aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAge
* cfg80211: wrap BSS krefJohannes Berg2013-02-04
| | | | | | | Add inline wrappers for the BSS struct krefs to be able to extend them easily later. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* cfg80211: remove a local variableJohannes Berg2013-02-04
| | | | | | This local variable is only used once, so remove it. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* cfg80211: use lockdep to assert lock is heldJohannes Berg2013-02-04
| | | | | | | Instead of annotating with a comment, add a lockdep annotation which also serves as documentation. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: send deauth if connection was lost during suspendJohannes Berg2013-02-04
| | | | | | | | | If the driver determined the connection was lost or that it couldn't securely maintain the connection when coming out of WoWLAN, send a deauth frame to the AP to also let it know. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: allow transmitting deauth with tainted keyJohannes Berg2013-02-04
| | | | | | | | | | | | | | | When we had a connection for WoWLAN and after resume it needed to be disconnected, the previous commit enabled sending a deauth frame to the AP. This frame would not go through on MFP-enabled networks as the key for it is marked tainted before the frame is transmitted. Allow a tainted key to be used for deauth frames. Worst case, we'll use a wrong key because the PTK was rekeyed while suspended, but more likely the PTK is still fine and the taint flag really only applies to the GTK(s). Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* cfg80211: move locking into cfg80211_bss_ageJohannes Berg2013-02-04
| | | | | | | There's no reason for it to require external locking, move it into the function. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* cfg80211: fix BSS IE allocation commentJohannes Berg2013-02-04
| | | | | | | | The comment about allocating the IEs together with the BSS struct is no longer true, remove it. Also fix a typo in the same area. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: remove unused SSID from BSSJohannes Berg2013-02-04
| | | | | | | | | The ssid/ssid_len fields in the private BSS struct are unused, contrary to the comment we do look up the SSID in the few cases we need it. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* cfg80211: remove unused cfg80211_get_meshJohannes Berg2013-02-04
| | | | | | | | As Thomas pointed out, cfg80211_get_mesh() is unused and can be removed. Cc: Thomas Pedersen <thomas@cozybit.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* cfg80211: simplify mesh BSS comparisonJohannes Berg2013-02-04
| | | | | | | | | | Instead of first checking if a BSS is an MBSS and then doing the comparisons, inline it all into the BSS comparison function. This avoids doing the IE searches twice and is also a lot less code. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* cfg80211: fix BSS list hidden SSID lookupJohannes Berg2013-02-04
| | | | | | | | | | | | | | | | | When trying to find a hidden SSID, the lookup function is done wrong; the code is trying to combine the two lookups into one, and as a consequence doesn't always find the entry at all. To understand this, consider a case where multiple BSS entries with the same channel and BSSID exist but have different SSID length. Then comparing against the probe response SSID length is bound to cause problems since the hidden one might be either zeroed out or zero-length. To fix this we need to do two lookups for the two ways to hide SSIDs. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* cfg80211: refactor hidden SSID findingJohannes Berg2013-02-04
| | | | | | | | Instead of duplicating the rbtree functions, pass an argument to the compare function. This removes the code duplication for the two searches. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* cfg80211: expand per-station byte counters to 64bitVladimir Kondratiev2013-02-04
| | | | | | | | | | | | | | | | | | | In per-station statistics, present 32bit counters are too small for practical purposes - with gigabit speeds, it get overlapped every few seconds. Expand counters in the struct station_info to be 64-bit. Driver can still fill only 32-bit and indicate in @filled only bits like STATION_INFO_[TR]X_BYTES; in case driver provides full 64-bit counter, it should also set in @filled bit STATION_INFO_[TR]RX_BYTES64 Netlink sends both 32-bit and 64-bit counters, if present, to not break userspace. Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com> [change to also have 32-bit counters if driver advertises 64-bit] Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: always allow calling ieee80211_connection_loss()Johannes Berg2013-02-04
| | | | | | | | | | | | | With multi-channel, there's a corner case where a driver doesn't receive a beacon soon enough to be able to sync its timers with the AP. In this case, the only recovery (after trying again) is to disconnect from the AP. Allow calling ieee80211_connection_loss() for such cases. To make that possible, modify the work function to not rely on the IEEE80211_HW_CONNECTION_MONITOR flag but use new state kept in the interface instead. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: send deauth when connection is lostJohannes Berg2013-02-04
| | | | | | | | | If the driver determines the connection is lost, send a deauth frame to the AP anyway just in case it still considers the connection alive. The frame might not go through, but at least we've tried. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: improve latency and throughput while software scanningStanislaw Gruszka2013-01-31
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Patch vastly improve latency while scanning. Slight throughput improvements were observed as well. Is intended for improve performance of voice and video applications, when scan is periodically requested by user space (i.e. default NetworkManager behaviour). Patch remove latency requirement based on PM_QOS_NETWORK_LATENCY, this value is 2000 seconds by default (i.e. approximately 0.5 hour !?!). Also remove listen interval requirement, which based on beaconing and depending on BSS parameters. It can make we stay off-channel for a second or more. Instead try to offer the best latency that we could, i.e. be off-channel no longer than PASSIVE channel scan time: 125 ms. That mean we will scan two ACTIVE channels and go back to on-channel, and one PASSIVE channel, and go back to on-channel. Patch also decrease PASSIVE channel scan time to about 110 ms. As drawback patch increase overall scan time. On my tests, when scanning both 2GHz and 5GHz bands, scanning time increase from 5 seconds up to 10 seconds. Since that increase happen only when we are associated, I think it can be acceptable. If eventually better scan time is needed for situations when we lose signal and quickly need to decide to which AP roam, additional scan flag or parameter can be introduced. I tested patch by doing: while true; do iw dev wlan0 scan; sleep 3; done > /dev/null and ping -i0.2 -c 1000 HOST on remote and local machine, results are as below: * Ping from local periodically scanning machine to AP: Unpatched: rtt min/avg/max/mdev = 0.928/24.946/182.135/36.873 ms Patched: rtt min/avg/max/mdev = 0.928/19.678/150.845/33.130 ms * Ping from remote machine to periodically scanning machine: Unpatched: rtt min/avg/max/mdev = 1.637/120.683/709.139/164.337 ms Patched: rtt min/avg/max/mdev = 1.807/26.893/201.435/40.284 ms Throughput measured by scp show following results. * Upload to periodically scanning machine: Unpatched: 3.9MB/s 03:15 Patched: 4.3MB/s 02:58 * Download from periodically scanning machine: Unpatched: 5.5MB/s 02:17 Patched: 6.2MB/s 02:02 Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: start auth/assoc timeout on frame statusJohannes Berg2013-01-31
| | | | | | | | | | | | | | | | | | | When sending authentication/association frames they might take a bit of time to go out because we may have to synchronise with the AP, in particular in the case where it's really a P2P GO. In this case the 200ms fixed timeout could potentially be too short if the beacon interval is relatively large. For drivers that report TX status we can do better. Instead of starting the timeout directly, start it only when the frame status arrives. Since then the frame was out on the air, we can wait shorter (the typical response time is supposed to be 30ms, wait 100ms.) Also, if the frame failed to be transmitted try again right away instead of waiting. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* cfg80211: remove free_priv BSS APIJohannes Berg2013-01-31
| | | | | | | | | Now that mac80211 no longer uses this API, remove it completely. If anyone needs it again, we can revert this patch of course, but mac80211 was the only user right now. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: remove unused mesh data from bssJohannes Berg2013-01-31
| | | | | | These pointers/values are never used, remove them. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: remove last_probe_resp from bssJohannes Berg2013-01-31
| | | | | | | We track this, but never use it, so we can just remove it. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: inform the driver about update of dtim_periodEmmanuel Grumbach2013-01-31
| | | | | | | | | | | | | | | | | | | | | | | Currently, when the driver requires the DTIM period, mac80211 will wait to hear a beacon before association. This behavior is suboptimal since some drivers may be able to deal with knowing the DTIM period after the association, if they get it at all. To address this, notify the drivers with bss_info_changed with the new BSS_CHANGED_DTIM_PERIOD flag when the DTIM becomes known. This might be when changing to associated, or later when the entire association was done with only probe response information. Rename the hardware flag for the current behaviour to IEEE80211_HW_NEED_DTIM_BEFORE_ASSOC to more accurately reflect its behaviour. IEEE80211_HW_NEED_DTIM_PERIOD is no longer accurate as all drivers get the DTIM period now, just not before association. Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: remove assoc data "sent_assoc"Johannes Berg2013-01-31
| | | | | | The field is never used, so remove it. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* cfg80211/mac80211: support reporting wakeup reasonJohannes Berg2013-01-31
| | | | | | | | | When waking up from WoWLAN, it is useful to know what triggered the wakeup. Support reporting the wakeup reason(s) in cfg80211 (and a pass-through in mac80211) to allow userspace to know. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: dynamic short slot time for MBSSsThomas Pedersen2013-01-29
| | | | | | | | | | | | | | The standard mandates mesh STAs to set the ERP Short Slot Time capability info bit in beacons to 0. Even though this is their way of disallowing short slot time for mesh STAs, there should be no harm in enabling it if we determine all STAs in the current MBSS support ERP rates. Increases throughput about 20% for legacy rates when enabled. Signed-off-by: Thomas Pedersen <thomas@cozybit.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* cfg80211: add SME state to warning in __cfg80211_mlme_disassocJohannes Berg2013-01-29
| | | | | | | | | The warning here occasionally triggers but we haven't found the cause yet. It's a valid warning since if it triggers the SME state got confused, so add the SME state to it to help narrow it down in the future. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: remove IEEE80211_HW_TEARDOWN_AGGR_ON_BAR_FAILStanislaw Gruszka2013-01-29
| | | | | | | | | | | | | | | | | | | | | | This is basically a revert of: commit 5b632fe85ec82e5c43740b52e74c66df50a37db3 Author: Stanislaw Gruszka <sgruszka@redhat.com> Date: Mon Dec 3 12:56:33 2012 +0100 mac80211: introduce IEEE80211_HW_TEARDOWN_AGGR_ON_BAR_FAIL We do not need this flag any longer, rt2x00 BAR/BA problem was fixed correctly by wireless-testing commit: commit 84e9e8ebd369679a958200a8baca96aafb2393bb Author: Helmut Schaa <helmut.schaa@googlemail.com> Date: Thu Jan 17 17:34:32 2013 +0100 rt2x00: Improve TX status handling for BlockAckReq frames Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* Merge remote-tracking branch 'wireless-next/master' into HEADJohannes Berg2013-01-29
|\
| * Merge branch 'master' of ↵John W. Linville2013-01-28
| |\ | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless Conflicts: drivers/net/wireless/ath/ath9k/main.c drivers/net/wireless/iwlwifi/dvm/tx.c
| | * mwifiex: fix typo in PCIe adapter NULL checkAvinash Patil2013-01-22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add missing "!" as we are supposed to check "!card->adapter" in PCIe suspend handler. Cc: "3.2+" <stable@vger.kernel.org> Signed-off-by: Avinash Patil <patila@marvell.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Reviewed-by: Sergey V. <sftp.mtuci@gmail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| | * ath9k: allow setting arbitrary antenna masks on AR9003+Felix Fietkau2013-01-22
| | | | | | | | | | | | | | | | | | Cc: stable@vger.kernel.org Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| | * ath9k_hw: fix chain swap setting when setting rx chainmask to 5Felix Fietkau2013-01-22
| | | | | | | | | | | | | | | | | | | | | | | | | | | Chain swapping should only be enabled when the EEPROM chainmask is set to 5, regardless of what the runtime chainmask is. Cc: stable@vger.kernel.org Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| | * ath9k_hw: fix calibration issues on chainmask that don't include chain 0Felix Fietkau2013-01-22
| | | | | | | | | | | | | | | | | | Cc: stable@vger.kernel.org Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| | * rtlwifi: Fix build warning introduced by commit a290593Larry Finger2013-01-22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The kbuild test robot reports the following warning with x86_64-randconfig-x955: warning: (RTL8192CE && RTL8192SE && RTL8192DE && RTL8723AE && RTL8192CU) selects RTLWIFI which has unmet direct dependencies (NETDEVICES && WLAN && (RTL8192CE || RTL8192CU || RTL8192SE || RTL8192DE)) This warning was introduced in commit a290593, "rtlwifi: Modify files for addition of rtl8723ae", and is d ue to a missing dependence of RTLWIFI on RTL8723AE. Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net> Cc: kbuild test robot <fengguang.wu@intel.com> Reported-by: kbuild test robot <fengguang.wu@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| | * iwlegacy: fix IBSS cleanupStanislaw Gruszka2013-01-22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We do not correctly change interface type when switching from IBSS mode to STA mode, that results in microcode errors. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=886946 Reported-by: Jaroslav Skarvada <jskarvad@redhat.com> Cc: stable@vger.kernel.org Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| | * Merge branch 'for-john' of ↵John W. Linville2013-01-22
