d/freebsd-nq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
507de8028f
freebsd-nq/sys/dev/ath/ath_hal/ar5416/ar5416_reset.c

2428 lines
80 KiB
C
Raw Normal View History

virgin import of ath hal
2008-11-28 00:03:41 +00:00
/*
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
* Copyright (c) 2002-2009 Sam Leffler, Errno Consulting
virgin import of ath hal
2008-11-28 00:03:41 +00:00
* Copyright (c) 2002-2008 Atheros Communications, Inc.
*
* Permission to use, copy, modify, and/or distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*
fix return status handling by ar5XXXReset; this is the reason the driver sometimes reports reset failed w/ status 0
2009-01-23 05:33:48 +00:00
* $FreeBSD$
virgin import of ath hal
2008-11-28 00:03:41 +00:00
*/
#include "opt_ah.h"
#include "ah.h"
#include "ah_internal.h"
#include "ah_devid.h"
#include "ah_eeprom_v14.h"
#include "ar5416/ar5416.h"
#include "ar5416/ar5416reg.h"
#include "ar5416/ar5416phy.h"
/* Eeprom versioning macros. Returns true if the version is equal or newer than the ver specified */
#define EEP_MINOR(_ah) \
(AH_PRIVATE(_ah)->ah_eeversion & AR5416_EEP_VER_MINOR_MASK)
#define IS_EEP_MINOR_V2(_ah) (EEP_MINOR(_ah) >= AR5416_EEP_MINOR_VER_2)
#define IS_EEP_MINOR_V3(_ah) (EEP_MINOR(_ah) >= AR5416_EEP_MINOR_VER_3)
/* Additional Time delay to wait after activiting the Base band */
#define BASE_ACTIVATE_DELAY 100 /* 100 usec */
#define PLL_SETTLE_DELAY 300 /* 300 usec */
#define RTC_PLL_SETTLE_DELAY 1000 /* 1 ms */
static void ar5416InitDMA(struct ath_hal *ah);
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
static void ar5416InitBB(struct ath_hal *ah, const struct ieee80211_channel *);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
static void ar5416InitIMR(struct ath_hal *ah, HAL_OPMODE opmode);
static void ar5416InitQoS(struct ath_hal *ah);
static void ar5416InitUserSettings(struct ath_hal *ah);
Initialise the chainmask from the EEPROM rather than the hard-coded defaults. The defaults enabled three chains on the AR5416 even if the card has two chains. This restores that and ensures that only the correct TX/RX chainmasks are used. When HT modes are enabled, all TX chains will be correctly enabled. This should now enable analog chain swapping with 2-chain cards. I'm not sure if this is needed for just the AR5416 or whether it also applies to AR9160, AR9280 and AR9287 (later on); I'll have to get clarification.
2011-01-27 09:26:37 +00:00
static void ar5416UpdateChainMasks(struct ath_hal *ah, HAL_BOOL is_ht);
Port over ar5416OverrideIni() from ath9k ar5008_hw_override_ini(). * change the BB gating logic to explicitly define which chips are covered; the ath9k method isn't as clear. * don't disable the BB gating for now, the ar5416 initvals have it, and the ar9160 initval sets it to 0x0. Figure out why before re-enabling this. * migrate the Merlin (ar9280) applicable WAR from the Kite (ar9285) code (which won't get called for Merlin!) and stuff it in here.
2011-03-03 08:38:31 +00:00
static void ar5416OverrideIni(struct ath_hal *ah, const struct ieee80211_channel *);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
#if 0
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
static HAL_BOOL ar5416ChannelChange(struct ath_hal *, const struct ieee80211_channel *);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
#endif
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
static void ar5416SetDeltaSlope(struct ath_hal *, const struct ieee80211_channel *);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
static HAL_BOOL ar5416SetResetPowerOn(struct ath_hal *ah);
static HAL_BOOL ar5416SetReset(struct ath_hal *ah, int type);
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
static void ar5416InitPLL(struct ath_hal *ah, const struct ieee80211_channel *chan);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
static HAL_BOOL ar5416SetPowerPerRateTable(struct ath_hal *ah,
struct ar5416eeprom *pEepData,
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
const struct ieee80211_channel *chan, int16_t *ratesArray,
virgin import of ath hal
2008-11-28 00:03:41 +00:00
uint16_t cfgCtl, uint16_t AntennaReduction,
uint16_t twiceMaxRegulatoryPower,
uint16_t powerLimit);
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
static uint16_t ar5416GetMaxEdgePower(uint16_t freq,
virgin import of ath hal
2008-11-28 00:03:41 +00:00
CAL_CTL_EDGES *pRdEdgesPower, HAL_BOOL is2GHz);
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
static void ar5416Set11nRegs(struct ath_hal *ah, const struct ieee80211_channel *chan);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/*
* Places the device in and out of reset and then places sane
* values in the registers based on EEPROM config, initialization
* vectors (as determined by the mode), and station configuration
*
* bChannelChange is used to preserve DMA/PCU registers across
* a HW Reset during channel change.
*/
HAL_BOOL
ar5416Reset(struct ath_hal *ah, HAL_OPMODE opmode,
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
struct ieee80211_channel *chan,
HAL_BOOL bChannelChange, HAL_STATUS *status)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
#define N(a) (sizeof (a) / sizeof (a[0]))
#define FAIL(_code) do { ecode = _code; goto bad; } while (0)
struct ath_hal_5212 *ahp = AH5212(ah);
HAL_CHANNEL_INTERNAL *ichan;
uint32_t saveDefAntenna, saveLedState;
uint32_t macStaId1;
uint16_t rfXpdGain[2];
HAL_STATUS ecode;
uint32_t powerVal, rssiThrReg;
uint32_t ackTpcPow, ctsTpcPow, chirpTpcPow;
preliminary ar9280 support: o add 9280 attach that sets up ini, cal, etc. o new rf backend for 9280 and later parts o split ini setup and spur mitigation support out to methods and provide 9280-specific support o minor fixups to shared code to handle 9280-specific work Obtained from: Atheros (ini values and some code)
2009-03-12 18:18:28 +00:00
int i;
For chips that are full reset in ar5416ChipReset(), save and restore the TSF. Merlin (ar9280) and later were full-reset if they're doing open-loop TX power control but the TSF wasn't being saved/restored. Add ar5212SetTsf64() which sets the 64 bit TSF appropriately.
2011-03-09 04:39:35 +00:00
uint64_t tsf = 0;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
OS_MARK(ah, AH_MARK_RESET, bChannelChange);
/* Bring out of sleep mode */
if (!ar5416SetPowerMode(ah, HAL_PM_AWAKE, AH_TRUE)) {
HALDEBUG(ah, HAL_DEBUG_ANY, "%s: chip did not wakeup\n",
__func__);
Revert a local change that should not have been in the last commit. Approved by: re (kib)
2009-06-28 11:32:52 +00:00
FAIL(HAL_EIO);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
/*
* Map public channel to private.
*/
ichan = ath_hal_checkchannel(ah, chan);
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (ichan == AH_NULL)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
FAIL(HAL_EINVAL);
switch (opmode) {
case HAL_M_STA:
case HAL_M_IBSS:
case HAL_M_HOSTAP:
case HAL_M_MONITOR:
break;
default:
HALDEBUG(ah, HAL_DEBUG_ANY, "%s: invalid operating mode %u\n",
__func__, opmode);
FAIL(HAL_EINVAL);
break;
}
HALASSERT(AH_PRIVATE(ah)->ah_eeversion >= AR_EEPROM_VER14_1);
/* XXX Turn on fast channel change for 5416 */
/*
* Preserve the bmiss rssi threshold and count threshold
* across resets
*/
rssiThrReg = OS_REG_READ(ah, AR_RSSI_THR);
/* If reg is zero, first time thru set to default val */
if (rssiThrReg == 0)
rssiThrReg = INIT_RSSI_THR;
/*
* Preserve the antenna on a channel change
*/
saveDefAntenna = OS_REG_READ(ah, AR_DEF_ANTENNA);
if (saveDefAntenna == 0) /* XXX magic constants */
saveDefAntenna = 1;
/* Save hardware flag before chip reset clears the register */
macStaId1 = OS_REG_READ(ah, AR_STA_ID1) &
(AR_STA_ID1_BASE_RATE_11B | AR_STA_ID1_USE_DEFANT);
/* Save led state from pci config register */
saveLedState = OS_REG_READ(ah, AR_MAC_LED) &
(AR_MAC_LED_ASSOC | AR_MAC_LED_MODE |
AR_MAC_LED_BLINK_THRESH_SEL | AR_MAC_LED_BLINK_SLOW);
For chips that are full reset in ar5416ChipReset(), save and restore the TSF. Merlin (ar9280) and later were full-reset if they're doing open-loop TX power control but the TSF wasn't being saved/restored. Add ar5212SetTsf64() which sets the 64 bit TSF appropriately.
2011-03-09 04:39:35 +00:00
/* For chips on which the RTC reset is done, save TSF before it gets cleared */
if (AR_SREV_MERLIN_20_OR_LATER(ah) && ath_hal_eepromGetFlag(ah, AR_EEP_OL_PWRCTRL))
tsf = ar5212GetTsf64(ah);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
if (!ar5416ChipReset(ah, chan)) {
HALDEBUG(ah, HAL_DEBUG_ANY, "%s: chip reset failed\n", __func__);
FAIL(HAL_EIO);
}
/* Restore bmiss rssi & count thresholds */
OS_REG_WRITE(ah, AR_RSSI_THR, rssiThrReg);
For chips that are full reset in ar5416ChipReset(), save and restore the TSF. Merlin (ar9280) and later were full-reset if they're doing open-loop TX power control but the TSF wasn't being saved/restored. Add ar5212SetTsf64() which sets the 64 bit TSF appropriately.
2011-03-09 04:39:35 +00:00
/* Restore TSF */
if (tsf)
ar5212SetTsf64(ah, tsf);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
OS_MARK(ah, AH_MARK_RESET_LINE, __LINE__);
Bring in a couple of fixes from the Linux ath9k related to chip hangs. While there, try to make the register write pattern look like what's done by ath9k. MFC after: 3 days
2010-06-01 15:33:10 +00:00
if (AR_SREV_MERLIN_10_OR_LATER(ah))
OS_REG_SET_BIT(ah, AR_GPIO_INPUT_EN_VAL, AR_GPIO_JTAG_DISABLE);
if (AR_SREV_KITE(ah)) {
uint32_t val;
val = OS_REG_READ(ah, AR_PHY_HEAVY_CLIP_FACTOR_RIFS);
val &= ~AR_PHY_RIFS_INIT_DELAY;
OS_REG_WRITE(ah, AR_PHY_HEAVY_CLIP_FACTOR_RIFS, val);
}
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
preliminary ar9280 support: o add 9280 attach that sets up ini, cal, etc. o new rf backend for 9280 and later parts o split ini setup and spur mitigation support out to methods and provide 9280-specific support o minor fixups to shared code to handle 9280-specific work Obtained from: Atheros (ini values and some code)
2009-03-12 18:18:28 +00:00
AH5416(ah)->ah_writeIni(ah, chan);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
Port over ar5416OverrideIni() from ath9k ar5008_hw_override_ini(). * change the BB gating logic to explicitly define which chips are covered; the ath9k method isn't as clear. * don't disable the BB gating for now, the ar5416 initvals have it, and the ar9160 initval sets it to 0x0. Figure out why before re-enabling this. * migrate the Merlin (ar9280) applicable WAR from the Kite (ar9285) code (which won't get called for Merlin!) and stuff it in here.
2011-03-03 08:38:31 +00:00
/* Override ini values (that can be overriden in this fashion) */
ar5416OverrideIni(ah, chan);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Setup 11n MAC/Phy mode registers */
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
ar5416Set11nRegs(ah, chan);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
OS_MARK(ah, AH_MARK_RESET_LINE, __LINE__);
HALDEBUG(ah, HAL_DEBUG_RESET, ">>>2 %s: AR_PHY_DAG_CTRLCCK=0x%x\n",
__func__, OS_REG_READ(ah,AR_PHY_DAG_CTRLCCK));
HALDEBUG(ah, HAL_DEBUG_RESET, ">>>2 %s: AR_PHY_ADC_CTL=0x%x\n",
__func__, OS_REG_READ(ah,AR_PHY_ADC_CTL));
/* Set the mute mask to the correct default */
if (AH_PRIVATE(ah)->ah_phyRev >= AR_PHY_CHIP_ID_REV_2)
OS_REG_WRITE(ah, AR_SEQ_MASK, 0x0000000F);
if (AH_PRIVATE(ah)->ah_phyRev >= AR_PHY_CHIP_ID_REV_3) {
/* Clear reg to alllow RX_CLEAR line debug */
OS_REG_WRITE(ah, AR_PHY_BLUETOOTH, 0);
}
if (AH_PRIVATE(ah)->ah_phyRev >= AR_PHY_CHIP_ID_REV_4) {
#ifdef notyet
/* Enable burst prefetch for the data queues */
OS_REG_RMW_FIELD(ah, AR_D_FPCTL, ... );
/* Enable double-buffering */
OS_REG_CLR_BIT(ah, AR_TXCFG, AR_TXCFG_DBL_BUF_DIS);
#endif
}
Initialise the chainmask from the EEPROM rather than the hard-coded defaults. The defaults enabled three chains on the AR5416 even if the card has two chains. This restores that and ensures that only the correct TX/RX chainmasks are used. When HT modes are enabled, all TX chains will be correctly enabled. This should now enable analog chain swapping with 2-chain cards. I'm not sure if this is needed for just the AR5416 or whether it also applies to AR9160, AR9280 and AR9287 (later on); I'll have to get clarification.
2011-01-27 09:26:37 +00:00
/*
* Setup ah_tx_chainmask / ah_rx_chainmask before we fiddle
* with enabling the TX/RX radio chains.
*/
ar5416UpdateChainMasks(ah, IEEE80211_IS_CHAN_HT(chan));
Break out the chainmask init code into a new function - ar5416InitChainMasks() . ath9k does a few different things here during config - if it's an early AR5416 with two chains, it enables all three chains for calibration and then restores the chainmask to the original values after initial calibration has completed. The reason behind this commit is to begin breaking out the chainmask configuration for this specific reason; follow-up commits will add the chainmask restore in the ar5416Reset() routine.
2011-01-26 10:08:37 +00:00
/*
* This routine swaps the analog chains - it should be done
* before any radio register twiddling is done.
*/
ar5416InitChainMasks(ah);
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
AH5416(ah)->ah_olcInit(ah);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Setup the transmit power values. */
Bring back AR9285 support. This fixes most of the issues and should be pretty usable. MFC after: 1 month
2010-02-15 17:49:49 +00:00
if (!ah->ah_setTxPower(ah, chan, rfXpdGain)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
HALDEBUG(ah, HAL_DEBUG_ANY,
"%s: error init'ing transmit power\n", __func__);
FAIL(HAL_EIO);
}
/* Write the analog registers */
preliminary ar9280 support: o add 9280 attach that sets up ini, cal, etc. o new rf backend for 9280 and later parts o split ini setup and spur mitigation support out to methods and provide 9280-specific support o minor fixups to shared code to handle 9280-specific work Obtained from: Atheros (ini values and some code)
2009-03-12 18:18:28 +00:00
if (!ahp->ah_rfHal->setRfRegs(ah, chan,
IEEE80211_IS_CHAN_2GHZ(chan) ? 2: 1, rfXpdGain)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
HALDEBUG(ah, HAL_DEBUG_ANY,
"%s: ar5212SetRfRegs failed\n", __func__);
FAIL(HAL_EIO);
}
/* Write delta slope for OFDM enabled modes (A, G, Turbo) */
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_OFDM(chan)|| IEEE80211_IS_CHAN_HT(chan))
ar5416SetDeltaSlope(ah, chan);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
preliminary ar9280 support: o add 9280 attach that sets up ini, cal, etc. o new rf backend for 9280 and later parts o split ini setup and spur mitigation support out to methods and provide 9280-specific support o minor fixups to shared code to handle 9280-specific work Obtained from: Atheros (ini values and some code)
2009-03-12 18:18:28 +00:00
AH5416(ah)->ah_spurMitigate(ah, chan);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Setup board specific options for EEPROM version 3 */
Bring back AR9285 support. This fixes most of the issues and should be pretty usable. MFC after: 1 month
2010-02-15 17:49:49 +00:00
if (!ah->ah_setBoardValues(ah, chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
HALDEBUG(ah, HAL_DEBUG_ANY,
"%s: error setting board options\n", __func__);
FAIL(HAL_EIO);
}
OS_MARK(ah, AH_MARK_RESET_LINE, __LINE__);
OS_REG_WRITE(ah, AR_STA_ID0, LE_READ_4(ahp->ah_macaddr));
OS_REG_WRITE(ah, AR_STA_ID1, LE_READ_2(ahp->ah_macaddr + 4)
| macStaId1
| AR_STA_ID1_RTS_USE_DEF
| ahp->ah_staId1Defaults
);
ar5212SetOperatingMode(ah, opmode);
/* Set Venice BSSID mask according to current state */
OS_REG_WRITE(ah, AR_BSSMSKL, LE_READ_4(ahp->ah_bssidmask));
OS_REG_WRITE(ah, AR_BSSMSKU, LE_READ_2(ahp->ah_bssidmask + 4));
/* Restore previous led state */
OS_REG_WRITE(ah, AR_MAC_LED, OS_REG_READ(ah, AR_MAC_LED) | saveLedState);
/* Restore previous antenna */
OS_REG_WRITE(ah, AR_DEF_ANTENNA, saveDefAntenna);
/* then our BSSID */
OS_REG_WRITE(ah, AR_BSS_ID0, LE_READ_4(ahp->ah_bssid));
OS_REG_WRITE(ah, AR_BSS_ID1, LE_READ_2(ahp->ah_bssid + 4));
/* Restore bmiss rssi & count thresholds */
OS_REG_WRITE(ah, AR_RSSI_THR, ahp->ah_rssiThr);
OS_REG_WRITE(ah, AR_ISR, ~0); /* cleared on write */
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (!ar5212SetChannel(ah, chan))
virgin import of ath hal
2008-11-28 00:03:41 +00:00
FAIL(HAL_EIO);
OS_MARK(ah, AH_MARK_RESET_LINE, __LINE__);
/* Set 1:1 QCU to DCU mapping for all queues */
for (i = 0; i < AR_NUM_DCU; i++)
OS_REG_WRITE(ah, AR_DQCUMASK(i), 1 << i);
ahp->ah_intrTxqs = 0;
for (i = 0; i < AH_PRIVATE(ah)->ah_caps.halTotalQueues; i++)
Use the HAL method rather than directly calling ar5212ResetTxQueue(). Since ath9k does some slightly different bit fiddling when setting up the TX queues, it may that the TX queue setup/reset functions will need overriding later on.
