Re: [PATCH v2] fstests: regression test for btrfs buffered read's repair

[Liu Bo <bo.li.liu@oracle.com>]

On Fri, Apr 28, 2017 at 10:52:12AM +0100, Filipe Manana wrote:
> On Wed, Apr 26, 2017 at 7:09 PM, Liu Bo <bo.li.liu@oracle.com> wrote:
> > This case tests whether buffered read can repair the bad copy if we
> > have a good copy.
> >
> > Commit 20a7db8ab3f2 ("btrfs: add dummy callback for readpage_io_failed
> > and drop checks") introduced the regression.
> >
> > The upstream fix is
> >         Btrfs: bring back repair during read
> 
> Same issue as reported for the other new test (btrfs/140), the test
> fails on a kernel with the mentioned patch. Seems like it does wrong
> assumptions somewhere (I haven't investigated, just run the test):

Thanks for running the test, what I've missed is that chunk start offset depends
on disk size, so I'll add a filesystem size limit to mkfs.

Thanks,

-liubo