svn commit: r286441 - in head: share/man/man4 sys/contrib/dev/iwm sys/dev/iwm sys/modules/iwm sys/modules/iwmfw sys/modules/iwmfw/iwm3160fw sys/modules/iwmfw/iwm7260fw sys/modules/iwmfw/iwm7265fw

[Gleb Smirnoff]

On Sat, Aug 08, 2015 at 06:06:49AM +0000, Rui Paulo wrote:
R> Author: rpaulo
R> Date: Sat Aug  8 06:06:48 2015
R> New Revision: 286441
R> URL: https://svnweb.freebsd.org/changeset/base/286441
R> 
R> Log:
R>   Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265.
R>   
R>   There are still several bugs, but I've been using it for a while now.
R>   Thanks to all the testers and to Adrian for his help with this
R>   driver.
R>   
R>   This driver isn't connected to the build yet, but it will be soon.
R>   
R>   There's no MFC planned because the driver isn't very stable yet.
R>   
R>   Reviewed by:	adrian
R>   Obtained from:	https://github.com/rpaulo/iwm
R>   Tested by:	adrian, gjb, dumbbell (others that I forgot).

Converting the driver to new KPI, I noticed that iwm_media_change()
expects the "parent ifnet" to be its argument:

static int
iwm_media_change(struct ifnet *ifp)
{
        struct iwm_softc *sc = ifp->if_softc;

I believe this is wrong, its argument is "vap ifnet", so the code
should be:

	struct ieee80211vap *vap = ifp->if_softc;
	struct ieee80211com *ic = vap->iv_ic;
	struct wi_softc *sc = ic->ic_softc;

The fact that it didn't panic is probably explained by the fact
that this function was never executed.

P.S. And this is an illustration on how convoluted the current KPI is :)
P.P.S. I plan to be ready with retry of KPI patch today. Including iwm,
and all ath debug fixes.

-- 
Totus tuus, Glebius.