[PATCH mm-unstable v2 09/10] kvm/x86: add kvm_arch_test_clear_young()
From: Yu Zhao <hidden>
Date: 2023-05-26 23:45:41
Also in:
kvm, kvmarm, linux-arm-kernel, linux-doc, linux-mm, linux-trace-kernel, lkml
Subsystem:
kernel virtual machine for x86 (kvm/x86), the rest, x86 architecture (32-bit and 64-bit) · Maintainers:
Sean Christopherson, Paolo Bonzini, Linus Torvalds, Thomas Gleixner, Ingo Molnar, Borislav Petkov, Dave Hansen
Implement kvm_arch_test_clear_young() to support the fast path in mmu_notifier_ops->test_clear_young(). It focuses on a simple case, i.e., TDP MMU sets the accessed bit in KVM PTEs and VMs are not nested, where it can rely on RCU and clear_bit() to safely clear the accessed bit without taking kvm->mmu_lock. Complex cases fall back to the existing slow path where kvm->mmu_lock is then taken. Signed-off-by: Yu Zhao <redacted> --- arch/x86/include/asm/kvm_host.h | 7 +++++++ arch/x86/kvm/mmu/tdp_mmu.c | 34 +++++++++++++++++++++++++++++++++ 2 files changed, 41 insertions(+)
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 753c67072c47..d6dfdebe3d94 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h@@ -2223,4 +2223,11 @@ int memslot_rmap_alloc(struct kvm_memory_slot *slot, unsigned long npages); */ #define KVM_EXIT_HYPERCALL_MBZ GENMASK_ULL(31, 1) +#define kvm_arch_has_test_clear_young kvm_arch_has_test_clear_young +static inline bool kvm_arch_has_test_clear_young(void) +{ + return IS_ENABLED(CONFIG_X86_64) && + (!IS_REACHABLE(CONFIG_KVM) || (tdp_mmu_enabled && shadow_accessed_mask)); +} + #endif /* _ASM_X86_KVM_HOST_H */
diff --git a/arch/x86/kvm/mmu/tdp_mmu.c b/arch/x86/kvm/mmu/tdp_mmu.c
index 08340219c35a..6875a819e007 100644
--- a/arch/x86/kvm/mmu/tdp_mmu.c
+++ b/arch/x86/kvm/mmu/tdp_mmu.c@@ -1232,6 +1232,40 @@ bool kvm_tdp_mmu_test_age_gfn(struct kvm *kvm, struct kvm_gfn_range *range) return kvm_tdp_mmu_handle_gfn(kvm, range, test_age_gfn); } +bool kvm_arch_test_clear_young(struct kvm *kvm, struct kvm_gfn_range *range) +{ + struct kvm_mmu_page *root; + int offset = ffs(shadow_accessed_mask) - 1; + + if (kvm_shadow_root_allocated(kvm)) + return true; + + rcu_read_lock(); + + list_for_each_entry_rcu(root, &kvm->arch.tdp_mmu_roots, link) { + struct tdp_iter iter; + + if (kvm_mmu_page_as_id(root) != range->slot->as_id) + continue; + + tdp_root_for_each_leaf_pte(iter, root, range->start, range->end) { + u64 *sptep = rcu_dereference(iter.sptep); + + VM_WARN_ON_ONCE(!page_count(virt_to_page(sptep))); + + if (!(iter.old_spte & shadow_accessed_mask)) + continue; + + if (kvm_should_clear_young(range, iter.gfn)) + clear_bit(offset, (unsigned long *)sptep); + } + } + + rcu_read_unlock(); + + return false; +} + static bool set_spte_gfn(struct kvm *kvm, struct tdp_iter *iter, struct kvm_gfn_range *range) {
--
2.41.0.rc0.172.g3f132b7071-goog