Re: [patch] softirq performance fixes, cleanups, 2.4.10.

[Josh MacDonald]

Quoting Rik van Riel ([e-mail]):
> On Fri, 28 Sep 2001 [e-mail] wrote:
> 
> > Please, explain who exactly obtains an advantage of looping.
> > net_rx_action()? Do you see drops in backlog?
> 
> > net_tx_action()? It does not look critical.
> 
> Then how would you explain the 10% speed difference
> Ben and others have seen with gigabit ethernet ?

Could this possibly be due to I-cache improvements?  If the same 
interrupt handling code is being run 10 times at once you should
expect an improvement of that kind.

-josh