fegray@socrates.berkeley.edu said:
> The two morals of this story are that the checksum is not the main problem,
> but that interrupt handling is. Moreover, it's not the e1000 interrupt
> handler that takes the time, it's the generic powerpc interrupt handling.
NAPI is pretty much essential if your CPU is having trouble keeping up with the network. However, it only helps with rx, not tx. If latency isn't critical, you can look using at the e1000 driver's TxIntDelay and TxAbsIntDelay to try and cut down on the frequency of tx interrupts. Using TxDescriptors to increase the size of the tx fifo might let you use more agressive delays, too.
actually, if your ring is full because your tx interrupts arent clearing the buffers quickly enough, you can have quite a lot of thrashing depending on how the application handles a "busy" transmit state.
Dennis
- : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
