Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Sat, 12 Mar 2022 20:04:12 -0800
From:   Richard Cochran <richardcochran@gmail.com>
To:     "Allan W. Nielsen" <allan.nielsen@microchip.com>
Cc:     Horatiu Vultur <horatiu.vultur@microchip.com>,
        "Russell King (Oracle)" <linux@armlinux.org.uk>,
        Andrew Lunn <andrew@lunn.ch>, Divya.Koppera@microchip.com,
        netdev@vger.kernel.org, hkallweit1@gmail.com, davem@davemloft.net,
        kuba@kernel.org, robh+dt@kernel.org, devicetree@vger.kernel.org,
        linux-kernel@vger.kernel.org, UNGLinuxDriver@microchip.com,
        Madhuri.Sripada@microchip.com, Manohar.Puri@microchip.com
Subject: Re: [PATCH net-next 2/3] dt-bindings: net: micrel: Configure latency
 values and timestamping check for LAN8814 phy
Message-ID: <20220313040412.GA30488@hoboy.vegasvil.org>
References: <20220308154345.l4mk2oab4u5ydn5r@soft-dev3-1.localhost>
 <YiecBKGhVui1Gtb/@lunn.ch>
 <20220308221404.bwhujvsdp253t4g3@soft-dev3-1.localhost>
 <YifoltDp4/Fs+9op@lunn.ch>
 <20220309132443.axyzcsc5kyb26su4@soft-dev3-1.localhost>
 <Yii/9RH67BEjNtLM@shell.armlinux.org.uk>
 <20220309195252.GB9663@hoboy.vegasvil.org>
 <20220311142814.z3h5nystnrkvbzek@soft-dev3-1.localhost>
 <20220311150842.GC7817@hoboy.vegasvil.org>
 <20220312193620.owhfd43dzzxtytgs@den-dk-m31684h>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20220312193620.owhfd43dzzxtytgs@den-dk-m31684h>
User-Agent: Mutt/1.10.1 (2018-07-13)
Precedence: bulk

On Sat, Mar 12, 2022 at 08:36:20PM +0100, Allan W. Nielsen wrote:
> I did skim through the articles, and as you hinted he does find small
> latency differences across different packets. (but as I understood, very
> few PHYs was tested).

There is also previous work cited in those articles.
 
> But this is not really an argument for not having _default_ values
> hard-coded in the driver (or DT, but lets forget about DT for now).

You put them in the DTS.  That means you expect them to need changes.

DTS is the WRONG place for such things.

If your numbers are perfect, then do the corrections in silicon/firmware.

If the numbers aren't 100% perfect, then provide your customers with a
test report providing the recommended numbers.  Include a proper
explanation of the test methodology and assumptions used in your
analysis.  Heck, you can even given them linuxptp config snippets (and
other for other popular PTP stacks, Oregano, ixat, ptpd, etc)

Don't hard code random, changing numbers into kernel drivers.

Thanks,
Richard