xen: add save/restore/migrate for Xen domains

@@ -1222,7 +1222,7 @@
 
 	/* Get mfn list */
 	if (!xen_feature(XENFEAT_auto_translated_physmap))
-		phys_to_machine_mapping = (unsigned long *)xen_start_info->mfn_list;
+		xen_build_dynamic_phys_to_machine();
 
 	pgd = (pgd_t *)xen_start_info->pt_base;
 

91 lines not on this page · raw message

@@ -27,8 +27,6 @@
 extern const char xen_hypervisor_callback[];
 extern const char xen_failsafe_callback[];
 
-unsigned long *phys_to_machine_mapping;
-EXPORT_SYMBOL(phys_to_machine_mapping);
 
 /**
  * machine_specific_memory_setup - Hook for machine specific memory setup.

@@ -21,6 +21,8 @@
 void __init xen_arch_setup(void);
 void __init xen_init_IRQ(void);
 void xen_enable_sysenter(void);
+
+void __init xen_build_dynamic_phys_to_machine(void);
 
 void xen_setup_timer(int cpu);
 void xen_setup_cpu_clockevents(void);

@@ -26,15 +26,15 @@
 #define FOREIGN_FRAME_BIT	(1UL<<31)
 #define FOREIGN_FRAME(m)	((m) | FOREIGN_FRAME_BIT)
 
-extern unsigned long *phys_to_machine_mapping;
+extern unsigned long get_phys_to_machine(unsigned long pfn);
+extern void set_phys_to_machine(unsigned long pfn, unsigned long mfn);
 
 static inline unsigned long pfn_to_mfn(unsigned long pfn)
 {
 	if (xen_feature(XENFEAT_auto_translated_physmap))
 		return pfn;
 
-	return phys_to_machine_mapping[(unsigned int)(pfn)] &
-		~FOREIGN_FRAME_BIT;
+	return get_phys_to_machine(pfn) & ~FOREIGN_FRAME_BIT;
 }
 
 static inline int phys_to_machine_mapping_valid(unsigned long pfn)

@@ -42,7 +42,7 @@
 	if (xen_feature(XENFEAT_auto_translated_physmap))
 		return 1;
 
-	return (phys_to_machine_mapping[pfn] != INVALID_P2M_ENTRY);
+	return get_phys_to_machine(pfn) != INVALID_P2M_ENTRY;
 }
 
 static inline unsigned long mfn_to_pfn(unsigned long mfn)

@@ -106,18 +106,10 @@
 	unsigned long pfn = mfn_to_pfn(mfn);
 	if ((pfn < max_mapnr)
 	    && !xen_feature(XENFEAT_auto_translated_physmap)
-	    && (phys_to_machine_mapping[pfn] != mfn))
+	    && (get_phys_to_machine(pfn) != mfn))
 		return max_mapnr; /* force !pfn_valid() */
+	/* XXX fixme; not true with sparsemem */
 	return pfn;
-}
-
-static inline void set_phys_to_machine(unsigned long pfn, unsigned long mfn)
-{
-	if (xen_feature(XENFEAT_auto_translated_physmap)) {
-		BUG_ON(pfn != mfn && mfn != INVALID_P2M_ENTRY);
-		return;
-	}
-	phys_to_machine_mapping[pfn] = mfn;
 }
 
 /* VIRT <-> MACHINE conversion */

@@ -56,19 +56,13 @@
 #include "multicalls.h"
 #include "mmu.h"
 
-/*
- * This should probably be a config option.  On 32-bit, it costs 1
- * page/gig of memory; on 64-bit its 2 pages/gig.  If we want it to be
- * completely unbounded we can add another level to the p2m structure.
- */
-#define MAX_GUEST_PAGES		(16ull * 1024*1024*1024 / PAGE_SIZE)
 #define P2M_ENTRIES_PER_PAGE	(PAGE_SIZE / sizeof(unsigned long))
 
-static unsigned long *p2m_top[MAX_GUEST_PAGES / P2M_ENTRIES_PER_PAGE];
+static unsigned long *p2m_top[MAX_DOMAIN_PAGES / P2M_ENTRIES_PER_PAGE];
 
 static inline unsigned p2m_top_index(unsigned long pfn)
 {
-	BUG_ON(pfn >= MAX_GUEST_PAGES);
+	BUG_ON(pfn >= MAX_DOMAIN_PAGES);
 	return pfn / P2M_ENTRIES_PER_PAGE;
 }
 

