On Wed, Jul 21, 2021 at 08:32:07AM +0900, Kuninori Morimoto wrote:
Why do we need these changes? I'm not wild about a new generic binding replacing an existing one which only has 2 or 3 users IIRC. Plus there's already the Renesas variant. (On the flip side, only a few users, easier to deprecate the old binding.)
Sorry I don't understand
- Who is "2 or 3 users" ?
Just that there's not that many users of the existing audio-graph-card (though it's a bit more than 2 or 3 and it's newer stuff rather than old).
- What is "Renesas variant" ?
I think that's the rsrc-card though that got removed. There's also the Tegra audio graph card though.
audio-graph-card2 is based on audio-graph-card, but different driver not minor variant. Becase these are different, it can't keep compatibility. This is the reason why we need audio-graph-card2 instead of expanding audio-graph-card.
I think what Rob is looking for here is a more detailed description of what the problems are with the existing binding that require a new binding - what's driving these big changes? TBH this is part of why I've been holding off on review, I need to get my head round why we can't fix the existing binding in place.
I also would like to see the graph card replace the simple card binding. Surely it can handle the 'simple' case too.
Do you mean you want to merge audio-graph-card and simple-card DT binding ?? audio-graph-card and simple-card are different drivers.
It's more about making sure that new users that currently use simple-card are using audio-graph-card instead - we need to keep simple-card around for existing users (or at least the binding but probably it's more effort than it's worth to merge the binding parsing code elsewhere) but we should be avoiding adding new users of it. I've been pushing people to use audio-graph-card for a while, TBH we should probably just go ahead and flag simple-card as deprecated in the binding now since I don't think there's any reason anyone is forced to use it at this point.