netdev - Re: [PATCH net-next] net: phy: avoid unneeded MDIO reads in genphy_read

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: <29ddafbb-ad65-f49d-7131-40ebc3083dd0@gmail.com>
Date:   Thu, 25 Apr 2019 07:02:05 +0200
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-next] net: phy: avoid unneeded MDIO reads in
 genphy_read_status

On 25.04.2019 00:41, Andrew Lunn wrote:
> On Wed, Apr 24, 2019 at 09:49:30PM +0200, Heiner Kallweit wrote:
>> Considering that in polling mode each link drop will be latched,
>> settings can't have changed if link was up and is up.
> 
> Hi Heiner
> 
Hi Andrew,

interesting questions.

> What about the case of the PHY performing a downshift?
> 
> Could it be up, then finds a pair fails, so performs a downshift. Does
> it always report link down and then up to the new speed? Or could it
> just shift down? And so we want to read the new speed?
> 
My experience with downshift is that it's part of an extended autoneg
process. The link stays down all the time until both link partners
have agreed on a downshifted speed.

> I suppose it could also perform an upshift? A broken pair comes back
> to life, so it returns to the higher speed?
> 
So far I didn't come across an upshift feature. But also then I'd expect
some autoneg process with the link being down. More likely seems to be
that upshifting requires an explicit autoneg restart.

>    Andrew
> 
Heiner