Re: MPC8272- Porting HDLC driver from 2.6.14 to 2.6.27- "no_irq_chip" error

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

On Tue, 2009-06-02 at 06:38 +0200, Wolfram Sang wrote:
> I fully agree that the proper way is using the device tree. I can't recall
> exactly why this wasn't an option to me a year ago, but I assume it had
> something to do with this ugly out-of-tree-driver. I wouldn't really argue
> about this patch, I just thought it might be useful as it seems there are
> people trying to do the same for some reason. Then again, maybe it should be
> skipped, as it makes it easier to not use the proper solution (= device tree)
> :)

Whether the driver is in or out of tree shouldn't affect the device-tree
which represents the -devices- on the system regardless of whether a
driver is in-tree or not.

So while some people might flinch, I don't necessarily object to adding
nodes in the device-trees we ship with the kernel for things for which
we don't have an in-tree driver.

It does mean however that the risk of such nodes containing busted
informations that aren't spotted because nobody has the driver to test
is higher but that's a risk you are taking with out of tree drivers
anyway.

Cheers,
Ben.