I was building a stage1 with a snapshot from this morning, and got the following error: i386-pc-linux-gnu-gcc -nostdlib -nostartfiles -o /var/tmp/portage/glibc-2.5/work /build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/rpcinfo -Wl,-dynamic-linke r=/lib/ld-linux.so.2 -Wl,-z,combreloc -Wl,-z,relro /var/tmp/portage/glibc-2.5 /work/build-default-i386-pc-linux-gnu-linuxthreads/csu/crt1.o /var/tmp/portage/g libc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/csu/crti.o `i386-pc-l inux-gnu-gcc --print-file-name=crtbegin.o` /var/tmp/portage/glibc-2.5/work/build -default-i386-pc-linux-gnu-linuxthreads/sunrpc/rpcinfo.o -Wl,-rpath-link=/var/t mp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads:/var/tmp/ portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/math:/var/tm p/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/elf:/var/t mp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/dlfcn:/va r/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/nss:/v ar/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/nis:/ var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/rt:/ var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/reso lv:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/ crypt:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthrea ds/linuxthreads /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu- linuxthreads/libc.so.6 /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-lin ux-gnu-linuxthreads/libc_nonshared.a -lgcc -Wl,--as-needed -lgcc_s -Wl,--no-as- needed `i386-pc-linux-gnu-gcc --print-file-name=crtend.o` /var/tmp/portage/glibc -2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/csu/crtn.o CPP='i386-pc-linux-gnu-gcc -E -x c-header' /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/elf/ld-linux.so.2 --library-path /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/math:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/elf:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/dlfcn:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/nss:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/nis:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/rt:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/resolv:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/crypt:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/linuxthreads /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/rpcgen -Y ../scripts -c rpcsvc/bootparam_prot.x -o /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/xbootparam_prot.T make[2]: *** [/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/xbootparam_prot.stmp] Segmentation fault make[2]: Leaving directory `/var/tmp/portage/glibc-2.5/work/glibc-2.5/sunrpc' make[1]: *** [sunrpc/others] Error 2 make[1]: Leaving directory `/var/tmp/portage/glibc-2.5/work/glibc-2.5' make !!! ERROR: sys-libs/glibc-2.5 failed. Call stack: ebuild.sh, line 1539: Called dyn_compile ebuild.sh, line 939: Called src_compile glibc-2.5.ebuild, line 1119: Called toolchain-glibc_src_compile glibc-2.5.ebuild, line 235: Called die !!! make for default failed !!! If you need support, post the topmost build error, and the call stack if relevant. : *** [all] Error 2 This was using a 2006.1 x86 stage3 as a seed stage in catalyst. The x86/no-nptl profile does not mask glibc-2.5, which was recently marked stable on x86.
I got a slightly different segfault, but same issue here. make[2]: *** [/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/xnlm_prot.stmp] Segmentation fault ... make[2]: *** [/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/xbootparam_prot.stmp] Segmentation fault make[2]: *** [/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/xrstat.stmp] Segmentation fault ...make[2]: *** [/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/xyppasswd.stmp] Segmentation fault ... CPP='i386-pc-linux-gnu-gcc -E -x c-header' /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/elf/ld-linux.so.2 --library-path /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/math:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/elf:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/dlfcn:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/nss:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/nis:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/rt:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/resolv:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/crypt:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/linuxthreads /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/rpcgen -Y ../scripts -h rpcsvc/rstat.x -o /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/rpcsvc/rstat.T make[2]: *** [/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/rpcsvc/bootparam_prot.stmp] Segmentation fault make[2]: *** [/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/rpcsvc/nlm_prot.stmp] Segmentation fault make[2]: *** [/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/rpcsvc/rstat.stmp] Segmentation fault make[2]: Leaving directory `/var/tmp/portage/glibc-2.5/work/glibc-2.5/sunrpc' make[1]: *** [sunrpc/others] Error 2 make[1]: Leaving directory `/var/tmp/portage/glibc-2.5/work/glibc-2.5' make: *** [all] Error 2 !!! ERROR: sys-libs/glibc-2.5 failed. Call stack: ebuild.sh, line 1546: Called dyn_compile ebuild.sh, line 937: Called src_compile glibc-2.5.ebuild, line 1119: Called toolchain-glibc_src_compile glibc-2.5.ebuild, line 235: Called die !!! make for default failed !!! If you need support, post the topmost build error, and the call stack if relevant. Here is my current profile: /etc/make.profile -> /usr/portage/profiles/default-linux/x86/no-nptl/2.4/ CHOST="i386-pc-linux-gnu" CFLAGS="-O2 -march=i386 -pipe" CXXFLAGS="${CFLAGS}" Wendall
