Re: [PATCH] staging: wlan-ng: silence incorrect type in argument 1 (different address spaces) warning

[Greg Kroah-Hartman]

On Sat, Apr 24, 2021 at 01:45:29PM +0530, Ashish Kalra wrote:
> On Fri, Apr 23, 2021 at 11:11:05PM -0700, Joe Perches wrote:
> > On Sat, 2021-04-24 at 08:00 +0200, Greg Kroah-Hartman wrote:
> > > On Fri, Apr 23, 2021 at 08:56:19PM +0530, Ashish Kalra wrote:
> > > > On Thu, Apr 22, 2021 at 10:43:13AM +0200, Greg Kroah-Hartman wrote:
> > > > > On Tue, Apr 20, 2021 at 02:31:42PM +0530, Ashish Kalra wrote:
> > > > > > Upon running sparse, "warning: incorrect type in argument 1 (different address spaces)
> > > > > > is brought to notice for this file.let's add correct typecast to make it cleaner and
> > > > > > silence the Sparse warning.
> > []
> > > >  struct p80211ioctl_req {
> > > >         char name[WLAN_DEVNAMELEN_MAX];
> > > > -       caddr_t data;
> > > > +       void __user *data;
> > > > 
> > > > Does this looks ok to you and is there any other check possible if this is ok?
> > > 
> > > Wait, what is "caddr_t"?  Try unwinding that mess first...
> > 
> > Might not be that simple.
> > 
> > include/linux/types.h:typedef __kernel_caddr_t  caddr_t;
> > include/uapi/linux/coda.h:typedef void * caddr_t;
> > include/uapi/asm-generic/posix_types.h:typedef char *           __kernel_caddr_t;
> > 
> > 
> data is part of p80211ioctl_req and is used at two places only inside p80211knetdev_do_ioctl
> it seems both places it will be used as void __user* only
> 
> 	msgbuf = memdup_user(req->data, req->len);
>         
> 	if (result == 0) {
>                 if (copy_to_user
>                     ((void __user *)req->data, msgbuf, req->len)) {
>                         result = -EFAULT;
>                 }
>         }
> 
> Will it still be problem if we change it from char * to void *.?

Why do you want to change it to void *?  Never use a void * unless it
has to point to unknown data.  That does not seem the case here.

> is there any way to check how caller of this function will be using it?

Look at the code to determine this...

thanks,

greg k-h