@@ -81,12 +75,9 @@
 {
 	unsigned pfn;
 	unsigned long *mfn_list = (unsigned long *)xen_start_info->mfn_list;
+	unsigned long max_pfn = min(MAX_DOMAIN_PAGES, xen_start_info->nr_pages);
 
-	BUG_ON(xen_start_info->nr_pages >= MAX_GUEST_PAGES);
-
-	for(pfn = 0;
-	    pfn < xen_start_info->nr_pages;
-	    pfn += P2M_ENTRIES_PER_PAGE) {
+	for(pfn = 0; pfn < max_pfn; pfn += P2M_ENTRIES_PER_PAGE) {
 		unsigned topidx = p2m_top_index(pfn);
 
 		p2m_top[topidx] = &mfn_list[pfn];

@@ -96,6 +87,9 @@
 unsigned long get_phys_to_machine(unsigned long pfn)
 {
 	unsigned topidx, idx;
+
+	if (unlikely(pfn >= MAX_DOMAIN_PAGES))
+		return INVALID_P2M_ENTRY;
 
 	topidx = p2m_top_index(pfn);
 	if (p2m_top[topidx] == NULL)

@@ -126,6 +120,11 @@
 
 	if (unlikely(xen_feature(XENFEAT_auto_translated_physmap))) {
 		BUG_ON(pfn != mfn && mfn != INVALID_P2M_ENTRY);
+		return;
+	}
+
+	if (unlikely(pfn >= MAX_DOMAIN_PAGES)) {
+		BUG_ON(mfn != INVALID_P2M_ENTRY);
 		return;
 	}
 

@@ -16,6 +16,7 @@
 #include <asm/xen/hypervisor.h>
 #include <asm/xen/hypercall.h>
 
+#include <xen/page.h>
 #include <xen/interface/callback.h>
 #include <xen/interface/physdev.h>
 #include <xen/features.h>

@@ -35,6 +36,8 @@
 char * __init xen_memory_setup(void)
 {
 	unsigned long max_pfn = xen_start_info->nr_pages;
+
+	max_pfn = min(MAX_DOMAIN_PAGES, max_pfn);
 
 	e820.nr_map = 0;
 	add_memory_region(0, LOWMEMSIZE(), E820_RAM);

@@ -25,6 +25,11 @@
 #define INVALID_P2M_ENTRY	(~0UL)
 #define FOREIGN_FRAME_BIT	(1UL<<31)
 #define FOREIGN_FRAME(m)	((m) | FOREIGN_FRAME_BIT)
+
+/* Maximum amount of memory we can handle in a domain in pages */
+#define MAX_DOMAIN_PAGES						\
+    ((unsigned long)((u64)CONFIG_XEN_MAX_DOMAIN_MEMORY * 1024 * 1024 * 1024 / PAGE_SIZE))
+
 
 extern unsigned long get_phys_to_machine(unsigned long pfn);
 extern void set_phys_to_machine(unsigned long pfn, unsigned long mfn);

@@ -57,8 +57,17 @@
 #include "mmu.h"
 
 #define P2M_ENTRIES_PER_PAGE	(PAGE_SIZE / sizeof(unsigned long))
+#define TOP_ENTRIES		(MAX_DOMAIN_PAGES / P2M_ENTRIES_PER_PAGE)
 
-static unsigned long *p2m_top[MAX_DOMAIN_PAGES / P2M_ENTRIES_PER_PAGE];
+/* Placeholder for holes in the address space */
+static unsigned long p2m_missing[P2M_ENTRIES_PER_PAGE]
+	__attribute__((section(".data.page_aligned"))) =
+		{ [ 0 ... P2M_ENTRIES_PER_PAGE-1 ] = ~0UL };
+
+ /* Array of pointers to pages containing p2m entries */
+static unsigned long *p2m_top[TOP_ENTRIES]
+	__attribute__((section(".data.page_aligned"))) =
+		{ [ 0 ... TOP_ENTRIES - 1] = &p2m_missing[0] };
 
 static inline unsigned p2m_top_index(unsigned long pfn)
 {

@@ -92,9 +101,6 @@
 		return INVALID_P2M_ENTRY;
 
 	topidx = p2m_top_index(pfn);
-	if (p2m_top[topidx] == NULL)
-		return INVALID_P2M_ENTRY;
-
 	idx = p2m_index(pfn);
 	return p2m_top[topidx][idx];
 }

