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, testsuite] Fix some testcases for nds32 target and provide new nds32 target specific tests


On 11/30/13, 9:07 AM, Mike Stump wrote:
> On Nov 28, 2013, at 2:03 AM, Chung-Ju Wu <jasonwucj@gmail.com> wrote:
>> There is a pending testsuite patch for nds32 target:
>>    http://gcc.gnu.org/ml/gcc-patches/2013-11/msg01584.html
>>
>> Is it OK for trunk? :)
> 
> Ok, but please remove:
> 
>   { target nds32*-*-* }
> 

Ah... I see.  Since I have following code fragment in nds32.exp:

  # Exit immediately if this isn't a nds32 target.
  if ![istarget nds32*-*-*] then {
    return
  }

all the "{ target nds32*-*-* }" from gcc.target test cases are unnecessary.

Thanks.  Remove it accordingly.


> from the gcc.target test cases, it is not redundant with the .exp file.
> 
> A few oddities I will note, if you can improve them… that'd be nice:
> 
>> +/* { dg-skip-if "Variadic funcs arguments will push by caller for current nds32 porting." { nds32*-*-* } } */
> 
> This is a bit weird.  There isn't a notion of current, there isn't a notion of porting.  Variadic funcs arguments are caller pushed?
> 

There are two approaches to deal with variadic function arguments
and I had a post on:
  http://gcc.gnu.org/ml/gcc-help/2013-03/msg00208.html

So far, in the nds32 port, we are using the approach like aarch64/avr does.
This test case is almost the same as gcc/testsuite/gcc.dg/builtin-apply2.c file.
Perhaps I should have used the following description, which seems much better:

+/* { dg-skip-if "Variadic funcs have all args on stack. Normal funcs have args in registers." { nds32*-*-* } "*" "" } */


>> +/* { dg-skip-if "nds32 target has special operations for 64-bit behavior" { nds32*-*-* }  { "*" } { "" } } */
> 
> This is a bit weird.
> 

Oops, you are right.
That is a legacy modification which is not
suitable for the nds32 port on trunk.
Thanks for the catch.  I will not commit that bit.

Thank you for the review and approval.
Attachment is the revised patch and I will apply it
tomorrow if there is no other comment. :)


Best regards,
jasonwucj




Attachment: nds32-testsuite.v2.patch
Description: Text document


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