Merge tag 'for-linus-7.0-rc3-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip

Pull xen fixes from Juergen Gross:

 - a cleanup of arch/x86/kernel/head_64.S removing the pre-built page
   tables for Xen guests

 - a small comment update

 - another cleanup for Xen PVH guests mode

 - fix an issue with Xen PV-devices backed by driver domains

* tag 'for-linus-7.0-rc3-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xen/xenbus: better handle backend crash
  xenbus: add xenbus_device parameter to xenbus_read_driver_state()
  x86/PVH: Use boot params to pass RSDP address in start_info page
  x86/xen: update outdated comment
  xen/acpi-processor: fix _CST detection using undersized evaluation buffer
  x86/xen: Build identity mapping page tables dynamically for XENPV

arch/x86/include/asm/pgtable_64.h

@@ -19,10 +19,8 @@
 extern p4d_t level4_kernel_pgt[512];
 extern p4d_t level4_ident_pgt[512];
 extern pud_t level3_kernel_pgt[512];
-extern pud_t level3_ident_pgt[512];
 extern pmd_t level2_kernel_pgt[512];
 extern pmd_t level2_fixmap_pgt[512];
-extern pmd_t level2_ident_pgt[512];
 extern pte_t level1_fixmap_pgt[512 * FIXMAP_PMD_NUM];
 extern pgd_t init_top_pgt[];
 

arch/x86/kernel/head_64.S

@@ -616,38 +616,10 @@ SYM_DATA(early_recursion_flag, .long 0)
 
 	.data
 
-#if defined(CONFIG_XEN_PV) || defined(CONFIG_PVH)
-SYM_DATA_START_PTI_ALIGNED(init_top_pgt)
-	.quad   level3_ident_pgt - __START_KERNEL_map + _KERNPG_TABLE_NOENC
-	.org    init_top_pgt + L4_PAGE_OFFSET*8, 0
-	.quad   level3_ident_pgt - __START_KERNEL_map + _KERNPG_TABLE_NOENC
-	.org    init_top_pgt + L4_START_KERNEL*8, 0
-	/* (2^48-(2*1024*1024*1024))/(2^39) = 511 */
-	.quad   level3_kernel_pgt - __START_KERNEL_map + _PAGE_TABLE_NOENC
-	.fill	PTI_USER_PGD_FILL,8,0
-SYM_DATA_END(init_top_pgt)
-
-SYM_DATA_START_PAGE_ALIGNED(level3_ident_pgt)
-	.quad	level2_ident_pgt - __START_KERNEL_map + _KERNPG_TABLE_NOENC
-	.fill	511, 8, 0
-SYM_DATA_END(level3_ident_pgt)
-SYM_DATA_START_PAGE_ALIGNED(level2_ident_pgt)
-	/*
-	 * Since I easily can, map the first 1G.
-	 * Don't set NX because code runs from these pages.
-	 *
-	 * Note: This sets _PAGE_GLOBAL despite whether
-	 * the CPU supports it or it is enabled.  But,
-	 * the CPU should ignore the bit.
-	 */
-	PMDS(0, __PAGE_KERNEL_IDENT_LARGE_EXEC, PTRS_PER_PMD)
-SYM_DATA_END(level2_ident_pgt)
-#else
 SYM_DATA_START_PTI_ALIGNED(init_top_pgt)
 	.fill	512,8,0
 	.fill	PTI_USER_PGD_FILL,8,0
 SYM_DATA_END(init_top_pgt)
-#endif
 
 SYM_DATA_START_PAGE_ALIGNED(level4_kernel_pgt)
 	.fill	511,8,0

arch/x86/platform/pvh/enlighten.c

@@ -25,11 +25,6 @@ struct hvm_start_info __initdata pvh_start_info;
 
 const unsigned int __initconst pvh_start_info_sz = sizeof(pvh_start_info);
 
-static u64 __init pvh_get_root_pointer(void)
-{
-	return pvh_start_info.rsdp_paddr;
-}
-
 /*
  * Xen guests are able to obtain the memory map from the hypervisor via the
  * HYPERVISOR_memory_op hypercall.
@@ -95,7 +90,7 @@ static void __init init_pvh_bootparams(bool xen_guest)
 	pvh_bootparams.hdr.version = (2 << 8) | 12;
 	pvh_bootparams.hdr.type_of_loader = ((xen_guest ? 0x9 : 0xb) << 4) | 0;
 
