drm/amdgpu: add PSP ip block for PSP 13.0.6

Add PSP IP handling for PSP 13.0.6

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Le Ma <Le.Ma@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: initialize ta ucode for psp v13_0_6

Initialize ta ucode for psp v13_0_6

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

drm/amdgpu: init sos microcode for psp v13_0_6

parse psp_v13_0_6_sos.bin and initialze various
psp ucode arraies respectively

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Le Ma <le.ma@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add psp early init for PSP 13.0.6

Initialize psp ip callbacks for PSP 13.0.6.

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Le Ma <le.ma@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: set sdma v4_4_2 ip block

Use sdma 4.4.2 IP block for chips with sdma 4.4.2 hardware.

Signed-off-by: Le Ma <le.ma@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add sdma v4_4_2 support (v4)

Add support for SDMA 4.4.2.

v1: Create sdma_v4_4_2.[c|h] for initial support (Le)
v2: update amdgpu_ring_init call with atomic score (Hawking)
v3: Squash in sdma_start fixes (Alex)
v4: Comment out currently unused RAS code (Alex)

Signed-off-by: Le Ma <le.ma@amd.com>
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add sdma v4_4_2 ip headers

Add sdma v4_4_2 register offset and shift masks
header files

v2: update headers (Alex)

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Le Ma <Le.Ma@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add HDP ip block for HDP 4.4.2

Add HDP IP handling for HDP 4.4.2

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Le Ma <Le.Ma@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: skip hdp invalidation for HDP 4.4.2

No mmHDP_READ_CACHE_INVALIDATE register on HDP 4.4.2.

Signed-off-by: Le Ma <le.ma@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add hdp v4_4_2 ip headers

Add hdp v4_4_2 register offset and shift masks
header files

v2: update headers (Alex)

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Le Ma <Le.Ma@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add IH ip block for IH 4.4.2

Add IH IP handling for IH 4.4.2

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Le Ma <Le.Ma@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: skip ih2 rb allocation for IH 4.4.2

No ih2 hardware on IH 4.4.2.

Signed-off-by: Le Ma <le.ma@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: set ih chicken bit for IH 4.4.2

Share same register address with IH 4.4.0.

Signed-off-by: Le Ma <le.ma@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: remove dead code

The less than zero comparison of unsigned variable "value" is never
true. Remove dead code.

Fixes: c3ed0e72c872 ("drm/amdgpu: added a sysfs interface for thermal throttling")
Signed-off-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Simplify same effect if/else blocks

The if / else block code has same effect irrespective of the logical
evaluation.  Hence, simply the implementation by removing the unnecessary
conditional evaluation. While at it, also fix the long line checkpatch
complaint. Issue identified using cond_no_effect.cocci Coccinelle
semantic patch script.

Fixes: 9114b55fabae ("drm/amd/display: Fix SubVP control flow in the MPO context")
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Deepak R Varma <drv@mailo.com>
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

amdkfd: Memory availability can never be negative

Our assumptions about how much KFD memory is currently available for
allocation may be violated by various complexities so we define the
reported value as advisory, however we should never report negative
availability.

Signed-off-by: Daniel Phillips <daniel.phillips@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Fix initialization mistake for NBIO 7.3.0

The same strapping initialization issue that happened on NBIO 7.5.1
appears to be happening on NBIO 7.3.0.
Apply the same fix to 7.3.0 as well.

Note: This workaround relies upon the integrated GPU being enabled
in BIOS. If the integrated GPU is disabled in BIOS a different
workaround will be required.

Reported-by: Thomas Glanzmann <thomas@glanzmann.de>
Cc: Basavaraj Natikar <Basavaraj.Natikar@amd.com>
Link: https://lore.kernel.org/linux-usb/Y%2Fz9GdHjPyF2rNG3@glanzmann.de/T/#u
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/amdgpu: Fix call trace warning and hang when removing amdgpu device

On GPUs with RAS enabled, below call trace and hang are observed when
shutting down device.

v2: use DRM device unplugged flag instead of shutdown flag as the check to
prevent memory wipe in shutdown stage.

[ +0.000000] RIP: 0010:amdgpu_vram_mgr_fini+0x18d/0x1c0 [amdgpu]
[ +0.000001] PKRU: 55555554
[ +0.000001] Call Trace:
[ +0.000001] <TASK>
[ +0.000002] amdgpu_ttm_fini+0x140/0x1c0 [amdgpu]
[ +0.000183] amdgpu_bo_fini+0x27/0xa0 [amdgpu]
[ +0.000184] gmc_v11_0_sw_fini+0x2b/0x40 [amdgpu]
[ +0.000163] amdgpu_device_fini_sw+0xb6/0x510 [amdgpu]
[ +0.000152] amdgpu_driver_release_kms+0x16/0x30 [amdgpu]
[ +0.000090] drm_dev_release+0x28/0x50 [drm]
[ +0.000016] devm_drm_dev_init_release+0x38/0x60 [drm]
[ +0.000011] devm_action_release+0x15/0x20
[ +0.000003] release_nodes+0x40/0xc0
[ +0.000001] devres_release_all+0x9e/0xe0
[ +0.000001] device_unbind_cleanup+0x12/0x80
[ +0.000003] device_release_driver_internal+0xff/0x160
[ +0.000001] driver_detach+0x4a/0x90
[ +0.000001] bus_remove_driver+0x6c/0xf0
[ +0.000001] driver_unregister+0x31/0x50
[ +0.000001] pci_unregister_driver+0x40/0x90
[ +0.000003] amdgpu_exit+0x15/0x120 [amdgpu]

Signed-off-by: lyndonli <Lyndon.Li@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: remove unused variable

building with gcc and W=1 reports
drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c:7292:11: error:
unused variable 'tmp' [-Werror=unused-variable]
  uint32_t tmp;
           ^~~

tmp is not used so remove it.

