drm/amd/display: Add more DC HW state info to underflow logging

[Why]
Debugging underflow issues frequently requires knowing the
HW state at the time of underflow. To enable capturing this
HW state information, interface functions are needed for the
various DC HW blocks.

[How]
This change adds the interface functions to read HW state for
the following DC HW blocks:
- HUBBUB
- HUBP
- DPP
- MPC
- OPP
- DSC
- OPTC
- DCCG

Reviewed-by: George Shen <george.shen@amd.com>
Signed-off-by: Karen Chen <Karen.Chen@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Rename dml2 to dml2_0 folder

[Why]
dml2 folder contains all logic for all versions of DML2
This is currently DML2.0 and DML2.1.
Rename dml2 to dml2_0 folder to reflect this better (dml2_0
for DML2.0).

[How]
Rename dml2 to dml2_0 folder and update dml2 references to
use dml2_0 folder.

Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Austin Zheng <Austin.Zheng@amd.com>
Signed-off-by: waynelin <Wayne.Lin@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix notification of vtotal to DMU for cursor offload

[Why]
It was placed after the early return and the notification is never sent.

[How]
Place it after .set_drr and before the return.

Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix wrong index for DCN401 cursor offload

[Why]
Payloads are ignored because the wrong index is written as part of the
pipe update implementation for DCN401.

[How]
Align it to the DCN35 implementation and ensure the + 1 is added.

Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add opp count validation to dml2.1

Newer asics can have mismatching dpp and opp counts and
dml needs to account for this.

Reviewed-by: Charlene Liu <charlene.liu@amd.com>
Signed-off-by: Dmytro Laktyushkin <dmytro.laktyushkin@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix DMUB reset sequence for DCN32

[WHY&HOW]
Backport reset sequence fixes implemented on DCN401 to DCN32 to address
stability issues when resetting the DMUB.

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Bump minimum for frame_warn_limit

[Why]
The bigger of CONFIG_FRAME_WARN and frame_warn_limit is used to
trigger warnings about large stack frames.  The dml_core_mode_support()
stack frame has grown to 2056.

[How]
Update frame_warn_limit to 2056 so that CONFIG_FRAME_WARN of 2048 doesn't
cause a failure.

Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4609
Reviewed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Mario Limonciello <superm1@kernel.org>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Re-introduce property to control adaptive backlight modulation

commit 0887054d14ae ("drm/amd: Drop abm_level property") dropped the
abm level property in favor of sysfs control. Since then there have
been discussions that compositors showed an interest in modifying
a vendor specific property instead.

So re-introduce the abm level property, but with different semantics.
Rather than being an integer it's now an enum. One of the enum options
is 'sysfs', and that is because there is still a sysfs file for use by
userspace when the compositor doesn't support this property.

If usespace has not modified this property, the default value will
be for sysfs to control it. Once userspace has set the property stop
allowing sysfs control.

The property is only attached to non-OLED eDP panels.

Cc: Xaver Hugl <xaver.hugl@kde.org>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Fix pointer casts when reading dynamic region sizes

The function amdgpu_virt_get_dynamic_data_info() writes a 64-bit size
value.  In two places (amdgpu_bios.c and amdgpu_discovery.c), the code
passed the address of a smaller variable by casting it to u64 *, which
is unsafe.

This could make the function write more bytes than the smaller variable
can hold, possibly overwriting nearby memory. Reported by static
analysis tools.

v2: Dynamic region size comes from the host (SR-IOV setup) and is always
fixed to 5 MB. (Lijo/Ellen)

5 MB easily fits inside a 32-bit value, so using a 64-bit type is not
needed. It also avoids extra type casts

Fixes: b4a8fcc7826a ("drm/amdgpu: Add logic for VF ipd and VF bios to init from dynamic crit_region offsets")
Reported by: Dan Carpenter <dan.carpenter@linaro.org>
Cc: Ellen Pan <yunru.pan@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: null check for hmm_pfns ptr before freeing it

Due to low memory or when num of pages is too big to be
accomodated, allocation could fail for pfn's.

