OSDN Git Service

uclinux-h8/linux.git
14 years agoath9k: Allow PSPOLL only when the interface is configured in AP mode
Senthil Balasubramanian [Fri, 18 Sep 2009 09:36:07 +0000 (15:06 +0530)]
ath9k: Allow PSPOLL only when the interface is configured in AP mode

Signed-off-by: Senthil Balasubramanian <senthilkumar@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: Update initvals
Vasanthakumar Thiagarajan [Fri, 18 Sep 2009 09:34:51 +0000 (15:04 +0530)]
ath9k: Update initvals

* Move 0xa274 and 0xa27c to the top of tx_gain table.
* Update initvals to fix random failure of noise floor calibration.

Signed-off-by: Vasanthakumar Thiagarajan <vasanth@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: Fix RTC reset for AR5416
Sujith [Fri, 18 Sep 2009 09:34:27 +0000 (15:04 +0530)]
ath9k: Fix RTC reset for AR5416

For AR5416 chipsets, clearing RTC_RESET_EN when setting
the chip to SLEEP mode results in high power consumption.
This patch fixes this issue by not clearing it for AR5416.

Signed-off-by: Sujith <Sujith.Manoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: Update INI release for AR9287
Sujith [Fri, 18 Sep 2009 09:34:07 +0000 (15:04 +0530)]
ath9k: Update INI release for AR9287

If the current channel is between 2412 and 2472 MHz and if the channel is
changing to 2484 MHz, then the registers 0xa1f4, 0xa1f8 and 0xa1fc need to be
programmed to the "japan_2484" values.  Conversely, if the current channel
is 2484 MHz and if the channel is changing to one between 2412 and 2472 MHz, then
the three registers need to be programmed to the "normal" values.

This is needed for compliance with Japanese regulatory requirements.

Signed-off-by: Sujith <Sujith.Manoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: Add Calibration checks
Vivek Natarajan [Fri, 18 Sep 2009 09:33:42 +0000 (15:03 +0530)]
ath9k: Add Calibration checks

* Prevent divide-by-zero errors in IQ Calibration.
* Do not run temperature compensation if initPDADC or currPDADC is zero.
* Also, introduce a separate function for handling OLC for AR9287.

Signed-off-by: Vivek Natarajan <vnatarajan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoar9170: handle overflow in tsf_low register during get_tsf
Joerg Albert [Tue, 15 Sep 2009 21:27:53 +0000 (23:27 +0200)]
ar9170: handle overflow in tsf_low register during get_tsf

ar9170_op_get_tsf: handle a carry from TSF_L into TSF_H
by reading TSF_H twice.

Signed-off-by: Joerg Albert <jal2@gmx.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoar9170: add heavy clip handling
Joerg Albert [Tue, 15 Sep 2009 20:23:06 +0000 (22:23 +0200)]
ar9170: add heavy clip handling

add heavy clip handling for 2.4GHz only (similar to the vendor driver).

Signed-off-by: Joerg Albert <jal2@gmx.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoar9170: fixed coding style, moved define
Joerg Albert [Tue, 15 Sep 2009 20:20:22 +0000 (22:20 +0200)]
ar9170: fixed coding style, moved define

This patch fixes some coding style issues and moves MAX_RATE_POWER into hw.h

Signed-off-by: Joerg Albert <jal2@gmx.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: add module firmware info for 1000 series
Huaxu Wan [Fri, 25 Sep 2009 21:24:28 +0000 (14:24 -0700)]
iwlwifi: add module firmware info for 1000 series

The module firmware information of 1000 series is missing from iwlagn.

Signed-off-by: Huaxu Wan <huaxu.wan@linux.intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: support idle for 6000 series hw
Johannes Berg [Fri, 25 Sep 2009 21:24:27 +0000 (14:24 -0700)]
iwlwifi: support idle for 6000 series hw

Using powersave while idle saves a lot of power, but
we've had problems with this on some cards (5150 has
been reported to be problematic). However, on the new
6000 series we're seeing no problems, so for now let
that hardware benefit from idle mode, we can look at
the problems with other hardware one by one and then
enable those once we figure out the problems.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: Chain Noise Calibration for 6000 series
Wey-Yi Guy [Fri, 25 Sep 2009 21:24:26 +0000 (14:24 -0700)]
iwlwifi: Chain Noise Calibration for 6000 series

Adding support of Chain Noise Calibration for 6000 series NICs.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: default to using all chains
Johannes Berg [Fri, 25 Sep 2009 21:24:25 +0000 (14:24 -0700)]
iwlwifi: default to using all chains

When instructing the microcode to use just a single
chain when we have power saving enabled, we should
also tell the AP that we are doing SM powersave.
However, using a single chain doesn't actually have
any power saving advantage while idle -- measurements
show that the power consumption is no different when
using one vs. two or three chains.

Therefore, always instruct the microcode to use all
chains.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: fix a typo
Johannes Berg [Fri, 25 Sep 2009 21:24:24 +0000 (14:24 -0700)]
iwlwifi: fix a typo

We never have four chains, but let's fix the typo
while we noticed it. You count 0, 1, 2, 3, not
0, 1, 2, 4 :)

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: add LED mode to support different LED behavior
Wey-Yi Guy [Thu, 17 Sep 2009 17:43:54 +0000 (10:43 -0700)]
iwlwifi: add LED mode to support different LED behavior

Today's implementation allow LED to blink based on the traffic
condition. We introduce an additional LED mode that reflects the RF
state.

