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