Re: [BUG] [2.4.21] 8139too 'too much work at interrupt'...

[Jeff Garzik]

Daniel Blueman wrote:
> Yes, the 8139too driver is doing it's job correctly - I was just thinking

I didn't say this, because it is obviously not doing it job correctly.

It needs NAPI to do this.  For now, the driver _should_ stay up, even 
with these 'too much work...' messages.


> about the source of the starvation, ie the cause of the effect. I.e. the
> brokeness is elsewhere, in other drivers etc.

Quoting from my original reply:

	It's due to the 8139 hardware

Tons of tiny packets can be sent before RTL8139 will throttle, which can 
easily be far more than your system can handle at that particular time, 
especially if it's experiencing high load elsewhere.

You need either NAPI (perfect solution) or "cap interrupt work limit at 
X packets" (stopgap solution) to prevent your 8139 hardware from being 
DoS'd off the network.

	Jeff



-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/