[tulip] Re: [tulip-bug] ADMtek Comet bug

Dan Hollis goemon@anime.net
Tue, 17 Oct 2000 22:18:13 -0700 (PDT)


On Wed, 18 Oct 2000, Donald Becker wrote:
> On Tue, 17 Oct 2000, Dan Hollis wrote:
> > > >  Port selection is 100mbps-SYM/PCS 100baseTx scrambler, half-duplex.
> > The claim of "half-duplex" is *definitely* wrong btw. Its full duplex for
> > sure.
> The Comet/Centaur are a special case -- the duplex is set and read only from
> the transceiver register.  It's a cleaner design than other Tulip chips, but
> it means that the standard CSR6 register report is inaccurate.

So should tulip-diag have special case code for admtek? Whats being
reported now by tulip-diag is very misleading.

> OK, we that's a little curious.  The chip's Tx FIFO underflowed enough times
> for the chip to switch to store-and-forward mode!  That's what the driver is
> supposed to do, but it indicates a PCI bandwidth problem.  What else are you
> running on the PCI bus?

Nothing. There are other cards in but they are simply inactive.

> > After a minute or two:
> > ADMtek AL985 Centaur-P chip registers at 0xe000:
> >   fff98000 ffffffff ffffffff 01165000 01165200 fc664010 ff97e117 ffffebff
> ...
> >   The transmit threshold is 1024.
> The driver has already increased the Tx threshold!  This isn't the problem,
> but it is suspicious.

It's also strange because this machine is doing almost NO traffic at all
when it adjusted the transmit threshold.

Oddly enough on another system which does not seem to exhibit the lockup
problem, it is left at 128 and never changes -- ever.

-Dan