[mirror_ubuntu-artful-kernel.git] / drivers / md / linear.c

/*
   linear.c : Multiple Devices driver for Linux
	      Copyright (C) 1994-96 Marc ZYNGIER
	      <zyngier@ufr-info-p7.ibp.fr> or
	      <maz@gloups.fdn.fr>

   Linear mode management functions.

   This program is free software; you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation; either version 2, or (at your option)
   any later version.
   
   You should have received a copy of the GNU General Public License
   (for example /usr/src/linux/COPYING); if not, write to the Free
   Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.  
*/

#include <linux/blkdev.h>
#include <linux/raid/md_u.h>
#include <linux/seq_file.h>
#include "md.h"
#include "linear.h"

/*
 * find which device holds a particular offset 
 */
static inline dev_info_t *which_dev(mddev_t *mddev, sector_t sector)
{
	dev_info_t *hash;
	linear_conf_t *conf = mddev->private;

	hash = conf->disks;

	while (sector >= hash->num_sectors + hash->start_sector)
		hash++;
	return hash;
}

/**
 *	linear_mergeable_bvec -- tell bio layer if two requests can be merged
 *	@q: request queue
 *	@bvm: properties of new bio
 *	@biovec: the request that could be merged to it.
 *
 *	Return amount of bytes we can take at this offset
 */
static int linear_mergeable_bvec(struct request_queue *q,
				 struct bvec_merge_data *bvm,
				 struct bio_vec *biovec)
{
	mddev_t *mddev = q->queuedata;
	dev_info_t *dev0;
	unsigned long maxsectors, bio_sectors = bvm->bi_size >> 9;
	sector_t sector = bvm->bi_sector + get_start_sect(bvm->bi_bdev);

	dev0 = which_dev(mddev, sector);
	maxsectors = dev0->num_sectors - (sector - dev0->start_sector);

	if (maxsectors < bio_sectors)
		maxsectors = 0;
	else
		maxsectors -= bio_sectors;

	if (maxsectors <= (PAGE_SIZE >> 9 ) && bio_sectors == 0)
		return biovec->bv_len;
	/* The bytes available at this offset could be really big,
	 * so we cap at 2^31 to avoid overflow */
	if (maxsectors > (1 << (31-9)))
		return 1<<31;
	return maxsectors << 9;
}

static void linear_unplug(struct request_queue *q)
{
	mddev_t *mddev = q->queuedata;
	linear_conf_t *conf = mddev->private;
	int i;

	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);
	}
}

static int linear_congested(void *data, int bits)
{
	mddev_t *mddev = data;
	linear_conf_t *conf = mddev->private;
	int i, ret = 0;

	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);
	}
	return ret;
}

static sector_t linear_size(mddev_t *mddev, sector_t sectors, int raid_disks)
{
	linear_conf_t *conf = mddev->private;

	WARN_ONCE(sectors || raid_disks,
		  "%s does not support generic reshape\n", __func__);

	return conf->array_sectors;
}

static linear_conf_t *linear_conf(mddev_t *mddev, int raid_disks)
{
	linear_conf_t *conf;
	mdk_rdev_t *rdev;
	int i, cnt;

	conf = kzalloc (sizeof (*conf) + raid_disks*sizeof(dev_info_t),
			GFP_KERNEL);
	if (!conf)
		return NULL;

	cnt = 0;
	conf->array_sectors = 0;

	list_for_each_entry(rdev, &mddev->disks, same_set) {
		int j = rdev->raid_disk;
		dev_info_t *disk = conf->disks + j;

		if (j < 0 || j >= raid_disks || disk->rdev) {
			printk("linear: disk numbering problem. Aborting!\n");
			goto out;
		}

		disk->rdev = rdev;

		blk_queue_stack_limits(mddev->queue,
				       rdev->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.
		 */
		if (rdev->bdev->bd_disk->queue->merge_bvec_fn &&
		    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;
	}

	/*
	 * Here we 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;

	return conf;

out:
	kfree(conf);
	return NULL;
}

static int linear_run (mddev_t *mddev)
{
	linear_conf_t *conf;

	mddev->queue->queue_lock = &mddev->queue->__queue_lock;
	conf = linear_conf(mddev, mddev->raid_disks);

	if (!conf)
		return 1;
	mddev->private = conf;
	md_set_array_sectors(mddev, linear_size(mddev, 0, 0));

	blk_queue_merge_bvec(mddev->queue, linear_mergeable_bvec);
	mddev->queue->unplug_fn = linear_unplug;
	mddev->queue->backing_dev_info.congested_fn = linear_congested;
	mddev->queue->backing_dev_info.congested_data = mddev;
	return 0;
}

static int linear_add(mddev_t *mddev, mdk_rdev_t *rdev)
{
	/* Adding a drive to a linear array allows the array to grow.
	 * It is permitted if the new drive has a matching superblock
	 * already on it, with raid_disk equal to raid_disks.
	 * It is achieved by creating a new linear_private_data structure
	 * and swapping it in in-place of the current one.
	 * The current one is never freed until the array is stopped.
	 * This avoids races.
	 */
	linear_conf_t *newconf;

	if (rdev->saved_raid_disk != mddev->raid_disks)
		return -EINVAL;

	rdev->raid_disk = rdev->saved_raid_disk;

	newconf = linear_conf(mddev,mddev->raid_disks+1);

	if (!newconf)
		return -ENOMEM;

	newconf->prev = mddev->private;
	mddev->private = newconf;
	mddev->raid_disks++;
	md_set_array_sectors(mddev, linear_size(mddev, 0, 0));
	set_capacity(mddev->gendisk, mddev->array_sectors);
	return 0;
}

static int linear_stop (mddev_t *mddev)
{
	linear_conf_t *conf = mddev->private;
  
	blk_sync_queue(mddev->queue); /* the unplug fn references 'conf'*/
	do {
		linear_conf_t *t = conf->prev;
		kfree(conf);
		conf = t;
	} while (conf);

	return 0;
}

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;
	int cpu;

	if (unlikely(bio_barrier(bio))) {
		bio_endio(bio, -EOPNOTSUPP);
		return 0;
	}

	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();

	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))) {
		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);
		bio_io_error(bio);
		return 0;
	}
	if (unlikely(bio->bi_sector + (bio->bi_size >> 9) >
		     tmp_dev->start_sector + tmp_dev->num_sectors)) {
		/* This bio crosses a device boundary, so we have to
		 * split it.
		 */
		struct bio_pair *bp;

		bp = bio_split(bio,
			       tmp_dev->start_sector + tmp_dev->num_sectors
			       - bio->bi_sector);

		if (linear_make_request(q, &bp->bio1))
			generic_make_request(&bp->bio1);
		if (linear_make_request(q, &bp->bio2))
			generic_make_request(&bp->bio2);
		bio_pair_release(bp);
		return 0;
	}
		    
	bio->bi_bdev = tmp_dev->rdev->bdev;
	bio->bi_sector = bio->bi_sector - tmp_dev->start_sector
		+ tmp_dev->rdev->data_offset;

	return 1;
}

static void linear_status (struct seq_file *seq, mddev_t *mddev)
{

	seq_printf(seq, " %dk rounding", mddev->chunk_size/1024);
}


static struct mdk_personality linear_personality =
{
	.name		= "linear",
	.level		= LEVEL_LINEAR,
	.owner		= THIS_MODULE,
	.make_request	= linear_make_request,
	.run		= linear_run,
	.stop		= linear_stop,
	.status		= linear_status,
	.hot_add_disk	= linear_add,
	.size		= linear_size,
};

static int __init linear_init (void)
{
	return register_md_personality (&linear_personality);
}

static void linear_exit (void)
{
	unregister_md_personality (&linear_personality);
}


