[PATCH 1/2] p2p: disable 11b rates only on p2p interface creation

Rajkumar Manoharan rmanohar at qca.qualcomm.com
Thu Oct 20 08:45:28 EDT 2011


On Thu, Oct 20, 2011 at 12:48:28PM +0200, Johannes Berg wrote:
> On Thu, 2011-10-20 at 16:15 +0530, Rajkumar Manoharan wrote:
> > On Thu, Oct 20, 2011 at 11:33:02AM +0200, Johannes Berg wrote:
> > > On Thu, 2011-10-20 at 14:51 +0530, Rajkumar Manoharan wrote:
> > > 
> > > > > We have the no-CCK thing in the kernel, did you ever make corresponding
> > > > > patches for wpa_supplicant?
> > > 
> > > > Sorry for the delayed response. I already posted the corresponding patch
> > > > for mgmt frame and scan trigger to make use of no-CCK attribute. But not
> > > > yet merged into hostap tree.
> > > > 
> > > > "nl80211: Add support to disable CCK rate for p2p probe and frames".
> > > 
> > > Oh, I missed that, thanks. But I'm a bit confused which patches should
> > > be applied now?
> > >
> > The above mentioned patch alone. because removing disable_11b_rates() from p2p_init
> > might affect firmware based drivers like ath9k_htc. So for the drivers that have
> > rate control in the firmware, we can disable 11b rates at init. Somwhow we need
> > to expose the driver capabilities to wpa_s to identify this. For other drivers
> > 11b rates can be disabled on iface change. Is that fine?
> 
> Even there I'm not convinced -- are those drivers not going to support
> multiple virtual interfaces?
>
ath9k_htc supports milti-vifs (max 4 limited by fw). As rate_control runs in fw,
currently we dont have wmi command to inform no-CCK info to fw. And also per
packet rate selection for mgmt frames is not supported. Due to these limitation,
we can disable 11b rate at init itself for these drivers.

--
Rajkumar


More information about the HostAP mailing list