summaryrefslogtreecommitdiff
path: root/source3/smbd/oplock.c
blob: 63f88b782456e889cb30cd3b9c434881a72af972 (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
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
/* 
   Unix SMB/CIFS implementation.
   oplock processing
   Copyright (C) Andrew Tridgell 1992-1998
   Copyright (C) Jeremy Allison 1998 - 2001
   Copyright (C) Volker Lendecke 2005
   
   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.
*/

#define DBGC_CLASS DBGC_LOCKING
#include "includes.h"

/* Current number of oplocks we have outstanding. */
static int32 exclusive_oplocks_open = 0;
static int32 level_II_oplocks_open = 0;
BOOL global_client_failed_oplock_break = False;

extern uint32 global_client_caps;
extern int smb_read_error;

static struct kernel_oplocks *koplocks;

/****************************************************************************
 Get the number of current exclusive oplocks.
****************************************************************************/

int32 get_number_of_exclusive_open_oplocks(void)
{
  return exclusive_oplocks_open;
}

/****************************************************************************
 Return True if an oplock message is pending.
****************************************************************************/

BOOL oplock_message_waiting(fd_set *fds)
{
	if (koplocks && koplocks->msg_waiting(fds)) {
		return True;
	}

	return False;
}

/****************************************************************************
 Find out if there are any kernel oplock messages waiting and process them
 if so. pfds is the fd_set from the main select loop (which contains any
 kernel oplock fd if that's what the system uses (IRIX). If may be NULL if
 we're calling this in a shutting down state.
****************************************************************************/

void process_kernel_oplocks(struct messaging_context *msg_ctx, fd_set *pfds)
{
	/*
	 * We need to check for kernel oplocks before going into the select
	 * here, as the EINTR generated by the linux kernel oplock may have
	 * already been eaten. JRA.
	 */

	if (!koplocks) {
		return;
	}

	while (koplocks->msg_waiting(pfds)) { 
		files_struct *fsp;
		char msg[MSG_SMB_KERNEL_BREAK_SIZE];

		fsp = koplocks->receive_message(pfds);

		if (fsp == NULL) {
			DEBUG(3, ("Kernel oplock message announced, but none "
				  "received\n"));
			return;
		}

		/* Put the kernel break info into the message. */
		SDEV_T_VAL(msg,0,fsp->dev);
		SINO_T_VAL(msg,8,fsp->inode);
		SIVAL(msg,16,fsp->fh->file_id);

		/* Don't need to be root here as we're only ever
		   sending to ourselves. */

		messaging_send_buf(msg_ctx, procid_self(),
				   MSG_SMB_KERNEL_BREAK,
				   (uint8 *)&msg, MSG_SMB_KERNEL_BREAK_SIZE);
	}
}

/****************************************************************************
 Attempt to set an oplock on a file. Always succeeds if kernel oplocks are
 disabled (just sets flags). Returns True if oplock set.
****************************************************************************/

BOOL set_file_oplock(files_struct *fsp, int oplock_type)
{
	if (koplocks && !koplocks->set_oplock(fsp, oplock_type)) {
		return False;
	}

	fsp->oplock_type = oplock_type;
	fsp->sent_oplock_break = NO_BREAK_SENT;
	if (oplock_type == LEVEL_II_OPLOCK) {
		level_II_oplocks_open++;
	} else {
		exclusive_oplocks_open++;
	}

	DEBUG(5,("set_file_oplock: granted oplock on file %s, 0x%x/%.0f/%lu, "
		    "tv_sec = %x, tv_usec = %x\n",
		 fsp->fsp_name, (unsigned int)fsp->dev, (double)fsp->inode,
		 fsp->fh->file_id, (int)fsp->open_time.tv_sec,
		 (int)fsp->open_time.tv_usec ));

	return True;
}

/****************************************************************************
 Attempt to release an oplock on a file. Decrements oplock count.
****************************************************************************/

void release_file_oplock(files_struct *fsp)
{
	if ((fsp->oplock_type != NO_OPLOCK) &&
	    (fsp->oplock_type != FAKE_LEVEL_II_OPLOCK) &&
	    koplocks) {
		koplocks->release_oplock(fsp);
	}

	if (fsp->oplock_type == LEVEL_II_OPLOCK) {
		level_II_oplocks_open--;
	} else if (EXCLUSIVE_OPLOCK_TYPE(fsp->oplock_type)) {
		exclusive_oplocks_open--;
	}

	SMB_ASSERT(exclusive_oplocks_open>=0);
	SMB_ASSERT(level_II_oplocks_open>=0);
	
	fsp->oplock_type = NO_OPLOCK;
	fsp->sent_oplock_break = NO_BREAK_SENT;
	
	flush_write_cache(fsp, OPLOCK_RELEASE_FLUSH);
}

/****************************************************************************
 Attempt to downgrade an oplock on a file. Doesn't decrement oplock count.
****************************************************************************/

static void downgrade_file_oplock(files_struct *fsp)
{
	if (koplocks) {
		koplocks->release_oplock(fsp);
	}
	fsp->oplock_type = LEVEL_II_OPLOCK;
	exclusive_oplocks_open--;
	level_II_oplocks_open++;
	fsp->sent_oplock_break = NO_BREAK_SENT;
}

/****************************************************************************
 Remove a file oplock. Copes with level II and exclusive.
 Locks then unlocks the share mode lock. Client can decide to go directly
 to none even if a "break-to-level II" was sent.
****************************************************************************/

BOOL remove_oplock(files_struct *fsp)
{
	SMB_DEV_T dev = fsp->dev;
	SMB_INO_T inode = fsp->inode;
	BOOL ret;
	struct share_mode_lock *lck;

	/* Remove the oplock flag from the sharemode. */
	lck = get_share_mode_lock(NULL, fsp->dev, fsp->inode, NULL, NULL);
	if (lck == NULL) {
		DEBUG(0,("remove_oplock: failed to lock share entry for "
			 "file %s\n", fsp->fsp_name ));
		return False;
	}
	ret = remove_share_oplock(lck, fsp);
	if (!ret) {
		DEBUG(0,("remove_oplock: failed to remove share oplock for "
			 "file %s fnum %d, 0x%x/%.0f\n",
			 fsp->fsp_name, fsp->fnum, (unsigned int)dev,
			 (double)inode));
	}
	release_file_oplock(fsp);
	TALLOC_FREE(lck);
	return ret;
}

/*
 * Deal with a reply when a break-to-level II was sent.
 */
BOOL downgrade_oplock(files_struct *fsp)
{
	SMB_DEV_T dev = fsp->dev;
	SMB_INO_T inode = fsp->inode;
	BOOL ret;
	struct share_mode_lock *lck;

	lck = get_share_mode_lock(NULL, fsp->dev, fsp->inode, NULL, NULL);
	if (lck == NULL) {
		DEBUG(0,("downgrade_oplock: failed to lock share entry for "
			 "file %s\n", fsp->fsp_name ));
		return False;
	}
	ret = downgrade_share_oplock(lck, fsp);
	if (!ret) {
		DEBUG(0,("downgrade_oplock: failed to downgrade share oplock "
			 "for file %s fnum %d, dev = %x, inode = %.0f\n",
			 fsp->fsp_name, fsp->fnum, (unsigned int)dev,
			 (double)inode));
	}

	downgrade_file_oplock(fsp);
	TALLOC_FREE(lck);
	return ret;
}

/****************************************************************************
 Return the fd (if any) used for receiving oplock notifications.
****************************************************************************/

int oplock_notify_fd(void)
{
	if (koplocks) {
		return koplocks->notification_fd;
	}

	return -1;
}

/****************************************************************************
 Set up an oplock break message.
****************************************************************************/

static char *new_break_smb_message(TALLOC_CTX *mem_ctx,
				   files_struct *fsp, uint8 cmd)
{
	char *result = TALLOC_ARRAY(mem_ctx, char, smb_size + 8*2 + 0);

	if (result == NULL) {
		DEBUG(0, ("talloc failed\n"));
		return NULL;
	}

	memset(result,'\0',smb_size);
	set_message(NULL,result,8,0,True);
	SCVAL(result,smb_com,SMBlockingX);
	SSVAL(result,smb_tid,fsp->conn->cnum);
	SSVAL(result,smb_pid,0xFFFF);
	SSVAL(result,smb_uid,0);
	SSVAL(result,smb_mid,0xFFFF);
	SCVAL(result,smb_vwv0,0xFF);
	SSVAL(result,smb_vwv2,fsp->fnum);
	SCVAL(result,smb_vwv3,LOCKING_ANDX_OPLOCK_RELEASE);
	SCVAL(result,smb_vwv3+1,cmd);
	return result;
}

/****************************************************************************
 Function to do the waiting before sending a local break.
****************************************************************************/

static void wait_before_sending_break(void)
{
	long wait_time = (long)lp_oplock_break_wait_time();

	if (wait_time) {
		smb_msleep(wait_time);
	}
}

/****************************************************************************
 Ensure that we have a valid oplock.
****************************************************************************/

static files_struct *initial_break_processing(SMB_DEV_T dev, SMB_INO_T inode, unsigned long file_id)
{
	files_struct *fsp = NULL;

	if( DEBUGLVL( 3 ) ) {
		dbgtext( "initial_break_processing: called for 0x%x/%.0f/%u\n",
			(unsigned int)dev, (double)inode, (int)file_id);
		dbgtext( "Current oplocks_open (exclusive = %d, levelII = %d)\n",
			exclusive_oplocks_open, level_II_oplocks_open );
	}

	/*
	 * We need to search the file open table for the
	 * entry containing this dev and inode, and ensure
	 * we have an oplock on it.
	 */

	fsp = file_find_dif(dev, inode, file_id);

	if(fsp == NULL) {
		/* The file could have been closed in the meantime - return success. */
		if( DEBUGLVL( 3 ) ) {
			dbgtext( "initial_break_processing: cannot find open file with " );
			dbgtext( "dev = 0x%x, inode = %.0f file_id = %lu", (unsigned int)dev,
				(double)inode, file_id);
			dbgtext( "allowing break to succeed.\n" );
		}
		return NULL;
	}

	/* Ensure we have an oplock on the file */

	/*
	 * There is a potential race condition in that an oplock could
	 * have been broken due to another udp request, and yet there are
	 * still oplock break messages being sent in the udp message
	 * queue for this file. So return true if we don't have an oplock,
	 * as we may have just freed it.
	 */

	if(fsp->oplock_type == NO_OPLOCK) {
		if( DEBUGLVL( 3 ) ) {
			dbgtext( "initial_break_processing: file %s ", fsp->fsp_name );
			dbgtext( "(dev = %x, inode = %.0f, file_id = %lu) has no oplock.\n",
				(unsigned int)dev, (double)inode, fsp->fh->file_id );
			dbgtext( "Allowing break to succeed regardless.\n" );
		}
		return NULL;
	}

	return fsp;
}

static void oplock_timeout_handler(struct event_context *ctx,
				   struct timed_event *te,
				   const struct timeval *now,
				   void *private_data)
{
	files_struct *fsp = (files_struct *)private_data;

	/* Ensure we always remove this event. */
	if (fsp->oplock_timeout != NULL) {
		/* Remove the timed event handler. */
		TALLOC_FREE(fsp->oplock_timeout);
		fsp->oplock_timeout = NULL;
	}
	DEBUG(0, ("Oplock break failed for file %s -- replying anyway\n", fsp->fsp_name));
	global_client_failed_oplock_break = True;
	remove_oplock(fsp);
	reply_to_oplock_break_requests(fsp);
}

/*******************************************************************
 Add a timeout handler waiting for the client reply.
*******************************************************************/

static void add_oplock_timeout_handler(files_struct *fsp)
{
	if (fsp->oplock_timeout != NULL) {
		DEBUG(0, ("Logic problem -- have an oplock event hanging "
			  "around\n"));
	}

	fsp->oplock_timeout =
		event_add_timed(smbd_event_context(), NULL,
				timeval_current_ofs(OPLOCK_BREAK_TIMEOUT, 0),
				"oplock_timeout_handler",
				oplock_timeout_handler, fsp);

	if (fsp->oplock_timeout == NULL) {
		DEBUG(0, ("Could not add oplock timeout handler\n"));
	}
}

/*******************************************************************
 This handles the case of a write triggering a break to none
 message on a level2 oplock.
 When we get this message we may be in any of three states :
 NO_OPLOCK, LEVEL_II, FAKE_LEVEL2. We only send a message to
 the client for LEVEL2.
*******************************************************************/

static void process_oplock_async_level2_break_message(struct messaging_context *msg_ctx,
						      void *private_data,
						      uint32_t msg_type,
						      struct server_id src,
						      DATA_BLOB *data)
{
	struct share_mode_entry msg;
	files_struct *fsp;
	char *break_msg;
	BOOL sign_state;

	if (data->data == NULL) {
		DEBUG(0, ("Got NULL buffer\n"));
		return;
	}

	if (data->length != MSG_SMB_SHARE_MODE_ENTRY_SIZE) {
		DEBUG(0, ("Got invalid msg len %d\n", (int)data->length));
		return;
	}

	/* De-linearize incoming message. */
	message_to_share_mode_entry(&msg, (char *)data->data);

	DEBUG(10, ("Got oplock async level 2 break message from pid %d: 0x%x/%.0f/%lu\n",
		   (int)procid_to_pid(&src), (unsigned int)msg.dev,
		   (double)msg.inode, msg.share_file_id));

	fsp = initial_break_processing(msg.dev, msg.inode,
				       msg.share_file_id);

	if (fsp == NULL) {
		/* We hit a race here. Break messages are sent, and before we
		 * get to process this message, we have closed the file. 
		 * No need to reply as this is an async message. */
		DEBUG(3, ("process_oplock_async_level2_break_message: Did not find fsp, ignoring\n"));
		return;
	}

	if (fsp->oplock_type == NO_OPLOCK) {
		/* We already got a "break to none" message and we've handled it.
		 * just ignore. */
		DEBUG(3, ("process_oplock_async_level2_break_message: already broken to none, ignoring.\n"));
		return;
	}

	if (fsp->oplock_type == FAKE_LEVEL_II_OPLOCK) {
		/* Don't tell the client, just downgrade. */
		DEBUG(3, ("process_oplock_async_level2_break_message: downgrading fake level 2 oplock.\n"));
		remove_oplock(fsp);
		return;
	}

	/* Ensure we're really at level2 state. */
	SMB_ASSERT(fsp->oplock_type == LEVEL_II_OPLOCK);

	/* Now send a break to none message to our client. */

	break_msg = new_break_smb_message(NULL, fsp, OPLOCKLEVEL_NONE);
	if (break_msg == NULL) {
		exit_server("Could not talloc break_msg\n");
	}

	/* Need to wait before sending a break message if we sent ourselves this message. */
	if (procid_to_pid(&src) == sys_getpid()) {
		wait_before_sending_break();
	}

	/* Save the server smb signing state. */
	sign_state = srv_oplock_set_signing(False);

	show_msg(break_msg);
	if (!send_smb(smbd_server_fd(), break_msg)) {
		exit_server_cleanly("oplock_break: send_smb failed.");
	}

	/* Restore the sign state to what it was. */
	srv_oplock_set_signing(sign_state);

	TALLOC_FREE(break_msg);

