netdev - Re: [PATCH net] net: phy: fix phy_init

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]

Date:   Sun, 23 Dec 2018 18:23:18 +0100
From:   Heiner Kallweit <hkallweit1@...il.com>
To:     Andrew Lunn <andrew@...n.ch>
Cc:     Florian Fainelli <f.fainelli@...il.com>,
        David Miller <davem@...emloft.net>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Re: [PATCH net] net: phy: fix phy_init_hw fixup handling

On 23.12.2018 18:07, Andrew Lunn wrote:
> On Sun, Dec 23, 2018 at 03:00:26PM +0100, Heiner Kallweit wrote:
>> Currently we return immediately if callback config_init isn't defined.
>> This prevents the fixups from being executed. I see no dependency
>> between fixups and config_init, therefore change the function to
>> run the fixups also if config_init isn't defined.
>>
>> Fixes: 2f5cb43406d0 ("phylib: Properly reinitialize PHYs after hibernation")
>> Signed-off-by: Heiner Kallweit <hkallweit1@...il.com>
> 
> Hi Heiner
> 
> Is this a real fix? It seems like it has been like this forever. Do
> you know of a PHY which is actually broken?
> 
Right, the current behavior has been there forever. I'm not aware of
any concrete case. I went for "net" because the current code could
break a device.

> I think the change does make sense, i just don't know if it should be
> a fix and included in stable. It might be better to wait until
> net-next opens again.
> 
Would be fine with me too. Maybe David can advise whether a fix for
a potential issue w/o known case qualifies for net.

>> ---
>>  drivers/net/phy/phy_device.c | 12 +++++++-----
>>  1 file changed, 7 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
>> index e10ac6075..07b1e6751 100644
>> --- a/drivers/net/phy/phy_device.c
>> +++ b/drivers/net/phy/phy_device.c
>> @@ -1035,20 +1035,22 @@ int phy_init_hw(struct phy_device *phydev)
>>  	/* Deassert the reset signal */
>>  	phy_device_reset(phydev, 0);
>>  
>> -	if (!phydev->drv || !phydev->drv->config_init)
>> +	if (!phydev->drv)
>>  		return 0;
>>  
>>  	if (phydev->drv->soft_reset)
>>  		ret = phydev->drv->soft_reset(phydev);
>> -
>> -	if (ret < 0)
>> +	if (ret)
>>  		return ret;
>>  
>>  	ret = phy_scan_fixups(phydev);
>> -	if (ret < 0)
>> +	if (ret)
>>  		return ret;
> 
> These changes from < 0 to any value other than zero should be in a
> separate patch. This is particularly true for a fix, where we want
> fixes to be as small as possible. Statistics show fixes more often
> break stuff than normal development, because they get less testing.
> So we don't really want to make such a change in a fix for stable.
>   
Agree. I'll provide a v2 then for either net or net-next (once it's
open again).

> Thanks
> 	Andrew
> 
>> -	return phydev->drv->config_init(phydev);
>> +	if (phydev->drv->config_init)
>> +		ret = phydev->drv->config_init(phydev);
>> +
>> +	return ret;
>>  }
>>  EXPORT_SYMBOL(phy_init_hw);
>>  
>> -- 
>> 2.20.1
>>
>