The supported LED modes after this are:
  IWL_LED_BLINK (current/default) - blink rate based on current Tx/Rx
                                    traffic
  IWL_LED_RF_STATE (new) -
    LED OFF: No power/RF disabled, the LED is emitting no light
    LED ON:  Powered/RF enabled, the LED is emitting light
             in a stable non-flashing state.

In order to provide the flexibility to support different LED
behavior per user/system preference we add "led_mode" iwlcore module
parameter.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: update PCI Subsystem ID for 6000 series
Wey-Yi Guy [Thu, 17 Sep 2009 17:43:53 +0000 (10:43 -0700)]
iwlwifi: update PCI Subsystem ID for 6000 series

Update PCI Subsystem ID for 60x0 series based on HW SKU. Adding new SKU
for "ABG" and "BG" only devices.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: update PCI Subsystem ID for 1000 series
Wey-Yi Guy [Thu, 17 Sep 2009 17:43:52 +0000 (10:43 -0700)]
iwlwifi: update PCI Subsystem ID for 1000 series

Update PCI Subsystem ID for 1000 series based on HW SKU. Adding new SKU
for "BG" only devices.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: send cmd to uCode to configure valid tx antenna
Wey-Yi Guy [Thu, 17 Sep 2009 17:43:51 +0000 (10:43 -0700)]
iwlwifi: send cmd to uCode to configure valid tx antenna

In order for uCode to select the valid antennas for transmit, driver
need to configure the allowed tx antennas through host command.

The TX_ANT_CONFIGURATION_CMD should be used for 5000 series and up

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: add aggregation tables to the rate scaling algorithm
Daniel C Halperin [Thu, 17 Sep 2009 17:43:50 +0000 (10:43 -0700)]
iwlwifi: add aggregation tables to the rate scaling algorithm

Current rate scale algorithm fluctuates between different MIMO modes fairly
rapidly, causing widely varying performance. These fluctuations occur because in
the rate_scale tables for expected throughput the values are not very different
for different modes.

However, when aggregation is turned on and MAC overhead is reduced, the
expected throughput for different MIMO modes grows and different modes have
vastly different performance. Add expected throughput tables for this case.

We also need to keep track of aggregation status per-station, so we add the
"is_agg" field to struct lq_sta.

Also includes cleanup of comments and variable names in/around the affected
code.

Signed-off-by: Daniel C Halperin <daniel.c.halperin@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: do not clear TX info flags when receiving BlockAckResponse
Daniel C Halperin [Thu, 17 Sep 2009 17:43:49 +0000 (10:43 -0700)]
iwlwifi: do not clear TX info flags when receiving BlockAckResponse

OR-in AMPDU flags rather than assigning them.  This lets the TX status for
aggregated packets be processed by rs_tx_status.

Signed-off-by: Daniel C Halperin <daniel.c.halperin@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: clean up rs_tx_status
Daniel C Halperin [Thu, 17 Sep 2009 17:43:48 +0000 (10:43 -0700)]
iwlwifi: clean up rs_tx_status

Cut down on redundant code, reorganize structure, and add/improve comments.
Should contain no functional changes.

Signed-off-by: Daniel C Halperin <daniel.c.halperin@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: use S_IRUGO and S_IWUSR in module parameters
Wey-Yi Guy [Thu, 17 Sep 2009 17:43:47 +0000 (10:43 -0700)]
iwlwifi: use S_IRUGO and S_IWUSR in module parameters

Instead of hardcode module parameter's permissions, use pre-defined.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: EEPROM version for 1000 and 6000 series
Wey-Yi Guy [Thu, 17 Sep 2009 17:43:46 +0000 (10:43 -0700)]
iwlwifi: EEPROM version for 1000 and 6000 series

Update EEPROM version requirement for 1000 and 6000 series of NIC
for EEPROM version verification.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: allow user change protection mechanism for HT
Wey-Yi Guy [Thu, 17 Sep 2009 17:43:45 +0000 (10:43 -0700)]
iwlwifi: allow user change protection mechanism for HT

Allow user to change protection mechanism for HT between RTS/CTS and
CTS-to-self through sysfs:

Show current protection mechanism for HT
   cat /sys/class/net/wlan0/device/rts_ht_protection

Change protection mechanism for HT (only allowed while not-associated)
CTS-to-self:
   echo 0 > /sys/class/net/wlan0/device/rts_ht_protection
RTS/CTS:
   echo 1 > /sys/class/net/wlan0/device/rts_ht_protection

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: Use RTS/CTS as the preferred protection mechanism for 6000 series
Wey-Yi Guy [Thu, 17 Sep 2009 17:43:44 +0000 (10:43 -0700)]
iwlwifi: Use RTS/CTS as the preferred protection mechanism for 6000 series

When 802.11g was introduced, we had RTS/CTS and CTS-to-Self protection
mechanisms. In an HT Beacon, HT stations use the "Operating Mode" field
in the HT Information Element to determine whether or not to use
protection.

The Operating Mode field has 4 possible settings: 0-3:
Mode 0: If all stations in the BSS are 20/40 MHz HT capable, or if the
BSS is 20/40 MHz capable, or if all stations in the BSS are 20 MHz HT
stations in a 20 MHz BSS
Mode 1: used if there are non-HT stations or APs using the primary or
secondary channels
Mode 2: if only HT stations are associated in the BSS and at least one
20 MHz HT station is associated.
Mode 3: used if one or more non-HT stations are associated in the BSS.

When in operating modes 1 or 3, and the Use_Protection field is 1 in the
Beacon's ERP IE, all HT transmissions must be protected using RTS/CTS or
CTS-to-Self.