module_init(linear_init);
module_exit(linear_exit);
MODULE_LICENSE("GPL");
MODULE_ALIAS("md-personality-1"); /* LINEAR - deprecated*/
MODULE_ALIAS("md-linear");
MODULE_ALIAS("md-level--1");
Commit	Line	Data
1da177e4 LT	1	/*
	2	linear.c : Multiple Devices driver for Linux
	3	Copyright (C) 1994-96 Marc ZYNGIER
	4	<zyngier@ufr-info-p7.ibp.fr> or
	5	<maz@gloups.fdn.fr>
	6
	7	Linear mode management functions.
	8
	9	This program is free software; you can redistribute it and/or modify
	10	it under the terms of the GNU General Public License as published by
	11	the Free Software Foundation; either version 2, or (at your option)
	12	any later version.
	13
	14	You should have received a copy of the GNU General Public License
	15	(for example /usr/src/linux/COPYING); if not, write to the Free
	16	Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
	17	*/
	18
bff61975 N	19	#include <linux/blkdev.h>
bff61975 N	20	#include <linux/raid/md_u.h>
bff61975	21	#include <linux/seq_file.h>
43b2e5d8	22	#include "md.h"
ef740c37	23	#include "linear.h"
1da177e4	24
1da177e4 LT	25	/*
	26	* find which device holds a particular offset
	27	*/
	28	static inline dev_info_t which_dev(mddev_t mddev, sector_t sector)
	29	{
	30	dev_info_t *hash;
070ec55d	31	linear_conf_t *conf = mddev->private;
1da177e4	32
45d4582f	33	hash = conf->disks;
1da177e4	34
6283815d	35	while (sector >= hash->num_sectors + hash->start_sector)
1da177e4 LT	36	hash++;
	37	return hash;
	38	}
	39
	40	/**
15945fee	41	* linear_mergeable_bvec -- tell bio layer if two requests can be merged
1da177e4	42	* @q: request queue
cc371e66	43	* @bvm: properties of new bio
1da177e4 LT	44	* @biovec: the request that could be merged to it.
	45	*
	46	* Return amount of bytes we can take at this offset
	47	*/
cc371e66 AK	48	static int linear_mergeable_bvec(struct request_queue *q,
	49	struct bvec_merge_data *bvm,
	50	struct bio_vec *biovec)
1da177e4 LT	51	{
	52	mddev_t *mddev = q->queuedata;
	53	dev_info_t *dev0;
cc371e66 AK	54	unsigned long maxsectors, bio_sectors = bvm->bi_size >> 9;
cc371e66 AK	55	sector_t sector = bvm->bi_sector + get_start_sect(bvm->bi_bdev);
1da177e4 LT	56
1da177e4 LT	57	dev0 = which_dev(mddev, sector);
6283815d	58	maxsectors = dev0->num_sectors - (sector - dev0->start_sector);
1da177e4 LT	59
	60	if (maxsectors < bio_sectors)
	61	maxsectors = 0;
	62	else
	63	maxsectors -= bio_sectors;
	64
	65	if (maxsectors <= (PAGE_SIZE >> 9 ) && bio_sectors == 0)
	66	return biovec->bv_len;
	67	/* The bytes available at this offset could be really big,
	68	* so we cap at 2^31 to avoid overflow */
	69	if (maxsectors > (1 << (31-9)))
	70	return 1<<31;
	71	return maxsectors << 9;
	72	}
	73
165125e1	74	static void linear_unplug(struct request_queue *q)
1da177e4 LT	75	{
1da177e4 LT	76	mddev_t *mddev = q->queuedata;
070ec55d	77	linear_conf_t *conf = mddev->private;
1da177e4 LT	78	int i;
	79
	80	for (i=0; i < mddev->raid_disks; i++) {
165125e1	81	struct request_queue *r_queue = bdev_get_queue(conf->disks[i].rdev->bdev);
2ad8b1ef	82	blk_unplug(r_queue);
1da177e4 LT	83	}
	84	}
	85
26be34dc N	86	static int linear_congested(void *data, int bits)
	87	{
	88	mddev_t *mddev = data;
070ec55d	89	linear_conf_t *conf = mddev->private;
26be34dc N	90	int i, ret = 0;
	91
	92	for (i = 0; i < mddev->raid_disks && !ret ; i++) {
165125e1	93	struct request_queue *q = bdev_get_queue(conf->disks[i].rdev->bdev);
26be34dc N	94	ret \|= bdi_congested(&q->backing_dev_info, bits);
	95	}
	96	return ret;
	97	}
	98
80c3a6ce DW	99	static sector_t linear_size(mddev_t *mddev, sector_t sectors, int raid_disks)
80c3a6ce DW	100	{
070ec55d	101	linear_conf_t *conf = mddev->private;
80c3a6ce DW	102
	103	WARN_ONCE(sectors \|\| raid_disks,
	104	"%s does not support generic reshape\n", __func__);
	105
	106	return conf->array_sectors;
	107	}
	108
7c7546cc	109	static linear_conf_t linear_conf(mddev_t mddev, int raid_disks)
1da177e4 LT	110	{
1da177e4 LT	111	linear_conf_t *conf;
1da177e4	112	mdk_rdev_t *rdev;
45d4582f	113	int i, cnt;
1da177e4	114
7c7546cc	115	conf = kzalloc (sizeof (conf) + raid_diskssizeof(dev_info_t),
1da177e4 LT	116	GFP_KERNEL);
1da177e4 LT	117	if (!conf)
7c7546cc N	118	return NULL;
7c7546cc N	119
1da177e4	120	cnt = 0;
d6e22150	121	conf->array_sectors = 0;
1da177e4	122
159ec1fc	123	list_for_each_entry(rdev, &mddev->disks, same_set) {
1da177e4 LT	124	int j = rdev->raid_disk;
	125	dev_info_t *disk = conf->disks + j;
	126
13864515	127	if (j < 0 \|\| j >= raid_disks \|\| disk->rdev) {
1da177e4 LT	128	printk("linear: disk numbering problem. Aborting!\n");
	129	goto out;
	130	}
	131
	132	disk->rdev = rdev;
	133
	134	blk_queue_stack_limits(mddev->queue,
	135	rdev->bdev->bd_disk->queue);
	136	/* as we don't honour merge_bvec_fn, we must never risk
	137	* violating it, so limit ->max_sector to one PAGE, as
	138	* a one page request is never in violation.
	139	*/
	140	if (rdev->bdev->bd_disk->queue->merge_bvec_fn &&
ae03bf63	141	queue_max_sectors(mddev->queue) > (PAGE_SIZE>>9))
1da177e4 LT	142	blk_queue_max_sectors(mddev->queue, PAGE_SIZE>>9);
1da177e4 LT	143
dd8ac336 AN	144	disk->num_sectors = rdev->sectors;
dd8ac336 AN	145	conf->array_sectors += rdev->sectors;
1da177e4	146
1da177e4 LT	147	cnt++;
1da177e4 LT	148	}
7c7546cc	149	if (cnt != raid_disks) {
1da177e4 LT	150	printk("linear: not enough drives present. Aborting!\n");
	151	goto out;
	152	}
	153
	154	/*
45d4582f	155	* Here we calculate the device offsets.
1da177e4	156	*/
6283815d	157	conf->disks[0].start_sector = 0;
a778b73f	158	for (i = 1; i < raid_disks; i++)
6283815d AN	159	conf->disks[i].start_sector =
	160	conf->disks[i-1].start_sector +
	161	conf->disks[i-1].num_sectors;
15945fee	162
7c7546cc N	163	return conf;
	164
	165	out:
	166	kfree(conf);
	167	return NULL;
	168	}
	169
	170	static int linear_run (mddev_t *mddev)
	171	{
	172	linear_conf_t *conf;
	173
e7e72bf6	174	mddev->queue->queue_lock = &mddev->queue->__queue_lock;
7c7546cc N	175	conf = linear_conf(mddev, mddev->raid_disks);
	176
	177	if (!conf)
	178	return 1;
	179	mddev->private = conf;
1f403624	180	md_set_array_sectors(mddev, linear_size(mddev, 0, 0));
7c7546cc	181
1da177e4 LT	182	blk_queue_merge_bvec(mddev->queue, linear_mergeable_bvec);
1da177e4 LT	183	mddev->queue->unplug_fn = linear_unplug;
26be34dc N	184	mddev->queue->backing_dev_info.congested_fn = linear_congested;
26be34dc N	185	mddev->queue->backing_dev_info.congested_data = mddev;
1da177e4	186	return 0;
7c7546cc	187	}
1da177e4	188
7c7546cc N	189	static int linear_add(mddev_t mddev, mdk_rdev_t rdev)
	190	{
	191	/* Adding a drive to a linear array allows the array to grow.
	192	* It is permitted if the new drive has a matching superblock
	193	* already on it, with raid_disk equal to raid_disks.
	194	* It is achieved by creating a new linear_private_data structure
	195	* and swapping it in in-place of the current one.
	196	* The current one is never freed until the array is stopped.
	197	* This avoids races.
	198	*/
	199	linear_conf_t *newconf;
	200
a778b73f	201	if (rdev->saved_raid_disk != mddev->raid_disks)
7c7546cc N	202	return -EINVAL;
7c7546cc N	203
a778b73f N	204	rdev->raid_disk = rdev->saved_raid_disk;
a778b73f N	205
7c7546cc N	206	newconf = linear_conf(mddev,mddev->raid_disks+1);
	207
	208	if (!newconf)
	209	return -ENOMEM;
	210
070ec55d	211	newconf->prev = mddev->private;
7c7546cc N	212	mddev->private = newconf;
7c7546cc N	213	mddev->raid_disks++;
1f403624	214	md_set_array_sectors(mddev, linear_size(mddev, 0, 0));