Signed-off-by: bobzhou <bob.zhou@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Fix BO offset for multi-VMA page migration

svm_migrate_ram_to_vram migrates a prange from sys ram to vram. The prange may
cross multiple vma. Need remember current dst vram offset in the TTM resource for
each migration.

v2: squash in warning fix (Alex)

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

drm/amd/display: Update clock table to include highest clock setting

[Why]
Currently, the clk manager matches SocVoltage with voltage from
fused settings (dfPstate clock table). And then corresponding clocks
are selected.

However in certain situations, this leads to clk manager not
including at least one entry with highest supported clock setting.

[How]
Update the clk manager to include at least one entry with highest
supported clock setting.

Reviewed-by: Pavle Kotarac <pavle.kotarac@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Swapnil Patel <Swapnil.Patel@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: stop waiting in amdgpu_uvd_send_msg

We have a wait in the amdgpu_bo_kmap() code for quite a while now, so
waiting here isn't needed any more.

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

drm/amd/pm: Enable ecc_info table support for smu v13_0_10

Support EccInfoTable which includes umc ras error count and
error address.

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

drm/amdgpu: Support umc node harvest config on umc v8_10

Don't need to query error count and error address on harvest umc nodes.
v2: Fix code bug, use active_mask instead of harvsest_config
    and remove unnecessary argument in LOOP macro.
v3: Leave adev->gmc.num_umc unchanged.

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

drm/amdgpu: fix no previous prototype warning

add static prefix for vangogh_set_apu_thermal_limit function

Signed-off-by: Kun Liu <Kun.Liu2@amd.com>
Reported-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/oe-kbuild-all/202303010827.c2N0yBGT-lkp@intel.com
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Stop clearing kiq position during fini

Do not clear kiq position in RLC_CP_SCHEDULER so that CP could perform
IDLE-SAVE after VF fini.
Otherwise it could cause GFX hang if another Win guest is rendering.

Signed-off-by: leiyaoyao <yaoyao.lei@amd.com>
Acked-by: ZhenGuo Yin <zhenguo.yin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/connector: print max_requested_bpc in state debugfs

This is useful to understand the bpc defaults and
support of a driver.

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Cc: Pekka Paalanen <ppaalanen@gmail.com>
Cc: Sebastian Wick <sebastian.wick@redhat.com>
Cc: Vitaly.Prosyak@amd.com
Cc: Uma Shankar <uma.shankar@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Joshua Ashton <joshua@froggi.es>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: dri-devel@lists.freedesktop.org
Cc: amd-gfx@lists.freedesktop.org
Reviewed-By: Joshua Ashton <joshua@froggi.es>
Link: https://patchwork.freedesktop.org/patch/msgid/20230113162428.33874-3-harry.wentland@amd.com
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/display: Don't block HDR_OUTPUT_METADATA on unknown EOTF

The EDID of an HDR display defines EOTFs that are supported
by the display and can be set in the HDR metadata infoframe.
Userspace is expected to read the EDID and set an appropriate
HDR_OUTPUT_METADATA.

In drm_parse_hdr_metadata_block the kernel reads the supported
EOTFs from the EDID and stores them in the
drm_connector->hdr_sink_metadata. While doing so it also
filters the EOTFs to the EOTFs the kernel knows about.
When an HDR_OUTPUT_METADATA is set it then checks to
make sure the EOTF is a supported EOTF. In cases where
the kernel doesn't know about a new EOTF this check will
fail, even if the EDID advertises support.

Since it is expected that userspace reads the EDID to understand
what the display supports it doesn't make sense for DRM to block
an HDR_OUTPUT_METADATA if it contains an EOTF the kernel doesn't
understand.

This comes with the added benefit of future-proofing metadata
support. If the spec defines a new EOTF there is no need to
update DRM and an compositor can immediately make use of it.

Bug: https://gitlab.freedesktop.org/wayland/weston/-/issues/609

v2: Distinguish EOTFs defind in kernel and ones defined
    in EDID in the commit description (Pekka)

v3: Rebase; drm_hdmi_infoframe_set_hdr_metadata moved
    to drm_hdmi_helper.c

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Cc: Pekka Paalanen <ppaalanen@gmail.com>
Cc: Sebastian Wick <sebastian.wick@redhat.com>
Cc: Vitaly.Prosyak@amd.com
Cc: Uma Shankar <uma.shankar@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Joshua Ashton <joshua@froggi.es>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: dri-devel@lists.freedesktop.org
Cc: amd-gfx@lists.freedesktop.org
Acked-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Reviewed-By: Joshua Ashton <joshua@froggi.es>
Link: https://patchwork.freedesktop.org/patch/msgid/20230113162428.33874-2-harry.wentland@amd.com
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Promote DAL to 3.2.224

This version brings along the following:
- Correct DML calculation
- Extend Freesync over Pcon support
- Fixes in pstate hang and more
- Code cleanup for dc_link.h and dc_link.c

Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix clock sequence logic for DCN32

[Why&How]
The newer commit sequence on DCN32 onwards did not finish clock
optimization sequence since the newer sequence did not end up calling
dc_post_update_surfaces_to_stream() which resets dc->optimized_required.

Call this function before passing control on to the new commit sequence.

Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Do not update DRR while BW optimizations pending

[why]
While bandwidth optimizations are pending, it's possible a pstate change
will occur.  During this time, VSYNC handler should not also try to update
DRR parameters causing pstate hang

[how]
Do not adjust DRR if optimize bandwidth is set.

Reviewed-by: Aric Cyr <aric.cyr@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Only wait for blank completion if OTG active

[why]
If OTG is not active, waiting for blank completion will always fail and
timeout resulting in unnecessary driver delays.

[how]
Check that OTG is enabled before waiting for blank.

Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: enable DPG when disabling plane for phantom pipe

[Why]
In disable_dangling_plane, for phantom pipes, we enable OTG so
disable programming gets the double buffer update.  But this
causes an underflow to occur.

[How]
Enable DPG prior to enabling OTG.

Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Samson Tam <samson.tam@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: remove empty dc_link.c

[why]
We kept an empty dc_link.c file due to external
build dependency. Now the last build dependency has
been removed. We can safely delete this file.

Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Correct DML calculation to align HW formula

[Why]
In 2560x1440@240p eDP panel, some use cases will enable MPC
combine with RGB MPO then underflow happened. This case is
not allowed from HW formula. 

[How]
Correct eDP, DP and DP2 output bpp calculation to align HW
formula.

Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Paul Hsieh <Paul.Hsieh@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/vcn: fix compilation issue with legacy gcc

This patch is used to fix following compilation issue with legacy gcc
error: ‘for’ loop initial declarations are only allowed in C99 mode
for (int i = 0; i < adev->vcn.num_vcn_inst; ++i) {

Signed-off-by: bobzhou <bob.zhou@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Implement DMA buf fd export from KFD

Exports a DMA buf fd of a given KFD buffer handle. This is intended for
being able to import KFD BOs into GEM contexts to leverage the
amdgpu_bo_va API for more flexible virtual address mappings. It will
also be used for the new upstreamable RDMA solution coming to UCX and
RCCL.

The corresponding user mode change (Thunk API and kfdtest) is here:
https://github.com/fxkamd/ROCT-Thunk-Interface/commits/fxkamd/dmabuf

Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Xiaogang Chen <Xiaogang.Chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Generalize KFD dmabuf import

Use proper amdgpu_gem_prime_import function to handle all kinds of
imports. Remember the dmabuf reference to enable proper multi-GPU
attachment to multiple VMs without erroneously re-exporting the
underlying BO multiple times.

Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: merge dc_link.h into dc.h and dc_types.h

[why]
Remove the need to include dc_link.h separately. dc.h should contain
everything needed on DM side.

[How]
Merge dc_link.h into dc.h and dc_types.h so DM only needs to include
dc.h to use all link public functions.

Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Update BW ALLOCATION Function declaration

[WHY & HOW]
Update the declaration to give a better idea of what the
function does.

Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Mustapha Ghaddar <mghaddar@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amd/display: Fix FreeSync active bit issue"

This reverts commit 6cfb6df2d645c00513ecf17832928e08979fa953.

[Why & How]
Original change causes black screen. Revert
until fix is available.

Reviewed-by: Aric Cyr <aric.cyr@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: DAL to program DISPCLK WDIVIDER if PMFW doesn't

[Why & How]
- If for any reason PMFW fails to set the expected (or valid)
  DISPCLK WDIVIDER, then DAL will program DENTIST DISPCLK
  WDIVIDER to correct for this issue

Reviewed-by: Samson Tam <Samson.Tam@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Extend Freesync over PCon support for more devices

[why]
More branch devices are able to support Freesync
over PCon so include them in the list of supporting devices.

[how]
Add more compatible PCon devices in the whitelist
for Freesync over Pcon.

Reviewed-by: Harry Wentland <Harry.Wentland@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Sung Joon Kim <sungjoon.kim@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: update pixel format in DP hw sequence

[WHY]
DP 420 formats do not light up because the pixel processing mode
of the DP_FORMAT is misprogrammed

[HOW]
Added appropriate programming for DP pixel format

Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Yihan Zhu <yihan.zhu@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: populate subvp cmd info only for the top pipe

[Why]
System restart observed while changing the display resolution
to 8k with extended mode. Sytem restart was caused by a page fault.

[How]
When the driver populates subvp info it did it for both the pipes using
vblank which caused an outof bounds array access causing the page fault.
added checks to allow the top pipe only to fix this issue.

