diff options
Diffstat (limited to 'drivers/md/raid0.c')
| -rw-r--r-- | drivers/md/raid0.c | 864 |
1 files changed, 540 insertions, 324 deletions
diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c index 818b4828409..407a99e46f6 100644 --- a/drivers/md/raid0.c +++ b/drivers/md/raid0.c @@ -18,35 +18,26 @@ Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ +#include <linux/blkdev.h> +#include <linux/seq_file.h> #include <linux/module.h> -#include <linux/raid/raid0.h> - -#define MAJOR_NR MD_MAJOR -#define MD_DRIVER -#define MD_PERSONALITY - -static void raid0_unplug(struct request_queue *q) -{ - mddev_t *mddev = q->queuedata; - raid0_conf_t *conf = mddev_to_conf(mddev); - mdk_rdev_t **devlist = conf->strip_zone[0].dev; - int i; - - for (i=0; i<mddev->raid_disks; i++) { - struct request_queue *r_queue = bdev_get_queue(devlist[i]->bdev); - - blk_unplug(r_queue); - } -} +#include <linux/slab.h> +#include "md.h" +#include "raid0.h" +#include "raid5.h" static int raid0_congested(void *data, int bits) { - mddev_t *mddev = data; - raid0_conf_t *conf = mddev_to_conf(mddev); - mdk_rdev_t **devlist = conf->strip_zone[0].dev; + struct mddev *mddev = data; + struct r0conf *conf = mddev->private; + struct md_rdev **devlist = conf->devlist; + int raid_disks = conf->strip_zone[0].nb_dev; int i, ret = 0; - for (i = 0; i < mddev->raid_disks && !ret ; i++) { + if (mddev_congested(mddev, bits)) + return 1; + + for (i = 0; i < raid_disks && !ret ; i++) { struct request_queue *q = bdev_get_queue(devlist[i]->bdev); ret |= bdi_congested(&q->backing_dev_info, bits); @@ -54,68 +45,110 @@ static int raid0_congested(void *data, int bits) return ret; } +/* + * inform the user of the raid configuration +*/ +static void dump_zones(struct mddev *mddev) +{ + int j, k; + sector_t zone_size = 0; + sector_t zone_start = 0; + char b[BDEVNAME_SIZE]; + struct r0conf *conf = mddev->private; + int raid_disks = conf->strip_zone[0].nb_dev; + printk(KERN_INFO "md: RAID0 configuration for %s - %d zone%s\n", + mdname(mddev), + conf->nr_strip_zones, conf->nr_strip_zones==1?"":"s"); + for (j = 0; j < conf->nr_strip_zones; j++) { + printk(KERN_INFO "md: zone%d=[", j); + for (k = 0; k < conf->strip_zone[j].nb_dev; k++) + printk(KERN_CONT "%s%s", k?"/":"", + bdevname(conf->devlist[j*raid_disks + + k]->bdev, b)); + printk(KERN_CONT "]\n"); + + zone_size = conf->strip_zone[j].zone_end - zone_start; + printk(KERN_INFO " zone-offset=%10lluKB, " + "device-offset=%10lluKB, size=%10lluKB\n", + (unsigned long long)zone_start>>1, + (unsigned long long)conf->strip_zone[j].dev_start>>1, + (unsigned long long)zone_size>>1); + zone_start = conf->strip_zone[j].zone_end; + } + printk(KERN_INFO "\n"); +} -static int create_strip_zones (mddev_t *mddev) +static int create_strip_zones(struct mddev *mddev, struct r0conf **private_conf) { - int i, c, j; - sector_t current_offset, curr_zone_offset; - sector_t min_spacing; - raid0_conf_t *conf = mddev_to_conf(mddev); - mdk_rdev_t *smallest, *rdev1, *rdev2, *rdev; - struct list_head *tmp1, *tmp2; + int i, c, err; + sector_t curr_zone_end, sectors; + struct md_rdev *smallest, *rdev1, *rdev2, *rdev, **dev; struct strip_zone *zone; int cnt; char b[BDEVNAME_SIZE]; - - /* - * The number of 'same size groups' - */ - conf->nr_strip_zones = 0; - - rdev_for_each(rdev1, tmp1, mddev) { - printk("raid0: looking at %s\n", - bdevname(rdev1->bdev,b)); + char b2[BDEVNAME_SIZE]; + struct r0conf *conf = kzalloc(sizeof(*conf), GFP_KERNEL); + bool discard_supported = false; + + if (!conf) + return -ENOMEM; + rdev_for_each(rdev1, mddev) { + pr_debug("md/raid0:%s: looking at %s\n", + mdname(mddev), + bdevname(rdev1->bdev, b)); c = 0; - rdev_for_each(rdev2, tmp2, mddev) { - printk("raid0: comparing %s(%llu)", - bdevname(rdev1->bdev,b), - (unsigned long long)rdev1->size); - printk(" with %s(%llu)\n", - bdevname(rdev2->bdev,b), - (unsigned long long)rdev2->size); + + /* round size to chunk_size */ + sectors = rdev1->sectors; + sector_div(sectors, mddev->chunk_sectors); + rdev1->sectors = sectors * mddev->chunk_sectors; + + rdev_for_each(rdev2, mddev) { + pr_debug("md/raid0:%s: comparing %s(%llu)" + " with %s(%llu)\n", + mdname(mddev), + bdevname(rdev1->bdev,b), + (unsigned long long)rdev1->sectors, + bdevname(rdev2->bdev,b2), + (unsigned long long)rdev2->sectors); if (rdev2 == rdev1) { - printk("raid0: END\n"); + pr_debug("md/raid0:%s: END\n", + mdname(mddev)); break; } - if (rdev2->size == rdev1->size) - { + if (rdev2->sectors == rdev1->sectors) { /* * Not unique, don't count it as a new * group */ - printk("raid0: EQUAL\n"); + pr_debug("md/raid0:%s: EQUAL\n", + mdname(mddev)); c = 1; break; } - printk("raid0: NOT EQUAL\n"); + pr_debug("md/raid0:%s: NOT EQUAL\n", + mdname(mddev)); } if (!c) { - printk("raid0: ==> UNIQUE\n"); + pr_debug("md/raid0:%s: ==> UNIQUE\n", + mdname(mddev)); conf->nr_strip_zones++; - printk("raid0: %d zones\n", conf->nr_strip_zones); + pr_debug("md/raid0:%s: %d zones\n", + mdname(mddev), conf->nr_strip_zones); } } - printk("raid0: FINAL %d zones\n", conf->nr_strip_zones); - + pr_debug("md/raid0:%s: FINAL %d zones\n", + mdname(mddev), conf->nr_strip_zones); + err = -ENOMEM; conf->strip_zone = kzalloc(sizeof(struct strip_zone)* conf->nr_strip_zones, GFP_KERNEL); if (!conf->strip_zone) - return 1; - conf->devlist = kzalloc(sizeof(mdk_rdev_t*)* + goto abort; + conf->devlist = kzalloc(sizeof(struct md_rdev*)* conf->nr_strip_zones*mddev->raid_disks, GFP_KERNEL); if (!conf->devlist) - return 1; + goto abort; /* The first zone must contain all devices, so here we check that * there is a proper alignment of slots to devices and find them all @@ -123,226 +156,306 @@ static int create_strip_zones (mddev_t *mddev) zone = &conf->strip_zone[0]; cnt = 0; smallest = NULL; - zone->dev = conf->devlist; - rdev_for_each(rdev1, tmp1, mddev) { + dev = conf->devlist; + err = -EINVAL; + rdev_for_each(rdev1, mddev) { int j = rdev1->raid_disk; - if (j < 0 || j >= mddev->raid_disks) { - printk("raid0: bad disk number %d - aborting!\n", j); + if (mddev->level == 10) { + /* taking over a raid10-n2 array */ + j /= 2; + rdev1->new_raid_disk = j; + } + + if (mddev->level == 1) { + /* taiking over a raid1 array- + * we have only one active disk + */ + j = 0; + rdev1->new_raid_disk = j; + } + + if (j < 0) { + printk(KERN_ERR + "md/raid0:%s: remove inactive devices before converting to RAID0\n", + mdname(mddev)); goto abort; } - if (zone->dev[j]) { - printk("raid0: multiple devices for %d - aborting!\n", - j); + if (j >= mddev->raid_disks) { + printk(KERN_ERR "md/raid0:%s: bad disk number %d - " + "aborting!\n", mdname(mddev), j); goto abort; } - zone->dev[j] = rdev1; + if (dev[j]) { + printk(KERN_ERR "md/raid0:%s: multiple devices for %d - " + "aborting!\n", mdname(mddev), j); + goto abort; + } + dev[j] = rdev1; - blk_queue_stack_limits(mddev->queue, - rdev1->bdev->bd_disk->queue); - /* as we don't honour merge_bvec_fn, we must never risk - * violating it, so limit ->max_sector to one PAGE, as - * a one page request is never in violation. - */ + disk_stack_limits(mddev->gendisk, rdev1->bdev, + rdev1->data_offset << 9); - if (rdev1->bdev->bd_disk->queue->merge_bvec_fn && - mddev->queue->max_sectors > (PAGE_SIZE>>9)) - blk_queue_max_sectors(mddev->queue, PAGE_SIZE>>9); + if (rdev1->bdev->bd_disk->queue->merge_bvec_fn) + conf->has_merge_bvec = 1; - if (!smallest || (rdev1->size <smallest->size)) + if (!smallest || (rdev1->sectors < smallest->sectors)) smallest = rdev1; cnt++; + + if (blk_queue_discard(bdev_get_queue(rdev1->bdev))) + discard_supported = true; } if (cnt != mddev->raid_disks) { - printk("raid0: too few disks (%d of %d) - aborting!