Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
*/
-#include <linux/module.h>
-#include <linux/raid/raid0.h>
+#include <linux/blkdev.h>
+#include <linux/seq_file.h>
+#include "md.h"
+#include "raid0.h"
-#define MAJOR_NR MD_MAJOR
-#define MD_DRIVER
-#define MD_PERSONALITY
-
-static void raid0_unplug(request_queue_t *q)
+static void raid0_unplug(struct request_queue *q)
{
mddev_t *mddev = q->queuedata;
raid0_conf_t *conf = mddev_to_conf(mddev);
int i;
for (i=0; i<mddev->raid_disks; i++) {
- request_queue_t *r_queue = bdev_get_queue(devlist[i]->bdev);
+ struct request_queue *r_queue = bdev_get_queue(devlist[i]->bdev);
- if (r_queue->unplug_fn)
- r_queue->unplug_fn(r_queue);
+ blk_unplug(r_queue);
}
}
-static int raid0_issue_flush(request_queue_t *q, struct gendisk *disk,
- sector_t *error_sector)
+static int raid0_congested(void *data, int bits)
{
- mddev_t *mddev = q->queuedata;
+ mddev_t *mddev = data;
raid0_conf_t *conf = mddev_to_conf(mddev);
mdk_rdev_t **devlist = conf->strip_zone[0].dev;
int i, ret = 0;
- for (i=0; i<mddev->raid_disks && ret == 0; i++) {
- struct block_device *bdev = devlist[i]->bdev;
- request_queue_t *r_queue = bdev_get_queue(bdev);
+ for (i = 0; i < mddev->raid_disks && !ret ; i++) {
+ struct request_queue *q = bdev_get_queue(devlist[i]->bdev);
- if (!r_queue->issue_flush_fn)
- ret = -EOPNOTSUPP;
- else
- ret = r_queue->issue_flush_fn(r_queue, bdev->bd_disk, error_sector);
+ ret |= bdi_congested(&q->backing_dev_info, bits);
}
return ret;
}
static int create_strip_zones (mddev_t *mddev)
{
int i, c, j;
- sector_t current_offset, curr_zone_offset;
+ sector_t current_start, curr_zone_start;
sector_t min_spacing;
raid0_conf_t *conf = mddev_to_conf(mddev);
mdk_rdev_t *smallest, *rdev1, *rdev2, *rdev;
- struct list_head *tmp1, *tmp2;
struct strip_zone *zone;
int cnt;
char b[BDEVNAME_SIZE];
*/
conf->nr_strip_zones = 0;
- ITERATE_RDEV(mddev,rdev1,tmp1) {
- printk("raid0: looking at %s\n",
+ list_for_each_entry(rdev1, &mddev->disks, same_set) {
+ printk(KERN_INFO "raid0: looking at %s\n",
bdevname(rdev1->bdev,b));
c = 0;
- ITERATE_RDEV(mddev,rdev2,tmp2) {
- printk("raid0: comparing %s(%llu)",
+ list_for_each_entry(rdev2, &mddev->disks, same_set) {
+ printk(KERN_INFO "raid0: comparing %s(%llu)",
bdevname(rdev1->bdev,b),
- (unsigned long long)rdev1->size);
- printk(" with %s(%llu)\n",
+ (unsigned long long)rdev1->sectors);
+ printk(KERN_INFO " with %s(%llu)\n",
bdevname(rdev2->bdev,b),
- (unsigned long long)rdev2->size);
+ (unsigned long long)rdev2->sectors);
if (rdev2 == rdev1) {
- printk("raid0: END\n");
+ printk(KERN_INFO "raid0: END\n");
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");
+ printk(KERN_INFO "raid0: EQUAL\n");
c = 1;
break;
}
- printk("raid0: NOT EQUAL\n");
+ printk(KERN_INFO "raid0: NOT EQUAL\n");
}
if (!c) {
- printk("raid0: ==> UNIQUE\n");
+ printk(KERN_INFO "raid0: ==> UNIQUE\n");
conf->nr_strip_zones++;
- printk("raid0: %d zones\n", conf->nr_strip_zones);
+ printk(KERN_INFO "raid0: %d zones\n",
+ conf->nr_strip_zones);
}
}
- printk("raid0: FINAL %d zones\n", conf->nr_strip_zones);
