changelog

无线网卡驱动程序 802.11b无线网卡可直接移植到2.6内核中去

2006-06-23 01:09  agx

	* at76c503a/: at76c503.c, at76c503.h: assume WIRELESS_EXT >= 16

2006-06-22 23:21  agx

	* at76c503a/at76c503.c: fix compilation for WE >19

2006-06-22 22:58  agx

	* at76c503a/at76c503.c: use iwreq_data instead of iw_point in
	  at76c503_iw_handler_set_scan as the ipw2X00 drivers do

2006-06-22 22:52  agx

	* at76c503a/at76c503.c: Pavel Roskin <[e-mail]>:

	  Scanning doesn't work when the interface is down.  When it's
	  brought up, the driver reports two assertion failures:

	  /home/proski/src/at76c503a/at76c503.c:
	  /home/proski/src/at76c503a/at76c503.c:2806 assertion dev->istate
	  == INIT failed/home/proski/src/at76c503a/at76c503.c:
	  /home/proski/src/at76c503a/at76c503.c:2777 assertion dev->istate
	  == SCANNING +failed

	  To prevent this, scanning should not be allowed when the
	  interface is down.

2006-06-22 22:48  agx

	* at76c503a/at76c503.c: vel Roskin <[e-mail]>:

	  The current CVS version of ignores ESSID settings.  When ESSID is
	  set using iwconfig, the device reports that it's disconnected,
	  and then associates to a random AP in the range, even it its
	  ESSID is different.

	  The simplest test is to set ESSID to a non-existing string (e.g.
	  "non-existing") and see it associate.  It happens whenever the
	  interface is up.  ESSID can be set when the interface is up or
	  down.

	  The fix it to set ESSID length properly

2006-06-22 15:14  maximsch2

	* at76c503a/at76c503.c:
	  Another attempt to fix wrqu error.

2006-06-22 14:58  maximsch2

	* at76c503a/at76c503.c:
	  Cleaned some more 2.4 staff.

2006-06-22 14:35  maximsch2

	* at76c503a/at76c503.h:
	  Removed some 2.4 specific staff

2006-06-22 14:18  maximsch2

	* at76c503a/at76c503.c:
	  Dirty hack to avoid problems with wrqu and min_scan_time and
	  max_scan_timenE

2006-06-22 14:09  maximsch2

	* at76c503a/at76c503.h:
	  Removed packed attribute from byte array. Thanks to Pavel Roskin.

2006-06-21 12:50  maximsch2

	* at76c503a/: Makefile, at76_usbdfu.c, at76c503-fw_skel.c,
	  at76c503.c, at76c503.h, scripts/fwbin2pkg.sh:
	  Applied fixes from debian patchset. Fixes compilation with 2.6.17
	  and other issuse. Added some fixes to WE code.

2006-02-12 20:02  tim_small

	* at76c503a/fw-rfmd-0.90.2-140.h: Backout bogus change.

2006-02-11 22:22  tim_small

	* at76c503a/fw-rfmd-0.90.2-140.h: Make this firmware file useful
	  for static inclusion.

2006-02-11 22:18  tim_small

	* at76c503a/: at76c503.c, at76c503.h:
	  Apply Pavel's patch from
	  https://lists.berlios.de/pipermail/at76c503a-develop/2005-September/000235.html

	  Have tested on ARM, and x86, with Netgear MA101, and a Linksys
	  WUSB11 v2.6 As I don't have time for more testing, I'm just going
	  to commit (Pavel should know what he's doing ;o) ).

2005-03-08 04:33  jal2

	* at76c503a/at76c503.c: replaced usb_unlink_urb by usb_kill_urb for
	  >= 2.6.10 (as posted by Deuce <[e-mail]> Jan 3, 2005)

2005-03-08 03:07  jal2

	* at76c503a/: at76c503-fw_skel.c, at76c503.c, at76c503.h,
	  at76c505a-rfmd2958.c: version 0.12beta23: - at76c505a-rfmd2958.c:
	  added Gigaset11 (0x1690/0x0701) (thanks to Mike Pinna) - added
	  patch to support HP iPAQ h5[45]xx (see email by Erik Hovland, 28
	  Febr 2005)

2004-10-20 00:45  jal2

	* at76c503a/: at76c503.c, at76c503.h: version 0.12beta22, fixed
	  compiler error: /usr/src/atm/at76c503a/at76c503.c:3903: error:
	  union has no member named 'ethernet'

