3d7a066a27
oops caused by not setting up network device queues. Change wlan_setup to use the kernel provided alloc_netdev() rather than hand coding it. This should help avoid any future problems. Also change register_wlandev to not call dev_alloc_name() as register_netdev() will do that if the device name has been setup. Signed-off-by: Richard Kennedy <richard@rsk.demon.co.uk>
1269 lines
66 KiB
Plaintext
1269 lines
66 KiB
Plaintext
* CHANGES
|
|
*
|
|
* Copyright (C) 1999 AbsoluteValue Systems, Inc. All Rights Reserved.
|
|
* --------------------------------------------------------------------
|
|
*
|
|
* linux-wlan
|
|
*
|
|
* The contents of this file are subject to the Mozilla Public
|
|
* License Version 1.1 (the "License"); you may not use this file
|
|
* except in compliance with the License. You may obtain a copy of
|
|
* the License at http://www.mozilla.org/MPL/
|
|
*
|
|
* Software distributed under the License is distributed on an "AS
|
|
* IS" basis, WITHOUT WARRANTY OF ANY KIND, either express or
|
|
* implied. See the License for the specific language governing
|
|
* rights and limitations under the License.
|
|
*
|
|
* Alternatively, the contents of this file may be used under the
|
|
* terms of the GNU Public License version 2 (the "GPL"), in which
|
|
* case the provisions of the GPL are applicable instead of the
|
|
* above. If you wish to allow the use of your version of this file
|
|
* only under the terms of the GPL and not to allow others to use
|
|
* your version of this file under the MPL, indicate your decision
|
|
* by deleting the provisions above and replace them with the notice
|
|
* and other provisions required by the GPL. If you do not delete
|
|
* the provisions above, a recipient may use your version of this
|
|
* file under either the MPL or the GPL.
|
|
*
|
|
* --------------------------------------------------------------------
|
|
*
|
|
* Inquiries regarding the linux-wlan Open Source project can be
|
|
* made directly to:
|
|
*
|
|
* AbsoluteValue Systems Inc.
|
|
* info@linux-wlan.com
|
|
* http://www.linux-wlan.com
|
|
*
|
|
* --------------------------------------------------------------------
|
|
*
|
|
* Portions of the development of this software were funded by
|
|
* Intersil Corporation as part of PRISM(R) chipset product development.
|
|
*
|
|
* --------------------------------------------------------------------
|
|
- Fix oops due to 2.6.27 API change (Richard Kennedy)
|
|
- Fix compiles with 2.6.26 (Pavel Roskin)
|
|
- Add compatibility with 2.6.27+ WEXT API (Pavel Roskin)
|
|
- Fix misplaced variable in PCMCIA code (Pavel Roskin)
|
|
- Add compatibility for gfp_t and mutex APIs (Pavel Roskin)
|
|
- Fix non-posix find arguments in script (Mark Mathews)
|
|
0.2.9
|
|
- Reduce stack usage in prism2mib_priv (Richard Kennedy)
|
|
- WEXT functions should return -EBUSY when the hardware is still
|
|
being initialized (Richard Kennedy)
|
|
- Use round_jiffies wherever possible (Richard Kennedy)
|
|
- Disallow ioctls from running until the hardware probe function has
|
|
finished. (Richard Kennedy)
|
|
- Fix userspace scripts to work on Fedora 8 (Chris Rankin)
|
|
- Perform the hardware reset/init sequence *before* we register
|
|
the netdevice. (Karl Renton)
|
|
- Always do an ifstate_disable before a firmware load cycle.
|
|
- Default to wireless extension mode.
|
|
- Make the SIWENCODE behaivor more closely match the spec. (Tormod Volden)
|
|
- Apply a patch from Karl Relton that improves USB reliability.
|
|
- When in wlan_wext_write=1 mode, auto-transition to ifstate_enable
|
|
and bypass userspace firmware load altogether.
|
|
- Fix bug in the GIWENCODE handler (Richard Kennedy & Tormod Volden)
|
|
- Support 2.6.24 kernels (modified patch from Tormod Volden)
|
|
- Support 2.6.23 kernels. (With apologies to Martin Dauskardt for a
|
|
misapplied patch)
|
|
- Support "Intersil Americas USB 802.11b WLAN DEVICE" (Peter Levart)
|
|
- Tweaks to support Fedora 7 (Dan Williams)
|
|
- Support 2.6.22 kernels
|
|
0.2.8
|
|
- Fix up a pile of sparse warnings (Pavel Roskin)
|
|
- Support 2.6.20 kernels (Pavel Roskin)
|
|
- Tweaks to the udev rules (Richard Kennedy)
|
|
- Silence a large pile of warnings with GCC4.1
|
|
- pt_regs is gone for USB as well on 2.6.19+ (Chris Rankin)
|
|
- Large WEXT patch that allows use of NetworkManager. (Richard Kennedy)
|
|
0.2.7
|
|
- Fix license text in prism2dl.c (it is in fact MPL/GPL)
|
|
- Fix oops in wext_autojoin (Richard Kennedy)
|
|
- Report use of dBm for WEXT requests (Richard Kennedy)
|
|
- Configure script better handles PREEMPT kernels (Chris Rankin)
|
|
- 2.6.19+ uses a different IRQ handler function prototype.
|
|
- linux/config.h is no more on 2.6.19+ kernels.
|
|
0.2.6
|
|
- Fix another error that affected <2.6.15 builds on some compilers
|
|
- Fix a cut-n-paste error that broke 2.6.18 builds.
|
|
0.2.5
|
|
- 2.6.18-rc fixes.
|
|
- A few tweaks for Wireless-Extensions-21
|
|
- Fix an interrupt-related race that could lead to a crash on shutdown.
|
|
- Tweaks to the prism2_usb code to improve suspendability on 2.6.15+
|
|
- Silence many warnings under GCC 4.1
|
|
- prism2_cs support for 2.6.16+ (untested)
|
|
- Makefile fixes for 2.6.17+
|
|
- udev support for USB targets (Richard Kennedy)
|
|
- semaphores are deprecated as of 2.6.16 (Chris Rankin)
|
|
- prism2_usb build fix for 2.6.16
|
|
- Patches from Debian (via Victor Seva)
|
|
- Fix bugs in /etc/wlan/shared, including a security hole
|
|
- Fix build on 2.4.27 kernel
|
|
- Added man page for prism2dl
|
|
- Generate a WEXT Event when we gain/lose association.
|
|
- Make prism2_defer query ssid and bssid, as IBSS mode has this change.
|
|
- Clean up prism2_defer_timer a bit.
|
|
- Propogate linkstatus into p80211, and make ethtool aware of it.
|
|
- Eliminate wland and entire indication infrastructure. If you want to
|
|
play with it, use 0.2.4. We can put it back in if necessary.
|
|
0.2.4
|
|
- USB Build fix from Chris Rankin.
|
|
- Add ability to install prism2 firmware seperately. (Petteri Raty)
|
|
- Theoretically the per-frame signal/noise figures are now in dBm.
|
|
0.2.3
|
|
- WEXT 16+ fixes, which came to a head with v2.6.14+ kernels.
|
|
- Fix a circular symbol dependency glitch
|
|
- Fix a potential panic when receiving unknown CONTROL frames.
|
|
- get rid of p80211frm.c entirely; it's all unused code.
|
|
- Fix prototype for pci_suspend functions to match 2.6.11+ APIs.
|
|
- Build fixes for 2.6.14+
|
|
- Change the type of 'request_pending' to take care of warnings.
|
|
- Fixes for warnings generated by the 'sparse' tool (Richard Kennedy)
|
|
- Add an explicit README.firmware containing copyright information
|
|
on the firmware images.
|
|
- Compile fixes for 2.6.8 (Victor Seva Lopez)
|
|
0.2.2
|
|
- Fix an array overflow in the hotplug helper function.
|
|
0.2.1-final
|
|
- Added in-kernel devicetable for 2.6.13+ pcmcia
|
|
- Updates to bring prism2_cs into the newer 2.6 kernel APIs.
|
|
- Fixes in the hotplug invocation with WEXT (Pavel Roskin)
|
|
- Experimentally try using usb_reset_device on 2.4 kernels too
|
|
- use usb_reset_device() on 2.6 kernels (Andreas Schulte)
|
|
- Included kludged 1.8.4 PCI/PCMCIA firmware (Pavel Roskin)
|
|
- PCMCIA card Ident fixes to coexist with orinioco driver (Pavel Roskin)
|
|
- Ignore unexpected ctlx responses. (Chris Rankin)
|
|
- Fix an off-by-one in scan results.
|
|
- Change some log priorities
|
|
- [usb] Fix a memory leak (Chris Rankin * Eroc Koenders)
|
|
- [usb] clean up the PDA Read function a bit. (Chris Rankin)
|
|
- [usb] Don't overwrite current CTLX URB before we're done with it.
|
|
(Eric Koenders & Chris Rankin)
|
|
- Fix a byte-order problem when parsing scan results (Eroc Koenders)
|
|
- Added a mib item (lnxRSNAIE) to get/set the WPA/RSNA IE.
|
|
- Applied a patch from Chris Rankin which should solve the hangs with
|
|
kernel preemption turned on.
|
|
- Cleaned up some compile warnings with GCC 3.4.x
|
|
- Add a background thread to fetch comms quality from hardware.
|
|
- Fixed a series of bad memsets in prism2dl
|
|
- Added id for Viewsonic USB widget (Ekin Meroglu)
|
|
- Eliminate the extra frame copy in the usb rx path.
|
|
- When issuing a start/join, set MACMODE to none as to disable
|
|
transmits through the hardware.
|
|
- Get rid of the prism2_pci/plx 2.2.x compatibility code (Pavel Roskin)
|
|
- Change default behaivor of top-level Makefile (Pavel Roskin)
|
|
- Get rid of the #ifndef DECLARE_TASKLET cases; 2.4+ supports it.
