netdev - Re: [PATCH] net: phy: fix autoneg invalid error state of GBSR register.

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>] [day] [month] [year] [list]

Date:   Wed, 1 Sep 2021 14:43:26 +0200
From:   Andrew Lunn <andrew@...n.ch>
To:     Qiang Ma <maqianga@...ontech.com>
Cc:     f.fainelli@...il.com, davem@...emloft.net, netdev@...r.kernel.org,
        linux-kernel@...r.kernel.org, Ming Wang <wangming01@...ngson.cn>
Subject: Re: [PATCH] net: phy: fix autoneg invalid error state of GBSR
 register.

On Wed, Sep 01, 2021 at 06:56:08PM +0800, Qiang Ma wrote:
> When the PHY is not link and working in polling mode, PHY state
> machine will periodically read GBSR register. Normally, the GBSR
> register value is 0. But as the log show, in very small cases the
> value is 0x8640. The value 0x8640 means Master/Slave resolution
> failed. The PHY state machine will enter PHY_HALTED state and the
> PHY will never be able to link.
> 
> [49176.903012] [debug]: lpagb:0x0 adv:0x300
> [49177.927025] [debug]: lpagb:0x8640 adv:0x300
> [49177.927034] Generic PHY stmmac-18:00: Master/Slave resolution failed
> [49177.927241] [debug]: lpagb:0x0 adv:0x300
> 
> According to the RTL8211E PHY chip datasheet, the contents of GBSR register
> are valid only when auto negotiation is completed(BMSR[5]: Auto-
> Negotiation Complete = 1). This patch adds the condition before
> reading GBSR register to fix this error state.
> 
> Signed-off-by: Ming Wang <wangming01@...ngson.cn>
> Signed-off-by: Qiang Ma <maqianga@...ontech.com>
> ---
>  drivers/net/phy/phy_device.c | 10 ++++------
>  include/linux/phy.h          |  1 +
>  2 files changed, 5 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
> index b884b681d5c5..b77ed5ec31c6 100644
> --- a/drivers/net/phy/phy_device.c
> +++ b/drivers/net/phy/phy_device.c
> @@ -1532,10 +1532,8 @@ int genphy_update_link(struct phy_device *phydev)
>  	if (status < 0)
>  		return status;
>  
> -	if ((status & BMSR_LSTATUS) == 0)
> -		phydev->link = 0;
> -	else
> -		phydev->link = 1;
> +	phydev->link = status & BMSR_LSTATUS ? 1 : 0;
> +	phydev->autoneg_complete = status & BMSR_ANEGCOMPLETE ? 1 : 0;
>  
>  	return 0;
>  }

What tree is this against? Both net-next/master and net/master have:

        if (status < 0)
                return status;
done:
        phydev->link = status & BMSR_LSTATUS ? 1 : 0;
        phydev->autoneg_complete = status & BMSR_ANEGCOMPLETE ? 1 : 0;

        /* Consider the case that autoneg was started and "aneg complete"
         * bit has been reset, but "link up" bit not yet.
         */
        if (phydev->autoneg == AUTONEG_ENABLE && !phydev->autoneg_complete)
                phydev->link = 0;

        return 0;
}

It looks like you are using an old tree.

	Andrew