Linux kernel -stable discussions
 help / color / mirror / Atom feed
From: David Laight <David.Laight@ACULAB.COM>
To: 'Christian Brauner' <brauner@kernel.org>,
	Linus Torvalds <torvalds@linux-foundation.org>
Cc: Jens Axboe <axboe@kernel.dk>, Christoph Hellwig <hch@lst.de>,
	Aleksa Sarai <cyphar@cyphar.com>,
	Al Viro <viro@zeniv.linux.org.uk>,
	Seth Forshee <sforshee@kernel.org>,
	"linux-fsdevel@vger.kernel.org" <linux-fsdevel@vger.kernel.org>,
	"stable@vger.kernel.org" <stable@vger.kernel.org>
Subject: RE: [PATCH] file: always lock position
Date: Wed, 26 Jul 2023 10:31:07 +0000	[thread overview]
Message-ID: <081f95b2428049999cc2c0f55a46075f@AcuMS.aculab.com> (raw)
In-Reply-To: <20230726-antik-abwinken-87647ff63ec8@brauner>

From: Christian Brauner
> Sent: 26 July 2023 09:37
...
> Yes, and to summarize which I tried in my description for the commit.
> The getdents support patchset would have introduced a bug because the
> patchset copied the fdget_pos() file_count(file) > 1 optimization into
> io_uring.
> 
> That works fine as long as the original file descriptor used to register
> the fixed file is kept. The locking will work correctly as
> file_count(file) > 1 and no races are possible neither via getdent calls
> using the original file descriptor nor via io_uring using the fixed file
> or even mixing both.
> 
> But as soon as the original file descriptor is closed the f_count for
> the file drops back to 1 but continues to be usable from io_uring via
> the fixed file. Now the optimization that the patchset wanted to copy
> over would cause bugs as multiple racing getdent requests would be
> possible using the fixed file.

Could the io_uring code grab two references?
That would stop the optimisation without affecting any
normal code paths?

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)

  reply	other threads:[~2023-07-26 10:31 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-07-24 15:00 [PATCH] file: always lock position Christian Brauner
2023-07-24 15:53 ` Linus Torvalds
2023-07-24 16:19   ` Christian Brauner
2023-07-24 16:36     ` Linus Torvalds
2023-07-24 16:51       ` Linus Torvalds
2023-09-02  4:44         ` Al Viro
2023-07-24 17:23       ` Christian Brauner
2023-07-24 17:34         ` Linus Torvalds
2023-07-24 17:46           ` Christian Brauner
2023-07-24 18:01             ` Linus Torvalds
2023-07-24 18:05               ` Jens Axboe
2023-07-24 18:27                 ` Linus Torvalds
2023-07-24 18:48                   ` Christian Brauner
2023-07-24 22:25                     ` Linus Torvalds
2023-07-24 22:56                       ` Jens Axboe
2023-07-25 18:30                         ` Linus Torvalds
2023-07-25 20:41                           ` Jens Axboe
2023-07-25 20:51                             ` Linus Torvalds
2023-07-25 20:58                               ` Jens Axboe
2023-07-26  8:36                               ` Christian Brauner
2023-07-26 10:31                                 ` David Laight [this message]
2023-07-26 12:53                                   ` Christian Brauner
2023-07-26  8:07                           ` Christian Brauner
2023-07-24 16:46   ` Christian Brauner
2023-07-24 16:59     ` Linus Torvalds
2023-07-24 17:18       ` Linus Torvalds
2023-08-03  9:53       ` Mateusz Guzik
2023-08-03 14:15         ` Christian Brauner
2023-08-03 15:17           ` Mateusz Guzik
2023-08-03 15:18             ` Mateusz Guzik
2023-08-03 15:45         ` Linus Torvalds
2023-08-03 17:54           ` Mateusz Guzik
2023-08-03 18:02           ` Christian Brauner
2023-08-03 18:35             ` Linus Torvalds
2023-08-04 13:43               ` Christian Brauner
2023-08-04 13:59                 ` Christoph Hellwig
2023-09-02  3:43               ` Al Viro
     [not found] <20230804-turnverein-helfer-ef07a4d7bbec@brauner>
2023-08-05 11:46 ` Christian Brauner
2023-08-05 18:47   ` Linus Torvalds
2023-08-05 19:46     ` Linus Torvalds
2023-08-06  6:10       ` Christian Brauner
2023-08-06 13:25         ` Christian Brauner
2023-08-06 17:48           ` Linus Torvalds

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=081f95b2428049999cc2c0f55a46075f@AcuMS.aculab.com \
    --to=david.laight@aculab.com \
    --cc=axboe@kernel.dk \
    --cc=brauner@kernel.org \
    --cc=cyphar@cyphar.com \
    --cc=hch@lst.de \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=sforshee@kernel.org \
    --cc=stable@vger.kernel.org \
    --cc=torvalds@linux-foundation.org \
    --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