vfs_lockf.c

来自「基于组件方式开发操作系统的OSKIT源代码」· C语言 代码 · 共 799 行 · 第 1/2 页

/*	$NetBSD: vfs_lockf.c,v 1.7 1996/02/04 02:18:21 christos Exp $	*/

/*
 * Copyright (c) 1982, 1986, 1989, 1993
 *	The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * Scooter Morris at Genentech Inc.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 *	@(#)ufs_lockf.c	8.3 (Berkeley) 1/6/94
 */

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/file.h>
#include <sys/proc.h>
#include <sys/vnode.h>
#include <sys/malloc.h>
#include <sys/fcntl.h>
#include <sys/lockf.h>

/*
 * This variable controls the maximum number of processes that will
 * be checked in doing deadlock detection.
 */
int maxlockdepth = MAXDEPTH;

#ifdef LOCKF_DEBUG
int	lockf_debug = 0;
#endif

#define NOLOCKF (struct lockf *)0
#define SELF	0x1
#define OTHERS	0x2

/*
 * Do an advisory lock operation.
 */
int
lf_advlock(head, size, id, op, fl, flags)
	struct lockf **head;
	off_t size;
	caddr_t id;
	int op;
	register struct flock *fl;
	int flags;
{
	register struct lockf *lock;
	off_t start, end;
	int error;

	/*
	 * Avoid the common case of unlocking when inode has no locks.
	 */
	if (*head == (struct lockf *)0) {
		if (op != F_SETLK) {
			fl->l_type = F_UNLCK;
			return (0);
		}
	}
	/*
	 * Convert the flock structure into a start and end.
	 */
	switch (fl->l_whence) {

	case SEEK_SET:
	case SEEK_CUR:
		/*
		 * Caller is responsible for adding any necessary offset
		 * when SEEK_CUR is used.
		 */
		start = fl->l_start;
		break;

	case SEEK_END:
		start = size + fl->l_start;
		break;

	default:
		return (EINVAL);
	}
	if (start < 0)
		return (EINVAL);
	if (fl->l_len == 0)
		end = -1;
	else
		end = start + fl->l_len - 1;
	/*
	 * Create the lockf structure.
	 */
	MALLOC(lock, struct lockf *, sizeof *lock, M_LOCKF, M_WAITOK);
	lock->lf_start = start;
	lock->lf_end = end;
	lock->lf_id = id;
	lock->lf_head = head;
	lock->lf_type = fl->l_type;
	lock->lf_next = (struct lockf *)0;
	lock->lf_block = (struct lockf *)0;
	lock->lf_flags = flags;
	/*
	 * Do the requested operation.
	 */
	switch (op) {

	case F_SETLK:
		return (lf_setlock(lock));

	case F_UNLCK:
		error = lf_clearlock(lock);
		FREE(lock, M_LOCKF);
		return (error);

	case F_GETLK:
		error = lf_getlock(lock, fl);
		FREE(lock, M_LOCKF);
		return (error);

	default:
		FREE(lock, M_LOCKF);
		return (EINVAL);
	}
	/* NOTREACHED */
}

/*
 * Set a byte-range lock.
 */
int
lf_setlock(lock)
	register struct lockf *lock;
{
	register struct lockf *block;
	struct lockf **head = lock->lf_head;
	struct lockf **prev, *overlap, *ltmp;
	static char lockstr[] = "lockf";
	int ovcase, priority, needtolink, error;

#ifdef LOCKF_DEBUG
	if (lockf_debug & 1)
		lf_print("lf_setlock", lock);
#endif /* LOCKF_DEBUG */

	/*
	 * Set the priority
	 */
	priority = PLOCK;
	if (lock->lf_type == F_WRLCK)
		priority += 4;
	priority |= PCATCH;
	/*
	 * Scan lock list for this file looking for locks that would block us.
	 */
	while ((block = lf_getblock(lock)) != NULL) {
		/*
		 * Free the structure and return if nonblocking.
		 */
		if ((lock->lf_flags & F_WAIT) == 0) {
			FREE(lock, M_LOCKF);
			return (EAGAIN);
		}
		/*
		 * We are blocked. Since flock style locks cover
		 * the whole file, there is no chance for deadlock.
		 * For byte-range locks we must check for deadlock.
		 *
		 * Deadlock detection is done by looking through the
		 * wait channels to see if there are any cycles that
		 * involve us. MAXDEPTH is set just to make sure we
		 * do not go off into neverland.
		 */
		if ((lock->lf_flags & F_POSIX) &&
		    (block->lf_flags & F_POSIX)) {
			register struct proc *wproc;
			register struct lockf *waitblock;
			int i = 0;

			/* The block is waiting on something */
			wproc = (struct proc *)block->lf_id;
			while (wproc->p_wchan &&
			       (wproc->p_wmesg == lockstr) &&
			       (i++ < maxlockdepth)) {
				waitblock = (struct lockf *)wproc->p_wchan;
				/* Get the owner of the blocking lock */
				waitblock = waitblock->lf_next;
				if ((waitblock->lf_flags & F_POSIX) == 0)
					break;
				wproc = (struct proc *)waitblock->lf_id;
				if (wproc == (struct proc *)lock->lf_id) {
					free(lock, M_LOCKF);
					return (EDEADLK);
				}
			}
		}
		/*
		 * For flock type locks, we must first remove
		 * any shared locks that we hold before we sleep
		 * waiting for an exclusive lock.
		 */
		if ((lock->lf_flags & F_FLOCK) &&
		    lock->lf_type == F_WRLCK) {
			lock->lf_type = F_UNLCK;
			(void) lf_clearlock(lock);
			lock->lf_type = F_WRLCK;
		}
		/*
		 * Add our lock to the blocked list and sleep until we're free.
		 * Remember who blocked us (for deadlock detection).
		 */
		lock->lf_next = block;
		lf_addblock(block, lock);
#ifdef LOCKF_DEBUG
		if (lockf_debug & 1) {
			lf_print("lf_setlock: blocking on", block);
			lf_printlist("lf_setlock", block);
		}
#endif /* LOCKF_DEBUG */
		error = tsleep((caddr_t)lock, priority, lockstr, 0);
		if (error) {
			/*
			 * Delete ourselves from the waiting to lock list.
			 */
			for (block = lock->lf_next;
			     block != NOLOCKF;
			     block = block->lf_block) {
				if (block->lf_block != lock)
					continue;
				block->lf_block = block->lf_block->lf_block;
				break;
			}
			/*
			 * If we did not find ourselves on the list, but
			 * are still linked onto a lock list, then something
			 * is very wrong.
			 */
			if (block == NOLOCKF && lock->lf_next != NOLOCKF)
				panic("lf_setlock: lost lock");
			free(lock, M_LOCKF);
			return (error);
		}
	}
	/*
	 * No blocks!!  Add the lock.  Note that we will
	 * downgrade or upgrade any overlapping locks this
	 * process already owns.
	 *
	 * Skip over locks owned by other processes.
	 * Handle any locks that overlap and are owned by ourselves.
	 */
	prev = head;
	block = *head;
	needtolink = 1;
	for (;;) {
		ovcase = lf_findoverlap(block, lock, SELF, &prev, &overlap);
		if (ovcase)
			block = overlap->lf_next;
		/*
		 * Six cases:
		 *	0) no overlap
		 *	1) overlap == lock
		 *	2) overlap contains lock
		 *	3) lock contains overlap
		 *	4) overlap starts before lock
		 *	5) overlap ends after lock
		 */
		switch (ovcase) {
		case 0: /* no overlap */
			if (needtolink) {
				*prev = lock;
				lock->lf_next = overlap;
			}
			break;

		case 1: /* overlap == lock */
			/*
			 * If downgrading lock, others may be
			 * able to acquire it.
			 */
			if (lock->lf_type == F_RDLCK &&
			    overlap->lf_type == F_WRLCK)
				lf_wakelock(overlap);
			overlap->lf_type = lock->lf_type;
			FREE(lock, M_LOCKF);
			lock = overlap; /* for debug output below */
			break;

		case 2: /* overlap contains lock */
			/*
			 * Check for common starting point and different types.
			 */
			if (overlap->lf_type == lock->lf_type) {
				free(lock, M_LOCKF);
				lock = overlap; /* for debug output below */
				break;
			}
			if (overlap->lf_start == lock->lf_start) {
				*prev = lock;
				lock->lf_next = overlap;
				overlap->lf_start = lock->lf_end + 1;
			} else
				lf_split(overlap, lock);
			lf_wakelock(overlap);
			break;

		case 3: /* lock contains overlap */
			/*
			 * If downgrading lock, others may be able to
			 * acquire it, otherwise take the list.
			 */
			if (lock->lf_type == F_RDLCK &&
			    overlap->lf_type == F_WRLCK) {
				lf_wakelock(overlap);
			} else {
				ltmp = lock->lf_block;
				lock->lf_block = overlap->lf_block;
				lf_addblock(lock, ltmp);
			}
			/*
			 * Add the new lock if necessary and delete the overlap.
			 */
			if (needtolink) {
				*prev = lock;
				lock->lf_next = overlap->lf_next;
				prev = &lock->lf_next;
				needtolink = 0;
			} else
				*prev = overlap->lf_next;
			free(overlap, M_LOCKF);
			continue;

		case 4: /* overlap starts before lock */
			/*
			 * Add lock after overlap on the list.
			 */
			lock->lf_next = overlap->lf_next;
			overlap->lf_next = lock;
			overlap->lf_end = lock->lf_start - 1;
			prev = &lock->lf_next;
			lf_wakelock(overlap);
			needtolink = 0;
			continue;

		case 5: /* overlap ends after lock */
			/*
			 * Add the new lock before overlap.
			 */
			if (needtolink) {
				*prev = lock;
				lock->lf_next = overlap;
			}
			overlap->lf_start = lock->lf_end + 1;
			lf_wakelock(overlap);
			break;
		}
		break;
	}
#ifdef LOCKF_DEBUG
	if (lockf_debug & 1) {
		lf_print("lf_setlock: got the lock", lock);
		lf_printlist("lf_setlock", lock);
	}
#endif /* LOCKF_DEBUG */
	return (0);
}

/*
 * Remove a byte-range lock on an inode.
 *
 * Generally, find the lock (or an overlap to that lock)
 * and remove it (or shrink it), then wakeup anyone we can.
 */
int