linux-kernel - Re: [PATCH printk 07/18] printk: Convert console list walks for readers to list lock

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 27 Sep 2022 16:07:09 +0200
From:   Petr Mladek <pmladek@...e.com>
To:     John Ogness <john.ogness@...utronix.de>
Cc:     Sergey Senozhatsky <senozhatsky@...omium.org>,
        Steven Rostedt <rostedt@...dmis.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        linux-kernel@...r.kernel.org,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Jiri Slaby <jirislaby@...nel.org>,
        linux-fsdevel@...r.kernel.org
Subject: Re: [PATCH printk 07/18] printk: Convert console list walks for
 readers to list lock

On Sat 2022-09-24 02:10:43, John Ogness wrote:
> From: Thomas Gleixner <tglx@...utronix.de>
> 
> Facilities which expose console information to sysfs or procfs can use the
> new list protection to keep the list stable. No need to hold console lock.
>
>  drivers/tty/tty_io.c   | 6 +++---
>  fs/proc/consoles.c     | 6 +++---
>  kernel/printk/printk.c | 8 ++++----

As described in the review of the 6th patch, the semantic of
the list_lock (module_mutex) is not well defined from my POV.
I would prefer to keep only one global console lock.

That said, the procf and sysfs interface is read-only. It seems
to be safe to show the info under the new console_srcu read lock.

On the other hand, console_device() should see the console
list in a consistent state. The first console with tty console->driver
should have the CON_CONSDEV flag set. Alternatively, we could
manipulate the list and the flag a safe way from the SRCU POV
but it is not worth it. So, I would keep console_lock()
in console_device() for now.

Best Regards,
Petr