2011-03-19 03:09:21 +00:00
ah->ah_resetTxQueue(ah, i);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
ar5416InitIMR(ah, opmode);
ar5212SetCoverageClass(ah, AH_PRIVATE(ah)->ah_coverageClass, 1);
ar5416InitQoS(ah);
ar5416InitUserSettings(ah);
/*
* disable seq number generation in hw
*/
OS_REG_WRITE(ah, AR_STA_ID1,
OS_REG_READ(ah, AR_STA_ID1) | AR_STA_ID1_PRESERVE_SEQNUM);
ar5416InitDMA(ah);
/*
* program OBS bus to see MAC interrupts
*/
OS_REG_WRITE(ah, AR_OBS, 8);
#ifdef AR5416_INT_MITIGATION
OS_REG_WRITE(ah, AR_MIRT, 0);
OS_REG_RMW_FIELD(ah, AR_RIMT, AR_RIMT_LAST, 500);
OS_REG_RMW_FIELD(ah, AR_RIMT, AR_RIMT_FIRST, 2000);
#endif
ar5416InitBB(ah, chan);
/* Setup compression registers */
ar5212SetCompRegs(ah); /* XXX not needed? */
/*
* 5416 baseband will check the per rate power table
* and select the lower of the two
*/
ackTpcPow = 63;
ctsTpcPow = 63;
chirpTpcPow = 63;
powerVal = SM(ackTpcPow, AR_TPC_ACK) |
SM(ctsTpcPow, AR_TPC_CTS) |
SM(chirpTpcPow, AR_TPC_CHIRP);
OS_REG_WRITE(ah, AR_TPC, powerVal);
if (!ar5416InitCal(ah, chan))
FAIL(HAL_ESELFTEST);
Add ar5416RestoreChainMask() which will undo any AR5416 specific chainmask overriding after calibration. This will get set for other two chain radios, such as AR9280 and later on, AR9287. It should however be a nul operation.
2011-01-26 10:48:29 +00:00
ar5416RestoreChainMask(ah);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
AH_PRIVATE(ah)->ah_opmode = opmode; /* record operating mode */
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (bChannelChange && !IEEE80211_IS_CHAN_DFS(chan))
chan->ic_state &= ~IEEE80211_CHANSTATE_CWINT;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
HALDEBUG(ah, HAL_DEBUG_RESET, "%s: done\n", __func__);
OS_MARK(ah, AH_MARK_RESET_DONE, 0);
return AH_TRUE;
bad:
OS_MARK(ah, AH_MARK_RESET_DONE, ecode);
fix return status handling by ar5XXXReset; this is the reason the driver sometimes reports reset failed w/ status 0
2009-01-23 05:33:48 +00:00
if (status != AH_NULL)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
*status = ecode;
return AH_FALSE;
#undef FAIL
#undef N
}
#if 0
/*
* This channel change evaluates whether the selected hardware can
* perform a synthesizer-only channel change (no reset). If the
* TX is not stopped, or the RFBus cannot be granted in the given
* time, the function returns false as a reset is necessary
*/
HAL_BOOL
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
ar5416ChannelChange(struct ath_hal *ah, const structu ieee80211_channel *chan)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
uint32_t ulCount;
uint32_t data, synthDelay, qnum;
uint16_t rfXpdGain[4];
struct ath_hal_5212 *ahp = AH5212(ah);
HAL_CHANNEL_INTERNAL *ichan;
/*
* Map public channel to private.
*/
ichan = ath_hal_checkchannel(ah, chan);
/* TX must be stopped or RF Bus grant will not work */
for (qnum = 0; qnum < AH_PRIVATE(ah)->ah_caps.halTotalQueues; qnum++) {
if (ar5212NumTxPending(ah, qnum)) {
HALDEBUG(ah, HAL_DEBUG_ANY,
"%s: frames pending on queue %d\n", __func__, qnum);
return AH_FALSE;
}
}
/*
* Kill last Baseband Rx Frame - Request analog bus grant
*/
OS_REG_WRITE(ah, AR_PHY_RFBUS_REQ, AR_PHY_RFBUS_REQ_REQUEST);
if (!ath_hal_wait(ah, AR_PHY_RFBUS_GNT, AR_PHY_RFBUS_GRANT_EN, AR_PHY_RFBUS_GRANT_EN)) {
HALDEBUG(ah, HAL_DEBUG_ANY, "%s: could not kill baseband rx\n",
__func__);
return AH_FALSE;
}
ar5416Set11nRegs(ah, chan); /* NB: setup 5416-specific regs */
/* Change the synth */
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (!ar5212SetChannel(ah, chan))
virgin import of ath hal
2008-11-28 00:03:41 +00:00
return AH_FALSE;
/* Setup the transmit power values. */
Don't call ar5416SetTransmitPower() directly from ar5416SetTxPowerLimit(); this is incorrect for Kite (AR9285) and any future chipsets that override the EEPROM related routines. It meant that a direct call to set the TX power would call the v14 EEPROM AR5416/AR9280 calibration routines, rather than the v4k EEPROM routines for the AR9285. It thus read the incorrect values from the EEPROM and programmed garbage PDADC and TX power values into the hardware.
2011-03-11 03:46:27 +00:00
if (!ah->ah_setTxPower(ah, chan, rfXpdGain)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
HALDEBUG(ah, HAL_DEBUG_ANY,
"%s: error init'ing transmit power\n", __func__);
return AH_FALSE;
}
/*
* Wait for the frequency synth to settle (synth goes on
* via PHY_ACTIVE_EN). Read the phy active delay register.
* Value is in 100ns increments.
*/
data = OS_REG_READ(ah, AR_PHY_RX_DELAY) & AR_PHY_RX_DELAY_DELAY;
if (IS_CHAN_CCK(ichan)) {
synthDelay = (4 * data) / 22;
} else {
synthDelay = data / 10;
}
OS_DELAY(synthDelay + BASE_ACTIVATE_DELAY);
/* Release the RFBus Grant */
OS_REG_WRITE(ah, AR_PHY_RFBUS_REQ, 0);
/* Write delta slope for OFDM enabled modes (A, G, Turbo) */
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_OFDM(ichan)|| IEEE80211_IS_CHAN_HT(chan)) {
HALASSERT(AH_PRIVATE(ah)->ah_eeversion >= AR_EEPROM_VER5_3);
ar5212SetSpurMitigation(ah, chan);
ar5416SetDeltaSlope(ah, chan);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
/* XXX spur mitigation for Melin */
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (!IEEE80211_IS_CHAN_DFS(chan))
chan->ic_state &= ~IEEE80211_CHANSTATE_CWINT;
ichan->channel_time = 0;
ichan->tsf_last = ar5212GetTsf64(ah);
ar5212TxEnable(ah, AH_TRUE);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
return AH_TRUE;
}
#endif
static void
ar5416InitDMA(struct ath_hal *ah)
{
Couple of suggestions from Sam regarding latest commit: o rename the new variables to comply with the naming scheme o move the new variables to an AR5212 specific struct o use ahp when available o revert to previous ts_flags check
2010-03-02 12:59:42 +00:00
struct ath_hal_5212 *ahp = AH5212(ah);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/*
* set AHB_MODE not to do cacheline prefetches
*/
OS_REG_SET_BIT(ah, AR_AHB_MODE, AR_AHB_PREFETCH_RD_EN);
/*
* let mac dma reads be in 128 byte chunks
*/
OS_REG_WRITE(ah, AR_TXCFG,
(OS_REG_READ(ah, AR_TXCFG) & ~AR_TXCFG_DMASZ_MASK) | AR_TXCFG_DMASZ_128B);
/*
* let mac dma writes be in 128 byte chunks
*/
OS_REG_WRITE(ah, AR_RXCFG,
(OS_REG_READ(ah, AR_RXCFG) & ~AR_RXCFG_DMASZ_MASK) | AR_RXCFG_DMASZ_128B);
Properly setup the TX FIFO threshold for AR5416 based chipsets, including the AR9285. This seems to fix some users's problems. Submitted by: Jorge Boncompte [DTI2] <jorge at dti2.net>
2010-03-01 17:04:19 +00:00
/* restore TX trigger level */
OS_REG_WRITE(ah, AR_TXCFG,
(OS_REG_READ(ah, AR_TXCFG) &~ AR_FTRIG) |
Couple of suggestions from Sam regarding latest commit: o rename the new variables to comply with the naming scheme o move the new variables to an AR5212 specific struct o use ahp when available o revert to previous ts_flags check
2010-03-02 12:59:42 +00:00
SM(ahp->ah_txTrigLev, AR_FTRIG));
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/*
* Setup receive FIFO threshold to hold off TX activities
*/
OS_REG_WRITE(ah, AR_RXFIFO_CFG, 0x200);
/*
* reduce the number of usable entries in PCU TXBUF to avoid
* wrap around.
*/
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
OS_REG_WRITE(ah, AR_PCU_TXBUF_CTRL, AR_PCU_TXBUF_CTRL_USABLE_SIZE);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
static void
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
ar5416InitBB(struct ath_hal *ah, const struct ieee80211_channel *chan)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
uint32_t synthDelay;
/*
* Wait for the frequency synth to settle (synth goes on
* via AR_PHY_ACTIVE_EN). Read the phy active delay register.
* Value is in 100ns increments.
*/
synthDelay = OS_REG_READ(ah, AR_PHY_RX_DELAY) & AR_PHY_RX_DELAY_DELAY;
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_CCK(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
synthDelay = (4 * synthDelay) / 22;
} else {
synthDelay /= 10;
}
/* Turn on PLL on 5416 */
HALDEBUG(ah, HAL_DEBUG_RESET, "%s %s channel\n",
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
__func__, IEEE80211_IS_CHAN_5GHZ(chan) ? "5GHz" : "2GHz");
virgin import of ath hal
2008-11-28 00:03:41 +00:00
ar5416InitPLL(ah, chan);
/* Activate the PHY (includes baseband activate and synthesizer on) */
OS_REG_WRITE(ah, AR_PHY_ACTIVE, AR_PHY_ACTIVE_EN);
/*
* If the AP starts the calibration before the base band timeout
Update to later code from my repository: o many bug fixes o add new periodic calibration api o break up 5416 periodic calibration code in preparation for 928x o move get noise floor to rf backends o 5416-specific ani (still disabled) o modularize 5210 eeprom format a la other eeprom formats o start cleaning up regdomain code o prepare for proper 1/2 and 1/4 width channel support o bring back 900MHz card support o clean up 5212 rf version handling o add 1/2 and 1/4 width channel support for 5212 parts o split 5212 rfgain handling out o improve ani debugging o add AH_USE_INIPDGAIN compile option o purge a bunch of dead 5212 state o add 1/2 and 1/4 rate modes o remove HAL_CAP_CHAN_HALFRATE and HAL_CAP_CHAN_QUARTERRATE; the same info can now be deduced from the set of supported modes
2008-11-28 00:48:05 +00:00
* completes we could get rx_clear false triggering. Add an
* extra BASE_ACTIVATE_DELAY usecs to ensure this condition
virgin import of ath hal
2008-11-28 00:03:41 +00:00
* does not happen.
*/
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_HALF(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
OS_DELAY((synthDelay << 1) + BASE_ACTIVATE_DELAY);
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
} else if (IEEE80211_IS_CHAN_QUARTER(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
OS_DELAY((synthDelay << 2) + BASE_ACTIVATE_DELAY);
} else {
OS_DELAY(synthDelay + BASE_ACTIVATE_DELAY);
}
}
static void
ar5416InitIMR(struct ath_hal *ah, HAL_OPMODE opmode)
{
struct ath_hal_5212 *ahp = AH5212(ah);
/*
* Setup interrupt handling. Note that ar5212ResetTxQueue
* manipulates the secondary IMR's as queues are enabled
* and disabled. This is done with RMW ops to insure the
* settings we make here are preserved.
*/
ahp->ah_maskReg = AR_IMR_TXERR | AR_IMR_TXURN
| AR_IMR_RXERR | AR_IMR_RXORN
| AR_IMR_BCNMISC;
#ifdef AR5416_INT_MITIGATION
ahp->ah_maskReg |= AR_IMR_TXINTM | AR_IMR_RXINTM
| AR_IMR_TXMINTR | AR_IMR_RXMINTR;
#else
ahp->ah_maskReg |= AR_IMR_TXOK | AR_IMR_RXOK;
#endif
if (opmode == HAL_M_HOSTAP)
ahp->ah_maskReg |= AR_IMR_MIB;
OS_REG_WRITE(ah, AR_IMR, ahp->ah_maskReg);
/* Enable bus errors that are OR'd to set the HIUERR bit */
#if 0
OS_REG_WRITE(ah, AR_IMR_S2,
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
OS_REG_READ(ah, AR_IMR_S2) | AR_IMR_S2_GTT | AR_IMR_S2_CST);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
#endif
}
static void
ar5416InitQoS(struct ath_hal *ah)
{
/* QoS support */
OS_REG_WRITE(ah, AR_QOS_CONTROL, 0x100aa); /* XXX magic */
OS_REG_WRITE(ah, AR_QOS_SELECT, 0x3210); /* XXX magic */
/* Turn on NOACK Support for QoS packets */
OS_REG_WRITE(ah, AR_NOACK,
SM(2, AR_NOACK_2BIT_VALUE) |
SM(5, AR_NOACK_BIT_OFFSET) |
SM(0, AR_NOACK_BYTE_OFFSET));
/*
* initialize TXOP for all TIDs
*/
OS_REG_WRITE(ah, AR_TXOP_X, AR_TXOP_X_VAL);
OS_REG_WRITE(ah, AR_TXOP_0_3, 0xFFFFFFFF);
OS_REG_WRITE(ah, AR_TXOP_4_7, 0xFFFFFFFF);
OS_REG_WRITE(ah, AR_TXOP_8_11, 0xFFFFFFFF);
OS_REG_WRITE(ah, AR_TXOP_12_15, 0xFFFFFFFF);
}
static void
ar5416InitUserSettings(struct ath_hal *ah)
{
struct ath_hal_5212 *ahp = AH5212(ah);
/* Restore user-specified settings */
if (ahp->ah_miscMode != 0)
Make sure that the AR_MISC_MODE value from the initvals are properly respected. This commit really is "fix the OFDM duration calculation to match reality when running in 802.11g mode." The AR5212 init vals set AR_MISC_MODE to 0x0 and all the bits that can be set are set through code. The AR5416 and later initvals set AR_MISC_MODE to various other values (with the AR5212 AR_MISC_MODE options cleared), which include AR_PCU_CCK_SIFS_MODE . This adds 6uS to SIFS on non-CCK frames when transmitting. This fixes the issue where _DATA_ 802.11g OFDM frames were being TX'ed with the ACK duration set to 38uS, not 44uS as on the AR5212 (and other devices.) The AR5212 TX pathway obeys the software-programmed duration field in the packet, but the 11n TX pathway overrides that with a hardware-calculated duration. This was getting it wrong because of the above AR_MISC_MODE setting. I've verified that 11g data OFDM frames are now being TXed with the correct ACK+SIFS duration programmed in.
2011-03-19 03:15:28 +00:00
OS_REG_WRITE(ah, AR_MISC_MODE, OS_REG_READ(ah, AR_MISC_MODE) | ahp->ah_miscMode);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
if (ahp->ah_sifstime != (u_int) -1)
ar5212SetSifsTime(ah, ahp->ah_sifstime);
if (ahp->ah_slottime != (u_int) -1)
ar5212SetSlotTime(ah, ahp->ah_slottime);
if (ahp->ah_acktimeout != (u_int) -1)
ar5212SetAckTimeout(ah, ahp->ah_acktimeout);
if (ahp->ah_ctstimeout != (u_int) -1)
ar5212SetCTSTimeout(ah, ahp->ah_ctstimeout);
if (AH_PRIVATE(ah)->ah_diagreg != 0)
OS_REG_WRITE(ah, AR_DIAG_SW, AH_PRIVATE(ah)->ah_diagreg);
#if 0 /* XXX Todo */
if (ahp->ah_globaltxtimeout != (u_int) -1)
ar5416SetGlobalTxTimeout(ah, ahp->ah_globaltxtimeout);
#endif
}
Break out the RF mode setup into ar5416SetRfMode(), mirroring what ath9k does.
2011-03-22 00:52:44 +00:00
static void
ar5416SetRfMode(struct ath_hal *ah, const struct ieee80211_channel *chan)
{
uint32_t rfMode;
if (chan == AH_NULL)
return;
/* treat channel B as channel G , no B mode suport in owl */
rfMode = IEEE80211_IS_CHAN_CCK(chan) ?
