k] >= 32) {
+                       //    DestinationLinesToRequestVMInVBlankEqualOrMoreThan32 = true;
+                       //}
+
+                       //if (locals->DestinationLinesToRequestRowInVBlank[k] >= 16) {
+                       //    DestinationLinesToRequestRowInVBlankEqualOrMoreThan16 = true;
+                       //}
+               }
+
+               locals->FractionOfUrgentBandwidth = s->MaxTotalRDBandwidthNoUrgentBurst / mode_lib->ms.ReturnBW;
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: MaxTotalRDBandwidthNoUrgentBurst=%f \n", __func__, s->MaxTotalRDBandwidthNoUrgentBurst);
+               dml_print("DML::%s: ReturnBW=%f \n", __func__, mode_lib->ms.ReturnBW);
+               dml_print("DML::%s: FractionOfUrgentBandwidth=%f \n", __func__, locals->FractionOfUrgentBandwidth);
+#endif
+
+               CalculatePrefetchBandwithSupport(
+                       mode_lib->ms.num_active_planes,
+                       mode_lib->ms.ReturnBW,
+                       mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange,
+                       locals->NoUrgentLatencyHidingPre,
+                       locals->ReadBandwidthSurfaceLuma,
+                       locals->ReadBandwidthSurfaceChroma,
+                       locals->RequiredPrefetchPixDataBWLuma,
+                       locals->RequiredPrefetchPixDataBWChroma,
+                       locals->cursor_bw,
+                       locals->meta_row_bw,
+                       locals->dpte_row_bw,
+                       locals->cursor_bw_pre,
+                       locals->prefetch_vmrow_bw,
+                       mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+                       locals->UrgBurstFactorLuma,
+                       locals->UrgBurstFactorChroma,
+                       locals->UrgBurstFactorCursor,
+                       locals->UrgBurstFactorLumaPre,
+                       locals->UrgBurstFactorChromaPre,
+                       locals->UrgBurstFactorCursorPre,
+
+                       /* output */
+                       &s->MaxTotalRDBandwidth, // dml_float_t *PrefetchBandwidth
+                       &s->MaxTotalRDBandwidthNotIncludingMALLPrefetch, // dml_float_t *PrefetchBandwidthNotIncludingMALLPrefetch
+                       &s->dummy_single[0], // dml_float_t *FractionOfUrgentBandwidth
+                       &locals->PrefetchModeSupported);
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k)
+                       s->dummy_unit_vector[k] = 1.0;
+
+               CalculatePrefetchBandwithSupport(mode_lib->ms.num_active_planes,
+                       mode_lib->ms.ReturnBW,
+                       mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange,
+                       locals->NoUrgentLatencyHidingPre,
+                       locals->ReadBandwidthSurfaceLuma,
+                       locals->ReadBandwidthSurfaceChroma,
+                       locals->RequiredPrefetchPixDataBWLuma,
+                       locals->RequiredPrefetchPixDataBWChroma,
+                       locals->cursor_bw,
+                       locals->meta_row_bw,
+                       locals->dpte_row_bw,
+                       locals->cursor_bw_pre,
+                       locals->prefetch_vmrow_bw,
+                       mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+                       s->dummy_unit_vector,
+                       s->dummy_unit_vector,
+                       s->dummy_unit_vector,
+                       s->dummy_unit_vector,
+                       s->dummy_unit_vector,
+                       s->dummy_unit_vector,
+
+                       /* output */
+                       &s->NonUrgentMaxTotalRDBandwidth, // dml_float_t *PrefetchBandwidth
+                       &s->NonUrgentMaxTotalRDBandwidthNotIncludingMALLPrefetch, // dml_float_t *PrefetchBandwidthNotIncludingMALLPrefetch
+                       &locals->FractionOfUrgentBandwidth,
+                       &s->dummy_boolean[0]); // dml_bool_t *PrefetchBandwidthSupport
+
+               if (s->VRatioPrefetchMoreThanMax != false || s->DestinationLineTimesForPrefetchLessThan2 != false) {
+                       dml_print("DML::%s: VRatioPrefetchMoreThanMax                   = %u\n", __func__, s->VRatioPrefetchMoreThanMax);
+                       dml_print("DML::%s: DestinationLineTimesForPrefetchLessThan2    = %u\n", __func__, s->DestinationLineTimesForPrefetchLessThan2);
+                       locals->PrefetchModeSupported = false;
+               }
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       if (locals->NoTimeToPrefetch[k] == true || locals->NotEnoughTimeForDynamicMetadata[k]) {
+                               dml_print("DML::%s: k=%u, NoTimeToPrefetch = %0d\n", __func__, k, locals->NoTimeToPrefetch[k]);
+                               dml_print("DML::%s: k=%u, NotEnoughTimeForDynamicMetadata=%u\n", __func__, k, locals->NotEnoughTimeForDynamicMetadata[k]);
+                               locals->PrefetchModeSupported = false;
+                       }
+               }
+
+               if (locals->PrefetchModeSupported == true && mode_lib->ms.support.ImmediateFlipSupport == true) {
+                       locals->BandwidthAvailableForImmediateFlip = CalculateBandwidthAvailableForImmediateFlip(
+                                                                                                                                       mode_lib->ms.num_active_planes,
+                                                                                                                                       mode_lib->ms.ReturnBW,
+                                                                                                                                       locals->ReadBandwidthSurfaceLuma,
+                                                                                                                                       locals->ReadBandwidthSurfaceChroma,
+                                                                                                                                       locals->RequiredPrefetchPixDataBWLuma,
+                                                                                                                                       locals->RequiredPrefetchPixDataBWChroma,
+                                                                                                                                       locals->cursor_bw,
+                                                                                                                                       locals->cursor_bw_pre,
+                                                                                                                                       mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+                                                                                                                                       locals->UrgBurstFactorLuma,
+                                                                                                                                       locals->UrgBurstFactorChroma,
+                                                                                                                                       locals->UrgBurstFactorCursor,
+                                                                                                                                       locals->UrgBurstFactorLumaPre,
+                                                                                                                                       locals->UrgBurstFactorChromaPre,
+                                                                                                                                       locals->UrgBurstFactorCursorPre);
+
+                       locals->TotImmediateFlipBytes = 0;
+                       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                               if (mode_lib->ms.policy.ImmediateFlipRequirement[k] != dml_immediate_flip_not_required) {
+                                       locals->TotImmediateFlipBytes = locals->TotImmediateFlipBytes + mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k] * (locals->PDEAndMetaPTEBytesFrame[k] + locals->MetaRowByte[k]);
+                                       if (locals->use_one_row_for_frame_flip[k]) {
+                                               locals->TotImmediateFlipBytes = locals->TotImmediateFlipBytes + mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k] * (2 * locals->PixelPTEBytesPerRow[k]);
+                                       } else {
+                                               locals->TotImmediateFlipBytes = locals->TotImmediateFlipBytes + mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k] * locals->PixelPTEBytesPerRow[k];
+                                       }
+#ifdef __DML_VBA_DEBUG__
+                                       dml_print("DML::%s: k = %u\n", __func__, k);
+                                       dml_print("DML::%s: DPPPerSurface = %u\n", __func__, mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k]);
+                                       dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %u\n", __func__, locals->PDEAndMetaPTEBytesFrame[k]);
+                                       dml_print("DML::%s: MetaRowByte = %u\n", __func__, locals->MetaRowByte[k]);
+                                       dml_print("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, locals->PixelPTEBytesPerRow[k]);
+                                       dml_print("DML::%s: TotImmediateFlipBytes = %u\n", __func__, locals->TotImmediateFlipBytes);
+#endif
+                               }
+                       }
+                       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                               CalculateFlipSchedule(
+                                               s->HostVMInefficiencyFactor,
+                                               locals->UrgentExtraLatency,
+                                               locals->UrgentLatency,
+                                               mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels,
+                                               mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+                                               mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels,
+                                               mode_lib->ms.cache_display_cfg.plane.GPUVMEnable,
+                                               mode_lib->ms.soc.hostvm_min_page_size_kbytes,
+                                               locals->PDEAndMetaPTEBytesFrame[k],
+                                               locals->MetaRowByte[k],
+                                               locals->PixelPTEBytesPerRow[k],
+                                               locals->BandwidthAvailableForImmediateFlip,
+                                               locals->TotImmediateFlipBytes,
+                                               mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k],
+                                               mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+                                               mode_lib->ms.cache_display_cfg.plane.VRatio[k],
+                                               mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k],
+                                               locals->Tno_bw[k],
+                                               mode_lib->ms.cache_display_cfg.surface.DCCEnable[k],
+                                               locals->dpte_row_height[k],
+                                               locals->meta_row_height[k],
+                                               locals->dpte_row_height_chroma[k],
+                                               locals->meta_row_height_chroma[k],
+                                               locals->use_one_row_for_frame_flip[k],
+
+                                               /* Output */
+                                               &locals->DestinationLinesToRequestVMInImmediateFlip[k],
+                                               &locals->DestinationLinesToRequestRowInImmediateFlip[k],
+                                               &locals->final_flip_bw[k],
+                                               &locals->ImmediateFlipSupportedForPipe[k]);
+                       }
+
+                       CalculateImmediateFlipBandwithSupport(mode_lib->ms.num_active_planes,
+                                                                                               mode_lib->ms.ReturnBW,
+                                                                                               mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange,
+                                                                                               mode_lib->ms.policy.ImmediateFlipRequirement,
+                                                                                               locals->final_flip_bw,
+                                                                                               locals->ReadBandwidthSurfaceLuma,
+                                                                                               locals->ReadBandwidthSurfaceChroma,
+                                                                                               locals->RequiredPrefetchPixDataBWLuma,
+                                                                                               locals->RequiredPrefetchPixDataBWChroma,
+                                                                                               locals->cursor_bw,
+                                                                                               locals->meta_row_bw,
+                                                                                               locals->dpte_row_bw,
+                                                                                               locals->cursor_bw_pre,
+                                                                                               locals->prefetch_vmrow_bw,
+                                                                                               mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+                                                                                               locals->UrgBurstFactorLuma,
+                                                                                               locals->UrgBurstFactorChroma,
+                                                                                               locals->UrgBurstFactorCursor,
+                                                                                               locals->UrgBurstFactorLumaPre,
+                                                                                               locals->UrgBurstFactorChromaPre,
+                                                                                               locals->UrgBurstFactorCursorPre,
+
+                                                                                               /* output */
+                                                                                               &locals->total_dcn_read_bw_with_flip, // dml_float_t *TotalBandwidth
+                                                                                               &locals->total_dcn_read_bw_with_flip_not_including_MALL_prefetch, // dml_float_t TotalBandwidthNotIncludingMALLPrefetch
+                                                                                               &s->dummy_single[0], // dml_float_t *FractionOfUrgentBandwidth
+                                                                                               &locals->ImmediateFlipSupported); // dml_bool_t *ImmediateFlipBandwidthSupport
+
+                       CalculateImmediateFlipBandwithSupport(mode_lib->ms.num_active_planes,
+                               mode_lib->ms.ReturnBW,
+                               mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange,
+                               mode_lib->ms.policy.ImmediateFlipRequirement,
+                               locals->final_flip_bw,
+                               locals->ReadBandwidthSurfaceLuma,
+                               locals->ReadBandwidthSurfaceChroma,
+                               locals->RequiredPrefetchPixDataBWLuma,
+                               locals->RequiredPrefetchPixDataBWChroma,
+                               locals->cursor_bw,
+                               locals->meta_row_bw,
+                               locals->dpte_row_bw,
+                               locals->cursor_bw_pre,
+                               locals->prefetch_vmrow_bw,
+                               mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+                               s->dummy_unit_vector,
+                               s->dummy_unit_vector,
+                               s->dummy_unit_vector,
+                               s->dummy_unit_vector,
+                               s->dummy_unit_vector,
+                               s->dummy_unit_vector,
+
+                               /* output */
+                               &locals->non_urgent_total_dcn_read_bw_with_flip, // dml_float_t *TotalBandwidth
+                               &locals->non_urgent_total_dcn_read_bw_with_flip_not_including_MALL_prefetch, // dml_float_t TotalBandwidthNotIncludingMALLPrefetch
+                               &locals->FractionOfUrgentBandwidthImmediateFlip, // dml_float_t *FractionOfUrgentBandwidth
+                               &s->dummy_boolean[0]); // dml_bool_t *ImmediateFlipBandwidthSupport
+
+                       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                               if (mode_lib->ms.policy.ImmediateFlipRequirement[k] != dml_immediate_flip_not_required && locals->ImmediateFlipSupportedForPipe[k] == false) {
+                                       locals->ImmediateFlipSupported = false;
+#ifdef __DML_VBA_DEBUG__
+                                       dml_print("DML::%s: Pipe %0d not supporing iflip\n", __func__, k);
+#endif
+                               }
+                       }
+               } else {
+                       locals->ImmediateFlipSupported = false;
+                       locals->total_dcn_read_bw_with_flip = s->MaxTotalRDBandwidth;
+                       locals->total_dcn_read_bw_with_flip_not_including_MALL_prefetch = s->MaxTotalRDBandwidthNotIncludingMALLPrefetch;
+                       locals->non_urgent_total_dcn_read_bw_with_flip = s->NonUrgentMaxTotalRDBandwidth;
+                       locals->non_urgent_total_dcn_read_bw_with_flip_not_including_MALL_prefetch = s->NonUrgentMaxTotalRDBandwidthNotIncludingMALLPrefetch;
+               }
+
+               /* consider flip support is okay if the flip bw is ok or (when user does't require a iflip and there is no host vm) */
+               locals->PrefetchAndImmediateFlipSupported = (locals->PrefetchModeSupported == true &&
+                                                                                                       ((!mode_lib->ms.support.ImmediateFlipSupport && !mode_lib->ms.cache_display_cfg.plane.HostVMEnable && !s->ImmediateFlipRequirementFinal) ||
+                                                                                                       locals->ImmediateFlipSupported)) ? true : false;
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: PrefetchModeSupported = %u\n", __func__, locals->PrefetchModeSupported);
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k)
+               dml_print("DML::%s: ImmediateFlipRequirement[%u] = %u\n", __func__, k, mode_lib->ms.policy.ImmediateFlipRequirement[k] == dml_immediate_flip_required);
+               dml_print("DML::%s: HostVMEnable = %u\n", __func__, mode_lib->ms.cache_display_cfg.plane.HostVMEnable);
+               dml_print("DML::%s: ImmediateFlipSupport = %u (from mode_support)\n", __func__, mode_lib->ms.support.ImmediateFlipSupport);
+               dml_print("DML::%s: ImmediateFlipSupported = %u\n", __func__, locals->ImmediateFlipSupported);
+               dml_print("DML::%s: PrefetchAndImmediateFlipSupported = %u\n", __func__, locals->PrefetchAndImmediateFlipSupported);
+#endif
+               dml_print("DML::%s: Done one iteration: VStartupLines=%u, MaxVStartupAllPlanes=%u\n", __func__, s->VStartupLines, s->MaxVStartupAllPlanes);
+
+               s->VStartupLines = s->VStartupLines + 1;
+
+               if (s->VStartupLines > s->MaxVStartupAllPlanes) {
+                       s->VStartupLines = __DML_VBA_MIN_VSTARTUP__;
+
+                       for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+                               s->NextPrefetchMode[k] = s->NextPrefetchMode[k] + 1;
+
+                               if (s->NextPrefetchMode[k] <= s->MaxPrefetchMode[k])
+                                       s->AllPrefetchModeTested = false;
+                               dml_print("DML::%s: VStartupLines=%u, reaches max vstartup, try next prefetch mode=%u\n", __func__, s->VStartupLines-1, s->AllPrefetchModeTested);
+                       }
+               } else {
+                       s->AllPrefetchModeTested = false;
+               }
+               s->iteration++;
+               if (s->iteration > 2500) {
+                       dml_print("ERROR: DML::%s: Too many errors, exit now\n", __func__);
+                       ASSERT(0);
+               }
+       } while (!(locals->PrefetchAndImmediateFlipSupported || s->AllPrefetchModeTested));
+
+       if (locals->PrefetchAndImmediateFlipSupported) {
+               dml_print("DML::%s: Good, Prefetch and flip scheduling solution found at VStartupLines=%u (MaxVStartupAllPlanes=%u)\n", __func__, s->VStartupLines-1, s->MaxVStartupAllPlanes);
+       } else {
+               dml_print("DML::%s: Bad, Prefetch and flip scheduling soluation NOT found solution! (MaxVStartupAllPlanes=%u)\n", __func__, s->MaxVStartupAllPlanes);
+       }
+
+       //Watermarks and NB P-State/DRAM Clock Change Support
+       {
+               s->mmSOCParameters.UrgentLatency = locals->UrgentLatency;
+               s->mmSOCParameters.ExtraLatency = locals->UrgentExtraLatency;
+               s->mmSOCParameters.WritebackLatency = mode_lib->ms.state.writeback_latency_us;
+               s->mmSOCParameters.DRAMClockChangeLatency = mode_lib->ms.state.dram_clock_change_latency_us;
+               s->mmSOCParameters.FCLKChangeLatency = mode_lib->ms.state.fclk_change_latency_us;
+               s->mmSOCParameters.SRExitTime = mode_lib->ms.state.sr_exit_time_us;
+               s->mmSOCParameters.SREnterPlusExitTime = mode_lib->ms.state.sr_enter_plus_exit_time_us;
+               s->mmSOCParameters.SRExitZ8Time = mode_lib->ms.state.sr_exit_z8_time_us;
+               s->mmSOCParameters.SREnterPlusExitZ8Time = mode_lib->ms.state.sr_enter_plus_exit_z8_time_us;
+               s->mmSOCParameters.USRRetrainingLatency = mode_lib->ms.state.usr_retraining_latency_us;
+               s->mmSOCParameters.SMNLatency = mode_lib->ms.soc.smn_latency_us;
+
+               CalculateWatermarks_params->USRRetrainingRequiredFinal = mode_lib->ms.policy.USRRetrainingRequiredFinal;
+               CalculateWatermarks_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+               CalculateWatermarks_params->PrefetchMode = locals->PrefetchMode;
+               CalculateWatermarks_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+               CalculateWatermarks_params->MaxLineBufferLines = mode_lib->ms.ip.max_line_buffer_lines;
+               CalculateWatermarks_params->LineBufferSize = mode_lib->ms.ip.line_buffer_size_bits;
+               CalculateWatermarks_params->WritebackInterfaceBufferSize = mode_lib->ms.ip.writeback_interface_buffer_size_kbytes;
+               CalculateWatermarks_params->DCFCLK = locals->Dcfclk;
+               CalculateWatermarks_params->ReturnBW = mode_lib->ms.ReturnBW;
+               CalculateWatermarks_params->SynchronizeTimingsFinal = mode_lib->ms.policy.SynchronizeTimingsFinal;
+               CalculateWatermarks_params->SynchronizeDRRDisplaysForUCLKPStateChangeFinal = mode_lib->ms.policy.SynchronizeDRRDisplaysForUCLKPStateChangeFinal;
+               CalculateWatermarks_params->DRRDisplay = mode_lib->ms.cache_display_cfg.timing.DRRDisplay;
+               CalculateWatermarks_params->dpte_group_bytes = locals->dpte_group_bytes;
+               CalculateWatermarks_params->meta_row_height = locals->meta_row_height;
+               CalculateWatermarks_params->meta_row_height_chroma = locals->meta_row_height_chroma;
+               CalculateWatermarks_params->mmSOCParameters = s->mmSOCParameters;
+               CalculateWatermarks_params->WritebackChunkSize = mode_lib->ms.ip.writeback_chunk_size_kbytes;
+               CalculateWatermarks_params->SOCCLK = mode_lib->ms.SOCCLK;
+               CalculateWatermarks_params->DCFClkDeepSleep = locals->DCFCLKDeepSleep;
+               CalculateWatermarks_params->DETBufferSizeY = locals->DETBufferSizeY;
+               CalculateWatermarks_params->DETBufferSizeC = locals->DETBufferSizeC;
+               CalculateWatermarks_params->SwathHeightY = locals->SwathHeightY;
+               CalculateWatermarks_params->SwathHeightC = locals->SwathHeightC;
+               CalculateWatermarks_params->LBBitPerPixel = mode_lib->ms.cache_display_cfg.plane.LBBitPerPixel;
+               CalculateWatermarks_params->SwathWidthY = locals->SwathWidthY;
+               CalculateWatermarks_params->SwathWidthC = locals->SwathWidthC;
+               CalculateWatermarks_params->HRatio = mode_lib->ms.cache_display_cfg.plane.HRatio;
+               CalculateWatermarks_params->HRatioChroma = mode_lib->ms.cache_display_cfg.plane.HRatioChroma;
+               CalculateWatermarks_params->VTaps = mode_lib->ms.cache_display_cfg.plane.VTaps;
+               CalculateWatermarks_params->VTapsChroma = mode_lib->ms.cache_display_cfg.plane.VTapsChroma;
+               CalculateWatermarks_params->VRatio = mode_lib->ms.cache_display_cfg.plane.VRatio;
+               CalculateWatermarks_params->VRatioChroma = mode_lib->ms.cache_display_cfg.plane.VRatioChroma;
+               CalculateWatermarks_params->HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal;
+               CalculateWatermarks_params->VTotal = mode_lib->ms.cache_display_cfg.timing.VTotal;
+               CalculateWatermarks_params->VActive = mode_lib->ms.cache_display_cfg.timing.VActive;
+               CalculateWatermarks_params->PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock;
+               CalculateWatermarks_params->BlendingAndTiming = mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming;
+               CalculateWatermarks_params->DPPPerSurface = mode_lib->ms.cache_display_cfg.hw.DPPPerSurface;
+               CalculateWatermarks_params->BytePerPixelDETY = locals->BytePerPixelDETY;
+               CalculateWatermarks_params->BytePerPixelDETC = locals->BytePerPixelDETC;
+               CalculateWatermarks_params->DSTXAfterScaler = locals->DSTXAfterScaler;
+               CalculateWatermarks_params->DSTYAfterScaler = locals->DSTYAfterScaler;
+               CalculateWatermarks_params->WritebackEnable = mode_lib->ms.cache_display_cfg.writeback.WritebackEnable;
+               CalculateWatermarks_params->WritebackPixelFormat = mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat;
+               CalculateWatermarks_params->WritebackDestinationWidth = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth;
+               CalculateWatermarks_params->WritebackDestinationHeight = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight;
+               CalculateWatermarks_params->WritebackSourceHeight = mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight;
+               CalculateWatermarks_params->UnboundedRequestEnabled = locals->UnboundedRequestEnabled;
+               CalculateWatermarks_params->CompressedBufferSizeInkByte = locals->CompressedBufferSizeInkByte;
+
+               // Output
+               CalculateWatermarks_params->Watermark = &s->dummy_watermark; // Watermarks *Watermark
+               CalculateWatermarks_params->DRAMClockChangeSupport = &mode_lib->ms.support.DRAMClockChangeSupport[j];
+               CalculateWatermarks_params->MaxActiveDRAMClockChangeLatencySupported = &s->dummy_single_array[0][0]; // dml_float_t *MaxActiveDRAMClockChangeLatencySupported[]
+               CalculateWatermarks_params->SubViewportLinesNeededInMALL = &mode_lib->ms.SubViewportLinesNeededInMALL[j]; // dml_uint_t SubViewportLinesNeededInMALL[]
+               CalculateWatermarks_params->FCLKChangeSupport = &mode_lib->ms.support.FCLKChangeSupport[j];
+               CalculateWatermarks_params->MaxActiveFCLKChangeLatencySupported = &s->dummy_single[0]; // dml_float_t *MaxActiveFCLKChangeLatencySupported
+               CalculateWatermarks_params->USRRetrainingSupport = &mode_lib->ms.support.USRRetrainingSupport[j];
+
+               CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
+                       &mode_lib->scratch,
+                       CalculateWatermarks_params);
+
+               /* Copy the calculated watermarks to mp.Watermark as the getter functions are
+                * implemented by the DML team to copy the calculated values from the mp.Watermark interface.
+                */
+               memcpy(&mode_lib->mp.Watermark, CalculateWatermarks_params->Watermark, sizeof(struct Watermarks));
+
+               for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+                       if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true) {
+                               locals->WritebackAllowDRAMClockChangeEndPosition[k] = dml_max(0, locals->VStartupMin[k] * mode_lib->ms.cache_display_cfg.timing.HTotal[k] /
+                                                                                                                                                       mode_lib->ms.cache_display_cfg.timing.PixelClock[k] - locals->Watermark.WritebackDRAMClockChangeWatermark);
+                               locals->WritebackAllowFCLKChangeEndPosition[k] = dml_max(0, locals->VStartupMin[k] * mode_lib->ms.cache_display_cfg.timing.HTotal[k] /
+                                                                                                                                                       mode_lib->ms.cache_display_cfg.timing.PixelClock[k] - locals->Watermark.WritebackFCLKChangeWatermark);
+                       } else {
+                               locals->WritebackAllowDRAMClockChangeEndPosition[k] = 0;
+                               locals->WritebackAllowFCLKChangeEndPosition[k] = 0;
+                       }
+               }
+       }
+
+       //Display Pipeline Delivery Time in Prefetch, Groups
+       CalculatePixelDeliveryTimes(
+                       mode_lib->ms.num_active_planes,
+                       mode_lib->ms.cache_display_cfg.plane.VRatio,
+                       mode_lib->ms.cache_display_cfg.plane.VRatioChroma,
+                       locals->VRatioPrefetchY,
+                       locals->VRatioPrefetchC,
+                       locals->swath_width_luma_ub,
+                       locals->swath_width_chroma_ub,
+                       mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+                       mode_lib->ms.cache_display_cfg.plane.HRatio,
+                       mode_lib->ms.cache_display_cfg.plane.HRatioChroma,
+                       mode_lib->ms.cache_display_cfg.timing.PixelClock,
+                       locals->PSCL_THROUGHPUT,
+                       locals->PSCL_THROUGHPUT_CHROMA,
+                       locals->Dppclk,
+                       locals->BytePerPixelC,
+                       mode_lib->ms.cache_display_cfg.plane.SourceScan,
+                       mode_lib->ms.cache_display_cfg.plane.NumberOfCursors,
+                       mode_lib->ms.cache_display_cfg.plane.CursorWidth,
+                       mode_lib->ms.cache_display_cfg.plane.CursorBPP,
+                       locals->BlockWidth256BytesY,
+                       locals->BlockHeight256BytesY,
+                       locals->BlockWidth256BytesC,
+                       locals->BlockHeight256BytesC,
+
+                       /* Output */
+                       locals->DisplayPipeLineDeliveryTimeLuma,
+                       locals->DisplayPipeLineDeliveryTimeChroma,
+                       locals->DisplayPipeLineDeliveryTimeLumaPrefetch,
+                       locals->DisplayPipeLineDeliveryTimeChromaPrefetch,
+                       locals->DisplayPipeRequestDeliveryTimeLuma,
+                       locals->DisplayPipeRequestDeliveryTimeChroma,
+                       locals->DisplayPipeRequestDeliveryTimeLumaPrefetch,
+                       locals->DisplayPipeRequestDeliveryTimeChromaPrefetch,
+                       locals->CursorRequestDeliveryTime,
+                       locals->CursorRequestDeliveryTimePrefetch);
+
+       CalculateMetaAndPTETimes(
+                       locals->use_one_row_for_frame,
+                       mode_lib->ms.num_active_planes,
+                       mode_lib->ms.cache_display_cfg.plane.GPUVMEnable,
+                       mode_lib->ms.ip.meta_chunk_size_kbytes,
+                       mode_lib->ms.ip.min_meta_chunk_size_bytes,
+                       mode_lib->ms.cache_display_cfg.timing.HTotal,
+                       mode_lib->ms.cache_display_cfg.plane.VRatio,
+                       mode_lib->ms.cache_display_cfg.plane.VRatioChroma,
+                       locals->DestinationLinesToRequestRowInVBlank,
+                       locals->DestinationLinesToRequestRowInImmediateFlip,
+                       mode_lib->ms.cache_display_cfg.surface.DCCEnable,
+                       mode_lib->ms.cache_display_cfg.timing.PixelClock,
+                       locals->BytePerPixelY,
+                       locals->BytePerPixelC,
+                       mode_lib->ms.cache_display_cfg.plane.SourceScan,
+                       locals->dpte_row_height,
+                       locals->dpte_row_height_chroma,
+                       locals->meta_row_width,
+                       locals->meta_row_width_chroma,
+                       locals->meta_row_height,
+                       locals->meta_row_height_chroma,
+                       locals->meta_req_width,
+                       locals->meta_req_width_chroma,
+                       locals->meta_req_height,
+                       locals->meta_req_height_chroma,
+                       locals->dpte_group_bytes,
+                       locals->PTERequestSizeY,
+                       locals->PTERequestSizeC,
+                       locals->PixelPTEReqWidthY,
+                       locals->PixelPTEReqHeightY,
+                       locals->PixelPTEReqWidthC,
+                       locals->PixelPTEReqHeightC,
+                       locals->dpte_row_width_luma_ub,
+                       locals->dpte_row_width_chroma_ub,
+
+                       /* Output */
+                       locals->DST_Y_PER_PTE_ROW_NOM_L,
+                       locals->DST_Y_PER_PTE_ROW_NOM_C,
+                       locals->DST_Y_PER_META_ROW_NOM_L,
+                       locals->DST_Y_PER_META_ROW_NOM_C,
+                       locals->TimePerMetaChunkNominal,
+                       locals->TimePerChromaMetaChunkNominal,
+                       locals->TimePerMetaChunkVBlank,
+                       locals->TimePerChromaMetaChunkVBlank,
+                       locals->TimePerMetaChunkFlip,
+                       locals->TimePerChromaMetaChunkFlip,
+                       locals->time_per_pte_group_nom_luma,
+                       locals->time_per_pte_group_vblank_luma,
+                       locals->time_per_pte_group_flip_luma,
+                       locals->time_per_pte_group_nom_chroma,
+                       locals->time_per_pte_group_vblank_chroma,
+                       locals->time_per_pte_group_flip_chroma);
+
+       CalculateVMGroupAndRequestTimes(
+                       mode_lib->ms.num_active_planes,
+                       mode_lib->ms.cache_display_cfg.plane.GPUVMEnable,
+                       mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels,
+                       mode_lib->ms.cache_display_cfg.timing.HTotal,
+                       locals->BytePerPixelC,
+                       locals->DestinationLinesToRequestVMInVBlank,
+                       locals->DestinationLinesToRequestVMInImmediateFlip,
+                       mode_lib->ms.cache_display_cfg.surface.DCCEnable,
+                       mode_lib->ms.cache_display_cfg.timing.PixelClock,
+                       locals->dpte_row_width_luma_ub,
+                       locals->dpte_row_width_chroma_ub,
+                       locals->vm_group_bytes,
+                       locals->dpde0_bytes_per_frame_ub_l,
+                       locals->dpde0_bytes_per_frame_ub_c,
+                       locals->meta_pte_bytes_per_frame_ub_l,
+                       locals->meta_pte_bytes_per_frame_ub_c,
+
+                       /* Output */
+                       locals->TimePerVMGroupVBlank,
+                       locals->TimePerVMGroupFlip,
+                       locals->TimePerVMRequestVBlank,
+                       locals->TimePerVMRequestFlip);
+
+       // Min TTUVBlank
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (locals->PrefetchMode[k] == 0) {
+                       locals->MinTTUVBlank[k] = dml_max4(
+                                       locals->Watermark.DRAMClockChangeWatermark,
+                                       locals->Watermark.FCLKChangeWatermark,
+                                       locals->Watermark.StutterEnterPlusExitWatermark,
+                                       locals->Watermark.UrgentWatermark);
+               } else if (locals->PrefetchMode[k] == 1) {
+                       locals->MinTTUVBlank[k] = dml_max3(
+                                       locals->Watermark.FCLKChangeWatermark,
+                                       locals->Watermark.StutterEnterPlusExitWatermark,
+                                       locals->Watermark.UrgentWatermark);
+               } else if (locals->PrefetchMode[k] == 2) {
+                       locals->MinTTUVBlank[k] = dml_max(
+                                       locals->Watermark.StutterEnterPlusExitWatermark,
+                                       locals->Watermark.UrgentWatermark);
+               } else {
+                       locals->MinTTUVBlank[k] = locals->Watermark.UrgentWatermark;
+               }
+               if (!mode_lib->ms.cache_display_cfg.plane.DynamicMetadataEnable[k])
+                       locals->MinTTUVBlank[k] = locals->TCalc + locals->MinTTUVBlank[k];
+       }
+
+       // DCC Configuration
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: Calculate DCC configuration for surface k=%u\n", __func__, k);
+#endif
+               CalculateDCCConfiguration(
+                       mode_lib->ms.cache_display_cfg.surface.DCCEnable[k],
+                       mode_lib->ms.policy.DCCProgrammingAssumesScanDirectionUnknownFinal,
+                       mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k],
+                       mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY[k],
+                       mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC[k],
+                       mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY[k],
+                       mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC[k],
+                       mode_lib->ms.NomDETInKByte,
+                       locals->BlockHeight256BytesY[k],
+                       locals->BlockHeight256BytesC[k],
+                       mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k],
+                       locals->BytePerPixelY[k],
+                       locals->BytePerPixelC[k],
+                       locals->BytePerPixelDETY[k],
+                       locals->BytePerPixelDETC[k],
+                       mode_lib->ms.cache_display_cfg.plane.SourceScan[k],
+                       /* Output */
+                       &locals->DCCYMaxUncompressedBlock[k],
+                       &locals->DCCCMaxUncompressedBlock[k],
+                       &locals->DCCYMaxCompressedBlock[k],
+                       &locals->DCCCMaxCompressedBlock[k],
+                       &locals->DCCYIndependentBlock[k],
+                       &locals->DCCCIndependentBlock[k]);
+       }
+
+       // VStartup Adjustment
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               s->Tvstartup_margin = (s->MaxVStartupLines[k] - locals->VStartupMin[k]) * mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k];
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, MinTTUVBlank = %f (before vstartup margin)\n", __func__, k, locals->MinTTUVBlank[k]);
+#endif
+
+               locals->MinTTUVBlank[k] = locals->MinTTUVBlank[k] + s->Tvstartup_margin;
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, Tvstartup_margin = %f\n", __func__, k, s->Tvstartup_margin);
+               dml_print("DML::%s: k=%u, MaxVStartupLines = %u\n", __func__, k, s->MaxVStartupLines[k]);
+               dml_print("DML::%s: k=%u, MinTTUVBlank = %f\n", __func__, k, locals->MinTTUVBlank[k]);
+#endif
+
+               locals->Tdmdl[k] = locals->Tdmdl[k] + s->Tvstartup_margin;
+               if (mode_lib->ms.cache_display_cfg.plane.DynamicMetadataEnable[k] && mode_lib->ms.ip.dynamic_metadata_vm_enabled) {
+                       locals->Tdmdl_vm[k] = locals->Tdmdl_vm[k] + s->Tvstartup_margin;
+               }
+
+               isInterlaceTiming        = (mode_lib->ms.cache_display_cfg.timing.Interlace[k] && !mode_lib->ms.ip.ptoi_supported);
+
+               // The actual positioning of the vstartup
+               locals->VStartup[k] = (isInterlaceTiming ? (2 * s->MaxVStartupLines[k]) : s->MaxVStartupLines[k]);
+
+               s->dlg_vblank_start        =  ((isInterlaceTiming ? dml_floor((mode_lib->ms.cache_display_cfg.timing.VTotal[k] - mode_lib->ms.cache_display_cfg.timing.VFrontPorch[k]) / 2.0, 1.0) :
+                                                                                                                                       mode_lib->ms.cache_display_cfg.timing.VTotal[k]) - mode_lib->ms.cache_display_cfg.timing.VFrontPorch[k]);
+               s->LSetup                  = dml_floor(4.0 * locals->TSetup[k] / ((dml_float_t) mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]), 1.0) / 4.0;
+               s->blank_lines_remaining   = (mode_lib->ms.cache_display_cfg.timing.VTotal[k] - mode_lib->ms.cache_display_cfg.timing.VActive[k]) - locals->VStartup[k];
+
+               if (s->blank_lines_remaining < 0) {
+                       dml_print("ERROR: Vstartup is larger than vblank!?\n");
+                       s->blank_lines_remaining = 0;
+                       ASSERT(0);
+               }
+               locals->MIN_DST_Y_NEXT_START[k] = s->dlg_vblank_start + s->blank_lines_remaining + s->LSetup;
+
+               // debug only
+               s->old_MIN_DST_Y_NEXT_START = ((isInterlaceTiming ? dml_floor((mode_lib->ms.cache_display_cfg.timing.VTotal[k] - mode_lib->ms.cache_display_cfg.timing.VFrontPorch[k]) / 2.0, 1.0) :
+                                                                                                                                       mode_lib->ms.cache_display_cfg.timing.VTotal[k]) - mode_lib->ms.cache_display_cfg.timing.VFrontPorch[k])
+                                                                                       + dml_max(1.0, dml_ceil((dml_float_t) locals->WritebackDelay[k] / ((dml_float_t) mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]), 1.0))
+                                                                                       + dml_floor(4.0 * locals->TSetup[k] / ((dml_float_t) mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]), 1.0) / 4.0;
+
+               if (((locals->VUpdateOffsetPix[k] + locals->VUpdateWidthPix[k] + locals->VReadyOffsetPix[k]) / mode_lib->ms.cache_display_cfg.timing.HTotal[k]) <=
+                       (isInterlaceTiming ?
+                               dml_floor((mode_lib->ms.cache_display_cfg.timing.VTotal[k] - mode_lib->ms.cache_display_cfg.timing.VActive[k] - mode_lib->ms.cache_display_cfg.timing.VFrontPorch[k] - locals->VStartup[k]) / 2.0, 1.0) :
+                               (int) (mode_lib->ms.cache_display_cfg.timing.VTotal[k] - mode_lib->ms.cache_display_cfg.timing.VActive[k] - mode_lib->ms.cache_display_cfg.timing.VFrontPorch[k] - locals->VStartup[k]))) {
+                       locals->VREADY_AT_OR_AFTER_VSYNC[k] = true;
+               } else {
+                       locals->VREADY_AT_OR_AFTER_VSYNC[k] = false;
+               }
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, VStartup = %u (max)\n", __func__, k, locals->VStartup[k]);
+               dml_print("DML::%s: k=%u, VStartupMin = %u (max)\n", __func__, k, locals->VStartupMin[k]);
+               dml_print("DML::%s: k=%u, VUpdateOffsetPix = %u\n", __func__, k, locals->VUpdateOffsetPix[k]);
+               dml_print("DML::%s: k=%u, VUpdateWidthPix = %u\n", __func__, k, locals->VUpdateWidthPix[k]);
+               dml_print("DML::%s: k=%u, VReadyOffsetPix = %u\n", __func__, k, locals->VReadyOffsetPix[k]);
+               dml_print("DML::%s: k=%u, HTotal = %u\n", __func__, k, mode_lib->ms.cache_display_cfg.timing.HTotal[k]);
+               dml_print("DML::%s: k=%u, VTotal = %u\n", __func__, k, mode_lib->ms.cache_display_cfg.timing.VTotal[k]);
+               dml_print("DML::%s: k=%u, VActive = %u\n", __func__, k, mode_lib->ms.cache_display_cfg.timing.VActive[k]);
+               dml_print("DML::%s: k=%u, VFrontPorch = %u\n", __func__, k, mode_lib->ms.cache_display_cfg.timing.VFrontPorch[k]);
+               dml_print("DML::%s: k=%u, TSetup = %f\n", __func__, k, locals->TSetup[k]);
+               dml_print("DML::%s: k=%u, MIN_DST_Y_NEXT_START = %f\n", __func__, k, locals->MIN_DST_Y_NEXT_START[k]);
+               dml_print("DML::%s: k=%u, MIN_DST_Y_NEXT_START = %f (old)\n", __func__, k, s->old_MIN_DST_Y_NEXT_START);
+               dml_print("DML::%s: k=%u, VREADY_AT_OR_AFTER_VSYNC = %u\n", __func__, k, locals->VREADY_AT_OR_AFTER_VSYNC[k]);
+#endif
+       }
+
+       //Maximum Bandwidth Used
+       s->TotalWRBandwidth = 0;
+       s->WRBandwidth = 0;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true && mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[k] == dml_444_32) {
+                       s->WRBandwidth = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k] * mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[k] /
+                                                       (mode_lib->ms.cache_display_cfg.timing.HTotal[k] * mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * 4;
+               } else if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true) {
+                       s->WRBandwidth = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k] * mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[k] /
+                                                       (mode_lib->ms.cache_display_cfg.timing.HTotal[k] * mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * 8;
+               }
+               s->TotalWRBandwidth = s->TotalWRBandwidth + s->WRBandwidth;
+       }
+
+       locals->TotalDataReadBandwidth = 0;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               locals->TotalDataReadBandwidth = locals->TotalDataReadBandwidth + locals->ReadBandwidthSurfaceLuma[k] + locals->ReadBandwidthSurfaceChroma[k];
+
+#ifdef __DML_VBA_DEBUG__
+               dml_print("DML::%s: k=%u, TotalDataReadBandwidth = %f\n", __func__, k, locals->TotalDataReadBandwidth);
+               dml_print("DML::%s: k=%u, ReadBandwidthSurfaceLuma = %f\n", __func__, k, locals->ReadBandwidthSurfaceLuma[k]);
+               dml_print("DML::%s: k=%u, ReadBandwidthSurfaceChroma = %f\n", __func__, k, locals->ReadBandwidthSurfaceChroma[k]);
+#endif
+       }
+
+       locals->TotalDataReadBandwidthNotIncludingMALLPrefetch = 0;
+       for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+               if (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) {
+                       locals->TotalDataReadBandwidthNotIncludingMALLPrefetch = locals->TotalDataReadBandwidthNotIncludingMALLPrefetch
+               + locals->ReadBandwidthSurfaceLuma[k] + locals->ReadBandwidthSurfaceChroma[k];
+               }
+       }
+
+       CalculateStutterEfficiency_params->CompressedBufferSizeInkByte = locals->CompressedBufferSizeInkByte;
+       CalculateStutterEfficiency_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+       CalculateStutterEfficiency_params->UnboundedRequestEnabled = locals->UnboundedRequestEnabled;
+       CalculateStutterEfficiency_params->MetaFIFOSizeInKEntries = mode_lib->ms.ip.meta_fifo_size_in_kentries;
+       CalculateStutterEfficiency_params->ZeroSizeBufferEntries = mode_lib->ms.ip.zero_size_buffer_entries;
+       CalculateStutterEfficiency_params->PixelChunkSizeInKByte = mode_lib->ms.ip.pixel_chunk_size_kbytes;
+       CalculateStutterEfficiency_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+       CalculateStutterEfficiency_params->ROBBufferSizeInKByte = mode_lib->ms.ip.rob_buffer_size_kbytes;
+       CalculateStutterEfficiency_params->TotalDataReadBandwidth = locals->TotalDataReadBandwidth;
+       CalculateStutterEfficiency_params->DCFCLK = locals->Dcfclk;
+       CalculateStutterEfficiency_params->ReturnBW = mode_lib->ms.ReturnBW;
+       CalculateStutterEfficiency_params->CompbufReservedSpace64B = locals->compbuf_reserved_space_64b;
+       CalculateStutterEfficiency_params->CompbufReservedSpaceZs = locals->compbuf_reserved_space_zs;
+       CalculateStutterEfficiency_params->SRExitTime = mode_lib->ms.state.sr_exit_time_us;
+       CalculateStutterEfficiency_params->SRExitZ8Time = mode_lib->ms.state.sr_exit_z8_time_us;
+       CalculateStutterEfficiency_params->SynchronizeTimingsFinal = mode_lib->ms.policy.SynchronizeTimingsFinal;
+       CalculateStutterEfficiency_params->BlendingAndTiming = mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming;
+       CalculateStutterEfficiency_params->StutterEnterPlusExitWatermark = locals->Watermark.StutterEnterPlusExitWatermark;
+       CalculateStutterEfficiency_params->Z8StutterEnterPlusExitWatermark = locals->Watermark.Z8StutterEnterPlusExitWatermark;
+       CalculateStutterEfficiency_params->ProgressiveToInterlaceUnitInOPP = mode_lib->ms.ip.ptoi_supported;
+       CalculateStutterEfficiency_params->Interlace = mode_lib->ms.cache_display_cfg.timing.Interlace;
+       CalculateStutterEfficiency_params->MinTTUVBlank = locals->MinTTUVBlank;
+       CalculateStutterEfficiency_params->DPPPerSurface = mode_lib->ms.cache_display_cfg.hw.DPPPerSurface;
+       CalculateStutterEfficiency_params->DETBufferSizeY = locals->DETBufferSizeY;
+       CalculateStutterEfficiency_params->BytePerPixelY = locals->BytePerPixelY;
+       CalculateStutterEfficiency_params->BytePerPixelDETY = locals->BytePerPixelDETY;
+       CalculateStutterEfficiency_params->SwathWidthY = locals->SwathWidthY;
+       CalculateStutterEfficiency_params->SwathHeightY = locals->SwathHeightY;
+       CalculateStutterEfficiency_params->SwathHeightC = locals->SwathHeightC;
+       CalculateStutterEfficiency_params->NetDCCRateLuma = mode_lib->ms.cache_display_cfg.surface.DCCRateLuma;
+       CalculateStutterEfficiency_params->NetDCCRateChroma = mode_lib->ms.cache_display_cfg.surface.DCCRateChroma;
+       CalculateStutterEfficiency_params->DCCFractionOfZeroSizeRequestsLuma = mode_lib->ms.cache_display_cfg.surface.DCCFractionOfZeroSizeRequestsLuma;
+       CalculateStutterEfficiency_params->DCCFractionOfZeroSizeRequestsChroma = mode_lib->ms.cache_display_cfg.surface.DCCFractionOfZeroSizeRequestsChroma;
+       CalculateStutterEfficiency_params->HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal;
+       CalculateStutterEfficiency_params->VTotal = mode_lib->ms.cache_display_cfg.timing.VTotal;
+       CalculateStutterEfficiency_params->PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock;
+       CalculateStutterEfficiency_params->VRatio = mode_lib->ms.cache_display_cfg.plane.VRatio;
+       CalculateStutterEfficiency_params->SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan;
+       CalculateStutterEfficiency_params->BlockHeight256BytesY = locals->BlockHeight256BytesY;
+       CalculateStutterEfficiency_params->BlockWidth256BytesY = locals->BlockWidth256BytesY;
+       CalculateStutterEfficiency_params->BlockHeight256BytesC = locals->BlockHeight256BytesC;
+       CalculateStutterEfficiency_params->BlockWidth256BytesC = locals->BlockWidth256BytesC;
+       CalculateStutterEfficiency_params->DCCYMaxUncompressedBlock = locals->DCCYMaxUncompressedBlock;
+       CalculateStutterEfficiency_params->DCCCMaxUncompressedBlock = locals->DCCCMaxUncompressedBlock;
+       CalculateStutterEfficiency_params->VActive = mode_lib->ms.cache_display_cfg.timing.VActive;
+       CalculateStutterEfficiency_params->DCCEnable = mode_lib->ms.cache_display_cfg.surface.DCCEnable;
+       CalculateStutterEfficiency_params->WritebackEnable = mode_lib->ms.cache_display_cfg.writeback.WritebackEnable;
+       CalculateStutterEfficiency_params->ReadBandwidthSurfaceLuma = locals->ReadBandwidthSurfaceLuma;
+       CalculateStutterEfficiency_params->ReadBandwidthSurfaceChroma = locals->ReadBandwidthSurfaceChroma;
+       CalculateStutterEfficiency_params->meta_row_bw = locals->meta_row_bw;
+       CalculateStutterEfficiency_params->dpte_row_bw = locals->dpte_row_bw;
+       CalculateStutterEfficiency_params->StutterEfficiencyNotIncludingVBlank = &locals->StutterEfficiencyNotIncludingVBlank;
+       CalculateStutterEfficiency_params->StutterEfficiency = &locals->StutterEfficiency;
+       CalculateStutterEfficiency_params->NumberOfStutterBurstsPerFrame = &locals->NumberOfStutterBurstsPerFrame;
+       CalculateStutterEfficiency_params->Z8StutterEfficiencyNotIncludingVBlank = &locals->Z8StutterEfficiencyNotIncludingVBlank;
+       CalculateStutterEfficiency_params->Z8StutterEfficiency = &locals->Z8StutterEfficiency;
+       CalculateStutterEfficiency_params->Z8NumberOfStutterBurstsPerFrame = &locals->Z8NumberOfStutterBurstsPerFrame;
+       CalculateStutterEfficiency_params->StutterPeriod = &locals->StutterPeriod;
+       CalculateStutterEfficiency_params->DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE = &locals->DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE;
+
+       // Stutter Efficiency
+       CalculateStutterEfficiency(&mode_lib->scratch,
+               CalculateStutterEfficiency_params);
+
+#ifdef __DML_VBA_ALLOW_DELTA__
+       {
+       dml_float_t dummy_single[2];
+       dml_uint_t dummy_integer[1];
+       dml_bool_t dummy_boolean[1];
+
+       // Calculate z8 stutter eff assuming 0 reserved space
+       CalculateStutterEfficiency(
+                       locals->CompressedBufferSizeInkByte,
+                       mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange,
+                       locals->UnboundedRequestEnabled,
+                       mode_lib->ms.ip.meta_fifo_size_in_kentries,
+                       mode_lib->ms.ip.zero_size_buffer_entries,
+                       mode_lib->ms.ip.pixel_chunk_size_kbytes,
+                       mode_lib->ms.num_active_planes,
+                       mode_lib->ms.ip.rob_buffer_size_kbytes,
+                       locals->TotalDataReadBandwidth,
+                       locals->Dcfclk,
+                       mode_lib->ms.ReturnBW,
+                       0, //mode_lib->ms.ip.compbuf_reserved_space_64b,
+                       0, //mode_lib->ms.ip.compbuf_reserved_space_zs,
+                       mode_lib->ms.state.sr_exit_time_us,
+                       mode_lib->ms.state.sr_exit_z8_time_us,
+                       mode_lib->ms.policy.SynchronizeTimingsFinal,
+                       mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming,
+                       locals->Watermark.StutterEnterPlusExitWatermark,
+                       locals->Watermark.Z8StutterEnterPlusExitWatermark,
+                       mode_lib->ms.ip.ptoi_supported,
+                       mode_lib->ms.cache_display_cfg.timing.Interlace,
+                       locals->MinTTUVBlank,
+                       mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+                       mode_lib->ms.DETBufferSizeY,
+                       locals->BytePerPixelY,
+                       locals->BytePerPixelDETY,
+                       locals->SwathWidthY,
+                       mode_lib->ms.SwathHeightY,
+                       mode_lib->ms.SwathHeightC,
+                       mode_lib->ms.cache_display_cfg.surface.DCCRateLuma,
+                       mode_lib->ms.cache_display_cfg.surface.DCCRateChroma,
+                       mode_lib->ms.cache_display_cfg.surface.DCCFractionOfZeroSizeRequestsLuma,
+                       mode_lib->ms.cache_display_cfg.surface.DCCFractionOfZeroSizeRequestsChroma,
+                       mode_lib->ms.cache_display_cfg.timing.HTotal,
+                       mode_lib->ms.cache_display_cfg.timing.VTotal,
+                       mode_lib->ms.cache_display_cfg.timing.PixelClock,
+                       mode_lib->ms.cache_display_cfg.plane.VRatio,
+                       mode_lib->ms.cache_display_cfg.plane.SourceScan,
+                       locals->BlockHeight256BytesY,
+                       locals->BlockWidth256BytesY,
+                       locals->BlockHeight256BytesC,
+                       locals->BlockWidth256BytesC,
+                       locals->DCCYMaxUncompressedBlock,
+                       locals->DCCCMaxUncompressedBlock,
+                       mode_lib->ms.cache_display_cfg.timing.VActive,
+                       mode_lib->ms.cache_display_cfg.surface.DCCEnable,
+                       mode_lib->ms.cache_display_cfg.writeback.WritebackEnable,
+                       locals->ReadBandwidthSurfaceLuma,
+                       locals->ReadBandwidthSurfaceChroma,
+                       locals->meta_row_bw,
+                       locals->dpte_row_bw,
+
+                       /* Output */
+                       &dummy_single[0],
+                       &dummy_single[1],
+                       &dummy_integer[0],
+                       &locals->Z8StutterEfficiencyNotIncludingVBlankBestCase,
+                       &locals->Z8StutterEfficiencyBestCase,
+                       &locals->Z8NumberOfStutterBurstsPerFrameBestCase,
+                       &locals->StutterPeriodBestCase,
+                       &dummy_boolean[0]);
+       }
+#else
+       locals->Z8StutterEfficiencyNotIncludingVBlankBestCase = locals->Z8StutterEfficiencyNotIncludingVBlank;
+       locals->Z8StutterEfficiencyBestCase                   = locals->Z8StutterEfficiency;
+       locals->Z8NumberOfStutterBurstsPerFrameBestCase       = locals->Z8NumberOfStutterBurstsPerFrame;
+       locals->StutterPeriodBestCase                         = locals->StutterPeriod;
+#endif
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: --- END --- \n",  __func__);
+#endif
+} // dml_core_mode_programming
+
+/// Function: dml_core_get_row_heights
+/// @brief Get row height for DPTE and META with minimal input.
+void dml_core_get_row_heights(
+                                               dml_uint_t                         *dpte_row_height,
+                                               dml_uint_t                         *meta_row_height,
+                                               const struct display_mode_lib_st   *mode_lib,
+                                               dml_bool_t                         is_plane1,
+                                               enum dml_source_format_class        SourcePixelFormat,
+                                               enum dml_swizzle_mode            SurfaceTiling,
+                                               enum dml_rotation_angle          ScanDirection,
+                                               dml_uint_t                         pitch,
+                                               dml_uint_t                         GPUVMMinPageSizeKBytes)
+{
+       dml_uint_t BytePerPixelY;
+       dml_uint_t BytePerPixelC;
+       dml_float_t BytePerPixelInDETY;
+       dml_float_t BytePerPixelInDETC;
+       dml_uint_t BlockHeight256BytesY;
+       dml_uint_t BlockHeight256BytesC;
+       dml_uint_t BlockWidth256BytesY;
+       dml_uint_t BlockWidth256BytesC;
+       dml_uint_t MacroTileWidthY;
+       dml_uint_t MacroTileWidthC;
+       dml_uint_t MacroTileHeightY;
+       dml_uint_t MacroTileHeightC;
+
+       dml_uint_t BytePerPixel;
+       dml_uint_t BlockHeight256Bytes;
+       dml_uint_t BlockWidth256Bytes;
+       dml_uint_t MacroTileWidth;
+       dml_uint_t MacroTileHeight;
+       dml_uint_t PTEBufferSizeInRequests;
+
+       dml_uint_t dummy_integer[16];
+
+       CalculateBytePerPixelAndBlockSizes(
+               SourcePixelFormat,
+               SurfaceTiling,
+
+               /* Output */
+               &BytePerPixelY,
+               &BytePerPixelC,
+               &BytePerPixelInDETY,
+               &BytePerPixelInDETC,
+               &BlockHeight256BytesY,
+               &BlockHeight256BytesC,
+               &BlockWidth256BytesY,
+               &BlockWidth256BytesC,
+               &MacroTileHeightY,
+               &MacroTileHeightC,
+               &MacroTileWidthY,
+               &MacroTileWidthC);
+
+       BytePerPixel = is_plane1 ? BytePerPixelC : BytePerPixelY;
+       BlockHeight256Bytes = is_plane1 ? BlockHeight256BytesC : BlockHeight256BytesY;
+       BlockWidth256Bytes = is_plane1 ? BlockWidth256BytesC : BlockWidth256BytesY;
+       MacroTileWidth = is_plane1 ? MacroTileWidthC : MacroTileWidthY;
+       MacroTileHeight = is_plane1 ? MacroTileHeightC : MacroTileHeightY;
+       PTEBufferSizeInRequests = is_plane1 ? mode_lib->ip.dpte_buffer_size_in_pte_reqs_chroma : mode_lib->ip.dpte_buffer_size_in_pte_reqs_luma;
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+       dml_print("DML_DLG: %s: is_plane1 = %u\n", __func__, is_plane1);
+       dml_print("DML_DLG: %s: BytePerPixel = %u\n", __func__, BytePerPixel);
+       dml_print("DML_DLG: %s: BlockHeight256Bytes = %u\n", __func__, BlockHeight256Bytes);
+       dml_print("DML_DLG: %s: BlockWidth256Bytes = %u\n", __func__, BlockWidth256Bytes);
+       dml_print("DML_DLG: %s: MacroTileWidth = %u\n", __func__, MacroTileWidth);
+       dml_print("DML_DLG: %s: MacroTileHeight = %u\n", __func__, MacroTileHeight);
+       dml_print("DML_DLG: %s: PTEBufferSizeInRequests = %u\n", __func__, PTEBufferSizeInRequests);
+       dml_print("DML_DLG: %s: dpte_buffer_size_in_pte_reqs_luma = %u\n", __func__, mode_lib->ip.dpte_buffer_size_in_pte_reqs_luma);
+       dml_print("DML_DLG: %s: dpte_buffer_size_in_pte_reqs_chroma = %u\n", __func__, mode_lib->ip.dpte_buffer_size_in_pte_reqs_chroma);
+       dml_print("DML_DLG: %s: GPUVMMinPageSizeKBytes = %u\n", __func__, GPUVMMinPageSizeKBytes);
+#endif
+
+       // just suppluy with enough parameters to calculate meta and dte
+       CalculateVMAndRowBytes(
+                       0, // dml_bool_t ViewportStationary,
+                       1, // dml_bool_t DCCEnable,
+                       1, // dml_uint_t NumberOfDPPs,
+                       BlockHeight256Bytes,
+                       BlockWidth256Bytes,
+                       SourcePixelFormat,
+                       SurfaceTiling,
+                       BytePerPixel,
+                       ScanDirection,
+                       0, // dml_uint_t SwathWidth,
+                       0, // dml_uint_t ViewportHeight, (Note: DML calculates one_row_for_frame height regardless, would need test input if that height is useful)
+                       0, // dml_uint_t ViewportXStart,
+                       0, // dml_uint_t ViewportYStart,
+                       1, // dml_bool_t GPUVMEnable,
+                       4, // dml_uint_t GPUVMMaxPageTableLevels,
+                       GPUVMMinPageSizeKBytes,
+                       PTEBufferSizeInRequests,
+                       pitch,
+                       0, // dml_uint_t DCCMetaPitch,
+                       MacroTileWidth,
+                       MacroTileHeight,
+
+                       // /* Output */
+                       &dummy_integer[0], // dml_uint_t *MetaRowByte,
+                       &dummy_integer[1], // dml_uint_t *PixelPTEBytesPerRow,
+                       &dummy_integer[2], // dml_uint_t *PixelPTEBytesPerRowStorage,
+                       &dummy_integer[3], // dml_uint_t *dpte_row_width_ub,
+                       dpte_row_height,
+                       &dummy_integer[4], // dml_uint_t *dpte_row_height_linear
+                       &dummy_integer[5], // dml_uint_t *PixelPTEBytesPerRow_one_row_per_frame,
+                       &dummy_integer[6], // dml_uint_t *dpte_row_width_ub_one_row_per_frame,
+                       &dummy_integer[7], // dml_uint_t *dpte_row_height_one_row_per_frame,
+                       &dummy_integer[8], // dml_uint_t *MetaRequestWidth,
+                       &dummy_integer[9], // dml_uint_t *MetaRequestHeight,
+                       &dummy_integer[10], // dml_uint_t *meta_row_width,
+                       meta_row_height,
+                       &dummy_integer[11], // dml_uint_t *PixelPTEReqWidth,
+                       &dummy_integer[12], // dml_uint_t *PixelPTEReqHeight,
+                       &dummy_integer[13], // dml_uint_t *PTERequestSize,
+                       &dummy_integer[14], // dml_uint_t *DPDE0BytesFrame,
+                       &dummy_integer[15]); // dml_uint_t *MetaPTEBytesFrame)
+
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+       dml_print("DML_DLG: %s: dpte_row_height = %u\n", __func__, *dpte_row_height);
+       dml_print("DML_DLG: %s: meta_row_height = %u\n", __func__, *meta_row_height);
+#endif
+}
+
+static struct soc_state_bounding_box_st dml_get_soc_state_bounding_box(
+       const struct soc_states_st *states,
+       dml_uint_t state_idx)
+{
+       dml_print("DML::%s: state_idx=%u (num_states=%u)\n", __func__, state_idx, states->num_states);
+
+       if (state_idx >= (dml_uint_t)states->num_states) {
+               dml_print("DML::%s: ERROR: Invalid state_idx=%u! num_states=%u\n", __func__, state_idx, states->num_states);
+               ASSERT(0);
+       }
+       return (states->state_array[state_idx]);
+}
+
+/// @brief Copy the parameters to a calculation struct, it actually only need when the DML needs to have
+///        the intelligence to re-calculate when any of display cfg, bbox, or policy changes since last calculated.
+///
+static void cache_ip_soc_cfg(struct display_mode_lib_st *mode_lib,
+                                               dml_uint_t state_idx)
+{
+       mode_lib->ms.state_idx = state_idx;
+       mode_lib->ms.max_state_idx = mode_lib->states.num_states - 1;
+       mode_lib->ms.soc = mode_lib->soc;
+       mode_lib->ms.ip = mode_lib->ip;
+       mode_lib->ms.policy = mode_lib->policy;
+       mode_lib->ms.state = dml_get_soc_state_bounding_box(&mode_lib->states, state_idx);
+       mode_lib->ms.max_state = dml_get_soc_state_bounding_box(&mode_lib->states, mode_lib->states.num_states - 1);
+}
+
+static void cache_display_cfg(struct display_mode_lib_st *mode_lib,
+       const struct dml_display_cfg_st *display_cfg)
+{
+       mode_lib->ms.cache_display_cfg = *display_cfg;
+}
+
+static void fetch_socbb_params(struct display_mode_lib_st *mode_lib)
+{
+       struct soc_state_bounding_box_st *state = &mode_lib->ms.state;
+
+       // Default values, SOCCLK, DRAMSpeed, and FabricClock will be reassigned to the same state value in mode_check step
+       // If UseMinimumRequiredDCFCLK is used, the DCFCLK will be the min dcflk for the mode support
+       mode_lib->ms.SOCCLK = (dml_float_t)state->socclk_mhz;
+       mode_lib->ms.DRAMSpeed = (dml_float_t)state->dram_speed_mts;
+       mode_lib->ms.FabricClock = (dml_float_t)state->fabricclk_mhz;
+       mode_lib->ms.DCFCLK = (dml_float_t)state->dcfclk_mhz;
+}
+
+/// @brief Use display_cfg directly for mode_support calculation
+///        Calculated values and informational output are stored in mode_lib.vba data struct
+///        The display configuration is described with pipes struct and num_pipes
+///        This function is used when physical resource mapping is not finalized (for example,
+///        don't know how many pipes to represent a surface)
+/// @param mode_lib Contains the bounding box and policy setting.
+/// @param state_idx Power state index
+/// @param display_cfg Display configurations. A display
+dml_bool_t dml_mode_support(
+       struct display_mode_lib_st *mode_lib,
+       dml_uint_t                        state_idx,
+       const struct dml_display_cfg_st *display_cfg)
+{
+       dml_bool_t is_mode_support;
+
+       dml_print("DML::%s: ------------- START ----------\n", __func__);
+       cache_ip_soc_cfg(mode_lib, state_idx);
+       cache_display_cfg(mode_lib, display_cfg);
+
+       fetch_socbb_params(mode_lib);
+
+       dml_print("DML::%s: state_idx          = %u\n", __func__, state_idx);
+
+       is_mode_support = dml_core_mode_support(mode_lib);
+
+       dml_print("DML::%s: is_mode_support = %u\n", __func__, is_mode_support);
+       dml_print("DML::%s: ------------- DONE ----------\n", __func__);
+       return is_mode_support;
+}
+
+/// @Brief A function to calculate the programming values for DCN DCHUB (Assume mode is supported)
+/// The output will be stored in the mode_lib.mp (mode_program_st) data struct and those can be accessed via the getter functions
+/// Calculated values include: watermarks, dlg, rq reg, different clock frequency
+/// This function returns 1 when there is no error.
+/// Note: In this function, it is assumed that DCFCLK, SOCCLK freq are the state values, and mode_program will just use the DML calculated DPPCLK and DISPCLK
+/// @param mode_lib mode_lib data struct that house all the input/output/bbox and calculation values.
+/// @param state_idx Power state idx chosen
+/// @param display_cfg Display Congiuration
+/// @param call_standalone Calling mode_programming without calling mode support.  Some of the "support" struct member will be pre-calculated before doing mode programming
+/// TODO: Add clk_cfg input, could be useful for standalone mode
+dml_bool_t dml_mode_programming(
+       struct display_mode_lib_st *mode_lib,
+       dml_uint_t                         state_idx,
+       const struct dml_display_cfg_st *display_cfg,
+       bool                               call_standalone)
+{
+       struct dml_clk_cfg_st clk_cfg;
+       memset(&clk_cfg, 0, sizeof(clk_cfg));
+
+       clk_cfg.dcfclk_option = dml_use_required_freq;
+       clk_cfg.dispclk_option = dml_use_required_freq;
+       for (dml_uint_t k = 0; k < __DML_NUM_PLANES__; ++k)
+               clk_cfg.dppclk_option[k] = dml_use_required_freq;
+
+       dml_print("DML::%s: ------------- START ----------\n", __func__);
+       dml_print("DML::%s: state_idx       = %u\n", __func__, state_idx);
+       dml_print("DML::%s: call_standalone = %u\n", __func__, call_standalone);
+
+       cache_ip_soc_cfg(mode_lib, state_idx);
+       cache_display_cfg(mode_lib, display_cfg);
+
+       fetch_socbb_params(mode_lib);
+       if (call_standalone) {
+               mode_lib->ms.support.ImmediateFlipSupport = 1; // assume mode support say immediate flip ok at max state/combine
+               dml_core_mode_support_partial(mode_lib);
+       }
+
+       dml_core_mode_programming(mode_lib, &clk_cfg);
+
+       dml_print("DML::%s: ------------- DONE ----------\n", __func__);
+       dml_print("DML::%s: PrefetchAndImmediateFlipSupported = %0d\n", __func__, mode_lib->mp.PrefetchAndImmediateFlipSupported);
+       return mode_lib->mp.PrefetchAndImmediateFlipSupported;
+}
+
+static dml_uint_t mode_support_pwr_states(
+       dml_uint_t *lowest_state_idx,
+       struct display_mode_lib_st *mode_lib,
+       const struct dml_display_cfg_st *display_cfg,
+       dml_uint_t start_state_idx,
+       dml_uint_t end_state_idx)
+{
+       dml_uint_t state_idx = 0;
+       dml_bool_t mode_is_supported = 0;
+       *lowest_state_idx = end_state_idx;
+
+       if (end_state_idx < start_state_idx)
+               ASSERT(0);
+
+       if (end_state_idx >= mode_lib->states.num_states) // idx is 0-based
+               ASSERT(0);
+
+       for (state_idx = start_state_idx; state_idx <= end_state_idx; state_idx++) {
+               if (dml_mode_support(mode_lib, state_idx, display_cfg)) {
+                       dml_print("DML::%s: Mode is supported at power state_idx = %u\n", __func__, state_idx);
+                       mode_is_supported = 1;
+                       *lowest_state_idx = state_idx;
+                       break;
+               }
+       }
+
+       return mode_is_supported;
+}
+
+dml_uint_t dml_mode_support_ex(struct dml_mode_support_ex_params_st *in_out_params)
+{
+       dml_uint_t result;
+
+       result = mode_support_pwr_states(&in_out_params->out_lowest_state_idx,
+               in_out_params->mode_lib,
+               in_out_params->in_display_cfg,
+               0,
+               in_out_params->mode_lib->states.num_states - 1);
+
+       if (result)
+               *in_out_params->out_evaluation_info = in_out_params->mode_lib->ms.support;
+
+       return result;
+}
+
+dml_bool_t dml_get_is_phantom_pipe(struct display_mode_lib_st *mode_lib, dml_uint_t pipe_idx)
+{
+       dml_uint_t plane_idx = mode_lib->mp.pipe_plane[pipe_idx];
+       dml_print("DML::%s: pipe_idx=%d UseMALLForPStateChange=%0d\n", __func__, pipe_idx, mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[plane_idx]);
+       return (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[plane_idx] == dml_use_mall_pstate_change_phantom_pipe);
+}
+
+#define dml_get_per_surface_var_func(variable, type, interval_var) type dml_get_##variable(struct display_mode_lib_st *mode_lib, dml_uint_t surface_idx) \
+{ \
+       dml_uint_t plane_idx; \
+       plane_idx = mode_lib->mp.pipe_plane[surface_idx]; \
+       return (type) interval_var[plane_idx]; \
+}
+
+#define dml_get_var_func(var, type, internal_var)  type dml_get_##var(struct display_mode_lib_st *mode_lib) \
+{ \
+       return (type) internal_var; \
+}
+
+dml_get_var_func(wm_urgent, dml_float_t, mode_lib->mp.Watermark.UrgentWatermark);
+dml_get_var_func(wm_stutter_exit, dml_float_t, mode_lib->mp.Watermark.StutterExitWatermark);
+dml_get_var_func(wm_stutter_enter_exit, dml_float_t, mode_lib->mp.Watermark.StutterEnterPlusExitWatermark);
+dml_get_var_func(wm_memory_trip, dml_float_t, mode_lib->mp.UrgentLatency);
+dml_get_var_func(wm_fclk_change, dml_float_t, mode_lib->mp.Watermark.FCLKChangeWatermark);
+dml_get_var_func(wm_usr_retraining, dml_float_t, mode_lib->mp.Watermark.USRRetrainingWatermark);
+dml_get_var_func(wm_dram_clock_change, dml_float_t, mode_lib->mp.Watermark.DRAMClockChangeWatermark);
+dml_get_var_func(fraction_of_urgent_bandwidth, dml_float_t, mode_lib->mp.FractionOfUrgentBandwidth);
+dml_get_var_func(fraction_of_urgent_bandwidth_imm_flip, dml_float_t, mode_lib->mp.FractionOfUrgentBandwidthImmediateFlip);
+dml_get_var_func(urgent_latency, dml_float_t, mode_lib->mp.UrgentLatency);
+dml_get_var_func(clk_dcf_deepsleep, dml_float_t, mode_lib->mp.DCFCLKDeepSleep);
+dml_get_var_func(wm_writeback_dram_clock_change, dml_float_t, mode_lib->mp.Watermark.WritebackDRAMClockChangeWatermark);
+dml_get_var_func(stutter_efficiency, dml_float_t, mode_lib->mp.StutterEfficiency);
+dml_get_var_func(stutter_efficiency_no_vblank, dml_float_t, mode_lib->mp.StutterEfficiencyNotIncludingVBlank);
+dml_get_var_func(stutter_efficiency_z8, dml_float_t, mode_lib->mp.Z8StutterEfficiency);
+dml_get_var_func(stutter_num_bursts_z8, dml_float_t, mode_lib->mp.Z8NumberOfStutterBurstsPerFrame);
+dml_get_var_func(stutter_period, dml_float_t, mode_lib->mp.StutterPeriod);
+dml_get_var_func(stutter_efficiency_z8_bestcase, dml_float_t, mode_lib->mp.Z8StutterEfficiencyBestCase);
+dml_get_var_func(stutter_num_bursts_z8_bestcase, dml_float_t, mode_lib->mp.Z8NumberOfStutterBurstsPerFrameBestCase);
+dml_get_var_func(stutter_period_bestcase, dml_float_t, mode_lib->mp.StutterPeriodBestCase);
+dml_get_var_func(urgent_extra_latency, dml_float_t, mode_lib->mp.UrgentExtraLatency);
+dml_get_var_func(dispclk_calculated, dml_float_t, mode_lib->mp.Dispclk_calculated);
+dml_get_var_func(total_data_read_bw, dml_float_t, mode_lib->mp.TotalDataReadBandwidth);
+dml_get_var_func(return_bw, dml_float_t, mode_lib->ms.ReturnBW);
+dml_get_var_func(tcalc, dml_float_t, mode_lib->mp.TCalc);
+dml_get_var_func(comp_buffer_size_kbytes, dml_uint_t, mode_lib->mp.CompressedBufferSizeInkByte);
+dml_get_var_func(pixel_chunk_size_in_kbyte, dml_uint_t, mode_lib->ms.ip.pixel_chunk_size_kbytes);
+dml_get_var_func(alpha_pixel_chunk_size_in_kbyte, dml_uint_t, mode_lib->ms.ip.alpha_pixel_chunk_size_kbytes);
+dml_get_var_func(meta_chunk_size_in_kbyte, dml_uint_t, mode_lib->ms.ip.meta_chunk_size_kbytes);
+dml_get_var_func(min_pixel_chunk_size_in_byte, dml_uint_t, mode_lib->ms.ip.min_pixel_chunk_size_bytes);
+dml_get_var_func(min_meta_chunk_size_in_byte, dml_uint_t, mode_lib->ms.ip.min_meta_chunk_size_bytes);
+dml_get_var_func(total_immediate_flip_bytes, dml_uint_t, mode_lib->mp.TotImmediateFlipBytes);
+
+dml_get_per_surface_var_func(dsc_delay, dml_uint_t, mode_lib->mp.DSCDelay); // this is the dsc latency
+dml_get_per_surface_var_func(dppclk_calculated, dml_float_t, mode_lib->mp.Dppclk_calculated);
+dml_get_per_surface_var_func(dscclk_calculated, dml_float_t, mode_lib->mp.DSCCLK_calculated);
+dml_get_per_surface_var_func(min_ttu_vblank_in_us, dml_float_t, mode_lib->mp.MinTTUVBlank);
+dml_get_per_surface_var_func(vratio_prefetch_l, dml_float_t, mode_lib->mp.VRatioPrefetchY);
+dml_get_per_surface_var_func(vratio_prefetch_c, dml_float_t, mode_lib->mp.VRatioPrefetchC);
+dml_get_per_surface_var_func(dst_x_after_scaler, dml_uint_t, mode_lib->mp.DSTXAfterScaler);
+dml_get_per_surface_var_func(dst_y_after_scaler, dml_uint_t, mode_lib->mp.DSTYAfterScaler);
+dml_get_per_surface_var_func(dst_y_per_vm_vblank, dml_float_t, mode_lib->mp.DestinationLinesToRequestVMInVBlank);
+dml_get_per_surface_var_func(dst_y_per_row_vblank, dml_float_t, mode_lib->mp.DestinationLinesToRequestRowInVBlank);
+dml_get_per_surface_var_func(dst_y_prefetch, dml_float_t, mode_lib->mp.DestinationLinesForPrefetch);
+dml_get_per_surface_var_func(dst_y_per_vm_flip, dml_float_t, mode_lib->mp.DestinationLinesToRequestVMInImmediateFlip);
+dml_get_per_surface_var_func(dst_y_per_row_flip, dml_float_t, mode_lib->mp.DestinationLinesToRequestRowInImmediateFlip);
+dml_get_per_surface_var_func(dst_y_per_pte_row_nom_l, dml_float_t, mode_lib->mp.DST_Y_PER_PTE_ROW_NOM_L);
+dml_get_per_surface_var_func(dst_y_per_pte_row_nom_c, dml_float_t, mode_lib->mp.DST_Y_PER_PTE_ROW_NOM_C);
+dml_get_per_surface_var_func(dst_y_per_meta_row_nom_l, dml_float_t, mode_lib->mp.DST_Y_PER_META_ROW_NOM_L);
+dml_get_per_surface_var_func(dst_y_per_meta_row_nom_c, dml_float_t, mode_lib->mp.DST_Y_PER_META_ROW_NOM_C);
+dml_get_per_surface_var_func(refcyc_per_vm_group_vblank_in_us, dml_float_t, mode_lib->mp.TimePerVMGroupVBlank);
+dml_get_per_surface_var_func(refcyc_per_vm_group_flip_in_us, dml_float_t, mode_lib->mp.TimePerVMGroupFlip);
+dml_get_per_surface_var_func(refcyc_per_vm_req_vblank_in_us, dml_float_t, mode_lib->mp.TimePerVMRequestVBlank);
+dml_get_per_surface_var_func(refcyc_per_vm_req_flip_in_us, dml_float_t, mode_lib->mp.TimePerVMRequestFlip);
+dml_get_per_surface_var_func(refcyc_per_vm_dmdata_in_us, dml_float_t, mode_lib->mp.Tdmdl_vm);
+dml_get_per_surface_var_func(dmdata_dl_delta_in_us, dml_float_t, mode_lib->mp.Tdmdl);
+dml_get_per_surface_var_func(refcyc_per_line_delivery_l_in_us, dml_float_t, mode_lib->mp.DisplayPipeLineDeliveryTimeLuma);
+dml_get_per_surface_var_func(refcyc_per_line_delivery_c_in_us, dml_float_t, mode_lib->mp.DisplayPipeLineDeliveryTimeChroma);
+dml_get_per_surface_var_func(refcyc_per_line_delivery_pre_l_in_us, dml_float_t, mode_lib->mp.DisplayPipeLineDeliveryTimeLumaPrefetch);
+dml_get_per_surface_var_func(refcyc_per_line_delivery_pre_c_in_us, dml_float_t, mode_lib->mp.DisplayPipeLineDeliveryTimeChromaPrefetch);
+dml_get_per_surface_var_func(refcyc_per_req_delivery_l_in_us, dml_float_t, mode_lib->mp.DisplayPipeRequestDeliveryTimeLuma);
+dml_get_per_surface_var_func(refcyc_per_req_delivery_c_in_us, dml_float_t, mode_lib->mp.DisplayPipeRequestDeliveryTimeChroma);
+dml_get_per_surface_var_func(refcyc_per_req_delivery_pre_l_in_us, dml_float_t, mode_lib->mp.DisplayPipeRequestDeliveryTimeLumaPrefetch);
+dml_get_per_surface_var_func(refcyc_per_req_delivery_pre_c_in_us, dml_float_t, mode_lib->mp.DisplayPipeRequestDeliveryTimeChromaPrefetch);
+dml_get_per_surface_var_func(refcyc_per_cursor_req_delivery_in_us, dml_float_t, mode_lib->mp.CursorRequestDeliveryTime);
+dml_get_per_surface_var_func(refcyc_per_cursor_req_delivery_pre_in_us, dml_float_t, mode_lib->mp.CursorRequestDeliveryTimePrefetch);
+dml_get_per_surface_var_func(refcyc_per_meta_chunk_nom_l_in_us, dml_float_t, mode_lib->mp.TimePerMetaChunkNominal);
+dml_get_per_surface_var_func(refcyc_per_meta_chunk_nom_c_in_us, dml_float_t, mode_lib->mp.TimePerChromaMetaChunkNominal);
+dml_get_per_surface_var_func(refcyc_per_meta_chunk_vblank_l_in_us, dml_float_t, mode_lib->mp.TimePerMetaChunkVBlank);
+dml_get_per_surface_var_func(refcyc_per_meta_chunk_vblank_c_in_us, dml_float_t, mode_lib->mp.TimePerChromaMetaChunkVBlank);
+dml_get_per_surface_var_func(refcyc_per_meta_chunk_flip_l_in_us, dml_float_t, mode_lib->mp.TimePerMetaChunkFlip);
+dml_get_per_surface_var_func(refcyc_per_meta_chunk_flip_c_in_us, dml_float_t, mode_lib->mp.TimePerChromaMetaChunkFlip);
+dml_get_per_surface_var_func(refcyc_per_pte_group_nom_l_in_us, dml_float_t, mode_lib->mp.time_per_pte_group_nom_luma);
+dml_get_per_surface_var_func(refcyc_per_pte_group_nom_c_in_us, dml_float_t, mode_lib->mp.time_per_pte_group_nom_chroma);
+dml_get_per_surface_var_func(refcyc_per_pte_group_vblank_l_in_us, dml_float_t, mode_lib->mp.time_per_pte_group_vblank_luma);
+dml_get_per_surface_var_func(refcyc_per_pte_group_vblank_c_in_us, dml_float_t, mode_lib->mp.time_per_pte_group_vblank_chroma);
+dml_get_per_surface_var_func(refcyc_per_pte_group_flip_l_in_us, dml_float_t, mode_lib->mp.time_per_pte_group_flip_luma);
+dml_get_per_surface_var_func(refcyc_per_pte_group_flip_c_in_us, dml_float_t, mode_lib->mp.time_per_pte_group_flip_chroma);
+dml_get_per_surface_var_func(dpte_group_size_in_bytes, dml_uint_t, mode_lib->mp.dpte_group_bytes);
+dml_get_per_surface_var_func(vm_group_size_in_bytes, dml_uint_t, mode_lib->mp.vm_group_bytes);
+dml_get_per_surface_var_func(swath_height_l, dml_uint_t, mode_lib->ms.SwathHeightY);
+dml_get_per_surface_var_func(swath_height_c, dml_uint_t, mode_lib->ms.SwathHeightC);
+dml_get_per_surface_var_func(dpte_row_height_l, dml_uint_t, mode_lib->mp.dpte_row_height);
+dml_get_per_surface_var_func(dpte_row_height_c, dml_uint_t, mode_lib->mp.dpte_row_height_chroma);
+dml_get_per_surface_var_func(dpte_row_height_linear_l, dml_uint_t, mode_lib->mp.dpte_row_height_linear);
+dml_get_per_surface_var_func(dpte_row_height_linear_c, dml_uint_t, mode_lib->mp.dpte_row_height_linear_chroma);
+dml_get_per_surface_var_func(meta_row_height_l, dml_uint_t, mode_lib->mp.meta_row_height);
+dml_get_per_surface_var_func(meta_row_height_c, dml_uint_t, mode_lib->mp.meta_row_height_chroma);
+
+dml_get_per_surface_var_func(vstartup_calculated, dml_uint_t, mode_lib->mp.VStartup);
+dml_get_per_surface_var_func(vupdate_offset, dml_uint_t, mode_lib->mp.VUpdateOffsetPix);
+dml_get_per_surface_var_func(vupdate_width, dml_uint_t, mode_lib->mp.VUpdateWidthPix);
+dml_get_per_surface_var_func(vready_offset, dml_uint_t, mode_lib->mp.VReadyOffsetPix);
+dml_get_per_surface_var_func(vready_at_or_after_vsync, dml_uint_t, mode_lib->mp.VREADY_AT_OR_AFTER_VSYNC);
+dml_get_per_surface_var_func(min_dst_y_next_start, dml_uint_t, mode_lib->mp.MIN_DST_Y_NEXT_START);
+dml_get_per_surface_var_func(det_stored_buffer_size_l_bytes, dml_uint_t, mode_lib->ms.DETBufferSizeY);
+dml_get_per_surface_var_func(det_stored_buffer_size_c_bytes, dml_uint_t, mode_lib->ms.DETBufferSizeC);
+dml_get_per_surface_var_func(use_mall_for_static_screen, dml_uint_t, mode_lib->mp.UsesMALLForStaticScreen);
+dml_get_per_surface_var_func(surface_size_for_mall, dml_uint_t, mode_lib->mp.SurfaceSizeInTheMALL);
+dml_get_per_surface_var_func(dcc_max_uncompressed_block_l, dml_uint_t, mode_lib->mp.DCCYMaxUncompressedBlock);
+dml_get_per_surface_var_func(dcc_max_compressed_block_l, dml_uint_t, mode_lib->mp.DCCYMaxCompressedBlock);
+dml_get_per_surface_var_func(dcc_independent_block_l, dml_uint_t, mode_lib->mp.DCCYIndependentBlock);
+dml_get_per_surface_var_func(dcc_max_uncompressed_block_c, dml_uint_t, mode_lib->mp.DCCCMaxUncompressedBlock);
+dml_get_per_surface_var_func(dcc_max_compressed_block_c, dml_uint_t, mode_lib->mp.DCCCMaxCompressedBlock);
+dml_get_per_surface_var_func(dcc_independent_block_c, dml_uint_t, mode_lib->mp.DCCCIndependentBlock);
+dml_get_per_surface_var_func(max_active_dram_clock_change_latency_supported, dml_uint_t, mode_lib->mp.MaxActiveDRAMClockChangeLatencySupported);
+dml_get_per_surface_var_func(pte_buffer_mode, dml_uint_t, mode_lib->mp.PTE_BUFFER_MODE);
+dml_get_per_surface_var_func(bigk_fragment_size, dml_uint_t, mode_lib->mp.BIGK_FRAGMENT_SIZE);
+dml_get_per_surface_var_func(dpte_bytes_per_row, dml_uint_t, mode_lib->mp.PixelPTEBytesPerRow);
+dml_get_per_surface_var_func(meta_bytes_per_row, dml_uint_t, mode_lib->mp.MetaRowByte);
+dml_get_per_surface_var_func(det_buffer_size_kbytes, dml_uint_t, mode_lib->ms.DETBufferSizeInKByte);
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.h b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.h
new file mode 100644 (file)
index 0000000..2a05458
--- /dev/null
@@ -0,0 +1,199 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DISPLAY_MODE_CORE_H__
+#define __DISPLAY_MODE_CORE_H__
+
+#include "display_mode_core_structs.h"
+
+struct display_mode_lib_st;
+
+dml_bool_t dml_core_mode_support(struct display_mode_lib_st *mode_lib);
+void dml_core_mode_support_partial(struct display_mode_lib_st *mode_lib);
+void dml_core_mode_programming(struct display_mode_lib_st *mode_lib, const struct dml_clk_cfg_st *clk_cfg);
+
+void dml_core_get_row_heights(
+                                               dml_uint_t *dpte_row_height,
+                                               dml_uint_t *meta_row_height,
+                                               const struct display_mode_lib_st *mode_lib,
+                                               dml_bool_t is_plane1,
+                                               enum dml_source_format_class SourcePixelFormat,
+                                               enum dml_swizzle_mode SurfaceTiling,
+                                               enum dml_rotation_angle ScanDirection,
+                                               dml_uint_t pitch,
+                                               dml_uint_t GPUVMMinPageSizeKBytes);
+
+dml_float_t dml_get_return_bw_mbps_vm_only(
+                                                                       const struct soc_bounding_box_st *soc,
+                                                                       dml_bool_t use_ideal_dram_bw_strobe,
+                                                                       dml_bool_t HostVMEnable,
+                                                                       dml_float_t DCFCLK,
+                                                                       dml_float_t FabricClock,
+                                                                       dml_float_t DRAMSpeed);
+
+dml_float_t dml_get_return_bw_mbps(
+                                                       const struct soc_bounding_box_st *soc,
+                                                       dml_bool_t use_ideal_dram_bw_strobe,
+                                                       dml_bool_t HostVMEnable,
+                                                       dml_float_t DCFCLK,
+                                                       dml_float_t FabricClock,
+                                                       dml_float_t DRAMSpeed);
+
+dml_bool_t dml_mode_support(
+       struct display_mode_lib_st *mode_lib,
+       dml_uint_t state_idx,
+       const struct dml_display_cfg_st *display_cfg);
+
+dml_bool_t dml_mode_programming(
+       struct display_mode_lib_st *mode_lib,
+       dml_uint_t state_idx,
+       const struct dml_display_cfg_st *display_cfg,
+       bool call_standalone);
+
+dml_uint_t dml_mode_support_ex(
+       struct dml_mode_support_ex_params_st *in_out_params);
+
+dml_bool_t dml_get_is_phantom_pipe(struct display_mode_lib_st *mode_lib, dml_uint_t pipe_idx);
+
+#define dml_get_per_surface_var_decl(variable, type) type dml_get_##variable(struct display_mode_lib_st *mode_lib, dml_uint_t surface_idx)
+#define dml_get_var_decl(var, type) type dml_get_##var(struct display_mode_lib_st *mode_lib)
+
+dml_get_var_decl(wm_urgent, dml_float_t);
+dml_get_var_decl(wm_stutter_exit, dml_float_t);
+dml_get_var_decl(wm_stutter_enter_exit, dml_float_t);
+dml_get_var_decl(wm_memory_trip, dml_float_t);
+dml_get_var_decl(wm_dram_clock_change, dml_float_t);
+dml_get_var_decl(urgent_latency, dml_float_t);
+dml_get_var_decl(clk_dcf_deepsleep, dml_float_t);
+dml_get_var_decl(wm_fclk_change, dml_float_t);
+dml_get_var_decl(wm_usr_retraining, dml_float_t);
+dml_get_var_decl(urgent_latency, dml_float_t);
+
+dml_get_var_decl(wm_writeback_dram_clock_change, dml_float_t);
+dml_get_var_decl(stutter_efficiency_no_vblank, dml_float_t);
+dml_get_var_decl(stutter_efficiency, dml_float_t);
+dml_get_var_decl(stutter_efficiency_z8, dml_float_t);
+dml_get_var_decl(stutter_num_bursts_z8, dml_float_t);
+dml_get_var_decl(stutter_period, dml_float_t);
+dml_get_var_decl(stutter_efficiency_z8_bestcase, dml_float_t);
+dml_get_var_decl(stutter_num_bursts_z8_bestcase, dml_float_t);
+dml_get_var_decl(stutter_period_bestcase, dml_float_t);
+dml_get_var_decl(urgent_latency, dml_float_t);
+dml_get_var_decl(urgent_extra_latency, dml_float_t);
+dml_get_var_decl(nonurgent_latency, dml_float_t);
+dml_get_var_decl(dispclk_calculated, dml_float_t);
+dml_get_var_decl(total_data_read_bw, dml_float_t);
+dml_get_var_decl(return_bw, dml_float_t);
+dml_get_var_decl(tcalc, dml_float_t);
+dml_get_var_decl(fraction_of_urgent_bandwidth, dml_float_t);
+dml_get_var_decl(fraction_of_urgent_bandwidth_imm_flip, dml_float_t);
+dml_get_var_decl(comp_buffer_size_kbytes, dml_uint_t);
+dml_get_var_decl(pixel_chunk_size_in_kbyte, dml_uint_t);
+dml_get_var_decl(alpha_pixel_chunk_size_in_kbyte, dml_uint_t);
+dml_get_var_decl(meta_chunk_size_in_kbyte, dml_uint_t);
+dml_get_var_decl(min_pixel_chunk_size_in_byte, dml_uint_t);
+dml_get_var_decl(min_meta_chunk_size_in_byte, dml_uint_t);
+dml_get_var_decl(total_immediate_flip_bytes, dml_uint_t);
+
+dml_get_per_surface_var_decl(dsc_delay, dml_uint_t);
+dml_get_per_surface_var_decl(dppclk_calculated, dml_float_t);
+dml_get_per_surface_var_decl(dscclk_calculated, dml_float_t);
+dml_get_per_surface_var_decl(min_ttu_vblank_in_us, dml_float_t);
+dml_get_per_surface_var_decl(vratio_prefetch_l, dml_float_t);
+dml_get_per_surface_var_decl(vratio_prefetch_c, dml_float_t);
+dml_get_per_surface_var_decl(dst_x_after_scaler, dml_uint_t);
+dml_get_per_surface_var_decl(dst_y_after_scaler, dml_uint_t);
+dml_get_per_surface_var_decl(dst_y_per_vm_vblank, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_per_row_vblank, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_prefetch, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_per_vm_flip, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_per_row_flip, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_per_pte_row_nom_l, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_per_pte_row_nom_c, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_per_meta_row_nom_l, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_per_meta_row_nom_c, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_vm_group_vblank_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_vm_group_flip_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_vm_req_vblank_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_vm_req_flip_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_vm_dmdata_in_us, dml_float_t);
+dml_get_per_surface_var_decl(dmdata_dl_delta_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_line_delivery_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_line_delivery_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_line_delivery_pre_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_line_delivery_pre_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_req_delivery_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_req_delivery_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_req_delivery_pre_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_req_delivery_pre_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_cursor_req_delivery_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_cursor_req_delivery_pre_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_meta_chunk_nom_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_meta_chunk_nom_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_meta_chunk_vblank_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_meta_chunk_vblank_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_meta_chunk_flip_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_meta_chunk_flip_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_pte_group_nom_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_pte_group_nom_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_pte_group_vblank_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_pte_group_vblank_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_pte_group_flip_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_pte_group_flip_c_in_us, dml_float_t);
+
+dml_get_per_surface_var_decl(dpte_group_size_in_bytes, dml_uint_t);
+dml_get_per_surface_var_decl(vm_group_size_in_bytes, dml_uint_t);
+dml_get_per_surface_var_decl(swath_height_l, dml_uint_t);
+dml_get_per_surface_var_decl(swath_height_c, dml_uint_t);
+dml_get_per_surface_var_decl(dpte_row_height_l, dml_uint_t);
+dml_get_per_surface_var_decl(dpte_row_height_c, dml_uint_t);
+dml_get_per_surface_var_decl(dpte_row_height_linear_l, dml_uint_t);
+dml_get_per_surface_var_decl(dpte_row_height_linear_c, dml_uint_t);
+dml_get_per_surface_var_decl(meta_row_height_l, dml_uint_t);
+dml_get_per_surface_var_decl(meta_row_height_c, dml_uint_t);
+dml_get_per_surface_var_decl(vstartup_calculated, dml_uint_t);
+dml_get_per_surface_var_decl(vupdate_offset, dml_uint_t);
+dml_get_per_surface_var_decl(vupdate_width, dml_uint_t);
+dml_get_per_surface_var_decl(vready_offset, dml_uint_t);
+dml_get_per_surface_var_decl(vready_at_or_after_vsync, dml_uint_t);
+dml_get_per_surface_var_decl(min_dst_y_next_start, dml_uint_t);
+dml_get_per_surface_var_decl(det_stored_buffer_size_l_bytes, dml_uint_t);
+dml_get_per_surface_var_decl(det_stored_buffer_size_c_bytes, dml_uint_t);
+dml_get_per_surface_var_decl(use_mall_for_static_screen, dml_uint_t);
+dml_get_per_surface_var_decl(surface_size_for_mall, dml_uint_t);
+dml_get_per_surface_var_decl(dcc_max_uncompressed_block_l, dml_uint_t);
+dml_get_per_surface_var_decl(dcc_max_uncompressed_block_c, dml_uint_t);
+dml_get_per_surface_var_decl(dcc_max_compressed_block_l, dml_uint_t);
+dml_get_per_surface_var_decl(dcc_max_compressed_block_c, dml_uint_t);
+dml_get_per_surface_var_decl(dcc_independent_block_l, dml_uint_t);
+dml_get_per_surface_var_decl(dcc_independent_block_c, dml_uint_t);
+dml_get_per_surface_var_decl(max_active_dram_clock_change_latency_supported, dml_uint_t);
+dml_get_per_surface_var_decl(pte_buffer_mode, dml_uint_t);
+dml_get_per_surface_var_decl(bigk_fragment_size, dml_uint_t);
+dml_get_per_surface_var_decl(dpte_bytes_per_row, dml_uint_t);
+dml_get_per_surface_var_decl(meta_bytes_per_row, dml_uint_t);
+dml_get_per_surface_var_decl(det_buffer_size_kbytes, dml_uint_t);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core_structs.h b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core_structs.h
new file mode 100644 (file)
index 0000000..d2e1510
--- /dev/null
@@ -0,0 +1,1968 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DISPLAY_MODE_CORE_STRUCT_H__
+#define __DISPLAY_MODE_CORE_STRUCT_H__
+
+#include "display_mode_lib_defines.h"
+
+enum dml_project_id {
+       dml_project_invalid = 0,
+       dml_project_default = 1,
+       dml_project_dcn32 = dml_project_default,
+       dml_project_dcn321 = 2,
+};
+enum dml_prefetch_modes {
+       dml_prefetch_support_uclk_fclk_and_stutter_if_possible = 0,
+       dml_prefetch_support_uclk_fclk_and_stutter = 1,
+       dml_prefetch_support_fclk_and_stutter = 2,
+       dml_prefetch_support_stutter = 3,
+       dml_prefetch_support_none = 4
+};
+enum dml_use_mall_for_pstate_change_mode {
+       dml_use_mall_pstate_change_disable = 0,
+       dml_use_mall_pstate_change_full_frame = 1,
+       dml_use_mall_pstate_change_sub_viewport = 2,
+       dml_use_mall_pstate_change_phantom_pipe = 3
+};
+enum dml_use_mall_for_static_screen_mode {
+       dml_use_mall_static_screen_disable = 0,
+       dml_use_mall_static_screen_enable = 1,
+       dml_use_mall_static_screen_optimize = 2
+};
+enum dml_output_encoder_class {
+       dml_dp = 0,
+       dml_edp = 1,
+       dml_dp2p0 = 2,
+       dml_hdmi = 3,
+       dml_hdmifrl = 4,
+       dml_none = 5
+};
+enum dml_output_link_dp_rate{
+       dml_dp_rate_na = 0,
+       dml_dp_rate_hbr = 1,
+       dml_dp_rate_hbr2 = 2,
+       dml_dp_rate_hbr3 = 3,
+       dml_dp_rate_uhbr10 = 4,
+       dml_dp_rate_uhbr13p5 = 5,
+       dml_dp_rate_uhbr20 = 6
+};
+enum dml_output_type_and_rate__type{
+       dml_output_type_unknown = 0,
+       dml_output_type_dp = 1,
+       dml_output_type_edp = 2,
+       dml_output_type_dp2p0 = 3,
+       dml_output_type_hdmi = 4,
+       dml_output_type_hdmifrl = 5
+};
+enum dml_output_type_and_rate__rate {
+       dml_output_rate_unknown = 0,
+       dml_output_rate_dp_rate_hbr = 1,
+       dml_output_rate_dp_rate_hbr2 = 2,
+       dml_output_rate_dp_rate_hbr3 = 3,
+       dml_output_rate_dp_rate_uhbr10 = 4,
+       dml_output_rate_dp_rate_uhbr13p5 = 5,
+       dml_output_rate_dp_rate_uhbr20 = 6,
+       dml_output_rate_hdmi_rate_3x3 = 7,
+       dml_output_rate_hdmi_rate_6x3 = 8,
+       dml_output_rate_hdmi_rate_6x4 = 9,
+       dml_output_rate_hdmi_rate_8x4 = 10,
+       dml_output_rate_hdmi_rate_10x4 = 11,
+       dml_output_rate_hdmi_rate_12x4 = 12
+};
+enum dml_output_format_class {
+       dml_444 = 0,
+       dml_s422 = 1,
+       dml_n422 = 2,
+       dml_420 = 3
+};
+enum dml_source_format_class {
+       dml_444_8 = 0,
+       dml_444_16 = 1,
+       dml_444_32 = 2,
+       dml_444_64 = 3,
+       dml_420_8 = 4,
+       dml_420_10 = 5,
+       dml_420_12 = 6,
+       dml_422_8 = 7,
+       dml_422_10 = 8,
+       dml_rgbe_alpha = 9,
+       dml_rgbe = 10,
+       dml_mono_8 = 11,
+       dml_mono_16 = 12
+};
+enum dml_output_bpc_class {
+       dml_out_6 = 0,
+       dml_out_8 = 1,
+       dml_out_10 = 2,
+       dml_out_12 = 3,
+       dml_out_16 = 4
+};
+enum dml_output_standard_class {
+       dml_std_cvt = 0,
+       dml_std_cea = 1,
+       dml_std_cvtr2 = 2
+};
+enum dml_rotation_angle {
+       dml_rotation_0 = 0,
+       dml_rotation_90 = 1,
+       dml_rotation_180 = 2,
+       dml_rotation_270 = 3,
+       dml_rotation_0m = 4,
+       dml_rotation_90m = 5,
+       dml_rotation_180m = 6,
+       dml_rotation_270m = 7
+};
+enum dml_swizzle_mode {
+       dml_sw_linear = 0,
+       dml_sw_256b_s = 1,
+       dml_sw_256b_d = 2,
+       dml_sw_256b_r = 3,
+       dml_sw_4kb_z = 4,
+       dml_sw_4kb_s = 5,
+       dml_sw_4kb_d = 6,
+       dml_sw_4kb_r = 7,
+       dml_sw_64kb_z = 8,
+       dml_sw_64kb_s = 9,
+       dml_sw_64kb_d = 10,
+       dml_sw_64kb_r = 11,
+       dml_sw_256kb_z = 12,
+       dml_sw_256kb_s = 13,
+       dml_sw_256kb_d = 14,
+       dml_sw_256kb_r = 15,
+       dml_sw_64kb_z_t = 16,
+       dml_sw_64kb_s_t = 17,
+       dml_sw_64kb_d_t = 18,
+       dml_sw_64kb_r_t = 19,
+       dml_sw_4kb_z_x = 20,
+       dml_sw_4kb_s_x = 21,
+       dml_sw_4kb_d_x = 22,
+       dml_sw_4kb_r_x = 23,
+       dml_sw_64kb_z_x = 24,
+       dml_sw_64kb_s_x = 25,
+       dml_sw_64kb_d_x = 26,
+       dml_sw_64kb_r_x = 27,
+       dml_sw_256kb_z_x = 28,
+       dml_sw_256kb_s_x = 29,
+       dml_sw_256kb_d_x = 30,
+       dml_sw_256kb_r_x = 31
+};
+enum dml_lb_depth {
+       dml_lb_6 = 0,
+       dml_lb_8 = 1,
+       dml_lb_10 = 2,
+       dml_lb_12 = 3,
+       dml_lb_16 = 4
+};
+enum dml_voltage_state {
+       dml_vmin_lv = 0,
+       dml_vmin = 1,
+       dml_vmid = 2,
+       dml_vnom = 3,
+       dml_vmax = 4
+};
+enum dml_source_macro_tile_size {
+       dml_4k_tile = 0,
+       dml_64k_tile = 1,
+       dml_256k_tile = 2
+};
+enum dml_cursor_bpp {
+       dml_cur_2bit = 0,
+       dml_cur_32bit = 1,
+       dml_cur_64bit = 2
+};
+enum dml_dram_clock_change_support {
+       dml_dram_clock_change_vactive = 0,
+       dml_dram_clock_change_vblank = 1,
+       dml_dram_clock_change_vblank_drr = 2,
+       dml_dram_clock_change_vactive_w_mall_full_frame = 3,
+       dml_dram_clock_change_vactive_w_mall_sub_vp = 4,
+       dml_dram_clock_change_vblank_w_mall_full_frame = 5,
+       dml_dram_clock_change_vblank_drr_w_mall_full_frame = 6,
+       dml_dram_clock_change_vblank_w_mall_sub_vp = 7,
+       dml_dram_clock_change_vblank_drr_w_mall_sub_vp = 8,
+       dml_dram_clock_change_unsupported = 9
+};
+enum dml_fclock_change_support {
+       dml_fclock_change_vactive = 0,
+       dml_fclock_change_vblank = 1,
+       dml_fclock_change_unsupported = 2
+};
+enum dml_dsc_enable {
+       dml_dsc_disable = 0,
+       dml_dsc_enable = 1,
+       dml_dsc_enable_if_necessary = 2
+};
+enum dml_mpc_use_policy {
+       dml_mpc_disabled = 0,
+       dml_mpc_as_possible = 1,
+       dml_mpc_as_needed_for_voltage = 2,
+       dml_mpc_as_needed_for_pstate_and_voltage = 3
+};
+enum dml_odm_use_policy {
+       dml_odm_use_policy_bypass = 0,
+       dml_odm_use_policy_combine_as_needed = 1,
+       dml_odm_use_policy_combine_2to1 = 2,
+       dml_odm_use_policy_combine_4to1 = 3,
+       dml_odm_use_policy_split_1to2 = 4,
+       dml_odm_use_policy_mso_1to2 = 5,
+       dml_odm_use_policy_mso_1to4 = 6
+};
+enum dml_odm_mode {
+       dml_odm_mode_bypass = 0,
+       dml_odm_mode_combine_2to1 = 1,
+       dml_odm_mode_combine_4to1 = 2,
+       dml_odm_mode_split_1to2 = 3,
+       dml_odm_mode_mso_1to2 = 4,
+       dml_odm_mode_mso_1to4 = 5
+};
+enum dml_writeback_configuration {
+       dml_whole_buffer_for_single_stream_no_interleave = 0,
+       dml_whole_buffer_for_single_stream_interleave = 1
+};
+enum dml_immediate_flip_requirement {
+       dml_immediate_flip_not_required = 0,
+       dml_immediate_flip_required = 1,
+       dml_immediate_flip_if_possible = 2
+};
+enum dml_unbounded_requesting_policy {
+       dml_unbounded_requesting_enable = 0,
+       dml_unbounded_requesting_edp_only = 1,
+       dml_unbounded_requesting_disable = 2
+};
+enum dml_clk_cfg_policy {
+       dml_use_required_freq = 0,
+       dml_use_override_freq = 1,
+       dml_use_state_freq = 2
+};
+
+
+struct soc_state_bounding_box_st {
+       dml_float_t socclk_mhz;
+       dml_float_t dscclk_mhz;
+       dml_float_t phyclk_mhz;
+       dml_float_t phyclk_d18_mhz;
+       dml_float_t phyclk_d32_mhz;
+       dml_float_t dtbclk_mhz;
+       dml_float_t fabricclk_mhz;
+       dml_float_t dcfclk_mhz;
+       dml_float_t dispclk_mhz;
+       dml_float_t dppclk_mhz;
+       dml_float_t dram_speed_mts;
+       dml_float_t urgent_latency_pixel_data_only_us;
+       dml_float_t urgent_latency_pixel_mixed_with_vm_data_us;
+       dml_float_t urgent_latency_vm_data_only_us;
+       dml_float_t writeback_latency_us;
+       dml_float_t urgent_latency_adjustment_fabric_clock_component_us;
+       dml_float_t urgent_latency_adjustment_fabric_clock_reference_mhz;
+       dml_float_t sr_exit_time_us;
+       dml_float_t sr_enter_plus_exit_time_us;
+       dml_float_t sr_exit_z8_time_us;
+       dml_float_t sr_enter_plus_exit_z8_time_us;
+       dml_float_t dram_clock_change_latency_us;
+       dml_float_t fclk_change_latency_us;
+       dml_float_t usr_retraining_latency_us;
+       dml_bool_t use_ideal_dram_bw_strobe;
+};
+
+struct soc_bounding_box_st {
+       dml_float_t dprefclk_mhz;
+       dml_float_t xtalclk_mhz;
+       dml_float_t pcierefclk_mhz;
+       dml_float_t refclk_mhz;
+       dml_float_t amclk_mhz;
+       dml_float_t max_outstanding_reqs;
+       dml_float_t pct_ideal_sdp_bw_after_urgent;
+       dml_float_t pct_ideal_fabric_bw_after_urgent;
+       dml_float_t pct_ideal_dram_bw_after_urgent_pixel_only;
+       dml_float_t pct_ideal_dram_bw_after_urgent_pixel_and_vm;
+       dml_float_t pct_ideal_dram_bw_after_urgent_vm_only;
+       dml_float_t pct_ideal_dram_bw_after_urgent_strobe;
+       dml_float_t max_avg_sdp_bw_use_normal_percent;
+       dml_float_t max_avg_fabric_bw_use_normal_percent;
+       dml_float_t max_avg_dram_bw_use_normal_percent;
+       dml_float_t max_avg_dram_bw_use_normal_strobe_percent;
+       dml_uint_t round_trip_ping_latency_dcfclk_cycles;
+       dml_uint_t urgent_out_of_order_return_per_channel_pixel_only_bytes;
+       dml_uint_t urgent_out_of_order_return_per_channel_pixel_and_vm_bytes;
+       dml_uint_t urgent_out_of_order_return_per_channel_vm_only_bytes;
+       dml_uint_t num_chans;
+       dml_uint_t return_bus_width_bytes;
+       dml_uint_t dram_channel_width_bytes;
+       dml_uint_t fabric_datapath_to_dcn_data_return_bytes;
+       dml_uint_t hostvm_min_page_size_kbytes;
+       dml_uint_t gpuvm_min_page_size_kbytes;
+       dml_float_t phy_downspread_percent;
+       dml_float_t dcn_downspread_percent;
+       dml_float_t smn_latency_us;
+       dml_uint_t mall_allocated_for_dcn_mbytes;
+       dml_float_t dispclk_dppclk_vco_speed_mhz;
+       dml_bool_t do_urgent_latency_adjustment;
+};
+
+struct ip_params_st {
+       dml_uint_t vblank_nom_default_us;
+       dml_uint_t rob_buffer_size_kbytes;
+       dml_uint_t config_return_buffer_size_in_kbytes;
+       dml_uint_t config_return_buffer_segment_size_in_kbytes;
+       dml_uint_t compressed_buffer_segment_size_in_kbytes;
+       dml_uint_t meta_fifo_size_in_kentries;
+       dml_uint_t zero_size_buffer_entries;
+       dml_uint_t dpte_buffer_size_in_pte_reqs_luma;
+       dml_uint_t dpte_buffer_size_in_pte_reqs_chroma;
+       dml_uint_t dcc_meta_buffer_size_bytes;
+       dml_bool_t gpuvm_enable;
+       dml_bool_t hostvm_enable;
+       dml_uint_t gpuvm_max_page_table_levels;
+       dml_uint_t hostvm_max_page_table_levels;
+       dml_uint_t pixel_chunk_size_kbytes;
+       dml_uint_t alpha_pixel_chunk_size_kbytes;
+       dml_uint_t min_pixel_chunk_size_bytes;
+       dml_uint_t meta_chunk_size_kbytes;
+       dml_uint_t min_meta_chunk_size_bytes;
+       dml_uint_t writeback_chunk_size_kbytes;
+       dml_uint_t line_buffer_size_bits;
+       dml_uint_t max_line_buffer_lines;
+       dml_uint_t writeback_interface_buffer_size_kbytes;
+       dml_uint_t max_num_dpp;
+       dml_uint_t max_num_otg;
+       dml_uint_t max_num_wb;
+       dml_uint_t max_dchub_pscl_bw_pix_per_clk;
+       dml_uint_t max_pscl_lb_bw_pix_per_clk;
+       dml_uint_t max_lb_vscl_bw_pix_per_clk;
+       dml_uint_t max_vscl_hscl_bw_pix_per_clk;
+       dml_float_t max_hscl_ratio;
+       dml_float_t max_vscl_ratio;
+       dml_uint_t max_hscl_taps;
+       dml_uint_t max_vscl_taps;
+       dml_uint_t num_dsc;
+       dml_uint_t maximum_dsc_bits_per_component;
+       dml_uint_t maximum_pixels_per_line_per_dsc_unit;
+       dml_bool_t dsc422_native_support;
+       dml_bool_t cursor_64bpp_support;
+       dml_float_t dispclk_ramp_margin_percent;
+       dml_uint_t dppclk_delay_subtotal;
+       dml_uint_t dppclk_delay_scl;
+       dml_uint_t dppclk_delay_scl_lb_only;
+       dml_uint_t dppclk_delay_cnvc_formatter;
+       dml_uint_t dppclk_delay_cnvc_cursor;
+       dml_uint_t cursor_buffer_size;
+       dml_uint_t cursor_chunk_size;
+       dml_uint_t dispclk_delay_subtotal;
+       dml_bool_t dynamic_metadata_vm_enabled;
+       dml_uint_t max_inter_dcn_tile_repeaters;
+       dml_uint_t max_num_hdmi_frl_outputs;
+       dml_uint_t max_num_dp2p0_outputs;
+       dml_uint_t max_num_dp2p0_streams;
+       dml_bool_t dcc_supported;
+       dml_bool_t ptoi_supported;
+       dml_float_t writeback_max_hscl_ratio;
+       dml_float_t writeback_max_vscl_ratio;
+       dml_float_t writeback_min_hscl_ratio;
+       dml_float_t writeback_min_vscl_ratio;
+       dml_uint_t writeback_max_hscl_taps;
+       dml_uint_t writeback_max_vscl_taps;
+       dml_uint_t writeback_line_buffer_buffer_size;
+};
+
+struct DmlPipe {
+       dml_float_t Dppclk;
+       dml_float_t Dispclk;
+       dml_float_t PixelClock;
+       dml_float_t DCFClkDeepSleep;
+       dml_uint_t DPPPerSurface;
+       dml_bool_t ScalerEnabled;
+       enum dml_rotation_angle SourceScan;
+       dml_uint_t ViewportHeight;
+       dml_uint_t ViewportHeightChroma;
+       dml_uint_t BlockWidth256BytesY;
+       dml_uint_t BlockHeight256BytesY;
+       dml_uint_t BlockWidth256BytesC;
+       dml_uint_t BlockHeight256BytesC;
+       dml_uint_t BlockWidthY;
+       dml_uint_t BlockHeightY;
+       dml_uint_t BlockWidthC;
+       dml_uint_t BlockHeightC;
+       dml_uint_t InterlaceEnable;
+       dml_uint_t NumberOfCursors;
+       dml_uint_t VBlank;
+       dml_uint_t HTotal;
+       dml_uint_t HActive;
+       dml_bool_t DCCEnable;
+       enum dml_odm_mode ODMMode;
+       enum dml_source_format_class SourcePixelFormat;
+       enum dml_swizzle_mode SurfaceTiling;
+       dml_uint_t BytePerPixelY;
+       dml_uint_t BytePerPixelC;
+       dml_bool_t ProgressiveToInterlaceUnitInOPP;
+       dml_float_t VRatio;
+       dml_float_t VRatioChroma;
+       dml_uint_t VTaps;
+       dml_uint_t VTapsChroma;
+       dml_uint_t PitchY;
+       dml_uint_t DCCMetaPitchY;
+       dml_uint_t PitchC;
+       dml_uint_t DCCMetaPitchC;
+       dml_bool_t ViewportStationary;
+       dml_uint_t ViewportXStart;
+       dml_uint_t ViewportYStart;
+       dml_uint_t ViewportXStartC;
+       dml_uint_t ViewportYStartC;
+       dml_bool_t FORCE_ONE_ROW_FOR_FRAME;
+       dml_uint_t SwathHeightY;
+       dml_uint_t SwathHeightC;
+};
+
+struct Watermarks {
+       dml_float_t UrgentWatermark;
+       dml_float_t WritebackUrgentWatermark;
+       dml_float_t DRAMClockChangeWatermark;
+       dml_float_t FCLKChangeWatermark;
+       dml_float_t WritebackDRAMClockChangeWatermark;
+       dml_float_t WritebackFCLKChangeWatermark;
+       dml_float_t StutterExitWatermark;
+       dml_float_t StutterEnterPlusExitWatermark;
+       dml_float_t Z8StutterExitWatermark;
+       dml_float_t Z8StutterEnterPlusExitWatermark;
+       dml_float_t USRRetrainingWatermark;
+};
+
+struct SOCParametersList {
+       dml_float_t UrgentLatency;
+       dml_float_t ExtraLatency;
+       dml_float_t WritebackLatency;
+       dml_float_t DRAMClockChangeLatency;
+       dml_float_t FCLKChangeLatency;
+       dml_float_t SRExitTime;
+       dml_float_t SREnterPlusExitTime;
+       dml_float_t SRExitZ8Time;
+       dml_float_t SREnterPlusExitZ8Time;
+       dml_float_t USRRetrainingLatency;
+       dml_float_t SMNLatency;
+};
+
+/// @brief Struct that represent Plane configration of a display cfg
+struct dml_plane_cfg_st {
+       //
+       // Pipe/Surface Parameters
+       //
+       dml_bool_t GPUVMEnable; /// <brief Set if any pipe has GPUVM enable
+       dml_bool_t HostVMEnable; /// <brief Set if any pipe has HostVM enable
+
+       dml_uint_t GPUVMMaxPageTableLevels; /// <brief GPUVM level; max of all pipes'
+       dml_uint_t HostVMMaxPageTableLevels; /// <brief HostVM level; max of all pipes'; that is the number of non-cache HVM level
+
+       dml_uint_t GPUVMMinPageSizeKBytes[__DML_NUM_PLANES__];
+       dml_bool_t ForceOneRowForFrame[__DML_NUM_PLANES__];
+       dml_bool_t PTEBufferModeOverrideEn[__DML_NUM_PLANES__]; //< brief when override enable; the DML will only check the given pte buffer and will use the pte buffer mode as is
+       dml_bool_t PTEBufferMode[__DML_NUM_PLANES__];
+       dml_uint_t ViewportWidth[__DML_NUM_PLANES__];
+       dml_uint_t ViewportHeight[__DML_NUM_PLANES__];
+       dml_uint_t ViewportWidthChroma[__DML_NUM_PLANES__];
+       dml_uint_t ViewportHeightChroma[__DML_NUM_PLANES__];
+       dml_uint_t ViewportXStart[__DML_NUM_PLANES__];
+       dml_uint_t ViewportXStartC[__DML_NUM_PLANES__];
+       dml_uint_t ViewportYStart[__DML_NUM_PLANES__];
+       dml_uint_t ViewportYStartC[__DML_NUM_PLANES__];
+       dml_bool_t ViewportStationary[__DML_NUM_PLANES__];
+
+       dml_bool_t ScalerEnabled[__DML_NUM_PLANES__];
+       dml_float_t HRatio[__DML_NUM_PLANES__];
+       dml_float_t VRatio[__DML_NUM_PLANES__];
+       dml_float_t HRatioChroma[__DML_NUM_PLANES__];
+       dml_float_t VRatioChroma[__DML_NUM_PLANES__];
+       dml_uint_t HTaps[__DML_NUM_PLANES__];
+       dml_uint_t VTaps[__DML_NUM_PLANES__];
+       dml_uint_t HTapsChroma[__DML_NUM_PLANES__];
+       dml_uint_t VTapsChroma[__DML_NUM_PLANES__];
+       dml_uint_t LBBitPerPixel[__DML_NUM_PLANES__];
+
+       enum dml_rotation_angle SourceScan[__DML_NUM_PLANES__];
+       dml_uint_t ScalerRecoutWidth[__DML_NUM_PLANES__];
+
+       dml_bool_t DynamicMetadataEnable[__DML_NUM_PLANES__];
+       dml_uint_t DynamicMetadataLinesBeforeActiveRequired[__DML_NUM_PLANES__];
+       dml_uint_t DynamicMetadataTransmittedBytes[__DML_NUM_PLANES__];
+       dml_uint_t DETSizeOverride[__DML_NUM_PLANES__]; /// <brief user can specify the desire DET buffer usage per-plane
+
+       dml_uint_t NumberOfCursors[__DML_NUM_PLANES__];
+       dml_uint_t CursorWidth[__DML_NUM_PLANES__];
+       dml_uint_t CursorBPP[__DML_NUM_PLANES__];
+
+       enum dml_use_mall_for_static_screen_mode UseMALLForStaticScreen[__DML_NUM_PLANES__];
+       enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[__DML_NUM_PLANES__];
+
+       dml_uint_t BlendingAndTiming[__DML_NUM_PLANES__]; /// <brief From which timing group (like OTG) that this plane is getting its timing from. Mode check also need this info for example to check num OTG; encoder; dsc etc.
+}; // dml_plane_cfg_st;
+
+/// @brief Surface Parameters
+struct dml_surface_cfg_st {
+       enum dml_swizzle_mode SurfaceTiling[__DML_NUM_PLANES__];
+       enum dml_source_format_class SourcePixelFormat[__DML_NUM_PLANES__];
+       dml_uint_t PitchY[__DML_NUM_PLANES__];
+       dml_uint_t SurfaceWidthY[__DML_NUM_PLANES__];
+       dml_uint_t SurfaceHeightY[__DML_NUM_PLANES__];
+       dml_uint_t PitchC[__DML_NUM_PLANES__];
+       dml_uint_t SurfaceWidthC[__DML_NUM_PLANES__];
+       dml_uint_t SurfaceHeightC[__DML_NUM_PLANES__];
+
+       dml_bool_t DCCEnable[__DML_NUM_PLANES__];
+       dml_uint_t DCCMetaPitchY[__DML_NUM_PLANES__];
+       dml_uint_t DCCMetaPitchC[__DML_NUM_PLANES__];
+
+       dml_float_t DCCRateLuma[__DML_NUM_PLANES__];
+       dml_float_t DCCRateChroma[__DML_NUM_PLANES__];
+       dml_float_t DCCFractionOfZeroSizeRequestsLuma[__DML_NUM_PLANES__];
+       dml_float_t DCCFractionOfZeroSizeRequestsChroma[__DML_NUM_PLANES__];
+}; // dml_surface_cfg_st
+
+/// @brief structure that represents the timing configuration
+struct dml_timing_cfg_st {
+       dml_uint_t HTotal[__DML_NUM_PLANES__];
+       dml_uint_t VTotal[__DML_NUM_PLANES__];
+       dml_uint_t HBlankEnd[__DML_NUM_PLANES__];
+       dml_uint_t VBlankEnd[__DML_NUM_PLANES__];
+       dml_uint_t RefreshRate[__DML_NUM_PLANES__];
+       dml_uint_t VFrontPorch[__DML_NUM_PLANES__];
+       dml_float_t PixelClock[__DML_NUM_PLANES__];
+       dml_uint_t HActive[__DML_NUM_PLANES__];
+       dml_uint_t VActive[__DML_NUM_PLANES__];
+       dml_bool_t Interlace[__DML_NUM_PLANES__];
+       dml_bool_t DRRDisplay[__DML_NUM_PLANES__];
+       dml_uint_t VBlankNom[__DML_NUM_PLANES__];
+}; // dml_timing_cfg_st;
+
+/// @brief structure that represents the output stream
+struct dml_output_cfg_st {
+       // Output Setting
+       dml_uint_t DSCInputBitPerComponent[__DML_NUM_PLANES__];
+       enum dml_output_format_class OutputFormat[__DML_NUM_PLANES__];
+       enum dml_output_encoder_class OutputEncoder[__DML_NUM_PLANES__];
+       dml_uint_t OutputMultistreamId[__DML_NUM_PLANES__];
+       dml_bool_t OutputMultistreamEn[__DML_NUM_PLANES__];
+       dml_float_t OutputBpp[__DML_NUM_PLANES__]; //< brief Use by mode_programming to specify a output bpp; user can use the output from mode_support (support.OutputBpp)
+       dml_float_t PixelClockBackEnd[__DML_NUM_PLANES__];
+       enum dml_dsc_enable DSCEnable[__DML_NUM_PLANES__]; //< brief for mode support check; use to determine if dsc is required
+       dml_uint_t OutputLinkDPLanes[__DML_NUM_PLANES__];
+       enum dml_output_link_dp_rate OutputLinkDPRate[__DML_NUM_PLANES__];
+       dml_float_t ForcedOutputLinkBPP[__DML_NUM_PLANES__];
+       dml_uint_t AudioSampleRate[__DML_NUM_PLANES__];
+       dml_uint_t AudioSampleLayout[__DML_NUM_PLANES__];
+       dml_bool_t OutputDisabled[__DML_NUM_PLANES__];
+}; // dml_timing_cfg_st;
+
+/// @brief Writeback Setting
+struct dml_writeback_cfg_st {
+       enum dml_source_format_class WritebackPixelFormat[__DML_NUM_PLANES__];
+       dml_bool_t WritebackEnable[__DML_NUM_PLANES__];
+       dml_uint_t ActiveWritebacksPerSurface[__DML_NUM_PLANES__];
+       dml_uint_t WritebackDestinationWidth[__DML_NUM_PLANES__];
+       dml_uint_t WritebackDestinationHeight[__DML_NUM_PLANES__];
+       dml_uint_t WritebackSourceWidth[__DML_NUM_PLANES__];
+       dml_uint_t WritebackSourceHeight[__DML_NUM_PLANES__];
+       dml_uint_t WritebackHTaps[__DML_NUM_PLANES__];
+       dml_uint_t WritebackVTaps[__DML_NUM_PLANES__];
+       dml_float_t WritebackHRatio[__DML_NUM_PLANES__];
+       dml_float_t WritebackVRatio[__DML_NUM_PLANES__];
+}; // dml_writeback_cfg_st;
+
+/// @brief Hardware resource specific; mainly used by mode_programming when test/sw wants to do some specific setting
+///        which are not the same as what the mode support stage derive.  When call mode_support with mode_programm; the hw-specific
+//         resource will be set to what the mode_support layer recommends
+struct dml_hw_resource_st {
+       enum dml_odm_mode ODMMode[__DML_NUM_PLANES__]; /// <brief ODM mode that is chosen in the mode check stage and will be used in mode programming stage
+       dml_uint_t DPPPerSurface[__DML_NUM_PLANES__]; /// <brief How many DPPs are needed drive the surface to output. If MPCC or ODMC could be 2 or 4.
+       dml_bool_t DSCEnabled[__DML_NUM_PLANES__]; /// <brief Indicate if the DSC is enabled; used in mode_programming
+       dml_uint_t NumberOfDSCSlices[__DML_NUM_PLANES__]; /// <brief Indicate how many slices needed to support the given mode
+       dml_float_t DLGRefClkFreqMHz; /// <brief DLG Global Reference timer
+};
+
+/// @brief DML display configuration.
+///        Describe how to display a surface in multi-plane setup and output to different output and writeback using the specified timgin
+struct dml_display_cfg_st {
+       struct dml_surface_cfg_st surface;
+       struct dml_plane_cfg_st plane;
+       struct dml_timing_cfg_st timing;
+       struct dml_output_cfg_st output;
+       struct dml_writeback_cfg_st writeback;
+       unsigned int num_surfaces;
+       unsigned int num_timings;
+
+       struct dml_hw_resource_st hw; //< brief for mode programming
+}; // dml_display_cfg_st
+
+/// @brief To control the clk usage for model programming
+struct dml_clk_cfg_st {
+       enum dml_clk_cfg_policy dcfclk_option; ///< brief Use for mode_program; user can select between use the min require clk req as calculated by DML or use the test-specific freq
+       enum dml_clk_cfg_policy dispclk_option; ///< brief Use for mode_program; user can select between use the min require clk req as calculated by DML or use the test-specific freq
+       enum dml_clk_cfg_policy dppclk_option[__DML_NUM_PLANES__];
+
+       dml_float_t dcfclk_freq_mhz;
+       dml_float_t dispclk_freq_mhz;
+       dml_float_t dppclk_freq_mhz[__DML_NUM_PLANES__];
+}; // dml_clk_cfg_st
+
+/// @brief DML mode evaluation and programming policy
+/// Those knobs that affect mode support and mode programming
+struct dml_mode_eval_policy_st {
+       // -------------------
+       // Policy
+       // -------------------
+       enum dml_mpc_use_policy MPCCombineUse[__DML_NUM_PLANES__]; /// <brief MPC Combine mode as selected by the user; used in mode check stage
+       enum dml_odm_use_policy ODMUse[__DML_NUM_PLANES__]; /// <brief ODM mode as selected by the user; used in mode check stage
+       enum dml_unbounded_requesting_policy UseUnboundedRequesting; ///< brief Unbounded request mode preference
+       enum dml_immediate_flip_requirement ImmediateFlipRequirement[__DML_NUM_PLANES__]; /// <brief Is immediate flip a requirement for this plane. When host vm is present iflip is needed regardless
+       enum dml_prefetch_modes AllowForPStateChangeOrStutterInVBlank[__DML_NUM_PLANES__]; /// <brief To specify if the DML should calculate the values for support different pwr saving features (cstate; pstate; etc.) during vblank
+
+       enum dml_prefetch_modes AllowForPStateChangeOrStutterInVBlankFinal;
+       bool UseOnlyMaxPrefetchModes;
+       dml_bool_t UseMinimumRequiredDCFCLK; //<brief When set the mode_check stage will figure the min DCFCLK freq to support the given display configuration. User can tell use the output DCFCLK for mode programming.
+       dml_bool_t DRAMClockChangeRequirementFinal;
+       dml_bool_t FCLKChangeRequirementFinal;
+       dml_bool_t USRRetrainingRequiredFinal;
+       dml_bool_t EnhancedPrefetchScheduleAccelerationFinal;
+
+       dml_bool_t NomDETInKByteOverrideEnable; //<brief Nomimal DET buffer size for a pipe. If this size fit the required 2 swathes; DML will use this DET size
+       dml_uint_t NomDETInKByteOverrideValue;
+
+       dml_bool_t DCCProgrammingAssumesScanDirectionUnknownFinal;
+       dml_bool_t SynchronizeTimingsFinal;
+       dml_bool_t SynchronizeDRRDisplaysForUCLKPStateChangeFinal;
+       dml_bool_t AssumeModeSupportAtMaxPwrStateEvenDRAMClockChangeNotSupported; //<brief if set; the mode support will say mode is supported even though the DRAM clock change is not support (assuming the soc will be stay in max power state)
+       dml_bool_t AssumeModeSupportAtMaxPwrStateEvenFClockChangeNotSupported; //<brief if set; the mode support will say mode is supported even though the Fabric clock change is not support (assuming the soc will be stay in max power state
+};
+
+/// @brief Contains important information after the mode support steps. Also why a mode is not supported.
+struct dml_mode_support_info_st {
+       //-----------------
+       // Mode Support Information
+       //-----------------
+       dml_bool_t ModeIsSupported; //<brief Is the mode support any voltage and combine setting
+       dml_bool_t ImmediateFlipSupport; //<brief Means mode support immediate flip at the max combine setting; determine in mode support and used in mode programming
+       dml_uint_t MaximumMPCCombine; //<brief If using MPC combine helps the power saving support; then this will be set to 1
+       dml_bool_t UnboundedRequestEnabled;
+       dml_uint_t CompressedBufferSizeInkByte;
+
+       /* Mode Support Reason */
+       dml_bool_t WritebackLatencySupport;
+       dml_bool_t ScaleRatioAndTapsSupport;
+       dml_bool_t SourceFormatPixelAndScanSupport;
+       dml_bool_t MPCCombineMethodIncompatible;
+       dml_bool_t P2IWith420;
+       dml_bool_t DSCOnlyIfNecessaryWithBPP;
+       dml_bool_t DSC422NativeNotSupported;
+       dml_bool_t LinkRateDoesNotMatchDPVersion;
+       dml_bool_t LinkRateForMultistreamNotIndicated;
+       dml_bool_t BPPForMultistreamNotIndicated;
+       dml_bool_t MultistreamWithHDMIOreDP;
+       dml_bool_t MSOOrODMSplitWithNonDPLink;
+       dml_bool_t NotEnoughLanesForMSO;
+       dml_bool_t NumberOfOTGSupport;
+       dml_bool_t NumberOfHDMIFRLSupport;
+       dml_bool_t NumberOfDP2p0Support;
+       dml_bool_t NonsupportedDSCInputBPC;
+       dml_bool_t WritebackScaleRatioAndTapsSupport;
+       dml_bool_t CursorSupport;
+       dml_bool_t PitchSupport;
+       dml_bool_t ViewportExceedsSurface;
+       dml_bool_t ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified;
+       dml_bool_t ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe;
+       dml_bool_t InvalidCombinationOfMALLUseForPStateAndStaticScreen;
+       dml_bool_t InvalidCombinationOfMALLUseForPState;
+       dml_bool_t ExceededMALLSize;
+       dml_bool_t EnoughWritebackUnits;
+
+       dml_bool_t ExceededMultistreamSlots;
+       dml_bool_t ODMCombineTwoToOneSupportCheckOK;
+       dml_bool_t ODMCombineFourToOneSupportCheckOK;
+       dml_bool_t NotEnoughDSCUnits;
+       dml_bool_t NotEnoughDSCSlices;
+       dml_bool_t PixelsPerLinePerDSCUnitSupport;
+       dml_bool_t DSCCLKRequiredMoreThanSupported;
+       dml_bool_t DTBCLKRequiredMoreThanSupported;
+       dml_bool_t LinkCapacitySupport;
+
+       dml_bool_t ROBSupport[2];
+       dml_bool_t PTEBufferSizeNotExceeded[2];
+       dml_bool_t DCCMetaBufferSizeNotExceeded[2];
+       dml_bool_t TotalVerticalActiveBandwidthSupport[2];
+       enum dml_dram_clock_change_support DRAMClockChangeSupport[2];
+       dml_float_t ActiveDRAMClockChangeLatencyMargin[__DML_NUM_PLANES__];
+       dml_uint_t SubViewportLinesNeededInMALL[__DML_NUM_PLANES__];
+       enum dml_fclock_change_support FCLKChangeSupport[2];
+       dml_bool_t USRRetrainingSupport[2];
+       dml_bool_t VActiveBandwithSupport[2];
+       dml_bool_t PrefetchSupported[2];
+       dml_bool_t DynamicMetadataSupported[2];
+       dml_bool_t VRatioInPrefetchSupported[2];
+       dml_bool_t DISPCLK_DPPCLK_Support[2];
+       dml_bool_t TotalAvailablePipesSupport[2];
+       dml_bool_t ModeSupport[2];
+       dml_bool_t ViewportSizeSupport[2];
+       dml_bool_t ImmediateFlipSupportedForState[2];
+
+       dml_bool_t NoTimeForPrefetch[2][__DML_NUM_PLANES__];
+       dml_bool_t NoTimeForDynamicMetadata[2][__DML_NUM_PLANES__];
+
+       dml_bool_t MPCCombineEnable[__DML_NUM_PLANES__]; /// <brief Indicate if the MPC Combine enable in the given state and optimize mpc combine setting
+       enum dml_odm_mode ODMMode[__DML_NUM_PLANES__]; /// <brief ODM mode that is chosen in the mode check stage and will be used in mode programming stage
+       dml_uint_t DPPPerSurface[__DML_NUM_PLANES__]; /// <brief How many DPPs are needed drive the surface to output. If MPCC or ODMC could be 2 or 4.
+       dml_bool_t DSCEnabled[__DML_NUM_PLANES__]; /// <brief Indicate if the DSC is actually required; used in mode_programming
+       dml_bool_t FECEnabled[__DML_NUM_PLANES__]; /// <brief Indicate if the FEC is actually required
+       dml_uint_t NumberOfDSCSlices[__DML_NUM_PLANES__]; /// <brief Indicate how many slices needed to support the given mode
+
+       dml_float_t OutputBpp[__DML_NUM_PLANES__];
+       enum dml_output_type_and_rate__type OutputType[__DML_NUM_PLANES__];
+       enum dml_output_type_and_rate__rate OutputRate[__DML_NUM_PLANES__];
+
+       dml_float_t AlignedDCCMetaPitchY[__DML_NUM_PLANES__]; /// <brief Pitch value that is aligned to tiling setting
+       dml_float_t AlignedDCCMetaPitchC[__DML_NUM_PLANES__];
+       dml_float_t AlignedYPitch[__DML_NUM_PLANES__];
+       dml_float_t AlignedCPitch[__DML_NUM_PLANES__];
+       dml_float_t MaxTotalVerticalActiveAvailableBandwidth[2]; /// <brief nominal bw available for display
+}; // dml_mode_support_info_st
+
+/// @brief Treat this as the intermediate values and outputs of mode check function. User can query the content of the struct to know more about the result of mode evaluation.
+struct mode_support_st {
+       struct ip_params_st ip;
+       struct soc_bounding_box_st soc;
+       struct soc_state_bounding_box_st state; //<brief Per-state bbox values; only 1 state per compute
+       struct dml_mode_eval_policy_st policy;
+
+       dml_uint_t state_idx; //<brief The power state idx for the power state under this computation
+       dml_uint_t max_state_idx; //<brief The MAX power state idx
+       struct soc_state_bounding_box_st max_state; //<brief The MAX power state; some algo needs to know the max state info to determine if
+       struct dml_display_cfg_st cache_display_cfg; // <brief A copy of the current display cfg in consideration
+
+       // Physical info; only using for programming
+       dml_uint_t num_active_planes; // <brief As determined by either e2e_pipe_param or display_cfg
+
+       // Calculated Clocks
+       dml_float_t RequiredDISPCLK[2]; /// <brief Required DISPCLK; depends on pixel rate; odm mode etc.
+       dml_float_t RequiredDPPCLKThisState[__DML_NUM_PLANES__];
+       dml_float_t DCFCLKState[2]; /// <brief recommended DCFCLK freq; calculated by DML. If UseMinimumRequiredDCFCLK is not set; then it will be just the state DCFCLK; else it will min DCFCLK for support
+       dml_float_t RequiredDISPCLKPerSurface[2][__DML_NUM_PLANES__];
+       dml_float_t RequiredDPPCLKPerSurface[2][__DML_NUM_PLANES__];
+
+       dml_float_t FabricClock; /// <brief Basically just the clock freq at the min (or given) state
+       dml_float_t DRAMSpeed; /// <brief Basically just the clock freq at the min (or given) state
+       dml_float_t SOCCLK; /// <brief Basically just the clock freq at the min (or given) state
+       dml_float_t DCFCLK; /// <brief Basically just the clock freq at the min (or given) state and max combine setting
+       dml_float_t GlobalDPPCLK; /// <brief the Max DPPCLK freq out of all pipes
+
+       // ----------------------------------
+       // Mode Support Info and fail reason
+       // ----------------------------------
+       struct dml_mode_support_info_st support;
+
+       // These are calculated before the ModeSupport and ModeProgram step
+       // They represent the bound for the return buffer sizing
+       dml_uint_t MaxTotalDETInKByte;
+       dml_uint_t NomDETInKByte;
+       dml_uint_t MinCompressedBufferSizeInKByte;
+
+       // Info obtained at the end of mode support calculations
+       // The reported info is at the "optimal" state and combine setting
+       dml_float_t ReturnBW;
+       dml_float_t ReturnDRAMBW;
+       dml_uint_t DETBufferSizeInKByte[__DML_NUM_PLANES__]; // <brief Recommended DET size configuration for this plane. All pipes under this plane should program the DET buffer size to the calculated value.
+       dml_uint_t DETBufferSizeY[__DML_NUM_PLANES__];
+       dml_uint_t DETBufferSizeC[__DML_NUM_PLANES__];
+       dml_uint_t SwathHeightY[__DML_NUM_PLANES__];
+       dml_uint_t SwathHeightC[__DML_NUM_PLANES__];
+
+       // ----------------------------------
+       // Intermediates/Informational
+       // ----------------------------------
+       dml_uint_t TotImmediateFlipBytes;
+       dml_bool_t DCCEnabledInAnySurface;
+       dml_float_t WritebackRequiredDISPCLK;
+       dml_float_t TimeCalc;
+       dml_float_t TWait;
+
+       dml_uint_t SwathWidthYAllStates[2][__DML_NUM_PLANES__];
+       dml_uint_t SwathWidthCAllStates[2][__DML_NUM_PLANES__];
+       dml_uint_t SwathHeightYAllStates[2][__DML_NUM_PLANES__];
+       dml_uint_t SwathHeightCAllStates[2][__DML_NUM_PLANES__];
+       dml_uint_t SwathWidthYThisState[__DML_NUM_PLANES__];
+       dml_uint_t SwathWidthCThisState[__DML_NUM_PLANES__];
+       dml_uint_t SwathHeightYThisState[__DML_NUM_PLANES__];
+       dml_uint_t SwathHeightCThisState[__DML_NUM_PLANES__];
+       dml_uint_t DETBufferSizeInKByteAllStates[2][__DML_NUM_PLANES__];
+       dml_uint_t DETBufferSizeYAllStates[2][__DML_NUM_PLANES__];
+       dml_uint_t DETBufferSizeCAllStates[2][__DML_NUM_PLANES__];
+       dml_bool_t UnboundedRequestEnabledAllStates[2];
+       dml_uint_t CompressedBufferSizeInkByteAllStates[2];
+       dml_bool_t UnboundedRequestEnabledThisState;
+       dml_uint_t CompressedBufferSizeInkByteThisState;
+       dml_uint_t DETBufferSizeInKByteThisState[__DML_NUM_PLANES__];
+       dml_uint_t DETBufferSizeYThisState[__DML_NUM_PLANES__];
+       dml_uint_t DETBufferSizeCThisState[__DML_NUM_PLANES__];
+       dml_float_t VRatioPreY[2][__DML_NUM_PLANES__];
+       dml_float_t VRatioPreC[2][__DML_NUM_PLANES__];
+       dml_uint_t swath_width_luma_ub_all_states[2][__DML_NUM_PLANES__];
+       dml_uint_t swath_width_chroma_ub_all_states[2][__DML_NUM_PLANES__];
+       dml_uint_t swath_width_luma_ub_this_state[__DML_NUM_PLANES__];
+       dml_uint_t swath_width_chroma_ub_this_state[__DML_NUM_PLANES__];
+       dml_uint_t RequiredSlots[__DML_NUM_PLANES__];
+       dml_uint_t PDEAndMetaPTEBytesPerFrame[2][__DML_NUM_PLANES__];
+       dml_uint_t MetaRowBytes[2][__DML_NUM_PLANES__];
+       dml_uint_t DPTEBytesPerRow[2][__DML_NUM_PLANES__];
+       dml_uint_t PrefetchLinesY[2][__DML_NUM_PLANES__];
+       dml_uint_t PrefetchLinesC[2][__DML_NUM_PLANES__];
+       dml_uint_t MaxNumSwY[__DML_NUM_PLANES__]; /// <brief Max number of swath for prefetch
+       dml_uint_t MaxNumSwC[__DML_NUM_PLANES__]; /// <brief Max number of swath for prefetch
+       dml_uint_t PrefillY[__DML_NUM_PLANES__];
+       dml_uint_t PrefillC[__DML_NUM_PLANES__];
+
+       dml_uint_t PrefetchLinesYThisState[__DML_NUM_PLANES__];
+       dml_uint_t PrefetchLinesCThisState[__DML_NUM_PLANES__];
+       dml_uint_t DPTEBytesPerRowThisState[__DML_NUM_PLANES__];
+       dml_uint_t PDEAndMetaPTEBytesPerFrameThisState[__DML_NUM_PLANES__];
+       dml_uint_t MetaRowBytesThisState[__DML_NUM_PLANES__];
+       dml_bool_t use_one_row_for_frame[2][__DML_NUM_PLANES__];
+       dml_bool_t use_one_row_for_frame_flip[2][__DML_NUM_PLANES__];
+       dml_bool_t use_one_row_for_frame_this_state[__DML_NUM_PLANES__];
+       dml_bool_t use_one_row_for_frame_flip_this_state[__DML_NUM_PLANES__];
+
+       dml_float_t LineTimesForPrefetch[__DML_NUM_PLANES__];
+       dml_float_t LinesForMetaPTE[__DML_NUM_PLANES__];
+       dml_float_t LinesForMetaAndDPTERow[__DML_NUM_PLANES__];
+       dml_float_t SwathWidthYSingleDPP[__DML_NUM_PLANES__];
+       dml_float_t SwathWidthCSingleDPP[__DML_NUM_PLANES__];
+       dml_uint_t BytePerPixelY[__DML_NUM_PLANES__];
+       dml_uint_t BytePerPixelC[__DML_NUM_PLANES__];
+       dml_float_t BytePerPixelInDETY[__DML_NUM_PLANES__];
+       dml_float_t BytePerPixelInDETC[__DML_NUM_PLANES__];
+
+       dml_uint_t Read256BlockHeightY[__DML_NUM_PLANES__];
+       dml_uint_t Read256BlockWidthY[__DML_NUM_PLANES__];
+       dml_uint_t Read256BlockHeightC[__DML_NUM_PLANES__];
+       dml_uint_t Read256BlockWidthC[__DML_NUM_PLANES__];
+       dml_uint_t MacroTileHeightY[__DML_NUM_PLANES__];
+       dml_uint_t MacroTileHeightC[__DML_NUM_PLANES__];
+       dml_uint_t MacroTileWidthY[__DML_NUM_PLANES__];
+       dml_uint_t MacroTileWidthC[__DML_NUM_PLANES__];
+       dml_float_t PSCL_FACTOR[__DML_NUM_PLANES__];
+       dml_float_t PSCL_FACTOR_CHROMA[__DML_NUM_PLANES__];
+       dml_float_t MaximumSwathWidthLuma[__DML_NUM_PLANES__];
+       dml_float_t MaximumSwathWidthChroma[__DML_NUM_PLANES__];
+       dml_float_t Tno_bw[__DML_NUM_PLANES__];
+       dml_float_t DestinationLinesToRequestVMInImmediateFlip[__DML_NUM_PLANES__];
+       dml_float_t DestinationLinesToRequestRowInImmediateFlip[__DML_NUM_PLANES__];
+       dml_float_t WritebackDelayTime[__DML_NUM_PLANES__];
+       dml_uint_t dpte_group_bytes[__DML_NUM_PLANES__];
+       dml_uint_t dpte_row_height[__DML_NUM_PLANES__];
+       dml_uint_t dpte_row_height_chroma[__DML_NUM_PLANES__];
+       dml_uint_t meta_row_height[__DML_NUM_PLANES__];
+       dml_uint_t meta_row_height_chroma[__DML_NUM_PLANES__];
+       dml_float_t UrgLatency;
+       dml_float_t UrgentBurstFactorCursor[__DML_NUM_PLANES__];
+       dml_float_t UrgentBurstFactorCursorPre[__DML_NUM_PLANES__];
+       dml_float_t UrgentBurstFactorLuma[__DML_NUM_PLANES__];
+       dml_float_t UrgentBurstFactorLumaPre[__DML_NUM_PLANES__];
+       dml_float_t UrgentBurstFactorChroma[__DML_NUM_PLANES__];
+       dml_float_t UrgentBurstFactorChromaPre[__DML_NUM_PLANES__];
+       dml_float_t MaximumSwathWidthInLineBufferLuma;
+       dml_float_t MaximumSwathWidthInLineBufferChroma;
+       dml_float_t ExtraLatency;
+
+       // Backend
+       dml_bool_t RequiresDSC[__DML_NUM_PLANES__];
+       dml_bool_t RequiresFEC[__DML_NUM_PLANES__];
+       dml_float_t OutputBppPerState[__DML_NUM_PLANES__];
+       dml_uint_t DSCDelayPerState[__DML_NUM_PLANES__];
+       enum dml_output_type_and_rate__type OutputTypePerState[__DML_NUM_PLANES__];
+       enum dml_output_type_and_rate__rate OutputRatePerState[__DML_NUM_PLANES__];
+
+       // Bandwidth Related Info
+       dml_float_t BandwidthAvailableForImmediateFlip;
+       dml_float_t ReadBandwidthLuma[__DML_NUM_PLANES__];
+       dml_float_t ReadBandwidthChroma[__DML_NUM_PLANES__];
+       dml_float_t WriteBandwidth[__DML_NUM_PLANES__];
+       dml_float_t RequiredPrefetchPixelDataBWLuma[__DML_NUM_PLANES__];
+       dml_float_t RequiredPrefetchPixelDataBWChroma[__DML_NUM_PLANES__];
+       dml_float_t cursor_bw[__DML_NUM_PLANES__];
+       dml_float_t cursor_bw_pre[__DML_NUM_PLANES__];
+       dml_float_t prefetch_vmrow_bw[__DML_NUM_PLANES__];
+       dml_float_t final_flip_bw[__DML_NUM_PLANES__];
+       dml_float_t meta_row_bandwidth_this_state[__DML_NUM_PLANES__];
+       dml_float_t dpte_row_bandwidth_this_state[__DML_NUM_PLANES__];
+       dml_float_t ReturnBWPerState[2];
+       dml_float_t ReturnDRAMBWPerState[2];
+       dml_float_t meta_row_bandwidth[2][__DML_NUM_PLANES__];
+       dml_float_t dpte_row_bandwidth[2][__DML_NUM_PLANES__];
+
+       // Something that should be feedback to caller
+       enum dml_odm_mode ODMModePerState[__DML_NUM_PLANES__];
+       enum dml_odm_mode ODMModeThisState[__DML_NUM_PLANES__];
+       dml_uint_t SurfaceSizeInMALL[__DML_NUM_PLANES__];
+       dml_uint_t NoOfDPP[2][__DML_NUM_PLANES__];
+       dml_uint_t NoOfDPPThisState[__DML_NUM_PLANES__];
+       dml_bool_t MPCCombine[2][__DML_NUM_PLANES__];
+       dml_bool_t MPCCombineThisState[__DML_NUM_PLANES__];
+       dml_float_t ProjectedDCFCLKDeepSleep[2];
+       dml_float_t MinDPPCLKUsingSingleDPP[__DML_NUM_PLANES__];
+       dml_bool_t SingleDPPViewportSizeSupportPerSurface[__DML_NUM_PLANES__];
+       dml_bool_t ImmediateFlipSupportedForPipe[__DML_NUM_PLANES__];
+       dml_bool_t NotUrgentLatencyHiding[__DML_NUM_PLANES__];
+       dml_bool_t NotUrgentLatencyHidingPre[__DML_NUM_PLANES__];
+       dml_bool_t PTEBufferSizeNotExceededPerState[__DML_NUM_PLANES__];
+       dml_bool_t DCCMetaBufferSizeNotExceededPerState[__DML_NUM_PLANES__];
+       dml_uint_t PrefetchMode[__DML_NUM_PLANES__];
+       dml_uint_t TotalNumberOfActiveDPP[2];
+       dml_uint_t TotalNumberOfSingleDPPSurfaces[2];
+       dml_uint_t TotalNumberOfDCCActiveDPP[2];
+
+       dml_uint_t SubViewportLinesNeededInMALL[__DML_NUM_PLANES__];
+
+}; // mode_support_st
+
+/// @brief A mega structure that houses various info for model programming step.
+struct mode_program_st {
+
+       //-------------
+       // Intermediate/Informational
+       //-------------
+       dml_float_t UrgentLatency;
+       dml_float_t UrgentLatencyWithUSRRetraining;
+       dml_uint_t VInitPreFillY[__DML_NUM_PLANES__];
+       dml_uint_t VInitPreFillC[__DML_NUM_PLANES__];
+       dml_uint_t MaxNumSwathY[__DML_NUM_PLANES__];
+       dml_uint_t MaxNumSwathC[__DML_NUM_PLANES__];
+
+       dml_float_t BytePerPixelDETY[__DML_NUM_PLANES__];
+       dml_float_t BytePerPixelDETC[__DML_NUM_PLANES__];
+       dml_uint_t BytePerPixelY[__DML_NUM_PLANES__];
+       dml_uint_t BytePerPixelC[__DML_NUM_PLANES__];
+       dml_uint_t SwathWidthY[__DML_NUM_PLANES__];
+       dml_uint_t SwathWidthC[__DML_NUM_PLANES__];
+       dml_uint_t SwathWidthSingleDPPY[__DML_NUM_PLANES__];
+       dml_uint_t SwathWidthSingleDPPC[__DML_NUM_PLANES__];
+       dml_float_t ReadBandwidthSurfaceLuma[__DML_NUM_PLANES__];
+       dml_float_t ReadBandwidthSurfaceChroma[__DML_NUM_PLANES__];
+
+       dml_uint_t PixelPTEBytesPerRow[__DML_NUM_PLANES__];
+       dml_uint_t PDEAndMetaPTEBytesFrame[__DML_NUM_PLANES__];
+       dml_uint_t MetaRowByte[__DML_NUM_PLANES__];
+       dml_uint_t PrefetchSourceLinesY[__DML_NUM_PLANES__];
+       dml_float_t RequiredPrefetchPixDataBWLuma[__DML_NUM_PLANES__];
+       dml_float_t RequiredPrefetchPixDataBWChroma[__DML_NUM_PLANES__];
+       dml_uint_t PrefetchSourceLinesC[__DML_NUM_PLANES__];
+       dml_float_t PSCL_THROUGHPUT[__DML_NUM_PLANES__];
+       dml_float_t PSCL_THROUGHPUT_CHROMA[__DML_NUM_PLANES__];
+       dml_uint_t DSCDelay[__DML_NUM_PLANES__];
+       dml_float_t DPPCLKUsingSingleDPP[__DML_NUM_PLANES__];
+
+       dml_uint_t MacroTileWidthY[__DML_NUM_PLANES__];
+       dml_uint_t MacroTileWidthC[__DML_NUM_PLANES__];
+       dml_uint_t BlockHeight256BytesY[__DML_NUM_PLANES__];
+       dml_uint_t BlockHeight256BytesC[__DML_NUM_PLANES__];
+       dml_uint_t BlockWidth256BytesY[__DML_NUM_PLANES__];
+       dml_uint_t BlockWidth256BytesC[__DML_NUM_PLANES__];
+
+       dml_uint_t BlockHeightY[__DML_NUM_PLANES__];
+       dml_uint_t BlockHeightC[__DML_NUM_PLANES__];
+       dml_uint_t BlockWidthY[__DML_NUM_PLANES__];
+       dml_uint_t BlockWidthC[__DML_NUM_PLANES__];
+
+       dml_uint_t SurfaceSizeInTheMALL[__DML_NUM_PLANES__];
+       dml_float_t VRatioPrefetchY[__DML_NUM_PLANES__];
+       dml_float_t VRatioPrefetchC[__DML_NUM_PLANES__];
+       dml_float_t Tno_bw[__DML_NUM_PLANES__];
+       dml_float_t final_flip_bw[__DML_NUM_PLANES__];
+       dml_float_t prefetch_vmrow_bw[__DML_NUM_PLANES__];
+       dml_float_t cursor_bw[__DML_NUM_PLANES__];
+       dml_float_t cursor_bw_pre[__DML_NUM_PLANES__];
+       dml_float_t WritebackDelay[__DML_NUM_PLANES__];
+       dml_uint_t dpte_row_height[__DML_NUM_PLANES__];
+       dml_uint_t dpte_row_height_linear[__DML_NUM_PLANES__];
+       dml_uint_t meta_req_width[__DML_NUM_PLANES__];
+       dml_uint_t meta_req_height[__DML_NUM_PLANES__];
+       dml_uint_t meta_row_width[__DML_NUM_PLANES__];
+       dml_uint_t meta_row_height[__DML_NUM_PLANES__];
+       dml_uint_t dpte_row_width_luma_ub[__DML_NUM_PLANES__];
+       dml_uint_t dpte_row_width_chroma_ub[__DML_NUM_PLANES__];
+       dml_uint_t dpte_row_height_chroma[__DML_NUM_PLANES__];
+       dml_uint_t dpte_row_height_linear_chroma[__DML_NUM_PLANES__];
+       dml_uint_t meta_req_width_chroma[__DML_NUM_PLANES__];
+       dml_uint_t meta_req_height_chroma[__DML_NUM_PLANES__];
+       dml_uint_t meta_row_width_chroma[__DML_NUM_PLANES__];
+       dml_uint_t meta_row_height_chroma[__DML_NUM_PLANES__];
+       dml_uint_t vm_group_bytes[__DML_NUM_PLANES__];
+       dml_uint_t dpte_group_bytes[__DML_NUM_PLANES__];
+       dml_float_t meta_row_bw[__DML_NUM_PLANES__];
+       dml_float_t dpte_row_bw[__DML_NUM_PLANES__];
+       dml_float_t UrgBurstFactorCursor[__DML_NUM_PLANES__];
+       dml_float_t UrgBurstFactorCursorPre[__DML_NUM_PLANES__];
+       dml_float_t UrgBurstFactorLuma[__DML_NUM_PLANES__];
+       dml_float_t UrgBurstFactorLumaPre[__DML_NUM_PLANES__];
+       dml_float_t UrgBurstFactorChroma[__DML_NUM_PLANES__];
+       dml_float_t UrgBurstFactorChromaPre[__DML_NUM_PLANES__];
+
+       dml_uint_t swath_width_luma_ub[__DML_NUM_PLANES__];
+       dml_uint_t swath_width_chroma_ub[__DML_NUM_PLANES__];
+       dml_uint_t PixelPTEReqWidthY[__DML_NUM_PLANES__];
+       dml_uint_t PixelPTEReqHeightY[__DML_NUM_PLANES__];
+       dml_uint_t PTERequestSizeY[__DML_NUM_PLANES__];
+       dml_uint_t PixelPTEReqWidthC[__DML_NUM_PLANES__];
+       dml_uint_t PixelPTEReqHeightC[__DML_NUM_PLANES__];
+       dml_uint_t PTERequestSizeC[__DML_NUM_PLANES__];
+
+       dml_float_t Tdmdl_vm[__DML_NUM_PLANES__];
+       dml_float_t Tdmdl[__DML_NUM_PLANES__];
+       dml_float_t TSetup[__DML_NUM_PLANES__];
+       dml_uint_t dpde0_bytes_per_frame_ub_l[__DML_NUM_PLANES__];
+       dml_uint_t meta_pte_bytes_per_frame_ub_l[__DML_NUM_PLANES__];
+       dml_uint_t dpde0_bytes_per_frame_ub_c[__DML_NUM_PLANES__];
+       dml_uint_t meta_pte_bytes_per_frame_ub_c[__DML_NUM_PLANES__];
+
+       dml_bool_t UnboundedRequestEnabled;
+       dml_uint_t compbuf_reserved_space_64b;
+       dml_uint_t compbuf_reserved_space_zs;
+       dml_uint_t CompressedBufferSizeInkByte;
+
+       dml_bool_t NoUrgentLatencyHiding[__DML_NUM_PLANES__];
+       dml_bool_t NoUrgentLatencyHidingPre[__DML_NUM_PLANES__];
+       dml_float_t UrgentExtraLatency;
+       dml_bool_t PrefetchAndImmediateFlipSupported;
+       dml_float_t TotalDataReadBandwidth;
+       dml_float_t BandwidthAvailableForImmediateFlip;
+       dml_bool_t NotEnoughTimeForDynamicMetadata[__DML_NUM_PLANES__];
+
+       dml_float_t ReadBandwidthLuma[__DML_NUM_PLANES__];
+       dml_float_t ReadBandwidthChroma[__DML_NUM_PLANES__];
+
+       dml_float_t total_dcn_read_bw_with_flip;
+       dml_float_t total_dcn_read_bw_with_flip_no_urgent_burst;
+       dml_float_t TotalDataReadBandwidthNotIncludingMALLPrefetch;
+       dml_float_t total_dcn_read_bw_with_flip_not_including_MALL_prefetch;
+       dml_float_t non_urgent_total_dcn_read_bw_with_flip;
+       dml_float_t non_urgent_total_dcn_read_bw_with_flip_not_including_MALL_prefetch;
+
+       dml_bool_t use_one_row_for_frame[__DML_NUM_PLANES__];
+       dml_bool_t use_one_row_for_frame_flip[__DML_NUM_PLANES__];
+
+       dml_float_t TCalc;
+       dml_uint_t TotImmediateFlipBytes;
+
+       // -------------------
+       // Output
+       // -------------------
+       dml_uint_t pipe_plane[__DML_NUM_PLANES__]; // <brief used mainly by dv to map the pipe inst to plane index within DML core; the plane idx of a pipe
+       dml_uint_t num_active_pipes;
+
+       dml_bool_t NoTimeToPrefetch[__DML_NUM_PLANES__]; /// <brief Prefetch schedule calculation result
+
+       // Support
+       dml_uint_t PrefetchMode[__DML_NUM_PLANES__]; /// <brief prefetch mode used for prefetch support check in mode programming step
+       dml_bool_t PrefetchModeSupported; /// <brief Is the prefetch mode (bandwidth and latency) supported
+       dml_bool_t ImmediateFlipSupported;
+       dml_bool_t ImmediateFlipSupportedForPipe[__DML_NUM_PLANES__];
+
+       // Clock
+       dml_float_t Dcfclk;
+       dml_float_t Dispclk; /// <brief dispclk being used in mode programming
+       dml_float_t Dppclk[__DML_NUM_PLANES__]; /// <brief dppclk being used in mode programming
+       dml_float_t WritebackDISPCLK;
+       dml_float_t GlobalDPPCLK;
+
+       //@ brief These "calculated" dispclk and dppclk clocks are calculated in the mode programming step.
+       // Depends on the dml_clk_cfg_st option; these calculated values may not used in subsequent calculation.
+       // Possible DV usage: Calculated values fetched by test once after mode_programming step and then possibly
+       // use the values as min and adjust the actual freq used for the 2nd pass
+       dml_float_t Dispclk_calculated;
+       dml_float_t Dppclk_calculated[__DML_NUM_PLANES__];
+
+       dml_float_t DSCCLK_calculated[__DML_NUM_PLANES__]; //< brief Required DSCCLK freq. Backend; not used in any subsequent calculations for now
+       dml_float_t DCFCLKDeepSleep;
+
+       // ARB reg
+       dml_bool_t DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE;
+       struct Watermarks Watermark;
+
+       // DCC compression control
+       dml_uint_t DCCYMaxUncompressedBlock[__DML_NUM_PLANES__];
+       dml_uint_t DCCYMaxCompressedBlock[__DML_NUM_PLANES__];
+       dml_uint_t DCCYIndependentBlock[__DML_NUM_PLANES__];
+       dml_uint_t DCCCMaxUncompressedBlock[__DML_NUM_PLANES__];
+       dml_uint_t DCCCMaxCompressedBlock[__DML_NUM_PLANES__];
+       dml_uint_t DCCCIndependentBlock[__DML_NUM_PLANES__];
+
+       // Stutter Efficiency
+       dml_float_t StutterEfficiency;
+       dml_float_t StutterEfficiencyNotIncludingVBlank;
+       dml_uint_t NumberOfStutterBurstsPerFrame;
+       dml_float_t Z8StutterEfficiency;
+       dml_uint_t Z8NumberOfStutterBurstsPerFrame;
+       dml_float_t Z8StutterEfficiencyNotIncludingVBlank;
+       dml_float_t StutterPeriod;
+       dml_float_t Z8StutterEfficiencyBestCase;
+       dml_uint_t Z8NumberOfStutterBurstsPerFrameBestCase;
+       dml_float_t Z8StutterEfficiencyNotIncludingVBlankBestCase;
+       dml_float_t StutterPeriodBestCase;
+
+       // DLG TTU reg
+       dml_float_t MIN_DST_Y_NEXT_START[__DML_NUM_PLANES__];
+       dml_bool_t VREADY_AT_OR_AFTER_VSYNC[__DML_NUM_PLANES__];
+       dml_uint_t DSTYAfterScaler[__DML_NUM_PLANES__];
+       dml_uint_t DSTXAfterScaler[__DML_NUM_PLANES__];
+       dml_float_t DestinationLinesForPrefetch[__DML_NUM_PLANES__];
+       dml_float_t DestinationLinesToRequestVMInVBlank[__DML_NUM_PLANES__];
+       dml_float_t DestinationLinesToRequestRowInVBlank[__DML_NUM_PLANES__];
+       dml_float_t DestinationLinesToRequestVMInImmediateFlip[__DML_NUM_PLANES__];
+       dml_float_t DestinationLinesToRequestRowInImmediateFlip[__DML_NUM_PLANES__];
+       dml_float_t MinTTUVBlank[__DML_NUM_PLANES__];
+       dml_float_t DisplayPipeLineDeliveryTimeLuma[__DML_NUM_PLANES__];
+       dml_float_t DisplayPipeLineDeliveryTimeChroma[__DML_NUM_PLANES__];
+       dml_float_t DisplayPipeLineDeliveryTimeLumaPrefetch[__DML_NUM_PLANES__];
+       dml_float_t DisplayPipeLineDeliveryTimeChromaPrefetch[__DML_NUM_PLANES__];
+       dml_float_t DisplayPipeRequestDeliveryTimeLuma[__DML_NUM_PLANES__];
+       dml_float_t DisplayPipeRequestDeliveryTimeChroma[__DML_NUM_PLANES__];
+       dml_float_t DisplayPipeRequestDeliveryTimeLumaPrefetch[__DML_NUM_PLANES__];
+       dml_float_t DisplayPipeRequestDeliveryTimeChromaPrefetch[__DML_NUM_PLANES__];
+       dml_float_t CursorRequestDeliveryTime[__DML_NUM_PLANES__];
+       dml_float_t CursorRequestDeliveryTimePrefetch[__DML_NUM_PLANES__];
+
+       dml_float_t DST_Y_PER_PTE_ROW_NOM_L[__DML_NUM_PLANES__];
+       dml_float_t DST_Y_PER_PTE_ROW_NOM_C[__DML_NUM_PLANES__];
+       dml_float_t DST_Y_PER_META_ROW_NOM_L[__DML_NUM_PLANES__];
+       dml_float_t DST_Y_PER_META_ROW_NOM_C[__DML_NUM_PLANES__];
+       dml_float_t TimePerMetaChunkNominal[__DML_NUM_PLANES__];
+       dml_float_t TimePerChromaMetaChunkNominal[__DML_NUM_PLANES__];
+       dml_float_t TimePerMetaChunkVBlank[__DML_NUM_PLANES__];
+       dml_float_t TimePerChromaMetaChunkVBlank[__DML_NUM_PLANES__];
+       dml_float_t TimePerMetaChunkFlip[__DML_NUM_PLANES__];
+       dml_float_t TimePerChromaMetaChunkFlip[__DML_NUM_PLANES__];
+       dml_float_t time_per_pte_group_nom_luma[__DML_NUM_PLANES__];
+       dml_float_t time_per_pte_group_nom_chroma[__DML_NUM_PLANES__];
+       dml_float_t time_per_pte_group_vblank_luma[__DML_NUM_PLANES__];
+       dml_float_t time_per_pte_group_vblank_chroma[__DML_NUM_PLANES__];
+       dml_float_t time_per_pte_group_flip_luma[__DML_NUM_PLANES__];
+       dml_float_t time_per_pte_group_flip_chroma[__DML_NUM_PLANES__];
+       dml_float_t TimePerVMGroupVBlank[__DML_NUM_PLANES__];
+       dml_float_t TimePerVMGroupFlip[__DML_NUM_PLANES__];
+       dml_float_t TimePerVMRequestVBlank[__DML_NUM_PLANES__];
+       dml_float_t TimePerVMRequestFlip[__DML_NUM_PLANES__];
+
+       dml_float_t FractionOfUrgentBandwidth;
+       dml_float_t FractionOfUrgentBandwidthImmediateFlip;
+
+       // RQ registers
+       dml_bool_t PTE_BUFFER_MODE[__DML_NUM_PLANES__];
+       dml_uint_t BIGK_FRAGMENT_SIZE[__DML_NUM_PLANES__];
+
+       dml_uint_t SubViewportLinesNeededInMALL[__DML_NUM_PLANES__];
+       dml_bool_t UsesMALLForStaticScreen[__DML_NUM_PLANES__];
+
+       // OTG
+       dml_uint_t VStartupMin[__DML_NUM_PLANES__]; /// <brief Minimum vstartup to meet the prefetch schedule (i.e. the prefetch solution can be found at this vstartup time); not the actual global sync vstartup pos.
+       dml_uint_t VStartup[__DML_NUM_PLANES__]; /// <brief The vstartup value for OTG programming (will set to max vstartup; but now bounded by min(vblank_nom. actual vblank))
+       dml_uint_t VUpdateOffsetPix[__DML_NUM_PLANES__];
+       dml_uint_t VUpdateWidthPix[__DML_NUM_PLANES__];
+       dml_uint_t VReadyOffsetPix[__DML_NUM_PLANES__];
+
+       // Latency and Support
+       dml_float_t MaxActiveFCLKChangeLatencySupported;
+       dml_bool_t USRRetrainingSupport;
+       enum dml_fclock_change_support FCLKChangeSupport;
+       enum dml_dram_clock_change_support DRAMClockChangeSupport;
+       dml_float_t MaxActiveDRAMClockChangeLatencySupported[__DML_NUM_PLANES__];
+       dml_float_t WritebackAllowFCLKChangeEndPosition[__DML_NUM_PLANES__];
+       dml_float_t WritebackAllowDRAMClockChangeEndPosition[__DML_NUM_PLANES__];
+
+       // buffer sizing
+       dml_uint_t DETBufferSizeInKByte[__DML_NUM_PLANES__];  // <brief Recommended DET size configuration for this plane.  All pipes under this plane should program the DET buffer size to the calculated value.
+       dml_uint_t DETBufferSizeY[__DML_NUM_PLANES__];
+       dml_uint_t DETBufferSizeC[__DML_NUM_PLANES__];
+       dml_uint_t SwathHeightY[__DML_NUM_PLANES__];
+       dml_uint_t SwathHeightC[__DML_NUM_PLANES__];
+}; // mode_program_st
+
+struct soc_states_st {
+       dml_uint_t num_states; /// <brief num of soc pwr states
+       struct soc_state_bounding_box_st state_array[__DML_MAX_STATE_ARRAY_SIZE__]; /// <brief fixed size array that holds states struct
+};
+
+struct UseMinimumDCFCLK_params_st {
+       enum dml_use_mall_for_pstate_change_mode *UseMALLForPStateChange;
+       dml_bool_t *DRRDisplay;
+       dml_bool_t SynchronizeDRRDisplaysForUCLKPStateChangeFinal;
+       dml_uint_t MaxInterDCNTileRepeaters;
+       dml_uint_t MaxPrefetchMode;
+       dml_float_t DRAMClockChangeLatencyFinal;
+       dml_float_t FCLKChangeLatency;
+       dml_float_t SREnterPlusExitTime;
+       dml_uint_t ReturnBusWidth;
+       dml_uint_t RoundTripPingLatencyCycles;
+       dml_uint_t ReorderingBytes;
+       dml_uint_t PixelChunkSizeInKByte;
+       dml_uint_t MetaChunkSize;
+       dml_bool_t GPUVMEnable;
+       dml_uint_t GPUVMMaxPageTableLevels;
+       dml_bool_t HostVMEnable;
+       dml_uint_t NumberOfActiveSurfaces;
+       dml_uint_t HostVMMinPageSize;
+       dml_uint_t HostVMMaxNonCachedPageTableLevels;
+       dml_bool_t DynamicMetadataVMEnabled;
+       dml_bool_t ImmediateFlipRequirement;
+       dml_bool_t ProgressiveToInterlaceUnitInOPP;
+       dml_float_t MaxAveragePercentOfIdealSDPPortBWDisplayCanUseInNormalSystemOperation;
+       dml_float_t PercentOfIdealSDPPortBWReceivedAfterUrgLatency;
+       dml_uint_t *VTotal;
+       dml_uint_t *VActive;
+       dml_uint_t *DynamicMetadataTransmittedBytes;
+       dml_uint_t *DynamicMetadataLinesBeforeActiveRequired;
+       dml_bool_t *Interlace;
+       dml_float_t (*RequiredDPPCLKPerSurface)[__DML_NUM_PLANES__];
+       dml_float_t *RequiredDISPCLK;
+       dml_float_t UrgLatency;
+       dml_uint_t (*NoOfDPP)[__DML_NUM_PLANES__];
+       dml_float_t *ProjectedDCFCLKDeepSleep;
+       dml_uint_t (*MaximumVStartup)[__DML_NUM_PLANES__];
+       dml_uint_t *TotalNumberOfActiveDPP;
+       dml_uint_t *TotalNumberOfDCCActiveDPP;
+       dml_uint_t *dpte_group_bytes;
+       dml_uint_t (*PrefetchLinesY)[__DML_NUM_PLANES__];
+       dml_uint_t (*PrefetchLinesC)[__DML_NUM_PLANES__];
+       dml_uint_t (*swath_width_luma_ub_all_states)[__DML_NUM_PLANES__];
+       dml_uint_t (*swath_width_chroma_ub_all_states)[__DML_NUM_PLANES__];
+       dml_uint_t *BytePerPixelY;
+       dml_uint_t *BytePerPixelC;
+       dml_uint_t *HTotal;
+       dml_float_t *PixelClock;
+       dml_uint_t (*PDEAndMetaPTEBytesPerFrame)[__DML_NUM_PLANES__];
+       dml_uint_t (*DPTEBytesPerRow)[__DML_NUM_PLANES__];
+       dml_uint_t (*MetaRowBytes)[__DML_NUM_PLANES__];
+       dml_bool_t *DynamicMetadataEnable;
+       dml_float_t *ReadBandwidthLuma;
+       dml_float_t *ReadBandwidthChroma;
+       dml_float_t DCFCLKPerState;
+       dml_float_t *DCFCLKState;
+};
+
+struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params_st {
+       dml_bool_t USRRetrainingRequiredFinal;
+       enum dml_use_mall_for_pstate_change_mode *UseMALLForPStateChange;
+       dml_uint_t *PrefetchMode;
+       dml_uint_t NumberOfActiveSurfaces;
+       dml_uint_t MaxLineBufferLines;
+       dml_uint_t LineBufferSize;
+       dml_uint_t WritebackInterfaceBufferSize;
+       dml_float_t DCFCLK;
+       dml_float_t ReturnBW;
+       dml_bool_t SynchronizeTimingsFinal;
+       dml_bool_t SynchronizeDRRDisplaysForUCLKPStateChangeFinal;
+       dml_bool_t *DRRDisplay;
+       dml_uint_t *dpte_group_bytes;
+       dml_uint_t *meta_row_height;
+       dml_uint_t *meta_row_height_chroma;
+       struct SOCParametersList mmSOCParameters;
+       dml_uint_t WritebackChunkSize;
+       dml_float_t SOCCLK;
+       dml_float_t DCFClkDeepSleep;
+       dml_uint_t *DETBufferSizeY;
+       dml_uint_t *DETBufferSizeC;
+       dml_uint_t *SwathHeightY;
+       dml_uint_t *SwathHeightC;
+       dml_uint_t *LBBitPerPixel;
+       dml_uint_t *SwathWidthY;
+       dml_uint_t *SwathWidthC;
+       dml_float_t *HRatio;
+       dml_float_t *HRatioChroma;
+       dml_uint_t *VTaps;
+       dml_uint_t *VTapsChroma;
+       dml_float_t *VRatio;
+       dml_float_t *VRatioChroma;
+       dml_uint_t *HTotal;
+       dml_uint_t *VTotal;
+       dml_uint_t *VActive;
+       dml_float_t *PixelClock;
+       dml_uint_t *BlendingAndTiming;
+       dml_uint_t *DPPPerSurface;
+       dml_float_t *BytePerPixelDETY;
+       dml_float_t *BytePerPixelDETC;
+       dml_uint_t *DSTXAfterScaler;
+       dml_uint_t *DSTYAfterScaler;
+       dml_bool_t *WritebackEnable;
+       enum dml_source_format_class *WritebackPixelFormat;
+       dml_uint_t *WritebackDestinationWidth;
+       dml_uint_t *WritebackDestinationHeight;
+       dml_uint_t *WritebackSourceHeight;
+       dml_bool_t UnboundedRequestEnabled;
+       dml_uint_t CompressedBufferSizeInkByte;
+
+       // Output
+       struct Watermarks *Watermark;
+       enum dml_dram_clock_change_support *DRAMClockChangeSupport;
+       dml_float_t *MaxActiveDRAMClockChangeLatencySupported;
+       dml_uint_t *SubViewportLinesNeededInMALL;
+       enum dml_fclock_change_support *FCLKChangeSupport;
+       dml_float_t *MaxActiveFCLKChangeLatencySupported;
+       dml_bool_t *USRRetrainingSupport;
+       dml_float_t *ActiveDRAMClockChangeLatencyMargin;
+};
+
+struct CalculateVMRowAndSwath_params_st {
+       dml_uint_t NumberOfActiveSurfaces;
+       struct DmlPipe *myPipe;
+       dml_uint_t *SurfaceSizeInMALL;
+       dml_uint_t PTEBufferSizeInRequestsLuma;
+       dml_uint_t PTEBufferSizeInRequestsChroma;
+       dml_uint_t DCCMetaBufferSizeBytes;
+       enum dml_use_mall_for_static_screen_mode *UseMALLForStaticScreen;
+       enum dml_use_mall_for_pstate_change_mode *UseMALLForPStateChange;
+       dml_uint_t MALLAllocatedForDCN;
+       dml_uint_t *SwathWidthY;
+       dml_uint_t *SwathWidthC;
+       dml_bool_t GPUVMEnable;
+       dml_bool_t HostVMEnable;
+       dml_uint_t HostVMMaxNonCachedPageTableLevels;
+       dml_uint_t GPUVMMaxPageTableLevels;
+       dml_uint_t *GPUVMMinPageSizeKBytes;
+       dml_uint_t HostVMMinPageSize;
+       dml_bool_t *PTEBufferModeOverrideEn;
+       dml_bool_t *PTEBufferModeOverrideVal;
+
+       // Output
+       dml_bool_t *PTEBufferSizeNotExceeded;
+       dml_bool_t *DCCMetaBufferSizeNotExceeded;
+       dml_uint_t *dpte_row_width_luma_ub;
+       dml_uint_t *dpte_row_width_chroma_ub;
+       dml_uint_t *dpte_row_height_luma;
+       dml_uint_t *dpte_row_height_chroma;
+       dml_uint_t *dpte_row_height_linear_luma; // VBA_DELTA
+       dml_uint_t *dpte_row_height_linear_chroma; // VBA_DELTA
+       dml_uint_t *meta_req_width;
+       dml_uint_t *meta_req_width_chroma;
+       dml_uint_t *meta_req_height;
+       dml_uint_t *meta_req_height_chroma;
+       dml_uint_t *meta_row_width;
+       dml_uint_t *meta_row_width_chroma;
+       dml_uint_t *meta_row_height;
+       dml_uint_t *meta_row_height_chroma;
+       dml_uint_t *vm_group_bytes;
+       dml_uint_t *dpte_group_bytes;
+       dml_uint_t *PixelPTEReqWidthY;
+       dml_uint_t *PixelPTEReqHeightY;
+       dml_uint_t *PTERequestSizeY;
+       dml_uint_t *PixelPTEReqWidthC;
+       dml_uint_t *PixelPTEReqHeightC;
+       dml_uint_t *PTERequestSizeC;
+       dml_uint_t *dpde0_bytes_per_frame_ub_l;
+       dml_uint_t *meta_pte_bytes_per_frame_ub_l;
+       dml_uint_t *dpde0_bytes_per_frame_ub_c;
+       dml_uint_t *meta_pte_bytes_per_frame_ub_c;
+       dml_uint_t *PrefetchSourceLinesY;
+       dml_uint_t *PrefetchSourceLinesC;
+       dml_uint_t *VInitPreFillY;
+       dml_uint_t *VInitPreFillC;
+       dml_uint_t *MaxNumSwathY;
+       dml_uint_t *MaxNumSwathC;
+       dml_float_t *meta_row_bw;
+       dml_float_t *dpte_row_bw;
+       dml_uint_t *PixelPTEBytesPerRow;
+       dml_uint_t *PDEAndMetaPTEBytesFrame;
+       dml_uint_t *MetaRowByte;
+       dml_bool_t *use_one_row_for_frame;
+       dml_bool_t *use_one_row_for_frame_flip;
+       dml_bool_t *UsesMALLForStaticScreen;
+       dml_bool_t *PTE_BUFFER_MODE;
+       dml_uint_t *BIGK_FRAGMENT_SIZE;
+};
+
+struct CalculateSwathAndDETConfiguration_params_st {
+       dml_uint_t *DETSizeOverride;
+       enum dml_use_mall_for_pstate_change_mode *UseMALLForPStateChange;
+       dml_uint_t ConfigReturnBufferSizeInKByte;
+       dml_uint_t ROBBufferSizeInKByte;
+       dml_uint_t MaxTotalDETInKByte;
+       dml_uint_t MinCompressedBufferSizeInKByte;
+       dml_uint_t PixelChunkSizeInKByte;
+       dml_bool_t ForceSingleDPP;
+       dml_uint_t NumberOfActiveSurfaces;
+       dml_uint_t nomDETInKByte;
+       enum dml_unbounded_requesting_policy UseUnboundedRequestingFinal;
+       dml_uint_t ConfigReturnBufferSegmentSizeInkByte;
+       dml_uint_t CompressedBufferSegmentSizeInkByteFinal;
+       enum dml_output_encoder_class *Output;
+       dml_float_t *ReadBandwidthLuma;
+       dml_float_t *ReadBandwidthChroma;
+       dml_float_t *MaximumSwathWidthLuma;
+       dml_float_t *MaximumSwathWidthChroma;
+       enum dml_rotation_angle *SourceScan;
+       dml_bool_t *ViewportStationary;
+       enum dml_source_format_class *SourcePixelFormat;
+       enum dml_swizzle_mode *SurfaceTiling;
+       dml_uint_t *ViewportWidth;
+       dml_uint_t *ViewportHeight;
+       dml_uint_t *ViewportXStart;
+       dml_uint_t *ViewportYStart;
+       dml_uint_t *ViewportXStartC;
+       dml_uint_t *ViewportYStartC;
+       dml_uint_t *SurfaceWidthY;
+       dml_uint_t *SurfaceWidthC;
+       dml_uint_t *SurfaceHeightY;
+       dml_uint_t *SurfaceHeightC;
+       dml_uint_t *Read256BytesBlockHeightY;
+       dml_uint_t *Read256BytesBlockHeightC;
+       dml_uint_t *Read256BytesBlockWidthY;
+       dml_uint_t *Read256BytesBlockWidthC;
+       enum dml_odm_mode *ODMMode;
+       dml_uint_t *BlendingAndTiming;
+       dml_uint_t *BytePerPixY;
+       dml_uint_t *BytePerPixC;
+       dml_float_t *BytePerPixDETY;
+       dml_float_t *BytePerPixDETC;
+       dml_uint_t *HActive;
+       dml_float_t *HRatio;
+       dml_float_t *HRatioChroma;
+       dml_uint_t *DPPPerSurface;
+       dml_uint_t *swath_width_luma_ub;
+       dml_uint_t *swath_width_chroma_ub;
+       dml_uint_t *SwathWidth;
+       dml_uint_t *SwathWidthChroma;
+       dml_uint_t *SwathHeightY;
+       dml_uint_t *SwathHeightC;
+       dml_uint_t *DETBufferSizeInKByte;
+       dml_uint_t *DETBufferSizeY;
+       dml_uint_t *DETBufferSizeC;
+       dml_bool_t *UnboundedRequestEnabled;
+       dml_uint_t *compbuf_reserved_space_64b;
+       dml_uint_t *compbuf_reserved_space_zs;
+       dml_uint_t *CompressedBufferSizeInkByte;
+       dml_bool_t *ViewportSizeSupportPerSurface;
+       dml_bool_t *ViewportSizeSupport;
+};
+
+struct CalculateStutterEfficiency_params_st {
+       dml_uint_t CompressedBufferSizeInkByte;
+       enum dml_use_mall_for_pstate_change_mode *UseMALLForPStateChange;
+       dml_bool_t UnboundedRequestEnabled;
+       dml_uint_t MetaFIFOSizeInKEntries;
+       dml_uint_t ZeroSizeBufferEntries;
+       dml_uint_t PixelChunkSizeInKByte;
+       dml_uint_t NumberOfActiveSurfaces;
+       dml_uint_t ROBBufferSizeInKByte;
+       dml_float_t TotalDataReadBandwidth;
+       dml_float_t DCFCLK;
+       dml_float_t ReturnBW;
+       dml_uint_t CompbufReservedSpace64B;
+       dml_uint_t CompbufReservedSpaceZs;
+       dml_float_t SRExitTime;
+       dml_float_t SRExitZ8Time;
+       dml_bool_t SynchronizeTimingsFinal;
+       dml_uint_t *BlendingAndTiming;
+       dml_float_t StutterEnterPlusExitWatermark;
+       dml_float_t Z8StutterEnterPlusExitWatermark;
+       dml_bool_t ProgressiveToInterlaceUnitInOPP;
+       dml_bool_t *Interlace;
+       dml_float_t *MinTTUVBlank;
+       dml_uint_t *DPPPerSurface;
+       dml_uint_t *DETBufferSizeY;
+       dml_uint_t *BytePerPixelY;
+       dml_float_t *BytePerPixelDETY;
+       dml_uint_t *SwathWidthY;
+       dml_uint_t *SwathHeightY;
+       dml_uint_t *SwathHeightC;
+       dml_float_t *NetDCCRateLuma;
+       dml_float_t *NetDCCRateChroma;
+       dml_float_t *DCCFractionOfZeroSizeRequestsLuma;
+       dml_float_t *DCCFractionOfZeroSizeRequestsChroma;
+       dml_uint_t *HTotal;
+       dml_uint_t *VTotal;
+       dml_float_t *PixelClock;
+       dml_float_t *VRatio;
+       enum dml_rotation_angle *SourceScan;
+       dml_uint_t *BlockHeight256BytesY;
+       dml_uint_t *BlockWidth256BytesY;
+       dml_uint_t *BlockHeight256BytesC;
+       dml_uint_t *BlockWidth256BytesC;
+       dml_uint_t *DCCYMaxUncompressedBlock;
+       dml_uint_t *DCCCMaxUncompressedBlock;
+       dml_uint_t *VActive;
+       dml_bool_t *DCCEnable;
+       dml_bool_t *WritebackEnable;
+       dml_float_t *ReadBandwidthSurfaceLuma;
+       dml_float_t *ReadBandwidthSurfaceChroma;
+       dml_float_t *meta_row_bw;
+       dml_float_t *dpte_row_bw;
+       dml_float_t *StutterEfficiencyNotIncludingVBlank;
+       dml_float_t *StutterEfficiency;
+       dml_uint_t *NumberOfStutterBurstsPerFrame;
+       dml_float_t *Z8StutterEfficiencyNotIncludingVBlank;
+       dml_float_t *Z8StutterEfficiency;
+       dml_uint_t *Z8NumberOfStutterBurstsPerFrame;
+       dml_float_t *StutterPeriod;
+       dml_bool_t *DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE;
+};
+
+struct CalculatePrefetchSchedule_params_st {
+       dml_bool_t EnhancedPrefetchScheduleAccelerationFinal;
+       dml_float_t HostVMInefficiencyFactor;
+       struct DmlPipe *myPipe;
+       dml_uint_t DSCDelay;
+       dml_float_t DPPCLKDelaySubtotalPlusCNVCFormater;
+       dml_float_t DPPCLKDelaySCL;
+       dml_float_t DPPCLKDelaySCLLBOnly;
+       dml_float_t DPPCLKDelayCNVCCursor;
+       dml_float_t DISPCLKDelaySubtotal;
+       dml_uint_t DPP_RECOUT_WIDTH;
+       enum dml_output_format_class OutputFormat;
+       dml_uint_t MaxInterDCNTileRepeaters;
+       dml_uint_t VStartup;
+       dml_uint_t MaxVStartup;
+       dml_uint_t GPUVMPageTableLevels;
+       dml_bool_t GPUVMEnable;
+       dml_bool_t HostVMEnable;
+       dml_uint_t HostVMMaxNonCachedPageTableLevels;
+       dml_uint_t HostVMMinPageSize;
+       dml_bool_t DynamicMetadataEnable;
+       dml_bool_t DynamicMetadataVMEnabled;
+       int DynamicMetadataLinesBeforeActiveRequired;
+       dml_uint_t DynamicMetadataTransmittedBytes;
+       dml_float_t UrgentLatency;
+       dml_float_t UrgentExtraLatency;
+       dml_float_t TCalc;
+       dml_uint_t PDEAndMetaPTEBytesFrame;
+       dml_uint_t MetaRowByte;
+       dml_uint_t PixelPTEBytesPerRow;
+       dml_float_t PrefetchSourceLinesY;
+       dml_uint_t VInitPreFillY;
+       dml_uint_t MaxNumSwathY;
+       dml_float_t PrefetchSourceLinesC;
+       dml_uint_t VInitPreFillC;
+       dml_uint_t MaxNumSwathC;
+       dml_uint_t swath_width_luma_ub;
+       dml_uint_t swath_width_chroma_ub;
+       dml_uint_t SwathHeightY;
+       dml_uint_t SwathHeightC;
+       dml_float_t TWait;
+       dml_uint_t *DSTXAfterScaler;
+       dml_uint_t *DSTYAfterScaler;
+       dml_float_t *DestinationLinesForPrefetch;
+       dml_float_t *DestinationLinesToRequestVMInVBlank;
+       dml_float_t *DestinationLinesToRequestRowInVBlank;
+       dml_float_t *VRatioPrefetchY;
+       dml_float_t *VRatioPrefetchC;
+       dml_float_t *RequiredPrefetchPixDataBWLuma;
+       dml_float_t *RequiredPrefetchPixDataBWChroma;
+       dml_bool_t *NotEnoughTimeForDynamicMetadata;
+       dml_float_t *Tno_bw;
+       dml_float_t *prefetch_vmrow_bw;
+       dml_float_t *Tdmdl_vm;
+       dml_float_t *Tdmdl;
+       dml_float_t *TSetup;
+       dml_uint_t *VUpdateOffsetPix;
+       dml_uint_t *VUpdateWidthPix;
+       dml_uint_t *VReadyOffsetPix;
+};
+
+struct dml_core_mode_support_locals_st {
+       dml_bool_t dummy_boolean[2];
+       dml_uint_t dummy_integer[3];
+       dml_uint_t dummy_integer_array[22][__DML_NUM_PLANES__];
+       enum dml_odm_mode dummy_odm_mode[__DML_NUM_PLANES__];
+       dml_bool_t dummy_boolean_array[2][__DML_NUM_PLANES__];
+       dml_uint_t MaxVStartupAllPlanes[2];
+       dml_uint_t MaximumVStartup[2][__DML_NUM_PLANES__];
+       dml_uint_t DSTYAfterScaler[__DML_NUM_PLANES__];
+       dml_uint_t DSTXAfterScaler[__DML_NUM_PLANES__];
+       dml_uint_t NextPrefetchMode[__DML_NUM_PLANES__];
+       dml_uint_t MinPrefetchMode[__DML_NUM_PLANES__];
+       dml_uint_t MaxPrefetchMode[__DML_NUM_PLANES__];
+       dml_float_t dummy_single[3];
+       dml_float_t dummy_single_array[__DML_NUM_PLANES__];
+       struct Watermarks dummy_watermark;
+       struct SOCParametersList mSOCParameters;
+       struct DmlPipe myPipe;
+       struct DmlPipe SurfParameters[__DML_NUM_PLANES__];
+       dml_uint_t TotalNumberOfActiveWriteback;
+       dml_uint_t MaximumSwathWidthSupportLuma;
+       dml_uint_t MaximumSwathWidthSupportChroma;
+       dml_bool_t MPCCombineMethodAsNeededForPStateChangeAndVoltage;
+       dml_bool_t MPCCombineMethodAsPossible;
+       dml_bool_t TotalAvailablePipesSupportNoDSC;
+       dml_uint_t NumberOfDPPNoDSC;
+       enum dml_odm_mode ODMModeNoDSC;
+       dml_float_t RequiredDISPCLKPerSurfaceNoDSC;
+       dml_bool_t TotalAvailablePipesSupportDSC;
+       dml_uint_t NumberOfDPPDSC;
+       enum dml_odm_mode ODMModeDSC;
+       dml_float_t RequiredDISPCLKPerSurfaceDSC;
+       dml_bool_t NoChromaOrLinear;
+       dml_float_t BWOfNonCombinedSurfaceOfMaximumBandwidth;
+       dml_uint_t NumberOfNonCombinedSurfaceOfMaximumBandwidth;
+       dml_uint_t TotalNumberOfActiveOTG;
+       dml_uint_t TotalNumberOfActiveHDMIFRL;
+       dml_uint_t TotalNumberOfActiveDP2p0;
+       dml_uint_t TotalNumberOfActiveDP2p0Outputs;
+       dml_uint_t TotalSlots;
+       dml_uint_t DSCFormatFactor;
+       dml_uint_t TotalDSCUnitsRequired;
+       dml_uint_t ReorderingBytes;
+       dml_bool_t ImmediateFlipRequiredFinal;
+       dml_bool_t FullFrameMALLPStateMethod;
+       dml_bool_t SubViewportMALLPStateMethod;
+       dml_bool_t PhantomPipeMALLPStateMethod;
+       dml_bool_t SubViewportMALLRefreshGreaterThan120Hz;
+       dml_float_t MaxTotalVActiveRDBandwidth;
+       dml_float_t VMDataOnlyReturnBWPerState;
+       dml_float_t HostVMInefficiencyFactor;
+       dml_uint_t NextMaxVStartup;
+       dml_uint_t MaxVStartup;
+       dml_bool_t AllPrefetchModeTested;
+       dml_bool_t AnyLinesForVMOrRowTooLarge;
+       dml_bool_t is_max_pwr_state;
+       dml_bool_t is_max_dram_pwr_state;
+       dml_bool_t dram_clock_change_support;
+       dml_bool_t f_clock_change_support;
+};
+
+struct dml_core_mode_programming_locals_st {
+       dml_uint_t DSCFormatFactor;
+       dml_uint_t dummy_integer_array[2][__DML_NUM_PLANES__];
+       enum dml_output_encoder_class dummy_output_encoder_array[__DML_NUM_PLANES__];
+       dml_float_t dummy_single_array[2][__DML_NUM_PLANES__];
+       dml_uint_t dummy_long_array[4][__DML_NUM_PLANES__];
+       dml_bool_t dummy_boolean_array[2][__DML_NUM_PLANES__];
+       dml_bool_t dummy_boolean[1];
+       struct DmlPipe SurfaceParameters[__DML_NUM_PLANES__];
+       dml_uint_t ReorderBytes;
+       dml_float_t VMDataOnlyReturnBW;
+       dml_float_t HostVMInefficiencyFactor;
+       dml_uint_t TotalDCCActiveDPP;
+       dml_uint_t TotalActiveDPP;
+       dml_uint_t VStartupLines;
+       dml_uint_t MaxVStartupLines[__DML_NUM_PLANES__]; /// <brief more like vblank for the plane's OTG
+       dml_uint_t MaxVStartupAllPlanes;
+       dml_bool_t ImmediateFlipRequirementFinal;
+       int iteration;
+       dml_float_t MaxTotalRDBandwidth;
+       dml_float_t MaxTotalRDBandwidthNoUrgentBurst;
+       dml_bool_t DestinationLineTimesForPrefetchLessThan2;
+       dml_bool_t VRatioPrefetchMoreThanMax;
+       dml_float_t MaxTotalRDBandwidthNotIncludingMALLPrefetch;
+       dml_uint_t NextPrefetchMode[__DML_NUM_PLANES__];
+       dml_uint_t MinPrefetchMode[__DML_NUM_PLANES__];
+       dml_uint_t MaxPrefetchMode[__DML_NUM_PLANES__];
+       dml_bool_t AllPrefetchModeTested;
+       dml_float_t dummy_unit_vector[__DML_NUM_PLANES__];
+       dml_float_t NonUrgentMaxTotalRDBandwidth;
+       dml_float_t NonUrgentMaxTotalRDBandwidthNotIncludingMALLPrefetch;
+       dml_float_t dummy_single[2];
+       struct SOCParametersList mmSOCParameters;
+       dml_float_t Tvstartup_margin;
+       dml_float_t dlg_vblank_start;
+       dml_float_t LSetup;
+       dml_float_t blank_lines_remaining;
+       dml_float_t old_MIN_DST_Y_NEXT_START;
+       dml_float_t TotalWRBandwidth;
+       dml_float_t WRBandwidth;
+       struct Watermarks dummy_watermark;
+       struct DmlPipe myPipe;
+};
+
+struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals_st {
+       dml_float_t ActiveDRAMClockChangeLatencyMargin[__DML_NUM_PLANES__];
+       dml_float_t ActiveFCLKChangeLatencyMargin[__DML_NUM_PLANES__];
+       dml_float_t USRRetrainingLatencyMargin[__DML_NUM_PLANES__];
+
+       dml_bool_t SynchronizedSurfaces[__DML_NUM_PLANES__][__DML_NUM_PLANES__];
+       dml_float_t EffectiveLBLatencyHidingY;
+       dml_float_t EffectiveLBLatencyHidingC;
+       dml_float_t LinesInDETY[__DML_NUM_PLANES__];
+       dml_float_t LinesInDETC[__DML_NUM_PLANES__];
+       dml_uint_t LinesInDETYRoundedDownToSwath[__DML_NUM_PLANES__];
+       dml_uint_t LinesInDETCRoundedDownToSwath[__DML_NUM_PLANES__];
+       dml_float_t FullDETBufferingTimeY;
+       dml_float_t FullDETBufferingTimeC;
+       dml_float_t WritebackDRAMClockChangeLatencyMargin;
+       dml_float_t WritebackFCLKChangeLatencyMargin;
+       dml_float_t WritebackLatencyHiding;
+
+       dml_uint_t TotalActiveWriteback;
+       dml_uint_t LBLatencyHidingSourceLinesY[__DML_NUM_PLANES__];
+       dml_uint_t LBLatencyHidingSourceLinesC[__DML_NUM_PLANES__];
+       dml_float_t TotalPixelBW;
+       dml_float_t EffectiveDETBufferSizeY;
+       dml_float_t ActiveClockChangeLatencyHidingY;
+       dml_float_t ActiveClockChangeLatencyHidingC;
+       dml_float_t ActiveClockChangeLatencyHiding;
+       dml_bool_t FoundCriticalSurface;
+       dml_uint_t LastSurfaceWithoutMargin;
+       dml_uint_t FCLKChangeSupportNumber;
+       dml_uint_t DRAMClockChangeMethod;
+       dml_uint_t DRAMClockChangeSupportNumber;
+       dml_uint_t dst_y_pstate;
+       dml_uint_t src_y_pstate_l;
+       dml_uint_t src_y_pstate_c;
+       dml_uint_t src_y_ahead_l;
+       dml_uint_t src_y_ahead_c;
+       dml_uint_t sub_vp_lines_l;
+       dml_uint_t sub_vp_lines_c;
+};
+
+struct CalculateVMRowAndSwath_locals_st {
+       dml_uint_t PTEBufferSizeInRequestsForLuma[__DML_NUM_PLANES__];
+       dml_uint_t PTEBufferSizeInRequestsForChroma[__DML_NUM_PLANES__];
+       dml_uint_t PDEAndMetaPTEBytesFrameY;
+       dml_uint_t PDEAndMetaPTEBytesFrameC;
+       dml_uint_t MetaRowByteY[__DML_NUM_PLANES__];
+       dml_uint_t MetaRowByteC[__DML_NUM_PLANES__];
+       dml_uint_t PixelPTEBytesPerRowY[__DML_NUM_PLANES__];
+       dml_uint_t PixelPTEBytesPerRowC[__DML_NUM_PLANES__];
+       dml_uint_t PixelPTEBytesPerRowStorageY[__DML_NUM_PLANES__];
+       dml_uint_t PixelPTEBytesPerRowStorageC[__DML_NUM_PLANES__];
+       dml_uint_t PixelPTEBytesPerRowY_one_row_per_frame[__DML_NUM_PLANES__];
+       dml_uint_t PixelPTEBytesPerRowC_one_row_per_frame[__DML_NUM_PLANES__];
+       dml_uint_t dpte_row_width_luma_ub_one_row_per_frame[__DML_NUM_PLANES__];
+       dml_uint_t dpte_row_height_luma_one_row_per_frame[__DML_NUM_PLANES__];
+       dml_uint_t dpte_row_width_chroma_ub_one_row_per_frame[__DML_NUM_PLANES__];
+       dml_uint_t dpte_row_height_chroma_one_row_per_frame[__DML_NUM_PLANES__];
+       dml_bool_t one_row_per_frame_fits_in_buffer[__DML_NUM_PLANES__];
+
+       dml_uint_t HostVMDynamicLevels;
+};
+
+struct UseMinimumDCFCLK_locals_st {
+       dml_uint_t dummy1;
+       dml_uint_t dummy2;
+       dml_uint_t dummy3;
+       dml_float_t NormalEfficiency;
+       dml_float_t TotalMaxPrefetchFlipDPTERowBandwidth[2];
+
+       dml_float_t PixelDCFCLKCyclesRequiredInPrefetch[__DML_NUM_PLANES__];
+       dml_float_t PrefetchPixelLinesTime[__DML_NUM_PLANES__];
+       dml_float_t DCFCLKRequiredForPeakBandwidthPerSurface[__DML_NUM_PLANES__];
+       dml_float_t DynamicMetadataVMExtraLatency[__DML_NUM_PLANES__];
+       dml_float_t MinimumTWait;
+       dml_float_t DPTEBandwidth;
+       dml_float_t DCFCLKRequiredForAverageBandwidth;
+       dml_uint_t ExtraLatencyBytes;
+       dml_float_t ExtraLatencyCycles;
+       dml_float_t DCFCLKRequiredForPeakBandwidth;
+       dml_uint_t NoOfDPPState[__DML_NUM_PLANES__];
+       dml_float_t MinimumTvmPlus2Tr0;
+};
+
+struct CalculatePrefetchSchedule_locals_st {
+       dml_bool_t MyError;
+       dml_uint_t DPPCycles;
+       dml_uint_t DISPCLKCycles;
+       dml_float_t DSTTotalPixelsAfterScaler;
+       dml_float_t LineTime;
+       dml_float_t dst_y_prefetch_equ;
+       dml_float_t prefetch_bw_oto;
+       dml_float_t Tvm_oto;
+       dml_float_t Tr0_oto;
+       dml_float_t Tvm_oto_lines;
+       dml_float_t Tr0_oto_lines;
+       dml_float_t dst_y_prefetch_oto;
+       dml_float_t TimeForFetchingMetaPTE;
+       dml_float_t TimeForFetchingRowInVBlank;
+       dml_float_t LinesToRequestPrefetchPixelData;
+       dml_uint_t HostVMDynamicLevelsTrips;
+       dml_float_t trip_to_mem;
+       dml_float_t Tvm_trips;
+       dml_float_t Tr0_trips;
+       dml_float_t Tvm_trips_rounded;
+       dml_float_t Tr0_trips_rounded;
+       dml_float_t max_Tsw;
+       dml_float_t Lsw_oto;
+       dml_float_t Tpre_rounded;
+       dml_float_t prefetch_bw_equ;
+       dml_float_t Tvm_equ;
+       dml_float_t Tr0_equ;
+       dml_float_t Tdmbf;
+       dml_float_t Tdmec;
+       dml_float_t Tdmsks;
+       dml_float_t prefetch_sw_bytes;
+       dml_float_t prefetch_bw_pr;
+       dml_float_t bytes_pp;
+       dml_float_t dep_bytes;
+       dml_float_t min_Lsw_oto;
+       dml_float_t Tsw_est1;
+       dml_float_t Tsw_est3;
+       dml_float_t PrefetchBandwidth1;
+       dml_float_t PrefetchBandwidth2;
+       dml_float_t PrefetchBandwidth3;
+       dml_float_t PrefetchBandwidth4;
+};
+
+/// @brief To minimize stack usage; function locals are instead placed into this scratch structure which is allocated per context
+struct display_mode_lib_scratch_st {
+       // Scratch space for function locals
+       struct dml_core_mode_support_locals_st dml_core_mode_support_locals;
+       struct dml_core_mode_programming_locals_st dml_core_mode_programming_locals;
+       struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals_st CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals;
+       struct CalculateVMRowAndSwath_locals_st CalculateVMRowAndSwath_locals;
+       struct UseMinimumDCFCLK_locals_st UseMinimumDCFCLK_locals;
+       struct CalculatePrefetchSchedule_locals_st CalculatePrefetchSchedule_locals;
+
+       // Scratch space for function params
+       struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params_st CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params;
+       struct CalculateVMRowAndSwath_params_st CalculateVMRowAndSwath_params;
+       struct UseMinimumDCFCLK_params_st UseMinimumDCFCLK_params;
+       struct CalculateSwathAndDETConfiguration_params_st CalculateSwathAndDETConfiguration_params;
+       struct CalculateStutterEfficiency_params_st CalculateStutterEfficiency_params;
+       struct CalculatePrefetchSchedule_params_st CalculatePrefetchSchedule_params;
+};
+
+/// @brief Represent the overall soc/ip enviroment. It contains data structure represent the soc/ip characteristic and also structures that hold calculation output
+struct display_mode_lib_st {
+       dml_uint_t project;
+
+       //@brief Mode evaluation and programming policy
+       struct dml_mode_eval_policy_st policy;
+
+       //@brief IP/SOC characteristic
+       struct ip_params_st ip;
+       struct soc_bounding_box_st soc;
+       struct soc_states_st states;
+
+       //@brief Mode Support and Mode programming struct
+       // Used to hold input; intermediate and output of the calculations
+       struct mode_support_st ms; // struct for mode support
+       struct mode_program_st mp; // struct for mode programming
+
+       struct display_mode_lib_scratch_st scratch;
+};
+
+struct dml_mode_support_ex_params_st {
+       struct display_mode_lib_st *mode_lib;
+       const struct dml_display_cfg_st *in_display_cfg;
+       dml_uint_t out_lowest_state_idx;
+       struct dml_mode_support_info_st *out_evaluation_info;
+};
+
+typedef struct _vcs_dpi_dml_display_rq_regs_st  dml_display_rq_regs_st;
+typedef struct _vcs_dpi_dml_display_dlg_regs_st dml_display_dlg_regs_st;
+typedef struct _vcs_dpi_dml_display_ttu_regs_st dml_display_ttu_regs_st;
+typedef struct _vcs_dpi_dml_display_arb_params_st   dml_display_arb_params_st;
+typedef struct _vcs_dpi_dml_display_plane_rq_regs_st    dml_display_plane_rq_regs_st;
+
+struct  _vcs_dpi_dml_display_dlg_regs_st {
+       dml_uint_t  refcyc_h_blank_end;
+       dml_uint_t  dlg_vblank_end;
+       dml_uint_t  min_dst_y_next_start;
+       dml_uint_t  refcyc_per_htotal;
+       dml_uint_t  refcyc_x_after_scaler;
+       dml_uint_t  dst_y_after_scaler;
+       dml_uint_t  dst_y_prefetch;
+       dml_uint_t  dst_y_per_vm_vblank;
+       dml_uint_t  dst_y_per_row_vblank;
+       dml_uint_t  dst_y_per_vm_flip;
+       dml_uint_t  dst_y_per_row_flip;
+       dml_uint_t  ref_freq_to_pix_freq;
+       dml_uint_t  vratio_prefetch;
+       dml_uint_t  vratio_prefetch_c;
+       dml_uint_t  refcyc_per_pte_group_vblank_l;
+       dml_uint_t  refcyc_per_pte_group_vblank_c;
+       dml_uint_t  refcyc_per_meta_chunk_vblank_l;
+       dml_uint_t  refcyc_per_meta_chunk_vblank_c;
+       dml_uint_t  refcyc_per_pte_group_flip_l;
+       dml_uint_t  refcyc_per_pte_group_flip_c;
+       dml_uint_t  refcyc_per_meta_chunk_flip_l;
+       dml_uint_t  refcyc_per_meta_chunk_flip_c;
+       dml_uint_t  dst_y_per_pte_row_nom_l;
+       dml_uint_t  dst_y_per_pte_row_nom_c;
+       dml_uint_t  refcyc_per_pte_group_nom_l;
+       dml_uint_t  refcyc_per_pte_group_nom_c;
+       dml_uint_t  dst_y_per_meta_row_nom_l;
+       dml_uint_t  dst_y_per_meta_row_nom_c;
+       dml_uint_t  refcyc_per_meta_chunk_nom_l;
+       dml_uint_t  refcyc_per_meta_chunk_nom_c;
+       dml_uint_t  refcyc_per_line_delivery_pre_l;
+       dml_uint_t  refcyc_per_line_delivery_pre_c;
+       dml_uint_t  refcyc_per_line_delivery_l;
+       dml_uint_t  refcyc_per_line_delivery_c;
+       dml_uint_t  refcyc_per_vm_group_vblank;
+       dml_uint_t  refcyc_per_vm_group_flip;
+       dml_uint_t  refcyc_per_vm_req_vblank;
+       dml_uint_t  refcyc_per_vm_req_flip;
+       dml_uint_t  dst_y_offset_cur0;
+       dml_uint_t  chunk_hdl_adjust_cur0;
+       dml_uint_t  dst_y_offset_cur1;
+       dml_uint_t  chunk_hdl_adjust_cur1;
+       dml_uint_t  vready_after_vcount0;
+       dml_uint_t  dst_y_delta_drq_limit;
+       dml_uint_t  refcyc_per_vm_dmdata;
+       dml_uint_t  dmdata_dl_delta;
+};
+
+struct  _vcs_dpi_dml_display_ttu_regs_st {
+       dml_uint_t  qos_level_low_wm;
+       dml_uint_t  qos_level_high_wm;
+       dml_uint_t  min_ttu_vblank;
+       dml_uint_t  qos_level_flip;
+       dml_uint_t  refcyc_per_req_delivery_l;
+       dml_uint_t  refcyc_per_req_delivery_c;
+       dml_uint_t  refcyc_per_req_delivery_cur0;
+       dml_uint_t  refcyc_per_req_delivery_cur1;
+       dml_uint_t  refcyc_per_req_delivery_pre_l;
+       dml_uint_t  refcyc_per_req_delivery_pre_c;
+       dml_uint_t  refcyc_per_req_delivery_pre_cur0;
+       dml_uint_t  refcyc_per_req_delivery_pre_cur1;
+       dml_uint_t  qos_level_fixed_l;
+       dml_uint_t  qos_level_fixed_c;
+       dml_uint_t  qos_level_fixed_cur0;
+       dml_uint_t  qos_level_fixed_cur1;
+       dml_uint_t  qos_ramp_disable_l;
+       dml_uint_t  qos_ramp_disable_c;
+       dml_uint_t  qos_ramp_disable_cur0;
+       dml_uint_t  qos_ramp_disable_cur1;
+};
+
+struct  _vcs_dpi_dml_display_arb_params_st {
+       dml_uint_t  max_req_outstanding;
+       dml_uint_t  min_req_outstanding;
+       dml_uint_t  sat_level_us;
+       dml_uint_t  hvm_max_qos_commit_threshold;
+       dml_uint_t  hvm_min_req_outstand_commit_threshold;
+       dml_uint_t  compbuf_reserved_space_kbytes;
+};
+
+struct  _vcs_dpi_dml_display_plane_rq_regs_st {
+       dml_uint_t  chunk_size;
+       dml_uint_t  min_chunk_size;
+       dml_uint_t  meta_chunk_size;
+       dml_uint_t  min_meta_chunk_size;
+       dml_uint_t  dpte_group_size;
+       dml_uint_t  mpte_group_size;
+       dml_uint_t  swath_height;
+       dml_uint_t  pte_row_height_linear;
+};
+
+struct  _vcs_dpi_dml_display_rq_regs_st {
+       dml_display_plane_rq_regs_st    rq_regs_l;
+       dml_display_plane_rq_regs_st    rq_regs_c;
+       dml_uint_t  drq_expansion_mode;
+       dml_uint_t  prq_expansion_mode;
+       dml_uint_t  mrq_expansion_mode;
+       dml_uint_t  crq_expansion_mode;
+       dml_uint_t  plane1_base_address;
+};
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_lib_defines.h b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_lib_defines.h
new file mode 100644 (file)
index 0000000..99bdb2d
--- /dev/null
@@ -0,0 +1,75 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DISPLAY_MODE_LIB_DEFINES_H__
+#define __DISPLAY_MODE_LIB_DEFINES_H__
+
+#define DCN_DML__DML_STANDALONE 1
+#define DCN_DML__DML_STANDALONE__1 1
+#define DCN_DML__PRESENT 1
+#define DCN_DML__PRESENT__1 1
+#define DCN_DML__NUM_PLANE 8
+#define DCN_DML__NUM_PLANE__8 1
+#define DCN_DML__NUM_CURSOR 1
+#define DCN_DML__NUM_CURSOR__1 1
+#define DCN_DML__NUM_PWR_STATE 30
+#define DCN_DML__NUM_PWR_STATE__30 1
+#define DCN_DML__VM_PRESENT 1
+#define DCN_DML__VM_PRESENT__1 1
+#define DCN_DML__HOST_VM_PRESENT 1
+#define DCN_DML__HOST_VM_PRESENT__1 1
+
+#include "dml_depedencies.h"
+
+#include "dml_logging.h"
+#include "dml_assert.h"
+
+// To enable a lot of debug msg
+#define __DML_VBA_DEBUG__
+#define __DML_VBA_ENABLE_INLINE_CHECK_                  0
+#define __DML_VBA_MIN_VSTARTUP__                        9       //<brief At which vstartup the DML start to try if the mode can be supported
+#define __DML_ARB_TO_RET_DELAY__                        7 + 95  //<brief Delay in DCFCLK from ARB to DET (1st num is ARB to SDPIF, 2nd number is SDPIF to DET)
+#define __DML_MIN_DCFCLK_FACTOR__                       1.15    //<brief fudge factor for min dcfclk calclation
+#define __DML_MAX_VRATIO_PRE__                          4.0     //<brief Prefetch schedule max vratio
+#define __DML_MAX_VRATIO_PRE_OTO__                      4.0     //<brief Prefetch schedule max vratio for one to one scheduling calculation for prefetch
+#define __DML_MAX_VRATIO_PRE_ENHANCE_PREFETCH_ACC__     6.0     //<brief Prefetch schedule max vratio when enhance prefetch schedule acceleration is enabled and vstartup is earliest possible already
+#define __DML_NUM_PLANES__                              DCN_DML__NUM_PLANE
+#define __DML_NUM_CURSORS__                             DCN_DML__NUM_CURSOR
+#define __DML_DPP_INVALID__                             0
+#define __DML_PIPE_NO_PLANE__                           99
+
+#define __DML_MAX_STATE_ARRAY_SIZE__        DCN_DML__NUM_PWR_STATE
+
+// Compilation define
+#define __DML_DLL_EXPORT__
+
+typedef          int   dml_int_t;   // int is 32-bit in C/C++, but Integer datatype is 16-bit in VBA. this should map to Long in VBA
+typedef unsigned int   dml_uint_t;
+typedef double         dml_float_t;
+
+// Note: bool is 8-bit in C/C++, but Boolean is 16-bit in VBA, use "short" in C/C++ DLL so the struct work when vba uses DLL
+// Or the VBA side don't use Boolean, just use "Byte", then C side can use bool
+typedef bool          dml_bool_t;
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_util.c b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_util.c
new file mode 100644 (file)
index 0000000..7dd1f8a
--- /dev/null
@@ -0,0 +1,796 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "display_mode_util.h"
+
+static dml_float_t _log(float in)
+{
+       int * const exp_ptr = (int *)(&in);
+       int x = *exp_ptr;
+       const int log_2 = ((x >> 23) & 255) - 128;
+
+       x &= ~(255 << 23);
+       x += 127 << 23;
+       *exp_ptr = x;
+
+       in = ((-1.0f / 3) * in + 2) * in - 2.0f / 3;
+
+       return (in + log_2);
+}
+
+dml_bool_t dml_util_is_420(enum dml_source_format_class source_format)
+{
+       dml_bool_t val = false;
+
+       switch (source_format) {
+       case dml_444_16:
+               val = 0;
+               break;
+       case dml_444_32:
+               val = 0;
+               break;
+       case dml_444_64:
+               val = 0;
+               break;
+       case dml_420_8:
+               val = 1;
+               break;
+       case dml_420_10:
+               val = 1;
+               break;
+       case dml_422_8:
+               val = 0;
+               break;
+       case dml_422_10:
+               val = 0;
+               break;
+       default:
+               ASSERT(0);
+               break;
+       }
+       return val;
+}
+
+static inline float dcn_bw_pow(float a, float exp)
+{
+       float temp;
+       /*ASSERT(exp == (int)exp);*/
+       if ((int)exp == 0)
+               return 1;
+       temp = dcn_bw_pow(a, (int)(exp / 2));
+       if (((int)exp % 2) == 0) {
+               return temp * temp;
+       } else {
+               if ((int)exp > 0)
+                       return a * temp * temp;
+               else
+                       return (temp * temp) / a;
+       }
+}
+
+static inline float dcn_bw_ceil2(const float arg, const float significance)
+{
+       ASSERT(significance != 0);
+
+       return ((int)(arg / significance + 0.99999)) * significance;
+}
+
+static inline float dcn_bw_floor2(const float arg, const float significance)
+{
+       ASSERT(significance != 0);
+
+       return ((int)(arg / significance)) * significance;
+}
+
+dml_float_t dml_ceil(dml_float_t x, dml_float_t granularity)
+{
+       if (granularity == 0)
+               return 0;
+       //return (dml_float_t) (ceil(x / granularity) * granularity);
+       return (dml_float_t)dcn_bw_ceil2(x, granularity);
+}
+
+dml_float_t dml_floor(dml_float_t x, dml_float_t granularity)
+{
+       if (granularity == 0)
+       return 0;
+       //return (dml_float_t) (floor(x / granularity) * granularity);
+       return (dml_float_t)dcn_bw_floor2(x, granularity);
+}
+
+dml_float_t dml_min(dml_float_t x, dml_float_t y)
+{
+       if (x != x)
+               return y;
+       if (y != y)
+               return x;
+       if (x < y)
+               return x;
+       else
+               return y;
+}
+
+dml_float_t dml_min3(dml_float_t x, dml_float_t y, dml_float_t z)
+{
+       return dml_min(dml_min(x, y), z);
+}
+
+dml_float_t dml_min4(dml_float_t x, dml_float_t y, dml_float_t z, dml_float_t w)
+{
+       return dml_min(dml_min(dml_min(x, y), z), w);
+}
+
+dml_float_t dml_max(dml_float_t x, dml_float_t y)
+{
+       if (x != x)
+               return y;
+       if (y != y)
+               return x;
+if (x > y)
+               return x;
+       else
+               return y;
+}
+dml_float_t dml_max3(dml_float_t x, dml_float_t y, dml_float_t z)
+{
+       return dml_max(dml_max(x, y), z);
+}
+dml_float_t dml_max4(dml_float_t a, dml_float_t b, dml_float_t c, dml_float_t d)
+{
+       return dml_max(dml_max(a, b), dml_max(c, d));
+}
+dml_float_t dml_max5(dml_float_t a, dml_float_t b, dml_float_t c, dml_float_t d, dml_float_t e)
+{
+       return dml_max(dml_max4(a, b, c, d), e);
+}
+dml_float_t dml_log(dml_float_t x, dml_float_t base)
+{
+       return (dml_float_t) (_log(x) / _log(base));
+}
+
+dml_float_t dml_log2(dml_float_t x)
+{
+       return (dml_float_t) (_log(x) / _log(2));
+}
+
+dml_float_t dml_round(dml_float_t val, dml_bool_t bankers_rounding)
+{
+//     if (bankers_rounding)
+//                     return (dml_float_t) lrint(val);
+//     else {
+//             return round(val);
+               double round_pt = 0.5;
+               double ceil = dml_ceil(val, 1);
+               double floor = dml_floor(val, 1);
+
+               if (val - floor >= round_pt)
+                       return ceil;
+               else
+                       return floor;
+//     }
+}
+
+dml_float_t dml_pow(dml_float_t base, int exp)
+{
+       return (dml_float_t) dcn_bw_pow(base, exp);
+}
+
+dml_uint_t dml_round_to_multiple(dml_uint_t num, dml_uint_t multiple, dml_bool_t up)
+{
+       dml_uint_t remainder;
+
+       if (multiple == 0)
+               return num;
+
+       remainder = num % multiple;
+       if (remainder == 0)
+               return num;
+
+       if (up)
+               return (num + multiple - remainder);
+       else
+               return (num - remainder);
+}
+
+void dml_print_data_rq_regs_st(const dml_display_plane_rq_regs_st *rq_regs)
+{
+       dml_print("DML: ===================================== \n");
+       dml_print("DML: DISPLAY_PLANE_RQ_REGS_ST\n");
+       dml_print("DML: chunk_size = 0x%x\n", rq_regs->chunk_size);
+       dml_print("DML: min_chunk_size = 0x%x\n", rq_regs->min_chunk_size);
+       dml_print("DML: meta_chunk_size = 0x%x\n", rq_regs->meta_chunk_size);
+       dml_print("DML: min_meta_chunk_size = 0x%x\n", rq_regs->min_meta_chunk_size);
+       dml_print("DML: dpte_group_size = 0x%x\n", rq_regs->dpte_group_size);
+       dml_print("DML: mpte_group_size = 0x%x\n", rq_regs->mpte_group_size);
+       dml_print("DML: swath_height = 0x%x\n", rq_regs->swath_height);
+       dml_print("DML: pte_row_height_linear = 0x%x\n", rq_regs->pte_row_height_linear);
+       dml_print("DML: ===================================== \n");
+}
+
+void dml_print_rq_regs_st(const dml_display_rq_regs_st *rq_regs)
+{
+       dml_print("DML: ===================================== \n");
+       dml_print("DML: DISPLAY_RQ_REGS_ST\n");
+       dml_print("DML: <LUMA> \n");
+       dml_print_data_rq_regs_st(&rq_regs->rq_regs_l);
+       dml_print("DML: <CHROMA> \n");
+       dml_print_data_rq_regs_st(&rq_regs->rq_regs_c);
+       dml_print("DML: drq_expansion_mode = 0x%x\n", rq_regs->drq_expansion_mode);
+       dml_print("DML: prq_expansion_mode = 0x%x\n", rq_regs->prq_expansion_mode);
+       dml_print("DML: mrq_expansion_mode = 0x%x\n", rq_regs->mrq_expansion_mode);
+       dml_print("DML: crq_expansion_mode = 0x%x\n", rq_regs->crq_expansion_mode);
+       dml_print("DML: plane1_base_address = 0x%x\n", rq_regs->plane1_base_address);
+       dml_print("DML: ===================================== \n");
+}
+
+void dml_print_dlg_regs_st(const dml_display_dlg_regs_st *dlg_regs)
+{
+       dml_print("DML: ===================================== \n");
+       dml_print("DML: DISPLAY_DLG_REGS_ST \n");
+       dml_print("DML: refcyc_h_blank_end = 0x%x\n", dlg_regs->refcyc_h_blank_end);
