aboutsummaryrefslogtreecommitdiff
path: root/drivers/gpu/drm
AgeCommit message (Collapse)AuthorFilesLines
2024-05-29drm/amdgpu: Make CPX mode auto default in NPS4Rajneesh Bhardwaj1-1/+1
On GFXIP9.4.3, make CPX mode as the default compute mode if the node is setup in NPS4 memory partition mode. This change is only applicable for dGPU, for APU, continue to use TPX mode. Reviewed-by: Felix Kuehling <[email protected]> Signed-off-by: Rajneesh Bhardwaj <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdkfd: simplify APU VRAM handlingAlex Deucher4-14/+11
With commit 89773b85599a ("drm/amdkfd: Let VRAM allocations go to GTT domain on small APUs") big and small APU "VRAM" handling in KFD was unified. Since AMD_IS_APU is set for both big and small APUs, we can simplify the checks in the code. v2: clean up a few more places (Lang) Acked-by: Felix Kuehling <[email protected]> Reviewed-by: Lang Yu <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29Revert "drm/amdkfd: fix gfx_target_version for certain 11.0.3 devices"Alex Deucher1-9/+2
This reverts commit 28ebbb4981cb1fad12e0b1227dbecc88810b1ee8. Revert this commit as apparently the LLVM code to take advantage of this never landed. Reviewed-by: Feifei Xu <[email protected]> Signed-off-by: Alex Deucher <[email protected]> Cc: Feifei Xu <[email protected]>
2024-05-29drm/amdgpu: fix dereference null return value for the function ↵Jesse Zhang1-1/+5
amdgpu_vm_pt_parent The pointer parent may be NULLed by the function amdgpu_vm_pt_parent. To make the code more robust, check the pointer parent. Signed-off-by: Jesse Zhang <[email protected]> Suggested-by: Christian König <[email protected]> Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdgpu: silence UBSAN warningAlex Deucher1-1/+1
Convert a variable sized array from [1] to []. Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdgpu: Adjust logic in amdgpu_device_partner_bandwidth()Alex Deucher1-7/+12
Use current speed/width on devices which don't support dynamic PCIe switching. Fixes: 466a7d115326 ("drm/amd: Use the first non-dGPU PCI device for BW limits") Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/3289 Acked-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: Move 'struct scaler_data' off stackArnd Bergmann1-25/+31
The scaler_data structure is implicitly copied onto the stack twice by being returned from a function. This is usually a bad idea, but it was not flagged by the compiler until a recent addition that pushed it over the 1024 byte function stack limit: drivers/gpu/drm/amd/amdgpu/../display/dc/dml2/dml2_translation_helper.c: In function 'populate_dml_plane_cfg_from_plane_state': drivers/gpu/drm/amd/amdgpu/../display/dc/dml2/dml2_translation_helper.c:1075:1: error: the frame size of 1032 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] Use an explicit kzalloc() and memcpy() instead here to keep it off the stack. Fixes: 00c391102abc ("drm/amd/display: Add misc DC changes for DCN401") Fixes: 7966f319c66d ("drm/amd/display: Introduce DML2") Signed-off-by: Arnd Bergmann <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: avoid large on-stack structuresArnd Bergmann2-10/+22
Putting excessively large objects on a function stack causes a warning about possibly overflowing the 8KiB of kernel stack: drivers/gpu/drm/amd/amdgpu/../display/dc/resource/dcn401/dcn401_resource.c: In function 'dcn401_update_bw_bounding_box': drivers/gpu/drm/amd/amdgpu/../display/dc/resource/dcn401/dcn401_resource.c:1599:1: error: the frame size of 1196 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] 1599 | } | ^ drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_state.c: In function 'dc_state_create': drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_state.c:221:1: error: the frame size of 1196 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] 221 | } | ^ Use dynamic allocation instead. Fixes: e779f4587f61 ("drm/amd/display: Add handling for DC power mode") Signed-off-by: Arnd Bergmann <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: fix graphics_object_id sizeArnd Bergmann1-2/+2
The graphics_object_id structure is meant to fit into 32 bits, as it's passed by value in and out of functions. A recent change increased the size to 128 bits, so it's now always passed by reference, which is clearly not intended and ends up producing a compile-time warning: drivers/gpu/drm/amd/amdgpu/../display/dc/link/link_factory.c: In function 'construct_phy': drivers/gpu/drm/amd/amdgpu/../display/dc/link/link_factory.c:743:1: error: the frame size of 1040 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] Add back the bitfields to revert to the original size, while keeping the 'enum' type change. Fixes: fec85f995a4b ("drm/amd/display: Fix compiler redefinition warnings for certain configs") Signed-off-by: Arnd Bergmann <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: dynamically allocate dml2_configuration_options structuresArnd Bergmann2-10/+22