I was building too a stage1 with a snapshot from this afternoon, and got the same following error: CPP='i386-pc-linux-gnu-gcc -E -x c-header' /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/elf/ld-linux.so.2 --library-path /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/math:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/elf:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/dlfcn:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/nss:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/nis:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/rt:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/resolv:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/crypt:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/linuxthreads /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/rpcgen -Y ../scripts -c rpcsvc/bootparam_prot.x -o /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/xbootparam_prot.T make[2]: *** [/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/xbootparam_prot.stmp] Segmentation fault make[2]: Leaving directory `/var/tmp/portage/glibc-2.5/work/glibc-2.5/sunrpc' make[1]: *** [sunrpc/others] Error 2 make[1]: Leaving directory `/var/tmp/portage/glibc-2.5/work/glibc-2.5' make: *** [all] Error 2 !!! ERROR: sys-libs/glibc-2.5 failed. Call stack: ebuild.sh, line 1539: Called dyn_compile ebuild.sh, line 939: Called src_compile glibc-2.5.ebuild, line 1119: Called toolchain-glibc_src_compile glibc-2.5.ebuild, line 235: Called die !!! make for default failed !!! If you need support, post the topmost build error, and the call stack if relevant. I'm using a 2006.1 stage3 as a seed stage in catalyst and the x86/no-nptl profile.
modernizing the system with "emerge -Du world", I have the following error : CPP='i386-pc-linux-gnu-gcc -E -x c-header' /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/elf/ld-linux.so.2 --library-path /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/math:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/elf:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/dlfcn:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/nss:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/nis:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/rt:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/resolv:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/crypt:/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/linuxthreads /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/rpcgen -Y ../scripts -c rpcsvc/bootparam_prot.x -o /var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/xbootparam_prot.T make[2]: *** [/var/tmp/portage/glibc-2.5/work/build-default-i386-pc-linux-gnu-linuxthreads/sunrpc/xbootparam_prot.stmp] Segmentation fault make[2]: Leaving directory `/var/tmp/portage/glibc-2.5/work/glibc-2.5/sunrpc' make[1]: *** [sunrpc/others] Error 2 make[1]: Leaving directory `/var/tmp/portage/glibc-2.5/work/glibc-2.5' make: *** [all] Error 2 !!! ERROR: sys-libs/glibc-2.5 failed. Call stack: ebuild.sh, line 1539: Called dyn_compile ebuild.sh, line 939: Called src_compile glibc-2.5.ebuild, line 1119: Called toolchain-glibc_src_compile glibc-2.5.ebuild, line 235: Called die !!! make for default failed !!! If you need support, post the topmost build error, and the call stack if relevant.
OK, guys. Is it really necessary to post the same stinking failure over and over? Just add yourselves to CC and have some patience. If your failure is different than the one posted in this bug originally, then it is a *different* bug. Either way, there's no need for adding more noise. Thanks
(In reply to comment #5) > OK, guys. Is it really necessary to post the same stinking failure over and > over? Just add yourselves to CC and have some patience. If your failure is > different than the one posted in this bug originally, then it is a *different* > bug. Either way, there's no need for adding more noise. > > Thanks > excuse if I have repeated the bugs, I thought could help the fact that to me comes later on given the same error "emerge -Du world. Regards
This seems to be an odd bug. I encountered this first on a workstation that is used for imaging the computer lab that I help manage, but we run another Gentoo LTSP server with the same arch/settings/etc. so I figured I'd try it there. The build succeeded, and then when I went back to the workstation and tried it again, it built successfully. I'm not sure what's causing the segfault, but it seems to be intermittent and/or is influenced by the good vibes of trying to merge it elsewhere.