@@ -110,7 +116,7 @@
 	for(i = 0; i < P2M_ENTRIES_PER_PAGE; i++)
 		p[i] = INVALID_P2M_ENTRY;
 
-	if (cmpxchg(pp, NULL, p) != NULL)
+	if (cmpxchg(pp, p2m_missing, p) != p2m_missing)
 		free_page((unsigned long)p);
 }
 

@@ -129,7 +135,7 @@
 	}
 
 	topidx = p2m_top_index(pfn);
-	if (p2m_top[topidx] == NULL) {
+	if (p2m_top[topidx] == p2m_missing) {
 		/* no need to allocate a page to store an invalid entry */
 		if (mfn == INVALID_P2M_ENTRY)
 			return;

@@ -75,13 +75,13 @@
 struct start_info *xen_start_info;
 EXPORT_SYMBOL_GPL(xen_start_info);
 
-static /* __initdata */ struct shared_info dummy_shared_info;
+struct shared_info xen_dummy_shared_info;
 
 /*
  * Point at some empty memory to start with. We map the real shared_info
  * page as soon as fixmap is up and running.
  */
-struct shared_info *HYPERVISOR_shared_info = (void *)&dummy_shared_info;
+struct shared_info *HYPERVISOR_shared_info = (void *)&xen_dummy_shared_info;
 
 /*
  * Flag to determine whether vcpu info placement is available on all

@@ -104,7 +104,7 @@
 	int err;
 	struct vcpu_info *vcpup;
 
-	BUG_ON(HYPERVISOR_shared_info == &dummy_shared_info);
+	BUG_ON(HYPERVISOR_shared_info == &xen_dummy_shared_info);
 	per_cpu(xen_vcpu, cpu) = &HYPERVISOR_shared_info->vcpu_info[cpu];
 
 	if (!have_vcpu_info_placement)

@@ -15,6 +15,7 @@
 DECLARE_PER_CPU(unsigned long, xen_current_cr3);
 
 extern struct start_info *xen_start_info;
+extern struct shared_info xen_dummy_shared_info;
 extern struct shared_info *HYPERVISOR_shared_info;
 
 char * __init xen_memory_setup(void);

@@ -880,6 +880,8 @@
 	/* In UP this is as good a place as any to set up shared info */
 	xen_setup_vcpu_info_placement();
 #endif
+
+	xen_setup_mfn_list_list();
 }
 
 static __init void xen_pagetable_setup_done(pgd_t *base)

@@ -69,6 +69,13 @@
 	__attribute__((section(".data.page_aligned"))) =
 		{ [ 0 ... TOP_ENTRIES - 1] = &p2m_missing[0] };
 
+/* Arrays of p2m arrays expressed in mfns used for save/restore */
+static unsigned long p2m_top_mfn[TOP_ENTRIES]
+	__attribute__((section(".bss.page_aligned")));
+
+static unsigned long p2m_top_mfn_list[TOP_ENTRIES / P2M_ENTRIES_PER_PAGE]
+	__attribute__((section(".bss.page_aligned")));
+
 static inline unsigned p2m_top_index(unsigned long pfn)
 {
 	BUG_ON(pfn >= MAX_DOMAIN_PAGES);

@@ -80,11 +87,35 @@
 	return pfn % P2M_ENTRIES_PER_PAGE;
 }
 
+/* Build the parallel p2m_top_mfn structures */
+void xen_setup_mfn_list_list(void)
+{
+	unsigned pfn, idx;
+
+	for(pfn = 0; pfn < MAX_DOMAIN_PAGES; pfn += P2M_ENTRIES_PER_PAGE) {
+		unsigned topidx = p2m_top_index(pfn);
+
+		p2m_top_mfn[topidx] = virt_to_mfn(p2m_top[topidx]);
+	}
+
+	for(idx = 0; idx < ARRAY_SIZE(p2m_top_mfn_list); idx++) {
+		unsigned topidx = idx * P2M_ENTRIES_PER_PAGE;
+		p2m_top_mfn_list[idx] = virt_to_mfn(&p2m_top_mfn[topidx]);
+	}
+
+	BUG_ON(HYPERVISOR_shared_info == &xen_dummy_shared_info);
+
+	HYPERVISOR_shared_info->arch.pfn_to_mfn_frame_list_list =
+		virt_to_mfn(p2m_top_mfn_list);
+	HYPERVISOR_shared_info->arch.max_pfn = xen_start_info->nr_pages;
+}
+
+/* Set up p2m_top to point to the domain-builder provided p2m pages */
 void __init xen_build_dynamic_phys_to_machine(void)
 {
-	unsigned pfn;
 	unsigned long *mfn_list = (unsigned long *)xen_start_info->mfn_list;
 	unsigned long max_pfn = min(MAX_DOMAIN_PAGES, xen_start_info->nr_pages);
+	unsigned pfn;
 
 	for(pfn = 0; pfn < max_pfn; pfn += P2M_ENTRIES_PER_PAGE) {
 		unsigned topidx = p2m_top_index(pfn);

@@ -105,7 +136,7 @@
 	return p2m_top[topidx][idx];
 }
 
-static void alloc_p2m(unsigned long **pp)
+static void alloc_p2m(unsigned long **pp, unsigned long *mfnp)
 {
 	unsigned long *p;
 	unsigned i;

@@ -118,6 +149,8 @@
 
 	if (cmpxchg(pp, p2m_missing, p) != p2m_missing)
 		free_page((unsigned long)p);
+	else
+		*mfnp = virt_to_mfn(p);
 }
 
 void set_phys_to_machine(unsigned long pfn, unsigned long mfn)

@@ -139,7 +172,7 @@
 		/* no need to allocate a page to store an invalid entry */
 		if (mfn == INVALID_P2M_ENTRY)
 			return;
-		alloc_p2m(&p2m_top[topidx]);
+		alloc_p2m(&p2m_top[topidx], &p2m_top_mfn[topidx]);
 	}
 
 	idx = p2m_index(pfn);

@@ -17,6 +17,8 @@
 extern struct start_info *xen_start_info;
 extern struct shared_info xen_dummy_shared_info;
 extern struct shared_info *HYPERVISOR_shared_info;
+
+void xen_setup_mfn_list_list(void);
 
 char * __init xen_memory_setup(void);
 void __init xen_arch_setup(void);

@@ -557,6 +557,33 @@
 	put_cpu();
 }
 
+/* Rebind a new event channel to an existing irq. */
+void rebind_evtchn_irq(int evtchn, int irq)
+{
+	/* Make sure the irq is masked, since the new event channel
+	   will also be masked. */
+	disable_irq(irq);
+
+	spin_lock(&irq_mapping_update_lock);
+
+	/* After resume the irq<->evtchn mappings are all cleared out */
+	BUG_ON(evtchn_to_irq[evtchn] != -1);
+	/* Expect irq to have been bound before,
+	   so the bindcount should be non-0 */
+	BUG_ON(irq_bindcount[irq] == 0);
+
+	evtchn_to_irq[evtchn] = irq;
+	irq_info[irq] = mk_irq_info(IRQT_EVTCHN, 0, evtchn);
+
+	spin_unlock(&irq_mapping_update_lock);
+
+	/* new event channels are always bound to cpu 0 */
+	irq_set_affinity(irq, cpumask_of_cpu(0));
+
+	/* Unmask the event channel. */
+	enable_irq(irq);
+}
+
 /* Rebind an evtchn so that it gets delivered to a specific cpu */
 static void rebind_irq_to_cpu(unsigned irq, unsigned tcpu)
 {

@@ -32,6 +32,7 @@
 
 void xen_send_IPI_one(unsigned int cpu, enum ipi_vector vector);
 int resend_irq_on_evtchn(unsigned int irq);
+void rebind_evtchn_irq(int evtchn, int irq);
 
 static inline void notify_remote_via_evtchn(int port)
 {

@@ -355,7 +355,7 @@
 
 	spin_lock(&irq_mapping_update_lock);
 
