Print this page
NEX-13140 DVA-throttle support for special-class
Reviewed by: Sanjay Nadkarni <sanjay.nadkarni@nexenta.com>
Reviewed by: Saso Kiselkov <saso.kiselkov@nexenta.com>
Reviewed by: Yuri Pankov <yuri.pankov@nexenta.com>
NEX-9752 backport illumos 6950 ARC should cache compressed data
Reviewed by: Saso Kiselkov <saso.kiselkov@nexenta.com>
Reviewed by: Yuri Pankov <yuri.pankov@nexenta.com>
6950 ARC should cache compressed data
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Dan Kimmel <dan.kimmel@delphix.com>
Reviewed by: Matt Ahrens <mahrens@delphix.com>
Reviewed by: Paul Dagnelie <pcd@delphix.com>
Reviewed by: Don Brady <don.brady@intel.com>
Reviewed by: Richard Elling <Richard.Elling@RichardElling.com>
Approved by: Richard Lowe <richlowe@richlowe.net>
NEX-5795 Rename 'wrc' as 'wbc' in the source and in the tech docs
Reviewed by: Alex Aizman <alex.aizman@nexenta.com>
Reviewed by: Sanjay Nadkarni <sanjay.nadkarni@nexenta.com>
Reviewed by: Alek Pinchuk <alek.pinchuk@nexenta.com>
NEX-5318 Cleanup specialclass property (obsolete, not used) and fix related meta-to-special case
Reviewed by: Roman Strashkin <roman.strashkin@nexenta.com>
Reviewed by: Alek Pinchuk <alek.pinchuk@nexenta.com>
NEX-5058 WBC: Race between the purging of window and opening new one
Reviewed by: Alek Pinchuk <alek.pinchuk@nexenta.com>
Reviewed by: Alex Aizman <alex.aizman@nexenta.com>
NEX-2830 ZFS smart compression
Reviewed by: Alek Pinchuk <alek.pinchuk@nexenta.com>
Reviewed by: Roman Strashkin <roman.strashkin@nexenta.com>
4185 add new cryptographic checksums to ZFS: SHA-512, Skein, Edon-R (fix studio build)
4185 add new cryptographic checksums to ZFS: SHA-512, Skein, Edon-R
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Saso Kiselkov <saso.kiselkov@nexenta.com>
Reviewed by: Richard Lowe <richlowe@richlowe.net>
Approved by: Garrett D'Amore <garrett@damore.org>
NEX-4582 update wrc test cases for allow to use write back cache per tree of datasets
Reviewed by: Steve Peng <steve.peng@nexenta.com>
Reviewed by: Alex Aizman <alex.aizman@nexenta.com>
5960 zfs recv should prefetch indirect blocks
5925 zfs receive -o origin=
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
NEX-4476 WRC: Allow to use write back cache per tree of datasets
Reviewed by: Alek Pinchuk <alek.pinchuk@nexenta.com>
Reviewed by: Alex Aizman <alex.aizman@nexenta.com>
Revert "NEX-4476 WRC: Allow to use write back cache per tree of datasets"
This reverts commit fe97b74444278a6f36fec93179133641296312da.
NEX-4476 WRC: Allow to use write back cache per tree of datasets
Reviewed by: Alek Pinchuk <alek.pinchuk@nexenta.com>
Reviewed by: Alex Aizman <alex.aizman@nexenta.com>
NEX-4044 remove sha1crc32 in preparation with upstream merge of edon-r and skien
Reviewed by: Saso Kiselkov <saso.kiselkov@nexenta.com>
Reviewed by: Alek Pinchuk <alek@nexenta.com>
Conflicts:
usr/src/uts/common/fs/zfs/sys/zio_checksum.h
NEX-4028 use lz4 by default
Reviewed by: Saso Kiselkov <saso.kiselkov@nexenta.com>
Reviewed by: Alek Pinchuk <alek.pinchuk@nexenta.com>
NEX-4003 WRC: System panics on debug build
Reviewed by: Alek Pinchuk <alek.pinchuk@nexenta.com>
Reviewed by: Josef 'Jeff' Sipek <josef.sipek@nexenta.com>
NEX-3508 CLONE - Port NEX-2946 Add UNMAP/TRIM functionality to ZFS and illumos
Reviewed by: Josef Sipek <josef.sipek@nexenta.com>
Reviewed by: Alek Pinchuk <alek.pinchuk@nexenta.com>
Conflicts:
usr/src/uts/common/io/scsi/targets/sd.c
usr/src/uts/common/sys/scsi/targets/sddef.h
OS-80 support for vdev and CoS properties for the new I/O scheduler
OS-95 lint warning introduced by OS-61
OS-70 remove zio timer code
Issues #7: Reconsile L2ARC and "special" use by datasets
Support for secondarycache=data option
Align mutex tables in arc.c and dbuf.c to 64 bytes (cache line), place each kmutex_t on cache line by itself to avoid false sharing
re #12643 rb4064 ZFS meta refactoring - vdev utilization tracking, auto-dedup
re #12585 rb4049 ZFS++ work port - refactoring to improve separation of open/closed code, bug fixes, performance improvements - open code
re #12393 rb3935 Kerberos and smbd disagree about who is our AD server (fix elf runtime attributes check)
re #11612 rb3907 Failing vdev of a mirrored pool should not take zfs operations out of action for extended periods of time.
re #8346 rb2639 KT disk failures
Bug 11205: add missing libzfs_closed_stubs.c to fix opensource-only build.
ZFS plus work: special vdevs, cos, cos/vdev properties
| Split |
Close |
| Expand all |
| Collapse all |
--- old/usr/src/uts/common/fs/zfs/sys/zio.h
+++ new/usr/src/uts/common/fs/zfs/sys/zio.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 *
13 13 * When distributing Covered Code, include this CDDL HEADER in each
|
↓ open down ↓ |
13 lines elided |
↑ open up ↑ |
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 /*
23 23 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
24 - * Copyright 2011 Nexenta Systems, Inc. All rights reserved.
24 + * Copyright 2016 Nexenta Systems, Inc. All rights reserved.
25 25 * Copyright (c) 2012, 2017 by Delphix. All rights reserved.
26 26 * Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
27 27 * Copyright (c) 2013, Joyent, Inc. All rights reserved.
28 28 * Copyright 2016 Toomas Soome <tsoome@me.com>
29 29 */
30 30
31 31 #ifndef _ZIO_H
32 32 #define _ZIO_H
33 33
34 34 #include <sys/zio_priority.h>
35 35 #include <sys/zfs_context.h>
36 36 #include <sys/spa.h>
|
↓ open down ↓ |
2 lines elided |
↑ open up ↑ |
37 37 #include <sys/txg.h>
38 38 #include <sys/avl.h>
39 39 #include <sys/fs/zfs.h>
40 40 #include <sys/zio_impl.h>
41 41
42 42 #ifdef __cplusplus
43 43 extern "C" {
44 44 #endif
45 45
46 46 /*
47 + * Checksum state w.r.t. SHA256 acceleration.
48 + */
49 +typedef enum {
50 + CKSTATE_NONE = 0,
51 + CKSTATE_WAITING,
52 + CKSTATE_CHECKSUMMING,
53 + CKSTATE_CHECKSUM_DONE
54 +} zio_checksum_state_t;
55 +
56 +/*
47 57 * Embedded checksum
48 58 */
49 59 #define ZEC_MAGIC 0x210da7ab10c7a11ULL
50 60
51 61 typedef struct zio_eck {
52 62 uint64_t zec_magic; /* for validation, endianness */
53 63 zio_cksum_t zec_cksum; /* 256-bit checksum */
54 64 } zio_eck_t;
55 65
56 66 /*
57 67 * Gang block headers are self-checksumming and contain an array
58 68 * of block pointers.
59 69 */
60 70 #define SPA_GANGBLOCKSIZE SPA_MINBLOCKSIZE
61 71 #define SPA_GBH_NBLKPTRS ((SPA_GANGBLOCKSIZE - \
62 72 sizeof (zio_eck_t)) / sizeof (blkptr_t))
63 73 #define SPA_GBH_FILLER ((SPA_GANGBLOCKSIZE - \
64 74 sizeof (zio_eck_t) - \
65 75 (SPA_GBH_NBLKPTRS * sizeof (blkptr_t))) /\
66 76 sizeof (uint64_t))
67 77
68 78 typedef struct zio_gbh {
69 79 blkptr_t zg_blkptr[SPA_GBH_NBLKPTRS];
70 80 uint64_t zg_filler[SPA_GBH_FILLER];
71 81 zio_eck_t zg_tail;
72 82 } zio_gbh_phys_t;
73 83
74 84 enum zio_checksum {
75 85 ZIO_CHECKSUM_INHERIT = 0,
76 86 ZIO_CHECKSUM_ON,
77 87 ZIO_CHECKSUM_OFF,
78 88 ZIO_CHECKSUM_LABEL,
79 89 ZIO_CHECKSUM_GANG_HEADER,
80 90 ZIO_CHECKSUM_ZILOG,
81 91 ZIO_CHECKSUM_FLETCHER_2,
82 92 ZIO_CHECKSUM_FLETCHER_4,
83 93 ZIO_CHECKSUM_SHA256,
84 94 ZIO_CHECKSUM_ZILOG2,
85 95 ZIO_CHECKSUM_NOPARITY,
86 96 ZIO_CHECKSUM_SHA512,
87 97 ZIO_CHECKSUM_SKEIN,
88 98 ZIO_CHECKSUM_EDONR,
89 99 ZIO_CHECKSUM_FUNCTIONS
90 100 };
91 101
92 102 /*
93 103 * The number of "legacy" compression functions which can be set on individual
94 104 * objects.
95 105 */
96 106 #define ZIO_CHECKSUM_LEGACY_FUNCTIONS ZIO_CHECKSUM_ZILOG2
97 107
98 108 #define ZIO_CHECKSUM_ON_VALUE ZIO_CHECKSUM_FLETCHER_4
99 109 #define ZIO_CHECKSUM_DEFAULT ZIO_CHECKSUM_ON
100 110
101 111 #define ZIO_CHECKSUM_MASK 0xffULL
102 112 #define ZIO_CHECKSUM_VERIFY (1 << 8)
103 113
104 114 #define ZIO_DEDUPCHECKSUM ZIO_CHECKSUM_SHA256
105 115 #define ZIO_DEDUPDITTO_MIN 100
106 116
107 117 /*
108 118 * The number of "legacy" compression functions which can be set on individual
109 119 * objects.
