Merge tag 'for-linus-4.5-rc0-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip

Pull xen updates from David Vrabel:
"Xen features and fixes for 4.5-rc0:

- Stolen ticks and PV wallclock support for arm/arm64

- Add grant copy ioctl to gntdev device"

* tag 'for-linus-4.5-rc0-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen/gntdev: add ioctl for grant copy
x86/xen: don't reset vcpu_info on a cancelled suspend
xen/gntdev: constify mmu_notifier_ops structures
xen/grant-table: constify gnttab_ops structure
xen/time: use READ_ONCE
xen/x86: convert remaining timespec to timespec64 in xen_pvclock_gtod_notify
xen/x86: support XENPF_settime64
xen/arm: set the system time in Xen via the XENPF_settime64 hypercall
xen/arm: introduce xen_read_wallclock
arm: extend pvclock_wall_clock with sec_hi
xen: introduce XENPF_settime64
xen/arm: introduce HYPERVISOR_platform_op on arm and arm64
xen: rename dom0_op to platform_op
xen/arm: account for stolen ticks
arm64: introduce CONFIG_PARAVIRT, PARAVIRT_TIME_ACCOUNTING and pv_time_ops
arm: introduce CONFIG_PARAVIRT, PARAVIRT_TIME_ACCOUNTING and pv_time_ops
missing include asm/paravirt.h in cputime.c
xen: move xen_setup_runstate_info and get_runstate_snapshot to drivers/xen/time.c

arch/arm/Kconfig

···
 config IOMMU_HELPER
 	def_bool SWIOTLB
 
+config PARAVIRT
+	bool "Enable paravirtualization code"
+	help
+	  This changes the kernel so it can modify itself when it is run
+	  under a hypervisor, potentially improving performance significantly
+	  over full virtualization.
+
+config PARAVIRT_TIME_ACCOUNTING
+	bool "Paravirtual steal time accounting"
+	select PARAVIRT
+	default n
+	help
+	  Select this option to enable fine granularity task steal time
+	  accounting. Time spent executing other tasks in parallel with
+	  the current vCPU is discounted from the vCPU power. To account for
+	  that, there can be a small performance impact.
+
+	  If in doubt, say N here.
+
 config XEN_DOM0
 	def_bool y
 	depends on XEN
···
 	select ARCH_DMA_ADDR_T_64BIT
 	select ARM_PSCI
 	select SWIOTLB_XEN
+	select PARAVIRT
 	help
 	  Say Y if you want to run Linux in a Virtual Machine on Xen on ARM.
 

arch/arm/include/asm/paravirt.h

···
+#ifndef _ASM_ARM_PARAVIRT_H
+#define _ASM_ARM_PARAVIRT_H
+
+#ifdef CONFIG_PARAVIRT
+struct static_key;
+extern struct static_key paravirt_steal_enabled;
+extern struct static_key paravirt_steal_rq_enabled;
+
+struct pv_time_ops {
+	unsigned long long (*steal_clock)(int cpu);
+};
+extern struct pv_time_ops pv_time_ops;
+
+static inline u64 paravirt_steal_clock(int cpu)
+{
+	return pv_time_ops.steal_clock(cpu);
+}
+#endif
+
+#endif

arch/arm/include/asm/xen/hypercall.h

···
 
 #include <xen/interface/xen.h>
 #include <xen/interface/sched.h>
+#include <xen/interface/platform.h>
 
 long privcmd_call(unsigned call, unsigned long a1,
 		unsigned long a2, unsigned long a3,
···
 int HYPERVISOR_physdev_op(int cmd, void *arg);
 int HYPERVISOR_vcpu_op(int cmd, int vcpuid, void *extra_args);
 int HYPERVISOR_tmem_op(void *arg);
+int HYPERVISOR_platform_op_raw(void *arg);
+static inline int HYPERVISOR_platform_op(struct xen_platform_op *op)
+{
+	op->interface_version = XENPF_INTERFACE_VERSION;
+	return HYPERVISOR_platform_op_raw(op);
+}
 int HYPERVISOR_multicall(struct multicall_entry *calls, uint32_t nr);
 
 static inline int

arch/arm/include/asm/xen/interface.h

···
 		(hnd).p = val;				\
 	} while (0)
 
+#define __HYPERVISOR_platform_op_raw __HYPERVISOR_platform_op
+
 #ifndef __ASSEMBLY__
 /* Explicitly size integers that represent pfns in the interface with
  * Xen so that we can have one ABI that works for 32 and 64 bit guests.
···
 	u32   version;
 	u32   sec;
 	u32   nsec;
+	u32   sec_hi;
 } __attribute__((__packed__));
 #endif
 

arch/arm/kernel/Makefile

···
 ifneq ($(CONFIG_ARCH_EBSA110),y)
   obj-y		+= io.o
 endif
+obj-$(CONFIG_PARAVIRT)	+= paravirt.o
 
 head-y			:= head$(MMUEXT).o
 obj-$(CONFIG_DEBUG_LL)	+= debug.o

arch/arm/kernel/paravirt.c

···
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * Copyright (C) 2013 Citrix Systems
+ *
+ * Author: Stefano Stabellini <[e-mail]>
+ */
+
+#include <linux/export.h>
+#include <linux/jump_label.h>
+#include <linux/types.h>
+#include <asm/paravirt.h>
+
+struct static_key paravirt_steal_enabled;
+struct static_key paravirt_steal_rq_enabled;
+
+struct pv_time_ops pv_time_ops;
+EXPORT_SYMBOL_GPL(pv_time_ops);

arch/arm/xen/enlighten.c

···
 #include <xen/page.h>
 #include <xen/interface/sched.h>
 #include <xen/xen-ops.h>
+#include <asm/paravirt.h>
 #include <asm/xen/hypervisor.h>
 #include <asm/xen/hypercall.h>
 #include <asm/system_misc.h>
···
 #include <linux/cpufreq.h>
 #include <linux/cpu.h>
 #include <linux/console.h>
+#include <linux/pvclock_gtod.h>
+#include <linux/time64.h>
+#include <linux/timekeeping.h>
+#include <linux/timekeeper_internal.h>
 
 #include <linux/mm.h>
 
···
 }
 EXPORT_SYMBOL_GPL(xen_unmap_domain_gfn_range);
 
+static unsigned long long xen_stolen_accounting(int cpu)
+{
+	struct vcpu_runstate_info state;
+
+	BUG_ON(cpu != smp_processor_id());
+
+	xen_get_runstate_snapshot(&state);
+
+	WARN_ON(state.state != RUNSTATE_running);
+
+	return state.time[RUNSTATE_runnable] + state.time[RUNSTATE_offline];
+}
+
+static void xen_read_wallclock(struct timespec64 *ts)
+{
+	u32 version;
+	struct timespec64 now, ts_monotonic;
+	struct shared_info *s = HYPERVISOR_shared_info;
+	struct pvclock_wall_clock *wall_clock = &(s->wc);
+
+	/* get wallclock at system boot */
+	do {
+		version = wall_clock->version;
+		rmb();		/* fetch version before time */
+		now.tv_sec  = ((uint64_t)wall_clock->sec_hi << 32) | wall_clock->sec;
+		now.tv_nsec = wall_clock->nsec;
+		rmb();		/* fetch time before checking version */
+	} while ((wall_clock->version & 1) || (version != wall_clock->version));
+
+	/* time since system boot */
+	ktime_get_ts64(&ts_monotonic);
+	*ts = timespec64_add(now, ts_monotonic);
+}
+
+static int xen_pvclock_gtod_notify(struct notifier_block *nb,
+				   unsigned long was_set, void *priv)
+{
+	/* Protected by the calling core code serialization */
+	static struct timespec64 next_sync;
+
+	struct xen_platform_op op;
+	struct timespec64 now, system_time;
+	struct timekeeper *tk = priv;
+
+	now.tv_sec = tk->xtime_sec;
+	now.tv_nsec = (long)(tk->tkr_mono.xtime_nsec >> tk->tkr_mono.shift);
+	system_time = timespec64_add(now, tk->wall_to_monotonic);
+
+	/*
+	 * We only take the expensive HV call when the clock was set
+	 * or when the 11 minutes RTC synchronization time elapsed.
+	 */
+	if (!was_set && timespec64_compare(&now, &next_sync) < 0)
+		return NOTIFY_OK;
+
+	op.cmd = XENPF_settime64;
+	op.u.settime64.mbz = 0;
+	op.u.settime64.secs = now.tv_sec;
+	op.u.settime64.nsecs = now.tv_nsec;
+	op.u.settime64.system_time = timespec64_to_ns(&system_time);
+	(void)HYPERVISOR_platform_op(&op);
+
+	/*
+	 * Move the next drift compensation time 11 minutes
+	 * ahead. That's emulating the sync_cmos_clock() update for
+	 * the hardware RTC.
+	 */
+	next_sync = now;
+	next_sync.tv_sec += 11 * 60;
+
+	return NOTIFY_OK;
+}
+
+static struct notifier_block xen_pvclock_gtod_notifier = {
+	.notifier_call = xen_pvclock_gtod_notify,
+};
+
 static void xen_percpu_init(void)
 {
 	struct vcpu_register_vcpu_info info;
···
 	err = HYPERVISOR_vcpu_op(VCPUOP_register_vcpu_info, cpu, &info);
 	BUG_ON(err);
 	per_cpu(xen_vcpu, cpu) = vcpup;
+
+	xen_setup_runstate_info(cpu);
 
 after_register_vcpu_info:
 	enable_percpu_irq(xen_events_irq, 0);
···
 
 	register_cpu_notifier(&xen_cpu_notifier);
 
+	pv_time_ops.steal_clock = xen_stolen_accounting;
+	static_key_slow_inc(&paravirt_steal_enabled);
+	if (xen_initial_domain())
+		pvclock_gtod_register_notifier(&xen_pvclock_gtod_notifier);
+
 	return 0;
 }
 early_initcall(xen_guest_init);
···
 
 	pm_power_off = xen_power_off;
 	arm_pm_restart = xen_restart;
+	if (!xen_initial_domain()) {
+		struct timespec64 ts;
+		xen_read_wallclock(&ts);
+		do_settimeofday64(&ts);
+	}
 
 	return 0;
 }
···
 EXPORT_SYMBOL_GPL(HYPERVISOR_physdev_op);
 EXPORT_SYMBOL_GPL(HYPERVISOR_vcpu_op);
 EXPORT_SYMBOL_GPL(HYPERVISOR_tmem_op);
+EXPORT_SYMBOL_GPL(HYPERVISOR_platform_op);
 EXPORT_SYMBOL_GPL(HYPERVISOR_multicall);
 EXPORT_SYMBOL_GPL(privcmd_call);

arch/arm/xen/hypercall.S

···
 HYPERCALL2(physdev_op);
 HYPERCALL3(vcpu_op);
 HYPERCALL1(tmem_op);
+HYPERCALL1(platform_op_raw);
 HYPERCALL2(multicall);
 
 ENTRY(privcmd_call)

arch/arm64/Kconfig

···
 	  and the task is only allowed to execute a few safe syscalls
 	  defined by each seccomp mode.
 