\n", - cnt, mddev->raid_disks); + printk(KERN_ERR "md/raid0:%s: too few disks (%d of %d) - " + "aborting!\n", mdname(mddev), cnt, mddev->raid_disks); goto abort; } zone->nb_dev = cnt; - zone->size = smallest->size * cnt; - zone->zone_offset = 0; + zone->zone_end = smallest->sectors * cnt; - current_offset = smallest->size; - curr_zone_offset = zone->size; + curr_zone_end = zone->zone_end; /* now do the other zones */ for (i = 1; i < conf->nr_strip_zones; i++) { + int j; + zone = conf->strip_zone + i; - zone->dev = conf->strip_zone[i-1].dev + mddev->raid_disks; + dev = conf->devlist + i * mddev->raid_disks; - printk("raid0: zone %d\n", i); - zone->dev_offset = current_offset; + pr_debug("md/raid0:%s: zone %d\n", mdname(mddev), i); + zone->dev_start = smallest->sectors; smallest = NULL; c = 0; for (j=0; j<cnt; j++) { - char b[BDEVNAME_SIZE]; - rdev = conf->strip_zone[0].dev[j]; - printk("raid0: checking %s ...", bdevname(rdev->bdev,b)); - if (rdev->size > current_offset) - { - printk(" contained as device %d\n", c); - zone->dev[c] = rdev; - c++; - if (!smallest || (rdev->size <smallest->size)) { - smallest = rdev; - printk(" (%llu) is smallest!.\n", - (unsigned long long)rdev->size); - } - } else - printk(" nope.\n"); + rdev = conf->devlist[j]; + if (rdev->sectors <= zone->dev_start) { + pr_debug("md/raid0:%s: checking %s ... nope\n", + mdname(mddev), + bdevname(rdev->bdev, b)); + continue; + } + pr_debug("md/raid0:%s: checking %s ..." + " contained as device %d\n", + mdname(mddev), + bdevname(rdev->bdev, b), c); + dev[c] = rdev; + c++; + if (!smallest || rdev->sectors < smallest->sectors) { + smallest = rdev; + pr_debug("md/raid0:%s: (%llu) is smallest!.\n", + mdname(mddev), + (unsigned long long)rdev->sectors); + } } zone->nb_dev = c; - zone->size = (smallest->size - current_offset) * c; - printk("raid0: zone->nb_dev: %d, size: %llu\n", - zone->nb_dev, (unsigned long long)zone->size); + sectors = (smallest->sectors - zone->dev_start) * c; + pr_debug("md/raid0:%s: zone->nb_dev: %d, sectors: %llu\n", + mdname(mddev), + zone->nb_dev, (unsigned long long)sectors); - zone->zone_offset = curr_zone_offset; - curr_zone_offset += zone->size; + curr_zone_end += sectors; + zone->zone_end = curr_zone_end; - current_offset = smallest->size; - printk("raid0: current zone offset: %llu\n", - (unsigned long long)current_offset); + pr_debug("md/raid0:%s: current zone start: %llu\n", + mdname(mddev), + (unsigned long long)smallest->sectors); } + mddev->queue->backing_dev_info.congested_fn = raid0_congested; + mddev->queue->backing_dev_info.congested_data = mddev; - /* Now find appropriate hash spacing. - * We want a number which causes most hash entries to cover - * at most two strips, but the hash table must be at most - * 1 PAGE. We choose the smallest strip, or contiguous collection - * of strips, that has big enough size. We never consider the last - * strip though as it's size has no bearing on the efficacy of the hash - * table. + /* + * now since we have the hard sector sizes, we can make sure + * chunk size is a multiple of that sector size */ - conf->hash_spacing = curr_zone_offset; - min_spacing = curr_zone_offset; - sector_div(min_spacing, PAGE_SIZE/sizeof(struct strip_zone*)); - for (i=0; i < conf->nr_strip_zones-1; i++) { - sector_t sz = 0; - for (j=i; j<conf->nr_strip_zones-1 && - sz < min_spacing ; j++) - sz += conf->strip_zone[j].size; - if (sz >= min_spacing && sz < conf->hash_spacing) - conf->hash_spacing = sz; + if ((mddev->chunk_sectors << 9) % queue_logical_block_size(mddev->queue)) { + printk(KERN_ERR "md/raid0:%s: chunk_size of %d not valid\n", + mdname(mddev), + mddev->chunk_sectors << 9); + goto abort; } - mddev->queue->unplug_fn = raid0_unplug; + blk_queue_io_min(mddev->queue, mddev->chunk_sectors << 9); + blk_queue_io_opt(mddev->queue, + (mddev->chunk_sectors << 9) * mddev->raid_disks); - mddev->queue->backing_dev_info.congested_fn = raid0_congested; - mddev->queue->backing_dev_info.congested_data = mddev; + if (!discard_supported) + queue_flag_clear_unlocked(QUEUE_FLAG_DISCARD, mddev->queue); + else + queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, mddev->queue); + + pr_debug("md/raid0:%s: done.\n", mdname(mddev)); + *private_conf = conf; - printk("raid0: done.\n"); return 0; - abort: - return 1; +abort: + kfree(conf->strip_zone); + kfree(conf->devlist); + kfree(conf); + *private_conf = ERR_PTR(err); + return err; +} + +/* Find the zone which holds a particular offset + * Update *sectorp to be an offset in that zone + */ +static struct strip_zone *find_zone(struct r0conf *conf, + sector_t *sectorp) +{ + int i; + struct strip_zone *z = conf->strip_zone; + sector_t sector = *sectorp; + + for (i = 0; i < conf->nr_strip_zones; i++) + if (sector < z[i].zone_end) { + if (i) + *sectorp = sector - z[i-1].zone_end; + return z + i; + } + BUG(); +} + +/* + * remaps the bio to the target device. we separate two flows. + * power 2 flow and a general flow for the sake of perfromance +*/ +static struct md_rdev *map_sector(struct mddev *mddev, struct strip_zone *zone, + sector_t sector, sector_t *sector_offset) +{ + unsigned int sect_in_chunk; + sector_t chunk; + struct r0conf *conf = mddev->private; + int raid_disks = conf->strip_zone[0].nb_dev; + unsigned int chunk_sects = mddev->chunk_sectors; + + if (is_power_of_2(chunk_sects)) { + int chunksect_bits = ffz(~chunk_sects); + /* find the sector offset inside the chunk */ + sect_in_chunk = sector & (chunk_sects - 1); + sector >>= chunksect_bits; + /* chunk in zone */ + chunk = *sector_offset; + /* quotient is the chunk in real device*/ + sector_div(chunk, zone->nb_dev << chunksect_bits); + } else{ + sect_in_chunk = sector_div(sector, chunk_sects); + chunk = *sector_offset; + sector_div(chunk, chunk_sects * zone->nb_dev); + } + /* + * position the bio over the real device + * real sector = chunk in device + starting of zone + * + the position in the chunk + */ + *sector_offset = (chunk * chunk_sects) + sect_in_chunk; + return conf->devlist[(zone - conf->strip_zone)*raid_disks + + sector_div(sector, zone->nb_dev)]; } /** - * raid0_mergeable_bvec -- tell bio layer if a two requests can be merged + * raid0_mergeable_bvec -- tell bio layer if two requests can be merged * @q: request queue - * @bio: the buffer head that's been built up so far + * @bvm: properties of new bio * @biovec: the request that could be merged to it. * * Return amount of bytes we can accept at this offset */ -static int raid0_mergeable_bvec(struct request_queue *q, struct bio *bio, struct bio_vec *biovec) +static int raid0_mergeable_bvec(struct request_queue *q, + struct bvec_merge_data *bvm, + struct bio_vec *biovec) { - mddev_t *mddev = q->queuedata; - sector_t sector = bio->bi_sector + get_start_sect(bio->bi_bdev); + struct mddev *mddev = q->queuedata; + struct r0conf *conf = mddev->private; + sector_t sector = bvm->bi_sector + get_start_sect(bvm->bi_bdev); + sector_t sector_offset = sector; int max; - unsigned int chunk_sectors = mddev->chunk_size >> 9; - unsigned int bio_sectors = bio->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 */ + unsigned int chunk_sectors = mddev->chunk_sectors; + unsigned