> I'm not aware of any changes in the qeth driver that would have affected
> the HiperSockets initialization (but I'm no expert on the Linux
> implementation).  It sounds like you are keeping your Linux users
> up-to-date.  What level of CP are you running (and what Guest LAN
> maintenance level is reported by Query VMLAN)?

We're running z/VM 4.4 at service level 401. We're at VMLAN maintenance level
VM63397.


> Is it possible that your old configuration file did not specifiy "noauto" ?
> If that were the case, your old configuration might have given the channel
> device layer more freedome to "guess" which devices were appropriate
> network types.
>
> I would suggest an experiment.  Add the following line to
> /etc/chandev.conf:
>
> add_parms,0x10,0x770,0x772
>
> Shutdown and restart the guest... If that does not correct the problem,
> change it to:
>
> add_parms,0x10,0x770,0x772,portname:anything
>
> Shutdown and restart the guest... The portname "anything" can be just about
> any string up to 8 characters.  It only has to match other interfaces on
> the same NIC (which is just your interface in the case of a virtual NIC).
>
> Those are a couple of things that might affect the way Linux initializes
> the HiperSockets interface.  If this works, it will give us a better idea
> where things changed.  If it does NOT work, I can suggest a virtual I/O
> trace that we can use to figure out why the initialization failed.

The exclusion of "noauto" and the inclusion of the add_parms line (both with
and without a "portname" statement) seemed to have no effect.


Thanks for your help,

Michael Lambert
Louisiana State University
Office of Computing Services

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to