-	if (VALID_EVTCHN(evtchn) && (--irq_bindcount[irq] == 0)) {
+	if ((--irq_bindcount[irq] == 0) && VALID_EVTCHN(evtchn)) {
 		close.port = evtchn;
 		if (HYPERVISOR_event_channel_op(EVTCHNOP_close, &close) != 0)
 			BUG();

@@ -375,7 +375,7 @@
 		evtchn_to_irq[evtchn] = -1;
 		irq_info[irq] = IRQ_UNBOUND;
 
-		dynamic_irq_init(irq);
+		dynamic_irq_cleanup(irq);
 	}
 
 	spin_unlock(&irq_mapping_update_lock);

@@ -39,9 +39,14 @@
 
 /* ------------------------------------------------------------------ */
 
+static unsigned long console_pfn = ~0ul;
+
 static inline struct xencons_interface *xencons_interface(void)
 {
-	return mfn_to_virt(xen_start_info->console.domU.mfn);
+	if (console_pfn == ~0ul)
+		return mfn_to_virt(xen_start_info->console.domU.mfn);
+	else
+		return __va(console_pfn << PAGE_SHIFT);
 }
 
 static inline void notify_daemon(void)

@@ -101,18 +106,30 @@
 {
 	struct hvc_struct *hp;
 
-	if (!is_running_on_xen())
-		return 0;
+	if (!is_running_on_xen() ||
+	    is_initial_xendomain() ||
+	    !xen_start_info->console.domU.evtchn)
+		return -ENODEV;
 
 	xencons_irq = bind_evtchn_to_irq(xen_start_info->console.domU.evtchn);
 	if (xencons_irq < 0)
-		xencons_irq = 0 /* NO_IRQ */;
+		xencons_irq = 0; /* NO_IRQ */
+
 	hp = hvc_alloc(HVC_COOKIE, xencons_irq, &hvc_ops, 256);
 	if (IS_ERR(hp))
 		return PTR_ERR(hp);
 
 	hvc = hp;
+
+	console_pfn = mfn_to_pfn(xen_start_info->console.domU.mfn);
+
 	return 0;
+}
+
+void xen_console_resume(void)
+{
+	if (xencons_irq)
+		rebind_evtchn_irq(xen_start_info->console.domU.evtchn, xencons_irq);
 }
 
 static void __exit xen_fini(void)

@@ -3,6 +3,8 @@
 
 extern struct console xenboot_console;
 
+void xen_console_resume(void);
+
 void xen_raw_console_write(const char *str);
 void xen_raw_printk(const char *fmt, ...);
 

Index: linux/arch/x86/xen/mmu.c
===================================================================
@@ -136,6 +136,7 @@ unsigned long get_phys_to_machine(unsign
 	idx = p2m_index(pfn);
 	return p2m_top[topidx][idx];
 }
+EXPORT_SYMBOL_GPL(get_phys_to_machine);
 
 static void alloc_p2m(unsigned long **pp, unsigned long *mfnp)
 {

Index: linux/kernel/printk.c
===================================================================
@@ -76,6 +76,8 @@ EXPORT_SYMBOL(oops_in_progress);
 static DECLARE_MUTEX(console_sem);
 static DECLARE_MUTEX(secondary_console_sem);
 struct console *console_drivers;
+EXPORT_SYMBOL_GPL(console_drivers);
+
 /*
  * This is used for debugging the mess that is the VT code by
  * keeping track if we have the console semaphore held. It's

@@ -203,7 +203,6 @@
 int xb_init_comms(void)
 {
 	struct xenstore_domain_interface *intf = xen_store_interface;
-	int err;
 
 	if (intf->req_prod != intf->req_cons)
 		printk(KERN_ERR "XENBUS request ring is not quiescent "

@@ -216,18 +215,20 @@
 		intf->rsp_cons = intf->rsp_prod;
 	}
 
-	if (xenbus_irq)
-		unbind_from_irqhandler(xenbus_irq, &xb_waitq);
+	if (xenbus_irq) {
+		/* Already have an irq; assume we're resuming */
+		rebind_evtchn_irq(xen_store_evtchn, xenbus_irq);
+	} else {
+		int err;
+		err = bind_evtchn_to_irqhandler(xen_store_evtchn, wake_waiting,
+						0, "xenbus", &xb_waitq);
+		if (err <= 0) {
+			printk(KERN_ERR "XENBUS request irq failed %i\n", err);
+			return err;
+		}
 
