Merge tag 'stable/for-linus-3.9-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen

Pull Xen bug-fixes from Konrad Rzeszutek Wilk:
 "This is mostly just the last stragglers of the regression bugs that
  this merge window had.  There are also two bug-fixes: one that adds an
  extra layer of security, and a regression fix for a change that was
  added in v3.7 (the v1 was faulty, the v2 works).

   - Regression fixes for C-and-P states not being parsed properly.
   - Fix possible security issue with guests triggering DoS via
     non-assigned MSI-Xs.
   - Fix regression (introduced in v3.7) with raising an event (v2).
   - Fix hastily introduced band-aid during c0 for the CR3 blowup."

* tag 'stable/for-linus-3.9-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen:
  xen/events: avoid race with raising an event in unmask_evtchn()
  xen/mmu: Move the setting of pvops.write_cr3 to later phase in bootup.
  xen/acpi-stub: Disable it b/c the acpi_processor_add is no longer called.
  xen-pciback: notify hypervisor about devices intended to be assigned to guests
  xen/acpi-processor: Don't dereference struct acpi_processor on all CPUs.

8 files changed, 74 insertions, 26 deletions

diff --git a/arch/x86/include/asm/xen/hypercall.h b/arch/x86/include/asm/xen/hypercall.h
index c20d1ce62dc6..e709884d0ef9 100644
--- a/arch/x86/include/asm/xen/hypercall.h
+++ b/arch/x86/include/asm/xen/hypercall.h
@@ -382,14 +382,14 @@ HYPERVISOR_console_io(int cmd, int count, char *str)
 	return _hypercall3(int, console_io, cmd, count, str);
 }
 
-extern int __must_check HYPERVISOR_physdev_op_compat(int, void *);
+extern int __must_check xen_physdev_op_compat(int, void *);
 
 static inline int
 HYPERVISOR_physdev_op(int cmd, void *arg)
 {
 	int rc = _hypercall2(int, physdev_op, cmd, arg);
 	if (unlikely(rc == -ENOSYS))
-		rc = HYPERVISOR_physdev_op_compat(cmd, arg);
+		rc = xen_physdev_op_compat(cmd, arg);
 	return rc;
 }
 
diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
index e8e34938c57d..6afbb2ca9a0a 100644
--- a/arch/x86/xen/mmu.c
+++ b/arch/x86/xen/mmu.c
@@ -1467,8 +1467,6 @@ static void __init xen_write_cr3_init(unsigned long cr3)
 	__xen_write_cr3(true, cr3);
 
 	xen_mc_issue(PARAVIRT_LAZY_CPU);  /* interrupts restored */
-
-	pv_mmu_ops.write_cr3 = &xen_write_cr3;
 }
 #endif
 
@@ -2122,6 +2120,7 @@ static void __init xen_post_allocator_init(void)
 #endif
 
 #ifdef CONFIG_X86_64
+	pv_mmu_ops.write_cr3 = &xen_write_cr3;
 	SetPagePinned(virt_to_page(level3_user_vsyscall));
 #endif
 	xen_mark_init_mm_pinned();
diff --git a/drivers/xen/Kconfig b/drivers/xen/Kconfig
index 5a32232cf7c1..67af155cf602 100644
--- a/drivers/xen/Kconfig
+++ b/drivers/xen/Kconfig
@@ -182,7 +182,7 @@ config XEN_PRIVCMD
 
 config XEN_STUB
 	bool "Xen stub drivers"
-	depends on XEN && X86_64
+	depends on XEN && X86_64 && BROKEN
 	default n
 	help
 	  Allow kernel to install stub drivers, to reserve space for Xen drivers,
diff --git a/drivers/xen/events.c b/drivers/xen/events.c
index d17aa41a9041..aa85881d17b2 100644
--- a/drivers/xen/events.c
+++ b/drivers/xen/events.c
@@ -403,11 +403,23 @@ static void unmask_evtchn(int port)
 
 	if (unlikely((cpu != cpu_from_evtchn(port))))
 		do_hypercall = 1;
-	else
+	else {
+		/*
+		 * Need to clear the mask before checking pending to
+		 * avoid a race with an event becoming pending.
+		 *
+		 * EVTCHNOP_unmask will only trigger an upcall if the
+		 * mask bit was set, so if a hypercall is needed
+		 * remask the event.
+		 */
+		sync_clear_bit(port, BM(&s->evtchn_mask[0]));
 		evtchn_pending = sync_test_bit(port, BM(&s->evtchn_pending[0]));
 
