drm/amdkfd: Support process XNACK mode dynamic change

Update queue qpd is done for the first queue creation of the process,
if the device support XNACK mode per process, update qpd setup
sh_mem_config based on the process XNACK mode, to support the process
destroy all queues, change XNACK mode, and then create queues.

Add helper macro KFD_SUPPORT_XNACK_PER_PROCESS to remove duplicate code
and add new ASICs support in future.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Evaluate early init for all IP blocks even if one fails

If early init fails for a single IP block, then no further IP blocks
are evaluated.  This means that if a user was missing more than one
firmware binary they would have to keep adding binaries and re-probing
until they discovered the ones missing.

To make this easier, run early init for each IP block and report a single
failure if not all passed.

Reviewed-by: Aaron Liu <aaron.liu@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Remove needless break for legacy IP discovery MP0 9.0.0

There is already a "default" case in the switch block, so there is
no need to have a break after the switch block.

Reviewed-by: Aaron Liu <aaron.liu@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix cleaning up reserved VMID on release

We need to reset this or otherwise run into list corruption later on.

Fixes: e44a0fe630c5 ("drm/amdgpu: rework reserved VMID handling")
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Tested-by: Candice Li <candice.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Adjust ras support check condition for special asic

[Why]:
     Amdgpu ras uses amdgpu_ras_is_supported to check whether
  the ras block supports the ras function. amdgpu_ras_is_supported
  uses .ras_enabled to determine whether the ras function of the
  block is enabled.
     But for special asic with mem ecc enabled but sram ecc not
  enabled, some ras blocks support poison mode but their ras function
  is not enabled on .ras_enabled, these ras blocks will run abnormally.

[How]:
    If the ras block is not supported on .ras_enabled but the asic
  supports poison mode and the ras block has ras configuration, it
  can be considered that the ras block supports ras function.

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: Remove unnecessary ras block support check

[Why]:
   For special asic with mem ecc enabled but sram ecc
not enabled, some ras blocks can register their ras
configuration to ras list, but these ras blocks are not
enabled on .ras_enabled, so it can not get ras block
object using amdgpu_ras_get_ras_block.

[How]:
   Remove ras block support check. Even if the ras block
checked is not in the ras list, it will return a null
pointer and will have no effect.

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: Perform gpu reset after gfx finishes processing ras poison consumption on gfx_v11_0_3

Perform gpu reset after gfx finishes processing
ras poison consumption on gfx_v11_0_3.

V2:
 Move gfx poison consumption handler from hw_ops to ip
 function level.

V3:
 Adjust the calling position of amdgpu_gfx_poison_consumation_handler.

V4:
   Since gfx v11_0_3 does not have .hw_ops instance, the .hw_ops null
 pointer check in amdgpu_ras_interrupt_poison_consumption_handler
 needs to be adjusted.

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

drm/amdgpu: Add gfx cp ecc error irq handling on gfx v11_0_3

V2:
  Optimize gfx_v11_0_set_cp_ecc_error_state function.

V3:
  Define macro constant for me pipe instance address interval.

V5:
  Register and handle gfx cp ecc error irq on gfx v11_0_3.

V6:
  Remove invalid intermediate function call.

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

drm/amdgpu: Add gfx ras poison consumption irq handling on gfx v11_0_3

Add gfx ras poison consumption irq handling on gfx v11_0_3.

V2:
  Move ras poison consumption irq handling code of gfx
     v11_0_3 to gfx_v11_0_3.c.
V5:
  Create dedicated irq handler for RLC_GC_FED_INTERRUPT.

V6:
  Remove invalid function call.

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

amd/amdgpu: Add RLC_RLCS_FED_STATUS_* to gc v11_0_3 ip headers

V2:
   Add RLC_RLCS_FED_STATUS_0 and RLC_RLCS_FED_STATUS_1 register
   offset and shift masks.

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

drm/amdgpu: Add gfx ras function on gfx v11_0_3

Add gfx ras function on gfx v11_0_3.

V2:
 1. Add separate source files for gfx v11_0_3.
 2. Create a common function to initialize gfx ras block.

V3:
 1. Rename amdgpu_gfx_ras_block_init to amdgpu_gfx_ras_sw_init.
 2. Adjust the calling position of amdgpu_gfx_ras_sw_init.
 3. Remove gfx_v11_0_3_ras_ops.

V4:
 Revert changes in amdgpu_ras_interrupt_poison_consumption_handler.

V5:
 1. Remove invalid include file in gfx_v11_0_3.c.
 2. Reduce the number of parameters of amdgpu_gfx_ras_sw_init.

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

drm/amd: Fix renoir/green sardine MP0 IP version detection

The existing codebase never had a case for detecting MP0 version on
Renoir and instead relied upon hardcoded chip name.  This was missed as
part of the changes to migrate all IP blocks to build filenames from
`amdgpu_ucode.c`.

Consequently, Renoir tries to fetch a binary with 11_0_3 in the filename
and since it's supposed to have "renoir" in the filename fails to probe.
The fbdev still works though so the series worked.

Add a case for Renoir into the legacy table to ensure the right ASD and
TA firmware load again.

Reported-by: Ekene Akuneme <Ekene.Akuneme@amd.com>
Reported-by: Nicholas Choi <Nicholas.Choi@amd.com>
Cc: Alex Hung <Alex.Hung@amd.com>
Fixes: 994a97447e38 ("drm/amd: Parse both v1 and v2 TA microcode headers using same function")
Fixes: 54a3e032340e ("drm/amd: Add a legacy mapping to "amdgpu_ucode_ip_version_decode"")
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: 3.2.218

This version brings along following fixes:

- Revert patches that caused regressions associated with audio and an
  old change that checks the DCN version.
- Refactor DDC and HDP.
- Move DPIA and DPCD logic to new files.
- Updates to DMUB.
- Optimization and bug fixes for SUBVP/DRR.
- Drop legacy code.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: hdcp not enabled on connector 0

[Why]
incorrect skip when drm_connector.index = 0 within
event_property_validate and update

[How] handle hdcp validate and update for connector 0

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: hersen wu <hersenxs.wu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: phase2 enable mst hdcp multiple displays

[why]
For MST topology with 1 physical link and multiple connectors (>=2),
e.g. daisy cahined MST + SST, or 1-to-multi MST hub, if userspace
set to enable the HDCP simultaneously on all connected outputs, the
commit tail iteratively call the hdcp_update_display() for each
display (connector). However, the hdcp workqueue data structure for
each link has only one DM connector and encryption status members,
which means the work queue of property_validate/update() would only
be triggered for the last connector within this physical link, and
therefore the HDCP property value of other connectors would stay on
DESIRED instead of switching to ENABLED, which is NOT as expected.

[how]
Use array of AMDGPU_DM_MAX_DISPLAY_INDEX for both aconnector and
encryption status in hdcp workqueue data structure for each physical
link. For property validate/update work queue, we iterates over the
array and do similar operation/check for each connected display.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Signed-off-by: hersen wu <hersenxs.wu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Account for MPO planes in dcn32 mall alloc calculations

[WHY?]
Cannot only consider the MALL required from top pipes because of the MPO
case.

[HOW?]
Only count a pipe if it fits the following criteria:
1) does not have a top pipe (is the topmost pipe for that plane)
2) it does have a top pipe, but that pipe is associated with a different
   plane

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Nevenko Stupar <Nevenko.Stupar@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Allow subvp on vactive pipes that are 2560x1440@60

Enable subvp on specifically 1440p@60hz displays even though it can
switch in vactive.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Request min clocks after disabling pipes on init

[Description]
- Request min clocks after disabling pipes on init
- This optimizes for power savings during init sequence
- Also handles the case where we boot up with no display connected

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix mapping to non-allocated address

[Why]
There is an issue mapping non-allocated location of memory.
It would allocate gpio registers from an array out of bounds.

[How]
Patch correct numbers of bounds for using.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Martin Leung <Martin.Leung@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Brandon Syu <Brandon.Syu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Skip backlight control delay on external powered links

[Why]
When an eDP panel is powered externally from a different GPU, we can avoid
waiting for hardware sequencing delays when switching the backlight on/off
as the display backlight is no longer powered by the original source.

[How]
This commit extends the 'link_powered_externally' variable to allow
bypassing hardware delays for additional backlight commands and force the
backlight on/off when a link is powered by another GPU.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Felipe Clark <felipe.clark@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Tony Tascioglu <tony.tascioglu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: contional remove disable dig_fifo when blank

keep dig_fifo enable for test pattern generation.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Chris Park <Chris.Park@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove DISPCLK dentist programming for dcn32

[WHY?]
For dcn32, SMU handles DISPCLK dentist programming.

[HOW?]
Only use calls to SMU to set DISPCLK.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add extra mblk for DCC

[Why]
DCC meta was found to be detached from usable pixel data. Due to this
DCC meta and the end of the fetched part of the frame will be on not
be on the same mblk. Furthermore if the meta is not aligned to the mblk
size, then we require an extra mblk in MALL to account for this.

[How]
Always add an additional mblk when DCC is enabled for detachment and
misalignment.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Dillon Varone <Dillon.Varone@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Saaem Rizvi <SyedSaaem.Rizvi@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: set active bit for desktop with VSDBv3

When using freesync on desktop, need to set freesync active bit for AMD
VSDBv3 infopacket.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Martin Leung <Martin.Leung@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove unused code

Remove some code that is never used from freesync file.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Optimize link power-down when link powered externally

[Why]
When an eDP panel is powered externally by a different GPU, we don't need
to wait for hardware sequencing delays when powering down a link, as the
display is not dependent on the GPU being powered down.

[How]
This commit adds a variable 'link_powered_externally' to indicate when a
link is being powered by another GPU.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Felipe Clark <felipe.clark@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Tony Tascioglu <tony.tascioglu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix an error check condition for synced pipes

Checking for disabled master pipe on a timing synchronized pipe is
incorrect in the case of ODM combine. This case is acceptable as long as
the disabled master pipe is part of the ODM tree. Skip printing error
message if this condition holds true.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Dillon Varone <Dillon.Varone@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amd/display: Demote Error Level When ODM Transition Supported"

This reverts commit e81b6a4427f3ca37859f5b9fdb6a66683bb84e2e.

Reverting to put in a better solution which does not involve checking
DCN version.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Dillon Varone <Dillon.Varone@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: move dp capability related logic to link_dp_capability

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Wesley Chalmers <Wesley.Chalmers@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: move dp phy related logic to link_dp_phy

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Wesley Chalmers <Wesley.Chalmers@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: move dp link training logic to link_dp_training

[why]
Extract dp link training logic out to their own files.
link_dp_training - high level training sequence and helper functions.
link_dp_training_8b_10b - dp1.x training
link_dp_training_auxless - aux-less training
link_dp_traininig_dpia - dpia training
link_dp_training_fixed_vs_pe_retimer - fixed vs pe retimer training
link_dp_training_128b_132b - dp2.1 training

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Wesley Chalmers <Wesley.Chalmers@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove SubVp support if src/dst rect does not equal stream timing

Current implementation of SubVP does not support cases where stream
timing matched neither the destination rect nor the source rect.

Will need to further debug to see how we can support these cases.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Saaem Rizvi <SyedSaaem.Rizvi@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Change i2c speed for hdcp

[why]
HDCP 1.4 failed on SL8800 SW w/a test driver use.

[how]
slower down the HW i2c speed when used by HW i2c.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Aurabindo Pillai <Aurabindo.Pillai@amd.com>
Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: cleanup function args in dml

Remove array size on array passed to CalculateDETSwathFillLatencyHiding.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Use DML for MALL SS and Subvp allocation calculations

MALL SS and Subvp use the same calculations for determining the size of
the required allocation for a given surface, which is already done in
DML. Add an interface to extract this information from VBA variables and
use in their respective helper functions. Also refactor existing code to
remove stale workarounds.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Account for Subvp Phantoms in DML MALL surface calculations

DML does not explicitly consider support for space in MALL required for
subvp phantom pipes. This adds a check to make sure portion of phantom
surface can fit in MALL.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Account for DCC Meta pitch in DML MALL surface calculations

DML incorrectly uses surface width for determining DCC meta size in MALL
allocation calculations.  Meta pitch should be used instead.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Optimize subvp and drr validation

Two issues existed:
1) Configs that support DRR, but have it disabled will fail subvp+vblank
validation incorrectly. Use subvp+vblank path for this case.
2) Configs that support DRR and have it enabled can use higher voltage level
than required if they also support subvp+vblank. Use lowest supported voltage
level for this case.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add a missing tab

This tab was deleted accidentally and triggers a Smatch warning:

    drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c:1006 gfx_v8_0_init_microcode()
    warn: inconsistent indenting

Add it back.