Same for me. glibc-2.5 segfault at the same moment on my xen domU (keyword x86). Segfault appears with both gcc-3.4.6-r2 and gcc-4.1.1-r3. summary of emerge --info : sys-devel/binutils: 2.16.1-r3 sys-devel/gcc-config: 1.3.14 sys-devel/libtool: 1.5.22 virtual/os-headers: 2.6.17-r2 ACCEPT_KEYWORDS="x86" CBUILD="i386-pc-linux-gnu" CFLAGS="-O2 -mtune=i686 -pipe" CHOST="i386-pc-linux-gnu"
In pentium 4s and such x686, the fault seems to be related to having set the wrong CHOST variable. Changing it to i686 instead of i386 and doing a emerge -av1 binutils gcc glibc && emerge -e system && emerge -e world and checking right routes like it is shown in http://www.gentoo.org/doc/en/change-chost.xml seems to fix the problem
The wrong CHOST alone shouldnt be the problem. I had the same problem, CHOST was set to i386 on pentium 3, i used 2006.1 stage 3 aswell. My USE-Flag included "nptl nptlonly" which is a nono with i386, dunno if that is the cause for the segfaults though. I changed the CHOST and am currently "reemerging" world. Don't ask me though why I downloaded the i386 stage3 instead of the i686 =)
Comment #9 solved my problem too. IMHO the problem could be checked in the ebuild of glibc and error the user giving him the choice to either disable nptl or changing the CHOST. Link to http://www.gentoo.org/doc/en/change-chost.xml should also be given. Thanks PS : for me the bug is closed since glibc now compiles
Comments #8 - #11 don't deal with this bug. nptl has not worked with i386 and glibc for some time now. This segfault happens with i386 regardless. i586 would also fix these compile errors. This may be related, but not the reason for the bug. With USE="-nptl nptlonly" and CHOST i386 this will still segfault. Wendall
OK, guys, please try to refrain from comments that aren't related to *my* reported bug. My bug is that glibc 2.5 fails to build with CHOST=i386-* and *without* NPTL, as shown by my use of the no-nptl profile. If your situation is not identical, it isn't the same bug. It also doesn't help get this bug solved to simply add noise.
While this isn't "fixed" we've decided to just use CHOST=i486-* for the stage1 for 2007.0, so I'm removing the snapshot tracker from this bug.
i386-* is no longer supported; upgrade to i486-*
Sorry, but you can't just "upgrade" to i486-*. There is a process by which you can change CHOST, which I've done many times, but far from a simple upgrade. It would be far better when deprecating an entire processor class in an arch to at least have the courtesy of providing some instructions on how to do this. I really do not see why this can't be patched, but since I don't make decisions here... This issue has been on all the previous builds of glibc, and would be trivial to do so here. If you want I can dig up a patch for this, or writeup some instructions for those effected if it is deemed by the local glibc gods that it won't be patched to compile properly. Wendall
you obviously dont know glibc as for CHOST changing, why dont you read our documentation http://www.gentoo.org/doc/en/change-chost.xml as for "deprecating an entire processor class", dont be such a drama queen ... name me the last time you saw a real i386 processor running Linux ? cant ? that's fine, i'm not surprised people who want to run on real i386 processors can use the uClibc stages which are better suited for the low cpu/low mem setup anyways
Sorry to sound like a "drama queen". Thanks for the reference to the doc. This doc however will only give you a working system maybe 50% of the time. A couple things missing: 1. It isn't stressed enough the importance of fix_libtool_files.sh It must be run repeatedly until nothing gets fixed. I typically do this: fix_libtool_files.sh on any older versions of gcc fix_libtool_files.sh same again with --oldarch {your_old_arch} fix_libtool_files.sh on current version of gcc fix_libtool_files.sh on current version with --oldarch {your_old_arch} 2. By only rebuilding binutils gcc glibc you often end up with a broken make and unable to compile packages. To get around this, I typically build binary packages for all effected packages such as: binutils-config gnuconfig m4 bison libintl gcc-config autoconf autoconf-wrapper automake pkgconfig make flex gdb patch glibc and emerge the binaries. Also iirc apache with apr gets missed with revdep-rebuild. That may be fixed now. 3. You can kill python and perl. emerge -e perl python and perl-cleaner reallyall is almost necessary. 4. The doc asks to check the env without mentioning that you might want to use env-update. 5. Instead of an after thought, running revdep-rebuild for libstdc++.so.5 and libstdc++.so.6 is a good idea, since it does catch problems on systems that have been around for a while. I hope the profile gets updated for this so it will mask glibc-2.5*. I realize that all these steps can be extraneous most of the time. But I have done this quite a bit and having a working system after changing CHOST is the important thing. Wendall
(In reply to comment #17) > name me the last time you saw a real i386 processor running Linux ? Yesterday! ;-) No, really, I'm trying to build a Gentoo system for an old 386 using a chroot on my Athlon! The hint with uClibc is fine, maybe I'll try that. Meanwhile I masked >=glibc-2.5 and now I'm compiling 2.3.6. If that works, maybe I won't have to switch to uClibc. But I'm still curious about why glibc won't compile for 386 anymore. What's the reason?
Mostly, it's because nobody cares.
i386 was already going to be deprecated by upstream ... we just did it a version sooner than previously planned because as noted, we dont really care to fix it the i386 architecture lacks a lot of locking primitives that were introduced starting with the i486 class