[Crackerjack-users 00063] Re: [LTP] eventfd testcases

[Michael K. <mtk...@go...>]

On Fri, Jul 25, 2008 at 2:26 AM, Subrata Modak
<su...@li...> wrote:
>
> On Fri, 2008-07-25 at 01:12 +0200, Michael Kerrisk wrote:
>> Vijay,
>>
>> On Thu, Jul 24, 2008 at 7:37 PM, Vijay Kumar <vij...@br...> wrote:
>> > Hi Subrata,
>> > I am in the process of writing test cases for eventfd()
>> > syscall. Please do let me know if anybody else is also working on the
>> > it.
>> >
>> > Here is a list of test cases I am implementing.
>> >
>> > 1. Reading counter value.
>> > 2. EAGAIN for read in non-blocking mode when counter value is zero.
>> > 3. Writing counter value.
>> > 4. EAGAIN for write in non-blocking mode when counter value exceed max.
>> > 5. EINVAL for write when buffer size is less than 8 in read and write.
>> > 6. EINVAL for write when the counter is set to 0xffffffffffffffff.
>> > 7. In select, check if readfds is set when counter value is non-zero.
>> > 8. In select, check if readfds is not set when counter value is zero.
>> > 9. In select, check if writefds is set when counter value is non-max.
>> > 10. In select, check if writefds is not set when counter value is max.
>> > 11. Check if file descriptor is inherited by child.
>>
>> Some others to consider adding:
>>
>> * Check whether counter is correctly initialized from initval.
>> * I'm not sure what kind of check you are doing in 11, but perhaps one
>> should somewhere have a test that checks whether writing to the
>> eventfd in the child means the child can read from the eventfd.
>> * exceptional condition (select(exceptfds) when the counter is
>> overflowed; and in this case read() shd return 0xffffffffffffffff.
>> * eventfd file descriptor is preserved across execve()
>>
>> If you want more detail on any of the above, let me know.
>
> Vijay,
>
> Please work with Michael to get more information and clarification
> before you start writing the tests.

Yes -- I should have added, I'm happy to help, and review the test.
Since I haven't actually written something that uses the LTP
framework, this could also be some useful education for me.  To make
that especially useful, it might be easiest for me if you start off
writing a very simple test suite (unless there is already something in
Crackerjack), perhaps one that tests one or two cases (and send it to
me for review/learning), and then build up from there.

Cheers,

Michael

> Also check whether the Crackerjack
> guys has already written them. If yes, then you can get direct reference
> from them and port the test case(s) from there to LTP. If necessary then
> you can also write additional tests. You may have observed Masatake
> doing that for couple of test cases already.
>
> Regards--
> Subrata
>
>>
>> Cheers,
>>
>> Michael
>>
>> -------------------------------------------------------------------------
>> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
>> Build the coolest Linux based applications with Moblin SDK & win great prizes
>> Grand prize is a trip for two to an Open Source event anywhere in the world
>> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>> _______________________________________________
>> Ltp-list mailing list
>> Ltp...@li...
>> https://lists.sourceforge.net/lists/listinfo/ltp-list
>
>



-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
man-pages online: http://www.kernel.org/doc/man-pages/online_pages.html
Found a bug? http://www.kernel.org/doc/man-pages/reporting_bugs.html