More on the Orinoco + XI-825 + StrongARM

[Ben Greear]

Jean Tourrilhes wrote:


>>No, not to me, because the packet printed out looked perfectly valid
>>as it was handed to the rest of the kernel (netif_rx), and the built-in
>>ethernet driver (for the 10bt interface) worked just fine.  Considering
>>I only re-compiled the orinoco, that seems to imply that the orinoco was
>>wrong in some way...but again, the bytes seemed to be correct.
>>
>>Can you detect any invalid bytes in the messages I posted last night?
>>
>>Any other reason you can think of that the compile option fixed
>>the problem?
>>
> 
> 	You said it yourself : The only thing you did was to recompile
> the driver, and after this recompile you had troubles with only this
> driver. I just hope the same law of causality applies where you live.


No, after recompile, the driver started working!  Before, it would
fail on RX for most packets (everything other than ARP, it seems).
Now, it seems to work just fine...though I do see those TX errors.
(I see the same TX errors on an x86 laptop running a slightly older,
and completely un-mucked-by-me orinoco driver too, so it does not
appear to be ARM related.)

Now, it's obvious that the recompile changed something so that
it started working.  However, I would like to know what.  Since
the compile w/out the option was generating what appeared to be
correct packets, I cannot explain why the recompile 'fixed'
the problem.

By the way, if you'd like to update your /etc/pcmcia/config, this
adds support for the XI-825 CF NIC:

device "orinoco_cs"
   class "network" module "hermes", "orinoco", "orinoco_cs"


card "IEEE 802.11 Wireless LAN/CF Card"
   manfid 0xd601, 0x0005
   bind "orinoco_cs"

Thanks,
Ben

-- 
Ben Greear <greearb at candelatech.com>       <Ben_Greear AT excite.com>
President of Candela Technologies Inc      http://www.candelatech.com
ScryMUD:  http://scry.wanfear.com     http://scry.wanfear.com/~greear