kvm: mmu: lazy collapse small sptes into large sptes

There are two scenarios for the requirement of collapsing small sptes into large sptes.
- dirty logging tracks sptes in 4k granularity, so large sptes are split, the large sptes will be reallocated in the destination machine and the guest in the source machine will be destroyed when live migration successfully. However, the guest in the source machine will continue to run if live migration fail due to some reasons, the sptes still keep small which lead to bad performance.
- our customers write tools to track the dirty speed of guests by EPT D bit/PML in order to determine the most appropriate one to be live migrated, however sptes will still keep small after tracking dirty speed.

This patch introduce lazy collapse small sptes into large sptes, the memory region will be scanned on the ioctl context when dirty log is stopped, the ones which can be collapsed into large pages will be dropped during the scan, it depends the on later #PF to reallocate all large sptes.

Reviewed-by Xiao Guangrong <[email protected]>

Signed-off-by Wanpeng Li <[email protected]>

---
v2 -> v3:
* update comments
* fix infinite for loop
v1 -> v2:
* use 'bool' instead of 'int'
* add more comments
* fix can not get the next spte after drop the current spte

 arch/x86/include/asm/kvm_host.h |  2 ++
 arch/x86/kvm/mmu.c              | 73 +++++++++++++++++++++++++++++++++++++++++
 arch/x86/kvm/x86.c              | 19 +++++++++++
 3 files changed, 94 insertions(+)

arch/x86/include/asm/kvm_host.h +2 -0

index 30b28dc..91b5bdb 100644
@@ -854,6 +854,8 @@ void kvm_mmu_set_mask_ptes(u64 user_mask, u64 accessed_mask,
 void kvm_mmu_reset_context(struct kvm_vcpu *vcpu);
 void kvm_mmu_slot_remove_write_access(struct kvm *kvm,
 				      struct kvm_memory_slot *memslot);
+void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
+					struct kvm_memory_slot *memslot);
 void kvm_mmu_slot_leaf_clear_dirty(struct kvm *kvm,
 				   struct kvm_memory_slot *memslot);
 void kvm_mmu_slot_largepage_remove_write_access(struct kvm *kvm,

arch/x86/kvm/mmu.c +73 -0

index cee7592..ba002a0 100644
@@ -4465,6 +4465,79 @@ void kvm_mmu_slot_remove_write_access(struct kvm *kvm,
 		kvm_flush_remote_tlbs(kvm);
 }
 
+static bool kvm_mmu_zap_collapsible_spte(struct kvm *kvm,
+		unsigned long *rmapp)
+{
+	u64 *sptep;
+	struct rmap_iterator iter;
+	int need_tlb_flush = 0;
+	pfn_t pfn;
+	struct kvm_mmu_page *sp;
+
+	for (sptep = rmap_get_first(*rmapp, &iter); sptep;) {
+		BUG_ON(!(*sptep & PT_PRESENT_MASK));
+
+		sp = page_header(__pa(sptep));
+		pfn = spte_to_pfn(*sptep);
+
+		/*
+		 * Lets support EPT only for now, there still needs to figure
+		 * out an efficient way to let these codes be aware what mapping
+		 * level used in guest.
+		 */
+		if (sp->role.direct &&
+			!kvm_is_reserved_pfn(pfn) &&
+			PageTransCompound(pfn_to_page(pfn))) {
+			drop_spte(kvm, sptep);
+			sptep = rmap_get_first(*rmapp, &iter);
+			need_tlb_flush = 1;
+		} else
+			sptep = rmap_get_next(&iter);
+	}
+
+	return need_tlb_flush;
+}
+
+void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
+			struct kvm_memory_slot *memslot)
+{
+	bool flush = false;
+	unsigned long *rmapp;
+	unsigned long last_index, index;
+	gfn_t gfn_start, gfn_end;
+
+	spin_lock(&kvm->mmu_lock);
+
+	gfn_start = memslot->base_gfn;
+	gfn_end = memslot->base_gfn + memslot->npages - 1;
+
+	if (gfn_start >= gfn_end)
+		goto out;
+
+	rmapp = memslot->arch.rmap[0];
+	last_index = gfn_to_index(gfn_end, memslot->base_gfn,
+					PT_PAGE_TABLE_LEVEL);
+
+	for (index = 0; index <= last_index; ++index, ++rmapp) {
+		if (*rmapp)
+			flush |= kvm_mmu_zap_collapsible_spte(kvm, rmapp);
+
+		if (need_resched() || spin_needbreak(&kvm->mmu_lock)) {
+			if (flush) {
+				kvm_flush_remote_tlbs(kvm);
+				flush = false;
+			}
+			cond_resched_lock(&kvm->mmu_lock);
+		}
+	}
+
+	if (flush)
+		kvm_flush_remote_tlbs(kvm);
+
+out:
+	spin_unlock(&kvm->mmu_lock);
+}
+
 void kvm_mmu_slot_leaf_clear_dirty(struct kvm *kvm,
 				   struct kvm_memory_slot *memslot)
 {

arch/x86/kvm/x86.c +12 -0

index 50861dd..a6cd10b 100644
@@ -7647,6 +7647,25 @@ void kvm_arch_commit_memory_region(struct kvm *kvm,
 	new = id_to_memslot(kvm->memslots, mem->slot);
 
 	/*
+	 * Dirty logging tracks sptes in 4k granularity, so large sptes are
+	 * split, the large sptes will be reallocated in the destination
+	 * machine and the guest in the source machine will be destroyed
+	 * when live migration successfully. However, the guest in the source
+	 * machine will continue to run if live migration fail due to some
+	 * reasons, the sptes still keep small which lead to bad performance.
+	 *
+	 * Lazy collapse small sptes into large sptes is intended to handle
+	 * this, the memory region will be scanned on the ioctl context when
+	 * dirty log is stopped, the ones which can be collapsed into large
+	 * pages will be dropped during the scan, it depends the on later #PF
+	 * to reallocate all large sptes.

There are two scenarios for the requirement of collapsing small sptes into large sptes.
- dirty logging tracks sptes in 4k granularity, so large sptes are split, the large sptes will be reallocated in the destination machine and the guest in the source machine will be destroyed when live migration successfully. However, the guest in the source machine will continue to run if live migration fail due to some reasons, the sptes still keep small which lead to bad performance.
- our customers write tools to track the dirty speed of guests by EPT D bit/PML in order to determine the most appropriate one to be live migrated, however sptes will still keep small after tracking dirty speed.

This patch introduce lazy collapse small sptes into large sptes, the memory region will be scanned on the ioctl context when dirty log is stopped, the ones which can be collapsed into large pages will be dropped during the scan, it depends the on later #PF to reallocate all large sptes.

Reviewed-by Xiao Guangrong <[email protected]>

Signed-off-by Wanpeng Li <[email protected]>

---
v2 -> v3:
* update comments
* fix infinite for loop
v1 -> v2:
* use 'bool' instead of 'int'
* add more comments
* fix can not get the next spte after drop the current spte

 arch/x86/include/asm/kvm_host.h |  2 ++
 arch/x86/kvm/mmu.c              | 73 +++++++++++++++++++++++++++++++++++++++++
 arch/x86/kvm/x86.c              | 19 +++++++++++
 3 files changed, 94 insertions(+)

arch/x86/include/asm/kvm_host.h +2 -0

index 30b28dc..91b5bdb 100644
@@ -854,6 +854,8 @@ void kvm_mmu_set_mask_ptes(u64 user_mask, u64 accessed_mask,
 void kvm_mmu_reset_context(struct kvm_vcpu *vcpu);
 void kvm_mmu_slot_remove_write_access(struct kvm *kvm,
                                      struct kvm_memory_slot *memslot);
+void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
+                                       struct kvm_memory_slot *memslot);
 void kvm_mmu_slot_leaf_clear_dirty(struct kvm *kvm,
                                   struct kvm_memory_slot *memslot);
 void kvm_mmu_slot_largepage_remove_write_access(struct kvm *kvm,

arch/x86/kvm/mmu.c +13 -0

index cee7592..ba002a0 100644
@@ -4465,6 +4465,79 @@ void kvm_mmu_slot_remove_write_access(struct kvm *kvm,
                kvm_flush_remote_tlbs(kvm);
 }
 
+static bool kvm_mmu_zap_collapsible_spte(struct kvm *kvm,
+               unsigned long *rmapp)
+{
+       u64 *sptep;
+       struct rmap_iterator iter;
+       int need_tlb_flush = 0;
+       pfn_t pfn;
+       struct kvm_mmu_page *sp;
+
+       for (sptep = rmap_get_first(*rmapp, &iter); sptep;) {
+               BUG_ON(!(*sptep & PT_PRESENT_MASK));
+
+               sp = page_header(__pa(sptep));

arch/x86/kvm/mmu.c +15 -0

+                       drop_spte(kvm, sptep);
+                       sptep = rmap_get_first(*rmapp, &iter);
+                       need_tlb_flush = 1;
+               } else
+                       sptep = rmap_get_next(&iter);
+       }
+
+       return need_tlb_flush;
+}
+
+void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
+                       struct kvm_memory_slot *memslot)
+{
+       bool flush = false;
+       unsigned long *rmapp;

arch/x86/kvm/mmu.c +10 -0

+               }
+       }
+
+       if (flush)
+               kvm_flush_remote_tlbs(kvm);
+
+out:
+       spin_unlock(&kvm->mmu_lock);
+}
+
 void kvm_mmu_slot_leaf_clear_dirty(struct kvm *kvm,
                                   struct kvm_memory_slot *memslot)
 {

arch/x86/kvm/x86.c +19 -1

index 50861dd..a6cd10b 100644
@@ -7647,6 +7647,25 @@ void kvm_arch_commit_memory_region(struct kvm *kvm,
        new = id_to_memslot(kvm->memslots, mem->slot);
 
        /*
+        * Dirty logging tracks sptes in 4k granularity, so large sptes are
+        * split, the large sptes will be reallocated in the destination
+        * machine and the guest in the source machine will be destroyed
+        * when live migration successfully. However, the guest in the source
+        * machine will continue to run if live migration fail due to some
+        * reasons, the sptes still keep small which lead to bad performance.
+        *
+        * Lazy collapse small sptes into large sptes is intended to handle
+        * this, the memory region will be scanned on the ioctl context when
+        * dirty log is stopped, the ones which can be collapsed into large
+        * pages will be dropped during the scan, it depends the on later #PF
+        * to reallocate all large sptes.
+        */
+       if ((change != KVM_MR_DELETE) &&
+               (old->flags & KVM_MEM_LOG_DIRTY_PAGES) &&
+               !(new->flags & KVM_MEM_LOG_DIRTY_PAGES))
+               kvm_mmu_zap_collapsible_sptes(kvm, new);
+
+       /*
         * Set up write protection and/or dirty logging for the new slot.
         *
         * For KVM_MR_DELETE and KVM_MR_MOVE, the shadow pages of old slot have
--

1.9.1

--
To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Quoted 18 lines · Wanpeng Li

There are two scenarios for the requirement of collapsing small sptes into large sptes.
- dirty logging tracks sptes in 4k granularity, so large sptes are split, the large sptes will be reallocated in the destination machine and the guest in the source machine will be destroyed when live migration successfully. However, the guest in the source machine will continue to run if live migration fail due to some reasons, the sptes still keep small which lead to bad performance.
- our customers write tools to track the dirty speed of guests by EPT D bit/PML in order to determine the most appropriate one to be live migrated, however sptes will still keep small after tracking dirty speed.

This patch introduce lazy collapse small sptes into large sptes, the memory region will be scanned on the ioctl context when dirty log is stopped, the ones which can be collapsed into large pages will be dropped during the scan, it depends the on later #PF to reallocate all large sptes.

Quoted 2 lines

Reviewed-by Xiao Guangrong <[email protected]>

Signed-off-by Wanpeng Li <[email protected]>

Hi, apologies for late review (vacation), but wanted to bring attention to a few matters:

Quoted 50 lines

---
v2 -> v3:
* update comments
* fix infinite for loop
v1 -> v2:
* use 'bool' instead of 'int'
* add more comments
* fix can not get the next spte after drop the current spte

  arch/x86/include/asm/kvm_host.h |  2 ++
  arch/x86/kvm/mmu.c              | 73 +++++++++++++++++++++++++++++++++++++++++
  arch/x86/kvm/x86.c              | 19 +++++++++++
  3 files changed, 94 insertions(+)

arch/x86/include/asm/kvm_host.h +2 -0

index 30b28dc..91b5bdb 100644
@@ -854,6 +854,8 @@ void kvm_mmu_set_mask_ptes(u64 user_mask, u64 accessed_mask,
  void kvm_mmu_reset_context(struct kvm_vcpu *vcpu);
  void kvm_mmu_slot_remove_write_access(struct kvm *kvm,
                                       struct kvm_memory_slot *memslot);
+void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
+                                       struct kvm_memory_slot *memslot);
  void kvm_mmu_slot_leaf_clear_dirty(struct kvm *kvm,
                                    struct kvm_memory_slot *memslot);
  void kvm_mmu_slot_largepage_remove_write_access(struct kvm *kvm,

arch/x86/kvm/mmu.c +13 -0

index cee7592..ba002a0 100644
@@ -4465,6 +4465,79 @@ void kvm_mmu_slot_remove_write_access(struct kvm *kvm,
                 kvm_flush_remote_tlbs(kvm);
  }
 
+static bool kvm_mmu_zap_collapsible_spte(struct kvm *kvm,
+               unsigned long *rmapp)
+{
+       u64 *sptep;
+       struct rmap_iterator iter;
+       int need_tlb_flush = 0;
+       pfn_t pfn;
+       struct kvm_mmu_page *sp;
+
+       for (sptep = rmap_get_first(*rmapp, &iter); sptep;) {
+               BUG_ON(!(*sptep & PT_PRESENT_MASK));
+
+               sp = page_header(__pa(sptep));

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4490

Quoted 10 lines

arch/x86/kvm/mmu.c +10 -0

+               pfn = spte_to_pfn(*sptep);
+
+               /*
+                * Lets support EPT only for now, there still needs to figure
+                * out an efficient way to let these codes be aware what mapping
+                * level used in guest.
+                */
+               if (sp->role.direct &&
+                       !kvm_is_reserved_pfn(pfn) &&
+                       PageTransCompound(pfn_to_page(pfn))) {

Not your fault, but PageTransCompound is very unhappy naming, as it also yields true for PageHuge. Suggestion: document this check covers static hugetlbfs, or switch to PageCompound() check.

A slightly bolder approach would be to refactor and reuse the nearly identical check done in transparent_hugepage_adjust, instead of open-coding here. In essence this code is asking for the same check, plus the out-of-band check for static hugepages.

Quoted 17 lines

arch/x86/kvm/mmu.c +15 -0

+                       drop_spte(kvm, sptep);
+                       sptep = rmap_get_first(*rmapp, &iter);
+                       need_tlb_flush = 1;
+               } else
+                       sptep = rmap_get_next(&iter);
+       }
+
+       return need_tlb_flush;
+}
+
+void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
+                       struct kvm_memory_slot *memslot)
+{
+       bool flush = false;
+       unsigned long *rmapp;

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4515

Quoted 10 lines

arch/x86/kvm/mmu.c +10 -0

+       unsigned long last_index, index;
+       gfn_t gfn_start, gfn_end;
+
+       spin_lock(&kvm->mmu_lock);
+
+       gfn_start = memslot->base_gfn;
+       gfn_end = memslot->base_gfn + memslot->npages - 1;
+
+       if (gfn_start >= gfn_end)
+               goto out;

I don't understand the value of this check here. Are we looking for a broken memslot? Shouldn't this be a BUG_ON? Is this the place to care about these things? npages is capped to KVM_MEM_MAX_NR_PAGES, i.e. 2^31. A 64 bit overflow would be caused by a gigantic gfn_start which would be trouble in many other ways.

All this to say: please remove the above 5 lines and make code simpler.

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4521

Quoted 7 lines

arch/x86/kvm/mmu.c +6 -0

+
+       rmapp = memslot->arch.rmap[0];
+       last_index = gfn_to_index(gfn_end, memslot->base_gfn,
+                                       PT_PAGE_TABLE_LEVEL);
+
+       for (index = 0; index <= last_index; ++index, ++rmapp) {

One could argue that the cleaner iteration should be over the gfn space covered by the memslot, thus leaving the gfn <--> rmap <--> spte interactions hidden under the hood of __gfn_to_rmap. That yields much cleaner (IMHO) code:

     for (gfn = memslot->base_gfn; gfn <= memslot->base_gfn +
memslot->npages; gfn++) {
         flush |= kvm_mmu_zap_collapsible_spte(kvm, __gfn_to_rmap(gfn,
1, memslot));
         ....

Now you can also get rid of index, last_index and rmapp. And more importantly, the code is more understandable, and follows pattern as established in x86/kvm/mmu.

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4530

Quoted 10 lines

arch/x86/kvm/mmu.c +9 -0

+               if (*rmapp)
+                       flush |= kvm_mmu_zap_collapsible_spte(kvm, rmapp);
+
+               if (need_resched() || spin_needbreak(&kvm->mmu_lock)) {
+                       if (flush) {
+                               kvm_flush_remote_tlbs(kvm);
+                               flush = false;
+                       }
+                       cond_resched_lock(&kvm->mmu_lock);

Relinquishing this spinlock is problematic, because commit_memory_region has not gotten around to removing write protection. Are you certain no new write-protected PTEs will be inserted by a racing fault that sneaks in while the spinlock is relinquished?

Quoted 19 lines · Andres Lagar-Cavilla

Quoted 18 lines · Wanpeng Li

There are two scenarios for the requirement of collapsing small sptes into large sptes.
- dirty logging tracks sptes in 4k granularity, so large sptes are split, the large sptes will be reallocated in the destination machine and the guest in the source machine will be destroyed when live migration successfully. However, the guest in the source machine will continue to run if live migration fail due to some reasons, the sptes still keep small which lead to bad performance.
- our customers write tools to track the dirty speed of guests by EPT D bit/PML in order to determine the most appropriate one to be live migrated, however sptes will still keep small after tracking dirty speed.

This patch introduce lazy collapse small sptes into large sptes, the memory region will be scanned on the ioctl context when dirty log is stopped, the ones which can be collapsed into large pages will be dropped during the scan, it depends the on later #PF to reallocate all large sptes.

Quoted 5 lines

Quoted 2 lines

Reviewed-by Xiao Guangrong <[email protected]>

Signed-off-by Wanpeng Li <[email protected]>

Hi, apologies for late review (vacation), but wanted to bring attention to a few matters:

No problem, your comments are really valuable to us. :)

Quoted 65 lines

Quoted 50 lines

---
v2 -> v3:
* update comments
* fix infinite for loop
v1 -> v2:
* use 'bool' instead of 'int'
* add more comments
* fix can not get the next spte after drop the current spte

 arch/x86/include/asm/kvm_host.h |  2 ++
 arch/x86/kvm/mmu.c              | 73 +++++++++++++++++++++++++++++++++++++++++
 arch/x86/kvm/x86.c              | 19 +++++++++++
 3 files changed, 94 insertions(+)

arch/x86/include/asm/kvm_host.h +2 -0

index 30b28dc..91b5bdb 100644
@@ -854,6 +854,8 @@ void kvm_mmu_set_mask_ptes(u64 user_mask, u64 accessed_mask,
 void kvm_mmu_reset_context(struct kvm_vcpu *vcpu);
 void kvm_mmu_slot_remove_write_access(struct kvm *kvm,
                                      struct kvm_memory_slot *memslot);
+void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
+                                       struct kvm_memory_slot *memslot);
 void kvm_mmu_slot_leaf_clear_dirty(struct kvm *kvm,
                                   struct kvm_memory_slot *memslot);
 void kvm_mmu_slot_largepage_remove_write_access(struct kvm *kvm,

arch/x86/kvm/mmu.c +13 -0

index cee7592..ba002a0 100644
@@ -4465,6 +4465,79 @@ void kvm_mmu_slot_remove_write_access(struct kvm *kvm,
                kvm_flush_remote_tlbs(kvm);
 }
 
+static bool kvm_mmu_zap_collapsible_spte(struct kvm *kvm,
+               unsigned long *rmapp)
+{
+       u64 *sptep;
+       struct rmap_iterator iter;
+       int need_tlb_flush = 0;
+       pfn_t pfn;
+       struct kvm_mmu_page *sp;
+
+       for (sptep = rmap_get_first(*rmapp, &iter); sptep;) {
+               BUG_ON(!(*sptep & PT_PRESENT_MASK));
+
+               sp = page_header(__pa(sptep));

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4490

Quoted 10 lines

arch/x86/kvm/mmu.c +10 -0

+               pfn = spte_to_pfn(*sptep);
+
+               /*
+                * Lets support EPT only for now, there still needs to figure
+                * out an efficient way to let these codes be aware what mapping
+                * level used in guest.
+                */
+               if (sp->role.direct &&
+                       !kvm_is_reserved_pfn(pfn) &&
+                       PageTransCompound(pfn_to_page(pfn))) {

Not your fault, but PageTransCompound is very unhappy naming, as it also yields true for PageHuge. Suggestion: document this check covers static hugetlbfs, or switch to PageCompound() check.

Quoted 4 lines

A slightly bolder approach would be to refactor and reuse the nearly identical check done in transparent_hugepage_adjust, instead of open-coding here. In essence this code is asking for the same check, plus the out-of-band check for static hugepages.

I agree.

Quoted 34 lines

Quoted 17 lines

arch/x86/kvm/mmu.c +15 -0

+                       drop_spte(kvm, sptep);
+                       sptep = rmap_get_first(*rmapp, &iter);
+                       need_tlb_flush = 1;
+               } else
+                       sptep = rmap_get_next(&iter);
+       }
+
+       return need_tlb_flush;
+}
+
+void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
+                       struct kvm_memory_slot *memslot)
+{
+       bool flush = false;
+       unsigned long *rmapp;

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4515

Quoted 10 lines

arch/x86/kvm/mmu.c +10 -0

+       unsigned long last_index, index;
+       gfn_t gfn_start, gfn_end;
+
+       spin_lock(&kvm->mmu_lock);
+
+       gfn_start = memslot->base_gfn;
+       gfn_end = memslot->base_gfn + memslot->npages - 1;
+
+       if (gfn_start >= gfn_end)
+               goto out;

I don't understand the value of this check here. Are we looking for a broken memslot? Shouldn't this be a BUG_ON? Is this the place to care about these things? npages is capped to KVM_MEM_MAX_NR_PAGES, i.e. 2^31. A 64 bit overflow would be caused by a gigantic gfn_start which would be trouble in many other ways.

Quoted 1 line

All this to say: please remove the above 5 lines and make code simpler.

Yes, this check is unnecessary indeed.

Quoted 19 lines

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4521

Quoted 7 lines

arch/x86/kvm/mmu.c +6 -0

+
+       rmapp = memslot->arch.rmap[0];
+       last_index = gfn_to_index(gfn_end, memslot->base_gfn,
+                                       PT_PAGE_TABLE_LEVEL);
+
+       for (index = 0; index <= last_index; ++index, ++rmapp) {

One could argue that the cleaner iteration should be over the gfn space covered by the memslot, thus leaving the gfn <--> rmap <--> spte interactions hidden under the hood of __gfn_to_rmap. That yields much cleaner (IMHO) code:

    for (gfn = memslot->base_gfn; gfn <= memslot->base_gfn +
memslot->npages; gfn++) {
        flush |= kvm_mmu_zap_collapsible_spte(kvm, __gfn_to_rmap(gfn,
1, memslot));
        ....

Quoted 3 lines

Now you can also get rid of index, last_index and rmapp. And more importantly, the code is more understandable, and follows pattern as established in x86/kvm/mmu.

Do not have strong opinion on it. Current code also has this style, please refer to kvm_mmu_slot_remove_write_access().

Quoted 11 lines

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4530

Quoted 10 lines

arch/x86/kvm/mmu.c +9 -0

+               if (*rmapp)
+                       flush |= kvm_mmu_zap_collapsible_spte(kvm, rmapp);
+
+               if (need_resched() || spin_needbreak(&kvm->mmu_lock)) {
+                       if (flush) {
+                               kvm_flush_remote_tlbs(kvm);
+                               flush = false;
+                       }
+                       cond_resched_lock(&kvm->mmu_lock);

Quoted 6 lines

Relinquishing this spinlock is problematic, because commit_memory_region has not gotten around to removing write protection. Are you certain no new write-protected PTEs will be inserted by a racing fault that sneaks in while the spinlock is relinquished?

I do not know clearly about the problem, new spte creation will be based on the host mapping, i.e, the huge mapping on shadow page table will be sync-ed with huge mapping on host. Could you please detail the problem?

Wanpeng, could you please post a patch to address Andres's comments?

Quoted 33 lines · Andres Lagar-Cavilla

Quoted 28 lines · Wanpeng Li

There are two scenarios for the requirement of collapsing small sptes into large sptes.
- dirty logging tracks sptes in 4k granularity, so large sptes are split, the large sptes will be reallocated in the destination machine and the guest in the source machine will be destroyed when live migration
successfully.
However, the guest in the source machine will continue to run if live
migration
fail due to some reasons, the sptes still keep small which lead to bad performance.
- our customers write tools to track the dirty speed of guests by EPT D bit/PML
in order to determine the most appropriate one to be live migrated,
however
sptes will still keep small after tracking dirty speed.

This patch introduce lazy collapse small sptes into large sptes, the memory region
will be scanned on the ioctl context when dirty log is stopped, the ones which can
be collapsed into large pages will be dropped during the scan, it depends the on
later #PF to reallocate all large sptes.

Reviewed-by Xiao Guangrong <[email protected]>

Signed-off-by Wanpeng Li <[email protected]>

Quoted 2 lines

Hi, apologies for late review (vacation), but wanted to bring attention to a few matters:

No problem, your comments are really valuable to us. :)

Quoted 74 lines

Quoted 63 lines

---
v2 -> v3:
* update comments
* fix infinite for loop
v1 -> v2:
* use 'bool' instead of 'int'
* add more comments
* fix can not get the next spte after drop the current spte

  arch/x86/include/asm/kvm_host.h |  2 ++
  arch/x86/kvm/mmu.c              | 73
+++++++++++++++++++++++++++++++++++++++++
  arch/x86/kvm/x86.c              | 19 +++++++++++
  3 files changed, 94 insertions(+)

arch/x86/include/asm/kvm_host.h +2 -0

index 30b28dc..91b5bdb 100644
@@ -854,6 +854,8 @@ void kvm_mmu_set_mask_ptes(u64 user_mask, u64
 accessed_mask,
  void kvm_mmu_reset_context(struct kvm_vcpu *vcpu);
  void kvm_mmu_slot_remove_write_access(struct kvm *kvm,
                                       struct kvm_memory_slot *memslot);
+void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
+                                       struct kvm_memory_slot *memslot);
  void kvm_mmu_slot_leaf_clear_dirty(struct kvm *kvm,
                                    struct kvm_memory_slot *memslot);

void kvm_mmu_slot_largepage_remove_write_access(struct kvm *kvm,

arch/x86/kvm/mmu.c +17 -0

index cee7592..ba002a0 100644
@@ -4465,6 +4465,79 @@ void kvm_mmu_slot_remove_write_access(struct kvm
 *kvm,
                 kvm_flush_remote_tlbs(kvm);
  }
 
+static bool kvm_mmu_zap_collapsible_spte(struct kvm *kvm,
+               unsigned long *rmapp)
+{
+       u64 *sptep;
+       struct rmap_iterator iter;
+       int need_tlb_flush = 0;
+       pfn_t pfn;
+       struct kvm_mmu_page *sp;
+
+       for (sptep = rmap_get_first(*rmapp, &iter); sptep;) {
+               BUG_ON(!(*sptep & PT_PRESENT_MASK));
+
+               sp = page_header(__pa(sptep));
+               pfn = spte_to_pfn(*sptep);
+
+               /*
+                * Lets support EPT only for now, there still needs to

figure
+ * out an efficient way to let these codes be aware what
mapping

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4490

Quoted 5 lines

arch/x86/kvm/mmu.c +5 -0

+                * level used in guest.
+                */
+               if (sp->role.direct &&
+                       !kvm_is_reserved_pfn(pfn) &&
+                       PageTransCompound(pfn_to_page(pfn))) {

Not your fault, but PageTransCompound is very unhappy naming, as it also yields true for PageHuge. Suggestion: document this check covers static hugetlbfs, or switch to PageCompound() check.

Quoted 4 lines

A slightly bolder approach would be to refactor and reuse the nearly identical check done in transparent_hugepage_adjust, instead of open-coding here. In essence this code is asking for the same check, plus the out-of-band check for static hugepages.

I agree.

Quoted 35 lines

Quoted 17 lines

arch/x86/kvm/mmu.c +15 -0

+                       drop_spte(kvm, sptep);
+                       sptep = rmap_get_first(*rmapp, &iter);
+                       need_tlb_flush = 1;
+               } else
+                       sptep = rmap_get_next(&iter);
+       }
+
+       return need_tlb_flush;
+}
+
+void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
+                       struct kvm_memory_slot *memslot)
+{
+       bool flush = false;
+       unsigned long *rmapp;

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4515

Quoted 10 lines

arch/x86/kvm/mmu.c +10 -0

+       unsigned long last_index, index;
+       gfn_t gfn_start, gfn_end;
+
+       spin_lock(&kvm->mmu_lock);
+
+       gfn_start = memslot->base_gfn;
+       gfn_end = memslot->base_gfn + memslot->npages - 1;
+
+       if (gfn_start >= gfn_end)
+               goto out;

I don't understand the value of this check here. Are we looking for a broken memslot? Shouldn't this be a BUG_ON? Is this the place to care about these things? npages is capped to KVM_MEM_MAX_NR_PAGES, i.e. 2^31. A 64 bit overflow would be caused by a gigantic gfn_start which would be trouble in many other ways.

Quoted 1 line

All this to say: please remove the above 5 lines and make code simpler.

Yes, this check is unnecessary indeed.

Quoted 20 lines

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4521

Quoted 7 lines

arch/x86/kvm/mmu.c +6 -0

+
+       rmapp = memslot->arch.rmap[0];
+       last_index = gfn_to_index(gfn_end, memslot->base_gfn,
+                                       PT_PAGE_TABLE_LEVEL);
+
+       for (index = 0; index <= last_index; ++index, ++rmapp) {

One could argue that the cleaner iteration should be over the gfn space covered by the memslot, thus leaving the gfn <--> rmap <--> spte interactions hidden under the hood of __gfn_to_rmap. That yields much cleaner (IMHO) code:

     for (gfn = memslot->base_gfn; gfn <= memslot->base_gfn +
memslot->npages; gfn++) {
         flush |= kvm_mmu_zap_collapsible_spte(kvm, __gfn_to_rmap(gfn,
1, memslot));
         ....

Quoted 3 lines

Now you can also get rid of index, last_index and rmapp. And more importantly, the code is more understandable, and follows pattern as established in x86/kvm/mmu.

Do not have strong opinion on it. Current code also has this style, please refer to kvm_mmu_slot_remove_write_access().

Quoted 13 lines

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4530

Quoted 11 lines

arch/x86/kvm/mmu.c +2 -0

+               if (*rmapp)
+                       flush |= kvm_mmu_zap_collapsible_spte(kvm,

rmapp);

arch/x86/kvm/mmu.c +7 -0

+
+               if (need_resched() || spin_needbreak(&kvm->mmu_lock)) {
+                       if (flush) {
+                               kvm_flush_remote_tlbs(kvm);
+                               flush = false;
+                       }
+                       cond_resched_lock(&kvm->mmu_lock);

Quoted 6 lines

Relinquishing this spinlock is problematic, because commit_memory_region has not gotten around to removing write protection. Are you certain no new write-protected PTEs will be inserted by a racing fault that sneaks in while the spinlock is relinquished?

I do not know clearly about the problem, new spte creation will be based on the host mapping, i.e, the huge mapping on shadow page table will be sync-ed with huge mapping on host. Could you please detail the problem?

Wanpeng, could you please post a patch to address Andres's comments?

arch/x86/kvm/mmu.c static bool kvm_mmu_zap_collapsible_spte(struct kvm *kvm, · line 4488

Quoted 10 lines

arch/x86/kvm/mmu.c +5 -3

-         * Only EPT supported for now; otherwise, one would need to
-         * find out efficiently whether the guest page tables are
-         * also using huge pages.
+         * We can not do huge page mapping for the indirect shadow
+         * page (sp) found on the last rmap (level = 1 ) since
+         * indirect sp is synced with the page table in guest and
+         * indirect sp->level = 1 means the guest page table is
+         * using 4K page size mapping.

What about:

arch/x86/kvm/mmu.c static bool kvm_mmu_zap_collapsible_spte(struct kvm *kvm, · line 4488

Quoted 10 lines

arch/x86/kvm/mmu.c +5 -3

-         * Only EPT supported for now; otherwise, one would need to
-         * find out efficiently whether the guest page tables are
-         * also using huge pages.
+         * We can not do huge page mapping for the indirect shadow
+         * page (sp) found on the last rmap (level = 1 ) since
+         * indirect sp is synced with the page table in guest and
+         * indirect sp->level = 1 means the guest page table is
+         * using 4K page size mapping.

What about:

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4490

Quoted 3 lines

arch/x86/kvm/mmu.c +3 -0

+               if (sp->role.direct &&
+                       !kvm_is_reserved_pfn(pfn) &&
+                       PageTransCompound(pfn_to_page(pfn))) {

Not your fault, but PageTransCompound is very unhappy naming, as it also yields true for PageHuge. Suggestion: document this check covers static hugetlbfs, or switch to PageCompound() check.

A slightly bolder approach would be to refactor and reuse the nearly identical check done in transparent_hugepage_adjust, instead of open-coding here. In essence this code is asking for the same check, plus the out-of-band check for static hugepages.

Quoted 17 lines

arch/x86/kvm/mmu.c +15 -0

+                       drop_spte(kvm, sptep);
+                       sptep = rmap_get_first(*rmapp, &iter);
+                       need_tlb_flush = 1;
+               } else
+                       sptep = rmap_get_next(&iter);
+       }
+
+       return need_tlb_flush;
+}
+
+void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
+                       struct kvm_memory_slot *memslot)
+{
+       bool flush = false;
+       unsigned long *rmapp;

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4515

Quoted 10 lines

arch/x86/kvm/mmu.c +10 -0

+       unsigned long last_index, index;
+       gfn_t gfn_start, gfn_end;
+
+       spin_lock(&kvm->mmu_lock);
+
+       gfn_start = memslot->base_gfn;
+       gfn_end = memslot->base_gfn + memslot->npages - 1;
+
+       if (gfn_start >= gfn_end)
+               goto out;

I don't understand the value of this check here. Are we looking for a broken memslot? Shouldn't this be a BUG_ON? Is this the place to care about these things? npages is capped to KVM_MEM_MAX_NR_PAGES, i.e. 2^31. A 64 bit overflow would be caused by a gigantic gfn_start which would be trouble in many other ways.

All this to say: please remove the above 5 lines and make code simpler.

Hi Andres,
On Fri, Apr 10, 2015 at 11:05:26AM -0700, Andres Lagar-Cavilla wrote:
[...]

Quoted 4 lines

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4490

Quoted 3 lines

arch/x86/kvm/mmu.c +3 -0

+               if (sp->role.direct &&
+                       !kvm_is_reserved_pfn(pfn) &&
+                       PageTransCompound(pfn_to_page(pfn))) {

Quoted 8 lines

Not your fault, but PageTransCompound is very unhappy naming, as it also yields true for PageHuge. Suggestion: document this check covers static hugetlbfs, or switch to PageCompound() check.

A slightly bolder approach would be to refactor and reuse the nearly identical check done in transparent_hugepage_adjust, instead of open-coding here. In essence this code is asking for the same check, plus the out-of-band check for static hugepages.

PageCompound() check still return true for both transparent huge pages and hugetlbfs pages, !PageHuge(page) && PageTransHuge(page) check can guarantee to catch the right transparent huge pages just as my old commit e76d30e20be5fc ("mm/hwpoison: fix test for a transparent huge page"). I will send a patch to fix this.

Quoted 34 lines

Quoted 17 lines

arch/x86/kvm/mmu.c +15 -0

+                       drop_spte(kvm, sptep);
+                       sptep = rmap_get_first(*rmapp, &iter);
+                       need_tlb_flush = 1;
+               } else
+                       sptep = rmap_get_next(&iter);
+       }
+
+       return need_tlb_flush;
+}
+
+void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
+                       struct kvm_memory_slot *memslot)
+{
+       bool flush = false;
+       unsigned long *rmapp;

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4515

Quoted 10 lines

arch/x86/kvm/mmu.c +10 -0

+       unsigned long last_index, index;
+       gfn_t gfn_start, gfn_end;
+
+       spin_lock(&kvm->mmu_lock);
+
+       gfn_start = memslot->base_gfn;
+       gfn_end = memslot->base_gfn + memslot->npages - 1;
+
+       if (gfn_start >= gfn_end)
+               goto out;

I don't understand the value of this check here. Are we looking for a broken memslot? Shouldn't this be a BUG_ON? Is this the place to care about these things? npages is capped to KVM_MEM_MAX_NR_PAGES, i.e. 2^31. A 64 bit overflow would be caused by a gigantic gfn_start which would be trouble in many other ways.

Quoted 1 line

All this to say: please remove the above 5 lines and make code simpler.

I will send a patch to cleanup it. Thanks for your review. :)

Regards,
Wanpeng Li

Quoted 16 lines · Wanpeng Li

Hi Andres,
On Fri, Apr 10, 2015 at 11:05:26AM -0700, Andres Lagar-Cavilla wrote:
[...]

Quoted 4 lines

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4490

Quoted 3 lines

arch/x86/kvm/mmu.c +3 -0

+               if (sp->role.direct &&
+                       !kvm_is_reserved_pfn(pfn) &&
+                       PageTransCompound(pfn_to_page(pfn))) {

Quoted 8 lines

Not your fault, but PageTransCompound is very unhappy naming, as it also yields true for PageHuge. Suggestion: document this check covers static hugetlbfs, or switch to PageCompound() check.

A slightly bolder approach would be to refactor and reuse the nearly identical check done in transparent_hugepage_adjust, instead of open-coding here. In essence this code is asking for the same check, plus the out-of-band check for static hugepages.

Quoted 6 lines

PageCompound() check still return true for both transparent huge pages and hugetlbfs pages, !PageHuge(page) && PageTransHuge(page) check can guarantee to catch the right transparent huge pages just as my old commit e76d30e20be5fc ("mm/hwpoison: fix test for a transparent huge page"). I will send a patch to fix this.

Why would you want to "fix" it that way? Aren't static hugepages supported?

(PageAnon is an inline check and much cheaper than !PageHuge(), which is an actual function call)

Please consider my suggestion about refactoring the similar checks in transparent_hugepage_adjust.

Thanks a ton
Andres

Quoted 41 lines

Quoted 34 lines

Quoted 17 lines

arch/x86/kvm/mmu.c +15 -0

+                       drop_spte(kvm, sptep);
+                       sptep = rmap_get_first(*rmapp, &iter);
+                       need_tlb_flush = 1;
+               } else
+                       sptep = rmap_get_next(&iter);
+       }
+
+       return need_tlb_flush;
+}
+
+void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
+                       struct kvm_memory_slot *memslot)
+{
+       bool flush = false;
+       unsigned long *rmapp;

arch/x86/kvm/mmu.c void kvm_mmu_slot_remove_write_access(struct kvm *kvm, · line 4515

Quoted 10 lines

arch/x86/kvm/mmu.c +10 -0

+       unsigned long last_index, index;
+       gfn_t gfn_start, gfn_end;
+
+       spin_lock(&kvm->mmu_lock);
+
+       gfn_start = memslot->base_gfn;
+       gfn_end = memslot->base_gfn + memslot->npages - 1;
+
+       if (gfn_start >= gfn_end)
+               goto out;

I don't understand the value of this check here. Are we looking for a broken memslot? Shouldn't this be a BUG_ON? Is this the place to care about these things? npages is capped to KVM_MEM_MAX_NR_PAGES, i.e. 2^31. A 64 bit overflow would be caused by a gigantic gfn_start which would be trouble in many other ways.

Quoted 1 line

All this to say: please remove the above 5 lines and make code simpler.

I will send a patch to cleanup it. Thanks for your review. :)

Regards,
Wanpeng Li

Signature · 1 line

Andres Lagar-Cavilla | Google Kernel Team | [email protected]