diff options
| author | Segher Boessenkool <segher@kernel.crashing.org> | 2018-04-03 17:38:43 +0200 |
|---|---|---|
| committer | Segher Boessenkool <segher@gcc.gnu.org> | 2018-04-03 17:38:43 +0200 |
| commit | 86d13a05fd2207c149311941b49b1866aeabfd5c (patch) | |
| tree | 6007df432848a2068ad638e55213afc9b8669cbf | |
| parent | 3bbeebd4d1ed20f1031d69e8ab646c687d58ee77 (diff) | |
| download | gcc-86d13a05fd2207c149311941b49b1866aeabfd5c.zip gcc-86d13a05fd2207c149311941b49b1866aeabfd5c.tar.gz gcc-86d13a05fd2207c149311941b49b1866aeabfd5c.tar.bz2 | |
rs6000: Fix pr69946.c testcase (PR85126)
After middle-end changes combine now gets fed different input, from
which it makes different (but just as efficient) code. So remove the
test for particular asm output.
gcc/testsuite/
PR target/85126
* gcc.target/powerpc/pr69946: Adjust comment. Remove
scan-assembler-times clause.
From-SVN: r259036
| -rw-r--r-- | gcc/testsuite/ChangeLog | 6 | ||||
| -rw-r--r-- | gcc/testsuite/gcc.target/powerpc/pr69946.c | 7 |
2 files changed, 9 insertions, 4 deletions
diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog index c40efc9..407bb4f 100644 --- a/gcc/testsuite/ChangeLog +++ b/gcc/testsuite/ChangeLog @@ -1,3 +1,9 @@ +2018-04-03 Segher Boessenkool <segher@kernel.crashing.org> + + PR target/85126 + * gcc.target/powerpc/pr69946: Adjust comment. Remove + scan-assembler-times clause. + 2018-04-03 Christophe Lyon <christophe.lyon@linaro.org> * gcc.target/arm/armv8_2-fp16-move-1.c: Move -mfloat-abi=hard to diff --git a/gcc/testsuite/gcc.target/powerpc/pr69946.c b/gcc/testsuite/gcc.target/powerpc/pr69946.c index eb0c365..e0ff422 100644 --- a/gcc/testsuite/gcc.target/powerpc/pr69946.c +++ b/gcc/testsuite/gcc.target/powerpc/pr69946.c @@ -2,9 +2,9 @@ /* { dg-skip-if "" { powerpc_elfv2 } } */ /* { dg-options "-O2" } */ -/* This generates a rotate:DI by 44, with mask 0xf00, which is implemented - using a rlwinm instruction. We used to write 44 for the shift count - there; it should be 12. */ +/* This used to generate a rotate:DI by 44, with mask 0xf00, which is + implemented using a rlwinm instruction. We used to write 44 for the + shift count there; it should be 12. */ struct A { @@ -35,4 +35,3 @@ foo (void) } /* { dg-final { scan-assembler-not {(?n)rlwinm.*,44,20,23} } } */ -/* { dg-final { scan-assembler-times {(?n)rlwinm.*,12,20,23} 1 } } */ |