summaryrefslogtreecommitdiffstats
path: root/fs/afs/callback.c
blob: bacf518c6fa80d891f24e1168a82abf1a48d2d62 (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
/*
 * Copyright (c) 2002, 2007 Red Hat, Inc. All rights reserved.
 *
 * This software may be freely redistributed under the terms of the
 * GNU General Public License.
 *
 * 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.
 *
 * Authors: David Woodhouse <dwmw2@cambridge.redhat.com>
 *          David Howells <dhowells@redhat.com>
 *
 */

#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/circ_buf.h>
#include <linux/sched.h>
#include "internal.h"

unsigned afs_vnode_update_timeout = 10;

#define afs_breakring_space(server) \
	CIRC_SPACE((server)->cb_break_head, (server)->cb_break_tail,	\
		   ARRAY_SIZE((server)->cb_break))

//static void afs_callback_updater(struct work_struct *);

static struct workqueue_struct *afs_callback_update_worker;

/*
 * allow the fileserver to request callback state (re-)initialisation
 */
void afs_init_callback_state(struct afs_server *server)
{
	struct afs_vnode *vnode;

	_enter("{%p}", server);

	spin_lock(&server->cb_lock);

	/* kill all the promises on record from this server */
	while (!RB_EMPTY_ROOT(&server->cb_promises)) {
		vnode = rb_entry(server->cb_promises.rb_node,
				 struct afs_vnode, cb_promise);
		_debug("UNPROMISE { vid=%x:%u uq=%u}",
		       vnode->fid.vid, vnode->fid.vnode, vnode->fid.unique);
		rb_erase(&vnode->cb_promise, &server->cb_promises);
		vnode->cb_promised = false;
	}

	spin_unlock(&server->cb_lock);
	_leave("");
}

/*
 * handle the data invalidation side of a callback being broken
 */
void afs_broken_callback_work(struct work_struct *work)
{
	struct afs_vnode *vnode =
		container_of(work, struct afs_vnode, cb_broken_work);

	_enter("");

	if (test_bit(AFS_VNODE_DELETED, &vnode->flags))
		return;

	/* we're only interested in dealing with a broken callback on *this*
	 * vnode and only if no-one else has dealt with it yet */
	if (!mutex_trylock(&vnode->validate_lock))
		return; /* someone else is dealing with it */

	if (test_bit(AFS_VNODE_CB_BROKEN, &vnode->flags)) {
		if (S_ISDIR(vnode->vfs_inode.i_mode))
			afs_clear_permits(vnode);

		if (afs_vnode_fetch_status(vnode, NULL, NULL) < 0)
			goto out;

		if (test_bit(AFS_VNODE_DELETED, &vnode->flags))
			goto out;

		/* if the vnode's data version number changed then its contents
		 * are different */
		if (test_and_clear_bit(AFS_VNODE_ZAP_DATA, &vnode->flags))
			afs_zap_data(vnode);
	}

out:
	mutex_unlock(&vnode->validate_lock);

	/* avoid the potential race whereby the mutex_trylock() in this
	 * function happens again between the clear_bit() and the
	 * mutex_unlock() */
	if (test_bit(AFS_VNODE_CB_BROKEN, &vnode->flags)) {
		_debug("requeue");
		queue_work(afs_callback_update_worker, &vnode->cb_broken_work);
	}
	_leave("");
}

/*
 * actually break a callback
 */
static void afs_break_callback(struct afs_server *server,
			       struct afs_vnode *vnode)
{
	_enter("");

	set_bit(AFS_VNODE_CB_BROKEN, &vnode->flags);

	if (vnode->cb_promised) {
		spin_lock(&vnode->lock);

		_debug("break callback");

		spin_lock(&server->cb_lock);
		if (vnode->cb_promised) {
			rb_erase(&vnode->cb_promise, &server->cb_promises);
			vnode->cb_promised = false;
		}
		spin_unlock(&server->cb_lock);

		queue_work(afs_callback_update_worker, &vnode->cb_broken_work);
		spin_unlock(&vnode->lock);
	}
}

/*
 * allow the fileserver to explicitly break one callback
 * - happens when
 *   - the backing file is changed
 *   - a lock is released
 */
static void afs_break_one_callback(struct afs_server *server,
				   struct afs_fid *fid)
{
	struct afs_vnode *vnode;
	struct rb_node *p;

