Re: [PATCH] serial: core: check if uart_get_info succeeds before using

[Greg KH]

On Sun, May 29, 2022 at 09:46:05AM -0400, Tom Rix wrote:
> clang static analysis reports this representative issue
> drivers/tty/serial/serial_core.c:2818:9: warning: 3rd function call argument is an uninitialized value [core.CallAndMessage]
>         return sprintf(buf, "%d\n", tmp.iomem_reg_shift);
>                ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> uart_get_info() is used the *show() functions.  When uart_get_info() fails, what is reported
> is garbage.  So check if uart_get_info() succeeded.
> 
> Fixes: 4047b37122d1 ("serial: core: Prevent unsafe uart port access, part 1")
> Signed-off-by: Tom Rix <trix@xxxxxxxxxx>
> ---
>  drivers/tty/serial/serial_core.c | 52 ++++++++++++++++++++++++--------
>  1 file changed, 39 insertions(+), 13 deletions(-)
> 
> diff --git a/drivers/tty/serial/serial_core.c b/drivers/tty/serial/serial_core.c
> index 9a85b41caa0a..4160f6711c5d 100644
> --- a/drivers/tty/serial/serial_core.c
> +++ b/drivers/tty/serial/serial_core.c
> @@ -2690,7 +2690,9 @@ static ssize_t uartclk_show(struct device *dev,
>  	struct serial_struct tmp;
>  	struct tty_port *port = dev_get_drvdata(dev);
>  
> -	uart_get_info(port, &tmp);
> +	if (uart_get_info(port, &tmp))
> +		return 0;

As Andy pointed out, this is an error, don't tell userspace that all
went well and yet you returned no data?  That will just confuse it.

thanks,

greg k-h