Print this page
2619 asynchronous destruction of ZFS file systems
2747 SPA versioning with zfs feature flags
Reviewed by: Matt Ahrens <mahrens@delphix.com>
Reviewed by: George Wilson <gwilson@delphix.com>
Reviewed by: Richard Lowe <richlowe@richlowe.net>
Reviewed by: Dan Kruchinin <dan.kruchinin@gmail.com>
Approved by: Dan McDonald <danmcd@nexenta.com>
| Split |
Close |
| Expand all |
| Collapse all |
--- old/usr/src/uts/common/fs/zfs/sys/spa.h
+++ new/usr/src/uts/common/fs/zfs/sys/spa.h
1 1 /*
2 2 * CDDL HEADER START
3 3 *
4 4 * The contents of this file are subject to the terms of the
5 5 * Common Development and Distribution License (the "License").
6 6 * You may not use this file except in compliance with the License.
7 7 *
8 8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
9 9 * or http://www.opensolaris.org/os/licensing.
10 10 * See the License for the specific language governing permissions
11 11 * and limitations under the License.
12 12 *
|
↓ open down ↓ |
12 lines elided |
↑ open up ↑ |
13 13 * When distributing Covered Code, include this CDDL HEADER in each
14 14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
15 15 * If applicable, add the following below this CDDL HEADER, with the
16 16 * fields enclosed by brackets "[]" replaced with your own identifying
17 17 * information: Portions Copyright [yyyy] [name of copyright owner]
18 18 *
19 19 * CDDL HEADER END
20 20 */
21 21 /*
22 22 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
23 - * Copyright (c) 2011 by Delphix. All rights reserved.
23 + * Copyright (c) 2012 by Delphix. All rights reserved.
24 24 * Copyright 2011 Nexenta Systems, Inc. All rights reserved.
25 25 */
26 26
27 27 #ifndef _SYS_SPA_H
28 28 #define _SYS_SPA_H
29 29
30 30 #include <sys/avl.h>
31 31 #include <sys/zfs_context.h>
32 32 #include <sys/nvpair.h>
33 33 #include <sys/sysmacros.h>
34 34 #include <sys/types.h>
35 35 #include <sys/fs/zfs.h>
36 36
37 37 #ifdef __cplusplus
38 38 extern "C" {
39 39 #endif
40 40
41 41 /*
42 42 * Forward references that lots of things need.
43 43 */
44 44 typedef struct spa spa_t;
45 45 typedef struct vdev vdev_t;
46 46 typedef struct metaslab metaslab_t;
47 47 typedef struct metaslab_group metaslab_group_t;
48 48 typedef struct metaslab_class metaslab_class_t;
49 49 typedef struct zio zio_t;
50 50 typedef struct zilog zilog_t;
51 51 typedef struct spa_aux_vdev spa_aux_vdev_t;
52 52 typedef struct ddt ddt_t;
53 53 typedef struct ddt_entry ddt_entry_t;
54 54 struct dsl_pool;
55 55
56 56 /*
57 57 * General-purpose 32-bit and 64-bit bitfield encodings.
58 58 */
59 59 #define BF32_DECODE(x, low, len) P2PHASE((x) >> (low), 1U << (len))
60 60 #define BF64_DECODE(x, low, len) P2PHASE((x) >> (low), 1ULL << (len))
61 61 #define BF32_ENCODE(x, low, len) (P2PHASE((x), 1U << (len)) << (low))
62 62 #define BF64_ENCODE(x, low, len) (P2PHASE((x), 1ULL << (len)) << (low))
63 63
64 64 #define BF32_GET(x, low, len) BF32_DECODE(x, low, len)
65 65 #define BF64_GET(x, low, len) BF64_DECODE(x, low, len)
66 66
67 67 #define BF32_SET(x, low, len, val) \
68 68 ((x) ^= BF32_ENCODE((x >> low) ^ (val), low, len))
69 69 #define BF64_SET(x, low, len, val) \
70 70 ((x) ^= BF64_ENCODE((x >> low) ^ (val), low, len))
71 71
72 72 #define BF32_GET_SB(x, low, len, shift, bias) \
73 73 ((BF32_GET(x, low, len) + (bias)) << (shift))
74 74 #define BF64_GET_SB(x, low, len, shift, bias) \
75 75 ((BF64_GET(x, low, len) + (bias)) << (shift))
76 76
77 77 #define BF32_SET_SB(x, low, len, shift, bias, val) \
78 78 BF32_SET(x, low, len, ((val) >> (shift)) - (bias))
79 79 #define BF64_SET_SB(x, low, len, shift, bias, val) \
80 80 BF64_SET(x, low, len, ((val) >> (shift)) - (bias))
81 81
82 82 /*
83 83 * We currently support nine block sizes, from 512 bytes to 128K.
84 84 * We could go higher, but the benefits are near-zero and the cost
85 85 * of COWing a giant block to modify one byte would become excessive.
86 86 */
|
↓ open down ↓ |
53 lines elided |
↑ open up ↑ |
87 87 #define SPA_MINBLOCKSHIFT 9
88 88 #define SPA_MAXBLOCKSHIFT 17
89 89 #define SPA_MINBLOCKSIZE (1ULL << SPA_MINBLOCKSHIFT)
90 90 #define SPA_MAXBLOCKSIZE (1ULL << SPA_MAXBLOCKSHIFT)
91 91
92 92 #define SPA_BLOCKSIZES (SPA_MAXBLOCKSHIFT - SPA_MINBLOCKSHIFT + 1)
93 93
94 94 /*
95 95 * Size of block to hold the configuration data (a packed nvlist)
96 96 */
97 -#define SPA_CONFIG_BLOCKSIZE (1 << 14)
97 +#define SPA_CONFIG_BLOCKSIZE (1ULL << 14)
98 98
99 99 /*
100 100 * The DVA size encodings for LSIZE and PSIZE support blocks up to 32MB.
101 101 * The ASIZE encoding should be at least 64 times larger (6 more bits)
102 102 * to support up to 4-way RAID-Z mirror mode with worst-case gang block
103 103 * overhead, three DVAs per bp, plus one more bit in case we do anything
104 104 * else that expands the ASIZE.
105 105 */
106 106 #define SPA_LSIZEBITS 16 /* LSIZE up to 32M (2^16 * 512) */
107 107 #define SPA_PSIZEBITS 16 /* PSIZE up to 32M (2^16 * 512) */
108 108 #define SPA_ASIZEBITS 24 /* ASIZE up to 64 times larger */
109 109
110 110 /*
111 111 * All SPA data is represented by 128-bit data virtual addresses (DVAs).
112 112 * The members of the dva_t should be considered opaque outside the SPA.
113 113 */
114 114 typedef struct dva {
115 115 uint64_t dva_word[2];
116 116 } dva_t;
117 117
118 118 /*
119 119 * Each block has a 256-bit checksum -- strong enough for cryptographic hashes.
120 120 */
121 121 typedef struct zio_cksum {
122 122 uint64_t zc_word[4];
123 123 } zio_cksum_t;
124 124
125 125 /*
126 126 * Each block is described by its DVAs, time of birth, checksum, etc.
127 127 * The word-by-word, bit-by-bit layout of the blkptr is as follows:
128 128 *
129 129 * 64 56 48 40 32 24 16 8 0
130 130 * +-------+-------+-------+-------+-------+-------+-------+-------+
131 131 * 0 | vdev1 | GRID | ASIZE |
132 132 * +-------+-------+-------+-------+-------+-------+-------+-------+
133 133 * 1 |G| offset1 |
134 134 * +-------+-------+-------+-------+-------+-------+-------+-------+
135 135 * 2 | vdev2 | GRID | ASIZE |
136 136 * +-------+-------+-------+-------+-------+-------+-------+-------+
137 137 * 3 |G| offset2 |
138 138 * +-------+-------+-------+-------+-------+-------+-------+-------+
139 139 * 4 | vdev3 | GRID | ASIZE |
140 140 * +-------+-------+-------+-------+-------+-------+-------+-------+
141 141 * 5 |G| offset3 |
142 142 * +-------+-------+-------+-------+-------+-------+-------+-------+
143 143 * 6 |BDX|lvl| type | cksum | comp | PSIZE | LSIZE |
144 144 * +-------+-------+-------+-------+-------+-------+-------+-------+
145 145 * 7 | padding |
146 146 * +-------+-------+-------+-------+-------+-------+-------+-------+
147 147 * 8 | padding |
148 148 * +-------+-------+-------+-------+-------+-------+-------+-------+
149 149 * 9 | physical birth txg |
150 150 * +-------+-------+-------+-------+-------+-------+-------+-------+
151 151 * a | logical birth txg |
152 152 * +-------+-------+-------+-------+-------+-------+-------+-------+
153 153 * b | fill count |
154 154 * +-------+-------+-------+-------+-------+-------+-------+-------+
155 155 * c | checksum[0] |
156 156 * +-------+-------+-------+-------+-------+-------+-------+-------+
157 157 * d | checksum[1] |
158 158 * +-------+-------+-------+-------+-------+-------+-------+-------+
159 159 * e | checksum[2] |
160 160 * +-------+-------+-------+-------+-------+-------+-------+-------+
161 161 * f | checksum[3] |
162 162 * +-------+-------+-------+-------+-------+-------+-------+-------+
163 163 *
164 164 * Legend:
165 165 *
166 166 * vdev virtual device ID
167 167 * offset offset into virtual device
168 168 * LSIZE logical size
169 169 * PSIZE physical size (after compression)
170 170 * ASIZE allocated size (including RAID-Z parity and gang block headers)
171 171 * GRID RAID-Z layout information (reserved for future use)
172 172 * cksum checksum function
173 173 * comp compression function
174 174 * G gang block indicator
175 175 * B byteorder (endianness)
176 176 * D dedup
177 177 * X unused
178 178 * lvl level of indirection
179 179 * type DMU object type
180 180 * phys birth txg of block allocation; zero if same as logical birth txg
181 181 * log. birth transaction group in which the block was logically born
182 182 * fill count number of non-zero blocks under this bp
183 183 * checksum[4] 256-bit checksum of the data this bp describes
184 184 */
185 185 #define SPA_BLKPTRSHIFT 7 /* blkptr_t is 128 bytes */
186 186 #define SPA_DVAS_PER_BP 3 /* Number of DVAs in a bp */
187 187
188 188 typedef struct blkptr {
189 189 dva_t blk_dva[SPA_DVAS_PER_BP]; /* Data Virtual Addresses */
190 190 uint64_t blk_prop; /* size, compression, type, etc */
191 191 uint64_t blk_pad[2]; /* Extra space for the future */
192 192 uint64_t blk_phys_birth; /* txg when block was allocated */
193 193 uint64_t blk_birth; /* transaction group at birth */
194 194 uint64_t blk_fill; /* fill count */
195 195 zio_cksum_t blk_cksum; /* 256-bit checksum */
196 196 } blkptr_t;
197 197
198 198 /*
199 199 * Macros to get and set fields in a bp or DVA.
