On Fri, 29 Oct 2004, John A. Goebel wrote:

> Knowing a card works well is village knowledge.

Don't I know the scenario!  We've had this kind of problem for over a
decade with audio cards - the answer is always "Well, this old card
works great, but you can't buy them any more" (or "they changed the
chipset and the existing driver doesn't work any more) etc. etc.

Sometimes I think we should buy a shedload of cards that we don't need
yet - and can't use because there are no drivers yet - and put them on
the shelf until the linux driver for them appears!!!  And now the same
for wireless cards.

But this is exactly why the ndiswrapper approach seemed so appealing,
since the vendor will update their (Win32) driver to match changes to
the underlying hardware/chipset.

> I don't know how to solve the instability problem without
> researching what card works best or taking the time to debug the
> existing driver/card pair.

Oh, fully understood.  With the previous card I tried (Texas
Instruments chipset) I had two Win32 drivers and two versions of
ndiswrapper.  I only got one of the four combinations to work.  But
that's better than 0, at least.

Anyhow - coming back to this EW-7108PCg card, I can say that once it's
up and running, it seems to do a good job.  And it was inexpensive;
and there's also said to be a native linux driver, so it seems there
are at least two possibilities.

The hangs/crashes were all associated with trying to get it started -
in fact, chiefly with trying to run dhclient on it, it seemed.

Hope that's of interest.  But I still don't really understand the
interaction of the commands I'm issuing manually to it, and whatever
the hotplug routines are diddling about with behind the scenes.