linux-kernel - Re: [PATCH net-next v3 5/8] net: phy: mscc: 1588 block initialization

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:   Mon, 22 Jun 2020 15:07:05 +0200
From:   Antoine Tenart <antoine.tenart@...tlin.com>
To:     Andrew Lunn <andrew@...n.ch>
Cc:     davem@...emloft.net, f.fainelli@...il.com, hkallweit1@...il.com,
        richardcochran@...il.com, alexandre.belloni@...tlin.com,
        UNGLinuxDriver@...rochip.com, netdev@...r.kernel.org,
        linux-kernel@...r.kernel.org, thomas.petazzoni@...tlin.com,
        allan.nielsen@...rochip.com, foss@...il.net
Subject: Re: [PATCH net-next v3 5/8] net: phy: mscc: 1588 block initialization

Hi Andrew,

Quoting Andrew Lunn (2020-06-20 17:10:01)
> On Fri, Jun 19, 2020 at 02:22:57PM +0200, Antoine Tenart wrote:
> > From: Quentin Schulz <quentin.schulz@...tlin.com>
> > 
> > This patch adds the first parts of the 1588 support in the MSCC PHY,
> > with registers definition and the 1588 block initialization.
> > 
> > Those PHYs are distributed in hardware packages containing multiple
> > times the PHY. The VSC8584 for example is composed of 4 PHYs. With
> > hardware packages, parts of the logic is usually common and one of the
> > PHY has to be used for some parts of the initialization. Following this
> > logic, the 1588 blocks of those PHYs are shared between two PHYs and
> > accessing the registers has to be done using the "base" PHY of the
> > group. This is handled thanks to helpers in the PTP code (and locks).
> > We also need the MDIO bus lock while performing a single read or write
> > to the 1588 registers as the read/write are composed of multiple MDIO
> > transactions (and we don't want other threads updating the page).
> 
> Locking sounds complex. I assume LOCKDEP was your friend in getting
> this correct and deadlock free.

I agree, locking is not straight forward. But it's actually not that
complex:

- The MDIO bus lock is used for all TS/PHC register access.
- There is one lock for PHC operations and one for timestamping
  operations. The two are never used in the same function. We could use
  the same lock; introducing more waiting.
- There is one shared lock for GPIO operations. It is only used in
  PHC functions, in two places.

And I realized I can remove the locks from vsc8584_ptp_init, as PHC/TS
helpers are not registered until the PHY is initialized.

Thanks,
Antoine

-- 
Antoine Ténart, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com