This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [testsuite, lto] Solaris 8/9 x86 testsuite fixes
Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE> writes:
> Richard Guenther <richard.guenther@gmail.com> writes:
>
>>> * Similarly, the older assemblers require .globl, not .global, so I've
>>> Âagain conditionalized the directive used.
>
> While re-testing the patch together with some other requested changes, I
> noticed that it still fails with Solaris 8 as:
>
> Assembler:
> "/var/tmp//ccX3aqsV.s", line 2 : Illegal register: "function"
> "/var/tmp//ccX3aqsV.s", line 2 : Syntax error
>
> No idea how this escaped me before. Unlike newer Sun as and gas,
> Solaris 8/x86 as needs @function, not %function, as the type specifier.
> Since this is an i386-specific test and the elfos.h definition (#define
> TYPE_OPERAND_FMT "@%s") is used unconditionally for them, ok to change
> this globally or better move to the #ifdef __sun__ part?
>
>>> diff -r e45414e86d51 gcc/testsuite/gcc.target/i386/pr25993.c
>>> --- a/gcc/testsuite/gcc.target/i386/pr25993.c  Mon May 17 19:09:01 2010 +0200
>>> +++ b/gcc/testsuite/gcc.target/i386/pr25993.c  Mon May 17 19:10:26 2010 +0200
>>> @@ -5,7 +5,11 @@
>>> Â#ifndef __ASSEMBLER__
>>> Âextern int func(void);
>>> Â#else
>>> +#ifdef __sun__
>>> +.globl func
>>> +#else
>>> Â.global func
>>> +#endif
>>> Â.type func,%function
>>> Â.align 4
>>> Âfunc:
Richard, do you have a preference how to fix the %function vs. @function
issue: use @function only for __sun__ or unconditionally?
Thanks.
Rainer
--
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University