[PATCH] w35und: fix registration with wlan stack

[PATCH] w35und: fix registration with wlan stack

[Pavel Machek]

Initialize few more fields in wireless device structure so that
wireless core actually accepts our registration.

Signed-off-by: Pavel Machek <pavel@suse.cz>

---
commit 87290671d60a4f0e734f389a266b13d71c275ce4
tree c8e312a3ef0b0d250c1e605ce86bff94c0254847
parent e6de9be58f118cba6ccfef28830db701b8cc9f46
author Pavel <pavel@suse.cz> Fri, 30 Jan 2009 09:08:36 +0100
committer Pavel <pavel@suse.cz> Fri, 30 Jan 2009 09:08:36 +0100

 drivers/staging/winbond/wbusb.c |    4 +++-
 1 files changed, 3 insertions(+), 1 deletions(-)

diff --git a/drivers/staging/winbond/wbusb.c b/drivers/staging/winbond/wbusb.c
index b003f9a..49f1bf0 100644
--- a/drivers/staging/winbond/wbusb.c
+++ b/drivers/staging/winbond/wbusb.c
@@ -369,9 +369,11 @@ static int wb35_probe(struct usb_interface *intf, const struct usb_device_id *id
 	}
 
 	dev->extra_tx_headroom = 12;	/* FIXME */
-	dev->flags = 0;
+	dev->flags = IEEE80211_HW_SIGNAL_UNSPEC;
+	dev->wiphy->interface_modes = BIT(NL80211_IFTYPE_STATION);
 
 	dev->channel_change_time = 1000;
+	dev->max_signal = 100;
 	dev->queues = 1;
 
 	dev->wiphy->bands[IEEE80211_BAND_2GHZ] = &wbsoft_band_2GHz;

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [PATCH] w35und: fix registration with wlan stack

[Pekka Enberg]

On Fri, 2009-01-30 at 10:05 +0100, Pavel Machek wrote:
> Initialize few more fields in wireless device structure so that
> wireless core actually accepts our registration.
> 
> Signed-off-by: Pavel Machek <pavel@suse.cz>

Acked-by: Pekka Enberg <penberg@cs.helsinki.fi>

Looking at the changes, lets see if we can get some more review on this:

> ---
> commit 87290671d60a4f0e734f389a266b13d71c275ce4
> tree c8e312a3ef0b0d250c1e605ce86bff94c0254847
> parent e6de9be58f118cba6ccfef28830db701b8cc9f46
> author Pavel <pavel@suse.cz> Fri, 30 Jan 2009 09:08:36 +0100
> committer Pavel <pavel@suse.cz> Fri, 30 Jan 2009 09:08:36 +0100
> 
>  drivers/staging/winbond/wbusb.c |    4 +++-
>  1 files changed, 3 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/staging/winbond/wbusb.c b/drivers/staging/winbond/wbusb.c
> index b003f9a..49f1bf0 100644
> --- a/drivers/staging/winbond/wbusb.c
> +++ b/drivers/staging/winbond/wbusb.c
> @@ -369,9 +369,11 @@ static int wb35_probe(struct usb_interface *intf, const struct usb_device_id *id
>  	}
>  
>  	dev->extra_tx_headroom = 12;	/* FIXME */
> -	dev->flags = 0;
> +	dev->flags = IEEE80211_HW_SIGNAL_UNSPEC;
> +	dev->wiphy->interface_modes = BIT(NL80211_IFTYPE_STATION);

The wiphy->interface_modes thing is probably fall-out from commit
8dffff216fcac4f79078478085e86d68db64922e ("mac80211: only create default
STA interface if supported"). I'm cc'ing Johannes.

>  
>  	dev->channel_change_time = 1000;
> +	dev->max_signal = 100;

And the IEEE80211_HW_SIGNAL_UNSPEC and ->max_signal from commit
566bfe5a8bcde13188a356f77666f8115813cf31 ("mac80211: use hardware flags
for signal/noise units"). Lets cc Bruno.