+ printk(KERN_INFO "raid0: FINAL %d zones\n", conf->nr_strip_zones);
conf->strip_zone = kzalloc(sizeof(struct strip_zone)*
conf->nr_strip_zones, GFP_KERNEL);
cnt = 0;
smallest = NULL;
zone->dev = conf->devlist;
- ITERATE_RDEV(mddev, rdev1, tmp1) {
+ list_for_each_entry(rdev1, &mddev->disks, same_set) {
int j = rdev1->raid_disk;
if (j < 0 || j >= mddev->raid_disks) {
- printk("raid0: bad disk number %d - aborting!\n", j);
+ printk(KERN_ERR "raid0: bad disk number %d - "
+ "aborting!\n", j);
goto abort;
}
if (zone->dev[j]) {
- printk("raid0: multiple devices for %d - aborting!\n",
- j);
+ printk(KERN_ERR "raid0: multiple devices for %d - "
+ "aborting!\n", j);
goto abort;
}
zone->dev[j] = rdev1;
*/
if (rdev1->bdev->bd_disk->queue->merge_bvec_fn &&
- mddev->queue->max_sectors > (PAGE_SIZE>>9))
+ queue_max_sectors(mddev->queue) > (PAGE_SIZE>>9))
blk_queue_max_sectors(mddev->queue, PAGE_SIZE>>9);
- if (!smallest || (rdev1->size <smallest->size))
+ if (!smallest || (rdev1->sectors < smallest->sectors))
smallest = rdev1;
cnt++;
}
if (cnt != mddev->raid_disks) {
- printk("raid0: too few disks (%d of %d) - aborting!\n",
- cnt, mddev->raid_disks);
+ printk(KERN_ERR "raid0: too few disks (%d of %d) - "
+ "aborting!\n", cnt, mddev->raid_disks);
goto abort;
}
zone->nb_dev = cnt;
- zone->size = smallest->size * cnt;
- zone->zone_offset = 0;
+ zone->sectors = smallest->sectors * cnt;
+ zone->zone_start = 0;
- current_offset = smallest->size;
- curr_zone_offset = zone->size;
+ current_start = smallest->sectors;
+ curr_zone_start = zone->sectors;
/* now do the other zones */
for (i = 1; i < conf->nr_strip_zones; i++)
zone = conf->strip_zone + i;
zone->dev = conf->strip_zone[i-1].dev + mddev->raid_disks;
- printk("raid0: zone %d\n", i);
- zone->dev_offset = current_offset;
+ printk(KERN_INFO "raid0: zone %d\n", i);
+ zone->dev_start = current_start;
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");
+ printk(KERN_INFO "raid0: checking %s ...",
+ bdevname(rdev->bdev, b));
+ if (rdev->sectors <= current_start) {
+ printk(KERN_INFO " nope.\n");
+ continue;
+ }
+ printk(KERN_INFO " contained as device %d\n", c);
+ zone->dev[c] = rdev;
+ c++;
+ if (!smallest || rdev->sectors < smallest->sectors) {
+ smallest = rdev;
+ printk(KERN_INFO " (%llu) is smallest!.\n",
+ (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);
+ zone->sectors = (smallest->sectors - current_start) * c;
+ printk(KERN_INFO "raid0: zone->nb_dev: %d, sectors: %llu\n",
+ zone->nb_dev, (unsigned long long)zone->sectors);
- zone->zone_offset = curr_zone_offset;
- curr_zone_offset += zone->size;
+ zone->zone_start = curr_zone_start;
+ curr_zone_start += zone->sectors;
- current_offset = smallest->size;
- printk("raid0: current zone offset: %llu\n",
- (unsigned long long)current_offset);
+ current_start = smallest->sectors;
+ printk(KERN_INFO "raid0: current zone start: %llu\n",
+ (unsigned long long)current_start);
}
/* Now find appropriate hash spacing.
* strip though as it's size has no bearing on the efficacy of the hash
* table.
