Re: [PATCH 12/15] arm: kvm: Move fake PGD handling to arch specific files

[Christoffer Dall <christoffer.dall@linaro.org>]

On Wed, Oct 07, 2015 at 11:23:52AM +0100, Marc Zyngier wrote:
> On 15/09/15 16:41, Suzuki K. Poulose wrote:
> > From: "Suzuki K. Poulose" <suzuki.poulose@arm.com>
> > 
> > Rearrange the code for fake pgd handling, which is applicable
> > to only ARM64. The intention is to keep the common code cleaner,
> > unaware of the underlying hacks.
> > 
> > Cc: kvmarm@lists.cs.columbia.edu
> > Cc: christoffer.dall@linaro.org
> > Cc: Marc.Zyngier@arm.com
> > Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
> > ---
> >  arch/arm/include/asm/kvm_mmu.h   |    7 ++++++
> >  arch/arm/kvm/mmu.c               |   44 +++++---------------------------------
> >  arch/arm64/include/asm/kvm_mmu.h |   43 +++++++++++++++++++++++++++++++++++++
> >  3 files changed, 55 insertions(+), 39 deletions(-)
> > 
> > diff --git a/arch/arm/include/asm/kvm_mmu.h b/arch/arm/include/asm/kvm_mmu.h
> > index 405aa18..1c9aa8a 100644
> > --- a/arch/arm/include/asm/kvm_mmu.h
> > +++ b/arch/arm/include/asm/kvm_mmu.h
> > @@ -173,6 +173,13 @@ static inline unsigned int kvm_get_hwpgd_size(void)
> >  	return PTRS_PER_S2_PGD * sizeof(pgd_t);
> >  }
> >  
> > +static inline pgd_t *kvm_setup_fake_pgd(pgd_t *pgd)
> > +{
> > +	return pgd;
> > +}
> > +
> > +static inline void kvm_free_fake_pgd(pgd_t *pgd) {}
> > +
> >  struct kvm;
> >  
> >  #define kvm_flush_dcache_to_poc(a,l)	__cpuc_flush_dcache_area((a), (l))
> > diff --git a/arch/arm/kvm/mmu.c b/arch/arm/kvm/mmu.c
> > index 7b42012..b210622 100644
> > --- a/arch/arm/kvm/mmu.c
> > +++ b/arch/arm/kvm/mmu.c
> > @@ -677,43 +677,11 @@ int kvm_alloc_stage2_pgd(struct kvm *kvm)
> >  	 * guest, we allocate a fake PGD and pre-populate it to point
> >  	 * to the next-level page table, which will be the real
> >  	 * initial page table pointed to by the VTTBR.
> > -	 *
> > -	 * When KVM_PREALLOC_LEVEL==2, we allocate a single page for
> > -	 * the PMD and the kernel will use folded pud.
> > -	 * When KVM_PREALLOC_LEVEL==1, we allocate 2 consecutive PUD
> > -	 * pages.
> >  	 */
> > -	if (KVM_PREALLOC_LEVEL > 0) {
> > -		int i;
> > -
> > -		/*
> > -		 * Allocate fake pgd for the page table manipulation macros to
> > -		 * work.  This is not used by the hardware and we have no
> > -		 * alignment requirement for this allocation.
> > -		 */
> > -		pgd = kmalloc(PTRS_PER_S2_PGD * sizeof(pgd_t),
> > -				GFP_KERNEL | __GFP_ZERO);
> > -
> > -		if (!pgd) {
> > -			kvm_free_hwpgd(hwpgd);
> > -			return -ENOMEM;
> > -		}
> > -
> > -		/* Plug the HW PGD into the fake one. */
> > -		for (i = 0; i < PTRS_PER_S2_PGD; i++) {
> > -			if (KVM_PREALLOC_LEVEL == 1)
> > -				pgd_populate(NULL, pgd + i,
> > -					     (pud_t *)hwpgd + i * PTRS_PER_PUD);
> > -			else if (KVM_PREALLOC_LEVEL == 2)