| | |\ | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211
| | | * mac80211: add encrypt headroom to PERR framesBob Copeland2013-01-16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Mesh PERR action frames are robust and thus may be encrypted, so add proper head/tailroom to allow this. Fixes this warning when operating a Mesh STA on ath5k: WARNING: at net/mac80211/wpa.c:427 ccmp_encrypt_skb.isra.5+0x7b/0x1a0 [mac80211]() Call Trace: [<c011c5e7>] warn_slowpath_common+0x63/0x78 [<c011c60b>] warn_slowpath_null+0xf/0x13 [<e090621d>] ccmp_encrypt_skb.isra.5+0x7b/0x1a0 [mac80211] [<e090685c>] ieee80211_crypto_ccmp_encrypt+0x1f/0x37 [mac80211] [<e0917113>] invoke_tx_handlers+0xcad/0x10bd [mac80211] [<e0917665>] ieee80211_tx+0x87/0xb3 [mac80211] [<e0918932>] ieee80211_tx_pending+0xcc/0x170 [mac80211] [<c0121c43>] tasklet_action+0x3e/0x65 Signed-off-by: Bob Copeland <me@bobcopeland.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
| | | * mac80211: set NEED_TXPROCESSING for PERR framesBob Copeland2013-01-16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | A user reported warnings in ath5k due to transmitting frames with no rates set up. The frames were Mesh PERR frames, and some debugging showed an empty control block with just the vif pointer: > [ 562.522682] XXX txinfo: 00000000: 00 00 00 00 00 00 00 00 00 00 00 > 00 00 00 00 00 ................ > [ 562.522688] XXX txinfo: 00000010: 00 00 00 00 00 00 00 00 54 b8 f2 > db 00 00 00 00 ........T....... > [ 562.522693] XXX txinfo: 00000020: 00 00 00 00 00 00 00 00 00 00 00 > 00 00 00 00 00 ................ Set the IEEE80211_TX_INTFL_NEED_TXPROCESSING flag to ensure that rate control gets run before the frame is sent. Signed-off-by: Bob Copeland <me@bobcopeland.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
| | | * mac80211: fix monitor mode injectionFelix Fietkau2013-01-16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Channel contexts are not always used with monitor interfaces. If no channel context is set, use the oper channel, otherwise tx fails. Signed-off-by: Felix Fietkau <nbd@openwrt.org> [check local->use_chanctx] Signed-off-by: Johannes Berg <johannes.berg@intel.com>
| | | * mac80211: synchronize scan off/on-channel and PS statesStanislaw Gruszka2013-01-16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Since: commit b23b025fe246f3acc2988eb6d400df34c27cb8ae Author: Ben Greear <greearb@candelatech.com> Date: Fri Feb 4 11:54:17 2011 -0800 mac80211: Optimize scans on current operating channel. we do not disable PS while going back to operational channel (on ieee80211_scan_state_suspend) and deffer that until scan finish. But since we are allowed to send frames, we can send a frame to AP without PM bit set, so disable PS on AP side. Then when we switch to off-channel (in ieee80211_scan_state_resume) we do not enable PS. Hence we are off-channel with PS disabled, frames are not buffered by AP. To fix remove offchannel_ps_disable argument and always enable PS when going off-channel and disable it when going on-channel, like it was before. Cc: stable@vger.kernel.org # 2.6.39+ Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com> Tested-by: Seth Forshee <seth.forshee@canonical.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
| | | * mac80211: fix FT roamingJohannes Berg2013-01-16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | During FT roaming, wpa_supplicant attempts to set the key before association. This used to be rejected, but as a side effect of my commit 66e67e418908442389d3a9e ("mac80211: redesign auth/assoc") the key was accepted causing hardware crypto to not be used for it as the station isn't added to the driver yet. It would be possible to accept the key and then add it to the driver when the station has been added. However, this may run into issues with drivers using the state- based station adding if they accept the key only after association like it used to be. For now, revert to the behaviour from before the auth and assoc change. Cc: stable@vger.kernel.org Reported-by: Cédric Debarge <cedric.debarge@acksys.fr> Tested-by: Cédric Debarge <cedric.debarge@acksys.fr> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
| | * | Merge branch 'for-john' of ↵John W. Linville2013-01-22
| | |\ \ | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi-fixes
