Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo


#14 DWL-520+ Error trying to open firmware package


All instructions followed to the letter, multiple firmware
versions for acx100 tried and failed, including
airplus.bin (renamed wlangen.bin) supplied with card.
Card D-link DWL-520+. I get the following message:

Nov 1 14:26:34 localhost kernel: init_module: TI
acx100_pci.o: Ver 0.1h Loaded
Nov 1 14:26:34 localhost kernel: init_module:
dev_info is: TI acx100_pci
Nov 1 14:26:34 localhost kernel: PCI: Found IRQ 11
for device 00:08.0
Nov 1 14:26:34 localhost kernel: A Acx100 PCI device
found, phymem1:0xd8010000, phymem2:0xd8000000,
irq:11, mem1:0xd2c3e000, mem2:0xd2c40000
Nov 1 14:26:35 localhost kernel: stop queue after
Nov 1 14:26:35 localhost kernel: ERROR 2 trying to
open firmware image file 'firmware/WLANGEN.BIN'.
Nov 1 14:26:35 localhost kernel: acx100_read_fw
Nov 1 14:26:35 localhost kernel: Failed to download
firmware to the ACX100
Nov 1 14:26:35 localhost kernel: acx100_probe_pci: TI
acx100_pci: MAC initialize failure!
Nov 1 14:26:35 localhost kernel: init_module:
acx100_pci: No devices found, driver not installed

I am probably stupidly overlooking the obvious, but any
help would be very much appreciated.

OS: Redhat 9


  • Rick Phillips
    Rick Phillips

    Logged In: YES

    Two of us using different packages but getting the
    identical result can't be stupid. After trying to extract
    the firmware image, my WLANGEN.BIN file ends up with a
    byte size of 0. I am using Mandrake 9.2 with kernel
    2.4.22-21. I too have tried renaming airplus.bin to
    WLANGEN.BIN with identical results - no extracted image
    and a byte size of 0. My card is an Edimax EW7105PC+ with
    the Texas Instruments chipset as confirmed by the
    manufacturers. They have confirmed that TI will not
    supply drivers to them.

    Rick Phillips

  • Andreas Mohr
    Andreas Mohr

    • labels: 508429 -->
  • Andreas Mohr
    Andreas Mohr

    Logged In: YES

    I better move this over to Bugs.
    There seem to be some non-trivial issues with firmware
    loading remaining...

    BTW: why did you use relative firmware directory?
    It's recommended to use absolute directory...

  • Logged In: NO

    I did not use a relative firmware directory. When I issued
    the command I used the absolute address of
    firmware_dir=/home/rick/acx100/firmware. AFAIK, I cannot
    get any more absolute than that unless I have
    misunderstood your meaning. All commands were issued as
    root. It probably gave the message that it could not
    open "firmware/WLANGEN.BIN" because the created file had
    0 bytes.



  • Andreas Mohr
    Andreas Mohr

    Logged In: YES

    acx100_upload_fw() calls acx100_read_fw() with a filename that's
    created from the complete firmware_dir value and the
    firmware file name to load (WLANGEN.BIN).
    So in THIS LOGGED CASE you MUST have given the relative
    "firmware" as the directory, otherwise the full path would
    have shown up in the log.
    Maybe that was simply a log part that's different from some
    other modprobe that you did with absolute path instead.
    Uh, and yup, an absolute path is an absolute path, that's
    absolutely correct :-)

    Hmm, anyway, so the firmware does load now, right?
    (except for special init problems with Edimax, right?)

  • Logged In: NO

    No - the insmod does not work at all. I now suspect that this
    is a variant chipset which does not work with your driver -
    nor that supplied by Mandrake.

    You may not have any interest in pursuing this one any
    further but I will attach 4 files. acx100_pci.o created during
    the compile, acx100_pci.o.gz as supplied by Mandrake,
    TIACXSN.bin (Windows) as supplied by the manufacturer
    and dmesg.txt - a snip from the file of the same name
    showing debug info.

    This was also the response from the manufacturer -
    "Dear customer,
    I am sorry that TI doesn't release EW-7105PC+ driver for
    Linux to us, we don't have such driver as yet. But you can
    surf web site as below http://www.wlana.org to look for help.

    manufacturer: Texas Instrument CO.
    chipset model:ACX100AGHK"

    All this info may help or hinder. I think I will throw the
    bloody card away!!


  • Andreas Mohr
    Andreas Mohr

    Logged In: YES

    Hmm, I really don't know why you created 4 separate new bug
    reports instead of using the perfectly fine (it is, or isn't
    it?) attachment capability of this bug report. I guess I'll
    delete the ones that aren't too important (I don't know what
    to do with driver files anyway).

    Anyway, of course the insmod still failed. But what worked
    is the firmware loading, right? (hopefully)

    Then one should probably try to find out why the damn eCPU
    didn't bother to start up properly :-\

    Could you try to use a different firmware version from some
    other Windows driver?
    Or maybe there's "prior art" in the forum with an Edimax
    card loading fine, and with a "known good" firmware version?
    (I suspect your firmware binary might be broken or
    unsuitable for this particular card, causing the card CPU to
    hang on firmware init sequence)


  • Logged In: NO

    RTFM, are you really sure that you have followed
    instructions? Check and make sure that filename on your
    firmware folder is WLANGEN.BIN not WLANGEN.bin


  • Ivor Hewitt
    Ivor Hewitt

    Logged In: YES

    Is the file called "wlangen.bin" or "WLANGEN.BIN" ? name is case
    sensitive. Are you passing the firmware path in an insmod command?
    or in the modules.conf.
    if in modules.conf make sure you give a full pathname. if in insmod,
    are you giving a full relative path.

  • Andreas Mohr
    Andreas Mohr

    • status: open --> closed
  • Andreas Mohr
    Andreas Mohr

    Logged In: YES

    Most likely user error, user doesn't use driver any more,