On Wed, 7 Jan 2009, Hin-Tak Leung wrote:
> --- On Wed, 7/1/09, jsassojr@... <jsassojr@...> wrote:
> > Thank you for your response! To answer your question, I know
> > the kernel
> > already supports the card via the iwl3945 driver, but I
> > have been having a
> > lot of trouble trying to get it configured to my
> > university's WLAN with the
> > specs: WPA w/ TKIP, EAP type of PEAP using MSCHAP v2 for
> > authentication,
> > and "IPS SERVIDORES" for the Trusted Root
> > Certificate. I went through
> > several configurations, as per documentation for
> > wpa_supplicant, turning on
> > debug mode, but I have had no luck. (IMHO, having upgraded
> > from Ubuntu Edgy
> > to Intrepid Ibex, I have found the wireless support to be
> > worse)
> > So, I heard that people were having more luck using
> > ndiswrapper, so I
> > thought I'd pursue this route instead of banging my
> > head against the wall
> > for days trying to resolve a problem I could possibly get
> > resolved w/
> > ndiswrappers.
> > --john
> But you would still be using wpa_supplicant for WPA, even if you are using ndiswrapper? So that part of the configuration should be similiar if not the same... maybe once you get ndiswrapper working, you can switch back.
> if you find the as-shipped in-kernel native driver less than satisfactory,
> maybe you could try compat-wireless (still at http://linuxwireless.org/) -
> a package for using the latest driver code against a stable kernel.
> good luck.
This is getting off-topic, but it might explain some of the interest in
ndiswrapper for a wireless NIC that is supposedly "fully supported"
natively in the kernel...
This has been a sore spot with me. I have a laptop with this wireless
3945 NIC in it. When I first got the laptop I set up Intel's ipw3945
Linux driver and it worked flawlessly AND fast. But Intel has pushed
everyone over to the newer iwl3945 driver instead and deprecated the old
one. Starting around kernel 2.6.24, the old driver no longer compiles.
Compared to ipw3945, the iwl3945 is far slower and exhibits far worse
wireless range, to put it kindly. This has been going on now for about
a year, and if you search Google, there's a lot of chatter about this,
but apparently no solution. Reading between the lines, the iwl3945
driver moves the regulatory functions from a closed-source user daemon
(which the ipw3945 driver relied on) into the wireless chip's processor
- and I wonder that if in doing so they overburdened the chip.
(Specifically, running WPA-PSK and talking to a Linksys WRT-54G WAP in
mixed mode *really* makes performance the problem apparent.) The
performance is so lousy for me that I too have been considering trying
out ndiswrapper, though for the moment I'm on a newer laptop with the
next generation Intel wirless NIC and it seems to perform better with
the iwl3945 driver.
A while back (last Spring?) I made several attempts at moving to the
compat-wirelss package you mentioned. It was an utter failure. While
it compiled and loaded, the results were buggy and crash-prone, i.e.
Yes, I know. I could have gotten more involved in the community and
tried to fix these things myself. But from what I gathered searching
the web, the overall impression I got is that this older 3945 part
simply isn't being supported very well anymore. And I already have my
hands full contributing to other aspects of the Linux community.
There's only so many hours in the day and I'm simply out of time. It
was easier for me to switch computers :-(
I have used ndiswrapper a lot in the past. It's a wonderfully clever
bit of engineering. It saved my butt back in the day before Broadcomm
support showed up in the kernel. And I might add that these days the
reverse-engineered Linux native (b43, IIRC) Broadcomm driver I use on
another laptop seems to work BETTER than Intel's directly supported
iwl3945 driver for my 3945-installed laptop...
isely @ pobox (dot) com
PGP: 03 54 43 4D 75 E5 CC 92 71 16 01 E2 B5 F5 C1 E8