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
|
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
*/
#include <sys/zfs_context.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/vdev_file.h>
#include <sys/vdev_impl.h>
#include <sys/vdev_trim.h>
#include <sys/zio.h>
#include <sys/fs/zfs.h>
#include <sys/fm/fs/zfs.h>
#include <sys/abd.h>
#include <sys/fcntl.h>
#include <sys/vnode.h>
#include <sys/zfs_file.h>
#ifdef _KERNEL
#include <linux/falloc.h>
#endif
/*
* Virtual device vector for files.
*/
static taskq_t *vdev_file_taskq;
/*
* By default, the logical/physical ashift for file vdevs is set to
* SPA_MINBLOCKSHIFT (9). This allows all file vdevs to use 512B (1 << 9)
* blocksizes. Users may opt to change one or both of these for testing
* or performance reasons. Care should be taken as these values will
* impact the vdev_ashift setting which can only be set at vdev creation
* time.
*/
static unsigned long vdev_file_logical_ashift = SPA_MINBLOCKSHIFT;
static unsigned long vdev_file_physical_ashift = SPA_MINBLOCKSHIFT;
static void
vdev_file_hold(vdev_t *vd)
{
ASSERT(vd->vdev_path != NULL);
}
static void
vdev_file_rele(vdev_t *vd)
{
ASSERT(vd->vdev_path != NULL);
}
static mode_t
vdev_file_open_mode(spa_mode_t spa_mode)
{
mode_t mode = 0;
if ((spa_mode & SPA_MODE_READ) && (spa_mode & SPA_MODE_WRITE)) {
mode = O_RDWR;
} else if (spa_mode & SPA_MODE_READ) {
mode = O_RDONLY;
} else if (spa_mode & SPA_MODE_WRITE) {
mode = O_WRONLY;
}
return (mode | O_LARGEFILE);
}
static int
vdev_file_open(vdev_t *vd, uint64_t *psize, uint64_t *max_psize,
uint64_t *logical_ashift, uint64_t *physical_ashift)
{
vdev_file_t *vf;
zfs_file_t *fp;
zfs_file_attr_t zfa;
int error;
/*
* Rotational optimizations only make sense on block devices.
*/
vd->vdev_nonrot = B_TRUE;
/*
* Allow TRIM on file based vdevs. This may not always be supported,
* since it depends on your kernel version and underlying filesystem
* type but it is always safe to attempt.
*/
vd->vdev_has_trim = B_TRUE;
/*
* Disable secure TRIM on file based vdevs. There is no way to
* request this behavior from the underlying filesystem.
*/
vd->vdev_has_securetrim = B_FALSE;
/*
* We must have a pathname, and it must be absolute.
*/
if (vd->vdev_path == NULL || vd->vdev_path[0] != '/') {
vd->vdev_stat.vs_aux = VDEV_AUX_BAD_LABEL;
return (SET_ERROR(EINVAL));
}
/*
* Reopen the device if it's not currently open. Otherwise,
* just update the physical size of the device.
*/
if (vd->vdev_tsd != NULL) {
ASSERT(vd->vdev_reopening);
vf = vd->vdev_tsd;
goto skip_open;
}
vf = vd->vdev_tsd = kmem_zalloc(sizeof (vdev_file_t), KM_SLEEP);
/*
* We always open the files from the root of the global zone, even if
* we're in a local zone. If the user has gotten to this point, the
* administrator has already decided that the pool should be available
* to local zone users, so the underlying devices should be as well.
*/
ASSERT(vd->vdev_path != NULL && vd->vdev_path[0] == '/');
error = zfs_file_open(vd->vdev_path,
vdev_file_open_mode(spa_mode(vd->vdev_spa)), 0, &fp);
if (error) {
vd->vdev_stat.vs_aux = VDEV_AUX_OPEN_FAILED;
return (error);
}
vf->vf_file = fp;
#ifdef _KERNEL
/*
* Make sure it's a regular file.
*/
if (zfs_file_getattr(fp, &zfa)) {
return (SET_ERROR(ENODEV));
}
if (!S_ISREG(zfa.zfa_mode)) {
vd->vdev_stat.vs_aux = VDEV_AUX_OPEN_FAILED;
return (SET_ERROR(ENODEV));
}
#endif
skip_open:
error = zfs_file_getattr(vf->vf_file, &zfa);
if (error) {
vd->vdev_stat.vs_aux = VDEV_AUX_OPEN_FAILED;
return (error);
}
*max_psize = *psize = zfa.zfa_size;
*logical_ashift = vdev_file_logical_ashift;
*physical_ashift = vdev_file_physical_ashift;
return (0);
}
static void
vdev_file_close(vdev_t *vd)
{
vdev_file_t *vf = vd->vdev_tsd;
if (vd->vdev_reopening || vf == NULL)
return;
if (vf->vf_file != NULL) {
(void) zfs_file_close(vf->vf_file);
}
vd->vdev_delayed_close = B_FALSE;
kmem_free(vf, sizeof (vdev_file_t));
vd->vdev_tsd = NULL;
}
static void
vdev_file_io_strategy(void *arg)
{
zio_t *zio = (zio_t *)arg;
vdev_t *vd = zio->io_vd;
vdev_file_t *vf = vd->vdev_tsd;
ssize_t resid;
void *buf;
loff_t off;
ssize_t size;
int err;
off = zio->io_offset;
size = zio->io_size;
resid = 0;
if (zio->io_type == ZIO_TYPE_READ) {
buf = abd_borrow_buf(zio->io_abd, zio->io_size);
err = zfs_file_pread(vf->vf_file, buf, size, off, &resid);
abd_return_buf_copy(zio->io_abd, buf, size);
} else {
buf = abd_borrow_buf_copy(zio->io_abd, zio->io_size);
err = zfs_file_pwrite(vf->vf_file, buf, size, off, &resid);
abd_return_buf(zio->io_abd, buf, size);
}
zio->io_error = err;
if (resid != 0 && zio->io_error == 0)
zio->io_error = SET_ERROR(ENOSPC);
zio_delay_interrupt(zio);
}
static void
vdev_file_io_fsync(void *arg)
{
zio_t *zio = (zio_t *)arg;
vdev_file_t *vf = zio->io_vd->vdev_tsd;
zio->io_error = zfs_file_fsync(vf->vf_file, O_SYNC | O_DSYNC);
zio_interrupt(zio);
}
static void
vdev_file_io_start(zio_t *zio)
{
vdev_t *vd = zio->io_vd;
vdev_file_t *vf = vd->vdev_tsd;
if (zio->io_type == ZIO_TYPE_IOCTL) {
/* XXPOLICY */
if (!vdev_readable(vd)) {
zio->io_error = SET_ERROR(ENXIO);
zio_interrupt(zio);
return;
}
switch (zio->io_cmd) {
case DKIOCFLUSHWRITECACHE:
if (zfs_nocacheflush)
break;
/*
* We cannot safely call vfs_fsync() when PF_FSTRANS
* is set in the current context. Filesystems like
* XFS include sanity checks to verify it is not
* already set, see xfs_vm_writepage(). Therefore
* the sync must be dispatched to a different context.
