From: Christoph Hellwig <hch@infradead.org>
To: Jan Kara <jack@suse.cz>
Cc: Ritesh Harjani <ritesh.list@gmail.com>,
linux-fsdevel@vger.kernel.org, linux-mm@kvack.org,
lsf-pc@lists.linux-foundation.org
Subject: Re: [RFCv1][WIP] ext2: Move direct-io to use iomap
Date: Wed, 29 Mar 2023 17:02:05 -0700 [thread overview]
Message-ID: <ZCTRfRVK7XdbovJK@infradead.org> (raw)
In-Reply-To: <20230323113030.ryne2oq27b6cx6xz@quack3>
On Thu, Mar 23, 2023 at 12:30:30PM +0100, Jan Kara wrote:
> >
> > One way which hch is suggesting is to use __iomap_dio_rw() -> unlock
> > inode -> call generic_write_sync(). I haven't yet worked on this part.
>
> So I see two problems with what Christoph suggests:
>
> a) It is unfortunate API design to require trivial (and low maintenance)
> filesystem to do these relatively complex locking games. But this can
> be solved by providing appropriate wrapper for them I guess.
Well, the problem is that this "trivial" file systems have a pretty
broken locking scheme for fsync.
The legacy dio code gets around this by unlocking i_rwsem inside of
__blockdev_direct_IO. Which force a particular and somewhat odd
locking scheme on the callers, and severly limits it what it can do.
> > Are you suggesting to rip of inode_lock from __generic_file_fsync()?
> > Won't it have a much larger implications?
>
> Yes and yes :). But inode writeback already happens from other paths
> without inode_lock so there's hardly any surprise there.
> sync_mapping_buffers() is impossible to "customize" by filesystems and the
> generic code is fine without inode_lock. So I have hard time imagining how
> any filesystem would really depend on inode_lock in this path (famous last
> words ;)).
Not holding the inode lock in ->fsync would solve all of this indeed.
next prev parent reply other threads:[~2023-03-30 0:02 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-01-29 4:46 LSF/MM/BPF 2023 IOMAP conversion status update Luis Chamberlain
2023-01-29 5:06 ` Matthew Wilcox
2023-01-29 5:39 ` Luis Chamberlain
2023-02-08 16:04 ` Jan Kara
2023-02-24 7:01 ` Zhang Yi
2023-02-26 20:16 ` Ritesh Harjani
2023-03-16 14:40 ` [RFCv1][WIP] ext2: Move direct-io to use iomap Ritesh Harjani (IBM)
2023-03-16 15:41 ` Darrick J. Wong
2023-03-20 16:11 ` Ritesh Harjani
2023-03-20 13:15 ` Christoph Hellwig
2023-03-20 17:51 ` Jan Kara
2023-03-22 6:34 ` Ritesh Harjani
2023-03-23 11:30 ` Jan Kara
2023-03-23 13:19 ` Ritesh Harjani
2023-03-30 0:02 ` Christoph Hellwig [this message]
2023-02-27 19:26 ` LSF/MM/BPF 2023 IOMAP conversion status update Darrick J. Wong
2023-02-27 21:02 ` Matthew Wilcox
2023-02-27 19:47 ` Darrick J. Wong
2023-02-27 20:24 ` Luis Chamberlain
2023-02-27 19:06 ` Darrick J. Wong
2023-02-27 19:58 ` Luis Chamberlain
2023-03-01 16:59 ` Ritesh Harjani
2023-03-01 17:08 ` Darrick J. Wong
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=ZCTRfRVK7XdbovJK@infradead.org \
--to=hch@infradead.org \
--cc=jack@suse.cz \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=lsf-pc@lists.linux-foundation.org \
--cc=ritesh.list@gmail.com \
/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).