|
|
- Fetch the BSSID when we get an AP_CHANGE event
|
|
- netif_carrier_off/netif_carrier_on at appropriate times
|
|
- Hook up the dot11CurrentTxPowerLevel MIB in the prism2 code.
|
|
- Add a default TMPDIR in case it isn't set.
|
|
- Removed some 2.2.x cruft
|
|
- CommsQuality now does the right thing with dBm values.
|
|
- More makefile work from Pavel Roskin
|
|
- A major USB patch from Chris Rankin. This rewrites the
|
|
command-queueing code to something considerably more sane.
|
|
- A few more Makefile tweaks
|
|
- Further compile warning fixes on 64-bit targets (Pavel Roskin)
|
|
- Fix prism2_cs on 64-bit targets (Pavel Roskin)
|
|
- Fix the 2.4 build (whoops)
|
|
-pre26
|
|
- Fix the "no hardware found" bug in 2.6.10+ kernels
|
|
- Cleanups in the Makefiles and build system
|
|
- Fix a typo in the shared.prism2 script (Chris Rankin)
|
|
- Clean up build warnings with gcc 3.3 (and possibly older too)
|
|
- Fix a nice race condition on device registration.
|
|
- Don't stop the netdev when we hit ifstate_disable.
|
|
- hw->state wasn't being set properly on drvr_stop, nor were we
|
|
clearing the port status in all cases.
|
|
- Fix the "Debug: sleeping function called from invalid context"
|
|
under 2.6 with prism2_usb (Chris Rankin)
|
|
- Wireless extension SCAN support (Dan Williams)
|
|
- 2.6 sysfs support in the form of SET_NETDEV_DEV (Dan Willimas)
|
|
-pre25
|
|
- Fix a brown-paper bag bug in the register_wlandev code. (Giacomo Lozito)
|
|
- Patch for the Sitecom WL-022 widget (Armijn Hemel)
|
|
- Build fix for ancient ETHTOOL versions
|
|
- build fix for old 2.4.x kernels
|
|
-pre24
|
|
- Fix build on 2.6.10-rc
|
|
- Fix USB build on 2.6.x (x < 7)
|
|
- Suspend the USB driver cleanly (Colin Leroy)
|
|
- Fix more USB badness (Colin Leroy)
|
|
- scan_timeout WAY too long. (Wolfram Gloger)
|
|
- dev_get is no more, as of 2.6.10-rc2 -- switch to dev_alloc_name
|
|
instead, which is present in older kernels. (reported by Colin Leroy)
|
|
-pre23
|
|
- Fix badness in 2.6.10-rc1 (Colin Leroy)
|
|
- Compile fix for 2.6.9-rc4+ (Jeff Chua)
|
|
-pre22
|
|
- Add support for x86-64 (and possibly other 64-bit targets) (J.D. Cole)
|
|
- Fix linkstate if driver reset without unloading. (Shiro Ninomiya)
|
|
- Rename the man pages to have a .1 extension
|
|
- Tweak the Configure script; change the default kernel build directory
|
|
- put a 'sleep 1' in the firmware load sequence to improve reliability.
|
|
(Arnold Liu)
|
|
- Add Linksys WUSB11 card ident
|
|
-pre21
|
|
- Add a per frame rx structure for OOB data to be passed to p80211
|
|
This is a step in the right direction for a major overhaul of
|
|
internal frame representation. The prism2 driver shouldn't care
|
|
about sniff headers or any of that crap; it should always pass
|
|
everything to p80211 (via this metadata) and the p80211conv_*
|
|
functions should do that work. It's also needed to eventually
|
|
support different encryption types.
|
|
- Added in beginnings of IWSPY support via patch from Josef Kriegl
|
|
- Install modules to $MOD_DIR/linux-wlan-ng instead of $MOD_DIR/kernel
|
|
- Primary/Secondary firmware (finally) bundled with the driver.
|
|
- Convert to newer 2.6 module parameter code.
|
|
- Fix an inverted test in GIWENCODE wireless extension (Shiro Ninomiya)
|
|
- Minor script fix for SSF cards (Pavel Roskin)
|
|
- Card idents.
|
|
- Fixes for SSF PCI cards; default to x16 SRAM as that's more common.
|
|
(download script now automatically detects hardware type)
|
|
- Fix script problems that result in a failed init returning success.
|
|
- Fix an endian-ness buglet in prism2sta.c (Federico Pellegrin)
|
|
- Compile fixes for 2.2.x kernels. (no tasklets or spin_lock_bh)
|
|
- Fix up pci resources on card init failure on plx card.
|
|
-pre20
|
|
- Fix up pci resources on card init failure
|
|
- Ident for "ROPEX FreeLan 802.11b USB Adapter" (Francesco Bochicchio)
|
|
- ETHTOOL support now optional, fixes build on older kernels.
|
|
- Fix prism2dl to not use kernel headers to fix 2.6.3 build (Pavel Roskin)
|
|
-pre19
|
|
- Add a 'lnxind_roam' indication.
|
|
- minor typo in the nwepgen makefile (Pavel Roskin)
|
|
- More makefile fixes (Pavel Roskin)
|
|
-pre18a
|
|
- brown-paper-bag fix for the 2.4 build. And it no longer builds on
|
|
2.2 thanks to an accidental makefile commit.
|
|
-pre18
|
|
- Fix broken makefiles so that prism2dl and the rc script install
|
|
- Get rid of the txfid lock, it's not necessary.
|
|
- Tasklet-driven RX handlers.
|
|
- Further work towards a Wireless Extension-enabled future.
|
|
- SIWFRAG, SIWRETRY, SIWFREQ, SIWRTS implemented
|
|
- SIWENCODE, SIWESSID taken from zdsta-1.0.3 (needs work)
|
|
- SIWCOMMIT Implemented ala autojoin.
|
|
- write support defaults to disabled, override via module parameter
|
|
- fix a warning in the wireless extension code.
|
|
- hfa384x.h header updates
|
|
- Fix prism2dl to handle the S7 record being elsewhere in the image.
|
|
(Pavel Roskin)
|
|
- Fix up brown paper bag script error for SSF hardware.
|
|
-pre17
|
|
- Spruce up the GIWNAME wireless extension ioctl. (Natsakis Konstantinos)
|
|
- Added in basic ethtool support; for link status reflection.
|
|
- Fixed a double-locked spinlock which deadlocked on SMP boxes.
|
|
- /proc/net/p80211/wlanX/nsd now has useful information about the
|
|
prism2 hardware/firmware revisions.
|
|
- eliminate the unused 'collptr' from the mib structures.
|
|
- Fix a bug with the pcmcia script and scanning (Dan Noe)
|
|
- Fix a crash on a VCC mismatch.
|
|
- Split out most pci/plx/pcmcia/usb specific code into their own .c files.
|
|
- Make prism2_cs build on 2.6.2-rc
|
|
- Changes to support SSF hardware.
|
|
- Card ident for D-Link DWL-650P1
|
|
- Move the prism2 link notification handler operate in scheduler
|
|
context to avoid a hang in softirq context.
|
|
- Melco WLI-USB-KB11 USB widget added (Bela Fenyvessy)
|
|
- Add in x86-64 support, plus other cleanups in wlan_compat.h
|
|
- Support the Netgear MA311 PCI card. (Steve Berry)
|
|
-pre16
|
|
- Implement wireless extension SETCHANNEL. (SIOCSIWFREQ)
|
|
- Currently broken. :)
|
|
- Fix several sniffing-related bugs:
|
|
- disable when we're already disabled
|
|
- no longer perform a full h/w reset when we're already enabled
|
|
- Vastly reduce CPU usage on channel hop.
|
|
- Fix a crash on unload for pci/plx widgets (Pavel Roskin)
|
|
-pre15
|
|
- Minor fix for the shared script (Jeff Chua)
|
|
- fix "bad: schedule while atomic" bug introduced by recent 2.6.0-test
|
|
kernels and wireless extensions.
|
|
- Switch to using spin_[un]lock_bh for the cmdlock; this way we
|
|
disable the transmit bh from running while issuing a card command.
|
|
Also have the txframe function obtain the lock.
|
|
- Hawking HighDB USB Widget (Allan Claghorn)
|
|
- Fix a double-lock in hfa384x_cmd_notify
|
|
- Identity for "T-Sinus 111 USB WlAN Adapter" (Roman Koutny)
|
|
- Fix up alignment problems on frame rx. (Ranjit Deshpande)
|
|
-pre14
|
|
- Update the HAS_HOTPLUG test to something considerably more foolproof
|
|
- Add ids for the Airvast Prism3 USB widget
|
|
- bap locking was broken when illegal parameters were passed. (Andre McCurdy)
|
|
- Don't use jiffies for timeouts when interrupts are disabled. (Andre McCurdy)
|
|
- Make the Configure script a little saner.
|
|
- Remove the Zydas 1201 USB widget. Use the Zydas modified
|
|
linux-wlan-ng release instead (we hope to merge it in eventually)
|
|
-pre13
|
|
- Fix two big-endian bugs in the scan code.
|
|
- Hotplug agent executes '/sbin/ifup INTERFACE' if possible.
|
|
- Planex GW-US11H USB adaptor (Henry Ip)
|
|
- Averatec USB Wlan Adapter (Stephan Miller)
|
|
- D-Link DWL-122 USB widget (Andrew Beresford)
|
|
- ASUS-WL140 Wireless USB widget
|
|
- Fix the WLAN_LOG_NOTICE macro for gcc 2.95.x
|
|
-pre12
|
|
- Fix up endian bugs in lnxerq_commsquality.
|
|
- Get rid of the pcmcia release timer for >2.6.0-test5
|
|
- Card ident for 'ZyDAS ZD1201 Wireless USB Adapter' (Albert Pauw)
|
|
- More tweaks for 2.6.0-test4
|
|
- fix some bash-isms in /etc/wlan/shared (Daniel Song)
|
|
- Collapsed the BOUNDEDINT and INT types into one.