By default, CTS-to-self is the preferred protection mechanism for less
overhead and higher throughput; but using the full RTS/CTS will better
protect the inner exchange from interference, especially in
highly-congested environment.

For 6000 series WIFI NIC, RTS/CTS protection mechanism is the
recommended choice for HT traffic based on the HW design.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: clarify and clean up chain settings
Johannes Berg [Fri, 11 Sep 2009 17:38:17 +0000 (10:38 -0700)]
iwlwifi: clarify and clean up chain settings

The chain settings we currently use in iwlwifi are
rather confusing -- and we also go by the wrong
settings entirely under certain circumstances. To
clean it up, create a new variable in the current
HT config -- single_chain_sufficient -- that tells
us whether we need more than one chain. Calculate
that based on the AP and operating mode (no IBSS
HT implemented -- so no need for multiple chains,
for station mode we use the AP's capabilities).

Additionally, since APs always send disabled SM PS
mode, keeping track of their sm_ps mode isn't very
useful -- doubly not so for our _own_ RX config
since that should depend on our, not the AP's, SM
PS mode.

Finally, document that our configuration of the
number of RX chains used is currently wrong when
in powersave (by adding a comment).

All together this removes the two remaining items
in struct iwl_ht_config that were done wrong there.

For the future, the number of RX chains and some
SM PS handshaking needs to be added to mac80211,
which then needs to tell us, and the new variable
current_ht_config.single_chain_sufficient should
also be calculated by mac80211.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: Daniel C Halperin <daniel.c.halperin@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: show NVM version in debugfs
Wey-Yi Guy [Fri, 11 Sep 2009 17:38:16 +0000 (10:38 -0700)]
iwlwifi: show NVM version in debugfs

Show version number along with dumping NVM data, the version information
being removed from sysfs, add it back to debugfs to help debugging.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: clean up ht config naming
Johannes Berg [Fri, 11 Sep 2009 17:38:15 +0000 (10:38 -0700)]
iwlwifi: clean up ht config naming

Daniel Halperin pointed out that the naming
here is rather inconsistent with at least 3
different names being used for one thing in
different contexts. Rename the struct to
iwl_ht_config (rather than iwl_ht_info) and
use ht_conf as a variable for it.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: Daniel C Halperin <daniel.c.halperin@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: Adjust blink rate to compensate Clock difference
Wey-Yi Guy [Fri, 11 Sep 2009 17:38:14 +0000 (10:38 -0700)]
iwlwifi: Adjust blink rate to compensate Clock difference

Adjust led blink rate to compensate on a MAC Clock difference on every
HW. Led blink rate analysis showed an average deviation of 0% on 3945,
5% on 4965 HW and 20% on 5000 series and up.
Need to compensate on the led on/off time per HW according to the
deviation to achieve the desired led frequency
The calculation is: (100-averageDeviation)/100 * blinkTime
For code efficiency the calculation will be:
    compensation = (100 - averageDeviation) * 64 / 100
    NewBlinkTime = (compensation * BlinkTime) / 64

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: clean up ht config a little
Johannes Berg [Fri, 11 Sep 2009 17:38:13 +0000 (10:38 -0700)]
iwlwifi: clean up ht config a little

is_ht can be bool instead of u8, and there's
no need to use IWL_CHANNEL_WIDTH_* constants
in supported_chan_width when that could just
be named is_40mhz instead.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: separate set_hw_params function for 6000 series
Wey-Yi Guy [Fri, 11 Sep 2009 17:38:11 +0000 (10:38 -0700)]
iwlwifi: separate set_hw_params function for 6000 series

Separate set_hw_params() function for 6000
series from 5000/1000 series because:
    1) 6000 series use different set of sensitivity range table
    2) 6000 series has different uCode image size

Also include the new sensitivity parameters needed by sensitivity
algorithm.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: separate nic_config for different NIC
Wey-Yi Guy [Fri, 11 Sep 2009 17:38:10 +0000 (10:38 -0700)]
iwlwifi: separate nic_config for different NIC

Different NIC has different requirements for configuration. Currently all
5000 series hardware and later share the same configuration function even
though they do not need the same configurations. Fix this by separating the
needed configuration actions for each hardware model.

.5000 series: L1-ASPM H/W bug work-around
              configure radio
              write CSR_HW_IF_CONFIG_REG for uCode use
              work-around for NIC get stuck after early PCIe power off

.1000 series: write CSR_HW_IF_CONFIG_REG for uCode use
              setting digital SVR for 1000 card to 1.32V

.6000 series: configure radio
              write CSR_HW_IF_CONFIG_REG for uCode use
              write CSR_GP_DRIVER_REG to indicate radio sku

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: remove un-supported eeprom parameters
Wey-Yi Guy [Fri, 11 Sep 2009 17:38:09 +0000 (10:38 -0700)]
iwlwifi: remove un-supported eeprom parameters

Remove few of the parameters not used and no longer valid in EEPROM.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: modify LED blink index table
Wey-Yi Guy [Fri, 11 Sep 2009 17:38:07 +0000 (10:38 -0700)]
iwlwifi: modify LED blink index table

Modify LED blink index table to include 1Mbps.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: make ath9k_common_ops const
Luis R. Rodriguez [Mon, 14 Sep 2009 09:09:38 +0000 (02:09 -0700)]
ath9k: make ath9k_common_ops const

As noted by Jiri.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoatheros: move bus ops to ath_common
Luis R. Rodriguez [Mon, 14 Sep 2009 07:55:09 +0000 (00:55 -0700)]
atheros: move bus ops to ath_common