Chekc hmm_pfns for NULL before calling the kvfree for the it.

Signed-off-by: Sunil Khatri <sunil.khatri@amd.com>
Acked-by: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: smu13: Enable VCN_RESET for pgm 7 with appropriate firmware version

This patch extends the VCN_RESET capability check to include pgm 7 when the firmware version is 0x07551400 or newer.

Signed-off-by: Jesse Zhang <Jesse.Zhang@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Make SR-IOV critical region checks overflow-safe

The function amdgpu_virt_init_critical_region() contained an invalid
check for a negative init_hdr_offset value:

    if (init_hdr_offset < 0)

Since init_hdr_offset is an unsigned 32-bit integer, this condition can
never be true and triggers a Smatch warning:

    warn: unsigned 'init_hdr_offset' is never less than zero

In addition, the subsequent bounds check: if ((init_hdr_offset +
init_hdr_size) > vram_size) was vulnerable to integer overflow when
adding the two unsigned values.  Thus, by promoting offset and size to
64-bit and using check_add_overflow() to safely validate the sum against
VRAM size.

Fixes: 07009df6494d ("drm/amdgpu: Introduce SRIOV critical regions v2 during VF init")
Reported by: Dan Carpenter <dan.carpenter@linaro.org>
Cc: Ellen Pan <yunru.pan@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Ellen Pan <yunru.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix SPDX header on cyan_skillfish_reg_init.c

This should be MIT.  The driver in general is MIT and
the license text at the top of the file is MIT so fix
it.

Fixes: e8529dbc75ca ("drm/amdgpu: add ip offset support for cyan skillfish")
Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4654
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix SPDX header on irqsrcs_vcn_5_0.h

This should be MIT.  The driver in general is MIT and
the license text at the top of the file is MIT so fix
it.

Fixes: d1bb64651095 ("drm/amdgpu: add irq source ids for VCN5_0/JPEG5_0")
Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4654
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix SPDX header on amd_cper.h

This should be MIT.  The driver in general is MIT and
the license text at the top of the file is MIT so fix
it.

Fixes: 523b69c65445 ("drm/amd/include: Add amd cper header")
Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4654
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix SPDX headers on amdgpu_cper.c/h

These should be MIT.  The driver in general is MIT and
the license text at the top of the files is MIT so fix
it.

Fixes: 92d5d2a09de1 ("drm/amdgpu: Introduce funcs for populating CPER")
Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4654
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/userqueue: Fix use after free in amdgpu_userq_buffer_vas_list_cleanup()

The amdgpu_userq_buffer_va_list_del() function frees "va_cursor" but it
is dereferenced on the next line when we print the debug message.  Print
the debug message first and then free it.

Fixes: 2a28f9665dca ("drm/amdgpu: track the userq bo va for its obj management")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm/powerplay/smumgr: Fix PCIeBootLinkLevel value on Iceland

Previously this was initialized with zero which represented PCIe Gen
1.0 instead of using the
maximum value from the speed table which is the behaviour of all other
smumgr implementations.

Fixes: 18aafc59b106 ("drm/amd/powerplay: implement fw related smu interface for iceland.")
Signed-off-by: John Smith <itistotalbotnet@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm/powerplay/smumgr: Fix PCIeBootLinkLevel value on Fiji

Previously this was initialized with zero which represented PCIe Gen
1.0 instead of using the
maximum value from the speed table which is the behaviour of all other
smumgr implementations.

Fixes: 18edef19ea44 ("drm/amd/powerplay: implement fw image related smu interface for Fiji.")
Signed-off-by: John Smith <itistotalbotnet@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: fix smu table id bound check issue in smu_cmn_update_table()

'table_index' is a variable defined by the smu driver (kmd)
'table_id' is a variable defined by the hw smu (pmfw)

This code should use table_index as a bounds check.

