linux-stable.git/drivers/gpu/drm/amd, branch v5.4.92

drm/amdgpu: fix a GPU hang issue when remove device

2021-01-19T17:26:16+00:00

[ Upstream commit 88e21af1b3f887d217f2fb14fc7e7d3cd87ebf57 ]

When GFXOFF is enabled and GPU is idle, driver will fail to access some
registers. Therefore change to disable power gating before all access
registers with MMIO.

Dmesg log is as following:
amdgpu 0000:03:00.0: amdgpu: amdgpu: finishing device.
amdgpu: cp queue pipe 4 queue 0 preemption failed
amdgpu 0000:03:00.0: amdgpu: failed to write reg 2890 wait reg 28a2
amdgpu 0000:03:00.0: amdgpu: failed to write reg 1a6f4 wait reg 1a706
amdgpu 0000:03:00.0: amdgpu: failed to write reg 2890 wait reg 28a2
amdgpu 0000:03:00.0: amdgpu: failed to write reg 1a6f4 wait reg 1a706

Signed-off-by: Dennis Li 
Reviewed-by: Hawking Zhang 
Signed-off-by: Alex Deucher 
Signed-off-by: Sasha Levin

Revert "drm/amd/display: Fix memory leaks in S3 resume"

2021-01-09T12:44:53+00:00

This reverts commit a135a1b4c4db1f3b8cbed9676a40ede39feb3362.

This leads to blank screens on some boards after replugging a
display.  Revert until we understand the root cause and can
fix both the leak and the blank screen after replug.

Cc: Stylon Wang 
Cc: Harry Wentland 
Cc: Nicholas Kazlauskas 
Cc: Andre Tomt 
Cc: Oleksandr Natalenko 
Signed-off-by: Alex Deucher 
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman

drm/amd/display: Fix memory leaks in S3 resume

2020-12-30T10:51:40+00:00

commit a135a1b4c4db1f3b8cbed9676a40ede39feb3362 upstream.

EDID parsing in S3 resume pushes new display modes
to probed_modes list but doesn't consolidate to actual
mode list. This creates a race condition when
amdgpu_dm_connector_ddc_get_modes() re-initializes the
list head without walking the list and results in  memory leak.

Bug: https://bugzilla.kernel.org/show_bug.cgi?id=209987
Acked-by: Harry Wentland 
Acked-by: Alex Deucher 
Reviewed-by: Nicholas Kazlauskas 
Signed-off-by: Stylon Wang 
Signed-off-by: Alex Deucher 
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman

drm/amdkfd: Fix leak in dmabuf import

2020-12-30T10:51:30+00:00

[ Upstream commit c897934da15f182ce99536007f8ef61c4748c07e ]

Release dmabuf reference before returning from kfd_ioctl_import_dmabuf.
amdgpu_amdkfd_gpuvm_import_dmabuf takes a reference to the underlying
GEM BO and doesn't keep the reference to the dmabuf wrapper.

Signed-off-by: Felix Kuehling 
Reviewed-by: Kent Russell 
Signed-off-by: Alex Deucher 
Signed-off-by: Sasha Levin

drm/amd/display: Prevent bandwidth overflow

2020-12-30T10:51:30+00:00

[ Upstream commit 80089dd8410f356d5104496d5ab71a66a4f4646b ]

[Why]
At very high pixel clock, bandwidth calculation exceeds 32 bit size
and overflow value. This causes the resulting selection of link rate
to be inaccurate.

[How]
Change order of operation and use fixed point to deal with integer
accuracy. Also address bug found when forcing link rate.

Signed-off-by: Chris Park 
Reviewed-by: Wenjing Liu 
Acked-by: Eryk Brol 
Signed-off-by: Alex Deucher 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

drm/amdgpu: fix build_coefficients() argument

2020-12-30T10:51:04+00:00

[ Upstream commit dbb60031dd0c2b85f10ce4c12ae604c28d3aaca4 ]

gcc -Wextra warns about a function taking an enum argument
being called with a bool:

drivers/gpu/drm/amd/amdgpu/../display/modules/color/color_gamma.c: In function 'apply_degamma_for_user_regamma':
drivers/gpu/drm/amd/amdgpu/../display/modules/color/color_gamma.c:1617:29: warning: implicit conversion from 'enum ' to 'enum dc_transfer_func_predefined' [-Wenum-conversion]
 1617 |  build_coefficients(&coeff, true);

It appears that a patch was added using the old calling conventions
after the type was changed, and the value should actually be 0
(TRANSFER_FUNCTION_SRGB) here instead of 1 (true).

Fixes: 55a01d4023ce ("drm/amd/display: Add user_regamma to color module")
Reviewed-by: Harry Wentland 
Signed-off-by: Arnd Bergmann 
Signed-off-by: Alex Deucher 
Signed-off-by: Sasha Levin

drm/amdgpu: fix incorrect enum type

2020-12-30T10:51:03+00:00

[ Upstream commit a110f3750bf8b93764f13bd1402c7cba03d15d61 ]

core_link_write_dpcd() returns enum dc_status, not ddc_result:

display/dc/core/dc_link_dp.c: In function 'dp_set_panel_mode':
display/dc/core/dc_link_dp.c:4237:11: warning: implicit conversion from 'enum dc_status' to 'enum ddc_result'
[-Wenum-conversion]

Avoid the warning by using the correct enum in the caller.

Fixes: 0b226322434c ("drm/amd/display: Synchronous DisplayPort Link Training")
Signed-off-by: Arnd Bergmann 
Signed-off-by: Alex Deucher 
Signed-off-by: Sasha Levin

drm/amd/display: Init clock value by current vbios CLKs

2020-12-30T10:50:57+00:00

[ Upstream commit 7e0b367db85ef7b91399006253759a024eab7653 ]

[Why]
While booting into OS, driver updates DPP/DISP CLKs.
But init clock value is zero which is invalid.

[How]
Get current clocks value to update init clocks.
To avoid underflow.

Signed-off-by: Brandon Syu 
Reviewed-by: Tony Cheng 
Acked-by: Aurabindo Pillai 
Signed-off-by: Alex Deucher 
Signed-off-by: Sasha Levin

drm/amd/display: Add missing pflip irq for dcn2.0

2020-11-24T12:29:23+00:00

commit 728321e53045d2668bf2b8627a8d61bc2c480d3b upstream.

If we have more than 4 displays we will run
into dummy irq calls or flip timout issues.

Reviewed-by: Nicholas Kazlauskas 
Signed-off-by: Alex Deucher 
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman

amd/amdgpu: Disable VCN DPG mode for Picasso

2020-11-18T18:20:25+00:00

[ Upstream commit c6d2b0fbb893d5c7dda405aa0e7bcbecf1c75f98 ]

Concurrent operation of VCN and JPEG decoder in DPG mode is
causing ring timeout due to power state.

Signed-off-by: Veerabadhran Gopalakrishnan 
Reviewed-by: Leo Liu 
Signed-off-by: Alex Deucher 
Signed-off-by: Sasha Levin