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: Consistently generate widening multiplies


On Thu, 22 Apr 2010, Bernd Schmidt wrote:

> On 04/22/2010 05:43 PM, H.J. Lu wrote:
> > gcc.target/i386/wmul-1.c is 32bit only. It fails on 64bit target:
> > 
> > Executing on host: /export/gnu/import/svn/gcc-test/bld/gcc/xgcc
> > -B/export/gnu/import/svn/gcc-test/bld/gcc/
> > /export/gnu/import/svn/gcc-test/src-trunk/gcc/testsuite/gcc.target/i386/wmul-1.c
> >   -O2 -S  -o wmul-1.s    (timeout = 300)
> > PASS: gcc.target/i386/wmul-1.c (test for excess errors)
> > FAIL: gcc.target/i386/wmul-1.c scan-assembler-times imull 2
> 
> Adding
> 
> /* { dg-options "-m32" } */
> 
> should fix it, shouldn't it?

It's essentially never correct for a gcc.target/i386 test to use -m32 or 
-m64 explicitly.  The multilib options go after the dg-options and 
override them, so if a test needs to run in 32-bit or 64-bit mode then 
specifying those options in dg-options will not achieve that effect; the 
tests will fail in multilib testing.  Instead, use

/* { dg-require-effective-target ilp32 } */

for a 32-bit-only test.

-- 
Joseph S. Myers
joseph@codesourcery.com


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