This is the last part to make ath9k hw code core driver agnostic.
I believe ath9k_htc can now use use the hw code unmodified.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: move ATH9K_RSSI_BAD to hw.h
Luis R. Rodriguez [Mon, 14 Sep 2009 06:55:05 +0000 (23:55 -0700)]
ath9k: move ATH9K_RSSI_BAD to hw.h

mac.c is now core driver independent.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: clarify what hw code is and remove ath9k.h from a few files
Luis R. Rodriguez [Mon, 14 Sep 2009 06:39:31 +0000 (23:39 -0700)]
ath9k: clarify what hw code is and remove ath9k.h from a few files

hw code will be shared between ath9k and ath9k_htc.
Just a few more files are left to clean up, mark them as well.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: remove driver ASSERT, just use BUG_ON()
Luis R. Rodriguez [Mon, 14 Sep 2009 06:28:00 +0000 (23:28 -0700)]
ath9k: remove driver ASSERT, just use BUG_ON()

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: move ATH_AMPDU_LIMIT_MAX to hw.h
Luis R. Rodriguez [Mon, 14 Sep 2009 06:11:13 +0000 (23:11 -0700)]
ath9k: move ATH_AMPDU_LIMIT_MAX to hw.h

This is used by hw code.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: remove ath9k_ht_macmode
Luis R. Rodriguez [Mon, 14 Sep 2009 06:04:44 +0000 (23:04 -0700)]
ath9k: remove ath9k_ht_macmode

This is used just to determine how to program the MAC,
either for 20 MHz operation of 40 MHz so just use conf_is_ht40()

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: remove ath9k 25 MHz HT40 spacing stuff
Luis R. Rodriguez [Mon, 14 Sep 2009 05:05:04 +0000 (22:05 -0700)]
ath9k: remove ath9k 25 MHz HT40 spacing stuff

This was for supporting 25 MHz spacing for HT40, this is not used
as we use 20 MHz spacing instead for HT40 as per 802.11n. The hardware
is capable of it though so we leave the phymode definition and EEPROM
parsing for it. If some experimenter wants to work on this stuff stuff
you can add an extension enabling bool on ath_common and perhaps some
debugfs knob to enable it. Keep in mind you'll also need to update the
phymode with the AR_PHY_FC_DYN2040_EXT_CH which has been left on the
driver.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoatheros: move tx/rx chainmask to ath_common
Luis R. Rodriguez [Mon, 14 Sep 2009 04:07:07 +0000 (21:07 -0700)]
atheros: move tx/rx chainmask to ath_common

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoatheros: add common debug printing
Luis R. Rodriguez [Sun, 13 Sep 2009 09:42:02 +0000 (02:42 -0700)]
atheros: add common debug printing

ath9k uses this for now, ath9k_htc is expected to re-use this
as well. We lave ath5k as is, but it certainly can also be
converted later.

The ath9k module parameter and debugfs entry is kept.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: separate core driver and hw timer code
Luis R. Rodriguez [Sun, 13 Sep 2009 09:08:34 +0000 (02:08 -0700)]
ath9k: separate core driver and hw timer code

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoatheros: add ieee80211_hw to ath_common
Luis R. Rodriguez [Sun, 13 Sep 2009 07:03:27 +0000 (00:03 -0700)]
atheros: add ieee80211_hw to ath_common

Make use of it on hw code in ath9k to avoid
using the ath9k ath_softc.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoatheros: define shared bssidmask setting
Luis R. Rodriguez [Fri, 11 Sep 2009 00:52:45 +0000 (17:52 -0700)]
atheros: define shared bssidmask setting

Also make ath5k and ath9k use it, and share register definitions.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath5k: define ath_common ops
Luis R. Rodriguez [Thu, 10 Sep 2009 23:55:11 +0000 (16:55 -0700)]
ath5k: define ath_common ops

Only common ath read/write ops go through the common ops.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath5k: allocate ath5k_hw prior to initializing hw
Luis R. Rodriguez [Fri, 11 Sep 2009 01:04:47 +0000 (18:04 -0700)]
ath5k: allocate ath5k_hw prior to initializing hw

We can propagate better errors upon failed hw initialization,
and set up the ath_common structure for attach purposes. This
will become important once we start using the ath_common
for read/write ops.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoatheros/ath9k: add common read/write ops and port ath9k to use it
Luis R. Rodriguez [Thu, 10 Sep 2009 23:11:21 +0000 (16:11 -0700)]
atheros/ath9k: add common read/write ops and port ath9k to use it

In an effort to make hw code driver core agnostic read
and write operations are defined on the ath_common structure.
This patch adds that and makes ath9k use it. This allows
drivers like ath9k_htc to define its own read/write ops and
still rely on the same hw code. This also paves the way for
sharing code between ath9k/ath5k/ath9k_htc.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: Define bus agnostic bluetooth coex prep helper
Luis R. Rodriguez [Thu, 10 Sep 2009 19:12:23 +0000 (12:12 -0700)]
ath9k: Define bus agnostic bluetooth coex prep helper

We disable ASPM when enabling bluetooth coexistance. Disabling
ASPM is a bus specific operation. In the future other buses may
support bluetooth coexistance, an example is USB. To this end
move the current routine which disables ASPM into pci.c, and declare
it the PCI bt_coex_prep() helper. Additionally, since ASPM is
a PCI-Express primitive ensure we don't ever try to muck with ASPM
registers on non PCI-express devices.

This also cleans up hw.c to not include bus specific headers or
utilities.

