man.bsd.lv manual page server

Manual Page Search Parameters
PIDLOCK(3) Library Functions Manual PIDLOCK(3)

pidlock, ttylock, ttyunlock — locks based on files containing PIDs

library “libutil”

#include <util.h>

int
pidlock(const char *lockfile, int flags, pid_t *locker, const char *info);

int
ttylock(const char *tty, int flags, pid_t *locker);

int
ttyunlock(const char *tty);

The () (), and () functions attempt to create a lockfile for an arbitrary resource that only one program may hold at a time. (In the case of ttylock(), this is access to a tty device.) If the function succeeds in creating the lockfile, it will succeed for no other program calling it with the same lockfile until the original calling program has removed the lockfile or exited. The ttyunlock() function will remove the lockfile created by ttylock().

These functions use the method of creating a lockfile traditionally used by UUCP software. This is described as follows in the documentation for Taylor UUCP:

The lock file normally contains the process ID of the locking process. This makes it easy to determine whether a lock is still valid. The algorithm is to create a temporary file and then link it to the name that must be locked. If the link fails because a file with that name already exists, the existing file is read to get the process ID. If the process still exists, the lock attempt fails. Otherwise the lock file is deleted and the locking algorithm is retried.

The PID is stored in ASCII format, with leading spaces to pad it out to ten characters, and a terminating newline. This implementation has been extended to put the hostname on the second line of the file, terminated with a newline, and optionally an arbitrary comment on the third line of the file, also terminated with a newline. If a comment is given, but PIDLOCK_NONBLOCK is not, a blank line will be written as the second line of the file.

The () function will attempt to create the file lockfile and put the current process's pid in it. The () function will do the same, but should be passed only the base name (with no leading directory prefix) of the tty to be locked; it will test that the tty exists in /dev and is a character device, and then create the file in the /var/spool/lock directory and prefix the filename with LCK... Use the () function to remove this lock.

The following flags may be passed in flags:

The function should return immediately when a lock is held by another active process. Otherwise the function will wait (forever, if necessary) for the lock to be freed.
The hostname should be compared against the hostname in the second line of the file (if present), and if they differ, no attempt at checking for a living process holding the lock will be made, and the lockfile will never be deleted. (The process is assumed to be alive.) This is used for locking on NFS or other remote filesystems. (The function will never create a lock if PIDLOCK_USEHOSTNAME is specified and no hostname is present.)

If locker is non-null, it will contain the PID of the locking process, if there is one, on return.

If info is non-null and the lock succeeds, the string it points to will be written as the third line of the lock file.

Zero is returned if the operation was successful; on an error a -1 is returned and a standard error code is left in the global location errno.

In addition to the errors that are returned from stat(2), open(2), read(2), write(2), and link(2), pidlock() or ttylock() can set errno to the following values on failure:

[]
The tty specified in ttylock() is not a character special device.
[]
Another running process has a lock and the PIDLOCK_NONBLOCK flag was specified.

The pidlock() and ttylock() functions appeared in NetBSD 1.3.

Curt Sampson ⟨cjs@NetBSD.org⟩.

The lockfile format breaks if a pid is longer than ten digits when printed in decimal form.

The PID returned will be the pid of the locker on the remote machine if PIDLOCK_USEHOSTNAME is specified, but there is no indication that this is not on the local machine.

March 19, 2006 NetBSD-9.2