netdev - Re: [PATCH net-next 3/3] net: phylink: add phylink_sfp_select_interface

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 for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <aG6Pd8sqgL5rILm-@shell.armlinux.org.uk>
Date: Wed, 9 Jul 2025 16:49:11 +0100
From: "Russell King (Oracle)" <linux@...linux.org.uk>
To: Alexander Duyck <alexander.duyck@...il.com>
Cc: Maxime Chevallier <maxime.chevallier@...tlin.com>,
	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>, netdev@...r.kernel.org,
	Paolo Abeni <pabeni@...hat.com>
Subject: Re: [PATCH net-next 3/3] net: phylink: add
 phylink_sfp_select_interface_speed()

On Wed, Jul 09, 2025 at 08:37:51AM -0700, Alexander Duyck wrote:
> On Wed, Jul 2, 2025 at 12:18 PM Russell King (Oracle)
> <linux@...linux.org.uk> wrote:
> >
> > On Wed, Jul 02, 2025 at 11:07:52AM -0700, Alexander Duyck wrote:
> > > On Wed, Jul 2, 2025 at 6:37 AM Russell King (Oracle)
> > > <linux@...linux.org.uk> wrote:
> > > >
> > > > On Wed, Jul 02, 2025 at 03:14:26PM +0200, Maxime Chevallier wrote:
> > > > > On Wed, 02 Jul 2025 10:44:34 +0100
> > > > > "Russell King (Oracle)" <rmk+kernel@...linux.org.uk> wrote:
> > > > >
> > > > > > Add phylink_sfp_select_interface_speed() which attempts to select the
> > > > > > SFP interface based on the ethtool speed when autoneg is turned off.
> > > > > > This allows users to turn off autoneg for SFPs that support multiple
> > > > > > interface modes, and have an appropriate interface mode selected.
> > > > > >
> > > > > > Signed-off-by: Russell King (Oracle) <rmk+kernel@...linux.org.uk>
> > > > >
> > > > > Reviewed-by: Maxime Chevallier <maxime.chevallier@...tlin.com>
> > > > >
> > > > > I don't have any hardware to perform relevant tests on this :(
> > > >
> > > > Me neither, I should've said. I'd like to see a t-b from
> > > > Alexander Duyck who originally had the problem before this is
> > > > merged.
> > >
> > > It will probably be several days before I can get around to testing it
> > > since I am slammed with meetings most of the next two days, then have
> > > a holiday weekend coming up.
> >
> > I, too, have a vacation - from tomorrow for three weeks. I may dip in
> > and out of kernel emails during that period, but it depends what
> > happens each day.
> 
> So I was able to go in and test it. I ended up just running the
> testing in QEMU w/ my patch set that currently enables QSFP support.
> From what I can tell it appears to be mostly working. Before when I
> tried to alter the speed to go from 100G to 50G it wouldn't change.
> After your patch set it appears to change, although I am noticing a
> slight difference from the default config.
> 
> So by default we come up in the 100G w/ the QSFP configuration:
> [root@...alhost fbnic]# ethtool enp1s0
> Settings for enp1s0:
>         Supported ports: [  ]
>         Supported link modes:   50000baseCR/Full
>                                 100000baseCR2/Full
>         Supported pause frame use: Symmetric Receive-only
>         Supports auto-negotiation: No
>         Supported FEC modes: RS
>         Advertised link modes:  100000baseCR2/Full
>         Advertised pause frame use: Symmetric Receive-only
>         Advertised auto-negotiation: No
>         Advertised FEC modes: RS
>         Link partner advertised link modes:  100000baseCR2/Full
>         Link partner advertised pause frame use: No
>         Link partner advertised auto-negotiation: No
>         Link partner advertised FEC modes: RS
>         Speed: 100000Mb/s
>         Duplex: Full
>         Auto-negotiation: off
>         Port: Other
>         PHYAD: 0
>         Transceiver: internal
>         Link detected: yes
> 
> I then change the speed to 50G and it links back up after a few
> seconds, however the "Advertised link modes" goes from
> "100000baseCR2/Full" to "Not reported" as shown here:
> [root@...alhost fbnic]# ethtool -s enp1s0 speed 50000
> [root@...alhost fbnic]# ethtool enp1s0
> Settings for enp1s0:
>         Supported ports: [  ]
>         Supported link modes:   50000baseCR/Full
>                                 100000baseCR2/Full
>         Supported pause frame use: Symmetric Receive-only
>         Supports auto-negotiation: No
>         Supported FEC modes: RS
>         Advertised link modes:  Not reported
>         Advertised pause frame use: Symmetric Receive-only
>         Advertised auto-negotiation: No
>         Advertised FEC modes: RS
>         Link partner advertised link modes:  100000baseCR2/Full
>         Link partner advertised pause frame use: No
>         Link partner advertised auto-negotiation: No
>         Link partner advertised FEC modes: RS
>         Speed: 50000Mb/s
>         Duplex: Full
>         Auto-negotiation: off
>         Port: Other
>         PHYAD: 0
>         Transceiver: internal
>         Link detected: yes
> 
> So all-in-all it is an improvement over the previous behavior although
> there may still need to be some work done to improve the consistency
> so that it more closely matches up with what happens when you
> initially configure the interface.

Likely, it's ethtool doing that. Autoneg-off (fixed speed) modes
generally don't have an advertisement. I suggest you debug at UAPI
level.

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