WATCHDOG(4) BSD Kernel Interfaces Manual WATCHDOG(4)NAME
watchdog -- hardware and software watchdog
SYNOPSIS
#include <sys/watchdog.h>
DESCRIPTION
The watchdog facility is used for controlling hardware and software watchdogs.
/dev/fido responds to a single ioctl(2) call, WDIOCPATPAT. It takes a single argument which represents a timeout value specified as a power
of two nanoseconds, or-ed with a flag selecting active or passive control of the watchdog.
WD_ACTIVE indicates that the watchdog will be kept from timing out from userland, for instance by the watchdogd(8) daemon. WD_PASSIVE indi-
cates that the watchdog will be kept from timing out from the kernel.
The ioctl(2) call will return success if just one of the available watchdog(9) implementations supports setting the timeout to the specified
timeout. This means that at least one watchdog is armed. If the call fails, for instance if none of watchdog(9) implementations support the
timeout length, all watchdogs are disabled and must be explicitly re-enabled.
To disable the watchdogs pass WD_TO_NEVER. If disarming the watchdog(s) failed an error is returned. The watchdog might still be armed!
RETURN VALUES
The ioctl returns zero on success and non-zero on failure.
[EOPNOTSUPP] No watchdog present in the kernel (timeout value other than 0).
[EOPNOTSUPP] Watchdog could not be disabled (timeout value of 0).
[EINVALID] Invalid flag combination passed.
[EINVALID] None of the watchdogs supports the requested timeout value.
EXAMPLES
#include <paths.h>
#include <sys/watchdog.h>
#define WDPATH "/dev/" _PATH_WATCHDOG
int wdfd = -1;
static void
wd_init(void)
{
wdfd = open(WDPATH, O_RDWR);
if (wdfd == -1)
err(1, WDPATH);
}
static void
wd_reset(u_int timeout)
{
if (ioctl(wdfd, WDIOCPATPAT, &timeout) == -1)
err(1, "WDIOCPATPAT");
}
/* in main() */
wd_init();
wd_reset(WD_ACTIVE|WD_TO_8SEC);
/* potential freeze point */
wd_reset(WD_TO_NEVER);
Enables a watchdog to recover from a potentially freezing piece of code.
options SW_WATCHDOG
in your kernel config adds a software watchdog in the kernel, dropping to KDB or panic-ing when firing.
SEE ALSO watchdogd(8), watchdog(9)HISTORY
The watchdog code first appeared in FreeBSD 5.1.
BUGS
The WD_PASSIVE option has not yet been implemented.
AUTHORS
The watchdog facility was written by Poul-Henning Kamp <phk@FreeBSD.org>. The software watchdog code and this manual page were written by
Sean Kelly <smkelly@FreeBSD.org>. Some contributions were made by Jeff Roberson <jeff@FreeBSD.org>.
BSD June 25, 2003 BSD
Check Out this Related Man Page
WDMD(8) System Manager's Manual WDMD(8)NAME
wdmd - watchdog multiplexing daemon
SYNOPSIS
wdmd [OPTIONS]
DESCRIPTION
This daemon opens /dev/watchdog and allows multiple independent sources to detmermine whether each KEEPALIVE is done. Every test interval
(10 seconds), the daemon tests each source. If any test fails, the KEEPALIVE is not done. In a standard configuration, the watchdog timer
will reset the system if no KEEPALIVE is done for 60 seconds ("fire timeout"). This means that if a single test fails 5-6 times in row,
the watchdog will fire and reset the system. With multiple test sources, fewer separate failures back to back can also cause a reset, e.g.
T seconds, P pass, F fail
T00: test1 P, test2 P, test3 P: KEEPALIVE done
T10: test1 F, test2 F, test3 P: KEEPALIVE skipped
T20: test1 F, test2 P, test3 P: KEEPALIVE skipped
T30: test1 P, test2 F, test3 P: KEEPALIVE skipped
T40: test1 P, test2 P, test3 F: KEEPALIVE skipped
T50: test1 F, test2 F, test3 P: KEEPALIVE skipped
T60: test1 P, test2 F, test3 P: KEEPALIVE skipped
T60: watchdog fires, system resets
(Depending on timings, the system may be reset sometime shortly before T60, and the tests at T60 would not be run.)
A crucial aspect to the design and function of wdmd is that if any single source does not pass tests for the fire timeout, the watchdog is
guaranteed to fire, regardless of whether other sources on the system have passed or failed. A spurious reset due to the combined effects
of multiple failing tests as shown above, is an accepted side effect.
The wdmd init script will load the softdog module if no other watchdog module has been loaded.
wdmd cannot be used on the system with any other program that needs to open /dev/watchdog, e.g. watchdog(8).
Test Source: clients
Using libwdmd, programs connect to wdmd via a unix socket, and send regular messages to wdmd to update an expiry time for their connection.
Every test interval, wdmd will check if the expiry time for a connection has been reached. If so, the test for that client fails.
Test Source: scripts
wdmd will run scripts from a designated directory every test interval. If a script exits with 0, the test is considered a success, other-
wise a failure. If a script does not exit by the end of the test interval, it is considered a failure.
OPTIONS --version, -V
Print version.
--help, -h
Print usage.
--dump, -d
Print debug information from the daemon.
--probe, -p
Print path of functional watchdog device. Exit code 0 indicates a
functional device was found. Exit code 1 indicates a functional device
was not found.
-D
Enable debugging to stderr and don't fork.
-H 0|1
Enable (1) or disable (0) high priority features such as realtime
scheduling priority and mlockall.
-G name
Group ownership for the socket.
-S 0|1
Enable (1) or disable (0) script tests.
-s path
Path to scripts dir.
-k num
Kill unfinished scripts after num seconds.
-w path
The path to the watchdog device to try first.
2011-08-01 WDMD(8)