On 3/6/20 6:07 PM, Segher Boessenkool wrote:
On Fri, Mar 06, 2020 at 02:07:15PM +0100, Martin Liška wrote:
After r9-1623-gc518c1025b435e1c593a745036fc9b8ed04c5819 the code was
changed to:
- if (align_jumps_max_skip <= 0)
- align_jumps_max_skip = 15;
- if (align_loops_max_skip <= 0)
- align_loops_max_skip = 15;
+
+ if (flag_align_jumps && !str_align_jumps)
+ str_align_jumps = "16";
+ if (flag_align_loops && !str_align_loops)
+ str_align_loops = "16";
which for situation where align_* was 0 caused that max_skip didn't play
any role.
My code wrongly changed that to str_align_jumps, which is now the was which
includes
both alignment (and max skip). The hunk should not be here.
Do you have a testsuite test as well? Or, how else was this checked?
Yes, I've extended the patch by addition of a test-case.
PR target/93800
* config/rs6000/rs6000.c (rs6000_option_override_internal):
Remove set of str_align_loops and str_align_jumps as these
should be set in previous 2 conditions in the function.
Okay for trunk, maybe with an added test. Thanks!
I'm going to install the patch.
Martin
Segher
