1 The Linux Watchdog driver API.
3 Copyright 2002 Christer Weingel <wingel@nano-system.com>
5 Some parts of this document are copied verbatim from the sbc60xxwdt
6 driver which is (c) Copyright 2000 Jakob Oestergaard <jakob@ostenfeld.dk>
8 This document describes the state of the Linux 2.4.18 kernel.
12 A Watchdog Timer (WDT) is a hardware circuit that can reset the
13 computer system in case of a software fault. You probably knew that
16 Usually a userspace daemon will notify the kernel watchdog driver via the
17 /dev/watchdog special device file that userspace is still alive, at
18 regular intervals. When such a notification occurs, the driver will
19 usually tell the hardware watchdog that everything is in order, and
20 that the watchdog should wait for yet another little while to reset
21 the system. If userspace fails (RAM error, kernel bug, whatever), the
22 notifications cease to occur, and the hardware watchdog will reset the
23 system (causing a reboot) after the timeout occurs.
25 The Linux watchdog API is a rather AD hoc construction and different
26 drivers implement different, and sometimes incompatible, parts of it.
27 This file is an attempt to document the existing usage and allow
28 future driver writers to use it as a reference.
32 All drivers support the basic mode of operation, where the watchdog
33 activates as soon as /dev/watchdog is opened and will reboot unless
34 the watchdog is pinged within a certain time, this time is called the
35 timeout or margin. The simplest way to ping the watchdog is to write
36 some data to the device. So a very simple watchdog daemon would look
39 int main(int argc, const char *argv[]) {
40 int fd=open("/dev/watchdog",O_WRONLY);
51 A more advanced driver could for example check that a HTTP server is
52 still responding before doing the write call to ping the watchdog.
54 When the device is closed, the watchdog is disabled. This is not
55 always such a good idea, since if there is a bug in the watchdog
56 daemon and it crashes the system will not reboot. Because of this,
57 some of the drivers support the configuration option "Disable watchdog
58 shutdown on close", CONFIG_WATCHDOG_NOWAYOUT. If it is set to Y when
59 compiling the kernel, there is no way of disabling the watchdog once
60 it has been started. So, if the watchdog dameon crashes, the system
61 will reboot after the timeout has passed.
63 Some other drivers will not disable the watchdog, unless a specific
64 magic character 'V' has been sent /dev/watchdog just before closing
65 the file. If the userspace daemon closes the file without sending
66 this special character, the driver will assume that the daemon (and
67 userspace in general) died, and will stop pinging the watchdog without
68 disabling it first. This will then cause a reboot.
72 All conforming drivers also support an ioctl API.
74 Pinging the watchdog using an ioctl:
76 All drivers that have an ioctl interface support at least one ioctl,
77 KEEPALIVE. This ioctl does exactly the same thing as a write to the
78 watchdog device, so the main loop in the above program could be
82 ioctl(fd, WDIOC_KEEPALIVE, 0);
86 the argument to the ioctl is ignored.
88 Setting and getting the timeout:
90 For some drivers it is possible to modify the watchdog timeout on the
91 fly with the SETTIMEOUT ioctl, those drivers have the WDIOF_SETTIMEOUT
92 flag set in their option field. The argument is an integer
93 representing the timeout in seconds. The driver returns the real
94 timeout used in the same variable, and this timeout might differ from
95 the requested one due to limitation of the hardware.
98 ioctl(fd, WDIOC_SETTIMEOUT, &timeout);
99 printf("The timeout was set to %d seconds\n", timeout);
101 This example might actually print "The timeout was set to 60 seconds"
102 if the device has a granularity of minutes for its timeout.
104 Starting with the Linux 2.4.18 kernel, it is possible to query the
105 current timeout using the GETTIMEOUT ioctl.
107 ioctl(fd, WDIOC_GETTIMEOUT, &timeout);
108 printf("The timeout was is %d seconds\n", timeout);
110 Envinronmental monitoring:
112 All watchdog drivers are required return more information about the system,
113 some do temperature, fan and power level monitoring, some can tell you
114 the reason for the last reboot of the system. The GETSUPPORT ioctl is
115 available to ask what the device can do:
117 struct watchdog_info ident;
118 ioctl(fd, WDIOC_GETSUPPORT, &ident);
120 the fields returned in the ident struct are:
122 identity a string identifying the watchdog driver
123 firmware_version the firmware version of the card if available
124 options a flags describing what the device supports
126 the options field can have the following bits set, and describes what
127 kind of information that the GET_STATUS and GET_BOOT_STATUS ioctls can
128 return. [FIXME -- Is this correct?]
130 WDIOF_OVERHEAT Reset due to CPU overheat
132 The machine was last rebooted by the watchdog because the thermal limit was
135 WDIOF_FANFAULT Fan failed
137 A system fan monitored by the watchdog card has failed
139 WDIOF_EXTERN1 External relay 1
141 External monitoring relay/source 1 was triggered. Controllers intended for
142 real world applications include external monitoring pins that will trigger
145 WDIOF_EXTERN2 External relay 2
147 External monitoring relay/source 2 was triggered
149 WDIOF_POWERUNDER Power bad/power fault
151 The machine is showing an undervoltage status
153 WDIOF_CARDRESET Card previously reset the CPU
155 The last reboot was caused by the watchdog card
157 WDIOF_POWEROVER Power over voltage
159 The machine is showing an overvoltage status. Note that if one level is
160 under and one over both bits will be set - this may seem odd but makes
163 WDIOF_KEEPALIVEPING Keep alive ping reply
165 The watchdog saw a keepalive ping since it was last queried.
167 WDIOF_SETTIMEOUT Can set/get the timeout
170 For those drivers that return any bits set in the option field, the
171 GETSTATUS and GETBOOTSTATUS ioctls can be used to ask for the current
172 status, and the status at the last reboot, respectively.
175 ioctl(fd, WDIOC_GETSTATUS, &flags);
179 ioctl(fd, WDIOC_GETBOOTSTATUS, &flags);
181 Note that not all devices support these two calls, and some only
182 support the GETBOOTSTATUS call.
184 Some drivers can measure the temperature using the GETTEMP ioctl. The
185 returned value is the temperature in degrees farenheit.
188 ioctl(fd, WDIOC_GETTEMP, &temperature);
190 Finally the SETOPTIONS ioctl can be used to control some aspects of
191 the cards operation; right now the pcwd driver is the only one
192 supporting thiss ioctl.
195 ioctl(fd, WDIOC_SETOPTIONS, options);
197 The following options are available:
199 WDIOS_DISABLECARD Turn off the watchdog timer
200 WDIOS_ENABLECARD Turn on the watchdog timer
201 WDIOS_TEMPPANIC Kernel panic on temperature trip
203 [FIXME -- better explanations]
205 Implementations in the current drivers in the kernel tree:
207 Here I have tried to summarize what the different drivers support and
208 where they do strange things compared to the other drivers.
210 acquirewdt.c -- Acquire Single Board Computer
212 This driver has a hardcoded timeout of 1 minute
214 Supports CONFIG_WATCHDOG_NOWAYOUT
216 GETSUPPORT returns KEEPALIVEPING. GETSTATUS will return 1 if
217 the device is open, 0 if not. [FIXME -- isn't this rather
218 silly? To be able to use the ioctl, the device must be open
219 and so GETSTATUS will always return 1].
221 advantechwdt.c -- Advantech Single Board Computer
223 Timeout that defaults to 60 seconds, supports SETTIMEOUT.
225 Supports CONFIG_WATCHDOG_NOWAYOUT
227 GETSUPPORT returns WDIOF_KEEPALIVEPING and WDIOF_SETTIMEOUT.
228 The GETSTATUS call returns if the device is open or not.
229 [FIXME -- silliness again?]
231 booke_wdt.c -- PowerPC BookE Watchdog Timer
233 Timeout default varies according to frequency, supports
236 Watchdog can not be turned off, CONFIG_WATCHDOG_NOWAYOUT
239 GETSUPPORT returns the watchdog_info struct, and
240 GETSTATUS returns the supported options. GETBOOTSTATUS
241 returns a 1 if the last reset was caused by the
242 watchdog and a 0 otherwise. This watchdog can not be
243 disabled once it has been started. The wdt_period kernel
244 parameter selects which bit of the time base changing
245 from 0->1 will trigger the watchdog exception. Changing
246 the timeout from the ioctl calls will change the
247 wdt_period as defined above. Finally if you would like to
248 replace the default Watchdog Handler you can implement the
249 WatchdogHandler() function in your own code.
251 eurotechwdt.c -- Eurotech CPU-1220/1410
253 The timeout can be set using the SETTIMEOUT ioctl and defaults
256 Also has a module parameter "ev", event type which controls
257 what should happen on a timeout, the string "int" or anything
258 else that causes a reboot. [FIXME -- better description]
260 Supports CONFIG_WATCHDOG_NOWAYOUT
262 GETSUPPORT returns CARDRESET and WDIOF_SETTIMEOUT but
263 GETSTATUS is not supported and GETBOOTSTATUS just returns 0.
265 i810-tco.c -- Intel 810 chipset
267 Also has support for a lot of other i8x0 stuff, but the
268 watchdog is one of the things.
270 The timeout is set using the module parameter "i810_margin",
271 which is in steps of 0.6 seconds where 2<i810_margin<64. The
272 driver supports the SETTIMEOUT ioctl.
274 Supports CONFIG_WATCHDOG_NOWAYOUT.
276 GETSUPPORT returns WDIOF_SETTIMEOUT. The GETSTATUS call
277 returns some kind of timer value which ist not compatible with
278 the other drivers. GETBOOT status returns some kind of
279 hardware specific boot status. [FIXME -- describe this]
281 ib700wdt.c -- IB700 Single Board Computer
283 Default timeout of 30 seconds and the timeout is settable
284 using the SETTIMEOUT ioctl. Note that only a few timeout
285 values are supported.
287 Supports CONFIG_WATCHDOG_NOWAYOUT
289 GETSUPPORT returns WDIOF_KEEPALIVEPING and WDIOF_SETTIMEOUT.
290 The GETSTATUS call returns if the device is open or not.
291 [FIXME -- silliness again?]
293 machzwd.c -- MachZ ZF-Logic
295 Hardcoded timeout of 10 seconds
297 Has a module parameter "action" that controls what happens
298 when the timeout runs out which can be 0 = RESET (default),
299 1 = SMI, 2 = NMI, 3 = SCI.
301 Supports CONFIG_WATCHDOG_NOWAYOUT and the magic character
304 GETSUPPORT returns WDIOF_KEEPALIVEPING, and the GETSTATUS call
305 returns if the device is open or not. [FIXME -- silliness
308 mixcomwd.c -- MixCom Watchdog
310 [FIXME -- I'm unable to tell what the timeout is]
312 Supports CONFIG_WATCHDOG_NOWAYOUT
314 GETSUPPORT returns WDIOF_KEEPALIVEPING, GETSTATUS returns if
315 the device is opened or not [FIXME -- I'm not really sure how
316 this works, there seems to be some magic connected to
317 CONFIG_WATCHDOG_NOWAYOUT]
319 pcwd.c -- Berkshire PC Watchdog
321 Hardcoded timeout of 1.5 seconds
323 Supports CONFIG_WATCHDOG_NOWAYOUT
325 GETSUPPORT returns WDIOF_OVERHEAT|WDIOF_CARDRESET and both
326 GETSTATUS and GETBOOTSTATUS return something useful.
328 The SETOPTIONS call can be used to enable and disable the card
329 and to ask the driver to call panic if the system overheats.
331 sbc60xxwdt.c -- 60xx Single Board Computer
333 Hardcoded timeout of 10 seconds
335 Does not support CONFIG_WATCHDOG_NOWAYOUT, but has the magic
336 character 'V' close handling.
338 No bits set in GETSUPPORT
340 scx200.c -- National SCx200 CPUs
342 Not in the kernel yet.
344 The timeout is set using a module parameter "margin" which
345 defaults to 60 seconds. The timeout can also be set using
346 SETTIMEOUT and read using GETTIMEOUT.
348 Supports a module parameter "nowayout" that is initialized
349 with the value of CONFIG_WATCHDOG_NOWAYOUT. Also supports the
350 magic character 'V' handling.
352 shwdt.c -- SuperH 3/4 processors
354 [FIXME -- I'm unable to tell what the timeout is]
356 Supports CONFIG_WATCHDOG_NOWAYOUT
358 GETSUPPORT returns WDIOF_KEEPALIVEPING, and the GETSTATUS call
359 returns if the device is open or not. [FIXME -- silliness
362 softdog.c -- Software watchdog
364 The timeout is set with the module parameter "soft_margin"
365 which defaults to 60 seconds, the timeout is also settable
366 using the SETTIMEOUT ioctl.
368 Supports CONFIG_WATCHDOG_NOWAYOUT
370 WDIOF_SETTIMEOUT bit set in GETSUPPORT
372 w83877f_wdt.c -- W83877F Computer
374 Hardcoded timeout of 30 seconds
376 Does not support CONFIG_WATCHDOG_NOWAYOUT, but has the magic
377 character 'V' close handling.
379 No bits set in GETSUPPORT
381 w83627hf_wdt.c -- w83627hf watchdog
383 Timeout that defaults to 60 seconds, supports SETTIMEOUT.
385 Supports CONFIG_WATCHDOG_NOWAYOUT
387 GETSUPPORT returns WDIOF_KEEPALIVEPING and WDIOF_SETTIMEOUT.
388 The GETSTATUS call returns if the device is open or not.
390 wdt.c -- ICS WDT500/501 ISA and
391 wdt_pci.c -- ICS WDT500/501 PCI
393 Default timeout of 60 seconds. The timeout is also settable
394 using the SETTIMEOUT ioctl.
396 Supports CONFIG_WATCHDOG_NOWAYOUT
398 GETSUPPORT returns with bits set depending on the actual
399 card. The WDT501 supports a lot of external monitoring, the
402 wdt285.c -- Footbridge watchdog
404 The timeout is set with the module parameter "soft_margin"
405 which defaults to 60 seconds. The timeout is also settable
406 using the SETTIMEOUT ioctl.
408 Does not support CONFIG_WATCHDOG_NOWAYOUT
410 WDIOF_SETTIMEOUT bit set in GETSUPPORT
412 wdt977.c -- Netwinder W83977AF chip
414 Hardcoded timeout of 3 minutes
416 Supports CONFIG_WATCHDOG_NOWAYOUT
418 Does not support any ioctls at all.