Co-authored-by: Ayush Gupta <ayush.gupta@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Ayush Gupta <ayush.gupta@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: dcn32/321 dsc_pg_control not executed properly

[why]
during boot up or resume from s3, hw default value of
domain_power_forceon is 1. when program domain_power_gate
to 1 to power down hw block, hw will not change to power
off due to domain_power_forceon = 1.

[how]
enable_power_gating_plane(true) should be executed to set
domain_power_forceon to 0 before dsc_pg_control.
dsc_pg_control is already called by dcn3x_init_hw-->
init_pipes--> dsc_pg_control. no need be programmed with
dcn3x_init_hw one more time.
to trigger dchub, dsc block power state change, need
program dc_ip_request_cntl to notify hw block.

Reviewed-by: Nevenko Stupar <Nevenko.Stupar@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Hersen Wu <hersenxs.wu@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Allocation at stream Enable

[WHY & HOW]
After we allocate BW at plug, we will de-alloc
and allocate only what stream needs at
stream_enable()

[HOW]
Introduce bw allocation check at link_enable()
for DPIA links

Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Mustapha Ghaddar <mghaddar@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amd/display: Do not set DRR on pipe commit"

This reverts commit 4f1b5e739dfd1edde33329e3f376733a131fb1ff.

[Why & How]
Original change causes a regression. Revert
until fix is available.

Reviewed-by: Aric Cyr <aric.cyr@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Updating Video Format Fall Back Policy.

[WHY]
Adding 1920x1080 as fail safe mode for
Video Format Fall Back Policy.

Reviewed-by: Jerry Zuo <Jerry.Zuo@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Jasdeep Dhillon <jdhillon@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Reduce CPU busy-waiting for long delays

[WHY]
udelay should not be used for long waits since it keeps CPU active,
wasting power.

[HOW]
Use fsleep where acceptable to allow CPU cores to be parked by the scheduler.

Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix shift-out-of-bounds in CalculateVMAndRowBytes

[WHY]
When PTEBufferSizeInRequests is zero, UBSAN reports the following
warning because dml_log2 returns an unexpected negative value:

  shift exponent 4294966273 is too large for 32-bit type 'int'

[HOW]

In the case PTEBufferSizeInRequests is zero, skip the dml_log2() and
assign the result directly.

Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Ext displays with dock can't recognized after resume

[Why]
Needs to set the default value of the LTTPR timeout after resume.

[How]
Set the default (3.2ms) timeout at resuming if the sink supports
LTTPR

Reviewed-by: Jerry Zuo <Jerry.Zuo@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Ryan Lin <tsung-hua.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix ttm_bo calltrace warning in psp_hw_fini

The call trace occurs when the amdgpu is removed after
the mode1 reset. During mode1 reset, from suspend to resume,
there is no need to reinitialize the ta firmware buffer
which caused the bo pin_count increase redundantly.

[  489.885525] Call Trace:
[  489.885525]  <TASK>
[  489.885526]  amdttm_bo_put+0x34/0x50 [amdttm]
[  489.885529]  amdgpu_bo_free_kernel+0xe8/0x130 [amdgpu]
[  489.885620]  psp_free_shared_bufs+0xb7/0x150 [amdgpu]
[  489.885720]  psp_hw_fini+0xce/0x170 [amdgpu]
[  489.885815]  amdgpu_device_fini_hw+0x2ff/0x413 [amdgpu]
[  489.885960]  ? blocking_notifier_chain_unregister+0x56/0xb0
[  489.885962]  amdgpu_driver_unload_kms+0x51/0x60 [amdgpu]
[  489.886049]  amdgpu_pci_remove+0x5a/0x140 [amdgpu]
[  489.886132]  ? __pm_runtime_resume+0x60/0x90
[  489.886134]  pci_device_remove+0x3e/0xb0
[  489.886135]  __device_release_driver+0x1ab/0x2a0
[  489.886137]  driver_detach+0xf3/0x140
[  489.886138]  bus_remove_driver+0x6c/0xf0
[  489.886140]  driver_unregister+0x31/0x60
[  489.886141]  pci_unregister_driver+0x40/0x90
[  489.886142]  amdgpu_exit+0x15/0x451 [amdgpu]

Signed-off-by: Horatio Zhang <Hongkun.Zhang@amd.com>
Signed-off-by: longlyao <Longlong.Yao@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: remove unused variable ring

building with gcc and W=1 reports
drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c:81:29: error: variable
  ‘ring’ set but not used [-Werror=unused-but-set-variable]
   81 |         struct amdgpu_ring *ring;
      |                             ^~~~

ring is not used so remove it.

Signed-off-by: Tom Rix <trix@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix dm irq error message in gpu recover

[Why]
Variable adev->crtc_irq.num_types was initialized as the value of
adev->mode_info.num_crtc at early_init stage, later at hw_init stage,
the num_crtc changed due to the display pipe harvest on some SKUs,
but the num_types was not updated accordingly, that cause below error
in gpu recover.

