On Thu, 13 Jul 2000, Jeff V. Merkey wrote:
>
...
> Yeah. What's going to happen when we are like Windows 2000 and have
> 20,000+ device drivers? Wait four years between each release of Linux
We are unlikely to reach that number anytime soon. Not because we will support
significantly fewer devices, but because your average Linux driver does support
a wide range of devices using similar/same chipsets.
There's *ONE* driver for every 3Com card I use, from my laptop to the office
machines. ONE driver. For 10-30 (didn't count) different cards.
And this is not just the 3Com driver, that was just an example. Pick just
about any driver that's not a driver for some very special kind of device. All
drivers for ``common'' hardware that I have ever seen, support a very large
number of different (but similar) devices, and those devices do require
different drivers on that other system you mention.
You almost have a point in what you're saying, but the _number_ of drivers
we actually need is at least an order of magnitude lower than on that other
platform.
> (and buy Linus 200 lbs. of Methamphetamine a month so he can stay awake
> around the clock and fix them all?). I guess poor Alan will have to
> give up his weekends and vacation for the next decade.
>
> People are motivated to update drivers because they release a new board
> or hardware device and want to SELL IT FOR $$$$ (which is probably an
> alien concept for someone at a non-profit organization like gnu.org --
> no offense). If their drivers are broken NOONE WILL BUY THEIR BUSTED
FSF does hire people, and I've heard rumers that they both eat and need to
pay for food as well.
Ever wondered why the Deluxe Distribution will cost you $5K ? :)
> BOARDS. THis is exactly how we enticed folks to upgrade drivers in
> NetWare. We'd bust them then tell our customers to buy some other
> vendors card that worked. You'd be amazed how fast these companies will
> their hot little asses in gear and fix stuff when their packetbooks are
> being affected.
Drivers are being updated in the stable series all the time. No problem,
because those are isolated patches.
The problem is interface changes that require an all-driver change. If such
changes are made a few times every two years in a development series, I fail to
see the problem. No such change is made for the fun of it.
Anyone who submits a patch that breaks the vast majority of drivers is going to
have a damn good reason, and if it is indeed such a damn good reason, I fail
to see the problem. Unlike others, we actually have the source for the
drivers and they can actually be fixed for real, not hooked up to some middle
layer that somewhat fakes the old semantics. Think ``softnet'' then think
``middle layer for old network drivers''. :)
--
................................................................
: jakob@ostenfeld.dtu.dk : And I see the elder races, :
:.........................: putrid forms of man :
: Jakob Østergaard : See him rise and claim the earth, :
: OZ9ABN : his downfall is at hand. :
:.........................:............{Konkhra}...............:
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Jul 15 2000 - 21:00:18 EST