Home Home > GIT Browse
summaryrefslogtreecommitdiff
blob: 6cdd7047c809377873578467169d13c6a63adbc1 (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
/*
 * 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@infradead.org>
 *          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"

/*
 * Create volume and callback interests on a server.
 */
static struct afs_cb_interest *afs_create_interest(struct afs_server *server,
						   struct afs_vnode *vnode)
{
	struct afs_vol_interest *new_vi, *vi;
	struct afs_cb_interest *new;
	struct hlist_node **pp;

	new_vi = kzalloc(sizeof(struct afs_vol_interest), GFP_KERNEL);
	if (!new_vi)
		return NULL;

	new = kzalloc(sizeof(struct afs_cb_interest), GFP_KERNEL);
	if (!new) {
		kfree(new_vi);
		return NULL;
	}

	new_vi->usage = 1;
	new_vi->vid = vnode->volume->vid;
	INIT_HLIST_NODE(&new_vi->srv_link);
	INIT_HLIST_HEAD(&new_vi->cb_interests);

	refcount_set(&new->usage, 1);
	new->sb = vnode->vfs_inode.i_sb;
	new->vid = vnode->volume->vid;
	new->server = afs_get_server(server, afs_server_trace_get_new_cbi);
	INIT_HLIST_NODE(&new->cb_vlink);

	write_lock(&server->cb_break_lock);

	for (pp = &server->cb_volumes.first; *pp; pp = &(*pp)->next) {
		vi = hlist_entry(*pp, struct afs_vol_interest, srv_link);
		if (vi->vid < new_vi->vid)
			continue;
		if (vi->vid > new_vi->vid)
			break;
		vi->usage++;
		goto found_vi;
	}

	new_vi->srv_link.pprev = pp;
	new_vi->srv_link.next = *pp;
	if (*pp)
		(*pp)->pprev = &new_vi->srv_link.next;
	*pp = &new_vi->srv_link;
	vi = new_vi;
	new_vi = NULL;
found_vi:

	new->vol_interest = vi;
	hlist_add_head(&new->cb_vlink, &vi->cb_interests);

	write_unlock(&server->cb_break_lock);
	kfree(new_vi);
	return new;
}

/*
 * Set up an interest-in-callbacks record for a volume on a server and
 * register it with the server.
 * - Called with vnode->io_lock held.
 */
int afs_register_server_cb_interest(struct afs_vnode *vnode,
				    struct afs_server_list *slist,
				    unsigned int index)
{
	struct afs_server_entry *entry = &slist->servers[index];
	struct afs_cb_interest *cbi, *vcbi, *new, *old;
	struct afs_server *server = entry->server;

again:
	vcbi = rcu_dereference_protected(vnode->cb_interest,
					 lockdep_is_held(&vnode->io_lock));
	if (vcbi && likely(vcbi == entry->cb_interest))
		return 0;

	read_lock(&slist->lock);
	cbi = afs_get_cb_interest(entry->cb_interest);
	read_unlock(&slist->lock);

	if (vcbi) {
		if (vcbi == cbi) {
			afs_put_cb_interest(afs_v2net(vnode), cbi);
			return 0;
		}

		/* Use a new interest in the server list for the same server
		 * rather than an old one that's still attached to a vnode.
		 */
		if (cbi && vcbi->server == cbi->server) {
			write_seqlock(&vnode->cb_lock);
			old = rcu_dereference_protected(vnode->cb_interest,
							lockdep_is_held(&vnode->cb_lock.lock));
			rcu_assign_pointer(vnode->cb_interest, cbi);
			write_sequnlock(&vnode->cb_lock);
			afs_put_cb_interest(afs_v2net(vnode), old);
			return 0;
		}

		/* Re-use the one attached to the vnode. */
		if (!cbi && vcbi->server == server) {
			write_lock(&slist->lock);
			if (entry->cb_interest) {
				write_unlock(&slist->lock);
				afs_put_cb_interest(afs_v2net(vnode), cbi);
				goto again;
			}

			entry->cb_interest = cbi;
			write_unlock(&slist->lock);
			return 0;
		}
	}

	if (!cbi) {
		new = afs_create_interest(server, vnode);
		if (!new)
			return -ENOMEM;

		write_lock(&slist->lock);
		if (!entry->cb_interest) {
			entry->cb_interest = afs_get_cb_interest(new);
			cbi = new;
			new = NULL;
		} else {
			cbi = afs_get_cb_interest(entry->cb_interest);
		}
		write_unlock(&slist->lock);
		afs_put_cb_interest(afs_v2net(vnode), new);
	}

	ASSERT(cbi);

