Re: [PATCH v2] mac80211: VHT peer STA caps

[Johannes Berg <johannes@sipsolutions.net>]

On Wed, 2012-09-26 at 22:52 +0530, Mahesh Palivela wrote:
> On 9/26/2012 5:23 PM, Johannes Berg wrote:
> >> you mean to say some drivers which dont provide its own VHT caps, will
> >> use remote STA VHT caps as its own?
> >> If so I feel not correct posing remote STA caps as local caps.
> >
> > No, no. If you look at HT, then you'll see that if the driver has HT but
> > isn't capable of doing everything the peer STA can do, then mac80211
> > restricts tells the driver
> 
> Sorry. Where is it telling the driver? I couldn't find this part of code 
> in mac80211, could you point me?

Well struct ieee80211_sta is what the driver uses. That's where it's
"telling" the driver, by making the data accessible to the driver. Your
new vht_cap field there is also "telling the driver" that way.


> > I agree that this isn't really necessary and can be rather complex in
> > VHT, but I think you need to document the difference very clearly in the
> > documentation for the ieee80211_sta VHT caps.
> 
> Can this go as different patch? take my v2 patch as framework. More 
> implementation of vht.c at later point of time?

I don't think this makes sense, since we're talking only about a
documentation change to your patch. See:

  * @ht_cap: HT capabilities of this STA; restricted to our own TX capabilities
+ * @vht_cap: VHT capabilities of this STA

The difference is that you don't say "restricted to own", but I think
you should explicitly say that it isn't restricted.

johannes