The actual snapshot gcc-4.1-20050501 fails to bootstrap if -O3 -ftree-vectorize is given as CFLAGS. Michael Cieslinski
I get the following bootstrap failure on powerpc-apple-darwin, with "-O2 -ftree- vectorize": ../../gcc/gcc/builtins.c: In function 'fold_builtin_1': ../../gcc/gcc/builtins.c:8451: internal compiler error: Bus error which seems to happen during tree-if-conversion: (gdb) Program received signal EXC_BAD_ACCESS, Could not access memory. 0x000cbe00 in clean_predicate_lists (loop=0x4144fda0) at ../../gcc/gcc/tree-if- conv.c:654 654 bb[i]->aux = NULL; (gdb) backtrace #0 0x000cbe00 in clean_predicate_lists (loop=0x4144fda0) at ../../gcc/gcc/tree- if-conv.c:654 #1 0x000caf0c in tree_if_conversion (loop=0x4d78b8, for_vectorizer=16 '\020') at ../../gcc/gcc/tree-if-conv.c:205 #2 0x000ccbf0 in main_tree_if_conversion () at ../../gcc/gcc/tree-if- conv.c:1116 #3 0x00065874 in execute_one_pass (pass=0x17311d0) at ../../gcc/gcc/tree- optimize.c:578 #4 0x00065958 in execute_pass_list (pass=0x4b7f2c) at ../../gcc/gcc/tree- optimize.c:615 #5 0x00065970 in execute_pass_list (pass=0x4b7cb0) at ../../gcc/gcc/tree- optimize.c:616 #6 0x00065970 in execute_pass_list (pass=0x4b74f0) at ../../gcc/gcc/tree- optimize.c:616 #7 0x00065c34 in tree_rest_of_compilation (fndecl=0xb7915c) at ../../gcc/gcc/tree-optimize.c:728 #8 0x0000e60c in c_expand_body (fndecl=0xde0bc8) at ../../gcc/gcc/c-decl.c:6411 #9 0x002f7998 in cgraph_expand_function (node=0xde0bc8) at ../../gcc/gcc/cgraphunit.c:924 #10 0x002f7ac4 in cgraph_expand_all_functions () at ../../gcc/gcc/cgraphunit.c:988 #11 0x002f7e0c in cgraph_optimize () at ../../gcc/gcc/cgraphunit.c:1090 #12 0x002a6ad4 in compile_file () at ../../gcc/gcc/toplev.c:1013 #13 0x002a8a80 in do_compile () at ../../gcc/gcc/toplev.c:2123 #14 0x002a8b04 in toplev_main (argc=43, argv=0xbffff868) at ../../gcc/gcc/toplev.c:2155 #15 0x00001df0 in _start (argc=43, argv=0xbffff868, envp=0xbffff918) at /SourceCache/Csu/Csu-47/crt.c:267 #16 0x8fe1a558 in __dyld__dyld_start () (gdb)
I am just going to mark this a meta-bug and have PR 24059 block this one.
I know this is a bootstrap failure, however is there a small reproducible test case for this crash ? Thanks.
(In reply to comment #3) > I know this is a bootstrap failure, however is there a small reproducible test > case for this crash ? Thanks. THe problem in comment #1 has already been fixed for a while now. See the bugs that this bug depends on.
I'd already verified that. How about "-O3 -ftree-vectorize" bootstrap failure ?
I can bootstrap current trunk (r128479) with -ftree-vectorize on x86_64-unknown-linux-gnu for some time now, and, according to http://gcc.gnu.org/ml/gcc-patches/2007-09/msg00327.html, this problem is gone on powerpc64 too. So closing as fixed.
(In reply to comment #6) > I can bootstrap current trunk (r128479) with -ftree-vectorize on > x86_64-unknown-linux-gnu for some time now, and, according to > http://gcc.gnu.org/ml/gcc-patches/2007-09/msg00327.html, this problem is gone > on powerpc64 too. actually the link you give above explicitly states that "we don't pass bootstrap with vectorization enabled on powerpc-linux", but there's already a PR for that (PR32582) so it's ok to close this one > So closing as fixed.