Fixes: caad2613dc4bd ("drm/amd/powerplay: move table setting common code to smu_cmn.c")
Signed-off-by: Yang Wang <kevinyang.wang@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Don't program BLNDGAM_MEM_PWR_FORCE when CM low-power is disabled on DCN30

Before commit 33056a97ae5e ("drm/amd/display: Remove double checks for
`debug.enable_mem_low_power.bits.cm`"), dpp3_program_blnd_lut(NULL)
checked the low-power debug flag before calling
dpp3_power_on_blnd_lut(false).

After commit 33056a97ae5e ("drm/amd/display: Remove double checks for
`debug.enable_mem_low_power.bits.cm`"), dpp3_program_blnd_lut(NULL)
unconditionally calls dpp3_power_on_blnd_lut(false). The BLNDGAM power
helper writes BLNDGAM_MEM_PWR_FORCE when CM low-power is disabled, causing
immediate SRAM power toggles instead of deferring at vupdate. This can
disrupt atomic color/LUT sequencing during transitions between
direct scanout and composition within gamescope's DRM backend on
Steam Deck OLED.

To fix this, leave the BLNDGAM power state unchanged when low-power is
disabled, matching dpp3_power_on_hdr3dlut and dpp3_power_on_shaper.

Fixes: 33056a97ae5e ("drm/amd/display: Remove double checks for `debug.enable_mem_low_power.bits.cm`")
Signed-off-by: Matthew Schwartz <matthew.schwartz@linux.dev>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add uniras version in sysfs

Display uniras version in sysfs version interface
when uniras enable.

v2: display ras version detail info

Signed-off-by: Jinzhou Su <jinzhou.su@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: get rev_id from strap register or IP-discovery table

Query the sub-revision field in the IP Discovery table for the VFs
to obtain their revision ID.
Meanwhile, read the revision ID from the strap register for the PF.

Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: clear bad page info of ras module

Clear bad page info of ras module.

Signed-off-by: Jinzhou Su <jinzhou.su@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: pause the workload setting in dm

v1:
Pause the workload setting in dm when doinn idle optimization

v2:
Rebase patch to latest kernel code base (kernel 6.16)

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Yang Wang <kevinyang.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: Remove calls to drm_put_dev()

Since the allocation of the drivers main structure was changed to
devm_drm_dev_alloc() drm_put_dev()'ing to trigger it to be free'd
should be done by devres.

However, drm_put_dev() is still in the probe error and device remove
paths. When the driver fails to probe warnings like the following are
shown because devres is trying to drm_put_dev() after the driver
already did it.

[    5.642230] radeon 0000:01:05.0: probe with driver radeon failed with error -22
[    5.649605] ------------[ cut here ]------------
[    5.649607] refcount_t: underflow; use-after-free.
[    5.649620] WARNING: CPU: 0 PID: 357 at lib/refcount.c:28 refcount_warn_saturate+0xbe/0x110

Fixes: a9ed2f052c5c ("drm/radeon: change drm_dev_alloc to devm_drm_dev_alloc")
Signed-off-by: Daniel Palmer <daniel@0x0f.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: Do not kfree() devres managed rdev

Since the allocation of the drivers main structure was changed to
devm_drm_dev_alloc() rdev is managed by devres and we shouldn't be calling
kfree() on it.

This fixes things exploding if the driver probe fails and devres cleans up
the rdev after we already free'd it.

Fixes: a9ed2f052c5c ("drm/radeon: change drm_dev_alloc to devm_drm_dev_alloc")
Signed-off-by: Daniel Palmer <daniel@0x0f.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: Clean up pdev->dev instances in probe

Get a struct device pointer from the start and use it.

Signed-off-by: Daniel Palmer <daniel@0x0f.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Check that VPE has reached DPM0 in idle handler

[Why]
Newer VPE microcode has functionality that will decrease DPM level
only when a workload has run for 2 or more seconds.  If VPE is turned
off before this DPM decrease and the PMFW doesn't reset it when
power gating VPE, the SOC can get stuck with a higher DPM level.

