https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70764
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Summary|PASS->FAIL: |FAIL:
|gcc.dg/guality/pr41447-1.c |gcc.dg/guality/pr41447-1.c
| -O2 -flto | fails on aarch64
|-fno-use-linker-plugin |
|-flto-partition=none |
--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to ktkachov from comment #0)
> Note that this test was already failing for other optimisation levels before
> that change:
>
> FAIL: gcc.dg/guality/pr41447-1.c -O2 execution test
> FAIL: gcc.dg/guality/pr41447-1.c -O3 -g execution test
> FAIL: gcc.dg/guality/pr41447-1.c -Os execution test
>
> That patch controlled the way -fomit-frame-pointer is turned on, which
> probably affected debug info generation, so it is very likely that the patch
> just triggered the same issue that caused the failure on the other
> optimisation levels in this test.
Right, in the case of -flto, the optimize attribute was being added to the
functions so the testcase was actually just accidently passing before the
patch. And since it was failing without -flto it is not a regression
