xen/paravirt_ops patches for 2.6.24

@@ -369,8 +369,8 @@ void apply_paravirt(struct paravirt_patc
 		BUG_ON(p->len > MAX_PATCH_LEN);
 		/* prep the buffer with the original instructions */
 		memcpy(insnbuf, p->instr, p->len);
-		used = paravirt_ops.patch(p->instrtype, p->clobbers, insnbuf,
-					  (unsigned long)p->instr, p->len);
+		used = pv_init_ops.patch(p->instrtype, p->clobbers, insnbuf,
+					 (unsigned long)p->instr, p->len);
 
 		BUG_ON(used > p->len);
 

@@ -116,12 +116,14 @@ void foo(void)
 
 #ifdef CONFIG_PARAVIRT
 	BLANK();
-	OFFSET(PARAVIRT_enabled, paravirt_ops, paravirt_enabled);
-	OFFSET(PARAVIRT_irq_disable, paravirt_ops, irq_disable);
-	OFFSET(PARAVIRT_irq_enable, paravirt_ops, irq_enable);
-	OFFSET(PARAVIRT_irq_enable_sysexit, paravirt_ops, irq_enable_sysexit);
-	OFFSET(PARAVIRT_iret, paravirt_ops, iret);
-	OFFSET(PARAVIRT_read_cr0, paravirt_ops, read_cr0);
+	OFFSET(PARAVIRT_enabled, pv_info, paravirt_enabled);
+	OFFSET(PARAVIRT_PATCH_pv_cpu_ops, paravirt_patch_template, pv_cpu_ops);
+	OFFSET(PARAVIRT_PATCH_pv_irq_ops, paravirt_patch_template, pv_irq_ops);
+	OFFSET(PV_IRQ_irq_disable, pv_irq_ops, irq_disable);
+	OFFSET(PV_IRQ_irq_enable, pv_irq_ops, irq_enable);
+	OFFSET(PV_CPU_iret, pv_cpu_ops, iret);
+	OFFSET(PV_CPU_irq_enable_sysexit, pv_cpu_ops, irq_enable_sysexit);
+	OFFSET(PV_CPU_read_cr0, pv_cpu_ops, read_cr0);
 #endif
 
 #ifdef CONFIG_XEN

295 lines not on this page · raw message

267 lines not on this page · raw message

255 lines not on this page · raw message

@@ -115,7 +115,7 @@ static struct hv_ops lguest_cons = {
  * (0), and the struct hv_ops containing the put_chars() function. */
 static int __init cons_init(void)
 {
-	if (strcmp(paravirt_ops.name, "lguest") != 0)
+	if (strcmp(pv_info.name, "lguest") != 0)
 		return 0;
 
 	return hvc_instantiate(0, 0, &lguest_cons);

@@ -248,8 +248,8 @@ static void unmap_switcher(void)
 }
 
 /*H:130 Our Guest is usually so well behaved; it never tries to do things it
- * isn't allowed to.  Unfortunately, "struct paravirt_ops" isn't quite
- * complete, because it doesn't contain replacements for the Intel I/O
+ * isn't allowed to.  Unfortunately, Linux's paravirtual infrastructure isn't
+ * quite complete, because it doesn't contain replacements for the Intel I/O
  * instructions.  As a result, the Guest sometimes fumbles across one during
  * the boot process as it probes for various things which are usually attached
  * to a PC.

@@ -694,7 +694,7 @@ static int __init init(void)
 
 	/* Lguest can't run under Xen, VMI or itself.  It does Tricky Stuff. */
 	if (paravirt_enabled()) {
-		printk("lguest is afraid of %s\n", paravirt_ops.name);
+		printk("lguest is afraid of %s\n", pv_info.name);
 		return -EPERM;
 	}
 

177 lines not on this page · raw message

@@ -201,7 +201,7 @@ static void scan_devices(void)
  * "struct lguest_device_desc" array. */
 static int __init lguest_bus_init(void)
 {
-	if (strcmp(paravirt_ops.name, "lguest") != 0)
+	if (strcmp(pv_info.name, "lguest") != 0)
 		return 0;
 
