Re: [PATCH V5 1/6] SLIMbus: Device management on SLIMbus

[Mark Brown <broonie@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 1565 bytes --]

On Thu, Apr 28, 2016 at 02:33:41PM +0200, Arnd Bergmann wrote:
> On Thursday 28 April 2016 12:53:37 Mark Brown wrote:
> > On Thu, Apr 28, 2016 at 12:00:26PM +0200, Arnd Bergmann wrote:

> > > This looks like an artifact of ancient pre-DT times. I'd say kill it off before
> > > someone starts using it.

> > Not every architecture uses DT, and even on architectures with DT
> > support it isn't always the only firmware.  In this specific case it's
> > questionable how many people are going to implement Slimbus at this
> > point but in general insisting that we go DT only doesn't seem great.

> Nothing wrong with adding support for manual board files later if
> we have a good reason for it, but at the moment, this seems completely
> ARM/ARM64 specific.

It's not in theory, but in practice nobody other that Qualcomm is ever
likely to release a controller.

> I don't foresee mobile phones with ACPI using this subsystem, but even
> if we got them, it would be a horrible idea to use hardcoded board
> specific tables in a platform file, and we should insist that whatever
> firmware is present has a way to describe the slimbus devices.

Right, in this particular case I don't think it makes a huge difference
but what you were talking about was "ancient pre-DT times" rather than
something specific to this particular case.  That's definitely a thing
that people keep thinking and it's good to push back on it since we do
have non-DT cases to worry about (some architectures, other firmwares,
things like PCI cards with other components on them and so on).

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 473 bytes --]