	/* Async level2 request, don't send a reply, just remove the oplock. */
	remove_oplock(fsp);
}

/*******************************************************************
 This handles the generic oplock break message from another smbd.
*******************************************************************/

static void process_oplock_break_message(struct messaging_context *msg_ctx,
					 void *private_data,
					 uint32_t msg_type,
					 struct server_id src,
					 DATA_BLOB *data)
{
	struct share_mode_entry msg;
	files_struct *fsp;
	char *break_msg;
	BOOL break_to_level2 = False;
	BOOL sign_state;

	if (data->data == NULL) {
		DEBUG(0, ("Got NULL buffer\n"));
		return;
	}

	if (data->length != MSG_SMB_SHARE_MODE_ENTRY_SIZE) {
		DEBUG(0, ("Got invalid msg len %d\n", (int)data->length));
		return;
	}

	/* De-linearize incoming message. */
	message_to_share_mode_entry(&msg, (char *)data->data);

	DEBUG(10, ("Got oplock break message from pid %d: 0x%x/%.0f/%lu\n",
		   (int)procid_to_pid(&src), (unsigned int)msg.dev,
		   (double)msg.inode, msg.share_file_id));

	fsp = initial_break_processing(msg.dev, msg.inode,
				       msg.share_file_id);

	if (fsp == NULL) {
		/* a We hit race here. Break messages are sent, and before we
		 * get to process this message, we have closed the file. Reply
		 * with 'ok, oplock broken' */
		DEBUG(3, ("Did not find fsp\n"));

		/* We just send the same message back. */
		messaging_send_buf(msg_ctx, src, MSG_SMB_BREAK_RESPONSE,
				   (uint8 *)data->data,
				   MSG_SMB_SHARE_MODE_ENTRY_SIZE);
		return;
	}

	if (fsp->sent_oplock_break != NO_BREAK_SENT) {
		/* Remember we have to inform the requesting PID when the
		 * client replies */
		msg.pid = src;
		ADD_TO_ARRAY(NULL, struct share_mode_entry, msg,
			     &fsp->pending_break_messages,
			     &fsp->num_pending_break_messages);
		return;
	}

	if (EXCLUSIVE_OPLOCK_TYPE(msg.op_type) &&
	    !EXCLUSIVE_OPLOCK_TYPE(fsp->oplock_type)) {
		DEBUG(3, ("Already downgraded oplock on 0x%x/%.0f: %s\n",
			  (unsigned int)fsp->dev, (double)fsp->inode,
			  fsp->fsp_name));
		/* We just send the same message back. */
		messaging_send_buf(msg_ctx, src, MSG_SMB_BREAK_RESPONSE,
				   (uint8 *)data->data,
				   MSG_SMB_SHARE_MODE_ENTRY_SIZE);
		return;
	}

	if ((global_client_caps & CAP_LEVEL_II_OPLOCKS) && 
	    !(msg.op_type & FORCE_OPLOCK_BREAK_TO_NONE) &&
	    !koplocks && /* NOTE: we force levelII off for kernel oplocks -
			  * this will change when it is supported */
	    lp_level2_oplocks(SNUM(fsp->conn))) {
		break_to_level2 = True;
	}

	break_msg = new_break_smb_message(NULL, fsp, break_to_level2 ?
					  OPLOCKLEVEL_II : OPLOCKLEVEL_NONE);
	if (break_msg == NULL) {
		exit_server("Could not talloc break_msg\n");
	}

	/* Need to wait before sending a break message if we sent ourselves this message. */
	if (procid_to_pid(&src) == sys_getpid()) {
		wait_before_sending_break();
	}

	/* Save the server smb signing state. */
	sign_state = srv_oplock_set_signing(False);

	show_msg(break_msg);
	if (!send_smb(smbd_server_fd(), break_msg)) {
		exit_server_cleanly("oplock_break: send_smb failed.");
	}

	/* Restore the sign state to what it was. */
	srv_oplock_set_signing(sign_state);

