From mboxrd@z Thu Jan  1 00:00:00 1970
From: Eric Dumazet <eric.dumazet@gmail.com>
Subject: Re: [PATCH net-next v2] net: Add sysctl to toggle early demux for
 tcp and udp
Date: Sun, 19 Mar 2017 12:18:39 -0700
Message-ID: <1489951119.16816.3.camel@edumazet-glaptop3.roam.corp.google.com>
References: <1489116660-4244-1-git-send-email-subashab@codeaurora.org>
         <CALx6S360Y3tmKMHVMUEuRKCskuwXi+9JBfce6WiSpko1sHj=Yg@mail.gmail.com>
         <674d67f5d76f761f3e872dff274a8bda@codeaurora.org>
         <CALx6S340TYEFDAbQfG6X4P0tPJWvyxyXpnLgp9JY7UoBuPEJVA@mail.gmail.com>
         <1489191742.28631.35.camel@edumazet-glaptop3.roam.corp.google.com>
         <CALx6S36eXpOn70paFnkSyafMkF7Q=MQWJKKb2=d6+hNpK-8V8A@mail.gmail.com>
         <e643b82c330eabfa0cef21c480eb6c2c@codeaurora.org>
         <CALx6S343iWzOKrKwHquG8qmyLyKOgQr_KroHyypE_9MFt2+1BQ@mail.gmail.com>
         <c965b4cc8b5f81cc17836f91423935f5@codeaurora.org>
Mime-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
Cc: Tom Herbert <tom@herbertland.com>,
        Linux Kernel Network Developers <netdev@vger.kernel.org>,
        Stephen Hemminger <stephen@networkplumber.org>,
        netdev-owner@vger.kernel.org
To: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Return-path: <netdev-owner@vger.kernel.org>
Received: from mail-pg0-f68.google.com ([74.125.83.68]:35641 "EHLO
        mail-pg0-f68.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1751643AbdCSTSn (ORCPT
        <rfc822;netdev@vger.kernel.org>); Sun, 19 Mar 2017 15:18:43 -0400
In-Reply-To: <c965b4cc8b5f81cc17836f91423935f5@codeaurora.org>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

On Sat, 2017-03-18 at 20:07 -0600, Subash Abhinov Kasiviswanathan wrote:
> > Less than 1% performance improvement in a benchmark doesn't justify
> > the complexity of the patch. Eric's hypothesis was that an unconnected
> > UDP socket may show issues because of cache misses in look-ups due to
> > so many different sources. This should be fairly easy to benchmark by
> > randomly setting source address in your test (IP any and routing my
> > need to be set appropriately).
> > 
> 
> With different source addresses, a larger increase is seen here
> (633->654Mbps).

Yes, also note the extra cache line might be not noticed if the cpu can
keep in its caches the whole UDP hash table (1 MB on 64bit kernels for
the portion touched in IP early demux)

So the exact slowdown depends on CPU generation / cache sizes.