https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109520
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |vmakarov at gcc dot gnu.org
--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
ISTR a duplicate. We're looping in lra ()
do
{
...
if (fails_p)
{
/* It is a very rare case. It is the last hope to
split a hard regno live range for a reload
pseudo. */
if (live_p)
lra_clear_live_ranges ();
live_p = false;
if (! lra_split_hard_reg_for ())
break;
lra_hard_reg_split_p = true;
}
}
while (fails_p && !lra_asm_error_p);
not sure if that's ever expected to execute the if (fails_p) case more
than once? That is, lra_split_hard_reg_for for some reason does something
twice?