Cc: Vasanthakumar Thiagarajan <vasanth@atheros.com>
Cc: Stephen Chen <stephen.chen@atheros.com>
Cc: Zhifeng Cai <zhifeng.cai@atheros.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath5k: move ath_common to ath5k_hw
Luis R. Rodriguez [Thu, 10 Sep 2009 18:20:57 +0000 (11:20 -0700)]
ath5k: move ath_common to ath5k_hw

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: move ath_common to ath_hw
Luis R. Rodriguez [Thu, 10 Sep 2009 18:08:14 +0000 (11:08 -0700)]
ath9k: move ath_common to ath_hw

This ensures that we can access common on hw related code
independent of the driver core.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath5k: initialize eeprom struct early on attach
Luis R. Rodriguez [Thu, 10 Sep 2009 17:57:00 +0000 (10:57 -0700)]
ath5k: initialize eeprom struct early on attach

This fixes this sparse warning:

  CHECK   drivers/net/wireless/ath/ath5k/attach.c
drivers/net/wireless/ath/ath5k/attach.c:288:42: warning: symbol 'ee' shadows an earlier one
drivers/net/wireless/ath/ath5k/attach.c:109:34: originally declared here

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath5k: use common curbssid, bssidmask and macaddr
Luis R. Rodriguez [Thu, 10 Sep 2009 17:51:33 +0000 (10:51 -0700)]
ath5k: use common curbssid, bssidmask and macaddr

The ah_sta_id was really being used as the macaddr.
ath5k still does not use the association ID now passed
up by mac80211, that can be fixed later.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoar9170: make use of common macaddr and curbssid
Luis R. Rodriguez [Thu, 10 Sep 2009 17:10:54 +0000 (10:10 -0700)]
ar9170: make use of common macaddr and curbssid

These are provided by ath_common.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoatheros/ath9k: move macaddr, curaid, curbssid and bssidmask to common
Luis R. Rodriguez [Thu, 10 Sep 2009 16:22:37 +0000 (09:22 -0700)]
atheros/ath9k: move macaddr, curaid, curbssid and bssidmask to common

These are common amongst ath9k and ath5k, so put them into the
common structure and make ath9k to use it. ar9170 can use macaddr,
and curbssid. We'll change ath5k and ar9170 separately.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: use ath9k_hw_write_associd() on reset
Luis R. Rodriguez [Thu, 10 Sep 2009 15:57:00 +0000 (08:57 -0700)]
ath9k: use ath9k_hw_write_associd() on reset

Use the already provided helper instead of rewriting the code
required in place.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: Use ath9k_hw_setbssidmask() on reset
Luis R. Rodriguez [Thu, 10 Sep 2009 15:54:56 +0000 (08:54 -0700)]
ath9k: Use ath9k_hw_setbssidmask() on reset

The same code was being implemented on reset for setting the bssidmask,
instead just use the already provided helper.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: make ath9k_hw_setbssidmask() and ath9k_hw_write_associd() use ath_hw
Luis R. Rodriguez [Thu, 10 Sep 2009 15:50:20 +0000 (08:50 -0700)]
ath9k: make ath9k_hw_setbssidmask() and ath9k_hw_write_associd() use ath_hw

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoatheros: use get_unaligned_le*() for bssid mask setting
Luis R. Rodriguez [Thu, 10 Sep 2009 05:43:17 +0000 (22:43 -0700)]
atheros: use get_unaligned_le*() for bssid mask setting

Historically some macro helpers have been users for this,
AR5K_LOW_ID() and AR5K_HIGH_ID(), use upstream unaligned
helpers instead. This applid to ath5k and ar9170. ath9k
already uses this.

Worth noting is ath5k uses an ah_sta_id but that is already
the MAC address combined with the associaiton ID, ah_sta_id
is really ETH_ALEN in size.

Cc: Bob Copeland <me@bobcopeland.com>
Cc: Nick Kossifidis <mick@madwifi-project.org>
Cc: Christian Lamparter <chunkeey@web.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath: move ath_bcast_mac to common header
Luis R. Rodriguez [Thu, 10 Sep 2009 05:19:26 +0000 (22:19 -0700)]
ath: move ath_bcast_mac to common header

This is used by both ath5k and ath9k to set the first bssid mask.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: rename driver core and hw power save helpers
Luis R. Rodriguez [Thu, 10 Sep 2009 04:10:09 +0000 (21:10 -0700)]
ath9k: rename driver core and hw power save helpers

ath9k_hw_setpower_nolock --> ath9k_hw_setpower()
ath9k_hw_setpower() --> ath9k_setpower()

Also change the param for ath9k_setpower() to pass the ath_softc.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: move ath9k_hw_setpower() to main.c
Luis R. Rodriguez [Thu, 10 Sep 2009 04:02:34 +0000 (21:02 -0700)]
ath9k: move ath9k_hw_setpower() to main.c

And we make it static.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: avoid usage of ath9k_hw_setpower() on hw.c
Luis R. Rodriguez [Thu, 10 Sep 2009 03:54:36 +0000 (20:54 -0700)]
ath9k: avoid usage of ath9k_hw_setpower() on hw.c

ath9k_hw_setpower() is a core driver helper with locking
protection. Locking protection should be left to the driver
core, not the hw code. Hardware code no longer contends for
locking when it needs to wake up the chip or put it to sleep.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: move ath9k_ps_wakeup() and ath9k_ps_restore() to main.c
Luis R. Rodriguez [Thu, 10 Sep 2009 03:29:18 +0000 (20:29 -0700)]
ath9k: move ath9k_ps_wakeup() and ath9k_ps_restore() to main.c

These are driver core helpers.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: move ps helpers onto core driver when reseting tsf
Luis R. Rodriguez [Thu, 10 Sep 2009 03:05:39 +0000 (20:05 -0700)]
ath9k: move ps helpers onto core driver when reseting tsf

When the TSF is reset power save state is disabled and
then restored. The helpers to disable power save and restore
it use a lock provided by the driver core. Move the callers
of the helpers outside of the hw code.

