On Sun, Oct 11, 2026 at 11:14 PM Torbjörn SVENSSON <[email protected]> wrote: > > Ok for trunk? Not exactly ok ...
> > -- > > The tests were added in commit r17-4921-g8b90726a28bb63, but > fp-trapping-cmp-7.c has never passed on Cortex-M targets that disable > non-short-circuit operations. > > The optimizations checked by these tests depend on the target's > logical-op-non-short-circuit setting. Enable it explicitly in both > tests so they check the intended ifcombine behavior independently of > target tuning. > > PR tree-optimization/53805 > > gcc/testsuite/ChangeLog: > > * gcc.dg/tree-ssa/fp-trapping-cmp-6.c: Pass --param > logical-op-non-short-circuit=1. > * gcc.dg/tree-ssa/fp-trapping-cmp-7.c: Likewise > > Signed-off-by: Torbjörn SVENSSON <[email protected]> > --- > gcc/testsuite/gcc.dg/tree-ssa/fp-trapping-cmp-6.c | 2 +- > gcc/testsuite/gcc.dg/tree-ssa/fp-trapping-cmp-7.c | 2 +- > 2 files changed, 2 insertions(+), 2 deletions(-) > > diff --git a/gcc/testsuite/gcc.dg/tree-ssa/fp-trapping-cmp-6.c > b/gcc/testsuite/gcc.dg/tree-ssa/fp-trapping-cmp-6.c > index 4ffb5a140c0..553b31b364a 100644 > --- a/gcc/testsuite/gcc.dg/tree-ssa/fp-trapping-cmp-6.c > +++ b/gcc/testsuite/gcc.dg/tree-ssa/fp-trapping-cmp-6.c > @@ -1,5 +1,5 @@ > /* { dg-do compile } */ > -/* { dg-options "-O1 -ftrapping-math -fdump-tree-ifcombine-details" } */ > +/* { dg-options "-O1 -ftrapping-math -fdump-tree-ifcombine-details --param > logical-op-non-short-circuit=1" } */ This should not depend on the setting of logical-op-non-short-circuit. > > int > f (double a, double b) > diff --git a/gcc/testsuite/gcc.dg/tree-ssa/fp-trapping-cmp-7.c > b/gcc/testsuite/gcc.dg/tree-ssa/fp-trapping-cmp-7.c > index 5916344877f..51a6f3bc761 100644 > --- a/gcc/testsuite/gcc.dg/tree-ssa/fp-trapping-cmp-7.c > +++ b/gcc/testsuite/gcc.dg/tree-ssa/fp-trapping-cmp-7.c > @@ -1,5 +1,5 @@ > /* { dg-do compile } */ > -/* { dg-options "-O1 -ftrapping-math -fdump-tree-ifcombine-details" } */ > +/* { dg-options "-O1 -ftrapping-math -fdump-tree-ifcombine-details --param > logical-op-non-short-circuit=1" } */ Can you make 2 versions of this testcase. one for logical-op-non-short-circuit = 1 and one for logical-op-non-short-circuit=0? The one for logical-op-non-short-circuit=0 should do scan-tree-dump-not for both dg-final? > > int > f (double a, double b) > -- > 2.43.0 >
