d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
81484cdb07
freebsd-dev/sys/dev/ath/if_ath.c

5614 lines
159 KiB
C
Raw Normal View History

Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/*-
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
* Copyright (c) 2002-2009 Sam Leffler, Errno Consulting
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer,
* without modification.
* 2. Redistributions in binary form must reproduce at minimum a disclaimer
* similar to the "NO WARRANTY" disclaimer below ("Disclaimer") and any
* redistribution must be conditioned upon including a substantially
* similar Disclaimer requirement for further binary redistribution.
*
* NO WARRANTY
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF NONINFRINGEMENT, MERCHANTIBILITY
* AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
* THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY,
* OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER
* IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
* THE POSSIBILITY OF SUCH DAMAGES.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
/*
* Driver for the Atheros Wireless LAN controller.
acknowledge the contribution of Atsushi Onoe
2003-06-30 04:51:11 +00:00
*
* This software is derived from work of Atsushi Onoe; his contribution
* is greatly appreciated.
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
*/
#include "opt_inet.h"
o add opt_ath.h enable tweaking various config parameters for the driver without modifying the source code o default debug msgs and diag support to off MFC after: 3 days
2006-04-03 18:14:02 +00:00
#include "opt_ath.h"
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#include "opt_wlan.h"
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
#include <sys/param.h>
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
#include <sys/systm.h>
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
#include <sys/sysctl.h>
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
#include <sys/mbuf.h>
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
#include <sys/malloc.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/kernel.h>
#include <sys/socket.h>
#include <sys/sockio.h>
#include <sys/errno.h>
#include <sys/callout.h>
#include <sys/bus.h>
#include <sys/endian.h>
use a private task queue thread MFC after: 2 weeks
2006-02-09 21:48:51 +00:00
#include <sys/kthread.h>
#include <sys/taskqueue.h>
add SIOCZATHSTATS ioctl to zero driver statistics
2009-02-13 05:38:03 +00:00
#include <sys/priv.h>
Break out the ath PCI logic into a separate device/module. Introduce the AHB glue for Atheros embedded systems. Right now it's hard-coded for the AR9130 chip whose support isn't yet in this HAL; it'll be added in a subsequent commit. Kernel configuration files now need both 'ath' and 'ath_pci' devices; both modules need to be loaded for the ath device to work.
2011-03-31 08:07:13 +00:00
#include <sys/module.h>
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
#include <machine/bus.h>
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
#include <net/if.h>
#include <net/if_dl.h>
#include <net/if_media.h>
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
#include <net/if_types.h>
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
#include <net/if_arp.h>
#include <net/ethernet.h>
#include <net/if_llc.h>
#include <net80211/ieee80211_var.h>
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
#include <net80211/ieee80211_regdomain.h>
Hoist 802.11 encapsulation up into net80211: o call ieee80211_encap in ieee80211_start so frames passed down to drivers are already encapsulated o remove ieee80211_encap calls in drivers o fixup wi so it recreates the 802.3 head it requires from the 802.11 header contents o move fast-frame aggregation from ath to net80211 (conditional on IEEE80211_SUPPORT_SUPERG): - aggregation is now done in ieee80211_start; it is enabled when the packets/sec exceeds ieee80211_ffppsmin (net.wlan.ffppsmin) and frames are held on a staging queue according to ieee80211_ffagemax (net.wlan.ffagemax) to wait for a frame to combine with - drivers must call back to age/flush the staging queue (ath does this on tx done, at swba, and on rx according to the state of the tx queues and/or the contents of the staging queue) - remove fast-frame-related data structures from ath - add ieee80211_ff_node_init and ieee80211_ff_node_cleanup to handle per-node fast-frames state (we reuse 11n tx ampdu state) o change ieee80211_encap calling convention to include an explicit vap so frames coming through a WDS vap are recognized w/o setting M_WDS With these changes any device able to tx/rx 3Kbyte+ frames can use fast-frames. Reviewed by: thompsa, rpaulo, avatar, imp, sephe
2009-03-30 21:53:27 +00:00
#ifdef IEEE80211_SUPPORT_SUPERG
#include <net80211/ieee80211_superg.h>
#endif
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
#include <net80211/ieee80211_tdma.h>
#endif
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
#include <net/bpf.h>
#ifdef INET
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
#include <netinet/in.h>
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
#include <netinet/if_ether.h>
#endif
#include <dev/ath/if_athvar.h>
Switch to ath hal source code. Note this removes the ath_hal module; the ath module now brings in the hal support. Kernel config files are almost backwards compatible; supplying device ath_hal gives you the same chip support that the binary hal did but you must also include options AH_SUPPORT_AR5416 to enable the extended format descriptors used by 11n parts. It is now possible to control the chip support included in a build by specifying exactly which chips are to be supported in the config file; consult ath_hal(4) for information.
2008-12-01 16:53:01 +00:00
#include <dev/ath/ath_hal/ah_devid.h> /* XXX for softled */
Break out the diagnostic codes from ah_internal.h and place them in ah_diagcodes.h. Since we now have the source code, there's no reason to hide the diag codes from other areas. They live in the HAL as they form part of the HAL API and should still be treate as "potentially flexible; don't publish as a public API." But since they're already used as a public API (see follow-up commit), we may as well use them in place of magic constants.
2011-01-20 04:57:26 +00:00
#include <dev/ath/ath_hal/ah_diagcodes.h>
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Break out the debug macros from if_ath.c into if_ath_debug.[ch] . This is prep work for breaking out the TX path into a separate set of source files.
2011-01-29 05:08:21 +00:00
#include <dev/ath/if_ath_debug.h>
Migrate the TX path code out of if_ath and into a separate source file. There's two reasons for this: * the raw and non-raw TX path shares a lot of duplicate code which should be refactored; * the 11n-ready chip TX path needs a little reworking.
2011-01-29 11:35:23 +00:00
#include <dev/ath/if_ath_misc.h>
#include <dev/ath/if_ath_tx.h>
Migrate the sysctl related routines (statistics, debugging, etc) out of if_ath.c and into if_ath_sysctl.c .
2011-03-02 16:03:19 +00:00
#include <dev/ath/if_ath_sysctl.h>
Break the keycache management functions out into if_ath_keycache.c .
2011-03-02 17:19:54 +00:00
#include <dev/ath/if_ath_keycache.h>
Break out the debug macros from if_ath.c into if_ath_debug.[ch] . This is prep work for breaking out the TX path into a separate set of source files.
2011-01-29 05:08:21 +00:00
add tx99 hooks MFC after: 2 weeks
2006-02-09 21:28:11 +00:00
#ifdef ATH_TX99_DIAG
#include <dev/ath/ath_tx99/ath_tx99.h>
#endif
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/*
* ATH_BCBUF determines the number of vap's that can transmit
* beacons and also (currently) the number of vap's that can
* have unique mac addresses/bssid. When staggering beacons
* 4 is probably a good max as otherwise the beacons become
* very closely spaced and there is limited time for cab q traffic
* to go out. You can burst beacons instead but that is not good
* for stations in power save and at some point you really want
* another radio (and channel).
*
* The limit on the number of mac addresses is tied to our use of
* the U/L bit and tracking addresses in a byte; it would be
* worthwhile to allow more for applications like proxy sta.
*/
CTASSERT(ATH_BCBUF <= 8);
static struct ieee80211vap *ath_vap_create(struct ieee80211com *,
const char name[IFNAMSIZ], int unit, int opmode,
int flags, const uint8_t bssid[IEEE80211_ADDR_LEN],
const uint8_t mac[IEEE80211_ADDR_LEN]);
static void ath_vap_delete(struct ieee80211vap *);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void ath_init(void *);
Update with last year of work.
2004-12-08 17:34:36 +00:00
static void ath_stop_locked(struct ifnet *);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void ath_stop(struct ifnet *);
static void ath_start(struct ifnet *);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static int ath_reset_vap(struct ieee80211vap *, u_long);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static int ath_media_change(struct ifnet *);
replace if_watchdog w/ private callout; probably can merge this with the calibration work sometime in the future
2009-03-09 23:10:19 +00:00
static void ath_watchdog(void *);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static int ath_ioctl(struct ifnet *, u_long, caddr_t);
static void ath_fatal_proc(void *, int);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static void ath_bmiss_vap(struct ieee80211vap *);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void ath_bmiss_proc(void *, int);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static void ath_key_update_begin(struct ieee80211vap *);
static void ath_key_update_end(struct ieee80211vap *);
static void ath_update_mcast(struct ifnet *);
static void ath_update_promisc(struct ifnet *);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void ath_mode_init(struct ath_softc *);
Update with last year of work.
2004-12-08 17:34:36 +00:00
static void ath_setslottime(struct ath_softc *);
static void ath_updateslot(struct ifnet *);
setup the beacon xmit queue to not interrupt; we don't use them and they make the led's flash unnecessarily in adhoc mode
2005-01-18 19:10:17 +00:00
static int ath_beaconq_setup(struct ath_hal *);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static int ath_beacon_alloc(struct ath_softc *, struct ieee80211_node *);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static void ath_beacon_update(struct ieee80211vap *, int item);
Update with last year of work.
2004-12-08 17:34:36 +00:00
static void ath_beacon_setup(struct ath_softc *, struct ath_buf *);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void ath_beacon_proc(void *, int);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static struct ath_buf *ath_beacon_generate(struct ath_softc *,
struct ieee80211vap *);
Update with last year of work.
2004-12-08 17:34:36 +00:00
static void ath_bstuck_proc(void *, int);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static void ath_beacon_return(struct ath_softc *, struct ath_buf *);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void ath_beacon_free(struct ath_softc *);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static void ath_beacon_config(struct ath_softc *, struct ieee80211vap *);
Update with last year of work.
2004-12-08 17:34:36 +00:00
static void ath_descdma_cleanup(struct ath_softc *sc,
struct ath_descdma *, ath_bufhead *);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static int ath_desc_alloc(struct ath_softc *);
static void ath_desc_free(struct ath_softc *);
Change the calling convention for ic_node_alloc to deal with some longstanding issues: o pass the vap since it's now the "coin of the realm" and required to do things like set initial tx parameters in private node state for use prior to association o pass the mac address as cards that maintain outboard station tables require this to create an entry (e.g. in ibss mode) o remove the node table reference, we only have one node table and it's unlikely this will change so this is not needed to find the com structure
2008-06-07 18:38:02 +00:00
static struct ieee80211_node *ath_node_alloc(struct ieee80211vap *,
const uint8_t [IEEE80211_ADDR_LEN]);
Update with last year of work.
2004-12-08 17:34:36 +00:00
static void ath_node_free(struct ieee80211_node *);
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
static void ath_node_getsignal(const struct ieee80211_node *,
int8_t *, int8_t *);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static int ath_rxbuf_init(struct ath_softc *, struct ath_buf *);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static void ath_recv_mgmt(struct ieee80211_node *ni, struct mbuf *m,
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
int subtype, int rssi, int nf);
Update with last year of work.
2004-12-08 17:34:36 +00:00
static void ath_setdefantenna(struct ath_softc *, u_int);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void ath_rx_proc(void *, int);
Close race in handling mcast traffic when operating as an ap with stations in power save: add a new q where mcast frames are stashed and on beacon update (at DTIM) move frames from the mcast q to the cabq and start it. This ensures the cabq is only manipulated in one place. Sponsored by: Hobnob MFC after: 2 weeks
2006-06-26 03:10:45 +00:00
static void ath_txq_init(struct ath_softc *sc, struct ath_txq *, int);
Update with last year of work.
2004-12-08 17:34:36 +00:00
static struct ath_txq *ath_txq_setup(struct ath_softc*, int qtype, int subtype);
static int ath_tx_setup(struct ath_softc *, int, int);
static int ath_wme_update(struct ieee80211com *);
static void ath_tx_cleanupq(struct ath_softc *, struct ath_txq *);
static void ath_tx_cleanup(struct ath_softc *);
static void ath_tx_proc_q0(void *, int);
static void ath_tx_proc_q0123(void *, int);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void ath_tx_proc(void *, int);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static void ath_tx_draintxq(struct ath_softc *, struct ath_txq *);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static int ath_chan_set(struct ath_softc *, struct ieee80211_channel *);
static void ath_draintxq(struct ath_softc *);
static void ath_stoprecv(struct ath_softc *);
static int ath_startrecv(struct ath_softc *);
Update with last year of work.
2004-12-08 17:34:36 +00:00
static void ath_chan_change(struct ath_softc *, struct ieee80211_channel *);
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
static void ath_scan_start(struct ieee80211com *);
static void ath_scan_end(struct ieee80211com *);
static void ath_set_channel(struct ieee80211com *);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void ath_calibrate(void *);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static int ath_newstate(struct ieee80211vap *, enum ieee80211_state, int);
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
static void ath_setup_stationkey(struct ieee80211_node *);
simplify ic_newassoc callback MFC after: 3 days
2005-07-22 17:57:16 +00:00
static void ath_newassoc(struct ieee80211_node *, int);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static int ath_setregdomain(struct ieee80211com *,
struct ieee80211_regdomain *, int,
struct ieee80211_channel []);
change ic_getradiocaps driver callback to include the max # channels so callers know the size of the array passed down
2009-01-27 23:19:36 +00:00
static void ath_getradiocaps(struct ieee80211com *, int, int *,
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211_channel []);
static int ath_getchannels(struct ath_softc *);
better led blinking
2005-01-18 19:03:04 +00:00
static void ath_led_event(struct ath_softc *, int);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
static int ath_rate_setup(struct ath_softc *, u_int mode);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void ath_setcurmode(struct ath_softc *, enum ieee80211_phymode);
Update with last year of work.
2004-12-08 17:34:36 +00:00
static void ath_announce(struct ath_softc *);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
static void ath_tdma_settimers(struct ath_softc *sc, u_int32_t nexttbtt,
u_int32_t bintval);
static void ath_tdma_bintvalsetup(struct ath_softc *sc,
const struct ieee80211_tdma_state *tdma);
static void ath_tdma_config(struct ath_softc *sc, struct ieee80211vap *vap);
static void ath_tdma_update(struct ieee80211_node *ni,
Minor cleanups of tdma protocol handling: o break out version-related code to simplify rev'ing the protocol o add parameter validation macros so checks that appear multiple places are consistent (and easy to change) o add protocol version check when looking for a scan candidate o improve scan debug output format o rewrite beacon update handling to calculate a bitmask of changed values and pass that down through the driver callback so drivers can optimize work o do slot bounds check before use when parsing received beacons
2009-03-18 19:28:17 +00:00
const struct ieee80211_tdma_param *tdma, int);
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
static void ath_tdma_beacon_send(struct ath_softc *sc,
struct ieee80211vap *vap);
static __inline void
ath_hal_setcca(struct ath_hal *ah, int ena)
{
/*
* NB: fill me in; this is not provided by default because disabling
* CCA in most locales violates regulatory.
*/
}
static __inline int
ath_hal_getcca(struct ath_hal *ah)
{
u_int32_t diag;
if (ath_hal_getcapability(ah, HAL_CAP_DIAG, 0, &diag) != HAL_OK)
return 1;
return ((diag & 0x500000) == 0);
}
#define TDMA_EP_MULTIPLIER (1<<10) /* pow2 to optimize out * and / */
#define TDMA_LPF_LEN 6
#define TDMA_DUMMY_MARKER 0x127
#define TDMA_EP_MUL(x, mul) ((x) * (mul))
#define TDMA_IN(x) (TDMA_EP_MUL((x), TDMA_EP_MULTIPLIER))
#define TDMA_LPF(x, y, len) \
((x != TDMA_DUMMY_MARKER) ? (((x) * ((len)-1) + (y)) / (len)) : (y))
#define TDMA_SAMPLE(x, y) do { \
x = TDMA_LPF((x), TDMA_IN(y), TDMA_LPF_LEN); \
} while (0)
#define TDMA_EP_RND(x,mul) \
((((x)%(mul)) >= ((mul)/2)) ? ((x) + ((mul) - 1)) / (mul) : (x)/(mul))
#define TDMA_AVG(x) TDMA_EP_RND(x, TDMA_EP_MULTIPLIER)
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#endif /* IEEE80211_SUPPORT_TDMA */
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
SYSCTL_DECL(_hw_ath);
/* XXX validate sysctl values */
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
static int ath_longcalinterval = 30; /* long cals every 30 secs */
SYSCTL_INT(_hw_ath, OID_AUTO, longcal, CTLFLAG_RW, &ath_longcalinterval,
0, "long chip calibration interval (secs)");
static int ath_shortcalinterval = 100; /* short cals every 100 ms */
SYSCTL_INT(_hw_ath, OID_AUTO, shortcal, CTLFLAG_RW, &ath_shortcalinterval,
0, "short chip calibration interval (msecs)");
static int ath_resetcalinterval = 20*60; /* reset cal state 20 mins */
SYSCTL_INT(_hw_ath, OID_AUTO, resetcal, CTLFLAG_RW, &ath_resetcalinterval,
0, "reset chip calibration results (secs)");
ANI changes #1 - split out the ANI polling from the RxMonitor hook. The rxmonitor hook is called on each received packet. This can get very, very busy as the tx/rx/chanbusy registers are thus read each time a packet is received. Instead, shuffle out the true per-packet processing which is needed and move the rest of the ANI processing into a periodic event which runs every 100ms by default.
2011-01-21 05:21:00 +00:00
static int ath_anicalinterval = 100; /* ANI calibration - 100 msec */
SYSCTL_INT(_hw_ath, OID_AUTO, anical, CTLFLAG_RW, &ath_anicalinterval,
0, "ANI calibration (msecs)");
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
allow the size of tx+rx buffer pools to be tuned MFC after: 2 weeks
2006-02-09 21:03:25 +00:00
static int ath_rxbuf = ATH_RXBUF; /* # rx buffers to allocate */
Add half/quarter rate 11a channel support: o change handling of regdomain-related mib knobs so they can be set post-attach: regdomain, countrycode, outdoor, and xchanmode; the hal will not permit changing the regdomain but we expose it for now o on regdomain/countrycode change recalculate the channel list and push it to the net80211 layer (NB: looks to need more tweaking) o setup rate tables for half/quarter rate channels o honor half/quarter rate channel configs when changing channels o honor half/quarter rate channel configs when setting the slot time o use hack/nonstandard channel numbering scheme for the public safety band to avoid overlapping 2.4G channels on dual-band cards o remove setup of ic_sup_rates; the net80211 layer can do this for us and it simplifies handling of half/quarter rate channels Tested only in Public Safety Band with cards that have RF5112.
2006-12-27 19:07:09 +00:00
SYSCTL_INT(_hw_ath, OID_AUTO, rxbuf, CTLFLAG_RW, &ath_rxbuf,
allow the size of tx+rx buffer pools to be tuned MFC after: 2 weeks
2006-02-09 21:03:25 +00:00
0, "rx buffers allocated");
TUNABLE_INT("hw.ath.rxbuf", &ath_rxbuf);
static int ath_txbuf = ATH_TXBUF; /* # tx buffers to allocate */
Add half/quarter rate 11a channel support: o change handling of regdomain-related mib knobs so they can be set post-attach: regdomain, countrycode, outdoor, and xchanmode; the hal will not permit changing the regdomain but we expose it for now o on regdomain/countrycode change recalculate the channel list and push it to the net80211 layer (NB: looks to need more tweaking) o setup rate tables for half/quarter rate channels o honor half/quarter rate channel configs when changing channels o honor half/quarter rate channel configs when setting the slot time o use hack/nonstandard channel numbering scheme for the public safety band to avoid overlapping 2.4G channels on dual-band cards o remove setup of ic_sup_rates; the net80211 layer can do this for us and it simplifies handling of half/quarter rate channels Tested only in Public Safety Band with cards that have RF5112.
2006-12-27 19:07:09 +00:00
SYSCTL_INT(_hw_ath, OID_AUTO, txbuf, CTLFLAG_RW, &ath_txbuf,
allow the size of tx+rx buffer pools to be tuned MFC after: 2 weeks
2006-02-09 21:03:25 +00:00
0, "tx buffers allocated");
TUNABLE_INT("hw.ath.txbuf", &ath_txbuf);
add hw.ath.bstuck to control the stuck beacon threshold
2009-02-10 19:26:42 +00:00
static int ath_bstuck_threshold = 4; /* max missed beacons */
SYSCTL_INT(_hw_ath, OID_AUTO, bstuck, CTLFLAG_RW, &ath_bstuck_threshold,
0, "max missed beacon xmits before chip reset");
Update with last year of work.
2004-12-08 17:34:36 +00:00
MALLOC_DEFINE(M_ATHDEV, "athdev", "ath driver dma buffers");
Add in the (very!) optional glue to flip the 11n bits for if_ath. There's still a lot of random issues to sort out with the radio side of things and AMPDU RX handling (and completely missing AMPDU TX handling!) but if people wish to give this a go and assist in debugging the issues, they can define ATH_DO_11N to enable it. I'm just re-iterating - this is here to allow people to assist in further 11n development; it is not any indication that the 11n support is complete and functional. Important notes: * This doesn't support 1-stream cards yet - (eg AR9285) - the various bits that negotiate TX/RX MCS don't know not to try >1 stream TX or negotiate 1-stream RX; so don't enable 11n unless you've first taught the rate control module and the net80211 stack to negotiate 1-stream stuff; * The only rate control module minimally 11n aware is ath_rate_sample; * ath_rate_sample doesn't know about HT/40; so airtime will be incorrectly calculated; * The AR9160 and AR9280 radio code is unreliable at the higher MCS rates for some reason; this will definitely impact 11n performance; * AMPDU-TX isn't yet implemented; * AMPDU-RX may be a bit buggy still and will definitely suffer from the radio unreliability mentioned above (ie, don't expect 150/300mbit RX just yet.)
2011-02-09 15:43:38 +00:00
#define HAL_MODE_HT20 (HAL_MODE_11NG_HT20 | HAL_MODE_11NA_HT20)
#define HAL_MODE_HT40 \
(HAL_MODE_11NG_HT40PLUS | HAL_MODE_11NG_HT40MINUS | \
HAL_MODE_11NA_HT40PLUS | HAL_MODE_11NA_HT40MINUS)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
int
ath_attach(u_int16_t devid, struct ath_softc *sc)
{
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211com *ic;
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ath_hal *ah = NULL;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
HAL_STATUS status;
Update with last year of work.
2004-12-08 17:34:36 +00:00
int error = 0, i;
prepare for a new hal
2008-10-27 18:30:33 +00:00
u_int wmodes;
Eliminate ic_myaddr so changing the mac address of a device works correctly: o remove ic_myaddr from ieee80211com o change ieee80211_ifattach to take the mac address of the physical device and use that to setup the lladdr. o replace all references to ic_myaddr in drivers by IF_LLADDR o related cleanups (e.g. kill dead code) PR: kern/133178 Reviewed by: thompsa, rpaulo
2009-03-29 17:59:14 +00:00
uint8_t macaddr[IEEE80211_ADDR_LEN];
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_ANY, "%s: devid 0x%x\n", __func__, devid);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ifp = sc->sc_ifp = if_alloc(IFT_IEEE80211);
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
if (ifp == NULL) {
device_printf(sc->sc_dev, "can not if_alloc()\n");
error = ENOSPC;
goto bad;
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ic = ifp->if_l2com;
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/* set these up early for if_printf use */
Replace the if_name and if_unit members of struct ifnet with new members if_xname, if_dname, and if_dunit. if_xname is the name of the interface and if_dname/unit are the driver name and instance. This change paves the way for interface renaming and enhanced pseudo device creation and configuration symantics. Approved By: re (in principle) Reviewed By: njl, imp Tested On: i386, amd64, sparc64 Obtained From: NetBSD (if_xname)
2003-10-31 18:32:15 +00:00
if_initname(ifp, device_get_name(sc->sc_dev),
Update with last year of work.
2004-12-08 17:34:36 +00:00
device_get_unit(sc->sc_dev));
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Include the initial support for external EEPROMs. The AR9100 at least doesn't have an external serial EEPROM attached to the MAC; it instead stores the calibration data in the normal system flash. I believe earlier parts can do something similar but I haven't experienced it first-hand. This commit introduces an eepromdata pointer into the API but doesn't at all commit to using it. A future commit will include the glue needed to allow the AR9100 support code to use this data pointer as the EEPROM.
2011-01-20 07:56:09 +00:00
ah = ath_hal_attach(devid, sc, sc->sc_st, sc->sc_sh, sc->sc_eepromdata, &status);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if (ah == NULL) {
if_printf(ifp, "unable to attach hardware; HAL status %u\n",
status);
error = ENXIO;
goto bad;
}
sc->sc_ah = ah;
Close a race where ath_intr is installed and may be called before the HAL is setup: use sc_invalid to discard such entries into ath_intr. This can easily happen if the device is assigned a shared IRQ.
2003-08-13 21:29:35 +00:00
sc->sc_invalid = 0; /* ready to go, enable interrupt handling */
shuffle debug setup to simplify debugging events during attach
2008-10-27 17:19:39 +00:00
#ifdef ATH_DEBUG
sc->sc_debug = ath_debug;
#endif
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Check if the MAC has multi-rate retry support.
* We do this by trying to setup a fake extended
* descriptor. MAC's that don't have support will
* return false w/o doing anything. MAC's that do
* support it will return true w/o doing anything.
*/
sc->sc_mrretry = ath_hal_setupxtxdesc(ah, NULL, 0,0, 0,0, 0,0);
/*
* Check if the device has hardware counters for PHY
* errors. If so we need to enable the MIB interrupt
* so we can act on stat triggers.
*/
if (ath_hal_hwphycounters(ah))
sc->sc_needmib = 1;
/*
* Get the hardware key cache size.
*/
sc->sc_keymax = ath_hal_keycachesize(ah);
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
if (sc->sc_keymax > ATH_KEYMAX) {
if_printf(ifp, "Warning, using only %u of %u key cache slots\n",
ATH_KEYMAX, sc->sc_keymax);
sc->sc_keymax = ATH_KEYMAX;
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
/*
* Reset the key cache since some parts do not
* reset the contents on initial power up.
*/
for (i = 0; i < sc->sc_keymax; i++)
ath_hal_keyreset(ah, i);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/*
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
* Collect the default channel list.
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
error = ath_getchannels(sc);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if (error != 0)
goto bad;
/*
* Setup rate tables for all potential media types.
*/
ath_rate_setup(sc, IEEE80211_MODE_11A);
ath_rate_setup(sc, IEEE80211_MODE_11B);
ath_rate_setup(sc, IEEE80211_MODE_11G);
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_rate_setup(sc, IEEE80211_MODE_TURBO_A);
ath_rate_setup(sc, IEEE80211_MODE_TURBO_G);
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
ath_rate_setup(sc, IEEE80211_MODE_STURBO_A);
ath_rate_setup(sc, IEEE80211_MODE_11NA);
ath_rate_setup(sc, IEEE80211_MODE_11NG);
Add initial support for 900MHz cards like the Ubiquiti SR9: o eliminate assumptions that half/quarter rate channels on exist in 11a o handle frequency mapping between hal and net80211; hal gives us freq's in the range 2422..2437 that we remap MFC after: 1 month
2007-01-15 01:15:57 +00:00
ath_rate_setup(sc, IEEE80211_MODE_HALF);
ath_rate_setup(sc, IEEE80211_MODE_QUARTER);
Add half/quarter rate 11a channel support: o change handling of regdomain-related mib knobs so they can be set post-attach: regdomain, countrycode, outdoor, and xchanmode; the hal will not permit changing the regdomain but we expose it for now o on regdomain/countrycode change recalculate the channel list and push it to the net80211 layer (NB: looks to need more tweaking) o setup rate tables for half/quarter rate channels o honor half/quarter rate channel configs when changing channels o honor half/quarter rate channel configs when setting the slot time o use hack/nonstandard channel numbering scheme for the public safety band to avoid overlapping 2.4G channels on dual-band cards o remove setup of ic_sup_rates; the net80211 layer can do this for us and it simplifies handling of half/quarter rate channels Tested only in Public Safety Band with cards that have RF5112.
2006-12-27 19:07:09 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
/* NB: setup here so ath_rate_update is happy */
ath_setcurmode(sc, IEEE80211_MODE_11A);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Allocate tx+rx descriptors and populate the lists.
*/
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
error = ath_desc_alloc(sc);
if (error != 0) {
if_printf(ifp, "failed to allocate descriptors: %d\n", error);
goto bad;
}
replace if_watchdog w/ private callout; probably can merge this with the calibration work sometime in the future
2009-03-09 23:10:19 +00:00
callout_init_mtx(&sc->sc_cal_ch, &sc->sc_mtx, 0);
callout_init_mtx(&sc->sc_wd_ch, &sc->sc_mtx, 0);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
o convert mutex calls to #defines for portability, etc. o destroy mutex's on detach (was missing)
2003-10-14 22:51:45 +00:00
ATH_TXBUF_LOCK_INIT(sc);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
use a private task queue thread MFC after: 2 weeks
2006-02-09 21:48:51 +00:00
sc->sc_tq = taskqueue_create("ath_taskq", M_NOWAIT,
taskqueue_thread_enqueue, &sc->sc_tq);
taskqueue_start_threads(&sc->sc_tq, 1, PI_NET,
"%s taskq", ifp->if_xname);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
TASK_INIT(&sc->sc_rxtask, 0, ath_rx_proc, sc);
TASK_INIT(&sc->sc_bmisstask, 0, ath_bmiss_proc, sc);
Update for rev 0.9.16.16 hal: o add dfs+radar hooks; DFS is presently disabled in the hal o channel and mode handling changes o various api changes o be more aggressive about iq calibration settling so ap mode operation is better immediately after startup o rfkill/rfsilent sysctl support o tpc ack/cts sysctl support MFC after: 2 weeks
2006-02-10 19:07:08 +00:00
TASK_INIT(&sc->sc_bstucktask,0, ath_bstuck_proc, sc);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/*
Update with last year of work.
2004-12-08 17:34:36 +00:00
* Allocate hardware transmit queues: one queue for
* beacon frames and one data queue for each QoS
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
* priority. Note that the hal handles resetting
Update with last year of work.
2004-12-08 17:34:36 +00:00
* these queues at the needed time.
*
* XXX PS-Poll
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
*/
setup the beacon xmit queue to not interrupt; we don't use them and they make the led's flash unnecessarily in adhoc mode
2005-01-18 19:10:17 +00:00
sc->sc_bhalq = ath_beaconq_setup(ah);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if (sc->sc_bhalq == (u_int) -1) {
if_printf(ifp, "unable to setup a beacon xmit queue!\n");
Update with last year of work.
2004-12-08 17:34:36 +00:00
error = EIO;
goto bad2;
}
sc->sc_cabq = ath_txq_setup(sc, HAL_TX_QUEUE_CAB, 0);
if (sc->sc_cabq == NULL) {
if_printf(ifp, "unable to setup CAB xmit queue!\n");
error = EIO;
goto bad2;
}
/* NB: insure BK queue is the lowest priority h/w queue */
if (!ath_tx_setup(sc, WME_AC_BK, HAL_WME_AC_BK)) {
if_printf(ifp, "unable to setup xmit queue for %s traffic!\n",
ieee80211_wme_acnames[WME_AC_BK]);
error = EIO;
goto bad2;
}
if (!ath_tx_setup(sc, WME_AC_BE, HAL_WME_AC_BE) ||
!ath_tx_setup(sc, WME_AC_VI, HAL_WME_AC_VI) ||
!ath_tx_setup(sc, WME_AC_VO, HAL_WME_AC_VO)) {
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
/*
Update with last year of work.
2004-12-08 17:34:36 +00:00
* Not enough hardware tx queues to properly do WME;
* just punt and assign them all to the same h/w queue.
* We could do a better job of this if, for example,
* we allocate queues when we switch from station to
* AP mode.
*/
if (sc->sc_ac2q[WME_AC_VI] != NULL)
ath_tx_cleanupq(sc, sc->sc_ac2q[WME_AC_VI]);
if (sc->sc_ac2q[WME_AC_BE] != NULL)
ath_tx_cleanupq(sc, sc->sc_ac2q[WME_AC_BE]);
sc->sc_ac2q[WME_AC_BE] = sc->sc_ac2q[WME_AC_BK];
sc->sc_ac2q[WME_AC_VI] = sc->sc_ac2q[WME_AC_BK];
sc->sc_ac2q[WME_AC_VO] = sc->sc_ac2q[WME_AC_BK];
}
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
/*
Update with last year of work.
2004-12-08 17:34:36 +00:00
* Special case certain configurations. Note the
* CAB queue is handled by these specially so don't
* include them when checking the txq setup mask.
*/
switch (sc->sc_txqsetup &~ (1<<sc->sc_cabq->axq_qnum)) {
case 0x01:
TASK_INIT(&sc->sc_txtask, 0, ath_tx_proc_q0, sc);
break;
case 0x0f:
TASK_INIT(&sc->sc_txtask, 0, ath_tx_proc_q0123, sc);
break;
default:
TASK_INIT(&sc->sc_txtask, 0, ath_tx_proc, sc);
break;
}
/*
* Setup rate control. Some rate control modules
* call back to change the anntena state so expose
* the necessary entry points.
* XXX maybe belongs in struct ath_ratectrl?
*/
sc->sc_setdefantenna = ath_setdefantenna;
sc->sc_rc = ath_rate_attach(sc);
if (sc->sc_rc == NULL) {
error = EIO;
cleanup descriptor allocation if attach fails Obtained from: madwifi
2004-04-02 23:47:39 +00:00
goto bad2;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
better led blinking
2005-01-18 19:03:04 +00:00
sc->sc_blinking = 0;
Update with last year of work.
2004-12-08 17:34:36 +00:00
sc->sc_ledstate = 1;
better led blinking
2005-01-18 19:03:04 +00:00
sc->sc_ledon = 0; /* low true */
sc->sc_ledidle = (2700*hz)/1000; /* 2.7sec */
callout_init(&sc->sc_ledtimer, CALLOUT_MPSAFE);
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Auto-enable soft led processing for IBM cards and for
* 5211 minipci cards. Users can also manually enable/disable
* support with a sysctl.
*/
sc->sc_softled = (devid == AR5212_DEVID_IBM || devid == AR5211_DEVID);
if (sc->sc_softled) {
5416 and later parts mux the gpio outputs; extend the api to include a signal type that's used to select the appropriate mux
2009-02-24 00:12:16 +00:00
ath_hal_gpioCfgOutput(ah, sc->sc_ledpin,
HAL_GPIO_MUX_MAC_NETWORK_LED);
better led blinking
2005-01-18 19:03:04 +00:00
ath_hal_gpioset(ah, sc->sc_ledpin, !sc->sc_ledon);
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ifp->if_softc = sc;
ifp->if_flags = IFF_SIMPLEX | IFF_BROADCAST | IFF_MULTICAST;
ifp->if_start = ath_start;
ifp->if_ioctl = ath_ioctl;
ifp->if_init = ath_init;
Add new tunable 'net.link.ifqmaxlen' to set default send interface queue length. The default value for this parameter is 50, which is quite low for many of today's uses and the only way to modify this parameter right now is to edit if_var.h file. Also add read-only sysctl with the same name, so that it's possible to retrieve the current value. MFC after: 1 month
2010-05-03 07:32:50 +00:00
IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
Second part of ALTQ driver modifications, covering: an(4), ath(4), hme(4), ndis(4), vr(4) and wi(4) Please help testing: http://people.freebsd.org/~mlaier/ALTQ_driver/ Tested by: Vaidas Damosevicius (an, ath, wi) Roman Divacky (vr) Submitted by: yongari (hme)
2004-08-01 23:58:04 +00:00
IFQ_SET_READY(&ifp->if_snd);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
ic->ic_ifp = ifp;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/* XXX not right but it's not used anywhere important */
ic->ic_phytype = IEEE80211_T_OFDM;
ic->ic_opmode = IEEE80211_M_STA;
Update with last year of work.
2004-12-08 17:34:36 +00:00
ic->ic_caps =
Minor cleanup of vap create work: o add IEEE80211_C_STA capability to indicate sta mode is supported (was previously assumed) and mark drivers as capable o add ieee80211_opcap array to map an opmode to the equivalent capability bit o move IEEE80211_C_OPMODE definition to where capabilities are defined so it's clear it should be kept in sync (on future additions) o check device capabilities in clone create before trying to create a vap; this makes driver checks unneeded o make error codes return on failed clone request unique o temporarily add console printfs on clone request failures to aid in debugging; these will move under DIAGNOSTIC or similar before release
2008-05-12 00:15:30 +00:00
IEEE80211_C_STA /* station mode */
| IEEE80211_C_IBSS /* ibss, nee adhoc, mode */
indicate device receives all management frames
2003-10-17 21:55:53 +00:00
| IEEE80211_C_HOSTAP /* hostap mode */
| IEEE80211_C_MONITOR /* monitor mode */
add adhoc demo mode support MFC after: 2 weeks
2006-02-09 21:42:53 +00:00
| IEEE80211_C_AHDEMO /* adhoc demo mode */
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
| IEEE80211_C_WDS /* 4-address traffic works */
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
| IEEE80211_C_MBSS /* mesh point link mode */
indicate device receives all management frames
2003-10-17 21:55:53 +00:00
| IEEE80211_C_SHPREAMBLE /* short preamble supported */
Update with last year of work.
2004-12-08 17:34:36 +00:00
| IEEE80211_C_SHSLOT /* short slot time supported */
| IEEE80211_C_WPA /* capable of WPA1+WPA2 */
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
| IEEE80211_C_BGSCAN /* capable of bg scanning */
| IEEE80211_C_TXFRAG /* handle tx frags */
remove use IEEE80211_C_RCVMGT
2004-04-02 23:37:00 +00:00
;
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Query the hal to figure out h/w crypto support.
*/
if (ath_hal_ciphersupported(ah, HAL_CIPHER_WEP))
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ic->ic_cryptocaps |= IEEE80211_CRYPTO_WEP;
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (ath_hal_ciphersupported(ah, HAL_CIPHER_AES_OCB))
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ic->ic_cryptocaps |= IEEE80211_CRYPTO_AES_OCB;
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (ath_hal_ciphersupported(ah, HAL_CIPHER_AES_CCM))
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ic->ic_cryptocaps |= IEEE80211_CRYPTO_AES_CCM;
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (ath_hal_ciphersupported(ah, HAL_CIPHER_CKIP))
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ic->ic_cryptocaps |= IEEE80211_CRYPTO_CKIP;
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (ath_hal_ciphersupported(ah, HAL_CIPHER_TKIP)) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ic->ic_cryptocaps |= IEEE80211_CRYPTO_TKIP;
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Check if h/w does the MIC and/or whether the
* separate key cache entries are required to
* handle both tx+rx MIC keys.
*/
if (ath_hal_ciphersupported(ah, HAL_CIPHER_MIC))
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ic->ic_cryptocaps |= IEEE80211_CRYPTO_TKIPMIC;
Add support for newer parts that do not require separate keycache entries for tx+rx mic keys. This requires a newer hal, but works fine with the current hal in cvs. MFC after: 2 weeks
2006-09-18 16:26:19 +00:00
/*
* If the h/w supports storing tx+rx MIC keys
* in one cache slot automatically enable use.
*/
if (ath_hal_hastkipsplit(ah) ||
!ath_hal_settkipsplit(ah, AH_FALSE))
Update with last year of work.
2004-12-08 17:34:36 +00:00
sc->sc_splitmic = 1;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/*
* If the h/w can do TKIP MIC together with WME then
* we use it; otherwise we force the MIC to be done
* in software by the net80211 layer.
*/
if (ath_hal_haswmetkipmic(ah))
sc->sc_wmetkipmic = 1;
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
sc->sc_hasclrkey = ath_hal_ciphersupported(ah, HAL_CIPHER_CLR);
Add multicast key search support. This fixes corrupted mcast packets when we have more than one hostap vap. Submitted by: Russell Yount <russell.yount at gmail.com> MFC after: 2 weeks
2010-02-08 20:23:20 +00:00
/*
Fix typo in comment. Pointed out by: danfe
2010-02-10 11:12:39 +00:00
* Check for multicast key search support.
Add multicast key search support. This fixes corrupted mcast packets when we have more than one hostap vap. Submitted by: Russell Yount <russell.yount at gmail.com> MFC after: 2 weeks
2010-02-08 20:23:20 +00:00
*/
if (ath_hal_hasmcastkeysearch(sc->sc_ah) &&
!ath_hal_getmcastkeysearch(sc->sc_ah)) {
ath_hal_setmcastkeysearch(sc->sc_ah, 1);
}
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
sc->sc_mcastkey = ath_hal_getmcastkeysearch(ah);
Add support for newer parts that do not require separate keycache entries for tx+rx mic keys. This requires a newer hal, but works fine with the current hal in cvs. MFC after: 2 weeks
2006-09-18 16:26:19 +00:00
/*
* Mark key cache slots associated with global keys
* as in use. If we knew TKIP was not to be used we
* could leave the +32, +64, and +32+64 slots free.
*/
for (i = 0; i < IEEE80211_WEP_NKID; i++) {
setbit(sc->sc_keymap, i);
setbit(sc->sc_keymap, i+64);
if (sc->sc_splitmic) {
setbit(sc->sc_keymap, i+32);
setbit(sc->sc_keymap, i+32+64);
}
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* TPC support can be done either with a global cap or
* per-packet support. The latter is not available on
* all parts. We're a bit pedantic here as all parts
* support a global cap.
*/
o fix setup of sc_diversity; the hal does not give us reliable status after attach, only after a reset o when setting diversity via the sysctl don't update sc_diversity until we know the hal requested worked o while here eliminate sc_hasdiversity and sc_hastpc; just query the hal each time since these are the only places we need to know MFC after: 3 days
2005-07-24 05:11:39 +00:00
if (ath_hal_hastpc(ah) || ath_hal_hastxpowlimit(ah))
Update with last year of work.
2004-12-08 17:34:36 +00:00
ic->ic_caps |= IEEE80211_C_TXPMGT;
/*
* Mark WME capability only if we have sufficient
* hardware queues to do proper priority scheduling.
*/
if (sc->sc_ac2q[WME_AC_BE] != sc->sc_ac2q[WME_AC_BK])
ic->ic_caps |= IEEE80211_C_WME;
/*
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
* Check for misc other capabilities.
Update with last year of work.
2004-12-08 17:34:36 +00:00
*/
if (ath_hal_hasbursting(ah))
ic->ic_caps |= IEEE80211_C_BURST;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
sc->sc_hasbmask = ath_hal_hasbssidmask(ah);
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
sc->sc_hasbmatch = ath_hal_hasbssidmatch(ah);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
sc->sc_hastsfadd = ath_hal_hastsfadjust(ah);
Add a HAL capability bit for supporting self-linked RX descriptors and disable it for the 11n chipsets. From the ath9k source: == 11N: we can no longer afford to self link the last descriptor. MAC acknowledges BA status as long as it copies frames to host buffer (or rx fifo). This can incorrectly acknowledge packets to a sender if last desc is self-linked. == Since this is useful for pre-AR5416 chips that communicate PHY errors via error frames rather than by on-chip counters, leave the support in there, but disable it for AR5416 and later.
2011-04-04 14:52:31 +00:00
sc->sc_rxslink = ath_hal_self_linked_final_rxdesc(ah);
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
if (ath_hal_hasfastframes(ah))
ic->ic_caps |= IEEE80211_C_FF;
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
wmodes = ath_hal_getwirelessmodes(ah);
prepare for a new hal
2008-10-27 18:30:33 +00:00
if (wmodes & (HAL_MODE_108G|HAL_MODE_TURBO))
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
ic->ic_caps |= IEEE80211_C_TURBOP;
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
if (ath_hal_macversion(ah) > 0x78) {
ic->ic_caps |= IEEE80211_C_TDMA; /* capable of TDMA */
ic->ic_tdma_update = ath_tdma_update;
}
#endif
Add in the (very!) optional glue to flip the 11n bits for if_ath. There's still a lot of random issues to sort out with the radio side of things and AMPDU RX handling (and completely missing AMPDU TX handling!) but if people wish to give this a go and assist in debugging the issues, they can define ATH_DO_11N to enable it. I'm just re-iterating - this is here to allow people to assist in further 11n development; it is not any indication that the 11n support is complete and functional. Important notes: * This doesn't support 1-stream cards yet - (eg AR9285) - the various bits that negotiate TX/RX MCS don't know not to try >1 stream TX or negotiate 1-stream RX; so don't enable 11n unless you've first taught the rate control module and the net80211 stack to negotiate 1-stream stuff; * The only rate control module minimally 11n aware is ath_rate_sample; * ath_rate_sample doesn't know about HT/40; so airtime will be incorrectly calculated; * The AR9160 and AR9280 radio code is unreliable at the higher MCS rates for some reason; this will definitely impact 11n performance; * AMPDU-TX isn't yet implemented; * AMPDU-RX may be a bit buggy still and will definitely suffer from the radio unreliability mentioned above (ie, don't expect 150/300mbit RX just yet.)
2011-02-09 15:43:38 +00:00
/*
* The if_ath 11n support is completely not ready for normal use.
* Enabling this option will likely break everything and everything.
* Don't think of doing that unless you know what you're doing.
*/
Rename AH_ENABLE_11N to ATH_ENABLE_11 - the HAL supports 11n by default but the ath driver doesn't. This is a much more consistent name.
2011-03-27 08:47:55 +00:00
#ifdef ATH_ENABLE_11N
Add in the (very!) optional glue to flip the 11n bits for if_ath. There's still a lot of random issues to sort out with the radio side of things and AMPDU RX handling (and completely missing AMPDU TX handling!) but if people wish to give this a go and assist in debugging the issues, they can define ATH_DO_11N to enable it. I'm just re-iterating - this is here to allow people to assist in further 11n development; it is not any indication that the 11n support is complete and functional. Important notes: * This doesn't support 1-stream cards yet - (eg AR9285) - the various bits that negotiate TX/RX MCS don't know not to try >1 stream TX or negotiate 1-stream RX; so don't enable 11n unless you've first taught the rate control module and the net80211 stack to negotiate 1-stream stuff; * The only rate control module minimally 11n aware is ath_rate_sample; * ath_rate_sample doesn't know about HT/40; so airtime will be incorrectly calculated; * The AR9160 and AR9280 radio code is unreliable at the higher MCS rates for some reason; this will definitely impact 11n performance; * AMPDU-TX isn't yet implemented; * AMPDU-RX may be a bit buggy still and will definitely suffer from the radio unreliability mentioned above (ie, don't expect 150/300mbit RX just yet.)
2011-02-09 15:43:38 +00:00
/*
* Query HT capabilities
*/
if (ath_hal_getcapability(ah, HAL_CAP_HT, 0, NULL) == HAL_OK &&
(wmodes & (HAL_MODE_HT20 | HAL_MODE_HT40))) {
int rxs, txs;
device_printf(sc->sc_dev, "[HT] enabling HT modes\n");
ic->ic_htcaps = IEEE80211_HTC_HT /* HT operation */
| IEEE80211_HTC_AMPDU /* A-MPDU tx/rx */
| IEEE80211_HTC_AMSDU /* A-MSDU tx/rx */
| IEEE80211_HTCAP_MAXAMSDU_3839 /* max A-MSDU length */
Disable short-GI in 20mhz mode - the hardware doesn't support this.
2011-02-17 17:31:36 +00:00
/* At the present time, the hardware doesn't support short-GI in 20mhz mode */
#if 0
Add in the (very!) optional glue to flip the 11n bits for if_ath. There's still a lot of random issues to sort out with the radio side of things and AMPDU RX handling (and completely missing AMPDU TX handling!) but if people wish to give this a go and assist in debugging the issues, they can define ATH_DO_11N to enable it. I'm just re-iterating - this is here to allow people to assist in further 11n development; it is not any indication that the 11n support is complete and functional. Important notes: * This doesn't support 1-stream cards yet - (eg AR9285) - the various bits that negotiate TX/RX MCS don't know not to try >1 stream TX or negotiate 1-stream RX; so don't enable 11n unless you've first taught the rate control module and the net80211 stack to negotiate 1-stream stuff; * The only rate control module minimally 11n aware is ath_rate_sample; * ath_rate_sample doesn't know about HT/40; so airtime will be incorrectly calculated; * The AR9160 and AR9280 radio code is unreliable at the higher MCS rates for some reason; this will definitely impact 11n performance; * AMPDU-TX isn't yet implemented; * AMPDU-RX may be a bit buggy still and will definitely suffer from the radio unreliability mentioned above (ie, don't expect 150/300mbit RX just yet.)
2011-02-09 15:43:38 +00:00
| IEEE80211_HTCAP_SHORTGI20 /* short GI in 20MHz */
Disable short-GI in 20mhz mode - the hardware doesn't support this.
2011-02-17 17:31:36 +00:00
#endif
Add in the (very!) optional glue to flip the 11n bits for if_ath. There's still a lot of random issues to sort out with the radio side of things and AMPDU RX handling (and completely missing AMPDU TX handling!) but if people wish to give this a go and assist in debugging the issues, they can define ATH_DO_11N to enable it. I'm just re-iterating - this is here to allow people to assist in further 11n development; it is not any indication that the 11n support is complete and functional. Important notes: * This doesn't support 1-stream cards yet - (eg AR9285) - the various bits that negotiate TX/RX MCS don't know not to try >1 stream TX or negotiate 1-stream RX; so don't enable 11n unless you've first taught the rate control module and the net80211 stack to negotiate 1-stream stuff; * The only rate control module minimally 11n aware is ath_rate_sample; * ath_rate_sample doesn't know about HT/40; so airtime will be incorrectly calculated; * The AR9160 and AR9280 radio code is unreliable at the higher MCS rates for some reason; this will definitely impact 11n performance; * AMPDU-TX isn't yet implemented; * AMPDU-RX may be a bit buggy still and will definitely suffer from the radio unreliability mentioned above (ie, don't expect 150/300mbit RX just yet.)
2011-02-09 15:43:38 +00:00
| IEEE80211_HTCAP_SMPS_OFF; /* SM power save off */
;
if (wmodes & HAL_MODE_HT40)
ic->ic_htcaps |= IEEE80211_HTCAP_CHWIDTH40
| IEEE80211_HTCAP_SHORTGI40;
/*
* rx/tx stream is not currently used anywhere; it needs to be taken
* into account when negotiating which MCS rates it'll receive and
* what MCS rates are available for TX.
*/
(void) ath_hal_getcapability(ah, HAL_CAP_STREAMS, 0, &rxs);
(void) ath_hal_getcapability(ah, HAL_CAP_STREAMS, 1, &txs);
ath_hal_getrxchainmask(ah, &sc->sc_rxchainmask);
ath_hal_gettxchainmask(ah, &sc->sc_txchainmask);
ic->ic_txstream = txs;
ic->ic_rxstream = rxs;
device_printf(sc->sc_dev, "[HT] %d RX streams; %d TX streams\n", rxs, txs);
}
#endif
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Indicate we need the 802.11 header padded to a
* 32-bit boundary for 4-address and QoS frames.
*/
ic->ic_flags |= IEEE80211_F_DATAPAD;
/*
* Query the hal about antenna support.
*/
sc->sc_defant = ath_hal_getdefantenna(ah);
/*
* Not all chips have the VEOL support we want to
* use with IBSS beacons; check here for it.
*/
sc->sc_hasveol = ath_hal_hasveol(ah);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/* get mac address from hardware */
Eliminate ic_myaddr so changing the mac address of a device works correctly: o remove ic_myaddr from ieee80211com o change ieee80211_ifattach to take the mac address of the physical device and use that to setup the lladdr. o replace all references to ic_myaddr in drivers by IF_LLADDR o related cleanups (e.g. kill dead code) PR: kern/133178 Reviewed by: thompsa, rpaulo
2009-03-29 17:59:14 +00:00
ath_hal_getmac(ah, macaddr);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (sc->sc_hasbmask)
ath_hal_getbssidmask(ah, sc->sc_hwbssidmask);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/* NB: used to size node table key mapping array */
ic->ic_max_keyix = sc->sc_keymax;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/* call MI attach routine. */
Eliminate ic_myaddr so changing the mac address of a device works correctly: o remove ic_myaddr from ieee80211com o change ieee80211_ifattach to take the mac address of the physical device and use that to setup the lladdr. o replace all references to ic_myaddr in drivers by IF_LLADDR o related cleanups (e.g. kill dead code) PR: kern/133178 Reviewed by: thompsa, rpaulo
2009-03-29 17:59:14 +00:00
ieee80211_ifattach(ic, macaddr);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ic->ic_setregdomain = ath_setregdomain;
ic->ic_getradiocaps = ath_getradiocaps;
sc->sc_opmode = HAL_M_STA;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/* override default methods */
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ic->ic_newassoc = ath_newassoc;
ic->ic_updateslot = ath_updateslot;
ic->ic_wme.wme_update = ath_wme_update;
ic->ic_vap_create = ath_vap_create;
ic->ic_vap_delete = ath_vap_delete;
ic->ic_raw_xmit = ath_raw_xmit;
ic->ic_update_mcast = ath_update_mcast;
ic->ic_update_promisc = ath_update_promisc;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ic->ic_node_alloc = ath_node_alloc;
do proper subclassing of node free+copy; the previous hack falls apart when the 802.11 layer does useful work Obtained from: madwifi
2004-04-03 00:06:23 +00:00
sc->sc_node_free = ic->ic_node_free;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ic->ic_node_free = ath_node_free;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
ic->ic_node_getsignal = ath_node_getsignal;
ic->ic_scan_start = ath_scan_start;
ic->ic_scan_end = ath_scan_end;
ic->ic_set_channel = ath_set_channel;
Add support for the experimental radiotap capture format. With this we no longer need the debugging code to dump packets.
2003-09-05 22:22:49 +00:00
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
ieee80211_radiotap_attach(ic,
&sc->sc_tx_th.wt_ihdr, sizeof(sc->sc_tx_th),
ATH_TX_RADIOTAP_PRESENT,
&sc->sc_rx_th.wr_ihdr, sizeof(sc->sc_rx_th),
ATH_RX_RADIOTAP_PRESENT);
o move ath_sysctlattach down so variables it depends on are setup o use any fixed tx antenna for beacons transmitted in adhoc mode Submitted by: David Young MFC after: 3 days
2005-07-23 05:34:29 +00:00
/*
* Setup dynamic sysctl's now that country code and
* regdomain are available from the hal.
*/
ath_sysctlattach(sc);
Export ath stats via snmp, rather than requiring a debugging interface and "athstats".
2010-08-14 14:18:02 +00:00
ath_sysctl_stats_attach(sc);
Add support for the experimental radiotap capture format. With this we no longer need the debugging code to dump packets.
2003-09-05 22:22:49 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (bootverbose)
ieee80211_announce(ic);
ath_announce(sc);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
return 0;
cleanup descriptor allocation if attach fails Obtained from: madwifi
2004-04-02 23:47:39 +00:00
bad2:
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_tx_cleanup(sc);
cleanup descriptor allocation if attach fails Obtained from: madwifi
2004-04-02 23:47:39 +00:00
ath_desc_free(sc);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
bad:
if (ah)
ath_hal_detach(ah);
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
if (ifp != NULL)
if_free(ifp);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_invalid = 1;
return error;
}
int
ath_detach(struct ath_softc *sc)
{
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_ANY, "%s: if_flags %x\n",
__func__, ifp->if_flags);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
/*
Update with last year of work.
2004-12-08 17:34:36 +00:00
* NB: the order of these is important:
close a race on detach by reordering bpfdetach and taskqueue_free
2008-05-28 23:55:11 +00:00
* o stop the chip so no more interrupts will fire
Update with last year of work.
2004-12-08 17:34:36 +00:00
* o call the 802.11 layer before detaching the hal to
* insure callbacks into the driver to delete global
* key cache entries can be handled
close a race on detach by reordering bpfdetach and taskqueue_free
2008-05-28 23:55:11 +00:00
* o free the taskqueue which drains any pending tasks
Update with last year of work.
2004-12-08 17:34:36 +00:00
* o reclaim the tx queue data structures after calling
* the 802.11 layer as we'll get called back to reclaim
* node state and potentially want to use them
* o to cleanup the tx queues the hal is called, so detach
* it last
* Other than that, it's straightforward...
*/
close a race on detach by reordering bpfdetach and taskqueue_free
2008-05-28 23:55:11 +00:00
ath_stop(ifp);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ieee80211_ifdetach(ifp->if_l2com);
close a race on detach by reordering bpfdetach and taskqueue_free
2008-05-28 23:55:11 +00:00
taskqueue_free(sc->sc_tq);
add tx99 hooks MFC after: 2 weeks
2006-02-09 21:28:11 +00:00
#ifdef ATH_TX99_DIAG
if (sc->sc_tx99 != NULL)
sc->sc_tx99->detach(sc->sc_tx99);
#endif
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_rate_detach(sc->sc_rc);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_desc_free(sc);
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_tx_cleanup(sc);
close a race on detach by reordering bpfdetach and taskqueue_free
2008-05-28 23:55:11 +00:00
ath_hal_detach(sc->sc_ah); /* NB: sets chip in full sleep */
Fix "struct ifnet" leak on detach.
2005-09-16 10:09:23 +00:00
if_free(ifp);
o convert mutex calls to #defines for portability, etc. o destroy mutex's on detach (was missing)
2003-10-14 22:51:45 +00:00
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
return 0;
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/*
* MAC address handling for multiple BSS on the same radio.
* The first vap uses the MAC address from the EEPROM. For
* subsequent vap's we set the U/L bit (bit 1) in the MAC
* address and use the next six bits as an index.
*/
static void
assign_address(struct ath_softc *sc, uint8_t mac[IEEE80211_ADDR_LEN], int clone)
{
int i;
if (clone && sc->sc_hasbmask) {
/* NB: we only do this if h/w supports multiple bssid */
for (i = 0; i < 8; i++)
if ((sc->sc_bssidmask & (1<<i)) == 0)
break;
if (i != 0)
mac[0] |= (i << 2)|0x2;
} else
i = 0;
sc->sc_bssidmask |= 1<<i;
sc->sc_hwbssidmask[0] &= ~mac[0];
if (i == 0)
sc->sc_nbssid0++;
}
static void
reclaim_address(struct ath_softc *sc, const uint8_t mac[IEEE80211_ADDR_LEN])
{
int i = mac[0] >> 2;
uint8_t mask;
if (i != 0 || --sc->sc_nbssid0 == 0) {
sc->sc_bssidmask &= ~(1<<i);
/* recalculate bssid mask from remaining addresses */
mask = 0xff;
for (i = 1; i < 8; i++)
if (sc->sc_bssidmask & (1<<i))
mask &= ~((i<<2)|0x2);
sc->sc_hwbssidmask[0] |= mask;
}
}
/*
* Assign a beacon xmit slot. We try to space out
* assignments so when beacons are staggered the
* traffic coming out of the cab q has maximal time
* to go out before the next beacon is scheduled.
*/
static int
assign_bslot(struct ath_softc *sc)
{
u_int slot, free;
free = 0;
for (slot = 0; slot < ATH_BCBUF; slot++)
if (sc->sc_bslot[slot] == NULL) {
if (sc->sc_bslot[(slot+1)%ATH_BCBUF] == NULL &&
sc->sc_bslot[(slot-1)%ATH_BCBUF] == NULL)
return slot;
free = slot;
/* NB: keep looking for a double slot */
}
return free;
}
static struct ieee80211vap *
ath_vap_create(struct ieee80211com *ic,
const char name[IFNAMSIZ], int unit, int opmode, int flags,
const uint8_t bssid[IEEE80211_ADDR_LEN],
const uint8_t mac0[IEEE80211_ADDR_LEN])
{
struct ath_softc *sc = ic->ic_ifp->if_softc;
struct ath_vap *avp;
struct ieee80211vap *vap;
uint8_t mac[IEEE80211_ADDR_LEN];
int ic_opmode, needbeacon, error;
avp = (struct ath_vap *) malloc(sizeof(struct ath_vap),
M_80211_VAP, M_WAITOK | M_ZERO);
needbeacon = 0;
IEEE80211_ADDR_COPY(mac, mac0);
ATH_LOCK(sc);
o cleanup checks for which vap combinations are permitted and what to use for ic_opmode o fixes the case where creating ahdemo+wds vaps caused ic_opmode to be set to hostap
2009-05-06 23:49:55 +00:00
ic_opmode = opmode; /* default to opmode of new vap */
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
switch (opmode) {
case IEEE80211_M_STA:
o cleanup checks for which vap combinations are permitted and what to use for ic_opmode o fixes the case where creating ahdemo+wds vaps caused ic_opmode to be set to hostap
2009-05-06 23:49:55 +00:00
if (sc->sc_nstavaps != 0) { /* XXX only 1 for now */
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
device_printf(sc->sc_dev, "only 1 sta vap supported\n");
goto bad;
}
if (sc->sc_nvaps) {
/*
o cleanup checks for which vap combinations are permitted and what to use for ic_opmode o fixes the case where creating ahdemo+wds vaps caused ic_opmode to be set to hostap
2009-05-06 23:49:55 +00:00
* With multiple vaps we must fall back
* to s/w beacon miss handling.
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
*/
flags |= IEEE80211_CLONE_NOBEACONS;
}
o cleanup checks for which vap combinations are permitted and what to use for ic_opmode o fixes the case where creating ahdemo+wds vaps caused ic_opmode to be set to hostap
2009-05-06 23:49:55 +00:00
if (flags & IEEE80211_CLONE_NOBEACONS) {
/*
* Station mode w/o beacons are implemented w/ AP mode.
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ic_opmode = IEEE80211_M_HOSTAP;
o cleanup checks for which vap combinations are permitted and what to use for ic_opmode o fixes the case where creating ahdemo+wds vaps caused ic_opmode to be set to hostap
2009-05-06 23:49:55 +00:00
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
break;
case IEEE80211_M_IBSS:
if (sc->sc_nvaps != 0) { /* XXX only 1 for now */
device_printf(sc->sc_dev,
"only 1 ibss vap supported\n");
goto bad;
}
needbeacon = 1;
break;
case IEEE80211_M_AHDEMO:
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
if (flags & IEEE80211_CLONE_TDMA) {
o cleanup checks for which vap combinations are permitted and what to use for ic_opmode o fixes the case where creating ahdemo+wds vaps caused ic_opmode to be set to hostap
2009-05-06 23:49:55 +00:00
if (sc->sc_nvaps != 0) {
device_printf(sc->sc_dev,
"only 1 tdma vap supported\n");
goto bad;
}
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
needbeacon = 1;
flags |= IEEE80211_CLONE_NOBEACONS;
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/* fall thru... */
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
#endif
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
case IEEE80211_M_MONITOR:
if (sc->sc_nvaps != 0 && ic->ic_opmode != opmode) {
o cleanup checks for which vap combinations are permitted and what to use for ic_opmode o fixes the case where creating ahdemo+wds vaps caused ic_opmode to be set to hostap
2009-05-06 23:49:55 +00:00
/*
* Adopt existing mode. Adding a monitor or ahdemo
* vap to an existing configuration is of dubious
* value but should be ok.
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/* XXX not right for monitor mode */
ic_opmode = ic->ic_opmode;
o cleanup checks for which vap combinations are permitted and what to use for ic_opmode o fixes the case where creating ahdemo+wds vaps caused ic_opmode to be set to hostap
2009-05-06 23:49:55 +00:00
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
break;
case IEEE80211_M_HOSTAP:
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
case IEEE80211_M_MBSS:
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
needbeacon = 1;
o cleanup checks for which vap combinations are permitted and what to use for ic_opmode o fixes the case where creating ahdemo+wds vaps caused ic_opmode to be set to hostap
2009-05-06 23:49:55 +00:00
break;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
case IEEE80211_M_WDS:
o cleanup checks for which vap combinations are permitted and what to use for ic_opmode o fixes the case where creating ahdemo+wds vaps caused ic_opmode to be set to hostap
2009-05-06 23:49:55 +00:00
if (sc->sc_nvaps != 0 && ic->ic_opmode == IEEE80211_M_STA) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
device_printf(sc->sc_dev,
"wds not supported in sta mode\n");
goto bad;
}
o cleanup checks for which vap combinations are permitted and what to use for ic_opmode o fixes the case where creating ahdemo+wds vaps caused ic_opmode to be set to hostap
2009-05-06 23:49:55 +00:00
/*
* Silently remove any request for a unique
* bssid; WDS vap's always share the local
* mac address.
*/
flags &= ~IEEE80211_CLONE_BSSID;
if (sc->sc_nvaps == 0)
ic_opmode = IEEE80211_M_HOSTAP;
else
ic_opmode = ic->ic_opmode;
Fix something bogus deletion that got it during mesh commit. Approved by: re (implicit)
2009-07-11 16:02:06 +00:00
break;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
default:
device_printf(sc->sc_dev, "unknown opmode %d\n", opmode);
goto bad;
}
/*
* Check that a beacon buffer is available; the code below assumes it.
*/
if (needbeacon & STAILQ_EMPTY(&sc->sc_bbuf)) {
device_printf(sc->sc_dev, "no beacon buffer available\n");
goto bad;
}
/* STA, AHDEMO? */
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
if (opmode == IEEE80211_M_HOSTAP || opmode == IEEE80211_M_MBSS) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
assign_address(sc, mac, flags & IEEE80211_CLONE_BSSID);
ath_hal_setbssidmask(sc->sc_ah, sc->sc_hwbssidmask);
}
vap = &avp->av_vap;
/* XXX can't hold mutex across if_alloc */
ATH_UNLOCK(sc);
error = ieee80211_vap_setup(ic, vap, name, unit, opmode, flags,
bssid, mac);
ATH_LOCK(sc);
if (error != 0) {
device_printf(sc->sc_dev, "%s: error %d creating vap\n",
__func__, error);
goto bad2;
}
/* h/w crypto support */
vap->iv_key_alloc = ath_key_alloc;
vap->iv_key_delete = ath_key_delete;
vap->iv_key_set = ath_key_set;
vap->iv_key_update_begin = ath_key_update_begin;
vap->iv_key_update_end = ath_key_update_end;
/* override various methods */
avp->av_recv_mgmt = vap->iv_recv_mgmt;
vap->iv_recv_mgmt = ath_recv_mgmt;
vap->iv_reset = ath_reset_vap;
vap->iv_update_beacon = ath_beacon_update;
avp->av_newstate = vap->iv_newstate;
vap->iv_newstate = ath_newstate;
avp->av_bmiss = vap->iv_bmiss;
vap->iv_bmiss = ath_bmiss_vap;
avp->av_bslot = -1;
if (needbeacon) {
/*
* Allocate beacon state and setup the q for buffered
* multicast frames. We know a beacon buffer is
* available because we checked above.
*/
avp->av_bcbuf = STAILQ_FIRST(&sc->sc_bbuf);
STAILQ_REMOVE_HEAD(&sc->sc_bbuf, bf_list);
if (opmode != IEEE80211_M_IBSS || !sc->sc_hasveol) {
/*
* Assign the vap to a beacon xmit slot. As above
* this cannot fail to find a free one.
*/
avp->av_bslot = assign_bslot(sc);
KASSERT(sc->sc_bslot[avp->av_bslot] == NULL,
("beacon slot %u not empty", avp->av_bslot));
sc->sc_bslot[avp->av_bslot] = vap;
sc->sc_nbcnvaps++;
}
if (sc->sc_hastsfadd && sc->sc_nbcnvaps > 0) {
/*
* Multple vaps are to transmit beacons and we
* have h/w support for TSF adjusting; enable
* use of staggered beacons.
*/
sc->sc_stagbeacons = 1;
}
ath_txq_init(sc, &avp->av_mcastq, ATH_TXQ_SWQ);
}
ic->ic_opmode = ic_opmode;
if (opmode != IEEE80211_M_WDS) {
sc->sc_nvaps++;
if (opmode == IEEE80211_M_STA)
sc->sc_nstavaps++;
track whether any mesh vaps are present to correctly setup the rx filter when, for example, an ap vap is created first Reviewed by: rpaulo Approved by: re (kib)
2009-07-21 19:01:04 +00:00
if (opmode == IEEE80211_M_MBSS)
sc->sc_nmeshvaps++;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
}
switch (ic_opmode) {
case IEEE80211_M_IBSS:
sc->sc_opmode = HAL_M_IBSS;
break;
case IEEE80211_M_STA:
sc->sc_opmode = HAL_M_STA;
break;
case IEEE80211_M_AHDEMO:
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
if (vap->iv_caps & IEEE80211_C_TDMA) {
sc->sc_tdma = 1;
/* NB: disable tsf adjust */
sc->sc_stagbeacons = 0;
}
/*
* NB: adhoc demo mode is a pseudo mode; to the hal it's
* just ap mode.
*/
/* fall thru... */
#endif
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
case IEEE80211_M_HOSTAP:
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
case IEEE80211_M_MBSS:
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
sc->sc_opmode = HAL_M_HOSTAP;
break;
case IEEE80211_M_MONITOR:
sc->sc_opmode = HAL_M_MONITOR;
break;
default:
/* XXX should not happen */
break;
}
if (sc->sc_hastsfadd) {
/*
* Configure whether or not TSF adjust should be done.
*/
ath_hal_settsfadjust(sc->sc_ah, sc->sc_stagbeacons);
}
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
if (flags & IEEE80211_CLONE_NOBEACONS) {
/*
* Enable s/w beacon miss handling.
*/
sc->sc_swbmiss = 1;
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ATH_UNLOCK(sc);
/* complete setup */
ieee80211_vap_attach(vap, ath_media_change, ieee80211_media_status);
return vap;
bad2:
reclaim_address(sc, mac);
ath_hal_setbssidmask(sc->sc_ah, sc->sc_hwbssidmask);
bad:
free(avp, M_80211_VAP);
ATH_UNLOCK(sc);
return NULL;
}
static void
ath_vap_delete(struct ieee80211vap *vap)
{
struct ieee80211com *ic = vap->iv_ic;
struct ifnet *ifp = ic->ic_ifp;
struct ath_softc *sc = ifp->if_softc;
struct ath_hal *ah = sc->sc_ah;
struct ath_vap *avp = ATH_VAP(vap);
if (ifp->if_drv_flags & IFF_DRV_RUNNING) {
/*
* Quiesce the hardware while we remove the vap. In
* particular we need to reclaim all references to
* the vap state by any frames pending on the tx queues.
*/
ath_hal_intrset(ah, 0); /* disable interrupts */
ath_draintxq(sc); /* stop xmit side */
ath_stoprecv(sc); /* stop recv side */
}
ieee80211_vap_detach(vap);
ATH_LOCK(sc);
/*
* Reclaim beacon state. Note this must be done before
* the vap instance is reclaimed as we may have a reference
* to it in the buffer for the beacon frame.
*/
if (avp->av_bcbuf != NULL) {
if (avp->av_bslot != -1) {
sc->sc_bslot[avp->av_bslot] = NULL;
sc->sc_nbcnvaps--;
}
ath_beacon_return(sc, avp->av_bcbuf);
avp->av_bcbuf = NULL;
if (sc->sc_nbcnvaps == 0) {
sc->sc_stagbeacons = 0;
if (sc->sc_hastsfadd)
ath_hal_settsfadjust(sc->sc_ah, 0);
}
/*
* Reclaim any pending mcast frames for the vap.
*/
ath_tx_draintxq(sc, &avp->av_mcastq);
ATH_TXQ_LOCK_DESTROY(&avp->av_mcastq);
}
/*
* Update bookkeeping.
*/
if (vap->iv_opmode == IEEE80211_M_STA) {
sc->sc_nstavaps--;
if (sc->sc_nstavaps == 0 && sc->sc_swbmiss)
sc->sc_swbmiss = 0;
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
} else if (vap->iv_opmode == IEEE80211_M_HOSTAP ||
vap->iv_opmode == IEEE80211_M_MBSS) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
reclaim_address(sc, vap->iv_myaddr);
ath_hal_setbssidmask(ah, sc->sc_hwbssidmask);
track whether any mesh vaps are present to correctly setup the rx filter when, for example, an ap vap is created first Reviewed by: rpaulo Approved by: re (kib)
2009-07-21 19:01:04 +00:00
if (vap->iv_opmode == IEEE80211_M_MBSS)
sc->sc_nmeshvaps--;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
}
if (vap->iv_opmode != IEEE80211_M_WDS)
sc->sc_nvaps--;
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
/* TDMA operation ceases when the last vap is destroyed */
if (sc->sc_tdma && sc->sc_nvaps == 0) {
sc->sc_tdma = 0;
sc->sc_swbmiss = 0;
}
#endif
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ATH_UNLOCK(sc);
free(avp, M_80211_VAP);
if (ifp->if_drv_flags & IFF_DRV_RUNNING) {
/*
* Restart rx+tx machines if still running (RUNNING will
* be reset if we just destroyed the last vap).
*/
if (ath_startrecv(sc) != 0)
if_printf(ifp, "%s: unable to restart recv logic\n",
__func__);
restart tdma beacons after vap destroy
2009-06-02 21:11:26 +00:00
if (sc->sc_beacons) { /* restart beacons */
#ifdef IEEE80211_SUPPORT_TDMA
if (sc->sc_tdma)
ath_tdma_config(sc, NULL);
else
#endif
ath_beacon_config(sc, NULL);
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_hal_intrset(ah, sc->sc_imask);
}
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
void
ath_suspend(struct ath_softc *sc)
{
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Cleanup power handling and fix suspend/resume: o do not put the chip into full sleep in ath_stop as it gains nothing and causes many parts to hang in ath_detach because we may touch the chip during vap teardown; this may also fix issues with unloading the module o add a note in ath_detach to explain ath_hal_detach puts the chip in low power mode; this is useful to know as it means unloading the module will place a pci device in the lowest possible power state o leave an #ifdef notyet marker for powering down the chip when a device is marked down; we can't do that until we handle all the ways the driver may be entered and touch the chip o fix resume by reloading the h/w key cache as it's been clobbered (for pci) by the socket being powered off; for station mode we directly stop+init the chip and then simulate a beacon miss to get the upper layers sync'd up; for other configs we must brute force stop+start the vaps so they go through the state machine
2008-05-29 00:10:48 +00:00
struct ieee80211com *ic = ifp->if_l2com;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_ANY, "%s: if_flags %x\n",
__func__, ifp->if_flags);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Cleanup power handling and fix suspend/resume: o do not put the chip into full sleep in ath_stop as it gains nothing and causes many parts to hang in ath_detach because we may touch the chip during vap teardown; this may also fix issues with unloading the module o add a note in ath_detach to explain ath_hal_detach puts the chip in low power mode; this is useful to know as it means unloading the module will place a pci device in the lowest possible power state o leave an #ifdef notyet marker for powering down the chip when a device is marked down; we can't do that until we handle all the ways the driver may be entered and touch the chip o fix resume by reloading the h/w key cache as it's been clobbered (for pci) by the socket being powered off; for station mode we directly stop+init the chip and then simulate a beacon miss to get the upper layers sync'd up; for other configs we must brute force stop+start the vaps so they go through the state machine
2008-05-29 00:10:48 +00:00
sc->sc_resume_up = (ifp->if_flags & IFF_UP) != 0;
if (ic->ic_opmode == IEEE80211_M_STA)
ath_stop(ifp);
else
ieee80211_suspend_all(ic);
/*
* NB: don't worry about putting the chip in low power
* mode; pci will power off our socket on suspend and
cardbus -> CardBus
2010-01-03 23:31:58 +00:00
* CardBus detaches the device.
Cleanup power handling and fix suspend/resume: o do not put the chip into full sleep in ath_stop as it gains nothing and causes many parts to hang in ath_detach because we may touch the chip during vap teardown; this may also fix issues with unloading the module o add a note in ath_detach to explain ath_hal_detach puts the chip in low power mode; this is useful to know as it means unloading the module will place a pci device in the lowest possible power state o leave an #ifdef notyet marker for powering down the chip when a device is marked down; we can't do that until we handle all the ways the driver may be entered and touch the chip o fix resume by reloading the h/w key cache as it's been clobbered (for pci) by the socket being powered off; for station mode we directly stop+init the chip and then simulate a beacon miss to get the upper layers sync'd up; for other configs we must brute force stop+start the vaps so they go through the state machine
2008-05-29 00:10:48 +00:00
*/
}
/*
* Reset the key cache since some parts do not reset the
* contents on resume. First we clear all entries, then
* re-load keys that the 802.11 layer assumes are setup
* in h/w.
*/
static void
ath_reset_keycache(struct ath_softc *sc)
{
struct ifnet *ifp = sc->sc_ifp;
struct ieee80211com *ic = ifp->if_l2com;
struct ath_hal *ah = sc->sc_ah;
int i;
for (i = 0; i < sc->sc_keymax; i++)
ath_hal_keyreset(ah, i);
ieee80211_crypto_reload_keys(ic);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
void
ath_resume(struct ath_softc *sc)
{
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Cleanup power handling and fix suspend/resume: o do not put the chip into full sleep in ath_stop as it gains nothing and causes many parts to hang in ath_detach because we may touch the chip during vap teardown; this may also fix issues with unloading the module o add a note in ath_detach to explain ath_hal_detach puts the chip in low power mode; this is useful to know as it means unloading the module will place a pci device in the lowest possible power state o leave an #ifdef notyet marker for powering down the chip when a device is marked down; we can't do that until we handle all the ways the driver may be entered and touch the chip o fix resume by reloading the h/w key cache as it's been clobbered (for pci) by the socket being powered off; for station mode we directly stop+init the chip and then simulate a beacon miss to get the upper layers sync'd up; for other configs we must brute force stop+start the vaps so they go through the state machine
2008-05-29 00:10:48 +00:00
struct ieee80211com *ic = ifp->if_l2com;
struct ath_hal *ah = sc->sc_ah;
HAL_STATUS status;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_ANY, "%s: if_flags %x\n",
__func__, ifp->if_flags);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Cleanup power handling and fix suspend/resume: o do not put the chip into full sleep in ath_stop as it gains nothing and causes many parts to hang in ath_detach because we may touch the chip during vap teardown; this may also fix issues with unloading the module o add a note in ath_detach to explain ath_hal_detach puts the chip in low power mode; this is useful to know as it means unloading the module will place a pci device in the lowest possible power state o leave an #ifdef notyet marker for powering down the chip when a device is marked down; we can't do that until we handle all the ways the driver may be entered and touch the chip o fix resume by reloading the h/w key cache as it's been clobbered (for pci) by the socket being powered off; for station mode we directly stop+init the chip and then simulate a beacon miss to get the upper layers sync'd up; for other configs we must brute force stop+start the vaps so they go through the state machine
2008-05-29 00:10:48 +00:00
/*
* Must reset the chip before we reload the
* keycache as we were powered down on suspend.
*/
on resume ah_curchan may be NULL if no channel change has been done; workaround this by passing net80211's channel as we know it'll never be null Submitted by: trasz
2009-02-10 19:25:11 +00:00
ath_hal_reset(ah, sc->sc_opmode,
sc->sc_curchan != NULL ? sc->sc_curchan : ic->ic_curchan,
AH_FALSE, &status);
Cleanup power handling and fix suspend/resume: o do not put the chip into full sleep in ath_stop as it gains nothing and causes many parts to hang in ath_detach because we may touch the chip during vap teardown; this may also fix issues with unloading the module o add a note in ath_detach to explain ath_hal_detach puts the chip in low power mode; this is useful to know as it means unloading the module will place a pci device in the lowest possible power state o leave an #ifdef notyet marker for powering down the chip when a device is marked down; we can't do that until we handle all the ways the driver may be entered and touch the chip o fix resume by reloading the h/w key cache as it's been clobbered (for pci) by the socket being powered off; for station mode we directly stop+init the chip and then simulate a beacon miss to get the upper layers sync'd up; for other configs we must brute force stop+start the vaps so they go through the state machine
2008-05-29 00:10:48 +00:00
ath_reset_keycache(sc);
if (sc->sc_resume_up) {
if (ic->ic_opmode == IEEE80211_M_STA) {
ath_init(sc);
On resume in sta mode program the beacon timers so when roaming (and the previous ap is no longer in range) the device will deliver bmiss interrupts and trigger the state machine. Also arrange to sync the beacon timers on the next received beacon frame so that when we don't roam we re-synchronize with the ap. Tested by: trasz MFC after: 1 week
2009-08-31 21:25:49 +00:00
/*
* Program the beacon registers using the last rx'd
* beacon frame and enable sync on the next beacon
* we see. This should handle the case where we
* wakeup and find the same AP and also the case where
* we wakeup and need to roam. For the latter we
* should get bmiss events that trigger a roam.
*/
ath_beacon_config(sc, NULL);
sc->sc_syncbeacon = 1;
Cleanup power handling and fix suspend/resume: o do not put the chip into full sleep in ath_stop as it gains nothing and causes many parts to hang in ath_detach because we may touch the chip during vap teardown; this may also fix issues with unloading the module o add a note in ath_detach to explain ath_hal_detach puts the chip in low power mode; this is useful to know as it means unloading the module will place a pci device in the lowest possible power state o leave an #ifdef notyet marker for powering down the chip when a device is marked down; we can't do that until we handle all the ways the driver may be entered and touch the chip o fix resume by reloading the h/w key cache as it's been clobbered (for pci) by the socket being powered off; for station mode we directly stop+init the chip and then simulate a beacon miss to get the upper layers sync'd up; for other configs we must brute force stop+start the vaps so they go through the state machine
2008-05-29 00:10:48 +00:00
} else
ieee80211_resume_all(ic);
o remove bmisshack no longer needed with the BSSID fix in v0.9.5.2 of the hal o add monitor mode support o fix short preamble handling in beacon setup (noop) o correct resume handling
2003-08-02 06:14:15 +00:00
}
restore led state on resume Submitted by: markus
2005-06-02 03:18:55 +00:00
if (sc->sc_softled) {
5416 and later parts mux the gpio outputs; extend the api to include a signal type that's used to select the appropriate mux
2009-02-24 00:12:16 +00:00
ath_hal_gpioCfgOutput(ah, sc->sc_ledpin,
HAL_GPIO_MUX_MAC_NETWORK_LED);
Cleanup power handling and fix suspend/resume: o do not put the chip into full sleep in ath_stop as it gains nothing and causes many parts to hang in ath_detach because we may touch the chip during vap teardown; this may also fix issues with unloading the module o add a note in ath_detach to explain ath_hal_detach puts the chip in low power mode; this is useful to know as it means unloading the module will place a pci device in the lowest possible power state o leave an #ifdef notyet marker for powering down the chip when a device is marked down; we can't do that until we handle all the ways the driver may be entered and touch the chip o fix resume by reloading the h/w key cache as it's been clobbered (for pci) by the socket being powered off; for station mode we directly stop+init the chip and then simulate a beacon miss to get the upper layers sync'd up; for other configs we must brute force stop+start the vaps so they go through the state machine
2008-05-29 00:10:48 +00:00
ath_hal_gpioset(ah, sc->sc_ledpin, !sc->sc_ledon);
restore led state on resume Submitted by: markus
2005-06-02 03:18:55 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
void
ath_shutdown(struct ath_softc *sc)
{
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_ANY, "%s: if_flags %x\n",
__func__, ifp->if_flags);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_stop(ifp);
Cleanup power handling and fix suspend/resume: o do not put the chip into full sleep in ath_stop as it gains nothing and causes many parts to hang in ath_detach because we may touch the chip during vap teardown; this may also fix issues with unloading the module o add a note in ath_detach to explain ath_hal_detach puts the chip in low power mode; this is useful to know as it means unloading the module will place a pci device in the lowest possible power state o leave an #ifdef notyet marker for powering down the chip when a device is marked down; we can't do that until we handle all the ways the driver may be entered and touch the chip o fix resume by reloading the h/w key cache as it's been clobbered (for pci) by the socket being powered off; for station mode we directly stop+init the chip and then simulate a beacon miss to get the upper layers sync'd up; for other configs we must brute force stop+start the vaps so they go through the state machine
2008-05-29 00:10:48 +00:00
/* NB: no point powering down chip as we're about to reboot */
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Interrupt handler. Most of the actual processing is deferred.
*/
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
void
ath_intr(void *arg)
{
struct ath_softc *sc = arg;
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ath_hal *ah = sc->sc_ah;
HAL_INT status;
if (sc->sc_invalid) {
/*
Close a race where ath_intr is installed and may be called before the HAL is setup: use sc_invalid to discard such entries into ath_intr. This can easily happen if the device is assigned a shared IRQ.
2003-08-13 21:29:35 +00:00
* The hardware is not ready/present, don't touch anything.
* Note this can happen early on if the IRQ is shared.
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
*/
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_ANY, "%s: invalid; ignored\n", __func__);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
return;
}
check more quickly (and directly) if an interrupt is pending; this reduces work done in ath_intr when the irq is shared Obtained from: madwifi
2004-04-02 23:49:15 +00:00
if (!ath_hal_intrpend(ah)) /* shared irq, not for us */
return;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
if ((ifp->if_flags & IFF_UP) == 0 ||
(ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) {
HAL_INT status;
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_ANY, "%s: if_flags 0x%x\n",
__func__, ifp->if_flags);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_hal_getisr(ah, &status); /* clear ISR */
ath_hal_intrset(ah, 0); /* disable further intr's */
return;
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Figure out the reason(s) for the interrupt. Note
* that the hal returns a pseudo-ISR that may include
* bits we haven't explicitly enabled so we mask the
* value to insure we only process bits we requested.
*/
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_hal_getisr(ah, &status); /* NB: clears ISR too */
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_INTR, "%s: status 0x%x\n", __func__, status);
o do not filter received frames based on type or length; pass 'em all up to the 802.11 layer if they are at least IEEE80211_MIN_LEN o mask off interrupt status bits that we don't care about so we don't do the wrong thing; this fixes a problem where the beacon miss interrupt status bit is delivered together with other status bits when operating in monitor mode (we would post a beacon miss swi and then do the wrong thing)
2003-09-15 19:41:54 +00:00
status &= sc->sc_imask; /* discard unasked for bits */
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if (status & HAL_INT_FATAL) {
sc->sc_stats.ast_hardware++;
ath_hal_intrset(ah, 0); /* disable intr's until reset */
o handle fatal errors directly instead of via the task queue o temporarily dump some h/w state for diagnosis; this will be removed once some issues are resolved MFC after: 2 weeks
2006-02-15 18:23:03 +00:00
ath_fatal_proc(sc, 0);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
} else {
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (status & HAL_INT_SWBA) {
/*
* Software beacon alert--time to send a beacon.
* Handle beacon transmission directly; deferring
* this is too slow to meet timing constraints
* under load.
*/
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
if (sc->sc_tdma) {
if (sc->sc_tdmaswba == 0) {
struct ieee80211com *ic = ifp->if_l2com;
struct ieee80211vap *vap =
TAILQ_FIRST(&ic->ic_vaps);
ath_tdma_beacon_send(sc, vap);
sc->sc_tdmaswba =
vap->iv_tdma->tdma_bintval;
} else
sc->sc_tdmaswba--;
} else
#endif
Hoist 802.11 encapsulation up into net80211: o call ieee80211_encap in ieee80211_start so frames passed down to drivers are already encapsulated o remove ieee80211_encap calls in drivers o fixup wi so it recreates the 802.3 head it requires from the 802.11 header contents o move fast-frame aggregation from ath to net80211 (conditional on IEEE80211_SUPPORT_SUPERG): - aggregation is now done in ieee80211_start; it is enabled when the packets/sec exceeds ieee80211_ffppsmin (net.wlan.ffppsmin) and frames are held on a staging queue according to ieee80211_ffagemax (net.wlan.ffagemax) to wait for a frame to combine with - drivers must call back to age/flush the staging queue (ath does this on tx done, at swba, and on rx according to the state of the tx queues and/or the contents of the staging queue) - remove fast-frame-related data structures from ath - add ieee80211_ff_node_init and ieee80211_ff_node_cleanup to handle per-node fast-frames state (we reuse 11n tx ampdu state) o change ieee80211_encap calling convention to include an explicit vap so frames coming through a WDS vap are recognized w/o setting M_WDS With these changes any device able to tx/rx 3Kbyte+ frames can use fast-frames. Reviewed by: thompsa, rpaulo, avatar, imp, sephe
2009-03-30 21:53:27 +00:00
{
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
ath_beacon_proc(sc, 0);
Hoist 802.11 encapsulation up into net80211: o call ieee80211_encap in ieee80211_start so frames passed down to drivers are already encapsulated o remove ieee80211_encap calls in drivers o fixup wi so it recreates the 802.3 head it requires from the 802.11 header contents o move fast-frame aggregation from ath to net80211 (conditional on IEEE80211_SUPPORT_SUPERG): - aggregation is now done in ieee80211_start; it is enabled when the packets/sec exceeds ieee80211_ffppsmin (net.wlan.ffppsmin) and frames are held on a staging queue according to ieee80211_ffagemax (net.wlan.ffagemax) to wait for a frame to combine with - drivers must call back to age/flush the staging queue (ath does this on tx done, at swba, and on rx according to the state of the tx queues and/or the contents of the staging queue) - remove fast-frame-related data structures from ath - add ieee80211_ff_node_init and ieee80211_ff_node_cleanup to handle per-node fast-frames state (we reuse 11n tx ampdu state) o change ieee80211_encap calling convention to include an explicit vap so frames coming through a WDS vap are recognized w/o setting M_WDS With these changes any device able to tx/rx 3Kbyte+ frames can use fast-frames. Reviewed by: thompsa, rpaulo, avatar, imp, sephe
2009-03-30 21:53:27 +00:00
#ifdef IEEE80211_SUPPORT_SUPERG
/*
* Schedule the rx taskq in case there's no
* traffic so any frames held on the staging
* queue are aged and potentially flushed.
*/
taskqueue_enqueue(sc->sc_tq, &sc->sc_rxtask);
#endif
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if (status & HAL_INT_RXEOL) {
/*
* NB: the hardware should re-read the link when
* RXE bit is written, but it doesn't work at
* least on older hardware revs.
*/
sc->sc_stats.ast_rxeol++;
sc->sc_rxlink = NULL;
}
if (status & HAL_INT_TXURN) {
sc->sc_stats.ast_txurn++;
/* bump tx trigger level */
ath_hal_updatetxtriglevel(ah, AH_TRUE);
}
if (status & HAL_INT_RX)
use a private task queue thread MFC after: 2 weeks
2006-02-09 21:48:51 +00:00
taskqueue_enqueue(sc->sc_tq, &sc->sc_rxtask);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if (status & HAL_INT_TX)
use a private task queue thread MFC after: 2 weeks
2006-02-09 21:48:51 +00:00
taskqueue_enqueue(sc->sc_tq, &sc->sc_txtask);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if (status & HAL_INT_BMISS) {
sc->sc_stats.ast_bmiss++;
use a private task queue thread MFC after: 2 weeks
2006-02-09 21:48:51 +00:00
taskqueue_enqueue(sc->sc_tq, &sc->sc_bmisstask);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (status & HAL_INT_MIB) {
sc->sc_stats.ast_mib++;
/*
* Disable interrupts until we service the MIB
* interrupt; otherwise it will continue to fire.
*/
ath_hal_intrset(ah, 0);
/*
* Let the hal handle the event. We assume it will
* clear whatever condition caused the interrupt.
*/
move hal statistics to softc; the per-node stats are overkill, they're only used when operating in station mode MFC after: 2 weeks
2006-02-09 21:23:44 +00:00
ath_hal_mibevent(ah, &sc->sc_halstats);
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_hal_intrset(ah, sc->sc_imask);
}
nuke special handling of RXORN interrupt; the hal marks the FATAL bit in the interrupt status when RXORN is hit and the chip requires a reset so our special handling was causing useless resets
2008-11-24 01:31:10 +00:00
if (status & HAL_INT_RXORN) {
/* NB: hal marks HAL_INT_FATAL when RXORN is fatal */
sc->sc_stats.ast_rxorn++;
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
}
static void
ath_fatal_proc(void *arg, int pending)
{
struct ath_softc *sc = arg;
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
o handle fatal errors directly instead of via the task queue o temporarily dump some h/w state for diagnosis; this will be removed once some issues are resolved MFC after: 2 weeks
2006-02-15 18:23:03 +00:00
u_int32_t *state;
u_int32_t len;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
void *sp;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
if_printf(ifp, "hardware error; resetting\n");
o handle fatal errors directly instead of via the task queue o temporarily dump some h/w state for diagnosis; this will be removed once some issues are resolved MFC after: 2 weeks
2006-02-15 18:23:03 +00:00
/*
* Fatal errors are unrecoverable. Typically these
* are caused by DMA errors. Collect h/w state from
* the hal so we can diagnose what's going on.
*/
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
if (ath_hal_getfatalstate(sc->sc_ah, &sp, &len)) {
o handle fatal errors directly instead of via the task queue o temporarily dump some h/w state for diagnosis; this will be removed once some issues are resolved MFC after: 2 weeks
2006-02-15 18:23:03 +00:00
KASSERT(len >= 6*sizeof(u_int32_t), ("len %u bytes", len));
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
state = sp;
o handle fatal errors directly instead of via the task queue o temporarily dump some h/w state for diagnosis; this will be removed once some issues are resolved MFC after: 2 weeks
2006-02-15 18:23:03 +00:00
if_printf(ifp, "0x%08x 0x%08x 0x%08x, 0x%08x 0x%08x 0x%08x\n",
state[0], state[1] , state[2], state[3],
state[4], state[5]);
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_reset(ifp);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static void
ath_bmiss_vap(struct ieee80211vap *vap)
{
/*
* Workaround phantom bmiss interrupts by sanity-checking
* the time of our last rx'd frame. If it is within the
* beacon miss interval then ignore the interrupt. If it's
* truly a bmiss we'll get another interrupt soon and that'll
don't do phantom beacon miss checking for s/w beacon miss handling, this can mistakenly drop events that cause the s/w bmiss timer to never get re-armed
2009-02-10 23:48:29 +00:00
* be dispatched up for processing. Note this applies only
* for h/w beacon miss events.
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
*/
don't do phantom beacon miss checking for s/w beacon miss handling, this can mistakenly drop events that cause the s/w bmiss timer to never get re-armed
2009-02-10 23:48:29 +00:00
if ((vap->iv_flags_ext & IEEE80211_FEXT_SWBMISS) == 0) {
struct ifnet *ifp = vap->iv_ic->ic_ifp;
struct ath_softc *sc = ifp->if_softc;
u_int64_t lastrx = sc->sc_lastrx;
u_int64_t tsf = ath_hal_gettsf64(sc->sc_ah);
u_int bmisstimeout =
vap->iv_bmissthreshold * vap->iv_bss->ni_intval * 1024;
DPRINTF(sc, ATH_DEBUG_BEACON,
"%s: tsf %llu lastrx %lld (%llu) bmiss %u\n",
__func__, (unsigned long long) tsf,
(unsigned long long)(tsf - lastrx),
(unsigned long long) lastrx, bmisstimeout);
if (tsf - lastrx <= bmisstimeout) {
sc->sc_stats.ast_bmiss_phantom++;
return;
}
}
ATH_VAP(vap)->av_bmiss(vap);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
}
some of the 11n parts can hang under certain conditions without necessary workarounds, add code to detect these hangs and distinguish them from other events; note this code is only invoked for anomalous conditions and (at the moment) is a noop because the hang detection code is in a new hal that's coming shortly
2008-11-30 18:34:27 +00:00
static int
ath_hal_gethangstate(struct ath_hal *ah, uint32_t mask, uint32_t *hangs)
{
uint32_t rsize;
void *sp;
Use the now-exposed diag code, rather than a hard-coded magic number.
2011-01-20 04:59:11 +00:00
if (!ath_hal_getdiagstate(ah, HAL_DIAG_CHECK_HANGS, &mask, sizeof(mask), &sp, &rsize))
some of the 11n parts can hang under certain conditions without necessary workarounds, add code to detect these hangs and distinguish them from other events; note this code is only invoked for anomalous conditions and (at the moment) is a noop because the hang detection code is in a new hal that's coming shortly
2008-11-30 18:34:27 +00:00
return 0;
KASSERT(rsize == sizeof(uint32_t), ("resultsize %u", rsize));
*hangs = *(uint32_t *)sp;
return 1;
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void
ath_bmiss_proc(void *arg, int pending)
{
struct ath_softc *sc = arg;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ifnet *ifp = sc->sc_ifp;
some of the 11n parts can hang under certain conditions without necessary workarounds, add code to detect these hangs and distinguish them from other events; note this code is only invoked for anomalous conditions and (at the moment) is a noop because the hang detection code is in a new hal that's coming shortly
2008-11-30 18:34:27 +00:00
uint32_t hangs;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_ANY, "%s: pending %u\n", __func__, pending);
some of the 11n parts can hang under certain conditions without necessary workarounds, add code to detect these hangs and distinguish them from other events; note this code is only invoked for anomalous conditions and (at the moment) is a noop because the hang detection code is in a new hal that's coming shortly
2008-11-30 18:34:27 +00:00
if (ath_hal_gethangstate(sc->sc_ah, 0xff, &hangs) && hangs != 0) {
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
if_printf(ifp, "bb hang detected (0x%x), resetting\n", hangs);
some of the 11n parts can hang under certain conditions without necessary workarounds, add code to detect these hangs and distinguish them from other events; note this code is only invoked for anomalous conditions and (at the moment) is a noop because the hang detection code is in a new hal that's coming shortly
2008-11-30 18:34:27 +00:00
ath_reset(ifp);
} else
ieee80211_beacon_miss(ifp->if_l2com);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/*
* Handle TKIP MIC setup to deal hardware that doesn't do MIC
* calcs together with WME. If necessary disable the crypto
* hardware and mark the 802.11 state so keys will be setup
* with the MIC work done in software.
*/
static void
ath_settkipmic(struct ath_softc *sc)
{
struct ifnet *ifp = sc->sc_ifp;
struct ieee80211com *ic = ifp->if_l2com;
if ((ic->ic_cryptocaps & IEEE80211_CRYPTO_TKIP) && !sc->sc_wmetkipmic) {
if (ic->ic_flags & IEEE80211_F_WME) {
ath_hal_settkipmic(sc->sc_ah, AH_FALSE);
ic->ic_cryptocaps &= ~IEEE80211_CRYPTO_TKIPMIC;
} else {
ath_hal_settkipmic(sc->sc_ah, AH_TRUE);
ic->ic_cryptocaps |= IEEE80211_CRYPTO_TKIPMIC;
}
}
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void
ath_init(void *arg)
{
struct ath_softc *sc = (struct ath_softc *) arg;
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211com *ic = ifp->if_l2com;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ath_hal *ah = sc->sc_ah;
HAL_STATUS status;
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_ANY, "%s: if_flags 0x%x\n",
__func__, ifp->if_flags);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
o convert mutex calls to #defines for portability, etc. o destroy mutex's on detach (was missing)
2003-10-14 22:51:45 +00:00
ATH_LOCK(sc);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/*
* Stop anything previously setup. This is safe
* whether this is the first time through or not.
*/
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_stop_locked(ifp);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/*
* The basic interface to setting the hardware in a good
* state is ``reset''. On return the hardware is known to
* be powered up and with interrupts disabled. This must
* be followed by initialization of the appropriate bits
* and then setup of the interrupt mask.
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_settkipmic(sc);
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 (!ath_hal_reset(ah, sc->sc_opmode, ic->ic_curchan, AH_FALSE, &status)) {
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if_printf(ifp, "unable to reset hardware; hal status %u\n",
status);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ATH_UNLOCK(sc);
return;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_chan_change(sc, ic->ic_curchan);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
o fix setup of sc_diversity; the hal does not give us reliable status after attach, only after a reset o when setting diversity via the sysctl don't update sc_diversity until we know the hal requested worked o while here eliminate sc_hasdiversity and sc_hastpc; just query the hal each time since these are the only places we need to know MFC after: 3 days
2005-07-24 05:11:39 +00:00
/*
* Likewise this is set during reset so update
* state cached in the driver.
*/
sc->sc_diversity = ath_hal_getdiversity(ah);
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
sc->sc_lastlongcal = 0;
sc->sc_resetcal = 1;
sc->sc_lastcalreset = 0;
ANI changes #1 - split out the ANI polling from the RxMonitor hook. The rxmonitor hook is called on each received packet. This can get very, very busy as the tx/rx/chanbusy registers are thus read each time a packet is received. Instead, shuffle out the true per-packet processing which is needed and move the rest of the ANI processing into a periodic event which runs every 100ms by default.
2011-01-21 05:21:00 +00:00
sc->sc_lastani = 0;
sc->sc_lastshortcal = 0;
sc->sc_doresetcal = AH_FALSE;
Update with last year of work.
2004-12-08 17:34:36 +00:00
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/*
* Setup the hardware after reset: the key cache
* is filled as needed and the receive engine is
* set going. Frame transmit is handled entirely
* in the frame output path; there's nothing to do
* here except setup the interrupt mask.
*/
if (ath_startrecv(sc) != 0) {
if_printf(ifp, "unable to start recv logic\n");
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ATH_UNLOCK(sc);
return;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
/*
* Enable interrupts.
*/
sc->sc_imask = HAL_INT_RX | HAL_INT_TX
| HAL_INT_RXEOL | HAL_INT_RXORN
| HAL_INT_FATAL | HAL_INT_GLOBAL;
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Enable MIB interrupts when there are hardware phy counters.
* Note we only do this (at the moment) for station mode.
*/
if (sc->sc_needmib && ic->ic_opmode == IEEE80211_M_STA)
sc->sc_imask |= HAL_INT_MIB;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
ifp->if_drv_flags |= IFF_DRV_RUNNING;
replace if_watchdog w/ private callout; probably can merge this with the calibration work sometime in the future
2009-03-09 23:10:19 +00:00
callout_reset(&sc->sc_wd_ch, hz, ath_watchdog, sc);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_hal_intrset(ah, sc->sc_imask);
ATH_UNLOCK(sc);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
add tx99 hooks MFC after: 2 weeks
2006-02-09 21:28:11 +00:00
#ifdef ATH_TX99_DIAG
if (sc->sc_tx99 != NULL)
sc->sc_tx99->start(sc->sc_tx99);
else
#endif
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ieee80211_start_all(ic); /* start all vap's */
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
static void
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_stop_locked(struct ifnet *ifp)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
{
struct ath_softc *sc = ifp->if_softc;
struct ath_hal *ah = sc->sc_ah;
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_ANY, "%s: invalid %u if_flags 0x%x\n",
__func__, sc->sc_invalid, ifp->if_flags);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
ATH_LOCK_ASSERT(sc);
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
if (ifp->if_drv_flags & IFF_DRV_RUNNING) {
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/*
* Shutdown the hardware and driver:
Update with last year of work.
2004-12-08 17:34:36 +00:00
* reset 802.11 state machine
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
* turn off timers
Update with last year of work.
2004-12-08 17:34:36 +00:00
* disable interrupts
* turn off the radio
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
* clear transmit machinery
* clear receive machinery
* drain and release tx queues
* reclaim beacon resources
* power down hardware
*
* Note that some of this work is not possible if the
* hardware is gone (invalid).
*/
add tx99 hooks MFC after: 2 weeks
2006-02-09 21:28:11 +00:00
#ifdef ATH_TX99_DIAG
if (sc->sc_tx99 != NULL)
sc->sc_tx99->stop(sc->sc_tx99);
#endif
replace if_watchdog w/ private callout; probably can merge this with the calibration work sometime in the future
2009-03-09 23:10:19 +00:00
callout_stop(&sc->sc_wd_ch);
sc->sc_wd_timer = 0;
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (!sc->sc_invalid) {
better led blinking
2005-01-18 19:03:04 +00:00
if (sc->sc_softled) {
callout_stop(&sc->sc_ledtimer);
ath_hal_gpioset(ah, sc->sc_ledpin,
!sc->sc_ledon);
sc->sc_blinking = 0;
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_hal_intrset(ah, 0);
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_draintxq(sc);
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (!sc->sc_invalid) {
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_stoprecv(sc);
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_hal_phydisable(ah);
} else
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_rxlink = NULL;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_beacon_free(sc); /* XXX not needed */
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
}
static void
ath_stop(struct ifnet *ifp)
{
struct ath_softc *sc = ifp->if_softc;
ATH_LOCK(sc);
ath_stop_locked(ifp);
o convert mutex calls to #defines for portability, etc. o destroy mutex's on detach (was missing)
2003-10-14 22:51:45 +00:00
ATH_UNLOCK(sc);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
/*
* Reset the hardware w/o losing operational state. This is
* basically a more efficient way of doing ath_stop, ath_init,
* followed by state transitions to the current 802.11
Update with last year of work.
2004-12-08 17:34:36 +00:00
* operational state. Used to recover from various errors and
* to reset or reload hardware state.
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
*/
Migrate the sysctl related routines (statistics, debugging, etc) out of if_ath.c and into if_ath_sysctl.c .
2011-03-02 16:03:19 +00:00
int
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_reset(struct ifnet *ifp)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
{
Update with last year of work.
2004-12-08 17:34:36 +00:00
struct ath_softc *sc = ifp->if_softc;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211com *ic = ifp->if_l2com;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ath_hal *ah = sc->sc_ah;
HAL_STATUS status;
ath_hal_intrset(ah, 0); /* disable interrupts */
ath_draintxq(sc); /* stop xmit side */
ath_stoprecv(sc); /* stop recv side */
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_settkipmic(sc); /* configure TKIP MIC handling */
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/* NB: indicate channel change so we do a full reset */
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 (!ath_hal_reset(ah, sc->sc_opmode, ic->ic_curchan, AH_TRUE, &status))
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if_printf(ifp, "%s: unable to reset hardware; hal status %u\n",
__func__, status);
o fix setup of sc_diversity; the hal does not give us reliable status after attach, only after a reset o when setting diversity via the sysctl don't update sc_diversity until we know the hal requested worked o while here eliminate sc_hasdiversity and sc_hastpc; just query the hal each time since these are the only places we need to know MFC after: 3 days
2005-07-24 05:11:39 +00:00
sc->sc_diversity = ath_hal_getdiversity(ah);
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
if (ath_startrecv(sc) != 0) /* restart recv */
if_printf(ifp, "%s: unable to start recv logic\n", __func__);
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* We may be doing a reset in response to an ioctl
* that changes the channel so update any state that
* might change as a result.
*/
Add initial support for 900MHz cards like the Ubiquiti SR9: o eliminate assumptions that half/quarter rate channels on exist in 11a o handle frequency mapping between hal and net80211; hal gives us freq's in the range 2422..2437 that we remap MFC after: 1 month
2007-01-15 01:15:57 +00:00
ath_chan_change(sc, ic->ic_curchan);
restart tdma beacons after vap destroy
2009-06-02 21:11:26 +00:00
if (sc->sc_beacons) { /* restart beacons */
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
if (sc->sc_tdma)
ath_tdma_config(sc, NULL);
else
#endif
restart tdma beacons after vap destroy
2009-06-02 21:11:26 +00:00
ath_beacon_config(sc, NULL);
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_hal_intrset(ah, sc->sc_imask);
ath_start(ifp); /* restart xmit */
return 0;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static int
ath_reset_vap(struct ieee80211vap *vap, u_long cmd)
{
intercept IEEE80211_IOC_TXPOWER and service tx power changes immediately
2008-10-27 17:12:41 +00:00
struct ieee80211com *ic = vap->iv_ic;
struct ifnet *ifp = ic->ic_ifp;
struct ath_softc *sc = ifp->if_softc;
struct ath_hal *ah = sc->sc_ah;
switch (cmd) {
case IEEE80211_IOC_TXPOWER:
/*
* If per-packet TPC is enabled, then we have nothing
* to do; otherwise we need to force the global limit.
* All this can happen directly; no need to reset.
*/
if (!ath_hal_gettpc(ah))
ath_hal_settxpowlimit(ah, ic->ic_txpowlimit);
return 0;
}
return ath_reset(ifp);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
}
Migrate the TX path code out of if_ath and into a separate source file. There's two reasons for this: * the raw and non-raw TX path shares a lot of duplicate code which should be refactored; * the 11n-ready chip TX path needs a little reworking.
2011-01-29 11:35:23 +00:00
struct ath_buf *
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
_ath_getbuf_locked(struct ath_softc *sc)
{
struct ath_buf *bf;
ATH_TXBUF_LOCK_ASSERT(sc);
bf = STAILQ_FIRST(&sc->sc_txbuf);
if (bf != NULL && (bf->bf_flags & ATH_BUF_BUSY) == 0)
STAILQ_REMOVE_HEAD(&sc->sc_txbuf, bf_list);
else
bf = NULL;
if (bf == NULL) {
DPRINTF(sc, ATH_DEBUG_XMIT, "%s: %s\n", __func__,
STAILQ_FIRST(&sc->sc_txbuf) == NULL ?
"out of xmit buffers" : "xmit buffer busy");
}
return bf;
}
Migrate the TX path code out of if_ath and into a separate source file. There's two reasons for this: * the raw and non-raw TX path shares a lot of duplicate code which should be refactored; * the 11n-ready chip TX path needs a little reworking.
2011-01-29 11:35:23 +00:00
struct ath_buf *
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
ath_getbuf(struct ath_softc *sc)
{
struct ath_buf *bf;
ATH_TXBUF_LOCK(sc);
bf = _ath_getbuf_locked(sc);
if (bf == NULL) {
struct ifnet *ifp = sc->sc_ifp;
DPRINTF(sc, ATH_DEBUG_XMIT, "%s: stop queue\n", __func__);
sc->sc_stats.ast_tx_qstop++;
ifp->if_drv_flags |= IFF_DRV_OACTIVE;
}
ATH_TXBUF_UNLOCK(sc);
return bf;
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void
ath_start(struct ifnet *ifp)
{
struct ath_softc *sc = ifp->if_softc;
struct ieee80211_node *ni;
struct ath_buf *bf;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
struct mbuf *m, *next;
ath_bufhead frags;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0 || sc->sc_invalid)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
return;
for (;;) {
/*
* Grab a TX buffer and associated resources.
*/
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
bf = ath_getbuf(sc);
if (bf == NULL)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
break;
Add missing bits that made bg scanning lame: o update ic_lastdata to reflect time of last outbound frame o outbound traffic must preempt/cancel bg scanning to avoid delays This stuff was somehow missed in the initial import. Reviewed by: thompsa, avatar, sephe (earlier version) Approved by: re (blanket wireless)
2007-09-05 23:00:27 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
IFQ_DEQUEUE(&ifp->if_snd, m);
if (m == NULL) {
ATH_TXBUF_LOCK(sc);
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
STAILQ_INSERT_HEAD(&sc->sc_txbuf, bf, bf_list);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ATH_TXBUF_UNLOCK(sc);
break;
}
ni = (struct ieee80211_node *) m->m_pkthdr.rcvif;
/*
* Check for fragmentation. If this frame
* has been broken up verify we have enough
* buffers to send all the fragments so all
* go out or none...
*/
Hoist 802.11 encapsulation up into net80211: o call ieee80211_encap in ieee80211_start so frames passed down to drivers are already encapsulated o remove ieee80211_encap calls in drivers o fixup wi so it recreates the 802.3 head it requires from the 802.11 header contents o move fast-frame aggregation from ath to net80211 (conditional on IEEE80211_SUPPORT_SUPERG): - aggregation is now done in ieee80211_start; it is enabled when the packets/sec exceeds ieee80211_ffppsmin (net.wlan.ffppsmin) and frames are held on a staging queue according to ieee80211_ffagemax (net.wlan.ffagemax) to wait for a frame to combine with - drivers must call back to age/flush the staging queue (ath does this on tx done, at swba, and on rx according to the state of the tx queues and/or the contents of the staging queue) - remove fast-frame-related data structures from ath - add ieee80211_ff_node_init and ieee80211_ff_node_cleanup to handle per-node fast-frames state (we reuse 11n tx ampdu state) o change ieee80211_encap calling convention to include an explicit vap so frames coming through a WDS vap are recognized w/o setting M_WDS With these changes any device able to tx/rx 3Kbyte+ frames can use fast-frames. Reviewed by: thompsa, rpaulo, avatar, imp, sephe
2009-03-30 21:53:27 +00:00
STAILQ_INIT(&frags);
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
if ((m->m_flags & M_FRAG) &&
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
!ath_txfrag_setup(sc, &frags, m, ni)) {
DPRINTF(sc, ATH_DEBUG_XMIT,
"%s: out of txfrag buffers\n", __func__);
add driver stat to count tx drops due to insufficient frag buffers
2009-02-13 05:30:22 +00:00
sc->sc_stats.ast_tx_nofrag++;
count frag tx failures as an ifnet error
2009-06-02 21:13:57 +00:00
ifp->if_oerrors++;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_freetx(m);
goto bad;
}
Hoist 802.11 encapsulation up into net80211: o call ieee80211_encap in ieee80211_start so frames passed down to drivers are already encapsulated o remove ieee80211_encap calls in drivers o fixup wi so it recreates the 802.3 head it requires from the 802.11 header contents o move fast-frame aggregation from ath to net80211 (conditional on IEEE80211_SUPPORT_SUPERG): - aggregation is now done in ieee80211_start; it is enabled when the packets/sec exceeds ieee80211_ffppsmin (net.wlan.ffppsmin) and frames are held on a staging queue according to ieee80211_ffagemax (net.wlan.ffagemax) to wait for a frame to combine with - drivers must call back to age/flush the staging queue (ath does this on tx done, at swba, and on rx according to the state of the tx queues and/or the contents of the staging queue) - remove fast-frame-related data structures from ath - add ieee80211_ff_node_init and ieee80211_ff_node_cleanup to handle per-node fast-frames state (we reuse 11n tx ampdu state) o change ieee80211_encap calling convention to include an explicit vap so frames coming through a WDS vap are recognized w/o setting M_WDS With these changes any device able to tx/rx 3Kbyte+ frames can use fast-frames. Reviewed by: thompsa, rpaulo, avatar, imp, sephe
2009-03-30 21:53:27 +00:00
ifp->if_opackets++;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
nextfrag:
/*
* Pass the frame to the h/w for transmission.
* Fragmented frames have each frag chained together
* with m_nextpkt. We know there are sufficient ath_buf's
* to send all the frags because of work done by
* ath_txfrag_setup. We leave m_nextpkt set while
* calling ath_tx_start so it can use it to extend the
* the tx duration to cover the subsequent frag and
* so it can reclaim all the mbufs in case of an error;
* ath_tx_start clears m_nextpkt once it commits to
* handing the frame to the hardware.
*/
next = m->m_nextpkt;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if (ath_tx_start(sc, ni, bf, m)) {
bad:
Update with last year of work.
2004-12-08 17:34:36 +00:00
ifp->if_oerrors++;
reclaim:
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
bf->bf_m = NULL;
bf->bf_node = NULL;
o convert mutex calls to #defines for portability, etc. o destroy mutex's on detach (was missing)
2003-10-14 22:51:45 +00:00
ATH_TXBUF_LOCK(sc);
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
STAILQ_INSERT_HEAD(&sc->sc_txbuf, bf, bf_list);
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
ath_txfrag_cleanup(sc, &frags, ni);
o convert mutex calls to #defines for portability, etc. o destroy mutex's on detach (was missing)
2003-10-14 22:51:45 +00:00
ATH_TXBUF_UNLOCK(sc);
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (ni != NULL)
ieee80211_free_node(ni);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
continue;
}
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
if (next != NULL) {
/*
* Beware of state changing between frags.
* XXX check sta power-save state?
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (ni->ni_vap->iv_state != IEEE80211_S_RUN) {
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
DPRINTF(sc, ATH_DEBUG_XMIT,
"%s: flush fragmented packet, state %s\n",
__func__,
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ieee80211_state_name[ni->ni_vap->iv_state]);
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
ath_freetx(next);
goto reclaim;
}
m = next;
bf = STAILQ_FIRST(&frags);
KASSERT(bf != NULL, ("no buf for txfrag"));
STAILQ_REMOVE_HEAD(&frags, bf_list);
goto nextfrag;
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
replace if_watchdog w/ private callout; probably can merge this with the calibration work sometime in the future
2009-03-09 23:10:19 +00:00
sc->sc_wd_timer = 5;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
}
static int
ath_media_change(struct ifnet *ifp)
{
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
int error = ieee80211_media_change(ifp);
/* NB: only the fixed rate can change and that doesn't need a reset */
return (error == ENETRESET ? 0 : error);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Block/unblock tx+rx processing while a key change is done.
* We assume the caller serializes key management operations
* so we only need to worry about synchronization with other
* uses that originate in the driver.
*/
static void
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_key_update_begin(struct ieee80211vap *vap)
Update with last year of work.
2004-12-08 17:34:36 +00:00
{
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ifnet *ifp = vap->iv_ic->ic_ifp;
Update with last year of work.
2004-12-08 17:34:36 +00:00
struct ath_softc *sc = ifp->if_softc;
DPRINTF(sc, ATH_DEBUG_KEYCACHE, "%s:\n", __func__);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
taskqueue_block(sc->sc_tq);
Update with last year of work.
2004-12-08 17:34:36 +00:00
IF_LOCK(&ifp->if_snd); /* NB: doesn't block mgmt frames */
}
static void
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_key_update_end(struct ieee80211vap *vap)
Update with last year of work.
2004-12-08 17:34:36 +00:00
{
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ifnet *ifp = vap->iv_ic->ic_ifp;
Update with last year of work.
2004-12-08 17:34:36 +00:00
struct ath_softc *sc = ifp->if_softc;
DPRINTF(sc, ATH_DEBUG_KEYCACHE, "%s:\n", __func__);
IF_UNLOCK(&ifp->if_snd);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
taskqueue_unblock(sc->sc_tq);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
o consolidate rx filter calculations in one place o enable beacon reception when operating in adhoc mode so the 802.11 layer can use them to create nodes for peers
2003-10-17 21:58:39 +00:00
/*
* Calculate the receive filter according to the
* operating mode and state:
*
* o always accept unicast, broadcast, and multicast traffic
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
* o accept PHY error frames when hardware doesn't have MIB support
prepare for a new hal
2008-10-27 18:30:33 +00:00
* to count and we need them for ANI (sta mode only until recently)
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
* and we are not scanning (ANI is disabled)
prepare for a new hal
2008-10-27 18:30:33 +00:00
* NB: older hal's add rx filter bits out of sight and we need to
* blindly preserve them
o consolidate rx filter calculations in one place o enable beacon reception when operating in adhoc mode so the 802.11 layer can use them to create nodes for peers
2003-10-17 21:58:39 +00:00
* o probe request frames are accepted only when operating in
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
* hostap, adhoc, mesh, or monitor modes
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
* o enable promiscuous mode
* - when in monitor mode
* - if interface marked PROMISC (assumes bridge setting is filtered)
o consolidate rx filter calculations in one place o enable beacon reception when operating in adhoc mode so the 802.11 layer can use them to create nodes for peers
2003-10-17 21:58:39 +00:00
* o accept beacons:
* - when operating in station mode for collecting rssi data when
* the station is otherwise quiet, or
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
* - when operating in adhoc mode so the 802.11 layer creates
* node table entries for peers,
o consolidate rx filter calculations in one place o enable beacon reception when operating in adhoc mode so the 802.11 layer can use them to create nodes for peers
2003-10-17 21:58:39 +00:00
* - when scanning
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
* - when doing s/w beacon miss (e.g. for ap+sta)
* - when operating in ap mode in 11g to detect overlapping bss that
* require protection
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
* - when operating in mesh mode to detect neighbors
enable rx of control frames when in monitor mode Submitted by: Andrea Bittau <a.bittau@cs.ucl.ac.uk> MFC after: 1 week
2006-06-26 04:31:36 +00:00
* o accept control frames:
* - when in monitor mode
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
* XXX HT protection for 11n
o consolidate rx filter calculations in one place o enable beacon reception when operating in adhoc mode so the 802.11 layer can use them to create nodes for peers
2003-10-17 21:58:39 +00:00
*/
static u_int32_t
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
ath_calcrxfilter(struct ath_softc *sc)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
{
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211com *ic = ifp->if_l2com;
o consolidate rx filter calculations in one place o enable beacon reception when operating in adhoc mode so the 802.11 layer can use them to create nodes for peers
2003-10-17 21:58:39 +00:00
u_int32_t rfilt;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
rfilt = HAL_RX_FILTER_UCAST | HAL_RX_FILTER_BCAST | HAL_RX_FILTER_MCAST;
prepare for a new hal
2008-10-27 18:30:33 +00:00
if (!sc->sc_needmib && !sc->sc_scanning)
rfilt |= HAL_RX_FILTER_PHYERR;
Explicitly enable probe request frame reception when not in station mode; this is needed for the 5212 which a separate filter bit for these frames. Submitted by: Stephane Laroche <stephane.laroche@colubris.com>
2003-09-01 03:12:19 +00:00
if (ic->ic_opmode != IEEE80211_M_STA)
rfilt |= HAL_RX_FILTER_PROBEREQ;
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
/* XXX ic->ic_monvaps != 0? */
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (ic->ic_opmode == IEEE80211_M_MONITOR || (ifp->if_flags & IFF_PROMISC))
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
rfilt |= HAL_RX_FILTER_PROM;
Maintain a history of data associated with received frames and use this to calculate smoothed signal quality data for each node. o add a 16-deep history buffer to each driver-private node storage that holds rssi and antenna info for received frames o override the default per-node "get rssi" method to return an average rssi value based on samples collected over the last second o enable beacon reception so even idle systems maintain a running history of signal quality This data may also be useful for improving the rate control algorithm. Based on work by Tom Marshall <tommy@home.tig-grr.com> for MADWIFI.
2003-09-15 22:34:46 +00:00
if (ic->ic_opmode == IEEE80211_M_STA ||
use the ic's opmode instead of our hal equivalent to check for adhoc mode; they are always the same
2008-10-27 17:14:39 +00:00
ic->ic_opmode == IEEE80211_M_IBSS ||
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
sc->sc_swbmiss || sc->sc_scanning)
rfilt |= HAL_RX_FILTER_BEACON;
/*
* NB: We don't recalculate the rx filter when
* ic_protmode changes; otherwise we could do
* this only when ic_protmode != NONE.
*/
if (ic->ic_opmode == IEEE80211_M_HOSTAP &&
IEEE80211_IS_CHAN_ANYG(ic->ic_curchan))
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
rfilt |= HAL_RX_FILTER_BEACON;
Add in the hardware PS-POLL frame reception setting, but leave it disabled by default. Adventourous souls with an AR9220/AR9280 or later and who have a device that sends PS-POLL frames may wish to try tinkering with this option and get back to me.
2011-03-26 10:52:37 +00:00
#if 0
/*
Fix typo.
2011-03-27 10:35:39 +00:00
* Enable hardware PS-POLL RX only for hostap mode;
Add in the hardware PS-POLL frame reception setting, but leave it disabled by default. Adventourous souls with an AR9220/AR9280 or later and who have a device that sends PS-POLL frames may wish to try tinkering with this option and get back to me.
2011-03-26 10:52:37 +00:00
* STA mode sends PS-POLL frames but never
Fix typo.
2011-03-27 10:35:39 +00:00
* receives them.
Add in the hardware PS-POLL frame reception setting, but leave it disabled by default. Adventourous souls with an AR9220/AR9280 or later and who have a device that sends PS-POLL frames may wish to try tinkering with this option and get back to me.
2011-03-26 10:52:37 +00:00
*/
if (ath_hal_getcapability(ah, HAL_CAP_HAS_PSPOLL,
0, NULL) == HAL_OK &&
ic->ic_opmode == IEEE80211_M_HOSTAP)
rfilt |= HAL_RX_FILTER_PSPOLL;
#endif
track whether any mesh vaps are present to correctly setup the rx filter when, for example, an ap vap is created first Reviewed by: rpaulo Approved by: re (kib)
2009-07-21 19:01:04 +00:00
if (sc->sc_nmeshvaps) {
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
rfilt |= HAL_RX_FILTER_BEACON;
if (sc->sc_hasbmatch)
rfilt |= HAL_RX_FILTER_BSSID;
else
rfilt |= HAL_RX_FILTER_PROM;
}
enable rx of control frames when in monitor mode Submitted by: Andrea Bittau <a.bittau@cs.ucl.ac.uk> MFC after: 1 week
2006-06-26 04:31:36 +00:00
if (ic->ic_opmode == IEEE80211_M_MONITOR)
rfilt |= HAL_RX_FILTER_CONTROL;
Add in the hardware PS-POLL frame reception setting, but leave it disabled by default. Adventourous souls with an AR9220/AR9280 or later and who have a device that sends PS-POLL frames may wish to try tinkering with this option and get back to me.
2011-03-26 10:52:37 +00:00
/*
* Enable RX of compressed BAR frames only when doing
* 802.11n. Required for A-MPDU.
*/
Enable AMPDU reorder processing and receiving BAR frames when doing 802.11n. Obtained from: rpaulo@
2011-01-31 22:44:28 +00:00
if (IEEE80211_IS_CHAN_HT(ic->ic_curchan))
rfilt |= HAL_RX_FILTER_COMPBAR;
Add in the hardware PS-POLL frame reception setting, but leave it disabled by default. Adventourous souls with an AR9220/AR9280 or later and who have a device that sends PS-POLL frames may wish to try tinkering with this option and get back to me.
2011-03-26 10:52:37 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
DPRINTF(sc, ATH_DEBUG_MODE, "%s: RX filter 0x%x, %s if_flags 0x%x\n",
__func__, rfilt, ieee80211_opmode_name[ic->ic_opmode], ifp->if_flags);
o consolidate rx filter calculations in one place o enable beacon reception when operating in adhoc mode so the 802.11 layer can use them to create nodes for peers
2003-10-17 21:58:39 +00:00
return rfilt;
}
static void
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_update_promisc(struct ifnet *ifp)
o consolidate rx filter calculations in one place o enable beacon reception when operating in adhoc mode so the 802.11 layer can use them to create nodes for peers
2003-10-17 21:58:39 +00:00
{
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ath_softc *sc = ifp->if_softc;
u_int32_t rfilt;
o consolidate rx filter calculations in one place o enable beacon reception when operating in adhoc mode so the 802.11 layer can use them to create nodes for peers
2003-10-17 21:58:39 +00:00
/* configure rx filter */
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
rfilt = ath_calcrxfilter(sc);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_hal_setrxfilter(sc->sc_ah, rfilt);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
DPRINTF(sc, ATH_DEBUG_MODE, "%s: RX filter 0x%x\n", __func__, rfilt);
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static void
ath_update_mcast(struct ifnet *ifp)
{
struct ath_softc *sc = ifp->if_softc;
u_int32_t mfilt[2];
o consolidate rx filter calculations in one place o enable beacon reception when operating in adhoc mode so the 802.11 layer can use them to create nodes for peers
2003-10-17 21:58:39 +00:00
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/* calculate and install multicast filter */
if ((ifp->if_flags & IFF_ALLMULTI) == 0) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ifmultiaddr *ifma;
/*
* Merge multicast addresses to form the hardware filter.
*/
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
mfilt[0] = mfilt[1] = 0;
Use if_maddr_rlock()/if_maddr_runlock() rather than IF_ADDR_LOCK()/ IF_ADDR_UNLOCK() across network device drivers when accessing the per-interface multicast address list, if_multiaddrs. This will allow us to change the locking strategy without affecting our driver programming interface or binary interface. For two wireless drivers, remove unnecessary locking, since they don't actually access the multicast address list. Approved by: re (kib) MFC after: 6 weeks
2009-06-26 11:45:06 +00:00
if_maddr_rlock(ifp); /* XXX need some fiddling to remove? */
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) {
caddr_t dl;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
u_int32_t val;
u_int8_t pos;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/* calculate XOR of eight 6bit values */
dl = LLADDR((struct sockaddr_dl *) ifma->ifma_addr);
val = LE_READ_4(dl + 0);
pos = (val >> 18) ^ (val >> 12) ^ (val >> 6) ^ val;
val = LE_READ_4(dl + 3);
pos ^= (val >> 18) ^ (val >> 12) ^ (val >> 6) ^ val;
pos &= 0x3f;
mfilt[pos / 32] |= (1 << (pos % 32));
}
Use if_maddr_rlock()/if_maddr_runlock() rather than IF_ADDR_LOCK()/ IF_ADDR_UNLOCK() across network device drivers when accessing the per-interface multicast address list, if_multiaddrs. This will allow us to change the locking strategy without affecting our driver programming interface or binary interface. For two wireless drivers, remove unnecessary locking, since they don't actually access the multicast address list. Approved by: re (kib) MFC after: 6 weeks
2009-06-26 11:45:06 +00:00
if_maddr_runlock(ifp);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
} else
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
mfilt[0] = mfilt[1] = ~0;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_hal_setmcastfilter(sc->sc_ah, mfilt[0], mfilt[1]);
DPRINTF(sc, ATH_DEBUG_MODE, "%s: MC filter %08x:%08x\n",
__func__, mfilt[0], mfilt[1]);
}
static void
ath_mode_init(struct ath_softc *sc)
{
struct ifnet *ifp = sc->sc_ifp;
struct ath_hal *ah = sc->sc_ah;
u_int32_t rfilt;
/* configure rx filter */
rfilt = ath_calcrxfilter(sc);
ath_hal_setrxfilter(ah, rfilt);
/* configure operational mode */
ath_hal_setopmode(ah);
Eliminate ic_myaddr so changing the mac address of a device works correctly: o remove ic_myaddr from ieee80211com o change ieee80211_ifattach to take the mac address of the physical device and use that to setup the lladdr. o replace all references to ic_myaddr in drivers by IF_LLADDR o related cleanups (e.g. kill dead code) PR: kern/133178 Reviewed by: thompsa, rpaulo
2009-03-29 17:59:14 +00:00
/* handle any link-level address change */
ath_hal_setmac(ah, IF_LLADDR(ifp));
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/* calculate and install multicast filter */
ath_update_mcast(ifp);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Set the slot time based on the current setting.
*/
static void
ath_setslottime(struct ath_softc *sc)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
{
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211com *ic = sc->sc_ifp->if_l2com;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ath_hal *ah = sc->sc_ah;
Add half/quarter rate 11a channel support: o change handling of regdomain-related mib knobs so they can be set post-attach: regdomain, countrycode, outdoor, and xchanmode; the hal will not permit changing the regdomain but we expose it for now o on regdomain/countrycode change recalculate the channel list and push it to the net80211 layer (NB: looks to need more tweaking) o setup rate tables for half/quarter rate channels o honor half/quarter rate channel configs when changing channels o honor half/quarter rate channel configs when setting the slot time o use hack/nonstandard channel numbering scheme for the public safety band to avoid overlapping 2.4G channels on dual-band cards o remove setup of ic_sup_rates; the net80211 layer can do this for us and it simplifies handling of half/quarter rate channels Tested only in Public Safety Band with cards that have RF5112.
2006-12-27 19:07:09 +00:00
u_int usec;
Update with last year of work.
2004-12-08 17:34:36 +00:00
Add initial support for 900MHz cards like the Ubiquiti SR9: o eliminate assumptions that half/quarter rate channels on exist in 11a o handle frequency mapping between hal and net80211; hal gives us freq's in the range 2422..2437 that we remap MFC after: 1 month
2007-01-15 01:15:57 +00:00
if (IEEE80211_IS_CHAN_HALF(ic->ic_curchan))
usec = 13;
else if (IEEE80211_IS_CHAN_QUARTER(ic->ic_curchan))
usec = 21;
else if (IEEE80211_IS_CHAN_ANYG(ic->ic_curchan)) {
/* honor short/long slot time only in 11g */
/* XXX shouldn't honor on pure g or turbo g channel */
if (ic->ic_flags & IEEE80211_F_SHSLOT)
Add half/quarter rate 11a channel support: o change handling of regdomain-related mib knobs so they can be set post-attach: regdomain, countrycode, outdoor, and xchanmode; the hal will not permit changing the regdomain but we expose it for now o on regdomain/countrycode change recalculate the channel list and push it to the net80211 layer (NB: looks to need more tweaking) o setup rate tables for half/quarter rate channels o honor half/quarter rate channel configs when changing channels o honor half/quarter rate channel configs when setting the slot time o use hack/nonstandard channel numbering scheme for the public safety band to avoid overlapping 2.4G channels on dual-band cards o remove setup of ic_sup_rates; the net80211 layer can do this for us and it simplifies handling of half/quarter rate channels Tested only in Public Safety Band with cards that have RF5112.
2006-12-27 19:07:09 +00:00
usec = HAL_SLOT_TIME_9;
Add initial support for 900MHz cards like the Ubiquiti SR9: o eliminate assumptions that half/quarter rate channels on exist in 11a o handle frequency mapping between hal and net80211; hal gives us freq's in the range 2422..2437 that we remap MFC after: 1 month
2007-01-15 01:15:57 +00:00
else
usec = HAL_SLOT_TIME_20;
} else
Add half/quarter rate 11a channel support: o change handling of regdomain-related mib knobs so they can be set post-attach: regdomain, countrycode, outdoor, and xchanmode; the hal will not permit changing the regdomain but we expose it for now o on regdomain/countrycode change recalculate the channel list and push it to the net80211 layer (NB: looks to need more tweaking) o setup rate tables for half/quarter rate channels o honor half/quarter rate channel configs when changing channels o honor half/quarter rate channel configs when setting the slot time o use hack/nonstandard channel numbering scheme for the public safety band to avoid overlapping 2.4G channels on dual-band cards o remove setup of ic_sup_rates; the net80211 layer can do this for us and it simplifies handling of half/quarter rate channels Tested only in Public Safety Band with cards that have RF5112.
2006-12-27 19:07:09 +00:00
usec = HAL_SLOT_TIME_9;
DPRINTF(sc, ATH_DEBUG_RESET,
"%s: chan %u MHz flags 0x%x %s slot, %u usec\n",
__func__, ic->ic_curchan->ic_freq, ic->ic_curchan->ic_flags,
ic->ic_flags & IEEE80211_F_SHSLOT ? "short" : "long", usec);
ath_hal_setslottime(ah, usec);
Update with last year of work.
2004-12-08 17:34:36 +00:00
sc->sc_updateslot = OK;
}
/*
* Callback from the 802.11 layer to update the
* slot time based on the current setting.
*/
static void
ath_updateslot(struct ifnet *ifp)
{
struct ath_softc *sc = ifp->if_softc;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211com *ic = ifp->if_l2com;
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* When not coordinating the BSS, change the hardware
* immediately. For other operation we defer the change
* until beacon updates have propagated to the stations.
*/
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
if (ic->ic_opmode == IEEE80211_M_HOSTAP ||
ic->ic_opmode == IEEE80211_M_MBSS)
Update with last year of work.
2004-12-08 17:34:36 +00:00
sc->sc_updateslot = UPDATE;
else
ath_setslottime(sc);
}
setup the beacon xmit queue to not interrupt; we don't use them and they make the led's flash unnecessarily in adhoc mode
2005-01-18 19:10:17 +00:00
/*
* Setup a h/w transmit queue for beacons.
*/
static int
ath_beaconq_setup(struct ath_hal *ah)
{
HAL_TXQ_INFO qi;
memset(&qi, 0, sizeof(qi));
qi.tqi_aifs = HAL_TXQ_USEDEFAULT;
qi.tqi_cwmin = HAL_TXQ_USEDEFAULT;
qi.tqi_cwmax = HAL_TXQ_USEDEFAULT;
Set the correct IFS parameters for the beacon tx queue when operating in ap and adhoc modes.
2005-06-07 00:12:40 +00:00
/* NB: for dynamic turbo, don't enable any other interrupts */
Update for rev 0.9.16.16 hal: o add dfs+radar hooks; DFS is presently disabled in the hal o channel and mode handling changes o various api changes o be more aggressive about iq calibration settling so ap mode operation is better immediately after startup o rfkill/rfsilent sysctl support o tpc ack/cts sysctl support MFC after: 2 weeks
2006-02-10 19:07:08 +00:00
qi.tqi_qflags = HAL_TXQ_TXDESCINT_ENABLE;
setup the beacon xmit queue to not interrupt; we don't use them and they make the led's flash unnecessarily in adhoc mode
2005-01-18 19:10:17 +00:00
return ath_hal_setuptxqueue(ah, HAL_TX_QUEUE_BEACON, &qi);
}
Set the correct IFS parameters for the beacon tx queue when operating in ap and adhoc modes.
2005-06-07 00:12:40 +00:00
/*
* Setup the transmit queue parameters for the beacon queue.
*/
static int
ath_beaconq_config(struct ath_softc *sc)
{
#define ATH_EXPONENT_TO_VALUE(v) ((1<<(v))-1)
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211com *ic = sc->sc_ifp->if_l2com;
Set the correct IFS parameters for the beacon tx queue when operating in ap and adhoc modes.
2005-06-07 00:12:40 +00:00
struct ath_hal *ah = sc->sc_ah;
HAL_TXQ_INFO qi;
ath_hal_gettxqueueprops(ah, sc->sc_bhalq, &qi);
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
if (ic->ic_opmode == IEEE80211_M_HOSTAP ||
ic->ic_opmode == IEEE80211_M_MBSS) {
Set the correct IFS parameters for the beacon tx queue when operating in ap and adhoc modes.
2005-06-07 00:12:40 +00:00
/*
* Always burst out beacon and CAB traffic.
*/
qi.tqi_aifs = ATH_BEACON_AIFS_DEFAULT;
qi.tqi_cwmin = ATH_BEACON_CWMIN_DEFAULT;
qi.tqi_cwmax = ATH_BEACON_CWMAX_DEFAULT;
} else {
struct wmeParams *wmep =
&ic->ic_wme.wme_chanParams.cap_wmeParams[WME_AC_BE];
/*
* Adhoc mode; important thing is to use 2x cwmin.
*/
qi.tqi_aifs = wmep->wmep_aifsn;
qi.tqi_cwmin = 2*ATH_EXPONENT_TO_VALUE(wmep->wmep_logcwmin);
qi.tqi_cwmax = ATH_EXPONENT_TO_VALUE(wmep->wmep_logcwmax);
}
if (!ath_hal_settxqueueprops(ah, sc->sc_bhalq, &qi)) {
device_printf(sc->sc_dev, "unable to update parameters for "
"beacon hardware queue!\n");
return 0;
} else {
ath_hal_resettxqueue(ah, sc->sc_bhalq); /* push to h/w */
return 1;
}
#undef ATH_EXPONENT_TO_VALUE
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Allocate and setup an initial beacon frame.
*/
static int
ath_beacon_alloc(struct ath_softc *sc, struct ieee80211_node *ni)
{
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211vap *vap = ni->ni_vap;
struct ath_vap *avp = ATH_VAP(vap);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ath_buf *bf;
struct mbuf *m;
Update with last year of work.
2004-12-08 17:34:36 +00:00
int error;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
bf = avp->av_bcbuf;
if (bf->bf_m != NULL) {
bus_dmamap_unload(sc->sc_dmat, bf->bf_dmamap);
m_freem(bf->bf_m);
bf->bf_m = NULL;
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (bf->bf_node != NULL) {
ieee80211_free_node(bf->bf_node);
bf->bf_node = NULL;
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/*
* NB: the beacon data buffer must be 32-bit aligned;
* we assume the mbuf routines will return us something
* with this alignment (perhaps should assert).
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
m = ieee80211_beacon_alloc(ni, &avp->av_boff);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if (m == NULL) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
device_printf(sc->sc_dev, "%s: cannot get mbuf\n", __func__);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_stats.ast_be_nombuf++;
return ENOMEM;
}
switch to use bus_dmamap_load_mbuf_sg
2005-01-24 19:51:59 +00:00
error = bus_dmamap_load_mbuf_sg(sc->sc_dmat, bf->bf_dmamap, m,
bf->bf_segs, &bf->bf_nseg,
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
BUS_DMA_NOWAIT);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (error != 0) {
device_printf(sc->sc_dev,
"%s: cannot map mbuf, bus_dmamap_load_mbuf_sg returns %d\n",
__func__, error);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
m_freem(m);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
return error;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/*
* Calculate a TSF adjustment factor required for staggered
* beacons. Note that we assume the format of the beacon
* frame leaves the tstamp field immediately following the
* header.
*/
if (sc->sc_stagbeacons && avp->av_bslot > 0) {
uint64_t tsfadjust;
struct ieee80211_frame *wh;
/*
* The beacon interval is in TU's; the TSF is in usecs.
* We figure out how many TU's to add to align the timestamp
* then convert to TSF units and handle byte swapping before
* inserting it in the frame. The hardware will then add this
* each time a beacon frame is sent. Note that we align vap's
* 1..N and leave vap 0 untouched. This means vap 0 has a
* timestamp in one beacon interval while the others get a
* timstamp aligned to the next interval.
*/
tsfadjust = ni->ni_intval *
(ATH_BCBUF - avp->av_bslot) / ATH_BCBUF;
tsfadjust = htole64(tsfadjust << 10); /* TU -> TSF */
DPRINTF(sc, ATH_DEBUG_BEACON,
"%s: %s beacons bslot %d intval %u tsfadjust %llu\n",
__func__, sc->sc_stagbeacons ? "stagger" : "burst",
fix compilation on 64-bit platform w/ ATH_DEBUG
2008-09-21 00:30:33 +00:00
avp->av_bslot, ni->ni_intval,
(long long unsigned) le64toh(tsfadjust));
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
wh = mtod(m, struct ieee80211_frame *);
memcpy(&wh[1], &tsfadjust, sizeof(tsfadjust));
}
bf->bf_m = m;
bf->bf_node = ieee80211_ref_node(ni);
return 0;
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
/*
* Setup the beacon frame for transmit.
*/
static void
ath_beacon_setup(struct ath_softc *sc, struct ath_buf *bf)
{
#define USE_SHPREAMBLE(_ic) \
(((_ic)->ic_flags & (IEEE80211_F_SHPREAMBLE | IEEE80211_F_USEBARKER))\
== IEEE80211_F_SHPREAMBLE)
struct ieee80211_node *ni = bf->bf_node;
struct ieee80211com *ic = ni->ni_ic;
struct mbuf *m = bf->bf_m;
struct ath_hal *ah = sc->sc_ah;
struct ath_desc *ds;
int flags, antenna;
move mgt frame tx rate responsibility from the rate control modules to the driver; this avoids redundant logic and will be necessary for future additions MFC after: 2 weeks
2006-02-09 20:49:55 +00:00
const HAL_RATE_TABLE *rt;
u_int8_t rix, rate;
Update with last year of work.
2004-12-08 17:34:36 +00:00
Improve ath_draintxq debug info: dump the packet as well as the descriptor and handle the beacon q like other q's MFC after: 1 month
2006-04-16 18:24:27 +00:00
DPRINTF(sc, ATH_DEBUG_BEACON_PROC, "%s: m %p len %u\n",
Update with last year of work.
2004-12-08 17:34:36 +00:00
__func__, m, m->m_len);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/* setup descriptors */
ds = bf->bf_desc;
Update with last year of work.
2004-12-08 17:34:36 +00:00
flags = HAL_TXDESC_NOACK;
if (ic->ic_opmode == IEEE80211_M_IBSS && sc->sc_hasveol) {
ds->ds_link = bf->bf_daddr; /* self-linked */
flags |= HAL_TXDESC_VEOL;
/*
* Let hardware handle antenna switching.
*/
o move ath_sysctlattach down so variables it depends on are setup o use any fixed tx antenna for beacons transmitted in adhoc mode Submitted by: David Young MFC after: 3 days
2005-07-23 05:34:29 +00:00
antenna = sc->sc_txantenna;
Update with last year of work.
2004-12-08 17:34:36 +00:00
} else {
ds->ds_link = 0;
/*
* Switch antenna every 4 beacons.
* XXX assumes two antenna
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (sc->sc_txantenna != 0)
antenna = sc->sc_txantenna;
else if (sc->sc_stagbeacons && sc->sc_nbcnvaps != 0)
antenna = ((sc->sc_stats.ast_be_xmit / sc->sc_nbcnvaps) & 4 ? 2 : 1);
else
antenna = (sc->sc_stats.ast_be_xmit & 4 ? 2 : 1);
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
KASSERT(bf->bf_nseg == 1,
("multi-segment beacon frame; nseg %u", bf->bf_nseg));
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ds->ds_data = bf->bf_segs[0].ds_addr;
/*
* Calculate rate code.
* XXX everything at min xmit rate
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
rix = 0;
move mgt frame tx rate responsibility from the rate control modules to the driver; this avoids redundant logic and will be necessary for future additions MFC after: 2 weeks
2006-02-09 20:49:55 +00:00
rt = sc->sc_currates;
rate = rt->info[rix].rateCode;
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (USE_SHPREAMBLE(ic))
move mgt frame tx rate responsibility from the rate control modules to the driver; this avoids redundant logic and will be necessary for future additions MFC after: 2 weeks
2006-02-09 20:49:55 +00:00
rate |= rt->info[rix].shortPreamble;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_hal_setuptxdesc(ah, ds
Update with last year of work.
2004-12-08 17:34:36 +00:00
, m->m_len + IEEE80211_CRC_LEN /* frame length */
, sizeof(struct ieee80211_frame)/* header length */
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
, HAL_PKT_TYPE_BEACON /* Atheros packet type */
Update with last year of work.
2004-12-08 17:34:36 +00:00
, ni->ni_txpower /* txpower XXX */
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
, rate, 1 /* series 0 rate/tries */
, HAL_TXKEYIX_INVALID /* no encryption */
Update with last year of work.
2004-12-08 17:34:36 +00:00
, antenna /* antenna mode */
, flags /* no ack, veol for beacons */
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
, 0 /* rts/cts rate */
, 0 /* rts/cts duration */
);
/* NB: beacon's BufLen must be a multiple of 4 bytes */
ath_hal_filltxdesc(ah, ds
Update with last year of work.
2004-12-08 17:34:36 +00:00
, roundup(m->m_len, 4) /* buffer length */
, AH_TRUE /* first segment */
, AH_TRUE /* last segment */
, ds /* first descriptor */
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
#if 0
ath_desc_swap(ds);
#endif
Update with last year of work.
2004-12-08 17:34:36 +00:00
#undef USE_SHPREAMBLE
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update beacon handling to sync w/ vap code base: o add driver callback to handle notification of beacon changes; this is required for devices that manage beacon frames themselves (devices must override the default handler which does nothing) o move beacon update-related flags from ieee80211com to the beacon offsets storage (or handle however a driver wants) o expand beacon offsets structure with members needed for 11h/dfs and appie's o change calling convention for ieee80211_beacon_alloc and ieee80211_beacon_update o add overlapping bss support for 11g; requires driver to pass beacon frames from overlapping bss up to net80211 which is not presently done by any driver o move HT beacon contents update to a routine in the HT code area Reviewed by: avatar, thompsa, sephe Approved by: re (blanket wireless)
2007-09-17 19:07:24 +00:00
static void
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_beacon_update(struct ieee80211vap *vap, int item)
Update beacon handling to sync w/ vap code base: o add driver callback to handle notification of beacon changes; this is required for devices that manage beacon frames themselves (devices must override the default handler which does nothing) o move beacon update-related flags from ieee80211com to the beacon offsets storage (or handle however a driver wants) o expand beacon offsets structure with members needed for 11h/dfs and appie's o change calling convention for ieee80211_beacon_alloc and ieee80211_beacon_update o add overlapping bss support for 11g; requires driver to pass beacon frames from overlapping bss up to net80211 which is not presently done by any driver o move HT beacon contents update to a routine in the HT code area Reviewed by: avatar, thompsa, sephe Approved by: re (blanket wireless)
2007-09-17 19:07:24 +00:00
{
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211_beacon_offsets *bo = &ATH_VAP(vap)->av_boff;
Update beacon handling to sync w/ vap code base: o add driver callback to handle notification of beacon changes; this is required for devices that manage beacon frames themselves (devices must override the default handler which does nothing) o move beacon update-related flags from ieee80211com to the beacon offsets storage (or handle however a driver wants) o expand beacon offsets structure with members needed for 11h/dfs and appie's o change calling convention for ieee80211_beacon_alloc and ieee80211_beacon_update o add overlapping bss support for 11g; requires driver to pass beacon frames from overlapping bss up to net80211 which is not presently done by any driver o move HT beacon contents update to a routine in the HT code area Reviewed by: avatar, thompsa, sephe Approved by: re (blanket wireless)
2007-09-17 19:07:24 +00:00
setbit(bo->bo_flags, item);
}
Close race in handling mcast traffic when operating as an ap with stations in power save: add a new q where mcast frames are stashed and on beacon update (at DTIM) move frames from the mcast q to the cabq and start it. This ensures the cabq is only manipulated in one place. Sponsored by: Hobnob MFC after: 2 weeks
2006-06-26 03:10:45 +00:00
/*
* Append the contents of src to dst; both queues
* are assumed to be locked.
*/
static void
ath_txqmove(struct ath_txq *dst, struct ath_txq *src)
{
STAILQ_CONCAT(&dst->axq_q, &src->axq_q);
dst->axq_link = src->axq_link;
src->axq_link = NULL;
dst->axq_depth += src->axq_depth;
src->axq_depth = 0;
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Transmit a beacon frame at SWBA. Dynamic updates to the
* frame contents are done as needed and the slot time is
* also adjusted based on current state.
*/
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void
ath_beacon_proc(void *arg, int pending)
{
struct ath_softc *sc = arg;
struct ath_hal *ah = sc->sc_ah;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211vap *vap;
struct ath_buf *bf;
int slot, otherant;
uint32_t bfaddr;
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_BEACON_PROC, "%s: pending %u\n",
__func__, pending);
/*
* Check if the previous beacon has gone out. If
fixup comments
2006-02-15 18:15:49 +00:00
* not don't try to post another, skip this period
* and wait for the next. Missed beacons indicate
* a problem and should not occur. If we miss too
* many consecutive beacons reset the device.
Update with last year of work.
2004-12-08 17:34:36 +00:00
*/
if (ath_hal_numtxpending(ah, sc->sc_bhalq) != 0) {
sc->sc_bmisscount++;
Add a global counter of missed beacons. The existing missed beacon count is reset once a beacon isn't missed.
2010-08-14 14:01:12 +00:00
sc->sc_stats.ast_be_missed++;
Improve ath_draintxq debug info: dump the packet as well as the descriptor and handle the beacon q like other q's MFC after: 1 month
2006-04-16 18:24:27 +00:00
DPRINTF(sc, ATH_DEBUG_BEACON,
Update with last year of work.
2004-12-08 17:34:36 +00:00
"%s: missed %u consecutive beacons\n",
__func__, sc->sc_bmisscount);
add hw.ath.bstuck to control the stuck beacon threshold
2009-02-10 19:26:42 +00:00
if (sc->sc_bmisscount >= ath_bstuck_threshold)
use a private task queue thread MFC after: 2 weeks
2006-02-09 21:48:51 +00:00
taskqueue_enqueue(sc->sc_tq, &sc->sc_bstucktask);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
return;
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (sc->sc_bmisscount != 0) {
DPRINTF(sc, ATH_DEBUG_BEACON,
"%s: resume beacon xmit after %u misses\n",
__func__, sc->sc_bmisscount);
sc->sc_bmisscount = 0;
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (sc->sc_stagbeacons) { /* staggered beacons */
struct ieee80211com *ic = sc->sc_ifp->if_l2com;
uint32_t tsftu;
tsftu = ath_hal_gettsf32(ah) >> 10;
/* XXX lintval */
slot = ((tsftu % ic->ic_lintval) * ATH_BCBUF) / ic->ic_lintval;
vap = sc->sc_bslot[(slot+1) % ATH_BCBUF];
bfaddr = 0;
treat IEEE80211_S_CSA as a "running state"; this fixes ap mode 11h channel switch announcements
2009-06-03 17:25:19 +00:00
if (vap != NULL && vap->iv_state >= IEEE80211_S_RUN) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
bf = ath_beacon_generate(sc, vap);
if (bf != NULL)
bfaddr = bf->bf_daddr;
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
} else { /* burst'd beacons */
uint32_t *bflink = &bfaddr;
for (slot = 0; slot < ATH_BCBUF; slot++) {
vap = sc->sc_bslot[slot];
treat IEEE80211_S_CSA as a "running state"; this fixes ap mode 11h channel switch announcements
2009-06-03 17:25:19 +00:00
if (vap != NULL && vap->iv_state >= IEEE80211_S_RUN) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
bf = ath_beacon_generate(sc, vap);
if (bf != NULL) {
*bflink = bf->bf_daddr;
bflink = &bf->bf_desc->ds_link;
}
}
}
*bflink = 0; /* terminate list */
Close race in handling mcast traffic when operating as an ap with stations in power save: add a new q where mcast frames are stashed and on beacon update (at DTIM) move frames from the mcast q to the cabq and start it. This ensures the cabq is only manipulated in one place. Sponsored by: Hobnob MFC after: 2 weeks
2006-06-26 03:10:45 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Handle slot time change when a non-ERP station joins/leaves
* an 11g network. The 802.11 layer notifies us via callback,
* we mark updateslot, then wait one beacon before effecting
* the change. This gives associated stations at least one
* beacon interval to note the state change.
*/
/* XXX locking */
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (sc->sc_updateslot == UPDATE) {
Update with last year of work.
2004-12-08 17:34:36 +00:00
sc->sc_updateslot = COMMIT; /* commit next beacon */
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
sc->sc_slotupdate = slot;
} else if (sc->sc_updateslot == COMMIT && sc->sc_slotupdate == slot)
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_setslottime(sc); /* commit change to h/w */
/*
* Check recent per-antenna transmit statistics and flip
* the default antenna if noticeably more frames went out
* on the non-default antenna.
* XXX assumes 2 anntenae
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (!sc->sc_diversity && (!sc->sc_stagbeacons || slot == 0)) {
otherant = sc->sc_defant & 1 ? 2 : 1;
if (sc->sc_ant_tx[otherant] > sc->sc_ant_tx[sc->sc_defant] + 2)
ath_setdefantenna(sc, otherant);
sc->sc_ant_tx[1] = sc->sc_ant_tx[2] = 0;
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (bfaddr != 0) {
/*
* Stop any current dma and put the new frame on the queue.
* This should never fail since we check above that no frames
* are still pending on the queue.
*/
if (!ath_hal_stoptxdma(ah, sc->sc_bhalq)) {
DPRINTF(sc, ATH_DEBUG_ANY,
"%s: beacon queue %u did not stop?\n",
__func__, sc->sc_bhalq);
}
/* NB: cabq traffic should already be queued and primed */
ath_hal_puttxbuf(ah, sc->sc_bhalq, bfaddr);
ath_hal_txstart(ah, sc->sc_bhalq);
sc->sc_stats.ast_be_xmit++;
}
}
static struct ath_buf *
ath_beacon_generate(struct ath_softc *sc, struct ieee80211vap *vap)
{
struct ath_vap *avp = ATH_VAP(vap);
struct ath_txq *cabq = sc->sc_cabq;
struct ath_buf *bf;
struct mbuf *m;
int nmcastq, error;
treat IEEE80211_S_CSA as a "running state"; this fixes ap mode 11h channel switch announcements
2009-06-03 17:25:19 +00:00
KASSERT(vap->iv_state >= IEEE80211_S_RUN,
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
("not running, state %d", vap->iv_state));
KASSERT(avp->av_bcbuf != NULL, ("no beacon buffer"));
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
* Update dynamic beacon contents. If this returns
* non-zero then we need to remap the memory because
* the beacon frame changed size (probably because
* of the TIM bitmap).
Update with last year of work.
2004-12-08 17:34:36 +00:00
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
bf = avp->av_bcbuf;
m = bf->bf_m;
nmcastq = avp->av_mcastq.axq_depth;
if (ieee80211_beacon_update(bf->bf_node, &avp->av_boff, m, nmcastq)) {
/* XXX too conservative? */
bus_dmamap_unload(sc->sc_dmat, bf->bf_dmamap);
error = bus_dmamap_load_mbuf_sg(sc->sc_dmat, bf->bf_dmamap, m,
bf->bf_segs, &bf->bf_nseg,
BUS_DMA_NOWAIT);
if (error != 0) {
if_printf(vap->iv_ifp,
"%s: bus_dmamap_load_mbuf_sg failed, error %u\n",
__func__, error);
return NULL;
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if ((avp->av_boff.bo_tim[4] & 1) && cabq->axq_depth) {
DPRINTF(sc, ATH_DEBUG_BEACON,
"%s: cabq did not drain, mcastq %u cabq %u\n",
__func__, nmcastq, cabq->axq_depth);
sc->sc_stats.ast_cabq_busy++;
if (sc->sc_nvaps > 1 && sc->sc_stagbeacons) {
/*
* CABQ traffic from a previous vap is still pending.
* We must drain the q before this beacon frame goes
* out as otherwise this vap's stations will get cab
* frames from a different vap.
* XXX could be slow causing us to miss DBA
*/
ath_tx_draintxq(sc, cabq);
}
}
ath_beacon_setup(sc, bf);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
bus_dmamap_sync(sc->sc_dmat, bf->bf_dmamap, BUS_DMASYNC_PREWRITE);
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Enable the CAB queue before the beacon queue to
* insure cab frames are triggered by this beacon.
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (avp->av_boff.bo_tim[4] & 1) {
struct ath_hal *ah = sc->sc_ah;
check tim is present in the beacon before defer'ing the mcast buffer bit; insures we don't do this when operating in adhoc mode Submitted by: avatar@ MFC after: 1 week
2006-07-26 03:48:41 +00:00
/* NB: only at DTIM */
Close race in handling mcast traffic when operating as an ap with stations in power save: add a new q where mcast frames are stashed and on beacon update (at DTIM) move frames from the mcast q to the cabq and start it. This ensures the cabq is only manipulated in one place. Sponsored by: Hobnob MFC after: 2 weeks
2006-06-26 03:10:45 +00:00
ATH_TXQ_LOCK(cabq);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ATH_TXQ_LOCK(&avp->av_mcastq);
Close race in handling mcast traffic when operating as an ap with stations in power save: add a new q where mcast frames are stashed and on beacon update (at DTIM) move frames from the mcast q to the cabq and start it. This ensures the cabq is only manipulated in one place. Sponsored by: Hobnob MFC after: 2 weeks
2006-06-26 03:10:45 +00:00
if (nmcastq) {
struct ath_buf *bfm;
/*
* Move frames from the s/w mcast q to the h/w cab q.
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
* XXX MORE_DATA bit
Close race in handling mcast traffic when operating as an ap with stations in power save: add a new q where mcast frames are stashed and on beacon update (at DTIM) move frames from the mcast q to the cabq and start it. This ensures the cabq is only manipulated in one place. Sponsored by: Hobnob MFC after: 2 weeks
2006-06-26 03:10:45 +00:00
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
bfm = STAILQ_FIRST(&avp->av_mcastq.axq_q);
Close race in handling mcast traffic when operating as an ap with stations in power save: add a new q where mcast frames are stashed and on beacon update (at DTIM) move frames from the mcast q to the cabq and start it. This ensures the cabq is only manipulated in one place. Sponsored by: Hobnob MFC after: 2 weeks
2006-06-26 03:10:45 +00:00
if (cabq->axq_link != NULL) {
*cabq->axq_link = bfm->bf_daddr;
} else
ath_hal_puttxbuf(ah, cabq->axq_qnum,
bfm->bf_daddr);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_txqmove(cabq, &avp->av_mcastq);
Close race in handling mcast traffic when operating as an ap with stations in power save: add a new q where mcast frames are stashed and on beacon update (at DTIM) move frames from the mcast q to the cabq and start it. This ensures the cabq is only manipulated in one place. Sponsored by: Hobnob MFC after: 2 weeks
2006-06-26 03:10:45 +00:00
sc->sc_stats.ast_cabq_xmit += nmcastq;
}
/* NB: gated by beacon so safe to start here */
ath_hal_txstart(ah, cabq->axq_qnum);
ATH_TXQ_UNLOCK(cabq);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ATH_TXQ_UNLOCK(&avp->av_mcastq);
Close race in handling mcast traffic when operating as an ap with stations in power save: add a new q where mcast frames are stashed and on beacon update (at DTIM) move frames from the mcast q to the cabq and start it. This ensures the cabq is only manipulated in one place. Sponsored by: Hobnob MFC after: 2 weeks
2006-06-26 03:10:45 +00:00
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
return bf;
}
static void
ath_beacon_start_adhoc(struct ath_softc *sc, struct ieee80211vap *vap)
{
struct ath_vap *avp = ATH_VAP(vap);
struct ath_hal *ah = sc->sc_ah;
struct ath_buf *bf;
struct mbuf *m;
int error;
KASSERT(avp->av_bcbuf != NULL, ("no beacon buffer"));
/*
* Update dynamic beacon contents. If this returns
* non-zero then we need to remap the memory because
* the beacon frame changed size (probably because
* of the TIM bitmap).
*/
bf = avp->av_bcbuf;
m = bf->bf_m;
if (ieee80211_beacon_update(bf->bf_node, &avp->av_boff, m, 0)) {
/* XXX too conservative? */
bus_dmamap_unload(sc->sc_dmat, bf->bf_dmamap);
error = bus_dmamap_load_mbuf_sg(sc->sc_dmat, bf->bf_dmamap, m,
bf->bf_segs, &bf->bf_nseg,
BUS_DMA_NOWAIT);
if (error != 0) {
if_printf(vap->iv_ifp,
"%s: bus_dmamap_load_mbuf_sg failed, error %u\n",
__func__, error);
return;
}
}
ath_beacon_setup(sc, bf);
bus_dmamap_sync(sc->sc_dmat, bf->bf_dmamap, BUS_DMASYNC_PREWRITE);
/* NB: caller is known to have already stopped tx dma */
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_hal_puttxbuf(ah, sc->sc_bhalq, bf->bf_daddr);
ath_hal_txstart(ah, sc->sc_bhalq);
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
/*
* Reset the hardware after detecting beacons have stopped.
*/
static void
ath_bstuck_proc(void *arg, int pending)
{
struct ath_softc *sc = arg;
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Update with last year of work.
2004-12-08 17:34:36 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if_printf(ifp, "stuck beacon; resetting (bmiss count %u)\n",
sc->sc_bmisscount);
count stuck beacon events
2009-02-07 05:34:41 +00:00
sc->sc_stats.ast_bstuck++;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_reset(ifp);
}
/*
* Reclaim beacon resources and return buffer to the pool.
*/
static void
ath_beacon_return(struct ath_softc *sc, struct ath_buf *bf)
{
if (bf->bf_m != NULL) {
bus_dmamap_unload(sc->sc_dmat, bf->bf_dmamap);
m_freem(bf->bf_m);
bf->bf_m = NULL;
}
if (bf->bf_node != NULL) {
ieee80211_free_node(bf->bf_node);
bf->bf_node = NULL;
}
STAILQ_INSERT_TAIL(&sc->sc_bbuf, bf, bf_list);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Reclaim beacon resources.
*/
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void
ath_beacon_free(struct ath_softc *sc)
{
Update with last year of work.
2004-12-08 17:34:36 +00:00
struct ath_buf *bf;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
beacon handling fixups for adhoc mode: o don't reclaim any previous beacon state in ath_beacon_alloc; do it explicitly in ath_newstate o reference count the node held in the beacon frame state block o process ibss merge more intelligently; let the state machine do the right thing instead of explicitly setting the new bssi id o explicitly stop tx dma before doing beacon setup to handle the ibss merge case
2005-01-24 20:05:03 +00:00
STAILQ_FOREACH(bf, &sc->sc_bbuf, bf_list) {
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (bf->bf_m != NULL) {
bus_dmamap_unload(sc->sc_dmat, bf->bf_dmamap);
m_freem(bf->bf_m);
bf->bf_m = NULL;
beacon handling fixups for adhoc mode: o don't reclaim any previous beacon state in ath_beacon_alloc; do it explicitly in ath_newstate o reference count the node held in the beacon frame state block o process ibss merge more intelligently; let the state machine do the right thing instead of explicitly setting the new bssi id o explicitly stop tx dma before doing beacon setup to handle the ibss merge case
2005-01-24 20:05:03 +00:00
}
if (bf->bf_node != NULL) {
ieee80211_free_node(bf->bf_node);
Update with last year of work.
2004-12-08 17:34:36 +00:00
bf->bf_node = NULL;
}
beacon handling fixups for adhoc mode: o don't reclaim any previous beacon state in ath_beacon_alloc; do it explicitly in ath_newstate o reference count the node held in the beacon frame state block o process ibss merge more intelligently; let the state machine do the right thing instead of explicitly setting the new bssi id o explicitly stop tx dma before doing beacon setup to handle the ibss merge case
2005-01-24 20:05:03 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
/*
* Configure the beacon and sleep timers.
*
* When operating as an AP this resets the TSF and sets
* up the hardware to notify us when we need to issue beacons.
*
* When operating in station mode this sets up the beacon
* timers according to the timestamp of the last received
* beacon and the current TSF, configures PCF and DTIM
* handling, programs the sleep registers so the hardware
* will wakeup in time to receive beacons, and configures
* the beacon miss handling so we'll receive a BMISS
* interrupt when we stop seeing beacons from the AP
* we've associated with.
*/
static void
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_beacon_config(struct ath_softc *sc, struct ieee80211vap *vap)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
{
Beacon timer setup fixes: o pull nexttbtt forward in adhoc mode too o resync beacon timers on joining a bss or ibss as the tstamp we collected while scanning is almost certainly out of date Note we may need to refine the ibss mode check in ath_recv_mgmt. Reviewed by: avatar, dyoung Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:21:53 +00:00
#define TSF_TO_TU(_h,_l) \
((((u_int32_t)(_h)) << 22) | (((u_int32_t)(_l)) >> 10))
#define FUDGE 2
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ath_hal *ah = sc->sc_ah;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211com *ic = sc->sc_ifp->if_l2com;
struct ieee80211_node *ni;
Beacon timer setup fixes: o pull nexttbtt forward in adhoc mode too o resync beacon timers on joining a bss or ibss as the tstamp we collected while scanning is almost certainly out of date Note we may need to refine the ibss mode check in ath_recv_mgmt. Reviewed by: avatar, dyoung Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:21:53 +00:00
u_int32_t nexttbtt, intval, tsftu;
u_int64_t tsf;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (vap == NULL)
vap = TAILQ_FIRST(&ic->ic_vaps); /* XXX */
ni = vap->iv_bss;
Change station mode beacon timer setup to insure the calculated nextTbtt is always ahead of the h/w TSF. Reviewed by: avatar
2005-06-09 04:10:01 +00:00
/* extract tstamp from last beacon and convert to TU */
nexttbtt = TSF_TO_TU(LE_READ_4(ni->ni_tstamp.data + 4),
LE_READ_4(ni->ni_tstamp.data));
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
if (ic->ic_opmode == IEEE80211_M_HOSTAP ||
ic->ic_opmode == IEEE80211_M_MBSS) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/*
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
* For multi-bss ap/mesh support beacons are either staggered
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
* evenly over N slots or burst together. For the former
* arrange for the SWBA to be delivered for each slot.
* Slots that are not occupied will generate nothing.
*/
/* NB: the beacon interval is kept internally in TU's */
intval = ni->ni_intval & HAL_BEACON_PERIOD;
if (sc->sc_stagbeacons)
intval /= ATH_BCBUF;
} else {
/* NB: the beacon interval is kept internally in TU's */
intval = ni->ni_intval & HAL_BEACON_PERIOD;
}
Correct beacon timer setup logic: o lintval is in ms; must convert to TU's for passing to the hal o roundup to calculate nexttbtt (should look at current tsf and pull the calculated nextbtt forward but this'll do for now) o don't or- in HAL_BEACON_RESET_TSF when doing station timer setup; this is not needed and messes up the sleep timer calcs, though it's unclear if it mattered as the hal masks these values before use Submitted by: Thorsten von Eicken
2004-12-31 20:20:02 +00:00
if (nexttbtt == 0) /* e.g. for ap mode */
nexttbtt = intval;
else if (intval) /* NB: can be 0 for monitor mode */
nexttbtt = roundup(nexttbtt, intval);
DPRINTF(sc, ATH_DEBUG_BEACON, "%s: nexttbtt %u intval %u (%u)\n",
__func__, nexttbtt, intval, ni->ni_intval);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (ic->ic_opmode == IEEE80211_M_STA && !sc->sc_swbmiss) {
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
HAL_BEACON_STATE bs;
Change station mode beacon timer setup to insure the calculated nextTbtt is always ahead of the h/w TSF. Reviewed by: avatar
2005-06-09 04:10:01 +00:00
int dtimperiod, dtimcount;
int cfpperiod, cfpcount;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Change station mode beacon timer setup to insure the calculated nextTbtt is always ahead of the h/w TSF. Reviewed by: avatar
2005-06-09 04:10:01 +00:00
/*
* Setup dtim and cfp parameters according to
* last beacon we received (which may be none).
*/
dtimperiod = ni->ni_dtim_period;
if (dtimperiod <= 0) /* NB: 0 if not known */
dtimperiod = 1;
dtimcount = ni->ni_dtim_count;
if (dtimcount >= dtimperiod) /* NB: sanity check */
dtimcount = 0; /* XXX? */
cfpperiod = 1; /* NB: no PCF support yet */
cfpcount = 0;
/*
* Pull nexttbtt forward to reflect the current
* TSF and calculate dtim+cfp state for the result.
*/
tsf = ath_hal_gettsf64(ah);
Beacon timer setup fixes: o pull nexttbtt forward in adhoc mode too o resync beacon timers on joining a bss or ibss as the tstamp we collected while scanning is almost certainly out of date Note we may need to refine the ibss mode check in ath_recv_mgmt. Reviewed by: avatar, dyoung Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:21:53 +00:00
tsftu = TSF_TO_TU(tsf>>32, tsf) + FUDGE;
Change station mode beacon timer setup to insure the calculated nextTbtt is always ahead of the h/w TSF. Reviewed by: avatar
2005-06-09 04:10:01 +00:00
do {
nexttbtt += intval;
if (--dtimcount < 0) {
dtimcount = dtimperiod - 1;
if (--cfpcount < 0)
cfpcount = cfpperiod - 1;
}
} while (nexttbtt < tsftu);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
memset(&bs, 0, sizeof(bs));
Correct beacon timer setup logic: o lintval is in ms; must convert to TU's for passing to the hal o roundup to calculate nexttbtt (should look at current tsf and pull the calculated nextbtt forward but this'll do for now) o don't or- in HAL_BEACON_RESET_TSF when doing station timer setup; this is not needed and messes up the sleep timer calcs, though it's unclear if it mattered as the hal masks these values before use Submitted by: Thorsten von Eicken
2004-12-31 20:20:02 +00:00
bs.bs_intval = intval;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
bs.bs_nexttbtt = nexttbtt;
Change station mode beacon timer setup to insure the calculated nextTbtt is always ahead of the h/w TSF. Reviewed by: avatar
2005-06-09 04:10:01 +00:00
bs.bs_dtimperiod = dtimperiod*intval;
bs.bs_nextdtim = bs.bs_nexttbtt + dtimcount*intval;
bs.bs_cfpperiod = cfpperiod*bs.bs_dtimperiod;
bs.bs_cfpnext = bs.bs_nextdtim + cfpcount*bs.bs_dtimperiod;
bs.bs_cfpmaxduration = 0;
#if 0
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* The 802.11 layer records the offset to the DTIM
* bitmap while receiving beacons; use it here to
* enable h/w detection of our AID being marked in
* the bitmap vector (to indicate frames for us are
* pending at the AP).
Change station mode beacon timer setup to insure the calculated nextTbtt is always ahead of the h/w TSF. Reviewed by: avatar
2005-06-09 04:10:01 +00:00
* XXX do DTIM handling in s/w to WAR old h/w bugs
* XXX enable based on h/w rev for newer chips
Update with last year of work.
2004-12-08 17:34:36 +00:00
*/
bs.bs_timoffset = ni->ni_timoff;
Change station mode beacon timer setup to insure the calculated nextTbtt is always ahead of the h/w TSF. Reviewed by: avatar
2005-06-09 04:10:01 +00:00
#endif
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/*
* Calculate the number of consecutive beacons to miss
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
* before taking a BMISS interrupt.
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
* Note that we clamp the result to at most 10 beacons.
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
bs.bs_bmissthreshold = vap->iv_bmissthreshold;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if (bs.bs_bmissthreshold > 10)
bs.bs_bmissthreshold = 10;
else if (bs.bs_bmissthreshold <= 0)
bs.bs_bmissthreshold = 1;
/*
* Calculate sleep duration. The configuration is
* given in ms. We insure a multiple of the beacon
* period is used. Also, if the sleep duration is
* greater than the DTIM period then it makes senses
* to make it a multiple of that.
*
* XXX fixed at 100ms
*/
o correct beacon interval calculation; the internal setting is in TU's not ms o replace the private macro to convert MS->TU with the common one
2005-01-24 19:45:59 +00:00
bs.bs_sleepduration =
roundup(IEEE80211_MS_TO_TU(100), bs.bs_intval);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if (bs.bs_sleepduration > bs.bs_dtimperiod)
bs.bs_sleepduration = roundup(bs.bs_sleepduration, bs.bs_dtimperiod);
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
DPRINTF(sc, ATH_DEBUG_BEACON,
Change station mode beacon timer setup to insure the calculated nextTbtt is always ahead of the h/w TSF. Reviewed by: avatar
2005-06-09 04:10:01 +00:00
"%s: tsf %ju tsf:tu %u intval %u nexttbtt %u dtim %u nextdtim %u bmiss %u sleep %u cfp:period %u maxdur %u next %u timoffset %u\n"
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
, __func__
Change station mode beacon timer setup to insure the calculated nextTbtt is always ahead of the h/w TSF. Reviewed by: avatar
2005-06-09 04:10:01 +00:00
, tsf, tsftu
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
, bs.bs_intval
, bs.bs_nexttbtt
, bs.bs_dtimperiod
, bs.bs_nextdtim
, bs.bs_bmissthreshold
, bs.bs_sleepduration
Update with last year of work.
2004-12-08 17:34:36 +00:00
, bs.bs_cfpperiod
, bs.bs_cfpmaxduration
, bs.bs_cfpnext
, bs.bs_timoffset
);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_hal_intrset(ah, 0);
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_hal_beacontimers(ah, &bs);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_imask |= HAL_INT_BMISS;
ath_hal_intrset(ah, sc->sc_imask);
} else {
ath_hal_intrset(ah, 0);
Correct beacon timer setup logic: o lintval is in ms; must convert to TU's for passing to the hal o roundup to calculate nexttbtt (should look at current tsf and pull the calculated nextbtt forward but this'll do for now) o don't or- in HAL_BEACON_RESET_TSF when doing station timer setup; this is not needed and messes up the sleep timer calcs, though it's unclear if it mattered as the hal masks these values before use Submitted by: Thorsten von Eicken
2004-12-31 20:20:02 +00:00
if (nexttbtt == intval)
Update with last year of work.
2004-12-08 17:34:36 +00:00
intval |= HAL_BEACON_RESET_TSF;
if (ic->ic_opmode == IEEE80211_M_IBSS) {
/*
* In IBSS mode enable the beacon timers but only
* enable SWBA interrupts if we need to manually
* prepare beacon frames. Otherwise we use a
* self-linked tx descriptor and let the hardware
* deal with things.
*/
intval |= HAL_BEACON_ENA;
if (!sc->sc_hasveol)
sc->sc_imask |= HAL_INT_SWBA;
Beacon timer setup fixes: o pull nexttbtt forward in adhoc mode too o resync beacon timers on joining a bss or ibss as the tstamp we collected while scanning is almost certainly out of date Note we may need to refine the ibss mode check in ath_recv_mgmt. Reviewed by: avatar, dyoung Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:21:53 +00:00
if ((intval & HAL_BEACON_RESET_TSF) == 0) {
/*
* Pull nexttbtt forward to reflect
* the current TSF.
*/
tsf = ath_hal_gettsf64(ah);
tsftu = TSF_TO_TU(tsf>>32, tsf) + FUDGE;
do {
nexttbtt += intval;
} while (nexttbtt < tsftu);
}
Set the correct IFS parameters for the beacon tx queue when operating in ap and adhoc modes.
2005-06-07 00:12:40 +00:00
ath_beaconq_config(sc);
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
} else if (ic->ic_opmode == IEEE80211_M_HOSTAP ||
ic->ic_opmode == IEEE80211_M_MBSS) {
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
* In AP/mesh mode we enable the beacon timers
* and SWBA interrupts to prepare beacon frames.
Update with last year of work.
2004-12-08 17:34:36 +00:00
*/
intval |= HAL_BEACON_ENA;
o remove bmisshack no longer needed with the BSSID fix in v0.9.5.2 of the hal o add monitor mode support o fix short preamble handling in beacon setup (noop) o correct resume handling
2003-08-02 06:14:15 +00:00
sc->sc_imask |= HAL_INT_SWBA; /* beacon prepare */
Set the correct IFS parameters for the beacon tx queue when operating in ap and adhoc modes.
2005-06-07 00:12:40 +00:00
ath_beaconq_config(sc);
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
ath_hal_beaconinit(ah, nexttbtt, intval);
sc->sc_bmisscount = 0;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_hal_intrset(ah, sc->sc_imask);
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* When using a self-linked beacon descriptor in
* ibss mode load it once here.
*/
if (ic->ic_opmode == IEEE80211_M_IBSS && sc->sc_hasveol)
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_beacon_start_adhoc(sc, vap);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Beacon timer setup fixes: o pull nexttbtt forward in adhoc mode too o resync beacon timers on joining a bss or ibss as the tstamp we collected while scanning is almost certainly out of date Note we may need to refine the ibss mode check in ath_recv_mgmt. Reviewed by: avatar, dyoung Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:21:53 +00:00
sc->sc_syncbeacon = 0;
#undef FUDGE
Change station mode beacon timer setup to insure the calculated nextTbtt is always ahead of the h/w TSF. Reviewed by: avatar
2005-06-09 04:10:01 +00:00
#undef TSF_TO_TU
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
static void
ath_load_cb(void *arg, bus_dma_segment_t *segs, int nsegs, int error)
{
bus_addr_t *paddr = (bus_addr_t*) arg;
cleanup some assertions
2004-12-31 20:11:23 +00:00
KASSERT(error == 0, ("error %u on bus_dma callback", error));
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
*paddr = segs->ds_addr;
}
static int
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_descdma_setup(struct ath_softc *sc,
struct ath_descdma *dd, ath_bufhead *head,
const char *name, int nbuf, int ndesc)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
{
Update with last year of work.
2004-12-08 17:34:36 +00:00
#define DS2PHYS(_dd, _ds) \
((_dd)->dd_desc_paddr + ((caddr_t)(_ds) - (caddr_t)(_dd)->dd_desc))
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ath_desc *ds;
struct ath_buf *bf;
Update with last year of work.
2004-12-08 17:34:36 +00:00
int i, bsize, error;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_RESET, "%s: %s DMA: %u buffers %u desc/buf\n",
__func__, name, nbuf, ndesc);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
dd->dd_name = name;
dd->dd_desc_len = sizeof(struct ath_desc) * nbuf * ndesc;
/*
* Setup DMA descriptor area.
*/
Change the remainder of the drivers for DMA'ing devices enabled in the sparc64 GENERIC and the sound device drivers known working on sparc64 to use bus_get_dma_tag() to obtain the parent DMA tag so we can get rid of the sparc64_root_dma_tag kludge eventually. Except for ath(4), sk(4), stge(4) and ti(4) these changes are runtime tested (unless I booted up the wrong kernels again...).
2007-01-21 19:32:51 +00:00
error = bus_dma_tag_create(bus_get_dma_tag(sc->sc_dev), /* parent */
Update with last year of work.
2004-12-08 17:34:36 +00:00
PAGE_SIZE, 0, /* alignment, bounds */
BUS_SPACE_MAXADDR_32BIT, /* lowaddr */
BUS_SPACE_MAXADDR, /* highaddr */
NULL, NULL, /* filter, filterarg */
dd->dd_desc_len, /* maxsize */
1, /* nsegments */
quiet tindexbox complaints about passing BUS_SPACE_MAXADDR as a bus_size_t to bus_dma_tag_create; when PAE is enabled this does not work Cluebat by: scottl MFC after: 2 weeks
2006-05-08 20:11:09 +00:00
dd->dd_desc_len, /* maxsegsize */
Update with last year of work.
2004-12-08 17:34:36 +00:00
BUS_DMA_ALLOCNOW, /* flags */
NULL, /* lockfunc */
NULL, /* lockarg */
&dd->dd_dmat);
if (error != 0) {
if_printf(ifp, "cannot allocate %s DMA tag\n", dd->dd_name);
return error;
}
/* allocate descriptors */
error = bus_dmamap_create(dd->dd_dmat, BUS_DMA_NOWAIT, &dd->dd_dmamap);
if (error != 0) {
if_printf(ifp, "unable to create dmamap for %s descriptors, "
"error %u\n", dd->dd_name, error);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
goto fail0;
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
error = bus_dmamem_alloc(dd->dd_dmat, (void**) &dd->dd_desc,
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
BUS_DMA_NOWAIT | BUS_DMA_COHERENT,
mark tx/rx descriptors COHERENT; we do not sync changes so on architectures like arm this is necessary MFC after: 1 month
2006-11-24 22:45:26 +00:00
&dd->dd_dmamap);
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (error != 0) {
if_printf(ifp, "unable to alloc memory for %u %s descriptors, "
"error %u\n", nbuf * ndesc, dd->dd_name, error);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
goto fail1;
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
error = bus_dmamap_load(dd->dd_dmat, dd->dd_dmamap,
dd->dd_desc, dd->dd_desc_len,
ath_load_cb, &dd->dd_desc_paddr,
BUS_DMA_NOWAIT);
if (error != 0) {
if_printf(ifp, "unable to map %s descriptors, error %u\n",
dd->dd_name, error);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
goto fail2;
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
ds = dd->dd_desc;
DPRINTF(sc, ATH_DEBUG_RESET, "%s: %s DMA map: %p (%lu) -> %p (%lu)\n",
__func__, dd->dd_name, ds, (u_long) dd->dd_desc_len,
(caddr_t) dd->dd_desc_paddr, /*XXX*/ (u_long) dd->dd_desc_len);
backout 1.136 until we can resolve report that it causes output to stall
2006-02-27 17:20:23 +00:00
/* allocate rx buffers */
Update with last year of work.
2004-12-08 17:34:36 +00:00
bsize = sizeof(struct ath_buf) * nbuf;
bf = malloc(bsize, M_ATHDEV, M_NOWAIT | M_ZERO);
if (bf == NULL) {
if_printf(ifp, "malloc of %s buffers failed, size %u\n",
dd->dd_name, bsize);
goto fail3;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
dd->dd_bufptr = bf;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
STAILQ_INIT(head);
for (i = 0; i < nbuf; i++, bf++, ds += ndesc) {
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
bf->bf_desc = ds;
Update with last year of work.
2004-12-08 17:34:36 +00:00
bf->bf_daddr = DS2PHYS(dd, ds);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
error = bus_dmamap_create(sc->sc_dmat, BUS_DMA_NOWAIT,
Update with last year of work.
2004-12-08 17:34:36 +00:00
&bf->bf_dmamap);
if (error != 0) {
if_printf(ifp, "unable to create dmamap for %s "
"buffer %u, error %u\n", dd->dd_name, i, error);
ath_descdma_cleanup(sc, dd, head);
return error;
}
STAILQ_INSERT_TAIL(head, bf, bf_list);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
return 0;
Update with last year of work.
2004-12-08 17:34:36 +00:00
fail3:
bus_dmamap_unload(dd->dd_dmat, dd->dd_dmamap);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
fail2:
Update with last year of work.
2004-12-08 17:34:36 +00:00
bus_dmamem_free(dd->dd_dmat, dd->dd_desc, dd->dd_dmamap);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
fail1:
Update with last year of work.
2004-12-08 17:34:36 +00:00
bus_dmamap_destroy(dd->dd_dmat, dd->dd_dmamap);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
fail0:
Update with last year of work.
2004-12-08 17:34:36 +00:00
bus_dma_tag_destroy(dd->dd_dmat);
memset(dd, 0, sizeof(*dd));
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
return error;
Update with last year of work.
2004-12-08 17:34:36 +00:00
#undef DS2PHYS
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
static void
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_descdma_cleanup(struct ath_softc *sc,
struct ath_descdma *dd, ath_bufhead *head)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
{
struct ath_buf *bf;
Update with last year of work.
2004-12-08 17:34:36 +00:00
struct ieee80211_node *ni;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
bus_dmamap_unload(dd->dd_dmat, dd->dd_dmamap);
bus_dmamem_free(dd->dd_dmat, dd->dd_desc, dd->dd_dmamap);
bus_dmamap_destroy(dd->dd_dmat, dd->dd_dmamap);
bus_dma_tag_destroy(dd->dd_dmat);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
STAILQ_FOREACH(bf, head, bf_list) {
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if (bf->bf_m) {
m_freem(bf->bf_m);
bf->bf_m = NULL;
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (bf->bf_dmamap != NULL) {
bus_dmamap_destroy(sc->sc_dmat, bf->bf_dmamap);
bf->bf_dmamap = NULL;
}
ni = bf->bf_node;
bf->bf_node = NULL;
if (ni != NULL) {
/*
* Reclaim node reference.
*/
ieee80211_free_node(ni);
}
}
STAILQ_INIT(head);
free(dd->dd_bufptr, M_ATHDEV);
memset(dd, 0, sizeof(*dd));
}
static int
ath_desc_alloc(struct ath_softc *sc)
{
int error;
error = ath_descdma_setup(sc, &sc->sc_rxdma, &sc->sc_rxbuf,
allow the size of tx+rx buffer pools to be tuned MFC after: 2 weeks
2006-02-09 21:03:25 +00:00
"rx", ath_rxbuf, 1);
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (error != 0)
return error;
error = ath_descdma_setup(sc, &sc->sc_txdma, &sc->sc_txbuf,
allow the size of tx+rx buffer pools to be tuned MFC after: 2 weeks
2006-02-09 21:03:25 +00:00
"tx", ath_txbuf, ATH_TXDESC);
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (error != 0) {
ath_descdma_cleanup(sc, &sc->sc_rxdma, &sc->sc_rxbuf);
return error;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
error = ath_descdma_setup(sc, &sc->sc_bdma, &sc->sc_bbuf,
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
"beacon", ATH_BCBUF, 1);
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (error != 0) {
ath_descdma_cleanup(sc, &sc->sc_txdma, &sc->sc_txbuf);
ath_descdma_cleanup(sc, &sc->sc_rxdma, &sc->sc_rxbuf);
return error;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
return 0;
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
static void
ath_desc_free(struct ath_softc *sc)
{
if (sc->sc_bdma.dd_desc_len != 0)
ath_descdma_cleanup(sc, &sc->sc_bdma, &sc->sc_bbuf);
if (sc->sc_txdma.dd_desc_len != 0)
ath_descdma_cleanup(sc, &sc->sc_txdma, &sc->sc_txbuf);
if (sc->sc_rxdma.dd_desc_len != 0)
ath_descdma_cleanup(sc, &sc->sc_rxdma, &sc->sc_rxbuf);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
static struct ieee80211_node *
Change the calling convention for ic_node_alloc to deal with some longstanding issues: o pass the vap since it's now the "coin of the realm" and required to do things like set initial tx parameters in private node state for use prior to association o pass the mac address as cards that maintain outboard station tables require this to create an entry (e.g. in ibss mode) o remove the node table reference, we only have one node table and it's unlikely this will change so this is not needed to find the com structure
2008-06-07 18:38:02 +00:00
ath_node_alloc(struct ieee80211vap *vap, const uint8_t mac[IEEE80211_ADDR_LEN])
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
{
Change the calling convention for ic_node_alloc to deal with some longstanding issues: o pass the vap since it's now the "coin of the realm" and required to do things like set initial tx parameters in private node state for use prior to association o pass the mac address as cards that maintain outboard station tables require this to create an entry (e.g. in ibss mode) o remove the node table reference, we only have one node table and it's unlikely this will change so this is not needed to find the com structure
2008-06-07 18:38:02 +00:00
struct ieee80211com *ic = vap->iv_ic;
Update with last year of work.
2004-12-08 17:34:36 +00:00
struct ath_softc *sc = ic->ic_ifp->if_softc;
const size_t space = sizeof(struct ath_node) + sc->sc_rc->arc_space;
struct ath_node *an;
an = malloc(space, M_80211_NODE, M_NOWAIT|M_ZERO);
if (an == NULL) {
/* XXX stat+msg */
Maintain a history of data associated with received frames and use this to calculate smoothed signal quality data for each node. o add a 16-deep history buffer to each driver-private node storage that holds rssi and antenna info for received frames o override the default per-node "get rssi" method to return an average rssi value based on samples collected over the last second o enable beacon reception so even idle systems maintain a running history of signal quality This data may also be useful for improving the rate control algorithm. Based on work by Tom Marshall <tommy@home.tig-grr.com> for MADWIFI.
2003-09-15 22:34:46 +00:00
return NULL;
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
ath_rate_node_init(sc, an);
DPRINTF(sc, ATH_DEBUG_NODE, "%s: an %p\n", __func__, an);
return &an->an_node;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
static void
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_node_free(struct ieee80211_node *ni)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
{
Update with last year of work.
2004-12-08 17:34:36 +00:00
struct ieee80211com *ic = ni->ni_ic;
struct ath_softc *sc = ic->ic_ifp->if_softc;
do proper subclassing of node free+copy; the previous hack falls apart when the 802.11 layer does useful work Obtained from: madwifi
2004-04-03 00:06:23 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_NODE, "%s: ni %p\n", __func__, ni);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_rate_node_cleanup(sc, ATH_NODE(ni));
sc->sc_node_free(ni);
}
Maintain a history of data associated with received frames and use this to calculate smoothed signal quality data for each node. o add a 16-deep history buffer to each driver-private node storage that holds rssi and antenna info for received frames o override the default per-node "get rssi" method to return an average rssi value based on samples collected over the last second o enable beacon reception so even idle systems maintain a running history of signal quality This data may also be useful for improving the rate control algorithm. Based on work by Tom Marshall <tommy@home.tig-grr.com> for MADWIFI.
2003-09-15 22:34:46 +00:00
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
static void
ath_node_getsignal(const struct ieee80211_node *ni, int8_t *rssi, int8_t *noise)
{
struct ieee80211com *ic = ni->ni_ic;
struct ath_softc *sc = ic->ic_ifp->if_softc;
struct ath_hal *ah = sc->sc_ah;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
*rssi = ic->ic_node_getrssi(ni);
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 (ni->ni_chan != IEEE80211_CHAN_ANYC)
*noise = ath_hal_getchannoise(ah, ni->ni_chan);
else
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
*noise = -95; /* nominally correct */
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static int
ath_rxbuf_init(struct ath_softc *sc, struct ath_buf *bf)
{
struct ath_hal *ah = sc->sc_ah;
int error;
struct mbuf *m;
struct ath_desc *ds;
m = bf->bf_m;
if (m == NULL) {
/*
* NB: by assigning a page to the rx dma buffer we
* implicitly satisfy the Atheros requirement that
* this buffer be cache-line-aligned and sized to be
* multiple of the cache line size. Not doing this
* causes weird stuff to happen (for the 5210 at least).
*/
m = m_getcl(M_DONTWAIT, MT_DATA, M_PKTHDR);
if (m == NULL) {
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_ANY,
"%s: no mbuf/cluster\n", __func__);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_stats.ast_rx_nombuf++;
return ENOMEM;
}
m->m_pkthdr.len = m->m_len = m->m_ext.ext_size;
switch to use bus_dmamap_load_mbuf_sg
2005-01-24 19:51:59 +00:00
error = bus_dmamap_load_mbuf_sg(sc->sc_dmat,
Update with last year of work.
2004-12-08 17:34:36 +00:00
bf->bf_dmamap, m,
switch to use bus_dmamap_load_mbuf_sg
2005-01-24 19:51:59 +00:00
bf->bf_segs, &bf->bf_nseg,
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
BUS_DMA_NOWAIT);
if (error != 0) {
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_ANY,
switch to use bus_dmamap_load_mbuf_sg
2005-01-24 19:51:59 +00:00
"%s: bus_dmamap_load_mbuf_sg failed; error %d\n",
__func__, error);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_stats.ast_rx_busdma++;
correct handling of mbuf allocation failure when replenishing the rx list (leave a printf for the moment, need to make a debug msg) Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:29:23 +00:00
m_freem(m);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
return error;
}
cleanup some assertions
2004-12-31 20:11:23 +00:00
KASSERT(bf->bf_nseg == 1,
("multi-segment packet; nseg %u", bf->bf_nseg));
correct handling of mbuf allocation failure when replenishing the rx list (leave a printf for the moment, need to make a debug msg) Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:29:23 +00:00
bf->bf_m = m;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
bus_dmamap_sync(sc->sc_dmat, bf->bf_dmamap, BUS_DMASYNC_PREREAD);
terminate the rx descriptor list with a self-linked entry so high phy error rates on a 5212 don't cause rx overruns
2003-10-22 04:37:34 +00:00
/*
* Setup descriptors. For receive we always terminate
* the descriptor list with a self-linked entry so we'll
* not get overrun under high load (as can happen with a
Update with last year of work.
2004-12-08 17:34:36 +00:00
* 5212 when ANI processing enables PHY error frames).
terminate the rx descriptor list with a self-linked entry so high phy error rates on a 5212 don't cause rx overruns
2003-10-22 04:37:34 +00:00
*
* To insure the last descriptor is self-linked we create
* each descriptor as self-linked and add it to the end. As
* each additional descriptor is added the previous self-linked
* entry is ``fixed'' naturally. This should be safe even
* if DMA is happening. When processing RX interrupts we
* never remove/process the last, self-linked, entry on the
* descriptor list. This insures the hardware always has
* someplace to write a new frame.
*/
Add a HAL capability bit for supporting self-linked RX descriptors and disable it for the 11n chipsets. From the ath9k source: == 11N: we can no longer afford to self link the last descriptor. MAC acknowledges BA status as long as it copies frames to host buffer (or rx fifo). This can incorrectly acknowledge packets to a sender if last desc is self-linked. == Since this is useful for pre-AR5416 chips that communicate PHY errors via error frames rather than by on-chip counters, leave the support in there, but disable it for AR5416 and later.
2011-04-04 14:52:31 +00:00
/*
* 11N: we can no longer afford to self link the last descriptor.
* MAC acknowledges BA status as long as it copies frames to host
* buffer (or rx fifo). This can incorrectly acknowledge packets
* to a sender if last desc is self-linked.
*/
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ds = bf->bf_desc;
Add a HAL capability bit for supporting self-linked RX descriptors and disable it for the 11n chipsets. From the ath9k source: == 11N: we can no longer afford to self link the last descriptor. MAC acknowledges BA status as long as it copies frames to host buffer (or rx fifo). This can incorrectly acknowledge packets to a sender if last desc is self-linked. == Since this is useful for pre-AR5416 chips that communicate PHY errors via error frames rather than by on-chip counters, leave the support in there, but disable it for AR5416 and later.
2011-04-04 14:52:31 +00:00
if (sc->sc_rxslink)
ds->ds_link = bf->bf_daddr; /* link to self */
else
ds->ds_link = 0; /* terminate the list */
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ds->ds_data = bf->bf_segs[0].ds_addr;
ath_hal_setuprxdesc(ah, ds
, m->m_len /* buffer size */
, 0
);
if (sc->sc_rxlink != NULL)
*sc->sc_rxlink = bf->bf_daddr;
sc->sc_rxlink = &ds->ds_link;
return 0;
}
extend the timestamp from the rx descriptor to calculate the tsf to use when checking for an ibss merge
2005-03-29 22:16:49 +00:00
/*
* Extend 15-bit time stamp from rx descriptor to
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
* a full 64-bit TSF using the specified TSF.
extend the timestamp from the rx descriptor to calculate the tsf to use when checking for an ibss merge
2005-03-29 22:16:49 +00:00
*/
static __inline u_int64_t
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
ath_extend_tsf(u_int32_t rstamp, u_int64_t tsf)
extend the timestamp from the rx descriptor to calculate the tsf to use when checking for an ibss merge
2005-03-29 22:16:49 +00:00
{
if ((tsf & 0x7fff) < rstamp)
tsf -= 0x8000;
return ((tsf &~ 0x7fff) | rstamp);
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Intercept management frames to collect beacon rssi data
* and to do ibss merges.
*/
static void
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_recv_mgmt(struct ieee80211_node *ni, struct mbuf *m,
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
int subtype, int rssi, int nf)
Update with last year of work.
2004-12-08 17:34:36 +00:00
{
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211vap *vap = ni->ni_vap;
struct ath_softc *sc = vap->iv_ic->ic_ifp->if_softc;
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Call up first so subsequent work can use information
* potentially stored in the node (e.g. for ibss merge).
*/
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
ATH_VAP(vap)->av_recv_mgmt(ni, m, subtype, rssi, nf);
Update with last year of work.
2004-12-08 17:34:36 +00:00
switch (subtype) {
case IEEE80211_FC0_SUBTYPE_BEACON:
/* update rssi statistics for use by the hal */
move hal statistics to softc; the per-node stats are overkill, they're only used when operating in station mode MFC after: 2 weeks
2006-02-09 21:23:44 +00:00
ATH_RSSI_LPF(sc->sc_halstats.ns_avgbrssi, rssi);
Beacon timer setup fixes: o pull nexttbtt forward in adhoc mode too o resync beacon timers on joining a bss or ibss as the tstamp we collected while scanning is almost certainly out of date Note we may need to refine the ibss mode check in ath_recv_mgmt. Reviewed by: avatar, dyoung Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:21:53 +00:00
if (sc->sc_syncbeacon &&
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ni == vap->iv_bss && vap->iv_state == IEEE80211_S_RUN) {
Beacon timer setup fixes: o pull nexttbtt forward in adhoc mode too o resync beacon timers on joining a bss or ibss as the tstamp we collected while scanning is almost certainly out of date Note we may need to refine the ibss mode check in ath_recv_mgmt. Reviewed by: avatar, dyoung Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:21:53 +00:00
/*
* Resync beacon timers using the tsf of the beacon
* frame we just received.
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_beacon_config(sc, vap);
Beacon timer setup fixes: o pull nexttbtt forward in adhoc mode too o resync beacon timers on joining a bss or ibss as the tstamp we collected while scanning is almost certainly out of date Note we may need to refine the ibss mode check in ath_recv_mgmt. Reviewed by: avatar, dyoung Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:21:53 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
/* fall thru... */
case IEEE80211_FC0_SUBTYPE_PROBE_RESP:
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (vap->iv_opmode == IEEE80211_M_IBSS &&
vap->iv_state == IEEE80211_S_RUN) {
sc_lastrs is also used in case the sending station is not known, for example in a split IBSS scenario. Therefore always assign sc_lastrs. This removes a hack I committed in r206457. Approved by: rpaulo (mentor)
2010-06-14 08:24:00 +00:00
uint32_t rstamp = sc->sc_lastrs->rs_tstamp;
uint64_t tsf = ath_extend_tsf(rstamp,
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
ath_hal_gettsf64(sc->sc_ah));
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Handle ibss merge as needed; check the tsf on the
* frame before attempting the merge. The 802.11 spec
* says the station should change it's bssid to match
* the oldest station with the same ssid, where oldest
beacon handling fixups for adhoc mode: o don't reclaim any previous beacon state in ath_beacon_alloc; do it explicitly in ath_newstate o reference count the node held in the beacon frame state block o process ibss merge more intelligently; let the state machine do the right thing instead of explicitly setting the new bssi id o explicitly stop tx dma before doing beacon setup to handle the ibss merge case
2005-01-24 20:05:03 +00:00
* is determined by the tsf. Note that hardware
* reconfiguration happens through callback to
extend the timestamp from the rx descriptor to calculate the tsf to use when checking for an ibss merge
2005-03-29 22:16:49 +00:00
* ath_newstate as the state machine will go from
* RUN -> RUN when this happens.
Update with last year of work.
2004-12-08 17:34:36 +00:00
*/
extend the timestamp from the rx descriptor to calculate the tsf to use when checking for an ibss merge
2005-03-29 22:16:49 +00:00
if (le64toh(ni->ni_tstamp.tsf) >= tsf) {
DPRINTF(sc, ATH_DEBUG_STATE,
Fixing kernel build on amd64 machines. Reviewed by: sam (mentor)
2005-03-30 02:33:33 +00:00
"ibss merge, rstamp %u tsf %ju "
"tstamp %ju\n", rstamp, (uintmax_t)tsf,
(uintmax_t)ni->ni_tstamp.tsf);
simplify ieee80211_ibss_merge api MFC after: 3 days
2005-07-22 17:54:40 +00:00
(void) ieee80211_ibss_merge(ni);
extend the timestamp from the rx descriptor to calculate the tsf to use when checking for an ibss merge
2005-03-29 22:16:49 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
break;
}
}
/*
* Set the default antenna.
*/
static void
ath_setdefantenna(struct ath_softc *sc, u_int antenna)
{
struct ath_hal *ah = sc->sc_ah;
/* XXX block beacon interrupts */
ath_hal_setdefantenna(ah, antenna);
if (sc->sc_defant != antenna)
sc->sc_stats.ast_ant_defswitch++;
sc->sc_defant = antenna;
sc->sc_rxotherant = 0;
}
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
static void
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_rx_tap(struct ifnet *ifp, struct mbuf *m,
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
const struct ath_rx_status *rs, u_int64_t tsf, int16_t nf)
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
{
correct rx radiotap channel flags construction for 11n frames
2008-05-29 00:14:35 +00:00
#define CHAN_HT20 htole32(IEEE80211_CHAN_HT20)
#define CHAN_HT40U htole32(IEEE80211_CHAN_HT40U)
#define CHAN_HT40D htole32(IEEE80211_CHAN_HT40D)
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
#define CHAN_HT (CHAN_HT20|CHAN_HT40U|CHAN_HT40D)
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ath_softc *sc = ifp->if_softc;
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
const HAL_RATE_TABLE *rt;
uint8_t rix;
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
rt = sc->sc_currates;
KASSERT(rt != NULL, ("no rate table, mode %u", sc->sc_curmode));
rix = rt->rateCodeToIndex[rs->rs_rate];
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
sc->sc_rx_th.wr_rate = sc->sc_hwmap[rix].ieeerate;
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
sc->sc_rx_th.wr_flags = sc->sc_hwmap[rix].rxflags;
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
#ifdef AH_SUPPORT_AR5416
correct rx radiotap channel flags construction for 11n frames
2008-05-29 00:14:35 +00:00
sc->sc_rx_th.wr_chan_flags &= ~CHAN_HT;
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
if (sc->sc_rx_th.wr_rate & IEEE80211_RATE_MCS) { /* HT rate */
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 ieee80211com *ic = ifp->if_l2com;
correct rx radiotap channel flags construction for 11n frames
2008-05-29 00:14:35 +00:00
if ((rs->rs_flags & HAL_RX_2040) == 0)
sc->sc_rx_th.wr_chan_flags |= CHAN_HT20;
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_HT40U(ic->ic_curchan))
correct rx radiotap channel flags construction for 11n frames
2008-05-29 00:14:35 +00:00
sc->sc_rx_th.wr_chan_flags |= CHAN_HT40U;
else
sc->sc_rx_th.wr_chan_flags |= CHAN_HT40D;
if ((rs->rs_flags & HAL_RX_GI) == 0)
sc->sc_rx_th.wr_flags |= IEEE80211_RADIOTAP_F_SHORTGI;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
}
#endif
sc->sc_rx_th.wr_tsf = htole64(ath_extend_tsf(rs->rs_tstamp, tsf));
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
if (rs->rs_status & HAL_RXERR_CRC)
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
sc->sc_rx_th.wr_flags |= IEEE80211_RADIOTAP_F_BADFCS;
/* XXX propagate other error flags from descriptor */
sc->sc_rx_th.wr_antnoise = nf;
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
sc->sc_rx_th.wr_antsignal = nf + rs->rs_rssi;
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
sc->sc_rx_th.wr_antenna = rs->rs_antenna;
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
#undef CHAN_HT
correct rx radiotap channel flags construction for 11n frames
2008-05-29 00:14:35 +00:00
#undef CHAN_HT20
#undef CHAN_HT40U
#undef CHAN_HT40D
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
static void
ath_handle_micerror(struct ieee80211com *ic,
struct ieee80211_frame *wh, int keyix)
{
struct ieee80211_node *ni;
/* XXX recheck MIC to deal w/ chips that lie */
/* XXX discard MIC errors on !data frames */
ni = ieee80211_find_rxnode(ic, (const struct ieee80211_frame_min *) wh);
if (ni != NULL) {
ieee80211_notify_michael_failure(ni->ni_vap, wh, keyix);
ieee80211_free_node(ni);
}
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void
ath_rx_proc(void *arg, int npending)
{
o track API change for HAL v0.9.6.1 o fix race condition when processing rx descriptors: because we use a self-linked descriptor at the end of the rx descriptor list to avoid rx overruns (which can easily happen for 5212 parts that enable PHY errors) we must carefully check that a descriptor is "done" by looking ahead to the next descriptor before believing the done bit in the current descriptor (this is all handled in the HAL since the rx descriptor format is chip-specific so we need to pass in two additional parameters--the physical address of the current descriptor and the virtual address of the next descriptor in the list) o check copyout return status for SIOCGATHSTATS ioctl Approved by: re (scottl)
2003-11-29 01:23:59 +00:00
#define PA2DESC(_sc, _pa) \
Update with last year of work.
2004-12-08 17:34:36 +00:00
((struct ath_desc *)((caddr_t)(_sc)->sc_rxdma.dd_desc + \
((_pa) - (_sc)->sc_rxdma.dd_desc_paddr)))
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ath_softc *sc = arg;
struct ath_buf *bf;
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211com *ic = ifp->if_l2com;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ath_hal *ah = sc->sc_ah;
struct ath_desc *ds;
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
struct ath_rx_status *rs;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct mbuf *m;
MFp4 changes to fix locking issues and correct reference count handling of station entries in hostap mode: Input path: o driver is now expected to find the node associated with the sender of a received frame; use ic_bss if none is located o driver passes the (referenced) node into ieee80211_input for use within the wlan module and is responsible for cleaning up on return o the antenna state is no longer passed up with each frame; this is now considered driver-private state and drivers are responsible for keeping it in the driver-private part of a node Output path: Revamp output path for management frames to eliminate redundant locking that causes problems and to correct reference counting bogosity that occurs when stations are timed out due to inactivity (in AP mode). On output the refcnt'd node is stashed in the pkthdr's recvif field (yech) and retrieved by the driver. This eliminates an unref/ref scenario and related node table unlock/lock due to the driver looking up the node. This is particularly important when stations are timed out as this causes a lock order reversal that can result in a deadlock. As a byproduct we also reduce the overhead for sending management frames (minimal). Additional fallout from this is a change to ieee80211_encap to return a refcn't node for tieing to the outbound frame. Node refcnts are not reclaimed until after a frame is completely processed (e.g. in the tx interrupt handler). This is especially important for timed out stations as this deref will be the final one causing the node entry to be reclaimed. Additional semi-related changes: o replace m_copym use with m_copypacket (optimization) o add assert to verify ic_bss is never free'd during normal operation o add comments explaining calling conventions by drivers for frames going in each direction o remove extraneous code that "cannot be executed" (e.g. because pointers may never be null)
2003-08-19 22:17:04 +00:00
struct ieee80211_node *ni;
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
int len, type, ngood;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
HAL_STATUS status;
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
int16_t nf;
u_int64_t tsf;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_RX_PROC, "%s: pending %u\n", __func__, npending);
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
ngood = 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
nf = ath_hal_getchannoise(ah, sc->sc_curchan);
Fixup statistics: o update tx rssi data only when an ACK was received o return tx rssi from sampled data instead of the last frame o track noise floor o return rx rssi and noise floor (was broken)
2008-10-27 17:51:24 +00:00
sc->sc_stats.ast_rx_noise = nf;
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
tsf = ath_hal_gettsf64(ah);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
do {
Update with last year of work.
2004-12-08 17:34:36 +00:00
bf = STAILQ_FIRST(&sc->sc_rxbuf);
Add a HAL capability bit for supporting self-linked RX descriptors and disable it for the 11n chipsets. From the ath9k source: == 11N: we can no longer afford to self link the last descriptor. MAC acknowledges BA status as long as it copies frames to host buffer (or rx fifo). This can incorrectly acknowledge packets to a sender if last desc is self-linked. == Since this is useful for pre-AR5416 chips that communicate PHY errors via error frames rather than by on-chip counters, leave the support in there, but disable it for AR5416 and later.
2011-04-04 14:52:31 +00:00
if (sc->sc_rxslink && bf == NULL) { /* NB: shouldn't happen */
Update with last year of work.
2004-12-08 17:34:36 +00:00
if_printf(ifp, "%s: no buffer!\n", __func__);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
break;
Add a HAL capability bit for supporting self-linked RX descriptors and disable it for the 11n chipsets. From the ath9k source: == 11N: we can no longer afford to self link the last descriptor. MAC acknowledges BA status as long as it copies frames to host buffer (or rx fifo). This can incorrectly acknowledge packets to a sender if last desc is self-linked. == Since this is useful for pre-AR5416 chips that communicate PHY errors via error frames rather than by on-chip counters, leave the support in there, but disable it for AR5416 and later.
2011-04-04 14:52:31 +00:00
} else if (bf == NULL) {
/*
* End of List:
* this can happen for non-self-linked RX chains
*/
sc->sc_stats.ast_rx_hitqueueend++;
break;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
correct handling of mbuf allocation failure when replenishing the rx list (leave a printf for the moment, need to make a debug msg) Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:29:23 +00:00
m = bf->bf_m;
if (m == NULL) { /* NB: shouldn't happen */
/*
* If mbuf allocation failed previously there
* will be no mbuf; try again to re-populate it.
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
*/
correct handling of mbuf allocation failure when replenishing the rx list (leave a printf for the moment, need to make a debug msg) Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:29:23 +00:00
/* XXX make debug msg */
if_printf(ifp, "%s: no mbuf!\n", __func__);
STAILQ_REMOVE_HEAD(&sc->sc_rxbuf, bf_list);
goto rx_next;
}
terminate the rx descriptor list with a self-linked entry so high phy error rates on a 5212 don't cause rx overruns
2003-10-22 04:37:34 +00:00
ds = bf->bf_desc;
if (ds->ds_link == bf->bf_daddr) {
/* NB: never process the self-linked entry at the end */
According to ath9k recv.c, one shouldn't be doing self-linked descriptors in the RX path when doing 11n and block-ack'ed frames. Apparently, the MAC will loop over that self-linked descriptor and treat it as "good enough" for (incorrectly!) ACKing the frames in the block-ack. Until I figure out how to work around this issue in the future, this counter will tell me if packet RX processing ever gets to the point where it's touching the self-linked descriptor. If there's ever enough packets to get to that point, BA's will be invalid and likely very unhappy.
2011-03-29 15:59:07 +00:00
sc->sc_stats.ast_rx_hitqueueend++;
terminate the rx descriptor list with a self-linked entry so high phy error rates on a 5212 don't cause rx overruns
2003-10-22 04:37:34 +00:00
break;
}
o track API change for HAL v0.9.6.1 o fix race condition when processing rx descriptors: because we use a self-linked descriptor at the end of the rx descriptor list to avoid rx overruns (which can easily happen for 5212 parts that enable PHY errors) we must carefully check that a descriptor is "done" by looking ahead to the next descriptor before believing the done bit in the current descriptor (this is all handled in the HAL since the rx descriptor format is chip-specific so we need to pass in two additional parameters--the physical address of the current descriptor and the virtual address of the next descriptor in the list) o check copyout return status for SIOCGATHSTATS ioctl Approved by: re (scottl)
2003-11-29 01:23:59 +00:00
/* XXX sync descriptor memory */
/*
* Must provide the virtual address of the current
* descriptor, the physical address, and the virtual
* address of the next descriptor in the h/w chain.
* This allows the HAL to look ahead to see if the
* hardware is done with a descriptor by checking the
* done bit in the following descriptor and the address
* of the current descriptor the DMA engine is working
* on. All this is necessary because of our use of
* a self-linked list to avoid rx overruns.
*/
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
rs = &bf->bf_status.ds_rxstat;
o track API change for HAL v0.9.6.1 o fix race condition when processing rx descriptors: because we use a self-linked descriptor at the end of the rx descriptor list to avoid rx overruns (which can easily happen for 5212 parts that enable PHY errors) we must carefully check that a descriptor is "done" by looking ahead to the next descriptor before believing the done bit in the current descriptor (this is all handled in the HAL since the rx descriptor format is chip-specific so we need to pass in two additional parameters--the physical address of the current descriptor and the virtual address of the next descriptor in the list) o check copyout return status for SIOCGATHSTATS ioctl Approved by: re (scottl)
2003-11-29 01:23:59 +00:00
status = ath_hal_rxprocdesc(ah, ds,
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
bf->bf_daddr, PA2DESC(sc, ds->ds_link), rs);
o add opt_ath.h enable tweaking various config parameters for the driver without modifying the source code o default debug msgs and diag support to off MFC after: 3 days
2006-04-03 18:14:02 +00:00
#ifdef ATH_DEBUG
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (sc->sc_debug & ATH_DEBUG_RECV_DESC)
print the extended tx/rx descriptor for 5416 and later parts
2008-11-24 01:34:56 +00:00
ath_printrxbuf(sc, bf, 0, status == HAL_OK);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
#endif
if (status == HAL_EINPROGRESS)
break;
Update with last year of work.
2004-12-08 17:34:36 +00:00
STAILQ_REMOVE_HEAD(&sc->sc_rxbuf, bf_list);
Some statistics additions - prepare for error codes > 32 (since the AR5416 error mask is > 5 bits) and add some extra CRC/HT40/ShortGI counters to help debug 802.11n issues.
2011-02-14 21:24:54 +00:00
/* These aren't specifically errors */
if (rs->rs_flags & HAL_RX_GI)
sc->sc_stats.ast_rx_halfgi++;
if (rs->rs_flags & HAL_RX_2040)
sc->sc_stats.ast_rx_2040++;
if (rs->rs_flags & HAL_RX_DELIM_CRC_PRE)
sc->sc_stats.ast_rx_pre_crc_err++;
if (rs->rs_flags & HAL_RX_DELIM_CRC_POST)
sc->sc_stats.ast_rx_post_crc_err++;
if (rs->rs_flags & HAL_RX_DECRYPT_BUSY)
sc->sc_stats.ast_rx_decrypt_busy_err++;
if (rs->rs_flags & HAL_RX_HI_RX_CHAIN)
sc->sc_stats.ast_rx_hi_rx_chain++;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
if (rs->rs_status != 0) {
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
if (rs->rs_status & HAL_RXERR_CRC)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_stats.ast_rx_crcerr++;
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
if (rs->rs_status & HAL_RXERR_FIFO)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_stats.ast_rx_fifoerr++;
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
if (rs->rs_status & HAL_RXERR_PHY) {
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_stats.ast_rx_phyerr++;
Some statistics additions - prepare for error codes > 32 (since the AR5416 error mask is > 5 bits) and add some extra CRC/HT40/ShortGI counters to help debug 802.11n issues.
2011-02-14 21:24:54 +00:00
/* Be suitably paranoid about receiving phy errors out of the stats array bounds */
if (rs->rs_phyerr < 64)
sc->sc_stats.ast_rx_phy[rs->rs_phyerr]++;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
goto rx_error; /* NB: don't count in ierrors */
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
if (rs->rs_status & HAL_RXERR_DECRYPT) {
Don't count PHY errors as input errors. This is important for 5212-based devices because PHY errors are used to collect data on environmental noise that and doesn't truly reflect the state of the communications media. The result is confused users. Folks that want to watch PHY errors can still get the statistics through the device ioctl (used by athstats).
2003-11-13 05:35:07 +00:00
/*
Update with last year of work.
2004-12-08 17:34:36 +00:00
* Decrypt error. If the error occurred
* because there was no hardware key, then
* let the frame through so the upper layers
* can process it. This is necessary for 5210
* parts which have no way to setup a ``clear''
* key cache entry.
*
* XXX do key cache faulting
Don't count PHY errors as input errors. This is important for 5212-based devices because PHY errors are used to collect data on environmental noise that and doesn't truly reflect the state of the communications media. The result is confused users. Folks that want to watch PHY errors can still get the statistics through the device ioctl (used by athstats).
2003-11-13 05:35:07 +00:00
*/
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
if (rs->rs_keyix == HAL_RXKEYIX_INVALID)
Update with last year of work.
2004-12-08 17:34:36 +00:00
goto rx_accept;
sc->sc_stats.ast_rx_badcrypt++;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
if (rs->rs_status & HAL_RXERR_MIC) {
Update with last year of work.
2004-12-08 17:34:36 +00:00
sc->sc_stats.ast_rx_badmic++;
/*
* Do minimal work required to hand off
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
* the 802.11 header for notification.
Update with last year of work.
2004-12-08 17:34:36 +00:00
*/
/* XXX frag's and qos frames */
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
len = rs->rs_datalen;
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (len >= sizeof (struct ieee80211_frame)) {
bus_dmamap_sync(sc->sc_dmat,
bf->bf_dmamap,
BUS_DMASYNC_POSTREAD);
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
ath_handle_micerror(ic,
Update with last year of work.
2004-12-08 17:34:36 +00:00
mtod(m, struct ieee80211_frame *),
for parts that require split keycache entries report the the index of the first entry on a mic error so we're consistent with parts that don't have split keycache
2004-12-31 20:21:31 +00:00
sc->sc_splitmic ?
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
rs->rs_keyix-32 : rs->rs_keyix);
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
}
ifp->if_ierrors++;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
rx_error:
/*
* Cleanup any pending partial frame.
*/
if (sc->sc_rxpending != NULL) {
m_freem(sc->sc_rxpending);
sc->sc_rxpending = NULL;
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
* When a tap is present pass error frames
* that have been requested. By default we
* pass decrypt+mic errors but others may be
* interesting (e.g. crc).
Update with last year of work.
2004-12-08 17:34:36 +00:00
*/
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
if (ieee80211_radiotap_active(ic) &&
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
(rs->rs_status & sc->sc_monpass)) {
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
bus_dmamap_sync(sc->sc_dmat, bf->bf_dmamap,
BUS_DMASYNC_POSTREAD);
/* NB: bpf needs the mbuf length setup */
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
len = rs->rs_datalen;
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
m->m_pkthdr.len = m->m_len = len;
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
ath_rx_tap(ifp, m, rs, tsf, nf);
ieee80211_radiotap_rx_all(ic, m);
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
}
/* XXX pass MIC errors up for s/w reclaculation */
goto rx_next;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
rx_accept:
/*
* Sync and unmap the frame. At this point we're
* committed to passing the mbuf somewhere so clear
fixup comments
2006-02-15 18:15:49 +00:00
* bf_m; this means a new mbuf must be allocated
Update with last year of work.
2004-12-08 17:34:36 +00:00
* when the rx descriptor is setup again to receive
* another frame.
*/
bus_dmamap_sync(sc->sc_dmat, bf->bf_dmamap,
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
BUS_DMASYNC_POSTREAD);
bus_dmamap_unload(sc->sc_dmat, bf->bf_dmamap);
bf->bf_m = NULL;
Update with last year of work.
2004-12-08 17:34:36 +00:00
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
len = rs->rs_datalen;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
m->m_len = len;
if (rs->rs_more) {
/*
* Frame spans multiple descriptors; save
* it for the next completed descriptor, it
* will be used to construct a jumbogram.
*/
if (sc->sc_rxpending != NULL) {
/* NB: max frame size is currently 2 clusters */
sc->sc_stats.ast_rx_toobig++;
m_freem(sc->sc_rxpending);
}
m->m_pkthdr.rcvif = ifp;
m->m_pkthdr.len = len;
sc->sc_rxpending = m;
goto rx_next;
} else if (sc->sc_rxpending != NULL) {
/*
* This is the second part of a jumbogram,
* chain it to the first mbuf, adjust the
* frame length, and clear the rxpending state.
*/
sc->sc_rxpending->m_next = m;
sc->sc_rxpending->m_pkthdr.len += len;
m = sc->sc_rxpending;
sc->sc_rxpending = NULL;
} else {
/*
* Normal single-descriptor receive; setup
* the rcvif and packet length.
*/
m->m_pkthdr.rcvif = ifp;
m->m_pkthdr.len = len;
}
Add support for the experimental radiotap capture format. With this we no longer need the debugging code to dump packets.
2003-09-05 22:22:49 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ifp->if_ipackets++;
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
sc->sc_stats.ast_ant_rx[rs->rs_antenna]++;
Update with last year of work.
2004-12-08 17:34:36 +00:00
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
/*
* Populate the rx status block. When there are bpf
* listeners we do the additional work to provide
* complete status. Otherwise we fill in only the
* material required by ieee80211_input. Note that
* noise setting is filled in above.
*/
if (ieee80211_radiotap_active(ic))
ath_rx_tap(ifp, m, rs, tsf, nf);
MFp4 changes to fix locking issues and correct reference count handling of station entries in hostap mode: Input path: o driver is now expected to find the node associated with the sender of a received frame; use ic_bss if none is located o driver passes the (referenced) node into ieee80211_input for use within the wlan module and is responsible for cleaning up on return o the antenna state is no longer passed up with each frame; this is now considered driver-private state and drivers are responsible for keeping it in the driver-private part of a node Output path: Revamp output path for management frames to eliminate redundant locking that causes problems and to correct reference counting bogosity that occurs when stations are timed out due to inactivity (in AP mode). On output the refcnt'd node is stashed in the pkthdr's recvif field (yech) and retrieved by the driver. This eliminates an unref/ref scenario and related node table unlock/lock due to the driver looking up the node. This is particularly important when stations are timed out as this causes a lock order reversal that can result in a deadlock. As a byproduct we also reduce the overhead for sending management frames (minimal). Additional fallout from this is a change to ieee80211_encap to return a refcn't node for tieing to the outbound frame. Node refcnts are not reclaimed until after a frame is completely processed (e.g. in the tx interrupt handler). This is especially important for timed out stations as this deref will be the final one causing the node entry to be reclaimed. Additional semi-related changes: o replace m_copym use with m_copypacket (optimization) o add assert to verify ic_bss is never free'd during normal operation o add comments explaining calling conventions by drivers for frames going in each direction o remove extraneous code that "cannot be executed" (e.g. because pointers may never be null)
2003-08-19 22:17:04 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* From this point on we assume the frame is at least
* as large as ieee80211_frame_min; verify that.
*/
if (len < IEEE80211_MIN_LEN) {
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
if (!ieee80211_radiotap_active(ic)) {
DPRINTF(sc, ATH_DEBUG_RECV,
"%s: short packet %d\n", __func__, len);
sc->sc_stats.ast_rx_tooshort++;
} else {
/* NB: in particular this captures ack's */
ieee80211_radiotap_rx_all(ic, m);
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
m_freem(m);
goto rx_next;
}
if (IFF_DUMPPKTS(sc, ATH_DEBUG_RECV)) {
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
const HAL_RATE_TABLE *rt = sc->sc_currates;
uint8_t rix = rt->rateCodeToIndex[rs->rs_rate];
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
ieee80211_dump_pkt(ic, mtod(m, caddr_t), len,
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
sc->sc_hwmap[rix].ieeerate, rs->rs_rssi);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
MFp4 changes to fix locking issues and correct reference count handling of station entries in hostap mode: Input path: o driver is now expected to find the node associated with the sender of a received frame; use ic_bss if none is located o driver passes the (referenced) node into ieee80211_input for use within the wlan module and is responsible for cleaning up on return o the antenna state is no longer passed up with each frame; this is now considered driver-private state and drivers are responsible for keeping it in the driver-private part of a node Output path: Revamp output path for management frames to eliminate redundant locking that causes problems and to correct reference counting bogosity that occurs when stations are timed out due to inactivity (in AP mode). On output the refcnt'd node is stashed in the pkthdr's recvif field (yech) and retrieved by the driver. This eliminates an unref/ref scenario and related node table unlock/lock due to the driver looking up the node. This is particularly important when stations are timed out as this causes a lock order reversal that can result in a deadlock. As a byproduct we also reduce the overhead for sending management frames (minimal). Additional fallout from this is a change to ieee80211_encap to return a refcn't node for tieing to the outbound frame. Node refcnts are not reclaimed until after a frame is completely processed (e.g. in the tx interrupt handler). This is especially important for timed out stations as this deref will be the final one causing the node entry to be reclaimed. Additional semi-related changes: o replace m_copym use with m_copypacket (optimization) o add assert to verify ic_bss is never free'd during normal operation o add comments explaining calling conventions by drivers for frames going in each direction o remove extraneous code that "cannot be executed" (e.g. because pointers may never be null)
2003-08-19 22:17:04 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
m_adj(m, -IEEE80211_CRC_LEN);
MFp4 changes to fix locking issues and correct reference count handling of station entries in hostap mode: Input path: o driver is now expected to find the node associated with the sender of a received frame; use ic_bss if none is located o driver passes the (referenced) node into ieee80211_input for use within the wlan module and is responsible for cleaning up on return o the antenna state is no longer passed up with each frame; this is now considered driver-private state and drivers are responsible for keeping it in the driver-private part of a node Output path: Revamp output path for management frames to eliminate redundant locking that causes problems and to correct reference counting bogosity that occurs when stations are timed out due to inactivity (in AP mode). On output the refcnt'd node is stashed in the pkthdr's recvif field (yech) and retrieved by the driver. This eliminates an unref/ref scenario and related node table unlock/lock due to the driver looking up the node. This is particularly important when stations are timed out as this causes a lock order reversal that can result in a deadlock. As a byproduct we also reduce the overhead for sending management frames (minimal). Additional fallout from this is a change to ieee80211_encap to return a refcn't node for tieing to the outbound frame. Node refcnts are not reclaimed until after a frame is completely processed (e.g. in the tx interrupt handler). This is especially important for timed out stations as this deref will be the final one causing the node entry to be reclaimed. Additional semi-related changes: o replace m_copym use with m_copypacket (optimization) o add assert to verify ic_bss is never free'd during normal operation o add comments explaining calling conventions by drivers for frames going in each direction o remove extraneous code that "cannot be executed" (e.g. because pointers may never be null)
2003-08-19 22:17:04 +00:00
/*
Update with last year of work.
2004-12-08 17:34:36 +00:00
* Locate the node for sender, track state, and then
* pass the (referenced) node up to the 802.11 layer
Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks
2005-08-08 18:46:36 +00:00
* for its use.
Maintain a history of data associated with received frames and use this to calculate smoothed signal quality data for each node. o add a 16-deep history buffer to each driver-private node storage that holds rssi and antenna info for received frames o override the default per-node "get rssi" method to return an average rssi value based on samples collected over the last second o enable beacon reception so even idle systems maintain a running history of signal quality This data may also be useful for improving the rate control algorithm. Based on work by Tom Marshall <tommy@home.tig-grr.com> for MADWIFI.
2003-09-15 22:34:46 +00:00
*/
Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks
2005-08-08 18:46:36 +00:00
ni = ieee80211_find_rxnode_withkey(ic,
mtod(m, const struct ieee80211_frame_min *),
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
rs->rs_keyix == HAL_RXKEYIX_INVALID ?
IEEE80211_KEYIX_NONE : rs->rs_keyix);
sc_lastrs is also used in case the sending station is not known, for example in a split IBSS scenario. Therefore always assign sc_lastrs. This removes a hack I committed in r206457. Approved by: rpaulo (mentor)
2010-06-14 08:24:00 +00:00
sc->sc_lastrs = rs;
Just tag all RX packets as needing reorder processing for now. This fixes two problems - * All packets need to be processed here, not just aggregate ones - as any received frames (AMPDU or otherwise) in the given TID (traffic class id) will update the sequence number and, implied with that, update the window; * It seems there's situations where packets aren't matching a current node but somehow need to be tracked. Thus just tag them all for now; I'll figure out the why later. Whilst I'm here, bump the stats counters whilst I'm at it. This fixes AMPDU RX in my tests; the main problems now stem from what look like PHY level error/retransmits which are impeding general throughput, incl. AMPDU.
2011-02-06 20:11:24 +00:00
/* Keep statistics on the number of aggregate packets received */
if (rs->rs_isaggr)
sc->sc_stats.ast_rx_agg++;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (ni != NULL) {
net80211 really doesn't want A_MPDU to appear on non-11n station node mbufs. Revert back to the previous method of doing it for where a node can be identified and it's an 11n node. I'll have to do some further research into exactly what is being messed up with the sequence number matching and I'll then revisit this.
2011-02-08 20:16:43 +00:00
/*
* Only punt packets for ampdu reorder processing for 11n nodes;
* net80211 enforces that M_AMPDU is only set for 11n nodes.
*/
if (ni->ni_flags & IEEE80211_NODE_HT)
m->m_flags |= M_AMPDU;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/*
* Sending station is known, dispatch directly.
*/
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
type = ieee80211_input(ni, m, rs->rs_rssi, nf);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ieee80211_free_node(ni);
/*
* Arrange to update the last rx timestamp only for
* frames from our ap when operating in station mode.
* This assumes the rx key is always setup when
* associated.
*/
if (ic->ic_opmode == IEEE80211_M_STA &&
rs->rs_keyix != HAL_RXKEYIX_INVALID)
ngood++;
} else {
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
type = ieee80211_input_all(ic, m, rs->rs_rssi, nf);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
}
Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks
2005-08-08 18:46:36 +00:00
/*
* Track rx rssi and do any rx antenna management.
*/
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
ATH_RSSI_LPF(sc->sc_halstats.ns_avgrssi, rs->rs_rssi);
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (sc->sc_diversity) {
/*
* When using fast diversity, change the default rx
* antenna if diversity chooses the other antenna 3
* times in a row.
*/
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
if (sc->sc_defant != rs->rs_antenna) {
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (++sc->sc_rxotherant >= 3)
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
ath_setdefantenna(sc, rs->rs_antenna);
Update with last year of work.
2004-12-08 17:34:36 +00:00
} else
sc->sc_rxotherant = 0;
}
better led blinking
2005-01-18 19:03:04 +00:00
if (sc->sc_softled) {
/*
* Blink for any data frame. Otherwise do a
* heartbeat-style blink when idle. The latter
* is mainly for station mode where we depend on
* periodic beacon frames to trigger the poll event.
*/
use frame type returned by ieee80211_input to drive softled code instead of monitoring the input packet count
2005-04-04 02:34:15 +00:00
if (type == IEEE80211_FC0_TYPE_DATA) {
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
const HAL_RATE_TABLE *rt = sc->sc_currates;
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
ath_led_event(sc,
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
rt->rateCodeToIndex[rs->rs_rate]);
better led blinking
2005-01-18 19:03:04 +00:00
} else if (ticks - sc->sc_ledevent >= sc->sc_ledidle)
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
ath_led_event(sc, 0);
better led blinking
2005-01-18 19:03:04 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
rx_next:
STAILQ_INSERT_TAIL(&sc->sc_rxbuf, bf, bf_list);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
} while (ath_rxbuf_init(sc, bf) == 0);
Update with last year of work.
2004-12-08 17:34:36 +00:00
/* rx signal state monitoring */
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
ath_hal_rxmonitor(ah, &sc->sc_halstats, sc->sc_curchan);
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
if (ngood)
sc->sc_lastrx = tsf;
Pickup Giant in ath_rx_proc and when handling a beacon miss in order to satisfy the assertion in if_start.
2004-08-07 00:45:05 +00:00
Hoist 802.11 encapsulation up into net80211: o call ieee80211_encap in ieee80211_start so frames passed down to drivers are already encapsulated o remove ieee80211_encap calls in drivers o fixup wi so it recreates the 802.3 head it requires from the 802.11 header contents o move fast-frame aggregation from ath to net80211 (conditional on IEEE80211_SUPPORT_SUPERG): - aggregation is now done in ieee80211_start; it is enabled when the packets/sec exceeds ieee80211_ffppsmin (net.wlan.ffppsmin) and frames are held on a staging queue according to ieee80211_ffagemax (net.wlan.ffagemax) to wait for a frame to combine with - drivers must call back to age/flush the staging queue (ath does this on tx done, at swba, and on rx according to the state of the tx queues and/or the contents of the staging queue) - remove fast-frame-related data structures from ath - add ieee80211_ff_node_init and ieee80211_ff_node_cleanup to handle per-node fast-frames state (we reuse 11n tx ampdu state) o change ieee80211_encap calling convention to include an explicit vap so frames coming through a WDS vap are recognized w/o setting M_WDS With these changes any device able to tx/rx 3Kbyte+ frames can use fast-frames. Reviewed by: thompsa, rpaulo, avatar, imp, sephe
2009-03-30 21:53:27 +00:00
if ((ifp->if_drv_flags & IFF_DRV_OACTIVE) == 0) {
#ifdef IEEE80211_SUPPORT_SUPERG
make superg/fast-frames state dynamically-allocated (and indirect off the com structure instead of embedded); this reduces the overhead when not configured and reduces visibility of the contents
2009-05-02 20:16:55 +00:00
ieee80211_ff_age_all(ic, 100);
Hoist 802.11 encapsulation up into net80211: o call ieee80211_encap in ieee80211_start so frames passed down to drivers are already encapsulated o remove ieee80211_encap calls in drivers o fixup wi so it recreates the 802.3 head it requires from the 802.11 header contents o move fast-frame aggregation from ath to net80211 (conditional on IEEE80211_SUPPORT_SUPERG): - aggregation is now done in ieee80211_start; it is enabled when the packets/sec exceeds ieee80211_ffppsmin (net.wlan.ffppsmin) and frames are held on a staging queue according to ieee80211_ffagemax (net.wlan.ffagemax) to wait for a frame to combine with - drivers must call back to age/flush the staging queue (ath does this on tx done, at swba, and on rx according to the state of the tx queues and/or the contents of the staging queue) - remove fast-frame-related data structures from ath - add ieee80211_ff_node_init and ieee80211_ff_node_cleanup to handle per-node fast-frames state (we reuse 11n tx ampdu state) o change ieee80211_encap calling convention to include an explicit vap so frames coming through a WDS vap are recognized w/o setting M_WDS With these changes any device able to tx/rx 3Kbyte+ frames can use fast-frames. Reviewed by: thompsa, rpaulo, avatar, imp, sephe
2009-03-30 21:53:27 +00:00
#endif
if (!IFQ_IS_EMPTY(&ifp->if_snd))
ath_start(ifp);
}
o track API change for HAL v0.9.6.1 o fix race condition when processing rx descriptors: because we use a self-linked descriptor at the end of the rx descriptor list to avoid rx overruns (which can easily happen for 5212 parts that enable PHY errors) we must carefully check that a descriptor is "done" by looking ahead to the next descriptor before believing the done bit in the current descriptor (this is all handled in the HAL since the rx descriptor format is chip-specific so we need to pass in two additional parameters--the physical address of the current descriptor and the virtual address of the next descriptor in the list) o check copyout return status for SIOCGATHSTATS ioctl Approved by: re (scottl)
2003-11-29 01:23:59 +00:00
#undef PA2DESC
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Close race in handling mcast traffic when operating as an ap with stations in power save: add a new q where mcast frames are stashed and on beacon update (at DTIM) move frames from the mcast q to the cabq and start it. This ensures the cabq is only manipulated in one place. Sponsored by: Hobnob MFC after: 2 weeks
2006-06-26 03:10:45 +00:00
static void
ath_txq_init(struct ath_softc *sc, struct ath_txq *txq, int qnum)
{
txq->axq_qnum = qnum;
Hoist 802.11 encapsulation up into net80211: o call ieee80211_encap in ieee80211_start so frames passed down to drivers are already encapsulated o remove ieee80211_encap calls in drivers o fixup wi so it recreates the 802.3 head it requires from the 802.11 header contents o move fast-frame aggregation from ath to net80211 (conditional on IEEE80211_SUPPORT_SUPERG): - aggregation is now done in ieee80211_start; it is enabled when the packets/sec exceeds ieee80211_ffppsmin (net.wlan.ffppsmin) and frames are held on a staging queue according to ieee80211_ffagemax (net.wlan.ffagemax) to wait for a frame to combine with - drivers must call back to age/flush the staging queue (ath does this on tx done, at swba, and on rx according to the state of the tx queues and/or the contents of the staging queue) - remove fast-frame-related data structures from ath - add ieee80211_ff_node_init and ieee80211_ff_node_cleanup to handle per-node fast-frames state (we reuse 11n tx ampdu state) o change ieee80211_encap calling convention to include an explicit vap so frames coming through a WDS vap are recognized w/o setting M_WDS With these changes any device able to tx/rx 3Kbyte+ frames can use fast-frames. Reviewed by: thompsa, rpaulo, avatar, imp, sephe
2009-03-30 21:53:27 +00:00
txq->axq_ac = 0;
Close race in handling mcast traffic when operating as an ap with stations in power save: add a new q where mcast frames are stashed and on beacon update (at DTIM) move frames from the mcast q to the cabq and start it. This ensures the cabq is only manipulated in one place. Sponsored by: Hobnob MFC after: 2 weeks
2006-06-26 03:10:45 +00:00
txq->axq_depth = 0;
txq->axq_intrcnt = 0;
txq->axq_link = NULL;
STAILQ_INIT(&txq->axq_q);
ATH_TXQ_LOCK_INIT(sc, txq);
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/*
Update with last year of work.
2004-12-08 17:34:36 +00:00
* Setup a h/w transmit queue.
*/
static struct ath_txq *
ath_txq_setup(struct ath_softc *sc, int qtype, int subtype)
{
#define N(a) (sizeof(a)/sizeof(a[0]))
struct ath_hal *ah = sc->sc_ah;
HAL_TXQ_INFO qi;
int qnum;
memset(&qi, 0, sizeof(qi));
qi.tqi_subtype = subtype;
qi.tqi_aifs = HAL_TXQ_USEDEFAULT;
qi.tqi_cwmin = HAL_TXQ_USEDEFAULT;
qi.tqi_cwmax = HAL_TXQ_USEDEFAULT;
/*
* Enable interrupts only for EOL and DESC conditions.
* We mark tx descriptors to receive a DESC interrupt
* when a tx queue gets deep; otherwise waiting for the
* EOL to reap descriptors. Note that this is done to
* reduce interrupt load and this only defers reaping
* descriptors, never transmitting frames. Aside from
* reducing interrupts this also permits more concurrency.
* The only potential downside is if the tx queue backs
* up in which case the top half of the kernel may backup
* due to a lack of tx descriptors.
*/
Update for rev 0.9.16.16 hal: o add dfs+radar hooks; DFS is presently disabled in the hal o channel and mode handling changes o various api changes o be more aggressive about iq calibration settling so ap mode operation is better immediately after startup o rfkill/rfsilent sysctl support o tpc ack/cts sysctl support MFC after: 2 weeks
2006-02-10 19:07:08 +00:00
qi.tqi_qflags = HAL_TXQ_TXEOLINT_ENABLE | HAL_TXQ_TXDESCINT_ENABLE;
Update with last year of work.
2004-12-08 17:34:36 +00:00
qnum = ath_hal_setuptxqueue(ah, qtype, &qi);
if (qnum == -1) {
/*
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
* NB: don't print a message, this happens
correct some typos Submitted by: Tai-hwa Liang
2004-12-31 20:35:05 +00:00
* normally on parts with too few tx queues
Update with last year of work.
2004-12-08 17:34:36 +00:00
*/
return NULL;
}
if (qnum >= N(sc->sc_txq)) {
Make this amd64-clean. sizeof is long on amd64, so things that do a printf of a sizeof, need to use %z to get the correct type on all our platforms. Also, convert integers<->pointers via uintptr_t. (I think Sam's instructions were for me to commit this. If I misunderstood, then I apologize in advance.)
2004-12-15 02:25:21 +00:00
device_printf(sc->sc_dev,
"hal qnum %u out of range, max %zu!\n",
Update with last year of work.
2004-12-08 17:34:36 +00:00
qnum, N(sc->sc_txq));
ath_hal_releasetxqueue(ah, qnum);
return NULL;
}
if (!ATH_TXQ_SETUP(sc, qnum)) {
Close race in handling mcast traffic when operating as an ap with stations in power save: add a new q where mcast frames are stashed and on beacon update (at DTIM) move frames from the mcast q to the cabq and start it. This ensures the cabq is only manipulated in one place. Sponsored by: Hobnob MFC after: 2 weeks
2006-06-26 03:10:45 +00:00
ath_txq_init(sc, &sc->sc_txq[qnum], qnum);
Update with last year of work.
2004-12-08 17:34:36 +00:00
sc->sc_txqsetup |= 1<<qnum;
}
return &sc->sc_txq[qnum];
#undef N
}
/*
* Setup a hardware data transmit queue for the specified
* access control. The hal may not support all requested
* queues in which case it will return a reference to a
* previously setup queue. We record the mapping from ac's
* to h/w queues for use by ath_tx_start and also track
* the set of h/w queues being used to optimize work in the
* transmit interrupt handler and related routines.
*/
static int
ath_tx_setup(struct ath_softc *sc, int ac, int haltype)
{
#define N(a) (sizeof(a)/sizeof(a[0]))
struct ath_txq *txq;
if (ac >= N(sc->sc_ac2q)) {
Make this amd64-clean. sizeof is long on amd64, so things that do a printf of a sizeof, need to use %z to get the correct type on all our platforms. Also, convert integers<->pointers via uintptr_t. (I think Sam's instructions were for me to commit this. If I misunderstood, then I apologize in advance.)
2004-12-15 02:25:21 +00:00
device_printf(sc->sc_dev, "AC %u out of range, max %zu!\n",
Update with last year of work.
2004-12-08 17:34:36 +00:00
ac, N(sc->sc_ac2q));
return 0;
}
txq = ath_txq_setup(sc, HAL_TX_QUEUE_DATA, haltype);
if (txq != NULL) {
Hoist 802.11 encapsulation up into net80211: o call ieee80211_encap in ieee80211_start so frames passed down to drivers are already encapsulated o remove ieee80211_encap calls in drivers o fixup wi so it recreates the 802.3 head it requires from the 802.11 header contents o move fast-frame aggregation from ath to net80211 (conditional on IEEE80211_SUPPORT_SUPERG): - aggregation is now done in ieee80211_start; it is enabled when the packets/sec exceeds ieee80211_ffppsmin (net.wlan.ffppsmin) and frames are held on a staging queue according to ieee80211_ffagemax (net.wlan.ffagemax) to wait for a frame to combine with - drivers must call back to age/flush the staging queue (ath does this on tx done, at swba, and on rx according to the state of the tx queues and/or the contents of the staging queue) - remove fast-frame-related data structures from ath - add ieee80211_ff_node_init and ieee80211_ff_node_cleanup to handle per-node fast-frames state (we reuse 11n tx ampdu state) o change ieee80211_encap calling convention to include an explicit vap so frames coming through a WDS vap are recognized w/o setting M_WDS With these changes any device able to tx/rx 3Kbyte+ frames can use fast-frames. Reviewed by: thompsa, rpaulo, avatar, imp, sephe
2009-03-30 21:53:27 +00:00
txq->axq_ac = ac;
Update with last year of work.
2004-12-08 17:34:36 +00:00
sc->sc_ac2q[ac] = txq;
return 1;
} else
return 0;
#undef N
}
/*
* Update WME parameters for a transmit queue.
*/
static int
ath_txq_update(struct ath_softc *sc, int ac)
{
#define ATH_EXPONENT_TO_VALUE(v) ((1<<v)-1)
#define ATH_TXOP_TO_US(v) (v<<5)
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ifnet *ifp = sc->sc_ifp;
struct ieee80211com *ic = ifp->if_l2com;
Update with last year of work.
2004-12-08 17:34:36 +00:00
struct ath_txq *txq = sc->sc_ac2q[ac];
struct wmeParams *wmep = &ic->ic_wme.wme_chanParams.cap_wmeParams[ac];
struct ath_hal *ah = sc->sc_ah;
HAL_TXQ_INFO qi;
ath_hal_gettxqueueprops(ah, txq->axq_qnum, &qi);
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
if (sc->sc_tdma) {
/*
* AIFS is zero so there's no pre-transmit wait. The
* burst time defines the slot duration and is configured
fix comment
2009-06-02 21:12:07 +00:00
* through net80211. The QCU is setup to not do post-xmit
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
* back off, lockout all lower-priority QCU's, and fire
* off the DMA beacon alert timer which is setup based
* on the slot configuration.
*/
qi.tqi_qflags = HAL_TXQ_TXOKINT_ENABLE
| HAL_TXQ_TXERRINT_ENABLE
| HAL_TXQ_TXURNINT_ENABLE
| HAL_TXQ_TXEOLINT_ENABLE
| HAL_TXQ_DBA_GATED
| HAL_TXQ_BACKOFF_DISABLE
| HAL_TXQ_ARB_LOCKOUT_GLOBAL
;
qi.tqi_aifs = 0;
/* XXX +dbaprep? */
qi.tqi_readyTime = sc->sc_tdmaslotlen;
qi.tqi_burstTime = qi.tqi_readyTime;
} else {
#endif
qi.tqi_qflags = HAL_TXQ_TXOKINT_ENABLE
| HAL_TXQ_TXERRINT_ENABLE
| HAL_TXQ_TXDESCINT_ENABLE
| HAL_TXQ_TXURNINT_ENABLE
;
qi.tqi_aifs = wmep->wmep_aifsn;
qi.tqi_cwmin = ATH_EXPONENT_TO_VALUE(wmep->wmep_logcwmin);
qi.tqi_cwmax = ATH_EXPONENT_TO_VALUE(wmep->wmep_logcwmax);
qi.tqi_readyTime = 0;
qi.tqi_burstTime = ATH_TXOP_TO_US(wmep->wmep_txopLimit);
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
}
#endif
DPRINTF(sc, ATH_DEBUG_RESET,
"%s: Q%u qflags 0x%x aifs %u cwmin %u cwmax %u burstTime %u\n",
__func__, txq->axq_qnum, qi.tqi_qflags,
qi.tqi_aifs, qi.tqi_cwmin, qi.tqi_cwmax, qi.tqi_burstTime);
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (!ath_hal_settxqueueprops(ah, txq->axq_qnum, &qi)) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if_printf(ifp, "unable to update hardware queue "
Update with last year of work.
2004-12-08 17:34:36 +00:00
"parameters for %s traffic!\n",
ieee80211_wme_acnames[ac]);
return 0;
} else {
ath_hal_resettxqueue(ah, txq->axq_qnum); /* push to h/w */
return 1;
}
#undef ATH_TXOP_TO_US
#undef ATH_EXPONENT_TO_VALUE
}
/*
* Callback from the 802.11 layer to update WME parameters.
*/
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
static int
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_wme_update(struct ieee80211com *ic)
{
struct ath_softc *sc = ic->ic_ifp->if_softc;
return !ath_txq_update(sc, WME_AC_BE) ||
!ath_txq_update(sc, WME_AC_BK) ||
!ath_txq_update(sc, WME_AC_VI) ||
!ath_txq_update(sc, WME_AC_VO) ? EIO : 0;
}
/*
* Reclaim resources for a setup queue.
*/
static void
ath_tx_cleanupq(struct ath_softc *sc, struct ath_txq *txq)
{
ath_hal_releasetxqueue(sc->sc_ah, txq->axq_qnum);
ATH_TXQ_LOCK_DESTROY(txq);
sc->sc_txqsetup &= ~(1<<txq->axq_qnum);
}
/*
* Reclaim all tx queue resources.
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
*/
Update with last year of work.
2004-12-08 17:34:36 +00:00
static void
ath_tx_cleanup(struct ath_softc *sc)
{
int i;
ATH_TXBUF_LOCK_DESTROY(sc);
for (i = 0; i < HAL_NUM_TX_QUEUES; i++)
if (ATH_TXQ_SETUP(sc, i))
ath_tx_cleanupq(sc, &sc->sc_txq[i]);
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
honor net80211 mcast tx rate MFC after: 2 weeks
2006-02-09 21:15:36 +00:00
/*
optimize ath_tx_findrix: there's no need to walk the rates table as sc_rixmap is an inverse map NB: could eliminate the check for an invalid rate by filling in 0 for invalid entries but the rate control modules use it to identify bogus rates so leave it for now
2009-05-07 00:35:32 +00:00
* Return h/w rate index for an IEEE rate (w/o basic rate bit)
* using the current rates in sc_rixmap.
honor net80211 mcast tx rate MFC after: 2 weeks
2006-02-09 21:15:36 +00:00
*/
Migrate the TX path code out of if_ath and into a separate source file. There's two reasons for this: * the raw and non-raw TX path shares a lot of duplicate code which should be refactored; * the 11n-ready chip TX path needs a little reworking.
2011-01-29 11:35:23 +00:00
int
optimize ath_tx_findrix: there's no need to walk the rates table as sc_rixmap is an inverse map NB: could eliminate the check for an invalid rate by filling in 0 for invalid entries but the rate control modules use it to identify bogus rates so leave it for now
2009-05-07 00:35:32 +00:00
ath_tx_findrix(const struct ath_softc *sc, uint8_t rate)
honor net80211 mcast tx rate MFC after: 2 weeks
2006-02-09 21:15:36 +00:00
{
optimize ath_tx_findrix: there's no need to walk the rates table as sc_rixmap is an inverse map NB: could eliminate the check for an invalid rate by filling in 0 for invalid entries but the rate control modules use it to identify bogus rates so leave it for now
2009-05-07 00:35:32 +00:00
int rix = sc->sc_rixmap[rate];
/* NB: return lowest rix for invalid rate */
return (rix == 0xff ? 0 : rix);
honor net80211 mcast tx rate MFC after: 2 weeks
2006-02-09 21:15:36 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Process completed xmit descriptors from the specified queue.
*/
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
static int
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_tx_processq(struct ath_softc *sc, struct ath_txq *txq)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
{
struct ath_hal *ah = sc->sc_ah;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ifnet *ifp = sc->sc_ifp;
struct ieee80211com *ic = ifp->if_l2com;
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
struct ath_buf *bf, *last;
o extend cts to cover packet burst when operating in 11g w/ protection o check current channel parameters, not shadow state, for acm policy on data frames
2005-03-30 20:13:08 +00:00
struct ath_desc *ds, *ds0;
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
struct ath_tx_status *ts;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ieee80211_node *ni;
struct ath_node *an;
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
int sr, lr, pri, nacked;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
HAL_STATUS status;
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_TX_PROC, "%s: tx queue %u head %p link %p\n",
__func__, txq->axq_qnum,
(caddr_t)(uintptr_t) ath_hal_gettxbuf(sc->sc_ah, txq->axq_qnum),
txq->axq_link);
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
nacked = 0;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
for (;;) {
Update with last year of work.
2004-12-08 17:34:36 +00:00
ATH_TXQ_LOCK(txq);
txq->axq_intrcnt = 0; /* reset periodic desc intr count */
bf = STAILQ_FIRST(&txq->axq_q);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if (bf == NULL) {
Update with last year of work.
2004-12-08 17:34:36 +00:00
ATH_TXQ_UNLOCK(txq);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
break;
}
o extend cts to cover packet burst when operating in 11g w/ protection o check current channel parameters, not shadow state, for acm policy on data frames
2005-03-30 20:13:08 +00:00
ds0 = &bf->bf_desc[0];
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ds = &bf->bf_desc[bf->bf_nseg - 1];
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
ts = &bf->bf_status.ds_txstat;
status = ath_hal_txprocdesc(ah, ds, ts);
o add opt_ath.h enable tweaking various config parameters for the driver without modifying the source code o default debug msgs and diag support to off MFC after: 3 days
2006-04-03 18:14:02 +00:00
#ifdef ATH_DEBUG
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (sc->sc_debug & ATH_DEBUG_XMIT_DESC)
print the extended tx/rx descriptor for 5416 and later parts
2008-11-24 01:34:56 +00:00
ath_printtxbuf(sc, bf, txq->axq_qnum, 0,
status == HAL_OK);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
#endif
if (status == HAL_EINPROGRESS) {
Update with last year of work.
2004-12-08 17:34:36 +00:00
ATH_TXQ_UNLOCK(txq);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
break;
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
ATH_TXQ_REMOVE_HEAD(txq, bf_list);
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
if (txq->axq_depth > 0) {
/*
* More frames follow. Mark the buffer busy
* so it's not re-used while the hardware may
* still re-read the link field in the descriptor.
*/
bf->bf_flags |= ATH_BUF_BUSY;
} else
#else
backout 1.136 until we can resolve report that it causes output to stall
2006-02-27 17:20:23 +00:00
if (txq->axq_depth == 0)
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
#endif
close race between ath_tx_start and ath_tx_processq Reviewed by: apatti MFC after: 1 week
2006-02-15 18:12:24 +00:00
txq->axq_link = NULL;
Update with last year of work.
2004-12-08 17:34:36 +00:00
ATH_TXQ_UNLOCK(txq);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ni = bf->bf_node;
if (ni != NULL) {
Update with last year of work.
2004-12-08 17:34:36 +00:00
an = ATH_NODE(ni);
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
if (ts->ts_status == 0) {
u_int8_t txant = ts->ts_antenna;
Update with last year of work.
2004-12-08 17:34:36 +00:00
sc->sc_stats.ast_ant_tx[txant]++;
sc->sc_ant_tx[txant]++;
purge HAL_TXSTAT_ALTRATE; you can figure this out by checking ts_finaltsi and it cannot be used with MCS rate codes
2009-06-13 23:36:54 +00:00
if (ts->ts_finaltsi != 0)
Update with last year of work.
2004-12-08 17:34:36 +00:00
sc->sc_stats.ast_tx_altrate++;
pri = M_WME_GETAC(bf->bf_m);
if (pri >= WME_AC_VO)
ic->ic_wme.wme_hipri_traffic++;
update the sta inactivity timer only if we actually received an ACK
2008-10-27 17:43:23 +00:00
if ((bf->bf_txflags & HAL_TXDESC_NOACK) == 0)
ni->ni_inact = ni->ni_inact_reload;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
} else {
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
if (ts->ts_status & HAL_TXERR_XRETRY)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_stats.ast_tx_xretries++;
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
if (ts->ts_status & HAL_TXERR_FIFO)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_stats.ast_tx_fifoerr++;
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
if (ts->ts_status & HAL_TXERR_FILT)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_stats.ast_tx_filtered++;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
if (bf->bf_m->m_flags & M_FF)
sc->sc_stats.ast_ff_txerr++;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
sr = ts->ts_shortretry;
lr = ts->ts_longretry;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_stats.ast_tx_shortretry += sr;
sc->sc_stats.ast_tx_longretry += lr;
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Hand the descriptor to the rate control algorithm.
*/
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
if ((ts->ts_status & HAL_TXERR_FILT) == 0 &&
rename bf_flags to bf_txflags in preparation for the addition of flags separate from the tx descriptor flags currently recorded
2008-10-27 17:16:46 +00:00
(bf->bf_txflags & HAL_TXDESC_NOACK) == 0) {
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
/*
Fixup statistics: o update tx rssi data only when an ACK was received o return tx rssi from sampled data instead of the last frame o track noise floor o return rx rssi and noise floor (was broken)
2008-10-27 17:51:24 +00:00
* If frame was ack'd update statistics,
* including the last rx time used to
* workaround phantom bmiss interrupts.
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
*/
Fixup statistics: o update tx rssi data only when an ACK was received o return tx rssi from sampled data instead of the last frame o track noise floor o return rx rssi and noise floor (was broken)
2008-10-27 17:51:24 +00:00
if (ts->ts_status == 0) {
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
nacked++;
Fixup statistics: o update tx rssi data only when an ACK was received o return tx rssi from sampled data instead of the last frame o track noise floor o return rx rssi and noise floor (was broken)
2008-10-27 17:51:24 +00:00
sc->sc_stats.ast_tx_rssi = ts->ts_rssi;
ATH_RSSI_LPF(sc->sc_halstats.ns_avgtxrssi,
ts->ts_rssi);
}
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
ath_rate_tx_complete(sc, an, bf);
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
}
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
/*
* Do any tx complete callback. Note this must
* be done before releasing the node reference.
*/
if (bf->bf_m->m_flags & M_TXCB)
ieee80211_process_callback(ni, bf->bf_m,
correct callback status parameter; only indicate success when an ACK was received
2008-10-27 17:52:41 +00:00
(bf->bf_txflags & HAL_TXDESC_NOACK) == 0 ?
ts->ts_status : HAL_TXERR_XRETRY);
Update with last year of work.
2004-12-08 17:34:36 +00:00
ieee80211_free_node(ni);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
bus_dmamap_sync(sc->sc_dmat, bf->bf_dmamap,
BUS_DMASYNC_POSTWRITE);
bus_dmamap_unload(sc->sc_dmat, bf->bf_dmamap);
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
m_freem(bf->bf_m);
bf->bf_m = NULL;
bf->bf_node = NULL;
o convert mutex calls to #defines for portability, etc. o destroy mutex's on detach (was missing)
2003-10-14 22:51:45 +00:00
ATH_TXBUF_LOCK(sc);
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
last = STAILQ_LAST(&sc->sc_txbuf, ath_buf, bf_list);
if (last != NULL)
last->bf_flags &= ~ATH_BUF_BUSY;
Update with last year of work.
2004-12-08 17:34:36 +00:00
STAILQ_INSERT_TAIL(&sc->sc_txbuf, bf, bf_list);
o convert mutex calls to #defines for portability, etc. o destroy mutex's on detach (was missing)
2003-10-14 22:51:45 +00:00
ATH_TXBUF_UNLOCK(sc);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Hoist 802.11 encapsulation up into net80211: o call ieee80211_encap in ieee80211_start so frames passed down to drivers are already encapsulated o remove ieee80211_encap calls in drivers o fixup wi so it recreates the 802.3 head it requires from the 802.11 header contents o move fast-frame aggregation from ath to net80211 (conditional on IEEE80211_SUPPORT_SUPERG): - aggregation is now done in ieee80211_start; it is enabled when the packets/sec exceeds ieee80211_ffppsmin (net.wlan.ffppsmin) and frames are held on a staging queue according to ieee80211_ffagemax (net.wlan.ffagemax) to wait for a frame to combine with - drivers must call back to age/flush the staging queue (ath does this on tx done, at swba, and on rx according to the state of the tx queues and/or the contents of the staging queue) - remove fast-frame-related data structures from ath - add ieee80211_ff_node_init and ieee80211_ff_node_cleanup to handle per-node fast-frames state (we reuse 11n tx ampdu state) o change ieee80211_encap calling convention to include an explicit vap so frames coming through a WDS vap are recognized w/o setting M_WDS With these changes any device able to tx/rx 3Kbyte+ frames can use fast-frames. Reviewed by: thompsa, rpaulo, avatar, imp, sephe
2009-03-30 21:53:27 +00:00
#ifdef IEEE80211_SUPPORT_SUPERG
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
/*
* Flush fast-frame staging queue when traffic slows.
*/
if (txq->axq_depth <= 1)
make superg/fast-frames state dynamically-allocated (and indirect off the com structure instead of embedded); this reduces the overhead when not configured and reduces visibility of the contents
2009-05-02 20:16:55 +00:00
ieee80211_ff_flush(ic, txq->axq_ac);
Hoist 802.11 encapsulation up into net80211: o call ieee80211_encap in ieee80211_start so frames passed down to drivers are already encapsulated o remove ieee80211_encap calls in drivers o fixup wi so it recreates the 802.3 head it requires from the 802.11 header contents o move fast-frame aggregation from ath to net80211 (conditional on IEEE80211_SUPPORT_SUPERG): - aggregation is now done in ieee80211_start; it is enabled when the packets/sec exceeds ieee80211_ffppsmin (net.wlan.ffppsmin) and frames are held on a staging queue according to ieee80211_ffagemax (net.wlan.ffagemax) to wait for a frame to combine with - drivers must call back to age/flush the staging queue (ath does this on tx done, at swba, and on rx according to the state of the tx queues and/or the contents of the staging queue) - remove fast-frame-related data structures from ath - add ieee80211_ff_node_init and ieee80211_ff_node_cleanup to handle per-node fast-frames state (we reuse 11n tx ampdu state) o change ieee80211_encap calling convention to include an explicit vap so frames coming through a WDS vap are recognized w/o setting M_WDS With these changes any device able to tx/rx 3Kbyte+ frames can use fast-frames. Reviewed by: thompsa, rpaulo, avatar, imp, sephe
2009-03-30 21:53:27 +00:00
#endif
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
return nacked;
}
static __inline int
txqactive(struct ath_hal *ah, int qnum)
{
use ath_hal_gettxintrtxqs so we only process h/w tx queues that have an interrupt pending MFC after: 2 weeks
2006-02-15 18:17:53 +00:00
u_int32_t txqs = 1<<qnum;
ath_hal_gettxintrtxqs(ah, &txqs);
add missing bit from 1.130
2006-02-15 18:24:22 +00:00
return (txqs & (1<<qnum));
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
/*
* Deferred processing of transmit interrupt; special-cased
* for a single hardware transmit queue (e.g. 5210 and 5211).
*/
static void
ath_tx_proc_q0(void *arg, int npending)
{
struct ath_softc *sc = arg;
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Update with last year of work.
2004-12-08 17:34:36 +00:00
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
if (txqactive(sc->sc_ah, 0) && ath_tx_processq(sc, &sc->sc_txq[0]))
sc->sc_lastrx = ath_hal_gettsf64(sc->sc_ah);
if (txqactive(sc->sc_ah, sc->sc_cabq->axq_qnum))
ath_tx_processq(sc, sc->sc_cabq);
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
ifp->if_drv_flags &= ~IFF_DRV_OACTIVE;
replace if_watchdog w/ private callout; probably can merge this with the calibration work sometime in the future
2009-03-09 23:10:19 +00:00
sc->sc_wd_timer = 0;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
better led blinking
2005-01-18 19:03:04 +00:00
if (sc->sc_softled)
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
ath_led_event(sc, sc->sc_txrix);
better led blinking
2005-01-18 19:03:04 +00:00
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_start(ifp);
}
/*
Update with last year of work.
2004-12-08 17:34:36 +00:00
* Deferred processing of transmit interrupt; special-cased
* for four hardware queues, 0-3 (e.g. 5212 w/ WME support).
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
*/
static void
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_tx_proc_q0123(void *arg, int npending)
{
struct ath_softc *sc = arg;
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
int nacked;
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Process each active queue.
*/
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
nacked = 0;
if (txqactive(sc->sc_ah, 0))
nacked += ath_tx_processq(sc, &sc->sc_txq[0]);
if (txqactive(sc->sc_ah, 1))
nacked += ath_tx_processq(sc, &sc->sc_txq[1]);
if (txqactive(sc->sc_ah, 2))
nacked += ath_tx_processq(sc, &sc->sc_txq[2]);
if (txqactive(sc->sc_ah, 3))
nacked += ath_tx_processq(sc, &sc->sc_txq[3]);
if (txqactive(sc->sc_ah, sc->sc_cabq->axq_qnum))
ath_tx_processq(sc, sc->sc_cabq);
if (nacked)
sc->sc_lastrx = ath_hal_gettsf64(sc->sc_ah);
Update with last year of work.
2004-12-08 17:34:36 +00:00
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
ifp->if_drv_flags &= ~IFF_DRV_OACTIVE;
replace if_watchdog w/ private callout; probably can merge this with the calibration work sometime in the future
2009-03-09 23:10:19 +00:00
sc->sc_wd_timer = 0;
Update with last year of work.
2004-12-08 17:34:36 +00:00
better led blinking
2005-01-18 19:03:04 +00:00
if (sc->sc_softled)
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
ath_led_event(sc, sc->sc_txrix);
better led blinking
2005-01-18 19:03:04 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_start(ifp);
}
/*
* Deferred processing of transmit interrupt.
*/
static void
ath_tx_proc(void *arg, int npending)
{
struct ath_softc *sc = arg;
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
int i, nacked;
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Process each active queue.
*/
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
nacked = 0;
Update with last year of work.
2004-12-08 17:34:36 +00:00
for (i = 0; i < HAL_NUM_TX_QUEUES; i++)
Phantom beacon miss workaround: track the tsf of the last received frame and if we get a beacon miss interrupt ignore it if we've received a frame within the beacon miss interval. This should never trigger and the handling at the net80211 layer should likewise deal with this but it doesn't hurt and can suppress extranous probe request frames. Note that we can legtimately get a bmiss when under heavy load. MFC after: 2 weeks
2006-02-09 22:03:26 +00:00
if (ATH_TXQ_SETUP(sc, i) && txqactive(sc->sc_ah, i))
nacked += ath_tx_processq(sc, &sc->sc_txq[i]);
if (nacked)
sc->sc_lastrx = ath_hal_gettsf64(sc->sc_ah);
Update with last year of work.
2004-12-08 17:34:36 +00:00
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
ifp->if_drv_flags &= ~IFF_DRV_OACTIVE;
replace if_watchdog w/ private callout; probably can merge this with the calibration work sometime in the future
2009-03-09 23:10:19 +00:00
sc->sc_wd_timer = 0;
Update with last year of work.
2004-12-08 17:34:36 +00:00
better led blinking
2005-01-18 19:03:04 +00:00
if (sc->sc_softled)
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
ath_led_event(sc, sc->sc_txrix);
better led blinking
2005-01-18 19:03:04 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_start(ifp);
}
static void
ath_tx_draintxq(struct ath_softc *sc, struct ath_txq *txq)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
{
o add opt_ath.h enable tweaking various config parameters for the driver without modifying the source code o default debug msgs and diag support to off MFC after: 3 days
2006-04-03 18:14:02 +00:00
#ifdef ATH_DEBUG
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ath_hal *ah = sc->sc_ah;
fix build w/o AR_DEBUG MFC after: 2 weeks
2006-02-15 18:35:09 +00:00
#endif
When draining the tx queue reclaim any node references held in packets. This fixes a problem when operating as an AP where clients would get stuck in the node table because the reference count never went to zero.
2004-01-07 19:16:49 +00:00
struct ieee80211_node *ni;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ath_buf *bf;
improve tx/rx buf printing routines MFC after: 2 weeks
2006-02-15 18:31:04 +00:00
u_int ix;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* NB: this assumes output has been stopped and
fix comment
2008-10-27 17:53:34 +00:00
* we do not need to block ath_tx_proc
Update with last year of work.
2004-12-08 17:34:36 +00:00
*/
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
ATH_TXBUF_LOCK(sc);
bf = STAILQ_LAST(&sc->sc_txbuf, ath_buf, bf_list);
if (bf != NULL)
bf->bf_flags &= ~ATH_BUF_BUSY;
ATH_TXBUF_UNLOCK(sc);
improve tx/rx buf printing routines MFC after: 2 weeks
2006-02-15 18:31:04 +00:00
for (ix = 0;; ix++) {
Update with last year of work.
2004-12-08 17:34:36 +00:00
ATH_TXQ_LOCK(txq);
bf = STAILQ_FIRST(&txq->axq_q);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
if (bf == NULL) {
backout 1.136 until we can resolve report that it causes output to stall
2006-02-27 17:20:23 +00:00
txq->axq_link = NULL;
Update with last year of work.
2004-12-08 17:34:36 +00:00
ATH_TXQ_UNLOCK(txq);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
break;
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
ATH_TXQ_REMOVE_HEAD(txq, bf_list);
ATH_TXQ_UNLOCK(txq);
o add opt_ath.h enable tweaking various config parameters for the driver without modifying the source code o default debug msgs and diag support to off MFC after: 3 days
2006-04-03 18:14:02 +00:00
#ifdef ATH_DEBUG
Improve ath_draintxq debug info: dump the packet as well as the descriptor and handle the beacon q like other q's MFC after: 1 month
2006-04-16 18:24:27 +00:00
if (sc->sc_debug & ATH_DEBUG_RESET) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211com *ic = sc->sc_ifp->if_l2com;
print the extended tx/rx descriptor for 5416 and later parts
2008-11-24 01:34:56 +00:00
ath_printtxbuf(sc, bf, txq->axq_qnum, ix,
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
ath_hal_txprocdesc(ah, bf->bf_desc,
&bf->bf_status.ds_txstat) == HAL_OK);
don't use caddr_t to match ieee80211_dump_pkt type; supplying the correct one costs nothing
2009-04-13 20:58:47 +00:00
ieee80211_dump_pkt(ic, mtod(bf->bf_m, const uint8_t *),
bf->bf_m->m_len, 0, -1);
Improve ath_draintxq debug info: dump the packet as well as the descriptor and handle the beacon q like other q's MFC after: 1 month
2006-04-16 18:24:27 +00:00
}
o add opt_ath.h enable tweaking various config parameters for the driver without modifying the source code o default debug msgs and diag support to off MFC after: 3 days
2006-04-03 18:14:02 +00:00
#endif /* ATH_DEBUG */
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
bus_dmamap_unload(sc->sc_dmat, bf->bf_dmamap);
When draining the tx queue reclaim any node references held in packets. This fixes a problem when operating as an AP where clients would get stuck in the node table because the reference count never went to zero.
2004-01-07 19:16:49 +00:00
ni = bf->bf_node;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
bf->bf_node = NULL;
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (ni != NULL) {
When draining the tx queue reclaim any node references held in packets. This fixes a problem when operating as an AP where clients would get stuck in the node table because the reference count never went to zero.
2004-01-07 19:16:49 +00:00
/*
Process tx callbacks when draining the tx q; this fixes a problem where a device timeout that occurs with a mgt frame on the tx q will leave the net80211 layer w/o any way to make progress. Reviewed by: thompsa, sephe Approved by: re (hrs)
2007-06-24 01:57:20 +00:00
* Do any callback and reclaim the node reference.
When draining the tx queue reclaim any node references held in packets. This fixes a problem when operating as an AP where clients would get stuck in the node table because the reference count never went to zero.
2004-01-07 19:16:49 +00:00
*/
Process tx callbacks when draining the tx q; this fixes a problem where a device timeout that occurs with a mgt frame on the tx q will leave the net80211 layer w/o any way to make progress. Reviewed by: thompsa, sephe Approved by: re (hrs)
2007-06-24 01:57:20 +00:00
if (bf->bf_m->m_flags & M_TXCB)
ieee80211_process_callback(ni, bf->bf_m, -1);
Update with last year of work.
2004-12-08 17:34:36 +00:00
ieee80211_free_node(ni);
When draining the tx queue reclaim any node references held in packets. This fixes a problem when operating as an AP where clients would get stuck in the node table because the reference count never went to zero.
2004-01-07 19:16:49 +00:00
}
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
m_freem(bf->bf_m);
bf->bf_m = NULL;
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
bf->bf_flags &= ~ATH_BUF_BUSY;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
o convert mutex calls to #defines for portability, etc. o destroy mutex's on detach (was missing)
2003-10-14 22:51:45 +00:00
ATH_TXBUF_LOCK(sc);
Update with last year of work.
2004-12-08 17:34:36 +00:00
STAILQ_INSERT_TAIL(&sc->sc_txbuf, bf, bf_list);
o convert mutex calls to #defines for portability, etc. o destroy mutex's on detach (was missing)
2003-10-14 22:51:45 +00:00
ATH_TXBUF_UNLOCK(sc);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
static void
ath_tx_stopdma(struct ath_softc *sc, struct ath_txq *txq)
{
struct ath_hal *ah = sc->sc_ah;
DPRINTF(sc, ATH_DEBUG_RESET, "%s: tx queue [%u] %p, link %p\n",
__func__, txq->axq_qnum,
Make this amd64-clean. sizeof is long on amd64, so things that do a printf of a sizeof, need to use %z to get the correct type on all our platforms. Also, convert integers<->pointers via uintptr_t. (I think Sam's instructions were for me to commit this. If I misunderstood, then I apologize in advance.)
2004-12-15 02:25:21 +00:00
(caddr_t)(uintptr_t) ath_hal_gettxbuf(ah, txq->axq_qnum),
txq->axq_link);
Improve ath_draintxq debug info: dump the packet as well as the descriptor and handle the beacon q like other q's MFC after: 1 month
2006-04-16 18:24:27 +00:00
(void) ath_hal_stoptxdma(ah, txq->axq_qnum);
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
/*
* Drain the transmit queues and reclaim resources.
*/
static void
ath_draintxq(struct ath_softc *sc)
{
struct ath_hal *ah = sc->sc_ah;
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Update with last year of work.
2004-12-08 17:34:36 +00:00
int i;
/* XXX return value */
if (!sc->sc_invalid) {
/* don't touch the hardware if marked invalid */
Improve ath_draintxq debug info: dump the packet as well as the descriptor and handle the beacon q like other q's MFC after: 1 month
2006-04-16 18:24:27 +00:00
DPRINTF(sc, ATH_DEBUG_RESET, "%s: tx queue [%u] %p, link %p\n",
__func__, sc->sc_bhalq,
(caddr_t)(uintptr_t) ath_hal_gettxbuf(ah, sc->sc_bhalq),
NULL);
Update with last year of work.
2004-12-08 17:34:36 +00:00
(void) ath_hal_stoptxdma(ah, sc->sc_bhalq);
for (i = 0; i < HAL_NUM_TX_QUEUES; i++)
if (ATH_TXQ_SETUP(sc, i))
ath_tx_stopdma(sc, &sc->sc_txq[i]);
}
for (i = 0; i < HAL_NUM_TX_QUEUES; i++)
if (ATH_TXQ_SETUP(sc, i))
ath_tx_draintxq(sc, &sc->sc_txq[i]);
Improve ath_draintxq debug info: dump the packet as well as the descriptor and handle the beacon q like other q's MFC after: 1 month
2006-04-16 18:24:27 +00:00
#ifdef ATH_DEBUG
if (sc->sc_debug & ATH_DEBUG_RESET) {
struct ath_buf *bf = STAILQ_FIRST(&sc->sc_bbuf);
if (bf != NULL && bf->bf_m != NULL) {
print the extended tx/rx descriptor for 5416 and later parts
2008-11-24 01:34:56 +00:00
ath_printtxbuf(sc, bf, sc->sc_bhalq, 0,
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
ath_hal_txprocdesc(ah, bf->bf_desc,
&bf->bf_status.ds_txstat) == HAL_OK);
don't use caddr_t to match ieee80211_dump_pkt type; supplying the correct one costs nothing
2009-04-13 20:58:47 +00:00
ieee80211_dump_pkt(ifp->if_l2com,
mtod(bf->bf_m, const uint8_t *), bf->bf_m->m_len,
0, -1);
Improve ath_draintxq debug info: dump the packet as well as the descriptor and handle the beacon q like other q's MFC after: 1 month
2006-04-16 18:24:27 +00:00
}
}
#endif /* ATH_DEBUG */
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
ifp->if_drv_flags &= ~IFF_DRV_OACTIVE;
replace if_watchdog w/ private callout; probably can merge this with the calibration work sometime in the future
2009-03-09 23:10:19 +00:00
sc->sc_wd_timer = 0;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
/*
* Disable the receive h/w in preparation for a reset.
*/
static void
ath_stoprecv(struct ath_softc *sc)
{
o track API change for HAL v0.9.6.1 o fix race condition when processing rx descriptors: because we use a self-linked descriptor at the end of the rx descriptor list to avoid rx overruns (which can easily happen for 5212 parts that enable PHY errors) we must carefully check that a descriptor is "done" by looking ahead to the next descriptor before believing the done bit in the current descriptor (this is all handled in the HAL since the rx descriptor format is chip-specific so we need to pass in two additional parameters--the physical address of the current descriptor and the virtual address of the next descriptor in the list) o check copyout return status for SIOCGATHSTATS ioctl Approved by: re (scottl)
2003-11-29 01:23:59 +00:00
#define PA2DESC(_sc, _pa) \
Update with last year of work.
2004-12-08 17:34:36 +00:00
((struct ath_desc *)((caddr_t)(_sc)->sc_rxdma.dd_desc + \
((_pa) - (_sc)->sc_rxdma.dd_desc_paddr)))
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ath_hal *ah = sc->sc_ah;
ath_hal_stoppcurecv(ah); /* disable PCU */
ath_hal_setrxfilter(ah, 0); /* clear recv filter */
ath_hal_stopdmarecv(ah); /* disable DMA engine */
Update with last year of work.
2004-12-08 17:34:36 +00:00
DELAY(3000); /* 3ms is long enough for 1 frame */
o add opt_ath.h enable tweaking various config parameters for the driver without modifying the source code o default debug msgs and diag support to off MFC after: 3 days
2006-04-03 18:14:02 +00:00
#ifdef ATH_DEBUG
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (sc->sc_debug & (ATH_DEBUG_RESET | ATH_DEBUG_FATAL)) {
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ath_buf *bf;
improve tx/rx buf printing routines MFC after: 2 weeks
2006-02-15 18:31:04 +00:00
u_int ix;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
split debugging messages up into classes; ah_debug is now treated as a bit vector
2004-01-07 18:57:38 +00:00
printf("%s: rx queue %p, link %p\n", __func__,
Make this compile on amd64. "I'll cope" by: sam
2004-02-06 00:10:54 +00:00
(caddr_t)(uintptr_t) ath_hal_getrxbuf(ah), sc->sc_rxlink);
improve tx/rx buf printing routines MFC after: 2 weeks
2006-02-15 18:31:04 +00:00
ix = 0;
Update with last year of work.
2004-12-08 17:34:36 +00:00
STAILQ_FOREACH(bf, &sc->sc_rxbuf, bf_list) {
o track API change for HAL v0.9.6.1 o fix race condition when processing rx descriptors: because we use a self-linked descriptor at the end of the rx descriptor list to avoid rx overruns (which can easily happen for 5212 parts that enable PHY errors) we must carefully check that a descriptor is "done" by looking ahead to the next descriptor before believing the done bit in the current descriptor (this is all handled in the HAL since the rx descriptor format is chip-specific so we need to pass in two additional parameters--the physical address of the current descriptor and the virtual address of the next descriptor in the list) o check copyout return status for SIOCGATHSTATS ioctl Approved by: re (scottl)
2003-11-29 01:23:59 +00:00
struct ath_desc *ds = bf->bf_desc;
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
struct ath_rx_status *rs = &bf->bf_status.ds_rxstat;
Update with last year of work.
2004-12-08 17:34:36 +00:00
HAL_STATUS status = ath_hal_rxprocdesc(ah, ds,
Track v0.9.20.3 hal: o no more ds_vdata in tx/rx descriptors o split h/w tx/rx descriptor from s/w status o as part of the descriptor split change the rate control module api so the ath_buf is passed in to the module so it can fetch both descriptor and status information as needed o add some const poisoning Also for sample rate control algorithm: o split debug msgs (node, rate, any) o uniformly bounds check rate indices (and in some cases correct checks) o move array index ops to after bounds checking o use final tsi from the status block instead of the h/w descriptor o replace h/w descriptor struct's with proper mask+shift defs (this doesn't belong here; everything is known by the driver and should just be sent down so there's no h/w-specific knowledge) MFC after: 1 month
2006-12-13 19:34:35 +00:00
bf->bf_daddr, PA2DESC(sc, ds->ds_link), rs);
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (status == HAL_OK || (sc->sc_debug & ATH_DEBUG_FATAL))
print the extended tx/rx descriptor for 5416 and later parts
2008-11-24 01:34:56 +00:00
ath_printrxbuf(sc, bf, ix, status == HAL_OK);
improve tx/rx buf printing routines MFC after: 2 weeks
2006-02-15 18:31:04 +00:00
ix++;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
}
#endif
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
if (sc->sc_rxpending != NULL) {
m_freem(sc->sc_rxpending);
sc->sc_rxpending = NULL;
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_rxlink = NULL; /* just in case */
o track API change for HAL v0.9.6.1 o fix race condition when processing rx descriptors: because we use a self-linked descriptor at the end of the rx descriptor list to avoid rx overruns (which can easily happen for 5212 parts that enable PHY errors) we must carefully check that a descriptor is "done" by looking ahead to the next descriptor before believing the done bit in the current descriptor (this is all handled in the HAL since the rx descriptor format is chip-specific so we need to pass in two additional parameters--the physical address of the current descriptor and the virtual address of the next descriptor in the list) o check copyout return status for SIOCGATHSTATS ioctl Approved by: re (scottl)
2003-11-29 01:23:59 +00:00
#undef PA2DESC
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
/*
* Enable the receive h/w following a reset.
*/
static int
ath_startrecv(struct ath_softc *sc)
{
struct ath_hal *ah = sc->sc_ah;
struct ath_buf *bf;
sc->sc_rxlink = NULL;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
sc->sc_rxpending = NULL;
Update with last year of work.
2004-12-08 17:34:36 +00:00
STAILQ_FOREACH(bf, &sc->sc_rxbuf, bf_list) {
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
int error = ath_rxbuf_init(sc, bf);
if (error != 0) {
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_RECV,
"%s: ath_rxbuf_init failed %d\n",
__func__, error);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
return error;
}
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
bf = STAILQ_FIRST(&sc->sc_rxbuf);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_hal_putrxbuf(ah, bf->bf_daddr);
ath_hal_rxena(ah); /* enable recv descriptors */
ath_mode_init(sc); /* set filters, etc. */
ath_hal_startpcurecv(ah); /* re-enable PCU/DMA engine */
return 0;
}
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
/*
Update with last year of work.
2004-12-08 17:34:36 +00:00
* Update internal state after a channel change.
*/
static void
ath_chan_change(struct ath_softc *sc, struct ieee80211_channel *chan)
{
enum ieee80211_phymode mode;
/*
* Change channels and update the h/w rate map
* if we're switching; e.g. 11a to 11b/g.
*/
remove private support for IEEE80211_MODE_HALF and IEEE80211_MODE_QUARTER now that net80211 has them
2009-02-19 05:22:40 +00:00
mode = ieee80211_chan2mode(chan);
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (mode != sc->sc_curmode)
ath_setcurmode(sc, 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
sc->sc_curchan = chan;
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/*
* Set/change channels. If the channel is really being changed,
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
* it's done by resetting the chip. To accomplish this we must
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
* first cleanup any pending DMA, then restart stuff after a la
* ath_init.
*/
static int
ath_chan_set(struct ath_softc *sc, struct ieee80211_channel *chan)
{
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ifnet *ifp = sc->sc_ifp;
struct ieee80211com *ic = ifp->if_l2com;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
struct ath_hal *ah = sc->sc_ah;
Update with last year of work.
2004-12-08 17:34:36 +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
DPRINTF(sc, ATH_DEBUG_RESET, "%s: %u (%u MHz, flags 0x%x)\n",
__func__, ieee80211_chan2ieee(ic, chan),
chan->ic_freq, chan->ic_flags);
if (chan != sc->sc_curchan) {
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
HAL_STATUS status;
/*
* To switch channels clear any pending DMA operations;
* wait long enough for the RX fifo to drain, reset the
* hardware at the new frequency, and then re-enable
* the relevant bits of the h/w.
*/
ath_hal_intrset(ah, 0); /* disable interrupts */
ath_draintxq(sc); /* clear pending tx frames */
ath_stoprecv(sc); /* turn off frame recv */
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 (!ath_hal_reset(ah, sc->sc_opmode, chan, AH_TRUE, &status)) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if_printf(ifp, "%s: unable to reset "
Spell "Hz" correctly wherever it is user-visible. PR: bin/142566 Submitted by: N.J. Mann njm njm.me.uk Approved by: ed (mentor) MFC after: 2 weeks
2010-01-12 17:59:58 +00:00
"channel %u (%u MHz, flags 0x%x), hal status %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__, ieee80211_chan2ieee(ic, chan),
chan->ic_freq, chan->ic_flags, status);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
return EIO;
}
o fix setup of sc_diversity; the hal does not give us reliable status after attach, only after a reset o when setting diversity via the sysctl don't update sc_diversity until we know the hal requested worked o while here eliminate sc_hasdiversity and sc_hastpc; just query the hal each time since these are the only places we need to know MFC after: 3 days
2005-07-24 05:11:39 +00:00
sc->sc_diversity = ath_hal_getdiversity(ah);
Update with last year of work.
2004-12-08 17:34:36 +00:00
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/*
* Re-enable rx framework.
*/
if (ath_startrecv(sc) != 0) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if_printf(ifp, "%s: unable to restart recv logic\n",
__func__);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
return EIO;
}
/*
* Change channels and update the h/w rate map
* if we're switching; e.g. 11a to 11b/g.
*/
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_chan_change(sc, chan);
MFp4 changes to fix locking issues and correct reference count handling of station entries in hostap mode: Input path: o driver is now expected to find the node associated with the sender of a received frame; use ic_bss if none is located o driver passes the (referenced) node into ieee80211_input for use within the wlan module and is responsible for cleaning up on return o the antenna state is no longer passed up with each frame; this is now considered driver-private state and drivers are responsible for keeping it in the driver-private part of a node Output path: Revamp output path for management frames to eliminate redundant locking that causes problems and to correct reference counting bogosity that occurs when stations are timed out due to inactivity (in AP mode). On output the refcnt'd node is stashed in the pkthdr's recvif field (yech) and retrieved by the driver. This eliminates an unref/ref scenario and related node table unlock/lock due to the driver looking up the node. This is particularly important when stations are timed out as this causes a lock order reversal that can result in a deadlock. As a byproduct we also reduce the overhead for sending management frames (minimal). Additional fallout from this is a change to ieee80211_encap to return a refcn't node for tieing to the outbound frame. Node refcnts are not reclaimed until after a frame is completely processed (e.g. in the tx interrupt handler). This is especially important for timed out stations as this deref will be the final one causing the node entry to be reclaimed. Additional semi-related changes: o replace m_copym use with m_copypacket (optimization) o add assert to verify ic_bss is never free'd during normal operation o add comments explaining calling conventions by drivers for frames going in each direction o remove extraneous code that "cannot be executed" (e.g. because pointers may never be null)
2003-08-19 22:17:04 +00:00
/*
* Re-enable interrupts.
*/
ath_hal_intrset(ah, sc->sc_imask);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
return 0;
}
/*
* Periodically recalibrate the PHY to account
* for temperature/environment changes.
*/
static void
ath_calibrate(void *arg)
{
struct ath_softc *sc = arg;
struct ath_hal *ah = sc->sc_ah;
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
struct ifnet *ifp = sc->sc_ifp;
don't run the calibration code if scanning, we won't be on the home channel
2009-01-23 03:15:28 +00:00
struct ieee80211com *ic = ifp->if_l2com;
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
HAL_BOOL longCal, isCalDone;
ANI changes #1 - split out the ANI polling from the RxMonitor hook. The rxmonitor hook is called on each received packet. This can get very, very busy as the tx/rx/chanbusy registers are thus read each time a packet is received. Instead, shuffle out the true per-packet processing which is needed and move the rest of the ANI processing into a periodic event which runs every 100ms by default.
2011-01-21 05:21:00 +00:00
HAL_BOOL aniCal, shortCal = AH_FALSE;
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
int nextcal;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
don't run the calibration code if scanning, we won't be on the home channel
2009-01-23 03:15:28 +00:00
if (ic->ic_flags & IEEE80211_F_SCAN) /* defer, off channel */
goto restart;
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
longCal = (ticks - sc->sc_lastlongcal >= ath_longcalinterval*hz);
ANI changes #1 - split out the ANI polling from the RxMonitor hook. The rxmonitor hook is called on each received packet. This can get very, very busy as the tx/rx/chanbusy registers are thus read each time a packet is received. Instead, shuffle out the true per-packet processing which is needed and move the rest of the ANI processing into a periodic event which runs every 100ms by default.
2011-01-21 05:21:00 +00:00
aniCal = (ticks - sc->sc_lastani >= ath_anicalinterval*hz/1000);
if (sc->sc_doresetcal)
shortCal = (ticks - sc->sc_lastshortcal >= ath_shortcalinterval*hz/1000);
DPRINTF(sc, ATH_DEBUG_CALIBRATE, "%s: shortCal=%d; longCal=%d; aniCal=%d\n", __func__, shortCal, longCal, aniCal);
if (aniCal) {
sc->sc_stats.ast_ani_cal++;
sc->sc_lastani = ticks;
ath_hal_ani_poll(ah, sc->sc_curchan);
}
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
if (longCal) {
sc->sc_stats.ast_per_cal++;
Don't delay updating the longcal timer - instead, update the longcal flag immediately so it's only set once per longcal interval. Without this, the current AR5416 code will continuously spam NF calibrations during a periodic calibration if the longcal flag is set. The longcal flag wouldn't be cleared until the calibration method indicates that calibrations are "complete". This drops the rate of NF calibration updates down from "once every shortcal" (ie, every 100ms) during a periodic calibration, to only once per "longcal" interval. Spamming NF calibrations every 100ms caused some potentially horrific issues in noisy environments as NF calibrations can take longer than 100ms and this spamming can cause invalid NF calibration results to be read back - leading to missed beacons, and thus leading to a stuck beacon situation. Stuck beacons cause interface resets, which restart calibrations. This means that the longcal calibration runs every 100ms (shortcal) until all initial calibrations are completed. This spamming can then cause the above issues which leads to stuck beacons, leading to interface resets, etc, etc. Quite annoying.
2010-08-10 07:56:56 +00:00
sc->sc_lastlongcal = ticks;
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
if (ath_hal_getrfgain(ah) == HAL_RFGAIN_NEED_CHANGE) {
/*
* Rfgain is out of bounds, reset the chip
* to load new gain values.
*/
DPRINTF(sc, ATH_DEBUG_CALIBRATE,
"%s: rfgain change\n", __func__);
sc->sc_stats.ast_per_rfgain++;
ath_reset(ifp);
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/*
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
* If this long cal is after an idle period, then
* reset the data collection state so we start fresh.
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
*/
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
if (sc->sc_resetcal) {
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
(void) ath_hal_calreset(ah, sc->sc_curchan);
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
sc->sc_lastcalreset = ticks;
ANI changes #1 - split out the ANI polling from the RxMonitor hook. The rxmonitor hook is called on each received packet. This can get very, very busy as the tx/rx/chanbusy registers are thus read each time a packet is received. Instead, shuffle out the true per-packet processing which is needed and move the rest of the ANI processing into a periodic event which runs every 100ms by default.
2011-01-21 05:21:00 +00:00
sc->sc_lastshortcal = ticks;
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
sc->sc_resetcal = 0;
ANI changes #1 - split out the ANI polling from the RxMonitor hook. The rxmonitor hook is called on each received packet. This can get very, very busy as the tx/rx/chanbusy registers are thus read each time a packet is received. Instead, shuffle out the true per-packet processing which is needed and move the rest of the ANI processing into a periodic event which runs every 100ms by default.
2011-01-21 05:21:00 +00:00
sc->sc_doresetcal = AH_TRUE;
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
ANI changes #1 - split out the ANI polling from the RxMonitor hook. The rxmonitor hook is called on each received packet. This can get very, very busy as the tx/rx/chanbusy registers are thus read each time a packet is received. Instead, shuffle out the true per-packet processing which is needed and move the rest of the ANI processing into a periodic event which runs every 100ms by default.
2011-01-21 05:21:00 +00:00
/* Only call if we're doing a short/long cal, not for ANI calibration */
if (shortCal || longCal) {
if (ath_hal_calibrateN(ah, sc->sc_curchan, longCal, &isCalDone)) {
if (longCal) {
/*
* Calibrate noise floor data again in case of change.
*/
ath_hal_process_noisefloor(ah);
}
} else {
DPRINTF(sc, ATH_DEBUG_ANY,
"%s: calibration of channel %u failed\n",
__func__, sc->sc_curchan->ic_freq);
sc->sc_stats.ast_per_calfail++;
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
}
ANI changes #1 - split out the ANI polling from the RxMonitor hook. The rxmonitor hook is called on each received packet. This can get very, very busy as the tx/rx/chanbusy registers are thus read each time a packet is received. Instead, shuffle out the true per-packet processing which is needed and move the rest of the ANI processing into a periodic event which runs every 100ms by default.
2011-01-21 05:21:00 +00:00
if (shortCal)
sc->sc_lastshortcal = ticks;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
if (!isCalDone) {
don't run the calibration code if scanning, we won't be on the home channel
2009-01-23 03:15:28 +00:00
restart:
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
/*
* Use a shorter interval to potentially collect multiple
* data samples required to complete calibration. Once
* we're told the work is done we drop back to a longer
* interval between requests. We're more aggressive doing
* work when operating as an AP to improve operation right
* after startup.
*/
ANI changes #1 - split out the ANI polling from the RxMonitor hook. The rxmonitor hook is called on each received packet. This can get very, very busy as the tx/rx/chanbusy registers are thus read each time a packet is received. Instead, shuffle out the true per-packet processing which is needed and move the rest of the ANI processing into a periodic event which runs every 100ms by default.
2011-01-21 05:21:00 +00:00
sc->sc_lastshortcal = ticks;
nextcal = ath_shortcalinterval*hz/1000;
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
if (sc->sc_opmode != HAL_M_HOSTAP)
nextcal *= 10;
ANI changes #1 - split out the ANI polling from the RxMonitor hook. The rxmonitor hook is called on each received packet. This can get very, very busy as the tx/rx/chanbusy registers are thus read each time a packet is received. Instead, shuffle out the true per-packet processing which is needed and move the rest of the ANI processing into a periodic event which runs every 100ms by default.
2011-01-21 05:21:00 +00:00
sc->sc_doresetcal = AH_TRUE;
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
} else {
ANI changes #1 - split out the ANI polling from the RxMonitor hook. The rxmonitor hook is called on each received packet. This can get very, very busy as the tx/rx/chanbusy registers are thus read each time a packet is received. Instead, shuffle out the true per-packet processing which is needed and move the rest of the ANI processing into a periodic event which runs every 100ms by default.
2011-01-21 05:21:00 +00:00
/* nextcal should be the shortest time for next event */
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
nextcal = ath_longcalinterval*hz;
if (sc->sc_lastcalreset == 0)
sc->sc_lastcalreset = sc->sc_lastlongcal;
else if (ticks - sc->sc_lastcalreset >= ath_resetcalinterval*hz)
sc->sc_resetcal = 1; /* setup reset next trip */
ANI changes #1 - split out the ANI polling from the RxMonitor hook. The rxmonitor hook is called on each received packet. This can get very, very busy as the tx/rx/chanbusy registers are thus read each time a packet is received. Instead, shuffle out the true per-packet processing which is needed and move the rest of the ANI processing into a periodic event which runs every 100ms by default.
2011-01-21 05:21:00 +00:00
sc->sc_doresetcal = AH_FALSE;
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
}
ANI changes #1 - split out the ANI polling from the RxMonitor hook. The rxmonitor hook is called on each received packet. This can get very, very busy as the tx/rx/chanbusy registers are thus read each time a packet is received. Instead, shuffle out the true per-packet processing which is needed and move the rest of the ANI processing into a periodic event which runs every 100ms by default.
2011-01-21 05:21:00 +00:00
/* ANI calibration may occur more often than short/long/resetcal */
if (ath_anicalinterval > 0)
nextcal = MIN(nextcal, ath_anicalinterval*hz/1000);
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
if (nextcal != 0) {
DPRINTF(sc, ATH_DEBUG_CALIBRATE, "%s: next +%u (%sisCalDone)\n",
__func__, nextcal, isCalDone ? "" : "!");
callout_reset(&sc->sc_cal_ch, nextcal, ath_calibrate, sc);
} else {
DPRINTF(sc, ATH_DEBUG_CALIBRATE, "%s: calibration disabled\n",
__func__);
/* NB: don't rearm timer */
Update for rev 0.9.16.16 hal: o add dfs+radar hooks; DFS is presently disabled in the hal o channel and mode handling changes o various api changes o be more aggressive about iq calibration settling so ap mode operation is better immediately after startup o rfkill/rfsilent sysctl support o tpc ack/cts sysctl support MFC after: 2 weeks
2006-02-10 19:07:08 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
static void
ath_scan_start(struct ieee80211com *ic)
{
struct ifnet *ifp = ic->ic_ifp;
struct ath_softc *sc = ifp->if_softc;
struct ath_hal *ah = sc->sc_ah;
u_int32_t rfilt;
/* XXX calibration timer? */
sc->sc_scanning = 1;
sc->sc_syncbeacon = 0;
rfilt = ath_calcrxfilter(sc);
ath_hal_setrxfilter(ah, rfilt);
ath_hal_setassocid(ah, ifp->if_broadcastaddr, 0);
DPRINTF(sc, ATH_DEBUG_STATE, "%s: RX filter 0x%x bssid %s aid 0\n",
__func__, rfilt, ether_sprintf(ifp->if_broadcastaddr));
}
static void
ath_scan_end(struct ieee80211com *ic)
{
struct ifnet *ifp = ic->ic_ifp;
struct ath_softc *sc = ifp->if_softc;
struct ath_hal *ah = sc->sc_ah;
u_int32_t rfilt;
sc->sc_scanning = 0;
rfilt = ath_calcrxfilter(sc);
ath_hal_setrxfilter(ah, rfilt);
ath_hal_setassocid(ah, sc->sc_curbssid, sc->sc_curaid);
ath_hal_process_noisefloor(ah);
DPRINTF(sc, ATH_DEBUG_STATE, "%s: RX filter 0x%x bssid %s aid 0x%x\n",
__func__, rfilt, ether_sprintf(sc->sc_curbssid),
sc->sc_curaid);
}
static void
ath_set_channel(struct ieee80211com *ic)
{
struct ifnet *ifp = ic->ic_ifp;
struct ath_softc *sc = ifp->if_softc;
(void) ath_chan_set(sc, ic->ic_curchan);
/*
* If we are returning to our bss channel then mark state
* so the next recv'd beacon's tsf will be used to sync the
* beacon timers. Note that since we only hear beacons in
* sta/ibss mode this has no effect in other operating modes.
*/
if (!sc->sc_scanning && ic->ic_curchan == ic->ic_bsschan)
sc->sc_syncbeacon = 1;
}
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
/*
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
* Walk the vap list and check if there any vap's in RUN state.
*/
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static int
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_isanyrunningvaps(struct ieee80211vap *this)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
{
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211com *ic = this->iv_ic;
struct ieee80211vap *vap;
IEEE80211_LOCK_ASSERT(ic);
TAILQ_FOREACH(vap, &ic->ic_vaps, iv_next) {
treat IEEE80211_S_CSA as a "running state"; this fixes ap mode 11h channel switch announcements
2009-06-03 17:25:19 +00:00
if (vap != this && vap->iv_state >= IEEE80211_S_RUN)
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
return 1;
}
return 0;
}
static int
ath_newstate(struct ieee80211vap *vap, enum ieee80211_state nstate, int arg)
{
struct ieee80211com *ic = vap->iv_ic;
struct ath_softc *sc = ic->ic_ifp->if_softc;
struct ath_vap *avp = ATH_VAP(vap);
track changes to 802.11 code: o override new_state method per new model o use ieee80211_state_name instead of private copy
2003-07-20 21:38:20 +00:00
struct ath_hal *ah = sc->sc_ah;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211_node *ni = NULL;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
int i, error, stamode;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
u_int32_t rfilt;
static const HAL_LED_STATE leds[] = {
HAL_LED_INIT, /* IEEE80211_S_INIT */
HAL_LED_SCAN, /* IEEE80211_S_SCAN */
HAL_LED_AUTH, /* IEEE80211_S_AUTH */
HAL_LED_ASSOC, /* IEEE80211_S_ASSOC */
fix led blinking in RUN state: the addition of the CAC state moved IEEE80211_S_RUN and broke the array lookup used to find the LED flags Approved by: re (blanket wireless)
2007-09-17 05:12:19 +00:00
HAL_LED_RUN, /* IEEE80211_S_CAC */
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
HAL_LED_RUN, /* IEEE80211_S_RUN */
fix led blinking in RUN state: the addition of the CAC state moved IEEE80211_S_RUN and broke the array lookup used to find the LED flags Approved by: re (blanket wireless)
2007-09-17 05:12:19 +00:00
HAL_LED_RUN, /* IEEE80211_S_CSA */
HAL_LED_RUN, /* IEEE80211_S_SLEEP */
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
};
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_STATE, "%s: %s -> %s\n", __func__,
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ieee80211_state_name[vap->iv_state],
Update with last year of work.
2004-12-08 17:34:36 +00:00
ieee80211_state_name[nstate]);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
replace if_watchdog w/ private callout; probably can merge this with the calibration work sometime in the future
2009-03-09 23:10:19 +00:00
callout_drain(&sc->sc_cal_ch);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_hal_setledstate(ah, leds[nstate]); /* set LED */
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (nstate == IEEE80211_S_SCAN) {
disable interrupts when transitioning to INIT state so we don't rx frames
2005-01-18 19:31:31 +00:00
/*
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
* Scanning: turn off beacon miss and don't beacon.
* Mark beacon state so when we reach RUN state we'll
* [re]setup beacons. Unblock the task q thread so
* deferred interrupt processing is done.
disable interrupts when transitioning to INIT state so we don't rx frames
2005-01-18 19:31:31 +00:00
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_hal_intrset(ah,
sc->sc_imask &~ (HAL_INT_SWBA | HAL_INT_BMISS));
Drain task q items when transitioning to INIT state; this closes a race seen on smp laptops when suspending where the rx task can be entered after the interface is detach'd. NB: use of taskqueue_drain while holding the softc mutex is problematic Submitted by: ambrisko MFC after: 1 month
2007-05-29 16:13:59 +00:00
sc->sc_imask &= ~(HAL_INT_SWBA | HAL_INT_BMISS);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
sc->sc_beacons = 0;
taskqueue_unblock(sc->sc_tq);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ni = vap->iv_bss;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
rfilt = ath_calcrxfilter(sc);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
stamode = (vap->iv_opmode == IEEE80211_M_STA ||
install bssid for ahdemo mode too
2008-10-27 17:54:17 +00:00
vap->iv_opmode == IEEE80211_M_AHDEMO ||
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
vap->iv_opmode == IEEE80211_M_IBSS);
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
if (stamode && nstate == IEEE80211_S_RUN) {
sc->sc_curaid = ni->ni_associd;
IEEE80211_ADDR_COPY(sc->sc_curbssid, ni->ni_bssid);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_hal_setassocid(ah, sc->sc_curbssid, sc->sc_curaid);
}
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
DPRINTF(sc, ATH_DEBUG_STATE, "%s: RX filter 0x%x bssid %s aid 0x%x\n",
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
__func__, rfilt, ether_sprintf(sc->sc_curbssid), sc->sc_curaid);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
ath_hal_setrxfilter(ah, rfilt);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/* XXX is this to restore keycache on resume? */
if (vap->iv_opmode != IEEE80211_M_STA &&
(vap->iv_flags & IEEE80211_F_PRIVACY)) {
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
for (i = 0; i < IEEE80211_WEP_NKID; i++)
if (ath_hal_keyisvalid(ah, i))
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
ath_hal_keysetmac(ah, i, ni->ni_bssid);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/*
* Invoke the parent method to do net80211 work.
*/
error = avp->av_newstate(vap, nstate, arg);
if (error != 0)
goto bad;
Update with last year of work.
2004-12-08 17:34:36 +00:00
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
if (nstate == IEEE80211_S_RUN) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/* NB: collect bss node again, it may have changed */
ni = vap->iv_bss;
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_STATE,
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
"%s(RUN): iv_flags 0x%08x bintvl %d bssid %s "
"capinfo 0x%04x chan %d\n", __func__,
vap->iv_flags, ni->ni_intval, ether_sprintf(ni->ni_bssid),
ni->ni_capinfo, ieee80211_chan2ieee(ic, ic->ic_curchan));
switch (vap->iv_opmode) {
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
case IEEE80211_M_AHDEMO:
if ((vap->iv_caps & IEEE80211_C_TDMA) == 0)
break;
/* fall thru... */
#endif
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
case IEEE80211_M_HOSTAP:
case IEEE80211_M_IBSS:
Implementation of the upcoming Wireless Mesh standard, 802.11s, on the net80211 wireless stack. This work is based on the March 2009 D3.0 draft standard. This standard is expected to become final next year. This includes two main net80211 modules, ieee80211_mesh.c which deals with peer link management, link metric calculation, routing table control and mesh configuration and ieee80211_hwmp.c which deals with the actually routing process on the mesh network. HWMP is the mandatory routing protocol on by the mesh standard, but others, such as RA-OLSR, can be implemented. Authentication and encryption are not implemented. There are several scripts under tools/tools/net80211/scripts that can be used to test different mesh network topologies and they also teach you how to setup a mesh vap (for the impatient: ifconfig wlan0 create wlandev ... wlanmode mesh). A new build option is available: IEEE80211_SUPPORT_MESH and it's enabled by default on GENERIC kernels for i386, amd64, sparc64 and pc98. Drivers that support mesh networks right now are: ath, ral and mwl. More information at: http://wiki.freebsd.org/WifiMesh Please note that this work is experimental. Also, please note that bridging a mesh vap with another network interface is not yet supported. Many thanks to the FreeBSD Foundation for sponsoring this project and to Sam Leffler for his support. Also, I would like to thank Gateworks Corporation for sending me a Cambria board which was used during the development of this project. Reviewed by: sam Approved by: re (kensmith) Obtained from: projects/mesh11s
2009-07-11 15:02:45 +00:00
case IEEE80211_M_MBSS:
beacon handling fixups for adhoc mode: o don't reclaim any previous beacon state in ath_beacon_alloc; do it explicitly in ath_newstate o reference count the node held in the beacon frame state block o process ibss merge more intelligently; let the state machine do the right thing instead of explicitly setting the new bssi id o explicitly stop tx dma before doing beacon setup to handle the ibss merge case
2005-01-24 20:05:03 +00:00
/*
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
* Allocate and setup the beacon frame.
*
beacon handling fixups for adhoc mode: o don't reclaim any previous beacon state in ath_beacon_alloc; do it explicitly in ath_newstate o reference count the node held in the beacon frame state block o process ibss merge more intelligently; let the state machine do the right thing instead of explicitly setting the new bssi id o explicitly stop tx dma before doing beacon setup to handle the ibss merge case
2005-01-24 20:05:03 +00:00
* Stop any previous beacon DMA. This may be
* necessary, for example, when an ibss merge
* causes reconfiguration; there will be a state
* transition from RUN->RUN that means we may
* be called with beacon transmission active.
*/
ath_hal_stoptxdma(ah, sc->sc_bhalq);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
error = ath_beacon_alloc(sc, ni);
if (error != 0)
goto bad;
add adhoc demo mode support MFC after: 2 weeks
2006-02-09 21:42:53 +00:00
/*
Beacon timer setup fixes: o pull nexttbtt forward in adhoc mode too o resync beacon timers on joining a bss or ibss as the tstamp we collected while scanning is almost certainly out of date Note we may need to refine the ibss mode check in ath_recv_mgmt. Reviewed by: avatar, dyoung Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:21:53 +00:00
* If joining an adhoc network defer beacon timer
* configuration to the next beacon frame so we
* have a current TSF to use. Otherwise we're
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
* starting an ibss/bss so there's no need to delay;
* if this is the first vap moving to RUN state, then
* beacon state needs to be [re]configured.
add adhoc demo mode support MFC after: 2 weeks
2006-02-09 21:42:53 +00:00
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (vap->iv_opmode == IEEE80211_M_IBSS &&
ni->ni_tstamp.tsf != 0) {
Beacon timer setup fixes: o pull nexttbtt forward in adhoc mode too o resync beacon timers on joining a bss or ibss as the tstamp we collected while scanning is almost certainly out of date Note we may need to refine the ibss mode check in ath_recv_mgmt. Reviewed by: avatar, dyoung Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:21:53 +00:00
sc->sc_syncbeacon = 1;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
} else if (!sc->sc_beacons) {
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
if (vap->iv_caps & IEEE80211_C_TDMA)
ath_tdma_config(sc, vap);
else
#endif
ath_beacon_config(sc, vap);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
sc->sc_beacons = 1;
}
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
break;
case IEEE80211_M_STA:
add adhoc demo mode support MFC after: 2 weeks
2006-02-09 21:42:53 +00:00
/*
Beacon timer setup fixes: o pull nexttbtt forward in adhoc mode too o resync beacon timers on joining a bss or ibss as the tstamp we collected while scanning is almost certainly out of date Note we may need to refine the ibss mode check in ath_recv_mgmt. Reviewed by: avatar, dyoung Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:21:53 +00:00
* Defer beacon timer configuration to the next
* beacon frame so we have a current TSF to use
* (any TSF collected when scanning is likely old).
add adhoc demo mode support MFC after: 2 weeks
2006-02-09 21:42:53 +00:00
*/
Beacon timer setup fixes: o pull nexttbtt forward in adhoc mode too o resync beacon timers on joining a bss or ibss as the tstamp we collected while scanning is almost certainly out of date Note we may need to refine the ibss mode check in ath_recv_mgmt. Reviewed by: avatar, dyoung Obtained from: atheros MFC after: 2 weeks
2006-02-09 22:21:53 +00:00
sc->sc_syncbeacon = 1;
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
break;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
case IEEE80211_M_MONITOR:
/*
* Monitor mode vaps have only INIT->RUN and RUN->RUN
* transitions so we must re-enable interrupts here to
* handle the case of a single monitor mode vap.
*/
ath_hal_intrset(ah, sc->sc_imask);
break;
case IEEE80211_M_WDS:
break;
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
default:
break;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update monitoring support: o record tsf in tx+rx frames o switch from raw rssi to dbm for signal data and record both signal and noise floor data (hacked for now to assume a fixed noise floor; is correct with new hal) o add monpass sysctl to control which rx'd frames are passed up with errors; especially useful to see frames with CRC errors o mark 'd packets w/ a CRC error with radiotap's BADFCS flag Also add placeholder code for calibrating the noise floor when using newer hals. Reviewed by: avatar MFC after: 1 week
2006-01-09 17:13:20 +00:00
/*
* Let the hal process statistics collected during a
* scan so it can provide calibrated noise floor data.
*/
ath_hal_process_noisefloor(ah);
move hal statistics to softc; the per-node stats are overkill, they're only used when operating in station mode MFC after: 2 weeks
2006-02-09 21:23:44 +00:00
/*
* Reset rssi stats; maybe not the best place...
*/
sc->sc_halstats.ns_avgbrssi = ATH_RSSI_DUMMY_MARKER;
sc->sc_halstats.ns_avgrssi = ATH_RSSI_DUMMY_MARKER;
sc->sc_halstats.ns_avgtxrssi = ATH_RSSI_DUMMY_MARKER;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/*
* Finally, start any timers and the task q thread
* (in case we didn't go through SCAN state).
*/
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
if (ath_longcalinterval != 0) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
/* start periodic recalibration timer */
New periodic calibration scheme needed for 11n parts that have multiple algorithms and potentially collect multiple samples. Instead of a single calibration interval we now have short and long intervals; the long interval roughly corresponds to the previous single interval. The short interval is used to speedup collection of samples and happens much quicker. We make calls using the short interval until we're told the calibration work is complete at which point we fallback to the long interval. In addition there is a much longer reset interval used to flush all calibration state and cause everthing to start anew. With these changes you can also disable calibration entirely by setting the long interval to zero.
2008-12-07 19:26:34 +00:00
callout_reset(&sc->sc_cal_ch, 1, ath_calibrate, sc);
} else {
DPRINTF(sc, ATH_DEBUG_CALIBRATE,
"%s: calibration disabled\n", __func__);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
}
taskqueue_unblock(sc->sc_tq);
} else if (nstate == IEEE80211_S_INIT) {
/*
* If there are no vaps left in RUN state then
* shutdown host/driver operation:
* o disable interrupts
* o disable the task queue thread
* o mark beacon processing as stopped
*/
if (!ath_isanyrunningvaps(vap)) {
sc->sc_imask &= ~(HAL_INT_SWBA | HAL_INT_BMISS);
/* disable interrupts */
ath_hal_intrset(ah, sc->sc_imask &~ HAL_INT_GLOBAL);
taskqueue_block(sc->sc_tq);
sc->sc_beacons = 0;
}
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
ath_hal_setcca(ah, AH_TRUE);
#endif
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
bad:
return error;
}
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
/*
* Allocate a key cache slot to the station so we can
* setup a mapping from key index to node. The key cache
* slot is needed for managing antenna state and for
* compression when stations do not use crypto. We do
* it uniliaterally here; if crypto is employed this slot
* will be reassigned.
*/
static void
ath_setup_stationkey(struct ieee80211_node *ni)
{
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211vap *vap = ni->ni_vap;
struct ath_softc *sc = vap->iv_ic->ic_ifp->if_softc;
Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks
2005-08-08 18:46:36 +00:00
ieee80211_keyix keyix, rxkeyix;
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (!ath_key_alloc(vap, &ni->ni_ucastkey, &keyix, &rxkeyix)) {
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
/*
* Key cache is full; we'll fall back to doing
* the more expensive lookup in software. Note
* this also means no h/w compression.
*/
/* XXX msg+statistic */
} else {
Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks
2005-08-08 18:46:36 +00:00
/* XXX locking? */
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
ni->ni_ucastkey.wk_keyix = keyix;
Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks
2005-08-08 18:46:36 +00:00
ni->ni_ucastkey.wk_rxkeyix = rxkeyix;
mark the CLR key installed for open auth stations such that it is reclaimed when net80211 tears down station state; without this we leak keycache slots
2009-02-10 19:27:50 +00:00
/* NB: must mark device key to get called back on delete */
ni->ni_ucastkey.wk_flags |= IEEE80211_KEY_DEVKEY;
Cleanup power handling and fix suspend/resume: o do not put the chip into full sleep in ath_stop as it gains nothing and causes many parts to hang in ath_detach because we may touch the chip during vap teardown; this may also fix issues with unloading the module o add a note in ath_detach to explain ath_hal_detach puts the chip in low power mode; this is useful to know as it means unloading the module will place a pci device in the lowest possible power state o leave an #ifdef notyet marker for powering down the chip when a device is marked down; we can't do that until we handle all the ways the driver may be entered and touch the chip o fix resume by reloading the h/w key cache as it's been clobbered (for pci) by the socket being powered off; for station mode we directly stop+init the chip and then simulate a beacon miss to get the upper layers sync'd up; for other configs we must brute force stop+start the vaps so they go through the state machine
2008-05-29 00:10:48 +00:00
IEEE80211_ADDR_COPY(ni->ni_ucastkey.wk_macaddr, ni->ni_macaddr);
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
/* NB: this will create a pass-thru key entry */
Cleanup power handling and fix suspend/resume: o do not put the chip into full sleep in ath_stop as it gains nothing and causes many parts to hang in ath_detach because we may touch the chip during vap teardown; this may also fix issues with unloading the module o add a note in ath_detach to explain ath_hal_detach puts the chip in low power mode; this is useful to know as it means unloading the module will place a pci device in the lowest possible power state o leave an #ifdef notyet marker for powering down the chip when a device is marked down; we can't do that until we handle all the ways the driver may be entered and touch the chip o fix resume by reloading the h/w key cache as it's been clobbered (for pci) by the socket being powered off; for station mode we directly stop+init the chip and then simulate a beacon miss to get the upper layers sync'd up; for other configs we must brute force stop+start the vaps so they go through the state machine
2008-05-29 00:10:48 +00:00
ath_keyset(sc, &ni->ni_ucastkey, vap->iv_bss);
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
}
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
/*
* Setup driver-specific state for a newly associated node.
* Note that we're called also on a re-associate, the isnew
* param tells us if this is the first time or not.
*/
static void
simplify ic_newassoc callback MFC after: 3 days
2005-07-22 17:57:16 +00:00
ath_newassoc(struct ieee80211_node *ni, int isnew)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
{
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ath_node *an = ATH_NODE(ni);
struct ieee80211vap *vap = ni->ni_vap;
struct ath_softc *sc = vap->iv_ic->ic_ifp->if_softc;
remove driver-private equivalent of ni_txparms; it's now superfluous
2008-10-27 17:03:24 +00:00
const struct ieee80211_txparam *tp = ni->ni_txparms;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
optimize ath_tx_findrix: there's no need to walk the rates table as sc_rixmap is an inverse map NB: could eliminate the check for an invalid rate by filling in 0 for invalid entries but the rate control modules use it to identify bogus rates so leave it for now
2009-05-07 00:35:32 +00:00
an->an_mcastrix = ath_tx_findrix(sc, tp->mcastrate);
an->an_mgmtrix = ath_tx_findrix(sc, tp->mgmtrate);
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ath_rate_newassoc(sc, an, isnew);
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
if (isnew &&
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
(vap->iv_flags & IEEE80211_F_PRIVACY) == 0 && sc->sc_hasclrkey &&
ni->ni_ucastkey.wk_keyix == IEEE80211_KEYIX_NONE)
Misc keycache changes: o purge ath_initkeytable; it's not needed o add multicast key search support for supporting multiple group keys (disabled for now; requires updated hal) o create keycache entry for stations using open auth so they get h/w antenna management support o add keycache -> node mapping table; eliminates mac-based lookup in the net80211 layer
2005-06-06 16:39:21 +00:00
ath_setup_stationkey(ni);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
static int
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
ath_setregdomain(struct ieee80211com *ic, struct ieee80211_regdomain *reg,
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
int nchans, struct ieee80211_channel chans[])
{
struct ath_softc *sc = ic->ic_ifp->if_softc;
struct ath_hal *ah = sc->sc_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
HAL_STATUS status;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
add regdomain debug msgs
2008-10-27 17:35:09 +00:00
DPRINTF(sc, ATH_DEBUG_REGDOMAIN,
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
"%s: rd %u cc %u location %c%s\n",
__func__, reg->regdomain, reg->country, reg->location,
reg->ecm ? " ecm" : "");
status = ath_hal_set_channels(ah, chans, nchans,
reg->country, reg->regdomain);
if (status != HAL_OK) {
DPRINTF(sc, ATH_DEBUG_REGDOMAIN, "%s: failed, status %u\n",
__func__, status);
return EINVAL; /* XXX */
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
}
return 0;
}
static void
ath_getradiocaps(struct ieee80211com *ic,
change ic_getradiocaps driver callback to include the max # channels so callers know the size of the array passed down
2009-01-27 23:19:36 +00:00
int maxchans, int *nchans, struct ieee80211_channel chans[])
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
{
struct ath_softc *sc = ic->ic_ifp->if_softc;
struct ath_hal *ah = sc->sc_ah;
add regdomain debug msgs
2008-10-27 17:35:09 +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
DPRINTF(sc, ATH_DEBUG_REGDOMAIN, "%s: use rd %u cc %d\n",
__func__, SKU_DEBUG, CTRY_DEFAULT);
add regdomain debug msgs
2008-10-27 17:35:09 +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
/* XXX check return */
(void) ath_hal_getchannels(ah, chans, maxchans, nchans,
HAL_MODE_ALL, CTRY_DEFAULT, SKU_DEBUG, AH_TRUE);
restore the hal's channel list when doing getradiocaps so it's in sync with the 802.11 layer's list
2008-04-27 22:03:56 +00:00
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
}
static int
ath_getchannels(struct ath_softc *sc)
{
struct ifnet *ifp = sc->sc_ifp;
struct ieee80211com *ic = ifp->if_l2com;
struct ath_hal *ah = sc->sc_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
HAL_STATUS status;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +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
* Collect channel set based on EEPROM contents.
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +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
status = ath_hal_init_channels(ah, ic->ic_channels, IEEE80211_CHAN_MAX,
&ic->ic_nchans, HAL_MODE_ALL, CTRY_DEFAULT, SKU_NONE, AH_TRUE);
if (status != HAL_OK) {
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if_printf(ifp, "%s: unable to collect channel list from hal, "
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
"status %d\n", __func__, status);
return EINVAL;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +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
(void) ath_hal_getregdomain(ah, &sc->sc_eerd);
ath_hal_getcountrycode(ah, &sc->sc_eecc); /* NB: cannot fail */
/* XXX map Atheros sku's to net80211 SKU's */
/* XXX net80211 types too small */
ic->ic_regdomain.regdomain = (uint16_t) sc->sc_eerd;
ic->ic_regdomain.country = (uint16_t) sc->sc_eecc;
ic->ic_regdomain.isocc[0] = ' '; /* XXX don't know */
ic->ic_regdomain.isocc[1] = ' ';
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
ic->ic_regdomain.ecm = 1;
ic->ic_regdomain.location = 'I';
add regdomain debug msgs
2008-10-27 17:35:09 +00:00
DPRINTF(sc, ATH_DEBUG_REGDOMAIN,
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
"%s: eeprom rd %u cc %u (mapped rd %u cc %u) location %c%s\n",
add regdomain debug msgs
2008-10-27 17:35:09 +00:00
__func__, sc->sc_eerd, sc->sc_eecc,
ic->ic_regdomain.regdomain, ic->ic_regdomain.country,
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
ic->ic_regdomain.location, ic->ic_regdomain.ecm ? " ecm" : "");
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
return 0;
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
static void
better led blinking
2005-01-18 19:03:04 +00:00
ath_led_done(void *arg)
Update with last year of work.
2004-12-08 17:34:36 +00:00
{
better led blinking
2005-01-18 19:03:04 +00:00
struct ath_softc *sc = arg;
Update with last year of work.
2004-12-08 17:34:36 +00:00
better led blinking
2005-01-18 19:03:04 +00:00
sc->sc_blinking = 0;
}
/*
* Turn the LED off: flip the pin and then set a timer so no
* update will happen for the specified duration.
*/
static void
ath_led_off(void *arg)
{
struct ath_softc *sc = arg;
ath_hal_gpioset(sc->sc_ah, sc->sc_ledpin, !sc->sc_ledon);
callout_reset(&sc->sc_ledtimer, sc->sc_ledoff, ath_led_done, sc);
}
/*
* Blink the LED according to the specified on/off times.
*/
static void
ath_led_blink(struct ath_softc *sc, int on, int off)
{
DPRINTF(sc, ATH_DEBUG_LED, "%s: on %u off %u\n", __func__, on, off);
ath_hal_gpioset(sc->sc_ah, sc->sc_ledpin, sc->sc_ledon);
sc->sc_blinking = 1;
sc->sc_ledoff = off;
callout_reset(&sc->sc_ledtimer, on, ath_led_off, sc);
}
static void
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
ath_led_event(struct ath_softc *sc, int rix)
better led blinking
2005-01-18 19:03:04 +00:00
{
sc->sc_ledevent = ticks; /* time of last event */
if (sc->sc_blinking) /* don't interrupt active blink */
return;
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
ath_led_blink(sc, sc->sc_hwmap[rix].ledon, sc->sc_hwmap[rix].ledoff);
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static int
ath_rate_setup(struct ath_softc *sc, u_int mode)
{
struct ath_hal *ah = sc->sc_ah;
const HAL_RATE_TABLE *rt;
switch (mode) {
case IEEE80211_MODE_11A:
cleanup rate setup MFC after: 2 weeks
2006-02-09 21:31:48 +00:00
rt = ath_hal_getratetable(ah, HAL_MODE_11A);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
break;
Add initial support for 900MHz cards like the Ubiquiti SR9: o eliminate assumptions that half/quarter rate channels on exist in 11a o handle frequency mapping between hal and net80211; hal gives us freq's in the range 2422..2437 that we remap MFC after: 1 month
2007-01-15 01:15:57 +00:00
case IEEE80211_MODE_HALF:
Add half/quarter rate 11a channel support: o change handling of regdomain-related mib knobs so they can be set post-attach: regdomain, countrycode, outdoor, and xchanmode; the hal will not permit changing the regdomain but we expose it for now o on regdomain/countrycode change recalculate the channel list and push it to the net80211 layer (NB: looks to need more tweaking) o setup rate tables for half/quarter rate channels o honor half/quarter rate channel configs when changing channels o honor half/quarter rate channel configs when setting the slot time o use hack/nonstandard channel numbering scheme for the public safety band to avoid overlapping 2.4G channels on dual-band cards o remove setup of ic_sup_rates; the net80211 layer can do this for us and it simplifies handling of half/quarter rate channels Tested only in Public Safety Band with cards that have RF5112.
2006-12-27 19:07:09 +00:00
rt = ath_hal_getratetable(ah, HAL_MODE_11A_HALF_RATE);
break;
Add initial support for 900MHz cards like the Ubiquiti SR9: o eliminate assumptions that half/quarter rate channels on exist in 11a o handle frequency mapping between hal and net80211; hal gives us freq's in the range 2422..2437 that we remap MFC after: 1 month
2007-01-15 01:15:57 +00:00
case IEEE80211_MODE_QUARTER:
Add half/quarter rate 11a channel support: o change handling of regdomain-related mib knobs so they can be set post-attach: regdomain, countrycode, outdoor, and xchanmode; the hal will not permit changing the regdomain but we expose it for now o on regdomain/countrycode change recalculate the channel list and push it to the net80211 layer (NB: looks to need more tweaking) o setup rate tables for half/quarter rate channels o honor half/quarter rate channel configs when changing channels o honor half/quarter rate channel configs when setting the slot time o use hack/nonstandard channel numbering scheme for the public safety band to avoid overlapping 2.4G channels on dual-band cards o remove setup of ic_sup_rates; the net80211 layer can do this for us and it simplifies handling of half/quarter rate channels Tested only in Public Safety Band with cards that have RF5112.
2006-12-27 19:07:09 +00:00
rt = ath_hal_getratetable(ah, HAL_MODE_11A_QUARTER_RATE);
break;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
case IEEE80211_MODE_11B:
cleanup rate setup MFC after: 2 weeks
2006-02-09 21:31:48 +00:00
rt = ath_hal_getratetable(ah, HAL_MODE_11B);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
break;
case IEEE80211_MODE_11G:
cleanup rate setup MFC after: 2 weeks
2006-02-09 21:31:48 +00:00
rt = ath_hal_getratetable(ah, HAL_MODE_11G);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
break;
Update with last year of work.
2004-12-08 17:34:36 +00:00
case IEEE80211_MODE_TURBO_A:
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
rt = ath_hal_getratetable(ah, HAL_MODE_108A);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
break;
Update with last year of work.
2004-12-08 17:34:36 +00:00
case IEEE80211_MODE_TURBO_G:
cleanup rate setup MFC after: 2 weeks
2006-02-09 21:31:48 +00:00
rt = ath_hal_getratetable(ah, HAL_MODE_108G);
Update with last year of work.
2004-12-08 17:34:36 +00:00
break;
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
case IEEE80211_MODE_STURBO_A:
rt = ath_hal_getratetable(ah, HAL_MODE_TURBO);
break;
case IEEE80211_MODE_11NA:
rt = ath_hal_getratetable(ah, HAL_MODE_11NA_HT20);
break;
case IEEE80211_MODE_11NG:
rt = ath_hal_getratetable(ah, HAL_MODE_11NG_HT20);
break;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
default:
Update with last year of work.
2004-12-08 17:34:36 +00:00
DPRINTF(sc, ATH_DEBUG_ANY, "%s: invalid mode %u\n",
__func__, mode);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
return 0;
}
cleanup rate setup MFC after: 2 weeks
2006-02-09 21:31:48 +00:00
sc->sc_rates[mode] = rt;
Add half/quarter rate 11a channel support: o change handling of regdomain-related mib knobs so they can be set post-attach: regdomain, countrycode, outdoor, and xchanmode; the hal will not permit changing the regdomain but we expose it for now o on regdomain/countrycode change recalculate the channel list and push it to the net80211 layer (NB: looks to need more tweaking) o setup rate tables for half/quarter rate channels o honor half/quarter rate channel configs when changing channels o honor half/quarter rate channel configs when setting the slot time o use hack/nonstandard channel numbering scheme for the public safety band to avoid overlapping 2.4G channels on dual-band cards o remove setup of ic_sup_rates; the net80211 layer can do this for us and it simplifies handling of half/quarter rate channels Tested only in Public Safety Band with cards that have RF5112.
2006-12-27 19:07:09 +00:00
return (rt != NULL);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
static void
ath_setcurmode(struct ath_softc *sc, enum ieee80211_phymode mode)
{
better led blinking
2005-01-18 19:03:04 +00:00
#define N(a) (sizeof(a)/sizeof(a[0]))
/* NB: on/off times from the Atheros NDIS driver, w/ permission */
static const struct {
u_int rate; /* tx/rx 802.11 rate */
u_int16_t timeOn; /* LED on time (ms) */
u_int16_t timeOff; /* LED off time (ms) */
} blinkrates[] = {
{ 108, 40, 10 },
{ 96, 44, 11 },
{ 72, 50, 13 },
{ 48, 57, 14 },
{ 36, 67, 16 },
{ 24, 80, 20 },
{ 22, 100, 25 },
{ 18, 133, 34 },
{ 12, 160, 40 },
{ 10, 200, 50 },
{ 6, 240, 58 },
{ 4, 267, 66 },
{ 2, 400, 100 },
{ 0, 500, 130 },
Add initial support for 900MHz cards like the Ubiquiti SR9: o eliminate assumptions that half/quarter rate channels on exist in 11a o handle frequency mapping between hal and net80211; hal gives us freq's in the range 2422..2437 that we remap MFC after: 1 month
2007-01-15 01:15:57 +00:00
/* XXX half/quarter rates */
better led blinking
2005-01-18 19:03:04 +00:00
};
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
const HAL_RATE_TABLE *rt;
better led blinking
2005-01-18 19:03:04 +00:00
int i, j;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
memset(sc->sc_rixmap, 0xff, sizeof(sc->sc_rixmap));
rt = sc->sc_rates[mode];
KASSERT(rt != NULL, ("no h/w rate set for phy mode %u", mode));
fix handling of HT rates; these overlap legacy rates and need to be marked as MCS in the inverse mapping table
2008-10-27 18:02:47 +00:00
for (i = 0; i < rt->rateCount; i++) {
uint8_t ieeerate = rt->info[i].dot11Rate & IEEE80211_RATE_VAL;
if (rt->info[i].phy != IEEE80211_T_HT)
sc->sc_rixmap[ieeerate] = i;
else
sc->sc_rixmap[ieeerate | IEEE80211_RATE_MCS] = i;
}
maintain a table for mapping hardware rate codes to 802.11 rates for calculating the rate for each rx'd frame
2003-08-19 21:24:16 +00:00
memset(sc->sc_hwmap, 0, sizeof(sc->sc_hwmap));
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
for (i = 0; i < N(sc->sc_hwmap); i++) {
if (i >= rt->rateCount) {
better led blinking
2005-01-18 19:03:04 +00:00
sc->sc_hwmap[i].ledon = (500 * hz) / 1000;
sc->sc_hwmap[i].ledoff = (130 * hz) / 1000;
Radiotap fixups: o catch one place where we were not using ath_chan_change to switch channels; this fixes a problem where the channel settings were not being correctly reported in captured packets o return unique channel identification in the channel flags; ethereal gets confused if you return merged flags (e.g. ofdm, cck, and 2Ghz) (this is workaround and should be removed if we can ever cleanup radiotap consumers) o correct short/long preamble flag state for rx and treat tx the same--use a new hwflags array that gives us the data based on the h/w rate index/cookie o add gross hack to handle radiotap capture of frames that come in with hardware padding; should be replaced by a flag in the radiotap header and more smarts in the apps that decode radiotap data
2004-12-31 20:32:40 +00:00
continue;
better led blinking
2005-01-18 19:03:04 +00:00
}
sc->sc_hwmap[i].ieeerate =
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
rt->info[i].dot11Rate & IEEE80211_RATE_VAL;
if (rt->info[i].phy == IEEE80211_T_HT)
prefer #define to naked constant
2008-10-27 18:05:26 +00:00
sc->sc_hwmap[i].ieeerate |= IEEE80211_RATE_MCS;
Fixup radiotap handling of FCS and QoS frames per discussion with David Young: o mark rx frames including FCS in the payload with the IEEE80211_RADIOTAP_F_FCS flag o remove hack to copy 802.11 headers with padding out of line; instead mark the frames with IEEE80211_RADIOTAP_F_DATAPAD and require applications to do the work o split precalculated radiotap flags into tx+rx now that they can be different Note the full usefulness of these changes depends on updates to applications that process radiotap data.
2005-01-24 20:31:24 +00:00
sc->sc_hwmap[i].txflags = IEEE80211_RADIOTAP_F_DATAPAD;
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
if (rt->info[i].shortPreamble ||
rt->info[i].phy == IEEE80211_T_OFDM)
Fixup radiotap handling of FCS and QoS frames per discussion with David Young: o mark rx frames including FCS in the payload with the IEEE80211_RADIOTAP_F_FCS flag o remove hack to copy 802.11 headers with padding out of line; instead mark the frames with IEEE80211_RADIOTAP_F_DATAPAD and require applications to do the work o split precalculated radiotap flags into tx+rx now that they can be different Note the full usefulness of these changes depends on updates to applications that process radiotap data.
2005-01-24 20:31:24 +00:00
sc->sc_hwmap[i].txflags |= IEEE80211_RADIOTAP_F_SHORTPRE;
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
sc->sc_hwmap[i].rxflags = sc->sc_hwmap[i].txflags;
better led blinking
2005-01-18 19:03:04 +00:00
for (j = 0; j < N(blinkrates)-1; j++)
if (blinkrates[j].rate == sc->sc_hwmap[i].ieeerate)
break;
/* NB: this uses the last entry if the rate isn't found */
/* XXX beware of overlow */
sc->sc_hwmap[i].ledon = (blinkrates[j].timeOn * hz) / 1000;
sc->sc_hwmap[i].ledoff = (blinkrates[j].timeOff * hz) / 1000;
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
sc->sc_currates = rt;
sc->sc_curmode = mode;
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* All protection frames are transmited at 2Mb/s for
* 11g, otherwise at 1Mb/s.
*/
lookup the protection tx rate index in the rate tables instead of using a known value MFC after: 2 weeks
2006-02-09 21:17:28 +00:00
if (mode == IEEE80211_MODE_11G)
optimize ath_tx_findrix: there's no need to walk the rates table as sc_rixmap is an inverse map NB: could eliminate the check for an invalid rate by filling in 0 for invalid entries but the rate control modules use it to identify bogus rates so leave it for now
2009-05-07 00:35:32 +00:00
sc->sc_protrix = ath_tx_findrix(sc, 2*2);
lookup the protection tx rate index in the rate tables instead of using a known value MFC after: 2 weeks
2006-02-09 21:17:28 +00:00
else
optimize ath_tx_findrix: there's no need to walk the rates table as sc_rixmap is an inverse map NB: could eliminate the check for an invalid rate by filling in 0 for invalid entries but the rate control modules use it to identify bogus rates so leave it for now
2009-05-07 00:35:32 +00:00
sc->sc_protrix = ath_tx_findrix(sc, 2*1);
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
/* NB: caller is responsible for resetting rate control state */
better led blinking
2005-01-18 19:03:04 +00:00
#undef N
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
static void
replace if_watchdog w/ private callout; probably can merge this with the calibration work sometime in the future
2009-03-09 23:10:19 +00:00
ath_watchdog(void *arg)
Update with last year of work.
2004-12-08 17:34:36 +00:00
{
replace if_watchdog w/ private callout; probably can merge this with the calibration work sometime in the future
2009-03-09 23:10:19 +00:00
struct ath_softc *sc = arg;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
replace if_watchdog w/ private callout; probably can merge this with the calibration work sometime in the future
2009-03-09 23:10:19 +00:00
if (sc->sc_wd_timer != 0 && --sc->sc_wd_timer == 0) {
struct ifnet *ifp = sc->sc_ifp;
some of the 11n parts can hang under certain conditions without necessary workarounds, add code to detect these hangs and distinguish them from other events; note this code is only invoked for anomalous conditions and (at the moment) is a noop because the hang detection code is in a new hal that's coming shortly
2008-11-30 18:34:27 +00:00
uint32_t hangs;
if (ath_hal_gethangstate(sc->sc_ah, 0xffff, &hangs) &&
hangs != 0) {
if_printf(ifp, "%s hang detected (0x%x)\n",
Correct spelling of reseting (found while researching the "bb hang detected" messages that are plaguing me). While I'm here, delete trailing whitespace.
2010-02-19 18:23:45 +00:00
hangs & 0xff ? "bb" : "mac", hangs);
some of the 11n parts can hang under certain conditions without necessary workarounds, add code to detect these hangs and distinguish them from other events; note this code is only invoked for anomalous conditions and (at the moment) is a noop because the hang detection code is in a new hal that's coming shortly
2008-11-30 18:34:27 +00:00
} else
if_printf(ifp, "device timeout\n");
Update 802.11 wireless support: o major overhaul of the way channels are handled: channels are now fully enumerated and uniquely identify the operating characteristics; these changes are visible to user applications which require changes o make scanning support independent of the state machine to enable background scanning and roaming o move scanning support into loadable modules based on the operating mode to enable different policies and reduce the memory footprint on systems w/ constrained resources o add background scanning in station mode (no support for adhoc/ibss mode yet) o significantly speedup sta mode scanning with a variety of techniques o add roaming support when background scanning is supported; for now we use a simple algorithm to trigger a roam: we threshold the rssi and tx rate, if either drops too low we try to roam to a new ap o add tx fragmentation support o add first cut at 802.11n support: this code works with forthcoming drivers but is incomplete; it's included now to establish a baseline for other drivers to be developed and for user applications o adjust max_linkhdr et. al. to reflect 802.11 requirements; this eliminates prepending mbufs for traffic generated locally o add support for Atheros protocol extensions; mainly the fast frames encapsulation (note this can be used with any card that can tx+rx large frames correctly) o add sta support for ap's that beacon both WPA1+2 support o change all data types from bsd-style to posix-style o propagate noise floor data from drivers to net80211 and on to user apps o correct various issues in the sta mode state machine related to handling authentication and association failures o enable the addition of sta mode power save support for drivers that need net80211 support (not in this commit) o remove old WI compatibility ioctls (wicontrol is officially dead) o change the data structures returned for get sta info and get scan results so future additions will not break user apps o fixed tx rate is now maintained internally as an ieee rate and not an index into the rate set; this needs to be extended to deal with multi-mode operation o add extended channel specifications to radiotap to enable 11n sniffing Drivers: o ath: add support for bg scanning, tx fragmentation, fast frames, dynamic turbo (lightly tested), 11n (sniffing only and needs new hal) o awi: compile tested only o ndis: lightly tested o ipw: lightly tested o iwi: add support for bg scanning (well tested but may have some rough edges) o ral, ural, rum: add suppoort for bg scanning, calibrate rssi data o wi: lightly tested This work is based on contributions by Atheros, kmacy, sephe, thompsa, mlaier, kevlo, and others. Much of the scanning work was supported by Atheros. The 11n work was supported by Marvell.
2007-06-11 03:36:55 +00:00
ath_reset(ifp);
ifp->if_oerrors++;
sc->sc_stats.ast_watchdog++;
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
replace if_watchdog w/ private callout; probably can merge this with the calibration work sometime in the future
2009-03-09 23:10:19 +00:00
callout_schedule(&sc->sc_wd_ch, hz);
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
o add opt_ath.h enable tweaking various config parameters for the driver without modifying the source code o default debug msgs and diag support to off MFC after: 3 days
2006-04-03 18:14:02 +00:00
#ifdef ATH_DIAGAPI
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Diagnostic interface to the HAL. This is used by various
* tools to do things like retrieve register contents for
* debugging. The mechanism is intentionally opaque so that
* it can change frequently w/o concern for compatiblity.
*/
static int
ath_ioctl_diag(struct ath_softc *sc, struct ath_diag *ad)
{
struct ath_hal *ah = sc->sc_ah;
u_int id = ad->ad_id & ATH_DIAG_ID;
void *indata = NULL;
void *outdata = NULL;
u_int32_t insize = ad->ad_in_size;
u_int32_t outsize = ad->ad_out_size;
int error = 0;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (ad->ad_id & ATH_DIAG_IN) {
/*
* Copy in data.
*/
indata = malloc(insize, M_TEMP, M_NOWAIT);
if (indata == NULL) {
error = ENOMEM;
goto bad;
}
error = copyin(ad->ad_in_data, indata, insize);
if (error)
goto bad;
}
if (ad->ad_id & ATH_DIAG_DYN) {
/*
* Allocate a buffer for the results (otherwise the HAL
* returns a pointer to a buffer where we can read the
* results). Note that we depend on the HAL leaving this
* pointer for us to use below in reclaiming the buffer;
* may want to be more defensive.
*/
outdata = malloc(outsize, M_TEMP, M_NOWAIT);
if (outdata == NULL) {
error = ENOMEM;
goto bad;
}
}
if (ath_hal_getdiagstate(ah, id, indata, insize, &outdata, &outsize)) {
if (outsize < ad->ad_out_size)
ad->ad_out_size = outsize;
if (outdata != NULL)
error = copyout(outdata, ad->ad_out_data,
ad->ad_out_size);
} else {
error = EINVAL;
}
bad:
if ((ad->ad_id & ATH_DIAG_IN) && indata != NULL)
free(indata, M_TEMP);
if ((ad->ad_id & ATH_DIAG_DYN) && outdata != NULL)
free(outdata, M_TEMP);
return error;
}
o add opt_ath.h enable tweaking various config parameters for the driver without modifying the source code o default debug msgs and diag support to off MFC after: 3 days
2006-04-03 18:14:02 +00:00
#endif /* ATH_DIAGAPI */
Update with last year of work.
2004-12-08 17:34:36 +00:00
static int
ath_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
{
#define IS_RUNNING(ifp) \
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
((ifp->if_flags & IFF_UP) && (ifp->if_drv_flags & IFF_DRV_RUNNING))
Update with last year of work.
2004-12-08 17:34:36 +00:00
struct ath_softc *sc = ifp->if_softc;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
struct ieee80211com *ic = ifp->if_l2com;
Update with last year of work.
2004-12-08 17:34:36 +00:00
struct ifreq *ifr = (struct ifreq *)data;
Fixup statistics: o update tx rssi data only when an ACK was received o return tx rssi from sampled data instead of the last frame o track noise floor o return rx rssi and noise floor (was broken)
2008-10-27 17:51:24 +00:00
const HAL_RATE_TABLE *rt;
Update with last year of work.
2004-12-08 17:34:36 +00:00
int error = 0;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
switch (cmd) {
case SIOCSIFFLAGS:
Unify all the wifi *_ioctl routines - Limit grabbing the lock to SIOCSIFFLAGS. - Move ieee80211_start_all() to SIOCSIFFLAGS. - Remove SIOCSIFMEDIA as it is not useful. - Limit ether_ioctl to only SIOCGIFADDR. SIOCSIFADDR and SIOCSIFMTU have no affect as there is no input/output path in the vap parent. The vap code will handle the reinit of the mac address changes. - Split off ndis_ioctl_80211 as it was getting too different to wired devices. This fixes a copyout while locked and a lock recursion. Reviewed by: sam
2008-05-01 04:55:00 +00:00
ATH_LOCK(sc);
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (IS_RUNNING(ifp)) {
/*
* To avoid rescanning another access point,
* do not call ath_init() here. Instead,
* only reflect promisc mode settings.
*/
ath_mode_init(sc);
} else if (ifp->if_flags & IFF_UP) {
/*
* Beware of being called during attach/detach
* to reset promiscuous mode. In that case we
* will still be marked UP but not RUNNING.
* However trying to re-init the interface
* is the wrong thing to do as we've already
* torn down much of our state. There's
* probably a better way to deal with this.
*/
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
if (!sc->sc_invalid)
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
ath_init(sc); /* XXX lose error */
Cleanup power handling and fix suspend/resume: o do not put the chip into full sleep in ath_stop as it gains nothing and causes many parts to hang in ath_detach because we may touch the chip during vap teardown; this may also fix issues with unloading the module o add a note in ath_detach to explain ath_hal_detach puts the chip in low power mode; this is useful to know as it means unloading the module will place a pci device in the lowest possible power state o leave an #ifdef notyet marker for powering down the chip when a device is marked down; we can't do that until we handle all the ways the driver may be entered and touch the chip o fix resume by reloading the h/w key cache as it's been clobbered (for pci) by the socket being powered off; for station mode we directly stop+init the chip and then simulate a beacon miss to get the upper layers sync'd up; for other configs we must brute force stop+start the vaps so they go through the state machine
2008-05-29 00:10:48 +00:00
} else {
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_stop_locked(ifp);
Cleanup power handling and fix suspend/resume: o do not put the chip into full sleep in ath_stop as it gains nothing and causes many parts to hang in ath_detach because we may touch the chip during vap teardown; this may also fix issues with unloading the module o add a note in ath_detach to explain ath_hal_detach puts the chip in low power mode; this is useful to know as it means unloading the module will place a pci device in the lowest possible power state o leave an #ifdef notyet marker for powering down the chip when a device is marked down; we can't do that until we handle all the ways the driver may be entered and touch the chip o fix resume by reloading the h/w key cache as it's been clobbered (for pci) by the socket being powered off; for station mode we directly stop+init the chip and then simulate a beacon miss to get the upper layers sync'd up; for other configs we must brute force stop+start the vaps so they go through the state machine
2008-05-29 00:10:48 +00:00
#ifdef notyet
/* XXX must wakeup in places like ath_vap_delete */
if (!sc->sc_invalid)
ath_hal_setpower(sc->sc_ah, HAL_PM_FULL_SLEEP);
#endif
}
Unify all the wifi *_ioctl routines - Limit grabbing the lock to SIOCSIFFLAGS. - Move ieee80211_start_all() to SIOCSIFFLAGS. - Remove SIOCSIFMEDIA as it is not useful. - Limit ether_ioctl to only SIOCGIFADDR. SIOCSIFADDR and SIOCSIFMTU have no affect as there is no input/output path in the vap parent. The vap code will handle the reinit of the mac address changes. - Split off ndis_ioctl_80211 as it was getting too different to wired devices. This fixes a copyout while locked and a lock recursion. Reviewed by: sam
2008-05-01 04:55:00 +00:00
ATH_UNLOCK(sc);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
break;
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
case SIOCGIFMEDIA:
case SIOCSIFMEDIA:
error = ifmedia_ioctl(ifp, ifr, &ic->ic_media, cmd);
break;
Update with last year of work.
2004-12-08 17:34:36 +00:00
case SIOCGATHSTATS:
/* NB: embed these numbers to get a consistent view */
sc->sc_stats.ast_tx_packets = ifp->if_opackets;
sc->sc_stats.ast_rx_packets = ifp->if_ipackets;
Fixup statistics: o update tx rssi data only when an ACK was received o return tx rssi from sampled data instead of the last frame o track noise floor o return rx rssi and noise floor (was broken)
2008-10-27 17:51:24 +00:00
sc->sc_stats.ast_tx_rssi = ATH_RSSI(sc->sc_halstats.ns_avgtxrssi);
sc->sc_stats.ast_rx_rssi = ATH_RSSI(sc->sc_halstats.ns_avgrssi);
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
sc->sc_stats.ast_tdma_tsfadjp = TDMA_AVG(sc->sc_avgtsfdeltap);
sc->sc_stats.ast_tdma_tsfadjm = TDMA_AVG(sc->sc_avgtsfdeltam);
#endif
Fixup statistics: o update tx rssi data only when an ACK was received o return tx rssi from sampled data instead of the last frame o track noise floor o return rx rssi and noise floor (was broken)
2008-10-27 17:51:24 +00:00
rt = sc->sc_currates;
o With the addition of HT rates the set of h/w codes has a much wider range making the use of sc_hwmap to do direct mapping impractical. Switch to indexing by the rate index instead of the rate code and adjust associated state and logic appropriately. This has several benefits including simplification of the led code. o fix radiotap capture of HT rates o fix conditional compilation of HT radiotap support to be based on the hal having 5416 support; not the ABI version as hal builds may or may not include 5416 support
2008-10-27 18:22:44 +00:00
sc->sc_stats.ast_tx_rate =
rt->info[sc->sc_txrix].dot11Rate &~ IEEE80211_RATE_BASIC;
Enable setting the MCS rate bit for ast_tx_rate. This allows ath_stats to print the MCS rate when TX'ing.
2011-03-22 22:59:09 +00:00
if (rt->info[sc->sc_txrix].phy & IEEE80211_T_HT)
sc->sc_stats.ast_tx_rate |= IEEE80211_RATE_MCS;
Update with last year of work.
2004-12-08 17:34:36 +00:00
return copyout(&sc->sc_stats,
Unify all the wifi *_ioctl routines - Limit grabbing the lock to SIOCSIFFLAGS. - Move ieee80211_start_all() to SIOCSIFFLAGS. - Remove SIOCSIFMEDIA as it is not useful. - Limit ether_ioctl to only SIOCGIFADDR. SIOCSIFADDR and SIOCSIFMTU have no affect as there is no input/output path in the vap parent. The vap code will handle the reinit of the mac address changes. - Split off ndis_ioctl_80211 as it was getting too different to wired devices. This fixes a copyout while locked and a lock recursion. Reviewed by: sam
2008-05-01 04:55:00 +00:00
ifr->ifr_data, sizeof (sc->sc_stats));
add SIOCZATHSTATS ioctl to zero driver statistics
2009-02-13 05:38:03 +00:00
case SIOCZATHSTATS:
error = priv_check(curthread, PRIV_DRIVER);
if (error == 0)
memset(&sc->sc_stats, 0, sizeof(sc->sc_stats));
break;
o add opt_ath.h enable tweaking various config parameters for the driver without modifying the source code o default debug msgs and diag support to off MFC after: 3 days
2006-04-03 18:14:02 +00:00
#ifdef ATH_DIAGAPI
Update with last year of work.
2004-12-08 17:34:36 +00:00
case SIOCGATHDIAG:
error = ath_ioctl_diag(sc, (struct ath_diag *) ifr);
break;
o add opt_ath.h enable tweaking various config parameters for the driver without modifying the source code o default debug msgs and diag support to off MFC after: 3 days
2006-04-03 18:14:02 +00:00
#endif
Unify all the wifi *_ioctl routines - Limit grabbing the lock to SIOCSIFFLAGS. - Move ieee80211_start_all() to SIOCSIFFLAGS. - Remove SIOCSIFMEDIA as it is not useful. - Limit ether_ioctl to only SIOCGIFADDR. SIOCSIFADDR and SIOCSIFMTU have no affect as there is no input/output path in the vap parent. The vap code will handle the reinit of the mac address changes. - Split off ndis_ioctl_80211 as it was getting too different to wired devices. This fixes a copyout while locked and a lock recursion. Reviewed by: sam
2008-05-01 04:55:00 +00:00
case SIOCGIFADDR:
Multi-bss (aka vap) support for 802.11 devices. Note this includes changes to all drivers and moves some device firmware loading to use firmware(9) and a separate module (e.g. ral). Also there no longer are separate wlan_scan* modules; this functionality is now bundled into the wlan module. Supported by: Hobnob and Marvell Reviewed by: many Obtained from: Atheros (some bits)
2008-04-20 20:35:46 +00:00
error = ether_ioctl(ifp, cmd, data);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
break;
Unify all the wifi *_ioctl routines - Limit grabbing the lock to SIOCSIFFLAGS. - Move ieee80211_start_all() to SIOCSIFFLAGS. - Remove SIOCSIFMEDIA as it is not useful. - Limit ether_ioctl to only SIOCGIFADDR. SIOCSIFADDR and SIOCSIFMTU have no affect as there is no input/output path in the vap parent. The vap code will handle the reinit of the mac address changes. - Split off ndis_ioctl_80211 as it was getting too different to wired devices. This fixes a copyout while locked and a lock recursion. Reviewed by: sam
2008-05-01 04:55:00 +00:00
default:
error = EINVAL;
break;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
Update with last year of work.
2004-12-08 17:34:36 +00:00
return error;
correct some typos Submitted by: Tai-hwa Liang
2004-12-31 20:35:05 +00:00
#undef IS_RUNNING
Update with last year of work.
2004-12-08 17:34:36 +00:00
}
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
Update with last year of work.
2004-12-08 17:34:36 +00:00
/*
* Announce various information on device/driver attach.
*/
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
static void
Update with last year of work.
2004-12-08 17:34:36 +00:00
ath_announce(struct ath_softc *sc)
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
{
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Update with last year of work.
2004-12-08 17:34:36 +00:00
struct ath_hal *ah = sc->sc_ah;
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
print mac+rf part names; drop the printing 2ghz rf stuff (might come back)
2009-02-23 23:41:12 +00:00
if_printf(ifp, "AR%s mac %d.%d RF%s phy %d.%d\n",
ath_hal_mac_name(ah), ah->ah_macVersion, ah->ah_macRev,
ath_hal_rf_name(ah), ah->ah_phyRev >> 4, ah->ah_phyRev & 0xf);
Update with last year of work.
2004-12-08 17:34:36 +00:00
if (bootverbose) {
int i;
for (i = 0; i <= WME_AC_VO; i++) {
struct ath_txq *txq = sc->sc_ac2q[i];
if_printf(ifp, "Use hw queue %u for %s traffic\n",
txq->axq_qnum, ieee80211_wme_acnames[i]);
}
if_printf(ifp, "Use hw queue %u for CAB traffic\n",
sc->sc_cabq->axq_qnum);
if_printf(ifp, "Use hw queue %u for beacons\n", sc->sc_bhalq);
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
allow the size of tx+rx buffer pools to be tuned MFC after: 2 weeks
2006-02-09 21:03:25 +00:00
if (ath_rxbuf != ATH_RXBUF)
if_printf(ifp, "using %u rx buffers\n", ath_rxbuf);
if (ath_txbuf != ATH_TXBUF)
if_printf(ifp, "using %u tx buffers\n", ath_txbuf);
Add multicast key search support. This fixes corrupted mcast packets when we have more than one hostap vap. Submitted by: Russell Yount <russell.yount at gmail.com> MFC after: 2 weeks
2010-02-08 20:23:20 +00:00
if (sc->sc_mcastkey && bootverbose)
if_printf(ifp, "using multicast key search\n");
Atheros 802.11 driver. Requires Atheros Hardware Access Lay (HAL). Supported by: Atheros Comunications
2003-06-23 17:01:19 +00:00
}
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#ifdef IEEE80211_SUPPORT_TDMA
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
static __inline uint32_t
ath_hal_getnexttbtt(struct ath_hal *ah)
{
#define AR_TIMER0 0x8028
return OS_REG_READ(ah, AR_TIMER0);
}
static __inline void
ath_hal_adjusttsf(struct ath_hal *ah, int32_t tsfdelta)
{
/* XXX handle wrap/overflow */
OS_REG_WRITE(ah, AR_TSF_L32, OS_REG_READ(ah, AR_TSF_L32) + tsfdelta);
}
static void
ath_tdma_settimers(struct ath_softc *sc, u_int32_t nexttbtt, u_int32_t bintval)
{
struct ath_hal *ah = sc->sc_ah;
HAL_BEACON_TIMERS bt;
bt.bt_intval = bintval | HAL_BEACON_ENA;
bt.bt_nexttbtt = nexttbtt;
bt.bt_nextdba = (nexttbtt<<3) - sc->sc_tdmadbaprep;
bt.bt_nextswba = (nexttbtt<<3) - sc->sc_tdmaswbaprep;
bt.bt_nextatim = nexttbtt+1;
ath_hal_beaconsettimers(ah, &bt);
}
/*
* Calculate the beacon interval. This is periodic in the
* superframe for the bss. We assume each station is configured
* identically wrt transmit rate so the guard time we calculate
* above will be the same on all stations. Note we need to
* factor in the xmit time because the hardware will schedule
* a frame for transmit if the start of the frame is within
* the burst time. When we get hardware that properly kills
* frames in the PCU we can reduce/eliminate the guard time.
*
* Roundup to 1024 is so we have 1 TU buffer in the guard time
* to deal with the granularity of the nexttbtt timer. 11n MAC's
* with 1us timer granularity should allow us to reduce/eliminate
* this.
*/
static void
ath_tdma_bintvalsetup(struct ath_softc *sc,
const struct ieee80211_tdma_state *tdma)
{
/* copy from vap state (XXX check all vaps have same value?) */
sc->sc_tdmaslotlen = tdma->tdma_slotlen;
sc->sc_tdmabintval = roundup((sc->sc_tdmaslotlen+sc->sc_tdmaguard) *
tdma->tdma_slotcnt, 1024);
sc->sc_tdmabintval >>= 10; /* TSF -> TU */
if (sc->sc_tdmabintval & 1)
sc->sc_tdmabintval++;
if (tdma->tdma_slot == 0) {
/*
* Only slot 0 beacons; other slots respond.
*/
sc->sc_imask |= HAL_INT_SWBA;
sc->sc_tdmaswba = 0; /* beacon immediately */
} else {
/* XXX all vaps must be slot 0 or slot !0 */
sc->sc_imask &= ~HAL_INT_SWBA;
}
}
/*
* Max 802.11 overhead. This assumes no 4-address frames and
* the encapsulation done by ieee80211_encap (llc). We also
* include potential crypto overhead.
*/
#define IEEE80211_MAXOVERHEAD \
(sizeof(struct ieee80211_qosframe) \
+ sizeof(struct llc) \
+ IEEE80211_ADDR_LEN \
+ IEEE80211_WEP_IVLEN \
+ IEEE80211_WEP_KIDLEN \
+ IEEE80211_WEP_CRCLEN \
+ IEEE80211_WEP_MICLEN \
+ IEEE80211_CRC_LEN)
/*
* Setup initially for tdma operation. Start the beacon
* timers and enable SWBA if we are slot 0. Otherwise
* we wait for slot 0 to arrive so we can sync up before
* starting to transmit.
*/
static void
ath_tdma_config(struct ath_softc *sc, struct ieee80211vap *vap)
{
struct ath_hal *ah = sc->sc_ah;
struct ifnet *ifp = sc->sc_ifp;
struct ieee80211com *ic = ifp->if_l2com;
const struct ieee80211_txparam *tp;
const struct ieee80211_tdma_state *tdma = NULL;
int rix;
if (vap == NULL) {
vap = TAILQ_FIRST(&ic->ic_vaps); /* XXX */
if (vap == NULL) {
if_printf(ifp, "%s: no vaps?\n", __func__);
return;
}
}
tp = vap->iv_bss->ni_txparms;
/*
* Calculate the guard time for each slot. This is the
* time to send a maximal-size frame according to the
* fixed/lowest transmit rate. Note that the interface
* mtu does not include the 802.11 overhead so we must
* tack that on (ath_hal_computetxtime includes the
* preamble and plcp in it's calculation).
*/
tdma = vap->iv_tdma;
if (tp->ucastrate != IEEE80211_FIXED_RATE_NONE)
optimize ath_tx_findrix: there's no need to walk the rates table as sc_rixmap is an inverse map NB: could eliminate the check for an invalid rate by filling in 0 for invalid entries but the rate control modules use it to identify bogus rates so leave it for now
2009-05-07 00:35:32 +00:00
rix = ath_tx_findrix(sc, tp->ucastrate);
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
else
optimize ath_tx_findrix: there's no need to walk the rates table as sc_rixmap is an inverse map NB: could eliminate the check for an invalid rate by filling in 0 for invalid entries but the rate control modules use it to identify bogus rates so leave it for now
2009-05-07 00:35:32 +00:00
rix = ath_tx_findrix(sc, tp->mcastrate);
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
/* XXX short preamble assumed */
sc->sc_tdmaguard = ath_hal_computetxtime(ah, sc->sc_currates,
ifp->if_mtu + IEEE80211_MAXOVERHEAD, rix, AH_TRUE);
ath_hal_intrset(ah, 0);
ath_beaconq_config(sc); /* setup h/w beacon q */
add a sysctl to ena/dis frobbing cca
2009-03-05 00:15:43 +00:00
if (sc->sc_setcca)
ath_hal_setcca(ah, AH_FALSE); /* disable CCA */
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
ath_tdma_bintvalsetup(sc, tdma); /* calculate beacon interval */
ath_tdma_settimers(sc, sc->sc_tdmabintval,
sc->sc_tdmabintval | HAL_BEACON_RESET_TSF);
sc->sc_syncbeacon = 0;
sc->sc_avgtsfdeltap = TDMA_DUMMY_MARKER;
sc->sc_avgtsfdeltam = TDMA_DUMMY_MARKER;
ath_hal_intrset(ah, sc->sc_imask);
DPRINTF(sc, ATH_DEBUG_TDMA, "%s: slot %u len %uus cnt %u "
"bsched %u guard %uus bintval %u TU dba prep %u\n", __func__,
tdma->tdma_slot, tdma->tdma_slotlen, tdma->tdma_slotcnt,
tdma->tdma_bintval, sc->sc_tdmaguard, sc->sc_tdmabintval,
sc->sc_tdmadbaprep);
}
/*
* Update tdma operation. Called from the 802.11 layer
* when a beacon is received from the TDMA station operating
* in the slot immediately preceding us in the bss. Use
* the rx timestamp for the beacon frame to update our
* beacon timers so we follow their schedule. Note that
* by using the rx timestamp we implicitly include the
* propagation delay in our schedule.
*/
static void
ath_tdma_update(struct ieee80211_node *ni,
Minor cleanups of tdma protocol handling: o break out version-related code to simplify rev'ing the protocol o add parameter validation macros so checks that appear multiple places are consistent (and easy to change) o add protocol version check when looking for a scan candidate o improve scan debug output format o rewrite beacon update handling to calculate a bitmask of changed values and pass that down through the driver callback so drivers can optimize work o do slot bounds check before use when parsing received beacons
2009-03-18 19:28:17 +00:00
const struct ieee80211_tdma_param *tdma, int changed)
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
{
#define TSF_TO_TU(_h,_l) \
((((u_int32_t)(_h)) << 22) | (((u_int32_t)(_l)) >> 10))
#define TU_TO_TSF(_tu) (((u_int64_t)(_tu)) << 10)
struct ieee80211vap *vap = ni->ni_vap;
struct ieee80211com *ic = ni->ni_ic;
struct ath_softc *sc = ic->ic_ifp->if_softc;
struct ath_hal *ah = sc->sc_ah;
const HAL_RATE_TABLE *rt = sc->sc_currates;
u_int64_t tsf, rstamp, nextslot;
u_int32_t txtime, nextslottu, timer0;
int32_t tudelta, tsfdelta;
const struct ath_rx_status *rs;
int rix;
sc->sc_stats.ast_tdma_update++;
/*
* Check for and adopt configuration changes.
*/
Minor cleanups of tdma protocol handling: o break out version-related code to simplify rev'ing the protocol o add parameter validation macros so checks that appear multiple places are consistent (and easy to change) o add protocol version check when looking for a scan candidate o improve scan debug output format o rewrite beacon update handling to calculate a bitmask of changed values and pass that down through the driver callback so drivers can optimize work o do slot bounds check before use when parsing received beacons
2009-03-18 19:28:17 +00:00
if (changed != 0) {
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
const struct ieee80211_tdma_state *ts = vap->iv_tdma;
ath_tdma_bintvalsetup(sc, ts);
o fix dynamic slave-side tdma slot length updating: we need to re-setup the burst length in the tx q's o remove re-config of the beaconq on update; it's not needed
2009-04-13 20:57:34 +00:00
if (changed & TDMA_UPDATE_SLOTLEN)
ath_wme_update(ic);
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
DPRINTF(sc, ATH_DEBUG_TDMA,
"%s: adopt slot %u slotcnt %u slotlen %u us "
"bintval %u TU\n", __func__,
ts->tdma_slot, ts->tdma_slotcnt, ts->tdma_slotlen,
sc->sc_tdmabintval);
/* XXX right? */
ath_hal_intrset(ah, sc->sc_imask);
/* NB: beacon timers programmed below */
}
/* extend rx timestamp to 64 bits */
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
rs = sc->sc_lastrs;
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
tsf = ath_hal_gettsf64(ah);
Overhaul monitor mode handling: o replace DLT_IEEE802_11 support in net80211 with DLT_IEEE802_11_RADIO and remove explicit bpf support from wireless drivers; drivers now use ieee80211_radiotap_attach to setup shared data structures that hold the radiotap header for each packet tx/rx o remove rx timestamp from the rx path; it was used only by the tdma support for debugging and was mostly useless due to it being 32-bits and mostly unavailable o track DLT_IEEE80211_RADIO bpf attachments and maintain per-vap and per-com state when there are active taps o track the number of monitor mode vaps o use bpf tap and monitor mode vap state to decide when to collect radiotap state and dispatch frames; drivers no longer explicitly directly check bpf state or use bpf calls to tap frames o handle radiotap state updates on channel change in net80211; drivers should not do this (unless they bypass net80211 which is almost always a mistake) o update various drivers to be more consistent/correct in handling radiotap o update ral to include TSF in radiotap'd frames o add promisc mode callback to wi Reviewed by: cbzimmer, rpaulo, thompsa
2009-05-20 20:00:40 +00:00
rstamp = ath_extend_tsf(rs->rs_tstamp, tsf);
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
/*
* The rx timestamp is set by the hardware on completing
* reception (at the point where the rx descriptor is DMA'd
* to the host). To find the start of our next slot we
* must adjust this time by the time required to send
* the packet just received.
*/
rix = rt->rateCodeToIndex[rs->rs_rate];
txtime = ath_hal_computetxtime(ah, rt, rs->rs_datalen, rix,
rt->info[rix].shortPreamble);
/* NB: << 9 is to cvt to TU and /2 */
nextslot = (rstamp - txtime) + (sc->sc_tdmabintval << 9);
nextslottu = TSF_TO_TU(nextslot>>32, nextslot) & HAL_BEACON_PERIOD;
/*
* TIMER0 is the h/w's idea of NextTBTT (in TU's). Convert
* to usecs and calculate the difference between what the
* other station thinks and what we have programmed. This
* lets us figure how to adjust our timers to match. The
* adjustments are done by pulling the TSF forward and possibly
* rewriting the beacon timers.
*/
timer0 = ath_hal_getnexttbtt(ah);
tsfdelta = (int32_t)((nextslot % TU_TO_TSF(HAL_BEACON_PERIOD+1)) - TU_TO_TSF(timer0));
DPRINTF(sc, ATH_DEBUG_TDMA_TIMER,
"tsfdelta %d avg +%d/-%d\n", tsfdelta,
TDMA_AVG(sc->sc_avgtsfdeltap), TDMA_AVG(sc->sc_avgtsfdeltam));
if (tsfdelta < 0) {
TDMA_SAMPLE(sc->sc_avgtsfdeltap, 0);
TDMA_SAMPLE(sc->sc_avgtsfdeltam, -tsfdelta);
tsfdelta = -tsfdelta % 1024;
nextslottu++;
} else if (tsfdelta > 0) {
TDMA_SAMPLE(sc->sc_avgtsfdeltap, tsfdelta);
TDMA_SAMPLE(sc->sc_avgtsfdeltam, 0);
tsfdelta = 1024 - (tsfdelta % 1024);
nextslottu++;
} else {
TDMA_SAMPLE(sc->sc_avgtsfdeltap, 0);
TDMA_SAMPLE(sc->sc_avgtsfdeltam, 0);
}
tudelta = nextslottu - timer0;
/*
* Copy sender's timetstamp into tdma ie so they can
* calculate roundtrip time. We submit a beacon frame
* below after any timer adjustment. The frame goes out
* at the next TBTT so the sender can calculate the
* roundtrip by inspecting the tdma ie in our beacon frame.
*
* NB: This tstamp is subtlely preserved when
* IEEE80211_BEACON_TDMA is marked (e.g. when the
* slot position changes) because ieee80211_add_tdma
* skips over the data.
*/
memcpy(ATH_VAP(vap)->av_boff.bo_tdma +
__offsetof(struct ieee80211_tdma_param, tdma_tstamp),
&ni->ni_tstamp.data, 8);
#if 0
DPRINTF(sc, ATH_DEBUG_TDMA_TIMER,
"tsf %llu nextslot %llu (%d, %d) nextslottu %u timer0 %u (%d)\n",
(unsigned long long) tsf, (unsigned long long) nextslot,
(int)(nextslot - tsf), tsfdelta,
nextslottu, timer0, tudelta);
#endif
/*
* Adjust the beacon timers only when pulling them forward
* or when going back by less than the beacon interval.
* Negative jumps larger than the beacon interval seem to
* cause the timers to stop and generally cause instability.
* This basically filters out jumps due to missed beacons.
*/
if (tudelta != 0 && (tudelta > 0 || -tudelta < sc->sc_tdmabintval)) {
ath_tdma_settimers(sc, nextslottu, sc->sc_tdmabintval);
sc->sc_stats.ast_tdma_timers++;
}
if (tsfdelta > 0) {
ath_hal_adjusttsf(ah, tsfdelta);
sc->sc_stats.ast_tdma_tsf++;
}
ath_tdma_beacon_send(sc, vap); /* prepare response */
#undef TU_TO_TSF
#undef TSF_TO_TU
}
/*
* Transmit a beacon frame at SWBA. Dynamic updates
* to the frame contents are done as needed.
*/
static void
ath_tdma_beacon_send(struct ath_softc *sc, struct ieee80211vap *vap)
{
struct ath_hal *ah = sc->sc_ah;
struct ath_buf *bf;
int otherant;
/*
* Check if the previous beacon has gone out. If
* not don't try to post another, skip this period
* and wait for the next. Missed beacons indicate
* a problem and should not occur. If we miss too
* many consecutive beacons reset the device.
*/
if (ath_hal_numtxpending(ah, sc->sc_bhalq) != 0) {
sc->sc_bmisscount++;
DPRINTF(sc, ATH_DEBUG_BEACON,
"%s: missed %u consecutive beacons\n",
__func__, sc->sc_bmisscount);
add hw.ath.bstuck to control the stuck beacon threshold
2009-02-10 19:26:42 +00:00
if (sc->sc_bmisscount >= ath_bstuck_threshold)
TDMA support for long distance point-to-point links using ath devices: o add net80211 support for a tdma vap that is built on top of the existing adhoc-demo support o add tdma scheduling of frame transmission to the ath driver; it's conceivable other devices might be capable of this too in which case they can make use of the 802.11 protocol additions etc. o add minor bits to user tools that need to know: ifconfig to setup and configure, new statistics in athstats, and new debug mask bits While the architecture can support >2 slots in a TDMA BSS the current design is intended (and tested) for only 2 slots. Sponsored by: Intel
2009-01-08 17:12:47 +00:00
taskqueue_enqueue(sc->sc_tq, &sc->sc_bstucktask);
return;
}
if (sc->sc_bmisscount != 0) {
DPRINTF(sc, ATH_DEBUG_BEACON,
"%s: resume beacon xmit after %u misses\n",
__func__, sc->sc_bmisscount);
sc->sc_bmisscount = 0;
}
/*
* Check recent per-antenna transmit statistics and flip
* the default antenna if noticeably more frames went out
* on the non-default antenna.
* XXX assumes 2 anntenae
*/
if (!sc->sc_diversity) {
otherant = sc->sc_defant & 1 ? 2 : 1;
if (sc->sc_ant_tx[otherant] > sc->sc_ant_tx[sc->sc_defant] + 2)
ath_setdefantenna(sc, otherant);
sc->sc_ant_tx[1] = sc->sc_ant_tx[2] = 0;
}
bf = ath_beacon_generate(sc, vap);
if (bf != NULL) {
/*
* Stop any current dma and put the new frame on the queue.
* This should never fail since we check above that no frames
* are still pending on the queue.
*/
if (!ath_hal_stoptxdma(ah, sc->sc_bhalq)) {
DPRINTF(sc, ATH_DEBUG_ANY,
"%s: beacon queue %u did not stop?\n",
__func__, sc->sc_bhalq);
/* NB: the HAL still stops DMA, so proceed */
}
ath_hal_puttxbuf(ah, sc->sc_bhalq, bf->bf_daddr);
ath_hal_txstart(ah, sc->sc_bhalq);
sc->sc_stats.ast_be_xmit++; /* XXX per-vap? */
/*
* Record local TSF for our last send for use
* in arbitrating slot collisions.
*/
vap->iv_bss->ni_tstamp.tsf = ath_hal_gettsf64(ah);
}
}
Remove ATH_SUPPORT_TDMA and use IEEE80211_SUPPORT_TDMA instead. It doesn't make much sense to configure driver support w/o net80211. Note this means ath now depends on opt_wlan.h.
2009-03-30 19:23:49 +00:00
#endif /* IEEE80211_SUPPORT_TDMA */
Export ath stats via snmp, rather than requiring a debugging interface and "athstats".
2010-08-14 14:18:02 +00:00
Break out the ath PCI logic into a separate device/module. Introduce the AHB glue for Atheros embedded systems. Right now it's hard-coded for the AR9130 chip whose support isn't yet in this HAL; it'll be added in a subsequent commit. Kernel configuration files now need both 'ath' and 'ath_pci' devices; both modules need to be loaded for the ath device to work.
2011-03-31 08:07:13 +00:00
MODULE_VERSION(if_ath, 1);
MODULE_DEPEND(if_ath, wlan, 1, 1, 1); /* 802.11 media layer */
Reference in New Issue Copy Permalink