Skip to content

Conversation

nros
Copy link

@nros nros commented Dec 14, 2022

When building Yocto kirkstone with this layer activated, the build fails with an error.
This PR tries to fix #2.

build error of Yocto Kirkwood if this layer is activated

ERROR: onevpl-intel-gpu-1_22.3.2-r0 do_patch: Applying patch '0001-Revert-Enable-CM-copy-by-default-on-TGL-RKL-ADL-Linu.patch' on target directory 'tmp/work/intel_skylake_64-poky-linux/onevpl-intel-gpu/1_22.3.2-r0/git'
CmdError('quilt --quiltrc tmp/work/intel_skylake_64-poky-linux/onevpl-intel-gpu/1_22.3.2-r0/recipe-sysroot-native/etc/quiltrc push', 0, 'stdout: Applying patch 0001-Revert-Enable-CM-copy-by-default-on-TGL-RKL-ADL-Linu.patch
patching file _studio/shared/include/libmfx_core_vaapi.h
Hunk #1 FAILED at 208.
1 out of 1 hunk FAILED -- rejects in file _studio/shared/include/libmfx_core_vaapi.h
patching file _studio/shared/src/libmfx_core_vaapi.cpp
Hunk #1 FAILED at 275.
Hunk #2 FAILED at 801.
Hunk #3 FAILED at 1294.
3 out of 3 hunks FAILED -- rejects in file _studio/shared/src/libmfx_core_vaapi.cpp
Patch 0001-Revert-Enable-CM-copy-by-default-on-TGL-RKL-ADL-Linu.patch does not apply (enforce with -f)

stderr: ')
ERROR: Logfile of failure stored in: tmp/work/intel_skylake_64-poky-linux/onevpl-intel-gpu/1_22.3.2-r0/temp/log.do_patch.36007
ERROR: Task (sources/meta-iotg-yocto-ese-main/recipes-multimedia/onevpl/onevpl-intel-gpu_22.3.2.bb:do_patch) failed with exit code '1'

Cause of the error

Applying patch 0001-Revert-Enable-CM-copy-by-default-on-TGL-RKL-ADL-Linu.patch to package onevpl-intel-gpu fails, as the version in Yocto Kirkstone has already imported this patch upstream in version 22.3. So, it can not be applied again and build will fail.

Looking at the history of the package onevpl-intel-gpu in "meta-intel" layer in yocto repository, it is totally unclear, why this patch has ended up here. It is not needed!

Version 22.1 from Feb 2nd 2022 does not include the bad commit to be reverted.

Version 22.3.2 introduced on March 3rd, 2022 already includes the revert commit. Therefore, no version in Yocto contains the bad commit. Hence, the patch is simply removed.

References

@nros nros force-pushed the nr/fix/onevpl-patch-does-not-apply branch 2 times, most recently from 88eb4fe to 45b83f1 Compare December 20, 2022 07:27
The applied patch has already been imported upstream in version 22.3.
So, it can not be applied again and build will fail.

Looking at the history of the package `onevpl-intel-gpu` in
"meta-intel" layer in yocto repository, it is totally unclear, why
this patch has ended up here. It is not needed!
Version 22.1 from Feb 2nd 2022 does not include the bad commit, that
needs to be reverted.
Version 22.3.2 introduced on March 3rd, 2022 already includes the
revert commit. Therefore, no version in Yocto contains the bad
commit and thus it seems no version check need to be performed
with the `.bbappend` file. Hence, the patch is simply removed.

see
----
Yocto Kirstone history:
https://git.yoctoproject.org/meta-intel/log/recipes-multimedia/onevpl?h=kirkstone

upstream import of removed patch
intel/vpl-gpu-rt@5650a11

Signed-off-by: Nikolaus Rosenmayr <[email protected]>
@nros nros force-pushed the nr/fix/onevpl-patch-does-not-apply branch from 45b83f1 to e49be97 Compare May 8, 2023 15:25
@nros
Copy link
Author

nros commented May 8, 2023

Stll not included in the latest updates. Building still fails on May, 8th 2023 with latest Yocto Kirkstone release 75239ddd8d from Arpil 28th 2023.

If you run into the same issue, just use my fork instead or create your own and include my patch: https://github.com/nros/iotg-yocto-ese-main/tree/kirkstone

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Build fails as patch to onevpl-intel-gpu is already in upstream repo
1 participant