summaryrefslogtreecommitdiff
path: root/source4/ntvfs/posix/pvfs_lock.c
blob: 9b5823e69e71f607b8608f0c2cb7751e1e0072e0 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
/* 
   Unix SMB/CIFS implementation.

   POSIX NTVFS backend - locking

   Copyright (C) Andrew Tridgell 2004

   This program is free software; you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation; either version 2 of the License, or
   (at your option) any later version.
   
   This program is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   GNU General Public License for more details.
   
   You should have received a copy of the GNU General Public License
   along with this program; if not, write to the Free Software
   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
*/

#include "includes.h"
#include "vfs_posix.h"
#include "system/time.h"
#include "lib/util/dlinklist.h"
#include "messaging/messaging.h"


/*
  check if we can perform IO on a range that might be locked
*/
NTSTATUS pvfs_check_lock(struct pvfs_state *pvfs,
			 struct pvfs_file *f,
			 uint16_t smbpid,
			 uint64_t offset, uint64_t count,
			 enum brl_type rw)
{
	if (!(pvfs->flags & PVFS_FLAG_STRICT_LOCKING)) {
		return NT_STATUS_OK;
	}

	return brl_locktest(pvfs->brl_context,
			    f->brl_handle,
			    smbpid,
			    offset, count, rw);
}

/* this state structure holds information about a lock we are waiting on */
struct pvfs_pending_lock {
	struct pvfs_pending_lock *next, *prev;
	struct pvfs_state *pvfs;
	union smb_lock *lck;
	struct pvfs_file *f;
	struct ntvfs_request *req;
	int pending_lock;
	void *wait_handle;
	struct timeval end_time;
};

/*
  a secondary attempt to setup a lock has failed - back out
  the locks we did get and send an error
*/
static void pvfs_lock_async_failed(struct pvfs_state *pvfs,
				   struct ntvfs_request *req,
				   struct pvfs_file *f,
				   struct smb_lock_entry *locks,
				   int i,
				   NTSTATUS status)
{
	/* in SMB2 mode we also try to unlock failing lock */ 
	if (req->ctx->protocol != PROTOCOL_SMB2) {
		i--;
	}

	/* undo the locks we just did */
	for (;i>=0;i--) {
		brl_unlock(pvfs->brl_context,
			   f->brl_handle,
			   locks[i].pid,
			   locks[i].offset,
			   locks[i].count);
		f->lock_count--;
	}
	req->async_states->status = status;
	req->async_states->send_fn(req);
}


/*
  called when we receive a pending lock notification. It means that
  either our lock timed out or someone else has unlocked a overlapping
  range, so we should try the lock again. Note that on timeout we
  do retry the lock, giving it a last chance.
*/
static void pvfs_pending_lock_continue(void *private, enum pvfs_wait_notice reason)
{
	struct pvfs_pending_lock *pending = private;
	struct pvfs_state *pvfs = pending->pvfs;
	struct pvfs_file *f = pending->f;
	struct ntvfs_request *req = pending->req;
	union smb_lock *lck = pending->lck;
	struct smb_lock_entry *locks;
	enum brl_type rw;
	NTSTATUS status;
	int i;
	BOOL timed_out;

	timed_out = (reason != PVFS_WAIT_EVENT);

	locks = lck->lockx.in.locks + lck->lockx.in.ulock_cnt;

	if (lck->lockx.in.mode & LOCKING_ANDX_SHARED_LOCK) {
		rw = READ_LOCK;
	} else {
		rw = WRITE_LOCK;
	}

	DLIST_REMOVE(f->pending_list, pending);

	/* we don't retry on a cancel */
	if (reason == PVFS_WAIT_CANCEL) {
		status = NT_STATUS_FILE_LOCK_CONFLICT;
	} else {
		/* 
		 * here it's important to pass the pending pointer
		 * because with this we'll get the correct error code
		 * FILE_LOCK_CONFLICT in the error case
		 */
		status = brl_lock(pvfs->brl_context,
				  f->brl_handle,
				  locks[pending->pending_lock].pid,
				  locks[pending->pending_lock].offset,
				  locks[pending->pending_lock].count,
				  rw, pending);
	}
	if (NT_STATUS_IS_OK(status)) {
		f->lock_count++;
		timed_out = False;
	}

