2 * Block Translation Table
3 * Copyright (c) 2014-2015, Intel Corporation.
5 * This program is free software; you can redistribute it and/or modify it
6 * under the terms and conditions of the GNU General Public License,
7 * version 2, as published by the Free Software Foundation.
9 * This program is distributed in the hope it will be useful, but WITHOUT
10 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
11 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
14 #include <linux/highmem.h>
15 #include <linux/debugfs.h>
16 #include <linux/blkdev.h>
17 #include <linux/module.h>
18 #include <linux/device.h>
19 #include <linux/mutex.h>
20 #include <linux/hdreg.h>
21 #include <linux/genhd.h>
22 #include <linux/sizes.h>
23 #include <linux/ndctl.h>
29 enum log_ent_request {
36 static int arena_read_bytes(struct arena_info *arena, resource_size_t offset,
39 struct nd_btt *nd_btt = arena->nd_btt;
40 struct nd_namespace_common *ndns = nd_btt->ndns;
42 /* arena offsets are 4K from the base of the device */
44 return nvdimm_read_bytes(ndns, offset, buf, n);
47 static int arena_write_bytes(struct arena_info *arena, resource_size_t offset,
50 struct nd_btt *nd_btt = arena->nd_btt;
51 struct nd_namespace_common *ndns = nd_btt->ndns;
53 /* arena offsets are 4K from the base of the device */
55 return nvdimm_write_bytes(ndns, offset, buf, n);
58 static int btt_info_write(struct arena_info *arena, struct btt_sb *super)
62 ret = arena_write_bytes(arena, arena->info2off, super,
63 sizeof(struct btt_sb));
67 return arena_write_bytes(arena, arena->infooff, super,
68 sizeof(struct btt_sb));
71 static int btt_info_read(struct arena_info *arena, struct btt_sb *super)
74 return arena_read_bytes(arena, arena->infooff, super,
75 sizeof(struct btt_sb));
79 * 'raw' version of btt_map write
81 * mapping is in little-endian
82 * mapping contains 'E' and 'Z' flags as desired
84 static int __btt_map_write(struct arena_info *arena, u32 lba, __le32 mapping)
86 u64 ns_off = arena->mapoff + (lba * MAP_ENT_SIZE);
88 WARN_ON(lba >= arena->external_nlba);
89 return arena_write_bytes(arena, ns_off, &mapping, MAP_ENT_SIZE);
92 static int btt_map_write(struct arena_info *arena, u32 lba, u32 mapping,
93 u32 z_flag, u32 e_flag)
99 * This 'mapping' is supposed to be just the LBA mapping, without
100 * any flags set, so strip the flag bits.
102 mapping &= MAP_LBA_MASK;
104 ze = (z_flag << 1) + e_flag;
108 * We want to set neither of the Z or E flags, and
109 * in the actual layout, this means setting the bit
110 * positions of both to '1' to indicate a 'normal'
113 mapping |= MAP_ENT_NORMAL;
116 mapping |= (1 << MAP_ERR_SHIFT);
119 mapping |= (1 << MAP_TRIM_SHIFT);
123 * The case where Z and E are both sent in as '1' could be
124 * construed as a valid 'normal' case, but we decide not to,
127 WARN_ONCE(1, "Invalid use of Z and E flags\n");
131 mapping_le = cpu_to_le32(mapping);
132 return __btt_map_write(arena, lba, mapping_le);
135 static int btt_map_read(struct arena_info *arena, u32 lba, u32 *mapping,
136 int *trim, int *error)
140 u32 raw_mapping, postmap, ze, z_flag, e_flag;
141 u64 ns_off = arena->mapoff + (lba * MAP_ENT_SIZE);
143 WARN_ON(lba >= arena->external_nlba);
145 ret = arena_read_bytes(arena, ns_off, &in, MAP_ENT_SIZE);
149 raw_mapping = le32_to_cpu(in);
151 z_flag = (raw_mapping & MAP_TRIM_MASK) >> MAP_TRIM_SHIFT;
152 e_flag = (raw_mapping & MAP_ERR_MASK) >> MAP_ERR_SHIFT;
153 ze = (z_flag << 1) + e_flag;
154 postmap = raw_mapping & MAP_LBA_MASK;
156 /* Reuse the {z,e}_flag variables for *trim and *error */
162 /* Initial state. Return postmap = premap */
188 static int btt_log_read_pair(struct arena_info *arena, u32 lane,
189 struct log_entry *ent)
192 return arena_read_bytes(arena,
193 arena->logoff + (2 * lane * LOG_ENT_SIZE), ent,
197 static struct dentry *debugfs_root;