-	if (unlikely(evtchn_pending && xen_hvm_domain()))
-		do_hypercall = 1;
+		if (unlikely(evtchn_pending && xen_hvm_domain())) {
+			sync_set_bit(port, BM(&s->evtchn_mask[0]));
+			do_hypercall = 1;
+		}
+	}
 
 	/* Slow path (hypercall) if this is a non-local port or if this is
 	 * an hvm domain and an event is pending (hvm domains don't have
@@ -418,8 +430,6 @@ static void unmask_evtchn(int port)
 	} else {
 		struct vcpu_info *vcpu_info = __this_cpu_read(xen_vcpu);
 
-		sync_clear_bit(port, BM(&s->evtchn_mask[0]));
-
 		/*
 		 * The following is basically the equivalent of
 		 * 'hw_resend_irq'. Just like a real IO-APIC we 'lose
diff --git a/drivers/xen/fallback.c b/drivers/xen/fallback.c
index 0ef7c4d40f86..b04fb64c5a91 100644
--- a/drivers/xen/fallback.c
+++ b/drivers/xen/fallback.c
@@ -44,7 +44,7 @@ int xen_event_channel_op_compat(int cmd, void *arg)
 }
 EXPORT_SYMBOL_GPL(xen_event_channel_op_compat);
 
-int HYPERVISOR_physdev_op_compat(int cmd, void *arg)
+int xen_physdev_op_compat(int cmd, void *arg)
 {
 	struct physdev_op op;
 	int rc;
@@ -78,3 +78,4 @@ int HYPERVISOR_physdev_op_compat(int cmd, void *arg)
 
 	return rc;
 }
+EXPORT_SYMBOL_GPL(xen_physdev_op_compat);
diff --git a/drivers/xen/xen-acpi-processor.c b/drivers/xen/xen-acpi-processor.c
index f3278a6603ca..90e34ac7e522 100644
--- a/drivers/xen/xen-acpi-processor.c
+++ b/drivers/xen/xen-acpi-processor.c
@@ -505,6 +505,9 @@ static int __init xen_acpi_processor_init(void)
 
 		pr = per_cpu(processors, i);
 		perf = per_cpu_ptr(acpi_perf_data, i);
+		if (!pr)
+			continue;
+
 		pr->performance = perf;
 		rc = acpi_processor_get_performance_info(pr);
 		if (rc)
diff --git a/drivers/xen/xen-pciback/pci_stub.c b/drivers/xen/xen-pciback/pci_stub.c
index 9204126f1560..a2278ba7fb27 100644
--- a/drivers/xen/xen-pciback/pci_stub.c
+++ b/drivers/xen/xen-pciback/pci_stub.c
@@ -17,6 +17,7 @@
 #include <xen/events.h>
 #include <asm/xen/pci.h>
 #include <asm/xen/hypervisor.h>
+#include <xen/interface/physdev.h>
 #include "pciback.h"
 #include "conf_space.h"
 #include "conf_space_quirks.h"
@@ -85,37 +86,52 @@ static struct pcistub_device *pcistub_device_alloc(struct pci_dev *dev)
 static void pcistub_device_release(struct kref *kref)
 {
 	struct pcistub_device *psdev;
+	struct pci_dev *dev;
 	struct xen_pcibk_dev_data *dev_data;
 
 	psdev = container_of(kref, struct pcistub_device, kref);
-	dev_data = pci_get_drvdata(psdev->dev);
+	dev = psdev->dev;
+	dev_data = pci_get_drvdata(dev);
 
-	dev_dbg(&psdev->dev->dev, "pcistub_device_release\n");
+	dev_dbg(&dev->dev, "pcistub_device_release\n");
 
-	xen_unregister_device_domain_owner(psdev->dev);
+	xen_unregister_device_domain_owner(dev);
 
 	/* Call the reset function which does not take lock as this
 	 * is called from "unbind" which takes a device_lock mutex.
 	 */
-	__pci_reset_function_locked(psdev->dev);
-	if (pci_load_and_free_saved_state(psdev->dev,
-					  &dev_data->pci_saved_state)) {
-		dev_dbg(&psdev->dev->dev, "Could not reload PCI state\n");
-	} else
-		pci_restore_state(psdev->dev);
+	__pci_reset_function_locked(dev);
+	if (pci_load_and_free_saved_state(dev, &dev_data->pci_saved_state))
+		dev_dbg(&dev->dev, "Could not reload PCI state\n");
+	else
+		pci_restore_state(dev);
+
+	if (pci_find_capability(dev, PCI_CAP_ID_MSIX)) {
+		struct physdev_pci_device ppdev = {
+			.seg = pci_domain_nr(dev->bus),
+			.bus = dev->bus->number,
+			.devfn = dev->devfn
+		};
+		int err = HYPERVISOR_physdev_op(PHYSDEVOP_release_msix,
+						&ppdev);
+
+		if (err)
+			dev_warn(&dev->dev, "MSI-X release failed (%d)\n",
+				 err);
+	}
 
 	/* Disable the device */
-	xen_pcibk_reset_device(psdev->dev);
+	xen_pcibk_reset_device(dev);
 
 	kfree(dev_data);
-	pci_set_drvdata(psdev->dev, NULL);
+	pci_set_drvdata(dev, NULL);
 
 	/* Clean-up the device */
-	xen_pcibk_config_free_dyn_fields(psdev->dev);
-	xen_pcibk_config_free_dev(psdev->dev);
+	xen_pcibk_config_free_dyn_fields(dev);
+	xen_pcibk_config_free_dev(dev);
 
-	psdev->dev->dev_flags &= ~PCI_DEV_FLAGS_ASSIGNED;
-	pci_dev_put(psdev->dev);
+	dev->dev_flags &= ~PCI_DEV_FLAGS_ASSIGNED;
+	pci_dev_put(dev);
 
 	kfree(psdev);
 }
@@ -355,6 +371,19 @@ static int pcistub_init_device(struct pci_dev *dev)
 	if (err)
 		goto config_release;
 
+	if (pci_find_capability(dev, PCI_CAP_ID_MSIX)) {
+		struct physdev_pci_device ppdev = {
+			.seg = pci_domain_nr(dev->bus),
+			.bus = dev->bus->number,
+			.devfn = dev->devfn
+		};
+
+		err = HYPERVISOR_physdev_op(PHYSDEVOP_prepare_msix, &ppdev);
+		if (err)
+			dev_err(&dev->dev, "MSI-X preparation failed (%d)\n",
+				err);
+	}
+
 	/* We need the device active to save the state. */
 	dev_dbg(&dev->dev, "save state of device\n");
 	pci_save_state(dev);
diff --git a/include/xen/interface/physdev.h b/include/xen/interface/physdev.h
index 1844d31f4552..7000bb1f6e96 100644
--- a/include/xen/interface/physdev.h
+++ b/include/xen/interface/physdev.h
@@ -251,6 +251,12 @@ struct physdev_pci_device_add {
 
 #define PHYSDEVOP_pci_device_remove     26
 #define PHYSDEVOP_restore_msi_ext       27
+/*
+ * Dom0 should use these two to announce MMIO resources assigned to
+ * MSI-X capable devices won't (prepare) or may (release) change.
+ */
+#define PHYSDEVOP_prepare_msix          30
+#define PHYSDEVOP_release_msix          31
 struct physdev_pci_device {
     /* IN */
     uint16_t seg;