More on the Orinoco + XI-825 + StrongARM platform...

[Ben Greear]

I just tried the 'testing' release of the orinoco_cs
driver...it works much better (I can see pings going
to my laptop, and being responded to.)

However, I never see the ping response on the cube's cmd line (I do
see the rx packet counters bumping correctly, though).

Arp is working though, for whatever reason...

Ben Greear wrote:

> 
> 
> Martijn van Oosterhout wrote:
> 
>> On Tue, Feb 19, 2002 at 09:51:43PM -0700, Ben Greear wrote:
>>
>>> The interesting thing is that when I ping between the other machines, I
>>> see packets received on the cube.  When I ping from the cube to the 
>>> laptop,
>>> while sniffing the wireless port on the laptop with tcpdump, I see this:
>>>
>>>
>>> [time-stamp] 0:60:b3:69:56:67 > Broadcast null I (s=85,r=85,C) len=50
>>> (more of the same)
>>>
>>
>> Was there not something about structure padding that caused some 
>> structures
>> to be padded differently on Arms than on other architectures, causing
>> transmitted packets to be unreadable to anyone else?
>>
>> I beleive it was fixed in a more recent version, the one shipped with the
>> kernel source is pretty old now.
> 
> 
> 
> I'll check that out in just a second (or maybe tomorrow :)
> 
> I put in some prints deep in the driver in the orinoco_xmit method:
> 
> 
>                if (1) {
>                   int i = 0;
>                   const unsigned char* dbg = (const unsigned char*)(&hdr);
>                   for (i = 0; i<(sizeof(hdr)); i++) {
>                      printk("%02hx ", (unsigned short)(dbg[i]));
>                   }
>                   printk("\n");
>                }
>         err  = hermes_bap_pwrite(hw, USER_BAP, &hdr, sizeof(hdr),
>                      txfid, 0);
> 
> 
> I see this printed out...it should be an arp from what I can tell:
> 
> Feb 20 06:46:46 CerfLinux kern.warn eh->h_proto: 0x608  ntohs: 0x806
> Feb 20 06:46:46 CerfLinux kern.warn Ethernet-II frame, len: 60, 
> sizeof(hdr)...
> Feb 20 06:46:46 CerfLinux kern.warn 00 00 00 00 00 00 00 00 00 00 00 00 
> 06 00 08 00 00 00 ff ff ff ff ff ff 00 60 b3 69 56 67 00 00 00
> 00 00 00 00 00 00 00 00 00 00 00 44 00 ff ff ff ff ff ff 00 60 b3 69 56 
> 67 00 44 00 00 aa aa 03 00 00 00 08 06
> 
> It really doesn't look too correct to me!
> 
> 
> I also receive this packet every now and then...not sure what it's 
> supposed to
> be.  This is for code that prints out 'hdr' in __orinoco_ev_rx:
> 
> Feb 20 06:48:46 CerfLinux kern.warn ERROR: rx pkt failed for some 
> reason, hdr: (len: 70)
> Feb 20 06:48:46 CerfLinux kern.warn 05 2e 50 31 6a 71 13 36 14 00 03 00 
> 00 00 a6 73 52 b8 d8 73 7a 70 31 3c 58 49 c4 44 c6 3a 52 9d 14
> 6d 89 44 7c e0 28 e0 a2 96 78 ee 00 00 af a0 90 69 ad 40 fd 40 b0 be 6a 
> 05 00 00 06 09 04 00 12 02 22 02 12 02
> 
> 
> Thanks for the link,
> Ben
> 
> 
>>
>> http://ozlabs.org/people/dgibson/dldwd/
>>
>> HTH,
>>
> 
> 


-- 
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