DWL 520 E1 (the continuing saga)

Pavel Roskin proski at gnu.org
Wed Jan 28 20:04:15 EST 2004


On Wed, 28 Jan 2004 cliffo at u.washington.edu wrote:

> I'm using the latest version from CVS (downloaded today) and using
> prism2_srec included in that update.  prism2_srec is failing on the
> second firmware update for the card and I'm perplexed as to my next
> step.

I'm fighting with this problem as well, except that I'm using a PCMCIA
card, DWL-650 revision P.  See this thread:
http://lists.shmoo.com/pipermail/hostap/2004-January/thread.html#5256

> [~]# ./utils/prism2_srec -gs wlan0 ~cliff/firmware/pm010102.hex
> <from dmesg>
> prism2_download: dl_cmd=4 start_addr=0x003f0000 num_areas=3
> wifi0: test Genesis mode with HCR 0x0f
> wifi0: Original COR value: 0x0
> Readback test succeeded, HCR 0x0f
> prism2_hw_init: initialized in 0 ms
> wifi0: valid PDA not found
> wifi0: NIC: id=0x8000 v0.0.1
> wifi0: PRI: id=0x15 v1.1.2
> Could not get RID for component STA
> wifi0: Failed to read STA f/w version - only Primary f/w present

You are a bit luckier.  You don't have "init command completed too
quickly" as I have with TI bridges, but you still have "valid PDA not
found".

The rest is obvious.  The second load is done so that the PDA is used.
If it's not available, you cannot go further.

What happens is that the primary firmware is loaded successfully.  If runs
after that and allows to run simple commands such initialization and
requests for version numbers.  It even returns its own version number
correctly.  However, it fails to execute the command to open AUX port,
which is needed to access the PDA.

In fact, I tested the same card in two machines.  One is AthlonXP 2000
with VIA KT133 motherboard.  The other is Pentium 800MHz with Intel 810
chipset.  The kernel, the driver, the card and even the PCMCIA-to-PCI
bridge are the same.  However, the card only works on the Athlon machine.

I tried to add delays in different places but it had no effect.  Maybe the
driver should initialize some register that is set randomly dependent on
the bus speed.

-- 
Regards,
Pavel Roskin



More information about the HostAP mailing list