200 200 */
201 201 #define DVA_GET_ASIZE(dva) \
202 202 BF64_GET_SB((dva)->dva_word[0], 0, 24, SPA_MINBLOCKSHIFT, 0)
203 203 #define DVA_SET_ASIZE(dva, x) \
204 204 BF64_SET_SB((dva)->dva_word[0], 0, 24, SPA_MINBLOCKSHIFT, 0, x)
205 205
206 206 #define DVA_GET_GRID(dva) BF64_GET((dva)->dva_word[0], 24, 8)
207 207 #define DVA_SET_GRID(dva, x) BF64_SET((dva)->dva_word[0], 24, 8, x)
208 208
209 209 #define DVA_GET_VDEV(dva) BF64_GET((dva)->dva_word[0], 32, 32)
210 210 #define DVA_SET_VDEV(dva, x) BF64_SET((dva)->dva_word[0], 32, 32, x)
211 211
212 212 #define DVA_GET_OFFSET(dva) \
213 213 BF64_GET_SB((dva)->dva_word[1], 0, 63, SPA_MINBLOCKSHIFT, 0)
214 214 #define DVA_SET_OFFSET(dva, x) \
215 215 BF64_SET_SB((dva)->dva_word[1], 0, 63, SPA_MINBLOCKSHIFT, 0, x)
216 216
217 217 #define DVA_GET_GANG(dva) BF64_GET((dva)->dva_word[1], 63, 1)
218 218 #define DVA_SET_GANG(dva, x) BF64_SET((dva)->dva_word[1], 63, 1, x)
219 219
220 220 #define BP_GET_LSIZE(bp) \
221 221 BF64_GET_SB((bp)->blk_prop, 0, 16, SPA_MINBLOCKSHIFT, 1)
222 222 #define BP_SET_LSIZE(bp, x) \
223 223 BF64_SET_SB((bp)->blk_prop, 0, 16, SPA_MINBLOCKSHIFT, 1, x)
224 224
225 225 #define BP_GET_PSIZE(bp) \
226 226 BF64_GET_SB((bp)->blk_prop, 16, 16, SPA_MINBLOCKSHIFT, 1)
227 227 #define BP_SET_PSIZE(bp, x) \
228 228 BF64_SET_SB((bp)->blk_prop, 16, 16, SPA_MINBLOCKSHIFT, 1, x)
229 229
230 230 #define BP_GET_COMPRESS(bp) BF64_GET((bp)->blk_prop, 32, 8)
231 231 #define BP_SET_COMPRESS(bp, x) BF64_SET((bp)->blk_prop, 32, 8, x)
232 232
233 233 #define BP_GET_CHECKSUM(bp) BF64_GET((bp)->blk_prop, 40, 8)
234 234 #define BP_SET_CHECKSUM(bp, x) BF64_SET((bp)->blk_prop, 40, 8, x)
235 235
236 236 #define BP_GET_TYPE(bp) BF64_GET((bp)->blk_prop, 48, 8)
237 237 #define BP_SET_TYPE(bp, x) BF64_SET((bp)->blk_prop, 48, 8, x)
238 238
239 239 #define BP_GET_LEVEL(bp) BF64_GET((bp)->blk_prop, 56, 5)
240 240 #define BP_SET_LEVEL(bp, x) BF64_SET((bp)->blk_prop, 56, 5, x)
241 241
242 242 #define BP_GET_PROP_BIT_61(bp) BF64_GET((bp)->blk_prop, 61, 1)
243 243 #define BP_SET_PROP_BIT_61(bp, x) BF64_SET((bp)->blk_prop, 61, 1, x)
244 244
245 245 #define BP_GET_DEDUP(bp) BF64_GET((bp)->blk_prop, 62, 1)
246 246 #define BP_SET_DEDUP(bp, x) BF64_SET((bp)->blk_prop, 62, 1, x)
247 247
248 248 #define BP_GET_BYTEORDER(bp) (0 - BF64_GET((bp)->blk_prop, 63, 1))
249 249 #define BP_SET_BYTEORDER(bp, x) BF64_SET((bp)->blk_prop, 63, 1, x)
250 250
251 251 #define BP_PHYSICAL_BIRTH(bp) \
252 252 ((bp)->blk_phys_birth ? (bp)->blk_phys_birth : (bp)->blk_birth)
253 253
254 254 #define BP_SET_BIRTH(bp, logical, physical) \
|
↓ open down ↓ |
147 lines elided |
↑ open up ↑ |
255 255 { \
256 256 (bp)->blk_birth = (logical); \
257 257 (bp)->blk_phys_birth = ((logical) == (physical) ? 0 : (physical)); \
258 258 }
259 259
260 260 #define BP_GET_ASIZE(bp) \
261 261 (DVA_GET_ASIZE(&(bp)->blk_dva[0]) + DVA_GET_ASIZE(&(bp)->blk_dva[1]) + \
262 262 DVA_GET_ASIZE(&(bp)->blk_dva[2]))
263 263
264 264 #define BP_GET_UCSIZE(bp) \
265 - ((BP_GET_LEVEL(bp) > 0 || dmu_ot[BP_GET_TYPE(bp)].ot_metadata) ? \
265 + ((BP_GET_LEVEL(bp) > 0 || DMU_OT_IS_METADATA(BP_GET_TYPE(bp))) ? \
266 266 BP_GET_PSIZE(bp) : BP_GET_LSIZE(bp))
267 267
268 268 #define BP_GET_NDVAS(bp) \
269 269 (!!DVA_GET_ASIZE(&(bp)->blk_dva[0]) + \
270 270 !!DVA_GET_ASIZE(&(bp)->blk_dva[1]) + \
271 271 !!DVA_GET_ASIZE(&(bp)->blk_dva[2]))
272 272
273 273 #define BP_COUNT_GANG(bp) \
274 274 (DVA_GET_GANG(&(bp)->blk_dva[0]) + \
275 275 DVA_GET_GANG(&(bp)->blk_dva[1]) + \
276 276 DVA_GET_GANG(&(bp)->blk_dva[2]))
277 277
278 278 #define DVA_EQUAL(dva1, dva2) \
279 279 ((dva1)->dva_word[1] == (dva2)->dva_word[1] && \
280 280 (dva1)->dva_word[0] == (dva2)->dva_word[0])
281 281
282 282 #define BP_EQUAL(bp1, bp2) \
283 283 (BP_PHYSICAL_BIRTH(bp1) == BP_PHYSICAL_BIRTH(bp2) && \
284 284 DVA_EQUAL(&(bp1)->blk_dva[0], &(bp2)->blk_dva[0]) && \
285 285 DVA_EQUAL(&(bp1)->blk_dva[1], &(bp2)->blk_dva[1]) && \
286 286 DVA_EQUAL(&(bp1)->blk_dva[2], &(bp2)->blk_dva[2]))
287 287
288 288 #define ZIO_CHECKSUM_EQUAL(zc1, zc2) \
289 289 (0 == (((zc1).zc_word[0] - (zc2).zc_word[0]) | \
290 290 ((zc1).zc_word[1] - (zc2).zc_word[1]) | \
291 291 ((zc1).zc_word[2] - (zc2).zc_word[2]) | \
292 292 ((zc1).zc_word[3] - (zc2).zc_word[3])))
293 293
294 294 #define DVA_IS_VALID(dva) (DVA_GET_ASIZE(dva) != 0)
295 295
296 296 #define ZIO_SET_CHECKSUM(zcp, w0, w1, w2, w3) \
297 297 { \
298 298 (zcp)->zc_word[0] = w0; \
299 299 (zcp)->zc_word[1] = w1; \
300 300 (zcp)->zc_word[2] = w2; \
301 301 (zcp)->zc_word[3] = w3; \
302 302 }
303 303
304 304 #define BP_IDENTITY(bp) (&(bp)->blk_dva[0])
305 305 #define BP_IS_GANG(bp) DVA_GET_GANG(BP_IDENTITY(bp))
306 306 #define BP_IS_HOLE(bp) ((bp)->blk_birth == 0)
307 307
308 308 /* BP_IS_RAIDZ(bp) assumes no block compression */
309 309 #define BP_IS_RAIDZ(bp) (DVA_GET_ASIZE(&(bp)->blk_dva[0]) > \
310 310 BP_GET_PSIZE(bp))
311 311
312 312 #define BP_ZERO(bp) \
313 313 { \
314 314 (bp)->blk_dva[0].dva_word[0] = 0; \
315 315 (bp)->blk_dva[0].dva_word[1] = 0; \
316 316 (bp)->blk_dva[1].dva_word[0] = 0; \
317 317 (bp)->blk_dva[1].dva_word[1] = 0; \
318 318 (bp)->blk_dva[2].dva_word[0] = 0; \
319 319 (bp)->blk_dva[2].dva_word[1] = 0; \
320 320 (bp)->blk_prop = 0; \
321 321 (bp)->blk_pad[0] = 0; \
322 322 (bp)->blk_pad[1] = 0; \
323 323 (bp)->blk_phys_birth = 0; \
324 324 (bp)->blk_birth = 0; \
325 325 (bp)->blk_fill = 0; \
326 326 ZIO_SET_CHECKSUM(&(bp)->blk_cksum, 0, 0, 0, 0); \
327 327 }
328 328
329 329 /*
330 330 * Note: the byteorder is either 0 or -1, both of which are palindromes.
331 331 * This simplifies the endianness handling a bit.
332 332 */
333 333 #ifdef _BIG_ENDIAN
334 334 #define ZFS_HOST_BYTEORDER (0ULL)
335 335 #else
336 336 #define ZFS_HOST_BYTEORDER (-1ULL)
337 337 #endif
338 338
339 339 #define BP_SHOULD_BYTESWAP(bp) (BP_GET_BYTEORDER(bp) != ZFS_HOST_BYTEORDER)
340 340
341 341 #define BP_SPRINTF_LEN 320
342 342
343 343 /*
344 344 * This macro allows code sharing between zfs, libzpool, and mdb.
345 345 * 'func' is either snprintf() or mdb_snprintf().
346 346 * 'ws' (whitespace) can be ' ' for single-line format, '\n' for multi-line.
