[PATCH 1/1] b43: do not call ieee80211_unregister_hw if we are not registred

[John W. Linville]

Could someone repost a clean version to linux-wireless at vger.kernel.org?

On Tue, Jun 05, 2012 at 11:49:45AM -0500, Seth Forshee wrote:
> (I don't have access to the original message, so I appologize if any
>  Cc's were dropped. I added some additional Cc's.)
> 
> On Fri, Jun 01, 2012 at 09:37:20AM +0200, Oleksij Rempel wrote:
> > this patch fixes kernel Oops on "rmmod b43" if firmware was not loaded:
> > 
> > BUG: unable to handle kernel NULL pointer dereference at 0000000000000088
> > IP: [<ffffffff8104e988>] drain_workqueue+0x25/0x142
> > PGD 153ac6067 PUD 153b82067 PMD 0
> > Oops: 0000 [#1] SMP
> > 
> > Signed-off-by: Oleksij Rempel <bug-track-M18mAb7Tlt0yCq4wW13eYl6hYfS7NtTn at public.gmane.org>
> 
> This issue is affecting the Ubuntu installer, and I've verified the
> patch fixes the problem on a Mac Mini 5,2.
> 
> Tested-by: Seth Forshee <seth.forshee at canonical.com>
> 
> This issue affects 3.4, probably introduced in 6b6fa58 (b43: Load
> firmware from a work queue and not from the probe routine), so I'd
> suggest a Cc for stable also be added.
> 
> > ---
> >  drivers/net/wireless/b43/b43.h  |    4 ++++
> >  drivers/net/wireless/b43/main.c |   19 ++++++++++++-------
> >  2 files changed, 16 insertions(+), 7 deletions(-)
> > 
> > diff --git a/drivers/net/wireless/b43/b43.h b/drivers/net/wireless/b43/b43.h
> > index 67c13af..c06b6cb 100644
> > --- a/drivers/net/wireless/b43/b43.h
> > +++ b/drivers/net/wireless/b43/b43.h
> > @@ -877,6 +877,10 @@ struct b43_wl {
> >  	 * from the mac80211 subsystem. */
> >  	u16 mac80211_initially_registered_queues;
> >  
> > +	/* Set this if we call ieee80211_register_hw() and check if we call
> > +	 * ieee80211_unregister_hw(). */
> > +	bool hw_registred;
> > +
> >  	/* We can only have one operating interface (802.11 core)
> >  	 * at a time. General information about this interface follows.
> >  	 */
> > diff --git a/drivers/net/wireless/b43/main.c b/drivers/net/wireless/b43/main.c
> > index e4d6dc2..747079b 100644
> > --- a/drivers/net/wireless/b43/main.c
> > +++ b/drivers/net/wireless/b43/main.c
> > @@ -2437,6 +2437,7 @@ start_ieee80211:
> >  	err = ieee80211_register_hw(wl->hw);
> >  	if (err)
> >  		goto err_one_core_detach;
> > +	wl->hw_registred = true;
> >  	b43_leds_register(wl->current_dev);
> >  	goto out;
> >  
> > @@ -5283,6 +5284,7 @@ static struct b43_wl *b43_wireless_init(struct b43_bus_dev *dev)
> >  
> >  	hw->queues = modparam_qos ? B43_QOS_QUEUE_NUM : 1;
> >  	wl->mac80211_initially_registered_queues = hw->queues;
> > +	wl->hw_registred = false;
> >  	hw->max_rates = 2;
> >  	SET_IEEE80211_DEV(hw, dev->dev);
> >  	if (is_valid_ether_addr(sprom->et1mac))
> > @@ -5354,12 +5356,15 @@ static void b43_bcma_remove(struct bcma_device *core)
> >  	 * as the ieee80211 unreg will destroy the workqueue. */
> >  	cancel_work_sync(&wldev->restart_work);
> >  
> > -	/* Restore the queues count before unregistering, because firmware detect
> > -	 * might have modified it. Restoring is important, so the networking
> > -	 * stack can properly free resources. */
> > -	wl->hw->queues = wl->mac80211_initially_registered_queues;
> > -	b43_leds_stop(wldev);
> > -	ieee80211_unregister_hw(wl->hw);
> > +	B43_WARN_ON(!wl);
> > +	if (wl->current_dev == wldev && wl->hw_registred) {
> > +		/* Restore the queues count before unregistering, because firmware detect
> > +		 * might have modified it. Restoring is important, so the networking
> > +		 * stack can properly free resources. */
> > +		wl->hw->queues = wl->mac80211_initially_registered_queues;
> > +		b43_leds_stop(wldev);
> > +		ieee80211_unregister_hw(wl->hw);
> > +	}
> >  
> >  	b43_one_core_detach(wldev->dev);
> >  
> > @@ -5430,7 +5435,7 @@ static void b43_ssb_remove(struct ssb_device *sdev)
> >  	cancel_work_sync(&wldev->restart_work);
> >  
> >  	B43_WARN_ON(!wl);
> > -	if (wl->current_dev == wldev) {
> > +	if (wl->current_dev == wldev && wl->hw_registred) {
> >  		/* Restore the queues count before unregistering, because firmware detect
> >  		 * might have modified it. Restoring is important, so the networking
> >  		 * stack can properly free resources. */
> > -- 
> > 1.7.9.5
> > 
> 

-- 
John W. Linville		Someday the world will need a hero, and you
linville at tuxdriver.com			might be all we have.  Be ready.