diff options
Diffstat (limited to 'drivers/md')
80 files changed, 2153 insertions, 1403 deletions
diff --git a/drivers/md/Kconfig b/drivers/md/Kconfig index 0b1870a09e1f..ddb37f6670de 100644 --- a/drivers/md/Kconfig +++ b/drivers/md/Kconfig @@ -139,7 +139,7 @@ config MD_RAID456 tristate "RAID-4/RAID-5/RAID-6 mode" depends on BLK_DEV_MD select RAID6_PQ - select LIBCRC32C + select CRC32 select ASYNC_MEMCPY select ASYNC_XOR select ASYNC_PQ @@ -267,6 +267,7 @@ config DM_CRYPT depends on BLK_DEV_DM depends on (ENCRYPTED_KEYS || ENCRYPTED_KEYS=n) depends on (TRUSTED_KEYS || TRUSTED_KEYS=n) + select CRC32 select CRYPTO select CRYPTO_CBC select CRYPTO_ESSIV diff --git a/drivers/md/bcache/Kconfig b/drivers/md/bcache/Kconfig index d4697e79d5a3..b2d10063d35f 100644 --- a/drivers/md/bcache/Kconfig +++ b/drivers/md/bcache/Kconfig @@ -5,7 +5,6 @@ config BCACHE select BLOCK_HOLDER_DEPRECATED if SYSFS select CRC64 select CLOSURES - select MIN_HEAP help Allows a block device to be used as cache for other devices; uses a btree for indexing and the layout is optimized for SSDs. diff --git a/drivers/md/bcache/alloc.c b/drivers/md/bcache/alloc.c index 8998e61efa40..48ce750bf70a 100644 --- a/drivers/md/bcache/alloc.c +++ b/drivers/md/bcache/alloc.c @@ -164,61 +164,40 @@ static void bch_invalidate_one_bucket(struct cache *ca, struct bucket *b) * prio is worth 1/8th of what INITIAL_PRIO is worth. */ -static inline unsigned int new_bucket_prio(struct cache *ca, struct bucket *b) -{ - unsigned int min_prio = (INITIAL_PRIO - ca->set->min_prio) / 8; - - return (b->prio - ca->set->min_prio + min_prio) * GC_SECTORS_USED(b); -} - -static inline bool new_bucket_max_cmp(const void *l, const void *r, void *args) -{ - struct bucket **lhs = (struct bucket **)l; - struct bucket **rhs = (struct bucket **)r; - struct cache *ca = args; - - return new_bucket_prio(ca, *lhs) > new_bucket_prio(ca, *rhs); -} - -static inline bool new_bucket_min_cmp(const void *l, const void *r, void *args) -{ - struct bucket **lhs = (struct bucket **)l; - struct bucket **rhs = (struct bucket **)r; - struct cache *ca = args; +#define bucket_prio(b) \ +({ \ + unsigned int min_prio = (INITIAL_PRIO - ca->set->min_prio) / 8; \ + \ + (b->prio - ca->set->min_prio + min_prio) * GC_SECTORS_USED(b); \ +}) - return new_bucket_prio(ca, *lhs) < new_bucket_prio(ca, *rhs); -} +#define bucket_max_cmp(l, r) (bucket_prio(l) < bucket_prio(r)) +#define bucket_min_cmp(l, r) (bucket_prio(l) > bucket_prio(r)) static void invalidate_buckets_lru(struct cache *ca) { struct bucket *b; - const struct min_heap_callbacks bucket_max_cmp_callback = { - .less = new_bucket_max_cmp, - .swp = NULL, - }; - const struct min_heap_callbacks bucket_min_cmp_callback = { - .less = new_bucket_min_cmp, - .swp = NULL, - }; + ssize_t i; - ca->heap.nr = 0; + ca->heap.used = 0; for_each_bucket(b, ca) { if (!bch_can_invalidate_bucket(ca, b)) continue; - if (!min_heap_full(&ca->heap)) - min_heap_push(&ca->heap, &b, &bucket_max_cmp_callback, ca); - else if (!new_bucket_max_cmp(&b, min_heap_peek(&ca->heap), ca)) { + if (!heap_full(&ca->heap)) + heap_add(&ca->heap, b, bucket_max_cmp); + else if (bucket_max_cmp(b, heap_peek(&ca->heap))) { ca->heap.data[0] = b; - min_heap_sift_down(&ca->heap, 0, &bucket_max_cmp_callback, ca); + heap_sift(&ca->heap, 0, bucket_max_cmp); } } - min_heapify_all(&ca->heap, &bucket_min_cmp_callback, ca); + for (i = ca->heap.used / 2 - 1; i >= 0; --i) + heap_sift(&ca->heap, i, bucket_min_cmp); while (!fifo_full(&ca->free_inc)) { - if (!ca->heap.nr) { + if (!heap_pop(&ca->heap, b, bucket_min_cmp)) { /* * We don't want to be calling invalidate_buckets() * multiple times when it can't do anything @@ -227,8 +206,6 @@ static void invalidate_buckets_lru(struct cache *ca) wake_up_gc(ca->set); return; } - b = min_heap_peek(&ca->heap)[0]; - min_heap_pop(&ca->heap, &bucket_min_cmp_callback, ca); bch_invalidate_one_bucket(ca, b); } diff --git a/drivers/md/bcache/bcache.h b/drivers/md/bcache/bcache.h index 785b0d9008fa..1d33e40d26ea 100644 --- a/drivers/md/bcache/bcache.h +++ b/drivers/md/bcache/bcache.h @@ -458,7 +458,7 @@ struct cache { /* Allocation stuff: */ struct bucket *buckets; - DEFINE_MIN_HEAP(struct bucket *, cache_heap) heap; + DECLARE_HEAP(struct bucket *, heap); /* * If nonzero, we know we aren't going to find any buckets to invalidate diff --git a/drivers/md/bcache/bset.c b/drivers/md/bcache/bset.c index 68258a16e125..463eb13bd0b2 100644 --- a/drivers/md/bcache/bset.c +++ b/drivers/md/bcache/bset.c @@ -54,11 +54,9 @@ void bch_dump_bucket(struct btree_keys *b) int __bch_count_data(struct btree_keys *b) { unsigned int ret = 0; - struct btree_iter iter; + struct btree_iter_stack iter; struct bkey *k; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - if (b->ops->is_extents) for_each_key(b, k, &iter) ret += KEY_SIZE(k); @@ -69,11 +67,9 @@ void __bch_check_keys(struct btree_keys *b, const char *fmt, ...) { va_list args; struct bkey *k, *p = NULL; - struct btree_iter iter; + struct btree_iter_stack iter; const char *err; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - for_each_key(b, k, &iter) { if (b->ops->is_extents) { err = "Keys out of order"; @@ -114,9 +110,9 @@ bug: static void bch_btree_iter_next_check(struct btree_iter *iter) { - struct bkey *k = iter->heap.data->k, *next = bkey_next(k); + struct bkey *k = iter->data->k, *next = bkey_next(k); - if (next < iter->heap.data->end && + if (next < iter->data->end && bkey_cmp(k, iter->b->ops->is_extents ? &START_KEY(next) : next) > 0) { bch_dump_bucket(iter->b); @@ -883,14 +879,12 @@ unsigned int bch_btree_insert_key(struct btree_keys *b, struct bkey *k, unsigned int status = BTREE_INSERT_STATUS_NO_INSERT; struct bset *i = bset_tree_last(b)->data; struct bkey *m, *prev = NULL; - struct btree_iter iter; + struct btree_iter_stack iter; struct bkey preceding_key_on_stack = ZERO_KEY; struct bkey *preceding_key_p = &preceding_key_on_stack; BUG_ON(b->ops->is_extents && !KEY_SIZE(k)); - min_heap_init(&iter.heap, NULL, MAX_BSETS); - /* * If k has preceding key, preceding_key_p will be set to address * of k's preceding key; otherwise preceding_key_p will be set @@ -901,9 +895,9 @@ unsigned int bch_btree_insert_key(struct btree_keys *b, struct bkey *k, else preceding_key(k, &preceding_key_p); - m = bch_btree_iter_init(b, &iter, preceding_key_p); + m = bch_btree_iter_stack_init(b, &iter, preceding_key_p); - if (b->ops->insert_fixup(b, k, &iter, replace_key)) + if (b->ops->insert_fixup(b, k, &iter.iter, replace_key)) return status; status = BTREE_INSERT_STATUS_INSERT; @@ -1083,94 +1077,79 @@ struct bkey *__bch_bset_search(struct btree_keys *b, struct bset_tree *t, /* Btree iterator */ -typedef bool (new_btree_iter_cmp_fn)(const void *, const void *, void *); +typedef bool (btree_iter_cmp_fn)(struct btree_iter_set, + struct btree_iter_set); -static inline bool new_btree_iter_cmp(const void *l, const void *r, void __always_unused *args) +static inline bool btree_iter_cmp(struct btree_iter_set l, + struct btree_iter_set r) { - const struct btree_iter_set *_l = l; - const struct btree_iter_set *_r = r; - - return bkey_cmp(_l->k, _r->k) <= 0; + return bkey_cmp(l.k, r.k) > 0; } static inline bool btree_iter_end(struct btree_iter *iter) { - return !iter->heap.nr; + return !iter->used; } void bch_btree_iter_push(struct btree_iter *iter, struct bkey *k, struct bkey *end) { - const struct min_heap_callbacks callbacks = { - .less = new_btree_iter_cmp, - .swp = NULL, - }; - if (k != end) - BUG_ON(!min_heap_push(&iter->heap, - &((struct btree_iter_set) { k, end }), - &callbacks, - NULL)); + BUG_ON(!heap_add(iter, + ((struct btree_iter_set) { k, end }), + btree_iter_cmp)); } -static struct bkey *__bch_btree_iter_init(struct btree_keys *b, - struct btree_iter *iter, - struct bkey *search, - struct bset_tree *start) +static struct bkey *__bch_btree_iter_stack_init(struct btree_keys *b, + struct btree_iter_stack *iter, + struct bkey *search, + struct bset_tree *start) { struct bkey *ret = NULL; - iter->heap.size = ARRAY_SIZE(iter->heap.preallocated); - iter->heap.nr = 0; + iter->iter.size = ARRAY_SIZE(iter->stack_data); + iter->iter.used = 0; #ifdef CONFIG_BCACHE_DEBUG - iter->b = b; + iter->iter.b = b; #endif for (; start <= bset_tree_last(b); start++) { ret = bch_bset_search(b, start, search); - bch_btree_iter_push(iter, ret, bset_bkey_last(start->data)); + bch_btree_iter_push(&iter->iter, ret, bset_bkey_last(start->data)); } return ret; } -struct bkey *bch_btree_iter_init(struct btree_keys *b, - struct btree_iter *iter, +struct bkey *bch_btree_iter_stack_init(struct btree_keys *b, + struct btree_iter_stack *iter, struct bkey *search) { - return __bch_btree_iter_init(b, iter, search, b->set); + return __bch_btree_iter_stack_init(b, iter, search, b->set); } static inline struct bkey *__bch_btree_iter_next(struct btree_iter *iter, - new_btree_iter_cmp_fn *cmp) + btree_iter_cmp_fn *cmp) { struct btree_iter_set b __maybe_unused; struct bkey *ret = NULL; - const struct min_heap_callbacks callbacks = { - .less = cmp, - .swp = NULL, - }; if (!btree_iter_end(iter)) { bch_btree_iter_next_check(iter); - ret = iter->heap.data->k; - iter->heap.data->k = bkey_next(iter->heap.data->k); + ret = iter->data->k; + iter->data->k = bkey_next(iter->data->k); - if (iter->heap.data->k > iter->heap.data->end) { + if (iter->data->k > iter->data->end) { WARN_ONCE(1, "bset was corrupt!\n"); - iter->heap.data->k = iter->heap.data->end; + iter->data->k = iter->data->end; } - if (iter->heap.data->k == iter->heap.data->end) { - if (iter->heap.nr) { - b = min_heap_peek(&iter->heap)[0]; - min_heap_pop(&iter->heap, &callbacks, NULL); - } - } + if (iter->data->k == iter->data->end) + heap_pop(iter, b, cmp); else - min_heap_sift_down(&iter->heap, 0, &callbacks, NULL); + heap_sift(iter, 0, cmp); } return ret; @@ -1178,7 +1157,7 @@ static inline struct bkey *__bch_btree_iter_next(struct btree_iter *iter, struct bkey *bch_btree_iter_next(struct btree_iter *iter) { - return __bch_btree_iter_next(iter, new_btree_iter_cmp); + return __bch_btree_iter_next(iter, btree_iter_cmp); } @@ -1216,18 +1195,16 @@ static void btree_mergesort(struct btree_keys *b, struct bset *out, struct btree_iter *iter, bool fixup, bool remove_stale) { + int i; struct bkey *k, *last = NULL; BKEY_PADDED(k) tmp; bool (*bad)(struct btree_keys *, const struct bkey *) = remove_stale ? bch_ptr_bad : bch_ptr_invalid; - const struct min_heap_callbacks callbacks = { - .less = b->ops->sort_cmp, - .swp = NULL, - }; /* Heapify the iterator, using our comparison function */ - min_heapify_all(&iter->heap, &callbacks, NULL); + for (i = iter->used / 2 - 1; i >= 0; --i) + heap_sift(iter, i, b->ops->sort_cmp); while (!btree_iter_end(iter)) { if (b->ops->sort_fixup && fixup) @@ -1316,11 +1293,10 @@ void bch_btree_sort_partial(struct btree_keys *b, unsigned int start, struct bset_sort_state *state) { size_t order = b->page_order, keys = 0; - struct btree_iter iter; + struct btree_iter_stack iter; int oldsize = bch_count_data(b); - min_heap_init(&iter.heap, NULL, MAX_BSETS); - __bch_btree_iter_init(b, &iter, NULL, &b->set[start]); + __bch_btree_iter_stack_init(b, &iter, NULL, &b->set[start]); if (start) { unsigned int i; @@ -1331,7 +1307,7 @@ void bch_btree_sort_partial(struct btree_keys *b, unsigned int start, order = get_order(__set_bytes(b->set->data, keys)); } - __btree_sort(b, &iter, start, order, false, state); + __btree_sort(b, &iter.iter, start, order, false, state); EBUG_ON(oldsize >= 0 && bch_count_data(b) != oldsize); } @@ -1347,13 +1323,11 @@ void bch_btree_sort_into(struct btree_keys *b, struct btree_keys *new, struct bset_sort_state *state) { uint64_t start_time = local_clock(); - struct btree_iter iter; - - min_heap_init(&iter.heap, NULL, MAX_BSETS); + struct btree_iter_stack iter; - bch_btree_iter_init(b, &iter, NULL); + bch_btree_iter_stack_init(b, &iter, NULL); - btree_mergesort(b, new->set->data, &iter, false, true); + btree_mergesort(b, new->set->data, &iter.iter, false, true); bch_time_stats_update(&state->time, start_time); diff --git a/drivers/md/bcache/bset.h b/drivers/md/bcache/bset.h index f79441acd4c1..011f6062c4c0 100644 --- a/drivers/md/bcache/bset.h +++ b/drivers/md/bcache/bset.h @@ -187,9 +187,8 @@ struct bset_tree { }; struct btree_keys_ops { - bool (*sort_cmp)(const void *l, - const void *r, - void *args); + bool (*sort_cmp)(struct btree_iter_set l, + struct btree_iter_set r); struct bkey *(*sort_fixup)(struct btree_iter *iter, struct bkey *tmp); bool (*insert_fixup)(struct btree_keys *b, @@ -313,17 +312,23 @@ enum { BTREE_INSERT_STATUS_FRONT_MERGE, }; -struct btree_iter_set { - struct bkey *k, *end; -}; - /* Btree key iteration */ struct btree_iter { + size_t size, used; #ifdef CONFIG_BCACHE_DEBUG struct btree_keys *b; #endif - MIN_HEAP_PREALLOCATED(struct btree_iter_set, btree_iter_heap, MAX_BSETS) heap; + struct btree_iter_set { + struct bkey *k, *end; + } data[]; +}; + +/* Fixed-size btree_iter that can be allocated on the stack */ + +struct btree_iter_stack { + struct btree_iter iter; + struct btree_iter_set stack_data[MAX_BSETS]; }; typedef bool (*ptr_filter_fn)(struct btree_keys *b, const struct bkey *k); @@ -335,9 +340,9 @@ struct bkey *bch_btree_iter_next_filter(struct btree_iter *iter, void bch_btree_iter_push(struct btree_iter *iter, struct bkey *k, struct bkey *end); -struct bkey *bch_btree_iter_init(struct btree_keys *b, - struct btree_iter *iter, - struct bkey *search); +struct bkey *bch_btree_iter_stack_init(struct btree_keys *b, + struct btree_iter_stack *iter, + struct bkey *search); struct bkey *__bch_bset_search(struct btree_keys *b, struct bset_tree *t, const struct bkey *search); @@ -352,13 +357,14 @@ static inline struct bkey *bch_bset_search(struct btree_keys *b, return search ? __bch_bset_search(b, t, search) : t->data->start; } -#define for_each_key_filter(b, k, iter, filter) \ - for (bch_btree_iter_init((b), (iter), NULL); \ - ((k) = bch_btree_iter_next_filter((iter), (b), filter));) +#define for_each_key_filter(b, k, stack_iter, filter) \ + for (bch_btree_iter_stack_init((b), (stack_iter), NULL); \ + ((k) = bch_btree_iter_next_filter(&((stack_iter)->iter), (b), \ + filter));) -#define for_each_key(b, k, iter) \ - for (bch_btree_iter_init((b), (iter), NULL); \ - ((k) = bch_btree_iter_next(iter));) +#define for_each_key(b, k, stack_iter) \ + for (bch_btree_iter_stack_init((b), (stack_iter), NULL); \ + ((k) = bch_btree_iter_next(&((stack_iter)->iter)));) /* Sorting */ diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c index ed40d8600656..210b59007d98 100644 --- a/drivers/md/bcache/btree.c +++ b/drivers/md/bcache/btree.c @@ -36,6 +36,7 @@ #include <linux/sched/clock.h> #include <linux/rculist.h> #include <linux/delay.h> +#include <linux/sort.h> #include <trace/events/bcache.h> /* @@ -88,8 +89,6 @@ * Test module load/unload */ -#define MAX_NEED_GC 64 -#define MAX_SAVE_PRIO 72 #define MAX_GC_TIMES 100 #define MIN_GC_NODES 100 #define GC_SLEEP_MS 100 @@ -149,19 +148,19 @@ void bch_btree_node_read_done(struct btree *b) { const char *err = "bad btree header"; struct bset *i = btree_bset_first(b); - struct btree_iter iter; + struct btree_iter *iter; /* * c->fill_iter can allocate an iterator with more memory space * than static MAX_BSETS. * See the comment arount cache_set->fill_iter. */ - iter.heap.data = mempool_alloc(&b->c->fill_iter, GFP_NOIO); - iter.heap.size = b->c->cache->sb.bucket_size / b->c->cache->sb.block_size; - iter.heap.nr = 0; + iter = mempool_alloc(&b->c->fill_iter, GFP_NOIO); + iter->size = b->c->cache->sb.bucket_size / b->c->cache->sb.block_size; + iter->used = 0; #ifdef CONFIG_BCACHE_DEBUG - iter.b = &b->keys; + iter->b = &b->keys; #endif if (!i->seq) @@ -199,7 +198,7 @@ void bch_btree_node_read_done(struct btree *b) if (i != b->keys.set[0].data && !i->keys) goto err; - bch_btree_iter_push(&iter, i->start, bset_bkey_last(i)); + bch_btree_iter_push(iter, i->start, bset_bkey_last(i)); b->written += set_blocks(i, block_bytes(b->c->cache)); } @@ -211,7 +210,7 @@ void bch_btree_node_read_done(struct btree *b) if (i->seq == b->keys.set[0].data->seq) goto err; - bch_btree_sort_and_fix_extents(&b->keys, &iter, &b->c->sort); + bch_btree_sort_and_fix_extents(&b->keys, iter, &b->c->sort); i = b->keys.set[0].data; err = "short btree key"; @@ -223,7 +222,7 @@ void bch_btree_node_read_done(struct btree *b) bch_bset_init_next(&b->keys, write_block(b), bset_magic(&b->c->cache->sb)); out: - mempool_free(iter.heap.data, &b->c->fill_iter); + mempool_free(iter, &b->c->fill_iter); return; err: set_btree_node_io_error(b); @@ -559,8 +558,6 @@ static void mca_data_alloc(struct btree *b, struct bkey *k, gfp_t gfp) } } -#define cmp_int(l, r) ((l > r) - (l < r)) - #ifdef CONFIG_PROVE_LOCKING static int btree_lock_cmp_fn(const struct lockdep_map *_a, const struct lockdep_map *_b) @@ -1309,11 +1306,9 @@ static bool btree_gc_mark_node(struct btree *b, struct gc_stat *gc) uint8_t stale = 0; unsigned int keys = 0, good_keys = 0; struct bkey *k; - struct btree_iter iter; + struct btree_iter_stack iter; struct bset_tree *t; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - gc->nodes++; for_each_key_filter(&b->keys, k, &iter, bch_ptr_invalid) { @@ -1572,11 +1567,9 @@ static int btree_gc_rewrite_node(struct btree *b, struct btree_op *op, static unsigned int btree_gc_count_keys(struct btree *b) { struct bkey *k; - struct btree_iter iter; + struct btree_iter_stack iter; unsigned int ret = 0; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - for_each_key_filter(&b->keys, k, &iter, bch_ptr_bad) ret += bkey_u64s(k); @@ -1615,18 +1608,18 @@ static int btree_gc_recurse(struct btree *b, struct btree_op *op, int ret = 0; bool should_rewrite; struct bkey *k; - struct btree_iter iter; + struct btree_iter_stack iter; struct gc_merge_info r[GC_MERGE_NODES]; struct gc_merge_info *i, *last = r + ARRAY_SIZE(r) - 1; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - bch_btree_iter_init(&b->keys, &iter, &b->c->gc_done); + bch_btree_iter_stack_init(&b->keys, &iter, &b->c->gc_done); for (i = r; i < r + ARRAY_SIZE(r); i++) i->b = ERR_PTR(-EINTR); while (1) { - k = bch_btree_iter_next_filter(&iter, &b->keys, bch_ptr_bad); + k = bch_btree_iter_next_filter(&iter.iter, &b->keys, + bch_ptr_bad); if (k) { r->b = bch_btree_node_get(b->c, op, k, b->level - 1, true, b); @@ -1921,9 +1914,7 @@ static int bch_btree_check_recurse(struct btree *b, struct btree_op *op) { int ret = 0; struct bkey *k, *p = NULL; - struct btree_iter iter; - - min_heap_init(&iter.heap, NULL, MAX_BSETS); + struct btree_iter_stack iter; for_each_key_filter(&b->keys, k, &iter, bch_ptr_invalid) bch_initial_mark_key(b->c, b->level, k); @@ -1931,10 +1922,10 @@ static int bch_btree_check_recurse(struct btree *b, struct btree_op *op) bch_initial_mark_key(b->c, b->level + 1, &b->key); if (b->level) { - bch_btree_iter_init(&b->keys, &iter, NULL); + bch_btree_iter_stack_init(&b->keys, &iter, NULL); do { - k = bch_btree_iter_next_filter(&iter, &b->keys, + k = bch_btree_iter_next_filter(&iter.iter, &b->keys, bch_ptr_bad); if (k) { btree_node_prefetch(b, k); @@ -1962,7 +1953,7 @@ static int bch_btree_check_thread(void *arg) struct btree_check_info *info = arg; struct btree_check_state *check_state = info->state; struct cache_set *c = check_state->c; - struct btree_iter iter; + struct btree_iter_stack iter; struct bkey *k, *p; int cur_idx, prev_idx, skip_nr; @@ -1970,11 +1961,9 @@ static int bch_btree_check_thread(void *arg) cur_idx = prev_idx = 0; ret = 0; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - /* root node keys are checked before thread created */ - bch_btree_iter_init(&c->root->keys, &iter, NULL); - k = bch_btree_iter_next_filter(&iter, &c->root->keys, bch_ptr_bad); + bch_btree_iter_stack_init(&c->root->keys, &iter, NULL); + k = bch_btree_iter_next_filter(&iter.iter, &c->root->keys, bch_ptr_bad); BUG_ON(!k); p = k; @@ -1992,7 +1981,7 @@ static int bch_btree_check_thread(void *arg) skip_nr = cur_idx - prev_idx; while (skip_nr) { - k = bch_btree_iter_next_filter(&iter, + k = bch_btree_iter_next_filter(&iter.iter, &c->root->keys, bch_ptr_bad); if (k) @@ -2065,11 +2054,9 @@ int bch_btree_check(struct cache_set *c) int ret = 0; int i; struct bkey *k = NULL; - struct btree_iter iter; + struct btree_iter_stack iter; struct btree_check_state check_state; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - /* check and mark root node keys */ for_each_key_filter(&c->root->keys, k, &iter, bch_ptr_invalid) bch_initial_mark_key(c, c->root->level, k); @@ -2563,12 +2550,11 @@ static int bch_btree_map_nodes_recurse(struct btree *b, struct btree_op *op, if (b->level) { struct bkey *k; - struct btree_iter iter; + struct btree_iter_stack iter; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - bch_btree_iter_init(&b->keys, &iter, from); + bch_btree_iter_stack_init(&b->keys, &iter, from); - while ((k = bch_btree_iter_next_filter(&iter, &b->keys, + while ((k = bch_btree_iter_next_filter(&iter.iter, &b->keys, bch_ptr_bad))) { ret = bcache_btree(map_nodes_recurse, k, b, op, from, fn, flags); @@ -2597,12 +2583,12 @@ int bch_btree_map_keys_recurse(struct btree *b, struct btree_op *op, { int ret = MAP_CONTINUE; struct bkey *k; - struct btree_iter iter; + struct btree_iter_stack iter; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - bch_btree_iter_init(&b->keys, &iter, from); + bch_btree_iter_stack_init(&b->keys, &iter, from); - while ((k = bch_btree_iter_next_filter(&iter, &b->keys, bch_ptr_bad))) { + while ((k = bch_btree_iter_next_filter(&iter.iter, &b->keys, + bch_ptr_bad))) { ret = !b->level ? fn(op, b, k) : bcache_btree(map_keys_recurse, k, diff --git a/drivers/md/bcache/extents.c b/drivers/md/bcache/extents.c index 4b84fda1530a..d626ffcbecb9 100644 --- a/drivers/md/bcache/extents.c +++ b/drivers/md/bcache/extents.c @@ -33,16 +33,15 @@ static void sort_key_next(struct btree_iter *iter, i->k = bkey_next(i->k); if (i->k == i->end) - *i = iter->heap.data[--iter->heap.nr]; + *i = iter->data[--iter->used]; } -static bool new_bch_key_sort_cmp(const void *l, const void *r, void *args) +static bool bch_key_sort_cmp(struct btree_iter_set l, + struct btree_iter_set r) { - struct btree_iter_set *_l = (struct btree_iter_set *)l; - struct btree_iter_set *_r = (struct btree_iter_set *)r; - int64_t c = bkey_cmp(_l->k, _r->k); + int64_t c = bkey_cmp(l.k, r.k); - return !(c ? c > 0 : _l->k < _r->k); + return c ? c > 0 : l.k < r.k; } static bool __ptr_invalid(struct cache_set *c, const struct bkey *k) @@ -239,7 +238,7 @@ static bool bch_btree_ptr_insert_fixup(struct btree_keys *bk, } const struct btree_keys_ops bch_btree_keys_ops = { - .sort_cmp = new_bch_key_sort_cmp, + .sort_cmp = bch_key_sort_cmp, .insert_fixup = bch_btree_ptr_insert_fixup, .key_invalid = bch_btree_ptr_invalid, .key_bad = bch_btree_ptr_bad, @@ -256,28 +255,22 @@ const struct btree_keys_ops bch_btree_keys_ops = { * Necessary for btree_sort_fixup() - if there are multiple keys that compare * equal in different sets, we have to process them newest to oldest. */ - -static bool new_bch_extent_sort_cmp(const void *l, const void *r, void __always_unused *args) +static bool bch_extent_sort_cmp(struct btree_iter_set l, + struct btree_iter_set r) { - struct btree_iter_set *_l = (struct btree_iter_set *)l; - struct btree_iter_set *_r = (struct btree_iter_set *)r; - int64_t c = bkey_cmp(&START_KEY(_l->k), &START_KEY(_r->k)); + int64_t c = bkey_cmp(&START_KEY(l.k), &START_KEY(r.k)); - return !(c ? c > 0 : _l->k < _r->k); + return c ? c > 0 : l.k < r.k; } static struct bkey *bch_extent_sort_fixup(struct btree_iter *iter, struct bkey *tmp) { - const struct min_heap_callbacks callbacks = { - .less = new_bch_extent_sort_cmp, - .swp = NULL, - }; - while (iter->heap.nr > 1) { - struct btree_iter_set *top = iter->heap.data, *i = top + 1; - - if (iter->heap.nr > 2 && - !new_bch_extent_sort_cmp(&i[0], &i[1], NULL)) + while (iter->used > 1) { + struct btree_iter_set *top = iter->data, *i = top + 1; + + if (iter->used > 2 && + bch_extent_sort_cmp(i[0], i[1])) i++; if (bkey_cmp(top->k, &START_KEY(i->k)) <= 0) @@ -285,7 +278,7 @@ static struct bkey *bch_extent_sort_fixup(struct btree_iter *iter, if (!KEY_SIZE(i->k)) { sort_key_next(iter, i); - min_heap_sift_down(&iter->heap, i - top, &callbacks, NULL); + heap_sift(iter, i - top, bch_extent_sort_cmp); continue; } @@ -295,7 +288,7 @@ static struct bkey *bch_extent_sort_fixup(struct btree_iter *iter, else bch_cut_front(top->k, i->k); - min_heap_sift_down(&iter->heap, i - top, &callbacks, NULL); + heap_sift(iter, i - top, bch_extent_sort_cmp); } else { /* can't happen because of comparison func */ BUG_ON(!bkey_cmp(&START_KEY(top->k), &START_KEY(i->k))); @@ -305,7 +298,7 @@ static struct bkey *bch_extent_sort_fixup(struct btree_iter *iter, bch_cut_back(&START_KEY(i->k), tmp); bch_cut_front(i->k, top->k); - min_heap_sift_down(&iter->heap, 0, &callbacks, NULL); + heap_sift(iter, 0, bch_extent_sort_cmp); return tmp; } else { @@ -625,7 +618,7 @@ static bool bch_extent_merge(struct btree_keys *bk, } const struct btree_keys_ops bch_extent_keys_ops = { - .sort_cmp = new_bch_extent_sort_cmp, + .sort_cmp = bch_extent_sort_cmp, .sort_fixup = bch_extent_sort_fixup, .insert_fixup = bch_extent_insert_fixup, .key_invalid = bch_extent_invalid, diff --git a/drivers/md/bcache/movinggc.c b/drivers/md/bcache/movinggc.c index 45ca134cbf02..26a6a535ec32 100644 --- a/drivers/md/bcache/movinggc.c +++ b/drivers/md/bcache/movinggc.c @@ -182,19 +182,16 @@ err: if (!IS_ERR_OR_NULL(w->private)) closure_sync(&cl); } -static bool new_bucket_cmp(const void *l, const void *r, void __always_unused *args) +static bool bucket_cmp(struct bucket *l, struct bucket *r) { - struct bucket **_l = (struct bucket **)l; - struct bucket **_r = (struct bucket **)r; - - return GC_SECTORS_USED(*_l) >= GC_SECTORS_USED(*_r); + return GC_SECTORS_USED(l) < GC_SECTORS_USED(r); } static unsigned int bucket_heap_top(struct cache *ca) { struct bucket *b; - return (b = min_heap_peek(&ca->heap)[0]) ? GC_SECTORS_USED(b) : 0; + return (b = heap_peek(&ca->heap)) ? GC_SECTORS_USED(b) : 0; } void bch_moving_gc(struct cache_set *c) @@ -202,10 +199,6 @@ void bch_moving_gc(struct cache_set *c) struct cache *ca = c->cache; struct bucket *b; unsigned long sectors_to_move, reserve_sectors; - const struct min_heap_callbacks callbacks = { - .less = new_bucket_cmp, - .swp = NULL, - }; if (!c->copy_gc_enabled) return; @@ -216,7 +209,7 @@ void bch_moving_gc(struct cache_set *c) reserve_sectors = ca->sb.bucket_size * fifo_used(&ca->free[RESERVE_MOVINGGC]); - ca->heap.nr = 0; + ca->heap.used = 0; for_each_bucket(b, ca) { if (GC_MARK(b) == GC_MARK_METADATA || @@ -225,31 +218,25 @@ void bch_moving_gc(struct cache_set *c) atomic_read(&b->pin)) continue; - if (!min_heap_full(&ca->heap)) { + if (!heap_full(&ca->heap)) { sectors_to_move += GC_SECTORS_USED(b); - min_heap_push(&ca->heap, &b, &callbacks, NULL); - } else if (!new_bucket_cmp(&b, min_heap_peek(&ca->heap), ca)) { + heap_add(&ca->heap, b, bucket_cmp); + } else if (bucket_cmp(b, heap_peek(&ca->heap))) { sectors_to_move -= bucket_heap_top(ca); sectors_to_move += GC_SECTORS_USED(b); ca->heap.data[0] = b; - min_heap_sift_down(&ca->heap, 0, &callbacks, NULL); + heap_sift(&ca->heap, 0, bucket_cmp); } } while (sectors_to_move > reserve_sectors) { - if (ca->heap.nr) { - b = min_heap_peek(&ca->heap)[0]; - min_heap_pop(&ca->heap, &callbacks, NULL); - } + heap_pop(&ca->heap, b, bucket_cmp); sectors_to_move -= GC_SECTORS_USED(b); } - while (ca->heap.nr) { - b = min_heap_peek(&ca->heap)[0]; - min_heap_pop(&ca->heap, &callbacks, NULL); + while (heap_pop(&ca->heap, b, bucket_cmp)) SET_GC_MOVE(b, 1); - } mutex_unlock(&c->bucket_lock); diff --git a/drivers/md/bcache/stats.c b/drivers/md/bcache/stats.c index 68b02216033d..0056106495a7 100644 --- a/drivers/md/bcache/stats.c +++ b/drivers/md/bcache/stats.c @@ -123,7 +123,7 @@ void bch_cache_accounting_destroy(struct cache_accounting *acc) kobject_put(&acc->day.kobj); atomic_set(&acc->closing, 1); - if (del_timer_sync(&acc->timer)) + if (timer_delete_sync(&acc->timer)) closure_return(&acc->cl); } @@ -149,7 +149,7 @@ static void scale_stats(struct cache_stats *stats, unsigned long rescale_at) static void scale_accounting(struct timer_list *t) { - struct cache_accounting *acc = from_timer(acc, t, timer); + struct cache_accounting *acc = timer_container_of(acc, t, timer); #define move_stat(name) do { \ unsigned int t = atomic_xchg(&acc->collector.name, 0); \ diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c index e42f1400cea9..2ea490b9d370 100644 --- a/drivers/md/bcache/super.c +++ b/drivers/md/bcache/super.c @@ -293,8 +293,7 @@ static void __write_super(struct cache_sb *sb, struct cache_sb_disk *out, bio->bi_opf = REQ_OP_WRITE | REQ_SYNC | REQ_META; bio->bi_iter.bi_sector = SB_SECTOR; - __bio_add_page(bio, virt_to_page(out), SB_SIZE, - offset_in_page(out)); + bio_add_virt_nofail(bio, out, SB_SIZE); out->offset = cpu_to_le64(sb->offset); @@ -546,7 +545,8 @@ static struct uuid_entry *uuid_find(struct cache_set *c, const char *uuid) static struct uuid_entry *uuid_find_empty(struct cache_set *c) { - static const char zero_uuid[16] = "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"; + static const char zero_uuid[16] __nonstring = + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }; return uuid_find(c, zero_uuid); } @@ -1733,7 +1733,12 @@ static CLOSURE_CALLBACK(cache_set_flush) mutex_unlock(&b->write_lock); } - if (ca->alloc_thread) + /* + * If the register_cache_set() call to bch_cache_set_alloc() failed, + * ca has not been assigned a value and return error. + * So we need check ca is not NULL during bch_cache_set_unregister(). + */ + if (ca && ca->alloc_thread) kthread_stop(ca->alloc_thread); if (c->journal.cur) { @@ -1907,7 +1912,8 @@ struct cache_set *bch_cache_set_alloc(struct cache_sb *sb) INIT_LIST_HEAD(&c->btree_cache_freed); INIT_LIST_HEAD(&c->data_buckets); - iter_size = ((meta_bucket_pages(sb) * PAGE_SECTORS) / sb->block_size) * + iter_size = sizeof(struct btree_iter) + + ((meta_bucket_pages(sb) * PAGE_SECTORS) / sb->block_size) * sizeof(struct btree_iter_set); c->devices = kcalloc(c->nr_uuids, sizeof(void *), GFP_KERNEL); @@ -2233,15 +2239,47 @@ static int cache_alloc(struct cache *ca) bio_init(&ca->journal.bio, NULL, ca->journal.bio.bi_inline_vecs, 8, 0); /* - * when ca->sb.njournal_buckets is not zero, journal exists, - * and in bch_journal_replay(), tree node may split, - * so bucket of RESERVE_BTREE type is needed, - * the worst situation is all journal buckets are valid journal, - * and all the keys need to replay, - * so the number of RESERVE_BTREE type buckets should be as much - * as journal buckets + * When the cache disk is first registered, ca->sb.njournal_buckets + * is zero, and it is assigned in run_cache_set(). + * + * When ca->sb.njournal_buckets is not zero, journal exists, + * and in bch_journal_replay(), tree node may split. + * The worst situation is all journal buckets are valid journal, + * and all the keys need to replay, so the number of RESERVE_BTREE + * type buckets should be as much as journal buckets. + * + * If the number of RESERVE_BTREE type buckets is too few, the + * bch_allocator_thread() may hang up and unable to allocate + * bucket. The situation is roughly as follows: + * + * 1. In bch_data_insert_keys(), if the operation is not op->replace, + * it will call the bch_journal(), which increments the journal_ref + * counter. This counter is only decremented after bch_btree_insert + * completes. + * + * 2. When calling bch_btree_insert, if the btree needs to split, + * it will call btree_split() and btree_check_reserve() to check + * whether there are enough reserved buckets in the RESERVE_BTREE + * slot. If not enough, bcache_btree_root() will repeatedly retry. + * + * 3. Normally, the bch_allocator_thread is responsible for filling + * the reservation slots from the free_inc bucket list. When the + * free_inc bucket list is exhausted, the bch_allocator_thread + * will call invalidate_buckets() until free_inc is refilled. + * Then bch_allocator_thread calls bch_prio_write() once. and + * bch_prio_write() will call bch_journal_meta() and waits for + * the journal write to complete. + * + * 4. During journal_write, journal_write_unlocked() is be called. + * If journal full occurs, journal_reclaim() and btree_flush_write() + * will be called sequentially, then retry journal_write. + * + * 5. When 2 and 4 occur together, IO will hung up and cannot recover. + * + * Therefore, reserve more RESERVE_BTREE type buckets. */ - btree_buckets = ca->sb.njournal_buckets ?: 8; + btree_buckets = clamp_t(size_t, ca->sb.nbuckets >> 7, + 32, SB_JOURNAL_BUCKETS); free = roundup_pow_of_two(ca->sb.nbuckets) >> 10; if (!free) { ret = -EPERM; diff --git a/drivers/md/bcache/sysfs.c b/drivers/md/bcache/sysfs.c index e8f696cb58c0..826b14cae4e5 100644 --- a/drivers/md/bcache/sysfs.c +++ b/drivers/md/bcache/sysfs.c @@ -660,9 +660,7 @@ static unsigned int bch_root_usage(struct cache_set *c) unsigned int bytes = 0; struct bkey *k; struct btree *b; - struct btree_iter iter; - - min_heap_init(&iter.heap, NULL, MAX_BSETS); + struct btree_iter_stack iter; goto lock_root; diff --git a/drivers/md/bcache/util.h b/drivers/md/bcache/util.h index 539454d8e2d0..f61ab1bada6c 100644 --- a/drivers/md/bcache/util.h +++ b/drivers/md/bcache/util.h @@ -9,7 +9,6 @@ #include <linux/kernel.h> #include <linux/sched/clock.h> #include <linux/llist.h> -#include <linux/min_heap.h> #include <linux/ratelimit.h> #include <linux/vmalloc.h> #include <linux/workqueue.h> @@ -31,10 +30,16 @@ struct closure; #endif +#define DECLARE_HEAP(type, name) \ + struct { \ + size_t size, used; \ + type *data; \ + } name + #define init_heap(heap, _size, gfp) \ ({ \ size_t _bytes; \ - (heap)->nr = 0; \ + (heap)->used = 0; \ (heap)->size = (_size); \ _bytes = (heap)->size * sizeof(*(heap)->data); \ (heap)->data = kvmalloc(_bytes, (gfp) & GFP_KERNEL); \ @@ -47,6 +52,64 @@ do { \ (heap)->data = NULL; \ } while (0) +#define heap_swap(h, i, j) swap((h)->data[i], (h)->data[j]) + +#define heap_sift(h, i, cmp) \ +do { \ + size_t _r, _j = i; \ + \ + for (; _j * 2 + 1 < (h)->used; _j = _r) { \ + _r = _j * 2 + 1; \ + if (_r + 1 < (h)->used && \ + cmp((h)->data[_r], (h)->data[_r + 1])) \ + _r++; \ + \ + if (cmp((h)->data[_r], (h)->data[_j])) \ + break; \ + heap_swap(h, _r, _j); \ + } \ +} while (0) + +#define heap_sift_down(h, i, cmp) \ +do { \ + while (i) { \ + size_t p = (i - 1) / 2; \ + if (cmp((h)->data[i], (h)->data[p])) \ + break; \ + heap_swap(h, i, p); \ + i = p; \ + } \ +} while (0) + +#define heap_add(h, d, cmp) \ +({ \ + bool _r = !heap_full(h); \ + if (_r) { \ + size_t _i = (h)->used++; \ + (h)->data[_i] = d; \ + \ + heap_sift_down(h, _i, cmp); \ + heap_sift(h, _i, cmp); \ + } \ + _r; \ +}) + +#define heap_pop(h, d, cmp) \ +({ \ + bool _r = (h)->used; \ + if (_r) { \ + (d) = (h)->data[0]; \ + (h)->used--; \ + heap_swap(h, 0, (h)->used); \ + heap_sift(h, 0, cmp); \ + } \ + _r; \ +}) + +#define heap_peek(h) ((h)->used ? (h)->data[0] : NULL) + +#define heap_full(h) ((h)->used == (h)->size) + #define DECLARE_FIFO(type, name) \ struct { \ size_t front, back, size, mask; \ diff --git a/drivers/md/bcache/writeback.c b/drivers/md/bcache/writeback.c index 453efbbdc8ee..302e75f1fc4b 100644 --- a/drivers/md/bcache/writeback.c +++ b/drivers/md/bcache/writeback.c @@ -908,16 +908,15 @@ static int bch_dirty_init_thread(void *arg) struct dirty_init_thrd_info *info = arg; struct bch_dirty_init_state *state = info->state; struct cache_set *c = state->c; - struct btree_iter iter; + struct btree_iter_stack iter; struct bkey *k, *p; int cur_idx, prev_idx, skip_nr; k = p = NULL; prev_idx = 0; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - bch_btree_iter_init(&c->root->keys, &iter, NULL); - k = bch_btree_iter_next_filter(&iter, &c->root->keys, bch_ptr_bad); + bch_btree_iter_stack_init(&c->root->keys, &iter, NULL); + k = bch_btree_iter_next_filter(&iter.iter, &c->root->keys, bch_ptr_bad); BUG_ON(!k); p = k; @@ -931,7 +930,7 @@ static int bch_dirty_init_thread(void *arg) skip_nr = cur_idx - prev_idx; while (skip_nr) { - k = bch_btree_iter_next_filter(&iter, + k = bch_btree_iter_next_filter(&iter.iter, &c->root->keys, bch_ptr_bad); if (k) @@ -980,13 +979,11 @@ void bch_sectors_dirty_init(struct bcache_device *d) int i; struct btree *b = NULL; struct bkey *k = NULL; - struct btree_iter iter; + struct btree_iter_stack iter; struct sectors_dirty_init op; struct cache_set *c = d->c; struct bch_dirty_init_state state; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - retry_lock: b = c->root; rw_lock(0, b, b->level); diff --git a/drivers/md/dm-bufio.c b/drivers/md/dm-bufio.c index aab8240429b0..ff7595caf440 100644 --- a/drivers/md/dm-bufio.c +++ b/drivers/md/dm-bufio.c @@ -41,16 +41,6 @@ #define DM_BUFIO_LOW_WATERMARK_RATIO 16 /* - * Check buffer ages in this interval (seconds) - */ -#define DM_BUFIO_WORK_TIMER_SECS 30 - -/* - * Free buffers when they are older than this (seconds) - */ -#define DM_BUFIO_DEFAULT_AGE_SECS 300 - -/* * The nr of bytes of cached data to keep around. */ #define DM_BUFIO_DEFAULT_RETAIN_BYTES (256 * 1024) @@ -68,6 +58,8 @@ #define LIST_DIRTY 1 #define LIST_SIZE 2 +#define SCAN_RESCHED_CYCLE 16 + /*--------------------------------------------------------------*/ /* @@ -1055,10 +1047,8 @@ static unsigned long dm_bufio_cache_size_latch; static DEFINE_SPINLOCK(global_spinlock); -/* - * Buffers are freed after this timeout - */ -static unsigned int dm_bufio_max_age = DM_BUFIO_DEFAULT_AGE_SECS; +static unsigned int dm_bufio_max_age; /* No longer does anything */ + static unsigned long dm_bufio_retain_bytes = DM_BUFIO_DEFAULT_RETAIN_BYTES; static unsigned long dm_bufio_peak_allocated; @@ -1086,7 +1076,6 @@ static LIST_HEAD(dm_bufio_all_clients); static DEFINE_MUTEX(dm_bufio_clients_lock); static struct workqueue_struct *dm_bufio_wq; -static struct delayed_work dm_bufio_cleanup_old_work; static struct work_struct dm_bufio_replacement_work; @@ -1362,7 +1351,7 @@ static void use_bio(struct dm_buffer *b, enum req_op op, sector_t sector, ptr = (char *)b->data + offset; len = n_sectors << SECTOR_SHIFT; - __bio_add_page(bio, virt_to_page(ptr), len, offset_in_page(ptr)); + bio_add_virt_nofail(bio, ptr, len); submit_bio(bio); } @@ -2234,7 +2223,7 @@ int dm_bufio_issue_discard(struct dm_bufio_client *c, sector_t block, sector_t c } EXPORT_SYMBOL_GPL(dm_bufio_issue_discard); -static bool forget_buffer(struct dm_bufio_client *c, sector_t block) +static void forget_buffer(struct dm_bufio_client *c, sector_t block) { struct dm_buffer *b; @@ -2249,8 +2238,6 @@ static bool forget_buffer(struct dm_bufio_client *c, sector_t block) cache_put_and_wake(c, b); } } - - return b ? true : false; } /* @@ -2426,7 +2413,12 @@ static void __scan(struct dm_bufio_client *c) atomic_long_dec(&c->need_shrink); freed++; - cond_resched(); + + if (unlikely(freed % SCAN_RESCHED_CYCLE == 0)) { + dm_bufio_unlock(c); + cond_resched(); + dm_bufio_lock(c); + } } } } @@ -2675,130 +2667,6 @@ EXPORT_SYMBOL_GPL(dm_bufio_set_sector_offset); /*--------------------------------------------------------------*/ -static unsigned int get_max_age_hz(void) -{ - unsigned int max_age = READ_ONCE(dm_bufio_max_age); - - if (max_age > UINT_MAX / HZ) - max_age = UINT_MAX / HZ; - - return max_age * HZ; -} - -static bool older_than(struct dm_buffer *b, unsigned long age_hz) -{ - return time_after_eq(jiffies, READ_ONCE(b->last_accessed) + age_hz); -} - -struct evict_params { - gfp_t gfp; - unsigned long age_hz; - - /* - * This gets updated with the largest last_accessed (ie. most - * recently used) of the evicted buffers. It will not be reinitialised - * by __evict_many(), so you can use it across multiple invocations. - */ - unsigned long last_accessed; -}; - -/* - * We may not be able to evict this buffer if IO pending or the client - * is still using it. - * - * And if GFP_NOFS is used, we must not do any I/O because we hold - * dm_bufio_clients_lock and we would risk deadlock if the I/O gets - * rerouted to different bufio client. - */ -static enum evict_result select_for_evict(struct dm_buffer *b, void *context) -{ - struct evict_params *params = context; - - if (!(params->gfp & __GFP_FS) || - (static_branch_unlikely(&no_sleep_enabled) && b->c->no_sleep)) { - if (test_bit_acquire(B_READING, &b->state) || - test_bit(B_WRITING, &b->state) || - test_bit(B_DIRTY, &b->state)) - return ER_DONT_EVICT; - } - - return older_than(b, params->age_hz) ? ER_EVICT : ER_STOP; -} - -static unsigned long __evict_many(struct dm_bufio_client *c, - struct evict_params *params, - int list_mode, unsigned long max_count) -{ - unsigned long count; - unsigned long last_accessed; - struct dm_buffer *b; - - for (count = 0; count < max_count; count++) { - b = cache_evict(&c->cache, list_mode, select_for_evict, params); - if (!b) - break; - - last_accessed = READ_ONCE(b->last_accessed); - if (time_after_eq(params->last_accessed, last_accessed)) - params->last_accessed = last_accessed; - - __make_buffer_clean(b); - __free_buffer_wake(b); - - cond_resched(); - } - - return count; -} - -static void evict_old_buffers(struct dm_bufio_client *c, unsigned long age_hz) -{ - struct evict_params params = {.gfp = 0, .age_hz = age_hz, .last_accessed = 0}; - unsigned long retain = get_retain_buffers(c); - unsigned long count; - LIST_HEAD(write_list); - - dm_bufio_lock(c); - - __check_watermark(c, &write_list); - if (unlikely(!list_empty(&write_list))) { - dm_bufio_unlock(c); - __flush_write_list(&write_list); - dm_bufio_lock(c); - } - - count = cache_total(&c->cache); - if (count > retain) - __evict_many(c, ¶ms, LIST_CLEAN, count - retain); - - dm_bufio_unlock(c); -} - -static void cleanup_old_buffers(void) -{ - unsigned long max_age_hz = get_max_age_hz(); - struct dm_bufio_client *c; - - mutex_lock(&dm_bufio_clients_lock); - - __cache_size_refresh(); - - list_for_each_entry(c, &dm_bufio_all_clients, client_list) - evict_old_buffers(c, max_age_hz); - - mutex_unlock(&dm_bufio_clients_lock); -} - -static void work_fn(struct work_struct *w) -{ - cleanup_old_buffers(); - - queue_delayed_work(dm_bufio_wq, &dm_bufio_cleanup_old_work, - DM_BUFIO_WORK_TIMER_SECS * HZ); -} - -/*--------------------------------------------------------------*/ - /* * Global cleanup tries to evict the oldest buffers from across _all_ * the clients. It does this by repeatedly evicting a few buffers from @@ -2836,27 +2704,55 @@ static void __insert_client(struct dm_bufio_client *new_client) list_add_tail(&new_client->client_list, h); } +static enum evict_result select_for_evict(struct dm_buffer *b, void *context) +{ + /* In no-sleep mode, we cannot wait on IO. */ + if (static_branch_unlikely(&no_sleep_enabled) && b->c->no_sleep) { + if (test_bit_acquire(B_READING, &b->state) || + test_bit(B_WRITING, &b->state) || + test_bit(B_DIRTY, &b->state)) + return ER_DONT_EVICT; + } + return ER_EVICT; +} + static unsigned long __evict_a_few(unsigned long nr_buffers) { - unsigned long count; struct dm_bufio_client *c; - struct evict_params params = { - .gfp = GFP_KERNEL, - .age_hz = 0, - /* set to jiffies in case there are no buffers in this client */ - .last_accessed = jiffies - }; + unsigned long oldest_buffer = jiffies; + unsigned long last_accessed; + unsigned long count; + struct dm_buffer *b; c = __pop_client(); if (!c) return 0; dm_bufio_lock(c); - count = __evict_many(c, ¶ms, LIST_CLEAN, nr_buffers); + + for (count = 0; count < nr_buffers; count++) { + b = cache_evict(&c->cache, LIST_CLEAN, select_for_evict, NULL); + if (!b) + break; + + last_accessed = READ_ONCE(b->last_accessed); + if (time_after_eq(oldest_buffer, last_accessed)) + oldest_buffer = last_accessed; + + __make_buffer_clean(b); + __free_buffer_wake(b); + + if (need_resched()) { + dm_bufio_unlock(c); + cond_resched(); + dm_bufio_lock(c); + } + } + dm_bufio_unlock(c); if (count) - c->oldest_buffer = params.last_accessed; + c->oldest_buffer = oldest_buffer; __insert_client(c); return count; @@ -2939,10 +2835,7 @@ static int __init dm_bufio_init(void) if (!dm_bufio_wq) return -ENOMEM; - INIT_DELAYED_WORK(&dm_bufio_cleanup_old_work, work_fn); INIT_WORK(&dm_bufio_replacement_work, do_global_cleanup); - queue_delayed_work(dm_bufio_wq, &dm_bufio_cleanup_old_work, - DM_BUFIO_WORK_TIMER_SECS * HZ); return 0; } @@ -2954,7 +2847,6 @@ static void __exit dm_bufio_exit(void) { int bug = 0; - cancel_delayed_work_sync(&dm_bufio_cleanup_old_work); destroy_workqueue(dm_bufio_wq); if (dm_bufio_client_count) { @@ -2991,7 +2883,7 @@ module_param_named(max_cache_size_bytes, dm_bufio_cache_size, ulong, 0644); MODULE_PARM_DESC(max_cache_size_bytes, "Size of metadata cache"); module_param_named(max_age_seconds, dm_bufio_max_age, uint, 0644); -MODULE_PARM_DESC(max_age_seconds, "Max age of a buffer in seconds"); +MODULE_PARM_DESC(max_age_seconds, "No longer does anything"); module_param_named(retain_bytes, dm_bufio_retain_bytes, ulong, 0644); MODULE_PARM_DESC(retain_bytes, "Try to keep at least this many bytes cached in memory"); diff --git a/drivers/md/dm-cache-target.c b/drivers/md/dm-cache-target.c index 9cb797a561d6..a10d75a562db 100644 --- a/drivers/md/dm-cache-target.c +++ b/drivers/md/dm-cache-target.c @@ -406,6 +406,12 @@ struct cache { mempool_t migration_pool; struct bio_set bs; + + /* + * Cache_size entries. Set bits indicate blocks mapped beyond the + * target length, which are marked for invalidation. + */ + unsigned long *invalid_bitset; }; struct per_bio_data { @@ -1922,6 +1928,9 @@ static void __destroy(struct cache *cache) if (cache->discard_bitset) free_bitset(cache->discard_bitset); + if (cache->invalid_bitset) + free_bitset(cache->invalid_bitset); + if (cache->copier) dm_kcopyd_client_destroy(cache->copier); @@ -2510,6 +2519,13 @@ static int cache_create(struct cache_args *ca, struct cache **result) } clear_bitset(cache->discard_bitset, from_dblock(cache->discard_nr_blocks)); + cache->invalid_bitset = alloc_bitset(from_cblock(cache->cache_size)); + if (!cache->invalid_bitset) { + *error = "could not allocate bitset for invalid blocks"; + goto bad; + } + clear_bitset(cache->invalid_bitset, from_cblock(cache->cache_size)); + cache->copier = dm_kcopyd_client_create(&dm_kcopyd_throttle); if (IS_ERR(cache->copier)) { *error = "could not create kcopyd client"; @@ -2808,6 +2824,24 @@ static int load_mapping(void *context, dm_oblock_t oblock, dm_cblock_t cblock, return policy_load_mapping(cache->policy, oblock, cblock, dirty, hint, hint_valid); } +static int load_filtered_mapping(void *context, dm_oblock_t oblock, dm_cblock_t cblock, + bool dirty, uint32_t hint, bool hint_valid) +{ + struct cache *cache = context; + + if (from_oblock(oblock) >= from_oblock(cache->origin_blocks)) { + if (dirty) { + DMERR("%s: unable to shrink origin; cache block %u is dirty", + cache_device_name(cache), from_cblock(cblock)); + return -EFBIG; + } + set_bit(from_cblock(cblock), cache->invalid_bitset); + return 0; + } + + return load_mapping(context, oblock, cblock, dirty, hint, hint_valid); +} + /* * The discard block size in the on disk metadata is not * necessarily the same as we're currently using. So we have to @@ -2899,6 +2933,27 @@ static dm_cblock_t get_cache_dev_size(struct cache *cache) return to_cblock(size); } +static bool can_resume(struct cache *cache) +{ + /* + * Disallow retrying the resume operation for devices that failed the + * first resume attempt, as the failure leaves the policy object partially + * initialized. Retrying could trigger BUG_ON when loading cache mappings + * into the incomplete policy object. + */ + if (cache->sized && !cache->loaded_mappings) { + if (get_cache_mode(cache) != CM_WRITE) + DMERR("%s: unable to resume a failed-loaded cache, please check metadata.", + cache_device_name(cache)); + else + DMERR("%s: unable to resume cache due to missing proper cache table reload", + cache_device_name(cache)); + return false; + } + + return true; +} + static bool can_resize(struct cache *cache, dm_cblock_t new_size) { if (from_cblock(new_size) > from_cblock(cache->cache_size)) { @@ -2941,12 +2996,33 @@ static int resize_cache_dev(struct cache *cache, dm_cblock_t new_size) return 0; } +static int truncate_oblocks(struct cache *cache) +{ + uint32_t nr_blocks = from_cblock(cache->cache_size); + uint32_t i; + int r; + + for_each_set_bit(i, cache->invalid_bitset, nr_blocks) { + r = dm_cache_remove_mapping(cache->cmd, to_cblock(i)); + if (r) { + DMERR_LIMIT("%s: invalidation failed; couldn't update on disk metadata", + cache_device_name(cache)); + return r; + } + } + + return 0; +} + static int cache_preresume(struct dm_target *ti) { int r = 0; struct cache *cache = ti->private; dm_cblock_t csize = get_cache_dev_size(cache); + if (!can_resume(cache)) + return -EINVAL; + /* * Check to see if the cache has resized. */ @@ -2962,11 +3038,25 @@ static int cache_preresume(struct dm_target *ti) } if (!cache->loaded_mappings) { + /* + * The fast device could have been resized since the last + * failed preresume attempt. To be safe we start by a blank + * bitset for cache blocks. + */ + clear_bitset(cache->invalid_bitset, from_cblock(cache->cache_size)); + r = dm_cache_load_mappings(cache->cmd, cache->policy, - load_mapping, cache); + load_filtered_mapping, cache); if (r) { DMERR("%s: could not load cache mappings", cache_device_name(cache)); - metadata_operation_failed(cache, "dm_cache_load_mappings", r); + if (r != -EFBIG) + metadata_operation_failed(cache, "dm_cache_load_mappings", r); + return r; + } + + r = truncate_oblocks(cache); + if (r) { + metadata_operation_failed(cache, "dm_cache_remove_mapping", r); return r; } @@ -3426,7 +3516,7 @@ static void cache_io_hints(struct dm_target *ti, struct queue_limits *limits) static struct target_type cache_target = { .name = "cache", - .version = {2, 2, 0}, + .version = {2, 3, 0}, .module = THIS_MODULE, .ctr = cache_ctr, .dtr = cache_dtr, diff --git a/drivers/md/dm-core.h b/drivers/md/dm-core.h index 3637761f3585..c889332e533b 100644 --- a/drivers/md/dm-core.h +++ b/drivers/md/dm-core.h @@ -141,6 +141,7 @@ struct mapped_device { #ifdef CONFIG_BLK_DEV_ZONED unsigned int nr_zones; void *zone_revalidate_map; + struct task_struct *revalidate_map_task; #endif #ifdef CONFIG_IMA @@ -162,9 +163,6 @@ struct mapped_device { #define DMF_POST_SUSPENDING 8 #define DMF_EMULATE_ZONE_APPEND 9 -void disable_discard(struct mapped_device *md); -void disable_write_zeroes(struct mapped_device *md); - static inline sector_t dm_get_size(struct mapped_device *md) { return get_capacity(md->disk); diff --git a/drivers/md/dm-crypt.c b/drivers/md/dm-crypt.c index 02a2919f4e5a..17157c4216a5 100644 --- a/drivers/md/dm-crypt.c +++ b/drivers/md/dm-crypt.c @@ -17,6 +17,7 @@ #include <linux/bio.h> #include <linux/blkdev.h> #include <linux/blk-integrity.h> +#include <linux/crc32.h> #include <linux/mempool.h> #include <linux/slab.h> #include <linux/crypto.h> @@ -125,7 +126,6 @@ struct iv_lmk_private { #define TCW_WHITENING_SIZE 16 struct iv_tcw_private { - struct crypto_shash *crc32_tfm; u8 *iv_seed; u8 *whitening; }; @@ -517,7 +517,10 @@ static int crypt_iv_lmk_one(struct crypt_config *cc, u8 *iv, { struct iv_lmk_private *lmk = &cc->iv_gen_private.lmk; SHASH_DESC_ON_STACK(desc, lmk->hash_tfm); - struct md5_state md5state; + union { + struct md5_state md5state; + u8 state[CRYPTO_MD5_STATESIZE]; + } u; __le32 buf[4]; int i, r; @@ -548,13 +551,13 @@ static int crypt_iv_lmk_one(struct crypt_config *cc, u8 *iv, return r; /* No MD5 padding here */ - r = crypto_shash_export(desc, &md5state); + r = crypto_shash_export(desc, &u.md5state); if (r) return r; for (i = 0; i < MD5_HASH_WORDS; i++) - __cpu_to_le32s(&md5state.hash[i]); - memcpy(iv, &md5state.hash, cc->iv_size); + __cpu_to_le32s(&u.md5state.hash[i]); + memcpy(iv, &u.md5state.hash, cc->iv_size); return 0; } @@ -607,10 +610,6 @@ static void crypt_iv_tcw_dtr(struct crypt_config *cc) tcw->iv_seed = NULL; kfree_sensitive(tcw->whitening); tcw->whitening = NULL; - - if (tcw->crc32_tfm && !IS_ERR(tcw->crc32_tfm)) - crypto_free_shash(tcw->crc32_tfm); - tcw->crc32_tfm = NULL; } static int crypt_iv_tcw_ctr(struct crypt_config *cc, struct dm_target *ti, @@ -628,13 +627,6 @@ static int crypt_iv_tcw_ctr(struct crypt_config *cc, struct dm_target *ti, return -EINVAL; } - tcw->crc32_tfm = crypto_alloc_shash("crc32", 0, - CRYPTO_ALG_ALLOCATES_MEMORY); - if (IS_ERR(tcw->crc32_tfm)) { - ti->error = "Error initializing CRC32 in TCW"; - return PTR_ERR(tcw->crc32_tfm); - } - tcw->iv_seed = kzalloc(cc->iv_size, GFP_KERNEL); tcw->whitening = kzalloc(TCW_WHITENING_SIZE, GFP_KERNEL); if (!tcw->iv_seed || !tcw->whitening) { @@ -668,36 +660,28 @@ static int crypt_iv_tcw_wipe(struct crypt_config *cc) return 0; } -static int crypt_iv_tcw_whitening(struct crypt_config *cc, - struct dm_crypt_request *dmreq, - u8 *data) +static void crypt_iv_tcw_whitening(struct crypt_config *cc, + struct dm_crypt_request *dmreq, u8 *data) { struct iv_tcw_private *tcw = &cc->iv_gen_private.tcw; __le64 sector = cpu_to_le64(dmreq->iv_sector); u8 buf[TCW_WHITENING_SIZE]; - SHASH_DESC_ON_STACK(desc, tcw->crc32_tfm); - int i, r; + int i; /* xor whitening with sector number */ crypto_xor_cpy(buf, tcw->whitening, (u8 *)§or, 8); crypto_xor_cpy(&buf[8], tcw->whitening + 8, (u8 *)§or, 8); /* calculate crc32 for every 32bit part and xor it */ - desc->tfm = tcw->crc32_tfm; - for (i = 0; i < 4; i++) { - r = crypto_shash_digest(desc, &buf[i * 4], 4, &buf[i * 4]); - if (r) - goto out; - } + for (i = 0; i < 4; i++) + put_unaligned_le32(crc32(0, &buf[i * 4], 4), &buf[i * 4]); crypto_xor(&buf[0], &buf[12], 4); crypto_xor(&buf[4], &buf[8], 4); /* apply whitening (8 bytes) to whole sector */ for (i = 0; i < ((1 << SECTOR_SHIFT) / 8); i++) crypto_xor(data + i * 8, buf, 8); -out: memzero_explicit(buf, sizeof(buf)); - return r; } static int crypt_iv_tcw_gen(struct crypt_config *cc, u8 *iv, @@ -707,13 +691,12 @@ static int crypt_iv_tcw_gen(struct crypt_config *cc, u8 *iv, struct iv_tcw_private *tcw = &cc->iv_gen_private.tcw; __le64 sector = cpu_to_le64(dmreq->iv_sector); u8 *src; - int r = 0; /* Remove whitening from ciphertext */ if (bio_data_dir(dmreq->ctx->bio_in) != WRITE) { sg = crypt_get_sg_data(cc, dmreq->sg_in); src = kmap_local_page(sg_page(sg)); - r = crypt_iv_tcw_whitening(cc, dmreq, src + sg->offset); + crypt_iv_tcw_whitening(cc, dmreq, src + sg->offset); kunmap_local(src); } @@ -723,7 +706,7 @@ static int crypt_iv_tcw_gen(struct crypt_config *cc, u8 *iv, crypto_xor_cpy(&iv[8], tcw->iv_seed + 8, (u8 *)§or, cc->iv_size - 8); - return r; + return 0; } static int crypt_iv_tcw_post(struct crypt_config *cc, u8 *iv, @@ -731,7 +714,6 @@ static int crypt_iv_tcw_post(struct crypt_config *cc, u8 *iv, { struct scatterlist *sg; u8 *dst; - int r; if (bio_data_dir(dmreq->ctx->bio_in) != WRITE) return 0; @@ -739,10 +721,10 @@ static int crypt_iv_tcw_post(struct crypt_config *cc, u8 *iv, /* Apply whitening on ciphertext */ sg = crypt_get_sg_data(cc, dmreq->sg_out); dst = kmap_local_page(sg_page(sg)); - r = crypt_iv_tcw_whitening(cc, dmreq, dst + sg->offset); + crypt_iv_tcw_whitening(cc, dmreq, dst + sg->offset); kunmap_local(dst); - return r; + return 0; } static int crypt_iv_random_gen(struct crypt_config *cc, u8 *iv, diff --git a/drivers/md/dm-delay.c b/drivers/md/dm-delay.c index 08f6387620c1..4bb6553278c7 100644 --- a/drivers/md/dm-delay.c +++ b/drivers/md/dm-delay.c @@ -14,11 +14,14 @@ #include <linux/bio.h> #include <linux/slab.h> #include <linux/kthread.h> +#include <linux/delay.h> #include <linux/device-mapper.h> #define DM_MSG_PREFIX "delay" +#define SLEEP_SHIFT 3 + struct delay_class { struct dm_dev *dev; sector_t start; @@ -34,6 +37,7 @@ struct delay_c { struct work_struct flush_expired_bios; struct list_head delayed_bios; struct task_struct *worker; + unsigned int worker_sleep_us; bool may_delay; struct delay_class read; @@ -52,7 +56,7 @@ struct dm_delay_info { static void handle_delayed_timer(struct timer_list *t) { - struct delay_c *dc = from_timer(dc, t, delay_timer); + struct delay_c *dc = timer_container_of(dc, t, delay_timer); queue_work(dc->kdelayd_wq, &dc->flush_expired_bios); } @@ -136,6 +140,7 @@ static int flush_worker_fn(void *data) schedule(); } else { spin_unlock(&dc->delayed_bios_lock); + fsleep(dc->worker_sleep_us); cond_resched(); } } @@ -212,7 +217,7 @@ static int delay_ctr(struct dm_target *ti, unsigned int argc, char **argv) { struct delay_c *dc; int ret; - unsigned int max_delay; + unsigned int max_delay, min_delay; if (argc != 3 && argc != 6 && argc != 9) { ti->error = "Requires exactly 3, 6 or 9 arguments"; @@ -235,7 +240,7 @@ static int delay_ctr(struct dm_target *ti, unsigned int argc, char **argv) ret = delay_class_ctr(ti, &dc->read, argv); if (ret) goto bad; - max_delay = dc->read.delay; + min_delay = max_delay = dc->read.delay; if (argc == 3) { ret = delay_class_ctr(ti, &dc->write, argv); @@ -251,6 +256,7 @@ static int delay_ctr(struct dm_target *ti, unsigned int argc, char **argv) if (ret) goto bad; max_delay = max(max_delay, dc->write.delay); + min_delay = min_not_zero(min_delay, dc->write.delay); if (argc == 6) { ret = delay_class_ctr(ti, &dc->flush, argv + 3); @@ -263,9 +269,14 @@ static int delay_ctr(struct dm_target *ti, unsigned int argc, char **argv) if (ret) goto bad; max_delay = max(max_delay, dc->flush.delay); + min_delay = min_not_zero(min_delay, dc->flush.delay); out: if (max_delay < 50) { + if (min_delay >> SLEEP_SHIFT) + dc->worker_sleep_us = 1000; + else + dc->worker_sleep_us = (min_delay * 1000) >> SLEEP_SHIFT; /* * In case of small requested delays, use kthread instead of * timers and workqueue to achieve better latency. @@ -369,6 +380,21 @@ static int delay_map(struct dm_target *ti, struct bio *bio) return delay_bio(dc, c, bio); } +#ifdef CONFIG_BLK_DEV_ZONED +static int delay_report_zones(struct dm_target *ti, + struct dm_report_zones_args *args, unsigned int nr_zones) +{ + struct delay_c *dc = ti->private; + struct delay_class *c = &dc->read; + + return dm_report_zones(c->dev->bdev, c->start, + c->start + dm_target_offset(ti, args->next_sector), + args, nr_zones); +} +#else +#define delay_report_zones NULL +#endif + #define DMEMIT_DELAY_CLASS(c) \ DMEMIT("%s %llu %u", (c)->dev->name, (unsigned long long)(c)->start, (c)->delay) @@ -423,12 +449,13 @@ out: static struct target_type delay_target = { .name = "delay", - .version = {1, 4, 0}, - .features = DM_TARGET_PASSES_INTEGRITY, + .version = {1, 5, 0}, + .features = DM_TARGET_PASSES_INTEGRITY | DM_TARGET_ZONED_HM, .module = THIS_MODULE, .ctr = delay_ctr, .dtr = delay_dtr, .map = delay_map, + .report_zones = delay_report_zones, .presuspend = delay_presuspend, .resume = delay_resume, .status = delay_status, diff --git a/drivers/md/dm-dust.c b/drivers/md/dm-dust.c index 1a33820c9f46..e75310232bbf 100644 --- a/drivers/md/dm-dust.c +++ b/drivers/md/dm-dust.c @@ -534,7 +534,9 @@ static void dust_status(struct dm_target *ti, status_type_t type, } } -static int dust_prepare_ioctl(struct dm_target *ti, struct block_device **bdev) +static int dust_prepare_ioctl(struct dm_target *ti, struct block_device **bdev, + unsigned int cmd, unsigned long arg, + bool *forward) { struct dust_device *dd = ti->private; struct dm_dev *dev = dd->dev; diff --git a/drivers/md/dm-ebs-target.c b/drivers/md/dm-ebs-target.c index b19b0142a690..6abb31ca9662 100644 --- a/drivers/md/dm-ebs-target.c +++ b/drivers/md/dm-ebs-target.c @@ -415,7 +415,8 @@ static void ebs_status(struct dm_target *ti, status_type_t type, } } -static int ebs_prepare_ioctl(struct dm_target *ti, struct block_device **bdev) +static int ebs_prepare_ioctl(struct dm_target *ti, struct block_device **bdev, + unsigned int cmd, unsigned long arg, bool *forward) { struct ebs_c *ec = ti->private; struct dm_dev *dev = ec->dev; diff --git a/drivers/md/dm-flakey.c b/drivers/md/dm-flakey.c index b690905ab89f..cf17fd46e255 100644 --- a/drivers/md/dm-flakey.c +++ b/drivers/md/dm-flakey.c @@ -47,14 +47,15 @@ enum feature_flag_bits { }; struct per_bio_data { - bool bio_submitted; + bool bio_can_corrupt; + struct bvec_iter saved_iter; }; static int parse_features(struct dm_arg_set *as, struct flakey_c *fc, struct dm_target *ti) { - int r; - unsigned int argc; + int r = 0; + unsigned int argc = 0; const char *arg_name; static const struct dm_arg _args[] = { @@ -65,14 +66,13 @@ static int parse_features(struct dm_arg_set *as, struct flakey_c *fc, {0, PROBABILITY_BASE, "Invalid random corrupt argument"}, }; - /* No feature arguments supplied. */ - if (!as->argc) - return 0; - - r = dm_read_arg_group(_args, as, &argc, &ti->error); - if (r) + if (as->argc && (r = dm_read_arg_group(_args, as, &argc, &ti->error))) return r; + /* No feature arguments supplied. */ + if (!argc) + goto error_all_io; + while (argc) { arg_name = dm_shift_arg(as); argc--; @@ -128,8 +128,11 @@ static int parse_features(struct dm_arg_set *as, struct flakey_c *fc, * corrupt_bio_byte <Nth_byte> <direction> <value> <bio_flags> */ if (!strcasecmp(arg_name, "corrupt_bio_byte")) { - if (!argc) { - ti->error = "Feature corrupt_bio_byte requires parameters"; + if (fc->corrupt_bio_byte) { + ti->error = "Feature corrupt_bio_byte duplicated"; + return -EINVAL; + } else if (argc < 4) { + ti->error = "Feature corrupt_bio_byte requires 4 parameters"; return -EINVAL; } @@ -176,7 +179,10 @@ static int parse_features(struct dm_arg_set *as, struct flakey_c *fc, } if (!strcasecmp(arg_name, "random_read_corrupt")) { - if (!argc) { + if (fc->random_read_corrupt) { + ti->error = "Feature random_read_corrupt duplicated"; + return -EINVAL; + } else if (!argc) { ti->error = "Feature random_read_corrupt requires a parameter"; return -EINVAL; } @@ -189,7 +195,10 @@ static int parse_features(struct dm_arg_set *as, struct flakey_c *fc, } if (!strcasecmp(arg_name, "random_write_corrupt")) { - if (!argc) { + if (fc->random_write_corrupt) { + ti->error = "Feature random_write_corrupt duplicated"; + return -EINVAL; + } else if (!argc) { ti->error = "Feature random_write_corrupt requires a parameter"; return -EINVAL; } @@ -205,18 +214,28 @@ static int parse_features(struct dm_arg_set *as, struct flakey_c *fc, return -EINVAL; } - if (test_bit(DROP_WRITES, &fc->flags) && (fc->corrupt_bio_rw == WRITE)) { - ti->error = "drop_writes is incompatible with corrupt_bio_byte with the WRITE flag set"; + if (test_bit(DROP_WRITES, &fc->flags) && + ((fc->corrupt_bio_byte && fc->corrupt_bio_rw == WRITE) || + fc->random_write_corrupt)) { + ti->error = "drop_writes is incompatible with random_write_corrupt or corrupt_bio_byte with the WRITE flag set"; return -EINVAL; - } else if (test_bit(ERROR_WRITES, &fc->flags) && (fc->corrupt_bio_rw == WRITE)) { - ti->error = "error_writes is incompatible with corrupt_bio_byte with the WRITE flag set"; + } else if (test_bit(ERROR_WRITES, &fc->flags) && + ((fc->corrupt_bio_byte && fc->corrupt_bio_rw == WRITE) || + fc->random_write_corrupt)) { + ti->error = "error_writes is incompatible with random_write_corrupt or corrupt_bio_byte with the WRITE flag set"; + return -EINVAL; + } else if (test_bit(ERROR_READS, &fc->flags) && + ((fc->corrupt_bio_byte && fc->corrupt_bio_rw == READ) || + fc->random_read_corrupt)) { + ti->error = "error_reads is incompatible with random_read_corrupt or corrupt_bio_byte with the READ flag set"; return -EINVAL; } if (!fc->corrupt_bio_byte && !test_bit(ERROR_READS, &fc->flags) && !test_bit(DROP_WRITES, &fc->flags) && !test_bit(ERROR_WRITES, &fc->flags) && !fc->random_read_corrupt && !fc->random_write_corrupt) { +error_all_io: set_bit(ERROR_WRITES, &fc->flags); set_bit(ERROR_READS, &fc->flags); } @@ -278,7 +297,7 @@ static int flakey_ctr(struct dm_target *ti, unsigned int argc, char **argv) if (r) goto bad; - r = dm_read_arg(_args, &as, &fc->down_interval, &ti->error); + r = dm_read_arg(_args + 1, &as, &fc->down_interval, &ti->error); if (r) goto bad; @@ -339,7 +358,8 @@ static void flakey_map_bio(struct dm_target *ti, struct bio *bio) } static void corrupt_bio_common(struct bio *bio, unsigned int corrupt_bio_byte, - unsigned char corrupt_bio_value) + unsigned char corrupt_bio_value, + struct bvec_iter start) { struct bvec_iter iter; struct bio_vec bvec; @@ -348,7 +368,7 @@ static void corrupt_bio_common(struct bio *bio, unsigned int corrupt_bio_byte, * Overwrite the Nth byte of the bio's data, on whichever page * it falls. */ - bio_for_each_segment(bvec, bio, iter) { + __bio_for_each_segment(bvec, bio, iter, start) { if (bio_iter_len(bio, iter) > corrupt_bio_byte) { unsigned char *segment = bvec_kmap_local(&bvec); segment[corrupt_bio_byte] = corrupt_bio_value; @@ -357,36 +377,31 @@ static void corrupt_bio_common(struct bio *bio, unsigned int corrupt_bio_byte, "(rw=%c bi_opf=%u bi_sector=%llu size=%u)\n", bio, corrupt_bio_value, corrupt_bio_byte, (bio_data_dir(bio) == WRITE) ? 'w' : 'r', bio->bi_opf, - (unsigned long long)bio->bi_iter.bi_sector, - bio->bi_iter.bi_size); + (unsigned long long)start.bi_sector, + start.bi_size); break; } corrupt_bio_byte -= bio_iter_len(bio, iter); } } -static void corrupt_bio_data(struct bio *bio, struct flakey_c *fc) +static void corrupt_bio_data(struct bio *bio, struct flakey_c *fc, + struct bvec_iter start) { unsigned int corrupt_bio_byte = fc->corrupt_bio_byte - 1; - if (!bio_has_data(bio)) - return; - - corrupt_bio_common(bio, corrupt_bio_byte, fc->corrupt_bio_value); + corrupt_bio_common(bio, corrupt_bio_byte, fc->corrupt_bio_value, start); } -static void corrupt_bio_random(struct bio *bio) +static void corrupt_bio_random(struct bio *bio, struct bvec_iter start) { unsigned int corrupt_byte; unsigned char corrupt_value; - if (!bio_has_data(bio)) - return; - - corrupt_byte = get_random_u32() % bio->bi_iter.bi_size; + corrupt_byte = get_random_u32() % start.bi_size; corrupt_value = get_random_u8(); - corrupt_bio_common(bio, corrupt_byte, corrupt_value); + corrupt_bio_common(bio, corrupt_byte, corrupt_value, start); } static void clone_free(struct bio *clone) @@ -481,7 +496,7 @@ static int flakey_map(struct dm_target *ti, struct bio *bio) unsigned int elapsed; struct per_bio_data *pb = dm_per_bio_data(bio, sizeof(struct per_bio_data)); - pb->bio_submitted = false; + pb->bio_can_corrupt = false; if (op_is_zone_mgmt(bio_op(bio))) goto map_bio; @@ -490,14 +505,15 @@ static int flakey_map(struct dm_target *ti, struct bio *bio) elapsed = (jiffies - fc->start_time) / HZ; if (elapsed % (fc->up_interval + fc->down_interval) >= fc->up_interval) { bool corrupt_fixed, corrupt_random; - /* - * Flag this bio as submitted while down. - */ - pb->bio_submitted = true; + + if (bio_has_data(bio)) { + pb->bio_can_corrupt = true; + pb->saved_iter = bio->bi_iter; + } /* - * Error reads if neither corrupt_bio_byte or drop_writes or error_writes are set. - * Otherwise, flakey_end_io() will decide if the reads should be modified. + * If ERROR_READS isn't set flakey_end_io() will decide if the + * reads should be modified. */ if (bio_data_dir(bio) == READ) { if (test_bit(ERROR_READS, &fc->flags)) @@ -516,6 +532,8 @@ static int flakey_map(struct dm_target *ti, struct bio *bio) return DM_MAPIO_SUBMITTED; } + if (!pb->bio_can_corrupt) + goto map_bio; /* * Corrupt matching writes. */ @@ -535,9 +553,11 @@ static int flakey_map(struct dm_target *ti, struct bio *bio) struct bio *clone = clone_bio(ti, fc, bio); if (clone) { if (corrupt_fixed) - corrupt_bio_data(clone, fc); + corrupt_bio_data(clone, fc, + clone->bi_iter); if (corrupt_random) - corrupt_bio_random(clone); + corrupt_bio_random(clone, + clone->bi_iter); submit_bio(clone); return DM_MAPIO_SUBMITTED; } @@ -559,28 +579,21 @@ static int flakey_end_io(struct dm_target *ti, struct bio *bio, if (op_is_zone_mgmt(bio_op(bio))) return DM_ENDIO_DONE; - if (!*error && pb->bio_submitted && (bio_data_dir(bio) == READ)) { + if (!*error && pb->bio_can_corrupt && (bio_data_dir(bio) == READ)) { if (fc->corrupt_bio_byte) { if ((fc->corrupt_bio_rw == READ) && all_corrupt_bio_flags_match(bio, fc)) { /* * Corrupt successful matching READs while in down state. */ - corrupt_bio_data(bio, fc); + corrupt_bio_data(bio, fc, pb->saved_iter); } } if (fc->random_read_corrupt) { u64 rnd = get_random_u64(); u32 rem = do_div(rnd, PROBABILITY_BASE); if (rem < fc->random_read_corrupt) - corrupt_bio_random(bio); - } - if (test_bit(ERROR_READS, &fc->flags)) { - /* - * Error read during the down_interval if drop_writes - * and error_writes were not configured. - */ - *error = BLK_STS_IOERR; + corrupt_bio_random(bio, pb->saved_iter); } } @@ -638,7 +651,9 @@ static void flakey_status(struct dm_target *ti, status_type_t type, } } -static int flakey_prepare_ioctl(struct dm_target *ti, struct block_device **bdev) +static int flakey_prepare_ioctl(struct dm_target *ti, struct block_device **bdev, + unsigned int cmd, unsigned long arg, + bool *forward) { struct flakey_c *fc = ti->private; diff --git a/drivers/md/dm-integrity.c b/drivers/md/dm-integrity.c index 65ab609ac0cb..4395657fa583 100644 --- a/drivers/md/dm-integrity.c +++ b/drivers/md/dm-integrity.c @@ -1540,7 +1540,8 @@ static void sleep_on_endio_wait(struct dm_integrity_c *ic) static void autocommit_fn(struct timer_list *t) { - struct dm_integrity_c *ic = from_timer(ic, t, autocommit_timer); + struct dm_integrity_c *ic = timer_container_of(ic, t, + autocommit_timer); if (likely(!dm_integrity_failed(ic))) queue_work(ic->commit_wq, &ic->commit_work); @@ -2557,14 +2558,8 @@ static void dm_integrity_inline_recheck(struct work_struct *w) char *mem; outgoing_bio = bio_alloc_bioset(ic->dev->bdev, 1, REQ_OP_READ, GFP_NOIO, &ic->recheck_bios); - - r = bio_add_page(outgoing_bio, virt_to_page(outgoing_data), ic->sectors_per_block << SECTOR_SHIFT, 0); - if (unlikely(r != (ic->sectors_per_block << SECTOR_SHIFT))) { - bio_put(outgoing_bio); - bio->bi_status = BLK_STS_RESOURCE; - bio_endio(bio); - return; - } + bio_add_virt_nofail(outgoing_bio, outgoing_data, + ic->sectors_per_block << SECTOR_SHIFT); bip = bio_integrity_alloc(outgoing_bio, GFP_NOIO, 1); if (IS_ERR(bip)) { @@ -2707,7 +2702,7 @@ static void integrity_commit(struct work_struct *w) unsigned int i, j, n; struct bio *flushes; - del_timer(&ic->autocommit_timer); + timer_delete(&ic->autocommit_timer); if (ic->mode == 'I') return; @@ -3211,7 +3206,8 @@ next_chunk: bio = bio_alloc_bioset(ic->dev->bdev, 1, REQ_OP_READ, GFP_NOIO, &ic->recalc_bios); bio->bi_iter.bi_sector = ic->start + SB_SECTORS + range.logical_sector; - __bio_add_page(bio, virt_to_page(recalc_buffer), range.n_sectors << SECTOR_SHIFT, offset_in_page(recalc_buffer)); + bio_add_virt_nofail(bio, recalc_buffer, + range.n_sectors << SECTOR_SHIFT); r = submit_bio_wait(bio); bio_put(bio); if (unlikely(r)) { @@ -3228,7 +3224,8 @@ next_chunk: bio = bio_alloc_bioset(ic->dev->bdev, 1, REQ_OP_WRITE, GFP_NOIO, &ic->recalc_bios); bio->bi_iter.bi_sector = ic->start + SB_SECTORS + range.logical_sector; - __bio_add_page(bio, virt_to_page(recalc_buffer), range.n_sectors << SECTOR_SHIFT, offset_in_page(recalc_buffer)); + bio_add_virt_nofail(bio, recalc_buffer, + range.n_sectors << SECTOR_SHIFT); bip = bio_integrity_alloc(bio, GFP_NOIO, 1); if (unlikely(IS_ERR(bip))) { @@ -3606,7 +3603,7 @@ static void dm_integrity_postsuspend(struct dm_target *ti) WARN_ON(unregister_reboot_notifier(&ic->reboot_notifier)); - del_timer_sync(&ic->autocommit_timer); + timer_delete_sync(&ic->autocommit_timer); if (ic->recalc_wq) drain_workqueue(ic->recalc_wq); @@ -4810,23 +4807,11 @@ static int dm_integrity_ctr(struct dm_target *ti, unsigned int argc, char **argv ti->error = "Cannot allocate bio set"; goto bad; } - r = bioset_integrity_create(&ic->recheck_bios, RECHECK_POOL_SIZE); - if (r) { - ti->error = "Cannot allocate bio integrity set"; - r = -ENOMEM; - goto bad; - } r = bioset_init(&ic->recalc_bios, 1, 0, BIOSET_NEED_BVECS); if (r) { ti->error = "Cannot allocate bio set"; goto bad; } - r = bioset_integrity_create(&ic->recalc_bios, 1); - if (r) { - ti->error = "Cannot allocate bio integrity set"; - r = -ENOMEM; - goto bad; - } } ic->metadata_wq = alloc_workqueue("dm-integrity-metadata", @@ -5176,7 +5161,7 @@ static void dm_integrity_dtr(struct dm_target *ti) BUG_ON(!RB_EMPTY_ROOT(&ic->in_progress)); BUG_ON(!list_empty(&ic->wait_list)); - if (ic->mode == 'B') + if (ic->mode == 'B' && ic->bitmap_flush_work.work.func) cancel_delayed_work_sync(&ic->bitmap_flush_work); if (ic->metadata_wq) destroy_workqueue(ic->metadata_wq); diff --git a/drivers/md/dm-ioctl.c b/drivers/md/dm-ioctl.c index d42eac944eb5..4165fef4c170 100644 --- a/drivers/md/dm-ioctl.c +++ b/drivers/md/dm-ioctl.c @@ -1885,6 +1885,7 @@ static ioctl_fn lookup_ioctl(unsigned int cmd, int *ioctl_flags) {DM_DEV_SET_GEOMETRY_CMD, 0, dev_set_geometry}, {DM_DEV_ARM_POLL_CMD, IOCTL_FLAGS_NO_PARAMS, dev_arm_poll}, {DM_GET_TARGET_VERSION_CMD, 0, get_target_version}, + {DM_MPATH_PROBE_PATHS_CMD, 0, NULL}, /* block device ioctl */ }; if (unlikely(cmd >= ARRAY_SIZE(_ioctls))) diff --git a/drivers/md/dm-linear.c b/drivers/md/dm-linear.c index 66318aba4bdb..15538ec58f8e 100644 --- a/drivers/md/dm-linear.c +++ b/drivers/md/dm-linear.c @@ -119,7 +119,9 @@ static void linear_status(struct dm_target *ti, status_type_t type, } } -static int linear_prepare_ioctl(struct dm_target *ti, struct block_device **bdev) +static int linear_prepare_ioctl(struct dm_target *ti, struct block_device **bdev, + unsigned int cmd, unsigned long arg, + bool *forward) { struct linear_c *lc = ti->private; struct dm_dev *dev = lc->dev; diff --git a/drivers/md/dm-log-writes.c b/drivers/md/dm-log-writes.c index 8d7df8303d0a..d484e8e1d48a 100644 --- a/drivers/md/dm-log-writes.c +++ b/drivers/md/dm-log-writes.c @@ -818,7 +818,9 @@ static void log_writes_status(struct dm_target *ti, status_type_t type, } static int log_writes_prepare_ioctl(struct dm_target *ti, - struct block_device **bdev) + struct block_device **bdev, + unsigned int cmd, unsigned long arg, + bool *forward) { struct log_writes_c *lc = ti->private; struct dm_dev *dev = lc->dev; diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c index 637977acc3dc..aaf4a0a4b0eb 100644 --- a/drivers/md/dm-mpath.c +++ b/drivers/md/dm-mpath.c @@ -79,6 +79,7 @@ struct multipath { struct pgpath *current_pgpath; struct priority_group *current_pg; struct priority_group *next_pg; /* Switch to this PG if set */ + struct priority_group *last_probed_pg; atomic_t nr_valid_paths; /* Total number of usable paths */ unsigned int nr_priority_groups; @@ -87,6 +88,7 @@ struct multipath { const char *hw_handler_name; char *hw_handler_params; wait_queue_head_t pg_init_wait; /* Wait for pg_init completion */ + wait_queue_head_t probe_wait; /* Wait for probing paths */ unsigned int pg_init_retries; /* Number of times to retry pg_init */ unsigned int pg_init_delay_msecs; /* Number of msecs before pg_init retry */ atomic_t pg_init_in_progress; /* Only one pg_init allowed at once */ @@ -100,6 +102,7 @@ struct multipath { struct bio_list queued_bios; struct timer_list nopath_timer; /* Timeout for queue_if_no_path */ + bool is_suspending; }; /* @@ -132,6 +135,8 @@ static void queue_if_no_path_timeout_work(struct timer_list *t); #define MPATHF_PG_INIT_DISABLED 4 /* pg_init is not currently allowed */ #define MPATHF_PG_INIT_REQUIRED 5 /* pg_init needs calling? */ #define MPATHF_PG_INIT_DELAY_RETRY 6 /* Delay pg_init retry? */ +#define MPATHF_DELAY_PG_SWITCH 7 /* Delay switching pg if it still has paths */ +#define MPATHF_NEED_PG_SWITCH 8 /* Need to switch pgs after the delay has ended */ static bool mpath_double_check_test_bit(int MPATHF_bit, struct multipath *m) { @@ -254,6 +259,7 @@ static int alloc_multipath_stage2(struct dm_target *ti, struct multipath *m) atomic_set(&m->pg_init_count, 0); m->pg_init_delay_msecs = DM_PG_INIT_DELAY_DEFAULT; init_waitqueue_head(&m->pg_init_wait); + init_waitqueue_head(&m->probe_wait); return 0; } @@ -413,13 +419,21 @@ static struct pgpath *choose_pgpath(struct multipath *m, size_t nr_bytes) goto failed; } + /* Don't change PG until it has no remaining paths */ + pg = READ_ONCE(m->current_pg); + if (pg) { + pgpath = choose_path_in_pg(m, pg, nr_bytes); + if (!IS_ERR_OR_NULL(pgpath)) + return pgpath; + } + /* Were we instructed to switch PG? */ if (READ_ONCE(m->next_pg)) { spin_lock_irqsave(&m->lock, flags); pg = m->next_pg; if (!pg) { spin_unlock_irqrestore(&m->lock, flags); - goto check_current_pg; + goto check_all_pgs; } m->next_pg = NULL; spin_unlock_irqrestore(&m->lock, flags); @@ -427,16 +441,7 @@ static struct pgpath *choose_pgpath(struct multipath *m, size_t nr_bytes) if (!IS_ERR_OR_NULL(pgpath)) return pgpath; } - - /* Don't change PG until it has no remaining paths */ -check_current_pg: - pg = READ_ONCE(m->current_pg); - if (pg) { - pgpath = choose_path_in_pg(m, pg, nr_bytes); - if (!IS_ERR_OR_NULL(pgpath)) - return pgpath; - } - +check_all_pgs: /* * Loop through priority groups until we find a valid path. * First time we skip PGs marked 'bypassed'. @@ -612,7 +617,6 @@ static void multipath_queue_bio(struct multipath *m, struct bio *bio) static struct pgpath *__map_bio(struct multipath *m, struct bio *bio) { struct pgpath *pgpath; - unsigned long flags; /* Do we need to select a new pgpath? */ pgpath = READ_ONCE(m->current_pgpath); @@ -620,12 +624,12 @@ static struct pgpath *__map_bio(struct multipath *m, struct bio *bio) pgpath = choose_pgpath(m, bio->bi_iter.bi_size); if (!pgpath) { - spin_lock_irqsave(&m->lock, flags); + spin_lock_irq(&m->lock); if (test_bit(MPATHF_QUEUE_IF_NO_PATH, &m->flags)) { __multipath_queue_bio(m, bio); pgpath = ERR_PTR(-EAGAIN); } - spin_unlock_irqrestore(&m->lock, flags); + spin_unlock_irq(&m->lock); } else if (mpath_double_check_test_bit(MPATHF_QUEUE_IO, m) || mpath_double_check_test_bit(MPATHF_PG_INIT_REQUIRED, m)) { @@ -688,7 +692,6 @@ static void process_queued_io_list(struct multipath *m) static void process_queued_bios(struct work_struct *work) { int r; - unsigned long flags; struct bio *bio; struct bio_list bios; struct blk_plug plug; @@ -697,16 +700,16 @@ static void process_queued_bios(struct work_struct *work) bio_list_init(&bios); - spin_lock_irqsave(&m->lock, flags); + spin_lock_irq(&m->lock); if (bio_list_empty(&m->queued_bios)) { - spin_unlock_irqrestore(&m->lock, flags); + spin_unlock_irq(&m->lock); return; } bio_list_merge_init(&bios, &m->queued_bios); - spin_unlock_irqrestore(&m->lock, flags); + spin_unlock_irq(&m->lock); blk_start_plug(&plug); while ((bio = bio_list_pop(&bios))) { @@ -787,7 +790,7 @@ static int queue_if_no_path(struct multipath *m, bool f_queue_if_no_path, */ static void queue_if_no_path_timeout_work(struct timer_list *t) { - struct multipath *m = from_timer(m, t, nopath_timer); + struct multipath *m = timer_container_of(m, t, nopath_timer); DMWARN("queue_if_no_path timeout on %s, failing queued IO", dm_table_device_name(m->ti->table)); @@ -815,7 +818,7 @@ static void enable_nopath_timeout(struct multipath *m) static void disable_nopath_timeout(struct multipath *m) { - del_timer_sync(&m->nopath_timer); + timer_delete_sync(&m->nopath_timer); } /* @@ -1190,7 +1193,6 @@ static int multipath_ctr(struct dm_target *ti, unsigned int argc, char **argv) struct dm_arg_set as; unsigned int pg_count = 0; unsigned int next_pg_num; - unsigned long flags; as.argc = argc; as.argv = argv; @@ -1255,9 +1257,9 @@ static int multipath_ctr(struct dm_target *ti, unsigned int argc, char **argv) goto bad; } - spin_lock_irqsave(&m->lock, flags); + spin_lock_irq(&m->lock); enable_nopath_timeout(m); - spin_unlock_irqrestore(&m->lock, flags); + spin_unlock_irq(&m->lock); ti->num_flush_bios = 1; ti->num_discard_bios = 1; @@ -1292,23 +1294,21 @@ static void multipath_wait_for_pg_init_completion(struct multipath *m) static void flush_multipath_work(struct multipath *m) { if (m->hw_handler_name) { - unsigned long flags; - if (!atomic_read(&m->pg_init_in_progress)) goto skip; - spin_lock_irqsave(&m->lock, flags); + spin_lock_irq(&m->lock); if (atomic_read(&m->pg_init_in_progress) && !test_and_set_bit(MPATHF_PG_INIT_DISABLED, &m->flags)) { - spin_unlock_irqrestore(&m->lock, flags); + spin_unlock_irq(&m->lock); flush_workqueue(kmpath_handlerd); multipath_wait_for_pg_init_completion(m); - spin_lock_irqsave(&m->lock, flags); + spin_lock_irq(&m->lock); clear_bit(MPATHF_PG_INIT_DISABLED, &m->flags); } - spin_unlock_irqrestore(&m->lock, flags); + spin_unlock_irq(&m->lock); } skip: if (m->queue_mode == DM_TYPE_BIO_BASED) @@ -1370,11 +1370,10 @@ out: static int reinstate_path(struct pgpath *pgpath) { int r = 0, run_queue = 0; - unsigned long flags; struct multipath *m = pgpath->pg->m; unsigned int nr_valid_paths; - spin_lock_irqsave(&m->lock, flags); + spin_lock_irq(&m->lock); if (pgpath->is_active) goto out; @@ -1404,7 +1403,7 @@ static int reinstate_path(struct pgpath *pgpath) schedule_work(&m->trigger_event); out: - spin_unlock_irqrestore(&m->lock, flags); + spin_unlock_irq(&m->lock); if (run_queue) { dm_table_run_md_queue_async(m->ti->table); process_queued_io_list(m); @@ -1439,15 +1438,19 @@ static int action_dev(struct multipath *m, dev_t dev, action_fn action) * Temporarily try to avoid having to use the specified PG */ static void bypass_pg(struct multipath *m, struct priority_group *pg, - bool bypassed) + bool bypassed, bool can_be_delayed) { unsigned long flags; spin_lock_irqsave(&m->lock, flags); pg->bypassed = bypassed; - m->current_pgpath = NULL; - m->current_pg = NULL; + if (can_be_delayed && test_bit(MPATHF_DELAY_PG_SWITCH, &m->flags)) + set_bit(MPATHF_NEED_PG_SWITCH, &m->flags); + else { + m->current_pgpath = NULL; + m->current_pg = NULL; + } spin_unlock_irqrestore(&m->lock, flags); @@ -1461,7 +1464,6 @@ static int switch_pg_num(struct multipath *m, const char *pgstr) { struct priority_group *pg; unsigned int pgnum; - unsigned long flags; char dummy; if (!pgstr || (sscanf(pgstr, "%u%c", &pgnum, &dummy) != 1) || !pgnum || @@ -1470,17 +1472,21 @@ static int switch_pg_num(struct multipath *m, const char *pgstr) return -EINVAL; } - spin_lock_irqsave(&m->lock, flags); + spin_lock_irq(&m->lock); list_for_each_entry(pg, &m->priority_groups, list) { pg->bypassed = false; if (--pgnum) continue; - m->current_pgpath = NULL; - m->current_pg = NULL; + if (test_bit(MPATHF_DELAY_PG_SWITCH, &m->flags)) + set_bit(MPATHF_NEED_PG_SWITCH, &m->flags); + else { + m->current_pgpath = NULL; + m->current_pg = NULL; + } m->next_pg = pg; } - spin_unlock_irqrestore(&m->lock, flags); + spin_unlock_irq(&m->lock); schedule_work(&m->trigger_event); return 0; @@ -1507,7 +1513,7 @@ static int bypass_pg_num(struct multipath *m, const char *pgstr, bool bypassed) break; } - bypass_pg(m, pg, bypassed); + bypass_pg(m, pg, bypassed, true); return 0; } @@ -1561,7 +1567,7 @@ static void pg_init_done(void *data, int errors) * Probably doing something like FW upgrade on the * controller so try the other pg. */ - bypass_pg(m, pg, true); + bypass_pg(m, pg, true, false); break; case SCSI_DH_RETRY: /* Wait before retrying. */ @@ -1742,6 +1748,9 @@ static void multipath_presuspend(struct dm_target *ti) { struct multipath *m = ti->private; + spin_lock_irq(&m->lock); + m->is_suspending = true; + spin_unlock_irq(&m->lock); /* FIXME: bio-based shouldn't need to always disable queue_if_no_path */ if (m->queue_mode == DM_TYPE_BIO_BASED || !dm_noflush_suspending(m->ti)) queue_if_no_path(m, false, true, __func__); @@ -1762,9 +1771,9 @@ static void multipath_postsuspend(struct dm_target *ti) static void multipath_resume(struct dm_target *ti) { struct multipath *m = ti->private; - unsigned long flags; - spin_lock_irqsave(&m->lock, flags); + spin_lock_irq(&m->lock); + m->is_suspending = false; if (test_bit(MPATHF_SAVED_QUEUE_IF_NO_PATH, &m->flags)) { set_bit(MPATHF_QUEUE_IF_NO_PATH, &m->flags); clear_bit(MPATHF_SAVED_QUEUE_IF_NO_PATH, &m->flags); @@ -1775,7 +1784,7 @@ static void multipath_resume(struct dm_target *ti) test_bit(MPATHF_QUEUE_IF_NO_PATH, &m->flags), test_bit(MPATHF_SAVED_QUEUE_IF_NO_PATH, &m->flags)); - spin_unlock_irqrestore(&m->lock, flags); + spin_unlock_irq(&m->lock); } /* @@ -1798,14 +1807,13 @@ static void multipath_status(struct dm_target *ti, status_type_t type, unsigned int status_flags, char *result, unsigned int maxlen) { int sz = 0, pg_counter, pgpath_counter; - unsigned long flags; struct multipath *m = ti->private; struct priority_group *pg; struct pgpath *p; unsigned int pg_num; char state; - spin_lock_irqsave(&m->lock, flags); + spin_lock_irq(&m->lock); /* Features */ if (type == STATUSTYPE_INFO) @@ -1845,10 +1853,10 @@ static void multipath_status(struct dm_target *ti, status_type_t type, DMEMIT("%u ", m->nr_priority_groups); - if (m->next_pg) - pg_num = m->next_pg->pg_num; - else if (m->current_pg) + if (m->current_pg) pg_num = m->current_pg->pg_num; + else if (m->next_pg) + pg_num = m->next_pg->pg_num; else pg_num = (m->nr_priority_groups ? 1 : 0); @@ -1951,7 +1959,7 @@ static void multipath_status(struct dm_target *ti, status_type_t type, break; } - spin_unlock_irqrestore(&m->lock, flags); + spin_unlock_irq(&m->lock); } static int multipath_message(struct dm_target *ti, unsigned int argc, char **argv, @@ -1961,7 +1969,6 @@ static int multipath_message(struct dm_target *ti, unsigned int argc, char **arg dev_t dev; struct multipath *m = ti->private; action_fn action; - unsigned long flags; mutex_lock(&m->work_mutex); @@ -1973,9 +1980,9 @@ static int multipath_message(struct dm_target *ti, unsigned int argc, char **arg if (argc == 1) { if (!strcasecmp(argv[0], "queue_if_no_path")) { r = queue_if_no_path(m, true, false, __func__); - spin_lock_irqsave(&m->lock, flags); + spin_lock_irq(&m->lock); enable_nopath_timeout(m); - spin_unlock_irqrestore(&m->lock, flags); + spin_unlock_irq(&m->lock); goto out; } else if (!strcasecmp(argv[0], "fail_if_no_path")) { r = queue_if_no_path(m, false, false, __func__); @@ -2021,14 +2028,132 @@ out: return r; } +/* + * Perform a minimal read from the given path to find out whether the + * path still works. If a path error occurs, fail it. + */ +static int probe_path(struct pgpath *pgpath) +{ + struct block_device *bdev = pgpath->path.dev->bdev; + unsigned int read_size = bdev_logical_block_size(bdev); + struct page *page; + struct bio *bio; + blk_status_t status; + int r = 0; + + if (WARN_ON_ONCE(read_size > PAGE_SIZE)) + return -EINVAL; + + page = alloc_page(GFP_KERNEL); + if (!page) + return -ENOMEM; + + /* Perform a minimal read: Sector 0, length read_size */ + bio = bio_alloc(bdev, 1, REQ_OP_READ, GFP_KERNEL); + if (!bio) { + r = -ENOMEM; + goto out; + } + + bio->bi_iter.bi_sector = 0; + __bio_add_page(bio, page, read_size, 0); + submit_bio_wait(bio); + status = bio->bi_status; + bio_put(bio); + + if (status && blk_path_error(status)) + fail_path(pgpath); + +out: + __free_page(page); + return r; +} + +/* + * Probe all active paths in current_pg to find out whether they still work. + * Fail all paths that do not work. + * + * Return -ENOTCONN if no valid path is left (even outside of current_pg). We + * cannot probe paths in other pgs without switching current_pg, so if valid + * paths are only in different pgs, they may or may not work. Additionally + * we should not probe paths in a pathgroup that is in the process of + * Initializing. Userspace can submit a request and we'll switch and wait + * for the pathgroup to be initialized. If the request fails, it may need to + * probe again. + */ +static int probe_active_paths(struct multipath *m) +{ + struct pgpath *pgpath; + struct priority_group *pg = NULL; + int r = 0; + + spin_lock_irq(&m->lock); + if (test_bit(MPATHF_DELAY_PG_SWITCH, &m->flags)) { + wait_event_lock_irq(m->probe_wait, + !test_bit(MPATHF_DELAY_PG_SWITCH, &m->flags), + m->lock); + /* + * if we waited because a probe was already in progress, + * and it probed the current active pathgroup, don't + * reprobe. Just return the number of valid paths + */ + if (m->current_pg == m->last_probed_pg) + goto skip_probe; + } + if (!m->current_pg || m->is_suspending || + test_bit(MPATHF_QUEUE_IO, &m->flags)) + goto skip_probe; + set_bit(MPATHF_DELAY_PG_SWITCH, &m->flags); + pg = m->last_probed_pg = m->current_pg; + spin_unlock_irq(&m->lock); + + list_for_each_entry(pgpath, &pg->pgpaths, list) { + if (pg != READ_ONCE(m->current_pg) || + READ_ONCE(m->is_suspending)) + goto out; + if (!pgpath->is_active) + continue; + + r = probe_path(pgpath); + if (r < 0) + goto out; + } + +out: + spin_lock_irq(&m->lock); + clear_bit(MPATHF_DELAY_PG_SWITCH, &m->flags); + if (test_and_clear_bit(MPATHF_NEED_PG_SWITCH, &m->flags)) { + m->current_pgpath = NULL; + m->current_pg = NULL; + } +skip_probe: + if (r == 0 && !atomic_read(&m->nr_valid_paths)) + r = -ENOTCONN; + spin_unlock_irq(&m->lock); + if (pg) + wake_up(&m->probe_wait); + return r; +} + static int multipath_prepare_ioctl(struct dm_target *ti, - struct block_device **bdev) + struct block_device **bdev, + unsigned int cmd, unsigned long arg, + bool *forward) { struct multipath *m = ti->private; struct pgpath *pgpath; - unsigned long flags; int r; + if (_IOC_TYPE(cmd) == DM_IOCTL) { + *forward = false; + switch (cmd) { + case DM_MPATH_PROBE_PATHS: + return probe_active_paths(m); + default: + return -ENOTTY; + } + } + pgpath = READ_ONCE(m->current_pgpath); if (!pgpath || !mpath_double_check_test_bit(MPATHF_QUEUE_IO, m)) pgpath = choose_pgpath(m, 0); @@ -2044,10 +2169,10 @@ static int multipath_prepare_ioctl(struct dm_target *ti, } else { /* No path is available */ r = -EIO; - spin_lock_irqsave(&m->lock, flags); + spin_lock_irq(&m->lock); if (test_bit(MPATHF_QUEUE_IF_NO_PATH, &m->flags)) r = -ENOTCONN; - spin_unlock_irqrestore(&m->lock, flags); + spin_unlock_irq(&m->lock); } if (r == -ENOTCONN) { @@ -2055,10 +2180,10 @@ static int multipath_prepare_ioctl(struct dm_target *ti, /* Path status changed, redo selection */ (void) choose_pgpath(m, 0); } - spin_lock_irqsave(&m->lock, flags); + spin_lock_irq(&m->lock); if (test_bit(MPATHF_PG_INIT_REQUIRED, &m->flags)) (void) __pg_init_all_paths(m); - spin_unlock_irqrestore(&m->lock, flags); + spin_unlock_irq(&m->lock); dm_table_run_md_queue_async(m->ti->table); process_queued_io_list(m); } @@ -2180,7 +2305,7 @@ static int multipath_busy(struct dm_target *ti) */ static struct target_type multipath_target = { .name = "multipath", - .version = {1, 14, 0}, + .version = {1, 15, 0}, .features = DM_TARGET_SINGLETON | DM_TARGET_IMMUTABLE | DM_TARGET_PASSES_INTEGRITY, .module = THIS_MODULE, diff --git a/drivers/md/dm-ps-historical-service-time.c b/drivers/md/dm-ps-historical-service-time.c index b49e10d76d03..2c8626a83de4 100644 --- a/drivers/md/dm-ps-historical-service-time.c +++ b/drivers/md/dm-ps-historical-service-time.c @@ -541,8 +541,10 @@ static int __init dm_hst_init(void) { int r = dm_register_path_selector(&hst_ps); - if (r < 0) + if (r < 0) { DMERR("register failed %d", r); + return r; + } DMINFO("version " HST_VERSION " loaded"); diff --git a/drivers/md/dm-ps-queue-length.c b/drivers/md/dm-ps-queue-length.c index e305f05ad1e5..eb543e6431e0 100644 --- a/drivers/md/dm-ps-queue-length.c +++ b/drivers/md/dm-ps-queue-length.c @@ -260,8 +260,10 @@ static int __init dm_ql_init(void) { int r = dm_register_path_selector(&ql_ps); - if (r < 0) + if (r < 0) { DMERR("register failed %d", r); + return r; + } DMINFO("version " QL_VERSION " loaded"); diff --git a/drivers/md/dm-ps-round-robin.c b/drivers/md/dm-ps-round-robin.c index d1745b123dc1..66a15ac0c22c 100644 --- a/drivers/md/dm-ps-round-robin.c +++ b/drivers/md/dm-ps-round-robin.c @@ -220,8 +220,10 @@ static int __init dm_rr_init(void) { int r = dm_register_path_selector(&rr_ps); - if (r < 0) + if (r < 0) { DMERR("register failed %d", r); + return r; + } DMINFO("version " RR_VERSION " loaded"); diff --git a/drivers/md/dm-ps-service-time.c b/drivers/md/dm-ps-service-time.c index 969d31c40272..f8c43aecdb27 100644 --- a/drivers/md/dm-ps-service-time.c +++ b/drivers/md/dm-ps-service-time.c @@ -341,8 +341,10 @@ static int __init dm_st_init(void) { int r = dm_register_path_selector(&st_ps); - if (r < 0) + if (r < 0) { DMERR("register failed %d", r); + return r; + } DMINFO("version " ST_VERSION " loaded"); diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c index 6adc55fd90d3..e8c0a8c6fb51 100644 --- a/drivers/md/dm-raid.c +++ b/drivers/md/dm-raid.c @@ -14,6 +14,7 @@ #include "raid5.h" #include "raid10.h" #include "md-bitmap.h" +#include "dm-core.h" #include <linux/device-mapper.h> @@ -1355,11 +1356,7 @@ static int parse_raid_params(struct raid_set *rs, struct dm_arg_set *as, return -EINVAL; } - /* - * In device-mapper, we specify things in sectors, but - * MD records this value in kB - */ - if (value < 0 || value / 2 > COUNTER_MAX) { + if (value < 0) { rs->ti->error = "Max write-behind limit out of range"; return -EINVAL; } @@ -2410,7 +2407,7 @@ static int super_init_validation(struct raid_set *rs, struct md_rdev *rdev) */ sb_retrieve_failed_devices(sb, failed_devices); rdev_for_each(r, mddev) { - if (test_bit(Journal, &rdev->flags) || + if (test_bit(Journal, &r->flags) || !r->sb_page) continue; sb2 = page_address(r->sb_page); @@ -3308,6 +3305,7 @@ size_check: /* Disable/enable discard support on raid set. */ configure_discard_support(rs); + rs->md.dm_gendisk = ti->table->md->disk; mddev_unlock(&rs->md); return 0; @@ -3327,6 +3325,7 @@ static void raid_dtr(struct dm_target *ti) mddev_lock_nointr(&rs->md); md_stop(&rs->md); + rs->md.dm_gendisk = NULL; mddev_unlock(&rs->md); if (work_pending(&rs->md.event_work)) diff --git a/drivers/md/dm-raid1.c b/drivers/md/dm-raid1.c index 8c6f1f7e6456..268f734ca9c3 100644 --- a/drivers/md/dm-raid1.c +++ b/drivers/md/dm-raid1.c @@ -103,7 +103,7 @@ static void wakeup_mirrord(void *context) static void delayed_wake_fn(struct timer_list *t) { - struct mirror_set *ms = from_timer(ms, t, timer); + struct mirror_set *ms = timer_container_of(ms, t, timer); clear_bit(0, &ms->timer_pending); wakeup_mirrord(ms); @@ -133,10 +133,9 @@ static void queue_bio(struct mirror_set *ms, struct bio *bio, int rw) spin_lock_irqsave(&ms->lock, flags); should_wake = !(bl->head); bio_list_add(bl, bio); - spin_unlock_irqrestore(&ms->lock, flags); - if (should_wake) wakeup_mirrord(ms); + spin_unlock_irqrestore(&ms->lock, flags); } static void dispatch_bios(void *context, struct bio_list *bio_list) @@ -646,9 +645,9 @@ static void write_callback(unsigned long error, void *context) if (!ms->failures.head) should_wake = 1; bio_list_add(&ms->failures, bio); - spin_unlock_irqrestore(&ms->lock, flags); if (should_wake) wakeup_mirrord(ms); + spin_unlock_irqrestore(&ms->lock, flags); } static void do_write(struct mirror_set *ms, struct bio *bio) @@ -1182,7 +1181,7 @@ static void mirror_dtr(struct dm_target *ti) { struct mirror_set *ms = ti->private; - del_timer_sync(&ms->timer); + timer_delete_sync(&ms->timer); flush_workqueue(ms->kmirrord_wq); flush_work(&ms->trigger_event); dm_kcopyd_client_destroy(ms->kcopyd_client); diff --git a/drivers/md/dm-rq.c b/drivers/md/dm-rq.c index e23076f7ece2..a6ca92049c10 100644 --- a/drivers/md/dm-rq.c +++ b/drivers/md/dm-rq.c @@ -217,10 +217,10 @@ static void dm_done(struct request *clone, blk_status_t error, bool mapped) if (unlikely(error == BLK_STS_TARGET)) { if (req_op(clone) == REQ_OP_DISCARD && !clone->q->limits.max_discard_sectors) - disable_discard(tio->md); + blk_queue_disable_discard(tio->md->queue); else if (req_op(clone) == REQ_OP_WRITE_ZEROES && !clone->q->limits.max_write_zeroes_sectors) - disable_write_zeroes(tio->md); + blk_queue_disable_write_zeroes(tio->md->queue); } switch (r) { diff --git a/drivers/md/dm-stripe.c b/drivers/md/dm-stripe.c index 3786ac67cefe..5bbbdf8fc1bd 100644 --- a/drivers/md/dm-stripe.c +++ b/drivers/md/dm-stripe.c @@ -405,7 +405,7 @@ static int stripe_end_io(struct dm_target *ti, struct bio *bio, blk_status_t *error) { unsigned int i; - char major_minor[16]; + char major_minor[22]; struct stripe_c *sc = ti->private; if (!