Hi!
Honza changed recently max-completely-peeled-insns param default from 400 to
100, which broke the testcase assumptions.
Fixed by forcing the old default, everywhere, committed to trunk as obvious.
2012-12-03 Jakub Jelinek <ja...@redhat.com>
PR testsuite/55452
* gfortran.dg/reassoc_4.f: Use --param max-completely-peeled-insns=400
on all targets, not just s390*.
--- gcc/testsuite/gfortran.dg/reassoc_4.f.jj 2012-01-30 00:09:59.000000000
+0100
+++ gcc/testsuite/gfortran.dg/reassoc_4.f 2012-12-03 16:14:56.429727381
+0100
@@ -1,7 +1,6 @@
! { dg-do compile }
-! { dg-options "-O3 -ffast-math -fdump-tree-reassoc1" }
+! { dg-options "-O3 -ffast-math -fdump-tree-reassoc1 --param
max-completely-peeled-insns=400" }
! { dg-additional-options "--param max-completely-peel-times=16" { target
spu-*-* } }
-! { dg-additional-options "--param max-completely-peeled-insns=400" { target
s390*-*-* } }
subroutine anisonl(w,vo,anisox,s,ii1,jj1,weight)
integer ii1,jj1,i1,iii1,j1,jjj1,k1,l1,m1,n1
real*8 w(3,3),vo(3,3),anisox(3,3,3,3),s(60,60),weight
Jakub
