- 06 Mar, 2012 24 commits
-
-
Johannes Berg authored
In the WoWLAN suspend flow, instead of accessing registers directly, ask the transport to do the required setup at the end of suspend. If the transport doesn't implement this, don't tell the stack we support WoWLAN. When the device suspends w/o WoWLAN, mac80211 will have stopped it already, which has already called iwl_apm_stop() via stop_hw(). Thus, it isn't necessary to call it again in pcie_suspend and we can simply do nothing there. This unifies the regular and WoWLAN suspend. Signed-off-by:
Johannes Berg <johannes.berg@intel.com> Signed-off-by:
Wey-Yi Guy <wey-yi.w.guy@intel.com> Signed-off-by:
John W. Linville <linville@tuxdriver.com>
-
Johannes Berg authored
The code in the two branches of the #ifdef is almost the same so move the common code outside the #ifdef. Signed-off-by:
-
Johannes Berg authored
Writing to the global config structures is always wrong. To protect against such mistakes in the future, mark them const. Signed-off-by:
-
Johannes Berg authored
It's not really a good idea to write to the global static configuration. Use the valid TX/RX antenna information only from the HW params struct except in the case where the values from the config are used to override the values from the EEPROM. Signed-off-by:
-
Johannes Berg authored
There's no SKU override, we always just use it from EEPROM. As such, we can remove it from the config and use it in hw_param only. Since iwl_eeprom_check_sku() really needs to fill it in also rename that to iwl_eeprom_init_hw_params(). Signed-off-by:
-
Johannes Berg authored
There's no need for the per-device debug level that we expose in debugfs since the module parameter is writable in sysfs. At the same time, simplify code by changing iwl_get_debug_level(shrd) & IWL_DL_ISR) to iwl_have_debug_level(IWL_DL_ISR) Signed-off-by: -
Johannes Berg authored
With mac80211 now giving us station information (via the sta_state callback) before auth/assoc we can get rid of tx_sync by adding the station early for the case of managed interfaces. Keep AP mode actions the same for now. As we now get the BSSID early, we can also get rid of iwl_reprogram_ap_sta(). We can still optimise the number of RXONs we send to the device, but that can be done later. Signed-off-by:
-
Johannes Berg authored
There's no reason to set EXIT_PENDING when we start removing the module, as mac80211 will cleanly shut down the device in this case. Additionally, there's no point in rejecting commands to the device when we're cleaning up as that only leads to unwanted errors from mac80211 being printed, such as failed to remove key (...) from hardware (-16) Signed-off-by:
-
Johannes Berg authored
Some data doesn't need protection, some of the lock places are simply useless, and some data can be protected with the mutex instead. Thus the shared lock can be removed by making those changes. Signed-off-by:
-
Johannes Berg authored
The statistics are currently only half-heartedly locked against concurrent reading & modification so introduce a lock to really protect them. Signed-off-by:
-
Johannes Berg authored
Now that the transport has its own locking, there's no need to have the sta_lock in the shared data. Also, it can be a BH lock as it's not used from IRQ handlers. Signed-off-by:
-
Johannes Berg authored
Instead of (ab)using the sta_lock, make the transport layer lock its own TX queue data structures with a lock per queue. This also unifies with the cmd queue lock. Signed-off-by:
-
Johannes Berg authored
In the powersave related functions we only need to set a few parameters for the station command and can otherwise leave it blank -- there's no need to copy it from the database. This allows us to not use the sta lock here. Signed-off-by:
-
Johannes Berg authored
Simplify the code by returning directly in the error case and replacing the switch/if with a single if statement. Signed-off-by:
-
Johannes Berg authored
When we TX, we hold the sta_lock for a long time, a lot of which isn't needed. Reduce the time we hold the lock. Note that this doesn't really change anything as the code is already under the other spinlock, but it makes the code easier to read. Signed-off-by:
-
Johannes Berg authored
The station ID argument to the function iwlagn_tx_cmd_build_hwcrypto isn't used so remove it. Signed-off-by:
-
Johannes Berg authored
There's only one place using this function, so move it where it's needed. Signed-off-by:
-
Dan Carpenter authored
We should be doing a shift (1 << FLAG_RADIO_DOWN) here before testing the flag. As luck would have it, this test works almost correctly. The current code tests for FLAG_RADIO_OFF instead of FLAG_RADIO_DOWN. #define FLAG_RADIO_OFF 0 /* User disabling of MAC */ #define FLAG_RADIO_DOWN 1 /* ifup/ifdown disabling of MAC */ Signed-off-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
-
Johannes Berg authored
If we only monitor while associated, the following can happen: - we're associated, and the queue stuck check runs, setting the queue "touch" time to X - we disassociate, stopping the monitoring, which leaves the time set to X - almost 2s later, we associate, and enqueue a frame - before the frame is transmitted, we monitor for stuck queues, and find the time set to X, although it is now later than X + 2000ms, so we decide that the queue is stuck and erroneously restart the device It happens more with P2P because there we can go between associated/unassociated frequently. Cc: stable@vger.kernel.org Reported-by:
Ben Cahill <ben.m.cahill@intel.com> Signed-off-by:
-
Christian Lamparter authored
Signed-off-by:
Christian Lamparter <chunkeey@googlemail.com> Signed-off-by:
-
Christian Lamparter authored
Signed-off-by:
-
Jingjun Wu authored
Driver rtl8192ce when used with the RTL8188CE device would start at about 20 Mbps on a 54 Mbps connection, but quickly drop to 1 Mbps. One of the symptoms is that the AP would need to retransmit each packet 4 of 5 times before the driver would acknowledge it. Recovery is possible only by unloading and reloading the driver. This problem was reported at https://bugzilla.redhat.com/show_bug.cgi?id=770207. The problem is due to a missing update of the gain setting. Signed-off-by:
Jingjun Wu <jingjun_wu@realsil.com.cn> Signed-off-by:
Larry Finger <Larry.Finger@lwfinger.net> Cc: Stable <stable@vger.kernel.org> Signed-off-by:
-
Chun-Yeow Yeoh authored
This patch allows the possibility of having the mesh point and AP to be operated simultaneously in one single radio. Previously, mesh point fails to generate the mesh beacon if virtual AP is created for the same radio. Signed-off-by:
Chun-Yeow Yeoh <yeohchunyeow@gmail.com> Signed-off-by:
-
Johannes Berg authored
Add the signal strength (in dBm only for now) to frames that are received via nl80211's various frame APIs. Signed-off-by:
Kalle Valo <kvalo@qca.qualcomm.com> Signed-off-by:
-
- 05 Mar, 2012 16 commits
-
-
Arend van Spriel authored
Tidying up some debug statements in brcms_c_ampdu_dotxstatus_complete() that got broken strings to satisfy checkpatch, but the rules changed. Reviewed-by:
Pieter-Paul Giesberts <pieterpg@broadcom.com> Reviewed-by:
Alwin Beukers <alwin@broadcom.com> Signed-off-by:
Arend van Spriel <arend@broadcom.com> Signed-off-by:
-
Arend van Spriel authored
As indicated in [1] on netdev mailing list drivers should not block on the init_module() syscall. This patch defers the actual driver registration to a workqueue so the init_module() syscall can complete without delay. [1] http://article.gmane.org/gmane.linux.network/217729/Reviewed-by:
-
Arend van Spriel authored
As indicated in [1] on netdev mailing list drivers should not block on the init_module() syscall. This patch defers the actual driver registration to a workqueue so the init_module() syscall can complete without delay. [1] http://article.gmane.org/gmane.linux.network/217729/Reviewed-by:
Franky (Zhenhui) Lin <frankyl@broadcom.com> Signed-off-by:
-
Arend van Spriel authored
The module init function of brcmfmac calls init functions for SDIO and USB doing driver registration. This patch removes terminating the module init when a driver registration for one host interface fails. Reviewed-by:
-
Arend van Spriel authored
The usb code defines a structure for counting statistics. However, it should use the statistics entry provided in brcmf_bus as that is exposed to the net_device. The usb private statistics counter only remains with counters for control packets between driver and usb device. Reviewed-by:
-
Arend van Spriel authored
Several fields in this structure are only written once or not used at all. Remaining two fields have been moved and brcmf_usb_attrib definition has been removed. Reviewed-by:
Kan Yan <kanyan@broadcom.com> Signed-off-by:
-
Arend van Spriel authored
In brcmf_usb_up() the variable devinfo was checked for being a NULL pointer, but this can not happen. Also the check was done after dereferencing the pointer. This patch removes the check. Reviewed-by:
-
Larry Finger authored
When driver rtl8192cu is used with the debug level set to 3 or greater, the result is "sleeping function called from invalid context" due to an rcu_read_lock() call in the DM refresh routine in driver rtl8192c. This lock is not necessary as the USB driver does not use the struct being protected, thus the lock is set only when a PCI interface is active. This bug is reported in https://bugzilla.kernel.org/show_bug.cgi?id=42775. Reported-by:
Ronald Wahl <ronald.wahl@raritan.com> Tested-by:
-
Larry Finger authored
In usb.c, routine _rtl_usb_init_sw() always returns 0, and rtl_usb_probe() never checks the returned value. Thus we can change _rtl_usb_init_sw() to a void routine. In addition, remove the underscore at the beginning of the name of the routine. Signed-off-by:
-
Larry Finger authored
In the unlikely event of a misread of the USB end point count, the driver generates a divide fault. To prevent this, add a check of the value returned by _rtl_usb_init(). In addition, add some logging to indicate why the condition occurred. Signed-off-by:
-
Johannes Berg authored
When associating and particularly when disassociating there's no need to notify the driver about changes with multiple calls to bss_info_changed, we should combine the QoS enabling/disabling into the same call as otherwise the driver could get confused about QoS suddenly getting disabled while connected. Signed-off-by:
-
Johannes Berg authored
There's no need to hardcode a subset of the radiotap header for cooked monitor receive, we can just reuse the normal monitor mode radiotap code. This simplifies the code and extends the information available on cooked monitor interfaces. Signed-off-by:
-
Vasanthakumar Thiagarajan authored
This patch adds an attribute, NL80211_ATTR_INACTIVITY_TIMEOUT, to set the inactivity timeout which can be used to remove the station in AP mode. This can be passed in NL80211_CMD_START_AP and used by the drivers which have AP MLME in firmware but don't support get_station() properly. To disable inactivity timer in userspace, wpa_s for example, there is a new flag, NL80211_FEATURE_INACTIVITY_TIMER, in nl80211_feature_flags through which drivers can register their capability to use the inactivity timeout to free the stations. Signed-off-by:
Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com> Acked-by:
Johannes Berg <johannes@sipsolutions.net> Signed-off-by:
-
Chun-Yeow Yeoh authored
RANN, PREP and PERR propagation should happen only if the dot11MeshForwarding is true. Besides, data frame should not be forwarded if dot11MeshForwarding is false. This redundant checking is necessary to avoid the broadcasted ARP breaking the non-forwarding rule. Signed-off-by:
-
Felix Fietkau authored
For A-MPDU rx it makes sense to only process the signal strength once per aggregate instead of once per subframe. Additonally, some hardware (e.g. Atheros) only provides valid signal strength information for the last subframe. Signed-off-by:
Felix Fietkau <nbd@openwrt.org> Signed-off-by:
-
Johannes Berg authored
I forgot to update the kernel-doc in my patch to redesign AP mode APIs, fix that now. Signed-off-by:
-