	TALLOC_FREE(break_msg);

	fsp->sent_oplock_break = break_to_level2 ? LEVEL_II_BREAK_SENT:BREAK_TO_NONE_SENT;

	msg.pid = src;
	ADD_TO_ARRAY(NULL, struct share_mode_entry, msg,
		     &fsp->pending_break_messages,
		     &fsp->num_pending_break_messages);

	add_oplock_timeout_handler(fsp);
}

/*******************************************************************
 This handles the kernel oplock break message.
*******************************************************************/

static void process_kernel_oplock_break(struct messaging_context *msg_ctx,
					void *private_data,
					uint32_t msg_type,
					struct server_id src,
					DATA_BLOB *data)
{
	SMB_DEV_T dev;
	SMB_INO_T inode;
	unsigned long file_id;
	files_struct *fsp;
	char *break_msg;
	BOOL sign_state;

	if (data->data == NULL) {
		DEBUG(0, ("Got NULL buffer\n"));
		return;
	}

	if (data->length != MSG_SMB_KERNEL_BREAK_SIZE) {
		DEBUG(0, ("Got invalid msg len %d\n", (int)data->length));
		return;
	}

	/* Pull the data from the message. */
	dev = DEV_T_VAL(data->data, 0);
	inode = INO_T_VAL(data->data, 8);
	file_id = (unsigned long)IVAL(data->data, 16);

	DEBUG(10, ("Got kernel oplock break message from pid %d: 0x%x/%.0f/%u\n",
		   (int)procid_to_pid(&src), (unsigned int)dev, (double)inode,
		   (unsigned int)file_id));

	fsp = initial_break_processing(dev, inode, file_id);

	if (fsp == NULL) {
		DEBUG(3, ("Got a kernel oplock break message for a file "
			  "I don't know about\n"));
		return;
	}

	if (fsp->sent_oplock_break != NO_BREAK_SENT) {
		/* This is ok, kernel oplocks come in completely async */
		DEBUG(3, ("Got a kernel oplock request while waiting for a "
			  "break reply\n"));
		return;
	}

	break_msg = new_break_smb_message(NULL, fsp, OPLOCKLEVEL_NONE);
	if (break_msg == NULL) {
		exit_server("Could not talloc break_msg\n");
	}

	/* Save the server smb signing state. */
	sign_state = srv_oplock_set_signing(False);

	show_msg(break_msg);
	if (!send_smb(smbd_server_fd(), break_msg)) {
		exit_server_cleanly("oplock_break: send_smb failed.");
	}

	/* Restore the sign state to what it was. */
	srv_oplock_set_signing(sign_state);

	TALLOC_FREE(break_msg);

	fsp->sent_oplock_break = BREAK_TO_NONE_SENT;

	add_oplock_timeout_handler(fsp);
}

void reply_to_oplock_break_requests(files_struct *fsp)
{
	int i;

	for (i=0; i<fsp->num_pending_break_messages; i++) {
		struct share_mode_entry *e = &fsp->pending_break_messages[i];
		char msg[MSG_SMB_SHARE_MODE_ENTRY_SIZE];

		share_mode_entry_to_message(msg, e);

		messaging_send_buf(smbd_messaging_context(), e->pid,
				   MSG_SMB_BREAK_RESPONSE,
				   (uint8 *)msg,
				   MSG_SMB_SHARE_MODE_ENTRY_SIZE);
	}

	SAFE_FREE(fsp->pending_break_messages);
	fsp->num_pending_break_messages = 0;
	if (fsp->oplock_timeout != NULL) {
		/* Remove the timed event handler. */
		TALLOC_FREE(fsp->oplock_timeout);
		fsp->oplock_timeout = NULL;
	}
	return;
}

static void process_oplock_break_response(struct messaging_context *msg_ctx,
					  void *private_data,
					  uint32_t msg_type,
					  struct server_id src,
					  DATA_BLOB *data)
{
	struct share_mode_entry msg;

	if (data->data == NULL) {
		DEBUG(0, ("Got NULL buffer\n"));
		return;
	}

	if (data->length != MSG_SMB_SHARE_MODE_ENTRY_SIZE) {
		DEBUG(0, ("Got invalid msg len %u\n",
			  (unsigned int)data->length));
		return;
	}

