summaryrefslogtreecommitdiffstats
path: root/fs/afs/volume.c
blob: 546f9d01710b5b66706331968edc557306ed848e (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
/* AFS volume management
 *
 * Copyright (C) 2002, 2007 Red Hat, Inc. All Rights Reserved.
 * Written by David Howells (dhowells@redhat.com)
 *
 * 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.
 */

#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/slab.h>
#include <linux/fs.h>
#include <linux/pagemap.h>
#include <linux/sched.h>
#include "internal.h"

static const char *afs_voltypes[] = { "R/W", "R/O", "BAK" };

/*
 * lookup a volume by name
 * - this can be one of the following:
 *	"%[cell:]volume[.]"		R/W volume
 *	"#[cell:]volume[.]"		R/O or R/W volume (rwparent=0),
 *					 or R/W (rwparent=1) volume
 *	"%[cell:]volume.readonly"	R/O volume
 *	"#[cell:]volume.readonly"	R/O volume
 *	"%[cell:]volume.backup"		Backup volume
 *	"#[cell:]volume.backup"		Backup volume
 *
 * The cell name is optional, and defaults to the current cell.
 *
 * See "The Rules of Mount Point Traversal" in Chapter 5 of the AFS SysAdmin
 * Guide
 * - Rule 1: Explicit type suffix forces access of that type or nothing
 *           (no suffix, then use Rule 2 & 3)
 * - Rule 2: If parent volume is R/O, then mount R/O volume by preference, R/W
 *           if not available
 * - Rule 3: If parent volume is R/W, then only mount R/W volume unless
 *           explicitly told otherwise
 */
struct afs_volume *afs_volume_lookup(struct afs_mount_params *params)
{
	struct afs_vlocation *vlocation = NULL;
	struct afs_volume *volume = NULL;
	struct afs_server *server = NULL;
	char srvtmask;
	int ret, loop;

	_enter("{%*.*s,%d}",
	       params->volnamesz, params->volnamesz, params->volname, params->rwpath);

	/* lookup the volume location record */
	vlocation = afs_vlocation_lookup(params->cell, params->key,
					 params->volname, params->volnamesz);
	if (IS_ERR(vlocation)) {
		ret = PTR_ERR(vlocation);
		vlocation = NULL;
		goto error;
	}

	/* make the final decision on the type we want */
	ret = -ENOMEDIUM;
	if (params->force && !(vlocation->vldb.vidmask & (1 << params->type)))
		goto error;

	srvtmask = 0;
	for (loop = 0; loop < vlocation->vldb.nservers; loop++)
		srvtmask |= vlocation->vldb.srvtmask[loop];

	if (params->force) {
		if (!(srvtmask & (1 << params->type)))
			goto error;
	} else if (srvtmask & AFS_VOL_VTM_RO) {
		params->type = AFSVL_ROVOL;
	} else if (srvtmask & AFS_VOL_VTM_RW) {
		params->type = AFSVL_RWVOL;
	} else {
		goto error;
	}

	down_write(&params->cell->vl_sem);

	/* is the volume already active? */
	if (vlocation->vols[params->type]) {
		/* yes - re-use it */
		volume = vlocation->vols[params->type];
		afs_get_volume(volume);
		goto success;
	}

	/* create a new volume record */
	_debug("creating new volume record");

	ret = -ENOMEM;
	volume = kzalloc(sizeof(struct afs_volume), GFP_KERNEL);
	if (!volume)
		goto error_up;

	atomic_set(&volume->usage, 1);
	volume->type		= params->type;
	volume->type_force	= params->force;
	volume->cell		= params->cell;
	volume->vid		= vlocation->vldb.vid[params->type];

	volume->bdi.ra_pages	= VM_MAX_READAHEAD*1024/PAGE_SIZE; 
	ret = bdi_setup_and_register(&volume->bdi, "afs");
	if (ret)
		goto error_bdi;

	init_rwsem(&volume->server_sem);

	/* look up all the applicable server records */
	for (loop = 0; loop < 8; loop++) {
		if (vlocation->vldb.srvtmask[loop] & (1 << volume->type)) {
			server = afs_lookup_server(
			       volume->cell, &vlocation->vldb.servers[loop]);
			if (IS_ERR(server)) {
				ret = PTR_ERR(server);
				goto error_discard;
			}

			volume->servers[volume->nservers] = server;
			volume->nservers++;
		}
	}

	/* attach the cache and volume location */
#ifdef CONFIG_AFS_FSCACHE
	volume->cache = fscache_acquire_cookie(vlocation->cache,
					       &afs_volume_cache_index_def,
					       volume, true);
#endif
	afs_get_vlocation(vlocation);
	volume->vlocation = vlocation;

	vlocation->vols[volume->type] = volume;

success:
	_debug("kAFS selected %s volume %08x",
	       afs_voltypes[volume->type], volume->vid);
	up_write(&params->cell->vl_sem);
	afs_put_vlocation(vlocation);
	_leave(" = %p", volume);
	return volume;

	/* clean up */
error_up:
	up_write(&params->cell->vl_sem);
error:
	afs_put_vlocation(vlocation);
	_leave(" = %d", ret);
	return ERR_PTR(ret);

error_discard:
	bdi_destroy(&volume->bdi);
error_bdi:
	up_write(&params->cell->vl_sem);

	for (loop = volume->nservers - 1; loop >= 0; loop--)
		afs_put_server(volume->servers[loop]);

	kfree(volume);
	goto error;
}

/*
 * destroy a volume record
 */
void afs_put_volume(struct afs_volume *volume)
{
	struct afs_vlocation *vlocation;
	int loop;

	if (!volume)
		return;

	_enter("%p", volume);

	ASSERTCMP(atomic_read(&volume->usage), >, 0);

	vlocation = volume->vlocation;

	/* to prevent a race, the decrement and the dequeue must be effectively
	 * atomic */
	down_write(&vlocation->cell->vl_sem);

	if (likely(!atomic_dec_and_test(&volume->usage))) {
		up_write(&vlocation->cell->vl_sem);
		_leave("");
		return;
	}

	vlocation->vols[volume->type] = NULL;

	up_write(&vlocation->cell->vl_sem);

	/* finish cleaning up the volume */
#ifdef CONFIG_AFS_FSCACHE
	fscache_relinquish_cookie(volume->cache, 0);
#endif
	afs_put_vlocation(vlocation);

	for (loop = volume->nservers - 1; loop >= 0; loop--)
		afs_put_server(volume->servers[loop]);

	bdi_destroy(&volume->bdi);
	kfree(volume);

	_leave(" [destroyed]");
}

/*
 * pick a server to use to try accessing this volume
 * - returns with an elevated usage count on the server chosen
 */
struct afs_server *afs_volume_pick_fileserver(struct afs_vnode *vnode)
{
	struct afs_volume *volume = vnode->volume;
	struct afs_server *server;
	int ret, state, loop;

	_enter("%s", volume->vlocation->vldb.name);

	/* stick with the server we're already using if we can */
	if (vnode->server && vnode->server->fs_state == 0) {
		afs_get_server(vnode->server);
		_leave(" = %p [current]", vnode->server);
		return vnode->server;
	}

	down_read(&volume->server_sem);

	/* handle the no-server case */
	if (volume->nservers == 0) {
		ret = volume->rjservers ? -ENOMEDIUM : -ESTALE;
		up_read(&volume->server_sem);
		_leave(" = %d [no servers]", ret);
		return ERR_PTR(ret);
	}

	/* basically, just search the list for the first live server and use
	 * that */
	ret = 0;
	for (loop = 0; loop < volume->nservers; loop++) {
		server = volume->servers[loop];
		state = server->fs_state;

