diff options
Diffstat (limited to 'drivers/md/raid10.c')
| -rw-r--r-- | drivers/md/raid10.c | 2250 |
1 files changed, 1770 insertions, 480 deletions
diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c index 6e8aa213f0d..cb882aae9e2 100644 --- a/drivers/md/raid10.c +++ b/drivers/md/raid10.c @@ -24,6 +24,7 @@ #include <linux/module.h> #include <linux/seq_file.h> #include <linux/ratelimit.h> +#include <linux/kthread.h> #include "md.h" #include "raid10.h" #include "raid0.h" @@ -37,21 +38,36 @@ * near_copies (stored in low byte of layout) * far_copies (stored in second byte of layout) * far_offset (stored in bit 16 of layout ) + * use_far_sets (stored in bit 17 of layout ) * - * The data to be stored is divided into chunks using chunksize. - * Each device is divided into far_copies sections. - * In each section, chunks are laid out in a style similar to raid0, but - * near_copies copies of each chunk is stored (each on a different drive). - * The starting device for each section is offset near_copies from the starting - * device of the previous section. - * Thus they are (near_copies*far_copies) of each chunk, and each is on a different - * drive. - * near_copies and far_copies must be at least one, and their product is at most - * raid_disks. + * The data to be stored is divided into chunks using chunksize. Each device + * is divided into far_copies sections. In each section, chunks are laid out + * in a style similar to raid0, but near_copies copies of each chunk is stored + * (each on a different drive). The starting device for each section is offset + * near_copies from the starting device of the previous section. Thus there + * are (near_copies * far_copies) of each chunk, and each is on a different + * drive. near_copies and far_copies must be at least one, and their product + * is at most raid_disks. * * If far_offset is true, then the far_copies are handled a bit differently. - * The copies are still in different stripes, but instead of be very far apart - * on disk, there are adjacent stripes. + * The copies are still in different stripes, but instead of being very far + * apart on disk, there are adjacent stripes. + * + * The far and offset algorithms are handled slightly differently if + * 'use_far_sets' is true. In this case, the array's devices are grouped into + * sets that are (near_copies * far_copies) in size. The far copied stripes + * are still shifted by 'near_copies' devices, but this shifting stays confined + * to the set rather than the entire array. This is done to improve the number + * of device combinations that can fail without causing the array to fail. + * Example 'far' algorithm w/o 'use_far_sets' (each letter represents a chunk + * on a device): + * A B C D A B C D E + * ... ... + * D A B C E A B C D + * Example 'far' algorithm w/ 'use_far_sets' enabled (sets illustrated w/ []'s): + * [A B] [C D] [A B] [C D E] + * |...| |...| |...| | ... | + * [B A] [D C] [B A] [E C D] */ /* @@ -59,7 +75,21 @@ */ #define NR_RAID10_BIOS 256 -/* When there are this many requests queue to be written by +/* when we get a read error on a read-only array, we redirect to another + * device without failing the first device, or trying to over-write to + * correct the read error. To keep track of bad blocks on a per-bio + * level, we store IO_BLOCKED in the appropriate 'bios' pointer + */ +#define IO_BLOCKED ((struct bio *)1) +/* When we successfully write to a known bad-block, we need to remove the + * bad-block marking which must be done from process context. So we record + * the success by setting devs[n].bio to IO_MADE_GOOD + */ +#define IO_MADE_GOOD ((struct bio *)2) + +#define BIO_SPECIAL(bio) ((unsigned long)bio <= 2) + +/* When there are this many requests queued to be written by * the raid10 thread, we become 'congested' to provide back-pressure * for writeback. */ @@ -67,6 +97,12 @@ static int max_queued_requests = 1024; static void allow_barrier(struct r10conf *conf); static void lower_barrier(struct r10conf *conf); +static int _enough(struct r10conf *conf, int previous, int ignore); +static sector_t reshape_request(struct mddev *mddev, sector_t sector_nr, + int *skipped); +static void reshape_request_write(struct mddev *mddev, struct r10bio *r10_bio); +static void end_reshape_write(struct bio *bio, int error); +static void end_reshape(struct r10conf *conf); static void * r10bio_pool_alloc(gfp_t gfp_flags, void *data) { @@ -111,7 +147,8 @@ static void * r10buf_pool_alloc(gfp_t gfp_flags, void *data) if (!r10_bio) return NULL; - if (test_bit(MD_RECOVERY_SYNC, &conf->mddev->recovery)) + if (test_bit(MD_RECOVERY_SYNC, &conf->mddev->recovery) || + test_bit(MD_RECOVERY_RESHAPE, &conf->mddev->recovery)) nalloc = conf->copies; /* resync */ else nalloc = 2; /* recovery */ @@ -139,9 +176,10 @@ static void * r10buf_pool_alloc(gfp_t gfp_flags, void *data) struct bio *rbio = r10_bio->devs[j].repl_bio; bio = r10_bio->devs[j].bio; for (i = 0; i < RESYNC_PAGES; i++) { - if (j == 1 && !test_bit(MD_RECOVERY_SYNC, - &conf->mddev->recovery)) { - /* we can share bv_page's during recovery */ + if (j > 0 && !test_bit(MD_RECOVERY_SYNC, + &conf->mddev->recovery)) { + /* we can share bv_page's during recovery + * and reshape */ struct bio *rbio = r10_bio->devs[0].bio; page = rbio->bi_io_vec[i].bv_page; get_page(page); @@ -164,10 +202,11 @@ out_free_pages: while (j--) for (i = 0; i < RESYNC_PAGES ; i++) safe_put_page(r10_bio->devs[j].bio->bi_io_vec[i].bv_page); - j = -1; + j = 0; out_free_bio: - while (++j < nalloc) { - bio_put(r10_bio->devs[j].bio); + for ( ; j < nalloc; j++) { + if (r10_bio->devs[j].bio) + bio_put(r10_bio->devs[j].bio); if (r10_bio->devs[j].repl_bio) bio_put(r10_bio->devs[j].repl_bio); } @@ -347,6 +386,17 @@ static void raid10_end_read_request(struct bio *bio, int error) * wait for the 'master' bio. */ set_bit(R10BIO_Uptodate, &r10_bio->state); + } else { + /* If all other devices that store this block 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" + */ + if (!_enough(conf, test_bit(R10BIO_Previous, &r10_bio->state), + rdev->raid_disk)) + uptodate = 1; + } + if (uptodate) { raid_end_bio_io(r10_bio); rdev_dec_pending(rdev, conf->mddev); } else { @@ -438,7 +488,17 @@ static void raid10_end_write_request(struct bio *bio, int error) sector_t first_bad; int bad_sectors; - set_bit(R10BIO_Uptodate, &r10_bio->state); + /* + * Do not set R10BIO_Uptodate if the current device is + * rebuilding or Faulty. This is because we cannot use + * such device for properly reading the data back (we could + * potentially use it, if the current write would have felt + * before rdev->recovery_offset, but for simplicity we don't + * check this here. + */ + if (test_bit(In_sync, &rdev->flags) && + !test_bit(Faulty, &rdev->flags)) + set_bit(R10BIO_Uptodate, &r10_bio->state); /* Maybe we can clear some bad blocks. */ if (is_badblock(rdev, @@ -462,7 +522,7 @@ static void raid10_end_write_request(struct bio *bio, int error) */ one_write_done(r10_bio); if (dec_rdev) - rdev_dec_pending(conf->mirrors[dev].rdev, conf->mddev); + rdev_dec_pending(rdev, conf->mddev); } /* @@ -490,79 +550,127 @@ static void raid10_end_write_request(struct bio *bio, int error) * sector offset to a virtual address */ -static void raid10_find_phys(struct r10conf *conf, struct r10bio *r10bio) +static void __raid10_find_phys(struct geom *geo, struct r10bio *r10bio) { int n,f; sector_t sector; sector_t chunk; sector_t stripe; int dev; - int slot = 0; + int last_far_set_start, last_far_set_size; + + last_far_set_start = (geo->raid_disks / geo->far_set_size) - 1; + last_far_set_start *= geo->far_set_size; + + last_far_set_size = geo->far_set_size; + last_far_set_size += (geo->raid_disks % geo->far_set_size); /* now calculate first sector/dev */ - chunk = r10bio->sector >> conf->chunk_shift; - sector = r10bio->sector & conf->chunk_mask; + chunk = r10bio->sector >> geo->chunk_shift; + sector = r10bio->sector & geo->chunk_mask; - chunk *= conf->near_copies; + chunk *= geo->near_copies; stripe = chunk; - dev = sector_div(stripe, conf->raid_disks); - if (conf->far_offset) - stripe *= conf->far_copies; + dev = sector_div(stripe, geo->raid_disks); + if (geo->far_offset) + stripe *= geo->far_copies; - sector += stripe << conf->chunk_shift; + sector += stripe << geo->chunk_shift; /* and calculate all the others */ - for (n=0; n < conf->near_copies; n++) { + for (n = 0; n < geo->near_copies; n++) { int d = dev; + int set; sector_t s = sector; - r10bio->devs[slot].addr = sector; r10bio->devs[slot].devnum = d; + r10bio->devs[slot].addr = s; slot++; - for (f = 1; f < conf->far_copies; f++) { - d += conf->near_copies; - if (d >= conf->raid_disks) - d -= conf->raid_disks; - s += conf->stride; + for (f = 1; f < geo->far_copies; f++) { + set = d / geo->far_set_size; + d += geo->near_copies; + + if ((geo->raid_disks % geo->far_set_size) && + (d > last_far_set_start)) { + d -= last_far_set_start; + d %= last_far_set_size; + d += last_far_set_start; + } else { + d %= geo->far_set_size; + d += geo->far_set_size * set; + } + s += geo->stride; r10bio->devs[slot].devnum = d; r10bio->devs[slot].addr = s; slot++; } dev++; - if (dev >= conf->raid_disks) { + if (dev >= geo->raid_disks) { dev = 0; - sector += (conf->chunk_mask + 1); + sector += (geo->chunk_mask + 1); } } - BUG_ON(slot != conf->copies); +} + +static void raid10_find_phys(struct r10conf *conf, struct r10bio *r10bio) +{ + struct geom *geo = &conf->geo; + + if (conf->reshape_progress != MaxSector && + ((r10bio->sector >= conf->reshape_progress) != + conf->mddev->reshape_backwards)) { + set_bit(R10BIO_Previous, &r10bio->state); + geo = &conf->prev; + } else + clear_bit(R10BIO_Previous, &r10bio->state); + + __raid10_find_phys(geo, r10bio); } static sector_t raid10_find_virt(struct r10conf *conf, sector_t sector, int dev) { sector_t offset, chunk, vchunk; + /* Never use conf->prev as this is only called during resync + * or recovery, so reshape isn't happening + */ + struct geom *geo = &conf->geo; + int far_set_start = (dev / geo->far_set_size) * geo->far_set_size; + int far_set_size = geo->far_set_size; + int last_far_set_start; + + if (geo->raid_disks % geo->far_set_size) { + last_far_set_start = (geo->raid_disks / geo->far_set_size) - 1; + last_far_set_start *= geo->far_set_size; + + if (dev >= last_far_set_start) { + far_set_size = geo->far_set_size; + far_set_size += (geo->raid_disks % geo->far_set_size); + far_set_start = last_far_set_start; + } + } - offset = sector & conf->chunk_mask; - if (conf->far_offset) { + offset = sector & geo->chunk_mask; + if (geo->far_offset) { int fc; - chunk = sector >> conf->chunk_shift; - fc = sector_div(chunk, conf->far_copies); - dev -= fc * conf->near_copies; - if (dev < 0) - dev += conf->raid_disks; + chunk = sector >> geo->chunk_shift; + fc = sector_div(chunk, geo->far_copies); + dev -= fc * geo->near_copies; + if (dev < far_set_start) + dev += far_set_size; } else { - while (sector >= conf->stride) { - sector -= conf->stride; - if (dev < conf->near_copies) - dev += conf->raid_disks - conf->near_copies; + while (sector >= geo->stride) { + sector -= geo->stride; + if (dev < (geo->near_copies + far_set_start)) + dev += far_set_size - geo->near_copies; else - dev -= conf->near_copies; + dev -= geo->near_copies; } - chunk = sector >> conf->chunk_shift; + chunk = sector >> geo->chunk_shift; } - vchunk = chunk * conf->raid_disks + dev; - sector_div(vchunk, conf->near_copies); - return (vchunk << conf->chunk_shift) + offset; + vchunk = chunk * geo->raid_disks + dev; + sector_div(vchunk, geo->near_copies); + return (vchunk << geo->chunk_shift) + offset; } /** @@ -572,25 +680,85 @@ static sector_t raid10_find_virt(struct r10conf *conf, sector_t sector, int dev) * @biovec: the request that could be merged to it. * * Return amount of bytes we can accept at this offset - * If near_copies == raid_disk, there are no striping issues, - * but in that case, the function isn't called at all. + * This requires checking for end-of-chunk if near_copies != raid_disks, + * and for subordinate merge_bvec_fns if merge_check_needed. */ static int raid10_mergeable_bvec(struct request_queue *q, struct bvec_merge_data *bvm, struct bio_vec *biovec) { struct mddev *mddev = q->queuedata; + struct r10conf *conf = mddev->private; sector_t sector = bvm->bi_sector + get_start_sect(bvm->bi_bdev); int max; - unsigned int chunk_sectors = mddev->chunk_sectors; + unsigned int chunk_sectors; unsigned int bio_sectors = bvm->bi_size >> 9; - - max = (chunk_sectors - ((sector & (chunk_sectors - 1)) + bio_sectors)) << 9; - if (max < 0) max = 0; /* bio_add cannot handle a negative return */ - if (max <= biovec->bv_len && bio_sectors == 0) - return biovec->bv_len; - else - return max; + struct geom *geo = &conf->geo; + + chunk_sectors = (conf->geo.chunk_mask & conf->prev.chunk_mask) + 1; + if (conf->reshape_progress != MaxSector && + ((sector >= conf->reshape_progress) != + conf->mddev->reshape_backwards)) + geo = &conf->prev; + + if (geo->near_copies < geo->raid_disks) { + max = (chunk_sectors - ((sector & (chunk_sectors - 1)) + + bio_sectors)) << 9; + if (max < 0) + /* bio_add cannot handle a negative return */ + max = 0; + if (max <= biovec->bv_len && bio_sectors == 0) + return biovec->bv_len; + } else + max = biovec->bv_len; + + if (mddev->merge_check_needed) { + struct { + struct r10bio r10_bio; + struct r10dev devs[conf->copies]; + } on_stack; + struct r10bio *r10_bio = &on_stack.r10_bio; + int s; + if (conf->reshape_progress != MaxSector) { + /* Cannot give any guidance during reshape */ + if (max <= biovec->bv_len && bio_sectors == 0) + return biovec->bv_len; + return 0; + } + r10_bio->sector = sector; + raid10_find_phys(conf, r10_bio); + rcu_read_lock(); + for (s = 0; s < conf->copies; s++) { + int disk = r10_bio->devs[s].devnum; + struct md_rdev *rdev = rcu_dereference( + conf->mirrors[disk].rdev); + if (rdev && !test_bit(Faulty, &rdev->flags)) { + struct request_queue *q = + bdev_get_queue(rdev->bdev); + if (q->merge_bvec_fn) { + bvm->bi_sector = r10_bio->devs[s].addr + + rdev->data_offset; + bvm->bi_bdev = rdev->bdev; + max = min(max, q->merge_bvec_fn( + q, bvm, biovec)); + } + } + rdev = rcu_dereference(conf->mirrors[disk].replacement); + if (rdev && !test_bit(Faulty, &rdev->flags)) { + struct request_queue *q = + bdev_get_queue(rdev->bdev); + if (q->merge_bvec_fn) { + bvm->bi_sector = r10_bio->devs[s].addr + + rdev->data_offset; + bvm->bi_bdev = rdev->bdev; + max = min(max, q->merge_bvec_fn( + q, bvm, biovec)); + } + } + } + rcu_read_unlock(); + } + return max; } /* @@ -621,9 +789,10 @@ static struct md_rdev *read_balance(struct r10conf *conf, int sectors = r10_bio->sectors; int best_good_sectors; sector_t new_distance, best_dist; - struct md_rdev *rdev, *best_rdev; + struct md_rdev *best_rdev, *rdev = NULL; int do_balance; int best_slot; + struct geom *geo = &conf->geo; raid10_find_phys(conf, r10_bio); rcu_read_lock(); @@ -654,11 +823,12 @@ retry: disk = r10_bio->devs[slot].devnum; rdev = rcu_dereference(conf->mirrors[disk].replacement); if (rdev == NULL || test_bit(Faulty, &rdev->flags) || + test_bit(Unmerged, &rdev->flags) || r10_bio->devs[slot].addr + sectors > rdev->recovery_offset) rdev = rcu_dereference(conf->mirrors[disk].rdev); - if (rdev == NULL) - continue; - if (test_bit(Faulty, &rdev->flags)) + if (rdev == NULL || + test_bit(Faulty, &rdev->flags) || + test_bit(Unmerged, &rdev->flags)) continue; if (!test_bit(In_sync, &rdev->flags) && r10_bio->devs[slot].addr + sectors > rdev->recovery_offset) @@ -703,11 +873,11 @@ retry: * sequential read speed for 'far copies' arrays. So only * keep it for 'near' arrays, and review those later. */ - if (conf->near_copies > 1 && !atomic_read(&rdev->nr_pending)) + if (geo->near_copies > 1 && !atomic_read(&rdev->nr_pending)) break; /* for far > 1 always use the lowest address */ - if (conf->far_copies > 1) + if (geo->far_copies > 1) new_distance = r10_bio->devs[slot].addr; else new_distance = abs(r10_bio->devs[slot].addr - @@ -741,9 +911,8 @@ retry: return rdev; } -static int raid10_congested(void *data, int bits) +int md_raid10_congested(struct mddev *mddev, int bits) { - struct mddev *mddev = data; struct r10conf *conf = mddev->private; int i, ret = 0; @@ -751,10 +920,11 @@ static int raid10_congested(void *data, int bits) conf->pending_count >= max_queued_requests) return 1; - if (mddev_congested(mddev, bits)) - return 1; rcu_read_lock(); - for (i = 0; i < conf->raid_disks && ret == 0; i++) { + for (i = 0; + (i < conf->geo.raid_disks || i < conf->prev.raid_disks) + && ret == 0; + i++) { struct md_rdev *rdev = rcu_dereference(conf->mirrors[i].rdev); if (rdev && !test_bit(Faulty, &rdev->flags)) { struct request_queue *q = bdev_get_queue(rdev->bdev); @@ -765,6 +935,15 @@ static int raid10_congested(void *data, int bits) rcu_read_unlock(); return ret; } +EXPORT_SYMBOL_GPL(md_raid10_congested); + +static int raid10_congested(void *data, int bits) +{ + struct mddev *mddev = data; + + return mddev_congested(mddev, bits) || + md_raid10_congested(mddev, bits); +} static void flush_pending_writes(struct r10conf *conf) { @@ -786,7 +965,12 @@ static void flush_pending_writes(struct r10conf *conf) while (bio) { /* submit pending writes */ struct bio *next = bio->bi_next; bio->bi_next = NULL; - generic_make_request(bio); + if (unlikely((bio->bi_rw & REQ_DISCARD) && + !blk_queue_discard(bdev_get_queue(bio->bi_bdev)))) + /* Just ignore it */ + bio_endio(bio, 0); + else + generic_make_request(bio); bio = next; } } else @@ -822,7 +1006,7 @@ static void raise_barrier(struct r10conf *conf, int force) /* Wait until no block IO is waiting (unless 'force') */ wait_event_lock_irq(conf->wait_barrier, force || !conf->nr_waiting, - conf->resync_lock, ); + conf->resync_lock); /* block any new IO from starting */ conf->barrier++; @@ -830,7 +1014,7 @@ static void raise_barrier(struct r10conf *conf, int force) /* Now wait for all pending IO to complete */ wait_event_lock_irq(conf->wait_barrier, !conf->nr_pending && conf->barrier < RESYNC_DEPTH, - conf->resync_lock, ); + conf->resync_lock); spin_unlock_irq(&conf->resync_lock); } @@ -849,9 +1033,21 @@ static void wait_barrier(struct r10conf *conf) spin_lock_irq(&conf->resync_lock); if (conf->barrier) { conf->nr_waiting++; - wait_event_lock_irq(conf->wait_barrier, !conf->barrier, - conf->resync_lock, - ); + /* Wait for the barrier to drop. + * However if there are already pending + * requests (preventing the barrier from + * rising completely), and the + * pre-process bio queue isn't empty, + * then don't wait, as we need to empty + * that queue to get the nr_pending + * count down. + */ + wait_event_lock_irq(conf->wait_barrier, + !conf->barrier || + (conf->nr_pending && + current->bio_list && + !bio_list_empty(current->bio_list)), + conf->resync_lock); conf->nr_waiting--; } conf->nr_pending++; @@ -867,27 +1063,27 @@ static void allow_barrier(struct r10conf *conf) wake_up(&conf->wait_barrier); } -static void freeze_array(struct r10conf *conf) +static void freeze_array(struct r10conf *conf, int extra) { /* stop syncio and normal IO and wait for everything to * go quiet. * We increment barrier and nr_waiting, and then - * wait until nr_pending match nr_queued+1 + * wait until nr_pending match nr_queued+extra * This is called in the context of one normal IO request * that has failed. Thus any sync request that might be pending * will be blocked by nr_pending, and we need to wait for * pending IO requests to complete or be queued for re-try. - * Thus the number queued (nr_queued) plus this request (1) + * Thus the number queued (nr_queued) plus this request (extra) * must match the number of pending IOs (nr_pending) before * we continue. */ spin_lock_irq(&conf->resync_lock); conf->barrier++; conf->nr_waiting++; - wait_event_lock_irq(conf->wait_barrier, - conf->nr_pending == conf->nr_queued+1, - conf->resync_lock, - flush_pending_writes(conf)); + wait_event_lock_irq_cmd(conf->wait_barrier, + conf->nr_pending == conf->nr_queued+extra, + conf->resync_lock, + flush_pending_writes(conf)); spin_unlock_irq(&conf->resync_lock); } @@ -902,76 +1098,79 @@ static void unfreeze_array(struct r10conf *conf) spin_unlock_irq(&conf->resync_lock); } -static void make_request(struct mddev *mddev, struct bio * bio) +static sector_t choose_data_offset(struct r10bio *r10_bio, + struct md_rdev *rdev) +{ + if (!test_bit(MD_RECOVERY_RESHAPE, &rdev->mddev->recovery) || + test_bit(R10BIO_Previous, &r10_bio->state)) + return rdev->data_offset; + else + return rdev->new_data_offset; +} + +struct raid10_plug_cb { + struct blk_plug_cb cb; + struct bio_list pending; + int pending_cnt; +}; + +static void raid10_unplug(struct blk_plug_cb *cb, bool from_schedule) +{ + struct raid10_plug_cb *plug = container_of(cb, struct raid10_plug_cb, + cb); + struct mddev *mddev = plug->cb.data; + struct r10conf *conf = mddev->private; + struct bio *bio; + + if (from_schedule || current->bio_list) { + spin_lock_irq(&conf->device_lock); + bio_list_merge(&conf->pending_bio_list, &plug->pending); + conf->pending_count += plug->pending_cnt; + spin_unlock_irq(&conf->device_lock); + wake_up(&conf->wait_barrier); + md_wakeup_thread(mddev->thread); + kfree(plug); + return; + } + + /* we aren't scheduling, so we can do the write-out directly. */ + bio = bio_list_get(&plug->pending); + bitmap_unplug(mddev->bitmap); + wake_up(&conf->wait_barrier); + + while (bio) { /* submit pending writes */ + struct bio *next = bio->bi_next; + bio->bi_next = NULL; + if (unlikely((bio->bi_rw & REQ_DISCARD) && + !blk_queue_discard(bdev_get_queue(bio->bi_bdev)))) + /* Just ignore it */ + bio_endio(bio, 0); + else + generic_make_request(bio); + bio = next; + } + kfree(plug); +} + +static void __make_request(struct mddev *mddev, struct bio *bio) { struct r10conf *conf = mddev->private; struct r10bio *r10_bio; struct bio *read_bio; int i; - int chunk_sects = conf->chunk_mask + 1; const int rw = bio_data_dir(bio); const unsigned long do_sync = (bio->bi_rw & REQ_SYNC); const unsigned long do_fua = (bio->bi_rw & REQ_FUA); + const unsigned long do_discard = (bio->bi_rw + & (REQ_DISCARD | REQ_SECURE)); + const unsigned long do_same = (bio->bi_rw & REQ_WRITE_SAME); unsigned long flags; struct md_rdev *blocked_rdev; - int plugged; + struct blk_plug_cb *cb; + struct raid10_plug_cb *plug = NULL; int sectors_handled; int max_sectors; - - if (unlikely(bio->bi_rw & REQ_FLUSH)) { - md_flush_request(mddev, bio); - return; - } - - /* If this request crosses a chunk boundary, we need to - * split it. This will only happen for 1 PAGE (or less) requests. - */ - if (unlikely( (bio->bi_sector & conf->chunk_mask) + (bio->bi_size >> 9) - > chunk_sects && - conf->near_copies < conf->raid_disks)) { - struct bio_pair *bp; - /* Sanity check -- queue functions should prevent this happening */ - if (bio->bi_vcnt != 1 || - bio->bi_idx != 0) - goto bad_map; - /* This is a one page bio that upper layers - * refuse to split for us, so we need to split it. - */ - bp = bio_split(bio, - chunk_sects - (bio->bi_sector & (chunk_sects - 1)) ); - - /* Each of these 'make_request' calls will call 'wait_barrier'. - * If the first succeeds but the second blocks due to the resync - * thread raising the barrier, we will deadlock because the - * IO to the underlying device will be queued in generic_make_request - * and will never complete, so will never reduce nr_pending. - * So increment nr_waiting here so no new raise_barriers will - * succeed, and so the second wait_barrier cannot block. - */ - spin_lock_irq(&conf->resync_lock); - conf->nr_waiting++; - spin_unlock_irq(&conf->resync_lock); - - make_request(mddev, &bp->bio1); - make_request(mddev, &bp->bio2); - - spin_lock_irq(&conf->resync_lock); - conf->nr_waiting--; - wake_up(&conf->wait_barrier); - spin_unlock_irq(&conf->resync_lock); - - bio_pair_release(bp); - return; - bad_map: - printk("md/raid10:%s: make_request bug: can't convert block across chunks" - " or bigger than %dk %llu %d\n", mdname(mddev), chunk_sects/2, - (unsigned long long)bio->bi_sector, bio->bi_size >> 10); - - bio_io_error(bio); - return; - } - - md_write_start(mddev, bio); + int sectors; /* * Register the new request and wait if the reconstruction @@ -980,13 +1179,45 @@ static void make_request(struct mddev *mddev, struct bio * bio) */ wait_barrier(conf); + sectors = bio_sectors(bio); + while (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery) && + bio->bi_iter.bi_sector < conf->reshape_progress && + bio->bi_iter.bi_sector + sectors > conf->reshape_progress) { + /* IO spans the reshape position. Need to wait for + * reshape to pass + */ + allow_barrier(conf); + wait_event(conf->wait_barrier, + conf->reshape_progress <= bio->bi_iter.bi_sector || + conf->reshape_progress >= bio->bi_iter.bi_sector + + sectors); + wait_barrier(conf); + } + if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery) && + bio_data_dir(bio) == WRITE && + (mddev->reshape_backwards + ? (bio->bi_iter.bi_sector < conf->reshape_safe && + bio->bi_iter.bi_sector + sectors > conf->reshape_progress) + : (bio->bi_iter.bi_sector + sectors > conf->reshape_safe && + bio->bi_iter.bi_sector < conf->reshape_progress))) { + /* Need to update reshape_position in metadata */ + mddev->reshape_position = conf->reshape_progress; + set_bit(MD_CHANGE_DEVS, &mddev->flags); + set_bit(MD_CHANGE_PENDING, &mddev->flags); + md_wakeup_thread(mddev->thread); + wait_event(mddev->sb_wait, + !test_bit(MD_CHANGE_PENDING, &mddev->flags)); + + conf->reshape_safe = mddev->reshape_position; + } + r10_bio = mempool_alloc(conf->r10bio_pool, GFP_NOIO); r10_bio->master_bio = bio; - r10_bio->sectors = bio->bi_size >> 9; + r10_bio->sectors = sectors; r10_bio->mddev = mddev; - r10_bio->sector = bio->bi_sector; + r10_bio->sector = bio->bi_iter.bi_sector; r10_bio->state = 0; /* We might need to issue multiple reads to different @@ -1015,14 +1246,14 @@ read_again: slot = r10_bio->read_slot; read_bio = bio_clone_mddev(bio, GFP_NOIO, mddev); - md_trim_bio(read_bio, r10_bio->sector - bio->bi_sector, - max_sectors); + bio_trim(read_bio, r10_bio->sector - bio->bi_iter.bi_sector, + max_sectors); r10_bio->devs[slot].bio = read_bio; r10_bio->devs[slot].rdev = rdev; - read_bio->bi_sector = r10_bio->devs[slot].addr + - rdev->data_offset; + read_bio->bi_iter.bi_sector = r10_bio->devs[slot].addr + + choose_data_offset(r10_bio, rdev); read_bio->bi_bdev = rdev->bdev; read_bio->bi_end_io = raid10_end_read_request; read_bio->bi_rw = READ | do_sync; @@ -1032,15 +1263,15 @@ read_again: /* Could not read all from this device, so we will * need another r10_bio. */ - sectors_handled = (r10_bio->sectors + max_sectors - - bio->bi_sector); + sectors_handled = (r10_bio->sector + max_sectors + - bio->bi_iter.bi_sector); r10_bio->sectors = max_sectors; spin_lock_irq(&conf->device_lock); if (bio->bi_phys_segments == 0) bio->bi_phys_segments = 2; else bio->bi_phys_segments++; - spin_unlock(&conf->device_lock); + spin_unlock_irq(&conf->device_lock); /* Cannot call generic_make_request directly * as that will be queued in __generic_make_request * and subsequent mempool_alloc might block @@ -1051,11 +1282,11 @@ read_again: r10_bio = mempool_alloc(conf->r10bio_pool, GFP_NOIO); r10_bio->master_bio = bio; - r10_bio->sectors = ((bio->bi_size >> 9) - - sectors_handled); + r10_bio->sectors = bio_sectors(bio) - sectors_handled; r10_bio->state = 0; r10_bio->mddev = mddev; - r10_bio->sector = bio->bi_sector + sectors_handled; + r10_bio->sector = bio->bi_iter.bi_sector + + sectors_handled; goto read_again; } else generic_make_request(read_bio); @@ -1081,7 +1312,6 @@ read_again: * of r10_bios is recored in bio->bi_phys_segments just as with * the read case. */ - plugged = mddev_check_plugged(mddev); r10_bio->read_slot = -1; /* make sure repl_bio gets freed */ raid10_find_phys(conf, r10_bio); @@ -1107,16 +1337,21 @@ retry_write: blocked_rdev = rrdev; break; } - if (rrdev && test_bit(Faulty, &rrdev->flags)) + if (rdev && (test_bit(Faulty, &rdev->flags) + || test_bit(Unmerged, &rdev->flags))) + rdev = NULL; + if (rrdev && (test_bit(Faulty, &rrdev->flags) + || test_bit(Unmerged, &rrdev->flags))) rrdev = NULL; r10_bio->devs[i].bio = NULL; r10_bio->devs[i].repl_bio = NULL; - if (!rdev || test_bit(Faulty, &rdev->flags)) { + + if (!rdev && !rrdev) { set_bit(R10BIO_Degraded, &r10_bio->state); continue; } - if (test_bit(WriteErrorSeen, &rdev->flags)) { + if (rdev && test_bit(WriteErrorSeen, &rdev->flags)) { sector_t first_bad; sector_t dev_sector = r10_bio->devs[i].addr; int bad_sectors; @@ -1158,8 +1393,10 @@ retry_write: max_sectors = good_sectors; } } - r10_bio->devs[i].bio = bio; - atomic_inc(&rdev->nr_pending); + if (rdev) { + r10_bio->devs[i].bio = bio; + atomic_inc(&rdev->nr_pending); + } if (rrdev) { r10_bio->devs[i].repl_bio = bio; atomic_inc(&rrdev->nr_pending); @@ -1207,7 +1444,8 @@ retry_write: bio->bi_phys_segments++; spin_unlock_irq(&conf->device_lock); } - sectors_handled = r10_bio->sector + max_sectors - bio->bi_sector; + sectors_handled = r10_bio->sector + max_sectors - + bio->bi_iter.bi_sector; atomic_set(&r10_bio->remaining, 1); bitmap_startwrite(mddev->bitmap, r10_bio->sector, r10_bio->sectors, 0); @@ -1215,58 +1453,80 @@ retry_write: for (i = 0; i < conf->copies; i++) { struct bio *mbio; int d = r10_bio->devs[i].devnum; - if (!r10_bio->devs[i].bio) - continue; - - mbio = bio_clone_mddev(bio, GFP_NOIO, mddev); - md_trim_bio(mbio, r10_bio->sector - bio->bi_sector, - max_sectors); - r10_bio->devs[i].bio = mbio; + if (r10_bio->devs[i].bio) { + struct md_rdev *rdev = conf->mirrors[d].rdev; + mbio = bio_clone_mddev(bio, GFP_NOIO, mddev); + bio_trim(mbio, r10_bio->sector - bio->bi_iter.bi_sector, + max_sectors); + r10_bio->devs[i].bio = mbio; + + mbio->bi_iter.bi_sector = (r10_bio->devs[i].addr+ + choose_data_offset(r10_bio, + rdev)); + mbio->bi_bdev = rdev->bdev; + mbio->bi_end_io = raid10_end_write_request; + mbio->bi_rw = + WRITE | do_sync | do_fua | do_discard | do_same; + mbio->bi_private = r10_bio; - mbio->bi_sector = (r10_bio->devs[i].addr+ - conf->mirrors[d].rdev->data_offset); - mbio->bi_bdev = conf->mirrors[d].rdev->bdev; - mbio->bi_end_io = raid10_end_write_request; - mbio->bi_rw = WRITE | do_sync | do_fua; - mbio->bi_private = r10_bio; - - atomic_inc(&r10_bio->remaining); - spin_lock_irqsave(&conf->device_lock, flags); - bio_list_add(&conf->pending_bio_list, mbio); - conf->pending_count++; - spin_unlock_irqrestore(&conf->device_lock, flags); - - if (!r10_bio->devs[i].repl_bio) - continue; + atomic_inc(&r10_bio->remaining); - mbio = bio_clone_mddev(bio, GFP_NOIO, mddev); - md_trim_bio(mbio, r10_bio->sector - bio->bi_sector, - max_sectors); - r10_bio->devs[i].repl_bio = mbio; + cb = blk_check_plugged(raid10_unplug, mddev, + sizeof(*plug)); + if (cb) + plug = container_of(cb, struct raid10_plug_cb, + cb); + else + plug = NULL; + spin_lock_irqsave(&conf->device_lock, flags); + if (plug) { + bio_list_add(&plug->pending, mbio); + plug->pending_cnt++; + } else { + bio_list_add(&conf->pending_bio_list, mbio); + conf->pending_count++; + } + spin_unlock_irqrestore(&conf->device_lock, flags); + if (!plug) + md_wakeup_thread(mddev->thread); + } - /* We are actively writing to the original device - * so it cannot disappear, so the replacement cannot - * become NULL here - */ - mbio->bi_sector = (r10_bio->devs[i].addr+ - conf->mirrors[d].replacement->data_offset); - mbio->bi_bdev = conf->mirrors[d].replacement->bdev; - mbio->bi_end_io = raid10_end_write_request; - mbio->bi_rw = WRITE | do_sync | do_fua; - mbio->bi_private = r10_bio; + if (r10_bio->devs[i].repl_bio) { + struct md_rdev *rdev = conf->mirrors[d].replacement; + if (rdev == NULL) { + /* Replacement just got moved to main 'rdev' */ + smp_mb(); + rdev = conf->mirrors[d].rdev; + } + mbio = bio_clone_mddev(bio, GFP_NOIO, mddev); + bio_trim(mbio, r10_bio->sector - bio->bi_iter.bi_sector, + max_sectors); + r10_bio->devs[i].repl_bio = mbio; + + mbio->bi_iter.bi_sector = (r10_bio->devs[i].addr + + choose_data_offset( + r10_bio, rdev)); + mbio->bi_bdev = rdev->bdev; + mbio->bi_end_io = raid10_end_write_request; + mbio->bi_rw = + WRITE | do_sync | do_fua | do_discard | do_same; + mbio->bi_private = r10_bio; - atomic_inc(&r10_bio->remaining); - spin_lock_irqsave(&conf->device_lock, flags); - bio_list_add(&conf->pending_bio_list, mbio); - conf->pending_count++; - spin_unlock_irqrestore(&conf->device_lock, flags); + atomic_inc(&r10_bio->remaining); + spin_lock_irqsave(&conf->device_lock, flags); + bio_list_add(&conf->pending_bio_list, mbio); + conf->pending_count++; + spin_unlock_irqrestore(&conf->device_lock, flags); + if (!mddev_check_plugged(mddev)) + md_wakeup_thread(mddev->thread); + } } /* Don't remove the bias on 'remaining' (one_write_done) until * after checking if we need to go around again. */ - if (sectors_handled < (bio->bi_size >> 9)) { + if (sectors_handled < bio_sectors(bio)) { one_write_done(r10_bio); /* We need another r10_bio. It has already been counted * in bio->bi_phys_segments. @@ -1274,20 +1534,57 @@ retry_write: r10_bio = mempool_alloc(conf->r10bio_pool, GFP_NOIO); r10_bio->master_bio = bio; - r10_bio->sectors = (bio->bi_size >> 9) - sectors_handled; + r10_bio->sectors = bio_sectors(bio) - sectors_handled; r10_bio->mddev = mddev; - r10_bio->sector = bio->bi_sector + sectors_handled; + r10_bio->sector = bio->bi_iter.bi_sector + sectors_handled; r10_bio->state = 0; goto retry_write; } one_write_done(r10_bio); +} + +static void make_request(struct mddev *mddev, struct bio *bio) +{ + struct r10conf *conf = mddev->private; + sector_t chunk_mask = (conf->geo.chunk_mask & conf->prev.chunk_mask); + int chunk_sects = chunk_mask + 1; + + struct bio *split; + + if (unlikely(bio->bi_rw & REQ_FLUSH)) { + md_flush_request(mddev, bio); + return; + } + + md_write_start(mddev, bio); + + + do { + + /* + * If this request crosses a chunk boundary, we need to split + * it. + */ + if (unlikely((bio->bi_iter.bi_sector & chunk_mask) + + bio_sectors(bio) > chunk_sects + && (conf->geo.near_copies < conf->geo.raid_disks + || conf->prev.near_copies < + conf->prev.raid_disks))) { + split = bio_split(bio, chunk_sects - + (bio->bi_iter.bi_sector & + (chunk_sects - 1)), + GFP_NOIO, fs_bio_set); + bio_chain(split, bio); + } else { + split = bio; + } + + __make_request(mddev, split); + } while (split != bio); /* In case raid10d snuck in to freeze_array */ wake_up(&conf->wait_barrier); - - if (do_sync || !mddev->bitmap || !plugged) - md_wakeup_thread(mddev->thread); } static void status(struct seq_file *seq, struct mddev *mddev) @@ -1295,19 +1592,19 @@ static void status(struct seq_file *seq, struct mddev *mddev) struct r10conf *conf = mddev->private; int i; - if (conf->near_copies < conf->raid_disks) + if (conf->geo.near_copies < conf->geo.raid_disks) seq_printf(seq, " %dK chunks", mddev->chunk_sectors / 2); - if (conf->near_copies > 1) - seq_printf(seq, " %d near-copies", conf->near_copies); - if (conf->far_copies > 1) { - if (conf->far_offset) - seq_printf(seq, " %d offset-copies", conf->far_copies); + if (conf->geo.near_copies > 1) + seq_printf(seq, " %d near-copies", conf->geo.near_copies); + if (conf->geo.far_copies > 1) { + if (conf->geo.far_offset) + seq_printf(seq, " %d offset-copies", conf->geo.far_copies); else - seq_printf(seq, " %d far-copies", conf->far_copies); + seq_printf(seq, " %d far-copies", conf->geo.far_copies); } - seq_printf(seq, " [%d/%d] [", conf->raid_disks, - conf->raid_disks - mddev->degraded); - for (i = 0; i < conf->raid_disks; i++) + seq_printf(seq, " [%d/%d] [", conf->geo.raid_disks, + conf->geo.raid_disks - mddev->degraded); + for (i = 0; i < conf->geo.raid_disks; i++) seq_printf(seq, "%s", conf->mirrors[i].rdev && test_bit(In_sync, &conf->mirrors[i].rdev->flags) ? "U" : "_"); @@ -1319,29 +1616,58 @@ static void status(struct seq_file *seq, struct mddev *mddev) * Don't consider the device numbered 'ignore' * as we might be about to remove it. */ -static int enough(struct r10conf *conf, int ignore) +static int _enough(struct r10conf *conf, int previous, int ignore) { int first = 0; + int has_enough = 0; + int disks, ncopies; + if (previous) { + disks = conf->prev.raid_disks; + ncopies = conf->prev.near_copies; + } else { + disks = conf->geo.raid_disks; + ncopies = conf->geo.near_copies; + } + rcu_read_lock(); do { int n = conf->copies; int cnt = 0; + int this = first; while (n--) { - if (conf->mirrors[first].rdev && - first != ignore) + struct md_rdev *rdev; + if (this != ignore && + (rdev = rcu_dereference(conf->mirrors[this].rdev)) && + test_bit(In_sync, &rdev->flags)) cnt++; - first = (first+1) % conf->raid_disks; + this = (this+1) % disks; } if (cnt == 0) - return 0; + goto out; + first = (first + ncopies) % disks; } while (first != 0); - return 1; + has_enough = 1; +out: + rcu_read_unlock(); + return has_enough; +} + +static int enough(struct r10conf *conf, int ignore) +{ + /* when calling 'enough', both 'prev' and 'geo' must + * be stable. + * This is ensured if ->reconfig_mutex or ->device_lock + * is held. + */ + return _enough(conf, 0, ignore) && + _enough(conf, 1, ignore); } static void error(struct mddev *mddev, struct md_rdev *rdev) { char b[BDEVNAME_SIZE]; struct r10conf *conf = mddev->private; + unsigned long flags; /* * If it is not operational, then we have already marked it as dead @@ -1349,18 +1675,18 @@ static void error(struct mddev *mddev, struct md_rdev *rdev) * next level up know. * else mark the drive as failed */ + spin_lock_irqsave(&conf->device_lock, flags); if (test_bit(In_sync, &rdev->flags) - && !enough(conf, rdev->raid_disk)) + && !enough(conf, rdev->raid_disk)) { /* * Don't fail the drive, just return an IO error. */ + spin_unlock_irqrestore(&conf->device_lock, flags); return; + } if (test_and_clear_bit(In_sync, &rdev->flags)) { - unsigned long flags; - spin_lock_irqsave(&conf->device_lock, flags); mddev->degraded++; - spin_unlock_irqrestore(&conf->device_lock, flags); - /* + /* * if recovery is running, make sure it aborts. */ set_bit(MD_RECOVERY_INTR, &mddev->recovery); @@ -1368,27 +1694,28 @@ static void error(struct mddev *mddev, struct md_rdev *rdev) set_bit(Blocked, &rdev->flags); set_bit(Faulty, &rdev->flags); set_bit(MD_CHANGE_DEVS, &mddev->flags); + spin_unlock_irqrestore(&conf->device_lock, flags); printk(KERN_ALERT "md/raid10:%s: Disk failure on %s, disabling device.\n" "md/raid10:%s: Operation continuing on %d devices.\n", mdname(mddev), bdevname(rdev->bdev, b), - mdname(mddev), conf->raid_disks - mddev->degraded); + mdname(mddev), conf->geo.raid_disks - mddev->degraded); } static void print_conf(struct r10conf *conf) { int i; - struct mirror_info *tmp; + struct raid10_info *tmp; printk(KERN_DEBUG "RAID10 conf printout:\n"); if (!conf) { printk(KERN_DEBUG "(!conf)\n"); return; } - printk(KERN_DEBUG " --- wd:%d rd:%d\n", conf->raid_disks - conf->mddev->degraded, - conf->raid_disks); + printk(KERN_DEBUG " --- wd:%d rd:%d\n", conf->geo.raid_disks - conf->mddev->degraded, + conf->geo.raid_disks); - for (i = 0; i < conf->raid_disks; i++) { + for (i = 0; i < conf->geo.raid_disks; i++) { char b[BDEVNAME_SIZE]; tmp = conf->mirrors + i; if (tmp->rdev) @@ -1412,7 +1739,7 @@ static int raid10_spare_active(struct mddev *mddev) { int i; struct r10conf *conf = mddev->private; - struct mirror_info *tmp; + struct raid10_info *tmp; int count = 0; unsigned long flags; @@ -1420,7 +1747,7 @@ static int raid10_spare_active(struct mddev *mddev) * Find all non-in_sync disks within the RAID10 configuration * and mark them in_sync */ - for (i = 0; i < conf->raid_disks; i++) { + for (i = 0; i < conf->geo.raid_disks; i++) { tmp = conf->mirrors + i; if (tmp->replacement && tmp->replacement->recovery_offset == MaxSector @@ -1441,10 +1768,11 @@ static int raid10_spare_active(struct mddev *mddev) } sysfs_notify_dirent_safe(tmp->replacement->sysfs_state); } else if (tmp->rdev + && tmp->rdev->recovery_offset == MaxSector && !test_bit(Faulty, &tmp->rdev->flags) && !test_and_set_bit(In_sync, &tmp->rdev->flags)) { count++; - sysfs_notify_dirent(tmp->rdev->sysfs_state); + sysfs_notify_dirent_safe(tmp->rdev->sysfs_state); } } spin_lock_irqsave(&conf->device_lock, flags); @@ -1462,26 +1790,32 @@ static int raid10_add_disk(struct mddev *mddev, struct md_rdev *rdev) int err = -EEXIST; int mirror; int first = 0; - int last = conf->raid_disks - 1; + int last = conf->geo.raid_disks - 1; + struct request_queue *q = bdev_get_queue(rdev->bdev); if (mddev->recovery_cp < MaxSector) /* only hot-add to in-sync arrays, as recovery is * very different from resync */ return -EBUSY; - if (!enough(conf, -1)) + if (rdev->saved_raid_disk < 0 && !_enough(conf, 1, -1)) return -EINVAL; if (rdev->raid_disk >= 0) first = last = rdev->raid_disk; + if (q->merge_bvec_fn) { + set_bit(Unmerged, &rdev->flags); + mddev->merge_check_needed = 1; + } + if (rdev->saved_raid_disk >= first && conf->mirrors[rdev->saved_raid_disk].rdev == NULL) mirror = rdev->saved_raid_disk; else mirror = first; for ( ; mirror <= last ; mirror++) { - struct mirror_info *p = &conf->mirrors[mirror]; + struct raid10_info *p = &conf->mirrors[mirror]; if (p->recovery_disabled == mddev->recovery_disabled) continue; if (p->rdev) { @@ -1492,31 +1826,17 @@ static int raid10_add_disk(struct mddev *mddev, struct md_rdev *rdev) set_bit(Replacement, &rdev->flags); rdev->raid_disk = mirror; err = 0; - disk_stack_limits(mddev->gendisk, rdev->bdev, - rdev->data_offset << 9); - if (rdev->bdev->bd_disk->queue->merge_bvec_fn) { - blk_queue_max_segments(mddev->queue, 1); - blk_queue_segment_boundary(mddev->queue, - PAGE_CACHE_SIZE - 1); - } + if (mddev->gendisk) + disk_stack_limits(mddev->gendisk, rdev->bdev, + rdev->data_offset << 9); conf->fullsync = 1; rcu_assign_pointer(p->replacement, rdev); break; } - disk_stack_limits(mddev->gendisk, rdev->bdev, - rdev->data_offset << 9); - /* as we don't honour merge_bvec_fn, we must - * never risk violating it, so limit - * ->max_segments to one lying with a single - * page, as a one page request is never in - * violation. - */ - if (rdev->bdev->bd_disk->queue->merge_bvec_fn) { - blk_queue_max_segments(mddev->queue, 1); - blk_queue_segment_boundary(mddev->queue, - PAGE_CACHE_SIZE - 1); - } + if (mddev->gendisk) + disk_stack_limits(mddev->gendisk, rdev->bdev, + rdev->data_offset << 9); p->head_position = 0; p->recovery_disabled = mddev->recovery_disabled - 1; @@ -1527,8 +1847,23 @@ static int raid10_add_disk(struct mddev *mddev, struct md_rdev *rdev) rcu_assign_pointer(p->rdev, rdev); break; } - + if (err == 0 && test_bit(Unmerged, &rdev->flags)) { + /* Some requests might not have seen this new + * merge_bvec_fn. We must wait for them to complete + * before merging the device fully. + * First we make sure any code which has tested + * our function has submitted the request, then + * we wait for all outstanding requests to complete. + */ + synchronize_sched(); + freeze_array(conf, 0); + unfreeze_array(conf); + clear_bit(Unmerged, &rdev->flags); + } md_integrity_add_rdev(rdev, mddev); + if (mddev->queue && blk_queue_discard(bdev_get_queue(rdev->bdev))) + queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, mddev->queue); + print_conf(conf); return err; } @@ -1539,7 +1874,7 @@ static int raid10_remove_disk(struct mddev *mddev, struct md_rdev *rdev) int err = 0; int number = rdev->raid_disk; struct md_rdev **rdevp; - struct mirror_info *p = conf->mirrors + number; + struct raid10_info *p = conf->mirrors + number; print_conf(conf); if (rdev == p->rdev) @@ -1560,6 +1895,7 @@ static int raid10_remove_disk(struct mddev *mddev, struct md_rdev *rdev) if (!test_bit(Faulty, &rdev->flags) && mddev->recovery_disabled != p->recovery_disabled && (!p->replacement || p->replacement == rdev) && + number < conf->geo.raid_disks && enough(conf, -1)) { err = -EBUSY; goto abort; @@ -1601,7 +1937,11 @@ static void end_sync_read(struct bio *bio, int error) struct r10conf *conf = r10_bio->mddev->private; int d; - d = find_bio_disk(conf, r10_bio, bio, NULL, NULL); + if (bio == r10_bio->master_bio) { + /* this is a reshape read */ + d = r10_bio->read_slot; /* really the read dev */ + } else + d = find_bio_disk(conf, r10_bio, bio, NULL, NULL); if (test_bit(BIO_UPTODATE, &bio->bi_flags)) set_bit(R10BIO_Uptodate, &r10_bio->state); @@ -1668,10 +2008,8 @@ static void end_sync_write(struct bio *bio, int error) d = find_bio_disk(conf, r10_bio, bio, &slot, &repl); if (repl) rdev = conf->mirrors[d].replacement; - if (!rdev) { - smp_mb(); + else rdev = conf->mirrors[d].rdev; - } if (!uptodate) { if (repl) @@ -1715,6 +2053,7 @@ static void sync_request_write(struct mddev *mddev, struct r10bio *r10_bio) struct r10conf *conf = mddev->private; int i, first; struct bio *tbio, *fbio; + int vcnt; atomic_set(&r10_bio->remaining, 1); @@ -1729,10 +2068,10 @@ static void sync_request_write(struct mddev *mddev, struct r10bio *r10_bio) first = i; fbio = r10_bio->devs[i].bio; + vcnt = (r10_bio->sectors + (PAGE_SIZE >> 9) - 1) >> (PAGE_SHIFT - 9); /* now find blocks with errors */ for (i=0 ; i < conf->copies ; i++) { int j, d; - int vcnt = r10_bio->sectors >> (PAGE_SHIFT-9); tbio = r10_bio->devs[i].bio; @@ -1745,14 +2084,20 @@ static void sync_request_write(struct mddev *mddev, struct r10bio *r10_bio) * both 'first' and 'i', so we just compare them. * All vec entries are PAGE_SIZE; */ - for (j = 0; j < vcnt; j++) + int sectors = r10_bio->sectors; + for (j = 0; j < vcnt; j++) { + int len = PAGE_SIZE; + if (sectors < (len / 512)) + len = sectors * 512; if (memcmp(page_address(fbio->bi_io_vec[j].bv_page), page_address(tbio->bi_io_vec[j].bv_page), - PAGE_SIZE)) + len)) break; + sectors -= len/512; + } if (j == vcnt) continue; - mddev->resync_mismatches += r10_bio->sectors; + atomic64_add(r10_bio->sectors, &mddev->resync_mismatches); if (test_bit(MD_RECOVERY_CHECK, &mddev->recovery)) /* Don't fix anything. */ continue; @@ -1762,16 +2107,13 @@ static void sync_request_write(struct mddev *mddev, struct r10bio *r10_bio) * First we need to fixup bv_offset, bv_len and * bi_vecs, as the read request might have corrupted these */ + bio_reset(tbio); + tbio->bi_vcnt = vcnt; - tbio->bi_size = r10_bio->sectors << 9; - tbio->bi_idx = 0; - tbio->bi_phys_segments = 0; - tbio->bi_flags &= ~(BIO_POOL_MASK - 1); - tbio->bi_flags |= 1 << BIO_UPTODATE; - tbio->bi_next = NULL; + tbio->bi_iter.bi_size = r10_bio->sectors << 9; tbio->bi_rw = WRITE; tbio->bi_private = r10_bio; - tbio->bi_sector = r10_bio->devs[i].addr; + tbio->bi_iter.bi_sector = r10_bio->devs[i].addr; for (j=0; j < vcnt ; j++) { tbio->bi_io_vec[j].bv_offset = 0; @@ -1786,9 +2128,9 @@ static void sync_request_write(struct mddev *mddev, struct r10bio *r10_bio) d = r10_bio->devs[i].devnum; atomic_inc(&conf->mirrors[d].rdev->nr_pending); atomic_inc(&r10_bio->remaining); - md_sync_acct(conf->mirrors[d].rdev->bdev, tbio->bi_size >> 9); + md_sync_acct(conf->mirrors[d].rdev->bdev, bio_sectors(tbio)); - tbio->bi_sector += conf->mirrors[d].rdev->data_offset; + tbio->bi_iter.bi_sector += conf->mirrors[d].rdev->data_offset; tbio->bi_bdev = conf->mirrors[d].rdev->bdev; generic_make_request(tbio); } @@ -1798,7 +2140,6 @@ static void sync_request_write(struct mddev *mddev, struct r10bio *r10_bio) */ for (i = 0; i < conf->copies; i++) { int j, d; - int vcnt = r10_bio->sectors >> (PAGE_SHIFT-9); tbio = r10_bio->devs[i].repl_bio; if (!tbio || !tbio->bi_end_io) @@ -1812,7 +2153,7 @@ static void sync_request_write(struct mddev *mddev, struct r10bio *r10_bio) d = r10_bio->devs[i].devnum; atomic_inc(&r10_bio->remaining); md_sync_acct(conf->mirrors[d].replacement->bdev, - tbio->bi_size >> 9); + bio_sectors(tbio)); generic_make_request(tbio); } @@ -1936,15 +2277,21 @@ static void recovery_request_write(struct mddev *mddev, struct r10bio *r10_bio) d = r10_bio->devs[1].devnum; wbio = r10_bio->devs[1].bio; wbio2 = r10_bio->devs[1].repl_bio; + /* Need to test wbio2->bi_end_io before we call + * generic_make_request as if the former is NULL, + * the latter is free to free wbio2. + */ + if (wbio2 && !wbio2->bi_end_io) + wbio2 = NULL; if (wbio->bi_end_io) { atomic_inc(&conf->mirrors[d].rdev->nr_pending); - md_sync_acct(conf->mirrors[d].rdev->bdev, wbio->bi_size >> 9); + md_sync_acct(conf->mirrors[d].rdev->bdev, bio_sectors(wbio)); generic_make_request(wbio); } - if (wbio2 && wbio2->bi_end_io) { + if (wbio2) { atomic_inc(&conf->mirrors[d].replacement->nr_pending); md_sync_acct(conf->mirrors[d].replacement->bdev, - wbio2->bi_size >> 9); + bio_sectors(wbio2)); generic_make_request(wbio2); } } @@ -2052,6 +2399,7 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10 "md/raid10:%s: %s: Failing raid device\n", mdname(mddev), b); md_error(mddev, conf->mirrors[d].rdev); + r10_bio->devs[r10_bio->read_slot].bio = IO_BLOCKED; return; } @@ -2072,6 +2420,7 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10 d = r10_bio->devs[sl].devnum; rdev = rcu_dereference(conf->mirrors[d].rdev); if (rdev && + !test_bit(Unmerged, &rdev->flags) && test_bit(In_sync, &rdev->flags) && is_badblock(rdev, r10_bio->devs[sl].addr + sect, s, &first_bad, &bad_sectors) == 0) { @@ -2105,8 +2454,11 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10 rdev, r10_bio->devs[r10_bio->read_slot].addr + sect, - s, 0)) + s, 0)) { md_error(mddev, rdev); + r10_bio->devs[r10_bio->read_slot].bio + = IO_BLOCKED; + } break; } @@ -2122,6 +2474,7 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10 d = r10_bio->devs[sl].devnum; rdev = rcu_dereference(conf->mirrors[d].rdev); if (!rdev || + test_bit(Unmerged, &rdev->flags) || !test_bit(In_sync, &rdev->flags)) continue; @@ -2130,7 +2483,7 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10 if (r10_sync_page_io(rdev, r10_bio->devs[sl].addr + sect, - s<<9, conf->tmppage, WRITE) + s, conf->tmppage, WRITE) == 0) { /* Well, this device is dead */ printk(KERN_NOTICE @@ -2139,7 +2492,9 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10 " (%d sectors at %llu on %s)\n", mdname(mddev), s, (unsigned long long)( - sect + rdev->data_offset), + sect + + choose_data_offset(r10_bio, + rdev)), bdevname(rdev->bdev, b)); printk(KERN_NOTICE "md/raid10:%s: %s: failing " "drive\n", @@ -2167,7 +2522,7 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10 switch (r10_sync_page_io(rdev, r10_bio->devs[sl].addr + sect, - s<<9, conf->tmppage, + s, conf->tmppage, READ)) { case 0: /* Well, this device is dead */ @@ -2177,7 +2532,8 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10 " (%d sectors at %llu on %s)\n", mdname(mddev), s, (unsigned long long)( - sect + rdev->data_offset), + sect + + choose_data_offset(r10_bio, rdev)), bdevname(rdev->bdev, b)); printk(KERN_NOTICE "md/raid10:%s: %s: failing " "drive\n", @@ -2190,7 +2546,8 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10 " (%d sectors at %llu on %s)\n", mdname(mddev), s, (unsigned long long)( - sect + rdev->data_offset), + sect + + choose_data_offset(r10_bio, rdev)), bdevname(rdev->bdev, b)); atomic_add(s, &rdev->corrected_errors); } @@ -2205,25 +2562,6 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10 } } -static void bi_complete(struct bio *bio, int error) -{ - complete((struct completion *)bio->bi_private); -} - -static int submit_bio_wait(int rw, struct bio *bio) -{ - struct completion event; - rw |= REQ_SYNC; - - init_completion(&event); - bio->bi_private = &event; - bio->bi_end_io = bi_complete; - submit_bio(rw, bio); - wait_for_completion(&event); - - return test_bit(BIO_UPTODATE, &bio->bi_flags); -} - static int narrow_write_error(struct r10bio *r10_bio, int i) { struct bio *bio = r10_bio->master_bio; @@ -2262,9 +2600,9 @@ static int narrow_write_error(struct r10bio *r10_bio, int i) sectors = sect_to_write; /* Write at 'sector' for 'sectors' */ wbio = bio_clone_mddev(bio, GFP_NOIO, mddev); - md_trim_bio(wbio, sector - bio->bi_sector, sectors); - wbio->bi_sector = (r10_bio->devs[i].addr+ - rdev->data_offset+ + bio_trim(wbio, sector - bio->bi_iter.bi_sector, sectors); + wbio->bi_iter.bi_sector = (r10_bio->devs[i].addr+ + choose_data_offset(r10_bio, rdev) + (sector - r10_bio->sector)); wbio->bi_bdev = rdev->bdev; if (submit_bio_wait(WRITE, wbio) == 0) @@ -2299,17 +2637,20 @@ static void handle_read_error(struct mddev *mddev, struct r10bio *r10_bio) * This is all done synchronously while the array is * frozen. */ + bio = r10_bio->devs[slot].bio; + bdevname(bio->bi_bdev, b); + bio_put(bio); + r10_bio->devs[slot].bio = NULL; + if (mddev->ro == 0) { - freeze_array(conf); + freeze_array(conf, 1); fix_read_error(conf, mddev, r10_bio); unfreeze_array(conf); - } + } else + r10_bio->devs[slot].bio = IO_BLOCKED; + rdev_dec_pending(rdev, mddev); - bio = r10_bio->devs[slot].bio; - bdevname(bio->bi_bdev, b); - r10_bio->devs[slot].bio = - mddev->ro ? IO_BLOCKED : NULL; read_more: rdev = read_balance(conf, r10_bio, &max_sectors); if (rdev == NULL) { @@ -2318,30 +2659,25 @@ read_more: mdname(mddev), b, (unsigned long long)r10_bio->sector); raid_end_bio_io(r10_bio); - bio_put(bio); return; } do_sync = (r10_bio->master_bio->bi_rw & REQ_SYNC); - if (bio) - bio_put(bio); slot = r10_bio->read_slot; printk_ratelimited( KERN_ERR - "md/raid10:%s: %s: redirecting" + "md/raid10:%s: %s: redirecting " "sector %llu to another mirror\n", mdname(mddev), bdevname(rdev->bdev, b), (unsigned long long)r10_bio->sector); bio = bio_clone_mddev(r10_bio->master_bio, GFP_NOIO, mddev); - md_trim_bio(bio, - r10_bio->sector - bio->bi_sector, - max_sectors); + bio_trim(bio, r10_bio->sector - bio->bi_iter.bi_sector, max_sectors); r10_bio->devs[slot].bio = bio; r10_bio->devs[slot].rdev = rdev; - bio->bi_sector = r10_bio->devs[slot].addr - + rdev->data_offset; + bio->bi_iter.bi_sector = r10_bio->devs[slot].addr + + choose_data_offset(r10_bio, rdev); bio->bi_bdev = rdev->bdev; bio->bi_rw = READ | do_sync; bio->bi_private = r10_bio; @@ -2351,7 +2687,7 @@ read_more: struct bio *mbio = r10_bio->master_bio; int sectors_handled = r10_bio->sector + max_sectors - - mbio->bi_sector; + - mbio->bi_iter.bi_sector; r10_bio->sectors = max_sectors; spin_lock_irq(&conf->device_lock); if (mbio->bi_phys_segments == 0) @@ -2360,18 +2696,16 @@ read_more: mbio->bi_phys_segments++; spin_unlock_irq(&conf->device_lock); generic_make_request(bio); - bio = NULL; r10_bio = mempool_alloc(conf->r10bio_pool, GFP_NOIO); r10_bio->master_bio = mbio; - r10_bio->sectors = (mbio->bi_size >> 9) - - sectors_handled; + r10_bio->sectors = bio_sectors(mbio) - sectors_handled; r10_bio->state = 0; set_bit(R10BIO_ReadError, &r10_bio->state); r10_bio->mddev = mddev; - r10_bio->sector = mbio->bi_sector + r10_bio->sector = mbio->bi_iter.bi_sector + sectors_handled; goto read_more; @@ -2402,7 +2736,7 @@ static void handle_write_completed(struct r10conf *conf, struct r10bio *r10_bio) rdev_clear_badblocks( rdev, r10_bio->devs[m].addr, - r10_bio->sectors); + r10_bio->sectors, 0); } else { if (!rdev_set_badblocks( rdev, @@ -2418,7 +2752,7 @@ static void handle_write_completed(struct r10conf *conf, struct r10bio *r10_bio) rdev_clear_badblocks( rdev, r10_bio->devs[m].addr, - r10_bio->sectors); + r10_bio->sectors, 0); } else { if (!rdev_set_badblocks( rdev, @@ -2437,7 +2771,7 @@ static void handle_write_completed(struct r10conf *conf, struct r10bio *r10_bio) rdev_clear_badblocks( rdev, r10_bio->devs[m].addr, - r10_bio->sectors); + r10_bio->sectors, 0); rdev_dec_pending(rdev, conf->mddev); } else if (bio != NULL && !test_bit(BIO_UPTODATE, &bio->bi_flags)) { @@ -2454,7 +2788,7 @@ static void handle_write_completed(struct r10conf *conf, struct r10bio *r10_bio) rdev_clear_badblocks( rdev, r10_bio->devs[m].addr, - r10_bio->sectors); + r10_bio->sectors, 0); rdev_dec_pending(rdev, conf->mddev); } } @@ -2465,8 +2799,9 @@ static void handle_write_completed(struct r10conf *conf, struct r10bio *r10_bio) } } -static void raid10d(struct mddev *mddev) +static void raid10d(struct md_thread *thread) { + struct mddev *mddev = thread->mddev; struct r10bio *r10_bio; unsigned long flags; struct r10conf *conf = mddev->private; @@ -2495,6 +2830,8 @@ static void raid10d(struct mddev *mddev) if (test_bit(R10BIO_MadeGood, &r10_bio->state) || test_bit(R10BIO_WriteError, &r10_bio->state)) handle_write_completed(conf, r10_bio); + else if (test_bit(R10BIO_IsReshape, &r10_bio->state)) + reshape_request_write(mddev, r10_bio); else if (test_bit(R10BIO_IsSync, &r10_bio->state)) sync_request_write(mddev, r10_bio); else if (test_bit(R10BIO_IsRecover, &r10_bio->state)) @@ -2525,7 +2862,7 @@ static int init_resync(struct r10conf *conf) buffs = RESYNC_WINDOW / RESYNC_BLOCK_SIZE; BUG_ON(conf->r10buf_pool); conf->have_replacement = 0; - for (i = 0; i < conf->raid_disks; i++) + for (i = 0; i < conf->geo.raid_disks; i++) if (conf->mirrors[i].replacement) conf->have_replacement = 1; conf->r10buf_pool = mempool_create(buffs, r10buf_pool_alloc, r10buf_pool_free, conf); @@ -2579,14 +2916,31 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, sector_t sync_blocks; sector_t sectors_skipped = 0; int chunks_skipped = 0; + sector_t chunk_mask = conf->geo.chunk_mask; if (!conf->r10buf_pool) if (init_resync(conf)) return 0; + /* + * Allow skipping a full rebuild for incremental assembly + * of a clean array, like RAID1 does. + */ + if (mddev->bitmap == NULL && + mddev->recovery_cp == MaxSector && + mddev->reshape_position == MaxSector && + !test_bit(MD_RECOVERY_SYNC, &mddev->recovery) && + !test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery) && + !test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery) && + conf->fullsync == 0) { + *skipped = 1; + return mddev->dev_sectors - sector_nr; + } + skipped: max_sector = mddev->dev_sectors; - if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery)) + if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery) || + test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery)) max_sector = mddev->resync_max_sectors; if (sector_nr >= max_sector) { /* If we aborted, we need to abort the @@ -2598,11 +2952,16 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, * we need to convert that to several * virtual addresses. */ + if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery)) { + end_reshape(conf); + return 0; + } + if (mddev->curr_resync < max_sector) { /* aborted */ if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery)) bitmap_end_sync(mddev->bitmap, mddev->curr_resync, &sync_blocks, 1); - else for (i=0; i<conf->raid_disks; i++) { + else for (i = 0; i < conf->geo.raid_disks; i++) { sector_t sect = raid10_find_virt(conf, mddev->curr_resync, i); bitmap_end_sync(mddev->bitmap, sect, @@ -2616,7 +2975,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, /* Completed a full sync so the replacements * are now fully recovered. */ - for (i = 0; i < conf->raid_disks; i++) + for (i = 0; i < conf->geo.raid_disks; i++) if (conf->mirrors[i].replacement) conf->mirrors[i].replacement ->recovery_offset @@ -2629,7 +2988,11 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, *skipped = 1; return sectors_skipped; } - if (chunks_skipped >= conf->raid_disks) { + + if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery)) + return reshape_request(mddev, sector_nr, skipped); + + if (chunks_skipped >= conf->geo.raid_disks) { /* if there has been nothing to do on any drive, * then there is nothing to do at all.. */ @@ -2643,9 +3006,9 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, /* make sure whole request will fit in a chunk - if chunks * are meaningful */ - if (conf->near_copies < conf->raid_disks && - max_sector > (sector_nr | conf->chunk_mask)) - max_sector = (sector_nr | conf->chunk_mask) + 1; + if (conf->geo.near_copies < conf->geo.raid_disks && + max_sector > (sector_nr | chunk_mask)) + max_sector = (sector_nr | chunk_mask) + 1; /* * If there is non-resync activity waiting for us then * put in a delay to throttle resync. @@ -2674,13 +3037,13 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, int j; r10_bio = NULL; - for (i=0 ; i<conf->raid_disks; i++) { + for (i = 0 ; i < conf->geo.raid_disks; i++) { int still_degraded; struct r10bio *rb2; sector_t sect; int must_sync; int any_working; - struct mirror_info *mirror = &conf->mirrors[i]; + struct raid10_info *mirror = &conf->mirrors[i]; if ((mirror->rdev == NULL || test_bit(In_sync, &mirror->rdev->flags)) @@ -2694,6 +3057,12 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, /* want to reconstruct this device */ rb2 = r10_bio; sect = raid10_find_virt(conf, sector_nr, i); + if (sect >= mddev->resync_max_sectors) { + /* last stripe is not complete - don't + * try to recover this sector. + */ + continue; + } /* Unless we are doing a full sync, or a replacement * we only need to recover the block if it is set in * the bitmap @@ -2728,7 +3097,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, /* Need to check if the array will still be * degraded */ - for (j=0; j<conf->raid_disks; j++) + for (j = 0; j < conf->geo.raid_disks; j++) if (conf->mirrors[j].rdev == NULL || test_bit(Faulty, &conf->mirrors[j].rdev->flags)) { still_degraded = 1; @@ -2767,13 +3136,15 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, } } bio = r10_bio->devs[0].bio; + bio_reset(bio); bio->bi_next = biolist; biolist = bio; bio->bi_private = r10_bio; bio->bi_end_io = end_sync_read; bio->bi_rw = READ; from_addr = r10_bio->devs[j].addr; - bio->bi_sector = from_addr + rdev->data_offset; + bio->bi_iter.bi_sector = from_addr + + rdev->data_offset; bio->bi_bdev = rdev->bdev; atomic_inc(&rdev->nr_pending); /* and we write to 'i' (if not in_sync) */ @@ -2791,12 +3162,13 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, rdev = mirror->rdev; if (!test_bit(In_sync, &rdev->flags)) { bio = r10_bio->devs[1].bio; + bio_reset(bio); bio->bi_next = biolist; biolist = bio; bio->bi_private = r10_bio; bio->bi_end_io = end_sync_write; bio->bi_rw = WRITE; - bio->bi_sector = to_addr + bio->bi_iter.bi_sector = to_addr + rdev->data_offset; bio->bi_bdev = rdev->bdev; atomic_inc(&r10_bio->remaining); @@ -2819,12 +3191,14 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, if (rdev == NULL || bio == NULL || test_bit(Faulty, &rdev->flags)) break; + bio_reset(bio); bio->bi_next = biolist; biolist = bio; bio->bi_private = r10_bio; bio->bi_end_io = end_sync_write; bio->bi_rw = WRITE; - bio->bi_sector = to_addr + rdev->data_offset; + bio->bi_iter.bi_sector = to_addr + + rdev->data_offset; bio->bi_bdev = rdev->bdev; atomic_inc(&r10_bio->remaining); break; @@ -2832,10 +3206,6 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, if (j == conf->copies) { /* Cannot recover, so abort the recovery or * record a bad block */ - put_buf(r10_bio); - if (rb2) - atomic_dec(&rb2->remaining); - r10_bio = rb2; if (any_working) { /* problem is that there are bad blocks * on other device(s) @@ -2867,6 +3237,10 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, mirror->recovery_disabled = mddev->recovery_disabled; } + put_buf(r10_bio); + if (rb2) + atomic_dec(&rb2->remaining); + r10_bio = rb2; break; } } @@ -2906,9 +3280,9 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, r10_bio->sector = sector_nr; set_bit(R10BIO_IsSync, &r10_bio->state); raid10_find_phys(conf, r10_bio); - r10_bio->sectors = (sector_nr | conf->chunk_mask) - sector_nr +1; + r10_bio->sectors = (sector_nr | chunk_mask) - sector_nr + 1; - for (i=0; i<conf->copies; i++) { + for (i = 0; i < conf->copies; i++) { int d = r10_bio->devs[i].devnum; sector_t first_bad, sector; int bad_sectors; @@ -2917,7 +3291,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, r10_bio->devs[i].repl_bio->bi_end_io = NULL; bio = r10_bio->devs[i].bio; - bio->bi_end_io = NULL; + bio_reset(bio); clear_bit(BIO_UPTODATE, &bio->bi_flags); if (conf->mirrors[d].rdev == NULL || test_bit(Faulty, &conf->mirrors[d].rdev->flags)) @@ -2931,7 +3305,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, else { bad_sectors -= (sector - first_bad); if (max_sync > bad_sectors) - max_sync = max_sync; + max_sync = bad_sectors; continue; } } @@ -2942,7 +3316,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, bio->bi_private = r10_bio; bio->bi_end_io = end_sync_read; bio->bi_rw = READ; - bio->bi_sector = sector + + bio->bi_iter.bi_sector = sector + conf->mirrors[d].rdev->data_offset; bio->bi_bdev = conf->mirrors[d].rdev->bdev; count++; @@ -2954,6 +3328,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, /* Need to set up for writing to the replacement */ bio = r10_bio->devs[i].repl_bio; + bio_reset(bio); clear_bit(BIO_UPTODATE, &bio->bi_flags); sector = r10_bio->devs[i].addr; @@ -2963,7 +3338,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, bio->bi_private = r10_bio; bio->bi_end_io = end_sync_write; bio->bi_rw = WRITE; - bio->bi_sector = sector + + bio->bi_iter.bi_sector = sector + conf->mirrors[d].replacement->data_offset; bio->bi_bdev = conf->mirrors[d].replacement->bdev; count++; @@ -2987,17 +3362,6 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, } } - for (bio = biolist; bio ; bio=bio->bi_next) { - - bio->bi_flags &= ~(BIO_POOL_MASK - 1); - if (bio->bi_end_io) - bio->bi_flags |= 1 << BIO_UPTODATE; - bio->bi_vcnt = 0; - bio->bi_idx = 0; - bio->bi_phys_segments = 0; - bio->bi_size = 0; - } - nr_sectors = 0; if (sector_nr + max_sync < max_sector) max_sector = sector_nr + max_sync; @@ -3021,7 +3385,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, bio2 = bio2->bi_next) { /* remove last page from this bio */ bio2->bi_vcnt--; - bio2->bi_size -= len; + bio2->bi_iter.bi_size -= len; bio2->bi_flags &= ~(1<< BIO_SEG_VALID); } goto bio_full; @@ -3042,6 +3406,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, if (bio->bi_end_io == end_sync_read) { md_sync_acct(bio->bi_bdev, nr_sectors); + set_bit(BIO_UPTODATE, &bio->bi_flags); generic_make_request(bio); } } @@ -3074,40 +3439,108 @@ raid10_size(struct mddev *mddev, sector_t sectors, int raid_disks) struct r10conf *conf = mddev->private; if (!raid_disks) - raid_disks = conf->raid_disks; + raid_disks = min(conf->geo.raid_disks, + conf->prev.raid_disks); if (!sectors) sectors = conf->dev_sectors; - size = sectors >> conf->chunk_shift; - sector_div(size, conf->far_copies); + size = sectors >> conf->geo.chunk_shift; + sector_div(size, conf->geo.far_copies); size = size * raid_disks; - sector_div(size, conf->near_copies); + sector_div(size, conf->geo.near_copies); + + return size << conf->geo.chunk_shift; +} + +static void calc_sectors(struct r10conf *conf, sector_t size) +{ + /* Calculate the number of sectors-per-device that will + * actually be used, and set conf->dev_sectors and + * conf->stride + */ + + size = size >> conf->geo.chunk_shift; + sector_div(size, conf->geo.far_copies); + size = size * conf->geo.raid_disks; + sector_div(size, conf->geo.near_copies); + /* 'size' is now the number of chunks in the array */ + /* calculate "used chunks per device" */ + size = size * conf->copies; + + /* We need to round up when dividing by raid_disks to + * get the stride size. + */ + size = DIV_ROUND_UP_SECTOR_T(size, conf->geo.raid_disks); + + conf->dev_sectors = size << conf->geo.chunk_shift; - return size << conf->chunk_shift; + if (conf->geo.far_offset) + conf->geo.stride = 1 << conf->geo.chunk_shift; + else { + sector_div(size, conf->geo.far_copies); + conf->geo.stride = size << conf->geo.chunk_shift; + } } +enum geo_type {geo_new, geo_old, geo_start}; +static int setup_geo(struct geom *geo, struct mddev *mddev, enum geo_type new) +{ + int nc, fc, fo; + int layout, chunk, disks; + switch (new) { + case geo_old: + layout = mddev->layout; + chunk = mddev->chunk_sectors; + disks = mddev->raid_disks - mddev->delta_disks; + break; + case geo_new: + layout = mddev->new_layout; + chunk = mddev->new_chunk_sectors; + disks = mddev->raid_disks; + break; + default: /* avoid 'may be unused' warnings */ + case geo_start: /* new when starting reshape - raid_disks not + * updated yet. */ + layout = mddev->new_layout; + chunk = mddev->new_chunk_sectors; + disks = mddev->raid_disks + mddev->delta_disks; + break; + } + if (layout >> 18) + return -1; + if (chunk < (PAGE_SIZE >> 9) || + !is_power_of_2(chunk)) + return -2; + nc = layout & 255; + fc = (layout >> 8) & 255; + fo = layout & (1<<16); + geo->raid_disks = disks; + geo->near_copies = nc; + geo->far_copies = fc; + geo->far_offset = fo; + geo->far_set_size = (layout & (1<<17)) ? disks / fc : disks; + geo->chunk_mask = chunk - 1; + geo->chunk_shift = ffz(~chunk); + return nc*fc; +} static struct r10conf *setup_conf(struct mddev *mddev) { struct r10conf *conf = NULL; - int nc, fc, fo; - sector_t stride, size; int err = -EINVAL; + struct geom geo; + int copies; - if (mddev->new_chunk_sectors < (PAGE_SIZE >> 9) || - !is_power_of_2(mddev->new_chunk_sectors)) { + copies = setup_geo(&geo, mddev, geo_new); + + if (copies == -2) { printk(KERN_ERR "md/raid10:%s: chunk size must be " "at least PAGE_SIZE(%ld) and be a power of 2.\n", mdname(mddev), PAGE_SIZE); goto out; } - nc = mddev->new_layout & 255; - fc = (mddev->new_layout >> 8) & 255; - fo = mddev->new_layout & (1<<16); - - if ((nc*fc) <2 || (nc*fc) > mddev->raid_disks || - (mddev->new_layout >> 17)) { + if (copies < 2 || copies > mddev->raid_disks) { printk(KERN_ERR "md/raid10:%s: unsupported raid10 layout: 0x%8x\n", mdname(mddev), mddev->new_layout); goto out; @@ -3118,7 +3551,9 @@ static struct r10conf *setup_conf(struct mddev *mddev) if (!conf) goto out; - conf->mirrors = kzalloc(sizeof(struct mirror_info)*mddev->raid_disks, + /* FIXME calc properly */ + conf->mirrors = kzalloc(sizeof(struct raid10_info)*(mddev->raid_disks + + max(0,-mddev->delta_disks)), GFP_KERNEL); if (!conf->mirrors) goto out; @@ -3127,50 +3562,36 @@ static struct r10conf *setup_conf(struct mddev *mddev) if (!conf->tmppage) goto out; - - conf->raid_disks = mddev->raid_disks; - conf->near_copies = nc; - conf->far_copies = fc; - conf->copies = nc*fc; - conf->far_offset = fo; - conf->chunk_mask = mddev->new_chunk_sectors - 1; - conf->chunk_shift = ffz(~mddev->new_chunk_sectors); - + conf->geo = geo; + conf->copies = copies; conf->r10bio_pool = mempool_create(NR_RAID10_BIOS, r10bio_pool_alloc, r10bio_pool_free, conf); if (!conf->r10bio_pool) goto out; - size = mddev->dev_sectors >> conf->chunk_shift; - sector_div(size, fc); - size = size * conf->raid_disks; - sector_div(size, nc); - /* 'size' is now the number of chunks in the array */ - /* calculate "used chunks per device" in 'stride' */ - stride = size * conf->copies; - - /* We need to round up when dividing by raid_disks to - * get the stride size. - */ - stride += conf->raid_disks - 1; - sector_div(stride, conf->raid_disks); - - conf->dev_sectors = stride << conf->chunk_shift; - - if (fo) - stride = 1; - else - sector_div(stride, fc); - conf->stride = stride << conf->chunk_shift; - - + calc_sectors(conf, mddev->dev_sectors); + if (mddev->reshape_position == MaxSector) { + conf->prev = conf->geo; + conf->reshape_progress = MaxSector; + } else { + if (setup_geo(&conf->prev, mddev, geo_old) != conf->copies) { + err = -EINVAL; + goto out; + } + conf->reshape_progress = mddev->reshape_position; + if (conf->prev.far_offset) + conf->prev.stride = 1 << conf->prev.chunk_shift; + else + /* far_copies must be 1 */ + conf->prev.stride = conf->dev_sectors; + } spin_lock_init(&conf->device_lock); INIT_LIST_HEAD(&conf->retry_list); spin_lock_init(&conf->resync_lock); init_waitqueue_head(&conf->wait_barrier); - conf->thread = md_register_thread(raid10d, mddev, NULL); + conf->thread = md_register_thread(raid10d, mddev, "raid10"); if (!conf->thread) goto out; @@ -3178,8 +3599,9 @@ static struct r10conf *setup_conf(struct mddev *mddev) return conf; out: - printk(KERN_ERR "md/raid10:%s: couldn't allocate memory.\n", - mdname(mddev)); + if (err == -ENOMEM) + printk(KERN_ERR "md/raid10:%s: couldn't allocate memory.\n", + mdname(mddev)); if (conf) { if (conf->r10bio_pool) mempool_destroy(conf->r10bio_pool); @@ -3194,15 +3616,12 @@ static int run(struct mddev *mddev) { struct r10conf *conf; int i, disk_idx, chunk_size; - struct mirror_info *disk; + struct raid10_info *disk; struct md_rdev *rdev; sector_t size; - - /* - * copy the already verified devices into our private RAID10 - * bookkeeping area. [whatever we allocate in run(), - * should be freed in stop()] - */ + sector_t min_offset_diff = 0; + int first = 1; + bool discard_supported = false; if (mddev->private == NULL) { conf = setup_conf(mddev); @@ -3218,18 +3637,27 @@ static int run(struct mddev *mddev) conf->thread = NULL; chunk_size = mddev->chunk_sectors << 9; - blk_queue_io_min(mddev->queue, chunk_size); - if (conf->raid_disks % conf->near_copies) - blk_queue_io_opt(mddev->queue, chunk_size * conf->raid_disks); - else - blk_queue_io_opt(mddev->queue, chunk_size * - (conf->raid_disks / conf->near_copies)); + if (mddev->queue) { + blk_queue_max_discard_sectors(mddev->queue, + mddev->chunk_sectors); + blk_queue_max_write_same_sectors(mddev->queue, 0); + blk_queue_io_min(mddev->queue, chunk_size); + if (conf->geo.raid_disks % conf->geo.near_copies) + blk_queue_io_opt(mddev->queue, chunk_size * conf->geo.raid_disks); + else + blk_queue_io_opt(mddev->queue, chunk_size * + (conf->geo.raid_disks / conf->geo.near_copies)); + } - list_for_each_entry(rdev, &mddev->disks, same_set) { + rdev_for_each(rdev, mddev) { + long long diff; + struct request_queue *q; disk_idx = rdev->raid_disk; - if (disk_idx >= conf->raid_disks - || disk_idx < 0) + if (disk_idx < 0) + continue; + if (disk_idx >= conf->geo.raid_disks && + disk_idx >= conf->prev.raid_disks) continue; disk = conf->mirrors + disk_idx; @@ -3242,21 +3670,34 @@ static int run(struct mddev *mddev) goto out_free_conf; disk->rdev = rdev; } - - disk->rdev = rdev; - disk_stack_limits(mddev->gendisk, rdev->bdev, - rdev->data_offset << 9); - /* as we don't honour merge_bvec_fn, we must never risk - * violating it, so limit max_segments to 1 lying - * within a single page. - */ - if (rdev->bdev->bd_disk->queue->merge_bvec_fn) { - blk_queue_max_segments(mddev->queue, 1); - blk_queue_segment_boundary(mddev->queue, - PAGE_CACHE_SIZE - 1); - } + q = bdev_get_queue(rdev->bdev); + if (q->merge_bvec_fn) + mddev->merge_check_needed = 1; + diff = (rdev->new_data_offset - rdev->data_offset); + if (!mddev->reshape_backwards) + diff = -diff; + if (diff < 0) + diff = 0; + if (first || diff < min_offset_diff) + min_offset_diff = diff; + + if (mddev->gendisk) + disk_stack_limits(mddev->gendisk, rdev->bdev, + rdev->data_offset << 9); disk->head_position = 0; + + if (blk_queue_discard(bdev_get_queue(rdev->bdev))) + discard_supported = true; + } + + if (mddev->queue) { + if (discard_supported) + queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, + mddev->queue); + else + queue_flag_clear_unlocked(QUEUE_FLAG_DISCARD, + mddev->queue); } /* need to check that every block has at least one working mirror */ if (!enough(conf, -1)) { @@ -3265,8 +3706,21 @@ static int run(struct mddev *mddev) goto out_free_conf; } + if (conf->reshape_progress != MaxSector) { + /* must ensure that shape change is supported */ + if (conf->geo.far_copies != 1 && + conf->geo.far_offset == 0) + goto out_free_conf; + if (conf->prev.far_copies != 1 && + conf->prev.far_offset == 0) + goto out_free_conf; + } + mddev->degraded = 0; - for (i = 0; i < conf->raid_disks; i++) { + for (i = 0; + i < conf->geo.raid_disks + || i < conf->prev.raid_disks; + i++) { disk = conf->mirrors + i; @@ -3281,7 +3735,8 @@ static int run(struct mddev *mddev) !test_bit(In_sync, &disk->rdev->flags)) { disk->head_position = 0; mddev->degraded++; - if (disk->rdev) + if (disk->rdev && + disk->rdev->saved_raid_disk < 0) conf->fullsync = 1; } disk->recovery_disabled = mddev->recovery_disabled - 1; @@ -3293,8 +3748,8 @@ static int run(struct mddev *mddev) mdname(mddev)); printk(KERN_INFO "md/raid10:%s: active with %d out of %d devices\n", - mdname(mddev), conf->raid_disks - mddev->degraded, - conf->raid_disks); + mdname(mddev), conf->geo.raid_disks - mddev->degraded, + conf->geo.raid_disks); /* * Ok, everything is just fine now */ @@ -3303,27 +3758,50 @@ static int run(struct mddev *mddev) md_set_array_sectors(mddev, size); mddev->resync_max_sectors = size; - mddev->queue->backing_dev_info.congested_fn = raid10_congested; - mddev->queue->backing_dev_info.congested_data = mddev; - - /* Calculate max read-ahead size. - * We need to readahead at least twice a whole stripe.... - * maybe... - */ - { - int stripe = conf->raid_disks * + if (mddev->queue) { + int stripe = conf->geo.raid_disks * ((mddev->chunk_sectors << 9) / PAGE_SIZE); - stripe /= conf->near_copies; - if (mddev->queue->backing_dev_info.ra_pages < 2* stripe) - mddev->queue->backing_dev_info.ra_pages = 2* stripe; - } + mddev->queue->backing_dev_info.congested_fn = raid10_congested; + mddev->queue->backing_dev_info.congested_data = mddev; - if (conf->near_copies < conf->raid_disks) + /* Calculate max read-ahead size. + * We need to readahead at least twice a whole stripe.... + * maybe... + */ + stripe /= conf->geo.near_copies; + if (mddev->queue->backing_dev_info.ra_pages < 2 * stripe) + mddev->queue->backing_dev_info.ra_pages = 2 * stripe; blk_queue_merge_bvec(mddev->queue, raid10_mergeable_bvec); + } + if (md_integrity_register(mddev)) goto out_free_conf; + if (conf->reshape_progress != MaxSector) { + unsigned long before_length, after_length; + + before_length = ((1 << conf->prev.chunk_shift) * + conf->prev.far_copies); + after_length = ((1 << conf->geo.chunk_shift) * + conf->geo.far_copies); + + if (max(before_length, after_length) > min_offset_diff) { + /* This cannot work */ + printk("md/raid10: offset difference not enough to continue reshape\n"); + goto out_free_conf; + } + conf->offset_diff = min_offset_diff; + + conf->reshape_safe = conf->reshape_progress; + clear_bit(MD_RECOVERY_SYNC, &mddev->recovery); + clear_bit(MD_RECOVERY_CHECK, &mddev->recovery); + set_bit(MD_RECOVERY_RESHAPE, &mddev->recovery); + set_bit(MD_RECOVERY_RUNNING, &mddev->recovery); + mddev->sync_thread = md_register_thread(md_do_sync, mddev, + "reshape"); + } + return 0; out_free_conf: @@ -3346,9 +3824,13 @@ static int stop(struct mddev *mddev) lower_barrier(conf); md_unregister_thread(&mddev->thread); - blk_sync_queue(mddev->queue); /* the unplug fn references 'conf'*/ + if (mddev->queue) + /* the unplug fn references 'conf'*/ + blk_sync_queue(mddev->queue); + if (conf->r10bio_pool) mempool_destroy(conf->r10bio_pool); + safe_put_page(conf->tmppage); kfree(conf->mirrors); kfree(conf); mddev->private = NULL; @@ -3369,6 +3851,53 @@ static void raid10_quiesce(struct mddev *mddev, int state) } } +static int raid10_resize(struct mddev *mddev, sector_t sectors) +{ + /* Resize of 'far' arrays is not supported. + * For 'near' and 'offset' arrays we can set the + * number of sectors used to be an appropriate multiple + * of the chunk size. + * For 'offset', this is far_copies*chunksize. + * For 'near' the multiplier is the LCM of + * near_copies and raid_disks. + * So if far_copies > 1 && !far_offset, fail. + * Else find LCM(raid_disks, near_copy)*far_copies and + * multiply by chunk_size. Then round to this number. + * This is mostly done by raid10_size() + */ + struct r10conf *conf = mddev->private; + sector_t oldsize, size; + + if (mddev->reshape_position != MaxSector) + return -EBUSY; + + if (conf->geo.far_copies > 1 && !conf->geo.far_offset) + return -EINVAL; + + oldsize = raid10_size(mddev, 0, 0); + size = raid10_size(mddev, sectors, 0); + if (mddev->external_size && + mddev->array_sectors > size) + return -EINVAL; + if (mddev->bitmap) { + int ret = bitmap_resize(mddev->bitmap, size, 0, 0); + if (ret) + return ret; + } + md_set_array_sectors(mddev, size); + set_capacity(mddev->gendisk, mddev->array_sectors); + revalidate_disk(mddev->gendisk); + if (sectors > mddev->dev_sectors && + mddev->recovery_cp > oldsize) { + mddev->recovery_cp = oldsize; + set_bit(MD_RECOVERY_NEEDED, &mddev->recovery); + } + calc_sectors(conf, sectors); + mddev->dev_sectors = conf->dev_sectors; + mddev->resync_max_sectors = size; + return 0; +} + static void *raid10_takeover_raid0(struct mddev *mddev) { struct md_rdev *rdev; @@ -3392,7 +3921,7 @@ static void *raid10_takeover_raid0(struct mddev *mddev) conf = setup_conf(mddev); if (!IS_ERR(conf)) { - list_for_each_entry(rdev, &mddev->disks, same_set) + rdev_for_each(rdev, mddev) if (rdev->raid_disk >= 0) rdev->new_raid_disk = rdev->raid_disk * 2; conf->barrier = 1; @@ -3422,6 +3951,763 @@ static void *raid10_takeover(struct mddev *mddev) return ERR_PTR(-EINVAL); } +static int raid10_check_reshape(struct mddev *mddev) +{ + /* Called when there is a request to change + * - layout (to ->new_layout) + * - chunk size (to ->new_chunk_sectors) + * - raid_disks (by delta_disks) + * or when trying to restart a reshape that was ongoing. + * + * We need to validate the request and possibly allocate + * space if that might be an issue later. + * + * Currently we reject any reshape of a 'far' mode array, + * allow chunk size to change if new is generally acceptable, + * allow raid_disks to increase, and allow + * a switch between 'near' mode and 'offset' mode. + */ + struct r10conf *conf = mddev->private; + struct geom geo; + + if (conf->geo.far_copies != 1 && !conf->geo.far_offset) + return -EINVAL; + + if (setup_geo(&geo, mddev, geo_start) != conf->copies) + /* mustn't change number of copies */ + return -EINVAL; + if (geo.far_copies > 1 && !geo.far_offset) + /* Cannot switch to 'far' mode */ + return -EINVAL; + + if (mddev->array_sectors & geo.chunk_mask) + /* not factor of array size */ + return -EINVAL; + + if (!enough(conf, -1)) + return -EINVAL; + + kfree(conf->mirrors_new); + conf->mirrors_new = NULL; + if (mddev->delta_disks > 0) { + /* allocate new 'mirrors' list */ + conf->mirrors_new = kzalloc( + sizeof(struct raid10_info) + *(mddev->raid_disks + + mddev->delta_disks), + GFP_KERNEL); + if (!conf->mirrors_new) + return -ENOMEM; + } + return 0; +} + +/* + * Need to check if array has failed when deciding whether to: + * - start an array + * - remove non-faulty devices + * - add a spare + * - allow a reshape + * This determination is simple when no reshape is happening. + * However if there is a reshape, we need to carefully check + * both the before and after sections. + * This is because some failed devices may only affect one + * of the two sections, and some non-in_sync devices may + * be insync in the section most affected by failed devices. + */ +static int calc_degraded(struct r10conf *conf) +{ + int degraded, degraded2; + int i; + + rcu_read_lock(); + degraded = 0; + /* 'prev' section first */ + for (i = 0; i < conf->prev.raid_disks; i++) { + struct md_rdev *rdev = rcu_dereference(conf->mirrors[i].rdev); + if (!rdev || test_bit(Faulty, &rdev->flags)) + degraded++; + else if (!test_bit(In_sync, &rdev->flags)) + /* When we can reduce the number of devices in + * an array, this might not contribute to + * 'degraded'. It does now. + */ + degraded++; + } + rcu_read_unlock(); + if (conf->geo.raid_disks == conf->prev.raid_disks) + return degraded; + rcu_read_lock(); + degraded2 = 0; + for (i = 0; i < conf->geo.raid_disks; i++) { + struct md_rdev *rdev = rcu_dereference(conf->mirrors[i].rdev); + if (!rdev || test_bit(Faulty, &rdev->flags)) + degraded2++; + else if (!test_bit(In_sync, &rdev->flags)) { + /* If reshape is increasing the number of devices, + * this section has already been recovered, so + * it doesn't contribute to degraded. + * else it does. + */ + if (conf->geo.raid_disks <= conf->prev.raid_disks) + degraded2++; + } + } + rcu_read_unlock(); + if (degraded2 > degraded) + return degraded2; + return degraded; +} + +static int raid10_start_reshape(struct mddev *mddev) +{ + /* A 'reshape' has been requested. This commits + * the various 'new' fields and sets MD_RECOVER_RESHAPE + * This also checks if there are enough spares and adds them + * to the array. + * We currently require enough spares to make the final + * array non-degraded. We also require that the difference + * between old and new data_offset - on each device - is + * enough that we never risk over-writing. + */ + + unsigned long before_length, after_length; + sector_t min_offset_diff = 0; + int first = 1; + struct geom new; + struct r10conf *conf = mddev->private; + struct md_rdev *rdev; + int spares = 0; + int ret; + + if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery)) + return -EBUSY; + + if (setup_geo(&new, mddev, geo_start) != conf->copies) + return -EINVAL; + + before_length = ((1 << conf->prev.chunk_shift) * + conf->prev.far_copies); + after_length = ((1 << conf->geo.chunk_shift) * + conf->geo.far_copies); + + rdev_for_each(rdev, mddev) { + if (!test_bit(In_sync, &rdev->flags) + && !test_bit(Faulty, &rdev->flags)) + spares++; + if (rdev->raid_disk >= 0) { + long long diff = (rdev->new_data_offset + - rdev->data_offset); + if (!mddev->reshape_backwards) + diff = -diff; + if (diff < 0) + diff = 0; + if (first || diff < min_offset_diff) + min_offset_diff = diff; + } + } + + if (max(before_length, after_length) > min_offset_diff) + return -EINVAL; + + if (spares < mddev->delta_disks) + return -EINVAL; + + conf->offset_diff = min_offset_diff; + spin_lock_irq(&conf->device_lock); + if (conf->mirrors_new) { + memcpy(conf->mirrors_new, conf->mirrors, + sizeof(struct raid10_info)*conf->prev.raid_disks); + smp_mb(); + kfree(conf->mirrors_old); /* FIXME and elsewhere */ + conf->mirrors_old = conf->mirrors; + conf->mirrors = conf->mirrors_new; + conf->mirrors_new = NULL; + } + setup_geo(&conf->geo, mddev, geo_start); + smp_mb(); + if (mddev->reshape_backwards) { + sector_t size = raid10_size(mddev, 0, 0); + if (size < mddev->array_sectors) { + spin_unlock_irq(&conf->device_lock); + printk(KERN_ERR "md/raid10:%s: array size must be reduce before number of disks\n", + mdname(mddev)); + return -EINVAL; + } + mddev->resync_max_sectors = size; + conf->reshape_progress = size; + } else + conf->reshape_progress = 0; + spin_unlock_irq(&conf->device_lock); + + if (mddev->delta_disks && mddev->bitmap) { + ret = bitmap_resize(mddev->bitmap, + raid10_size(mddev, 0, + conf->geo.raid_disks), + 0, 0); + if (ret) + goto abort; + } + if (mddev->delta_disks > 0) { + rdev_for_each(rdev, mddev) + if (rdev->raid_disk < 0 && + !test_bit(Faulty, &rdev->flags)) { + if (raid10_add_disk(mddev, rdev) == 0) { + if (rdev->raid_disk >= + conf->prev.raid_disks) + set_bit(In_sync, &rdev->flags); + else + rdev->recovery_offset = 0; + + if (sysfs_link_rdev(mddev, rdev)) + /* Failure here is OK */; + } + } else if (rdev->raid_disk >= conf->prev.raid_disks + && !test_bit(Faulty, &rdev->flags)) { + /* This is a spare that was manually added */ + set_bit(In_sync, &rdev->flags); + } + } + /* When a reshape changes the number of devices, + * ->degraded is measured against the larger of the + * pre and post numbers. + */ + spin_lock_irq(&conf->device_lock); + mddev->degraded = calc_degraded(conf); + spin_unlock_irq(&conf->device_lock); + mddev->raid_disks = conf->geo.raid_disks; + mddev->reshape_position = conf->reshape_progress; + set_bit(MD_CHANGE_DEVS, &mddev->flags); + + clear_bit(MD_RECOVERY_SYNC, &mddev->recovery); + clear_bit(MD_RECOVERY_CHECK, &mddev->recovery); + set_bit(MD_RECOVERY_RESHAPE, &mddev->recovery); + set_bit(MD_RECOVERY_RUNNING, &mddev->recovery); + + mddev->sync_thread = md_register_thread(md_do_sync, mddev, + "reshape"); + if (!mddev->sync_thread) { + ret = -EAGAIN; + goto abort; + } + conf->reshape_checkpoint = jiffies; + md_wakeup_thread(mddev->sync_thread); + md_new_event(mddev); + return 0; + +abort: + mddev->recovery = 0; + spin_lock_irq(&conf->device_lock); + conf->geo = conf->prev; + mddev->raid_disks = conf->geo.raid_disks; + rdev_for_each(rdev, mddev) + rdev->new_data_offset = rdev->data_offset; + smp_wmb(); + conf->reshape_progress = MaxSector; + mddev->reshape_position = MaxSector; + spin_unlock_irq(&conf->device_lock); + return ret; +} + +/* Calculate the last device-address that could contain + * any block from the chunk that includes the array-address 's' + * and report the next address. + * i.e. the address returned will be chunk-aligned and after + * any data that is in the chunk containing 's'. + */ +static sector_t last_dev_address(sector_t s, struct geom *geo) +{ + s = (s | geo->chunk_mask) + 1; + s >>= geo->chunk_shift; + s *= geo->near_copies; + s = DIV_ROUND_UP_SECTOR_T(s, geo->raid_disks); + s *= geo->far_copies; + s <<= geo->chunk_shift; + return s; +} + +/* Calculate the first device-address that could contain + * any block from the chunk that includes the array-address 's'. + * This too will be the start of a chunk + */ +static sector_t first_dev_address(sector_t s, struct geom *geo) +{ + s >>= geo->chunk_shift; + s *= geo->near_copies; + sector_div(s, geo->raid_disks); + s *= geo->far_copies; + s <<= geo->chunk_shift; + return s; +} + +static sector_t reshape_request(struct mddev *mddev, sector_t sector_nr, + int *skipped) +{ + /* We simply copy at most one chunk (smallest of old and new) + * at a time, possibly less if that exceeds RESYNC_PAGES, + * or we hit a bad block or something. + * This might mean we pause for normal IO in the middle of + * a chunk, but that is not a problem was mddev->reshape_position + * can record any location. + * + * If we will want to write to a location that isn't + * yet recorded as 'safe' (i.e. in metadata on disk) then + * we need to flush all reshape requests and update the metadata. + * + * When reshaping forwards (e.g. to more devices), we interpret + * 'safe' as the earliest block which might not have been copied + * down yet. We divide this by previous stripe size and multiply + * by previous stripe length to get lowest device offset that we + * cannot write to yet. + * We interpret 'sector_nr' as an address that we want to write to. + * From this we use last_device_address() to find where we might + * write to, and first_device_address on the 'safe' position. + * If this 'next' write position is after the 'safe' position, + * we must update the metadata to increase the 'safe' position. + * + * When reshaping backwards, we round in the opposite direction + * and perform the reverse test: next write position must not be + * less than current safe position. + * + * In all this the minimum difference in data offsets + * (conf->offset_diff - always positive) allows a bit of slack, + * so next can be after 'safe', but not by more than offset_disk + * + * We need to prepare all the bios here before we start any IO + * to ensure the size we choose is acceptable to all devices. + * The means one for each copy for write-out and an extra one for + * read-in. + * We store the read-in bio in ->master_bio and the others in + * ->devs[x].bio and ->devs[x].repl_bio. + */ + struct r10conf *conf = mddev->private; + struct r10bio *r10_bio; + sector_t next, safe, last; + int max_sectors; + int nr_sectors; + int s; + struct md_rdev *rdev; + int need_flush = 0; + struct bio *blist; + struct bio *bio, *read_bio; + int sectors_done = 0; + + if (sector_nr == 0) { + /* If restarting in the middle, skip the initial sectors */ + if (mddev->reshape_backwards && + conf->reshape_progress < raid10_size(mddev, 0, 0)) { + sector_nr = (raid10_size(mddev, 0, 0) + - conf->reshape_progress); + } else if (!mddev->reshape_backwards && + conf->reshape_progress > 0) + sector_nr = conf->reshape_progress; + if (sector_nr) { + mddev->curr_resync_completed = sector_nr; + sysfs_notify(&mddev->kobj, NULL, "sync_completed"); + *skipped = 1; + return sector_nr; + } + } + + /* We don't use sector_nr to track where we are up to + * as that doesn't work well for ->reshape_backwards. + * So just use ->reshape_progress. + */ + if (mddev->reshape_backwards) { + /* 'next' is the earliest device address that we might + * write to for this chunk in the new layout + */ + next = first_dev_address(conf->reshape_progress - 1, + &conf->geo); + + /* 'safe' is the last device address that we might read from + * in the old layout after a restart + */ + safe = last_dev_address(conf->reshape_safe - 1, + &conf->prev); + + if (next + conf->offset_diff < safe) + need_flush = 1; + + last = conf->reshape_progress - 1; + sector_nr = last & ~(sector_t)(conf->geo.chunk_mask + & conf->prev.chunk_mask); + if (sector_nr + RESYNC_BLOCK_SIZE/512 < last) + sector_nr = last + 1 - RESYNC_BLOCK_SIZE/512; + } else { + /* 'next' is after the last device address that we + * might write to for this chunk in the new layout + */ + next = last_dev_address(conf->reshape_progress, &conf->geo); + + /* 'safe' is the earliest device address that we might + * read from in the old layout after a restart + */ + safe = first_dev_address(conf->reshape_safe, &conf->prev); + + /* Need to update metadata if 'next' might be beyond 'safe' + * as that would possibly corrupt data + */ + if (next > safe + conf->offset_diff) + need_flush = 1; + + sector_nr = conf->reshape_progress; + last = sector_nr | (conf->geo.chunk_mask + & conf->prev.chunk_mask); + + if (sector_nr + RESYNC_BLOCK_SIZE/512 <= last) + last = sector_nr + RESYNC_BLOCK_SIZE/512 - 1; + } + + if (need_flush || + time_after(jiffies, conf->reshape_checkpoint + 10*HZ)) { + /* Need to update reshape_position in metadata */ + wait_barrier(conf); + mddev->reshape_position = conf->reshape_progress; + if (mddev->reshape_backwards) + mddev->curr_resync_completed = raid10_size(mddev, 0, 0) + - conf->reshape_progress; + else + mddev->curr_resync_completed = conf->reshape_progress; + conf->reshape_checkpoint = jiffies; + set_bit(MD_CHANGE_DEVS, &mddev->flags); + md_wakeup_thread(mddev->thread); + wait_event(mddev->sb_wait, mddev->flags == 0 || + test_bit(MD_RECOVERY_INTR, &mddev->recovery)); + if (test_bit(MD_RECOVERY_INTR, &mddev->recovery)) { + allow_barrier(conf); + return sectors_done; + } + conf->reshape_safe = mddev->reshape_position; + allow_barrier(conf); + } + +read_more: + /* Now schedule reads for blocks from sector_nr to last */ + r10_bio = mempool_alloc(conf->r10buf_pool, GFP_NOIO); + raise_barrier(conf, sectors_done != 0); + atomic_set(&r10_bio->remaining, 0); + r10_bio->mddev = mddev; + r10_bio->sector = sector_nr; + set_bit(R10BIO_IsReshape, &r10_bio->state); + r10_bio->sectors = last - sector_nr + 1; + rdev = read_balance(conf, r10_bio, &max_sectors); + BUG_ON(!test_bit(R10BIO_Previous, &r10_bio->state)); + + if (!rdev) { + /* Cannot read from here, so need to record bad blocks + * on all the target devices. + */ + // FIXME + set_bit(MD_RECOVERY_INTR, &mddev->recovery); + return sectors_done; + } + + read_bio = bio_alloc_mddev(GFP_KERNEL, RESYNC_PAGES, mddev); + + read_bio->bi_bdev = rdev->bdev; + read_bio->bi_iter.bi_sector = (r10_bio->devs[r10_bio->read_slot].addr + + rdev->data_offset); + read_bio->bi_private = r10_bio; + read_bio->bi_end_io = end_sync_read; + read_bio->bi_rw = READ; + read_bio->bi_flags &= ~(BIO_POOL_MASK - 1); + read_bio->bi_flags |= 1 << BIO_UPTODATE; + read_bio->bi_vcnt = 0; + read_bio->bi_iter.bi_size = 0; + r10_bio->master_bio = read_bio; + r10_bio->read_slot = r10_bio->devs[r10_bio->read_slot].devnum; + + /* Now find the locations in the new layout */ + __raid10_find_phys(&conf->geo, r10_bio); + + blist = read_bio; + read_bio->bi_next = NULL; + + for (s = 0; s < conf->copies*2; s++) { + struct bio *b; + int d = r10_bio->devs[s/2].devnum; + struct md_rdev *rdev2; + if (s&1) { + rdev2 = conf->mirrors[d].replacement; + b = r10_bio->devs[s/2].repl_bio; + } else { + rdev2 = conf->mirrors[d].rdev; + b = r10_bio->devs[s/2].bio; + } + if (!rdev2 || test_bit(Faulty, &rdev2->flags)) + continue; + + bio_reset(b); + b->bi_bdev = rdev2->bdev; + b->bi_iter.bi_sector = r10_bio->devs[s/2].addr + + rdev2->new_data_offset; + b->bi_private = r10_bio; + b->bi_end_io = end_reshape_write; + b->bi_rw = WRITE; + b->bi_next = blist; + blist = b; + } + + /* Now add as many pages as possible to all of these bios. */ + + nr_sectors = 0; + for (s = 0 ; s < max_sectors; s += PAGE_SIZE >> 9) { + struct page *page = r10_bio->devs[0].bio->bi_io_vec[s/(PAGE_SIZE>>9)].bv_page; + int len = (max_sectors - s) << 9; + if (len > PAGE_SIZE) + len = PAGE_SIZE; + for (bio = blist; bio ; bio = bio->bi_next) { + struct bio *bio2; + if (bio_add_page(bio, page, len, 0)) + continue; + + /* Didn't fit, must stop */ + for (bio2 = blist; + bio2 && bio2 != bio; + bio2 = bio2->bi_next) { + /* Remove last page from this bio */ + bio2->bi_vcnt--; + bio2->bi_iter.bi_size -= len; + bio2->bi_flags &= ~(1<<BIO_SEG_VALID); + } + goto bio_full; + } + sector_nr += len >> 9; + nr_sectors += len >> 9; + } +bio_full: + r10_bio->sectors = nr_sectors; + + /* Now submit the read */ + md_sync_acct(read_bio->bi_bdev, r10_bio->sectors); + atomic_inc(&r10_bio->remaining); + read_bio->bi_next = NULL; + generic_make_request(read_bio); + sector_nr += nr_sectors; + sectors_done += nr_sectors; + if (sector_nr <= last) + goto read_more; + + /* Now that we have done the whole section we can + * update reshape_progress + */ + if (mddev->reshape_backwards) + conf->reshape_progress -= sectors_done; + else + conf->reshape_progress += sectors_done; + + return sectors_done; +} + +static void end_reshape_request(struct r10bio *r10_bio); +static int handle_reshape_read_error(struct mddev *mddev, + struct r10bio *r10_bio); +static void reshape_request_write(struct mddev *mddev, struct r10bio *r10_bio) +{ + /* Reshape read completed. Hopefully we have a block + * to write out. + * If we got a read error then we do sync 1-page reads from + * elsewhere until we find the data - or give up. + */ + struct r10conf *conf = mddev->private; + int s; + + if (!test_bit(R10BIO_Uptodate, &r10_bio->state)) + if (handle_reshape_read_error(mddev, r10_bio) < 0) { + /* Reshape has been aborted */ + md_done_sync(mddev, r10_bio->sectors, 0); + return; + } + + /* We definitely have the data in the pages, schedule the + * writes. + */ + atomic_set(&r10_bio->remaining, 1); + for (s = 0; s < conf->copies*2; s++) { + struct bio *b; + int d = r10_bio->devs[s/2].devnum; + struct md_rdev *rdev; + if (s&1) { + rdev = conf->mirrors[d].replacement; + b = r10_bio->devs[s/2].repl_bio; + } else { + rdev = conf->mirrors[d].rdev; + b = r10_bio->devs[s/2].bio; + } + if (!rdev || test_bit(Faulty, &rdev->flags)) + continue; + atomic_inc(&rdev->nr_pending); + md_sync_acct(b->bi_bdev, r10_bio->sectors); + atomic_inc(&r10_bio->remaining); + b->bi_next = NULL; + generic_make_request(b); + } + end_reshape_request(r10_bio); +} + +static void end_reshape(struct r10conf *conf) +{ + if (test_bit(MD_RECOVERY_INTR, &conf->mddev->recovery)) + return; + + spin_lock_irq(&conf->device_lock); + conf->prev = conf->geo; + md_finish_reshape(conf->mddev); + smp_wmb(); + conf->reshape_progress = MaxSector; + spin_unlock_irq(&conf->device_lock); + + /* read-ahead size must cover two whole stripes, which is + * 2 * (datadisks) * chunksize where 'n' is the number of raid devices + */ + if (conf->mddev->queue) { + int stripe = conf->geo.raid_disks * + ((conf->mddev->chunk_sectors << 9) / PAGE_SIZE); + stripe /= conf->geo.near_copies; + if (conf->mddev->queue->backing_dev_info.ra_pages < 2 * stripe) + conf->mddev->queue->backing_dev_info.ra_pages = 2 * stripe; + } + conf->fullsync = 0; +} + + +static int handle_reshape_read_error(struct mddev *mddev, + struct r10bio *r10_bio) +{ + /* Use sync reads to get the blocks from somewhere else */ + int sectors = r10_bio->sectors; + struct r10conf *conf = mddev->private; + struct { + struct r10bio r10_bio; + struct r10dev devs[conf->copies]; + } on_stack; + struct r10bio *r10b = &on_stack.r10_bio; + int slot = 0; + int idx = 0; + struct bio_vec *bvec = r10_bio->master_bio->bi_io_vec; + + r10b->sector = r10_bio->sector; + __raid10_find_phys(&conf->prev, r10b); + + while (sectors) { + int s = sectors; + int success = 0; + int first_slot = slot; + + if (s > (PAGE_SIZE >> 9)) + s = PAGE_SIZE >> 9; + + while (!success) { + int d = r10b->devs[slot].devnum; + struct md_rdev *rdev = conf->mirrors[d].rdev; + sector_t addr; + if (rdev == NULL || + test_bit(Faulty, &rdev->flags) || + !test_bit(In_sync, &rdev->flags)) + goto failed; + + addr = r10b->devs[slot].addr + idx * PAGE_SIZE; + success = sync_page_io(rdev, + addr, + s << 9, + bvec[idx].bv_page, + READ, false); + if (success) + break; + failed: + slot++; + if (slot >= conf->copies) + slot = 0; + if (slot == first_slot) + break; + } + if (!success) { + /* couldn't read this block, must give up */ + set_bit(MD_RECOVERY_INTR, + &mddev->recovery); + return -EIO; + } + sectors -= s; + idx++; + } + return 0; +} + +static void end_reshape_write(struct bio *bio, int error) +{ + int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags); + struct r10bio *r10_bio = bio->bi_private; + struct mddev *mddev = r10_bio->mddev; + struct r10conf *conf = mddev->private; + int d; + int slot; + int repl; + struct md_rdev *rdev = NULL; + + d = find_bio_disk(conf, r10_bio, bio, &slot, &repl); + if (repl) + rdev = conf->mirrors[d].replacement; + if (!rdev) { + smp_mb(); + rdev = conf->mirrors[d].rdev; + } + + if (!uptodate) { + /* FIXME should record badblock */ + md_error(mddev, rdev); + } + + rdev_dec_pending(rdev, mddev); + end_reshape_request(r10_bio); +} + +static void end_reshape_request(struct r10bio *r10_bio) +{ + if (!atomic_dec_and_test(&r10_bio->remaining)) + return; + md_done_sync(r10_bio->mddev, r10_bio->sectors, 1); + bio_put(r10_bio->master_bio); + put_buf(r10_bio); +} + +static void raid10_finish_reshape(struct mddev *mddev) +{ + struct r10conf *conf = mddev->private; + + if (test_bit(MD_RECOVERY_INTR, &mddev->recovery)) + return; + + if (mddev->delta_disks > 0) { + sector_t size = raid10_size(mddev, 0, 0); + md_set_array_sectors(mddev, size); + if (mddev->recovery_cp > mddev->resync_max_sectors) { + mddev->recovery_cp = mddev->resync_max_sectors; + set_bit(MD_RECOVERY_NEEDED, &mddev->recovery); + } + mddev->resync_max_sectors = size; + set_capacity(mddev->gendisk, mddev->array_sectors); + revalidate_disk(mddev->gendisk); + } else { + int d; + for (d = conf->geo.raid_disks ; + d < conf->geo.raid_disks - mddev->delta_disks; + d++) { + struct md_rdev *rdev = conf->mirrors[d].rdev; + if (rdev) + clear_bit(In_sync, &rdev->flags); + rdev = conf->mirrors[d].replacement; + if (rdev) + clear_bit(In_sync, &rdev->flags); + } + } + mddev->layout = mddev->new_layout; + mddev->chunk_sectors = 1 << conf->geo.chunk_shift; + mddev->reshape_position = MaxSector; + mddev->delta_disks = 0; + mddev->reshape_backwards = 0; +} + static struct md_personality raid10_personality = { .name = "raid10", @@ -3438,7 +4724,11 @@ static struct md_personality raid10_personality = .sync_request = sync_request, .quiesce = raid10_quiesce, .size = raid10_size, + .resize = raid10_resize, .takeover = raid10_takeover, + .check_reshape = raid10_check_reshape, + .start_reshape = raid10_start_reshape, + .finish_reshape = raid10_finish_reshape, }; static int __init raid_init(void) |