Re: net/twinkle stuck in _umtx_op syscall

From: Sam Leffler <sam_at_errno.com>
Date: Thu, 26 Jul 2007 12:11:26 -0700
Ulrich Spoerlein wrote:
> Gentlemen,
> 
> I had trouble running twinkle back on 6-STABLE but now on -CURRENT it's
> similar. A ktrace of the session reveals the following
> 
>   4019 twinkle  0.000273 CALL  gettimeofday(0xbf2f7cec,0)
>   4019 twinkle  0.000009 RET   gettimeofday 0
>   4019 twinkle  0.000518 CALL  sigprocmask(SIG_BLOCK,0,0x29e09508)
>   4019 twinkle  0.000009 RET   sigprocmask 0
>   4019 twinkle  0.000067 CALL  ioctl(0x8,FIONREAD,0xbf2f7bb8)
>   4019 twinkle  0.000017 RET   ioctl 0
>   4019 twinkle  0.000015 CALL  ioctl(0x8,FIONREAD,0xbf2f7ba8)
>   4019 twinkle  0.000015 RET   ioctl 0
>   4019 twinkle  0.000143 CALL  ioctl(0x8,FIONREAD,0xbf2f7b38)
>   4019 twinkle  0.000016 RET   ioctl 0
>   4019 twinkle  0.000007 CALL  ioctl(0x8,FIONREAD,0xbf2f7b28)
>   4019 twinkle  0.000005 RET   ioctl 0
>   4019 twinkle  0.000006 CALL  ioctl(0x8,FIONREAD,0xbf2f7b28)
>   4019 twinkle  0.000014 RET   ioctl 0
>   4019 twinkle  0.000011 CALL  ioctl(0x8,FIONREAD,0xbf2f7b18)
>   4019 twinkle  0.000006 RET   ioctl 0
>   4019 twinkle  0.000014 CALL  _umtx_op(0x2a2e7e80,0x5,0,0,0)
>   4019 twinkle  0.001670 RET   _umtx_op 0
>   4019 twinkle  0.000013 CALL  getitimer(0,0xbf6fbf4c)
>   4019 twinkle  0.000005 RET   getitimer 0
>   4019 twinkle  0.000023 CALL  _umtx_op(0x29e1985c,0x2,0,0,0)
>   4019 twinkle  4.361916 RET   _umtx_op -1 errno 4 Interrupted system call
>   4019 twinkle  0.000044 PSIG  SIGKILL SIG_DFL
> 
> It can only be killed -9 and will otherwise stick in _umtx_op() forever.
> Any clue from the threading guys on what tricks I should try?
> 
> My libmap.conf is emtpy, twinkle is linked against the following
> binaries:
> 
> /usr/local/bin/twinkle:
>         libsndfile.so.1 => /usr/local/lib/libsndfile.so.1 (0x283a0000)
>         libccext2-1.5.so.0 => /usr/local/lib/libccext2-1.5.so.0 (0x283ff000)
>         libgnutls.so.15 => /usr/local/lib/libgnutls.so.15 (0x28441000)
>         libgcrypt.so.13 => /usr/local/lib/libgcrypt.so.13 (0x284bb000)
>         libz.so.4 => /lib/libz.so.4 (0x2850a000)
>         libccrtp1-1.5.so.0 => /usr/local/lib/libccrtp1-1.5.so.0 (0x2851c000)
>         libccgnu2-1.5.so.0 => /usr/local/lib/libccgnu2-1.5.so.0 (0x28542000)
>         librt.so.1 => /usr/lib/librt.so.1 (0x28594000)
>         libkdecore.so.6 => /usr/local/lib/libkdecore.so.6 (0x28599000)
>         libkdeui.so.6 => /usr/local/lib/libkdeui.so.6 (0x287d3000)
>         libkabc.so.3 => /usr/local/lib/libkabc.so.3 (0x28aaa000)
>         libspeex.so.1 => /usr/local/lib/libspeex.so.1 (0x28b5d000)
>         libilbc.so.0 => /usr/local/lib/libilbc.so.0 (0x28b7d000)
>         libzrtpcpp-0.9.so.0 => /usr/local/lib/libzrtpcpp-0.9.so.0 (0x28b8c000)
>         libboost_regex.so => /usr/local/lib/libboost_regex.so (0x28bab000)
>         libqt-mt.so.3 => /usr/local/lib/libqt-mt.so.3 (0x28c39000)
>         libXext.so.6 => /usr/local/lib/libXext.so.6 (0x2930b000)
>         libX11.so.6 => /usr/local/lib/libX11.so.6 (0x29319000)
>         libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0x29405000)
