#define MSNFS /* HACK HACK */
/*
* linux/fs/locks.c
*
* Provide support for fcntl()'s F_GETLK, F_SETLK, and F_SETLKW calls.
* Doug Evans (dje@spiff.uucp), August 07, 1992
*
* Deadlock detection added.
* FIXME: one thing isn't handled yet:
* - mandatory locks (requires lots of changes elsewhere)
* Kelly Carmichael (kelly@[142.24.8.65]), September 17, 1994.
*
* Miscellaneous edits, and a total rewrite of posix_lock_file() code.
* Kai Petzke (wpp@marie.physik.tu-berlin.de), 1994
*
* Converted file_lock_table to a linked list from an array, which eliminates
* the limits on how many active file locks are open.
* Chad Page (pageone@netcom.com), November 27, 1994
*
* Removed dependency on file descriptors. dup()'ed file descriptors now
* get the same locks as the original file descriptors, and a close() on
* any file descriptor removes ALL the locks on the file for the current
* process. Since locks still depend on the process id, locks are inherited
* after an exec() but not after a fork(). This agrees with POSIX, and both
* BSD and SVR4 practice.
* Andy Walker (andy@lysaker.kvaerner.no), February 14, 1995
*
* Scrapped free list which is redundant now that we allocate locks
* dynamically with kmalloc()/kfree().
* Andy Walker (andy@lysaker.kvaerner.no), February 21, 1995
*
* Implemented two lock personalities - FL_FLOCK and FL_POSIX.
*
* FL_POSIX locks are created with calls to fcntl() and lockf() through the
* fcntl() system call. They have the semantics described above.
*
* FL_FLOCK locks are created with calls to flock(), through the flock()
* system call, which is new. Old C libraries implement flock() via fcntl()
* and will continue to use the old, broken implementation.
*
* FL_FLOCK locks follow the 4.4 BSD flock() semantics. They are associated
* with a file pointer (filp). As a result they can be shared by a parent
* process and its children after a fork(). They are removed when the last
* file descriptor referring to the file pointer is closed (unless explicitly
* unlocked).
*
* FL_FLOCK locks never deadlock, an existing lock is always removed before
* upgrading from shared to exclusive (or vice versa). When this happens
* any processes blocked by the current lock are woken up and allowed to
* run before the new lock is applied.
* Andy Walker (andy@lysaker.kvaerner.no), June 09, 1995
*
* Removed some race conditions in flock_lock_file(), marked other possible
* races. Just grep for FIXME to see them.
* Dmitry Gorodchanin (pgmdsg@ibi.com), February 09, 1996.
*
* Addressed Dmitry's concerns. Deadlock checking no longer recursive.
* Lock allocation changed to GFP_ATOMIC as we can't afford to sleep
* once we've checked for blocking and deadlocking.
* Andy Walker (andy@lysaker.kvaerner.no), April 03, 1996.
*
* Initial implementation of mandatory locks. SunOS turned out to be
* a rotten model, so I implemented the "obvious" semantics.
* See 'linux/Documentation/mandatory.txt' for details.
* Andy Walker (andy@lysaker.kvaerner.no), April 06, 1996.
*
* Don't allow mandatory locks on mmap()'ed files. Added simple functions to
* check if a file has mandatory locks, used by mmap(), open() and creat() to
* see if system call should be rejected. Ref. HP-UX/SunOS/Solaris Reference
* Manual, Section 2.
* Andy Walker (andy@lysaker.kvaerner.no), April 09, 1996.
*
* Tidied up block list handling. Added '/proc/locks' interface.
* Andy Walker (andy@lysaker.kvaerner.no), April 24, 1996.
*
* Fixed deadlock condition for pathological code that mixes calls to
* flock() and fcntl().
* Andy Walker (andy@lysaker.kvaerner.no), April 29, 1996.
*
* Allow only one type of locking scheme (FL_POSIX or FL_FLOCK) to be in use
* for a given file at a time. Changed the CONFIG_LOCK_MANDATORY scheme to
* guarantee sensible behaviour in the case where file system modules might
* be compiled with different options than the kernel itself.
* Andy Walker (andy@lysaker.kvaerner.no), May 15, 1996.
*
* Added a couple of missing wake_up() calls. Thanks to Thomas Meckel
* (Thomas.Meckel@mni.fh-giessen.de) for spotting this.
* Andy Walker (andy@lysaker.kvaerner.no), May 15, 1996.
*
* Changed FL_POSIX locks to use the block list in the same way as FL_FLOCK
* locks. Changed process synchronisation to avoid dereferencing locks that
* have already been freed.
* Andy Walker (andy@lysaker.kvaerner.no), Sep 21, 1996.
*
* Made the block list a circular list to minimise searching in the list.
* Andy Walker (andy@lysaker.kvaerner.no), Sep 25, 1996.
*
* Made mandatory locking a mount option. Default is not to allow mandatory
* locking.
* Andy Walker (andy@lysaker.kvaerner.no), Oct 04, 1996.
*
* Some adaptations for NFS support.
* Olaf Kirch (okir@monad.swb.de), Dec 1996,
*
* Fixed /proc/locks interface so that we can't overrun the buffer we are handed.
* Andy Walker (andy@lysaker.kvaerner.no), May 12, 1997.
*
* Use slab allocator instead of kmalloc/kfree.
* Use generic list implementation from <linux/list.h>.
* Sped up posix_locks_deadlock by only considering blocked locks.
* Matthew Wilcox <willy@thepuffingroup.com>, March, 2000.
*
* Leases and LOCK_MAND
* Matthew Wilcox <willy@linuxcare.com>, June, 2000.
* Stephen Rothwell <sfr@linuxcare.com>, June, 2000.
*/
#include <linux/malloc.h>
#include <linux/file.h>
#include <linux/smp_lock.h>
#include <linux/init.h>
#include <linux/capability.h>
#include <linux/sched.h>
#include <asm/semaphore.h>
#include <asm/uaccess.h>
int leases_enable = 1;
int lease_break_time = 45;
LIST_HEAD(file_lock_list);
static LIST_HEAD(blocked_list);
static kmem_cache_t *filelock_cache;
/* Allocate an empty lock structure. */
137 static struct file_lock *locks_alloc_lock(int account)
{
struct file_lock *fl;
140 if (account && current->locks >= current->rlim[RLIMIT_LOCKS].rlim_cur)
141 return NULL;
fl = kmem_cache_alloc(filelock_cache, SLAB_KERNEL);
143 if (fl)
current->locks++;
145 return fl;
}
/* Free a lock which is not in use. */
149 static inline void locks_free_lock(struct file_lock *fl)
{
151 if (fl == NULL) {
152 BUG();
153 return;
}
current->locks--;
156 if (waitqueue_active(&fl->fl_wait))
panic("Attempting to free lock with active wait queue");
159 if (!list_empty(&fl->fl_block))
panic("Attempting to free lock with active block list");
162 if (!list_empty(&fl->fl_link))
panic("Attempting to free lock on active lock list");
kmem_cache_free(filelock_cache, fl);
}
168 void locks_init_lock(struct file_lock *fl)
{
170 INIT_LIST_HEAD(&fl->fl_link);
171 INIT_LIST_HEAD(&fl->fl_block);
init_waitqueue_head(&fl->fl_wait);
fl->fl_next = NULL;
fl->fl_fasync = NULL;
fl->fl_owner = 0;
fl->fl_pid = 0;
fl->fl_file = NULL;
fl->fl_flags = 0;
fl->fl_type = 0;
fl->fl_start = fl->fl_end = 0;
fl->fl_notify = NULL;
fl->fl_insert = NULL;
fl->fl_remove = NULL;
}
/*
* Initialises the fields of the file lock which are invariant for
* free file_locks.
