Am 27.04.23 um 05:23 schrieb Lin.Cao:
Vmbo->shadow is used to back vram bo up when vram lost. So that we should set
shadow as vmbo->shadow to recover vmbo->bo.
Good catch.
Fix: 'commit e18aaea733da ("drm/amdgpu: move shadow_list to amdgpu_bo_vm")'
Signed-off-by: Lin.Cao <lincao12@xxxxxxx>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 8 +++++++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index e536886f6d42..a3c8216f554a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -4501,7 +4501,13 @@ static int amdgpu_device_recover_vram(struct amdgpu_device *adev)
dev_info(adev->dev, "recover vram bo from shadow start\n");
mutex_lock(&adev->shadow_list_lock);
list_for_each_entry(vmbo, &adev->shadow_list, shadow_list) {
- shadow = &vmbo->bo;
+
+ /* If vm is compute context or adev is APU, shadow will be NULL */
+ if (vmbo->shadow)
+ shadow = vmbo->shadow;
+ else
+ continue;
Better write that as "if (!vmbo->shadow) continue;", apart from that
looks good to me.
Christian.
+
/* No need to recover an evicted BO */
if (shadow->tbo.resource->mem_type != TTM_PL_TT ||
shadow->tbo.resource->start == AMDGPU_BO_INVALID_OFFSET ||
