[PATCH v3 06/13] KVM: arm64: Use LPA2 page-tables for stage2 if HW supports it

[Ryan Roberts]

On 27/09/2023 08:02, Oliver Upton wrote:
> On Mon, Sep 18, 2023 at 07:57:33AM +0100, Ryan Roberts wrote:
> 
> [...]
> 
>>  static inline u64 kvm_pte_to_phys(kvm_pte_t pte)
>>  {
>> -	u64 pa = pte & KVM_PTE_ADDR_MASK;
>> -
>> -	if (PAGE_SHIFT == 16)
>> -		pa |= FIELD_GET(KVM_PTE_ADDR_51_48, pte) << 48;
>> +	u64 pa;
>> +
>> +	if (system_supports_lpa2()) {
>> +		pa = pte & KVM_PTE_ADDR_MASK_LPA2;
>> +		pa |= FIELD_GET(KVM_PTE_ADDR_51_50_LPA2, pte) << 50;
>> +	} else {
>> +		pa = pte & KVM_PTE_ADDR_MASK;
>> +		if (PAGE_SHIFT == 16)
>> +			pa |= FIELD_GET(KVM_PTE_ADDR_51_48, pte) << 48;
>> +	}
>>  
>>  	return pa;
>>  }
>>  
>>  static inline kvm_pte_t kvm_phys_to_pte(u64 pa)
>>  {
>> -	kvm_pte_t pte = pa & KVM_PTE_ADDR_MASK;
>> -
>> -	if (PAGE_SHIFT == 16) {
>> -		pa &= GENMASK(51, 48);
>> -		pte |= FIELD_PREP(KVM_PTE_ADDR_51_48, pa >> 48);
>> +	kvm_pte_t pte;
>> +
>> +	if (system_supports_lpa2()) {
>> +		pte = pa & KVM_PTE_ADDR_MASK_LPA2;
>> +		pa &= GENMASK(51, 50);
>> +		pte |= FIELD_PREP(KVM_PTE_ADDR_51_50_LPA2, pa >> 50);
>> +	} else {
>> +		pte = pa & KVM_PTE_ADDR_MASK;
>> +		if (PAGE_SHIFT == 16) {
>> +			pa &= GENMASK(51, 48);
>> +			pte |= FIELD_PREP(KVM_PTE_ADDR_51_48, pa >> 48);
>> +		}
>>  	}
> 
> These accessors are common between stage-1 and stage-2 walks... Aren't
> bits [9:8] of the descriptor still being used to encode shareability for
> hyp stage-1 at this point?

Yes, that's correct. I split the commits just to be able to see what changes are
in respect of which stage, although as you point out, this part is actually shared.

> 
> You may as well squash the subsequent patch (LPA2 format for S1) into this
> one such that the change is atomic.
> 

OK, I'll squash them in the next version.