watchdog-kernel-api.txt 8.8 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174
  1. The Linux WatchDog Timer Driver Core kernel API.
  2. ===============================================
  3. Last reviewed: 16-Mar-2012
  4. Wim Van Sebroeck <wim@iguana.be>
  5. Introduction
  6. ------------
  7. This document does not describe what a WatchDog Timer (WDT) Driver or Device is.
  8. It also does not describe the API which can be used by user space to communicate
  9. with a WatchDog Timer. If you want to know this then please read the following
  10. file: Documentation/watchdog/watchdog-api.txt .
  11. So what does this document describe? It describes the API that can be used by
  12. WatchDog Timer Drivers that want to use the WatchDog Timer Driver Core
  13. Framework. This framework provides all interfacing towards user space so that
  14. the same code does not have to be reproduced each time. This also means that
  15. a watchdog timer driver then only needs to provide the different routines
  16. (operations) that control the watchdog timer (WDT).
  17. The API
  18. -------
  19. Each watchdog timer driver that wants to use the WatchDog Timer Driver Core
  20. must #include <linux/watchdog.h> (you would have to do this anyway when
  21. writing a watchdog device driver). This include file contains following
  22. register/unregister routines:
  23. extern int watchdog_register_device(struct watchdog_device *);
  24. extern void watchdog_unregister_device(struct watchdog_device *);
  25. The watchdog_register_device routine registers a watchdog timer device.
  26. The parameter of this routine is a pointer to a watchdog_device structure.
  27. This routine returns zero on success and a negative errno code for failure.
  28. The watchdog_unregister_device routine deregisters a registered watchdog timer
  29. device. The parameter of this routine is the pointer to the registered
  30. watchdog_device structure.
  31. The watchdog device structure looks like this:
  32. struct watchdog_device {
  33. const struct watchdog_info *info;
  34. const struct watchdog_ops *ops;
  35. unsigned int bootstatus;
  36. unsigned int timeout;
  37. unsigned int min_timeout;
  38. unsigned int max_timeout;
  39. void *driver_data;
  40. unsigned long status;
  41. };
  42. It contains following fields:
  43. * info: a pointer to a watchdog_info structure. This structure gives some
  44. additional information about the watchdog timer itself. (Like it's unique name)
  45. * ops: a pointer to the list of watchdog operations that the watchdog supports.
  46. * timeout: the watchdog timer's timeout value (in seconds).
  47. * min_timeout: the watchdog timer's minimum timeout value (in seconds).
  48. * max_timeout: the watchdog timer's maximum timeout value (in seconds).
  49. * bootstatus: status of the device after booting (reported with watchdog
  50. WDIOF_* status bits).
  51. * driver_data: a pointer to the drivers private data of a watchdog device.
  52. This data should only be accessed via the watchdog_set_drvadata and
  53. watchdog_get_drvdata routines.
  54. * status: this field contains a number of status bits that give extra
  55. information about the status of the device (Like: is the watchdog timer
  56. running/active, is the nowayout bit set, is the device opened via
  57. the /dev/watchdog interface or not, ...).
  58. The list of watchdog operations is defined as:
  59. struct watchdog_ops {
  60. struct module *owner;
  61. /* mandatory operations */
  62. int (*start)(struct watchdog_device *);
  63. int (*stop)(struct watchdog_device *);
  64. /* optional operations */
  65. int (*ping)(struct watchdog_device *);
  66. unsigned int (*status)(struct watchdog_device *);
  67. int (*set_timeout)(struct watchdog_device *, unsigned int);
  68. unsigned int (*get_timeleft)(struct watchdog_device *);
  69. long (*ioctl)(struct watchdog_device *, unsigned int, unsigned long);
  70. };
  71. It is important that you first define the module owner of the watchdog timer
  72. driver's operations. This module owner will be used to lock the module when
  73. the watchdog is active. (This to avoid a system crash when you unload the
  74. module and /dev/watchdog is still open).
  75. Some operations are mandatory and some are optional. The mandatory operations
  76. are:
  77. * start: this is a pointer to the routine that starts the watchdog timer
  78. device.
  79. The routine needs a pointer to the watchdog timer device structure as a
  80. parameter. It returns zero on success or a negative errno code for failure.
  81. * stop: with this routine the watchdog timer device is being stopped.
  82. The routine needs a pointer to the watchdog timer device structure as a
  83. parameter. It returns zero on success or a negative errno code for failure.
