Merge branch 'for-linus' of git://neil.brown.name/md

* 'for-linus' of git://neil.brown.name/md: (39 commits)
  md/raid5: correctly update sync_completed when we reach max_resync
  md/raid5: add missing call to schedule() after prepare_to_wait()
  md/linear: use call_rcu to free obsolete 'conf' structures.
  md linear: Protecting mddev with rcu locks to avoid races
  md: Move check for bitmap presence to personality code.
  md: remove chunksize rounding from common code.
  md: raid0/linear: ensure device sizes are rounded to chunk size.
  md: move assignment of ->utime so that it never gets skipped.
  md: Push down reconstruction log message to personality code.
  md: merge reconfig and check_reshape methods.
  md: remove unnecessary arguments from ->reconfig method.
  md: raid5: check stripe cache is large enough in start_reshape
  md: raid0: chunk_sectors cleanups.
  md: fix some comments.
  md/raid5: Use is_power_of_2() in raid5_reconfig()/raid6_reconfig().
  md: convert conf->chunk_size and conf->prev_chunk to sectors.
  md: Convert mddev->new_chunk to sectors.
  md: Make mddev->chunk_size sector-based.
  md: raid0 :Enables chunk size other than powers of 2.
  md: prepare for non-power-of-two chunk sizes
  ...

1 files changed, 92 insertions, 126 deletions

diff --git a/drivers/md/linear.c b/drivers/md/linear.c
index 64f1f3e046e0..15c8b7b25a9b 100644
--- a/drivers/md/linear.c
+++ b/drivers/md/linear.c
@@ -27,19 +27,27 @@
  */
 static inline dev_info_t *which_dev(mddev_t *mddev, sector_t sector)
 {
-	dev_info_t *hash;
-	linear_conf_t *conf = mddev_to_conf(mddev);
-	sector_t idx = sector >> conf->sector_shift;
+	int lo, mid, hi;
+	linear_conf_t *conf;
+
+	lo = 0;
+	hi = mddev->raid_disks - 1;
+	conf = rcu_dereference(mddev->private);
 
 	/*
-	 * sector_div(a,b) returns the remainer and sets a to a/b
+	 * Binary Search
 	 */
-	(void)sector_div(idx, conf->spacing);
-	hash = conf->hash_table[idx];
 
-	while (sector >= hash->num_sectors + hash->start_sector)
-		hash++;
-	return hash;
+	while (hi > lo) {
+
+		mid = (hi + lo) / 2;
+		if (sector < conf->disks[mid].end_sector)
+			hi = mid;
+		else
+			lo = mid + 1;
+	}
+
+	return conf->disks + lo;
 }
 
 /**
@@ -59,8 +67,10 @@ static int linear_mergeable_bvec(struct request_queue *q,
 	unsigned long maxsectors, bio_sectors = bvm->bi_size >> 9;
 	sector_t sector = bvm->bi_sector + get_start_sect(bvm->bi_bdev);
 
+	rcu_read_lock();
 	dev0 = which_dev(mddev, sector);
-	maxsectors = dev0->num_sectors - (sector - dev0->start_sector);
+	maxsectors = dev0->end_sector - sector;
+	rcu_read_unlock();
 
 	if (maxsectors < bio_sectors)
 		maxsectors = 0;
@@ -79,46 +89,57 @@ static int linear_mergeable_bvec(struct request_queue *q,
 static void linear_unplug(struct request_queue *q)
 {
 	mddev_t *mddev = q->queuedata;
-	linear_conf_t *conf = mddev_to_conf(mddev);
+	linear_conf_t *conf;
 	int i;
 
+	rcu_read_lock();
+	conf = rcu_dereference(mddev->private);
+
 	for (i=0; i < mddev->raid_disks; i++) {
 		struct request_queue *r_queue = bdev_get_queue(conf->disks[i].rdev->bdev);
 		blk_unplug(r_queue);
 	}
+	rcu_read_unlock();
 }
 
 static int linear_congested(void *data, int bits)
 {
 	mddev_t *mddev = data;
-	linear_conf_t *conf = mddev_to_conf(mddev);
+	linear_conf_t *conf;
 	int i, ret = 0;
 
+	rcu_read_lock();
+	conf = rcu_dereference(mddev->private);
+
 	for (i = 0; i < mddev->raid_disks && !ret ; i++) {
 		struct request_queue *q = bdev_get_queue(conf->disks[i].rdev->bdev);
 		ret |= bdi_congested(&q->backing_dev_info, bits);
 	}
+
+	rcu_read_unlock();
 	return ret;
 }
 
 static sector_t linear_size(mddev_t *mddev, sector_t sectors, int raid_disks)
 {
-	linear_conf_t *conf = mddev_to_conf(mddev);
+	linear_conf_t *conf;
+	sector_t array_sectors;
 
+	rcu_read_lock();
+	conf = rcu_dereference(mddev->private);
 	WARN_ONCE(sectors || raid_disks,
 		  "%s does not support generic reshape\n", __func__);
+	array_sectors = conf->array_sectors;
+	rcu_read_unlock();
 
-	return conf->array_sectors;
+	return array_sectors;
 }
 
 static linear_conf_t *linear_conf(mddev_t *mddev, int raid_disks)
 {
 	linear_conf_t *conf;
-	dev_info_t **table;
 	mdk_rdev_t *rdev;
-	int i, nb_zone, cnt;
-	sector_t min_sectors;
-	sector_t curr_sector;
+	int i, cnt;
 
 	conf = kzalloc (sizeof (*conf) + raid_disks*sizeof(dev_info_t),
 			GFP_KERNEL);
@@ -131,6 +152,7 @@ static linear_conf_t *linear_conf(mddev_t *mddev, int raid_disks)
 	list_for_each_entry(rdev, &mddev->disks, same_set) {
 		int j = rdev->raid_disk;
 		dev_info_t *disk = conf->disks + j;
+		sector_t sectors;
 
 		if (j < 0 || j >= raid_disks || disk->rdev) {
 			printk("linear: disk numbering problem. Aborting!\n");
@@ -138,6 +160,11 @@ static linear_conf_t *linear_conf(mddev_t *mddev, int raid_disks)
 		}
 
 		disk->rdev = rdev;
+		if (mddev->chunk_sectors) {
+			sectors = rdev->sectors;
+			sector_div(sectors, mddev->chunk_sectors);
+			rdev->sectors = sectors * mddev->chunk_sectors;
+		}
 
 		blk_queue_stack_limits(mddev->queue,
 				       rdev->bdev->bd_disk->queue);
@@ -149,102 +176,24 @@ static linear_conf_t *linear_conf(mddev_t *mddev, int raid_disks)
 		    queue_max_sectors(mddev->queue) > (PAGE_SIZE>>9))
 			blk_queue_max_sectors(mddev->queue, PAGE_SIZE>>9);
 
-		disk->num_sectors = rdev->sectors;
 		conf->array_sectors += rdev->sectors;
-
 		cnt++;
+
 	}
 	if (cnt != raid_disks) {
 		printk("linear: not enough drives present. Aborting!\n");
 		goto out;
 	}
 
-	min_sectors = conf->array_sectors;
-	sector_div(min_sectors, PAGE_SIZE/sizeof(struct dev_info *));
-	if (min_sectors == 0)
-		min_sectors = 1;
-
-	/* min_sectors is the minimum spacing that will fit the hash
-	 * table in one PAGE.  This may be much smaller than needed.
-	 * We find the smallest non-terminal set of consecutive devices
-	 * that is larger than min_sectors and use the size of that as
-	 * the actual spacing
-	 */
-	conf->spacing = conf->array_sectors;
-	for (i=0; i < cnt-1 ; i++) {
-		sector_t tmp = 0;
-		int j;
-		for (j = i; j < cnt - 1 && tmp < min_sectors; j++)
-			tmp += conf->disks[j].num_sectors;
-		if (tmp >= min_sectors && tmp < conf->spacing)
-			conf->spacing = tmp;
-	}
-
-	/* spacing may be too large for sector_div to work with,
-	 * so we might need to pre-shift
-	 */
-	conf->sector_shift = 0;
-	if (sizeof(sector_t) > sizeof(u32)) {
-		sector_t space = conf->spacing;
-		while (space > (sector_t)(~(u32)0)) {
-			space >>= 1;
-			conf->sector_shift++;
-		}
-	}
 	/*
-	 * This code was restructured to work around a gcc-2.95.3 internal
-	 * compiler error.  Alter it with care.
+	 * Here we calculate the device offsets.
 	 */
-	{
-		sector_t sz;
-		unsigned round;
-		unsigned long base;
-
-		sz = conf->array_sectors >> conf->sector_shift;
-		sz += 1; /* force round-up */
-		base = conf->spacing >> conf->sector_shift;
-		round = sector_div(sz, base);
-		nb_zone = sz + (round ? 1 : 0);
-	}
-	BUG_ON(nb_zone > PAGE_SIZE / sizeof(struct dev_info *));
-
-	conf->hash_table = kmalloc (sizeof (struct dev_info *) * nb_zone,
-					GFP_KERNEL);
-	if (!conf->hash_table)
-		goto out;
+	conf->disks[0].end_sector = conf->disks[0].rdev->sectors;
 
-	/*
-	 * Here we generate the linear hash table
-	 * First calculate the device offsets.
-	 */
-	conf->disks[0].start_sector = 0;
 	for (i = 1; i < raid_disks; i++)
-		conf->disks[i].start_sector =
-			conf->disks[i-1].start_sector +
-			conf->disks[i-1].num_sectors;
-
-	table = conf->hash_table;
-	i = 0;
-	for (curr_sector = 0;
-	     curr_sector < conf->array_sectors;
-	     curr_sector += conf->spacing) {
-
-		while (i < raid_disks-1 &&
-		       curr_sector >= conf->disks[i+1].start_sector)
-			i++;
-
-		*table ++ = conf->disks + i;
-	}
-
-	if (conf->sector_shift) {
-		conf->spacing >>= conf->sector_shift;
-		/* round spacing up so that when we divide by it,
-		 * we err on the side of "too-low", which is safest.
-		 */
-		conf->spacing++;
-	}
-
-	BUG_ON(table - conf->hash_table > nb_zone);
+		conf->disks[i].end_sector =
+			conf->disks[i-1].end_sector +
+			conf->disks[i].rdev->sectors;
 
 	return conf;
 
@@ -257,6 +206,8 @@ static int linear_run (mddev_t *mddev)
 {
 	linear_conf_t *conf;
 
+	if (md_check_no_bitmap(mddev))
+		return -EINVAL;
 	mddev->queue->queue_lock = &mddev->queue->__queue_lock;
 	conf = linear_conf(mddev, mddev->raid_disks);
 
@@ -272,6 +223,12 @@ static int linear_run (mddev_t *mddev)
 	return 0;
 }
 
+static void free_conf(struct rcu_head *head)
+{
+	linear_conf_t *conf = container_of(head, linear_conf_t, rcu);
+	kfree(conf);
+}
+
 static int linear_add(mddev_t *mddev, mdk_rdev_t *rdev)
 {
 	/* Adding a drive to a linear array allows the array to grow.
@@ -282,7 +239,7 @@ static int linear_add(mddev_t *mddev, mdk_rdev_t *rdev)
 	 * The current one is never freed until the array is stopped.
 	 * This avoids races.
 	 */
-	linear_conf_t *newconf;
+	linear_conf_t *newconf, *oldconf;
 
 	if (rdev->saved_raid_disk != mddev->raid_disks)
 		return -EINVAL;
@@ -294,25 +251,29 @@ static int linear_add(mddev_t *mddev, mdk_rdev_t *rdev)
 	if (!newconf)
 		return -ENOMEM;
 
-	newconf->prev = mddev_to_conf(mddev);
-	mddev->private = newconf;
+	oldconf = rcu_dereference(mddev->private);
 	mddev->raid_disks++;
+	rcu_assign_pointer(mddev->private, newconf);
 	md_set_array_sectors(mddev, linear_size(mddev, 0, 0));
 	set_capacity(mddev->gendisk, mddev->array_sectors);
+	call_rcu(&oldconf->rcu, free_conf);
 	return 0;
 }
 
 static int linear_stop (mddev_t *mddev)
 {
-	linear_conf_t *conf = mddev_to_conf(mddev);
-  
+	linear_conf_t *conf = mddev->private;
+
+	/*
+	 * We do not require rcu protection here since
+	 * we hold reconfig_mutex for both linear_add and
+	 * linear_stop, so they cannot race.
+	 * We should make sure any old 'conf's are properly
+	 * freed though.
+	 */
+	rcu_barrier();
 	blk_sync_queue(mddev->queue); /* the unplug fn references 'conf'*/
-	do {
-		linear_conf_t *t = conf->prev;
-		kfree(conf->hash_table);
-		kfree(conf);
-		conf = t;
-	} while (conf);
+	kfree(conf);
 
 	return 0;
 }
@@ -322,6 +283,7 @@ static int linear_make_request (struct request_queue *q, struct bio *bio)
 	const int rw = bio_data_dir(bio);
 	mddev_t *mddev = q->queuedata;
 	dev_info_t *tmp_dev;
+	sector_t start_sector;
 	int cpu;
 
 	if (unlikely(bio_barrier(bio))) {
@@ -335,33 +297,36 @@ static int linear_make_request (struct request_queue *q, struct bio *bio)
 		      bio_sectors(bio));
 	part_stat_unlock();
 
+	rcu_read_lock();
 	tmp_dev = which_dev(mddev, bio->bi_sector);
-    
-	if (unlikely(bio->bi_sector >= (tmp_dev->num_sectors +
-					tmp_dev->start_sector)
-		     || (bio->bi_sector <
-			 tmp_dev->start_sector))) {
+	start_sector = tmp_dev->end_sector - tmp_dev->rdev->sectors;
+
+
+	if (unlikely(bio->bi_sector >= (tmp_dev->end_sector)
+		     || (bio->bi_sector < start_sector))) {
 		char b[BDEVNAME_SIZE];
 
 		printk("linear_make_request: Sector %llu out of bounds on "
 			"dev %s: %llu sectors, offset %llu\n",
 			(unsigned long long)bio->bi_sector,
 			bdevname(tmp_dev->rdev->bdev, b),
-			(unsigned long long)tmp_dev->num_sectors,
-			(unsigned long long)tmp_dev->start_sector);
+			(unsigned long long)tmp_dev->rdev->sectors,
+			(unsigned long long)start_sector);
+		rcu_read_unlock();
 		bio_io_error(bio);
 		return 0;
 	}
 	if (unlikely(bio->bi_sector + (bio->bi_size >> 9) >
-		     tmp_dev->start_sector + tmp_dev->num_sectors)) {
+		     tmp_dev->end_sector)) {
 		/* This bio crosses a device boundary, so we have to
 		 * split it.
 		 */
 		struct bio_pair *bp;
+		sector_t end_sector = tmp_dev->end_sector;
+
+		rcu_read_unlock();
 
-		bp = bio_split(bio,
-			       tmp_dev->start_sector + tmp_dev->num_sectors
-			       - bio->bi_sector);
+		bp = bio_split(bio, end_sector - bio->bi_sector);
 
 		if (linear_make_request(q, &bp->bio1))
 			generic_make_request(&bp->bio1);
@@ -372,8 +337,9 @@ static int linear_make_request (struct request_queue *q, struct bio *bio)
 	}
 		    
 	bio->bi_bdev = tmp_dev->rdev->bdev;
-	bio->bi_sector = bio->bi_sector - tmp_dev->start_sector
+	bio->bi_sector = bio->bi_sector - start_sector
 		+ tmp_dev->rdev->data_offset;
+	rcu_read_unlock();
 
 	return 1;
 }
@@ -381,7 +347,7 @@ static int linear_make_request (struct request_queue *q, struct bio *bio)
 static void linear_status (struct seq_file *seq, mddev_t *mddev)
 {
 
-	seq_printf(seq, " %dk rounding", mddev->chunk_size/1024);
+	seq_printf(seq, " %dk rounding", mddev->chunk_sectors / 2);
 }