- 05 Jul, 2022 4 commits
-
-
Stanley.Yang authored
GFX is the only IP block that RAS TA needs to program the hardware when receiving enable_feature command. Changed from V1: remove amdgpu_ras_need_send_ras_feature inline function, use GFX RAS block check directly. Signed-off-by:Stanley.Yang <Stanley.Yang@amd.com> Reviewed-by:
Hawking Zhang <Hawking.Zhang@amd.com> Signed-off-by:
Alex Deucher <alexander.deucher@amd.com>
-
Evan Quan authored
And bump the version to 0x2A. Signed-off-by:
Evan Quan <evan.quan@amd.com> Acked-by:
-
Lang Yu authored
We don't need to validate and map root PD specially here, it would be validated and mapped by amdgpu_vm_validate_pt_bos if it is evicted. The special case is when turning a GFX VM to a compute VM, if vm_update_mode changed, we should make sure root PD gets mapped. So just map root PD after updating vm->update_funcs in amdgpu_vm_make_compute whether the vm_update_mode changed or not. v3: - Add some comments suggested by Christian. v2: - Don't rename vm_validate_pt_pd_bos and make it public. Signed-off-by:
Lang Yu <Lang.Yu@amd.com> Acked-by:
Christian König <christian.koenig@amd.com> Reviewed-by:
Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by:
-
Aurabindo Pillai authored
[Why&How] Expose a new dc debug mask enum to force a subviewport memory clock switch to facilitate easy testing. Signed-off-by:
Aurabindo Pillai <aurabindo.pillai@amd.com> Acked-by:
Harry Wentland <harry.wentland@amd.com> Signed-off-by:
-
- 30 Jun, 2022 21 commits
-
-
Alex Deucher authored
Properly handle FP code in dcn32_clk_mgr.c. Fixes: 265280b9 ("drm/amd/display: add CLKMGR changes for DCN32/321") Reviewed-by:
Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Signed-off-by:
-
Rodrigo Siqueira authored
When we tried to compile DCN32/321 for 32-bit architecture, we got this error message: ERROR: modpost: "__nedf2" [drivers/gpu/drm/amd/amdgpu/amdgpu.ko] undefined! This commit fixes this issue by rewriting a small part of the dcn32_build_wm_range_table. Cc: Aurabindo Pillai <aurabindo.pillai@amd.com> Cc: Harry Wentland <harry.wentland@amd.com> Cc: Alex Deucher <alexander.deucher@amd.com> Cc: Randy Dunlap <rdunlap@infradead.org> Fixes: 265280b9 ("drm/amd/display: add CLKMGR changes for DCN32/321") Acked-by:
Randy Dunlap <rdunlap@infradead.org> Tested-by: Randy Dunlap <rdunlap@infradead.org> # build-tested Reviewed-by:
-
Rodrigo Siqueira authored
Sometimes when trying to enable some feature, we have to define some values with educated guesses, but we mark those values as TBD, which means "To Be Determined". However, the correct way to approach it is by loading that information from the firmware. Anyway, some of the values that we were experimenting with caused this issue: ERROR: modpost: "__muldf3" [drivers/gpu/drm/amd/amdgpu/amdgpu.ko] undefined! This commit fixes this issue by removing the division by two since it is harmless in this case. Cc: Aurabindo Pillai <aurabindo.pillai@amd.com> Cc: Harry Wentland <harry.wentland@amd.com> Cc: Alex Deucher <alexander.deucher@amd.com> Cc: Randy Dunlap <rdunlap@infradead.org> Fixes: 265280b9 ("drm/amd/display: add CLKMGR changes for DCN32/321") Acked-by:
-
Rodrigo Siqueira authored
When we tried to compile DCN32/321 for 32-bit architecture, we got this error message: ERROR: modpost: "__floatunsidf" [drivers/gpu/drm/amd/amdgpu/amdgpu.ko] undefined! This was caused because we were trying to assign an unsigned int to a double value which causes issues for 32-bit architecture. This issue can be fixed by changing the value type. Cc: Aurabindo Pillai <aurabindo.pillai@amd.com> Cc: Harry Wentland <harry.wentland@amd.com> Cc: Alex Deucher <alexander.deucher@amd.com> Cc: Randy Dunlap <rdunlap@infradead.org> Fixes: 265280b9 ("drm/amd/display: add CLKMGR changes for DCN32/321") Reported-by:
Stephen Rothwell <sfr@canb.auug.org.au> Acked-by:
-
Rodrigo Siqueira authored
