linux-kernel - Re: [PATCH printk 06/18] printk: Protect [un]register

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, 4 Oct 2022 11:06:16 +0900
From:   Sergey Senozhatsky <senozhatsky@...omium.org>
To:     John Ogness <john.ogness@...utronix.de>
Cc:     Petr Mladek <pmladek@...e.com>,
        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>
Subject: Re: [PATCH printk 06/18] printk: Protect [un]register_console() with
 a mutex

On (22/10/03 21:41), John Ogness wrote:
> A semaphore has been needed because we are performing global locking for
> ambiguous reasons in all possible contexts. We should be using
> fine-grained lock and synchronization mechanisms that are appropriate
> for their used contexts to precisely lock/synchronize exactly what needs
> to be locked/synchronized.
> 
> Your first question is literally, "what is wrong with a BKL".
> 
> And the answer to that is: A BKL is preventing us from optimizing the
> kernel by decoupling unrelated activities.
> 
> > The above proposal suggests that it might be something like:
> >
> > register_console()
> > {
> > 	console_list_lock();
> >
> > 	if (!need_console())
> > 		goto out;
> >
> > 	if (!try_enable_console())
> > 		goto out;
> >
> > 	if (!(con->flags & CON_NOBLK))
> > 		console_lock()
> 
> Why are you taking the console_lock here? The console_list_lock needs to
> replace this responsibility. I realize the RFC and this v1 series does
> not do this. For v2, it will be clear.

So tty/VT code also needs to take list_lock? list_lock does not look
precisely relevant to vt, which has it's own "list" of "struct vc" to
maintain.