* [RFC 1/2] mac80211: disconnect on suspend in managed mode @ 2013-02-19 16:47 Stanislaw Gruszka 2013-02-19 16:48 ` [RFC 2/2] mac80211: sync suspend and stop interface Stanislaw Gruszka 2013-02-20 10:06 ` [RFC 1/2] mac80211: disconnect on suspend in managed mode Johannes Berg 0 siblings, 2 replies; 6+ messages in thread From: Stanislaw Gruszka @ 2013-02-19 16:47 UTC (permalink / raw) To: Johannes Berg; +Cc: linux-wireless If possible that after suspend cfg80211 receive request to disconnect what require action on interface that was removed during suspend. Problem can manifest itself by various warnings similar to below one: WARNING: at net/mac80211/driver-ops.h:12 ieee80211_bss_info_change_notify+0x2f9/0x300 [mac80211]() wlan0: Failed check-sdata-in-driver check, flags: 0x4 Call Trace: [<c043e0b3>] warn_slowpath_fmt+0x33/0x40 [<f83707c9>] ieee80211_bss_info_change_notify+0x2f9/0x300 [mac80211] [<f83a660a>] ieee80211_recalc_ps_vif+0x2a/0x30 [mac80211] [<f83a6706>] ieee80211_set_disassoc+0xf6/0x500 [mac80211] [<f83a9441>] ieee80211_mgd_deauth+0x1f1/0x280 [mac80211] [<f8381b36>] ieee80211_deauth+0x16/0x20 [mac80211] [<f8261e70>] cfg80211_mlme_down+0x70/0xc0 [cfg80211] [<f8264de1>] __cfg80211_disconnect+0x1b1/0x1d0 [cfg80211] To fix the problem disconnect from AP during suspend. Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com> --- net/mac80211/mlme.c | 58 +++++++++-------------------------------------------- 1 file changed, 10 insertions(+), 48 deletions(-) diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 9f6464f..19c926d 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -2060,7 +2060,7 @@ struct sk_buff *ieee80211_ap_probereq_get(struct ieee80211_hw *hw, } EXPORT_SYMBOL(ieee80211_ap_probereq_get); -static void __ieee80211_disconnect(struct ieee80211_sub_if_data *sdata) +void ieee80211_disconnect(struct ieee80211_sub_if_data *sdata, u8 reason) { struct ieee80211_if_managed *ifmgd = &sdata->u.mgd; u8 frame_buf[IEEE80211_DEAUTH_FRAME_LEN]; @@ -2071,8 +2071,7 @@ static void __ieee80211_disconnect(struct ieee80211_sub_if_data *sdata) return; } - ieee80211_set_disassoc(sdata, IEEE80211_STYPE_DEAUTH, - WLAN_REASON_DISASSOC_DUE_TO_INACTIVITY, + ieee80211_set_disassoc(sdata, IEEE80211_STYPE_DEAUTH, reason, true, frame_buf); ifmgd->flags &= ~IEEE80211_STA_CSA_RECEIVED; ieee80211_wake_queues_by_reason(&sdata->local->hw, @@ -2105,7 +2104,8 @@ static void ieee80211_beacon_connection_loss_work(struct work_struct *work) if (ifmgd->connection_loss) { sdata_info(sdata, "Connection to AP %pM lost\n", ifmgd->bssid); - __ieee80211_disconnect(sdata); + ieee80211_disconnect(sdata, + WLAN_REASON_DISASSOC_DUE_TO_INACTIVITY); } else { ieee80211_mgd_probe_ap(sdata, true); } @@ -2117,7 +2117,7 @@ static void ieee80211_csa_connection_drop_work(struct work_struct *work) container_of(work, struct ieee80211_sub_if_data, u.mgd.csa_connection_drop_work); - __ieee80211_disconnect(sdata); + ieee80211_disconnect(sdata, WLAN_REASON_DISASSOC_DUE_TO_INACTIVITY); } void ieee80211_beacon_loss(struct ieee80211_vif *vif) @@ -3502,60 +3502,22 @@ void ieee80211_sta_quiesce(struct ieee80211_sub_if_data *sdata) { struct ieee80211_if_managed *ifmgd = &sdata->u.mgd; + ieee80211_disconnect(sdata, WLAN_REASON_DISASSOC_STA_HAS_LEFT); + /* - * we need to use atomic bitops for the running bits - * only because both timers might fire at the same - * time -- the code here is properly synchronised. + * As we disassociate all timers are canceled and pending works will + * perform no action, but we need to wait until works will finish, + * before we can freeze. */ - cancel_work_sync(&ifmgd->request_smps_work); - cancel_work_sync(&ifmgd->monitor_work); cancel_work_sync(&ifmgd->beacon_connection_loss_work); cancel_work_sync(&ifmgd->csa_connection_drop_work); - if (del_timer_sync(&ifmgd->timer)) - set_bit(TMR_RUNNING_TIMER, &ifmgd->timers_running); - cancel_work_sync(&ifmgd->chswitch_work); - if (del_timer_sync(&ifmgd->chswitch_timer)) - set_bit(TMR_RUNNING_CHANSW, &ifmgd->timers_running); - - /* these will just be re-established on connection */ - del_timer_sync(&ifmgd->conn_mon_timer); - del_timer_sync(&ifmgd->bcn_mon_timer); } void ieee80211_sta_restart(struct ieee80211_sub_if_data *sdata) { - struct ieee80211_if_managed *ifmgd = &sdata->u.mgd; - - mutex_lock(&ifmgd->mtx); - if (!ifmgd->associated) { - mutex_unlock(&ifmgd->mtx); - return; - } - - if (sdata->flags & IEEE80211_SDATA_DISCONNECT_RESUME) { - sdata->flags &= ~IEEE80211_SDATA_DISCONNECT_RESUME; - mlme_dbg(sdata, "driver requested disconnect after resume\n"); - ieee80211_sta_connection_lost(sdata, - ifmgd->associated->bssid, - WLAN_REASON_UNSPECIFIED, - true); - mutex_unlock(&ifmgd->mtx); - return; - } - mutex_unlock(&ifmgd->mtx); - - if (test_and_clear_bit(TMR_RUNNING_TIMER, &ifmgd->timers_running)) - add_timer(&ifmgd->timer); - if (test_and_clear_bit(TMR_RUNNING_CHANSW, &ifmgd->timers_running)) - add_timer(&ifmgd->chswitch_timer); - ieee80211_sta_reset_beacon_monitor(sdata); - - mutex_lock(&sdata->local->mtx); - ieee80211_restart_sta_timer(sdata); - mutex_unlock(&sdata->local->mtx); } #endif -- 1.7.11.7 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* [RFC 2/2] mac80211: sync suspend and stop interface 2013-02-19 16:47 [RFC 1/2] mac80211: disconnect on suspend in managed mode Stanislaw Gruszka @ 2013-02-19 16:48 ` Stanislaw Gruszka 2013-02-20 10:08 ` Johannes Berg 2013-02-20 10:06 ` [RFC 1/2] mac80211: disconnect on suspend in managed mode Johannes Berg 1 sibling, 1 reply; 6+ messages in thread From: Stanislaw Gruszka @ 2013-02-19 16:48 UTC (permalink / raw) To: Johannes Berg; +Cc: linux-wireless Is possible that we close interface while we are suspended, that results in warning like below (and some other similar): WARNING: at net/mac80211/driver-ops.h:12 ieee80211_do_stop+0x62e/0x670 [mac80211]() wlan0: Failed check-sdata-in-driver check, flags: 0x4 Call Trace: [<ffffffff8105c9d6>] warn_slowpath_fmt+0x46/0x50 [<ffffffffa045d46e>] ieee80211_do_stop+0x62e/0x670 [mac80211] [<ffffffffa045d4ca>] ieee80211_stop+0x1a/0x20 [mac80211] [<ffffffff815122ed>] __dev_close_many+0x7d/0xc0 [<ffffffff81513af8>] dev_close_many+0x88/0x100 [<ffffffff81513f2a>] dev_close+0x3a/0x50 [<ffffffffa03c90ae>] cfg80211_rfkill_set_block+0x6e/0xa0 [cfg80211] [<ffffffffa03c9106>] cfg80211_rfkill_sync_work+0x26/0x30 [cfg80211] To fix check if interface was removed during suspend in ieee80211_do_stop() before perform any action that require operation on the interface. Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com> --- net/mac80211/iface.c | 21 ++++++++++++--------- net/mac80211/pm.c | 2 ++ 2 files changed, 14 insertions(+), 9 deletions(-) diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c index 2c059e5..7703141 100644 --- a/net/mac80211/iface.c +++ b/net/mac80211/iface.c @@ -419,7 +419,8 @@ static void ieee80211_del_virtual_monitor(struct ieee80211_local *local) ieee80211_vif_release_channel(sdata); - drv_remove_interface(local, sdata); + if (!local->suspended || local->wowlan) + drv_remove_interface(local, sdata); kfree(sdata); out_unlock: @@ -834,16 +835,18 @@ static void ieee80211_do_stop(struct ieee80211_sub_if_data *sdata, skb_queue_purge(&sdata->skb_queue); - /* - * Free all remaining keys, there shouldn't be any, - * except maybe group keys in AP more or WDS? - */ - ieee80211_free_keys(sdata); - drv_remove_interface_debugfs(local, sdata); - if (going_down) - drv_remove_interface(local, sdata); + if (!local->suspended || local->wowlan) { + /* + * Free all remaining keys, there shouldn't be any, + * except maybe group keys in AP more or WDS? + */ + ieee80211_free_keys(sdata); + + if (going_down) + drv_remove_interface(local, sdata); + } } sdata->bss = NULL; diff --git a/net/mac80211/pm.c b/net/mac80211/pm.c index d0275f3..51b4ec7 100644 --- a/net/mac80211/pm.c +++ b/net/mac80211/pm.c @@ -182,6 +182,8 @@ int __ieee80211_suspend(struct ieee80211_hw *hw, struct cfg80211_wowlan *wowlan) mutex_unlock(&local->chanctx_mtx); } + + ieee80211_free_keys(sdata); drv_remove_interface(local, sdata); } -- 1.7.11.7 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [RFC 2/2] mac80211: sync suspend and stop interface 2013-02-19 16:48 ` [RFC 2/2] mac80211: sync suspend and stop interface Stanislaw Gruszka @ 2013-02-20 10:08 ` Johannes Berg 0 siblings, 0 replies; 6+ messages in thread From: Johannes Berg @ 2013-02-20 10:08 UTC (permalink / raw) To: Stanislaw Gruszka; +Cc: linux-wireless On Tue, 2013-02-19 at 17:48 +0100, Stanislaw Gruszka wrote: > Is possible that we close interface while we are suspended, that > results in warning like below (and some other similar): > > WARNING: at net/mac80211/driver-ops.h:12 ieee80211_do_stop+0x62e/0x670 [mac80211]() > wlan0: Failed check-sdata-in-driver check, flags: 0x4 > Call Trace: > [<ffffffff8105c9d6>] warn_slowpath_fmt+0x46/0x50 > [<ffffffffa045d46e>] ieee80211_do_stop+0x62e/0x670 [mac80211] > [<ffffffffa045d4ca>] ieee80211_stop+0x1a/0x20 [mac80211] > [<ffffffff815122ed>] __dev_close_many+0x7d/0xc0 > [<ffffffff81513af8>] dev_close_many+0x88/0x100 > [<ffffffff81513f2a>] dev_close+0x3a/0x50 > [<ffffffffa03c90ae>] cfg80211_rfkill_set_block+0x6e/0xa0 [cfg80211] > [<ffffffffa03c9106>] cfg80211_rfkill_sync_work+0x26/0x30 [cfg80211] > > To fix check if interface was removed during suspend in > ieee80211_do_stop() before perform any action that require operation > on the interface. This would also become simpler then because the key stuff for instance wouldn't be necessary any more. johannes ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [RFC 1/2] mac80211: disconnect on suspend in managed mode 2013-02-19 16:47 [RFC 1/2] mac80211: disconnect on suspend in managed mode Stanislaw Gruszka 2013-02-19 16:48 ` [RFC 2/2] mac80211: sync suspend and stop interface Stanislaw Gruszka @ 2013-02-20 10:06 ` Johannes Berg 2013-02-21 10:29 ` Stanislaw Gruszka 1 sibling, 1 reply; 6+ messages in thread From: Johannes Berg @ 2013-02-20 10:06 UTC (permalink / raw) To: Stanislaw Gruszka; +Cc: linux-wireless On Tue, 2013-02-19 at 17:47 +0100, Stanislaw Gruszka wrote: > If possible that after suspend cfg80211 receive request to disconnect > what require action on interface that was removed during suspend. > Problem can manifest itself by various warnings similar to below one: [...] > To fix the problem disconnect from AP during suspend. This will fix the immediate problem, but not in general. What if we're joined to an IBSS while suspending? A similar thing could happen. Or