199 static void arena_debugfs_init(struct arena_info *a, struct dentry *parent,
205 /* If for some reason, parent bttN was not created, exit */
209 snprintf(dirname, 32, "arena%d", idx);
210 d = debugfs_create_dir(dirname, parent);
211 if (IS_ERR_OR_NULL(d))
215 debugfs_create_x64("size", S_IRUGO, d, &a->size);
216 debugfs_create_x64("external_lba_start", S_IRUGO, d,
217 &a->external_lba_start);
218 debugfs_create_x32("internal_nlba", S_IRUGO, d, &a->internal_nlba);
219 debugfs_create_u32("internal_lbasize", S_IRUGO, d,
220 &a->internal_lbasize);
221 debugfs_create_x32("external_nlba", S_IRUGO, d, &a->external_nlba);
222 debugfs_create_u32("external_lbasize", S_IRUGO, d,
223 &a->external_lbasize);
224 debugfs_create_u32("nfree", S_IRUGO, d, &a->nfree);
225 debugfs_create_u16("version_major", S_IRUGO, d, &a->version_major);
226 debugfs_create_u16("version_minor", S_IRUGO, d, &a->version_minor);
227 debugfs_create_x64("nextoff", S_IRUGO, d, &a->nextoff);
228 debugfs_create_x64("infooff", S_IRUGO, d, &a->infooff);
229 debugfs_create_x64("dataoff", S_IRUGO, d, &a->dataoff);
230 debugfs_create_x64("mapoff", S_IRUGO, d, &a->mapoff);
231 debugfs_create_x64("logoff", S_IRUGO, d, &a->logoff);
232 debugfs_create_x64("info2off", S_IRUGO, d, &a->info2off);
233 debugfs_create_x32("flags", S_IRUGO, d, &a->flags);
236 static void btt_debugfs_init(struct btt *btt)
239 struct arena_info *arena;
241 btt->debugfs_dir = debugfs_create_dir(dev_name(&btt->nd_btt->dev),
243 if (IS_ERR_OR_NULL(btt->debugfs_dir))
246 list_for_each_entry(arena, &btt->arena_list, list) {
247 arena_debugfs_init(arena, btt->debugfs_dir, i);
253 * This function accepts two log entries, and uses the
254 * sequence number to find the 'older' entry.
255 * It also updates the sequence number in this old entry to
256 * make it the 'new' one if the mark_flag is set.
257 * Finally, it returns which of the entries was the older one.
259 * TODO The logic feels a bit kludge-y. make it better..
261 static int btt_log_get_old(struct log_entry *ent)
266 * the first ever time this is seen, the entry goes into [0]
267 * the next time, the following logic works out to put this
268 * (next) entry into [1]
270 if (ent[0].seq == 0) {
271 ent[0].seq = cpu_to_le32(1);
275 if (ent[0].seq == ent[1].seq)
277 if (le32_to_cpu(ent[0].seq) + le32_to_cpu(ent[1].seq) > 5)
280 if (le32_to_cpu(ent[0].seq) < le32_to_cpu(ent[1].seq)) {
281 if (le32_to_cpu(ent[1].seq) - le32_to_cpu(ent[0].seq) == 1)
286 if (le32_to_cpu(ent[0].seq) - le32_to_cpu(ent[1].seq) == 1)
295 static struct device *to_dev(struct arena_info *arena)
297 return &arena->nd_btt->dev;
301 * This function copies the desired (old/new) log entry into ent if
302 * it is not NULL. It returns the sub-slot number (0 or 1)
303 * where the desired log entry was found. Negative return values
306 static int btt_log_read(struct arena_info *arena, u32 lane,
307 struct log_entry *ent, int old_flag)
310 int old_ent, ret_ent;
311 struct log_entry log[2];
313 ret = btt_log_read_pair(arena, lane, log);
317 old_ent = btt_log_get_old(log);
318 if (old_ent < 0 || old_ent > 1) {
319 dev_info(to_dev(arena),
320 "log corruption (%d): lane %d seq [%d, %d]\n",
321 old_ent, lane, log[0].seq, log[1].seq);
322 /* TODO set error state? */
326 ret_ent = (old_flag ? old_ent : (1 - old_ent));
329 memcpy(ent, &log[ret_ent], LOG_ENT_SIZE);
335 * This function commits a log entry to media
336 * It does _not_ prepare the freelist entry for the next write
337 * btt_flog_write is the wrapper for updating the freelist elements
339 static int __btt_log_write(struct arena_info *arena, u32 lane,
340 u32 sub, struct log_entry *ent)
344 * Ignore the padding in log_entry for calculating log_half.
345 * The entry is 'committed' when we write the sequence number,
346 * and we want to ensure that that is the last thing written.
347 * We don't bother writing the padding as that would be extra
348 * media wear and write amplification
350 unsigned int log_half = (LOG_ENT_SIZE - 2 * sizeof(u64)) / 2;
351 u64 ns_off = arena->logoff + (((2 * lane) + sub) * LOG_ENT_SIZE);
354 /* split the 16B write into atomic, durable halves */
355 ret = arena_write_bytes(arena, ns_off, src, log_half);
361 return arena_write_bytes(arena, ns_off, src, log_half);
364 static int btt_flog_write(struct arena_info *arena, u32 lane, u32 sub,
365 struct log_entry *ent)
369 ret = __btt_log_write(arena, lane, sub, ent);
373 /* prepare the next free entry */
374 arena->freelist[lane].sub = 1 - arena->freelist[lane].sub;
375 if (++(arena->freelist[lane].seq) == 4)
376 arena->freelist[lane].seq = 1;
377 arena->freelist[lane].block = le32_to_cpu(ent->old_map);
383 * This function initializes the BTT map to the initial state, which is
384 * all-zeroes, and indicates an identity mapping
386 static int btt_map_init(struct arena_info *arena)
391 size_t chunk_size = SZ_2M;
392 size_t mapsize = arena->logoff - arena->mapoff;
394 zerobuf = kzalloc(chunk_size, GFP_KERNEL);
399 size_t size = min(mapsize, chunk_size);
401 ret = arena_write_bytes(arena, arena->mapoff + offset, zerobuf,
417 * This function initializes the BTT log with 'fake' entries pointing
418 * to the initial reserved set of blocks as being free
420 static int btt_log_init(struct arena_info *arena)
424 struct log_entry log, zerolog;
426 memset(&zerolog, 0, sizeof(zerolog));
428 for (i = 0; i < arena->nfree; i++) {
429 log.lba = cpu_to_le32(i);
430 log.old_map = cpu_to_le32(arena->external_nlba + i);
431 log.new_map = cpu_to_le32(arena->external_nlba + i);
432 log.seq = cpu_to_le32(LOG_SEQ_INIT);
433 ret = __btt_log_write(arena, i, 0, &log);
436 ret = __btt_log_write(arena, i, 1, &zerolog);
444 static int btt_freelist_init(struct arena_info *arena)
448 struct log_entry log_new, log_old;
450 arena->freelist = kcalloc(arena->nfree, sizeof(struct free_entry),
452 if (!arena->freelist)
455 for (i = 0; i < arena->nfree; i++) {
456 old = btt_log_read(arena, i, &log_old, LOG_OLD_ENT);
460 new = btt_log_read(arena, i, &log_new, LOG_NEW_ENT);
464 /* sub points to the next one to be overwritten */
465 arena->freelist[i].sub = 1 - new;
466 arena->freelist[i].seq = nd_inc_seq(le32_to_cpu(log_new.seq));
467 arena->freelist[i].block = le32_to_cpu(log_new.old_map);
469 /* This implies a newly created or untouched flog entry */
470 if (log_new.old_map == log_new.new_map)
473 /* Check if map recovery is needed */
474 ret = btt_map_read(arena, le32_to_cpu(log_new.lba), &map_entry,
478 if ((le32_to_cpu(log_new.new_map) != map_entry) &&
479 (le32_to_cpu(log_new.old_map) == map_entry)) {
481 * Last transaction wrote the flog, but wasn't able
482 * to complete the map write. So fix up the map.
484 ret = btt_map_write(arena, le32_to_cpu(log_new.lba),
485 le32_to_cpu(log_new.new_map), 0, 0);
495 static int btt_rtt_init(struct arena_info *arena)
497 arena->rtt = kcalloc(arena->nfree, sizeof(u32), GFP_KERNEL);
498 if (arena->rtt == NULL)
504 static int btt_maplocks_init(struct arena_info *arena)
508 arena->map_locks = kcalloc(arena->nfree, sizeof(struct aligned_lock),
510 if (!arena->map_locks)
513 for (i = 0; i < arena->nfree; i++)
514 spin_lock_init(&arena->map_locks[i].lock);
519 static struct arena_info *alloc_arena(struct btt *btt, size_t size,
520 size_t start, size_t arena_off)
522 struct arena_info *arena;
523 u64 logsize, mapsize, datasize;
524 u64 available = size;
526 arena = kzalloc(sizeof(struct arena_info), GFP_KERNEL);
529 arena->nd_btt = btt->nd_btt;
535 arena->external_lba_start = start;
536 arena->external_lbasize = btt->lbasize;
537 arena->internal_lbasize = roundup(arena->external_lbasize,
538 INT_LBASIZE_ALIGNMENT);
539 arena->nfree = BTT_DEFAULT_NFREE;
540 arena->version_major = 1;
541 arena->version_minor = 1;
543 if (available % BTT_PG_SIZE)
544 available -= (available % BTT_PG_SIZE);
546 /* Two pages are reserved for the super block and its copy */
