Opened 9 months ago
Closed 9 months ago
#5315 closed defect (fixed)
Glibc-2.38 posix_memalign is sometimes very slow
| Reported by: | Xi Ruoyao | Owned by: | lfs-book |
|---|---|---|---|
| Priority: | normal | Milestone: | 12.0 |
| Component: | Book | Version: | git |
| Severity: | normal | Keywords: | |
| Cc: |
Description
https://sourceware.org/bugzilla/show_bug.cgi?id=30723
I'm not sure how severe this issue is. It's causing a Mesa test failure but Mesa works fine in practice. Arch users have reported that this can blow up ffmpeg-based applications (https://github.com/mpv-player/mpv/issues/12076) if caching is enabled, but we've not received complains about ffmpeg yet.
I expect the fix to be committed upstream on next Monday (Aug 14, 2023). Maybe it's worthy to apply the fix before package freeze (to prevent subtle issues from suddenly showing up during tagging, or user complains after 12.0 release).
Change History (3)
comment:1 by , 9 months ago
comment:2 by , 9 months ago
Yes, that looks like we should apply that patch. I was going to prepare for package freeze on the 14th, but this looks important enough for us to hold off a day or two.
comment:3 by , 9 months ago
| Resolution: | → fixed |
|---|---|
| Status: | new → closed |
Fixed at commit f7d3b6ed213e8957b3c32318fc240c9f73ffe7ea
The first part of the fix is committed as https://sourceware.org/git/?p=glibc.git;a=commit;h=542b1105852568c3ebc712225ae78b8c8ba31a78. The second part is under review at https://sourceware.org/pipermail/libc-alpha/2023-August/150857.html.