After 'update world', the 32bit applications stops working: $ quake3 /usr/games/bin/quake3: line 4: /opt/quake3/quake3.x86: Accessing a corrupted shared library /usr/games/bin/quake3: line 4: /opt/quake3/quake3.x86: Success Packages gcc and glibc was upgraded in this update: [...] [ebuild U ] sys-devel/gcc-3.4.4-r1 [3.4.3-r1] (-altivec) -bootstrap -boundschecking -build +fortran -gcj +gtk -hardened -ip28 (-multilib) -multislot (-n32) (-n64) +nls -nocxx -nopie -nossp -objc -static -vanilla 27,036 kB [ebuild U ] sys-libs/glibc-2.3.5-r1 [2.3.5] -build -erandom -glibc-compat20 -glibc-omitfp -hardened -linuxthreads-tls (-multilib) +nls -nptl -nptlonly -pic -profile (-selinux) +userlocales 1,267 kB [...] Before update I've made "backup" package of glibc (2.3.5) with quickpkg command. When I emerged it, everything is working fine. I'm using 2005.0 profile from the begining. I've never used 2004.x. # emerge --info Portage 2.0.51.22-r2 (default-linux/amd64/2005.0, gcc-3.4.4, glibc-2.3.5-r0, 2.6.12-gentoo-r6 x86_64) ================================================================= System uname: 2.6.12-gentoo-r6 x86_64 AMD Athlon(tm) 64 Processor 3200+ Gentoo Base System version 1.6.13 dev-lang/python: 2.3.5 sys-apps/sandbox: 1.2.10 sys-devel/autoconf: 2.13, 2.59-r6 sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6 sys-devel/binutils: 2.15.92.0.2-r10 sys-devel/libtool: 1.5.18-r1 virtual/os-headers: 2.6.11-r2 ACCEPT_KEYWORDS="amd64" AUTOCLEAN="yes" CBUILD="x86_64-pc-linux-gnu" CFLAGS="-O2 -march=athlon64 -fomit-frame-pointer -fweb -pipe" CHOST="x86_64-pc-linux-gnu" CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3.3/env /usr/kde/3.3/share/config /usr/kde/3.3/shutdown /usr/kde/3.4/env /usr/kde/3.4/share/config /usr/kde/3.4/shutdown /usr/kde/3/share/config /usr/lib/X11/xkb /usr/share/config /var/qmail/control" CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d" CXXFLAGS="-O2 -march=athlon64 -fomit-frame-pointer -fweb -pipe" DISTDIR="/usr/portage/distfiles" FEATURES="autoconfig distlocks sandbox sfperms strict" GENTOO_MIRRORS="http://ftp.belnet.be/mirror/rsync.gentoo.org/gentoo http://distfiles.gentoo.org http://www.ibiblio.org/pub/Linux/distributions/gentoo" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/opt/reserve/install/portage/hs /opt/reserve/install/portage/mplayer32" SYNC="rsync://rsync.gentoo.org/gentoo-portage" USE="amd64 3dnowex X Xaw3d a52 aac aalib aim alsa avi bash-completion berkdb bitmap-fonts bluetooth bonobo bzlib cdparanoia cdr crypt cups curl dga dts dv dvd dvdr edl emacs emboss encode ethereal exif fbcon ffmpeg flac flash font-server foomaticdb fortran ftp gdbm gif gimpprint gphoto2 gpm gtk gtk2 gtkhtml guile iconv icq imagemagic imagemagick imap imlib ipv6 jabber jack jbig jpeg junit ladcca lcms leim libcaca libwww live lm_sensors lzw lzw-tiff mad maildir matroska mikmod mime mng motif mozsvg mp3 mpeg msn mule ncurses network new-login nls nowin nvidia ofx ogg oggvorbis opengl oscar oss pam pda pdflib perl png pnp portaudio ppds python qt quicktime readline real recode rtc samba sdl slang snmp sockets sox speex spell ssl svg tcltk tcpd theora tiff truetype truetype-fonts type1-fonts unicode usb userlocales vcd vorbis wmf wxwindows xface xml xml2 xmms xosd xpm xscreensaver xv xvmc yahoo zlib userland_GNU kernel_linux elibc_glibc" Unset: ASFLAGS, CTARGET, LANG, LC_ALL, LDFLAGS, LINGUAS
I found another workaround in gentoo-amd64 mailing list archive (http://marc.theaimsgroup.com/?l=gentoo-amd64&m=112329353322198&w=2). It can be more helpful. It is about that symlink /lib/ld-linux.so.2 points to wrong destination. Situation in glibc-2.3.5: lrwxrwxrwx 1 root root 22 Sep 11 18:02 /lib/ld-linux.so.2 -> ../lib32/ld-linux.so.2 Situation in glibc-2.3.5-r1: lrwxrwxrwx 1 root root 11 Sep 11 18:21 /lib/ld-linux.so.2 -> ld-2.3.5.so
*** Bug 108248 has been marked as a duplicate of this bug. ***
can someone run this: emerge glibc >& log and then post the log file as an attachment
Created attachment 69993 [details] emerge glibc
Comment on attachment 69993 [details] emerge glibc sorry, lemme be more specific ... can someone post a log of glibc emerging which results in the system breaking ... it has to show the compile/install/merge stages in other words if anyone has portage logging enabled, you could find probably find one in /var/log/portage/ ...
