[epic] eth0: Too much work at interrupt - PCI IRQ-sharing problem?

[Thomas Stroesslin]

I also get this error. It blocks everything, I don't receive/transmit
any more packets at all...

Anthony: hmm, I'll try increasing max_interrupt_work. but then again,
this is like ignoring a big WARNING sign, isn't it?

Till: where did you get your working driver from?


Apr 30 14:14:28 mantrapc1 kernel: eth0: Too much work at interrupt,
IntrStatus=0x008d0004.
Apr 30 14:14:43 mantrapc1 last message repeated 33 times
etc.

and also:
Apr 30 15:25:33 mantrapc1 kernel: NETDEV WATCHDOG: eth0: transmit timed out
Apr 30 15:25:33 mantrapc1 kernel: eth0: Transmit timeout using MII device,
Tx status 4003.
Apr 30 15:25:33 mantrapc1 kernel: eth0: Restarting the EPIC chip, Rx 1/1 Tx
2/12.
Apr 30 15:25:33 mantrapc1 kernel: eth0: epic_restart() done, cmd status
000a, ctl 0512 interrupt 240000.
Apr 30 15:25:33 mantrapc1 kernel: eth0: Too much work at interrupt,
IntrStatus=0x008d0004.
Apr 30 15:25:34 mantrapc1 kernel: eth0: Setting half-duplex based on MII #3
link partner capability of 0001.

cheers,
tom

-- 
T. Stroesslin
DI-MANTRA (INF 130)
EPFL