Re: [PATCH 3/7] kvm/mmu: rename is_largepage_backed to mapping_level

[Joerg Roedel <joerg.roedel@amd.com>]

Hi Marcello,

On Mon, Apr 27, 2009 at 02:12:45PM -0300, Marcelo Tosatti wrote:
> On Fri, Apr 24, 2009 at 01:58:43PM +0200, Joerg Roedel wrote:
> > With the new name and the corresponding backend changes this function
> > can now support multiple hugepage sizes.
> > 
> > Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
> > ---
> >  arch/x86/kvm/mmu.c         |  100 +++++++++++++++++++++++++++++--------------
> >  arch/x86/kvm/paging_tmpl.h |    4 +-
> >  2 files changed, 69 insertions(+), 35 deletions(-)
> > 
> > diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
> > index e3421d8..56cd7c2 100644
> > --- a/arch/x86/kvm/mmu.c
> > +++ b/arch/x86/kvm/mmu.c
> > @@ -382,37 +382,52 @@ static void mmu_free_rmap_desc(struct kvm_rmap_desc *rd)
> >   * Return the pointer to the largepage write count for a given
> >   * gfn, handling slots that are not large page aligned.
> >   */
> > -static int *slot_largepage_idx(gfn_t gfn, struct kvm_memory_slot *slot)
> > +static int *slot_largepage_idx(gfn_t gfn,
> > +			       struct kvm_memory_slot *slot,
> > +			       int level)
> >  {
> >  	unsigned long idx;
> >  
> > -	idx = (gfn / KVM_PAGES_PER_HPAGE(PT_DIRECTORY_LEVEL)) -
> > -	      (slot->base_gfn / KVM_PAGES_PER_HPAGE(PT_DIRECTORY_LEVEL));
> > -	return &slot->lpage_info[0][idx].write_count;
> > +	idx = (gfn / KVM_PAGES_PER_HPAGE(level)) -
> > +	      (slot->base_gfn / KVM_PAGES_PER_HPAGE(level));
> > +	return &slot->lpage_info[level - 2][idx].write_count;
> >  }
> >  
> >  static void account_shadowed(struct kvm *kvm, gfn_t gfn)
> >  {
> > +	struct kvm_memory_slot *slot;
> >  	int *write_count;
> > +	int i;
> >  
> >  	gfn = unalias_gfn(kvm, gfn);
> > -	write_count = slot_largepage_idx(gfn,
> > -					 gfn_to_memslot_unaliased(kvm, gfn));
> > -	*write_count += 1;
> > +
> > +	for (i = PT_DIRECTORY_LEVEL;
> > +	     i < PT_PAGE_TABLE_LEVEL + KVM_NR_PAGE_SIZES; ++i) {
> > +		slot          = gfn_to_memslot_unaliased(kvm, gfn);
> > +		write_count   = slot_largepage_idx(gfn, slot, i);
> > +		*write_count += 1;
> > +	}
> >  }
> >  
> >  static void unaccount_shadowed(struct kvm *kvm, gfn_t gfn)
> >  {
> > +	struct kvm_memory_slot *slot;
> >  	int *write_count;
> > +	int i;
> >  
> >  	gfn = unalias_gfn(kvm, gfn);
> > -	write_count = slot_largepage_idx(gfn,
> > -					 gfn_to_memslot_unaliased(kvm, gfn));
> > -	*write_count -= 1;
> > -	WARN_ON(*write_count < 0);
> > +	for (i = PT_DIRECTORY_LEVEL;
> > +	     i < PT_PAGE_TABLE_LEVEL + KVM_NR_PAGE_SIZES; ++i) {
> > +		slot          = gfn_to_memslot_unaliased(kvm, gfn);
> > +		write_count   = slot_largepage_idx(gfn, slot, i);
> > +		*write_count -= 1;
> > +		WARN_ON(*write_count < 0);
> > +	}
> >  }
> >  
> > -static int has_wrprotected_page(struct kvm *kvm, gfn_t gfn)
