linux-kernel - Re: [RFC PATCH net v2] net: phy: Fix the issue that netif always links up after resuming

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: <47b5606c-564c-2fe8-0683-65288dafe028@gmail.com>
Date:   Fri, 30 Nov 2018 19:47:37 +0100
From:   Heiner Kallweit <hkallweit1@...il.com>
To:     Florian Fainelli <f.fainelli@...il.com>,
        Kunihiko Hayashi <hayashi.kunihiko@...ionext.com>,
        Andrew Lunn <andrew@...n.ch>,
        "David S. Miller" <davem@...emloft.net>
Cc:     netdev@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH net v2] net: phy: Fix the issue that netif always
 links up after resuming

On 30.11.2018 18:46, Florian Fainelli wrote:
> 
> 
> On 11/30/2018 1:25 AM, Kunihiko Hayashi wrote:
>> Even though the link is down before entering hibernation,
>> there is an issue that the network interface always links up after resuming
>> from hibernation.
>>
>> The phydev->state is PHY_READY before enabling the network interface, so
>> the link is down. After resuming from hibernation, the phydev->state is
>> forcibly set to PHY_UP in mdio_bus_phy_restore(), and the link becomes up.
>>
>> This patch adds a new convenient function to check whether the PHY is in
>> a started state, and expects to solve the issue by changing phydev->state
>> to PHY_UP and calling phy_start_machine() only when the PHY is started.
>>
>> Suggested-by: Heiner Kallweit <hkallweit1@...il.com>
>> Signed-off-by: Kunihiko Hayashi <hayashi.kunihiko@...ionext.com>
>> ---
>>  drivers/net/phy/phy.c        |  2 +-
>>  drivers/net/phy/phy_device.c |  9 +++++----
>>  include/linux/phy.h          | 10 ++++++++++
>>  3 files changed, 16 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c
>> index 1d73ac3..f484d03 100644
>> --- a/drivers/net/phy/phy.c
>> +++ b/drivers/net/phy/phy.c
>> @@ -670,7 +670,7 @@ void phy_stop_machine(struct phy_device *phydev)
>>  	cancel_delayed_work_sync(&phydev->state_queue);
>>  
>>  	mutex_lock(&phydev->lock);
>> -	if (phydev->state > PHY_UP && phydev->state != PHY_HALTED)
>> +	if (phy_is_started(phydev))
>>  		phydev->state = PHY_UP;
>>  	mutex_unlock(&phydev->lock);
>>  }
>> diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
>> index ab33d17..2c39717 100644
>> --- a/drivers/net/phy/phy_device.c
>> +++ b/drivers/net/phy/phy_device.c
>> @@ -309,10 +309,11 @@ static int mdio_bus_phy_restore(struct device *dev)
>>  		return ret;
>>  
>>  	/* The PHY needs to renegotiate. */
>> -	phydev->link = 0;
>> -	phydev->state = PHY_UP;
>> -
>> -	phy_start_machine(phydev);
>> +	if (phy_is_started(phydev)) {
>> +		phydev->link = 0;
>> +		phydev->state = PHY_UP;
>> +		phy_start_machine(phydev);
>> +	}
> 
> Don't you need some of these steps to be performed under phydev->lock
> being held? See comment below.
> 
Yes, locking should be done. The old code sets phydev->state
w/o holding the lock, I'd says this was wrong.

>>  
>>  	return 0;
>>  }
>> diff --git a/include/linux/phy.h b/include/linux/phy.h
>> index 3ea87f7..c194b45 100644
>> --- a/include/linux/phy.h
>> +++ b/include/linux/phy.h
>> @@ -898,6 +898,16 @@ static inline bool phy_is_pseudo_fixed_link(struct phy_device *phydev)
>>  }
>>  
>>  /**
>> + * phy_is_started - Convenience function for testing whether a PHY is in
>> + * a started state
>> + * @phydev: the phy_device struct
>> + */
>> +static inline bool phy_is_started(struct phy_device *phydev)
>> +{
> 
> An assert with the phydev->lock mutex being held here would greatly
> help, because otherwise this is possibly racy.
> 
Have a look at __phy_resume() to see what is meant with this comment.

>> +	return phydev->state >= PHY_UP && phydev->state != PHY_HALTED;
>> +}
>> +
>> +/**
>>   * phy_write_mmd - Convenience function for writing a register
>>   * on an MMD on a given PHY.
>>   * @phydev: The phy_device struct
>>
>