netdev - Re: [PATCH] net: fec: Fix PHY init after phy_reset_after_clk

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: <b6397b39-c897-6e0a-6bf7-b6b24908de1a@denx.de>
Date:   Thu, 3 Sep 2020 23:36:39 +0200
From:   Marek Vasut <marex@...x.de>
To:     Andrew Lunn <andrew@...n.ch>
Cc:     netdev@...r.kernel.org,
        Christoph Niedermaier <cniedermaier@...electronics.com>,
        "David S . Miller" <davem@...emloft.net>,
        NXP Linux Team <linux-imx@....com>,
        Richard Leitner <richard.leitner@...data.com>,
        Shawn Guo <shawnguo@...nel.org>
Subject: Re: [PATCH] net: fec: Fix PHY init after phy_reset_after_clk_enable()

On 9/3/20 11:00 PM, Andrew Lunn wrote:
> On Thu, Sep 03, 2020 at 10:27:12PM +0200, Marek Vasut wrote:
>> The phy_reset_after_clk_enable() does a PHY reset, which means the PHY
>> loses its register settings. The fec_enet_mii_probe() starts the PHY
>> and does the necessary calls to configure the PHY via PHY framework,
>> and loads the correct register settings into the PHY. Therefore,
>> fec_enet_mii_probe() should be called only after the PHY has been
>> reset, not before as it is now.
> 
> I think this patch is related to what Florian is currently doing with
> PHY clocks.

Which is what ? Details please.

> I think a better fix for the original problem is for the SMSC PHY
> driver to control the clock itself. If it clk_prepare_enables() the
> clock, it knows it will not be shut off again by the FEC run time
> power management.

The FEC MAC is responsible for generating the clock, the PHY clock are
not part of the clock framework as far as I can tell.

> All this phy_reset_after_clk_enable() can then go away.

I'm not sure about that. Also, this is a much simpler fix which can be
backported easily.