| Submitter | Mikael Pettersson |
|---|---|
| Date | July 10, 2010, 9:05 p.m. |
| Message ID | <19512.57519.360717.486839@pilspetsen.it.uu.se> |
| Download | mbox | patch |
| Permalink | /patch/58495/ |
| State | New |
| Headers | show |
Comments
On Sat, Jul 10, 2010 at 11:05 PM, Mikael Pettersson <mikpe@it.uu.se> wrote: > PR testsuite/44325 is about the gcc.dg/vect/vect-109.c FAIL on numerous > targets since the PR44284 fix in late May. This test case now also fails > with 4.5 on those targets since the PR44284 backport a few days ago. > > Recently the test case was fixed on trunk as part of r161797, a vectorizer > realignment improvement patch. This patch backports just the vect-109.c fix, > which suffices to prevent the regression. > > Tested on {powerpc64,sparc64}-linux where it eliminated the new FAIL, and > on i686-linux where it caused no change (since i686 is a vect_hw_misalign > target). > > Ok for 4.5? (I don't have svn write access.) Ok and committed. Richard. > gcc/testsuite/ > > 2010-07-10 Mikael Pettersson <mikpe@it.uu.se> > > PR testsuite/44325 > > Backport from mainline > 2010-07-04 Ira Rosen <irar@il.ibm.com> > Revital Eres <eres@il.ibm.com> > > * gcc.dg/vect/vect-109.c: Expect vectorization only on targets that > support misaligned stores. > > --- gcc-4.5-20100708/gcc/testsuite/gcc.dg/vect/vect-109.c.~1~ 2010-07-09 11:07:19.000000000 +0200 > +++ gcc-4.5-20100708/gcc/testsuite/gcc.dg/vect/vect-109.c 2010-07-10 22:32:58.000000000 +0200 > @@ -72,7 +72,7 @@ int main (void) > return 0; > } > > -/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 2 "vect" } } */ > +/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 2 "vect" { target vect_hw_misalign } } } */ > /* { dg-final { scan-tree-dump-times "not vectorized: unsupported unaligned store" 2 "vect" { xfail vect_hw_misalign } } } */ > /* { dg-final { scan-tree-dump-times "Vectorizing an unaligned access" 10 "vect" { target vect_hw_misalign } } } */ > /* { dg-final { cleanup-tree-dump "vect" } } */ >
Patch
--- gcc-4.5-20100708/gcc/testsuite/gcc.dg/vect/vect-109.c.~1~ 2010-07-09 11:07:19.000000000 +0200 +++ gcc-4.5-20100708/gcc/testsuite/gcc.dg/vect/vect-109.c 2010-07-10 22:32:58.000000000 +0200 @@ -72,7 +72,7 @@ int main (void) return 0; } -/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 2 "vect" } } */ +/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 2 "vect" { target vect_hw_misalign } } } */ /* { dg-final { scan-tree-dump-times "not vectorized: unsupported unaligned store" 2 "vect" { xfail vect_hw_misalign } } } */ /* { dg-final { scan-tree-dump-times "Vectorizing an unaligned access" 10 "vect" { target vect_hw_misalign } } } */ /* { dg-final { cleanup-tree-dump "vect" } } */
PR testsuite/44325 is about the gcc.dg/vect/vect-109.c FAIL on numerous targets since the PR44284 fix in late May. This test case now also fails with 4.5 on those targets since the PR44284 backport a few days ago. Recently the test case was fixed on trunk as part of r161797, a vectorizer realignment improvement patch. This patch backports just the vect-109.c fix, which suffices to prevent the regression. Tested on {powerpc64,sparc64}-linux where it eliminated the new FAIL, and on i686-linux where it caused no change (since i686 is a vect_hw_misalign target). Ok for 4.5? (I don't have svn write access.) gcc/testsuite/ 2010-07-10 Mikael Pettersson <mikpe@it.uu.se> PR testsuite/44325 Backport from mainline 2010-07-04 Ira Rosen <irar@il.ibm.com> Revital Eres <eres@il.ibm.com> * gcc.dg/vect/vect-109.c: Expect vectorization only on targets that support misaligned stores.