netdev - Re: [PATCH v2 net-next 1/2] phy: add phy fixup unregister functions

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] [day] [month] [year] [list]

Date:   Fri, 16 Dec 2016 10:05:36 +0800
From:   Dongpo Li <lidongpo@...ilicon.com>
To:     <Woojung.Huh@...rochip.com>, <davem@...emloft.net>,
        <f.fainelli@...il.com>
CC:     <andrew@...n.ch>, <netdev@...r.kernel.org>,
        <UNGLinuxDriver@...rochip.com>
Subject: Re: [PATCH v2 net-next 1/2] phy: add phy fixup unregister functions


On 2016/12/14 23:34, Woojung.Huh@...rochip.com wrote:
>> I just want to commit the unregister patch and found this patch. Good job!
>> But I consider this patch may miss something.
>> If one SoC has 2 MAC ports and each port uses the different network driver,
>> the 2 drivers may register fixup for the same PHY chip with different
>> "run" function because the PHY chip works in different mode.
>> In such a case, this patch doesn't consider "run" function and may cause
>> problem.
>> When removing the driver which register fixup at last, it will remove another
>> driver's fixup.
>> Should this condition be considered and fixed?
> Good point.
> Current phy fixup is independent LIST from phydev structure,
> and, fixup runs in two places of phy_device_register() and phy_init_hw().
> It's not clear that it needs two separate fixup, but it may be good idea to
> pass phy fixup when calling phy_attach() or phy_attach_direct() and
> put it under phydev structure.
> So, fixup can be called at phy_init_hw() per phy device and remove
> When phy detached.
> Welcome any comments.

I rethink this problem and find that the "fixup->bus_id" may be a flag to
distinguish different PHY device.
In such condition, the driver should call "phy_register_fixup/phy_unregister_fixup" directly
instead of "*_for_uid" interface.


    Regards,
    Dongpo

.