110 120 */
111 121 #define ZIO_COMPRESS_LEGACY_FUNCTIONS ZIO_COMPRESS_LZ4
112 122
113 123 /*
114 124 * The meaning of "compress = on" selected by the compression features enabled
115 125 * on a given pool.
116 126 */
117 127 #define ZIO_COMPRESS_LEGACY_ON_VALUE ZIO_COMPRESS_LZJB
118 128 #define ZIO_COMPRESS_LZ4_ON_VALUE ZIO_COMPRESS_LZ4
119 129
120 130 #define ZIO_COMPRESS_DEFAULT ZIO_COMPRESS_OFF
121 131
122 132 #define BOOTFS_COMPRESS_VALID(compress) \
123 133 ((compress) == ZIO_COMPRESS_LZJB || \
124 134 (compress) == ZIO_COMPRESS_LZ4 || \
125 135 (compress) == ZIO_COMPRESS_GZIP_1 || \
126 136 (compress) == ZIO_COMPRESS_GZIP_2 || \
127 137 (compress) == ZIO_COMPRESS_GZIP_3 || \
128 138 (compress) == ZIO_COMPRESS_GZIP_4 || \
129 139 (compress) == ZIO_COMPRESS_GZIP_5 || \
130 140 (compress) == ZIO_COMPRESS_GZIP_6 || \
|
↓ open down ↓ |
74 lines elided |
↑ open up ↑ |
131 141 (compress) == ZIO_COMPRESS_GZIP_7 || \
132 142 (compress) == ZIO_COMPRESS_GZIP_8 || \
133 143 (compress) == ZIO_COMPRESS_GZIP_9 || \
134 144 (compress) == ZIO_COMPRESS_ON || \
135 145 (compress) == ZIO_COMPRESS_OFF)
136 146
137 147 #define ZIO_FAILURE_MODE_WAIT 0
138 148 #define ZIO_FAILURE_MODE_CONTINUE 1
139 149 #define ZIO_FAILURE_MODE_PANIC 2
140 150
151 +/*
152 + * Macro for asserting validity of the priorities obtained by conversion
153 + * from CoS/vdev properties
154 + */
155 +#define ZIO_PRIORITY_QUEUEABLE_VALID(prio) \
156 + (((prio) >= ZIO_PRIORITY_SYNC_READ) && \
157 + ((prio) < ZIO_PRIORITY_NUM_QUEUEABLE))
158 +
159 +#define ZIO_PIPELINE_CONTINUE 0x100
160 +#define ZIO_PIPELINE_STOP 0x101
161 +#define ZIO_PIPELINE_RESTART_STAGE 0x102
162 +
141 163 enum zio_flag {
142 164 /*
143 165 * Flags inherited by gang, ddt, and vdev children,
144 166 * and that must be equal for two zios to aggregate
145 167 */
146 168 ZIO_FLAG_DONT_AGGREGATE = 1 << 0,
147 169 ZIO_FLAG_IO_REPAIR = 1 << 1,
148 170 ZIO_FLAG_SELF_HEAL = 1 << 2,
149 171 ZIO_FLAG_RESILVER = 1 << 3,
150 172 ZIO_FLAG_SCRUB = 1 << 4,
151 173 ZIO_FLAG_SCAN_THREAD = 1 << 5,
152 174 ZIO_FLAG_PHYSICAL = 1 << 6,
153 175
154 176 #define ZIO_FLAG_AGG_INHERIT (ZIO_FLAG_CANFAIL - 1)
155 177
156 178 /*
157 179 * Flags inherited by ddt, gang, and vdev children.
158 180 */
159 181 ZIO_FLAG_CANFAIL = 1 << 7, /* must be first for INHERIT */
160 182 ZIO_FLAG_SPECULATIVE = 1 << 8,
161 183 ZIO_FLAG_CONFIG_WRITER = 1 << 9,
162 184 ZIO_FLAG_DONT_RETRY = 1 << 10,
163 185 ZIO_FLAG_DONT_CACHE = 1 << 11,
164 186 ZIO_FLAG_NODATA = 1 << 12,
165 187 ZIO_FLAG_INDUCE_DAMAGE = 1 << 13,
166 188 ZIO_FLAG_IO_ALLOCATING = 1 << 14,
167 189
168 190 #define ZIO_FLAG_DDT_INHERIT (ZIO_FLAG_IO_RETRY - 1)
169 191 #define ZIO_FLAG_GANG_INHERIT (ZIO_FLAG_IO_RETRY - 1)
170 192
171 193 /*
172 194 * Flags inherited by vdev children.
173 195 */
174 196 ZIO_FLAG_IO_RETRY = 1 << 15, /* must be first for INHERIT */
175 197 ZIO_FLAG_PROBE = 1 << 16,
176 198 ZIO_FLAG_TRYHARD = 1 << 17,
177 199 ZIO_FLAG_OPTIONAL = 1 << 18,
178 200
179 201 #define ZIO_FLAG_VDEV_INHERIT (ZIO_FLAG_DONT_QUEUE - 1)
180 202
181 203 /*
182 204 * Flags not inherited by any children.
