| Message ID | 1436877119-17577-6-git-send-email-me@bobcopeland.com (mailing list archive) |
|---|---|
| State | Accepted |
| Delegated to: | Johannes Berg |
| Headers | show |
On Tue, 2015-07-14 at 08:31 -0400, Bob Copeland wrote: > Instead of using peer link id for AID, generate a new > AID when creating mesh STAs in the kernel peering manager. > This enables smaller TIM elements and more closely follows > the standard, and it also enables mesh to work on drivers > that require a valid AID when the STA is inserted (ath10k > firmware has this requirement, for example). > > In the case of userspace-managed stations, we use the AID > from NL80211_CMD_NEW_STATION. > Applied the remaining 3 patches to mac80211-next (after merging mac80211) johannes -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, 2015-07-17 at 15:48 +0200, Johannes Berg wrote: > On Tue, 2015-07-14 at 08:31 -0400, Bob Copeland wrote: > > Instead of using peer link id for AID, generate a new > > AID when creating mesh STAs in the kernel peering manager. > > This enables smaller TIM elements and more closely follows > > the standard, and it also enables mesh to work on drivers > > that require a valid AID when the STA is inserted (ath10k > > firmware has this requirement, for example). > > > > In the case of userspace-managed stations, we use the AID > > from NL80211_CMD_NEW_STATION. > > > Applied the remaining 3 patches to mac80211-next (after merging > mac80211) > Actually, I'm dropping this one (the 5th) as it broke all the wpa_s mesh test cases. Perhaps wpa_s isn't assigning an AID and mac80211 needs a fallback? johannes -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, 2015-07-17 at 16:48 +0200, Johannes Berg wrote: > On Fri, 2015-07-17 at 15:48 +0200, Johannes Berg wrote: > > On Tue, 2015-07-14 at 08:31 -0400, Bob Copeland wrote: > > > Instead of using peer link id for AID, generate a new > > > AID when creating mesh STAs in the kernel peering manager. > > > This enables smaller TIM elements and more closely follows > > > the standard, and it also enables mesh to work on drivers > > > that require a valid AID when the STA is inserted (ath10k > > > firmware has this requirement, for example). > > > > > > In the case of userspace-managed stations, we use the AID > > > from NL80211_CMD_NEW_STATION. > > > > > Applied the remaining 3 patches to mac80211-next (after merging > > mac80211) > > > > Actually, I'm dropping this one (the 5th) as it broke all the wpa_s > mesh test cases. Perhaps wpa_s isn't assigning an AID and mac80211 > needs a fallback? > Actually, it also fails without any of your patches, so I guess it's some other issue. I'll keep these and sort it out when I'm back. johannes -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, Jul 17, 2015 at 04:50:56PM +0200, Johannes Berg wrote: > > Actually, I'm dropping this one (the 5th) as it broke all the wpa_s > > mesh test cases. Perhaps wpa_s isn't assigning an AID and mac80211 > > needs a fallback? > > > > Actually, it also fails without any of your patches, so I guess it's > some other issue. I'll keep these and sort it out when I'm back. Ok, thanks for the heads-up. I ran them against wpa_s and didn't notice any breakage, but will retest against your tree over the weekend.
diff --git a/net/mac80211/mesh_plink.c b/net/mac80211/mesh_plink.c index 3323413acb77..e12be2e4e8df 100644 --- a/net/mac80211/mesh_plink.c +++ b/net/mac80211/mesh_plink.c @@ -422,20 +422,54 @@ out: spin_unlock_bh(&sta->mesh->plink_lock); } +static int mesh_allocate_aid(struct ieee80211_sub_if_data *sdata) +{ + struct sta_info *sta; + unsigned long *aid_map; + int aid; + + aid_map = kcalloc(BITS_TO_LONGS(IEEE80211_MAX_AID + 1), + sizeof(*aid_map), GFP_KERNEL); + if (!aid_map) + return -ENOMEM; + + /* reserve aid 0 for mcast indication */ + __set_bit(0, aid_map); + + rcu_read_lock(); + list_for_each_entry_rcu(sta, &sdata->local->sta_list, list) + __set_bit(sta->sta.aid, aid_map); + rcu_read_unlock(); + + aid = find_first_zero_bit(aid_map, IEEE80211_MAX_AID + 1); + kfree(aid_map); + + if (aid > IEEE80211_MAX_AID) + return -ENOBUFS; + + return aid; +} + static struct sta_info * __mesh_sta_info_alloc(struct ieee80211_sub_if_data *sdata, u8 *hw_addr) { struct sta_info *sta; + int aid; if (sdata->local->num_sta >= MESH_MAX_PLINKS) return NULL; + aid = mesh_allocate_aid(sdata); + if (aid < 0) + return NULL; + sta = sta_info_alloc(sdata, hw_addr, GFP_KERNEL); if (!sta) return NULL; sta->mesh->plink_state = NL80211_PLINK_LISTEN; sta->sta.wme = true; + sta->sta.aid = aid; sta_info_pre_move_state(sta, IEEE80211_STA_AUTH); sta_info_pre_move_state(sta, IEEE80211_STA_ASSOC); @@ -659,8 +693,6 @@ static u16 mesh_get_new_llid(struct ieee80211_sub_if_data *sdata) do { get_random_bytes(&llid, sizeof(llid)); - /* for mesh PS we still only have the AID range for TIM bits */ - llid = (llid % IEEE80211_MAX_AID) + 1; } while (llid_in_use(sdata, llid)); return llid; @@ -1069,7 +1101,6 @@ mesh_process_plink_frame(struct ieee80211_sub_if_data *sdata, goto unlock_rcu; } sta->mesh->plid = plid; - sta->sta.aid = plid; } else if (!sta && event == OPN_RJCT) { mesh_plink_frame_tx(sdata, NULL, WLAN_SP_MESH_PEERING_CLOSE, mgmt->sa, 0, plid, @@ -1082,10 +1113,8 @@ mesh_process_plink_frame(struct ieee80211_sub_if_data *sdata, if (event == CNF_ACPT) { /* 802.11-2012 13.3.7.2 - update plid on CNF if not set */ - if (!sta->mesh->plid) { + if (!sta->mesh->plid) sta->mesh->plid = plid; - sta->sta.aid = sta->mesh->plid; - } sta->mesh->aid = get_unaligned_le16(PLINK_CNF_AID(mgmt)); }
Instead of using peer link id for AID, generate a new AID when creating mesh STAs in the kernel peering manager. This enables smaller TIM elements and more closely follows the standard, and it also enables mesh to work on drivers that require a valid AID when the STA is inserted (ath10k firmware has this requirement, for example). In the case of userspace-managed stations, we use the AID from NL80211_CMD_NEW_STATION. Signed-off-by: Bob Copeland <me@bobcopeland.com> --- v2: generate the bitmap as needed inside mesh_allocate_aid rather than maintaining the whole time (Johannes) v3: unchanged from v2 net/mac80211/mesh_plink.c | 41 +++++++++++++++++++++++++++++++++++------ 1 file changed, 35 insertions(+), 6 deletions(-)