On 05/18/10 12:32 AM, Al wrote:
Thanks a lot Garrett!! Your response is very helpful. I have some follow-up
questions that hopefully you and others may be able to help me with.
I was hoping for something that would work in general including on Solaris 10 /
bge (not sure if the any of the drivers with PHY disabling capabilities is part
of Solaris 10).
It's possible that something less severe than completely turning off the PHY
will still do the job (I doubt turning just the MAC off will). I just want to
be able to make one side detect that a link is down based on a software action
on the other side of the link.
For example if I just connect 2 Solaris machines with a crossover cable (no
switches or anything in between), will it be possible for one machine to
manipulate the link state in a way that the other machine will report (with
ifconfig or kstat) that the link is down?
The only way I know of to achieve this is to either power down the PHY,
or put it in some kind of loopback state. (And I'm not positive that
loopback will have the desired effect on the remote peer.)
Is it possible to use ndd to advertise a faulty or offline link state using the
adv_rem_fault parameter?
Not generally, no. That parameter should be read-only.
(Fault value reported by the local system to the peer) or perhaps try to force
a speed mismatch on the link (also using ndd) hoping this will lead the peer to
detect the link is not usable.
Certainly, forcing incompatible speed settings would have that effect.
Not sure why I didn't think of it. So force one peer to 1G, and the
other 100 M, and link should not be established.
Taking bge driver as an example, I was not able to confirm whether it supports
adv_rem_fault.
I don't think it does. adv_rem_fault is *supposed* to be something set
automatically, not something that gets manually tweaked.
Thanks,
Al.
-- Garrett
_______________________________________________
networking-discuss mailing list
[email protected]