While we tried to build amdgpu on i386, we got this error: ERROR: modpost: "__umoddi3" [drivers/gpu/drm/amd/amdgpu/amdgpu.ko] undefined! This commit fixes this issue by replacing the standard module operator with div_u64_rem. Cc: Aurabindo Pillai <aurabindo.pillai@amd.com> Cc: Harry Wentland <harry.wentland@amd.com> Cc: Alex Deucher <alexander.deucher@amd.com> Cc: Randy Dunlap <rdunlap@infradead.org> Fixes: d3dfceb5 ("drm/amd/display: Add dependant changes for DCN32/321") Reported-by:
-
Philip Yang authored
Indicate SMI profiling events available. Signed-off-by:
Philip Yang <Philip.Yang@amd.com> Reviewed-by:
-
Philip Yang authored
The synchronize_rcu may take several ms, which noticeably slows down applications close SMI event handle. Use call_rcu to free client->fifo and client asynchronously and eliminate the synchronize_rcu call in the user thread. Signed-off-by:
-
Philip Yang authored
SVM range unmapped from GPUs when range is unmapped from CPU, or with xnack on from MMU notifier when range is evicted or migrated. Signed-off-by:
-
Philip Yang authored
Output user queue eviction and restore event. User queue eviction may be triggered by svm or userptr MMU notifier, TTM eviction, device suspend and CRIU checkpoint and restore. User queue restore may be rescheduled if eviction happens again while restore. Signed-off-by:
-
Philip Yang authored
For migration start and end event, output timestamp when migration starts, ends, svm range address and size, GPU id of migration source and destination and svm range attributes, Migration trigger could be prefetch, CPU or GPU page fault and TTM eviction. Signed-off-by:
-
Philip Yang authored
Use ktime_get_boottime_ns() as timestamp to correlate with other APIs. Output timestamp when GPU recoverable fault starts and ends to recover the fault, if migration happened or only GPU page table is updated to recover, fault address, if read or write fault. Signed-off-by:
-
Philip Yang authored
Process receive event from same process by default. Add a flag to be able to receive event from all processes, this requires super user permission. Event using pid 0 to send the event to all processes, to keep the default behavior of existing SMI events. Signed-off-by:
-
Philip Yang authored
Define new system management interface event IDs for migration, GPU recoverable page fault, user queues eviction, restore and unmap from GPU events and corresponding event triggers, those will be implemented in the following patches. Signed-off-by:
-
Jack Xiao authored
This reverts commit 8748de87 since drv enabled mes to access registers. Signed-off-by:
Jack Xiao <Jack.Xiao@amd.com> Reviewed-by:
-
Jack Xiao authored
Enable mes to access registers. v2: squash mes sched ring enablement flag Signed-off-by:
-
Jack Xiao authored
Add mes register access routines: 1. read register 2. write register 3. wait register 4. write and wait register Signed-off-by:
-
Jack Xiao authored
Add misc op commands in mes11. Signed-off-by:
-
Jonathan Kim authored
GFX10 and up have work group processors (WGP) and WGP mode is the native compile mode. KFD and ROCr have no visibility into whether a dispatch is operating in CU or WGP mode. Enforce CU masking to be pairwise continguous in enablement and round robin distribute CUs across the SEs in a pairwise manner to assume WGP mode at all times. Signed-off-by:
Jonathan Kim <jonathan.kim@amd.com> Reviewed-by:
-
Jack Xiao authored
Add common interface for mes misc op, including accessing register interface. Signed-off-by:
-
Jack Xiao authored
Update MES firmware api for accessing registers. Signed-off-by:
-
Alex Deucher authored
Fixes this issue: drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:5094: warning: expecting prototype for amdgpu_device_gpu_recover_imp(). Prototype was for amdgpu_device_gpu_recover() instead Fixes: cf727044 ("drm/amdgpu: Rename amdgpu_device_gpu_recover_imp back to amdgpu_device_gpu_recover") Reviewed-by:
Kent Russell <kent.russell@amd.com> Reported-by:
-
- 29 Jun, 2022 10 commits
-
-
Chandan Vurdigere Nataraj authored
[Why] Redundant if-else cases for repeater and non-repeater checks [How] Without changing the core logic, rearranged the code by removing redundant checks Signed-off-by:
Chandan Vurdigere Nataraj <chandan.vurdigerenataraj@amd.com> Reviewed-by:
Leo Li <sunpeng.li@amd.com> Signed-off-by:
-
Aurabindo Pillai authored
[Why&How] Some userspace expect a backwards compatible modifier on DCN32/321. For hardware with num_pipes more than 16, we expose the most efficient modifier first. As a fall back method, we need to expose slightly inefficient modifier AMD_FMT_MOD_TILE_GFX9_64K_R_X after the best option. Also set the number of packers to fixed value as required per hardware documentation. This value is cached during hardware initialization and can be read through the base driver. Signed-off-by:
Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl> Signed-off-by:
-
Aurabindo Pillai authored
[Why&How] TA firmware is needed to enable HDCP. Changes in v2: Load separate firmware for PSP 13.0.0 Signed-off-by:
-
Mauro Carvalho Chehab authored
This symbol is missing documentation: drivers/gpu/drm/amd/include/amd_shared.h:224: warning: Enum value 'PP_GFX_DCS_MASK' not described in enum 'PP_FEATURE_MASK' Document it. Fixes: 680602d6 ("drm/amd/pm: enable DCS") Signed-off-by:
Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by:
-
Mauro Carvalho Chehab authored
There are 4 undocumented fields at struct amdgpu_display_manager. Add documentation for them, fixing those warnings: drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h:544: warning: Function parameter or member 'dmub_outbox_params' not described in 'amdgpu_display_manager' drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h:544: warning: Function parameter or member 'num_of_edps' not described in 'amdgpu_display_manager' drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h:544: warning: Function parameter or member 'disable_hpd_irq' not described in 'amdgpu_display_manager' drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h:544: warning: Function parameter or member 'dmub_aux_transfer_done' not described in 'amdgpu_display_manager' drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h:544: warning: Function parameter or member 'delayed_hpd_wq' not described in 'amdgpu_display_manager' Signed-off-by:
-
Tom Rix authored
sparse reports drivers/gpu/drm/amd/amdgpu/../display/dc/irq/dcn32/irq_service_dcn32.c:39:20: warning: symbol 'to_dal_irq_source_dcn32' was not declared. Should it be static? to_dal_irq_source_dnc32() is only referenced in irq_service_dnc32.c, so change its storage class specifier to static. Fixes: 0efd4374 ("drm/amd/display: add dcn32 IRQ changes") Reviewed-by:
Tom Rix <trix@redhat.com> Signed-off-by:
-
Tom Rix authored
sparse reports drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_link_dp.c:3885:6: warning: symbol 'FORCE_RATE' was not declared. Should it be static? drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_link_dp.c:3886:10: warning: symbol 'FORCE_LANE_COUNT' was not declared. Should it be static? Neither of thse variables is used in dc_link_dp.c. Reviewing the commit listed in the fixes tag shows neither was used in the original patch. So remove them. Fixes: 265280b9 ("drm/amd/display: add CLKMGR changes for DCN32/321") Reviewed-by:
-
Alex Deucher authored
No longer used so drop it. Fixes: ec457f83 ("drm/amd/display: Drop unnecessary detect link code") Reported-by:
kernel test robot <lkp@intel.com> Reviewed-by:
-
Kent Russell authored
Change amdggpu to amdgpu and pedning to pending Signed-off-by:
-
Souptick Joarder (HPE) authored
Kernel test robot throws below warning -> drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_link.c: In function 'dc_link_reduce_mst_payload': drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_link.c:3782:32: warning: variable 'ret' set but not used [-Wunused-but-set-variable] 3782 | enum act_return_status ret; Removed the unused ret variable. Reported-by:Souptick Joarder (HPE) <jrdr.linux@gmail.com> Signed-off-by:
-
- 28 Jun, 2022 5 commits
-
-
Philip Yang authored
This reverts commit ab8529b0. This causes KFDTest KFDMemoryTest.MemoryRegister test failed on gfx9. Signed-off-by:
-
Rahul Kumar authored
