lineage-rockchip/android_external_minigbm
1
0
Fork 0
Code Activity Actions

amdgpu: Only align video BOs to 512 for Android

Browse source 
crrev/c/3633249 compares libva and minigbm allocated Surfaces/BOs
for accelerated video decoding. Currently it fails on e.g. zork
because libva (mesa-amd backend) aligns those buffers to 256,
whereas minigbm aligns them to 512. This CL reduces the 512-align
provision in amdgpu to Android-originated buffers, where IIUC
from the bug there is such restriction.

BUG=b:200680176
BUG=b:171013552
TEST=vaapi_unittest on zork w/ crrev/c/3633249

Change-Id: I703037c70dded8596408c475518a29de3b59f039
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/minigbm/+/3647157
Tested-by: Miguel Casas-Sanchez <mcasas@chromium.org>
Reviewed-by: Drew Davenport <ddavenport@chromium.org>
Commit-Queue: Miguel Casas-Sanchez <mcasas@chromium.org>
This commit is contained in:
Miguel Casas 2022-05-13 17:55:49 -07:00 • committed by Chromeos LUCI
parent 365e5b4a9e
commit 26960909ad
1 changed files with 3 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

6
amdgpu.c
View file

@ -517,10 +517,10 @@ static int amdgpu_create_bo_linear(struct bo *bo, uint32_t width, uint32_t heigh
* aligned. This uses more memory than necessary since the first plane only needs to be
* 256 aligned, but it's acceptable for a short-term fix. It's probably safe for other gpu
* families, but let's restrict it to Raven and Stoney for now (b/171013552, b/190484589).
* This only applies to the Android YUV (multiplane) format.
* */
if (num_planes > 1 &&
(priv->dev_info.family == AMDGPU_FAMILY_RV ||
(priv->dev_info.family == AMDGPU_FAMILY_CZ && !(use_flags & BO_USE_HW_VIDEO_ENCODER))))
if (format == DRM_FORMAT_YVU420_ANDROID &&
(priv->dev_info.family == AMDGPU_FAMILY_RV || priv->dev_info.family == AMDGPU_FAMILY_CZ))
stride = ALIGN(stride, 512);
else
stride = ALIGN(stride, 256);