On Fri, 2015-09-25 at 16:28 +0100, Catalin Marinas wrote: > > This only works as long as they target an existing driver with prior DT > support (usually with reviewed bindings). If they have a new driver and > only ACPI in mind, I'm pretty sure we'll end up with new insane things. > That's why we need some form of _DSD properties review and "compatible" > is one such property.
Sure, that makes a lot of sense. My main concern is that we don't end up with gratuitously *different* property sets for DT vs. ACPI. That way we end up with either two separate drivers, or abstracting the core of the driver out and having two bindings for it (much as we do for PCI vs. platform/etc for some devices already). We don't want that pain where we can avoid it. And we don't want people to *have* to hack the kernel driver to migrate to ACPI, if we can avoid it. Sometimes it might be worth being different — if the DT binding is utterly crap, and deserves to be thrown away. In that case, by all means invent a new binding. But if we're going to accept the pain of having multiple bindings, why not make the *new* one work via DT too anyway. I'd be happy to see the existing DT bindings put through the nascent _DSD review process — such as it is — and into the database. One at a time on a case-by-case basis as they get used, perhaps. -- dwmw2
smime.p7s
Description: S/MIME cryptographic signature