*/
190 static void init_once(void *foo, kmem_cache_t *cache, unsigned long flags)
{
struct file_lock *lock = (struct file_lock *) foo;
if ((flags & (SLAB_CTOR_VERIFY|SLAB_CTOR_CONSTRUCTOR)) !=
195 SLAB_CTOR_CONSTRUCTOR)
196 return;
locks_init_lock(lock);
}
/*
* Initialize a new lock from an existing file_lock structure.
*/
204 void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
{
new->fl_owner = fl->fl_owner;
new->fl_pid = fl->fl_pid;
new->fl_file = fl->fl_file;
new->fl_flags = fl->fl_flags;
new->fl_type = fl->fl_type;
new->fl_start = fl->fl_start;
new->fl_end = fl->fl_end;
new->fl_notify = fl->fl_notify;
new->fl_insert = fl->fl_insert;
new->fl_remove = fl->fl_remove;
new->fl_u = fl->fl_u;
}
/* Fill in a file_lock structure with an appropriate FLOCK lock. */
220 static struct file_lock *flock_make_lock(struct file *filp, unsigned int type)
{
struct file_lock *fl = locks_alloc_lock(1);
223 if (fl == NULL)
224 return NULL;
fl->fl_owner = NULL;
fl->fl_file = filp;
fl->fl_pid = current->pid;
fl->fl_flags = FL_FLOCK;
fl->fl_type = type;
fl->fl_start = 0;
fl->fl_end = OFFSET_MAX;
fl->fl_notify = NULL;
fl->fl_insert = NULL;
fl->fl_remove = NULL;
237 return fl;
}
240 static int assign_type(struct file_lock *fl, int type)
{
242 switch (type) {
243 case F_RDLCK:
244 case F_WRLCK:
245 case F_UNLCK:
fl->fl_type = type;
247 break;
248 default:
249 return -EINVAL;
}
251 return 0;
}
/* Verify a "struct flock" and copy it to a "struct file_lock" as a POSIX
* style lock.
*/
257 static int flock_to_posix_lock(struct file *filp, struct file_lock *fl,
struct flock *l)
{
loff_t start;
262 switch (l->l_whence) {
263 case 0: /*SEEK_SET*/
start = 0;
265 break;
266 case 1: /*SEEK_CUR*/
start = filp->f_pos;
268 break;
269 case 2: /*SEEK_END*/
start = filp->f_dentry->d_inode->i_size;
271 break;
272 default:
273 return (0);
}
276 if (((start += l->l_start) < 0) || (l->l_len < 0))
277 return (0);
fl->fl_end = start + l->l_len - 1;
279 if (l->l_len > 0 && fl->fl_end < 0)
280 return (0);
281 if (fl->fl_end > OFFT_OFFSET_MAX)
282 return 0;
fl->fl_start = start; /* we record the absolute position */
284 if (l->l_len == 0)
fl->fl_end = OFFSET_MAX;
fl->fl_owner = current->files;
fl->fl_pid = current->pid;
fl->fl_file = filp;
fl->fl_flags = FL_POSIX;
fl->fl_notify = NULL;
fl->fl_insert = NULL;
fl->fl_remove = NULL;
295 return (assign_type(fl, l->l_type) == 0);
}
#if BITS_PER_LONG == 32
299 static int flock64_to_posix_lock(struct file *filp, struct file_lock *fl,
struct flock64 *l)
{
loff_t start;
304 switch (l->l_whence) {
305 case 0: /*SEEK_SET*/
start = 0;
307 break;
308 case 1: /*SEEK_CUR*/
start = filp->f_pos;
310 break;
311 case 2: /*SEEK_END*/
start = filp->f_dentry->d_inode->i_size;
313 break;
314 default:
315 return (0);
}
318 if (((start += l->l_start) < 0) || (l->l_len < 0))
319 return (0);
fl->fl_end = start + l->l_len - 1;
321 if (l->l_len > 0 && fl->fl_end < 0)
322 return (0);
fl->fl_start = start; /* we record the absolute position */
324 if (l->l_len == 0)
fl->fl_end = OFFSET_MAX;
fl->fl_owner = current->files;
fl->fl_pid = current->pid;
fl->fl_file = filp;
fl->fl_flags = FL_POSIX;
fl->fl_notify = NULL;
fl->fl_insert = NULL;
fl->fl_remove = NULL;
335 switch (l->l_type) {
336 case F_RDLCK:
337 case F_WRLCK:
338 case F_UNLCK:
fl->fl_type = l->l_type;
340 break;
341 default:
342 return (0);
}
345 return (1);
}
#endif
/* Allocate a file_lock initialised to this type of lease */
350 static int lease_alloc(struct file *filp, int type, struct file_lock **flp)
{
struct file_lock *fl = locks_alloc_lock(1);
353 if (fl == NULL)
354 return -ENOMEM;
fl->fl_owner = current->files;
fl->fl_pid = current->pid;
fl->fl_file = filp;
fl->fl_flags = FL_LEASE;
361 if (assign_type(fl, type) != 0) {
locks_free_lock(fl);
363 return -EINVAL;
}
fl->fl_start = 0;
fl->fl_end = OFFSET_MAX;
fl->fl_notify = NULL;
fl->fl_insert = NULL;
fl->fl_remove = NULL;
*flp = fl;
372 return 0;
}
/* Check if two locks overlap each other.
*/
377 static inline int locks_overlap(struct file_lock *fl1, struct file_lock *fl2)
{
return ((fl1->fl_end >= fl2->fl_start) &&
380 (fl2->fl_end >= fl1->fl_start));
}
/*
* Check whether two locks have the same owner
* N.B. Do we need the test on PID as well as owner?
* (Clone tasks should be considered as one "owner".)
*/
static inline int
389 locks_same_owner(struct file_lock *fl1, struct file_lock *fl2)
{
return (fl1->fl_owner == fl2->fl_owner) &&
392 (fl1->fl_pid == fl2->fl_pid);
}
/* Remove waiter from blocker's block list.
* When blocker ends up pointing to itself then the list is empty.
*/
398 static void locks_delete_block(struct file_lock *waiter)
{
list_del(&waiter->fl_block);
401 INIT_LIST_HEAD(&waiter->fl_block);
list_del(&waiter->fl_link);
403 INIT_LIST_HEAD(&waiter->fl_link);
waiter->fl_next = NULL;
}
/* Insert waiter into blocker's block list.
* We use a circular list so that processes can be easily woken up in
* the order they blocked. The documentation doesn't require this but
* it seems like the reasonable thing to do.
*/
412 static void locks_insert_block(struct file_lock *blocker,
struct file_lock *waiter)
{
415 if (!list_empty(&waiter->fl_block)) {
printk(KERN_ERR "locks_insert_block: removing duplicated lock "
"(pid=%d %Ld-%Ld type=%d)\n", waiter->fl_pid,
waiter->fl_start, waiter->fl_end, waiter->fl_type);
locks_delete_block(waiter);
}
list_add_tail(&waiter->fl_block, &blocker->fl_block);
waiter->fl_next = blocker;
list_add(&waiter->fl_link, &blocked_list);
}
static inline
427 void locks_notify_blocked(struct file_lock *waiter)
{
429 if (waiter->fl_notify)
waiter->fl_notify(waiter);
431 else
wake_up(&waiter->fl_wait);
}
/* Wake up processes blocked waiting for blocker.
* If told to wait then schedule the processes until the block list
* is empty, otherwise empty the block list ourselves.
*/
439 static void locks_wake_up_blocks(struct file_lock *blocker, unsigned int wait)
{
441 while (!list_empty(&blocker->fl_block)) {
struct file_lock *waiter = list_entry(blocker->fl_block.next, struct file_lock, fl_block);
444 if (wait) {
locks_notify_blocked(waiter);
/* Let the blocked process remove waiter from the
* block list when it gets scheduled.
*/
current->policy |= SCHED_YIELD;
schedule();
451 } else {
/* Remove waiter from the block list, because by the
* time it wakes up blocker won't exist any more.
*/
locks_delete_block(waiter);
locks_notify_blocked(waiter);
}
}
}
/* Insert file lock fl into an inode's lock list at the position indicated
* by pos. At the same time add the lock to the global file lock list.
*/
464 static void locks_insert_lock(struct file_lock **pos, struct file_lock *fl)
{
list_add(&fl->fl_link, &file_lock_list);
/* insert into file's list */
fl->fl_next = *pos;
*pos = fl;
472 if (fl->fl_insert)
fl->fl_insert(fl);
}
/* Delete a lock and then free it.
* Remove our lock from the lock lists, wake up processes that are blocked
* waiting for this lock, notify the FS that the lock has been cleared and
* finally free the lock.
*/
481 static void locks_delete_lock(struct file_lock **thisfl_p, unsigned int wait)
{
struct file_lock *fl = *thisfl_p;
*thisfl_p = fl->fl_next;
fl->fl_next = NULL;
list_del(&fl->fl_link);
489 INIT_LIST_HEAD(&fl->fl_link);
fasync_helper(0, fl->fl_file, 0, &fl->fl_fasync);
492 if (fl->fl_fasync != NULL){
printk(KERN_ERR "locks_delete_lock: fasync == %p\n", fl->fl_fasync);
fl->fl_fasync = NULL;
}
497 if (fl->fl_remove)
fl->fl_remove(fl);
locks_wake_up_blocks(fl, wait);
locks_free_lock(fl);
}
/*
* Call back client filesystem in order to get it to unregister a lock,
* then delete lock. Essentially useful only in locks_remove_*().
* Note: this must be called with the semaphore already held!
*/
509 static inline void locks_unlock_delete(struct file_lock **thisfl_p)
{
struct file_lock *fl = *thisfl_p;
int (*lock)(struct file *, int, struct file_lock *);
if (fl->fl_file->f_op &&
515 (lock = fl->fl_file->f_op->lock) != NULL) {
fl->fl_type = F_UNLCK;
lock(fl->fl_file, F_SETLK, fl);
}
locks_delete_lock(thisfl_p, 0);
}
/* Determine if lock sys_fl blocks lock caller_fl. Common functionality
* checks for shared/exclusive status of overlapping locks.
*/
525 static int locks_conflict(struct file_lock *caller_fl, struct file_lock *sys_fl)
{
527 switch (caller_fl->fl_type) {
528 case F_RDLCK:
529 return (sys_fl->fl_type == F_WRLCK);
531 case F_WRLCK:
532 return (1);
534 default:
printk("locks_conflict(): impossible lock type - %d\n",
caller_fl->fl_type);
537 break;
}
539 return (0); /* This should never happen */
}
/* Determine if lock sys_fl blocks lock caller_fl. POSIX specific
* checking before calling the locks_conflict().
*/
545 static int posix_locks_conflict(struct file_lock *caller_fl, struct file_lock *sys_fl)
{
/* POSIX locks owned by the same process do not conflict with
* each other.
*/
if (!(sys_fl->fl_flags & FL_POSIX) ||
551 locks_same_owner(caller_fl, sys_fl))
552 return (0);
/* Check whether they overlap */
555 if (!locks_overlap(caller_fl, sys_fl))
556 return 0;
558 return (locks_conflict(caller_fl, sys_fl));
}
/* Determine if lock sys_fl blocks lock caller_fl. FLOCK specific
* checking before calling the locks_conflict().
*/
564 static int flock_locks_conflict(struct file_lock *caller_fl, struct file_lock *sys_fl)
{
/* FLOCK locks referring to the same filp do not conflict with
* each other.
*/
if (!(sys_fl->fl_flags & FL_FLOCK) ||
570 (caller_fl->fl_file == sys_fl->fl_file))
571 return (0);
#ifdef MSNFS
573 if ((caller_fl->fl_type & LOCK_MAND) || (sys_fl->fl_type & LOCK_MAND))
574 return 0;
#endif
577 return (locks_conflict(caller_fl, sys_fl));
}
580 static int interruptible_sleep_on_locked(wait_queue_head_t *fl_wait, int timeout)
{
int result = 0;
DECLARE_WAITQUEUE(wait, current);
current->state = TASK_INTERRUPTIBLE;
add_wait_queue(fl_wait, &wait);
587 if (timeout == 0)
schedule();
589 else
result = schedule_timeout(timeout);
591 if (signal_pending(current))
result = -ERESTARTSYS;
remove_wait_queue(fl_wait, &wait);
current->state = TASK_RUNNING;
595 return result;
}
598 static int locks_block_on(struct file_lock *blocker, struct file_lock *waiter)
{
int result;
locks_insert_block(blocker, waiter);
result = interruptible_sleep_on_locked(&waiter->fl_wait, 0);
locks_delete_block(waiter);
604 return result;
}
607 static int locks_block_on_timeout(struct file_lock *blocker, struct file_lock *waiter, int time)
{
int result;
locks_insert_block(blocker, waiter);
result = interruptible_sleep_on_locked(&waiter->fl_wait, time);
locks_delete_block(waiter);
613 return result;
}
struct file_lock *
617 posix_test_lock(struct file *filp, struct file_lock *fl)
{
struct file_lock *cfl;
621 lock_kernel();
622 for (cfl = filp->f_dentry->d_inode->i_flock; cfl; cfl = cfl->fl_next) {
623 if (!(cfl->fl_flags & FL_POSIX))
624 continue;
625 if (posix_locks_conflict(cfl, fl))
626 break;
}
628 unlock_kernel();
630 return (cfl);
}
/* This function tests for deadlock condition before putting a process to
* sleep. The detection scheme is no longer recursive. Recursive was neat,
* but dangerous - we risked stack corruption if the lock data was bad, or
* if the recursion was too deep for any other reason.
*
* We rely on the fact that a task can only be on one lock's wait queue
* at a time. When we find blocked_task on a wait queue we can re-search
* with blocked_task equal to that queue's owner, until either blocked_task
* isn't found, or blocked_task is found on a queue owned by my_task.
*
* Note: the above assumption may not be true when handling lock requests
* from a broken NFS client. But broken NFS clients have a lot more to
* worry about than proper deadlock detection anyway... --okir
*/
647 static int posix_locks_deadlock(struct file_lock *caller_fl,
struct file_lock *block_fl)
{
struct list_head *tmp;
fl_owner_t caller_owner, blocked_owner;
unsigned int caller_pid, blocked_pid;
caller_owner = caller_fl->fl_owner;
caller_pid = caller_fl->fl_pid;
blocked_owner = block_fl->fl_owner;
blocked_pid = block_fl->fl_pid;
next_task:
660 if (caller_owner == blocked_owner && caller_pid == blocked_pid)
661 return 1;
662 list_for_each(tmp, &blocked_list) {
struct file_lock *fl = list_entry(tmp, struct file_lock, fl_link);
if ((fl->fl_owner == blocked_owner)
665 && (fl->fl_pid == blocked_pid)) {
fl = fl->fl_next;
blocked_owner = fl->fl_owner;
blocked_pid = fl->fl_pid;
669 goto next_task;
}
}
672 return 0;
}
675 int locks_mandatory_locked(struct inode *inode)
{
fl_owner_t owner = current->files;
struct file_lock *fl;
/*
* Search the lock list for this inode for any POSIX locks.
*/
683 lock_kernel();
684 for (fl = inode->i_flock; fl != NULL; fl = fl->fl_next) {
685 if (!(fl->fl_flags & FL_POSIX))
686 continue;
687 if (fl->fl_owner != owner)
688 break;
}
690 unlock_kernel();
691 return fl ? -EAGAIN : 0;
}
694 int locks_mandatory_area(int read_write, struct inode *inode,
struct file *filp, loff_t offset,
size_t count)
{
struct file_lock *fl;
struct file_lock *new_fl = locks_alloc_lock(0);
int error;
new_fl->fl_owner = current->files;
new_fl->fl_pid = current->pid;
new_fl->fl_file = filp;
new_fl->fl_flags = FL_POSIX | FL_ACCESS;
new_fl->fl_type = (read_write == FLOCK_VERIFY_WRITE) ? F_WRLCK : F_RDLCK;
new_fl->fl_start = offset;
new_fl->fl_end = offset + count - 1;
error = 0;
711 lock_kernel();
repeat:
/* Search the lock list for this inode for locks that conflict with
* the proposed read/write.
*/
717 for (fl = inode->i_flock; fl != NULL; fl = fl->fl_next) {
718 if (!(fl->fl_flags & FL_POSIX))
719 continue;
720 if (fl->fl_start > new_fl->fl_end)
721 break;
722 if (posix_locks_conflict(new_fl, fl)) {
error = -EAGAIN;
724 if (filp && (filp->f_flags & O_NONBLOCK))
725 break;
error = -EDEADLK;
727 if (posix_locks_deadlock(new_fl, fl))
728 break;
error = locks_block_on(fl, new_fl);
731 if (error != 0)
732 break;
/*
* If we've been sleeping someone might have
* changed the permissions behind our back.
*/
738 if ((inode->i_mode & (S_ISGID | S_IXGRP)) != S_ISGID)
739 break;
740 goto repeat;
}
}
locks_free_lock(new_fl);
744 unlock_kernel();
745 return error;
}
/* Try to create a FLOCK lock on filp. We always insert new FLOCK locks
* at the head of the list, but that's secret knowledge known only to
* flock_lock_file and posix_lock_file.
*/
752 static int flock_lock_file(struct file *filp, unsigned int lock_type,
unsigned int wait)
{
struct file_lock *fl;
struct file_lock *new_fl = NULL;
struct file_lock **before;
struct inode * inode = filp->f_dentry->d_inode;
int error, change;
int unlock = (lock_type == F_UNLCK);
/*
* If we need a new lock, get it in advance to avoid races.
*/
765 if (!unlock) {
error = -ENOLCK;
new_fl = flock_make_lock(filp, lock_type);
768 if (!new_fl)
769 return error;
}
error = 0;
search:
change = 0;
before = &inode->i_flock;
776 while (((fl = *before) != NULL) && (fl->fl_flags & FL_FLOCK)) {
777 if (filp == fl->fl_file) {
778 if (lock_type == fl->fl_type)
779 goto out;
change = 1;
781 break;
}
before = &fl->fl_next;
}
/* change means that we are changing the type of an existing lock,
* or else unlocking it.
*/
788 if (change) {
/* N.B. What if the wait argument is false? */
locks_delete_lock(before, !unlock);
/*
* If we waited, another lock may have been added ...
*/
794 if (!unlock)
795 goto search;
}
797 if (unlock)
798 goto out;
repeat:
801 for (fl = inode->i_flock; (fl != NULL) && (fl->fl_flags & FL_FLOCK);
fl = fl->fl_next) {
803 if (!flock_locks_conflict(new_fl, fl))
804 continue;
error = -EAGAIN;
806 if (!wait)
807 goto out;
error = locks_block_on(fl, new_fl);
809 if (error != 0)
810 goto out;
811 goto repeat;
}
locks_insert_lock(&inode->i_flock, new_fl);
new_fl = NULL;
error = 0;
out:
818 if (new_fl)
locks_free_lock(new_fl);
820 return error;
}
/**
* posix_lock_file:
* @filp: The file to apply the lock to
* @caller: The lock to be applied
* @wait: 1 to retry automatically, 0 to return -EAGAIN
*
* Add a POSIX style lock to a file.
* We merge adjacent locks whenever possible. POSIX locks are sorted by owner
* task, then by starting address
*
* Kai Petzke writes:
* To make freeing a lock much faster, we keep a pointer to the lock before the
* actual one. But the real gain of the new coding was, that lock_it() and
* unlock_it() became one function.
*
* To all purists: Yes, I use a few goto's. Just pass on to the next function.
*/
841 int posix_lock_file(struct file *filp, struct file_lock *caller,
unsigned int wait)
{
struct file_lock *fl;
struct file_lock *new_fl, *new_fl2;
struct file_lock *left = NULL;
struct file_lock *right = NULL;
struct file_lock **before;
struct inode * inode = filp->f_dentry->d_inode;
int error, added = 0;
/*
* We may need two file_lock structures for this operation,
* so we get them in advance to avoid races.
*/
new_fl = locks_alloc_lock(0);
new_fl2 = locks_alloc_lock(0);
error = -ENOLCK; /* "no luck" */
859 if (!(new_fl && new_fl2))
860 goto out;
862 lock_kernel();
863 if (caller->fl_type != F_UNLCK) {
repeat:
865 for (fl = inode->i_flock; fl != NULL; fl = fl->fl_next) {
866 if (!(fl->fl_flags & FL_POSIX))
867 continue;
868 if (!posix_locks_conflict(caller, fl))
869 continue;
error = -EAGAIN;
871 if (!wait)
872 goto out;
error = -EDEADLK;
874 if (posix_locks_deadlock(caller, fl))
875 goto out;
error = locks_block_on(fl, caller);
878 if (error != 0)
879 goto out;
880 goto repeat;
}
}
/*
* We've allocated the new locks in advance, so there are no
* errors possible (and no blocking operations) from here on.
*
* Find the first old lock with the same owner as the new lock.
*/
before = &inode->i_flock;
/* First skip locks owned by other processes.
*/
while ((fl = *before) && (!(fl->fl_flags & FL_POSIX) ||
896 !locks_same_owner(caller, fl))) {
before = &fl->fl_next;
}
/* Process locks with this owner.
*/
902 while ((fl = *before) && locks_same_owner(caller, fl)) {
/* Detect adjacent or overlapping regions (if same lock type)
*/
905 if (caller->fl_type == fl->fl_type) {
906 if (fl->fl_end < caller->fl_start - 1)
907 goto next_lock;
/* If the next lock in the list has entirely bigger
* addresses than the new one, insert the lock here.
