Re: [PATCH v7 0/2] arm64: dts: qcom: extend the register range for ICE on sm8[56]50

[Bjorn Andersson]

On Fri, Oct 04, 2024 at 04:13:15PM GMT, Bartosz Golaszewski wrote:
> On Fri, Oct 4, 2024 at 3:56 PM Bjorn Andersson
> <quic_bjorande@xxxxxxxxxxx> wrote:
> >
> > On Tue, Oct 01, 2024 at 10:35:29AM +0200, Bartosz Golaszewski wrote:
> > > The following changes extend the register range for ICE IPs on sm8550
> > > and sm8650 in order to cover the registers used for wrapped key support
> > > on these platforms.
> > >
> > > Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@xxxxxxxxxx>
> > >
> > > Changes in v7:
> > > - bring the ICE register range up to its full size of 0x18000
> > > - Link to v6: https://lore.kernel.org/r/20240906-wrapped-keys-dts-v6-0-3f0287cf167e@xxxxxxxxxx
> > >
> > > Changes in v6:
> > > - split out the DT changes into a separate series
> >
> > Bartosz, this strategy of "let's split things such that the maintainers
> > can't see the full picture" is just BS. It needs to stop.
> >
> 
> You're exaggerating, I'm not doing anything like this. You're still
> Cc'ed on the single big series containing the code changes for wrapped
> keys. The full picture is over there.
> 
> > Now you will argue that these patches stands on their own, and that
> 
> Yes, that's precisely what I'm going to say, because it's true.

Good, I fully agree with you here.

> 0x18000 is the true register size (as per QCom docs) for ICE on
> sm8[56]50 and sa8775p too and that alone warrants this change. If we
> can get the DTS changes out of the way of wrapped keys, then that's
> just a bonus.
> 
> > might be a valid case, but the argumentation you're making in the commit
> > message clearly ties them to the code changes you're making somewhere
> > else.
> >
> 
> Are you referring to "(...) registers used for wrapped key support on
> these platforms"? I could argue that it just says what the additional
> registers are used for but I can drop the mention of this from the
> message and just say "0x18000 is the true register range for ICE on
> sm8650" if that works for you.
> 

Please update the commit messages of the two patches to not say that
they are enabling HWKM on the platforms, but that you're expanding the
mmio region to cover the whole IP-block. I think it makes sense to
mention that this will allow us to enable HWKM.

Thanks,
Bjorn