Re: [PATCH v3] drm/amdgpu: refine amdgpu_bo_create_kernel_at()

["Yu, Lang" <Lang.Yu@xxxxxxx>]

[Public]

Ping

---

发件人: Yu, Lang <Lang.Yu@xxxxxxx>
发送时间: 星期二, 八月 1, 2023 15:16
收件人: amd-gfx@xxxxxxxxxxxxxxxxxxxxx <amd-gfx@xxxxxxxxxxxxxxxxxxxxx>
抄送: Koenig, Christian <Christian.Koenig@xxxxxxx>; Paneer Selvam, Arunpravin <Arunpravin.PaneerSelvam@xxxxxxx>; Zhang, Yifan <Yifan1.Zhang@xxxxxxx>; Yu, Lang <Lang.Yu@xxxxxxx>
主题: [PATCH v3] drm/amdgpu: refine amdgpu_bo_create_kernel_at()

 

Use amdgpu_bo_create_reserved() to create a BO in VRAM
domain would fail if requested VRAM size is large(>128MB)
on APU which usually has a default 512MB VRAM.

That's because VRAM is framgented after several allocations.

The approach is using amdgpu_bo_create_reserved() to
create a BO in CPU domain first, it will always succeed.

v2: Don't overwrite the contents at specific offset.
v3: Don't return GPU addr.

Signed-off-by: Lang Yu <Lang.Yu@xxxxxxx>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 19 ++++++++++++++-----
 1 file changed, 14 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index ff73cc11d47e..df5ba9509a41 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -377,27 +377,36 @@ int amdgpu_bo_create_kernel_at(struct amdgpu_device *adev,
         size = ALIGN(size, PAGE_SIZE);
 
         r = amdgpu_bo_create_reserved(adev, size, PAGE_SIZE,
-                                     AMDGPU_GEM_DOMAIN_VRAM, bo_ptr, NULL,
-                                     cpu_addr);
+                                     AMDGPU_GEM_DOMAIN_CPU,
+                                     bo_ptr, NULL, NULL);
         if (r)
                 return r;
 
         if ((*bo_ptr) == NULL)
                 return 0;
 
+       (*bo_ptr)->preferred_domains = AMDGPU_GEM_DOMAIN_VRAM;
+       (*bo_ptr)->allowed_domains = (*bo_ptr)->preferred_domains;
+       (*bo_ptr)->flags = AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
+       (*bo_ptr)->flags |= cpu_addr ? AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED
+               : AMDGPU_GEM_CREATE_NO_CPU_ACCESS;
+
         /*
          * Remove the original mem node and create a new one at the request
          * position.
          */
-       if (cpu_addr)
-               amdgpu_bo_kunmap(*bo_ptr);
-
         ttm_resource_free(&(*bo_ptr)->tbo, &(*bo_ptr)->tbo.resource);
 
         for (i = 0; i < (*bo_ptr)->placement.num_placement; ++i) {
                 (*bo_ptr)->placements[i].fpfn = offset >> PAGE_SHIFT;
                 (*bo_ptr)->placements[i].lpfn = (offset + size) >> PAGE_SHIFT;
+               (*bo_ptr)->placements[i].mem_type = TTM_PL_VRAM;
+               (*bo_ptr)->placements[i].flags = TTM_PL_FLAG_CONTIGUOUS;
+
+               if (!((*bo_ptr)->flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED))
+                       (*bo_ptr)->placements[i].flags |= TTM_PL_FLAG_TOPDOWN;
         }
+
         r = ttm_bo_mem_space(&(*bo_ptr)->tbo, &(*bo_ptr)->placement,
                              &(*bo_ptr)->tbo.resource, &ctx);
         if (r)
--
2.25.1