FreeBSD 5.2-RELEASE buildworld failure.

erek erek at necrotec.com
Fri Jan 23 18:14:10 PST 2004


I cvsuped today using tag RELENG_5_2 (i'm already using 5.2-RELEASE), i
noticed there was only documentation changes, and no real source changes,
i was preparing to recompile everything using some different
optimizations:
"
erek# cat /etc/make.conf
CFLAGS+= -pipe -march=athlon-xp -mcpu=athlon-xp -mmmx -m3dnow -ffast-math
-msse -mfpmath=sse,387
CXXFLAGS+= -pipe -march=athlon-xp -mcpu=athlon-xp -mmmx -m3dnow
-ffast-math   -msse -mfpmath=sse,387
COPTFLAGS+= -pipe -march=athlon-xp -mcpu=athlon-xp -mmmx -m3dnow
-ffast-math -msse -mfpmath=sse,387
"
Previously i had unroll-loops in my optimizations and i heard this could
actually be slower on modern CPUs, so i removed it.

During the buildworld I get this VERY odd error:

"mkdep -f .depend -a    -DIN_GCC -DHAVE_CONFIG_H
-DPREFIX=\"/usr/obj/usr/src/i386
/usr\" -I/usr/obj/usr/src/i386/usr/src/gnu/usr.bin/cc/cc1plus/../cc_tools
-I/usr
/src/gnu/usr.bin/cc/cc1plus/../cc_tools
-I/usr/src/gnu/usr.bin/cc/cc1plus/../../
../../contrib/gcc
-I/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/conf
ig -I/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp -I.
-I/usr/obj/u
sr/src/i386/legacy/usr/include  parse+%DIKED.c
/usr/src/gnu/usr.bin/cc/cc1plus/.
./../../../contrib/gcc/main.c
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contri
b/gcc/cp/cp-lang.c
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/c-opt
s.c /usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/call.c
/usr/src/g
nu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/class.c
/usr/src/gnu/usr.bin/cc
/cc1plus/../../../../contrib/gcc/cp/cvt.c
/usr/src/gnu/usr.bin/cc/cc1plus/../../
../../contrib/gcc/cp/decl.c
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/
gcc/cp/decl2.c
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/error.
c /usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/except.c
/usr/src/g
nu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/expr.c
/usr/src/gnu/usr.bin/cc/
cc1plus/../../../../contrib/gcc/cp/friend.c
/usr/src/gnu/usr.bin/cc/cc1plus/../.
./../../contrib/gcc/cp/init.c
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contri
b/gcc/cp/lex.c
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/mangle
.c /usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/method.c
/usr/src/
gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/pt.c
/usr/src/gnu/usr.bin/cc/c
c1plus/../../../../contrib/gcc/cp/ptree.c
/usr/src/gnu/usr.bin/cc/cc1plus/../../
../../contrib/gcc/cp/repo.c
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/
gcc/cp/rtti.c
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/search.
c /usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/semantics.c
/usr/sr
c/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/spew.c
/usr/src/gnu/usr.bin/
cc/cc1plus/../../../../contrib/gcc/cp/tree.c
/usr/src/gnu/usr.bin/cc/cc1plus/../
../../../contrib/gcc/cp/typeck.c
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../con
trib/gcc/cp/typeck2.c
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp
/dump.c /usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/optimize.c
In file included from
/usr/obj/usr/src/i386/usr/src/gnu/usr.bin/cc/cc_tools/tcon
fig.h:17,
                 from
/usr/obj/usr/src/i386/usr/src/gnu/usr.bin/cc/cc_tools/hcon
fig.h:2,
                 from
/usr/obj/usr/src/i386/usr/src/gnu/usr.bin/cc/cc_tools/conf
ig.h:1,
                 from
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp
/parse.y:31,
                 from /usr/src/contrib/gcc/cp/lex.c:33:
/usr/src/gnu/usr.bin/cc/cc_tools/freebsd-native.h:62:25: attempt to use
poisoned
 "malloc"
/usr/src/gnu/usr.bin/cc/cc_tools/freebsd-native.h:63:25: attempt to use
poisoned
 "calloc"
/usr/src/gnu/usr.bin/cc/cc_tools/freebsd-native.h:64:25: attempt to use
poisoned
 "realloc"
/usr/src/gnu/usr.bin/cc/cc_tools/freebsd-native.h:65:25: attempt to use
poisoned
 "strdup"
In file included from
/usr/obj/usr/src/i386/usr/src/gnu/usr.bin/cc/cc_tools/tcon
fig.h:17,
                 from
/usr/obj/usr/src/i386/usr/src/gnu/usr.bin/cc/cc_tools/hcon
fig.h:2,
                 from
/usr/obj/usr/src/i386/usr/src/gnu/usr.bin/cc/cc_tools/conf
ig.h:1,
                 from
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp
/parse.y:31,
                 from /usr/src/contrib/gcc/cp/pt.c:37:
/usr/src/gnu/usr.bin/cc/cc_tools/freebsd-native.h:62:25: attempt to use
poisoned
 "malloc"
/usr/src/gnu/usr.bin/cc/cc_tools/freebsd-native.h:63:25: attempt to use
poisoned
 "calloc"
/usr/src/gnu/usr.bin/cc/cc_tools/freebsd-native.h:64:25: attempt to use
poisoned
 "realloc"
/usr/src/gnu/usr.bin/cc/cc_tools/freebsd-native.h:65:25: attempt to use
poisoned
 "strdup"
In file included from
/usr/obj/usr/src/i386/usr/src/gnu/usr.bin/cc/cc_tools/tcon
fig.h:17,
                 from
/usr/obj/usr/src/i386/usr/src/gnu/usr.bin/cc/cc_tools/hcon
fig.h:2,
                 from
/usr/obj/usr/src/i386/usr/src/gnu/usr.bin/cc/cc_tools/conf
ig.h:1,
                 from
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp
/parse.y:31,
                 from /usr/src/contrib/gcc/cp/spew.c:34:
/usr/src/gnu/usr.bin/cc/cc_tools/freebsd-native.h:62:25: attempt to use
poisoned
 "malloc"
/usr/src/gnu/usr.bin/cc/cc_tools/freebsd-native.h:63:25: attempt to use
poisoned
 "calloc"
/usr/src/gnu/usr.bin/cc/cc_tools/freebsd-native.h:64:25: attempt to use
poisoned
 "realloc"
/usr/src/gnu/usr.bin/cc/cc_tools/freebsd-native.h:65:25: attempt to use
poisoned
 "strdup"
mkdep: compile failed
*** Error code 1
"

I simply proceeded to try the build again, it did the same thing at the
exact same spot, so I then underclocked my cpu and ram by 300 or mhz,
1912Mhz (Stock) to 1533mhz and set my memory timings from "Optimal" to "By
SPD" (manufacturer defaults).  With the cpu being underclocked and the
vcore being 1.65V (for stock) the cpu was being overvolted for it's clock
so if there was any stability problems they "should" be gone (even though
i dind't experience stability problems or crash/locking up anyhow).   This
move resulted in the same problem. so i removed the optimizations in my
/etc/make.conf, attempted to do it again, same results as above. I was
further instructed to rm -rf /usr/obj, did that, same results.  I verified
the integrity of my disks with a low level utility inside my controller,
and it revealed no bad blocks on any disks.  (mprime95 running torture
test for a long time at stock speeds - far more stressful than buildworld,
and no a single error)...

any suggestions?



More information about the freebsd-questions mailing list