Bug 74720 – hald segfaults after 10 seconds with -O2 AND -fomit-frame-pointer

• Product: Gentoo Linux
• Status: VERIFIED NEEDINFO

[Description Ernst Sjöstrand 2004-12-16 20:16:36 UTC]

-O3 alone works.
-fomit-frame-pointer alone works.
But when using both it segfault after trying to send some messages.
(I assume that the messages are normal?)
Filter -fomit-frame-pointer?

sys-apps/hal-0.4.2

Can't think of anything else to add right now.

libhal.c 911 : Error sending msg: No property info.category on device with id /o rg/freedesktop/Hal/devices/ide_1_0
libhal.c 911 : Error sending msg: No property info.category on device with id /o rg/freedesktop/Hal/devices/ide_1_0
libhal.c 911 : Error sending msg: No property info.category on device with id /o rg/freedesktop/Hal/devices/ide_1_0
libhal.c 911 : Error sending msg: No property info.category on device with id /o rg/freedesktop/Hal/devices/ide_1_0
libhal.c 911 : Error sending msg: No property info.category on device with id /org/freedesktop/Hal/devices/ide_1_0
libhal.c 911 : Error sending msg: No property info.category on device with id /org/freedesktop/Hal/devices/ide_0_0
libhal.c 911 : Error sending msg: No property info.category on device with id /org/freedesktop/Hal/devices/scsi_4_0_0_0
libhal.c 911 : Error sending msg: No property info.category on device with id /org/freedesktop/Hal/devices/scsi_4_0_0_0
libhal.c 911 : Error sending msg: No property info.category on device with id /org/freedesktop/Hal/devices/scsi_4_0_0_0
libhal.c 911 : Error sending msg: No property info.category on device with id /org/freedesktop/Hal/devices/scsi_4_0_0_0
libhal.c 911 : Error sending msg: No property info.category on device with id /org/freedesktop/Hal/devices/scsi_4_0_0_0
libhal.c 911 : Error sending msg: No property info.category on device with id /org/freedesktop/Hal/devices/scsi_4_0_0_0
libhal.c 911 : Error sending msg: No property info.category on device with id /org/freedesktop/Hal/devices/scsi_4_0_0_0
libhal.c 911 : Error sending msg: No property info.category on device with id /org/freedesktop/Hal/devices/scsi_4_0_0_0
libhal.c 911 : Error sending msg: No property info.category on device with id /org/freedesktop/Hal/devices/scsi_4_0_0_0
Segmenteringsfel

