On Wed, Aug 22, 2012 at 12:55:25PM +0100, Lee Jones wrote:

quoted

I *have* asked you to communicate more clearly about what you're doing
but that doesn't mean to stop sending code, it means to have clearer
words around what you're sending. 

That's not how I interpreted your words:

"What you can do here is to commmunicate about what you're doing more.
Don't just think about the code, think about the communication
surrounding the code - this is the core of the issue."

Just to be clear I'd include things like commit messages, cover letters
and so on in the general area of communication.

I think it's great that you like the idea and want to extend that
functionality to other MFDs which perhaps don't support DT, or the
ones that do but don't want to provide compatible strings or device
nodes for all the MFD's child devices. But that is all we're doing
here. There was no breakage. It served a purpose and it worked well.
So well in fact that you've now provided the intended functionality
to other devices.

I'm looking at this from the point of view of adding the compatible
strings and then finding that the core starts remapping things in the
case where you're probing from DT - and this behaviour will vary
depending on the device tree that the user is using so the driver can't
even make a decision based on if device tree is being used by the
system.

You only noticed it 2 days ago and I had a patch ready to go yesterday.
I'm confused because I don't understand why would you even complain about
it if you intended to work on it yourself? Surely, "Ah, I see an issue with
xyz, patch to follow." Would have been more appropriate, instead of 
complaining about it, then I go and waste my time trying to fix something
you intend on fixing yourself.

I didn't originally intend to do anything, if I had done I'd just have
sent a patch as you say.  Originally I'd just noticed it as being an
awkwardly designed interface at the wrong abstraction layer, it was only
later on that I realised how it could blow up.