*/
- conf->hash_spacing = curr_zone_offset;
- min_spacing = curr_zone_offset;
+ conf->spacing = curr_zone_start;
+ min_spacing = curr_zone_start;
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;
+ sector_t s = 0;
+ for (j = i; j < conf->nr_strip_zones - 1 &&
+ s < min_spacing; j++)
+ s += conf->strip_zone[j].sectors;
+ if (s >= min_spacing && s < conf->spacing)
+ conf->spacing = s;
}
mddev->queue->unplug_fn = raid0_unplug;
- mddev->queue->issue_flush_fn = raid0_issue_flush;
+ mddev->queue->backing_dev_info.congested_fn = raid0_congested;
+ mddev->queue->backing_dev_info.congested_data = mddev;
- printk("raid0: done.\n");
+ printk(KERN_INFO "raid0: done.\n");
return 0;
abort:
return 1;
/**
* raid0_mergeable_bvec -- tell bio layer if a 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(request_queue_t *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);
+ sector_t sector = bvm->bi_sector + get_start_sect(bvm->bi_bdev);
int max;
unsigned int chunk_sectors = mddev->chunk_size >> 9;
- unsigned int bio_sectors = bio->bi_size >> 9;
+ 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 */
return max;
}
+static sector_t raid0_size(mddev_t *mddev, sector_t sectors, int raid_disks)
+{
+ sector_t array_sectors = 0;
+ mdk_rdev_t *rdev;
+
+ WARN_ONCE(sectors || raid_disks,
+ "%s does not support generic reshape\n", __func__);
+
+ list_for_each_entry(rdev, &mddev->disks, same_set)
+ array_sectors += rdev->sectors;
+
+ return array_sectors;
+}
+
static int raid0_run (mddev_t *mddev)
{
unsigned cur=0, i=0, nb_zone;
- s64 size;
+ s64 sectors;
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");
(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);
+ mddev->queue->queue_lock = &mddev->queue->__queue_lock;
conf = kmalloc(sizeof (raid0_conf_t), GFP_KERNEL);
if (!conf)
goto out_free_conf;
/* calculate array device size */
- mddev->array_size = 0;
- ITERATE_RDEV(mddev,rdev,tmp)
- 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);
+ md_set_array_sectors(mddev, raid0_size(mddev, 0, 0));
+
+ printk(KERN_INFO "raid0 : md_size is %llu sectors.\n",
+ (unsigned long long)mddev->array_sectors);
+ printk(KERN_INFO "raid0 : conf->spacing is %llu sectors.\n",
+ (unsigned long long)conf->spacing);
{
- sector_t s = mddev->array_size;
- sector_t space = conf->hash_spacing;
+ sector_t s = raid0_size(mddev, 0, 0);
+ sector_t space = conf->spacing;
int round;
- conf->preshift = 0;
+ conf->sector_shift = 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++;
+ conf->sector_shift++;
}
}
round = sector_div(s, (u32)space) ? 1 : 0;
nb_zone = s + round;
}
- printk("raid0 : nb_zone is %d.\n", nb_zone);
+ printk(KERN_INFO "raid0 : nb_zone is %d.\n", nb_zone);
- printk("raid0 : Allocating %Zd bytes for hash.\n",
+ printk(KERN_INFO "raid0 : Allocating %zu 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;
+ sectors = conf->strip_zone[cur].sectors;
- for (i=0; i< nb_zone; i++) {
- conf->hash_table[i] = conf->strip_zone + cur;
- while (size <= conf->hash_spacing) {
+ conf->hash_table[0] = conf->strip_zone + cur;
+ for (i=1; i< nb_zone; i++) {
+ while (sectors <= conf->spacing) {
cur++;
- size += conf->strip_zone[cur].size;
+ sectors += conf->strip_zone[cur].sectors;
}
- size -= conf->hash_spacing;
+ sectors -= conf->spacing;
+ conf->hash_table[i] = conf->strip_zone + cur;
}
- if (conf->preshift) {
- conf->hash_spacing >>= conf->preshift;
- /* round hash_spacing up so when we divide by it, we
+ if (conf->sector_shift) {
+ conf->spacing >>= conf->sector_shift;
+ /* round spacing up so when we divide by it, we
* err on the side of too-low, which is safest
*/
- conf->hash_spacing++;
+ conf->spacing++;
}
/* calculate the max read-ahead size.
return 0;
}
-static int raid0_make_request (request_queue_t *q, struct bio *bio)
+static int raid0_make_request (struct request_queue *q, struct bio *bio)
{
mddev_t *mddev = q->queuedata;
- unsigned int sect_in_chunk, chunksize_bits, chunk_size, chunk_sects;
+ unsigned int sect_in_chunk, chunksect_bits, chunk_sects;
raid0_conf_t *conf = mddev_to_conf(mddev);
struct strip_zone *zone;
mdk_rdev_t *tmp_dev;
- unsigned long chunk;
- sector_t block, rsect;
+ sector_t chunk;
+ sector_t sector, rsect;
const int rw = bio_data_dir(bio);
+ int cpu;
if (unlikely(bio_barrier(bio))) {
- bio_endio(bio, bio->bi_size, -EOPNOTSUPP);
+ bio_endio(bio, -EOPNOTSUPP);
return 0;
}
- disk_stat_inc(mddev->gendisk, ios[rw]);
- disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bio));
+ cpu = part_stat_lock();
+ part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]);
+ part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw],
+ bio_sectors(bio));
+ part_stat_unlock();
- chunk_size = mddev->chunk_size >> 10;
chunk_sects = mddev->chunk_size >> 9;
- chunksize_bits = ffz(~chunk_size);
- block = bio->bi_sector >> 1;
-
+ chunksect_bits = ffz(~chunk_sects);
+ sector = bio->bi_sector;
if (unlikely(chunk_sects < (bio->bi_sector & (chunk_sects - 1)) + (bio->bi_size >> 9))) {
struct bio_pair *bp;
/* 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)) );
+ bp = bio_split(bio, 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))
{
- sector_t x = block >> conf->preshift;
- sector_div(x, (u32)conf->hash_spacing);
+ sector_t x = sector >> conf->sector_shift;
+ sector_div(x, (u32)conf->spacing);
zone = conf->hash_table[x];
}
-
- while (block >= (zone->zone_offset + zone->size))
+
+ while (sector >= zone->zone_start + zone->sectors)
zone++;
-
- sect_in_chunk = bio->bi_sector & ((chunk_size<<1) -1);
+
+ sect_in_chunk = bio->bi_sector & (chunk_sects - 1);
{
- sector_t x = (block - zone->zone_offset) >> chunksize_bits;
+ sector_t x = (sector - zone->zone_start) >> chunksect_bits;
sector_div(x, zone->nb_dev);
chunk = x;
- BUG_ON(x != (sector_t)chunk);
- x = block >> chunksize_bits;
+ x = sector >> chunksect_bits;
tmp_dev = zone->dev[sector_div(x, zone->nb_dev)];
}
- rsect = (((chunk << chunksize_bits) + zone->dev_offset)<<1)
- + sect_in_chunk;
+ rsect = (chunk << chunksect_bits) + zone->dev_start + sect_in_chunk;
bio->bi_bdev = tmp_dev->bdev;
bio->bi_sector = rsect + tmp_dev->data_offset;
bad_map:
printk("raid0_make_request bug: can't convert block across chunks"
- " or bigger than %dk %llu %d\n", chunk_size,
+ " or bigger than %dk %llu %d\n", chunk_sects / 2,
(unsigned long long)bio->bi_sector, bio->bi_size >> 10);
- bio_io_error(bio, bio->bi_size);
+ bio_io_error(bio);
return 0;
}
-
+
static void raid0_status (struct seq_file *seq, mddev_t *mddev)
{
#undef MD_DEBUG
int j, k, h;
char b[BDEVNAME_SIZE];
raid0_conf_t *conf = mddev_to_conf(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("(h%d)", h++);
+ 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(
+ seq_printf(seq, "%s/", bdevname(
conf->strip_zone[j].dev[k]->bdev,b));
- 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);
+ seq_printf(seq, "] zs=%d ds=%d s=%d\n",
+ conf->strip_zone[j].zone_start,
+ conf->strip_zone[j].dev_start,
+ conf->strip_zone[j].sectors);
}
#endif
seq_printf(seq, " %dk chunks", mddev->chunk_size/1024);
.run = raid0_run,
.stop = raid0_stop,
.status = raid0_status,
+ .size = raid0_size,
};
static int __init raid0_init (void)
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff