Re: Locking API testsuite problem - Microblaze

[Michal Simek <monstr@seznam.cz>]

Hi Matthew,

> On Sat, May 17, 2008 at 09:07:36AM +0200, Michal Simek wrote:
>> Hi Richard, Tony, Arnd and others,
>>
>> I turn on API testsuite for Microblaze CPU for stable 2.6.24 and I got nice log
>> with some faults. The same problem I have on latest version.
>>
>> I found the similar problem for alpha CPU and ia64 but I haven't found solving
>> these issues. Could you give me a hint how to solve this problem.
> 
> You missed the crucial line:

oou.

> 143 out of 218 testcases failed, as expected.
> 
> Some of these testcases are /expected/ to fail.  Succeeding would be an
> bug.  There are no bugs here.

Thanks for your help. Are these tescases architecture dependent? Because I saw
some logs from other arch and they have everything OK.

Does it mean that everything what I supported works well?

I look to source code and I found these two lines.

DO_TESTCASE_6("double unlock", double_unlock);
DO_TESTCASE_6("initialize held", init_held);

and macro

#define DO_TESTCASE_6(desc, name)                               \
         print_testname(desc);                                   \
         dotest(name##_spin, FAILURE, LOCKTYPE_SPIN);            \
         dotest(name##_wlock, FAILURE, LOCKTYPE_RWLOCK);         \
         dotest(name##_rlock, FAILURE, LOCKTYPE_RWLOCK);         \
         dotest(name##_mutex, FAILURE, LOCKTYPE_MUTEX);          \
         dotest(name##_wsem, FAILURE, LOCKTYPE_RWSEM);           \
         dotest(name##_rsem, FAILURE, LOCKTYPE_RWSEM);           \
         printk("\n");

My results
                             | spin |wlock |rlock |mutex | wsem | rsem |
                double unlock:  ok  |  ok  |failed|failed|failed|failed|
              initialize held:failed|failed|failed|failed|failed|failed|

Is this result OK? Because macro say that expected value is failed.

Regards,
Michal