183 205 */
184 206 ZIO_FLAG_DONT_QUEUE = 1 << 19, /* must be first for INHERIT */
185 207 ZIO_FLAG_DONT_PROPAGATE = 1 << 20,
186 208 ZIO_FLAG_IO_BYPASS = 1 << 21,
187 209 ZIO_FLAG_IO_REWRITE = 1 << 22,
188 210 ZIO_FLAG_RAW = 1 << 23,
189 211 ZIO_FLAG_GANG_CHILD = 1 << 24,
190 212 ZIO_FLAG_DDT_CHILD = 1 << 25,
191 213 ZIO_FLAG_GODFATHER = 1 << 26,
192 214 ZIO_FLAG_NOPWRITE = 1 << 27,
193 215 ZIO_FLAG_REEXECUTED = 1 << 28,
194 216 ZIO_FLAG_DELEGATED = 1 << 29,
195 217 };
196 218
197 219 #define ZIO_FLAG_MUSTSUCCEED 0
198 220
|
↓ open down ↓ |
48 lines elided |
↑ open up ↑ |
199 221 #define ZIO_DDT_CHILD_FLAGS(zio) \
200 222 (((zio)->io_flags & ZIO_FLAG_DDT_INHERIT) | \
201 223 ZIO_FLAG_DDT_CHILD | ZIO_FLAG_CANFAIL)
202 224
203 225 #define ZIO_GANG_CHILD_FLAGS(zio) \
204 226 (((zio)->io_flags & ZIO_FLAG_GANG_INHERIT) | \
205 227 ZIO_FLAG_GANG_CHILD | ZIO_FLAG_CANFAIL)
206 228
207 229 #define ZIO_VDEV_CHILD_FLAGS(zio) \
208 230 (((zio)->io_flags & ZIO_FLAG_VDEV_INHERIT) | \
209 - ZIO_FLAG_DONT_PROPAGATE | ZIO_FLAG_CANFAIL)
231 + ZIO_FLAG_CANFAIL)
210 232
211 -#define ZIO_CHILD_BIT(x) (1 << (x))
212 -#define ZIO_CHILD_BIT_IS_SET(val, x) ((val) & (1 << (x)))
213 -
214 233 enum zio_child {
215 234 ZIO_CHILD_VDEV = 0,
216 235 ZIO_CHILD_GANG,
217 236 ZIO_CHILD_DDT,
218 237 ZIO_CHILD_LOGICAL,
219 238 ZIO_CHILD_TYPES
220 239 };
221 240
222 -#define ZIO_CHILD_VDEV_BIT ZIO_CHILD_BIT(ZIO_CHILD_VDEV)
223 -#define ZIO_CHILD_GANG_BIT ZIO_CHILD_BIT(ZIO_CHILD_GANG)
224 -#define ZIO_CHILD_DDT_BIT ZIO_CHILD_BIT(ZIO_CHILD_DDT)
225 -#define ZIO_CHILD_LOGICAL_BIT ZIO_CHILD_BIT(ZIO_CHILD_LOGICAL)
226 -#define ZIO_CHILD_ALL_BITS \
227 - (ZIO_CHILD_VDEV_BIT | ZIO_CHILD_GANG_BIT | \
228 - ZIO_CHILD_DDT_BIT | ZIO_CHILD_LOGICAL_BIT)
229 -
230 241 enum zio_wait_type {
231 242 ZIO_WAIT_READY = 0,
232 243 ZIO_WAIT_DONE,
233 244 ZIO_WAIT_TYPES
234 245 };
235 246
236 247 /*
237 248 * We'll take the unused errnos, 'EBADE' and 'EBADR' (from the Convergent
238 249 * graveyard) to indicate checksum errors and fragmentation.
239 250 */
240 251 #define ECKSUM EBADE
241 252 #define EFRAGS EBADR
242 253
243 254 typedef void zio_done_func_t(zio_t *zio);
244 255
245 256 extern boolean_t zio_dva_throttle_enabled;
246 257 extern const char *zio_type_name[ZIO_TYPES];
247 258
259 +struct range_tree;
260 +
248 261 /*
249 262 * A bookmark is a four-tuple <objset, object, level, blkid> that uniquely
250 263 * identifies any block in the pool. By convention, the meta-objset (MOS)
251 264 * is objset 0, and the meta-dnode is object 0. This covers all blocks
252 265 * except root blocks and ZIL blocks, which are defined as follows:
253 266 *
254 267 * Root blocks (objset_phys_t) are object 0, level -1: <objset, 0, -1, 0>.
255 268 * ZIL blocks are bookmarked <objset, 0, -2, blkid == ZIL sequence number>.
256 269 * dmu_sync()ed ZIL data blocks are bookmarked <objset, object, -2, blkid>.
257 270 * dnode visit bookmarks are <objset, object id of dnode, -3, 0>.
258 271 *
259 272 * Note: this structure is called a bookmark because its original purpose
260 273 * was to remember where to resume a pool-wide traverse.
261 274 *
262 275 * Note: this structure is passed between userland and the kernel, and is
263 276 * stored on disk (by virtue of being incorporated into other on-disk
264 277 * structures, e.g. dsl_scan_phys_t).