Portage 2.0.51-r8 (gcc34-x86-2004.2, gcc-3.4.3, glibc-2.3.4.20041102-r0, 2.6.9-gentoo-r10 i686)
=================================================================
System uname: 2.6.9-gentoo-r10 i686 AMD Athlon(tm) 64 Processor 3200+
Gentoo Base System version 1.6.8
Python:              dev-lang/python-2.3.4 [2.3.4 (#1, Sep  2 2004, 18:07:59)]
dev-lang/python:     2.3.4
sys-devel/autoconf:  2.59-r6, 2.13
sys-devel/automake:  1.5, 1.9.3, 1.8.5-r2, 1.6.3, 1.7.9, 1.4_p6
sys-devel/binutils:  2.15.92.0.2-r2
sys-devel/libtool:   1.5.10-r1
virtual/os-headers:  2.6.8.1-r1
ACCEPT_KEYWORDS="x86 ~x86"
AUTOCLEAN="yes"
CFLAGS="-O3 -pipe -march=k8 -fomit-frame-pointer -ftracer"
CHOST="i686-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/X11R6/lib/X11/xkb /usr/kde/2/share/config /usr/kde/3.3/env /usr/kde/3.3/share/config /usr/kde/3.3/shutdown /usr/kde/3/share/config /usr/lib/mozilla/defaults/pref /usr/share/config /usr/share/texmf/dvipdfm/config/ /usr/share/texmf/dvips/config/ /usr/share/texmf/tex/generic/config/ /usr/share/texmf/tex/platex/config/ /usr/share/texmf/xdvi/ /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d"
CXXFLAGS="-O3 -pipe -march=k8 -fomit-frame-pointer -ftracer"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoaddcvs ccache distlocks sandbox"
GENTOO_MIRRORS="http://ftp.du.se/pub/os/gentoo http://mirror.pudas.net/gentoo http://trumpetti.atm.tut.fi/gentoo/"
MAKEOPTS="-j2"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
PORTDIR_OVERLAY="/usr/local/private_overlay /usr/local/kde_overlay"
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
USE="3dnow 3dnowex X aac acpi alsa apache2 apm arts avi berkdb bitmap-fonts bootsplash bzlib caps cddb cdr crypt cups curl dnd doc dvd eds encode esd evo f77 faac fbcon ffmpeg foomaticdb fortran gcj gdbm gif gimp gimpprint glade glitz glut gmail gmp gnome gnomedb gnutls gphoto2 gpm gstreamer gtk gtk2 gtkhtml hal howl icq imap imlib imlib2 jabber java jit jpeg kde libg++ libwww mad mikmod mime mmap mmx mmx2 mng mono mozilla mpeg mpeg4 msn mysql ncurses nls nptl nptlonly nvidia objc offensive oggvorbis openal opengl oscar pam pdflib perl plotutils png postgres pthreads python qt quicktime readline samba scanner sdk sdl slang spell ssl stencil-buffer svg tcltk tcpd tetex theora threads tiff truetype type1 uml unicode usb x86 xml2 xv zlib"


Reproducible: Always
Steps to Reproduce:

[Comment 1 foser (RETIRED) 2004-12-17 06:30:11 UTC]

get a backtrace with debug info 

i'd say the first thing to remove from cflags would be tracer, have you experimented with that ?

[Comment 2 Ernst Sjöstrand 2004-12-17 08:54:52 UTC]

I've tested them all pretty thoroughly, it's not tracer.
Details:

WORKS:
-03
-fomit-frame-pointer
-O3 -pipe -march=k8 -ftracer
-O1 -fomit-frame-pointer
-O1 -fomit-frame-pointer -ftracer
-O1 -pipe -march=k8 -fomit-frame-pointer -ftracer

CRASH:
-O3 -pipe -march=k8 -fomit-frame-pointer -ftracer
-O3 -pipe -march=k8 -fomit-frame-pointer
-O2 -pipe -march=k8 -fomit-frame-pointer -ftracer
-O3 -fomit-frame-pointer
-O2 -fomit-frame-pointer

[Comment 3 Ernst Sjöstrand 2004-12-17 09:33:43 UTC]

Some of those CFLAGS made hald not crash anymore.

BUT, the crash was in glib. (2.4.8 that is ;-)

0xb7fc367a in g_signal_has_handler_pending () from /usr/lib/libgobject-2.0.so.0
(gdb) backtrace
#0  0xb7fc367a in g_signal_has_handler_pending () from /usr/lib/libgobject-2.0.so.0

Compiling glib with these cflags makes the crash go away also!
-O2 -pipe -march=k8 -fomit-frame-pointer -ftracer
-O3 -pipe -fomit-frame-pointer -ftracer

This turned into a mess. But those CFLAGS have been surprisingly stable on my box for a long time now.

[Comment 4 foser (RETIRED) 2004-12-17 09:40:43 UTC]

recompile at least hal & glib with debug info : USE=debug, add "-g" to CFLAGS & set FEATURES=nostrip , then get a backtrace.

It might be a compiler bug, could you also try with arch i686 and see if the bug shows there too ?

[Comment 5 foser (RETIRED) 2005-01-06 04:06:13 UTC]

needinfo

[Comment 6 Ernst Sjöstrand 2005-01-13 15:56:12 UTC]

Hmm, I didn't have time to look at this for a few weeks, and now with the latest update to hal, I can no longer reproduce it.

Closing. (Specially since noone else seems to be affected.)