Hi Emily,
have you also tested this? I don't have the hardware to test it so that
would be rather nice to have.
Thanks,
Christian.
Am 25.09.19 um 11:31 schrieb Deng, Emily:
Reviewed-by: Emily Deng <Emily.Deng@xxxxxxx>
-----Original Message-----
From: Christian König <ckoenig.leichtzumerken@xxxxxxxxx>
Sent: Tuesday, September 24, 2019 7:56 PM
To: Deng, Emily <Emily.Deng@xxxxxxx>; amd-gfx@xxxxxxxxxxxxxxxxxxxxx
Subject: [PATCH] drm7amdgpu: once more fix amdgpu_bo_create_kernel_at
When CPU access is needed we should tell that to
amdgpu_bo_create_reserved() or otherwise the access is denied later on.
Signed-off-by: Christian König <christian.koenig@xxxxxxx>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 9 ++++++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 12d2adcdf14e..f10b6175e20c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -369,7 +369,7 @@ int amdgpu_bo_create_kernel_at(struct
amdgpu_device *adev,
size = ALIGN(size, PAGE_SIZE);
r = amdgpu_bo_create_reserved(adev, size, PAGE_SIZE, domain,
bo_ptr,
- NULL, NULL);
+ NULL, cpu_addr);
if (r)
return r;
@@ -377,12 +377,15 @@ int amdgpu_bo_create_kernel_at(struct
amdgpu_device *adev,
* Remove the original mem node and create a new one at the
request
* position.
*/
+ if (cpu_addr)
+ amdgpu_bo_kunmap(*bo_ptr);
+
+ ttm_bo_mem_put(&(*bo_ptr)->tbo, &(*bo_ptr)->tbo.mem);
+
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;
}
-
- ttm_bo_mem_put(&(*bo_ptr)->tbo, &(*bo_ptr)->tbo.mem);
r = ttm_bo_mem_space(&(*bo_ptr)->tbo, &(*bo_ptr)->placement,
&(*bo_ptr)->tbo.mem, &ctx);
if (r)
--
2.14.1
_______________________________________________
amd-gfx mailing list
amd-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