This structure is too large to fit on a stack, as shown by the newly introduced warnings from a recent code change: drivers/gpu/drm/amd/amdgpu/../display/dc/resource/dcn32/dcn32_resource.c: In function 'dcn32_update_bw_bounding_box': drivers/gpu/drm/amd/amdgpu/../display/dc/resource/dcn32/dcn32_resource.c:2019:1: error: the frame size of 1180 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] drivers/gpu/drm/amd/amdgpu/../display/dc/resource/dcn321/dcn321_resource.c: In function 'dcn321_update_bw_bounding_box': drivers/gpu/drm/amd/amdgpu/../display/dc/resource/dcn321/dcn321_resource.c:1597:1: error: the frame size of 1180 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_state.c: In function 'dc_state_create': drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_state.c:219:1: error: the frame size of 1184 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] Instead of open-coding the assignment of a large structure to a stack variable, use an explicit kmemdup() in each case to move it off the stack. Fixes: e779f4587f61 ("drm/amd/display: Add handling for DC power mode") Signed-off-by: Arnd Bergmann <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drivers/gpu: Fix misalignment in comment blockBruno Rocha Levi1-1/+1
This patch fixes a warning from checkpatch by ensuring the trailing */ is aligned with the rest of the *, improving readability. Co-developed-by: Lucas Antonio <[email protected]> Signed-off-by: Lucas Antonio <[email protected]> Signed-off-by: Bruno Rocha Levi <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdkfd: Replace deprecated gfx12 trap handler instructionsJay Cornwall2-95/+97
Newer assemblers reject S_WAITCNT. All instances of S_WAITCNT can be replaced by S_WAITCNT 0 (< gfx12) or S_WAIT_IDLE (>= gfx12) since there is no concurrency of different memory instruction classes. Signed-off-by: Jay Cornwall <[email protected]> Reviewed-by: Lancelot Six <[email protected]> Acked-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdkfd: Sync trap handler binary with sourceJay Cornwall1-33/+24
Source and binary have become mismatched during branch activity. Signed-off-by: Jay Cornwall <[email protected]> Reviewed-by: Lancelot Six <[email protected]> Acked-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd: Fix shutdown (again) on some SMU v13.0.4/11 platformsMario Limonciello1-9/+11
commit cd94d1b182d2 ("dm/amd/pm: Fix problems with reboot/shutdown for some SMU 13.0.4/13.0.11 users") attempted to fix shutdown issues that were reported since commit 31729e8c21ec ("drm/amd/pm: fixes a random hang in S4 for SMU v13.0.4/11") but caused issues for some people. Adjust the workaround flow to properly only apply in the S4 case: -> For shutdown go through SMU_MSG_PrepareMp1ForUnload -> For S4 go through SMU_MSG_GfxDeviceDriverReset and SMU_MSG_PrepareMp1ForUnload Reported-and-tested-by: lectrode <[email protected]> Closes: https://github.com/void-linux/void-packages/issues/50417 Cc: [email protected] Fixes: cd94d1b182d2 ("dm/amd/pm: Fix problems with reboot/shutdown for some SMU 13.0.4/13.0.11 users") Reviewed-by: Tim Huang <[email protected]> Signed-off-by: Mario Limonciello <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdgpu/gfx11: enable gfx pipe1 hardware supportAlex Deucher1-3/+3
Enable gfx pipe1 hardware support. Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdgpu: Make CPX mode auto default in NPS4Rajneesh Bhardwaj1-1/+1
On GFXIP9.4.3, make CPX mode as the default compute mode if the node is setup in NPS4 memory partition mode. This change is only applicable for dGPU, for APU, continue to use TPX mode. Reviewed-by: Felix Kuehling <[email protected]> Signed-off-by: Rajneesh Bhardwaj <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdgpu/gfx11: handle priority setup for gfx pipe1Alex Deucher1-11/+25
Set up pipe1 as a high priority queue. Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdgpu/gfx11: select HDP ref/mask according to gfx ring pipeAlex Deucher1-1/+1
Use correct ref/mask for differnent gfx ring pipe. Ported from ZhenGuo's patch for gfx10. Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdgpu: Add lock around VF RLCG interfaceVictor Skvortsov3-0/+9
flush_gpu_tlb may be called from another thread while device_gpu_recover is running. Both of these threads access registers through the VF RLCG interface during VF Full Access. Add a lock around this interface to prevent race conditions between these threads. Signed-off-by: Victor Skvortsov <[email protected]> Reviewed-by: Zhigang Luo <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdkfd: simplify APU VRAM handlingAlex Deucher4-14/+11
