Merge tag 'kvm-3.6-1' of git://git.kernel.org/pub/scm/virt/kvm/kvm

Pull KVM updates from Avi Kivity:
 "Highlights include
  - full big real mode emulation on pre-Westmere Intel hosts (can be
    disabled with emulate_invalid_guest_state=0)
  - relatively small ppc and s390 updates
  - PCID/INVPCID support in guests
  - EOI avoidance; 3.6 guests should perform better on 3.6 hosts on
    interrupt intensive workloads)
  - Lockless write faults during live migration
  - EPT accessed/dirty bits support for new Intel processors"

Fix up conflicts in:
 - Documentation/virtual/kvm/api.txt:

   Stupid subchapter numbering, added next to each other.

 - arch/powerpc/kvm/booke_interrupts.S:

   PPC asm changes clashing with the KVM fixes

 - arch/s390/include/asm/sigp.h, arch/s390/kvm/sigp.c:

   Duplicated commits through the kvm tree and the s390 tree, with
   subsequent edits in the KVM tree.

* tag 'kvm-3.6-1' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (93 commits)
  KVM: fix race with level interrupts
  x86, hyper: fix build with !CONFIG_KVM_GUEST
  Revert "apic: fix kvm build on UP without IOAPIC"
  KVM guest: switch to apic_set_eoi_write, apic_write
  apic: add apic_set_eoi_write for PV use
  KVM: VMX: Implement PCID/INVPCID for guests with EPT
  KVM: Add x86_hyper_kvm to complete detect_hypervisor_platform check
  KVM: PPC: Critical interrupt emulation support
  KVM: PPC: e500mc: Fix tlbilx emulation for 64-bit guests
  KVM: PPC64: booke: Set interrupt computation mode for 64-bit host
  KVM: PPC: bookehv: Add ESR flag to Data Storage Interrupt
  KVM: PPC: bookehv64: Add support for std/ld emulation.
  booke: Added crit/mc exception handler for e500v2
  booke/bookehv: Add host crit-watchdog exception support
  KVM: MMU: document mmu-lock and fast page fault
  KVM: MMU: fix kvm_mmu_pagetable_walk tracepoint
  KVM: MMU: trace fast page fault
  KVM: MMU: fast path of handling guest page fault
  KVM: MMU: introduce SPTE_MMU_WRITEABLE bit
  KVM: MMU: fold tlb flush judgement into mmu_spte_update
  ...

3 files changed, 81 insertions, 3 deletions

diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c
index c421512ca5e..98e24131ff3 100644
--- a/arch/x86/kernel/apic/apic.c
+++ b/arch/x86/kernel/apic/apic.c
@@ -2143,6 +2143,23 @@ int default_cpu_mask_to_apicid_and(const struct cpumask *cpumask,
 }
 
 /*
+ * Override the generic EOI implementation with an optimized version.
+ * Only called during early boot when only one CPU is active and with
+ * interrupts disabled, so we know this does not race with actual APIC driver
+ * use.
+ */
+void __init apic_set_eoi_write(void (*eoi_write)(u32 reg, u32 v))
+{
+	struct apic **drv;
+
+	for (drv = __apicdrivers; drv < __apicdrivers_end; drv++) {
+		/* Should happen once for each apic */
+		WARN_ON((*drv)->eoi_write == eoi_write);
+		(*drv)->eoi_write = eoi_write;
+	}
+}
+
+/*
  * Power management
  */
 #ifdef CONFIG_PM
diff --git a/arch/x86/kernel/cpu/hypervisor.c b/arch/x86/kernel/cpu/hypervisor.c
index 755f64fb074..a8f8fa9769d 100644
--- a/arch/x86/kernel/cpu/hypervisor.c
+++ b/arch/x86/kernel/cpu/hypervisor.c
@@ -37,6 +37,9 @@ static const __initconst struct hypervisor_x86 * const hypervisors[] =
 #endif
 	&x86_hyper_vmware,
 	&x86_hyper_ms_hyperv,
+#ifdef CONFIG_KVM_GUEST
+	&x86_hyper_kvm,
+#endif
 };
 
 const struct hypervisor_x86 *x86_hyper;
diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
index e554e5ad2fe..c1d61ee4b4f 100644
--- a/arch/x86/kernel/kvm.c
+++ b/arch/x86/kernel/kvm.c
@@ -39,6 +39,9 @@
 #include <asm/desc.h>
 #include <asm/tlbflush.h>
 #include <asm/idle.h>
+#include <asm/apic.h>
+#include <asm/apicdef.h>
+#include <asm/hypervisor.h>
 
 static int kvmapf = 1;
 
@@ -283,6 +286,22 @@ static void kvm_register_steal_time(void)
 		cpu, __pa(st));
 }
 
+static DEFINE_PER_CPU(unsigned long, kvm_apic_eoi) = KVM_PV_EOI_DISABLED;
+
+static void kvm_guest_apic_eoi_write(u32 reg, u32 val)
+{
+	/**
+	 * This relies on __test_and_clear_bit to modify the memory
+	 * in a way that is atomic with respect to the local CPU.
+	 * The hypervisor only accesses this memory from the local CPU so
+	 * there's no need for lock or memory barriers.
+	 * An optimization barrier is implied in apic write.
+	 */
+	if (__test_and_clear_bit(KVM_PV_EOI_BIT, &__get_cpu_var(kvm_apic_eoi)))
+		return;
+	apic_write(APIC_EOI, APIC_EOI_ACK);
+}
+
 void __cpuinit kvm_guest_cpu_init(void)
 {
 	if (!kvm_para_available())
@@ -300,11 +319,20 @@ void __cpuinit kvm_guest_cpu_init(void)
 		       smp_processor_id());
 	}
 
+	if (kvm_para_has_feature(KVM_FEATURE_PV_EOI)) {
+		unsigned long pa;
+		/* Size alignment is implied but just to make it explicit. */
+		BUILD_BUG_ON(__alignof__(kvm_apic_eoi) < 4);
+		__get_cpu_var(kvm_apic_eoi) = 0;
+		pa = __pa(&__get_cpu_var(kvm_apic_eoi)) | KVM_MSR_ENABLED;
+		wrmsrl(MSR_KVM_PV_EOI_EN, pa);
+	}
+
 	if (has_steal_clock)
 		kvm_register_steal_time();
 }
 
-static void kvm_pv_disable_apf(void *unused)
+static void kvm_pv_disable_apf(void)
 {
 	if (!__get_cpu_var(apf_reason).enabled)
 		return;
@@ -316,11 +344,23 @@ static void kvm_pv_disable_apf(void *unused)
 	       smp_processor_id());
 }
 
+static void kvm_pv_guest_cpu_reboot(void *unused)
+{
+	/*
+	 * We disable PV EOI before we load a new kernel by kexec,
+	 * since MSR_KVM_PV_EOI_EN stores a pointer into old kernel's memory.
+	 * New kernel can re-enable when it boots.
+	 */
+	if (kvm_para_has_feature(KVM_FEATURE_PV_EOI))
+		wrmsrl(MSR_KVM_PV_EOI_EN, 0);
+	kvm_pv_disable_apf();
+}
+
 static int kvm_pv_reboot_notify(struct notifier_block *nb,
 				unsigned long code, void *unused)
 {
 	if (code == SYS_RESTART)
-		on_each_cpu(kvm_pv_disable_apf, NULL, 1);
+		on_each_cpu(kvm_pv_guest_cpu_reboot, NULL, 1);
 	return NOTIFY_DONE;
 }
 
@@ -371,7 +411,9 @@ static void __cpuinit kvm_guest_cpu_online(void *dummy)
 static void kvm_guest_cpu_offline(void *dummy)
 {
 	kvm_disable_steal_time();
-	kvm_pv_disable_apf(NULL);
+	if (kvm_para_has_feature(KVM_FEATURE_PV_EOI))
+		wrmsrl(MSR_KVM_PV_EOI_EN, 0);
+	kvm_pv_disable_apf();
 	apf_task_wake_all();
 }
 
@@ -424,6 +466,9 @@ void __init kvm_guest_init(void)
 		pv_time_ops.steal_clock = kvm_steal_clock;
 	}
 
+	if (kvm_para_has_feature(KVM_FEATURE_PV_EOI))
+		apic_set_eoi_write(kvm_guest_apic_eoi_write);
+
 #ifdef CONFIG_SMP
 	smp_ops.smp_prepare_boot_cpu = kvm_smp_prepare_boot_cpu;
 	register_cpu_notifier(&kvm_cpu_notifier);
@@ -432,6 +477,19 @@ void __init kvm_guest_init(void)
 #endif
 }
 
+static bool __init kvm_detect(void)
+{
+	if (!kvm_para_available())
+		return false;
+	return true;
+}
+
+const struct hypervisor_x86 x86_hyper_kvm __refconst = {
+	.name			= "KVM",
+	.detect			= kvm_detect,
+};
+EXPORT_SYMBOL_GPL(x86_hyper_kvm);
+
 static __init int activate_jump_labels(void)
 {
 	if (has_steal_clock) {