AR_PHY_MODE_DYNAMIC : AR_PHY_MODE_OFDM;
if (AR_SREV_MERLIN_20(ah) && IS_5GHZ_FAST_CLOCK_EN(ah, chan)) {
/* phy mode bits for 5GHz channels require Fast Clock */
rfMode |= AR_PHY_MODE_DYNAMIC
| AR_PHY_MODE_DYN_CCK_DISABLE;
} else if (!AR_SREV_MERLIN_10_OR_LATER(ah)) {
rfMode |= IEEE80211_IS_CHAN_5GHZ(chan) ?
AR_PHY_MODE_RF5GHZ : AR_PHY_MODE_RF2GHZ;
}
OS_REG_WRITE(ah, AR_PHY_MODE, rfMode);
}
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/*
* Places the hardware into reset and then pulls it out of reset
*/
HAL_BOOL
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
ar5416ChipReset(struct ath_hal *ah, const struct ieee80211_channel *chan)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
OS_MARK(ah, AH_MARK_CHIPRESET, chan ? chan->ic_freq : 0);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/*
* Warm reset is optimistic.
*/
if (AR_SREV_MERLIN_20_OR_LATER(ah) &&
ath_hal_eepromGetFlag(ah, AR_EEP_OL_PWRCTRL)) {
if (!ar5416SetResetReg(ah, HAL_RESET_POWER_ON))
return AH_FALSE;
} else {
if (!ar5416SetResetReg(ah, HAL_RESET_WARM))
return AH_FALSE;
}
/* Bring out of sleep mode (AGAIN) */
if (!ar5416SetPowerMode(ah, HAL_PM_AWAKE, AH_TRUE))
return AH_FALSE;
ar5416InitPLL(ah, chan);
/*
* Perform warm reset before the mode/PLL/turbo registers
* are changed in order to deactivate the radio. Mode changes
* with an active radio can result in corrupted shifts to the
* radio device.
*/
Break out the RF mode setup into ar5416SetRfMode(), mirroring what ath9k does.
2011-03-22 00:52:44 +00:00
if (chan != AH_NULL)
ar5416SetRfMode(ah, chan);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
return AH_TRUE;
}
/*
* Delta slope coefficient computation.
* Required for OFDM operation.
*/
static void
ar5416GetDeltaSlopeValues(struct ath_hal *ah, uint32_t coef_scaled,
uint32_t *coef_mantissa, uint32_t *coef_exponent)
{
#define COEF_SCALE_S 24
uint32_t coef_exp, coef_man;
/*
* ALGO -> coef_exp = 14-floor(log2(coef));
* floor(log2(x)) is the highest set bit position
*/
for (coef_exp = 31; coef_exp > 0; coef_exp--)
if ((coef_scaled >> coef_exp) & 0x1)
break;
/* A coef_exp of 0 is a legal bit position but an unexpected coef_exp */
HALASSERT(coef_exp);
coef_exp = 14 - (coef_exp - COEF_SCALE_S);
/*
* ALGO -> coef_man = floor(coef* 2^coef_exp+0.5);
* The coefficient is already shifted up for scaling
*/
coef_man = coef_scaled + (1 << (COEF_SCALE_S - coef_exp - 1));
*coef_mantissa = coef_man >> (COEF_SCALE_S - coef_exp);
*coef_exponent = coef_exp - 16;
#undef COEF_SCALE_S
}
void
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
ar5416SetDeltaSlope(struct ath_hal *ah, const struct ieee80211_channel *chan)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
#define INIT_CLOCKMHZSCALED 0x64000000
uint32_t coef_scaled, ds_coef_exp, ds_coef_man;
misc fixups, mostly for code not compiled yet
2009-02-24 00:29:32 +00:00
uint32_t clockMhzScaled;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
CHAN_CENTERS centers;
/* half and quarter rate can divide the scaled clock by 2 or 4 respectively */
/* scale for selected channel bandwidth */
misc fixups, mostly for code not compiled yet
2009-02-24 00:29:32 +00:00
clockMhzScaled = INIT_CLOCKMHZSCALED;
if (IEEE80211_IS_CHAN_TURBO(chan))
clockMhzScaled <<= 1;
else if (IEEE80211_IS_CHAN_HALF(chan))
clockMhzScaled >>= 1;
else if (IEEE80211_IS_CHAN_QUARTER(chan))
clockMhzScaled >>= 2;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/*
* ALGO -> coef = 1e8/fcarrier*fclock/40;
* scaled coef to provide precision for this floating calculation
*/
ar5416GetChannelCenters(ah, chan, &centers);
coef_scaled = clockMhzScaled / centers.synth_center;
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
virgin import of ath hal
2008-11-28 00:03:41 +00:00
ar5416GetDeltaSlopeValues(ah, coef_scaled, &ds_coef_man, &ds_coef_exp);
OS_REG_RMW_FIELD(ah, AR_PHY_TIMING3,
AR_PHY_TIMING3_DSC_MAN, ds_coef_man);
OS_REG_RMW_FIELD(ah, AR_PHY_TIMING3,
AR_PHY_TIMING3_DSC_EXP, ds_coef_exp);
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/*
* For Short GI,
* scaled coeff is 9/10 that of normal coeff
*/
coef_scaled = (9 * coef_scaled)/10;
ar5416GetDeltaSlopeValues(ah, coef_scaled, &ds_coef_man, &ds_coef_exp);
/* for short gi */
OS_REG_RMW_FIELD(ah, AR_PHY_HALFGI,
AR_PHY_HALFGI_DSC_MAN, ds_coef_man);
OS_REG_RMW_FIELD(ah, AR_PHY_HALFGI,
AR_PHY_HALFGI_DSC_EXP, ds_coef_exp);
#undef INIT_CLOCKMHZSCALED
}
/*
* Set a limit on the overall output power. Used for dynamic
* transmit power control and the like.
*
* NB: limit is in units of 0.5 dbM.
*/
HAL_BOOL
ar5416SetTxPowerLimit(struct ath_hal *ah, uint32_t limit)
{
uint16_t dummyXpdGains[2];
AH_PRIVATE(ah)->ah_powerLimit = AH_MIN(limit, MAX_RATE_POWER);
Don't call ar5416SetTransmitPower() directly from ar5416SetTxPowerLimit(); this is incorrect for Kite (AR9285) and any future chipsets that override the EEPROM related routines. It meant that a direct call to set the TX power would call the v14 EEPROM AR5416/AR9280 calibration routines, rather than the v4k EEPROM routines for the AR9285. It thus read the incorrect values from the EEPROM and programmed garbage PDADC and TX power values into the hardware.
2011-03-11 03:46:27 +00:00
return ah->ah_setTxPower(ah, AH_PRIVATE(ah)->ah_curchan,
virgin import of ath hal
2008-11-28 00:03:41 +00:00
dummyXpdGains);
}
HAL_BOOL
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
ar5416GetChipPowerLimits(struct ath_hal *ah,
struct ieee80211_channel *chan)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
struct ath_hal_5212 *ahp = AH5212(ah);
int16_t minPower, maxPower;
/*
* Get Pier table max and min powers.
*/
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (ahp->ah_rfHal->getChannelMaxMinPower(ah, chan, &maxPower, &minPower)) {
/* NB: rf code returns 1/4 dBm units, convert */
chan->ic_maxpower = maxPower / 2;
chan->ic_minpower = minPower / 2;
} else {
HALDEBUG(ah, HAL_DEBUG_ANY,
"%s: no min/max power for %u/0x%x\n",
__func__, chan->ic_freq, chan->ic_flags);
chan->ic_maxpower = AR5416_MAX_RATE_POWER;
chan->ic_minpower = 0;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
HALDEBUG(ah, HAL_DEBUG_RESET,
"Chan %d: MaxPow = %d MinPow = %d\n",
chan->ic_freq, chan->ic_maxpower, chan->ic_minpower);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
return AH_TRUE;
}
/* XXX gag, this is sick */
typedef enum Ar5416_Rates {
rate6mb, rate9mb, rate12mb, rate18mb,
rate24mb, rate36mb, rate48mb, rate54mb,
rate1l, rate2l, rate2s, rate5_5l,
rate5_5s, rate11l, rate11s, rateXr,
rateHt20_0, rateHt20_1, rateHt20_2, rateHt20_3,
rateHt20_4, rateHt20_5, rateHt20_6, rateHt20_7,
rateHt40_0, rateHt40_1, rateHt40_2, rateHt40_3,
rateHt40_4, rateHt40_5, rateHt40_6, rateHt40_7,
rateDupCck, rateDupOfdm, rateExtCck, rateExtOfdm,
Ar5416RateSize
} AR5416_RATES;
/**************************************************************
* ar5416SetTransmitPower
*
* Set the transmit power in the baseband for the given
* operating channel and mode.
*/
Bring back AR9285 support. This fixes most of the issues and should be pretty usable. MFC after: 1 month
2010-02-15 17:49:49 +00:00
HAL_BOOL
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
ar5416SetTransmitPower(struct ath_hal *ah,
const struct ieee80211_channel *chan, uint16_t *rfXpdGain)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
#define POW_SM(_r, _s) (((_r) & 0x3f) << (_s))
#define N(a) (sizeof (a) / sizeof (a[0]))
MODAL_EEP_HEADER *pModal;
struct ath_hal_5212 *ahp = AH5212(ah);
int16_t ratesArray[Ar5416RateSize];
int16_t txPowerIndexOffset = 0;
uint8_t ht40PowerIncForPdadc = 2;
int i;
uint16_t cfgCtl;
uint16_t powerLimit;
uint16_t twiceAntennaReduction;
uint16_t twiceMaxRegulatoryPower;
int16_t maxPower;
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
HAL_EEPROM_v14 *ee = AH_PRIVATE(ah)->ah_eeprom;
struct ar5416eeprom *pEepData = &ee->ee_base;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
HALASSERT(AH_PRIVATE(ah)->ah_eeversion >= AR_EEPROM_VER14_1);
/* Setup info for the actual eeprom */
kill more portability functions that are no longer useful
2009-05-08 00:23:00 +00:00
OS_MEMZERO(ratesArray, sizeof(ratesArray));
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
cfgCtl = ath_hal_getctl(ah, chan);
powerLimit = chan->ic_maxregpower * 2;
twiceAntennaReduction = chan->ic_maxantgain;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
twiceMaxRegulatoryPower = AH_MIN(MAX_RATE_POWER, AH_PRIVATE(ah)->ah_powerLimit);
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
pModal = &pEepData->modalHeader[IEEE80211_IS_CHAN_2GHZ(chan)];
virgin import of ath hal
2008-11-28 00:03:41 +00:00
HALDEBUG(ah, HAL_DEBUG_RESET, "%s Channel=%u CfgCtl=%u\n",
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
__func__,chan->ic_freq, cfgCtl );
virgin import of ath hal
2008-11-28 00:03:41 +00:00
if (IS_EEP_MINOR_V2(ah)) {
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
ht40PowerIncForPdadc = pModal->ht40PowerIncForPdadc;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
if (!ar5416SetPowerPerRateTable(ah, pEepData, chan,
virgin import of ath hal
2008-11-28 00:03:41 +00:00
&ratesArray[0],cfgCtl,
twiceAntennaReduction,
twiceMaxRegulatoryPower, powerLimit)) {
HALDEBUG(ah, HAL_DEBUG_ANY,
"%s: unable to set tx power per rate table\n", __func__);
return AH_FALSE;
}
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
if (!AH5416(ah)->ah_setPowerCalTable(ah, pEepData, chan, &txPowerIndexOffset)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
HALDEBUG(ah, HAL_DEBUG_ANY, "%s: unable to set power table\n",
__func__);
return AH_FALSE;
}
maxPower = AH_MAX(ratesArray[rate6mb], ratesArray[rateHt20_0]);
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_2GHZ(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
maxPower = AH_MAX(maxPower, ratesArray[rate1l]);
}
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_HT40(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
maxPower = AH_MAX(maxPower, ratesArray[rateHt40_0]);
}
ahp->ah_tx6PowerInHalfDbm = maxPower;
AH_PRIVATE(ah)->ah_maxPowerLevel = maxPower;
ahp->ah_txPowerIndexOffset = txPowerIndexOffset;
/*
* txPowerIndexOffset is set by the SetPowerTable() call -
* adjust the rate table (0 offset if rates EEPROM not loaded)
*/
for (i = 0; i < N(ratesArray); i++) {
ratesArray[i] = (int16_t)(txPowerIndexOffset + ratesArray[i]);
if (ratesArray[i] > AR5416_MAX_RATE_POWER)
ratesArray[i] = AR5416_MAX_RATE_POWER;
}
#ifdef AH_EEPROM_DUMP
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
/*
* Dump the rate array whilst it represents the intended dBm*2
* values versus what's being adjusted before being programmed
* in. Keep this in mind if you code up this function and enable
* this debugging; the values won't necessarily be what's being
* programmed into the hardware.
*/
virgin import of ath hal
2008-11-28 00:03:41 +00:00
ar5416PrintPowerPerRate(ah, ratesArray);
#endif
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
/*
* Merlin and later have a power offset, so subtract
* pwr_table_offset * 2 from each value. The default
* power offset is -5 dBm - ie, a register value of 0
* equates to a TX power of -5 dBm.
*/
if (AR_SREV_MERLIN_20_OR_LATER(ah)) {
int8_t pwr_table_offset;
(void) ath_hal_eepromGet(ah, AR_EEP_PWR_TABLE_OFFSET,
&pwr_table_offset);
/* Underflow power gets clamped at raw value 0 */
/* Overflow power gets camped at AR5416_MAX_RATE_POWER */
for (i = 0; i < N(ratesArray); i++) {
/*
* + pwr_table_offset is in dBm
* + ratesArray is in 1/2 dBm
*/
ratesArray[i] -= (pwr_table_offset * 2);
if (ratesArray[i] < 0)
ratesArray[i] = 0;
else if (ratesArray[i] > AR5416_MAX_RATE_POWER)
ratesArray[i] = AR5416_MAX_RATE_POWER;
}
}
/*
* Adjust rates for OLC where needed
*
* The following CCK rates need adjusting when doing 2.4ghz
* CCK transmission.
*
* + rate2s, rate2l, rate1l, rate11s, rate11l, rate5_5s, rate5_5l
* + rateExtCck, rateDupCck
*
* They're adjusted here regardless. The hardware then gets
* programmed as needed. 5GHz operation doesn't program in CCK
* rates for legacy mode but they seem to be initialised for
* HT40 regardless of channel type.
*/
if (AR_SREV_MERLIN_20_OR_LATER(ah) &&
ath_hal_eepromGetFlag(ah, AR_EEP_OL_PWRCTRL)) {
int adj[] = {
rate2s, rate2l, rate1l, rate11s, rate11l,
rate5_5s, rate5_5l, rateExtCck, rateDupCck
};
int cck_ofdm_delta = 2;
int i;
for (i = 0; i < N(adj); i++) {
ratesArray[i] -= cck_ofdm_delta;
if (ratesArray[i] < 0)
ratesArray[i] = 0;
}
}
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Write the OFDM power per rate set */
OS_REG_WRITE(ah, AR_PHY_POWER_TX_RATE1,
POW_SM(ratesArray[rate18mb], 24)
| POW_SM(ratesArray[rate12mb], 16)
| POW_SM(ratesArray[rate9mb], 8)
| POW_SM(ratesArray[rate6mb], 0)
);
OS_REG_WRITE(ah, AR_PHY_POWER_TX_RATE2,
POW_SM(ratesArray[rate54mb], 24)
| POW_SM(ratesArray[rate48mb], 16)
| POW_SM(ratesArray[rate36mb], 8)
| POW_SM(ratesArray[rate24mb], 0)
);
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_2GHZ(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Write the CCK power per rate set */
OS_REG_WRITE(ah, AR_PHY_POWER_TX_RATE3,
POW_SM(ratesArray[rate2s], 24)
| POW_SM(ratesArray[rate2l], 16)
| POW_SM(ratesArray[rateXr], 8) /* XR target power */
| POW_SM(ratesArray[rate1l], 0)
);
OS_REG_WRITE(ah, AR_PHY_POWER_TX_RATE4,
POW_SM(ratesArray[rate11s], 24)
| POW_SM(ratesArray[rate11l], 16)
| POW_SM(ratesArray[rate5_5s], 8)
| POW_SM(ratesArray[rate5_5l], 0)
);
HALDEBUG(ah, HAL_DEBUG_RESET,
"%s AR_PHY_POWER_TX_RATE3=0x%x AR_PHY_POWER_TX_RATE4=0x%x\n",
__func__, OS_REG_READ(ah,AR_PHY_POWER_TX_RATE3),
OS_REG_READ(ah,AR_PHY_POWER_TX_RATE4));
}
/* Write the HT20 power per rate set */
OS_REG_WRITE(ah, AR_PHY_POWER_TX_RATE5,
POW_SM(ratesArray[rateHt20_3], 24)
| POW_SM(ratesArray[rateHt20_2], 16)
| POW_SM(ratesArray[rateHt20_1], 8)
| POW_SM(ratesArray[rateHt20_0], 0)
);
OS_REG_WRITE(ah, AR_PHY_POWER_TX_RATE6,
POW_SM(ratesArray[rateHt20_7], 24)
| POW_SM(ratesArray[rateHt20_6], 16)
| POW_SM(ratesArray[rateHt20_5], 8)
| POW_SM(ratesArray[rateHt20_4], 0)
);
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_HT40(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Write the HT40 power per rate set */
/* Correct PAR difference between HT40 and HT20/LEGACY */
OS_REG_WRITE(ah, AR_PHY_POWER_TX_RATE7,
POW_SM(ratesArray[rateHt40_3] + ht40PowerIncForPdadc, 24)
| POW_SM(ratesArray[rateHt40_2] + ht40PowerIncForPdadc, 16)
| POW_SM(ratesArray[rateHt40_1] + ht40PowerIncForPdadc, 8)
| POW_SM(ratesArray[rateHt40_0] + ht40PowerIncForPdadc, 0)
);
OS_REG_WRITE(ah, AR_PHY_POWER_TX_RATE8,
POW_SM(ratesArray[rateHt40_7] + ht40PowerIncForPdadc, 24)
| POW_SM(ratesArray[rateHt40_6] + ht40PowerIncForPdadc, 16)
| POW_SM(ratesArray[rateHt40_5] + ht40PowerIncForPdadc, 8)
| POW_SM(ratesArray[rateHt40_4] + ht40PowerIncForPdadc, 0)
);
/* Write the Dup/Ext 40 power per rate set */
OS_REG_WRITE(ah, AR_PHY_POWER_TX_RATE9,
POW_SM(ratesArray[rateExtOfdm], 24)
| POW_SM(ratesArray[rateExtCck], 16)
| POW_SM(ratesArray[rateDupOfdm], 8)
| POW_SM(ratesArray[rateDupCck], 0)
);
}
/* Write the Power subtraction for dynamic chain changing, for per-packet powertx */
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
OS_REG_WRITE(ah, AR_PHY_POWER_TX_SUB,
POW_SM(pModal->pwrDecreaseFor3Chain, 6)
| POW_SM(pModal->pwrDecreaseFor2Chain, 0)
);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
return AH_TRUE;
#undef POW_SM
#undef N
}
/*
* Exported call to check for a recent gain reading and return
* the current state of the thermal calibration gain engine.
