[PATCH v2 0/3] xfs: fix quota related problems

[PATCH v2 0/3] xfs: fix quota related problems

[Jeff Liu]

Hi Folks,

Here is the revised version for fixing the infinite loop when turning group/project
quota off but keep user quota being active.  After this fix, I ran into another
assertion failure which can be triggered by running fsstress and turning group/project
quota off concurrently.

Since the tests will fail without applying the previous fix:
[PATCH v2] xfs: fix assertion failure at xfs_setattr_nonsize

So I wrapped it up in this series as the 1st fix, maintainers please ignore that in
another post, the 2nd patch is used to fix the infinite loop, the 3rd one is trying
to solve the assertion issue.

Dave,

Since the infinite loop and assertion issue are failed due to different issues, I
wonder if we'd better to add two test cases to cover them separately or not? i.e,
One sanity check to verify turning group/project off quota with user quota is active,
another test to verify turning group/project quota off with fsstress is running.

Thanks,
-Jeff

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: [PATCH v2 0/3] xfs: fix quota related problems

[Christoph Hellwig]

On Tue, Nov 26, 2013 at 09:38:17PM +0800, Jeff Liu wrote:
> Since the infinite loop and assertion issue are failed due to different issues, I
> wonder if we'd better to add two test cases to cover them separately or not? i.e,
> One sanity check to verify turning group/project off quota with user quota is active,
> another test to verify turning group/project quota off with fsstress is running.

I always prefer to have isolated cases for bugs that are trivially
reprodible instead of relying on heavy stress tests with multiple
components. Vote for two test cases from me!

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs