[dm-devel] xfstests generic/347 was never correct [was: Re: dm: fix inflight IO check]

[Mike Snitzer]

On Tue, Dec 11 2018 at  9:58pm -0500,
Eric Sandeen <sandeen at sandeen.net> wrote:

> On 12/11/18 8:34 PM, Mike Snitzer wrote:
> > On Mon, Dec 10 2018 at  7:32pm -0500,
> > Mike Snitzer <snitzer at redhat.com> wrote:
> > 
> > SO generic/347 needs to account for 2 things:
> > 1) issue $XFS_IO_PROG IO in the background with other process(es)
> > 2) in parallel, in main generic/347 shell, run _dmthin_grow but the
> > "thin" device must be suspended with dmsetup suspend --noflush
> > 
> > Eric Sandeen: can you or someone else xfstests inclined use this info to
> > fix generic/347 please?
> 
> It's not at all clear to me what is wrong with the test.  "Do a bunch of IO
> including a sync, completely filling the backing store, then grow the backing store"
> is surely a real-world scenario that dm-thinp might encounter, no?  Things
> do recover, subsequent writes succeed, and when all is said and done the filesystem
> is checked and it's still consistent.  That sounds like some version of success to me.
> 
> Even if it's emulating a dumb user/administrator, that doesn't seem invalid.
> 
> What is wrong here?

I told you in elaborate detail.

Your willingness to leave generic/347 as is makes little sense; but that
is your call.  I'll just move on...