linux-kernel - Re: [PATCH] net: phy: phy_ethtool_ksettings

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <Z02GCGMOuiwZ4qvA@shell.armlinux.org.uk>
Date: Mon, 2 Dec 2024 10:03:52 +0000
From: "Russell King (Oracle)" <linux@...linux.org.uk>
To: Nikita Yushchenko <nikita.yoush@...entembedded.com>
Cc: Andrew Lunn <andrew@...n.ch>, Heiner Kallweit <hkallweit1@...il.com>,
	"David S. Miller" <davem@...emloft.net>,
	Eric Dumazet <edumazet@...gle.com>,
	Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
	netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
	Michael Dege <michael.dege@...esas.com>,
	Christian Mardmoeller <christian.mardmoeller@...esas.com>,
	Dennis Ostermann <dennis.ostermann@...esas.com>
Subject: Re: [PATCH] net: phy: phy_ethtool_ksettings_set: Allow any supported
 speed

On Mon, Dec 02, 2024 at 01:33:52PM +0500, Nikita Yushchenko wrote:
> When auto-negotiation is not used, allow any speed/duplex pair
> supported by the PHY, not only 10/100/1000 half/full.
> 
> This enables drivers to use phy_ethtool_set_link_ksettings() in their
> ethtool_ops and still support configuring PHYs for speeds above 1 GBps.
> 
> Also this will cause an error return on attempt to manually set
> speed/duplex pair that is not supported by the PHY.

Does IEEE 802.3 allow auto-negotiation to be turned off for speeds
greater than 1Gbps?

My research for 1G speeds indicated that AN is required as part of the
establishment of link parameters other than the capabilities of each
end. We have PHYs that require AN to be turned on for 1G speeds, and
other PHYs that allow the AN enable bit to be cleared, but internally
keep it enabled for 1G. To eliminate patches in drivers that force
AN for 1G or error out the ksettings_set call, phylib now emulates the
advertisement for all PHYs and keeps AN enabled when the user requests
fixed-speed 1G, which is what Marvell PHYs do and is the most sensible
solution.

Presently, I don't think it makes sense to turn off AN for speeds
beyond 1G. You need to provide a very good reason for why this is
desired, a real use for it, and indicate why it would be safe to
do.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!