linux-kernel - Re: [RFC PATCH net] 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: <325a028e-ce95-1055-1b59-ed6120eeb229@gmail.com>
Date:   Fri, 30 Nov 2018 07:20:27 +0100
From:   Heiner Kallweit <hkallweit1@...il.com>
To:     Kunihiko Hayashi <hayashi.kunihiko@...ionext.com>,
        Florian Fainelli <f.fainelli@...il.com>
Cc:     Andrew Lunn <andrew@...n.ch>,
        "David S. Miller" <davem@...emloft.net>, netdev@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH net] net: phy: fix the issue that netif always links
 up after resuming

On 30.11.2018 05:37, Kunihiko Hayashi wrote:
> Hi Heiner Florian,
> 
> Thank you for your comments.
> 
> On Thu, 29 Nov 2018 16:37:48 -0800 <f.fainelli@...il.com> wrote:
> 
>>
>>
>> On 11/29/2018 2:47 PM, Heiner Kallweit wrote:
>>> On 29.11.2018 09:12, 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 expects to solve the issue by changing phydev->state to PHY_UP
>>>> only when the link is up.
>>>>
>>>> Signed-off-by: Kunihiko Hayashi <hayashi.kunihiko@...ionext.com>
>>>> ---
>>>>  drivers/net/phy/phy_device.c | 6 ++++--
>>>>  1 file changed, 4 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
>>>> index ab33d17..d5bba0f 100644
>>>> --- a/drivers/net/phy/phy_device.c
>>>> +++ b/drivers/net/phy/phy_device.c
>>>> @@ -309,8 +309,10 @@ static int mdio_bus_phy_restore(struct device *dev)
>>>>  		return ret;
>>>>  
>>>>  	/* The PHY needs to renegotiate. */
>>>> -	phydev->link = 0;
>>>> -	phydev->state = PHY_UP;
>>>> +	if (phydev->link) {
>>>> +		phydev->link = 0;
>>>> +		phydev->state = PHY_UP;
>>>> +	}
>>>>  
>>> Thanks for reporting. I agree that it isn't right to unconditionally set
>>> PHY_UP, because we don't know whether the PHY was started before
>>> hibernation. However I don't think using phydev->link as criteria is
>>> right. Example would be: PHY was started before hibernation, but w/o link.
>>> In this case we want to set PHY_UP to start an aneg, because a cable may
>>> have been plugged in whilst system was sleeping.
> 
> Indeed. I didn't consider the case that the PHY was started but a cable was
> unplugged before hibernation.
> 
>>> So I think, similar to phy_stop_machine, we should use state >= UP and
>>> state != HALTED as criteria, and also phy_start_machine() would need to
>>> be called only if this criteria is met.
>>>
>>> It may make sense to add a helper for checking whether PHY is in a
>>> started state (>=UP && !=HALTED), because we need this in more than
>>> one place.
>>
>> Agreed, that would make sense.
> 
> I agree, too.
> I'll try this in v2 patch that changes the PHY state to PHY_UP and calls
> phy_start_machine(), only when the PHY was started before hibernation.
> If I understand correctly, it will be like that:
> 
> 	phydev->link = 0;
Even this may go into the if clause. If PHY isn't started then
phydev->link should be 0 anyway.

> 	if (phy_is_started(phydev)) {
> 		phydev->state = PHY_UP;
> 		phy_start_machine(phydev);
> 	}
> 
Yes, this is what was meant. Thanks.

> Thank you,
> 
> ---
> Best Regards,
> Kunihiko Hayashi
> 
> 
>