Revert "mesa: Backport UBWC fix for PIPE_FORMAT_R8_G8B8_420_UNORM for…#2591
Conversation
…mat" This reverts commit 540437f. Patch now available via mesa 26.1.2, currently available in oe-core. Signed-off-by: Ricardo Salveti <ricardo.salveti@oss.qualcomm.com>
|
We are not yet using the oe-core rev that contains latest mesa, but that is part of #2461 (due the series of changes required, from multiple layers), but separated this change as it is required to unblock autobuilder, which is now failing with meta-qcom. |
Test Results 105 files + 36 634 suites +278 5h 27m 50s ⏱️ + 2h 11m 40s Results for commit 41deeb7. ± Comparison against base commit 28629bd. This pull request removes 2 tests.♻️ This comment has been updated with latest results. |
Koen Kooi (koenkooi)
left a comment
Koen Kooi (koenkooi)
left a comment
There was a problem hiding this comment.
In the future we could start using versioned bbappends and discuss if we want to downgrade the dangle error to warning.
Anyway, approved
|
That would mean that the master branch is broken if built using kas. I'm fine with landing this one as a desperate measure, but what does it take to land #2461 ? |
Dmitry Baryshkov (lumag)
merged commit a5b2d1e
into
qualcomm-linux:master
Going over the multiple failures and bringing the fixes as part of #2461, taking a bit to validate the entire run due the lack of cache and constant build failures caused by long clang/rust builds. The only actual known issue (linux-yocto-dev fails to build) will be fixed once https://lists.openembedded.org/g/openembedded-core/message/239324 is merged upstream. The current workaround breaks yocto-check-layer, I could bring the kas patch but hopefully it will be merged soon. |
3 participants
…mat"
This reverts commit 540437f.
Patch now available via mesa 26.1.2, currently available in oe-core.