265 278 */
266 279 typedef struct zbookmark_phys {
267 280 uint64_t zb_objset;
268 281 uint64_t zb_object;
269 282 int64_t zb_level;
270 283 uint64_t zb_blkid;
271 284 } zbookmark_phys_t;
272 285
273 286 #define SET_BOOKMARK(zb, objset, object, level, blkid) \
274 287 { \
275 288 (zb)->zb_objset = objset; \
276 289 (zb)->zb_object = object; \
277 290 (zb)->zb_level = level; \
278 291 (zb)->zb_blkid = blkid; \
279 292 }
280 293
281 294 #define ZB_DESTROYED_OBJSET (-1ULL)
282 295
283 296 #define ZB_ROOT_OBJECT (0ULL)
284 297 #define ZB_ROOT_LEVEL (-1LL)
285 298 #define ZB_ROOT_BLKID (0ULL)
286 299
287 300 #define ZB_ZIL_OBJECT (0ULL)
288 301 #define ZB_ZIL_LEVEL (-2LL)
289 302
290 303 #define ZB_DNODE_LEVEL (-3LL)
291 304 #define ZB_DNODE_BLKID (0ULL)
292 305
293 306 #define ZB_IS_ZERO(zb) \
294 307 ((zb)->zb_objset == 0 && (zb)->zb_object == 0 && \
295 308 (zb)->zb_level == 0 && (zb)->zb_blkid == 0)
296 309 #define ZB_IS_ROOT(zb) \
|
↓ open down ↓ |
39 lines elided |
↑ open up ↑ |
297 310 ((zb)->zb_object == ZB_ROOT_OBJECT && \
298 311 (zb)->zb_level == ZB_ROOT_LEVEL && \
299 312 (zb)->zb_blkid == ZB_ROOT_BLKID)
300 313
301 314 typedef struct zio_prop {
302 315 enum zio_checksum zp_checksum;
303 316 enum zio_compress zp_compress;
304 317 dmu_object_type_t zp_type;
305 318 uint8_t zp_level;
306 319 uint8_t zp_copies;
307 - boolean_t zp_dedup;
308 - boolean_t zp_dedup_verify;
320 + uint8_t zp_dedup;
321 + uint8_t zp_dedup_verify;
309 322 boolean_t zp_nopwrite;
323 + boolean_t zp_metadata;
324 + boolean_t zp_usesc;
325 + boolean_t zp_usewbc;
326 + uint64_t zp_zpl_meta_to_special;
310 327 } zio_prop_t;
311 328
312 329 typedef struct zio_cksum_report zio_cksum_report_t;
313 330
314 331 typedef void zio_cksum_finish_f(zio_cksum_report_t *rep,
315 332 const void *good_data);
316 333 typedef void zio_cksum_free_f(void *cbdata, size_t size);
317 334
318 335 struct zio_bad_cksum; /* defined in zio_checksum.h */
319 336 struct dnode_phys;
320 337 struct abd;
321 338
322 339 struct zio_cksum_report {
323 340 struct zio_cksum_report *zcr_next;
324 341 nvlist_t *zcr_ereport;
325 342 nvlist_t *zcr_detector;
326 343 void *zcr_cbdata;
327 344 size_t zcr_cbinfo; /* passed to zcr_free() */
328 345 uint64_t zcr_align;
329 346 uint64_t zcr_length;
330 347 zio_cksum_finish_f *zcr_finish;
331 348 zio_cksum_free_f *zcr_free;
332 349
333 350 /* internal use only */
334 351 struct zio_bad_cksum *zcr_ckinfo; /* information from failure */
335 352 };
336 353
337 354 typedef void zio_vsd_cksum_report_f(zio_t *zio, zio_cksum_report_t *zcr,
338 355 void *arg);
339 356
340 357 zio_vsd_cksum_report_f zio_vsd_default_cksum_report;
341 358
342 359 typedef struct zio_vsd_ops {
343 360 zio_done_func_t *vsd_free;
344 361 zio_vsd_cksum_report_f *vsd_cksum_report;
345 362 } zio_vsd_ops_t;
346 363
347 364 typedef struct zio_gang_node {
348 365 zio_gbh_phys_t *gn_gbh;
349 366 struct zio_gang_node *gn_child[SPA_GBH_NBLKPTRS];
350 367 } zio_gang_node_t;
351 368
352 369 typedef zio_t *zio_gang_issue_func_t(zio_t *zio, blkptr_t *bp,
353 370 zio_gang_node_t *gn, struct abd *data, uint64_t offset);
354 371
355 372 typedef void zio_transform_func_t(zio_t *zio, struct abd *data, uint64_t size);
356 373
357 374 typedef struct zio_transform {
358 375 struct abd *zt_orig_abd;
359 376 uint64_t zt_orig_size;
360 377 uint64_t zt_bufsize;
361 378 zio_transform_func_t *zt_transform;
362 379 struct zio_transform *zt_next;
363 380 } zio_transform_t;
364 381
365 382 typedef int zio_pipe_stage_t(zio_t *zio);
366 383
367 384 /*
368 385 * The io_reexecute flags are distinct from io_flags because the child must
369 386 * be able to propagate them to the parent. The normal io_flags are local
370 387 * to the zio, not protected by any lock, and not modifiable by children;
371 388 * the reexecute flags are protected by io_lock, modifiable by children,
372 389 * and always propagated -- even when ZIO_FLAG_DONT_PROPAGATE is set.
