Re: NE2000 driver

Bob Tracy - TDS (rct@merkin.csap.af.mil)
Mon, 4 Aug 1997 08:39:43 -0500 (CDT)


Alan Cox wrote:
> Bob Tracy wrote:
> > I would recommend enabling that block of code by default unless there'
> > some compelling reason why that would be a bad idea.
>
> Well one perfectly good reason is that it slows every other card down.

I would buy the above if you had said "every other NEx000 card".

> Another
> is that we don't know if it will upset some other clones.

The evidence I've seen so far suggests that my experience is the rule
rather than the exception. I'd love to be proven wrong on this. What
say you people that have *real* NE2000 cards? Am I fighting "bad" clone
hardware, or is the NE2000 simply a bad design to emulate?

> Linux can't spend
> its entire time bandaging crap vendor hardware. At best it'll simply
> make the vendor shift more and more complete junk.

Although I agree in principle with these sentiments, the stance being
promoted isn't terribly pragmatic. Wintel hardware ain't popular because
it's good -- it's popular because it's *cheap*. We can protect ourselves
somewhat by buying from vendors that seem to be aware of and responsive
to the Linux user community, e.g., Mylex/Buslogic, but the majority of
the hardware manufacturers are not and will not be affected one way or
the other by our decision to not support their shoddy merchandise. Even
at its best, Wintel hardware is hardly a paragon of hardware design virtue.

About that shoddy merchandise anyway... The clones I've seen the problem
with ALL bear the Novell certification label. Isn't the whole purpose of
companies like Microsoft and Novell certifying stuff to protect us from
the worst of the bad hardware manufacturers? Too often I hear us saying
"good enough for DOS/Windows doesn't mean good enough for Linux", but my
position is that we take that viewpoint and shove it into the nearest
Linux advocacy group access point. Comments sprinkled liberally
throughout the Linux kernel source attest to the behind-the-scenes
hair-pulling and head-scratching going on, so my contention is that the
precedent for coping with "junk" hardware has been established. Besides,
one of the MAJOR good things that Linux has going for it is how flexible
we are in accommodating the incredible variety of PeeCee hardware
available. I agree that we shouldn't support hardware that barely
functions under *any* OS, e.g., the 3c501 and the 80286, but if it
functions well under DOS/Windows, I don't see that the "broken hardware"
charge reflects well on the Linux community. It might be defensible to
the technically savvy, but doesn't fly when presented to the average
PeeCee user.

-- 
--------Bob Tracy       AFIWC/AFCERT    rct@merkin.csap.af.mil--------
"If you make any money, the government shoves you in the creek once a
year with it in your pockets, and all that don't get wet you can keep."
                                -- Will Rogers