KVM: Handle async PF in a guest.

When async PF capability is detected hook up special page fault handler
that will handle async page fault events and bypass other page faults to
regular page fault handler. Also add async PF handling to nested SVM
emulation. Async PF always generates exit to L1 where vcpu thread will
be scheduled out until page is available.

Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

1 files changed, 36 insertions, 9 deletions

diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index b81a9b7c2ca4..93e8120b8021 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -31,6 +31,7 @@
 
 #include <asm/tlbflush.h>
 #include <asm/desc.h>
+#include <asm/kvm_para.h>
 
 #include <asm/virtext.h>
 #include "trace.h"
@@ -133,6 +134,7 @@ struct vcpu_svm {
 
         unsigned int3_injected;
         unsigned long int3_rip;
+        u32 apf_reason;
 };
 
 #define MSR_INVALID                     0xffffffffU
@@ -1383,16 +1385,33 @@ static void svm_set_dr7(struct kvm_vcpu *vcpu, unsigned long value)
 
 static int pf_interception(struct vcpu_svm *svm)
 {
-        u64 fault_address;
+        u64 fault_address = svm->vmcb->control.exit_info_2;
         u32 error_code;
+        int r = 1;
 
-        fault_address  = svm->vmcb->control.exit_info_2;
-        error_code = svm->vmcb->control.exit_info_1;
+        switch (svm->apf_reason) {
+        default:
+                error_code = svm->vmcb->control.exit_info_1;
 
-        trace_kvm_page_fault(fault_address, error_code);
-        if (!npt_enabled && kvm_event_needs_reinjection(&svm->vcpu))
-                kvm_mmu_unprotect_page_virt(&svm->vcpu, fault_address);
-        return kvm_mmu_page_fault(&svm->vcpu, fault_address, error_code);
+                trace_kvm_page_fault(fault_address, error_code);
+                if (!npt_enabled && kvm_event_needs_reinjection(&svm->vcpu))
+                        kvm_mmu_unprotect_page_virt(&svm->vcpu, fault_address);
+                r = kvm_mmu_page_fault(&svm->vcpu, fault_address, error_code);
+                break;
+        case KVM_PV_REASON_PAGE_NOT_PRESENT:
+                svm->apf_reason = 0;
+                local_irq_disable();
+                kvm_async_pf_task_wait(fault_address);
+                local_irq_enable();
+                break;
+        case KVM_PV_REASON_PAGE_READY:
+                svm->apf_reason = 0;
+                local_irq_disable();
+                kvm_async_pf_task_wake(fault_address);
+                local_irq_enable();
+                break;
+        }
+        return r;
 }
 
 static int db_interception(struct vcpu_svm *svm)
@@ -1836,8 +1855,8 @@ static int nested_svm_exit_special(struct vcpu_svm *svm)
                         return NESTED_EXIT_HOST;
                 break;
         case SVM_EXIT_EXCP_BASE + PF_VECTOR:
-                /* When we're shadowing, trap PFs */
-                if (!npt_enabled)
+                /* When we're shadowing, trap PFs, but not async PF */
+                if (!npt_enabled && svm->apf_reason == 0)
                         return NESTED_EXIT_HOST;
                 break;
         case SVM_EXIT_EXCP_BASE + NM_VECTOR:
@@ -1893,6 +1912,10 @@ static int nested_svm_intercept(struct vcpu_svm *svm)
                 u32 excp_bits = 1 << (exit_code - SVM_EXIT_EXCP_BASE);
                 if (svm->nested.intercept_exceptions & excp_bits)
                         vmexit = NESTED_EXIT_DONE;
+                /* async page fault always cause vmexit */
+                else if ((exit_code == SVM_EXIT_EXCP_BASE + PF_VECTOR) &&
+                         svm->apf_reason != 0)
+                        vmexit = NESTED_EXIT_DONE;
                 break;
         }
         case SVM_EXIT_ERR: {
@@ -3414,6 +3437,10 @@ static void svm_vcpu_run(struct kvm_vcpu *vcpu)
 
         svm->next_rip = 0;
 
+        /* if exit due to PF check for async PF */
+        if (svm->vmcb->control.exit_code == SVM_EXIT_EXCP_BASE + PF_VECTOR)
+                svm->apf_reason = kvm_read_and_reset_pf_reason();
+
         if (npt_enabled) {
                 vcpu->arch.regs_avail &= ~(1 << VCPU_EXREG_PDPTR);
                 vcpu->arch.regs_dirty &= ~(1 << VCPU_EXREG_PDPTR);