Created attachment 70005 [details] 3086-glibc-2.3.5-r2.log I hope this is what you needed. It's made with PORT_LOGDIR set in make.conf.
Created attachment 70006 [details] emerge glibc-2.3.5-r1 This is full log of emerge glibc-2.3.5-r1 which produces wrong symlink and breaks 32bit compatibility. # ll /lib/ld-linux.so.2 lrwxrwxrwx 1 root root 11 Oct 6 18:44 /lib/ld-linux.so.2 -> ld-2.3.5.so While installing it seems that symlink is created correctly but then overwritten with wrong symlink: # grep "ld-linux.so.2 ->" emerge-glibc.log >>> /lib/ld-linux.so.2 -> ../lib32/ld-linux.so.2 >>> /lib32/ld-linux.so.2 -> ld-2.3.5.so >>> /lib64/ld-linux.so.2 -> ld-2.3.5.so
err, what does this show on your machine ? ls -ld /lib*
yeah, that last bit shouldnt exist ... correct (i.e. what i have): /lib/ld-linux.so.2 -> ../lib32/ld-linux.so.2 /lib32/ld-2.3.5.so /lib32/ld-linux.so.2 -> ld-2.3.5.so /lib64/ld-2.3.5.so /lib64/ld-linux-x86-64.so.2 -> ld-2.3.5.so wrong (i.e. what you guys are being screwed with): /lib/ld-linux.so.2 -> ../lib32/ld-linux.so.2 /lib32/ld-2.3.5.so /lib32/ld-linux.so.2 -> ld-2.3.5.so /lib64/ld-2.3.5.so /lib64/ld-linux-x86-64.so.2 -> ld-2.3.5.so /lib64/ld-linux.so.2 -> ld-2.3.5.so
$ ls -ld /lib* lrwxrwxrwx 1 root root 5 Aug 13 13:18 /lib -> lib64 drwxr-xr-x 2 root root 4096 Oct 6 19:58 /lib32 drwxr-xr-x 9 root root 4096 Oct 6 19:58 /lib64
Robert, thanks for the link-repair workaround the bug was reported a month ago ... how is it possible that my system still suffered from it after a fresh (today) update, especially when the workaround is so easy? :-( well, I do not want to say that the Gentoo developers are doing bad job, no, I would like to thank them a lot, but sometimes, a bit more testing and repairing would be appreticiated more ... even if we do not want to end like stable Debian ;-)
if you dont have anything helpful to add, dont bother commenting
Created attachment 70824 [details] emerge glibc >& emerge-glibc.log
Experiencing same problem here, except that I don't have either of these files: /lib32/ld-2.3.5.so /lib32/ld-linux.so.2 -> ld-2.3.5.so glibc-2.3.5-r2 /lib64 symlinked to /lib #ls -l /lib/ld-* -rwxr-xr-x 1 root root 113669 Oct 16 22:59 /lib/ld-2.3.5.so lrwxrwxrwx 1 root root 11 Oct 16 22:59 /lib/ld-linux-x86-64.so.2 -> ld-2.3. 5.so #ls -l /lib32/* -rwxr-xr-x 1 root root 28724 Apr 19 18:24 /lib32/libsandbox.so obviously a lot of stuff is missing.. I emerged most of the emul- stuff just in case even though I think it's not relevant since 2005.0.. I just attached: emerge glibc >& emerge-glibc.log Sorry if it's before comment, it's my first attach. in bugzilla.. hmm.. #emerge info Portage 2.0.51.22-r3 (default-linux/x86/2005.1, gcc-3.4.4, glibc-2.3.5-r2, 2.6. 