*error) @@ -417,8 +417,7 @@ static int stripe_end_io(struct dm_target *ti, struct bio *bio, if (*error == BLK_STS_NOTSUPP) return DM_ENDIO_DONE; - memset(major_minor, 0, sizeof(major_minor)); - sprintf(major_minor, "%d:%d", MAJOR(bio_dev(bio)), MINOR(bio_dev(bio))); + format_dev_t(major_minor, bio_dev(bio)); /* * Test to see which stripe drive triggered the event @@ -459,6 +458,7 @@ static void stripe_io_hints(struct dm_target *ti, struct stripe_c *sc = ti->private; unsigned int chunk_size = sc->chunk_size << SECTOR_SHIFT; + limits->chunk_sectors = sc->chunk_size; limits->io_min = chunk_size; limits->io_opt = chunk_size * sc->stripes; } @@ -467,7 +467,7 @@ static struct target_type stripe_target = { .name = "striped", .version = {1, 7, 0}, .features = DM_TARGET_PASSES_INTEGRITY | DM_TARGET_NOWAIT | - DM_TARGET_ATOMIC_WRITES, + DM_TARGET_ATOMIC_WRITES | DM_TARGET_PASSES_CRYPTO, .module = THIS_MODULE, .ctr = stripe_ctr, .dtr = stripe_dtr, diff --git a/drivers/md/dm-switch.c b/drivers/md/dm-switch.c index dfd9fb52a6f3..bb1a70b5a215 100644 --- a/drivers/md/dm-switch.c +++ b/drivers/md/dm-switch.c @@ -517,7 +517,9 @@ static void switch_status(struct dm_target *ti, status_type_t type, * * Passthrough all ioctls to the path for sector 0 */ -static int switch_prepare_ioctl(struct dm_target *ti, struct block_device **bdev) +static int switch_prepare_ioctl(struct dm_target *ti, struct block_device **bdev, + unsigned int cmd, unsigned long arg, + bool *forward) { struct switch_ctx *sctx = ti->private; unsigned int path_nr; diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c index 0ef5203387b2..79ba4bacd0f9 100644 --- a/drivers/md/dm-table.c +++ b/drivers/md/dm-table.c @@ -117,7 +117,6 @@ static int alloc_targets(struct dm_table *t, unsigned int num) n_targets = (struct dm_target *) (n_highs + num); memset(n_highs, -1, sizeof(*n_highs) * num); - kvfree(t->highs); t->num_allocated = num; t->highs = n_highs; @@ -257,7 +256,7 @@ static int device_area_is_invalid(struct dm_target *ti, struct dm_dev *dev, if (bdev_is_zoned(bdev)) { unsigned int zone_sectors = bdev_zone_sectors(bdev); - if (start & (zone_sectors - 1)) { + if (!bdev_is_zone_aligned(bdev, start)) { DMERR("%s: start=%llu not aligned to h/w zone size %u of %pg", dm_device_name(ti->table->md), (unsigned long long)start, @@ -274,7 +273,7 @@ static int device_area_is_invalid(struct dm_target *ti, struct dm_dev *dev, * devices do not end up with a smaller zone in the middle of * the sector range. */ - if (len & (zone_sectors - 1)) { + if (!bdev_is_zone_aligned(bdev, len)) { DMERR("%s: len=%llu not aligned to h/w zone size %u of %pg", dm_device_name(ti->table->md), (unsigned long long)len, @@ -431,6 +430,13 @@ static int dm_set_device_limits(struct dm_target *ti, struct dm_dev *dev, return 0; } + mutex_lock(&q->limits_lock); + /* + * BLK_FEAT_ATOMIC_WRITES is not inherited from the bottom device in + * blk_stack_limits(), so do it manually. + */ + limits->features |= (q->limits.features & BLK_FEAT_ATOMIC_WRITES); + if (blk_stack_limits(limits, &q->limits, get_start_sect(bdev) + start) < 0) DMWARN("%s: adding target device %pg caused an alignment inconsistency: " @@ -448,6 +454,7 @@ static int dm_set_device_limits(struct dm_target *ti, struct dm_dev *dev, */ if (!dm_target_has_integrity(ti->type)) queue_limits_stack_integrity_bdev(limits, bdev); + mutex_unlock(&q->limits_lock); return 0; } @@ -523,8 +530,9 @@ static char **realloc_argv(unsigned int *size, char **old_argv) gfp = GFP_NOIO; } argv = kmalloc_array(new_size, sizeof(*argv), gfp); - if (argv && old_argv) { - memcpy(argv, old_argv, *size * sizeof(*argv)); + if (argv) { + if (old_argv) + memcpy(argv, old_argv, *size * sizeof(*argv)); *size = new_size; } @@ -697,6 +705,10 @@ int dm_table_add_target(struct dm_table *t, const char *type, DMERR("%s: zero-length target", dm_device_name(t->md)); return -EINVAL; } + if (start + len < start || start + len > LLONG_MAX >> SECTOR_SHIFT) { + DMERR("%s: too large device", dm_device_name(t->md)); + return -EINVAL; + } ti->type = dm_get_target_type(type); if (!ti->type) { @@ -887,17 +899,17 @@ static bool dm_table_supports_dax(struct dm_table *t, return true; } -static int device_is_rq_stackable(struct dm_target *ti, struct dm_dev *dev, - sector_t start, sector_t len, void *data) +static int device_is_not_rq_stackable(struct dm_target *ti, struct dm_dev *dev, + sector_t start, sector_t len, void *data) { struct block_device *bdev = dev->bdev; struct request_queue *q = bdev_get_queue(bdev); /* request-based cannot stack on partitions! */ if (bdev_is_partition(bdev)) - return false; + return true; - return queue_is_mq(q); + return !queue_is_mq(q); } static int dm_table_determine_type(struct dm_table *t) @@ -993,7 +1005,7 @@ verify_rq_based: /* Non-request-stackable devices can't be used for request-based dm */ if (!ti->type->iterate_devices || - !ti->type->iterate_devices(ti, device_is_rq_stackable, NULL)) { + ti->type->iterate_devices(ti, device_is_not_rq_stackable, NULL)) { DMERR("table load rejected: including non-request-stackable devices"); return -EINVAL; } @@ -1045,7 +1057,6 @@ static int dm_table_alloc_md_mempools(struct dm_table *t, struct mapped_device * unsigned int min_pool_size = 0, pool_size; struct dm_md_mempools *pools; unsigned int bioset_flags = 0; - bool mempool_needs_integrity = t->integrity_supported; if (unlikely(type == DM_TYPE_NONE)) { DMERR("no table type is set, can't allocate mempools"); @@ -1070,8 +1081,6 @@ static int dm_table_alloc_md_mempools(struct dm_table *t, struct mapped_device * per_io_data_size = max(per_io_data_size, ti->per_io_data_size); min_pool_size = max(min_pool_size, ti->num_flush_bios); - - mempool_needs_integrity |= ti->mempool_needs_integrity; } pool_size = max(dm_get_reserved_bio_based_ios(), min_pool_size); front_pad = roundup(per_io_data_size, @@ -1081,15 +1090,9 @@ static int dm_table_alloc_md_mempools(struct dm_table *t, struct mapped_device * __alignof__(struct dm_io)) + DM_IO_BIO_OFFSET; if (bioset_init(&pools->io_bs, pool_size, io_front_pad, bioset_flags)) goto out_free_pools; - if (mempool_needs_integrity && - bioset_integrity_create(&pools->io_bs, pool_size)) - goto out_free_pools; init_bs: if (bioset_init(&pools->bs, pool_size, front_pad, 0)) goto out_free_pools; - if (mempool_needs_integrity && - bioset_integrity_create(&pools->bs, pool_size)) - goto out_free_pools; t->mempools = pools; return 0; @@ -1177,7 +1180,7 @@ static int dm_keyslot_evict(struct blk_crypto_profile *profile, t = dm_get_live_table(md, &srcu_idx); if (!t) - return 0; + goto put_live_table; for (unsigned int i = 0; i < t->num_targets; i++) { struct dm_target *ti = dm_table_get_target(t, i); @@ -1188,10 +1191,181 @@ static int dm_keyslot_evict(struct blk_crypto_profile *profile, (void *)key); } +put_live_table: dm_put_live_table(md, srcu_idx); return 0; } +enum dm_wrappedkey_op { + DERIVE_SW_SECRET, + IMPORT_KEY, + GENERATE_KEY, + PREPARE_KEY, +}; + +struct dm_wrappedkey_op_args { + enum dm_wrappedkey_op op; + int err; + union { + struct { + const u8 *eph_key; + size_t eph_key_size; + u8 *sw_secret; + } derive_sw_secret; + struct { + const u8 *raw_key; + size_t raw_key_size; + u8 *lt_key; + } import_key; + struct { + u8 *lt_key; + } generate_key; + struct { + const u8 *lt_key; + size_t lt_key_size; + u8 *eph_key; + } prepare_key; + }; +}; + +static int dm_wrappedkey_op_callback(struct dm_target *ti, struct dm_dev *dev, + sector_t start, sector_t len, void *data) +{ + struct dm_wrappedkey_op_args *args = data; + struct block_device *bdev = dev->bdev; + struct blk_crypto_profile *profile = + bdev_get_queue(bdev)->crypto_profile; + int err = -EOPNOTSUPP; + + if (!args->err) + return 0; + + switch (args->op) { + case DERIVE_SW_SECRET: + err = blk_crypto_derive_sw_secret( + bdev, + args->derive_sw_secret.eph_key, + args->derive_sw_secret.eph_key_size, + args->derive_sw_secret.sw_secret); + break; + case IMPORT_KEY: + err = blk_crypto_import_key(profile, + args->import_key.raw_key, + args->import_key.raw_key_size, + args->import_key.lt_key); + break; + case GENERATE_KEY: + err = blk_crypto_generate_key(profile, + args->generate_key.lt_key); + break; + case PREPARE_KEY: + err = blk_crypto_prepare_key(profile, + args->prepare_key.lt_key, + args->prepare_key.lt_key_size, + args->prepare_key.eph_key); + break; + } + args->err = err; + + /* Try another device in case this fails. */ + return 0; +} + +static int dm_exec_wrappedkey_op(struct blk_crypto_profile *profile, + struct dm_wrappedkey_op_args *args) +{ + struct mapped_device *md = + container_of(profile, struct dm_crypto_profile, profile)->md; + struct dm_target *ti; + struct dm_table *t; + int srcu_idx; + int i; + + args->err = -EOPNOTSUPP; + + t = dm_get_live_table(md, &srcu_idx); + if (!t) + goto out; + + /* + * blk-crypto currently has no support for multiple incompatible + * implementations of wrapped inline crypto keys on a single system. + * It was already checked earlier that support for wrapped keys was + * declared on all underlying devices. Thus, all the underlying devices + * should support all wrapped key operations and they should behave + * identically, i.e. work with the same keys. So, just executing the + * operation on the first device on which it works suffices for now. + */ + for (i = 0; i < t->num_targets; i++) { + ti = dm_table_get_target(t, i); + if (!ti->type->iterate_devices) + continue; + ti->type->iterate_devices(ti, dm_wrappedkey_op_callback, args); + if (!args->err) + break; + } +out: + dm_put_live_table(md, srcu_idx); + return args->err; +} + +static int dm_derive_sw_secret(struct blk_crypto_profile *profile, + const u8 *eph_key, size_t eph_key_size, + u8 sw_secret[BLK_CRYPTO_SW_SECRET_SIZE]) +{ + struct dm_wrappedkey_op_args args = { + .op = DERIVE_SW_SECRET, + .derive_sw_secret = { + .eph_key = eph_key, + .eph_key_size = eph_key_size, + .sw_secret = sw_secret, + }, + }; + return dm_exec_wrappedkey_op(profile, &args); +} + +static int dm_import_key(struct blk_crypto_profile *profile, + const u8 *raw_key, size_t raw_key_size, + u8 lt_key[BLK_CRYPTO_MAX_HW_WRAPPED_KEY_SIZE]) +{ + struct dm_wrappedkey_op_args args = { + .op = IMPORT_KEY, + .import_key = { + .raw_key = raw_key, + .raw_key_size = raw_key_size, + .lt_key = lt_key, + }, + }; + return dm_exec_wrappedkey_op(profile, &args); +} + +static int dm_generate_key(struct blk_crypto_profile *profile, + u8 lt_key[BLK_CRYPTO_MAX_HW_WRAPPED_KEY_SIZE]) +{ + struct dm_wrappedkey_op_args args = { + .op = GENERATE_KEY, + .generate_key = { + .lt_key = lt_key, + }, + }; + return dm_exec_wrappedkey_op(profile, &args); +} + +static int dm_prepare_key(struct blk_crypto_profile *profile, + const u8 *lt_key, size_t lt_key_size, + u8 eph_key[BLK_CRYPTO_MAX_HW_WRAPPED_KEY_SIZE]) +{ + struct dm_wrappedkey_op_args args = { + .op = PREPARE_KEY, + .prepare_key = { + .lt_key = lt_key, + .lt_key_size = lt_key_size, + .eph_key = eph_key, + }, + }; + return dm_exec_wrappedkey_op(profile, &args); +} + static int device_intersect_crypto_capabilities(struct dm_target *ti, struct dm_dev *dev, sector_t start, sector_t len, void *data) @@ -1250,6 +1424,7 @@ static int dm_table_construct_crypto_profile(struct dm_table *t) profile->max_dun_bytes_supported = UINT_MAX; memset(profile->modes_supported, 0xFF, sizeof(profile->modes_supported)); + profile->key_types_supported = ~0; for (i = 0; i < t->num_targets; i++) { struct dm_target *ti = dm_table_get_target(t, i); @@ -1265,6 +1440,13 @@ static int dm_table_construct_crypto_profile(struct dm_table *t) profile); } + if (profile->key_types_supported & BLK_CRYPTO_KEY_TYPE_HW_WRAPPED) { + profile->ll_ops.derive_sw_secret = dm_derive_sw_secret; + profile->ll_ops.import_key = dm_import_key; + profile->ll_ops.generate_key = dm_generate_key; + profile->ll_ops.prepare_key = dm_prepare_key; + } + if (t->md->queue && !blk_crypto_has_capabilities(profile, t->md->queue->crypto_profile)) { @@ -1492,6 +1674,18 @@ bool dm_table_has_no_data_devices(struct dm_table *t) return true; } +bool dm_table_is_wildcard(struct dm_table *t) +{ + for (unsigned int i = 0; i < t->num_targets; i++) { + struct dm_target *ti = dm_table_get_target(t, i); + + if (!dm_target_is_wildcard(ti->type)) + return false; + } + + return true; +} + static int device_not_zoned(struct dm_target *ti, struct dm_dev *dev, sector_t start, sector_t len, void *data) { @@ -1723,8 +1917,12 @@ static int device_not_write_zeroes_capable(struct dm_target *ti, struct dm_dev * sector_t start, sector_t len, void *data) { struct request_queue *q = bdev_get_queue(dev->bdev); + int b; - return !q->limits.max_write_zeroes_sectors; + mutex_lock(&q->limits_lock); + b = !q->limits.max_write_zeroes_sectors; + mutex_unlock(&q->limits_lock); + return b; } static bool dm_table_supports_write_zeroes(struct dm_table *t) @@ -1832,10 +2030,24 @@ static bool dm_table_supports_atomic_writes(struct dm_table *t) return true; } +bool dm_table_supports_size_change(struct dm_table *t, sector_t old_size, + sector_t new_size) +{ + if (IS_ENABLED(CONFIG_BLK_DEV_ZONED) && dm_has_zone_plugs(t->md) && + old_size != new_size) { + DMWARN("%s: device has zone write plug resources. " + "Cannot change size", + dm_device_name(t->md)); + return false; + } + return true; +} + int dm_table_set_restrictions(struct dm_table *t, struct request_queue *q, struct queue_limits *limits) { int r; + struct queue_limits old_limits; if (!dm_table_supports_nowait(t)) limits->features &= ~BLK_FEAT_NOWAIT; @@ -1862,28 +2074,30 @@ int dm_table_set_restrictions(struct dm_table *t, struct request_queue *q, if (dm_table_supports_flush(t)) limits->features |= BLK_FEAT_WRITE_CACHE | BLK_FEAT_FUA; - if (dm_table_supports_dax(t, device_not_dax_capable)) { + if (dm_table_supports_dax(t, device_not_dax_capable)) limits->features |= BLK_FEAT_DAX; - if (dm_table_supports_dax(t, device_not_dax_synchronous_capable)) - set_dax_synchronous(t->md->dax_dev); - } else + else limits->features &= ~BLK_FEAT_DAX; - if (dm_table_any_dev_attr(t, device_dax_write_cache_enabled, NULL)) - dax_write_cache(t->md->dax_dev, true); - /* For a zoned table, setup the zone related queue attributes. */ - if (IS_ENABLED(CONFIG_BLK_DEV_ZONED) && - (limits->features & BLK_FEAT_ZONED)) { - r = dm_set_zones_restrictions(t, q, limits); - if (r) - return r; + if (IS_ENABLED(CONFIG_BLK_DEV_ZONED)) { + if (limits->features & BLK_FEAT_ZONED) { + r = dm_set_zones_restrictions(t, q, limits); + if (r) + return r; + } else if (dm_has_zone_plugs(t->md)) { + DMWARN("%s: device has zone write plug resources. " + "Cannot switch to non-zoned table.", + dm_device_name(t->md)); + return -EINVAL; + } } if (dm_table_supports_atomic_writes(t)) limits->features |= BLK_FEAT_ATOMIC_WRITES; - r = queue_limits_set(q, limits); + old_limits = queue_limits_start_update(q); + r = queue_limits_commit_update(q, limits); if (r) return r; @@ -1894,10 +2108,21 @@ int dm_table_set_restrictions(struct dm_table *t, struct request_queue *q, if (IS_ENABLED(CONFIG_BLK_DEV_ZONED) && (limits->features & BLK_FEAT_ZONED)) { r = dm_revalidate_zones(t, q); - if (r) + if (r) { + queue_limits_set(q, &old_limits); return r; + } } + if (IS_ENABLED(CONFIG_BLK_DEV_ZONED)) + dm_finalize_zone_settings(t, limits); + + if (dm_table_supports_dax(t, device_not_dax_synchronous_capable)) + set_dax_synchronous(t->md->dax_dev); + + if (dm_table_any_dev_attr(t, device_dax_write_cache_enabled, NULL)) + dax_write_cache(t->md->dax_dev, true); + dm_update_crypto_profile(q, t); return 0; } diff --git a/drivers/md/dm-vdo/block-map.c b/drivers/md/dm-vdo/block-map.c index 89cb7942ec5c..baf683cabb1b 100644 --- a/drivers/md/dm-vdo/block-map.c +++ b/drivers/md/dm-vdo/block-map.c @@ -451,7 +451,7 @@ static struct page_info * __must_check find_page(struct vdo_page_cache *cache, * select_lru_page() - Determine which page is least recently used. * * Picks the least recently used from among the non-busy entries at the front of each of the lru - * ring. Since whenever we mark a page busy we also put it to the end of the ring it is unlikely + * list. Since whenever we mark a page busy we also put it to the end of the list it is unlikely * that the entries at the front are busy unless the queue is very short, but not impossible. * * Return: A pointer to the info structure for a relevant page, or NULL if no such page can be @@ -1544,7 +1544,7 @@ static void write_page_if_not_dirtied(struct vdo_waiter *waiter, void *context) static void return_to_pool(struct block_map_zone *zone, struct pooled_vio *vio) { - return_vio_to_pool(zone->vio_pool, vio); + return_vio_to_pool(vio); check_for_drain_complete(zone); } @@ -1837,7 +1837,7 @@ static void finish_block_map_page_load(struct vdo_completion *completion) if (!vdo_copy_valid_page(vio->data, nonce, pbn, page)) vdo_format_block_map_page(page, nonce, pbn, false); - return_vio_to_pool(zone->vio_pool, pooled); + return_vio_to_pool(pooled); /* Release our claim to the load and wake any waiters */ release_page_lock(data_vio, "load"); @@ -1851,10 +1851,9 @@ static void handle_io_error(struct vdo_completion *completion) struct vio *vio = as_vio(completion); struct pooled_vio *pooled = container_of(vio, struct pooled_vio, vio); struct data_vio *data_vio = completion->parent; - struct block_map_zone *zone = pooled->context; vio_record_metadata_io_error(vio); - return_vio_to_pool(zone->vio_pool, pooled); + return_vio_to_pool(pooled); abort_load(data_vio, result); } @@ -2499,7 +2498,7 @@ static void finish_cursor(struct cursor *cursor) struct cursors *cursors = cursor->parent; struct vdo_completion *completion = cursors->completion; - return_vio_to_pool(cursors->pool, vdo_forget(cursor->vio)); + return_vio_to_pool(vdo_forget(cursor->vio)); if (--cursors->active_roots > 0) return; @@ -2746,7 +2745,7 @@ static int __must_check initialize_block_map_zone(struct block_map *map, if (result != VDO_SUCCESS) return result; - result = make_vio_pool(vdo, BLOCK_MAP_VIO_POOL_SIZE, + result = make_vio_pool(vdo, BLOCK_MAP_VIO_POOL_SIZE, 1, zone->thread_id, VIO_TYPE_BLOCK_MAP_INTERIOR, VIO_PRIORITY_METADATA, zone, &zone->vio_pool); if (result != VDO_SUCCESS) diff --git a/drivers/md/dm-vdo/constants.h b/drivers/md/dm-vdo/constants.h index a8c4d6e24b38..2a8b03779f87 100644 --- a/drivers/md/dm-vdo/constants.h +++ b/drivers/md/dm-vdo/constants.h @@ -44,9 +44,6 @@ enum { /* The default size of each slab journal, in blocks */ DEFAULT_VDO_SLAB_JOURNAL_SIZE = 224, - /* Unit test minimum */ - MINIMUM_VDO_SLAB_JOURNAL_BLOCKS = 2, - /* * The initial size of lbn_operations and pbn_operations, which is based upon the expected * maximum number of outstanding VIOs. This value was chosen to make it highly unlikely diff --git a/drivers/md/dm-vdo/dedupe.c b/drivers/md/dm-vdo/dedupe.c index 3f3d29af1be4..4d983092a152 100644 --- a/drivers/md/dm-vdo/dedupe.c +++ b/drivers/md/dm-vdo/dedupe.c @@ -226,7 +226,7 @@ struct hash_lock { * A list containing the data VIOs sharing this lock, all having the same record name and * data block contents, linked by their hash_lock_node fields. */ - struct list_head duplicate_ring; + struct list_head duplicate_vios; /* The number of data_vios sharing this lock instance */ data_vio_count_t reference_count; @@ -343,7 +343,7 @@ static void return_hash_lock_to_pool(struct hash_zone *zone, struct hash_lock *l { memset(lock, 0, sizeof(*lock)); INIT_LIST_HEAD(&lock->pool_node); - INIT_LIST_HEAD(&lock->duplicate_ring); + INIT_LIST_HEAD(&lock->duplicate_vios); vdo_waitq_init(&lock->waiters); list_add_tail(&lock->pool_node, &zone->lock_pool); } @@ -441,7 +441,7 @@ static void set_hash_lock(struct data_vio *data_vio, struct hash_lock *new_lock) VDO_ASSERT_LOG_ONLY(data_vio->hash_zone != NULL, "must have a hash zone when holding a hash lock"); VDO_ASSERT_LOG_ONLY(!list_empty(&data_vio->hash_lock_entry), - "must be on a hash lock ring when holding a hash lock"); + "must be on a hash lock list when holding a hash lock"); VDO_ASSERT_LOG_ONLY(old_lock->reference_count > 0, "hash lock reference must be counted"); @@ -464,10 +464,10 @@ static void set_hash_lock(struct data_vio *data_vio, struct hash_lock *new_lock) if (new_lock != NULL) { /* - * Keep all data_vios sharing the lock on a ring since they can complete in any + * Keep all data_vios sharing the lock on a list since they can complete in any * order and we'll always need a pointer to one to compare data. */ - list_move_tail(&data_vio->hash_lock_entry, &new_lock->duplicate_ring); + list_move_tail(&data_vio->hash_lock_entry, &new_lock->duplicate_vios); new_lock->reference_count += 1; if (new_lock->max_references < new_lock->reference_count) new_lock->max_references = new_lock->reference_count; @@ -1789,10 +1789,10 @@ static bool is_hash_collision(struct hash_lock *lock, struct data_vio *candidate struct hash_zone *zone; bool collides; - if (list_empty(&lock->duplicate_ring)) + if (list_empty(&lock->duplicate_vios)) return false; - lock_holder = list_first_entry(&lock->duplicate_ring, struct data_vio, + lock_holder = list_first_entry(&lock->duplicate_vios, struct data_vio, hash_lock_entry); zone = candidate->hash_zone; collides = !blocks_equal(lock_holder->vio.data, candidate->vio.data); @@ -1815,7 +1815,7 @@ static inline int assert_hash_lock_preconditions(const struct data_vio *data_vio return result; result = VDO_ASSERT(list_empty(&data_vio->hash_lock_entry), - "must not already be a member of a hash lock ring"); + "must not already be a member of a hash lock list"); if (result != VDO_SUCCESS) return result; @@ -1942,8 +1942,8 @@ void vdo_release_hash_lock(struct data_vio *data_vio) "returned hash lock must not be in use with state %s", get_hash_lock_state_name(lock->state)); VDO_ASSERT_LOG_ONLY(list_empty(&lock->pool_node), - "hash lock returned to zone must not be in a pool ring"); - VDO_ASSERT_LOG_ONLY(list_empty(&lock->duplicate_ring), + "hash lock returned to zone must not be in a pool list"); + VDO_ASSERT_LOG_ONLY(list_empty(&lock->duplicate_vios), "hash lock returned to zone must not reference DataVIOs"); return_hash_lock_to_pool(zone, lock); @@ -2261,7 +2261,7 @@ static void check_for_drain_complete(struct hash_zone *zone) if ((atomic_read(&zone->timer_state) == DEDUPE_QUERY_TIMER_IDLE) || change_timer_state(zone, DEDUPE_QUERY_TIMER_RUNNING, DEDUPE_QUERY_TIMER_IDLE)) { - del_timer_sync(&zone->timer); + timer_delete_sync(&zone->timer); } else { /* * There is an in flight time-out, which must get processed before we can continue. @@ -2337,7 +2337,7 @@ static void timeout_index_operations_callback(struct vdo_completion *completion) static void timeout_index_operations(struct timer_list *t) { - struct hash_zone *zone = from_timer(zone, t, timer); + struct hash_zone *zone = timer_container_of(zone, t, timer); if (change_timer_state(zone, DEDUPE_QUERY_TIMER_RUNNING, DEDUPE_QUERY_TIMER_FIRED)) diff --git a/drivers/md/dm-vdo/encodings.c b/drivers/md/dm-vdo/encodings.c index 100e92f8f866..b7cc0f41caca 100644 --- a/drivers/md/dm-vdo/encodings.c +++ b/drivers/md/dm-vdo/encodings.c @@ -711,24 +711,11 @@ int vdo_configure_slab(block_count_t slab_size, block_count_t slab_journal_block ref_blocks = vdo_get_saved_reference_count_size(slab_size - slab_journal_blocks); meta_blocks = (ref_blocks + slab_journal_blocks); - /* Make sure test code hasn't configured slabs to be too small. */ + /* Make sure configured slabs are not too small. */ if (meta_blocks >= slab_size) return VDO_BAD_CONFIGURATION; - /* - * If the slab size is very small, assume this must be a unit test and override the number - * of data blocks to be a power of two (wasting blocks in the slab). Many tests need their - * data_blocks fields to be the exact capacity of the configured volume, and that used to - * fall out since they use a power of two for the number of data blocks, the slab size was - * a power of two, and every block in a slab was a data block. - * - * TODO: Try to figure out some way of structuring testParameters and unit tests so this - * hack isn't needed without having to edit several unit tests every time the metadata size - * changes by one block. - */ data_blocks = slab_size - meta_blocks; - if ((slab_size < 1024) && !is_power_of_2(data_blocks)) - data_blocks = ((block_count_t) 1 << ilog2(data_blocks)); /* * Configure the slab journal thresholds. The flush threshold is 168 of 224 blocks in @@ -1221,11 +1208,6 @@ int vdo_validate_config(const struct vdo_config *config, if (result != VDO_SUCCESS) return result; - result = VDO_ASSERT(config->slab_journal_blocks >= MINIMUM_VDO_SLAB_JOURNAL_BLOCKS, - "slab journal size meets minimum size"); - if (result != VDO_SUCCESS) - return result; - result = VDO_ASSERT(config->slab_journal_blocks <= config->slab_size, "slab journal size is within expected bound"); if (result != VDO_SUCCESS) diff --git a/drivers/md/dm-vdo/indexer/index-layout.c b/drivers/md/dm-vdo/indexer/index-layout.c index af8fab83b0f3..61edf2b72427 100644 --- a/drivers/md/dm-vdo/indexer/index-layout.c +++ b/drivers/md/dm-vdo/indexer/index-layout.c @@ -54,7 +54,6 @@ * Each save also has a unique nonce. */ -#define MAGIC_SIZE 32 #define NONCE_INFO_SIZE 32 #define MAX_SAVES 2 @@ -98,9 +97,11 @@ enum region_type { #define SUPER_VERSION_CURRENT 3 #define SUPER_VERSION_MAXIMUM 7 -static const u8 LAYOUT_MAGIC[MAGIC_SIZE] = "*ALBIREO*SINGLE*FILE*LAYOUT*001*"; +static const u8 LAYOUT_MAGIC[] = "*ALBIREO*SINGLE*FILE*LAYOUT*001*"; static const u64 REGION_MAGIC = 0x416c6252676e3031; /* 'AlbRgn01' */ +#define MAGIC_SIZE (sizeof(LAYOUT_MAGIC) - 1) + struct region_header { u64 magic; u64 region_blocks; diff --git a/drivers/md/dm-vdo/indexer/index-session.c b/drivers/md/dm-vdo/indexer/index-session.c index aee0914d604a..aa575a24e0b2 100644 --- a/drivers/md/dm-vdo/indexer/index-session.c +++ b/drivers/md/dm-vdo/indexer/index-session.c @@ -100,7 +100,6 @@ static int get_index_session(struct uds_index_session *index_session) int uds_launch_request(struct uds_request *request) { - size_t internal_size; int result; if (request->callback == NULL) { @@ -121,10 +120,7 @@ int uds_launch_request(struct uds_request *request) } /* Reset all internal fields before processing. */ - internal_size = - sizeof(struct uds_request) - offsetof(struct uds_request, zone_number); - // FIXME should be using struct_group for this instead - memset((char *) request + sizeof(*request) - internal_size, 0, internal_size); + memset(&request->internal, 0, sizeof(request->internal)); result = get_index_session(request->session); if (result != UDS_SUCCESS) diff --git a/drivers/md/dm-vdo/indexer/indexer.h b/drivers/md/dm-vdo/indexer/indexer.h index 183a94eb7e92..7c1fc4577f5b 100644 --- a/drivers/md/dm-vdo/indexer/indexer.h +++ b/drivers/md/dm-vdo/indexer/indexer.h @@ -8,6 +8,7 @@ #include <linux/mutex.h> #include <linux/sched.h> +#include <linux/stddef.h> #include <linux/types.h> #include <linux/wait.h> @@ -73,7 +74,7 @@ enum uds_request_type { /* Remove any mapping for a name. */ UDS_DELETE, -}; +} __packed; enum uds_open_index_type { /* Create a new index. */ @@ -226,7 +227,7 @@ struct uds_zone_message { enum uds_zone_message_type type; /* The virtual chapter number to which the message applies */ u64 virtual_chapter; -}; +} __packed; struct uds_index_session; struct uds_index; @@ -253,34 +254,32 @@ struct uds_request { /* The existing data associated with the request name, if any */ struct uds_record_data old_metadata; - /* Either UDS_SUCCESS or an error code for the request */ - int status; /* True if the record name had an existing entry in the index */ bool found; + /* Either UDS_SUCCESS or an error code for the request */ + int status; - /* - * The remaining fields are used internally and should not be altered by clients. The index - * relies on zone_number being the first field in this section. - */ - - /* The number of the zone which will process this request*/ - unsigned int zone_number; - /* A link for adding a request to a lock-free queue */ - struct funnel_queue_entry queue_link; - /* A link for adding a request to a standard linked list */ - struct uds_request *next_request; - /* A pointer to the index processing this request */ - struct uds_index *index; - /* Control message for coordinating between zones */ - struct uds_zone_message zone_message; - /* If true, process request immediately by waking the worker thread */ - bool unbatched; - /* If true, continue this request before processing newer requests */ - bool requeued; - /* The virtual chapter containing the record name, if known */ - u64 virtual_chapter; - /* The region of the index containing the record name */ - enum uds_index_region location; + /* The remaining fields are used internally and should not be altered by clients. */ + struct_group(internal, + /* The virtual chapter containing the record name, if known */ + u64 virtual_chapter; + /* The region of the index containing the record name */ + enum uds_index_region location; + /* If true, process request immediately by waking the worker thread */ + bool unbatched; + /* If true, continue this request before processing newer requests */ + bool requeued; + /* Control message for coordinating between zones */ + struct uds_zone_message zone_message; + /* The number of the zone which will process this request*/ + unsigned int zone_number; + /* A link for adding a request to a lock-free queue */ + struct funnel_queue_entry queue_link; + /* A link for adding a request to a standard linked list */ + struct uds_request *next_request; + /* A pointer to the index processing this request */ + struct uds_index *index; + ); }; /* A session is required for most index operations. */ diff --git a/drivers/md/dm-vdo/indexer/volume.c b/drivers/md/dm-vdo/indexer/volume.c index 655453bb276b..425b3a74f4db 100644 --- a/drivers/md/dm-vdo/indexer/volume.c +++ b/drivers/md/dm-vdo/indexer/volume.c @@ -754,10 +754,11 @@ static int get_volume_page_protected(struct volume *volume, struct uds_request * u32 physical_page, struct cached_page **page_ptr) { struct cached_page *page; + unsigned int zone_number = request->zone_number; get_page_from_cache(&volume->page_cache, physical_page, &page); if (page != NULL) { - if (request->zone_number == 0) { + if (zone_number == 0) { /* Only one zone is allowed to update the LRU. */ make_page_most_recent(&volume->page_cache, page); } @@ -767,7 +768,7 @@ static int get_volume_page_protected(struct volume *volume, struct uds_request * } /* Prepare to enqueue a read for the page. */ - end_pending_search(&volume->page_cache, request->zone_number); + end_pending_search(&volume->page_cache, zone_number); mutex_lock(&volume->read_threads_mutex); /* @@ -787,8 +788,7 @@ static int get_volume_page_protected(struct volume *volume, struct uds_request * * the order does not matter for correctness as it does below. */ mutex_unlock(&volume->read_threads_mutex); - begin_pending_search(&volume->page_cache, physical_page, - request->zone_number); + begin_pending_search(&volume->page_cache, physical_page, zone_number); return UDS_QUEUED; } @@ -797,7 +797,7 @@ static int get_volume_page_protected(struct volume *volume, struct uds_request * * "search pending" state in careful order so no other thread can mess with the data before * the caller gets to look at it. */ - begin_pending_search(&volume->page_cache, physical_page, request->zone_number); + begin_pending_search(&volume->page_cache, physical_page, zone_number); mutex_unlock(&volume->read_threads_mutex); *page_ptr = page; return UDS_SUCCESS; @@ -849,6 +849,7 @@ static int search_cached_index_page(struct volume *volume, struct uds_request *r { int result; struct cached_page *page = NULL; + unsigned int zone_number = request->zone_number; u32 physical_page = map_to_physical_page(volume->geometry, chapter, index_page_number); @@ -858,18 +859,18 @@ static int search_cached_index_page(struct volume *volume, struct uds_request *r * invalidation by the reader thread, before the reader thread has noticed that the * invalidate_counter has been incremented. */ - begin_pending_search(&volume->page_cache, physical_page, request->zone_number); + begin_pending_search(&volume->page_cache, physical_page, zone_number); result = get_volume_page_protected(volume, request, physical_page, &page); if (result != UDS_SUCCESS) { - end_pending_search(&volume->page_cache, request->zone_number); + end_pending_search(&volume->page_cache, zone_number); return result; } result = uds_search_chapter_index_page(&page->index_page, volume->geometry, &request->record_name, record_page_number); - end_pending_search(&volume->page_cache, request->zone_number); + end_pending_search(&volume->page_cache, zone_number); return result; } @@ -882,6 +883,7 @@ int uds_search_cached_record_page(struct volume *volume, struct uds_request *req { struct cached_page *record_page; struct index_geometry *geometry = volume->geometry; + unsigned int zone_number = request->zone_number; int result; u32 physical_page, page_number; @@ -905,11 +907,11 @@ int uds_search_cached_record_page(struct volume *volume, struct uds_request *req * invalidation by the reader thread, before the reader thread has noticed that the * invalidate_counter has been incremented. */ - begin_pending_search(&volume->page_cache, physical_page, request->zone_number); + begin_pending_search(&volume->page_cache, physical_page, zone_number); result = get_volume_page_protected(volume, request, physical_page, &record_page); if (result != UDS_SUCCESS) { - end_pending_search(&volume->page_cache, request->zone_number); + end_pending_search(&volume->page_cache, zone_number); return result; } @@ -917,7 +919,7 @@ int uds_search_cached_record_page(struct volume *volume, struct uds_request *req &request->record_name, geometry, &request->old_metadata)) *found = true; - end_pending_search(&volume->page_cache, request->zone_number); + end_pending_search(&volume->page_cache, zone_number); return UDS_SUCCESS; } diff --git a/drivers/md/dm-vdo/io-submitter.c b/drivers/md/dm-vdo/io-submitter.c index 421e5436c32c..11d47770b54d 100644 --- a/drivers/md/dm-vdo/io-submitter.c +++ b/drivers/md/dm-vdo/io-submitter.c @@ -327,6 +327,7 @@ void vdo_submit_data_vio(struct data_vio *data_vio) * @error_handler: the handler for submission or I/O errors (may be NULL) * @operation: the type of I/O to perform * @data: the buffer to read or write (may be NULL) + * @size: the I/O amount in bytes * * The vio is enqueued on a vdo bio queue so that bio submission (which may block) does not block * other vdo threads. @@ -338,7 +339,7 @@ void vdo_submit_data_vio(struct data_vio *data_vio) */ void __submit_metadata_vio(struct vio *vio, physical_block_number_t physical, bio_end_io_t callback, vdo_action_fn error_handler, - blk_opf_t operation, char *data) + blk_opf_t operation, char *data, int size) { int result; struct vdo_completion *completion = &vio->completion; @@ -349,7 +350,8 @@ void __submit_metadata_vio(struct vio *vio, physical_block_number_t physical, vdo_reset_completion(completion); completion->error_handler = error_handler; - result = vio_reset_bio(vio, data, callback, operation | REQ_META, physical); + result = vio_reset_bio_with_size(vio, data, size, callback, operation | REQ_META, + physical); if (result != VDO_SUCCESS) { continue_vio(vio, result); return; diff --git a/drivers/md/dm-vdo/io-submitter.h b/drivers/md/dm-vdo/io-submitter.h index 80748699496f..3088f11055fd 100644 --- a/drivers/md/dm-vdo/io-submitter.h +++ b/drivers/md/dm-vdo/io-submitter.h @@ -8,6 +8,7 @@ #include <linux/bio.h> +#include "constants.h" #include "types.h" struct io_submitter; @@ -26,14 +27,25 @@ void vdo_submit_data_vio(struct data_vio *data_vio); void __submit_metadata_vio(struct vio *vio, physical_block_number_t physical, bio_end_io_t callback, vdo_action_fn error_handler, - blk_opf_t operation, char *data); + blk_opf_t operation, char *data, int size); static inline void vdo_submit_metadata_vio(struct vio *vio, physical_block_number_t physical, bio_end_io_t callback, vdo_action_fn error_handler, blk_opf_t operation) { __submit_metadata_vio(vio, physical, callback, error_handler, - operation, vio->data); + operation, vio->data, vio->block_count * VDO_BLOCK_SIZE); +} + +static inline void vdo_submit_metadata_vio_with_size(struct vio *vio, + physical_block_number_t physical, + bio_end_io_t callback, + vdo_action_fn error_handler, + blk_opf_t operation, + int size) +{ + __submit_metadata_vio(vio, physical, callback, error_handler, + operation, vio->data, size); } static inline void vdo_submit_flush_vio(struct vio *vio, bio_end_io_t callback, @@ -41,7 +53,7 @@ static inline void vdo_submit_flush_vio(struct vio *vio, bio_end_io_t callback, { /* FIXME: Can we just use REQ_OP_FLUSH? */ __submit_metadata_vio(vio, 0, callback, error_handler, - REQ_OP_WRITE | REQ_PREFLUSH, NULL); + REQ_OP_WRITE | REQ_PREFLUSH, NULL, 0); } #endif /* VDO_IO_SUBMITTER_H */ diff --git a/drivers/md/dm-vdo/packer.h b/drivers/md/dm-vdo/packer.h index 0f3be44710b5..8c8d6892582d 100644 --- a/drivers/md/dm-vdo/packer.h +++ b/drivers/md/dm-vdo/packer.h @@ -46,7 +46,7 @@ struct compressed_block { /* * Each packer_bin holds an incomplete batch of data_vios that only partially fill a compressed - * block. The bins are kept in a ring sorted by the amount of unused space so the first bin with + * block. The bins are kept in a list sorted by the amount of unused space so the first bin with * enough space to hold a newly-compressed data_vio can easily be found. When the bin fills up or * is flushed, the first uncanceled data_vio in the bin is selected to be the agent for that bin. * Upon entering the packer, each data_vio already has its compressed data in the first slot of the diff --git a/drivers/md/dm-vdo/priority-table.c b/drivers/md/dm-vdo/priority-table.c index 42d3d8d0e4b5..9bae8256ba4e 100644 --- a/drivers/md/dm-vdo/priority-table.c +++ b/drivers/md/dm-vdo/priority-table.c @@ -199,7 +199,7 @@ void vdo_priority_table_remove(struct priority_table *table, struct list_head *e /* * Remove the entry from the bucket list, remembering a pointer to another entry in the - * ring. + * list. */ next_entry = entry->next; list_del_init(entry); diff --git a/drivers/md/dm-vdo/recovery-journal.h b/drivers/md/dm-vdo/recovery-journal.h index 899071173015..25e7ec6d19f6 100644 --- a/drivers/md/dm-vdo/recovery-journal.h +++ b/drivers/md/dm-vdo/recovery-journal.h @@ -43,9 +43,9 @@ * has a vio which is used to commit that block to disk. The vio's data is the on-disk * representation of the journal block. In addition each in-memory block has a buffer which is used * to accumulate entries while a partial commit of the block is in progress. In-memory blocks are - * kept on two rings. Free blocks live on the 'free_tail_blocks' ring. When a block becomes active - * (see below) it is moved to the 'active_tail_blocks' ring. When a block is fully committed, it is - * moved back to the 'free_tail_blocks' ring. + * kept on two lists. Free blocks live on the 'free_tail_blocks' list. When a block becomes active + * (see below) it is moved to the 'active_tail_blocks' list. When a block is fully committed, it is + * moved back to the 'free_tail_blocks' list. * * When entries are added to the journal, they are added to the active in-memory block, as * indicated by the 'active_block' field. If the caller wishes to wait for the entry to be diff --git a/drivers/md/dm-vdo/slab-depot.c b/drivers/md/dm-vdo/slab-depot.c index 8f0a35c63af6..f3d80ff7bef5 100644 --- a/drivers/md/dm-vdo/slab-depot.c +++ b/drivers/md/dm-vdo/slab-depot.c @@ -139,7 +139,7 @@ static bool is_slab_journal_blank(const struct vdo_slab *slab) } /** - * mark_slab_journal_dirty() - Put a slab journal on the dirty ring of its allocator in the correct + * mark_slab_journal_dirty() - Put a slab journal on the dirty list of its allocator in the correct * order. * @journal: The journal to be marked dirty. * @lock: The recovery journal lock held by the slab journal. @@ -414,8 +414,7 @@ static void complete_reaping(struct vdo_completion *completion) { struct slab_journal *journal = completion->parent; - return_vio_to_pool(journal->slab->allocator->vio_pool, - vio_as_pooled_vio(as_vio(vdo_forget(completion)))); + return_vio_to_pool(vio_as_pooled_vio(as_vio(completion))); finish_reaping(journal); reap_slab_journal(journal); } @@ -698,7 +697,7 @@ static void complete_write(struct vdo_completion *completion) sequence_number_t committed = get_committing_sequence_number(pooled); list_del_init(&pooled->list_entry); - return_vio_to_pool(journal->slab->allocator->vio_pool, vdo_forget(pooled)); + return_vio_to_pool(pooled); if (result != VDO_SUCCESS) { vio_record_metadata_io_error(as_vio(completion)); @@ -822,7 +821,7 @@ static void commit_tail(struct slab_journal *journal) /* * Since we are about to commit the tail block, this journal no longer needs to be on the - * ring of journals which the recovery journal might ask to commit. + * list of journals which the recovery journal might ask to commit. */ mark_slab_journal_clean(journal); @@ -1076,7 +1075,7 @@ static void finish_reference_block_write(struct vdo_completion *completion) /* Release the slab journal lock. */ adjust_slab_journal_block_reference(&slab->journal, block->slab_journal_lock_to_release, -1); - return_vio_to_pool(slab->allocator->vio_pool, pooled); + return_vio_to_pool(pooled); /* * We can't clear the is_writing flag earlier as releasing the slab journal lock may cause @@ -1170,8 +1169,8 @@ static void handle_io_error(struct vdo_completion *completion) struct vdo_slab *slab = ((struct reference_block *) completion->parent)->slab; vio_record_metadata_io_error(vio); - return_vio_to_pool(slab->allocator->vio_pool, vio_as_pooled_vio(vio)); - slab->active_count--; + return_vio_to_pool(vio_as_pooled_vio(vio)); + slab->active_count -= vio->io_size / VDO_BLOCK_SIZE; vdo_enter_read_only_mode(slab->allocator->depot->vdo, result); check_if_slab_drained(slab); } @@ -1372,7 +1371,7 @@ static unsigned int calculate_slab_priority(struct vdo_slab *slab) static void prioritize_slab(struct vdo_slab *slab) { VDO_ASSERT_LOG_ONLY(list_empty(&slab->allocq_entry), - "a slab must not already be on a ring when prioritizing"); + "a slab must not already be on a list when prioritizing"); slab->priority = calculate_slab_priority(slab); vdo_priority_table_enqueue(slab->allocator->prioritized_slabs, slab->priority, &slab->allocq_entry); @@ -2165,28 +2164,95 @@ static void dirty_all_reference_blocks(struct vdo_slab *slab) dirty_block(&slab->reference_blocks[i]); } +static inline bool journal_points_equal(struct journal_point first, + struct journal_point second) +{ + return ((first.sequence_number == second.sequence_number) && + (first.entry_count == second.entry_count)); +} + /** - * clear_provisional_references() - Clear the provisional reference counts from a reference block. - * @block: The block to clear. + * match_bytes() - Check an 8-byte word for bytes matching the value specified + * @input: A word to examine the bytes of + * @match: The byte value sought + * + * Return: 1 in each byte when the corresponding input byte matched, 0 otherwise */ -static void clear_provisional_references(struct reference_block *block) +static inline u64 match_bytes(u64 input, u8 match) { - vdo_refcount_t *counters = get_reference_counters_for_block(block); - block_count_t j; + u64 temp = input ^ (match * 0x0101010101010101ULL); + /* top bit of each byte is set iff top bit of temp byte is clear; rest are 0 */ + u64 test_top_bits = ~temp & 0x8080808080808080ULL; + /* top bit of each byte is set iff low 7 bits of temp byte are clear; rest are useless */ + u64 test_low_bits = 0x8080808080808080ULL - (temp & 0x7f7f7f7f7f7f7f7fULL); + /* return 1 when both tests indicate temp byte is 0 */ + return (test_top_bits & test_low_bits) >> 7; +} + +/** + * count_valid_references() - Process a newly loaded refcount array + * @counters: the array of counters from a metadata block + * + * Scan a 8-byte-aligned array of counters, fixing up any "provisional" values that weren't + * cleaned up at shutdown, changing them internally to "empty". + * + * Return: the number of blocks that are referenced (counters not "empty") + */ +static unsigned int count_valid_references(vdo_refcount_t *counters) +{ + u64 *words = (u64 *)counters; + /* It's easier to count occurrences of a specific byte than its absences. */ + unsigned int empty_count = 0; + /* For speed, we process 8 bytes at once. */ + unsigned int words_left = COUNTS_PER_BLOCK / sizeof(u64); + + /* + * Sanity check assumptions used for optimizing this code: Counters are bytes. The counter + * array is a multiple of the word size. + */ + BUILD_BUG_ON(sizeof(vdo_refcount_t) != 1); + BUILD_BUG_ON((COUNTS_PER_BLOCK % sizeof(u64)) != 0); + + while (words_left > 0) { + /* + * This is used effectively as 8 byte-size counters. Byte 0 counts how many words + * had the target value found in byte 0, etc. We just have to avoid overflow. + */ + u64 split_count = 0; + /* + * The counter "% 255" trick used below to fold split_count into empty_count + * imposes a limit of 254 bytes examined each iteration of the outer loop. We + * process a word at a time, so that limit gets rounded down to 31 u64 words. + */ + const unsigned int max_words_per_iteration = 254 / sizeof(u64); + unsigned int iter_words_left = min_t(unsigned int, words_left, + max_words_per_iteration); + + words_left -= iter_words_left; + + while (iter_words_left--) { + u64 word = *words; + u64 temp; + + /* First, if we have any provisional refcount values, clear them. */ + temp = match_bytes(word, PROVISIONAL_REFERENCE_COUNT); + if (temp) { + /* + * 'temp' has 0x01 bytes where 'word' has PROVISIONAL; this xor + * will alter just those bytes, changing PROVISIONAL to EMPTY. + */ + word ^= temp * (PROVISIONAL_REFERENCE_COUNT ^ EMPTY_REFERENCE_COUNT); + *words = word; + } - for (j = 0; j < COUNTS_PER_BLOCK; j++) { - if (counters[j] == PROVISIONAL_REFERENCE_COUNT) { - counters[j] = EMPTY_REFERENCE_COUNT; - block->allocated_count--; + /* Now count the EMPTY_REFERENCE_COUNT bytes, updating the 8 counters. */ + split_count += match_bytes(word, EMPTY_REFERENCE_COUNT); + words++; } + empty_count += split_count % 255; } -} -static inline bool journal_points_equal(struct journal_point first, - struct journal_point second) -{ - return ((first.sequence_number == second.sequence_number) && - (first.entry_count == second.entry_count)); + return COUNTS_PER_BLOCK - empty_count; } /** @@ -2197,7 +2263,6 @@ static inline bool journal_points_equal(struct journal_point first, static void unpack_reference_block(struct packed_reference_block *packed, struct reference_block *block) { - block_count_t index; sector_count_t i; struct vdo_slab *slab = block->slab; vdo_refcount_t *counters = get_reference_counters_for_block(block); @@ -2223,11 +2288,7 @@ static void unpack_reference_block(struct packed_reference_block *packed, } } - block->allocated_count = 0; - for (index = 0; index < COUNTS_PER_BLOCK; index++) { - if (counters[index] != EMPTY_REFERENCE_COUNT) - block->allocated_count++; - } + block->allocated_count = count_valid_references(counters); } /** @@ -2240,13 +2301,19 @@ static void finish_reference_block_load(struct vdo_completion *completion) struct pooled_vio *pooled = vio_as_pooled_vio(vio); struct reference_block *block = completion->parent; struct vdo_slab *slab = block->slab; + unsigned int block_count = vio->io_size / VDO_BLOCK_SIZE; + unsigned int i; + char *data = vio->data; - unpack_reference_block((struct packed_reference_block *) vio->data, block); - return_vio_to_pool(slab->allocator->vio_pool, pooled); - slab->active_count--; - clear_provisional_references(block); + for (i = 0; i < block_count; i++, block++, data += VDO_BLOCK_SIZE) { + struct packed_reference_block *packed = (struct packed_reference_block *) data; + + unpack_reference_block(packed, block); + slab->free_blocks -= block->allocated_count; + } + return_vio_to_pool(pooled); + slab->active_count -= block_count; - slab->free_blocks -= block->allocated_count; check_if_slab_drained(slab); } @@ -2260,23 +2327,25 @@ static void load_reference_block_endio(struct bio *bio) } /** - * load_reference_block() - After a block waiter has gotten a VIO from the VIO pool, load the - * block. - * @waiter: The waiter of the block to load. + * load_reference_block_group() - After a block waiter has gotten a VIO from the VIO pool, load + * a set of blocks. + * @waiter: The waiter of the first block to load. * @context: The VIO returned by the pool. */ -static void load_reference_block(struct vdo_waiter *waiter, void *context) +static void load_reference_block_group(struct vdo_waiter *waiter, void *context) { struct pooled_vio *pooled = context; struct vio *vio = &pooled->vio; struct reference_block *block = container_of(waiter, struct reference_block, waiter); - size_t block_offset = (block - block->slab->reference_blocks); + u32 block_offset = block - block->slab->reference_blocks; + u32 max_block_count = block->slab->reference_block_count - block_offset; + u32 block_count = min_t(int, vio->block_count, max_block_count); vio->completion.parent = block; - vdo_submit_metadata_vio(vio, block->slab->ref_counts_origin + block_offset, - load_reference_block_endio, handle_io_error, - REQ_OP_READ); + vdo_submit_metadata_vio_with_size(vio, block->slab->ref_counts_origin + block_offset, + load_reference_block_endio, handle_io_error, + REQ_OP_READ, block_count * VDO_BLOCK_SIZE); } /** @@ -2286,14 +2355,21 @@ static void load_reference_block(struct vdo_waiter *waiter, void *context) static void load_reference_blocks(struct vdo_slab *slab) { block_count_t i; + u64 blocks_per_vio = slab->allocator->refcount_blocks_per_big_vio; + struct vio_pool *pool = slab->allocator->refcount_big_vio_pool; + + if (!pool) { + pool = slab->allocator->vio_pool; + blocks_per_vio = 1; + } slab->free_blocks = slab->block_count; slab->active_count = slab->reference_block_count; - for (i = 0; i < slab->reference_block_count; i++) { + for (i = 0; i < slab->reference_block_count; i += blocks_per_vio) { struct vdo_waiter *waiter = &slab->reference_blocks[i].waiter; - waiter->callback = load_reference_block; - acquire_vio_from_pool(slab->allocator->vio_pool, waiter); + waiter->callback = load_reference_block_group; + acquire_vio_from_pool(pool, waiter); } } @@ -2429,7 +2505,7 @@ static void finish_loading_journal(struct vdo_completion *completion) initialize_journal_state(journal); } - return_vio_to_pool(slab->allocator->vio_pool, vio_as_pooled_vio(vio)); + return_vio_to_pool(vio_as_pooled_vio(vio)); vdo_finish_loading_with_result(&slab->state, allocate_counters_if_clean(slab)); } @@ -2449,7 +2525,7 @@ static void handle_load_error(struct vdo_completion *completion) struct vio *vio = as_vio(completion); vio_record_metadata_io_error(vio); - return_vio_to_pool(journal->slab->allocator->vio_pool, vio_as_pooled_vio(vio)); + return_vio_to_pool(vio_as_pooled_vio(vio)); vdo_finish_loading_with_result(&journal->slab->state, result); } @@ -2547,7 +2623,7 @@ static void queue_slab(struct vdo_slab *slab) int result; VDO_ASSERT_LOG_ONLY(list_empty(&slab->allocq_entry), - "a requeued slab must not already be on a ring"); + "a requeued slab must not already be on a list"); if (vdo_is_read_only(allocator->depot->vdo)) return; @@ -2700,6 +2776,7 @@ static void finish_scrubbing(struct slab_scrubber *scrubber, int result) vdo_log_info("VDO commencing normal operation"); else if (prior_state == VDO_RECOVERING) vdo_log_info("Exiting recovery mode"); + free_vio_pool(vdo_forget(allocator->refcount_big_vio_pool)); } /* @@ -3281,7 +3358,7 @@ int vdo_release_block_reference(struct block_allocator *allocator, * This is a min_heap callback function orders slab_status structures using the 'is_clean' field as * the primary key and the 'emptiness' field as the secondary key. * - * Slabs need to be pushed onto the rings in the same order they are to be popped off. Popping + * Slabs need to be pushed onto the lists in the same order they are to be popped off. Popping * should always get the most empty first, so pushing should be from most empty to least empty. * Thus, the ordering is reversed from the usual sense since min_heap returns smaller elements * before larger ones. @@ -3983,6 +4060,7 @@ static int __must_check initialize_block_allocator(struct slab_depot *depot, struct vdo *vdo = depot->vdo; block_count_t max_free_blocks = depot->slab_config.data_blocks; unsigned int max_priority = (2 + ilog2(max_free_blocks)); + u32 reference_block_count, refcount_reads_needed, refcount_blocks_per_vio; *allocator = (struct block_allocator) { .depot = depot, @@ -4000,12 +4078,24 @@ static int __must_check initialize_block_allocator(struct slab_depot *depot, return result; vdo_initialize_completion(&allocator->completion, vdo, VDO_BLOCK_ALLOCATOR_COMPLETION); - result = make_vio_pool(vdo, BLOCK_ALLOCATOR_VIO_POOL_SIZE, allocator->thread_id, + result = make_vio_pool(vdo, BLOCK_ALLOCATOR_VIO_POOL_SIZE, 1, allocator->thread_id, VIO_TYPE_SLAB_JOURNAL, VIO_PRIORITY_METADATA, allocator, &allocator->vio_pool); if (result != VDO_SUCCESS) return result; + /* Initialize the refcount-reading vio pool. */ + reference_block_count = vdo_get_saved_reference_count_size(depot->slab_config.slab_blocks); + refcount_reads_needed = DIV_ROUND_UP(reference_block_count, MAX_BLOCKS_PER_VIO); + refcount_blocks_per_vio = DIV_ROUND_UP(reference_block_count, refcount_reads_needed); + allocator->refcount_blocks_per_big_vio = refcount_blocks_per_vio; + result = make_vio_pool(vdo, BLOCK_ALLOCATOR_REFCOUNT_VIO_POOL_SIZE, + allocator->refcount_blocks_per_big_vio, allocator->thread_id, + VIO_TYPE_SLAB_JOURNAL, VIO_PRIORITY_METADATA, + NULL, &allocator->refcount_big_vio_pool); + if (result != VDO_SUCCESS) + return result; + result = initialize_slab_scrubber(allocator); if (result != VDO_SUCCESS) return result; @@ -4223,6 +4313,7 @@ void vdo_free_slab_depot(struct slab_depot *depot) uninitialize_allocator_summary(allocator); uninitialize_scrubber_vio(&allocator->scrubber); free_vio_pool(vdo_forget(allocator->vio_pool)); + free_vio_pool(vdo_forget(allocator->refcount_big_vio_pool)); vdo_free_priority_table(vdo_forget(allocator->prioritized_slabs)); } diff --git a/drivers/md/dm-vdo/slab-depot.h b/drivers/md/dm-vdo/slab-depot.h index f234853501ca..fadc0c9d4dc4 100644 --- a/drivers/md/dm-vdo/slab-depot.h +++ b/drivers/md/dm-vdo/slab-depot.h @@ -45,6 +45,13 @@ enum { /* The number of vios in the vio pool is proportional to the throughput of the VDO. */ BLOCK_ALLOCATOR_VIO_POOL_SIZE = 128, + + /* + * The number of vios in the vio pool used for loading reference count data. A slab's + * refcounts is capped at ~8MB, and we process one at a time in a zone, so 9 should be + * plenty. + */ + BLOCK_ALLOCATOR_REFCOUNT_VIO_POOL_SIZE = 9, }; /* @@ -248,7 +255,7 @@ struct vdo_slab { /* A list of the dirty blocks waiting to be written out */ struct vdo_wait_queue dirty_blocks; - /* The number of blocks which are currently writing */ + /* The number of blocks which are currently reading or writing */ size_t active_count; /* A waiter object for updating the slab summary */ @@ -425,6 +432,10 @@ struct block_allocator { /* The vio pool for reading and writing block allocator metadata */ struct vio_pool *vio_pool; + /* The vio pool for large initial reads of ref count areas */ + struct vio_pool *refcount_big_vio_pool; + /* How many ref count blocks are read per vio at initial load */ + u32 refcount_blocks_per_big_vio; /* The dm_kcopyd client for erasing slab journals */ struct dm_kcopyd_client *eraser; /* Iterator over the slabs to be erased */ diff --git a/drivers/md/dm-vdo/types.h b/drivers/md/dm-vdo/types.h index dbe892b10f26..cdf36e7d7702 100644 --- a/drivers/md/dm-vdo/types.h +++ b/drivers/md/dm-vdo/types.h @@ -376,6 +376,9 @@ struct vio { /* The size of this vio in blocks */ unsigned int block_count; + /* The amount of data to be read or written, in bytes */ + unsigned int io_size; + /* The data being read or written. */ char *data; diff --git a/drivers/md/dm-vdo/vdo.c b/drivers/md/dm-vdo/vdo.c index a7e32baab4af..80b608674022 100644 --- a/drivers/md/dm-vdo/vdo.c +++ b/drivers/md/dm-vdo/vdo.c @@ -31,9 +31,7 @@ #include <linux/completion.h> #include <linux/device-mapper.h> -#include <linux/kernel.h> #include <linux/lz4.h> -#include <linux/module.h> #include <linux/mutex.h> #include <linux/spinlock.h> #include <linux/types.h> @@ -142,12 +140,6 @@ static void finish_vdo_request_queue(void *ptr) vdo_unregister_allocating_thread(); } -#ifdef MODULE -#define MODULE_NAME THIS_MODULE->name -#else -#define MODULE_NAME "dm-vdo" -#endif /* MODULE */ - static const struct vdo_work_queue_type default_queue_type = { .start = start_vdo_request_queue, .finish = finish_vdo_request_queue, @@ -559,8 +551,7 @@ int vdo_make(unsigned int instance, struct device_config *config, char **reason, *vdo_ptr = vdo; snprintf(vdo->thread_name_prefix, sizeof(vdo->thread_name_prefix), - "%s%u", MODULE_NAME, instance); - BUG_ON(vdo->thread_name_prefix[0] == '\0'); + "vdo%u", instance); result = vdo_allocate(vdo->thread_config.thread_count, struct vdo_thread, __func__, &vdo->threads); if (result != VDO_SUCCESS) { diff --git a/drivers/md/dm-vdo/vio.c b/drivers/md/dm-vdo/vio.c index e710f3c5a972..e7f4153e55e3 100644 --- a/drivers/md/dm-vdo/vio.c +++ b/drivers/md/dm-vdo/vio.c @@ -188,14 +188,23 @@ void vdo_set_bio_properties(struct bio *bio, struct vio *vio, bio_end_io_t callb /* * Prepares the bio to perform IO with the specified buffer. May only be used on a VDO-allocated - * bio, as it assumes the bio wraps a 4k buffer that is 4k aligned, but there does not have to be a - * vio associated with the bio. + * bio, as it assumes the bio wraps a 4k-multiple buffer that is 4k aligned, but there does not + * have to be a vio associated with the bio. */ int vio_reset_bio(struct vio *vio, char *data, bio_end_io_t callback, blk_opf_t bi_opf, physical_block_number_t pbn) { - int bvec_count, offset, len, i; + return vio_reset_bio_with_size(vio, data, vio->block_count * VDO_BLOCK_SIZE, + callback, bi_opf, pbn); +} + +int vio_reset_bio_with_size(struct vio *vio, char *data, int size, bio_end_io_t callback, + blk_opf_t bi_opf, physical_block_number_t pbn) +{ + int bvec_count, offset, i; struct bio *bio = vio->bio; + int vio_size = vio->block_count * VDO_BLOCK_SIZE; + int remaining; bio_reset(bio, bio->bi_bdev, bi_opf); vdo_set_bio_properties(bio, vio, callback, bi_opf, pbn); @@ -205,22 +214,21 @@ int vio_reset_bio(struct vio *vio, char *data, bio_end_io_t callback, bio->bi_ioprio = 0; bio->bi_io_vec = bio->bi_inline_vecs; bio->bi_max_vecs = vio->block_count + 1; - len = VDO_BLOCK_SIZE * vio->block_count; + if (VDO_ASSERT(size <= vio_size, "specified size %d is not greater than allocated %d", + size, vio_size) != VDO_SUCCESS) + size = vio_size; + vio->io_size = size; offset = offset_in_page(data); - bvec_count = DIV_ROUND_UP(offset + len, PAGE_SIZE); + bvec_count = DIV_ROUND_UP(offset + size, PAGE_SIZE); + remaining = size; - /* - * If we knew that data was always on one page, or contiguous pages, we wouldn't need the - * loop. But if we're using vmalloc, it's not impossible that the data is in different - * pages that can't be merged in bio_add_page... - */ - for (i = 0; (i < bvec_count) && (len > 0); i++) { + for (i = 0; (i < bvec_count) && (remaining > 0); i++) { struct page *page; int bytes_added; int bytes = PAGE_SIZE - offset; - if (bytes > len) - bytes = len; + if (bytes > remaining) + bytes = remaining; page = is_vmalloc_addr(data) ? vmalloc_to_page(data) : virt_to_page(data); bytes_added = bio_add_page(bio, page, bytes, offset); @@ -232,7 +240,7 @@ int vio_reset_bio(struct vio *vio, char *data, bio_end_io_t callback, } data += bytes; - len -= bytes; + remaining -= bytes; offset = 0; } @@ -301,6 +309,7 @@ void vio_record_metadata_io_error(struct vio *vio) * make_vio_pool() - Create a new vio pool. * @vdo: The vdo. * @pool_size: The number of vios in the pool. + * @block_count: The number of 4k blocks per vio. * @thread_id: The ID of the thread using this pool. * @vio_type: The type of vios in the pool. * @priority: The priority with which vios from the pool should be enqueued. @@ -309,13 +318,14 @@ void vio_record_metadata_io_error(struct vio *vio) * * Return: A success or error code. */ -int make_vio_pool(struct vdo *vdo, size_t pool_size, thread_id_t thread_id, +int make_vio_pool(struct vdo *vdo, size_t pool_size, size_t block_count, thread_id_t thread_id, enum vio_type vio_type, enum vio_priority priority, void *context, struct vio_pool **pool_ptr) { struct vio_pool *pool; char *ptr; int result; + size_t per_vio_size = VDO_BLOCK_SIZE * block_count; result = vdo_allocate_extended(struct vio_pool, pool_size, struct pooled_vio, __func__, &pool); @@ -326,7 +336,7 @@ int make_vio_pool(struct vdo *vdo, size_t pool_size, thread_id_t thread_id, INIT_LIST_HEAD(&pool->available); INIT_LIST_HEAD(&pool->busy); - result = vdo_allocate(pool_size * VDO_BLOCK_SIZE, char, + result = vdo_allocate(pool_size * per_vio_size, char, "VIO pool buffer", &pool->buffer); if (result != VDO_SUCCESS) { free_vio_pool(pool); @@ -334,10 +344,10 @@ int make_vio_pool(struct vdo *vdo, size_t pool_size, thread_id_t thread_id, } ptr = pool->buffer; - for (pool->size = 0; pool->size < pool_size; pool->size++, ptr += VDO_BLOCK_SIZE) { + for (pool->size = 0; pool->size < pool_size; pool->size++, ptr += per_vio_size) { struct pooled_vio *pooled = &pool->vios[pool->size]; - result = allocate_vio_components(vdo, vio_type, priority, NULL, 1, ptr, + result = allocate_vio_components(vdo, vio_type, priority, NULL, block_count, ptr, &pooled->vio); if (result != VDO_SUCCESS) { free_vio_pool(pool); @@ -345,6 +355,7 @@ int make_vio_pool(struct vdo *vdo, size_t pool_size, thread_id_t thread_id, } pooled->context = context; + pooled->pool = pool; list_add_tail(&pooled->pool_entry, &pool->available); } @@ -419,12 +430,13 @@ void acquire_vio_from_pool(struct vio_pool *pool, struct vdo_waiter *waiter) } /** - * return_vio_to_pool() - Return a vio to the pool - * @pool: The vio pool. + * return_vio_to_pool() - Return a vio to its pool * @vio: The pooled vio to return. */ -void return_vio_to_pool(struct vio_pool *pool, struct pooled_vio *vio) +void return_vio_to_pool(struct pooled_vio *vio) { + struct vio_pool *pool = vio->pool; + VDO_ASSERT_LOG_ONLY((pool->thread_id == vdo_get_callback_thread_id()), "vio pool entry returned on same thread as it was acquired"); diff --git a/drivers/md/dm-vdo/vio.h b/drivers/md/dm-vdo/vio.h index 3490e9f59b04..4bfcb21901f1 100644 --- a/drivers/md/dm-vdo/vio.h +++ b/drivers/md/dm-vdo/vio.h @@ -30,6 +30,8 @@ struct pooled_vio { void *context; /* The list entry used by the pool */ struct list_head pool_entry; + /* The pool this vio is allocated from */ + struct vio_pool *pool; }; /** @@ -123,6 +125,8 @@ void vdo_set_bio_properties(struct bio *bio, struct vio *vio, bio_end_io_t callb int vio_reset_bio(struct vio *vio, char *data, bio_end_io_t callback, blk_opf_t bi_opf, physical_block_number_t pbn); +int vio_reset_bio_with_size(struct vio *vio, char *data, int size, bio_end_io_t callback, + blk_opf_t bi_opf, physical_block_number_t pbn); void update_vio_error_stats(struct vio *vio, const char *format, ...) __printf(2, 3); @@ -188,12 +192,13 @@ static inline struct pooled_vio *vio_as_pooled_vio(struct vio *vio) struct vio_pool; -int __must_check make_vio_pool(struct vdo *vdo, size_t pool_size, thread_id_t thread_id, - enum vio_type vio_type, enum vio_priority priority, - void *context, struct vio_pool **pool_ptr); +int __must_check make_vio_pool(struct vdo *vdo, size_t pool_size, size_t block_count, + thread_id_t thread_id, enum vio_type vio_type, + enum vio_priority priority, void *context, + struct vio_pool **pool_ptr); void free_vio_pool(struct vio_pool *pool); bool __must_check is_vio_pool_busy(struct vio_pool *pool); void acquire_vio_from_pool(struct vio_pool *pool, struct vdo_waiter *waiter); -void return_vio_to_pool(struct vio_pool *pool, struct pooled_vio *vio); +void return_vio_to_pool(struct pooled_vio *vio); #endif /* VIO_H */ diff --git a/drivers/md/dm-vdo/wait-queue.c b/drivers/md/dm-vdo/wait-queue.c index 6e1e739277ef..f81ed0cee2bf 100644 --- a/drivers/md/dm-vdo/wait-queue.c +++ b/drivers/md/dm-vdo/wait-queue.c @@ -34,7 +34,7 @@ void vdo_waitq_enqueue_waiter(struct vdo_wait_queue *waitq, struct vdo_waiter *w waitq->last_waiter->next_waiter = waiter; } - /* In both cases, the waiter we added to the ring becomes the last waiter. */ + /* In both cases, the waiter we added to the list becomes the last waiter. */ waitq->last_waiter = waiter; waitq->length += 1; } diff --git a/drivers/md/dm-verity-fec.c b/drivers/md/dm-verity-fec.c index 0c41949db784..631a887b487c 100644 --- a/drivers/md/dm-verity-fec.c +++ b/drivers/md/dm-verity-fec.c @@ -593,6 +593,10 @@ int verity_fec_parse_opt_args(struct dm_arg_set *as, struct dm_verity *v, (*argc)--; if (!strcasecmp(arg_name, DM_VERITY_OPT_FEC_DEV)) { + if (v->fec->dev) { + ti->error = "FEC device already specified"; + return -EINVAL; + } r = dm_get_device(ti, arg_value, BLK_OPEN_READ, &v->fec->dev); if (r) { ti->error = "FEC device lookup failed"; diff --git a/drivers/md/dm-verity-target.c b/drivers/md/dm-verity-target.c index 24b167f71c5f..81186bded1ce 100644 --- a/drivers/md/dm-verity-target.c +++ b/drivers/md/dm-verity-target.c @@ -30,6 +30,7 @@ #define DM_VERITY_ENV_VAR_NAME "DM_VERITY_ERR_BLOCK_NR" #define DM_VERITY_DEFAULT_PREFETCH_SIZE 262144 +#define DM_VERITY_USE_BH_DEFAULT_BYTES 8192 #define DM_VERITY_MAX_CORRUPTED_ERRS 100 @@ -49,6 +50,15 @@ static unsigned int dm_verity_prefetch_cluster = DM_VERITY_DEFAULT_PREFETCH_SIZE module_param_named(prefetch_cluster, dm_verity_prefetch_cluster, uint, 0644); +static unsigned int dm_verity_use_bh_bytes[4] = { + DM_VERITY_USE_BH_DEFAULT_BYTES, // IOPRIO_CLASS_NONE + DM_VERITY_USE_BH_DEFAULT_BYTES, // IOPRIO_CLASS_RT + DM_VERITY_USE_BH_DEFAULT_BYTES, // IOPRIO_CLASS_BE + 0 // IOPRIO_CLASS_IDLE +}; + +module_param_array_named(use_bh_bytes, dm_verity_use_bh_bytes, uint, NULL, 0644); + static DEFINE_STATIC_KEY_FALSE(use_bh_wq_enabled); /* Is at least one dm-verity instance using ahash_tfm instead of shash_tfm? */ @@ -311,7 +321,7 @@ static int verity_verify_level(struct dm_verity *v, struct dm_verity_io *io, if (static_branch_unlikely(&use_bh_wq_enabled) && io->in_bh) { data = dm_bufio_get(v->bufio, hash_block, &buf); - if (data == NULL) { + if (IS_ERR_OR_NULL(data)) { /* * In tasklet and the hash was not in the bufio cache. * Return early and resume execution from a work-queue @@ -324,8 +334,24 @@ static int verity_verify_level(struct dm_verity *v, struct dm_verity_io *io, &buf, bio->bi_ioprio); } - if (IS_ERR(data)) - return PTR_ERR(data); + if (IS_ERR(data)) { + if (skip_unverified) + return 1; + r = PTR_ERR(data); + data = dm_bufio_new(v->bufio, hash_block, &buf); + if (IS_ERR(data)) + return r; + if (verity_fec_decode(v, io, DM_VERITY_BLOCK_TYPE_METADATA, + hash_block, data) == 0) { + aux = dm_bufio_get_aux_data(buf); + aux->hash_verified = 1; + goto release_ok; + } else { + dm_bufio_release(buf); + dm_bufio_forget(v->bufio, hash_block); + return r; + } + } aux = dm_bufio_get_aux_data(buf); @@ -366,6 +392,7 @@ static int verity_verify_level(struct dm_verity *v, struct dm_verity_io *io, } } +release_ok: data += offset; memcpy(want_digest, data, v->digest_size); r = 0; @@ -652,9 +679,18 @@ static void verity_bh_work(struct work_struct *w) verity_finish_io(io, errno_to_blk_status(err)); } +static inline bool verity_use_bh(unsigned int bytes, unsigned short ioprio) +{ + return ioprio <= IOPRIO_CLASS_IDLE && + bytes <= READ_ONCE(dm_verity_use_bh_bytes[ioprio]) && + !need_resched(); +} + static void verity_end_io(struct bio *bio) { struct dm_verity_io *io = bio->bi_private; + unsigned short ioprio = IOPRIO_PRIO_CLASS(bio->bi_ioprio); + unsigned int bytes = io->n_blocks << io->v->data_dev_block_bits; if (bio->bi_status && (!verity_fec_is_enabled(io->v) || @@ -664,9 +700,14 @@ static void verity_end_io(struct bio *bio) return; } - if (static_branch_unlikely(&use_bh_wq_enabled) && io->v->use_bh_wq) { - INIT_WORK(&io->bh_work, verity_bh_work); - queue_work(system_bh_wq, &io->bh_work); + if (static_branch_unlikely(&use_bh_wq_enabled) && io->v->use_bh_wq && + verity_use_bh(bytes, ioprio)) { + if (in_hardirq() || irqs_disabled()) { + INIT_WORK(&io->bh_work, verity_bh_work); + queue_work(system_bh_wq, &io->bh_work); + } else { + verity_bh_work(&io->bh_work); + } } else { INIT_WORK(&io->work, verity_work); queue_work(io->v->verify_wq, &io->work); @@ -953,7 +994,9 @@ static void verity_status(struct dm_target *ti, status_type_t type, } } -static int verity_prepare_ioctl(struct dm_target *ti, struct block_device **bdev) +static int verity_prepare_ioctl(struct dm_target *ti, struct block_device **bdev, + unsigned int cmd, unsigned long arg, + bool *forward) { struct dm_verity *v = ti->private; @@ -1080,6 +1123,9 @@ static int verity_alloc_most_once(struct dm_verity *v) { struct dm_target *ti = v->ti; + if (v->validated_blocks) + return 0; + /* the bitset can only handle INT_MAX blocks */ if (v->data_blocks > INT_MAX) { ti->error = "device too large to use check_at_most_once"; @@ -1103,6 +1149,9 @@ static int verity_alloc_zero_digest(struct dm_verity *v) struct dm_verity_io *io; u8 *zero_data; + if (v->zero_digest) + return 0; + v->zero_digest = kmalloc(v->digest_size, GFP_KERNEL); if (!v->zero_digest) @@ -1537,7 +1586,7 @@ static int verity_ctr(struct dm_target *ti, unsigned int argc, char **argv) goto bad; } - /* Root hash signature is a optional parameter*/ + /* Root hash signature is an optional parameter */ r = verity_verify_root_hash(root_hash_digest_to_validate, strlen(root_hash_digest_to_validate), verify_args.sig, @@ -1768,7 +1817,7 @@ static struct target_type verity_target = { .name = "verity", /* Note: the LSMs depend on the singleton and immutable features */ .features = DM_TARGET_SINGLETON | DM_TARGET_IMMUTABLE, - .version = {1, 10, 0}, + .version = {1, 11, 0}, .module = THIS_MODULE, .ctr = verity_ctr, .dtr = verity_dtr, diff --git a/drivers/md/dm-verity-verify-sig.c b/drivers/md/dm-verity-verify-sig.c index a9e2c6c0a33c..d5261a0e4232 100644 --- a/drivers/md/dm-verity-verify-sig.c +++ b/drivers/md/dm-verity-verify-sig.c @@ -71,9 +71,14 @@ int verity_verify_sig_parse_opt_args(struct dm_arg_set *as, const char *arg_name) { struct dm_target *ti = v->ti; - int ret = 0; + int ret; const char *sig_key = NULL; + if (v->signature_key_desc) { + ti->error = DM_VERITY_VERIFY_ERR("root_hash_sig_key_desc already specified"); + return -EINVAL; + } + if (!*argc) { ti->error = DM_VERITY_VERIFY_ERR("Signature key not specified"); return -EINVAL; @@ -83,14 +88,18 @@ int verity_verify_sig_parse_opt_args(struct dm_arg_set *as, (*argc)--; ret = verity_verify_get_sig_from_key(sig_key, sig_opts); - if (ret < 0) + if (ret < 0) { ti->error = DM_VERITY_VERIFY_ERR("Invalid key specified"); + return ret; + } v->signature_key_desc = kstrdup(sig_key, GFP_KERNEL); - if (!v->signature_key_desc) + if (!v->signature_key_desc) { + ti->error = DM_VERITY_VERIFY_ERR("Could not allocate memory for signature key"); return -ENOMEM; + } - return ret; + return 0; } /* diff --git a/drivers/md/dm-writecache.c b/drivers/md/dm-writecache.c index 7ce8847b3404..a428e1cacf07 100644 --- a/drivers/md/dm-writecache.c +++ b/drivers/md/dm-writecache.c @@ -706,7 +706,7 @@ static inline void writecache_verify_watermark(struct dm_writecache *wc) static void writecache_max_age_timer(struct timer_list *t) { - struct dm_writecache *wc = from_timer(wc, t, max_age_timer); + struct dm_writecache *wc = timer_container_of(wc, t, max_age_timer); if (!dm_suspended(wc->ti) && !writecache_has_error(wc)) { queue_work(wc->writeback_wq, &wc->writeback_work); @@ -797,7 +797,7 @@ static void writecache_flush(struct dm_writecache *wc) bool need_flush_after_free; wc->uncommitted_blocks = 0; - del_timer(&wc->autocommit_timer); + timer_delete(&wc->autocommit_timer); if (list_empty(&wc->lru)) return; @@ -866,7 +866,7 @@ static void writecache_flush_work(struct work_struct *work) static void writecache_autocommit_timer(struct timer_list *t) { - struct dm_writecache *wc = from_timer(wc, t, autocommit_timer); + struct dm_writecache *wc = timer_container_of(wc, t, autocommit_timer); if (!writecache_has_error(wc)) queue_work(wc->writeback_wq, &wc->flush_work); @@ -927,8 +927,8 @@ static void writecache_suspend(struct dm_target *ti) struct dm_writecache *wc = ti->private; bool flush_on_suspend; - del_timer_sync(&wc->autocommit_timer); - del_timer_sync(&wc->max_age_timer); + timer_delete_sync(&wc->autocommit_timer); + timer_delete_sync(&wc->max_age_timer); wc_lock(wc); writecache_flush(wc); diff --git a/drivers/md/dm-zone.c b/drivers/md/dm-zone.c index 20edd3fabbab..3d31b82e0730 100644 --- a/drivers/md/dm-zone.c +++ b/drivers/md/dm-zone.c @@ -56,24 +56,31 @@ int dm_blk_report_zones(struct gendisk *disk, sector_t sector, { struct mapped_device *md = disk->private_data; struct dm_table *map; - int srcu_idx, ret; + struct dm_table *zone_revalidate_map = md->zone_revalidate_map; + int srcu_idx, ret = -EIO; + bool put_table = false; - if (!md->zone_revalidate_map) { - /* Regular user context */ + if (!zone_revalidate_map || md->revalidate_map_task != current) { + /* + * Regular user context or + * Zone revalidation during __bind() is in progress, but this + * call is from a different process + */ if (dm_suspended_md(md)) return -EAGAIN; map = dm_get_live_table(md, &srcu_idx); - if (!map) - return -EIO; + put_table = true; } else { /* Zone revalidation during __bind() */ - map = md->zone_revalidate_map; + map = zone_revalidate_map; } - ret = dm_blk_do_report_zones(md, map, sector, nr_zones, cb, data); + if (map) + ret = dm_blk_do_report_zones(md, map, sector, nr_zones, cb, + data); - if (!md->zone_revalidate_map) + if (put_table) dm_put_live_table(md, srcu_idx); return ret; @@ -153,33 +160,36 @@ int dm_revalidate_zones(struct dm_table *t, struct request_queue *q) { struct mapped_device *md = t->md; struct gendisk *disk = md->disk; + unsigned int nr_zones = disk->nr_zones; int ret; if (!get_capacity(disk)) return 0; - /* Revalidate only if something changed. */ - if (!disk->nr_zones || disk->nr_zones != md->nr_zones) { - DMINFO("%s using %s zone append", - disk->disk_name, - queue_emulates_zone_append(q) ? "emulated" : "native"); - md->nr_zones = 0; - } - - if (md->nr_zones) + /* + * Do not revalidate if zone write plug resources have already + * been allocated. + */ + if (dm_has_zone_plugs(md)) return 0; + DMINFO("%s using %s zone append", disk->disk_name, + queue_emulates_zone_append(q) ? "emulated" : "native"); + /* * Our table is not live yet. So the call to dm_get_live_table() * in dm_blk_report_zones() will fail. Set a temporary pointer to * our table for dm_blk_report_zones() to use directly. */ md->zone_revalidate_map = t; + md->revalidate_map_task = current; ret = blk_revalidate_disk_zones(disk); + md->revalidate_map_task = NULL; md->zone_revalidate_map = NULL; if (ret) { DMERR("Revalidate zones failed %d", ret); + disk->nr_zones = nr_zones; return ret; } @@ -337,15 +347,15 @@ int dm_set_zones_restrictions(struct dm_table *t, struct request_queue *q, /* * Check if zone append is natively supported, and if not, set the - * mapped device queue as needing zone append emulation. + * mapped device queue as needing zone append emulation. If zone + * append is natively supported, make sure that + * max_hw_zone_append_sectors is not set to 0. */ WARN_ON_ONCE(queue_is_mq(q)); - if (dm_table_supports_zone_append(t)) { - clear_bit(DMF_EMULATE_ZONE_APPEND, &md->flags); - } else { - set_bit(DMF_EMULATE_ZONE_APPEND, &md->flags); + if (!dm_table_supports_zone_append(t)) lim->max_hw_zone_append_sectors = 0; - } + else if (lim->max_hw_zone_append_sectors == 0) + lim->max_hw_zone_append_sectors = lim->max_zone_append_sectors; /* * Determine the max open and max active zone limits for the mapped @@ -380,15 +390,28 @@ int dm_set_zones_restrictions(struct dm_table *t, struct request_queue *q, lim->max_open_zones = 0; lim->max_active_zones = 0; lim->max_hw_zone_append_sectors = 0; + lim->max_zone_append_sectors = 0; lim->zone_write_granularity = 0; lim->chunk_sectors = 0; lim->features &= ~BLK_FEAT_ZONED; - clear_bit(DMF_EMULATE_ZONE_APPEND, &md->flags); - md->nr_zones = 0; - disk->nr_zones = 0; return 0; } + if (get_capacity(disk) && dm_has_zone_plugs(t->md)) { + if (q->limits.chunk_sectors != lim->chunk_sectors) { + DMWARN("%s: device has zone write plug resources. " + "Cannot change zone size", + disk->disk_name); + return -EINVAL; + } + if (lim->max_hw_zone_append_sectors != 0 && + !dm_table_is_wildcard(t)) { + DMWARN("%s: device has zone write plug resources. " + "New table must emulate zone append", + disk->disk_name); + return -EINVAL; + } + } /* * Warn once (when the capacity is not yet set) if the mapped device is * partially using zone resources of the target devices as that leads to @@ -408,6 +431,23 @@ int dm_set_zones_restrictions(struct dm_table *t, struct request_queue *q, return 0; } +void dm_finalize_zone_settings(struct dm_table *t, struct queue_limits *lim) +{ + struct mapped_device *md = t->md; + + if (lim->features & BLK_FEAT_ZONED) { + if (dm_table_supports_zone_append(t)) + clear_bit(DMF_EMULATE_ZONE_APPEND, &md->flags); + else + set_bit(DMF_EMULATE_ZONE_APPEND, &md->flags); + } else { + clear_bit(DMF_EMULATE_ZONE_APPEND, &md->flags); + md->nr_zones = 0; + md->disk->nr_zones = 0; + } +} + + /* * IO completion callback called from clone_endio(). */ @@ -423,9 +463,9 @@ void dm_zone_endio(struct dm_io *io, struct bio *clone) */ if (clone->bi_status == BLK_STS_OK && bio_op(clone) == REQ_OP_ZONE_APPEND) { - sector_t mask = bdev_zone_sectors(disk->part0) - 1; - - orig_bio->bi_iter.bi_sector += clone->bi_iter.bi_sector & mask; + orig_bio->bi_iter.bi_sector += + bdev_offset_from_zone_start(disk->part0, + clone->bi_iter.bi_sector); } return; diff --git a/drivers/md/dm-zoned-target.c b/drivers/md/dm-zoned-target.c index 6141fc25d842..9da329078ea4 100644 --- a/drivers/md/dm-zoned-target.c +++ b/drivers/md/dm-zoned-target.c @@ -1015,7 +1015,8 @@ static void dmz_io_hints(struct dm_target *ti, struct queue_limits *limits) /* * Pass on ioctl to the backend device. */ -static int dmz_prepare_ioctl(struct dm_target *ti, struct block_device **bdev) +static int dmz_prepare_ioctl(struct dm_target *ti, struct block_device **bdev, + unsigned int cmd, unsigned long arg, bool *forward) { struct dmz_target *dmz = ti->private; struct dmz_dev *dev = &dmz->dev[0]; @@ -1061,7 +1062,7 @@ static int dmz_iterate_devices(struct dm_target *ti, struct dmz_target *dmz = ti->private; unsigned int zone_nr_sectors = dmz_zone_nr_sectors(dmz->metadata); sector_t capacity; - int i, r; + int i, r = 0; for (i = 0; i < dmz->nr_ddevs; i++) { capacity = dmz->dev[i].capacity & ~(zone_nr_sectors - 1); diff --git a/drivers/md/dm.c b/drivers/md/dm.c index 4d1e42891d24..9f6d88ea60e6 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -411,7 +411,8 @@ static int dm_blk_getgeo(struct block_device *bdev, struct hd_geometry *geo) } static int dm_prepare_ioctl(struct mapped_device *md, int *srcu_idx, - struct block_device **bdev) + struct block_device **bdev, unsigned int cmd, + unsigned long arg, bool *forward) { struct dm_target *ti; struct dm_table *map; @@ -434,8 +435,8 @@ retry: if (dm_suspended_md(md)) return -EAGAIN; - r = ti->type->prepare_ioctl(ti, bdev); - if (r == -ENOTCONN && !fatal_signal_pending(current)) { + r = ti->type->prepare_ioctl(ti, bdev, cmd, arg, forward); + if (r == -ENOTCONN && *forward && !fatal_signal_pending(current)) { dm_put_live_table(md, *srcu_idx); fsleep(10000); goto retry; @@ -454,9 +455,10 @@ static int dm_blk_ioctl(struct block_device *bdev, blk_mode_t mode, { struct mapped_device *md = bdev->bd_disk->private_data; int r, srcu_idx; + bool forward = true; - r = dm_prepare_ioctl(md, &srcu_idx, &bdev); - if (r < 0) + r = dm_prepare_ioctl(md, &srcu_idx, &bdev, cmd, arg, &forward); + if (!forward || r < 0) goto out; if (r > 0) { @@ -1082,22 +1084,6 @@ static inline struct queue_limits *dm_get_queue_limits(struct mapped_device *md) return &md->queue->limits; } -void disable_discard(struct mapped_device *md) -{ - struct queue_limits *limits = dm_get_queue_limits(md); - - /* device doesn't really support DISCARD, disable it */ - limits->max_hw_discard_sectors = 0; -} - -void disable_write_zeroes(struct mapped_device *md) -{ - struct queue_limits *limits = dm_get_queue_limits(md); - - /* device doesn't really support WRITE ZEROES, disable it */ - limits->max_write_zeroes_sectors = 0; -} - static bool swap_bios_limit(struct dm_target *ti, struct bio *bio) { return unlikely((bio->bi_opf & REQ_SWAP) != 0) && unlikely(ti->limit_swap_bios); @@ -1115,10 +1101,10 @@ static void clone_endio(struct bio *bio) if (unlikely(error == BLK_STS_TARGET)) { if (bio_op(bio) == REQ_OP_DISCARD && !bdev_max_discard_sectors(bio->bi_bdev)) - disable_discard(md); + blk_queue_disable_discard(md->queue); else if (bio_op(bio) == REQ_OP_WRITE_ZEROES && !bdev_write_zeroes_sectors(bio->bi_bdev)) - disable_write_zeroes(md); + blk_queue_disable_write_zeroes(md->queue); } if (static_branch_unlikely(&zoned_enabled) && @@ -1540,14 +1526,18 @@ static void __send_empty_flush(struct clone_info *ci) { struct dm_table *t = ci->map; struct bio flush_bio; + blk_opf_t opf = REQ_OP_WRITE | REQ_PREFLUSH | REQ_SYNC; + + if ((ci->io->orig_bio->bi_opf & (REQ_IDLE | REQ_SYNC)) == + (REQ_IDLE | REQ_SYNC)) + opf |= REQ_IDLE; /* * Use an on-stack bio for this, it's safe since we don't * need to reference it after submit. It's just used as * the basis for the clone(s). */ - bio_init(&flush_bio, ci->io->md->disk->part0, NULL, 0, - REQ_OP_WRITE | REQ_PREFLUSH | REQ_SYNC); + bio_init(&flush_bio, ci->io->md->disk->part0, NULL, 0, opf); ci->bio = &flush_bio; ci->sector_count = 0; @@ -1786,19 +1776,35 @@ static void init_clone_info(struct clone_info *ci, struct dm_io *io, } #ifdef CONFIG_BLK_DEV_ZONED -static inline bool dm_zone_bio_needs_split(struct mapped_device *md, - struct bio *bio) +static inline bool dm_zone_bio_needs_split(struct bio *bio) { /* - * For mapped device that need zone append emulation, we must - * split any large BIO that straddles zone boundaries. + * Special case the zone operations that cannot or should not be split. + */ + switch (bio_op(bio)) { + case REQ_OP_ZONE_APPEND: + case REQ_OP_ZONE_FINISH: + case REQ_OP_ZONE_RESET: + case REQ_OP_ZONE_RESET_ALL: + return false; + default: + break; + } + + /* + * When mapped devices use the block layer zone write plugging, we must + * split any large BIO to the mapped device limits to not submit BIOs + * that span zone boundaries and to avoid potential deadlocks with + * queue freeze operations. */ - return dm_emulate_zone_append(md) && bio_straddles_zones(bio) && - !bio_flagged(bio, BIO_ZONE_WRITE_PLUGGING); + return bio_needs_zone_write_plugging(bio) || bio_straddles_zones(bio); } + static inline bool dm_zone_plug_bio(struct mapped_device *md, struct bio *bio) { - return dm_emulate_zone_append(md) && blk_zone_plug_bio(bio, 0); + if (!bio_needs_zone_write_plugging(bio)) + return false; + return blk_zone_plug_bio(bio, 0); } static blk_status_t __send_zone_reset_all_emulated(struct clone_info *ci, @@ -1914,8 +1920,7 @@ static blk_status_t __send_zone_reset_all(struct clone_info *ci) } #else -static inline bool dm_zone_bio_needs_split(struct mapped_device *md, - struct bio *bio) +static inline bool dm_zone_bio_needs_split(struct bio *bio) { return false; } @@ -1942,9 +1947,7 @@ static void dm_split_and_process_bio(struct mapped_device *md, is_abnormal = is_abnormal_io(bio); if (static_branch_unlikely(&zoned_enabled)) { - /* Special case REQ_OP_ZONE_RESET_ALL as it cannot be split. */ - need_split = (bio_op(bio) != REQ_OP_ZONE_RESET_ALL) && - (is_abnormal || dm_zone_bio_needs_split(md, bio)); + need_split = is_abnormal || dm_zone_bio_needs_split(bio); } else { need_split = is_abnormal; } @@ -2417,21 +2420,35 @@ static struct dm_table *__bind(struct mapped_device *md, struct dm_table *t, struct queue_limits *limits) { struct dm_table *old_map; - sector_t size; + sector_t size, old_size; int ret; lockdep_assert_held(&md->suspend_lock); size = dm_table_get_size(t); + old_size = dm_get_size(md); + + if (!dm_table_supports_size_change(t, old_size, size)) { + old_map = ERR_PTR(-EINVAL); + goto out; + } + + set_capacity(md->disk, size); + + ret = dm_table_set_restrictions(t, md->queue, limits); + if (ret) { + set_capacity(md->disk, old_size); + old_map = ERR_PTR(ret); + goto out; + } + /* * Wipe any geometry if the size of the table changed. */ - if (size != dm_get_size(md)) + if (size != old_size) memset(&md->geometry, 0, sizeof(md->geometry)); - set_capacity(md->disk, size); - dm_table_event_callback(t, event_callback, md); if (dm_table_request_based(t)) { @@ -2449,10 +2466,10 @@ static struct dm_table *__bind(struct mapped_device *md, struct dm_table *t, * requests in the queue may refer to bio from the old bioset, * so you must walk through the queue to unprep. */ - if (!md->mempools) { + if (!md->mempools) md->mempools = t->mempools; - t->mempools = NULL; - } + else + dm_free_md_mempools(t->mempools); } else { /* * The md may already have mempools that need changing. @@ -2461,14 +2478,8 @@ static struct dm_table *__bind(struct mapped_device *md, struct dm_table *t, */ dm_free_md_mempools(md->mempools); md->mempools = t->mempools; - t->mempools = NULL; - } - - ret = dm_table_set_restrictions(t, md->queue, limits); - if (ret) { - old_map = ERR_PTR(ret); - goto out; } + t->mempools = NULL; old_map = rcu_dereference_protected(md->map, lockdep_is_held(&md->suspend_lock)); rcu_assign_pointer(md->map, (void *)t); @@ -3634,10 +3645,13 @@ static int dm_pr_clear(struct block_device *bdev, u64 key) struct mapped_device *md = bdev->bd_disk->private_data; const struct pr_ops *ops; int r, srcu_idx; + bool forward = true; - r = dm_prepare_ioctl(md, &srcu_idx, &bdev); + /* Not a real ioctl, but targets must not interpret non-DM ioctls */ + r = dm_prepare_ioctl(md, &srcu_idx, &bdev, 0, 0, &forward); if (r < 0) goto out; + WARN_ON_ONCE(!forward); ops = bdev->bd_disk->fops->pr_ops; if (ops && ops->pr_clear) diff --git a/drivers/md/dm.h b/drivers/md/dm.h index a0a8ff119815..245f52b59215 100644 --- a/drivers/md/dm.h +++ b/drivers/md/dm.h @@ -58,6 +58,7 @@ void dm_table_event_callback(struct dm_table *t, void (*fn)(void *), void *context); struct dm_target *dm_table_find_target(struct dm_table *t, sector_t sector); bool dm_table_has_no_data_devices(struct dm_table *table); +bool dm_table_is_wildcard(struct dm_table *t); int dm_calculate_queue_limits(struct dm_table *table, struct queue_limits *limits); int dm_table_set_restrictions(struct dm_table *t, struct request_queue *q, @@ -72,6 +73,8 @@ struct target_type *dm_table_get_immutable_target_type(struct dm_table *t); struct dm_target *dm_table_get_immutable_target(struct dm_table *t); struct dm_target *dm_table_get_wildcard_target(struct dm_table *t); bool dm_table_request_based(struct dm_table *t); +bool dm_table_supports_size_change(struct dm_table *t, sector_t old_size, + sector_t new_size); void dm_lock_md_type(struct mapped_device *md); void dm_unlock_md_type(struct mapped_device *md); @@ -102,6 +105,7 @@ int dm_setup_md_queue(struct mapped_device *md, struct dm_table *t); int dm_set_zones_restrictions(struct dm_table *t, struct request_queue *q, struct queue_limits *lim); int dm_revalidate_zones(struct dm_table *t, struct request_queue *q); +void dm_finalize_zone_settings(struct dm_table *t, struct queue_limits *lim); void dm_zone_endio(struct dm_io *io, struct bio *clone); #ifdef CONFIG_BLK_DEV_ZONED int dm_blk_report_zones(struct gendisk *disk, sector_t sector, @@ -110,12 +114,14 @@ bool dm_is_zone_write(struct mapped_device *md, struct bio *bio); int dm_zone_get_reset_bitmap(struct mapped_device *md, struct dm_table *t, sector_t sector, unsigned int nr_zones, unsigned long *need_reset); +#define dm_has_zone_plugs(md) ((md)->disk->zone_wplugs_hash != NULL) #else #define dm_blk_report_zones NULL static inline bool dm_is_zone_write(struct mapped_device *md, struct bio *bio) { return false; } +#define dm_has_zone_plugs(md) false #endif /* diff --git a/drivers/md/md-bitmap.c b/drivers/md/md-bitmap.c index 27409d05f053..7f524a26cebc 100644 --- a/drivers/md/md-bitmap.c +++ b/drivers/md/md-bitmap.c @@ -29,8 +29,10 @@ #include <linux/buffer_head.h> #include <linux/seq_file.h> #include <trace/events/block.h> + #include "md.h" #include "md-bitmap.h" +#include "md-cluster.h" #define BITMAP_MAJOR_LO 3 /* version 4 insists the bitmap is in little-endian order @@ -103,9 +105,19 @@ * */ +typedef __u16 bitmap_counter_t; + #define PAGE_BITS (PAGE_SIZE << 3) #define PAGE_BIT_SHIFT (PAGE_SHIFT + 3) +#define COUNTER_BITS 16 +#define COUNTER_BIT_SHIFT 4 +#define COUNTER_BYTE_SHIFT (COUNTER_BIT_SHIFT - 3) + +#define NEEDED_MASK ((bitmap_counter_t) (1 << (COUNTER_BITS - 1))) +#define RESYNC_MASK ((bitmap_counter_t) (1 << (COUNTER_BITS - 2))) +#define COUNTER_MAX ((bitmap_counter_t) RESYNC_MASK - 1) + #define NEEDED(x) (((bitmap_counter_t) x) & NEEDED_MASK) #define RESYNC(x) (((bitmap_counter_t) x) & RESYNC_MASK) #define COUNTER(x) (((bitmap_counter_t) x) & COUNTER_MAX) @@ -787,7 +799,7 @@ static int md_bitmap_new_disk_sb(struct bitmap *bitmap) * is a good choice? We choose COUNTER_MAX / 2 arbitrarily. */ write_behind = bitmap->mddev->bitmap_info.max_write_behind; - if (write_behind > COUNTER_MAX) + if (write_behind > COUNTER_MAX / 2) write_behind = COUNTER_MAX / 2; sb->write_behind = cpu_to_le32(write_behind); bitmap->mddev->bitmap_info.max_write_behind = write_behind; @@ -942,7 +954,7 @@ out: bmname(bitmap), err); goto out_no_sb; } - bitmap->cluster_slot = md_cluster_ops->slot_number(bitmap->mddev); + bitmap->cluster_slot = bitmap->mddev->cluster_ops->slot_number(bitmap->mddev); goto re_read; } @@ -1670,13 +1682,13 @@ __acquires(bitmap->lock) &(bitmap->bp[page].map[pageoff]); } -static int bitmap_startwrite(struct mddev *mddev, sector_t offset, - unsigned long sectors) +static void bitmap_start_write(struct mddev *mddev, sector_t offset, + unsigned long sectors) { struct bitmap *bitmap = mddev->bitmap; if (!bitmap) - return 0; + return; while (sectors) { sector_t blocks; @@ -1686,7 +1698,7 @@ static int bitmap_startwrite(struct mddev *mddev, sector_t offset, bmc = md_bitmap_get_counter(&bitmap->counts, offset, &blocks, 1); if (!bmc) { spin_unlock_irq(&bitmap->counts.lock); - return 0; + return; } if (unlikely(COUNTER(*bmc) == COUNTER_MAX)) { @@ -1722,11 +1734,10 @@ static int bitmap_startwrite(struct mddev *mddev, sector_t offset, else sectors = 0; } - return 0; } -static void bitmap_endwrite(struct mddev *mddev, sector_t offset, - unsigned long sectors) +static void bitmap_end_write(struct mddev *mddev, sector_t offset, + unsigned long sectors) { struct bitmap *bitmap = mddev->bitmap; @@ -2021,7 +2032,7 @@ static void md_bitmap_free(void *data) sysfs_put(bitmap->sysfs_can_clear); if (mddev_is_clustered(bitmap->mddev) && bitmap->mddev->cluster_info && - bitmap->cluster_slot == md_cluster_ops->slot_number(bitmap->mddev)) + bitmap->cluster_slot == bitmap->mddev->cluster_ops->slot_number(bitmap->mddev)) md_cluster_stop(bitmap->mddev); /* Shouldn't be needed - but just in case.... */ @@ -2203,9 +2214,9 @@ static struct bitmap *__bitmap_create(struct mddev *mddev, int slot) return ERR_PTR(err); } -static int bitmap_create(struct mddev *mddev, int slot) +static int bitmap_create(struct mddev *mddev) { - struct bitmap *bitmap = __bitmap_create(mddev, slot); + struct bitmap *bitmap = __bitmap_create(mddev, -1); if (IS_ERR(bitmap)) return PTR_ERR(bitmap); @@ -2229,7 +2240,7 @@ static int bitmap_load(struct mddev *mddev) mddev_create_serial_pool(mddev, rdev); if (mddev_is_clustered(mddev)) - md_cluster_ops->load_bitmaps(mddev, mddev->bitmap_info.nodes); + mddev->cluster_ops->load_bitmaps(mddev, mddev->bitmap_info.nodes); /* Clear out old bitmap info first: Either there is none, or we * are resuming after someone else has possibly changed things, @@ -2355,8 +2366,7 @@ static int bitmap_get_stats(void *data, struct md_bitmap_stats *stats) if (!bitmap) return -ENOENT; - if (!bitmap->mddev->bitmap_info.external && - !bitmap->storage.sb_page) + if (!bitmap->storage.sb_page) return -EINVAL; sb = kmap_local_page(bitmap->storage.sb_page); stats->sync_size = le64_to_cpu(sb->sync_size); @@ -2668,7 +2678,7 @@ location_store(struct mddev *mddev, const char *buf, size_t len) } mddev->bitmap_info.offset = offset; - rv = bitmap_create(mddev, -1); + rv = bitmap_create(mddev); if (rv) goto out; @@ -3001,8 +3011,8 @@ static struct bitmap_operations bitmap_ops = { .end_behind_write = bitmap_end_behind_write, .wait_behind_writes = bitmap_wait_behind_writes, - .startwrite = bitmap_startwrite, - .endwrite = bitmap_endwrite, + .start_write = bitmap_start_write, + .end_write = bitmap_end_write, .start_sync = bitmap_start_sync, .end_sync = bitmap_end_sync, .cond_end_sync = bitmap_cond_end_sync, diff --git a/drivers/md/md-bitmap.h b/drivers/md/md-bitmap.h index 31c93019c76b..59e9dd45cfde 100644 --- a/drivers/md/md-bitmap.h +++ b/drivers/md/md-bitmap.h @@ -9,15 +9,6 @@ #define BITMAP_MAGIC 0x6d746962 -typedef __u16 bitmap_counter_t; -#define COUNTER_BITS 16 -#define COUNTER_BIT_SHIFT 4 -#define COUNTER_BYTE_SHIFT (COUNTER_BIT_SHIFT - 3) - -#define NEEDED_MASK ((bitmap_counter_t) (1 << (COUNTER_BITS - 1))) -#define RESYNC_MASK ((bitmap_counter_t) (1 << (COUNTER_BITS - 2))) -#define COUNTER_MAX ((bitmap_counter_t) RESYNC_MASK - 1) - /* use these for bitmap->flags and bitmap->sb->state bit-fields */ enum bitmap_state { BITMAP_STALE = 1, /* the bitmap file is out of date or had -EIO */ @@ -72,7 +63,7 @@ struct md_bitmap_stats { struct bitmap_operations { bool (*enabled)(struct mddev *mddev); - int (*create)(struct mddev *mddev, int slot); + int (*create)(struct mddev *mddev); int (*resize)(struct mddev *mddev, sector_t blocks, int chunksize, bool init); @@ -89,10 +80,10 @@ struct bitmap_operations { void (*end_behind_write)(struct mddev *mddev); void (*wait_behind_writes)(struct mddev *mddev); - int (*startwrite)(struct mddev *mddev, sector_t offset, + void (*start_write)(struct mddev *mddev, sector_t offset, + unsigned long sectors); + void (*end_write)(struct mddev *mddev, sector_t offset, unsigned long sectors); - void (*endwrite)(struct mddev *mddev, sector_t offset, - unsigned long sectors); bool (*start_sync)(struct mddev *mddev, sector_t offset, sector_t *blocks, bool degraded); void (*end_sync)(struct mddev *mddev, sector_t offset, sector_t *blocks); diff --git a/drivers/md/md-cluster.c b/drivers/md/md-cluster.c index 6595f89becdb..94221d964d4f 100644 --- a/drivers/md/md-cluster.c +++ b/drivers/md/md-cluster.c @@ -1166,7 +1166,7 @@ static int resize_bitmaps(struct mddev *mddev, sector_t newsize, sector_t oldsiz struct dlm_lock_resource *bm_lockres; char str[64]; - if (i == md_cluster_ops->slot_number(mddev)) + if (i == slot_number(mddev)) continue; bitmap = mddev->bitmap_ops->get_from_slot(mddev, i); @@ -1216,7 +1216,7 @@ out: */ static int cluster_check_sync_size(struct mddev *mddev) { - int current_slot = md_cluster_ops->slot_number(mddev); + int current_slot = slot_number(mddev); int node_num = mddev->bitmap_info.nodes; struct dlm_lock_resource *bm_lockres; struct md_bitmap_stats stats; @@ -1612,7 +1612,14 @@ out: return err; } -static const struct md_cluster_operations cluster_ops = { +static struct md_cluster_operations cluster_ops = { + .head = { + .type = MD_CLUSTER, + .id = ID_CLUSTER, + .name = "cluster", + .owner = THIS_MODULE, + }, + .join = join, .leave = leave, .slot_number = slot_number, @@ -1642,13 +1649,12 @@ static int __init cluster_init(void) { pr_warn("md-cluster: support raid1 and raid10 (limited support)\n"); pr_info("Registering Cluster MD functions\n"); - register_md_cluster_operations(&cluster_ops, THIS_MODULE); - return 0; + return register_md_submodule(&cluster_ops.head); } static void cluster_exit(void) { - unregister_md_cluster_operations(); + unregister_md_submodule(&cluster_ops.head); } module_init(cluster_init); diff --git a/drivers/md/md-cluster.h b/drivers/md/md-cluster.h index 470bf18ffde5..8fb06d853173 100644 --- a/drivers/md/md-cluster.h +++ b/drivers/md/md-cluster.h @@ -10,6 +10,8 @@ struct mddev; struct md_rdev; struct md_cluster_operations { + struct md_submodule_head head; + int (*join)(struct mddev *mddev, int nodes); int (*leave)(struct mddev *mddev); int (*slot_number)(struct mddev *mddev); @@ -35,4 +37,8 @@ struct md_cluster_operations { void (*update_size)(struct mddev *mddev, sector_t old_dev_sectors); }; +extern int md_setup_cluster(struct mddev *mddev, int nodes); +extern void md_cluster_stop(struct mddev *mddev); +extern void md_reload_sb(struct mddev *mddev, int raid_disk); + #endif /* _MD_CLUSTER_H */ diff --git a/drivers/md/md-linear.c b/drivers/md/md-linear.c index 369aed044b40..5d9b08115375 100644 --- a/drivers/md/md-linear.c +++ b/drivers/md/md-linear.c @@ -5,7 +5,6 @@ */ #include <linux/blkdev.h> -#include <linux/raid/md_u.h> #include <linux/seq_file.h> #include <linux/module.h> #include <linux/slab.h> @@ -320,9 +319,13 @@ static void linear_quiesce(struct mddev *mddev, int state) } static struct md_personality linear_personality = { - .name = "linear", - .level = LEVEL_LINEAR, - .owner = THIS_MODULE, + .head = { + .type = MD_PERSONALITY, + .id = ID_LINEAR, + .name = "linear", + .owner = THIS_MODULE, + }, + .make_request = linear_make_request, .run = linear_run, .free = linear_free, @@ -335,12 +338,12 @@ static struct md_personality linear_personality = { static int __init linear_init(void) { - return register_md_personality(&linear_personality); + return register_md_submodule(&linear_personality.head); } static void linear_exit(void) { - unregister_md_personality(&linear_personality); + unregister_md_submodule(&linear_personality.head); } module_init(linear_init); diff --git a/drivers/md/md.c b/drivers/md/md.c index ef859ccb0366..10670c62b09e 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -79,16 +79,10 @@ static const char *action_name[NR_SYNC_ACTIONS] = { [ACTION_IDLE] = "idle", }; -/* pers_list is a list of registered personalities protected by pers_lock. */ -static LIST_HEAD(pers_list); -static DEFINE_SPINLOCK(pers_lock); +static DEFINE_XARRAY(md_submodule); static const struct kobj_type md_ktype; -const struct md_cluster_operations *md_cluster_ops; -EXPORT_SYMBOL(md_cluster_ops); -static struct module *md_cluster_mod; - static DECLARE_WAIT_QUEUE_HEAD(resync_wait); static struct workqueue_struct *md_wq; @@ -117,32 +111,48 @@ static void md_wakeup_thread_directly(struct md_thread __rcu *thread); /* Default safemode delay: 200 msec */ #define DEFAULT_SAFEMODE_DELAY ((200 * HZ)/1000 +1) /* - * Current RAID-1,4,5 parallel reconstruction 'guaranteed speed limit' - * is 1000 KB/sec, so the extra system load does not show up that much. - * Increase it if you want to have more _guaranteed_ speed. Note that - * the RAID driver will use the maximum available bandwidth if the IO - * subsystem is idle. There is also an 'absolute maximum' reconstruction - * speed limit - in case reconstruction slows down your system despite - * idle IO detection. + * Current RAID-1,4,5,6,10 parallel reconstruction 'guaranteed speed limit' + * is sysctl_speed_limit_min, 1000 KB/sec by default, so the extra system load + * does not show up that much. Increase it if you want to have more guaranteed + * speed. Note that the RAID driver will use the maximum bandwidth + * sysctl_speed_limit_max, 200 MB/sec by default, if the IO subsystem is idle. + * + * Background sync IO speed control: + * + * - below speed min: + * no limit; + * - above speed min and below speed max: + * a) if mddev is idle, then no limit; + * b) if mddev is busy handling normal IO, then limit inflight sync IO + * to sync_io_depth; + * - above speed max: + * sync IO can't be issued; * - * you can change it via /proc/sys/dev/raid/speed_limit_min and _max. - * or /sys/block/mdX/md/sync_speed_{min,max} + * Following configurations can be changed via /proc/sys/dev/raid/ for system + * or /sys/block/mdX/md/ for one array. */ - static int sysctl_speed_limit_min = 1000; static int sysctl_speed_limit_max = 200000; -static inline int speed_min(struct mddev *mddev) +static int sysctl_sync_io_depth = 32; + +static int speed_min(struct mddev *mddev) { return mddev->sync_speed_min ? mddev->sync_speed_min : sysctl_speed_limit_min; } -static inline int speed_max(struct mddev *mddev) +static int speed_max(struct mddev *mddev) { return mddev->sync_speed_max ? mddev->sync_speed_max : sysctl_speed_limit_max; } +static int sync_io_depth(struct mddev *mddev) +{ + return mddev->sync_io_depth ? + mddev->sync_io_depth : sysctl_sync_io_depth; +} + static void rdev_uninit_serial(struct md_rdev *rdev) { if (!test_and_clear_bit(CollisionCheck, &rdev->flags)) @@ -299,14 +309,21 @@ static const struct ctl_table raid_table[] = { .procname = "speed_limit_min", .data = &sysctl_speed_limit_min, .maxlen = sizeof(int), - .mode = S_IRUGO|S_IWUSR, + .mode = 0644, .proc_handler = proc_dointvec, }, { .procname = "speed_limit_max", .data = &sysctl_speed_limit_max, .maxlen = sizeof(int), - .mode = S_IRUGO|S_IWUSR, + .mode = 0644, + .proc_handler = proc_dointvec, + }, + { + .procname = "sync_io_depth", + .data = &sysctl_sync_io_depth, + .maxlen = sizeof(int), + .mode = 0644, .proc_handler = proc_dointvec, }, }; @@ -894,16 +911,40 @@ struct md_rdev *md_find_rdev_rcu(struct mddev *mddev, dev_t dev) } EXPORT_SYMBOL_GPL(md_find_rdev_rcu); -static struct md_personality *find_pers(int level, char *clevel) +static struct md_personality *get_pers(int level, char *clevel) { - struct md_personality *pers; - list_for_each_entry(pers, &pers_list, list) { - if (level != LEVEL_NONE && pers->level == level) - return pers; - if (strcmp(pers->name, clevel)==0) - return pers; + struct md_personality *ret = NULL; + struct md_submodule_head *head; + unsigned long i; + + xa_lock(&md_submodule); + xa_for_each(&md_submodule, i, head) { + if (head->type != MD_PERSONALITY) + continue; + if ((level != LEVEL_NONE && head->id == level) || + !strcmp(head->name, clevel)) { + if (try_module_get(head->owner)) + ret = (void *)head; + break; + } } - return NULL; + xa_unlock(&md_submodule); + + if (!ret) { + if (level != LEVEL_NONE) + pr_warn("md: personality for level %d is not loaded!