547 available -= 2 * BTT_PG_SIZE;
549 /* The log takes a fixed amount of space based on nfree */
550 logsize = roundup(2 * arena->nfree * sizeof(struct log_entry),
552 available -= logsize;
554 /* Calculate optimal split between map and data area */
555 arena->internal_nlba = div_u64(available - BTT_PG_SIZE,
556 arena->internal_lbasize + MAP_ENT_SIZE);
557 arena->external_nlba = arena->internal_nlba - arena->nfree;
559 mapsize = roundup((arena->external_nlba * MAP_ENT_SIZE), BTT_PG_SIZE);
560 datasize = available - mapsize;
562 /* 'Absolute' values, relative to start of storage space */
563 arena->infooff = arena_off;
564 arena->dataoff = arena->infooff + BTT_PG_SIZE;
565 arena->mapoff = arena->dataoff + datasize;
566 arena->logoff = arena->mapoff + mapsize;
567 arena->info2off = arena->logoff + logsize;
571 static void free_arenas(struct btt *btt)
573 struct arena_info *arena, *next;
575 list_for_each_entry_safe(arena, next, &btt->arena_list, list) {
576 list_del(&arena->list);
578 kfree(arena->map_locks);
579 kfree(arena->freelist);
580 debugfs_remove_recursive(arena->debugfs_dir);
586 * This function reads an existing valid btt superblock and
587 * populates the corresponding arena_info struct
589 static void parse_arena_meta(struct arena_info *arena, struct btt_sb *super,
592 arena->internal_nlba = le32_to_cpu(super->internal_nlba);
593 arena->internal_lbasize = le32_to_cpu(super->internal_lbasize);
594 arena->external_nlba = le32_to_cpu(super->external_nlba);
595 arena->external_lbasize = le32_to_cpu(super->external_lbasize);
596 arena->nfree = le32_to_cpu(super->nfree);
597 arena->version_major = le16_to_cpu(super->version_major);
598 arena->version_minor = le16_to_cpu(super->version_minor);
600 arena->nextoff = (super->nextoff == 0) ? 0 : (arena_off +
601 le64_to_cpu(super->nextoff));
602 arena->infooff = arena_off;
603 arena->dataoff = arena_off + le64_to_cpu(super->dataoff);
604 arena->mapoff = arena_off + le64_to_cpu(super->mapoff);
605 arena->logoff = arena_off + le64_to_cpu(super->logoff);
606 arena->info2off = arena_off + le64_to_cpu(super->info2off);
608 arena->size = (le64_to_cpu(super->nextoff) > 0)
609 ? (le64_to_cpu(super->nextoff))
610 : (arena->info2off - arena->infooff + BTT_PG_SIZE);
612 arena->flags = le32_to_cpu(super->flags);
615 static int discover_arenas(struct btt *btt)
618 struct arena_info *arena;
619 struct btt_sb *super;
620 size_t remaining = btt->rawsize;
625 super = kzalloc(sizeof(*super), GFP_KERNEL);
630 /* Alloc memory for arena */
631 arena = alloc_arena(btt, 0, 0, 0);
637 arena->infooff = cur_off;
638 ret = btt_info_read(arena, super);
642 if (!nd_btt_arena_is_valid(btt->nd_btt, super)) {
643 if (remaining == btt->rawsize) {
644 btt->init_state = INIT_NOTFOUND;
645 dev_info(to_dev(arena), "No existing arenas\n");
648 dev_info(to_dev(arena),
649 "Found corrupted metadata!\n");
655 arena->external_lba_start = cur_nlba;
656 parse_arena_meta(arena, super, cur_off);
658 ret = btt_freelist_init(arena);
662 ret = btt_rtt_init(arena);
666 ret = btt_maplocks_init(arena);
670 list_add_tail(&arena->list, &btt->arena_list);
672 remaining -= arena->size;
673 cur_off += arena->size;
674 cur_nlba += arena->external_nlba;
677 if (arena->nextoff == 0)
680 btt->num_arenas = num_arenas;
681 btt->nlba = cur_nlba;
682 btt->init_state = INIT_READY;
695 static int create_arenas(struct btt *btt)
697 size_t remaining = btt->rawsize;
701 struct arena_info *arena;
702 size_t arena_size = min_t(u64, ARENA_MAX_SIZE, remaining);
704 remaining -= arena_size;
705 if (arena_size < ARENA_MIN_SIZE)
708 arena = alloc_arena(btt, arena_size, btt->nlba, cur_off);
713 btt->nlba += arena->external_nlba;
714 if (remaining >= ARENA_MIN_SIZE)
715 arena->nextoff = arena->size;
718 cur_off += arena_size;
719 list_add_tail(&arena->list, &btt->arena_list);
726 * This function completes arena initialization by writing
728 * It is only called for an uninitialized arena when a write
729 * to that arena occurs for the first time.
