> > Ah, I overlooked the implementation. What we care is to only run mmx if
> > there is no FPU data to clobber (but we can run mmx even if the current
> > task used math in some previous timeslice). So the right implementation
> > should be this according to me:
> >
> > if (!(current->flags & PF_USEDFPU))
> > checksum_mmx();
> > else
> > checksum_nommx();
>
> Wont this have cache trashing problems?
I really don't think so -- if you have done taskswitch from/to
different task, your caches are probably lost anyways. [And -- we've
only increased cache footprint twice. If cache is so big it can hold
task switching routine, it is probably big enough for two copies of
packet checksummer.]
Pavel
-- I'm really pavel@ucw.cz. Look at http://195.113.31.123/~pavel. Pavel Hi! I'm a .signature virus! Copy me into your ~/.signature, please!- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/