> On 5. Feb 2026, at 13:23, Alex Bennée <[email protected]> wrote:
> 
> Richard Henderson <[email protected]> writes:
> 
>> On 2/4/26 23:13, Alex Bennée wrote:
>>> FEAT_E2H0 is a formalisation of the existing behaviour of HCR_EL2.E2H
>>> being programmable to switch between EL2 host mode and the
>>> "traditional" nVHE EL2 mode. This implies at some point we might want
>>> to model CPUs without FEAT_E2H0 which will always have EL2 host mode
>>> enabled.
>>> There are two values to represent no E2H0 systems of which 0b1110
>>> will
>>> make HCR_EL2.NV1 RES0 for FEAT_NV systems. For FEAT_NV2 the NV1 bit is
>>> always valid.
>>> Message-ID: <[email protected]>
>>> Signed-off-by: Alex Bennée <[email protected]>
>>> ---
>>> v2
>>>   - new helper and properly handling NV1
>>> ---
> <snip>
>> 
>>> @@ -3801,10 +3802,13 @@ static void do_hcr_write(CPUARMState *env, uint64_t 
>>> value, uint64_t valid_mask)
>>>              valid_mask |= HCR_GPF;
>>>          }
>>>          if (cpu_isar_feature(aa64_nv, cpu)) {
>>> -            valid_mask |= HCR_NV | HCR_NV1 | HCR_AT;
>>> +            valid_mask |= HCR_NV | HCR_AT;
>>> +            if (!cpu_isar_feature(aa64_noe2h0_and_nv1_res0, cpu)) {
>>> +                valid_mask |= HCR_NV1;
>>> +            }
>>>          }
>>>          if (cpu_isar_feature(aa64_nv2, cpu)) {
>>> -            valid_mask |= HCR_NV2;
>>> +            valid_mask |= HCR_NV1 | HCR_NV2;
>> 
>> Why add NV1 here?
> 
> My reading was if you had FEAT_NV2 then HCR_NV1 was always valid
> whatever E2H0 might say. Unless we can validate we don't get
> contradictory ID values elsewhere?
KVM with nested virt enabled out of the box exposes a virtual machine with
VHE-only L1 and above, with no allowance for an nVHE nested guest past L1.

In the Linux kernel in arch/arm64/kvm/nested.c:

case SYS_ID_AA64MMFR4_EL1:
/*
* You get EITHER
*
* - FEAT_VHE without FEAT_E2H0
* - FEAT_NV limited to FEAT_NV2
* - HCR_EL2.NV1 being RES0
*
* OR
*
* - FEAT_E2H0 without FEAT_VHE nor FEAT_NV
*
* Life is too short for anything else.
*/

And after taking a further look at the machine readable spec*, 0b1110 is valid 
together
with FEAT_NV2, as nothing there says otherwise.

* https://developer.arm.com/Architectures/A-Profile%20Architecture#Downloads
> 
> 
> -- 
> Alex Bennée
> Virtualisation Tech Lead @ Linaro



Reply via email to