	_debug("find");
	spin_lock(&server->fs_lock);
	p = server->fs_vnodes.rb_node;
	while (p) {
		vnode = rb_entry(p, struct afs_vnode, server_rb);
		if (fid->vid < vnode->fid.vid)
			p = p->rb_left;
		else if (fid->vid > vnode->fid.vid)
			p = p->rb_right;
		else if (fid->vnode < vnode->fid.vnode)
			p = p->rb_left;
		else if (fid->vnode > vnode->fid.vnode)
			p = p->rb_right;
		else if (fid->unique < vnode->fid.unique)
			p = p->rb_left;
		else if (fid->unique > vnode->fid.unique)
			p = p->rb_right;
		else
			goto found;
	}

	/* not found so we just ignore it (it may have moved to another
	 * server) */
not_available:
	_debug("not avail");
	spin_unlock(&server->fs_lock);
	_leave("");
	return;

found:
	_debug("found");
	ASSERTCMP(server, ==, vnode->server);

	if (!igrab(AFS_VNODE_TO_I(vnode)))
		goto not_available;
	spin_unlock(&server->fs_lock);

	afs_break_callback(server, vnode);
	iput(&vnode->vfs_inode);
	_leave("");
}

/*
 * allow the fileserver to break callback promises
 */
void afs_break_callbacks(struct afs_server *server, size_t count,
			 struct afs_callback callbacks[])
{
	_enter("%p,%zu,", server, count);

	ASSERT(server != NULL);
	ASSERTCMP(count, <=, AFSCBMAX);

	for (; count > 0; callbacks++, count--) {
		_debug("- Fid { vl=%08x n=%u u=%u }  CB { v=%u x=%u t=%u }",
		       callbacks->fid.vid,
		       callbacks->fid.vnode,
		       callbacks->fid.unique,
		       callbacks->version,
		       callbacks->expiry,
		       callbacks->type
		       );
		afs_break_one_callback(server, &callbacks->fid);
	}

	_leave("");
	return;
}

/*
 * record the callback for breaking
 * - the caller must hold server->cb_lock
 */
static void afs_do_give_up_callback(struct afs_server *server,
				    struct afs_vnode *vnode)
{
	struct afs_callback *cb;

	_enter("%p,%p", server, vnode);

	cb = &server->cb_break[server->cb_break_head];
	cb->fid		= vnode->fid;
	cb->version	= vnode->cb_version;
	cb->expiry	= vnode->cb_expiry;
	cb->type	= vnode->cb_type;
	smp_wmb();
	server->cb_break_head =
		(server->cb_break_head + 1) &
		(ARRAY_SIZE(server->cb_break) - 1);

	/* defer the breaking of callbacks to try and collect as many as
	 * possible to ship in one operation */
	switch (atomic_inc_return(&server->cb_break_n)) {
	case 1 ... AFSCBMAX - 1:
		queue_delayed_work(afs_callback_update_worker,
				   &server->cb_break_work, HZ * 2);
		break;
	case AFSCBMAX:
		afs_flush_callback_breaks(server);
		break;
	default:
		break;
	}

	ASSERT(server->cb_promises.rb_node != NULL);
	rb_erase(&vnode->cb_promise, &server->cb_promises);
	vnode->cb_promised = false;
	_leave("");
}

/*
 * discard the callback on a deleted item
 */
void afs_discard_callback_on_delete(struct afs_vnode *vnode)
{
	struct afs_server *server = vnode->server;

	_enter("%d", vnode->cb_promised);

	if (!vnode->cb_promised) {
		_leave(" [not promised]");
		return;
	}

	ASSERT(server != NULL);

	spin_lock(&server->cb_lock);
	if (vnode->cb_promised) {
		ASSERT(server->cb_promises.rb_node != NULL);
		rb_erase(&vnode->cb_promise, &server->cb_promises);
		vnode->cb_promised = false;
	}
	spin_unlock(&server->cb_lock);
	_leave("");
}

/*
 * give up the callback registered for a vnode on the file server when the
 * inode is being cleared
 */
void afs_give_up_callback(struct afs_vnode *vnode)
{
	struct afs_server *server = vnode->server;

	DECLARE_WAITQUEUE(myself, current);

	_enter("%d", vnode->cb_promised);

	_debug("GIVE UP INODE %p", &vnode->vfs_inode);

	if (!vnode->cb_promised) {
		_leave(" [not promised]");
		return;
	}

	ASSERT(server != NULL);

	spin_lock(&server->cb_lock);
	if (vnode->cb_promised && afs_breakring_space(server) == 0) {
		add_wait_queue(&server->cb_break_waitq, &myself);
		for (;;) {
			set_current_state(TASK_UNINTERRUPTIBLE);
			if (!vnode->cb_promised ||
			    afs_breakring_space(server) != 0)
				break;
			spin_unlock(&server->cb_lock);
			schedule();
			spin_lock(&server->cb_lock);
		}
		remove_wait_queue(&server->cb_break_waitq, &myself);
		__set_current_state(TASK_RUNNING);
	}