+config PARAVIRT
+	bool "Enable paravirtualization code"
+	help
+	  This changes the kernel so it can modify itself when it is run
+	  under a hypervisor, potentially improving performance significantly
+	  over full virtualization.
+
+config PARAVIRT_TIME_ACCOUNTING
+	bool "Paravirtual steal time accounting"
+	select PARAVIRT
+	default n
+	help
+	  Select this option to enable fine granularity task steal time
+	  accounting. Time spent executing other tasks in parallel with
+	  the current vCPU is discounted from the vCPU power. To account for
+	  that, there can be a small performance impact.
+
+	  If in doubt, say N here.
+
 config XEN_DOM0
 	def_bool y
 	depends on XEN
···
 	bool "Xen guest support on ARM64"
 	depends on ARM64 && OF
 	select SWIOTLB_XEN
+	select PARAVIRT
 	help
 	  Say Y if you want to run Linux in a Virtual Machine on Xen on ARM64.
 

arch/arm64/include/asm/paravirt.h

···
+#ifndef _ASM_ARM64_PARAVIRT_H
+#define _ASM_ARM64_PARAVIRT_H
+
+#ifdef CONFIG_PARAVIRT
+struct static_key;
+extern struct static_key paravirt_steal_enabled;
+extern struct static_key paravirt_steal_rq_enabled;
+
+struct pv_time_ops {
+	unsigned long long (*steal_clock)(int cpu);
+};
+extern struct pv_time_ops pv_time_ops;
+
+static inline u64 paravirt_steal_clock(int cpu)
+{
+	return pv_time_ops.steal_clock(cpu);
+}
+#endif
+
+#endif

arch/arm64/kernel/Makefile

···
 arm64-obj-$(CONFIG_PCI)			+= pci.o
 arm64-obj-$(CONFIG_ARMV8_DEPRECATED)	+= armv8_deprecated.o
 arm64-obj-$(CONFIG_ACPI)		+= acpi.o
+arm64-obj-$(CONFIG_PARAVIRT)		+= paravirt.o
 
 obj-y					+= $(arm64-obj-y) vdso/
 obj-m					+= $(arm64-obj-m)

arch/arm64/kernel/paravirt.c

···
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * Copyright (C) 2013 Citrix Systems
+ *
+ * Author: Stefano Stabellini <[e-mail]>
+ */
+
+#include <linux/export.h>
+#include <linux/jump_label.h>
+#include <linux/types.h>
+#include <asm/paravirt.h>
+
+struct static_key paravirt_steal_enabled;
+struct static_key paravirt_steal_rq_enabled;
+
+struct pv_time_ops pv_time_ops;
+EXPORT_SYMBOL_GPL(pv_time_ops);

arch/arm64/xen/hypercall.S

···
 HYPERCALL2(physdev_op);
 HYPERCALL3(vcpu_op);
 HYPERCALL1(tmem_op);
+HYPERCALL1(platform_op_raw);
 HYPERCALL2(multicall);
 
 ENTRY(privcmd_call)

arch/x86/include/asm/xen/hypercall.h

···
 }
 
 static inline int
-HYPERVISOR_dom0_op(struct xen_platform_op *platform_op)
+HYPERVISOR_platform_op(struct xen_platform_op *op)
 {
-	platform_op->interface_version = XENPF_INTERFACE_VERSION;
-	return _hypercall1(int, dom0_op, platform_op);
+	op->interface_version = XENPF_INTERFACE_VERSION;
+	return _hypercall1(int, platform_op, op);
 }
 
 static inline int

arch/x86/xen/apic.c

···
 	if (reg != APIC_ID)
 		return 0;
 
-	ret = HYPERVISOR_dom0_op(&op);
+	ret = HYPERVISOR_platform_op(&op);
 	if (ret)
 		return 0;
 

arch/x86/xen/enlighten.c

···
 
 	set_xen_guest_handle(op.u.set_pminfo.pdc, buf);
 
-	if ((HYPERVISOR_dom0_op(&op) == 0) &&
+	if ((HYPERVISOR_platform_op(&op) == 0) &&
 	    (buf[2] & (ACPI_PDC_C_C1_FFH | ACPI_PDC_C_C2C3_FFH))) {
 		cpuid_leaf5_ecx_val = cx;
 		cpuid_leaf5_edx_val = dx;
···
 		info->params.length = sizeof(info->params);
 		set_xen_guest_handle(op.u.firmware_info.u.disk_info.edd_params,
 				     &info->params);
-		ret = HYPERVISOR_dom0_op(&op);
+		ret = HYPERVISOR_platform_op(&op);
 		if (ret)
 			break;
 
···
 	op.u.firmware_info.type = XEN_FW_DISK_MBR_SIGNATURE;
 	for (nr = 0; nr < EDD_MBR_SIG_MAX; nr++) {
 		op.u.firmware_info.index = nr;
-		ret = HYPERVISOR_dom0_op(&op);
+		ret = HYPERVISOR_platform_op(&op);
 		if (ret)
 			break;
 		mbr_signature[nr] = op.u.firmware_info.u.disk_mbr_signature.mbr_signature;
···
 		xen_start_info->console.domU.mfn = 0;
 		xen_start_info->console.domU.evtchn = 0;
 
-		if (HYPERVISOR_dom0_op(&op) == 0)
+		if (HYPERVISOR_platform_op(&op) == 0)
 			boot_params.kbd_status = op.u.firmware_info.u.kbd_shift_flags;
 
 		/* Make sure ACS will be enabled */

arch/x86/xen/suspend.c

···
 {
 #ifdef CONFIG_XEN_PVHVM
 	int cpu;
-	xen_hvm_init_shared_info();
+	if (!suspend_cancelled)
+	    xen_hvm_init_shared_info();
 	xen_callback_vector();
 	xen_unplug_emulated_devices();
 	if (xen_feature(XENFEAT_hvm_safe_pvclock)) {

arch/x86/xen/time.c

···
 #include <linux/gfp.h>
 #include <linux/slab.h>
 #include <linux/pvclock_gtod.h>
+#include <linux/timekeeper_internal.h>
 
 #include <asm/pvclock.h>
 #include <asm/xen/hypervisor.h>
···
 #define TIMER_SLOP	100000
 #define NS_PER_TICK	(1000000000LL / HZ)
 
-/* runstate info updated by Xen */
-static DEFINE_PER_CPU(struct vcpu_runstate_info, xen_runstate);
-
 /* snapshots of runstate info */
 static DEFINE_PER_CPU(struct vcpu_runstate_info, xen_runstate_snapshot);
 
 /* unused ns of stolen time */
 static DEFINE_PER_CPU(u64, xen_residual_stolen);
-
-/* return an consistent snapshot of 64-bit time/counter value */
-static u64 get64(const u64 *p)
-{
-	u64 ret;
-
-	if (BITS_PER_LONG < 64) {
-		u32 *p32 = (u32 *)p;
-		u32 h, l;
-
-		/*
-		 * Read high then low, and then make sure high is
-		 * still the same; this will only loop if low wraps
-		 * and carries into high.
-		 * XXX some clean way to make this endian-proof?
-		 */
-		do {
-			h = p32[1];
-			barrier();
-			l = p32[0];
-			barrier();
-		} while (p32[1] != h);
-
-		ret = (((u64)h) << 32) | l;
-	} else
-		ret = *p;
-
-	return ret;
-}
-
-/*
- * Runstate accounting
- */
-static void get_runstate_snapshot(struct vcpu_runstate_info *res)
-{
-	u64 state_time;
-	struct vcpu_runstate_info *state;
-
-	BUG_ON(preemptible());
-
-	state = this_cpu_ptr(&xen_runstate);
-
-	/*
-	 * The runstate info is always updated by the hypervisor on
-	 * the current CPU, so there's no need to use anything
-	 * stronger than a compiler barrier when fetching it.
-	 */
-	do {
-		state_time = get64(&state->state_entry_time);
-		barrier();
-		*res = *state;
-		barrier();
-	} while (get64(&state->state_entry_time) != state_time);
-}
-
-/* return true when a vcpu could run but has no real cpu to run on */
-bool xen_vcpu_stolen(int vcpu)
-{
-	return per_cpu(xen_runstate, vcpu).state == RUNSTATE_runnable;
-}
-
-void xen_setup_runstate_info(int cpu)
-{
-	struct vcpu_register_runstate_memory_area area;
-
-	area.addr.v = &per_cpu(xen_runstate, cpu);
-
-	if (HYPERVISOR_vcpu_op(VCPUOP_register_runstate_memory_area,
-			       cpu, &area))
-		BUG();
-}
 
 static void do_stolen_accounting(void)
 {
···
 	s64 runnable, offline, stolen;
 	cputime_t ticks;
 
-	get_runstate_snapshot(&state);
+	xen_get_runstate_snapshot(&state);
 
 	WARN_ON(state.state != RUNSTATE_running);
 
···
 				   unsigned long was_set, void *priv)
 {
 	/* Protected by the calling core code serialization */
-	static struct timespec next_sync;
+	static struct timespec64 next_sync;
 
 	struct xen_platform_op op;
-	struct timespec now;
+	struct timespec64 now;
+	struct timekeeper *tk = priv;
+	static bool settime64_supported = true;
+	int ret;
 
-	now = __current_kernel_time();
+	now.tv_sec = tk->xtime_sec;
+	now.tv_nsec = (long)(tk->tkr_mono.xtime_nsec >> tk->tkr_mono.shift);
 
 	/*
 	 * We only take the expensive HV call when the clock was set
 	 * or when the 11 minutes RTC synchronization time elapsed.
 	 */
-	if (!was_set && timespec_compare(&now, &next_sync) < 0)
+	if (!was_set && timespec64_compare(&now, &next_sync) < 0)
 		return NOTIFY_OK;
 
-	op.cmd = XENPF_settime;
-	op.u.settime.secs = now.tv_sec;
-	op.u.settime.nsecs = now.tv_nsec;
-	op.u.settime.system_time = xen_clocksource_read();
+again:
+	if (settime64_supported) {
+		op.cmd = XENPF_settime64;
+		op.u.settime64.mbz = 0;
+		op.u.settime64.secs = now.tv_sec;
+		op.u.settime64.nsecs = now.tv_nsec;
+		op.u.settime64.system_time = xen_clocksource_read();
+	} else {
+		op.cmd = XENPF_settime32;
+		op.u.settime32.secs = now.tv_sec;
+		op.u.settime32.nsecs = now.tv_nsec;
+		op.u.settime32.system_time = xen_clocksource_read();
+	}
 
-	(void)HYPERVISOR_dom0_op(&op);
+	ret = HYPERVISOR_platform_op(&op);
+
+	if (ret == -ENOSYS && settime64_supported) {
+		settime64_supported = false;
+		goto again;
+	}
+	if (ret < 0)
+		return NOTIFY_BAD;
 
 	/*
 	 * Move the next drift compensation time 11 minutes

drivers/xen/Makefile

···
 obj-$(CONFIG_HOTPLUG_CPU)		+= cpu_hotplug.o
 obj-$(CONFIG_X86)			+= fallback.o
-obj-y	+= grant-table.o features.o balloon.o manage.o preempt.o
+obj-y	+= grant-table.o features.o balloon.o manage.o preempt.o time.o
 obj-y	+= events/
 obj-y	+= xenbus/
 

drivers/xen/acpi.c

···
 		 bits, val_a, val_b))
 		return -1;
 
-	HYPERVISOR_dom0_op(&op);
+	HYPERVISOR_platform_op(&op);
 	return 1;
 }
 

drivers/xen/efi.c

···
 {
 	struct xen_platform_op op = INIT_EFI_OP(get_time);
 
-	if (HYPERVISOR_dom0_op(&op) < 0)
+	if (HYPERVISOR_platform_op(&op) < 0)
 		return EFI_UNSUPPORTED;
 
 	if (tm) {
···
 	BUILD_BUG_ON(sizeof(*tm) != sizeof(efi_data(op).u.set_time));
 	memcpy(&efi_data(op).u.set_time, tm, sizeof(*tm));
 
-	if (HYPERVISOR_dom0_op(&op) < 0)
+	if (HYPERVISOR_platform_op(&op) < 0)
 		return EFI_UNSUPPORTED;
 
 	return efi_data(op).status;
···
 {
 	struct xen_platform_op op = INIT_EFI_OP(get_wakeup_time);
 
-	if (HYPERVISOR_dom0_op(&op) < 0)
+	if (HYPERVISOR_platform_op(&op) < 0)
 		return EFI_UNSUPPORTED;
 
 	if (tm) {
···
 	else
 		efi_data(op).misc |= XEN_EFI_SET_WAKEUP_TIME_ENABLE_ONLY;
 
-	if (HYPERVISOR_dom0_op(&op) < 0)
+	if (HYPERVISOR_platform_op(&op) < 0)
 		return EFI_UNSUPPORTED;
 
 	return efi_data(op).status;
···
 	efi_data(op).u.get_variable.size = *data_size;
 	set_xen_guest_handle(efi_data(op).u.get_variable.data, data);
 
-	if (HYPERVISOR_dom0_op(&op) < 0)
+	if (HYPERVISOR_platform_op(&op) < 0)
 		return EFI_UNSUPPORTED;
 
 	*data_size = efi_data(op).u.get_variable.size;
···
 	memcpy(&efi_data(op).u.get_next_variable_name.vendor_guid, vendor,
 	       sizeof(*vendor));
 
-	if (HYPERVISOR_dom0_op(&op) < 0)
+	if (HYPERVISOR_platform_op(&op) < 0)
 		return EFI_UNSUPPORTED;
 
 	*name_size = efi_data(op).u.get_next_variable_name.size;
···
 	efi_data(op).u.set_variable.size = data_size;
 	set_xen_guest_handle(efi_data(op).u.set_variable.data, data);
 
-	if (HYPERVISOR_dom0_op(&op) < 0)
+	if (HYPERVISOR_platform_op(&op) < 0)
 		return EFI_UNSUPPORTED;
 
 	return efi_data(op).status;
···
 
 	efi_data(op).u.query_variable_info.attr = attr;
 
-	if (HYPERVISOR_dom0_op(&op) < 0)
+	if (HYPERVISOR_platform_op(&op) < 0)
 		return EFI_UNSUPPORTED;
 
 	*storage_space = efi_data(op).u.query_variable_info.max_store_size;
···
 {
 	struct xen_platform_op op = INIT_EFI_OP(get_next_high_monotonic_count);
 
-	if (HYPERVISOR_dom0_op(&op) < 0)
+	if (HYPERVISOR_platform_op(&op) < 0)
 		return EFI_UNSUPPORTED;
 
 	*count = efi_data(op).misc;
···
 	efi_data(op).u.update_capsule.capsule_count = count;
 	efi_data(op).u.update_capsule.sg_list = sg_list;
 
-	if (HYPERVISOR_dom0_op(&op) < 0)
+	if (HYPERVISOR_platform_op(&op) < 0)
 		return EFI_UNSUPPORTED;
 
 	return efi_data(op).status;
···
 					capsules);
 	efi_data(op).u.query_capsule_capabilities.capsule_count = count;
 
-	if (HYPERVISOR_dom0_op(&op) < 0)
+	if (HYPERVISOR_platform_op(&op) < 0)
 		return EFI_UNSUPPORTED;
 
 	*max_size = efi_data(op).u.query_capsule_capabilities.max_capsule_size;
···
 	};
 	union xenpf_efi_info *info = &op.u.firmware_info.u.efi_info;
 
-	if (!xen_initial_domain() || HYPERVISOR_dom0_op(&op) < 0)
+	if (!xen_initial_domain() || HYPERVISOR_platform_op(&op) < 0)
 		return NULL;
 
 	/* Here we know that Xen runs on EFI platform. */
···
 	info->vendor.bufsz = sizeof(vendor);
 	set_xen_guest_handle(info->vendor.name, vendor);
 
-	if (HYPERVISOR_dom0_op(&op) == 0) {
+	if (HYPERVISOR_platform_op(&op) == 0) {
 		efi_systab_xen.fw_vendor = __pa_symbol(vendor);
 		efi_systab_xen.fw_revision = info->vendor.revision;
 	} else
···
 	op.u.firmware_info.type = XEN_FW_EFI_INFO;
 	op.u.firmware_info.index = XEN_FW_EFI_VERSION;
 
-	if (HYPERVISOR_dom0_op(&op) == 0)
+	if (HYPERVISOR_platform_op(&op) == 0)
 		efi_systab_xen.hdr.revision = info->version;
 
 	op.cmd = XENPF_firmware_info;
 	op.u.firmware_info.type = XEN_FW_EFI_INFO;
 	op.u.firmware_info.index = XEN_FW_EFI_RT_VERSION;
 
-	if (HYPERVISOR_dom0_op(&op) == 0)
+	if (HYPERVISOR_platform_op(&op) == 0)
 		efi.runtime_version = info->version;
 
 	return &efi_systab_xen;

drivers/xen/gntdev.c

···
 	mutex_unlock(&priv->lock);
 }
 
-static struct mmu_notifier_ops gntdev_mmu_ops = {
+static const struct mmu_notifier_ops gntdev_mmu_ops = {
 	.release                = mn_release,
 	.invalidate_page        = mn_invl_page,
 	.invalidate_range_start = mn_invl_range_start,
···
 	return rc;
 }
 
+#define GNTDEV_COPY_BATCH 24
+
+struct gntdev_copy_batch {
+	struct gnttab_copy ops[GNTDEV_COPY_BATCH];
+	struct page *pages[GNTDEV_COPY_BATCH];
+	s16 __user *status[GNTDEV_COPY_BATCH];
+	unsigned int nr_ops;
+	unsigned int nr_pages;
+};
+
+static int gntdev_get_page(struct gntdev_copy_batch *batch, void __user *virt,
+			   bool writeable, unsigned long *gfn)
+{
+	unsigned long addr = (unsigned long)virt;
+	struct page *page;
+	unsigned long xen_pfn;
+	int ret;
+
+	ret = get_user_pages_fast(addr, 1, writeable, &page);
+	if (ret < 0)
+		return ret;
+
+	batch->pages[batch->nr_pages++] = page;
+
+	xen_pfn = page_to_xen_pfn(page) + XEN_PFN_DOWN(addr & ~PAGE_MASK);
+	*gfn = pfn_to_gfn(xen_pfn);
+
+	return 0;
+}
+
+static void gntdev_put_pages(struct gntdev_copy_batch *batch)
+{
+	unsigned int i;
+
+	for (i = 0; i < batch->nr_pages; i++)
+		put_page(batch->pages[i]);
+	batch->nr_pages = 0;
+}
+
+static int gntdev_copy(struct gntdev_copy_batch *batch)
+{
+	unsigned int i;
+
+	gnttab_batch_copy(batch->ops, batch->nr_ops);
+	gntdev_put_pages(batch);
+
+	/*
+	 * For each completed op, update the status if the op failed
+	 * and all previous ops for the segment were successful.
+	 */
+	for (i = 0; i < batch->nr_ops; i++) {
+		s16 status = batch->ops[i].status;
+		s16 old_status;
+
+		if (status == GNTST_okay)
+			continue;
+
+		if (__get_user(old_status, batch->status[i]))
+			return -EFAULT;
+
+		if (old_status != GNTST_okay)
+			continue;
+
+		if (__put_user(status, batch->status[i]))
+			return -EFAULT;
+	}
+
+	batch->nr_ops = 0;
+	return 0;
+}
+
+static int gntdev_grant_copy_seg(struct gntdev_copy_batch *batch,
+				 struct gntdev_grant_copy_segment *seg,
+				 s16 __user *status)
+{
+	uint16_t copied = 0;
+
+	/*
+	 * Disallow local -> local copies since there is only space in
+	 * batch->pages for one page per-op and this would be a very
+	 * expensive memcpy().
+	 */
+	if (!(seg->flags & (GNTCOPY_source_gref | GNTCOPY_dest_gref)))
+		return -EINVAL;
+
+	/* Can't cross page if source/dest is a grant ref. */
+	if (seg->flags & GNTCOPY_source_gref) {
+		if (seg->source.foreign.offset + seg->len > XEN_PAGE_SIZE)
+			return -EINVAL;
+	}
+	if (seg->flags & GNTCOPY_dest_gref) {
+		if (seg->dest.foreign.offset + seg->len > XEN_PAGE_SIZE)
+			return -EINVAL;
+	}
+
+	if (put_user(GNTST_okay, status))
+		return -EFAULT;
+
+	while (copied < seg->len) {
+		struct gnttab_copy *op;
+		void __user *virt;
+		size_t len, off;
+		unsigned long gfn;
+		int ret;
+
+		if (batch->nr_ops >= GNTDEV_COPY_BATCH) {
+			ret = gntdev_copy(batch);
+			if (ret < 0)
+				return ret;
+		}
+
+		len = seg->len - copied;
+
+		op = &batch->ops[batch->nr_ops];
+		op->flags = 0;
+
+		if (seg->flags & GNTCOPY_source_gref) {
+			op->source.u.ref = seg->source.foreign.ref;
+			op->source.domid = seg->source.foreign.domid;
+			op->source.offset = seg->source.foreign.offset + copied;
+			op->flags |= GNTCOPY_source_gref;
+		} else {
+			virt = seg->source.virt + copied;
+			off = (unsigned long)virt & ~XEN_PAGE_MASK;
+			len = min(len, (size_t)XEN_PAGE_SIZE - off);
+
+			ret = gntdev_get_page(batch, virt, false, &gfn);
+			if (ret < 0)
+				return ret;
+
+			op->source.u.gmfn = gfn;
+			op->source.domid = DOMID_SELF;
+			op->source.offset = off;
+		}
+
+		if (seg->flags & GNTCOPY_dest_gref) {
+			op->dest.u.ref = seg->dest.foreign.ref;
+			op->dest.domid = seg->dest.foreign.domid;
+			op->dest.offset = seg->dest.foreign.offset + copied;
+			op->flags |= GNTCOPY_dest_gref;
+		} else {
+			virt = seg->dest.virt + copied;
+			off = (unsigned long)virt & ~XEN_PAGE_MASK;
+			len = min(len, (size_t)XEN_PAGE_SIZE - off);
+
+			ret = gntdev_get_page(batch, virt, true, &gfn);
+			if (ret < 0)
+				return ret;
+
+			op->dest.u.gmfn = gfn;
+			op->dest.domid = DOMID_SELF;
+			op->dest.offset = off;
+		}
+
+		op->len = len;
+		copied += len;
+
+		batch->status[batch->nr_ops] = status;
+		batch->nr_ops++;
+	}
+
+	return 0;
+}
+
+static long gntdev_ioctl_grant_copy(struct gntdev_priv *priv, void __user *u)
+{
+	struct ioctl_gntdev_grant_copy copy;
+	struct gntdev_copy_batch batch;
+	unsigned int i;
+	int ret = 0;
+
+	if (copy_from_user(&copy, u, sizeof(copy)))
+		return -EFAULT;
+
+	batch.nr_ops = 0;
+	batch.nr_pages = 0;
+
+	for (i = 0; i < copy.count; i++) {
+		struct gntdev_grant_copy_segment seg;
+
+		if (copy_from_user(&seg, &copy.segments[i], sizeof(seg))) {
+			ret = -EFAULT;
+			goto out;
+		}
+
+		ret = gntdev_grant_copy_seg(&batch, &seg, &copy.segments[i].status);
+		if (ret < 0)
+			goto out;
+
+		cond_resched();
+	}
+	if (batch.nr_ops)
+		ret = gntdev_copy(&batch);
+	return ret;
+
+  out:
+	gntdev_put_pages(&batch);
+	return ret;
+}
+
 static long gntdev_ioctl(struct file *flip,
 			 unsigned int cmd, unsigned long arg)
 {
···
 
 	case IOCTL_GNTDEV_SET_UNMAP_NOTIFY:
 		return gntdev_ioctl_notify(priv, ptr);
+
+	case IOCTL_GNTDEV_GRANT_COPY:
+		return gntdev_ioctl_grant_copy(priv, ptr);
 
 	default:
 		pr_debug("priv %p, unknown cmd %x\n", priv, cmd);

drivers/xen/grant-table.c

···
 	int result;
 };
 
-static struct gnttab_ops *gnttab_interface;
+static const struct gnttab_ops *gnttab_interface;
 
 static int grant_table_version;
 static int grefs_per_grant_frame;
···
 	return rc;
 }
 
-static struct gnttab_ops gnttab_v1_ops = {
+static const struct gnttab_ops gnttab_v1_ops = {
 	.map_frames			= gnttab_map_frames_v1,
 	.unmap_frames			= gnttab_unmap_frames_v1,
 	.update_entry			= gnttab_update_entry_v1,

drivers/xen/pcpu.c

···
 		.u.cpu_ol.cpuid		= cpu_id,
 	};
 
-	return HYPERVISOR_dom0_op(&op);
+	return HYPERVISOR_platform_op(&op);
 }
 
 static int xen_pcpu_up(uint32_t cpu_id)
···
 		.u.cpu_ol.cpuid		= cpu_id,
 	};
 
-	return HYPERVISOR_dom0_op(&op);
+	return HYPERVISOR_platform_op(&op);
 }
 
 static ssize_t show_online(struct device *dev,
···
 		.u.pcpu_info.xen_cpuid = cpu,
 	};
 
-	ret = HYPERVISOR_dom0_op(&op);
+	ret = HYPERVISOR_platform_op(&op);
 	if (ret)
 		return ret;
 
···
 	op.cmd = XENPF_get_cpuinfo;
 	while (cpu_id <= max_id) {
 		op.u.pcpu_info.xen_cpuid = cpu_id;
-		if (HYPERVISOR_dom0_op(&op)) {
+		if (HYPERVISOR_platform_op(&op)) {
 			cpu_id++;
 			continue;
 		}

drivers/xen/time.c

···
+/*
+ * Xen stolen ticks accounting.
+ */
+#include <linux/kernel.h>
+#include <linux/kernel_stat.h>
+#include <linux/math64.h>
+#include <linux/gfp.h>
+
+#include <asm/xen/hypervisor.h>
+#include <asm/xen/hypercall.h>
+
+#include <xen/events.h>
+#include <xen/features.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/vcpu.h>
+#include <xen/xen-ops.h>
+
+/* runstate info updated by Xen */
+static DEFINE_PER_CPU(struct vcpu_runstate_info, xen_runstate);
+
+/* return an consistent snapshot of 64-bit time/counter value */
+static u64 get64(const u64 *p)
+{
+	u64 ret;
+
+	if (BITS_PER_LONG < 64) {
+		u32 *p32 = (u32 *)p;
+		u32 h, l, h2;
+
+		/*
+		 * Read high then low, and then make sure high is
+		 * still the same; this will only loop if low wraps
+		 * and carries into high.
+		 * XXX some clean way to make this endian-proof?
+		 */
+		do {
+			h = READ_ONCE(p32[1]);
+			l = READ_ONCE(p32[0]);
+			h2 = READ_ONCE(p32[1]);
+		} while(h2 != h);
+
+		ret = (((u64)h) << 32) | l;
+	} else
+		ret = READ_ONCE(*p);
+
+	return ret;
+}
+
+/*
+ * Runstate accounting
+ */
+void xen_get_runstate_snapshot(struct vcpu_runstate_info *res)
+{
+	u64 state_time;
+	struct vcpu_runstate_info *state;
+
+	BUG_ON(preemptible());
+
+	state = this_cpu_ptr(&xen_runstate);
+
+	/*
+	 * The runstate info is always updated by the hypervisor on
+	 * the current CPU, so there's no need to use anything
+	 * stronger than a compiler barrier when fetching it.
+	 */
+	do {
+		state_time = get64(&state->state_entry_time);
+		*res = READ_ONCE(*state);
+	} while (get64(&state->state_entry_time) != state_time);
+}
+
+/* return true when a vcpu could run but has no real cpu to run on */
+bool xen_vcpu_stolen(int vcpu)
+{
+	return per_cpu(xen_runstate, vcpu).state == RUNSTATE_runnable;
+}
+
+void xen_setup_runstate_info(int cpu)
+{
+	struct vcpu_register_runstate_memory_area area;
+
+	area.addr.v = &per_cpu(xen_runstate, cpu);
+
+	if (HYPERVISOR_vcpu_op(VCPUOP_register_runstate_memory_area,
+			       cpu, &area))
+		BUG();
+}
+

drivers/xen/xen-acpi-cpuhotplug.c

···
 	op.u.cpu_add.acpi_id = pr->acpi_id;
 	op.u.cpu_add.pxm = pxm;
 
-	cpu_id = HYPERVISOR_dom0_op(&op);
+	cpu_id = HYPERVISOR_platform_op(&op);
 	if (cpu_id < 0)
 		pr_err(PREFIX "Failed to hotadd CPU for acpi_id %d\n",
 				pr->acpi_id);

drivers/xen/xen-acpi-pad.c

···
 	op.u.core_parking.type = XEN_CORE_PARKING_SET;
 	op.u.core_parking.idle_nums = idle_nums;
 
-	return HYPERVISOR_dom0_op(&op);
+	return HYPERVISOR_platform_op(&op);
 }
 
 static int xen_acpi_pad_idle_cpus_num(void)
···
 	op.cmd = XENPF_core_parking;
 	op.u.core_parking.type = XEN_CORE_PARKING_GET;
 
-	return HYPERVISOR_dom0_op(&op)
+	return HYPERVISOR_platform_op(&op)
 	       ?: op.u.core_parking.idle_nums;
 }
 

drivers/xen/xen-acpi-processor.c

···
 	set_xen_guest_handle(op.u.set_pminfo.power.states, dst_cx_states);
 
 	if (!no_hypercall)
-		ret = HYPERVISOR_dom0_op(&op);
+		ret = HYPERVISOR_platform_op(&op);
 
 	if (!ret) {
 		pr_debug("ACPI CPU%u - C-states uploaded.\n", _pr->acpi_id);
···
 	}
 
 	if (!no_hypercall)
-		ret = HYPERVISOR_dom0_op(&op);
+		ret = HYPERVISOR_platform_op(&op);
 
 	if (!ret) {
 		struct acpi_processor_performance *perf;
···
 	info = &op.u.pcpu_info;
 	info->xen_cpuid = 0;
 
-	ret = HYPERVISOR_dom0_op(&op);
+	ret = HYPERVISOR_platform_op(&op);
 	if (ret)
 		return NR_CPUS;
 
···
 	last_cpu = op.u.pcpu_info.max_present;
 	for (i = 0; i <= last_cpu; i++) {
 		info->xen_cpuid = i;
-		ret = HYPERVISOR_dom0_op(&op);
+		ret = HYPERVISOR_platform_op(&op);
 		if (ret)
 			continue;
 		max_acpi_id = max(info->acpi_id, max_acpi_id);

drivers/xen/xenfs/xensyms.c

···
 
 	symnum = symdata->symnum;
 
-	ret = HYPERVISOR_dom0_op(&xs->op);
+	ret = HYPERVISOR_platform_op(&xs->op);
 	if (ret < 0)
 		return ret;
 
···
 		set_xen_guest_handle(symdata->name, xs->name);
 		symdata->symnum--; /* Rewind */
 
-		ret = HYPERVISOR_dom0_op(&xs->op);
+		ret = HYPERVISOR_platform_op(&xs->op);
 		if (ret < 0)
 			return ret;
 	}

include/uapi/xen/gntdev.h

···
 	__u32 event_channel_port;
 };
 
+struct gntdev_grant_copy_segment {
+	union {
+		void __user *virt;
+		struct {
+			grant_ref_t ref;
+			__u16 offset;
+			domid_t domid;
+		} foreign;
+	} source, dest;
+	__u16 len;
+
+	__u16 flags;  /* GNTCOPY_* */
+	__s16 status; /* GNTST_* */
+};
+
+/*
+ * Copy between grant references and local buffers.
+ *
+ * The copy is split into @count @segments, each of which can copy
+ * to/from one grant reference.
+ *
+ * Each segment is similar to struct gnttab_copy in the hypervisor ABI
+ * except the local buffer is specified using a virtual address
+ * (instead of a GFN and offset).
+ *
+ * The local buffer may cross a Xen page boundary -- the driver will
+ * split segments into multiple ops if required.
+ *
+ * Returns 0 if all segments have been processed and @status in each
+ * segment is valid.  Note that one or more segments may have failed
+ * (status != GNTST_okay).
+ *
+ * If the driver had to split a segment into two or more ops, @status
+ * includes the status of the first failed op for that segment (or
+ * GNTST_okay if all ops were successful).
+ *
+ * If -1 is returned, the status of all segments is undefined.
+ *
+ * EINVAL: A segment has local buffers for both source and
+ *         destination.
+ * EINVAL: A segment crosses the boundary of a foreign page.
+ * EFAULT: A segment's local buffer is not accessible.
+ */
+#define IOCTL_GNTDEV_GRANT_COPY \
+	_IOC(_IOC_NONE, 'G', 8, sizeof(struct ioctl_gntdev_grant_copy))
+struct ioctl_gntdev_grant_copy {
+	unsigned int count;
+	struct gntdev_grant_copy_segment __user *segments;
+};
+
 /* Clear (set to zero) the byte specified by index */
 #define UNMAP_NOTIFY_CLEAR_BYTE 0x1
 /* Send an interrupt on the indicated event channel */

include/xen/interface/platform.h

···
  * Set clock such that it would read <secs,nsecs> after 00:00:00 UTC,
  * 1 January, 1970 if the current system time was <system_time>.
  */
-#define XENPF_settime             17
-struct xenpf_settime {
+#define XENPF_settime32             17
+struct xenpf_settime32 {
 	/* IN variables. */
 	uint32_t secs;
 	uint32_t nsecs;
 	uint64_t system_time;
 };
-DEFINE_GUEST_HANDLE_STRUCT(xenpf_settime_t);
+DEFINE_GUEST_HANDLE_STRUCT(xenpf_settime32_t);
+#define XENPF_settime64           62
+struct xenpf_settime64 {
+    /* IN variables. */
+    uint64_t secs;
+    uint32_t nsecs;
+    uint32_t mbz;
+    uint64_t system_time;
+};
+DEFINE_GUEST_HANDLE_STRUCT(xenpf_settime64_t);
 
 /*
  * Request memory range (@mfn, @mfn+@nr_mfns-1) to have type @type.
···
 	uint32_t cmd;
 	uint32_t interface_version; /* XENPF_INTERFACE_VERSION */
 	union {
-		struct xenpf_settime           settime;
+		struct xenpf_settime32         settime32;
+		struct xenpf_settime64         settime64;
 		struct xenpf_add_memtype       add_memtype;
 		struct xenpf_del_memtype       del_memtype;
 		struct xenpf_read_memtype      read_memtype;

include/xen/interface/xen.h

···
 #define __HYPERVISOR_set_callbacks         4
 #define __HYPERVISOR_fpu_taskswitch        5
 #define __HYPERVISOR_sched_op_compat       6
-#define __HYPERVISOR_dom0_op               7
+#define __HYPERVISOR_platform_op           7
 #define __HYPERVISOR_set_debugreg          8
 #define __HYPERVISOR_get_debugreg          9
 #define __HYPERVISOR_update_descriptor    10

include/xen/xen-ops.h

···
 #include <linux/notifier.h>
 #include <linux/efi.h>
 #include <asm/xen/interface.h>
+#include <xen/interface/vcpu.h>
 
 DECLARE_PER_CPU(struct vcpu_info *, xen_vcpu);
 
···
 
 void xen_resume_notifier_register(struct notifier_block *nb);
 void xen_resume_notifier_unregister(struct notifier_block *nb);
+
+bool xen_vcpu_stolen(int vcpu);
+void xen_setup_runstate_info(int cpu);
+void xen_get_runstate_snapshot(struct vcpu_runstate_info *res);
 
 int xen_setup_shutdown_event(void);
 

kernel/sched/cputime.c

···
 #include <linux/static_key.h>
 #include <linux/context_tracking.h>
 #include "sched.h"
+#ifdef CONFIG_PARAVIRT
+#include <asm/paravirt.h>
+#endif
 
 
 #ifdef CONFIG_IRQ_TIME_ACCOUNTING