[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20230206112246.pazwn7r75oru5iq3@skbuf>
Date: Mon, 6 Feb 2023 13:22:46 +0200
From: Vladimir Oltean <olteanv@...il.com>
To: Oleksij Rempel <o.rempel@...gutronix.de>
Cc: Woojung Huh <woojung.huh@...rochip.com>,
UNGLinuxDriver@...rochip.com, Andrew Lunn <andrew@...n.ch>,
Vivien Didelot <vivien.didelot@...il.com>,
Florian Fainelli <f.fainelli@...il.com>,
"David S. Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>, Wei Fang <wei.fang@....com>,
Heiner Kallweit <hkallweit1@...il.com>, kernel@...gutronix.de,
linux-kernel@...r.kernel.org, netdev@...r.kernel.org,
Arun.Ramadoss@...rochip.com, intel-wired-lan@...ts.osuosl.org
Subject: Re: [PATCH net-next v4 02/23] net: phy: add
genphy_c45_read_eee_abilities() function
On Mon, Feb 06, 2023 at 11:49:55AM +0100, Oleksij Rempel wrote:
> > Why stop at 10GBase-KR? Register 3.20 defines EEE abilities up to 100G
> > (for speeds >10G, there seem to be 2 modes, "deep sleep" or "fast wake",
> > with "deep sleep" being essentially equivalent to the only mode
> > available for <=10G modes).
>
> Hm,
>
> If i take only deep sleep, missing modes are:
> 3.20.13 100GBASE-R deep sleep
> family of Physical Layer devices using 100GBASE-R encoding:
> 100000baseCR4_Full
> 100000baseKR4_Full
> 100000baseCR10_Full (missing)
> 100000baseSR4_Full
> 100000baseSR10_Full (missing)
> 100000baseLR4_ER4_Full
>
> 3.20.11 25GBASE-R deep sleep
> family of Physical Layer devices using 25GBASE-R encoding:
> 25000baseCR_Full
> 25000baseER_Full (missing)
> 25000baseKR_Full
> 25000baseLR_Full (missing)
> 25000baseSR_Full
>
> 3.20.9 40GBASE-R deep sleep
> family of Physical Layer devices using 40GBASE-R encoding:
> 40000baseKR4_Full
> 40000baseCR4_Full
> 40000baseSR4_Full
> 40000baseLR4_Full
>
> 3.20.7 40GBASE-T
> 40000baseT_Full (missing)
>
> I have no experience with modes > 1Gbit. Do all of them correct? What
> should we do with missing modes? Or may be it make sense to implement >
> 10G modes separately?
Given the fact that UAPI needs an extension to cover supported/advertisement
bits > 31, I think it makes sense to add these separately. I had not
realized this when I commented on this patch. I don't think we want the
kernel to advertise EEE for some link modes without user space seeing it.
Powered by blists - more mailing lists