[staging-next-25.11] alsa-lib: fix patch for CVE-2026-25068 for v1.2.14#493885
Conversation
There was a problem hiding this comment.
This report is automatically generated by the PR / Check / cherry-pick CI workflow.
Some of the commits in this PR require the author's and reviewer's attention.
If you need to merge this PR despite the warnings, please dismiss this review shortly before merging.
Important
7b1d330 is not a cherry-pick, because: fix that does not apply to unstable since it is on a later version of package. Please review this commit manually.
Hint: The full diffs are also available in the runner logs with slightly better highlighting.
NixOS#492079 introduced a patch for CVE-2026-25068, which was backported to staging-25.11 in NixOS#492453. However, the patch fails to compile when ported directly to 25.11 since the way of doing error handling changed from an `SNDERR` macro to an `snd_error` function between v1.2.14 (which is on 25.11) and on v1.2.15. In order to fix this, we vendor the patch and change the offending line like so: ```diff - + snd_error(TOPOLOGY, "mixer: unexpected channel count %d", map->num_channels); + + SNDERR("mixer: unexpected channel count %d", map->num_channels); ``` Not-cherry-picked-because: fix that does not apply to unstable since it is on a later version of package
1b64f18 to
7b1d330
Compare
mdaniels5757
left a comment
There was a problem hiding this comment.
The spacing in your patch is different, but I verified that it's the otherwise the same as Ubuntu's (from alsa-lib_1.2.14-1ubuntu1.1.debian.tar.xz, so not suitable for fetchpatch).
|
Built on aarch64-linux at HEAD, I should've mentioned. |
fadd0c9
#492079 introduced a patch for CVE-2026-25068, which was backported to staging-25.11 in #492453. However, the patch fails to compile when ported directly to 25.11 since the way of doing error handling changed from an
SNDERRmacro to ansnd_errorfunction between v1.2.14 (which is on 25.11) and on v1.2.15. In order to fix this, we vendor the patch and change the offending line like so:Alternatively, the package itself could be updated? Not familiar enough with alsa-lib to assess whether that should happen/is liable to cause breakage. This is probably the lowest impact option, but I don't know if it's the best.
Things done
passthru.tests.nixpkgs-reviewon this PR. See nixpkgs-review usage../result/bin/.