347 347 */
348 348 #define SPRINTF_BLKPTR(func, ws, buf, bp, type, checksum, compress) \
349 349 { \
350 350 static const char *copyname[] = \
351 351 { "zero", "single", "double", "triple" }; \
352 352 int size = BP_SPRINTF_LEN; \
353 353 int len = 0; \
354 354 int copies = 0; \
355 355 \
356 356 if (bp == NULL) { \
357 357 len = func(buf + len, size - len, "<NULL>"); \
358 358 } else if (BP_IS_HOLE(bp)) { \
359 359 len = func(buf + len, size - len, "<hole>"); \
360 360 } else { \
361 361 for (int d = 0; d < BP_GET_NDVAS(bp); d++) { \
362 362 const dva_t *dva = &bp->blk_dva[d]; \
363 363 if (DVA_IS_VALID(dva)) \
364 364 copies++; \
365 365 len += func(buf + len, size - len, \
366 366 "DVA[%d]=<%llu:%llx:%llx>%c", d, \
367 367 (u_longlong_t)DVA_GET_VDEV(dva), \
368 368 (u_longlong_t)DVA_GET_OFFSET(dva), \
369 369 (u_longlong_t)DVA_GET_ASIZE(dva), \
370 370 ws); \
371 371 } \
372 372 if (BP_IS_GANG(bp) && \
373 373 DVA_GET_ASIZE(&bp->blk_dva[2]) <= \
374 374 DVA_GET_ASIZE(&bp->blk_dva[1]) / 2) \
375 375 copies--; \
376 376 len += func(buf + len, size - len, \
377 377 "[L%llu %s] %s %s %s %s %s %s%c" \
378 378 "size=%llxL/%llxP birth=%lluL/%lluP fill=%llu%c" \
379 379 "cksum=%llx:%llx:%llx:%llx", \
380 380 (u_longlong_t)BP_GET_LEVEL(bp), \
381 381 type, \
382 382 checksum, \
383 383 compress, \
384 384 BP_GET_BYTEORDER(bp) == 0 ? "BE" : "LE", \
385 385 BP_IS_GANG(bp) ? "gang" : "contiguous", \
386 386 BP_GET_DEDUP(bp) ? "dedup" : "unique", \
387 387 copyname[copies], \
388 388 ws, \
389 389 (u_longlong_t)BP_GET_LSIZE(bp), \
390 390 (u_longlong_t)BP_GET_PSIZE(bp), \
391 391 (u_longlong_t)bp->blk_birth, \
392 392 (u_longlong_t)BP_PHYSICAL_BIRTH(bp), \
393 393 (u_longlong_t)bp->blk_fill, \
394 394 ws, \
395 395 (u_longlong_t)bp->blk_cksum.zc_word[0], \
|
↓ open down ↓ |
120 lines elided |
↑ open up ↑ |
396 396 (u_longlong_t)bp->blk_cksum.zc_word[1], \
397 397 (u_longlong_t)bp->blk_cksum.zc_word[2], \
398 398 (u_longlong_t)bp->blk_cksum.zc_word[3]); \
399 399 } \
400 400 ASSERT(len < size); \
401 401 }
402 402
403 403 #include <sys/dmu.h>
404 404
405 405 #define BP_GET_BUFC_TYPE(bp) \
406 - (((BP_GET_LEVEL(bp) > 0) || (dmu_ot[BP_GET_TYPE(bp)].ot_metadata)) ? \
407 - ARC_BUFC_METADATA : ARC_BUFC_DATA);
406 + (((BP_GET_LEVEL(bp) > 0) || (DMU_OT_IS_METADATA(BP_GET_TYPE(bp)))) ? \
407 + ARC_BUFC_METADATA : ARC_BUFC_DATA)
408 408
409 409 typedef enum spa_import_type {
410 410 SPA_IMPORT_EXISTING,
411 411 SPA_IMPORT_ASSEMBLE
412 412 } spa_import_type_t;
413 413
414 414 /* state manipulation functions */
415 415 extern int spa_open(const char *pool, spa_t **, void *tag);
416 416 extern int spa_open_rewind(const char *pool, spa_t **, void *tag,
417 417 nvlist_t *policy, nvlist_t **config);
418 -extern int spa_get_stats(const char *pool, nvlist_t **config,
419 - char *altroot, size_t buflen);
418 +extern int spa_get_stats(const char *pool, nvlist_t **config, char *altroot,
419 + size_t buflen);
420 420 extern int spa_create(const char *pool, nvlist_t *config, nvlist_t *props,
421 421 const char *history_str, nvlist_t *zplprops);
422 422 extern int spa_import_rootpool(char *devpath, char *devid);
423 423 extern int spa_import(const char *pool, nvlist_t *config, nvlist_t *props,
424 424 uint64_t flags);
425 425 extern nvlist_t *spa_tryimport(nvlist_t *tryconfig);
426 426 extern int spa_destroy(char *pool);
427 427 extern int spa_export(char *pool, nvlist_t **oldconfig, boolean_t force,
428 428 boolean_t hardforce);
429 429 extern int spa_reset(char *pool);
430 430 extern void spa_async_request(spa_t *spa, int flag);
431 431 extern void spa_async_unrequest(spa_t *spa, int flag);
432 432 extern void spa_async_suspend(spa_t *spa);
433 433 extern void spa_async_resume(spa_t *spa);
434 434 extern spa_t *spa_inject_addref(char *pool);
435 435 extern void spa_inject_delref(spa_t *spa);
436 436 extern void spa_scan_stat_init(spa_t *spa);
437 437 extern int spa_scan_get_stats(spa_t *spa, pool_scan_stat_t *ps);
438 438
439 439 #define SPA_ASYNC_CONFIG_UPDATE 0x01
440 440 #define SPA_ASYNC_REMOVE 0x02
441 441 #define SPA_ASYNC_PROBE 0x04
442 442 #define SPA_ASYNC_RESILVER_DONE 0x08
443 443 #define SPA_ASYNC_RESILVER 0x10
444 444 #define SPA_ASYNC_AUTOEXPAND 0x20
445 445 #define SPA_ASYNC_REMOVE_DONE 0x40
446 446 #define SPA_ASYNC_REMOVE_STOP 0x80
447 447
448 448 /*
449 449 * Controls the behavior of spa_vdev_remove().
