Re: [PATCH v10 7/9] KVM: Update lpage info when private/shared memory are mixed

[Sean Christopherson]

On Fri, Dec 02, 2022, Chao Peng wrote:
> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
> index 283cbb83d6ae..7772ab37ac89 100644
> --- a/arch/x86/include/asm/kvm_host.h
> +++ b/arch/x86/include/asm/kvm_host.h
> @@ -38,6 +38,7 @@
>  #include <asm/hyperv-tlfs.h>
>  
>  #define __KVM_HAVE_ARCH_VCPU_DEBUGFS
> +#define __KVM_HAVE_ARCH_SET_MEMORY_ATTRIBUTES

No need for this, I think we should just make it mandatory to implement the
arch hook when CONFIG_KVM_GENERIC_MEMORY_ATTRIBUTES=y.  If another arch gains
support for mem attributes and doesn't need the hook, then we can simply add a
weak helper (or maybe add a #define then if we feel that's the way to go).

>  #define KVM_MAX_VCPUS 1024
>  
> @@ -1011,6 +1012,13 @@ struct kvm_vcpu_arch {
>  #endif
>  };
>  
> +/*
> + * Use a bit in disallow_lpage to indicate private/shared pages mixed at the
> + * level. The remaining bits are used as a reference count.
> + */
> +#define KVM_LPAGE_PRIVATE_SHARED_MIXED		(1U << 31)

Similar to the need to unmap, I think we should just say "mixed" and ignore the
private vs. shared, i.e. make this a flag for all memory attributes.

> +#define KVM_LPAGE_COUNT_MAX			((1U << 31) - 1)

"MAX" is technically correct, but it's more of a mask.  I think we can make it a
moot point though.  There's no need to mask the count, we just want to assert that
adjusting the counting doesn't change the flag.

I would also say throw these defines into mmu.c, at least pending the bug fix
for kvm_alloc_memslot_metadata() (more on that below).

>  struct kvm_lpage_info {
>  	int disallow_lpage;
>  };
> diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
> index e2c70b5afa3e..2190fd8c95c0 100644
> --- a/arch/x86/kvm/mmu/mmu.c
> +++ b/arch/x86/kvm/mmu/mmu.c
> @@ -763,11 +763,16 @@ static void update_gfn_disallow_lpage_count(const struct kvm_memory_slot *slot,
>  {
>  	struct kvm_lpage_info *linfo;
>  	int i;
> +	int disallow_count;
>  
>  	for (i = PG_LEVEL_2M; i <= KVM_MAX_HUGEPAGE_LEVEL; ++i) {
>  		linfo = lpage_info_slot(gfn, slot, i);
> +
> +		disallow_count = linfo->disallow_lpage & KVM_LPAGE_COUNT_MAX;
> +		WARN_ON(disallow_count + count < 0 ||
> +			disallow_count > KVM_LPAGE_COUNT_MAX - count);
> +
>  		linfo->disallow_lpage += count;
> -		WARN_ON(linfo->disallow_lpage < 0);

It's been a long week so don't trust my math, but I believe this can simply be:

		old = linfo->disallow_lpage;
		linfo->disallow_lpage += count;

		WARN_ON_ONCE((old ^ linfo->disallow_lpage) & KVM_LPAGE_MIXED_FLAG);
>  	}
>  }
>  
> @@ -6986,3 +6991,130 @@ void kvm_mmu_pre_destroy_vm(struct kvm *kvm)
>  	if (kvm->arch.nx_huge_page_recovery_thread)
>  		kthread_stop(kvm->arch.nx_huge_page_recovery_thread);
>  }
> +
> +static bool linfo_is_mixed(struct kvm_lpage_info *linfo)
> +{
> +	return linfo->disallow_lpage & KVM_LPAGE_PRIVATE_SHARED_MIXED;
> +}
> +
> +static void linfo_set_mixed(gfn_t gfn, struct kvm_memory_slot *slot,
> +			    int level, bool mixed)
> +{
> +	struct kvm_lpage_info *linfo = lpage_info_slot(gfn, slot, level);
> +
> +	if (mixed)
> +		linfo->disallow_lpage |= KVM_LPAGE_PRIVATE_SHARED_MIXED;
> +	else
> +		linfo->disallow_lpage &= ~KVM_LPAGE_PRIVATE_SHARED_MIXED;
> +}
> +
> +static bool is_expected_attr_entry(void *entry, unsigned long expected_attrs)
> +{
> +	bool expect_private = expected_attrs & KVM_MEMORY_ATTRIBUTE_PRIVATE;
> +
> +	if (xa_to_value(entry) & KVM_MEMORY_ATTRIBUTE_PRIVATE) {
> +		if (!expect_private)
> +			return false;
> +	} else if (expect_private)
> +		return false;

