- 29 Jan, 2009 40 commits
-
-
Samuel Ortiz authored
Most of the power (not TX power, but power management) structures and definitions are duplicated accross iwl-power.h and iwl-3945.h. We should try to only use the iwl header. Signed-off-by:
Samuel Ortiz <samuel.ortiz@intel.com> Acked-by:
Abhijeet Kolekar <abhijeet.kolekar@intel.com> Signed-off-by:
Reinette Chatre <reinette.chatre@intel.com> Signed-off-by:
John W. Linville <linville@tuxdriver.com>
-
Samuel Ortiz authored
The duplicated LED definitions prevent one from including iwl-dev.h from the 3945 specific C files. Moreover, we are sharing many definitions between iwl-3945-led.h and iwl-led.h, so let's just use the iwl one. Note that this file will get more cleanups once we share a common iwl_priv structure. Signed-off-by:
-
Samuel Ortiz authored
A lot of rate relates definition are shared between iwl-3945-rs.h and iwl-agn-rs.h. Let's just use the agn version, and add the 3945 specific constants there. Signed-off-by:
-
Samuel Ortiz authored
SRAM addresses are different for 3945, 4065, and 5000, let's give them different names. Also, the RSSI_OFFSET is different for 3945 and 4965, thus they should be named differently. Signed-off-by:
-
Samuel Ortiz authored
The corresponding iwl structure is identical. Signed-off-by:
-
Samuel Ortiz authored
As DRV_NAME is defined in 2 different header files, including both is not possible. This patch defines this constant from iwl3945-base.c and iwl-agn.c. It also redefines the IWL_ERROR and IWL_WARNING macros to use dev_printk, as the IWL_DEBUG_* macros do. Signed-off-by:
-
Samuel Ortiz authored
At the cost of adding a debug_level field to iwl3945_priv, we are now able to get rid of iwl-3945-debug.h, and use iwl-debug.h instead. Signed-off-by:
-
Tomas Winkler authored
This patch remove iwl-3945-commands.h eliminating duplicated and moving all definitions to iwl-commands.h Signed-off-by:
Tomas Winkler <tomas.winkler@intel.com> Acked-by:
Samuel Ortiz <sameo@linux.intel.com> Signed-off-by:
-
Tomas Winkler authored
This patch replaces union tsf with u64 This also allows to use iwl_error_res and iwl_rxon_time_cmd instead of 3945 structures Signed-off-by:
-
Tomas Winkler authored
This patch remove trivial (renames) commands and defines from iwl-3945-commands.h Signed-off-by:
-
Tomas Winkler authored
This patch uses iwl-commands.h also for iwl3945 more clean ups are required but this get to stage where it compiles cleanly. Most massive changes are in spectrum and power managment. Signed-off-by:
-
Winkler, Tomas authored
This patch removes current_rate from station_entry it was write only variable Signed-off-by:
-
Tomas Winkler authored
This patch unfolds includes of iwl-3945-commands.h Signed-off-by:
-
Tomas Winkler authored
This patch moves 3945 definitions into iwl-3945-fh.h It renames FH_ to FH39 to help inclusion of 3945 into iwlcore framework Signed-off-by:
-
Vivek Natarajan authored
Since the station info is flushed before calling set_disassoc in ieee80211_stop, the power save timer is never cancelled when the driver is unloaded. Hence the timer cancellation has to be done in ieee80211_stop itself. Signed-off-by:
Vivek Natarajan <vnatarajan@atheros.com> Signed-off-by:
-
Vivek Natarajan authored
This patch enables mac80211 to send a null frame and also to check for tim in the beacon if dynamic power save is enabled. Signed-off-by:
-
Vivek Natarajan authored
a) hw_config() should not be called from siwpower() for the drivers which do not support dynamic powersave. b) IEEE80211_HW_NO_STACK_DYNAMIC_PS needs to be verified in set_associated() also before enabling the power save timers. Signed-off-by:
-
Vivek Natarajan authored
When a null data frame is generated from mac80211, it goes through master_start_xmit and not through subif_start_xmit. Hence for the power save timer to be triggered while sending this null data frame also, the timer has to be reset from master_start_xmit. Signed-off-by:
-
Vasanthakumar Thiagarajan authored
As TKIP is not updated to new security needs which arise when TKIP is used to encrypt A-MPDU aggregated data frames, IEEE802.11n does not allow any cipher other than CCMP (Which has new extensions defined) as pairwise cipher between HT peers. When such configuration (TKIP/WEP in HT) is forced, we still associate in non-HT mode (11a/b/g). Signed-off-by:
Vasanthakumar Thiagarajan <vasanth@atheros.com> Acked-by:
Johannes Berg <johannes@sipsolutions.net> Signed-off-by:
-
David Kilroy authored
Added mappings for FHSS, DSSS and OFDM channels - with macros to point HR DSSS and ERP to the DSSS mappings. Currently just static inline functions. Use the new functions in the older fullmac drivers. This eliminates a number of const static buffers and removes a couple of range checks that are now redundant. Signed-off-by:
David Kilroy <kilroyd@googlemail.com> Acked-by:
Arnaldo Carvalho de Melo <acme@redhat.com> Acked-by:
Richard Farina <sidhayn@gmail.com> Acked-by:
Jeroen Vreeken <pe1rxq@amsat.org> Signed-off-by:
-
Nick Kossifidis authored
* Don't scale power values on RF5111 EEPROMs because they get out of bounds (power is u8, so multiplying power by 50 is too much and there is no reason to do so -we don't do it on other chips anyway-). HAL does it as a technique to handle 0.5 dbm steps but i believe it's not the right thing to do and certainly not the right place to do it. We 'll work this out on interpolation code for all chips (0.5 or 0.25 steps etc) in a generic way. Signed-Off-by:
Nick Kossifidis <mickflemm@gmail.com> Signed-off-by:
-
Winkler, Tomas authored
This patch kills retry_rate in sysfs for iwlagn. It's not used. Signed-off-by:
Zhu Yi <yi.zhu@intel.com> Signed-off-by:
-
Winkler, Tomas authored
This patch moves priv->status sysfs entry to debugfs. It is for debugging only anyway. Signed-off-by:
-
Christian Lamparter authored
This patch enables frame injection in monitor mode for all p54 devices. As a result, any user can finally use the aircrack-ng suite out of the box. e.g: aireplay-ng --test wlan0 Trying broadcast probe requests... Injection is working! Found 1 AP Trying directed probe requests... XX:XX:XX:XX:XX:XX - channel: i - 'SSID' Ping (min/avg/max): 1.536ms/3.193ms/4.377ms Power: 193.00 30/30: 100% Signed-off-by:
Christian Lamparter <chunkeey@web.de> Signed-off-by:
-
Christian Lamparter authored
This patch introduce new shiny named labels for our 8 (4 - on old firmware) queues. Signed-off-by:
-
Nick Kossifidis authored
*Read misc2...6 values from eeprom since we want to use them (fixes wrong power calibration info offset on RF2413+ chips) *Initialize num_piers to 0 for RF2413 chips (note that we read 2GHz frequency piers while reading mode sections, we have to ignore them -usualy they are 0xff anyway but during my tests i got a 1 on b mode with no data- and use the newer eemap. *Add some more comments (please forgive my poor English ;-( ) and some minor code cleanup *Tested on 2425 and 2112 and has the same data with ath_info (i wrote some debug code on debug.c to print everything like ath_info but i haven't tested it yet on 5111 and it's full of > 80 col lines, if anyone wants to play with it let me know). Signed-Off-by:
Felix Fietkau <nbd@openwrt.org> Signed-off-by:
-
Ivo van Doorn authored
Signed-off-by:
Ivo van Doorn <IvDoorn@gmail.com> Signed-off-by:
-
Ivo van Doorn authored
Improve error message reporting when a frame was received with unknown rate. Instead of using the boolean check if the frame is supposed to be a PLCP value or not, we should add a new mask (RXDONE_SIGNAL_MASK) which returns the type identification for a signal value (i.e. PLCP). At the moment we only have 2 different types, but more will arrive when support for 11n is added. Signed-off-by:
-
Ivo van Doorn authored
Signed-off-by:
-
Ivo van Doorn authored
The flag ENTRY_TXD_OFDM_RATE isn't flexible enough to indicate which rate modulation should be used for a frame. This will become a problem when 11n support is added. Remove the flag and replace it with an enum value which can better indicate the exact rate modulation. Signed-off-by:
-
Ivo van Doorn authored
Some functions have grown rapidly in size over the last time, some of those functions (like the rt2x00queue_create_tx_descriptor) will further increase in size soon, so it is best to start cutting it into logical pieces. Signed-off-by:
-
Ivo van Doorn authored
The 2 bits in EEPROM_NIC_TX_RX_FIXED each influence a different antenna. We might as well split the definition and directly read the correct bit. Signed-off-by:
-
Ivo van Doorn authored
WDS support should be very easy to handle, mac80211 handles everything for us, so all that is needed is to set the support flags and handle it in the add_interface() callback. Signed-off-by:
-
Ivo van Doorn authored
CONFIG_CRYPTO_COPY_IV is a bad name since it is part of the driver requirements instead of a configuration option. Signed-off-by:
-
Ivo van Doorn authored
Some very rare Ralink USB hardware exists which features the RFKILL switch on the USB stick. This patch adds the EEPROM check function to see if RFKILL is supported and the polling function to rt2500usb and rt73usb in order to support RFKILL for that hardware. Signed-off-by:
-
Andrey Yurovsky authored
This adds initial support for Mesh Point mode. For this we tell mac80211 that we support NL80211_IFTYPE_MESH_POINT. We also need to send beacons. mac80211 will configure our RX filter accordingly. Signed-off-by:
Andrey Yurovsky <andrey@cozybit.com> Signed-off-by:
-
Ivo van Doorn authored
Restrict drivers to only access link_qual structure during link tuning. The contents of these fields are for the drivers and all fields are allowed to be changed to values the driver considers correct. This means that some fields need to be moved outside of this structure to restrict access only to rt2x00link itself. This allows some code to be moved outside of the rt2x00.h header and into rt2x00link.c. Signed-off-by:
-
Ivo van Doorn authored
The link_tuner() function will always call bbp_read() at the start of the function. Because this is an indirect register access has some costs attached to it (especially for USB hardware). We already store the value read from the register into the vgc_level value inside the link structure. Instead of reading from the register we can read that field directly and base the tuner on that value. This reduces the time the registers are locked with the csr_mutex and speeds up the link_tuner processing. Signed-off-by:
-
Ivo van Doorn authored
Move link and antenna tuning into a seperate file named rt2x00link.c, this makes the interface to the link tuner a lot cleaner. Signed-off-by:
-
Ivo van Doorn authored
Listen to IEEE80211_CONF_PS to determine if the device should drop into powersaving mode. This feature depends on the dynamic power save functionality in mac80211. Signed-off-by:
-