>         libm.so.5 => /lib/libm.so.5 (0x294ed000)
>         libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x29503000)
>         libthr.so.3 => /lib/libthr.so.3 (0x2950e000)
>         libc.so.7 => /lib/libc.so.7 (0x29521000)
>         libFLAC.so.7 => /usr/local/lib/libFLAC.so.7 (0x29623000)
>         libgpg-error.so.0 => /usr/local/lib/libgpg-error.so.0 (0x29656000)
>         libintl.so.8 => /usr/local/lib/libintl.so.8 (0x2965a000)
>         libiconv.so.3 => /usr/local/lib/libiconv.so.3 (0x29663000)
>         libDCOP.so.6 => /usr/local/lib/libDCOP.so.6 (0x29751000)
>         libutil.so.7 => /lib/libutil.so.7 (0x29783000)
>         libart_lgpl_2.so.5 => /usr/local/lib/libart_lgpl_2.so.5 (0x29790000)
>         libidn.so.16 => /usr/local/lib/libidn.so.16 (0x297a6000)
>         libkdefx.so.6 => /usr/local/lib/libkdefx.so.6 (0x297d7000)
>         libjpeg.so.9 => /usr/local/lib/libjpeg.so.9 (0x297ff000)
>         libvcard.so.0 => /usr/local/lib/libvcard.so.0 (0x2981e000)
>         libkio.so.6 => /usr/local/lib/libkio.so.6 (0x29842000)
>         libkresources.so.3 => /usr/local/lib/libkresources.so.3 (0x29b82000)
>         libmng.so.1 => /usr/local/lib/libmng.so.1 (0x29ba4000)
>         libpng.so.5 => /usr/local/lib/libpng.so.5 (0x29c05000)
>         libXi.so.6 => /usr/local/lib/libXi.so.6 (0x29c2a000)
>         libXrender.so.1 => /usr/local/lib/libXrender.so.1 (0x29c32000)
>         libXrandr.so.2 => /usr/local/lib/libXrandr.so.2 (0x29c3a000)
>         libXcursor.so.1 => /usr/local/lib/libXcursor.so.1 (0x29c41000)
>         libXinerama.so.1 => /usr/local/lib/libXinerama.so.1 (0x29c4a000)
>         libXft.so.2 => /usr/local/lib/libXft.so.2 (0x29c4d000)
>         libfreetype.so.9 => /usr/local/lib/libfreetype.so.9 (0x29c5f000)
>         libfontconfig.so.1 => /usr/local/lib/libfontconfig.so.1 (0x29cc9000)
>         libSM.so.6 => /usr/local/lib/libSM.so.6 (0x29cf3000)
>         libICE.so.6 => /usr/local/lib/libICE.so.6 (0x29cfc000)
>         libXau.so.6 => /usr/local/lib/libXau.so.6 (0x29d13000)
>         libXdmcp.so.6 => /usr/local/lib/libXdmcp.so.6 (0x29d16000)
>         librpcsvc.so.4 => /usr/lib/librpcsvc.so.4 (0x29d1b000)
>         libkdesu.so.6 => /usr/local/lib/libkdesu.so.6 (0x29d24000)
>         libkwalletclient.so.1 => /usr/local/lib/libkwalletclient.so.1 (0x29d3c000)
>         libfam.so.0 => /usr/local/lib/libfam.so.0 (0x29d4d000)
>         liblcms.so.1 => /usr/local/lib/liblcms.so.1 (0x29d55000)
>         libXfixes.so.3 => /usr/local/lib/libXfixes.so.3 (0x29d84000)
>         libexpat.so.6 => /usr/local/lib/libexpat.so.6 (0x29d89000)

I may be seeing similar issues with thunderbird on UP i386 w/ HEAD. 
Periodically thunderbird will take 100% of cpu for something like 30-60 
seconds (maybe longer).  When I ktrace the process consuming the cpu I 
see very similar traces (umtx_op returning errors).

This is HEAD as of Jun 2.  I need to update the machine but haven't seen 
any commits that seemed relevant.  Been too busy to pursue the problem 
but would gladly help someone trying to track this down.

	Sam
Received on Thu Jul 26 2007 - 17:07:16 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:15 UTC