Received-SPF: pass (google.com: domain of linux-kernel+bounces-128416-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
Date: Tue, 2 Apr 2024 13:18:27 -0400
From: Hugo Villeneuve <hugo@hugovil.com>
To: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
 linux-kernel@vger.kernel.org, linux-serial@vger.kernel.org, Jiri Slaby
 <jirislaby@kernel.org>
Message-Id: <20240402131827.fdc429dfb6f62db4d291688f@hugovil.com>
In-Reply-To: <20240402154219.3583679-3-andriy.shevchenko@linux.intel.com>
References: <20240402154219.3583679-1-andriy.shevchenko@linux.intel.com>
	<20240402154219.3583679-3-andriy.shevchenko@linux.intel.com>
Precedence: bulk
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Subject: Re: [PATCH v1 02/16] serial: max3100: Update uart_driver_registered
 on driver removal

On Tue,  2 Apr 2024 18:38:08 +0300
Andy Shevchenko <andriy.shevchenko@linux.intel.com> wrote:

Hi Andy,

> The removal of the last MAX3100 device triggers the removal of
> the driver. However, code doesn't update the respective global
> variable and after insmod — rmmod — insmod cycle the kernel
> oopses:
> 
>   max3100 spi-PRP0001:01: max3100_probe: adding port 0
>   BUG: kernel NULL pointer dereference, address: 0000000000000408
>   ...
>   RIP: 0010:serial_core_register_port+0xa0/0x840
>   ...
>    max3100_probe+0x1b6/0x280 [max3100]
>    spi_probe+0x8d/0xb0
> 
> Update the actual state so next time UART driver will be registered
> again.
> 
> Fixes: 7831d56b0a35 ("tty: MAX3100")
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> ---
>  drivers/tty/serial/max3100.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/tty/serial/max3100.c b/drivers/tty/serial/max3100.c
> index 45022f2909f0..efe26f6d5ebf 100644
> --- a/drivers/tty/serial/max3100.c
> +++ b/drivers/tty/serial/max3100.c
> @@ -841,6 +841,7 @@ static void max3100_remove(struct spi_device *spi)
>  		}
>  	pr_debug("removing max3100 driver\n");
>  	uart_unregister_driver(&max3100_uart_driver);
> +	uart_driver_registered = 0;

At the beginning of the probe function, we have:

-----------------------
if (!uart_driver_registered) {
		uart_driver_registered = 1;
		retval = uart_register_driver(&max3100_uart_driver);
		if (retval) {
			printk(KERN_ERR "Couldn't register max3100 uart
driver\n"); mutex_unlock(&max3100s_lock);
			return retval;
..
-----------------------

If uart_register_driver() fails, uart_driver_registered would still be
true and would it prevent any other subsequent devices from being
properly registered? If yes, then maybe "uart_driver_registered = 1"
should be set only after a sucessfull call to uart_register_driver()?

Hugo.


>  
>  	mutex_unlock(&max3100s_lock);
>  }
> -- 
> 2.43.0.rc1.1.gbec44491f096
> 
>