We reset the TSF when mac80211 tells us and on the beacon.c
helper ath9k_hw_beaconinit() when it is made explicitly required.
Add a helper on beacon.c which will deal with ps awake/restore
if we need to reset the TSF upon ath9k_hw_beaconinit().

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: rename ath_btcoex_supported() to ath9k_hw_btcoex_supported()
Luis R. Rodriguez [Wed, 9 Sep 2009 23:05:32 +0000 (16:05 -0700)]
ath9k: rename ath_btcoex_supported() to ath9k_hw_btcoex_supported()

Also just pass the ath_hw as the parameter.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: move ath_btcoex_config and ath_bt_mode to btcoex.c
Luis R. Rodriguez [Wed, 9 Sep 2009 22:24:02 +0000 (15:24 -0700)]
ath9k: move ath_btcoex_config and ath_bt_mode to btcoex.c

These are only used by btcoex.c on one routine, so stuff them
into that file.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: now move ath9k_hw_btcoex_set_weight() to btcoex.c
Luis R. Rodriguez [Wed, 9 Sep 2009 22:15:55 +0000 (15:15 -0700)]
ath9k: now move ath9k_hw_btcoex_set_weight() to btcoex.c

After some necessary cleanups we now move ath9k_hw_btcoex_set_weight()
to where it belongs.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: simplify ath_btcoex_bt_stomp()
Luis R. Rodriguez [Wed, 9 Sep 2009 22:05:00 +0000 (15:05 -0700)]
ath9k: simplify ath_btcoex_bt_stomp()

The second argument is always the hardware bt coex struct, so
remove it, and rename the function on the path with a ath9k_ prefix.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: rename ath_btcoex_info to ath_btcoex_hw
Luis R. Rodriguez [Wed, 9 Sep 2009 21:52:02 +0000 (14:52 -0700)]
ath9k: rename ath_btcoex_info to ath_btcoex_hw

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: rename btcoex_scheme to just scheme
Luis R. Rodriguez [Wed, 9 Sep 2009 21:34:22 +0000 (14:34 -0700)]
ath9k: rename btcoex_scheme to just scheme

btcoex_scheme is already part of a btcoex struct, its implied
this is btcoex related.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: remove unused bt_duty_cycle
Luis R. Rodriguez [Wed, 9 Sep 2009 21:31:24 +0000 (14:31 -0700)]
ath9k: remove unused bt_duty_cycle

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: move bt_stomp_type to driver core
Luis R. Rodriguez [Wed, 9 Sep 2009 21:26:15 +0000 (14:26 -0700)]
ath9k: move bt_stomp_type to driver core

The bt_stomp_type defines the bt coex weight, it has a one-to-one
mapping. In the future we may want to just use the weight directly.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: replaces SC_OP_BTCOEX_ENABLED with a bool
Luis R. Rodriguez [Wed, 9 Sep 2009 20:44:23 +0000 (13:44 -0700)]
ath9k: replaces SC_OP_BTCOEX_ENABLED with a bool

Whether or not bluetooth coex has been enabled is a hardware
state and only the hardware helpers will be able to set this.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: split ath9k_hw_btcoex_enable() into two helpers
Luis R. Rodriguez [Wed, 9 Sep 2009 11:17:45 +0000 (04:17 -0700)]
ath9k: split ath9k_hw_btcoex_enable() into two helpers

One for 2-wire and another for 3-wire.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: move driver core helpers to main.c
Luis R. Rodriguez [Wed, 9 Sep 2009 11:00:10 +0000 (04:00 -0700)]
ath9k: move driver core helpers to main.c

Keep on btcoex.c only hardware access helpers, move the
driver core specific code to main.c. To accomplish
this we had to split ath_init_btcoex_info() into two parts,
the driver core part -- ath_init_btcoex_timer() and the hw
specific part -- ath9k_hw_init_btcoex_hw_info(). This
highlights how ath_gen_timer is part of the driver core, not
hw related, so stuff that into ath_btcoex struct.

The ath9k_hw_btcoex_init() code is now put inline on
ath_init_softc() through a switch to it easier to follow,
since we did that we can now call ath_tx_get_qnum() from
the main.c instead of btcoex.c

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: split bluetooth hardware coex init into two helpers
Luis R. Rodriguez [Wed, 9 Sep 2009 09:54:40 +0000 (02:54 -0700)]
ath9k: split bluetooth hardware coex init into two helpers

Use a helper for 2-wire and another for 3-wire.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: move hw specific btcoex info to ath_hw
Luis R. Rodriguez [Wed, 9 Sep 2009 09:33:11 +0000 (02:33 -0700)]
ath9k: move hw specific btcoex info to ath_hw

Since we now access it via the ath_hw declare the ath_hw pointer
at the header of some routines and se it. ath9k.h no longer needs to
access btcoex.h and to adjust for this move ath_btcoex_set_weight()
into btcoex.h and instead give main.c a helper for setting initial
values upon drv_start()

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: move btcoex core driver info to its own struct
Luis R. Rodriguez [Wed, 9 Sep 2009 08:18:09 +0000 (01:18 -0700)]
ath9k: move btcoex core driver info to its own struct

There is some bluetooth coexistance data which is driver
specific, stuff that into its own structure.

