eth0: Too much work at interrupt, status=0x4050

"Joe Teixeira" <[email protected]> Wed, 30 Jul 2003 09:49:02 -0400
Newsgroups gmane.linux.drivers.eepro100.devel,gmane.linux.drivers.eepro100.user
Message-ID <[email protected]>
Hi,  I am using the eepro100 driver (version 1.09j) with an Intel 82559 ER 
on board chipset with a Mandrake 8.1 (Kernel 2.4.8-26mdk).  It seems to work 
fine until we test to see if it recovers from a broadcast storm using a 
Fluke Optiview Integrated Network Analyzer.  Which continuously broadcasts a 
Benign IP packet with a frame size of 64 at 100Mb/s.  This obviously will 
cause a lot of errors, perhaps missed packets etc.  We are just looking for 
it (the driver) to recover properly once the broadcast has stopped.  In this 
case it does not ... after repeated  "eth0: Too much work at interrupt, 
status=0x4050" and "eth0: card reports no resources." messages in the system 
log the eth0 ineterface seems to stop functioning.  The "ifconfig" command 
reportes:
  "error fetching interface information: Device not found". After typing 
'/etc/rc.d/init.d/network restart' brought "eth0" back to normal.

Also we noted that changing the broadcast test to send out frame size of 
1500 bytes instead of 64 does not generate any errors and eth0 interface 
works fine.

So, has any one run into this type of problem before and is this a bug?  I 
know that I can increase "max_interrupt_work" from 20 to somethign higher 
and it might prevent the error but this is not the solution I'm looking for. 
  I hope I'm not wrong in assuming that no errors generated should cause the 
driver to stop working?

Thanks for any help.
joe

_________________________________________________________________
Help STOP SPAM with the new MSN 8 and get 2 months FREE*   
http://join.msn.com/?page=features/junkmail