2 * Copyright (c) 2010 Alexander Motin <mav@FreeBSD.org>
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
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.
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
27 #include <sys/cdefs.h>
28 __FBSDID("$FreeBSD$");
30 #include <sys/param.h>
32 #include <sys/endian.h>
33 #include <sys/kernel.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"
43 static MALLOC_DEFINE(M_TR_CONCAT, "tr_concat_data", "GEOM_RAID CONCAT data");
45 struct g_raid_tr_concat_object {
46 struct g_raid_tr_object trso_base;
51 static g_raid_tr_taste_t g_raid_tr_taste_concat;
52 static g_raid_tr_event_t g_raid_tr_event_concat;
53 static g_raid_tr_start_t g_raid_tr_start_concat;
54 static g_raid_tr_stop_t g_raid_tr_stop_concat;
55 static g_raid_tr_iostart_t g_raid_tr_iostart_concat;
56 static g_raid_tr_iodone_t g_raid_tr_iodone_concat;
57 static g_raid_tr_kerneldump_t g_raid_tr_kerneldump_concat;
58 static g_raid_tr_free_t g_raid_tr_free_concat;
60 static kobj_method_t g_raid_tr_concat_methods[] = {
61 KOBJMETHOD(g_raid_tr_taste, g_raid_tr_taste_concat),
62 KOBJMETHOD(g_raid_tr_event, g_raid_tr_event_concat),
63 KOBJMETHOD(g_raid_tr_start, g_raid_tr_start_concat),
64 KOBJMETHOD(g_raid_tr_stop, g_raid_tr_stop_concat),
65 KOBJMETHOD(g_raid_tr_iostart, g_raid_tr_iostart_concat),
66 KOBJMETHOD(g_raid_tr_iodone, g_raid_tr_iodone_concat),
67 KOBJMETHOD(g_raid_tr_kerneldump, g_raid_tr_kerneldump_concat),
68 KOBJMETHOD(g_raid_tr_free, g_raid_tr_free_concat),
72 static struct g_raid_tr_class g_raid_tr_concat_class = {
74 g_raid_tr_concat_methods,
75 sizeof(struct g_raid_tr_concat_object),
78 .trc_accept_unmapped = 1
82 g_raid_tr_taste_concat(struct g_raid_tr_object *tr, struct g_raid_volume *volume)
84 struct g_raid_tr_concat_object *trs;
86 trs = (struct g_raid_tr_concat_object *)tr;
87 if (tr->tro_volume->v_raid_level != G_RAID_VOLUME_RL_SINGLE &&
88 tr->tro_volume->v_raid_level != G_RAID_VOLUME_RL_CONCAT &&
89 !(tr->tro_volume->v_disks_count == 1 &&
90 tr->tro_volume->v_raid_level != G_RAID_VOLUME_RL_UNKNOWN))
91 return (G_RAID_TR_TASTE_FAIL);
92 trs->trso_starting = 1;
93 return (G_RAID_TR_TASTE_SUCCEED);
97 g_raid_tr_update_state_concat(struct g_raid_volume *vol)
99 struct g_raid_tr_concat_object *trs;
100 struct g_raid_softc *sc;
106 trs = (struct g_raid_tr_concat_object *)vol->v_tr;
107 if (trs->trso_stopped)
108 s = G_RAID_VOLUME_S_STOPPED;
109 else if (trs->trso_starting)
110 s = G_RAID_VOLUME_S_STARTING;
112 n = g_raid_nsubdisks(vol, G_RAID_SUBDISK_S_ACTIVE);
113 f = g_raid_nsubdisks(vol, G_RAID_SUBDISK_S_FAILED);
114 if (n + f == vol->v_disks_count) {
116 s = G_RAID_VOLUME_S_OPTIMAL;
118 s = G_RAID_VOLUME_S_SUBOPTIMAL;
120 s = G_RAID_VOLUME_S_BROKEN;
122 if (s != vol->v_state) {
125 * Some metadata modules may not know CONCAT volume
126 * mediasize until all disks connected. Recalculate.
128 if (vol->v_raid_level == G_RAID_VOLUME_RL_CONCAT &&
129 G_RAID_VOLUME_S_ALIVE(s) &&
130 !G_RAID_VOLUME_S_ALIVE(vol->v_state)) {
132 for (i = 0; i < vol->v_disks_count; i++) {
133 if (vol->v_subdisks[i].sd_state !=
134 G_RAID_SUBDISK_S_NONE)
135 size += vol->v_subdisks[i].sd_size;
137 vol->v_mediasize = size;
140 g_raid_event_send(vol, G_RAID_VOLUME_S_ALIVE(s) ?
141 G_RAID_VOLUME_E_UP : G_RAID_VOLUME_E_DOWN,
142 G_RAID_EVENT_VOLUME);
143 g_raid_change_volume_state(vol, s);
144 if (!trs->trso_starting && !trs->trso_stopped)
145 g_raid_write_metadata(sc, vol, NULL, NULL);
151 g_raid_tr_event_concat(struct g_raid_tr_object *tr,
152 struct g_raid_subdisk *sd, u_int event)
154 struct g_raid_tr_concat_object *trs;
155 struct g_raid_softc *sc;
156 struct g_raid_volume *vol;
159 trs = (struct g_raid_tr_concat_object *)tr;
160 vol = tr->tro_volume;
163 state = sd->sd_state;
164 if (state != G_RAID_SUBDISK_S_NONE &&
165 state != G_RAID_SUBDISK_S_FAILED &&
166 state != G_RAID_SUBDISK_S_ACTIVE) {
168 "Promote subdisk %s:%d from %s to ACTIVE.",
169 vol->v_name, sd->sd_pos,
170 g_raid_subdisk_state2str(sd->sd_state));
171 g_raid_change_subdisk_state(sd, G_RAID_SUBDISK_S_ACTIVE);
173 if (state != sd->sd_state &&
174 !trs->trso_starting && !trs->trso_stopped)
175 g_raid_write_metadata(sc, vol, sd, NULL);
176 g_raid_tr_update_state_concat(vol);
181 g_raid_tr_start_concat(struct g_raid_tr_object *tr)
183 struct g_raid_tr_concat_object *trs;
184 struct g_raid_volume *vol;
186 trs = (struct g_raid_tr_concat_object *)tr;
187 vol = tr->tro_volume;
188 trs->trso_starting = 0;
189 g_raid_tr_update_state_concat(vol);
194 g_raid_tr_stop_concat(struct g_raid_tr_object *tr)
196 struct g_raid_tr_concat_object *trs;
197 struct g_raid_volume *vol;
199 trs = (struct g_raid_tr_concat_object *)tr;
200 vol = tr->tro_volume;
201 trs->trso_starting = 0;
202 trs->trso_stopped = 1;
203 g_raid_tr_update_state_concat(vol);
208 g_raid_tr_iostart_concat(struct g_raid_tr_object *tr, struct bio *bp)
210 struct g_raid_volume *vol;
211 struct g_raid_subdisk *sd;
212 struct bio_queue_head queue;
215 off_t offset, length, remain;
218 vol = tr->tro_volume;
219 if (vol->v_state != G_RAID_VOLUME_S_OPTIMAL &&
220 vol->v_state != G_RAID_VOLUME_S_SUBOPTIMAL) {
221 g_raid_iodone(bp, EIO);
224 if (bp->bio_cmd == BIO_FLUSH) {
225 g_raid_tr_flush_common(tr, bp);
229 offset = bp->bio_offset;
230 remain = bp->bio_length;
231 if ((bp->bio_flags & BIO_UNMAPPED) != 0)
236 while (no < vol->v_disks_count &&
237 offset >= vol->v_subdisks[no].sd_size) {
238 offset -= vol->v_subdisks[no].sd_size;
241 KASSERT(no < vol->v_disks_count,
242 ("Request starts after volume end (%ju)", bp->bio_offset));
245 sd = &vol->v_subdisks[no];
246 length = MIN(sd->sd_size - offset, remain);
247 cbp = g_clone_bio(bp);
250 cbp->bio_offset = offset;
251 cbp->bio_length = length;
252 if ((bp->bio_flags & BIO_UNMAPPED) != 0 &&
253 bp->bio_cmd != BIO_DELETE) {
254 cbp->bio_ma_offset += (uintptr_t)addr;
255 cbp->bio_ma += cbp->bio_ma_offset / PAGE_SIZE;
256 cbp->bio_ma_offset %= PAGE_SIZE;
257 cbp->bio_ma_n = round_page(cbp->bio_ma_offset +
258 cbp->bio_length) / PAGE_SIZE;
260 cbp->bio_data = addr;
261 cbp->bio_caller1 = sd;
262 bioq_insert_tail(&queue, cbp);
264 if (bp->bio_cmd != BIO_DELETE)
268 KASSERT(no < vol->v_disks_count || remain == 0,
269 ("Request ends after volume end (%ju, %ju)",
270 bp->bio_offset, bp->bio_length));
271 } while (remain > 0);
272 while ((cbp = bioq_takefirst(&queue)) != NULL) {
273 sd = cbp->bio_caller1;
274 cbp->bio_caller1 = NULL;
275 g_raid_subdisk_iostart(sd, cbp);
279 while ((cbp = bioq_takefirst(&queue)) != NULL)
281 if (bp->bio_error == 0)
282 bp->bio_error = ENOMEM;
283 g_raid_iodone(bp, bp->bio_error);
287 g_raid_tr_kerneldump_concat(struct g_raid_tr_object *tr,
288 void *virtual, vm_offset_t physical, off_t boffset, size_t blength)
290 struct g_raid_volume *vol;
291 struct g_raid_subdisk *sd;
293 off_t offset, length, remain;
296 vol = tr->tro_volume;
297 if (vol->v_state != G_RAID_VOLUME_S_OPTIMAL)
304 while (no < vol->v_disks_count &&
305 offset >= vol->v_subdisks[no].sd_size) {
306 offset -= vol->v_subdisks[no].sd_size;
309 KASSERT(no < vol->v_disks_count,
310 ("Request starts after volume end (%ju)", boffset));
312 sd = &vol->v_subdisks[no];
313 length = MIN(sd->sd_size - offset, remain);
314 error = g_raid_subdisk_kerneldump(&vol->v_subdisks[no],
315 addr, 0, offset, length);
322 KASSERT(no < vol->v_disks_count || remain == 0,
323 ("Request ends after volume end (%ju, %zu)",
325 } while (remain > 0);
330 g_raid_tr_iodone_concat(struct g_raid_tr_object *tr,
331 struct g_raid_subdisk *sd,struct bio *bp)
335 pbp = bp->bio_parent;
336 if (pbp->bio_error == 0)
337 pbp->bio_error = bp->bio_error;
340 if (pbp->bio_children == pbp->bio_inbed) {
341 pbp->bio_completed = pbp->bio_length;
342 g_raid_iodone(pbp, bp->bio_error);
347 g_raid_tr_free_concat(struct g_raid_tr_object *tr)
353 G_RAID_TR_DECLARE(concat, "CONCAT");