Subject: Re: fxp full-duplex broken (and tlp too)
To: Jeff Rizzo <email@example.com>
From: Dave Sainty <firstname.lastname@example.org>
Date: 07/23/2000 14:02:17
Jeff Rizzo writes:
> On Fri, Jul 14, 2000 at 07:37:46PM -0500, Erich T. Enke wrote:
> > On Fri, 14 Jul 2000, Jason R Thorpe wrote:
> > > > Is this a known problem? Unfortunately, I won't be able to get boot
> > > > messages from the machine until either later today, or Monday. I
> > > > wanted to make sure I didn't forget the issue between now and then...
> > >
> > > Not that I know of. I've been using fxps in full-duplex w/
> > > 1.5_ALPHA constantly, just 10 minutes ago, even.
> > I assume this is an nsphy?
> Yes, I've just verified that. (We just brought the machine up on the
> Here's the boot messages:
> fxp0 at pci0 dev 13 function 0: Intel i82557 Ethernet, rev 1
> fxp0: interrupting at irq 11
> fxp0: Ethernet address 00:a0:c9:1c:88:d3, 10/100 Mb/s
> nsphy0 at fxp0 phy 1: DP83840 10/100 media interface, rev. 0
> nsphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto
I have some Tulip cards with nsphy's on them too, and they also
misbehave. I keep meaning to try and track it down, but haven't had
much time to do so.
I have found that this little hack helps a bit, it seems just looking
at the nsphy registers might help... I run this at boot time :)
# Problems with tlp0
echo -n "Kicking tulip NIC: disconnecting"
/sbin/ifconfig tlp0 media none
echo -n ", reconnecting"
/sbin/ifconfig tlp0 media autoselect
echo -n ", pausing"
echo -n ", probe"
/sbin/ifconfig tlp0 >/dev/null 2>&1
echo ", Done."
However, I'm still seeing significant numbers of errors, and I suspect
it's specific to the nsphy's...