|
|
- Eliminated the WLAN_LOG_***0() macros at long last thanks to CPP
|
|
trickery. Also eliminated more cruft from wlan_compat.h
|
|
- Card ident for the Samsung SWL-2210P PCI card
|
|
- Try not to schedule a USB endpoint reset if there's one already
|
|
pending. (Chris Rankin)
|
|
- Fix pci/plx operation on architechures with >32bit physical address
|
|
sizes. (namely mips64 & some mips32)
|
|
- get rid of save_flags/cli/restore_flags in prism2_cs on >2.2 builds
|
|
(tweaked from patch by Pavel Roskin)
|
|
- pci_device_id tables should not be marked __devinitdata
|
|
-pre11
|
|
- support genesis mode for RAM download of primary fimware
|
|
- prism2_usb now builds on 2.5/2.6.
|
|
- endpoint numeration removed.
|
|
- hfa384x_usb_corereset unimplemented for 2.5/2.6
|
|
- hfa384x.h updates.
|
|
- p80211conv_* functions should work with netdev->mtu, not MAX_ETHFRM.
|
|
- Simply erase the bogus '3rdparty' drivers on Mandrake systems.
|
|
- Card idents for Adaptec AWN-8030/AWN-8020, and Netgear MA111
|
|
- Strip out all 2.2.x stuff from prism2_usb, it doesn't work.
|
|
- add proper __devinit/__devexit/etc tags to prism2sta.c
|
|
- Minor USB janitorial patch from Chris Rankin.
|
|
- Make most card commands require superuser access (exception is
|
|
mibget) (Original patch by Pavel Roskin)
|
|
-pre10
|
|
- Make the wakeup flag in the usbctlx structure volatile, as well as
|
|
other wait condition variables.
|
|
- Make the rx_urb buffer dynamically allocated so we can re-post the
|
|
URB as soon as possible.
|
|
- Get rid of completion.h
|
|
- More wlan_compat updates.
|
|
- Disable TXEX reporting on USB as well as TXOK. should improve
|
|
stability a tad.
|
|
- Further 2.5 USB work. Not quite there yet.
|
|
-pre9
|
|
- Properly initialize the prism2_cs release timer (Pavel Roskin)
|
|
- Eliminate the CONFIG_ISA requirement for prism2_cs (Pavel Roskin)
|
|
- netlink build fixes for 2.5.71
|
|
- Hopefully fix the 2.2.x build.
|
|
- More USB fixes (Chris Rankin)
|
|
- Compile fixes for < 2.4.20
|
|
-pre8
|
|
- Fix the pcmcia breakage introduced in -pre2.
|
|
- On USB widgets, only signal txcomplete if there was a tx error.
|
|
- Minor cleanups in the PCMCIA code.
|
|
-pre7
|
|
- Add in the 'JVC MP-XP7250" USB Widget.
|
|
- Don't call ev_alloc on tx_complete event; only the tx urb completion.
|
|
- Minor tweaks to tx_timeout code
|
|
- More USB updates; primarily with error-handling. (Chris Rankin)
|
|
-pre6
|
|
- On USB transmits, don't return 'success' if the tx_urb is still
|
|
pending. Also don't wake the netdev queue if the tx fails.
|
|
- Hook up the tx watchdog timer and add a p80211-level handler.
|
|
- Fix compilation with older (<= 2.4.18) kernels.
|
|
- Another massive USB+other patch from Chris Rankin:
|
|
- a separate list for "finished" CTLX objects.
|
|
- correct (hopefully!) handling of -EPIPE errors, shamelessly copied
|
|
from^W^Winspired by David Brownell's usbnet driver.
|
|
- finished support for asynchronous commands, with proper tracking
|
|
through the CTLX lifecycle, clean-up etc. Each asynchronous command now
|
|
completes in its own tasklet.
|
|
- CTLX cancellation and error handling.
|
|
- further work on the timers and spin locks.
|
|
- clean up setconfig16() and setconfig32() functions so that they
|
|
take appropriate value parameters, even if the underlying
|
|
setconfig() commands don't.
|
|
- use an asynchronous command in setmulticast(). (CS, PCI and PLX
|
|
are still synchronous; they just pretend not to be.)
|
|
- further clean up in the disconnect() handler. However, it's still
|
|
not a good idea to remove the adapter while the
|
|
prism2sta_inf_hostscanresults() function is sleeping, and the wlandevice
|
|
struct could easily have been freed before the p80211req_dorequest()
|
|
function manages to clear the request_pending bit. It would be
|
|
nice if we could cancel both of these operations when we call
|
|
prism2sta_ifstate(P80211ENUM_ifstate_disable).
|
|
- extra support for Linux 2.5 (task queue -> work queue).
|
|
- reset the link status when we stop the driver.
|
|
- Some work on module locking. Not there yet, but getting better.
|
|
- Remove magic numbers from USB endpoint probe (Chris Rankin)
|
|
- Eliminate useless timer in p80211do_request, as all calls are
|
|
synchronous. Also clena up more function pointers, and a few other
|
|
doodads. (Modified from patch by Chris Rankin)
|
|
- Track the current SSID in the wlandev; Make the WEXT stuff key from
|
|
this. Update the SSID when we get a linkstatus_connect notification.
|
|
- A couple of misc cleanups from Chris Wilson
|
|
- Cleanups in the 2.5 makefile code. (Pavel Roskin)
|
|
- Add in another PCI device ID, to handle Sony VAIO PCG-V505AP
|
|
- Make sure CONFIG_ISA is enabled for the prism2_cs build. (x86 only)
|
|
-pre5
|
|
- Get rid of generic manfids in /etc/pcmcia/wlan.conf (Pavel Roskin)
|
|
- Fix the 'make install' target.
|
|
- only call flush_scheduled_tasks() if DECLARE_TASKLET not defined.
|
|
(fixes 2.5.x compile)
|
|
-pre4
|
|
- More script work; mostly cleanups and other enhancements.
|
|
- Substantial prism2_usb patch from Chris Rankin.
|
|
- using 2 Linux lists for CTLXs
|
|
- remove race condition and double-free with CTLXs in disconnect()
|
|
- initialise URBs correctly (vital for Linux 2.5+)
|
|
- separate callbacks for data/CTLX OUT URBs
|
|
- flush task queue on shutdown
|
|
- more thorough use of spinlocks with CTLXs
|
|
- set the ASYNC_UNLINK flag when it is time to unlink an URB
|
|
asynchronously, and not before.
|
|
- ensure we stop submitting URBs once the disconnect function
|
|
is called.
|
|
- More USB idents (Jeff Chua)
|
|
- Scan now uses active scans only; should speed things up a bit.
|
|
- Proper module init/cleanup in prism2sta.c, plus more work for 2.5.69
|
|
- Added in descriptions for module parameters.
|
|
- Major overhaul of the hotplug stuff. Now we use our own hotplug
|
|
event class, and use that to kick off configuration and whatnot.
|
|
All device types (pci/plx/pcmcia/usb) now use this mechanism.
|
|
Also brings us things like unified power management (suspend/resume)
|
|
- Supress spurious output when setting WEP parameters.
|
|
- Only query PrivacyOptionImplemented if wep was enabled in the config
|
|
Works around an apparent firmware bug, see http://bugs.debian.org/190851
|
|
- Eliminate hw->name.
|
|
- Card ident for NL-2511CD Plus pcmcia card.
|
|
- Minor tweak on pcmcia removal.
|
|
- Fix the USB compile on <2.4.20
|
|
-pre3
|
|
- get rid of wlandev_get_index and whatnot.
|
|
- URB cleanup in the USB code. (Chris Rankin)
|
|
- Moved host auth sequence out of interrupt context.
|
|
- USB paranoia checks (Chris Rankin)
|
|
- Partial rewrite of the USB shutdown sequence
|
|
- eliminate wlandev->hwremovedfn().
|
|
- Minor tweaks to the mib.
|
|
- Get rid of the horridly outdated skeleton driver.
|
|
-pre2
|
|
- Fix a crash-on-unload affecting pci and plx adapters
|
|
- have the init script load the module if needed.
|
|
- Handle the new 2.5.69+ irq handler semantics
|
|
- Fix 2.5 with pcmcia (Pavel Roskin)
|
|
- Massive cleanup of the prism2/driver/Makefile (Pavel Roskin)
|
|
- Don't wrap the [un]register_netdevice() calls in rtnl_[un]lock()
|
|
And use the [un]register_netdev() calls instead.
|
|
- Call the hwremoved() functions on pci/plx_remove
|
|
- Eliminate the prism2sta_priv_t structure.
|
|
- Fix a memory leak on ifstate_disable
|
|
- Minor changes to the USB driver
|
|
- Widget id for "Melco WLI-USB-KS11G" (Ichiro Doi)
|
|
- Function pointer syntax cleanup (Chris Rankin)
|
|
- Push the multicast/promisc set into the prism2 driver.
|
|
- Move type/multicast filtering into the p80211 layer.
|
|
- Format fix for p802addr_to_str (Pavel Roskin)
|
|
- Add a check for non-Intersil firmware.
|
|
- Fix a infinite recursion in the usb lowlevel code. (Chris Rankin)
|
|
-pre1
|
|
- Header updates.
|
|
- Add a card ident for the USR 1120 USB widget (Juan Conde)
|
|
- Cleaner shutdown; fixed a possible leak and a race condition.
|
|
- Another rearrangement of the command locks, to eliminate use of the
|
|
spin_is_locked() call.
|
|
- Fix a rare deadlock with cmd_initialize (Pavel Roskin)
|
|
- Add a Lucent-compatible key-generator in add-ons/lwepgen
|
|
- Fix a stack overflow problem. (Pavel Roskin)
|
|
- Further monitor mode fix for control frames (Pavel Roskin)
|
|
- Cleanups in the dependency generation code
|
|
- Initial 2.5 support. USB currently broken.
|
|
- Minor build system cleanups.
