Re: [PATCH 3/3] KVM: arm64: nv: Handle all _EL02 and _EL12 registers

[Marc Zyngier <maz@kernel.org>]

On Mon, 18 Sep 2023 13:41:45 +0100,
Miguel Luis <miguel.luis@oracle.com> wrote:
> 
> Hi Marc,
> 
> > On 18 Sep 2023, at 09:44, Marc Zyngier <maz@kernel.org> wrote:
> > 
> > On Wed, 13 Sep 2023 19:52:08 +0100,
> > Miguel Luis <miguel.luis@oracle.com> wrote:
> >> 
> >> Specify both _EL02 and _EL12 system registers.
> >> 
> >> Signed-off-by: Miguel Luis <miguel.luis@oracle.com>
> >> ---
> >> arch/arm64/kvm/emulate-nested.c | 35 +++++++++++++++++++++++++++++----
> >> 1 file changed, 31 insertions(+), 4 deletions(-)
> >> 
> >> diff --git a/arch/arm64/kvm/emulate-nested.c b/arch/arm64/kvm/emulate-nested.c
> >> index 9aa1c06abdb7..957afd97e488 100644
> >> --- a/arch/arm64/kvm/emulate-nested.c
> >> +++ b/arch/arm64/kvm/emulate-nested.c
> >> @@ -690,10 +690,37 @@ static const struct encoding_to_trap_config encoding_to_cgt[] __initconst = {
> >> SR_RANGE_TRAP(sys_reg(3, 4, 14, 0, 3),
> >> sys_reg(3, 4, 14, 5, 2), CGT_HCR_NV),
> >> /* All _EL02, _EL12 registers */
> >> - SR_RANGE_TRAP(sys_reg(3, 5, 0, 0, 0),
> >> -       sys_reg(3, 5, 10, 15, 7), CGT_HCR_NV),
> >> - SR_RANGE_TRAP(sys_reg(3, 5, 12, 0, 0),
> >> -       sys_reg(3, 5, 14, 15, 7), CGT_HCR_NV),
> >> + SR_TRAP(SYS_SCTLR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_CPACR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_SCTLR2_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_ZCR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_TRFCR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_SMCR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_TTBR0_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_TTBR1_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_TCR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_TCR2_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_SPSR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_ELR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_AFSR0_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_AFSR1_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_ESR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_TFSR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_FAR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_BRBCR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_PMSCR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_MAIR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_AMAIR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_VBAR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_CONTEXTIDR_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_SCXTNUM_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_CNTKCTL_EL12, CGT_HCR_NV),
> >> + SR_TRAP(SYS_CNTP_TVAL_EL02, CGT_HCR_NV),
> >> + SR_TRAP(SYS_CNTP_CTL_EL02, CGT_HCR_NV),
> >> + SR_TRAP(SYS_CNTP_CVAL_EL02, CGT_HCR_NV),
> >> + SR_TRAP(SYS_CNTV_TVAL_EL02, CGT_HCR_NV),
> >> + SR_TRAP(SYS_CNTV_CTL_EL02, CGT_HCR_NV),
> >> + SR_TRAP(SYS_CNTV_CVAL_EL02, CGT_HCR_NV),
> >> SR_TRAP(OP_AT_S1E2R, CGT_HCR_NV),
> >> SR_TRAP(OP_AT_S1E2W, CGT_HCR_NV),
> >> SR_TRAP(OP_AT_S12E1R, CGT_HCR_NV),
> > 
> > While I could see the problem with the EL2 registers, I'm not
> > convinced by this patch. Is there an actual case for non _EL02, non
> > _EL12 registers that are included in the two ranges above?
> > 
> 
> Having DDI0487Ja as reference, there is none. It is not clear to me having two
> separate ranges. If it is to cover _EL02 and _EL12 ranges separately then the
> second range is covering both aliases. I couldn't find the reason for these
> aliases start and end other than SYS_SCTLR_EL12 and SYS_CNTV_CVAL_EL02,
> respectively.

The reason we have two ranges is to explicitly exclude the IMPDEF
range, which is trapped by HCR_EL2.TIDCP:

	SR_RANGE_TRAP(sys_reg(3, 5, 11, 0, 0),
		      sys_reg(3, 5, 11, 15, 7), CGT_HCR_TIDCP),

Thanks,

	M.

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