From: Jean Tourrilhes (jt_at_bougret.hpl.hp.com)
Date: 2002-06-14 16:50:49 UTC
On Fri, Jun 14, 2002 at 10:46:11AM +0300, Jouni Malinen wrote:
> On Thu, Jun 13, 2002 at 02:57:39PM -0700, Jean Tourrilhes wrote:
> > I went quickly through the Host-AP driver and converted it the
> > the new driver API of Wireless Extension (Wireless Extensions
> > v13). That was a very quick job and I only converted the standard
> > calls (not the private one).
> Thanks! I was going to take a closer look at the new WE API this weekend,
> so this will certainly free some time for taking a closer look
> at private ioctls and wireless events for AP scanning etc.
Good. For scanning, I will be adding a new event type called"string" or "extra" or "priv" which will be a free form string. This will allow you to export all the info about other APs that I've not yet standardised in scanning.
I was also looking at the private extension, and the way to generalise that. I think I've got a sane way to do it. More on this later.
> > To test it, you will just need a kernel later than 2.4.19-pre5
> > and apply the patch to the driver. It should also be backward
> > compatible, but you won't get any benefit there.
> I'll have to do some testing first and I will then merge the changes
> into the driver. However, it looked like the patch would break
> at least permission checking for older kernels (the driver is still
> supposed to work with any Linux 2.2.x/2.4.x version), so this will
> change a bit.
Please find me a kernel that doesn't do permission checking. Those have been around since a very long time in the Wireless Extensions (at least as far as v6, which mean 2.0.X kernel). The only one to be careful about is GETENCODE (and all the private ioctls in the old range).
The code looks like this :
if((IW_IS_SET(cmd)) && (!suser())) return -EPERM; --------------------------------------------
> Jouni Malinen