*/
911 if (fl->fl_start > caller->fl_end + 1)
912 break;
/* If we come here, the new and old lock are of the
* same type and adjacent or overlapping. Make one
* lock yielding from the lower start address of both
* locks to the higher end address.
*/
919 if (fl->fl_start > caller->fl_start)
fl->fl_start = caller->fl_start;
921 else
caller->fl_start = fl->fl_start;
923 if (fl->fl_end < caller->fl_end)
fl->fl_end = caller->fl_end;
925 else
caller->fl_end = fl->fl_end;
927 if (added) {
locks_delete_lock(before, 0);
929 continue;
}
caller = fl;
added = 1;
}
934 else {
/* Processing for different lock types is a bit
* more complex.
*/
938 if (fl->fl_end < caller->fl_start)
939 goto next_lock;
940 if (fl->fl_start > caller->fl_end)
941 break;
942 if (caller->fl_type == F_UNLCK)
added = 1;
944 if (fl->fl_start < caller->fl_start)
left = fl;
/* If the next lock in the list has a higher end
* address than the new one, insert the new one here.
*/
949 if (fl->fl_end > caller->fl_end) {
right = fl;
951 break;
}
953 if (fl->fl_start >= caller->fl_start) {
/* The new lock completely replaces an old
* one (This may happen several times).
*/
957 if (added) {
locks_delete_lock(before, 0);
959 continue;
}
/* Replace the old lock with the new one.
* Wake up anybody waiting for the old one,
* as the change in lock type might satisfy
* their needs.
*/
locks_wake_up_blocks(fl, 0); /* This cannot schedule()! */
fl->fl_start = caller->fl_start;
fl->fl_end = caller->fl_end;
fl->fl_type = caller->fl_type;
fl->fl_u = caller->fl_u;
caller = fl;
added = 1;
}
}
/* Go on to next lock.
*/
next_lock:
before = &fl->fl_next;
}
error = 0;
982 if (!added) {
983 if (caller->fl_type == F_UNLCK)
984 goto out;
locks_copy_lock(new_fl, caller);
locks_insert_lock(before, new_fl);
new_fl = NULL;
}
989 if (right) {
990 if (left == right) {
/* The new lock breaks the old one in two pieces,
* so we have to use the second new lock.
*/
left = new_fl2;
new_fl2 = NULL;
locks_copy_lock(left, right);
locks_insert_lock(before, left);
}
right->fl_start = caller->fl_end + 1;
locks_wake_up_blocks(right, 0);
}
1002 if (left) {
left->fl_end = caller->fl_start - 1;
locks_wake_up_blocks(left, 0);
}
out:
1007 unlock_kernel();
/*
* Free any unused locks.
*/
1011 if (new_fl)
locks_free_lock(new_fl);
1013 if (new_fl2)
locks_free_lock(new_fl2);
1015 return error;
}
1018 static inline int flock_translate_cmd(int cmd) {
#ifdef MSNFS
1020 if (cmd & LOCK_MAND)
1021 return cmd & (LOCK_MAND | LOCK_RW);
#endif
1023 switch (cmd &~ LOCK_NB) {
1024 case LOCK_SH:
1025 return F_RDLCK;
1026 case LOCK_EX:
1027 return F_WRLCK;
1028 case LOCK_UN:
1029 return F_UNLCK;
}
1031 return -EINVAL;
}
/**
* __get_lease - revoke all outstanding leases on file
* @inode: the inode of the file to return
* @mode: the open mode (read or write)
*
* get_lease (inlined for speed) has checked there already
* is a lease on this file. Leases are broken on a call to open()
* or truncate(). This function can sleep unless you
* specified %O_NONBLOCK to your open().
*/
1044 int __get_lease(struct inode *inode, unsigned int mode)
{
int error = 0, future;
struct file_lock *new_fl, *flock;
struct file_lock *fl;
int alloc_err;
alloc_err = lease_alloc(NULL, 0, &new_fl);
1053 lock_kernel();
flock = inode->i_flock;
1055 if (flock->fl_type & F_INPROGRESS) {
if ((mode & O_NONBLOCK)
1057 || (flock->fl_owner == current->files)) {
error = -EWOULDBLOCK;
1059 goto out;
}
1061 if (alloc_err != 0) {
error = alloc_err;
1063 goto out;
}
1065 do {
error = locks_block_on(flock, new_fl);
1067 if (error != 0)
1068 goto out;
flock = inode->i_flock;
1070 if (!(flock && (flock->fl_flags & FL_LEASE)))
1071 goto out;
1072 } while (flock->fl_type & F_INPROGRESS);
}
1075 if (mode & FMODE_WRITE) {
/* If we want write access, we have to revoke any lease. */
future = F_UNLCK | F_INPROGRESS;
1078 } else if (flock->fl_type & F_WRLCK) {
/* Downgrade the exclusive lease to a read-only lease. */
future = F_RDLCK | F_INPROGRESS;
1081 } else {
/* the existing lease was read-only, so we can read too. */
1083 goto out;
}
1086 if (alloc_err && (flock->fl_owner != current->files)) {
error = alloc_err;
1088 goto out;
}
fl = flock;
1092 do {
fl->fl_type = future;
fl = fl->fl_next;
1095 } while (fl != NULL && (fl->fl_flags & FL_LEASE));
kill_fasync(&flock->fl_fasync, SIGIO, POLL_MSG);
1099 if ((mode & O_NONBLOCK) || (flock->fl_owner == current->files)) {
error = -EWOULDBLOCK;
1101 goto out;
}
1104 if (lease_break_time > 0)
error = lease_break_time * HZ;
1106 else
error = 0;
restart:
error = locks_block_on_timeout(flock, new_fl, error);
1110 if (error == 0) {
/* We timed out. Unilaterally break the lease. */
locks_delete_lock(&inode->i_flock, 0);
printk(KERN_WARNING "lease timed out\n");
1114 } else if (error > 0) {
flock = inode->i_flock;
1116 if (flock && (flock->fl_flags & FL_LEASE))
1117 goto restart;
error = 0;
}
out:
1122 unlock_kernel();
1123 if (!alloc_err)
locks_free_lock(new_fl);
1125 return error;
}
/**
* lease_get_mtime
* @inode: the inode
*
* This is to force NFS clients to flush their caches for files with
* exclusive leases. The justification is that if someone has an
* exclusive lease, then they could be modifiying it.
*/
1136 time_t lease_get_mtime(struct inode *inode)
{
struct file_lock *flock = inode->i_flock;
1139 if (flock && (flock->fl_flags & FL_LEASE) && (flock->fl_type & F_WRLCK))
1140 return CURRENT_TIME;
1141 return inode->i_mtime;
}
/**
* fcntl_getlease - Enquire what lease is currently active
* @filp: the file
*
* The value returned by this function will be one of
*
* %F_RDLCK to indicate a read-only (type II) lease is held.
*
* %F_WRLCK to indicate an exclusive lease is held.
*
* XXX: sfr & i disagree over whether F_INPROGRESS
* should be returned to userspace.
