Re: For the love of god!!! PLEASE somebody update 3com vortex config listing.

Michael H. Warfield (
Thu, 26 Jun 1997 10:38:17 -0400 (EDT)

Dave Cinege enscribed thusly:

> "Does linux support/Is there a driver for the 3c900 or 3c905. It's not
> listed as a compile option."

> I must see at least one of these posts *everyday* between the linux mailing
> lists and NGs. These are extremely popular cards!

They are extremely popular because 3Com has been extremely popular
in the past (it should be interesting to see if that continues after this
fiasco) and they have discontinued the 3c59x cards.

> PLEASE, somebody add: and 3c90x 'boomerang'
> To the 3c509 vortex kernel configuration description line. I would do it
> myself if I knew how, and where to submit it.

It's being worked on. David Miller ( if I can speak for Dave... Hi
Dave :-) ) has already stated an intention on including the boomerang driver
in a upcoming kernel (2.0.31 I hope!). The great debate (flame) is what
version. The most "stable" version at this time appears to be 0.30 with
some patches. But, AFAIK, it does not support the 100Mbit 3c905 but only
the 10Mbit 3c900. I may be wrong on that and people are welcome to roundly
flame me in private if I am. Version 0.40 had promise but proved to be
radioactive. Version 0.41 is out and is much much more stable but still
has a cable disconnect and 16 retrys failure problem. For me, 0.41 is at
least as stable as 0.30 as long as I don't disconnect the coax on a 10base2
line. Obviously, this does NOT impact the 3c905 100baseT configuration
but the 16 retry failure error could. This can be a problem on highly
congested networks but appears to NOT be a problem on switched ethernet
with full duplex (where collisions are theoretically impossible).

I have just seen a patch on the vortex mailing list (vortex was the
code name for the 3c59x cards while boomerang is the code name for their
successors, the 3c90x cards) which claims to now fix both the cable
disconnect and the 16 retries problem. The explanation seems plausable and
the patch looks reasonable. I'll be checking this out later today. If this
shoots these two dragons in the 0.41 driver, we need the subsequent 0.42
driver in the kernel ASAP.

It's really up to Donald Becker (the maintainer of the vortex driver)
and David Miller to call the shots on this one. I wouldn't be suprised (but
would be disappointed) if Dave has been put-off by the recent instabilities
in the driver and decided to delay its integration. I think we are on the
verge of having a stable, fully featured, driver at this point.

I'm adding the mailing list to
the Cc list of this message due to its relevance.

I too am anxiously awaiting the arrival of support for the 3c90x
cards in the stock kernel.

> The bandwidth you save could be your own!! : >

> ------------------------------------------------------------------
> Elite MicroComputers 908-541-4214


 Michael H. Warfield    |  (770) 985-6132   |
  (The Mad Wizard)      |  (770) 925-8248   |
  NIC whois:  MHW9      |  An optimist believes we live in the best of all
 PGP Key: 0xDF1DD471    |  possible worlds.  A pessimist is sure of it!