[PATCH] ARM: move firmware_ops to drivers/firmware

[Catalin Marinas]

On Mon, Nov 18, 2013 at 05:00:32PM +0000, Stephen Warren wrote:
> On 11/17/2013 08:59 AM, Catalin Marinas wrote:
> > On 17 November 2013 08:49, Alexandre Courbot <acourbot at nvidia.com> wrote:
> >> The ARM tree includes a firmware_ops interface that is designed to
> >> implement support for simple, TrustZone-based firmwares but could
> >> also cover other use-cases. It has been suggested that this
> >> interface might be useful to other architectures (e.g. arm64) and
> >> that it should be moved out of arch/arm.
> > 
> > NAK. I'm for code sharing with arm via common locations but this API
> > goes against the ARMv8 firmware standardisation efforts like PSCI,
> > encouraging each platform to define there own non-standard interface.
> 
> Surely PSCI is *an* implementation of firmware_ops?
> 
> Couldn't firmware_ops be relevant to non-ARM architectures too?

There are similarities but you don't ask other architectures to
implement an l2x0_init function. If we find other things we want to
describe in here, does this structure become a pool of function pointers
to be shared by other architectures? What's the common functionality
that you want to place in this structure?

> If so, that would support my previous point; we're presumably not
> requiring non-ARM architectures to implement PSCI?

So you think non-ARM architectures could make use of the firmware_ops?

> On a practical note, unless ARM mandates by ARM architecture licensing
> condition that mechanisms other than PSCI are not allowed, then they're
> going to exist even if the upstream Linux community doesn't like it.
> History has certainly shown that.

I'm pretty sure they will exist, as there is tons of kernel code in
production devices that never reach mainline. And I already stated a few
times, this interface can be extended, I just don't want one per SoC
just because some vendors want to use different SMC numbers.

-- 
Catalin