summaryrefslogtreecommitdiffstats
path: root/fs/afs/callback.c
blob: 99b2c8172021a2f1ac8bc1d8b4e36d947b971f48 (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
/*
 * 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"

/*
 * Handle invalidation of an mmap'd file.  We invalidate all the PTEs referring
 * to the pages in this file's pagecache, forcing the kernel to go through
 * ->fault() or ->page_mkwrite() - at which point we can handle invalidation
 * more fully.
 */
void afs_invalidate_mmap_work(struct work_struct *work)
{
	struct afs_vnode *vnode = container_of(work, struct afs_vnode, cb_work);

	unmap_mapping_pages(vnode->netfs.inode.i_mapping, 0, 0, false);
}

static void afs_volume_init_callback(struct afs_volume *volume)
{
	struct afs_vnode *vnode;

	down_read(&volume->open_mmaps_lock);

	list_for_each_entry(vnode, &volume->open_mmaps, cb_mmap_link) {
		if (vnode->cb_v_check != atomic_read(&volume->cb_v_break)) {
			atomic64_set(&vnode->cb_expires_at, AFS_NO_CB_PROMISE);
			queue_work(system_unbound_wq, &vnode->cb_work);
		}
	}

	up_read(&volume->open_mmaps_lock);
}

/*
 * Allow the fileserver to request callback state (re-)initialisation.
 * Unfortunately, UUIDs are not guaranteed unique.
 */
void afs_init_callback_state(struct afs_server *server)
{
	struct afs_server_entry *se;

	down_read(&server->cell->vs_lock);

	list_for_each_entry(se, &server->volumes, slink) {
		se->cb_expires_at = AFS_NO_CB_PROMISE;
		se->volume->cb_expires_at = AFS_NO_CB_PROMISE;
		trace_afs_cb_v_break(se->volume->vid, atomic_read(&se->volume->cb_v_break),
				     afs_cb_break_for_s_reinit);
		if (!list_empty(&se->volume->open_mmaps))
			afs_volume_init_callback(se->volume);
	}

	up_read(&server->cell->vs_lock);
}

/*
 * 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 (atomic64_xchg(&vnode->cb_expires_at, AFS_NO_CB_PROMISE) != AFS_NO_CB_PROMISE) {
		vnode->cb_break++;
		vnode->cb_v_check = atomic_read(&vnode->volume->cb_v_break);
		afs_clear_permits(vnode);

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

		if (reason != afs_cb_break_for_deleted &&
		    vnode->status.type == AFS_FTYPE_FILE &&
		    atomic_read(&vnode->cb_nr_mmap))
			queue_work(system_unbound_wq, &vnode->cb_work);

		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);
}

/*
 * Look up a volume by volume ID under RCU conditions.
 */
static struct afs_volume *afs_lookup_volume_rcu(struct afs_cell *cell,
						afs_volid_t vid)
{
	struct afs_volume *volume = NULL;
	struct rb_node *p;
	int seq = 1;

	for (;;) {
		/* Unfortunately, rbtree walking doesn't give reliable results
		 * under just the RCU read lock, so we have to check for
		 * changes.
		 */
		seq++; /* 2 on the 1st/lockless path, otherwise odd */
		read_seqbegin_or_lock(&cell->volume_lock, &seq);

		p = rcu_dereference_raw(cell->volumes.rb_node);
		while (p) {
			volume = rb_entry(p, struct afs_volume, cell_node);

			if (volume->vid < vid)
				p = rcu_dereference_raw(p->rb_left);
			else if (volume->vid > vid)
				p = rcu_dereference_raw(p->rb_right);
			else
				break;
			volume = NULL;
		}

		if (volume && afs_try_get_volume(volume, afs_volume_trace_get_callback))
			break;
		if (!need_seqretry(&cell->volume_lock, seq))
			break;
		seq |= 1; /* Want a lock next time */
	}

	done_seqretry(&cell->volume_lock, seq);
	return volume;
}

/*
 * Allow the fileserver to break callbacks at the volume-level.  This is
 * typically done when, for example, a R/W volume is snapshotted to a R/O
 * volume (the only way to change an R/O volume).  It may also, however, happen
 * when a volserver takes control of a volume (offlining it, moving it, etc.).
 *
 * Every file in that volume will need to be reevaluated.
 */
static void afs_break_volume_callback(struct afs_server *server,
				      struct afs_volume *volume)
	__releases(RCU)
{
	struct afs_server_list *slist = rcu_dereference(volume->servers);
	unsigned int i, cb_v_break;

	write_lock(&volume->cb_v_break_lock);

	for (i = 0; i < slist->nr_servers; i++)
		if (slist->servers[i].server == server)
			slist->servers[i].cb_expires_at = AFS_NO_CB_PROMISE;
	volume->cb_expires_at = AFS_NO_CB_PROMISE;

	cb_v_break = atomic_inc_return_release(&volume->cb_v_break);
	trace_afs_cb_v_break(volume->vid, cb_v_break, afs_cb_break_for_volume_callback);

	write_unlock(&volume->cb_v_break_lock);
	rcu_read_unlock();

	if (!list_empty(&volume->open_mmaps))
		afs_volume_init_callback(volume);
}

/*
 * 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_volume *volume,
				   struct afs_fid *fid)
{
	struct super_block *sb;
	struct afs_vnode *vnode;
	struct inode *inode;

	/* See if we can find a matching inode - even an I_NEW inode needs to
	 * be marked as it can have its callback broken before we finish
	 * setting up the local inode.
	 */
	sb = rcu_dereference(volume->sb);
	if (!sb)
		return;

	inode = find_inode_rcu(sb, fid->vnode, afs_ilookup5_test_by_fid, fid);
	if (inode) {
		vnode = AFS_FS_I(inode);
		afs_break_callback(vnode, afs_cb_break_for_callback);
	} else {
		trace_afs_cb_miss(fid, afs_cb_break_for_callback);
	}
}

static void afs_break_some_callbacks(struct afs_server *server,
				     struct afs_callback_break *cbb,
				     size_t *_count)
{
	struct afs_callback_break *residue = cbb;
	struct afs_volume *volume;
	afs_volid_t vid = cbb->fid.vid;
	size_t i;

	rcu_read_lock();
	volume = afs_lookup_volume_rcu(server->cell, vid);
	if (cbb->fid.vnode == 0 && cbb->fid.unique == 0) {
		afs_break_volume_callback(server, volume);
		*_count -= 1;
		if (*_count)
			memmove(cbb, cbb + 1, sizeof(*cbb) * *_count);
	} else {
		/* TODO: Find all matching volumes if we couldn't match the server and
		 * break them anyway.
		 */

		for (i = *_count; i > 0; cbb++, i--) {
			if (cbb->fid.vid == vid) {
				_debug("- Fid { vl=%08llx n=%llu u=%u }",
				       cbb->fid.vid,
				       cbb->fid.vnode,
				       cbb->fid.unique);
				--*_count;
				if (volume)
					afs_break_one_callback(server, volume, &cbb->fid);
			} else {
				*residue++ = *cbb;
			}
		}
		rcu_read_unlock();
	}

	afs_put_volume(volume, afs_volume_trace_put_callback);
}

/*
 * 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);

	while (count > 0)
		afs_break_some_callbacks(server, callbacks, &count);
}