
*** For driver installation documentation, please see main README file ***

Available cards with Alantro/TI ACX100 chip (seems to also be called
TNETW1100):
- mini-PCI adapter GL2422VP PCI card (made by Global Sun Technology Inc.)
  Distributed by Alloy, Eusso
  See also http://www.loneguard.de/wlan/gl2422vp/ (WARNING, German! :)
- mini-PCI -> PCI Level One WNC-0200
- PCMCIA Level One WPC-0200
  PCMCIA cards *probably* have a MMCX connector (although it looks a bit
  strange...) inside the case, like tons of other PCMCIA cards are said
  to have it...

The successor of the ACX100 chip is TNETW1100B (pin-for-pin
compatible with ACX100, much lower power consumption)

Further chips are TNETW1130 (ACX111) and TNETW1230. A new chip for
mobile phones is TNETW1250 (which will probably be used together with
TI's TNETW3422M radio). We think that we will be able to support all of
those chips, also the 1250.
The TNETW1130 frequently uses the 0x16 radio Radia RC2321H / RC2422.
Hmm, however the USR(80)5410 uses:
TI TNETW1130GVF
   42Z3468 D4

RADIA RC2422B
      HK0407
      A5D0

RADIA RC2326
      SG0405
      410L2CK

FIXME: what is the TNETW2522 and RC2432?
The RC2432/RC2436 are said to support 802.11a/b/g

Transfer quality can be influenced by the following parameters:
- txpower setting via iwconfig [implemented, 16.5 and 18dB]
- antenna selection [not implemented yet; ctlDot11CurrentAntenna()]
- regulatory domain settings (number of channels allowed for use within
  certain regions, not implemented yet; ctlDot11CurrentRegDomainWrite())
- sensitivity setting
- channel agility, to avoid In-Band interference from household appliances, such as microwave ovens and cordless phones.
- "short retry limit" and "long retry limit" [currently hardcoded in
  driver, could be exported; see 802.11 spec 9.2.5.3 for details]
- ED (energy detect) threshold. It detects whether the channel is
  unusable due to existing energy on the channel (microwave, ...).
  Setting it to a higher value might still allow for communication
  (ctlDot11EDThresholdWrite(), ctlDot11CurrentCCAModeWrite())
- ???

Other tunable parameters (to be implemented via iwpriv):
- dot11CFPPeriod
- dot11CFPMaxDuration
- dot11BeaconPeriod (see ctlDot11BeaconPeriodWrite())
- max transmit MSDU lifetime (ctlDot11MaxTransmitMSDULifetimeWrite())

According to Luie Matthee, the ACX100 PCI card (DWL-520+?) might have a
requirement for PCI 2.2 compliant boards, thus PCI 2.1 may not work.
Problems with IRQ assignment, spontaneous rebooting when "problematic"
cards from other vendors are being used etc. and people having a hard time
to get it running at all are pretty well-known, so I suspect it is true.


The TI binary drivers have been developed by the following people:

* William F. Alexander (Windows NDIS driver for PCI, USB):
  http://www.geocities.com/SiliconValley/Haven/4824/Resume.txt
* Larry A. Cawley (mentioned on Alexander's site)
* Lancelot Wang (Linux driver)

Tim Rikers is the Linux Technology Guy at TI. Might be worthwhile to
contact him about getting ACX1xx support from TI...

ACX100 chip specifications are being listed at eupuls.com (maybe we can
ask them to deliver them??):
http://www.eupuls.com/product/eu_prod_wireless11.html
----
 Eupuls S/W deliverables
    for Access Points Drivers Development

Drivers development Kit

 Programmers Guide
   - ACX100 Register I/F Guide
   - Host Interface Guide
   - Driver Specification AP
   - Firmware Behavior
 Examples Driver Code (C Code)
 Example PCI Loader
----

The Maxim radio in most cases probably is the MAX2820 (at least it's
been confirmed for DWL-120+).
Candidates might be the MAX2820, MAX2821 or MAX2822...
Some cards *might* also contain the MAX2644 2.4GHz amplifier perhaps eventually.
However, it's said to be a "one-chip" radio, so that probably isn't the case...

But OTOH the 2820 doesn't have enough output power, so an amplifier is
needed. This could be the MAX2242, which has 22.5dBm linear power output
(+2.5dBm is to compensate loss due to T/R switch and bandpass filter).
As such, I guess we can assume that a Maxim radio based card has a
maximum power output of 20dBm.

The RFMD radio might be the RF2948B.

DWL-650+ cards seem to be exceptionally unreliable:
I've heard of at least 5 cards with severe or slight hardware problems
(PC lockup on card insert or firmware upload issues due to RAM problems).
One card which deadlocked the PC (i.e.: final and fatal card damage) was:
H/W Ver.: B4
F/W Ver.: 1.9.3.101
FCC ID: KA2DWL-650PLUSB1
Another deadlocked card:
H/W Ver.: B3
F/W Ver.: 1.9 (can't find any more subversion numbers)
FCC ID: KA2DWL-650PLUSB1

Not even a single card other than the DWL-650+ is known to me for causing
hardware trouble...