-	x86_init.acpi.get_root_pointer = pvh_get_root_pointer;
+	pvh_bootparams.acpi_rsdp_addr = pvh_start_info.rsdp_paddr;
 }
 
 /*

arch/x86/xen/enlighten_pv.c

@@ -392,7 +392,7 @@ static void __init xen_init_capabilities(void)
 
 	/*
 	 * Xen PV would need some work to support PCID: CR3 handling as well
-	 * as xen_flush_tlb_others() would need updating.
+	 * as xen_flush_tlb_multi() would need updating.
 	 */
 	setup_clear_cpu_cap(X86_FEATURE_PCID);
 

arch/x86/xen/mmu_pv.c

@@ -105,6 +105,9 @@ pte_t xen_make_pte_init(pteval_t pte);
 static pud_t level3_user_vsyscall[PTRS_PER_PUD] __page_aligned_bss;
 #endif
 
+static pud_t level3_ident_pgt[PTRS_PER_PUD] __page_aligned_bss;
+static pmd_t level2_ident_pgt[PTRS_PER_PMD] __page_aligned_bss;
+
 /*
  * Protects atomic reservation decrease/increase against concurrent increases.
  * Also protects non-atomic updates of current_pages and balloon lists.
@@ -1777,6 +1780,12 @@ void __init xen_setup_kernel_pagetable(pgd_t *pgd, unsigned long max_pfn)
 	/* Zap identity mapping */
 	init_top_pgt[0] = __pgd(0);
 
+	init_top_pgt[pgd_index(__PAGE_OFFSET_BASE_L4)].pgd =
+		__pa_symbol(level3_ident_pgt) + _KERNPG_TABLE_NOENC;
+	init_top_pgt[pgd_index(__START_KERNEL_map)].pgd =
+		__pa_symbol(level3_kernel_pgt) + _PAGE_TABLE_NOENC;
+	level3_ident_pgt[0].pud = __pa_symbol(level2_ident_pgt) + _KERNPG_TABLE_NOENC;
+
 	/* Pre-constructed entries are in pfn, so convert to mfn */
 	/* L4[273] -> level3_ident_pgt  */
 	/* L4[511] -> level3_kernel_pgt */

drivers/net/xen-netfront.c

@@ -1646,7 +1646,7 @@ static int xennet_xdp_set(struct net_device *dev, struct bpf_prog *prog,
 
 	/* avoid the race with XDP headroom adjustment */
 	wait_event(module_wq,
-		   xenbus_read_driver_state(np->xbdev->otherend) ==
+		   xenbus_read_driver_state(np->xbdev, np->xbdev->otherend) ==
 		   XenbusStateReconfigured);
 	np->netfront_xdp_enabled = true;
 
@@ -1764,9 +1764,9 @@ static struct net_device *xennet_create_dev(struct xenbus_device *dev)
 	do {
 		xenbus_switch_state(dev, XenbusStateInitialising);
 		err = wait_event_timeout(module_wq,
-				 xenbus_read_driver_state(dev->otherend) !=
+				 xenbus_read_driver_state(dev, dev->otherend) !=
 				 XenbusStateClosed &&
-				 xenbus_read_driver_state(dev->otherend) !=
+				 xenbus_read_driver_state(dev, dev->otherend) !=
 				 XenbusStateUnknown, XENNET_TIMEOUT);
 	} while (!err);
 
@@ -2626,31 +2626,31 @@ static void xennet_bus_close(struct xenbus_device *dev)
 {
 	int ret;
 
-	if (xenbus_read_driver_state(dev->otherend) == XenbusStateClosed)
+	if (xenbus_read_driver_state(dev, dev->otherend) == XenbusStateClosed)
 		return;
 	do {
 		xenbus_switch_state(dev, XenbusStateClosing);
 		ret = wait_event_timeout(module_wq,
-				   xenbus_read_driver_state(dev->otherend) ==
+				xenbus_read_driver_state(dev, dev->otherend) ==
-				   XenbusStateClosing ||
+				XenbusStateClosing ||
-				   xenbus_read_driver_state(dev->otherend) ==
+				xenbus_read_driver_state(dev, dev->otherend) ==
-				   XenbusStateClosed ||
+				XenbusStateClosed ||
-				   xenbus_read_driver_state(dev->otherend) ==
+				xenbus_read_driver_state(dev, dev->otherend) ==
-				   XenbusStateUnknown,
+				XenbusStateUnknown,
-				   XENNET_TIMEOUT);
+				XENNET_TIMEOUT);
 	} while (!ret);
 
