changes

RTL8187 for linux驱动 提供内核支持802.11g的模块

V 0.21
Make sure the ssid beacon field is long enought(Thanks to 'Deuce' and Markus Waldeck)
Fixed auto load of WEP modules (Thanks to Lamarque Eric)
2.4 compatibility fixes (Thanks to Lamarque Eric)
BASIC support for AP mode
use the AP wx to make possible change the IBSS cell mac(use ifconfig hw ether for master mode)
Fixed oops when remmod and workqueue was running
Fuxed oops on old kernels when loading with no devices present (Thansk to Lamarque Eric)
Avoid to respond to probe request until the nic is correctly settled in ad-hoc or master mode (fix some HW error when creating IBSS and possibly BSS)
New created IBSS cell have randomized MAC address (hope this is ok..)
Tx fragmentation fix for HW seq number
Static allocation for SSID (Thanks to 'Deuce')
Statistics in proc fs reworked
Support for optional prism-header in monitor mode (Thanks to Christian)
Addedd module-author and module-description tags (Thanks to Lamarque Eric)
Some code clean-ups
Added WX for setting retry limit
Fixed wrong declarations for some inline functions (Thanks to Benjamin Collins)
Fixed iw_handler_def for compatiblity with newer WX (Thanks to Phil Jones)
Setting essid - only in managed mode you can set null essid
Ensure locking in WX
Fixed a typo: reiceve -> receive (Thanks to 'Deuce')
Support for active scan (Thanks to 'Deuce')
Make the WEP module complains if ARC4 not loaded (Thanks to 'Thenihilist')
increased PHY settle time, hopefully helps who seen errors on PHY write
Fixed swapped arguments in TX buffer allocation (Thanks to Aryix who pointed out the original patch for the ieee 802.11 stack in -mm kernel patch)
Improved association logic behaviour with varius user-set-essid, user-set-ap-bssid, network is broadcasting essid etc..

V 0.20.2
Did make clean

V 0.20.1
Deleted (again) an outdated README.gct

V 0.20
Experimental support for GCT radio
RSSI fixes. Tested on Maxim - Philips still broken - (Thanks to Deuce)
Some code for master (AP) mode included. Not finished yet
Reassociation improvement (Thanks to Peter Fales for some reports)
Support for AP that send deauth frame instead of disassoc frame (Thanks to 'ad209' for reporting the problem and to Andrea Fava who lend me his AP)
When association procedure fails then the nic restart to scan and eventually retryes.. This is also a workaround for some case in witch the reassociation fail (don't ask me why..)
Fixed problem with DHCP: bringing down the iface now keep the essid (Thanks do Deuce to have pointed it out and to have worked on it)
Fixed memory leak in scan routines
Support for the wx for setting ap. Should work for hidden essid network. 
Association code reworked
Fixed bug in probe response capability-WEP
Make use of IOMMAP instead of PIO by default - performance improved.
Fixed a bug in encoding the RX-ed rate for passing to the stack - Thanks to 'Christian' for pointing this out
Changed the number of RX and TX descriptors to an half to reduce DMAable consistent memory requirement
Fixed mgmt frames have reasonable sequence number even if we are using SW seq number handling
Fixed module_unload24 script (Thanks to Ivano Primi)

V 0.19
Killed some nonsense warning for maxim radio
Fixed maxim radio doesn't work with debug disabled (Thanks to Deuce)
Added initial hook for wireless statistics (Thanks to Csredeagle )
Fixed basic rate at 1Mbps (was unitialized in some circumnstances)

V 0.18
Fix: MAXIM radio fixes
Fix: Module probe doesn't fail with MAXIM radio (Thanks to Deuce) 
Fix: MAXIM / PHILIPS write_phy_antenna passed wrong argument
Fix: Wep module autoload (Thanks to Deuce)
Fix: MAXIM radio now should work (Thanks to Deuce)
Fix: Changed some locking mechanism
Added module load/unload scripts for 2.4 kernel

V 0.17
Fix: Makefile for 2.4 kernel
Fix: Makefile26
Fix: Ad-hoc mode channel logic improvement
Fix: Driver should tolerate hw hotplug without complain too much
Implemented memory free routines for buffers/descriptors
SA2400 cleanups
SA2400 digital PHY experimental fix (PLEASE REPORT!!!!!!!!!)
Should recognize OK all PHY type (included MAXIM and GCT (previously recognized as RFMD?))
Old EEPROM default values fix
UNTESTED initial support for MAXIM radio
Fixed another bastard bug in TX descriptor handling

V 0.16
Merged in a lot of cleanups (Thanks to Mariusz Matuszek)
Added module parameter for channel scan bitmap (thanks to Mariusz Matuszek)
Code for start/stop scan reworked
Fix in sa2400 code (firdac- experimental, if you have firdac card please report)
Fixes in the probe fail resource release
Beacon TX support via HW beacon queue
Initial ad-hoc support - Should basically work, if you treat it with care -
Addedd a dummy WX for a not supported wx.
Changed ieee stack name, so on the same PC I can have an rtl8180 and an ipw2100 running at the same time ;-P
initial attempt to have duration field correct in TXed frame (code not active yet)
Rx ISR is invoked with RX nic int masked.
Added kernel tx_timeout handler and we update last_trans in dev struct
Fix in beacon parsing
Addedd lot of stats
Addedd register image in proc fs
Fix: insmodding with hwseqnum=1 should work now
Reworked TX queue control flow
Reworked TX descriptor handling
Fix: compatibility with >2.6.9 kernel, we use pci_dma_sync_single_for_cpu/ _for_device. (Thanks to 'dignome')
Fix: RX max DMA burst was wrong in register definition
Fix: TX config take care of some more params
Fix: makefile for 2.4 kernel (Thanks to Kamil Zidek)

V 0.15
We collects a lot of statistics
Basic rtx stats returned by ifconfig should be ok.
We don't want CTRL frames if we are not in monitor mode
Fixed memory leak in TX
Addedd /proc/net/rtl8180/<ifacename> proc enrty with stats
Set the NIC to perform early rx as little as possible (This causes less RX pointer lost and missing last descriptor)

V 0.14
Some cleanups and PM skeleton (Thanks to Mariusz Matuszek)
RX workaround to avoid RX lockup 

V 0.13
Addedd full register definition with Realtek's name (Thanks to Mariusz Matuszek)
fixed 5.5Mbps (Thanks to Mihnea Stoenescu)
fixed memory leak in RX isr in some case when packet reception isn't fully completed
fixed memory leak in RX 802.11 stack when handling management frames
Added support for changing mac address
fixed sleeping in irq while resetting the card due to an int error (experimental:don't reset at all)
Added workaround for RX fifo reset (from Mihnea's rtl8180-emb driver)
We call the rx process with enabled interrupt (experimental) 

V 0.12
Improvement for monitor mode
Explicit support for promiscuos mode in managed mode (implemented by workaround, might fail)
Fixed mad numbers in iwconfig
Support for changing rate via iwconfig (5.5Mbps is broken!)
Compile fixes,cleanups and cosmetics (Thanks to Trisk)
Lot of code cleanup and Updated README (Mariusz Matuszek)
Better NIC restart/commit logics
Empty the TX descriptors when reset/commit (instead of when init)

V 0.11
Now we reset/reconfigure card when some config param changed
Fixed switching from monitor to managed
Fixed crc-error-accept monitor ipwriv
Fix (I hope) for who had association error code 0xa (Thanks to Harry Glinos for reporting this)

V 0.10
Addedd PCI id for Belkin F5D6020 (Thanks to Rick Bronson)
Some experimental improvement in sa2400 programming
Some major bugfix in sa2400 programming: Digital and Analog PHY was wrongly detected
Fixed Analog PHY calibration procedure
Added check/diagnostic for AP authentication failure
Added check/diagnostic for AP association failure
Added INSTALL file (Thanks to Rick Bronson)
Fixed sleeping in softirq context during scan

V 0.9.1
Fix for hardlock when bringing up the interface (Thanks to Rick Bronson)

V 0.9
Updated README file
Fixed spin_(lock/unlock)_irq(save/restore) typo (called without &) - (Thanks to Bjoern Riemer)
Initial support for reassociation if association lost 
Experimental skb change to see fix problem for who see hard reboot
2.4 kernel compatibility fix (thanks to Rafael Darder Calvo)
Fix in TX: now HI priority (and hopefully low priority) queue is working
Management frames are sent on high priority queue

V 0.8.1
REALLY copyed ieee802_11.h from 2.6 kernel tree (because it has been removed in > 2.6.6)

V 0.8
Fix in the reset routine (From David Young)
Minor code clean-ups