From mboxrd@z Thu Jan  1 00:00:00 1970
From: David Miller <davem@davemloft.net>
Subject: Re: r8169 tx batching(?) causing performance problems
Date: Wed, 03 Oct 2018 09:31:46 -0700 (PDT)
Message-ID: <20181003.093146.431165270358997515.davem@davemloft.net>
References: <23097.1538562476@warthog.procyon.org.uk>
Mime-Version: 1.0
Content-Type: Text/Plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Cc: nic_swsd@realtek.com, netdev@vger.kernel.org
To: dhowells@redhat.com
Return-path: <netdev-owner@vger.kernel.org>
Received: from shards.monkeyblade.net ([23.128.96.9]:34382 "EHLO
        shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1726829AbeJCXUy (ORCPT
        <rfc822;netdev@vger.kernel.org>); Wed, 3 Oct 2018 19:20:54 -0400
In-Reply-To: <23097.1538562476@warthog.procyon.org.uk>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>


Probably you are seeing some interrupt mitigation.

It seems there is a difference in how the interrupt mitigation is
programmed on for 8168 chips vs. others by default.  Most get
all zeros in the IntrMitigate register, whilst for 8168 chips
a value of 0x5151 is programmed.

You can play with ethtool to mess with the coalescing settings
to see if this is part of the problem.

I bet this might explain the behavior you see after including
even Heiner's TXCFG_AUTO_FIFO patch.