  *ERROR* amdgpu_dm_set_crtc_irq_state: crtc is NULL at id :3
  *ERROR* amdgpu_dm_set_crtc_irq_state: crtc is NULL at id :3
  *ERROR* amdgpu_dm_set_crtc_irq_state: crtc is NULL at id :3
  *ERROR* amdgpu_dm_set_pflip_irq_state: crtc is NULL at id :3
  *ERROR* amdgpu_dm_set_pflip_irq_state: crtc is NULL at id :3
  *ERROR* amdgpu_dm_set_pflip_irq_state: crtc is NULL at id :3
  *ERROR* amdgpu_dm_set_pflip_irq_state: crtc is NULL at id :3
  *ERROR* amdgpu_dm_set_vupdate_irq_state: crtc is NULL at id :3
  *ERROR* amdgpu_dm_set_vupdate_irq_state: crtc is NULL at id :3
  *ERROR* amdgpu_dm_set_vupdate_irq_state: crtc is NULL at id :3

[How]
Defer the initialization of num_types to eliminate the error logs.

Signed-off-by: tiancyin <tianci.yin@amd.com>
Reviewed-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Fix initialization for nbio 7.5.1

A mistake has been made in the BIOS for some ASICs with NBIO 7.5.1
where some NBIO registers aren't properly setup.

Ensure that they're set during initialization.

Tested-by: Richard Gong <richard.gong@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Format input and output CSC matrix

Format the input and output CSC matrix so they
look like 3x4 matrixes. This will make parsing them
much easier and allows us to quickly spot potential
mistakes.

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Cc: Pekka Paalanen <ppaalanen@gmail.com>
Cc: Sebastian Wick <sebastian.wick@redhat.com>
Cc: Vitaly.Prosyak@amd.com
Cc: Joshua Ashton <joshua@froggi.es>
Cc: dri-devel@lists.freedesktop.org
Cc: amd-gfx@lists.freedesktop.org
Reviewed-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Don't restrict bpc to 8 bpc

This will let us pass the kms_hdr.bpc_switch IGT
test.

The reason the bpc restriction was required is
historical. At one point in time we were not falling
back to a lower bpc when we didn't have enough
bandwidth for the maximum bpc reported by a display.
This meant that we couldn't enable some high refresh
modes unless we limitted the bpc.

Starting with this patch the issue is fixed:
commit cbd14ae7ea93 ("drm/amd/display: Fix
incorrectly pruned modes with deep color")

This patch implemented a fallback mechanism if mode
validation failed at the max bpc. This means users
now automatically get all modes that can be supported
by at least 6 bpc. The driver will enable the mode
with the highest possible bpc that is supported by
the display.

v2:
 - explain why this is no longer needed (Michel)
 - refer to commit that fixed bpc fallback (Michel)

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Cc: Pekka Paalanen <ppaalanen@gmail.com>
Cc: Sebastian Wick <sebastian.wick@redhat.com>
Cc: Vitaly.Prosyak@amd.com
Cc: Joshua Ashton <joshua@froggi.es>
Cc: dri-devel@lists.freedesktop.org
Cc: amd-gfx@lists.freedesktop.org
Cc: Michel Dänzer <michel.daenzer@mailbox.org>
Reviewed-by: Joshua Ashton <joshua@froggi.es>
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Make umc_v8_10_convert_error_address static and remove unused variable

Fixes following warnings:
warning: no previous prototype for 'umc_v8_10_convert_error_address'
warning: variable 'channel_index' set but not used

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

drm/amdgpu: added a sysfs interface for thermal throttling

implement apu_thermal_cap r/w callback for vangogh

Signed-off-by: Kun Liu <Kun.Liu2@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: added a sysfs interface for thermal throttling

added a sysfs interface for thermal throttling, then userspace
can get/update thermal limit

Signed-off-by: Kun Liu <Kun.Liu2@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove unused local variables and function

Remove a couple of local variables that are only set but never used,
also remove an static utility function that is never used in consequence
of the variable removal.

This decrease the number of -Wunused-but-set-variable warnings.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Arthur Grillo <arthurgrillo@riseup.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove unused local variables

Remove local variables that were just set but were never used. This
decrease the number of -Wunused-but-set-variable warnings.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Arthur Grillo <arthurgrillo@riseup.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: Fix eDP for single-display iMac11,2

Apple iMac11,2 (mid 2010) also with Radeon HD-4670 that has the same
issue as iMac10,1 (late 2009) where the internal eDP panel stays dark on
driver load.  This patch treats iMac11,2 the same as iMac10,1,
so the eDP panel stays active.

Additional steps:
Kernel boot parameter radeon.nomodeset=0 required to keep the eDP
panel active.

This patch is an extension of
commit 564d8a2cf3ab ("drm/radeon: Fix eDP for single-display iMac10,1 (v2)")
Link: https://lore.kernel.org/all/lsq.1507553064.833262317@decadent.org.uk/
Signed-off-by: Mark Hawrylak <mark.hawrylak@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Make kobj_type structures constant

Since commit ee6d3dd4ed48 ("driver core: make kobj_type constant.")
the driver core allows the usage of const struct kobj_type.

