Problem in vt6655 driver

Greg KH greg at kroah.com
Sat May 14 14:32:46 UTC 2011


On Sat, May 14, 2011 at 08:50:58AM +0200, Piotr Kaczorek wrote:
> W dniu 13.05.2011 21:06, Greg KH pisze:
> >On Fri, May 13, 2011 at 03:46:56PM +0200, Piotr Kaczorek wrote:
> >>I'm using vt6655 driver from 2.6.38 kernel for my wifi card. It's
> >>set up as access point (using modified script from VIA's native
> >>drivers - attached in package).
> >Right there you are probably alone in doing this.  Running the device as
> >an access point is something that I doubt the code really was ever
> >tested for at all.
> >
> >>I'd like to know if there is a possibility to run this card in
> >>monitor mode? It would be best if I could run AP and monitor modes
> >>at the same time (maybe on separate interfaces?).
> >>
> >>I also found a problem with sent packets. My setup is quite strange
> >>- my AP doesn't give access to the internet but substitutes any
> >>request with my page hosted on AP itself (kind of captive portal).
> >>There is a dnsmasq set to provide DHCP and DNS pointing to hotspot
> >>itself. All packets coming from eth1 are intercepted by iptables (-A
> >>PREROUTING -i eth1 -p tcp -j DNAT --to-destination
> >>192.168.100.1:8180). On port 8180 i have nginx serving website.
> >>The page is sent correct but images seem to break while being sent
> >>(on mobile phones i can't see them in the browser or can see only
> >>part of them, on windows however whole content is sent correctly).
> >>I attach packet dumps from three devices i'm testing it on -
> >>SonyEricsson P1i (Opera Mobile), BlackBerry Curve 8520 (their native
> >>browser) and laptop with Window VIsta installed (FF4.0). Each
> >>contains one request to some 'made up' url.
> >You should be happey that you are even getting this to work at all :)
> >
> >And I think you are really on your own here, any patches that you could
> >provide to resolve any of these issues would be greatly appreciated.
> >
> >good luck,
> >
> >greg k-h
> >
> I am really happy that it works, and thanks guys for doing a great work.
> Problem is that our customers won't be satisfied with 'thats the way
> it works' and i'm responsible for setting it all up :)

If you have "customers" for this type of thing, I would suggest you get
different hardware that you know works properly for Linux.  Right now,
this driver is not at the level of which I would feel comfortable having
to support it for anyone, so you really are on your own here.

Best of luck,

greg k-h



More information about the devel mailing list