Previously, I had assumed this was not possible
so it was OK to not handle it, but now we got a report
from a user who has a board that is configured this way.
When the VCPU BO is already located in a low 32-bit address
in VRAM (eg. when VRAM is mapped to the low address space),
don't do the workaround.
Fixes: 66a80158aa2a ("amdgpu/vce: use amdgpu_gtt_mgr_alloc_entries")
Signed-off-by: Timur Kristóf <[email protected]>
Reviewed-by: Christian König <[email protected]>
---
drivers/gpu/drm/amd/amdgpu/vce_v1_0.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v1_0.c
b/drivers/gpu/drm/amd/amdgpu/vce_v1_0.c
index 884f24be3685..a49f11be74b2 100644
--- a/drivers/gpu/drm/amd/amdgpu/vce_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v1_0.c
@@ -542,6 +542,9 @@ static int vce_v1_0_ensure_vcpu_bo_32bit_addr(struct
amdgpu_device *adev)
u64 vce_gart_start_offs;
int r;
+ if (adev->gmc.vram_start < adev->gmc.gart_start)
+ return amdgpu_bo_gpu_offset(adev->vce.vcpu_bo) <=
max_vcpu_bo_addr ? 0 : -EINVAL;
+
r = amdgpu_gtt_mgr_alloc_entries(&adev->mman.gtt_mgr,
&adev->vce.gart_node, num_pages,
DRM_MM_INSERT_LOW);
--
2.54.0
