Bug 22078 - [3.4 Regression] Cannot profilebootstrap 3.4.4 on OS X 10.3, 3.4.3 works
Summary: [3.4 Regression] Cannot profilebootstrap 3.4.4 on OS X 10.3, 3.4.3 works
Status: RESOLVED WONTFIX
Alias: None
Product: gcc
Classification: Unclassified
Component: target (show other bugs)
Version: 3.4.4
: P2 normal
Target Milestone: 3.4.6
Assignee: Not yet assigned to anyone
URL:
Keywords: build, ice-on-valid-code
Depends on:
Blocks:
 
Reported: 2005-06-15 15:05 UTC by Michal Suchanek
Modified: 2006-02-28 15:48 UTC (History)
1 user (show)

See Also:
Host: powerpc-apple-darwin
Target: powerpc-apple-darwin
Build: powerpc-apple-darwin
Known to work:
Known to fail:
Last reconfirmed:


Attachments
saved .i file (103.37 KB, text/plain)
2005-06-16 11:14 UTC, Michal Suchanek
Details
.gcno file in case it is useful (47.75 KB, application/octet-stream)
2005-06-16 11:15 UTC, Michal Suchanek
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Michal Suchanek 2005-06-15 15:05:07 UTC
So far I have tried bootstrapping 3.4.4 three times, and it always fails in
something named rs6000. I also tried 3.4.3 several times, and it works (compiles
gcj).

Re-running make in the failed directory does not reproduce the error, at least
not quickly.

Configure parameters from config.status:
# ../gcc-3.4.4/configure  --prefix=/sw --enable-languages=c++,java
'--infodir=${prefix}/share/info' '--libexecdir=${prefix}/lib' --disable-shared
--progra
m-suffix=-3.4
--with-gcc-version-trigger=/Volumes/Fink/sw/src/gcj-3.4.4-0/gcc-3.4.4/gcc/version.c

From Makefile in gcc:
XCFLAGS =
TCFLAGS =
CFLAGS = -g
STAGE1_CFLAGS = -g -no-cpp-precomp -DHAVE_DESIGNATED_INITIALIZERS=0
BOOT_CFLAGS = -g -O2


Error message:
stage1/xgcc -Bstage1/ -B/sw/powerpc-apple-darwin7.9.0/bin/ -c   -g -O2
-fprofile-generate -DIN_GCC   -W -Wall -Wwrite-strings -Wstrict-prototypes
-Wmissing-prototypes -pedantic -Wno-long-long -Wold-style-definition    
-DHAVE_CONFIG_H -DGENERATOR_FILE -I/sw/include   -I. -I. -I../../gcc-3.4.4/gcc
-I../../gcc-3.4.4/gcc/. -I../../gcc-3.4.4/gcc/../include 
../../gcc-3.4.4/gcc/genattr.c -o genattr.o
stage1/xgcc -Bstage1/ -B/sw/powerpc-apple-darwin7.9.0/bin/   -g -O2
-fprofile-generate -DIN_GCC   -W -Wall -Wwrite-strings -Wstrict-prototypes
-Wmissing-prototypes -pedantic -Wno-long-long -Wold-style-definition    
-DHAVE_CONFIG_H -DGENERATOR_FILE  -o genattr \
        genattr.o rtl.o read-rtl.o bitmap.o ggc-none.o min-insn-modes.o
gensupport.o insn-conditions.o print-rtl1.o \
        errors.o ../libiberty/libiberty.a
./genattr ../../gcc-3.4.4/gcc/config/rs6000/rs6000.md > tmp-attr.h
/bin/sh ../../gcc-3.4.4/gcc/move-if-change tmp-attr.h insn-attr.h
echo timestamp > s-attr
stage1/xgcc -Bstage1/ -B/sw/powerpc-apple-darwin7.9.0/bin/ -c   -g -O2
-fprofile-generate -DIN_GCC   -W -Wall -Wwrite-strings -Wstrict-prototypes
-Wmissing-prototypes -pedantic -Wno-long-long -Wold-style-definition    
-DHAVE_CONFIG_H -I. -I. -I../../gcc-3.4.4/gcc -I../../gcc-3.4.4/gcc/.
-I../../gcc-3.4.4/gcc/../include  -I/sw/include   ../../gcc-3.4.4/gcc/combine.c
-o combine.o
../../gcc-3.4.4/gcc/combine.c: In function `subst':
../../gcc-3.4.4/gcc/combine.c:3540: internal compiler error: in
rs6000_split_multireg_move, at config/rs6000/rs6000.c:10613
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
Comment 1 Andrew Pinski 2005-06-15 15:22:17 UTC
So rerunning "make bootstrap", makes this work, then there is something wrong with your hardware.
Comment 2 Michal Suchanek 2005-06-15 16:56:48 UTC
hmm, I forgot bootstrap. That's why it fails differently when I continue manually.

The build is normally controlled by a package management system, I will try with
full make command line next time.
Comment 3 Michal Suchanek 2005-06-16 11:12:45 UTC
The command is

make CFLAGS='-O -fno-common' LIBCFLAGS='-g -O2 -fno-common' LIBCXXFLAGS='-g -O2
-fno-implicit-templates' profiledbootstrap

and it reproduces the error


Comment 4 Michal Suchanek 2005-06-16 11:14:15 UTC
Created attachment 9098 [details]
saved .i file
Comment 5 Michal Suchanek 2005-06-16 11:15:58 UTC
Created attachment 9099 [details]
.gcno file in case it is useful
Comment 6 Michal Suchanek 2005-06-16 11:19:42 UTC
continuing with bootstrap instead of profiledbootstrap does not fail
Comment 7 Andrew Pinski 2005-06-16 12:08:31 UTC
Hmm, weird, might have a look later next week.
Comment 8 Gabriel Dos Reis 2006-02-28 15:48:07 UTC
Unlikely to be fixed for 3.4.6