This can happen from amdgpu's ring buffer test because it's a short
quick workload for VPE and VPE is turned off after 1s.

[How]
In idle handler besides checking fences are drained check PMFW version
to determine if it will reset DPM when power gating VPE.  If PMFW will
not do this, then check VPE DPM level. If it is not DPM0 reschedule
delayed work again until it is.

v2: squash in return fix (Alex)

Cc: Peyton.Lee@amd.com
Reported-by: Sultan Alsawaf <sultan@kerneltoast.com>
Reviewed-by: Sultan Alsawaf <sultan@kerneltoast.com>
Tested-by: Sultan Alsawaf <sultan@kerneltoast.com>
Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4615
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Remove unused members in amdgpu_mman

Discovery related members are now part of amdgpu_discovery_info.

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: query block error count of ras module

Query block error count of ras module.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add logic for VF data exchange region to init from dynamic crit_region offsets

1. Added VF logic to init data exchange region using the offsets from dynamic(v2) critical regions;

Signed-off-by: Ellen Pan <yunru.pan@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add logic for VF ipd and VF bios to init from dynamic crit_region offsets

1. Added VF logic in amdgpu_virt to init IP discovery using the offsets from dynamic(v2) critical regions;
2. Added VF logic in amdgpu_virt to init bios image using the offsets from dynamic(v2) critical regions;

Signed-off-by: Ellen Pan <yunru.pan@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Reuse fw_vram_usage_* for dynamic critical region in SRIOV

- During guest driver init, asa VFs receive PF msg to
init dynamic critical region(v2), VFs reuse fw_vram_usage_*
 from ttm to store critical region tables in a 5MB chunk.

Signed-off-by: Ellen Pan <yunru.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Introduce SRIOV critical regions v2 during VF init

    1. Introduced amdgpu_virt_init_critical_region during VF init.
     - VFs use init_data_header_offset and init_data_header_size_kb
            transmitted via PF2VF mailbox to fetch the offset of
            critical regions' offsets/sizes in VRAM and save to
            adev->virt.crit_region_offsets and adev->virt.crit_region_sizes_kb.

Signed-off-by: Ellen Pan <yunru.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add SRIOV crit_region_version support

1. Added enum amd_sriov_crit_region_version to support multi versions
2. Added logic in SRIOV mailbox to regonize crit_region version during
   req_gpu_init_data

Signed-off-by: Ellen Pan <yunru.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Updated naming of SRIOV critical region offsets/sizes with _V1 suffix

 - This change prepares the later patches to intro  _v2 suffix to SRIOV critical regions

Signed-off-by: Ellen Pan <yunru.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: query bad page info of ras module

Query bad page info of ras module.

V2:
  Update code to reuse bad page output code.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: ras module supports error injection

ras module supports error injection.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/ras: Update function and remove redundant code

Update function and remove redundant code:
1. Update function to prepare for internal use.
2. Remove unused function code previously prepared
   for ioctl.

V2:
  Update commit message content.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/ras: Update ras command context structure name

According to the actual usage of this structure,
it is more appropriate to call it context, the
structure name with ioctl is easy to cause
misunderstanding.

V2:
  Update commit message content.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Promote DC to 3.2.355

This version brings along following update:

-Fix GFP_ATOMIC abuse
-Fix several checkpatch issues
-Set DCN32 to use update planes and stream version 3
-Write segment pointer with mot enabled for MST
-Control BW allocation in FW side
-Change clean dsc blocks condition in accelerated mode
-Check disable_fec flag before enabling FEC

Acked-by: Wayne Lin <wayne.lin@amd.com>
Signed-off-by: Taimur Hassan <Syed.Hassan@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: write segment pointer with mot enabled for mst

[Why]
Some mst branches NAK's segment pointer writes with mot disabled.
So reset of segment pointer to 0 should be performed with mot enabled.

