On 2017å¹´05æ??31æ?¥ 20:38, Christian König wrote: > From: Christian König <christian.koenig at amd.com> > > The fragment bits work differently for Vega10 compared to previous generations. > > Increase the fragment size to 2MB for now to better handle that. I checked the fragment bits in PTE, don't find difference, Could you point what difference is? And 2MB fragment seems not a good option, the buffer people allocated may be often smaller than 2MB, so 64KB seems be more effective. Regards, David Zhou > > Signed-off-by: Christian König <christian.koenig at amd.com> > --- > drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 5 +++-- > drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 5 +++-- > drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 4 +++- > 3 files changed, 9 insertions(+), 5 deletions(-) > > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c > index 4dd83a3..36de8a5 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c > @@ -551,8 +551,9 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file > dev_info.virtual_address_offset = AMDGPU_VA_RESERVED_SIZE; > dev_info.virtual_address_max = (uint64_t)adev->vm_manager.max_pfn * AMDGPU_GPU_PAGE_SIZE; > dev_info.virtual_address_alignment = max((int)PAGE_SIZE, AMDGPU_GPU_PAGE_SIZE); > - dev_info.pte_fragment_size = (1 << AMDGPU_LOG2_PAGES_PER_FRAG) * > - AMDGPU_GPU_PAGE_SIZE; > + dev_info.pte_fragment_size = > + (1 << AMDGPU_LOG2_PAGES_PER_FRAG(adev)) * > + AMDGPU_GPU_PAGE_SIZE; > dev_info.gart_page_size = AMDGPU_GPU_PAGE_SIZE; > > dev_info.cu_active_number = adev->gfx.cu_info.number; > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c > index b56e24a..d07a28c 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c > @@ -1239,8 +1239,9 @@ static int amdgpu_vm_frag_ptes(struct amdgpu_pte_update_params *params, > */ > > /* SI and newer are optimized for 64KB */ > - uint64_t frag_flags = AMDGPU_PTE_FRAG(AMDGPU_LOG2_PAGES_PER_FRAG); > - uint64_t frag_align = 1 << AMDGPU_LOG2_PAGES_PER_FRAG; > + unsigned pages_per_frag = AMDGPU_LOG2_PAGES_PER_FRAG(params->adev); > + uint64_t frag_flags = AMDGPU_PTE_FRAG(pages_per_frag); > + uint64_t frag_align = 1 << pages_per_frag; > > uint64_t frag_start = ALIGN(start, frag_align); > uint64_t frag_end = end & ~(frag_align - 1); > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h > index a49c608..aabe815 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h > @@ -51,7 +51,9 @@ struct amdgpu_bo_list_entry; > #define AMDGPU_VM_PTB_ALIGN_SIZE 32768 > > /* LOG2 number of continuous pages for the fragment field */ > -#define AMDGPU_LOG2_PAGES_PER_FRAG 4 > +#define AMDGPU_LOG2_PAGES_PER_FRAG(adev) \ > + ((adev)->asic_type < CHIP_VEGA10 ? 4 : \ > + (adev)->vm_manager.block_size) > > #define AMDGPU_PTE_VALID (1ULL << 0) > #define AMDGPU_PTE_SYSTEM (1ULL << 1)
