wi0 and adhoc

[Rick van der Zwet]

2011/5/11 Lystopad Olexandr <laa at laa.zp.ua>:
> I have two freebsd boxes with 4.11 (yes, really!). They both
> serve wifi adhoc link (wi0 + pcic0: <Vadem 469>) during many
> years with no problems. One box has died, and I change it to
> new one with 8.0-rel. New box have cbb0 and same wi0 card. So,
> now I have two boxes 4.11 and 8.0. Both with wi0 devices detected.

[snip: ifconfig]

> # ifconfig wlan0 list caps
> drivercaps=10301<STA,IBSS,PMGT,MONITOR>
> cryptocaps=1<WEP>
>
> pciconf -lvc:
> ...
> cbb0 at pci0:2:9:0:        class=0x060700 card=0x14101524 chip=0x14101524 rev=0x01 hdr=0x02
>    vendor     = 'ENE Technology Inc'
>    device     = 'CardBus Controller (CB-1420)'
>    class      = bridge
>    subclass   = PCI-CardBus
>    cap 01[a0] = powerspec 1  supports D0 D1 D2 D3  current D0
> ...
>
> dmesg:
> wi0: <Lucent Technologies WaveLAN/IEEE> at port 0x100-0x13f irq 17 function 0 config 1 on pccard0
> wi0: [ITHREAD]
> wi0: device timeout
> wi0: device timeout
> wi0: record read mismatch, rid=fd44, got=fc80
> wi0: device timeout
> ....

Interesting it does not seems to be able to fetch the firmware
information of the card. As it would normally list something like:

wi0: <Intersil Prism2.5> mem 0xa0000000-0xa0000fff irq 10 at device 16.0 on pci0
wi0: [ITHREAD]
wi0: using RF:PRISM2.5 MAC:ISL3874A(Mini-PCI)
wi0: Intersil Firmware: Primary (1.1.1), Station (1.8.2)
wi0: Ethernet address: 00:02:6f:4a:cd:f5

> my questions are:
> 1. why txpower 0 in 8.0 box?
> 2. why too many ``device timeout'' messages in dmesg on 8.0 box?
> 3. why all pings lost in that link?

Usually the firmware on the card is this biggest issue, causing the
card not getting detected, generating all the device timeout messages.
Running new firmware seems to fix it for a while. ```man 4 wi''' lists
the minimum required versions. Some firmware upgrade hints are listed
over here:
    http://www.wirelessleiden.nl/projects/nodefactory/ticket/23

We had 200+ wi(4) cards running in our Wireless Metropolitan Network
and where running into the same walls as you did, every new release
issued more problems with those wi(4) cards. By the end we desired to
replace them all by ath(4) based cards (active development on the
driver :-)).

Br. /Rick
-- 
http://wirelessleiden.nl