f233ea5c	215	set_capacity(mddev->gendisk, mddev->array_sectors);
7c7546cc	216	return 0;
1da177e4 LT	217	}
	218
	219	static int linear_stop (mddev_t *mddev)
	220	{
070ec55d	221	linear_conf_t *conf = mddev->private;
1da177e4 LT	222
1da177e4 LT	223	blk_sync_queue(mddev->queue); /* the unplug fn references 'conf'*/
7c7546cc N	224	do {
7c7546cc N	225	linear_conf_t *t = conf->prev;
7c7546cc N	226	kfree(conf);
	227	conf = t;
	228	} while (conf);
1da177e4 LT	229
	230	return 0;
	231	}
	232
165125e1	233	static int linear_make_request (struct request_queue q, struct bio bio)
1da177e4	234	{
a362357b	235	const int rw = bio_data_dir(bio);
1da177e4 LT	236	mddev_t *mddev = q->queuedata;
1da177e4 LT	237	dev_info_t *tmp_dev;
c9959059	238	int cpu;
1da177e4	239
e5dcdd80	240	if (unlikely(bio_barrier(bio))) {
6712ecf8	241	bio_endio(bio, -EOPNOTSUPP);
e5dcdd80 N	242	return 0;
	243	}
	244
074a7aca TH	245	cpu = part_stat_lock();
	246	part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]);
	247	part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw],
	248	bio_sectors(bio));
	249	part_stat_unlock();
1da177e4 LT	250
1da177e4 LT	251	tmp_dev = which_dev(mddev, bio->bi_sector);
1da177e4	252
6283815d AN	253	if (unlikely(bio->bi_sector >= (tmp_dev->num_sectors +
	254	tmp_dev->start_sector)
	255	\|\| (bio->bi_sector <
	256	tmp_dev->start_sector))) {
1da177e4 LT	257	char b[BDEVNAME_SIZE];
1da177e4 LT	258
6283815d AN	259	printk("linear_make_request: Sector %llu out of bounds on "
	260	"dev %s: %llu sectors, offset %llu\n",
	261	(unsigned long long)bio->bi_sector,
1da177e4	262	bdevname(tmp_dev->rdev->bdev, b),
6283815d AN	263	(unsigned long long)tmp_dev->num_sectors,
6283815d AN	264	(unsigned long long)tmp_dev->start_sector);
6712ecf8	265	bio_io_error(bio);
1da177e4 LT	266	return 0;
	267	}
	268	if (unlikely(bio->bi_sector + (bio->bi_size >> 9) >
6283815d	269	tmp_dev->start_sector + tmp_dev->num_sectors)) {
1da177e4 LT	270	/* This bio crosses a device boundary, so we have to
	271	* split it.
	272	*/
	273	struct bio_pair *bp;
6283815d	274
6feef531	275	bp = bio_split(bio,
6283815d AN	276	tmp_dev->start_sector + tmp_dev->num_sectors
	277	- bio->bi_sector);
	278
1da177e4 LT	279	if (linear_make_request(q, &bp->bio1))
	280	generic_make_request(&bp->bio1);
	281	if (linear_make_request(q, &bp->bio2))
	282	generic_make_request(&bp->bio2);
	283	bio_pair_release(bp);
	284	return 0;
	285	}
	286
	287	bio->bi_bdev = tmp_dev->rdev->bdev;
6283815d AN	288	bio->bi_sector = bio->bi_sector - tmp_dev->start_sector
6283815d AN	289	+ tmp_dev->rdev->data_offset;
1da177e4 LT	290
	291	return 1;
	292	}
	293
	294	static void linear_status (struct seq_file seq, mddev_t mddev)
	295	{
	296
1da177e4 LT	297	seq_printf(seq, " %dk rounding", mddev->chunk_size/1024);
	298	}
	299
	300
2604b703	301	static struct mdk_personality linear_personality =
1da177e4 LT	302	{
1da177e4 LT	303	.name = "linear",
2604b703	304	.level = LEVEL_LINEAR,
1da177e4 LT	305	.owner = THIS_MODULE,
	306	.make_request = linear_make_request,
	307	.run = linear_run,
	308	.stop = linear_stop,
	309	.status = linear_status,
7c7546cc	310	.hot_add_disk = linear_add,
80c3a6ce	311	.size = linear_size,
1da177e4 LT	312	};
	313
	314	static int __init linear_init (void)
	315	{
2604b703	316	return register_md_personality (&linear_personality);
1da177e4 LT	317	}
	318
	319	static void linear_exit (void)
	320	{
2604b703	321	unregister_md_personality (&linear_personality);
1da177e4 LT	322	}
	323
	324
	325	module_init(linear_init);
	326	module_exit(linear_exit);
	327	MODULE_LICENSE("GPL");
d9d166c2 N	328	MODULE_ALIAS("md-personality-1"); /* LINEAR - deprecated*/
d9d166c2 N	329	MODULE_ALIAS("md-linear");
2604b703	330	MODULE_ALIAS("md-level--1");