450 450 */
451 451 #define SPA_REMOVE_UNSPARE 0x01
452 452 #define SPA_REMOVE_DONE 0x02
453 453
454 454 /* device manipulation */
455 455 extern int spa_vdev_add(spa_t *spa, nvlist_t *nvroot);
456 456 extern int spa_vdev_attach(spa_t *spa, uint64_t guid, nvlist_t *nvroot,
457 457 int replacing);
458 458 extern int spa_vdev_detach(spa_t *spa, uint64_t guid, uint64_t pguid,
459 459 int replace_done);
460 460 extern int spa_vdev_remove(spa_t *spa, uint64_t guid, boolean_t unspare);
461 461 extern boolean_t spa_vdev_remove_active(spa_t *spa);
462 462 extern int spa_vdev_setpath(spa_t *spa, uint64_t guid, const char *newpath);
463 463 extern int spa_vdev_setfru(spa_t *spa, uint64_t guid, const char *newfru);
464 464 extern int spa_vdev_split_mirror(spa_t *spa, char *newname, nvlist_t *config,
465 465 nvlist_t *props, boolean_t exp);
466 466
467 467 /* spare state (which is global across all pools) */
468 468 extern void spa_spare_add(vdev_t *vd);
469 469 extern void spa_spare_remove(vdev_t *vd);
470 470 extern boolean_t spa_spare_exists(uint64_t guid, uint64_t *pool, int *refcnt);
471 471 extern void spa_spare_activate(vdev_t *vd);
472 472
473 473 /* L2ARC state (which is global across all pools) */
474 474 extern void spa_l2cache_add(vdev_t *vd);
475 475 extern void spa_l2cache_remove(vdev_t *vd);
476 476 extern boolean_t spa_l2cache_exists(uint64_t guid, uint64_t *pool);
477 477 extern void spa_l2cache_activate(vdev_t *vd);
478 478 extern void spa_l2cache_drop(spa_t *spa);
479 479
480 480 /* scanning */
481 481 extern int spa_scan(spa_t *spa, pool_scan_func_t func);
482 482 extern int spa_scan_stop(spa_t *spa);
483 483
484 484 /* spa syncing */
485 485 extern void spa_sync(spa_t *spa, uint64_t txg); /* only for DMU use */
486 486 extern void spa_sync_allpools(void);
487 487
488 488 /*
489 489 * DEFERRED_FREE must be large enough that regular blocks are not
490 490 * deferred. XXX so can't we change it back to 1?
491 491 */
492 492 #define SYNC_PASS_DEFERRED_FREE 2 /* defer frees after this pass */
493 493 #define SYNC_PASS_DONT_COMPRESS 4 /* don't compress after this pass */
494 494 #define SYNC_PASS_REWRITE 1 /* rewrite new bps after this pass */
495 495
496 496 /* spa namespace global mutex */
497 497 extern kmutex_t spa_namespace_lock;
498 498
499 499 /*
500 500 * SPA configuration functions in spa_config.c
501 501 */
502 502
503 503 #define SPA_CONFIG_UPDATE_POOL 0
504 504 #define SPA_CONFIG_UPDATE_VDEVS 1
505 505
506 506 extern void spa_config_sync(spa_t *, boolean_t, boolean_t);
507 507 extern void spa_config_load(void);
508 508 extern nvlist_t *spa_all_configs(uint64_t *);
509 509 extern void spa_config_set(spa_t *spa, nvlist_t *config);
510 510 extern nvlist_t *spa_config_generate(spa_t *spa, vdev_t *vd, uint64_t txg,
511 511 int getstats);
512 512 extern void spa_config_update(spa_t *spa, int what);
513 513
514 514 /*
515 515 * Miscellaneous SPA routines in spa_misc.c
516 516 */
517 517
518 518 /* Namespace manipulation */
519 519 extern spa_t *spa_lookup(const char *name);
520 520 extern spa_t *spa_add(const char *name, nvlist_t *config, const char *altroot);
521 521 extern void spa_remove(spa_t *spa);
522 522 extern spa_t *spa_next(spa_t *prev);
523 523
524 524 /* Refcount functions */
525 525 extern void spa_open_ref(spa_t *spa, void *tag);
526 526 extern void spa_close(spa_t *spa, void *tag);
527 527 extern boolean_t spa_refcount_zero(spa_t *spa);
528 528
529 529 #define SCL_NONE 0x00
530 530 #define SCL_CONFIG 0x01
531 531 #define SCL_STATE 0x02
532 532 #define SCL_L2ARC 0x04 /* hack until L2ARC 2.0 */
533 533 #define SCL_ALLOC 0x08
534 534 #define SCL_ZIO 0x10
535 535 #define SCL_FREE 0x20
536 536 #define SCL_VDEV 0x40
537 537 #define SCL_LOCKS 7
538 538 #define SCL_ALL ((1 << SCL_LOCKS) - 1)
539 539 #define SCL_STATE_ALL (SCL_STATE | SCL_L2ARC | SCL_ZIO)
540 540
541 541 /* Pool configuration locks */
542 542 extern int spa_config_tryenter(spa_t *spa, int locks, void *tag, krw_t rw);
543 543 extern void spa_config_enter(spa_t *spa, int locks, void *tag, krw_t rw);