*/
if (__spl_pf_fstrans_check()) {
VERIFY3U(taskq_dispatch(vdev_file_taskq,
vdev_file_io_fsync, zio, TQ_SLEEP), !=,
TASKQID_INVALID);
return;
}
zio->io_error = zfs_file_fsync(vf->vf_file,
O_SYNC | O_DSYNC);
break;
default:
zio->io_error = SET_ERROR(ENOTSUP);
}
zio_execute(zio);
return;
} else if (zio->io_type == ZIO_TYPE_TRIM) {
int mode = 0;
ASSERT3U(zio->io_size, !=, 0);
#ifdef __linux__
mode = FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE;
#endif
zio->io_error = zfs_file_fallocate(vf->vf_file,
mode, zio->io_offset, zio->io_size);
zio_execute(zio);
return;
}
zio->io_target_timestamp = zio_handle_io_delay(zio);
VERIFY3U(taskq_dispatch(vdev_file_taskq, vdev_file_io_strategy, zio,
TQ_SLEEP), !=, TASKQID_INVALID);
}
static void
vdev_file_io_done(zio_t *zio)
{
(void) zio;
}
vdev_ops_t vdev_file_ops = {
.vdev_op_init = NULL,
.vdev_op_fini = NULL,
.vdev_op_open = vdev_file_open,
.vdev_op_close = vdev_file_close,
.vdev_op_asize = vdev_default_asize,
.vdev_op_min_asize = vdev_default_min_asize,
.vdev_op_min_alloc = NULL,
.vdev_op_io_start = vdev_file_io_start,
.vdev_op_io_done = vdev_file_io_done,
.vdev_op_state_change = NULL,
.vdev_op_need_resilver = NULL,
.vdev_op_hold = vdev_file_hold,
.vdev_op_rele = vdev_file_rele,
.vdev_op_remap = NULL,
.vdev_op_xlate = vdev_default_xlate,
.vdev_op_rebuild_asize = NULL,
.vdev_op_metaslab_init = NULL,
.vdev_op_config_generate = NULL,
.vdev_op_nparity = NULL,
.vdev_op_ndisks = NULL,
.vdev_op_type = VDEV_TYPE_FILE, /* name of this vdev type */
.vdev_op_leaf = B_TRUE /* leaf vdev */
};
void
vdev_file_init(void)
{
vdev_file_taskq = taskq_create("z_vdev_file", MAX(boot_ncpus, 16),
minclsyspri, boot_ncpus, INT_MAX, TASKQ_DYNAMIC);
VERIFY(vdev_file_taskq);
}
void
vdev_file_fini(void)
{
taskq_destroy(vdev_file_taskq);
}
/*
* From userland we access disks just like files.
*/
#ifndef _KERNEL
vdev_ops_t vdev_disk_ops = {
.vdev_op_init = NULL,
.vdev_op_fini = NULL,
.vdev_op_open = vdev_file_open,
.vdev_op_close = vdev_file_close,
.vdev_op_asize = vdev_default_asize,
.vdev_op_min_asize = vdev_default_min_asize,
.vdev_op_min_alloc = NULL,
.vdev_op_io_start = vdev_file_io_start,
.vdev_op_io_done = vdev_file_io_done,
.vdev_op_state_change = NULL,
.vdev_op_need_resilver = NULL,
.vdev_op_hold = vdev_file_hold,
.vdev_op_rele = vdev_file_rele,
.vdev_op_remap = NULL,
.vdev_op_xlate = vdev_default_xlate,
.vdev_op_rebuild_asize = NULL,
.vdev_op_metaslab_init = NULL,
.vdev_op_config_generate = NULL,
.vdev_op_nparity = NULL,
.vdev_op_ndisks = NULL,
.vdev_op_type = VDEV_TYPE_DISK, /* name of this vdev type */
.vdev_op_leaf = B_TRUE /* leaf vdev */
};
#endif
ZFS_MODULE_PARAM(zfs_vdev_file, vdev_file_, logical_ashift, ULONG, ZMOD_RW,
"Logical ashift for file-based devices");
ZFS_MODULE_PARAM(zfs_vdev_file, vdev_file_, physical_ashift, ULONG, ZMOD_RW,
"Physical ashift for file-based devices");
|