1 files changed, 38 insertions, 17 deletions
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index 75fd36971d0b..e334bf7cb327 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -53,7 +53,23 @@ EXPORT_SYMBOL_GPL(hypercall_page);
 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+ * out of date, but it will never be set early).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);
@@ -610,32 +626,36 @@ 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)) {
+        mcs = xen_mc_entry(sizeof(*op));  /* disables interrupts */
-                /* just a simple tlb flush */
-                xen_flush_tlb();
-                return;
-        }
+        /* Update while interrupts are disabled, so its atomic with
+           respect to ipis */
        x86_write_percpu(xen_cr3, cr3);
+        op = mcs.args;
+        op->cmd = MMUEXT_NEW_BASEPTR;
+        op->arg1.mfn = mfn;
-        {
+        MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);
-                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);
+        /* 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);
+        xen_mc_issue(PARAVIRT_LAZY_CPU);  /* interrupts restored */
-        }
 }
 /* Early in boot, while setting up the initial pagetable, assume
@@ -1120,6 +1140,7 @@ asmlinkage void __init xen_start_kernel(void)
        /* 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

diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index 75fd36971d0b..e334bf7cb327 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c
@@ -53,7 +53,23 @@ EXPORT_SYMBOL_GPL(hypercall_page);
53		53
54	DEFINE_PER_CPU(struct vcpu_info *, xen_vcpu);	54	DEFINE_PER_CPU(struct vcpu_info *, xen_vcpu);
55	DEFINE_PER_CPU(struct vcpu_info, xen_vcpu_info);	55	DEFINE_PER_CPU(struct vcpu_info, xen_vcpu_info);
56	DEFINE_PER_CPU(unsigned long, xen_cr3);	56
		57	/*
		58	* Note about cr3 (pagetable base) values:
		59	*
		60	* xen_cr3 contains the current logical cr3 value; it contains the
		61	* last set cr3. This may not be the current effective cr3, because
		62	* its update may be being lazily deferred. However, a vcpu looking
		63	* at its own cr3 can use this value knowing that it everything will
		64	* be self-consistent.
		65	*
		66	* xen_current_cr3 contains the actual vcpu cr3; it is set once the
		67	* hypercall to set the vcpu cr3 is complete (so it may be a little
		68	* out of date, but it will never be set early). If one vcpu is
		69	* looking at another vcpu's cr3 value, it should use this variable.
		70	*/
		71	DEFINE_PER_CPU(unsigned long, xen_cr3); /* cr3 stored as physaddr */
		72	DEFINE_PER_CPU(unsigned long, xen_current_cr3); /* actual vcpu cr3 */
57		73
58	struct start_info *xen_start_info;	74	struct start_info *xen_start_info;
59	EXPORT_SYMBOL_GPL(xen_start_info);	75	EXPORT_SYMBOL_GPL(xen_start_info);
@@ -610,32 +626,36 @@ static unsigned long xen_read_cr3(void)
610	return x86_read_percpu(xen_cr3);	626	return x86_read_percpu(xen_cr3);
611	}	627	}
612		628
		629	static void set_current_cr3(void *v)
		630	{
		631	x86_write_percpu(xen_current_cr3, (unsigned long)v);
		632	}
		633
613	static void xen_write_cr3(unsigned long cr3)	634	static void xen_write_cr3(unsigned long cr3)
614	{	635	{
		636	struct mmuext_op *op;
		637	struct multicall_space mcs;
		638	unsigned long mfn = pfn_to_mfn(PFN_DOWN(cr3));
		639
615	BUG_ON(preemptible());	640	BUG_ON(preemptible());
616		641
617	if (cr3 == x86_read_percpu(xen_cr3)) {	642	mcs = xen_mc_entry(sizeof(op)); / disables interrupts */
618	/* just a simple tlb flush */
619	xen_flush_tlb();
620	return;
621	}
622		643
		644	/* Update while interrupts are disabled, so its atomic with
		645	respect to ipis */
623	x86_write_percpu(xen_cr3, cr3);	646	x86_write_percpu(xen_cr3, cr3);
624		647
		648	op = mcs.args;
		649	op->cmd = MMUEXT_NEW_BASEPTR;
		650	op->arg1.mfn = mfn;
625		651
626	{	652	MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);
627	struct mmuext_op *op;
628	struct multicall_space mcs = xen_mc_entry(sizeof(*op));
629	unsigned long mfn = pfn_to_mfn(PFN_DOWN(cr3));
630
631	op = mcs.args;
632	op->cmd = MMUEXT_NEW_BASEPTR;
633	op->arg1.mfn = mfn;
634		653
635	MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);	654	/* Update xen_update_cr3 once the batch has actually
		655	been submitted. */
		656	xen_mc_callback(set_current_cr3, (void *)cr3);
636		657
637	xen_mc_issue(PARAVIRT_LAZY_CPU);	658	xen_mc_issue(PARAVIRT_LAZY_CPU); /* interrupts restored */
638	}
639	}	659	}
640		660
641	/* Early in boot, while setting up the initial pagetable, assume	661	/* Early in boot, while setting up the initial pagetable, assume
@@ -1120,6 +1140,7 @@ asmlinkage void __init xen_start_kernel(void)
1120	/* keep using Xen gdt for now; no urgent need to change it */	1140	/* keep using Xen gdt for now; no urgent need to change it */
1121		1141
1122	x86_write_percpu(xen_cr3, __pa(pgd));	1142	x86_write_percpu(xen_cr3, __pa(pgd));
		1143	x86_write_percpu(xen_current_cr3, __pa(pgd));
1123		1144
1124	#ifdef CONFIG_SMP	1145	#ifdef CONFIG_SMP
1125	/* Don't do the full vcpu_info placement stuff until we have a	1146	/* Don't do the full vcpu_info placement stuff until we have a