Re: errors on my interfaces

From: Donald Becker (becker@scyld.com)
Date: Sat Jan 05 2002 - 09:07:28 EST


On Fri, 4 Jan 2002, Madhav Diwan wrote:

> Its very odd but this error behavior is showing up on a rtl8139 card
> linked into a ciso switch.
> The switch itself reports frame checksum errors from the port into
> which the nic is plugged. The switch port is forced at 100 Full , the
> Realtec tries to Autonegotiate.

I write this many times a day:
      Don't use forced-full-duplex!

Anytime the speed or duplex is forced, autonegotation is disabled. Read
  http://scyld.com/expert/NWay.html

It's almost always Cisco switch administrators that misconfigure their
network. It's based on a Cisco recommendation initially made when Cisco
produced a bunch of hardware with broken autonegotiation. Rather than
replace the flawed hardware, Cisco claimed that using autonegotiation
was the problem and recommended forcing the speed and duplex.

When using forced full duplex every network device must manually
configured before installation. And when it's moved to a
standards-conforming network you must remember to change the setting
back to autonegotiate. You see, that's why you need a professional
network administrator -- only the guru knows the magic and is allowed to
touch the machines. And only the system operator is allowed to put your
punched card deck in the hopper.

> I am going to try forcing the media type to see what happens.

You can use a module option, or better
   mii-diag eth0 --force 100baseTx-FDX

> by the way .. i'd like to report some odd behavior from a rtl8139 card
> i have .. I am useing the rtl8139-diag program from your site to
> adjust the media type . It seems that if i force the media type i have
> to discontinue power to the nic card before the new register entry takes
> effect ... No biggie .. but its anoying that i actually have to turn the
> boxes power off before the new setting takes effect.

The *-diag programs modify the persistent EEPROM contents. The changed
configuration won't take effect until the chip is reset in a way that
causes the EEPROM to be re-read. This is usually a hardware reset,
although some drivers read the EEPROM every time the module is loaded.

> (sometimes this occurs natuarlly after a reboot command .. ie i see the
> light onb the hub go out for a second and then come back. then i know it
> took without me having to turn the power off)

Donald Becker becker@scyld.com
Scyld Computing Corporation http://www.scyld.com
410 Severn Ave. Suite 210 Second Generation Beowulf Clusters
Annapolis MD 21403 410-990-9993

-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html



This archive was generated by hypermail 2b29 : Mon Jan 07 2002 - 21:00:40 EST