[RFC: timer passthrough 1/9] KVM: vmx: hook set_next_event for getting the host tscd

[Zhimin Feng]

In order to get the host tscd value,
we need to hook set_next_event function

Signed-off-by: Zhimin Feng <fengzhimin@xxxxxxxxxxxxx>
---
 arch/x86/include/asm/kvm_host.h | 21 +++++++++++++++++
 arch/x86/kvm/vmx/vmx.c          | 51 +++++++++++++++++++++++++++++++++++++++++
 arch/x86/kvm/x86.c              |  1 +
 kernel/time/tick-common.c       |  1 +
 4 files changed, 74 insertions(+)

diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 7e5f33a0d0e2..eb6a611963b7 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -34,6 +34,7 @@
 #include <asm/kvm_page_track.h>
 #include <asm/kvm_vcpu_regs.h>
 #include <asm/hyperv-tlfs.h>
+#include <linux/clockchips.h>
 
 #define __KVM_HAVE_ARCH_VCPU_DEBUGFS
 
@@ -520,6 +521,24 @@ struct kvm_vcpu_hv {
 	cpumask_t tlb_flush;
 };
 
+enum tick_device_mode {
+	TICKDEV_MODE_PERIODIC,
+	TICKDEV_MODE_ONESHOT,
+};
+
+struct tick_device {
+	struct clock_event_device *evtdev;
+	enum tick_device_mode mode;
+};
+
+struct timer_passth_info {
+	u64 host_tscd;
+	struct clock_event_device *curr_dev;
+
+	void (*orig_event_handler)(struct clock_event_device *dev);
+	int (*orig_set_next_event)(unsigned long evt, struct clock_event_device *dev);
+};
+
 struct kvm_vcpu_arch {
 	/*
 	 * rip and regs accesses must go through
@@ -805,6 +824,8 @@ struct kvm_vcpu_arch {
 		 */
 		bool enforce;
 	} pv_cpuid;
+
+	bool timer_passth_enable;
 };
 
 struct kvm_lpage_info {
diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
index 47b8357b9751..38b8d80fa157 100644
--- a/arch/x86/kvm/vmx/vmx.c
+++ b/arch/x86/kvm/vmx/vmx.c
@@ -128,6 +128,11 @@ static bool __read_mostly enable_preemption_timer = 1;
 module_param_named(preemption_timer, enable_preemption_timer, bool, S_IRUGO);
 #endif
 
+static bool __read_mostly enable_timer_passth;
+#ifdef CONFIG_X86_64
+module_param_named(timer_passth, enable_timer_passth, bool, 0444);
+#endif
+
 extern bool __read_mostly allow_smaller_maxphyaddr;
 module_param(allow_smaller_maxphyaddr, bool, S_IRUGO);
 
@@ -220,6 +225,46 @@ static const struct {
 	[VMENTER_L1D_FLUSH_NOT_REQUIRED] = {"not required", false},
 };
 
+#define TSC_DIVISOR  8
+static DEFINE_PER_CPU(struct timer_passth_info, passth_info);
+
+static int override_lapic_next_event(unsigned long delta,
+		struct clock_event_device *evt)
+{
+	struct timer_passth_info *local_timer_info;
+	u64 tsc;
+	u64 tscd;
+
+	local_timer_info = &per_cpu(passth_info, smp_processor_id());
+	tsc = rdtsc();
+	tscd = tsc + (((u64) delta) * TSC_DIVISOR);
+	local_timer_info->host_tscd = tscd;
+	wrmsrl(MSR_IA32_TSCDEADLINE, tscd);
+
+	return 0;
+}
+
+static void vmx_host_timer_passth_init(void *junk)
+{
+	struct timer_passth_info *local_timer_info;
+	int cpu = smp_processor_id();
+
+	local_timer_info = &per_cpu(passth_info, cpu);
+	local_timer_info->curr_dev = per_cpu(tick_cpu_device, cpu).evtdev;
+	local_timer_info->orig_set_next_event =
+		local_timer_info->curr_dev->set_next_event;
+	local_timer_info->curr_dev->set_next_event = override_lapic_next_event;
+}
+
+static void vmx_host_timer_restore(void *junk)
+{
+	struct timer_passth_info *local_timer_info;
+
+	local_timer_info = &per_cpu(passth_info, smp_processor_id());
+	local_timer_info->curr_dev->set_next_event =
+		local_timer_info->orig_set_next_event;
+}
+
 #define L1D_CACHE_ORDER 4
 static void *vmx_l1d_flush_pages;
 
@@ -7573,6 +7618,9 @@ static void vmx_migrate_timers(struct kvm_vcpu *vcpu)
 
 static void hardware_unsetup(void)
 {
+	if (enable_timer_passth)
+		on_each_cpu(vmx_host_timer_restore, NULL, 1);
+
 	if (nested)
 		nested_vmx_hardware_unsetup();
 
@@ -7884,6 +7932,9 @@ static __init int hardware_setup(void)
 
 	vmx_set_cpu_caps();
 
+	if (enable_timer_passth)
+		on_each_cpu(vmx_host_timer_passth_init, NULL, 1);
+
 	r = alloc_kvm_area();
 	if (r)
 		nested_vmx_hardware_unsetup();
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index e545a8a613b1..5d353a9c9881 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -9911,6 +9911,7 @@ int kvm_arch_vcpu_create(struct kvm_vcpu *vcpu)
 
 	vcpu->arch.pending_external_vector = -1;
 	vcpu->arch.preempted_in_kernel = false;
+	vcpu->arch.timer_passth_enable = false;
 
 	kvm_hv_vcpu_init(vcpu);
 
diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c
index 6c9c342dd0e5..bc50f4a1a7c0 100644
--- a/kernel/time/tick-common.c
+++ b/kernel/time/tick-common.c
@@ -26,6 +26,7 @@
  * Tick devices
  */
 DEFINE_PER_CPU(struct tick_device, tick_cpu_device);
+EXPORT_SYMBOL_GPL(tick_cpu_device);
 /*
  * Tick next event: keeps track of the tick time
  */
-- 
2.11.0