|
|
0.2.0
|
|
- Use PSUEDOIBSS mode for monitor mode; this should eliminate spurious
|
|
beacon generation. (Ryan Veety)
|
|
- Don't drop "unrecognized" frame types in monitor mode.
|
|
- Added ident for the Linksys WUSB12 (Brad Davidson)
|
|
- Eliminated most of the WLAN_LOG_???n macros.
|
|
- Minor fixups in /etc/wlan/shared
|
|
- Cleaned up some of the error messages in the conversion code.
|
|
- changed the "Host de-WEP failed" error to a debug message, as the
|
|
/proc/net/wireless stats are updated to reflect undecryptable frames
|
|
-pre10
|
|
- Added the p2TxPowerMax MIB item to specify the TxPower level.
|
|
0-30 dBm. Needs STA>=1.7.0 or AP>=1.4.0.
|
|
- Clean up some spurious warnings in prism2sta.c (Pavel Roskin)
|
|
- A series of patches from Pavel Kankovsky, somewhat tweaked. :)
|
|
- Properly set skb->mac.raw in non-monitor mode
|
|
- Enhancments to the p80211 frame conversion code
|
|
- Handle A4 frames.
|
|
- Don't issue linkstatus notifications in monitor mode
|
|
- Supress Linkstatus messages in monitor mode
|
|
- hfa384x.h updates for latest firmware.
|
|
-pre9
|
|
- ZyXEL ZyAir B200 Wireless USB widget ID added (Paul Lacatus)
|
|
- Only enable interrupts in one place (hfa384x_drvr_start)
|
|
- Fix the startup scripts to disable WEP completely if not enabled.
|
|
- ALLNET 0193 USB widget ID (Kurt Huwig)
|
|
- Disable the hfa384x port before issuing the autojoin stuff.
|
|
- Fixed a subtle flaw in the 802.11->802.3 conversion code.
|
|
- Updates to hfa384x.h
|
|
- Add productinfo for dlink DRC650 from Petr Slansky.
|
|
- Minor changes to the rfmon code.
|
|
- Only notify us on linkstatus changes if it actually changes.
|
|
- Fixed the path the manpages were installed to.
|
|
-pre8
|
|
- Properly reset the hardware state after a flash/ram download.
|
|
- Patch to allow escaped characters in the scan list (Derek Atkins)
|
|
- Re-added the txfid queue locking which somehow got dropped.
|
|
- Added DellTrueMobile 1180 USB ident. (Michael Hackett)
|
|
- Changed linkstatus messages to INFO level reporting.
|
|
- Ident info for the Intel Anypoint II 802.11b PCMCIA card.
|
|
- Added support for the SH architechure and many changes necessary for
|
|
2.5 support. Modified from the patches by Ryan Veety.
|
|
- Compile fixes for 2.2.x introduced as part of the new hotplug support.
|
|
- Added man pages contributed by Joey Hess
|
|
- Add ident info for the Z-Com 725/726 USB Widgets.
|
|
- Fixes to silence GCC 3.2 warnings. (Pavel Roskin)
|
|
-pre7
|
|
- Patch to fix int array mib manipulation. (Pavel Roskin)
|
|
- Minor improvements to the build system.
|
|
- Moved RX processing to a bottom half instead of hard irq context.
|
|
- Further script improvements: signal strength filtering, and some
|
|
space handling stuff. The latter is not complete yet.
|
|
- lnxreq_commsquality wasn't setting the status field properly on the
|
|
response messages.
|
|
- Workaround for spurious interrupts generated before initialization
|
|
is complete. (Clay Jones)
|
|
- hotplug script now invokes 'ifup $DEVICE' if present.
|
|
- Minor cleanups (Pavel Roskin)
|
|
- Tenative SPARC support. (Olivier Bornet)
|
|
- After leaving monitor mode, disable the port if we had to enable it.
|
|
- Updates to hfa384x.h
|
|
-pre6
|
|
- Further script enhancements; automatically use scan mode on hardware
|
|
that is known to support it properly.
|
|
- Fixed errors in the hotplug script, and other misc fixes.
|
|
- 2.2.x fixes for the new proc stuff and deferred processing stuff.
|
|
- Add support for the new sniff frame capture. (doc/capturefrm.txt)
|
|
- Get rid of some bitrot with the sniffing code.
|
|
- Populate the tx dropped frames statistics.
|
|
-pre5
|
|
- Fixes for host scan when not currently joined.
|
|
- /proc support; some informational stuff is kept in /proc/wlandev/wlanX
|
|
- Complete rewrite of the configuration files and startup scripts.
|
|
hotplug(pci+usb), pcmcia, rc, everything uses same config file(s).
|
|
per-network configuration
|
|
supports scanning and automatically choosing a network from a set.
|
|
- Fix the TMD7160/ncp130 support so it acually works. :)
|
|
- Make the BUG() call in cmd_access enabled only in debug builds.
|
|
- Minor makefile fixes.
|
|
- Microsoft MN520 PCMCIA ident info added.
|
|
- p2cnfShortPreamble is not a boolean value. (Clay Jones)
|
|
-pre4
|
|
- BAP access fixes for PCI platforms. This mostly affects SMP
|
|
arrangements.
|
|
- Back out W200 USB ident info; it's not prism-based.
|
|
- Patch from Tom Prado to make 'keygen' more user-friendly.
|
|
- bap_timeout parameter only valid on non-usb platforms
|
|
-pre3
|
|
- Added Ident info for the USB Compaq/Intel W200 widget
|
|
- dot11req_scan requires firmware >=1.3.2.
|
|
- Added the "Acer Warplink USB Adapter" id.
|
|
- Rework the prism2sta_inf_linkstatus call to defer processing of
|
|
results until outside interrupt context. Should fix the SMP
|
|
deadlocks seen on some machines.
|
|
- Rework the usb_disconnect code to hopefully alleviate the occasional
|
|
crash-on-disconnect some people see. Solution is not SMP-safe yet.
|
|
- Add the D-Link DCF-660W ident info.
|
|
-pre2
|
|
- New parameter, 'prism2_bap_timeout' to specify the timeout on
|
|
bap setup. timeout*10 == timeout in microseconds. Note that this
|
|
timeout occurs in spinlock context, so increasing it arbitrarily is
|
|
BAD! Defaults to 1000, ie 10000us.
|
|
- compile fix for prism2sta in debug mode.
|
|
- Only prompt for pcmcia sources if we're using non-kernel pcmcia.
|
|
(original patch by Beat Bolli, rewritten a bit..)
|
|
- Add the Linksys WCF12 CF ident info.
|
|
- A bunch of compile fixes for older versions of wireless extensions.
|
|
- Added implementations of GIWTXPOWER and GIWRETRY
|
|
- Some fixes to the wireless extension code (thanks to Jean Tourrilhes)
|
|
-pre1
|
|
- wland was incorrectly using kernel headers in its compile.
|
|
- Adhoc mode was auto-roaming onto "any" SSID. Fixed.
|
|
- Implemented the dot11req_scan & dot11req_scan_results functions.
|
|
This, combined with more intelligent startup scripts, will allow
|
|
you per-network profiles. Scripts are unwritten as yet. :)
|
|
- hfa384x.h and MIB updates.
|
|
- Rewrote the wireless extension support code; moved to p80211 layer.
|
|
Now we're compatible with at least v6-v14, and everything
|
|
inbetween. We hook up to the new iw_request structure, which
|
|
makes further wireless extension support much easier to add.
|
|
- Add the Microsoft MN510 USB device ID.
|
|
- Finally fix the 'crash on unload' problem for 2.2.x kernels with PCI.
|
|
Turns out the kcompat24 code wasn't quite complete. (Matthew Rush)
|
|
- Cleaned up our usage of __FUNCTION__ to make gcc 3.x happy.
|
|
- Fix a typo in the prism2 makefile.
|
|
0.1.15
|
|
- Fix an obscure PCMCIA build problem. (modversions disabled but
|
|
header present, and using non-kernel pcmcia)
|
|
-pre8
|
|
- Added code to optionally truncate packets in monitor mode. (Clay Jones)
|
|
- Fix a polarity problem with the stripfcs argument to monitor mode.
|
|
- Add a test to make sure wireless.h is included.
|
|
- Netgear MA401A card ident.
|
|
- Further cleanups of the build system (Makefiles)
|
|
-pre7
|
|
- Include wireless.h so wireless extensions work again. :)
|
|
- Work around the braindead RedHat kernel build system.
|
|
- Fix a class of unaligned accesses in the message structures
|
|
(kernel driver) and user space parsers (wlanctl-ng, etc)
|
|
- Fix an unpacked struct problem that affected ARM platforms
|
|
(thanks to Clay Jones for spotting this one)
|
|
-pre6
|
|
- Support the new auto-unknown mode present in firmware >1.3.3
|
|
When issuing an autojoin, we try to join a BSS, then an IBSS, and if
|
|
all fails, we create a new IBSS with the given parameters.
|
|
*note* currently disabled.
|
|
- Updates to hfa384x.h to reflect latest documentation
|
|
- Build system mini-enema. Automagically detect kernel version,
|
|
pcmcia status, and modversions. Pick up the kernel compile flags.
|
|
- Hopefully fixed the 'crash on unload' for 2.2.x kernels and PCI.
|
|
- 'p2req_join' command, see doc/wlanctl-ng.p2req_join.txt
|
|
This lets you join a specific SSID. Thanks to Clay Jones.
|
|
- A couple more pcmcia/cf card idents
|
|
- More deletions/fixes in wlan_compat.h
|
|
-pre5
|
|
- Due to new hotplug code, atmel device detection removed
|
|
- Further cleanups in wlan_compat.h and non-kernel makefiles
|
|
- Debian package stuff yanked in favor of the "official" deb packages
|
|
- Rearranged the wireless extension support code a bit
|
|
- A fix for the promisc/allmulti stuff on USB devices
|
|
-pre4
|
|
- Hotplug support for PCI/PLX/USB widgets
|
|
- A compile fix for 2.2.x (x < 18) kernels and PCI adapters
|
|
- Support for TMD7160-based "almost but not quite PLX" boards
|
|
- HP-PA support. In theory. Anyone care to test?