> > -				pud_populate(NULL, pud_offset(pgd, 0) + i,
> > -					     (pmd_t *)hwpgd + i * PTRS_PER_PMD);
> > -		}
> > -	} else {
> > -		/*
> > -		 * Allocate actual first-level Stage-2 page table used by the
> > -		 * hardware for Stage-2 page table walks.
> > -		 */
> > -		pgd = (pgd_t *)hwpgd;
> > +	pgd = kvm_setup_fake_pgd(hwpgd);
> > +	if (IS_ERR(pgd)) {
> > +		kvm_free_hwpgd(hwpgd);
> > +		return PTR_ERR(pgd);
> >  	}
> >  
> >  	kvm_clean_pgd(pgd);
> > @@ -820,9 +788,7 @@ void kvm_free_stage2_pgd(struct kvm *kvm)
> >  
> >  	unmap_stage2_range(kvm, 0, KVM_PHYS_SIZE);
> >  	kvm_free_hwpgd(kvm_get_hwpgd(kvm));
> > -	if (KVM_PREALLOC_LEVEL > 0)
> > -		kfree(kvm->arch.pgd);
> > -
> > +	kvm_free_fake_pgd(kvm->arch.pgd);
> >  	kvm->arch.pgd = NULL;
> >  }
> >  
> > diff --git a/arch/arm64/include/asm/kvm_mmu.h b/arch/arm64/include/asm/kvm_mmu.h
> > index 6150567..2567fe8 100644
> > --- a/arch/arm64/include/asm/kvm_mmu.h
> > +++ b/arch/arm64/include/asm/kvm_mmu.h
> > @@ -198,6 +198,49 @@ static inline unsigned int kvm_get_hwpgd_size(void)
> >  	return PTRS_PER_S2_PGD * sizeof(pgd_t);
> >  }
> >  
> > +/*
> > + * Allocate fake pgd for the page table manipulation macros to
> > + * work.  This is not used by the hardware and we have no
> > + * alignment requirement for this allocation.
> > + */
> > +static inline pgd_t* kvm_setup_fake_pgd(pgd_t *hwpgd)
> > +{
> > +	int i;
> > +	pgd_t *pgd;
> > +
> > +	if (!KVM_PREALLOC_LEVEL)
> > +		return hwpgd;
> > +	/*
> > +	 * When KVM_PREALLOC_LEVEL==2, we allocate a single page for
> > +	 * the PMD and the kernel will use folded pud.
> > +	 * When KVM_PREALLOC_LEVEL==1, we allocate 2 consecutive PUD
> > +	 * pages.
> > +	 */
> > +	pgd = kmalloc(PTRS_PER_S2_PGD * sizeof(pgd_t),
> > +			GFP_KERNEL | __GFP_ZERO);
> > +
> > +	if (!pgd)
> > +		return ERR_PTR(-ENOMEM);
> > +
> > +	/* Plug the HW PGD into the fake one. */
> > +	for (i = 0; i < PTRS_PER_S2_PGD; i++) {
> > +		if (KVM_PREALLOC_LEVEL == 1)
> > +			pgd_populate(NULL, pgd + i,
> > +				     (pud_t *)hwpgd + i * PTRS_PER_PUD);
> > +		else if (KVM_PREALLOC_LEVEL == 2)
> > +			pud_populate(NULL, pud_offset(pgd, 0) + i,
> > +				     (pmd_t *)hwpgd + i * PTRS_PER_PMD);
> > +	}
> > +
> > +	return pgd;
> > +}
> > +
> > +static inline void kvm_free_fake_pgd(pgd_t *pgd)
> > +{
> > +	if (KVM_PREALLOC_LEVEL > 0)
> > +		kfree(pgd);
> > +}
> > +
> >  static inline bool kvm_page_empty(void *ptr)
> >  {
> >  	struct page *ptr_page = virt_to_page(ptr);
> > 
> 
> Reviewed-by: Marc Zyngier <marc.zyngier@arm.com>

I see we like moving this code around:
a987370 (arm64: KVM: Fix stage-2 PGD allocation to have per-page refcounting, 2015-03-10)

But I think the end result from this patch looks nice and it seems
correct to me:

Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>


Thanks,
-Christoffer