I've gotten a bunch of these errors with 3.3 that don't happen with gcc 3.2.3 20030415 (Debian prerelease). Reading specs from /usr/lib/gcc-lib/mipsel-linux/3.3/specs Configured with: ../src/configure -v --enable-languages=c,c++,f77,pascal,objc,treelang --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-gxx-include-dir=/usr/include/c++/3.3 --enable-shared --with-system-zlib --enable-nls --without-included-gettext --enable-__cxa_atexit --enable-clocale=gnu --enable-debug --enable-objc-gc mips-linux Thread model: posix gcc version 3.3 (Debian) gcc -save-temps -c -g -O2 -fno-strict-aliasing -Wall -Wstrict-prototypes -I. -I../Include -DHAVE_CONFIG_H -o Python/errors.o ../Python/errors.c ../Python/errors.c: In function `PyErr_Format': ../Python/errors.c:514: error: Attempt to delete prologue/epilogue insn: (insn 854 656 855 0 (nil) (set (mem:SI (plus:SI (reg/f:SI 29 sp) (const_int 84 [0x54])) [0 S4 A32]) (reg:SI 5 a1)) -1 (nil) (nil)) ../Python/errors.c:514: internal compiler error: in propagate_one_insn, at flow.c:1639
Created attachment 4158 [details] preprocessed source
Created attachment 4171 [details] reduced testcase.i
Confirmed with gcc 3.3 branch and mainline (20030604). Took me a while, because this is not reproducible with mipsel-elf target. Hopefully the reduced testcase I've attached will help track this one down... Dara
With permission. Testing a fix now.
Used the wrong bugzilla option. Let's try that again...
Oh for the love of a duck. ;) Why does changing the assignee make the status go from ASSIGNED to NEW?
Subject: Bug 11084 CVSROOT: /cvs/gcc Module name: gcc Branch: gcc-3_3-branch Changes by: rsandifo@gcc.gnu.org 2003-06-25 06:50:57 Modified files: gcc : ChangeLog gcc/config/mips: mips.c gcc/testsuite : ChangeLog Added files: gcc/testsuite/gcc.c-torture/compile: 20030625-1.c Log message: PR target/11084 * config/mips/mips.c (mips_expand_prologue): Fix setting of regno for the end of a variable argument list. Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=1.16114.2.614&r2=1.16114.2.615 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/config/mips/mips.c.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=1.241.2.2&r2=1.241.2.3 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=1.2261.2.201&r2=1.2261.2.202 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.c-torture/compile/20030625-1.c.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=NONE&r2=1.1.2.1
The commit message in the audit trail is for 3.3. See: http://gcc.gnu.org/ml/gcc-patches/2003-06/msg02766.html for a full description. I submitted a different version for 3.4: http://gcc.gnu.org/ml/gcc-patches/2003-06/msg02768.html However, an earlier patch means that the bug no longer causes an ICE in 3.4 sources. I'll go ahead and close this as fixed in both branch and trunk.