*/
1157 int fcntl_getlease(struct file *filp)
{
struct file_lock *fl;
fl = filp->f_dentry->d_inode->i_flock;
1162 if ((fl == NULL) || ((fl->fl_flags & FL_LEASE) == 0))
1163 return F_UNLCK;
1164 return fl->fl_type & ~F_INPROGRESS;
}
/* We already had a lease on this file; just change its type */
1168 static int lease_modify(struct file_lock **before, int arg, int fd, struct file *filp)
{
struct file_lock *fl = *before;
int error = assign_type(fl, arg);
1172 if (error < 0)
1173 goto out;
locks_wake_up_blocks(fl, 0);
1177 if (arg == F_UNLCK) {
filp->f_owner.pid = 0;
filp->f_owner.uid = 0;
filp->f_owner.euid = 0;
filp->f_owner.signum = 0;
locks_delete_lock(before, 0);
fasync_helper(fd, filp, 0, &fl->fl_fasync);
}
out:
1187 return error;
}
/**
* fcntl_setlease - sets a lease on an open file
* @fd: open file descriptor
* @filp: file pointer
* @arg: type of lease to obtain
*
* Call this fcntl to establish a lease on the file.
* Note that you also need to call %F_SETSIG to
* receive a signal when the lease is broken.
*/
1200 int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
{
struct file_lock *fl, **before, **my_before = NULL;
struct dentry *dentry;
struct inode *inode;
int error, rdlease_count = 0, wrlease_count = 0;
dentry = filp->f_dentry;
inode = dentry->d_inode;
1210 if ((current->fsuid != inode->i_uid) && !capable(CAP_LEASE))
1211 return -EACCES;
1212 if (!S_ISREG(inode->i_mode))
1213 return -EINVAL;
/*
* FIXME: What about F_RDLCK and files open for writing?
*/
if ((arg == F_WRLCK)
&& ((atomic_read(&dentry->d_count) > 1)
1220 || (atomic_read(&inode->i_count) > 1)))
1221 return -EAGAIN;
before = &inode->i_flock;
1225 lock_kernel();
1227 while ((fl = *before) != NULL) {
1228 if (fl->fl_flags != FL_LEASE)
1229 break;
1230 if (fl->fl_file == filp)
my_before = before;
1232 else if (fl->fl_type & F_WRLCK)
wrlease_count++;
1234 else
rdlease_count++;
before = &fl->fl_next;
}
if ((arg == F_RDLCK && (wrlease_count > 0)) ||
1240 (arg == F_WRLCK && ((rdlease_count + wrlease_count) > 0))) {
error = -EAGAIN;
1242 goto out_unlock;
}
1245 if (my_before != NULL) {
error = lease_modify(my_before, arg, fd, filp);
1247 goto out_unlock;
}
1250 if (arg == F_UNLCK) {
error = 0;
1252 goto out_unlock;
}
1255 if (!leases_enable) {
error = -EINVAL;
1257 goto out_unlock;
}
error = lease_alloc(filp, arg, &fl);
1261 if (error)
1262 goto out_unlock;
error = fasync_helper(fd, filp, 1, &fl->fl_fasync);
1265 if (error < 0) {
locks_free_lock(fl);
1267 goto out_unlock;
}
fl->fl_next = *before;
*before = fl;
list_add(&fl->fl_link, &file_lock_list);
filp->f_owner.pid = current->pid;
filp->f_owner.uid = current->uid;
filp->f_owner.euid = current->euid;
out_unlock:
1276 unlock_kernel();
1277 return error;
}
/**
* sys_flock: - flock() system call.
* @fd: the file descriptor to lock.
* @cmd: the type of lock to apply.
*
* Apply a %FL_FLOCK style lock to an open file descriptor.
* The @cmd can be one of
*
* %LOCK_SH -- a shared lock.
*
* %LOCK_EX -- an exclusive lock.
*
* %LOCK_UN -- remove an existing lock.
*
* %LOCK_MAND -- a `mandatory' flock. This exists to emulate Windows Share Modes.
*
* %LOCK_MAND can be combined with %LOCK_READ or %LOCK_WRITE to allow other
* processes read and write access respectively.
*/
1299 asmlinkage long sys_flock(unsigned int fd, unsigned int cmd)
{
struct file *filp;
int error, type;
error = -EBADF;
filp = fget(fd);
1306 if (!filp)
1307 goto out;
error = flock_translate_cmd(cmd);
1310 if (error < 0)
1311 goto out_putf;
type = error;
error = -EBADF;
if ((type != F_UNLCK)
#ifdef MSNFS
&& !(type & LOCK_MAND)
#endif
1319 && !(filp->f_mode & 3))
1320 goto out_putf;
1322 lock_kernel();
error = flock_lock_file(filp, type,
(cmd & (LOCK_UN | LOCK_NB)) ? 0 : 1);
1325 unlock_kernel();
out_putf:
fput(filp);
out:
1330 return error;
}
/* Report the first existing lock that would conflict with l.
* This implements the F_GETLK command of fcntl().
*/
1336 int fcntl_getlk(unsigned int fd, struct flock *l)
{
struct file *filp;
struct file_lock *fl, file_lock;
struct flock flock;
int error;
error = -EFAULT;
1344 if (copy_from_user(&flock, l, sizeof(flock)))
1345 goto out;
error = -EINVAL;
1347 if ((flock.l_type != F_RDLCK) && (flock.l_type != F_WRLCK))
1348 goto out;
error = -EBADF;
filp = fget(fd);
1352 if (!filp)
1353 goto out;
error = -EINVAL;
1356 if (!flock_to_posix_lock(filp, &file_lock, &flock))
1357 goto out_putf;
1359 if (filp->f_op && filp->f_op->lock) {
error = filp->f_op->lock(filp, F_GETLK, &file_lock);
1361 if (error < 0)
1362 goto out_putf;
1363 else if (error == LOCK_USE_CLNT)
/* Bypass for NFS with no locking - 2.0.36 compat */
fl = posix_test_lock(filp, &file_lock);
1366 else
fl = (file_lock.fl_type == F_UNLCK ? NULL : &file_lock);
1368 } else {
fl = posix_test_lock(filp, &file_lock);
}
flock.l_type = F_UNLCK;
1373 if (fl != NULL) {
flock.l_pid = fl->fl_pid;
#if BITS_PER_LONG == 32
/*
* Make sure we can represent the posix lock via
* legacy 32bit flock.
*/
error = -EOVERFLOW;
1381 if (fl->fl_start > OFFT_OFFSET_MAX)
1382 goto out_putf;
if ((fl->fl_end != OFFSET_MAX)
1384 && (fl->fl_end > OFFT_OFFSET_MAX))
1385 goto out_putf;
#endif
flock.l_start = fl->fl_start;
flock.l_len = fl->fl_end == OFFSET_MAX ? 0 :
fl->fl_end - fl->fl_start + 1;
flock.l_whence = 0;
flock.l_type = fl->fl_type;
}
error = -EFAULT;
1394 if (!copy_to_user(l, &flock, sizeof(flock)))
error = 0;
out_putf:
fput(filp);
out:
1400 return error;
}
/* Apply the lock described by l to an open file descriptor.
* This implements both the F_SETLK and F_SETLKW commands of fcntl().
*/
1406 int fcntl_setlk(unsigned int fd, unsigned int cmd, struct flock *l)
{
struct file *filp;
struct file_lock *file_lock = locks_alloc_lock(0);
struct flock flock;
struct inode *inode;
int error;
/*
* This might block, so we do it before checking the inode.
*/
error = -EFAULT;
1418 if (copy_from_user(&flock, l, sizeof(flock)))
1419 goto out;
/* Get arguments and validate them ...