+       dml_print("DML: dlg_vblank_end = 0x%x\n", dlg_regs->dlg_vblank_end);
+       dml_print("DML: min_dst_y_next_start = 0x%x\n", dlg_regs->min_dst_y_next_start);
+       dml_print("DML: refcyc_per_htotal = 0x%x\n", dlg_regs->refcyc_per_htotal);
+       dml_print("DML: refcyc_x_after_scaler = 0x%x\n", dlg_regs->refcyc_x_after_scaler);
+       dml_print("DML: dst_y_after_scaler = 0x%x\n", dlg_regs->dst_y_after_scaler);
+       dml_print("DML: dst_y_prefetch = 0x%x\n", dlg_regs->dst_y_prefetch);
+       dml_print("DML: dst_y_per_vm_vblank = 0x%x\n", dlg_regs->dst_y_per_vm_vblank);
+       dml_print("DML: dst_y_per_row_vblank = 0x%x\n", dlg_regs->dst_y_per_row_vblank);
+       dml_print("DML: dst_y_per_vm_flip = 0x%x\n", dlg_regs->dst_y_per_vm_flip);
+       dml_print("DML: dst_y_per_row_flip = 0x%x\n", dlg_regs->dst_y_per_row_flip);
+       dml_print("DML: ref_freq_to_pix_freq = 0x%x\n", dlg_regs->ref_freq_to_pix_freq);
+       dml_print("DML: vratio_prefetch = 0x%x\n", dlg_regs->vratio_prefetch);
+       dml_print("DML: vratio_prefetch_c = 0x%x\n", dlg_regs->vratio_prefetch_c);
+       dml_print("DML: refcyc_per_pte_group_vblank_l = 0x%x\n", dlg_regs->refcyc_per_pte_group_vblank_l);
+       dml_print("DML: refcyc_per_pte_group_vblank_c = 0x%x\n", dlg_regs->refcyc_per_pte_group_vblank_c);
+       dml_print("DML: refcyc_per_meta_chunk_vblank_l = 0x%x\n", dlg_regs->refcyc_per_meta_chunk_vblank_l);
+       dml_print("DML: refcyc_per_meta_chunk_vblank_c = 0x%x\n", dlg_regs->refcyc_per_meta_chunk_vblank_c);
+       dml_print("DML: refcyc_per_pte_group_flip_l = 0x%x\n", dlg_regs->refcyc_per_pte_group_flip_l);
+       dml_print("DML: refcyc_per_pte_group_flip_c = 0x%x\n", dlg_regs->refcyc_per_pte_group_flip_c);
+       dml_print("DML: refcyc_per_meta_chunk_flip_l = 0x%x\n", dlg_regs->refcyc_per_meta_chunk_flip_l);
+       dml_print("DML: refcyc_per_meta_chunk_flip_c = 0x%x\n", dlg_regs->refcyc_per_meta_chunk_flip_c);
+       dml_print("DML: dst_y_per_pte_row_nom_l = 0x%x\n", dlg_regs->dst_y_per_pte_row_nom_l);
+       dml_print("DML: dst_y_per_pte_row_nom_c = 0x%x\n", dlg_regs->dst_y_per_pte_row_nom_c);
+       dml_print("DML: refcyc_per_pte_group_nom_l = 0x%x\n", dlg_regs->refcyc_per_pte_group_nom_l);
+       dml_print("DML: refcyc_per_pte_group_nom_c = 0x%x\n", dlg_regs->refcyc_per_pte_group_nom_c);
+       dml_print("DML: dst_y_per_meta_row_nom_l = 0x%x\n", dlg_regs->dst_y_per_meta_row_nom_l);
+       dml_print("DML: dst_y_per_meta_row_nom_c = 0x%x\n", dlg_regs->dst_y_per_meta_row_nom_c);
+       dml_print("DML: refcyc_per_meta_chunk_nom_l = 0x%x\n", dlg_regs->refcyc_per_meta_chunk_nom_l);
+       dml_print("DML: refcyc_per_meta_chunk_nom_c = 0x%x\n", dlg_regs->refcyc_per_meta_chunk_nom_c);
+       dml_print("DML: refcyc_per_line_delivery_pre_l = 0x%x\n", dlg_regs->refcyc_per_line_delivery_pre_l);
+       dml_print("DML: refcyc_per_line_delivery_pre_c = 0x%x\n", dlg_regs->refcyc_per_line_delivery_pre_c);
+       dml_print("DML: refcyc_per_line_delivery_l = 0x%x\n", dlg_regs->refcyc_per_line_delivery_l);
+       dml_print("DML: refcyc_per_line_delivery_c = 0x%x\n", dlg_regs->refcyc_per_line_delivery_c);
+       dml_print("DML: refcyc_per_vm_group_vblank = 0x%x\n", dlg_regs->refcyc_per_vm_group_vblank);
+       dml_print("DML: refcyc_per_vm_group_flip = 0x%x\n", dlg_regs->refcyc_per_vm_group_flip);
+       dml_print("DML: refcyc_per_vm_req_vblank = 0x%x\n", dlg_regs->refcyc_per_vm_req_vblank);
+       dml_print("DML: refcyc_per_vm_req_flip = 0x%x\n", dlg_regs->refcyc_per_vm_req_flip);
+       dml_print("DML: chunk_hdl_adjust_cur0 = 0x%x\n", dlg_regs->chunk_hdl_adjust_cur0);
+       dml_print("DML: dst_y_offset_cur1 = 0x%x\n", dlg_regs->dst_y_offset_cur1);
+       dml_print("DML: chunk_hdl_adjust_cur1 = 0x%x\n", dlg_regs->chunk_hdl_adjust_cur1);
+       dml_print("DML: vready_after_vcount0 = 0x%x\n", dlg_regs->vready_after_vcount0);
+       dml_print("DML: dst_y_delta_drq_limit = 0x%x\n", dlg_regs->dst_y_delta_drq_limit);
+       dml_print("DML: refcyc_per_vm_dmdata = 0x%x\n", dlg_regs->refcyc_per_vm_dmdata);
+       dml_print("DML: ===================================== \n");
+}
+
+void dml_print_ttu_regs_st(const dml_display_ttu_regs_st *ttu_regs)
+{
+       dml_print("DML: ===================================== \n");
+       dml_print("DML: DISPLAY_TTU_REGS_ST \n");
+       dml_print("DML: qos_level_low_wm = 0x%x\n", ttu_regs->qos_level_low_wm);
+       dml_print("DML: qos_level_high_wm = 0x%x\n", ttu_regs->qos_level_high_wm);
+       dml_print("DML: min_ttu_vblank = 0x%x\n", ttu_regs->min_ttu_vblank);
+       dml_print("DML: qos_level_flip = 0x%x\n", ttu_regs->qos_level_flip);
+       dml_print("DML: refcyc_per_req_delivery_pre_l = 0x%x\n", ttu_regs->refcyc_per_req_delivery_pre_l);
+       dml_print("DML: refcyc_per_req_delivery_l = 0x%x\n", ttu_regs->refcyc_per_req_delivery_l);
+       dml_print("DML: refcyc_per_req_delivery_pre_c = 0x%x\n", ttu_regs->refcyc_per_req_delivery_pre_c);
+       dml_print("DML: refcyc_per_req_delivery_c = 0x%x\n", ttu_regs->refcyc_per_req_delivery_c);
+       dml_print("DML: refcyc_per_req_delivery_cur0 = 0x%x\n", ttu_regs->refcyc_per_req_delivery_cur0);
+       dml_print("DML: refcyc_per_req_delivery_pre_cur0 = 0x%x\n", ttu_regs->refcyc_per_req_delivery_pre_cur0);
+       dml_print("DML: refcyc_per_req_delivery_cur1 = 0x%x\n", ttu_regs->refcyc_per_req_delivery_cur1);
+       dml_print("DML: refcyc_per_req_delivery_pre_cur1 = 0x%x\n", ttu_regs->refcyc_per_req_delivery_pre_cur1);
+       dml_print("DML: qos_level_fixed_l = 0x%x\n", ttu_regs->qos_level_fixed_l);
+       dml_print("DML: qos_ramp_disable_l = 0x%x\n", ttu_regs->qos_ramp_disable_l);
+       dml_print("DML: qos_level_fixed_c = 0x%x\n", ttu_regs->qos_level_fixed_c);
+       dml_print("DML: qos_ramp_disable_c = 0x%x\n", ttu_regs->qos_ramp_disable_c);
+       dml_print("DML: qos_level_fixed_cur0 = 0x%x\n", ttu_regs->qos_level_fixed_cur0);
+       dml_print("DML: qos_ramp_disable_cur0 = 0x%x\n", ttu_regs->qos_ramp_disable_cur0);
+       dml_print("DML: qos_level_fixed_cur1 = 0x%x\n", ttu_regs->qos_level_fixed_cur1);
+       dml_print("DML: qos_ramp_disable_cur1 = 0x%x\n", ttu_regs->qos_ramp_disable_cur1);
+       dml_print("DML: ===================================== \n");
+}
+
+void dml_print_dml_policy(const struct dml_mode_eval_policy_st *policy)
+{
+       dml_print("DML: ===================================== \n");
+       dml_print("DML: DML_MODE_EVAL_POLICY_ST\n");
+       dml_print("DML: Policy: UseUnboundedRequesting = 0x%x\n", policy->UseUnboundedRequesting);
+       dml_print("DML: Policy: UseMinimumRequiredDCFCLK = 0x%x\n", policy->UseMinimumRequiredDCFCLK);
+       dml_print("DML: Policy: DRAMClockChangeRequirementFinal = 0x%x\n", policy->DRAMClockChangeRequirementFinal);
+       dml_print("DML: Policy: FCLKChangeRequirementFinal = 0x%x\n", policy->FCLKChangeRequirementFinal);
+       dml_print("DML: Policy: USRRetrainingRequiredFinal = 0x%x\n", policy->USRRetrainingRequiredFinal);
+       dml_print("DML: Policy: EnhancedPrefetchScheduleAccelerationFinal = 0x%x\n", policy->EnhancedPrefetchScheduleAccelerationFinal);
+       dml_print("DML: Policy: NomDETInKByteOverrideEnable = 0x%x\n", policy->NomDETInKByteOverrideEnable);
+       dml_print("DML: Policy: NomDETInKByteOverrideValue = 0x%x\n", policy->NomDETInKByteOverrideValue);
+       dml_print("DML: Policy: DCCProgrammingAssumesScanDirectionUnknownFinal = 0x%x\n", policy->DCCProgrammingAssumesScanDirectionUnknownFinal);
+       dml_print("DML: Policy: SynchronizeTimingsFinal = 0x%x\n", policy->SynchronizeTimingsFinal);
+       dml_print("DML: Policy: SynchronizeDRRDisplaysForUCLKPStateChangeFinal = 0x%x\n", policy->SynchronizeDRRDisplaysForUCLKPStateChangeFinal);
+       dml_print("DML: Policy: AssumeModeSupportAtMaxPwrStateEvenDRAMClockChangeNotSupported = 0x%x\n", policy->AssumeModeSupportAtMaxPwrStateEvenDRAMClockChangeNotSupported);
+       dml_print("DML: Policy: AssumeModeSupportAtMaxPwrStateEvenFClockChangeNotSupported = 0x%x\n", policy->AssumeModeSupportAtMaxPwrStateEvenFClockChangeNotSupported);
+
+       for (dml_uint_t i = 0; i < DCN_DML__NUM_PLANE; i++) {
+               dml_print("DML: i=%0d, Policy: MPCCombineUse = 0x%x\n", i, policy->MPCCombineUse[i]);
+               dml_print("DML: i=%0d, Policy: ODMUse = 0x%x\n", i, policy->ODMUse[i]);
+               dml_print("DML: i=%0d, Policy: ImmediateFlipRequirement = 0x%x\n", i, policy->ImmediateFlipRequirement[i]);
+               dml_print("DML: i=%0d, Policy: AllowForPStateChangeOrStutterInVBlank = 0x%x\n", i, policy->AllowForPStateChangeOrStutterInVBlank[i]);
+       }
+       dml_print("DML: ===================================== \n");
+}
+
+void dml_print_mode_support(struct display_mode_lib_st *mode_lib, dml_uint_t j)
+{
+       dml_print("DML: MODE SUPPORT: ===============================================\n");
+       dml_print("DML: MODE SUPPORT: Voltage State %d\n", j);
+       dml_print("DML: MODE SUPPORT:     Mode Supported              : %s\n", mode_lib->ms.support.ModeSupport[j] == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Scale Ratio And Taps                : %s\n", mode_lib->ms.support.ScaleRatioAndTapsSupport == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Source Format Pixel And Scan        : %s\n", mode_lib->ms.support.SourceFormatPixelAndScanSupport == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Viewport Size                       : %s\n", mode_lib->ms.support.ViewportSizeSupport[j] == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Link Rate Does Not Match DP Version        : %s\n", mode_lib->ms.support.LinkRateDoesNotMatchDPVersion == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Link Rate For Multistream Not Indicated    : %s\n", mode_lib->ms.support.LinkRateForMultistreamNotIndicated == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     BPP For Multi stream Not Indicated         : %s\n", mode_lib->ms.support.BPPForMultistreamNotIndicated == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Multistream With HDMI Or eDP               : %s\n", mode_lib->ms.support.MultistreamWithHDMIOreDP == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Exceeded Multistream Slots                 : %s\n", mode_lib->ms.support.ExceededMultistreamSlots == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     MSO Or ODM Split With Non DP Link          : %s\n", mode_lib->ms.support.MSOOrODMSplitWithNonDPLink == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Not Enough Lanes For MSO                   : %s\n", mode_lib->ms.support.NotEnoughLanesForMSO == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     LinkCapacitySupport                        : %s\n", mode_lib->ms.support.LinkCapacitySupport == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     P2IWith420                                 : %s\n", mode_lib->ms.support.P2IWith420 == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     DSCOnlyIfNecessaryWithBPP                  : %s\n", mode_lib->ms.support.DSCOnlyIfNecessaryWithBPP == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     DSC422NativeNotSupported                   : %s\n", mode_lib->ms.support.DSC422NativeNotSupported == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     MPCCombineMethodIncompatible               : %s\n", mode_lib->ms.support.MPCCombineMethodIncompatible == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     ODMCombineTwoToOneSupportCheckOK           : %s\n", mode_lib->ms.support.ODMCombineTwoToOneSupportCheckOK == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     ODMCombineFourToOneSupportCheckOK          : %s\n", mode_lib->ms.support.ODMCombineFourToOneSupportCheckOK == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     NotEnoughDSCUnits                          : %s\n", mode_lib->ms.support.NotEnoughDSCUnits == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     NotEnoughDSCSlices                         : %s\n", mode_lib->ms.support.NotEnoughDSCSlices == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe : %s\n", mode_lib->ms.support.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     InvalidCombinationOfMALLUseForPStateAndStaticScreen          : %s\n", mode_lib->ms.support.InvalidCombinationOfMALLUseForPStateAndStaticScreen == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     DSCCLKRequiredMoreThanSupported            : %s\n", mode_lib->ms.support.DSCCLKRequiredMoreThanSupported == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     PixelsPerLinePerDSCUnitSupport             : %s\n", mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     DTBCLKRequiredMoreThanSupported            : %s\n", mode_lib->ms.support.DTBCLKRequiredMoreThanSupported == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     InvalidCombinationOfMALLUseForPState       : %s\n", mode_lib->ms.support.InvalidCombinationOfMALLUseForPState == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified : %s\n", mode_lib->ms.support.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     ROB Support                                : %s\n", mode_lib->ms.support.ROBSupport[j] == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     DISPCLK DPPCLK Support                     : %s\n", mode_lib->ms.support.DISPCLK_DPPCLK_Support[j] == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Total Available Pipes Support              : %s\n", mode_lib->ms.support.TotalAvailablePipesSupport[j] == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Number Of OTG Support                      : %s\n", mode_lib->ms.support.NumberOfOTGSupport == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Number Of DP2p0 Support                    : %s\n", mode_lib->ms.support.NumberOfDP2p0Support == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Writeback Latency Support                  : %s\n", mode_lib->ms.support.WritebackLatencySupport == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Writeback Scale Ratio And Taps Support     : %s\n", mode_lib->ms.support.WritebackScaleRatioAndTapsSupport == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Cursor Support                             : %s\n", mode_lib->ms.support.CursorSupport == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Pitch Support                              : %s\n", mode_lib->ms.support.PitchSupport == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Viewport Exceeds Surface                   : %s\n", mode_lib->ms.support.ViewportExceedsSurface == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Prefetch Supported                         : %s\n", mode_lib->ms.support.PrefetchSupported[j] == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     VActive Bandwith Support                   : %s\n", mode_lib->ms.support.VActiveBandwithSupport[j] == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Dynamic Metadata Supported                 : %s\n", mode_lib->ms.support.DynamicMetadataSupported[j] == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Total Vertical Active Bandwidth Support    : %s\n", mode_lib->ms.support.TotalVerticalActiveBandwidthSupport[j] == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     VRatio In Prefetch Supported               : %s\n", mode_lib->ms.support.VRatioInPrefetchSupported[j] == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     PTE Buffer Size Not Exceeded               : %s\n", mode_lib->ms.support.PTEBufferSizeNotExceeded[j] == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     DCC Meta Buffer Size Not Exceeded          : %s\n", mode_lib->ms.support.DCCMetaBufferSizeNotExceeded[j] == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Non supported DSC Input BPC                : %s\n", mode_lib->ms.support.NonsupportedDSCInputBPC == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Exceeded MALL Size                         : %s\n", mode_lib->ms.support.ExceededMALLSize == false ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     Host VM or Immediate Flip Supported        : %s\n", ((mode_lib->ms.cache_display_cfg.plane.HostVMEnable == false && !mode_lib->scratch.dml_core_mode_support_locals.ImmediateFlipRequiredFinal) || mode_lib->ms.support.ImmediateFlipSupportedForState[j]) ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     dram clock change support                  : %s\n", mode_lib->scratch.dml_core_mode_support_locals.dram_clock_change_support == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     f_clock change support                     : %s\n", mode_lib->scratch.dml_core_mode_support_locals.f_clock_change_support == true ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT:     USR Retraining Support                     : %s\n", (!mode_lib->ms.policy.USRRetrainingRequiredFinal || &mode_lib->ms.support.USRRetrainingSupport[j]) ? "Supported" : "NOT Supported");
+       dml_print("DML: MODE SUPPORT: ===============================================\n");
+}
+
+void dml_print_dml_mode_support_info(const struct dml_mode_support_info_st *support, dml_bool_t fail_only)
+{
+       dml_print("DML: ===================================== \n");
+       dml_print("DML: DML_MODE_SUPPORT_INFO_ST\n");
+       if (!fail_only || support->ModeIsSupported == 0)
+               dml_print("DML: support: ModeIsSupported = 0x%x\n", support->ModeIsSupported);
+       if (!fail_only || support->ImmediateFlipSupport == 0)
+               dml_print("DML: support: ImmediateFlipSupport = 0x%x\n", support->ImmediateFlipSupport);
+       if (!fail_only || support->WritebackLatencySupport == 0)
+               dml_print("DML: support: WritebackLatencySupport = 0x%x\n", support->WritebackLatencySupport);
+       if (!fail_only || support->ScaleRatioAndTapsSupport == 0)
+               dml_print("DML: support: ScaleRatioAndTapsSupport = 0x%x\n", support->ScaleRatioAndTapsSupport);
+       if (!fail_only || support->SourceFormatPixelAndScanSupport == 0)
+               dml_print("DML: support: SourceFormatPixelAndScanSupport = 0x%x\n", support->SourceFormatPixelAndScanSupport);
+       if (!fail_only || support->MPCCombineMethodIncompatible == 1)
+               dml_print("DML: support: MPCCombineMethodIncompatible = 0x%x\n", support->MPCCombineMethodIncompatible);
+       if (!fail_only || support->P2IWith420 == 1)
+               dml_print("DML: support: P2IWith420 = 0x%x\n", support->P2IWith420);
+       if (!fail_only || support->DSCOnlyIfNecessaryWithBPP == 1)
+               dml_print("DML: support: DSCOnlyIfNecessaryWithBPP = 0x%x\n", support->DSCOnlyIfNecessaryWithBPP);
+       if (!fail_only || support->DSC422NativeNotSupported == 1)
+               dml_print("DML: support: DSC422NativeNotSupported = 0x%x\n", support->DSC422NativeNotSupported);
+       if (!fail_only || support->LinkRateDoesNotMatchDPVersion == 1)
+               dml_print("DML: support: LinkRateDoesNotMatchDPVersion = 0x%x\n", support->LinkRateDoesNotMatchDPVersion);
+       if (!fail_only || support->LinkRateForMultistreamNotIndicated == 1)
+               dml_print("DML: support: LinkRateForMultistreamNotIndicated = 0x%x\n", support->LinkRateForMultistreamNotIndicated);
+       if (!fail_only || support->BPPForMultistreamNotIndicated == 1)
+               dml_print("DML: support: BPPForMultistreamNotIndicated = 0x%x\n", support->BPPForMultistreamNotIndicated);
+       if (!fail_only || support->MultistreamWithHDMIOreDP == 1)
+               dml_print("DML: support: MultistreamWithHDMIOreDP = 0x%x\n", support->MultistreamWithHDMIOreDP);
+       if (!fail_only || support->MSOOrODMSplitWithNonDPLink == 1)
+               dml_print("DML: support: MSOOrODMSplitWithNonDPLink = 0x%x\n", support->MSOOrODMSplitWithNonDPLink);
+       if (!fail_only || support->NotEnoughLanesForMSO == 1)
+               dml_print("DML: support: NotEnoughLanesForMSO = 0x%x\n", support->NotEnoughLanesForMSO);
+       if (!fail_only || support->NumberOfOTGSupport == 0)
+               dml_print("DML: support: NumberOfOTGSupport = 0x%x\n", support->NumberOfOTGSupport);
+       if (!fail_only || support->NumberOfDP2p0Support == 0)
+               dml_print("DML: support: NumberOfDP2p0Support = 0x%x\n", support->NumberOfDP2p0Support);
+       if (!fail_only || support->NonsupportedDSCInputBPC == 1)
+               dml_print("DML: support: NonsupportedDSCInputBPC = 0x%x\n", support->NonsupportedDSCInputBPC);
+       if (!fail_only || support->WritebackScaleRatioAndTapsSupport == 0)
+               dml_print("DML: support: WritebackScaleRatioAndTapsSupport = 0x%x\n", support->WritebackScaleRatioAndTapsSupport);
+       if (!fail_only || support->CursorSupport == 0)
+               dml_print("DML: support: CursorSupport = 0x%x\n", support->CursorSupport);
+       if (!fail_only || support->PitchSupport == 0)
+               dml_print("DML: support: PitchSupport = 0x%x\n", support->PitchSupport);
+       if (!fail_only || support->ViewportExceedsSurface == 1)
+               dml_print("DML: support: ViewportExceedsSurface = 0x%x\n", support->ViewportExceedsSurface);
+       if (!fail_only || support->ExceededMALLSize == 1)
+               dml_print("DML: support: ExceededMALLSize = 0x%x\n", support->ExceededMALLSize);
+       if (!fail_only || support->EnoughWritebackUnits == 0)
+               dml_print("DML: support: EnoughWritebackUnits = 0x%x\n", support->EnoughWritebackUnits);
+       if (!fail_only || support->ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified == 1)
+               dml_print("DML: support: ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified = 0x%x\n", support->ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified);
+       if (!fail_only || support->ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe == 1)
+               dml_print("DML: support: ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe = 0x%x\n", support->ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe);
+       if (!fail_only || support->InvalidCombinationOfMALLUseForPStateAndStaticScreen == 1)
+               dml_print("DML: support: InvalidCombinationOfMALLUseForPStateAndStaticScreen = 0x%x\n", support->InvalidCombinationOfMALLUseForPStateAndStaticScreen);
+       if (!fail_only || support->InvalidCombinationOfMALLUseForPState == 1)
+               dml_print("DML: support: InvalidCombinationOfMALLUseForPState = 0x%x\n", support->InvalidCombinationOfMALLUseForPState);
+
+       if (!fail_only || support->ExceededMultistreamSlots == 1)
+               dml_print("DML: support: ExceededMultistreamSlots = 0x%x\n", support->ExceededMultistreamSlots);
+       if (!fail_only || support->ODMCombineTwoToOneSupportCheckOK == 0)
+               dml_print("DML: support: ODMCombineTwoToOneSupportCheckOK = 0x%x\n", support->ODMCombineTwoToOneSupportCheckOK);
+       if (!fail_only || support->ODMCombineFourToOneSupportCheckOK == 0)
+               dml_print("DML: support: ODMCombineFourToOneSupportCheckOK = 0x%x\n", support->ODMCombineFourToOneSupportCheckOK);
+       if (!fail_only || support->NotEnoughDSCUnits == 1)
+               dml_print("DML: support: NotEnoughDSCUnits = 0x%x\n", support->NotEnoughDSCUnits);
+       if (!fail_only || support->NotEnoughDSCSlices == 1)
+               dml_print("DML: support: NotEnoughDSCSlices = 0x%x\n", support->NotEnoughDSCSlices);
+       if (!fail_only || support->PixelsPerLinePerDSCUnitSupport == 0)
+               dml_print("DML: support: PixelsPerLinePerDSCUnitSupport = 0x%x\n", support->PixelsPerLinePerDSCUnitSupport);
+       if (!fail_only || support->DSCCLKRequiredMoreThanSupported == 1)
+               dml_print("DML: support: DSCCLKRequiredMoreThanSupported = 0x%x\n", support->DSCCLKRequiredMoreThanSupported);
+       if (!fail_only || support->DTBCLKRequiredMoreThanSupported == 1)
+               dml_print("DML: support: DTBCLKRequiredMoreThanSupported = 0x%x\n", support->DTBCLKRequiredMoreThanSupported);
+       if (!fail_only || support->LinkCapacitySupport == 0)
+               dml_print("DML: support: LinkCapacitySupport = 0x%x\n", support->LinkCapacitySupport);
+
+       for (dml_uint_t j = 0; j < 2; j++) {
+               if (!fail_only || support->DRAMClockChangeSupport[j] == dml_dram_clock_change_unsupported)
+                       dml_print("DML: support: combine=%d, DRAMClockChangeSupport = %d\n", j, support->DRAMClockChangeSupport[j]);
+               if (!fail_only || support->FCLKChangeSupport[j] == dml_fclock_change_unsupported)
+                       dml_print("DML: support: combine=%d, FCLKChangeSupport = %d\n", j, support->FCLKChangeSupport[j]);
+               if (!fail_only || support->ROBSupport[j] == 0)
+                       dml_print("DML: support: combine=%d, ROBSupport = %d\n", j, support->ROBSupport[j]);
+               if (!fail_only || support->PTEBufferSizeNotExceeded[j] == 0)
+                       dml_print("DML: support: combine=%d, PTEBufferSizeNotExceeded = %d\n", j, support->PTEBufferSizeNotExceeded[j]);
+               if (!fail_only || support->DCCMetaBufferSizeNotExceeded[j] == 0)
+                       dml_print("DML: support: combine=%d, DCCMetaBufferSizeNotExceeded = %d\n", j, support->DCCMetaBufferSizeNotExceeded[j]);
+               if (!fail_only || support->TotalVerticalActiveBandwidthSupport[j] == 0)
+                       dml_print("DML: support: combine=%d, TotalVerticalActiveBandwidthSupport = %d\n", j, support->TotalVerticalActiveBandwidthSupport[j]);
+               if (!fail_only || support->USRRetrainingSupport[j] == 0)
+                       dml_print("DML: support: combine=%d, USRRetrainingSupport = %d\n", j, support->USRRetrainingSupport[j]);
+               if (!fail_only || support->VActiveBandwithSupport[j] == 0)
+                       dml_print("DML: support: combine=%d, VActiveBandwithSupport = %d\n", j, support->VActiveBandwithSupport[j]);
+               if (!fail_only || support->PrefetchSupported[j] == 0)
+                       dml_print("DML: support: combine=%d, PrefetchSupported = %d\n", j, support->PrefetchSupported[j]);
+               if (!fail_only || support->DynamicMetadataSupported[j] == 0)
+                       dml_print("DML: support: combine=%d, DynamicMetadataSupported = %d\n", j, support->DynamicMetadataSupported[j]);
+               if (!fail_only || support->VRatioInPrefetchSupported[j] == 0)
+                       dml_print("DML: support: combine=%d, VRatioInPrefetchSupported = %d\n", j, support->VRatioInPrefetchSupported[j]);
+               if (!fail_only || support->DISPCLK_DPPCLK_Support[j] == 0)
+                       dml_print("DML: support: combine=%d, DISPCLK_DPPCLK_Support = %d\n", j, support->DISPCLK_DPPCLK_Support[j]);
+               if (!fail_only || support->TotalAvailablePipesSupport[j] == 0)
+                       dml_print("DML: support: combine=%d, TotalAvailablePipesSupport = %d\n", j, support->TotalAvailablePipesSupport[j]);
+               if (!fail_only || support->ModeSupport[j] == 0)
+                       dml_print("DML: support: combine=%d, ModeSupport = %d\n", j, support->ModeSupport[j]);
+               if (!fail_only || support->ViewportSizeSupport[j] == 0)
+                       dml_print("DML: support: combine=%d, ViewportSizeSupport = %d\n", j, support->ViewportSizeSupport[j]);
+               if (!fail_only || support->ImmediateFlipSupportedForState[j] == 0)
+                       dml_print("DML: support: combine=%d, ImmediateFlipSupportedForState = %d\n", j, support->ImmediateFlipSupportedForState[j]);
+       }
+}
+
+void dml_print_dml_display_cfg_timing(const struct dml_timing_cfg_st *timing, dml_uint_t num_plane)
+{
+       for (dml_uint_t i = 0; i < num_plane; i++) {
+               dml_print("DML: timing_cfg: plane=%d, HTotal = %d\n", i, timing->HTotal[i]);
+               dml_print("DML: timing_cfg: plane=%d, VTotal = %d\n", i, timing->VTotal[i]);
+               dml_print("DML: timing_cfg: plane=%d, HActive = %d\n", i, timing->HActive[i]);
+               dml_print("DML: timing_cfg: plane=%d, VActive = %d\n", i, timing->VActive[i]);
+               dml_print("DML: timing_cfg: plane=%d, VFrontPorch = %d\n", i, timing->VFrontPorch[i]);
+               dml_print("DML: timing_cfg: plane=%d, VBlankNom = %d\n", i, timing->VBlankNom[i]);
+               dml_print("DML: timing_cfg: plane=%d, RefreshRate = %d\n", i, timing->RefreshRate[i]);
+               dml_print("DML: timing_cfg: plane=%d, PixelClock = %f\n", i, timing->PixelClock[i]);
+               dml_print("DML: timing_cfg: plane=%d, Interlace = %d\n", i, timing->Interlace[i]);
+               dml_print("DML: timing_cfg: plane=%d, DRRDisplay = %d\n", i, timing->DRRDisplay[i]);
+       }
+}
+
+void dml_print_dml_display_cfg_plane(const struct dml_plane_cfg_st *plane, dml_uint_t num_plane)
+{
+       dml_print("DML: plane_cfg: num_plane = %d\n", num_plane);
+       dml_print("DML: plane_cfg: GPUVMEnable = %d\n", plane->GPUVMEnable);
+       dml_print("DML: plane_cfg: HostVMEnable = %d\n", plane->HostVMEnable);
+       dml_print("DML: plane_cfg: GPUVMMaxPageTableLevels = %d\n", plane->GPUVMMaxPageTableLevels);
+       dml_print("DML: plane_cfg: HostVMMaxPageTableLevels = %d\n", plane->HostVMMaxPageTableLevels);
+
+       for (dml_uint_t i = 0; i < num_plane; i++) {
+               dml_print("DML: plane_cfg: plane=%d, GPUVMMinPageSizeKBytes = %d\n", i, plane->GPUVMMinPageSizeKBytes[i]);
+               dml_print("DML: plane_cfg: plane=%d, ForceOneRowForFrame = %d\n", i, plane->ForceOneRowForFrame[i]);
+               dml_print("DML: plane_cfg: plane=%d, PTEBufferModeOverrideEn = %d\n", i, plane->PTEBufferModeOverrideEn[i]);
+               dml_print("DML: plane_cfg: plane=%d, PTEBufferMode = %d\n", i, plane->PTEBufferMode[i]);
+               dml_print("DML: plane_cfg: plane=%d, DETSizeOverride = %d\n", i, plane->DETSizeOverride[i]);
+               dml_print("DML: plane_cfg: plane=%d, UseMALLForStaticScreen = %d\n", i, plane->UseMALLForStaticScreen[i]);
+               dml_print("DML: plane_cfg: plane=%d, UseMALLForPStateChange = %d\n", i, plane->UseMALLForPStateChange[i]);
+               dml_print("DML: plane_cfg: plane=%d, BlendingAndTiming = %d\n", i, plane->BlendingAndTiming[i]);
+               dml_print("DML: plane_cfg: plane=%d, ViewportWidth = %d\n", i, plane->ViewportWidth[i]);
+               dml_print("DML: plane_cfg: plane=%d, ViewportHeight = %d\n", i, plane->ViewportHeight[i]);
+               dml_print("DML: plane_cfg: plane=%d, ViewportWidthChroma = %d\n", i, plane->ViewportWidthChroma[i]);
+               dml_print("DML: plane_cfg: plane=%d, ViewportHeightChroma = %d\n", i, plane->ViewportHeightChroma[i]);
+               dml_print("DML: plane_cfg: plane=%d, ViewportXStart = %d\n", i, plane->ViewportXStart[i]);
+               dml_print("DML: plane_cfg: plane=%d, ViewportXStartC = %d\n", i, plane->ViewportXStartC[i]);
+               dml_print("DML: plane_cfg: plane=%d, ViewportYStart = %d\n", i, plane->ViewportYStart[i]);
+               dml_print("DML: plane_cfg: plane=%d, ViewportYStartC = %d\n", i, plane->ViewportYStartC[i]);
+               dml_print("DML: plane_cfg: plane=%d, ViewportStationary = %d\n", i, plane->ViewportStationary[i]);
+               dml_print("DML: plane_cfg: plane=%d, ScalerEnabled = %d\n", i, plane->ScalerEnabled[i]);
+               dml_print("DML: plane_cfg: plane=%d, HRatio = %3.2f\n", i, plane->HRatio[i]);
+               dml_print("DML: plane_cfg: plane=%d, VRatio = %3.2f\n", i, plane->VRatio[i]);
+               dml_print("DML: plane_cfg: plane=%d, HRatioChroma = %3.2f\n", i, plane->HRatioChroma[i]);
+               dml_print("DML: plane_cfg: plane=%d, VRatioChroma = %3.2f\n", i, plane->VRatioChroma[i]);
+               dml_print("DML: plane_cfg: plane=%d, HTaps = %d\n", i, plane->HTaps[i]);
+               dml_print("DML: plane_cfg: plane=%d, VTaps = %d\n", i, plane->VTaps[i]);
+               dml_print("DML: plane_cfg: plane=%d, HTapsChroma = %d\n", i, plane->HTapsChroma[i]);
+               dml_print("DML: plane_cfg: plane=%d, VTapsChroma = %d\n", i, plane->VTapsChroma[i]);
+               dml_print("DML: plane_cfg: plane=%d, LBBitPerPixel = %d\n", i, plane->LBBitPerPixel[i]);
+               dml_print("DML: plane_cfg: plane=%d, SourceScan = %d\n", i, plane->SourceScan[i]);
+               dml_print("DML: plane_cfg: plane=%d, ScalerRecoutWidth = %d\n", i, plane->ScalerRecoutWidth[i]);
+               dml_print("DML: plane_cfg: plane=%d, NumberOfCursors = %d\n", i, plane->NumberOfCursors[i]);
+               dml_print("DML: plane_cfg: plane=%d, CursorWidth = %d\n", i, plane->CursorWidth[i]);
+               dml_print("DML: plane_cfg: plane=%d, CursorBPP = %d\n", i, plane->CursorBPP[i]);
+
+               dml_print("DML: plane_cfg: plane=%d, DynamicMetadataEnable = %d\n", i, plane->DynamicMetadataEnable[i]);
+               dml_print("DML: plane_cfg: plane=%d, DynamicMetadataLinesBeforeActiveRequired = %d\n", i, plane->DynamicMetadataLinesBeforeActiveRequired[i]);
+               dml_print("DML: plane_cfg: plane=%d, DynamicMetadataTransmittedBytes = %d\n", i, plane->DynamicMetadataTransmittedBytes[i]);
+       }
+}
+
+void dml_print_dml_display_cfg_surface(const struct dml_surface_cfg_st *surface, dml_uint_t num_plane)
+{
+       for (dml_uint_t i = 0; i < num_plane; i++) {
+               dml_print("DML: surface_cfg: plane=%d, PitchY = %d\n", i, surface->PitchY[i]);
+               dml_print("DML: surface_cfg: plane=%d, SurfaceWidthY = %d\n", i, surface->SurfaceWidthY[i]);
+               dml_print("DML: surface_cfg: plane=%d, SurfaceHeightY = %d\n", i, surface->SurfaceHeightY[i]);
+               dml_print("DML: surface_cfg: plane=%d, PitchC = %d\n", i, surface->PitchC[i]);
+               dml_print("DML: surface_cfg: plane=%d, SurfaceWidthC = %d\n", i, surface->SurfaceWidthC[i]);
+               dml_print("DML: surface_cfg: plane=%d, SurfaceHeightC = %d\n", i, surface->SurfaceHeightC[i]);
+               dml_print("DML: surface_cfg: plane=%d, DCCEnable = %d\n", i, surface->DCCEnable[i]);
+               dml_print("DML: surface_cfg: plane=%d, DCCMetaPitchY = %d\n", i, surface->DCCMetaPitchY[i]);
+               dml_print("DML: surface_cfg: plane=%d, DCCMetaPitchC = %d\n", i, surface->DCCMetaPitchC[i]);
+               dml_print("DML: surface_cfg: plane=%d, DCCRateLuma = %f\n", i, surface->DCCRateLuma[i]);
+               dml_print("DML: surface_cfg: plane=%d, DCCRateChroma = %f\n", i, surface->DCCRateChroma[i]);
+               dml_print("DML: surface_cfg: plane=%d, DCCFractionOfZeroSizeRequestsLuma = %f\n", i, surface->DCCFractionOfZeroSizeRequestsLuma[i]);
+               dml_print("DML: surface_cfg: plane=%d, DCCFractionOfZeroSizeRequestsChroma= %f\n", i, surface->DCCFractionOfZeroSizeRequestsChroma[i]);
+       }
+}
+
+void dml_print_dml_display_cfg_hw_resource(const struct dml_hw_resource_st *hw, dml_uint_t num_plane)
+{
+       for (dml_uint_t i = 0; i < num_plane; i++) {
+               dml_print("DML: hw_resource: plane=%d, ODMMode = %d\n", i, hw->ODMMode[i]);
+               dml_print("DML: hw_resource: plane=%d, DPPPerSurface = %d\n", i, hw->DPPPerSurface[i]);
+               dml_print("DML: hw_resource: plane=%d, DSCEnabled = %d\n", i, hw->DSCEnabled[i]);
+               dml_print("DML: hw_resource: plane=%d, NumberOfDSCSlices = %d\n", i, hw->NumberOfDSCSlices[i]);
+       }
+       dml_print("DML: hw_resource: DLGRefClkFreqMHz   = %f\n", hw->DLGRefClkFreqMHz);
+}
+
+__DML_DLL_EXPORT__ void dml_print_soc_state_bounding_box(const struct soc_state_bounding_box_st *state)
+{
+       dml_print("DML: state_bbox: socclk_mhz = %f\n", state->socclk_mhz);
+       dml_print("DML: state_bbox: dscclk_mhz = %f\n", state->dscclk_mhz);
+       dml_print("DML: state_bbox: phyclk_mhz = %f\n", state->phyclk_mhz);
+       dml_print("DML: state_bbox: phyclk_d18_mhz = %f\n", state->phyclk_d18_mhz);
+       dml_print("DML: state_bbox: phyclk_d32_mhz = %f\n", state->phyclk_d32_mhz);
+       dml_print("DML: state_bbox: dtbclk_mhz = %f\n", state->dtbclk_mhz);
+       dml_print("DML: state_bbox: dispclk_mhz = %f\n", state->dispclk_mhz);
+       dml_print("DML: state_bbox: dppclk_mhz = %f\n", state->dppclk_mhz);
+       dml_print("DML: state_bbox: fabricclk_mhz = %f\n", state->fabricclk_mhz);
+       dml_print("DML: state_bbox: dcfclk_mhz = %f\n", state->dcfclk_mhz);
+       dml_print("DML: state_bbox: dram_speed_mts = %f\n", state->dram_speed_mts);
+       dml_print("DML: state_bbox: urgent_latency_pixel_data_only_us = %f\n", state->urgent_latency_pixel_data_only_us);
+       dml_print("DML: state_bbox: urgent_latency_pixel_mixed_with_vm_data_us = %f\n", state->urgent_latency_pixel_mixed_with_vm_data_us);
+       dml_print("DML: state_bbox: urgent_latency_vm_data_only_us = %f\n", state->urgent_latency_vm_data_only_us);
+       dml_print("DML: state_bbox: writeback_latency_us = %f\n", state->writeback_latency_us);
+       dml_print("DML: state_bbox: urgent_latency_adjustment_fabric_clock_component_us = %f\n", state->urgent_latency_adjustment_fabric_clock_component_us);
+       dml_print("DML: state_bbox: urgent_latency_adjustment_fabric_clock_reference_mhz= %f\n", state->urgent_latency_adjustment_fabric_clock_reference_mhz);
+       dml_print("DML: state_bbox: sr_exit_time_us = %f\n", state->sr_exit_time_us);
+       dml_print("DML: state_bbox: sr_enter_plus_exit_time_us = %f\n", state->sr_enter_plus_exit_time_us);
+       dml_print("DML: state_bbox: sr_exit_z8_time_us = %f\n", state->sr_exit_z8_time_us);
+       dml_print("DML: state_bbox: sr_enter_plus_exit_z8_time_us = %f\n", state->sr_enter_plus_exit_z8_time_us);
+       dml_print("DML: state_bbox: dram_clock_change_latency_us = %f\n", state->dram_clock_change_latency_us);
+       dml_print("DML: state_bbox: fclk_change_latency_us = %f\n", state->fclk_change_latency_us);
+       dml_print("DML: state_bbox: usr_retraining_latency_us = %f\n", state->usr_retraining_latency_us);
+       dml_print("DML: state_bbox: use_ideal_dram_bw_strobe = %d\n", state->use_ideal_dram_bw_strobe);
+}
+
+__DML_DLL_EXPORT__ void dml_print_soc_bounding_box(const struct soc_bounding_box_st *soc)
+{
+       dml_print("DML: soc_bbox: dprefclk_mhz = %f\n", soc->dprefclk_mhz);
+       dml_print("DML: soc_bbox: xtalclk_mhz = %f\n", soc->xtalclk_mhz);
+       dml_print("DML: soc_bbox: pcierefclk_mhz = %f\n", soc->pcierefclk_mhz);
+       dml_print("DML: soc_bbox: refclk_mhz = %f\n", soc->refclk_mhz);
+       dml_print("DML: soc_bbox: amclk_mhz = %f\n", soc->amclk_mhz);
+
+       dml_print("DML: soc_bbox: max_outstanding_reqs = %f\n", soc->max_outstanding_reqs);
+       dml_print("DML: soc_bbox: pct_ideal_sdp_bw_after_urgent = %f\n", soc->pct_ideal_sdp_bw_after_urgent);
+       dml_print("DML: soc_bbox: pct_ideal_fabric_bw_after_urgent = %f\n", soc->pct_ideal_fabric_bw_after_urgent);
+       dml_print("DML: soc_bbox: pct_ideal_dram_bw_after_urgent_pixel_only = %f\n", soc->pct_ideal_dram_bw_after_urgent_pixel_only);
+       dml_print("DML: soc_bbox: pct_ideal_dram_bw_after_urgent_pixel_and_vm = %f\n", soc->pct_ideal_dram_bw_after_urgent_pixel_and_vm);
+       dml_print("DML: soc_bbox: pct_ideal_dram_bw_after_urgent_vm_only = %f\n", soc->pct_ideal_dram_bw_after_urgent_vm_only);
+       dml_print("DML: soc_bbox: pct_ideal_dram_bw_after_urgent_strobe = %f\n", soc->pct_ideal_dram_bw_after_urgent_strobe);
+       dml_print("DML: soc_bbox: max_avg_sdp_bw_use_normal_percent = %f\n", soc->max_avg_sdp_bw_use_normal_percent);
+       dml_print("DML: soc_bbox: max_avg_fabric_bw_use_normal_percent = %f\n", soc->max_avg_fabric_bw_use_normal_percent);
+       dml_print("DML: soc_bbox: max_avg_dram_bw_use_normal_percent = %f\n", soc->max_avg_dram_bw_use_normal_percent);
+       dml_print("DML: soc_bbox: max_avg_dram_bw_use_normal_strobe_percent = %f\n", soc->max_avg_dram_bw_use_normal_strobe_percent);
+       dml_print("DML: soc_bbox: round_trip_ping_latency_dcfclk_cycles = %d\n", soc->round_trip_ping_latency_dcfclk_cycles);
+       dml_print("DML: soc_bbox: urgent_out_of_order_return_per_channel_pixel_only_bytes = %d\n", soc->urgent_out_of_order_return_per_channel_pixel_only_bytes);
+       dml_print("DML: soc_bbox: urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = %d\n", soc->urgent_out_of_order_return_per_channel_pixel_and_vm_bytes);
+       dml_print("DML: soc_bbox: urgent_out_of_order_return_per_channel_vm_only_bytes = %d\n", soc->urgent_out_of_order_return_per_channel_vm_only_bytes);
+       dml_print("DML: soc_bbox: num_chans = %d\n", soc->num_chans);
+       dml_print("DML: soc_bbox: return_bus_width_bytes = %d\n", soc->return_bus_width_bytes);
+       dml_print("DML: soc_bbox: dram_channel_width_bytes = %d\n", soc->dram_channel_width_bytes);
+       dml_print("DML: soc_bbox: fabric_datapath_to_dcn_data_return_bytes = %d\n", soc->fabric_datapath_to_dcn_data_return_bytes);
+       dml_print("DML: soc_bbox: hostvm_min_page_size_kbytes = %d\n", soc->hostvm_min_page_size_kbytes);
+       dml_print("DML: soc_bbox: gpuvm_min_page_size_kbytes = %d\n", soc->gpuvm_min_page_size_kbytes);
+       dml_print("DML: soc_bbox: phy_downspread_percent = %f\n", soc->phy_downspread_percent);
+       dml_print("DML: soc_bbox: dcn_downspread_percent = %f\n", soc->dcn_downspread_percent);
+       dml_print("DML: soc_bbox: smn_latency_us = %f\n", soc->smn_latency_us);
+       dml_print("DML: soc_bbox: mall_allocated_for_dcn_mbytes = %d\n", soc->mall_allocated_for_dcn_mbytes);
+       dml_print("DML: soc_bbox: dispclk_dppclk_vco_speed_mhz = %f\n", soc->dispclk_dppclk_vco_speed_mhz);
+       dml_print("DML: soc_bbox: do_urgent_latency_adjustment = %d\n", soc->do_urgent_latency_adjustment);
+}
+
+__DML_DLL_EXPORT__ void dml_print_clk_cfg(const struct dml_clk_cfg_st *clk_cfg)
+{
+       dml_print("DML: clk_cfg: 0-use_required, 1-use pipe.clks_cfg, 2-use state bbox\n");
+       dml_print("DML: clk_cfg: dcfclk_option = %d\n", clk_cfg->dcfclk_option);
+       dml_print("DML: clk_cfg: dispclk_option = %d\n", clk_cfg->dispclk_option);
+
+       dml_print("DML: clk_cfg: dcfclk_freq_mhz = %f\n", clk_cfg->dcfclk_freq_mhz);
+       dml_print("DML: clk_cfg: dispclk_freq_mhz = %f\n", clk_cfg->dispclk_freq_mhz);
+
+       for (dml_uint_t i = 0; i < DCN_DML__NUM_PLANE; i++) {
+               dml_print("DML: clk_cfg: i=%d, dppclk_option = %d\n", i, clk_cfg->dppclk_option[i]);
+               dml_print("DML: clk_cfg: i=%d, dppclk_freq_mhz = %f\n", i, clk_cfg->dppclk_freq_mhz[i]);
+       }
+}
+
+dml_bool_t dml_is_vertical_rotation(enum dml_rotation_angle Scan)
+{
+       dml_bool_t is_vert = false;
+       if (Scan == dml_rotation_90 || Scan == dml_rotation_90m || Scan == dml_rotation_270 || Scan == dml_rotation_270m) {
+               is_vert = true;
+       } else {
+               is_vert = false;
+       }
+       return is_vert;
+} // dml_is_vertical_rotation
+
+dml_uint_t dml_get_cursor_bit_per_pixel(enum dml_cursor_bpp ebpp)
+{
+       switch (ebpp) {
+       case dml_cur_2bit:
+               return 2;
+       case dml_cur_32bit:
+               return 32;
+       case dml_cur_64bit:
+               return 64;
+       default:
+               return 0;
+       }
+}
+
+/// @brief Determine the physical pipe to logical plane mapping using the display_cfg
+dml_uint_t dml_get_num_active_planes(const struct dml_display_cfg_st *display_cfg)
+{
+       dml_uint_t num_active_planes = 0;
+
+       for (dml_uint_t k = 0; k < __DML_NUM_PLANES__; k++) {
+               if (display_cfg->plane.ViewportWidth[k] > 0)
+               num_active_planes = num_active_planes + 1;
+       }
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: num_active_planes = %d\n", __func__, num_active_planes);
+#endif
+       return num_active_planes;
+}
+
+/// @brief Determine the physical pipe to logical plane mapping using the display_cfg
+dml_uint_t dml_get_num_active_pipes(const struct dml_display_cfg_st *display_cfg)
+{
+       dml_uint_t num_active_pipes = 0;
+
+       for (dml_uint_t j = 0; j < dml_get_num_active_planes(display_cfg); j++) {
+               num_active_pipes = num_active_pipes + display_cfg->hw.DPPPerSurface[j];
+       }
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML::%s: num_active_pipes    = %d\n", __func__, num_active_pipes);
+#endif
+       return num_active_pipes;
+}
+
+dml_uint_t dml_get_plane_idx(const struct display_mode_lib_st *mode_lib, dml_uint_t pipe_idx)
+{
+       dml_uint_t plane_idx = mode_lib->mp.pipe_plane[pipe_idx];
+       return plane_idx;
+}
+
+dml_uint_t dml_get_pipe_idx(const struct display_mode_lib_st *mode_lib, dml_uint_t plane_idx)
+{
+       dml_uint_t     pipe_idx = 0;
+       dml_bool_t     pipe_found = 0;
+
+       ASSERT(plane_idx < __DML_NUM_PLANES__);
+
+       for (dml_uint_t i = 0; i < __DML_NUM_PLANES__; i++) {
+               if (plane_idx == mode_lib->mp.pipe_plane[i]) {
+                       pipe_idx = i;
+                       pipe_found = 1;
+                       break;
+               }
+       }
+       ASSERT(pipe_found != 0);
+
+       return pipe_idx;
+}
+
+void dml_calc_pipe_plane_mapping(const struct dml_hw_resource_st *hw, dml_uint_t *pipe_plane)
+{
+       dml_uint_t pipe_idx = 0;
+
+       for (dml_uint_t k = 0; k < __DML_NUM_PLANES__; ++k) {
+               pipe_plane[k] = __DML_PIPE_NO_PLANE__;
+       }
+
+       for (dml_uint_t plane_idx = 0; plane_idx < __DML_NUM_PLANES__; plane_idx++) {
+               for (dml_uint_t i = 0; i < hw->DPPPerSurface[plane_idx]; i++) {
+                       pipe_plane[pipe_idx] = plane_idx;
+                       pipe_idx++;
+               }
+       }
+}
+
+
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_util.h b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_util.h
new file mode 100644 (file)
index 0000000..fb74385
--- /dev/null
@@ -0,0 +1,74 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DISPLAY_MODE_UTIL_H__
+#define __DISPLAY_MODE_UTIL_H__
+
+#include "display_mode_core_structs.h"
+#include "cmntypes.h"
+
+
+#include "dml_assert.h"
+#include "dml_logging.h"
+
+__DML_DLL_EXPORT__ dml_bool_t dml_util_is_420(enum dml_source_format_class source_format);
+__DML_DLL_EXPORT__ dml_float_t dml_ceil(dml_float_t x, dml_float_t granularity);
+__DML_DLL_EXPORT__ dml_float_t dml_floor(dml_float_t x, dml_float_t granularity);
+__DML_DLL_EXPORT__ dml_float_t dml_min(dml_float_t x, dml_float_t y);
+__DML_DLL_EXPORT__ dml_float_t dml_min3(dml_float_t x, dml_float_t y, dml_float_t z);
+__DML_DLL_EXPORT__ dml_float_t dml_min4(dml_float_t x, dml_float_t y, dml_float_t z, dml_float_t w);
+__DML_DLL_EXPORT__ dml_float_t dml_max(dml_float_t x, dml_float_t y);
+__DML_DLL_EXPORT__ dml_float_t dml_max3(dml_float_t x, dml_float_t y, dml_float_t z);
+__DML_DLL_EXPORT__ dml_float_t dml_max4(dml_float_t a, dml_float_t b, dml_float_t c, dml_float_t d);
+__DML_DLL_EXPORT__ dml_float_t dml_max5(dml_float_t a, dml_float_t b, dml_float_t c, dml_float_t d, dml_float_t e);
+__DML_DLL_EXPORT__ dml_float_t dml_log(dml_float_t x, dml_float_t base);
+__DML_DLL_EXPORT__ dml_float_t dml_log2(dml_float_t x);
+__DML_DLL_EXPORT__ dml_float_t dml_round(dml_float_t val, dml_bool_t bankers_rounding);
+__DML_DLL_EXPORT__ dml_float_t dml_pow(dml_float_t base, int exp);
+__DML_DLL_EXPORT__ dml_uint_t dml_round_to_multiple(dml_uint_t num, dml_uint_t multiple, dml_bool_t up);
+__DML_DLL_EXPORT__ dml_bool_t dml_is_vertical_rotation(enum dml_rotation_angle scan);
+__DML_DLL_EXPORT__ dml_uint_t  dml_get_cursor_bit_per_pixel(enum dml_cursor_bpp ebpp);
+__DML_DLL_EXPORT__ void dml_print_data_rq_regs_st(const dml_display_plane_rq_regs_st *data_rq_regs);
+__DML_DLL_EXPORT__ void dml_print_rq_regs_st(const dml_display_rq_regs_st *rq_regs);
+__DML_DLL_EXPORT__ void dml_print_dlg_regs_st(const dml_display_dlg_regs_st *dlg_regs);
+__DML_DLL_EXPORT__ void dml_print_ttu_regs_st(const dml_display_ttu_regs_st *ttu_regs);
+__DML_DLL_EXPORT__ void dml_print_dml_policy(const struct dml_mode_eval_policy_st *policy);
+__DML_DLL_EXPORT__ void dml_print_mode_support(struct display_mode_lib_st *mode_lib, dml_uint_t j);
+__DML_DLL_EXPORT__ void dml_print_dml_mode_support_info(const struct dml_mode_support_info_st *support, dml_bool_t fail_only);
+__DML_DLL_EXPORT__ void dml_print_dml_display_cfg_timing(const struct dml_timing_cfg_st *timing, dml_uint_t num_plane);
+__DML_DLL_EXPORT__ void dml_print_dml_display_cfg_plane(const struct dml_plane_cfg_st *plane, dml_uint_t num_plane);
+__DML_DLL_EXPORT__ void dml_print_dml_display_cfg_surface(const struct dml_surface_cfg_st *surface, dml_uint_t num_plane);
+__DML_DLL_EXPORT__ void dml_print_dml_display_cfg_hw_resource(const struct dml_hw_resource_st *hw, dml_uint_t num_plane);
+__DML_DLL_EXPORT__ void dml_print_soc_state_bounding_box(const struct soc_state_bounding_box_st *state);
+__DML_DLL_EXPORT__ void dml_print_soc_bounding_box(const struct soc_bounding_box_st *soc);
+__DML_DLL_EXPORT__ void dml_print_clk_cfg(const struct dml_clk_cfg_st *clk_cfg);
+
+__DML_DLL_EXPORT__ dml_uint_t dml_get_num_active_planes(const struct dml_display_cfg_st *display_cfg);
+__DML_DLL_EXPORT__ dml_uint_t dml_get_num_active_pipes(const struct dml_display_cfg_st *display_cfg);
+__DML_DLL_EXPORT__ dml_uint_t dml_get_plane_idx(const struct display_mode_lib_st *mode_lib, dml_uint_t pipe_idx);
+__DML_DLL_EXPORT__ dml_uint_t dml_get_pipe_idx(const struct display_mode_lib_st *mode_lib, dml_uint_t plane_idx);
+__DML_DLL_EXPORT__ void       dml_calc_pipe_plane_mapping(const struct dml_hw_resource_st *hw, dml_uint_t *pipe_plane);
+
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c
new file mode 100644 (file)
index 0000000..7fd0e1c
--- /dev/null
@@ -0,0 +1,734 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "dml2_mall_phantom.h"
+
+#include "dml2_dc_types.h"
+#include "dml2_internal_types.h"
+#include "dml2_utils.h"
+#include "dml2_dc_resource_mgmt.h"
+
+#define MAX_ODM_FACTOR 4
+#define MAX_MPCC_FACTOR 4
+
+struct dc_plane_pipe_pool {
+       int pipes_assigned_to_plane[MAX_ODM_FACTOR][MAX_MPCC_FACTOR];
+       bool pipe_used[MAX_ODM_FACTOR][MAX_MPCC_FACTOR];
+       int num_pipes_assigned_to_plane_for_mpcc_combine;
+       int num_pipes_assigned_to_plane_for_odm_combine;
+};
+
+struct dc_pipe_mapping_scratch {
+       struct {
+               unsigned int odm_factor;
+               unsigned int odm_slice_end_x[MAX_PIPES];
+               struct pipe_ctx *next_higher_pipe_for_odm_slice[MAX_PIPES];
+       } odm_info;
+       struct {
+               unsigned int mpc_factor;
+               struct pipe_ctx *prev_odm_pipe;
+       } mpc_info;
+
+       struct dc_plane_pipe_pool pipe_pool;
+};
+
+static bool get_plane_id(const struct dc_state *state, const struct dc_plane_state *plane, unsigned int *plane_id)
+{
+       int i, j;
+
+       if (!plane_id)
+               return false;
+
+       for (i = 0; i < state->stream_count; i++) {
+               for (j = 0; j < state->stream_status[i].plane_count; j++) {
+                       if (state->stream_status[i].plane_states[j] == plane) {
+                               *plane_id = (i << 16) | j;
+                               return true;
+                       }
+               }
+       }
+
+       return false;
+}
+
+static int find_disp_cfg_idx_by_plane_id(struct dml2_dml_to_dc_pipe_mapping *mapping, unsigned int plane_id)
+{
+       int i;
+
+       for (i = 0; i < __DML2_WRAPPER_MAX_STREAMS_PLANES__; i++) {
+               if (mapping->disp_cfg_to_plane_id_valid[i] && mapping->disp_cfg_to_plane_id[i] == plane_id)
+                       return  i;
+       }
+
+       return -1;
+}
+
+static int find_disp_cfg_idx_by_stream_id(struct dml2_dml_to_dc_pipe_mapping *mapping, unsigned int stream_id)
+{
+       int i;
+
+       for (i = 0; i < __DML2_WRAPPER_MAX_STREAMS_PLANES__; i++) {
+               if (mapping->disp_cfg_to_stream_id_valid[i] && mapping->disp_cfg_to_stream_id[i] == stream_id)
+                       return  i;
+       }
+
+       return -1;
+}
+
+// The master pipe of a stream is defined as the top pipe in odm slice 0
+static struct pipe_ctx *find_master_pipe_of_stream(struct dml2_context *ctx, struct dc_state *state, unsigned int stream_id)
+{
+       int i;
+
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               if (state->res_ctx.pipe_ctx[i].stream && state->res_ctx.pipe_ctx[i].stream->stream_id == stream_id) {
+                       if (!state->res_ctx.pipe_ctx[i].prev_odm_pipe && !state->res_ctx.pipe_ctx[i].top_pipe)
+                               return &state->res_ctx.pipe_ctx[i];
+               }
+       }
+
+       return NULL;
+}
+
+static struct pipe_ctx *find_master_pipe_of_plane(struct dml2_context *ctx, struct dc_state *state, unsigned int plane_id)
+{
+       int i;
+       unsigned int plane_id_assigned_to_pipe;
+
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               if (state->res_ctx.pipe_ctx[i].plane_state && get_plane_id(state, state->res_ctx.pipe_ctx[i].plane_state, &plane_id_assigned_to_pipe)) {
+                       if (plane_id_assigned_to_pipe == plane_id)
+                               return &state->res_ctx.pipe_ctx[i];
+               }
+       }
+
+       return NULL;
+}
+
+static unsigned int find_pipes_assigned_to_plane(struct dml2_context *ctx, struct dc_state *state, unsigned int plane_id, unsigned int *pipes)
+{
+       int i;
+       unsigned int num_found = 0;
+       unsigned int plane_id_assigned_to_pipe;
+
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               if (state->res_ctx.pipe_ctx[i].plane_state && get_plane_id(state, state->res_ctx.pipe_ctx[i].plane_state, &plane_id_assigned_to_pipe)) {
+                       if (plane_id_assigned_to_pipe == plane_id)
+                               pipes[num_found++] = i;
+               }
+       }
+
+       return num_found;
+}
+
+static bool validate_pipe_assignment(const struct dml2_context *ctx, const struct dc_state *state, const struct dml_display_cfg_st *disp_cfg, const struct dml2_dml_to_dc_pipe_mapping *mapping)
+{
+//     int i, j, k;
+//
+//     unsigned int plane_id;
+//
+//     unsigned int disp_cfg_index;
+//
+//     unsigned int pipes_assigned_to_plane[MAX_PIPES];
+//     unsigned int num_pipes_assigned_to_plane;
+//
+//     struct pipe_ctx *top_pipe;
+//
+//     for (i = 0; i < state->stream_count; i++) {
+//             for (j = 0; j < state->stream_status[i]->plane_count; j++) {
+//                     if (get_plane_id(state, state->stream_status.plane_states[j], &plane_id)) {
+//                             disp_cfg_index = find_disp_cfg_idx_by_plane_id(mapping, plane_id);
+//                             num_pipes_assigned_to_plane = find_pipes_assigned_to_plane(ctx, state, plane_id, pipes_assigned_to_plane);
+//
+//                             if (disp_cfg_index >= 0 && num_pipes_assigned_to_plane > 0) {
+//                                     // Verify the number of pipes assigned matches
+//                                     if (disp_cfg->hw.DPPPerSurface != num_pipes_assigned_to_plane)
+//                                             return false;
+//
+//                                     top_pipe = find_top_pipe_in_tree(state->res_ctx.pipe_ctx[pipes_assigned_to_plane[0]]);
+//
+//                                     // Verify MPC and ODM combine
+//                                     if (disp_cfg->hw.ODMMode == dml_odm_mode_bypass) {
+//                                             verify_combine_tree(top_pipe, state->streams[i]->stream_id, plane_id, state, false);
+//                                     } else {
+//                                             verify_combine_tree(top_pipe, state->streams[i]->stream_id, plane_id, state, true);
+//                                     }
+//
+//                                     // TODO: could also do additional verification that the pipes in tree are the same as
+//                                     // pipes_assigned_to_plane
+//                             } else {
+//                                     ASSERT(false);
+//                                     return false;
+//                             }
+//                     } else {
+//                             ASSERT(false);
+//                             return false;
+//                     }
+//             }
+//     }
+       return true;
+}
+
+static bool is_plane_using_pipe(const struct pipe_ctx *pipe)
+{
+       if (pipe->plane_state)
+               return true;
+
+       return false;
+}
+
+static bool is_pipe_free(const struct pipe_ctx *pipe)
+{
+       if (!pipe->plane_state && !pipe->stream)
+               return true;
+
+       return false;
+}
+
+static bool find_more_pipes_for_stream(struct dml2_context *ctx,
+       struct dc_state *state, // The state we want to find a free mapping in
+       unsigned int stream_id, // The stream we want this pipe to drive
+       int *assigned_pipes,
+       int *assigned_pipe_count,
+       int pipes_needed,
+       const struct dc_state *existing_state) // The state (optional) that we want to minimize remapping relative to
+{
+       struct pipe_ctx *pipe = NULL;
+       unsigned int preferred_pipe_candidates[MAX_PIPES];
+       unsigned int num_preferred_candidates = 0;
+       int i;
+
+       if (existing_state) {
+               // To minimize prioritize candidates from existing stream
+               for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+                       if (existing_state->res_ctx.pipe_ctx[i].stream && existing_state->res_ctx.pipe_ctx[i].stream->stream_id == stream_id)
+                               preferred_pipe_candidates[num_preferred_candidates++] = existing_state->res_ctx.pipe_ctx[i].pipe_idx;
+               }
+       }
+
+       // First see if any of the preferred are unmapped, and choose those instead
+       for (i = 0; pipes_needed > 0 && i < num_preferred_candidates; i++) {
+               pipe = &state->res_ctx.pipe_ctx[preferred_pipe_candidates[i]];
+               if (!is_plane_using_pipe(pipe)) {
+                       pipes_needed--;
+                       // TODO: This doens't make sense really, pipe_idx should always be valid
+                       pipe->pipe_idx = preferred_pipe_candidates[i];
+                       assigned_pipes[(*assigned_pipe_count)++] = pipe->pipe_idx;
+               }
+       }
+
+       // We like to pair pipes starting from the higher order indicies for combining
+       for (i = ctx->config.dcn_pipe_count - 1; pipes_needed > 0 && i >= 0; i--) {
+               pipe = &state->res_ctx.pipe_ctx[i];
+               if (!is_plane_using_pipe(pipe)) {
+                       pipes_needed--;
+                       // TODO: This doens't make sense really, pipe_idx should always be valid
+                       pipe->pipe_idx = i;
+                       assigned_pipes[(*assigned_pipe_count)++] = pipe->pipe_idx;
+               }
+       }
+
+       ASSERT(pipes_needed <= 0); // Validation should prevent us from building a pipe context that exceeds the number of HW resoruces available
+
+       return pipes_needed <= 0;
+}
+
+static bool find_more_free_pipes(struct dml2_context *ctx,
+       struct dc_state *state, // The state we want to find a free mapping in
+       unsigned int stream_id, // The stream we want this pipe to drive
+       int *assigned_pipes,
+       int *assigned_pipe_count,
+       int pipes_needed,
+       const struct dc_state *existing_state) // The state (optional) that we want to minimize remapping relative to
+{
+       struct pipe_ctx *pipe = NULL;
+       unsigned int preferred_pipe_candidates[MAX_PIPES];
+       unsigned int num_preferred_candidates = 0;
+       int i;
+
+       if (existing_state) {
+               // To minimize prioritize candidates from existing stream
+               for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+                       if (existing_state->res_ctx.pipe_ctx[i].stream && existing_state->res_ctx.pipe_ctx[i].stream->stream_id == stream_id)
+                               preferred_pipe_candidates[num_preferred_candidates++] = existing_state->res_ctx.pipe_ctx[i].pipe_idx;
+               }
+       }
+
+       // First see if any of the preferred are unmapped, and choose those instead
+       for (i = 0; pipes_needed > 0 && i < num_preferred_candidates; i++) {
+               pipe = &state->res_ctx.pipe_ctx[preferred_pipe_candidates[i]];
+               if (is_pipe_free(pipe)) {
+                       pipes_needed--;
+                       // TODO: This doens't make sense really, pipe_idx should always be valid
+                       pipe->pipe_idx = preferred_pipe_candidates[i];
+                       assigned_pipes[(*assigned_pipe_count)++] = pipe->pipe_idx;
+               }
+       }
+
+       // We like to pair pipes starting from the higher order indicies for combining
+       for (i = ctx->config.dcn_pipe_count - 1; pipes_needed > 0 && i >= 0; i--) {
+               pipe = &state->res_ctx.pipe_ctx[i];
+               if (is_pipe_free(pipe)) {
+                       pipes_needed--;
+                       // TODO: This doens't make sense really, pipe_idx should always be valid
+                       pipe->pipe_idx = i;
+                       assigned_pipes[(*assigned_pipe_count)++] = pipe->pipe_idx;
+               }
+       }
+
+       ASSERT(pipes_needed == 0); // Validation should prevent us from building a pipe context that exceeds the number of HW resoruces available
+
+       return pipes_needed == 0;
+}
+
+static void sort_pipes_for_splitting(struct dc_plane_pipe_pool *pipes)
+{
+       bool sorted, swapped;
+       unsigned int cur_index;
+       unsigned int temp;
+       int odm_slice_index;
+
+       for (odm_slice_index = 0; odm_slice_index < pipes->num_pipes_assigned_to_plane_for_odm_combine; odm_slice_index++) {
+               // Sort each MPCC set
+               //Un-optimized bubble sort, but that's okay for array sizes <= 6
+
+               if (pipes->num_pipes_assigned_to_plane_for_mpcc_combine <= 1)
+                       sorted = true;
+               else
+                       sorted = false;
+
+               cur_index = 0;
+               swapped = false;
+               while (!sorted) {
+                       if (pipes->pipes_assigned_to_plane[odm_slice_index][cur_index] > pipes->pipes_assigned_to_plane[odm_slice_index][cur_index + 1]) {
+                               temp = pipes->pipes_assigned_to_plane[odm_slice_index][cur_index];
+                               pipes->pipes_assigned_to_plane[odm_slice_index][cur_index] = pipes->pipes_assigned_to_plane[odm_slice_index][cur_index + 1];
+                               pipes->pipes_assigned_to_plane[odm_slice_index][cur_index + 1] = temp;
+
+                               swapped = true;
+                       }
+
+                       cur_index++;
+
+                       if (cur_index == pipes->num_pipes_assigned_to_plane_for_mpcc_combine - 1) {
+                               cur_index = 0;
+
+                               if (swapped)
+                                       sorted = false;
+                               else
+                                       sorted = true;
+
+                               swapped = false;
+                       }
+
+               };
+       }
+}
+
+// For example, 3840 x 2160, ODM2:1 has a slice array of [1919, 3839], meaning, slice0 spans h_pixels 0->1919, and slice1 spans 1920->3840
+static void calculate_odm_slices(const struct dc_stream_state *stream, unsigned int odm_factor, unsigned int *odm_slice_end_x)
+{
+       unsigned int slice_size = 0;
+       int i;
+
+       if (odm_factor < 1 || odm_factor > 4) {
+               ASSERT(false);
+               return;
+       }
+
+       slice_size = stream->src.width / odm_factor;
+
+       for (i = 0; i < odm_factor; i++)
+               odm_slice_end_x[i] = (slice_size * (i + 1)) - 1;
+
+       odm_slice_end_x[odm_factor - 1] = stream->src.width - 1;
+}
+
+static bool is_plane_in_odm_slice(const struct dc_plane_state *plane, unsigned int slice_index, unsigned int *odm_slice_end_x, unsigned int num_slices)
+{
+       unsigned int slice_start_x, slice_end_x;
+
+       if (slice_index == 0)
+               slice_start_x = 0;
+       else
+               slice_start_x = odm_slice_end_x[slice_index - 1] + 1;
+
+       slice_end_x = odm_slice_end_x[slice_index];
+
+       if (plane->clip_rect.x + plane->clip_rect.width < slice_start_x)
+               return false;
+
+       if (plane->clip_rect.x > slice_end_x)
+               return false;
+
+       return true;
+}
+
+static void add_odm_slice_to_odm_tree(struct dml2_context *ctx,
+               struct dc_state *state,
+               struct dc_pipe_mapping_scratch *scratch,
+               unsigned int odm_slice_index)
+{
+       struct pipe_ctx *pipe = NULL;
+       int i;
+
+       // MPCC Combine + ODM Combine is not supported, so there should never be a case where the current plane
+       // has more than 1 pipe mapped to it for a given slice.