*/
HAL_RFGAIN
ar5416GetRfgain(struct ath_hal *ah)
{
return HAL_RFGAIN_INACTIVE;
}
/*
* Places all of hardware into reset
*/
HAL_BOOL
ar5416Disable(struct ath_hal *ah)
{
if (!ar5212SetPowerMode(ah, HAL_PM_AWAKE, AH_TRUE))
return AH_FALSE;
return ar5416SetResetReg(ah, HAL_RESET_COLD);
}
/*
* Places the PHY and Radio chips into reset. A full reset
* must be called to leave this state. The PCI/MAC/PCU are
* not placed into reset as we must receive interrupt to
* re-enable the hardware.
*/
HAL_BOOL
ar5416PhyDisable(struct ath_hal *ah)
{
return ar5416SetResetReg(ah, HAL_RESET_WARM);
}
/*
* Write the given reset bit mask into the reset register
*/
HAL_BOOL
ar5416SetResetReg(struct ath_hal *ah, uint32_t type)
{
switch (type) {
case HAL_RESET_POWER_ON:
return ar5416SetResetPowerOn(ah);
case HAL_RESET_WARM:
case HAL_RESET_COLD:
return ar5416SetReset(ah, type);
default:
preliminary ar9280 support: o add 9280 attach that sets up ini, cal, etc. o new rf backend for 9280 and later parts o split ini setup and spur mitigation support out to methods and provide 9280-specific support o minor fixups to shared code to handle 9280-specific work Obtained from: Atheros (ini values and some code)
2009-03-12 18:18:28 +00:00
HALASSERT(AH_FALSE);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
return AH_FALSE;
}
}
static HAL_BOOL
ar5416SetResetPowerOn(struct ath_hal *ah)
{
/* Power On Reset (Hard Reset) */
/*
* Set force wake
*
* If the MAC was running, previously calling
* reset will wake up the MAC but it may go back to sleep
* before we can start polling.
* Set force wake stops that
* This must be called before initiating a hard reset.
*/
OS_REG_WRITE(ah, AR_RTC_FORCE_WAKE,
AR_RTC_FORCE_WAKE_EN | AR_RTC_FORCE_WAKE_ON_INT);
/*
* RTC reset and clear
*/
Bring in a couple of fixes from the Linux ath9k related to chip hangs. While there, try to make the register write pattern look like what's done by ath9k. MFC after: 3 days
2010-06-01 15:33:10 +00:00
OS_REG_WRITE(ah, AR_RC, AR_RC_AHB);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
OS_REG_WRITE(ah, AR_RTC_RESET, 0);
OS_DELAY(20);
Bring in a couple of fixes from the Linux ath9k related to chip hangs. While there, try to make the register write pattern look like what's done by ath9k. MFC after: 3 days
2010-06-01 15:33:10 +00:00
OS_REG_WRITE(ah, AR_RC, 0);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
OS_REG_WRITE(ah, AR_RTC_RESET, 1);
/*
* Poll till RTC is ON
*/
if (!ath_hal_wait(ah, AR_RTC_STATUS, AR_RTC_PM_STATUS_M, AR_RTC_STATUS_ON)) {
HALDEBUG(ah, HAL_DEBUG_ANY, "%s: RTC not waking up\n", __func__);
return AH_FALSE;
}
return ar5416SetReset(ah, HAL_RESET_COLD);
}
static HAL_BOOL
ar5416SetReset(struct ath_hal *ah, int type)
{
Fix ar5416 and later parts on big-endian platforms: setup the h/w byte swizzler using the same technique used everywhere else. Approved by: re (kib)
2009-07-07 18:11:05 +00:00
uint32_t tmpReg, mask;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/*
* Force wake
*/
OS_REG_WRITE(ah, AR_RTC_FORCE_WAKE,
AR_RTC_FORCE_WAKE_EN | AR_RTC_FORCE_WAKE_ON_INT);
/*
* Reset AHB
*/
tmpReg = OS_REG_READ(ah, AR_INTR_SYNC_CAUSE);
if (tmpReg & (AR_INTR_SYNC_LOCAL_TIMEOUT|AR_INTR_SYNC_RADM_CPL_TIMEOUT)) {
OS_REG_WRITE(ah, AR_INTR_SYNC_ENABLE, 0);
OS_REG_WRITE(ah, AR_RC, AR_RC_AHB|AR_RC_HOSTIF);
} else {
OS_REG_WRITE(ah, AR_RC, AR_RC_AHB);
}
/*
* Set Mac(BB,Phy) Warm Reset
*/
switch (type) {
case HAL_RESET_WARM:
OS_REG_WRITE(ah, AR_RTC_RC, AR_RTC_RC_MAC_WARM);
break;
preliminary ar9280 support: o add 9280 attach that sets up ini, cal, etc. o new rf backend for 9280 and later parts o split ini setup and spur mitigation support out to methods and provide 9280-specific support o minor fixups to shared code to handle 9280-specific work Obtained from: Atheros (ini values and some code)
2009-03-12 18:18:28 +00:00
case HAL_RESET_COLD:
virgin import of ath hal
2008-11-28 00:03:41 +00:00
OS_REG_WRITE(ah, AR_RTC_RC, AR_RTC_RC_MAC_WARM|AR_RTC_RC_MAC_COLD);
break;
preliminary ar9280 support: o add 9280 attach that sets up ini, cal, etc. o new rf backend for 9280 and later parts o split ini setup and spur mitigation support out to methods and provide 9280-specific support o minor fixups to shared code to handle 9280-specific work Obtained from: Atheros (ini values and some code)
2009-03-12 18:18:28 +00:00
default:
HALASSERT(AH_FALSE);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
break;
}
/*
* Clear resets and force wakeup
*/
OS_REG_WRITE(ah, AR_RTC_RC, 0);
if (!ath_hal_wait(ah, AR_RTC_RC, AR_RTC_RC_M, 0)) {
HALDEBUG(ah, HAL_DEBUG_ANY, "%s: RTC stuck in MAC reset\n", __func__);
return AH_FALSE;
}
/* Clear AHB reset */
OS_REG_WRITE(ah, AR_RC, 0);
Fix ar5416 and later parts on big-endian platforms: setup the h/w byte swizzler using the same technique used everywhere else. Approved by: re (kib)
2009-07-07 18:11:05 +00:00
if (type == HAL_RESET_COLD) {
if (isBigEndian()) {
/*
* Set CFG, little-endian for register
* and descriptor accesses.
*/
mask = INIT_CONFIG_STATUS | AR_CFG_SWRD | AR_CFG_SWRG;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
#ifndef AH_NEED_DESC_SWAP
Fix ar5416 and later parts on big-endian platforms: setup the h/w byte swizzler using the same technique used everywhere else. Approved by: re (kib)
2009-07-07 18:11:05 +00:00
mask |= AR_CFG_SWTD;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
#endif
Fix ar5416 and later parts on big-endian platforms: setup the h/w byte swizzler using the same technique used everywhere else. Approved by: re (kib)
2009-07-07 18:11:05 +00:00
HALDEBUG(ah, HAL_DEBUG_RESET,
"%s Applying descriptor swap\n", __func__);
OS_REG_WRITE(ah, AR_CFG, LE_READ_4(&mask));
} else
OS_REG_WRITE(ah, AR_CFG, INIT_CONFIG_STATUS);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
ar5416InitPLL(ah, AH_NULL);
return AH_TRUE;
}
Break out the chainmask init code into a new function - ar5416InitChainMasks() . ath9k does a few different things here during config - if it's an early AR5416 with two chains, it enables all three chains for calibration and then restores the chainmask to the original values after initial calibration has completed. The reason behind this commit is to begin breaking out the chainmask configuration for this specific reason; follow-up commits will add the chainmask restore in the ar5416Reset() routine.
2011-01-26 10:08:37 +00:00
void
ar5416InitChainMasks(struct ath_hal *ah)
{
if (AH5416(ah)->ah_rx_chainmask == 0x5 ||
AH5416(ah)->ah_tx_chainmask == 0x5)
OS_REG_WRITE(ah, AR_PHY_ANALOG_SWAP, AR_PHY_SWAP_ALT_CHAIN);
/* Setup Chain Masks */
OS_REG_WRITE(ah, AR_PHY_RX_CHAINMASK, AH5416(ah)->ah_rx_chainmask);
OS_REG_WRITE(ah, AR_PHY_CAL_CHAINMASK, AH5416(ah)->ah_rx_chainmask);
OS_REG_WRITE(ah, AR_SELFGEN_MASK, AH5416(ah)->ah_tx_chainmask);
}
Add ar5416RestoreChainMask() which will undo any AR5416 specific chainmask overriding after calibration. This will get set for other two chain radios, such as AR9280 and later on, AR9287. It should however be a nul operation.
2011-01-26 10:48:29 +00:00
void
ar5416RestoreChainMask(struct ath_hal *ah)
{
int rx_chainmask = AH5416(ah)->ah_rx_chainmask;
if ((rx_chainmask == 0x5) || (rx_chainmask == 0x3)) {
OS_REG_WRITE(ah, AR_PHY_RX_CHAINMASK, rx_chainmask);
OS_REG_WRITE(ah, AR_PHY_CAL_CHAINMASK, rx_chainmask);
}
}
Initialise the chainmask from the EEPROM rather than the hard-coded defaults. The defaults enabled three chains on the AR5416 even if the card has two chains. This restores that and ensures that only the correct TX/RX chainmasks are used. When HT modes are enabled, all TX chains will be correctly enabled. This should now enable analog chain swapping with 2-chain cards. I'm not sure if this is needed for just the AR5416 or whether it also applies to AR9160, AR9280 and AR9287 (later on); I'll have to get clarification.
2011-01-27 09:26:37 +00:00
/*
* Update the chainmask based on the current channel configuration.
*
* XXX ath9k checks bluetooth co-existence here
* XXX ath9k checks whether the current state is "off-channel".
* XXX ath9k sticks the hardware into 1x1 mode for legacy;
* we're going to leave multi-RX on for multi-path cancellation.
*/
static void
ar5416UpdateChainMasks(struct ath_hal *ah, HAL_BOOL is_ht)
{
struct ath_hal_private *ahpriv = AH_PRIVATE(ah);
HAL_CAPABILITIES *pCap = &ahpriv->ah_caps;
if (is_ht) {
AH5416(ah)->ah_tx_chainmask = pCap->halTxChainMask;
} else {
AH5416(ah)->ah_tx_chainmask = 1;
}
AH5416(ah)->ah_rx_chainmask = pCap->halRxChainMask;
HALDEBUG(ah, HAL_DEBUG_ANY, "TX chainmask: 0x%x; RX chainmask: 0x%x\n",
AH5416(ah)->ah_tx_chainmask,
AH5416(ah)->ah_rx_chainmask);
}
Add ar5416RestoreChainMask() which will undo any AR5416 specific chainmask overriding after calibration. This will get set for other two chain radios, such as AR9280 and later on, AR9287. It should however be a nul operation.
2011-01-26 10:48:29 +00:00
virgin import of ath hal
2008-11-28 00:03:41 +00:00
#ifndef IS_5GHZ_FAST_CLOCK_EN
#define IS_5GHZ_FAST_CLOCK_EN(ah, chan) AH_FALSE
#endif
static void
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
ar5416InitPLL(struct ath_hal *ah, const struct ieee80211_channel *chan)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
uint32_t pll;
preliminary ar9280 support: o add 9280 attach that sets up ini, cal, etc. o new rf backend for 9280 and later parts o split ini setup and spur mitigation support out to methods and provide 9280-specific support o minor fixups to shared code to handle 9280-specific work Obtained from: Atheros (ini values and some code)
2009-03-12 18:18:28 +00:00
if (AR_SREV_MERLIN_20(ah) &&
chan != AH_NULL && IEEE80211_IS_CHAN_5GHZ(chan)) {
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
/*
* PLL WAR for Merlin 2.0/2.1
* When doing fast clock, set PLL to 0x142c
* Else, set PLL to 0x2850 to prevent reset-to-reset variation
*/
pll = IS_5GHZ_FAST_CLOCK_EN(ah, chan) ? 0x142c : 0x2850;
} else if (AR_SREV_MERLIN_10_OR_LATER(ah)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
pll = SM(0x5, AR_RTC_SOWL_PLL_REFDIV);
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (chan != AH_NULL) {
if (IEEE80211_IS_CHAN_HALF(chan))
pll |= SM(0x1, AR_RTC_SOWL_PLL_CLKSEL);
else if (IEEE80211_IS_CHAN_QUARTER(chan))
pll |= SM(0x2, AR_RTC_SOWL_PLL_CLKSEL);
else if (IEEE80211_IS_CHAN_5GHZ(chan))
pll |= SM(0x28, AR_RTC_SOWL_PLL_DIV);
else
pll |= SM(0x2c, AR_RTC_SOWL_PLL_DIV);
} else
virgin import of ath hal
2008-11-28 00:03:41 +00:00
pll |= SM(0x2c, AR_RTC_SOWL_PLL_DIV);
} else if (AR_SREV_SOWL_10_OR_LATER(ah)) {
pll = SM(0x5, AR_RTC_SOWL_PLL_REFDIV);
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (chan != AH_NULL) {
if (IEEE80211_IS_CHAN_HALF(chan))
pll |= SM(0x1, AR_RTC_SOWL_PLL_CLKSEL);
else if (IEEE80211_IS_CHAN_QUARTER(chan))
pll |= SM(0x2, AR_RTC_SOWL_PLL_CLKSEL);
else if (IEEE80211_IS_CHAN_5GHZ(chan))
pll |= SM(0x50, AR_RTC_SOWL_PLL_DIV);
else
pll |= SM(0x58, AR_RTC_SOWL_PLL_DIV);
} else
virgin import of ath hal
2008-11-28 00:03:41 +00:00
pll |= SM(0x58, AR_RTC_SOWL_PLL_DIV);
} else {
pll = AR_RTC_PLL_REFDIV_5 | AR_RTC_PLL_DIV2;
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (chan != AH_NULL) {
if (IEEE80211_IS_CHAN_HALF(chan))
pll |= SM(0x1, AR_RTC_PLL_CLKSEL);
else if (IEEE80211_IS_CHAN_QUARTER(chan))
pll |= SM(0x2, AR_RTC_PLL_CLKSEL);
else if (IEEE80211_IS_CHAN_5GHZ(chan))
pll |= SM(0xa, AR_RTC_PLL_DIV);
else
pll |= SM(0xb, AR_RTC_PLL_DIV);
} else
virgin import of ath hal
2008-11-28 00:03:41 +00:00
pll |= SM(0xb, AR_RTC_PLL_DIV);
}
OS_REG_WRITE(ah, AR_RTC_PLL_CONTROL, pll);
/* TODO:
* For multi-band owl, switch between bands by reiniting the PLL.
*/
OS_DELAY(RTC_PLL_SETTLE_DELAY);
OS_REG_WRITE(ah, AR_RTC_SLEEP_CLK, AR_RTC_SLEEP_DERIVED_CLK);
}
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
static void
ar5416SetDefGainValues(struct ath_hal *ah,
const MODAL_EEP_HEADER *pModal,
const struct ar5416eeprom *eep,
uint8_t txRxAttenLocal, int regChainOffset, int i)
{
if (IS_EEP_MINOR_V3(ah)) {
txRxAttenLocal = pModal->txRxAttenCh[i];
if (AR_SREV_MERLIN_20_OR_LATER(ah)) {
OS_REG_RMW_FIELD(ah, AR_PHY_GAIN_2GHZ + regChainOffset,
AR_PHY_GAIN_2GHZ_XATTEN1_MARGIN,
pModal->bswMargin[i]);
OS_REG_RMW_FIELD(ah, AR_PHY_GAIN_2GHZ + regChainOffset,
AR_PHY_GAIN_2GHZ_XATTEN1_DB,
pModal->bswAtten[i]);
OS_REG_RMW_FIELD(ah, AR_PHY_GAIN_2GHZ + regChainOffset,
AR_PHY_GAIN_2GHZ_XATTEN2_MARGIN,
pModal->xatten2Margin[i]);
OS_REG_RMW_FIELD(ah, AR_PHY_GAIN_2GHZ + regChainOffset,
AR_PHY_GAIN_2GHZ_XATTEN2_DB,
pModal->xatten2Db[i]);
} else {
Do a bit of spring cleaning in the board setup code, just to bring it in line with the rest of the register initialisation. I've verified that the 2/5ghz board values written to the chip match what was previously written.