[How]
Write segment pointer of mst branch devices with mot enabled.

Reviewed-by: Cruise Hung <cruise.hung@amd.com>
Signed-off-by: Meenakshikumar Somasundaram <meenakshikumar.somasundaram@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Control BW allocation in FW side

[Why]
The BW allocation feature should be controlled in FW side.

[How]
Pass the control bit to FW boot option.

Reviewed-by: Meenakshikumar Somasundaram <meenakshikumar.somasundaram@amd.com>
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Cruise Hung <Cruise.Hung@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix misc. checkpatch issues

[Why/How]
Addresses various checkpatch issues related to the HWSS block sequence
function change.

Reviewed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Ilya Bakoulin <Ilya.Bakoulin@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Change clean dsc blocks condition in accelerated mode

[Why]
On system resume from S4 with the lid closed,
DSC was not cleared because DPMS was already off.

[How]
In accelerated mode, to clean up DSC blocks if eDP dpms off is true
to align the DSC and dpms state when we are not in fast boot and
seamless boot.

Reviewed-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Lewis Huang <Lewis.Huang@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Set DCN32 to use update planes and stream version 3

[Why]
Old minimal transition does not always wait for updates to complete
before proceeding, which can lead to corruption in multi display
scenarios for DCN32.

[How]
Set DCN32 to use update_planes_and_stream_v3 for better pipe transition
handling.

Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Nicholas Carbones <Nicholas.Carbones@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Check disable_fec flag before enabling fec.

[Why]
dc debug option disable_fec was not working.

[How]
Check dc debug option disable_fec flag before
enabling fec in dp_should_enable_fec().

Reviewed-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Meenakshikumar Somasundaram <meenakshikumar.somasundaram@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix GFP_ATOMIC abuse

There is a lot GFP_ATOMIC allocations which are not in interrupt
context. Change them to use GFP_KERNEL instead.

Reviewed-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Enable ras module

Enable ras module, disabled by default.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Fix set but not used warnings

There are many set but not used warnings under drivers/gpu/drm/amd when
compiling with the latest upstream mainline GCC:

  drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c:305:18: warning: variable ‘p’ set but not used [-Wunused-but-set-variable=]
  drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h:103:26: warning: variable ‘internal_reg_offset’ set but not used [-Wunused-but-set-variable=]
  ...
  drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h:164:26: warning: variable ‘internal_reg_offset’ set but not used [-Wunused-but-set-variable=]
  ...
  drivers/gpu/drm/amd/amdgpu/../display/dc/dc_dmub_srv.c:445:13: warning: variable ‘pipe_idx’ set but not used [-Wunused-but-set-variable=]
  drivers/gpu/drm/amd/amdgpu/../display/dc/dc_dmub_srv.c:875:21: warning: variable ‘pipe_idx’ set but not used [-Wunused-but-set-variable=]

Remove the variables actually not used or add __maybe_unused attribute for
the variables actually used to fix them, compile tested only.

Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add ras module ip block to amdgpu discovery

Add ras module ip block to amdgpu discovery.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: check save count before RAS bad page saving

It's possible that unit_num is larger than 0 but save_count is zero,
since we do get bad page address but the address is invalid. Check
unit_num and save_count together.

Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Candice Li <candice.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add the kernel docs for alloc/free/valid range

Add kernel docs for the functions related to hmm_range.

Documents added for functions:
amdgpu_hmm_range_valid
amdgpu_hmm_range_alloc
amdgpu_hmm_range_free

Signed-off-by: Sunil Khatri <sunil.khatri@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: use GPU_HDP_FLUSH for sriov

Currently SRIOV runtime will use kiq to write HDP_MEM_FLUSH_CNTL for
hdp flush. This register need to be write from CPU for nbif to aware,
otherwise it will not work.