-	if (xenbus_read_driver_state(dev->otherend) == XenbusStateClosed)
+	if (xenbus_read_driver_state(dev, dev->otherend) == XenbusStateClosed)
 		return;
 
 	do {
 		xenbus_switch_state(dev, XenbusStateClosed);
 		ret = wait_event_timeout(module_wq,
-				   xenbus_read_driver_state(dev->otherend) ==
+				xenbus_read_driver_state(dev, dev->otherend) ==
-				   XenbusStateClosed ||
+				XenbusStateClosed ||
-				   xenbus_read_driver_state(dev->otherend) ==
+				xenbus_read_driver_state(dev, dev->otherend) ==
-				   XenbusStateUnknown,
+				XenbusStateUnknown,
-				   XENNET_TIMEOUT);
+				XENNET_TIMEOUT);
 	} while (!ret);
 }
 

drivers/pci/xen-pcifront.c

@@ -856,7 +856,7 @@ static void pcifront_try_connect(struct pcifront_device *pdev)
 	int err;
 
 	/* Only connect once */
-	if (xenbus_read_driver_state(pdev->xdev->nodename) !=
+	if (xenbus_read_driver_state(pdev->xdev, pdev->xdev->nodename) !=
 	    XenbusStateInitialised)
 		return;
 
@@ -876,7 +876,7 @@ static int pcifront_try_disconnect(struct pcifront_device *pdev)
 	enum xenbus_state prev_state;
 
 
-	prev_state = xenbus_read_driver_state(pdev->xdev->nodename);
+	prev_state = xenbus_read_driver_state(pdev->xdev, pdev->xdev->nodename);
 
 	if (prev_state >= XenbusStateClosing)
 		goto out;
@@ -895,7 +895,7 @@ out:
 
 static void pcifront_attach_devices(struct pcifront_device *pdev)
 {
-	if (xenbus_read_driver_state(pdev->xdev->nodename) ==
+	if (xenbus_read_driver_state(pdev->xdev, pdev->xdev->nodename) ==
 	    XenbusStateReconfiguring)
 		pcifront_connect(pdev);
 }
@@ -909,7 +909,7 @@ static int pcifront_detach_devices(struct pcifront_device *pdev)
 	struct pci_dev *pci_dev;
 	char str[64];
 
-	state = xenbus_read_driver_state(pdev->xdev->nodename);
+	state = xenbus_read_driver_state(pdev->xdev, pdev->xdev->nodename);
 	if (state == XenbusStateInitialised) {
 		dev_dbg(&pdev->xdev->dev, "Handle skipped connect.\n");
 		/* We missed Connected and need to initialize. */

drivers/scsi/xen-scsifront.c

@@ -1175,7 +1175,7 @@ static void scsifront_backend_changed(struct xenbus_device *dev,
 			return;
 		}
 
-		if (xenbus_read_driver_state(dev->nodename) ==
+		if (xenbus_read_driver_state(dev, dev->nodename) ==
 		    XenbusStateInitialised)
 			scsifront_do_lun_hotplug(info, VSCSIFRONT_OP_ADD_LUN);
 

drivers/xen/xen-acpi-processor.c

@@ -378,11 +378,8 @@ read_acpi_id(acpi_handle handle, u32 lvl, void *context, void **rv)
 			 acpi_psd[acpi_id].domain);
 	}
 
-	status = acpi_evaluate_object(handle, "_CST", NULL, &buffer);
+	if (!pblk && !acpi_has_method(handle, "_CST"))
-	if (ACPI_FAILURE(status)) {
+		return AE_OK;
-		if (!pblk)
-			return AE_OK;
-	}
 	/* .. and it has a C-state */
 	__set_bit(acpi_id, acpi_id_cst_present);
 

drivers/xen/xen-pciback/xenbus.c

@@ -149,12 +149,12 @@ static int xen_pcibk_attach(struct xen_pcibk_device *pdev)
 
 	mutex_lock(&pdev->dev_lock);
 	/* Make sure we only do this setup once */
-	if (xenbus_read_driver_state(pdev->xdev->nodename) !=
+	if (xenbus_read_driver_state(pdev->xdev, pdev->xdev->nodename) !=
 	    XenbusStateInitialised)
 		goto out;
 