2004-10-14 05:21  jal2

	* at76c503a/: at76c503.c, at76c503.h: - version 0.12beta21 - iwlist
	  wlanx scanning: fixed bug which showed only parts of the result:
	  it seems like iwlist is polling by SIOCGIWSCAN, i.e. invoking
	  _get_scan,   which must return -EAGAIN if the scan is still
	  running.  - introd. SCAN_POLLING_INTERVAL and poll faster to
	  decrease startup time

2004-10-14 03:58  jal2

	* at76c503a/: at76c503.c, at76c503.h: version 0.12beta20, fixed
	  compiler bug with undefined IW_MONITOR_MODE for kernel 2.4.19 (WE
	  < 15)

2004-09-19 23:17  jal2

	* at76c503a/Makefile: added target install-modules for make-kpkg

2004-09-05 18:07  jal2

	* at76c503a/: at76_usbdfu.c, at76_usbdfu.h, at76c503.c, at76c503.h:
	  version 0.12beta19: fixed some problems with at76c505a and 2.6.x
	  kernels, probs with ohci remain here :(

2004-08-29 15:41  jal2

	* at76c503a/: at76c503.c, at76c503.h: - version 0.12beta18 - added
	  support for international roaming - added support for
	  "pseudo-monitor" mode, i.e. Kismet displays	the contents of
	  beacon and probe responses   (both items thanks to the patch by
	  Balint Seeber) - removed polling for scan completion in kevent()
	  - this will	hopefully remove the temporary system locks in
	  managed mode	 when no AP is found. Removed RESCAN_TIME etc.

2004-08-27 00:41  jal2

	* at76c503a/: Makefile, at76c503.c: Makefile: ignore errors at
	  rmmod ; at76c503.c: range of signal is 0..255 (fix -XXX dBm
	  displays in iwconfig for Fiberline WL-240

2004-08-19 02:01  jal2

	* at76c503a/: Makefile, Makefile.k26, at76_ieee802_11.h,
	  at76_usbdfu.c, at76_usbdfu.h, at76c503-fw_skel.c,
	  at76c503-i3861.c, at76c503-i3863.c, at76c503-rfmd-acc.c,
	  at76c503-rfmd.c, at76c503.c, at76c503.h, at76c505-rfmd.c,
	  at76c505-rfmd2958.c, at76c505a-rfmd2958.c, fw-empty.h,
	  ieee802_11.h, kernel_patch.sh, usbdfu.c, usbdfu.h: version
	  0.12beta17: - Makefile: new targets:	  * rmmod - remove all
	  loaded modules in correct order (for tests etc.)   * kernel_patch
	  - try to copy the source files and patch Kconfig
	  and Makefile, for 2.6.x only - use with care !  * VERSION went
	  into at76c503.h

	  - file names cleaned up:	  * usbdfu.*  --> at76_usbdfu.*   *
	  removed fw-empty.h	* ieee802* --> at76_ieee802*

	  - no debugs even if CONFIG_USB_DEBUG is defined - fixed copyright
	  entries - added ref. to Berlios project's homepage

2004-08-14 02:45  jal2

	* at76c503a/at76c503.c: *do_probe(): firmware 0.90.x (Intersil)
	  shows error "control timeout on ep0in"   when querying the
	  firmware version from the device and a subsequent call to
	  get_op_mode() fail. Fortunately we can overwrite an existing
	  external firmware, so we force the external fw download here if
	  the device   has no internal flash.

2004-08-14 02:42  jal2

	* at76c503a/Makefile: - version 0.12beta16 - removed -g for 2.6.x
	  build (make the .o .ko files smaller) - install: don't rename old
	  driver into *.orig as the 2.4.x depmod seems	 to parse _all_
	  files under /lib/modules/`uname -r`/; remove old version now
	  (and all *.orig)

2004-08-13 04:17  jal2

	* at76c503a/: Makefile, at76c503.c, at76c503.h: - version
	  0.12beta15 - added support for >= 2.6.8: USB subsys will
	  disconnect after the device is   reset. We get reconnected
	  afterwards with the internal fw loaded (only), which	 we can
	  recognize by a failure in getting the firmware version.    added
	  new state WAIT_FOR_DISCONNECT, added new KEVENT_RESET_DEVICE

