Re: [PATCH] xfstests: Add message indicating btrfs-progs support FST in read-only mode

From: Eryu Guan <eguan@redhat.com>
To: "Gu, Jinxiang" <gujx@cn.fujitsu.com>
Cc: "fstests@vger.kernel.org" <fstests@vger.kernel.org>,
	"linux-btrfs@vger.kernel.org" <linux-btrfs@vger.kernel.org>
Subject: Re: [PATCH] xfstests: Add message indicating btrfs-progs support FST in read-only mode
Date: Thu, 26 Oct 2017 15:34:28 +0800	[thread overview]
Message-ID: <20171026073428.GL3235@eguan.usersys.redhat.com> (raw)
In-Reply-To: <516DDBE5B1D92D42BCF7A2E37F045A5D01479E2127@G08CNEXMBPEKD02.g08.fujitsu.local>

On Thu, Oct 26, 2017 at 07:16:02AM +0000, Gu, Jinxiang wrote:
> Hi,
> 
> > -----Original Message-----
> > From: Eryu Guan [mailto:eguan@redhat.com]
> > Sent: Thursday, October 26, 2017 2:52 PM
> > To: Gu, Jinxiang/顾 金香 <gujx@cn.fujitsu.com>
> > Cc: fstests@vger.kernel.org; linux-btrfs@vger.kernel.org
> > Subject: Re: [PATCH] xfstests: Add message indicating btrfs-progs support FST in read-only mode
> > 
> > On Thu, Oct 26, 2017 at 01:57:46PM +0800, Gu Jinxiang wrote:
> > > From: Gu JinXiang <gujx@cn.fujitsu.com>
> > >
> > > btrfs-progs now support FST in read-only mode, so when space_cache=v2
> > > enabled, this test case will fail.
> > > Add message to help user to understand this status.
> > 
> > Sorry, I don't quite understand the new 'FST' feature. But is it a bug that we want to fix when mounting with space_cache=v2
> > option, or we just couldn't do btrfs-convert in this case? If it's a real bug, I'd say let the test fail as it is, and
> > track bug in tools like bugzilla not comments/messages in the test; if it's the latter case, then just _notrun the test
> > if space_cache=v2 option is specified, e.g.
> > 
> > _exclude_scratch_mount_option "space_cache=v2"
> Thank you for your suggestion.
> As an known behavior when use btrfs-convert when space_cache=v2 enabled, I prefer to modify this case not to be run.
> I all send a new patch for this.

But from what Qu said in his reply, it looks like this is a real bug in
btrfs-convert, so I don't think _notrun is a good idea. I'd leave the
test as it is, because that's a bug and this test finds it.

Thanks,
Eryu