- 12 Sep, 2017 27 commits
-
-
Christian König authored
All users of a VM must always wait for updates with always valid BOs to be completed. v2: remove debugging leftovers, rename struct member Signed-off-by:
Christian König <christian.koenig@amd.com> Reviewed-by:
Roger He <Hongbo.He@amd.com> Reviewed-by:
Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by:
Alex Deucher <alexander.deucher@amd.com>
-
Evan Quan authored
- v2: share code with CHIP_VEGA10 case Signed-off-by:
Evan Quan <evan.quan@amd.com> Reviewed-by:
Junwei Zhang <Jerry.Zhang@amd.com> Acked-by:
Huang Rui <ray.huang@amd.com> Signed-off-by:
-
Evan Quan authored
Otherwise, the ring will fail to create on next resume. Signed-off-by:
-
Evan Quan authored
- v2: reuse the ring stop api in ring destory Signed-off-by:
-
Evan Quan authored
- fw_size in psp_v10_0_prep_cmd_buf is wrongly set as 0 - fixed the wrong calculation of psp_write_ptr_reg in psp_v10_0_cmd_submit Signed-off-by:
-
Christian König authored
Just some cleanup. Signed-off-by:
Michel Dänzer <michel.daenzer@amd.com> Reviewed-by:
-
Christian König authored
Just some cleanup. Signed-off-by:
-
Christian König authored
This is quite controversial because it adds another lock which is held during page table updates, but I don't see much other option. v2: allow multiple updates to be in flight at the same time v3: simplify the patch, take the read side only once v4: correctly fix rebase conflict Signed-off-by:
-
Christian König authored
Use the VM instead of the BO list to find the BO for a virtual address. This fixes UVD/VCE in physical mode with VM local BOs. Signed-off-by:
Leo Liu <leo.liu@amd.com> Signed-off-by:
-
Christian König authored
When we need to find the mapping we need sysvm access anyway. Signed-off-by:
-
Christian König authored
With shared reservation objects it is possible that we want to temporary kmap an BO while it is still on the swap list. Signed-off-by:
-
Christian König authored
Instead take the callback lock during the final parts of CS. This should solve the last remaining locking order problems with BO reservations. v2: rebase, make dummy functions static inline v3: add one more missing inline and comments Signed-off-by:
-
Christian König authored
Allow at least some parallel processing. Signed-off-by:
-
Christian König authored
Instead of moving them in the MMU notifier move them during CS. v2: still mark pages as accessed/dirty Signed-off-by:
-
Christian König authored
Instead use a counter to figure out if we need to set new pages or not. Signed-off-by:
-
Christian König authored
This didn't helped as intended, just simplify the code. v2: unlock mmap_sem in the error path as well Signed-off-by:
-
Bas Nieuwenhuizen authored
When amdgpu_vm_frag_ptes calls amdgpu_vm_update_ptes and the pt has a shadow PT we mirror all the write to the shadow PT too, which results in twice the commands. Signed-off-by:
Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl> Reviewed-by:
-
Christian König authored
With shared reservation objects __ttm_bo_reserve() can easily fail even on destroyed BOs. This prevents correct handling when we need to individualize the reservation object. Fix this by individualizing the object before even trying to reserve it. Signed-off-by:
Chunming Zhou <david1.zhou@amd.com> Signed-off-by:
-
Christian König authored
This reverts commit 10e709cb. The patch doesn't work at all: 1. The CS can still be blocked because of amdgpu_ctx_add_fence(). 2. The order of submission isn't correct any more. 3. We could end up using freed up memory because we now drop the ctx reference to early. This needs to be fixed cleanly by doing the context handling after the BO handling, but this is a larger task just avoid the obvious crashes for now. Signed-off-by:
-
Tom St Denis authored
We now pass a pointer to a pointer which seems to be what they meant in the first place. The previous version was modifying a pointer passed by value. Fixes bug that was introduced by commit 332798d40c2e91:drm/amd/powerplay: delete eventmgr layer in poweprlay Signed-off-by:
Tom St Denis <tom.stdenis@amd.com> Reviewed-By:
Rex Zhu <Rex.Zhu@amd.com> Signed-off-by:
-
Christian König authored
Move calling put_page into the unpopulate callback. Otherwise we mess up the pages reference count when it is unbound multiple times. Signed-off-by:
-
Tom St Denis authored
Fixed kbuild errors when IOMMU/SWIOTLB are disabled. Signed-off-by:
-
Monk Liu authored
first is incorrect if hit NULL/signaled fence Signed-off-by:
Monk Liu <monk.liu@amd.com> Reviewed-by:
-
Rex Zhu authored
Reviewed-by:
-
Rex Zhu authored
This feature is for UMD to run benchmark in a power state that is as steady as possible. kmd need to fix the power state as stable as possible. now, kmd support four level: profile_standard,peak,min_sclk,min_mclk move common related code to amd_powerplay.c Reviewed-by:
-
Rex Zhu authored
Reviewed-by:
-
Rex Zhu authored
Reviewed-by:
-
- 01 Sep, 2017 13 commits
-
-
Alex Deucher authored
Need to select dpm0 to avoid clock fluctuations. Reviewed-by:
-
Thierry Reding authored
The header comment in include/trace/define_trace.h specifies that the TRACE_INCLUDE_PATH needs to be relative to the define_trace.h header rather than the trace file including it. Most instances get that wrong and work around it by adding the $(src) directory to the include path. While this works, it is preferable to refer to the correct path to the trace file in the first place and avoid any workaround. Reviewed-by:
Thierry Reding <treding@nvidia.com> Signed-off-by:
-
Thierry Reding authored
The header comment in include/trace/define_trace.h specifies that the TRACE_INCLUDE_PATH needs to be relative to the define_trace.h header rather than the trace file including it. Most instances get that wrong and work around it by adding the $(src) directory to the include path. While this works, it is preferable to refer to the correct path to the trace file in the first place and avoid any workaround. Reviewed-by:
-
Tom St Denis authored
Reviewed-by:
-
Tom St Denis authored
Even though fini returns 0 always it could theoretically fail in the future. Might as well return it instead of 0. Signed-off-by:
-
Tom St Denis authored
Signed-off-by:
-
Tom St Denis authored
Signed-off-by:
-
Tom St Denis authored
Signed-off-by:
-
Tom St Denis authored
Right now there's only one but the rest of the code is being setup to support more so might as well fix this up too. Signed-off-by:
-
Christian König authored
Otherwise we lose the NO_EVICT flag and can try to evict pinned BOs. Signed-off-by:
-
Christian König authored
Only move BOs to the moved/relocated list when they aren't already on a list. This prevents accidential removal from the evicted list. Signed-off-by:
-
Evan Quan authored
Signed-off-by:
Flora Cui <Flora.Cui@amd.com> Signed-off-by:
-
Roger He authored
This can improve performance for some cases. v2 (chk): handle all sizes, simplify the patch quite a bit v3 (chk): adjust dw estimation as well v4 (chk): use single loop, make end mask 64bit Signed-off-by:
-