2011-03-22 00:43:58 +00:00
OS_REG_RMW_FIELD(ah, AR_PHY_GAIN_2GHZ + regChainOffset,
AR_PHY_GAIN_2GHZ_BSW_MARGIN,
pModal->bswMargin[i]);
OS_REG_RMW_FIELD(ah, AR_PHY_GAIN_2GHZ + regChainOffset,
AR_PHY_GAIN_2GHZ_BSW_ATTEN,
pModal->bswAtten[i]);
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
}
}
if (AR_SREV_MERLIN_20_OR_LATER(ah)) {
OS_REG_RMW_FIELD(ah,
AR_PHY_RXGAIN + regChainOffset,
AR9280_PHY_RXGAIN_TXRX_ATTEN, txRxAttenLocal);
OS_REG_RMW_FIELD(ah,
AR_PHY_RXGAIN + regChainOffset,
AR9280_PHY_RXGAIN_TXRX_MARGIN, pModal->rxTxMarginCh[i]);
} else {
Do a bit of spring cleaning in the board setup code, just to bring it in line with the rest of the register initialisation. I've verified that the 2/5ghz board values written to the chip match what was previously written.
2011-03-22 00:43:58 +00:00
OS_REG_RMW_FIELD(ah,
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
AR_PHY_RXGAIN + regChainOffset,
Do a bit of spring cleaning in the board setup code, just to bring it in line with the rest of the register initialisation. I've verified that the 2/5ghz board values written to the chip match what was previously written.
2011-03-22 00:43:58 +00:00
AR_PHY_RXGAIN_TXRX_ATTEN, txRxAttenLocal);
OS_REG_RMW_FIELD(ah,
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
AR_PHY_GAIN_2GHZ + regChainOffset,
Do a bit of spring cleaning in the board setup code, just to bring it in line with the rest of the register initialisation. I've verified that the 2/5ghz board values written to the chip match what was previously written.
2011-03-22 00:43:58 +00:00
AR_PHY_GAIN_2GHZ_RXTX_MARGIN, pModal->rxTxMarginCh[i]);
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
}
}
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
/*
* Get the register chain offset for the given chain.
*
* Take into account the register chain swapping with AR5416 v2.0.
*
* XXX make sure that the reg chain swapping is only done for
* XXX AR5416 v2.0 or greater, and not later chips?
*/
int
ar5416GetRegChainOffset(struct ath_hal *ah, int i)
{
int regChainOffset;
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
if (AR_SREV_OWL_20_OR_LATER(ah) &&
(AH5416(ah)->ah_rx_chainmask == 0x5 ||
AH5416(ah)->ah_tx_chainmask == 0x5) && (i != 0)) {
/* Regs are swapped from chain 2 to 1 for 5416 2_0 with
* only chains 0 and 2 populated
*/
regChainOffset = (i == 1) ? 0x2000 : 0x1000;
} else {
regChainOffset = i * 0x1000;
}
return regChainOffset;
}
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/*
* Read EEPROM header info and program the device for correct operation
* given the channel value.
*/
Bring back AR9285 support. This fixes most of the issues and should be pretty usable. MFC after: 1 month
2010-02-15 17:49:49 +00:00
HAL_BOOL
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
ar5416SetBoardValues(struct ath_hal *ah, const struct ieee80211_channel *chan)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
const HAL_EEPROM_v14 *ee = AH_PRIVATE(ah)->ah_eeprom;
const struct ar5416eeprom *eep = &ee->ee_base;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
const MODAL_EEP_HEADER *pModal;
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
int i, regChainOffset;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
uint8_t txRxAttenLocal; /* workaround for eeprom versions <= 14.2 */
HALASSERT(AH_PRIVATE(ah)->ah_eeversion >= AR_EEPROM_VER14_1);
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
pModal = &eep->modalHeader[IEEE80211_IS_CHAN_2GHZ(chan)];
virgin import of ath hal
2008-11-28 00:03:41 +00:00
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
/* NB: workaround for eeprom versions <= 14.2 */
txRxAttenLocal = IEEE80211_IS_CHAN_2GHZ(chan) ? 23 : 44;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
OS_REG_WRITE(ah, AR_PHY_SWITCH_COM, pModal->antCtrlCommon);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
for (i = 0; i < AR5416_MAX_CHAINS; i++) {
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
if (AR_SREV_MERLIN(ah)) {
if (i >= 2) break;
}
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
regChainOffset = ar5416GetRegChainOffset(ah, i);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
OS_REG_WRITE(ah, AR_PHY_SWITCH_CHAIN_0 + regChainOffset, pModal->antCtrlChain[i]);
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
virgin import of ath hal
2008-11-28 00:03:41 +00:00
OS_REG_WRITE(ah, AR_PHY_TIMING_CTRL4 + regChainOffset,
(OS_REG_READ(ah, AR_PHY_TIMING_CTRL4 + regChainOffset) &
~(AR_PHY_TIMING_CTRL4_IQCORR_Q_Q_COFF | AR_PHY_TIMING_CTRL4_IQCORR_Q_I_COFF)) |
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
SM(pModal->iqCalICh[i], AR_PHY_TIMING_CTRL4_IQCORR_Q_I_COFF) |
SM(pModal->iqCalQCh[i], AR_PHY_TIMING_CTRL4_IQCORR_Q_Q_COFF));
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/*
* Large signal upgrade.
Update to later code from my repository: o many bug fixes o add new periodic calibration api o break up 5416 periodic calibration code in preparation for 928x o move get noise floor to rf backends o 5416-specific ani (still disabled) o modularize 5210 eeprom format a la other eeprom formats o start cleaning up regdomain code o prepare for proper 1/2 and 1/4 width channel support o bring back 900MHz card support o clean up 5212 rf version handling o add 1/2 and 1/4 width channel support for 5212 parts o split 5212 rfgain handling out o improve ani debugging o add AH_USE_INIPDGAIN compile option o purge a bunch of dead 5212 state o add 1/2 and 1/4 rate modes o remove HAL_CAP_CHAN_HALFRATE and HAL_CAP_CHAN_QUARTERRATE; the same info can now be deduced from the set of supported modes
2008-11-28 00:48:05 +00:00
* XXX update
virgin import of ath hal
2008-11-28 00:03:41 +00:00
*/
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
if ((i == 0) || AR_SREV_OWL_20_OR_LATER(ah))
ar5416SetDefGainValues(ah, pModal, eep, txRxAttenLocal, regChainOffset, i);
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
if (AR_SREV_MERLIN_20_OR_LATER(ah)) {
if (IEEE80211_IS_CHAN_2GHZ(chan)) {
OS_A_REG_RMW_FIELD(ah, AR_AN_RF2G1_CH0, AR_AN_RF2G1_CH0_OB, pModal->ob);
OS_A_REG_RMW_FIELD(ah, AR_AN_RF2G1_CH0, AR_AN_RF2G1_CH0_DB, pModal->db);
OS_A_REG_RMW_FIELD(ah, AR_AN_RF2G1_CH1, AR_AN_RF2G1_CH1_OB, pModal->ob_ch1);
OS_A_REG_RMW_FIELD(ah, AR_AN_RF2G1_CH1, AR_AN_RF2G1_CH1_DB, pModal->db_ch1);
} else {
OS_A_REG_RMW_FIELD(ah, AR_AN_RF5G1_CH0, AR_AN_RF5G1_CH0_OB5, pModal->ob);
OS_A_REG_RMW_FIELD(ah, AR_AN_RF5G1_CH0, AR_AN_RF5G1_CH0_DB5, pModal->db);
OS_A_REG_RMW_FIELD(ah, AR_AN_RF5G1_CH1, AR_AN_RF5G1_CH1_OB5, pModal->ob_ch1);
OS_A_REG_RMW_FIELD(ah, AR_AN_RF5G1_CH1, AR_AN_RF5G1_CH1_DB5, pModal->db_ch1);
}
OS_A_REG_RMW_FIELD(ah, AR_AN_TOP2, AR_AN_TOP2_XPABIAS_LVL, pModal->xpaBiasLvl);
OS_A_REG_RMW_FIELD(ah, AR_AN_TOP2, AR_AN_TOP2_LOCALBIAS, pModal->flagBits & AR5416_EEP_FLAG_LOCALBIAS);
OS_A_REG_RMW_FIELD(ah, AR_PHY_XPA_CFG, AR_PHY_FORCE_XPA_CFG, pModal->flagBits & AR5416_EEP_FLAG_FORCEXPAON);
}
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
OS_REG_RMW_FIELD(ah, AR_PHY_SETTLING, AR_PHY_SETTLING_SWITCH, pModal->switchSettling);
OS_REG_RMW_FIELD(ah, AR_PHY_DESIRED_SZ, AR_PHY_DESIRED_SZ_ADC, pModal->adcDesiredSize);
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
if (! AR_SREV_MERLIN_20_OR_LATER(ah))
OS_REG_RMW_FIELD(ah, AR_PHY_DESIRED_SZ, AR_PHY_DESIRED_SZ_PGA, pModal->pgaDesiredSize);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
OS_REG_WRITE(ah, AR_PHY_RF_CTL4,
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
SM(pModal->txEndToXpaOff, AR_PHY_RF_CTL4_TX_END_XPAA_OFF)
| SM(pModal->txEndToXpaOff, AR_PHY_RF_CTL4_TX_END_XPAB_OFF)
| SM(pModal->txFrameToXpaOn, AR_PHY_RF_CTL4_FRAME_XPAA_ON)
| SM(pModal->txFrameToXpaOn, AR_PHY_RF_CTL4_FRAME_XPAB_ON));
OS_REG_RMW_FIELD(ah, AR_PHY_RF_CTL3, AR_PHY_TX_END_TO_A2_RX_ON, pModal->txEndToRxOn);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
if (AR_SREV_MERLIN_10_OR_LATER(ah)) {
OS_REG_RMW_FIELD(ah, AR_PHY_CCA, AR9280_PHY_CCA_THRESH62,
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
pModal->thresh62);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
OS_REG_RMW_FIELD(ah, AR_PHY_EXT_CCA0, AR_PHY_EXT_CCA0_THRESH62,
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
pModal->thresh62);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
} else {
OS_REG_RMW_FIELD(ah, AR_PHY_CCA, AR_PHY_CCA_THRESH62,
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
pModal->thresh62);
Set the "right" CCA register. Obtained From: ath9k
2011-03-22 05:47:48 +00:00
OS_REG_RMW_FIELD(ah, AR_PHY_EXT_CCA, AR_PHY_EXT_CCA_THRESH62,
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
pModal->thresh62);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
/* Minor Version Specific application */
if (IS_EEP_MINOR_V2(ah)) {
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
OS_REG_RMW_FIELD(ah, AR_PHY_RF_CTL2, AR_PHY_TX_FRAME_TO_DATA_START, pModal->txFrameToDataStart);
OS_REG_RMW_FIELD(ah, AR_PHY_RF_CTL2, AR_PHY_TX_FRAME_TO_PA_ON, pModal->txFrameToPaOn);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
if (IS_EEP_MINOR_V3(ah) && IEEE80211_IS_CHAN_HT40(chan))
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Overwrite switch settling with HT40 value */
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
OS_REG_RMW_FIELD(ah, AR_PHY_SETTLING, AR_PHY_SETTLING_SWITCH, pModal->swSettleHt40);
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
if (AR_SREV_MERLIN_20_OR_LATER(ah) && EEP_MINOR(ah) >= AR5416_EEP_MINOR_VER_19)
OS_REG_RMW_FIELD(ah, AR_PHY_CCK_TX_CTRL, AR_PHY_CCK_TX_CTRL_TX_DAC_SCALE_CCK, pModal->miscBits);
if (AR_SREV_MERLIN_20(ah) && EEP_MINOR(ah) >= AR5416_EEP_MINOR_VER_20) {
if (IEEE80211_IS_CHAN_2GHZ(chan))
OS_A_REG_RMW_FIELD(ah, AR_AN_TOP1, AR_AN_TOP1_DACIPMODE, eep->baseEepHeader.dacLpMode);
else if (eep->baseEepHeader.dacHiPwrMode_5G)
OS_A_REG_RMW_FIELD(ah, AR_AN_TOP1, AR_AN_TOP1_DACIPMODE, 0);
else
OS_A_REG_RMW_FIELD(ah, AR_AN_TOP1, AR_AN_TOP1_DACIPMODE, eep->baseEepHeader.dacLpMode);
OS_REG_RMW_FIELD(ah, AR_PHY_FRAME_CTL, AR_PHY_FRAME_CTL_TX_CLIP, pModal->miscBits >> 2);
OS_REG_RMW_FIELD(ah, AR_PHY_TX_PWRCTRL9, AR_PHY_TX_DESIRED_SCALE_CCK, eep->baseEepHeader.desiredScaleCCK);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
Add in some AR9280 specific board configuration options. * The existing radio config code was for the AR5416/AR9160 and missed out on some of the AR9280 specific stuff. Include said stuff from ath9k. * Refactor out the gain control settings into a new function, again pilfered from ath9k. * Use the analog register RMW macro when touching analog registers. Obtained from: Linux ath9k
2011-02-07 22:00:31 +00:00
virgin import of ath hal
2008-11-28 00:03:41 +00:00
return AH_TRUE;
}
/*
* Helper functions common for AP/CB/XB
*/
/*
* ar5416SetPowerPerRateTable
*
* Sets the transmit power in the baseband for the given
* operating channel and mode.
*/
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
static HAL_BOOL
virgin import of ath hal
2008-11-28 00:03:41 +00:00
ar5416SetPowerPerRateTable(struct ath_hal *ah, struct ar5416eeprom *pEepData,
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
const struct ieee80211_channel *chan,
virgin import of ath hal
2008-11-28 00:03:41 +00:00
int16_t *ratesArray, uint16_t cfgCtl,
uint16_t AntennaReduction,
uint16_t twiceMaxRegulatoryPower,
uint16_t powerLimit)
{
#define N(a) (sizeof(a)/sizeof(a[0]))
/* Local defines to distinguish between extension and control CTL's */
#define EXT_ADDITIVE (0x8000)
#define CTL_11A_EXT (CTL_11A | EXT_ADDITIVE)
#define CTL_11G_EXT (CTL_11G | EXT_ADDITIVE)
#define CTL_11B_EXT (CTL_11B | EXT_ADDITIVE)
uint16_t twiceMaxEdgePower = AR5416_MAX_RATE_POWER;
int i;
int16_t twiceLargestAntenna;
CAL_CTL_DATA *rep;
CAL_TARGET_POWER_LEG targetPowerOfdm, targetPowerCck = {0, {0, 0, 0, 0}};
CAL_TARGET_POWER_LEG targetPowerOfdmExt = {0, {0, 0, 0, 0}}, targetPowerCckExt = {0, {0, 0, 0, 0}};
CAL_TARGET_POWER_HT targetPowerHt20, targetPowerHt40 = {0, {0, 0, 0, 0}};
int16_t scaledPower, minCtlPower;
#define SUB_NUM_CTL_MODES_AT_5G_40 2 /* excluding HT40, EXT-OFDM */
#define SUB_NUM_CTL_MODES_AT_2G_40 3 /* excluding HT40, EXT-OFDM, EXT-CCK */
static const uint16_t ctlModesFor11a[] = {
CTL_11A, CTL_5GHT20, CTL_11A_EXT, CTL_5GHT40
};
static const uint16_t ctlModesFor11g[] = {
CTL_11B, CTL_11G, CTL_2GHT20, CTL_11B_EXT, CTL_11G_EXT, CTL_2GHT40
};
const uint16_t *pCtlMode;
uint16_t numCtlModes, ctlMode, freq;
CHAN_CENTERS centers;
ar5416GetChannelCenters(ah, chan, &centers);
/* Compute TxPower reduction due to Antenna Gain */
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
twiceLargestAntenna = AH_MAX(AH_MAX(
pEepData->modalHeader[IEEE80211_IS_CHAN_2GHZ(chan)].antennaGainCh[0],
pEepData->modalHeader[IEEE80211_IS_CHAN_2GHZ(chan)].antennaGainCh[1]),
pEepData->modalHeader[IEEE80211_IS_CHAN_2GHZ(chan)].antennaGainCh[2]);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
#if 0
/* Turn it back on if we need to calculate per chain antenna gain reduction */
/* Use only if the expected gain > 6dbi */
/* Chain 0 is always used */
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
twiceLargestAntenna = pEepData->modalHeader[IEEE80211_IS_CHAN_2GHZ(chan)].antennaGainCh[0];
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Look at antenna gains of Chains 1 and 2 if the TX mask is set */
if (ahp->ah_tx_chainmask & 0x2)
twiceLargestAntenna = AH_MAX(twiceLargestAntenna,
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
pEepData->modalHeader[IEEE80211_IS_CHAN_2GHZ(chan)].antennaGainCh[1]);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
if (ahp->ah_tx_chainmask & 0x4)
twiceLargestAntenna = AH_MAX(twiceLargestAntenna,
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
pEepData->modalHeader[IEEE80211_IS_CHAN_2GHZ(chan)].antennaGainCh[2]);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
#endif
twiceLargestAntenna = (int16_t)AH_MIN((AntennaReduction) - twiceLargestAntenna, 0);
/* XXX setup for 5212 use (really used?) */
ath_hal_eepromSet(ah,
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
IEEE80211_IS_CHAN_2GHZ(chan) ? AR_EEP_ANTGAINMAX_2 : AR_EEP_ANTGAINMAX_5,
virgin import of ath hal
2008-11-28 00:03:41 +00:00
twiceLargestAntenna);
/*
* scaledPower is the minimum of the user input power level and
* the regulatory allowed power level
*/
scaledPower = AH_MIN(powerLimit, twiceMaxRegulatoryPower + twiceLargestAntenna);
/* Reduce scaled Power by number of chains active to get to per chain tx power level */
/* TODO: better value than these? */
switch (owl_get_ntxchains(AH5416(ah)->ah_tx_chainmask)) {
case 1:
break;
case 2:
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
scaledPower -= pEepData->modalHeader[IEEE80211_IS_CHAN_2GHZ(chan)].pwrDecreaseFor2Chain;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
break;
case 3:
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
scaledPower -= pEepData->modalHeader[IEEE80211_IS_CHAN_2GHZ(chan)].pwrDecreaseFor3Chain;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
break;
default:
return AH_FALSE; /* Unsupported number of chains */
}
scaledPower = AH_MAX(0, scaledPower);
/* Get target powers from EEPROM - our baseline for TX Power */
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_2GHZ(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Setup for CTL modes */
numCtlModes = N(ctlModesFor11g) - SUB_NUM_CTL_MODES_AT_2G_40; /* CTL_11B, CTL_11G, CTL_2GHT20 */
pCtlMode = ctlModesFor11g;
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
ar5416GetTargetPowersLeg(ah, chan, pEepData->calTargetPowerCck,
AR5416_NUM_2G_CCK_TARGET_POWERS, &targetPowerCck, 4, AH_FALSE);
ar5416GetTargetPowersLeg(ah, chan, pEepData->calTargetPower2G,
AR5416_NUM_2G_20_TARGET_POWERS, &targetPowerOfdm, 4, AH_FALSE);
ar5416GetTargetPowers(ah, chan, pEepData->calTargetPower2GHT20,
AR5416_NUM_2G_20_TARGET_POWERS, &targetPowerHt20, 8, AH_FALSE);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_HT40(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
numCtlModes = N(ctlModesFor11g); /* All 2G CTL's */
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
ar5416GetTargetPowers(ah, chan, pEepData->calTargetPower2GHT40,
AR5416_NUM_2G_40_TARGET_POWERS, &targetPowerHt40, 8, AH_TRUE);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Get target powers for extension channels */
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
ar5416GetTargetPowersLeg(ah, chan, pEepData->calTargetPowerCck,
AR5416_NUM_2G_CCK_TARGET_POWERS, &targetPowerCckExt, 4, AH_TRUE);
ar5416GetTargetPowersLeg(ah, chan, pEepData->calTargetPower2G,
AR5416_NUM_2G_20_TARGET_POWERS, &targetPowerOfdmExt, 4, AH_TRUE);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
} else {
/* Setup for CTL modes */
numCtlModes = N(ctlModesFor11a) - SUB_NUM_CTL_MODES_AT_5G_40; /* CTL_11A, CTL_5GHT20 */
pCtlMode = ctlModesFor11a;
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
ar5416GetTargetPowersLeg(ah, chan, pEepData->calTargetPower5G,
virgin import of ath hal
2008-11-28 00:03:41 +00:00
AR5416_NUM_5G_20_TARGET_POWERS, &targetPowerOfdm, 4, AH_FALSE);
ar5416GetTargetPowers(ah, chan, pEepData->calTargetPower5GHT20,
AR5416_NUM_5G_20_TARGET_POWERS, &targetPowerHt20, 8, AH_FALSE);
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_HT40(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
numCtlModes = N(ctlModesFor11a); /* All 5G CTL's */
ar5416GetTargetPowers(ah, chan, pEepData->calTargetPower5GHT40,
AR5416_NUM_5G_40_TARGET_POWERS, &targetPowerHt40, 8, AH_TRUE);
ar5416GetTargetPowersLeg(ah, chan, pEepData->calTargetPower5G,
AR5416_NUM_5G_20_TARGET_POWERS, &targetPowerOfdmExt, 4, AH_TRUE);
}
}
/*
* For MIMO, need to apply regulatory caps individually across dynamically
* running modes: CCK, OFDM, HT20, HT40
*
* The outer loop walks through each possible applicable runtime mode.