Fixes: 0aaafb7359d2 ("drm/amd: Use `amdgpu_ucode_*` helpers for GFX8")
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: add hubbub_init related

Required by display init, also update get_dig_mode

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Hansen Dsouza <hansen.dsouza@amd.com>
Reviewed-by: Duncan Ma <duncan.ma@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Update dmub header to match DMUB

[WHY]
Last PR missed name of a struct to match in DMUB

[HOW]
Update the logic in dmub_cmh.h header

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Meenakshikumar Somasundaram <Meenakshikumar.Somasundaram@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Mustapha Ghaddar <mghaddar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Skip specific mmhub and sdma registers accessing under sriov

[Why]
SDMA0_CNTL and MMHUB system aperture related registers are blocked by L1 Policy.
Therefore, they cannot be accessed by VF and loged in violation.

[How]
For MMHUB registers, they will be programmed by PF. So VF will skip to program them in mmhubv3_0.
For SDMA0_CNTL which is a PF_only register, VF don't need to program it in sdma_v6_0.

Signed-off-by: Yifan Zha <Yifan.Zha@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: Support RAS fatal error mode1 reset on smu v13_0_0 and v13_0_10

Support RAS fatal error mode1 reset on smu v13_0_0 and v13_0_10.

Signed-off-by: Candice Li <candice.li@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Correct the power calcultion for Renior/Cezanne.

From smu firmware,the value of power is transferred  in units of watts.

Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/2321
Fixes: 137aac26a2ed ("drm/amdgpu/smu12: fix power reporting on renoir")

Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Jesse Zhang <Jesse.Zhang@amd.com>
Reviewed-by: Aaron Liu <aaron.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: fix some dead code in `gfx_v9_0_init_cp_compute_microcode`

Some dead code was introduced as part of utilizing the `amdgpu_ucode_*`
helpers. Adjust the control flow to make sure that firmware is released
in the appropriate error flows.

Reported-by: coverity-bot <keescook+coverity-bot@chromium.org>
Addresses-Coverity-ID: 1530548 ("Control flow issues")
Fixes: ec787deb2ddf ("drm/amd: Use `amdgpu_ucode_*` helpers for GFX9")
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: move dc_link_dpia logic to link_dp_dpia

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: George Shen <George.Shen@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: move dpcd logic from dc_link_dpcd to link_dpcd

[why]
Moving dpcd logic from dc_link_dpcd to link_dpcd as part of link file
restructure

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: George Shen <George.Shen@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: refactor ddc logic from dc_link_ddc to link_ddc

[why]
1. Move dd_link_ddc functions to link_ddc.
2. Move link ddc functions declaration exposed in dc to link.h
3. Move link ddc functions declaration exposed in dm to dc_link.h
4. Remove i2caux_interface.h file

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: refactor hpd logic from dc_link to link_hpd

[why]
Factor out hpd handling logic from generic dc link file.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Implement FIFO enable sequence on DCN32

[WHY?]
FIFO enable sequence is incomplete as it is currently implemented in FW,
and requires reset to prevent the FIFO to be enabled in an invalid
state. This cannot be done until DIG FE is connected to the BE.

[HOW?]
Add FIFO enable sequence in driver for dcn32 with reset after DIG FE is
connected to BE.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix DPIA link encoder assignment issue

[Why]
The DPIA link encoder (DIG) was not released when a Display XDR 6K
monitor was disconnected. That's because the DPIA link encoder logic
checked the BW when it removed the link encoder for it. And the timing
BW was less than DP link BW.  So, it failed to release the link encoder.
And that caused the DIG was occupied and can't be used.  The reason it
checked the BW is it wants to identify whether this link was using DIG
or HPO. It skips if it's not DIG.

[How]
Not to check the BW when it removes the link encoder because it's
already in the assignment table, it must be the DIG.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Cruise Hung <Cruise.Hung@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix multi edp panel instancing

A previous fix attempted to correct mismatch between DM display
targets and dc panel instancing by only counting connected panels.
This behaviour breaks a feature, thus this is an alternative solution
that allows mapping display targets to dc links during mod_power_create.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amd/display: Speed up DML fast_validate path"

This reverts commit 1b5d0e7e15430aecbf2bb0ac634a44aec971895c.

It caused corruption in some media players.