Take advantage of this to constify the structure definitions to prevent
modification at runtime.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: make kobj_type structures constant

Since commit ee6d3dd4ed48 ("driver core: make kobj_type constant.")
the driver core allows the usage of const struct kobj_type.

Take advantage of this to constify the structure definitions to prevent
modification at runtime.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Modify mismatched function name

No functional modification involved.

drivers/gpu/drm/amd/amdgpu/../display/dc/link/link_detection.c:1199: warning: expecting prototype for dc_link_detect_connection_type(). Prototype was for link_detect_connection_type() instead.

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=4103
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Pass proper parent for DM backlight device registration

The parent for the backlight device should be the drm-connector object,
not the PCI device.

Userspace relies on this to be able to detect which backlight class device
to use on hybrid gfx devices where there may be multiple native (raw)
backlight devices registered.

Specifically gnome-settings-daemon expects the parent device to have
an "enabled" sysfs attribute (as drm_connector devices do) and tests
that this returns "enabled" when read.

This aligns the parent of the backlight device with i915, nouveau, radeon.
Note that drivers/gpu/drm/amd/amdgpu/atombios_encoders.c also already
uses the drm_connector as parent, only amdgpu_dm.c used the PCI device
as parent before this change.

Note this is marked as a RFC because I don't have hw to test, so this
has only been compile tested! If someone can test this on actual
hw which hits the changed code path that would be great.

Link: https://gitlab.gnome.org/GNOME/gnome-settings-daemon/-/issues/730
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: downgrade log level upon SMU IF version mismatch

SMU IF version mismatch as a warning message exists widely
after asic production, however, due to this log level setting,
such mismatch warning will be caught by automation test like
IGT and reported as a fake error after checking. As such mismatch
does not break anything, to reduce confusion, downgrade it from
dev_warn to dev_info.

Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add ecc info query interface for umc v8_10

Support ecc info query for umc v8_10.

v2: Simplied by convert_error_address.
v3: Remove unused variable and invalid checking.

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

drm/amdgpu: Add convert_error_address function for umc v8_10

Add convert_error_address for umc v8_10.

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

drm/amdgpu: add bad_page_threshold check in ras_eeprom_check_err

bad_page_threshold controls page retirement behavior and it should be
also checked.

v2: simplify the condition of bad page handling path.

Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Stanley.Yang <Stanley.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: change default behavior of bad_page_threshold parameter

Ignore ras umc bad page threshold by default, GPU initialization won't
be stopped in this mode.

v2: refine the description of bad_page_threshold.

Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Stanley.Yang <Stanley.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: exclude duplicate pages from UMC RAS UE count

If a UMC bad page is reserved but not freed by an application, the
application may trigger uncorrectable error repeatly by accessing the page.

v2: add specific function to do the check.
v3: remove duplicate pages, calculate new added bad page number.
v4: reuse save_bad_pages to calculate new added bad page number.

Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Stanley.Yang <Stanley.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add umc retire unit element

It records how many bad pages are retired in one uncorrectable error.

Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Stanley.Yang <Stanley.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: no pptable resetup on runpm exiting

It is assumed the pptable used before runpm is same as
the one used afterwards. Thus, we can reuse the stored
copy and do not need to resetup the pptable again.

Signed-off-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Feifei Xu <feifei.xu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: correct the baco state setting for ArmD3 scenario

The check for baco support relies on the correct baco state.

Signed-off-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Feifei Xu <feifei.xu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add more fields into device info, caches sizes, etc.

AMDGPU_IDS_FLAGS_CONFORMANT_TRUNC_COORD: important for conformance on gfx11
Other fields are exposed from IP discovery.
enabled_rb_pipes_mask_hi is added for future chips, currently 0.

Mesa MR: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21403

Signed-off-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Fix an illegal memory access

In the kfd_wait_on_events() function, the kfd_event_waiter structure is
allocated by alloc_event_waiters(), but the event field of the waiter
structure is not initialized; When copy_from_user() fails in the
kfd_wait_on_events() function, it will enter exception handling to
release the previously allocated memory of the waiter structure;
Due to the event field of the waiters structure being accessed
in the free_waiters() function, this results in illegal memory access
and system crash, here is the crash log:

localhost kernel: RIP: 0010:native_queued_spin_lock_slowpath+0x185/0x1e0
localhost kernel: RSP: 0018:ffffaa53c362bd60 EFLAGS: 00010082
localhost kernel: RAX: ff3d3d6bff4007cb RBX: 0000000000000282 RCX: 00000000002c0000
localhost kernel: RDX: ffff9e855eeacb80 RSI: 000000000000279c RDI: ffffe7088f6a21d0
localhost kernel: RBP: ffffe7088f6a21d0 R08: 00000000002c0000 R09: ffffaa53c362be64
localhost kernel: R10: ffffaa53c362bbd8 R11: 0000000000000001 R12: 0000000000000002
localhost kernel: R13: ffff9e7ead15d600 R14: 0000000000000000 R15: ffff9e7ead15d698
localhost kernel: FS:  0000152a3d111700(0000) GS:ffff9e855ee80000(0000) knlGS:0000000000000000
localhost kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
localhost kernel: CR2: 0000152938000010 CR3: 000000044d7a4000 CR4: 00000000003506e0
localhost kernel: Call Trace:
localhost kernel: _raw_spin_lock_irqsave+0x30/0x40
localhost kernel: remove_wait_queue+0x12/0x50
localhost kernel: kfd_wait_on_events+0x1b6/0x490 [hydcu]
localhost kernel: ? ftrace_graph_caller+0xa0/0xa0
localhost kernel: kfd_ioctl+0x38c/0x4a0 [hydcu]
localhost kernel: ? kfd_ioctl_set_trap_handler+0x70/0x70 [hydcu]
localhost kernel: ? kfd_ioctl_create_queue+0x5a0/0x5a0 [hydcu]
localhost kernel: ? ftrace_graph_caller+0xa0/0xa0
localhost kernel: __x64_sys_ioctl+0x8e/0xd0
localhost kernel: ? syscall_trace_enter.isra.18+0x143/0x1b0
localhost kernel: do_syscall_64+0x33/0x80
localhost kernel: entry_SYSCALL_64_after_hwframe+0x44/0xa9
localhost kernel: RIP: 0033:0x152a4dff68d7

Allocate the structure with kcalloc, and remove redundant 0-initialization
and a redundant loop condition check.

Signed-off-by: Qu Huang <qu.huang@linux.dev>
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/vcn: set and use harvest config

in early init to set harvest config if the vcn0/1 is disabled
rather than hard-code the ring attributes as before did

Signed-off-by: Jane Jian <Jane.Jian@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Don't allow s0ix on APUs older than Raven

APUs before Raven didn't support s0ix.  As we just relieved some
of the safety checks for s0ix to improve power consumption on
APUs that support it but that are missing BIOS support a new
blind spot was introduced that a user could "try" to run s0ix.

Plug this hole so that if users try to run s0ix on anything older
than Raven it will just skip suspend of the GPU.

Fixes: cf488dcd0ab7 ("drm/amd: Allow s0ix without BIOS support")
Suggested-by: Alexander Deucher <Alexander.Deucher@amd.com>
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/amdgpu: fix incorrect active rb bitmap for gfx11

GFX v11 changes RB_BACKEND_DISABLE related registers
from per SA to global ones. The approach to query active
rb bitmap needs to be changed accordingly. Query per
SE setting returns wrong active RB bitmap especially
in the case when some of SA are disabled. With the new
approach, driver will generate the active rb bitmap
based on active SA bitmap and global active RB bitmap.

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: optimize VRAM allocation when using drm buddy

Since the VRAM manager changed from drm mm to drm buddy. It's
not necessary to allocate 2MB aligned VRAM for more than 2MB
unaligned size, and then do trim. This method improves the
allocation efficiency and reduces memory fragmentation.

v2: Correct the remainder operation

Signed-off-by: Shane Xiao <shane.xiao@amd.com>
Reviewed-by: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: remove TOPDOWN flags when allocating VRAM in large bar system

Since VRAM manager is changed from drm mm to drm buddy, the
TOP_DOWN flag should not be set by default in the large bar system.
Removing this flag helps improve drm buddy allocator efficiency and
reduce the risk of splitting higher order block into lower order.

Signed-off-by: Shane Xiao <shane.xiao@amd.com>
Reviewed-by: Christian K�nig <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Select DRM_DISPLAY_HDCP_HELPER in amdgpu

Keeps this selection with the rest of the DRM HELPER
selection.

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

drm/amd/display: only warn once in dce110_edp_wait_for_hpd_ready()

Since, hot plugging eDP displays isn't supported, it is sufficient for
us to warn about the lack of a connected display once. So, use ASSERT()
in dce110_edp_wait_for_hpd_ready() instead of DC_LOG_WARNING().

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: re-enable ac/dc on smu_v13_0_0/10

re-enable ac/dc on smu_v13_0_0/10

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

drm/amdgpu: add tmz support for GC 10.3.6

this patch to add tmz support for GC 10.3.6

Signed-off-by: Jesse Zhang <Jesse.Zhang@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: To fix sdma page fault issue for GC 11

For the MQD memory, KMD would always allocate 4K memory,
and mes scheduler would write to the end of MQD for unmap flag.

Signed-off-by: Ruili Ji <ruiliji2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

msm/fbdev: fix unused variable warning with clang.

clang builds showed this:
drivers/gpu/drm/msm/msm_fbdev.c:144:6: error: variable 'helper' is used uninitialized whenever 'if' condition is true [-Werror,-Wsometimes-uninitialized]
        if (!fbdev)
            ^~~~~~

Fixes: 3fb1f62f80a1 ("drm/fb-helper: Remove drm_fb_helper_unprepare() from drm_fb_helper_fini()")
Signed-off-by: Dave Airlie <airlied@redhat.com>

Merge tag 'drm-misc-next-fixes-2023-02-21' of git://anongit.freedesktop.org/drm/drm-misc into drm-next

Short summary of fixes pull:

Fixes GEM SHMEM locking and generic fbdev hotplugging. Constifies
dma_buf kobj type.