+       ASSERT(scratch->pipe_pool.num_pipes_assigned_to_plane_for_mpcc_combine == 1 || scratch->pipe_pool.num_pipes_assigned_to_plane_for_odm_combine == 1);
+
+       for (i = 0; i < scratch->pipe_pool.num_pipes_assigned_to_plane_for_mpcc_combine; i++) {
+               pipe = &state->res_ctx.pipe_ctx[scratch->pipe_pool.pipes_assigned_to_plane[odm_slice_index][i]];
+
+               if (scratch->mpc_info.prev_odm_pipe)
+                       scratch->mpc_info.prev_odm_pipe->next_odm_pipe = pipe;
+
+               pipe->prev_odm_pipe = scratch->mpc_info.prev_odm_pipe;
+               pipe->next_odm_pipe = NULL;
+       }
+       scratch->mpc_info.prev_odm_pipe = pipe;
+}
+
+static struct pipe_ctx *add_plane_to_blend_tree(struct dml2_context *ctx,
+       struct dc_state *state,
+       const struct dc_plane_state *plane,
+       struct dc_plane_pipe_pool *pipe_pool,
+       unsigned int odm_slice,
+       struct pipe_ctx *top_pipe)
+{
+       int i;
+
+       for (i = 0; i < pipe_pool->num_pipes_assigned_to_plane_for_mpcc_combine; i++) {
+               if (top_pipe)
+                       top_pipe->bottom_pipe = &state->res_ctx.pipe_ctx[pipe_pool->pipes_assigned_to_plane[odm_slice][i]];
+
+               pipe_pool->pipe_used[odm_slice][i] = true;
+
+               state->res_ctx.pipe_ctx[pipe_pool->pipes_assigned_to_plane[odm_slice][i]].top_pipe = top_pipe;
+               state->res_ctx.pipe_ctx[pipe_pool->pipes_assigned_to_plane[odm_slice][i]].bottom_pipe = NULL;
+
+               top_pipe = &state->res_ctx.pipe_ctx[pipe_pool->pipes_assigned_to_plane[odm_slice][i]];
+       }
+
+       // After running the above loop, the top pipe actually ends up pointing to the bottom of this MPCC combine tree, so we are actually
+       // returning the bottom pipe here
+       return top_pipe;
+}
+
+static unsigned int find_pipes_assigned_to_stream(struct dml2_context *ctx, struct dc_state *state, unsigned int stream_id, unsigned int *pipes)
+{
+       int i;
+       unsigned int num_found = 0;
+
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               if (state->res_ctx.pipe_ctx[i].stream && state->res_ctx.pipe_ctx[i].stream->stream_id == stream_id) {
+                       pipes[num_found++] = i;
+               }
+       }
+
+       return num_found;
+}
+
+static struct pipe_ctx *assign_pipes_to_stream(struct dml2_context *ctx, struct dc_state *state,
+               const struct dc_stream_state *stream,
+               int odm_factor,
+               struct dc_plane_pipe_pool *pipe_pool,
+               const struct dc_state *existing_state)
+{
+       struct pipe_ctx *master_pipe;
+       unsigned int pipes_needed;
+       unsigned int pipes_assigned;
+       unsigned int pipes[MAX_PIPES];
+       unsigned int next_pipe_to_assign;
+       int odm_slice;
+
+       pipes_needed = odm_factor;
+
+       master_pipe = find_master_pipe_of_stream(ctx, state, stream->stream_id);
+       ASSERT(master_pipe);
+
+       pipes_assigned = find_pipes_assigned_to_stream(ctx, state, stream->stream_id, pipes);
+
+       find_more_free_pipes(ctx, state, stream->stream_id, pipes, &pipes_assigned, pipes_needed - pipes_assigned, existing_state);
+
+       ASSERT(pipes_assigned == pipes_needed);
+
+       next_pipe_to_assign = 0;
+       for (odm_slice = 0; odm_slice < odm_factor; odm_slice++)
+               pipe_pool->pipes_assigned_to_plane[odm_slice][0] = pipes[next_pipe_to_assign++];
+
+       pipe_pool->num_pipes_assigned_to_plane_for_mpcc_combine = 1;
+       pipe_pool->num_pipes_assigned_to_plane_for_odm_combine = odm_factor;
+
+       return master_pipe;
+}
+
+static struct pipe_ctx *assign_pipes_to_plane(struct dml2_context *ctx, struct dc_state *state,
+               const struct dc_stream_state *stream,
+               const struct dc_plane_state *plane,
+               int odm_factor,
+               int mpc_factor,
+               struct dc_plane_pipe_pool *pipe_pool,
+               const struct dc_state *existing_state)
+{
+       struct pipe_ctx *master_pipe = NULL;
+       unsigned int plane_id;
+       unsigned int pipes_needed;
+       unsigned int pipes_assigned;
+       unsigned int pipes[MAX_PIPES];
+       unsigned int next_pipe_to_assign;
+       int odm_slice, mpc_slice;
+
+       if (!get_plane_id(state, plane, &plane_id)) {
+               ASSERT(false);
+               return master_pipe;
+       }
+
+       pipes_needed = mpc_factor * odm_factor;
+
+       master_pipe = find_master_pipe_of_plane(ctx, state, plane_id);
+       ASSERT(master_pipe);
+
+       pipes_assigned = find_pipes_assigned_to_plane(ctx, state, plane_id, pipes);
+
+       find_more_pipes_for_stream(ctx, state, stream->stream_id, pipes, &pipes_assigned, pipes_needed - pipes_assigned, existing_state);
+
+       ASSERT(pipes_assigned >= pipes_needed);
+
+       next_pipe_to_assign = 0;
+       for (odm_slice = 0; odm_slice < odm_factor; odm_slice++)
+               for (mpc_slice = 0; mpc_slice < mpc_factor; mpc_slice++)
+                       pipe_pool->pipes_assigned_to_plane[odm_slice][mpc_slice] = pipes[next_pipe_to_assign++];
+
+       pipe_pool->num_pipes_assigned_to_plane_for_mpcc_combine = mpc_factor;
+       pipe_pool->num_pipes_assigned_to_plane_for_odm_combine = odm_factor;
+
+       return master_pipe;
+}
+
+static bool is_pipe_used(const struct dc_plane_pipe_pool *pool, unsigned int pipe_idx)
+{
+       int i, j;
+
+       for (i = 0; i < pool->num_pipes_assigned_to_plane_for_odm_combine; i++) {
+               for (j = 0; j < pool->num_pipes_assigned_to_plane_for_mpcc_combine; j++) {
+                       if (pool->pipes_assigned_to_plane[i][j] == pipe_idx && pool->pipe_used[i][j])
+                               return true;
+               }
+       }
+
+       return false;
+}
+
+static void free_pipe(struct pipe_ctx *pipe)
+{
+       memset(pipe, 0, sizeof(struct pipe_ctx));
+}
+
+static void free_unused_pipes_for_plane(struct dml2_context *ctx, struct dc_state *state, const struct dc_plane_state *plane, const struct dc_plane_pipe_pool *pool)
+{
+       int i;
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               if (state->res_ctx.pipe_ctx[i].plane_state == plane && !is_pipe_used(pool, state->res_ctx.pipe_ctx[i].pipe_idx)) {
+                       free_pipe(&state->res_ctx.pipe_ctx[i]);
+               }
+       }
+}
+
+static void remove_pipes_from_blend_trees(struct dml2_context *ctx, struct dc_state *state, struct dc_plane_pipe_pool *pipe_pool, unsigned int odm_slice)
+{
+       struct pipe_ctx *pipe;
+       int i;
+
+       for (i = 0; i < pipe_pool->num_pipes_assigned_to_plane_for_mpcc_combine; i++) {
+               pipe = &state->res_ctx.pipe_ctx[pipe_pool->pipes_assigned_to_plane[odm_slice][0]];
+               if (pipe->top_pipe)
+                       pipe->top_pipe->bottom_pipe = pipe->bottom_pipe;
+
+               if (pipe->bottom_pipe)
+                       pipe->bottom_pipe = pipe->top_pipe;
+
+               pipe_pool->pipe_used[odm_slice][i] = true;
+       }
+}
+
+static void map_pipes_for_stream(struct dml2_context *ctx, struct dc_state *state, const struct dc_stream_state *stream,
+               struct dc_pipe_mapping_scratch *scratch, const struct dc_state *existing_state)
+{
+       int odm_slice_index;
+       struct pipe_ctx *master_pipe = NULL;
+
+
+       master_pipe = assign_pipes_to_stream(ctx, state, stream, scratch->odm_info.odm_factor, &scratch->pipe_pool, existing_state);
+       sort_pipes_for_splitting(&scratch->pipe_pool);
+
+       for (odm_slice_index = 0; odm_slice_index < scratch->odm_info.odm_factor; odm_slice_index++) {
+               remove_pipes_from_blend_trees(ctx, state, &scratch->pipe_pool, odm_slice_index);
+
+               add_odm_slice_to_odm_tree(ctx, state, scratch, odm_slice_index);
+
+               ctx->config.callbacks.acquire_secondary_pipe_for_mpc_odm(ctx->config.callbacks.dc, state,
+                       master_pipe, &state->res_ctx.pipe_ctx[scratch->pipe_pool.pipes_assigned_to_plane[odm_slice_index][0]], true);
+       }
+}
+
+static void map_pipes_for_plane(struct dml2_context *ctx, struct dc_state *state, const struct dc_stream_state *stream, const struct dc_plane_state *plane,
+               struct dc_pipe_mapping_scratch *scratch, const struct dc_state *existing_state)
+{
+       int odm_slice_index;
+       unsigned int plane_id;
+       struct pipe_ctx *master_pipe = NULL;
+       int i;
+
+       if (!get_plane_id(state, plane, &plane_id)) {
+               ASSERT(false);
+               return;
+       }
+
+       master_pipe = assign_pipes_to_plane(ctx, state, stream, plane, scratch->odm_info.odm_factor, scratch->mpc_info.mpc_factor, &scratch->pipe_pool, existing_state);
+       sort_pipes_for_splitting(&scratch->pipe_pool);
+
+       for (odm_slice_index = 0; odm_slice_index < scratch->odm_info.odm_factor; odm_slice_index++) {
+               // We build the tree for one ODM slice at a time.
+               // Each ODM slice shares a common OPP
+               if (!is_plane_in_odm_slice(plane, odm_slice_index, scratch->odm_info.odm_slice_end_x, scratch->odm_info.odm_factor)) {
+                       continue;
+               }
+
+               // Now we have a list of all pipes to be used for this plane/stream, now setup the tree.
+               scratch->odm_info.next_higher_pipe_for_odm_slice[odm_slice_index] = add_plane_to_blend_tree(ctx, state,
+                               plane,
+                               &scratch->pipe_pool,
+                               odm_slice_index,
+                               scratch->odm_info.next_higher_pipe_for_odm_slice[odm_slice_index]);
+
+               add_odm_slice_to_odm_tree(ctx, state, scratch, odm_slice_index);
+
+               for (i = 0; i < scratch->pipe_pool.num_pipes_assigned_to_plane_for_mpcc_combine; i++) {
+
+                       ctx->config.callbacks.acquire_secondary_pipe_for_mpc_odm(ctx->config.callbacks.dc, state,
+                               master_pipe, &state->res_ctx.pipe_ctx[scratch->pipe_pool.pipes_assigned_to_plane[odm_slice_index][i]], true);
+               }
+       }
+
+       free_unused_pipes_for_plane(ctx, state, plane, &scratch->pipe_pool);
+}
+
+bool dml2_map_dc_pipes(struct dml2_context *ctx, struct dc_state *state, const struct dml_display_cfg_st *disp_cfg, struct dml2_dml_to_dc_pipe_mapping *mapping, const struct dc_state *existing_state)
+{
+       int stream_index, plane_index, i;
+
+       unsigned int stream_disp_cfg_index;
+       unsigned int plane_disp_cfg_index;
+
+       unsigned int plane_id;
+       unsigned int stream_id;
+
+       const unsigned int *ODMMode, *DPPPerSurface;
+       unsigned int odm_mode_array[__DML2_WRAPPER_MAX_STREAMS_PLANES__] = {0}, dpp_per_surface_array[__DML2_WRAPPER_MAX_STREAMS_PLANES__] = {0};
+       struct dc_pipe_mapping_scratch scratch;
+
+       if (ctx->architecture == dml2_architecture_21) {
+               /*
+                * Extract ODM and DPP outputs from DML2.1 and map them in an array as required for pipe mapping in dml2_map_dc_pipes.
+                * As data cannot be directly extracted in const pointers, assign these arrays to const pointers before proceeding to
+                * maximize the reuse of existing code. Const pointers are required because dml2.0 dml_display_cfg_st is const.
+                *
+                */
+               ODMMode = (const unsigned int *)odm_mode_array;
+               DPPPerSurface = (const unsigned int *)dpp_per_surface_array;
+       } else {
+               ODMMode = (unsigned int *)disp_cfg->hw.ODMMode;
+               DPPPerSurface = disp_cfg->hw.DPPPerSurface;
+       }
+
+       for (stream_index = 0; stream_index < state->stream_count; stream_index++) {
+               memset(&scratch, 0, sizeof(struct dc_pipe_mapping_scratch));
+
+               stream_id = state->streams[stream_index]->stream_id;
+               stream_disp_cfg_index = find_disp_cfg_idx_by_stream_id(mapping, stream_id);
+
+               if (ODMMode[stream_disp_cfg_index] == dml_odm_mode_bypass) {
+                       scratch.odm_info.odm_factor = 1;
+               } else if (ODMMode[stream_disp_cfg_index] == dml_odm_mode_combine_2to1) {
+                       scratch.odm_info.odm_factor = 2;
+               } else if (ODMMode[stream_disp_cfg_index] == dml_odm_mode_combine_4to1) {
+                       scratch.odm_info.odm_factor = 4;
+               } else {
+                       ASSERT(false);
+                       scratch.odm_info.odm_factor = 1;
+               }
+
+               calculate_odm_slices(state->streams[stream_index], scratch.odm_info.odm_factor, scratch.odm_info.odm_slice_end_x);
+
+               // If there are no planes, you still want to setup ODM...
+               if (state->stream_status[stream_index].plane_count == 0) {
+                       map_pipes_for_stream(ctx, state, state->streams[stream_index], &scratch, existing_state);
+               }
+
+               for (plane_index = 0; plane_index < state->stream_status[stream_index].plane_count; plane_index++) {
+                       // Planes are ordered top to bottom.
+                       if (get_plane_id(state, state->stream_status[stream_index].plane_states[plane_index], &plane_id)) {
+                               plane_disp_cfg_index = find_disp_cfg_idx_by_plane_id(mapping, plane_id);
+
+                               // Setup mpc_info for this plane
+                               scratch.mpc_info.prev_odm_pipe = NULL;
+                               if (scratch.odm_info.odm_factor == 1) {
+                                       // If ODM combine is not inuse, then the number of pipes
+                                       // per plane is determined by MPC combine factor
+                                       scratch.mpc_info.mpc_factor = DPPPerSurface[plane_disp_cfg_index];
+                               } else {
+                                       // If ODM combine is enabled, then we use at most 1 pipe per
+                                       // odm slice per plane, i.e. MPC combine is never used
+                                       scratch.mpc_info.mpc_factor = 1;
+                               }
+
+                               ASSERT(scratch.odm_info.odm_factor * scratch.mpc_info.mpc_factor > 0);
+
+                               // Clear the pool assignment scratch (which is per plane)
+                               memset(&scratch.pipe_pool, 0, sizeof(struct dc_plane_pipe_pool));
+
+                               map_pipes_for_plane(ctx, state, state->streams[stream_index], state->stream_status[stream_index].plane_states[plane_index], &scratch, existing_state);
+                       } else {
+                               // Plane ID cannot be generated, therefore no DML mapping can be performed.
+                               ASSERT(false);
+                       }
+               }
+
+       }
+
+       if (!validate_pipe_assignment(ctx, state, disp_cfg, mapping))
+               ASSERT(false);
+
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               struct pipe_ctx *pipe = &state->res_ctx.pipe_ctx[i];
+
+               if (pipe->plane_state) {
+                       if (!ctx->config.callbacks.build_scaling_params(pipe)) {
+                               ASSERT(false);
+                       }
+               }
+       }
+
+       return true;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.h b/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.h
new file mode 100644 (file)
index 0000000..a78de19
--- /dev/null
@@ -0,0 +1,48 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DML2_DC_RESOURCE_MGMT_H__
+#define __DML2_DC_RESOURCE_MGMT_H__
+
+#include "dml2_dc_types.h"
+
+struct dml2_context;
+
+/*
+ * dml2_map_dc_pipes - Creates a pipe linkage in dc_state based on current display config.
+ * @ctx: Input dml2 context
+ * @state: Current dc_state to be updated.
+ * @disp_cfg: Current display config.
+ * @mapping: Pipe mapping logic structure to keep a track of pipes to be used.
+ *
+ * Based on ODM and DPPPersurface outputs calculated by the DML for the current display
+ * config, create a pipe linkage in dc_state which is then used by DC core.
+ * Make this function generic to be used by multiple DML versions.
+ *
+ * Return: True if pipe mapping and linking is successful, false otherwise.
+ */
+
+bool dml2_map_dc_pipes(struct dml2_context *ctx, struct dc_state *state, const struct dml_display_cfg_st *disp_cfg, struct dml2_dml_to_dc_pipe_mapping *mapping, const struct dc_state *existing_state);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_types.h
new file mode 100644 (file)
index 0000000..191b2e6
--- /dev/null
@@ -0,0 +1,40 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+
+/*
+ * Wrapper header for externally defined types from DC.  These types come from
+ * dc headers when building DML2 as part of DC, but are defined here when building
+ * DML2 as a standalone library (such as for unit testing).
+ */
+
+#ifndef __DML2_DC_TYPES_H__
+#define __DML2_DC_TYPES_H__
+
+#include "resource.h"
+#include "core_types.h"
+#include "dsc.h"
+#include "clk_mgr.h"
+
+#endif //__DML2_DC_TYPES_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_internal_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml2_internal_types.h
new file mode 100644 (file)
index 0000000..bcb59bc
--- /dev/null
@@ -0,0 +1,121 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __DML2_INTERNAL_TYPES_H__
+#define __DML2_INTERNAL_TYPES_H__
+
+#include "dml2_dc_types.h"
+#include "display_mode_core.h"
+#include "dml2_wrapper.h"
+#include "dml2_policy.h"
+
+
+struct dml2_wrapper_optimize_configuration_params {
+       struct display_mode_lib_st *dml_core_ctx;
+       struct dml2_configuration_options *config;
+       struct ip_params_st *ip_params;
+       struct dml_display_cfg_st *cur_display_config;
+       struct dml_display_cfg_st *new_display_config;
+       const struct dml_mode_support_info_st *cur_mode_support_info;
+       struct dml_mode_eval_policy_st *cur_policy;
+       struct dml_mode_eval_policy_st *new_policy;
+};
+
+struct dml2_calculate_lowest_supported_state_for_temp_read_scratch {
+       struct dml_mode_support_info_st evaluation_info;
+       dml_float_t uclk_change_latencies[__DML_MAX_STATE_ARRAY_SIZE__];
+       struct dml_display_cfg_st cur_display_config;
+       struct dml_display_cfg_st new_display_config;
+       struct dml_mode_eval_policy_st new_policy;
+       struct dml_mode_eval_policy_st cur_policy;
+};
+
+struct dml2_create_scratch {
+       struct dml2_policy_build_synthetic_soc_states_scratch build_synthetic_socbb_scratch;
+       struct soc_states_st in_states;
+};
+
+struct dml2_calculate_rq_and_dlg_params_scratch {
+       struct _vcs_dpi_dml_display_rq_regs_st rq_regs;
+       struct _vcs_dpi_dml_display_dlg_regs_st disp_dlg_regs;
+       struct _vcs_dpi_dml_display_ttu_regs_st disp_ttu_regs;
+};
+
+#define __DML2_WRAPPER_MAX_STREAMS_PLANES__ 6
+
+struct dml2_dml_to_dc_pipe_mapping {
+       unsigned int disp_cfg_to_stream_id[__DML2_WRAPPER_MAX_STREAMS_PLANES__];
+       bool disp_cfg_to_stream_id_valid[__DML2_WRAPPER_MAX_STREAMS_PLANES__];
+       unsigned int disp_cfg_to_plane_id[__DML2_WRAPPER_MAX_STREAMS_PLANES__];
+       bool disp_cfg_to_plane_id_valid[__DML2_WRAPPER_MAX_STREAMS_PLANES__];
+       unsigned int dml_pipe_idx_to_stream_id[__DML2_WRAPPER_MAX_STREAMS_PLANES__];
+       bool dml_pipe_idx_to_stream_id_valid[__DML2_WRAPPER_MAX_STREAMS_PLANES__];
+       unsigned int dml_pipe_idx_to_plane_id[__DML2_WRAPPER_MAX_STREAMS_PLANES__];
+       bool dml_pipe_idx_to_plane_id_valid[__DML2_WRAPPER_MAX_STREAMS_PLANES__];
+};
+
+struct dml2_wrapper_scratch {
+       struct dml_display_cfg_st cur_display_config;
+       struct dml_display_cfg_st new_display_config;
+       struct dml_mode_eval_policy_st new_policy;
+       struct dml_mode_eval_policy_st cur_policy;
+       struct dml_mode_support_info_st mode_support_info;
+       struct dml_mode_support_ex_params_st mode_support_params;
+
+       struct dummy_pstate_entry dummy_pstate_table[4];
+
+       struct dml2_create_scratch create_scratch;
+       struct dml2_calculate_lowest_supported_state_for_temp_read_scratch dml2_calculate_lowest_supported_state_for_temp_read_scratch;
+       struct dml2_calculate_rq_and_dlg_params_scratch calculate_rq_and_dlg_params_scratch;
+
+       struct dml2_wrapper_optimize_configuration_params optimize_configuration_params;
+       struct dml2_policy_build_synthetic_soc_states_params build_synthetic_socbb_params;
+
+       struct dml2_dml_to_dc_pipe_mapping dml_to_dc_pipe_mapping;
+       bool enable_flexible_pipe_mapping;
+};
+
+struct dml2_helper_det_policy_scratch {
+       int dpps_per_surface[MAX_PLANES];
+};
+
+enum dml2_architecture {
+       dml2_architecture_20,
+       dml2_architecture_21
+};
+
+struct dml2_context {
+       enum dml2_architecture architecture;
+       struct dml2_configuration_options config;
+       struct dml2_helper_det_policy_scratch det_helper_scratch;
+       union {
+               struct {
+                       struct display_mode_lib_st dml_core_ctx;
+                       struct dml2_wrapper_scratch scratch;
+                       struct dcn_watermarks g6_temp_read_watermark_set;
+               } v20;
+       };
+};
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_mall_phantom.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_mall_phantom.c
new file mode 100644 (file)
index 0000000..02797cb
--- /dev/null
@@ -0,0 +1,913 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "dml2_dc_types.h"
+#include "dml2_internal_types.h"
+#include "dml2_utils.h"
+#include "dml2_mall_phantom.h"
+
+unsigned int dml2_helper_calculate_num_ways_for_subvp(struct dml2_context *ctx, struct dc_state *context)
+{
+       uint32_t num_ways = 0;
+       uint32_t bytes_per_pixel = 0;
+       uint32_t cache_lines_used = 0;
+       uint32_t lines_per_way = 0;
+       uint32_t total_cache_lines = 0;
+       uint32_t bytes_in_mall = 0;
+       uint32_t num_mblks = 0;
+       uint32_t cache_lines_per_plane = 0;
+       uint32_t i = 0;
+       uint32_t mblk_width = 0;
+       uint32_t mblk_height = 0;
+       uint32_t full_vp_width_blk_aligned = 0;
+       uint32_t mall_alloc_width_blk_aligned = 0;
+       uint32_t mall_alloc_height_blk_aligned = 0;
+
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+               // Find the phantom pipes
+               if (pipe->stream && pipe->plane_state && !pipe->top_pipe && !pipe->prev_odm_pipe &&
+                               pipe->stream->mall_stream_config.type == SUBVP_PHANTOM) {
+                       bytes_per_pixel = pipe->plane_state->format >= SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616 ? 8 : 4;
+                       mblk_width = ctx->config.mall_cfg.mblk_width_pixels;
+                       mblk_height = bytes_per_pixel == 4 ? mblk_width = ctx->config.mall_cfg.mblk_height_4bpe_pixels : ctx->config.mall_cfg.mblk_height_8bpe_pixels;
+
+                       /* full_vp_width_blk_aligned = FLOOR(vp_x_start + full_vp_width + blk_width - 1, blk_width) -
+                        * FLOOR(vp_x_start, blk_width)
+                        */
+                       full_vp_width_blk_aligned = ((pipe->plane_res.scl_data.viewport.x +
+                                       pipe->plane_res.scl_data.viewport.width + mblk_width - 1) / mblk_width * mblk_width) +
+                                       (pipe->plane_res.scl_data.viewport.x / mblk_width * mblk_width);
+
+                       /* mall_alloc_width_blk_aligned_l/c = full_vp_width_blk_aligned_l/c */
+                       mall_alloc_width_blk_aligned = full_vp_width_blk_aligned;
+
+                       /* mall_alloc_height_blk_aligned_l/c = CEILING(sub_vp_height_l/c - 1, blk_height_l/c) + blk_height_l/c */
+                       mall_alloc_height_blk_aligned = (pipe->stream->timing.v_addressable - 1 + mblk_height - 1) /
+                                       mblk_height * mblk_height + mblk_height;
+
+                       /* full_mblk_width_ub_l/c = malldml2_mall_phantom.c_alloc_width_blk_aligned_l/c;
+                        * full_mblk_height_ub_l/c = mall_alloc_height_blk_aligned_l/c;
+                        * num_mblk_l/c = (full_mblk_width_ub_l/c / mblk_width_l/c) * (full_mblk_height_ub_l/c / mblk_height_l/c);
+                        * (Should be divisible, but round up if not)
+                        */
+                       num_mblks = ((mall_alloc_width_blk_aligned + mblk_width - 1) / mblk_width) *
+                                       ((mall_alloc_height_blk_aligned + mblk_height - 1) / mblk_height);
+                       bytes_in_mall = num_mblks * ctx->config.mall_cfg.mblk_size_bytes;
+                       // cache lines used is total bytes / cache_line size. Add +2 for worst case alignment
+                       // (MALL is 64-byte aligned)
+                       cache_lines_per_plane = bytes_in_mall / ctx->config.mall_cfg.cache_line_size_bytes + 2;
+
+                       // For DCC we must cache the meat surface, so double cache lines required
+                       if (pipe->plane_state->dcc.enable)
+                               cache_lines_per_plane *= 2;
+                       cache_lines_used += cache_lines_per_plane;
+               }
+       }
+
+       total_cache_lines = ctx->config.mall_cfg.max_cab_allocation_bytes / ctx->config.mall_cfg.cache_line_size_bytes;
+       lines_per_way = total_cache_lines / ctx->config.mall_cfg.cache_num_ways;
+       num_ways = cache_lines_used / lines_per_way;
+       if (cache_lines_used % lines_per_way > 0)
+               num_ways++;
+
+       return num_ways;
+}
+
+static void merge_pipes_for_subvp(struct dml2_context *ctx, struct dc_state *context)
+{
+       int i;
+
+       /* merge pipes if necessary */
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+               // For now merge all pipes for SubVP since pipe split case isn't supported yet
+
+               /* if ODM merge we ignore mpc tree, mpo pipes will have their own flags */
+               if (pipe->prev_odm_pipe) {
+                       /*split off odm pipe*/
+                       pipe->prev_odm_pipe->next_odm_pipe = pipe->next_odm_pipe;
+                       if (pipe->next_odm_pipe)
+                               pipe->next_odm_pipe->prev_odm_pipe = pipe->prev_odm_pipe;
+
+                       pipe->bottom_pipe = NULL;
+                       pipe->next_odm_pipe = NULL;
+                       pipe->plane_state = NULL;
+                       pipe->stream = NULL;
+                       pipe->top_pipe = NULL;
+                       pipe->prev_odm_pipe = NULL;
+                       if (pipe->stream_res.dsc)
+                               ctx->config.svp_pstate.callbacks.release_dsc(&context->res_ctx, ctx->config.svp_pstate.callbacks.dc->res_pool, &pipe->stream_res.dsc);
+                       memset(&pipe->plane_res, 0, sizeof(pipe->plane_res));
+                       memset(&pipe->stream_res, 0, sizeof(pipe->stream_res));
+               } else if (pipe->top_pipe && pipe->top_pipe->plane_state == pipe->plane_state) {
+                       struct pipe_ctx *top_pipe = pipe->top_pipe;
+                       struct pipe_ctx *bottom_pipe = pipe->bottom_pipe;
+
+                       top_pipe->bottom_pipe = bottom_pipe;
+                       if (bottom_pipe)
+                               bottom_pipe->top_pipe = top_pipe;
+
+                       pipe->top_pipe = NULL;
+                       pipe->bottom_pipe = NULL;
+                       pipe->plane_state = NULL;
+                       pipe->stream = NULL;
+                       memset(&pipe->plane_res, 0, sizeof(pipe->plane_res));
+                       memset(&pipe->stream_res, 0, sizeof(pipe->stream_res));
+               }
+       }
+}
+
+static bool all_pipes_have_stream_and_plane(struct dml2_context *ctx, const struct dc_state *context)
+{
+       int i;
+
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               const struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+               if (!pipe->stream)
+                       continue;
+
+               if (!pipe->plane_state)
+                       return false;
+       }
+       return true;
+}
+
+static bool mpo_in_use(const struct dc_state *context)
+{
+       int i;
+
+       for (i = 0; i < context->stream_count; i++) {
+               if (context->stream_status[i].plane_count > 1)
+                       return true;
+       }
+       return false;
+}
+
+/*
+ * dcn32_get_num_free_pipes: Calculate number of free pipes
+ *
+ * This function assumes that a "used" pipe is a pipe that has
+ * both a stream and a plane assigned to it.
+ *
+ * @dc: current dc state
+ * @context: new dc state
+ *
+ * Return:
+ * Number of free pipes available in the context
+ */
+static unsigned int get_num_free_pipes(struct dml2_context *ctx, struct dc_state *state)
+{
+       unsigned int i;
+       unsigned int free_pipes = 0;
+       unsigned int num_pipes = 0;
+
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               struct pipe_ctx *pipe = &state->res_ctx.pipe_ctx[i];
+
+               if (pipe->stream && !pipe->top_pipe) {
+                       while (pipe) {
+                               num_pipes++;
+                               pipe = pipe->bottom_pipe;
+                       }
+               }
+       }
+
+       free_pipes = ctx->config.dcn_pipe_count - num_pipes;
+       return free_pipes;
+}
+
+/*
+ * dcn32_assign_subvp_pipe: Function to decide which pipe will use Sub-VP.
+ *
+ * We enter this function if we are Sub-VP capable (i.e. enough pipes available)
+ * and regular P-State switching (i.e. VACTIVE/VBLANK) is not supported, or if
+ * we are forcing SubVP P-State switching on the current config.
+ *
+ * The number of pipes used for the chosen surface must be less than or equal to the
+ * number of free pipes available.
+ *
+ * In general we choose surfaces with the longest frame time first (better for SubVP + VBLANK).
+ * For multi-display cases the ActiveDRAMClockChangeMargin doesn't provide enough info on its own
+ * for determining which should be the SubVP pipe (need a way to determine if a pipe / plane doesn't
+ * support MCLK switching naturally [i.e. ACTIVE or VBLANK]).
+ *
+ * @param dc: current dc state
+ * @param context: new dc state
+ * @param index: [out] dc pipe index for the pipe chosen to have phantom pipes assigned
+ *
+ * Return:
+ * True if a valid pipe assignment was found for Sub-VP. Otherwise false.
+ */
+static bool assign_subvp_pipe(struct dml2_context *ctx, struct dc_state *context, unsigned int *index)
+{
+       unsigned int i, pipe_idx;
+       unsigned int max_frame_time = 0;
+       bool valid_assignment_found = false;
+       unsigned int free_pipes = 2; //dcn32_get_num_free_pipes(dc, context);
+       bool current_assignment_freesync = false;
+       struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+
+       for (i = 0, pipe_idx = 0; i < ctx->config.dcn_pipe_count; i++) {
+               struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+               unsigned int num_pipes = 0;
+               unsigned int refresh_rate = 0;
+
+               if (!pipe->stream)
+                       continue;
+
+               // Round up
+               refresh_rate = (pipe->stream->timing.pix_clk_100hz * 100 +
+                               pipe->stream->timing.v_total * pipe->stream->timing.h_total - 1)
+                               / (double)(pipe->stream->timing.v_total * pipe->stream->timing.h_total);
+               /* SubVP pipe candidate requirements:
+                * - Refresh rate < 120hz
+                * - Not able to switch in vactive naturally (switching in active means the
+                *   DET provides enough buffer to hide the P-State switch latency -- trying
+                *   to combine this with SubVP can cause issues with the scheduling).
+                */
+               if (pipe->plane_state && !pipe->top_pipe &&
+                               pipe->stream->mall_stream_config.type == SUBVP_NONE && refresh_rate < 120 &&
+                               vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] <= 0) {
+                       while (pipe) {
+                               num_pipes++;
+                               pipe = pipe->bottom_pipe;
+                       }
+
+                       pipe = &context->res_ctx.pipe_ctx[i];
+                       if (num_pipes <= free_pipes) {
+                               struct dc_stream_state *stream = pipe->stream;
+                               unsigned int frame_us = (stream->timing.v_total * stream->timing.h_total /
+                                               (double)(stream->timing.pix_clk_100hz * 100)) * 1000000;
+                               if (frame_us > max_frame_time && !stream->ignore_msa_timing_param) {
+                                       *index = i;
+                                       max_frame_time = frame_us;
+                                       valid_assignment_found = true;
+                                       current_assignment_freesync = false;
+                               /* For the 2-Freesync display case, still choose the one with the
+                            * longest frame time
+                            */
+                               } else if (stream->ignore_msa_timing_param && (!valid_assignment_found ||
+                                               (current_assignment_freesync && frame_us > max_frame_time))) {
+                                       *index = i;
+                                       valid_assignment_found = true;
+                                       current_assignment_freesync = true;
+                               }
+                       }
+               }
+               pipe_idx++;
+       }
+       return valid_assignment_found;
+}
+
+/*
+ * enough_pipes_for_subvp: Function to check if there are "enough" pipes for SubVP.
+ *
+ * This function returns true if there are enough free pipes
+ * to create the required phantom pipes for any given stream
+ * (that does not already have phantom pipe assigned).
+ *
+ * e.g. For a 2 stream config where the first stream uses one
+ * pipe and the second stream uses 2 pipes (i.e. pipe split),
+ * this function will return true because there is 1 remaining
+ * pipe which can be used as the phantom pipe for the non pipe
+ * split pipe.
+ *
+ * @dc: current dc state
+ * @context: new dc state
+ *
+ * Return:
+ * True if there are enough free pipes to assign phantom pipes to at least one
+ * stream that does not already have phantom pipes assigned. Otherwise false.
+ */
+static bool enough_pipes_for_subvp(struct dml2_context *ctx, struct dc_state *state)
+{
+       unsigned int i, split_cnt, free_pipes;
+       unsigned int min_pipe_split = ctx->config.dcn_pipe_count + 1; // init as max number of pipes + 1
+       bool subvp_possible = false;
+
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               struct pipe_ctx *pipe = &state->res_ctx.pipe_ctx[i];
+
+               // Find the minimum pipe split count for non SubVP pipes
+               if (pipe->stream && !pipe->top_pipe &&
+                   pipe->stream->mall_stream_config.type == SUBVP_NONE) {
+                       split_cnt = 0;
+                       while (pipe) {
+                               split_cnt++;
+                               pipe = pipe->bottom_pipe;
+                       }
+
+                       if (split_cnt < min_pipe_split)
+                               min_pipe_split = split_cnt;
+               }
+       }
+
+       free_pipes = get_num_free_pipes(ctx, state);
+
+       // SubVP only possible if at least one pipe is being used (i.e. free_pipes
+       // should not equal to the pipe_count)
+       if (free_pipes >= min_pipe_split && free_pipes < ctx->config.dcn_pipe_count)
+               subvp_possible = true;
+
+       return subvp_possible;
+}
+
+/*
+ * subvp_subvp_schedulable: Determine if SubVP + SubVP config is schedulable
+ *
+ * High level algorithm:
+ * 1. Find longest microschedule length (in us) between the two SubVP pipes
+ * 2. Check if the worst case overlap (VBLANK in middle of ACTIVE) for both
+ * pipes still allows for the maximum microschedule to fit in the active
+ * region for both pipes.
+ *
+ * @dc: current dc state
+ * @context: new dc state
+ *
+ * Return:
+ * bool - True if the SubVP + SubVP config is schedulable, false otherwise
+ */
+static bool subvp_subvp_schedulable(struct dml2_context *ctx, struct dc_state *context)
+{
+       struct pipe_ctx *subvp_pipes[2];
+       struct dc_stream_state *phantom = NULL;
+       uint32_t microschedule_lines = 0;
+       uint32_t index = 0;
+       uint32_t i;
+       uint32_t max_microschedule_us = 0;
+       int32_t vactive1_us, vactive2_us, vblank1_us, vblank2_us;
+
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+               uint32_t time_us = 0;
+
+               /* Loop to calculate the maximum microschedule time between the two SubVP pipes,
+                * and also to store the two main SubVP pipe pointers in subvp_pipes[2].
+                */
+               if (pipe->stream && pipe->plane_state && !pipe->top_pipe &&
+                   pipe->stream->mall_stream_config.type == SUBVP_MAIN) {
+                       phantom = pipe->stream->mall_stream_config.paired_stream;
+                       microschedule_lines = (phantom->timing.v_total - phantom->timing.v_front_porch) +
+                                       phantom->timing.v_addressable;
+
+                       // Round up when calculating microschedule time (+ 1 at the end)
+                       time_us = (microschedule_lines * phantom->timing.h_total) /
+                                       (double)(phantom->timing.pix_clk_100hz * 100) * 1000000 +
+                                       ctx->config.svp_pstate.subvp_prefetch_end_to_mall_start_us +
+                                       ctx->config.svp_pstate.subvp_fw_processing_delay_us + 1;
+                       if (time_us > max_microschedule_us)
+                               max_microschedule_us = time_us;
+
+                       subvp_pipes[index] = pipe;
+                       index++;
+
+                       // Maximum 2 SubVP pipes
+                       if (index == 2)
+                               break;
+               }
+       }
+       vactive1_us = ((subvp_pipes[0]->stream->timing.v_addressable * subvp_pipes[0]->stream->timing.h_total) /
+                       (double)(subvp_pipes[0]->stream->timing.pix_clk_100hz * 100)) * 1000000;
+       vactive2_us = ((subvp_pipes[1]->stream->timing.v_addressable * subvp_pipes[1]->stream->timing.h_total) /
+                               (double)(subvp_pipes[1]->stream->timing.pix_clk_100hz * 100)) * 1000000;
+       vblank1_us = (((subvp_pipes[0]->stream->timing.v_total - subvp_pipes[0]->stream->timing.v_addressable) *
+                       subvp_pipes[0]->stream->timing.h_total) /
+                       (double)(subvp_pipes[0]->stream->timing.pix_clk_100hz * 100)) * 1000000;
+       vblank2_us = (((subvp_pipes[1]->stream->timing.v_total - subvp_pipes[1]->stream->timing.v_addressable) *
+                       subvp_pipes[1]->stream->timing.h_total) /
+                       (double)(subvp_pipes[1]->stream->timing.pix_clk_100hz * 100)) * 1000000;
+
+       if ((vactive1_us - vblank2_us) / 2 > max_microschedule_us &&
+           (vactive2_us - vblank1_us) / 2 > max_microschedule_us)
+               return true;
+
+       return false;
+}
+
+/*
+ * dml2_svp_drr_schedulable: Determine if SubVP + DRR config is schedulable
+ *
+ * High level algorithm:
+ * 1. Get timing for SubVP pipe, phantom pipe, and DRR pipe
+ * 2. Determine the frame time for the DRR display when adding required margin for MCLK switching
+ * (the margin is equal to the MALL region + DRR margin (500us))
+ * 3.If (SubVP Active - Prefetch > Stretched DRR frame + max(MALL region, Stretched DRR frame))
+ * then report the configuration as supported
+ *
+ * @dc: current dc state
+ * @context: new dc state
+ * @drr_pipe: DRR pipe_ctx for the SubVP + DRR config
+ *
+ * Return:
+ * bool - True if the SubVP + DRR config is schedulable, false otherwise
+ */
+bool dml2_svp_drr_schedulable(struct dml2_context *ctx, struct dc_state *context, struct dc_crtc_timing *drr_timing)
+{
+       bool schedulable = false;
+       uint32_t i;
+       struct pipe_ctx *pipe = NULL;
+       struct dc_crtc_timing *main_timing = NULL;
+       struct dc_crtc_timing *phantom_timing = NULL;
+       int16_t prefetch_us = 0;
+       int16_t mall_region_us = 0;
+       int16_t drr_frame_us = 0;       // nominal frame time
+       int16_t subvp_active_us = 0;
+       int16_t stretched_drr_us = 0;
+       int16_t drr_stretched_vblank_us = 0;
+       int16_t max_vblank_mallregion = 0;
+
+       // Find SubVP pipe
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               pipe = &context->res_ctx.pipe_ctx[i];
+
+               // We check for master pipe, but it shouldn't matter since we only need
+               // the pipe for timing info (stream should be same for any pipe splits)
+               if (!pipe->stream || !pipe->plane_state || pipe->top_pipe || pipe->prev_odm_pipe)
+                       continue;
+
+               // Find the SubVP pipe
+               if (pipe->stream->mall_stream_config.type == SUBVP_MAIN)
+                       break;
+       }
+
+       main_timing = &pipe->stream->timing;
+       phantom_timing = &pipe->stream->mall_stream_config.paired_stream->timing;
+       prefetch_us = (phantom_timing->v_total - phantom_timing->v_front_porch) * phantom_timing->h_total /
+                       (double)(phantom_timing->pix_clk_100hz * 100) * 1000000 +
+                       ctx->config.svp_pstate.subvp_prefetch_end_to_mall_start_us;
+       subvp_active_us = main_timing->v_addressable * main_timing->h_total /
+                       (double)(main_timing->pix_clk_100hz * 100) * 1000000;
+       drr_frame_us = drr_timing->v_total * drr_timing->h_total /
+                       (double)(drr_timing->pix_clk_100hz * 100) * 1000000;
+       // P-State allow width and FW delays already included phantom_timing->v_addressable
+       mall_region_us = phantom_timing->v_addressable * phantom_timing->h_total /
+                       (double)(phantom_timing->pix_clk_100hz * 100) * 1000000;
+       stretched_drr_us = drr_frame_us + mall_region_us + SUBVP_DRR_MARGIN_US;
+       drr_stretched_vblank_us = (drr_timing->v_total - drr_timing->v_addressable) * drr_timing->h_total /
+                       (double)(drr_timing->pix_clk_100hz * 100) * 1000000 + (stretched_drr_us - drr_frame_us);
+       max_vblank_mallregion = drr_stretched_vblank_us > mall_region_us ? drr_stretched_vblank_us : mall_region_us;
+
+       /* We consider SubVP + DRR schedulable if the stretched frame duration of the DRR display (i.e. the
+        * highest refresh rate + margin that can support UCLK P-State switch) passes the static analysis
+        * for VBLANK: (VACTIVE region of the SubVP pipe can fit the MALL prefetch, VBLANK frame time,
+        * and the max of (VBLANK blanking time, MALL region)).
+        */
+       if (stretched_drr_us < (1 / (double)drr_timing->min_refresh_in_uhz) * 1000000 * 1000000 &&
+                       subvp_active_us - prefetch_us - stretched_drr_us - max_vblank_mallregion > 0)
+               schedulable = true;
+
+       return schedulable;
+}
+
+
+/*
+ * subvp_vblank_schedulable: Determine if SubVP + VBLANK config is schedulable
+ *
+ * High level algorithm:
+ * 1. Get timing for SubVP pipe, phantom pipe, and VBLANK pipe
+ * 2. If (SubVP Active - Prefetch > Vblank Frame Time + max(MALL region, Vblank blanking time))
+ * then report the configuration as supported
+ * 3. If the VBLANK display is DRR, then take the DRR static schedulability path
+ *
+ * @dc: current dc state
+ * @context: new dc state
+ *
+ * Return:
+ * bool - True if the SubVP + VBLANK/DRR config is schedulable, false otherwise
+ */
+static bool subvp_vblank_schedulable(struct dml2_context *ctx, struct dc_state *context)
+{
+       struct pipe_ctx *pipe = NULL;
+       struct pipe_ctx *subvp_pipe = NULL;
+       bool found = false;
+       bool schedulable = false;
+       uint32_t i = 0;
+       uint8_t vblank_index = 0;
+       uint16_t prefetch_us = 0;
+       uint16_t mall_region_us = 0;
+       uint16_t vblank_frame_us = 0;
+       uint16_t subvp_active_us = 0;
+       uint16_t vblank_blank_us = 0;
+       uint16_t max_vblank_mallregion = 0;
+       struct dc_crtc_timing *main_timing = NULL;
+       struct dc_crtc_timing *phantom_timing = NULL;
+       struct dc_crtc_timing *vblank_timing = NULL;
+
+       /* For SubVP + VBLANK/DRR cases, we assume there can only be
+        * a single VBLANK/DRR display. If DML outputs SubVP + VBLANK
+        * is supported, it is either a single VBLANK case or two VBLANK
+        * displays which are synchronized (in which case they have identical
+        * timings).
+        */
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               pipe = &context->res_ctx.pipe_ctx[i];
+
+               // We check for master pipe, but it shouldn't matter since we only need
+               // the pipe for timing info (stream should be same for any pipe splits)
+               if (!pipe->stream || !pipe->plane_state || pipe->top_pipe || pipe->prev_odm_pipe)
+                       continue;
+
+               if (!found && pipe->stream->mall_stream_config.type == SUBVP_NONE) {
+                       // Found pipe which is not SubVP or Phantom (i.e. the VBLANK pipe).
+                       vblank_index = i;
+                       found = true;
+               }
+
+               if (!subvp_pipe && pipe->stream->mall_stream_config.type == SUBVP_MAIN)
+                       subvp_pipe = pipe;
+       }
+       // Use ignore_msa_timing_param flag to identify as DRR
+       if (found && context->res_ctx.pipe_ctx[vblank_index].stream->ignore_msa_timing_param) {
+               // SUBVP + DRR case
+               schedulable = dml2_svp_drr_schedulable(ctx, context, &context->res_ctx.pipe_ctx[vblank_index].stream->timing);
+       } else if (found) {
+               main_timing = &subvp_pipe->stream->timing;
+               phantom_timing = &subvp_pipe->stream->mall_stream_config.paired_stream->timing;
+               vblank_timing = &context->res_ctx.pipe_ctx[vblank_index].stream->timing;
+               // Prefetch time is equal to VACTIVE + BP + VSYNC of the phantom pipe
+               // Also include the prefetch end to mallstart delay time
+               prefetch_us = (phantom_timing->v_total - phantom_timing->v_front_porch) * phantom_timing->h_total /
+                               (double)(phantom_timing->pix_clk_100hz * 100) * 1000000 +
+                               ctx->config.svp_pstate.subvp_prefetch_end_to_mall_start_us;
+               // P-State allow width and FW delays already included phantom_timing->v_addressable
+               mall_region_us = phantom_timing->v_addressable * phantom_timing->h_total /
+                               (double)(phantom_timing->pix_clk_100hz * 100) * 1000000;
+               vblank_frame_us = vblank_timing->v_total * vblank_timing->h_total /
+                               (double)(vblank_timing->pix_clk_100hz * 100) * 1000000;
+               vblank_blank_us =  (vblank_timing->v_total - vblank_timing->v_addressable) * vblank_timing->h_total /
+                               (double)(vblank_timing->pix_clk_100hz * 100) * 1000000;
+               subvp_active_us = main_timing->v_addressable * main_timing->h_total /
+                               (double)(main_timing->pix_clk_100hz * 100) * 1000000;
+               max_vblank_mallregion = vblank_blank_us > mall_region_us ? vblank_blank_us : mall_region_us;
+
+               // Schedulable if VACTIVE region of the SubVP pipe can fit the MALL prefetch, VBLANK frame time,
+               // and the max of (VBLANK blanking time, MALL region)
+               // TODO: Possibly add some margin (i.e. the below conditions should be [...] > X instead of [...] > 0)
+               if (subvp_active_us - prefetch_us - vblank_frame_us - max_vblank_mallregion > 0)
+                       schedulable = true;
+       }
+       return schedulable;
+}
+
+/*
+ * subvp_validate_static_schedulability: Check which SubVP case is calculated and handle
+ * static analysis based on the case.
+ *
+ * Three cases:
+ * 1. SubVP + SubVP
+ * 2. SubVP + VBLANK (DRR checked internally)
+ * 3. SubVP + VACTIVE (currently unsupported)
+ *
+ * @dc: current dc state
+ * @context: new dc state
+ * @vlevel: Voltage level calculated by DML
+ *
+ * Return:
+ * bool - True if statically schedulable, false otherwise
+ */
+bool dml2_svp_validate_static_schedulability(struct dml2_context *ctx, struct dc_state *context, enum dml_dram_clock_change_support pstate_change_type)
+{
+       bool schedulable = true;        // true by default for single display case
+       struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+       uint32_t i, pipe_idx;
+       uint8_t subvp_count = 0;
+       uint8_t vactive_count = 0;
+
+       for (i = 0, pipe_idx = 0; i < ctx->config.dcn_pipe_count; i++) {
+               struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+               if (!pipe->stream)
+                       continue;
+
+               if (pipe->plane_state && !pipe->top_pipe &&
+                               pipe->stream->mall_stream_config.type == SUBVP_MAIN)
+                       subvp_count++;
+
+               // Count how many planes that aren't SubVP/phantom are capable of VACTIVE
+               // switching (SubVP + VACTIVE unsupported). In situations where we force
+               // SubVP for a VACTIVE plane, we don't want to increment the vactive_count.
+               if (vba->ActiveDRAMClockChangeLatencyMargin[vba->pipe_plane[pipe_idx]] > 0 &&
+                   pipe->stream->mall_stream_config.type == SUBVP_NONE) {
+                       vactive_count++;
+               }
+               pipe_idx++;
+       }
+
+       if (subvp_count == 2) {
+               // Static schedulability check for SubVP + SubVP case
+               schedulable = subvp_subvp_schedulable(ctx, context);
+       } else if (pstate_change_type == dml_dram_clock_change_vblank_w_mall_sub_vp) {
+               // Static schedulability check for SubVP + VBLANK case. Also handle the case where
+               // DML outputs SubVP + VBLANK + VACTIVE (DML will report as SubVP + VBLANK)
+               if (vactive_count > 0)
+                       schedulable = false;
+               else
+                       schedulable = subvp_vblank_schedulable(ctx, context);
+       } else if (pstate_change_type == dml_dram_clock_change_vactive_w_mall_sub_vp &&
+                       vactive_count > 0) {
+               // For single display SubVP cases, DML will output dm_dram_clock_change_vactive_w_mall_sub_vp by default.
+               // We tell the difference between SubVP vs. SubVP + VACTIVE by checking the vactive_count.
+               // SubVP + VACTIVE currently unsupported
+               schedulable = false;
+       }
+       return schedulable;
+}
+
+static void set_phantom_stream_timing(struct dml2_context *ctx, struct dc_state *state,
+                                    struct pipe_ctx *ref_pipe,
+                                    struct dc_stream_state *phantom_stream,
+                                    unsigned int dc_pipe_idx,
+                                    unsigned int svp_height,
+                                    unsigned int svp_vstartup)
+{
+       unsigned int i, pipe_idx;
+       double line_time, fp_and_sync_width_time;
+       struct pipe_ctx *pipe;
+       uint32_t phantom_vactive, phantom_bp, pstate_width_fw_delay_lines;
+       static const double cvt_rb_vblank_max = ((double) 460 / (1000 * 1000));
+
+       // Find DML pipe index (pipe_idx) using dc_pipe_idx
+       for (i = 0, pipe_idx = 0; i < ctx->config.dcn_pipe_count; i++) {
+               pipe = &state->res_ctx.pipe_ctx[i];
+
+               if (!pipe->stream)
+                       continue;
+
+               if (i == dc_pipe_idx)
+                       break;
+
+               pipe_idx++;
+       }
+
+       // Calculate lines required for pstate allow width and FW processing delays
+       pstate_width_fw_delay_lines = ((double)(ctx->config.svp_pstate.subvp_fw_processing_delay_us +
+                       ctx->config.svp_pstate.subvp_pstate_allow_width_us) / 1000000) *
+                       (ref_pipe->stream->timing.pix_clk_100hz * 100) /
+                       (double)ref_pipe->stream->timing.h_total;
+
+       // DML calculation for MALL region doesn't take into account FW delay
+       // and required pstate allow width for multi-display cases
+       /* Add 16 lines margin to the MALL REGION because SUB_VP_START_LINE must be aligned
+        * to 2 swaths (i.e. 16 lines)
+        */
+       phantom_vactive = svp_height + pstate_width_fw_delay_lines + ctx->config.svp_pstate.subvp_swath_height_margin_lines;
+
+       phantom_stream->timing.v_front_porch = 1;
+
+       line_time = phantom_stream->timing.h_total / ((double)phantom_stream->timing.pix_clk_100hz * 100);
+       fp_and_sync_width_time = (phantom_stream->timing.v_front_porch + phantom_stream->timing.v_sync_width) * line_time;
+
+       if ((svp_vstartup * line_time) + fp_and_sync_width_time > cvt_rb_vblank_max) {
+               svp_vstartup = (cvt_rb_vblank_max - fp_and_sync_width_time) / line_time;
+       }
+
+       // For backporch of phantom pipe, use vstartup of the main pipe
+       phantom_bp = svp_vstartup;
+
+       phantom_stream->dst.y = 0;
+       phantom_stream->dst.height = phantom_vactive;
+       phantom_stream->src.y = 0;
+       phantom_stream->src.height = phantom_vactive;
+
+       phantom_stream->timing.v_addressable = phantom_vactive;
+
+       phantom_stream->timing.v_total = phantom_stream->timing.v_addressable +
+                                               phantom_stream->timing.v_front_porch +
+                                               phantom_stream->timing.v_sync_width +
+                                               phantom_bp;
+       phantom_stream->timing.flags.DSC = 0; // Don't need DSC for phantom timing
+}
+
+static struct dc_stream_state *enable_phantom_stream(struct dml2_context *ctx, struct dc_state *state, unsigned int dc_pipe_idx, unsigned int svp_height, unsigned int vstartup)
+{
+       struct pipe_ctx *ref_pipe = &state->res_ctx.pipe_ctx[dc_pipe_idx];
+       struct dc_stream_state *phantom_stream = ctx->config.svp_pstate.callbacks.create_stream_for_sink(ref_pipe->stream->sink);
+
+       phantom_stream->signal = SIGNAL_TYPE_VIRTUAL;
+       phantom_stream->dpms_off = true;
+       phantom_stream->mall_stream_config.type = SUBVP_PHANTOM;
+       phantom_stream->mall_stream_config.paired_stream = ref_pipe->stream;
+       ref_pipe->stream->mall_stream_config.type = SUBVP_MAIN;
+       ref_pipe->stream->mall_stream_config.paired_stream = phantom_stream;
+
+       /* stream has limited viewport and small timing */
+       memcpy(&phantom_stream->timing, &ref_pipe->stream->timing, sizeof(phantom_stream->timing));
+       memcpy(&phantom_stream->src, &ref_pipe->stream->src, sizeof(phantom_stream->src));
+       memcpy(&phantom_stream->dst, &ref_pipe->stream->dst, sizeof(phantom_stream->dst));
+       set_phantom_stream_timing(ctx, state, ref_pipe, phantom_stream, dc_pipe_idx, svp_height, vstartup);
+
+       ctx->config.svp_pstate.callbacks.add_stream_to_ctx(ctx->config.svp_pstate.callbacks.dc, state, phantom_stream);
+       return phantom_stream;
+}
+
+static void enable_phantom_plane(struct dml2_context *ctx,
+               struct dc_state *state,
+               struct dc_stream_state *phantom_stream,
+               unsigned int dc_pipe_idx)
+{
+       struct dc_plane_state *phantom_plane = NULL;
+       struct dc_plane_state *prev_phantom_plane = NULL;
+       struct pipe_ctx *curr_pipe = &state->res_ctx.pipe_ctx[dc_pipe_idx];
+
+       while (curr_pipe) {
+               if (curr_pipe->top_pipe && curr_pipe->top_pipe->plane_state == curr_pipe->plane_state) {
+                       phantom_plane = prev_phantom_plane;
+               } else {
+                       phantom_plane = ctx->config.svp_pstate.callbacks.create_plane(ctx->config.svp_pstate.callbacks.dc);
+               }
+
+               memcpy(&phantom_plane->address, &curr_pipe->plane_state->address, sizeof(phantom_plane->address));
+               memcpy(&phantom_plane->scaling_quality, &curr_pipe->plane_state->scaling_quality,
+                               sizeof(phantom_plane->scaling_quality));
+               memcpy(&phantom_plane->src_rect, &curr_pipe->plane_state->src_rect, sizeof(phantom_plane->src_rect));
+               memcpy(&phantom_plane->dst_rect, &curr_pipe->plane_state->dst_rect, sizeof(phantom_plane->dst_rect));
+               memcpy(&phantom_plane->clip_rect, &curr_pipe->plane_state->clip_rect, sizeof(phantom_plane->clip_rect));
+               memcpy(&phantom_plane->plane_size, &curr_pipe->plane_state->plane_size,
+                               sizeof(phantom_plane->plane_size));
+               memcpy(&phantom_plane->tiling_info, &curr_pipe->plane_state->tiling_info,
+                               sizeof(phantom_plane->tiling_info));
+               memcpy(&phantom_plane->dcc, &curr_pipe->plane_state->dcc, sizeof(phantom_plane->dcc));
+               //phantom_plane->tiling_info.gfx10compatible.compat_level = curr_pipe->plane_state->tiling_info.gfx10compatible.compat_level;
+               phantom_plane->format = curr_pipe->plane_state->format;
+               phantom_plane->rotation = curr_pipe->plane_state->rotation;
+               phantom_plane->visible = curr_pipe->plane_state->visible;
+
+               /* Shadow pipe has small viewport. */
+               phantom_plane->clip_rect.y = 0;
+               phantom_plane->clip_rect.height = phantom_stream->timing.v_addressable;
+
+               phantom_plane->is_phantom = true;
+
+               ctx->config.svp_pstate.callbacks.add_plane_to_context(ctx->config.svp_pstate.callbacks.dc, phantom_stream, phantom_plane, state);
+
+               curr_pipe = curr_pipe->bottom_pipe;
+               prev_phantom_plane = phantom_plane;
+       }
+}
+
+static void add_phantom_pipes_for_main_pipe(struct dml2_context *ctx, struct dc_state *state, unsigned int main_pipe_idx, unsigned int svp_height, unsigned int vstartup)
+{
+       struct dc_stream_state *phantom_stream = NULL;
+       unsigned int i;
+
+       // The index of the DC pipe passed into this function is guarenteed to
+       // be a valid candidate for SubVP (i.e. has a plane, stream, doesn't
+       // already have phantom pipe assigned, etc.) by previous checks.
+       phantom_stream = enable_phantom_stream(ctx, state, main_pipe_idx, svp_height, vstartup);
+       enable_phantom_plane(ctx, state, phantom_stream, main_pipe_idx);
+
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               struct pipe_ctx *pipe = &state->res_ctx.pipe_ctx[i];
+
+               // Build scaling params for phantom pipes which were newly added.
+               // We determine which phantom pipes were added by comparing with
+               // the phantom stream.
+               if (pipe->plane_state && pipe->stream && pipe->stream == phantom_stream &&
+                               pipe->stream->mall_stream_config.type == SUBVP_PHANTOM) {
+                       pipe->stream->use_dynamic_meta = false;
+                       pipe->plane_state->flip_immediate = false;
+                       if (!ctx->config.svp_pstate.callbacks.build_scaling_params(pipe)) {
+                               // Log / remove phantom pipes since failed to build scaling params
+                       }
+               }
+       }
+}
+
+static bool remove_all_planes_for_stream(struct dml2_context *ctx, struct dc_stream_state *stream, struct dc_state *context)
+{
+       int i, old_plane_count;
+       struct dc_stream_status *stream_status = NULL;
+       struct dc_plane_state *del_planes[MAX_SURFACE_NUM] = { 0 };
+
+       for (i = 0; i < context->stream_count; i++)
+                       if (context->streams[i] == stream) {
+                               stream_status = &context->stream_status[i];
+                               break;
+                       }
+
+       if (stream_status == NULL) {
+               return false;
+       }
+
+       old_plane_count = stream_status->plane_count;
+
+       for (i = 0; i < old_plane_count; i++)
+               del_planes[i] = stream_status->plane_states[i];
+
+       for (i = 0; i < old_plane_count; i++)
+               if (!ctx->config.svp_pstate.callbacks.remove_plane_from_context(ctx->config.svp_pstate.callbacks.dc, stream, del_planes[i], context))
+                       return false;
+
+       return true;
+}
+
+bool dml2_svp_remove_all_phantom_pipes(struct dml2_context *ctx, struct dc_state *state)
+{
+       int i;
+       bool removed_pipe = false;
+       struct dc_plane_state *phantom_plane = NULL;
+       struct dc_stream_state *phantom_stream = NULL;
+
+       for (i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               struct pipe_ctx *pipe = &state->res_ctx.pipe_ctx[i];
+               // build scaling params for phantom pipes
+               if (pipe->plane_state && pipe->stream && pipe->stream->mall_stream_config.type == SUBVP_PHANTOM) {
+                       phantom_plane = pipe->plane_state;
+                       phantom_stream = pipe->stream;
+
+                       remove_all_planes_for_stream(ctx, pipe->stream, state);
+                       ctx->config.svp_pstate.callbacks.remove_stream_from_ctx(ctx->config.svp_pstate.callbacks.dc, state, pipe->stream);
+
+                       /* Ref count is incremented on allocation and also when added to the context.
+                        * Therefore we must call release for the the phantom plane and stream once
+                        * they are removed from the ctx to finally decrement the refcount to 0 to free.
+                        */
+                       ctx->config.svp_pstate.callbacks.plane_state_release(phantom_plane);
+                       ctx->config.svp_pstate.callbacks.stream_release(phantom_stream);
+
+                       removed_pipe = true;
+               }
+
+               // Clear all phantom stream info
+               if (pipe->stream) {
+                       pipe->stream->mall_stream_config.type = SUBVP_NONE;
+                       pipe->stream->mall_stream_config.paired_stream = NULL;
+               }
+
+               if (pipe->plane_state) {
+                       pipe->plane_state->is_phantom = false;
+               }
+       }
+       return removed_pipe;
+}
+
+
+/* Conditions for setting up phantom pipes for SubVP:
+ * 1. Not force disable SubVP
+ * 2. Full update (i.e. !fast_validate)
+ * 3. Enough pipes are available to support SubVP (TODO: Which pipes will use VACTIVE / VBLANK / SUBVP?)
