Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   John Ogness <john.ogness@linutronix.de>
To:     Petr Mladek <pmladek@suse.com>
Cc:     Sergey Senozhatsky <senozhatsky@chromium.org>,
        Steven Rostedt <rostedt@goodmis.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        linux-kernel@vger.kernel.org
Subject: Re: replay log: Re: [PATCH printk v4 38/39] printk: relieve
 console_lock of list synchronization duties
In-Reply-To: <Y3Oxck0/LAHFLYip@alley>
References: <20221114162932.141883-1-john.ogness@linutronix.de>
 <20221114162932.141883-39-john.ogness@linutronix.de>
 <Y3Oxck0/LAHFLYip@alley>
Date:   Tue, 15 Nov 2022 17:47:12 +0106
Message-ID: <87tu30maqf.fsf@jogness.linutronix.de>
MIME-Version: 1.0
Content-Type: text/plain
Precedence: bulk

On 2022-11-15, Petr Mladek <pmladek@suse.com> wrote:
> On Mon 2022-11-14 17:35:31, John Ogness wrote:
>> --- a/kernel/printk/printk.c
>> +++ b/kernel/printk/printk.c
>> @@ -3334,6 +3330,11 @@ void register_console(struct console *newcon)
>>  		 * boot console that is the furthest behind.
>>  		 */
>>  		if (bootcon_registered && !keep_bootcon) {
>> +			/*
>> +			 * Hold the console_lock to guarantee safe access to
>> +			 * console->seq.
>> +			 */
>> +			console_lock();
>>  			for_each_console(con) {
>>  				if ((con->flags & CON_BOOT) &&
>>  				    (con->flags & CON_ENABLED) &&
>> @@ -3341,6 +3342,7 @@ void register_console(struct console *newcon)
>>  					newcon->seq = con->seq;
>>  				}
>>  			}
>> +			console_unlock();
>
> Thinking more about it. This console_unlock() will actually cause
> flushing the boot consoles. A solution would be to call
> console_flush_all() here.

console_flush_all() requires the console_lock, so I don't think it would
be different.

The correct solution would be to recognize if nextcon is taking over a
bootcon. If yes, that bootcon could be unregistered right here with
unregister_console_locked() and then seq for nextcon set appropriately
to perfectly take over.

But we will need to think about how we could recognize the same
device. I was thinking about if consoles hat some attribute showing
their io-membase or something so that it could be clear that the two are
the same hardware.

> And we could/should solve this in a separate patch. This code was not
> locked before. It is a corner case. It could be solved later.

Agreed.

John