*/
error = -EBADF;
filp = fget(fd);
1426 if (!filp)
1427 goto out;
error = -EINVAL;
inode = filp->f_dentry->d_inode;
/* Don't allow mandatory locks on files that may be memory mapped
* and shared.
*/
if (IS_MANDLOCK(inode) &&
1436 (inode->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID) {
struct address_space *mapping = inode->i_mapping;
1439 if (mapping->i_mmap_shared != NULL) {
error = -EAGAIN;
1441 goto out_putf;
}
}
error = -EINVAL;
1446 if (!flock_to_posix_lock(filp, file_lock, &flock))
1447 goto out_putf;
error = -EBADF;
1450 switch (flock.l_type) {
1451 case F_RDLCK:
1452 if (!(filp->f_mode & FMODE_READ))
1453 goto out_putf;
1454 break;
1455 case F_WRLCK:
1456 if (!(filp->f_mode & FMODE_WRITE))
1457 goto out_putf;
1458 break;
1459 case F_UNLCK:
1460 break;
1461 case F_SHLCK:
1462 case F_EXLCK:
#ifdef __sparc__
/* warn a bit for now, but don't overdo it */
{
static int count = 0;
if (!count) {
count=1;
printk(KERN_WARNING
"fcntl_setlk() called by process %d (%s) with broken flock() emulation\n",
current->pid, current->comm);
}
}
if (!(filp->f_mode & 3))
goto out_putf;
break;
#endif
1478 default:
error = -EINVAL;
1480 goto out_putf;
}
1483 if (filp->f_op && filp->f_op->lock != NULL) {
error = filp->f_op->lock(filp, cmd, file_lock);
1485 if (error < 0)
1486 goto out_putf;
}
error = posix_lock_file(filp, file_lock, cmd == F_SETLKW);
out_putf:
fput(filp);
out:
locks_free_lock(file_lock);
1494 return error;
}
#if BITS_PER_LONG == 32
/* Report the first existing lock that would conflict with l.
* This implements the F_GETLK command of fcntl().
*/
1501 int fcntl_getlk64(unsigned int fd, struct flock64 *l)
{
struct file *filp;
struct file_lock *fl, file_lock;
struct flock64 flock;
int error;
error = -EFAULT;
1509 if (copy_from_user(&flock, l, sizeof(flock)))
1510 goto out;
error = -EINVAL;
1512 if ((flock.l_type != F_RDLCK) && (flock.l_type != F_WRLCK))
1513 goto out;
error = -EBADF;
filp = fget(fd);
1517 if (!filp)
1518 goto out;
error = -EINVAL;
1521 if (!flock64_to_posix_lock(filp, &file_lock, &flock))
1522 goto out_putf;
1524 if (filp->f_op && filp->f_op->lock) {
error = filp->f_op->lock(filp, F_GETLK, &file_lock);
1526 if (error < 0)
1527 goto out_putf;
1528 else if (error == LOCK_USE_CLNT)
/* Bypass for NFS with no locking - 2.0.36 compat */
fl = posix_test_lock(filp, &file_lock);
1531 else
fl = (file_lock.fl_type == F_UNLCK ? NULL : &file_lock);
1533 } else {
fl = posix_test_lock(filp, &file_lock);
}
flock.l_type = F_UNLCK;
1538 if (fl != NULL) {
flock.l_pid = fl->fl_pid;
flock.l_start = fl->fl_start;
flock.l_len = fl->fl_end == OFFSET_MAX ? 0 :
fl->fl_end - fl->fl_start + 1;
flock.l_whence = 0;
flock.l_type = fl->fl_type;
}
error = -EFAULT;
1547 if (!copy_to_user(l, &flock, sizeof(flock)))
error = 0;
out_putf:
fput(filp);
out:
1553 return error;
}
/* Apply the lock described by l to an open file descriptor.
* This implements both the F_SETLK and F_SETLKW commands of fcntl().
*/
1559 int fcntl_setlk64(unsigned int fd, unsigned int cmd, struct flock64 *l)
{
struct file *filp;
struct file_lock *file_lock = locks_alloc_lock(0);
struct flock64 flock;
struct inode *inode;
int error;
/*
* This might block, so we do it before checking the inode.
*/
error = -EFAULT;
1571 if (copy_from_user(&flock, l, sizeof(flock)))
1572 goto out;
/* Get arguments and validate them ...
*/
error = -EBADF;
filp = fget(fd);
1579 if (!filp)
1580 goto out;
error = -EINVAL;
inode = filp->f_dentry->d_inode;
/* Don't allow mandatory locks on files that may be memory mapped
* and shared.
*/
if (IS_MANDLOCK(inode) &&
1589 (inode->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID) {
struct address_space *mapping = inode->i_mapping;
1592 if (mapping->i_mmap_shared != NULL) {
error = -EAGAIN;
1594 goto out_putf;
}
}
error = -EINVAL;
1599 if (!flock64_to_posix_lock(filp, file_lock, &flock))
1600 goto out_putf;
error = -EBADF;
1603 switch (flock.l_type) {
1604 case F_RDLCK:
1605 if (!(filp->f_mode & FMODE_READ))
1606 goto out_putf;
1607 break;
1608 case F_WRLCK:
1609 if (!(filp->f_mode & FMODE_WRITE))
1610 goto out_putf;
1611 break;
1612 case F_UNLCK:
1613 break;
1614 case F_SHLCK:
1615 case F_EXLCK:
1616 default:
error = -EINVAL;
1618 goto out_putf;
}
1621 if (filp->f_op && filp->f_op->lock != NULL) {
error = filp->f_op->lock(filp, cmd, file_lock);
1623 if (error < 0)
1624 goto out_putf;
}
error = posix_lock_file(filp, file_lock, cmd == F_SETLKW64);
out_putf:
fput(filp);
out:
locks_free_lock(file_lock);
1632 return error;
}
#endif /* BITS_PER_LONG == 32 */
/*
* This function is called when the file is being removed
* from the task's fd array.
*/
1640 void locks_remove_posix(struct file *filp, fl_owner_t owner)
{
struct inode * inode = filp->f_dentry->d_inode;
struct file_lock *fl;
struct file_lock **before;
/*
* For POSIX locks we free all locks on this file for the given task.
*/
1649 if (!inode->i_flock) {
/*
* Notice that something might be grabbing a lock right now.
* Consider it as a race won by us - event is async, so even if
* we miss the lock added we can trivially consider it as added
* after we went through this call.
*/
1656 return;
}
1658 lock_kernel();
before = &inode->i_flock;
1660 while ((fl = *before) != NULL) {
1661 if ((fl->fl_flags & FL_POSIX) && fl->fl_owner == owner) {
locks_unlock_delete(before);
1663 continue;
}
before = &fl->fl_next;
}
1667 unlock_kernel();
}
/*
* This function is called on the last close of an open file.
*/
1673 void locks_remove_flock(struct file *filp)
{
struct inode * inode = filp->f_dentry->d_inode;
struct file_lock *fl;
struct file_lock **before;
1679 if (!inode->i_flock)
1680 return;
1682 lock_kernel();
before = &inode->i_flock;
1685 while ((fl = *before) != NULL) {
if ((fl->fl_flags & (FL_FLOCK|FL_LEASE))
1687 && (fl->fl_file == filp)) {
locks_delete_lock(before, 0);
1689 continue;
}
before = &fl->fl_next;
}
1693 unlock_kernel();
}
/**
* posix_block_lock - blocks waiting for a file lock
* @blocker: the lock which is blocking
* @waiter: the lock which conflicts and has to wait
*
* lockd needs to block waiting for locks.
*/
void
1704 posix_block_lock(struct file_lock *blocker, struct file_lock *waiter)
{
locks_insert_block(blocker, waiter);
}
/**
* posix_unblock_lock - stop waiting for a file lock
* @waiter: the lock which was waiting
*
* lockd needs to block waiting for locks.
*/
void
1716 posix_unblock_lock(struct file_lock *waiter)
{
1718 if (!list_empty(&waiter->fl_block))
locks_delete_block(waiter);
}
1722 static void lock_get_status(char* out, struct file_lock *fl, int id, char *pfx)
{
struct inode *inode = NULL;
1726 if (fl->fl_file != NULL)
inode = fl->fl_file->f_dentry->d_inode;
out += sprintf(out, "%d:%s ", id, pfx);
1730 if (fl->fl_flags & FL_POSIX) {
out += sprintf(out, "%6s %s ",
(fl->fl_flags & FL_ACCESS) ? "ACCESS" : "POSIX ",
(inode == NULL) ? "*NOINODE*" :
(IS_MANDLOCK(inode) &&
(inode->i_mode & (S_IXGRP | S_ISGID)) == S_ISGID) ?
"MANDATORY" : "ADVISORY ");
1737 } else if (fl->fl_flags & FL_FLOCK) {
#ifdef MSNFS
1739 if (fl->fl_type & LOCK_MAND) {
out += sprintf(out, "FLOCK MSNFS ");
1741 } else
#endif
out += sprintf(out, "FLOCK ADVISORY ");
1744 } else if (fl->fl_flags & FL_LEASE) {
out += sprintf(out, "LEASE MANDATORY ");
1746 } else {
out += sprintf(out, "UNKNOWN UNKNOWN ");
}
#ifdef MSNFS
1750 if (fl->fl_type & LOCK_MAND) {
out += sprintf(out, "%s ",
(fl->fl_type & LOCK_READ)
? (fl->fl_type & LOCK_WRITE) ? "RW " : "READ "
: (fl->fl_type & LOCK_WRITE) ? "WRITE" : "NONE ");
1755 } else
#endif
out += sprintf(out, "%s ",
(fl->fl_type & F_WRLCK) ? "WRITE" : "READ ");
out += sprintf(out, "%d %s:%ld ",
fl->fl_pid,
inode ? kdevname(inode->i_dev) : "<none>",
inode ? inode->i_ino : 0);
out += sprintf(out, "%Ld ", fl->fl_start);
1764 if (fl->fl_end == OFFSET_MAX)
out += sprintf(out, "EOF ");
1766 else
out += sprintf(out, "%Ld ", fl->fl_end);
sprintf(out, "%08lx %08lx %08lx %08lx %08lx\n",
(long)fl, (long)fl->fl_link.prev, (long)fl->fl_link.next,
(long)fl->fl_next, (long)fl->fl_block.next);
}
1773 static void move_lock_status(char **p, off_t* pos, off_t offset)
{
int len;
len = strlen(*p);
1777 if(*pos >= offset) {
/* the complete line is valid */
*p += len;
*pos += len;
1781 return;
}
1783 if(*pos+len > offset) {
/* use the second part of the line */
int i = offset-*pos;
memmove(*p,*p+i,len-i);
*p += len-i;
*pos += len;
1789 return;
}
/* discard the complete line */
*pos += len;
}
/**
* get_locks_status - reports lock usage in /proc/locks
* @buffer: address in userspace to write into
* @start: ?
* @offset: how far we are through the buffer
* @length: how much to read
*/
1803 int get_locks_status(char *buffer, char **start, off_t offset, int length)
{
struct list_head *tmp;
char *q = buffer;
off_t pos = 0;
int i = 0;
1810 lock_kernel();
1811 list_for_each(tmp, &file_lock_list) {
struct list_head *btmp;
struct file_lock *fl = list_entry(tmp, struct file_lock, fl_link);
lock_get_status(q, fl, ++i, "");
move_lock_status(&q, &pos, offset);
1817 if(pos >= offset+length)
1818 goto done;
1820 list_for_each(btmp, &fl->fl_block) {
struct file_lock *bfl = list_entry(btmp,
struct file_lock, fl_block);
lock_get_status(q, bfl, i, " ->");
move_lock_status(&q, &pos, offset);
1826 if(pos >= offset+length)
1827 goto done;
}
}
done:
1831 unlock_kernel();
*start = buffer;
1833 if(q-buffer < length)
1834 return (q-buffer);
1835 return length;
}
#ifdef MSNFS
/**
* lock_may_read - checks that the region is free of locks
* @inode: the inode that is being read
* @start: the first byte to read
* @len: the number of bytes to read
*
* Emulates Windows locking requirements. Whole-file
* mandatory locks (share modes) can prohibit a read and
* byte-range POSIX locks can prohibit a read if they overlap.
*
* N.B. this function is only ever called
* from knfsd and ownership of locks is never checked.
*/
1852 int lock_may_read(struct inode *inode, loff_t start, unsigned long len)
{
struct file_lock *fl;
int result = 1;
1856 lock_kernel();
1857 for (fl = inode->i_flock; fl != NULL; fl = fl->fl_next) {
1858 if (fl->fl_flags == FL_POSIX) {
1859 if (fl->fl_type == F_RDLCK)
1860 continue;
1861 if ((fl->fl_end < start) || (fl->fl_start > (start + len)))
1862 continue;
1863 } else if (fl->fl_flags == FL_FLOCK) {
1864 if (!(fl->fl_type & LOCK_MAND))
1865 continue;
1866 if (fl->fl_type & LOCK_READ)
1867 continue;
1868 } else
1869 continue;
result = 0;
1871 break;
}
1873 unlock_kernel();
1874 return result;
}
/**
* lock_may_write - checks that the region is free of locks
* @inode: the inode that is being written
* @start: the first byte to write
* @len: the number of bytes to write
*
* Emulates Windows locking requirements. Whole-file
* mandatory locks (share modes) can prohibit a write and
* byte-range POSIX locks can prohibit a write if they overlap.
*
* N.B. this function is only ever called
* from knfsd and ownership of locks is never checked.
*/
1890 int lock_may_write(struct inode *inode, loff_t start, unsigned long len)
{
struct file_lock *fl;
int result = 1;
1894 lock_kernel();
1895 for (fl = inode->i_flock; fl != NULL; fl = fl->fl_next) {
1896 if (fl->fl_flags == FL_POSIX) {
1897 if ((fl->fl_end < start) || (fl->fl_start > (start + len)))
1898 continue;
1899 } else if (fl->fl_flags == FL_FLOCK) {
1900 if (!(fl->fl_type & LOCK_MAND))
1901 continue;
1902 if (fl->fl_type & LOCK_WRITE)
1903 continue;
1904 } else
1905 continue;
result = 0;
1907 break;
}
1909 unlock_kernel();
1910 return result;
}
#endif
1914 static int __init filelock_init(void)
{
filelock_cache = kmem_cache_create("file lock cache",
sizeof(struct file_lock), 0, 0, init_once, NULL);
1918 if (!filelock_cache)
panic("cannot create file lock slab cache");
1920 return 0;
}
module_init(filelock_init)