Message ID | d998ddb5-04c2-ab83-794b-f1a7a75a2325@linux.vnet.ibm.com |
---|---|
State | New |
Headers | show |
Series | [rs6000] Skip safe-indirect-jump-8.c for AIX | expand |
On Sun, Jan 21, 2018 at 8:46 AM, Bill Schmidt <wschmidt@linux.vnet.ibm.com> wrote: > Hi, > > Segher discovered that one of the new safe-indirect-jump tests fails on AIX > because of a difference between AIX_V4 and AIX_ABI for 32-bit code. The > former generates sibcalls for nonlocal function calls, but the latter does > not. Thus this test should be skipped for AIX. > > Okay for trunk and possible backport to 7 after a quick AIX sniff test? > > Thanks, > Bill > > > 2018-01-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com> > > PR target/83946 > * gcc.target/powerpc/safe-indirect-jump-8.c: Skip for AIX. I noticed the same failure. The current testcase partially succeeds. The test either can be completely skipped or some of the expected assembler output ignored on AIX. Thanks, David
Hi David, On Sun, Jan 21, 2018 at 09:27:42AM -0500, David Edelsohn wrote: > On Sun, Jan 21, 2018 at 8:46 AM, Bill Schmidt > <wschmidt@linux.vnet.ibm.com> wrote: > > Segher discovered that one of the new safe-indirect-jump tests fails on AIX > > because of a difference between AIX_V4 and AIX_ABI for 32-bit code. The > > former generates sibcalls for nonlocal function calls, but the latter does > > not. Thus this test should be skipped for AIX. > > > > Okay for trunk and possible backport to 7 after a quick AIX sniff test? > > > > Thanks, > > Bill > > > > > > 2018-01-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com> > > > > PR target/83946 > > * gcc.target/powerpc/safe-indirect-jump-8.c: Skip for AIX. > > I noticed the same failure. The current testcase partially succeeds. > The test either can be completely skipped or some of the expected > assembler output ignored on AIX. Yes, the crset test works. Maybe we should have safe-indirect-jump-1.c run on 32-bit AIX as well? Segher
Hi! On Sun, Jan 21, 2018 at 07:46:26AM -0600, Bill Schmidt wrote: > Segher discovered that one of the new safe-indirect-jump tests fails on AIX > because of a difference between AIX_V4 and AIX_ABI for 32-bit code. The > former generates sibcalls for nonlocal function calls, but the latter does > not. Thus this test should be skipped for AIX. > > Okay for trunk and possible backport to 7 after a quick AIX sniff test? That's fine. Maybe figure out how to make the -1 test run on 32-bit AIX as well (see other thread)? But it's okay either way. For the 7 backport you need RM approval, you know the drill :-) Thanks, Segher > 2018-01-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com> > > PR target/83946 > * gcc.target/powerpc/safe-indirect-jump-8.c: Skip for AIX. > > > Index: gcc/testsuite/gcc.target/powerpc/safe-indirect-jump-8.c > =================================================================== > --- gcc/testsuite/gcc.target/powerpc/safe-indirect-jump-8.c (revision 256931) > +++ gcc/testsuite/gcc.target/powerpc/safe-indirect-jump-8.c (working copy) > @@ -1,4 +1,5 @@ > /* { dg-do compile { target { ilp32 } } } */ > +/* { dg-skip-if "" { powerpc*-*-aix* } } */ > /* { dg-additional-options "-O2 -mno-speculate-indirect-jumps" } */ > > /* Test for deliberate misprediction of -m32 sibcalls. */
Index: gcc/testsuite/gcc.target/powerpc/safe-indirect-jump-8.c =================================================================== --- gcc/testsuite/gcc.target/powerpc/safe-indirect-jump-8.c (revision 256931) +++ gcc/testsuite/gcc.target/powerpc/safe-indirect-jump-8.c (working copy) @@ -1,4 +1,5 @@ /* { dg-do compile { target { ilp32 } } } */ +/* { dg-skip-if "" { powerpc*-*-aix* } } */ /* { dg-additional-options "-O2 -mno-speculate-indirect-jumps" } */ /* Test for deliberate misprediction of -m32 sibcalls. */