diff mbox series

[01/15] KVM: x86/mmu: Remove redundant flushes when disabling dirty logging

Message ID 20211115234603.2908381-2-bgardon@google.com (mailing list archive)
State New, archived
Headers show
Series Currently disabling dirty logging with the TDP MMU is extremely slow. On a 96 vCPU / 96G VM it takes ~45 seconds to disable dirty logging with the TDP MMU, as opposed to ~3.5 seconds with the legacy MMU. This series optimizes TLB flushes and intro | expand

Commit Message

Ben Gardon Nov. 15, 2021, 11:45 p.m. UTC 
tdp_mmu_zap_spte_atomic flushes on every zap already, so no need to
flush again after it's done.

Reviewed-by: David Matlack <dmatlack@google.com>

Signed-off-by: Ben Gardon <bgardon@google.com>
---
 arch/x86/kvm/mmu/mmu.c     |  4 +---
 arch/x86/kvm/mmu/tdp_mmu.c | 21 ++++++---------------
 arch/x86/kvm/mmu/tdp_mmu.h |  5 ++---
 3 files changed, 9 insertions(+), 21 deletions(-)

Comments

Paolo Bonzini Nov. 18, 2021, 8:26 a.m. UTC | #1 
On 11/16/21 00:45, Ben Gardon wrote:
> tdp_mmu_zap_spte_atomic flushes on every zap already, so no need to
> flush again after it's done.
> 
> Reviewed-by: David Matlack <dmatlack@google.com>
> 
> Signed-off-by: Ben Gardon <bgardon@google.com>
> ---
>   arch/x86/kvm/mmu/mmu.c     |  4 +---
>   arch/x86/kvm/mmu/tdp_mmu.c | 21 ++++++---------------
>   arch/x86/kvm/mmu/tdp_mmu.h |  5 ++---
>   3 files changed, 9 insertions(+), 21 deletions(-)
> 
> diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
> index 354d2ca92df4..baa94acab516 100644
> --- a/arch/x86/kvm/mmu/mmu.c
> +++ b/arch/x86/kvm/mmu/mmu.c
> @@ -5870,9 +5870,7 @@ void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
>   
>   	if (is_tdp_mmu_enabled(kvm)) {
>   		read_lock(&kvm->mmu_lock);
> -		flush = kvm_tdp_mmu_zap_collapsible_sptes(kvm, slot, flush);
> -		if (flush)
> -			kvm_arch_flush_remote_tlbs_memslot(kvm, slot);
> +		kvm_tdp_mmu_zap_collapsible_sptes(kvm, slot);
>   		read_unlock(&kvm->mmu_lock);
>   	}
>   }
> diff --git a/arch/x86/kvm/mmu/tdp_mmu.c b/arch/x86/kvm/mmu/tdp_mmu.c
> index 7c5dd83e52de..b3c78568ae60 100644
> --- a/arch/x86/kvm/mmu/tdp_mmu.c
> +++ b/arch/x86/kvm/mmu/tdp_mmu.c
> @@ -1364,10 +1364,9 @@ void kvm_tdp_mmu_clear_dirty_pt_masked(struct kvm *kvm,
>    * Clear leaf entries which could be replaced by large mappings, for
>    * GFNs within the slot.
>    */
> -static bool zap_collapsible_spte_range(struct kvm *kvm,
> +static void zap_collapsible_spte_range(struct kvm *kvm,
>   				       struct kvm_mmu_page *root,
> -				       const struct kvm_memory_slot *slot,
> -				       bool flush)
> +				       const struct kvm_memory_slot *slot)
>   {
>   	gfn_t start = slot->base_gfn;
>   	gfn_t end = start + slot->npages;
> @@ -1378,10 +1377,8 @@ static bool zap_collapsible_spte_range(struct kvm *kvm,
>   
>   	tdp_root_for_each_pte(iter, root, start, end) {
>   retry:
> -		if (tdp_mmu_iter_cond_resched(kvm, &iter, flush, true)) {
> -			flush = false;
> +		if (tdp_mmu_iter_cond_resched(kvm, &iter, false, true))
>   			continue;
> -		}
>   
>   		if (!is_shadow_present_pte(iter.old_spte) ||
>   		    !is_last_spte(iter.old_spte, iter.level))
> @@ -1401,30 +1398,24 @@ static bool zap_collapsible_spte_range(struct kvm *kvm,
>   			iter.old_spte = READ_ONCE(*rcu_dereference(iter.sptep));
>   			goto retry;
>   		}
> -		flush = true;
>   	}
>   
>   	rcu_read_unlock();
> -
> -	return flush;
>   }
>   
>   /*
>    * Clear non-leaf entries (and free associated page tables) which could
>    * be replaced by large mappings, for GFNs within the slot.
>    */
> -bool kvm_tdp_mmu_zap_collapsible_sptes(struct kvm *kvm,
> -				       const struct kvm_memory_slot *slot,
> -				       bool flush)
> +void kvm_tdp_mmu_zap_collapsible_sptes(struct kvm *kvm,
> +				       const struct kvm_memory_slot *slot)
>   {
>   	struct kvm_mmu_page *root;
>   
>   	lockdep_assert_held_read(&kvm->mmu_lock);
>   
>   	for_each_tdp_mmu_root_yield_safe(kvm, root, slot->as_id, true)
> -		flush = zap_collapsible_spte_range(kvm, root, slot, flush);
> -
> -	return flush;
> +		zap_collapsible_spte_range(kvm, root, slot);
>   }
>   
>   /*
> diff --git a/arch/x86/kvm/mmu/tdp_mmu.h b/arch/x86/kvm/mmu/tdp_mmu.h
> index 476b133544dd..3899004a5d91 100644
> --- a/arch/x86/kvm/mmu/tdp_mmu.h
> +++ b/arch/x86/kvm/mmu/tdp_mmu.h
> @@ -64,9 +64,8 @@ void kvm_tdp_mmu_clear_dirty_pt_masked(struct kvm *kvm,
>   				       struct kvm_memory_slot *slot,
>   				       gfn_t gfn, unsigned long mask,
>   				       bool wrprot);
> -bool kvm_tdp_mmu_zap_collapsible_sptes(struct kvm *kvm,
> -				       const struct kvm_memory_slot *slot,
> -				       bool flush);
> +void kvm_tdp_mmu_zap_collapsible_sptes(struct kvm *kvm,
> +				       const struct kvm_memory_slot *slot);
>   
>   bool kvm_tdp_mmu_write_protect_gfn(struct kvm *kvm,
>   				   struct kvm_memory_slot *slot, gfn_t gfn,
> 

Queued, thanks.

Paolo
diff mbox series

Patch

diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
index 354d2ca92df4..baa94acab516 100644
--- a/arch/x86/kvm/mmu/mmu.c
+++ b/arch/x86/kvm/mmu/mmu.c
@@ -5870,9 +5870,7 @@  void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
 
 	if (is_tdp_mmu_enabled(kvm)) {
 		read_lock(&kvm->mmu_lock);
-		flush = kvm_tdp_mmu_zap_collapsible_sptes(kvm, slot, flush);
-		if (flush)
-			kvm_arch_flush_remote_tlbs_memslot(kvm, slot);
+		kvm_tdp_mmu_zap_collapsible_sptes(kvm, slot);
 		read_unlock(&kvm->mmu_lock);
 	}
 }
diff --git a/arch/x86/kvm/mmu/tdp_mmu.c b/arch/x86/kvm/mmu/tdp_mmu.c
index 7c5dd83e52de..b3c78568ae60 100644
--- a/arch/x86/kvm/mmu/tdp_mmu.c
+++ b/arch/x86/kvm/mmu/tdp_mmu.c
@@ -1364,10 +1364,9 @@  void kvm_tdp_mmu_clear_dirty_pt_masked(struct kvm *kvm,
  * Clear leaf entries which could be replaced by large mappings, for
  * GFNs within the slot.
  */
-static bool zap_collapsible_spte_range(struct kvm *kvm,
+static void zap_collapsible_spte_range(struct kvm *kvm,
 				       struct kvm_mmu_page *root,
-				       const struct kvm_memory_slot *slot,
-				       bool flush)
+				       const struct kvm_memory_slot *slot)
 {
 	gfn_t start = slot->base_gfn;
 	gfn_t end = start + slot->npages;
@@ -1378,10 +1377,8 @@  static bool zap_collapsible_spte_range(struct kvm *kvm,
 
 	tdp_root_for_each_pte(iter, root, start, end) {
 retry:
-		if (tdp_mmu_iter_cond_resched(kvm, &iter, flush, true)) {
-			flush = false;
+		if (tdp_mmu_iter_cond_resched(kvm, &iter, false, true))
 			continue;
-		}
 
 		if (!is_shadow_present_pte(iter.old_spte) ||
 		    !is_last_spte(iter.old_spte, iter.level))
@@ -1401,30 +1398,24 @@  static bool zap_collapsible_spte_range(struct kvm *kvm,
 			iter.old_spte = READ_ONCE(*rcu_dereference(iter.sptep));
 			goto retry;
 		}
-		flush = true;
 	}
 
 	rcu_read_unlock();
-
-	return flush;
 }
 
 /*
  * Clear non-leaf entries (and free associated page tables) which could
  * be replaced by large mappings, for GFNs within the slot.
  */
-bool kvm_tdp_mmu_zap_collapsible_sptes(struct kvm *kvm,
-				       const struct kvm_memory_slot *slot,
-				       bool flush)
+void kvm_tdp_mmu_zap_collapsible_sptes(struct kvm *kvm,
+				       const struct kvm_memory_slot *slot)
 {
 	struct kvm_mmu_page *root;
 
 	lockdep_assert_held_read(&kvm->mmu_lock);
 
 	for_each_tdp_mmu_root_yield_safe(kvm, root, slot->as_id, true)
-		flush = zap_collapsible_spte_range(kvm, root, slot, flush);
-
-	return flush;
+		zap_collapsible_spte_range(kvm, root, slot);
 }
 
 /*
diff --git a/arch/x86/kvm/mmu/tdp_mmu.h b/arch/x86/kvm/mmu/tdp_mmu.h
index 476b133544dd..3899004a5d91 100644
--- a/arch/x86/kvm/mmu/tdp_mmu.h
+++ b/arch/x86/kvm/mmu/tdp_mmu.h
@@ -64,9 +64,8 @@  void kvm_tdp_mmu_clear_dirty_pt_masked(struct kvm *kvm,
 				       struct kvm_memory_slot *slot,
 				       gfn_t gfn, unsigned long mask,
 				       bool wrprot);
-bool kvm_tdp_mmu_zap_collapsible_sptes(struct kvm *kvm,
-				       const struct kvm_memory_slot *slot,
-				       bool flush);
+void kvm_tdp_mmu_zap_collapsible_sptes(struct kvm *kvm,
+				       const struct kvm_memory_slot *slot);
 
 bool kvm_tdp_mmu_write_protect_gfn(struct kvm *kvm,
 				   struct kvm_memory_slot *slot, gfn_t gfn,