-	err = bind_evtchn_to_irqhandler(
-		xen_store_evtchn, wake_waiting,
-		0, "xenbus", &xb_waitq);
-	if (err <= 0) {
-		printk(KERN_ERR "XENBUS request irq failed %i\n", err);
-		return err;
+		xenbus_irq = err;
 	}
-
-	xenbus_irq = err;
 
 	return 0;
 }

@@ -857,7 +857,7 @@
 			  PFN_DOWN(__pa(xen_start_info->pt_base)));
 }
 
-static __init void setup_shared_info(void)
+void xen_setup_shared_info(void)
 {
 	if (!xen_feature(XENFEAT_auto_translated_physmap)) {
 		unsigned long addr = fix_to_virt(FIX_PARAVIRT_BOOTMAP);

@@ -894,7 +894,7 @@
 	pv_mmu_ops.release_pmd = xen_release_pmd;
 	pv_mmu_ops.set_pte = xen_set_pte;
 
-	setup_shared_info();
+	xen_setup_shared_info();
 
 	/* Actually pin the pagetable down, but we can't set PG_pinned
 	   yet because the page structures don't exist yet. */

@@ -902,7 +902,7 @@
 }
 
 /* This is called once we have the cpu_possible_map */
-void __init xen_setup_vcpu_info_placement(void)
+void xen_setup_vcpu_info_placement(void)
 {
 	int cpu;
 

58 lines not on this page · raw message

@@ -35,7 +35,7 @@
 #include "xen-ops.h"
 #include "mmu.h"
 
-static cpumask_t xen_cpu_initialized_map;
+cpumask_t xen_cpu_initialized_map;
 static DEFINE_PER_CPU(int, resched_irq) = -1;
 static DEFINE_PER_CPU(int, callfunc_irq) = -1;
 static DEFINE_PER_CPU(int, debug_irq) = -1;

42 lines not on this page · raw message

@@ -565,6 +565,14 @@
 	clockevents_register_device(&__get_cpu_var(xen_clock_events));
 }
 
+void xen_time_suspend(void)
+{
+}
+
+void xen_time_resume(void)
+{
+}
+
 __init void xen_time_init(void)
 {
 	int cpu = smp_processor_id();

@@ -9,6 +9,7 @@
 extern const char xen_hypervisor_callback[];
 extern const char xen_failsafe_callback[];
 
+struct trap_info;
 void xen_copy_trap_info(struct trap_info *traps);
 
 DECLARE_PER_CPU(unsigned long, xen_cr3);

@@ -19,6 +20,7 @@
 extern struct shared_info *HYPERVISOR_shared_info;
 
 void xen_setup_mfn_list_list(void);
+void xen_setup_shared_info(void);
 
 char * __init xen_memory_setup(void);
 void __init xen_arch_setup(void);

@@ -59,6 +61,8 @@
 int xen_smp_call_function_mask(cpumask_t mask, void (*func)(void *),
 			       void *info, int wait);
 
+extern cpumask_t xen_cpu_initialized_map;
+
 
 /* Declare an asm function, along with symbols needed to make it
    inlineable */

89 lines not on this page · raw message

@@ -471,14 +471,14 @@
 	return 0;
 }
 
-static int gnttab_resume(void)
+int gnttab_resume(void)
 {
 	if (max_nr_grant_frames() < nr_grant_frames)
 		return -ENOSYS;
 	return gnttab_map(0, nr_grant_frames - 1);
 }
 
-static int gnttab_suspend(void)
+int gnttab_suspend(void)
 {
 	arch_gnttab_unmap_shared(shared, nr_grant_frames);
 	return 0;