operating as an AP? Or joined to a mesh? etc. I think it'd be better to implement this as explicit disconnect, leave, teardown, whatever ... in cfg80211 rather than mac80211, although then of course mac80211 could have a bunch of cleanups, e.g. the various _quiesce functions could be reduced then I think. johannes ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [RFC 1/2] mac80211: disconnect on suspend in managed mode 2013-02-20 10:06 ` [RFC 1/2] mac80211: disconnect on suspend in managed mode Johannes Berg @ 2013-02-21 10:29 ` Stanislaw Gruszka 2013-02-22 20:31 ` Johannes Berg 0 siblings, 1 reply; 6+ messages in thread From: Stanislaw Gruszka @ 2013-02-21 10:29 UTC (permalink / raw) To: Johannes Berg; +Cc: linux-wireless On Wed, Feb 20, 2013 at 11:06:44AM +0100, Johannes Berg wrote: > On Tue, 2013-02-19 at 17:47 +0100, Stanislaw Gruszka wrote: > > If possible that after suspend cfg80211 receive request to disconnect > > what require action on interface that was removed during suspend. > > Problem can manifest itself by various warnings similar to below one: > [...] > > To fix the problem disconnect from AP during suspend. > > This will fix the immediate problem, but not in general. What if we're > joined to an IBSS while suspending? A similar thing could happen. Or > operating as an AP? Or joined to a mesh? etc. > > I think it'd be better to implement this as explicit disconnect, leave, > teardown, whatever ... in cfg80211 rather than mac80211, although then > of course mac80211 could have a bunch of cleanups, e.g. the various > _quiesce functions could be reduced then I think. Ok, I have below patch: diff --git a/net/wireless/core.c b/net/wireless/core.c index 5ffff03..5a4dc09 100644 --- a/net/wireless/core.c +++ b/net/wireless/core.c @@ -815,6 +815,46 @@ void cfg80211_update_iface_num(struct cfg80211_registered_device *rdev, rdev->num_running_monitor_ifaces += num; } +void cfg80211_leave(struct cfg80211_registered_device *rdev, + struct wireless_dev *wdev) +{ + struct net_device *dev = wdev->netdev; + + switch (wdev->iftype) { + case NL80211_IFTYPE_ADHOC: + cfg80211_leave_ibss(rdev, dev, true); + break; + case NL80211_IFTYPE_P2P_CLIENT: + case NL80211_IFTYPE_STATION: + mutex_lock(&rdev->sched_scan_mtx); + __cfg80211_stop_sched_scan(rdev, false); + mutex_unlock(&rdev->sched_scan_mtx); + + wdev_lock(wdev); +#ifdef CONFIG_CFG80211_WEXT + kfree(wdev->wext.ie); + wdev->wext.ie = NULL; + wdev->wext.ie_len = 0; + wdev->wext.connect.auth_type = NL80211_AUTHTYPE_AUTOMATIC; +#endif + __cfg80211_disconnect(rdev, dev, + WLAN_REASON_DEAUTH_LEAVING, true); + cfg80211_mlme_down(rdev, dev); + wdev_unlock(wdev); + break; + case NL80211_IFTYPE_MESH_POINT: + cfg80211_leave_mesh(rdev, dev); + break; + case NL80211_IFTYPE_AP: + cfg80211_stop_ap(rdev, dev); + break; + default: + break; + } + + wdev->beacon_interval = 0; +} + static int cfg80211_netdev_notifier_call(struct notifier_block *nb, unsigned long state, void *ndev) @@ -883,38 +923,7 @@ static int cfg80211_netdev_notifier_call(struct notifier_block *nb, dev->priv_flags |= IFF_DONT_BRIDGE; break; case NETDEV_GOING_DOWN: - switch (wdev->iftype) { - case NL80211_IFTYPE_ADHOC: - cfg80211_leave_ibss(rdev, dev, true); - break; - case NL80211_IFTYPE_P2P_CLIENT: - case NL80211_IFTYPE_STATION: - mutex_lock(&rdev->sched_scan_mtx); - __cfg80211_stop_sched_scan(rdev, false); - mutex_unlock(&rdev->sched_scan_mtx); - - wdev_lock(wdev); -#ifdef CONFIG_CFG80211_WEXT - kfree(wdev->wext.ie); - wdev->wext.ie = NULL; - wdev->wext.ie_len = 0; - wdev->wext.connect.auth_type = NL80211_AUTHTYPE_AUTOMATIC; -#endif - __cfg80211_disconnect(rdev, dev, - WLAN_REASON_DEAUTH_LEAVING, true); - cfg80211_mlme_down(rdev, dev); - wdev_unlock(wdev); - break; - case NL80211_IFTYPE_MESH_POINT: - cfg80211_leave_mesh(rdev, dev); - break; - case NL80211_IFTYPE_AP: - cfg80211_stop_ap(rdev, dev); - break; - default: - break; - } - wdev->beacon_interval = 0; + cfg80211_leave(rdev, wdev); break; case NETDEV_DOWN: cfg80211_update_iface_num(rdev, wdev->iftype, -1); diff --git a/net/wireless/core.h b/net/wireless/core.h index 3aec0e4..878d204 100644 --- a/net/wireless/core.h +++ b/net/wireless/core.h @@ -503,6 +503,9 @@ int cfg80211_validate_beacon_int(struct cfg80211_registered_device *rdev, void cfg80211_update_iface_num(struct cfg80211_registered_device *rdev, enum nl80211_iftype iftype, int num); +void cfg80211_leave(struct cfg80211_registered_device *rdev, + struct wireless_dev *wdev); + #define CFG80211_MAX_NUM_DIFFERENT_CHANNELS 10 #ifdef CONFIG_CFG80211_DEVELOPER_WARNINGS diff --git a/net/wireless/sysfs.c b/net/wireless/sysfs.c index 238ee49..64e951f 100644 --- a/net/wireless/sysfs.c +++ b/net/wireless/sysfs.c @@ -86,16 +86,21 @@ static int wiphy_uevent(struct device *dev, struct kobj_uevent_env *env) static int wiphy_suspend(struct device *dev, pm_message_t state) { struct cfg80211_registered_device *rdev = dev_to_rdev(dev); + struct wireless_dev *wdev; int ret = 0; rdev->suspend_at = get_seconds(); - if (rdev->ops->suspend) { - rtnl_lock(); - if (rdev->wiphy.registered) + rtnl_lock(); + if (rdev->wiphy.registered) { + list_for_each_entry_rcu(wdev, &rdev->wdev_list, list) + cfg80211_leave(rdev, wdev); + + if (rdev->ops->suspend) ret = rdev_suspend(rdev); - rtnl_unlock(); + } + rtnl_unlock(); return ret; } ^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [RFC 1/2] mac80211: disconnect on suspend in managed mode 2013-02-21 10:29 ` Stanislaw Gruszka @ 2013-02-22 20:31 ` Johannes Berg 0 siblings, 0 replies; 6+ messages in thread From: Johannes Berg @ 2013-02-22 20:31 UTC (permalink / raw) To: Stanislaw Gruszka; +Cc: linux-wireless On Thu, 2013-02-21 at 11:29 +0100, Stanislaw Gruszka wrote: > --- a/net/wireless/sysfs.c > +++ b/net/wireless/sysfs.c > @@ -86,16 +86,21 @@ static int wiphy_uevent(struct device *dev, struct kobj_uevent_env *env) > static int wiphy_suspend(struct device *dev, pm_message_t state) > { > struct cfg80211_registered_device *rdev = dev_to_rdev(dev); > + struct wireless_dev *wdev; > int ret = 0; > > rdev->suspend_at = get_seconds(); > > - if (rdev->ops->suspend) { > - rtnl_lock(); > - if (rdev->wiphy.registered) > + rtnl_lock(); > + if (rdev->wiphy.registered) { > + list_for_each_entry_rcu(wdev, &rdev->wdev_list, list) > + cfg80211_leave(rdev, wdev); > + > + if (rdev->ops->suspend) > ret = rdev_suspend(rdev); I think this would break WoWLAN? johannes ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2013-02-22 20:31 UTC | newest] Thread overview: 6+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2013-02-19 16:47 [RFC 1/2] mac80211: disconnect on suspend in managed mode Stanislaw Gruszka 2013-02-19 16:48 ` [RFC 2/2] mac80211: sync suspend and stop interface Stanislaw Gruszka 2013-02-20 10:08 ` Johannes Berg 2013-02-20 10:06 ` [RFC 1/2] mac80211: disconnect on suspend in managed mode Johannes Berg 2013-02-21 10:29 ` Stanislaw Gruszka 2013-02-22 20:31 ` Johannes Berg
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).