Re: [PATCH v4 0/7] firmware: smccc: Add basic SMCCC v1.2 + ARCH_SOC_ID support

[Will Deacon]

On Thu, May 21, 2020 at 11:06:23AM +0200, Arnd Bergmann wrote:
> On Thu, May 21, 2020 at 10:11 AM Sudeep Holla <sudeep.holla@xxxxxxx> wrote:
> > Indeed, it is also last patch in the series. However if Arnd is happy
> > with the sysfs names, we can move to generic code later without breaking
> > anything.
> >
> > We need not revert or drop it now. I will leave that to you or Arnd to
> > decide. Just that it may be too late to get acks for all the soc sysfs
> > drivers in time for v5.8
> >
> > I am fine if you want to drop the last patch.
> 
> Ok, let's drop that patch then and make sure we do something that
> everyone is happy with later on. I'm already in favor of adding
> a more reliable soc_device instance based on this, but we need to
> be sure we don't screw up the contents of the attributes when we
> can't change them later.
> 
> > > > >> drivers/firmware/smccc/smccc.c:14:13: warning: no previous prototype for function 'arm_smccc_version_init' [-Wmissing-prototypes]
> > > > void __init arm_smccc_version_init(u32 version, enum arm_smccc_conduit conduit)
> > > > ^
> > > > drivers/firmware/smccc/smccc.c:14:1: note: declare 'static' if the
> > > > function is not intended to be used outside of this translation unit
> > > > void __init arm_smccc_version_init(u32 version, enum arm_smccc_conduit conduit)
> > >
> > > I saw that when I applied the patches, but since the function is called from
> > > another compilation unit (psci/psci.o), I just ignored it as we have loads
> > > of these already and it only screams if you build with W=1.
> > >
> >
> > /me confused. Do you need the fix for this warning or you are happy to ignore?
> 
> I want a fix for that, as I hope we can eventually turn this warning on by
> default and stop playing whack-a-mole when they come up. Most of these
> warnings are harmless, but occasionally the prototypes don't match exactly
> and cause real bugs depending on the configuration, and ensuring both
> sides include a common header file is an easy way to make it work
> more reliably.
> 
> Note that the warning should come up for either W=1 or C=1, and I also
> think that
> new code should generally be written sparse-clean and have no warnings with
> 'make C=1' as a rule.

Fair enough. Is anybody working on a tree-wide sweep for this, like we've
done for other things such as zero-length arrays? If so, I can start
enforcing this in the arch code as well (I haven't been so far, even though
I do run sparse on every commit).

Anyway, I've dropped the last patch from the branch, and we can put a fix
for the missing prototype on top.

Will