143 lines not on this page · raw message

@@ -157,6 +157,7 @@
 __PAGEFLAG(Slab, slab)
 PAGEFLAG(Checked, owner_priv_1)		/* Used by some filesystems */
 PAGEFLAG(Pinned, owner_priv_1) TESTSCFLAG(Pinned, owner_priv_1) /* Xen */
+PAGEFLAG(SavePinned, dirty);					/* Xen */
 PAGEFLAG(Reserved, reserved) __CLEARPAGEFLAG(Reserved, reserved)
 PAGEFLAG(Private, private) __CLEARPAGEFLAG(Private, private)
 	__SETPAGEFLAG(Private, private)

@@ -41,4 +41,7 @@
 }
 
 extern void notify_remote_via_irq(int irq);
+
+extern void xen_irq_resume(void);
+
 #endif	/* _XEN_EVENTS_H */

@@ -50,6 +50,9 @@
 	void *arg;
 	u16 count;
 };
+
+int gnttab_suspend(void);
+int gnttab_resume(void);
 
 int gnttab_grant_foreign_access(domid_t domid, unsigned long frame,
 				int readonly);

@@ -5,4 +5,13 @@
 
 DECLARE_PER_CPU(struct vcpu_info *, xen_vcpu);
 
+void xen_pre_suspend(void);
+void xen_post_suspend(int suspend_cancelled);
+
+void xen_mm_pin_all(void);
+void xen_mm_unpin_all(void);
+
+void xen_time_suspend(void);
+void xen_time_resume(void);
+
 #endif /* INCLUDE_XEN_OPS_H */

Index: linux/include/xen/hvc-console.h
===================================================================
@@ -3,7 +3,11 @@
 
 extern struct console xenboot_console;
 
+#ifdef CONFIG_HVC_XEN
 void xen_console_resume(void);
+#else
+static inline void xen_console_resume(void) { }
+#endif
 
 void xen_raw_console_write(const char *str);
 void xen_raw_printk(const char *fmt, ...);

@@ -31,6 +31,7 @@

index 675c3dd..28bb618 100644
@@ -38,8 +38,6 @@ static int xen_suspend(void *data)
 
 	BUG_ON(!irqs_disabled());
 
-	load_cr3(swapper_pg_dir);
-
 	err = device_power_down(PMSG_SUSPEND);
 	if (err) {
 		printk(KERN_ERR "xen_suspend: device_power_down failed: %d\n",

@@ -565,14 +565,6 @@
 	clockevents_register_device(&__get_cpu_var(xen_clock_events));
 }
 
-void xen_time_suspend(void)
-{
-}
-
-void xen_time_resume(void)
-{
-}
-
 __init void xen_time_init(void)
 {
 	int cpu = smp_processor_id();

@@ -34,14 +34,21 @@
 static int xen_suspend(void *data)
 {
 	int *cancelled = data;
+	int err;
 
 	BUG_ON(!irqs_disabled());
 
 	load_cr3(swapper_pg_dir);
 
+	err = device_power_down(PMSG_SUSPEND);
+	if (err) {
+		printk(KERN_ERR "xen_suspend: device_power_down failed: %d\n",
+		       err);
+		return err;
+	}
+
 	xen_mm_pin_all();
 	gnttab_suspend();
-	xen_time_suspend();
 	xen_pre_suspend();
 
 	/*

@@ -52,9 +59,10 @@
 	*cancelled = HYPERVISOR_suspend(virt_to_mfn(xen_start_info));
 
 	xen_post_suspend(*cancelled);
-	xen_time_resume();
 	gnttab_resume();
 	xen_mm_unpin_all();
+
+	device_power_up();
 
 	if (!*cancelled) {
 		xen_irq_resume();

@@ -105,7 +113,8 @@
 
 	device_resume();
 
-
+	/* Make sure timer events get retriggered on all CPUs */
+	clock_was_set();
 out:
 #ifdef CONFIG_PREEMPT
 	thaw_processes();

@@ -11,7 +11,4 @@
 void xen_mm_pin_all(void);
 void xen_mm_unpin_all(void);
 
-void xen_time_suspend(void);
-void xen_time_resume(void);
-
 #endif /* INCLUDE_XEN_OPS_H */

@@ -632,8 +632,6 @@
  */
 void hres_timers_resume(void)
 {
-	WARN_ON_ONCE(num_online_cpus() > 1);
-
 	/* Retrigger the CPU local events: */
 	retrigger_next_event(NULL);
 }