>  	dev->queues = 1;
>  
>  	dev->wiphy->bands[IEEE80211_BAND_2GHZ] = &wbsoft_band_2GHz;

Re: [PATCH] w35und: fix registration with wlan stack

[Johannes Berg]

[-- Attachment #1: Type: text/plain, Size: 922 bytes --]

On Fri, 2009-01-30 at 11:13 +0200, Pekka Enberg wrote:


> >  	dev->extra_tx_headroom = 12;	/* FIXME */
> > -	dev->flags = 0;
> > +	dev->flags = IEEE80211_HW_SIGNAL_UNSPEC;
> > +	dev->wiphy->interface_modes = BIT(NL80211_IFTYPE_STATION);
> 
> The wiphy->interface_modes thing is probably fall-out from commit
> 8dffff216fcac4f79078478085e86d68db64922e ("mac80211: only create default
> STA interface if supported"). I'm cc'ing Johannes.

Yes. That looks reasonable, though it means you'll only support STA, not
ad-hoc.

> >  
> >  	dev->channel_change_time = 1000;
> > +	dev->max_signal = 100;
> 
> And the IEEE80211_HW_SIGNAL_UNSPEC and ->max_signal from commit
> 566bfe5a8bcde13188a356f77666f8115813cf31 ("mac80211: use hardware flags
> for signal/noise units"). Lets cc Bruno.

Seems reasonable, though does it would be good to see if it can do DBM,
which we very very much prefer.

johannes

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH] w35und: fix registration with wlan stack

[Pavel Machek]

> On Fri, 2009-01-30 at 11:13 +0200, Pekka Enberg wrote:
> 
> 
> > >  	dev->extra_tx_headroom = 12;	/* FIXME */
> > > -	dev->flags = 0;
> > > +	dev->flags = IEEE80211_HW_SIGNAL_UNSPEC;
> > > +	dev->wiphy->interface_modes = BIT(NL80211_IFTYPE_STATION);
> > 
> > The wiphy->interface_modes thing is probably fall-out from commit
> > 8dffff216fcac4f79078478085e86d68db64922e ("mac80211: only create default
> > STA interface if supported"). I'm cc'ing Johannes.
> 
> Yes. That looks reasonable, though it means you'll only support STA, not
> ad-hoc.

Ok. The driver is very basic for now (can only do channel 1), so we
certainly have bigger problems than ad-hoc.

> > >  	dev->channel_change_time = 1000;
> > > +	dev->max_signal = 100;
> > 
> > And the IEEE80211_HW_SIGNAL_UNSPEC and ->max_signal from commit
> > 566bfe5a8bcde13188a356f77666f8115813cf31 ("mac80211: use hardware flags
> > for signal/noise units"). Lets cc Bruno.
> 
> Seems reasonable, though does it would be good to see if it can do DBM,
> which we very very much prefer.

Well, currenlty it reports all zeros...
									Pavel

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [PATCH] w35und: fix registration with wlan stack

[Johannes Berg]

[-- Attachment #1: Type: text/plain, Size: 607 bytes --]

On Fri, 2009-01-30 at 10:38 +0100, Pavel Machek wrote:

> > > >  	dev->channel_change_time = 1000;
> > > > +	dev->max_signal = 100;
> > > 
> > > And the IEEE80211_HW_SIGNAL_UNSPEC and ->max_signal from commit
> > > 566bfe5a8bcde13188a356f77666f8115813cf31 ("mac80211: use hardware flags
> > > for signal/noise units"). Lets cc Bruno.
> > 
> > Seems reasonable, though does it would be good to see if it can do DBM,
> > which we very very much prefer.
> 
> Well, currenlty it reports all zeros...

In that case you shouldn't do this patch to avoid userspace getting
bogus values.

johannes

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]