Summary: | x11-libs/pixman-0.15.20 segfaults when fixefox is used | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Andrew Savchenko <bircoph> |
Component: | Current packages | Assignee: | Gentoo X packagers <x11> |
Status: | RESOLVED DUPLICATE | ||
Severity: | normal | CC: | esigra, jlec, mozilla |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
Andrew Savchenko
2009-08-30 08:23:32 UTC
1) Try with _sane_ CFLAGS 2) Disable ccache 3) Try downgrading pixman back to 0.14.0-r1 (In reply to comment #1) > 1) Try with _sane_ CFLAGS I tried with "-O2 -g", it still fails. And now, please, point me to even single unsafe flag in my options. Actually this is nothing more, but distcc-friendly notation of "-O3 -fno-inline-functions -march=native -fomit-frame-pointer -pipe -mfpmath=sse". > 2) Disable ccache I disabled ccache and distcc, with no positive result. Yes, the latter is used, but not via portage FEATURES due to a numerous reasons. But requirement of ccache disabling is nonsense. It is used worldwide for years and no single problem is known due to ccache failure of any recent ccache versions. > 3) Try downgrading pixman back to 0.14.0-r1 I tried 0.16.0 instead. Things are unchanged. However, I looked into the code. The problem occurs at memory access in user-supplied buffer, so this is not libpixman fault, but improper usage by firefox or xullrunner or some another library. Thus I close this bug because it seems to be related to another software. I really have no intention to debug xulrunner to verify precisely what is wrong. (In reply to comment #2) > But requirement of ccache disabling is nonsense. It is used worldwide for years > and no single problem is known due to ccache failure of any recent ccache > versions. FYI, just take a look through our bugzilla and you'll see how many bugs are caused by ccache. It's not because ccache is used worldwide that it suddenly bug-free. I asked you to disable ccache, not because I don't like it but, because it _is_ a source of problems. Anyhow, thanks for the follow up, we appreciate it. Thanks (In reply to comment #0) [snip] > $ emerge --info > Portage 2.2_rc39 (default/linux/x86/2008.0, gcc-4.4.1, glibc-2.10.1-r0, > 2.6.29.2-yoruichi i686) [snip] > CFLAGS="-march=athlon-xp -m32 --param l1-cache-line-size=64 --param > l1-cache-size=64 --param l2-cache-size=512 -O2 -funswitch-loops > -fpredictive-commoning -fgcse-after-reload -ftree-vectorize > -fomit-frame-pointer -mfpmath=sse -pipe" gcc-4.4 + x86 + -ftree-vectorize == gcc fail Dupe of bug 270120 per previous comment. *** This bug has been marked as a duplicate of bug 270120 *** (In reply to comment #4) > gcc-4.4 + x86 + -ftree-vectorize == gcc fail Yes, that's it! Now I get rid of firefox, ld segfaults and some other weird problems. I suspected binutils, my hardware, but not unmasked gcc 8-). Thank you for pointing out. |