Re: [PATCH 2.6] natsemi.c NAPI

[jamal <hadi@xxxxxxxxxx>]

On Tue, 2004-09-28 at 00:23, Manfred Spraul wrote:
> jamal wrote:
[..]
> >Does this mean that the interupt clears when read?
> >Why would reading a NIC register clear a soundcard interupt?
> >
> >  
> >
> No, of course not. But each soundcard interrupt would cause a 
> netif_rx_schedule, with the following callback from tasklet context. 
> IMHO the overhead of that operation is too large. Additionally it breaks 
> the stuck irq detection.

Ok, In such a case it is advised that a e1000 style NAPI is used
(as opposed to the tulip varinat).
Sorry, I should actually read the code first - I ma hoping i am in
context here.

cheers,
jamal