|
|
- pcmcia script cleanups
|
|
- Add ident info for the "Intel PRO/Wireless 2011B LAN USB Device"
|
|
- Include an almost-sane fix to the modversion madness
|
|
- Tweak the init order of the PCI driver structs
|
|
-pre3
|
|
- Added ident info for ASUS WL-110/WL-100 cards
|
|
- Support for PCI power managment
|
|
- Fixed a wep hang on the transmit side
|
|
- Compile fixes; mainly for 2.2 kernels
|
|
- Add support for the SIOCGIWRATE wireless extension
|
|
-pre2
|
|
- Restructured the driver locking. SMP should be MUCH better now.
|
|
- Stripped out more dead code. (mainly FID_STACK)
|
|
- Fix an skb access after we call netif_rx
|
|
- Fixes for XScale CPUs.
|
|
- Added added a MiniUSB ident info
|
|
- Restructure of the docmd_* functions
|
|
-pre1
|
|
- Host-based WEP decryption and encryption
|
|
- Partial rewrite of copy_to_bap.
|
|
- Debian package metadata updated to 3.0-STABLE. (David Everley)
|
|
- Stripped out a lot of crap from wlan_compat.h
|
|
- More atmel adapters added to the USB exclude list.
|
|
- James Goodwin's patch to fix a bug in the 80211conv code.
|
|
- The 3COM AirConnect board is a PLX adapter.
|
|
0.1.14
|
|
- Added PCI ident info for the 3Com AirConnect PCI.
|
|
- FCS toggle now (really) defaults to false.
|
|
- Added PCMCIA ident info for the 3Com AirConnect 3CRW737A/3CRW777A
|
|
- David Everly's patch to kernel detection in the Configure script
|
|
- David Everly's patch to add better debian packaging bits.
|
|
-pre7
|
|
- Added kernel tags stuff.
|
|
- Added PCMCIA ident info for the Actiontec 802CI2 card.
|
|
- Added PCMCIA ident info for the Proxim RangeLAN-DS/LAN card.
|
|
- Endian "fixes" for PCI PowerMacs. This fix may be relevant for other
|
|
BigEndian/PCI (and PLX) platforms.
|
|
- FCS in monitor mode now optional; defaults to off.
|
|
- Populate the monitor channel in the prismheader.
|
|
-pre6
|
|
- monitor mode now works on USB widgets.
|
|
- Fix an alignment bug in the rx path.
|
|
- Fix a brown-paper-bag bug in the usb flash code.
|
|
- Explicitly enumerate the USB widget endpoints, and use 'em.
|
|
- REALLY fix the wlan.conf overwriting.
|
|
- More minor cleanups in the PCI/PLX init code.
|
|
- endian-ness fixes in the monitor code
|
|
- Applied a patch from David Everly to fix another 2.2.x compile problem.
|
|
-pre5
|
|
- Add USB entries for the "Melco WLI-USB-S11 11Mbps WLAN Adapter"
|
|
- Don't overwrite /etc/wlan.conf if it already exists.
|
|
- Compile sniffing code if packet socket is modularized.
|
|
- Fix a typo with the init script.
|
|
- Fix a compile problem with 2.2.x kernels
|
|
-pre4
|
|
- First bits of the SMP/locking fixes.
|
|
- Partial rewrite of the PCI/PLX/USB driver internals.
|
|
- Added a makefile for the 'addons/keygen' tool.
|
|
- Fixed a race condition in the USB code.
|
|
- Applied Godmar Back's "SSIDs with spaces" patch.
|
|
-pre3
|
|
- Applied pizza's patch to remove the pb_t structure and make all
|
|
802.11<->ether conversions happen in-place. Noticeably reduces
|
|
cpu load. Also includes a few misc bugfixes to the conversion code.
|
|
- Workaround for the kernel panic on bringing up a USB interface
|
|
- Applied Mike Klar's USB patch that adds support for the MIPS
|
|
target and should, among other things, fix WEP under USB.
|
|
- Added Remy Cool's patch to add the Dynalink prism2 card ids.
|
|
- Tim Miller's pcmcia script install path patch
|
|
- Applied Tim Fletcher's patch to allow compilation against
|
|
the newest versions of the wireless extensions. (V13)
|
|
- Added jwyatt's patch for the Compaq Evo N600C USB adapter thing.
|
|
- Per mwelk's suggestion, changed the displaystring totext to single
|
|
quote the strings. Hopefully this will take care of some of the
|
|
"SSID w/ space" problems.
|
|
-pre2
|
|
- tseward's ID code patch for Siemens SpeedStream USB adapters.
|
|
- pizza's patch for dropping/ignoring frames in promisc and allmulti
|
|
modes.
|
|
- crankin's new Actiontec USB IDs.
|
|
- crankin's fix for wext /proc zero vs. garbage problem.
|
|
-pre1
|
|
- pizza's chkconfig comment in rc.wlan.
|
|
- pizza's patch to add true promiscious support to 0.1.13. Also includes
|
|
Experimental ALLMULTI support.
|
|
- pizza's patch to add PF_PACKET capture and raw transmit to 0.1.13-pre2,
|
|
(with toggles for WEP in sniffing and prism header)
|
|
NOTE: this patch _removes_ the old netlink monitor mode (sniff) interface.
|
|
If you're sniffing, you need to use the latest libpcap (v7.2-pre..from CVS),
|
|
you'll also need to build your own ethereal (0.9.0 or greater) using the
|
|
latest libpcap.
|
|
- bhuang's fix for the hfa384x_test_command() function.
|
|
0.1.13
|
|
-pre4
|
|
- crankin's revised USB ID patch.
|
|
- Fixed a little buggage in the p80211knetdev_set_mac_address() function.
|
|
-pre3
|
|
- crankin's patch for a bug in usb rrid, DBFENTER/EXIT, and Actiontec
|
|
USB IDs.
|
|
- hmuurimaa's ID patch for the Omnibook500 integrated USB adapter
|
|
- proskin's patch for MODULE_LICENSE conditional
|
|
- dstates (JediElite) submission (others submitted it too) for the
|
|
NDC PLX card.
|
|
- crankin's patch for src/prism2/driver/Makefile. PCI and USB were
|
|
linking with PLX's prism2wext.o file. DOH!
|
|
-pre2
|
|
- Synch'd the etc/pcmcia/wlan-ng scripts with the etc/wlan* scripts.
|
|
This should help PCI/PLX/USB Adhoc users.
|
|
- Added cmulliner's patch for set_mac_address(). Also changed it to
|
|
use dot11StationID instead of the prism2 specific mib item...Oh
|
|
Yeah, I also changed to comments to /*...*/ ;-)
|
|
- Fixed PDA reads and PDA validation.
|
|
- KNOWNBUG: prism2_usb + linux-2.4.17 + uhci.o : Very slow performance.
|
|
We need to look into having multiple posted tx and rx urbs (this is
|
|
also a TODO).
|
|
- KNOWNBUG: prism2_usb + linux-2.2.20 + uhci.o : Physically
|
|
disconnecting the device causes a Panic. The logs look like the
|
|
usbdev is still holding some buffers (urbs?) associated with the
|
|
device despite the fact that I've explicitly unlinked them. This
|
|
problem does not occur on 2.4.17 and, unfortunately, I don't have
|
|
time to debug 2.2.20 this time around.
|
|
- Made all corereset() calls dependent on the prism2_doreset module
|
|
argument.
|
|
- Added corereset() support for all device types.
|
|
- Added the lnxreq_ifstate request message. This request is used
|
|
to bring the device up to an operable state _after_ the
|
|
[hostbus]_probe() function is finished. It can also be used to
|
|
shutdown and/or restart the driver+device. This takes the place
|
|
of the "activate hardware on ifconfig up" code that was introduced
|
|
in 0.1.11. This is the change that should fix most of our dhcp
|
|
client problems.
|
|
- Massive changes/cleanups to the driver init/shutdown code for all
|
|
device types (pcmcia|pci|plx|usb).
|
|
- Removed the add-ons/mibedit application from the linux-wlan-ng package.
|
|
We'll be re-releasing it as a separate tarball.
|
|
- Modified hfa384x_usb.c per dbrownell's suggestion to mark the request
|
|
urb's with the USB_ASYNC_UNLINK flag. This because we're calling
|
|
unlink_urb() from the timer callback (intterupt context).
|
|
- Added arichter's patch for the RRID framelength. It's a good catch
|
|
to make things truly correct. Unfortunately, I don't think it will
|
|
have any effect. Because of the 64-byte minimum, the actual RID value
|
|
is being transferred regardless. Also, I don't think the RRID handler
|
|
in the MAC even looks at the framelen field of RRID frames.
|
|
- Added mwelwarsky's patch for the tx path that may fix some of the
|
|
"queue empty" errors. Had to update it for 0.1.13.
|
|
- Added new PLX device IDs from cyokoyama's patch, but deferred the rest
|
|
until we can get a clean patch against the latest release.
|
|
- Added pizza's v2.1 NETLINK enable/disable patch. Note that this whole
|
|
NETLINK enable/disable thing is only barely tested by me. I _always_
|
|
have netlink available so maintaining a separate kernel setup just to
|
|
test this feature is a pain.
|
|
- Added jdiedrich's pcmcia card ids for the Compaq ipaq PCMCIA card.
|
|
- Added crankin's patch fixing some more wext induced problems and adding
|
|
the Actiontec USB ids.
|
|
- Added abridgett's patch for PRIV_GENSTRING handling.
|
|
- Added emckee's malloc.h -> slab.h patch and made the mods to the usb
|
|
and wext files he missed. Some kind list soul told me
|
|
that the change to slab.h goes way back in the 2.2.x history, so we
|
|
should be OK. If this change proves to be a problem for someone with
|
|
an old kernel, please submit a patch with ifdefs for the version where
|
|
things changed (not just the version you're using....thanks).
|
|
- Added jtourrhiles patch for wext reporting level and noise in dbm.
|
|
To enable the reporting in dbm, you must add "#define WEXT_IN_DB"
|
|
to the top of src/prism2/driver/prism2wext.c.
|
|
- Added support for the dbm-normalized RID for CommsQuality in prep
|
|
for working with Jean's patch. The mibs are called p2dbmCommsQuality*.
|
|
Note that this will only work for secondary firmware versions 1.2.0
|
|
and higher. Earlier versions will return with a
|
|
'resultcode=implementation_failure'. Note also: The values returned
|
|
are signed 16-bit integers that have been stuffed into unsigned 32-bit
|
|
integers for the trip up to userland. The following shell expression
|
|
will convert them back to the signed quantity you want:
|
|
if [ $val -lt 32767 ]; then
|
|
echo $1
|
|
else
|
|
echo $[ $1 | 0xffff0000 ]
|
|
fi
|
|
- Added mkershaw's suggestion to lengthen the wait interval for
|
|
commands in hfa384x.c.
|
|
- Added a note to the README about dleffler's observation regarding
|
|
netlink troubles and using 'make mrproper'.
|
|
- Added gback's patch to initialize wireless extensions for non-pcmcia
|
|
targets.
|
|
-pre1
|
|
- Fixed a patch induced (not sure which patch) Makefile syntax error.
|
|
- Added rfloeter's cool Wireless Extensions patch which was kindly
|
|
updated to 0.1.12 by pizza. Note that my own testing of this patch has
|
|
been minimal but I've had reports of folks who are using it successfully.
|
|
NOTE: The addition of this patch adds the requirement that Wireless
|
|
Extensions (version >= 10) be compiled into one's kernel.
|
|
- DOH! some somewhat kind soul on the lists pointed out that the keygen
|
|
program wasn't in the distributed tarball. Basically, I forgot to
|
|
check it into CVS, therefore the distribution build script couldn't
|
|
find it. Ahem...it's now in CVS and should be included in the tarballs.
|
|
- Added pizza's patch fixing the "netlink disable" patch. The
|
|
p80211ind_* functions need to be handled individually instead of two
|
|
giant #if's, but that's a task for another day.
|
|
- Added theBS's patch for making the /etc stuff installation relative and
|
|
installing the FIXINST scripts for USB.
|
|
- Added crankin's SMP deadlock fix patch.
|
|
0.1.12
|
|
- Added dlyddy's string.h includes and comma fix. Changing all malloc.h's
|
|
to slab.h's will have to wait until I figure which kernel version marks
|
|
the switch so we can ifdef it.
|
|
- Added bgertfield's Corega support patch.
|
|
- I added a check to Configure so we add the CONFIG_NETLINK symbol
|
|
on the compile command line if necessary.
|
|
- proskin's patches for netlink disable when not configured in kernel,
|
|
mkprintstr fix, and turn off skeleton install. The netlink patch
|
|
pointed out something that might be the root cause of some problems
|
|
with the _cs driver. pcmcia_cs now strips the kernel CONFIG items
|
|
it cares about and puts them in it's own include/linux/config.h
|
|
(which is symlinked to include/pcmcia/config.h).
|
|
- Added mtaylor's suggested close() to do_ioctl() in wlanctl.
|
|
- Added mkarmak's suggested change for the GL24110P device ID as
|
|
PCI_DEVICE_GL24110P_ALT and an additional item in the device table
|
|
for PLX devices.
|
|
- Added crankin's suggested MODULE_LICENSE statements. Had to figure
|
|
out the version ifdefs myself though. ;-)
|
|
- Added vragor's patch for arm updates to the Makefiles and Configure.
|
|
- Added jsuhr's patch of the PLX9052 irq enable code from the orinoco
|
|
driver. Note: PLX contributions are highly valued here because we rarely
|
|
use these devices ourselves.
|
|
- Added cfang's PCMCIA ident info for AirEZY card.
|
|
- Made USB port reset in open() an option, DEFAULT IS OFF.
|
|
- Cleaned up the top edge of hfa384x*.c (needs some more).
|
|
- Added the notion of CMD vs. AUX format MAC chip addresses.
|
|
- Modified isgoodpdr() to accept anything less than 0x1000.
|
|
- Updated the wlan-ng scripts for the new open/close behavior.
|
|
- Fixed a bunch of issues that caused bad breakage for hot-plugging
|
|
the USB module.
|
|
- Added a module argument to enable/disable the MAC (or usbport) reset
|
|
performed in prism2sta_open().
|
|
- Moved the port reset corereset() and the call to the open(). This
|
|
gets the port reset out of the probe_usb() context and appears to
|
|
solve the recursive probe problem in some cases.
|
|
- Cleaned up the overlength usb packet problems
|
|
- Added rmemreq/resp and wmemreq/resp support
|
|
- Fixed some of the 64-byte packet problems in a few commands.
|
|
0.1.11-usbonly
|
|
- Note: hfa384x*.c are currently undergoing some changes. I believe
|
|
these changes have broken some things for pcmcia/pci/plx support. I'm
|
|
not certain since I haven't tested them. As such, this is a USB ONLY
|
|
release.
|
|
- Cleaned up the use of hfa384x_drvr_*() vs. hfa384x_cmd_*() a little.
|
|
This is a work in progress.
|
|
- Rewrote the USB request/response machine again.
|
|
- Added tnewsham's keygen program to the add-ons directory. I don't
|
|
include it in the overall build because I'm not comfortable with making
|
|
this package dependent on openssl right now. It is supposed to generate
|
|
keys in a manner identical to the windows software. I have not personally
|
|
verified this. If someone wants to contribute more complete build
|
|
instructions for this program, please do.
|
|
- IMPORTANT! Note that we've changed the up/down behavior. You now
|
|
must 'up' the linux netdevice interface (e.g. ifconfig wlan0 up) prior
|
|
to using wlanctl-ng. Also, wlanctl commands aren't available after
|
|
"ifconfig down". From a low-level perspective, this is how most
|
|
linux netdevices work. We originally started trying to create a
|
|
setup where "ifconfig up" on an interface that wasn't associated would
|
|
fail. With PCI and even more so with USB, that has proven unworkable.
|
|
The idea now is:
|
|
Step 1: ifconfig wlan0 up
|
|
Enables the hardware so it's ready to handle wlanctl-ng commands.
|
|
Step 2: wlanctl-ng <various commands>, repeat
|
|
Set up the dot11 configuration and get yourself associated or
|
|
whatever.
|
|
Step 3: ifconfig wlan0 <address>
|
|
After getting a "real" 802.11 connection, assign the address
|
|
to the interface (via static numbers, dhcp or whatever).
|
|
|
|
Note that it is very important that you do _not_ down the interface
|
|
between the initial 'ifconfig up' and the later ifconfig that assigns
|
|
the address. 'ifconfig down' will shut down and reset the hardware.
|
|
If you down the interface, all that good work in step 2 will be lost.
|
|
- Rewrote the USB request/response machine
|
|
- Moved all memmapped/irq interface elements from prism2*.c to hfa384x.c
|
|
in preparation for adding USB support.
|
|
- Fixed 'make clean' so it doesn't generate .depend files.
|
|
- Removed some leftover pcmcia-cs dependencies for non-pcmcia targets.
|
|
0.1.10
|
|
- Added module param "prism2_ignorevcc". If set to non-zero, we just
|
|
accept the Voltage we get from pcmcia-cs without checking. USE WITH
|
|
CAUTION!!!!
|
|
- Changed the voltage setting code such that we never try to change it.
|
|
- Added NETGEAR PCI info.
|
|
- Fixed a problem with the interrrupt enable in mlme_start().
|
|
- Fixed the netdevice_t problem.
|
|
- Added some new CIS entries.
|
|
- Fixed some byte order things re: dblythe's patch. The problem went
|
|
a little farther than the patch so I had to change some things.
|
|
- Added lye's fix for the netdevice_t problem on various targets.
|
|
- Added an Ad-Hoc support patch from ????? (please identify yourself,
|
|
it's a nice bit of code and stays within the spirit of things). Adhoc
|
|
still needs work though. Remember, only the _first_ STA in an IBSS
|
|
should issue an MLME-Start.request. All the rest should scan/join
|
|
- Re-added the monitor mode stuff. At this point I guess Intersil won't
|
|
be taking it out. Many thanks to those who maintained the patches and
|
|
my apologies for the hassle.
|
|
0.1.9
|
|
- As part of the PCI stuff, rewrote the scripts for 'fixed' (i.e. non-
|
|
hotswappable) cards like the ISL3874 and PLX based PCI adapters.
|
|
- Added support for the dot11req_reset command. It should perform a
|
|
complete reset of the MAC and driver. Note that the MAC needs to be
|
|
completely reconfigured from userland after the reset has occurred.
|
|
Also added an internal call to the reset function when unloading a
|
|
PCI or PLX driver. We need this reset to make sure that the card is
|
|
no longer generating interrupts. Fixed this problem: If the driver
|
|
is unloaded, the card is still generating interrupts, and another active
|
|
device+driver is using the same interrupt....system locks up. DDT.
|
|
- Rewrote the PCI support for PLX and ISL3874. The rewrite was targeted
|
|
at 2.4.7 with additional code to provide backward compatibility for
|
|
2.2.x. NOTE: the PCI code still doesn't support the PCI power
|
|
management stuff. It will probably need more work when laptops w/
|
|
the ISL3874 based mini-PCI cards show up.
|
|
- Whoops, forgot to put this in the original 0.1.9 release
|
|
0.1.9-pre1
|
|
- Added support for the ISL3874 based mini-pci card. Also cleaned
|
|
up some of the PLX stuff (needs more work though).
|
|
0.1.8
|
|
- Fixed some script bugs relating to WEP and generated keys.
