[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] memory fault

> " It's a message from Xen. It only gets printed if the network device
> " driver's transmit function returned an error. This should never happen
> " (if the driver isn't ready to accept packets there is a way to signal
> " that).
> "   
> " Maybe it's a driver or hardware bug. Maybe Xen is running low on
> " memory (unlikely).
> I am seeing xen crash again with the 'Weird' log message.
> This time it is xen-1.3-devel using an eepro100 NIC:
>     (XEN) Weird failure in hard_start_xmit!

Couple of questions that I think might be a big help to us in
tracking this down:

 Can you repeat with hardware other than eepro100? (Our test
machines have tg3 and e1000 cards).

 Can you reproduce reliably using the ttcp test program? 

 What about in UDP mode "ttcp -s -u -fm", or does it require data
to be flowing in both directions to trigger it? (I presume it's
the transmitter that blows up?)

 Does reducing the packet size (and hence taking the packet rate
up) cause the bug to trigger more frequently?

 Can you cause the bug to trigger with just one domain running on
each machine?

 Can you cause the bug to trigger with the machine booted "nosmp"?

Of course, the new IO stuff should be ready soon and the bug
might just disappear ;-)


This SF.net email is sponsored by: The Robotic Monkeys at ThinkGeek
For a limited time only, get FREE Ground shipping on all orders of $35
or more. Hurry up and shop folks, this offer expires April 30th!
Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.