+ * 4. Display configuration passes validation
+ * 5. (Config doesn't support MCLK in VACTIVE/VBLANK || dc->debug.force_subvp_mclk_switch)
+ */
+bool dml2_svp_add_phantom_pipe_to_dc_state(struct dml2_context *ctx, struct dc_state *state, struct dml_mode_support_info_st *mode_support_info)
+{
+       unsigned int dc_pipe_idx, dml_pipe_idx;
+       unsigned int svp_height, vstartup;
+
+       if (ctx->config.svp_pstate.force_disable_subvp)
+               return false;
+
+       if (!all_pipes_have_stream_and_plane(ctx, state))
+               return false;
+
+       if (mpo_in_use(state))
+               return false;
+
+       merge_pipes_for_subvp(ctx, state);
+       // to re-initialize viewport after the pipe merge
+       for (int i = 0; i < ctx->config.dcn_pipe_count; i++) {
+               struct pipe_ctx *pipe_ctx = &state->res_ctx.pipe_ctx[i];
+
+               if (!pipe_ctx->plane_state || !pipe_ctx->stream)
+                       continue;
+
+               ctx->config.svp_pstate.callbacks.build_scaling_params(pipe_ctx);
+       }
+
+       if (enough_pipes_for_subvp(ctx, state) && assign_subvp_pipe(ctx, state, &dc_pipe_idx)) {
+               dml_pipe_idx = dml2_helper_find_dml_pipe_idx_by_stream_id(ctx, state->res_ctx.pipe_ctx[dc_pipe_idx].stream->stream_id);
+               svp_height = mode_support_info->SubViewportLinesNeededInMALL[dml_pipe_idx];
+               vstartup = dml_get_vstartup_calculated(&ctx->v20.dml_core_ctx, dml_pipe_idx);
+
+               add_phantom_pipes_for_main_pipe(ctx, state, dc_pipe_idx, svp_height, vstartup);
+
+               return true;
+       }
+
+       return false;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_mall_phantom.h b/drivers/gpu/drm/amd/display/dc/dml2/dml2_mall_phantom.h
new file mode 100644 (file)
index 0000000..c14b4de
--- /dev/null
@@ -0,0 +1,50 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DML2_MALL_PHANTOM_H__
+#define __DML2_MALL_PHANTOM_H__
+
+#include "dml2_dc_types.h"
+#include "display_mode_core_structs.h"
+
+struct dml2_svp_helper_select_best_svp_candidate_params {
+       const struct dml_display_cfg_st *dml_config;
+       const struct dml_mode_support_info_st *mode_support_info;
+       const unsigned int blacklist;
+       unsigned int *candidate_index;
+};
+
+struct dml2_context;
+
+unsigned int dml2_helper_calculate_num_ways_for_subvp(struct dml2_context *ctx, struct dc_state *context);
+
+bool dml2_svp_add_phantom_pipe_to_dc_state(struct dml2_context *ctx, struct dc_state *state, struct dml_mode_support_info_st *mode_support_info);
+
+bool dml2_svp_remove_all_phantom_pipes(struct dml2_context *ctx, struct dc_state *state);
+
+bool dml2_svp_validate_static_schedulability(struct dml2_context *ctx, struct dc_state *context, enum dml_dram_clock_change_support pstate_change_type);
+
+bool dml2_svp_drr_schedulable(struct dml2_context *ctx, struct dc_state *context, struct dc_crtc_timing *drr_timing);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_policy.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_policy.c
new file mode 100644 (file)
index 0000000..0c71a8a
--- /dev/null
@@ -0,0 +1,301 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "dml2_policy.h"
+
+static void get_optimal_ntuple(
+       const struct soc_bounding_box_st *socbb,
+       struct soc_state_bounding_box_st *entry)
+{
+       if (entry->dcfclk_mhz > 0) {
+               float bw_on_sdp = (float)(entry->dcfclk_mhz * socbb->return_bus_width_bytes * ((float)socbb->pct_ideal_sdp_bw_after_urgent / 100));
+
+               entry->fabricclk_mhz = bw_on_sdp / (socbb->return_bus_width_bytes * ((float)socbb->pct_ideal_fabric_bw_after_urgent / 100));
+               entry->dram_speed_mts = bw_on_sdp / (socbb->num_chans *
+                       socbb->dram_channel_width_bytes * ((float)socbb->pct_ideal_dram_bw_after_urgent_pixel_only / 100));
+       } else if (entry->fabricclk_mhz > 0) {
+               float bw_on_fabric = (float)(entry->fabricclk_mhz * socbb->return_bus_width_bytes * ((float)socbb->pct_ideal_fabric_bw_after_urgent / 100));
+
+               entry->dcfclk_mhz = bw_on_fabric / (socbb->return_bus_width_bytes * ((float)socbb->pct_ideal_sdp_bw_after_urgent / 100));
+               entry->dram_speed_mts = bw_on_fabric / (socbb->num_chans *
+                       socbb->dram_channel_width_bytes * ((float)socbb->pct_ideal_dram_bw_after_urgent_pixel_only / 100));
+       } else if (entry->dram_speed_mts > 0) {
+               float bw_on_dram = (float)(entry->dram_speed_mts * socbb->num_chans *
+                       socbb->dram_channel_width_bytes * ((float)socbb->pct_ideal_dram_bw_after_urgent_pixel_only / 100));
+
+               entry->fabricclk_mhz = bw_on_dram / (socbb->return_bus_width_bytes * ((float)socbb->pct_ideal_fabric_bw_after_urgent / 100));
+               entry->dcfclk_mhz = bw_on_dram / (socbb->return_bus_width_bytes * ((float)socbb->pct_ideal_sdp_bw_after_urgent / 100));
+       }
+}
+
+static float calculate_net_bw_in_mbytes_sec(const struct soc_bounding_box_st *socbb,
+       struct soc_state_bounding_box_st *entry)
+{
+       float memory_bw_mbytes_sec = (float)(entry->dram_speed_mts *  socbb->num_chans *
+               socbb->dram_channel_width_bytes * ((float)socbb->pct_ideal_dram_bw_after_urgent_pixel_only / 100));
+
+       float fabric_bw_mbytes_sec = (float)(entry->fabricclk_mhz * socbb->return_bus_width_bytes * ((float)socbb->pct_ideal_fabric_bw_after_urgent / 100));
+
+       float sdp_bw_mbytes_sec = (float)(entry->dcfclk_mhz * socbb->return_bus_width_bytes * ((float)socbb->pct_ideal_sdp_bw_after_urgent / 100));
+
+       float limiting_bw_mbytes_sec = memory_bw_mbytes_sec;
+
+       if (fabric_bw_mbytes_sec < limiting_bw_mbytes_sec)
+               limiting_bw_mbytes_sec = fabric_bw_mbytes_sec;
+
+       if (sdp_bw_mbytes_sec < limiting_bw_mbytes_sec)
+               limiting_bw_mbytes_sec = sdp_bw_mbytes_sec;
+
+       return limiting_bw_mbytes_sec;
+}
+
+static void insert_entry_into_table_sorted(const struct soc_bounding_box_st *socbb,
+       struct soc_states_st *table,
+       struct soc_state_bounding_box_st *entry)
+{
+       int index = 0;
+       int i = 0;
+       float net_bw_of_new_state = 0;
+
+       get_optimal_ntuple(socbb, entry);
+
+       if (table->num_states == 0) {
+               index = 0;
+       } else {
+               net_bw_of_new_state = calculate_net_bw_in_mbytes_sec(socbb, entry);
+               while (net_bw_of_new_state > calculate_net_bw_in_mbytes_sec(socbb, &table->state_array[index])) {
+                       index++;
+                       if (index >= (int) table->num_states)
+                               break;
+               }
+
+               for (i = table->num_states; i > index; i--) {
+                       table->state_array[i] = table->state_array[i - 1];
+               }
+               //ASSERT(index < MAX_CLK_TABLE_SIZE);
+       }
+
+       table->state_array[index] = *entry;
+       table->state_array[index].dcfclk_mhz = (int)entry->dcfclk_mhz;
+       table->state_array[index].fabricclk_mhz = (int)entry->fabricclk_mhz;
+       table->state_array[index].dram_speed_mts = (int)entry->dram_speed_mts;
+       table->num_states++;
+}
+
+static void remove_entry_from_table_at_index(struct soc_states_st *table,
+       unsigned int index)
+{
+       int i;
+
+       if (table->num_states == 0)
+               return;
+
+       for (i = index; i < (int) table->num_states - 1; i++) {
+               table->state_array[i] = table->state_array[i + 1];
+       }
+       memset(&table->state_array[--table->num_states], 0, sizeof(struct soc_state_bounding_box_st));
+}
+
+int dml2_policy_build_synthetic_soc_states(struct dml2_policy_build_synthetic_soc_states_scratch *s,
+       struct dml2_policy_build_synthetic_soc_states_params *p)
+{
+       int i, j;
+       unsigned int min_fclk_mhz = p->in_states->state_array[0].fabricclk_mhz;
+       unsigned int min_dcfclk_mhz = p->in_states->state_array[0].dcfclk_mhz;
+       unsigned int min_socclk_mhz = p->in_states->state_array[0].socclk_mhz;
+
+       int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0,
+               max_phyclk_mhz = 0, max_dtbclk_mhz = 0, max_fclk_mhz = 0,
+               max_uclk_mhz = 0, max_socclk_mhz = 0;
+
+       int num_uclk_dpms = 0, num_fclk_dpms = 0;
+
+       for (i = 0; i < __DML_MAX_STATE_ARRAY_SIZE__; i++) {
+               if (p->in_states->state_array[i].dcfclk_mhz > max_dcfclk_mhz)
+                       max_dcfclk_mhz = (int) p->in_states->state_array[i].dcfclk_mhz;
+               if (p->in_states->state_array[i].fabricclk_mhz > max_fclk_mhz)
+                       max_fclk_mhz = (int) p->in_states->state_array[i].fabricclk_mhz;
+               if (p->in_states->state_array[i].socclk_mhz > max_socclk_mhz)
+                       max_socclk_mhz = (int) p->in_states->state_array[i].socclk_mhz;
+               if (p->in_states->state_array[i].dram_speed_mts > max_uclk_mhz)
+                       max_uclk_mhz = (int) p->in_states->state_array[i].dram_speed_mts;
+               if (p->in_states->state_array[i].dispclk_mhz > max_dispclk_mhz)
+                       max_dispclk_mhz = (int) p->in_states->state_array[i].dispclk_mhz;
+               if (p->in_states->state_array[i].dppclk_mhz > max_dppclk_mhz)
+                       max_dppclk_mhz = (int) p->in_states->state_array[i].dppclk_mhz;
+               if (p->in_states->state_array[i].phyclk_mhz > max_phyclk_mhz)
+                       max_phyclk_mhz = (int)p->in_states->state_array[i].phyclk_mhz;
+               if (p->in_states->state_array[i].dtbclk_mhz > max_dtbclk_mhz)
+                       max_dtbclk_mhz = (int)p->in_states->state_array[i].dtbclk_mhz;
+
+               if (p->in_states->state_array[i].fabricclk_mhz > 0)
+                       num_fclk_dpms++;
+               if (p->in_states->state_array[i].dram_speed_mts > 0)
+                       num_uclk_dpms++;
+       }
+
+       if (!max_dcfclk_mhz || !max_dispclk_mhz || !max_dppclk_mhz || !max_phyclk_mhz || !max_dtbclk_mhz)
+               return -1;
+
+       p->out_states->num_states = 0;
+
+       s->entry = p->in_states->state_array[0];
+
+       s->entry.dispclk_mhz = max_dispclk_mhz;
+       s->entry.dppclk_mhz = max_dppclk_mhz;
+       s->entry.dtbclk_mhz = max_dtbclk_mhz;
+       s->entry.phyclk_mhz = max_phyclk_mhz;
+
+       s->entry.dscclk_mhz = max_dispclk_mhz / 3;
+       s->entry.phyclk_mhz = max_phyclk_mhz;
+       s->entry.dtbclk_mhz = max_dtbclk_mhz;
+
+       // Insert all the DCFCLK STAs first
+       for (i = 0; i < p->num_dcfclk_stas; i++) {
+               s->entry.dcfclk_mhz = p->dcfclk_stas_mhz[i];
+               s->entry.fabricclk_mhz = 0;
+               s->entry.dram_speed_mts = 0;
+               if (i > 0)
+                       s->entry.socclk_mhz = max_socclk_mhz;
+
+               insert_entry_into_table_sorted(p->in_bbox, p->out_states, &s->entry);
+       }
+
+       // Insert the UCLK DPMS
+       for (i = 0; i < num_uclk_dpms; i++) {
+               s->entry.dcfclk_mhz = 0;
+               s->entry.fabricclk_mhz = 0;
+               s->entry.dram_speed_mts = p->in_states->state_array[i].dram_speed_mts;
+               if (i == 0) {
+                       s->entry.socclk_mhz = min_socclk_mhz;
+               } else {
+                       s->entry.socclk_mhz = max_socclk_mhz;
+               }
+
+               insert_entry_into_table_sorted(p->in_bbox, p->out_states, &s->entry);
+       }
+
+       // Insert FCLK DPMs (if present)
+       if (num_fclk_dpms > 2) {
+               for (i = 0; i < num_fclk_dpms; i++) {
+                       s->entry.dcfclk_mhz = 0;
+                       s->entry.fabricclk_mhz = p->in_states->state_array[i].fabricclk_mhz;
+                       s->entry.dram_speed_mts = 0;
+
+               insert_entry_into_table_sorted(p->in_bbox, p->out_states, &s->entry);
+               }
+       }
+       // Add max FCLK
+       else {
+               s->entry.dcfclk_mhz = 0;
+               s->entry.fabricclk_mhz = p->in_states->state_array[num_fclk_dpms - 1].fabricclk_mhz;
+               s->entry.dram_speed_mts = 0;
+
+               insert_entry_into_table_sorted(p->in_bbox, p->out_states, &s->entry);
+       }
+
+       // Remove states that require higher clocks than are supported
+       for (i = p->out_states->num_states - 1; i >= 0; i--) {
+               if (p->out_states->state_array[i].dcfclk_mhz > max_dcfclk_mhz ||
+                       p->out_states->state_array[i].fabricclk_mhz > max_fclk_mhz ||
+                       p->out_states->state_array[i].dram_speed_mts > max_uclk_mhz)
+                       remove_entry_from_table_at_index(p->out_states, i);
+       }
+
+       // At this point, the table contains all "points of interest" based on
+       // DPMs from PMFW, and STAs. Table is sorted by BW, and all clock
+       // ratios (by derate, are exact).
+
+       // Round up UCLK to DPMs
+       for (i = p->out_states->num_states - 1; i >= 0; i--) {
+               for (j = 0; j < num_uclk_dpms; j++) {
+                       if (p->in_states->state_array[j].dram_speed_mts >= p->out_states->state_array[i].dram_speed_mts) {
+                               p->out_states->state_array[i].dram_speed_mts = p->in_states->state_array[j].dram_speed_mts;
+                               break;
+                       }
+               }
+       }
+
+       // If FCLK is coarse grained, round up to next DPMs
+       if (num_fclk_dpms > 2) {
+               for (i = p->out_states->num_states - 1; i >= 0; i--) {
+                       for (j = 0; j < num_fclk_dpms; j++) {
+                               if (p->in_states->state_array[j].fabricclk_mhz >= p->out_states->state_array[i].fabricclk_mhz) {
+                                       p->out_states->state_array[i].fabricclk_mhz = p->in_states->state_array[j].fabricclk_mhz;
+                                       break;
+                               }
+                       }
+               }
+       }
+
+       // Clamp to min FCLK/DCFCLK
+       for (i = p->out_states->num_states - 1; i >= 0; i--) {
+               if (p->out_states->state_array[i].fabricclk_mhz < min_fclk_mhz) {
+                       p->out_states->state_array[i].fabricclk_mhz = min_fclk_mhz;
+               }
+               if (p->out_states->state_array[i].dcfclk_mhz < min_dcfclk_mhz) {
+                       p->out_states->state_array[i].dcfclk_mhz = min_dcfclk_mhz;
+               }
+       }
+
+       // Remove duplicate states, note duplicate states are always neighbouring since table is sorted.
+       i = 0;
+       while (i < (int) p->out_states->num_states - 1) {
+               if (p->out_states->state_array[i].dcfclk_mhz == p->out_states->state_array[i + 1].dcfclk_mhz &&
+                       p->out_states->state_array[i].fabricclk_mhz == p->out_states->state_array[i + 1].fabricclk_mhz &&
+                       p->out_states->state_array[i].dram_speed_mts == p->out_states->state_array[i + 1].dram_speed_mts)
+                       remove_entry_from_table_at_index(p->out_states, i);
+       else
+               i++;
+       }
+
+       return 0;
+}
+
+void build_unoptimized_policy_settings(enum dml_project_id project, struct dml_mode_eval_policy_st *policy)
+{
+       for (int i = 0; i < __DML_NUM_PLANES__; i++) {
+               policy->MPCCombineUse[i] = dml_mpc_as_needed_for_voltage; // TOREVIEW: Is this still needed?  When is MPCC useful for pstate given CRB?
+               policy->ODMUse[i] = dml_odm_use_policy_combine_as_needed;
+               policy->ImmediateFlipRequirement[i] = dml_immediate_flip_required;
+               policy->AllowForPStateChangeOrStutterInVBlank[i] = dml_prefetch_support_uclk_fclk_and_stutter_if_possible;
+       }
+
+       /* Change the default policy initializations as per spreadsheet. We might need to
+        * review and change them later on as per Jun's earlier comments.
+        */
+       policy->UseUnboundedRequesting = dml_unbounded_requesting_enable;
+       policy->UseMinimumRequiredDCFCLK = false;
+       policy->DRAMClockChangeRequirementFinal = true; // TOREVIEW: What does this mean?
+       policy->FCLKChangeRequirementFinal = true; // TOREVIEW: What does this mean?
+       policy->USRRetrainingRequiredFinal = true;
+       policy->EnhancedPrefetchScheduleAccelerationFinal = true; // TOREVIEW: What does this mean?
+       policy->NomDETInKByteOverrideEnable = false;
+       policy->NomDETInKByteOverrideValue = 0;
+       policy->DCCProgrammingAssumesScanDirectionUnknownFinal = true;
+       policy->SynchronizeTimingsFinal = true;
+       policy->SynchronizeDRRDisplaysForUCLKPStateChangeFinal = true;
+       policy->AssumeModeSupportAtMaxPwrStateEvenDRAMClockChangeNotSupported = true; // TOREVIEW: What does this mean?
+       policy->AssumeModeSupportAtMaxPwrStateEvenFClockChangeNotSupported = true; // TOREVIEW: What does this mean?
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_policy.h b/drivers/gpu/drm/amd/display/dc/dml2/dml2_policy.h
new file mode 100644 (file)
index 0000000..e83e052
--- /dev/null
@@ -0,0 +1,47 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DML2_POLICY_H__
+#define __DML2_POLICY_H__
+
+#include "display_mode_core_structs.h"
+
+struct dml2_policy_build_synthetic_soc_states_params {
+       const struct soc_bounding_box_st *in_bbox;
+       struct soc_states_st *in_states;
+       struct soc_states_st *out_states;
+       int *dcfclk_stas_mhz;
+       int num_dcfclk_stas;
+};
+
+struct dml2_policy_build_synthetic_soc_states_scratch {
+       struct soc_state_bounding_box_st entry;
+};
+
+int dml2_policy_build_synthetic_soc_states(struct dml2_policy_build_synthetic_soc_states_scratch *s,
+       struct dml2_policy_build_synthetic_soc_states_params *p);
+
+void build_unoptimized_policy_settings(enum dml_project_id project, struct dml_mode_eval_policy_st *policy);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c
new file mode 100644 (file)
index 0000000..353f22a
--- /dev/null
@@ -0,0 +1,1109 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "display_mode_core.h"
+#include "dml2_internal_types.h"
+#include "dml2_translation_helper.h"
+
+#define NUM_DCFCLK_STAS 5
+
+void dml2_init_ip_params(struct dml2_context *dml2, const struct dc *in_dc, struct ip_params_st *out)
+{
+       switch (dml2->v20.dml_core_ctx.project) {
+
+       case dml_project_dcn32:
+       case dml_project_dcn321:
+       default:
+               // Hardcoded values for DCN32x
+               out->vblank_nom_default_us = 600;
+               out->rob_buffer_size_kbytes = 128;
+               out->config_return_buffer_size_in_kbytes = 1280;
+               out->config_return_buffer_segment_size_in_kbytes = 64;
+               out->compressed_buffer_segment_size_in_kbytes = 64;
+               out->meta_fifo_size_in_kentries = 22;
+               out->zero_size_buffer_entries = 512;
+               out->dpte_buffer_size_in_pte_reqs_luma = 68;
+               out->dpte_buffer_size_in_pte_reqs_chroma = 36;
+               out->dcc_meta_buffer_size_bytes = 6272;
+               out->gpuvm_max_page_table_levels = 4;
+               out->hostvm_max_page_table_levels = 0;
+               out->pixel_chunk_size_kbytes = 8;
+               //out->alpha_pixel_chunk_size_kbytes;
+               out->min_pixel_chunk_size_bytes = 1024;
+               out->meta_chunk_size_kbytes = 2;
+               out->min_meta_chunk_size_bytes = 256;
+               out->writeback_chunk_size_kbytes = 8;
+               out->line_buffer_size_bits = 1171920;
+               out->max_line_buffer_lines = 32;
+               out->writeback_interface_buffer_size_kbytes = 90;
+               //Number of pipes after DCN Pipe harvesting
+               out->max_num_dpp = dml2->config.dcn_pipe_count;
+               out->max_num_otg = dml2->config.dcn_pipe_count;
+               out->max_num_wb = 1;
+               out->max_dchub_pscl_bw_pix_per_clk = 4;
+               out->max_pscl_lb_bw_pix_per_clk = 2;
+               out->max_lb_vscl_bw_pix_per_clk = 4;
+               out->max_vscl_hscl_bw_pix_per_clk = 4;
+               out->max_hscl_ratio = 6;
+               out->max_vscl_ratio = 6;
+               out->max_hscl_taps = 8;
+               out->max_vscl_taps = 8;
+               out->dispclk_ramp_margin_percent = 1;
+               out->dppclk_delay_subtotal = 47;
+               out->dppclk_delay_scl = 50;
+               out->dppclk_delay_scl_lb_only = 16;
+               out->dppclk_delay_cnvc_formatter = 28;
+               out->dppclk_delay_cnvc_cursor = 6;
+               out->cursor_buffer_size = 16;
+               out->cursor_chunk_size = 2;
+               out->dispclk_delay_subtotal = 125;
+               out->max_inter_dcn_tile_repeaters = 8;
+               out->writeback_max_hscl_ratio = 1;
+               out->writeback_max_vscl_ratio = 1;
+               out->writeback_min_hscl_ratio = 1;
+               out->writeback_min_vscl_ratio = 1;
+               out->writeback_max_hscl_taps = 1;
+               out->writeback_max_vscl_taps = 1;
+               out->writeback_line_buffer_buffer_size = 0;
+               out->num_dsc = 4;
+               out->maximum_dsc_bits_per_component = 12;
+               out->maximum_pixels_per_line_per_dsc_unit = 6016;
+               out->dsc422_native_support = true;
+               out->dcc_supported = true;
+               out->ptoi_supported = false;
+
+               out->gpuvm_enable = false;
+               out->hostvm_enable = false;
+               out->cursor_64bpp_support = false;
+               out->dynamic_metadata_vm_enabled = false;
+
+               out->max_num_hdmi_frl_outputs = 1;
+               out->max_num_dp2p0_outputs = 2;
+               out->max_num_dp2p0_streams = 4;
+               break;
+
+       }
+}
+
+void dml2_init_socbb_params(struct dml2_context *dml2, const struct dc *in_dc, struct soc_bounding_box_st *out)
+{
+       out->dprefclk_mhz = dml2->config.bbox_overrides.dprefclk_mhz;
+       out->xtalclk_mhz = dml2->config.bbox_overrides.xtalclk_mhz;
+       out->pcierefclk_mhz = 100;
+       out->refclk_mhz = dml2->config.bbox_overrides.dchub_refclk_mhz;
+
+       out->max_outstanding_reqs = 512;
+       out->pct_ideal_sdp_bw_after_urgent = 100;
+       out->pct_ideal_fabric_bw_after_urgent = 67;
+       out->pct_ideal_dram_bw_after_urgent_pixel_only = 20;
+       out->pct_ideal_dram_bw_after_urgent_pixel_and_vm = 60;
+       out->pct_ideal_dram_bw_after_urgent_vm_only = 30;
+       out->pct_ideal_dram_bw_after_urgent_strobe = 67;
+       out->max_avg_sdp_bw_use_normal_percent = 80;
+       out->max_avg_fabric_bw_use_normal_percent = 60;
+       out->max_avg_dram_bw_use_normal_percent = 15;
+       out->max_avg_dram_bw_use_normal_strobe_percent = 50;
+
+       out->urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096;
+       out->urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096;
+       out->urgent_out_of_order_return_per_channel_vm_only_bytes = 4096;
+       out->return_bus_width_bytes = 64;
+       out->dram_channel_width_bytes = 2;
+       out->fabric_datapath_to_dcn_data_return_bytes = 64;
+       out->hostvm_min_page_size_kbytes = 0;
+       out->gpuvm_min_page_size_kbytes = 256;
+       out->phy_downspread_percent = 0.38;
+       out->dcn_downspread_percent = 0.5;
+       out->dispclk_dppclk_vco_speed_mhz = dml2->config.bbox_overrides.disp_pll_vco_speed_mhz;
+       out->mall_allocated_for_dcn_mbytes = dml2->config.mall_cfg.max_cab_allocation_bytes / 1048576; // 64 or 32 MB;
+
+       out->do_urgent_latency_adjustment = true;
+
+       switch (dml2->v20.dml_core_ctx.project) {
+
+       case dml_project_dcn32:
+       default:
+               out->num_chans = 24;
+               out->round_trip_ping_latency_dcfclk_cycles = 263;
+               out->smn_latency_us = 2;
+               break;
+
+       case dml_project_dcn321:
+               out->num_chans = 8;
+               out->round_trip_ping_latency_dcfclk_cycles = 207;
+               out->smn_latency_us = 0;
+               break;
+
+       }
+       /* ---Overrides if available--- */
+       if (dml2->config.bbox_overrides.dram_num_chan)
+               out->num_chans = dml2->config.bbox_overrides.dram_num_chan;
+
+       if (dml2->config.bbox_overrides.dram_chanel_width_bytes)
+               out->dram_channel_width_bytes = dml2->config.bbox_overrides.dram_chanel_width_bytes;
+}
+
+void dml2_init_soc_states(struct dml2_context *dml2, const struct dc *in_dc,
+       const struct soc_bounding_box_st *in_bbox, struct soc_states_st *out)
+{
+       struct dml2_policy_build_synthetic_soc_states_scratch *s = &dml2->v20.scratch.create_scratch.build_synthetic_socbb_scratch;
+       struct dml2_policy_build_synthetic_soc_states_params *p = &dml2->v20.scratch.build_synthetic_socbb_params;
+       unsigned int dcfclk_stas_mhz[NUM_DCFCLK_STAS];
+       unsigned int i = 0;
+       unsigned int transactions_per_mem_clock = 16; // project specific, depends on used Memory type
+
+       p->dcfclk_stas_mhz = dcfclk_stas_mhz;
+       p->num_dcfclk_stas = NUM_DCFCLK_STAS;
+       p->in_bbox = in_bbox;
+       p->out_states = out;
+       p->in_states = &dml2->v20.scratch.create_scratch.in_states;
+
+
+       /* Initial hardcoded values */
+       switch (dml2->v20.dml_core_ctx.project) {
+
+       case dml_project_dcn32:
+       default:
+               p->in_states->num_states = 2;
+               transactions_per_mem_clock = 16;
+               p->in_states->state_array[0].socclk_mhz = 620.0;
+               p->in_states->state_array[0].dscclk_mhz = 716.667;
+               p->in_states->state_array[0].phyclk_mhz = 810;
+               p->in_states->state_array[0].phyclk_d18_mhz = 667;
+               p->in_states->state_array[0].phyclk_d32_mhz = 625;
+               p->in_states->state_array[0].dtbclk_mhz = 1564.0;
+               p->in_states->state_array[0].fabricclk_mhz = 450.0;
+               p->in_states->state_array[0].dcfclk_mhz = 300.0;
+               p->in_states->state_array[0].dispclk_mhz = 2150.0;
+               p->in_states->state_array[0].dppclk_mhz = 2150.0;
+               p->in_states->state_array[0].dram_speed_mts = 100 * transactions_per_mem_clock;
+
+               p->in_states->state_array[0].urgent_latency_pixel_data_only_us = 4;
+               p->in_states->state_array[0].urgent_latency_pixel_mixed_with_vm_data_us = 0;
+               p->in_states->state_array[0].urgent_latency_vm_data_only_us = 0;
+               p->in_states->state_array[0].writeback_latency_us = 12;
+               p->in_states->state_array[0].urgent_latency_adjustment_fabric_clock_component_us = 1;
+               p->in_states->state_array[0].urgent_latency_adjustment_fabric_clock_reference_mhz = 3000;
+               p->in_states->state_array[0].sr_exit_z8_time_us = 0;
+               p->in_states->state_array[0].sr_enter_plus_exit_z8_time_us = 0;
+               p->in_states->state_array[0].dram_clock_change_latency_us = 400;
+               p->in_states->state_array[0].use_ideal_dram_bw_strobe = true;
+               p->in_states->state_array[0].sr_exit_time_us = 42.97;
+               p->in_states->state_array[0].sr_enter_plus_exit_time_us = 49.94;
+               p->in_states->state_array[0].fclk_change_latency_us = 20;
+               p->in_states->state_array[0].usr_retraining_latency_us = 2;
+
+               p->in_states->state_array[1].socclk_mhz = 1200.0;
+               p->in_states->state_array[1].fabricclk_mhz = 2500.0;
+               p->in_states->state_array[1].dcfclk_mhz = 1564.0;
+               p->in_states->state_array[1].dram_speed_mts = 1125 * transactions_per_mem_clock;
+               break;
+
+       case dml_project_dcn321:
+               p->in_states->num_states = 2;
+               transactions_per_mem_clock = 16;
+               p->in_states->state_array[0].socclk_mhz = 582.0;
+               p->in_states->state_array[0].dscclk_mhz = 573.333;
+               p->in_states->state_array[0].phyclk_mhz = 810;
+               p->in_states->state_array[0].phyclk_d18_mhz = 667;
+               p->in_states->state_array[0].phyclk_d32_mhz = 313;
+               p->in_states->state_array[0].dtbclk_mhz = 1564.0;
+               p->in_states->state_array[0].fabricclk_mhz = 450.0;
+               p->in_states->state_array[0].dcfclk_mhz = 300.0;
+               p->in_states->state_array[0].dispclk_mhz = 1720.0;
+               p->in_states->state_array[0].dppclk_mhz = 1720.0;
+               p->in_states->state_array[0].dram_speed_mts = 100 * transactions_per_mem_clock;
+
+               p->in_states->state_array[0].urgent_latency_pixel_data_only_us = 4;
+               p->in_states->state_array[0].urgent_latency_pixel_mixed_with_vm_data_us = 0;
+               p->in_states->state_array[0].urgent_latency_vm_data_only_us = 0;
+               p->in_states->state_array[0].writeback_latency_us = 12;
+               p->in_states->state_array[0].urgent_latency_adjustment_fabric_clock_component_us = 1;
+               p->in_states->state_array[0].urgent_latency_adjustment_fabric_clock_reference_mhz = 3000;
+               p->in_states->state_array[0].sr_exit_z8_time_us = 0;
+               p->in_states->state_array[0].sr_enter_plus_exit_z8_time_us = 0;
+               p->in_states->state_array[0].dram_clock_change_latency_us = 400;
+               p->in_states->state_array[0].use_ideal_dram_bw_strobe = true;
+               p->in_states->state_array[0].sr_exit_time_us = 19.95;
+               p->in_states->state_array[0].sr_enter_plus_exit_time_us = 24.36;
+               p->in_states->state_array[0].fclk_change_latency_us = 7;
+               p->in_states->state_array[0].usr_retraining_latency_us = 0;
+
+               p->in_states->state_array[1].socclk_mhz = 1200.0;
+               p->in_states->state_array[1].fabricclk_mhz = 2250.0;
+               p->in_states->state_array[1].dcfclk_mhz = 1434.0;
+               p->in_states->state_array[1].dram_speed_mts = 1000 * transactions_per_mem_clock;
+               break;
+
+       }
+
+       /* Override from passed values, mainly for debugging purposes, if available */
+       if (dml2->config.bbox_overrides.sr_exit_latency_us) {
+               p->in_states->state_array[0].sr_exit_time_us = dml2->config.bbox_overrides.sr_exit_latency_us;
+       }
+
+       if (dml2->config.bbox_overrides.sr_enter_plus_exit_latency_us) {
+               p->in_states->state_array[0].sr_enter_plus_exit_time_us = dml2->config.bbox_overrides.sr_enter_plus_exit_latency_us;
+       }
+
+       if (dml2->config.bbox_overrides.urgent_latency_us) {
+               p->in_states->state_array[0].urgent_latency_pixel_data_only_us = dml2->config.bbox_overrides.urgent_latency_us;
+       }
+
+       if (dml2->config.bbox_overrides.dram_clock_change_latency_us) {
+               p->in_states->state_array[0].dram_clock_change_latency_us = dml2->config.bbox_overrides.dram_clock_change_latency_us;
+       }
+
+       if (dml2->config.bbox_overrides.fclk_change_latency_us) {
+               p->in_states->state_array[0].fclk_change_latency_us = dml2->config.bbox_overrides.fclk_change_latency_us;
+       }
+
+       /* DCFCLK stas values are project specific */
+       if ((dml2->v20.dml_core_ctx.project == dml_project_dcn32) ||
+               (dml2->v20.dml_core_ctx.project == dml_project_dcn321)) {
+               p->dcfclk_stas_mhz[0] = p->in_states->state_array[0].dcfclk_mhz;
+               p->dcfclk_stas_mhz[1] = 615;
+               p->dcfclk_stas_mhz[2] = 906;
+               p->dcfclk_stas_mhz[3] = 1324;
+               p->dcfclk_stas_mhz[4] = p->in_states->state_array[1].dcfclk_mhz;
+       }
+       /* Copy clocks tables entries, if available */
+       if (dml2->config.bbox_overrides.clks_table.num_states) {
+               p->in_states->num_states = dml2->config.bbox_overrides.clks_table.num_states;
+
+               for (i = 0; i < dml2->config.bbox_overrides.clks_table.num_entries_per_clk.num_dcfclk_levels; i++) {
+                       p->in_states->state_array[i].dcfclk_mhz = dml2->config.bbox_overrides.clks_table.clk_entries[i].dcfclk_mhz;
+               }
+
+               p->dcfclk_stas_mhz[0] = dml2->config.bbox_overrides.clks_table.clk_entries[0].dcfclk_mhz;
+               if (i > 1)
+                       p->dcfclk_stas_mhz[4] = dml2->config.bbox_overrides.clks_table.clk_entries[i-1].dcfclk_mhz;
+
+               for (i = 0; i < dml2->config.bbox_overrides.clks_table.num_entries_per_clk.num_fclk_levels; i++) {
+                       p->in_states->state_array[i].fabricclk_mhz =
+                               dml2->config.bbox_overrides.clks_table.clk_entries[i].fclk_mhz;
+               }
+
+               for (i = 0; i < dml2->config.bbox_overrides.clks_table.num_entries_per_clk.num_memclk_levels; i++) {
+                       p->in_states->state_array[i].dram_speed_mts =
+                               dml2->config.bbox_overrides.clks_table.clk_entries[i].memclk_mhz * transactions_per_mem_clock;
+               }
+
+               for (i = 0; i < dml2->config.bbox_overrides.clks_table.num_entries_per_clk.num_socclk_levels; i++) {
+                       p->in_states->state_array[i].socclk_mhz =
+                               dml2->config.bbox_overrides.clks_table.clk_entries[i].socclk_mhz;
+               }
+
+               for (i = 0; i < dml2->config.bbox_overrides.clks_table.num_entries_per_clk.num_dtbclk_levels; i++) {
+                       p->in_states->state_array[i].dtbclk_mhz =
+                               dml2->config.bbox_overrides.clks_table.clk_entries[i].dtbclk_mhz;
+               }
+
+               for (i = 0; i < dml2->config.bbox_overrides.clks_table.num_entries_per_clk.num_dispclk_levels; i++) {
+                       p->in_states->state_array[i].dispclk_mhz =
+                               dml2->config.bbox_overrides.clks_table.clk_entries[i].dispclk_mhz;
+                       p->in_states->state_array[i].dppclk_mhz =
+                               dml2->config.bbox_overrides.clks_table.clk_entries[i].dppclk_mhz;
+               }
+       }
+
+       dml2_policy_build_synthetic_soc_states(s, p);
+}
+
+void dml2_translate_ip_params(const struct dc *in, struct ip_params_st *out)
+{
+       const struct _vcs_dpi_ip_params_st *in_ip_params = &in->dml.ip;
+       /* Copy over the IP params tp dml2_ctx */
+       out->compressed_buffer_segment_size_in_kbytes = in_ip_params->compressed_buffer_segment_size_in_kbytes;
+       out->config_return_buffer_size_in_kbytes = in_ip_params->config_return_buffer_size_in_kbytes;
+       out->cursor_buffer_size = in_ip_params->cursor_buffer_size;
+       out->cursor_chunk_size = in_ip_params->cursor_chunk_size;
+       out->dcc_meta_buffer_size_bytes = in_ip_params->dcc_meta_buffer_size_bytes;
+       out->dcc_supported = in_ip_params->dcc_supported;
+       out->dispclk_delay_subtotal = in_ip_params->dispclk_delay_subtotal;
+       out->dispclk_ramp_margin_percent = in_ip_params->dispclk_ramp_margin_percent;
+       out->dppclk_delay_cnvc_cursor = in_ip_params->dppclk_delay_cnvc_cursor;
+       out->dppclk_delay_cnvc_formatter = in_ip_params->dppclk_delay_cnvc_formatter;
+       out->dppclk_delay_scl = in_ip_params->dppclk_delay_scl;
+       out->dppclk_delay_scl_lb_only = in_ip_params->dppclk_delay_scl_lb_only;
+       out->dppclk_delay_subtotal = in_ip_params->dppclk_delay_subtotal;
+       out->dpte_buffer_size_in_pte_reqs_chroma = in_ip_params->dpte_buffer_size_in_pte_reqs_chroma;
+       out->dpte_buffer_size_in_pte_reqs_luma = in_ip_params->dpte_buffer_size_in_pte_reqs_luma;
+       out->dsc422_native_support = in_ip_params->dsc422_native_support;
+       out->dynamic_metadata_vm_enabled = in_ip_params->dynamic_metadata_vm_enabled;
+       out->gpuvm_enable = in_ip_params->gpuvm_enable;
+       out->gpuvm_max_page_table_levels = in_ip_params->gpuvm_max_page_table_levels;
+       out->hostvm_enable = in_ip_params->hostvm_enable;
+       out->hostvm_max_page_table_levels = in_ip_params->hostvm_max_page_table_levels;
+       out->line_buffer_size_bits = in_ip_params->line_buffer_size_bits;
+       out->maximum_dsc_bits_per_component = in_ip_params->maximum_dsc_bits_per_component;
+       out->maximum_pixels_per_line_per_dsc_unit = in_ip_params->maximum_pixels_per_line_per_dsc_unit;
+       out->max_dchub_pscl_bw_pix_per_clk = in_ip_params->max_dchub_pscl_bw_pix_per_clk;
+       out->max_hscl_ratio = in_ip_params->max_hscl_ratio;
+       out->max_hscl_taps = in_ip_params->max_hscl_taps;
+       out->max_inter_dcn_tile_repeaters = in_ip_params->max_inter_dcn_tile_repeaters;
+       out->max_lb_vscl_bw_pix_per_clk = in_ip_params->max_lb_vscl_bw_pix_per_clk;
+       out->max_line_buffer_lines = in_ip_params->max_line_buffer_lines;
+       out->max_num_dp2p0_outputs = in_ip_params->max_num_dp2p0_outputs;
+       out->max_num_dp2p0_streams = in_ip_params->max_num_dp2p0_streams;
+       out->max_num_dpp = in_ip_params->max_num_dpp;
+       out->max_num_hdmi_frl_outputs = in_ip_params->max_num_hdmi_frl_outputs;
+       out->max_num_otg = in_ip_params->max_num_otg;
+       out->max_num_wb = in_ip_params->max_num_wb;
+       out->max_pscl_lb_bw_pix_per_clk = in_ip_params->max_pscl_lb_bw_pix_per_clk;
+       out->max_vscl_hscl_bw_pix_per_clk = in_ip_params->max_vscl_hscl_bw_pix_per_clk;
+       out->max_vscl_ratio = in_ip_params->max_vscl_ratio;
+       out->max_vscl_taps = in_ip_params->max_vscl_taps;
+       out->meta_chunk_size_kbytes = in_ip_params->meta_chunk_size_kbytes;
+       out->meta_fifo_size_in_kentries = in_ip_params->meta_fifo_size_in_kentries;
+       out->min_meta_chunk_size_bytes = in_ip_params->min_meta_chunk_size_bytes;
+       out->min_pixel_chunk_size_bytes = in_ip_params->min_pixel_chunk_size_bytes;
+       out->num_dsc = in_ip_params->num_dsc;
+       out->pixel_chunk_size_kbytes = in_ip_params->pixel_chunk_size_kbytes;
+       out->ptoi_supported = in_ip_params->ptoi_supported;
+       out->rob_buffer_size_kbytes = in_ip_params->rob_buffer_size_kbytes;
+       out->writeback_chunk_size_kbytes = in_ip_params->writeback_chunk_size_kbytes;
+       out->writeback_interface_buffer_size_kbytes = in_ip_params->writeback_interface_buffer_size_kbytes;
+       out->writeback_line_buffer_buffer_size = in_ip_params->writeback_line_buffer_buffer_size;
+       out->writeback_max_hscl_ratio = in_ip_params->writeback_max_hscl_ratio;
+       out->writeback_max_hscl_taps = in_ip_params->writeback_max_hscl_taps;
+       out->writeback_max_vscl_ratio = in_ip_params->writeback_max_vscl_ratio;
+       out->writeback_max_vscl_taps = in_ip_params->writeback_max_vscl_taps;
+       out->writeback_min_hscl_ratio = in_ip_params->writeback_min_hscl_ratio;
+       out->writeback_min_vscl_ratio = in_ip_params->writeback_min_vscl_ratio;
+       out->zero_size_buffer_entries = in_ip_params->zero_size_buffer_entries;
+
+       /* As per hardcoded reference / discussions */
+       out->config_return_buffer_segment_size_in_kbytes = 64;
+       //out->vblank_nom_default_us = 600;
+       out->vblank_nom_default_us = in_ip_params->VBlankNomDefaultUS;
+}
+
+void dml2_translate_socbb_params(const struct dc *in, struct soc_bounding_box_st *out)
+{
+       const struct _vcs_dpi_soc_bounding_box_st *in_soc_params = &in->dml.soc;
+       /* Copy over the SOCBB params to dml2_ctx */
+       out->dispclk_dppclk_vco_speed_mhz = in_soc_params->dispclk_dppclk_vco_speed_mhz;
+       out->do_urgent_latency_adjustment = in_soc_params->do_urgent_latency_adjustment;
+       out->dram_channel_width_bytes = (dml_uint_t)in_soc_params->dram_channel_width_bytes;
+       out->fabric_datapath_to_dcn_data_return_bytes = (dml_uint_t)in_soc_params->fabric_datapath_to_dcn_data_return_bytes;
+       out->gpuvm_min_page_size_kbytes = in_soc_params->gpuvm_min_page_size_bytes * 1024;
+       out->hostvm_min_page_size_kbytes = in_soc_params->hostvm_min_page_size_bytes * 1024;
+       out->mall_allocated_for_dcn_mbytes = (dml_uint_t)in_soc_params->mall_allocated_for_dcn_mbytes;
+       out->max_avg_dram_bw_use_normal_percent = in_soc_params->max_avg_dram_bw_use_normal_percent;
+       out->max_avg_fabric_bw_use_normal_percent = in_soc_params->max_avg_fabric_bw_use_normal_percent;
+       out->max_avg_dram_bw_use_normal_strobe_percent = in_soc_params->max_avg_dram_bw_use_normal_strobe_percent;
+       out->max_avg_sdp_bw_use_normal_percent = in_soc_params->max_avg_sdp_bw_use_normal_percent;
+       out->max_outstanding_reqs = in_soc_params->max_request_size_bytes;
+       out->num_chans = in_soc_params->num_chans;
+       out->pct_ideal_dram_bw_after_urgent_strobe = in_soc_params->pct_ideal_dram_bw_after_urgent_strobe;
+       out->pct_ideal_dram_bw_after_urgent_vm_only = in_soc_params->pct_ideal_dram_sdp_bw_after_urgent_vm_only;
+       out->pct_ideal_fabric_bw_after_urgent = in_soc_params->pct_ideal_fabric_bw_after_urgent;
+       out->pct_ideal_sdp_bw_after_urgent = in_soc_params->pct_ideal_sdp_bw_after_urgent;
+       out->phy_downspread_percent = in_soc_params->downspread_percent;
+       out->refclk_mhz = 50; // As per hardcoded reference.
+       out->return_bus_width_bytes = in_soc_params->return_bus_width_bytes;
+       out->round_trip_ping_latency_dcfclk_cycles = in_soc_params->round_trip_ping_latency_dcfclk_cycles;
+       out->smn_latency_us = in_soc_params->smn_latency_us;
+       out->urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = in_soc_params->urgent_out_of_order_return_per_channel_pixel_and_vm_bytes;
+       out->urgent_out_of_order_return_per_channel_pixel_only_bytes = in_soc_params->urgent_out_of_order_return_per_channel_pixel_only_bytes;
+       out->urgent_out_of_order_return_per_channel_vm_only_bytes = in_soc_params->urgent_out_of_order_return_per_channel_vm_only_bytes;
+       out->pct_ideal_dram_bw_after_urgent_pixel_and_vm = in_soc_params->pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm;
+       out->pct_ideal_dram_bw_after_urgent_pixel_only = in_soc_params->pct_ideal_dram_sdp_bw_after_urgent_pixel_only;
+       out->dcn_downspread_percent = in_soc_params->dcn_downspread_percent;
+}
+
+void dml2_translate_soc_states(const struct dc *dc, struct soc_states_st *out, int num_states)
+{
+       unsigned int i = 0;
+       out->num_states = num_states;
+
+       for (i = 0; i < out->num_states; i++) {
+               out->state_array[i].dcfclk_mhz = dc->dml.soc.clock_limits[i].dcfclk_mhz;
+               out->state_array[i].dispclk_mhz = dc->dml.soc.clock_limits[i].dispclk_mhz;
+               out->state_array[i].dppclk_mhz = dc->dml.soc.clock_limits[i].dppclk_mhz;
+               out->state_array[i].dram_speed_mts = dc->dml.soc.clock_limits[i].dram_speed_mts;
+               out->state_array[i].dtbclk_mhz = dc->dml.soc.clock_limits[i].dtbclk_mhz;
+               out->state_array[i].socclk_mhz = dc->dml.soc.clock_limits[i].socclk_mhz;
+               out->state_array[i].fabricclk_mhz = dc->dml.soc.clock_limits[i].fabricclk_mhz;
+               out->state_array[i].dscclk_mhz = dc->dml.soc.clock_limits[i].dscclk_mhz;
+               out->state_array[i].phyclk_d18_mhz = dc->dml.soc.clock_limits[i].phyclk_d18_mhz;
+               out->state_array[i].phyclk_d32_mhz = dc->dml.soc.clock_limits[i].phyclk_d32_mhz;
+               out->state_array[i].phyclk_mhz = dc->dml.soc.clock_limits[i].phyclk_mhz;
+               out->state_array[i].sr_enter_plus_exit_time_us = dc->dml.soc.sr_enter_plus_exit_time_us;
+               out->state_array[i].sr_exit_time_us = dc->dml.soc.sr_exit_time_us;
+               out->state_array[i].fclk_change_latency_us = dc->dml.soc.fclk_change_latency_us;
+               out->state_array[i].dram_clock_change_latency_us = dc->dml.soc.dram_clock_change_latency_us;
+               out->state_array[i].usr_retraining_latency_us = dc->dml.soc.usr_retraining_latency_us;
+               out->state_array[i].writeback_latency_us = dc->dml.soc.writeback_latency_us;
+               /* Driver initialized values for these are different than the spreadsheet. Use the
+                * spreadsheet ones for now. We need to decided which ones to use.
+                */
+               out->state_array[i].sr_exit_z8_time_us = dc->dml.soc.sr_exit_z8_time_us;
+               out->state_array[i].sr_enter_plus_exit_z8_time_us = dc->dml.soc.sr_enter_plus_exit_z8_time_us;
+               //out->state_array[i].sr_exit_z8_time_us = 5.20;
+               //out->state_array[i].sr_enter_plus_exit_z8_time_us = 9.60;
+               out->state_array[i].use_ideal_dram_bw_strobe = true;
+               out->state_array[i].urgent_latency_pixel_data_only_us = dc->dml.soc.urgent_latency_pixel_data_only_us;
+               out->state_array[i].urgent_latency_pixel_mixed_with_vm_data_us = dc->dml.soc.urgent_latency_pixel_mixed_with_vm_data_us;
+               out->state_array[i].urgent_latency_vm_data_only_us = dc->dml.soc.urgent_latency_vm_data_only_us;
+               out->state_array[i].urgent_latency_adjustment_fabric_clock_component_us = dc->dml.soc.urgent_latency_adjustment_fabric_clock_component_us;
+               out->state_array[i].urgent_latency_adjustment_fabric_clock_reference_mhz = dc->dml.soc.urgent_latency_adjustment_fabric_clock_reference_mhz;
+       }
+}
+
+static void populate_dml_timing_cfg_from_stream_state(struct dml_timing_cfg_st *out, unsigned int location, const struct dc_stream_state *in)
+{
+       dml_uint_t hblank_start, vblank_start;
+
+       out->HActive[location] = in->timing.h_addressable + in->timing.h_border_left + in->timing.h_border_right;
+       out->VActive[location] = in->timing.v_addressable + in->timing.v_border_bottom + in->timing.v_border_top;
+       out->RefreshRate[location] = ((in->timing.pix_clk_100hz * 100) / in->timing.h_total) / in->timing.v_total;
+       out->VFrontPorch[location] = in->timing.v_front_porch;
+       out->PixelClock[location] = in->timing.pix_clk_100hz / 10000.00;
+       out->HTotal[location] = in->timing.h_total;
+       out->VTotal[location] = in->timing.v_total;
+       out->Interlace[location] = in->timing.flags.INTERLACE;
+       hblank_start = in->timing.h_total - in->timing.h_front_porch;
+       out->HBlankEnd[location] = hblank_start
+                                       - in->timing.h_addressable
+                                       - in->timing.h_border_left
+                                       - in->timing.h_border_right;
+       vblank_start = in->timing.v_total - in->timing.v_front_porch;
+       out->VBlankEnd[location] = vblank_start
+                                       - in->timing.v_addressable
+                                       - in->timing.v_border_top
+                                       - in->timing.v_border_bottom;
+       out->DRRDisplay[location] = false;
+}
+
+static void populate_dml_output_cfg_from_stream_state(struct dml_output_cfg_st *out, unsigned int location,
+                               const struct dc_stream_state *in, const struct pipe_ctx *pipe)
+{
+       unsigned int output_bpc;
+
+       out->DSCEnable[location] = (enum dml_dsc_enable)in->timing.flags.DSC;
+       out->OutputLinkDPLanes[location] = 4; // As per code in dcn20_resource.c
+       out->DSCInputBitPerComponent[location] = 12; // As per code in dcn20_resource.c
+
+       switch (in->signal) {
+       case SIGNAL_TYPE_DISPLAY_PORT_MST:
+       case SIGNAL_TYPE_DISPLAY_PORT:
+               out->OutputEncoder[location] = dml_dp;
+               if (is_dp2p0_output_encoder(pipe))
+                       out->OutputEncoder[location] = dml_dp2p0;
+               break;
+               out->OutputEncoder[location] = dml_edp;
+       case SIGNAL_TYPE_EDP:
+               break;
+       case SIGNAL_TYPE_HDMI_TYPE_A:
+       case SIGNAL_TYPE_DVI_SINGLE_LINK:
+       case SIGNAL_TYPE_DVI_DUAL_LINK:
+               out->OutputEncoder[location] = dml_hdmi;
+               break;
+       default:
+               out->OutputEncoder[location] = dml_dp;
+       }
+
+       switch (in->timing.display_color_depth) {
+       case COLOR_DEPTH_666:
+               output_bpc = 6;
+               break;
+       case COLOR_DEPTH_888:
+               output_bpc = 8;
+               break;
+       case COLOR_DEPTH_101010:
+               output_bpc = 10;
+               break;
+       case COLOR_DEPTH_121212:
+               output_bpc = 12;
+               break;
+       case COLOR_DEPTH_141414:
+               output_bpc = 14;
+               break;
+       case COLOR_DEPTH_161616:
+               output_bpc = 16;
+               break;
+       case COLOR_DEPTH_999:
+               output_bpc = 9;
+               break;
+       case COLOR_DEPTH_111111:
+               output_bpc = 11;
+               break;
+       default:
+               output_bpc = 8;
+               break;
+       }
+
+       switch (in->timing.pixel_encoding) {
+       case PIXEL_ENCODING_RGB:
+       case PIXEL_ENCODING_YCBCR444:
+               out->OutputFormat[location] = dml_444;
+               out->OutputBpp[location] = (dml_float_t)output_bpc * 3;
+               break;
+       case PIXEL_ENCODING_YCBCR420:
+               out->OutputFormat[location] = dml_420;
+               out->OutputBpp[location] = (output_bpc * 3.0) / 2;
+               break;
+       case PIXEL_ENCODING_YCBCR422:
+               if (in->timing.flags.DSC && !in->timing.dsc_cfg.ycbcr422_simple)
+                       out->OutputFormat[location] = dml_n422;
+               else
+                       out->OutputFormat[location] = dml_s422;
+               out->OutputBpp[location] = (dml_float_t)output_bpc * 2;
+               break;
+       default:
+               out->OutputFormat[location] = dml_444;
+               out->OutputBpp[location] = (dml_float_t)output_bpc * 3;
+               break;
+       }
+
+       if (in->timing.flags.DSC) {
+               out->OutputBpp[location] = in->timing.dsc_cfg.bits_per_pixel / 16.0;
+       }
+
+       // This has been false throughout DCN32x development. If needed we can change this later on.
+       out->OutputMultistreamEn[location] = false;
+
+       switch (in->signal) {
+       case SIGNAL_TYPE_NONE:
+       case SIGNAL_TYPE_DVI_SINGLE_LINK:
+       case SIGNAL_TYPE_DVI_DUAL_LINK:
+       case SIGNAL_TYPE_HDMI_TYPE_A:
+       case SIGNAL_TYPE_LVDS:
+       case SIGNAL_TYPE_RGB:
+       case SIGNAL_TYPE_DISPLAY_PORT:
+       case SIGNAL_TYPE_DISPLAY_PORT_MST:
+       case SIGNAL_TYPE_EDP:
+       case SIGNAL_TYPE_VIRTUAL:
+       default:
+               out->OutputLinkDPRate[location] = dml_dp_rate_na;
+               break;
+       }
+
+       out->PixelClockBackEnd[location] = in->timing.pix_clk_100hz / 10000.00;
+
+       out->AudioSampleLayout[location] = in->audio_info.modes->sample_size;
+       out->AudioSampleRate[location] = in->audio_info.modes->max_bit_rate;
+
+       out->OutputDisabled[location] = true;
+}
+
+static void populate_dummy_dml_surface_cfg(struct dml_surface_cfg_st *out, unsigned int location, const struct dc_stream_state *in)
+{
+       out->SurfaceWidthY[location] = in->timing.h_addressable;
+       out->SurfaceHeightY[location] = in->timing.v_addressable;
+       out->SurfaceWidthC[location] = in->timing.h_addressable;
+       out->SurfaceHeightC[location] = in->timing.v_addressable;
+       out->PitchY[location] = ((out->SurfaceWidthY[location] + 127) / 128) * 128;
+       out->PitchC[location] = 0;
+       out->DCCEnable[location] = false;
+       out->DCCMetaPitchY[location] = 0;
+       out->DCCMetaPitchC[location] = 0;
+       out->DCCRateLuma[location] = 1.0;
+       out->DCCRateChroma[location] = 1.0;
+       out->DCCFractionOfZeroSizeRequestsLuma[location] = 0;
+       out->DCCFractionOfZeroSizeRequestsChroma[location] = 0;
+       out->SurfaceTiling[location] = dml_sw_64kb_r_x;
+       out->SourcePixelFormat[location] = dml_444_32;
+}
+
+static void populate_dml_surface_cfg_from_plane_state(enum dml_project_id dml2_project, struct dml_surface_cfg_st *out, unsigned int location, const struct dc_plane_state *in)
+{
+       out->PitchY[location] = in->plane_size.surface_pitch;
+       out->SurfaceHeightY[location] = in->plane_size.surface_size.height;
+       out->SurfaceWidthY[location] = in->plane_size.surface_size.width;
+       out->SurfaceHeightC[location] = in->plane_size.surface_size.height;
+       out->SurfaceWidthC[location] = in->plane_size.surface_size.width;
+       out->PitchC[location] = in->plane_size.chroma_pitch;
+       out->DCCEnable[location] = in->dcc.enable;
+       out->DCCMetaPitchY[location] = in->dcc.meta_pitch;
+       out->DCCMetaPitchC[location] = in->dcc.meta_pitch_c;
+       out->DCCRateLuma[location] = 1.0;
+       out->DCCRateChroma[location] = 1.0;
+       out->DCCFractionOfZeroSizeRequestsLuma[location] = in->dcc.independent_64b_blks;
+       out->DCCFractionOfZeroSizeRequestsChroma[location] = in->dcc.independent_64b_blks_c;
+
+       switch (dml2_project) {
+       default:
+               out->SurfaceTiling[location] = (enum dml_swizzle_mode)in->tiling_info.gfx9.swizzle;
+               break;
+       }
+
+       switch (in->format) {
+       case SURFACE_PIXEL_FORMAT_VIDEO_420_YCbCr:
+       case SURFACE_PIXEL_FORMAT_VIDEO_420_YCrCb:
+               out->SourcePixelFormat[location] = dml_420_8;
+               break;
+       case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCbCr:
+       case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCrCb:
+               out->SourcePixelFormat[location] = dml_420_10;
+               break;
+       case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616:
+       case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616F:
+       case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F:
+               out->SourcePixelFormat[location] = dml_444_64;
+               break;
+       case SURFACE_PIXEL_FORMAT_GRPH_ARGB1555:
+       case SURFACE_PIXEL_FORMAT_GRPH_RGB565:
+               out->SourcePixelFormat[location] = dml_444_16;
+               break;
+       case SURFACE_PIXEL_FORMAT_GRPH_PALETA_256_COLORS:
+               out->SourcePixelFormat[location] = dml_444_8;
+               break;
+       case SURFACE_PIXEL_FORMAT_GRPH_RGBE_ALPHA:
+               out->SourcePixelFormat[location] = dml_rgbe_alpha;
+               break;
+       default:
+               out->SourcePixelFormat[location] = dml_444_32;
+               break;
+       }
+}
+
+/*TODO no support for mpc combine, need rework - should calculate scaling params based on plane+stream*/
+static struct scaler_data get_scaler_data_for_plane(const struct dc_plane_state *in, const struct dc_state *context)
+{
+       int i;
+       struct scaler_data data = { 0 };
+
+       for (i = 0; i < MAX_PIPES; i++) {
+               const struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+               if (pipe->plane_state == in && !pipe->prev_odm_pipe) {
+                       const struct pipe_ctx *next_pipe = pipe->next_odm_pipe;
+
+                       data = context->res_ctx.pipe_ctx[i].plane_res.scl_data;
+                       while (next_pipe) {
+                               data.h_active += next_pipe->plane_res.scl_data.h_active;
+                               data.recout.width += next_pipe->plane_res.scl_data.recout.width;
+                               if (in->rotation == ROTATION_ANGLE_0 || in->rotation == ROTATION_ANGLE_180) {
+                                       data.viewport.width += next_pipe->plane_res.scl_data.viewport.width;
+                               } else {
+                                       data.viewport.height += next_pipe->plane_res.scl_data.viewport.height;
+                               }
+                               next_pipe = next_pipe->next_odm_pipe;
+                       }
+                       break;
+               }
+       }
+
+       ASSERT(i < MAX_PIPES);
+       return data;
+}
+
+static void populate_dummy_dml_plane_cfg(struct dml_plane_cfg_st *out, unsigned int location, const struct dc_stream_state *in)
+{
+       out->CursorBPP[location] = dml_cur_32bit;
+       out->CursorWidth[location] = 256;
+
+       out->GPUVMMinPageSizeKBytes[location] = 256;
+
+       out->ViewportWidth[location] = in->timing.h_addressable;
+       out->ViewportHeight[location] = in->timing.v_addressable;
+       out->ViewportStationary[location] = false;
+       out->ViewportWidthChroma[location] = 0;
+       out->ViewportHeightChroma[location] = 0;
+       out->ViewportXStart[location] = 0;
+       out->ViewportXStartC[location] = 0;
+       out->ViewportYStart[location] = 0;
+       out->ViewportYStartC[location] = 0;
+
+       out->ScalerEnabled[location] = false;
+       out->HRatio[location] = 1.0;
+       out->VRatio[location] = 1.0;
+       out->HRatioChroma[location] = 0;
+       out->VRatioChroma[location] = 0;
+       out->HTaps[location] = 1;
+       out->VTaps[location] = 1;
+       out->HTapsChroma[location] = 0;
+       out->VTapsChroma[location] = 0;
+       out->SourceScan[location] = dml_rotation_0;
+       out->ScalerRecoutWidth[location] = in->timing.h_addressable;
+
+       out->LBBitPerPixel[location] = 57;
+
+       out->DynamicMetadataEnable[location] = false;
+
+       out->NumberOfCursors[location] = 1;
+       out->UseMALLForStaticScreen[location] = dml_use_mall_static_screen_disable;
+       out->UseMALLForPStateChange[location] = dml_use_mall_pstate_change_disable;
+
+       out->DETSizeOverride[location] = 256;
+
+       out->ScalerEnabled[location] = false;
+}
+
+static void populate_dml_plane_cfg_from_plane_state(struct dml_plane_cfg_st *out, unsigned int location, const struct dc_plane_state *in, const struct dc_state *context)
+{
+       const struct scaler_data scaler_data = get_scaler_data_for_plane(in, context);
+
+       out->CursorBPP[location] = dml_cur_32bit;
+       out->CursorWidth[location] = 256;
+
+       out->GPUVMMinPageSizeKBytes[location] = 256;
+
+       out->ViewportWidth[location] = scaler_data.viewport.width;
+       out->ViewportHeight[location] = scaler_data.viewport.height;
+       out->ViewportWidthChroma[location] = scaler_data.viewport_c.width;
+       out->ViewportHeightChroma[location] = scaler_data.viewport_c.height;
+       out->ViewportXStart[location] = scaler_data.viewport.x;
+       out->ViewportYStart[location] = scaler_data.viewport.y;
+       out->ViewportXStartC[location] = scaler_data.viewport_c.x;
+       out->ViewportYStartC[location] = scaler_data.viewport_c.y;
+       out->ViewportStationary[location] = false;
+
+       out->ScalerEnabled[location] = scaler_data.ratios.horz.value != dc_fixpt_one.value ||
+                               scaler_data.ratios.horz_c.value != dc_fixpt_one.value ||
+                               scaler_data.ratios.vert.value != dc_fixpt_one.value ||
+                               scaler_data.ratios.vert_c.value != dc_fixpt_one.value;
+
+       /* Current driver code base uses LBBitPerPixel as 57. There is a discrepancy
+        * from the HW/DML teams about this value. Initialize LBBitPerPixel with the
+        * value current used in Navi3x .