373 390 */
374 391 #define ZIO_REEXECUTE_NOW 0x01
375 392 #define ZIO_REEXECUTE_SUSPEND 0x02
376 393
377 394 typedef struct zio_alloc_list {
378 395 list_t zal_list;
|
↓ open down ↓ |
59 lines elided |
↑ open up ↑ |
379 396 uint64_t zal_size;
380 397 } zio_alloc_list_t;
381 398
382 399 typedef struct zio_link {
383 400 zio_t *zl_parent;
384 401 zio_t *zl_child;
385 402 list_node_t zl_parent_node;
386 403 list_node_t zl_child_node;
387 404 } zio_link_t;
388 405
406 +/*
407 + * When smart compression is enabled, this callback info structure is
408 + * passed to write zio's to monitor per-object compression performance.
409 + *
410 + * When zio_write determines that the `compression' setting for the dataset
411 + * is not `off', if `sc_ask' is not NULL, it will call the `sc_ask' callback
412 + * function, asking the upper layers whether it should really try to compress
413 + * the object in question. If the function returns B_TRUE, compression is
414 + * attempted. Once compression is done, sc_result is called to inform the
415 + * upper layers of the compression result. By comparing the zio's io_size to
416 + * io_orig_size it can monitor compression performance on the particular
417 + * object in question (if io_size == io_orig_size, then compression failed).
418 + * It is not legal to pass a NULL sc_ask but non-NULL sc_result to zio_write.
419 + */
420 +typedef struct zio_smartcomp_info {
421 + boolean_t (*sc_ask)(void *userinfo, const zio_t *);
422 + void (*sc_result)(void *userinfo, const zio_t *);
423 + void *sc_userinfo;
424 +} zio_smartcomp_info_t;
425 +
426 +#define ZIO_SHOULD_COMPRESS(zio) \
427 + ((zio)->io_smartcomp.sc_ask == NULL || \
428 + (zio)->io_smartcomp.sc_ask((zio)->io_smartcomp.sc_userinfo, (zio)))
429 +
389 430 struct zio {
390 431 /* Core information about this I/O */
391 432 zbookmark_phys_t io_bookmark;
392 433 zio_prop_t io_prop;
393 434 zio_type_t io_type;
394 435 enum zio_child io_child_type;
395 436 int io_cmd;
396 437 zio_priority_t io_priority;
397 438 uint8_t io_reexecute;
398 439 uint8_t io_state[ZIO_WAIT_TYPES];
399 440 uint64_t io_txg;
400 441 spa_t *io_spa;
401 442 blkptr_t *io_bp;
402 443 blkptr_t *io_bp_override;
403 444 blkptr_t io_bp_copy;
404 445 list_t io_parent_list;
405 446 list_t io_child_list;
406 447 zio_t *io_logical;
407 448 zio_transform_t *io_transform_stack;
449 + zio_smartcomp_info_t io_smartcomp;
408 450
409 451 /* Callback info */
410 452 zio_done_func_t *io_ready;
411 453 zio_done_func_t *io_children_ready;
412 454 zio_done_func_t *io_physdone;
413 455 zio_done_func_t *io_done;
414 456 void *io_private;
415 457 int64_t io_prev_space_delta; /* DMU private */
416 458 blkptr_t io_bp_orig;
417 459
418 460 /* Data represented by this I/O */
419 461 struct abd *io_abd;
420 462 struct abd *io_orig_abd;
421 463 uint64_t io_size;
422 464 uint64_t io_orig_size;
423 465 /* io_lsize != io_orig_size iff this is a raw write */
424 466 uint64_t io_lsize;
425 467
426 468 /* Stuff for the vdev stack */
427 469 vdev_t *io_vd;
428 470 void *io_vsd;
429 471 const zio_vsd_ops_t *io_vsd_ops;
430 472
431 473 uint64_t io_offset;
432 474 hrtime_t io_timestamp;
433 475 hrtime_t io_queued_timestamp;
434 476 hrtime_t io_target_timestamp;
435 477 avl_node_t io_queue_node;
436 478 avl_node_t io_offset_node;
437 479 avl_node_t io_alloc_node;
438 480 zio_alloc_list_t io_alloc_list;
439 481
440 482 /* Internal pipeline state */
441 483 enum zio_flag io_flags;
442 484 enum zio_stage io_stage;
443 485 enum zio_stage io_pipeline;
444 486 enum zio_flag io_orig_flags;
445 487 enum zio_stage io_orig_stage;
446 488 enum zio_stage io_orig_pipeline;
447 489 enum zio_stage io_pipeline_trace;
448 490 int io_error;
449 491 int io_child_error[ZIO_CHILD_TYPES];
450 492 uint64_t io_children[ZIO_CHILD_TYPES][ZIO_WAIT_TYPES];
451 493 uint64_t io_child_count;
452 494 uint64_t io_phys_children;
453 495 uint64_t io_parent_count;
454 496 uint64_t *io_stall;
455 497 zio_t *io_gang_leader;
456 498 zio_gang_node_t *io_gang_tree;
457 499 void *io_executor;
|
↓ open down ↓ |
40 lines elided |
↑ open up ↑ |
458 500 void *io_waiter;
459 501 kmutex_t io_lock;
460 502 kcondvar_t io_cv;
461 503
462 504 /* FMA state */
463 505 zio_cksum_report_t *io_cksum_report;
464 506 uint64_t io_ena;
465 507
