- 22 May, 2009 1 commit
-
-
Johannes Berg authored
The requirement for wireless stats to be atomic is now mostly artificial since we hold the rtnl _and_ the dev_base_lock for iterating the device list. Doing that is not required, just the rtnl is sufficient (and the rtnl is required for other reasons outlined in commit "wext: fix get_wireless_stats locking"). This will fix http://bugzilla.kernel.org/show_bug.cgi?id=13344 and make things easier for drivers. Signed-off-by:
Johannes Berg <johannes@sipsolutions.net> Signed-off-by:
John W. Linville <linville@tuxdriver.com>
-
- 20 May, 2009 39 commits
-
-
Johannes Berg authored
Currently, we will ask the driver to configure right away when somebody changes the desired BSSID. That's totally strange because then we will configure the driver without even knowing whether the BSS exists. Change this to only configure the BSSID when associated, and configure a zero BSSID when not associated. As a side effect, this fixes an issue with the iwlwifi driver which doesn't implement sta_notify properly and uses the BSSID instead and gets very confused if the BSSID is cleared before we disassociate, which results in the warning Marcel posted [1] and iwlwifi bug 1995 [2]. [1] http://thread.gmane.org/gmane.linux.kernel.wireless.general/32598 [2] http://www.intellinuxwireless.org/bugzilla/show_bug.cgi?id=1995 Cc: Marcel Holtmann <marcel@holtmann.org> Signed-off-by:
-
Luis R. Rodriguez authored
Device drivers using wiphy_apply_custom_regulatory() want some regulatory settings applied to their wiphy, if no bands were configured on the wiphy then something went wrong. Signed-off-by:
Luis R. Rodriguez <lrodriguez@atheros.com> Signed-off-by:
-
Luis R. Rodriguez authored
This is always discarded anyway but lets just set this to our safest lowest. Signed-off-by:
-
Johannes Berg authored
When I fixed the crypto bit I must have done the negative test only -- it is quite clearly impossible to find _any_ IBSS to join with the parameters put the wrong way around. Signed-off-by:
-
Luis R. Rodriguez authored
ath_regd_init() needs to be called with the wiphy already properly set with the bands. Without this the custom regulatory settings were not taking effect, and the device would get the default channel settings from ath9k_[25]ghz_chantable. Reported-by:
-
Johannes Berg authored
When the genIE hasn't changed there's no reason to kick the state machine since it won't be able to do anything new -- doing this decreases the useless work we do for reassociating because if we do kick the state machine it will try to find a usable BSS but there might not be one because wpa_supplicant will only change the BSSID a little later. In a sense this is a workaround for userspace behaviour, but on the other hand userspace cannot really keep track of what the kernel currently has for genIE since any process could have changed that while wpa_supplicant wasn't looking. Signed-off-by:
-
Alexandre Becholey authored
Report status unknown as if there were successfully transmitted. This will avoid hostapd to disassociate because it doesn't understand what a status unknown is. Signed-off-by:
Alexandre Becholey <alexandre.becholey@epfl.ch> Signed-off-by:
-
Jouni Malinen authored
When updating the duration field for TX frames, skip the update for PS-Poll frames that use this field for other purposes (AID). Signed-off-by:
Jouni Malinen <jouni.malinen@atheros.com> Signed-off-by:
-
Jouni Malinen authored
If the AP includes our AID in the TIM IE, we need to process the Beacon frame as far as PS is concerned (send PS-Poll or nullfunc data with PM=0). The previous code skipped this in cases where the CRC value did not change and it would not change if the AP continues including our AID in the TIM.. There is no need to count the crc32 value for directed_tim with this change, so we can remove that part. In order not to change the order of operations (i.e., update WMM parameters prior to sending PS-Poll), the CRC match is checked twice as only after the PS processing step, the rest of the function is skipped if nothing changed in the Beacon. Signed-off-by:
-
Jouni Malinen authored
mac80211 does not set PM field for normal data frames, so we need to update that based on the current PS mode when using PS-Poll (timeout=0) power save mode. This allows the AP to remain in sync with our PS state. However, there is still a potential race condition between PS state changes when multiple TX queues are used and nullfunc and PS-Poll frames use different queue. That corner case may need to be handled separately by changing which queue is used either in ath9k or mac80211. Signed-off-by:
-
Jouni Malinen authored
