[Click] RX Overruns with eepro100

[Barry Gowan]

Hello, 
Can someone please shed some light on how interface
overruns can happen when using click with PollDevice. 

The little bit of information that I can find about
interface overruns is in the context of interrupts.
Since PollDevice disables interrupts on the device,
this seems like the only place I can ask about this!

I have an eepro100 4 port copper card on each of 2
boxes. The ports are connected back to back with 4
crossover cables.  The Click configuration has 4
ToDevice and 4 PollDevice elements, one for each
eepro100 interface. 

If I use this configuration to send, say 5000 packets
per second per interface, 1500 byte packets, both
direction, what happens is this:
The TX interface stats on both machines show 5000pps
ok.
The RX intrerface stats for all interfaces on both
machines show only 3000-4000 pps, and the RX Overruns
consequently increase by ~1000 per second. 

Nothing I do with the Click configuration (including
simplification and BURST parameters) seems to make any
difference to the amount of overruns happening. 

Is this simply a matter of the eepro100 device running
out of available memory or processing speed or is
there something I can do about it?

I would be grateful if someone would explain what is
happening between PollDevice, the eepro100 driver, the
PCI bus, and the card itself in this situation. Then
maybe I could figure out what I can do to stop the
overruns. 

Also with smaller packet sizes, e.g. 64 bytes, I can
send 30,000 packets per second with the same setup. 

BG



________________________________________________________________________
Want to chat instantly with your online friends?  Get the FREE Yahoo!
Messenger http://uk.messenger.yahoo.com/