Re: [RFC PATCH net v3] net: phy: Fix the issue that netif always links up after resuming

[Heiner Kallweit]

On 18.12.2018 07:25, Kunihiko Hayashi wrote:
> Hi Heiner,
> 
> On Mon, 17 Dec 2018 19:43:31 +0100 <hkallweit1@xxxxxxxxx> wrote:
> 
>> On 17.12.2018 19:41, Heiner Kallweit wrote:
>>> On 17.12.2018 07:41, Kunihiko Hayashi wrote:
>>>> Hi,
>>>>
>>>> Gentle ping...
>>>> Are there any comments about changes since v2?
>>>>
>>>> v2: https://www.spinics.net/lists/netdev/msg536926.html
>>>>
>>>> Thank you,
>>>>
>>>> On Mon, 3 Dec 2018 17:22:29 +0900 <hayashi.kunihiko@xxxxxxxxxxxxx> 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.
>>>>>
>>> This convenience function and the related change to phy_stop() are part of
>>> the following already and don't need to be part of your patch.
>>> https://patchwork.ozlabs.org/patch/1014171/
> 
> I see. I'll follow your patch when necessary.
> 
>>>>> Suggested-by: Heiner Kallweit <hkallweit1@xxxxxxxxx>
>>>>> Signed-off-by: Kunihiko Hayashi <hayashi.kunihiko@xxxxxxxxxxxxx>
>>>>> ---
>>>>>
>>>>> Changes since v2:
>>>>>  - add mutex lock/unlock for changing phydev->state
>>>>>  - check whether the mutex is locked in phy_is_started()
>>>>>  
>>>>> Changes since v1:
>>>>>  - introduce a new helper function phy_is_started() and use it instead of
>>>>>    checking link status
>>>>>  - replace checking phydev->state with phy_is_started() in
>>>>>    phy_stop_machine()
>>>>>
>>>>> drivers/net/phy/phy.c        |  2 +-
>>>>>  drivers/net/phy/phy_device.c | 12 +++++++++---
>>>>>  include/linux/phy.h          | 13 +++++++++++++
>>>>>  3 files changed, 23 insertions(+), 4 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;
>>>
>>> I'm wondering whether we need to do this. If the PHY is attached,
>>> then mdio_bus_phy_suspend() calls phy_stop_machine() which does
>>> exactly the same. If the PHY is not attached, then we don't have
>>> to do anything. Therefore I think we just have to do the same as
>>> in mdio_bus_phy_resume():
>>>
>>> if (phydev->attached_dev && phydev->adjust_link)
>>> 	phy_start_machine(phydev);
> 
> Agreed.
> 
> Although the original code changed phydev->state,
> it seems that it's only enough to
> - call phy_stop_machine() in mdio_bus_phy_suspend()
> - call phy_start_machine() in mdio_bus_phy_resume() and mdio_bus_phy_restore()
> if the PHY is attached.
> 
>>> Can you test this?
> 
> I tested your code instead of applying my entire patch, and I confirmed
> that the code solved the issue in my environment.
> 
> Do you make new patch instead of my patch?
> (and I can add Reported-by: for the issue and Tested-by:)
> 
Up to you. It's fine with me if you submit the patch, but I can also do it
and mention you in Reported-by and Tested-by. Just let me know.

> Thank you,
> 
> 
>>>
>> Sorry for the confusion, this comment is related to the next part
>> of your patch.
>>
>>>>>  	mutex_unlock(&phydev->lock);
>>>>>  }
>>>>> diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
>>>>> index ab33d17..4897d24 100644
>>>>> --- a/drivers/net/phy/phy_device.c
>>>>> +++ b/drivers/net/phy/phy_device.c
>>>>> @@ -309,10 +309,16 @@ static int mdio_bus_phy_restore(struct device *dev)
>>>>>  		return ret;
>>>>>  
>>>>>  	/* The PHY needs to renegotiate. */
>>>>> -	phydev->link = 0;
>>>>> -	phydev->state = PHY_UP;
>>>>> +	mutex_lock(&phydev->lock);
>>>>> +	if (phy_is_started(phydev)) {
>>>>> +		phydev->state = PHY_UP;
>>>>> +		mutex_unlock(&phydev->lock);
>>>>> +		phydev->link = 0;
>>>>> +		phy_start_machine(phydev);
>>>>> +	} else {
>>>>> +		mutex_unlock(&phydev->lock);
>>>>> +	}
>>>>>  
>>>>> -	phy_start_machine(phydev);
>>>>>  
>>>>>  	return 0;
>>>>>  }
>>>>> diff --git a/include/linux/phy.h b/include/linux/phy.h
>>>>> index 3ea87f7..dd21537 100644
>>>>> --- a/include/linux/phy.h
>>>>> +++ b/include/linux/phy.h
>>>>> @@ -898,6 +898,19 @@ 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
>>>>> + *
>>>>> + * The caller must have taken the phy_device mutex lock.
>>>>> + */
>>>>> +static inline bool phy_is_started(struct phy_device *phydev)
>>>>> +{
>>>>> +	WARN_ON(!mutex_is_locked(&phydev->lock));
>>>>> +	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
>>>>> -- 
>>>>> 2.7.4
>>>>
>>>> ---
>>>> Best Regards,
>>>> Kunihiko Hayashi
>>>>
>>>>
>>>>
>>>
> 
> ---
> Best Regards,
> Kunihiko Hayashi
> 
> 
>