+        */
+
+       out->LBBitPerPixel[location] = 57;
+
+       if (out->ScalerEnabled[location] == false) {
+               out->HRatio[location] = 1;
+               out->HRatioChroma[location] = 1;
+               out->VRatio[location] = 1;
+               out->VRatioChroma[location] = 1;
+       } else {
+               /* Follow the original dml_wrapper.c code direction to fix scaling issues */
+               out->HRatio[location] = (dml_float_t)scaler_data.ratios.horz.value / (1ULL << 32);
+               out->HRatioChroma[location] = (dml_float_t)scaler_data.ratios.horz_c.value / (1ULL << 32);
+               out->VRatio[location] = (dml_float_t)scaler_data.ratios.vert.value / (1ULL << 32);
+               out->VRatioChroma[location] = (dml_float_t)scaler_data.ratios.vert_c.value / (1ULL << 32);
+       }
+
+       if (!scaler_data.taps.h_taps) {
+               out->HTaps[location] = 1;
+               out->HTapsChroma[location] = 1;
+       } else {
+               out->HTaps[location] = scaler_data.taps.h_taps;
+               out->HTapsChroma[location] = scaler_data.taps.h_taps_c;
+       }
+       if (!scaler_data.taps.v_taps) {
+               out->VTaps[location] = 1;
+               out->VTapsChroma[location] = 1;
+       } else {
+               out->VTaps[location] = scaler_data.taps.v_taps;
+               out->VTapsChroma[location] = scaler_data.taps.v_taps_c;
+       }
+
+       out->SourceScan[location] = (enum dml_rotation_angle)in->rotation;
+       out->ScalerRecoutWidth[location] = in->dst_rect.width;
+
+       out->DynamicMetadataEnable[location] = false;
+       out->DynamicMetadataLinesBeforeActiveRequired[location] = 0;
+       out->DynamicMetadataTransmittedBytes[location] = 0;
+
+       out->NumberOfCursors[location] = 1;
+}
+
+static unsigned int map_stream_to_dml_display_cfg(const struct dml2_context *dml2,
+               const struct dc_stream_state *stream, const struct dml_display_cfg_st *dml_dispcfg)
+{
+       int i = 0;
+       int location = -1;
+
+       for (i = 0; i < __DML2_WRAPPER_MAX_STREAMS_PLANES__; i++) {
+               if (dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_stream_id_valid[i] && dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_stream_id[i] == stream->stream_id) {
+                       location = i;
+                       break;
+               }
+       }
+
+       return location;
+}
+
+static bool get_plane_id(const struct dc_state *context, const struct dc_plane_state *plane, unsigned int *plane_id)
+{
+       int i, j;
+
+       if (!plane_id)
+               return false;
+
+       for (i = 0; i < context->stream_count; i++) {
+               for (j = 0; j < context->stream_status[i].plane_count; j++) {
+                       if (context->stream_status[i].plane_states[j] == plane) {
+                               *plane_id = (i << 16) | j;
+                               return true;
+                       }
+               }
+       }
+
+       return false;
+}
+
+static unsigned int map_plane_to_dml_display_cfg(const struct dml2_context *dml2,
+               const struct dc_plane_state *plane, const struct dc_state *context, const struct dml_display_cfg_st *dml_dispcfg)
+{
+       unsigned int plane_id;
+       int i = 0;
+       int location = -1;
+
+       if (!get_plane_id(context, plane, &plane_id)) {
+               ASSERT(false);
+               return -1;
+       }
+
+       for (i = 0; i < __DML2_WRAPPER_MAX_STREAMS_PLANES__; i++) {
+               if (dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id_valid[i] && dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id[i] == plane_id) {
+                       location = i;
+                       break;
+               }
+       }
+
+       return location;
+}
+
+static void apply_legacy_svp_drr_settings(struct dml2_context *dml2, const struct dc_state *state, struct dml_display_cfg_st *dml_dispcfg)
+{
+       int i;
+
+       if (state->bw_ctx.bw.dcn.clk.fw_based_mclk_switching) {
+               ASSERT(state->stream_count == 1);
+               dml_dispcfg->timing.DRRDisplay[0] = true;
+       } else if (state->bw_ctx.bw.dcn.legacy_svp_drr_stream_index_valid) {
+
+               for (i = 0; i < dml_dispcfg->num_timings; i++) {
+                       if (dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_stream_id[i] == state->streams[state->bw_ctx.bw.dcn.legacy_svp_drr_stream_index]->stream_id)
+                               dml_dispcfg->timing.DRRDisplay[i] = true;
+               }
+       }
+}
+
+void map_dc_state_into_dml_display_cfg(struct dml2_context *dml2, const struct dc_state *context, struct dml_display_cfg_st *dml_dispcfg)
+{
+       int i = 0, j = 0;
+       int disp_cfg_stream_location, disp_cfg_plane_location;
+
+       for (i = 0; i < __DML2_WRAPPER_MAX_STREAMS_PLANES__; i++) {
+               dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_stream_id_valid[i] = false;
+               dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id_valid[i] = false;
+               dml2->v20.scratch.dml_to_dc_pipe_mapping.dml_pipe_idx_to_stream_id_valid[i] = false;
+               dml2->v20.scratch.dml_to_dc_pipe_mapping.dml_pipe_idx_to_plane_id_valid[i] = false;
+       }
+
+       //Generally these are set by referencing our latest BB/IP params in dcn32_resource.c file
+       dml_dispcfg->plane.GPUVMEnable = true;
+       dml_dispcfg->plane.GPUVMMaxPageTableLevels = 4;
+       dml_dispcfg->plane.HostVMEnable = false;
+
+       for (i = 0; i < context->stream_count; i++) {
+               disp_cfg_stream_location = map_stream_to_dml_display_cfg(dml2, context->streams[i], dml_dispcfg);
+
+               if (disp_cfg_stream_location < 0)
+                       disp_cfg_stream_location = dml_dispcfg->num_timings++;
+
+               ASSERT(disp_cfg_stream_location >= 0 && disp_cfg_stream_location <= __DML2_WRAPPER_MAX_STREAMS_PLANES__);
+
+               populate_dml_timing_cfg_from_stream_state(&dml_dispcfg->timing, disp_cfg_stream_location, context->streams[i]);
+               populate_dml_output_cfg_from_stream_state(&dml_dispcfg->output, disp_cfg_stream_location, context->streams[i], &context->res_ctx.pipe_ctx[i]);
+               switch (context->streams[i]->debug.force_odm_combine_segments) {
+               case 2:
+                       dml2->v20.dml_core_ctx.policy.ODMUse[disp_cfg_stream_location] = dml_odm_use_policy_combine_2to1;
+                       break;
+               case 4:
+                       dml2->v20.dml_core_ctx.policy.ODMUse[disp_cfg_stream_location] = dml_odm_use_policy_combine_4to1;
+                       break;
+               default:
+                       break;
+               }
+
+               dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_stream_id[disp_cfg_stream_location] = context->streams[i]->stream_id;
+               dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_stream_id_valid[disp_cfg_stream_location] = true;
+
+               if (context->stream_status[i].plane_count == 0) {
+                       disp_cfg_plane_location = dml_dispcfg->num_surfaces++;
+
+                       populate_dummy_dml_surface_cfg(&dml_dispcfg->surface, disp_cfg_plane_location, context->streams[i]);
+                       populate_dummy_dml_plane_cfg(&dml_dispcfg->plane, disp_cfg_plane_location, context->streams[i]);
+
+                       dml_dispcfg->plane.BlendingAndTiming[disp_cfg_plane_location] = disp_cfg_stream_location;
+
+                       dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id_valid[disp_cfg_plane_location] = true;
+               } else {
+                       for (j = 0; j < context->stream_status[i].plane_count; j++) {
+                               disp_cfg_plane_location = map_plane_to_dml_display_cfg(dml2, context->stream_status[i].plane_states[j], context, dml_dispcfg);
+
+                               if (disp_cfg_plane_location < 0)
+                                       disp_cfg_plane_location = dml_dispcfg->num_surfaces++;
+
+                               ASSERT(disp_cfg_plane_location >= 0 && disp_cfg_plane_location <= __DML2_WRAPPER_MAX_STREAMS_PLANES__);
+
+                               populate_dml_surface_cfg_from_plane_state(dml2->v20.dml_core_ctx.project, &dml_dispcfg->surface, disp_cfg_plane_location, context->stream_status[i].plane_states[j]);
+                               populate_dml_plane_cfg_from_plane_state(&dml_dispcfg->plane, disp_cfg_plane_location, context->stream_status[i].plane_states[j], context);
+
+                               if (context->streams[i]->mall_stream_config.type == SUBVP_MAIN) {
+                                       dml_dispcfg->plane.UseMALLForPStateChange[disp_cfg_plane_location] = dml_use_mall_pstate_change_sub_viewport;
+                                       dml_dispcfg->plane.UseMALLForStaticScreen[disp_cfg_plane_location] = dml_use_mall_static_screen_optimize;
+                               } else if (context->streams[i]->mall_stream_config.type == SUBVP_PHANTOM) {
+                                       dml_dispcfg->plane.UseMALLForPStateChange[disp_cfg_plane_location] = dml_use_mall_pstate_change_phantom_pipe;
+                                       dml_dispcfg->plane.UseMALLForStaticScreen[disp_cfg_plane_location] = dml_use_mall_static_screen_disable;
+                                       dml2->v20.dml_core_ctx.policy.ImmediateFlipRequirement[disp_cfg_plane_location] = dml_immediate_flip_not_required;
+                               } else {
+                                       dml_dispcfg->plane.UseMALLForPStateChange[disp_cfg_plane_location] = dml_use_mall_pstate_change_disable;
+                                       dml_dispcfg->plane.UseMALLForStaticScreen[disp_cfg_plane_location] = dml_use_mall_static_screen_optimize;
+                               }
+
+                               dml_dispcfg->plane.BlendingAndTiming[disp_cfg_plane_location] = disp_cfg_stream_location;
+
+                               if (get_plane_id(context, context->stream_status[i].plane_states[j], &dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id[disp_cfg_plane_location]))
+                                       dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id_valid[disp_cfg_plane_location] = true;
+
+                               if (j >= 1) {
+                                       populate_dml_timing_cfg_from_stream_state(&dml_dispcfg->timing, disp_cfg_plane_location, context->streams[i]);
+                                       populate_dml_output_cfg_from_stream_state(&dml_dispcfg->output, disp_cfg_plane_location, context->streams[i], &context->res_ctx.pipe_ctx[i]);
+                                       switch (context->streams[i]->debug.force_odm_combine_segments) {
+                                       case 2:
+                                               dml2->v20.dml_core_ctx.policy.ODMUse[disp_cfg_plane_location] = dml_odm_use_policy_combine_2to1;
+                                               break;
+                                       case 4:
+                                               dml2->v20.dml_core_ctx.policy.ODMUse[disp_cfg_plane_location] = dml_odm_use_policy_combine_4to1;
+                                               break;
+                                       default:
+                                               break;
+                                       }
+
+                                       if (context->streams[i]->mall_stream_config.type == SUBVP_MAIN)
+                                               dml_dispcfg->plane.UseMALLForPStateChange[disp_cfg_plane_location] = dml_use_mall_pstate_change_sub_viewport;
+                                       else if (context->streams[i]->mall_stream_config.type == SUBVP_PHANTOM)
+                                               dml_dispcfg->plane.UseMALLForPStateChange[disp_cfg_plane_location] = dml_use_mall_pstate_change_phantom_pipe;
+
+                                       dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_stream_id[disp_cfg_plane_location] = context->streams[i]->stream_id;
+                                       dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_stream_id_valid[disp_cfg_plane_location] = true;
+
+                                       dml_dispcfg->num_timings++;
+                               }
+                       }
+               }
+       }
+
+       if (!dml2->config.use_native_pstate_optimization)
+               apply_legacy_svp_drr_settings(dml2, context, dml_dispcfg);
+}
+
+void dml2_update_pipe_ctx_dchub_regs(struct _vcs_dpi_dml_display_rq_regs_st *rq_regs,
+       struct _vcs_dpi_dml_display_dlg_regs_st *disp_dlg_regs,
+       struct _vcs_dpi_dml_display_ttu_regs_st *disp_ttu_regs,
+       struct pipe_ctx *out)
+{
+       memset(&out->rq_regs, 0, sizeof(out->rq_regs));
+       out->rq_regs.rq_regs_l.chunk_size = rq_regs->rq_regs_l.chunk_size;
+       out->rq_regs.rq_regs_l.min_chunk_size = rq_regs->rq_regs_l.min_chunk_size;
+       out->rq_regs.rq_regs_l.meta_chunk_size = rq_regs->rq_regs_l.meta_chunk_size;
+       out->rq_regs.rq_regs_l.min_meta_chunk_size = rq_regs->rq_regs_l.min_meta_chunk_size;
+       out->rq_regs.rq_regs_l.dpte_group_size = rq_regs->rq_regs_l.dpte_group_size;
+       out->rq_regs.rq_regs_l.mpte_group_size = rq_regs->rq_regs_l.mpte_group_size;
+       out->rq_regs.rq_regs_l.swath_height = rq_regs->rq_regs_l.swath_height;
+       out->rq_regs.rq_regs_l.pte_row_height_linear = rq_regs->rq_regs_l.pte_row_height_linear;
+
+       out->rq_regs.rq_regs_c.chunk_size = rq_regs->rq_regs_c.chunk_size;
+       out->rq_regs.rq_regs_c.min_chunk_size = rq_regs->rq_regs_c.min_chunk_size;
+       out->rq_regs.rq_regs_c.meta_chunk_size = rq_regs->rq_regs_c.meta_chunk_size;
+       out->rq_regs.rq_regs_c.min_meta_chunk_size = rq_regs->rq_regs_c.min_meta_chunk_size;
+       out->rq_regs.rq_regs_c.dpte_group_size = rq_regs->rq_regs_c.dpte_group_size;
+       out->rq_regs.rq_regs_c.mpte_group_size = rq_regs->rq_regs_c.mpte_group_size;
+       out->rq_regs.rq_regs_c.swath_height = rq_regs->rq_regs_c.swath_height;
+       out->rq_regs.rq_regs_c.pte_row_height_linear = rq_regs->rq_regs_c.pte_row_height_linear;
+
+       out->rq_regs.drq_expansion_mode = rq_regs->drq_expansion_mode;
+       out->rq_regs.prq_expansion_mode = rq_regs->prq_expansion_mode;
+       out->rq_regs.mrq_expansion_mode = rq_regs->mrq_expansion_mode;
+       out->rq_regs.crq_expansion_mode = rq_regs->crq_expansion_mode;
+       out->rq_regs.plane1_base_address = rq_regs->plane1_base_address;
+
+       memset(&out->dlg_regs, 0, sizeof(out->dlg_regs));
+       out->dlg_regs.refcyc_h_blank_end = disp_dlg_regs->refcyc_h_blank_end;
+       out->dlg_regs.dlg_vblank_end = disp_dlg_regs->dlg_vblank_end;
+       out->dlg_regs.min_dst_y_next_start = disp_dlg_regs->min_dst_y_next_start;
+       out->dlg_regs.refcyc_per_htotal = disp_dlg_regs->refcyc_per_htotal;
+       out->dlg_regs.refcyc_x_after_scaler = disp_dlg_regs->refcyc_x_after_scaler;
+       out->dlg_regs.dst_y_after_scaler = disp_dlg_regs->dst_y_after_scaler;
+       out->dlg_regs.dst_y_prefetch = disp_dlg_regs->dst_y_prefetch;
+       out->dlg_regs.dst_y_per_vm_vblank = disp_dlg_regs->dst_y_per_vm_vblank;
+       out->dlg_regs.dst_y_per_row_vblank = disp_dlg_regs->dst_y_per_row_vblank;
+       out->dlg_regs.dst_y_per_vm_flip = disp_dlg_regs->dst_y_per_vm_flip;
+       out->dlg_regs.dst_y_per_row_flip = disp_dlg_regs->dst_y_per_row_flip;
+       out->dlg_regs.ref_freq_to_pix_freq = disp_dlg_regs->ref_freq_to_pix_freq;
+       out->dlg_regs.vratio_prefetch = disp_dlg_regs->vratio_prefetch;
+       out->dlg_regs.vratio_prefetch_c = disp_dlg_regs->vratio_prefetch_c;
+       out->dlg_regs.refcyc_per_pte_group_vblank_l = disp_dlg_regs->refcyc_per_pte_group_vblank_l;
+       out->dlg_regs.refcyc_per_pte_group_vblank_c = disp_dlg_regs->refcyc_per_pte_group_vblank_c;
+       out->dlg_regs.refcyc_per_meta_chunk_vblank_l = disp_dlg_regs->refcyc_per_meta_chunk_vblank_l;
+       out->dlg_regs.refcyc_per_meta_chunk_vblank_c = disp_dlg_regs->refcyc_per_meta_chunk_vblank_c;
+       out->dlg_regs.refcyc_per_pte_group_flip_l = disp_dlg_regs->refcyc_per_pte_group_flip_l;
+       out->dlg_regs.refcyc_per_pte_group_flip_c = disp_dlg_regs->refcyc_per_pte_group_flip_c;
+       out->dlg_regs.refcyc_per_meta_chunk_flip_l = disp_dlg_regs->refcyc_per_meta_chunk_flip_l;
+       out->dlg_regs.refcyc_per_meta_chunk_flip_c = disp_dlg_regs->refcyc_per_meta_chunk_flip_c;
+       out->dlg_regs.dst_y_per_pte_row_nom_l = disp_dlg_regs->dst_y_per_pte_row_nom_l;
+       out->dlg_regs.dst_y_per_pte_row_nom_c = disp_dlg_regs->dst_y_per_pte_row_nom_c;
+       out->dlg_regs.refcyc_per_pte_group_nom_l = disp_dlg_regs->refcyc_per_pte_group_nom_l;
+       out->dlg_regs.refcyc_per_pte_group_nom_c = disp_dlg_regs->refcyc_per_pte_group_nom_c;
+       out->dlg_regs.dst_y_per_meta_row_nom_l = disp_dlg_regs->dst_y_per_meta_row_nom_l;
+       out->dlg_regs.dst_y_per_meta_row_nom_c = disp_dlg_regs->dst_y_per_meta_row_nom_c;
+       out->dlg_regs.refcyc_per_meta_chunk_nom_l = disp_dlg_regs->refcyc_per_meta_chunk_nom_l;
+       out->dlg_regs.refcyc_per_meta_chunk_nom_c = disp_dlg_regs->refcyc_per_meta_chunk_nom_c;
+       out->dlg_regs.refcyc_per_line_delivery_pre_l = disp_dlg_regs->refcyc_per_line_delivery_pre_l;
+       out->dlg_regs.refcyc_per_line_delivery_pre_c = disp_dlg_regs->refcyc_per_line_delivery_pre_c;
+       out->dlg_regs.refcyc_per_line_delivery_l = disp_dlg_regs->refcyc_per_line_delivery_l;
+       out->dlg_regs.refcyc_per_line_delivery_c = disp_dlg_regs->refcyc_per_line_delivery_c;
+       out->dlg_regs.refcyc_per_vm_group_vblank = disp_dlg_regs->refcyc_per_vm_group_vblank;
+       out->dlg_regs.refcyc_per_vm_group_flip = disp_dlg_regs->refcyc_per_vm_group_flip;
+       out->dlg_regs.refcyc_per_vm_req_vblank = disp_dlg_regs->refcyc_per_vm_req_vblank;
+       out->dlg_regs.refcyc_per_vm_req_flip = disp_dlg_regs->refcyc_per_vm_req_flip;
+       out->dlg_regs.dst_y_offset_cur0 = disp_dlg_regs->dst_y_offset_cur0;
+       out->dlg_regs.chunk_hdl_adjust_cur0 = disp_dlg_regs->chunk_hdl_adjust_cur0;
+       out->dlg_regs.dst_y_offset_cur1 = disp_dlg_regs->dst_y_offset_cur1;
+       out->dlg_regs.chunk_hdl_adjust_cur1 = disp_dlg_regs->chunk_hdl_adjust_cur1;
+       out->dlg_regs.vready_after_vcount0 = disp_dlg_regs->vready_after_vcount0;
+       out->dlg_regs.dst_y_delta_drq_limit = disp_dlg_regs->dst_y_delta_drq_limit;
+       out->dlg_regs.refcyc_per_vm_dmdata = disp_dlg_regs->refcyc_per_vm_dmdata;
+       out->dlg_regs.dmdata_dl_delta = disp_dlg_regs->dmdata_dl_delta;
+
+       memset(&out->ttu_regs, 0, sizeof(out->ttu_regs));
+       out->ttu_regs.qos_level_low_wm = disp_ttu_regs->qos_level_low_wm;
+       out->ttu_regs.qos_level_high_wm = disp_ttu_regs->qos_level_high_wm;
+       out->ttu_regs.min_ttu_vblank = disp_ttu_regs->min_ttu_vblank;
+       out->ttu_regs.qos_level_flip = disp_ttu_regs->qos_level_flip;
+       out->ttu_regs.refcyc_per_req_delivery_l = disp_ttu_regs->refcyc_per_req_delivery_l;
+       out->ttu_regs.refcyc_per_req_delivery_c = disp_ttu_regs->refcyc_per_req_delivery_c;
+       out->ttu_regs.refcyc_per_req_delivery_cur0 = disp_ttu_regs->refcyc_per_req_delivery_cur0;
+       out->ttu_regs.refcyc_per_req_delivery_cur1 = disp_ttu_regs->refcyc_per_req_delivery_cur1;
+       out->ttu_regs.refcyc_per_req_delivery_pre_l = disp_ttu_regs->refcyc_per_req_delivery_pre_l;
+       out->ttu_regs.refcyc_per_req_delivery_pre_c = disp_ttu_regs->refcyc_per_req_delivery_pre_c;
+       out->ttu_regs.refcyc_per_req_delivery_pre_cur0 = disp_ttu_regs->refcyc_per_req_delivery_pre_cur0;
+       out->ttu_regs.refcyc_per_req_delivery_pre_cur1 = disp_ttu_regs->refcyc_per_req_delivery_pre_cur1;
+       out->ttu_regs.qos_level_fixed_l = disp_ttu_regs->qos_level_fixed_l;
+       out->ttu_regs.qos_level_fixed_c = disp_ttu_regs->qos_level_fixed_c;
+       out->ttu_regs.qos_level_fixed_cur0 = disp_ttu_regs->qos_level_fixed_cur0;
+       out->ttu_regs.qos_level_fixed_cur1 = disp_ttu_regs->qos_level_fixed_cur1;
+       out->ttu_regs.qos_ramp_disable_l = disp_ttu_regs->qos_ramp_disable_l;
+       out->ttu_regs.qos_ramp_disable_c = disp_ttu_regs->qos_ramp_disable_c;
+       out->ttu_regs.qos_ramp_disable_cur0 = disp_ttu_regs->qos_ramp_disable_cur0;
+       out->ttu_regs.qos_ramp_disable_cur1 = disp_ttu_regs->qos_ramp_disable_cur1;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.h b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.h
new file mode 100644 (file)
index 0000000..1bcfca5
--- /dev/null
@@ -0,0 +1,39 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DML2_TRANSLATION_HELPER_H__
+#define __DML2_TRANSLATION_HELPER_H__
+
+void dml2_init_ip_params(struct dml2_context *dml2, const struct dc *in_dc, struct ip_params_st *out);
+void dml2_init_socbb_params(struct dml2_context *dml2, const struct dc *in_dc, struct soc_bounding_box_st *out);
+void dml2_init_soc_states(struct dml2_context *dml2, const struct dc *in_dc,
+               const struct soc_bounding_box_st *in_bbox, struct soc_states_st *out);
+void dml2_translate_ip_params(const struct dc *in_dc, struct ip_params_st *out);
+void dml2_translate_socbb_params(const struct dc *in_dc, struct soc_bounding_box_st *out);
+void dml2_translate_soc_states(const struct dc *in_dc, struct soc_states_st *out, int num_states);
+void map_dc_state_into_dml_display_cfg(struct dml2_context *dml2, const struct dc_state *context, struct dml_display_cfg_st *dml_dispcfg);
+void dml2_update_pipe_ctx_dchub_regs(struct _vcs_dpi_dml_display_rq_regs_st *rq_regs, struct _vcs_dpi_dml_display_dlg_regs_st *disp_dlg_regs, struct _vcs_dpi_dml_display_ttu_regs_st *disp_ttu_regs, struct pipe_ctx *out);
+bool is_dp2p0_output_encoder(const struct pipe_ctx *pipe);
+
+#endif //__DML2_TRANSLATION_HELPER_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_utils.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_utils.c
new file mode 100644 (file)
index 0000000..5bd6956
--- /dev/null
@@ -0,0 +1,452 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+//#include "dml2_utils.h"
+#include "display_mode_core.h"
+#include "dml_display_rq_dlg_calc.h"
+#include "dml2_internal_types.h"
+#include "dml2_translation_helper.h"
+#include "dml2_utils.h"
+
+void dml2_util_copy_dml_timing(struct dml_timing_cfg_st *dml_timing_array, unsigned int dst_index, unsigned int src_index)
+{
+       dml_timing_array->HTotal[dst_index] = dml_timing_array->HTotal[src_index];
+       dml_timing_array->VTotal[dst_index] = dml_timing_array->VTotal[src_index];
+       dml_timing_array->HBlankEnd[dst_index] = dml_timing_array->HBlankEnd[src_index];
+       dml_timing_array->VBlankEnd[dst_index] = dml_timing_array->VBlankEnd[src_index];
+       dml_timing_array->RefreshRate[dst_index] = dml_timing_array->RefreshRate[src_index];
+       dml_timing_array->VFrontPorch[dst_index] = dml_timing_array->VFrontPorch[src_index];
+       dml_timing_array->PixelClock[dst_index] = dml_timing_array->PixelClock[src_index];
+       dml_timing_array->HActive[dst_index] = dml_timing_array->HActive[src_index];
+       dml_timing_array->VActive[dst_index] = dml_timing_array->VActive[src_index];
+       dml_timing_array->Interlace[dst_index] = dml_timing_array->Interlace[src_index];
+       dml_timing_array->DRRDisplay[dst_index] = dml_timing_array->DRRDisplay[src_index];
+       dml_timing_array->VBlankNom[dst_index] = dml_timing_array->VBlankNom[src_index];
+}
+
+void dml2_util_copy_dml_plane(struct dml_plane_cfg_st *dml_plane_array, unsigned int dst_index, unsigned int src_index)
+{
+       dml_plane_array->GPUVMMinPageSizeKBytes[dst_index] = dml_plane_array->GPUVMMinPageSizeKBytes[src_index];
+       dml_plane_array->ForceOneRowForFrame[dst_index] = dml_plane_array->ForceOneRowForFrame[src_index];
+       dml_plane_array->PTEBufferModeOverrideEn[dst_index] = dml_plane_array->PTEBufferModeOverrideEn[src_index];
+       dml_plane_array->PTEBufferMode[dst_index] = dml_plane_array->PTEBufferMode[src_index];
+       dml_plane_array->ViewportWidth[dst_index] = dml_plane_array->ViewportWidth[src_index];
+       dml_plane_array->ViewportHeight[dst_index] = dml_plane_array->ViewportHeight[src_index];
+       dml_plane_array->ViewportWidthChroma[dst_index] = dml_plane_array->ViewportWidthChroma[src_index];
+       dml_plane_array->ViewportHeightChroma[dst_index] = dml_plane_array->ViewportHeightChroma[src_index];
+       dml_plane_array->ViewportXStart[dst_index] = dml_plane_array->ViewportXStart[src_index];
+       dml_plane_array->ViewportXStartC[dst_index] = dml_plane_array->ViewportXStartC[src_index];
+       dml_plane_array->ViewportYStart[dst_index] = dml_plane_array->ViewportYStart[src_index];
+       dml_plane_array->ViewportYStartC[dst_index] = dml_plane_array->ViewportYStartC[src_index];
+       dml_plane_array->ViewportStationary[dst_index] = dml_plane_array->ViewportStationary[src_index];
+
+       dml_plane_array->ScalerEnabled[dst_index] = dml_plane_array->ScalerEnabled[src_index];
+       dml_plane_array->HRatio[dst_index] = dml_plane_array->HRatio[src_index];
+       dml_plane_array->VRatio[dst_index] = dml_plane_array->VRatio[src_index];
+       dml_plane_array->HRatioChroma[dst_index] = dml_plane_array->HRatioChroma[src_index];
+       dml_plane_array->VRatioChroma[dst_index] = dml_plane_array->VRatioChroma[src_index];
+       dml_plane_array->HTaps[dst_index] = dml_plane_array->HTaps[src_index];
+       dml_plane_array->VTaps[dst_index] = dml_plane_array->VTaps[src_index];
+       dml_plane_array->HTapsChroma[dst_index] = dml_plane_array->HTapsChroma[src_index];
+       dml_plane_array->VTapsChroma[dst_index] = dml_plane_array->VTapsChroma[src_index];
+       dml_plane_array->LBBitPerPixel[dst_index] = dml_plane_array->LBBitPerPixel[src_index];
+
+       dml_plane_array->SourceScan[dst_index] = dml_plane_array->SourceScan[src_index];
+       dml_plane_array->ScalerRecoutWidth[dst_index] = dml_plane_array->ScalerRecoutWidth[src_index];
+
+       dml_plane_array->DynamicMetadataEnable[dst_index] = dml_plane_array->DynamicMetadataEnable[src_index];
+       dml_plane_array->DynamicMetadataLinesBeforeActiveRequired[dst_index] = dml_plane_array->DynamicMetadataLinesBeforeActiveRequired[src_index];
+       dml_plane_array->DynamicMetadataTransmittedBytes[dst_index] = dml_plane_array->DynamicMetadataTransmittedBytes[src_index];
+       dml_plane_array->DETSizeOverride[dst_index] = dml_plane_array->DETSizeOverride[src_index];
+
+       dml_plane_array->NumberOfCursors[dst_index] = dml_plane_array->NumberOfCursors[src_index];
+       dml_plane_array->CursorWidth[dst_index] = dml_plane_array->CursorWidth[src_index];
+       dml_plane_array->CursorBPP[dst_index] = dml_plane_array->CursorBPP[src_index];
+
+       dml_plane_array->UseMALLForStaticScreen[dst_index] = dml_plane_array->UseMALLForStaticScreen[src_index];
+       dml_plane_array->UseMALLForPStateChange[dst_index] = dml_plane_array->UseMALLForPStateChange[src_index];
+
+       dml_plane_array->BlendingAndTiming[dst_index] = dml_plane_array->BlendingAndTiming[src_index];
+}
+
+void dml2_util_copy_dml_surface(struct dml_surface_cfg_st *dml_surface_array, unsigned int dst_index, unsigned int src_index)
+{
+       dml_surface_array->SurfaceTiling[dst_index] = dml_surface_array->SurfaceTiling[src_index];
+       dml_surface_array->SourcePixelFormat[dst_index] = dml_surface_array->SourcePixelFormat[src_index];
+       dml_surface_array->PitchY[dst_index] = dml_surface_array->PitchY[src_index];
+       dml_surface_array->SurfaceWidthY[dst_index] = dml_surface_array->SurfaceWidthY[src_index];
+       dml_surface_array->SurfaceHeightY[dst_index] = dml_surface_array->SurfaceHeightY[src_index];
+       dml_surface_array->PitchC[dst_index] = dml_surface_array->PitchC[src_index];
+       dml_surface_array->SurfaceWidthC[dst_index] = dml_surface_array->SurfaceWidthC[src_index];
+       dml_surface_array->SurfaceHeightC[dst_index] = dml_surface_array->SurfaceHeightC[src_index];
+
+       dml_surface_array->DCCEnable[dst_index] = dml_surface_array->DCCEnable[src_index];
+       dml_surface_array->DCCMetaPitchY[dst_index] = dml_surface_array->DCCMetaPitchY[src_index];
+       dml_surface_array->DCCMetaPitchC[dst_index] = dml_surface_array->DCCMetaPitchC[src_index];
+
+       dml_surface_array->DCCRateLuma[dst_index] = dml_surface_array->DCCRateLuma[src_index];
+       dml_surface_array->DCCRateChroma[dst_index] = dml_surface_array->DCCRateChroma[src_index];
+       dml_surface_array->DCCFractionOfZeroSizeRequestsLuma[dst_index] = dml_surface_array->DCCFractionOfZeroSizeRequestsLuma[src_index];
+       dml_surface_array->DCCFractionOfZeroSizeRequestsChroma[dst_index] = dml_surface_array->DCCFractionOfZeroSizeRequestsChroma[src_index];
+}
+
+void dml2_util_copy_dml_output(struct dml_output_cfg_st *dml_output_array, unsigned int dst_index, unsigned int src_index)
+{
+       dml_output_array->DSCInputBitPerComponent[dst_index] = dml_output_array->DSCInputBitPerComponent[src_index];
+       dml_output_array->OutputFormat[dst_index] = dml_output_array->OutputFormat[src_index];
+       dml_output_array->OutputEncoder[dst_index] = dml_output_array->OutputEncoder[src_index];
+       dml_output_array->OutputMultistreamId[dst_index] = dml_output_array->OutputMultistreamId[src_index];
+       dml_output_array->OutputMultistreamEn[dst_index] = dml_output_array->OutputMultistreamEn[src_index];
+       dml_output_array->OutputBpp[dst_index] = dml_output_array->OutputBpp[src_index];
+       dml_output_array->PixelClockBackEnd[dst_index] = dml_output_array->PixelClockBackEnd[src_index];
+       dml_output_array->DSCEnable[dst_index] = dml_output_array->DSCEnable[src_index];
+       dml_output_array->OutputLinkDPLanes[dst_index] = dml_output_array->OutputLinkDPLanes[src_index];
+       dml_output_array->OutputLinkDPRate[dst_index] = dml_output_array->OutputLinkDPRate[src_index];
+       dml_output_array->ForcedOutputLinkBPP[dst_index] = dml_output_array->ForcedOutputLinkBPP[src_index];
+       dml_output_array->AudioSampleRate[dst_index] = dml_output_array->AudioSampleRate[src_index];
+       dml_output_array->AudioSampleLayout[dst_index] = dml_output_array->AudioSampleLayout[src_index];
+}
+
+unsigned int dml2_util_get_maximum_odm_combine_for_output(bool force_odm_4to1, enum dml_output_encoder_class encoder, bool dsc_enabled)
+{
+       switch (encoder) {
+       case dml_dp:
+       case dml_edp:
+               return 2;
+       case dml_dp2p0:
+       if (dsc_enabled || force_odm_4to1)
+               return 4;
+       else
+               return 2;
+       case dml_hdmi:
+               return 1;
+       case dml_hdmifrl:
+       if (force_odm_4to1)
+               return 4;
+       else
+               return 2;
+       default:
+               return 1;
+       }
+}
+
+bool is_dp2p0_output_encoder(const struct pipe_ctx *pipe_ctx)
+{
+       /* If this assert is hit then we have a link encoder dynamic management issue */
+       ASSERT(pipe_ctx->stream_res.hpo_dp_stream_enc ? pipe_ctx->link_res.hpo_dp_link_enc != NULL : true);
+       return (pipe_ctx->stream_res.hpo_dp_stream_enc &&
+               pipe_ctx->link_res.hpo_dp_link_enc &&
+               dc_is_dp_signal(pipe_ctx->stream->signal));
+}
+
+bool is_dtbclk_required(const struct dc *dc, struct dc_state *context)
+{
+       int i;
+
+       for (i = 0; i < dc->res_pool->pipe_count; i++) {
+               if (!context->res_ctx.pipe_ctx[i].stream)
+                       continue;
+               if (is_dp2p0_output_encoder(&context->res_ctx.pipe_ctx[i]))
+                       return true;
+       }
+       return false;
+}
+
+void dml2_copy_clocks_to_dc_state(struct dml2_dcn_clocks *out_clks, struct dc_state *context)
+{
+       context->bw_ctx.bw.dcn.clk.dispclk_khz = out_clks->dispclk_khz;
+       context->bw_ctx.bw.dcn.clk.dcfclk_khz = out_clks->dcfclk_khz;
+       context->bw_ctx.bw.dcn.clk.dramclk_khz = out_clks->uclk_mts / 16;
+       context->bw_ctx.bw.dcn.clk.fclk_khz = out_clks->fclk_khz;
+       context->bw_ctx.bw.dcn.clk.phyclk_khz = out_clks->phyclk_khz;
+       context->bw_ctx.bw.dcn.clk.socclk_khz = out_clks->socclk_khz;
+       context->bw_ctx.bw.dcn.clk.ref_dtbclk_khz = out_clks->ref_dtbclk_khz;
+       context->bw_ctx.bw.dcn.clk.p_state_change_support = out_clks->p_state_supported;
+}
+
+int dml2_helper_find_dml_pipe_idx_by_stream_id(struct dml2_context *ctx, unsigned int stream_id)
+{
+       int i;
+       for (i = 0; i < __DML2_WRAPPER_MAX_STREAMS_PLANES__; i++) {
+               if (ctx->v20.scratch.dml_to_dc_pipe_mapping.dml_pipe_idx_to_stream_id_valid[i] && ctx->v20.scratch.dml_to_dc_pipe_mapping.dml_pipe_idx_to_stream_id[i] == stream_id)
+                       return  i;
+       }
+
+       return -1;
+}
+
+static int find_dml_pipe_idx_by_plane_id(struct dml2_context *ctx, unsigned int plane_id)
+{
+       int i;
+       for (i = 0; i < __DML2_WRAPPER_MAX_STREAMS_PLANES__; i++) {
+               if (ctx->v20.scratch.dml_to_dc_pipe_mapping.dml_pipe_idx_to_plane_id_valid[i] && ctx->v20.scratch.dml_to_dc_pipe_mapping.dml_pipe_idx_to_plane_id[i] == plane_id)
+                       return  i;
+       }
+
+       return -1;
+}
+
+static bool get_plane_id(const struct dc_state *state, const struct dc_plane_state *plane, unsigned int *plane_id)
+{
+       int i, j;
+
+       if (!plane_id)
+               return false;
+
+       for (i = 0; i < state->stream_count; i++) {
+               for (j = 0; j < state->stream_status[i].plane_count; j++) {
+                       if (state->stream_status[i].plane_states[j] == plane) {
+                               *plane_id = (i << 16) | j;
+                               return true;
+                       }
+               }
+       }
+
+       return false;
+}
+
+static void populate_pipe_ctx_dlg_params_from_dml(struct pipe_ctx *pipe_ctx, struct display_mode_lib_st *mode_lib, dml_uint_t pipe_idx)
+{
+       unsigned int hactive, vactive, hblank_start, vblank_start, hblank_end, vblank_end;
+       struct dc_crtc_timing *timing = &pipe_ctx->stream->timing;
+
+       hactive = timing->h_addressable + timing->h_border_left + timing->h_border_right;
+       vactive = timing->v_addressable + timing->v_border_bottom + timing->v_border_top;
+       hblank_start = pipe_ctx->stream->timing.h_total - pipe_ctx->stream->timing.h_front_porch;
+       vblank_start = pipe_ctx->stream->timing.v_total - pipe_ctx->stream->timing.v_front_porch;
+
+       hblank_end = hblank_start - timing->h_addressable - timing->h_border_left - timing->h_border_right;
+       vblank_end = vblank_start - timing->v_addressable - timing->v_border_top - timing->v_border_bottom;
+
+       pipe_ctx->pipe_dlg_param.vstartup_start = dml_get_vstartup_calculated(mode_lib, pipe_idx);
+       pipe_ctx->pipe_dlg_param.vupdate_offset = dml_get_vupdate_offset(mode_lib, pipe_idx);
+       pipe_ctx->pipe_dlg_param.vupdate_width = dml_get_vupdate_width(mode_lib, pipe_idx);
+       pipe_ctx->pipe_dlg_param.vready_offset = dml_get_vready_offset(mode_lib, pipe_idx);
+
+       pipe_ctx->pipe_dlg_param.otg_inst = pipe_ctx->stream_res.tg->inst;
+
+       pipe_ctx->pipe_dlg_param.hactive = hactive;
+       pipe_ctx->pipe_dlg_param.vactive = vactive;
+       pipe_ctx->pipe_dlg_param.htotal = pipe_ctx->stream->timing.h_total;
+       pipe_ctx->pipe_dlg_param.vtotal = pipe_ctx->stream->timing.v_total;
+       pipe_ctx->pipe_dlg_param.hblank_end = hblank_end;
+       pipe_ctx->pipe_dlg_param.vblank_end = vblank_end;
+       pipe_ctx->pipe_dlg_param.hblank_start = hblank_start;
+       pipe_ctx->pipe_dlg_param.vblank_start = vblank_start;
+       pipe_ctx->pipe_dlg_param.vfront_porch = pipe_ctx->stream->timing.v_front_porch;
+       pipe_ctx->pipe_dlg_param.pixel_rate_mhz = pipe_ctx->stream->timing.pix_clk_100hz / 10000.00;
+       pipe_ctx->pipe_dlg_param.refresh_rate = ((timing->pix_clk_100hz * 100) / timing->h_total) / timing->v_total;
+       pipe_ctx->pipe_dlg_param.vtotal_max = pipe_ctx->stream->adjust.v_total_max;
+       pipe_ctx->pipe_dlg_param.vtotal_min = pipe_ctx->stream->adjust.v_total_min;
+       pipe_ctx->pipe_dlg_param.recout_height = pipe_ctx->plane_res.scl_data.recout.height;
+       pipe_ctx->pipe_dlg_param.recout_width = pipe_ctx->plane_res.scl_data.recout.width;
+       pipe_ctx->pipe_dlg_param.full_recout_height = pipe_ctx->plane_res.scl_data.recout.height;
+       pipe_ctx->pipe_dlg_param.full_recout_width = pipe_ctx->plane_res.scl_data.recout.width;
+}
+
+void dml2_calculate_rq_and_dlg_params(const struct dc *dc, struct dc_state *context, struct resource_context *out_new_hw_state, struct dml2_context *in_ctx, unsigned int pipe_cnt)
+{
+       unsigned int dc_pipe_ctx_index, dml_pipe_idx, plane_id;
+       bool unbounded_req_enabled = false;
+       struct dml2_calculate_rq_and_dlg_params_scratch *s = &in_ctx->v20.scratch.calculate_rq_and_dlg_params_scratch;
+
+       context->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz = (unsigned int)in_ctx->v20.dml_core_ctx.mp.DCFCLKDeepSleep * 1000;
+       context->bw_ctx.bw.dcn.clk.dppclk_khz = 0;
+
+       if (in_ctx->v20.dml_core_ctx.ms.support.FCLKChangeSupport[in_ctx->v20.scratch.mode_support_params.out_lowest_state_idx] == dml_fclock_change_unsupported)
+               context->bw_ctx.bw.dcn.clk.fclk_p_state_change_support = false;
+       else
+               context->bw_ctx.bw.dcn.clk.fclk_p_state_change_support = true;
+
+       if (context->bw_ctx.bw.dcn.clk.dispclk_khz < dc->debug.min_disp_clk_khz)
+               context->bw_ctx.bw.dcn.clk.dispclk_khz = dc->debug.min_disp_clk_khz;
+
+       unbounded_req_enabled = in_ctx->v20.dml_core_ctx.ms.UnboundedRequestEnabledThisState;
+
+       if (unbounded_req_enabled && pipe_cnt > 1) {
+               // Unbounded requesting should not ever be used when more than 1 pipe is enabled.
+               //ASSERT(false);
+               unbounded_req_enabled = false;
+       }
+
+       context->bw_ctx.bw.dcn.compbuf_size_kb = in_ctx->v20.dml_core_ctx.ip.config_return_buffer_size_in_kbytes;
+
+       for (dc_pipe_ctx_index = 0; dc_pipe_ctx_index < pipe_cnt; dc_pipe_ctx_index++) {
+               if (!context->res_ctx.pipe_ctx[dc_pipe_ctx_index].stream)
+                       continue;
+               /* The DML2 and the DC logic of determining pipe indices are different from each other so
+                * there is a need to know which DML pipe index maps to which DC pipe. The code below
+                * finds a dml_pipe_index from the plane id if a plane is valid. If a plane is not valid then
+                * it finds a dml_pipe_index from the stream id. */
+               if (get_plane_id(context, context->res_ctx.pipe_ctx[dc_pipe_ctx_index].plane_state, &plane_id)) {
+                       dml_pipe_idx = find_dml_pipe_idx_by_plane_id(in_ctx, plane_id);
+               } else {
+                       dml_pipe_idx = dml2_helper_find_dml_pipe_idx_by_stream_id(in_ctx, context->res_ctx.pipe_ctx[dc_pipe_ctx_index].stream->stream_id);
+               }
+
+               ASSERT(in_ctx->v20.scratch.dml_to_dc_pipe_mapping.dml_pipe_idx_to_stream_id_valid[dml_pipe_idx]);
+               ASSERT(in_ctx->v20.scratch.dml_to_dc_pipe_mapping.dml_pipe_idx_to_stream_id[dml_pipe_idx] == context->res_ctx.pipe_ctx[dc_pipe_ctx_index].stream->stream_id);
+
+               /* Use the dml_pipe_index here for the getters to fetch the correct values and dc_pipe_index in the pipe_ctx to populate them
+                * at the right locations.
+                */
+               populate_pipe_ctx_dlg_params_from_dml(&context->res_ctx.pipe_ctx[dc_pipe_ctx_index], &context->bw_ctx.dml2->v20.dml_core_ctx, dml_pipe_idx);
+
+               if (context->res_ctx.pipe_ctx[dc_pipe_ctx_index].stream->mall_stream_config.type == SUBVP_PHANTOM) {
+                       // Phantom pipe requires that DET_SIZE = 0 and no unbounded requests
+                       context->res_ctx.pipe_ctx[dc_pipe_ctx_index].det_buffer_size_kb = 0;
+                       context->res_ctx.pipe_ctx[dc_pipe_ctx_index].unbounded_req = false;
+               } else {
+                       context->res_ctx.pipe_ctx[dc_pipe_ctx_index].det_buffer_size_kb = dml_get_det_buffer_size_kbytes(&context->bw_ctx.dml2->v20.dml_core_ctx, dml_pipe_idx);
+                       context->res_ctx.pipe_ctx[dc_pipe_ctx_index].unbounded_req = unbounded_req_enabled;
+               }
+
+               context->bw_ctx.bw.dcn.compbuf_size_kb -= context->res_ctx.pipe_ctx[dc_pipe_ctx_index].det_buffer_size_kb;
+               context->res_ctx.pipe_ctx[dc_pipe_ctx_index].plane_res.bw.dppclk_khz = dml_get_dppclk_calculated(&context->bw_ctx.dml2->v20.dml_core_ctx, dml_pipe_idx) * 1000;
+               if (context->bw_ctx.bw.dcn.clk.dppclk_khz < context->res_ctx.pipe_ctx[dc_pipe_ctx_index].plane_res.bw.dppclk_khz)
+                       context->bw_ctx.bw.dcn.clk.dppclk_khz = context->res_ctx.pipe_ctx[dc_pipe_ctx_index].plane_res.bw.dppclk_khz;
+
+               dml_rq_dlg_get_rq_reg(&s->rq_regs, &in_ctx->v20.dml_core_ctx, dml_pipe_idx);
+               dml_rq_dlg_get_dlg_reg(&s->disp_dlg_regs, &s->disp_ttu_regs, &in_ctx->v20.dml_core_ctx, dml_pipe_idx);
+               dml2_update_pipe_ctx_dchub_regs(&s->rq_regs, &s->disp_dlg_regs, &s->disp_ttu_regs, &out_new_hw_state->pipe_ctx[dc_pipe_ctx_index]);
+
+               context->res_ctx.pipe_ctx[dc_pipe_ctx_index].surface_size_in_mall_bytes = dml_get_surface_size_for_mall(&context->bw_ctx.dml2->v20.dml_core_ctx, dml_pipe_idx);
+
+               /* Reuse MALL Allocation Sizes logic from dcn32_fpu.c */
+               /* Count from active, top pipes per plane only. Only add mall_ss_size_bytes for each unique plane. */
+               if (context->res_ctx.pipe_ctx[dc_pipe_ctx_index].stream && context->res_ctx.pipe_ctx[dc_pipe_ctx_index].plane_state &&
+                       (context->res_ctx.pipe_ctx[dc_pipe_ctx_index].top_pipe == NULL ||
+                       context->res_ctx.pipe_ctx[dc_pipe_ctx_index].plane_state != context->res_ctx.pipe_ctx[dc_pipe_ctx_index].top_pipe->plane_state) &&
+                       context->res_ctx.pipe_ctx[dc_pipe_ctx_index].prev_odm_pipe == NULL) {
+                       /* SS: all active surfaces stored in MALL */
+                       if (context->res_ctx.pipe_ctx[dc_pipe_ctx_index].stream->mall_stream_config.type != SUBVP_PHANTOM) {
+                               context->bw_ctx.bw.dcn.mall_ss_size_bytes += context->res_ctx.pipe_ctx[dc_pipe_ctx_index].surface_size_in_mall_bytes;
+                       } else {
+                               /* SUBVP: phantom surfaces only stored in MALL */
+                               context->bw_ctx.bw.dcn.mall_subvp_size_bytes += context->res_ctx.pipe_ctx[dc_pipe_ctx_index].surface_size_in_mall_bytes;
+                       }
+               }
+       }
+
+       context->bw_ctx.bw.dcn.clk.bw_dppclk_khz = context->bw_ctx.bw.dcn.clk.dppclk_khz;
+       context->bw_ctx.bw.dcn.clk.bw_dispclk_khz = context->bw_ctx.bw.dcn.clk.dispclk_khz;
+       context->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz = in_ctx->v20.dml_core_ctx.states.state_array[in_ctx->v20.scratch.mode_support_params.out_lowest_state_idx].dppclk_mhz
+               * 1000;
+       context->bw_ctx.bw.dcn.clk.max_supported_dispclk_khz = in_ctx->v20.dml_core_ctx.states.state_array[in_ctx->v20.scratch.mode_support_params.out_lowest_state_idx].dispclk_mhz
+               * 1000;
+}
+
+void dml2_extract_watermark_set(struct dcn_watermarks *watermark, struct display_mode_lib_st *dml_core_ctx)
+{
+       watermark->urgent_ns = dml_get_wm_urgent(dml_core_ctx) * 1000;
+       watermark->cstate_pstate.cstate_enter_plus_exit_ns = dml_get_wm_stutter_enter_exit(dml_core_ctx) * 1000;
+       watermark->cstate_pstate.cstate_exit_ns = dml_get_wm_stutter_exit(dml_core_ctx) * 1000;
+       watermark->cstate_pstate.pstate_change_ns = dml_get_wm_dram_clock_change(dml_core_ctx) * 1000;
+       watermark->pte_meta_urgent_ns = dml_get_wm_memory_trip(dml_core_ctx) * 1000;
+       watermark->frac_urg_bw_nom = dml_get_fraction_of_urgent_bandwidth(dml_core_ctx) * 1000;
+       watermark->frac_urg_bw_flip = dml_get_fraction_of_urgent_bandwidth_imm_flip(dml_core_ctx) * 1000;
+       watermark->urgent_latency_ns = dml_get_urgent_latency(dml_core_ctx) * 1000;
+       watermark->cstate_pstate.fclk_pstate_change_ns = dml_get_wm_fclk_change(dml_core_ctx) * 1000;
+       watermark->usr_retraining_ns = dml_get_wm_usr_retraining(dml_core_ctx) * 1000;
+}
+
+void dml2_initialize_det_scratch(struct dml2_context *in_ctx)
+{
+       int i;
+
+       for (i = 0; i < MAX_PLANES; i++) {
+               in_ctx->det_helper_scratch.dpps_per_surface[i] = 1;
+       }
+}
+
+static unsigned int find_planes_per_stream_and_stream_count(struct dml2_context *in_ctx, struct dml_display_cfg_st *dml_dispcfg, int *num_of_planes_per_stream)
+{
+       unsigned int plane_index, stream_index = 0, num_of_streams;
+
+       for (plane_index = 0; plane_index < dml_dispcfg->num_surfaces; plane_index++) {
+               /* Number of planes per stream */
+               num_of_planes_per_stream[stream_index] += 1;
+
+               if (plane_index + 1 < dml_dispcfg->num_surfaces && dml_dispcfg->plane.BlendingAndTiming[plane_index] != dml_dispcfg->plane.BlendingAndTiming[plane_index + 1])
+                       stream_index++;
+       }
+
+       num_of_streams = stream_index + 1;
+
+       return num_of_streams;
+}
+
+void dml2_apply_det_buffer_allocation_policy(struct dml2_context *in_ctx, struct dml_display_cfg_st *dml_dispcfg)
+{
+       unsigned int num_of_streams = 0, plane_index = 0, max_det_size, stream_index = 0;
+       int num_of_planes_per_stream[__DML_NUM_PLANES__] = { 0 };
+
+       max_det_size = in_ctx->config.det_segment_size * in_ctx->config.max_segments_per_hubp;
+
+       num_of_streams = find_planes_per_stream_and_stream_count(in_ctx, dml_dispcfg, num_of_planes_per_stream);
+
+       for (plane_index = 0; plane_index < dml_dispcfg->num_surfaces; plane_index++) {
+
+               dml_dispcfg->plane.DETSizeOverride[plane_index] = ((max_det_size / num_of_streams) / num_of_planes_per_stream[stream_index] / in_ctx->det_helper_scratch.dpps_per_surface[plane_index]);
+
+               /* If the override size is not divisible by det_segment_size then round off to nearest number divisible by det_segment_size as
+                * this is a requirement.
+                */
+               if (dml_dispcfg->plane.DETSizeOverride[plane_index] % in_ctx->config.det_segment_size != 0) {
+                       dml_dispcfg->plane.DETSizeOverride[plane_index] = dml_dispcfg->plane.DETSizeOverride[plane_index] & ~0x3F;
+               }
+
+               if (plane_index + 1 < dml_dispcfg->num_surfaces && dml_dispcfg->plane.BlendingAndTiming[plane_index] != dml_dispcfg->plane.BlendingAndTiming[plane_index + 1])
+                       stream_index++;
+       }
+}
+
+bool dml2_verify_det_buffer_configuration(struct dml2_context *in_ctx, struct dc_state *display_state, struct dml2_helper_det_policy_scratch *det_scratch)
+{
+       unsigned int i = 0, dml_pipe_idx = 0, plane_id = 0;
+       unsigned int max_det_size, total_det_allocated = 0;
+       bool need_recalculation = false;
+
+       max_det_size = in_ctx->config.det_segment_size * in_ctx->config.max_segments_per_hubp;
+
+       for (i = 0; i < MAX_PIPES; i++) {
+               if (!display_state->res_ctx.pipe_ctx[i].stream)
+                       continue;
+               if (get_plane_id(display_state, display_state->res_ctx.pipe_ctx[i].plane_state, &plane_id))
+                       dml_pipe_idx = find_dml_pipe_idx_by_plane_id(in_ctx, plane_id);
+               else
+                       dml_pipe_idx = dml2_helper_find_dml_pipe_idx_by_stream_id(in_ctx, display_state->res_ctx.pipe_ctx[i].stream->stream_id);
+               total_det_allocated += dml_get_det_buffer_size_kbytes(&in_ctx->v20.dml_core_ctx, dml_pipe_idx);
+               if (total_det_allocated > max_det_size) {
+                       need_recalculation = true;
+               }
+       }
+
+       /* Store the DPPPerSurface for correctly determining the number of planes in the next call. */
+       for (i = 0; i < MAX_PLANES; i++) {
+               det_scratch->dpps_per_surface[i] = in_ctx->v20.scratch.cur_display_config.hw.DPPPerSurface[i];
+       }
+
+       return need_recalculation;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_utils.h b/drivers/gpu/drm/amd/display/dc/dml2/dml2_utils.h
new file mode 100644 (file)
index 0000000..342d640
--- /dev/null
@@ -0,0 +1,143 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef _DML2_UTILS_H_
+#define _DML2_UTILS_H_
+
+#include "os_types.h"
+#include "dml2_dc_types.h"
+
+struct dc;
+struct dml_timing_cfg_st;
+struct dml2_dcn_clocks;
+struct dc_state;
+
+void dml2_util_copy_dml_timing(struct dml_timing_cfg_st *dml_timing_array, unsigned int dst_index, unsigned int src_index);
+void dml2_util_copy_dml_plane(struct dml_plane_cfg_st *dml_plane_array, unsigned int dst_index, unsigned int src_index);
+void dml2_util_copy_dml_surface(struct dml_surface_cfg_st *dml_surface_array, unsigned int dst_index, unsigned int src_index);
+void dml2_util_copy_dml_output(struct dml_output_cfg_st *dml_output_array, unsigned int dst_index, unsigned int src_index);
+unsigned int dml2_util_get_maximum_odm_combine_for_output(bool force_odm_4to1, enum dml_output_encoder_class encoder, bool dsc_enabled);
+void dml2_copy_clocks_to_dc_state(struct dml2_dcn_clocks *out_clks, struct dc_state *context);
+void dml2_extract_watermark_set(struct dcn_watermarks *watermark, struct display_mode_lib_st *dml_core_ctx);
+int dml2_helper_find_dml_pipe_idx_by_stream_id(struct dml2_context *ctx, unsigned int stream_id);
+bool is_dtbclk_required(const struct dc *dc, struct dc_state *context);
+
+/*
+ * dml2_dc_construct_pipes - This function will determine if we need additional pipes based
+ * on the DML calculated outputs for MPC, ODM and allocate them as necessary. This function
+ * could be called after in dml_validate_build_resource after dml_mode_pragramming like :
+ * {
+ *   ...
+ * map_hw_resources(&s->cur_display_config, &s->mode_support_info);
+ * result = dml_mode_programming(&in_ctx->dml_core_ctx, s->mode_support_params.out_lowest_state_idx, &s->cur_display_config, true);
+ * dml2_dc_construct_pipes(in_display_state, s->mode_support_info, out_hw_context);
+ * ...
+ * }
+ *
+ * @context: To obtain res_ctx and read other information like stream ID etc.
+ * @dml_mode_support_st : To get the ODM, MPC outputs as determined by the DML.
+ * @out_hw_context : Handle to the new hardware context.
+ *
+ *
+ * Return: None.
+ */
+void dml2_dc_construct_pipes(struct dc_state *context, struct dml_mode_support_info_st *dml_mode_support_st,
+               struct resource_context *out_hw_context);
+
+/*
+ * dml2_predict_pipe_split - This function is the dml2 version of predict split pipe. It predicts a
+ * if pipe split is required or not and returns the output as a bool.
+ * @context : dc_state.
+ * @pipe : old_index is the index of the pipe as derived from pipe_idx.
+ * @index : index of the pipe
+ *
+ *
+ * Return: Returns the result in boolean.
+ */
+bool dml2_predict_pipe_split(struct dc_state *context, display_pipe_params_st pipe, int index);
+
+/*
+ * dml2_build_mapped_resource - This function is the dml2 version of build_mapped_resource.
+ * In case of ODM, we need to build pipe hardware params again as done in dcn20_build_mapped_resource.
+ * @dc : struct dc
+ * @context : struct dc_state.
+ * @stream : stream whoose corresponding pipe params need to be modified.
+ *
+ *
+ * Return: Returns DC_OK if successful.
+ */
+enum dc_status dml2_build_mapped_resource(const struct dc *dc, struct dc_state *context, struct dc_stream_state *stream);
+
+/*
+ * dml2_extract_rq_regs - This function will extract information needed for struct _vcs_dpi_display_rq_regs_st
+ * and populate it.
+ * @context: To obtain and populate the res_ctx->pipe_ctx->rq_regs with DML outputs.
+ * @support : This structure has the DML intermediate outputs required to populate rq_regs.
+ *
+ *
+ * Return: None.
+ */
+
+ /*
+  * dml2_calculate_rq_and_dlg_params - This function will call into DML2 functions needed
+  * for populating rq, ttu and dlg param structures and populate it.
+  * @dc : struct dc
+  * @context : dc_state provides a handle to selectively populate pipe_ctx
+  * @out_new_hw_state: To obtain and populate the rq, dlg and ttu regs in
+  *                    out_new_hw_state->pipe_ctx with DML outputs.
+  * @in_ctx : This structure has the pointer to display_mode_lib_st.
+  * @pipe_cnt : DML functions to obtain RQ, TTu and DLG params need a pipe_index.
+  *                            This helps provide pipe_index in the pipe_cnt loop.
+  *
+  *
+  * Return: None.
+  */
+void dml2_calculate_rq_and_dlg_params(const struct dc *dc, struct dc_state *context, struct resource_context *out_new_hw_state, struct dml2_context *in_ctx, unsigned int pipe_cnt);
+
+/*
+ * dml2_apply_det_buffer_allocation_policy - This function will determine the DET Buffer size
+ * and return the number of streams.
+ * @dml2 : Handle for dml2 context
+ * @dml_dispcfg : dml_dispcfg is the DML2 struct representing the current display config
+ * Return : None.
+ */
+void dml2_apply_det_buffer_allocation_policy(struct dml2_context *in_ctx, struct dml_display_cfg_st *dml_dispcfg);
+
+/*
+ * dml2_verify_det_buffer_configuration - This function will verify if the allocated DET buffer exceeds
+ * the total available DET size available and outputs a boolean to indicate if recalulation is needed.
+ * @dml2 : Handle for dml2 context
+ * @dml_dispcfg : dml_dispcfg is the DML2 struct representing the current display config
+ * @struct dml2_helper_det_policy_scratch : Pointer to DET helper scratch
+ * Return : returns true if recalculation is required, false otherwise.
+ */
+bool dml2_verify_det_buffer_configuration(struct dml2_context *in_ctx, struct dc_state *display_state, struct dml2_helper_det_policy_scratch *det_scratch);
+
+/*
+ * dml2_initialize_det_scratch - This function will initialize the DET scratch space as per requirements.
+ * @dml2 : Handle for dml2 context
+ * Return : None
+ */
+void dml2_initialize_det_scratch(struct dml2_context *in_ctx);
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c
new file mode 100644 (file)
index 0000000..aed3e42
--- /dev/null
@@ -0,0 +1,730 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "display_mode_core.h"
+#include "dml2_internal_types.h"
+#include "dml2_utils.h"
+#include "dml2_policy.h"
+#include "dml2_translation_helper.h"
+#include "dml2_mall_phantom.h"
+#include "dml2_dc_resource_mgmt.h"
+
+
+static void initialize_dml2_ip_params(struct dml2_context *dml2, const struct dc *in_dc, struct ip_params_st *out)
+{
+       if (dml2->config.use_native_soc_bb_construction)
+               dml2_init_ip_params(dml2, in_dc, out);
+       else
+               dml2_translate_ip_params(in_dc, out);
+}
+
+static void initialize_dml2_soc_bbox(struct dml2_context *dml2, const struct dc *in_dc, struct soc_bounding_box_st *out)
+{
+       if (dml2->config.use_native_soc_bb_construction)
+               dml2_init_socbb_params(dml2, in_dc, out);
+       else
+               dml2_translate_socbb_params(in_dc, out);
+}
+
+static void initialize_dml2_soc_states(struct dml2_context *dml2,
+       const struct dc *in_dc, const struct soc_bounding_box_st *in_bbox, struct soc_states_st *out)
+{
+       if (dml2->config.use_native_soc_bb_construction)
+               dml2_init_soc_states(dml2, in_dc, in_bbox, out);
+       else
+               dml2_translate_soc_states(in_dc, out, in_dc->dml.soc.num_states);
+}
+
+static void map_hw_resources(struct dml2_context *dml2,
+               struct dml_display_cfg_st *in_out_display_cfg, struct dml_mode_support_info_st *mode_support_info)
+{
+       unsigned int num_pipes = 0;
+       int i, j;
+
+       for (i = 0; i < __DML_NUM_PLANES__; i++) {
+               in_out_display_cfg->hw.ODMMode[i] = mode_support_info->ODMMode[i];
+               in_out_display_cfg->hw.DPPPerSurface[i] = mode_support_info->DPPPerSurface[i];
+               in_out_display_cfg->hw.DSCEnabled[i] = mode_support_info->DSCEnabled[i];
+               in_out_display_cfg->hw.NumberOfDSCSlices[i] = mode_support_info->NumberOfDSCSlices[i];
+               in_out_display_cfg->hw.DLGRefClkFreqMHz = 50;
+
+               for (j = 0; j < mode_support_info->DPPPerSurface[i]; j++) {
+                       dml2->v20.scratch.dml_to_dc_pipe_mapping.dml_pipe_idx_to_stream_id[num_pipes] = dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_stream_id[i];
+                       dml2->v20.scratch.dml_to_dc_pipe_mapping.dml_pipe_idx_to_stream_id_valid[num_pipes] = true;
+                       dml2->v20.scratch.dml_to_dc_pipe_mapping.dml_pipe_idx_to_plane_id[num_pipes] = dml2->v20.scratch.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id[i];
+                       dml2->v20.scratch.dml_to_dc_pipe_mapping.dml_pipe_idx_to_plane_id_valid[num_pipes] = true;
+                       num_pipes++;
+               }
+       }
+}
+
+static unsigned int pack_and_call_dml_mode_support_ex(struct dml2_context *dml2,
+       const struct dml_display_cfg_st *display_cfg,
+       struct dml_mode_support_info_st *evaluation_info)
+{
+       struct dml2_wrapper_scratch *s = &dml2->v20.scratch;
+
+       s->mode_support_params.mode_lib = &dml2->v20.dml_core_ctx;
+       s->mode_support_params.in_display_cfg = display_cfg;
+       s->mode_support_params.out_evaluation_info = evaluation_info;
+
+       memset(evaluation_info, 0, sizeof(struct dml_mode_support_info_st));
+       s->mode_support_params.out_lowest_state_idx = 0;
+
+       return dml_mode_support_ex(&s->mode_support_params);
+}
+
+static bool optimize_configuration(struct dml2_context *dml2, struct dml2_wrapper_optimize_configuration_params *p)
+{
+       int unused_dpps = p->ip_params->max_num_dpp;
+       int i, j;
+       int odms_needed, refresh_rate_hz, dpps_needed, subvp_height, pstate_width_fw_delay_lines, surface_count;
+       int subvp_timing_to_add, new_timing_index, subvp_surface_to_add, new_surface_index;
+       float frame_time_sec, max_frame_time_sec;
+       int largest_blend_and_timing = 0;
+       bool optimization_done = false;
+
+       for (i = 0; i < (int) p->cur_display_config->num_timings; i++) {
+               if (p->cur_display_config->plane.BlendingAndTiming[i] > largest_blend_and_timing)
+                       largest_blend_and_timing = p->cur_display_config->plane.BlendingAndTiming[i];
+       }
+
+       if (p->new_policy != p->cur_policy)
+               *p->new_policy = *p->cur_policy;
+
+       if (p->new_display_config != p->cur_display_config)
+               *p->new_display_config = *p->cur_display_config;
+
+       // Optimize P-State Support
+       if (dml2->config.use_native_pstate_optimization) {
+               if (p->cur_mode_support_info->DRAMClockChangeSupport[0] == dml_dram_clock_change_unsupported) {
+                       // Find a display with < 120Hz refresh rate with maximal refresh rate that's not already subvp
+                       subvp_timing_to_add = -1;
+                       subvp_surface_to_add = -1;
+                       max_frame_time_sec = 0;
+                       surface_count = 0;
+                       for (i = 0; i < (int) p->cur_display_config->num_timings; i++) {
+                               refresh_rate_hz = (int)div_u64((unsigned long long) p->cur_display_config->timing.PixelClock[i] * 1000 * 1000,
+                                       (p->cur_display_config->timing.HTotal[i] * p->cur_display_config->timing.VTotal[i]));
+                               if (refresh_rate_hz < 120) {
+                                       // Check its upstream surfaces to see if this one could be converted to subvp.
+                                       dpps_needed = 0;
+                               for (j = 0; j < (int) p->cur_display_config->num_surfaces; j++) {
+                                       if (p->cur_display_config->plane.BlendingAndTiming[j] == i &&
+                                               p->cur_display_config->plane.UseMALLForPStateChange[j] == dml_use_mall_pstate_change_disable) {
+                                               dpps_needed += p->cur_mode_support_info->DPPPerSurface[j];
+                                               subvp_surface_to_add = j;
+                                               surface_count++;
+                                       }
+                               }
+
+                               if (surface_count == 1 && dpps_needed > 0 && dpps_needed <= unused_dpps) {
+                                       frame_time_sec = (float)1 / refresh_rate_hz;
+                                       if (frame_time_sec > max_frame_time_sec) {
+                                               max_frame_time_sec = frame_time_sec;
+                                               subvp_timing_to_add = i;
+                                               }
+                                       }
+                               }
+                       }
+                       if (subvp_timing_to_add >= 0) {
+                               new_timing_index = p->new_display_config->num_timings++;
+                               new_surface_index = p->new_display_config->num_surfaces++;
+                               // Add a phantom pipe reflecting the main pipe's timing
+                               dml2_util_copy_dml_timing(&p->new_display_config->timing, new_timing_index, subvp_timing_to_add);
+
+                               pstate_width_fw_delay_lines = (int)(((double)(p->config->svp_pstate.subvp_fw_processing_delay_us +
+                                       p->config->svp_pstate.subvp_pstate_allow_width_us) / 1000000) *
+                               (p->new_display_config->timing.PixelClock[subvp_timing_to_add] * 1000 * 1000) /
+                               (double)p->new_display_config->timing.HTotal[subvp_timing_to_add]);
+
+                               subvp_height = p->cur_mode_support_info->SubViewportLinesNeededInMALL[subvp_timing_to_add] + pstate_width_fw_delay_lines;
+
+                               p->new_display_config->timing.VActive[new_timing_index] = subvp_height;
+                               p->new_display_config->timing.VTotal[new_timing_index] = subvp_height +
+                               p->new_display_config->timing.VTotal[subvp_timing_to_add] - p->new_display_config->timing.VActive[subvp_timing_to_add];
+
+                               p->new_display_config->output.OutputDisabled[new_timing_index] = true;
+
+                               p->new_display_config->plane.UseMALLForPStateChange[subvp_surface_to_add] = dml_use_mall_pstate_change_sub_viewport;
+
+                               dml2_util_copy_dml_plane(&p->new_display_config->plane, new_surface_index, subvp_surface_to_add);
+                               dml2_util_copy_dml_surface(&p->new_display_config->surface, new_surface_index, subvp_surface_to_add);
+
+                               p->new_display_config->plane.ViewportHeight[new_surface_index] = subvp_height;
+                               p->new_display_config->plane.ViewportHeightChroma[new_surface_index] = subvp_height;
+                               p->new_display_config->plane.ViewportStationary[new_surface_index] = false;
+
+                               p->new_display_config->plane.UseMALLForStaticScreen[new_surface_index] = dml_use_mall_static_screen_disable;
+                               p->new_display_config->plane.UseMALLForPStateChange[new_surface_index] = dml_use_mall_pstate_change_phantom_pipe;
+
+                               p->new_display_config->plane.NumberOfCursors[new_surface_index] = 0;
+
+                               p->new_policy->ImmediateFlipRequirement[new_surface_index] = dml_immediate_flip_not_required;
+
+                               p->new_display_config->plane.BlendingAndTiming[new_surface_index] = new_timing_index;
+
+                               optimization_done = true;
+                       }
+               }
+       }
+
+       // Optimize Clocks
+       if (!optimization_done) {
+               if (largest_blend_and_timing == 0 && p->cur_policy->ODMUse[0] == dml_odm_use_policy_combine_as_needed && dml2->config.minimize_dispclk_using_odm) {
+                       odms_needed = dml2_util_get_maximum_odm_combine_for_output(dml2->config.optimize_odm_4to1,
+                               p->cur_display_config->output.OutputEncoder[0], p->cur_mode_support_info->DSCEnabled[0]) - 1;
+
+                       if (odms_needed <= unused_dpps) {
+                               unused_dpps -= odms_needed;
+
+                               if (odms_needed == 1) {
+                                       p->new_policy->ODMUse[0] = dml_odm_use_policy_combine_2to1;
+                                       optimization_done = true;
+                               } else if (odms_needed == 3) {
+                                       p->new_policy->ODMUse[0] = dml_odm_use_policy_combine_4to1;
+                                       optimization_done = true;
+                               } else
+                                       optimization_done = false;
+                       }
+               }
+       }
+
+       return optimization_done;
+}
+
+static int calculate_lowest_supported_state_for_temp_read(struct dml2_context *dml2, struct dc_state *display_state)
+{
+       struct dml2_calculate_lowest_supported_state_for_temp_read_scratch *s = &dml2->v20.scratch.dml2_calculate_lowest_supported_state_for_temp_read_scratch;
+       struct dml2_wrapper_scratch *s_global = &dml2->v20.scratch;
+
+       unsigned int dml_result = 0;
+       int result = -1, i, j;
+
+       build_unoptimized_policy_settings(dml2->v20.dml_core_ctx.project, &dml2->v20.dml_core_ctx.policy);
+
+       /* Zero out before each call before proceeding */
+       memset(s, 0, sizeof(struct dml2_calculate_lowest_supported_state_for_temp_read_scratch));
+       memset(&s_global->mode_support_params, 0, sizeof(struct dml_mode_support_ex_params_st));
+       memset(&s_global->dml_to_dc_pipe_mapping, 0, sizeof(struct dml2_dml_to_dc_pipe_mapping));
+
+       for (i = 0; i < dml2->config.dcn_pipe_count; i++) {
+               /* Calling resource_build_scaling_params will populate the pipe params
+                * with the necessary information needed for correct DML calculations
+                * This is also done in DML1 driver code path and hence display_state
+                * cannot be const.
+                */
+               struct pipe_ctx *pipe = &display_state->res_ctx.pipe_ctx[i];
+
+               if (pipe->plane_state) {
+                       if (!dml2->config.callbacks.build_scaling_params(pipe)) {
+                               ASSERT(false);
+                               return false;
+                       }
+               }
+       }
+
+       map_dc_state_into_dml_display_cfg(dml2, display_state, &s->cur_display_config);
+
+       for (i = 0; i < dml2->v20.dml_core_ctx.states.num_states; i++) {
+               s->uclk_change_latencies[i] = dml2->v20.dml_core_ctx.states.state_array[i].dram_clock_change_latency_us;
+       }
+
+       for (i = 0; i < 4; i++) {
+               for (j = 0; j < dml2->v20.dml_core_ctx.states.num_states; j++) {
+                       dml2->v20.dml_core_ctx.states.state_array[j].dram_clock_change_latency_us = s_global->dummy_pstate_table[i].dummy_pstate_latency_us;
+               }
+
+               dml_result = pack_and_call_dml_mode_support_ex(dml2, &s->cur_display_config, &s->evaluation_info);
+
+               if (dml_result && s->evaluation_info.DRAMClockChangeSupport[0] == dml_dram_clock_change_vactive) {
+                       map_hw_resources(dml2, &s->cur_display_config, &s->evaluation_info);
+                       dml_result = dml_mode_programming(&dml2->v20.dml_core_ctx, s_global->mode_support_params.out_lowest_state_idx, &s->cur_display_config, true);
+
+                       ASSERT(dml_result);
+
+                       dml2_extract_watermark_set(&dml2->v20.g6_temp_read_watermark_set, &dml2->v20.dml_core_ctx);
+                       dml2->v20.g6_temp_read_watermark_set.cstate_pstate.fclk_pstate_change_ns = dml2->v20.g6_temp_read_watermark_set.cstate_pstate.pstate_change_ns;
+
+                       result = s_global->mode_support_params.out_lowest_state_idx;
+
+                       while (dml2->v20.dml_core_ctx.states.state_array[result].dram_speed_mts < s_global->dummy_pstate_table[i].dram_speed_mts)
+                               result++;
+
+                       break;
+               }
+       }
+
+       for (i = 0; i < dml2->v20.dml_core_ctx.states.num_states; i++) {
+               dml2->v20.dml_core_ctx.states.state_array[i].dram_clock_change_latency_us = s->uclk_change_latencies[i];
+       }
+
+       return result;
+}
+
+static void copy_dummy_pstate_table(struct dummy_pstate_entry *dest, struct dummy_pstate_entry *src, unsigned int num_entries)
+{
+       for (int i = 0; i < num_entries; i++) {
+               dest[i] = src[i];
+       }
+}
+
+static bool are_timings_requiring_odm_doing_blending(const struct dml_display_cfg_st *display_cfg,
+               const struct dml_mode_support_info_st *evaluation_info)
+{
+       unsigned int planes_per_timing[__DML_NUM_PLANES__] = {0};
+       int i;
+
+       for (i = 0; i < display_cfg->num_surfaces; i++)
+               planes_per_timing[display_cfg->plane.BlendingAndTiming[i]]++;
+
+       for (i = 0; i < __DML_NUM_PLANES__; i++) {
+               if (planes_per_timing[i] > 1 && evaluation_info->ODMMode[i] != dml_odm_mode_bypass)
+                       return true;
+       }
+
+       return false;
+}
+
+static bool does_configuration_meet_sw_policies(struct dml2_context *ctx, const struct dml_display_cfg_st *display_cfg,
+       const struct dml_mode_support_info_st *evaluation_info)
+{
+       bool pass = true;
+
+       if (!ctx->config.enable_windowed_mpo_odm) {
+               if (are_timings_requiring_odm_doing_blending(display_cfg, evaluation_info))
+                       pass = false;
+       }
+
+       return pass;
+}
+
+static bool dml_mode_support_wrapper(struct dml2_context *dml2,
+               struct dc_state *display_state)
+{
+       struct dml2_wrapper_scratch *s = &dml2->v20.scratch;
+       unsigned int result = 0, i;
+       unsigned int optimized_result = true;
+
+       build_unoptimized_policy_settings(dml2->v20.dml_core_ctx.project, &dml2->v20.dml_core_ctx.policy);
+
+       /* Zero out before each call before proceeding */
+       memset(&s->cur_display_config, 0, sizeof(struct dml_display_cfg_st));
+       memset(&s->mode_support_params, 0, sizeof(struct dml_mode_support_ex_params_st));
+       memset(&s->dml_to_dc_pipe_mapping, 0, sizeof(struct dml2_dml_to_dc_pipe_mapping));
+       memset(&s->optimize_configuration_params, 0, sizeof(struct dml2_wrapper_optimize_configuration_params));
+
+       for (i = 0; i < dml2->config.dcn_pipe_count; i++) {
+               /* Calling resource_build_scaling_params will populate the pipe params
+                * with the necessary information needed for correct DML calculations
+                * This is also done in DML1 driver code path and hence display_state
+                * cannot be const.
