This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
arm failure in g++.other/eh2.C
- To: gcc-bugs at gcc dot gnu dot org
- Subject: arm failure in g++.other/eh2.C
- From: Philip Blundell <philb at gnu dot org>
- Date: Mon, 28 May 2001 23:18:07 +0100
This seems to be a regression from 2.95. I can't reproduce it with a
cross-compiler, but it happens both for me and for the netwinder.org
autobuilder with a native bootstrap.
Here's the crash:
Program received signal SIGSEGV, Segmentation fault.
0x14c14c in sjlj_mark_call_sites (lp_info=0x4a6af8)
at /home/pb/egcs/3.0/gcc/gcc/except.c:2113
2113 if (GET_CODE (before) == CODE_LABEL)
(gdb) bt
#0 0x14c14c in sjlj_mark_call_sites (lp_info=0x4a6af8)
at /home/pb/egcs/3.0/gcc/gcc/except.c:2113
#1 0x14c95c in sjlj_build_landing_pads ()
at /home/pb/egcs/3.0/gcc/gcc/except.c:2323
#2 0x14ca00 in finish_eh_generation ()
at /home/pb/egcs/3.0/gcc/gcc/except.c:2364
#3 0xfc5f0 in rest_of_compilation (decl=0x4018dc98)
at /home/pb/egcs/3.0/gcc/gcc/toplev.c:2901
#4 0xce1fc in genrtl_finish_function (fn=0x4018dc98)
at /home/pb/egcs/3.0/gcc/gcc/cp/semantics.c:2662
#5 0xcdc68 in expand_body (fn=0x4018dc98)
at /home/pb/egcs/3.0/gcc/gcc/cp/semantics.c:2448
#6 0x8b520 in yyparse_1 () at parse.y:759
#7 0xf4a40 in yyparse () at /home/pb/egcs/3.0/gcc/gcc/c-lex.c:158
#8 0xfb700 in compile_file (
name=0x49acc4 "/home/pb/egcs/3.0/gcc/gcc/testsuite/g++.old-deja/g++.other/eh2.C") at /home/pb/egcs/3.0/gcc/gcc/toplev.c:2332
#9 0x1008b8 in main (argc=3, argv=0xbffffa1c)
at /home/pb/egcs/3.0/gcc/gcc/toplev.c:4944
#10 0x400401e0 in __libc_start_main () from /lib/libc.so.6
The problem seems to be that it walks off the front of the chain of insns:
"before" gets set to NULL when nparm_regs is still 1.
I'm not sure if this only occurs with a compiler that's been 3-staged, or if I
can make it happen when cc1plus was built with a known-working compiler. I'll
try to establish this.
p.
PGP signature