Linux Serial subsystem development
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Igor Putko" <igorpetindev@gmail.com>
Cc: linux-serial@vger.kernel.org
Subject: Re: [PATCH] serial: core: shut down initialized port on removal
Date: Thu, 24 Sep 2026 19:39:04 +0000	[thread overview]
Message-ID: <20260924193904.984FD1F000FF@smtp.kernel.org> (raw)
In-Reply-To: <20260924192224.3175-1-igorpetindev@gmail.com>

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] tty: data race on port->console in serial_core_remove_one_port()
--

commit ff681977bacb42b18c6925085fd5fa965af4c283
Author: Igor Putko <igorpetindev@gmail.com>

serial: core: shut down initialized port on removal

This commit addresses a use-after-free read during IRQ setup by ensuring
initialized console ports are properly shut down before removal. It clears
the console flag and calls uart_shutdown() under the port mutex.

> diff --git a/drivers/tty/serial/serial_core.c b/drivers/tty/serial/serial_core.c
> index 95774b0f1484b..030735da6b8c5 100644
> --- a/drivers/tty/serial/serial_core.c
> +++ b/drivers/tty/serial/serial_core.c
> @@ -3210,8 +3210,15 @@ static void serial_core_remove_one_port(struct uart_driver *drv,
>  	/*
>  	 * If the port is used as a console, unregister it
>  	 */
> -	if (uart_console(uport))
> +	if (uart_console(uport)) {
>  		unregister_console(uport->cons);
> +		port->console = false;

[Severity: Medium]
Could this assignment cause a data race on the port->console bitfield?

If a user process holds the serial console open while the underlying device
is removed, serial_core_remove_one_port() will synchronously hang up the
tty by calling tty_port_tty_vhangup().

If the user process then calls close(), the syscall path will execute
tty_release() -> uart_close() -> tty_port_close() -> tty_port_shutdown().

In tty_port_shutdown(), the code acquires port->mutex and evaluates
port->console. Concurrently, the assignment above happens without
holding port->mutex.

While this might be practically benign at the machine level since byte stores
are usually atomic, concurrent read/write access without synchronization is
undefined behavior in C and can trigger KCSAN warnings.

> +	}
> +
> +	guard(mutex)(&port->mutex);
> +
> +	if (tty_port_initialized(port))
> +		uart_shutdown(NULL, state);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260924192224.3175-1-igorpetindev@gmail.com?part=1

  reply	other threads:[~2026-09-24 19:39 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-24 19:22 [PATCH] serial: core: shut down initialized port on removal Igor Putko
2026-09-24 19:39 ` sashiko-bot [this message]
2026-09-25  5:34 ` Greg Kroah-Hartman
2026-09-26 13:05 ` IgorpetinDev
2026-10-01  8:58   ` Greg KH
2026-10-03 16:33 ` [PATCH v2] " Igor Putko
2026-10-03 17:56   ` sashiko-bot

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=20260924193904.984FD1F000FF@smtp.kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=igorpetindev@gmail.com \
    --cc=linux-serial@vger.kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    /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