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, 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. |