* The inner loop walks through each ctlIndex entry in EEPROM.
* The ctl value is encoded as [7:4] == test group, [3:0] == test mode.
*
*/
for (ctlMode = 0; ctlMode < numCtlModes; ctlMode++) {
HAL_BOOL isHt40CtlMode = (pCtlMode[ctlMode] == CTL_5GHT40) ||
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
(pCtlMode[ctlMode] == CTL_2GHT40);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
if (isHt40CtlMode) {
freq = centers.ctl_center;
} else if (pCtlMode[ctlMode] & EXT_ADDITIVE) {
freq = centers.ext_center;
} else {
freq = centers.ctl_center;
}
/* walk through each CTL index stored in EEPROM */
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
for (i = 0; (i < AR5416_NUM_CTLS) && pEepData->ctlIndex[i]; i++) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
uint16_t twiceMinEdgePower;
/* compare test group from regulatory channel list with test mode from pCtlMode list */
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
if ((((cfgCtl & ~CTL_MODE_M) | (pCtlMode[ctlMode] & CTL_MODE_M)) == pEepData->ctlIndex[i]) ||
virgin import of ath hal
2008-11-28 00:03:41 +00:00
(((cfgCtl & ~CTL_MODE_M) | (pCtlMode[ctlMode] & CTL_MODE_M)) ==
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
((pEepData->ctlIndex[i] & CTL_MODE_M) | SD_NO_CTL))) {
rep = &(pEepData->ctlData[i]);
twiceMinEdgePower = ar5416GetMaxEdgePower(freq,
rep->ctlEdges[owl_get_ntxchains(AH5416(ah)->ah_tx_chainmask) - 1],
IEEE80211_IS_CHAN_2GHZ(chan));
virgin import of ath hal
2008-11-28 00:03:41 +00:00
if ((cfgCtl & ~CTL_MODE_M) == SD_NO_CTL) {
/* Find the minimum of all CTL edge powers that apply to this channel */
twiceMaxEdgePower = AH_MIN(twiceMaxEdgePower, twiceMinEdgePower);
} else {
/* specific */
twiceMaxEdgePower = twiceMinEdgePower;
break;
}
}
}
minCtlPower = (uint8_t)AH_MIN(twiceMaxEdgePower, scaledPower);
/* Apply ctl mode to correct target power set */
switch(pCtlMode[ctlMode]) {
case CTL_11B:
for (i = 0; i < N(targetPowerCck.tPow2x); i++) {
targetPowerCck.tPow2x[i] = (uint8_t)AH_MIN(targetPowerCck.tPow2x[i], minCtlPower);
}
break;
case CTL_11A:
case CTL_11G:
for (i = 0; i < N(targetPowerOfdm.tPow2x); i++) {
targetPowerOfdm.tPow2x[i] = (uint8_t)AH_MIN(targetPowerOfdm.tPow2x[i], minCtlPower);
}
break;
case CTL_5GHT20:
case CTL_2GHT20:
for (i = 0; i < N(targetPowerHt20.tPow2x); i++) {
targetPowerHt20.tPow2x[i] = (uint8_t)AH_MIN(targetPowerHt20.tPow2x[i], minCtlPower);
}
break;
case CTL_11B_EXT:
targetPowerCckExt.tPow2x[0] = (uint8_t)AH_MIN(targetPowerCckExt.tPow2x[0], minCtlPower);
break;
case CTL_11A_EXT:
case CTL_11G_EXT:
targetPowerOfdmExt.tPow2x[0] = (uint8_t)AH_MIN(targetPowerOfdmExt.tPow2x[0], minCtlPower);
break;
case CTL_5GHT40:
case CTL_2GHT40:
for (i = 0; i < N(targetPowerHt40.tPow2x); i++) {
targetPowerHt40.tPow2x[i] = (uint8_t)AH_MIN(targetPowerHt40.tPow2x[i], minCtlPower);
}
break;
default:
return AH_FALSE;
break;
}
} /* end ctl mode checking */
/* Set rates Array from collected data */
ratesArray[rate6mb] = ratesArray[rate9mb] = ratesArray[rate12mb] = ratesArray[rate18mb] = ratesArray[rate24mb] = targetPowerOfdm.tPow2x[0];
ratesArray[rate36mb] = targetPowerOfdm.tPow2x[1];
ratesArray[rate48mb] = targetPowerOfdm.tPow2x[2];
ratesArray[rate54mb] = targetPowerOfdm.tPow2x[3];
ratesArray[rateXr] = targetPowerOfdm.tPow2x[0];
for (i = 0; i < N(targetPowerHt20.tPow2x); i++) {
ratesArray[rateHt20_0 + i] = targetPowerHt20.tPow2x[i];
}
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_2GHZ(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
ratesArray[rate1l] = targetPowerCck.tPow2x[0];
ratesArray[rate2s] = ratesArray[rate2l] = targetPowerCck.tPow2x[1];
ratesArray[rate5_5s] = ratesArray[rate5_5l] = targetPowerCck.tPow2x[2];
ratesArray[rate11s] = ratesArray[rate11l] = targetPowerCck.tPow2x[3];
}
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_HT40(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
for (i = 0; i < N(targetPowerHt40.tPow2x); i++) {
ratesArray[rateHt40_0 + i] = targetPowerHt40.tPow2x[i];
}
ratesArray[rateDupOfdm] = targetPowerHt40.tPow2x[0];
ratesArray[rateDupCck] = targetPowerHt40.tPow2x[0];
ratesArray[rateExtOfdm] = targetPowerOfdmExt.tPow2x[0];
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_2GHZ(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
ratesArray[rateExtCck] = targetPowerCckExt.tPow2x[0];
}
}
return AH_TRUE;
#undef EXT_ADDITIVE
#undef CTL_11A_EXT
#undef CTL_11G_EXT
#undef CTL_11B_EXT
#undef SUB_NUM_CTL_MODES_AT_5G_40
#undef SUB_NUM_CTL_MODES_AT_2G_40
#undef N
}
/**************************************************************************
* fbin2freq
*
* Get channel value from binary representation held in eeprom
* RETURNS: the frequency in MHz
*/
static uint16_t
fbin2freq(uint8_t fbin, HAL_BOOL is2GHz)
{
/*
* Reserved value 0xFF provides an empty definition both as
* an fbin and as a frequency - do not convert
*/
if (fbin == AR5416_BCHAN_UNUSED) {
return fbin;
}
return (uint16_t)((is2GHz) ? (2300 + fbin) : (4800 + 5 * fbin));
}
/*
* ar5416GetMaxEdgePower
*
* Find the maximum conformance test limit for the given channel and CTL info
*/
static uint16_t
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
ar5416GetMaxEdgePower(uint16_t freq, CAL_CTL_EDGES *pRdEdgesPower, HAL_BOOL is2GHz)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
uint16_t twiceMaxEdgePower = AR5416_MAX_RATE_POWER;
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
int i;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Get the edge power */
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
for (i = 0; (i < AR5416_NUM_BAND_EDGES) && (pRdEdgesPower[i].bChannel != AR5416_BCHAN_UNUSED) ; i++) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/*
* If there's an exact channel match or an inband flag set
* on the lower channel use the given rdEdgePower
*/
if (freq == fbin2freq(pRdEdgesPower[i].bChannel, is2GHz)) {
twiceMaxEdgePower = MS(pRdEdgesPower[i].tPowerFlag, CAL_CTL_EDGES_POWER);
break;
} else if ((i > 0) && (freq < fbin2freq(pRdEdgesPower[i].bChannel, is2GHz))) {
if (fbin2freq(pRdEdgesPower[i - 1].bChannel, is2GHz) < freq && (pRdEdgesPower[i - 1].tPowerFlag & CAL_CTL_EDGES_FLAG) != 0) {
twiceMaxEdgePower = MS(pRdEdgesPower[i - 1].tPowerFlag, CAL_CTL_EDGES_POWER);
}
/* Leave loop - no more affecting edges possible in this monotonic increasing list */
break;
}
}
HALASSERT(twiceMaxEdgePower > 0);
return twiceMaxEdgePower;
}
/**************************************************************
* ar5416GetTargetPowers
*
* Return the rates of target power for the given target power table
* channel, and number of channels
*/
Bring back AR9285 support. This fixes most of the issues and should be pretty usable. MFC after: 1 month
2010-02-15 17:49:49 +00:00
void
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
ar5416GetTargetPowers(struct ath_hal *ah, const struct ieee80211_channel *chan,
virgin import of ath hal
2008-11-28 00:03:41 +00:00
CAL_TARGET_POWER_HT *powInfo, uint16_t numChannels,
CAL_TARGET_POWER_HT *pNewPower, uint16_t numRates,
HAL_BOOL isHt40Target)
{
uint16_t clo, chi;
int i;
int matchIndex = -1, lowIndex = -1;
uint16_t freq;
CHAN_CENTERS centers;
ar5416GetChannelCenters(ah, chan, &centers);
freq = isHt40Target ? centers.synth_center : centers.ctl_center;
/* Copy the target powers into the temp channel list */
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (freq <= fbin2freq(powInfo[0].bChannel, IEEE80211_IS_CHAN_2GHZ(chan))) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
matchIndex = 0;
} else {
for (i = 0; (i < numChannels) && (powInfo[i].bChannel != AR5416_BCHAN_UNUSED); i++) {
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (freq == fbin2freq(powInfo[i].bChannel, IEEE80211_IS_CHAN_2GHZ(chan))) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
matchIndex = i;
break;
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
} else if ((freq < fbin2freq(powInfo[i].bChannel, IEEE80211_IS_CHAN_2GHZ(chan))) &&
(freq > fbin2freq(powInfo[i - 1].bChannel, IEEE80211_IS_CHAN_2GHZ(chan))))
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
lowIndex = i - 1;
break;
}
}
if ((matchIndex == -1) && (lowIndex == -1)) {
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
HALASSERT(freq > fbin2freq(powInfo[i - 1].bChannel, IEEE80211_IS_CHAN_2GHZ(chan)));
virgin import of ath hal
2008-11-28 00:03:41 +00:00
matchIndex = i - 1;
}
}
if (matchIndex != -1) {
OS_MEMCPY(pNewPower, &powInfo[matchIndex], sizeof(*pNewPower));
} else {
HALASSERT(lowIndex != -1);
/*
* Get the lower and upper channels, target powers,
* and interpolate between them.
*/
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
clo = fbin2freq(powInfo[lowIndex].bChannel, IEEE80211_IS_CHAN_2GHZ(chan));
chi = fbin2freq(powInfo[lowIndex + 1].bChannel, IEEE80211_IS_CHAN_2GHZ(chan));
virgin import of ath hal
2008-11-28 00:03:41 +00:00
for (i = 0; i < numRates; i++) {
Move out some of the shared eeprom board value calculation routines into ah.c rather than duplicating them for the v14 (ar5416+) and v4k (ar9285) codebases. Further chipsets (eg the AR9287) have yet another EEPROM format which will use these routines to calculate things.
2011-03-13 05:54:05 +00:00
pNewPower->tPow2x[i] = (uint8_t)ath_ee_interpolate(freq, clo, chi,
virgin import of ath hal
2008-11-28 00:03:41 +00:00
powInfo[lowIndex].tPow2x[i], powInfo[lowIndex + 1].tPow2x[i]);
}
}
}
/**************************************************************
* ar5416GetTargetPowersLeg
*
* Return the four rates of target power for the given target power table
* channel, and number of channels
*/
Bring back AR9285 support. This fixes most of the issues and should be pretty usable. MFC after: 1 month
2010-02-15 17:49:49 +00:00
void
virgin import of ath hal
2008-11-28 00:03:41 +00:00
ar5416GetTargetPowersLeg(struct ath_hal *ah,
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
const struct ieee80211_channel *chan,
virgin import of ath hal
2008-11-28 00:03:41 +00:00
CAL_TARGET_POWER_LEG *powInfo, uint16_t numChannels,
CAL_TARGET_POWER_LEG *pNewPower, uint16_t numRates,
HAL_BOOL isExtTarget)
{
uint16_t clo, chi;
int i;
int matchIndex = -1, lowIndex = -1;
uint16_t freq;
CHAN_CENTERS centers;
ar5416GetChannelCenters(ah, chan, &centers);
freq = (isExtTarget) ? centers.ext_center :centers.ctl_center;
/* Copy the target powers into the temp channel list */
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (freq <= fbin2freq(powInfo[0].bChannel, IEEE80211_IS_CHAN_2GHZ(chan))) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
matchIndex = 0;
} else {
for (i = 0; (i < numChannels) && (powInfo[i].bChannel != AR5416_BCHAN_UNUSED); i++) {
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (freq == fbin2freq(powInfo[i].bChannel, IEEE80211_IS_CHAN_2GHZ(chan))) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
matchIndex = i;
break;
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
} else if ((freq < fbin2freq(powInfo[i].bChannel, IEEE80211_IS_CHAN_2GHZ(chan))) &&
(freq > fbin2freq(powInfo[i - 1].bChannel, IEEE80211_IS_CHAN_2GHZ(chan))))
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
lowIndex = i - 1;
break;
}
}
if ((matchIndex == -1) && (lowIndex == -1)) {
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
HALASSERT(freq > fbin2freq(powInfo[i - 1].bChannel, IEEE80211_IS_CHAN_2GHZ(chan)));
virgin import of ath hal
2008-11-28 00:03:41 +00:00
matchIndex = i - 1;
}
}
if (matchIndex != -1) {
OS_MEMCPY(pNewPower, &powInfo[matchIndex], sizeof(*pNewPower));
} else {
HALASSERT(lowIndex != -1);
/*
* Get the lower and upper channels, target powers,
* and interpolate between them.
*/
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
clo = fbin2freq(powInfo[lowIndex].bChannel, IEEE80211_IS_CHAN_2GHZ(chan));
chi = fbin2freq(powInfo[lowIndex + 1].bChannel, IEEE80211_IS_CHAN_2GHZ(chan));
virgin import of ath hal
2008-11-28 00:03:41 +00:00
for (i = 0; i < numRates; i++) {
Move out some of the shared eeprom board value calculation routines into ah.c rather than duplicating them for the v14 (ar5416+) and v4k (ar9285) codebases. Further chipsets (eg the AR9287) have yet another EEPROM format which will use these routines to calculate things.