int bio_sectors = bvm->bi_size >> 9; + struct strip_zone *zone; + struct md_rdev *rdev; + struct request_queue *subq; + + if (is_power_of_2(chunk_sectors)) + max = (chunk_sectors - ((sector & (chunk_sectors-1)) + + bio_sectors)) << 9; + else + max = (chunk_sectors - (sector_div(sector, chunk_sectors) + + 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 + if (max < biovec->bv_len) + /* too small already, no need to check further */ + return max; + if (!conf->has_merge_bvec) + return max; + + /* May need to check subordinate device */ + sector = sector_offset; + zone = find_zone(mddev->private, §or_offset); + rdev = map_sector(mddev, zone, sector, §or_offset); + subq = bdev_get_queue(rdev->bdev); + if (subq->merge_bvec_fn) { + bvm->bi_bdev = rdev->bdev; + bvm->bi_sector = sector_offset + zone->dev_start + + rdev->data_offset; + return min(max, subq->merge_bvec_fn(subq, bvm, biovec)); + } else return max; } -static int raid0_run (mddev_t *mddev) +static sector_t raid0_size(struct mddev *mddev, sector_t sectors, int raid_disks) { - unsigned cur=0, i=0, nb_zone; - s64 size; - raid0_conf_t *conf; - mdk_rdev_t *rdev; - struct list_head *tmp; - - if (mddev->chunk_size == 0) { - printk(KERN_ERR "md/raid0: non-zero chunk size required.\n"); - return -EINVAL; - } - printk(KERN_INFO "%s: setting max_sectors to %d, segment boundary to %d\n", - mdname(mddev), - mddev->chunk_size >> 9, - (mddev->chunk_size>>1)-1); - blk_queue_max_sectors(mddev->queue, mddev->chunk_size >> 9); - blk_queue_segment_boundary(mddev->queue, (mddev->chunk_size>>1) - 1); + sector_t array_sectors = 0; + struct md_rdev *rdev; - conf = kmalloc(sizeof (raid0_conf_t), GFP_KERNEL); - if (!conf) - goto out; - mddev->private = (void *)conf; - - conf->strip_zone = NULL; - conf->devlist = NULL; - if (create_strip_zones (mddev)) - goto out_free_conf; + WARN_ONCE(sectors || raid_disks, + "%s does not support generic reshape\n", __func__); - /* calculate array device size */ - mddev->array_size = 0; - rdev_for_each(rdev, tmp, mddev) - mddev->array_size += rdev->size; - - printk("raid0 : md_size is %llu blocks.\n", - (unsigned long long)mddev->array_size); - printk("raid0 : conf->hash_spacing is %llu blocks.\n", - (unsigned long long)conf->hash_spacing); - { - sector_t s = mddev->array_size; - sector_t space = conf->hash_spacing; - int round; - conf->preshift = 0; - if (sizeof(sector_t) > sizeof(u32)) { - /*shift down space and s so that sector_div will work */ - while (space > (sector_t) (~(u32)0)) { - s >>= 1; - space >>= 1; - s += 1; /* force round-up */ - conf->preshift++; - } - } - round = sector_div(s, (u32)space) ? 1 : 0; - nb_zone = s + round; - } - printk("raid0 : nb_zone is %d.\n", nb_zone); - - printk("raid0 : Allocating %Zd bytes for hash.\n", - nb_zone*sizeof(struct strip_zone*)); - conf->hash_table = kmalloc (sizeof (struct strip_zone *)*nb_zone, GFP_KERNEL); - if (!conf->hash_table) - goto out_free_conf; - size = conf->strip_zone[cur].size; - - conf->hash_table[0] = conf->strip_zone + cur; - for (i=1; i< nb_zone; i++) { - while (size <= conf->hash_spacing) { - cur++; - size += conf->strip_zone[cur].size; - } - size -= conf->hash_spacing; - conf->hash_table[i] = conf->strip_zone + cur; + rdev_for_each(rdev, mddev) + array_sectors += (rdev->sectors & + ~(sector_t)(mddev->chunk_sectors-1)); + + return array_sectors; +} + +static int raid0_stop(struct mddev *mddev); + +static int raid0_run(struct mddev *mddev) +{ + struct r0conf *conf; + int ret; + + if (mddev->chunk_sectors == 0) { + printk(KERN_ERR "md/raid0:%s: chunk size must be set.\n", + mdname(mddev)); + return -EINVAL; } - if (conf->preshift) { - conf->hash_spacing >>= conf->preshift; - /* round hash_spacing up so when we divide by it, we - * err on the side of too-low, which is safest - */ - conf->hash_spacing++; + if (md_check_no_bitmap(mddev)) + return -EINVAL; + blk_queue_max_hw_sectors(mddev->queue, mddev->chunk_sectors); + blk_queue_max_write_same_sectors(mddev->queue, mddev->chunk_sectors); + blk_queue_max_discard_sectors(mddev->queue, mddev->chunk_sectors); + + /* if private is not null, we are here after takeover */ + if (mddev->private == NULL) { + ret = create_strip_zones(mddev, &conf); + if (ret < 0) + return ret; + mddev->private = conf; } + conf = mddev->private; + + /* calculate array device size */ + md_set_array_sectors(mddev, raid0_size(mddev, 0, 0)); + printk(KERN_INFO "md/raid0:%s: md_size is %llu sectors.\n", + mdname(mddev), + (unsigned long long)mddev->array_sectors); /* calculate the max read-ahead size. * For read-ahead of large files to be effective, we need to * readahead at least twice a whole stripe. i.e. number of devices @@ -353,154 +466,253 @@ static int raid0_run (mddev_t *mddev) * chunksize should be used in that case. */ { - int stripe = mddev->raid_disks * mddev->chunk_size / PAGE_SIZE; + int stripe = mddev->raid_disks * + (mddev->chunk_sectors << 9) / PAGE_SIZE; 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, raid0_mergeable_bvec); - return 0; + dump_zones(mddev); -out_free_conf: - kfree(conf->strip_zone); - kfree(conf->devlist); - kfree(conf); - mddev->private = NULL; -out: - return -ENOMEM; + ret = md_integrity_register(mddev); + if (ret) + raid0_stop(mddev); + + return ret; } -static int raid0_stop (mddev_t *mddev) +static int raid0_stop(struct mddev *mddev) { - raid0_conf_t *conf = mddev_to_conf(mddev); + struct r0conf *conf = mddev->private; blk_sync_queue(mddev->queue); /* the unplug fn references 'conf'*/ - kfree(conf->hash_table); - conf->hash_table = NULL; kfree(conf->strip_zone); - conf->strip_zone = NULL; + kfree(conf->devlist); kfree(conf); mddev->private = NULL; - return 0; } -static int raid0_make_request (struct request_queue *q, struct bio *bio) +/* + * Is io distribute over 1 or more chunks ? +*/ +static inline int is_io_in_chunk_boundary(struct mddev *mddev, + unsigned int chunk_sects, struct bio *bio) +{ + if (likely(is_power_of_2(chunk_sects))) { + return chunk_sects >= + ((bio->bi_iter.bi_sector & (chunk_sects-1)) + + bio_sectors(bio)); + } else{ + sector_t sector = bio->bi_iter.bi_sector; + return chunk_sects >= (sector_div(sector, chunk_sects) + + bio_sectors(bio)); + } +} + +static void raid0_make_request(struct mddev *mddev, struct bio *bio) { - mddev_t *mddev = q->queuedata; - unsigned int sect_in_chunk, chunksize_bits, chunk_size, chunk_sects; - raid0_conf_t *conf = mddev_to_conf(mddev); struct strip_zone *zone; - mdk_rdev_t *tmp_dev; - sector_t chunk; - sector_t block, rsect; - const int rw = bio_data_dir(bio); + struct md_rdev *tmp_dev; + struct bio *split; - if (unlikely(bio_barrier(bio))) { - bio_endio(bio, -EOPNOTSUPP); - return 0; + if (unlikely(bio->bi_rw & REQ_FLUSH)) { + md_flush_request(mddev, bio); + return; } - disk_stat_inc(mddev->gendisk, ios[rw]); - disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bio)); - - chunk_size = mddev->chunk_size >> 10; - chunk_sects = mddev->chunk_size >> 9; - chunksize_bits = ffz(~chunk_size); - block = bio->bi_sector >> 1; - - - if (unlikely(chunk_sects < (bio->bi_sector & (chunk_sects - 1)) + (bio->bi_size >> 9))) { - 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, bio_split_pool, chunk_sects - (bio->bi_sector & (chunk_sects - 1)) ); - if (raid0_make_request(q, &bp->bio1)) - generic_make_request(&bp->bio1); - if (raid0_make_request(q, &bp->bio2)) - generic_make_request(&bp->bio2); - - bio_pair_release(bp); - return 