[Ocfs2-test-devel] Hello, any suggestion and update for ocfs2-test?

[Eric Ren]

Hi Junxiao,

>> 1. tell ocfs2-test a specified list of testcases to run
>>
>> This has been achieved by your patch introducing "-t" option. However,
>> this option can only specify one case at a time, making the log of test
>> results distributed among many directories. That's inconvenient to get a
>> whole picture of the result at a glance;-)
> OK, please fix the check part.

Done. I will send together with series v2;-)

>>
>> 2. eliminating 2-layer loop for blocksize and clustersize by introducing
>> "-b" (blocksize) and "-c" (clustersize) options so that we can loop
>> different blocksize and clustersize outside of ocfs2-test
> Looks good to test with specified block/cluster size, but I think it'd
> better leave the original block/cluster coverage test there, it can be
> made as a full test, it will be run if block/cluster is not given.

OK, totally agree with you, thanks.

>>
>> This 2 patches [2][3] are for it.
>>
>> 3. different combinations of features and options for mkfs and mount
>> operations, like
>> "inline, sparse, unwritten, ...", and "journal={data | order|
>> writeback}" etc
>>
>> I intend to untouch this for now. Maybe a global configuration file can
>> be introduced to save different combinations for different test cases in
>> the future. Thus, make it more configurable in a common file;-)
>>
>> [1] https://oss.oracle.com/pipermail/ocfs2-test-devel/2016-June/000849.html
>> [2] https://oss.oracle.com/pipermail/ocfs2-test-devel/2016-June/000850.html
>> [3] https://oss.oracle.com/pipermail/ocfs2-test-devel/2016-June/000851.html
>>
>> Could you help review those fixes? Thus, I can push the fixes with a
>> "reviewed-by" to
>> upstream, and improve/resend the others.

How about this?

>>
>> BTW, any update for ocfs2-test from your side? I'd be very appreciated
>> to see them;-)
> No new patches from my side, all have been pushed to upstream.

Thanks! I asked, to make sure no overlapping work between us. ^o^

Regards,
Eric

>
> Thanks,
> Junxiao.
>
>>
>> Regards,
>> Eric
>
>