kernel/linux.git/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c, branch v4.12.9

drm/amdgpu: fix NULL pointer panic of emit_gds_switch

2017-05-24T20:49:09+00:00

[ 338.384770] BUG: unable to handle kernel NULL pointer dereference at (null) [ 338.384817] IP: [< (null)>] (null) [ 338.385505] RIP: 0010:[<0000000000000000>] [< (null)>] (null) [ 338.385950] Call Trace: [ 338.385993] [] ? amdgpu_vm_flush+0x283/0x400 [amdgpu] [ 338.386025] [] ? printk+0x4d/0x4f [ 338.386074] [] amdgpu_ib_schedule+0x4a6/0x4d0 [amdgpu] [ 338.386140] [] amdgpu_job_run+0x64/0x180 [amdgpu] [ 338.386203] [] amd_sched_main+0x2e9/0x4a0 [amdgpu] [ 338.386232] [] ? prepare_to_wait_event+0x110/0x110 [ 338.386295] [] ? amd_sched_select_entity+0xe0/0xe0 [amdgpu] [ 338.386327] [] kthread+0xd3/0xf0 [ 338.386349] [] ? kthread_park+0x60/0x60 [ 338.386376] [] ret_from_fork+0x25/0x30 [ 338.386401] Code: Bad RIP value. [ 338.386420] RIP [< (null)>] (null) [ 338.386443] RSP [ 338.386458] CR2: 0000000000000000 [ 338.398508] ---[ end trace 4c66fcdc74b9a0a2 ]--- Signed-off-by: Chunming Zhou Reviewed-by: Christian König Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher

drm/amdgpu: fix fundamental suspend/resume issue

2017-05-24T19:49:24+00:00

Reinitializing the VM manager during suspend/resume is a very very bad idea since all the VMs are still active and kicking. This can lead to random VM faults after resume when new processes become the same client ID assigned. Signed-off-by: Christian König Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org

drm/amdgpu: fix dependency issue

2017-05-10T17:23:53+00:00

The problem is that executing the jobs in the right order doesn't give you the right result because consecutive jobs executed on the same engine are pipelined. In other words job B does it buffer read before job A has written it's result. Signed-off-by: Chunming Zhou Reviewed-by: Christian König Signed-off-by: Alex Deucher

drm/amdgpu: PRT support for gfx9 (v3)

2017-04-28T21:32:47+00:00

Fix PRT handling on gfx9 v2: unify PRT bit for all ASICs v3: move PRT flag checking in amdgpu_vm_bo_split_mapping() Signed-off-by: Junwei Zhang Acked-by: David Zhou Reviewed-by: Christian König Signed-off-by: Alex Deucher

drm/amdgpu: fix amdgpu_vm_clear_freed v2

2017-04-28T21:32:47+00:00

Use amdgpu_vm_bo_update_mapping() instead of amdgpu_vm_bo_split_mapping() here. We don't want any flags set in the cleared areas and splitting shouldn't be necessary. v2: fix typo in commit message Signed-off-by: Christian König Reviewed-by: Alex Deucher Reviewed-by: Junwei Zhang Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org

drm/amdgpu: trace vm hub during flush as well v2

2017-04-28T21:32:21+00:00

Trace on which hub we are doing the flush. v2: fix typo in commit message Signed-off-by: Christian König Reviewed-by: Alex Deucher Reviewed-by: Andres Rodriguez Signed-off-by: Alex Deucher

drm/amdgpu: trace the vmhub in grab_id as well

2017-04-28T21:32:20+00:00

Trace on which VMHUB we assigned an VMID. Signed-off-by: Christian König Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher

drm/amdgpu: allow concurrent VM flushes

2017-04-28T21:32:19+00:00

Enable concurrent VM flushes for Vega10. Signed-off-by: Christian König Acked-by: Alex Deucher Reviewed-by: Andres Rodriguez Signed-off-by: Alex Deucher

drm/amdgpu: invalidate only the currently needed VMHUB v2

2017-04-28T21:32:18+00:00

Drop invalidating both hubs from each engine. v2: don't use hardcoded values Signed-off-by: Christian König Reviewed-by: Alex Deucher Reviewed-by: Andres Rodriguez Signed-off-by: Alex Deucher

drm/amdgpu: split VMID management by VMHUB

2017-04-28T21:32:18+00:00

This way GFX and MM won't fight for VMIDs any more. Initially disabled since we need to stop flushing all HUBS at the same time as well. Signed-off-by: Christian König Reviewed-by: Andres Rodriguez Acked-by: Alex Deucher Signed-off-by: Alex Deucher