netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Martyn Welch <martyn.welch@collabora.com>
To: Andrew Lunn <andrew@lunn.ch>
Cc: Vivien Didelot <vivien.didelot@gmail.com>,
	Florian Fainelli <f.fainelli@gmail.com>,
	Vladimir Oltean <olteanv@gmail.com>,
	netdev@vger.kernel.org, kernel@collabora.com,
	Russell King <rmk+kernel@armlinux.org.uk>
Subject: Re: mv88e6240 configuration broken for B850v3
Date: Mon, 06 Dec 2021 18:26:25 +0000	[thread overview]
Message-ID: <fb6370266a71fdd855d6cf4d147780e0f9e1f5e4.camel@collabora.com> (raw)
In-Reply-To: <b0643124f372db5e579b11237b65336430a71474.camel@collabora.com>

On Mon, 2021-12-06 at 17:44 +0000, Martyn Welch wrote:
> On Fri, 2021-12-03 at 17:25 +0100, Andrew Lunn wrote:
> > > Hi Andrew,
> > 
> > Adding Russell to Cc:
> > 
> > > I'm currently in the process of updating the GE B850v3 [1] to run
> > > a
> > > newer kernel than the one it's currently running. 
> > 
> > Which kernel exactly. We like bug reports against net-next, or at
> > least the last -rc.
> > 
> 
> I tested using v5.15-rc3 and that was also affected.
> 

I've just tested v5.16-rc4 (sorry - just realised I previously wrote
v5.15-rc3, it was v5.16-rc3...) and that was exactly the same.

> > > 
> > > This device (and others in the same family) use a mv88e6240
> > > switch to
> > > provide a number of their ethernet ports. The CPU link on the
> > > switch
> > > is
> > > connected via a PHY, as the network port on the SoM used is
> > > exposed
> > > via
> > > a PHY.
> > > 
> > > The ports of the B850v3 stopped working when I upgraded,
> > > bisecting
> > > resulted in me finding that this commit was the root cause:
> > > 
> > > 3be98b2d5fbc (refs/bisect/bad) net: dsa: Down cpu/dsa ports
> > > phylink
> > > will control
> > > 
> > > I think this is causing the PHY on the mv88e6240 side of the CPU
> > > link
> > > to be forced down in our use case.
> > > 
> > > I assume an extra check is needed here to stop that in cases like
> > > ours,
> > > though I'm not sure what at this point. Any ideas?
> > 
> > From the commit message.
> > 
> >     DSA and CPU ports can be configured in two ways. By default,
> > the
> >     driver should configure such ports to there maximum bandwidth.
> > For
> >     most use cases, this is sufficient. When this default is
> > insufficient,
> >     a phylink instance can be bound to such ports, and phylink will
> >     configure the port,
> > 
> > You have a phy-handle in your node:
> > 
> >         port@4 {
> >                 reg = <4>;
> >                 label = "cpu";
> >                 ethernet = <&switch_nic>;
> >                 phy-handle = <&switchphy4>;
> >         };
> > 
> > so i would expect there to be a phylink instance. The commit
> > message
> > continues to say:
> > 
> >                                           and phylink will
> >     configure the port, e.g. based on fixed-link properties.
> > 
> > So i think you are asking the wrong question. It is not an extra
> > check
> > is needed here, we need to understand why phylink is not
> > configuring
> > the MAC. Or is that configuration wrong.
> > 
> > I suggest you add #define DEBUG 1 to the very top of
> > drivers/net/phy/phylink.c so we can see what phylink is doing.
> > 
> 
> The pertinent parts of the logs (from v5.16-rc3, with the above debug
> added) appear to be:
> 
> # dmesg | grep -e mv88e6085 -e DSA -e libphy
> [    2.119282] libphy: Fixed MDIO Bus: probed
> [    2.124547] libphy: GPIO Bitbanged MDIO: probed
> [    2.129795] mv88e6085 gpio-0:00: switch 0x2400 detected: Marvell
> 88E6240, revision 1
> [    2.150568] libphy: mdio: probed
> [    2.224233] libphy: fec_enet_mii_bus: probed
> [    3.064455] mv88e6085 gpio-0:00: switch 0x2400 detected: Marvell
> 88E6240, revision 1
> [    3.083930] libphy: mdio: probed
> [    3.844882] mv88e6085 gpio-0:00 eneport1 (uninitialized): PHY
> [!mdio-gpio!switch@0!mdio:00] driver [Marvell 88E1540] (irq=362)
> [    3.856335] mv88e6085 gpio-0:00 eneport1 (uninitialized): phy:
> setting supported 0000000,00000000,000022ef advertising
> 0000000,00000000,000022ef
> [    3.962649] mv88e6085 gpio-0:00 eneport2 (uninitialized): PHY
> [!mdio-gpio!switch@0!mdio:01] driver [Marvell 88E1540] (irq=363)
> [    3.974091] mv88e6085 gpio-0:00 eneport2 (uninitialized): phy:
> setting supported 0000000,00000000,000022ef advertising
> 0000000,00000000,000022ef
> [    4.080405] mv88e6085 gpio-0:00 enix (uninitialized): PHY
> [!mdio-gpio!switch@0!mdio:02] driver [Marvell 88E1540] (irq=364)
> [    4.091510] mv88e6085 gpio-0:00 enix (uninitialized): phy: setting
> supported 0000000,00000000,000022ef advertising
> 0000000,00000000,000022ef
> [    4.202683] mv88e6085 gpio-0:00 enid (uninitialized): PHY
> [!mdio-gpio!switch@0!mdio:03] driver [Marvell 88E1540] (irq=365)
> [    4.213774] mv88e6085 gpio-0:00 enid (uninitialized): phy: setting
> supported 0000000,00000000,000022ef advertising
> 0000000,00000000,000022ef
> [    4.298209] mv88e6085 gpio-0:00: PHY [!mdio-gpio!switch@0!mdio:04]
> driver [Marvell 88E1540] (irq=366)
> [    4.307475] mv88e6085 gpio-0:00: phy: setting supported
> 0000000,00000000,000022ef advertising 0000000,00000000,000022ef
> [    4.314285] mv88e6085 gpio-0:00: configuring for phy/ link mode
> [    4.320251] mv88e6085 gpio-0:00: major config 
> [    4.320262] mv88e6085 gpio-0:00: phylink_mac_config:
> mode=phy//Unknown/Unknown adv=0000000,00000000,00000000 pause=00
> link=0
> an=0
> [    4.324265] DSA: tree 0 setup
> [    4.399423] mv88e6085 gpio-0:00: phy link down
> /Unknown/Unknown/off
> [   15.600977] mv88e6085 gpio-0:00 enix: configuring for phy/ link
> mode
> [   15.607417] mv88e6085 gpio-0:00 enix: major config 
> [   15.607443] mv88e6085 gpio-0:00 enix: phylink_mac_config:
> mode=phy//Unknown/Unknown adv=0000000,00000000,00000000 pause=00
> link=0
> an=0
> [   15.678811] mv88e6085 gpio-0:00 enix: phy link down
> /Unknown/Unknown/off
> [   15.961559] mv88e6085 gpio-0:00 enid: configuring for phy/ link
> mode
> [   15.967945] mv88e6085 gpio-0:00 enid: major config 
> [   15.967958] mv88e6085 gpio-0:00 enid: phylink_mac_config:
> mode=phy//Unknown/Unknown adv=0000000,00000000,00000000 pause=00
> link=0
> an=0
> [   15.986370] mv88e6085 gpio-0:00 enix: configuring for phy/ link
> mode
> [   15.992829] mv88e6085 gpio-0:00 enix: major config 
> [   15.992843] mv88e6085 gpio-0:00 enix: phylink_mac_config:
> mode=phy//Unknown/Unknown adv=0000000,00000000,00000000 pause=00
> link=0
> an=0
> [   16.019628] mv88e6085 gpio-0:00 eneport2: configuring for phy/
> link
> mode
> [   16.026370] mv88e6085 gpio-0:00 eneport2: major config 
> [   16.026382] mv88e6085 gpio-0:00 eneport2: phylink_mac_config:
> mode=phy//Unknown/Unknown adv=0000000,00000000,00000000 pause=00
> link=0
> an=0
> [   16.057585] mv88e6085 gpio-0:00 eneport1: configuring for phy/
> link
> mode
> [   16.064367] mv88e6085 gpio-0:00 eneport1: major config 
> [   16.064381] mv88e6085 gpio-0:00 eneport1: phylink_mac_config:
> mode=phy//Unknown/Unknown adv=0000000,00000000,00000000 pause=00
> link=0
> an=0
> [   16.132700] mv88e6085 gpio-0:00 enid: phy link up /10Mbps/Full/off
> [   16.134210] mv88e6085 gpio-0:00 enid: phylink_mac_config:
> mode=phy//10Mbps/Full adv=0000000,00000000,00000000 pause=00 link=1
> an=0
> [   16.141177] mv88e6085 gpio-0:00 enid: Link is Up - 10Mbps/Full -
> flow control off
> [   16.195201] mv88e6085 gpio-0:00 enix: phy link down
> /Unknown/Unknown/off
> [   16.215131] mv88e6085 gpio-0:00 eneport2: phy link down
> /Unknown/Unknown/off
> [   16.254004] mv88e6085 gpio-0:00 eneport1: phy link up
> /10Mbps/Full/off
> [   16.254027] mv88e6085 gpio-0:00 eneport1: phylink_mac_config:
> mode=phy//10Mbps/Full adv=0000000,00000000,00000000 pause=00 link=1
> an=0
> [   16.254056] mv88e6085 gpio-0:00 eneport1: Link is Up - 10Mbps/Full
> -
> flow control off
> [   18.706700] mv88e6085 gpio-0:00: phy link up /1Gbps/Full/rx/tx
> [   18.708976] mv88e6085 gpio-0:00: phylink_mac_config:
> mode=phy//1Gbps/Full adv=0000000,00000000,00000000 pause=03 link=1
> an=0
> [   18.709002] mv88e6085 gpio-0:00: Link is Up - 1Gbps/Full - flow
> control rx/tx
> 
> Despite the last few lines suggesting to me the phy link is up, I'm
> unable to access the network I'd expect to be able to access.
> 
> >         Andrew
> > 
> 


  reply	other threads:[~2021-12-06 18:28 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-12-03  9:06 mv88e6240 configuration broken for B850v3 Martyn Welch
2021-12-03 16:25 ` Andrew Lunn
2021-12-06 17:44   ` Martyn Welch
2021-12-06 18:26     ` Martyn Welch [this message]
2021-12-06 18:31       ` Vladimir Oltean
2021-12-06 18:37         ` Martyn Welch
2021-12-06 18:50           ` Vladimir Oltean
2021-12-06 19:24             ` Martyn Welch
2021-12-06 19:37               ` Vladimir Oltean
2021-12-06 19:53                 ` Andrew Lunn
2021-12-06 20:01                   ` Vladimir Oltean
2021-12-06 20:18                     ` Russell King (Oracle)
2021-12-06 20:29                       ` Vladimir Oltean
2021-12-07 14:09                         ` Andrew Lunn
2021-12-06 21:44                       ` Vladimir Oltean
2021-12-06 22:13                         ` Russell King (Oracle)
2021-12-06 20:07                 ` Russell King (Oracle)
2021-12-06 20:23                   ` Vladimir Oltean
2021-12-06 20:51                     ` Russell King (Oracle)
2021-12-06 21:13                       ` Vladimir Oltean
2021-12-06 21:27                         ` Russell King (Oracle)
2021-12-06 21:49                           ` Russell King (Oracle)
2021-12-06 23:27                             ` Vladimir Oltean
2021-12-07  0:58                               ` Russell King (Oracle)
2021-12-07 13:24                                 ` Vladimir Oltean
2021-12-07 13:59                                   ` Russell King (Oracle)
2021-12-07 14:37                                     ` Vladimir Oltean
2021-12-07 14:53                                       ` Russell King (Oracle)
2021-12-06 21:51                           ` Vladimir Oltean
2021-12-06 22:17                             ` Andrew Lunn
2021-12-06 22:22                             ` Russell King (Oracle)
2021-12-06 23:44                               ` Vladimir Oltean
2021-12-07  2:06                                 ` Andrew Lunn
2021-12-07 12:48                                   ` Vladimir Oltean

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=fb6370266a71fdd855d6cf4d147780e0f9e1f5e4.camel@collabora.com \
    --to=martyn.welch@collabora.com \
    --cc=andrew@lunn.ch \
    --cc=f.fainelli@gmail.com \
    --cc=kernel@collabora.com \
    --cc=netdev@vger.kernel.org \
    --cc=olteanv@gmail.com \
    --cc=rmk+kernel@armlinux.org.uk \
    --cc=vivien.didelot@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).