[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 05 Nov 2007 22:04:11 +1100
From: Benjamin Herrenschmidt <benh@...nel.crashing.org>
To: Stefan Roese <sr@...x.de>
Cc: Olof Johansson <olof@...om.net>, netdev@...r.kernel.org,
linuxppc-dev@...abs.org, jwboyer@...ux.vnet.ibm.com
Subject: Re: [PATCH] net: Add 405EX support to new EMAC driver
On Mon, 2007-11-05 at 10:19 +0100, Stefan Roese wrote:
>
> > Somewhat yeah. There are subtle variations here or there we haven't
> > totally indenfified... It might be a better option in our case here
> to
> > add "has-mdio" to the rgmii nodes indeed.
>
> So how exactly do you want me to handle this (I'm still new to this
> device
> tree stuff, so please bear with me)? Like this?
>
> RGMII0: emac-rgmii@...01000 {
> device_type = "rgmii-interface";
> compatible = "ibm,rgmii-405ex",
> "ibm,rgmii";
> reg = <ef601000 8>;
> has-mdio;
> };
>
The above.
Properties without values are typically used for such "flags". I'll
fixup the driver to also take that for the inverted STACR and will post
a patch fixing that up asap.
> It's not only the OC bit-flip on AXON, but also the different STACR
> register
> layout for read/write op-codes (STAOPC). This seems to be the same on
> all new
> EMAC core's like on AXON, 440EPx/GRx and 405EX. So "stacr-oc-inverted"
> is not
> enough here. This is what is needed for 440SPe, which "only" has the
> bit-flip
> and the "old" STAOPC layout.
Ok.
> So perhaps most flexible would be to add individual properties,
> like "stacr-oc-inverted" and "stacr-staopc-19-20". What do you think?
> And
> again the additional question: Should the be added as an new property
> or
> added to the compatible property?
That's always the main question imho ... When it gets nasty like that I
tend to think the compatible property is a good compromise. It's mostly
a matter of taste. Unless you can come up with some more pleasant way to
do it... maybe a stacr-type property with multiple values but it's
really not worth complicating things when a compatible property will do
the job just fine. In that case, it's not really a "feature" of a given
implementation, but more about subtle differences between
implementations.
Ben.
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists