On 2025-12-01 09:28, Philip Yang wrote:
To reduce queue switch latency further, move MQD to VRAM domain, add AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS flag to allocate contiguous pages using one buddy block.
Why does it need to be contiguous? In the next patch you're mapping it in the GART anyway.
Regards, Felix
Signed-off-by: Philip Yang <[email protected]> --- drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c | 3 ++- drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c | 2 +- 2 files changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index 9cd1660b8f60..c11e37915365 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -329,7 +329,8 @@ int amdgpu_amdkfd_alloc_gtt_mem(struct amdgpu_device *adev, size_t size, bp.size = size; bp.byte_align = PAGE_SIZE; bp.domain = domain; - bp.flags = AMDGPU_GEM_CREATE_CPU_GTT_USWC; + bp.flags = AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS | + AMDGPU_GEM_CREATE_CPU_GTT_USWC; bp.type = ttm_bo_type_kernel; bp.resv = NULL; bp.bo_ptr_size = sizeof(struct amdgpu_bo); diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c index a489d43d5f64..c6945c842267 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c @@ -139,7 +139,7 @@ static struct kfd_mem_obj *allocate_mqd(struct kfd_node *node, (ALIGN(q->ctl_stack_size, PAGE_SIZE) + ALIGN(sizeof(struct v9_mqd), PAGE_SIZE)) * NUM_XCC(node->xcc_mask), - AMDGPU_GEM_DOMAIN_GTT, + AMDGPU_GEM_DOMAIN_VRAM, &(mqd_mem_obj->gtt_mem), &(mqd_mem_obj->gpu_addr), (void *)&(mqd_mem_obj->cpu_ptr), true);