	/* if we have failed and timed out, or succeeded, then we
	   don't need the pending lock any more */
	if (NT_STATUS_IS_OK(status) || timed_out) {
		NTSTATUS status2;
		status2 = brl_remove_pending(pvfs->brl_context, 
					     f->brl_handle, pending);
		if (!NT_STATUS_IS_OK(status2)) {
			DEBUG(0,("pvfs_lock: failed to remove pending lock - %s\n", nt_errstr(status2)));
		}
		talloc_free(pending->wait_handle);
	}

	if (!NT_STATUS_IS_OK(status)) {
		if (timed_out) {
			/* no more chances */
			pvfs_lock_async_failed(pvfs, req, f, locks, pending->pending_lock, status);
		} else {
			/* we can try again */
			DLIST_ADD(f->pending_list, pending);
		}
		return;
	}

	/* if we haven't timed out yet, then we can do more pending locks */
	if (rw == READ_LOCK) {
		rw = PENDING_READ_LOCK;
	} else {
		rw = PENDING_WRITE_LOCK;
	}

	/* we've now got the pending lock. try and get the rest, which might
	   lead to more pending locks */
	for (i=pending->pending_lock+1;i<lck->lockx.in.lock_cnt;i++) {		
		if (pending) {
			pending->pending_lock = i;
		}

		status = brl_lock(pvfs->brl_context,
				  f->brl_handle,
				  locks[i].pid,
				  locks[i].offset,
				  locks[i].count,
				  rw, pending);
		if (!NT_STATUS_IS_OK(status)) {
			if (pending) {
				/* a timed lock failed - setup a wait message to handle
				   the pending lock notification or a timeout */
				pending->wait_handle = pvfs_wait_message(pvfs, req, MSG_BRL_RETRY, 
									 pending->end_time,
									 pvfs_pending_lock_continue,
									 pending);
				if (pending->wait_handle == NULL) {
					pvfs_lock_async_failed(pvfs, req, f, locks, i, NT_STATUS_NO_MEMORY);
				} else {
					talloc_steal(pending, pending->wait_handle);
					DLIST_ADD(f->pending_list, pending);
				}
				return;
			}
			pvfs_lock_async_failed(pvfs, req, f, locks, i, status);
			return;
		}

		f->lock_count++;
	}

	/* we've managed to get all the locks. Tell the client */
	req->async_states->status = NT_STATUS_OK;
	req->async_states->send_fn(req);
}


/*
  called when we close a file that might have locks
*/
void pvfs_lock_close(struct pvfs_state *pvfs, struct pvfs_file *f)
{
	struct pvfs_pending_lock *p, *next;

	if (f->lock_count || f->pending_list) {
		DEBUG(5,("pvfs_lock: removing %.0f locks on close\n", 
			 (double)f->lock_count));
		brl_close(f->pvfs->brl_context, f->brl_handle);
		f->lock_count = 0;
	}

	/* reply to all the pending lock requests, telling them the 
	   lock failed */
	for (p=f->pending_list;p;p=next) {
		next = p->next;
		DLIST_REMOVE(f->pending_list, p);
		p->req->async_states->status = NT_STATUS_RANGE_NOT_LOCKED;
		p->req->async_states->send_fn(p->req);
	}
}


/*
  cancel a set of locks
*/
static NTSTATUS pvfs_lock_cancel(struct pvfs_state *pvfs, struct ntvfs_request *req, union smb_lock *lck,
				 struct pvfs_file *f)
{
	struct pvfs_pending_lock *p;

	for (p=f->pending_list;p;p=p->next) {
		/* check if the lock request matches exactly - you can only cancel with exact matches */
		if (p->lck->lockx.in.ulock_cnt == lck->lockx.in.ulock_cnt &&
		    p->lck->lockx.in.lock_cnt  == lck->lockx.in.lock_cnt &&
		    p->lck->lockx.in.file.ntvfs== lck->lockx.in.file.ntvfs &&
		    p->lck->lockx.in.mode      == (lck->lockx.in.mode & ~LOCKING_ANDX_CANCEL_LOCK)) {
			int i;

			for (i=0;i<lck->lockx.in.ulock_cnt + lck->lockx.in.lock_cnt;i++) {
				if (p->lck->lockx.in.locks[i].pid != lck->lockx.in.locks[i].pid ||
				    p->lck->lockx.in.locks[i].offset != lck->lockx.in.locks[i].offset ||
				    p->lck->lockx.in.locks[i].count != lck->lockx.in.locks[i].count) {
					break;
				}
			}
			if (i < lck->lockx.in.ulock_cnt) continue;

