Before anyone suggests it RAM or hardware errors, I have run memtest86 up and down, all tests, zero RAM errors reported. Also, the error always happens reproducibly for the "ldconfig" in the same spot in (so far) only these two ebuilds and nothing else: From gawk: * Installing filefuncs module ... install -m 0755 -d /var/tmp/portage/gawk-3.1.3-r1/image//lib/rcscripts install -m 0755 filefuncs.so.0.0.1 /var/tmp/portage/gawk-3.1.3-r1/image//lib/rcscripts ldconfig -l /var/tmp/portage/gawk-3.1.3-r1/image//lib/rcscripts/filefuncs.so.0.0.1 make: *** [install] Segmentation fault running an strace of the specific offending command yields: laptop root # strace ldconfig -l /var/tmp/portage/gawk-3.1.3-r1/image/lib/rcscripts/filefuncs.so.0.0.1 execve("/sbin/ldconfig", ["ldconfig", "-l", "/var/tmp/portage/gawk-3.1.3-r1/image/lib/rcscripts/filefuncs.so.0.0.1"], [/* 43 vars */]) = 0 uname({sys="Linux", node="laptop", ...}) = 0 brk(0) = 0x80d0000 brk(0x80d0848) = 0x80d0848 set_thread_area({entry_number:-1 -> 6, base_addr:0x80d0830, limit:1048575, seg_32bit:1, contents:0, read_exec_only:0, limit_in_pages:1, seg_not_present:0, useable:1}) = 0 brk(0x80f1848) = 0x80f1848 brk(0x80f2000) = 0x80f2000 open("/dev/urandom", O_RDONLY) = 3 read(3, "\306\207?\333", 4) = 4 close(3) = 0 --- SIGSEGV (Segmentation fault) @ 0 (0) --- +++ killed by SIGSEGV +++ The procps failure is very similar: install -D --owner 0 --group 0 --mode a=rx --strip proc/libproc.so.3.2.1 /var/tmp/portage/procps-3.2.1/image//lib//libproc.so.3.2.1 cd /var/tmp/portage/procps-3.2.1/image//lib/ && ln -sf libproc.so.3.2.1 libproc.so ldconfig make: *** [/var/tmp/portage/procps-3.2.1/image//lib//libproc.so.3.2.1] Segmentation fault make: *** Deleting file `/var/tmp/portage/procps-3.2.1/image//lib//libproc.so.3.2.1' The strace for this is essentialy identical to the one above. Other info: glibc-2.3.4.20040619 linux26-headers-2.6.7-r1 All the rest of system compiled fine with the emerge info shown below. I did try compiling gawk and procps with CFLAGS="-pipe" so this isn't a cflags issue. IN particular is doesn't seem to have anything to do with "-march=pentium-m" mentioned in a similar bug. I should also mention that this error does not happen on my AMD system, both gawk and procps build just fine with an identical setup (CFLAGS, kernel, glibc, headers, etc) otherwise. Reproducible: Always Steps to Reproduce: 1. 2. 3. laptop lib # emerge info Portage 2.0.50-r9 (default-x86-1.4, gcc-3.4.1, glibc-2.3.4.20040619-r0, 2.6.7-gentoo-r8) ================================================================= System uname: 2.6.7-gentoo-r8 i686 Intel(R) Pentium(R) 4 CPU 3.00GHz Gentoo Base System version 1.5.1 ccache version 2.3 [enabled] Autoconf: sys-devel/autoconf-2.59-r4 Automake: sys-devel/automake-1.8.5-r1 ACCEPT_KEYWORDS="x86 ~x86" AUTOCLEAN="yes" CFLAGS="-O3 -fweb -ffast-math -march=pentium4 -ftracer -pipe -fomit-frame-pointer" CHOST="i686-pc-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/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 -fweb -ffast-math -march=pentium4 -ftracer -pipe -fomit-frame-pointer -fvisibility-inlines-hidden" DISTDIR="/usr/portage/distfiles" FEATURES="autoaddcvs ccache fixpackages sandbox" GENTOO_MIRRORS="ftp:///ftp-stud.fht-esslingen.de/pub/Mirrors/gentoo/ http://gentoo.binarycompass.org http://gentoo.ccccom.com http://gentoo.llarian.net/ http://mirror.tucdemonic.org/gentoo/" MAKEOPTS="-j4" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/local/portage/ebuilds" SYNC="rsync://rsync.gentoo.org/gentoo-portage" USE="3dnow X Xaw3d aalib acpi alsa apm arts artswrappersuid atlas avi berkdb cddb cdr crypt cups curl dga directfb dnd dv dvd emacs encode esd ext-png ext-zlib f77 faad fbcon fbdev ffmpeg fftw flac foomaticdb freetype fs gc gd gd-external gdbm gif gimpprint gphoto2 gpm gstreamer gtk gtk2 gtkhtml imagemagick imap imlib imlib2 java javascript jdepend jikes jpeg kde lcms libg++ libwww mad mikmod mmx mng motif mozdomi mozilla moznocompose moznoirc moznomail mozsvg mozxmlterm mpeg mplayer ncurses nls nptl offensive oggvorbis opengl oss pam passfile pcmcia pcre pdflib perl pic plotutils png ppds python qt qtmt quicktime readline ruby samba sdl slang spell sse ssl svg svga tcltk tcpd tetex threads tiff transcode truetype type1 usb vim-with-x wifi wmf x86 xml xml2 xmms xv zlib"
Same problem here.
please change to critical, since this is blocking bootstrapping.
Could you try compiling ldconfig with debug symboles and run it in gdb, and do a backtrace?
Yikes... when I add "-g" to CFLAGS and "debug" to USE and rebuild glibc (for ldconfig), the error goes away. Talk about pernicious. genlop reports * sys-libs/glibc-2.3.4.20040619 Install date: Wed Jul 14 16:37:11 2004 USE="nls pic nptl debug -build -erandom -hardened -makecheck -multilib" CFLAGS=" -march=pentium4 -pipe -g -freorder-blocks -O2" I don't mess with the automatic flag stipping, etc. so the only thing that should be different from my previous build is the -g and debug. But, I guess just to verify this I will rebuild glibc yet again without the -g and without debug.
OK, after re-emerging glibc without debug and without "-g" the problem remains fixed. Weird since these are the exact same CFLAGS, etc that they were compiled with before. Anyway I guess I don't have much more to add since I can't reproduce this problem any longer.
...is anyone here using a pentium-m, perhaps?
ok I had left this open since Bjarke also seemed to have the problem, but it's gone away for me, and Bjarke has disappeared. For the record, I was not using pentium-m.
Problem has gone away here too, just forgot to update the bug, sorry for that.
