diff options
| author |   Dave Airlie <airlied@redhat.com> | 2019-12-04 10:13:16 +1000 |
|---|---|---|
| committer | Dave Airlie <airlied@redhat.com> | 2019-12-04 10:13:16 +1000 |
| commit | 909a60652624b837aff16825197c6a21d84fec82 (patch) | |
| tree | e565f940a2b1a135915c001490956376dc711085 /drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | |
| parent | Merge tag 'drm-intel-next-fixes-2019-11-28' of git://anongit.freedesktop.org/drm/drm-intel into drm-next (diff) | |
| parent | drm/radeon: fix r1xx/r2xx register checker for POT textures (diff) | |
| download | linux-dev-909a60652624b837aff16825197c6a21d84fec82.tar.xz linux-dev-909a60652624b837aff16825197c6a21d84fec82.zip | |
Merge tag 'drm-next-5.5-2019-12-03' of git://people.freedesktop.org/~agd5f/linux into drm-next
drm-next-5.5-2019-12-03:
amdgpu:
- Fix vram lost handling with BACO on VI/CI asics
- DC fixes for Navi14
- Misc gfx10 fixes
- SR-IOV fixes
- Fix driver unload
- Fix XGMI limits on Arcturus
amdkfd:
- Enable KFD on PPC
- Optimize KFD page table reservations
radeon:
- Fix register checker for r1xx/r2xx
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191203204135.5437-1-alexander.deucher@amd.com
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c')
| -rw-r--r-- | drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 15 |
1 files changed, 14 insertions, 1 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index ae6f5446262c..12dbcfaa34b8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -105,11 +105,24 @@ void amdgpu_amdkfd_gpuvm_init_mem_limits(void) (kfd_mem_limit.max_ttm_mem_limit >> 20)); } +/* Estimate page table size needed to represent a given memory size + * + * With 4KB pages, we need one 8 byte PTE for each 4KB of memory + * (factor 512, >> 9). With 2MB pages, we need one 8 byte PTE for 2MB + * of memory (factor 256K, >> 18). ROCm user mode tries to optimize + * for 2MB pages for TLB efficiency. However, small allocations and + * fragmented system memory still need some 4KB pages. We choose a + * compromise that should work in most cases without reserving too + * much memory for page tables unnecessarily (factor 16K, >> 14). + */ +#define ESTIMATE_PT_SIZE(mem_size) ((mem_size) >> 14) + static int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev, uint64_t size, u32 domain, bool sg) { + uint64_t reserved_for_pt = + ESTIMATE_PT_SIZE(amdgpu_amdkfd_total_mem_size); size_t acc_size, system_mem_needed, ttm_mem_needed, vram_needed; - uint64_t reserved_for_pt = amdgpu_amdkfd_total_mem_size >> 9; int ret = 0; acc_size = ttm_bo_dma_acc_size(&adev->mman.bdev, size, |