From: "Måns Rullgård" <mans@mansr.com>
To: Christian Riesch <christian.riesch@omicron.at>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
<linux-kernel@vger.kernel.org>,
Peter Hurley <peter@hurleysoftware.com>, <stable@vger.kernel.org>
Subject: Re: [PATCH v2] n_tty: Fix read_buf race condition, increment read_head after pushing data
Date: Tue, 11 Nov 2014 13:04:28 +0000 [thread overview]
Message-ID: <yw1x4mu5j3c3.fsf@unicorn.mansr.com> (raw)
In-Reply-To: <a2e126c9-3001-4255-91c9-6311b54988c5@EXC02-ATKLA.omicron.at> (Christian Riesch's message of "Tue, 11 Nov 2014 08:24:17 +0100")
Christian Riesch <christian.riesch@omicron.at> writes:
> Commit 19e2ad6a09f0c06dbca19c98e5f4584269d913dd ("n_tty: Remove overflow
> tests from receive_buf() path") moved the increment of read_head into
> the arguments list of read_buf_addr(). Function calls represent a
> sequence point in C. Therefore read_head is incremented before the
> character c is placed in the buffer. Since the circular read buffer is
> a lock-less design since commit 6d76bd2618535c581f1673047b8341fd291abc67
> ("n_tty: Make N_TTY ldisc receive path lockless"), this creates a race
> condition that leads to communication errors.
>
> This patch modifies the code to increment read_head _after_ the data
> is placed in the buffer and thus fixes the race for non-SMP machines.
> To fix the problem for SMP machines, memory barriers must be added in
> a separate patch.
>
> Signed-off-by: Christian Riesch <christian.riesch@omicron.at>
> Cc: <stable@vger.kernel.org>
> ---
>
> This is version 2 of the patch in [1].
>
> Changes for v2:
> - Rewrote commit message. Since I did not know better, I blamed the compiler
> in v1, but actually the code was wrong. See the discussion in [1].
> - Removed memory barriers. For non-SMP machines they are not required,
> for SMP machines more brainwork and discussions are needed.
>
> Best regards,
> Christian
>
> [1] https://lkml.org/lkml/2014/11/6/216
>
> drivers/tty/n_tty.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/tty/n_tty.c b/drivers/tty/n_tty.c
> index 2e900a9..b09f326 100644
> --- a/drivers/tty/n_tty.c
> +++ b/drivers/tty/n_tty.c
> @@ -321,7 +321,9 @@ static void n_tty_check_unthrottle(struct tty_struct *tty)
>
> static inline void put_tty_queue(unsigned char c, struct n_tty_data *ldata)
> {
> - *read_buf_addr(ldata, ldata->read_head++) = c;
> + *read_buf_addr(ldata, ldata->read_head) = c;
> + /* increment read_head _after_ placing the character in the buffer */
> + ldata->read_head++;
> }
Is that comment really necessary?
--
Måns Rullgård
mans@mansr.com
next prev parent reply other threads:[~2014-11-11 13:04 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-11-11 7:24 [PATCH v2] n_tty: Fix read_buf race condition, increment read_head after pushing data Christian Riesch
2014-11-11 13:04 ` Måns Rullgård [this message]
2014-11-12 7:28 ` Christian Riesch
2014-11-12 11:53 ` Måns Rullgård
2014-11-12 20:03 ` Christian Riesch
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=yw1x4mu5j3c3.fsf@unicorn.mansr.com \
--to=mans@mansr.com \
--cc=christian.riesch@omicron.at \
--cc=gregkh@linuxfoundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=peter@hurleysoftware.com \
--cc=stable@vger.kernel.org \
/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).