On Thu, May 23, 2013 at 11:11:12AM -0600, Jason Gunthorpe wrote:

On Thu, May 23, 2013 at 12:01:11PM -0400, Jason Cooper wrote:

quoted

quoted

quoted

+	/* Kirkwood resets some registers on gated clocks. Especially
+	 * CLK125_BYPASS_EN must be cleared but is not available on
+	 * all other SoCs/System Controllers using this driver.
+	 */
+	if (of_machine_is_compatible("marvell,kirkwood"))
+		wrlp(mp, PORT_SERIAL_CONTROL1,
+		     rdlp(mp, PORT_SERIAL_CONTROL1) & ~CLK125_BYPASS_EN);

of_machine_is_compatible seems heavy handed, I would expect this to be
based on the compatible string of the ethernet node itself, not the
machine??

Is there a model number variation between IP that needs this and IP that
doesn't?  If not, I'm fine with of_machine_is_compatible().

Well the name 'mv643xx' is a family of system controller SOC's
from ages ago, it seems reasonble to continue the trend and label the
IP variations with the SOC name:

 compatible = "marvell,kirwood,ethernet", "marvell,mv643xx_eth"

Shouldn't it rather be

	compatible = "marvell,kirkwood-eth", "marvell,orion-eth";

I'm inclined to go with of_machine_is_compatible() since the only
concrete difference we know is that the tweak is needed on kirkwood and
nowhere else.

If we had an errata, or a datasheet saying specifically flavor X needs
this and none other does, then we could trigger on the ethernet node
compatible string or a boolean in the node.  But we don't have that...

thx,

Jason.