2011-03-13 05:54:05 +00:00
pNewPower->tPow2x[i] = (uint8_t)ath_ee_interpolate(freq, clo, chi,
virgin import of ath hal
2008-11-28 00:03:41 +00:00
powInfo[lowIndex].tPow2x[i], powInfo[lowIndex + 1].tPow2x[i]);
}
}
}
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
/*
* Set the gain boundaries for the given radio chain.
*
* The gain boundaries tell the hardware at what point in the
* PDADC array to "switch over" from one PD gain setting
* to another. There's also a gain overlap between two
* PDADC array gain curves where there's valid PD values
* for 2 gain settings.
*
* The hardware uses the gain overlap and gain boundaries
* to determine which gain curve to use for the given
* target TX power.
*/
void
* Add in some board settings debugging to log what's being written to the TX closed-loop power control registers. * Modify a couple of functions to take the register chain number, rather than the regChainOffset value. This allows for the register chain to be logged.
2011-03-13 05:30:14 +00:00
ar5416SetGainBoundariesClosedLoop(struct ath_hal *ah, int i,
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
uint16_t pdGainOverlap_t2, uint16_t gainBoundaries[])
{
* Add in some board settings debugging to log what's being written to the TX closed-loop power control registers. * Modify a couple of functions to take the register chain number, rather than the regChainOffset value. This allows for the register chain to be logged.
2011-03-13 05:30:14 +00:00
int regChainOffset;
regChainOffset = ar5416GetRegChainOffset(ah, i);
HALDEBUG(ah, HAL_DEBUG_EEPROM, "%s: chain %d: gainOverlap_t2: %d,"
" gainBoundaries: %d, %d, %d, %d\n", __func__, i, pdGainOverlap_t2,
gainBoundaries[0], gainBoundaries[1], gainBoundaries[2],
gainBoundaries[3]);
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
OS_REG_WRITE(ah, AR_PHY_TPCRG5 + regChainOffset,
SM(pdGainOverlap_t2, AR_PHY_TPCRG5_PD_GAIN_OVERLAP) |
SM(gainBoundaries[0], AR_PHY_TPCRG5_PD_GAIN_BOUNDARY_1) |
SM(gainBoundaries[1], AR_PHY_TPCRG5_PD_GAIN_BOUNDARY_2) |
SM(gainBoundaries[2], AR_PHY_TPCRG5_PD_GAIN_BOUNDARY_3) |
SM(gainBoundaries[3], AR_PHY_TPCRG5_PD_GAIN_BOUNDARY_4));
}
/*
* Get the gain values and the number of gain levels given
* in xpdMask.
*
* The EEPROM xpdMask determines which power detector gain
* levels were used during calibration. Each of these mask
* bits maps to a fixed gain level in hardware.
*/
uint16_t
ar5416GetXpdGainValues(struct ath_hal *ah, uint16_t xpdMask,
uint16_t xpdGainValues[])
{
int i;
uint16_t numXpdGain = 0;
for (i = 1; i <= AR5416_PD_GAINS_IN_MASK; i++) {
if ((xpdMask >> (AR5416_PD_GAINS_IN_MASK - i)) & 1) {
if (numXpdGain >= AR5416_NUM_PD_GAINS) {
HALASSERT(0);
break;
}
xpdGainValues[numXpdGain] = (uint16_t)(AR5416_PD_GAINS_IN_MASK - i);
numXpdGain++;
}
}
return numXpdGain;
}
/*
* Write the detector gain and biases.
*
* There are four power detector gain levels. The xpdMask in the EEPROM
* determines which power detector gain levels have TX power calibration
* data associated with them. This function writes the number of
* PD gain levels and their values into the hardware.
*
* This is valid for all TX chains - the calibration data itself however
* will likely differ per-chain.
*/
void
ar5416WriteDetectorGainBiases(struct ath_hal *ah, uint16_t numXpdGain,
uint16_t xpdGainValues[])
{
* Add in some board settings debugging to log what's being written to the TX closed-loop power control registers. * Modify a couple of functions to take the register chain number, rather than the regChainOffset value. This allows for the register chain to be logged.
2011-03-13 05:30:14 +00:00
HALDEBUG(ah, HAL_DEBUG_EEPROM, "%s: numXpdGain: %d,"
" xpdGainValues: %d, %d, %d\n", __func__, numXpdGain,
xpdGainValues[0], xpdGainValues[1], xpdGainValues[2]);
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
OS_REG_WRITE(ah, AR_PHY_TPCRG1, (OS_REG_READ(ah, AR_PHY_TPCRG1) &
~(AR_PHY_TPCRG1_NUM_PD_GAIN | AR_PHY_TPCRG1_PD_GAIN_1 |
AR_PHY_TPCRG1_PD_GAIN_2 | AR_PHY_TPCRG1_PD_GAIN_3)) |
SM(numXpdGain - 1, AR_PHY_TPCRG1_NUM_PD_GAIN) |
SM(xpdGainValues[0], AR_PHY_TPCRG1_PD_GAIN_1 ) |
SM(xpdGainValues[1], AR_PHY_TPCRG1_PD_GAIN_2) |
SM(xpdGainValues[2], AR_PHY_TPCRG1_PD_GAIN_3));
}
/*
* Add in some board settings debugging to log what's being written to the TX closed-loop power control registers. * Modify a couple of functions to take the register chain number, rather than the regChainOffset value. This allows for the register chain to be logged.
2011-03-13 05:30:14 +00:00
* Write the PDADC array to the given radio chain i.
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
*
* The 32 PDADC registers are written without any care about
* their contents - so if various chips treat values as "special",
* this routine will not care.
*/
void
* Add in some board settings debugging to log what's being written to the TX closed-loop power control registers. * Modify a couple of functions to take the register chain number, rather than the regChainOffset value. This allows for the register chain to be logged.
2011-03-13 05:30:14 +00:00
ar5416WritePdadcValues(struct ath_hal *ah, int i, uint8_t pdadcValues[])
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
{
* Add in some board settings debugging to log what's being written to the TX closed-loop power control registers. * Modify a couple of functions to take the register chain number, rather than the regChainOffset value. This allows for the register chain to be logged.
2011-03-13 05:30:14 +00:00
int regOffset, regChainOffset;
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
int j;
int reg32;
* Add in some board settings debugging to log what's being written to the TX closed-loop power control registers. * Modify a couple of functions to take the register chain number, rather than the regChainOffset value. This allows for the register chain to be logged.
2011-03-13 05:30:14 +00:00
regChainOffset = ar5416GetRegChainOffset(ah, i);
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
regOffset = AR_PHY_BASE + (672 << 2) + regChainOffset;
for (j = 0; j < 32; j++) {
reg32 = ((pdadcValues[4*j + 0] & 0xFF) << 0) |
((pdadcValues[4*j + 1] & 0xFF) << 8) |
((pdadcValues[4*j + 2] & 0xFF) << 16) |
((pdadcValues[4*j + 3] & 0xFF) << 24) ;
OS_REG_WRITE(ah, regOffset, reg32);
* Add in some board settings debugging to log what's being written to the TX closed-loop power control registers. * Modify a couple of functions to take the register chain number, rather than the regChainOffset value. This allows for the register chain to be logged.
2011-03-13 05:30:14 +00:00
HALDEBUG(ah, HAL_DEBUG_EEPROM, "PDADC: Chain %d |"
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
" PDADC %3d Value %3d | PDADC %3d Value %3d | PDADC %3d"
* Add in some board settings debugging to log what's being written to the TX closed-loop power control registers. * Modify a couple of functions to take the register chain number, rather than the regChainOffset value. This allows for the register chain to be logged.
2011-03-13 05:30:14 +00:00
" Value %3d | PDADC %3d Value %3d |\n",
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
i,
4*j, pdadcValues[4*j],
4*j+1, pdadcValues[4*j + 1],
4*j+2, pdadcValues[4*j + 2],
4*j+3, pdadcValues[4*j + 3]);
regOffset += 4;
}
}
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/**************************************************************
* ar5416SetPowerCalTable
*
* Pull the PDADC piers from cal data and interpolate them across the given
* points as well as from the nearest pier(s) to get a power detector
* linear voltage to power level table.
*/
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
HAL_BOOL
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
ar5416SetPowerCalTable(struct ath_hal *ah, struct ar5416eeprom *pEepData,
const struct ieee80211_channel *chan, int16_t *pTxPowerIndexOffset)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
CAL_DATA_PER_FREQ *pRawDataset;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
uint8_t *pCalBChans = AH_NULL;
uint16_t pdGainOverlap_t2;
static uint8_t pdadcValues[AR5416_NUM_PDADC_VALUES];
uint16_t gainBoundaries[AR5416_PD_GAINS_IN_MASK];
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
uint16_t numPiers, i;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
int16_t tMinCalPower;
uint16_t numXpdGain, xpdMask;
uint16_t xpdGainValues[AR5416_NUM_PD_GAINS];
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
uint32_t regChainOffset;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
kill more portability functions that are no longer useful
2009-05-08 00:23:00 +00:00
OS_MEMZERO(xpdGainValues, sizeof(xpdGainValues));
virgin import of ath hal
2008-11-28 00:03:41 +00:00
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
xpdMask = pEepData->modalHeader[IEEE80211_IS_CHAN_2GHZ(chan)].xpdGain;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
if (IS_EEP_MINOR_V2(ah)) {
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
pdGainOverlap_t2 = pEepData->modalHeader[IEEE80211_IS_CHAN_2GHZ(chan)].pdGainOverlap;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
} else {
pdGainOverlap_t2 = (uint16_t)(MS(OS_REG_READ(ah, AR_PHY_TPCRG5), AR_PHY_TPCRG5_PD_GAIN_OVERLAP));
}
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_2GHZ(chan)) {
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
pCalBChans = pEepData->calFreqPier2G;
numPiers = AR5416_NUM_2G_CAL_PIERS;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
} else {
pCalBChans = pEepData->calFreqPier5G;
numPiers = AR5416_NUM_5G_CAL_PIERS;
}
/* Calculate the value of xpdgains from the xpdGain Mask */
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
numXpdGain = ar5416GetXpdGainValues(ah, xpdMask, xpdGainValues);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Write the detector gain biases and their number */
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
ar5416WriteDetectorGainBiases(ah, numXpdGain, xpdGainValues);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
for (i = 0; i < AR5416_MAX_CHAINS; i++) {
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
regChainOffset = ar5416GetRegChainOffset(ah, i);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
if (pEepData->baseEepHeader.txMask & (1 << i)) {
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_2GHZ(chan)) {
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
pRawDataset = pEepData->calPierData2G[i];
virgin import of ath hal
2008-11-28 00:03:41 +00:00
} else {
pRawDataset = pEepData->calPierData5G[i];
}
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
/* Fetch the gain boundaries and the PDADC values */
ar5416GetGainBoundariesAndPdadcs(ah, chan, pRawDataset,
virgin import of ath hal
2008-11-28 00:03:41 +00:00
pCalBChans, numPiers,
pdGainOverlap_t2,
&tMinCalPower, gainBoundaries,
pdadcValues, numXpdGain);
if ((i == 0) || AR_SREV_OWL_20_OR_LATER(ah)) {
* Add in some board settings debugging to log what's being written to the TX closed-loop power control registers. * Modify a couple of functions to take the register chain number, rather than the regChainOffset value. This allows for the register chain to be logged.
2011-03-13 05:30:14 +00:00
ar5416SetGainBoundariesClosedLoop(ah, i, pdGainOverlap_t2,
gainBoundaries);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
/* Write the power values into the baseband power table */
* Add in some board settings debugging to log what's being written to the TX closed-loop power control registers. * Modify a couple of functions to take the register chain number, rather than the regChainOffset value. This allows for the register chain to be logged.
2011-03-13 05:30:14 +00:00
ar5416WritePdadcValues(ah, i, pdadcValues);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
}
*pTxPowerIndexOffset = 0;
return AH_TRUE;
}
/**************************************************************
* ar5416GetGainBoundariesAndPdadcs
*
* Uses the data points read from EEPROM to reconstruct the pdadc power table
* Called by ar5416SetPowerCalTable only.
*/
Implement open-loop TX power control (OLC) for Merlin (AR9280) and generally tidy up the TX power programming code. Enforce that the TX power offset for Merlin is -5 dBm, rather than any other value programmable in the EEPROM. This requires some further code to be ported over from ath9k, so until that is done and tested, fail to attach NICs whose TX power offset isn't -5 dBm. This improves both legacy and HT transmission on my merlin board. It allows for stable MCS TX up to MCS15. Specifics: * Refactor out a bunch of the TX power calibration code - setting/obtaining the power detector / gain boundaries, programming the PDADC * Take the -5 dBm TX power offset into account on Merlin - "0" in the per-rate TX power register means -5 dBm, not 0 dBm * When doing OLC * Enforce min (0) and max (AR5416_MAX_RATE_POWER) when fiddling with the TX power, to avoid the TX power values from wrapping when low. * Implement the 1 dBm cck power offset when doing OLC * Implement temperature compensation for 2.4ghz mode when doing OLC * Implement an AR9280 specific TX power calibration routine which includes the OLC twiddles, leaving the earlier chipset path (AR5416, AR9160) alone Whilst here, use these refactored routines for the AR9285 TX power calibration/programming code and enforce correct overflow/underflow handling when fiddling with TX power values. Obtained from: linux ath9k
2011-03-08 06:59:59 +00:00
void
virgin import of ath hal
2008-11-28 00:03:41 +00:00
ar5416GetGainBoundariesAndPdadcs(struct ath_hal *ah,
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
const struct ieee80211_channel *chan,
CAL_DATA_PER_FREQ *pRawDataSet,
virgin import of ath hal
2008-11-28 00:03:41 +00:00
uint8_t * bChans, uint16_t availPiers,
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
uint16_t tPdGainOverlap, int16_t *pMinCalPower, uint16_t * pPdGainBoundaries,
virgin import of ath hal
2008-11-28 00:03:41 +00:00
uint8_t * pPDADCValues, uint16_t numXpdGains)
{
int i, j, k;
int16_t ss; /* potentially -ve index for taking care of pdGainOverlap */
uint16_t idxL, idxR, numPiers; /* Pier indexes */
/* filled out Vpd table for all pdGains (chanL) */
static uint8_t vpdTableL[AR5416_NUM_PD_GAINS][AR5416_MAX_PWR_RANGE_IN_HALF_DB];
/* filled out Vpd table for all pdGains (chanR) */
static uint8_t vpdTableR[AR5416_NUM_PD_GAINS][AR5416_MAX_PWR_RANGE_IN_HALF_DB];
/* filled out Vpd table for all pdGains (interpolated) */
static uint8_t vpdTableI[AR5416_NUM_PD_GAINS][AR5416_MAX_PWR_RANGE_IN_HALF_DB];
uint8_t *pVpdL, *pVpdR, *pPwrL, *pPwrR;
uint8_t minPwrT4[AR5416_NUM_PD_GAINS];
uint8_t maxPwrT4[AR5416_NUM_PD_GAINS];
int16_t vpdStep;
int16_t tmpVal;
uint16_t sizeCurrVpdTable, maxIndex, tgtIndex;
HAL_BOOL match;
int16_t minDelta = 0;
CHAN_CENTERS centers;
ar5416GetChannelCenters(ah, chan, &centers);
/* Trim numPiers for the number of populated channel Piers */
for (numPiers = 0; numPiers < availPiers; numPiers++) {
if (bChans[numPiers] == AR5416_BCHAN_UNUSED) {
break;
}
}
/* Find pier indexes around the current channel */
Move out some of the shared eeprom board value calculation routines into ah.c rather than duplicating them for the v14 (ar5416+) and v4k (ar9285) codebases. Further chipsets (eg the AR9287) have yet another EEPROM format which will use these routines to calculate things.
2011-03-13 05:54:05 +00:00
match = ath_ee_getLowerUpperIndex((uint8_t)FREQ2FBIN(centers.synth_center,
IEEE80211_IS_CHAN_2GHZ(chan)), bChans, numPiers, &idxL, &idxR);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
if (match) {
/* Directly fill both vpd tables from the matching index */
for (i = 0; i < numXpdGains; i++) {
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
minPwrT4[i] = pRawDataSet[idxL].pwrPdg[i][0];
maxPwrT4[i] = pRawDataSet[idxL].pwrPdg[i][4];
Move out some of the shared eeprom board value calculation routines into ah.c rather than duplicating them for the v14 (ar5416+) and v4k (ar9285) codebases. Further chipsets (eg the AR9287) have yet another EEPROM format which will use these routines to calculate things.
2011-03-13 05:54:05 +00:00
ath_ee_FillVpdTable(minPwrT4[i], maxPwrT4[i], pRawDataSet[idxL].pwrPdg[i],
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
pRawDataSet[idxL].vpdPdg[i], AR5416_PD_GAIN_ICEPTS, vpdTableI[i]);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
} else {
for (i = 0; i < numXpdGains; i++) {
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
pVpdL = pRawDataSet[idxL].vpdPdg[i];
pPwrL = pRawDataSet[idxL].pwrPdg[i];
pVpdR = pRawDataSet[idxR].vpdPdg[i];
pPwrR = pRawDataSet[idxR].pwrPdg[i];
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Start Vpd interpolation from the max of the minimum powers */
minPwrT4[i] = AH_MAX(pPwrL[0], pPwrR[0]);
/* End Vpd interpolation from the min of the max powers */
maxPwrT4[i] = AH_MIN(pPwrL[AR5416_PD_GAIN_ICEPTS - 1], pPwrR[AR5416_PD_GAIN_ICEPTS - 1]);
HALASSERT(maxPwrT4[i] > minPwrT4[i]);
/* Fill pier Vpds */
Move out some of the shared eeprom board value calculation routines into ah.c rather than duplicating them for the v14 (ar5416+) and v4k (ar9285) codebases. Further chipsets (eg the AR9287) have yet another EEPROM format which will use these routines to calculate things.
