On Wed, 29 Apr 2015, Gujulan Elango, Hari Prasath (H.) wrote:
> The return value of memdup_user is a pointer to errno.Freeing it will cause
> error.Hence set it to NULL before branching to free the pointer.smatch also
> raises the same warning.
Perhaps there is no point to jumping to the label if there is nothing to
do there.
The code contains other strange things, like:
if (status) {
ret = -EFAULT;
goto _r871x_mp_ioctl_hdl_exit;
}
_r871x_mp_ioctl_hdl_exit:
Overall, it would be nice to see if the code could be reorganized so that
there is only a goto if there is something to do to cleanup, so that
return ret would only be executed in an error case, and so that the code
would return 0 directly on success.
julia
> Signed-off-by: Hari Prasath Gujulan Elango <[email protected]>
> ---
> drivers/staging/rtl8712/rtl871x_ioctl_linux.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/staging/rtl8712/rtl871x_ioctl_linux.c
> b/drivers/staging/rtl8712/rtl871x_ioctl_linux.c
> index 42fba3f..9d59be8 100644
> --- a/drivers/staging/rtl8712/rtl871x_ioctl_linux.c
> +++ b/drivers/staging/rtl8712/rtl871x_ioctl_linux.c
> @@ -1915,6 +1915,7 @@ static int r871x_mp_ioctl_hdl(struct net_device *dev,
> pparmbuf = memdup_user(p->pointer, len);
> if (IS_ERR(pparmbuf)) {
> ret = PTR_ERR(pparmbuf);
> + pparmbuf = NULL;
> goto _r871x_mp_ioctl_hdl_exit;
> }
> poidparam = (struct mp_ioctl_param *)pparmbuf;
> --
> 1.9.1
>
_______________________________________________
devel mailing list
[email protected]
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
