Ive just finished the bootstrap stage of my first gentoo install. I encountered some problems along the way, which you might find interesting. This install was on a 200MHz k6, 160M ram, 400M swap. 1. I booted from the 16Mb cd image, and proceeded with the install instructions. I edited make.conf with -march=i586 -mcpu=k6 -O3, and ran bootstrap.sh. Part way through (2 hours or so; Im not sure exactly what it was doing) make reported that some process (sorry; I didnt note which one) terminated with signal 11. 2. I repeated bootstrap.sh. This time is reported some compile error. It looked like the interesting details about *what* was wrong had scrolled off the top of the terminal, so I decided to repeat the process running chroot inside my normal RedHat install. (xterm has a bigger scroll buffer ;-) 3. This time bootstrap.sh terminated with a *different* compile error. I kept this one, appended below. 4. I figured maybe this problem could be due to gcc's -mcpu=k6 optimisation. That cant be the most tested optimisation mode. I changed make.conf to -march=i586 -O3, and re-ran bootstrap.sh. This time make reported during build of glibc that the compiler termined with signal 4, and syslog showed it had been hit by the oom killer. 5. I guessed maybe one of those early abortive compiles had left a broken gcc, so I re-extracted the stage 1 compressed tarball, I also changed the make.conf optimisation setting to -Os (using less memory, although this should not have been a problem, and building faster too). This time bootstrap.sh completed successfully (and "emerge system" is well underway) I hope you find this useful. error from stage 3: gcc ansi_x3.110.c -c -O2 -O3 -Wall -Winline -Wstrict-prototypes -Wwrite-strings -march=i586 -pipe -fPIC -I../include -I. -I/var/tmp/portage/glibc-2.2.5-r4/work/glibc-2.2.5/buildhere/iconvdata -I.. -I../libio -I/var/tmp/portage/glibc-2.2.5-r4/work/glibc-2.2.5/buildhere -I../sysdeps/i386/elf -I../linuxthreads/sysdeps/unix/sysv/linux/i386 -I../linuxthreads/sysdeps/unix/sysv/linux -I../linuxthreads/sysdeps/pthread -I../sysdeps/pthread -I../linuxthreads/sysdeps/unix/sysv -I../linuxthreads/sysdeps/unix -I../linuxthreads/sysdeps/i386/i586 -I../linuxthreads/sysdeps/i386 -I../sysdeps/unix/sysv/linux/i386 -I../sysdeps/unix/sysv/linux -I../sysdeps/gnu -I../sysdeps/unix/common -I../sysdeps/unix/mman -I../sysdeps/unix/inet -I../sysdeps/unix/sysv/i386 -I../sysdeps/unix/sysv -I../sysdeps/unix/i386 -I../sysdeps/unix -I../sysdeps/posix -I../sysdeps/i386/i586 -I../sysdeps/i386/i486 -I../sysdeps/i386/fpu -I../sysdeps/i386 -I../sysdeps/wordsize-32 -I../sysdeps/ieee754/ldbl-96 -I../sysdeps/ieee754/dbl-64 -I../sysdeps/ieee754/flt-32 -I../sysdeps/ieee754 -I../sysdeps/generic/elf -I../sysdeps/generic -D_LIBC_REENTRANT -include ../include/libc-symbols.h -DPIC -DSHARED -o /var/tmp/portage/glibc-2.2.5-r4/work/glibc-2.2.5/buildhere/iconvdata/ansi_x3.110.os {standard input}: Assembler messages: {standard input}:2711: Error: invalid character '-' in mnemonic {standard input}:2962: Error: no such instruction: `cmpkl 20(%ebp),%edi' {standard input}:2963: Error: unbalanced parenthesis in operand 1. make[2]: *** [/var/tmp/portage/glibc-2.2.5-r4/work/glibc-2.2.5/buildhere/iconvdata/ansi_x3.110.os] Error 1 make[2]: *** Waiting for unfinished jobs.... make[2]: Leaving directory `/var/tmp/portage/glibc-2.2.5-r4/work/glibc-2.2.5/iconvdata' make[1]: *** [iconvdata/others] Error 2 make[1]: Leaving directory `/var/tmp/portage/glibc-2.2.5-r4/work/glibc-2.2.5' make: *** [all] Error 2 !!! ERROR: The ebuild did not complete successfully. !!! Function src_compile, Line 14, Exitcode 2 !!! (no error message) !!! emerge aborting on /usr/portage/sys-libs/glibc/glibc-2.2.5-r4.ebuild .
what gcc version?
gcc version 2.95.3-r7 (which is what portage gave me)
Hrm, Only reason I can think of is the unmatched values you had for -march and -mcpu (-march=i586 -mcpu=k6). If you have some time, could you try the same steps as the first time you had the errors, and try with "-march=k6 -mcpu=k6", or just "-mcpu=k6" if possible ? Thanks
I could go right back to the beginning and try -mcpu=k6, but that will take some time. This machine is _still_ busy building an installation that I hope to use. Alternatively, I could recompile gcc with -mcpu=k6 and continue emerging. Would that be interesting? This machine has been busy today with "emerge system" using "-march=i586 -Os -fomit-frame-pointer". It got half way through and terminated with a 'file not found' message. A superficial glance at the make output suggested that file should have been present. Ive seen similar wierdness with parallel makes before, so I have added a MAKEOPTS="-j1". so far no problems. Im not sure if that could have caused the other problems....
I have transplanted the disk containing this install into a different machine; an Athlon XP 1700. Ive tried re-bootstrapping with various optimisation settings, but couldnt reproduce any problems. Transplanting it back into the original machine, the k6, and the problems return. Maybe this could be a hardware problem..... but this machine has been stable in the past. Tomorrow I will try the gcc3 profile.
The gcc3 profile didnt get me anywhere - I guess my fault but I didnt debug further 2.93.3-r7 with "-O2 -march=i586" is stable so far on this problematic k6
Moving these so we can remove the "Install CD" component from "Gentoo Linux". I apologize to everyone for this spam, but according to the bugzilla developers, this is the only reasonable way to do this.