> > +static int has_wrprotected_page(struct kvm *kvm,
> > +				gfn_t gfn,
> > +				int level)
> >  {
> >  	struct kvm_memory_slot *slot;
> >  	int *largepage_idx;
> > @@ -420,47 +435,67 @@ static int has_wrprotected_page(struct kvm *kvm, gfn_t gfn)
> >  	gfn = unalias_gfn(kvm, gfn);
> >  	slot = gfn_to_memslot_unaliased(kvm, gfn);
> >  	if (slot) {
> > -		largepage_idx = slot_largepage_idx(gfn, slot);
> > +		largepage_idx = slot_largepage_idx(gfn, slot, level);
> >  		return *largepage_idx;
> >  	}
> >  
> >  	return 1;
> >  }
> >  
> > -static int host_largepage_backed(struct kvm *kvm, gfn_t gfn)
> > +static int host_mapping_level(struct kvm *kvm, gfn_t gfn)
> >  {
> > +	unsigned long page_size = PAGE_SIZE;
> >  	struct vm_area_struct *vma;
> >  	unsigned long addr;
> > -	int ret = 0;
> > +	int i, ret = 0;
> >  
> >  	addr = gfn_to_hva(kvm, gfn);
> >  	if (kvm_is_error_hva(addr))
> > -		return ret;
> > +		return page_size;
> >  
> >  	down_read(&current->mm->mmap_sem);
> >  	vma = find_vma(current->mm, addr);
> > -	if (vma && is_vm_hugetlb_page(vma))
> > -		ret = 1;
> > +	if (!vma)
> > +		goto out;
> > +
> > +	page_size = vma_kernel_pagesize(vma);
> > +
> > +out:
> >  	up_read(&current->mm->mmap_sem);
> >  
> > +	for (i = PT_PAGE_TABLE_LEVEL;
> > +	     i < (PT_PAGE_TABLE_LEVEL + KVM_NR_PAGE_SIZES); ++i) {
> > +		if (page_size >= KVM_HPAGE_SIZE(i))
> > +			ret = i;
> > +		else
> > +			break;
> > +	}
> > +
> >  	return ret;
> >  }
> >  
> > -static int is_largepage_backed(struct kvm_vcpu *vcpu, gfn_t large_gfn)
> > +static int mapping_level(struct kvm_vcpu *vcpu, gfn_t large_gfn)
> >  {
> >  	struct kvm_memory_slot *slot;
> > -
> > -	if (has_wrprotected_page(vcpu->kvm, large_gfn))
> > -		return 0;
> > -
> > -	if (!host_largepage_backed(vcpu->kvm, large_gfn))
> > -		return 0;
> > +	int host_level;
> > +	int level = PT_PAGE_TABLE_LEVEL;
> >  
> >  	slot = gfn_to_memslot(vcpu->kvm, large_gfn);
> >  	if (slot && slot->dirty_bitmap)
> > -		return 0;
> > +		return PT_PAGE_TABLE_LEVEL;
> >  
> > -	return 1;
> > +	host_level = host_mapping_level(vcpu->kvm, large_gfn);
> > +
> > +	if (host_level == PT_PAGE_TABLE_LEVEL)
> > +		return host_level;
> > +
> > +	for (level = PT_DIRECTORY_LEVEL; level <= host_level; ++level) {
> > +
> > +		if (has_wrprotected_page(vcpu->kvm, large_gfn, level))
> > +			break;
> 
> Is there any reason for checking for 4k wrprotect pages in the 2MB
> lpage_info if you're going to map 1GB pages?

Yes. The reason is that if we can't map the page with 1GB than we can
fall back to 2MB and not down to 4kb pages.

Joerg

-- 
           | Advanced Micro Devices GmbH
 Operating | Karl-Hammerschmidt-Str. 34, 85609 Dornach bei München
 System    | 
 Research  | Geschäftsführer: Jochen Polster, Thomas M. McCoy, Giuliano Meroni
 Center    | Sitz: Dornach, Gemeinde Aschheim, Landkreis München
           | Registergericht München, HRB Nr. 43632