12-gentoo-r9becks x86_64) ================================================================= System uname: 2.6.12-gentoo-r9becks x86_64 AMD Athlon(tm) 64 Processor 3000+ Gentoo Base System version 1.6.13 dev-lang/python: 2.3.5-r2, 2.4.2 sys-apps/sandbox: 1.2.11 sys-devel/autoconf: 2.13, 2.59-r6 sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r1 sys-devel/binutils: 2.15.92.0.2-r10 sys-devel/libtool: 1.5.20 virtual/os-headers: 2.6.11-r2 ACCEPT_KEYWORDS="amd64 x86" AUTOCLEAN="yes" CBUILD="x86_64-pc-linux-gnu" CFLAGS="-march=k8 -O2 -pipe" CHOST="x86_64-pc-linux-gnu" CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3.4/env /usr/kde/3.4/ share/config /usr/kde/3.4/shutdown /usr/kde/3/share/config /usr/lib/X11/xkb / usr/lib64/mozilla/defaults/pref /usr/share/config /var/qmail/control" CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d" CXXFLAGS="-march=k8 -O2 -pipe" DISTDIR="/usr/portage/distfiles" FEATURES="autoconfig distlocks sandbox sfperms strict" GENTOO_MIRRORS="http://mirror.datapipe.net/gentoo http://mirrors.tds.net/gentoo ftp://mirrors.tds.net/gentoo" LANG="en_CA.UTF-8" LC_ALL="en_CA.UTF-8" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" SYNC="rsync://rsync.gentoo.org/gentoo-portage" USE="x86 3dnow 3dnowext X aac aalib acpi alsa apm arts avi berkdb bitmap-fonts bonobo cdr crypt cups curl divx4linux doc dvd dvdread eds emboss encode esd fam fame ffmpeg flac foomaticdb fortran gd gdbm gif gimpprint gnome gnutls gphoto2 gpm gstreamer gtk gtk2 gtkhtml imagemagick imlib inkjar ipv6 java joystick jpeg junit kde ldap libg++ libwww mad matroska mikmod mng motif mozilla mp3 mpeg musicbrainz mysql ncurses nls offensive ogg oggvorbis opengl oss pam pdflib perl plugin png ppds python qt quicktime readline real ruby samba sdl slang spell ssl svg tcltk tcpd theora tiff truetype truetype-fonts type1-fonts unicode userlocales utf8 v4l v4l2 visualization vorbis win32codecs xine xinerama xml xml2 xmms xv xvid zlib userland_GNU kernel_linux elibc_glibc" Unset: ASFLAGS, CTARGET, LDFLAGS, LINGUAS, PORTDIR_OVERLAY
Ignore my report (Olivier Bilodeau). I made a silly mistake when I switched to profile 2005.1. I pointed to [...]/ x86/2005.1 instead of [...]/amd64/2005.1 as you can see in my emerge info. Sorry for the bad report.
*** Bug 110773 has been marked as a duplicate of this bug. ***
Created attachment 71714 [details] log of the configure script Here is config.log, size of long double test fails with seg fault. If you need additional info let me know.
Checked symlinks(https://bugs.gentoo.org/show_bug.cgi?id=105535#c9). 1) /lib in my system is symlink for /lib64 2) /lib64/ld-linux.so.2 is symlink to ../lib32/ld-linux.so.2 I tried to change this symlink to cubic-pc lib64 # ls -ld ld* -rwxr-xr-x 1 root root 107498 Окт 14 20:11 ld-2.3.5.so lrwxrwxrwx 1 root root 11 Окт 30 13:21 ld-linux.so.2 -> ld-2.3.5.so lrwxrwxrwx 1 root root 11 Окт 14 20:12 ld-linux-x86-64.so.2 -> ld-2.3.5.so This didn't help, I still couldn't emerge glibc. Also want to note that I don't see problems with 32bit applications, at least openoffice-bin and acroread work as expected with no problems.
*** Bug 112558 has been marked as a duplicate of this bug. ***
Having the exact same problem as Ruslan U. Zakirov. No 32bit troubles, glibc is actually failing to build, not producing bad symlinks.
Just for the information: problem (as described in original post) remains also in glibc-2.3.5-r2.
*** Bug 115549 has been marked as a duplicate of this bug. ***
I've fixed my issue with help from SpanKY. Big thanks. Issue on my system was that I had /usr/lib64 as relative symlink to other place, as I've changed symlink to 'mount --bind' solution and it all works. Also I've found thread http://forums.gentoo.org/viewtopic-t-293507.html that could be very helpful.
*** Bug 118135 has been marked as a duplicate of this bug. ***
I've debugged it and found the reason of this bug. The ground of the problem is different sorting behavior in sk_SK (and probably also cs_CZ) locale. I do not know if it is a bug but in sk_SK locale the letters have precedence before digits, and "." have precedence before "-": $ LC_COLLATE=C ls ld-2.3.5.so ld-linux.so.2 libthread_db-1.0.so libthread_db.so.1 $ LC_COLLATE=sk_SK ls ld-linux.so.2 ld-2.3.5.so libthread_db.so.1 libthread_db-1.0.so Unfortunately glibc ebuild relies on this. In function toolchain-glibc_src_install(), just before 'prepallstrip', are loops where some libs are temporarily moved to thread-backup dir for x in ${D}$(alt_libdir)/lib{pthread,thread_db}* \ ${D}$(alt_libdir)/ld-* ; do [[ -f ${x} ]] && mv -f ${x} ${T}/thread-backup/ done if want_linuxthreads && want_nptl ; then mkdir -p ${T}/thread-backup/tls for x in ${D}$(alt_libdir)/tls/lib{pthread,thread_db}* ; do [[ -f ${x} ]] && mv -f ${x} ${T}/thread-backup/tls done fi env -uRESTRICT CHOST=${CTARGET} prepallstrip cp -a -- ${T}/thread-backup/* ${D}$(alt_libdir)/ || die In the "correct" environment (C, POSIX locale) the order in which files are moved is: libpthread-0.10.so libpthread.so.0 (symlink to libpthread-0.10.so) libthread_db-1.0.so libthread_db.so.1 (symlink to libthread_db-1.0.so) ld-2.3.5.so ld-linux-x86-64.so.2 (symlink to ld-2.3.5.so) So, the libpthread-0.10.so file is moved first, then the symlink libpthread.so.0 is about to be moved but is not, because is not valid anymore, and so on. In the end, regular files are moved but symlinks not. In the "incorrect" environment (LC_COLLATE=sk_SK) the order is that symlinks are moved _before_ regular files (because of "wrong" sorting), thus in the end, regular files _and_ symlinks are moved. That's it. This causes that symlink ld-linux.so.2 appears in image/lib64 directory and later in qmerge phase replaces the correct one. Strange is, that 'unset LC_COLLATE' in the ebuild does not help. I've tried it - the variable is unset (I've chcecked it by printing its value) but sorting have still sk_SK behavior (bash magic?!?). But 'LC_COLLATE=C' _works_ (see diff below). Unset this variable in the command line (before running emerge command), works also. This is diff -p glibc-2.3.5-r2.ebuild glibc-2.3.5-r2.ebuild.new *************** toolchain-glibc_src_install() { *** 284,289 **** --- 284,290 ---- # These should not be set, else the # zoneinfo do not always get installed ... unset LANGUAGE LANG LC_ALL + LC_COLLATE=C if want_linuxthreads ; then cd "${WORKDIR}"/build-${ABI}-${CTARGET}-linuxthreads Hope this helps. ;-)
that code no longer exists in newer glibc as i rewrote it for other things try one of the ebuilds which has a 'src_strip' func in it also, rather than set specific variables, might as well just export LC_ALL=C
Great! Bug is not present in recent ebuilds. After sync and re-emerging glibc-2.3.5-r2, 32-bit apps works fine.
thanks for testing