 	/* Devices are in a single page above top of "normal" mem */

1,015 lines not on this page · raw message

@@ -2,7 +2,7 @@
 #define _I386_PGTABLE_3LEVEL_DEFS_H
 
 #ifdef CONFIG_PARAVIRT
-#define SHARED_KERNEL_PMD	(paravirt_ops.shared_kernel_pmd)
+#define SHARED_KERNEL_PMD	(pv_info.shared_kernel_pmd)
 #else
 #define SHARED_KERNEL_PMD	1
 #endif

@@ -164,7 +164,6 @@ static void *get_call_destination(u8 typ
 {
 	struct paravirt_patch_template tmpl = {
 		.pv_init_ops = pv_init_ops,
-		.pv_misc_ops = pv_misc_ops,
 		.pv_time_ops = pv_time_ops,
 		.pv_cpu_ops = pv_cpu_ops,
 		.pv_irq_ops = pv_irq_ops,

@@ -280,6 +279,49 @@ int paravirt_disable_iospace(void)
 	}
 
 	return ret;
+}
+
+static DEFINE_PER_CPU(enum paravirt_lazy_mode, paravirt_lazy_mode) = PARAVIRT_LAZY_NONE;
+
+static inline void enter_lazy(enum paravirt_lazy_mode mode)
+{
+	BUG_ON(x86_read_percpu(paravirt_lazy_mode) != PARAVIRT_LAZY_NONE);
+	BUG_ON(preemptible());
+
+	x86_write_percpu(paravirt_lazy_mode, mode);
+}
+
+void paravirt_leave_lazy(enum paravirt_lazy_mode mode)
+{
+	BUG_ON(x86_read_percpu(paravirt_lazy_mode) != mode);
+	BUG_ON(preemptible());
+
+	x86_write_percpu(paravirt_lazy_mode, PARAVIRT_LAZY_NONE);
+}
+
+void paravirt_enter_lazy_mmu(void)
+{
+	enter_lazy(PARAVIRT_LAZY_MMU);
+}
+
+void paravirt_leave_lazy_mmu(void)
+{
+	paravirt_leave_lazy(PARAVIRT_LAZY_MMU);
+}
+
+void paravirt_enter_lazy_cpu(void)
+{
+	enter_lazy(PARAVIRT_LAZY_CPU);
+}
+
+void paravirt_leave_lazy_cpu(void)
+{
+	paravirt_leave_lazy(PARAVIRT_LAZY_CPU);
+}
+
+enum paravirt_lazy_mode paravirt_get_lazy_mode(void)
+{
+	return x86_read_percpu(paravirt_lazy_mode);
 }
 
 struct pv_info pv_info = {

@@ -347,6 +389,11 @@ struct pv_cpu_ops pv_cpu_ops = {
 
 	.set_iopl_mask = native_set_iopl_mask,
 	.io_delay = native_io_delay,
+
+	.lazy_mode = {
+		.enter = paravirt_nop,
+		.leave = paravirt_nop,
+	},
 };
 
 struct pv_apic_ops pv_apic_ops = {

@@ -360,10 +407,6 @@ struct pv_apic_ops pv_apic_ops = {
 #endif
 };
 
-struct pv_misc_ops pv_misc_ops = {
-	.set_lazy_mode = paravirt_nop,
-};
-
 struct pv_mmu_ops pv_mmu_ops = {
 	.pagetable_setup_start = native_pagetable_setup_start,
 	.pagetable_setup_done = native_pagetable_setup_done,

@@ -414,6 +457,11 @@ struct pv_mmu_ops pv_mmu_ops = {
 	.dup_mmap = paravirt_nop,
 	.exit_mmap = paravirt_nop,
 	.activate_mm = paravirt_nop,
+
+	.lazy_mode = {
+		.enter = paravirt_nop,
+		.leave = paravirt_nop,
+	},
 };
 
 EXPORT_SYMBOL_GPL(pv_time_ops);