| | | * | iwlwifi: audit single frames from AGG queue in RSEmmanuel Grumbach2013-01-16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The rate scaling won't treat the information in a frame with IEEE80211_TX_CTL_AMPDU set if IEEE80211_TX_STAT_AMPDU is cleared. But all the frames coming from an AGG tx queue have IEEE80211_TX_CTL_AMPDU set, and IEEE80211_TX_STAT_AMPDU is set only if the frame was sent in an AMPDU. This means that all the data in frames in AGG tx queues that aren't sent as an AMPDU is thrown away. This is even more harmful when in bad link conditions, the frames are sent in an AMPDU and then finally sent as single frame. So a lot of failures weren't reported and the rate scaling got stuck in high rates leading to very poor connectivity. Fix that by clearing IEEE80211_TX_CTL_AMPDU when the frame isn't part of an AMPDU. This bug was introduced by 2eb81a40aa521035ff9c8c8309e482dff523f8c9 iwlwifi: don't clear CTL_AMPDU on frame status This fix basically reverts the aforementioned commit. Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
| | * | | ath9k: disable the tasklet before taking the PCU lockFelix Fietkau2013-01-14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Fixes a reported CPU soft lockup where the tasklet tries to acquire the lock and blocks while ath_prepare_reset (holding the lock) waits for it to complete. Cc: stable@vger.kernel.org Reported-by: Robert Shade <robert.shade@gmail.com> Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| | * | | ath9k: remove sc->rx.rxbuflock to fix a deadlockFelix Fietkau2013-01-14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The commit "ath9k: fix rx flush handling" added a deadlock that happens because ath_rx_tasklet is called in a section that has already taken the rx buffer lock. It seems that the only purpose of the rxbuflock was a band-aid fix to the reset vs rx tasklet race, which has been properly fixed in the commit "ath9k: add a better fix for the rx tasklet vs rx flush race". Now that the fix is in, we can safely remove the lock to avoid such issues. Cc: stable@vger.kernel.org Reported-by: Sujith Manoharan <c_manoha@qca.qualcomm.com> Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| | * | | brcmsmac: increase timer reference count for new timers onlyPiotr Haber2013-01-11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | On hardware reintialization reference count of already existing timers would be increased again. This leads to problems on module unloading. Cc: stable@vger.kernel.org Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com> Reviewed-by: Hante Meuleman <meuleman@broadcom.com> Reviewed-by: Arend van Spriel <arend@broadcom.com> Signed-off-by: Piotr Haber <phaber@broadcom.com> Signed-off-by: Arend van Spriel <arend@broadcom.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| | * | | ath9k: fix rx flush handlingFelix Fietkau2013-01-11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Right now the rx flush is not doing anything useful on AR9003+, as it only works if the buffers in the rx FIFO have not been purged yet, as is done by ath_stoprecv. To fix this, always call ath_flushrecv from within ath_stoprecv before the FIFO is emptied, but still after the hw receive path has been stopped. This ensures that frames received (and ACKed by the hardware) shortly before a reset will be seen by the software, which should improve A-MPDU session stability. Cc: stable@vger.kernel.org Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| | * | | ath9k: add a better fix for the rx tasklet vs rx flush raceFelix Fietkau2013-01-11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Ensure that the rx tasklet is no longer running when entering the reset path. Also remove the distinction between flush and no-flush frame processing. If a frame has been received and ACKed by the hardware, the stack needs to see it, so that the BA receive window does not go out of sync. Cc: stable@vger.kernel.org Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| | * | | ath9k: remove the WARN_ON that triggers if generating a beacon failsFelix Fietkau2013-01-11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | During teardown, mac80211 will not return a new beacon. This is normal and handled properly in the driver, so there's no need to spam the user with a kernel warning here. Cc: stable@vger.kernel.org Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| | * | | ath9k: fix double-free bug on beacon generate failureFelix Fietkau2013-01-11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When the next beacon is sent, the ath_buf from the previous run is reused. If getting a new beacon from mac80211 fails, bf->bf_mpdu is not reset, yet the skb is freed, leading to a double-free on the next beacon tx attempt, resulting in a system crash. Cc: stable@vger.kernel.org Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| | * | | ath9k: do not link receive buffers during flushFelix Fietkau2013-01-11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | On AR9300 the rx FIFO needs to be empty during reset to ensure that no further DMA activity is generated, otherwise it might lead to memory corruption issues. Cc: stable@vger.kernel.org Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
generated by cgit v1.2.2 (git 2.25.0) at 2025-09-09 19:23:03 -0400