[Bug c/11734] New: Optimization errors using -m64 above -O1

["gerryt at vcn dot bc dot ca" <gcc-bugzilla at gcc dot gnu dot org>]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11734

           Summary: Optimization errors using -m64 above -O1
           Product: gcc
           Version: 3.3
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: gerryt at vcn dot bc dot ca
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: Configured with: ../configure --prefix=/opt/gcc --
                    enable-multili
  GCC host triplet: SunOS brad 5.9 Generic_112233-06 sun4u sparc SUNW,Ultra-
                    2
GCC target triplet: sparc-sun-solaris2.9

Waited a while and no one seems to have noticed.
Sorry I dont understand what triplet means above : >

Try running make test after compiling openssl-0.9.7b
Use: ./Configure solaris64-sparcv9-gcc

The complete command line that triggers the bug:

cd test
gcc -I.. -I../include  -DOPENSSL_SYSNAME_ULTRASPARC -DOPENSSL_THREADS -D_REENTRA
NT -DDSO_DLFCN -DHAVE_DLFCN_H -DOPENSSL_NO_KRB5 -m64 -mcpu=ultrasparc -O3 -Wall
-DB_ENDIAN -DMD5_ASM  -c  rmdtest.c

gcc -o rmdtest -I.. -I../include  -DOPENSSL_SYSNAME_ULTRASPARC -DOPENSSL_THREADS
 -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H -DOPENSSL_NO_KRB5 -m64 -mcpu=ultrasparc
 -O3 -Wall -DB_ENDIAN -DMD5_ASM rmdtest.o  -L.. -lcrypto -lsocket -lnsl -ldl

./rmdtest
error calculating RIPEMD160 on ''
got c12836ad0d061da6ccde02fb0b5be87f0c62a4a5 instead of 9c1185a5c5e9fc5461280897
7ee8f548b2258d31

I have saved all .i and .s files. The .i files are identical
but there are MANY diffs with the .s files between -O1 and -O3
Too many to drop in here.

To summarize -O passes ALL tests -O1 good as well -O2 bad -O3 bad

gerryt@vcn.bc.ca