			/* an exact match! we can cancel it, which is equivalent
			   to triggering the timeout early */
			pvfs_pending_lock_continue(p, PVFS_WAIT_TIMEOUT);
			return NT_STATUS_OK;
		}
	}

	return NT_STATUS_DOS(ERRDOS, ERRcancelviolation);
}


/*
  lock or unlock a byte range
*/
NTSTATUS pvfs_lock(struct ntvfs_module_context *ntvfs,
		   struct ntvfs_request *req, union smb_lock *lck)
{
	struct pvfs_state *pvfs = ntvfs->private_data;
	struct pvfs_file *f;
	struct smb_lock_entry *locks;
	int i;
	enum brl_type rw;
	struct pvfs_pending_lock *pending = NULL;
	NTSTATUS status;

	if (lck->generic.level != RAW_LOCK_GENERIC) {
		return ntvfs_map_lock(ntvfs, req, lck);
	}

	f = pvfs_find_fd(pvfs, req, lck->lockx.in.file.ntvfs);
	if (!f) {
		return NT_STATUS_INVALID_HANDLE;
	}

	if (f->handle->fd == -1) {
		return NT_STATUS_FILE_IS_A_DIRECTORY;
	}

	if (lck->lockx.in.timeout != 0 && 
	    (req->async_states->state & NTVFS_ASYNC_STATE_MAY_ASYNC)) {
		pending = talloc(f, struct pvfs_pending_lock);
		if (pending == NULL) {
			return NT_STATUS_NO_MEMORY;
		}

		pending->pvfs = pvfs;
		pending->lck = lck;
		pending->f = f;
		pending->req = req;

		pending->end_time = 
			timeval_current_ofs(lck->lockx.in.timeout/1000,
					    1000*(lck->lockx.in.timeout%1000));
	}

	if (lck->lockx.in.mode & LOCKING_ANDX_SHARED_LOCK) {
		rw = pending? PENDING_READ_LOCK : READ_LOCK;
	} else {
		rw = pending? PENDING_WRITE_LOCK : WRITE_LOCK;
	}

	if (lck->lockx.in.mode & LOCKING_ANDX_CANCEL_LOCK) {
		return pvfs_lock_cancel(pvfs, req, lck, f);
	}

	if (lck->lockx.in.mode & LOCKING_ANDX_CHANGE_LOCKTYPE) {
		/* this seems to not be supported by any windows server,
		   or used by any clients */
		return NT_STATUS_DOS(ERRDOS, ERRnoatomiclocks);
	}

	if (lck->lockx.in.mode & LOCKING_ANDX_OPLOCK_RELEASE) {
		DEBUG(0,("received unexpected oplock break\n"));
		return NT_STATUS_NOT_IMPLEMENTED;
	}


	/* the unlocks happen first */
	locks = lck->lockx.in.locks;

	for (i=0;i<lck->lockx.in.ulock_cnt;i++) {
		status = brl_unlock(pvfs->brl_context,
				    f->brl_handle,
				    locks[i].pid,
				    locks[i].offset,
				    locks[i].count);
		if (!NT_STATUS_IS_OK(status)) {
			return status;
		}
		f->lock_count--;
	}

	locks += i;

	for (i=0;i<lck->lockx.in.lock_cnt;i++) {
		if (pending) {
			pending->pending_lock = i;
		}

		status = brl_lock(pvfs->brl_context,
				  f->brl_handle,
				  locks[i].pid,
				  locks[i].offset,
				  locks[i].count,
				  rw, pending);
		if (!NT_STATUS_IS_OK(status)) {
			if (pending) {
				/* a timed lock failed - setup a wait message to handle
				   the pending lock notification or a timeout */
				pending->wait_handle = pvfs_wait_message(pvfs, req, MSG_BRL_RETRY, 
									 pending->end_time,
									 pvfs_pending_lock_continue,
									 pending);
				if (pending->wait_handle == NULL) {
					return NT_STATUS_NO_MEMORY;
				}
				talloc_steal(pending, pending->wait_handle);
				DLIST_ADD(f->pending_list, pending);
				return NT_STATUS_OK;
			}
			/* in SMB2 mode we also try to unlock failing lock */ 
			if (req->ctx->protocol != PROTOCOL_SMB2) {
				i--;
			}
			/* undo the locks we just did */
			for (;i>=0;i--) {
				brl_unlock(pvfs->brl_context,
					   f->brl_handle,
					   locks[i].pid,
					   locks[i].offset,
					   locks[i].count);
				f->lock_count--;
			}
			return status;
		}
		f->lock_count++;
	}

	return NT_STATUS_OK;
}