[PATCH v3 20/21] KVM: arm64: Remove unused 'pgd' field from 'struct kvm_s2_mmu'

[Gavin Shan]

Hi Will,

On 9/4/20 2:50 AM, Will Deacon wrote:
> On Thu, Sep 03, 2020 at 03:07:17PM +1000, Gavin Shan wrote:
>> On 8/25/20 7:39 PM, Will Deacon wrote:
>>> The stage-2 page-tables are entirely encapsulated by the 'pgt' field of
>>> 'struct kvm_s2_mmu', so remove the unused 'pgd' field.
>>>
>>> Cc: Marc Zyngier <maz at kernel.org>
>>> Cc: Quentin Perret <qperret at google.com>
>>> Signed-off-by: Will Deacon <will at kernel.org>
>>> ---
>>>    arch/arm64/include/asm/kvm_host.h | 1 -
>>>    arch/arm64/kvm/mmu.c              | 2 --
>>>    2 files changed, 3 deletions(-)
>>>
>>
>> I think this might be folded into PATCH[18] as both patches are
>> simple enough. I'm not sure the changes introduced in PATCH[19]
>> prevent us doing this.
>>
>> There is another question below.
>>
>> Reviewed-by: Gavin Shan <gshan at redhat.com>
>>
>>> diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
>>> index 0b7c702b2151..41caf29bd93c 100644
>>> --- a/arch/arm64/include/asm/kvm_host.h
>>> +++ b/arch/arm64/include/asm/kvm_host.h
>>> @@ -79,7 +79,6 @@ struct kvm_s2_mmu {
>>>    	 * for vEL1/EL0 with vHCR_EL2.VM == 0.  In that case, we use the
>>>    	 * canonical stage-2 page tables.
>>>    	 */
>>> -	pgd_t		*pgd;
>>>    	phys_addr_t	pgd_phys;
>>>    	struct kvm_pgtable *pgt;
>>> diff --git a/arch/arm64/kvm/mmu.c b/arch/arm64/kvm/mmu.c
>>> index ddeec0b03666..f28e03dcb897 100644
>>> --- a/arch/arm64/kvm/mmu.c
>>> +++ b/arch/arm64/kvm/mmu.c
>>> @@ -384,7 +384,6 @@ int kvm_init_stage2_mmu(struct kvm *kvm, struct kvm_s2_mmu *mmu)
>>>    	mmu->kvm = kvm;
>>>    	mmu->pgt = pgt;
>>>    	mmu->pgd_phys = __pa(pgt->pgd);
>>> -	mmu->pgd = (void *)pgt->pgd;
>>>    	mmu->vmid.vmid_gen = 0;
>>>    	return 0;
>>> @@ -470,7 +469,6 @@ void kvm_free_stage2_pgd(struct kvm_s2_mmu *mmu)
>>>    	spin_lock(&kvm->mmu_lock);
>>>    	pgt = mmu->pgt;
>>>    	if (pgt) {
>>> -		mmu->pgd = NULL;
>>>    		mmu->pgd_phys = 0;
>>>    		mmu->pgt = NULL;
>>>    		free_percpu(mmu->last_vcpu_ran);
>>>
>>
>> I guess mmu->pgd_phys might be removed either because kvm_get_vttbr()
>> is the only consumer.
> 
> Hmm, but kvm_get_vttbr() is still used after these patches, so I think
> the pgd_phys field needs to stick around.
> 

Yes, kvm_get_vttbr() is the only consumer. The corresponding physical
address can be figured out in the function, we needn't have the physical
address and cache it in advance. However, it's not a big deal. I probably
post one patch to remove it after this series gets merged.

    baddr = __pa(mmu->pgt->pgd);

Thanks,
Gavin