Received-SPF: pass (google.com: domain of linux-kernel+bounces-182145-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Date: Fri, 17 May 2024 15:31:28 +0200
From: Petr Mladek <pmladek@suse.com>
To: John Ogness <john.ogness@linutronix.de>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>,
	Steven Rostedt <rostedt@goodmis.org>,
	Thomas Gleixner <tglx@linutronix.de>, linux-kernel@vger.kernel.org,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Subject: Re: [PATCH printk v5 06/30] printk: nbcon: Add callbacks to
 synchronize with driver
Message-ID: <ZkdcFxW-e9LVmMd8@pathway.suse.cz>
References: <20240502213839.376636-1-john.ogness@linutronix.de>
 <20240502213839.376636-7-john.ogness@linutronix.de>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240502213839.376636-7-john.ogness@linutronix.de>

Hi,

On Thu 2024-05-02 23:44:15, John Ogness wrote:
> Console drivers typically must deal with access to the hardware
> via user input/output (such as an interactive login shell) and
> output of kernel messages via printk() calls. To provide the
> necessary synchronization, usually some driver-specific locking
> mechanism is used (for example, the port spinlock for uart
> serial consoles).
>
> Until now, usage of this driver-specific locking has been hidden
> from the printk-subsystem and implemented within the various
> console callbacks. However, for nbcon consoles, it is necessary
> that the printk-subsystem uses the driver-specific locking so
> that nbcon console ownership can be acquired _after_ the
> driver-specific locking has succeeded. This allows for lock
> contention to exist on the more context-friendly driver-specific
> locking rather than nbcon console ownership (for non-emergency
> and non-panic cases).

Honestly, the above paragraph is kind of a puzzle. My first
understanding was that the driver-specific lock will always
need to be taken before acquiring the nbcon console ownership.

I believe that it is actually correct. But the right meaning
is hidden in the wording. I had to read it many times to get
it correctly and I knew what I was looking for.

I have to admit that providing a good explanation is probably
very hard. A proof is the long discussion in v4, see
https://lore.kernel.org/r/20240402221129.2613843-7-john.ogness@linutronix.de

BTW: I wonder if you use AI for generating the commit message.
     My experience is that AI produces longer fancy sentences
     which might be good for a novel but they sometimes hide
     the important details.


> Require nbcon consoles to implement two new callbacks
> (device_lock(), device_unlock()) that will use whatever
> synchronization mechanism the driver is using for itself.
> 
> Signed-off-by: John Ogness <john.ogness@linutronix.de>

My attempt of a more strightforwward explanation:

<proposal>
Console drivers typically must deal with access to the hardware
via user input/output (such as an interactive login shell) and
output of kernel messages via printk() calls. To provide the
necessary synchronization, usually some driver-specific locking
mechanism is used (for example, the port spinlock for uart
serial consoles).

Until now, usage of this driver-specific locking has been hidden
from the printk-subsystem and implemented within the various
console callbacks. However, nbcon consoles would need to use it
even in the generic code.

Add device_lock() and device_unlock() callback which will need
to get implemented by nbcon consoles.

The callbacks will use whatever synchronization mechanism the driver
is using for itself. The minimum requirement is to prevent CPU
migration. It would allow a context friendly acquiring of nbcon
console ownership in non-emergency and non-panic context.
</proposal>

That said, I could live even with the original commit message.
Eitherway:

Reviewed-by: Petr Mladek <pmladek@suse.com>

Best Regards,
Petr