With commit 89773b85599a ("drm/amdkfd: Let VRAM allocations go to GTT domain on small APUs") big and small APU "VRAM" handling in KFD was unified. Since AMD_IS_APU is set for both big and small APUs, we can simplify the checks in the code. v2: clean up a few more places (Lang) Acked-by: Felix Kuehling <[email protected]> Reviewed-by: Lang Yu <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29Revert "drm/amdkfd: fix gfx_target_version for certain 11.0.3 devices"Alex Deucher1-9/+2
This reverts commit 28ebbb4981cb1fad12e0b1227dbecc88810b1ee8. Revert this commit as apparently the LLVM code to take advantage of this never landed. Reviewed-by: Feifei Xu <[email protected]> Signed-off-by: Alex Deucher <[email protected]> Cc: Feifei Xu <[email protected]>
2024-05-29drm/amd/display: Enable copying of bounding box data from VBIOS DMUBAurabindo Pillai7-24/+146
Allocate some memory, send the address in chunks to dmub, and finally ask it to copy the bounding box data into the newly allocated memory. Signed-off-by: Aurabindo Pillai <[email protected]> Acked-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: Add new GPINT command definitionsAurabindo Pillai1-0/+31
New commands for enabling copy of DC bounding box values from VBIOS DMUB Signed-off-by: Aurabindo Pillai <[email protected]> Acked-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: Move some init routines to dm_sw_init()Aurabindo Pillai1-9/+11
Move creation of CGS device node and the DAL allocation list from amdgpu_dm_init() to dm_sw_init() which runs before dmub's sw init hook. This is required for communicating with the VBIOS DMUB image from the VBIOS that was loaded for early pre-os boot. Signed-off-by: Aurabindo Pillai <[email protected]> Acked-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: Add missing registers for DCN401Rodrigo Siqueira1-1/+50
Add some additional registers. Signed-off-by: Rodrigo Siqueira <[email protected]> Acked-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: Fix null pointer dereference for dcn401Aurabindo Pillai1-3/+5
When ODM slice happens on DCN401, there is a null pointer exception caused by that. This commit address this issue by checking if the required data structures are initialized. Reviewed-by: Rodrigo Siqueira <[email protected]> Signed-off-by: Aurabindo Pillai <[email protected]> Acked-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: Remove incorrect FP context startAurabindo Pillai1-2/+0
All the DC_FP_START/END should be used before call anything from DML2, for this reason, the use of those guards inside DML it is not correct. This commit removes two unnecessary DC_FP_START/END from a dml2 function. Reviewed-by: Rodrigo Siqueira <[email protected]> Signed-off-by: Aurabindo Pillai <[email protected]> Acked-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: Use DCN 410 includes for DCN401Aurabindo Pillai1-2/+2
DCN401 is using DCN 320 headers, which does not have all the right registers for DCN401. This commit just replace DCN320 includes with the one from DCN410. Reviewed-by: Rodrigo Siqueira <[email protected]> Signed-off-by: Aurabindo Pillai <[email protected]> Acked-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: Remove unnecessary HPD entry for DCN401Aurabindo Pillai1-2/+0
Drop the extra HPD irq entry for DCN401. Reviewed-by: Rodrigo Siqueira <[email protected]> Signed-off-by: Aurabindo Pillai <[email protected]> Acked-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: Update DML2.1 generated codeRodrigo Siqueira25-1172/+2062
Most of the DML code is generated, and it is necessary to update some parts of it from time to time. This commit brings the latest generated code for DML 2.1. Signed-off-by: Rodrigo Siqueira <[email protected]> Acked-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: Refactor HUBP into component folder.Pinninti4-3/+7
[why] cleaning up the code refactor requires hubp to be in its own component. [how] move all files under newly created hubp folder and fixing the makefiles. Reviewed-by: Rodrigo Siqueira <[email protected]> Signed-off-by: Pinninti <[email protected]> Acked-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: Remove unused codeRodrigo Siqueira4-12/+2
This commit removes some unused code with the required adjustments. Signed-off-by: Rodrigo Siqueira <[email protected]> Acked-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: Enable ISHARP support for DCN401Samson Tam8-65/+223
[Why] Enable sharpener support for DCN401 [How] - Removed memcmp check that was preventing ISHARP from being enabled. - Add missing ISHARP register defines, masks, and writes. - Add programming of Blur and Scale coefficients. - Program FMT_MODE and NLDELTA registers based on LLS_PREF and pixel - format - Only enable ISHARP for YUV420 - Add disabling of ISHARP - Add debug flags and registry keys for debugging ISHARP. - Set default to medium level Acked-by: Rodrigo Siqueira <[email protected]> Signed-off-by: Samson Tam <[email protected]> Acked-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: Fix incorrect cursor position for dcn401Sridevi2-14/+34