2011-03-13 05:54:05 +00:00
ath_ee_FillVpdTable(minPwrT4[i], maxPwrT4[i], pPwrL, pVpdL, AR5416_PD_GAIN_ICEPTS, vpdTableL[i]);
ath_ee_FillVpdTable(minPwrT4[i], maxPwrT4[i], pPwrR, pVpdR, AR5416_PD_GAIN_ICEPTS, vpdTableR[i]);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Interpolate the final vpd */
for (j = 0; j <= (maxPwrT4[i] - minPwrT4[i]) / 2; j++) {
Move out some of the shared eeprom board value calculation routines into ah.c rather than duplicating them for the v14 (ar5416+) and v4k (ar9285) codebases. Further chipsets (eg the AR9287) have yet another EEPROM format which will use these routines to calculate things.
2011-03-13 05:54:05 +00:00
vpdTableI[i][j] = (uint8_t)(ath_ee_interpolate((uint16_t)FREQ2FBIN(centers.synth_center,
IEEE80211_IS_CHAN_2GHZ(chan)),
virgin import of ath hal
2008-11-28 00:03:41 +00:00
bChans[idxL], bChans[idxR], vpdTableL[i][j], vpdTableR[i][j]));
}
}
}
*pMinCalPower = (int16_t)(minPwrT4[0] / 2);
k = 0; /* index for the final table */
for (i = 0; i < numXpdGains; i++) {
if (i == (numXpdGains - 1)) {
pPdGainBoundaries[i] = (uint16_t)(maxPwrT4[i] / 2);
} else {
pPdGainBoundaries[i] = (uint16_t)((maxPwrT4[i] + minPwrT4[i+1]) / 4);
}
pPdGainBoundaries[i] = (uint16_t)AH_MIN(AR5416_MAX_RATE_POWER, pPdGainBoundaries[i]);
Update to later code from my repository: o many bug fixes o add new periodic calibration api o break up 5416 periodic calibration code in preparation for 928x o move get noise floor to rf backends o 5416-specific ani (still disabled) o modularize 5210 eeprom format a la other eeprom formats o start cleaning up regdomain code o prepare for proper 1/2 and 1/4 width channel support o bring back 900MHz card support o clean up 5212 rf version handling o add 1/2 and 1/4 width channel support for 5212 parts o split 5212 rfgain handling out o improve ani debugging o add AH_USE_INIPDGAIN compile option o purge a bunch of dead 5212 state o add 1/2 and 1/4 rate modes o remove HAL_CAP_CHAN_HALFRATE and HAL_CAP_CHAN_QUARTERRATE; the same info can now be deduced from the set of supported modes
2008-11-28 00:48:05 +00:00
/* NB: only applies to owl 1.0 */
virgin import of ath hal
2008-11-28 00:03:41 +00:00
if ((i == 0) && !AR_SREV_OWL_20_OR_LATER(ah) ) {
/*
* fix the gain delta, but get a delta that can be applied to min to
* keep the upper power values accurate, don't think max needs to
* be adjusted because should not be at that area of the table?
*/
minDelta = pPdGainBoundaries[0] - 23;
pPdGainBoundaries[0] = 23;
}
else {
minDelta = 0;
}
/* Find starting index for this pdGain */
if (i == 0) {
Merlin fix - first pdadc gain index is 0 - minpwr/2 . Obtained from: Linux ath9k
2011-03-10 06:06:26 +00:00
if (AR_SREV_MERLIN_20_OR_LATER(ah))
ss = (int16_t)(0 - (minPwrT4[i] / 2));
else
ss = 0; /* for the first pdGain, start from index 0 */
virgin import of ath hal
2008-11-28 00:03:41 +00:00
} else {
/* need overlap entries extrapolated below. */
ss = (int16_t)((pPdGainBoundaries[i-1] - (minPwrT4[i] / 2)) - tPdGainOverlap + 1 + minDelta);
}
vpdStep = (int16_t)(vpdTableI[i][1] - vpdTableI[i][0]);
vpdStep = (int16_t)((vpdStep < 1) ? 1 : vpdStep);
/*
*-ve ss indicates need to extrapolate data below for this pdGain
*/
while ((ss < 0) && (k < (AR5416_NUM_PDADC_VALUES - 1))) {
tmpVal = (int16_t)(vpdTableI[i][0] + ss * vpdStep);
pPDADCValues[k++] = (uint8_t)((tmpVal < 0) ? 0 : tmpVal);
ss++;
}
Revert part of the 9285 support because it breaks the 9280 support. I'll try to do the 9285 support without interfering with any other chipset revisions support.
2010-02-14 16:26:32 +00:00
sizeCurrVpdTable = (uint8_t)((maxPwrT4[i] - minPwrT4[i]) / 2 +1);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
tgtIndex = (uint8_t)(pPdGainBoundaries[i] + tPdGainOverlap - (minPwrT4[i] / 2));
maxIndex = (tgtIndex < sizeCurrVpdTable) ? tgtIndex : sizeCurrVpdTable;
while ((ss < maxIndex) && (k < (AR5416_NUM_PDADC_VALUES - 1))) {
pPDADCValues[k++] = vpdTableI[i][ss++];
}
vpdStep = (int16_t)(vpdTableI[i][sizeCurrVpdTable - 1] - vpdTableI[i][sizeCurrVpdTable - 2]);
vpdStep = (int16_t)((vpdStep < 1) ? 1 : vpdStep);
/*
* for last gain, pdGainBoundary == Pmax_t2, so will
* have to extrapolate
*/
Merge in a fix for the power/(gain?) calculation. Apply it to both the 5416/9160 and 9285 code paths. Obtained from: OpenWRT r22123, 522-ath9k_pwrcal_fix.patch
2010-08-14 15:29:21 +00:00
if (tgtIndex >= maxIndex) { /* need to extrapolate above */
virgin import of ath hal
2008-11-28 00:03:41 +00:00
while ((ss <= tgtIndex) && (k < (AR5416_NUM_PDADC_VALUES - 1))) {
tmpVal = (int16_t)((vpdTableI[i][sizeCurrVpdTable - 1] +
(ss - maxIndex +1) * vpdStep));
pPDADCValues[k++] = (uint8_t)((tmpVal > 255) ? 255 : tmpVal);
ss++;
}
} /* extrapolated above */
} /* for all pdGainUsed */
/* Fill out pdGainBoundaries - only up to 2 allowed here, but hardware allows up to 4 */
while (i < AR5416_PD_GAINS_IN_MASK) {
pPdGainBoundaries[i] = pPdGainBoundaries[i-1];
i++;
}
while (k < AR5416_NUM_PDADC_VALUES) {
pPDADCValues[k] = pPDADCValues[k-1];
k++;
}
return;
}
Enable the 11n PHY by default whether or not 11n is configured. The linux ath9k driver and (from what I've been told) the atheros reference driver does this; it then leaves discarding 11n frames to the 802.11 layer. Whilst I'm here, merge in a fix from ath9k which maintains a turbo register setting when enabling the 11n register; and remove an un-needed (duplicate) flag setting.
2011-01-23 14:49:50 +00:00
/*
* The linux ath9k driver and (from what I've been told) the reference
* Atheros driver enables the 11n PHY by default whether or not it's
* configured.
*/
virgin import of ath hal
2008-11-28 00:03:41 +00:00
static void
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
ar5416Set11nRegs(struct ath_hal *ah, const struct ieee80211_channel *chan)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
uint32_t phymode;
Enable the 11n PHY by default whether or not 11n is configured. The linux ath9k driver and (from what I've been told) the atheros reference driver does this; it then leaves discarding 11n frames to the 802.11 layer. Whilst I'm here, merge in a fix from ath9k which maintains a turbo register setting when enabling the 11n register; and remove an un-needed (duplicate) flag setting.
2011-01-23 14:49:50 +00:00
uint32_t enableDacFifo = 0;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
HAL_HT_MACMODE macmode; /* MAC - 20/40 mode */
Enable the 11n PHY by default whether or not 11n is configured. The linux ath9k driver and (from what I've been told) the atheros reference driver does this; it then leaves discarding 11n frames to the 802.11 layer. Whilst I'm here, merge in a fix from ath9k which maintains a turbo register setting when enabling the 11n register; and remove an un-needed (duplicate) flag setting.
2011-01-23 14:49:50 +00:00
if (AR_SREV_KITE_10_OR_LATER(ah))
enableDacFifo = (OS_REG_READ(ah, AR_PHY_TURBO) & AR_PHY_FC_ENABLE_DAC_FIFO);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Enable 11n HT, 20 MHz */
phymode = AR_PHY_FC_HT_EN | AR_PHY_FC_SHORT_GI_40
Enable the 11n PHY by default whether or not 11n is configured. The linux ath9k driver and (from what I've been told) the atheros reference driver does this; it then leaves discarding 11n frames to the 802.11 layer. Whilst I'm here, merge in a fix from ath9k which maintains a turbo register setting when enabling the 11n register; and remove an un-needed (duplicate) flag setting.
2011-01-23 14:49:50 +00:00
| AR_PHY_FC_SINGLE_HT_LTF1 | AR_PHY_FC_WALSH | enableDacFifo;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Configure baseband for dynamic 20/40 operation */
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_HT40(chan)) {
Enable the 11n PHY by default whether or not 11n is configured. The linux ath9k driver and (from what I've been told) the atheros reference driver does this; it then leaves discarding 11n frames to the 802.11 layer. Whilst I'm here, merge in a fix from ath9k which maintains a turbo register setting when enabling the 11n register; and remove an un-needed (duplicate) flag setting.
2011-01-23 14:49:50 +00:00
phymode |= AR_PHY_FC_DYN2040_EN;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/* Configure control (primary) channel at +-10MHz */
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_HT40U(chan))
virgin import of ath hal
2008-11-28 00:03:41 +00:00
phymode |= AR_PHY_FC_DYN2040_PRI_CH;
#if 0
/* Configure 20/25 spacing */
if (ht->ht_extprotspacing == HAL_HT_EXTPROTSPACING_25)
phymode |= AR_PHY_FC_DYN2040_EXT_CH;
#endif
macmode = HAL_HT_MACMODE_2040;
} else
macmode = HAL_HT_MACMODE_20;
OS_REG_WRITE(ah, AR_PHY_TURBO, phymode);
/* Configure MAC for 20/40 operation */
ar5416Set11nMac2040(ah, macmode);
/* global transmit timeout (25 TUs default)*/
/* XXX - put this elsewhere??? */
OS_REG_WRITE(ah, AR_GTXTO, 25 << AR_GTXTO_TIMEOUT_LIMIT_S) ;
/* carrier sense timeout */
OS_REG_SET_BIT(ah, AR_GTTM, AR_GTTM_CST_USEC);
bring this in line with what ath9k does.
2011-02-14 21:35:11 +00:00
OS_REG_WRITE(ah, AR_CST, 0xF << AR_CST_TIMEOUT_LIMIT_S);
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
void
ar5416GetChannelCenters(struct ath_hal *ah,
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
const struct ieee80211_channel *chan, CHAN_CENTERS *centers)
virgin import of ath hal
2008-11-28 00:03:41 +00:00
{
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
uint16_t freq = ath_hal_gethwchannel(ah, chan);
centers->ctl_center = freq;
centers->synth_center = freq;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
/*
* In 20/40 phy mode, the center frequency is
* "between" the control and extension channels.
*/
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
if (IEEE80211_IS_CHAN_HT40U(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
centers->synth_center += HT40_CHANNEL_CENTER_SHIFT;
centers->ext_center =
centers->synth_center + HT40_CHANNEL_CENTER_SHIFT;
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
} else if (IEEE80211_IS_CHAN_HT40D(chan)) {
virgin import of ath hal
2008-11-28 00:03:41 +00:00
centers->synth_center -= HT40_CHANNEL_CENTER_SHIFT;
centers->ext_center =
centers->synth_center - HT40_CHANNEL_CENTER_SHIFT;
} else {
Overhaul regulatory support: o remove HAL_CHANNEL; convert the hal to use net80211 channels; this mostly involves mechanical changes to variable names and channel attribute macros o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant with the net80211 channel available o change api for ath_hal_init_channels: no more reglass id's, no more outdoor indication (was a noop), anM contents o add ath_hal_getchannels to have the hal construct a channel list without altering runtime state; this is used to retrieve the calibration list for the device in ath_getradiocaps o add ath_hal_set_channels to take a channel list and regulatory data from above and construct internal state to match (maps frequencies for 900MHz cards, setup for CTL lookups, etc) o compact the private channel table: we keep one private channel per frequency instead of one per HAL_CHANNEL; this gives a big space savings and potentially improves ani and calibration by sharing state (to be seen; didn't see anything in testing); a new config option AH_MAXCHAN controls the table size (default to 96 which was chosen to be ~3x the largest expected size) o shrink ani state and change to mirror private channel table (one entry per frequency indexed by ic_devdata) o move ani state flags to private channel state o remove country codes; use net80211 definitions instead o remove GSM regulatory support; it's no longer needed now that we pass in channel lists from above o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A o simplify initial channel list construction based on the EEPROM contents; we preserve country code support for now but may want to just fallback to a WWR sku and dispatch the discovered country code up to user space so the channel list can be constructed using the master regdomain tables o defer to net80211 for max antenna gain o eliminate sorting of internal channel table; now that we use ic_devdata as an index, table lookups are O(1) o remove internal copy of the country code; the public one is sufficient o remove AH_SUPPORT_11D conditional compilation; we always support 11d o remove ath_hal_ispublicsafetysku; not needed any more o remove ath_hal_isgsmsku; no more GSM stuff o move Conformance Test Limit (CTL) state from private channel to a lookup using per-band pointers cached in the private state block o remove regulatory class id support; was unused and belongs in net80211 o fix channel list construction to set IEEE80211_CHAN_NOADHOC, IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are now set in the constructed net80211 channel o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one of the driver-private flag bits of the net80211 channel o move 900MHz frequency mapping into the hal; the mapped frequency is stored in the private channel and used throughout the hal (no more mapping in the driver and/or net80211) o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the calculation and available in the net80211 channel o change noise floor calibration logic to work with compacted private channel table setup; this may require revisiting as we no longer can distinguish channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used only to calculate status data we can live with it for now o change ah_getChipPowerLimits internal method to operate on a single channel instead of all channels in the private channel table o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency (always the same except for 900MHz channels) o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory problems o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now o change ath_hal_getwirelessmodes to really return wireless modes supported by the hardware (was previously applying regulatory constraints) o return channel interference status with IEEE80211_CHANSTATE_CWINT (should change to a callback so hal api's can take const pointers) o remove some #define's no longer needed with the inclusion of <net80211/_ieee80211.h> Sponsored by: Carlson Wireless
2009-01-28 18:00:22 +00:00
centers->ext_center = freq;
virgin import of ath hal
2008-11-28 00:03:41 +00:00
}
}
Port over ar5416OverrideIni() from ath9k ar5008_hw_override_ini(). * change the BB gating logic to explicitly define which chips are covered; the ath9k method isn't as clear. * don't disable the BB gating for now, the ar5416 initvals have it, and the ar9160 initval sets it to 0x0. Figure out why before re-enabling this. * migrate the Merlin (ar9280) applicable WAR from the Kite (ar9285) code (which won't get called for Merlin!) and stuff it in here.
2011-03-03 08:38:31 +00:00
/*
* Override the INI vals being programmed.
*/
static void
ar5416OverrideIni(struct ath_hal *ah, const struct ieee80211_channel *chan)
{
uint32_t val;
/*
* Set the RX_ABORT and RX_DIS and clear if off only after
* RXE is set for MAC. This prevents frames with corrupted
* descriptor status.
*/
OS_REG_SET_BIT(ah, AR_DIAG_SW, (AR_DIAG_RX_DIS | AR_DIAG_RX_ABORT));
if (AR_SREV_MERLIN_20_OR_LATER(ah)) {
val = OS_REG_READ(ah, AR_PCU_MISC_MODE2);
if (!AR_SREV_9271(ah))
val &= ~AR_PCU_MISC_MODE2_HWWAR1;
if (AR_SREV_9287_11_OR_LATER(ah))
val = val & (~AR_PCU_MISC_MODE2_HWWAR2);
OS_REG_WRITE(ah, AR_PCU_MISC_MODE2, val);
}
/*
* The AR5416 initvals have this already set to 0x11; AR9160 has
* the register set to 0x0. Figure out whether AR9100/AR9160 needs
* this before moving forward with it.
*/
#if 0
/* Disable BB clock gating for AR5416v2, AR9100, AR9160 */
if (AR_SREV_OWL_20_OR_LATER(ah) || AR_SREV_9100(ah) || AR_SREV_SOWL(ah)) {
/*
* Disable BB clock gating
* Necessary to avoid issues on AR5416 2.0
*/
OS_REG_WRITE(ah, 0x9800 + (651 << 2), 0x11);
}
#endif
/*
* Disable RIFS search on some chips to avoid baseband
* hang issues.
*/
if (AR_SREV_9100(ah) || AR_SREV_SOWL(ah)) {
val = OS_REG_READ(ah, AR_PHY_HEAVY_CLIP_FACTOR_RIFS);
val &= ~AR_PHY_RIFS_INIT_DELAY;
OS_REG_WRITE(ah, AR_PHY_HEAVY_CLIP_FACTOR_RIFS, val);
}
}
Reference in New Issue Copy Permalink