	/* of course, it's always possible for the server to break this vnode's
	 * callback first... */
	if (vnode->cb_promised)
		afs_do_give_up_callback(server, vnode);

	spin_unlock(&server->cb_lock);
	_leave("");
}

/*
 * dispatch a deferred give up callbacks operation
 */
void afs_dispatch_give_up_callbacks(struct work_struct *work)
{
	struct afs_server *server =
		container_of(work, struct afs_server, cb_break_work.work);

	_enter("");

	/* tell the fileserver to discard the callback promises it has
	 * - in the event of ENOMEM or some other error, we just forget that we
	 *   had callbacks entirely, and the server will call us later to break
	 *   them
	 */
	afs_fs_give_up_callbacks(server, &afs_async_call);
}

/*
 * flush the outstanding callback breaks on a server
 */
void afs_flush_callback_breaks(struct afs_server *server)
{
	cancel_delayed_work(&server->cb_break_work);
	queue_delayed_work(afs_callback_update_worker,
			   &server->cb_break_work, 0);
}

#if 0
/*
 * update a bunch of callbacks
 */
static void afs_callback_updater(struct work_struct *work)
{
	struct afs_server *server;
	struct afs_vnode *vnode, *xvnode;
	time_t now;
	long timeout;
	int ret;

	server = container_of(work, struct afs_server, updater);

	_enter("");

	now = get_seconds();

	/* find the first vnode to update */
	spin_lock(&server->cb_lock);
	for (;;) {
		if (RB_EMPTY_ROOT(&server->cb_promises)) {
			spin_unlock(&server->cb_lock);
			_leave(" [nothing]");
			return;
		}

		vnode = rb_entry(rb_first(&server->cb_promises),
				 struct afs_vnode, cb_promise);
		if (atomic_read(&vnode->usage) > 0)
			break;
		rb_erase(&vnode->cb_promise, &server->cb_promises);
		vnode->cb_promised = false;
	}

	timeout = vnode->update_at - now;
	if (timeout > 0) {
		queue_delayed_work(afs_vnode_update_worker,
				   &afs_vnode_update, timeout * HZ);
		spin_unlock(&server->cb_lock);
		_leave(" [nothing]");
		return;
	}

	list_del_init(&vnode->update);
	atomic_inc(&vnode->usage);
	spin_unlock(&server->cb_lock);

	/* we can now perform the update */
	_debug("update %s", vnode->vldb.name);
	vnode->state = AFS_VL_UPDATING;
	vnode->upd_rej_cnt = 0;
	vnode->upd_busy_cnt = 0;

	ret = afs_vnode_update_record(vl, &vldb);
	switch (ret) {
	case 0:
		afs_vnode_apply_update(vl, &vldb);
		vnode->state = AFS_VL_UPDATING;
		break;
	case -ENOMEDIUM:
		vnode->state = AFS_VL_VOLUME_DELETED;
		break;
	default:
		vnode->state = AFS_VL_UNCERTAIN;
		break;
	}

	/* and then reschedule */
	_debug("reschedule");
	vnode->update_at = get_seconds() + afs_vnode_update_timeout;

	spin_lock(&server->cb_lock);

	if (!list_empty(&server->cb_promises)) {
		/* next update in 10 minutes, but wait at least 1 second more
		 * than the newest record already queued so that we don't spam
		 * the VL server suddenly with lots of requests
		 */
		xvnode = list_entry(server->cb_promises.prev,
				    struct afs_vnode, update);
		if (vnode->update_at <= xvnode->update_at)
			vnode->update_at = xvnode->update_at + 1;
		xvnode = list_entry(server->cb_promises.next,
				    struct afs_vnode, update);
		timeout = xvnode->update_at - now;
		if (timeout < 0)
			timeout = 0;
	} else {
		timeout = afs_vnode_update_timeout;
	}

	list_add_tail(&vnode->update, &server->cb_promises);

	_debug("timeout %ld", timeout);
	queue_delayed_work(afs_vnode_update_worker,
			   &afs_vnode_update, timeout * HZ);
	spin_unlock(&server->cb_lock);
	afs_put_vnode(vl);
}
#endif

/*
 * initialise the callback update process
 */
int __init afs_callback_update_init(void)
{
	afs_callback_update_worker =
		create_singlethread_workqueue("kafs_callbackd");
	return afs_callback_update_worker ? 0 : -ENOMEM;
}

/*
 * shut down the callback update process
 */
void afs_callback_update_kill(void)
{
	destroy_workqueue(afs_callback_update_worker);
}
OpenPOWER on IntegriCloud