[Why] Incorrect cursor position calculation in some scenarios. Also for mirror and rotation cases. [How] Fix for incorrect cursor position. Added new test scenarios for diags cursor test. Updated CRC for few of the diags cursor test scenarios. Reviewed-by: Rodrigo Siqueira <[email protected]> Signed-off-by: Sridevi <[email protected]> Acked-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdgpu: fix typo in amdgpu_ras_aca_sysfs_read() functionYang Wang1-1/+1
fix typo "info.ue_count" in amdgpu_ras_aca_sysfs_read() function. Fixes: 865d3397630b ("drm/amdgpu: add aca deferred error type support") Signed-off-by: Yang Wang <[email protected]> Reviewed-by: Tao Zhou <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdgpu: fix dereference null return value for the function ↵Jesse Zhang1-1/+5
amdgpu_vm_pt_parent The pointer parent may be NULLed by the function amdgpu_vm_pt_parent. To make the code more robust, check the pointer parent. Signed-off-by: Jesse Zhang <[email protected]> Suggested-by: Christian König <[email protected]> Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/amdgpu: add AMD_PG_SUPPORT_VCN_DPG flagDavid (Ming Qiang) Wu1-1/+2
AMD_PG_SUPPORT_VCN_DPG is needed for secure parts and should/can be enabled by now. Signed-off-by: David (Ming Qiang) Wu <[email protected]> Reviewed-by: Sonny Jiang <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdgpu: drop MES 10.1 support v3Alex Deucher8-2021/+72
It was an enablement vehicle for MES 11 and was never productized. Remove it. v2: drop additional checks in the GFX10 code. v3: drop mes_api_def.h Acked-by: Christian König <[email protected]> Reviewed-by: Harish Kasiviswanathan <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amd/display: clean up some inconsistent indentingJiapeng Chong1-3/+1
No functional modification involved. drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc.c:5200 dc_power_down_on_boot() warn: inconsistent indenting. Reported-by: Abaci Robot <[email protected]> Closes: https://bugzilla.openanolis.cn/show_bug.cgi?id=9166 Signed-off-by: Jiapeng Chong <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdgpu: silence UBSAN warningAlex Deucher1-1/+1
Convert a variable sized array from [1] to []. Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/amdgpu: Adjust logic in amdgpu_device_partner_bandwidth()Alex Deucher1-7/+12
Use current speed/width on devices which don't support dynamic PCIe switching. Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/3289 Acked-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2024-05-29drm/xe: Drop undesired prefix from the platform nameMichal Wajdeczko1-12/+12
We don't have to use exact names of the enumerators as the potentially user-facing platform names. When constructing platform descriptor fields, use the unique platform tag and add the XE_ prefix only to the generated enum field. Signed-off-by: Michal Wajdeczko <[email protected]> Cc: Lucas De Marchi <[email protected]> Reviewed-by: Rodrigo Vivi <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] Signed-off-by: Lucas De Marchi <[email protected]>
2024-05-29drm/xe/hwmon: Expose card power and energy attributes of BMGKarthik Poosa2-8/+18
In BMG there are separate registers for card/platform power and energy. These are exposed through channel 0 i.e power_1/energy1_xxx. Signed-off-by: Karthik Poosa <[email protected]> Reviewed-by: Badal Nilawar <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Balasubramani Vivekanandan <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] Signed-off-by: Rodrigo Vivi <[email protected]>
2024-05-29drm/xe/hwmon: Add HWMON support for BMGKarthik Poosa2-4/+17
Add HWMON support for BMG. Exposing the pkg power, current, energy info. Signed-off-by: Karthik Poosa <[email protected]> Reviewed-by: Badal Nilawar <[email protected]> Reviewed-by: Rodrigo Vivi <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Balasubramani Vivekanandan <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] Signed-off-by: Rodrigo Vivi <[email protected]>
2024-05-29drm/display: Fix HDMI state helper dependencyMaxime Ripard1-1/+1