 	/* Wait for frontend to state that it has published the configuration */
-	if (xenbus_read_driver_state(pdev->xdev->otherend) !=
+	if (xenbus_read_driver_state(pdev->xdev, pdev->xdev->otherend) !=
 	    XenbusStateInitialised)
 		goto out;
 
@@ -374,7 +374,7 @@ static int xen_pcibk_reconfigure(struct xen_pcibk_device *pdev,
 	dev_dbg(&pdev->xdev->dev, "Reconfiguring device ...\n");
 
 	mutex_lock(&pdev->dev_lock);
-	if (xenbus_read_driver_state(pdev->xdev->nodename) != state)
+	if (xenbus_read_driver_state(pdev->xdev, pdev->xdev->nodename) != state)
 		goto out;
 
 	err = xenbus_scanf(XBT_NIL, pdev->xdev->nodename, "num_devs", "%d",
@@ -572,7 +572,7 @@ static int xen_pcibk_setup_backend(struct xen_pcibk_device *pdev)
 	/* It's possible we could get the call to setup twice, so make sure
 	 * we're not already connected.
 	 */
-	if (xenbus_read_driver_state(pdev->xdev->nodename) !=
+	if (xenbus_read_driver_state(pdev->xdev, pdev->xdev->nodename) !=
 	    XenbusStateInitWait)
 		goto out;
 
@@ -662,7 +662,7 @@ static void xen_pcibk_be_watch(struct xenbus_watch *watch,
 	struct xen_pcibk_device *pdev =
 	    container_of(watch, struct xen_pcibk_device, be_watch);
 
-	switch (xenbus_read_driver_state(pdev->xdev->nodename)) {
+	switch (xenbus_read_driver_state(pdev->xdev, pdev->xdev->nodename)) {
 	case XenbusStateInitWait:
 		xen_pcibk_setup_backend(pdev);
 		break;

drivers/xen/xenbus/xenbus_client.c

@@ -226,8 +226,9 @@ __xenbus_switch_state(struct xenbus_device *dev,
 	struct xenbus_transaction xbt;
 	int current_state;
 	int err, abort;
+	bool vanished = false;
 
-	if (state == dev->state)
+	if (state == dev->state || dev->vanished)
 		return 0;
 
 again:
@@ -242,6 +243,10 @@ again:
 	err = xenbus_scanf(xbt, dev->nodename, "state", "%d", &current_state);
 	if (err != 1)
 		goto abort;
+	if (current_state != dev->state && current_state == XenbusStateInitialising) {
+		vanished = true;
+		goto abort;
+	}
 
 	err = xenbus_printf(xbt, dev->nodename, "state", "%d", state);
 	if (err) {
@@ -256,7 +261,7 @@ abort:
 		if (err == -EAGAIN && !abort)
 			goto again;
 		xenbus_switch_fatal(dev, depth, err, "ending transaction");
-	} else
+	} else if (!vanished)
 		dev->state = state;
 
 	return 0;
@@ -931,14 +936,20 @@ static int xenbus_unmap_ring_hvm(struct xenbus_device *dev, void *vaddr)
 
 /**
  * xenbus_read_driver_state - read state from a store path
+ * @dev: xenbus device pointer
  * @path: path for driver
  *
  * Returns: the state of the driver rooted at the given store path, or
  * XenbusStateUnknown if no state can be read.
  */
-enum xenbus_state xenbus_read_driver_state(const char *path)
+enum xenbus_state xenbus_read_driver_state(const struct xenbus_device *dev,
+					   const char *path)
 {
 	enum xenbus_state result;
+
+	if (dev && dev->vanished)
+		return XenbusStateUnknown;
+
 	int err = xenbus_gather(XBT_NIL, path, "state", "%d", &result, NULL);
 	if (err)
 		result = XenbusStateUnknown;

drivers/xen/xenbus/xenbus_probe.c

@@ -191,7 +191,7 @@ void xenbus_otherend_changed(struct xenbus_watch *watch,
 		return;
 	}
 
-	state = xenbus_read_driver_state(dev->otherend);
+	state = xenbus_read_driver_state(dev, dev->otherend);
 
 	dev_dbg(&dev->dev, "state is %d, (%s), %s, %s\n",
 		state, xenbus_strstate(state), dev->otherend_watch.node, path);
@@ -364,7 +364,7 @@ void xenbus_dev_remove(struct device *_dev)
 	 * closed.
 	 */
 	if (!drv->allow_rebind ||
-	    xenbus_read_driver_state(dev->nodename) == XenbusStateClosing)
+	    xenbus_read_driver_state(dev, dev->nodename) == XenbusStateClosing)
 		xenbus_switch_state(dev, XenbusStateClosed);
 }
 EXPORT_SYMBOL_GPL(xenbus_dev_remove);
@@ -444,6 +444,9 @@ static void xenbus_cleanup_devices(const char *path, struct bus_type *bus)
 		info.dev = NULL;
 		bus_for_each_dev(bus, NULL, &info, cleanup_dev);
 		if (info.dev) {
+			dev_warn(&info.dev->dev,
+				 "device forcefully removed from xenstore\n");
+			info.dev->vanished = true;
 			device_unregister(&info.dev->dev);
 			put_device(&info.dev->dev);
 		}
@@ -514,7 +517,7 @@ int xenbus_probe_node(struct xen_bus_type *bus,
 	size_t stringlen;
 	char *tmpstring;
 
-	enum xenbus_state state = xenbus_read_driver_state(nodename);
+	enum xenbus_state state = xenbus_read_driver_state(NULL, nodename);
 
 	if (state != XenbusStateInitialising) {
 		/* Device is not new, so ignore it.  This can happen if a
@@ -659,6 +662,39 @@ void xenbus_dev_changed(const char *node, struct xen_bus_type *bus)
 		return;
 
 	dev = xenbus_device_find(root, &bus->bus);
+	/*
+	 * Backend domain crash results in not coordinated frontend removal,
+	 * without going through XenbusStateClosing. If this is a new instance
+	 * of the same device Xen tools will have reset the state to
+	 * XenbusStateInitializing.
+	 * It might be that the backend crashed early during the init phase of
+	 * device setup, in which case the known state would have been
+	 * XenbusStateInitializing. So test the backend domid to match the
+	 * saved one. In case the new backend happens to have the same domid as
+	 * the old one, we can just carry on, as there is no inconsistency
+	 * resulting in this case.
+	 */
+	if (dev && !strcmp(bus->root, "device")) {
+		enum xenbus_state state = xenbus_read_driver_state(dev, dev->nodename);
+		unsigned int backend = xenbus_read_unsigned(root, "backend-id",
+							    dev->otherend_id);
+
+		if (state == XenbusStateInitialising &&
+		    (state != dev->state || backend != dev->otherend_id)) {
+			/*
+			 * State has been reset, assume the old one vanished
+			 * and new one needs to be probed.
+			 */
+			dev_warn(&dev->dev,
+				 "state reset occurred, reconnecting\n");
+			dev->vanished = true;
+		}
+		if (dev->vanished) {
+			device_unregister(&dev->dev);
+			put_device(&dev->dev);
+			dev = NULL;
+		}
+	}
 	if (!dev)
 		xenbus_probe_node(bus, type, root);
 	else

drivers/xen/xenbus/xenbus_probe_frontend.c

@@ -253,7 +253,7 @@ static int print_device_status(struct device *dev, void *data)
 	} else if (xendev->state < XenbusStateConnected) {
 		enum xenbus_state rstate = XenbusStateUnknown;
 		if (xendev->otherend)
-			rstate = xenbus_read_driver_state(xendev->otherend);
+			rstate = xenbus_read_driver_state(xendev, xendev->otherend);
 		pr_warn("Timeout connecting to device: %s (local state %d, remote state %d)\n",
 			xendev->nodename, xendev->state, rstate);
 	}

include/xen/xenbus.h

@@ -80,6 +80,7 @@ struct xenbus_device {
 	const char *devicetype;
 	const char *nodename;
 	const char *otherend;
+	bool vanished;
 	int otherend_id;
 	struct xenbus_watch otherend_watch;
 	struct device dev;
@@ -228,7 +229,8 @@ int xenbus_unmap_ring_vfree(struct xenbus_device *dev, void *vaddr);
 int xenbus_alloc_evtchn(struct xenbus_device *dev, evtchn_port_t *port);
 int xenbus_free_evtchn(struct xenbus_device *dev, evtchn_port_t port);
 
-enum xenbus_state xenbus_read_driver_state(const char *path);
+enum xenbus_state xenbus_read_driver_state(const struct xenbus_device *dev,
+					   const char *path);
 
 __printf(3, 4)
 void xenbus_dev_error(struct xenbus_device *dev, int err, const char *fmt, ...);