adm8211 (from linville wireless-2.6) in xen guest

adm8211 (from linville wireless-2.6) in xen guest

[Jan Evert van Grootheest]

Hi,

Just writing to thank all of you that made the adm8211 driver in John 
Linvilles wireless-2.6 development tree (not the dscape one, which I 
will try also).

It is this card:
00:01.0 Network controller: ADMtek ADM8211 802.11b Wireless Interface 
(rev 11)
        Subsystem: Standard Microsystems Corp [SMC] Unknown device 2602
        Flags: bus master, medium devsel, latency 96, IRQ 16
        I/O ports at d800 [size=256]
        Memory at cfffe400 (32-bit, non-prefetchable) [size=1K]
        Expansion ROM at <ignored> [disabled]
        Capabilities: [c0] Power Management version 2

which is reported by the driver as
Jan 16 22:05:40 janus kernel: adm8211: release 20060414
Jan 16 22:05:40 janus kernel: PCI: Enabling device 0000:00:01.0 (0000 -> 
0003)
Jan 16 22:05:40 janus kernel: 0000:00:01.0 (adm8211): EEPROM type 93C46
Jan 16 22:05:40 janus kernel: 0000:00:01.0 (adm8211): Channel range: 1 - 13
Jan 16 22:05:40 janus kernel: 0000:00:01.0 (adm8211): RFtype=1 BBPtype=1 
Specific BBP=0 Transceiver=0
Jan 16 22:05:40 janus kernel: eth0: hwaddr 00:04:e2:5e:b3:08, IRQ 16, 
Rev 0x11

I've taken two commits from that tree (the initial one and one about 
statistics) to make a driver that would compile against debian etch 
2.6.18 and moved the Makefile to be Kbuild. That allowed me to compile 
it as an external module (with something like 'make -C 
/lib/modules/2.6.18-xxxxxx/build M=/usr/src/wireless-adm8211 
CONFIG_ADM8211=m').
 
So I now have an AMD64 with xen on it. The adm8211 card is used in one 
of the guest domains!

One question, though. I'm trying to use it in ad-hoc mode. If there's no 
other station with the same essid, it just keeps mentioning that 
(besides the BCNTC and TSFTF messages)

wlan0: No matching adhoc sta found. Creating IBSS 00:04:e2:3f:2a:70 CHAN=11

And the IBSS changes every time, which is about every 30 seconds. Is it 
really supposed to change the IBSS each time?
The reason I ask is that my wife has a laptop with an Atheros chip in 
it. If it is configured for ad-hoc it creates the IBSS and sticks with 
it. So I now have two computers that have different behaviour in this 
respect. That in itself is not, I guess, a problem. But it seems to me 
that the random IBSS created by the adm8211 would result in two stations 
not seeing eachother or taking longer than necessary to find eachother?

And one more question that I don't understand... the wlan0 interface 
does not appear in snmp queries. Any idea about that?

Please CC me; I'm not subscribed.

Thanks,
Jan Evert

Re: adm8211 (from linville wireless-2.6) in xen guest

[Michael Wu]

[-- Attachment #1: Type: text/plain, Size: 1550 bytes --]

On Saturday 20 January 2007 07:19, Jan Evert van Grootheest wrote:
> Hi,
>
> Just writing to thank all of you that made the adm8211 driver in John
> Linvilles wireless-2.6 development tree (not the dscape one, which I
> will try also).
>
CC me then. :) Note that I only maintain the d80211 based adm8211 driver now, 
but that driver doesn't have adhoc support yet.

> One question, though. I'm trying to use it in ad-hoc mode. If there's no
> other station with the same essid, it just keeps mentioning that
> (besides the BCNTC and TSFTF messages)
>
> wlan0: No matching adhoc sta found. Creating IBSS 00:04:e2:3f:2a:70 CHAN=11
>
> And the IBSS changes every time, which is about every 30 seconds. Is it
> really supposed to change the IBSS each time?
Probably not, but I never debugged adhoc mode enough.

