| Submitter | Jakub Jelinek |
|---|---|
| Date | Nov. 27, 2012, 2:09 p.m. |
| Message ID | <20121127140952.GC2315@tucnak.redhat.com> |
| Download | mbox | patch |
| Permalink | /patch/202236/ |
| State | New |
| Headers | show |
Comments
On Tue, Nov 27, 2012 at 3:09 PM, Jakub Jelinek <jakub@redhat.com> wrote: > Hi! > > I've noticed we don't schedule asan/tsan passes for -Og, which results in > ICEs. For -O0 that is handled by pass_[at]san_O0, but for -Og which also > doesn't use the normal optimization queue there is nothing. > > Ok for trunk? Ok. Thanks, Richard. > 2012-11-27 Jakub Jelinek <jakub@redhat.com> > > * passes.c (init_optimization_passes): Add pass_asan and pass_tsan > to -Og optimization passes. > > --- gcc/passes.c.jj 2012-11-27 12:08:02.000000000 +0100 > +++ gcc/passes.c 2012-11-27 15:04:59.722410179 +0100 > @@ -1536,6 +1536,8 @@ init_optimization_passes (void) > /* Copy propagation also copy-propagates constants, this is necessary > to forward object-size results properly. */ > NEXT_PASS (pass_copy_prop); > + NEXT_PASS (pass_asan); > + NEXT_PASS (pass_tsan); > NEXT_PASS (pass_rename_ssa_copies); > NEXT_PASS (pass_dce); > /* Fold remaining builtins. */ > > Jakub
Patch
--- gcc/passes.c.jj 2012-11-27 12:08:02.000000000 +0100 +++ gcc/passes.c 2012-11-27 15:04:59.722410179 +0100 @@ -1536,6 +1536,8 @@ init_optimization_passes (void) /* Copy propagation also copy-propagates constants, this is necessary to forward object-size results properly. */ NEXT_PASS (pass_copy_prop); + NEXT_PASS (pass_asan); + NEXT_PASS (pass_tsan); NEXT_PASS (pass_rename_ssa_copies); NEXT_PASS (pass_dce); /* Fold remaining builtins. */