Cc: Vasanthakumar Thiagarajan <vasanth@atheros.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: use ath_hw for DPRINTF() and debug init/exit
Luis R. Rodriguez [Mon, 7 Sep 2009 11:52:26 +0000 (04:52 -0700)]
ath9k: use ath_hw for DPRINTF() and debug init/exit

DPRINTF() is used in hw specific related code, as such
ensure we don't rely on the private driver core ath_softc
struct when calling it. Drivers can then implement their
own DPRINTF() as they see fit.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agonet: export device speed and duplex via sysfs
Andy Gospodarek [Fri, 2 Oct 2009 09:26:12 +0000 (09:26 +0000)]
net: export device speed and duplex via sysfs

This patch exports the link-speed (in Mbps) and duplex of an interface
via sysfs.  This eliminates the need to use ethtool just to check the
link-speed.  Not requiring 'ethtool' and not relying on the SIOCETHTOOL
ioctl should be helpful in an embedded environment where space is at a
premium as well.

NOTE: This patch also intentionally allows non-root users to check the link
speed and duplex -- something not possible with ethtool.

Here's some sample output:

# cat /sys/class/net/eth0/speed
100
# cat /sys/class/net/eth0/duplex
half
# ethtool eth0
Settings for eth0:
        Supported ports: [ TP ]
        Supported link modes:   10baseT/Half 10baseT/Full
                                100baseT/Half 100baseT/Full
                                1000baseT/Half 1000baseT/Full
        Supports auto-negotiation: Yes
        Advertised link modes:  Not reported
        Advertised auto-negotiation: No
        Speed: 100Mb/s
        Duplex: Half
        Port: Twisted Pair
        PHYAD: 1
        Transceiver: internal
        Auto-negotiation: off
        Supports Wake-on: g
        Wake-on: g
        Current message level: 0x000000ff (255)
        Link detected: yes

Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agocfg80211: assign device type in netdev notifier callback
Marcel Holtmann [Fri, 2 Oct 2009 05:15:28 +0000 (05:15 +0000)]
cfg80211: assign device type in netdev notifier callback

Instead of having to modify every non-mac80211 for device type assignment,
do this inside the netdev notifier callback of cfg80211. So all drivers
that integrate with cfg80211 will export a proper device type.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agonet: introduce NETDEV_POST_INIT notifier
Johannes Berg [Fri, 2 Oct 2009 05:15:27 +0000 (05:15 +0000)]
net: introduce NETDEV_POST_INIT notifier

For various purposes including a wireless extensions
bugfix, we need to hook into the netdev creation before
before netdev_register_kobject(). This will also ease
doing the dev type assignment that Marcel was working
on for cfg80211 drivers w/o touching them all.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agousbnet: Set device type for wlan and wwan devices
Marcel Holtmann [Fri, 2 Oct 2009 05:15:26 +0000 (05:15 +0000)]
usbnet: Set device type for wlan and wwan devices

For usbnet devices with FLAG_WLAN and FLAG_WWAN set the proper device
type so that uevent contains the correct value. This then allows an easy
identification of the actual underlying technology of the Ethernet device.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agousbnet: Use wwan%d interface name for mobile broadband devices
Marcel Holtmann [Fri, 2 Oct 2009 05:15:25 +0000 (05:15 +0000)]
usbnet: Use wwan%d interface name for mobile broadband devices

Add support for usbnet based devices like CDC-Ether to indicate that they
are actually mobile broadband devices. In that case use wwan%d as default
interface name.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agotunnels: Optimize tx path
Eric Dumazet [Mon, 5 Oct 2009 07:11:22 +0000 (00:11 -0700)]
tunnels: Optimize tx path

We currently dirty a cache line to update tunnel device stats
(tx_packets/tx_bytes). We better use the txq->tx_bytes/tx_packets
counters that already are present in cpu cache, in the cache
line shared with txq->_xmit_lock

This patch extends IPTUNNEL_XMIT() macro to use txq pointer
provided by the caller.

Also &tunnel->dev->stats can be replaced by &dev->stats

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoipv4: fib table algorithm performance improvement
Stephen Hemminger [Sun, 20 Sep 2009 10:35:36 +0000 (10:35 +0000)]
ipv4: fib table algorithm performance improvement

The FIB algorithim for IPV4 is set at compile time, but kernel goes through
the overhead of function call indirection at runtime. Save some
cycles by turning the indirect calls to direct calls to either
hash or trie code.

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoaf_packet: add interframe drop cmsg (v6)
Neil Horman [Fri, 2 Oct 2009 06:56:41 +0000 (06:56 +0000)]
af_packet: add interframe drop cmsg (v6)

Add Ancilliary data to better represent loss information

I've had a few requests recently to provide more detail regarding frame loss
during an AF_PACKET packet capture session.  Specifically the requestors want to
see where in a packet sequence frames were lost, i.e. they want to see that 40
frames were lost between frames 302 and 303 in a packet capture file.  In order
to do this we need:

1) The kernel to export this data to user space
2) The applications to make use of it

This patch addresses item (1).  It does this by doing the following:

A) Anytime we drop a frame for which we would increment po->stats.tp_drops, we
also no increment a stats called po->stats.tp_gap.