We observed hard hang due to NULL derefrence This issue is seen after running system all the time after two or three days struct dc *dc = plane_state->ctx->dc; Randomly in long run we found plane_state or plane_state->ctx is found NULL which causes exception. BUG: kernel NULL pointer dereference, address: 0000000000000000 PF: supervisor read access in kernel mode PF: error_code(0x0000) - not-present page PGD 1dc7f2067 P4D 1dc7f2067 PUD 222c75067 PMD 0 Oops: 0000 [#1] SMP NOPTI CPU: 5 PID: 29855 Comm: kworker/u16:4 ... ... Workqueue: events_unbound commit_work [drm_kms_helper] RIP: 0010:dcn10_update_pending_status+0x1f/0xee [amdgpu] Code: 41 5f c3 0f 1f 44 00 00 b0 01 c3 0f 1f 44 00 00 41 55 41 54 55 53 48 8b 1f 4c 8b af f8 00 00 00 48 8b 83 88 03 00 00 48 85 db <4c> 8b 20 0f 84 bf 00 00 00 48 89 fd 48 8b bf b8 00 00 00 48 8b 07 RSP: 0018:ffff942941997ab8 EFLAGS: 00010286 RAX: 0000000000000000 RBX: ffff8d7fd98d2000 RCX: 0000000000000000 RDX: 0000000000000000 RSI: ffff8d7e3e87c708 RDI: ffff8d7f2d8c0690 RBP: ffff8d7f2d8c0000 R08: ffff942941997a34 R09: 00000000ffffffff R10: 0000000000005000 R11: 00000000000000f0 R12: ffff8d7f2d8c0690 R13: ffff8d8035a41680 R14: 00000000000186a0 R15: ffff8d7f2d8c1dd8 FS: 0000000000000000(0000) GS:ffff8d8037340000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000000 CR3: 0000000148030000 CR4: 00000000003406e0 Call Trace: dc_commit_state+0x6a2/0x7f0 [amdgpu] amdgpu_dm_atomic_commit_tail+0x460/0x19bb [amdgpu] Tested-by:
Rahul Kumar <rahul.kumar1@amd.com> Signed-off-by:
-
Andrey Grodzovsky authored
Align refcount behaviour for amdgpu_job embedded HW fence with classic pointer style HW fences by increasing refcount each time emit is called so amdgpu code doesn't need to make workarounds using amdgpu_job.job_run_counter to keep the HW fence refcount balanced. Also since in the previous patch we resumed setting s_fence->parent to NULL in drm_sched_stop switch to directly checking if job->hw_fence is signaled to short circuit reset if already signed. Signed-off-by:
Andrey Grodzovsky <andrey.grodzovsky@amd.com> Tested-by:
Yiqing Yao <yiqing.yao@amd.com> Acked-by:
-
Andrey Grodzovsky authored
Problem: This patch caused negative refcount as described in [1] because for that case parent fence did not signal by the time of drm_sched_stop and hence kept in pending list the assumption was they will not signal and so fence was put to account for the s_fence->parent refcount but for amdgpu which has embedded HW fence (always same parent fence) drm_sched_fence_release_scheduled was always called and would still drop the count for parent fence once more. For jobs that never signaled this imbalance was masked by refcount bug in amdgpu_fence_driver_clear_job_fences that would not drop refcount on the fences that were removed from fence drive fences array (against prevois insertion into the array in get in amdgpu_fence_emit). Fix: Revert this patch and by setting s_job->s_fence->parent to NULL as before prevent the extra refcount drop in amdgpu when drm_sched_fence_release_scheduled is called on job release. Also - align behaviour in drm_sched_resubmit_jobs_ext with that of drm_sched_main when submitting jobs - take a refcount for the new parent fence pointer and drop refcount for original kref_init for new HW fence creation (or fake new HW fence in amdgpu - see next patch). [1] - https://lore.kernel.org/all/731b7ff1-3cc9-e314-df2a-7c51b76d4db0@amd.com/t/#r00c728fcc069b1276642c325bfa9d82bf8fa21a3Signed-off-by:
-
Andrey Grodzovsky authored
Problem: After we start handling timed out jobs we assume there fences won't be signaled but we cannot be sure and sometimes they fire late. We need to prevent concurrent accesses to fence array from amdgpu_fence_driver_clear_job_fences during GPU reset and amdgpu_fence_process from a late EOP interrupt. Fix: Before accessing fence array in GPU disable EOP interrupt and flush all pending interrupt handlers for amdgpu device's interrupt line. v2: Switch from irq_get/put to full enable/disable_irq for amdgpu Signed-off-by:
-