544 544 extern void spa_config_exit(spa_t *spa, int locks, void *tag);
545 545 extern int spa_config_held(spa_t *spa, int locks, krw_t rw);
546 546
547 547 /* Pool vdev add/remove lock */
548 548 extern uint64_t spa_vdev_enter(spa_t *spa);
549 549 extern uint64_t spa_vdev_config_enter(spa_t *spa);
550 550 extern void spa_vdev_config_exit(spa_t *spa, vdev_t *vd, uint64_t txg,
551 551 int error, char *tag);
552 552 extern int spa_vdev_exit(spa_t *spa, vdev_t *vd, uint64_t txg, int error);
553 553
554 554 /* Pool vdev state change lock */
555 555 extern void spa_vdev_state_enter(spa_t *spa, int oplock);
556 556 extern int spa_vdev_state_exit(spa_t *spa, vdev_t *vd, int error);
557 557
558 558 /* Log state */
559 559 typedef enum spa_log_state {
560 560 SPA_LOG_UNKNOWN = 0, /* unknown log state */
561 561 SPA_LOG_MISSING, /* missing log(s) */
562 562 SPA_LOG_CLEAR, /* clear the log(s) */
563 563 SPA_LOG_GOOD, /* log(s) are good */
564 564 } spa_log_state_t;
565 565
|
↓ open down ↓ |
136 lines elided |
↑ open up ↑ |
566 566 extern spa_log_state_t spa_get_log_state(spa_t *spa);
567 567 extern void spa_set_log_state(spa_t *spa, spa_log_state_t state);
568 568 extern int spa_offline_log(spa_t *spa);
569 569
570 570 /* Log claim callback */
571 571 extern void spa_claim_notify(zio_t *zio);
572 572
573 573 /* Accessor functions */
574 574 extern boolean_t spa_shutting_down(spa_t *spa);
575 575 extern struct dsl_pool *spa_get_dsl(spa_t *spa);
576 +extern boolean_t spa_is_initializing(spa_t *spa);
576 577 extern blkptr_t *spa_get_rootblkptr(spa_t *spa);
577 578 extern void spa_set_rootblkptr(spa_t *spa, const blkptr_t *bp);
578 579 extern void spa_altroot(spa_t *, char *, size_t);
579 580 extern int spa_sync_pass(spa_t *spa);
580 581 extern char *spa_name(spa_t *spa);
581 582 extern uint64_t spa_guid(spa_t *spa);
582 583 extern uint64_t spa_load_guid(spa_t *spa);
583 584 extern uint64_t spa_last_synced_txg(spa_t *spa);
584 585 extern uint64_t spa_first_txg(spa_t *spa);
585 586 extern uint64_t spa_syncing_txg(spa_t *spa);
586 587 extern uint64_t spa_version(spa_t *spa);
587 588 extern pool_state_t spa_state(spa_t *spa);
588 589 extern spa_load_state_t spa_load_state(spa_t *spa);
589 590 extern uint64_t spa_freeze_txg(spa_t *spa);
590 591 extern uint64_t spa_get_asize(spa_t *spa, uint64_t lsize);
591 592 extern uint64_t spa_get_dspace(spa_t *spa);
592 593 extern void spa_update_dspace(spa_t *spa);
593 594 extern uint64_t spa_version(spa_t *spa);
594 595 extern boolean_t spa_deflate(spa_t *spa);
595 596 extern metaslab_class_t *spa_normal_class(spa_t *spa);
596 597 extern metaslab_class_t *spa_log_class(spa_t *spa);
|
↓ open down ↓ |
11 lines elided |
↑ open up ↑ |
597 598 extern int spa_max_replication(spa_t *spa);
598 599 extern int spa_prev_software_version(spa_t *spa);
599 600 extern int spa_busy(void);
600 601 extern uint8_t spa_get_failmode(spa_t *spa);
601 602 extern boolean_t spa_suspended(spa_t *spa);
602 603 extern uint64_t spa_bootfs(spa_t *spa);
603 604 extern uint64_t spa_delegation(spa_t *spa);
604 605 extern objset_t *spa_meta_objset(spa_t *spa);
605 606
606 607 /* Miscellaneous support routines */
608 +extern void spa_activate_mos_feature(spa_t *spa, const char *feature);
609 +extern void spa_deactivate_mos_feature(spa_t *spa, const char *feature);
607 610 extern int spa_rename(const char *oldname, const char *newname);
608 611 extern spa_t *spa_by_guid(uint64_t pool_guid, uint64_t device_guid);
609 612 extern boolean_t spa_guid_exists(uint64_t pool_guid, uint64_t device_guid);
610 613 extern char *spa_strdup(const char *);
611 614 extern void spa_strfree(char *);
612 615 extern uint64_t spa_get_random(uint64_t range);
613 616 extern uint64_t spa_generate_guid(spa_t *spa);
614 617 extern void sprintf_blkptr(char *buf, const blkptr_t *bp);
615 618 extern void spa_freeze(spa_t *spa);
616 619 extern int spa_change_guid(spa_t *spa);
617 620 extern void spa_upgrade(spa_t *spa, uint64_t version);
618 621 extern void spa_evict_all(void);
619 622 extern vdev_t *spa_lookup_by_guid(spa_t *spa, uint64_t guid,
620 623 boolean_t l2cache);
621 624 extern boolean_t spa_has_spare(spa_t *, uint64_t guid);
622 625 extern uint64_t dva_get_dsize_sync(spa_t *spa, const dva_t *dva);
623 626 extern uint64_t bp_get_dsize_sync(spa_t *spa, const blkptr_t *bp);
624 627 extern uint64_t bp_get_dsize(spa_t *spa, const blkptr_t *bp);
625 628 extern boolean_t spa_has_slogs(spa_t *spa);
626 629 extern boolean_t spa_is_root(spa_t *spa);
627 630 extern boolean_t spa_writeable(spa_t *spa);
628 631
629 632 extern int spa_mode(spa_t *spa);
630 633 extern uint64_t strtonum(const char *str, char **nptr);
631 634
632 635 /* history logging */
633 636 typedef enum history_log_type {
634 637 LOG_CMD_POOL_CREATE,
635 638 LOG_CMD_NORMAL,
636 639 LOG_INTERNAL
637 640 } history_log_type_t;
638 641
639 642 typedef struct history_arg {
640 643 char *ha_history_str;
641 644 history_log_type_t ha_log_type;
642 645 history_internal_events_t ha_event;
643 646 char *ha_zone;
644 647 uid_t ha_uid;
645 648 } history_arg_t;
646 649
647 650 extern char *spa_his_ievent_table[];
648 651
649 652 extern void spa_history_create_obj(spa_t *spa, dmu_tx_t *tx);
650 653 extern int spa_history_get(spa_t *spa, uint64_t *offset, uint64_t *len_read,
651 654 char *his_buf);
652 655 extern int spa_history_log(spa_t *spa, const char *his_buf,
653 656 history_log_type_t what);
654 657 extern void spa_history_log_internal(history_internal_events_t event,
655 658 spa_t *spa, dmu_tx_t *tx, const char *fmt, ...);
656 659 extern void spa_history_log_version(spa_t *spa, history_internal_events_t evt);
657 660
658 661 /* error handling */
659 662 struct zbookmark;
660 663 extern void spa_log_error(spa_t *spa, zio_t *zio);
661 664 extern void zfs_ereport_post(const char *class, spa_t *spa, vdev_t *vd,
662 665 zio_t *zio, uint64_t stateoroffset, uint64_t length);
663 666 extern void zfs_post_remove(spa_t *spa, vdev_t *vd);
664 667 extern void zfs_post_state_change(spa_t *spa, vdev_t *vd);
665 668 extern void zfs_post_autoreplace(spa_t *spa, vdev_t *vd);
666 669 extern uint64_t spa_get_errlog_size(spa_t *spa);
667 670 extern int spa_get_errlog(spa_t *spa, void *uaddr, size_t *count);
668 671 extern void spa_errlog_rotate(spa_t *spa);
669 672 extern void spa_errlog_drain(spa_t *spa);
670 673 extern void spa_errlog_sync(spa_t *spa, uint64_t txg);
671 674 extern void spa_get_errlists(spa_t *spa, avl_tree_t *last, avl_tree_t *scrub);
672 675
673 676 /* vdev cache */
674 677 extern void vdev_cache_stat_init(void);
675 678 extern void vdev_cache_stat_fini(void);
676 679
677 680 /* Initialization and termination */
678 681 extern void spa_init(int flags);
679 682 extern void spa_fini(void);
680 683 extern void spa_boot_init();
681 684
682 685 /* properties */
683 686 extern int spa_prop_set(spa_t *spa, nvlist_t *nvp);
684 687 extern int spa_prop_get(spa_t *spa, nvlist_t **nvp);
685 688 extern void spa_prop_clear_bootfs(spa_t *spa, uint64_t obj, dmu_tx_t *tx);
686 689 extern void spa_configfile_set(spa_t *, nvlist_t *, boolean_t);
687 690
688 691 /* asynchronous event notification */
689 692 extern void spa_event_notify(spa_t *spa, vdev_t *vdev, const char *name);
690 693
691 694 #ifdef ZFS_DEBUG
692 695 #define dprintf_bp(bp, fmt, ...) do { \
693 696 if (zfs_flags & ZFS_DEBUG_DPRINTF) { \
694 697 char *__blkbuf = kmem_alloc(BP_SPRINTF_LEN, KM_SLEEP); \
695 698 sprintf_blkptr(__blkbuf, (bp)); \
696 699 dprintf(fmt " %s\n", __VA_ARGS__, __blkbuf); \
697 700 kmem_free(__blkbuf, BP_SPRINTF_LEN); \
698 701 } \
699 702 _NOTE(CONSTCOND) } while (0)
700 703 #else
701 704 #define dprintf_bp(bp, fmt, ...)
702 705 #endif
703 706
704 707 extern boolean_t spa_debug_enabled(spa_t *spa);
705 708 #define spa_dbgmsg(spa, ...) \
706 709 { \
707 710 if (spa_debug_enabled(spa)) \
708 711 zfs_dbgmsg(__VA_ARGS__); \
709 712 }
710 713
711 714 extern int spa_mode_global; /* mode, e.g. FREAD | FWRITE */
712 715
713 716 #ifdef __cplusplus
714 717 }
715 718 #endif
716 719
717 720 #endif /* _SYS_SPA_H */
|
↓ open down ↓ |
101 lines elided |
↑ open up ↑ |
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX