-march=c3 gets clobbered with -march=i686 which will generate invalid code. With 2.6.1 and 2.7: * Installing Hardened Gentoo SSP handler >>> Source unpacked. >>> Compiling source in /var/tmp/portage/sys-libs/glibc-2.7/work/glibc-2.7 ... * ABI: default * CBUILD: i686-pc-linux-gnu * CHOST: i686-pc-linux-gnu * CTARGET: i686-pc-linux-gnu * CBUILD_OPT: * CTARGET_OPT: * CC: * CFLAGS: -march=i686 -pipe -O2 -fno-strict-aliasing -fno-stack-protector With 2.6, -march=i686 was prepended: * Installing Hardened Gentoo SSP handler * Using GNU config files from /usr/share/libtool * Updating scripts/config.sub [ ok ] * Updating scripts/config.guess [ ok ] >>> Source unpacked. >>> Compiling source in /var/tmp/portage/sys-libs/glibc-2.6/work/glibc-2.6 ... * ABI: default * CBUILD: i686-pc-linux-gnu * CHOST: i686-pc-linux-gnu * CTARGET: i686-pc-linux-gnu * CBUILD_OPT: * CTARGET_OPT: * CC: * CFLAGS: -march=i686 -march=c3 -pipe -O2 -fno-strict-aliasing -fno-stack-protector Replacing -march=c3 with -mtune=c3 in my CFLAGS leads to CFLAGS: -march=i686 -mtune=c3 ... Not sure what kind of binary I'd get oreilly ~ # emerge --info Portage 2.1.3.19 (hardened/x86/2.6, gcc-3.4.6, glibc-2.6-r0, 2.6.17-hardened-r1 i686) ================================================================= System uname: 2.6.17-hardened-r1 i686 VIA Ezra Timestamp of tree: Thu, 15 Nov 2007 17:46:01 +0000 app-shells/bash: 3.2_p17-r1 dev-lang/python: 2.4.4-r4, 2.5.1-r3 dev-python/pycrypto: 2.0.1-r6 sys-apps/baselayout: 1.12.10-r4 sys-apps/sandbox: 1.2.18.1-r2 sys-devel/autoconf: 2.13, 2.61-r1 sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2, 1.10 sys-devel/binutils: 2.17 sys-devel/gcc-config: 1.3.16 sys-devel/libtool: 1.5.24 virtual/os-headers: 2.6.23-r1 ACCEPT_KEYWORDS="x86 ~x86" CBUILD="i686-pc-linux-gnu" CFLAGS="-march=c3 -Os -pipe -fomit-frame-pointer" CHOST="i686-pc-linux-gnu" CONFIG_PROTECT="/etc" CONFIG_PROTECT_MASK="/etc/env.d /etc/gconf /etc/revdep-rebuild /etc/terminfo" CXXFLAGS="-march=c3 -Os -pipe -fomit-frame-pointer" DISTDIR="/usr/portage/distfiles" EMERGE_DEFAULT_OPTS="--alphabetical --nospinner" FEATURES="ccache distlocks metadata-transfer parallel-fetch sandbox sfperms strict unmerge-orphans userfetch" GENTOO_MIRRORS="ftp://polly.a.la.maison/gentoo http://www.mirrorservice.org/sites/www.ibiblio.org/gentoo" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --delete-after --stats --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages --filter=H_**/files/digest-*" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" SYNC="rsync://polly/portage" USE="apache2 bash-completion berkdb bzip2 cracklib crypt curl dpms glibc-omitfp hardened imap imlib maildir mailwrapper midi moznocompose moznoirc ncurses network nptl nptlonly pic python readline rtc ruby slang ssl symlink unicode urandom usb userlocales x86 xml2 xorg zlib" ALSA_CARDS="ali5451 als4000 atiixp atiixp-modem bt87x ca0106 cmipci emu10k1 emu10k1x ens1370 ens1371 es1938 es1968 fm801 hda-intel intel8x0 intel8x0m maestro3 trident usb-audio via82xx via82xx-modem ymfpci" ALSA_PCM_PLUGINS="adpcm alaw asym copy dmix dshare dsnoop empty extplug file hooks iec958 ioplug ladspa lfloat linear meter mulaw multi null plug rate route share shm softvol" ELIBC="glibc" INPUT_DEVICES="mouse keyboard" KERNEL="linux" LCD_DEVICES="bayrad cfontz cfontz633 glk hd44780 lb216 lcdm001 mtxorb ncurses text" USERLAND="GNU" VIDEO_CARDS="apm ark chips cirrus cyrix dummy fbdev glint i128 i740 i810 imstt mach64 mga neomagic nsc nv r128 radeon rendition s3 s3virge savage siliconmotion sis sisusb tdfx tga trident tseng v4l vesa vga via vmware voodoo" Unset: CPPFLAGS, CTARGET, INSTALL_MASK, LANG, LC_ALL, LDFLAGS, LINGUAS, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, PORTAGE_RSYNC_EXTRA_OPTS, PORTDIR_OVERLAY
-march=c3 targets i486 which does not match your CHOST which will break things when building glibc
(In reply to comment #1) > -march=c3 targets i486 which does not match your CHOST which will break things > when building glibc Yeah right, say that to the boxes that have been running fine for years... How difficult is it not to override user-defined settings? Portage 2.1.3.19 (hardened/x86/2.6, gcc-3.4.6, glibc-2.6-r0, 2.6.17-hardened-r1 i686) System uname: 2.6.17-hardened-r1 i686 VIA Ezra CFLAGS="-march=c3 -Os -pipe -fomit-frame-pointer" CHOST="i686-pc-linux-gnu" echo '>sys-libs/glibc-2.6' >> /etc/portage/package.mask will do the trick
neither of those statements have any relevance at all user settings are not being ignored, they're purposefully being changed. just because it used to work doesnt mean it will continue to work ... glibc changes code with every release
(In reply to comment #3) > user settings are not being ignored, they're purposefully being changed. just > because it used to work doesnt mean it will continue to work ... glibc changes > code with every release Not purposefully changed, purposefully broken. The only reason it breaks is setup_flags() in /usr/portage/sys-libs/glibc/files/eblits/common.eblit replaces a -march that works with one that does not. Portage 2.1.3.19 (hardened/x86/2.6, gcc-3.4.6, glibc-2.7-r0, 2.6.23-hardened-r1 i686) ================================================================= System uname: 2.6.23-hardened-r1 i686 VIA Ezra Timestamp of tree: Fri, 16 Nov 2007 08:46:01 +0000 app-shells/bash: 3.2_p17-r1 dev-lang/python: 2.4.4-r4, 2.5.1-r3 dev-python/pycrypto: 2.0.1-r6 sys-apps/baselayout: 1.12.10-r5 sys-apps/sandbox: 1.2.18.1-r2 sys-devel/autoconf: 2.13, 2.61-r1 sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2, 1.10 sys-devel/binutils: 2.18-r1 sys-devel/gcc-config: 1.4.0-r4 sys-devel/libtool: 1.5.24 virtual/os-headers: 2.6.23-r1 ACCEPT_KEYWORDS="x86 ~x86" CBUILD="i686-pc-linux-gnu" CFLAGS="-march=c3 -Os -pipe -fomit-frame-pointer" CHOST="i686-pc-linux-gnu" CONFIG_PROTECT="/etc" CONFIG_PROTECT_MASK="/etc/env.d /etc/gconf /etc/revdep-rebuild /etc/terminfo /etc/udev/rules.d" CXXFLAGS="-march=c3 -Os -pipe -fomit-frame-pointer" DISTDIR="/usr/portage/distfiles" EMERGE_DEFAULT_OPTS="--alphabetical --nospinner" FEATURES="ccache distlocks metadata-transfer parallel-fetch sandbox sfperms strict unmerge-orphans userfetch" GENTOO_MIRRORS="ftp://polly.a.la.maison/gentoo http://www.mirrorservice.org/sites/www.ibiblio.org/gentoo" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --delete-after --stats --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages --filter=H_**/files/digest-*" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" SYNC="rsync://polly/portage" USE="apache2 bash-completion berkdb bzip2 cracklib crypt curl dpms glibc-omitfp hardened imap imlib maildir mailwrapper midi moznocompose moznoirc ncurses network nptl nptlonly pic python readline rtc ruby slang ssl symlink unicode urandom usb userlocales x86 xml2 xorg zlib" ELIBC="glibc" KERNEL="linux" USERLAND="GNU" Unset: CPPFLAGS, CTARGET, INSTALL_MASK, LANG, LC_ALL, LDFLAGS, LINGUAS, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, PORTAGE_RSYNC_EXTRA_OPTS, PORTDIR_OVERLAY # cat /proc/cpuinfo processor : 0 vendor_id : CentaurHauls cpu family : 6 model : 7 model name : VIA Ezra stepping : 8 cpu MHz : 800.065 cache size : 64 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu de tsc msr cx8 mtrr pge mmx 3dnow bogomips : 1602.02 clflush size : 32 oreilly ~ # ls -l /lib/libc.so.6 lrwxrwxrwx 1 root root 11 Nov 19 09:12 /lib/libc.so.6 -> libc-2.7.so oreilly ~ # /lib/libc.so.6 GNU C Library stable release version 2.7, by Roland McGrath et al. Copyright (C) 2007 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Compiled by GNU CC version 3.4.6 (Gentoo Hardened 3.4.6-r2, ssp-3.4.6-1.0, pie-8.7.9). Compiled on a Linux >>2.6.23-hardened-r1<< system on 2007-11-18. Available extensions: C stubs add-on version 2.1.2 crypt add-on version 2.1 by Michael Glad and others Gentoo patchset 1.2 GNU Libidn by Simon Josefsson Native POSIX Threads Library by Ulrich Drepper et al Support for some architectures added on, not maintained in glibc core. BIND-8.2.3-T5B
Why does -march=c3 target i486, anyway? The VIA C3 CPU is i686-compatible (minus CMOV, I think)...
(In reply to comment #5) > Why does -march=c3 target i486, anyway? The VIA C3 CPU is i686-compatible > (minus CMOV, I think)... > Not as far as what GCC will do when you specify -march=c3 (I'm assuming you are asking why the flags get changed) {"c3", PROCESSOR_I486, PTA_MMX | PTA_3DNOW}, {"c3-2", PROCESSOR_PENTIUMPRO, PTA_MMX | PTA_SSE},
(In reply to comment #5) > Why does -march=c3 target i486, anyway? The VIA C3 CPU is i686-compatible > (minus CMOV, I think)... > There are multiple versions of the C3 processor. Read http://en.wikipedia.org/wiki/VIA_C3 If you are sure you have a C3 with support for CMOV, then use -march=c3-2 and CHOST="i686-pc-linux-gnu". Otherwise, you have to use -march=c3 and CHOST="i486-pc-linux-gnu". I suggest to change resolution of this bug to INVALID.
(In reply to comment #7) > If you are sure you have a C3 with support for CMOV, then use -march=c3-2 and > CHOST="i686-pc-linux-gnu". Otherwise, you have to use -march=c3 and > CHOST="i486-pc-linux-gnu". There are several things I'm sure of: 1. you did not read the comments above, no worries, I'll sum it up below 2. as shown above, my C3 does not have cmov 3. setting -march to any value for glibc atm is pointless since gentoo's current build clobbers it with the value taken from CHOST 4. -march=c3 is the right CFLAG 5. CHOST="i686-pc-linux-gnu" need not be a problem 6. replacing a -march that works with one that does not is what this bug is about 7. downgrading plenty of -march values to i686 is fine with me, I can understand it saves a lot of potential problems but upgrading -march from c3 to i686 is not OK. I hope this sums it up in a clear and concise way. I had upgraded to 2.7 by hacking out the x86 case in setup_flags() in /usr/portage/sys-libs/glibc/files/eblits/common.eblit No issue for months on a box that is up 24x7 Now running 2.8 # /lib/libc.so.6 GNU C Library stable release version 2.8, by Roland McGrath et al. ....
(In reply to comment #8) > There are several things I'm sure of: > 1. you did not read the comments above, no worries, I'll sum it up below > 2. as shown above, my C3 does not have cmov > 3. setting -march to any value for glibc atm is pointless since gentoo's > current build clobbers it with the value taken from CHOST > 4. -march=c3 is the right CFLAG > 5. CHOST="i686-pc-linux-gnu" need not be a problem Sure it is. It causes bug #199334. > 6. replacing a -march that works with one that does not is what this bug is > about > 7. downgrading plenty of -march values to i686 is fine with me, I can > understand it saves a lot of potential problems but upgrading -march from c3 to > i686 is not OK. Then use a CHOST that reflects the architecture you're on. http://www.gentoo.org/doc/en/change-chost.xml
(In reply to comment #9) > > 5. CHOST="i686-pc-linux-gnu" need not be a problem > > Sure it is. It causes bug #199334. Which would be fixed as soon as you fix this very bug > > 7. downgrading plenty of -march values to i686 is fine with me, I can > > understand it saves a lot of potential problems but upgrading -march from c3 to > > i686 is not OK. > Then use a CHOST that reflects the architecture you're on. I am and I am not changing my CHOST even though it would be faster than getting you guys understand. Don't bother, removing Gentoo's defective code is easy enough to make it work whenever I need to upgrade glibc on this box.
*** Bug 250527 has been marked as a duplicate of this bug. ***
*** Bug 225237 has been marked as a duplicate of this bug. ***
too many arch<->chost crap combos out there
Use a CHOST that makes sense and this isn't a problem. This has gone no where, and I'm not going to leave it open to rot.
should be fixed by using a bit more specific test http://sources.gentoo.org/sys-libs/glibc/files/eblits/common.eblit?r1=1.19&r2=1.20