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Reviewed-by: Nevenko Stupar <Nevenko.Stupar@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Martin Leung <Martin.Leung@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Update BW alloc after new DMUB logic

[WHY]
After introducing new DPIA NOTIFICATION we will need
to update the exiting BW allocation logic

[HOW]
Updated the BW alloc source and header files

Tested-by: Daniel Wheeler <Daniel.Wheeler@amd.com>
Reviewed-by: Meenakshikumar Somasundaram <Meenakshikumar.Somasundaram@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Mustapha Ghaddar <mghaddar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm: amd: display: Fix memory leakage

This commit fixes memory leakage in dc_construct_ctx() function.

Signed-off-by: Konstantin Meskhidze <konstantin.meskhidze@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Conversion to bool not necessary

A logical evaluation already results in bool. There is no need for using
a ternary operator based evaluation and bool conversion of the outcome.
Issue identified using boolconv.cocci Coccinelle semantic patch.
This was also reported by the Kernel Test Robot. Hence

Fixes: 473683a03495 ("drm/amd/display: Create a file dedicated for CRTC")

Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Deepak R Varma <drv@mailo.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove useless else if

The assignment of the else and if branches is the same, so the if else
here is redundant, so we remove it.

./drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c:1951:2-4: WARNING: possible condition with no effect (if == else).

Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=3719
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Avoid ASSERT for some message failures

On DCN314 when resuming from s0i3 an ASSERT is shown indicating that
`VBIOSSMC_MSG_SetHardMinDcfclkByFreq` returned `VBIOSSMC_Result_Failed`.

This isn't a driver bug; it's a BIOS/configuration bug. To make this
easier to triage, add an explicit warning when this issue happens.

This matches the behavior utilized for failures with
`VBIOSSMC_MSG_TransferTableDram2Smu` configuration.

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

drm/amd/display: Fix set scaling doesn's work

[Why]
Setting scaling does not correctly update CRTC state. As a result
dc stream state's src (composition area) && dest (addressable area)
was not calculated as expected. This causes set scaling doesn's work.

[How]
Correctly update CRTC state when setting scaling property.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Tested-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: hongao <hongao@uniontech.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix possible buffer overflow relating to secure display

It is possible that adev->dm.dc->caps.max_links is greater than
AMDGPU_MAX_CRTCS. So, to not potentially access unallocated memory use
adev->mode_info.num_crtc to do the bounds check instead of
adev->dm.dc->caps.max_links.

Fixes: 1b11ff764aef ("drm/amd/display: Implement multiple secure display")
Fixes: b8ff7e08bab9 ("drm/amd/display: Fix when disabling secure_display")
Reviewed-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Page aligned memory reserve size

Use page aligned size to reserve memory usage because page aligned TTM
BO size is used to unreserve memory usage, otherwise no page aligned
size causes memory usage accounting unbalanced.

Change vram_used definition type to int64_t to be able to trigger
WARN_ONCE(adev && adev->kfd.vram_used < 0, "..."), to help debug the
accounting issue with warning and backtrace.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Cleanup vm process info if init vm failed

If acquire_vm failed when initializing KFD vm, set vm->process_info to
NULL and free process info, otherwise, the future acquire_vm will
always fail as vm->process_info is not NULL.

Pass avm as parameter to remove the duplicate code.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Fix NULL pointer error for GC 11.0.1 on mGPU

The point bo->kfd_bo is NULL for queue's write pointer BO
when creating queue on mGPU. To avoid using the pointer
fixes the error.

Signed-off-by: Eric Huang <jinhuieric.huang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: No need for Null pointer check before kfree

kfree() & vfree() internally performs NULL check on the pointer handed
to it and take no action if it indeed is NULL. Hence there is no need
for a pre-check of the memory pointer before handing it to
kfree()/vfree().

Issue reported by ifnullfree.cocci Coccinelle semantic patch script.

Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Deepak R Varma <drv@mailo.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm/smu13: BACO is supported when it's in BACO state

This leverages the logic in smu11. No need to talk to SMU to
check BACO enablement as it's in BACO state already.

Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Add sync after creating vram bo

There will be data corruption on vram allocated by svm
if the initialization is not complete and application is
writting on the memory. Adding sync to wait for the
initialization completion is to resolve this issue.

Signed-off-by: Eric Huang <jinhuieric.huang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Use resource_size() helper function

Use the resource_size() function instead of a open coded computation
resource size. It makes the code more readable.

Issue identified using resource_size.cocci coccinelle semantic patch.