Implement amdgpu_kiq_hdp_flush and use kiq to do gpu hdp flush during
sriov runtime.

v2:
- fallback to amdgpu_asic_flush_hdp when amdgpu_kiq_hdp_flush failed
- add function amdgpu_mes_hdp_flush

v3:
- changed returned error

Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Victor Zhao <Victor.Zhao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add kiq hdp flush callbacks

Add kiq hdp flush callbacks for gfx ips to support gpu hdp flush when no
ring presents

Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Victor Zhao <Victor.Zhao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Add a helper to tell whether an IP block HW is enabled

There is already a helper for telling if a block is valid, but if
IP handling wants to check if it's HW is enabled no such helper
exists.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Fix vram_usage underflow

vram_usage was subtracting non-vram memory size,
which caused it to become negative.

Signed-off-by: Alysa Liu <Alysa.Liu@amd.com>
Reviewed-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: Avoid writing nulls into `pp_od_clk_voltage`

Calling `smu_cmn_get_sysfs_buf` aligns the
offset used by `sysfs_emit_at` to the current page boundary, which was
previously directly returned from the various `print_clk_levels`
implementations to be added to the buffer position.
Instead, only the relative offset showing how much was written
to the buffer should be returned, regardless of how it was changed
for alignment purposes.

Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Ilya Zlobintsev <ilya.zlobintsev@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Use memset32 for IB padding

Use memset32 instead of open coding it, just because it is
that bit nicer.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Promote DC to 3.2.354

Display Core v3.2.354 release highlights:

* DCN35 dispclk, dppclk & other fixes
* DCN401 cursor offload fix
* Add new block seqeunce-building/executing functions
* null ptr fixes
* DPIA hpd fix
* debug improvements
* Fix performance regression from full updates
* Firmware Release 0.1.31.0

Signed-off-by: Taimur Hassan <Syed.Hassan@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Reviewed-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: [FW Promotion] Release 0.1.31.0

Release highlights:

DCN35/351/36:
* fix video lag with replay
* DPP DTO programming sequence fix
* IPS exit programming sequence fix

DCN 3.1.5:
* fix video lag with replay

Signed-off-by: Taimur Hassan <Syed.Hassan@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Reviewed-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: add new block sequence-building/executing functions

[Why/How]
Create functions for building/executing HW block programming steps

Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Signed-off-by: Ilya Bakoulin <Ilya.Bakoulin@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: add additional hdcp traces

[why]
Current hdcp trace only tracks hdcp errors. We need to expand the trace
structure for more tracing information.

[how]
Add following traces for hdcp1:
- attempt_count
- downstream_device_count
Add following traces for hdcp2:
- attempt_count
- downstream_device_count
- hdcp1_device_downstream
- hdcp2_legacy_device_downstream

Reviewed-by: Sung Lee <sung.lee@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix performance regression from full updates

[Why]
full_update_required is too strict at update_planes_and_stream_state,
causing a performance regression due to too many updates being full.

[How]
* Carve out weak version of full_update_required for use inside
update_planes_and_stream_state.

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Dominik Kaszewski <dominik.kaszewski@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove dc state from check_update

[Why]
dc_check_update_surfaces_for_stream should not have access to entire
DC, especially not a mutable one. Concurrent checks should be able
to run independently of one another, without risk of changing state.

[How]
* Remove access to dc state other than debug and capacity.
* Move some checks from DC to DM caller.

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Dominik Kaszewski <dominik.kaszewski@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: use GFP_NOWAIT for allocation in interrupt handler

schedule_dc_vmin_vmax() is called by dm_crtc_high_irq(). Hence, we
cannot have the former sleep. Use GFP_NOWAIT for allocation in this
function.

Fixes: c210b757b400 ("drm/amd/display: fix dmub access race condition")
Cc: Mario Limonciello <mario.limonciello@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Sun peng (Leo) Li <sunpeng.li@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add sink/link debug logs

Add some extra logs to better help triage blackscreen issues.

* Dump all the links to see if they have sinks associated.
* Print the edid manufacturer & product id associated with a stream that
  was just created.

Reviewed-by: Jerry Zuo <jerry.zuo@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Move all DCCG RCG into HWSS root_clock_control

[why & how]
Enabling/disabling DCCG RCG should be done as a last-level step when
enabling/disable blocks. This is handled by HWSS root_clock_control
already during optimize_bandwidth.
However, dccg35_dpp_root_clock_control was missing the RCG enable
call on the disable path.

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Ovidiu Bunea <ovidiu.bunea@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: update perfmon measurement interfaces

[how]
The commit update interfaces for dchubbub perfmon meansurement to better
reflect our requirements.

Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix dppclk rcg poweron check

[WHY & HOW]
dppclk rcg power down will flip the poweron flag in the cache to cause dppclk rcg will never
run the rcg ungate sequence in some condition. Wait 10us to let dpp dto fully ramp.

Reviewed-by: Ovidiu (Ovi) Bunea <ovidiu.bunea@amd.com>
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Yihan Zhu <Yihan.Zhu@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: not skip hpd irq for bw alloc mode

[WHY]
Driver only process hpd irq when a branch device or when
the link is established. It would cause some irq for bw_alloc
mode of dp tunneling are ignored.

[HOW]
Driver should process hpd irq if bw_alloc and dp tunneling
are enabled.

Reviewed-by: Cruise Hung <cruise.hung@amd.com>
Signed-off-by: Peichen Huang <PeiChen.Huang@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Update spacing in struct

Update spacing so that fields with longer name will
still be aligned correctly (new fields to be added).

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Update DCN401 path for cursor offload

[Description]
The DCN401 cursor offload path needs to take into account
use_mall_for_cursor, and also need to ensure the dcn32
function assigns the cursor cache fields (DCN401 uses the
dcn32 implementation).

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: increase max link count and fix link->enc NULL pointer access

[why]
1.) dc->links[MAX_LINKS] array size smaller than actual requested.
max_connector + max_dpia + 4 virtual = 14.
increase from 12 to 14.

2.) hw_init() access null LINK_ENC for dpia non display_endpoint.

Cc: Mario Limonciello <mario.limonciello@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Meenakshikumar Somasundaram <meenakshikumar.somasundaram@amd.com>
Reviewed-by: Chris Park <chris.park@amd.com>
Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Rework HDMI data channel reads

Fix the HDMI data channel reads to respect scdc_present field
to pass compliance test.

Reviewed-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Relja Vojvodic <rvojvodi@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix typo in display_mode_core_structs.h

Fix a typo in a comment, change "enviroment" to "environment" in
drivers/gpu/drm/amd/display/dc/dml2/display_mode_core_structs.h

Signed-off-by: Aditya Gollamudi <adigollamudi@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: add dccg dfs mask def

[why]
add some register masks for DCCG

Reviewed-by: Yihan Zhu <yihan.zhu@amd.com>
Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove unused field in DML

Remove unused fields.

Reviewed-by: Austin Zheng <austin.zheng@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix NULL pointer dereference

[Why]
On a mst branch with multi display setup, dc context is obselete
after updating the first stream. Referencing the same dc context
for the next stream update to fetch dc pointer leads to NULL
pointer dereference.

[How]
Get the dc pointer from the link rather than context.

Cc: Mario Limonciello <mario.limonciello@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Charlene Liu <charlene.liu@amd.com>
Signed-off-by: Meenakshikumar Somasundaram <meenakshikumar.somasundaram@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: add dispclk ramping to dcn35.

[why]
this is a required logic based on HW programming guide.
tested/ported on dcn401.

Reviewed-by: Yihan Zhu <yihan.zhu@amd.com>
Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add debug option to override EASF scaler taps

[Why & How]
Add new option override_easf to use in_taps instead of internal
 taps policy for debugging

Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Signed-off-by: Samson Tam <Samson.Tam@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix duplicate aux command with AMD aux backlight

when using AMD aux backlight control, we avoid sending backlight
update commands to DMUB firmware because it is controlled by aux commands
in driver.