2004-08-11 03:10  jal2

	* at76c503a/at76c503.c: take USB_CTRL_GET_TIMEOUT from the kernel
	  include if defined there

2004-08-11 03:05  jal2

	* at76c503a/at76c503.c: - fixed "Bad board type set (7).  Unable to
	  get hardware config." for   505a-rfmd2958 - fixed "control
	  timeout on ep0in" for 505a-rfmd2958	(both thanks to
	  <[e-mail]>) - added some debug code to catch a spurious
	  error where submit_tx   seems to return -EINVAL from the USB
	  subsys

2004-08-02 02:52  jal2

	* at76c503a/Makefile: install: rename any old driver file into
	  *.orig

2004-07-24 03:28  jal2

	* at76c503a/Makefile: fixed typo in extension

2004-06-30 01:26  jal2

	* at76c503a/at76c503.c: fixed bug with le16_to_cpus showing up on
	  bigendian archs only (which should have been fixed by version
	  1.54)

2004-06-16 00:02  jal2

	* at76c503a/at76c505a-rfmd2958.c: added boardtype
	  BOARDTYPE_505A_RFMD_2958 and new firmware and driver variant file
	  (not tested yet - probably needs some special handling in
	  at76c503.c)

2004-06-15 23:53  jal2

	* at76c503a/: Makefile, at76c503.h, at76c505-rfmd.c,
	  fw-pkg-505a-rfmd2958-1.102.0-113.h, gen_fw.c: added boardtype
	  BOARDTYPE_505A_RFMD_2958 and new firmware and driver variant file
	  (not tested yet - probably needs some special handling in
	  at76c503.c)

2004-06-15 22:20  jal2

	* at76c503a/at76c503.c: fixed compile bug for WE version <= 13

2004-06-14 04:05  jal2

	* at76c503a/: at76c503.c, at76c503.h: remove debug (hexdump) buffer
	  to further reduce stack usage

2004-06-14 02:23  jal2

	* at76c503a/at76c503.c: at76c503_ioctl: cleaned up the case:

2004-06-14 02:03  jal2

	* at76c503a/: Makefile, at76c503.c: Nick Jones' patch to fix the
	  crash with kernel versions with 4KSTACKS set (2.6.5+) when
	  iwconfgi is called

2004-06-12 14:59  jal2

	* at76c503a/: Makefile, at76c503.c, ieee802_11.h: - version
	  0.12beta13 - tx data: send correct SNAP header, which should fix
	  bug with WindowsXP as peer - rx data, ieee802_11_hdr: cleaned it
	  up a bit, removed addr4 which is   not used by non-APs

2004-05-31 17:59  jal2

	* at76c503a/: Makefile, at76c503.c: version 0.12beta12: added
	  wireless events

2004-05-30 02:16  jal2

	* at76c503a/at76c503.c: fixed bug with compiling for >= 2.4.25 and
	  on big endian machines (Nick's patch from email "Fixup for >
	  2.4.25", 22/5/2004)

2004-04-16 23:06  jal2

	* at76c503a/at76c503.c: iwlist wlanX scan showed APs as Managed.
	  Changed to Master

2004-04-15 02:15  jal2

	* at76c503a/: Makefile, at76c503.c, at76c503.h: version 0.12beta11,
	  added patch by Nick Jones to specify the wanted BSSID with
	  iwconfig

2004-04-15 01:06  jal2

	* at76c503a/at76c503-rfmd.c: added devices: Gigabyte GN-WLBM101,
	  Planex GW-US11S, Compaq iPAQs

2004-04-10 22:56  jal2

	* at76c503a/at76c503.c: at76c503_iw_handler_set_encode: returns
	  -EIWCOMMIT (was: 0)

2004-04-10 00:56  jal2

	* at76c503a/at76c503.c: fixed bug in iw_handler list; CGIWRANGE:
	  max retry is 255 (1 byte value, no idea if 255 will work)

2004-04-10 00:47  jal2

	* at76c503a/at76c503.c: enabled set retry in iwconfig

2004-04-04 02:06  jal2

	* at76c503a/Makefile: version 0.12beta10 (needed for documentation
	  on firmware download proc.)

2004-03-29 02:12  jal2

	* at76c503a/Makefile: mark versions without compiled-in firmware by
	  suffix -fwdl

2004-03-27 00:54  jal2

	* at76c503a/at76c503.c: made beacon scanning more robust (Nick
	  Jones' patch)

2004-03-27 00:53  jal2

	* at76c503a/Makefile: if no file fw-pkg*.h is avail., compile for
	  firmware download, create non-firmware distribution tgz, too

2004-03-22 20:10  jal2

	* at76c503a/at76c503.c: fixed SET_NETDEV_DEV for 2.6.x kernels
	  (patch by Pavel Roskin, Sat, 20 Mar 2004, [PATCH] Fix for
	  SET_NETDEV_DEV

2004-03-18 23:54  jal2

	* at76c503a/: Makefile, README, at76c503-fw_skel.c, at76c503.c,
	  at76c503.h: - version 0.12beta9 - Nick's new iw_handler
	  implementation, supports "iwlist ..." and more - authentication
	  method is now choosen by encr XXX restr, not by iwpriv anymore -
	  removed "iwpriv wlan0 list_bss" command

2004-03-18 01:35  jal2

	* at76c503a/: at76c503.c, at76c503.h: replaced kfree on a netdev by
	  free_netdev (patch provided by Pavel Roskin

2004-03-16 23:15  jal2

	* at76c503a/at76c503.c: - treat ' ' as a cloaked SSID, too -
	  rx_mgmt_beacon: write \0 after copying the SSID

2004-03-16 18:23  njones

	* at76c503a/at76c503.c: Some code cleanups: - removed a stray
	  comment near the top - Improved the debugging output in the
	  _get/set_encode iw_handlers

2004-03-16 18:06  njones

	* at76c503a/at76c503.c: Changes made based on information from
	  Jean, both in _get_scan: - When returning ESSID information,
	  flags = 1 -> SSID on,   flags = 0 -> SSID off/any - With the
	  encoding information, we add to the flags: | IW_ENCODE_NOKEY
	  This indicates that the encryption keys are not returned as part
	  of the   scan information.

2004-03-15 20:27  njones

	* at76c503a/at76c503.c: Passing through the various TODOs embedded
	  in the code.	Some TODOs have been removed because fixes were
	  made or it was agreed that no changes were needed in those cases:
	  - handle_scan: used EPERM as a more suitable return code if the
	  device is   found to be in the wrong state when this function is
	  called - _get_range: max/avg_qual.qual is set to zero unless the
	  board has an	 intersil radio chipset.  - min/max_r_time settings
	  removed as the device doesn't support these.	- _set_scan now
	  returns a more appropriate error code.  - in get essid and
	  nickname handlers, 1 is added to the length of the   returned
	  string to include the '\0'.  This is the same behaviour as
	  displayed by iwconfig when these values are set.

2004-03-14 19:02  njones

	* at76c503a/at76c503.c: - Cleaned up a lot of the mess that was
	  present in the spy handlers.	The   code is a bit more pleasant
	  to look at now

2004-03-13 16:43  njones

	* at76c503a/at76c503.c: - In the ioctl for SIOCGIWSPY, fixed up the
	  process of copying the data	back to user space.  Frankly it was
	  absurd before - my fault.  The whole	 array of spy statistics is
	  translated in one go.

2004-03-13 16:24  njones

	* at76c503a/at76c503.c: Changes centre around the TODOs that were
	  left in the iw_handler code.	- Removed several TODOs that I felt
	  were no longer relevant - in get_rts, rts->disabled = (rts->value
	  >= MAX_RTS_THRESHOLD); rather   than == MAX_RTS_THRESHOLD

2004-03-11 21:40  njones

	* at76c503a/at76c503.c: - yet again, removed the copy_from_user
	  stuff from the SET_DEBUG private   ioctl.  This should work now
	  because the private ioctl commands have   been renumbered as
	  suggested by the comments in net/core/wireless.c   copy_from_user
	  should be done before this function is reached and this
	  function should work properly now.	If it doesn't, Joerg is
	  going to kill me!  - Because the private ioctl command number is
	  used as an index into the   table of private ioctl functions, and
	  Because the private ioctl command   numbers have been changed in
	  at76c503.h, it is necessary to insert   'gaps' in the private
	  ioctl function table so that the functions match   the new
	  command numbers.

2004-03-11 21:31  njones

	* at76c503a/at76c503.h: - 'setter' iw_handler command codes must
	  have an even value, otherwise   they will not be set up and
	  called properly from the wireless subsystem

2004-03-10 00:13  jal2