netdev - Re: [BUG] fec mdio times out under system stress

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, 11 Aug 2019 18:03:04 +0200
From:   Andrew Lunn <andrew@...n.ch>
To:     Vladimir Oltean <olteanv@...il.com>
Cc:     Russell King - ARM Linux admin <linux@...linux.org.uk>,
        Florian Fainelli <f.fainelli@...il.com>,
        netdev <netdev@...r.kernel.org>,
        Hubert Feurstein <h.feurstein@...il.com>,
        Fabio Estevam <festevam@...il.com>,
        linux-arm-kernel@...ts.infradead.org,
        Heiner Kallweit <hkallweit1@...il.com>
Subject: Re: [BUG] fec mdio times out under system stress

> I think a better question is why is the FEC MDIO controller configured
> to emit interrupts anyway (especially since the API built on top does
> not benefit in any way from this)? Hubert (copied) sent an interesting
> email very recently where he pointed out that this is one of the main
> sources of jitter when reading the PTP clock on a Marvell switch
> attached over MDIO.

Hi Vladimir

One reason is runtime power management.

For a write operation, you could set it going and return
immediately. Many drivers do, and then when the next read/write
operation comes along, they poll in a loop waiting for the device to
go idle, if it was still busy from the last operation.

However, FEC supports runtime PM. When an MDIO read/write call is
made, it calls runtime PM functions to indicate the device is active.
Once it has completed the MDIO transaction, it calls runtime PM
functions to indicate the device is inactive. These transitions can
cause clocks to be enabled/disabled. If we don't wait around for the
operation to complete, the clock could be disabled too early, and bad
things would happen.

You could replace the interrupt with a sleeping poll, but my guess
would be, that has more jitter than using an interrupt.

      Andrew