	/* De-linearize incoming message. */
	message_to_share_mode_entry(&msg, (char *)data->data);

	DEBUG(10, ("Got oplock break response from pid %d: 0x%x/%.0f/%lu mid %u\n",
		   (int)procid_to_pid(&src), (unsigned int)msg.dev,
		   (double)msg.inode, msg.share_file_id,
		   (unsigned int)msg.op_mid));

	/* Here's the hack from open.c, store the mid in the 'port' field */
	schedule_deferred_open_smb_message(msg.op_mid);
}

static void process_open_retry_message(struct messaging_context *msg_ctx,
				       void *private_data,
				       uint32_t msg_type,
				       struct server_id src,
				       DATA_BLOB *data)
{
	struct share_mode_entry msg;
	
	if (data->data == NULL) {
		DEBUG(0, ("Got NULL buffer\n"));
		return;
	}

	if (data->length != MSG_SMB_SHARE_MODE_ENTRY_SIZE) {
		DEBUG(0, ("Got invalid msg len %d\n", (int)data->length));
		return;
	}

	/* De-linearize incoming message. */
	message_to_share_mode_entry(&msg, (char *)data->data);

	DEBUG(10, ("Got open retry msg from pid %d: 0x%x/%.0f/%lu mid %u\n",
		   (int)procid_to_pid(&src), (unsigned int)msg.dev,
		   (double)msg.inode, msg.share_file_id,
		   (unsigned int)msg.op_mid));

	schedule_deferred_open_smb_message(msg.op_mid);
}

/****************************************************************************
 This function is called on any file modification or lock request. If a file
 is level 2 oplocked then it must tell all other level 2 holders to break to
 none.
****************************************************************************/

void release_level_2_oplocks_on_change(files_struct *fsp)
{
	int i;
	struct share_mode_lock *lck;

	/*
	 * If this file is level II oplocked then we need
	 * to grab the shared memory lock and inform all
	 * other files with a level II lock that they need
	 * to flush their read caches. We keep the lock over
	 * the shared memory area whilst doing this.
	 */

	if (!LEVEL_II_OPLOCK_TYPE(fsp->oplock_type))
		return;

	lck = get_share_mode_lock(NULL, fsp->dev, fsp->inode, NULL, NULL);
	if (lck == NULL) {
		DEBUG(0,("release_level_2_oplocks_on_change: failed to lock "
			 "share mode entry for file %s.\n", fsp->fsp_name ));
		return;
	}

	DEBUG(10,("release_level_2_oplocks_on_change: num_share_modes = %d\n", 
		  lck->num_share_modes ));

	for(i = 0; i < lck->num_share_modes; i++) {
		struct share_mode_entry *share_entry = &lck->share_modes[i];
		char msg[MSG_SMB_SHARE_MODE_ENTRY_SIZE];

		if (!is_valid_share_mode_entry(share_entry)) {
			continue;
		}

		/*
		 * As there could have been multiple writes waiting at the
		 * lock_share_entry gate we may not be the first to
		 * enter. Hence the state of the op_types in the share mode
		 * entries may be partly NO_OPLOCK and partly LEVEL_II or FAKE_LEVEL_II
		 * oplock. It will do no harm to re-send break messages to
		 * those smbd's that are still waiting their turn to remove
		 * their LEVEL_II state, and also no harm to ignore existing
		 * NO_OPLOCK states. JRA.
		 */