|
|
- Fixed the autojoin auth type control
|
|
- Added/Changed some PDR numbers (thanks to BroMax for supplying hardware)
|
|
- Received a large patch from bjames of Rebel Computing. Here's a
|
|
summary in his words:
|
|
|
|
> Minor bugs (not already found in 0.1.8-pre10):
|
|
>
|
|
> 1. p80211metamsg.c:
|
|
> - Wrong conversion functions for "p2req_mmi_read" (value). enumint ==>
|
|
> int
|
|
>
|
|
> 2. prism2sta.c:
|
|
> - "dot11_desired_bss_type" in "priv" structure left at 0 (i.e. an
|
|
> invalid value) after initialization so mibget could not retrieve a value
|
|
> unless the network was started. Initialized to 1.
|
|
>
|
|
> 3. prism2mgmt.c:
|
|
> - "prism2mgmt_channel_info()" uses HFA384x_RID_SCANREQUEST_LEN instead
|
|
> of ...CHANNELINFOREQUEST... (However, they're the same size.)
|
|
>
|
|
> 4. prism2mgmt.c:
|
|
> - "prism2mgmt_auxport_read()" did not set the resultcode to "success" if
|
|
> the read succeeded.
|
|
>
|
|
> 5. p80211types.c:
|
|
> - "p80211_fromtext_displaystr()" tests for length < maxlen. This
|
|
> probably should have been <= maxlen (i.e. excluding the \0), based on
|
|
> the maxlen values in "p80211metamib.c" and based in what
|
|
> "p80211_isvalid_displaystr()" checks. Also,
|
|
> "p80211_isvalid_displaystr()" assumes that "pstr->len" includes the \0
|
|
> when it does the minlen check. It probably shouldn't do this.
|
|
>
|
|
> 6. p80211metamsg.c:
|
|
> - dot11_reassociate command has a "dot11req_reassociatefailuretimeout"
|
|
> argument which should probably be "reassociatefailuretimeout".
|
|
>
|
|
> 7. prism2sta.c:
|
|
> - WLAN_PCI ==> WLAN_PLX in "init_module()".
|
|
>
|
|
> Enhancements:
|
|
>
|
|
> 1. All Prism2 RIDs are now accessible. Unfortunately, rather a lot was
|
|
> changed to do this:
|
|
> - Rather than cut and paste many more cases into the monster
|
|
> "prism2mib.c" switch statement, "prism2mib.c" was completely reworked to
|
|
> be table-driven with a bunch of semi-generic processing functions. That
|
|
> made it a lot easier to add new DIDs. Several RID lengths in hfa384x.h
|
|
> had to be changed from 0 to their actual value since more of the lengths
|
|
> get used, now. (Changed: prism2mib.c, hfa384x.h)
|
|
> - 2 new data types were added to "p80211types.c". One to handle integer
|
|
> arrays (i.e. multi-word RID records...so that a DID didn't need to be
|
|
> added for each element) and bit arrays (just to make them easier to deal
|
|
> with). However, there wasn't enough room in the packed DID to add new
|
|
> types, so the type was removed from the DID (and from P80211DID_MKID)
|
|
> and a function written to determine the type based on the conversion
|
|
> functions. (Changed: p80211meta.h, p80211types.h, p80211meta.c,
|
|
> p80211types.c, mkmetastruct.c, p80211metamib.c p80211metamsg.c,
|
|
> wlanctl/help.c)
|
|
> Integer arrays look like: p2PRIIdentity=21,0,0,3
|
|
> "meta->maxlen" specifies the number of integers.
|
|
> Bit arrays look like: p2ChannelList=0,1,2,3,4,5,6,7,8,9,10
|
|
> for bits 0-10. "meta->min" and "meta->max" give the range of bits
|
|
> allowed.
|
|
> - The "p2Table" DIDs were split into p2Static, p2Dynamic, p2Behavior,
|
|
> etc., in order to correspond with the Prism Driver Programmer's Manual.
|
|
> A few of the names were also changed in order to correspond to the
|
|
> manual. (Changed: p80211metamib.c, prism2mib.c) The changed names (of
|
|
> the supported DIDs) are:
|
|
> p2SystemScale ==> p2CnfSystemScale
|
|
> p2MaxDataLen ==> p2CnfMaxDataLength
|
|
> p2EnhancePowerSaveMode ==> p2CnfPMEPS
|
|
> p2MaxSleepDurection ==> p2CnfMaxSleepDuration
|
|
> p2PMHoldoverDuration ==> p2CnfPMHoldoverDuration
|
|
> p2OwnName ==> p2CnfOwnName
|
|
> p2McastPMBuffer ==> p2CnfMulticastPMBuffering
|
|
> p2TxControlDepth ==>p2CnfTxControl
|
|
> p2RoamingMode ==> p2CnfRoamingMode
|
|
> p2MMLife ==> p2CnfMMLife
|
|
> p2AltRetryCount ==> p2CnfAltRetryCount
|
|
> p2ReserveDuration ==> p2CnfAPPCFInfo (bit 1)
|
|
> p2ExcludeLongPreamble ==> p2CnfExcludeLongPreamble
|
|
> - Added new RIDS from the Prism Driver Programmer's Manual 1.70
|
|
> (2000-11-29). Added p2CnfPriorityQUsage, p2CnfTimCtrl,
|
|
> p2CnfThirty2Tally, p2CnfEnhSecurity. Added "algorithm" to AuthRequest.
|
|
>
|
|
> 2. A user utility program was written to do such things as query all
|
|
> read-able MIBs, query all write-able MIBs, set all write-able MIBs,
|
|
> etc. A "p2req_enable" command was added (similar to the
|
|
> "dot11req_start" command but without requiring all the 802.11
|
|
> arguments). This allows the driver to be started in a straight forward
|
|
> manner, simply by setting all write-able MIBs and then enabling it.
|
|
> (Changed: p80211metamsg.c, prism2sta.c, prism2mgmt.c, prism2mgmt.h)
|
|
>
|
|
> 3. Support added for Prism2 CommTallies and a DID added to query the
|
|
> values. The values will come back as an array of integers. Support
|
|
> also added for 32-bit tallies (i.e. p2CnfThirty2Tally). (Changed:
|
|
> prism2mib.c, prism2sta.c, prism2mgmt.h, p80211metamib.c, hfa384x.h)
|
|
>
|
|
> 4. Access Point support:
|
|
> - Added support for maintaining a list of authenticated/associated
|
|
> stations and a DID added to query the list. (Changed: prism2mib.c,
|
|
> prism2sta.c, prism2mgmt.h, p80211metamib.c)
|
|
> - Added support for host control of authentications (allowed and denied
|
|
> lists) and a DID added to set and query the lists. Support for Prism2
|
|
> AuthenticationRequest information frames. Not tested yet. (Changed:
|
|
> prism2mib.c, prism2sta.c, prism2mgmt.h, p80211metamib.c)
|
|
> - Added support for Prism2 PowerSaveUserCount information frames and a
|
|
> DID added to query the value. (Changed: prism2mib.c, prism2sta.c,
|
|
> prism2mgmt.h, p80211metamib.c)
|
|
> - Added a new data type to "p80211types.c" to handle arrays of MAC
|
|
> address (used to set/query the above lists). (Changed: p80211types.h,
|
|
> p80211types.c, mkmetastruct.c, wlanctl/help.c)
|
|
> Address arrays look like:
|
|
> p2Authenticated=00:50:51:52:53:54,00:60:61:62:63:64,00:70:71:72:73:74
|
|
> "meta->maxlen" specifies the maximum number of addresses.
|
|
> - Unforetunately, all these lists are limited in length by the length of
|
|
> the mibattribute parameter in mibget/mibset. This is not very nice but
|
|
> it was the quickest/easiest way to implement it.
|
|
>
|
|
> 5. Changed maximum mibattribute length from 256 to 384 to support
|
|
> querying commtallies and MAC address arrays. (Changed: p80211types.h)
|
|
>
|
|
> 6. 128-bit WEP key support:
|
|
> - Rather than creating different MIB's for 64-bit and 128-bit WEP keys,
|
|
> the "octetstr" definition for dot11WEPDefaultKey0, etc., was changed to
|
|
> allow for a range of lengths (e.g. from 5 to 13). The
|
|
> dot11WEP128DefaultKey0, etc., MIBs were removed. (Changed:
|
|
> p80211meta.c, p80211types.c, p80211metamib.c, p80211metamsg.c,
|
|
> mkmetastruct.c, wlanctl/help.c)
|
|
> - WLAN_WEP_KEYLEN changed to WLAN_WEP_MAXKEYLEN and set to 13.
|
|
> (Changed: p80211hdr.h)
|
|
> - Optional "length" argument added to "nwepgen.c" so that either 5 or 13
|
|
> byte keys can be handled. (Changed: nwepgen.c)
|
|
> - 128-bit-specific parts of "etc/pcmcia/wlan-ng" are no longer
|
|
> necessary. Genstr changes to handle generation of 13-byte keys using
|
|
> nwepgen. (Changed: etc/pcmcia/wlan-ng)
|
|
> - WEPDefaultKey's were made write-only. A new
|
|
> "P80211ENUM_resultcode_cant_get_writeonly_mib" was added. (Changed:
|
|
> p80211types.h, p80211types.c, p80211metamib.c, prism2mib.c)
|
|
>
|
|
> 7. Access code in DID:
|
|
> - This was changed from a single bit flag (P80211DID_ACCESS_READONLY and
|
|
> P80211DID_ACCESS_READWRITE) to a pair of flags (P80211DID_ACCESS_READ
|
|
> and P80211DID_ACCESS_WRITE) since some MIBs are read/write, some are
|
|
> read-only, and some are write-only (e.g. WEP keys). This makes it
|
|
> easier to determine which MIBs are read-able and write-able without
|
|
> having to go all the way to the driver and then interpret a failure
|
|
> error code. (Changed: p80211types.h, p80211metamib.c)
|
|
>
|
|
> 8. It was a bit of a hassle switching between the station driver and the
|
|
> AP driver...which we seemed to be doing a lot. The additional cost was
|
|
> not particulary high to simply compile in both station functionality and
|
|
> AP functionality. A change was made so that "prism2sta_initmac()" set
|
|
> an "ap" flag when it detected AP firmware. "prism2mgmt.c" then uses the
|
|
> flag to determine whether or not to allow station-only or AP-only
|
|
> commands to be executed. "prism2mib.c" also uses this to handle the
|
|
> mode-specific MIBs. This made it possible for the same module to
|
|
> support either stations or APs, thereby simplifying the "make", the
|
|
> install, and switching between stations and APs. The only thing that
|
|
> needs to change, is the value of "IS_AP" in "wlan-ng.opts".
|
|
>
|
|
> 9. Miscellaneous small additions:
|
|
> - Network state (stopped, started) added and a DID to query it.
|
|
> - User-settable comment string added and a DID to set/query it.
|
|
> - Some event logging (authentications, associations, etc.) can be
|
|
> enabled/disabled at run time rather than only at compile time. A DID
|
|
> was added to set/query it.
|
|
> - A "p2req_readcis" command (similar to "p2req_readpda") was added.
|
|
>
|
|
> 10. Small changes:
|
|
> - Added bounds checking to "p80211item_fromtext_boundedint". This was
|
|
> the only "fromtext" function that did not already do complete
|
|
> validation. It should no longer be necessary to call the "isvalid"
|
|
> function if "fromtext" has just been called. The call to "isvalid" was
|
|
> removed from "wlanctl". (Changed: p80211types.c, wlanctl.c)
|
|
> - "p80211item_maxitemlen()" wasn't really adding much of value so it was
|
|
> amalgamated into "p80211item_getoffset()". The DID type is now only
|
|
> checked once instead of three times. (Changed: p80211meta.h,
|
|
> p80211meta.c)
|
|
> - "hfa384x_drvr_setconfig16()" and "hfa384x_drvr_setconfig32()" changed
|
|
> their input parameters when converting to hf384x byte order. Potential
|
|
> source of bugs. Changed to use a local variable. (Changed: hfa384x.c)
|
|
> - ScanRequest, JoinRequest, etc., moved from the "Information Record"
|
|
> section to the "Configuration Record" section. (Changed: hfa384x.h)
|
|
> - Re-worked "p80211_fromtext_setmibattribute()" to make it a little
|
|
> easier to follow the logic. (Changed: p80211types.c)
|
|
> - Type of "channellist" argument in "p2req_channel_info" command changed
|
|
> from "octetstr" to "bitarray" to make it easier to deal with (can
|
|
> specify a variable number of channels and don't need to use hex
|
|
> values). (Changed: p80211metamsg.c, prism2mgmt.c)
|
|
> - "auxctl" parameter set to HFA384x_AUX_CTL_EXTDS in
|
|
> "prism2mgmt_auxport_read()" and "prism2mgmt_auxport_write()" so that
|
|
> they are still compatible with 0.1.7. (Changed: prism2mgmt.c)
|
|
> - Re-worked "prism2mgmt_set_grpaddr()" and added comments so that it's
|
|
> more obvious what is going on. Also, deleted the check which produced
|
|
> an error if an attempt was made to delete an address from an empty
|
|
> list. This makes things more consistent and predictable (i.e. the same
|
|
> as attempting to delete an address which is above the last group
|
|
> address...the attempt is simply ignored). Replaced the "memcpy()" call
|
|
> with "memmove()" since "memcpy()" is not guaranteed to work with
|
|
> overlapping ranges. (Changed: prism2mib.c)
|
|
> - Minimum length of "dot11OperationRateSet" changed from 126 to 1 to
|
|
> match the 802.11 standard. (Changed: p80211metamib.c)
|
|
|
|
- Fixed some ci/co related problems pointed out by wbinjie.
|
|
- Added ygalayda's startup scripts for PCI in the directory
|
|
./etc/wlan_pci. They aren't installed, that will have to be done
|
|
by hand. I have not tested these scripts myself.
|
|
- Added sbaer's 'NULL check' fix in p80211_indicate_shutdown().
|
|
- Added ctedrow's 'alloc not checked' fix in prism2sta_int_rx().
|
|
- Added pkundrat's auth mib changes.
|
|
- fgrau's conf.modules/modules.conf fix.
|
|
- Added ygalayda's submission for the Global Sun GL24110P PLX based
|
|
adapter. Also included a Makefile fix.
|
|
- Fixed some script bugs relating to 128-bit WEP based on smmclusky's
|
|
observations. These changes have not yet been tested by me.
|
|
- Added rlazarev's additions to the pcmcia startup scripts
|
|
- Added pkundrat's wlandev-name fix for 2.3.99+ kernels
|
|
- Fixed some type/(to|from)text mismatches per note from pkundrat
|
|
- Fixed some missing MKITEMNAMEs per note from pkundrat
|
|
- Added D-LINK and SMC to the linux-wlan-ng.conf file
|
|
- Added Zoom, Linksys and Addtron to the linux-wlan-ng.conf file
|
|
- Added pkundrat's meta-item minlen stuff.
|
|
- Added conditionals for PLX vs. PCMCIA build
|
|
- Changed _pci driver filename to _plx
|
|
- Fixed undconditional MOD_INC in prismt2sta_open()
|
|
- Removed identification of old PCF firmware.
|
|
- Fixed version.h generation bug from pre5
|
|
- Added dhsu's (Eumitcom) code for supporting the PLX based PCI cards.
|
|
- pkundrat's (whoops, sorry peter) PDA display program in ./scripts.
|
|
I haven't used it because my perl setup is messed up and I'm perl
|
|
challenged.
|
|
- bcarr's patches for alpha boxes.
|
|
- wwoods's patches for 2.2.18 and adding bytes received to the
|
|
statistics.
|
|
- proskin's patch adding 'network stop'
|
|
- Most of proskin's patch containing Configure fixes, adding file
|
|
existence checks, and making sure directories are present.
|
|
- tgriffin's patch for mrproper/memleak/nullcheck/freeonsuccess.
|
|
conf.modules change is already there.
|
|
- Added most of jhicks ARM related patches. I had to leave out the
|
|
HOST vs. TARGET compile variables in the Makefiles though. Some folks
|
|
are using that stuff.
|
|
- Added script changes and new mib items to support 128-bit WEP per hong's
|
|
suggestion.
|
|
- Added conf.modules alias instead of symlink
|
|
- Added a compat macro for kfree_s to wlan_compat.h per the suggestion
|
|
from rbraun.
|
|
- Added pkundrat's 'chinfo' patch
|
|
- Added pkundrat's 'small fixes' patch
|
|
- Added some contributed patches supporting a fix to our 'reassoc
|
|
under load' problem
|
|
- Added wlannoenable pcmcia scheme support to prevent card initialization.
|
|
- Removed a PDR patch requirement.
|
|
0.1.7
|
|
- Fixed the "converting eth in unknown mode" bug
|
|
- Added APM/ACPI suspend/resume handling
|
|
- Added PPC support
|
|
- Added Prism test command support
|
|
- Integrated patch to fix some structure alignment problems for ARM
|
|
- Added multi-transmit buffer support
|
|
- And much, much more.....
|
|
0.1.6
|
|
- Fixed usage of pcmcia scheme file so we're not stomping on the user's
|
|
scheme setting.
|
|
- Updated the wlan-ng script for the new config variable names and added
|
|
configurable download utility and a set of user MIB settings.
|
|
- Added support for the Samsung and Z-Com cards.
|
|
- Changed the name of the module (from pcmcia's point of view).
|
|
- Fixed the "echo \c" problem in Configure
|
|
- Added new and rearranged existing wlan-ng.opts configuration variables.
|
|
- Created a scheme for "addon software" in the ./src/Makefile
|
|
- Made some changes to the target detection conditionals in wlan_compat.h
|
|
to better support LinuxPPC (port isn't finished yet).
|
|
- Finally fixed the byteorder stuff to use (duh) byteorder.h
|
|
- Fixed a bad free of the TX skb when we signal an error to higher layers.
|
|
- Moved some code around and added rules to the prism2/driver (new directory
|
|
too) Makefile. This is so we build two separate drivers; one for STA,
|
|
one for AP. Both drivers are built from the same source files with
|
|
the code conditional on the WLAN_AP and WLAN_STA defines.
|
|
- Fixed the plugging in prism2dl so it reports ALL missing PDRs.
|
|
- Added an anonymously contributed patch that adds support for the -m
|
|
and -s cmdline options for prism2dl. Thanks! You know who you are.@-)
|
|
- Added some fixes to the dependency file generation.
|
|
- Fixed a flash programming timeout miscalculation...and fixed it again
|
|
when I introduced an integer promotion problem.
|
|
- Added the STA vs. AP conditionals to prism2mgmt. Note that this causes
|
|
a few compile time warnings. They're harmless and will go away in the
|
|
next release.
|
|
- Created the file prism2mib.c to hold all the MIB get/set stuff which is
|
|
getting a little out of hand.
|
|
- Added pcmcia config support for the Z-Com (supposedly) dual voltage card.
|
|
Unfortunately the engineering sample I have doesn't seem to work at 5v,
|
|
or it might be the code.....not sure yet. We're still a little
|
|
bleeding edge with that card.
|
|
- Cleaned up the version, compatibility range, and serial number
|
|
log messages.
|
|
- Added a loop limit to the interrupt handler. (duh)
|
|
- Changed the names of the driver modules for STA and AP. make install
|
|
then creates a softlink depending on the make config selection. One
|
|
note, both drivers report the same dev_info string to pcmcia-cs. Trying
|
|
to load both of them simultaneously would be bad (I have no idea what
|
|
would happen, I just thought of it).
|
|
- Changed a bunch of types and constants for hfa384x.
|
|
- Added functions for a couple of new commands in the hfa384x.
|
|
- Began adding support for the Prism2 unique MIB items.
|