linux-kernel - Re: [PATCH] serial: ma35d1: Fix resource leaks on driver removal

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] [day] [month] [year] [list]

Message-ID: <2025103115-humped-body-d522@gregkh>
Date: Fri, 31 Oct 2025 17:31:32 +0100
From: Greg KH <gregkh@...uxfoundation.org>
To: Haotian Zhang <vulab@...as.ac.cn>
Cc: jirislaby@...nel.org, ychuang3@...oton.com, schung@...oton.com,
	linux-serial@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] serial: ma35d1: Fix resource leaks on driver removal

On Wed, Oct 29, 2025 at 04:00:39PM +0800, Haotian Zhang wrote:
> The driver fails to release resources acquired during probe. It calls
> of_clk_get() to obtain the clock reference but never calls clk_put()
> in the remove path, leaking the clock reference count. Similarly, it
> calls ioremap() to map UART registers but never calls iounmap() during
> removal, leaking the I/O memory mapping.
> 
> Switch to devm_clk_get() for automatic clock resource management and
> add iounmap() call in ma35d1serial_remove() to properly release the
> memory mapping.
> 
> Fixes: 930cbf92db01 ("tty: serial: Add Nuvoton ma35d1 serial driver support")
> Signed-off-by: Haotian Zhang <vulab@...as.ac.cn>
> ---
>  drivers/tty/serial/ma35d1_serial.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/tty/serial/ma35d1_serial.c b/drivers/tty/serial/ma35d1_serial.c
> index 285b0fe41a86..37e9f6166290 100644
> --- a/drivers/tty/serial/ma35d1_serial.c
> +++ b/drivers/tty/serial/ma35d1_serial.c
> @@ -711,7 +711,7 @@ static int ma35d1serial_probe(struct platform_device *pdev)
>  
>  	spin_lock_init(&up->port.lock);
>  
> -	up->clk = of_clk_get(pdev->dev.of_node, 0);
> +	up->clk = devm_clk_get(&pdev->dev, NULL);
>  	if (IS_ERR(up->clk)) {
>  		ret = PTR_ERR(up->clk);
>  		dev_err(&pdev->dev, "failed to get core clk: %d\n", ret);
> @@ -762,6 +762,7 @@ static void ma35d1serial_remove(struct platform_device *dev)
>  
>  	uart_remove_one_port(&ma35d1serial_reg, port);
>  	clk_disable_unprepare(up->clk);
> +	iounmap(up->port.membase);

You are doing two different things in this patch, shouldn't this be 2
different commits?

thanks,

greg k-h