unresolved's on gcc profiling tests

Hi,

The other source of noise in RTEMS test results
are the profiling tests.  Again the latest
powerpc-rtems results:

http://gcc.gnu.org/ml/gcc-testresults/2008-12/msg01477.html

A typical noise entry is:

FAIL: gcc.dg/matrix/transpose-6.c compilation,  -fprofile-generate -O3
UNRESOLVED: gcc.dg/matrix/transpose-6.c execution,    -fprofile-generate -O3
UNRESOLVED: gcc.dg/matrix/transpose-6.c compilation,  -fprofile-use
-fipa-matrix-reorg -fdump-ipa-matrix-reorg -O3 -fwhole-program -combine
UNRESOLVED: gcc.dg/matrix/transpose-6.c execution,    -fprofile-use
-fipa-matrix-reorg -fdump-ipa-matrix-reorg -O3 -fwhole-program -combine

So generally saying profiling isn't going to
work on this target would eliminate a bunch of
noise.

My goal is to get results that accurately reflect
the target so I can then begin to focus more on the
real failures.

Thanks.

--
Joel Sherrill, Ph.D.             Director of Research & Development
joel.sherrill@...        On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
   Support Available             [phone]




_______________________________________________
DejaGnu mailing list
DejaGnu@...
http://lists.gnu.org/mailman/listinfo/dejagnu

[Janis Johnson]

On Wed, 2008-12-17 at 16:09 -0600, Joel Sherrill wrote:

> Hi,
>
> The other source of noise in RTEMS test results
> are the profiling tests.  Again the latest
> powerpc-rtems results:
>
> http://gcc.gnu.org/ml/gcc-testresults/2008-12/msg01477.html
>
> A typical noise entry is:
>
> FAIL: gcc.dg/matrix/transpose-6.c compilation,  -fprofile-generate -O3
> UNRESOLVED: gcc.dg/matrix/transpose-6.c execution,    -fprofile-generate -O3
> UNRESOLVED: gcc.dg/matrix/transpose-6.c compilation,  -fprofile-use
> -fipa-matrix-reorg -fdump-ipa-matrix-reorg -O3 -fwhole-program -combine
> UNRESOLVED: gcc.dg/matrix/transpose-6.c execution,    -fprofile-use
> -fipa-matrix-reorg -fdump-ipa-matrix-reorg -O3 -fwhole-program -combine
>
> So generally saying profiling isn't going to
> work on this target would eliminate a bunch of
> noise.
>
> My goal is to get results that accurately reflect
> the target so I can then begin to focus more on the
> real failures.

This is specific to the GCC testsuite, which already has the support
you need.  Add your target(s) to check_profiling_available in
gcc/lib/target-supports.exp.

Janis



_______________________________________________
DejaGnu mailing list
DejaGnu@...
http://lists.gnu.org/mailman/listinfo/dejagnu

Janis Johnson wrote:

> On Wed, 2008-12-17 at 16:09 -0600, Joel Sherrill wrote:
>  
>> Hi,
>>
>> The other source of noise in RTEMS test results
>> are the profiling tests.  Again the latest
>> powerpc-rtems results:
>>
>> http://gcc.gnu.org/ml/gcc-testresults/2008-12/msg01477.html
>>
>> A typical noise entry is:
>>
>> FAIL: gcc.dg/matrix/transpose-6.c compilation,  -fprofile-generate -O3
>> UNRESOLVED: gcc.dg/matrix/transpose-6.c execution,    -fprofile-generate -O3
>> UNRESOLVED: gcc.dg/matrix/transpose-6.c compilation,  -fprofile-use
>> -fipa-matrix-reorg -fdump-ipa-matrix-reorg -O3 -fwhole-program -combine
>> UNRESOLVED: gcc.dg/matrix/transpose-6.c execution,    -fprofile-use
>> -fipa-matrix-reorg -fdump-ipa-matrix-reorg -O3 -fwhole-program -combine
>>
>> So generally saying profiling isn't going to
>> work on this target would eliminate a bunch of
>> noise.
>>
>> My goal is to get results that accurately reflect
>> the target so I can then begin to focus more on the
>> real failures.
>>    
>
> This is specific to the GCC testsuite, which already has the support
> you need.  Add your target(s) to check_profiling_available in
> gcc/lib/target-supports.exp.
>  

Thanks for the tip.  Does this look about right?  I have a
powerpc run going now and will repeat it with this patch
in if you think it looks right. :)

Index: lib/target-supports.exp
===================================================================
--- lib/target-supports.exp    (revision 142815)
+++ lib/target-supports.exp    (working copy)
@@ -452,6 +452,11 @@
     return 0
     }
 
+    # At present, there is no profiling support on RTEMS.
+    if { [istarget *-*-rtems*] } {
+    return 0
+    }
+
     # At present, there is no profiling support on NetWare.
     if { [istarget *-*-netware*] } {
     return 0


> Janis
>
>  


--
Joel Sherrill, Ph.D.             Director of Research & Development
joel.sherrill@...        On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
   Support Available             [phone]




_______________________________________________
DejaGnu mailing list
DejaGnu@...
http://lists.gnu.org/mailman/listinfo/dejagnu