This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: New mainline bootstrap failure
"Vladimir N. Makarov" <vmakarov@redhat.com> writes:
|> Andreas Schwab wrote:
|>
|> > I'm currently getting the following error during bootstrap for ia64-linux:
|> >
|> > ./xgcc -B./ -B/usr/local/ia64-suse-linux/bin/ -isystem /usr/local/ia64-suse-linux/include -isystem /usr/local/ia64-suse-linux/sys-include -O2 -DIN_GCC -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -isystem ./include -fPIC -g -DHAVE_GTHR_DEFAULT -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -I. -I. -I../../gcc -I../../gcc/. -I../../gcc/config -I../../gcc/../include -DL_ffsdi2 -c ../../gcc/libgcc2.c -o libgcc/./_ffsdi2.o
|> > ../../gcc/libgcc2.c: In function `__ffsti2':
|> > ../../gcc/libgcc2.c:350: internal compiler error: RTL flag check: SCHED_GROUP_P used with unexpected rtx code `code_label' in init_ready_list, at sched-ebb.c:97
|> >
|> > Andreas.
|>
|> Here, the patch is. Its is obvious, therefore I've commited it into main line.
Thanks, this works. Now I'm faced with some kind of memory corruption
during the compilation of tradcpp.c with the stage1 compiler. It only
happens when called during bootstrap, not when I run the compiler
manually, which indicates that it depends on the exact memory layout. The
symptoms are that the compiler crashes during GC, due to an invalid tree
node.
Btw., at the time of the crash there were over 1900 call frames active,
99% being calls to gt_ggc_mx_lang_tree_node. Most of them seem to come
from recursion through (*x).generic.*.common.chain.
Andreas.
--
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux AG, Deutschherrnstr. 15-19, D-90429 Nürnberg
Key fingerprint = 58CA 54C7 6D53 942B 1756 01D3 44D5 214B 8276 4ED5
"And now for something completely different."