From: Ross Zwisler <ross.zwisler@linux.intel.com>
To: Ross Zwisler <ross.zwisler@linux.intel.com>,
Jan Kara <jack@suse.cz>,
linux-kernel@vger.kernel.org, "Theodore Ts'o" <tytso@mit.edu>,
Alexander Viro <viro@zeniv.linux.org.uk>,
Andreas Dilger <adilger.kernel@dilger.ca>,
Andrew Morton <akpm@linux-foundation.org>,
Christoph Hellwig <hch@lst.de>,
Dan Williams <dan.j.williams@intel.com>,
Dave Chinner <david@fromorbit.com>, Jan Kara <jack@suse.com>,
Matthew Wilcox <mawilcox@microsoft.com>,
linux-ext4@vger.kernel.org, linux-fsdevel@vger.kernel.org,
linux-mm@kvack.org, linux-nvdimm@ml01.01.org,
linux-xfs@vger.kernel.org
Subject: Re: [PATCH v4 10/12] dax: add struct iomap based DAX PMD support
Date: Thu, 6 Oct 2016 20:58:33 -0600 [thread overview]
Message-ID: <20161007025833.GA2934@linux.intel.com> (raw)
In-Reply-To: <20161006213424.GA4569@linux.intel.com>
On Thu, Oct 06, 2016 at 03:34:24PM -0600, Ross Zwisler wrote:
> Interesting - adding iomap_end() calls to the DAX PTE fault handler causes an
> AA deadlock because we try and retake ei->dax_sem. We take dax_sem in
> ext2_dax_fault() before calling into the DAX code, then if we end up going
> through the error path in ext2_iomap_end(), we call
> ext2_write_failed()
> ext2_truncate_blocks()
> dax_sem_down_write()
>
> Where we try and take dax_sem again. This error path is really only valid for
> I/O operations, but we happen to call it for page faults because 'written' in
> ext2_iomap_end() is just 0.
>
> So...how should we handle this? A few ideas:
>
> 1) Just continue to omit the calls to iomap_end() in the DAX page fault
> handlers for now, and add them when there is useful work to be done in one of
> the filesystems.
>
> 2) Add an IOMAP_FAULT flag to the flags passed into iomap_begin() and
> iomap_end() so make it explicit that we are calling as part of a fault handler
> and not an I/O operation, and use this to adjust the error handling in
> ext2_iomap_end().
>
> 3) Just work around the existing error handling in ext2_iomap_end() by either
> unsetting IOMAP_WRITE or by setting 'written' to the size of the fault.
>
> For #2 or #3, probably add a comment explaining the deadlock and why we need
> to never call ext2_write_failed() while handling a page fault.
>
> Thoughts?
Never mind, #3 it is, I think it was just a plain bug to call iomap_end() with
'length' != 'written'.
next prev parent reply other threads:[~2016-10-07 2:58 UTC|newest]
Thread overview: 46+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-09-29 22:49 [PATCH v4 00/12] re-enable DAX PMD support Ross Zwisler
2016-09-29 22:49 ` [PATCH v4 01/12] ext4: allow DAX writeback for hole punch Ross Zwisler
2016-09-29 22:49 ` [PATCH v4 02/12] ext4: tell DAX the size of allocation holes Ross Zwisler
2016-09-29 22:49 ` [PATCH v4 03/12] dax: remove buffer_size_valid() Ross Zwisler
2016-09-30 8:49 ` Christoph Hellwig
2016-09-29 22:49 ` [PATCH v4 04/12] ext2: remove support for DAX PMD faults Ross Zwisler
2016-09-30 8:49 ` Christoph Hellwig
2016-10-03 9:35 ` Jan Kara
2016-09-29 22:49 ` [PATCH v4 05/12] dax: make 'wait_table' global variable static Ross Zwisler
2016-09-30 8:50 ` Christoph Hellwig
2016-10-03 9:36 ` Jan Kara
2016-09-29 22:49 ` [PATCH v4 06/12] dax: consistent variable naming for DAX entries Ross Zwisler
2016-09-30 8:50 ` Christoph Hellwig
2016-10-03 9:37 ` Jan Kara
2016-09-29 22:49 ` [PATCH v4 07/12] dax: coordinate locking for offsets in PMD range Ross Zwisler
2016-09-30 9:44 ` Christoph Hellwig
2016-10-03 9:55 ` Jan Kara
2016-10-03 18:40 ` Ross Zwisler
2016-09-29 22:49 ` [PATCH v4 08/12] dax: remove dax_pmd_fault() Ross Zwisler
2016-09-30 8:50 ` Christoph Hellwig
2016-10-03 9:56 ` Jan Kara
2016-09-29 22:49 ` [PATCH v4 09/12] dax: correct dax iomap code namespace Ross Zwisler
2016-09-30 8:51 ` Christoph Hellwig
2016-10-03 9:57 ` Jan Kara
2016-09-29 22:49 ` [PATCH v4 10/12] dax: add struct iomap based DAX PMD support Ross Zwisler
2016-09-30 9:56 ` Christoph Hellwig
2016-10-03 21:16 ` Ross Zwisler
2016-10-03 10:59 ` Jan Kara
2016-10-03 16:37 ` Christoph Hellwig
2016-10-03 21:05 ` Ross Zwisler
2016-10-04 5:55 ` Jan Kara
2016-10-04 15:39 ` Ross Zwisler
2016-10-05 5:50 ` Jan Kara
2016-10-06 21:34 ` Ross Zwisler
2016-10-07 2:58 ` Ross Zwisler [this message]
2016-10-07 7:24 ` Jan Kara
2016-09-29 22:49 ` [PATCH v4 11/12] xfs: use struct iomap based DAX PMD fault path Ross Zwisler
2016-09-29 22:49 ` [PATCH v4 12/12] dax: remove "depends on BROKEN" from FS_DAX_PMD Ross Zwisler
2016-09-29 23:43 ` [PATCH v4 00/12] re-enable DAX PMD support Dave Chinner
2016-09-30 3:03 ` Ross Zwisler
2016-09-30 4:00 ` Darrick J. Wong
2016-10-03 18:54 ` Ross Zwisler
2016-09-30 6:48 ` Dave Chinner
2016-10-03 21:11 ` Ross Zwisler
2016-10-03 23:05 ` Ross Zwisler
2016-09-30 11:46 ` Christoph Hellwig
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=20161007025833.GA2934@linux.intel.com \
--to=ross.zwisler@linux.intel.com \
--cc=adilger.kernel@dilger.ca \
--cc=akpm@linux-foundation.org \
--cc=dan.j.williams@intel.com \
--cc=david@fromorbit.com \
--cc=hch@lst.de \
--cc=jack@suse.com \
--cc=jack@suse.cz \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-nvdimm@ml01.01.org \
--cc=linux-xfs@vger.kernel.org \
--cc=mawilcox@microsoft.com \
--cc=tytso@mit.edu \
--cc=viro@zeniv.linux.org.uk \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).