Re: [PATCH v2 0/3] xfs/348: test handling of malformed inode mode

[Eryu Guan <eguan@redhat.com>]

On Sun, Dec 25, 2016 at 08:09:33PM +0200, Amir Goldstein wrote:
> Eryu,
> 
> I beefed up the initial test sent earlier today with more checks
> on mounted fs.
> 
> The original patch testing only xfs_repair remains patch 1 in this series.
> 
> Patch 2 adds fstat tests on mounted fs, which are safe on my test system.
> 
> Patch 3 is explosive. It exposes an XFS assert, but I left a "safety pin"
> that needs to be commented out to reproduce the assert.

My kenrel config doesn't turn DEBUG on, so I don't see a kernel crash :)

I can push this test out after the fix lands in upstream, then I think
there's no need to leave a switch in the test.

Otherwise tests look good to me. But I'd like to have Darrick to review
too, as he had written many fuzzer tests and suggested this test :)

Thanks,
Eryu

> 
> I will soon post my kernel configs and dmesg logs.
> 
> Amir.
> 
> v2:
> - test fstat with malformed inode mode
> - test readlink/readdir with malformed inode mode
> 
> v1:
> - test xfs_repair with malformed inode mode
> 
> Amir Goldstein (3):
>   xfs/348: test handling of invalid inode modes
>   xfs/348: test fstat with malformed inode mode
>   xfs/348: test readlink/readdir with malformed inode mode
> 
>  tests/xfs/348     | 154 +++++++++++++++++++++++++++
>  tests/xfs/348.out | 312 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
>  tests/xfs/group   |   1 +
>  3 files changed, 467 insertions(+)
>  create mode 100755 tests/xfs/348
>  create mode 100644 tests/xfs/348.out
> 
> -- 
> 2.7.4
>