We must make sure the chip is awake when changing the RX filter parameters. This could have caused problems, e.g., when changing the interface to promiscuous mode while in sleep mode. Signed-off-by:
-
Jouni Malinen authored
If the chip complains about TSF sync, make sure we remain awake to sync with the next Beacon frame. In theory, this should not be needed since we are currently trying to receive all Beacon frames, anyway, better have this code ready should we ever change that. Signed-off-by:
-
Jouni Malinen authored
When the chip is in sleep mode, there is no point trying to calibrate the radio since it will just results in incorrect values being read from registers and other potential issues. In addition, if we actually start processing calibrate, do not allow the chip to be put into sleep until we have completed the calibration step. Signed-off-by:
-
Jouni Malinen authored
When using timeout=0 (PS-Poll) with mac80211, the driver will need to wake up for TX requests and remain awake until the TX has been completed (ACK received or timeout) or until the buffer frame(s) have been received (in case the TX is for a PS-Poll frame). Signed-off-by:
-
Johannes Berg authored
Moving information from config_interface to bss_info_changed removed struct ieee80211_if_conf which the documentation still refers to, additionally there's one kernel-doc description too much and one other missing, fix all this. Signed-off-by:
-
Ivo van Doorn authored
Signed-off-by:
Ivo van Doorn <IvDoorn@gmail.com> Signed-off-by:
-
Pavel Roskin authored
If CONFIG_CFG80211_DEBUGFS is enabled and CONFIG_MAC80211_DEBUGFS is not, compilation fails in net/wireless/debugfs.c: net/wireless/debugfs.c: In function 'cfg80211_debugfs_drv_add': net/wireless/debugfs.c:117: error: 'struct cfg80211_registered_device' has no member named 'debugfs' The debugfs filed is needed if and only if CONFIG_CFG80211_DEBUGFS is enabled, so use that instead of CONFIG_MAC80211_DEBUGFS. Signed-off-by:
Pavel Roskin <proski@gnu.org> Signed-off-by:
-
Bob Copeland authored
If we are trying to interpolate a curve with slope == 0, the return value will always be the y-coordinate. In this code we are looping until we reach a minimum y-coordinate on a line, which in the 0-slope case can never happen, thus the loop never terminates. The PCDAC steps come from the EEPROM and should never be equal, but we should gracefully handle that case, so warn and bail out. Reported-by:
Steven Rostedt <rostedt@goodmis.org> Signed-off-by:
Bob Copeland <me@bobcopeland.com> Signed-off-by:
-
Luis R. Rodriguez authored
There is a race on access to last_request and its alpha2 through reg_is_valid_request() and us possibly processing first another regulatory request on another CPU. We avoid this improbably race by locking with the cfg80211_mutex as we should have done in the first place. While at it add the assert on locking on reg_is_valid_request(). Cc: stable@kernel.org Signed-off-by:
-
Luis R. Rodriguez authored
This has no functional change, but it will make the race fix easier to spot in my next patch. Cc: stable@kernel.org Signed-off-by:
-
Luis R. Rodriguez authored
This has no functional change except we save a kfree(rd) and allows us to clean this code up a bit after this. We do avoid an unnecessary kfree(NULL) but calling that was OK too. Cc: stable@kernel.org Signed-off-by:
-
Wei Yongjun authored
This patch fix to set dev->broadcast correctly, since dev->broadcast is defined as: unsigned char broadcast[MAX_ADDR_LEN]; Signed-off-by:
Wei Yongjun <yjwei@cn.fujitsu.com> Signed-off-by:
-
Alexander Beregalov authored
Fix this build error: drivers/staging/agnx/pci.c: In function 'agnx_config_interface': drivers/staging/agnx/pci.c:315: error: dereferencing pointer to incomplete type .. drivers/staging/agnx/pci.c:424: error: unknown field 'config_interface' specified in initializer config_interface was removed in 2d0ddec5 ("mac80211: unify config_interface and bss_info_changed") Signed-off-by:
Alexander Beregalov <a.beregalov@gmail.com> Signed-off-by:
-
Vasanthakumar Thiagarajan authored
We do not need to do this in ath_isr() and it looks like the modified version ends up being more stable as far as being able receive beacon frames is concerned. Furthermore, this reduces need to move between AWAKE and NETWORK SLEEP states when processing some unrelated interrupts. Signed-off-by:
Vasanthakumar Thiagarajan <vasanth@atheros.com> Signed-off-by:
-
Johannes Berg authored
