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: altivec testsuite dejagnuified and committed


> With powerpc-*-linux-gnu and powerpc64-*-linux-gnu it's always possible
> to generate Altivec instructions, but not always possible to run that
> code.  Would it be OK for the new vmx tests to use altivec_check() from
> gcc.dg/altivec_check.h, and to be enabled for powerpc*linux?

This would be great, however the powerpc-eabi simulated target will
die a horrible death when executing altivec code.  Powerpc-eabispe on
the other hand, has conflicting opcodes for altivec and e500, so that
won't work either.

Powerpc-eabi gives:

	     cpu 1, cia 0x10000228: program interrupt - illegal instruction
	     a.out: Caught signal -1 at address 0x10000228

and it doesn't catch the signal :-(.

I'll approve your patch, if it disables the tests on powerpc-eabi and
powerpc-eabispe.

How about AIX.  Have y'all tested altivec_check() on AIX?

Aldy


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