On Wed, 2005-07-12 at 14:11 -0800, David S. Miller wrote:
> From: Eric Dumazet <[EMAIL PROTECTED]>
> Date: Wed, 07 Dec 2005 23:04:04 +0100
> 
> > Another try could be to do some benchmarks about NET_IP_ALIGN being a valid 
> > optimization nowadays :
> > Maybe setting it to 0 in e1000 driver could give better results.
> > Could somebody give it a try ?
> 
> NET_IP_ALIGN is a two-folded thing:
> 
> 1) Getting IP headers aligned
> 2) Getting good DMA strides (this is why ppc sets it to zero).
> 
> So, keep that in mind.
> 
> On x86 you won't get the unaligned traps, but on other platforms
> you will and I know that the ia64 folks will get a couple
> megabytes of unaligned memory access trap kernel logs when
> their machine boots and will rightfully scream bloody murder
> as a result.
> 
> So you can set it to zero if you want on x86 systems, but even
> there I really don't recommend it.

I can give this a try tommorow as well if Robert doesnt beat me to it.

cheers,
jamal

-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to