On Mon, Jul 13, 2020 at 11:08:15AM +0100, Steven Price wrote: > On 03/07/2020 16:36, Catalin Marinas wrote: > > From: Vincenzo Frascino <vincenzo.frascino@xxxxxxx> > > > > Add the cpufeature and hwcap entries to detect the presence of MTE on > > the boot CPUs (primary and secondary). Any late secondary CPU not > > supporting the feature, if detected during boot, will be parked. > > > > In addition, add the minimum SCTLR_EL1 and HCR_EL2 bits for enabling > > MTE. Without subsequent setting of MAIR, these bits do not have an > > effect on tag checking. > > > > Signed-off-by: Vincenzo Frascino <vincenzo.frascino@xxxxxxx> > > Co-developed-by: Catalin Marinas <catalin.marinas@xxxxxxx> > > Signed-off-by: Catalin Marinas <catalin.marinas@xxxxxxx> > > Cc: Will Deacon <will@xxxxxxxxxx> > > Cc: Suzuki K Poulose <Suzuki.Poulose@xxxxxxx> > > This commit causes the feature bit to be exposed to a guest, but we > don't at this point have any way of handling a guest which attempts to > use MTE. > > This is 'fixed' by the first patch of my KVM MTE series[1], but perhaps > the chunk modifying arch/arm64/kvm/sys_regs.c (see below) should be included here > instead? That way we hide the feature until we're ready for a guest with > MTE support. > > Steve > [1] https://lore.kernel.org/r/20200713100102.53664-2-steven.price@xxxxxxx > > ----8<---- > diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c > index baf5ce9225ce..5ca974c93bd4 100644 > --- a/arch/arm64/kvm/sys_regs.c > +++ b/arch/arm64/kvm/sys_regs.c > @@ -1104,6 +1104,8 @@ static u64 read_id_reg(const struct kvm_vcpu *vcpu, > if (!vcpu_has_sve(vcpu)) > val &= ~(0xfUL << ID_AA64PFR0_SVE_SHIFT); > val &= ~(0xfUL << ID_AA64PFR0_AMU_SHIFT); > + } else if (id == SYS_ID_AA64PFR1_EL1) { > + val &= ~(0xfUL << ID_AA64PFR1_MTE_SHIFT); > } else if (id == SYS_ID_AA64ISAR1_EL1 && !vcpu_has_ptrauth(vcpu)) { > val &= ~((0xfUL << ID_AA64ISAR1_APA_SHIFT) | > (0xfUL << ID_AA64ISAR1_API_SHIFT) | Thanks Steven. I thought this worked on explicitly enabling the CPUID for guests but I think I only checked with an old host kernel which was masking the ID field anyway. I tried it again now and it indeed fails. I'll fold this in. -- Catalin
