summaryrefslogtreecommitdiffstats
path: root/abs/core-testing/wlan-ng26-utils/tmp/trunk/CHANGES
diff options
context:
space:
mode:
Diffstat (limited to 'abs/core-testing/wlan-ng26-utils/tmp/trunk/CHANGES')
-rw-r--r--abs/core-testing/wlan-ng26-utils/tmp/trunk/CHANGES1267
1 files changed, 1267 insertions, 0 deletions
diff --git a/abs/core-testing/wlan-ng26-utils/tmp/trunk/CHANGES b/abs/core-testing/wlan-ng26-utils/tmp/trunk/CHANGES
new file mode 100644
index 0000000..f286bf6
--- /dev/null
+++ b/abs/core-testing/wlan-ng26-utils/tmp/trunk/CHANGES
@@ -0,0 +1,1267 @@
+* 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 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.