diff options
Diffstat (limited to 'drivers/md')
| -rw-r--r-- | drivers/md/bcache/bcache.h | 6 | ||||
| -rw-r--r-- | drivers/md/bcache/btree.c | 49 | ||||
| -rw-r--r-- | drivers/md/bcache/debug.c | 6 | ||||
| -rw-r--r-- | drivers/md/bcache/journal.c | 1 | ||||
| -rw-r--r-- | drivers/md/bcache/request.c | 16 | ||||
| -rw-r--r-- | drivers/md/bcache/super.c | 50 | ||||
| -rw-r--r-- | drivers/md/bcache/writeback.c | 34 | ||||
| -rw-r--r-- | drivers/md/bcache/writeback.h | 19 | ||||
| -rw-r--r-- | drivers/md/dm.c | 6 | ||||
| -rw-r--r-- | drivers/md/md.c | 12 |
10 files changed, 139 insertions, 60 deletions
diff --git a/drivers/md/bcache/bcache.h b/drivers/md/bcache/bcache.h index d6bf294f3907..872ef4d67711 100644 --- a/drivers/md/bcache/bcache.h +++ b/drivers/md/bcache/bcache.h @@ -474,6 +474,7 @@ struct cache { struct gc_stat { size_t nodes; + size_t nodes_pre; size_t key_bytes; size_t nkeys; @@ -525,6 +526,7 @@ struct cache_set { unsigned devices_max_used; struct list_head cached_devs; uint64_t cached_dev_sectors; + atomic_long_t flash_dev_dirty_sectors; struct closure caching; struct closure sb_write; @@ -603,6 +605,10 @@ struct cache_set { */ atomic_t rescale; /* + * used for GC, identify if any front side I/Os is inflight + */ + atomic_t search_inflight; + /* * When we invalidate buckets, we use both the priority and the amount * of good data to determine which buckets to reuse first - to weight * those together consistently we keep track of the smallest nonzero diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c index 547c9eedc2f4..475008fbbaab 100644 --- a/drivers/md/bcache/btree.c +++ b/drivers/md/bcache/btree.c @@ -90,6 +90,9 @@ #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 #define PTR_DIRTY_BIT (((uint64_t) 1 << 36)) @@ -1520,6 +1523,32 @@ static unsigned btree_gc_count_keys(struct btree *b) return ret; } +static size_t btree_gc_min_nodes(struct cache_set *c) +{ + size_t min_nodes; + + /* + * Since incremental GC would stop 100ms when front + * side I/O comes, so when there are many btree nodes, + * if GC only processes constant (100) nodes each time, + * GC would last a long time, and the front side I/Os + * would run out of the buckets (since no new bucket + * can be allocated during GC), and be blocked again. + * So GC should not process constant nodes, but varied + * nodes according to the number of btree nodes, which + * realized by dividing GC into constant(100) times, + * so when there are many btree nodes, GC can process + * more nodes each time, otherwise, GC will process less + * nodes each time (but no less than MIN_GC_NODES) + */ + min_nodes = c->gc_stats.nodes / MAX_GC_TIMES; + if (min_nodes < MIN_GC_NODES) + min_nodes = MIN_GC_NODES; + + return min_nodes; +} + + static int btree_gc_recurse(struct btree *b, struct btree_op *op, struct closure *writes, struct gc_stat *gc) { @@ -1585,6 +1614,13 @@ static int btree_gc_recurse(struct btree *b, struct btree_op *op, memmove(r + 1, r, sizeof(r[0]) * (GC_MERGE_NODES - 1)); r->b = NULL; + if (atomic_read(&b->c->search_inflight) && + gc->nodes >= gc->nodes_pre + btree_gc_min_nodes(b->c)) { + gc->nodes_pre = gc->nodes; + ret = -EAGAIN; + break; + } + if (need_resched()) { ret = -EAGAIN; break; @@ -1753,7 +1789,10 @@ static void bch_btree_gc(struct cache_set *c) closure_sync(&writes); cond_resched(); - if (ret && ret != -EAGAIN) + if (ret == -EAGAIN) + schedule_timeout_interruptible(msecs_to_jiffies + (GC_SLEEP_MS)); + else if (ret) pr_warn("gc failed!"); } while (ret && !test_bit(CACHE_SET_IO_DISABLE, &c->flags)); @@ -1834,8 +1873,14 @@ static int bch_btree_check_recurse(struct btree *b, struct btree_op *op) do { k = bch_btree_iter_next_filter(&iter, &b->keys, bch_ptr_bad); - if (k) + if (k) { btree_node_prefetch(b, k); + /* + * initiallize c->gc_stats.nodes + * for incremental GC + */ + b->c->gc_stats.nodes++; + } if (p) ret = btree(check_recurse, p, b, op); diff --git a/drivers/md/bcache/debug.c b/drivers/md/bcache/debug.c index d030ce3025a6..04d146711950 100644 --- a/drivers/md/bcache/debug.c +++ b/drivers/md/bcache/debug.c @@ -110,11 +110,15 @@ void bch_data_verify(struct cached_dev *dc, struct bio *bio) struct bio_vec bv, cbv; struct bvec_iter iter, citer = { 0 }; - check = bio_clone_kmalloc(bio, GFP_NOIO); + check = bio_kmalloc(GFP_NOIO, bio_segments(bio)); if (!check) return; + check->bi_disk = bio->bi_disk; check->bi_opf = REQ_OP_READ; + check->bi_iter.bi_sector = bio->bi_iter.bi_sector; + check->bi_iter.bi_size = bio->bi_iter.bi_size; + bch_bio_map(check, NULL); if (bch_bio_alloc_pages(check, GFP_NOIO)) goto out_put; diff --git a/drivers/md/bcache/journal.c b/drivers/md/bcache/journal.c index 18f1b5239620..10748c626a1d 100644 --- a/drivers/md/bcache/journal.c +++ b/drivers/md/bcache/journal.c @@ -828,6 +828,7 @@ void bch_journal_free(struct cache_set *c) free_pages((unsigned long) c->journal.w[1].data, JSET_BITS); free_pages((unsigned long) c->journal.w[0].data, JSET_BITS); free_fifo(&c->journal.pin); + free_heap(&c->flush_btree); } int bch_journal_alloc(struct cache_set *c) diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c index ae67f5fa8047..43af905920f5 100644 --- a/drivers/md/bcache/request.c +++ b/drivers/md/bcache/request.c @@ -667,8 +667,7 @@ static void backing_request_endio(struct bio *bio) static void bio_complete(struct search *s) { if (s->orig_bio) { - generic_end_io_acct(s->d->disk->queue, - bio_data_dir(s->orig_bio), + generic_end_io_acct(s->d->disk->queue, bio_op(s->orig_bio), &s->d->disk->part0, s->start_time); trace_bcache_request_end(s->d, s->orig_bio); @@ -702,6 +701,8 @@ static void search_free(struct closure *cl) { struct search *s = container_of(cl, struct search, cl); + atomic_dec(&s->d->c->search_inflight); + if (s->iop.bio) bio_put(s->iop.bio); @@ -719,6 +720,7 @@ static inline struct search *search_alloc(struct bio *bio, closure_init(&s->cl, NULL); do_bio_hook(s, bio, request_endio); + atomic_inc(&d->c->search_inflight); s->orig_bio = bio; s->cache_miss = NULL; @@ -1062,8 +1064,7 @@ static void detached_dev_end_io(struct bio *bio) bio->bi_end_io = ddip->bi_end_io; bio->bi_private = ddip->bi_private; - generic_end_io_acct(ddip->d->disk->queue, - bio_data_dir(bio), + generic_end_io_acct(ddip->d->disk->queue, bio_op(bio), &ddip->d->disk->part0, ddip->start_time); if (bio->bi_status) { @@ -1120,7 +1121,7 @@ static blk_qc_t cached_dev_make_request(struct request_queue *q, } atomic_set(&dc->backing_idle, 0); - generic_start_io_acct(q, rw, bio_sectors(bio), &d->disk->part0); + generic_start_io_acct(q, bio_op(bio), bio_sectors(bio), &d->disk->part0); bio_set_dev(bio, dc->bdev); bio->bi_iter.bi_sector += dc->sb.data_offset; @@ -1229,7 +1230,6 @@ static blk_qc_t flash_dev_make_request(struct request_queue *q, struct search *s; struct closure *cl; struct bcache_device *d = bio->bi_disk->private_data; - int rw = bio_data_dir(bio); if (unlikely(d->c && test_bit(CACHE_SET_IO_DISABLE, &d->c->flags))) { bio->bi_status = BLK_STS_IOERR; @@ -1237,7 +1237,7 @@ static blk_qc_t flash_dev_make_request(struct request_queue *q, return BLK_QC_T_NONE; } - generic_start_io_acct(q, rw, bio_sectors(bio), &d->disk->part0); + generic_start_io_acct(q, bio_op(bio), bio_sectors(bio), &d->disk->part0); s = search_alloc(bio, d); cl = &s->cl; @@ -1254,7 +1254,7 @@ static blk_qc_t flash_dev_make_request(struct request_queue *q, flash_dev_nodata, bcache_wq); return BLK_QC_T_NONE; - } else if (rw) { + } else if (bio_data_dir(bio)) { bch_keybuf_check_overlapping(&s->iop.c->moving_gc_keys, &KEY(d->id, bio->bi_iter.bi_sector, 0), &KEY(d->id, bio_end_sector(bio), 0)); diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c index fa4058e43202..e0a92104ca23 100644 --- a/drivers/md/bcache/super.c +++ b/drivers/md/bcache/super.c @@ -181,7 +181,7 @@ static const char *read_super(struct cache_sb *sb, struct block_device *bdev, goto err; } - sb->last_mount = get_seconds(); + sb->last_mount = (u32)ktime_get_real_seconds(); err = NULL; get_page(bh->b_page); @@ -701,7 +701,7 @@ static void bcache_device_detach(struct bcache_device *d) SET_UUID_FLASH_ONLY(u, 0); memcpy(u->uuid, invalid_uuid, 16); - u->invalidated = cpu_to_le32(get_seconds()); + u->invalidated = cpu_to_le32((u32)ktime_get_real_seconds()); bch_uuid_write(d->c); } @@ -796,11 +796,12 @@ static int bcache_device_init(struct bcache_device *d, unsigned block_size, return idx; if (bioset_init(&d->bio_split, 4, offsetof(struct bbio, bio), - BIOSET_NEED_BVECS|BIOSET_NEED_RESCUER) || - !(d->disk = alloc_disk(BCACHE_MINORS))) { - ida_simple_remove(&bcache_device_idx, idx); - return -ENOMEM; - } + BIOSET_NEED_BVECS|BIOSET_NEED_RESCUER)) + goto err; + + d->disk = alloc_disk(BCACHE_MINORS); + if (!d->disk) + goto err; set_capacity(d->disk, sectors); snprintf(d->disk->disk_name, DISK_NAME_LEN, "bcache%i", idx); @@ -834,6 +835,11 @@ static int bcache_device_init(struct bcache_device *d, unsigned block_size, blk_queue_write_cache(q, true, true); return 0; + +err: + ida_simple_remove(&bcache_device_idx, idx); + return -ENOMEM; + } /* Cached device */ @@ -1027,7 +1033,7 @@ void bch_cached_dev_detach(struct cached_dev *dc) int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c, uint8_t *set_uuid) { - uint32_t rtime = cpu_to_le32(get_seconds()); + uint32_t rtime = cpu_to_le32((u32)ktime_get_real_seconds()); struct uuid_entry *u; struct cached_dev *exist_dc, *t; @@ -1070,7 +1076,7 @@ int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c, (BDEV_STATE(&dc->sb) == BDEV_STATE_STALE || BDEV_STATE(&dc->sb) == BDEV_STATE_NONE)) { memcpy(u->uuid, invalid_uuid, 16); - u->invalidated = cpu_to_le32(get_seconds()); + u->invalidated = cpu_to_le32((u32)ktime_get_real_seconds()); u = NULL; } @@ -1311,6 +1317,8 @@ static void flash_dev_free(struct closure *cl) { struct bcache_device *d = container_of(cl, struct bcache_device, cl); mutex_lock(&bch_register_lock); + atomic_long_sub(bcache_dev_sectors_dirty(d), + &d->c->flash_dev_dirty_sectors); bcache_device_free(d); mutex_unlock(&bch_register_lock); kobject_put(&d->kobj); @@ -1390,7 +1398,7 @@ int bch_flash_dev_create(struct cache_set *c, uint64_t size) get_random_bytes(u->uuid, 16); memset(u->label, 0, 32); - u->first_reg = u->last_reg = cpu_to_le32(get_seconds()); + u->first_reg = u->last_reg = cpu_to_le32((u32)ktime_get_real_seconds()); SET_UUID_FLASH_ONLY(u, 1); u->sectors = size >> 9; @@ -1894,7 +1902,7 @@ static void run_cache_set(struct cache_set *c) goto err; closure_sync(&cl); - c->sb.last_mount = get_seconds(); + c->sb.last_mount = (u32)ktime_get_real_seconds(); bcache_write_super(c); list_for_each_entry_safe(dc, t, &uncached_devices, list) @@ -2163,8 +2171,12 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr, if (!try_module_get(THIS_MODULE)) return -EBUSY; - if (!(path = kstrndup(buffer, size, GFP_KERNEL)) || - !(sb = kmalloc(sizeof(struct cache_sb), GFP_KERNEL))) + path = kstrndup(buffer, size, GFP_KERNEL); + if (!path) + goto err; + + sb = kmalloc(sizeof(struct cache_sb), GFP_KERNEL); + if (!sb) goto err; err = "failed to open device"; @@ -2324,9 +2336,15 @@ static int __init bcache_init(void) return bcache_major; } - if (!(bcache_wq = alloc_workqueue("bcache", WQ_MEM_RECLAIM, 0)) || - !(bcache_kobj = kobject_create_and_add("bcache", fs_kobj)) || - bch_request_init() || + bcache_wq = alloc_workqueue("bcache", WQ_MEM_RECLAIM, 0); + if (!bcache_wq) + goto err; + + bcache_kobj = kobject_create_and_add("bcache", fs_kobj); + if (!bcache_kobj) + goto err; + + if (bch_request_init() || bch_debug_init(bcache_kobj) || closure_debug_init() || sysfs_create_files(bcache_kobj, files)) goto err; diff --git a/drivers/md/bcache/writeback.c b/drivers/md/bcache/writeback.c index ad45ebe1a74b..912e969fedba 100644 --- a/drivers/md/bcache/writeback.c +++ b/drivers/md/bcache/writeback.c @@ -27,7 +27,7 @@ static uint64_t __calc_target_rate(struct cached_dev *dc) * flash-only devices */ uint64_t cache_sectors = c->nbuckets * c->sb.bucket_size - - bcache_flash_devs_sectors_dirty(c); + atomic_long_read(&c->flash_dev_dirty_sectors); /* * Unfortunately there is no control of global dirty data. If the @@ -476,6 +476,9 @@ void bcache_dev_sectors_dirty_add(struct cache_set *c, unsigned inode, if (!d) return; + if (UUID_FLASH_ONLY(&c->uuids[inode])) + atomic_long_add(nr_sectors, &c->flash_dev_dirty_sectors); + stripe = offset_to_stripe(d, offset); stripe_offset = offset & (d->stripe_size - 1); @@ -673,10 +676,14 @@ static int bch_writeback_thread(void *arg) } /* Init */ +#define INIT_KEYS_EACH_TIME 500000 +#define INIT_KEYS_SLEEP_MS 100 struct sectors_dirty_init { struct btree_op op; unsigned inode; + size_t count; + struct bkey start; }; static int sectors_dirty_init_fn(struct btree_op *_op, struct btree *b, @@ -691,18 +698,37 @@ static int sectors_dirty_init_fn(struct btree_op *_op, struct btree *b, bcache_dev_sectors_dirty_add(b->c, KEY_INODE(k), KEY_START(k), KEY_SIZE(k)); + op->count++; + if (atomic_read(&b->c->search_inflight) && + !(op->count % INIT_KEYS_EACH_TIME)) { + bkey_copy_key(&op->start, k); + return -EAGAIN; + } + return MAP_CONTINUE; } void bch_sectors_dirty_init(struct bcache_device *d) { struct sectors_dirty_init op; + int ret; bch_btree_op_init(&op.op, -1); op.inode = d->id; - - bch_btree_map_keys(&op.op, d->c, &KEY(op.inode, 0, 0), - sectors_dirty_init_fn, 0); + op.count = 0; + op.start = KEY(op.inode, 0, 0); + + do { + ret = bch_btree_map_keys(&op.op, d->c, &op.start, + sectors_dirty_init_fn, 0); + if (ret == -EAGAIN) + schedule_timeout_interruptible( + msecs_to_jiffies(INIT_KEYS_SLEEP_MS)); + else if (ret < 0) { + pr_warn("sectors dirty init failed, ret=%d!", ret); + break; + } + } while (ret == -EAGAIN); } void bch_cached_dev_writeback_init(struct cached_dev *dc) diff --git a/drivers/md/bcache/writeback.h b/drivers/md/bcache/writeback.h index 610fb01de629..3745d7004c47 100644 --- a/drivers/md/bcache/writeback.h +++ b/drivers/md/bcache/writeback.h @@ -28,25 +28,6 @@ static inline uint64_t bcache_dev_sectors_dirty(struct bcache_device *d) return ret; } -static inline uint64_t bcache_flash_devs_sectors_dirty(struct cache_set *c) -{ - uint64_t i, ret = 0; - - mutex_lock(&bch_register_lock); - - for (i = 0; i < c->devices_max_used; i++) { - struct bcache_device *d = c->devices[i]; - - if (!d || !UUID_FLASH_ONLY(&c->uuids[i])) - continue; - ret += bcache_dev_sectors_dirty(d); - } - - mutex_unlock(&bch_register_lock); - - return ret; -} - static inline unsigned offset_to_stripe(struct bcache_device *d, uint64_t offset) { diff --git a/drivers/md/dm.c b/drivers/md/dm.c index b0dd7027848b..20f7e4ef5342 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -609,7 +609,8 @@ static void start_io_acct(struct dm_io *io) io->start_time = jiffies; - generic_start_io_acct(md->queue, rw, bio_sectors(bio), &dm_disk(md)->part0); + generic_start_io_acct(md->queue, bio_op(bio), bio_sectors(bio), + &dm_disk(md)->part0); atomic_set(&dm_disk(md)->part0.in_flight[rw], atomic_inc_return(&md->pending[rw])); @@ -628,7 +629,8 @@ static void end_io_acct(struct dm_io *io) int pending; int rw = bio_data_dir(bio); - generic_end_io_acct(md->queue, rw, &dm_disk(md)->part0, io->start_time); + generic_end_io_acct(md->queue, bio_op(bio), &dm_disk(md)->part0, + io->start_time); if (unlikely(dm_stats_used(&md->stats))) dm_stats_account_io(&md->stats, bio_data_dir(bio), diff --git a/drivers/md/md.c b/drivers/md/md.c index 994aed2f9dff..cb4eb5faa519 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -204,10 +204,6 @@ static int start_readonly; */ static bool create_on_open = true; -/* bio_clone_mddev - * like bio_clone_bioset, but with a local bio set - */ - struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs, struct mddev *mddev) { @@ -335,6 +331,7 @@ EXPORT_SYMBOL(md_handle_request); static blk_qc_t md_make_request(struct request_queue *q, struct bio *bio) { const int rw = bio_data_dir(bio); + const int sgrp = op_stat_group(bio_op(bio)); struct mddev *mddev = q->queuedata; unsigned int sectors; int cpu; @@ -363,8 +360,8 @@ static blk_qc_t md_make_request(struct request_queue *q, struct bio *bio) md_handle_request(mddev, bio); cpu = part_stat_lock(); - part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]); - part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw], sectors); + part_stat_inc(cpu, &mddev->gendisk->part0, ios[sgrp]); + part_stat_add(cpu, &mddev->gendisk->part0, sectors[sgrp], sectors); part_stat_unlock(); return BLK_QC_T_NONE; @@ -8046,8 +8043,7 @@ static int is_mddev_idle(struct mddev *mddev, int init) rcu_read_lock(); rdev_for_each_rcu(rdev, mddev) { struct gendisk *disk = rdev->bdev->bd_contains->bd_disk; - curr_events = (int)part_stat_read(&disk->part0, sectors[0]) + - (int)part_stat_read(&disk->part0, sectors[1]) - + 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 |