Link up/down messages from 5200 ethernet driver

Grant Likely grant.likely at secretlab.ca
Thu Dec 6 17:08:00 EST 2007


On 12/5/07, Jon Smirl <jonsmirl at gmail.com> wrote:
> On 12/6/07, Jon Smirl <jonsmirl at gmail.com> wrote:
> > On 12/5/07, Grant Likely <grant.likely at secretlab.ca> wrote:
> > > On 12/5/07, Jon Smirl <jonsmirl at gmail.com> wrote:
> > > > Why does the fec5200 driver sometimes not print 'PHY' correctly?
> > > >
> > > > phyCORE-MPC5200B-tiny login: TؤV� f0003000:10 - Link is Down
> > > > T�Y: f0003000:10 - Link is Up - 100/Full
> > > > T�Y: f0003000:10 - Link is Down
> > > > PHY: f0003000:10 - Link is Up - 100/Full
> > >
> > > Weird; that line comes from drivers/net/phy/phy.c in
> > > phy_print_status(), and it's just a a pr_info statement.  I suppose it
> > > might be a stack overrun, but I'd expect more severe symptoms if that
> > > was the case.
> > >
> > > That *looks* like a serial stream corruption.  Is that the output of
> > > your serial port?  Does dmesg show the same corruption?
> >
> > That's the serial port output.
> > dmesg is ok.
> >
> > Nothing else I do can mess up the serial port.
>
> Maybe the KERN_INFO prefix is sending out control characters that mess
> up the serial output?

More likely the Ethernet link state change is interfering with the
serial line at the electrical level.  Try tweaking the serial clock up
or down a tiny bit, that might make it more stable.  (I've had to do
that on one of my boards here to get stable serial output).

Cheers,
g.


-- 
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.
grant.likely at secretlab.ca
(403) 399-0195


More information about the Linuxppc-dev mailing list