Re: [PATCH] rtl8723bs/ioctl_linux: Add a security check to copy_from_user()

[Dan Carpenter]

On Sun, Dec 23, 2018 at 06:13:02PM -0600, Aditya Pakki wrote:
> diff --git a/drivers/staging/rtl8723bs/os_dep/ioctl_linux.c b/drivers/staging/rtl8723bs/os_dep/ioctl_linux.c
> index b8631baf128d..9992caa8c839 100644
> --- a/drivers/staging/rtl8723bs/os_dep/ioctl_linux.c
> +++ b/drivers/staging/rtl8723bs/os_dep/ioctl_linux.c
> @@ -2577,14 +2577,19 @@ static int rtw_wps_start(struct net_device *dev,
>  	struct adapter *padapter = (struct adapter *)rtw_netdev_priv(dev);
>  	struct iw_point *pdata = &wrqu->data;
>  	u32   u32wps_start = 0;
> -        unsigned int uintRet = 0;
>  
>  	if ((true == padapter->bDriverStopped) ||(true ==padapter->bSurpriseRemoved) || (NULL == pdata)) {
>  		ret = -EINVAL;
>  		goto exit;
>  	}
>  
> -	uintRet = copy_from_user((void*)&u32wps_start, pdata->pointer, 4);
> +	ret = copy_from_user((void *)&u32wps_start, pdata->pointer, 4);
> +
> +	if (ret) {
> +		ret = -EINVAL;
> +		goto exit;

Good eye for spotting this bug.  :)

Really this function is not useful though so we should just delete it.
All the CONFIG_INTEL_WIDI stuff is dead code.

Also if copy_from_user() the correct error code is -EFAULT.  And we
would normally write it like this:

	if (copy_from_user((void *)&u32wps_start, pdata->pointer, 4)) {
		ret = -EFAULT;
		goto exit;
	}

But in this case, since this is dead code we should just delete
rtw_wps_start() and put a NULL in the rtw_private_handler[] array where
it used to be.

regards,
dan carpenter