Tethereal 0.10.0a (from Ethereal 0.10.0a) fails with segmentation fault in my iBook G4. It works perfectly in my x86 box. Actual Results: americana root # tethereal Segmentation fault Expected Results: No errors ;-) Emerge Info Output: americana root # emerge info Portage 2.0.50_pre19 (default-ppc-1.4, gcc-3.2.3, glibc-2.3.2-r9, 2.6.1-benh1) ================================================================= System uname: 2.6.1-benh1 ppc 7455, altivec supported Gentoo Base System version 1.4.3.12 Autoconf: sys-devel/autoconf-2.59 Automake: sys-devel/automake-1.7.8 ACCEPT_KEYWORDS="ppc ~ppc" AUTOCLEAN="yes" CFLAGS="-O2 -pipe -mcpu=7450 -maltivec -mabi=altivec" CHOST="powerpc-unknown-linux-gnu" COMPILER="gcc3" CONFIG_PROTECT="/etc /usr/X11R6/lib/X11/xkb /usr/kde/2/share/config /usr/kde/3.2/share/config /usr/kde/3/share/config /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/lib/pbbuttonsd /var/qmail/control" CONFIG_PROTECT_MASK="/etc/gconf /etc/env.d" CXXFLAGS="-O2 -pipe -mcpu=7450 -maltivec -mabi=altivec" DISTDIR="/usr/portage/distfiles" FEATURES="ccache fixpackages" GENTOO_MIRRORS="http://192.168.232.1 http://ftp.gentoo.skynet.be/pub/gentoo http://gentoo.oregonstate.edu http://www.ibiblio.org/pub/Linux/distributions/gentoo" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/local/portage" SYNC="rsync://192.168.232.1/portage" USE="X aalib alsa altivec arts berkdb cdr crypt cscope cups dga dvd encode fbcon flash foomaticdb gd gdbm gif gphoto2 gpm gtk gtk2 imagemagick imap imlib innodb ipv6 java javascript joystick jpeg kde lame lcms libwww mad maildir mitshm motif mpeg ncurses net nls oggvorbis opengl oss pam perl png ppc python qt qtmt radeon readline samba sdl slang socks5 spell ssl tcpd tetex tiff truetype usb vim-with-x xml xml2 xv xvid zlib linguas_es"
Strace here: http://thor.epsig.uniovi.es/~skyhusker/strace.tethereal.txt
Does 0.9.16 do the same thing?
americana root # tethereal -v tethereal 0.9.16 Compiled with GLib 2.2.3, with libpcap 0.8, with libz 1.2.1, without UCD-SNMP or Net-SNMP, without ADNS Running with libpcap version 0.8 on Linux 2.6.1-benh1 americana root # tethereal Segmentation fault
As this bug wasn't going anywhere I've posted the information here on the ethereal-dev mailing list in case someone there might be able to help.
Looking at Guy Harris's reply to my question on ethereal-dev he wants a strack trace rather than a system trace. Can you do this? http://www.ethereal.com/lists/ethereal-dev/200401/msg00848.html
- run the app from gdb and paste the stack trace (backtrace) gdb [app] press r and then bt - alternatively enable core dumping and make gdb read it ulimit -c unlimited [app] gdb [app] corefile bt
Could you try the ebuild over in bug 40660 and see if that cures the problem? It may do or may not.
I have the same problem of segfault. I've tried to rebuild ethereal to get debug informations using : CFLAGS="-O2 -pipe -mcpu=7400 -maltivec -mabi=altivec -g" FEATURES="noclean nostrip" emerge ethereal but I wasn't able to complete the build, getting an "Internal compiler error"... Removing -mabi=altivec : CFLAGS="-O2 -pipe -mcpu=7400 -maltivec -g" FEATURES="noclean nostrip" emerge ethereal The build succeeded and tethereal does not segfault anymore. Rebuilding without debug : CFLAGS="-O2 -pipe -mcpu=7400 -maltivec" emerge ethereal gives a perfectly working tethereal I've already encountered exactly the same problems (1. -mabi=altivec and -g incompatibility, 2. removing -mabi=altivec solves the problem) as I have reported in bug #36127. It looks like a general problem in gcc with Altivec extensions for ABI... Should I fill a new bug for this ? Concerning ethereal doing "filter-flags -mabi-altivec" should solve the problem.
Works flawlessly with gcc 3.3.3. Seems a gcc problem instead of an ethereal problem.
reopened, gcc-3.3.3 isn't yet stable enough to be put even on ~ppc
gcc-3.3.3_pre20040408 is now stable, I think we can close this bug.
Closing
