1The Linux WatchDog Timer Driver Core kernel API. 2=============================================== 3Last reviewed: 12-Feb-2013 4 5Wim Van Sebroeck <wim@iguana.be> 6 7Introduction 8------------ 9This document does not describe what a WatchDog Timer (WDT) Driver or Device is. 10It also does not describe the API which can be used by user space to communicate 11with a WatchDog Timer. If you want to know this then please read the following 12file: Documentation/watchdog/watchdog-api.txt . 13 14So what does this document describe? It describes the API that can be used by 15WatchDog Timer Drivers that want to use the WatchDog Timer Driver Core 16Framework. This framework provides all interfacing towards user space so that 17the same code does not have to be reproduced each time. This also means that 18a watchdog timer driver then only needs to provide the different routines 19(operations) that control the watchdog timer (WDT). 20 21The API 22------- 23Each watchdog timer driver that wants to use the WatchDog Timer Driver Core 24must #include <linux/watchdog.h> (you would have to do this anyway when 25writing a watchdog device driver). This include file contains following 26register/unregister routines: 27 28extern int watchdog_register_device(struct watchdog_device *); 29extern void watchdog_unregister_device(struct watchdog_device *); 30 31The watchdog_register_device routine registers a watchdog timer device. 32The parameter of this routine is a pointer to a watchdog_device structure. 33This routine returns zero on success and a negative errno code for failure. 34 35The watchdog_unregister_device routine deregisters a registered watchdog timer 36device. The parameter of this routine is the pointer to the registered 37watchdog_device structure. 38 39The watchdog device structure looks like this: 40 41struct watchdog_device { 42 int id; 43 struct cdev cdev; 44 struct device *dev; 45 struct device *parent; 46 const struct watchdog_info *info; 47 const struct watchdog_ops *ops; 48 unsigned int bootstatus; 49 unsigned int timeout; 50 unsigned int min_timeout; 51 unsigned int max_timeout; 52 void *driver_data; 53 struct mutex lock; 54 unsigned long status; 55}; 56 57It contains following fields: 58* id: set by watchdog_register_device, id 0 is special. It has both a 59 /dev/watchdog0 cdev (dynamic major, minor 0) as well as the old 60 /dev/watchdog miscdev. The id is set automatically when calling 61 watchdog_register_device. 62* cdev: cdev for the dynamic /dev/watchdog<id> device nodes. This 63 field is also populated by watchdog_register_device. 64* dev: device under the watchdog class (created by watchdog_register_device). 65* parent: set this to the parent device (or NULL) before calling 66 watchdog_register_device. 67* info: a pointer to a watchdog_info structure. This structure gives some 68 additional information about the watchdog timer itself. (Like it's unique name) 69* ops: a pointer to the list of watchdog operations that the watchdog supports. 70* timeout: the watchdog timer's timeout value (in seconds). 71* min_timeout: the watchdog timer's minimum timeout value (in seconds). 72* max_timeout: the watchdog timer's maximum timeout value (in seconds). 73* bootstatus: status of the device after booting (reported with watchdog 74 WDIOF_* status bits). 75* driver_data: a pointer to the drivers private data of a watchdog device. 76 This data should only be accessed via the watchdog_set_drvdata and 77 watchdog_get_drvdata routines. 78* lock: Mutex for WatchDog Timer Driver Core internal use only. 79* status: this field contains a number of status bits that give extra 80 information about the status of the device (Like: is the watchdog timer 81 running/active, is the nowayout bit set, is the device opened via 82 the /dev/watchdog interface or not, ...). 83 84The list of watchdog operations is defined as: 85 86struct watchdog_ops { 87 struct module *owner; 88 /* mandatory operations */ 89 int (*start)(struct watchdog_device *); 90 int (*stop)(struct watchdog_device *); 91 /* optional operations */ 92 int (*ping)(struct watchdog_device *); 93 unsigned int (*status)(struct watchdog_device *); 94 int (*set_timeout)(struct watchdog_device *, unsigned int); 95 unsigned int (*get_timeleft)(struct watchdog_device *); 96 void (*ref)(struct watchdog_device *); 97 void (*unref)(struct watchdog_device *); 98 long (*ioctl)(struct watchdog_device *, unsigned int, unsigned long); 99}; 100 101It is important that you first define the module owner of the watchdog timer 102driver's operations. This module owner will be used to lock the module when 103the watchdog is active. (This to avoid a system crash when you unload the 104module and /dev/watchdog is still open). 105 106If the watchdog_device struct is dynamically allocated, just locking the module 107is not enough and a driver also needs to define the ref and unref operations to 108ensure the structure holding the watchdog_device does not go away. 109 110The simplest (and usually sufficient) implementation of this is to: 1111) Add a kref struct to the same structure which is holding the watchdog_device 1122) Define a release callback for the kref which frees the struct holding both 1133) Call kref_init on this kref *before* calling watchdog_register_device() 1144) Define a ref operation calling kref_get on this kref 1155) Define a unref operation calling kref_put on this kref 1166) When it is time to cleanup: 117 * Do not kfree() the struct holding both, the last kref_put will do this! 118 * *After* calling watchdog_unregister_device() call kref_put on the kref 119 120Some operations are mandatory and some are optional. The mandatory operations 121are: 122* start: this is a pointer to the routine that starts the watchdog timer 123 device. 124 The routine needs a pointer to the watchdog timer device structure as a 125 parameter. It returns zero on success or a negative errno code for failure. 126* stop: with this routine the watchdog timer device is being stopped. 127 The routine needs a pointer to the watchdog timer device structure as a 128 parameter. It returns zero on success or a negative errno code for failure. 129 Some watchdog timer hardware can only be started and not be stopped. The 130 driver supporting this hardware needs to make sure that a start and stop 131 routine is being provided. This can be done by using a timer in the driver 132 that regularly sends a keepalive ping to the watchdog timer hardware. 133 134Not all watchdog timer hardware supports the same functionality. That's why 135all other routines/operations are optional. They only need to be provided if 136they are supported. These optional routines/operations are: 137* ping: this is the routine that sends a keepalive ping to the watchdog timer 138 hardware. 139 The routine needs a pointer to the watchdog timer device structure as a 140 parameter. It returns zero on success or a negative errno code for failure. 141 Most hardware that does not support this as a separate function uses the 142 start function to restart the watchdog timer hardware. And that's also what 143 the watchdog timer driver core does: to send a keepalive ping to the watchdog 144 timer hardware it will either use the ping operation (when available) or the 145 start operation (when the ping operation is not available). 146 (Note: the WDIOC_KEEPALIVE ioctl call will only be active when the 147 WDIOF_KEEPALIVEPING bit has been set in the option field on the watchdog's 148 info structure). 149* status: this routine checks the status of the watchdog timer device. The 150 status of the device is reported with watchdog WDIOF_* status flags/bits. 151* set_timeout: this routine checks and changes the timeout of the watchdog 152 timer device. It returns 0 on success, -EINVAL for "parameter out of range" 153 and -EIO for "could not write value to the watchdog". On success this 154 routine should set the timeout value of the watchdog_device to the 155 achieved timeout value (which may be different from the requested one 156 because the watchdog does not necessarily has a 1 second resolution). 157 (Note: the WDIOF_SETTIMEOUT needs to be set in the options field of the 158 watchdog's info structure). 159* get_timeleft: this routines returns the time that's left before a reset. 160* ref: the operation that calls kref_get on the kref of a dynamically 161 allocated watchdog_device struct. 162* unref: the operation that calls kref_put on the kref of a dynamically 163 allocated watchdog_device struct. 164* ioctl: if this routine is present then it will be called first before we do 165 our own internal ioctl call handling. This routine should return -ENOIOCTLCMD 166 if a command is not supported. The parameters that are passed to the ioctl 167 call are: watchdog_device, cmd and arg. 168 169The status bits should (preferably) be set with the set_bit and clear_bit alike 170bit-operations. The status bits that are defined are: 171* WDOG_ACTIVE: this status bit indicates whether or not a watchdog timer device 172 is active or not. When the watchdog is active after booting, then you should 173 set this status bit (Note: when you register the watchdog timer device with 174 this bit set, then opening /dev/watchdog will skip the start operation) 175* WDOG_DEV_OPEN: this status bit shows whether or not the watchdog device 176 was opened via /dev/watchdog. 177 (This bit should only be used by the WatchDog Timer Driver Core). 178* WDOG_ALLOW_RELEASE: this bit stores whether or not the magic close character 179 has been sent (so that we can support the magic close feature). 180 (This bit should only be used by the WatchDog Timer Driver Core). 181* WDOG_NO_WAY_OUT: this bit stores the nowayout setting for the watchdog. 182 If this bit is set then the watchdog timer will not be able to stop. 183* WDOG_UNREGISTERED: this bit gets set by the WatchDog Timer Driver Core 184 after calling watchdog_unregister_device, and then checked before calling 185 any watchdog_ops, so that you can be sure that no operations (other then 186 unref) will get called after unregister, even if userspace still holds a 187 reference to /dev/watchdog 188 189 To set the WDOG_NO_WAY_OUT status bit (before registering your watchdog 190 timer device) you can either: 191 * set it statically in your watchdog_device struct with 192 .status = WATCHDOG_NOWAYOUT_INIT_STATUS, 193 (this will set the value the same as CONFIG_WATCHDOG_NOWAYOUT) or 194 * use the following helper function: 195 static inline void watchdog_set_nowayout(struct watchdog_device *wdd, int nowayout) 196 197Note: The WatchDog Timer Driver Core supports the magic close feature and 198the nowayout feature. To use the magic close feature you must set the 199WDIOF_MAGICCLOSE bit in the options field of the watchdog's info structure. 200The nowayout feature will overrule the magic close feature. 201 202To get or set driver specific data the following two helper functions should be 203used: 204 205static inline void watchdog_set_drvdata(struct watchdog_device *wdd, void *data) 206static inline void *watchdog_get_drvdata(struct watchdog_device *wdd) 207 208The watchdog_set_drvdata function allows you to add driver specific data. The 209arguments of this function are the watchdog device where you want to add the 210driver specific data to and a pointer to the data itself. 211 212The watchdog_get_drvdata function allows you to retrieve driver specific data. 213The argument of this function is the watchdog device where you want to retrieve 214data from. The function returns the pointer to the driver specific data. 215 216To initialize the timeout field, the following function can be used: 217 218extern int watchdog_init_timeout(struct watchdog_device *wdd, 219 unsigned int timeout_parm, struct device *dev); 220 221The watchdog_init_timeout function allows you to initialize the timeout field 222using the module timeout parameter or by retrieving the timeout-sec property from 223the device tree (if the module timeout parameter is invalid). Best practice is 224to set the default timeout value as timeout value in the watchdog_device and 225then use this function to set the user "preferred" timeout value. 226This routine returns zero on success and a negative errno code for failure. 227