https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100767
--- Comment #6 from Gabriel Marcano ---
I was the one that originally reported the issue on gcc-help, and I can confirm
that with the patch the segfault is no longer happening. I do appear to have
found a different bug (undefined references to l
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119953
--- Comment #1 from Gabriel Marcano ---
I've also run into this bug, and I've done some debugging. The problem, as the
log states, appears to be a constraint failure. Specifically, it looks like GCC
is trying to use R13 and R14 for the addsi3 in
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119953
--- Comment #3 from Gabriel Marcano ---
Created attachment 61569
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=61569&action=edit
msp430_enforce_register_pair_alignment.patch
This is a patch I made that forces GCC to only use even registe
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119953
--- Comment #4 from Gabriel Marcano ---
Created attachment 61570
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=61570&action=edit
reduced_cow-fs_path.ii
I managed to use cvise to reduce the necessary code to reproduce the problem,
which I
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119953
--- Comment #5 from Gabriel Marcano ---
Oh, this may be relevant. This is the output of -v for the generated xgcc that
reproduces the problem:
# gcc/xgcc -v -B. -B gcc/
Reading specs from gcc/specs
COLLECT_GCC=gcc/xgcc
COLLECT_LTO_WRAPPER=gcc/l
