FS#27737 - glibc 2.15-1 causes applications to hang

Attached to Project: Arch Linux
Opened by Martin Matlag (newtype) - Tuesday, 27 December 2011, 15:41 GMT
Last edited by Allan McRae (Allan) - Wednesday, 28 December 2011, 20:59 GMT
Task Type Bug Report
Category Packages: Testing
Status Closed
Assigned To Allan McRae (Allan)
Architecture x86_64
Severity Critical
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 5
Private No

Details

After update to glibc 2.15-1 some applications hangs after about 30 sec. Apps that I found problematic:
virt-manager,
firefox,
midori,
chromium

All in newest version avalible in testing repo.
This task depends upon

Closed by  Allan McRae (Allan)
Wednesday, 28 December 2011, 20:59 GMT
Reason for closing:  Fixed
Additional comments about closing:  glibc-2.15-2
Comment by Ionut Biru (wonder) - Tuesday, 27 December 2011, 15:48 GMT
I noticed this as well and at first I was thinking that nouveau started to be unstable. Seems is not the nouveau fault since after downgrading to core version, everything is fine

Comment by Anarconda (Anarconda) - Tuesday, 27 December 2011, 16:23 GMT
With an ATI graphics card I had similar problems. All problems gone downgrading from glibc 2.15-1 to 2.14.1-4.
Comment by Michal Bien (Glorf) - Tuesday, 27 December 2011, 16:30 GMT
I have the same problem, firefox hangs.
Additional, with glibc-2.15-1 gnome always starting in fallback mode. After downgrade these problems disappeared.
Comment by Allan McRae (Allan) - Tuesday, 27 December 2011, 21:56 GMT
Can I get some strace output of a hang?
Comment by Allan McRae (Allan) - Tuesday, 27 December 2011, 22:12 GMT
Also, please provide details of which architecture you are using.
Comment by Frederic Bezies (fredbezies) - Tuesday, 27 December 2011, 22:57 GMT
Had to downgrade too.

Using x86_64 version, with testing fully up to date, but not for gcc and glibc.

Freeze when trying to close firefox for example. Can't do a strace sorry. But I noticed it busted also qemu and a lot of programs.
Comment by Vladimir (_v_l) - Tuesday, 27 December 2011, 23:32 GMT
Hi, I don't think that my problem related with glibc but after upgrade to new glibc (2-15-1) and gcc (unrelated completely for the case) svn crashes:
$ LANG=C svn --version
Ошибка сегментирования [segfault]

Running svn under gdb gives following (attached svn-version.log)...

P.S. I didn't reboot after upgrading. Now I'm recompiling kernel (I need support for elantech touchpad, not in 3.1 version) after installing "new" kernel package I'll try again and report.
   svn-version.log (3.1 KiB)
Comment by Vladimir (_v_l) - Wednesday, 28 December 2011, 00:25 GMT
Problem with svn still here.

P.S. Forgot to add: Archlinux x86_64.
Comment by Allan McRae (Allan) - Wednesday, 28 December 2011, 01:22 GMT
Please file a separate bug report for subversion - it is entirely unrelated to this report.
Comment by Vladimir (_v_l) - Wednesday, 28 December 2011, 05:23 GMT
Done:
https://bugs.archlinux.org/task/27743

P.S. Of course I suspected glibc because problem arose after update.
Comment by Allan McRae (Allan) - Wednesday, 28 December 2011, 08:03 GMT
The hanging issue should be fixed in glibc-2.15-2.
Comment by Anarconda (Anarconda) - Wednesday, 28 December 2011, 13:30 GMT
While problems with Firefox and Chromium are gone with new glibc-2.15-2 still I have one problem (downgrading the problem disappear) this time building AUR packages using svn:

[amarcord@Gaia qlandkartegt-svn]$strace makepkg

... some unnecessary lines (I think) ...

wait4(-1, ==> Determinando ultima revisión de svn...
==> Iniciando build()...
==> Connecting to https://qlandkartegt.svn.sourceforge.net/svnroot/qlandkartegt/QLandkarteGT/trunk ...
/home/amarcord/Programas/qlandkartegt-svn/PKGBUILD: línea 20: 1928 Violación de segmento svn up
==> ERROR: Se produjo un error en build().
Cancelando...
[{WIFEXITED(s) && WEXITSTATUS(s) == 2}], 0, NULL) = 1838
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
--- {si_signo=SIGCHLD, si_code=CLD_EXITED, si_pid=1838, si_status=2, si_utime=0, si_stime=0} (Child exited) ---
wait4(-1, 0x7fff47a45058, WNOHANG, NULL) = -1 ECHILD (No child processes)
rt_sigreturn(0xffffffffffffffff) = 0
rt_sigaction(SIGINT, {0x44ec70, [], SA_RESTORER, 0x7f47a1b189b0}, {0x438cc0, [], SA_RESTORER, 0x7f47a1b189b0}, 8) = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
exit_group(2) = ?
[amarcord@Gaia qlandkartegt-svn]$


Checked with tint2-svn and qlandkartegt-svn (both works with 2.14.1-4).

If more information (or translation) is necessary, please, don't hesitate to ask.

TIA.,
Comment by Ionut Biru (wonder) - Wednesday, 28 December 2011, 13:42 GMT
subversion crash is tracked in  FS#27743 
Comment by Vladimir (_v_l) - Wednesday, 28 December 2011, 13:43 GMT
Hi Anarconda (Anarconda).
Please, see my comments above.
Comment by Anarconda (Anarconda) - Wednesday, 28 December 2011, 13:47 GMT
Shame on me. Sorry.

Loading...