On Tue, 25 Jan 2022 12:25:47 +0000,
Mark Brown <broonie@xxxxxxxxxx> wrote:
>
> [1 <text/plain; us-ascii (7bit)>]
> On Tue, Jan 25, 2022 at 11:27:55AM +0000, Marc Zyngier wrote:
> > Mark Brown <broonie@xxxxxxxxxx> wrote:
>
> > > + if (IS_ENABLED(CONFIG_ARM64_SME) && cpus_have_final_cap(ARM64_SME))
>
> > Please drop the IS_ENABLED(). We purposely avoid conditional
> > compilation in KVM in order to avoid bitrot, and the amount of code
> > you save isn't significant. Having a static key is more than enough to
> > avoid runtime costs.
>
> Sure, I wanted to be extra careful here as this is all in hot paths and
> going to get moved elsewhere when we have real guest support.
>
> > > + if (IS_ENABLED(CONFIG_ARM64_SME) && cpus_have_final_cap(ARM64_SME) &&
> > > + cpus_have_final_cap(ARM64_HAS_FGT)) {
> > > + val = read_sysreg_s(SYS_HFGRTR_EL2);
> > > + val &= ~(HFGxTR_EL2_nTPIDR_EL0_MASK |
> > > + HFGxTR_EL2_nSMPRI_EL1_MASK);
> > > + write_sysreg_s(val, SYS_HFGRTR_EL2);
> > > +
> > > + val = read_sysreg_s(SYS_HFGWTR_EL2);
> > > + val &= ~(HFGxTR_EL2_nTPIDR_EL0_MASK |
> > > + HFGxTR_EL2_nSMPRI_EL1_MASK);
> > > + write_sysreg_s(val, SYS_HFGWTR_EL2);
> > > + }
>
> > If the CPUs do not have FGT, what provides the equivalent trapping?
>
> Nothing for nVHE mode.
That's what I feared.
>
> > If FGT is mandatory when SME exists, then you should simplify the
> > condition.
>
> OK, I'll remove the defensiveness here. FGT is mandatory from v8.6 and
> SME is a v9 feature so people shouldn't build a SME implementation that
> lacks FGT.
Can you then please make it that SME doesn't get enabled at all if FGT
isn't present? It would also be good to have a clarification in the
architecture that it isn't allowed to build SME without FGT (specially
given that v9.0 is congruent to v8.5, and thus doesn't have FGT).
Thanks,
M.
--
Without deviation from the norm, progress is not possible.
