netdev - Re: [PATCH v1] ethtool: provide UAPI for PHY master/slave configuration.

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: <20200415215739.GI657811@lunn.ch>
Date:   Wed, 15 Apr 2020 23:57:39 +0200
From:   Andrew Lunn <andrew@...n.ch>
To:     Oleksij Rempel <o.rempel@...gutronix.de>
Cc:     "David S. Miller" <davem@...emloft.net>,
        Florian Fainelli <f.fainelli@...il.com>,
        Heiner Kallweit <hkallweit1@...il.com>,
        Jakub Kicinski <kuba@...nel.org>,
        Jonathan Corbet <corbet@....net>,
        Michal Kubecek <mkubecek@...e.cz>,
        David Jander <david@...tonic.nl>, kernel@...gutronix.de,
        linux-kernel@...r.kernel.org, netdev@...r.kernel.org,
        Russell King <linux@...linux.org.uk>, mkl@...gutronix.de
Subject: Re: [PATCH v1] ethtool: provide UAPI for PHY master/slave
 configuration.

>  ``ETHTOOL_A_LINKMODES_OURS`` bit set allows setting advertised link modes. If
> diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c
> index d76e038cf2cb5..9f48141f1e701 100644
> --- a/drivers/net/phy/phy.c
> +++ b/drivers/net/phy/phy.c
> @@ -294,7 +294,7 @@ int phy_ethtool_ksettings_set(struct phy_device *phydev,
>  			 phydev->advertising, autoneg == AUTONEG_ENABLE);
>  
>  	phydev->duplex = duplex;
> -
> +	phydev->master_slave = cmd->base.master_slave;
>  	phydev->mdix_ctrl = cmd->base.eth_tp_mdix_ctrl;
>  
>  	/* Restart the PHY */
> @@ -313,6 +313,7 @@ void phy_ethtool_ksettings_get(struct phy_device *phydev,
>  
>  	cmd->base.speed = phydev->speed;
>  	cmd->base.duplex = phydev->duplex;
> +	cmd->base.master_slave = phydev->master_slave;
>  	if (phydev->interface == PHY_INTERFACE_MODE_MOCA)
>  		cmd->base.port = PORT_BNC;
>  	else
> diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
> index c8b0c34030d32..d5edf2bc40e43 100644
> --- a/drivers/net/phy/phy_device.c
> +++ b/drivers/net/phy/phy_device.c
> @@ -604,6 +604,7 @@ struct phy_device *phy_device_create(struct mii_bus *bus, int addr, u32 phy_id,
>  	dev->asym_pause = 0;
>  	dev->link = 0;
>  	dev->interface = PHY_INTERFACE_MODE_GMII;
> +	dev->master_slave = PORT_MODE_UNKNOWN;

phydev->master_slave is how we want the PHY to be configured. I don't
think PORT_MODE_UNKNOWN makes any sense in that contest. 802.3 gives
some defaults. 9.12 should be 0, meaning manual master/slave
configuration is disabled. The majority of linux devices are end
systems. So we should default to a single point device. So i would
initialise PORT_MODE_SLAVE, or whatever we end up calling that.
>  
>  	dev->autoneg = AUTONEG_ENABLE;
>  
> @@ -1772,6 +1773,68 @@ int genphy_setup_forced(struct phy_device *phydev)
>  }
>  EXPORT_SYMBOL(genphy_setup_forced);
>  
> +static int genphy_setup_master_slave(struct phy_device *phydev)
> +{
> +	u16 ctl = 0;
> +
> +	if (!phydev->is_gigabit_capable)
> +		return 0;
> +
> +	switch (phydev->master_slave) {
> +	case PORT_MODE_MASTER:
> +		ctl |= CTL1000_PREFER_MASTER;
> +		/* fallthrough */
> +	case PORT_MODE_SLAVE:
> +		/* CTL1000_ENABLE_MASTER is zero */
> +		break;
> +	case PORT_MODE_MASTER_FORCE:
> +		ctl |= CTL1000_AS_MASTER;
> +		/* fallthrough */
> +	case PORT_MODE_SLAVE_FORCE:
> +		ctl |= CTL1000_ENABLE_MASTER;
> +		break;
> +	case PORT_MODE_UNKNOWN:
> +		return 0;
> +	default:
> +		phydev_warn(phydev, "Unsupported Master/Slave mode\n");
> +		return 0;
> +	}
> +
> +	return phy_modify_changed(phydev, MII_CTRL1000,
> +				  (CTL1000_ENABLE_MASTER | CTL1000_AS_MASTER |
> +				   CTL1000_PREFER_MASTER), ctl);
> +}
> +
> +static int genphy_read_master_slave(struct phy_device *phydev)
> +{
> +	u16 ctl, stat;
> +
> +	if (!phydev->is_gigabit_capable)
> +		return 0;
> +
> +	ctl = phy_read(phydev, MII_CTRL1000);
> +	if (ctl < 0)
> +		return ctl;
> +
> +	stat = phy_read(phydev, MII_STAT1000);
> +	if (stat < 0)
> +		return stat;
> +
> +	if (ctl & CTL1000_ENABLE_MASTER) {
> +		if (stat & LPA_1000MSRES)
> +			phydev->master_slave = PORT_MODE_MASTER_FORCE;
> +		else
> +			phydev->master_slave = PORT_MODE_SLAVE_FORCE;
> +	} else {
> +		if (stat & LPA_1000MSRES)
> +			phydev->master_slave = PORT_MODE_MASTER;
> +		else
> +			phydev->master_slave = PORT_MODE_SLAVE;
> +	}

This seems wrong. phydev->master_slave should be about how we want the
PHY to be configured. genphy_read_master_slave() should be about what
we actually ended up using. It should not be over-writing
phydev->master_slave, it needs to put it into some other variable.

phy_ethtool_ksettings_set() should allow us to set how we want the
device to behave. phy_ethtool_ksettings_get() should return both how
we have configured it, and if master/slave has been resolved, what the
result of the resolution is. Here something like PORT_MODE_UNKNOWN
does make sense, when the link is down, to resolution has not yet
completed.

       Andrew