B) Every time we successfully enqueue a frame to sk_receive_queue, we record the
value of po->stats.tp_gap in skb->mark.  skb->cb would nominally be the place to
record this, but since all the space there is used up, we're overloading
skb->mark.  Its safe to do since any enqueued packet is guaranteed to be
unshared at this point, and skb->mark isn't used for anything else in the rx
path to the application.  After we record tp_gap in the skb, we zero
po->stats.tp_gap.  This allows us to keep a counter of the number of frames lost
between any two enqueued packets

C) When the application goes to dequeue a frame from the packet socket, we look
at skb->mark for that frame.  If it is non-zero, we add a cmsg chunk to the
msghdr of level SOL_PACKET and type PACKET_GAPDATA.  Its a 32 bit integer that
represents the number of frames lost between this packet and the last previous
frame received.

Note there is a chance that if there is frame loss after a receive, and then the
socket is closed, some gap data might be lost.  This is covered by the use of
the PACKET_AUXDATA socket option, which gives total loss data.  With a bit of
math, the final gap can be determined that way.

I've tested this patch myself, and it works well.

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
 include/linux/if_packet.h |    2 ++
 net/packet/af_packet.c    |   33 +++++++++++++++++++++++++++++++++
 2 files changed, 35 insertions(+)
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoTI DaVinci EMAC: Minor macro related updates
chaithrika@ti.com [Thu, 1 Oct 2009 10:25:19 +0000 (10:25 +0000)]
TI DaVinci EMAC: Minor macro related updates

Use BIT for macro definitions wherever possible, remove
unused and redundant macros.

Signed-off-by: Chaithrika U S <chaithrika@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agopktgen: Avoid dirtying skb->users when txq is full
Eric Dumazet [Wed, 30 Sep 2009 13:03:33 +0000 (13:03 +0000)]
pktgen: Avoid dirtying skb->users when txq is full

We can avoid two atomic ops on skb->users if packet is not going to be
sent to the device (because hardware txqueue is full)

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoicmp: No need to call sk_write_space()
Eric Dumazet [Thu, 24 Sep 2009 12:16:51 +0000 (12:16 +0000)]
icmp: No need to call sk_write_space()

We can make icmp messages tx completion callback a litle bit faster.

Setting SOCK_USE_WRITE_QUEUE sk flag tells sock_wfree() to
not call sk_write_space() on a socket we know no thread is posssibly
waiting for write space. (on per cpu kernel internal icmp sockets only)

This avoids the sock_def_write_space() call and
read_lock(&sk->sk_callback_lock)/read_unlock(&sk->sk_callback_lock) calls
as well.

We avoid three atomic ops.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoethtool: Remove support for obsolete string query operations
Ben Hutchings [Thu, 1 Oct 2009 11:33:03 +0000 (11:33 +0000)]
ethtool: Remove support for obsolete string query operations

The in-tree implementations have all been converted to
get_sset_count().

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agonet: Convert ethtool {get_stats, self_test}_count() ops to get_sset_count()
Ben Hutchings [Thu, 1 Oct 2009 11:58:24 +0000 (11:58 +0000)]
net: Convert ethtool {get_stats, self_test}_count() ops to get_sset_count()

These string query operations were supposed to be replaced by the
generic get_sset_count() starting in 2007.  Convert the remaining
implementations.

Also remove calls to these operations to initialise drvinfo->n_stats.
The ethtool core code already does that.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Acked-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agotehuti: Convert ethtool get_stats_count() ops to get_sset_count()
Ben Hutchings [Thu, 1 Oct 2009 11:27:59 +0000 (11:27 +0000)]
tehuti: Convert ethtool get_stats_count() ops to get_sset_count()

This string query operation was supposed to be replaced by the
generic get_sset_count() starting in 2007.  Convert tehuti's
implementation.

Also remove the dummy self-test name which was not used since tehuti
does not advertise any self-tests.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoqeth: Convert ethtool get_stats_count() ops to get_sset_count()
Ben Hutchings [Thu, 1 Oct 2009 11:24:32 +0000 (11:24 +0000)]
qeth: Convert ethtool get_stats_count() ops to get_sset_count()

This string query operation was supposed to be replaced by the
generic get_sset_count() starting in 2007.  Convert qeth's
implementation.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoLinux 2.6.32-rc3 v2.6.32-rc3
Linus Torvalds [Mon, 5 Oct 2009 00:12:30 +0000 (17:12 -0700)]
Linux 2.6.32-rc3

I'm skipping -rc2 because the -rc1 Makefile mistakenly said -rc2, so in
order to avoid confusion, I'm jumping from -rc1 to -rc3.  That way, when
'uname' (or an oops report) says 2.6.32-rc2, there's no confusion about
whether people perhaps meant -rc1 or -rc2.

14 years agoheaders: remove sched.h from poll.h
Alexey Dobriyan [Sun, 4 Oct 2009 12:11:37 +0000 (16:11 +0400)]
headers: remove sched.h from poll.h

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge branch 'acpi-pad' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux...
Linus Torvalds [Sun, 4 Oct 2009 22:03:00 +0000 (15:03 -0700)]
Merge branch 'acpi-pad' of git://git./linux/kernel/git/lenb/linux-acpi-2.6

* 'acpi-pad' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6:
  acpi_pad: build only on X86
  ACPI: create Processor Aggregator Device driver

Fixup trivial conflicts in MAINTAINERS file.

14 years agoMerge branch 'sfi-release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb...
Linus Torvalds [Sun, 4 Oct 2009 22:00:08 +0000 (15:00 -0700)]
Merge branch 'sfi-release' of git://git./linux/kernel/git/lenb/linux-sfi-2.6

* 'sfi-release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-sfi-2.6:
  SFI: remove __init from sfi_verify_table
  SFI: fix section mismatch warnings in sfi_core.c