0; + do { + sector_t sector = bio->bi_iter.bi_sector; + unsigned chunk_sects = mddev->chunk_sectors; + + unsigned sectors = chunk_sects - + (likely(is_power_of_2(chunk_sects)) + ? (sector & (chunk_sects-1)) + : sector_div(sector, chunk_sects)); + + if (sectors < bio_sectors(bio)) { + split = bio_split(bio, sectors, GFP_NOIO, fs_bio_set); + bio_chain(split, bio); + } else { + split = bio; + } + + zone = find_zone(mddev->private, §or); + tmp_dev = map_sector(mddev, zone, sector, §or); + split->bi_bdev = tmp_dev->bdev; + split->bi_iter.bi_sector = sector + zone->dev_start + + tmp_dev->data_offset; + + if (unlikely((split->bi_rw & REQ_DISCARD) && + !blk_queue_discard(bdev_get_queue(split->bi_bdev)))) { + /* Just ignore it */ + bio_endio(split, 0); + } else + generic_make_request(split); + } while (split != bio); +} + +static void raid0_status(struct seq_file *seq, struct mddev *mddev) +{ + seq_printf(seq, " %dk chunks", mddev->chunk_sectors / 2); + return; +} + +static void *raid0_takeover_raid45(struct mddev *mddev) +{ + struct md_rdev *rdev; + struct r0conf *priv_conf; + + if (mddev->degraded != 1) { + printk(KERN_ERR "md/raid0:%s: raid5 must be degraded! Degraded disks: %d\n", + mdname(mddev), + mddev->degraded); + return ERR_PTR(-EINVAL); } - - { - sector_t x = block >> conf->preshift; - sector_div(x, (u32)conf->hash_spacing); - zone = conf->hash_table[x]; + rdev_for_each(rdev, mddev) { + /* check slot number for a disk */ + if (rdev->raid_disk == mddev->raid_disks-1) { + printk(KERN_ERR "md/raid0:%s: raid5 must have missing parity disk!\n", + mdname(mddev)); + return ERR_PTR(-EINVAL); + } + rdev->sectors = mddev->dev_sectors; } - - while (block >= (zone->zone_offset + zone->size)) - zone++; - - sect_in_chunk = bio->bi_sector & ((chunk_size<<1) -1); + /* Set new parameters */ + mddev->new_level = 0; + mddev->new_layout = 0; + mddev->new_chunk_sectors = mddev->chunk_sectors; + mddev->raid_disks--; + mddev->delta_disks = -1; + /* make sure it will be not marked as dirty */ + mddev->recovery_cp = MaxSector; + + create_strip_zones(mddev, &priv_conf); + return priv_conf; +} - { - sector_t x = (block - zone->zone_offset) >> chunksize_bits; +static void *raid0_takeover_raid10(struct mddev *mddev) +{ + struct r0conf *priv_conf; + + /* Check layout: + * - far_copies must be 1 + * - near_copies must be 2 + * - disks number must be even + * - all mirrors must be already degraded + */ + if (mddev->layout != ((1 << 8) + 2)) { + printk(KERN_ERR "md/raid0:%s:: Raid0 cannot takover layout: 0x%x\n", + mdname(mddev), + mddev->layout); + return ERR_PTR(-EINVAL); + } + if (mddev->raid_disks & 1) { + printk(KERN_ERR "md/raid0:%s: Raid0 cannot takover Raid10 with odd disk number.\n", + mdname(mddev)); + return ERR_PTR(-EINVAL); + } + if (mddev->degraded != (mddev->raid_disks>>1)) { + printk(KERN_ERR "md/raid0:%s: All mirrors must be already degraded!\n", + mdname(mddev)); + return ERR_PTR(-EINVAL); + } + + /* Set new parameters */ + mddev->new_level = 0; + mddev->new_layout = 0; + mddev->new_chunk_sectors = mddev->chunk_sectors; + mddev->delta_disks = - mddev->raid_disks / 2; + mddev->raid_disks += mddev->delta_disks; + mddev->degraded = 0; + /* make sure it will be not marked as dirty */ + mddev->recovery_cp = MaxSector; + + create_strip_zones(mddev, &priv_conf); + return priv_conf; +} - sector_div(x, zone->nb_dev); - chunk = x; +static void *raid0_takeover_raid1(struct mddev *mddev) +{ + struct r0conf *priv_conf; + int chunksect; - x = block >> chunksize_bits; - tmp_dev = zone->dev[sector_div(x, zone->nb_dev)]; + /* Check layout: + * - (N - 1) mirror drives must be already faulty + */ + if ((mddev->raid_disks - 1) != mddev->degraded) { + printk(KERN_ERR "md/raid0:%s: (N - 1) mirrors drives must be already faulty!