watchdog: Xen watchdog driver

While the hypervisor change adding SCHEDOP_watchdog support included a
daemon to make use of the new functionality, having a kernel driver
for /dev/watchdog so that user space code doesn't need to distinguish
non-Xen and Xen seems to be preferable.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>

drivers/watchdog/Kconfig

@@ -1137,6 +1137,16 @@ config WATCHDOG_RIO
 
 # XTENSA Architecture
 
+# Xen Architecture
+
+config XEN_WDT
+	tristate "Xen Watchdog support"
+	depends on XEN
+	help
+	  Say Y here to support the hypervisor watchdog capability provided
+	  by Xen 4.0 and newer.  The watchdog timeout period is normally one
+	  minute but can be changed with a boot-time parameter.
+
 #
 # ISA-based Watchdog Cards
 #

drivers/watchdog/Makefile

@@ -150,6 +150,9 @@ obj-$(CONFIG_WATCHDOG_CP1XXX)		+= cpwd.o
 
 # XTENSA Architecture
 
+# Xen
+obj-$(CONFIG_XEN_WDT) += xen_wdt.o
+
 # Architecture Independant
 obj-$(CONFIG_WM831X_WATCHDOG) += wm831x_wdt.o
 obj-$(CONFIG_WM8350_WATCHDOG) += wm8350_wdt.o

drivers/watchdog/xen_wdt.c

+/*
+ *	Xen Watchdog Driver
+ *
+ *	(c) Copyright 2010 Novell, Inc.
+ *
+ *	This program is free software; you can redistribute it and/or
+ *	modify it under the terms of the GNU General Public License
+ *	as published by the Free Software Foundation; either version
+ *	2 of the License, or (at your option) any later version.
+ */
+
+#define DRV_NAME	"wdt"
+#define DRV_VERSION	"0.01"
+#define PFX		DRV_NAME ": "
+
+#include <linux/bug.h>
+#include <linux/errno.h>
+#include <linux/fs.h>
+#include <linux/hrtimer.h>
+#include <linux/kernel.h>
+#include <linux/ktime.h>
+#include <linux/init.h>
+#include <linux/miscdevice.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/platform_device.h>
+#include <linux/spinlock.h>
+#include <linux/uaccess.h>
+#include <linux/watchdog.h>
+#include <xen/xen.h>
+#include <asm/xen/hypercall.h>
+#include <xen/interface/sched.h>
+
+static struct platform_device *platform_device;
+static DEFINE_SPINLOCK(wdt_lock);
+static struct sched_watchdog wdt;
+static __kernel_time_t wdt_expires;
+static bool is_active, expect_release;
+
+#define WATCHDOG_TIMEOUT 60 /* in seconds */
+static unsigned int timeout = WATCHDOG_TIMEOUT;
+module_param(timeout, uint, S_IRUGO);
+MODULE_PARM_DESC(timeout, "Watchdog timeout in seconds "
+	"(default=" __MODULE_STRING(WATCHDOG_TIMEOUT) ")");
+
+static bool nowayout = WATCHDOG_NOWAYOUT;
+module_param(nowayout, bool, S_IRUGO);
+MODULE_PARM_DESC(nowayout, "Watchdog cannot be stopped once started "
+	"(default=" __MODULE_STRING(WATCHDOG_NOWAYOUT) ")");
+
+static inline __kernel_time_t set_timeout(void)
+{
+	wdt.timeout = timeout;
+	return ktime_to_timespec(ktime_get()).tv_sec + timeout;
+}
+
+static int xen_wdt_start(void)
+{
+	__kernel_time_t expires;
+	int err;
+
+	spin_lock(&wdt_lock);
+
+	expires = set_timeout();
+	if (!wdt.id)
+		err = HYPERVISOR_sched_op(SCHEDOP_watchdog, &wdt);
+	else
+		err = -EBUSY;
+	if (err > 0) {
+		wdt.id = err;
+		wdt_expires = expires;
+		err = 0;
+	} else
+		BUG_ON(!err);
+
+	spin_unlock(&wdt_lock);
+
+	return err;
+}
+
+static int xen_wdt_stop(void)
+{
+	int err = 0;
+
+	spin_lock(&wdt_lock);
+
+	wdt.timeout = 0;
+	if (wdt.id)
+		err = HYPERVISOR_sched_op(SCHEDOP_watchdog, &wdt);
+	if (!err)
+		wdt.id = 0;
+
+	spin_unlock(&wdt_lock);
+
+	return err;
+}
+
+static int xen_wdt_kick(void)
+{
+	__kernel_time_t expires;
+	int err;
+
+	spin_lock(&wdt_lock);
+
+	expires = set_timeout();
+	if (wdt.id)
+		err = HYPERVISOR_sched_op(SCHEDOP_watchdog, &wdt);
+	else
+		err = -ENXIO;
+	if (!err)
+		wdt_expires = expires;
+
+	spin_unlock(&wdt_lock);
+
+	return err;
+}
+
+static int xen_wdt_open(struct inode *inode, struct file *file)
+{
+	int err;
+
+	/* /dev/watchdog can only be opened once */
+	if (xchg(&is_active, true))
+		return -EBUSY;
+
+	err = xen_wdt_start();
+	if (err == -EBUSY)
+		err = xen_wdt_kick();
+	return err ?: nonseekable_open(inode, file);
+}
+
+static int xen_wdt_release(struct inode *inode, struct file *file)
+{
+	if (expect_release)
+		xen_wdt_stop();
+	else {
+		printk(KERN_CRIT PFX
+		       "unexpected close, not stopping watchdog!\n");
+		xen_wdt_kick();
+	}
+	is_active = false;
+	expect_release = false;
+	return 0;
+}
+
+static ssize_t xen_wdt_write(struct file *file, const char __user *data,
+			     size_t len, loff_t *ppos)
+{
+	/* See if we got the magic character 'V' and reload the timer */
+	if (len) {
+		if (!nowayout) {
+			size_t i;
+
+			/* in case it was set long ago */
+			expect_release = false;
+
+			/* scan to see whether or not we got the magic
+			   character */
+			for (i = 0; i != len; i++) {
+				char c;
+				if (get_user(c, data + i))
+					return -EFAULT;
+				if (c == 'V')
+					expect_release = true;
+			}
+		}
+
+		/* someone wrote to us, we should reload the timer */
+		xen_wdt_kick();
+	}
+	return len;
+}
+
+static long xen_wdt_ioctl(struct file *file, unsigned int cmd,
+			  unsigned long arg)
+{
+	int new_options, retval = -EINVAL;
+	int new_timeout;
+	int __user *argp = (void __user *)arg;
+	static const struct watchdog_info ident = {
+		.options =		WDIOF_SETTIMEOUT | WDIOF_MAGICCLOSE,
+		.firmware_version =	0,
+		.identity =		DRV_NAME,
+	};
+
+	switch (cmd) {
+	case WDIOC_GETSUPPORT:
+		return copy_to_user(argp, &ident, sizeof(ident)) ? -EFAULT : 0;
+
+	case WDIOC_GETSTATUS:
+	case WDIOC_GETBOOTSTATUS:
+		return put_user(0, argp);
+
+	case WDIOC_SETOPTIONS:
+		if (get_user(new_options, argp))
+			return -EFAULT;
+
+		if (new_options & WDIOS_DISABLECARD)
+			retval = xen_wdt_stop();
+		if (new_options & WDIOS_ENABLECARD) {
+			retval = xen_wdt_start();
+			if (retval == -EBUSY)
+				retval = xen_wdt_kick();
+		}
+		return retval;
+
+	case WDIOC_KEEPALIVE:
+		xen_wdt_kick();
+		return 0;
+
+	case WDIOC_SETTIMEOUT:
+		if (get_user(new_timeout, argp))
+			return -EFAULT;
+		if (!new_timeout)
+			return -EINVAL;
+		timeout = new_timeout;
+		xen_wdt_kick();
+		/* fall through */
+	case WDIOC_GETTIMEOUT:
+		return put_user(timeout, argp);
+
+	case WDIOC_GETTIMELEFT:
+		retval = wdt_expires - ktime_to_timespec(ktime_get()).tv_sec;
+		return put_user(retval, argp);
+	}
+
+	return -ENOTTY;
+}
+
+static const struct file_operations xen_wdt_fops = {
+	.owner =		THIS_MODULE,
+	.llseek =		no_llseek,
+	.write =		xen_wdt_write,
+	.unlocked_ioctl =	xen_wdt_ioctl,
+	.open =			xen_wdt_open,
+	.release =		xen_wdt_release,
+};
+
+static struct miscdevice xen_wdt_miscdev = {
+	.minor =	WATCHDOG_MINOR,
+	.name =		"watchdog",
+	.fops =		&xen_wdt_fops,
+};
+
+static int __devinit xen_wdt_probe(struct platform_device *dev)
+{
+	struct sched_watchdog wd = { .id = ~0 };
+	int ret = HYPERVISOR_sched_op(SCHEDOP_watchdog, &wd);
+
+	switch (ret) {
+	case -EINVAL:
+		if (!timeout) {
+			timeout = WATCHDOG_TIMEOUT;
+			printk(KERN_INFO PFX
+			       "timeout value invalid, using %d\n", timeout);
+		}
+
+		ret = misc_register(&xen_wdt_miscdev);
+		if (ret) {
+			printk(KERN_ERR PFX
+			       "cannot register miscdev on minor=%d (%d)\n",
+			       WATCHDOG_MINOR, ret);
+			break;
+		}
+
+		printk(KERN_INFO PFX
+		       "initialized (timeout=%ds, nowayout=%d)\n",
+		       timeout, nowayout);
+		break;
+
+	case -ENOSYS:
+		printk(KERN_INFO PFX "not supported\n");
+		ret = -ENODEV;
+		break;
+
+	default:
+		printk(KERN_INFO PFX "bogus return value %d\n", ret);
+		break;
+	}
+
+	return ret;
+}
+
+static int __devexit xen_wdt_remove(struct platform_device *dev)
+{
+	/* Stop the timer before we leave */
+	if (!nowayout)
+		xen_wdt_stop();
+
+	misc_deregister(&xen_wdt_miscdev);
+
+	return 0;
+}
+
+static void xen_wdt_shutdown(struct platform_device *dev)
+{
+	xen_wdt_stop();
+}
+
+static int xen_wdt_suspend(struct platform_device *dev, pm_message_t state)
+{
+	return xen_wdt_stop();
+}
+
+static int xen_wdt_resume(struct platform_device *dev)
+{
+	return xen_wdt_start();
+}
+
+static struct platform_driver xen_wdt_driver = {
+	.probe          = xen_wdt_probe,
+	.remove         = __devexit_p(xen_wdt_remove),
+	.shutdown       = xen_wdt_shutdown,
+	.suspend        = xen_wdt_suspend,
+	.resume         = xen_wdt_resume,
+	.driver         = {
+		.owner  = THIS_MODULE,
+		.name   = DRV_NAME,
+	},
+};
+
+static int __init xen_wdt_init_module(void)
+{
+	int err;
+
+	if (!xen_domain())
+		return -ENODEV;
+
+	printk(KERN_INFO PFX "Xen WatchDog Timer Driver v%s\n", DRV_VERSION);
+
+	err = platform_driver_register(&xen_wdt_driver);
+	if (err)
+		return err;
+
+	platform_device = platform_device_register_simple(DRV_NAME,
+								  -1, NULL, 0);
+	if (IS_ERR(platform_device)) {
+		err = PTR_ERR(platform_device);
+		platform_driver_unregister(&xen_wdt_driver);
+	}
+
+	return err;
+}
+
+static void __exit xen_wdt_cleanup_module(void)
+{
+	platform_device_unregister(platform_device);
+	platform_driver_unregister(&xen_wdt_driver);
+	printk(KERN_INFO PFX "module unloaded\n");
+}
+
+module_init(xen_wdt_init_module);
+module_exit(xen_wdt_cleanup_module);
+
+MODULE_AUTHOR("Jan Beulich <jbeulich@novell.com>");
+MODULE_DESCRIPTION("Xen WatchDog Timer Driver");
+MODULE_VERSION(DRV_VERSION);
+MODULE_LICENSE("GPL");
+MODULE_ALIAS_MISCDEV(WATCHDOG_MINOR);

include/xen/interface/sched.h

@@ -64,6 +64,39 @@ struct sched_poll {
 };
 DEFINE_GUEST_HANDLE_STRUCT(sched_poll);
 
+/*
+ * Declare a shutdown for another domain. The main use of this function is
+ * in interpreting shutdown requests and reasons for fully-virtualized
+ * domains.  A para-virtualized domain may use SCHEDOP_shutdown directly.
+ * @arg == pointer to sched_remote_shutdown structure.
+ */
+#define SCHEDOP_remote_shutdown        4
+struct sched_remote_shutdown {
+    domid_t domain_id;         /* Remote domain ID */
+    unsigned int reason;       /* SHUTDOWN_xxx reason */
+};
+
+/*
+ * Latch a shutdown code, so that when the domain later shuts down it
+ * reports this code to the control tools.
+ * @arg == as for SCHEDOP_shutdown.
+ */
+#define SCHEDOP_shutdown_code 5
+
+/*
+ * Setup, poke and destroy a domain watchdog timer.
+ * @arg == pointer to sched_watchdog structure.
+ * With id == 0, setup a domain watchdog timer to cause domain shutdown
+ *               after timeout, returns watchdog id.
+ * With id != 0 and timeout == 0, destroy domain watchdog timer.
+ * With id != 0 and timeout != 0, poke watchdog timer and set new timeout.
+ */
+#define SCHEDOP_watchdog    6
+struct sched_watchdog {
+    uint32_t id;                /* watchdog ID */
+    uint32_t timeout;           /* timeout */
+};
+
 /*
  * Reason codes for SCHEDOP_shutdown. These may be interpreted by control
  * software to determine the appropriate action. For the most part, Xen does
@@ -73,5 +106,6 @@ DEFINE_GUEST_HANDLE_STRUCT(sched_poll);
 #define SHUTDOWN_reboot     1  /* Clean up, kill, and then restart.          */
 #define SHUTDOWN_suspend    2  /* Clean up, save suspend info, kill.         */
 #define SHUTDOWN_crash      3  /* Tell controller we've crashed.             */
+#define SHUTDOWN_watchdog   4  /* Restart because watchdog time expired.     */
 
 #endif /* __XEN_PUBLIC_SCHED_H__ */