466 508 /* Taskq dispatching state */
467 509 taskq_ent_t io_tqent;
510 +
511 + /* Timestamp for tracking vdev I/O latency */
512 + hrtime_t io_vd_timestamp;
513 +
514 + /* Checksum acceleration */
515 + zio_checksum_state_t zio_checksum_state;
516 + zio_cksum_t *zio_checksump;
517 + void *zio_checksum_datap;
518 + uint64_t zio_checksum_data_size;
519 + struct zio *zio_checksum_next;
520 + zio_cksum_t actual_cksum;
521 +
522 + /* Metaslab class that will be used */
523 + metaslab_class_t *io_mc;
468 524 };
469 525
470 526 extern int zio_bookmark_compare(const void *, const void *);
471 527
472 528 extern zio_t *zio_null(zio_t *pio, spa_t *spa, vdev_t *vd,
473 529 zio_done_func_t *done, void *private, enum zio_flag flags);
474 530
475 531 extern zio_t *zio_root(spa_t *spa,
476 532 zio_done_func_t *done, void *private, enum zio_flag flags);
477 533
478 534 extern zio_t *zio_read(zio_t *pio, spa_t *spa, const blkptr_t *bp,
479 535 struct abd *data, uint64_t lsize, zio_done_func_t *done, void *private,
480 536 zio_priority_t priority, enum zio_flag flags, const zbookmark_phys_t *zb);
481 537
482 538 extern zio_t *zio_write(zio_t *pio, spa_t *spa, uint64_t txg, blkptr_t *bp,
483 539 struct abd *data, uint64_t size, uint64_t psize, const zio_prop_t *zp,
484 540 zio_done_func_t *ready, zio_done_func_t *children_ready,
485 541 zio_done_func_t *physdone, zio_done_func_t *done,
486 542 void *private, zio_priority_t priority, enum zio_flag flags,
487 - const zbookmark_phys_t *zb);
543 + const zbookmark_phys_t *zb,
544 + const zio_smartcomp_info_t *smartcomp);
488 545
489 546 extern zio_t *zio_rewrite(zio_t *pio, spa_t *spa, uint64_t txg, blkptr_t *bp,
490 547 struct abd *data, uint64_t size, zio_done_func_t *done, void *private,
491 548 zio_priority_t priority, enum zio_flag flags, zbookmark_phys_t *zb);
492 549
493 550 extern void zio_write_override(zio_t *zio, blkptr_t *bp, int copies,
494 551 boolean_t nopwrite);
495 552
496 553 extern void zio_free(spa_t *spa, uint64_t txg, const blkptr_t *bp);
497 554
498 555 extern zio_t *zio_claim(zio_t *pio, spa_t *spa, uint64_t txg,
499 556 const blkptr_t *bp,
500 557 zio_done_func_t *done, void *private, enum zio_flag flags);
501 558
502 559 extern zio_t *zio_ioctl(zio_t *pio, spa_t *spa, vdev_t *vd, int cmd,
503 560 zio_done_func_t *done, void *private, enum zio_flag flags);
504 561
562 +extern zio_t *zio_trim(spa_t *spa, vdev_t *vd, struct range_tree *tree,
563 + zio_done_func_t *done, void *private, enum zio_flag flags,
564 + int dkiocfree_flags, metaslab_t *msp);
565 +
505 566 extern zio_t *zio_read_phys(zio_t *pio, vdev_t *vd, uint64_t offset,
506 567 uint64_t size, struct abd *data, int checksum,
507 568 zio_done_func_t *done, void *private, zio_priority_t priority,
508 569 enum zio_flag flags, boolean_t labels);
509 570
510 571 extern zio_t *zio_write_phys(zio_t *pio, vdev_t *vd, uint64_t offset,
511 572 uint64_t size, struct abd *data, int checksum,
512 573 zio_done_func_t *done, void *private, zio_priority_t priority,
513 574 enum zio_flag flags, boolean_t labels);
514 575
576 +extern zio_t *zio_wbc(zio_type_t type, vdev_t *vd, abd_t *data,
577 + uint64_t size, uint64_t offset);
578 +
515 579 extern zio_t *zio_free_sync(zio_t *pio, spa_t *spa, uint64_t txg,
516 580 const blkptr_t *bp, enum zio_flag flags);
517 581
518 582 extern int zio_alloc_zil(spa_t *spa, uint64_t txg, blkptr_t *new_bp,
519 583 blkptr_t *old_bp, uint64_t size, boolean_t *slog);
520 584 extern void zio_free_zil(spa_t *spa, uint64_t txg, blkptr_t *bp);
521 585 extern void zio_flush(zio_t *zio, vdev_t *vd);
522 586 extern void zio_shrink(zio_t *zio, uint64_t size);
523 587
524 588 extern int zio_wait(zio_t *zio);
525 589 extern void zio_nowait(zio_t *zio);
526 590 extern void zio_execute(zio_t *zio);
527 591 extern void zio_interrupt(zio_t *zio);
528 592 extern void zio_delay_init(zio_t *zio);
529 593 extern void zio_delay_interrupt(zio_t *zio);
530 594
531 595 extern zio_t *zio_walk_parents(zio_t *cio, zio_link_t **);
532 596 extern zio_t *zio_walk_children(zio_t *pio, zio_link_t **);
533 597 extern zio_t *zio_unique_parent(zio_t *cio);
534 598 extern void zio_add_child(zio_t *pio, zio_t *cio);
535 599
536 600 extern void *zio_buf_alloc(size_t size);
537 601 extern void zio_buf_free(void *buf, size_t size);
