Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Mon, 16 Nov 2020 15:56:05 -0800
From:   Jakub Kicinski <kuba@kernel.org>
To:     Vladimir Oltean <olteanv@gmail.com>
Cc:     Oleksij Rempel <o.rempel@pengutronix.de>,
        Andrew Lunn <andrew@lunn.ch>,
        Vivien Didelot <vivien.didelot@gmail.com>,
        Florian Fainelli <f.fainelli@gmail.com>,
        "David S. Miller" <davem@davemloft.net>,
        Russell King <linux@armlinux.org.uk>,
        Pengutronix Kernel Team <kernel@pengutronix.de>,
        netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
        linux-mips@vger.kernel.org
Subject: Re: [PATCH v1 net-next] net: dsa: qca: ar9331: add ethtool stats
 support
Message-ID: <20201116155605.1309c4eb@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
In-Reply-To: <20201116232731.4utpige7fguzghsi@skbuf>
References: <20201115073533.1366-1-o.rempel@pengutronix.de>
        <20201116133453.270b8db5@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
        <20201116222146.znetv5u2q2q2vk2j@skbuf>
        <20201116143544.036baf58@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
        <20201116230053.ddub7p6lvvszz7ic@skbuf>
        <20201116151347.591925ca@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
        <20201116232731.4utpige7fguzghsi@skbuf>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Tue, 17 Nov 2020 01:27:31 +0200 Vladimir Oltean wrote:
> > Note that I said _forwarded_. Frames are either forwarded by the HW or
> > SW (former never hit the CPU, while the latter do hit the CPU or
> > originate from it).  
> 
> Ah, you were just thinking out loud, I really did not understand what
> you meant by the separation between "forwarded in software" and
> "forwarded in hardware".
> Yes, the hardware typically only gives us MAC-level counters anyway.
> Another way to look at it is that the number of packets forwarded in
> hardware from a given port are equal to the total number of RX packets
> on that MAC minus the packets seen by the CPU coming from that port.
> So all in all, it's the MAC-level counters we should expose in
> .ndo_get_stats64, I'm glad you agree. As for the error packets, I
> suppose that would be a driver-specific aggregate.

Yup, sorry about the confusion, I was only working on those stats
with SDN/OvS/tc hardware, which explains the slight difference in
terminology.