On Sep 17 2018, John David Anglin <dave.ang...@bell.net> wrote: > On 2018-09-17 5:08 AM, Andreas Schwab wrote: >> PR rtl-optimization/85458 >> * sel-sched.c (sel_target_adjust_priority): Remove wrong >> assertion. >> >> diff --git a/gcc/sel-sched.c b/gcc/sel-sched.c >> index 824f1ec340..1be977d70b 100644 >> --- a/gcc/sel-sched.c >> +++ b/gcc/sel-sched.c >> @@ -3330,8 +3330,6 @@ sel_target_adjust_priority (expr_t expr) >> else >> new_priority = priority; >> - gcc_assert (new_priority >= 0); >> - >> /* If the priority has changed, adjust EXPR_PRIORITY_ADJ accordingly. */ >> EXPR_PRIORITY_ADJ (expr) = new_priority - EXPR_PRIORITY (expr); >> > I added the assert because the hppa implementation of > TARGET_SCHED_ADJUST_PRIORITY assumes > scheduling priorities are non negative. If that is not the case, I tend > to think this should be documented. > > It seems ia64 is the only target tripping on the assert.
The assertion only happens at -O3, see <http://gcc.gnu.org/ml/gcc-testresults/2018-09/msg01218.html>. Andreas. -- Andreas Schwab, SUSE Labs, sch...@suse.de GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7 "And now for something completely different."