57 lines not on this page · raw message

81 lines not on this page · raw message

@@ -154,7 +154,7 @@ void xen_set_pte_at(struct mm_struct *mm
 		    pte_t *ptep, pte_t pteval)
 {
 	if (mm == current->mm || mm == &init_mm) {
-		if (xen_get_lazy_mode() == PARAVIRT_LAZY_MMU) {
+		if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU) {
 			struct multicall_space mcs;
 			mcs = xen_mc_entry(0);
 

@@ -35,7 +35,7 @@ void xen_mc_flush(void);
 /* Issue a multicall if we're not in a lazy mode */
 static inline void xen_mc_issue(unsigned mode)
 {
-	if ((xen_get_lazy_mode() & mode) == 0)
+	if ((paravirt_get_lazy_mode() & mode) == 0)
 		xen_mc_flush();
 
 	/* restore flags saved in xen_mc_batch */

@@ -32,13 +32,6 @@ bool xen_vcpu_stolen(int vcpu);
 
 void xen_mark_init_mm_pinned(void);
 
-DECLARE_PER_CPU(enum paravirt_lazy_mode, xen_lazy_mode);
-
-static inline unsigned xen_get_lazy_mode(void)
-{
-	return x86_read_percpu(xen_lazy_mode);
-}
-
 void __init xen_fill_possible_map(void);
 
 void __init xen_setup_vcpu_info_placement(void);

@@ -97,29 +97,17 @@ static cycle_t clock_base;
  * them as a batch when lazy_mode is eventually turned off.  Because hypercalls
  * are reasonably expensive, batching them up makes sense.  For example, a
  * large mmap might update dozens of page table entries: that code calls
- * lguest_lazy_mode(PARAVIRT_LAZY_MMU), does the dozen updates, then calls
- * lguest_lazy_mode(PARAVIRT_LAZY_NONE).
+ * paravirt_enter_lazy_mmu(), does the dozen updates, then calls
+ * lguest_leave_lazy_mode().
  *
  * So, when we're in lazy mode, we call async_hypercall() to store the call for
  * future processing.  When lazy mode is turned off we issue a hypercall to
  * flush the stored calls.
- *
- * There's also a hack where "mode" is set to "PARAVIRT_LAZY_FLUSH" which
- * indicates we're to flush any outstanding calls immediately.  This is used
- * when an interrupt handler does a kmap_atomic(): the page table changes must
- * happen immediately even if we're in the middle of a batch.  Usually we're
- * not, though, so there's nothing to do. */
-static enum paravirt_lazy_mode lazy_mode; /* Note: not SMP-safe! */
-static void lguest_lazy_mode(enum paravirt_lazy_mode mode)
-{
-	if (mode == PARAVIRT_LAZY_FLUSH) {
-		if (unlikely(lazy_mode != PARAVIRT_LAZY_NONE))
-			hcall(LHCALL_FLUSH_ASYNC, 0, 0, 0);
-	} else {
-		lazy_mode = mode;
-		if (mode == PARAVIRT_LAZY_NONE)
-			hcall(LHCALL_FLUSH_ASYNC, 0, 0, 0);
-	}
+ */
+static void lguest_leave_lazy_mode(void)
+{
+	paravirt_leave_lazy(paravirt_get_lazy_mode());
+	hcall(LHCALL_FLUSH_ASYNC, 0, 0, 0);
 }
 
 static void lazy_hcall(unsigned long call,

@@ -127,7 +115,7 @@ static void lazy_hcall(unsigned long cal
 		       unsigned long arg2,
 		       unsigned long arg3)
 {
-	if (lazy_mode == PARAVIRT_LAZY_NONE)
+	if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_NONE)
 		hcall(call, arg1, arg2, arg3);
 	else
 		async_hcall(call, arg1, arg2, arg3);

@@ -1011,6 +999,8 @@ __init void lguest_init(void *boot)
 	pv_cpu_ops.write_gdt_entry = lguest_write_gdt_entry;
 	pv_cpu_ops.write_idt_entry = lguest_write_idt_entry;
 	pv_cpu_ops.wbinvd = lguest_wbinvd;
+	pv_cpu_ops.lazy_mode.enter = paravirt_enter_lazy_cpu;
+	pv_cpu_ops.lazy_mode.leave = lguest_leave_lazy_mode;
 
