All of lore.kernel.org
 help / color / mirror / Atom feed
From: Eryu Guan <eguan@redhat.com>
To: Dave Chinner <david@fromorbit.com>
Cc: fstests@vger.kernel.org
Subject: Re: [PATCH 4/4] filter: inode size output of mkfs.xfs can change
Date: Thu, 7 May 2015 11:05:38 +0800	[thread overview]
Message-ID: <20150507030538.GH1187@dhcp-13-216.nay.redhat.com> (raw)
In-Reply-To: <20150506225627.GJ21261@dastard>

On Thu, May 07, 2015 at 08:56:27AM +1000, Dave Chinner wrote:
> On Wed, May 06, 2015 at 03:19:34PM +0800, Eryu Guan wrote:
> > On Tue, May 05, 2015 at 08:01:33AM +1000, Dave Chinner wrote:
> > > From: Dave Chinner <dchinner@redhat.com>
> > > 
> > > With the change to CRCs by default, the mkfs inode size is defaults
> > > to 512 bytes and the minimum block size changes to 1024 bytes. This
> > > causes mismatches with golden output that expects the inode size to
> > > be 256 bytes, and some tests are tailored around the amount of space
> > > inside a 256 byte inode. Fix them with appropriate filtering or mkfs
> > > parameters to allow 256 byte inodes to be used.
> > > 
> > > Signed-off-by: Dave Chinner <dchinner@redhat.com>
> > > ---
[snip]
> > > -_scratch_mkfs_xfs -dsize=41m,agcount=2 | _filter_mkfs 2>/dev/null
> > > +_scratch_mkfs_xfs -m crc=0 -dsize=41m,agcount=2 | _filter_mkfs 2>/dev/null
> > 
> > Adding "-m crc=0" to mkfs directly will break testing on older distros
> > where mkfs.xfs doesn't have crc support, e.g. RHEL6
> 
> I know we need to keep older distros working, but right now we need
> to make upstream xfstests work with upstream xfsprogs changes.  Keep
> in mind that we do allow upstream changes for upstream support to
> break older distro support. The responsibility for keeping upstream
> xfstests working on older distros falls to the people running it on
> older distros, not the people who are working on upstream.

Understood, so I'll look into this based on how much work is needed and
how hardly we want it on RHEL6.8/9/etc.

> 
> Of course, people testing older distros always have the option of
> not upgrading xfstests every time a commit is made.... :P

Yes, we use a fixed version of xfstests to test RHEL5, and RHEL6 is
almost reaching that maintaince-only phase too.

> 
> > I haven't figured out a better way, but I was thinking about tweaking
> > _scratch_mkfs_xfs or $MKFS_XFS_PROG based on mkfs.xfs supports crc or
> > not.
> 
> I'd suggest that we should look at filtering this out in
> _scratch_options, based on some environment variable that you set in
> the .config files for the older distros.  e.g.
> 
> 	mkfs)
> +		if [ -n "$MKFS_HAS_NO_META_SUPPORT" ]; then
> +			# strip "-m ..." options
> +			SCRATCH_OPTIONS=`<insert sed magic here>`
> +		fi

I was thinking about _scratch_options too, I'll follow this way and see.

Thanks,
Eryu

      reply	other threads:[~2015-05-07  3:05 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-05-04 22:01 [PATCH 0/4] fstests: xfs utility output has changed Dave Chinner
2015-05-04 22:01 ` [PATCH 1/4] filter: latest mkfs.xfs makes logical sector size noise Dave Chinner
2015-05-04 22:45   ` Eric Sandeen
2015-05-04 22:01 ` [PATCH 2/4] filter: latest xfs_repair is more verbose Dave Chinner
2015-05-04 22:46   ` Eric Sandeen
2015-05-04 22:01 ` [PATCH 3/4] xfs/045: can't change UUID on v5 filesystems Dave Chinner
2015-05-04 22:48   ` Eric Sandeen
2015-05-05  0:51     ` Dave Chinner
2015-05-05  2:04       ` Eric Sandeen
2015-05-04 22:01 ` [PATCH 4/4] filter: inode size output of mkfs.xfs can change Dave Chinner
2015-05-05  2:06   ` Eric Sandeen
2015-05-05  5:34     ` Dave Chinner
2015-05-06  7:19   ` Eryu Guan
2015-05-06 22:56     ` Dave Chinner
2015-05-07  3:05       ` Eryu Guan [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20150507030538.GH1187@dhcp-13-216.nay.redhat.com \
    --to=eguan@redhat.com \
    --cc=david@fromorbit.com \
    --cc=fstests@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.