731 static int btt_arena_write_layout(struct arena_info *arena)
735 struct btt_sb *super;
736 struct nd_btt *nd_btt = arena->nd_btt;
737 const u8 *parent_uuid = nd_dev_to_uuid(&nd_btt->ndns->dev);
739 ret = btt_map_init(arena);
743 ret = btt_log_init(arena);
747 super = kzalloc(sizeof(struct btt_sb), GFP_NOIO);
751 strncpy(super->signature, BTT_SIG, BTT_SIG_LEN);
752 memcpy(super->uuid, nd_btt->uuid, 16);
753 memcpy(super->parent_uuid, parent_uuid, 16);
754 super->flags = cpu_to_le32(arena->flags);
755 super->version_major = cpu_to_le16(arena->version_major);
756 super->version_minor = cpu_to_le16(arena->version_minor);
757 super->external_lbasize = cpu_to_le32(arena->external_lbasize);
758 super->external_nlba = cpu_to_le32(arena->external_nlba);
759 super->internal_lbasize = cpu_to_le32(arena->internal_lbasize);
760 super->internal_nlba = cpu_to_le32(arena->internal_nlba);
761 super->nfree = cpu_to_le32(arena->nfree);
762 super->infosize = cpu_to_le32(sizeof(struct btt_sb));
763 super->nextoff = cpu_to_le64(arena->nextoff);
765 * Subtract arena->infooff (arena start) so numbers are relative
768 super->dataoff = cpu_to_le64(arena->dataoff - arena->infooff);
769 super->mapoff = cpu_to_le64(arena->mapoff - arena->infooff);
770 super->logoff = cpu_to_le64(arena->logoff - arena->infooff);
771 super->info2off = cpu_to_le64(arena->info2off - arena->infooff);
774 sum = nd_sb_checksum((struct nd_gen_sb *) super);
775 super->checksum = cpu_to_le64(sum);
777 ret = btt_info_write(arena, super);
784 * This function completes the initialization for the BTT namespace
785 * such that it is ready to accept IOs
787 static int btt_meta_init(struct btt *btt)
790 struct arena_info *arena;
792 mutex_lock(&btt->init_lock);
793 list_for_each_entry(arena, &btt->arena_list, list) {
794 ret = btt_arena_write_layout(arena);
798 ret = btt_freelist_init(arena);
802 ret = btt_rtt_init(arena);
806 ret = btt_maplocks_init(arena);
811 btt->init_state = INIT_READY;
814 mutex_unlock(&btt->init_lock);
818 static u32 btt_meta_size(struct btt *btt)
820 return btt->lbasize - btt->sector_size;
824 * This function calculates the arena in which the given LBA lies
825 * by doing a linear walk. This is acceptable since we expect only
826 * a few arenas. If we have backing devices that get much larger,
827 * we can construct a balanced binary tree of arenas at init time
828 * so that this range search becomes faster.
830 static int lba_to_arena(struct btt *btt, sector_t sector, __u32 *premap,
831 struct arena_info **arena)
833 struct arena_info *arena_list;
834 __u64 lba = div_u64(sector << SECTOR_SHIFT, btt->sector_size);
836 list_for_each_entry(arena_list, &btt->arena_list, list) {
837 if (lba < arena_list->external_nlba) {
842 lba -= arena_list->external_nlba;
849 * The following (lock_map, unlock_map) are mostly just to improve
850 * readability, since they index into an array of locks
852 static void lock_map(struct arena_info *arena, u32 premap)
853 __acquires(&arena->map_locks[idx].lock)
855 u32 idx = (premap * MAP_ENT_SIZE / L1_CACHE_BYTES) % arena->nfree;
857 spin_lock(&arena->map_locks[idx].lock);
860 static void unlock_map(struct arena_info *arena, u32 premap)
861 __releases(&arena->map_locks[idx].lock)
863 u32 idx = (premap * MAP_ENT_SIZE / L1_CACHE_BYTES) % arena->nfree;
865 spin_unlock(&arena->map_locks[idx].lock);
868 static u64 to_namespace_offset(struct arena_info *arena, u64 lba)
870 return arena->dataoff + ((u64)lba * arena->internal_lbasize);
873 static int btt_data_read(struct arena_info *arena, struct page *page,
874 unsigned int off, u32 lba, u32 len)
877 u64 nsoff = to_namespace_offset(arena, lba);
878 void *mem = kmap_atomic(page);
880 ret = arena_read_bytes(arena, nsoff, mem + off, len);
886 static int btt_data_write(struct arena_info *arena, u32 lba,
887 struct page *page, unsigned int off, u32 len)
890 u64 nsoff = to_namespace_offset(arena, lba);
891 void *mem = kmap_atomic(page);
893 ret = arena_write_bytes(arena, nsoff, mem + off, len);
899 static void zero_fill_data(struct page *page, unsigned int off, u32 len)
901 void *mem = kmap_atomic(page);
903 memset(mem + off, 0, len);
907 #ifdef CONFIG_BLK_DEV_INTEGRITY
908 static int btt_rw_integrity(struct btt *btt, struct bio_integrity_payload *bip,
909 struct arena_info *arena, u32 postmap, int rw)
911 unsigned int len = btt_meta_size(btt);
918 meta_nsoff = to_namespace_offset(arena, postmap) + btt->sector_size;
921 unsigned int cur_len;
925 bv = bvec_iter_bvec(bip->bip_vec, bip->bip_iter);
927 * The 'bv' obtained from bvec_iter_bvec has its .bv_len and
928 * .bv_offset already adjusted for iter->bi_bvec_done, and we
929 * can use those directly
932 cur_len = min(len, bv.bv_len);
933 mem = kmap_atomic(bv.bv_page);
935 ret = arena_write_bytes(arena, meta_nsoff,
936 mem + bv.bv_offset, cur_len);
938 ret = arena_read_bytes(arena, meta_nsoff,
939 mem + bv.bv_offset, cur_len);
946 meta_nsoff += cur_len;
947 bvec_iter_advance(bip->bip_vec, &bip->bip_iter, cur_len);
953 #else /* CONFIG_BLK_DEV_INTEGRITY */
954 static int btt_rw_integrity(struct btt *btt, struct bio_integrity_payload *bip,
955 struct arena_info *arena, u32 postmap, int rw)
961 static int btt_read_pg(struct btt *btt, struct bio_integrity_payload *bip,
962 struct page *page, unsigned int off, sector_t sector,
967 struct arena_info *arena = NULL;
968 u32 lane = 0, premap, postmap;
973 lane = nd_region_acquire_lane(btt->nd_region);
975 ret = lba_to_arena(btt, sector, &premap, &arena);
979 cur_len = min(btt->sector_size, len);
981 ret = btt_map_read(arena, premap, &postmap, &t_flag, &e_flag);
986 * We loop to make sure that the post map LBA didn't change
987 * from under us between writing the RTT and doing the actual
994 zero_fill_data(page, off, cur_len);
1003 arena->rtt[lane] = RTT_VALID | postmap;
1005 * Barrier to make sure this write is not reordered
1006 * to do the verification map_read before the RTT store
1010 ret = btt_map_read(arena, premap, &new_map, &t_flag,
1015 if (postmap == new_map)
1021 ret = btt_data_read(arena, page, off, postmap, cur_len);
1026 ret = btt_rw_integrity(btt, bip, arena, postmap, READ);
1031 arena->rtt[lane] = RTT_INVALID;
1032 nd_region_release_lane(btt->nd_region, lane);
1036 sector += btt->sector_size >> SECTOR_SHIFT;
1042 arena->rtt[lane] = RTT_INVALID;
1044 nd_region_release_lane(btt->nd_region, lane);
1048 static int btt_write_pg(struct btt *btt, struct bio_integrity_payload *bip,
1049 sector_t sector, struct page *page, unsigned int off,
1053 struct arena_info *arena = NULL;
1054 u32 premap = 0, old_postmap, new_postmap, lane = 0, i;
1055 struct log_entry log;
1061 lane = nd_region_acquire_lane(btt->nd_region);
1063 ret = lba_to_arena(btt, sector, &premap, &arena);
1066 cur_len = min(btt->sector_size, len);
1068 if ((arena->flags & IB_FLAG_ERROR_MASK) != 0) {
1073 new_postmap = arena->freelist[lane].block;
1075 /* Wait if the new block is being read from */
1076 for (i = 0; i < arena->nfree; i++)
1077 while (arena->rtt[i] == (RTT_VALID | new_postmap))
1081 if (new_postmap >= arena->internal_nlba) {
1086 ret = btt_data_write(arena, new_postmap, page, off, cur_len);
1091 ret = btt_rw_integrity(btt, bip, arena, new_postmap,
1097 lock_map(arena, premap);
1098 ret = btt_map_read(arena, premap, &old_postmap, NULL, NULL);
1101 if (old_postmap >= arena->internal_nlba) {
1106 log.lba = cpu_to_le32(premap);
1107 log.old_map = cpu_to_le32(old_postmap);
1108 log.new_map = cpu_to_le32(new_postmap);
1109 log.seq = cpu_to_le32(arena->freelist[lane].seq);
1110 sub = arena->freelist[lane].sub;
1111 ret = btt_flog_write(arena, lane, sub, &log);
1115 ret = btt_map_write(arena, premap, new_postmap, 0, 0);
1119 unlock_map(arena, premap);
1120 nd_region_release_lane(btt->nd_region, lane);
1124 sector += btt->sector_size >> SECTOR_SHIFT;
1130 unlock_map(arena, premap);
1132 nd_region_release_lane(btt->nd_region, lane);
1136 static int btt_do_bvec(struct btt *btt, struct bio_integrity_payload *bip,
1137 struct page *page, unsigned int len, unsigned int off,
1138 int rw, sector_t sector)
1143 ret = btt_read_pg(btt, bip, page, off, sector, len);
1144 flush_dcache_page(page);
1146 flush_dcache_page(page);
1147 ret = btt_write_pg(btt, bip, sector, page, off, len);
1153 static void btt_make_request(struct request_queue *q, struct bio *bio)
1155 struct bio_integrity_payload *bip = bio_integrity(bio);
1156 struct btt *btt = q->queuedata;
1157 struct bvec_iter iter;
1158 unsigned long start;
1159 struct bio_vec bvec;
1164 * bio_integrity_enabled also checks if the bio already has an
1165 * integrity payload attached. If it does, we *don't* do a
1166 * bio_integrity_prep here - the payload has been generated by
1167 * another kernel subsystem, and we just pass it through.
1169 if (bio_integrity_enabled(bio) && bio_integrity_prep(bio)) {
1170 bio->bi_error = -EIO;
1174 do_acct = nd_iostat_start(bio, &start);
1175 rw = bio_data_dir(bio);
1176 bio_for_each_segment(bvec, bio, iter) {
1177 unsigned int len = bvec.bv_len;
1179 BUG_ON(len > PAGE_SIZE);
1180 /* Make sure len is in multiples of sector size. */
1181 /* XXX is this right? */
1182 BUG_ON(len < btt->sector_size);
1183 BUG_ON(len % btt->sector_size);
1185 err = btt_do_bvec(btt, bip, bvec.bv_page, len, bvec.bv_offset,
1186 rw, iter.bi_sector);
1188 dev_info(&btt->nd_btt->dev,
1189 "io error in %s sector %lld, len %d,\n",
1190 (rw == READ) ? "READ" : "WRITE",
1191 (unsigned long long) iter.bi_sector, len);
1192 bio->bi_error = err;
1197 nd_iostat_end(bio, start);
1203 static int btt_rw_page(struct block_device *bdev, sector_t sector,
1204 struct page *page, int rw)
1206 struct btt *btt = bdev->bd_disk->private_data;
1208 btt_do_bvec(btt, NULL, page, PAGE_CACHE_SIZE, 0, rw, sector);
1209 page_endio(page, rw & WRITE, 0);
1214 static int btt_getgeo(struct block_device *bd, struct hd_geometry *geo)
1216 /* some standard values */
1217 geo->heads = 1 << 6;
1218 geo->sectors = 1 << 5;
1219 geo->cylinders = get_capacity(bd->bd_disk) >> 11;
1223 static const struct block_device_operations btt_fops = {
1224 .owner = THIS_MODULE,
1225 .rw_page = btt_rw_page,
1226 .getgeo = btt_getgeo,
1227 .revalidate_disk = nvdimm_revalidate_disk,
1230 static int btt_blk_init(struct btt *btt)
1232 struct nd_btt *nd_btt = btt->nd_btt;
1233 struct nd_namespace_common *ndns = nd_btt->ndns;
1235 /* create a new disk and request queue for btt */
1236 btt->btt_queue = blk_alloc_queue(GFP_KERNEL);
1237 if (!btt->btt_queue)
1240 btt->btt_disk = alloc_disk(0);
1241 if (!btt->btt_disk) {
1242 blk_cleanup_queue(btt->btt_queue);
1246 nvdimm_namespace_disk_name(ndns, btt->btt_disk->disk_name);
1247 btt->btt_disk->driverfs_dev = &btt->nd_btt->dev;
1248 btt->btt_disk->major = btt_major;
1249 btt->btt_disk->first_minor = 0;
1250 btt->btt_disk->fops = &btt_fops;
1251 btt->btt_disk->private_data = btt;
1252 btt->btt_disk->queue = btt->btt_queue;
1253 btt->btt_disk->flags = GENHD_FL_EXT_DEVT;
1255 blk_queue_make_request(btt->btt_queue, btt_make_request);
1256 blk_queue_logical_block_size(btt->btt_queue, btt->sector_size);
1257 blk_queue_max_hw_sectors(btt->btt_queue, UINT_MAX);
1258 blk_queue_bounce_limit(btt->btt_queue, BLK_BOUNCE_ANY);
1259 queue_flag_set_unlocked(QUEUE_FLAG_NONROT, btt->btt_queue);
1260 btt->btt_queue->queuedata = btt;
1262 set_capacity(btt->btt_disk, 0);
1263 add_disk(btt->btt_disk);
1264 if (btt_meta_size(btt)) {
1265 int rc = nd_integrity_init(btt->btt_disk, btt_meta_size(btt));
1268 del_gendisk(btt->btt_disk);
1269 put_disk(btt->btt_disk);
1270 blk_cleanup_queue(btt->btt_queue);
1274 set_capacity(btt->btt_disk, btt->nlba * btt->sector_size >> 9);
1275 revalidate_disk(btt->btt_disk);
1280 static void btt_blk_cleanup(struct btt *btt)
1282 del_gendisk(btt->btt_disk);
1283 put_disk(btt->btt_disk);
1284 blk_cleanup_queue(btt->btt_queue);
1288 * btt_init - initialize a block translation table for the given device
1289 * @nd_btt: device with BTT geometry and backing device info
1290 * @rawsize: raw size in bytes of the backing device
1291 * @lbasize: lba size of the backing device
1292 * @uuid: A uuid for the backing device - this is stored on media
1293 * @maxlane: maximum number of parallel requests the device can handle
1295 * Initialize a Block Translation Table on a backing device to provide
1296 * single sector power fail atomicity.
1302 * Pointer to a new struct btt on success, NULL on failure.
1304 static struct btt *btt_init(struct nd_btt *nd_btt, unsigned long long rawsize,
1305 u32 lbasize, u8 *uuid, struct nd_region *nd_region)
1309 struct device *dev = &nd_btt->dev;
1311 btt = kzalloc(sizeof(struct btt), GFP_KERNEL);
1315 btt->nd_btt = nd_btt;
1316 btt->rawsize = rawsize;
1317 btt->lbasize = lbasize;
1318 btt->sector_size = ((lbasize >= 4096) ? 4096 : 512);
1319 INIT_LIST_HEAD(&btt->arena_list);
1320 mutex_init(&btt->init_lock);
1321 btt->nd_region = nd_region;
1323 ret = discover_arenas(btt);
1325 dev_err(dev, "init: error in arena_discover: %d\n", ret);
1329 if (btt->init_state != INIT_READY && nd_region->ro) {
1330 dev_info(dev, "%s is read-only, unable to init btt metadata\n",
1331 dev_name(&nd_region->dev));
1333 } else if (btt->init_state != INIT_READY) {
1334 btt->num_arenas = (rawsize / ARENA_MAX_SIZE) +
1335 ((rawsize % ARENA_MAX_SIZE) ? 1 : 0);
1336 dev_dbg(dev, "init: %d arenas for %llu rawsize\n",
1337 btt->num_arenas, rawsize);
1339 ret = create_arenas(btt);
1341 dev_info(dev, "init: create_arenas: %d\n", ret);
1345 ret = btt_meta_init(btt);
1347 dev_err(dev, "init: error in meta_init: %d\n", ret);
1352 ret = btt_blk_init(btt);
1354 dev_err(dev, "init: error in blk_init: %d\n", ret);
1358 btt_debugfs_init(btt);
1368 * btt_fini - de-initialize a BTT
1369 * @btt: the BTT handle that was generated by btt_init
1371 * De-initialize a Block Translation Table on device removal
1376 static void btt_fini(struct btt *btt)
1379 btt_blk_cleanup(btt);
1381 debugfs_remove_recursive(btt->debugfs_dir);
1386 int nvdimm_namespace_attach_btt(struct nd_namespace_common *ndns)
1388 struct nd_btt *nd_btt = to_nd_btt(ndns->claim);
1389 struct nd_region *nd_region;
1393 if (!nd_btt->uuid || !nd_btt->ndns || !nd_btt->lbasize)
1396 rawsize = nvdimm_namespace_capacity(ndns) - SZ_4K;
1397 if (rawsize < ARENA_MIN_SIZE) {
1400 nd_region = to_nd_region(nd_btt->dev.parent);
1401 btt = btt_init(nd_btt, rawsize, nd_btt->lbasize, nd_btt->uuid,
1409 EXPORT_SYMBOL(nvdimm_namespace_attach_btt);
1411 int nvdimm_namespace_detach_btt(struct nd_namespace_common *ndns)
1413 struct nd_btt *nd_btt = to_nd_btt(ndns->claim);
1414 struct btt *btt = nd_btt->btt;
1421 EXPORT_SYMBOL(nvdimm_namespace_detach_btt);
1423 static int __init nd_btt_init(void)
1427 btt_major = register_blkdev(0, "btt");
1431 debugfs_root = debugfs_create_dir("btt", NULL);
1432 if (IS_ERR_OR_NULL(debugfs_root)) {
1440 unregister_blkdev(btt_major, "btt");
1445 static void __exit nd_btt_exit(void)
1447 debugfs_remove_recursive(debugfs_root);
1448 unregister_blkdev(btt_major, "btt");
1451 MODULE_ALIAS_ND_DEVICE(ND_DEVICE_BTT);
1452 MODULE_AUTHOR("Vishal Verma <vishal.l.verma@linux.intel.com>");
1453 MODULE_LICENSE("GPL v2");
1454 module_init(nd_btt_init);
1455 module_exit(nd_btt_exit);