Merge tag 'for-linus-7.0-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip

torvalds · torvalds · commit 33120a2f8fc4 · 2026-02-09T20:38:27.000-08:00
Pull xen updates from Juergen Gross:

 - fix running as Xen PVH guest in 32-bit mode without PAE

 - fix PV device handling for suspend/resume when running as
   a Xen guest

 - clean up workqueue usage

 - fix the Xen balloon driver for PVH dom0

 - introduce the possibility to use hypercalls for console
   messages in unprivileged guests

 - enable Xen dom0 use of virtio devices in nested virtualization
   setups

 - simplify the xen-mcelog driver

* tag 'for-linus-7.0-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xenbus: Rename helpers to freeze/thaw/restore
  xenbus: Use .freeze/.thaw to handle xenbus devices
  xen/mcelog: simplify MCE_GETCLEAR_FLAGS using xchg()
  xen/balloon: improve accuracy of initial balloon target for dom0
  Partial revert "x86/xen: fix balloon target initialization for PVH dom0"
  xen: introduce xen_console_io option
  xen/virtio: Don't use grant-dma-ops when running as Dom0
  x86/xen/pvh: Enable PAE mode for 32-bit guest only when CONFIG_X86_PAE is set
  xen: privcmd: WQ_PERCPU added to alloc_workqueue users
  xen/events: replace use of system_wq with system_percpu_wq
diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
@@ -8437,6 +8437,11 @@ Kernel parameters
 			save/restore/migration must be enabled to handle larger
 			domains.
 
+	xen_console_io	[XEN,EARLY]
+			Boolean option to enable/disable the usage of the Xen
+			console_io hypercalls to read and write to the console.
+			Mostly useful for debugging and development.
+
 	xen_emul_unplug=		[HW,X86,XEN,EARLY]
 			Unplug Xen emulated devices
 			Format: [unplug0,][unplug1]
diff --git a/arch/x86/platform/pvh/head.S b/arch/x86/platform/pvh/head.S
@@ -91,10 +91,12 @@ SYM_CODE_START(pvh_start_xen)
 
 	leal rva(early_stack_end)(%ebp), %esp
 
+#if defined(CONFIG_X86_64) || defined(CONFIG_X86_PAE)
 	/* Enable PAE mode. */
 	mov %cr4, %eax
 	orl $X86_CR4_PAE, %eax
 	mov %eax, %cr4
+#endif
 
 #ifdef CONFIG_X86_64
 	/* Enable Long mode. */
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
@@ -470,7 +470,7 @@ int __init arch_xen_unpopulated_init(struct resource **res)
 		 * driver to know how much of the physmap is unpopulated and
 		 * set an accurate initial memory target.
 		 */
-		xen_released_pages += xen_extra_mem[i].n_pfns;
+		xen_unpopulated_pages += xen_extra_mem[i].n_pfns;
 		/* Zero so region is not also added to the balloon driver. */
 		xen_extra_mem[i].n_pfns = 0;
 	}
diff --git a/drivers/tty/hvc/hvc_xen.c b/drivers/tty/hvc/hvc_xen.c
@@ -51,6 +51,22 @@ static DEFINE_SPINLOCK(xencons_lock);
 
 /* ------------------------------------------------------------------ */
 
+static bool xen_console_io = false;
+static int __initdata opt_console_io = -1;
+
+static int __init parse_xen_console_io(char *arg)
+{
+	bool val;
+	int ret;
+
+	ret = kstrtobool(arg, &val);
+	if (ret == 0)
+		opt_console_io = (int)val;
+
+	return ret;
+}
+early_param("xen_console_io", parse_xen_console_io);
+
 static struct xencons_info *vtermno_to_xencons(int vtermno)
 {
 	struct xencons_info *entry, *ret = NULL;
@@ -331,7 +347,7 @@ static int xen_initial_domain_console_init(void)
 	struct xencons_info *info;
 	unsigned long flags;
 
