netdev
[Top] [All Lists]

Re: RFC: NAPI packet weighting patch

To: P@xxxxxxxxxxxxxx
Subject: Re: RFC: NAPI packet weighting patch
From: jamal <hadi@xxxxxxxxxx>
Date: Wed, 22 Jun 2005 15:37:46 -0400
Cc: "David S. Miller" <davem@xxxxxxxxxxxxx>, gandalf@xxxxxxxxxxxxxx, shemminger@xxxxxxxx, mitch.a.williams@xxxxxxxxx, john.ronciak@xxxxxxxxx, mchan@xxxxxxxxxxxx, buytenh@xxxxxxxxxxxxxx, jdmason@xxxxxxxxxx, netdev@xxxxxxxxxxx, Robert.Olsson@xxxxxxxxxxx, ganesh.venkatesan@xxxxxxxxx, jesse.brandeburg@xxxxxxxxx
In-reply-to: <42B92490.40005@draigBrady.com>
Organization: unknown
References: <42A5284C.3060808@osdl.org> <1118147904.6320.108.camel@localhost.localdomain> <Pine.LNX.4.58.0506071351080.16594@tux.rsn.bth.se> <20050621.133704.08321534.davem@davemloft.net> <42B92490.40005@draigBrady.com>
Reply-to: hadi@xxxxxxxxxx
Sender: netdev-bounce@xxxxxxxxxxx
On Wed, 2005-22-06 at 09:42 +0100, P@xxxxxxxxxxxxxx wrote:

> 
> Yes the copy is essentially free here as the data is already cached.
> 
> As a data point, I went the whole hog and used buffer recycling
> in my essentially packet sniffing application. I.E. there are no
> allocs per packet at all, and this make a HUGE difference. On a
> 2x3.4GHz 2xe1000 system I can receive 620Kpps per port sustained
> into my userspace app which does a LOT of processing per packet.
> Without the buffer recycling is was around 250Kpps.
> Note I don't reuse an skb until the packet is copied into a
> PACKET_MMAP buffer.

Was this machine SMP? NAPI involved? I take it nothing interfering in
the middle with the headers?

cheers,
jamal


<Prev in Thread] Current Thread [Next in Thread>