During the life of the series that introduced the DRM_DISPLAY_HDMI_STATE_HELPER option, we reworked the Kconfig option dependency setup to rely on depends on with commit f6d2dc03fa85 ("drm: Switch DRM_DISPLAY_HDMI_HELPER to depends on") which got reverted later on because it was creating too many issues by commit d7c128cb775e ("Revert "drm: Switch DRM_DISPLAY_HDMI_HELPER to depends on""). However, since the series was out of tree at that time, DRM_DISPLAY_HDMI_STATE_HELPER wasn't properly updated to take the revert into account and is now creating build issues. Let's switch the depends on to a select to fix this. Reported-by: kernel test robot <[email protected]> Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/ Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/ Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/ Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/ Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/ Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/ Fixes: 54cb39e2293b ("drm/connector: hdmi: Create an HDMI sub-state") Reviewed-by: Javier Martinez Canillas <[email protected]> Reviewed-by: Dmitry Baryshkov <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] Signed-off-by: Maxime Ripard <[email protected]>
2024-05-29drm/xe: Add engine name to the engine reset and cat-err logNirmoy Das1-2/+7
Add engine name to the engine reset and cat error log which should be useful while debugging. v2: Add logical mask and engine class(Matt) Use xe_gt_{info|dbg} (Michal) Cc: Matthew Brost <[email protected]> Cc: Michal Wajdeczko <[email protected]> Reviewed-by: Matthew Brost <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] Signed-off-by: Nirmoy Das <[email protected]>
2024-05-29drm/xe: Check empty pinned BO list with lock held.Nirmoy Das1-6/+4
Use lock that is meant to use for accessing the BO pin list. Reviewed-by: Matthew Brost <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] Signed-off-by: Nirmoy Das <[email protected]>
2024-05-29drm/i915: Fix audio component initializationImre Deak3-11/+24
After registering the audio component in i915_audio_component_init() the audio driver may call i915_audio_component_get_power() via the component ops. This could program AUD_FREQ_CNTRL with an uninitialized value if the latter function is called before display.audio.freq_cntrl gets initialized. The get_power() function also does a modeset which in the above case happens too early before the initialization step and triggers the "Reject display access from task" error message added by the Fixes: commit below. Fix the above issue by registering the audio component only after the initialization step. Fixes: 87c1694533c9 ("drm/i915: save AUD_FREQ_CNTRL state at audio domain suspend") Closes: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/10291 Cc: [email protected] # v5.5+ Signed-off-by: Imre Deak <[email protected]> Reviewed-by: Jani Nikula <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] (cherry picked from commit fdd0b80172758ce284f19fa8a26d90c61e4371d2) Signed-off-by: Jani Nikula <[email protected]>
2024-05-29drm/i915/dpt: Make DPT object unshrinkableVidya Srinivas1-1/+3
In some scenarios, the DPT object gets shrunk but the actual framebuffer did not and thus its still there on the DPT's vm->bound_list. Then it tries to rewrite the PTEs via a stale CPU mapping. This causes panic. Cc: [email protected] Reported-by: Shawn Lee <[email protected]> Fixes: 0dc987b699ce ("drm/i915/display: Add smem fallback allocation for dpt") Signed-off-by: Vidya Srinivas <[email protected]> [vsyrjala: Add TODO comment] Signed-off-by: Ville Syrjälä <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] (cherry picked from commit 51064d471c53dcc8eddd2333c3f1c1d9131ba36c) Signed-off-by: Jani Nikula <[email protected]>
2024-05-29drm/i915/gt: Fix CCS id's calculation for CCS mode settingAndi Shyti3-1/+15
The whole point of the previous fixes has been to change the CCS hardware configuration to generate only one stream available to the compute users. We did this by changing the info.engine_mask that is set during device probe, reset during the detection of the fused engines, and finally reset again when choosing the CCS mode. We can't use the engine_mask variable anymore, as with the current configuration, it imposes only one CCS no matter what the hardware configuration is. Before changing the engine_mask for the third time, save it and use it for calculating the CCS mode. After the previous changes, the user reported a performance drop to around 1/4. We have tested that the compute operations, with the current patch, have improved by the same factor. Fixes: 6db31251bb26 ("drm/i915/gt: Enable only one CCS for compute workload") Signed-off-by: Andi Shyti <[email protected]> Cc: Chris Wilson <[email protected]> Cc: Gnattu OC <[email protected]> Cc: Joonas Lahtinen <[email protected]> Cc: Matt Roper <[email protected]> Tested-by: Jian Ye <[email protected]> Reviewed-by: Umesh Nerlige Ramappa <[email protected]> Tested-by: Gnattu OC <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] (cherry picked from commit a09d2327a9ba8e3f5be238bc1b7ca2809255b464) Signed-off-by: Jani Nikula <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.