 	/* pagetable management */
 	pv_mmu_ops.write_cr3 = lguest_write_cr3;

@@ -1022,6 +1012,8 @@ __init void lguest_init(void *boot)
 	pv_mmu_ops.set_pmd = lguest_set_pmd;
 	pv_mmu_ops.read_cr2 = lguest_read_cr2;
 	pv_mmu_ops.read_cr3 = lguest_read_cr3;
+	pv_mmu_ops.lazy_mode.enter = paravirt_enter_lazy_mmu;
+	pv_mmu_ops.lazy_mode.leave = lguest_leave_lazy_mode;
 
 #ifdef CONFIG_X86_LOCAL_APIC
 	/* apic read/write intercepts */

@@ -1033,8 +1025,6 @@ __init void lguest_init(void *boot)
 	/* time operations */
 	pv_time_ops.get_wallclock = lguest_get_wallclock;
 	pv_time_ops.time_init = lguest_time_init;
-
-	pv_misc_ops.set_lazy_mode = lguest_lazy_mode;
 
 	/* Now is a good time to look at the implementations of these functions
 	 * before returning to the rest of lguest_init(). */

120 lines not on this page · raw message

@@ -746,24 +746,12 @@ struct kmem_cache *pmd_cache;
 
 void __init pgtable_cache_init(void)
 {
-	size_t pgd_size = PTRS_PER_PGD*sizeof(pgd_t);
-
-	if (PTRS_PER_PMD > 1) {
+	if (PTRS_PER_PMD > 1)
 		pmd_cache = kmem_cache_create("pmd",
-					PTRS_PER_PMD*sizeof(pmd_t),
-					PTRS_PER_PMD*sizeof(pmd_t),
-					SLAB_PANIC,
-					pmd_ctor);
-		if (!SHARED_KERNEL_PMD) {
-			/* If we're in PAE mode and have a non-shared
-			   kernel pmd, then the pgd size must be a
-			   page size.  This is because the pgd_list
-			   links through the page structure, so there
-			   can only be one pgd per page for this to
-			   work. */
-			pgd_size = PAGE_SIZE;
-		}
-	}
+					      PTRS_PER_PMD*sizeof(pmd_t),
+					      PTRS_PER_PMD*sizeof(pmd_t),
+					      SLAB_PANIC,
+					      pmd_ctor);
 }
 
 /*

@@ -25,7 +25,6 @@
 #include <linux/mm.h>
 #include <linux/page-flags.h>
 #include <linux/highmem.h>
-#include <linux/smp.h>
 
 #include <xen/interface/xen.h>
 #include <xen/interface/physdev.h>

index 4ae038a..35691a2 100644

@@ -41,7 +41,6 @@
 #include <linux/sched.h>
 #include <linux/highmem.h>
 #include <linux/bug.h>
-#include <linux/sched.h>
 
 #include <asm/pgtable.h>
 #include <asm/tlbflush.h>

@@ -360,7 +360,8 @@ int xen_smp_call_function_mask(cpumask_t
 			       void *info, int wait)
 {
 	struct call_data_struct data;
-	int cpus;
+	int cpus, cpu;
+	bool yield;
 
 	/* Holding any lock stops cpus from going down. */
 	spin_lock(&call_lock);

@@ -389,9 +390,14 @@ int xen_smp_call_function_mask(cpumask_t
 	/* Send a message to other CPUs and wait for them to respond */
 	xen_send_IPI_mask(mask, XEN_CALL_FUNCTION_VECTOR);
 
-	/* Make sure other vcpus get a chance to run.
-	   XXX too severe?  Maybe we should check the other CPU's states? */
-	HYPERVISOR_sched_op(SCHEDOP_yield, 0);
+	/* Make sure other vcpus get a chance to run if they need to. */
+	yield = false;
+	for_each_cpu_mask(cpu, mask)
+		if (xen_vcpu_stolen(cpu))
+			yield = true;
+
+	if (yield)
+		HYPERVISOR_sched_op(SCHEDOP_yield, 0);
 
 	/* Wait for response */
 	while (atomic_read(&data.started) != cpus ||

@@ -103,6 +103,12 @@ static void get_runstate_snapshot(struct
 		*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(runstate, vcpu).state == RUNSTATE_runnable;
 }
 
 static void setup_runstate_info(int cpu)

@@ -27,6 +27,8 @@ int xen_set_wallclock(unsigned long time
 int xen_set_wallclock(unsigned long time);
 unsigned long long xen_sched_clock(void);
 
+bool xen_vcpu_stolen(int vcpu);
+
 void xen_mark_init_mm_pinned(void);
 
 DECLARE_PER_CPU(enum paravirt_lazy_mode, xen_lazy_mode);

@@ -32,7 +32,11 @@ struct mc_buffer {
 struct mc_buffer {
 	struct multicall_entry entries[MC_BATCH];
 	u64 args[MC_ARGS];
-	unsigned mcidx, argidx;
+	struct callback {
+		void (*fn)(void *);
+		void *data;
+	} callbacks[MC_BATCH];
+	unsigned mcidx, argidx, cbidx;
 };
 
 static DEFINE_PER_CPU(struct mc_buffer, mc_buffer);

@@ -43,6 +47,7 @@ void xen_mc_flush(void)
 	struct mc_buffer *b = &__get_cpu_var(mc_buffer);
 	int ret = 0;
 	unsigned long flags;
+	int i;
 
 	BUG_ON(preemptible());
 

@@ -51,8 +56,6 @@ void xen_mc_flush(void)
 	local_irq_save(flags);
 
 	if (b->mcidx) {
-		int i;
-
 		if (HYPERVISOR_multicall(b->entries, b->mcidx) != 0)
 			BUG();
 		for (i = 0; i < b->mcidx; i++)

@@ -64,6 +67,13 @@ void xen_mc_flush(void)
 		BUG_ON(b->argidx != 0);
 
 	local_irq_restore(flags);
+
+	for(i = 0; i < b->cbidx; i++) {
+		struct callback *cb = &b->callbacks[i];
+
+		(*cb->fn)(cb->data);
+	}
+	b->cbidx = 0;
 
 	BUG_ON(ret);
 }

@@ -88,3 +98,16 @@ struct multicall_space __xen_mc_entry(si
 
 	return ret;
 }
+
+void xen_mc_callback(void (*fn)(void *), void *data)
+{
+	struct mc_buffer *b = &__get_cpu_var(mc_buffer);
+	struct callback *cb;
+
+	if (b->cbidx == MC_BATCH)
+		xen_mc_flush();
+
+	cb = &b->callbacks[b->cbidx++];
+	cb->fn = fn;
+	cb->data = data;
+}

@@ -42,4 +42,7 @@ static inline void xen_mc_issue(unsigned
 	local_irq_restore(x86_read_percpu(xen_mc_irq_flags));
 }
 
+/* Set up a callback to be called when the current batch is flushed */
+void xen_mc_callback(void (*fn)(void *), void *data);
+
 #endif /* _XEN_MULTICALLS_H */

@@ -55,7 +55,23 @@ DEFINE_PER_CPU(enum paravirt_lazy_mode, 
 
 DEFINE_PER_CPU(struct vcpu_info *, xen_vcpu);
 DEFINE_PER_CPU(struct vcpu_info, xen_vcpu_info);
-DEFINE_PER_CPU(unsigned long, xen_cr3);
+
+/*
+ * Note about cr3 (pagetable base) values:
+ *
+ * xen_cr3 contains the current logical cr3 value; it contains the
+ * last set cr3.  This may not be the current effective cr3, because
+ * its update may be being lazily deferred.  However, a vcpu looking
+ * at its own cr3 can use this value knowing that it everything will
+ * be self-consistent.
+ *
+ * xen_current_cr3 contains the actual vcpu cr3; it is set once the
+ * hypercall to set the vcpu cr3 is complete (so it may be a little
+ * out of date, but it will never be set early).  If one vcpu is
+ * looking at another vcpu's cr3 value, it should use this variable.
+ */
+DEFINE_PER_CPU(unsigned long, xen_cr3);	 /* cr3 stored as physaddr */
+DEFINE_PER_CPU(unsigned long, xen_current_cr3);	 /* actual vcpu cr3 */
 
 struct start_info *xen_start_info;
 EXPORT_SYMBOL_GPL(xen_start_info);

@@ -631,32 +647,36 @@ static unsigned long xen_read_cr3(void)
 	return x86_read_percpu(xen_cr3);
 }
 
+static void set_current_cr3(void *v)
+{
+	x86_write_percpu(xen_current_cr3, (unsigned long)v);
+}
+
 static void xen_write_cr3(unsigned long cr3)
 {
+	struct mmuext_op *op;
+	struct multicall_space mcs;
+	unsigned long mfn = pfn_to_mfn(PFN_DOWN(cr3));
+
 	BUG_ON(preemptible());
 
-	if (cr3 == x86_read_percpu(xen_cr3)) {
-		/* just a simple tlb flush */
-		xen_flush_tlb();
-		return;
-	}
-
+	mcs = xen_mc_entry(sizeof(*op));  /* disables interrupts */
+
+	/* Update while interrupts are disabled, so its atomic with
+	   respect to ipis */
 	x86_write_percpu(xen_cr3, cr3);
 
-
-	{
-		struct mmuext_op *op;
-		struct multicall_space mcs = xen_mc_entry(sizeof(*op));
-		unsigned long mfn = pfn_to_mfn(PFN_DOWN(cr3));
-
-		op = mcs.args;
-		op->cmd = MMUEXT_NEW_BASEPTR;
-		op->arg1.mfn = mfn;
-
-		MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);
-
-		xen_mc_issue(PARAVIRT_LAZY_CPU);
-	}
+	op = mcs.args;
+	op->cmd = MMUEXT_NEW_BASEPTR;
+	op->arg1.mfn = mfn;
+
+	MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);
+
+	/* Update xen_update_cr3 once the batch has actually
+	   been submitted. */
+	xen_mc_callback(set_current_cr3, (void *)cr3);
+
+	xen_mc_issue(PARAVIRT_LAZY_CPU);  /* interrupts restored */
 }
 
 /* Early in boot, while setting up the initial pagetable, assume

@@ -1124,6 +1144,7 @@ asmlinkage void __init xen_start_kernel(
 	/* keep using Xen gdt for now; no urgent need to change it */
 
 	x86_write_percpu(xen_cr3, __pa(pgd));
+	x86_write_percpu(xen_current_cr3, __pa(pgd));
 
 #ifdef CONFIG_SMP
 	/* Don't do the full vcpu_info placement stuff until we have a

@@ -564,20 +564,43 @@ static void drop_other_mm_ref(void *info
 
 	if (__get_cpu_var(cpu_tlbstate).active_mm == mm)
 		leave_mm(smp_processor_id());
+
+	/* If this cpu still has a stale cr3 reference, then make sure
+	   it has been flushed. */
+	if (x86_read_percpu(xen_current_cr3) == __pa(mm->pgd)) {
+		load_cr3(swapper_pg_dir);
+		arch_flush_lazy_cpu_mode();
+	}
 }
 
 static void drop_mm_ref(struct mm_struct *mm)
 {
+	cpumask_t mask;
+	unsigned cpu;
+
 	if (current->active_mm == mm) {
 		if (current->mm == mm)
 			load_cr3(swapper_pg_dir);
 		else
 			leave_mm(smp_processor_id());
-	}
-
-	if (!cpus_empty(mm->cpu_vm_mask))
-		xen_smp_call_function_mask(mm->cpu_vm_mask, drop_other_mm_ref,
-					   mm, 1);
+		arch_flush_lazy_cpu_mode();
+	}
+
+	/* Get the "official" set of cpus referring to our pagetable. */
+	mask = mm->cpu_vm_mask;
+
+	/* It's possible that a vcpu may have a stale reference to our
+	   cr3, because its in lazy mode, and it hasn't yet flushed
+	   its set of pending hypercalls yet.  In this case, we can
+	   look at its actual current cr3 value, and force it to flush
+	   if needed. */
+	for_each_online_cpu(cpu) {
+		if (per_cpu(xen_current_cr3, cpu) == __pa(mm->pgd))
+			cpu_set(cpu, mask);
+	}
+
+	if (!cpus_empty(mask))
+		xen_smp_call_function_mask(mask, drop_other_mm_ref, mm, 1);
 }
 #else
 static void drop_mm_ref(struct mm_struct *mm)

@@ -11,6 +11,7 @@ void xen_copy_trap_info(struct trap_info
 
 DECLARE_PER_CPU(struct vcpu_info *, xen_vcpu);
 DECLARE_PER_CPU(unsigned long, xen_cr3);
+DECLARE_PER_CPU(unsigned long, xen_current_cr3);
 
 extern struct start_info *xen_start_info;
 extern struct shared_info *HYPERVISOR_shared_info;

@@ -687,6 +687,15 @@ static __init void xen_alloc_pt_init(str
 	make_lowmem_page_readonly(__va(PFN_PHYS(pfn)));
 }
 
+static void pin_pagetable_pfn(unsigned level, unsigned long pfn)
+{
+	struct mmuext_op op;
+	op.cmd = level;
+	op.arg1.mfn = pfn_to_mfn(pfn);
+	if (HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF))
+		BUG();
+}
+
 /* This needs to make sure the new pte page is pinned iff its being
    attached to a pinned pagetable. */
 static void xen_alloc_pt(struct mm_struct *mm, u32 pfn)

@@ -696,9 +705,10 @@ static void xen_alloc_pt(struct mm_struc
 	if (PagePinned(virt_to_page(mm->pgd))) {
 		SetPagePinned(page);
 
-		if (!PageHighMem(page))
+		if (!PageHighMem(page)) {
 			make_lowmem_page_readonly(__va(PFN_PHYS(pfn)));
-		else
+			pin_pagetable_pfn(MMUEXT_PIN_L1_TABLE, pfn);
+		} else
 			/* make sure there are no stray mappings of
 			   this page */
 			kmap_flush_unused();

@@ -711,8 +721,10 @@ static void xen_release_pt(u32 pfn)
 	struct page *page = pfn_to_page(pfn);
 
 	if (PagePinned(page)) {
-		if (!PageHighMem(page))
+		if (!PageHighMem(page)) {
+			pin_pagetable_pfn(MMUEXT_UNPIN_TABLE, pfn);
 			make_lowmem_page_readwrite(__va(PFN_PHYS(pfn)));
+		}
 	}
 }
 

@@ -827,15 +839,15 @@ static __init void xen_pagetable_setup_d
 	/* Actually pin the pagetable down, but we can't set PG_pinned
 	   yet because the page structures don't exist yet. */
 	{
-		struct mmuext_op op;
+		unsigned level;
+
 #ifdef CONFIG_X86_PAE
-		op.cmd = MMUEXT_PIN_L3_TABLE;
+		level = MMUEXT_PIN_L3_TABLE;
 #else
-		op.cmd = MMUEXT_PIN_L3_TABLE;
+		level = MMUEXT_PIN_L2_TABLE;
 #endif
-		op.arg1.mfn = pfn_to_mfn(PFN_DOWN(__pa(base)));
-		if (HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF))
-			BUG();
+
+		pin_pagetable_pfn(level, PFN_DOWN(__pa(base)));
 	}
 }
 

211 lines not on this page · raw message

@@ -1085,6 +1085,17 @@ static const struct machine_ops __initda
 };
 
 
+static void __init xen_reserve_top(void)
+{
+	unsigned long top = HYPERVISOR_VIRT_START;
+	struct xen_platform_parameters pp;
+
+	if (HYPERVISOR_xen_version(XENVER_platform_parameters, &pp) == 0)
+		top = pp.virt_start;
+
+	reserve_top_address(-top + 2 * PAGE_SIZE);
+}
+
 /* First C function to be called on Xen boot */
 asmlinkage void __init xen_start_kernel(void)
 {

@@ -1134,7 +1145,7 @@ asmlinkage void __init xen_start_kernel(
 		paravirt_ops.kernel_rpl = 0;
 
 	/* set the limit of our address space */
-	reserve_top_address(-HYPERVISOR_VIRT_START + 2 * PAGE_SIZE);
+	xen_reserve_top();
 
 	/* set up basic CPUID stuff */
 	cpu_detect(&new_cpu_data);

@@ -115,7 +115,7 @@ static void __init xen_vcpu_setup(int cp
 	info.mfn = virt_to_mfn(vcpup);
 	info.offset = offset_in_page(vcpup);
 
-	printk(KERN_DEBUG "trying to map vcpu_info %d at %p, mfn %x, offset %d\n",
+	printk(KERN_DEBUG "trying to map vcpu_info %d at %p, mfn %llx, offset %d\n",
 	       cpu, vcpup, info.mfn, info.offset);
 
 	/* Check to see if the hypervisor will put the vcpu_info

@@ -160,8 +160,9 @@ struct vcpu_set_singleshot_timer {
  */
 #define VCPUOP_register_vcpu_info   10  /* arg == struct vcpu_info */
 struct vcpu_register_vcpu_info {
-    uint32_t mfn;               /* mfn of page to place vcpu_info */
-    uint32_t offset;            /* offset within page */
+    uint64_t mfn;    /* mfn of page to place vcpu_info */
+    uint32_t offset; /* offset within page */
+    uint32_t rsvd;   /* unused */
 };
 
 #endif /* __XEN_PUBLIC_VCPU_H__ */

@@ -26,11 +26,16 @@
 
 #include "multicalls.h"
 
+#define MC_DEBUG	1
+
 #define MC_BATCH	32
 #define MC_ARGS		(MC_BATCH * 16 / sizeof(u64))
 
 struct mc_buffer {
 	struct multicall_entry entries[MC_BATCH];
+#if MC_DEBUG
+	struct multicall_entry debug[MC_BATCH];
+#endif
 	u64 args[MC_ARGS];
 	struct callback {
 		void (*fn)(void *);

@@ -56,11 +61,31 @@ void xen_mc_flush(void)
 	local_irq_save(flags);
 
 	if (b->mcidx) {
+#if MC_DEBUG
+		memcpy(b->debug, b->entries,
+		       b->mcidx * sizeof(struct multicall_entry));
+#endif
+
 		if (HYPERVISOR_multicall(b->entries, b->mcidx) != 0)
 			BUG();
 		for (i = 0; i < b->mcidx; i++)
 			if (b->entries[i].result < 0)
 				ret++;
+
+#if MC_DEBUG
+		if (ret) {
+			printk(KERN_ERR "%d multicall(s) failed: cpu %d\n",
+			       ret, smp_processor_id());
+			for(i = 0; i < b->mcidx; i++) {
+				printk("  call %2d/%d: op=%lu arg=[%lx] result=%ld\n",
+				       i+1, b->mcidx,
+				       b->debug[i].op,
+				       b->debug[i].args[0],
+				       b->entries[i].result);
+			}
+		}
+#endif
+
 		b->mcidx = 0;
 		b->argidx = 0;
 	} else

@@ -186,6 +186,19 @@ free_address(
 	void		*addr)
 {
 	a_list_t	*aentry;
+
+#ifdef CONFIG_XEN
+	/*
+	 * Xen needs to be able to make sure it can get an exclusive
+	 * RO mapping of pages it wants to turn into a pagetable.  If
+	 * a newly allocated page is also still being vmap()ed by xfs,
+	 * it will cause pagetable construction to fail.  This is a
+	 * quick workaround to always eagerly unmap pages so that Xen
+	 * is happy.
+	 */
+	vunmap(addr);
+	return;
+#endif
 
 	aentry = kmalloc(sizeof(a_list_t), GFP_NOWAIT);
 	if (likely(aentry)) {