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/Y/S6tu3gdQ0VizR+@linux-uq9g

drm/fb-helper: Remove drm_fb_helper_unprepare() from drm_fb_helper_fini()

Move drm_fb_helper_unprepare() from drm_fb_helper_fini() into the
calling fbdev implementation. Avoids a possible stale mutex with
generic fbdev code.

As indicated by its name, drm_fb_helper_prepare() prepares struct
drm_fb_helper before setting up the fbdev support with a call to
drm_fb_helper_init(). In legacy fbdev emulation, this happens next
to each other. If successful, drm_fb_helper_fini() later tear down
the fbdev device and also unprepare via drm_fb_helper_unprepare().

Generic fbdev emulation prepares struct drm_fb_helper immediately
after allocating the instance. It only calls drm_fb_helper_init()
as part of processing a hotplug event. If the hotplug-handling fails,
it runs drm_fb_helper_fini(). This unprepares the fb-helper instance
and the next hotplug event runs on stale data.

Solve this by moving drm_fb_helper_unprepare() from drm_fb_helper_fini()
into the fbdev implementations. Call it right before freeing the
fb-helper instance.

Fixes: 643231b28380 ("drm/fbdev-generic: Minimize hotplug error handling")
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Javier Martinez Canillas <javierm@redhat.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: David Airlie <airlied@gmail.com>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230216140620.17699-1-tzimmermann@suse.de

Merge tag 'drm-intel-next-fixes-2023-02-17' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

drm/i915 fixes for the v6.3 merge window:
- Fix eDP+DSI dual panel systems
- Fix system suspend when fbdev isn't initialized
- Fix memory leaks in scatterlist
- Fix some MCR register annotations
- Fix documentation build warnings

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/87v8k0xyx4.fsf@intel.com

Merge tag 'amd-drm-next-6.3-2023-02-17' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-6.3-2023-02-17:

amdgpu:
- GC 11 fixes
- Display fixes
- Backlight cleanup
- SMU13 fixes
- SMU7 regression fix
- GFX9 sw queues fix
- AGP fix for GMC 11
- W1 warning fixes
- S/G display fixes
- Misc spelling fixes
- Driver unload fix
- DCN 3.1.4 fixes
- Display code reorg fixes
- Rotation fixes

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230217230930.64821-1-alexander.deucher@amd.com

Merge tag 'drm-misc-next-fixes-2023-02-16' of git://anongit.freedesktop.org/drm/drm-misc into drm-next

Short summary of fixes pull:

Contains fixes for DP MST and the panel orientation on an Lenovo
IdeaPad model.

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/Y+4H4C4E6cZcM9+J@linux-uq9g

dma-buf: make kobj_type structure constant

Since commit ee6d3dd4ed48 ("driver core: make kobj_type constant.")
the driver core allows the usage of const struct kobj_type.

Take advantage of this to constify the structure definition to prevent
modification at runtime.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20230217-kobj_type-dma-buf-v1-1-b84a3616522c@weissschuh.net

drm/shmem-helper: Fix locking for drm_gem_shmem_get_pages_sgt()

Other functions touching shmem->sgt take the pages lock, so do that here
too. drm_gem_shmem_get_pages() & co take the same lock, so move to the
_locked() variants to avoid recursive locking.

Discovered while auditing locking to write the Rust abstractions.

Fixes: 2194a63a818d ("drm: Add library for shmem backed GEM objects")
Fixes: 4fa3d66f132b ("drm/shmem: Do dma_unmap_sg before purging pages")
Signed-off-by: Asahi Lina <lina@asahilina.net>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230205125124.2260-1-lina@asahilina.net
(cherry picked from commit aa8c85affe3facd3842c8912186623415931cc72)
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>

Merge tag 'drm-misc-next-fixes-2023-02-09' of git://anongit.freedesktop.org/drm/drm-misc into drm-next

Short summary of fixes pull:

Contains a number of fixes to vc4 and ivpu. The patches to the probe
helpers were cherry-picked from the regular development branch.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
From: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/Y+S6HBmaRJNPYiBG@linux-uq9g

drm/amd/display: disable SubVP + DRR to prevent underflow

[Why&How]
Temporarily disable SubVP+DRR since Xorg has an architectural limitation
where freesync will not work in a multi monitor configuration. SubVP+DRR
requires that freesync be working.

Whether OS has variable refresh setting enabled or not, the state on
the crtc remains same unless an application requests VRR. Due to this,
there is no way to know whether freesync will actually work or not
while we are on the desktop from the kernel's perspective.

If userspace does not have a limitation with multi-display freesync (for
example wayland), then this feature can be enabled by adding a
dcfeaturemask option to amdgpu on the kernel cmdline like:

amdgpu.dcfeaturemask=0x200

Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fail atomic_check early on normalize_zpos error

[Why]

drm_atomic_normalize_zpos() can return an error code when there's
modeset lock contention. This was being ignored.

[How]

Bail out of atomic check if normalize_zpos() returns an error.

Fixes: b261509952bc ("drm/amd/display: Fix double cursor on non-video RGB MPO")
Signed-off-by: Leo Li <sunpeng.li@amd.com>
Tested-by: Mikhail Gavrilov <mikhail.v.gavrilov@gmail.com>
Reviewed-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>