Some applications using wireless extensions expect to be able to remove a key that doesn't exist. One example is wpa_supplicant which doesn't actually change behaviour when running into an error while trying to do that, but it prints an error message which users interpret as wpa_supplicant having problems. The safe thing to do is not change the behaviour of wireless extensions any more, so when the driver reports -ENOENT let the wext bridge code return success to userspace. To guarantee this, also document that drivers should return -ENOENT when the key doesn't exist. Signed-off-by:
-
Daniel Wagner authored
Currently all radios receive all traffic on the simulated air if they are tuned to the same channel. This patch introduces the concept of grouping, which allows to assign a radio to certain group. Only radios in the same group can 'see' each other. Each bit in /debug/ieee80211/phy*/hwsim/group represents one group. By default all radios belong to the same group "1", e.g. bit 1 is set. Additionally a radio can belong to several groups. Signed-off-by:
Daniel Wagner <wagi@monom.org> Signed-off-by:
-
Bob Copeland authored
ath5k only generated the beacon when bss_info_changed() was called, but for AP mode this is not enough, because the TIM IE would never get updated and consequently PS mode clients wouldn't know about buffered frames. Instead, get a new beacon on every SWBA interrupt. Signed-off-by:
-
Max Filippov authored
Drop test for FW_STATE_RESET in p54spi_work as fw_state is never assigned this value. Signed-off-by:
Roel Kluin <roel.kluin@gmail.com> Signed-off-by:
Max Filippov <jcmvbkbc@gmail.com> Signed-off-by:
-
Johannes Berg authored
Process beacon change even if the BSSID doesn't change at the same time. Also fix what I think is a small locking error in b43legacy, there's a spin_unlock_irqrestore that looks out of place. Signed-off-by:
-
Max Filippov authored
Under high load first data word, read after available data size is sometimes lost in p54spi_rx. It seems to depend on frequency of interrupts and latency of data read request relatively to 'data available' interrupt. The worst consequence of this bug is loss of packet transmission acknowledgement, which in turn causes overflow of tx queues and permanent link loss. Read data size and first data word in one SPI transaction. No packets from LMAC should have length less than 1 word, so this shouldn't interfere with the next read transaction. Also call p54spi_sleep if p54spi_wake succeeded. Signed-off-by:
-
Max Filippov authored
Put chip into sleep state, once it's been awaken. Also, propagate error code to the caller. Signed-off-by:
-
Max Filippov authored
Return whether wakeup operation succeeded. Make use of this return value. Signed-off-by:
-
Max Filippov authored
When SPI write of odd length is requested, p54spi_write splits it into two parts: one for all data, except the last byte, and one for last byte and padding byte. Unfortunately, the length of first part is not amended. It works because all meaningful bytes have proper value and the last byte of odd length SPI write transaction is ignored. p54spi_work has dummy HOST_INTERRUPTS register read at the end. Drop it, as its result is not used and it has no side effects. Signed-off-by:
-
Max Filippov authored
Host is not allowed to modify DMA_WRITE_CTRL register if bit HOST_ALLOWED in it is not set. Wait for HOST_ALLOWED first. Also get rid of timeout in p54spi_wait_bit as it's been playing a role of workaround for such an incorrect register access. Signed-off-by:
-
Julia Lawall authored
Eliminate direct accesses to the driver_data field. cf 82ab13b26f15f49be45f15ccc96bfa0b81dfd015 The semantic patch that makes this change is as follows: (http://www.emn.fr/x-info/coccinelle/) // <smpl> @@ struct device *dev; expression E; type T; @@ - dev->driver_data = (T)E + dev_set_drvdata(dev, E) @@ struct device *dev; type T; @@ - (T)dev->driver_data + dev_get_drvdata(dev) // </smpl> Signed-off-by:
Julia Lawall <julia@diku.dk> Acked-by:
Zhu Yi <yi.zhu@intel.com> Signed-off-by:
-
David Kilroy authored
This allows drivers to use a const pointer as the privid without a cast. Signed-off-by:
David Kilroy <kilroyd@gmail.com> Acked-by:
-
David Kilroy authored
This allows drivers to mark their cfg80211_ops tables const. Signed-off-by:
David Kilroy <kilroyd@googlemail.com> Acked-by:
-
Tomas Winkler authored
This patch is one of the incremental steps for unifying iwl_station_entry for all HWs, i.e. removing of iwl3945_station_entry This patch drops iwl3945_tid_data and use iwl_tid_data instead. Signed-off-by:
Tomas Winkler <tomas.winkler@intel.com> Signed-off-by:
Reinette Chatre <reinette.chatre@intel.com> Signed-off-by:
-
Tomas Winkler authored
This patch replaces struct iwl3945_hw_key by struct iwl_hw_key. It's not used directly with any host command therefore removal is trivial Signed-off-by:
-
