[PATCH v3] mac80211: Add WME information element for IBSS

[PATCH v3] mac80211: Add WME information element for IBSS

[Bruno Randolf]

Enable WME QoS in IBSS mode by adding a WME information element to beacons and
probe respones and by checking for it and marking stations as WME capable if it
is present.

Signed-off-by: Bruno Randolf <br1@einfach.org>

---
v3: use GFP_ATOMIC and re-organize nesting of "if" statements. allthough in
practice we would probaby always have the rates element i think it's more
correct to code it like this.
---
 net/mac80211/ibss.c |   65 ++++++++++++++++++++++++++++++++++-----------------
 1 files changed, 43 insertions(+), 22 deletions(-)

diff --git a/net/mac80211/ibss.c b/net/mac80211/ibss.c
index 1a3aae5..ff60c02 100644
--- a/net/mac80211/ibss.c
+++ b/net/mac80211/ibss.c
@@ -173,6 +173,19 @@ static void __ieee80211_sta_join_ibss(struct ieee80211_sub_if_data *sdata,
 		memcpy(skb_put(skb, ifibss->ie_len),
 		       ifibss->ie, ifibss->ie_len);
 
+	if (local->hw.queues >= 4) {
+		pos = skb_put(skb, 9);
+		*pos++ = WLAN_EID_VENDOR_SPECIFIC;
+		*pos++ = 7; /* len */
+		*pos++ = 0x00; /* Microsoft OUI 00:50:F2 */
+		*pos++ = 0x50;
+		*pos++ = 0xf2;
+		*pos++ = 2; /* WME */
+		*pos++ = 0; /* WME info */
+		*pos++ = 1; /* WME ver */
+		*pos++ = 0; /* U-APSD no in use */
+	}
+
 	rcu_assign_pointer(ifibss->presp, skb);
 
 	sdata->vif.bss_conf.beacon_int = beacon_int;
@@ -266,37 +279,45 @@ static void ieee80211_rx_bss_info(struct ieee80211_sub_if_data *sdata,
 	if (!channel || channel->flags & IEEE80211_CHAN_DISABLED)
 		return;
 
-	if (sdata->vif.type == NL80211_IFTYPE_ADHOC && elems->supp_rates &&
+	if (sdata->vif.type == NL80211_IFTYPE_ADHOC &&
 	    memcmp(mgmt->bssid, sdata->u.ibss.bssid, ETH_ALEN) == 0) {
-		supp_rates = ieee80211_sta_get_rates(local, elems, band);
 
 		rcu_read_lock();
-
 		sta = sta_info_get(sdata, mgmt->sa);
-		if (sta) {
-			u32 prev_rates;
 
-			prev_rates = sta->sta.supp_rates[band];
-			/* make sure mandatory rates are always added */
-			sta->sta.supp_rates[band] = supp_rates |
-				ieee80211_mandatory_rates(local, band);
+		if (elems->supp_rates) {
+			supp_rates = ieee80211_sta_get_rates(local, elems,
+							     band);
+			if (sta) {
+				u32 prev_rates;
 
-			if (sta->sta.supp_rates[band] != prev_rates) {
+				prev_rates = sta->sta.supp_rates[band];
+				/* make sure mandatory rates are always added */
+				sta->sta.supp_rates[band] = supp_rates |
+					ieee80211_mandatory_rates(local, band);
+
+				if (sta->sta.supp_rates[band] != prev_rates) {
 #ifdef CONFIG_MAC80211_IBSS_DEBUG
-				printk(KERN_DEBUG "%s: updated supp_rates set "
-				    "for %pM based on beacon/probe_response "
-				    "(0x%x -> 0x%x)\n",
-				    sdata->name, sta->sta.addr,
-				    prev_rates, sta->sta.supp_rates[band]);
+					printk(KERN_DEBUG
+						"%s: updated supp_rates set "
+						"for %pM based on beacon"
+						"/probe_resp (0x%x -> 0x%x)\n",
+						sdata->name, sta->sta.addr,
+						prev_rates,
+						sta->sta.supp_rates[band]);
 #endif
-				rate_control_rate_init(sta);
-			}
-			rcu_read_unlock();
-		} else {
-			rcu_read_unlock();
-			ieee80211_ibss_add_sta(sdata, mgmt->bssid, mgmt->sa,
-					       supp_rates, GFP_KERNEL);
+					rate_control_rate_init(sta);
+				}
+			} else
+				sta = ieee80211_ibss_add_sta(sdata, mgmt->bssid,
+						mgmt->sa, supp_rates,
+						GFP_ATOMIC);
 		}
+
+		if (sta && elems->wmm_info)
+			set_sta_flags(sta, WLAN_STA_WME);
+
+		rcu_read_unlock();
 	}
 
