Re: [PATCH v4] KVM: arm64: Advertise ID_AA64PFR0_EL1.CSV2/3 to protected VMs

[Marc Zyngier <maz@kernel.org>]

On Mon, 03 Apr 2023 14:29:53 +0100,
Fuad Tabba <tabba@google.com> wrote:
> 
> The existing pKVM code attempts to advertise CSV2/3 using values
> initialized to 0, but never set. To advertise CSV2/3 to protected
> guests, pass the CSV2/3 values to hyp when initializing hyp's
> view of guests' ID_AA64PFR0_EL1.
> 
> Similar to non-protected KVM, these are system-wide, rather than
> per cpu, for simplicity.
> 
> Fixes: 6c30bfb18d0b ("KVM: arm64: Add handlers for protected VM System Registers")
> Signed-off-by: Fuad Tabba <tabba@google.com>
> 
> ---
> 
> Based on: 197b6b60ae7b (6.3-rc4)
> Changes from V3: Fix masking to constrain CSV2/3 to 1 (Oliver).
> 
>  arch/arm64/kvm/arm.c                          | 25 ++++++++++++++++++-
>  .../arm64/kvm/hyp/include/nvhe/fixed_config.h |  5 +++-
>  arch/arm64/kvm/hyp/nvhe/sys_regs.c            |  7 ------
>  3 files changed, 28 insertions(+), 9 deletions(-)
> 
> diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c
> index 3bd732eaf087..04401c2851f8 100644
> --- a/arch/arm64/kvm/arm.c
> +++ b/arch/arm64/kvm/arm.c
> @@ -1889,9 +1889,32 @@ static int __init do_pkvm_init(u32 hyp_va_bits)
>  	return ret;
>  }
>  
> +static u64 get_hyp_id_aa64pfr0_el1(void)
> +{
> +	u64 mask = ~(ARM64_FEATURE_MASK(ID_AA64PFR0_EL1_CSV2) |
> +		     ARM64_FEATURE_MASK(ID_AA64PFR0_EL1_CSV3));
> +
> +	/*
> +	 * Track whether the system isn't affected by spectre/meltown in the
> +	 * hypervisor's view of id_aa64pfr0_el1, used for protected VMs.
> +	 * Although this is per-CPU, we make it global for simplicity, e.g., not
> +	 * to have to worry about vcpu migration.
> +	 *
> +	 * Unlike for non-protected VMs, userspace cannot override this for
> +	 * protected VMs.
> +	 */
> +	mask |= FIELD_PREP(ARM64_FEATURE_MASK(ID_AA64PFR0_EL1_CSV2),
> +			   arm64_get_spectre_v2_state() == SPECTRE_UNAFFECTED);
> +
> +	mask |= FIELD_PREP(ARM64_FEATURE_MASK(ID_AA64PFR0_EL1_CSV3),
> +			   arm64_get_meltdown_state() == SPECTRE_UNAFFECTED);
> +
> +	return read_sanitised_ftr_reg(SYS_ID_AA64PFR0_EL1) & mask;

I'm confused. PFR0.CSV2 is only set if the HW advertises it. But KVM
should advertises it to guests if arm64_get_spectre_v2_state() reports
"UNAFFECTED".

In that later case, "read_sanitised_ftr_reg(SYS_ID_AA64PFR0_EL1) &
mask" will always leave CSV2 to 0. Same thing for CSV3.

I'd expect something like:

	u64 val = read_sanitised_ftr_reg(SYS_ID_AA64PFR0_EL1);

	val &= ~(ARM64_FEATURE_MASK(ID_AA64PFR0_EL1_CSV2) |
		 ARM64_FEATURE_MASK(ID_AA64PFR0_EL1_CSV3));
	val |= FIELD_PREP(ARM64_FEATURE_MASK(ID_AA64PFR0_EL1_CSV2),
			  arm64_get_spectre_v2_state() == SPECTRE_UNAFFECTED);
	val |= FIELD_PREP(ARM64_FEATURE_MASK(ID_AA64PFR0_EL1_CSV3),
			  arm64_get_meltdown_state() == SPECTRE_UNAFFECTED);

	return val;

Or did I read that completely wrong?

	M.

-- 
Without deviation from the norm, progress is not possible.