Consistant buildworld Segmentation fault at '===> gnu/usr.bin/cc/cpp' after latest RELENG_5 cvsup

[Aaron Daubman]

Hi all,

I'm an oldtime FreeBSD user but am new to the amd64 world.  I'm trying
to put one of my Sunfire V20z single proc Opteron w/2G RAM servers to
better use (previously ran Win2k3) and am having some difficulty
updating it.

I started off using the latest amd64 snapshot:
ftp://ftp.freebsd.org/pub/FreeBSD/snapshots/Feb_2005/5.3-STABLE-SNAP001-amd64-miniinst.iso
and that has been up and running (albeit not doing much) for several days.

After cvsup'ing to RELENG_5 and tweaking my kernel config (note -
since buildworld never completed I'm still running with the snapshot
GENERIC kernel) I prepared to experience my fastest buildworld yet. 
Well, it was fast alright...

About 15 minutes into the build process buildworld died - I couldn't
tell exactly where the first time since I used -j 4, so I re-ran
buildworld with no -j option and saw that make was coreing here:

---snip---
makeinfo -I /usr/src/gnu/usr.bin/cc/doc/../../../../contrib/gcc/doc -I
/usr/src/gnu/usr.bin/cc/doc/../../../../contrib/gcc/doc/include
--no-split -I /usr/src/gnu/usr.bin/cc/doc -I
/usr/src/gnu/usr.bin/cc/doc
/usr/src/gnu/usr.bin/cc/doc/../../../../contrib/gcc/doc/cppinternals.texi
 -o cppinternals.info
gzip -cn cppinternals.info > cppinternals.info.gz
===> gnu/usr.bin/cc/cpp
Segmentation fault (core dumped)
*** Error code 139

Stop in /usr/src/gnu/usr.bin/cc.
*** Error code 1

Stop in /usr/src/gnu/usr.bin.
*** Error code 1

Stop in /usr/src/gnu.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.
---snip---

after a few more buildworld attempts I turned to IRC where somebody
suggested I may have just hit a bad CVSUP mirror.  So, I nuked
/usr/src, grabbed a new one and re-cvsupd from a different mirror. 
Same segfault, same place.

After some searching I came across a somewhat similar post to my problem:
http://lists.freebsd.org/mailman/htdig/freebsd-amd64/2004-December/003040.html

Even though it seems the bug would have been fixed in the latest
stable RELENG_5 that I'd just cvsup'd to, I tried the suggestion
anyway.  It made no difference - same segfault in make, same place.

Another post suggested adjusting the HT settings in BIOS (although did
not say to what) - I have not tried that yet - is this a valid
suggestion?  Why would the defaults (that Sun set up for Linux
install) not work?

Any other troubleshooting suggestions?

My /etc/make.conf is as follows:
---snip---
PERL_VER=5.8.5
PERL_VERSION=5.8.5
PERL_ARCH=mach
NOPERL=yo
NO_PERL=yo
NO_PERL_WRAPPER=yo

CFLAGS= -O -pipe
COPTFLAGS= -O -pipe
TOP_TABLE_SIZE= 101
BOOTWAIT=       4

NOGAMES=        true    # do not build games (games/ subdir)
NO_FORTRAN=     true
NO_LPR=         true
NO_X=           true
NOPROFILE=      true

MAKE_IDEA=      YES     # IDEA (128 bit symmetric encryption)
USA_RESIDENT=   YES

SUP_UPDATE=     YES
SUP=            /usr/local/bin/cvsup
SUPFLAGS=       -g -L 2
SUPHOST?=       cvsup3.freebsd.org
SUPFILE=        /usr/share/examples/cvsup/standard-supfile
PORTSSUPFILE=   /usr/share/examples/cvsup/ports-supfile
DOCSUPFILE=     /usr/share/examples/cvsup/doc-supfile

APACHE2=        true
WITH_APACHE2=   true
WITH_MODPERL2=  true

WITHOUT_EXTRA_ENCODINGS=yes 

WITH_OPENSSL_PORT=      YES
---snip---

Thanks,
     ~Aaron