		DEBUG(10,("release_level_2_oplocks_on_change: "
			  "share_entry[%i]->op_type == %d\n",
			  i, share_entry->op_type ));

		if (share_entry->op_type == NO_OPLOCK) {
			continue;
		}

		/* Paranoia .... */
		if (EXCLUSIVE_OPLOCK_TYPE(share_entry->op_type)) {
			DEBUG(0,("release_level_2_oplocks_on_change: PANIC. "
				 "share mode entry %d is an exlusive "
				 "oplock !\n", i ));
			TALLOC_FREE(lck);
			abort();
		}

		share_mode_entry_to_message(msg, share_entry);

		messaging_send_buf(smbd_messaging_context(), share_entry->pid,
				   MSG_SMB_ASYNC_LEVEL2_BREAK,
				   (uint8 *)msg,
				   MSG_SMB_SHARE_MODE_ENTRY_SIZE);
	}

	/* We let the message receivers handle removing the oplock state
	   in the share mode lock db. */

	TALLOC_FREE(lck);
}

/****************************************************************************
 Linearize a share mode entry struct to an internal oplock break message.
****************************************************************************/

void share_mode_entry_to_message(char *msg, struct share_mode_entry *e)
{
	SIVAL(msg,0,(uint32)e->pid.pid);
	SSVAL(msg,4,e->op_mid);
	SSVAL(msg,6,e->op_type);
	SIVAL(msg,8,e->access_mask);
	SIVAL(msg,12,e->share_access);
	SIVAL(msg,16,e->private_options);
	SIVAL(msg,20,(uint32)e->time.tv_sec);
	SIVAL(msg,24,(uint32)e->time.tv_usec);
	SDEV_T_VAL(msg,28,e->dev);
	SINO_T_VAL(msg,36,e->inode);
	SIVAL(msg,44,e->share_file_id);
	SIVAL(msg,48,e->uid);
	SSVAL(msg,52,e->flags);
}

/****************************************************************************
 De-linearize an internal oplock break message to a share mode entry struct.
****************************************************************************/

void message_to_share_mode_entry(struct share_mode_entry *e, char *msg)
{
	e->pid.pid = (pid_t)IVAL(msg,0);
	e->op_mid = SVAL(msg,4);
	e->op_type = SVAL(msg,6);
	e->access_mask = IVAL(msg,8);
	e->share_access = IVAL(msg,12);
	e->private_options = IVAL(msg,16);
	e->time.tv_sec = (time_t)IVAL(msg,20);
	e->time.tv_usec = (int)IVAL(msg,24);
	e->dev = DEV_T_VAL(msg,28);
	e->inode = INO_T_VAL(msg,36);
	e->share_file_id = (unsigned long)IVAL(msg,44);
	e->uid = (uint32)IVAL(msg,48);
	e->flags = (uint16)SVAL(msg,52);
}

/****************************************************************************
 Setup oplocks for this process.
****************************************************************************/

BOOL init_oplocks(struct messaging_context *msg_ctx)
{
	DEBUG(3,("init_oplocks: initializing messages.\n"));

	messaging_register(msg_ctx, NULL, MSG_SMB_BREAK_REQUEST,
			   process_oplock_break_message);
	messaging_register(msg_ctx, NULL, MSG_SMB_ASYNC_LEVEL2_BREAK,
			   process_oplock_async_level2_break_message);
	messaging_register(msg_ctx, NULL, MSG_SMB_BREAK_RESPONSE,
			   process_oplock_break_response);
	messaging_register(msg_ctx, NULL, MSG_SMB_KERNEL_BREAK,
			   process_kernel_oplock_break);
	messaging_register(msg_ctx, NULL, MSG_SMB_OPEN_RETRY,
			   process_open_retry_message);

	if (lp_kernel_oplocks()) {
#if HAVE_KERNEL_OPLOCKS_IRIX
		koplocks = irix_init_kernel_oplocks();
#elif HAVE_KERNEL_OPLOCKS_LINUX
		koplocks = linux_init_kernel_oplocks();
#endif
	}

	return True;
}