On 05/25/2015 11:51 AM, Cyril Hrubis wrote:
> Hi!
>> As of now, if we run test cases which use the tst_checkpoint framework on 
>> NFS,
>> they will fail with errors like:
>>
>> creat07     0  TWARN  :  tst_tmpdir.c:206: tst_rmdir: 
>> rmobj(/tmpdir/ltp-0IVQPP0NK6/creigi0kV) failed: 
>> remove(/tmpdir/ltp-0IVQPP0NK6/creigi0kV) failed; errno=66: Directory not 
>> empty
>>
>> This happens because we keep the futex page mapped while tst_rmdir() is 
>> running.
>>
>> So on we need to unmap the page before tst_rmdir().
>>
>> What do you think about the proposed solution?
>
> Why don't we hide this from user and unmap() the file in the tst_rmdir()
> if checkpoint was initialized?

Ok, we can do it this way.

One question regarding the typedef of futex_t. As of now the futex_t 
type is defined in two places:
  * testcases/kernel/syscalls/futex/futextest.h
  * lib/tst_checkpoint.h

If I put unmap() in tst_rmdir(), I will add futex_t to tst_tmpdir.c 
either, so it will be the third occurrence.

It seems it would be good to move the typedef of futex_t into a common 
header.

I'm unsure which one to use :) Or just create include/lapi/futex.h?

Thanks.


>
> Otherwise I'm all for this fix.
>

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud 
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Reply via email to