Signed-off-by: Deepak R Varma <drv@mailo.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Fixed bug on error when unloading amdgpu

Fixed bug on error when unloading amdgpu.

The error message is as follows:
[  377.706202] kernel BUG at drivers/gpu/drm/drm_buddy.c:278!
[  377.706215] invalid opcode: 0000 [#1] PREEMPT SMP NOPTI
[  377.706222] CPU: 4 PID: 8610 Comm: modprobe Tainted: G          IOE      6.0.0-thomas #1
[  377.706231] Hardware name: ASUS System Product Name/PRIME Z390-A, BIOS 2004 11/02/2021
[  377.706238] RIP: 0010:drm_buddy_free_block+0x26/0x30 [drm_buddy]
[  377.706264] Code: 00 00 00 90 0f 1f 44 00 00 48 8b 0e 89 c8 25 00 0c 00 00 3d 00 04 00 00 75 10 48 8b 47 18 48 d3 e0 48 01 47 28 e9 fa fe ff ff <0f> 0b 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 41 54 55 48 89 f5 53
[  377.706282] RSP: 0018:ffffad2dc4683cb8 EFLAGS: 00010287
[  377.706289] RAX: 0000000000000000 RBX: ffff8b1743bd5138 RCX: 0000000000000000
[  377.706297] RDX: ffff8b1743bd5160 RSI: ffff8b1743bd5c78 RDI: ffff8b16d1b25f70
[  377.706304] RBP: ffff8b1743bd59e0 R08: 0000000000000001 R09: 0000000000000001
[  377.706311] R10: ffff8b16c8572400 R11: ffffad2dc4683cf0 R12: ffff8b16d1b25f70
[  377.706318] R13: ffff8b16d1b25fd0 R14: ffff8b1743bd59c0 R15: ffff8b16d1b25f70
[  377.706325] FS:  00007fec56c72c40(0000) GS:ffff8b1836500000(0000) knlGS:0000000000000000
[  377.706334] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  377.706340] CR2: 00007f9b88c1ba50 CR3: 0000000110450004 CR4: 00000000003706e0
[  377.706347] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  377.706354] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[  377.706361] Call Trace:
[  377.706365]  <TASK>
[  377.706369]  drm_buddy_free_list+0x2a/0x60 [drm_buddy]
[  377.706376]  amdgpu_vram_mgr_fini+0xea/0x180 [amdgpu]
[  377.706572]  amdgpu_ttm_fini+0x12e/0x1a0 [amdgpu]
[  377.706650]  amdgpu_bo_fini+0x22/0x90 [amdgpu]
[  377.706727]  gmc_v11_0_sw_fini+0x26/0x30 [amdgpu]
[  377.706821]  amdgpu_device_fini_sw+0xa1/0x3c0 [amdgpu]
[  377.706897]  amdgpu_driver_release_kms+0x12/0x30 [amdgpu]
[  377.706975]  drm_dev_release+0x20/0x40 [drm]
[  377.707006]  release_nodes+0x35/0xb0
[  377.707014]  devres_release_all+0x8b/0xc0
[  377.707020]  device_unbind_cleanup+0xe/0x70
[  377.707027]  device_release_driver_internal+0xee/0x160
[  377.707033]  driver_detach+0x44/0x90
[  377.707039]  bus_remove_driver+0x55/0xe0
[  377.707045]  pci_unregister_driver+0x3b/0x90
[  377.707052]  amdgpu_exit+0x11/0x6c [amdgpu]
[  377.707194]  __x64_sys_delete_module+0x142/0x2b0
[  377.707201]  ? fpregs_assert_state_consistent+0x22/0x50
[  377.707208]  ? exit_to_user_mode_prepare+0x3e/0x190
[  377.707215]  do_syscall_64+0x38/0x90
[  377.707221]  entry_SYSCALL_64_after_hwframe+0x63/0xcd

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: make amdgpu_ucode_validate static

No consumers outside of amdgpu_ucode.c use amdgpu_ucode_validate
anymore, so make the function static.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_release` helper for si

The `amdgpu_ucode_release` helper is replacing all calls
to release_firmware.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_release` helper for powerplay

The `amdgpu_ucode_release` helper is replacing all calls to
release_firmware.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for DMCU

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
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/amd: Use `amdgpu_ucode_*` helpers for GPU info bin