Reviewed-by: Iswara Nagulendran <iswara.nagulendran@amd.com>
Reviewed-by: Aric Cyr <aric.cyr@amd.com>
Signed-off-by: Harry VanZyllDeJong <hvanzyll@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add ras module eeprom safety watermark check

Add ras module eeprom safety watermark check.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Avoid hive seqno increment in legacy ras

The hive->event_mgr variable is used by both ras module
and legacy ras. To ensure the continuity of hive seqno
growth, after enabling ras module, it is forbidden to
operate the event_mgr variable in legacy ras.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add poison consumption sequence numbers for gfx and sdma

Add poison consumption sequence numbers for
gfx and sdma.

V3:
  Use RAS_EVENT_LOG to print ras log info.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Avoid loading bad pages into legacy ras

When ras module is enabled, the bad pages will
be loaded by ras module.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add ras module rma check

Add ras module rma check.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Improve ras fatal error handling function

In multi-gpu case, a fatal error will generate several
fatal error interrupts. After improving this function,
the ras module can reuse this function to only
handle the first interrupt.

V3:
  Initialize event_id using RAS_EVENT_INVALID_ID.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Intercept ras interrupts to ras module

Intercept ras interrupts to ras module.

V2:
  Change function names in ras module.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Drop calls to restore power limit and clock from smu_resume()

User requested power limits and clock settings are already restored as
part of smu_restore_dpm_user_profile(). It's unnecessary to call the
same restore as part of smu_resume().

Revert the following commits to drop that extra restore:
commit ed4efe426a49 ("drm/amd: Restore cached power limit during resume")
commit 796ff8a7e01b ("drm/amd: Restore cached manual clock settings during resume")
commit f9b80514a722 ("drm/amd: Only restore cached manual clock settings in restore if OD enabled")

Suggested-by: Lijo Lazar <Lijo.Lazar@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: update remove after reset flag for MES remove queue

Remove queue after reset flag is required to remove a queue that has
been successfully reset to clean up the MES' internal state.

Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add ras module files into amdgpu

Add ras module files into amdgpu.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/userqueue: validate userptrs for userqueues

userptrs could be changed by the user at any time and
hence while locking all the bos before GPU start processing
validate all the userptr bos.

Signed-off-by: Sunil Khatri <sunil.khatri@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: update the functions to use amdgpu version of hmm

At times we need a bo reference for hmm and for that add
a new struct amdgpu_hmm_range which will hold an optional
bo member and hmm_range.

Use amdgpu_hmm_range instead of hmm_range and let the bo
as an optional argument for the caller if they want to
the bo reference to be taken or they want to handle that
explicitly.

Signed-off-by: Sunil Khatri <sunil.khatri@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Reserve discovery TMR only if needed

For legacy SOCs, discovery binary is sideloaded. Instead of checking for
binary blob, use a flag to determine if discovery region needs to be
reserved.

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: export a function amdgpu_smu_ras_send_msg to allow send msg directly

provide a interface that allows ras client send msg to smu/pmfw directly.

Signed-off-by: Yang Wang <kevinyang.wang@amd.com>
Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: Grant interface access after full init

Allow access to user interfaces like sysfs/hwmon only after full
initialization of the device. When device is part of XGMI hive and a
reset is required during initialization, the inteface files will be
created as part of minimal device initialization. Full initialization of
the device will be done only after all devices in XGMI hive are probed
and a reset is done together on all.

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Asad Kamal <asad.kamal@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Move reset-on-init sequence earlier

Complete reset-on-init sequence before sysfs interfaces are created.
Devices get properly initiaized only after reset, and then only sysfs
interfaces should be made available.

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Asad Kamal <asad.kamal@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add amdgpu_discovery_info

Add amdgpu_discovery_info structure to keep all discovery related
information.

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>