gcc 2.95.2 dies with signal 11, internal compiler error

[mulix]

hello,

i hope i'm writing to the correct gcc mailing list - if not, please
advise.

i'm compiling some kernel code [1], with gcc 3.0, gcc 2.95.2 and egcs
2.91.66. 3.0 and 2.91.6 are fine, 2.95.2 dies with a signal 11,
internal compiler error.

i've read the signal 11 faq, and i still think it's a compiler error.
i'd like to know what makes it barf, so that i can avoid it (the code is
legal, as evident by the fact that it compiles fine with the other
compilers).

gcc -v is:
Reading specs from /usr/local/lib/gcc-lib/i686-pc-linux-gnu/2.95.2/specs
gcc version 2.95.2 19991024 (release)

the command line and the errors are:
gcc -D__KERNEL__ -DMODULE -I/archive/kernel/uml/linux/include -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -g -fomit-frame-pointer
-fno-strict-aliasing -fno-common -pipe -I../st_rules   -c module_main.c
-o module_main.o
gcc: Internal compiler error: program cc1 got fatal signal 11
make: *** [module_main.o] Error 1
[mulix@alhambra st_rules_module]$ {standard input}: Assembler messages:
{standard input}:1010: Warning: end of file not at end of a line;
newline inserted
{standard input}:1876: Error: unbalanced parenthesis in operand 1.
cpp: output pipe has been closed

unfortunately, the file that causes this error is autogenerated and
quite big, if anyone wants to compile it, you can get it from cvs [2] or
i'll be happy to send you a tarball.

any ideas on how to debug this further?

[1] shameless plug: check out the syscalltracker project, at
http://www.sf.net/projects/syscalltrack/

[2] goto [1]
-- 
mulix
http://www.advogato.com/person/mulix

linux/reboot.h: #define LINUX_REBOOT_MAGIC1 0xfee1dead