	/* Change the server the vnode is using.  This entails scrubbing any
	 * interest the vnode had in the previous server it was using.
	 */
	write_seqlock(&vnode->cb_lock);

	old = rcu_dereference_protected(vnode->cb_interest,
					lockdep_is_held(&vnode->cb_lock.lock));
	rcu_assign_pointer(vnode->cb_interest, cbi);
	vnode->cb_s_break = cbi->server->cb_s_break;
	vnode->cb_v_break = vnode->volume->cb_v_break;
	clear_bit(AFS_VNODE_CB_PROMISED, &vnode->flags);

	write_sequnlock(&vnode->cb_lock);
	afs_put_cb_interest(afs_v2net(vnode), old);
	return 0;
}

/*
 * Remove an interest on a server.
 */
void afs_put_cb_interest(struct afs_net *net, struct afs_cb_interest *cbi)
{
	struct afs_vol_interest *vi;

	if (cbi && refcount_dec_and_test(&cbi->usage)) {
		if (!hlist_unhashed(&cbi->cb_vlink)) {
			write_lock(&cbi->server->cb_break_lock);

			hlist_del_init(&cbi->cb_vlink);
			vi = cbi->vol_interest;
			cbi->vol_interest = NULL;
			if (--vi->usage == 0)
				hlist_del(&vi->srv_link);
			else
				vi = NULL;

			write_unlock(&cbi->server->cb_break_lock);
			if (vi)
				kfree_rcu(vi, rcu);
			afs_put_server(net, cbi->server, afs_server_trace_put_cbi);
		}
		kfree_rcu(cbi, rcu);
	}
}

/*
 * allow the fileserver to request callback state (re-)initialisation
 */
void afs_init_callback_state(struct afs_server *server)
{
	server->cb_s_break++;
}

/*
 * actually break a callback
 */
void __afs_break_callback(struct afs_vnode *vnode, enum afs_cb_break_reason reason)
{
	_enter("");

	clear_bit(AFS_VNODE_NEW_CONTENT, &vnode->flags);
	if (test_and_clear_bit(AFS_VNODE_CB_PROMISED, &vnode->flags)) {
		vnode->cb_break++;
		afs_clear_permits(vnode);

		if (vnode->lock_state == AFS_VNODE_LOCK_WAITING_FOR_CB)
			afs_lock_may_be_available(vnode);

		trace_afs_cb_break(&vnode->fid, vnode->cb_break, reason, true);
	} else {
		trace_afs_cb_break(&vnode->fid, vnode->cb_break, reason, false);
	}
}

void afs_break_callback(struct afs_vnode *vnode, enum afs_cb_break_reason reason)
{
	write_seqlock(&vnode->cb_lock);
	__afs_break_callback(vnode, reason);
	write_sequnlock(&vnode->cb_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_vol_interest *vi;
	struct afs_cb_interest *cbi;
	struct afs_iget_data data;
	struct afs_vnode *vnode;
	struct inode *inode;

	read_lock(&server->cb_break_lock);
	hlist_for_each_entry(vi, &server->cb_volumes, srv_link) {
		if (vi->vid < fid->vid)
			continue;
		if (vi->vid > fid->vid) {
			vi = NULL;
			break;
		}
		//atomic_inc(&vi->usage);
		break;
	}

	/* TODO: Find all matching volumes if we couldn't match the server and
	 * break them anyway.
	 */
	if (!vi)
		goto out;

	/* Step through all interested superblocks.  There may be more than one
	 * because of cell aliasing.
	 */
	hlist_for_each_entry(cbi, &vi->cb_interests, cb_vlink) {
		if (fid->vnode == 0 && fid->unique == 0) {
			/* The callback break applies to an entire volume. */
			struct afs_super_info *as = AFS_FS_S(cbi->sb);
			struct afs_volume *volume = as->volume;

			write_lock(&volume->cb_v_break_lock);
			volume->cb_v_break++;
			trace_afs_cb_break(fid, volume->cb_v_break,
					   afs_cb_break_for_volume_callback, false);
			write_unlock(&volume->cb_v_break_lock);
		} else {
			data.volume = NULL;
			data.fid = *fid;
			inode = ilookup5_nowait(cbi->sb, fid->vnode,
						afs_iget5_test, &data);
			if (inode) {
				vnode = AFS_FS_I(inode);
				afs_break_callback(vnode, afs_cb_break_for_callback);
				iput(inode);
			} else {
				trace_afs_cb_miss(fid, afs_cb_break_for_callback);
			}
		}
	}

out:
	read_unlock(&server->cb_break_lock);
}

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

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

	/* TODO: Sort the callback break list by volume ID */

	for (; count > 0; callbacks++, count--) {
		_debug("- Fid { vl=%08llx n=%llu u=%u }",
		       callbacks->fid.vid,
		       callbacks->fid.vnode,
		       callbacks->fid.unique);
		afs_break_one_callback(server, &callbacks->fid);
	}

	_leave("");
	return;
}

/*
 * Clear the callback interests in a server list.
 */
void afs_clear_callback_interests(struct afs_net *net, struct afs_server_list *slist)
{
	int i;

	for (i = 0; i < slist->nr_servers; i++) {
		afs_put_cb_interest(net, slist->servers[i].cb_interest);
		slist->servers[i].cb_interest = NULL;
	}
}