This is messy.  If we drop the private vs. shared specifity, this can go away if
we add a helper to get attributes

	static inline unsigned long kvm_get_memory_attributes(struct kvm *kvm, gfn_t gfn)
	{
		return xa_to_value(xa_load(&kvm->mem_attr_array, gfn));
	}

and then we can do


		if (KVM_BUG_ON(gfn != xas.xa_index, kvm) ||
		    attrs != kvm_get_memory_attributes(kvm, gfn)) {
			mixed = true;
			break;
		}

and

		if (linfo_is_mixed(lpage_info_slot(gfn, slot, level - 1)) ||
		    attrs != kvm_get_memory_attributes(kvm, gfn))
			return true;


> +
> +	return true;
> +}
> +
> +static bool mem_attrs_mixed_2m(struct kvm *kvm, unsigned long attrs,
> +			       gfn_t start, gfn_t end)
> +{
> +	XA_STATE(xas, &kvm->mem_attr_array, start);
> +	gfn_t gfn = start;
> +	void *entry;
> +	bool mixed = false;
> +
> +	rcu_read_lock();
> +	entry = xas_load(&xas);
> +	while (gfn < end) {
> +		if (xas_retry(&xas, entry))
> +			continue;
> +
> +		KVM_BUG_ON(gfn != xas.xa_index, kvm);

As above, I think it's worth bailing immediately if there's a mismatch.

> +
> +		if (!is_expected_attr_entry(entry, attrs)) {
> +			mixed = true;
> +			break;
> +		}
> +
> +		entry = xas_next(&xas);
> +		gfn++;
> +	}
> +
> +	rcu_read_unlock();
> +	return mixed;
> +}
> +
> +static bool mem_attrs_mixed(struct kvm *kvm, struct kvm_memory_slot *slot,

s/mem_attrs_mixed/has_mixed_attrs to make it clear this is querying, not setting.
And has_mixed_attrs_2m() above.

> +			    int level, unsigned long attrs,
> +			    gfn_t start, gfn_t end)
> +{
> +	unsigned long gfn;
> +
> +	if (level == PG_LEVEL_2M)
> +		return mem_attrs_mixed_2m(kvm, attrs, start, end);
> +
> +	for (gfn = start; gfn < end; gfn += KVM_PAGES_PER_HPAGE(level - 1))

Curly braces needed on the for-loop.

> +		if (linfo_is_mixed(lpage_info_slot(gfn, slot, level - 1)) ||
> +		    !is_expected_attr_entry(xa_load(&kvm->mem_attr_array, gfn),
> +					    attrs))
> +			return true;
> +	return false;
> +}
> +
> +static void kvm_update_lpage_private_shared_mixed(struct kvm *kvm,
> +						  struct kvm_memory_slot *slot,
> +						  unsigned long attrs,
> +						  gfn_t start, gfn_t end)
> +{
> +	unsigned long pages, mask;
> +	gfn_t gfn, gfn_end, first, last;
> +	int level;
> +	bool mixed;
> +
> +	/*
> +	 * The sequence matters here: we set the higher level basing on the
> +	 * lower level's scanning result.
> +	 */
> +	for (level = PG_LEVEL_2M; level <= KVM_MAX_HUGEPAGE_LEVEL; level++) {
> +		pages = KVM_PAGES_PER_HPAGE(level);
> +		mask = ~(pages - 1);
> +		first = start & mask;
> +		last = (end - 1) & mask;
> +
> +		/*
> +		 * We only need to scan the head and tail page, for middle pages
> +		 * we know they will not be mixed.
> +		 */
> +		gfn = max(first, slot->base_gfn);
> +		gfn_end = min(first + pages, slot->base_gfn + slot->npages);
> +		mixed = mem_attrs_mixed(kvm, slot, level, attrs, gfn, gfn_end);
> +		linfo_set_mixed(gfn, slot, level, mixed);
> +
> +		if (first == last)
> +			return;
> +
> +		for (gfn = first + pages; gfn < last; gfn += pages)
> +			linfo_set_mixed(gfn, slot, level, false);
> +
> +		gfn = last;
> +		gfn_end = min(last + pages, slot->base_gfn + slot->npages);
> +		mixed = mem_attrs_mixed(kvm, slot, level, attrs, gfn, gfn_end);
> +		linfo_set_mixed(gfn, slot, level, mixed);
> +	}
> +}
> +
> +void kvm_arch_set_memory_attributes(struct kvm *kvm,
> +				    struct kvm_memory_slot *slot,
> +				    unsigned long attrs,
> +				    gfn_t start, gfn_t end)
> +{
> +	if (kvm_slot_can_be_private(slot))

Make this an early return optimization, with a comment explaining that KVM x86
doesn't yet support other attributes.

	/*
	 * KVM x86 currently only supports KVM_MEMORY_ATTRIBUTE_PRIVATE, skip
	 * the slot if the slot will never consume the PRIVATE attribute.
	 */
	if (!kvm_slot_can_be_private(slot))
		return;


> +		kvm_update_lpage_private_shared_mixed(kvm, slot, attrs,
> +						      start, end);
> +}
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index 9a07380f8d3c..5aefcff614d2 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -12362,6 +12362,8 @@ static int kvm_alloc_memslot_metadata(struct kvm *kvm,
>  		if ((slot->base_gfn + npages) & (KVM_PAGES_PER_HPAGE(level) - 1))
>  			linfo[lpages - 1].disallow_lpage = 1;
>  		ugfn = slot->userspace_addr >> PAGE_SHIFT;
> +		if (kvm_slot_can_be_private(slot))
> +			ugfn |= slot->restricted_offset >> PAGE_SHIFT;

I would rather reject memslot if the gfn has lesser alignment than the offset.
I'm totally ok with this approach _if_ there's a use case.  Until such a use case
presents itself, I would rather be conservative from a uAPI perspective.

>  		/*
>  		 * If the gfn and userspace address are not aligned wrt each
>  		 * other, disable large page support for this slot.
> diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
> index 3331c0c92838..25099c94e770 100644
> --- a/include/linux/kvm_host.h
> +++ b/include/linux/kvm_host.h
> @@ -592,6 +592,11 @@ struct kvm_memory_slot {
>  	struct restrictedmem_notifier notifier;
>  };
>  
> +static inline bool kvm_slot_can_be_private(const struct kvm_memory_slot *slot)
> +{
> +	return slot && (slot->flags & KVM_MEM_PRIVATE);

KVM_MEM_PRIVATE should really be defined only when private memory is exposed to
userspace.  For this patch, even though it means we have untestable code, I think
it makes sense to "return false".

> +}
> +
>  static inline bool kvm_slot_dirty_track_enabled(const struct kvm_memory_slot *slot)
>  {
>  	return slot->flags & KVM_MEM_LOG_DIRTY_PAGES;
> @@ -2316,4 +2321,18 @@ static inline void kvm_account_pgtable_pages(void *virt, int nr)
>  /* Max number of entries allowed for each kvm dirty ring */
>  #define  KVM_DIRTY_RING_MAX_ENTRIES  65536
>  
> +#ifdef __KVM_HAVE_ARCH_SET_MEMORY_ATTRIBUTES
> +void kvm_arch_set_memory_attributes(struct kvm *kvm,
> +				    struct kvm_memory_slot *slot,
> +				    unsigned long attrs,
> +				    gfn_t start, gfn_t end);
> +#else
> +static inline void kvm_arch_set_memory_attributes(struct kvm *kvm,
> +						  struct kvm_memory_slot *slot,
> +						  unsigned long attrs,
> +						  gfn_t start, gfn_t end)
> +{
> +}
> +#endif /* __KVM_HAVE_ARCH_SET_MEMORY_ATTRIBUTES */

As above, no stub is necessary.

>  #endif
> diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
> index 4e1e1e113bf0..e107afea32f0 100644
> --- a/virt/kvm/kvm_main.c
> +++ b/virt/kvm/kvm_main.c
> @@ -2354,7 +2354,8 @@ static u64 kvm_supported_mem_attributes(struct kvm *kvm)
>  	return 0;
>  }
>  
> -static void kvm_unmap_mem_range(struct kvm *kvm, gfn_t start, gfn_t end)

Feedback for an earlier patch (to avoid churn): this should be kvm_mem_attrs_changed()
or so now that this does more than just unmap.

> +static void kvm_unmap_mem_range(struct kvm *kvm, gfn_t start, gfn_t end,
> +				unsigned long attrs)

Weird nit.  I think we should keep the prototypes for kvm_mem_attrs_changed()
and kvm_arch_set_memory_attributes() somewhat similar, i.e. squeeze in @attrs
before @start.

>  {
>  	struct kvm_gfn_range gfn_range;
>  	struct kvm_memory_slot *slot;
> @@ -2378,6 +2379,10 @@ static void kvm_unmap_mem_range(struct kvm *kvm, gfn_t start, gfn_t end)
>  			gfn_range.slot = slot;
>  
>  			r |= kvm_unmap_gfn_range(kvm, &gfn_range);
> +
> +			kvm_arch_set_memory_attributes(kvm, slot, attrs,
> +						       gfn_range.start,
> +						       gfn_range.end);
>  		}
>  	}
>  
> @@ -2427,7 +2432,7 @@ static int kvm_vm_ioctl_set_mem_attributes(struct kvm *kvm,
>  		idx = srcu_read_lock(&kvm->srcu);
>  		KVM_MMU_LOCK(kvm);
>  		if (i > start)
> -			kvm_unmap_mem_range(kvm, start, i);
> +			kvm_unmap_mem_range(kvm, start, i, attrs->attributes);
>  		kvm_mmu_invalidate_end(kvm);
>  		KVM_MMU_UNLOCK(kvm);
>  		srcu_read_unlock(&kvm->srcu, idx);
> -- 
> 2.25.1
>