538 602 extern void *zio_data_buf_alloc(size_t size);
539 603 extern void zio_data_buf_free(void *buf, size_t size);
540 604
541 605 extern void zio_push_transform(zio_t *zio, struct abd *abd, uint64_t size,
542 606 uint64_t bufsize, zio_transform_func_t *transform);
543 607 extern void zio_pop_transforms(zio_t *zio);
544 608
545 609 extern void zio_resubmit_stage_async(void *);
546 610
547 611 extern zio_t *zio_vdev_child_io(zio_t *zio, blkptr_t *bp, vdev_t *vd,
548 612 uint64_t offset, struct abd *data, uint64_t size, int type,
549 613 zio_priority_t priority, enum zio_flag flags,
550 614 zio_done_func_t *done, void *private);
551 615
552 616 extern zio_t *zio_vdev_delegated_io(vdev_t *vd, uint64_t offset,
553 617 struct abd *data, uint64_t size, int type, zio_priority_t priority,
554 618 enum zio_flag flags, zio_done_func_t *done, void *private);
555 619
556 620 extern void zio_vdev_io_bypass(zio_t *zio);
557 621 extern void zio_vdev_io_reissue(zio_t *zio);
558 622 extern void zio_vdev_io_redone(zio_t *zio);
559 623
560 624 extern void zio_checksum_verified(zio_t *zio);
561 625 extern int zio_worst_error(int e1, int e2);
562 626
563 627 extern enum zio_checksum zio_checksum_select(enum zio_checksum child,
564 628 enum zio_checksum parent);
565 629 extern enum zio_checksum zio_checksum_dedup_select(spa_t *spa,
566 630 enum zio_checksum child, enum zio_checksum parent);
567 631 extern enum zio_compress zio_compress_select(spa_t *spa,
568 632 enum zio_compress child, enum zio_compress parent);
569 633
570 634 extern void zio_suspend(spa_t *spa, zio_t *zio);
571 635 extern int zio_resume(spa_t *spa);
572 636 extern void zio_resume_wait(spa_t *spa);
573 637
574 638 /*
575 639 * Initial setup and teardown.
576 640 */
577 641 extern void zio_init(void);
578 642 extern void zio_fini(void);
579 643
580 644 /*
581 645 * Fault injection
582 646 */
583 647 struct zinject_record;
584 648 extern uint32_t zio_injection_enabled;
585 649 extern int zio_inject_fault(char *name, int flags, int *id,
586 650 struct zinject_record *record);
587 651 extern int zio_inject_list_next(int *id, char *name, size_t buflen,
588 652 struct zinject_record *record);
589 653 extern int zio_clear_fault(int id);
590 654 extern void zio_handle_panic_injection(spa_t *spa, char *tag, uint64_t type);
591 655 extern int zio_handle_fault_injection(zio_t *zio, int error);
592 656 extern int zio_handle_device_injection(vdev_t *vd, zio_t *zio, int error);
593 657 extern int zio_handle_label_injection(zio_t *zio, int error);
594 658 extern void zio_handle_ignored_writes(zio_t *zio);
595 659 extern hrtime_t zio_handle_io_delay(zio_t *zio);
596 660
|
↓ open down ↓ |
72 lines elided |
↑ open up ↑ |
597 661 /*
598 662 * Checksum ereport functions
599 663 */
600 664 extern void zfs_ereport_start_checksum(spa_t *spa, vdev_t *vd, struct zio *zio,
601 665 uint64_t offset, uint64_t length, void *arg, struct zio_bad_cksum *info);
602 666 extern void zfs_ereport_finish_checksum(zio_cksum_report_t *report,
603 667 const void *good_data, const void *bad_data, boolean_t drop_if_identical);
604 668
605 669 extern void zfs_ereport_send_interim_checksum(zio_cksum_report_t *report);
606 670 extern void zfs_ereport_free_checksum(zio_cksum_report_t *report);
607 -
608 671 /* If we have the good data in hand, this function can be used */
609 672 extern void zfs_ereport_post_checksum(spa_t *spa, vdev_t *vd,
610 673 struct zio *zio, uint64_t offset, uint64_t length,
611 674 const void *good_data, const void *bad_data, struct zio_bad_cksum *info);
612 675
613 676 /* Called from spa_sync(), but primarily an injection handler */
614 677 extern void spa_handle_ignored_writes(spa_t *spa);
615 678
616 679 /* zbookmark_phys functions */
617 680 boolean_t zbookmark_subtree_completed(const struct dnode_phys *dnp,
618 681 const zbookmark_phys_t *subtree_root, const zbookmark_phys_t *last_block);
619 682 int zbookmark_compare(uint16_t dbss1, uint8_t ibs1, uint16_t dbss2,
620 683 uint8_t ibs2, const zbookmark_phys_t *zb1, const zbookmark_phys_t *zb2);
621 684
685 +/* best effort dedup */
686 +void zio_best_effort_dedup(zio_t *zio);
687 +
622 688 #ifdef __cplusplus
623 689 }
624 690 #endif
625 691
626 692 #endif /* _ZIO_H */
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX