tr_raid0.c revision 240465
1/*- 2 * Copyright (c) 2010 Alexander Motin <mav@FreeBSD.org> 3 * All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 14 * THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND 15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 17 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE 18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 24 * SUCH DAMAGE. 25 */ 26 27#include <sys/cdefs.h> 28__FBSDID("$FreeBSD: head/sys/geom/raid/tr_raid0.c 240465 2012-09-13 13:27:09Z mav $"); 29 30#include <sys/param.h> 31#include <sys/bio.h> 32#include <sys/endian.h> 33#include <sys/kernel.h> 34#include <sys/kobj.h> 35#include <sys/lock.h> 36#include <sys/malloc.h> 37#include <sys/mutex.h> 38#include <sys/systm.h> 39#include <geom/geom.h> 40#include "geom/raid/g_raid.h" 41#include "g_raid_tr_if.h" 42 43static MALLOC_DEFINE(M_TR_RAID0, "tr_raid0_data", "GEOM_RAID RAID0 data"); 44 45struct g_raid_tr_raid0_object { 46 struct g_raid_tr_object trso_base; 47 int trso_starting; 48 int trso_stopped; 49}; 50 51static g_raid_tr_taste_t g_raid_tr_taste_raid0; 52static g_raid_tr_event_t g_raid_tr_event_raid0; 53static g_raid_tr_start_t g_raid_tr_start_raid0; 54static g_raid_tr_stop_t g_raid_tr_stop_raid0; 55static g_raid_tr_iostart_t g_raid_tr_iostart_raid0; 56static g_raid_tr_iodone_t g_raid_tr_iodone_raid0; 57static g_raid_tr_kerneldump_t g_raid_tr_kerneldump_raid0; 58static g_raid_tr_free_t g_raid_tr_free_raid0; 59 60static kobj_method_t g_raid_tr_raid0_methods[] = { 61 KOBJMETHOD(g_raid_tr_taste, g_raid_tr_taste_raid0), 62 KOBJMETHOD(g_raid_tr_event, g_raid_tr_event_raid0), 63 KOBJMETHOD(g_raid_tr_start, g_raid_tr_start_raid0), 64 KOBJMETHOD(g_raid_tr_stop, g_raid_tr_stop_raid0), 65 KOBJMETHOD(g_raid_tr_iostart, g_raid_tr_iostart_raid0), 66 KOBJMETHOD(g_raid_tr_iodone, g_raid_tr_iodone_raid0), 67 KOBJMETHOD(g_raid_tr_kerneldump, g_raid_tr_kerneldump_raid0), 68 KOBJMETHOD(g_raid_tr_free, g_raid_tr_free_raid0), 69 { 0, 0 } 70}; 71 72static struct g_raid_tr_class g_raid_tr_raid0_class = { 73 "RAID0", 74 g_raid_tr_raid0_methods, 75 sizeof(struct g_raid_tr_raid0_object), 76 .trc_enable = 1, 77 .trc_priority = 100 78}; 79 80static int 81g_raid_tr_taste_raid0(struct g_raid_tr_object *tr, struct g_raid_volume *volume) 82{ 83 struct g_raid_tr_raid0_object *trs; 84 85 trs = (struct g_raid_tr_raid0_object *)tr; 86 if (tr->tro_volume->v_raid_level != G_RAID_VOLUME_RL_RAID0 || 87 tr->tro_volume->v_raid_level_qualifier != G_RAID_VOLUME_RLQ_NONE) 88 return (G_RAID_TR_TASTE_FAIL); 89 trs->trso_starting = 1; 90 return (G_RAID_TR_TASTE_SUCCEED); 91} 92 93static int 94g_raid_tr_update_state_raid0(struct g_raid_volume *vol) 95{ 96 struct g_raid_tr_raid0_object *trs; 97 struct g_raid_softc *sc; 98 u_int s; 99 int n, f; 100 101 sc = vol->v_softc; 102 trs = (struct g_raid_tr_raid0_object *)vol->v_tr; 103 if (trs->trso_stopped) 104 s = G_RAID_VOLUME_S_STOPPED; 105 else if (trs->trso_starting) 106 s = G_RAID_VOLUME_S_STARTING; 107 else { 108 n = g_raid_nsubdisks(vol, G_RAID_SUBDISK_S_ACTIVE); 109 f = g_raid_nsubdisks(vol, G_RAID_SUBDISK_S_FAILED); 110 if (n + f == vol->v_disks_count) { 111 if (f == 0) 112 s = G_RAID_VOLUME_S_OPTIMAL; 113 else 114 s = G_RAID_VOLUME_S_SUBOPTIMAL; 115 } else 116 s = G_RAID_VOLUME_S_BROKEN; 117 } 118 if (s != vol->v_state) { 119 g_raid_event_send(vol, G_RAID_VOLUME_S_ALIVE(s) ? 120 G_RAID_VOLUME_E_UP : G_RAID_VOLUME_E_DOWN, 121 G_RAID_EVENT_VOLUME); 122 g_raid_change_volume_state(vol, s); 123 if (!trs->trso_starting && !trs->trso_stopped) 124 g_raid_write_metadata(sc, vol, NULL, NULL); 125 } 126 return (0); 127} 128 129static int 130g_raid_tr_event_raid0(struct g_raid_tr_object *tr, 131 struct g_raid_subdisk *sd, u_int event) 132{ 133 struct g_raid_tr_raid0_object *trs; 134 struct g_raid_softc *sc; 135 struct g_raid_volume *vol; 136 int state; 137 138 trs = (struct g_raid_tr_raid0_object *)tr; 139 vol = tr->tro_volume; 140 sc = vol->v_softc; 141 142 state = sd->sd_state; 143 if (state != G_RAID_SUBDISK_S_NONE && 144 state != G_RAID_SUBDISK_S_FAILED && 145 state != G_RAID_SUBDISK_S_ACTIVE) { 146 G_RAID_DEBUG1(1, sc, 147 "Promote subdisk %s:%d from %s to ACTIVE.", 148 vol->v_name, sd->sd_pos, 149 g_raid_subdisk_state2str(sd->sd_state)); 150 g_raid_change_subdisk_state(sd, G_RAID_SUBDISK_S_ACTIVE); 151 } 152 if (state != sd->sd_state && 153 !trs->trso_starting && !trs->trso_stopped) 154 g_raid_write_metadata(sc, vol, sd, NULL); 155 g_raid_tr_update_state_raid0(vol); 156 return (0); 157} 158 159static int 160g_raid_tr_start_raid0(struct g_raid_tr_object *tr) 161{ 162 struct g_raid_tr_raid0_object *trs; 163 struct g_raid_volume *vol; 164 165 trs = (struct g_raid_tr_raid0_object *)tr; 166 vol = tr->tro_volume; 167 trs->trso_starting = 0; 168 g_raid_tr_update_state_raid0(vol); 169 return (0); 170} 171 172static int 173g_raid_tr_stop_raid0(struct g_raid_tr_object *tr) 174{ 175 struct g_raid_tr_raid0_object *trs; 176 struct g_raid_volume *vol; 177 178 trs = (struct g_raid_tr_raid0_object *)tr; 179 vol = tr->tro_volume; 180 trs->trso_starting = 0; 181 trs->trso_stopped = 1; 182 g_raid_tr_update_state_raid0(vol); 183 return (0); 184} 185 186static void 187g_raid_tr_iostart_raid0(struct g_raid_tr_object *tr, struct bio *bp) 188{ 189 struct g_raid_volume *vol; 190 struct g_raid_subdisk *sd; 191 struct bio_queue_head queue; 192 struct bio *cbp; 193 char *addr; 194 off_t offset, start, length, nstripe, remain; 195 u_int no, strip_size; 196 197 vol = tr->tro_volume; 198 if (vol->v_state != G_RAID_VOLUME_S_OPTIMAL && 199 vol->v_state != G_RAID_VOLUME_S_SUBOPTIMAL) { 200 g_raid_iodone(bp, EIO); 201 return; 202 } 203 if (bp->bio_cmd == BIO_FLUSH) { 204 g_raid_tr_flush_common(tr, bp); 205 return; 206 } 207 addr = bp->bio_data; 208 strip_size = vol->v_strip_size; 209 210 /* Stripe number. */ 211 nstripe = bp->bio_offset / strip_size; 212 /* Start position in stripe. */ 213 start = bp->bio_offset % strip_size; 214 /* Disk number. */ 215 no = nstripe % vol->v_disks_count; 216 /* Stripe start position in disk. */ 217 offset = (nstripe / vol->v_disks_count) * strip_size; 218 /* Length of data to operate. */ 219 remain = bp->bio_length; 220 221 bioq_init(&queue); 222 do { 223 length = MIN(strip_size - start, remain); 224 cbp = g_clone_bio(bp); 225 if (cbp == NULL) 226 goto failure; 227 cbp->bio_offset = offset + start; 228 cbp->bio_data = addr; 229 cbp->bio_length = length; 230 cbp->bio_caller1 = &vol->v_subdisks[no]; 231 bioq_insert_tail(&queue, cbp); 232 if (++no >= vol->v_disks_count) { 233 no = 0; 234 offset += strip_size; 235 } 236 remain -= length; 237 addr += length; 238 start = 0; 239 } while (remain > 0); 240 for (cbp = bioq_first(&queue); cbp != NULL; 241 cbp = bioq_first(&queue)) { 242 bioq_remove(&queue, cbp); 243 sd = cbp->bio_caller1; 244 cbp->bio_caller1 = NULL; 245 g_raid_subdisk_iostart(sd, cbp); 246 } 247 return; 248failure: 249 for (cbp = bioq_first(&queue); cbp != NULL; 250 cbp = bioq_first(&queue)) { 251 bioq_remove(&queue, cbp); 252 g_destroy_bio(cbp); 253 } 254 if (bp->bio_error == 0) 255 bp->bio_error = ENOMEM; 256 g_raid_iodone(bp, bp->bio_error); 257} 258 259static int 260g_raid_tr_kerneldump_raid0(struct g_raid_tr_object *tr, 261 void *virtual, vm_offset_t physical, off_t boffset, size_t blength) 262{ 263 struct g_raid_volume *vol; 264 char *addr; 265 off_t offset, start, length, nstripe, remain; 266 u_int no, strip_size; 267 int error; 268 269 vol = tr->tro_volume; 270 if (vol->v_state != G_RAID_VOLUME_S_OPTIMAL) 271 return (ENXIO); 272 addr = virtual; 273 strip_size = vol->v_strip_size; 274 275 /* Stripe number. */ 276 nstripe = boffset / strip_size; 277 /* Start position in stripe. */ 278 start = boffset % strip_size; 279 /* Disk number. */ 280 no = nstripe % vol->v_disks_count; 281 /* Stripe tart position in disk. */ 282 offset = (nstripe / vol->v_disks_count) * strip_size; 283 /* Length of data to operate. */ 284 remain = blength; 285 286 do { 287 length = MIN(strip_size - start, remain); 288 error = g_raid_subdisk_kerneldump(&vol->v_subdisks[no], 289 addr, 0, offset + start, length); 290 if (error != 0) 291 return (error); 292 if (++no >= vol->v_disks_count) { 293 no = 0; 294 offset += strip_size; 295 } 296 remain -= length; 297 addr += length; 298 start = 0; 299 } while (remain > 0); 300 return (0); 301} 302 303static void 304g_raid_tr_iodone_raid0(struct g_raid_tr_object *tr, 305 struct g_raid_subdisk *sd,struct bio *bp) 306{ 307 struct bio *pbp; 308 309 pbp = bp->bio_parent; 310 if (pbp->bio_error == 0) 311 pbp->bio_error = bp->bio_error; 312 g_destroy_bio(bp); 313 pbp->bio_inbed++; 314 if (pbp->bio_children == pbp->bio_inbed) { 315 pbp->bio_completed = pbp->bio_length; 316 g_raid_iodone(pbp, bp->bio_error); 317 } 318} 319 320static int 321g_raid_tr_free_raid0(struct g_raid_tr_object *tr) 322{ 323 324 return (0); 325} 326 327G_RAID_TR_DECLARE(raid0, "RAID0"); 328