\n", + mdname(mddev)); + return ERR_PTR(-EINVAL); } - rsect = (((chunk << chunksize_bits) + zone->dev_offset)<<1) - + sect_in_chunk; - - bio->bi_bdev = tmp_dev->bdev; - bio->bi_sector = rsect + tmp_dev->data_offset; /* - * Let the main block layer submit the IO and resolve recursion: + * a raid1 doesn't have the notion of chunk size, so + * figure out the largest suitable size we can use. */ - return 1; - -bad_map: - printk("raid0_make_request bug: can't convert block across chunks" - " or bigger than %dk %llu %d\n", chunk_size, - (unsigned long long)bio->bi_sector, bio->bi_size >> 10); - - bio_io_error(bio); - return 0; + chunksect = 64 * 2; /* 64K by default */ + + /* The array must be an exact multiple of chunksize */ + while (chunksect && (mddev->array_sectors & (chunksect - 1))) + chunksect >>= 1; + + if ((chunksect << 9) < PAGE_SIZE) + /* array size does not allow a suitable chunk size */ + return ERR_PTR(-EINVAL); + + /* Set new parameters */ + mddev->new_level = 0; + mddev->new_layout = 0; + mddev->new_chunk_sectors = chunksect; + mddev->chunk_sectors = chunksect; + mddev->delta_disks = 1 - mddev->raid_disks; + mddev->raid_disks = 1; + /* make sure it will be not marked as dirty */ + mddev->recovery_cp = MaxSector; + + create_strip_zones(mddev, &priv_conf); + return priv_conf; } -static void raid0_status (struct seq_file *seq, mddev_t *mddev) +static void *raid0_takeover(struct mddev *mddev) { -#undef MD_DEBUG -#ifdef MD_DEBUG - int j, k, h; - char b[BDEVNAME_SIZE]; - raid0_conf_t *conf = mddev_to_conf(mddev); + /* raid0 can take over: + * raid4 - if all data disks are active. + * raid5 - providing it is Raid4 layout and one disk is faulty + * raid10 - assuming we have all necessary active disks + * raid1 - with (N -1) mirror drives faulty + */ + if (mddev->level == 4) + return raid0_takeover_raid45(mddev); - h = 0; - for (j = 0; j < conf->nr_strip_zones; j++) { - seq_printf(seq, " z%d", j); - if (conf->hash_table[h] == conf->strip_zone+j) - seq_printf(seq, "(h%d)", h++); - seq_printf(seq, "=["); - for (k = 0; k < conf->strip_zone[j].nb_dev; k++) - seq_printf(seq, "%s/", bdevname( - conf->strip_zone[j].dev[k]->bdev,b)); + if (mddev->level == 5) { + if (mddev->layout == ALGORITHM_PARITY_N) + return raid0_takeover_raid45(mddev); - seq_printf(seq, "] zo=%d do=%d s=%d\n", - conf->strip_zone[j].zone_offset, - conf->strip_zone[j].dev_offset, - conf->strip_zone[j].size); + printk(KERN_ERR "md/raid0:%s: Raid can only takeover Raid5 with layout: %d\n", + mdname(mddev), ALGORITHM_PARITY_N); } -#endif - seq_printf(seq, " %dk chunks", mddev->chunk_size/1024); - return; + + if (mddev->level == 10) + return raid0_takeover_raid10(mddev); + + if (mddev->level == 1) + return raid0_takeover_raid1(mddev); + + printk(KERN_ERR "Takeover from raid%i to raid0 not supported\n", + mddev->level); + + return ERR_PTR(-EINVAL); +} + +static void raid0_quiesce(struct mddev *mddev, int state) +{ } -static struct mdk_personality raid0_personality= +static struct md_personality raid0_personality= { .name = "raid0", .level = 0, @@ -509,6 +721,9 @@ static struct mdk_personality raid0_personality= .run = raid0_run, .stop = raid0_stop, .status = raid0_status, + .size = raid0_size, + .takeover = raid0_takeover, + .quiesce = raid0_quiesce, }; static int __init raid0_init (void) @@ -524,6 +739,7 @@ static void raid0_exit (void) module_init(raid0_init); module_exit(raid0_exit); MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("RAID0 (striping) personality for MD"); MODULE_ALIAS("md-personality-2"); /* RAID0 */ MODULE_ALIAS("md-raid0"); MODULE_ALIAS("md-level-0"); |