This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH 3/3] testsuite: Use -Wno-psabi instead of pruning compiler messages


On Thu, Oct 06, 2016 at 11:08:52PM +0200, Jakub Jelinek wrote:
> On Thu, Oct 06, 2016 at 07:57:37PM +0000, Segher Boessenkool wrote:
> > It is much nicer to avoid a warning than to manually prune it from the
> > compiler output.  This patch adds -Wno-psabi to various testcases, and
> > removes dg-prune-output, for the testcases where x86 and/or PowerPC had
> > those (or did warn, so the testcase failed).
> > 
> > Tested on powerpc64-linux {-m32,-m64}, and on x86_64-linux.  Should I
> > test something else as well, maybe some options for x86 (which?)
> 
> If you could test on x86_64 with -m32/-mno-sse, it would help, but if you
> don't do that, I'll check it after it is checked in in my i686-linux
> bootstraps.

I'll fire one off, np.

> > --- a/gcc/testsuite/c-c++-common/scal-to-vec2.c
> > +++ b/gcc/testsuite/c-c++-common/scal-to-vec2.c
> > @@ -1,9 +1,7 @@
> >  /* { dg-do compile } */   
> >  /* { dg-options "-fno-common" { target hppa*-*-hpux* } } */
> > -/* { dg-options "-mabi=altivec" { target { { powerpc*-*-linux* } && ilp32 } } } */
> 
> Why are you removing these lines (note, some tests still have them)?

I removed it because I thought it was only there to avoid the problems.
Note it only enables -mabi=altivec for 32-bit.  We also get better test
coverage by *not* overriding the options.  But I later changed my mind,
figured we want it because most 32-bit configurations do not have the
altivec abi enabled by default.  And then I forgot to undo these few, sigh.

> I'd think it would be just enough to add the dg-additional-options.
> For i386 it also adds -msse2, so that even the default is CPU without
> vector ISAs, it still tests those.

Right.

> >  /* { dg-options "-msse2" { target { i?86-*-* x86_64-*-* } } } */
> > -/* Ignore warning on some powerpc-ibm-aix configurations. */
> > -/* { dg-prune-output "non-standard ABI extension" } */
> > +/* { dg-additional-options "-Wno-psabi" } */
> >  
> >  /* Test for C_MAYBE_CONST are folded correctly when 
> >     expanding an expression to vector.  */
> 
> Otherwise LGTM, but I'd say it is not all that we want to do.
> It would be nice to grep also for all tests that have both -Wpsabi and -w
> in dg-options or dg-additional-options, and try to remove the -w
> which got usually added because of these rs6000 warnings.

Both -Wno-psabi and -w, okay.  I looked for just -w and that didn't seem
so feasible to update.


Segher


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]