bitmap.c 66.7 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
/*
 * bitmap.c two-level bitmap (C) Peter T. Breuer (ptb@ot.uc3m.es) 2003
 *
 * bitmap_create  - sets up the bitmap structure
 * bitmap_destroy - destroys the bitmap structure
 *
 * additions, Copyright (C) 2003-2004, Paul Clements, SteelEye Technology, Inc.:
 * - added disk storage for bitmap
 * - changes to allow various bitmap chunk sizes
 */

/*
 * Still to do:
 *
 * flush after percent set rather than just time based. (maybe both).
 */

18
#include <linux/blkdev.h>
19 20 21 22 23 24 25 26 27 28
#include <linux/module.h>
#include <linux/errno.h>
#include <linux/slab.h>
#include <linux/init.h>
#include <linux/timer.h>
#include <linux/sched.h>
#include <linux/list.h>
#include <linux/file.h>
#include <linux/mount.h>
#include <linux/buffer_head.h>
29
#include <linux/seq_file.h>
30
#include "md.h"
31
#include "bitmap.h"
32

33
static inline char *bmname(struct bitmap *bitmap)
34 35 36 37 38 39 40 41 42 43 44 45 46 47
{
	return bitmap->mddev ? mdname(bitmap->mddev) : "mdX";
}

/*
 * check a page and, if necessary, allocate it (or hijack it if the alloc fails)
 *
 * 1) check to see if this page is allocated, if it's not then try to alloc
 * 2) if the alloc fails, set the page's hijacked flag so we'll use the
 *    page pointer directly as a counter
 *
 * if we find our page, we increment the page's refcount so that it stays
 * allocated while we're using it
 */
48
static int bitmap_checkpage(struct bitmap_counts *bitmap,
49
			    unsigned long page, int create, int no_hijack)
50 51
__releases(bitmap->lock)
__acquires(bitmap->lock)
52 53 54 55
{
	unsigned char *mappage;

	if (page >= bitmap->pages) {
56 57 58 59
		/* This can happen if bitmap_start_sync goes beyond
		 * End-of-device while looking for a whole page.
		 * It is harmless.
		 */
60 61 62 63 64 65 66 67 68 69 70 71 72 73
		return -EINVAL;
	}

	if (bitmap->bp[page].hijacked) /* it's hijacked, don't try to alloc */
		return 0;

	if (bitmap->bp[page].map) /* page is already allocated, just return */
		return 0;

	if (!create)
		return -ENOENT;

	/* this page has not been allocated yet */

74
	spin_unlock_irq(&bitmap->lock);
75 76 77 78 79 80 81 82 83 84 85 86 87
	/* It is possible that this is being called inside a
	 * prepare_to_wait/finish_wait loop from raid5c:make_request().
	 * In general it is not permitted to sleep in that context as it
	 * can cause the loop to spin freely.
	 * That doesn't apply here as we can only reach this point
	 * once with any loop.
	 * When this function completes, either bp[page].map or
	 * bp[page].hijacked.  In either case, this function will
	 * abort before getting to this point again.  So there is
	 * no risk of a free-spin, and so it is safe to assert
	 * that sleeping here is allowed.
	 */
	sched_annotate_sleep();
88
	mappage = kzalloc(PAGE_SIZE, GFP_NOIO);
89 90 91
	spin_lock_irq(&bitmap->lock);

	if (mappage == NULL) {
92
		pr_debug("md/bitmap: map page allocation failed, hijacking\n");
93 94 95
		/* We don't support hijack for cluster raid */
		if (no_hijack)
			return -ENOMEM;
96 97 98 99
		/* failed - set the hijacked flag so that we can use the
		 * pointer as a counter */
		if (!bitmap->bp[page].map)
			bitmap->bp[page].hijacked = 1;
100 101
	} else if (bitmap->bp[page].map ||
		   bitmap->bp[page].hijacked) {
102
		/* somebody beat us to getting the page */
103
		kfree(mappage);
104
	} else {
105

106
		/* no page was in place and we have one, so install it */
107

108 109 110
		bitmap->bp[page].map = mappage;
		bitmap->missing_pages--;
	}
111 112 113 114 115 116
	return 0;
}

/* if page is completely empty, put it back on the free list, or dealloc it */
/* if page was hijacked, unmark the flag so it might get alloced next time */
/* Note: lock should be held when calling this */
117
static void bitmap_checkfree(struct bitmap_counts *bitmap, unsigned long page)
118 119 120 121 122 123 124 125 126 127 128
{
	char *ptr;

	if (bitmap->bp[page].count) /* page is still busy */
		return;

	/* page is no longer in use, it can be released */

	if (bitmap->bp[page].hijacked) { /* page was hijacked, undo this now */
		bitmap->bp[page].hijacked = 0;
		bitmap->bp[page].map = NULL;
129 130 131 132 133
	} else {
		/* normal case, free the page */
		ptr = bitmap->bp[page].map;
		bitmap->bp[page].map = NULL;
		bitmap->missing_pages++;
134
		kfree(ptr);
135 136 137 138 139 140 141 142 143 144 145
	}
}

/*
 * bitmap file handling - read and write the bitmap file and its superblock
 */

/*
 * basic page I/O operations
 */

146
/* IO operations when bitmap is stored near all superblocks */
147 148 149
static int read_sb_page(struct mddev *mddev, loff_t offset,
			struct page *page,
			unsigned long index, int size)
150 151 152
{
	/* choose a good rdev and read the page from there */

153
	struct md_rdev *rdev;
154 155
	sector_t target;

N
NeilBrown 已提交
156
	rdev_for_each(rdev, mddev) {
157 158
		if (! test_bit(In_sync, &rdev->flags)
		    || test_bit(Faulty, &rdev->flags))
159 160
			continue;

J
Jonathan Brassow 已提交
161
		target = offset + index * (PAGE_SIZE/512);
162

163
		if (sync_page_io(rdev, target,
164
				 roundup(size, bdev_logical_block_size(rdev->bdev)),
M
Mike Christie 已提交
165
				 page, REQ_OP_READ, 0, true)) {
166
			page->index = index;
167
			return 0;
168 169
		}
	}
170
	return -EIO;
171 172
}

173
static struct md_rdev *next_active_rdev(struct md_rdev *rdev, struct mddev *mddev)
174 175 176 177 178 179 180
{
	/* Iterate the disks of an mddev, using rcu to protect access to the
	 * linked list, and raising the refcount of devices we return to ensure
	 * they don't disappear while in use.
	 * As devices are only added or removed when raid_disk is < 0 and
	 * nr_pending is 0 and In_sync is clear, the entries we return will
	 * still be in the same position on the list when we re-enter
181
	 * list_for_each_entry_continue_rcu.
182 183 184 185 186
	 *
	 * Note that if entered with 'rdev == NULL' to start at the
	 * beginning, we temporarily assign 'rdev' to an address which
	 * isn't really an rdev, but which can be used by
	 * list_for_each_entry_continue_rcu() to find the first entry.
187 188 189 190
	 */
	rcu_read_lock();
	if (rdev == NULL)
		/* start at the beginning */
191
		rdev = list_entry(&mddev->disks, struct md_rdev, same_set);
192 193 194 195
	else {
		/* release the previous rdev and start from there. */
		rdev_dec_pending(rdev, mddev);
	}
196
	list_for_each_entry_continue_rcu(rdev, &mddev->disks, same_set) {
197 198 199 200 201 202 203 204 205 206 207 208
		if (rdev->raid_disk >= 0 &&
		    !test_bit(Faulty, &rdev->flags)) {
			/* this is a usable devices */
			atomic_inc(&rdev->nr_pending);
			rcu_read_unlock();
			return rdev;
		}
	}
	rcu_read_unlock();
	return NULL;
}

209
static int write_sb_page(struct bitmap *bitmap, struct page *page, int wait)
210
{
211
	struct md_rdev *rdev = NULL;
212
	struct block_device *bdev;
213
	struct mddev *mddev = bitmap->mddev;
214
	struct bitmap_storage *store = &bitmap->storage;
215

216
	while ((rdev = next_active_rdev(rdev, mddev)) != NULL) {
217 218
		int size = PAGE_SIZE;
		loff_t offset = mddev->bitmap_info.offset;
219 220 221

		bdev = (rdev->meta_bdev) ? rdev->meta_bdev : rdev->bdev;

222 223 224 225 226
		if (page->index == store->file_pages-1) {
			int last_page_size = store->bytes & (PAGE_SIZE-1);
			if (last_page_size == 0)
				last_page_size = PAGE_SIZE;
			size = roundup(last_page_size,
227
				       bdev_logical_block_size(bdev));
228
		}
229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268
		/* Just make sure we aren't corrupting data or
		 * metadata
		 */
		if (mddev->external) {
			/* Bitmap could be anywhere. */
			if (rdev->sb_start + offset + (page->index
						       * (PAGE_SIZE/512))
			    > rdev->data_offset
			    &&
			    rdev->sb_start + offset
			    < (rdev->data_offset + mddev->dev_sectors
			     + (PAGE_SIZE/512)))
				goto bad_alignment;
		} else if (offset < 0) {
			/* DATA  BITMAP METADATA  */
			if (offset
			    + (long)(page->index * (PAGE_SIZE/512))
			    + size/512 > 0)
				/* bitmap runs in to metadata */
				goto bad_alignment;
			if (rdev->data_offset + mddev->dev_sectors
			    > rdev->sb_start + offset)
				/* data runs in to bitmap */
				goto bad_alignment;
		} else if (rdev->sb_start < rdev->data_offset) {
			/* METADATA BITMAP DATA */
			if (rdev->sb_start
			    + offset
			    + page->index*(PAGE_SIZE/512) + size/512
			    > rdev->data_offset)
				/* bitmap runs in to data */
				goto bad_alignment;
		} else {
			/* DATA METADATA BITMAP - no problems */
		}
		md_super_write(mddev, rdev,
			       rdev->sb_start + offset
			       + page->index * (PAGE_SIZE/512),
			       size,
			       page);
269
	}
270 271

	if (wait)
272
		md_super_wait(mddev);
273
	return 0;
274 275 276

 bad_alignment:
	return -EINVAL;
277 278
}

279
static void bitmap_file_kick(struct bitmap *bitmap);
280
/*
281
 * write out a page to a file
282
 */
283
static void write_page(struct bitmap *bitmap, struct page *page, int wait)
284
{
285
	struct buffer_head *bh;
286

287
	if (bitmap->storage.file == NULL) {
288 289
		switch (write_sb_page(bitmap, page, wait)) {
		case -EINVAL:
290
			set_bit(BITMAP_WRITE_ERROR, &bitmap->flags);
291
		}
292
	} else {
293

294
		bh = page_buffers(page);
295

296 297 298 299
		while (bh && bh->b_blocknr) {
			atomic_inc(&bitmap->pending_writes);
			set_buffer_locked(bh);
			set_buffer_mapped(bh);
300
			submit_bh(REQ_OP_WRITE, REQ_SYNC, bh);
301 302
			bh = bh->b_this_page;
		}
303

304
		if (wait)
305 306
			wait_event(bitmap->write_wait,
				   atomic_read(&bitmap->pending_writes)==0);
307
	}
308
	if (test_bit(BITMAP_WRITE_ERROR, &bitmap->flags))
309
		bitmap_file_kick(bitmap);
310 311 312 313 314
}

static void end_bitmap_write(struct buffer_head *bh, int uptodate)
{
	struct bitmap *bitmap = bh->b_private;
315

316 317
	if (!uptodate)
		set_bit(BITMAP_WRITE_ERROR, &bitmap->flags);
318 319 320
	if (atomic_dec_and_test(&bitmap->pending_writes))
		wake_up(&bitmap->write_wait);
}
321

322 323 324 325 326 327
/* copied from buffer.c */
static void
__clear_page_buffers(struct page *page)
{
	ClearPagePrivate(page);
	set_page_private(page, 0);
328
	put_page(page);
329 330 331
}
static void free_buffers(struct page *page)
{
332
	struct buffer_head *bh;
333

334 335 336 337
	if (!PagePrivate(page))
		return;

	bh = page_buffers(page);
338 339 340 341
	while (bh) {
		struct buffer_head *next = bh->b_this_page;
		free_buffer_head(bh);
		bh = next;
342
	}
343 344
	__clear_page_buffers(page);
	put_page(page);
345 346
}

347 348 349 350 351 352 353
/* read a page from a file.
 * We both read the page, and attach buffers to the page to record the
 * address of each block (using bmap).  These addresses will be used
 * to write the block later, completely bypassing the filesystem.
 * This usage is similar to how swap files are handled, and allows us
 * to write to a file with no concerns of memory allocation failing.
 */
354 355 356 357
static int read_page(struct file *file, unsigned long index,
		     struct bitmap *bitmap,
		     unsigned long count,
		     struct page *page)
358
{
359
	int ret = 0;
A
Al Viro 已提交
360
	struct inode *inode = file_inode(file);
361 362
	struct buffer_head *bh;
	sector_t block;
363

364 365
	pr_debug("read bitmap file (%dB @ %llu)\n", (int)PAGE_SIZE,
		 (unsigned long long)index << PAGE_SHIFT);
366

367 368
	bh = alloc_page_buffers(page, 1<<inode->i_blkbits, 0);
	if (!bh) {
369
		ret = -ENOMEM;
370 371
		goto out;
	}
372 373 374 375 376 377 378 379 380
	attach_page_buffers(page, bh);
	block = index << (PAGE_SHIFT - inode->i_blkbits);
	while (bh) {
		if (count == 0)
			bh->b_blocknr = 0;
		else {
			bh->b_blocknr = bmap(inode, block);
			if (bh->b_blocknr == 0) {
				/* Cannot use this file! */
381
				ret = -EINVAL;
382 383 384 385 386 387 388 389 390 391
				goto out;
			}
			bh->b_bdev = inode->i_sb->s_bdev;
			if (count < (1<<inode->i_blkbits))
				count = 0;
			else
				count -= (1<<inode->i_blkbits);

			bh->b_end_io = end_bitmap_write;
			bh->b_private = bitmap;
392 393 394
			atomic_inc(&bitmap->pending_writes);
			set_buffer_locked(bh);
			set_buffer_mapped(bh);
395
			submit_bh(REQ_OP_READ, 0, bh);
396 397 398 399 400
		}
		block++;
		bh = bh->b_this_page;
	}
	page->index = index;
401 402 403

	wait_event(bitmap->write_wait,
		   atomic_read(&bitmap->pending_writes)==0);
404
	if (test_bit(BITMAP_WRITE_ERROR, &bitmap->flags))
405
		ret = -EIO;
406
out:
407 408
	if (ret)
		printk(KERN_ALERT "md: bitmap read error: (%dB @ %llu): %d\n",
409 410
			(int)PAGE_SIZE,
			(unsigned long long)index << PAGE_SHIFT,
411 412
			ret);
	return ret;
413 414 415 416 417 418 419
}

/*
 * bitmap file superblock operations
 */

/* update the event counter and sync the superblock to disk */
420
void bitmap_update_sb(struct bitmap *bitmap)
421 422 423 424
{
	bitmap_super_t *sb;

	if (!bitmap || !bitmap->mddev) /* no bitmap for this array */
425
		return;
426 427
	if (bitmap->mddev->bitmap_info.external)
		return;
428
	if (!bitmap->storage.sb_page) /* no superblock */
429
		return;
430
	sb = kmap_atomic(bitmap->storage.sb_page);
431
	sb->events = cpu_to_le64(bitmap->mddev->events);
432
	if (bitmap->mddev->events < bitmap->events_cleared)
433 434
		/* rocking back to read-only */
		bitmap->events_cleared = bitmap->mddev->events;
435 436
	sb->events_cleared = cpu_to_le64(bitmap->events_cleared);
	sb->state = cpu_to_le32(bitmap->flags);
437 438 439
	/* Just in case these have been changed via sysfs: */
	sb->daemon_sleep = cpu_to_le32(bitmap->mddev->bitmap_info.daemon_sleep/HZ);
	sb->write_behind = cpu_to_le32(bitmap->mddev->bitmap_info.max_write_behind);
440 441 442
	/* This might have been changed by a reshape */
	sb->sync_size = cpu_to_le64(bitmap->mddev->resync_max_sectors);
	sb->chunksize = cpu_to_le32(bitmap->mddev->bitmap_info.chunksize);
G
Goldwyn Rodrigues 已提交
443
	sb->nodes = cpu_to_le32(bitmap->mddev->bitmap_info.nodes);
444 445
	sb->sectors_reserved = cpu_to_le32(bitmap->mddev->
					   bitmap_info.space);
446
	kunmap_atomic(sb);
447
	write_page(bitmap, bitmap->storage.sb_page, 1);
448 449 450 451 452 453 454
}

/* print out the bitmap file superblock */
void bitmap_print_sb(struct bitmap *bitmap)
{
	bitmap_super_t *sb;

455
	if (!bitmap || !bitmap->storage.sb_page)
456
		return;
457
	sb = kmap_atomic(bitmap->storage.sb_page);
458
	printk(KERN_DEBUG "%s: bitmap file superblock:\n", bmname(bitmap));
459 460 461
	printk(KERN_DEBUG "         magic: %08x\n", le32_to_cpu(sb->magic));
	printk(KERN_DEBUG "       version: %d\n", le32_to_cpu(sb->version));
	printk(KERN_DEBUG "          uuid: %08x.%08x.%08x.%08x\n",
462 463 464 465
					*(__u32 *)(sb->uuid+0),
					*(__u32 *)(sb->uuid+4),
					*(__u32 *)(sb->uuid+8),
					*(__u32 *)(sb->uuid+12));
466
	printk(KERN_DEBUG "        events: %llu\n",
467
			(unsigned long long) le64_to_cpu(sb->events));
468
	printk(KERN_DEBUG "events cleared: %llu\n",
469
			(unsigned long long) le64_to_cpu(sb->events_cleared));
470 471 472 473 474
	printk(KERN_DEBUG "         state: %08x\n", le32_to_cpu(sb->state));
	printk(KERN_DEBUG "     chunksize: %d B\n", le32_to_cpu(sb->chunksize));
	printk(KERN_DEBUG "  daemon sleep: %ds\n", le32_to_cpu(sb->daemon_sleep));
	printk(KERN_DEBUG "     sync size: %llu KB\n",
			(unsigned long long)le64_to_cpu(sb->sync_size)/2);
475
	printk(KERN_DEBUG "max write behind: %d\n", le32_to_cpu(sb->write_behind));
476
	kunmap_atomic(sb);
477 478
}

479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494
/*
 * bitmap_new_disk_sb
 * @bitmap
 *
 * This function is somewhat the reverse of bitmap_read_sb.  bitmap_read_sb
 * reads and verifies the on-disk bitmap superblock and populates bitmap_info.
 * This function verifies 'bitmap_info' and populates the on-disk bitmap
 * structure, which is to be written to disk.
 *
 * Returns: 0 on success, -Exxx on error
 */
static int bitmap_new_disk_sb(struct bitmap *bitmap)
{
	bitmap_super_t *sb;
	unsigned long chunksize, daemon_sleep, write_behind;

495
	bitmap->storage.sb_page = alloc_page(GFP_KERNEL | __GFP_ZERO);
496 497
	if (bitmap->storage.sb_page == NULL)
		return -ENOMEM;
498
	bitmap->storage.sb_page->index = 0;
499

500
	sb = kmap_atomic(bitmap->storage.sb_page);
501 502 503 504 505 506 507

	sb->magic = cpu_to_le32(BITMAP_MAGIC);
	sb->version = cpu_to_le32(BITMAP_MAJOR_HI);

	chunksize = bitmap->mddev->bitmap_info.chunksize;
	BUG_ON(!chunksize);
	if (!is_power_of_2(chunksize)) {
508
		kunmap_atomic(sb);
509 510 511 512 513 514
		printk(KERN_ERR "bitmap chunksize not a power of 2\n");
		return -EINVAL;
	}
	sb->chunksize = cpu_to_le32(chunksize);

	daemon_sleep = bitmap->mddev->bitmap_info.daemon_sleep;
515
	if (!daemon_sleep || (daemon_sleep > MAX_SCHEDULE_TIMEOUT)) {
516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536
		printk(KERN_INFO "Choosing daemon_sleep default (5 sec)\n");
		daemon_sleep = 5 * HZ;
	}
	sb->daemon_sleep = cpu_to_le32(daemon_sleep);
	bitmap->mddev->bitmap_info.daemon_sleep = daemon_sleep;

	/*
	 * FIXME: write_behind for RAID1.  If not specified, what
	 * is a good choice?  We choose COUNTER_MAX / 2 arbitrarily.
	 */
	write_behind = bitmap->mddev->bitmap_info.max_write_behind;
	if (write_behind > COUNTER_MAX)
		write_behind = COUNTER_MAX / 2;
	sb->write_behind = cpu_to_le32(write_behind);
	bitmap->mddev->bitmap_info.max_write_behind = write_behind;

	/* keep the array size field of the bitmap superblock up to date */
	sb->sync_size = cpu_to_le64(bitmap->mddev->resync_max_sectors);

	memcpy(sb->uuid, bitmap->mddev->uuid, 16);

537
	set_bit(BITMAP_STALE, &bitmap->flags);
538
	sb->state = cpu_to_le32(bitmap->flags);
539 540
	bitmap->events_cleared = bitmap->mddev->events;
	sb->events_cleared = cpu_to_le64(bitmap->mddev->events);
541
	bitmap->mddev->bitmap_info.nodes = 0;
542

543
	kunmap_atomic(sb);
544 545 546 547

	return 0;
}

548 549 550 551 552
/* read the superblock from the bitmap file and initialize some bitmap fields */
static int bitmap_read_sb(struct bitmap *bitmap)
{
	char *reason = NULL;
	bitmap_super_t *sb;
553
	unsigned long chunksize, daemon_sleep, write_behind;
554
	unsigned long long events;
G
Goldwyn Rodrigues 已提交
555
	int nodes = 0;
556
	unsigned long sectors_reserved = 0;
557
	int err = -EINVAL;
558
	struct page *sb_page;
559
	loff_t offset = bitmap->mddev->bitmap_info.offset;
560

561
	if (!bitmap->storage.file && !bitmap->mddev->bitmap_info.offset) {
562 563 564
		chunksize = 128 * 1024 * 1024;
		daemon_sleep = 5 * HZ;
		write_behind = 0;
565
		set_bit(BITMAP_STALE, &bitmap->flags);
566 567 568
		err = 0;
		goto out_no_sb;
	}
569
	/* page 0 is the superblock, read it... */
570 571 572
	sb_page = alloc_page(GFP_KERNEL);
	if (!sb_page)
		return -ENOMEM;
573
	bitmap->storage.sb_page = sb_page;
574

575
re_read:
576 577
	/* If cluster_slot is set, the cluster is setup */
	if (bitmap->cluster_slot >= 0) {
578
		sector_t bm_blocks = bitmap->mddev->resync_max_sectors;
579

580 581
		sector_div(bm_blocks,
			   bitmap->mddev->bitmap_info.chunksize >> 9);
582 583 584
		/* bits to bytes */
		bm_blocks = ((bm_blocks+7) >> 3) + sizeof(bitmap_super_t);
		/* to 4k blocks */
585
		bm_blocks = DIV_ROUND_UP_SECTOR_T(bm_blocks, 4096);
586
		offset = bitmap->mddev->bitmap_info.offset + (bitmap->cluster_slot * (bm_blocks << 3));
587
		pr_info("%s:%d bm slot: %d offset: %llu\n", __func__, __LINE__,
588
			bitmap->cluster_slot, offset);
589 590
	}

591 592
	if (bitmap->storage.file) {
		loff_t isize = i_size_read(bitmap->storage.file->f_mapping->host);
593 594
		int bytes = isize > PAGE_SIZE ? PAGE_SIZE : isize;

595
		err = read_page(bitmap->storage.file, 0,
596
				bitmap, bytes, sb_page);
597
	} else {
598
		err = read_sb_page(bitmap->mddev,
599
				   offset,
600 601
				   sb_page,
				   0, sizeof(bitmap_super_t));
602
	}
603
	if (err)
604 605
		return err;

606
	err = -EINVAL;
607
	sb = kmap_atomic(sb_page);
608 609

	chunksize = le32_to_cpu(sb->chunksize);
610
	daemon_sleep = le32_to_cpu(sb->daemon_sleep) * HZ;
611
	write_behind = le32_to_cpu(sb->write_behind);
612
	sectors_reserved = le32_to_cpu(sb->sectors_reserved);
613 614
	/* Setup nodes/clustername only if bitmap version is
	 * cluster-compatible
615
	 */
616
	if (sb->version == cpu_to_le32(BITMAP_MAJOR_CLUSTERED)) {
617 618 619 620
		nodes = le32_to_cpu(sb->nodes);
		strlcpy(bitmap->mddev->bitmap_info.cluster_name,
				sb->cluster_name, 64);
	}
621 622 623 624

	/* verify that the bitmap-specific fields are valid */
	if (sb->magic != cpu_to_le32(BITMAP_MAGIC))
		reason = "bad magic";
625
	else if (le32_to_cpu(sb->version) < BITMAP_MAJOR_LO ||
626
		 le32_to_cpu(sb->version) > BITMAP_MAJOR_CLUSTERED)
627
		reason = "unrecognized superblock version";
628
	else if (chunksize < 512)
629
		reason = "bitmap chunksize too small";
J
Jonathan Brassow 已提交
630
	else if (!is_power_of_2(chunksize))
631
		reason = "bitmap chunksize not a power of 2";
632
	else if (daemon_sleep < 1 || daemon_sleep > MAX_SCHEDULE_TIMEOUT)
633
		reason = "daemon sleep period out of range";
634 635
	else if (write_behind > COUNTER_MAX)
		reason = "write-behind limit out of range (0 - 16383)";
636 637 638 639 640 641 642 643 644
	if (reason) {
		printk(KERN_INFO "%s: invalid bitmap file superblock: %s\n",
			bmname(bitmap), reason);
		goto out;
	}

	/* keep the array size field of the bitmap superblock up to date */
	sb->sync_size = cpu_to_le64(bitmap->mddev->resync_max_sectors);

645 646 647 648 649 650 651 652 653 654 655 656
	if (bitmap->mddev->persistent) {
		/*
		 * We have a persistent array superblock, so compare the
		 * bitmap's UUID and event counter to the mddev's
		 */
		if (memcmp(sb->uuid, bitmap->mddev->uuid, 16)) {
			printk(KERN_INFO
			       "%s: bitmap superblock UUID mismatch\n",
			       bmname(bitmap));
			goto out;
		}
		events = le64_to_cpu(sb->events);
657
		if (!nodes && (events < bitmap->mddev->events)) {
658 659 660 661 662
			printk(KERN_INFO
			       "%s: bitmap file is out of date (%llu < %llu) "
			       "-- forcing full recovery\n",
			       bmname(bitmap), events,
			       (unsigned long long) bitmap->mddev->events);
663
			set_bit(BITMAP_STALE, &bitmap->flags);
664
		}
665
	}
666

667
	/* assign fields using values from superblock */
668
	bitmap->flags |= le32_to_cpu(sb->state);
669
	if (le32_to_cpu(sb->version) == BITMAP_MAJOR_HOSTENDIAN)
670
		set_bit(BITMAP_HOSTENDIAN, &bitmap->flags);
671
	bitmap->events_cleared = le64_to_cpu(sb->events_cleared);
G
Goldwyn Rodrigues 已提交
672
	strlcpy(bitmap->mddev->bitmap_info.cluster_name, sb->cluster_name, 64);
673
	err = 0;
674

675
out:
676
	kunmap_atomic(sb);
677 678
	/* Assiging chunksize is required for "re_read" */
	bitmap->mddev->bitmap_info.chunksize = chunksize;
679
	if (err == 0 && nodes && (bitmap->cluster_slot < 0)) {
680 681 682 683 684 685 686 687 688 689 690
		err = md_setup_cluster(bitmap->mddev, nodes);
		if (err) {
			pr_err("%s: Could not setup cluster service (%d)\n",
					bmname(bitmap), err);
			goto out_no_sb;
		}
		bitmap->cluster_slot = md_cluster_ops->slot_number(bitmap->mddev);
		goto re_read;
	}


691
out_no_sb:
692
	if (test_bit(BITMAP_STALE, &bitmap->flags))
693 694 695 696
		bitmap->events_cleared = bitmap->mddev->events;
	bitmap->mddev->bitmap_info.chunksize = chunksize;
	bitmap->mddev->bitmap_info.daemon_sleep = daemon_sleep;
	bitmap->mddev->bitmap_info.max_write_behind = write_behind;
G
Goldwyn Rodrigues 已提交
697
	bitmap->mddev->bitmap_info.nodes = nodes;
698 699 700
	if (bitmap->mddev->bitmap_info.space == 0 ||
	    bitmap->mddev->bitmap_info.space > sectors_reserved)
		bitmap->mddev->bitmap_info.space = sectors_reserved;
701
	if (err) {
702
		bitmap_print_sb(bitmap);
703
		if (bitmap->cluster_slot < 0)
704 705
			md_cluster_stop(bitmap->mddev);
	}
706 707 708 709 710 711 712
	return err;
}

/*
 * general bitmap file operations
 */

713 714 715 716 717 718
/*
 * on-disk bitmap:
 *
 * Use one bit per "chunk" (block set). We do the disk I/O on the bitmap
 * file a page at a time. There's a superblock at the start of the file.
 */
719
/* calculate the index of the page that contains this bit */
720 721
static inline unsigned long file_page_index(struct bitmap_storage *store,
					    unsigned long chunk)
722
{
723
	if (store->sb_page)
724 725
		chunk += sizeof(bitmap_super_t) << 3;
	return chunk >> PAGE_BIT_SHIFT;
726 727 728
}

/* calculate the (bit) offset of this bit within a page */
729 730
static inline unsigned long file_page_offset(struct bitmap_storage *store,
					     unsigned long chunk)
731
{
732
	if (store->sb_page)
733 734
		chunk += sizeof(bitmap_super_t) << 3;
	return chunk & (PAGE_BITS - 1);
735 736 737 738 739 740
}

/*
 * return a pointer to the page in the filemap that contains the given bit
 *
 */
741
static inline struct page *filemap_get_page(struct bitmap_storage *store,
742
					    unsigned long chunk)
743
{
744
	if (file_page_index(store, chunk) >= store->file_pages)
745
		return NULL;
746
	return store->filemap[file_page_index(store, chunk)];
747 748
}

749
static int bitmap_storage_alloc(struct bitmap_storage *store,
750 751
				unsigned long chunks, int with_super,
				int slot_number)
752
{
753
	int pnum, offset = 0;
754 755 756 757 758 759 760 761
	unsigned long num_pages;
	unsigned long bytes;

	bytes = DIV_ROUND_UP(chunks, 8);
	if (with_super)
		bytes += sizeof(bitmap_super_t);

	num_pages = DIV_ROUND_UP(bytes, PAGE_SIZE);
762
	offset = slot_number * num_pages;
763 764 765 766 767 768 769

	store->filemap = kmalloc(sizeof(struct page *)
				 * num_pages, GFP_KERNEL);
	if (!store->filemap)
		return -ENOMEM;

	if (with_super && !store->sb_page) {
770
		store->sb_page = alloc_page(GFP_KERNEL|__GFP_ZERO);
771 772 773
		if (store->sb_page == NULL)
			return -ENOMEM;
	}
774

775 776 777 778
	pnum = 0;
	if (store->sb_page) {
		store->filemap[0] = store->sb_page;
		pnum = 1;
779
		store->sb_page->index = offset;
780
	}
781

782
	for ( ; pnum < num_pages; pnum++) {
783
		store->filemap[pnum] = alloc_page(GFP_KERNEL|__GFP_ZERO);
784 785 786 787
		if (!store->filemap[pnum]) {
			store->file_pages = pnum;
			return -ENOMEM;
		}
788
		store->filemap[pnum]->index = pnum + offset;
789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804
	}
	store->file_pages = pnum;

	/* We need 4 bits per page, rounded up to a multiple
	 * of sizeof(unsigned long) */
	store->filemap_attr = kzalloc(
		roundup(DIV_ROUND_UP(num_pages*4, 8), sizeof(unsigned long)),
		GFP_KERNEL);
	if (!store->filemap_attr)
		return -ENOMEM;

	store->bytes = bytes;

	return 0;
}

805
static void bitmap_file_unmap(struct bitmap_storage *store)
806 807 808
{
	struct page **map, *sb_page;
	int pages;
809
	struct file *file;
810

811
	file = store->file;
812 813 814
	map = store->filemap;
	pages = store->file_pages;
	sb_page = store->sb_page;
815 816

	while (pages--)
817
		if (map[pages] != sb_page) /* 0 is sb_page, release it below */
818
			free_buffers(map[pages]);
819
	kfree(map);
820
	kfree(store->filemap_attr);
821

822 823
	if (sb_page)
		free_buffers(sb_page);
824

825
	if (file) {
A
Al Viro 已提交
826
		struct inode *inode = file_inode(file);
827
		invalidate_mapping_pages(inode->i_mapping, 0, -1);
828
		fput(file);
829
	}
830 831 832 833 834 835 836 837 838 839 840
}

/*
 * bitmap_file_kick - if an error occurs while manipulating the bitmap file
 * then it is no longer reliable, so we stop using it and we mark the file
 * as failed in the superblock
 */
static void bitmap_file_kick(struct bitmap *bitmap)
{
	char *path, *ptr = NULL;

841
	if (!test_and_set_bit(BITMAP_STALE, &bitmap->flags)) {
842
		bitmap_update_sb(bitmap);
843

844
		if (bitmap->storage.file) {
845 846
			path = kmalloc(PAGE_SIZE, GFP_KERNEL);
			if (path)
M
Miklos Szeredi 已提交
847
				ptr = file_path(bitmap->storage.file,
848
					     path, PAGE_SIZE);
C
Christoph Hellwig 已提交
849

850 851
			printk(KERN_ALERT
			      "%s: kicking failed bitmap file %s from array!\n",
C
Christoph Hellwig 已提交
852
			      bmname(bitmap), IS_ERR(ptr) ? "" : ptr);
853

854 855 856 857 858
			kfree(path);
		} else
			printk(KERN_ALERT
			       "%s: disabling internal bitmap due to errors\n",
			       bmname(bitmap));
859
	}
860 861 862
}

enum bitmap_page_attr {
863
	BITMAP_PAGE_DIRTY = 0,     /* there are set bits that need to be synced */
864 865
	BITMAP_PAGE_PENDING = 1,   /* there are bits that are being cleaned.
				    * i.e. counter is 1 or 2. */
866
	BITMAP_PAGE_NEEDWRITE = 2, /* there are cleared bits that need to be synced */
867 868
};

869 870
static inline void set_page_attr(struct bitmap *bitmap, int pnum,
				 enum bitmap_page_attr attr)
871
{
872
	set_bit((pnum<<2) + attr, bitmap->storage.filemap_attr);
873 874
}

875 876
static inline void clear_page_attr(struct bitmap *bitmap, int pnum,
				   enum bitmap_page_attr attr)
877
{
878
	clear_bit((pnum<<2) + attr, bitmap->storage.filemap_attr);
879 880
}

881 882
static inline int test_page_attr(struct bitmap *bitmap, int pnum,
				 enum bitmap_page_attr attr)
883
{
884
	return test_bit((pnum<<2) + attr, bitmap->storage.filemap_attr);
885 886
}

887 888 889 890 891 892
static inline int test_and_clear_page_attr(struct bitmap *bitmap, int pnum,
					   enum bitmap_page_attr attr)
{
	return test_and_clear_bit((pnum<<2) + attr,
				  bitmap->storage.filemap_attr);
}
893 894 895 896 897 898 899 900 901 902
/*
 * bitmap_file_set_bit -- called before performing a write to the md device
 * to set (and eventually sync) a particular bit in the bitmap file
 *
 * we set the bit immediately, then we record the page number so that
 * when an unplug occurs, we can flush the dirty pages out to disk
 */
static void bitmap_file_set_bit(struct bitmap *bitmap, sector_t block)
{
	unsigned long bit;
903
	struct page *page;
904
	void *kaddr;
905
	unsigned long chunk = block >> bitmap->counts.chunkshift;
906 907 908 909 910
	struct bitmap_storage *store = &bitmap->storage;
	unsigned long node_offset = 0;

	if (mddev_is_clustered(bitmap->mddev))
		node_offset = bitmap->cluster_slot * store->file_pages;
911

912
	page = filemap_get_page(&bitmap->storage, chunk);
913 914
	if (!page)
		return;
915
	bit = file_page_offset(&bitmap->storage, chunk);
916

917
	/* set the bit */
918
	kaddr = kmap_atomic(page);
919
	if (test_bit(BITMAP_HOSTENDIAN, &bitmap->flags))
920 921
		set_bit(bit, kaddr);
	else
922
		set_bit_le(bit, kaddr);
923
	kunmap_atomic(kaddr);
924
	pr_debug("set file bit %lu page %lu\n", bit, page->index);
925
	/* record page number so it gets flushed to disk when unplug occurs */
926
	set_page_attr(bitmap, page->index - node_offset, BITMAP_PAGE_DIRTY);
927 928
}

929 930 931 932 933
static void bitmap_file_clear_bit(struct bitmap *bitmap, sector_t block)
{
	unsigned long bit;
	struct page *page;
	void *paddr;
934
	unsigned long chunk = block >> bitmap->counts.chunkshift;
935 936 937 938 939
	struct bitmap_storage *store = &bitmap->storage;
	unsigned long node_offset = 0;

	if (mddev_is_clustered(bitmap->mddev))
		node_offset = bitmap->cluster_slot * store->file_pages;
940

941
	page = filemap_get_page(&bitmap->storage, chunk);
942 943
	if (!page)
		return;
944
	bit = file_page_offset(&bitmap->storage, chunk);
945
	paddr = kmap_atomic(page);
946
	if (test_bit(BITMAP_HOSTENDIAN, &bitmap->flags))
947 948
		clear_bit(bit, paddr);
	else
949
		clear_bit_le(bit, paddr);
950
	kunmap_atomic(paddr);
951 952
	if (!test_page_attr(bitmap, page->index - node_offset, BITMAP_PAGE_NEEDWRITE)) {
		set_page_attr(bitmap, page->index - node_offset, BITMAP_PAGE_PENDING);
953 954 955 956
		bitmap->allclean = 0;
	}
}

957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978
static int bitmap_file_test_bit(struct bitmap *bitmap, sector_t block)
{
	unsigned long bit;
	struct page *page;
	void *paddr;
	unsigned long chunk = block >> bitmap->counts.chunkshift;
	int set = 0;

	page = filemap_get_page(&bitmap->storage, chunk);
	if (!page)
		return -EINVAL;
	bit = file_page_offset(&bitmap->storage, chunk);
	paddr = kmap_atomic(page);
	if (test_bit(BITMAP_HOSTENDIAN, &bitmap->flags))
		set = test_bit(bit, paddr);
	else
		set = test_bit_le(bit, paddr);
	kunmap_atomic(paddr);
	return set;
}


979 980 981
/* this gets called when the md device is ready to unplug its underlying
 * (slave) device queues -- before we let any writes go down, we need to
 * sync the dirty pages of the bitmap file to disk */
982
void bitmap_unplug(struct bitmap *bitmap)
983
{
984
	unsigned long i;
985
	int dirty, need_write;
986

987 988
	if (!bitmap || !bitmap->storage.filemap ||
	    test_bit(BITMAP_STALE, &bitmap->flags))
989
		return;
990 991 992

	/* look at each page to see if there are any set bits that need to be
	 * flushed out to disk */
993
	for (i = 0; i < bitmap->storage.file_pages; i++) {
994
		if (!bitmap->storage.filemap)
995
			return;
996 997 998 999
		dirty = test_and_clear_page_attr(bitmap, i, BITMAP_PAGE_DIRTY);
		need_write = test_and_clear_page_attr(bitmap, i,
						      BITMAP_PAGE_NEEDWRITE);
		if (dirty || need_write) {
1000
			clear_page_attr(bitmap, i, BITMAP_PAGE_PENDING);
1001 1002
			write_page(bitmap, bitmap->storage.filemap[i], 0);
		}
1003
	}
1004 1005 1006 1007 1008 1009
	if (bitmap->storage.file)
		wait_event(bitmap->write_wait,
			   atomic_read(&bitmap->pending_writes)==0);
	else
		md_super_wait(bitmap->mddev);

1010
	if (test_bit(BITMAP_WRITE_ERROR, &bitmap->flags))
1011
		bitmap_file_kick(bitmap);
1012
}
1013
EXPORT_SYMBOL(bitmap_unplug);
1014

1015
static void bitmap_set_memory_bits(struct bitmap *bitmap, sector_t offset, int needed);
1016 1017 1018 1019 1020 1021 1022
/* * bitmap_init_from_disk -- called at bitmap_create time to initialize
 * the in-memory bitmap from the on-disk bitmap -- also, sets up the
 * memory mapping of the bitmap file
 * Special cases:
 *   if there's no bitmap file, or if the bitmap file had been
 *   previously kicked from the array, we mark all the bits as
 *   1's in order to cause a full resync.
1023 1024 1025
 *
 * We ignore all bits for sectors that end earlier than 'start'.
 * This is used when reading an out-of-date bitmap...
1026
 */
1027
static int bitmap_init_from_disk(struct bitmap *bitmap, sector_t start)
1028
{
1029
	unsigned long i, chunks, index, oldindex, bit, node_offset = 0;
1030
	struct page *page = NULL;
1031
	unsigned long bit_cnt = 0;
1032
	struct file *file;
1033
	unsigned long offset;
1034 1035
	int outofdate;
	int ret = -ENOSPC;
1036
	void *paddr;
1037
	struct bitmap_storage *store = &bitmap->storage;
1038

1039
	chunks = bitmap->counts.chunks;
1040
	file = store->file;
1041

1042 1043
	if (!file && !bitmap->mddev->bitmap_info.offset) {
		/* No permanent bitmap - fill with '1s'. */
1044 1045
		store->filemap = NULL;
		store->file_pages = 0;
1046 1047
		for (i = 0; i < chunks ; i++) {
			/* if the disk bit is set, set the memory bit */
1048
			int needed = ((sector_t)(i+1) << (bitmap->counts.chunkshift)
1049 1050
				      >= start);
			bitmap_set_memory_bits(bitmap,
1051
					       (sector_t)i << bitmap->counts.chunkshift,
1052 1053 1054 1055
					       needed);
		}
		return 0;
	}
1056

1057
	outofdate = test_bit(BITMAP_STALE, &bitmap->flags);
1058 1059 1060 1061
	if (outofdate)
		printk(KERN_INFO "%s: bitmap file is out of date, doing full "
			"recovery\n", bmname(bitmap));

1062
	if (file && i_size_read(file->f_mapping->host) < store->bytes) {
1063
		printk(KERN_INFO "%s: bitmap file too short %lu < %lu\n",
1064 1065 1066
		       bmname(bitmap),
		       (unsigned long) i_size_read(file->f_mapping->host),
		       store->bytes);
1067
		goto err;
1068
	}
1069

1070
	oldindex = ~0L;
1071
	offset = 0;
1072
	if (!bitmap->mddev->bitmap_info.external)
1073
		offset = sizeof(bitmap_super_t);
1074

1075 1076 1077
	if (mddev_is_clustered(bitmap->mddev))
		node_offset = bitmap->cluster_slot * (DIV_ROUND_UP(store->bytes, PAGE_SIZE));

1078
	for (i = 0; i < chunks; i++) {
1079
		int b;
1080 1081
		index = file_page_index(&bitmap->storage, i);
		bit = file_page_offset(&bitmap->storage, i);
1082
		if (index != oldindex) { /* this is a new page, read it in */
1083
			int count;
1084
			/* unmap the old page, we're done with it */
1085 1086
			if (index == store->file_pages-1)
				count = store->bytes - index * PAGE_SIZE;
1087 1088
			else
				count = PAGE_SIZE;
1089
			page = store->filemap[index];
1090 1091 1092 1093 1094 1095 1096 1097
			if (file)
				ret = read_page(file, index, bitmap,
						count, page);
			else
				ret = read_sb_page(
					bitmap->mddev,
					bitmap->mddev->bitmap_info.offset,
					page,
1098
					index + node_offset, count);
1099 1100

			if (ret)
1101
				goto err;
1102

1103 1104 1105 1106 1107
			oldindex = index;

			if (outofdate) {
				/*
				 * if bitmap is out of date, dirty the
1108
				 * whole page and write it out
1109
				 */
1110
				paddr = kmap_atomic(page);
1111
				memset(paddr + offset, 0xff,
1112
				       PAGE_SIZE - offset);
1113
				kunmap_atomic(paddr);
1114 1115 1116
				write_page(bitmap, page, 1);

				ret = -EIO;
1117 1118
				if (test_bit(BITMAP_WRITE_ERROR,
					     &bitmap->flags))
1119
					goto err;
1120 1121
			}
		}
1122
		paddr = kmap_atomic(page);
1123
		if (test_bit(BITMAP_HOSTENDIAN, &bitmap->flags))
1124
			b = test_bit(bit, paddr);
1125
		else
A
Akinobu Mita 已提交
1126
			b = test_bit_le(bit, paddr);
1127
		kunmap_atomic(paddr);
1128
		if (b) {
1129
			/* if the disk bit is set, set the memory bit */
1130
			int needed = ((sector_t)(i+1) << bitmap->counts.chunkshift
1131 1132
				      >= start);
			bitmap_set_memory_bits(bitmap,
1133
					       (sector_t)i << bitmap->counts.chunkshift,
1134
					       needed);
1135 1136
			bit_cnt++;
		}
1137
		offset = 0;
1138 1139 1140
	}

	printk(KERN_INFO "%s: bitmap initialized from disk: "
1141
	       "read %lu pages, set %lu of %lu bits\n",
1142
	       bmname(bitmap), store->file_pages,
1143
	       bit_cnt, chunks);
1144 1145

	return 0;
1146

1147 1148 1149
 err:
	printk(KERN_INFO "%s: bitmap initialisation failed: %d\n",
	       bmname(bitmap), ret);
1150 1151 1152
	return ret;
}

1153 1154 1155 1156 1157
void bitmap_write_all(struct bitmap *bitmap)
{
	/* We don't actually write all bitmap blocks here,
	 * just flag them as needing to be written
	 */
1158
	int i;
1159

1160
	if (!bitmap || !bitmap->storage.filemap)
1161
		return;
1162
	if (bitmap->storage.file)
1163 1164 1165
		/* Only one copy, so nothing needed */
		return;

1166
	for (i = 0; i < bitmap->storage.file_pages; i++)
1167
		set_page_attr(bitmap, i,
1168
			      BITMAP_PAGE_NEEDWRITE);
1169
	bitmap->allclean = 0;
1170 1171
}

1172 1173
static void bitmap_count_page(struct bitmap_counts *bitmap,
			      sector_t offset, int inc)
1174
{
1175
	sector_t chunk = offset >> bitmap->chunkshift;
1176 1177 1178 1179
	unsigned long page = chunk >> PAGE_COUNTER_SHIFT;
	bitmap->bp[page].count += inc;
	bitmap_checkfree(bitmap, page);
}
1180

1181
static void bitmap_set_pending(struct bitmap_counts *bitmap, sector_t offset)
1182 1183 1184 1185 1186 1187 1188 1189 1190
{
	sector_t chunk = offset >> bitmap->chunkshift;
	unsigned long page = chunk >> PAGE_COUNTER_SHIFT;
	struct bitmap_page *bp = &bitmap->bp[page];

	if (!bp->pending)
		bp->pending = 1;
}

1191
static bitmap_counter_t *bitmap_get_counter(struct bitmap_counts *bitmap,
N
NeilBrown 已提交
1192
					    sector_t offset, sector_t *blocks,
1193 1194 1195 1196 1197 1198 1199
					    int create);

/*
 * bitmap daemon -- periodically wakes up to clean bits and flush pages
 *			out to disk
 */

1200
void bitmap_daemon_work(struct mddev *mddev)
1201
{
1202
	struct bitmap *bitmap;
1203
	unsigned long j;
1204
	unsigned long nextpage;
N
NeilBrown 已提交
1205
	sector_t blocks;
1206
	struct bitmap_counts *counts;
1207

1208 1209 1210
	/* Use a mutex to guard daemon_work against
	 * bitmap_destroy.
	 */
1211
	mutex_lock(&mddev->bitmap_info.mutex);
1212 1213
	bitmap = mddev->bitmap;
	if (bitmap == NULL) {
1214
		mutex_unlock(&mddev->bitmap_info.mutex);
1215
		return;
1216
	}
1217
	if (time_before(jiffies, bitmap->daemon_lastrun
N
NeilBrown 已提交
1218
			+ mddev->bitmap_info.daemon_sleep))
1219 1220
		goto done;

1221
	bitmap->daemon_lastrun = jiffies;
1222
	if (bitmap->allclean) {
N
NeilBrown 已提交
1223
		mddev->thread->timeout = MAX_SCHEDULE_TIMEOUT;
1224
		goto done;
1225 1226
	}
	bitmap->allclean = 1;
1227

1228 1229 1230 1231
	/* Any file-page which is PENDING now needs to be written.
	 * So set NEEDWRITE now, then after we make any last-minute changes
	 * we will write it.
	 */
1232
	for (j = 0; j < bitmap->storage.file_pages; j++)
1233 1234
		if (test_and_clear_page_attr(bitmap, j,
					     BITMAP_PAGE_PENDING))
1235
			set_page_attr(bitmap, j,
1236 1237 1238 1239 1240 1241 1242 1243
				      BITMAP_PAGE_NEEDWRITE);

	if (bitmap->need_sync &&
	    mddev->bitmap_info.external == 0) {
		/* Arrange for superblock update as well as
		 * other changes */
		bitmap_super_t *sb;
		bitmap->need_sync = 0;
1244 1245
		if (bitmap->storage.filemap) {
			sb = kmap_atomic(bitmap->storage.sb_page);
1246 1247 1248
			sb->events_cleared =
				cpu_to_le64(bitmap->events_cleared);
			kunmap_atomic(sb);
1249
			set_page_attr(bitmap, 0,
1250 1251
				      BITMAP_PAGE_NEEDWRITE);
		}
1252 1253 1254 1255
	}
	/* Now look at the bitmap counters and if any are '2' or '1',
	 * decrement and handle accordingly.
	 */
1256 1257
	counts = &bitmap->counts;
	spin_lock_irq(&counts->lock);
1258
	nextpage = 0;
1259
	for (j = 0; j < counts->chunks; j++) {
1260
		bitmap_counter_t *bmc;
1261
		sector_t  block = (sector_t)j << counts->chunkshift;
1262

1263 1264
		if (j == nextpage) {
			nextpage += PAGE_COUNTER_RATIO;
1265
			if (!counts->bp[j >> PAGE_COUNTER_SHIFT].pending) {
1266
				j |= PAGE_COUNTER_MASK;
1267 1268
				continue;
			}
1269
			counts->bp[j >> PAGE_COUNTER_SHIFT].pending = 0;
1270
		}
1271
		bmc = bitmap_get_counter(counts,
1272
					 block,
1273
					 &blocks, 0);
1274 1275

		if (!bmc) {
1276
			j |= PAGE_COUNTER_MASK;
1277 1278 1279 1280 1281
			continue;
		}
		if (*bmc == 1 && !bitmap->need_sync) {
			/* We can clear the bit */
			*bmc = 0;
1282
			bitmap_count_page(counts, block, -1);
1283
			bitmap_file_clear_bit(bitmap, block);
1284 1285
		} else if (*bmc && *bmc <= 2) {
			*bmc = 1;
1286
			bitmap_set_pending(counts, block);
1287
			bitmap->allclean = 0;
1288
		}
1289
	}
1290
	spin_unlock_irq(&counts->lock);
1291

1292 1293 1294 1295 1296 1297 1298 1299
	/* Now start writeout on any page in NEEDWRITE that isn't DIRTY.
	 * DIRTY pages need to be written by bitmap_unplug so it can wait
	 * for them.
	 * If we find any DIRTY page we stop there and let bitmap_unplug
	 * handle all the rest.  This is important in the case where
	 * the first blocking holds the superblock and it has been updated.
	 * We mustn't write any other blocks before the superblock.
	 */
1300 1301 1302 1303
	for (j = 0;
	     j < bitmap->storage.file_pages
		     && !test_bit(BITMAP_STALE, &bitmap->flags);
	     j++) {
1304
		if (test_page_attr(bitmap, j,
1305 1306 1307
				   BITMAP_PAGE_DIRTY))
			/* bitmap_unplug will handle the rest */
			break;
1308 1309
		if (test_and_clear_page_attr(bitmap, j,
					     BITMAP_PAGE_NEEDWRITE)) {
1310
			write_page(bitmap, bitmap->storage.filemap[j], 0);
1311 1312 1313
		}
	}

1314
 done:
1315
	if (bitmap->allclean == 0)
N
NeilBrown 已提交
1316 1317
		mddev->thread->timeout =
			mddev->bitmap_info.daemon_sleep;
1318
	mutex_unlock(&mddev->bitmap_info.mutex);
1319 1320
}

1321
static bitmap_counter_t *bitmap_get_counter(struct bitmap_counts *bitmap,
N
NeilBrown 已提交
1322
					    sector_t offset, sector_t *blocks,
1323
					    int create)
1324 1325
__releases(bitmap->lock)
__acquires(bitmap->lock)
1326 1327 1328 1329 1330
{
	/* If 'create', we might release the lock and reclaim it.
	 * The lock must have been taken with interrupts enabled.
	 * If !create, we don't release the lock.
	 */
1331
	sector_t chunk = offset >> bitmap->chunkshift;
1332 1333 1334
	unsigned long page = chunk >> PAGE_COUNTER_SHIFT;
	unsigned long pageoff = (chunk & PAGE_COUNTER_MASK) << COUNTER_BYTE_SHIFT;
	sector_t csize;
1335
	int err;
1336

1337
	err = bitmap_checkpage(bitmap, page, create, 0);
1338 1339 1340

	if (bitmap->bp[page].hijacked ||
	    bitmap->bp[page].map == NULL)
1341
		csize = ((sector_t)1) << (bitmap->chunkshift +
1342 1343
					  PAGE_COUNTER_SHIFT - 1);
	else
1344
		csize = ((sector_t)1) << bitmap->chunkshift;
1345 1346 1347
	*blocks = csize - (offset & (csize - 1));

	if (err < 0)
1348
		return NULL;
1349

1350 1351 1352 1353 1354 1355 1356 1357
	/* now locked ... */

	if (bitmap->bp[page].hijacked) { /* hijacked pointer */
		/* should we use the first or second counter field
		 * of the hijacked pointer? */
		int hi = (pageoff > PAGE_COUNTER_MASK);
		return  &((bitmap_counter_t *)
			  &bitmap->bp[page].map)[hi];
1358
	} else /* page is allocated */
1359 1360 1361 1362
		return (bitmap_counter_t *)
			&(bitmap->bp[page].map[pageoff]);
}

1363
int bitmap_startwrite(struct bitmap *bitmap, sector_t offset, unsigned long sectors, int behind)
1364
{
1365 1366
	if (!bitmap)
		return 0;
1367 1368

	if (behind) {
1369
		int bw;
1370
		atomic_inc(&bitmap->behind_writes);
1371 1372 1373 1374
		bw = atomic_read(&bitmap->behind_writes);
		if (bw > bitmap->behind_writes_used)
			bitmap->behind_writes_used = bw;

1375 1376
		pr_debug("inc write-behind count %d/%lu\n",
			 bw, bitmap->mddev->bitmap_info.max_write_behind);
1377 1378
	}

1379
	while (sectors) {
N
NeilBrown 已提交
1380
		sector_t blocks;
1381 1382
		bitmap_counter_t *bmc;

1383 1384
		spin_lock_irq(&bitmap->counts.lock);
		bmc = bitmap_get_counter(&bitmap->counts, offset, &blocks, 1);
1385
		if (!bmc) {
1386
			spin_unlock_irq(&bitmap->counts.lock);
1387 1388 1389
			return 0;
		}

1390
		if (unlikely(COUNTER(*bmc) == COUNTER_MAX)) {
1391 1392 1393 1394 1395 1396 1397
			DEFINE_WAIT(__wait);
			/* note that it is safe to do the prepare_to_wait
			 * after the test as long as we do it before dropping
			 * the spinlock.
			 */
			prepare_to_wait(&bitmap->overflow_wait, &__wait,
					TASK_UNINTERRUPTIBLE);
1398
			spin_unlock_irq(&bitmap->counts.lock);
1399
			schedule();
1400 1401 1402 1403
			finish_wait(&bitmap->overflow_wait, &__wait);
			continue;
		}

1404
		switch (*bmc) {
1405 1406
		case 0:
			bitmap_file_set_bit(bitmap, offset);
1407
			bitmap_count_page(&bitmap->counts, offset, 1);
1408 1409 1410 1411
			/* fall through */
		case 1:
			*bmc = 2;
		}
1412

1413 1414
		(*bmc)++;

1415
		spin_unlock_irq(&bitmap->counts.lock);
1416 1417 1418 1419

		offset += blocks;
		if (sectors > blocks)
			sectors -= blocks;
1420 1421
		else
			sectors = 0;
1422 1423 1424
	}
	return 0;
}
1425
EXPORT_SYMBOL(bitmap_startwrite);
1426 1427

void bitmap_endwrite(struct bitmap *bitmap, sector_t offset, unsigned long sectors,
1428
		     int success, int behind)
1429
{
1430 1431
	if (!bitmap)
		return;
1432
	if (behind) {
1433 1434
		if (atomic_dec_and_test(&bitmap->behind_writes))
			wake_up(&bitmap->behind_wait);
1435 1436 1437
		pr_debug("dec write-behind count %d/%lu\n",
			 atomic_read(&bitmap->behind_writes),
			 bitmap->mddev->bitmap_info.max_write_behind);
1438 1439
	}

1440
	while (sectors) {
N
NeilBrown 已提交
1441
		sector_t blocks;
1442 1443 1444
		unsigned long flags;
		bitmap_counter_t *bmc;

1445 1446
		spin_lock_irqsave(&bitmap->counts.lock, flags);
		bmc = bitmap_get_counter(&bitmap->counts, offset, &blocks, 0);
1447
		if (!bmc) {
1448
			spin_unlock_irqrestore(&bitmap->counts.lock, flags);
1449 1450 1451
			return;
		}

1452
		if (success && !bitmap->mddev->degraded &&
1453 1454 1455
		    bitmap->events_cleared < bitmap->mddev->events) {
			bitmap->events_cleared = bitmap->mddev->events;
			bitmap->need_sync = 1;
1456
			sysfs_notify_dirent_safe(bitmap->sysfs_can_clear);
1457 1458
		}

1459
		if (!success && !NEEDED(*bmc))
1460 1461
			*bmc |= NEEDED_MASK;

1462
		if (COUNTER(*bmc) == COUNTER_MAX)
1463 1464
			wake_up(&bitmap->overflow_wait);

1465
		(*bmc)--;
1466
		if (*bmc <= 2) {
1467
			bitmap_set_pending(&bitmap->counts, offset);
1468 1469
			bitmap->allclean = 0;
		}
1470
		spin_unlock_irqrestore(&bitmap->counts.lock, flags);
1471 1472 1473
		offset += blocks;
		if (sectors > blocks)
			sectors -= blocks;
1474 1475
		else
			sectors = 0;
1476 1477
	}
}
1478
EXPORT_SYMBOL(bitmap_endwrite);
1479

N
NeilBrown 已提交
1480
static int __bitmap_start_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks,
1481
			       int degraded)
1482 1483 1484 1485 1486 1487 1488
{
	bitmap_counter_t *bmc;
	int rv;
	if (bitmap == NULL) {/* FIXME or bitmap set as 'failed' */
		*blocks = 1024;
		return 1; /* always resync if no bitmap */
	}
1489 1490
	spin_lock_irq(&bitmap->counts.lock);
	bmc = bitmap_get_counter(&bitmap->counts, offset, blocks, 0);
1491 1492 1493 1494 1495 1496 1497
	rv = 0;
	if (bmc) {
		/* locked */
		if (RESYNC(*bmc))
			rv = 1;
		else if (NEEDED(*bmc)) {
			rv = 1;
1498 1499 1500 1501
			if (!degraded) { /* don't set/clear bits if degraded */
				*bmc |= RESYNC_MASK;
				*bmc &= ~NEEDED_MASK;
			}
1502 1503
		}
	}
1504
	spin_unlock_irq(&bitmap->counts.lock);
1505 1506 1507
	return rv;
}

N
NeilBrown 已提交
1508
int bitmap_start_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks,
1509 1510 1511 1512 1513 1514 1515 1516 1517 1518
		      int degraded)
{
	/* bitmap_start_sync must always report on multiples of whole
	 * pages, otherwise resync (which is very PAGE_SIZE based) will
	 * get confused.
	 * So call __bitmap_start_sync repeatedly (if needed) until
	 * At least PAGE_SIZE>>9 blocks are covered.
	 * Return the 'or' of the result.
	 */
	int rv = 0;
N
NeilBrown 已提交
1519
	sector_t blocks1;
1520 1521 1522 1523 1524 1525 1526 1527 1528 1529

	*blocks = 0;
	while (*blocks < (PAGE_SIZE>>9)) {
		rv |= __bitmap_start_sync(bitmap, offset,
					  &blocks1, degraded);
		offset += blocks1;
		*blocks += blocks1;
	}
	return rv;
}
1530
EXPORT_SYMBOL(bitmap_start_sync);
1531

N
NeilBrown 已提交
1532
void bitmap_end_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks, int aborted)
1533 1534 1535
{
	bitmap_counter_t *bmc;
	unsigned long flags;
1536 1537

	if (bitmap == NULL) {
1538 1539 1540
		*blocks = 1024;
		return;
	}
1541 1542
	spin_lock_irqsave(&bitmap->counts.lock, flags);
	bmc = bitmap_get_counter(&bitmap->counts, offset, blocks, 0);
1543 1544 1545 1546 1547 1548 1549 1550 1551
	if (bmc == NULL)
		goto unlock;
	/* locked */
	if (RESYNC(*bmc)) {
		*bmc &= ~RESYNC_MASK;

		if (!NEEDED(*bmc) && aborted)
			*bmc |= NEEDED_MASK;
		else {
1552
			if (*bmc <= 2) {
1553
				bitmap_set_pending(&bitmap->counts, offset);
1554 1555
				bitmap->allclean = 0;
			}
1556 1557 1558
		}
	}
 unlock:
1559
	spin_unlock_irqrestore(&bitmap->counts.lock, flags);
1560
}
1561
EXPORT_SYMBOL(bitmap_end_sync);
1562 1563 1564 1565 1566 1567 1568 1569

void bitmap_close_sync(struct bitmap *bitmap)
{
	/* Sync has finished, and any bitmap chunks that weren't synced
	 * properly have been aborted.  It remains to us to clear the
	 * RESYNC bit wherever it is still on
	 */
	sector_t sector = 0;
N
NeilBrown 已提交
1570
	sector_t blocks;
N
NeilBrown 已提交
1571 1572
	if (!bitmap)
		return;
1573 1574
	while (sector < bitmap->mddev->resync_max_sectors) {
		bitmap_end_sync(bitmap, sector, &blocks, 0);
N
NeilBrown 已提交
1575 1576 1577
		sector += blocks;
	}
}
1578
EXPORT_SYMBOL(bitmap_close_sync);
N
NeilBrown 已提交
1579

1580
void bitmap_cond_end_sync(struct bitmap *bitmap, sector_t sector, bool force)
N
NeilBrown 已提交
1581 1582
{
	sector_t s = 0;
N
NeilBrown 已提交
1583
	sector_t blocks;
N
NeilBrown 已提交
1584 1585 1586 1587 1588 1589 1590

	if (!bitmap)
		return;
	if (sector == 0) {
		bitmap->last_end_sync = jiffies;
		return;
	}
1591
	if (!force && time_before(jiffies, (bitmap->last_end_sync
1592
				  + bitmap->mddev->bitmap_info.daemon_sleep)))
N
NeilBrown 已提交
1593 1594 1595 1596
		return;
	wait_event(bitmap->mddev->recovery_wait,
		   atomic_read(&bitmap->mddev->recovery_active) == 0);

1597
	bitmap->mddev->curr_resync_completed = sector;
1598
	set_bit(MD_CHANGE_CLEAN, &bitmap->mddev->flags);
1599
	sector &= ~((1ULL << bitmap->counts.chunkshift) - 1);
N
NeilBrown 已提交
1600 1601 1602 1603
	s = 0;
	while (s < sector && s < bitmap->mddev->resync_max_sectors) {
		bitmap_end_sync(bitmap, s, &blocks, 0);
		s += blocks;
1604
	}
N
NeilBrown 已提交
1605
	bitmap->last_end_sync = jiffies;
1606
	sysfs_notify(&bitmap->mddev->kobj, NULL, "sync_completed");
1607
}
1608
EXPORT_SYMBOL(bitmap_cond_end_sync);
1609

1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630
void bitmap_sync_with_cluster(struct mddev *mddev,
			      sector_t old_lo, sector_t old_hi,
			      sector_t new_lo, sector_t new_hi)
{
	struct bitmap *bitmap = mddev->bitmap;
	sector_t sector, blocks = 0;

	for (sector = old_lo; sector < new_lo; ) {
		bitmap_end_sync(bitmap, sector, &blocks, 0);
		sector += blocks;
	}
	WARN((blocks > new_lo) && old_lo, "alignment is not correct for lo\n");

	for (sector = old_hi; sector < new_hi; ) {
		bitmap_start_sync(bitmap, sector, &blocks, 0);
		sector += blocks;
	}
	WARN((blocks > new_hi) && old_hi, "alignment is not correct for hi\n");
}
EXPORT_SYMBOL(bitmap_sync_with_cluster);

1631
static void bitmap_set_memory_bits(struct bitmap *bitmap, sector_t offset, int needed)
1632 1633
{
	/* For each chunk covered by any of these sectors, set the
1634
	 * counter to 2 and possibly set resync_needed.  They should all
1635 1636
	 * be 0 at this point
	 */
1637

N
NeilBrown 已提交
1638
	sector_t secs;
1639
	bitmap_counter_t *bmc;
1640 1641
	spin_lock_irq(&bitmap->counts.lock);
	bmc = bitmap_get_counter(&bitmap->counts, offset, &secs, 1);
1642
	if (!bmc) {
1643
		spin_unlock_irq(&bitmap->counts.lock);
1644
		return;
1645
	}
1646
	if (!*bmc) {
1647
		*bmc = 2;
1648 1649
		bitmap_count_page(&bitmap->counts, offset, 1);
		bitmap_set_pending(&bitmap->counts, offset);
1650
		bitmap->allclean = 0;
1651
	}
1652 1653
	if (needed)
		*bmc |= NEEDED_MASK;
1654
	spin_unlock_irq(&bitmap->counts.lock);
1655 1656
}

1657 1658 1659 1660 1661 1662
/* dirty the memory and file bits for bitmap chunks "s" to "e" */
void bitmap_dirty_bits(struct bitmap *bitmap, unsigned long s, unsigned long e)
{
	unsigned long chunk;

	for (chunk = s; chunk <= e; chunk++) {
1663
		sector_t sec = (sector_t)chunk << bitmap->counts.chunkshift;
1664 1665
		bitmap_set_memory_bits(bitmap, sec, 1);
		bitmap_file_set_bit(bitmap, sec);
1666 1667 1668 1669 1670 1671
		if (sec < bitmap->mddev->recovery_cp)
			/* We are asserting that the array is dirty,
			 * so move the recovery_cp address back so
			 * that it is obvious that it is dirty
			 */
			bitmap->mddev->recovery_cp = sec;
1672 1673 1674
	}
}

1675 1676 1677
/*
 * flush out any pending updates
 */
1678
void bitmap_flush(struct mddev *mddev)
1679 1680
{
	struct bitmap *bitmap = mddev->bitmap;
1681
	long sleep;
1682 1683 1684 1685 1686 1687 1688

	if (!bitmap) /* there was no bitmap */
		return;

	/* run the daemon_work three time to ensure everything is flushed
	 * that can be
	 */
1689
	sleep = mddev->bitmap_info.daemon_sleep * 2;
1690
	bitmap->daemon_lastrun -= sleep;
1691
	bitmap_daemon_work(mddev);
1692
	bitmap->daemon_lastrun -= sleep;
1693
	bitmap_daemon_work(mddev);
1694
	bitmap->daemon_lastrun -= sleep;
1695
	bitmap_daemon_work(mddev);
1696 1697 1698
	bitmap_update_sb(bitmap);
}

1699 1700 1701
/*
 * free memory that was allocated
 */
1702
static void bitmap_free(struct bitmap *bitmap)
1703 1704 1705 1706 1707 1708 1709
{
	unsigned long k, pages;
	struct bitmap_page *bp;

	if (!bitmap) /* there was no bitmap */
		return;

1710 1711 1712
	if (bitmap->sysfs_can_clear)
		sysfs_put(bitmap->sysfs_can_clear);

1713 1714
	if (mddev_is_clustered(bitmap->mddev) && bitmap->mddev->cluster_info &&
		bitmap->cluster_slot == md_cluster_ops->slot_number(bitmap->mddev))
1715 1716
		md_cluster_stop(bitmap->mddev);

1717 1718 1719 1720 1721 1722
	/* Shouldn't be needed - but just in case.... */
	wait_event(bitmap->write_wait,
		   atomic_read(&bitmap->pending_writes) == 0);

	/* release the bitmap file  */
	bitmap_file_unmap(&bitmap->storage);
1723

1724 1725
	bp = bitmap->counts.bp;
	pages = bitmap->counts.pages;
1726 1727 1728 1729 1730 1731 1732 1733 1734 1735

	/* free all allocated memory */

	if (bp) /* deallocate the page memory */
		for (k = 0; k < pages; k++)
			if (bp[k].map && !bp[k].hijacked)
				kfree(bp[k].map);
	kfree(bp);
	kfree(bitmap);
}
1736

1737
void bitmap_destroy(struct mddev *mddev)
1738 1739 1740 1741 1742 1743
{
	struct bitmap *bitmap = mddev->bitmap;

	if (!bitmap) /* there was no bitmap */
		return;

1744
	mutex_lock(&mddev->bitmap_info.mutex);
1745
	spin_lock(&mddev->lock);
1746
	mddev->bitmap = NULL; /* disconnect from the md device */
1747
	spin_unlock(&mddev->lock);
1748
	mutex_unlock(&mddev->bitmap_info.mutex);
1749 1750
	if (mddev->thread)
		mddev->thread->timeout = MAX_SCHEDULE_TIMEOUT;
1751 1752 1753

	bitmap_free(bitmap);
}
1754 1755 1756 1757

/*
 * initialize the bitmap structure
 * if this returns an error, bitmap_destroy must be called to do clean up
1758
 * once mddev->bitmap is set
1759
 */
1760
struct bitmap *bitmap_create(struct mddev *mddev, int slot)
1761 1762
{
	struct bitmap *bitmap;
1763
	sector_t blocks = mddev->resync_max_sectors;
1764
	struct file *file = mddev->bitmap_info.file;
1765
	int err;
1766
	struct kernfs_node *bm = NULL;
1767

A
Alexey Dobriyan 已提交
1768
	BUILD_BUG_ON(sizeof(bitmap_super_t) != 256);
1769

1770
	BUG_ON(file && mddev->bitmap_info.offset);
1771

1772
	bitmap = kzalloc(sizeof(*bitmap), GFP_KERNEL);
1773
	if (!bitmap)
1774
		return ERR_PTR(-ENOMEM);
1775

1776
	spin_lock_init(&bitmap->counts.lock);
1777 1778
	atomic_set(&bitmap->pending_writes, 0);
	init_waitqueue_head(&bitmap->write_wait);
1779
	init_waitqueue_head(&bitmap->overflow_wait);
1780
	init_waitqueue_head(&bitmap->behind_wait);
1781

1782
	bitmap->mddev = mddev;
1783
	bitmap->cluster_slot = slot;
1784

1785
	if (mddev->kobj.sd)
T
Tejun Heo 已提交
1786
		bm = sysfs_get_dirent(mddev->kobj.sd, "bitmap");
1787
	if (bm) {
T
Tejun Heo 已提交
1788
		bitmap->sysfs_can_clear = sysfs_get_dirent(bm, "can_clear");
1789 1790 1791 1792
		sysfs_put(bm);
	} else
		bitmap->sysfs_can_clear = NULL;

1793
	bitmap->storage.file = file;
1794 1795
	if (file) {
		get_file(file);
1796 1797 1798 1799
		/* As future accesses to this file will use bmap,
		 * and bypass the page cache, we must sync the file
		 * first.
		 */
1800
		vfs_fsync(file, 1);
1801
	}
1802
	/* read superblock from bitmap file (this sets mddev->bitmap_info.chunksize) */
1803 1804 1805 1806 1807 1808 1809 1810 1811 1812
	if (!mddev->bitmap_info.external) {
		/*
		 * If 'MD_ARRAY_FIRST_USE' is set, then device-mapper is
		 * instructing us to create a new on-disk bitmap instance.
		 */
		if (test_and_clear_bit(MD_ARRAY_FIRST_USE, &mddev->flags))
			err = bitmap_new_disk_sb(bitmap);
		else
			err = bitmap_read_sb(bitmap);
	} else {
1813 1814 1815 1816 1817 1818 1819
		err = 0;
		if (mddev->bitmap_info.chunksize == 0 ||
		    mddev->bitmap_info.daemon_sleep == 0)
			/* chunksize and time_base need to be
			 * set first. */
			err = -EINVAL;
	}
1820
	if (err)
1821
		goto error;
1822

1823
	bitmap->daemon_lastrun = jiffies;
1824 1825
	err = bitmap_resize(bitmap, blocks, mddev->bitmap_info.chunksize, 1);
	if (err)
1826
		goto error;
1827

1828
	printk(KERN_INFO "created bitmap (%lu pages) for device %s\n",
1829
	       bitmap->counts.pages, bmname(bitmap));
1830

1831 1832 1833
	err = test_bit(BITMAP_WRITE_ERROR, &bitmap->flags) ? -EIO : 0;
	if (err)
		goto error;
1834

1835
	return bitmap;
1836 1837
 error:
	bitmap_free(bitmap);
1838
	return ERR_PTR(err);
1839 1840
}

1841
int bitmap_load(struct mddev *mddev)
1842 1843
{
	int err = 0;
1844
	sector_t start = 0;
1845 1846 1847 1848 1849 1850
	sector_t sector = 0;
	struct bitmap *bitmap = mddev->bitmap;

	if (!bitmap)
		goto out;

1851 1852 1853
	if (mddev_is_clustered(mddev))
		md_cluster_ops->load_bitmaps(mddev, mddev->bitmap_info.nodes);

1854 1855 1856 1857 1858 1859
	/* Clear out old bitmap info first:  Either there is none, or we
	 * are resuming after someone else has possibly changed things,
	 * so we should forget old cached info.
	 * All chunks should be clean, but some might need_sync.
	 */
	while (sector < mddev->resync_max_sectors) {
N
NeilBrown 已提交
1860
		sector_t blocks;
1861 1862 1863 1864 1865
		bitmap_start_sync(bitmap, sector, &blocks, 0);
		sector += blocks;
	}
	bitmap_close_sync(bitmap);

1866 1867 1868 1869 1870 1871
	if (mddev->degraded == 0
	    || bitmap->events_cleared == mddev->events)
		/* no need to keep dirty bits to optimise a
		 * re-add of a missing device */
		start = mddev->recovery_cp;

1872
	mutex_lock(&mddev->bitmap_info.mutex);
1873
	err = bitmap_init_from_disk(bitmap, start);
1874
	mutex_unlock(&mddev->bitmap_info.mutex);
1875

1876
	if (err)
1877
		goto out;
1878
	clear_bit(BITMAP_STALE, &bitmap->flags);
1879 1880 1881

	/* Kick recovery in case any bits were set */
	set_bit(MD_RECOVERY_NEEDED, &bitmap->mddev->recovery);
1882

1883
	mddev->thread->timeout = mddev->bitmap_info.daemon_sleep;
1884
	md_wakeup_thread(mddev->thread);
1885

1886 1887
	bitmap_update_sb(bitmap);

1888
	if (test_bit(BITMAP_WRITE_ERROR, &bitmap->flags))
1889 1890
		err = -EIO;
out:
1891
	return err;
1892
}
1893
EXPORT_SYMBOL_GPL(bitmap_load);
1894

1895 1896 1897 1898
/* Loads the bitmap associated with slot and copies the resync information
 * to our bitmap
 */
int bitmap_copy_from_slot(struct mddev *mddev, int slot,
1899
		sector_t *low, sector_t *high, bool clear_bits)
1900 1901 1902 1903 1904 1905
{
	int rv = 0, i, j;
	sector_t block, lo = 0, hi = 0;
	struct bitmap_counts *counts;
	struct bitmap *bitmap = bitmap_create(mddev, slot);

S
Shaohua Li 已提交
1906
	if (IS_ERR(bitmap))
1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925
		return PTR_ERR(bitmap);

	rv = bitmap_init_from_disk(bitmap, 0);
	if (rv)
		goto err;

	counts = &bitmap->counts;
	for (j = 0; j < counts->chunks; j++) {
		block = (sector_t)j << counts->chunkshift;
		if (bitmap_file_test_bit(bitmap, block)) {
			if (!lo)
				lo = block;
			hi = block;
			bitmap_file_clear_bit(bitmap, block);
			bitmap_set_memory_bits(mddev->bitmap, block, 1);
			bitmap_file_set_bit(mddev->bitmap, block);
		}
	}

1926 1927
	if (clear_bits) {
		bitmap_update_sb(bitmap);
1928 1929
		/* BITMAP_PAGE_PENDING is set, but bitmap_unplug needs
		 * BITMAP_PAGE_DIRTY or _NEEDWRITE to write ... */
1930
		for (i = 0; i < bitmap->storage.file_pages; i++)
1931 1932
			if (test_page_attr(bitmap, i, BITMAP_PAGE_PENDING))
				set_page_attr(bitmap, i, BITMAP_PAGE_NEEDWRITE);
1933 1934
		bitmap_unplug(bitmap);
	}
1935
	bitmap_unplug(mddev->bitmap);
1936 1937 1938 1939 1940 1941 1942 1943 1944
	*low = lo;
	*high = hi;
err:
	bitmap_free(bitmap);
	return rv;
}
EXPORT_SYMBOL_GPL(bitmap_copy_from_slot);


1945 1946 1947
void bitmap_status(struct seq_file *seq, struct bitmap *bitmap)
{
	unsigned long chunk_kb;
1948
	struct bitmap_counts *counts;
1949 1950 1951 1952

	if (!bitmap)
		return;

1953 1954
	counts = &bitmap->counts;