 	bss = ieee80211_bss_info_update(local, rx_status, mgmt, len, elems,

Re: [PATCH v3] mac80211: Add WME information element for IBSS

[Johannes Berg]

On Mon, 2010-10-04 at 11:17 +0900, Bruno Randolf wrote:
> Enable WME QoS in IBSS mode by adding a WME information element to beacons and
> probe respones and by checking for it and marking stations as WME capable if it
> is present.
> 
> Signed-off-by: Bruno Randolf <br1@einfach.org>
> 
> ---
> v3: use GFP_ATOMIC and re-organize nesting of "if" statements. allthough in
> practice we would probaby always have the rates element i think it's more
> correct to code it like this.

You're probably right, but shouldn't that be a separate patch? It
doesn't relate to WME at all, afaict.

johannes

Re: [PATCH v3] mac80211: Add WME information element for IBSS

[Bruno Randolf]

On Mon October 4 2010 20:21:40 Johannes Berg wrote:
> On Mon, 2010-10-04 at 11:17 +0900, Bruno Randolf wrote:
> > Enable WME QoS in IBSS mode by adding a WME information element to
> > beacons and probe respones and by checking for it and marking stations
> > as WME capable if it is present.
> > 
> > Signed-off-by: Bruno Randolf <br1@einfach.org>
> > 
> > ---
> > v3: use GFP_ATOMIC and re-organize nesting of "if" statements. allthough
> > in practice we would probaby always have the rates element i think it's
> > more correct to code it like this.
> 
> You're probably right, but shouldn't that be a separate patch? It
> doesn't relate to WME at all, afaict.

it does. before we only looked at elems->supp_rates, so it was o.k. to combine 
that check into the first if (adhoc mode && bssid matches). now we also check 
elems->wmm_info so we should make separate cases for each. as i noted, it 
probably does not matter much in practice, since elems->supp_rates should be 
present in all proper beacons and probe responses.

bruno

Re: [PATCH v3] mac80211: Add WME information element for IBSS

[Johannes Berg]

On Tue, 2010-10-05 at 10:24 +0900, Bruno Randolf wrote:
> On Mon October 4 2010 20:21:40 Johannes Berg wrote:
> > On Mon, 2010-10-04 at 11:17 +0900, Bruno Randolf wrote:
> > > Enable WME QoS in IBSS mode by adding a WME information element to
> > > beacons and probe respones and by checking for it and marking stations
> > > as WME capable if it is present.
> > > 
> > > Signed-off-by: Bruno Randolf <br1@einfach.org>
> > > 
> > > ---
> > > v3: use GFP_ATOMIC and re-organize nesting of "if" statements. allthough
> > > in practice we would probaby always have the rates element i think it's
> > > more correct to code it like this.
> > 
> > You're probably right, but shouldn't that be a separate patch? It
> > doesn't relate to WME at all, afaict.
> 
> it does. before we only looked at elems->supp_rates, so it was o.k. to combine 
> that check into the first if (adhoc mode && bssid matches). now we also check 
> elems->wmm_info so we should make separate cases for each. as i noted, it 
> probably does not matter much in practice, since elems->supp_rates should be 
> present in all proper beacons and probe responses.

Fair enough, looks fine to me.

johannes