changes

RTL8187 for linux驱动 提供内核支持802.11g的模块

Added more severe check for the RF chip (trivial)
Compatibility fixes for 2.4 kernels (From Mihnea Stoenescu)
Added module parm to choose if use HW for sequnce number (see README file)
updated README file
added other workarounds for compiling issues - see the firsts lines in r8180_core.c (Thanks to 'rooter2' who reported errors)..
some initial code for managed mode & to make the nic tx beacons (not working yet!!)
added experimental calibration of AGC by keeping high AGC_RESET on radio
fixed bug while reading eeprom  value for energy trheshold BB reg (was masked with 3.. i'm not sure about this)
Added initial BUGGY support for statistics in /proc/net/dev 
Hooked in SW wep support
fixed a bug that caused system crash while reiceving packet larger than a descriptor by less than 4 bytes
copyed ieee802_11.h from 2.6 kernel tree (because it has been removed in > 2.6.6)

V 0.7
Added PCI id for DLINK DWL510 (From Markus Gothe)
Some code clean-ups and compatibility fixes (From Markus Gothe)
If-name defaulted to "wlan0" and make settable via module param (From Markus Gothe)
Makefile clean improved (From Markus Gothe)
Initial disassociation code
some improvement to the hw init routine
some code for future WEP support (not working yet)
some improvements for baseband and sa2400 configuration
updated README file 

V 0.6
Fixed TX to work with paket larger than a descriptor.
Added WX for fragmentation.Now we can set fragmentation treshold
in iwconfig you should see the bssid of the associated AP (if any)
fix in RX (we called kmalloc with GFP_KERNEL in irq context)
When the TX descriptors becomes full, the driver buffers the remaining TX packet and stop the net_if queue, then it wait the nic to complete a transfer and TX the buffered packet, the wake the queue again.
I've copied a 560Mb file using my RTL8180 nic.. 8-)


V 0.5
Fix in the Tx routine (about fragment and descriptors)
Increased number of TX descriptors and size of the TX buffers
We take care of correctly unload the ieee80211 stack on rmmod, we also destroy eventually scheduled works
Added Wireless Extension for getname
some code clean-ups
Fixed scannig routines (didn't release lock)
We switched to WLAN header sw seq number handling: hw seemeed to give problems
RX routines should handle packet bigger than a descriptor now
Changes in the TX fragmentation code
Fixed ieee80211 header size in the ieee80211 stack


V 0.4
Fix len in the association request frame builder routine
Fix in the tx routines, now we should have the right seq number in the txed packet
Other association logics implemented
Basic association should work now
README file updated
Update ieee80211 stack. Merge of my changes whith the new ipw2100 v0.51 stack
Added some basic IEEE802.11 header construction code to the IEEE802.11 stack
Modified TX routine to get packet already fragmented by the IEEE802.11 stack
Initial network support..
Initial ping! On my system the NIC associates and is able to ping another host! ;-)


V 0.3
Improved scan routine: we have channel, rate, mode, etc..
Improved scan routine: now runnig iwlist scan returns immediately, results are collected continuosly
Improved makefile: you have to run make, mk script does not longer exist
Some initial association logics. You may ask the NIC to associate by running iwconfig eth1 essid XXXXX. This is not functional yet, however your AP should at least answer with a failed association message
Various fixes in rx/tx configuration code
Some code cleanups
added handler for network ifdown
Fix in the TxPower EEPROM read routine
Fix in the TX routine. Now PLCP duration should be OK
TX should work correctly now
fix in the set_channel wx handler (check for valid ch number)


v 0.2
Added some wireless extension handler and set nettype, this make kismet work
Fix in the RX routine and in the FIFO fixing routine (set buffer lenght every time we got a frame/fix fifo)
We remove CRC at the end of the packet
We make the RF sleep before remove the module
We can switch between monitor and managed mode. In managed mode we have basic/passive net scanning
You may choose if accept bad-crc frames in monitor mode via iwpriv


V 0.1 - First Sourceforge available. - it still should be considered a development release
Added inital support for wireless extension. you may change channel with iwconfig
Code cleanup
Added README file
Some additional check prevents the driver to load for nic that are know to don't work
Fix in RX descriptor fixing code
Removed some debug code


V 0.0.9
We use uniderectional DMA single mapping for rx buffers (not for descriptors) instead of consistent memory
Some PCI posting fixes
Some code celan-ups
Sa2400 RF code rewritten
Fix in the RX code - We shouldn't get oopses anymore!!


V 0.0.8
Attempt to fix hardlock on module remove
Some code clean ups
Fix in the buffer dump routine - if you have beacons in the air, you will se the essid in the dump!
Initial merge with the 802.11 stack from the Intel Pro Wireless 2100 project
Monitor Mode working on channel 11 - with ethereal you should see the raw ieee802.11 packets !! (we still have oopses!!!!!!)
Better irq handler
Specify the max DMA burst size for rx
some code restructurations


V 0.0.7
Code clean ups - **now you have to insmod r8180.ko instead of r8180_pci_init.ko **
Support for reading the card eeprom (tested on 93c46, but should work also on 93c56)
Init routine now sets analog parameter reading it from the eprom - you should really see TX ok confirmations now
Initial, experimental, support for PHILIPS sa2400 radio. (note that most probably this works only on MY card)
Added basic RX logics and buffer handling
Fixes to the RX descriptor allocator
Initial *very* rough monitor mode via dmesg on ch11 (most probably frames are corrupted and data is messy. Under relatively heavy load there are oopses and some amenity)


V 0.0.6
We have TX descriptor and TX buffers pointer handling funcions and logics
We have a rough initial TX funcion. this lack of certains things (mostly we set a dummy tx required time)
Bringing down and up the interface make the drv invoke TX dma. The nic successfully TX the (dummy) frame (altought the RF part is not configured, so the frame isn't TXed in the air)
We still don't see anything looping back..


V 0.0.5
Request irq earlyer
reworked EEPROM config code
changes in TX descriptor allocator (can allocate for all the 3 queues)
we set correctly (at least better) the eprom mode before setting registers
We use explicit asking to the nic to poll the descriptor queues
DMA transfer now (try to) start!!! You can see TX and RX failure int from the NIC!


V 0.0.4
Write to CONFIG reg needs enabling write 
We force the media link status to managed mode link
Major fix in IOMAP write routines
For now we always returns IRQ_HANDLED in the irq handler for safety reasons
we allocate the tx descriptor ring (I forgot to do that in 0.0.3)
we now write addresses for rx and tx rings in the right location
fixed tx conf
we dump the nic register map (page 0)


V 0.0.3
Fixed interrupt handler: now we reset properly irq condition
some fix in rx descriptor allocation code
contains some initial code for rough normal priority frame tx 
set nic to MAC loopback mode


V 0.0.2
Added PCI ID for Belkin 8180 wlan card
Fixed module probe fail
Read & write MEM/IO routines fixed
Read MAC correctly
reset the NIC on card init
added dummy irq handler
contains some initial code for rough frame rx


V 0.0.1
Inital test release.
Recognizes the card and read MAC nearly-correctly