xfs
[Top] [All Lists]

Re: Test 274 in xfstests

To: Eric Sandeen <sandeen@xxxxxxxxxxx>
Subject: Re: Test 274 in xfstests
From: Jan Kara <jack@xxxxxxx>
Date: Tue, 15 May 2012 22:16:45 +0200
Cc: Jan Kara <jack@xxxxxxx>, wu.bo@xxxxxxxxxxxxxx, xfs@xxxxxxxxxxx
In-reply-to: <4FB29C85.5000603@xxxxxxxxxxx>
References: <20120515180521.GE26579@xxxxxxxxxxxxx> <4FB29C85.5000603@xxxxxxxxxxx>
User-agent: Mutt/1.5.20 (2009-06-14)
On Tue 15-05-12 13:12:21, Eric Sandeen wrote:
> On 5/15/12 1:05 PM, Jan Kara wrote:
> >   Hello,
> > 
> >   test 274 fails e.g. for ext3 because it does not support fallocate(). So
> > I would think the best way to fix the is to add to the test
> >   _require_xfs_io_falloc()
> > so that the test is run only for filesystems which do support fallocate.
> > But with that is connected one slightly related question - why does the
> > test use fallocate(1) binary instead of "xfs_io falloc" command which is
> > more common in xfstests?
> > 
> >                                                             Honza
> 
> Argh I'm behind.  I had:
> 
> [PATCH V2] xfstests: several 274 fixups
> 
> on the list way too long ago.
> 
> Want to test that out?  It has 1 review on the list, I'll merge it
> soon - if you want to give it a quick test first, go for it :)
  Yup, the patch works for me (since I'm mostly interested in ext3 not
being tested ;). I also reviewed your patch and it looks OK so feel free to
add:
  Reviewed-by: Jan Kara <jack@xxxxxxx>

  Just one comment. Messages like:
echo "Fill fs with 1M IOs; EIO expected" >> $seq.full
  should probably speak of ENOSPC, not EIO, shouldn't they?

                                                                Honza
-- 
Jan Kara <jack@xxxxxxx>
SUSE Labs, CR

<Prev in Thread] Current Thread [Next in Thread>