From mboxrd@z Thu Jan  1 00:00:00 1970
From: Matthew Wilcox <willy@infradead.org>
Date: Tue, 16 Jun 2020 17:32:16 -0700
Subject: [Ocfs2-devel] [Cluster-devel] [PATCH v11 16/25] fs: Convert
 mpage_readpages to mpage_readahead
In-Reply-To: <CAHc6FU4m1M7Tv4scX0UxSiVBqkL=Vcw_z-R7SufL8k7Bw=qPOw@mail.gmail.com>
References: <20200414150233.24495-1-willy@infradead.org>
	<20200414150233.24495-17-willy@infradead.org>
	<CAHc6FU4m1M7Tv4scX0UxSiVBqkL=Vcw_z-R7SufL8k7Bw=qPOw@mail.gmail.com>
Message-ID: <20200617003216.GC8681@bombadil.infradead.org>
List-Id: <ocfs2-devel.oss.oracle.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: Andreas Gruenbacher <agruenba@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>, linux-xfs@vger.kernel.org, Junxiao Bi <junxiao.bi@oracle.com>, William Kucharski <william.kucharski@oracle.com>, Joseph Qi <joseph.qi@linux.alibaba.com>, John Hubbard <jhubbard@nvidia.com>, LKML <linux-kernel@vger.kernel.org>, linux-f2fs-devel@lists.sourceforge.net, cluster-devel <cluster-devel@redhat.com>, Linux-MM <linux-mm@kvack.org>, ocfs2-devel@oss.oracle.com, linux-fsdevel <linux-fsdevel@vger.kernel.org>, linux-ext4 <linux-ext4@vger.kernel.org>, linux-erofs@lists.ozlabs.org, Christoph Hellwig <hch@lst.de>, linux-btrfs@vger.kernel.org, Steven Whitehouse <swhiteho@redhat.com>, Bob Peterson <rpeterso@redhat.com>

On Wed, Jun 17, 2020 at 12:36:13AM +0200, Andreas Gruenbacher wrote:
> Am Mi., 15. Apr. 2020 um 23:39 Uhr schrieb Matthew Wilcox <willy@infradead.org>:
> > From: "Matthew Wilcox (Oracle)" <willy@infradead.org>
> >
> > Implement the new readahead aop and convert all callers (block_dev,
> > exfat, ext2, fat, gfs2, hpfs, isofs, jfs, nilfs2, ocfs2, omfs, qnx6,
> > reiserfs & udf).  The callers are all trivial except for GFS2 & OCFS2.
> 
> This patch leads to an ABBA deadlock in xfstest generic/095 on gfs2.
> 
> Our lock hierarchy is such that the inode cluster lock ("inode glock")
> for an inode needs to be taken before any page locks in that inode's
> address space.

How does that work for ...

writepage:              yes, unlocks (see below)
readpage:               yes, unlocks
invalidatepage:         yes
releasepage:            yes
freepage:               yes
isolate_page:           yes
migratepage:            yes (both)
putback_page:           yes
launder_page:           yes
is_partially_uptodate:  yes
error_remove_page:      yes

Is there a reason that you don't take the glock in the higher level
ops which are called before readhead gets called?  I'm looking at XFS,
and it takes the xfs_ilock SHARED in xfs_file_buffered_aio_read()
(called from xfs_file_read_iter).

Not that after -rc1 is a great time to be upending the locking model in
a filesystem ... but then, this has been baking in -mm for ten weeks and
the GFS2 mailing list has been on the cc for the patches for five months,
so I don't have a lot of sympathy for this.