Re: [PATCH v3 2/3] firmware: arm_scmi: Add Quirks framework

[Cristian Marussi <cristian.marussi@arm.com>]

On Wed, Apr 30, 2025 at 12:36:40PM +0100, Marc Zyngier wrote:
> On Wed, 30 Apr 2025 08:54:06 +0100,
> Cristian Marussi <cristian.marussi@arm.com> wrote:
> > 
> > On Tue, Apr 29, 2025 at 03:11:07PM +0100, Cristian Marussi wrote:
> > > Add a common framework to describe SCMI quirks and associate them with a
> > > specific platform or a specific set of SCMI firmware versions.
> > > 
> > > All the matching SCMI quirks will be enabled when the SCMI core stack
> > > probes and after all the needed SCMI firmware versioning information was
> > > retrieved using Base protocol.
> > > 
> > > Tested-by: Johan Hovold <johan+linaro@kernel.org>
> > > Signed-off-by: Cristian Marussi <cristian.marussi@arm.com>
> > 
> > Hi
> > 
> > just a quick remarks that a bot spotted the usage of __VA_OPT__ which is
> > only available since GCC >= 8.0 :< ... so I will probably revert to use the
> > previous, less clean, mechanism to build the NULL terminated array in
> > which the compats array WILL HAVE to be explicitly NULL terminated when
> > provided (even the empty ones...)
> 
> See 20250407094116.1339199-1-arnd@kernel.org, which is slated for
> 6.16. The TL;DR is that GCC 8.1 and binutils 2.30 should be the
> minimal versions from 6.16 onwards.
> 
> So it's probably not worth using ugly hacks that will eventually be
> reverted.

Great news, thanks for the heads-up !

Thanks,
Cristian