> The reason I ask is that my wife has a laptop with an Atheros chip in
> it. If it is configured for ad-hoc it creates the IBSS and sticks with
> it. So I now have two computers that have different behaviour in this
> respect. That in itself is not, I guess, a problem. But it seems to me
> that the random IBSS created by the adm8211 would result in two stations
> not seeing eachother or taking longer than necessary to find eachother?
>
Have you checked? I think the bssid should stop changing once another station 
with the same ssid comes into range.

> And one more question that I don't understand... the wlan0 interface
> does not appear in snmp queries. Any idea about that?
>
Nope.

-Michael Wu

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: adm8211 (from linville wireless-2.6) in xen guest

[Jan Evert van Grootheest]

Michael,

Michael Wu wrote:
> On Saturday 20 January 2007 07:19, Jan Evert van Grootheest wrote:
>   
>> One question, though. I'm trying to use it in ad-hoc mode. If there's no
>> other station with the same essid, it just keeps mentioning that
>> (besides the BCNTC and TSFTF messages)
>>
>> wlan0: No matching adhoc sta found. Creating IBSS 00:04:e2:3f:2a:70 CHAN=11
>>
>> And the IBSS changes every time, which is about every 30 seconds. Is it
>> really supposed to change the IBSS each time?
>>     
> Probably not, but I never debugged adhoc mode enough.
>   
If there're some tests I can help with, just ask.
I've decided yesterday that I really should get myself a WRT54G or 
somesuch. So there's not much pressure for this adm8211 anymore.
>> The reason I ask is that my wife has a laptop with an Atheros chip in
>> it. If it is configured for ad-hoc it creates the IBSS and sticks with
>> it. So I now have two computers that have different behaviour in this
>> respect. That in itself is not, I guess, a problem. But it seems to me
>> that the random IBSS created by the adm8211 would result in two stations
>> not seeing eachother or taking longer than necessary to find eachother?
>>
>>     
> Have you checked? I think the bssid should stop changing once another station 
> with the same ssid comes into range.
>   
Well, it took more than 1 minute and I left for coffee. When I returned 
they had found eachother.

Do you perhaps know whether there's a sourcetree somewhere of the dscape 
code that can be compiled as external modules for 2.6.18?
At least I can help test the dscape code, which I understand is the 
current future of wireless in linux.

Thanks,
Jan Evert

Re: adm8211 (from linville wireless-2.6) in xen guest

[Dan Williams]

On Mon, 2007-01-22 at 14:30 -0500, Michael Wu wrote:
> On Saturday 20 January 2007 07:19, Jan Evert van Grootheest wrote:
> > Hi,
> >
> > Just writing to thank all of you that made the adm8211 driver in John
> > Linvilles wireless-2.6 development tree (not the dscape one, which I
> > will try also).
> >
> CC me then. :) Note that I only maintain the d80211 based adm8211 driver now, 
> but that driver doesn't have adhoc support yet.
> 
> > One question, though. I'm trying to use it in ad-hoc mode. If there's no
> > other station with the same essid, it just keeps mentioning that
> > (besides the BCNTC and TSFTF messages)
> >
> > wlan0: No matching adhoc sta found. Creating IBSS 00:04:e2:3f:2a:70 CHAN=11
> >
> > And the IBSS changes every time, which is about every 30 seconds. Is it
> > really supposed to change the IBSS each time?
> Probably not, but I never debugged adhoc mode enough.
> 
> > The reason I ask is that my wife has a laptop with an Atheros chip in
> > it. If it is configured for ad-hoc it creates the IBSS and sticks with
> > it. So I now have two computers that have different behaviour in this
> > respect. That in itself is not, I guess, a problem. But it seems to me
> > that the random IBSS created by the adm8211 would result in two stations
> > not seeing eachother or taking longer than necessary to find eachother?
> >
> Have you checked? I think the bssid should stop changing once another station 
> with the same ssid comes into range.

Hmm; that's odd.  If the card is set to adhoc mode and it cannot find
another station, then it should just start up as the only station in the
adhoc network.  It ideally shouldn't be searching for another one since
of course you don't really need more than one STA to form an adhoc
network.

In any case, concentration should go into the d80211 version of
course :)

Dan

> > And one more question that I don't understand... the wlan0 interface
> > does not appear in snmp queries. Any idea about that?
> >
> Nope.
> 
> -Michael Wu