Nicholas Piggin <npiggin@xxxxxxxxx> writes:
> This bit only applies to hash partitions.
>
> Signed-off-by: Nicholas Piggin <npiggin@xxxxxxxxx>
> ---
> arch/powerpc/kvm/book3s_hv.c | 6 ++++--
> arch/powerpc/kvm/book3s_hv_nested.c | 2 +-
> 2 files changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/arch/powerpc/kvm/book3s_hv.c b/arch/powerpc/kvm/book3s_hv.c
> index c40eeb20be39..2e29b96ef775 100644
> --- a/arch/powerpc/kvm/book3s_hv.c
> +++ b/arch/powerpc/kvm/book3s_hv.c
> @@ -1666,10 +1666,12 @@ static void kvmppc_set_lpcr(struct kvm_vcpu *vcpu, u64 new_lpcr,
>
> /*
> * Userspace can only modify DPFD (default prefetch depth),
> - * ILE (interrupt little-endian) and TC (translation control).
> + * ILE (interrupt little-endian) and TC (translation control) if HPT.
> * On POWER8 and POWER9 userspace can also modify AIL (alt. interrupt loc.).
> */
> - mask = LPCR_DPFD | LPCR_ILE | LPCR_TC;
> + mask = LPCR_DPFD | LPCR_ILE;
> + if (!kvm_is_radix(kvm))
> + mask |= LPCR_TC;
I think in theory there is a possibility that userspace sets the LPCR
while we running Radix and then calls the KVM_PPC_CONFIGURE_V3_MMU ioctl
right after to switch to HPT. I'm not sure if that would make sense but
maybe it's something to consider...
> if (cpu_has_feature(CPU_FTR_ARCH_207S)) {
> mask |= LPCR_AIL;
> /* LPCR[AIL]=1/2 is disallowed */
> diff --git a/arch/powerpc/kvm/book3s_hv_nested.c b/arch/powerpc/kvm/book3s_hv_nested.c
> index b496079e02f7..0e6cf650cbfe 100644
> --- a/arch/powerpc/kvm/book3s_hv_nested.c
> +++ b/arch/powerpc/kvm/book3s_hv_nested.c
> @@ -141,7 +141,7 @@ static void sanitise_hv_regs(struct kvm_vcpu *vcpu, struct hv_guest_state *hr)
> * Don't let L1 change LPCR bits for the L2 except these:
> * Keep this in sync with kvmppc_set_lpcr.
> */
> - mask = LPCR_DPFD | LPCR_ILE | LPCR_TC | LPCR_LD | LPCR_LPES | LPCR_MER;
> + mask = LPCR_DPFD | LPCR_ILE | LPCR_LD | LPCR_LPES | LPCR_MER;
> /* LPCR[AIL]=1/2 is disallowed */
> if ((hr->lpcr & LPCR_AIL) && (hr->lpcr & LPCR_AIL) != LPCR_AIL_3)
> hr->lpcr &= ~LPCR_AIL;
