netdev - Re: create drivers/net/mdio and move mdio drivers into it

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: <6cec7d67-3e99-e97d-3720-2ae6884ce150@gmail.com>
Date:   Mon, 20 Feb 2017 14:30:10 -0800
From:   Florian Fainelli <f.fainelli@...il.com>
To:     YUAN Linyu <Linyu.Yuan@...atel-sbell.com.cn>,
        "David S . Miller" <davem@...emloft.net>,
        Andrew Lunn <andrew@...n.ch>
Cc:     "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "cugyly@....com" <cugyly@....com>
Subject: Re: create drivers/net/mdio and move mdio drivers into it



On 02/19/2017 10:26 PM, YUAN Linyu wrote:
> 
> 
>> -----Original Message-----
>> From: Florian Fainelli [mailto:f.fainelli@...il.com]
>> Sent: Monday, February 20, 2017 2:16 PM
>> To: YUAN Linyu; David S . Miller; Andrew Lunn
>> Cc: netdev@...r.kernel.org; cugyly@....com
>> Subject: Re: create drivers/net/mdio and move mdio drivers into it
>>
>>
>>
>> On 02/19/2017 10:10 PM, YUAN Linyu wrote:
>>>
>>>
>>>> -----Original Message-----
>>>> From: Florian Fainelli [mailto:f.fainelli@...il.com]
>>>> Sent: Monday, February 20, 2017 1:42 PM
>>>> To: YUAN Linyu; David S . Miller; Andrew Lunn
>>>> Cc: netdev@...r.kernel.org; cugyly@....com
>>>> Subject: Re: create drivers/net/mdio and move mdio drivers into it
>>>>> 3. another idea is bind mdio device to network device
>>>>
>>>> You would have to be more specific about what you want to do here. If
>>>> the MDIO device is e.g: a switch, what we recommend doing is provide a
>>>> fixed-link node that describes how the Ethernet MAC and the switch's
>>>> CPU/management ports are connected (that way the MAC always "sees"
>> the
>>>> link as UP, running with a specific speed and duplex).
>>>>
>>> Yes, some system will configured the phy to fixed speed/duplex at boot time,
>>> no phy driver used in kernel at all.
>>
>> There is always a PHY driver used, either is a dedicated one, or its the
>> generic PHY driver in drivers/net/phy/phy_device.c, even when fixed
>> PHYs/link are used.
>>
>>> If network device know mdio device it used, we can do phy dump through this
>>> mdio device driver.
>>
>> We are not going to accept MDIO device drivers whose only purpose is to
>> allow PHY devices register dumps. Implement a proper PHY driver for
>> these devices, if nothing needs to be done, you just need to call into
>> genphy_* functions, and just override how to do register dumps.
>>
> No, we discuss mdio driver here, not phy driver.
> I mean if a network device know mdio device it used, we use mdio driver to 
> dump any register of phy device even it's driver is not build.

The network device is supposed to attach to a PHY, which in turns has a
backing PHY driver, which is a superset of a MDIO device driver.

If the thing you are after is to be able to do PHY device register dumps
without a PHY driver to assist you with that, then we need to add whole
lot of things:

- have an ioctl() which allows stopping/freezing the PHY state machine
so dumps don't interfere with the PHY state

- have the ability to reliably perform PHY dumps that e.g: require page
programming etc. etc.

Then again, all of this could be done by a PHY driver reliably and in a
way that is consistent and predictable.
-- 
Florian