+                */
+               struct pipe_ctx *pipe = &display_state->res_ctx.pipe_ctx[i];
+
+               if (pipe->plane_state) {
+                       if (!dml2->config.callbacks.build_scaling_params(pipe)) {
+                               ASSERT(false);
+                               return false;
+                       }
+               }
+       }
+
+       map_dc_state_into_dml_display_cfg(dml2, display_state, &s->cur_display_config);
+       if (!dml2->config.skip_hw_state_mapping)
+               dml2_apply_det_buffer_allocation_policy(dml2, &s->cur_display_config);
+
+       result = pack_and_call_dml_mode_support_ex(dml2,
+               &s->cur_display_config,
+               &s->mode_support_info);
+
+       if (result)
+               result = does_configuration_meet_sw_policies(dml2, &s->cur_display_config, &s->mode_support_info);
+
+       // Try to optimize
+       if (result) {
+               s->cur_policy = dml2->v20.dml_core_ctx.policy;
+               s->optimize_configuration_params.dml_core_ctx = &dml2->v20.dml_core_ctx;
+               s->optimize_configuration_params.config = &dml2->config;
+               s->optimize_configuration_params.ip_params = &dml2->v20.dml_core_ctx.ip;
+               s->optimize_configuration_params.cur_display_config = &s->cur_display_config;
+               s->optimize_configuration_params.cur_mode_support_info = &s->mode_support_info;
+               s->optimize_configuration_params.cur_policy = &s->cur_policy;
+               s->optimize_configuration_params.new_display_config = &s->new_display_config;
+               s->optimize_configuration_params.new_policy = &s->new_policy;
+
+               while (optimized_result && optimize_configuration(dml2, &s->optimize_configuration_params)) {
+                       dml2->v20.dml_core_ctx.policy = s->new_policy;
+                       optimized_result = pack_and_call_dml_mode_support_ex(dml2,
+                               &s->new_display_config,
+                               &s->mode_support_info);
+
+                       if (optimized_result)
+                               optimized_result = does_configuration_meet_sw_policies(dml2, &s->new_display_config, &s->mode_support_info);
+
+                       // If the new optimized state is supposed, then set current = new
+                       if (optimized_result) {
+                               s->cur_display_config = s->new_display_config;
+                               s->cur_policy = s->new_policy;
+                       } else {
+                               // Else, restore policy to current
+                               dml2->v20.dml_core_ctx.policy = s->cur_policy;
+                       }
+               }
+
+               // Optimize ended with a failed config, so we need to restore DML state to last passing
+               if (!optimized_result) {
+                       result = pack_and_call_dml_mode_support_ex(dml2,
+                               &s->cur_display_config,
+                               &s->mode_support_info);
+               }
+       }
+
+       if (result)
+               map_hw_resources(dml2, &s->cur_display_config, &s->mode_support_info);
+
+       return result;
+}
+
+static int find_drr_eligible_stream(struct dc_state *display_state)
+{
+       int i;
+
+       for (i = 0; i < display_state->stream_count; i++) {
+               if (display_state->streams[i]->mall_stream_config.type == SUBVP_NONE
+                       && display_state->streams[i]->ignore_msa_timing_param) {
+                       // Use ignore_msa_timing_param flag to identify as DRR
+                       return i;
+               }
+       }
+
+       return -1;
+}
+
+static bool optimize_pstate_with_svp_and_drr(struct dml2_context *dml2, struct dc_state *display_state)
+{
+       struct dml2_wrapper_scratch *s = &dml2->v20.scratch;
+       bool pstate_optimization_done = false;
+       bool pstate_optimization_success = false;
+       bool result = false;
+       int drr_display_index = 0, non_svp_streams = 0;
+       bool force_svp = dml2->config.svp_pstate.force_enable_subvp;
+       bool advanced_pstate_switching = false;
+
+       display_state->bw_ctx.bw.dcn.clk.fw_based_mclk_switching = false;
+       display_state->bw_ctx.bw.dcn.legacy_svp_drr_stream_index_valid = false;
+
+       result = dml_mode_support_wrapper(dml2, display_state);
+
+       if (!result) {
+               pstate_optimization_done = true;
+       } else if (!advanced_pstate_switching ||
+               (s->mode_support_info.DRAMClockChangeSupport[0] != dml_dram_clock_change_unsupported && !force_svp)) {
+               pstate_optimization_success = true;
+               pstate_optimization_done = true;
+       }
+
+       if (display_state->stream_count == 1 && dml2->config.callbacks.can_support_mclk_switch_using_fw_based_vblank_stretch(dml2->config.callbacks.dc, display_state)) {
+                       display_state->bw_ctx.bw.dcn.clk.fw_based_mclk_switching = true;
+
+                       result = dml_mode_support_wrapper(dml2, display_state);
+       } else {
+               non_svp_streams = display_state->stream_count;
+
+               while (!pstate_optimization_done) {
+                       result = dml_mode_programming(&dml2->v20.dml_core_ctx, s->mode_support_params.out_lowest_state_idx, &s->cur_display_config, true);
+
+                       // Always try adding SVP first
+                       if (result)
+                               result = dml2_svp_add_phantom_pipe_to_dc_state(dml2, display_state, &s->mode_support_info);
+                       else
+                               pstate_optimization_done = true;
+
+
+                       if (result) {
+                               result = dml_mode_support_wrapper(dml2, display_state);
+                       } else {
+                               pstate_optimization_done = true;
+                       }
+
+                       if (result) {
+                               non_svp_streams--;
+
+                               if (s->mode_support_info.DRAMClockChangeSupport[0] != dml_dram_clock_change_unsupported) {
+                                       if (dml2_svp_validate_static_schedulability(dml2, display_state, s->mode_support_info.DRAMClockChangeSupport[0])) {
+                                               pstate_optimization_success = true;
+                                               pstate_optimization_done = true;
+                                       } else {
+                                               pstate_optimization_success = false;
+                                               pstate_optimization_done = false;
+                                       }
+                               } else {
+                                       drr_display_index = find_drr_eligible_stream(display_state);
+
+                                       // If there is only 1 remaining non SubVP pipe that is DRR, check static
+                                       // schedulability for SubVP + DRR.
+                                       if (non_svp_streams == 1 && drr_display_index >= 0) {
+                                               if (dml2_svp_drr_schedulable(dml2, display_state, &display_state->streams[drr_display_index]->timing)) {
+                                                       display_state->bw_ctx.bw.dcn.legacy_svp_drr_stream_index_valid = true;
+                                                       display_state->bw_ctx.bw.dcn.legacy_svp_drr_stream_index = drr_display_index;
+                                                       result = dml_mode_support_wrapper(dml2, display_state);
+                                               }
+
+                                               if (result && s->mode_support_info.DRAMClockChangeSupport[0] != dml_dram_clock_change_unsupported) {
+                                                       pstate_optimization_success = true;
+                                                       pstate_optimization_done = true;
+                                               } else {
+                                                       pstate_optimization_success = false;
+                                                       pstate_optimization_done = false;
+                                               }
+                                       }
+
+                                       if (pstate_optimization_success) {
+                                               pstate_optimization_done = true;
+                                       } else {
+                                               pstate_optimization_done = false;
+                                       }
+                               }
+                       }
+               }
+       }
+
+       if (!pstate_optimization_success) {
+               dml2_svp_remove_all_phantom_pipes(dml2, display_state);
+               display_state->bw_ctx.bw.dcn.clk.fw_based_mclk_switching = false;
+               display_state->bw_ctx.bw.dcn.legacy_svp_drr_stream_index_valid = false;
+               result = dml_mode_support_wrapper(dml2, display_state);
+       }
+
+       return result;
+}
+
+static bool call_dml_mode_support_and_programming(struct dc_state *context)
+{
+       unsigned int result = 0;
+       unsigned int min_state;
+       int min_state_for_g6_temp_read = 0;
+       struct dml2_context *dml2 = context->bw_ctx.dml2;
+       struct dml2_wrapper_scratch *s = &dml2->v20.scratch;
+
+       min_state_for_g6_temp_read = calculate_lowest_supported_state_for_temp_read(dml2, context);
+
+       ASSERT(min_state_for_g6_temp_read >= 0);
+
+       if (!dml2->config.use_native_pstate_optimization) {
+               result = optimize_pstate_with_svp_and_drr(dml2, context);
+       } else {
+               result = dml_mode_support_wrapper(dml2, context);
+       }
+
+       /* Upon trying to sett certain frequencies in FRL, min_state_for_g6_temp_read is reported as -1. This leads to an invalid value of min_state causing crashes later on.
+        * Use the default logic for min_state only when min_state_for_g6_temp_read is a valid value. In other cases, use the value calculated by the DML directly.
+        */
+       if (min_state_for_g6_temp_read >= 0)
+               min_state = min_state_for_g6_temp_read > s->mode_support_params.out_lowest_state_idx ? min_state_for_g6_temp_read : s->mode_support_params.out_lowest_state_idx;
+       else
+               min_state = s->mode_support_params.out_lowest_state_idx;
+
+       if (result)
+               result = dml_mode_programming(&dml2->v20.dml_core_ctx, min_state, &s->cur_display_config, true);
+
+       return result;
+}
+
+static bool dml2_validate_and_build_resource(const struct dc *in_dc, struct dc_state *context)
+{
+       struct dml2_context *dml2 = context->bw_ctx.dml2;
+       struct dml2_wrapper_scratch *s = &dml2->v20.scratch;
+       struct dml2_dcn_clocks out_clks;
+       unsigned int result = 0;
+       bool need_recalculation = false;
+
+       if (!context || context->stream_count == 0)
+               return true;
+
+       /* Zero out before each call before proceeding */
+       memset(&dml2->v20.scratch, 0, sizeof(struct dml2_wrapper_scratch));
+       memset(&dml2->v20.dml_core_ctx.policy, 0, sizeof(struct dml_mode_eval_policy_st));
+       memset(&dml2->v20.dml_core_ctx.ms, 0, sizeof(struct mode_support_st));
+       memset(&dml2->v20.dml_core_ctx.mp, 0, sizeof(struct mode_program_st));
+
+       /* Initialize DET scratch */
+       dml2_initialize_det_scratch(dml2);
+
+       copy_dummy_pstate_table(s->dummy_pstate_table, in_dc->clk_mgr->bw_params->dummy_pstate_table, 4);
+
+       result = call_dml_mode_support_and_programming(context);
+       /* Call map dc pipes to map the pipes based on the DML output. For correctly determining if recalculation
+        * is required or not, the resource context needs to correctly reflect the number of active pipes. We would
+        * only know the correct number if active pipes after dml2_map_dc_pipes is called.
+        */
+       if (result && !dml2->config.skip_hw_state_mapping)
+               dml2_map_dc_pipes(dml2, context, &s->cur_display_config, &s->dml_to_dc_pipe_mapping, in_dc->current_state);
+
+       /* Verify and update DET Buffer configuration if needed. dml2_verify_det_buffer_configuration will check if DET Buffer
+        * size needs to be updated. If yes it will update the DETOverride variable and set need_recalculation flag to true.
+        * Based on that flag, run mode support again. Verification needs to be run after dml_mode_programming because the getters
+        * return correct det buffer values only after dml_mode_programming is called.
+        */
+       if (result && !dml2->config.skip_hw_state_mapping) {
+               need_recalculation = dml2_verify_det_buffer_configuration(dml2, context, &dml2->det_helper_scratch);
+               if (need_recalculation) {
+                       /* Engage the DML again if recalculation is required. */
+                       call_dml_mode_support_and_programming(context);
+                       if (!dml2->config.skip_hw_state_mapping) {
+                               dml2_map_dc_pipes(dml2, context, &s->cur_display_config, &s->dml_to_dc_pipe_mapping, in_dc->current_state);
+                       }
+                       need_recalculation = dml2_verify_det_buffer_configuration(dml2, context, &dml2->det_helper_scratch);
+                       ASSERT(need_recalculation == false);
+               }
+       }
+
+       if (result) {
+               unsigned int lowest_state_idx = s->mode_support_params.out_lowest_state_idx;
+               out_clks.dispclk_khz = (unsigned int)dml2->v20.dml_core_ctx.mp.Dispclk_calculated * 1000;
+               out_clks.p_state_supported = s->mode_support_info.DRAMClockChangeSupport[0] != dml_dram_clock_change_unsupported;
+               if (in_dc->config.use_default_clock_table &&
+                       (lowest_state_idx < dml2->v20.dml_core_ctx.states.num_states - 1)) {
+                       lowest_state_idx = dml2->v20.dml_core_ctx.states.num_states - 1;
+                       out_clks.dispclk_khz = (unsigned int)dml2->v20.dml_core_ctx.states.state_array[lowest_state_idx].dispclk_mhz * 1000;
+               }
+
+               out_clks.dcfclk_khz = (unsigned int)dml2->v20.dml_core_ctx.states.state_array[lowest_state_idx].dcfclk_mhz * 1000;
+               out_clks.fclk_khz = (unsigned int)dml2->v20.dml_core_ctx.states.state_array[lowest_state_idx].fabricclk_mhz * 1000;
+               out_clks.uclk_mts = (unsigned int)dml2->v20.dml_core_ctx.states.state_array[lowest_state_idx].dram_speed_mts;
+               out_clks.phyclk_khz = (unsigned int)dml2->v20.dml_core_ctx.states.state_array[lowest_state_idx].phyclk_mhz * 1000;
+               out_clks.socclk_khz = (unsigned int)dml2->v20.dml_core_ctx.states.state_array[lowest_state_idx].socclk_mhz * 1000;
+               out_clks.ref_dtbclk_khz = (unsigned int)dml2->v20.dml_core_ctx.states.state_array[lowest_state_idx].dtbclk_mhz * 1000;
+               context->bw_ctx.bw.dcn.clk.dtbclk_en = is_dtbclk_required(in_dc, context);
+
+               if (!dml2->config.skip_hw_state_mapping) {
+                       /* Call dml2_calculate_rq_and_dlg_params */
+                       dml2_calculate_rq_and_dlg_params(in_dc, context, &context->res_ctx, dml2, in_dc->res_pool->pipe_count);
+               }
+
+               dml2_copy_clocks_to_dc_state(&out_clks, context);
+               dml2_extract_watermark_set(&context->bw_ctx.bw.dcn.watermarks.a, &dml2->v20.dml_core_ctx);
+               dml2_extract_watermark_set(&context->bw_ctx.bw.dcn.watermarks.b, &dml2->v20.dml_core_ctx);
+               memcpy(&context->bw_ctx.bw.dcn.watermarks.c, &dml2->v20.g6_temp_read_watermark_set, sizeof(context->bw_ctx.bw.dcn.watermarks.c));
+               dml2_extract_watermark_set(&context->bw_ctx.bw.dcn.watermarks.d, &dml2->v20.dml_core_ctx);
+       }
+
+       return result;
+}
+
+static bool dml2_validate_only(const struct dc_state *context)
+{
+       struct dml2_context *dml2 = context->bw_ctx.dml2;
+       unsigned int result = 0;
+
+       if (!context || context->stream_count == 0)
+               return true;
+
+       /* Zero out before each call before proceeding */
+       memset(&dml2->v20.scratch, 0, sizeof(struct dml2_wrapper_scratch));
+       memset(&dml2->v20.dml_core_ctx.policy, 0, sizeof(struct dml_mode_eval_policy_st));
+       memset(&dml2->v20.dml_core_ctx.ms, 0, sizeof(struct mode_support_st));
+       memset(&dml2->v20.dml_core_ctx.mp, 0, sizeof(struct mode_program_st));
+
+       build_unoptimized_policy_settings(dml2->v20.dml_core_ctx.project, &dml2->v20.dml_core_ctx.policy);
+
+       map_dc_state_into_dml_display_cfg(dml2, context, &dml2->v20.scratch.cur_display_config);
+
+       result = pack_and_call_dml_mode_support_ex(dml2,
+               &dml2->v20.scratch.cur_display_config,
+               &dml2->v20.scratch.mode_support_info);
+
+       return (result == 1) ? true : false;
+}
+
+static void dml2_apply_debug_options(const struct dc *dc, struct dml2_context *dml2)
+{
+       if (dc->debug.override_odm_optimization) {
+               dml2->config.minimize_dispclk_using_odm = dc->debug.minimize_dispclk_using_odm;
+       }
+}
+
+bool dml2_validate(const struct dc *in_dc, struct dc_state *context, bool fast_validate)
+{
+       bool out = false;
+
+       if (!(context->bw_ctx.dml2))
+               return false;
+       dml2_apply_debug_options(in_dc, context->bw_ctx.dml2);
+
+
+       /* Use dml_validate_only for fast_validate path */
+       if (fast_validate)
+               out = dml2_validate_only(context);
+       else
+               out = dml2_validate_and_build_resource(in_dc, context);
+       return out;
+}
+
+bool dml2_create(const struct dc *in_dc, const struct dml2_configuration_options *config, struct dml2_context **dml2)
+{
+       // Allocate Mode Lib Ctx
+       *dml2 = (struct dml2_context *) kzalloc(sizeof(struct dml2_context), GFP_KERNEL);
+
+       if (!(*dml2))
+               return false;
+
+       // Store config options
+       (*dml2)->config = *config;
+
+       switch (in_dc->ctx->dce_version) {
+       case DCN_VERSION_3_2:
+               (*dml2)->v20.dml_core_ctx.project = dml_project_dcn32;
+               break;
+       case DCN_VERSION_3_21:
+               (*dml2)->v20.dml_core_ctx.project = dml_project_dcn321;
+               break;
+       default:
+               (*dml2)->v20.dml_core_ctx.project = dml_project_default;
+               break;
+       }
+
+       initialize_dml2_ip_params(*dml2, in_dc, &(*dml2)->v20.dml_core_ctx.ip);
+
+       initialize_dml2_soc_bbox(*dml2, in_dc, &(*dml2)->v20.dml_core_ctx.soc);
+
+       initialize_dml2_soc_states(*dml2, in_dc, &(*dml2)->v20.dml_core_ctx.soc, &(*dml2)->v20.dml_core_ctx.states);
+
+       return true;
+}
+
+void dml2_destroy(struct dml2_context *dml2)
+{
+       if (!dml2)
+               return;
+
+       kfree(dml2);
+}
+
+void dml2_extract_dram_and_fclk_change_support(struct dml2_context *dml2,
+       unsigned int *fclk_change_support, unsigned int *dram_clk_change_support)
+{
+       *fclk_change_support = (unsigned int) dml2->v20.dml_core_ctx.ms.support.FCLKChangeSupport[0];
+       *dram_clk_change_support = (unsigned int) dml2->v20.dml_core_ctx.ms.support.DRAMClockChangeSupport[0];
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h b/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h
new file mode 100644 (file)
index 0000000..e767260
--- /dev/null
@@ -0,0 +1,210 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef _DML2_WRAPPER_H_
+#define _DML2_WRAPPER_H_
+
+#include "os_types.h"
+
+#define DML2_MAX_NUM_DPM_LVL 30
+
+struct dml2_context;
+struct display_mode_lib_st;
+struct dc;
+struct pipe_ctx;
+struct dc_plane_state;
+struct dc_sink;
+struct dc_stream_state;
+struct resource_context;
+struct display_stream_compressor;
+
+// Configuration of the MALL on the SoC
+struct dml2_soc_mall_info {
+       // Cache line size of 0 means MALL is not enabled/present
+       unsigned int cache_line_size_bytes;
+       unsigned int cache_num_ways;
+       unsigned int max_cab_allocation_bytes;
+
+       unsigned int mblk_width_pixels;
+       unsigned int mblk_size_bytes;
+       unsigned int mblk_height_4bpe_pixels;
+       unsigned int mblk_height_8bpe_pixels;
+};
+
+// Output of DML2 for clock requirements
+struct dml2_dcn_clocks {
+       unsigned int dispclk_khz;
+       unsigned int dcfclk_khz;
+       unsigned int fclk_khz;
+       unsigned int uclk_mts;
+       unsigned int phyclk_khz;
+       unsigned int socclk_khz;
+       unsigned int ref_dtbclk_khz;
+       bool p_state_supported;
+       unsigned int cab_num_ways_required;
+};
+
+struct dml2_dc_callbacks {
+       struct dc *dc;
+       bool (*build_scaling_params)(struct pipe_ctx *pipe_ctx);
+       bool (*can_support_mclk_switch_using_fw_based_vblank_stretch)(struct dc *dc, struct dc_state *context);
+       bool (*acquire_secondary_pipe_for_mpc_odm)(const struct dc *dc, struct dc_state *state, struct pipe_ctx *pri_pipe, struct pipe_ctx *sec_pipe, bool odm);
+};
+
+struct dml2_dc_svp_callbacks {
+       struct dc *dc;
+       bool (*build_scaling_params)(struct pipe_ctx *pipe_ctx);
+       struct dc_stream_state* (*create_stream_for_sink)(struct dc_sink *dc_sink_data);
+       struct dc_plane_state* (*create_plane)(struct dc *dc);
+       enum dc_status (*add_stream_to_ctx)(struct dc *dc, struct dc_state *new_ctx, struct dc_stream_state *dc_stream);
+       bool (*add_plane_to_context)(const struct dc *dc, struct dc_stream_state *stream, struct dc_plane_state *plane_state, struct dc_state *context);
+       bool (*remove_plane_from_context)(const struct dc *dc, struct dc_stream_state *stream, struct dc_plane_state *plane_state, struct dc_state *context);
+       enum dc_status (*remove_stream_from_ctx)(struct dc *dc, struct dc_state *new_ctx, struct dc_stream_state *stream);
+       void (*plane_state_release)(struct dc_plane_state *plane_state);
+       void (*stream_release)(struct dc_stream_state *stream);
+       void (*release_dsc)(struct resource_context *res_ctx, const struct resource_pool *pool, struct display_stream_compressor **dsc);
+};
+
+struct dml2_clks_table_entry {
+       unsigned int dcfclk_mhz;
+       unsigned int fclk_mhz;
+       unsigned int memclk_mhz;
+       unsigned int socclk_mhz;
+       unsigned int dtbclk_mhz;
+       unsigned int dispclk_mhz;
+       unsigned int dppclk_mhz;
+};
+
+struct dml2_clks_num_entries {
+       unsigned int num_dcfclk_levels;
+       unsigned int num_fclk_levels;
+       unsigned int num_memclk_levels;
+       unsigned int num_socclk_levels;
+       unsigned int num_dtbclk_levels;
+       unsigned int num_dispclk_levels;
+       unsigned int num_dppclk_levels;
+};
+
+struct dml2_clks_limit_table {
+       struct dml2_clks_table_entry clk_entries[DML2_MAX_NUM_DPM_LVL];
+       struct dml2_clks_num_entries num_entries_per_clk;
+       unsigned int num_states;
+};
+
+// Various overrides, per ASIC or per SKU specific, or for debugging purpose when/if available
+struct dml2_soc_bbox_overrides {
+       double xtalclk_mhz;
+       double dchub_refclk_mhz;
+       double dprefclk_mhz;
+       double disp_pll_vco_speed_mhz;
+       double urgent_latency_us;
+       double sr_exit_latency_us;
+       double sr_enter_plus_exit_latency_us;
+       double dram_clock_change_latency_us;
+       double fclk_change_latency_us;
+       unsigned int dram_num_chan;
+       unsigned int dram_chanel_width_bytes;
+       struct dml2_clks_limit_table clks_table;
+};
+
+struct dml2_configuration_options {
+       int dcn_pipe_count;
+       bool use_native_pstate_optimization;
+       bool enable_windowed_mpo_odm;
+       bool use_native_soc_bb_construction;
+       bool skip_hw_state_mapping;
+       bool optimize_odm_4to1;
+       bool minimize_dispclk_using_odm;
+       struct dml2_dc_callbacks callbacks;
+       struct {
+               bool force_disable_subvp;
+               bool force_enable_subvp;
+               unsigned int subvp_fw_processing_delay_us;
+               unsigned int subvp_pstate_allow_width_us;
+               unsigned int subvp_prefetch_end_to_mall_start_us;
+               unsigned int subvp_swath_height_margin_lines;
+               struct dml2_dc_svp_callbacks callbacks;
+       } svp_pstate;
+       struct dml2_soc_mall_info mall_cfg;
+       struct dml2_soc_bbox_overrides bbox_overrides;
+       unsigned int max_segments_per_hubp;
+       unsigned int det_segment_size;
+};
+
+/*
+ * dml2_create - Creates dml2_context.
+ * @in_dc: dc.
+ * @config: dml2 configuration options.
+ * @dml2: Created dml2 context.
+ *
+ * Create and destroy of DML2 is done as part of dc_state creation
+ * and dc_state_free. DML2 IP, SOC and STATES are initialized at
+ * creation time.
+ *
+ * Return: True if dml2 is successfully created, false otherwise.
+ */
+bool dml2_create(const struct dc *in_dc,
+                                const struct dml2_configuration_options *config,
+                                struct dml2_context **dml2);
+
+void dml2_destroy(struct dml2_context *dml2);
+
+/*
+ * dml2_validate - Determines if a display configuration is supported or not.
+ * @in_dc: dc.
+ * @context: dc_state to be validated.
+ * @fast_validate: Fast validate will not populate context.res_ctx.
+ *
+ * DML1.0 compatible interface for validation.
+ *
+ * Based on fast_validate option internally would call:
+ *
+ * -dml2_validate_and_build_resource - for non fast_validate option
+ * Calculates if dc_state can be supported on the SOC, and attempts to
+ * optimize the power management feature supports versus minimum clocks.
+ * If supported, also builds out_new_hw_state to represent the hw programming
+ * for the new dc state.
+ *
+ * -dml2_validate_only - for fast_validate option
+ * Calculates if dc_state can be supported on the SOC (i.e. at maximum
+ * clocks) with all mandatory power features enabled.
+
+ * Context: Two threads may not invoke this function concurrently unless they reference
+ *          separate dc_states for validation.
+ * Return: True if mode is supported, false otherwise.
+ */
+bool dml2_validate(const struct dc *in_dc,
+                                  struct dc_state *context,
+                                  bool fast_validate);
+
+/*
+ * dml2_extract_dram_and_fclk_change_support - Extracts the FCLK and UCLK change support info.
+ * @dml2: input dml2 context pointer.
+ * @fclk_change_support: output pointer holding the fclk change support info (vactive, vblank, unsupported).
+ * @dram_clk_change_support: output pointer holding the uclk change support info (vactive, vblank, unsupported).
+ */
+void dml2_extract_dram_and_fclk_change_support(struct dml2_context *dml2,
+       unsigned int *fclk_change_support, unsigned int *dram_clk_change_support);
+
+#endif //_DML2_WRAPPER_H_
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml_assert.h b/drivers/gpu/drm/amd/display/dc/dml2/dml_assert.h
new file mode 100644 (file)
index 0000000..b5e463a
--- /dev/null
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DML_ASSERT_H__
+#define __DML_ASSERT_H__
+
+#include "os_types.h"
+
+#endif //__DML_ASSERT_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml_depedencies.h b/drivers/gpu/drm/amd/display/dc/dml2/dml_depedencies.h
new file mode 100644 (file)
index 0000000..d8c7f74
--- /dev/null
@@ -0,0 +1,31 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+/* This header intentinally does not include an #ifdef guard as it only contains includes for other headers*/
+
+/*
+ * Standard Types
+ */
+#include "os_types.h"
+#include "cmntypes.h"
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml_display_rq_dlg_calc.c b/drivers/gpu/drm/amd/display/dc/dml2/dml_display_rq_dlg_calc.c
new file mode 100644 (file)
index 0000000..377ef6d
--- /dev/null
@@ -0,0 +1,585 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "dml_display_rq_dlg_calc.h"
+#include "display_mode_core.h"
+#include "display_mode_util.h"
+
+static dml_bool_t is_dual_plane(enum dml_source_format_class source_format)
+{
+       dml_bool_t ret_val = 0;
+
+       if ((source_format == dml_420_12) || (source_format == dml_420_8) || (source_format == dml_420_10) || (source_format == dml_rgbe_alpha))
+               ret_val = 1;
+
+       return ret_val;
+}
+
+void dml_rq_dlg_get_rq_reg(dml_display_rq_regs_st                      *rq_regs,
+                                                       struct display_mode_lib_st       *mode_lib,
+                                                       const dml_uint_t                          pipe_idx)
+{
+       dml_uint_t plane_idx                                    = dml_get_plane_idx(mode_lib, pipe_idx);
+       enum dml_source_format_class source_format      = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[plane_idx];
+       enum dml_swizzle_mode     sw_mode                = mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[plane_idx];
+       dml_bool_t dual_plane                                   = is_dual_plane((enum dml_source_format_class)(source_format));
+
+       uint32 pixel_chunk_bytes = 0;
+       uint32 min_pixel_chunk_bytes = 0;
+       uint32 meta_chunk_bytes = 0;
+       uint32 min_meta_chunk_bytes = 0;
+       uint32 dpte_group_bytes = 0;
+       uint32 mpte_group_bytes = 0;
+
+       uint32 p1_pixel_chunk_bytes = 0;
+       uint32 p1_min_pixel_chunk_bytes = 0;
+       uint32 p1_meta_chunk_bytes = 0;
+       uint32 p1_min_meta_chunk_bytes = 0;
+       uint32 p1_dpte_group_bytes = 0;
+       uint32 p1_mpte_group_bytes = 0;
+
+       dml_uint_t detile_buf_size_in_bytes;
+       dml_uint_t detile_buf_plane1_addr = 0;
+
+       dml_float_t stored_swath_l_bytes;
+       dml_float_t stored_swath_c_bytes;
+       dml_bool_t      is_phantom_pipe;
+
+       dml_uint_t pte_row_height_linear;
+
+       dml_print("DML_DLG::%s: Calculation for pipe[%d] start\n", __func__, pipe_idx);
+
+       memset(rq_regs, 0, sizeof(*rq_regs));
+
+       pixel_chunk_bytes               = (dml_uint_t)(dml_get_pixel_chunk_size_in_kbyte(mode_lib) * 1024);
+       min_pixel_chunk_bytes   = (dml_uint_t)(dml_get_min_pixel_chunk_size_in_byte(mode_lib));
+
+       if (pixel_chunk_bytes == 64 * 1024)
+               min_pixel_chunk_bytes = 0;
+
+       meta_chunk_bytes                = (dml_uint_t)(dml_get_meta_chunk_size_in_kbyte(mode_lib) * 1024);
+       min_meta_chunk_bytes    = (dml_uint_t)(dml_get_min_meta_chunk_size_in_byte(mode_lib));
+
+       dpte_group_bytes = (dml_uint_t)(dml_get_dpte_group_size_in_bytes(mode_lib, pipe_idx));
+       mpte_group_bytes = (dml_uint_t)(dml_get_vm_group_size_in_bytes(mode_lib, pipe_idx));
+
+       p1_pixel_chunk_bytes            =  pixel_chunk_bytes;
+       p1_min_pixel_chunk_bytes        =  min_pixel_chunk_bytes;
+       p1_meta_chunk_bytes                     =  meta_chunk_bytes;
+       p1_min_meta_chunk_bytes         =  min_meta_chunk_bytes;
+       p1_dpte_group_bytes                     =  dpte_group_bytes;
+       p1_mpte_group_bytes                     =  mpte_group_bytes;
+
+       if (source_format == dml_rgbe_alpha)
+               p1_pixel_chunk_bytes = (dml_uint_t)(dml_get_alpha_pixel_chunk_size_in_kbyte(mode_lib) * 1024);
+
+       rq_regs->rq_regs_l.chunk_size = (dml_uint_t)(dml_log2((dml_float_t) pixel_chunk_bytes) - 10);
+       rq_regs->rq_regs_c.chunk_size = (dml_uint_t)(dml_log2((dml_float_t) p1_pixel_chunk_bytes) - 10);
+
+       if (min_pixel_chunk_bytes == 0)
+               rq_regs->rq_regs_l.min_chunk_size = 0;
+       else
+               rq_regs->rq_regs_l.min_chunk_size = (dml_uint_t)(dml_log2((dml_float_t) min_pixel_chunk_bytes) - 8 + 1);
+
+       if (p1_min_pixel_chunk_bytes == 0)
+               rq_regs->rq_regs_c.min_chunk_size = 0;
+       else
+               rq_regs->rq_regs_c.min_chunk_size = (dml_uint_t)(dml_log2((dml_float_t) p1_min_pixel_chunk_bytes) - 8 + 1);
+
+       rq_regs->rq_regs_l.meta_chunk_size = (dml_uint_t)(dml_log2((dml_float_t) meta_chunk_bytes) - 10);
+       rq_regs->rq_regs_c.meta_chunk_size = (dml_uint_t)(dml_log2((dml_float_t) p1_meta_chunk_bytes) - 10);
+
+       if (min_meta_chunk_bytes == 0)
+               rq_regs->rq_regs_l.min_meta_chunk_size = 0;
+       else
+               rq_regs->rq_regs_l.min_meta_chunk_size = (dml_uint_t)(dml_log2((dml_float_t) min_meta_chunk_bytes) - 6 + 1);
+
+       if (min_meta_chunk_bytes == 0)
+               rq_regs->rq_regs_c.min_meta_chunk_size = 0;
+       else
+               rq_regs->rq_regs_c.min_meta_chunk_size = (dml_uint_t)(dml_log2((dml_float_t) p1_min_meta_chunk_bytes) - 6 + 1);
+
+       rq_regs->rq_regs_l.dpte_group_size = (dml_uint_t)(dml_log2((dml_float_t) dpte_group_bytes) - 6);
+       rq_regs->rq_regs_l.mpte_group_size = (dml_uint_t)(dml_log2((dml_float_t) mpte_group_bytes) - 6);
+       rq_regs->rq_regs_c.dpte_group_size = (dml_uint_t)(dml_log2((dml_float_t) p1_dpte_group_bytes) - 6);
+       rq_regs->rq_regs_c.mpte_group_size = (dml_uint_t)(dml_log2((dml_float_t) p1_mpte_group_bytes) - 6);
+
+       detile_buf_size_in_bytes = (dml_uint_t)(dml_get_det_buffer_size_kbytes(mode_lib, pipe_idx) * 1024);
+
+       pte_row_height_linear = (dml_uint_t)(dml_get_dpte_row_height_linear_l(mode_lib, pipe_idx));
+
+       if (sw_mode == dml_sw_linear)
+               ASSERT(pte_row_height_linear >= 8);
+
+       rq_regs->rq_regs_l.pte_row_height_linear = (dml_uint_t)(dml_floor(dml_log2((dml_float_t) pte_row_height_linear), 1) - 3);
+
+       if (dual_plane) {
+               dml_uint_t p1_pte_row_height_linear = (dml_uint_t)(dml_get_dpte_row_height_linear_c(mode_lib, pipe_idx));
+               if (sw_mode == dml_sw_linear)
+                       ASSERT(p1_pte_row_height_linear >= 8);
+
+               rq_regs->rq_regs_c.pte_row_height_linear = (dml_uint_t)(dml_floor(dml_log2((dml_float_t) p1_pte_row_height_linear), 1) - 3);
+       }
+
+       rq_regs->rq_regs_l.swath_height = (dml_uint_t)(dml_log2((dml_float_t) dml_get_swath_height_l(mode_lib, pipe_idx)));
+       rq_regs->rq_regs_c.swath_height = (dml_uint_t)(dml_log2((dml_float_t) dml_get_swath_height_c(mode_lib, pipe_idx)));
+
+       if (pixel_chunk_bytes >= 32 * 1024 || (dual_plane && p1_pixel_chunk_bytes >= 32 * 1024)) { //32kb
+               rq_regs->drq_expansion_mode = 0;
+       } else {
+               rq_regs->drq_expansion_mode = 2;
+       }
+       rq_regs->prq_expansion_mode = 1;
+       rq_regs->mrq_expansion_mode = 1;
+       rq_regs->crq_expansion_mode = 1;
+
+       stored_swath_l_bytes = dml_get_det_stored_buffer_size_l_bytes(mode_lib, pipe_idx);
+       stored_swath_c_bytes = dml_get_det_stored_buffer_size_c_bytes(mode_lib, pipe_idx);
+       is_phantom_pipe          = dml_get_is_phantom_pipe(mode_lib, pipe_idx);
+
+       // Note: detile_buf_plane1_addr is in unit of 1KB
+       if (dual_plane) {
+               if (is_phantom_pipe) {
+                       detile_buf_plane1_addr = (dml_uint_t)((1024.0*1024.0) / 2.0 / 1024.0); // half to chroma
+               } else {
+                       if (stored_swath_l_bytes / stored_swath_c_bytes <= 1.5) {
+                               detile_buf_plane1_addr = (dml_uint_t)(detile_buf_size_in_bytes / 2.0 / 1024.0); // half to chroma
+#ifdef __DML_VBA_DEBUG__
+                               dml_print("DML_DLG: %s: detile_buf_plane1_addr = %d (1/2 to chroma)\n", __func__, detile_buf_plane1_addr);
+#endif
+                       } else {
+                               detile_buf_plane1_addr = (dml_uint_t)(dml_round_to_multiple((dml_uint_t)((2.0 * detile_buf_size_in_bytes) / 3.0), 1024, 0) / 1024.0); // 2/3 to luma
+#ifdef __DML_VBA_DEBUG__
+                               dml_print("DML_DLG: %s: detile_buf_plane1_addr = %d (1/3 chroma)\n", __func__, detile_buf_plane1_addr);
+#endif
+                       }
+               }
+       }
+       rq_regs->plane1_base_address = detile_buf_plane1_addr;
+
+#ifdef __DML_VBA_DEBUG__
+       dml_print("DML_DLG: %s: is_phantom_pipe = %d\n", __func__, is_phantom_pipe);
+       dml_print("DML_DLG: %s: stored_swath_l_bytes = %f\n", __func__, stored_swath_l_bytes);
+       dml_print("DML_DLG: %s: stored_swath_c_bytes = %f\n", __func__, stored_swath_c_bytes);
+       dml_print("DML_DLG: %s: detile_buf_size_in_bytes = %d\n", __func__, detile_buf_size_in_bytes);
+       dml_print("DML_DLG: %s: detile_buf_plane1_addr = %d\n", __func__, detile_buf_plane1_addr);
+       dml_print("DML_DLG: %s: plane1_base_address = %d\n", __func__, rq_regs->plane1_base_address);
+#endif
+       dml_print_rq_regs_st(rq_regs);
+       dml_print("DML_DLG::%s: Calculation for pipe[%d] done\n", __func__, pipe_idx);
+}
+
+// Note: currently taken in as is.
+// Nice to decouple code from hw register implement and extract code that are repeated for luma and chroma.
+
+
+void dml_rq_dlg_get_dlg_reg(dml_display_dlg_regs_st               *disp_dlg_regs,
+       dml_display_ttu_regs_st            *disp_ttu_regs,
+                                                       struct display_mode_lib_st *mode_lib,
+                                                       const dml_uint_t                        pipe_idx)
+{
+       dml_uint_t                                      plane_idx               = dml_get_plane_idx(mode_lib, pipe_idx);
+       enum dml_source_format_class    source_format   = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[plane_idx];
+       struct dml_timing_cfg_st                   *timing                 = &mode_lib->ms.cache_display_cfg.timing;
+       struct dml_plane_cfg_st                    *plane                  = &mode_lib->ms.cache_display_cfg.plane;
+       struct dml_hw_resource_st                  *hw                     = &mode_lib->ms.cache_display_cfg.hw;
+       dml_bool_t dual_plane                                           = is_dual_plane(source_format);
+       dml_uint_t num_cursors                                          = plane->NumberOfCursors[plane_idx];
+       enum dml_odm_mode                               odm_mode                = hw->ODMMode[plane_idx];
+
+       dml_uint_t        htotal                                = timing->HTotal[plane_idx];
+       dml_uint_t        hactive                               = timing->HActive[plane_idx];
+       dml_uint_t        hblank_end                    = timing->HBlankEnd[plane_idx];
+       dml_uint_t        vblank_end                    = timing->VBlankEnd[plane_idx];
+       dml_bool_t        interlaced                    = timing->Interlace[plane_idx];
+       dml_float_t   pclk_freq_in_mhz          = (dml_float_t) timing->PixelClock[plane_idx];
+       dml_float_t   refclk_freq_in_mhz        = (hw->DLGRefClkFreqMHz > 0) ? (dml_float_t) hw->DLGRefClkFreqMHz : mode_lib->soc.refclk_mhz;
+       dml_float_t   ref_freq_to_pix_freq      = refclk_freq_in_mhz / pclk_freq_in_mhz;
+
+       dml_uint_t vready_after_vcount0;
+
+       dml_uint_t dst_x_after_scaler;
+       dml_uint_t dst_y_after_scaler;
+
+       dml_float_t dst_y_prefetch;
+       dml_float_t dst_y_per_vm_vblank;
+       dml_float_t dst_y_per_row_vblank;
+       dml_float_t dst_y_per_vm_flip;
+       dml_float_t dst_y_per_row_flip;
+
+       dml_float_t max_dst_y_per_vm_vblank = 32.0;             //U5.2
+       dml_float_t max_dst_y_per_row_vblank = 16.0;    //U4.2
+
+       dml_float_t vratio_pre_l;
+       dml_float_t vratio_pre_c;
+
+       dml_float_t refcyc_per_line_delivery_pre_l;
+       dml_float_t refcyc_per_line_delivery_l;
+       dml_float_t refcyc_per_line_delivery_pre_c = 0.;
+       dml_float_t refcyc_per_line_delivery_c = 0.;
+       dml_float_t refcyc_per_req_delivery_pre_l;
+       dml_float_t refcyc_per_req_delivery_l;
+       dml_float_t refcyc_per_req_delivery_pre_c = 0.;
+       dml_float_t refcyc_per_req_delivery_c     = 0.;
+       dml_float_t refcyc_per_req_delivery_pre_cur0 = 0.;
+       dml_float_t refcyc_per_req_delivery_cur0 = 0.;
+
+       dml_float_t dst_y_per_pte_row_nom_l;
+       dml_float_t dst_y_per_pte_row_nom_c;
+       dml_float_t dst_y_per_meta_row_nom_l;
+       dml_float_t dst_y_per_meta_row_nom_c;
+       dml_float_t refcyc_per_pte_group_nom_l;
+       dml_float_t refcyc_per_pte_group_nom_c;
+       dml_float_t refcyc_per_pte_group_vblank_l;
+       dml_float_t refcyc_per_pte_group_vblank_c;
+       dml_float_t refcyc_per_pte_group_flip_l;
+       dml_float_t refcyc_per_pte_group_flip_c;
+       dml_float_t refcyc_per_meta_chunk_nom_l;
+       dml_float_t refcyc_per_meta_chunk_nom_c;
+       dml_float_t refcyc_per_meta_chunk_vblank_l;
+       dml_float_t refcyc_per_meta_chunk_vblank_c;
+       dml_float_t refcyc_per_meta_chunk_flip_l;
+       dml_float_t refcyc_per_meta_chunk_flip_c;
+
+       dml_float_t temp;
+       dml_float_t min_ttu_vblank;
+       dml_uint_t min_dst_y_next_start;
+
+       dml_print("DML_DLG::%s: Calculation for pipe_idx=%d\n", __func__, pipe_idx);
+       dml_print("DML_DLG::%s: plane_idx                               = %d\n", __func__, plane_idx);
+       dml_print("DML_DLG: %s: htotal                                  = %d\n", __func__, htotal);
+       dml_print("DML_DLG: %s: refclk_freq_in_mhz              = %3.2f\n", __func__, refclk_freq_in_mhz);
+       dml_print("DML_DLG: %s: hw->DLGRefClkFreqMHz    = %3.2f\n", __func__, hw->DLGRefClkFreqMHz);
+       dml_print("DML_DLG: %s: soc.refclk_mhz                  = %3.2f\n", __func__, mode_lib->soc.refclk_mhz);
+       dml_print("DML_DLG: %s: pclk_freq_in_mhz                = %3.2f\n", __func__, pclk_freq_in_mhz);
+       dml_print("DML_DLG: %s: ref_freq_to_pix_freq    = %3.2f\n", __func__, ref_freq_to_pix_freq);
+       dml_print("DML_DLG: %s: interlaced                              = %d\n", __func__, interlaced);
+
+       memset(disp_dlg_regs, 0, sizeof(*disp_dlg_regs));
+       memset(disp_ttu_regs, 0, sizeof(*disp_ttu_regs));
+
+       ASSERT(refclk_freq_in_mhz != 0);
+       ASSERT(pclk_freq_in_mhz != 0);
+       ASSERT(ref_freq_to_pix_freq < 4.0);
+
+       // Need to figure out which side of odm combine we're in
+       // Assume the pipe instance under the same plane is in order
+
+       if (odm_mode == dml_odm_mode_bypass) {
+               disp_dlg_regs->refcyc_h_blank_end = (dml_uint_t)((dml_float_t) hblank_end * ref_freq_to_pix_freq);
+       } else if (odm_mode == dml_odm_mode_combine_2to1 || odm_mode == dml_odm_mode_combine_4to1) {
+               // find out how many pipe are in this plane
+               dml_uint_t num_active_pipes                     = dml_get_num_active_pipes(&mode_lib->ms.cache_display_cfg);
+               dml_uint_t first_pipe_idx_in_plane      = __DML_NUM_PLANES__;
+               dml_uint_t pipe_idx_in_combine          = 0; // pipe index within the plane
+               dml_uint_t odm_combine_factor           = (odm_mode == dml_odm_mode_combine_2to1 ? 2 : 4);
+
+               for (dml_uint_t i = 0; i < num_active_pipes; i++) {
+                       if (dml_get_plane_idx(mode_lib, i) == plane_idx) {
+                               if (i < first_pipe_idx_in_plane) {
+                                       first_pipe_idx_in_plane = i;
+                               }
+                       }
+               }
+               pipe_idx_in_combine = pipe_idx - first_pipe_idx_in_plane; // DML assumes the pipes in the same plane will have continuous indexing (i.e. plane 0 use pipe 0, 1, and plane 1 uses pipe 2, 3, etc.)
+
+               disp_dlg_regs->refcyc_h_blank_end = (dml_uint_t)(((dml_float_t) hblank_end + (dml_float_t) pipe_idx_in_combine * (dml_float_t) hactive / (dml_float_t) odm_combine_factor) * ref_freq_to_pix_freq);
+               dml_print("DML_DLG: %s: pipe_idx = %d\n", __func__, pipe_idx);
+               dml_print("DML_DLG: %s: first_pipe_idx_in_plane = %d\n", __func__, first_pipe_idx_in_plane);
+               dml_print("DML_DLG: %s: pipe_idx_in_combine = %d\n", __func__, pipe_idx_in_combine);
+               dml_print("DML_DLG: %s: odm_combine_factor = %d\n", __func__, odm_combine_factor);
+       }
+       dml_print("DML_DLG: %s: refcyc_h_blank_end = %d\n", __func__, disp_dlg_regs->refcyc_h_blank_end);
+
+       ASSERT(disp_dlg_regs->refcyc_h_blank_end < (dml_uint_t)dml_pow(2, 13));
+
+       disp_dlg_regs->ref_freq_to_pix_freq = (dml_uint_t)(ref_freq_to_pix_freq * dml_pow(2, 19));
+       temp = dml_pow(2, 8);
+       disp_dlg_regs->refcyc_per_htotal = (dml_uint_t)(ref_freq_to_pix_freq * (dml_float_t)htotal * temp);
+       disp_dlg_regs->dlg_vblank_end = interlaced ? (vblank_end / 2) : vblank_end; // 15 bits
+
+       min_ttu_vblank          = dml_get_min_ttu_vblank_in_us(mode_lib, pipe_idx);
+       min_dst_y_next_start = (dml_uint_t)(dml_get_min_dst_y_next_start(mode_lib, pipe_idx));
+
+       dml_print("DML_DLG: %s: min_ttu_vblank (us)    = %3.2f\n", __func__, min_ttu_vblank);
+       dml_print("DML_DLG: %s: min_dst_y_next_start   = %d\n", __func__, min_dst_y_next_start);
+       dml_print("DML_DLG: %s: ref_freq_to_pix_freq   = %3.2f\n", __func__, ref_freq_to_pix_freq);
+
+       vready_after_vcount0            = (dml_uint_t)(dml_get_vready_at_or_after_vsync(mode_lib, pipe_idx));
+       disp_dlg_regs->vready_after_vcount0 = vready_after_vcount0;
+
+       dml_print("DML_DLG: %s: vready_after_vcount0 = %d\n", __func__, disp_dlg_regs->vready_after_vcount0);
+
+       dst_x_after_scaler = (dml_uint_t)(dml_get_dst_x_after_scaler(mode_lib, pipe_idx));
+       dst_y_after_scaler = (dml_uint_t)(dml_get_dst_y_after_scaler(mode_lib, pipe_idx));
+
+       dml_print("DML_DLG: %s: dst_x_after_scaler         = %d\n", __func__, dst_x_after_scaler);
+       dml_print("DML_DLG: %s: dst_y_after_scaler         = %d\n", __func__, dst_y_after_scaler);
+
+       dst_y_prefetch                  = dml_get_dst_y_prefetch(mode_lib, pipe_idx);
+       dst_y_per_vm_vblank             = dml_get_dst_y_per_vm_vblank(mode_lib, pipe_idx);
+       dst_y_per_row_vblank    = dml_get_dst_y_per_row_vblank(mode_lib, pipe_idx);
+       dst_y_per_vm_flip               = dml_get_dst_y_per_vm_flip(mode_lib, pipe_idx);
+       dst_y_per_row_flip              = dml_get_dst_y_per_row_flip(mode_lib, pipe_idx);
+
+       // magic!
+       if (htotal <= 75) {
+               max_dst_y_per_vm_vblank = 100.0;
+               max_dst_y_per_row_vblank = 100.0;
+       }
+
+       dml_print("DML_DLG: %s: dst_y_prefetch (after rnd) = %3.2f\n", __func__, dst_y_prefetch);
+       dml_print("DML_DLG: %s: dst_y_per_vm_flip        = %3.2f\n", __func__, dst_y_per_vm_flip);
+       dml_print("DML_DLG: %s: dst_y_per_row_flip       = %3.2f\n", __func__, dst_y_per_row_flip);
+       dml_print("DML_DLG: %s: dst_y_per_vm_vblank  = %3.2f\n", __func__, dst_y_per_vm_vblank);
+       dml_print("DML_DLG: %s: dst_y_per_row_vblank = %3.2f\n", __func__, dst_y_per_row_vblank);
+
+       ASSERT(dst_y_per_vm_vblank < max_dst_y_per_vm_vblank);
+       ASSERT(dst_y_per_row_vblank < max_dst_y_per_row_vblank);
+       ASSERT(dst_y_prefetch > (dst_y_per_vm_vblank + dst_y_per_row_vblank));
+
+       vratio_pre_l = dml_get_vratio_prefetch_l(mode_lib, pipe_idx);
+       vratio_pre_c = dml_get_vratio_prefetch_c(mode_lib, pipe_idx);
+
+       dml_print("DML_DLG: %s: vratio_pre_l = %3.2f\n", __func__, vratio_pre_l);
+       dml_print("DML_DLG: %s: vratio_pre_c = %3.2f\n", __func__, vratio_pre_c);
+
+       // Active
+       refcyc_per_line_delivery_pre_l = dml_get_refcyc_per_line_delivery_pre_l_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+       refcyc_per_line_delivery_l         = dml_get_refcyc_per_line_delivery_l_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+
+       dml_print("DML_DLG: %s: refcyc_per_line_delivery_pre_l = %3.2f\n", __func__, refcyc_per_line_delivery_pre_l);
+       dml_print("DML_DLG: %s: refcyc_per_line_delivery_l         = %3.2f\n", __func__, refcyc_per_line_delivery_l);
+
+       if (dual_plane) {
+               refcyc_per_line_delivery_pre_c = dml_get_refcyc_per_line_delivery_pre_c_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+               refcyc_per_line_delivery_c         = dml_get_refcyc_per_line_delivery_c_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+
+               dml_print("DML_DLG: %s: refcyc_per_line_delivery_pre_c = %3.2f\n", __func__, refcyc_per_line_delivery_pre_c);
+               dml_print("DML_DLG: %s: refcyc_per_line_delivery_c         = %3.2f\n", __func__, refcyc_per_line_delivery_c);
+       }
+
+       disp_dlg_regs->refcyc_per_vm_dmdata = (dml_uint_t)(dml_get_refcyc_per_vm_dmdata_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz);
+       disp_dlg_regs->dmdata_dl_delta = (dml_uint_t)(dml_get_dmdata_dl_delta_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz);
+
+       refcyc_per_req_delivery_pre_l = dml_get_refcyc_per_req_delivery_pre_l_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+       refcyc_per_req_delivery_l         = dml_get_refcyc_per_req_delivery_l_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+
+       dml_print("DML_DLG: %s: refcyc_per_req_delivery_pre_l = %3.2f\n", __func__, refcyc_per_req_delivery_pre_l);
+       dml_print("DML_DLG: %s: refcyc_per_req_delivery_l         = %3.2f\n", __func__, refcyc_per_req_delivery_l);
+
+       if (dual_plane) {
+               refcyc_per_req_delivery_pre_c = dml_get_refcyc_per_req_delivery_pre_c_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+               refcyc_per_req_delivery_c         = dml_get_refcyc_per_req_delivery_c_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+
+               dml_print("DML_DLG: %s: refcyc_per_req_delivery_pre_c = %3.2f\n", __func__, refcyc_per_req_delivery_pre_c);
+               dml_print("DML_DLG: %s: refcyc_per_req_delivery_c         = %3.2f\n", __func__, refcyc_per_req_delivery_c);
+       }
+
+       // TTU - Cursor
+       ASSERT(num_cursors <= 1);
+       if (num_cursors > 0) {
+               refcyc_per_req_delivery_pre_cur0 = dml_get_refcyc_per_cursor_req_delivery_pre_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+               refcyc_per_req_delivery_cur0     = dml_get_refcyc_per_cursor_req_delivery_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+
+               dml_print("DML_DLG: %s: refcyc_per_req_delivery_pre_cur0 = %3.2f\n", __func__, refcyc_per_req_delivery_pre_cur0);
+               dml_print("DML_DLG: %s: refcyc_per_req_delivery_cur0     = %3.2f\n", __func__, refcyc_per_req_delivery_cur0);
+       }
+
+       // Assign to register structures
+       disp_dlg_regs->min_dst_y_next_start = (dml_uint_t)((dml_float_t) min_dst_y_next_start * dml_pow(2, 2));
+       ASSERT(disp_dlg_regs->min_dst_y_next_start < (dml_uint_t)dml_pow(2, 18));
+
+       disp_dlg_regs->dst_y_after_scaler = dst_y_after_scaler; // in terms of line
+       disp_dlg_regs->refcyc_x_after_scaler = (dml_uint_t)((dml_float_t) dst_x_after_scaler * ref_freq_to_pix_freq); // in terms of refclk
+       disp_dlg_regs->dst_y_prefetch                   = (dml_uint_t)(dst_y_prefetch * dml_pow(2, 2));
+       disp_dlg_regs->dst_y_per_vm_vblank              = (dml_uint_t)(dst_y_per_vm_vblank * dml_pow(2, 2));
+       disp_dlg_regs->dst_y_per_row_vblank             = (dml_uint_t)(dst_y_per_row_vblank * dml_pow(2, 2));
+       disp_dlg_regs->dst_y_per_vm_flip                = (dml_uint_t)(dst_y_per_vm_flip * dml_pow(2, 2));
+       disp_dlg_regs->dst_y_per_row_flip               = (dml_uint_t)(dst_y_per_row_flip * dml_pow(2, 2));
+
+       disp_dlg_regs->vratio_prefetch = (dml_uint_t)(vratio_pre_l * dml_pow(2, 19));
+       disp_dlg_regs->vratio_prefetch_c = (dml_uint_t)(vratio_pre_c * dml_pow(2, 19));
+
+       dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_vm_vblank      = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_vm_vblank);
+       dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_vblank = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_vblank);
+       dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_vm_flip        = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_vm_flip);
+       dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_flip       = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_flip);
+
+       // hack for FPGA
+       /* NOTE: We dont have getenv defined in driver and it does not make any sense in the driver */
+       /*char* fpga_env = getenv("FPGA_FPDIV");
+       if(fpga_env !=NULL)
+       {
+               if(disp_dlg_regs->vratio_prefetch >= (dml_uint_t)dml_pow(2, 22))
+               {
+                       disp_dlg_regs->vratio_prefetch = (dml_uint_t)dml_pow(2, 22)-1;
+                       dml_print("FPGA msg: vratio_prefetch exceed the max value, the register field is [21:0]\n");
+               }
+       }*/
+
+       disp_dlg_regs->refcyc_per_vm_group_vblank               = (dml_uint_t)(dml_get_refcyc_per_vm_group_vblank_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz);
+       disp_dlg_regs->refcyc_per_vm_group_flip                 = (dml_uint_t)(dml_get_refcyc_per_vm_group_flip_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz);
+       disp_dlg_regs->refcyc_per_vm_req_vblank                 = (dml_uint_t)(dml_get_refcyc_per_vm_req_vblank_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz * dml_pow(2, 10));
+       disp_dlg_regs->refcyc_per_vm_req_flip                   = (dml_uint_t)(dml_get_refcyc_per_vm_req_flip_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz * dml_pow(2, 10));
+
+       dst_y_per_pte_row_nom_l = dml_get_dst_y_per_pte_row_nom_l(mode_lib, pipe_idx);
+       dst_y_per_pte_row_nom_c = dml_get_dst_y_per_pte_row_nom_c(mode_lib, pipe_idx);
+       dst_y_per_meta_row_nom_l = dml_get_dst_y_per_meta_row_nom_l(mode_lib, pipe_idx);
+       dst_y_per_meta_row_nom_c = dml_get_dst_y_per_meta_row_nom_c(mode_lib, pipe_idx);
+
+       refcyc_per_pte_group_nom_l              = dml_get_refcyc_per_pte_group_nom_l_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+       refcyc_per_pte_group_nom_c              = dml_get_refcyc_per_pte_group_nom_c_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+       refcyc_per_pte_group_vblank_l   = dml_get_refcyc_per_pte_group_vblank_l_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+       refcyc_per_pte_group_vblank_c   = dml_get_refcyc_per_pte_group_vblank_c_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+       refcyc_per_pte_group_flip_l             = dml_get_refcyc_per_pte_group_flip_l_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+       refcyc_per_pte_group_flip_c             = dml_get_refcyc_per_pte_group_flip_c_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+
+       refcyc_per_meta_chunk_nom_l             = dml_get_refcyc_per_meta_chunk_nom_l_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+       refcyc_per_meta_chunk_nom_c             = dml_get_refcyc_per_meta_chunk_nom_c_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+       refcyc_per_meta_chunk_vblank_l  = dml_get_refcyc_per_meta_chunk_vblank_l_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+       refcyc_per_meta_chunk_vblank_c  = dml_get_refcyc_per_meta_chunk_vblank_c_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+       refcyc_per_meta_chunk_flip_l    = dml_get_refcyc_per_meta_chunk_flip_l_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+       refcyc_per_meta_chunk_flip_c    = dml_get_refcyc_per_meta_chunk_flip_c_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz;
+
+       disp_dlg_regs->dst_y_per_pte_row_nom_l                  = (dml_uint_t)(dst_y_per_pte_row_nom_l * dml_pow(2, 2));
+       disp_dlg_regs->dst_y_per_pte_row_nom_c                  = (dml_uint_t)(dst_y_per_pte_row_nom_c * dml_pow(2, 2));
+       disp_dlg_regs->dst_y_per_meta_row_nom_l                 = (dml_uint_t)(dst_y_per_meta_row_nom_l * dml_pow(2, 2));
+       disp_dlg_regs->dst_y_per_meta_row_nom_c                 = (dml_uint_t)(dst_y_per_meta_row_nom_c * dml_pow(2, 2));
+       disp_dlg_regs->refcyc_per_pte_group_nom_l               = (dml_uint_t)(refcyc_per_pte_group_nom_l);
+       disp_dlg_regs->refcyc_per_pte_group_nom_c               = (dml_uint_t)(refcyc_per_pte_group_nom_c);
+       disp_dlg_regs->refcyc_per_pte_group_vblank_l    = (dml_uint_t)(refcyc_per_pte_group_vblank_l);
+       disp_dlg_regs->refcyc_per_pte_group_vblank_c    = (dml_uint_t)(refcyc_per_pte_group_vblank_c);
+       disp_dlg_regs->refcyc_per_pte_group_flip_l              = (dml_uint_t)(refcyc_per_pte_group_flip_l);
+       disp_dlg_regs->refcyc_per_pte_group_flip_c              = (dml_uint_t)(refcyc_per_pte_group_flip_c);
+       disp_dlg_regs->refcyc_per_meta_chunk_nom_l              = (dml_uint_t)(refcyc_per_meta_chunk_nom_l);
+       disp_dlg_regs->refcyc_per_meta_chunk_nom_c              = (dml_uint_t)(refcyc_per_meta_chunk_nom_c);
+       disp_dlg_regs->refcyc_per_meta_chunk_vblank_l   = (dml_uint_t)(refcyc_per_meta_chunk_vblank_l);
+       disp_dlg_regs->refcyc_per_meta_chunk_vblank_c   = (dml_uint_t)(refcyc_per_meta_chunk_vblank_c);
+       disp_dlg_regs->refcyc_per_meta_chunk_flip_l             = (dml_uint_t)(refcyc_per_meta_chunk_flip_l);
+       disp_dlg_regs->refcyc_per_meta_chunk_flip_c             = (dml_uint_t)(refcyc_per_meta_chunk_flip_c);
+       disp_dlg_regs->refcyc_per_line_delivery_pre_l   = (dml_uint_t)dml_floor(refcyc_per_line_delivery_pre_l, 1);
+       disp_dlg_regs->refcyc_per_line_delivery_l               = (dml_uint_t)dml_floor(refcyc_per_line_delivery_l, 1);
+       disp_dlg_regs->refcyc_per_line_delivery_pre_c   = (dml_uint_t)dml_floor(refcyc_per_line_delivery_pre_c, 1);
+       disp_dlg_regs->refcyc_per_line_delivery_c               = (dml_uint_t)dml_floor(refcyc_per_line_delivery_c, 1);
+
+       disp_dlg_regs->chunk_hdl_adjust_cur0    = 3;
+       disp_dlg_regs->dst_y_offset_cur0                = 0;
+       disp_dlg_regs->chunk_hdl_adjust_cur1    = 3;
+       disp_dlg_regs->dst_y_offset_cur1                = 0;
+
+       disp_dlg_regs->dst_y_delta_drq_limit = 0x7fff; // off
+
+       disp_ttu_regs->refcyc_per_req_delivery_pre_l    = (dml_uint_t)(refcyc_per_req_delivery_pre_l      * dml_pow(2, 10));
+       disp_ttu_regs->refcyc_per_req_delivery_l                = (dml_uint_t)(refcyc_per_req_delivery_l                  * dml_pow(2, 10));
+       disp_ttu_regs->refcyc_per_req_delivery_pre_c    = (dml_uint_t)(refcyc_per_req_delivery_pre_c      * dml_pow(2, 10));
+       disp_ttu_regs->refcyc_per_req_delivery_c                = (dml_uint_t)(refcyc_per_req_delivery_c                  * dml_pow(2, 10));
+       disp_ttu_regs->refcyc_per_req_delivery_pre_cur0 = (dml_uint_t)(refcyc_per_req_delivery_pre_cur0   * dml_pow(2, 10));
+       disp_ttu_regs->refcyc_per_req_delivery_cur0             = (dml_uint_t)(refcyc_per_req_delivery_cur0               * dml_pow(2, 10));
+       disp_ttu_regs->refcyc_per_req_delivery_pre_cur1 = 0;
+       disp_ttu_regs->refcyc_per_req_delivery_cur1             = 0;
+       disp_ttu_regs->qos_level_low_wm = 0;
+
+       disp_ttu_regs->qos_level_high_wm = (dml_uint_t)(4.0 * (dml_float_t)htotal * ref_freq_to_pix_freq);
+
+       disp_ttu_regs->qos_level_flip = 14;
+       disp_ttu_regs->qos_level_fixed_l = 8;
+       disp_ttu_regs->qos_level_fixed_c = 8;
+       disp_ttu_regs->qos_level_fixed_cur0 = 8;
+       disp_ttu_regs->qos_ramp_disable_l = 0;
+       disp_ttu_regs->qos_ramp_disable_c = 0;
+       disp_ttu_regs->qos_ramp_disable_cur0 = 0;
+       disp_ttu_regs->min_ttu_vblank = (dml_uint_t)(min_ttu_vblank * refclk_freq_in_mhz);
+
+       // CHECK for HW registers' range, assert or clamp
+       ASSERT(refcyc_per_req_delivery_pre_l < dml_pow(2, 13));
+       ASSERT(refcyc_per_req_delivery_l < dml_pow(2, 13));
+       ASSERT(refcyc_per_req_delivery_pre_c < dml_pow(2, 13));
+       ASSERT(refcyc_per_req_delivery_c < dml_pow(2, 13));
+       if (disp_dlg_regs->refcyc_per_vm_group_vblank >= (dml_uint_t)dml_pow(2, 23))
+               disp_dlg_regs->refcyc_per_vm_group_vblank = (dml_uint_t)(dml_pow(2, 23) - 1);
+
+       if (disp_dlg_regs->refcyc_per_vm_group_flip >= (dml_uint_t)dml_pow(2, 23))
+               disp_dlg_regs->refcyc_per_vm_group_flip = (dml_uint_t)(dml_pow(2, 23) - 1);
+
+       if (disp_dlg_regs->refcyc_per_vm_req_vblank >= (dml_uint_t)dml_pow(2, 23))
+               disp_dlg_regs->refcyc_per_vm_req_vblank = (dml_uint_t)(dml_pow(2, 23) - 1);
+
+       if (disp_dlg_regs->refcyc_per_vm_req_flip >= (dml_uint_t)dml_pow(2, 23))
+               disp_dlg_regs->refcyc_per_vm_req_flip = (dml_uint_t)(dml_pow(2, 23) - 1);
+
+
+       ASSERT(disp_dlg_regs->dst_y_after_scaler < (dml_uint_t)8);
+       ASSERT(disp_dlg_regs->refcyc_x_after_scaler < (dml_uint_t)dml_pow(2, 13));
+       ASSERT(disp_dlg_regs->dst_y_per_pte_row_nom_l < (dml_uint_t)dml_pow(2, 17));
+       if (dual_plane) {
+               if (disp_dlg_regs->dst_y_per_pte_row_nom_c >= (dml_uint_t)dml_pow(2, 17)) { // FIXME what so special about chroma, can we just assert?
+                       dml_print("DML_DLG: %s: Warning dst_y_per_pte_row_nom_c %u > register max U15.2 %u\n", __func__, disp_dlg_regs->dst_y_per_pte_row_nom_c, (dml_uint_t)dml_pow(2, 17) - 1);
+               }
+       }
+       ASSERT(disp_dlg_regs->dst_y_per_meta_row_nom_l < (dml_uint_t)dml_pow(2, 17));
+       ASSERT(disp_dlg_regs->dst_y_per_meta_row_nom_c < (dml_uint_t)dml_pow(2, 17));
+
+       if (disp_dlg_regs->refcyc_per_pte_group_nom_l >= (dml_uint_t)dml_pow(2, 23))
+               disp_dlg_regs->refcyc_per_pte_group_nom_l = (dml_uint_t)(dml_pow(2, 23) - 1);
+       if (dual_plane) {
+               if (disp_dlg_regs->refcyc_per_pte_group_nom_c >= (dml_uint_t)dml_pow(2, 23))
+                       disp_dlg_regs->refcyc_per_pte_group_nom_c = (dml_uint_t)(dml_pow(2, 23) - 1);
+       }
+       ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_l < (dml_uint_t)dml_pow(2, 13));
+       if (dual_plane) {
+               ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_c < (dml_uint_t)dml_pow(2, 13));
+       }
+
+       if (disp_dlg_regs->refcyc_per_meta_chunk_nom_l >= (dml_uint_t)dml_pow(2, 23))
+               disp_dlg_regs->refcyc_per_meta_chunk_nom_l = (dml_uint_t)(dml_pow(2, 23) - 1);
+       if (dual_plane) {
+               if (disp_dlg_regs->refcyc_per_meta_chunk_nom_c >= (dml_uint_t)dml_pow(2, 23))
+                       disp_dlg_regs->refcyc_per_meta_chunk_nom_c = (dml_uint_t)(dml_pow(2, 23) - 1);
+       }
+       ASSERT(disp_dlg_regs->refcyc_per_meta_chunk_vblank_l    < (dml_uint_t)dml_pow(2, 13));
+       ASSERT(disp_dlg_regs->refcyc_per_meta_chunk_vblank_c    < (dml_uint_t)dml_pow(2, 13));
+       ASSERT(disp_dlg_regs->refcyc_per_line_delivery_pre_l    < (dml_uint_t)dml_pow(2, 13));
+       ASSERT(disp_dlg_regs->refcyc_per_line_delivery_l                < (dml_uint_t)dml_pow(2, 13));
+       ASSERT(disp_dlg_regs->refcyc_per_line_delivery_pre_c    < (dml_uint_t)dml_pow(2, 13));
+       ASSERT(disp_dlg_regs->refcyc_per_line_delivery_c                < (dml_uint_t)dml_pow(2, 13));
+       ASSERT(disp_ttu_regs->qos_level_low_wm                                  < (dml_uint_t) dml_pow(2, 14));
+       ASSERT(disp_ttu_regs->qos_level_high_wm                                 < (dml_uint_t) dml_pow(2, 14));
+       ASSERT(disp_ttu_regs->min_ttu_vblank                                    < (dml_uint_t) dml_pow(2, 24));
+
+       dml_print_ttu_regs_st(disp_ttu_regs);
+       dml_print_dlg_regs_st(disp_dlg_regs);
+       dml_print("DML_DLG::%s: Calculation for pipe[%d] done\n", __func__, pipe_idx);
+}
+
+void dml_rq_dlg_get_arb_params(struct display_mode_lib_st *mode_lib, dml_display_arb_params_st *arb_param)
+{
+       memset(arb_param, 0, sizeof(*arb_param));
+       arb_param->max_req_outstanding = 256;
+       arb_param->min_req_outstanding = 256; // turn off the sat level feature if this set to max
+       arb_param->sat_level_us = 60;
+       arb_param->hvm_max_qos_commit_threshold = 0xf;
+       arb_param->hvm_min_req_outstand_commit_threshold = 0xa;
+       arb_param->compbuf_reserved_space_kbytes = 2 * 8;  // assume max data chunk size of 8K
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml_display_rq_dlg_calc.h b/drivers/gpu/drm/amd/display/dc/dml2/dml_display_rq_dlg_calc.h
new file mode 100644 (file)
index 0000000..bf491cf
--- /dev/null
@@ -0,0 +1,63 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DML_DISPLAY_RQ_DLG_CALC_H__
+#define __DML_DISPLAY_RQ_DLG_CALC_H__
+
+#include "display_mode_core_structs.h"
+#include "display_mode_lib_defines.h"
+
+struct display_mode_lib_st;
+
+// Function: dml_rq_dlg_get_rq_reg
+//  Main entry point for test to get the register values out of this DML class.
+//  This function calls <get_rq_param> and <extract_rq_regs> fucntions to calculate
+//  and then populate the rq_regs struct
+// Input:
+//  Assume mode_program is already called
+// Output:
+//  rq_regs - struct that holds all the RQ registers field value.
+//            See also: <display_rq_regs_st>
+
+void dml_rq_dlg_get_rq_reg(dml_display_rq_regs_st *rq_regs,
+                       struct display_mode_lib_st *mode_lib,
+                       const dml_uint_t pipe_idx);
+
+// Function: dml_rq_dlg_get_dlg_reg
+//   Calculate and return DLG and TTU register struct given the system setting
+// Output:
+//  dlg_regs - output DLG register struct
+//  ttu_regs - output DLG TTU register struct
+// Input:
+//  Assume mode_program is already called
+//  pipe_idx - index that identifies the e2e_pipe_param that corresponding to this dlg
+void dml_rq_dlg_get_dlg_reg(dml_display_dlg_regs_st        *dlg_regs,
+                               dml_display_ttu_regs_st        *ttu_regs,
+                               struct display_mode_lib_st        *mode_lib,
+                               const dml_uint_t            pipe_idx);
+
+// Function: dml_rq_dlg_get_arb_params
+void dml_rq_dlg_get_arb_params(struct display_mode_lib_st *mode_lib, dml_display_arb_params_st *arb_param);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml_logging.h b/drivers/gpu/drm/amd/display/dc/dml2/dml_logging.h
new file mode 100644 (file)
index 0000000..890c0a0
--- /dev/null
@@ -0,0 +1,29 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __DML_LOGGING_H__
+#define __DML_LOGGING_H__
+
+#define dml_print(...) ((void)0)
+
+#endif //__DML_LOGGING_H__
index c095ccb..2e11586 100644 (file)
@@ -491,6 +491,7 @@ union bw_output {
 struct bw_context {
        union bw_output bw;
        struct display_mode_lib dml;
+       struct dml2_context *dml2;
 };
 
 struct dc_dmub_cmd {