linux-kernel - Re: [PATCH v2] tty: ttynull: implement console write

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 for Android: free password hash cracker in your pocket

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

Date:   Wed, 15 Feb 2023 16:24:23 +0106
From:   John Ogness <john.ogness@...utronix.de>
To:     Petr Mladek <pmladek@...e.com>,
        Vincent Whitchurch <vincent.whitchurch@...s.com>
Cc:     Michael Thalmeier <michael.thalmeier@...e.at>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Jiri Slaby <jirislaby@...nel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] tty: ttynull: implement console write

On 2023-02-15, Petr Mladek <pmladek@...e.com> wrote:
> That said, the current code is error-prone. The check for non-NULL
> con->write is too far from the caller.

con->write is supposed to be immutable after registration, so having the
check "far from the caller" is not a real danger.

console_emit_next_record() is the toplevel function responsible for
printing on a particular console so I think it is appropriate that the
check is made when determining if this function should be called. I also
think console_is_usable() is the proper place for the NULL check to
reside since that is the function that determines if printing is
allowed.

> I would prefer to make it more safe. For example, I would prevent
> registration of consoles without con->write callback in the first
> place. It would require, to implement the empty write() callback
> for ttynull console as done by this patch.

I would prefer that we do not start encouraging dummy implementations.
If you insist on con->write having _some_ value other than NULL, then we
could define some macro with a special value (CONSOLE_NO_WRITE). But
then we have to check that value. IMHO, allowing NULL is not an issue.

John Ogness