[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20160428150705.GA30217@jcartwri.amer.corp.natinst.com>
Date: Thu, 28 Apr 2016 10:07:05 -0500
From: Josh Cartwright <joshc@...com>
To: Nathan Sullivan <nathan.sullivan@...com>
Cc: nicolas.ferre@...el.com, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] net: macb: do not scan PHYs manually
On Thu, Apr 28, 2016 at 09:19:47AM -0500, Nathan Sullivan wrote:
> Since of_mdiobus_register and mdiobus_register will scan automatically,
This is only partially true. of_mdiobus_register() only scans for PHYs
with device tree presence (starting with nodes which specify an address,
then continuing for nodes without an address specifier).
So, this patch will regress any platform currently relying on a PHY
being found/probed even though it has no representation as a device tree
node. I don't have enough history with this driver to make a judgement
as to whether or not there are any users of this functionality.
> do not manually scan for PHY devices in the macb ethernet driver. Doing
> so will result in many nonexistent PHYs on the MDIO bus if the MDIO
> lines are floating or grounded, such as when they are not used.
Just a restatement/elaboration of the problem in an attempt to provide
more context:
The assumption being made is that the phy associated with a given macb
instance exists on that instance's controlled mdio bus. This assumption
doesn't hold true on some dual-ethernet Zynq 7000-series based designs
where both phys exist on the same MDIO bus (for pinning reasons, or
whatever).
MDIO
macb0 ----+----- phy0, addrX
|
+----- phy1, addrY
macb1 ----- {MDIO lines NC, not pinmuxed}
(In this case, the phy associated with macb1 is phy1, which exists on
the MDIO bus controlled by the macb0 instance).
HTH,
Josh
Download attachment "signature.asc" of type "application/pgp-signature" (474 bytes)
Powered by blists - more mailing lists