Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!

Bug 54415

Summary: Stage 1 bootstrapping on athlon-xp had a internal compiler error while building coffgen.c
Product: Gentoo Linux Reporter: Jeff Cooper <jacooper>
Component: [OLD] Core systemAssignee: Gentoo Release Team <releng>
Status: RESOLVED INVALID    
Severity: blocker CC: toolchain
Priority: High    
Version: unspecified   
Hardware: x86   
OS: Linux   
Whiteboard:
Package list:
Runtime testing required: ---
Attachments: Preprocessed source for coffgen.c

Description Jeff Cooper 2004-06-19 06:30:06 UTC
During a stage1 bootstrapping on a Athlon-XP, I got the following error from the 'scripts/bootstrap.sh' step:

coffgen.c: In function `coff_write_symbol':
coffgen.c:954: internal compiler error: Didn't find a coloring.

The last entry in emerge.log is:

1087605823:  >>> emerge (6 of 6) sys-devel/binutils-2.14.90.0.8-r1 to /
1087605823:  === (6 of 6) Cleaning (sys-devel/binutils-2.14.90.0.8-r1::/usr/portage/sys-devel/binutils/bin
utils-2.14.90.0.8-r1.ebuild)
1087605824:  === (6 of 6) Compiling/Merging (sys-devel/binutils-2.14.90.0.8-r1::/usr/portage/sys-devel/bin
utils/binutils-2.14.90.0.8-r1.ebuild)
1087606110:  *** terminating.

My CFLAGS are listed below.

Reproducible: Always
Steps to Reproduce:
1.Start doing a stage1 install (2004.1), following the instructions at: http://www.gentoo.org/doc/en/handbook/handbook-x86.xml?part=1&chap=6
2.Get to the step of Bootstrapping the system and run '
scripts/bootstrap.sh'

Actual Results:  
bootstrap failed, internal compiler error.

Expected Results:  
It should have compiled without failing! :-)

Portage 2.0.50-r8 (default-x86-2004.0, gcc-3.3.2, glibc-2.3.2-r9, 2.6.5-gentoo)
=================================================================
System uname: 2.6.5-gentoo i686 AMD Athlon(tm)
Gentoo Base System version 1.4.3.13
Autoconf:
Automake:
ACCEPT_KEYWORDS="x86"
AUTOCLEAN="yes"
CFLAGS="-O3 -fforce-addr -fmerge-all-constants -ffast-math -fnew-ra -ftracer
-funroll-all-loops -fprefetch-loop-arrays -fstrict-aliasing -falign-functions=64
-falign-labels=1 -falign-loops=16 -falign-jumps=16 -march=athlon-xp -mfpmath=sse
-mpreferred-stack-boundary=4 -mno-push-args -maccumulate-outgoing-args
-minline-all-stringops -pipe"
CHOST="i386-pc-linux-gnu"
COMPILER="gcc3"
CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3/share/config
/usr/share/config /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d"
CXXFLAGS="-O3 -fforce-addr -fmerge-all-constants -ffast-math -fnew-ra -ftracer
-funroll-all-loops -fprefetch-loop-arrays -fstrict-aliasing -falign-functions=64
-falign-labels=1 -falign-loops=16 -falign-jumps=16 -march=athlon-xp -mfpmath=sse
-mpreferred-stack-boundary=4 -mno-push-args -maccumulate-outgoing-args
-minline-all-stringops -pipe"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoaddcvs ccache sandbox"
GENTOO_MIRRORS="http://gentoo.oregonstate.edu
http://distro.ibiblio.org/pub/Linux/distributions/gentoo"
MAKEOPTS="-j2"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
PORTDIR_OVERLAY=""
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
USE="X acpi alsa arts avi berkdb cdr crypt cups dga directfb dvd encode
foomaticdb gdbm gif gnome gphoto2 gpm gtk gtk2 imagemagick imlib java jpeg kde
libg++ libwww mad mbox mikmod motif mozilla mpeg ncurses nls oggvorbis opengl
oss pam pcmcia pdflib perl png ppds python qt quicktime readline samba sdl slang
spell ssl svga tcpd tiff truetype trusted x86 xml2 xmms xv zlib"
Comment 1 Jeff Cooper 2004-06-19 06:33:00 UTC
Created attachment 33561 [details]
Preprocessed source for coffgen.c
Comment 2 Stephen Bennett (RETIRED) gentoo-dev 2004-06-19 08:35:33 UTC
-fnew-ra is known to be buggy. Take it out and you should be fine.
Comment 3 Benjamin Judas (RETIRED) gentoo-dev 2004-07-04 23:50:29 UTC
Is this bug still an issue or is the problem solved?
Comment 4 Travis Tilley (RETIRED) gentoo-dev 2004-07-14 09:34:34 UTC
fnew-ra
    Use a graph coloring register allocator. Currently this option is meant only for testing. Users should not specify this option, since it is not yet ready for production use. 

that should say it all. plus you should never use -ffast-math as a global cflag, it's almost as broken when used that way.