This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: s390 & other targets misc ACATS failures
- From: Ulrich Weigand <weigand at i1 dot informatik dot uni-erlangen dot de>
- To: laurent at guerby dot net (Laurent GUERBY)
- Cc: weigand at i1 dot informatik dot uni-erlangen dot de (Ulrich Weigand), gcc at gcc dot gnu dot org
- Date: Sat, 24 Apr 2004 18:29:15 +0200 (CEST)
- Subject: Re: s390 & other targets misc ACATS failures
Laurent Guerby wrote:
> Just to let you know, the following tests fail on 64 bit
> targets because of a test configuration problem:
>
> FAIL: cd2a83c
> FAIL: cd2a91c
>
> In testsuite/ada/acats/support/macro.dfs the two
> harcdoded 32 (ACC_SIZE and TASK_SIZE) need
> to be computed for the target (I suspect 64
> for s390, same for amd64).
I assume ACC_SIZE means the size of a pointer, and likewise
TASK_SIZE (I'm not sure how exactly a 'TASK OBJECT' is represented)?
If so, this needs to be 64 for s390x, and 32 for s390.
> FAIL: ad8011a
>
> This one tests an asm inline statement and it is
> configured by macro.dfs to use:
>
> MACHINE_CODE_STATEMENT Asm_Insn'(Asm ("nop"));
>
> So may be "nop" is not supported on s390,
> could you provide me with a suitable s390 replacement
> if the testsuite log confirms that it is complaining
> about "nop"?
Yes, it is indeed the "nop". As valid nop on s390 you can use
"nop 0". See also the configure.ac section used to define
gcc_cv_as_dwarf2_debug_line: there's a selection of nop insns
for various platforms there.
Thanks for looking after this!
Bye,
Ulrich
--
Dr. Ulrich Weigand
weigand@informatik.uni-erlangen.de