Subject: Re: wip/kismet vs. wi(4): kernel eats CPU
To: Thomas Klausner <wiz@NetBSD.org>
From: pancake <pancake@phreaker.net>
List: tech-net
Date: 06/11/2006 18:10:47
I had some similar problems writing the wireless stumbler
(http://news.nopcode.org/wist3.png), i'll take a look this week about
if this is a driver problem or what.

--pancake

On Sun, 11 Jun 2006 17:51:08 +0200
Thomas Klausner <wiz@NetBSD.org> wrote:

> Hi!
> 
> I recently tried the wip/kismet package with a wi(4) card:
> 
> wi0 at pci0 dev 10 function 0: Intersil PRISM2.5 Mini-PCI WLAN (rev. 0x01)
> wi0: interrupting at irq 11
> wi0: 802.11 address 00:09:5b:2f:0e:76
> wi0: using RF:PRISM2.5 MAC:ISL3874A(Mini-PCI)
> wi0: Intersil Firmware: Primary (1.0.7), Station (1.3.6)
> wi0: 11b rates: 1Mbps 2Mbps 5.5Mbps 11Mbps
> 
> AIUI, kismet does channelhopping to find all available networks.
> On my Athlon XP 2400 with NetBSD-3.99.18, this made the machine
> nearly completely unresponsible since it seemed to spend quite a
> lot of time in the kernel (I assume the wi(4) driver).
> 
> I was a bit surprised, since I thought that an Athlon XP 2400 would
> have quite enough CPU power for this task. I guess that the wi(4)
> driver or hardware is not well-suited for the task.
> 
> Did anyone else see this?
> 
> Is this a bug in the driver, or is it a problem with the hardware?
> 
> If it's the latter, what do you recommend as wireless card under NetBSD?
> 
> Cheers,
>  Thomas