		_debug("consider %d [%d]", loop, state);

		switch (state) {
			/* found an apparently healthy server */
		case 0:
			afs_get_server(server);
			up_read(&volume->server_sem);
			_leave(" = %p (picked %08x)",
			       server, ntohl(server->addr.s_addr));
			return server;

		case -ENETUNREACH:
			if (ret == 0)
				ret = state;
			break;

		case -EHOSTUNREACH:
			if (ret == 0 ||
			    ret == -ENETUNREACH)
				ret = state;
			break;

		case -ECONNREFUSED:
			if (ret == 0 ||
			    ret == -ENETUNREACH ||
			    ret == -EHOSTUNREACH)
				ret = state;
			break;

		default:
		case -EREMOTEIO:
			if (ret == 0 ||
			    ret == -ENETUNREACH ||
			    ret == -EHOSTUNREACH ||
			    ret == -ECONNREFUSED)
				ret = state;
			break;
		}
	}

	/* no available servers
	 * - TODO: handle the no active servers case better
	 */
	up_read(&volume->server_sem);
	_leave(" = %d", ret);
	return ERR_PTR(ret);
}

/*
 * release a server after use
 * - releases the ref on the server struct that was acquired by picking
 * - records result of using a particular server to access a volume
 * - return 0 to try again, 1 if okay or to issue error
 * - the caller must release the server struct if result was 0
 */
int afs_volume_release_fileserver(struct afs_vnode *vnode,
				  struct afs_server *server,
				  int result)
{
	struct afs_volume *volume = vnode->volume;
	unsigned loop;

	_enter("%s,%08x,%d",
	       volume->vlocation->vldb.name, ntohl(server->addr.s_addr),
	       result);

	switch (result) {
		/* success */
	case 0:
		server->fs_act_jif = jiffies;
		server->fs_state = 0;
		_leave("");
		return 1;

		/* the fileserver denied all knowledge of the volume */
	case -ENOMEDIUM:
		server->fs_act_jif = jiffies;
		down_write(&volume->server_sem);

		/* firstly, find where the server is in the active list (if it
		 * is) */
		for (loop = 0; loop < volume->nservers; loop++)
			if (volume->servers[loop] == server)
				goto present;

		/* no longer there - may have been discarded by another op */
		goto try_next_server_upw;

	present:
		volume->nservers--;
		memmove(&volume->servers[loop],
			&volume->servers[loop + 1],
			sizeof(volume->servers[loop]) *
			(volume->nservers - loop));
		volume->servers[volume->nservers] = NULL;
		afs_put_server(server);
		volume->rjservers++;

		if (volume->nservers > 0)
			/* another server might acknowledge its existence */
			goto try_next_server_upw;

		/* handle the case where all the fileservers have rejected the
		 * volume
		 * - TODO: try asking the fileservers for volume information
		 * - TODO: contact the VL server again to see if the volume is
		 *         no longer registered
		 */
		up_write(&volume->server_sem);
		afs_put_server(server);
		_leave(" [completely rejected]");
		return 1;

		/* problem reaching the server */
	case -ENETUNREACH:
	case -EHOSTUNREACH:
	case -ECONNREFUSED:
	case -ETIME:
	case -ETIMEDOUT:
	case -EREMOTEIO:
		/* mark the server as dead
		 * TODO: vary dead timeout depending on error
		 */
		spin_lock(&server->fs_lock);
		if (!server->fs_state) {
			server->fs_dead_jif = jiffies + HZ * 10;
			server->fs_state = result;
			printk("kAFS: SERVER DEAD state=%d\n", result);
		}
		spin_unlock(&server->fs_lock);
		goto try_next_server;

		/* miscellaneous error */
	default:
		server->fs_act_jif = jiffies;
	case -ENOMEM:
	case -ENONET:
		/* tell the caller to accept the result */
		afs_put_server(server);
		_leave(" [local failure]");
		return 1;
	}

	/* tell the caller to loop around and try the next server */
try_next_server_upw:
	up_write(&volume->server_sem);
try_next_server:
	afs_put_server(server);
	_leave(" [try next server]");
	return 0;
}