\n", + level); + else + pr_warn("md: personality for level %s is not loaded!\n", + clevel); + } + + return ret; +} + +static void put_pers(struct md_personality *pers) +{ + module_put(pers->head.owner); } /* return the offset of the super block in 512byte sectors */ @@ -1186,7 +1227,7 @@ int md_check_no_bitmap(struct mddev *mddev) if (!mddev->bitmap_info.file && !mddev->bitmap_info.offset) return 0; pr_warn("%s: bitmaps are not supported for %s\n", - mdname(mddev), mddev->pers->name); + mdname(mddev), mddev->pers->head.name); return 1; } EXPORT_SYMBOL(md_check_no_bitmap); @@ -2365,19 +2406,6 @@ int md_integrity_register(struct mddev *mddev) return 0; /* shouldn't register */ pr_debug("md: data integrity enabled on %s\n", mdname(mddev)); - if (bioset_integrity_create(&mddev->bio_set, BIO_POOL_SIZE) || - (mddev->level != 1 && mddev->level != 10 && - bioset_integrity_create(&mddev->io_clone_set, BIO_POOL_SIZE))) { - /* - * No need to handle the failure of bioset_integrity_create, - * because the function is called by md_run() -> pers->run(), - * md_run calls bioset_exit -> bioset_integrity_free in case - * of failure case. - */ - pr_err("md: failed to create integrity pool for %s\n", - mdname(mddev)); - return -EINVAL; - } return 0; } EXPORT_SYMBOL(md_integrity_register); @@ -2645,11 +2673,11 @@ repeat: force_change = 1; if (test_and_clear_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags)) nospares = 1; - ret = md_cluster_ops->metadata_update_start(mddev); + ret = mddev->cluster_ops->metadata_update_start(mddev); /* Has someone else has updated the sb */ if (!does_sb_need_changing(mddev)) { if (ret == 0) - md_cluster_ops->metadata_update_cancel(mddev); + mddev->cluster_ops->metadata_update_cancel(mddev); bit_clear_unless(&mddev->sb_flags, BIT(MD_SB_CHANGE_PENDING), BIT(MD_SB_CHANGE_DEVS) | BIT(MD_SB_CHANGE_CLEAN)); @@ -2789,7 +2817,7 @@ rewrite: /* if there was a failure, MD_SB_CHANGE_DEVS was set, and we re-write super */ if (mddev_is_clustered(mddev) && ret == 0) - md_cluster_ops->metadata_update_finish(mddev); + mddev->cluster_ops->metadata_update_finish(mddev); if (mddev->in_sync != sync_req || !bit_clear_unless(&mddev->sb_flags, BIT(MD_SB_CHANGE_PENDING), @@ -2948,7 +2976,7 @@ state_store(struct md_rdev *rdev, const char *buf, size_t len) else { err = 0; if (mddev_is_clustered(mddev)) - err = md_cluster_ops->remove_disk(mddev, rdev); + err = mddev->cluster_ops->remove_disk(mddev, rdev); if (err == 0) { md_kick_rdev_from_array(rdev); @@ -3058,7 +3086,7 @@ state_store(struct md_rdev *rdev, const char *buf, size_t len) * by this node eventually */ if (!mddev_is_clustered(rdev->mddev) || - (err = md_cluster_ops->gather_bitmaps(rdev)) == 0) { + (err = mddev->cluster_ops->gather_bitmaps(rdev)) == 0) { clear_bit(Faulty, &rdev->flags); err = add_bound_rdev(rdev); } @@ -3866,7 +3894,7 @@ level_show(struct mddev *mddev, char *page) spin_lock(&mddev->lock); p = mddev->pers; if (p) - ret = sprintf(page, "%s\n", p->name); + ret = sprintf(page, "%s\n", p->head.name); else if (mddev->clevel[0]) ret = sprintf(page, "%s\n", mddev->clevel); else if (mddev->level != LEVEL_NONE) @@ -3923,7 +3951,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len) rv = -EINVAL; if (!mddev->pers->quiesce) { pr_warn("md: %s: %s does not support online personality change\n", - mdname(mddev), mddev->pers->name); + mdname(mddev), mddev->pers->head.name); goto out_unlock; } @@ -3937,24 +3965,20 @@ level_store(struct mddev *mddev, const char *buf, size_t len) if (request_module("md-%s", clevel) != 0) request_module("md-level-%s", clevel); - spin_lock(&pers_lock); - pers = find_pers(level, clevel); - if (!pers || !try_module_get(pers->owner)) { - spin_unlock(&pers_lock); - pr_warn("md: personality %s not loaded\n", clevel); + pers = get_pers(level, clevel); + if (!pers) { rv = -EINVAL; goto out_unlock; } - spin_unlock(&pers_lock); if (pers == mddev->pers) { /* Nothing to do! */ - module_put(pers->owner); + put_pers(pers); rv = len; goto out_unlock; } if (!pers->takeover) { - module_put(pers->owner); + put_pers(pers); pr_warn("md: %s: %s does not support personality takeover\n", mdname(mddev), clevel); rv = -EINVAL; @@ -3975,7 +3999,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len) mddev->raid_disks -= mddev->delta_disks; mddev->delta_disks = 0; mddev->reshape_backwards = 0; - module_put(pers->owner); + put_pers(pers); pr_warn("md: %s: %s would not accept array\n", mdname(mddev), clevel); rv = PTR_ERR(priv); @@ -3990,7 +4014,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len) oldpriv = mddev->private; mddev->pers = pers; mddev->private = priv; - strscpy(mddev->clevel, pers->name, sizeof(mddev->clevel)); + strscpy(mddev->clevel, pers->head.name, sizeof(mddev->clevel)); mddev->level = mddev->new_level; mddev->layout = mddev->new_layout; mddev->chunk_sectors = mddev->new_chunk_sectors; @@ -4032,7 +4056,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len) mddev->to_remove = &md_redundancy_group; } - module_put(oldpers->owner); + put_pers(oldpers); rdev_for_each(rdev, mddev) { if (rdev->raid_disk < 0) @@ -4063,7 +4087,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len) * it must always be in_sync */ mddev->in_sync = 1; - del_timer_sync(&mddev->safemode_timer); + timer_delete_sync(&mddev->safemode_timer); } pers->run(mddev); set_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags); @@ -5090,7 +5114,7 @@ static ssize_t sync_min_show(struct mddev *mddev, char *page) { return sprintf(page, "%d (%s)\n", speed_min(mddev), - mddev->sync_speed_min ? "local": "system"); + mddev->sync_speed_min ? "local" : "system"); } static ssize_t @@ -5099,7 +5123,7 @@ sync_min_store(struct mddev *mddev, const char *buf, size_t len) unsigned int min; int rv; - if (strncmp(buf, "system", 6)==0) { + if (strncmp(buf, "system", 6) == 0) { min = 0; } else { rv = kstrtouint(buf, 10, &min); @@ -5119,7 +5143,7 @@ static ssize_t sync_max_show(struct mddev *mddev, char *page) { return sprintf(page, "%d (%s)\n", speed_max(mddev), - mddev->sync_speed_max ? "local": "system"); + mddev->sync_speed_max ? "local" : "system"); } static ssize_t @@ -5128,7 +5152,7 @@ sync_max_store(struct mddev *mddev, const char *buf, size_t len) unsigned int max; int rv; - if (strncmp(buf, "system", 6)==0) { + if (strncmp(buf, "system", 6) == 0) { max = 0; } else { rv = kstrtouint(buf, 10, &max); @@ -5145,6 +5169,35 @@ static struct md_sysfs_entry md_sync_max = __ATTR(sync_speed_max, S_IRUGO|S_IWUSR, sync_max_show, sync_max_store); static ssize_t +sync_io_depth_show(struct mddev *mddev, char *page) +{ + return sprintf(page, "%d (%s)\n", sync_io_depth(mddev), + mddev->sync_io_depth ? "local" : "system"); +} + +static ssize_t +sync_io_depth_store(struct mddev *mddev, const char *buf, size_t len) +{ + unsigned int max; + int rv; + + if (strncmp(buf, "system", 6) == 0) { + max = 0; + } else { + rv = kstrtouint(buf, 10, &max); + if (rv < 0) + return rv; + if (max == 0) + return -EINVAL; + } + mddev->sync_io_depth = max; + return len; +} + +static struct md_sysfs_entry md_sync_io_depth = +__ATTR_RW(sync_io_depth); + +static ssize_t degraded_show(struct mddev *mddev, char *page) { return sprintf(page, "%d\n", mddev->degraded); @@ -5590,7 +5643,7 @@ __ATTR(fail_last_dev, S_IRUGO | S_IWUSR, fail_last_dev_show, static ssize_t serialize_policy_show(struct mddev *mddev, char *page) { - if (mddev->pers == NULL || (mddev->pers->level != 1)) + if (mddev->pers == NULL || (mddev->pers->head.id != ID_RAID1)) return sprintf(page, "n/a\n"); else return sprintf(page, "%d\n", mddev->serialize_policy); @@ -5616,7 +5669,7 @@ serialize_policy_store(struct mddev *mddev, const char *buf, size_t len) err = mddev_suspend_and_lock(mddev); if (err) return err; - if (mddev->pers == NULL || (mddev->pers->level != 1)) { + if (mddev->pers == NULL || (mddev->pers->head.id != ID_RAID1)) { pr_err("md: serialize_policy is only effective for raid1\n"); err = -EINVAL; goto unlock; @@ -5670,6 +5723,7 @@ static struct attribute *md_redundancy_attrs[] = { &md_mismatches.attr, &md_sync_min.attr, &md_sync_max.attr, + &md_sync_io_depth.attr, &md_sync_speed.attr, &md_sync_force_parallel.attr, &md_sync_completed.attr, @@ -5994,7 +6048,7 @@ static int add_named_array(const char *val, const struct kernel_param *kp) static void md_safemode_timeout(struct timer_list *t) { - struct mddev *mddev = from_timer(mddev, t, safemode_timer); + struct mddev *mddev = timer_container_of(mddev, t, safemode_timer); mddev->safemode = 1; if (mddev->external) @@ -6102,30 +6156,21 @@ int md_run(struct mddev *mddev) goto exit_sync_set; } - spin_lock(&pers_lock); - pers = find_pers(mddev->level, mddev->clevel); - if (!pers || !try_module_get(pers->owner)) { - spin_unlock(&pers_lock); - if (mddev->level != LEVEL_NONE) - pr_warn("md: personality for level %d is not loaded!\n", - mddev->level); - else - pr_warn("md: personality for level %s is not loaded!\n", - mddev->clevel); + pers = get_pers(mddev->level, mddev->clevel); + if (!pers) { err = -EINVAL; goto abort; } - spin_unlock(&pers_lock); - if (mddev->level != pers->level) { - mddev->level = pers->level; - mddev->new_level = pers->level; + if (mddev->level != pers->head.id) { + mddev->level = pers->head.id; + mddev->new_level = pers->head.id; } - strscpy(mddev->clevel, pers->name, sizeof(mddev->clevel)); + strscpy(mddev->clevel, pers->head.name, sizeof(mddev->clevel)); if (mddev->reshape_position != MaxSector && pers->start_reshape == NULL) { /* This personality cannot handle reshaping... */ - module_put(pers->owner); + put_pers(pers); err = -EINVAL; goto abort; } @@ -6180,7 +6225,7 @@ int md_run(struct mddev *mddev) } if (err == 0 && pers->sync_request && (mddev->bitmap_info.file || mddev->bitmap_info.offset)) { - err = mddev->bitmap_ops->create(mddev, -1); + err = mddev->bitmap_ops->create(mddev); if (err) pr_warn("%s: failed to create bitmap (%d)\n", mdname(mddev), err); @@ -6252,7 +6297,7 @@ bitmap_abort: if (mddev->private) pers->free(mddev, mddev->private); mddev->private = NULL; - module_put(pers->owner); + put_pers(pers); mddev->bitmap_ops->destroy(mddev); abort: bioset_exit(&mddev->io_clone_set); @@ -6413,7 +6458,7 @@ static void md_clean(struct mddev *mddev) static void __md_stop_writes(struct mddev *mddev) { - del_timer_sync(&mddev->safemode_timer); + timer_delete_sync(&mddev->safemode_timer); if (mddev->pers && mddev->pers->quiesce) { mddev->pers->quiesce(mddev, 1); @@ -6473,7 +6518,7 @@ static void __md_stop(struct mddev *mddev) mddev->private = NULL; if (pers->sync_request && mddev->to_remove == NULL) mddev->to_remove = &md_redundancy_group; - module_put(pers->owner); + put_pers(pers); clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery); bioset_exit(&mddev->bio_set); @@ -6989,7 +7034,7 @@ int md_add_new_disk(struct mddev *mddev, struct mdu_disk_info_s *info) set_bit(Candidate, &rdev->flags); else if (info->state & (1 << MD_DISK_CLUSTER_ADD)) { /* --add initiated by this node */ - err = md_cluster_ops->add_new_disk(mddev, rdev); + err = mddev->cluster_ops->add_new_disk(mddev, rdev); if (err) { export_rdev(rdev, mddev); return err; @@ -7006,14 +7051,14 @@ int md_add_new_disk(struct mddev *mddev, struct mdu_disk_info_s *info) if (mddev_is_clustered(mddev)) { if (info->state & (1 << MD_DISK_CANDIDATE)) { if (!err) { - err = md_cluster_ops->new_disk_ack(mddev, - err == 0); + err = mddev->cluster_ops->new_disk_ack( + mddev, err == 0); if (err) md_kick_rdev_from_array(rdev); } } else { if (err) - md_cluster_ops->add_new_disk_cancel(mddev); + mddev->cluster_ops->add_new_disk_cancel(mddev); else err = add_bound_rdev(rdev); } @@ -7093,10 +7138,9 @@ static int hot_remove_disk(struct mddev *mddev, dev_t dev) goto busy; kick_rdev: - if (mddev_is_clustered(mddev)) { - if (md_cluster_ops->remove_disk(mddev, rdev)) - goto busy; - } + if (mddev_is_clustered(mddev) && + mddev->cluster_ops->remove_disk(mddev, rdev)) + goto busy; md_kick_rdev_from_array(rdev); set_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags); @@ -7241,7 +7285,7 @@ static int set_bitmap_file(struct mddev *mddev, int fd) err = 0; if (mddev->pers) { if (fd >= 0) { - err = mddev->bitmap_ops->create(mddev, -1); + err = mddev->bitmap_ops->create(mddev); if (!err) err = mddev->bitmap_ops->load(mddev); @@ -7399,7 +7443,7 @@ static int update_size(struct mddev *mddev, sector_t num_sectors) rv = mddev->pers->resize(mddev, num_sectors); if (!rv) { if (mddev_is_clustered(mddev)) - md_cluster_ops->update_size(mddev, old_dev_sectors); + mddev->cluster_ops->update_size(mddev, old_dev_sectors); else if (!mddev_is_dm(mddev)) set_capacity_and_notify(mddev->gendisk, mddev->array_sectors); @@ -7447,6 +7491,28 @@ static int update_raid_disks(struct mddev *mddev, int raid_disks) return rv; } +static int get_cluster_ops(struct mddev *mddev) +{ + xa_lock(&md_submodule); + mddev->cluster_ops = xa_load(&md_submodule, ID_CLUSTER); + if (mddev->cluster_ops && + !try_module_get(mddev->cluster_ops->head.owner)) + mddev->cluster_ops = NULL; + xa_unlock(&md_submodule); + + return mddev->cluster_ops == NULL ? -ENOENT : 0; +} + +static void put_cluster_ops(struct mddev *mddev) +{ + if (!mddev->cluster_ops) + return; + + mddev->cluster_ops->leave(mddev); + module_put(mddev->cluster_ops->head.owner); + mddev->cluster_ops = NULL; +} + /* * update_array_info is used to change the configuration of an * on-line array. @@ -7535,7 +7601,7 @@ static int update_array_info(struct mddev *mddev, mdu_array_info_t *info) mddev->bitmap_info.default_offset; mddev->bitmap_info.space = mddev->bitmap_info.default_space; - rv = mddev->bitmap_ops->create(mddev, -1); + rv = mddev->bitmap_ops->create(mddev); if (!rv) rv = mddev->bitmap_ops->load(mddev); @@ -7555,16 +7621,15 @@ static int update_array_info(struct mddev *mddev, mdu_array_info_t *info) if (mddev->bitmap_info.nodes) { /* hold PW on all the bitmap lock */ - if (md_cluster_ops->lock_all_bitmaps(mddev) <= 0) { + if (mddev->cluster_ops->lock_all_bitmaps(mddev) <= 0) { pr_warn("md: can't change bitmap to none since the array is in use by more than one node\n"); rv = -EPERM; - md_cluster_ops->unlock_all_bitmaps(mddev); + mddev->cluster_ops->unlock_all_bitmaps(mddev); goto err; } mddev->bitmap_info.nodes = 0; - md_cluster_ops->leave(mddev); - module_put(md_cluster_mod); + put_cluster_ops(mddev); mddev->safemode_delay = DEFAULT_SAFEMODE_DELAY; } mddev->bitmap_ops->destroy(mddev); @@ -7848,7 +7913,7 @@ static int md_ioctl(struct block_device *bdev, blk_mode_t mode, case CLUSTERED_DISK_NACK: if (mddev_is_clustered(mddev)) - md_cluster_ops->new_disk_ack(mddev, false); + mddev->cluster_ops->new_disk_ack(mddev, false); else err = -EINVAL; goto unlock; @@ -8130,7 +8195,8 @@ void md_error(struct mddev *mddev, struct md_rdev *rdev) return; mddev->pers->error_handler(mddev, rdev); - if (mddev->pers->level == 0 || mddev->pers->level == LEVEL_LINEAR) + if (mddev->pers->head.id == ID_RAID0 || + mddev->pers->head.id == ID_LINEAR) return; if (mddev->degraded && !test_bit(MD_BROKEN, &mddev->flags)) @@ -8168,14 +8234,17 @@ static void status_unused(struct seq_file *seq) static void status_personalities(struct seq_file *seq) { - struct md_personality *pers; + struct md_submodule_head *head; + unsigned long i; seq_puts(seq, "Personalities : "); - spin_lock(&pers_lock); - list_for_each_entry(pers, &pers_list, list) - seq_printf(seq, "[%s] ", pers->name); - spin_unlock(&pers_lock); + xa_lock(&md_submodule); + xa_for_each(&md_submodule, i, head) + if (head->type == MD_PERSONALITY) + seq_printf(seq, "[%s] ", head->name); + xa_unlock(&md_submodule); + seq_puts(seq, "\n"); } @@ -8398,7 +8467,7 @@ static int md_seq_show(struct seq_file *seq, void *v) seq_printf(seq, " (read-only)"); if (mddev->ro == MD_AUTO_READ) seq_printf(seq, " (auto-read-only)"); - seq_printf(seq, " %s", mddev->pers->name); + seq_printf(seq, " %s", mddev->pers->head.name); } else { seq_printf(seq, "inactive"); } @@ -8518,67 +8587,34 @@ static const struct proc_ops mdstat_proc_ops = { .proc_poll = mdstat_poll, }; -int register_md_personality(struct md_personality *p) -{ - pr_debug("md: %s personality registered for level %d\n", - p->name, p->level); - spin_lock(&pers_lock); - list_add_tail(&p->list, &pers_list); - spin_unlock(&pers_lock); - return 0; -} -EXPORT_SYMBOL(register_md_personality); - -int unregister_md_personality(struct md_personality *p) +int register_md_submodule(struct md_submodule_head *msh) { - pr_debug("md: %s personality unregistered\n", p->name); - spin_lock(&pers_lock); - list_del_init(&p->list); - spin_unlock(&pers_lock); - return 0; + return xa_insert(&md_submodule, msh->id, msh, GFP_KERNEL); } -EXPORT_SYMBOL(unregister_md_personality); +EXPORT_SYMBOL_GPL(register_md_submodule); -int register_md_cluster_operations(const struct md_cluster_operations *ops, - struct module *module) +void unregister_md_submodule(struct md_submodule_head *msh) { - int ret = 0; - spin_lock(&pers_lock); - if (md_cluster_ops != NULL) - ret = -EALREADY; - else { - md_cluster_ops = ops; - md_cluster_mod = module; - } - spin_unlock(&pers_lock); - return ret; + xa_erase(&md_submodule, msh->id); } -EXPORT_SYMBOL(register_md_cluster_operations); - -int unregister_md_cluster_operations(void) -{ - spin_lock(&pers_lock); - md_cluster_ops = NULL; - spin_unlock(&pers_lock); - return 0; -} -EXPORT_SYMBOL(unregister_md_cluster_operations); +EXPORT_SYMBOL_GPL(unregister_md_submodule); int md_setup_cluster(struct mddev *mddev, int nodes) { - int ret; - if (!md_cluster_ops) + int ret = get_cluster_ops(mddev); + + if (ret) { request_module("md-cluster"); - spin_lock(&pers_lock); + ret = get_cluster_ops(mddev); + } + /* ensure module won't be unloaded */ - if (!md_cluster_ops || !try_module_get(md_cluster_mod)) { + if (ret) { pr_warn("can't find md-cluster module or get its reference.\n"); - spin_unlock(&pers_lock); - return -ENOENT; + return ret; } - spin_unlock(&pers_lock); - ret = md_cluster_ops->join(mddev, nodes); + ret = mddev->cluster_ops->join(mddev, nodes); if (!ret) mddev->safemode_delay = 0; return ret; @@ -8586,56 +8622,58 @@ int md_setup_cluster(struct mddev *mddev, int nodes) void md_cluster_stop(struct mddev *mddev) { - if (!md_cluster_ops) - return; - md_cluster_ops->leave(mddev); - module_put(md_cluster_mod); + put_cluster_ops(mddev); } -static int is_mddev_idle(struct mddev *mddev, int init) +static bool is_rdev_holder_idle(struct md_rdev *rdev, bool init) +{ + unsigned long last_events = rdev->last_events; + + if (!bdev_is_partition(rdev->bdev)) + return true; + + /* + * If rdev is partition, and user doesn't issue IO to the array, the + * array is still not idle if user issues IO to other partitions. + */ + rdev->last_events = part_stat_read_accum(rdev->bdev->bd_disk->part0, + sectors) - + part_stat_read_accum(rdev->bdev, sectors); + + return init || rdev->last_events <= last_events; +} + +/* + * mddev is idle if following conditions are matched since last check: + * 1) mddev doesn't have normal IO completed; + * 2) mddev doesn't have inflight normal IO; + * 3) if any member disk is partition, and other partitions don't have IO + * completed; + * + * Noted this checking rely on IO accounting is enabled. + */ +static bool is_mddev_idle(struct mddev *mddev, int init) { + unsigned long last_events = mddev->normal_io_events; + struct gendisk *disk; struct md_rdev *rdev; - int idle; - int curr_events; + bool idle = true; - idle = 1; - rcu_read_lock(); - rdev_for_each_rcu(rdev, mddev) { - struct gendisk *disk = rdev->bdev->bd_disk; + disk = mddev_is_dm(mddev) ? mddev->dm_gendisk : mddev->gendisk; + if (!disk) + return true; - if (!init && !blk_queue_io_stat(disk->queue)) - continue; + mddev->normal_io_events = part_stat_read_accum(disk->part0, sectors); + if (!init && (mddev->normal_io_events > last_events || + bdev_count_inflight(disk->part0))) + idle = false; - curr_events = (int)part_stat_read_accum(disk->part0, sectors) - - atomic_read(&disk->sync_io); - /* sync IO will cause sync_io to increase before the disk_stats - * as sync_io is counted when a request starts, and - * disk_stats is counted when it completes. - * So resync activity will cause curr_events to be smaller than - * when there was no such activity. - * non-sync IO will cause disk_stat to increase without - * increasing sync_io so curr_events will (eventually) - * be larger than it was before. Once it becomes - * substantially larger, the test below will cause - * the array to appear non-idle, and resync will slow - * down. - * If there is a lot of outstanding resync activity when - * we set last_event to curr_events, then all that activity - * completing might cause the array to appear non-idle - * and resync will be slowed down even though there might - * not have been non-resync activity. This will only - * happen once though. 'last_events' will soon reflect - * the state where there is little or no outstanding - * resync requests, and further resync activity will - * always make curr_events less than last_events. - * - */ - if (init || curr_events - rdev->last_events > 64) { - rdev->last_events = curr_events; - idle = 0; - } - } + rcu_read_lock(); + rdev_for_each_rcu(rdev, mddev) + if (!is_rdev_holder_idle(rdev, init)) + idle = false; rcu_read_unlock(); + return idle; } @@ -8761,14 +8799,14 @@ static void md_bitmap_start(struct mddev *mddev, mddev->pers->bitmap_sector(mddev, &md_io_clone->offset, &md_io_clone->sectors); - mddev->bitmap_ops->startwrite(mddev, md_io_clone->offset, - md_io_clone->sectors); + mddev->bitmap_ops->start_write(mddev, md_io_clone->offset, + md_io_clone->sectors); } static void md_bitmap_end(struct mddev *mddev, struct md_io_clone *md_io_clone) { - mddev->bitmap_ops->endwrite(mddev, md_io_clone->offset, - md_io_clone->sectors); + mddev->bitmap_ops->end_write(mddev, md_io_clone->offset, + md_io_clone->sectors); } static void md_end_clone_io(struct bio *bio) @@ -8947,6 +8985,23 @@ static sector_t md_sync_position(struct mddev *mddev, enum sync_action action) } } +static bool sync_io_within_limit(struct mddev *mddev) +{ + int io_sectors; + + /* + * For raid456, sync IO is stripe(4k) per IO, for other levels, it's + * RESYNC_PAGES(64k) per IO. + */ + if (mddev->level == 4 || mddev->level == 5 || mddev->level == 6) + io_sectors = 8; + else + io_sectors = 128; + + return atomic_read(&mddev->recovery_active) < + io_sectors * sync_io_depth(mddev); +} + #define SYNC_MARKS 10 #define SYNC_MARK_STEP (3*HZ) #define UPDATE_FREQUENCY (5*60*HZ) @@ -8982,7 +9037,7 @@ void md_do_sync(struct md_thread *thread) } if (mddev_is_clustered(mddev)) { - ret = md_cluster_ops->resync_start(mddev); + ret = mddev->cluster_ops->resync_start(mddev); if (ret) goto skip; @@ -9009,7 +9064,7 @@ void md_do_sync(struct md_thread *thread) * */ if (mddev_is_clustered(mddev)) - md_cluster_ops->resync_start_notify(mddev); + mddev->cluster_ops->resync_start_notify(mddev); do { int mddev2_minor = -1; mddev->curr_resync = MD_RESYNC_DELAYED; @@ -9215,7 +9270,8 @@ void md_do_sync(struct md_thread *thread) msleep(500); goto repeat; } - if (!is_mddev_idle(mddev, 0)) { + if (!sync_io_within_limit(mddev) && + !is_mddev_idle(mddev, 0)) { /* * Give other IO more of a chance. * The faster the devices, the less we wait. @@ -9362,6 +9418,12 @@ static bool rdev_is_spare(struct md_rdev *rdev) static bool rdev_addable(struct md_rdev *rdev) { + struct mddev *mddev; + + mddev = READ_ONCE(rdev->mddev); + if (!mddev) + return false; + /* rdev is already used, don't add it again. */ if (test_bit(Candidate, &rdev->flags) || rdev->raid_disk >= 0 || test_bit(Faulty, &rdev->flags)) @@ -9372,7 +9434,7 @@ static bool rdev_addable(struct md_rdev *rdev) return true; /* Allow to add if array is read-write. */ - if (md_is_rdwr(rdev->mddev)) + if (md_is_rdwr(mddev)) return true; /* @@ -9400,17 +9462,11 @@ static bool md_spares_need_change(struct mddev *mddev) return false; } -static int remove_and_add_spares(struct mddev *mddev, - struct md_rdev *this) +static int remove_spares(struct mddev *mddev, struct md_rdev *this) { struct md_rdev *rdev; - int spares = 0; int removed = 0; - if (this && test_bit(MD_RECOVERY_RUNNING, &mddev->recovery)) - /* Mustn't remove devices when resync thread is running */ - return 0; - rdev_for_each(rdev, mddev) { if ((this == NULL || rdev == this) && rdev_removeable(rdev) && !mddev->pers->hot_remove_disk(mddev, rdev)) { @@ -9424,6 +9480,21 @@ static int remove_and_add_spares(struct mddev *mddev, if (removed && mddev->kobj.sd) sysfs_notify_dirent_safe(mddev->sysfs_degraded); + return removed; +} + +static int remove_and_add_spares(struct mddev *mddev, + struct md_rdev *this) +{ + struct md_rdev *rdev; + int spares = 0; + int removed = 0; + + if (this && test_bit(MD_RECOVERY_RUNNING, &mddev->recovery)) + /* Mustn't remove devices when resync thread is running */ + return 0; + + removed = remove_spares(mddev, this); if (this && removed) goto no_add; @@ -9466,6 +9537,7 @@ static bool md_choose_sync_action(struct mddev *mddev, int *spares) /* Check if resync is in progress. */ if (mddev->recovery_cp < MaxSector) { + remove_spares(mddev, NULL); set_bit(MD_RECOVERY_SYNC, &mddev->recovery); clear_bit(MD_RECOVERY_RECOVER, &mddev->recovery); return true; @@ -9702,8 +9774,8 @@ void md_check_recovery(struct mddev *mddev) * remove disk. */ rdev_for_each_safe(rdev, tmp, mddev) { - if (test_and_clear_bit(ClusterRemove, &rdev->flags) && - rdev->raid_disk < 0) + if (rdev->raid_disk < 0 && + test_and_clear_bit(ClusterRemove, &rdev->flags)) md_kick_rdev_from_array(rdev); } } @@ -9793,7 +9865,7 @@ void md_reap_sync_thread(struct mddev *mddev) * call resync_finish here if MD_CLUSTER_RESYNC_LOCKED is set by * clustered raid */ if (test_and_clear_bit(MD_CLUSTER_RESYNC_LOCKED, &mddev->flags)) - md_cluster_ops->resync_finish(mddev); + mddev->cluster_ops->resync_finish(mddev); clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery); clear_bit(MD_RECOVERY_DONE, &mddev->recovery); clear_bit(MD_RECOVERY_SYNC, &mddev->recovery); @@ -9801,13 +9873,13 @@ void md_reap_sync_thread(struct mddev *mddev) clear_bit(MD_RECOVERY_REQUESTED, &mddev->recovery); clear_bit(MD_RECOVERY_CHECK, &mddev->recovery); /* - * We call md_cluster_ops->update_size here because sync_size could + * We call mddev->cluster_ops->update_size here because sync_size could * be changed by md_update_sb, and MD_RECOVERY_RESHAPE is cleared, * so it is time to update size across cluster. */ if (mddev_is_clustered(mddev) && is_reshaped && !test_bit(MD_CLOSING, &mddev->flags)) - md_cluster_ops->update_size(mddev, old_dev_sectors); + mddev->cluster_ops->update_size(mddev, old_dev_sectors); /* flag recovery needed just to double check */ set_bit(MD_RECOVERY_NEEDED, &mddev->recovery); sysfs_notify_dirent_safe(mddev->sysfs_completed); @@ -9845,9 +9917,9 @@ EXPORT_SYMBOL(md_finish_reshape); /* Bad block management */ -/* Returns 1 on success, 0 on failure */ -int rdev_set_badblocks(struct md_rdev *rdev, sector_t s, int sectors, - int is_new) +/* Returns true on success, false on failure */ +bool rdev_set_badblocks(struct md_rdev *rdev, sector_t s, int sectors, + int is_new) { struct mddev *mddev = rdev->mddev; @@ -9859,7 +9931,7 @@ int rdev_set_badblocks(struct md_rdev *rdev, sector_t s, int sectors, * avoid it. */ if (test_bit(Faulty, &rdev->flags)) - return 1; + return true; if (is_new) s += rdev->new_data_offset; @@ -9867,7 +9939,7 @@ int rdev_set_badblocks(struct md_rdev *rdev, sector_t s, int sectors, s += rdev->data_offset; if (!badblocks_set(&rdev->badblocks, s, sectors, 0)) - return 0; + return false; /* Make sure they get written out promptly */ if (test_bit(ExternalBbl, &rdev->flags)) @@ -9876,12 +9948,12 @@ int rdev_set_badblocks(struct md_rdev *rdev, sector_t s, int sectors, set_mask_bits(&mddev->sb_flags, 0, BIT(MD_SB_CHANGE_CLEAN) | BIT(MD_SB_CHANGE_PENDING)); md_wakeup_thread(rdev->mddev->thread); - return 1; + return true; } EXPORT_SYMBOL_GPL(rdev_set_badblocks); -int rdev_clear_badblocks(struct md_rdev *rdev, sector_t s, int sectors, - int is_new) +void rdev_clear_badblocks(struct md_rdev *rdev, sector_t s, int sectors, + int is_new) { if (is_new) s += rdev->new_data_offset; @@ -9889,11 +9961,10 @@ int rdev_clear_badblocks(struct md_rdev *rdev, sector_t s, int sectors, s += rdev->data_offset; if (!badblocks_clear(&rdev->badblocks, s, sectors)) - return 0; + return; if (test_bit(ExternalBbl, &rdev->flags)) sysfs_notify_dirent_safe(rdev->sysfs_badblocks); - return 1; } EXPORT_SYMBOL_GPL(rdev_clear_badblocks); @@ -10010,8 +10081,11 @@ static void check_sb_changes(struct mddev *mddev, struct md_rdev *rdev) /* Check for change of roles in the active devices */ rdev_for_each_safe(rdev2, tmp, mddev) { - if (test_bit(Faulty, &rdev2->flags)) + if (test_bit(Faulty, &rdev2->flags)) { + if (test_bit(ClusterRemove, &rdev2->flags)) + set_bit(MD_RECOVERY_NEEDED, &mddev->recovery); continue; + } /* Check if the roles changed */ role = le16_to_cpu(sb->dev_roles[rdev2->desc_nr]); @@ -10034,7 +10108,7 @@ static void check_sb_changes(struct mddev *mddev, struct md_rdev *rdev) if (rdev2->raid_disk == -1 && role != MD_DISK_ROLE_SPARE && !(le32_to_cpu(sb->feature_map) & MD_FEATURE_RESHAPE_ACTIVE) && - !md_cluster_ops->resync_status_get(mddev)) { + !mddev->cluster_ops->resync_status_get(mddev)) { /* * -1 to make raid1_add_disk() set conf->fullsync * to 1. This could avoid skipping sync when the diff --git a/drivers/md/md.h b/drivers/md/md.h index cc31c795369d..d45a9e6ead80 100644 --- a/drivers/md/md.h +++ b/drivers/md/md.h @@ -18,11 +18,37 @@ #include <linux/timer.h> #include <linux/wait.h> #include <linux/workqueue.h> +#include <linux/raid/md_u.h> #include <trace/events/block.h> -#include "md-cluster.h" #define MaxSector (~(sector_t)0) +enum md_submodule_type { + MD_PERSONALITY = 0, + MD_CLUSTER, + MD_BITMAP, /* TODO */ +}; + +enum md_submodule_id { + ID_LINEAR = LEVEL_LINEAR, + ID_RAID0 = 0, + ID_RAID1 = 1, + ID_RAID4 = 4, + ID_RAID5 = 5, + ID_RAID6 = 6, + ID_RAID10 = 10, + ID_CLUSTER, + ID_BITMAP, /* TODO */ + ID_LLBITMAP, /* TODO */ +}; + +struct md_submodule_head { + enum md_submodule_type type; + enum md_submodule_id id; + const char *name; + struct module *owner; +}; + /* * These flags should really be called "NO_RETRY" rather than * "FAILFAST" because they don't make any promise about time lapse, @@ -106,7 +132,7 @@ struct md_rdev { sector_t sectors; /* Device size (in 512bytes sectors) */ struct mddev *mddev; /* RAID array if running */ - int last_events; /* IO event timestamp */ + unsigned long last_events; /* IO event timestamp */ /* * If meta_bdev is non-NULL, it means that a separate device is @@ -289,11 +315,12 @@ static inline int rdev_has_badblock(struct md_rdev *rdev, sector_t s, return is_badblock(rdev, s, sectors, &first_bad, &bad_sectors); } -extern int rdev_set_badblocks(struct md_rdev *rdev, sector_t s, int sectors, - int is_new); -extern int rdev_clear_badblocks(struct md_rdev *rdev, sector_t s, int sectors, - int is_new); +extern bool rdev_set_badblocks(struct md_rdev *rdev, sector_t s, int sectors, + int is_new); +extern void rdev_clear_badblocks(struct md_rdev *rdev, sector_t s, int sectors, + int is_new); struct md_cluster_info; +struct md_cluster_operations; /** * enum mddev_flags - md device flags. @@ -377,7 +404,8 @@ struct mddev { * are happening, so run/ * takeover/stop are not safe */ - struct gendisk *gendisk; + struct gendisk *gendisk; /* mdraid gendisk */ + struct gendisk *dm_gendisk; /* dm-raid gendisk */ struct kobject kobj; int hold_active; @@ -456,6 +484,7 @@ struct mddev { /* if zero, use the system-wide default */ int sync_speed_min; int sync_speed_max; + int sync_io_depth; /* resync even though the same disks are shared among md-devices */ int parallel_resync; @@ -491,6 +520,7 @@ struct mddev { * adding a spare */ + unsigned long normal_io_events; /* IO event timestamp */ atomic_t recovery_active; /* blocks scheduled, but not written */ wait_queue_head_t recovery_wait; sector_t recovery_cp; @@ -576,6 +606,7 @@ struct mddev { mempool_t *serial_info_pool; void (*sync_super)(struct mddev *mddev, struct md_rdev *rdev); struct md_cluster_info *cluster_info; + struct md_cluster_operations *cluster_ops; unsigned int good_device_nr; /* good device num within cluster raid */ unsigned int noio_flag; /* for memalloc scope API */ @@ -686,23 +717,10 @@ static inline int mddev_trylock(struct mddev *mddev) } extern void mddev_unlock(struct mddev *mddev); -static inline void md_sync_acct(struct block_device *bdev, unsigned long nr_sectors) -{ - if (blk_queue_io_stat(bdev->bd_disk->queue)) - atomic_add(nr_sectors, &bdev->bd_disk->sync_io); -} - -static inline void md_sync_acct_bio(struct bio *bio, unsigned long nr_sectors) -{ - md_sync_acct(bio->bi_bdev, nr_sectors); -} - struct md_personality { - char *name; - int level; - struct list_head list; - struct module *owner; + struct md_submodule_head head; + bool __must_check (*make_request)(struct mddev *mddev, struct bio *bio); /* * start up works that do NOT require md_thread. tasks that @@ -843,13 +861,9 @@ static inline void safe_put_page(struct page *p) if (p) put_page(p); } -extern int register_md_personality(struct md_personality *p); -extern int unregister_md_personality(struct md_personality *p); -extern int register_md_cluster_operations(const struct md_cluster_operations *ops, - struct module *module); -extern int unregister_md_cluster_operations(void); -extern int md_setup_cluster(struct mddev *mddev, int nodes); -extern void md_cluster_stop(struct mddev *mddev); +int register_md_submodule(struct md_submodule_head *msh); +void unregister_md_submodule(struct md_submodule_head *msh); + extern struct md_thread *md_register_thread( void (*run)(struct md_thread *thread), struct mddev *mddev, @@ -906,7 +920,6 @@ extern void md_idle_sync_thread(struct mddev *mddev); extern void md_frozen_sync_thread(struct mddev *mddev); extern void md_unfrozen_sync_thread(struct mddev *mddev); -extern void md_reload_sb(struct mddev *mddev, int raid_disk); extern void md_update_sb(struct mddev *mddev, int force); extern void mddev_create_serial_pool(struct mddev *mddev, struct md_rdev *rdev); extern void mddev_destroy_serial_pool(struct mddev *mddev, @@ -928,7 +941,6 @@ static inline void rdev_dec_pending(struct md_rdev *rdev, struct mddev *mddev) } } -extern const struct md_cluster_operations *md_cluster_ops; static inline int mddev_is_clustered(struct mddev *mddev) { return mddev->cluster_info && mddev->bitmap_info.nodes > 1; diff --git a/drivers/md/persistent-data/Kconfig b/drivers/md/persistent-data/Kconfig index f4f948b0e173..dbb97a7233ab 100644 --- a/drivers/md/persistent-data/Kconfig +++ b/drivers/md/persistent-data/Kconfig @@ -2,7 +2,7 @@ config DM_PERSISTENT_DATA tristate depends on BLK_DEV_DM - select LIBCRC32C + select CRC32 select DM_BUFIO help Library providing immutable on-disk data structure support for diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c index 70bcc3cdf2cd..d8f639f4ae12 100644 --- a/drivers/md/raid0.c +++ b/drivers/md/raid0.c @@ -809,9 +809,13 @@ static void raid0_quiesce(struct mddev *mddev, int quiesce) static struct md_personality raid0_personality= { - .name = "raid0", - .level = 0, - .owner = THIS_MODULE, + .head = { + .type = MD_PERSONALITY, + .id = ID_RAID0, + .name = "raid0", + .owner = THIS_MODULE, + }, + .make_request = raid0_make_request, .run = raid0_run, .free = raid0_free, @@ -822,14 +826,14 @@ static struct md_personality raid0_personality= .error_handler = raid0_error, }; -static int __init raid0_init (void) +static int __init raid0_init(void) { - return register_md_personality (&raid0_personality); + return register_md_submodule(&raid0_personality.head); } -static void raid0_exit (void) +static void __exit raid0_exit(void) { - unregister_md_personality (&raid0_personality); + unregister_md_submodule(&raid0_personality.head); } module_init(raid0_init); diff --git a/drivers/md/raid1-10.c b/drivers/md/raid1-10.c index 62b980b12f93..b8b3a9069701 100644 --- a/drivers/md/raid1-10.c +++ b/drivers/md/raid1-10.c @@ -287,9 +287,19 @@ static inline bool raid1_should_read_first(struct mddev *mddev, return true; if (mddev_is_clustered(mddev) && - md_cluster_ops->area_resyncing(mddev, READ, this_sector, - this_sector + len)) + mddev->cluster_ops->area_resyncing(mddev, READ, this_sector, + this_sector + len)) return true; return false; } + +/* + * bio with REQ_RAHEAD or REQ_NOWAIT can fail at anytime, before such IO is + * submitted to the underlying disks, hence don't record badblocks or retry + * in this case. + */ +static inline bool raid1_should_handle_error(struct bio *bio) +{ + return !(bio->bi_opf & (REQ_RAHEAD | REQ_NOWAIT)); +} diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c index 15829ab192d2..64b8176907a9 100644 --- a/drivers/md/raid1.c +++ b/drivers/md/raid1.c @@ -36,6 +36,7 @@ #include "md.h" #include "raid1.h" #include "md-bitmap.h" +#include "md-cluster.h" #define UNSUPPORTED_MDDEV_FLAGS \ ((1L << MD_HAS_JOURNAL) | \ @@ -372,14 +373,16 @@ static void raid1_end_read_request(struct bio *bio) */ update_head_pos(r1_bio->read_disk, r1_bio); - if (uptodate) + if (uptodate) { set_bit(R1BIO_Uptodate, &r1_bio->state); - else if (test_bit(FailFast, &rdev->flags) && - test_bit(R1BIO_FailFast, &r1_bio->state)) + } else if (test_bit(FailFast, &rdev->flags) && + test_bit(R1BIO_FailFast, &r1_bio->state)) { /* This was a fail-fast read so we definitely * want to retry */ ; - else { + } else if (!raid1_should_handle_error(bio)) { + uptodate = 1; + } else { /* If all other devices have failed, we want to return * the error upwards rather than fail the last device. * Here we redefine "uptodate" to mean "Don't want to retry" @@ -450,16 +453,15 @@ static void raid1_end_write_request(struct bio *bio) struct bio *to_put = NULL; int mirror = find_bio_disk(r1_bio, bio); struct md_rdev *rdev = conf->mirrors[mirror].rdev; - bool discard_error; sector_t lo = r1_bio->sector; sector_t hi = r1_bio->sector + r1_bio->sectors; - - discard_error = bio->bi_status && bio_op(bio) == REQ_OP_DISCARD; + bool ignore_error = !raid1_should_handle_error(bio) || + (bio->bi_status && bio_op(bio) == REQ_OP_DISCARD); /* * 'one mirror IO has finished' event handler: */ - if (bio->bi_status && !discard_error) { + if (bio->bi_status && !ignore_error) { set_bit(WriteErrorSeen, &rdev->flags); if (!test_and_set_bit(WantReplacement, &rdev->flags)) set_bit(MD_RECOVERY_NEEDED, & @@ -510,7 +512,7 @@ static void raid1_end_write_request(struct bio *bio) /* Maybe we can clear some bad blocks. */ if (rdev_has_badblock(rdev, r1_bio->sector, r1_bio->sectors) && - !discard_error) { + !ignore_error) { r1_bio->bios[mirror] = IO_MADE_GOOD; set_bit(R1BIO_MadeGood, &r1_bio->state); } @@ -1397,7 +1399,7 @@ static void raid1_read_request(struct mddev *mddev, struct bio *bio, } read_bio = bio_alloc_clone(mirror->rdev->bdev, bio, gfp, &mddev->bio_set); - + read_bio->bi_opf &= ~REQ_NOWAIT; r1_bio->bios[rdisk] = read_bio; read_bio->bi_iter.bi_sector = r1_bio->sector + @@ -1465,7 +1467,7 @@ static void raid1_write_request(struct mddev *mddev, struct bio *bio, bool is_discard = (bio_op(bio) == REQ_OP_DISCARD); if (mddev_is_clustered(mddev) && - md_cluster_ops->area_resyncing(mddev, WRITE, + mddev->cluster_ops->area_resyncing(mddev, WRITE, bio->bi_iter.bi_sector, bio_end_sector(bio))) { DEFINE_WAIT(w); @@ -1476,7 +1478,7 @@ static void raid1_write_request(struct mddev *mddev, struct bio *bio, for (;;) { prepare_to_wait(&conf->wait_barrier, &w, TASK_IDLE); - if (!md_cluster_ops->area_resyncing(mddev, WRITE, + if (!mddev->cluster_ops->area_resyncing(mddev, WRITE, bio->bi_iter.bi_sector, bio_end_sector(bio))) break; @@ -1647,6 +1649,7 @@ static void raid1_write_request(struct mddev *mddev, struct bio *bio, wait_for_serialization(rdev, r1_bio); } + mbio->bi_opf &= ~REQ_NOWAIT; r1_bio->bios[i] = mbio; mbio->bi_iter.bi_sector = (r1_bio->sector + rdev->data_offset); @@ -2199,14 +2202,9 @@ static int fix_sync_read_error(struct r1bio *r1_bio) if (!rdev_set_badblocks(rdev, sect, s, 0)) abort = 1; } - if (abort) { - conf->recovery_disabled = - mddev->recovery_disabled; - set_bit(MD_RECOVERY_INTR, &mddev->recovery); - md_done_sync(mddev, r1_bio->sectors, 0); - put_buf(r1_bio); + if (abort) return 0; - } + /* Try next page */ sectors -= s; sect += s; @@ -2345,10 +2343,21 @@ static void sync_request_write(struct mddev *mddev, struct r1bio *r1_bio) int disks = conf->raid_disks * 2; struct bio *wbio; - if (!test_bit(R1BIO_Uptodate, &r1_bio->state)) - /* ouch - failed to read all of that. */ - if (!fix_sync_read_error(r1_bio)) + if (!test_bit(R1BIO_Uptodate, &r1_bio->state)) { + /* + * ouch - failed to read all of that. + * No need to fix read error for check/repair + * because all member disks are read. + */ + if (test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery) || + !fix_sync_read_error(r1_bio)) { + conf->recovery_disabled = mddev->recovery_disabled; + set_bit(MD_RECOVERY_INTR, &mddev->recovery); + md_done_sync(mddev, r1_bio->sectors, 0); + put_buf(r1_bio); return; + } + } if (test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery)) process_checks(r1_bio); @@ -2375,7 +2384,6 @@ static void sync_request_write(struct mddev *mddev, struct r1bio *r1_bio) wbio->bi_end_io = end_sync_write; atomic_inc(&r1_bio->remaining); - md_sync_acct(conf->mirrors[i].rdev->bdev, bio_sectors(wbio)); submit_bio_noacct(wbio); } @@ -2482,7 +2490,7 @@ static void fix_read_error(struct r1conf *conf, struct r1bio *r1_bio) } } -static int narrow_write_error(struct r1bio *r1_bio, int i) +static bool narrow_write_error(struct r1bio *r1_bio, int i) { struct mddev *mddev = r1_bio->mddev; struct r1conf *conf = mddev->private; @@ -2503,10 +2511,10 @@ static int narrow_write_error(struct r1bio *r1_bio, int i) sector_t sector; int sectors; int sect_to_write = r1_bio->sectors; - int ok = 1; + bool ok = true; if (rdev->badblocks.shift < 0) - return 0; + return false; block_sectors = roundup(1 << rdev->badblocks.shift, bdev_logical_block_size(rdev->bdev) >> 9); @@ -3034,9 +3042,9 @@ static sector_t raid1_sync_request(struct mddev *mddev, sector_t sector_nr, conf->cluster_sync_low = mddev->curr_resync_completed; conf->cluster_sync_high = conf->cluster_sync_low + CLUSTER_RESYNC_WINDOW_SECTORS; /* Send resync message */ - md_cluster_ops->resync_info_update(mddev, - conf->cluster_sync_low, - conf->cluster_sync_high); + mddev->cluster_ops->resync_info_update(mddev, + conf->cluster_sync_low, + conf->cluster_sync_high); } /* For a user-requested sync, we read all readable devices and do a @@ -3048,7 +3056,6 @@ static sector_t raid1_sync_request(struct mddev *mddev, sector_t sector_nr, bio = r1_bio->bios[i]; if (bio->bi_end_io == end_sync_read) { read_targets--; - md_sync_acct_bio(bio, nr_sectors); if (read_targets == 1) bio->bi_opf &= ~MD_FAILFAST; submit_bio_noacct(bio); @@ -3057,7 +3064,6 @@ static sector_t raid1_sync_request(struct mddev *mddev, sector_t sector_nr, } else { atomic_set(&r1_bio->remaining, 1); bio = r1_bio->bios[r1_bio->read_disk]; - md_sync_acct_bio(bio, nr_sectors); if (read_targets == 1) bio->bi_opf &= ~MD_FAILFAST; submit_bio_noacct(bio); @@ -3423,6 +3429,7 @@ static int raid1_reshape(struct mddev *mddev) /* ok, everything is stopped */ oldpool = conf->r1bio_pool; conf->r1bio_pool = newpool; + init_waitqueue_head(&conf->r1bio_pool.wait); for (d = d2 = 0; d < conf->raid_disks; d++) { struct md_rdev *rdev = conf->mirrors[d].rdev; @@ -3492,9 +3499,13 @@ static void *raid1_takeover(struct mddev *mddev) static struct md_personality raid1_personality = { - .name = "raid1", - .level = 1, - .owner = THIS_MODULE, + .head = { + .type = MD_PERSONALITY, + .id = ID_RAID1, + .name = "raid1", + .owner = THIS_MODULE, + }, + .make_request = raid1_make_request, .run = raid1_run, .free = raid1_free, @@ -3511,18 +3522,18 @@ static struct md_personality raid1_personality = .takeover = raid1_takeover, }; -static int __init raid_init(void) +static int __init raid1_init(void) { - return register_md_personality(&raid1_personality); + return register_md_submodule(&raid1_personality.head); } -static void raid_exit(void) +static void __exit raid1_exit(void) { - unregister_md_personality(&raid1_personality); + unregister_md_submodule(&raid1_personality.head); } -module_init(raid_init); -module_exit(raid_exit); +module_init(raid1_init); +module_exit(raid1_exit); MODULE_LICENSE("GPL"); MODULE_DESCRIPTION("RAID1 (mirroring) personality for MD"); MODULE_ALIAS("md-personality-3"); /* RAID1 */ diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c index 76a75925b713..95dc354a86a0 100644 --- a/drivers/md/raid10.c +++ b/drivers/md/raid10.c @@ -24,6 +24,7 @@ #include "raid10.h" #include "raid0.h" #include "md-bitmap.h" +#include "md-cluster.h" /* * RAID10 provides a combination of RAID0 and RAID1 functionality. @@ -398,6 +399,8 @@ static void raid10_end_read_request(struct bio *bio) * wait for the 'master' bio. */ set_bit(R10BIO_Uptodate, &r10_bio->state); + } else if (!raid1_should_handle_error(bio)) { + uptodate = 1; } else { /* If all other devices that store this block have * failed, we want to return the error upwards rather @@ -455,9 +458,8 @@ static void raid10_end_write_request(struct bio *bio) int slot, repl; struct md_rdev *rdev = NULL; struct bio *to_put = NULL; - bool discard_error; - - discard_error = bio->bi_status && bio_op(bio) == REQ_OP_DISCARD; + bool ignore_error = !raid1_should_handle_error(bio) || + (bio->bi_status && bio_op(bio) == REQ_OP_DISCARD); dev = find_bio_disk(conf, r10_bio, bio, &slot, &repl); @@ -471,7 +473,7 @@ static void raid10_end_write_request(struct bio *bio) /* * this branch is our 'one mirror IO has finished' event handler: */ - if (bio->bi_status && !discard_error) { + if (bio->bi_status && !ignore_error) { if (repl) /* Never record new bad blocks to replacement, * just fail it. @@ -526,7 +528,7 @@ static void raid10_end_write_request(struct bio *bio) /* Maybe we can clear some bad blocks. */ if (rdev_has_badblock(rdev, r10_bio->devs[slot].addr, r10_bio->sectors) && - !discard_error) { + !ignore_error) { bio_put(bio); if (repl) r10_bio->devs[slot].repl_bio = IO_MADE_GOOD; @@ -1180,8 +1182,11 @@ static void raid10_read_request(struct mddev *mddev, struct bio *bio, } } - if (!regular_request_wait(mddev, conf, bio, r10_bio->sectors)) + if (!regular_request_wait(mddev, conf, bio, r10_bio->sectors)) { + raid_end_bio_io(r10_bio); return; + } + rdev = read_balance(conf, r10_bio, &max_sectors); if (!rdev) { if (err_rdev) { @@ -1219,6 +1224,7 @@ static void raid10_read_request(struct mddev *mddev, struct bio *bio, r10_bio->master_bio = bio; } read_bio = bio_alloc_clone(rdev->bdev, bio, gfp, &mddev->bio_set); + read_bio->bi_opf &= ~REQ_NOWAIT; r10_bio->devs[slot].bio = read_bio; r10_bio->devs[slot].rdev = rdev; @@ -1254,6 +1260,7 @@ static void raid10_write_one_disk(struct mddev *mddev, struct r10bio *r10_bio, conf->mirrors[devnum].rdev; mbio = bio_alloc_clone(rdev->bdev, bio, GFP_NOIO, &mddev->bio_set); + mbio->bi_opf &= ~REQ_NOWAIT; if (replacement) r10_bio->devs[n_copy].repl_bio = mbio; else @@ -1347,9 +1354,9 @@ static void raid10_write_request(struct mddev *mddev, struct bio *bio, int error; if ((mddev_is_clustered(mddev) && - md_cluster_ops->area_resyncing(mddev, WRITE, - bio->bi_iter.bi_sector, - bio_end_sector(bio)))) { + mddev->cluster_ops->area_resyncing(mddev, WRITE, + bio->bi_iter.bi_sector, + bio_end_sector(bio)))) { DEFINE_WAIT(w); /* Bail out if REQ_NOWAIT is set for the bio */ if (bio->bi_opf & REQ_NOWAIT) { @@ -1359,7 +1366,7 @@ static void raid10_write_request(struct mddev *mddev, struct bio *bio, for (;;) { prepare_to_wait(&conf->wait_barrier, &w, TASK_IDLE); - if (!md_cluster_ops->area_resyncing(mddev, WRITE, + if (!mddev->cluster_ops->area_resyncing(mddev, WRITE, bio->bi_iter.bi_sector, bio_end_sector(bio))) break; schedule(); @@ -1368,8 +1375,11 @@ static void raid10_write_request(struct mddev *mddev, struct bio *bio, } sectors = r10_bio->sectors; - if (!regular_request_wait(mddev, conf, bio, sectors)) + if (!regular_request_wait(mddev, conf, bio, sectors)) { + raid_end_bio_io(r10_bio); return; + } + if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery) && (mddev->reshape_backwards ? (bio->bi_iter.bi_sector < conf->reshape_safe && @@ -2425,7 +2435,6 @@ static void sync_request_write(struct mddev *mddev, struct r10bio *r10_bio) atomic_inc(&conf->mirrors[d].rdev->nr_pending); atomic_inc(&r10_bio->remaining); - md_sync_acct(conf->mirrors[d].rdev->bdev, bio_sectors(tbio)); if (test_bit(FailFast, &conf->mirrors[d].rdev->flags)) tbio->bi_opf |= MD_FAILFAST; @@ -2437,18 +2446,13 @@ static void sync_request_write(struct mddev *mddev, struct r10bio *r10_bio) * that are active */ for (i = 0; i < conf->copies; i++) { - int d; - tbio = r10_bio->devs[i].repl_bio; if (!tbio || !tbio->bi_end_io) continue; if (r10_bio->devs[i].bio->bi_end_io != end_sync_write && r10_bio->devs[i].bio != fbio) bio_copy_data(tbio, fbio); - d = r10_bio->devs[i].devnum; atomic_inc(&r10_bio->remaining); - md_sync_acct(conf->mirrors[d].replacement->bdev, - bio_sectors(tbio)); submit_bio_noacct(tbio); } @@ -2582,13 +2586,10 @@ static void recovery_request_write(struct mddev *mddev, struct r10bio *r10_bio) d = r10_bio->devs[1].devnum; if (wbio->bi_end_io) { atomic_inc(&conf->mirrors[d].rdev->nr_pending); - md_sync_acct(conf->mirrors[d].rdev->bdev, bio_sectors(wbio)); submit_bio_noacct(wbio); } if (wbio2) { atomic_inc(&conf->mirrors[d].replacement->nr_pending); - md_sync_acct(conf->mirrors[d].replacement->bdev, - bio_sectors(wbio2)); submit_bio_noacct(wbio2); } } @@ -2778,7 +2779,7 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10 } } -static int narrow_write_error(struct r10bio *r10_bio, int i) +static bool narrow_write_error(struct r10bio *r10_bio, int i) { struct bio *bio = r10_bio->master_bio; struct mddev *mddev = r10_bio->mddev; @@ -2799,10 +2800,10 @@ static int narrow_write_error(struct r10bio *r10_bio, int i) sector_t sector; int sectors; int sect_to_write = r10_bio->sectors; - int ok = 1; + bool ok = true; if (rdev->badblocks.shift < 0) - return 0; + return false; block_sectors = roundup(1 << rdev->badblocks.shift, bdev_logical_block_size(rdev->bdev) >> 9); @@ -3708,7 +3709,7 @@ static sector_t raid10_sync_request(struct mddev *mddev, sector_t sector_nr, conf->cluster_sync_low = mddev->curr_resync_completed; raid10_set_cluster_sync_high(conf); /* Send resync message */ - md_cluster_ops->resync_info_update(mddev, + mddev->cluster_ops->resync_info_update(mddev, conf->cluster_sync_low, conf->cluster_sync_high); } @@ -3741,7 +3742,7 @@ static sector_t raid10_sync_request(struct mddev *mddev, sector_t sector_nr, } if (broadcast_msg) { raid10_set_cluster_sync_high(conf); - md_cluster_ops->resync_info_update(mddev, + mddev->cluster_ops->resync_info_update(mddev, conf->cluster_sync_low, conf->cluster_sync_high); } @@ -3756,7 +3757,6 @@ static sector_t raid10_sync_request(struct mddev *mddev, sector_t sector_nr, r10_bio->sectors = nr_sectors; if (bio->bi_end_io == end_sync_read) { - md_sync_acct_bio(bio, nr_sectors); bio->bi_status = 0; submit_bio_noacct(bio); } @@ -4009,6 +4009,7 @@ static int raid10_set_queue_limits(struct mddev *mddev) md_init_stacking_limits(&lim); lim.max_write_zeroes_sectors = 0; lim.io_min = mddev->chunk_sectors << 9; + lim.chunk_sectors = mddev->chunk_sectors; lim.io_opt = lim.io_min * raid10_nr_stripes(conf); lim.features |= BLK_FEAT_ATOMIC_WRITES; err = mddev_stack_rdev_limits(mddev, &lim, MDDEV_STACK_INTEGRITY); @@ -4533,7 +4534,7 @@ static int raid10_start_reshape(struct mddev *mddev) if (ret) goto abort; - ret = md_cluster_ops->resize_bitmaps(mddev, newsize, oldsize); + ret = mddev->cluster_ops->resize_bitmaps(mddev, newsize, oldsize); if (ret) { mddev->bitmap_ops->resize(mddev, oldsize, 0, false); goto abort; @@ -4824,7 +4825,7 @@ read_more: conf->cluster_sync_low = sb_reshape_pos; } - md_cluster_ops->resync_info_update(mddev, conf->cluster_sync_low, + mddev->cluster_ops->resync_info_update(mddev, conf->cluster_sync_low, conf->cluster_sync_high); } @@ -4879,7 +4880,6 @@ read_more: r10_bio->sectors = nr_sectors; /* Now submit the read */ - md_sync_acct_bio(read_bio, r10_bio->sectors); atomic_inc(&r10_bio->remaining); read_bio->bi_next = NULL; submit_bio_noacct(read_bio); @@ -4939,7 +4939,6 @@ static void reshape_request_write(struct mddev *mddev, struct r10bio *r10_bio) continue; atomic_inc(&rdev->nr_pending); - md_sync_acct_bio(b, r10_bio->sectors); atomic_inc(&r10_bio->remaining); b->bi_next = NULL; submit_bio_noacct(b); @@ -4969,7 +4968,7 @@ static void raid10_update_reshape_pos(struct mddev *mddev) struct r10conf *conf = mddev->private; sector_t lo, hi; - md_cluster_ops->resync_info_get(mddev, &lo, &hi); + mddev->cluster_ops->resync_info_get(mddev, &lo, &hi); if (((mddev->reshape_position <= hi) && (mddev->reshape_position >= lo)) || mddev->reshape_position == MaxSector) conf->reshape_progress = mddev->reshape_position; @@ -5115,9 +5114,13 @@ static void raid10_finish_reshape(struct mddev *mddev) static struct md_personality raid10_personality = { - .name = "raid10", - .level = 10, - .owner = THIS_MODULE, + .head = { + .type = MD_PERSONALITY, + .id = ID_RAID10, + .name = "raid10", + .owner = THIS_MODULE, + }, + .make_request = raid10_make_request, .run = raid10_run, .free = raid10_free, @@ -5137,18 +5140,18 @@ static struct md_personality raid10_personality = .update_reshape_pos = raid10_update_reshape_pos, }; -static int __init raid_init(void) +static int __init raid10_init(void) { - return register_md_personality(&raid10_personality); + return register_md_submodule(&raid10_personality.head); } -static void raid_exit(void) +static void __exit raid10_exit(void) { - unregister_md_personality(&raid10_personality); + unregister_md_submodule(&raid10_personality.head); } -module_init(raid_init); -module_exit(raid_exit); +module_init(raid10_init); +module_exit(raid10_exit); MODULE_LICENSE("GPL"); MODULE_DESCRIPTION("RAID10 (striped mirror) personality for MD"); MODULE_ALIAS("md-personality-9"); /* RAID10 */ diff --git a/drivers/md/raid5-cache.c b/drivers/md/raid5-cache.c index e530271cb86b..ba768ca7f422 100644 --- a/drivers/md/raid5-cache.c +++ b/drivers/md/raid5-cache.c @@ -714,7 +714,7 @@ static void r5l_submit_current_io(struct r5l_log *log) block = page_address(io->meta_page); block->meta_size = cpu_to_le32(io->meta_offset); - crc = crc32c_le(log->uuid_checksum, block, PAGE_SIZE); + crc = crc32c(log->uuid_checksum, block, PAGE_SIZE); block->checksum = cpu_to_le32(crc); log->current_io = NULL; @@ -1020,8 +1020,8 @@ int r5l_write_stripe(struct r5l_log *log, struct stripe_head *sh) if (test_bit(STRIPE_LOG_TRAPPED, &sh->state)) continue; addr = kmap_local_page(sh->dev[i].page); - sh->dev[i].log_checksum = crc32c_le(log->uuid_checksum, - addr, PAGE_SIZE); + sh->dev[i].log_checksum = crc32c(log->uuid_checksum, + addr, PAGE_SIZE); kunmap_local(addr); } parity_pages = 1 + !!(sh->qd_idx >= 0); @@ -1741,7 +1741,7 @@ static int r5l_recovery_read_meta_block(struct r5l_log *log, le64_to_cpu(mb->position) != ctx->pos) return -EINVAL; - crc = crc32c_le(log->uuid_checksum, mb, PAGE_SIZE); + crc = crc32c(log->uuid_checksum, mb, PAGE_SIZE); if (stored_crc != crc) return -EINVAL; @@ -1780,8 +1780,7 @@ static int r5l_log_write_empty_meta_block(struct r5l_log *log, sector_t pos, return -ENOMEM; r5l_recovery_create_empty_meta_block(log, page, pos, seq); mb = page_address(page); - mb->checksum = cpu_to_le32(crc32c_le(log->uuid_checksum, - mb, PAGE_SIZE)); + mb->checksum = cpu_to_le32(crc32c(log->uuid_checksum, mb, PAGE_SIZE)); if (!sync_page_io(log->rdev, pos, PAGE_SIZE, page, REQ_OP_WRITE | REQ_SYNC | REQ_FUA, false)) { __free_page(page); @@ -1976,7 +1975,7 @@ r5l_recovery_verify_data_checksum(struct r5l_log *log, r5l_recovery_read_page(log, ctx, page, log_offset); addr = kmap_local_page(page); - checksum = crc32c_le(log->uuid_checksum, addr, PAGE_SIZE); + checksum = crc32c(log->uuid_checksum, addr, PAGE_SIZE); kunmap_local(addr); return (le32_to_cpu(log_checksum) == checksum) ? 0 : -EINVAL; } @@ -2379,8 +2378,8 @@ r5c_recovery_rewrite_data_only_stripes(struct r5l_log *log, raid5_compute_blocknr(sh, i, 0)); addr = kmap_local_page(dev->page); payload->checksum[0] = cpu_to_le32( - crc32c_le(log->uuid_checksum, addr, - PAGE_SIZE)); + crc32c(log->uuid_checksum, addr, + PAGE_SIZE)); kunmap_local(addr); sync_page_io(log->rdev, write_pos, PAGE_SIZE, dev->page, REQ_OP_WRITE, false); @@ -2392,8 +2391,8 @@ r5c_recovery_rewrite_data_only_stripes(struct r5l_log *log, } } mb->meta_size = cpu_to_le32(offset); - mb->checksum = cpu_to_le32(crc32c_le(log->uuid_checksum, - mb, PAGE_SIZE)); + mb->checksum = cpu_to_le32(crc32c(log->uuid_checksum, + mb, PAGE_SIZE)); sync_page_io(log->rdev, ctx->pos, PAGE_SIZE, page, REQ_OP_WRITE | REQ_SYNC | REQ_FUA, false); sh->log_start = ctx->pos; @@ -2885,8 +2884,8 @@ int r5c_cache_data(struct r5l_log *log, struct stripe_head *sh) if (!test_bit(R5_Wantwrite, &sh->dev[i].flags)) continue; addr = kmap_local_page(sh->dev[i].page); - sh->dev[i].log_checksum = crc32c_le(log->uuid_checksum, - addr, PAGE_SIZE); + sh->dev[i].log_checksum = crc32c(log->uuid_checksum, + addr, PAGE_SIZE); kunmap_local(addr); pages++; } @@ -2969,7 +2968,7 @@ static int r5l_load_log(struct r5l_log *log) } stored_crc = le32_to_cpu(mb->checksum); mb->checksum = 0; - expected_crc = crc32c_le(log->uuid_checksum, mb, PAGE_SIZE); + expected_crc = crc32c(log->uuid_checksum, mb, PAGE_SIZE); if (stored_crc != expected_crc) { create_super = true; goto create; @@ -3077,8 +3076,8 @@ int r5l_init_log(struct r5conf *conf, struct md_rdev *rdev) return -ENOMEM; log->rdev = rdev; log->need_cache_flush = bdev_write_cache(rdev->bdev); - log->uuid_checksum = crc32c_le(~0, rdev->mddev->uuid, - sizeof(rdev->mddev->uuid)); + log->uuid_checksum = crc32c(~0, rdev->mddev->uuid, + sizeof(rdev->mddev->uuid)); mutex_init(&log->io_mutex); diff --git a/drivers/md/raid5-ppl.c b/drivers/md/raid5-ppl.c index 37c4da5311ca..c0fb335311aa 100644 --- a/drivers/md/raid5-ppl.c +++ b/drivers/md/raid5-ppl.c @@ -346,9 +346,9 @@ static int ppl_log_stripe(struct ppl_log *log, struct stripe_head *sh) if (!test_bit(STRIPE_FULL_WRITE, &sh->state)) { le32_add_cpu(&e->pp_size, PAGE_SIZE); io->pp_size += PAGE_SIZE; - e->checksum = cpu_to_le32(crc32c_le(le32_to_cpu(e->checksum), - page_address(sh->ppl_page), - PAGE_SIZE)); + e->checksum = cpu_to_le32(crc32c(le32_to_cpu(e->checksum), + page_address(sh->ppl_page), + PAGE_SIZE)); } list_add_tail(&sh->log_list, &io->stripe_list); @@ -454,7 +454,7 @@ static void ppl_submit_iounit(struct ppl_io_unit *io) } pplhdr->entries_count = cpu_to_le32(io->entries_count); - pplhdr->checksum = cpu_to_le32(~crc32c_le(~0, pplhdr, PPL_HEADER_SIZE)); + pplhdr->checksum = cpu_to_le32(~crc32c(~0, pplhdr, PPL_HEADER_SIZE)); /* Rewind the buffer if current PPL is larger then remaining space */ if (log->use_multippl && @@ -998,7 +998,7 @@ static int ppl_recover(struct ppl_log *log, struct ppl_header *pplhdr, goto out; } - crc = crc32c_le(crc, page_address(page), s); + crc = crc32c(crc, page_address(page), s); pp_size -= s; sector += s >> 9; @@ -1052,7 +1052,7 @@ static int ppl_write_empty_header(struct ppl_log *log) log->rdev->ppl.size, GFP_NOIO, 0); memset(pplhdr->reserved, 0xff, PPL_HDR_RESERVED); pplhdr->signature = cpu_to_le32(log->ppl_conf->signature); - pplhdr->checksum = cpu_to_le32(~crc32c_le(~0, pplhdr, PAGE_SIZE)); + pplhdr->checksum = cpu_to_le32(~crc32c(~0, pplhdr, PAGE_SIZE)); if (!sync_page_io(rdev, rdev->ppl.sector - rdev->data_offset, PPL_HEADER_SIZE, page, REQ_OP_WRITE | REQ_SYNC | @@ -1106,7 +1106,7 @@ static int ppl_load_distributed(struct ppl_log *log) /* check header validity */ crc_stored = le32_to_cpu(pplhdr->checksum); pplhdr->checksum = 0; - crc = ~crc32c_le(~0, pplhdr, PAGE_SIZE); + crc = ~crc32c(~0, pplhdr, PAGE_SIZE); if (crc_stored != crc) { pr_debug("%s: ppl header crc does not match: stored: 0x%x calculated: 0x%x (offset: %llu)\n", @@ -1390,7 +1390,7 @@ int ppl_init_log(struct r5conf *conf) spin_lock_init(&ppl_conf->no_mem_stripes_lock); if (!mddev->external) { - ppl_conf->signature = ~crc32c_le(~0, mddev->uuid, sizeof(mddev->uuid)); + ppl_conf->signature = ~crc32c(~0, mddev->uuid, sizeof(mddev->uuid)); ppl_conf->block_size = 512; } else { ppl_conf->block_size = diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 5c79429acc64..ca5b0e8ba707 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -1240,10 +1240,6 @@ again: } if (rdev) { - if (s->syncing || s->expanding || s->expanded - || s->replacing) - md_sync_acct(rdev->bdev, RAID5_STRIPE_SECTORS(conf)); - set_bit(STRIPE_IO_STARTED, &sh->state); bio_init(bi, rdev->bdev, &dev->vec, 1, op | op_flags); @@ -1300,10 +1296,6 @@ again: submit_bio_noacct(bi); } if (rrdev) { - if (s->syncing || s->expanding || s->expanded - || s->replacing) - md_sync_acct(rrdev->bdev, RAID5_STRIPE_SECTORS(conf)); - set_bit(STRIPE_IO_STARTED, &sh->state); bio_init(rbi, rrdev->bdev, &dev->rvec, 1, op | op_flags); @@ -5858,6 +5850,9 @@ static enum reshape_loc get_reshape_loc(struct mddev *mddev, struct r5conf *conf, sector_t logical_sector) { sector_t reshape_progress, reshape_safe; + + if (likely(conf->reshape_progress == MaxSector)) + return LOC_NO_RESHAPE; /* * Spinlock is needed as reshape_progress may be * 64bit on a 32bit platform, and so it might be @@ -5935,22 +5930,19 @@ static enum stripe_result make_stripe_request(struct mddev *mddev, const int rw = bio_data_dir(bi); enum stripe_result ret; struct stripe_head *sh; + enum reshape_loc loc; sector_t new_sector; int previous = 0, flags = 0; int seq, dd_idx; seq = read_seqcount_begin(&conf->gen_lock); - - if (unlikely(conf->reshape_progress != MaxSector)) { - enum reshape_loc loc = get_reshape_loc(mddev, conf, - logical_sector); - if (loc == LOC_INSIDE_RESHAPE) { - ret = STRIPE_SCHEDULE_AND_RETRY; - goto out; - } - if (loc == LOC_AHEAD_OF_RESHAPE) - previous = 1; + loc = get_reshape_loc(mddev, conf, logical_sector); + if (loc == LOC_INSIDE_RESHAPE) { + ret = STRIPE_SCHEDULE_AND_RETRY; + goto out; } + if (loc == LOC_AHEAD_OF_RESHAPE) + previous = 1; new_sector = raid5_compute_sector(conf, logical_sector, previous, &dd_idx, NULL); @@ -6127,7 +6119,6 @@ static bool raid5_make_request(struct mddev *mddev, struct bio * bi) /* Bail out if conflicts with reshape and REQ_NOWAIT is set */ if ((bi->bi_opf & REQ_NOWAIT) && - (conf->reshape_progress != MaxSector) && get_reshape_loc(mddev, conf, logical_sector) == LOC_INSIDE_RESHAPE) { bio_wouldblock_error(bi); if (rw == WRITE) @@ -8954,9 +8945,13 @@ static void raid5_prepare_suspend(struct mddev *mddev) static struct md_personality raid6_personality = { - .name = "raid6", - .level = 6, - .owner = THIS_MODULE, + .head = { + .type = MD_PERSONALITY, + .id = ID_RAID6, + .name = "raid6", + .owner = THIS_MODULE, + }, + .make_request = raid5_make_request, .run = raid5_run, .start = raid5_start, @@ -8980,9 +8975,13 @@ static struct md_personality raid6_personality = }; static struct md_personality raid5_personality = { - .name = "raid5", - .level = 5, - .owner = THIS_MODULE, + .head = { + .type = MD_PERSONALITY, + .id = ID_RAID5, + .name = "raid5", + .owner = THIS_MODULE, + }, + .make_request = raid5_make_request, .run = raid5_run, .start = raid5_start, @@ -9007,9 +9006,13 @@ static struct md_personality raid5_personality = static struct md_personality raid4_personality = { - .name = "raid4", - .level = 4, - .owner = THIS_MODULE, + .head = { + .type = MD_PERSONALITY, + .id = ID_RAID4, + .name = "raid4", + .owner = THIS_MODULE, + }, + .make_request = raid5_make_request, .run = raid5_run, .start = raid5_start, @@ -9045,21 +9048,39 @@ static int __init raid5_init(void) "md/raid5:prepare", raid456_cpu_up_prepare, raid456_cpu_dead); - if (ret) { - destroy_workqueue(raid5_wq); - return ret; - } - register_md_personality(&raid6_personality); - register_md_personality(&raid5_personality); - register_md_personality(&raid4_personality); + if (ret) + goto err_destroy_wq; + + ret = register_md_submodule(&raid6_personality.head); + if (ret) + goto err_cpuhp_remove; + + ret = register_md_submodule(&raid5_personality.head); + if (ret) + goto err_unregister_raid6; + + ret = register_md_submodule(&raid4_personality.head); + if (ret) + goto err_unregister_raid5; + return 0; + +err_unregister_raid5: + unregister_md_submodule(&raid5_personality.head); +err_unregister_raid6: + unregister_md_submodule(&raid6_personality.head); +err_cpuhp_remove: + cpuhp_remove_multi_state(CPUHP_MD_RAID5_PREPARE); +err_destroy_wq: + destroy_workqueue(raid5_wq); + return ret; } -static void raid5_exit(void) +static void __exit raid5_exit(void) { - unregister_md_personality(&raid6_personality); - unregister_md_personality(&raid5_personality); - unregister_md_personality(&raid4_personality); + unregister_md_submodule(&raid6_personality.head); + unregister_md_submodule(&raid5_personality.head); + unregister_md_submodule(&raid4_personality.head); cpuhp_remove_multi_state(CPUHP_MD_RAID5_PREPARE); destroy_workqueue(raid5_wq); } |