-	if (!xen_initial_domain())
+	if (!xen_console_io)
 		return -ENODEV;
 
 	info = vtermno_to_xencons(HVC_COOKIE);
@@ -369,7 +385,7 @@ void xen_console_resume(void)
 {
 	struct xencons_info *info = vtermno_to_xencons(HVC_COOKIE);
 	if (info != NULL && info->irq) {
-		if (!xen_initial_domain())
+		if (!xen_console_io)
 			xen_console_update_evtchn(info);
 		rebind_evtchn_irq(info->evtchn, info->irq);
 	}
@@ -601,7 +617,7 @@ static int __init xen_hvc_init(void)
 	if (!xen_domain())
 		return -ENODEV;
 
-	if (xen_initial_domain()) {
+	if (xen_console_io) {
 		ops = &dom0_hvc_ops;
 		r = xen_initial_domain_console_init();
 		if (r < 0)
@@ -647,14 +663,17 @@ static int __init xen_hvc_init(void)
 }
 device_initcall(xen_hvc_init);
 
-static int xen_cons_init(void)
+static int __init xen_cons_init(void)
 {
 	const struct hv_ops *ops;
 
+	xen_console_io = opt_console_io >= 0 ? opt_console_io :
+					       xen_initial_domain();
+
 	if (!xen_domain())
 		return 0;
 
-	if (xen_initial_domain())
+	if (xen_console_io)
 		ops = &dom0_hvc_ops;
 	else {
 		int r;
diff --git a/drivers/xen/balloon.c b/drivers/xen/balloon.c
@@ -724,19 +724,33 @@ static int __init balloon_add_regions(void)
 static int __init balloon_init(void)
 {
 	struct task_struct *task;
+	long current_pages = 0;
+	domid_t domid = DOMID_SELF;
 	int rc;
 
 	if (!xen_domain())
 		return -ENODEV;
 
 	pr_info("Initialising balloon driver\n");
 
-	if (xen_released_pages >= get_num_physpages()) {
-		WARN(1, "Released pages underflow current target");
-		return -ERANGE;
+	if (xen_initial_domain())
+		current_pages = HYPERVISOR_memory_op(XENMEM_current_reservation,
+		                                     &domid);
+	if (current_pages <= 0) {
+		if (xen_pv_domain()) {
+			if (xen_released_pages >= xen_start_info->nr_pages)
+				goto underflow;
+			current_pages = min(xen_start_info->nr_pages -
+			                    xen_released_pages, max_pfn);
+		} else {
+			if (xen_unpopulated_pages >= get_num_physpages())
+				goto underflow;
+			current_pages = get_num_physpages() -
+			                xen_unpopulated_pages;
+		}
 	}
 
-	balloon_stats.current_pages = get_num_physpages() - xen_released_pages;
+	balloon_stats.current_pages = current_pages;
 	balloon_stats.target_pages  = balloon_stats.current_pages;
 	balloon_stats.balloon_low   = 0;
 	balloon_stats.balloon_high  = 0;
@@ -767,6 +781,10 @@ static int __init balloon_init(void)
 	xen_balloon_init();
 
 	return 0;
+
+ underflow:
+	WARN(1, "Released pages underflow current target");
+	return -ERANGE;
 }
 subsys_initcall(balloon_init);
 
diff --git a/drivers/xen/events/events_base.c b/drivers/xen/events/events_base.c
@@ -581,7 +581,7 @@ static void lateeoi_list_add(struct irq_info *info)
 					eoi_list);
 	if (!elem || info->eoi_time < elem->eoi_time) {
 		list_add(&info->eoi_list, &eoi->eoi_list);
-		mod_delayed_work_on(info->eoi_cpu, system_wq,
+		mod_delayed_work_on(info->eoi_cpu, system_percpu_wq,
 				    &eoi->delayed, delay);
 	} else {
 		list_for_each_entry_reverse(elem, &eoi->eoi_list, eoi_list) {
@@ -666,7 +666,7 @@ static void xen_irq_lateeoi_worker(struct work_struct *work)
 			break;
 
 		if (now < info->eoi_time) {
-			mod_delayed_work_on(info->eoi_cpu, system_wq,
+			mod_delayed_work_on(info->eoi_cpu, system_percpu_wq,
 					    &eoi->delayed,
 					    info->eoi_time - now);
 			break;
@@ -782,7 +782,7 @@ static void xen_free_irq(struct irq_info *info)
 
 	WARN_ON(info->refcnt > 0);
 
-	queue_rcu_work(system_wq, &info->rwork);
+	queue_rcu_work(system_percpu_wq, &info->rwork);
 }
 
 /* Not called for lateeoi events. */
diff --git a/drivers/xen/grant-dma-ops.c b/drivers/xen/grant-dma-ops.c
@@ -366,7 +366,8 @@ static int xen_grant_init_backend_domid(struct device *dev,
 	if (np) {
 		ret = xen_dt_grant_init_backend_domid(dev, np, backend_domid);
 		of_node_put(np);
-	} else if (IS_ENABLED(CONFIG_XEN_VIRTIO_FORCE_GRANT) || xen_pv_domain()) {
+	} else if (!xen_initial_domain() &&
+		   (IS_ENABLED(CONFIG_XEN_VIRTIO_FORCE_GRANT) || xen_pv_domain())) {
 		dev_info(dev, "Using dom0 as backend\n");
 		*backend_domid = 0;
 		ret = 0;
diff --git a/drivers/xen/mcelog.c b/drivers/xen/mcelog.c
@@ -165,9 +165,7 @@ static long xen_mce_chrdev_ioctl(struct file *f, unsigned int cmd,
 	case MCE_GETCLEAR_FLAGS: {
 		unsigned flags;
 
-		do {
-			flags = xen_mcelog.flags;
-		} while (cmpxchg(&xen_mcelog.flags, flags, 0) != flags);
+		flags = xchg(&xen_mcelog.flags, 0);
 
 		return put_user(flags, p);
 	}
diff --git a/drivers/xen/privcmd.c b/drivers/xen/privcmd.c
@@ -1091,7 +1091,8 @@ static long privcmd_ioctl_irqfd(struct file *file, void __user *udata)
 
 static int privcmd_irqfd_init(void)
 {
-	irqfd_cleanup_wq = alloc_workqueue("privcmd-irqfd-cleanup", 0, 0);
+	irqfd_cleanup_wq = alloc_workqueue("privcmd-irqfd-cleanup", WQ_PERCPU,
+					   0);
 	if (!irqfd_cleanup_wq)
 		return -ENOMEM;
 
diff --git a/drivers/xen/unpopulated-alloc.c b/drivers/xen/unpopulated-alloc.c
@@ -18,6 +18,9 @@ static unsigned int list_count;
 
 static struct resource *target_resource;
 
+/* Pages to subtract from the memory count when setting balloon target. */
+unsigned long xen_unpopulated_pages __initdata;
+
 /*
  * If arch is not happy with system "iomem_resource" being used for
  * the region allocation it can provide it's own view by creating specific
diff --git a/drivers/xen/xenbus/xenbus.h b/drivers/xen/xenbus/xenbus.h
@@ -120,9 +120,9 @@ int xenbus_probe_devices(struct xen_bus_type *bus);
 
 void xenbus_dev_changed(const char *node, struct xen_bus_type *bus);
 
-int xenbus_dev_suspend(struct device *dev);
-int xenbus_dev_resume(struct device *dev);
-int xenbus_dev_cancel(struct device *dev);
+int xenbus_dev_freeze(struct device *dev);
+int xenbus_dev_restore(struct device *dev);
+int xenbus_dev_thaw(struct device *dev);
 
 void xenbus_otherend_changed(struct xenbus_watch *watch,
 			     const char *path, const char *token,
diff --git a/drivers/xen/xenbus/xenbus_probe.c b/drivers/xen/xenbus/xenbus_probe.c
@@ -668,7 +668,7 @@ void xenbus_dev_changed(const char *node, struct xen_bus_type *bus)
 }
 EXPORT_SYMBOL_GPL(xenbus_dev_changed);
 
-int xenbus_dev_suspend(struct device *dev)
+int xenbus_dev_freeze(struct device *dev)
 {
 	int err = 0;
 	struct xenbus_driver *drv;
@@ -683,12 +683,12 @@ int xenbus_dev_suspend(struct device *dev)
 	if (drv->suspend)
 		err = drv->suspend(xdev);
 	if (err)
-		dev_warn(dev, "suspend failed: %i\n", err);
+		dev_warn(dev, "freeze failed: %i\n", err);
 	return 0;
 }
-EXPORT_SYMBOL_GPL(xenbus_dev_suspend);
+EXPORT_SYMBOL_GPL(xenbus_dev_freeze);
 
-int xenbus_dev_resume(struct device *dev)
+int xenbus_dev_restore(struct device *dev)
 {
 	int err;
 	struct xenbus_driver *drv;
@@ -702,7 +702,7 @@ int xenbus_dev_resume(struct device *dev)
 	drv = to_xenbus_driver(dev->driver);
 	err = talk_to_otherend(xdev);
 	if (err) {
-		dev_warn(dev, "resume (talk_to_otherend) failed: %i\n", err);
+		dev_warn(dev, "restore (talk_to_otherend) failed: %i\n", err);
 		return err;
 	}
 
@@ -711,28 +711,28 @@ int xenbus_dev_resume(struct device *dev)
 	if (drv->resume) {
 		err = drv->resume(xdev);
 		if (err) {
-			dev_warn(dev, "resume failed: %i\n", err);
+			dev_warn(dev, "restore failed: %i\n", err);
 			return err;
 		}
 	}
 
 	err = watch_otherend(xdev);
 	if (err) {
-		dev_warn(dev, "resume (watch_otherend) failed: %d\n", err);
+		dev_warn(dev, "restore (watch_otherend) failed: %d\n", err);
 		return err;
 	}
 
 	return 0;
 }
-EXPORT_SYMBOL_GPL(xenbus_dev_resume);
+EXPORT_SYMBOL_GPL(xenbus_dev_restore);
 
-int xenbus_dev_cancel(struct device *dev)
+int xenbus_dev_thaw(struct device *dev)
 {
 	/* Do nothing */
-	DPRINTK("cancel");
+	DPRINTK("thaw");
 	return 0;
 }
-EXPORT_SYMBOL_GPL(xenbus_dev_cancel);
+EXPORT_SYMBOL_GPL(xenbus_dev_thaw);
 
 /* A flag to determine if xenstored is 'ready' (i.e. has started) */
 int xenstored_ready;
diff --git a/drivers/xen/xenbus/xenbus_probe_frontend.c b/drivers/xen/xenbus/xenbus_probe_frontend.c
@@ -91,14 +91,14 @@ static void backend_changed(struct xenbus_watch *watch,
 	xenbus_otherend_changed(watch, path, token, 1);
 }
 
-static void xenbus_frontend_delayed_resume(struct work_struct *w)
+static void xenbus_frontend_delayed_restore(struct work_struct *w)
 {
 	struct xenbus_device *xdev = container_of(w, struct xenbus_device, work);
 
-	xenbus_dev_resume(&xdev->dev);
+	xenbus_dev_restore(&xdev->dev);
 }
 
-static int xenbus_frontend_dev_resume(struct device *dev)
+static int xenbus_frontend_dev_restore(struct device *dev)
 {
 	/*
 	 * If xenstored is running in this domain, we cannot access the backend
@@ -112,14 +112,14 @@ static int xenbus_frontend_dev_resume(struct device *dev)
 		return 0;
 	}
 
-	return xenbus_dev_resume(dev);
+	return xenbus_dev_restore(dev);
 }
 
 static int xenbus_frontend_dev_probe(struct device *dev)
 {
 	if (xen_store_domain_type == XS_LOCAL) {
 		struct xenbus_device *xdev = to_xenbus_device(dev);
-		INIT_WORK(&xdev->work, xenbus_frontend_delayed_resume);
+		INIT_WORK(&xdev->work, xenbus_frontend_delayed_restore);
 	}
 
 	return xenbus_dev_probe(dev);
@@ -148,11 +148,9 @@ static void xenbus_frontend_dev_shutdown(struct device *_dev)
 }
 
 static const struct dev_pm_ops xenbus_pm_ops = {
-	.suspend	= xenbus_dev_suspend,
-	.resume		= xenbus_frontend_dev_resume,
-	.freeze		= xenbus_dev_suspend,
-	.thaw		= xenbus_dev_cancel,
-	.restore	= xenbus_dev_resume,
+	.freeze		= xenbus_dev_freeze,
+	.thaw		= xenbus_dev_thaw,
+	.restore	= xenbus_frontend_dev_restore,
 };
 
 static struct xen_bus_type xenbus_frontend = {
diff --git a/include/xen/xen.h b/include/xen/xen.h
@@ -69,11 +69,13 @@ extern u64 xen_saved_max_mem_size;
 #endif
 
 #ifdef CONFIG_XEN_UNPOPULATED_ALLOC
+extern unsigned long xen_unpopulated_pages;
 int xen_alloc_unpopulated_pages(unsigned int nr_pages, struct page **pages);
 void xen_free_unpopulated_pages(unsigned int nr_pages, struct page **pages);
 #include <linux/ioport.h>
 int arch_xen_unpopulated_init(struct resource **res);
 #else
+#define xen_unpopulated_pages 0UL
 #include <xen/balloon.h>
 static inline int xen_alloc_unpopulated_pages(unsigned int nr_pages,
 		struct page **pages)

Original file line number	Diff line number	Diff line change
`@@ -470,7 +470,7 @@ int __init arch_xen_unpopulated_init(struct resource **res)`
`470`	`470`	`* driver to know how much of the physmap is unpopulated and`
`471`	`471`	`* set an accurate initial memory target.`
`472`	`472`	`*/`
`473`		`- xen_released_pages += xen_extra_mem[i].n_pfns;`
	`473`	`+ xen_unpopulated_pages += xen_extra_mem[i].n_pfns;`
`474`	`474`	`/* Zero so region is not also added to the balloon driver. */`
`475`	`475`	`xen_extra_mem[i].n_pfns = 0;`
`476`	`476`	`}`
Original file line number	Diff line number	Diff line change
`@@ -165,9 +165,7 @@ static long xen_mce_chrdev_ioctl(struct file *f, unsigned int cmd,`
`165`	`165`	`case MCE_GETCLEAR_FLAGS: {`
`166`	`166`	`unsigned flags;`
`167`	`167`
`168`		`- do {`
`169`		`- flags = xen_mcelog.flags;`
`170`		`- } while (cmpxchg(&xen_mcelog.flags, flags, 0) != flags);`
	`168`	`+ flags = xchg(&xen_mcelog.flags, 0);`
`171`	`169`
`172`	`170`	`return put_user(flags, p);`
`173`	`171`	`}`
Original file line number	Diff line number	Diff line change
`@@ -1091,7 +1091,8 @@ static long privcmd_ioctl_irqfd(struct file file, void __user udata)`
`1091`	`1091`
`1092`	`1092`	`static int privcmd_irqfd_init(void)`
`1093`	`1093`	`{`
`1094`		`- irqfd_cleanup_wq = alloc_workqueue("privcmd-irqfd-cleanup", 0, 0);`
	`1094`	`+ irqfd_cleanup_wq = alloc_workqueue("privcmd-irqfd-cleanup", WQ_PERCPU,`
	`1095`	`+ 0);`
`1095`	`1096`	`if (!irqfd_cleanup_wq)`
`1096`	`1097`	`return -ENOMEM;`
`1097`	`1098`