  84. Some watchdog timer hardware can only be started and not be stopped. The
  85. driver supporting this hardware needs to make sure that a start and stop
  86. routine is being provided. This can be done by using a timer in the driver
  87. that regularly sends a keepalive ping to the watchdog timer hardware.
  88. Not all watchdog timer hardware supports the same functionality. That's why
  89. all other routines/operations are optional. They only need to be provided if
  90. they are supported. These optional routines/operations are:
  91. * ping: this is the routine that sends a keepalive ping to the watchdog timer
  92. hardware.
  93. The routine needs a pointer to the watchdog timer device structure as a
  94. parameter. It returns zero on success or a negative errno code for failure.
  95. Most hardware that does not support this as a separate function uses the
  96. start function to restart the watchdog timer hardware. And that's also what
  97. the watchdog timer driver core does: to send a keepalive ping to the watchdog
  98. timer hardware it will either use the ping operation (when available) or the
  99. start operation (when the ping operation is not available).
  100. (Note: the WDIOC_KEEPALIVE ioctl call will only be active when the
  101. WDIOF_KEEPALIVEPING bit has been set in the option field on the watchdog's
  102. info structure).
  103. * status: this routine checks the status of the watchdog timer device. The
  104. status of the device is reported with watchdog WDIOF_* status flags/bits.
  105. * set_timeout: this routine checks and changes the timeout of the watchdog
  106. timer device. It returns 0 on success, -EINVAL for "parameter out of range"
  107. and -EIO for "could not write value to the watchdog". On success this
  108. routine should set the timeout value of the watchdog_device to the
  109. achieved timeout value (which may be different from the requested one
  110. because the watchdog does not necessarily has a 1 second resolution).
  111. (Note: the WDIOF_SETTIMEOUT needs to be set in the options field of the
  112. watchdog's info structure).
  113. * get_timeleft: this routines returns the time that's left before a reset.
  114. * ioctl: if this routine is present then it will be called first before we do
  115. our own internal ioctl call handling. This routine should return -ENOIOCTLCMD
  116. if a command is not supported. The parameters that are passed to the ioctl
  117. call are: watchdog_device, cmd and arg.
  118. The status bits should (preferably) be set with the set_bit and clear_bit alike
  119. bit-operations. The status bits that are defined are:
  120. * WDOG_ACTIVE: this status bit indicates whether or not a watchdog timer device
  121. is active or not. When the watchdog is active after booting, then you should
  122. set this status bit (Note: when you register the watchdog timer device with
  123. this bit set, then opening /dev/watchdog will skip the start operation)
  124. * WDOG_DEV_OPEN: this status bit shows whether or not the watchdog device
  125. was opened via /dev/watchdog.
  126. (This bit should only be used by the WatchDog Timer Driver Core).
  127. * WDOG_ALLOW_RELEASE: this bit stores whether or not the magic close character
  128. has been sent (so that we can support the magic close feature).
  129. (This bit should only be used by the WatchDog Timer Driver Core).
  130. * WDOG_NO_WAY_OUT: this bit stores the nowayout setting for the watchdog.
  131. If this bit is set then the watchdog timer will not be able to stop.
  132. To set the WDOG_NO_WAY_OUT status bit (before registering your watchdog
  133. timer device) you can either:
  134. * set it statically in your watchdog_device struct with
  135. .status = WATCHDOG_NOWAYOUT_INIT_STATUS,
  136. (this will set the value the same as CONFIG_WATCHDOG_NOWAYOUT) or
  137. * use the following helper function:
  138. static inline void watchdog_set_nowayout(struct watchdog_device *wdd, int nowayout)
  139. Note: The WatchDog Timer Driver Core supports the magic close feature and
  140. the nowayout feature. To use the magic close feature you must set the
  141. WDIOF_MAGICCLOSE bit in the options field of the watchdog's info structure.
  142. The nowayout feature will overrule the magic close feature.
  143. To get or set driver specific data the following two helper functions should be
  144. used:
  145. static inline void watchdog_set_drvdata(struct watchdog_device *wdd, void *data)
  146. static inline void *watchdog_get_drvdata(struct watchdog_device *wdd)
  147. The watchdog_set_drvdata function allows you to add driver specific data. The
  148. arguments of this function are the watchdog device where you want to add the
  149. driver specific data to and a pointer to the data itself.
  150. The watchdog_get_drvdata function allows you to retrieve driver specific data.
  151. The argument of this function is the watchdog device where you want to retrieve
  152. data from. The function returns the pointer to the driver specific data.