http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57896
--- Comment #18 from Uroš Bizjak ---
And finally, it looks that expand_vector_operations doesn't call init_emit ().
Backtrace:
Breakpoint 1, gen_reg_rtx (mode=V4SImode) at
../../gcc-svn/branches/gcc-4_8-branch/gcc/emit-rtl.c:864
864 unsign
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57896
--- Comment #17 from Uroš Bizjak ---
Somehow we get old_size = 0 in gen_reg_rtx. It is calculated as:
int old_size = crtl->emit.regno_pointer_align_length;
Trivial patch that prevents this situation and fixes the problem. However, it
is no
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57896
--- Comment #16 from Uroš Bizjak ---
(In reply to Jakub Jelinek from comment #15)
> I'll have a look (unless you want to continue poking).
According to ChangeLogs, you have much more experience in this part, so I would
be much grateful if you can
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57896
--- Comment #15 from Jakub Jelinek ---
I'll have a look (unless you want to continue poking).
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57896
--- Comment #14 from Uroš Bizjak ---
I have found where corruption of const_int_rtx array happens.
1. Put breakpoint on gen_rtx_CONST_INT to determine address of
const_int_rtx[24], a.k.a (const_int -40) and put HW watchpoint on its address:
Brea
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57896
--- Comment #13 from Marek Polacek ---
Seems like r196890 made this latent.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57896
Uroš Bizjak changed:
What|Removed |Added
CC||ebotcazou at gcc dot gnu.org,