Re: drivers/net/can/usb/etas_es58x/es58x_fd.c:174:8: warning: Uninitialized variable: rx_event_msg [uninitvar]

[Marc Kleine-Budde]

On 02.03.2022 17:47:08, kernel test robot wrote:
> tree:   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> head:   575115360652e9920cc56a028a286ebe9bf82694
> commit: c664e2137a27680922d8aeb64fb10313416b254f can: etas_es58x: add support for the ETAS ES58X_FD CAN USB interfaces
> date:   11 months ago
> compiler: powerpc64-linux-gcc (GCC) 11.2.0
> 
> If you fix the issue, kindly add following tag as appropriate
> Reported-by: kernel test robot <yujie.liu@xxxxxxxxx>
> 
> 
> cppcheck possible warnings: (new ones prefixed by >>, may not be real problems)
> 
>    In file included from drivers/net/can/usb/etas_es58x/es58x_fd.c:
> >> drivers/net/can/usb/etas_es58x/es58x_fd.c:174:8: warning: Uninitialized variable: rx_event_msg [uninitvar]
>     ret = es58x_check_msg_len(es58x_dev->dev, *rx_event_msg, msg_len);
>           ^
> 
> vim +174 drivers/net/can/usb/etas_es58x/es58x_fd.c
> 
> c664e2137a2768 Vincent Mailhol 2021-04-10  165
> c664e2137a2768 Vincent Mailhol 2021-04-10  166  static int es58x_fd_rx_event_msg(struct net_device *netdev,
> c664e2137a2768 Vincent Mailhol 2021-04-10  167  				 const struct es58x_fd_urb_cmd *es58x_fd_urb_cmd)
> c664e2137a2768 Vincent Mailhol 2021-04-10  168  {
> c664e2137a2768 Vincent Mailhol 2021-04-10  169  	struct es58x_device *es58x_dev = es58x_priv(netdev)->es58x_dev;
> c664e2137a2768 Vincent Mailhol 2021-04-10  170  	u16 msg_len = get_unaligned_le16(&es58x_fd_urb_cmd->msg_len);
> c664e2137a2768 Vincent Mailhol 2021-04-10 @171  	const struct es58x_fd_rx_event_msg *rx_event_msg;
> c664e2137a2768 Vincent Mailhol 2021-04-10  172  	int ret;
> c664e2137a2768 Vincent Mailhol 2021-04-10  173
> c664e2137a2768 Vincent Mailhol 2021-04-10 @174  	ret = es58x_check_msg_len(es58x_dev->dev, *rx_event_msg, msg_len);
> c664e2137a2768 Vincent Mailhol 2021-04-10  175  	if (ret)
> c664e2137a2768 Vincent Mailhol 2021-04-10  176  		return ret;
> c664e2137a2768 Vincent Mailhol 2021-04-10  177
> c664e2137a2768 Vincent Mailhol 2021-04-10  178  	rx_event_msg = &es58x_fd_urb_cmd->rx_event_msg;
> c664e2137a2768 Vincent Mailhol 2021-04-10  179
> c664e2137a2768 Vincent Mailhol 2021-04-10  180  	return es58x_rx_err_msg(netdev, rx_event_msg->error_code,
> c664e2137a2768 Vincent Mailhol 2021-04-10  181  				rx_event_msg->event_code,
> c664e2137a2768 Vincent Mailhol 2021-04-10  182  				get_unaligned_le64(&rx_event_msg->timestamp));
> c664e2137a2768 Vincent Mailhol 2021-04-10  183  }
> c664e2137a2768 Vincent Mailhol 2021-04-10  184

Thanks for the report.

This looks like a false positive to me, as es58x_check_msg_len() is not
a function, but a macro:

| #define es58x_check_msg_len(dev, msg, actual_len)			\
| 	__es58x_check_msg_len(dev, __stringify(msg),			\
| 			      actual_len, sizeof(msg))

__es58x_check_msg_len() don't use "rx_event_msg" directly, but only a
string representation of it and a "sizeof()".

I think it's possible to assign rx_event_msg before the
es58x_check_msg_len(). I think (hope?) the compiler will not optimize
anything away. :)

regards,
Marc

-- 
Pengutronix e.K.                 | Marc Kleine-Budde           |
Embedded Linux                   | https://www.pengutronix.de  |
Vertretung West/Dortmund         | Phone: +49-231-2826-924     |
Amtsgericht Hildesheim, HRA 2686 | Fax:   +49-5121-206917-5555 |

Attachment: signature.asc
Description: PGP signature