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for CGS

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for VCE

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for UVD

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for SDMA on CIK

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for SDMA3.0

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for SDMA2.4

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for GMC8

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for GMC7

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for GMC6

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for GFX8

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for GFX7

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for GFX6

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Optimize SRIOV switch/case for PSP microcode load

Now that IP version decoding is used, a number of case statements
can be combined.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@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/amd: Load SMU microcode during early_init

This will ensure that the microcode is available before the firmware
framebuffer has been destroyed.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for SMU

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_release` helper for DMUB

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
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/amd/display: Load DMUB microcode during early_init

If DMUB is required for an ASIC, ensure that the microcode is available
and validates during early_init.

Any failures will cause the driver to fail to probe before the firmware
framebuffer has been removed.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@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/amd: Use `amdgpu_ucode_*` helpers for PSP

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unloading.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Load PSP microcode during early_init

Simplifies the code so that all PSP versions will get the firmware
name from `amdgpu_ucode_ip_version_decode` and then use this filename
to load microcode as part of the early_init process.

Any failures will cause the driver to fail to probe before the firmware
framebuffer has been removed.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Avoid BUG() for case of SRIOV missing IP version

No need to crash the kernel.  AMDGPU will now fail to probe.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Parse both v1 and v2 TA microcode headers using same function

Several IP versions duplicate code and can't use the common helpers.
Move this code into a single function so that the helpers can be used.

v2: squash in fix from Mario to remove duplicate ta parsing

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

drm/radeon: free iio for atombios when driver shutdown

Fix below kmemleak when unload radeon driver:

unreferenced object 0xffff9f8608ede200 (size 512):
  comm "systemd-udevd", pid 326, jiffies 4294682822 (age 716.338s)
  hex dump (first 32 bytes):
    00 00 00 00 c4 aa ec aa 14 ab 00 00 00 00 00 00  ................
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
  backtrace:
    [<0000000062fadebe>] kmem_cache_alloc_trace+0x2f1/0x500
    [<00000000b6883cea>] atom_parse+0x117/0x230 [radeon]
    [<00000000158c23fd>] radeon_atombios_init+0xab/0x170 [radeon]
    [<00000000683f672e>] si_init+0x57/0x750 [radeon]
    [<00000000566cc31f>] radeon_device_init+0x559/0x9c0 [radeon]
    [<0000000046efabb3>] radeon_driver_load_kms+0xc1/0x1a0 [radeon]
    [<00000000b5155064>] drm_dev_register+0xdd/0x1d0
    [<0000000045fec835>] radeon_pci_probe+0xbd/0x100 [radeon]
    [<00000000e69ecca3>] pci_device_probe+0xe1/0x160
    [<0000000019484b76>] really_probe.part.0+0xc1/0x2c0
    [<000000003f2649da>] __driver_probe_device+0x96/0x130
    [<00000000231c5bb1>] driver_probe_device+0x24/0xf0
    [<0000000000a42377>] __driver_attach+0x77/0x190
    [<00000000d7574da6>] bus_for_each_dev+0x7f/0xd0
    [<00000000633166d2>] driver_attach+0x1e/0x30
    [<00000000313b05b8>] bus_add_driver+0x12c/0x1e0

iio was allocated in atom_index_iio() called by atom_parse(),
but it doesn't got released when the dirver is shutdown.
Fix this kmemleak by free it in radeon_atombios_fini().

Signed-off-by: Liwei Song <liwei.song@windriver.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: clean up some inconsistent indentings

drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c:65 amdgpu_gem_fault() warn: inconsistent indenting

Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=3639
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Load GFX11 microcode during early_init

If GFX11 microcode is required but not available during early init, the
firmware framebuffer will have already been released and the screen will
freeze.

Move the request for GFX11 microcode into the early_init phase
so that if it's not available, driver init will fail.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for GFX11

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper will provide symmetery on unload.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Load GFX10 microcode during early_init

Simplifies the code so that GFX10 will get the firmware
name from `amdgpu_ucode_ip_version_decode` and then use this filename
to load microcode as part of the early_init process.

Any failures will cause the driver to fail to probe before the firmware
framebuffer has been removed.

Reviewed-by: Alex Deucher <alexander.deucher@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/amd: Use `amdgpu_ucode_*` helpers for GFX10

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unload.

Reviewed-by: Alex Deucher <alexander.deucher@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>