1955 1956 1957
	chunk_kb = bitmap->mddev->bitmap_info.chunksize >> 10;
	seq_printf(seq, "bitmap: %lu/%lu pages [%luKB], "
		   "%lu%s chunk",
1958 1959 1960
		   counts->pages - counts->missing_pages,
		   counts->pages,
		   (counts->pages - counts->missing_pages)
1961 1962 1963
		   << (PAGE_SHIFT - 10),
		   chunk_kb ? chunk_kb : bitmap->mddev->bitmap_info.chunksize,
		   chunk_kb ? "KB" : "B");
1964
	if (bitmap->storage.file) {
1965
		seq_printf(seq, ", file: ");
M
Miklos Szeredi 已提交
1966
		seq_file_path(seq, bitmap->storage.file, " \t\n");
1967 1968 1969 1970 1971
	}

	seq_printf(seq, "\n");
}

1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028
int bitmap_resize(struct bitmap *bitmap, sector_t blocks,
		  int chunksize, int init)
{
	/* If chunk_size is 0, choose an appropriate chunk size.
	 * Then possibly allocate new storage space.
	 * Then quiesce, copy bits, replace bitmap, and re-start
	 *
	 * This function is called both to set up the initial bitmap
	 * and to resize the bitmap while the array is active.
	 * If this happens as a result of the array being resized,
	 * chunksize will be zero, and we need to choose a suitable
	 * chunksize, otherwise we use what we are given.
	 */
	struct bitmap_storage store;
	struct bitmap_counts old_counts;
	unsigned long chunks;
	sector_t block;
	sector_t old_blocks, new_blocks;
	int chunkshift;
	int ret = 0;
	long pages;
	struct bitmap_page *new_bp;

	if (chunksize == 0) {
		/* If there is enough space, leave the chunk size unchanged,
		 * else increase by factor of two until there is enough space.
		 */
		long bytes;
		long space = bitmap->mddev->bitmap_info.space;

		if (space == 0) {
			/* We don't know how much space there is, so limit
			 * to current size - in sectors.
			 */
			bytes = DIV_ROUND_UP(bitmap->counts.chunks, 8);
			if (!bitmap->mddev->bitmap_info.external)
				bytes += sizeof(bitmap_super_t);
			space = DIV_ROUND_UP(bytes, 512);
			bitmap->mddev->bitmap_info.space = space;
		}
		chunkshift = bitmap->counts.chunkshift;
		chunkshift--;
		do {
			/* 'chunkshift' is shift from block size to chunk size */
			chunkshift++;
			chunks = DIV_ROUND_UP_SECTOR_T(blocks, 1 << chunkshift);
			bytes = DIV_ROUND_UP(chunks, 8);
			if (!bitmap->mddev->bitmap_info.external)
				bytes += sizeof(bitmap_super_t);
		} while (bytes > (space << 9));
	} else
		chunkshift = ffz(~chunksize) - BITMAP_BLOCK_SHIFT;

	chunks = DIV_ROUND_UP_SECTOR_T(blocks, 1 << chunkshift);
	memset(&store, 0, sizeof(store));
	if (bitmap->mddev->bitmap_info.offset || bitmap->mddev->bitmap_info.file)
		ret = bitmap_storage_alloc(&store, chunks,
2029
					   !bitmap->mddev->bitmap_info.external,
2030 2031
					   mddev_is_clustered(bitmap->mddev)
					   ? bitmap->cluster_slot : 0);
2032 2033
	if (ret) {
		bitmap_file_unmap(&store);
2034
		goto err;
2035
	}
2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071

	pages = DIV_ROUND_UP(chunks, PAGE_COUNTER_RATIO);

	new_bp = kzalloc(pages * sizeof(*new_bp), GFP_KERNEL);
	ret = -ENOMEM;
	if (!new_bp) {
		bitmap_file_unmap(&store);
		goto err;
	}

	if (!init)
		bitmap->mddev->pers->quiesce(bitmap->mddev, 1);

	store.file = bitmap->storage.file;
	bitmap->storage.file = NULL;

	if (store.sb_page && bitmap->storage.sb_page)
		memcpy(page_address(store.sb_page),
		       page_address(bitmap->storage.sb_page),
		       sizeof(bitmap_super_t));
	bitmap_file_unmap(&bitmap->storage);
	bitmap->storage = store;

	old_counts = bitmap->counts;
	bitmap->counts.bp = new_bp;
	bitmap->counts.pages = pages;
	bitmap->counts.missing_pages = pages;
	bitmap->counts.chunkshift = chunkshift;
	bitmap->counts.chunks = chunks;
	bitmap->mddev->bitmap_info.chunksize = 1 << (chunkshift +
						     BITMAP_BLOCK_SHIFT);

	blocks = min(old_counts.chunks << old_counts.chunkshift,
		     chunks << chunkshift);

	spin_lock_irq(&bitmap->counts.lock);
2072 2073 2074 2075 2076 2077 2078 2079 2080 2081
	/* For cluster raid, need to pre-allocate bitmap */
	if (mddev_is_clustered(bitmap->mddev)) {
		unsigned long page;
		for (page = 0; page < pages; page++) {
			ret = bitmap_checkpage(&bitmap->counts, page, 1, 1);
			if (ret) {
				unsigned long k;

				/* deallocate the page memory */
				for (k = 0; k < page; k++) {
2082
					kfree(new_bp[k].map);
2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100
				}

				/* restore some fields from old_counts */
				bitmap->counts.bp = old_counts.bp;
				bitmap->counts.pages = old_counts.pages;
				bitmap->counts.missing_pages = old_counts.pages;
				bitmap->counts.chunkshift = old_counts.chunkshift;
				bitmap->counts.chunks = old_counts.chunks;
				bitmap->mddev->bitmap_info.chunksize = 1 << (old_counts.chunkshift +
									     BITMAP_BLOCK_SHIFT);
				blocks = old_counts.chunks << old_counts.chunkshift;
				pr_err("Could not pre-allocate in-memory bitmap for cluster raid\n");
				break;
			} else
				bitmap->counts.bp[page].count += 1;
		}
	}

2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168
	for (block = 0; block < blocks; ) {
		bitmap_counter_t *bmc_old, *bmc_new;
		int set;

		bmc_old = bitmap_get_counter(&old_counts, block,
					     &old_blocks, 0);
		set = bmc_old && NEEDED(*bmc_old);

		if (set) {
			bmc_new = bitmap_get_counter(&bitmap->counts, block,
						     &new_blocks, 1);
			if (*bmc_new == 0) {
				/* need to set on-disk bits too. */
				sector_t end = block + new_blocks;
				sector_t start = block >> chunkshift;
				start <<= chunkshift;
				while (start < end) {
					bitmap_file_set_bit(bitmap, block);
					start += 1 << chunkshift;
				}
				*bmc_new = 2;
				bitmap_count_page(&bitmap->counts,
						  block, 1);
				bitmap_set_pending(&bitmap->counts,
						   block);
			}
			*bmc_new |= NEEDED_MASK;
			if (new_blocks < old_blocks)
				old_blocks = new_blocks;
		}
		block += old_blocks;
	}

	if (!init) {
		int i;
		while (block < (chunks << chunkshift)) {
			bitmap_counter_t *bmc;
			bmc = bitmap_get_counter(&bitmap->counts, block,
						 &new_blocks, 1);
			if (bmc) {
				/* new space.  It needs to be resynced, so
				 * we set NEEDED_MASK.
				 */
				if (*bmc == 0) {
					*bmc = NEEDED_MASK | 2;
					bitmap_count_page(&bitmap->counts,
							  block, 1);
					bitmap_set_pending(&bitmap->counts,
							   block);
				}
			}
			block += new_blocks;
		}
		for (i = 0; i < bitmap->storage.file_pages; i++)
			set_page_attr(bitmap, i, BITMAP_PAGE_DIRTY);
	}
	spin_unlock_irq(&bitmap->counts.lock);

	if (!init) {
		bitmap_unplug(bitmap);
		bitmap->mddev->pers->quiesce(bitmap->mddev, 0);
	}
	ret = 0;
err:
	return ret;
}
EXPORT_SYMBOL_GPL(bitmap_resize);

2169
static ssize_t
2170
location_show(struct mddev *mddev, char *page)
2171 2172
{
	ssize_t len;
2173
	if (mddev->bitmap_info.file)
2174
		len = sprintf(page, "file");
2175
	else if (mddev->bitmap_info.offset)
2176
		len = sprintf(page, "%+lld", (long long)mddev->bitmap_info.offset);
2177
	else
2178 2179 2180 2181 2182 2183
		len = sprintf(page, "none");
	len += sprintf(page+len, "\n");
	return len;
}

static ssize_t
2184
location_store(struct mddev *mddev, const char *buf, size_t len)
2185
{
2186
	int rv;
2187

2188 2189 2190
	rv = mddev_lock(mddev);
	if (rv)
		return rv;
2191
	if (mddev->pers) {
2192 2193 2194 2195 2196 2197 2198 2199
		if (!mddev->pers->quiesce) {
			rv = -EBUSY;
			goto out;
		}
		if (mddev->recovery || mddev->sync_thread) {
			rv = -EBUSY;
			goto out;
		}
2200 2201 2202 2203 2204
	}

	if (mddev->bitmap || mddev->bitmap_info.file ||
	    mddev->bitmap_info.offset) {
		/* bitmap already configured.  Only option is to clear it */
2205 2206 2207 2208
		if (strncmp(buf, "none", 4) != 0) {
			rv = -EBUSY;
			goto out;
		}
2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226
		if (mddev->pers) {
			mddev->pers->quiesce(mddev, 1);
			bitmap_destroy(mddev);
			mddev->pers->quiesce(mddev, 0);
		}
		mddev->bitmap_info.offset = 0;
		if (mddev->bitmap_info.file) {
			struct file *f = mddev->bitmap_info.file;
			mddev->bitmap_info.file = NULL;
			fput(f);
		}
	} else {
		/* No bitmap, OK to set a location */
		long long offset;
		if (strncmp(buf, "none", 4) == 0)
			/* nothing to be done */;
		else if (strncmp(buf, "file:", 5) == 0) {
			/* Not supported yet */
2227 2228
			rv = -EINVAL;
			goto out;
2229 2230
		} else {
			if (buf[0] == '+')
2231
				rv = kstrtoll(buf+1, 10, &offset);
2232
			else
2233
				rv = kstrtoll(buf, 10, &offset);
2234
			if (rv)
2235 2236 2237 2238 2239
				goto out;
			if (offset == 0) {
				rv = -EINVAL;
				goto out;
			}
2240 2241
			if (mddev->bitmap_info.external == 0 &&
			    mddev->major_version == 0 &&
2242 2243 2244 2245
			    offset != mddev->bitmap_info.default_offset) {
				rv = -EINVAL;
				goto out;
			}
2246 2247
			mddev->bitmap_info.offset = offset;
			if (mddev->pers) {
2248
				struct bitmap *bitmap;
2249
				mddev->pers->quiesce(mddev, 1);
2250 2251 2252 2253 2254
				bitmap = bitmap_create(mddev, -1);
				if (IS_ERR(bitmap))
					rv = PTR_ERR(bitmap);
				else {
					mddev->bitmap = bitmap;
2255
					rv = bitmap_load(mddev);
2256
					if (rv)
2257
						mddev->bitmap_info.offset = 0;
2258 2259
				}
				mddev->pers->quiesce(mddev, 0);
2260 2261
				if (rv) {
					bitmap_destroy(mddev);
2262
					goto out;
2263
				}
2264 2265 2266 2267 2268 2269 2270 2271 2272 2273
			}
		}
	}
	if (!mddev->external) {
		/* Ensure new bitmap info is stored in
		 * metadata promptly.
		 */
		set_bit(MD_CHANGE_DEVS, &mddev->flags);
		md_wakeup_thread(mddev->thread);
	}
2274 2275 2276 2277 2278
	rv = 0;
out:
	mddev_unlock(mddev);
	if (rv)
		return rv;
2279 2280 2281 2282 2283 2284
	return len;
}

static struct md_sysfs_entry bitmap_location =
__ATTR(location, S_IRUGO|S_IWUSR, location_show, location_store);

2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308
/* 'bitmap/space' is the space available at 'location' for the
 * bitmap.  This allows the kernel to know when it is safe to
 * resize the bitmap to match a resized array.
 */
static ssize_t
space_show(struct mddev *mddev, char *page)
{
	return sprintf(page, "%lu\n", mddev->bitmap_info.space);
}

static ssize_t
space_store(struct mddev *mddev, const char *buf, size_t len)
{
	unsigned long sectors;
	int rv;

	rv = kstrtoul(buf, 10, &sectors);
	if (rv)
		return rv;

	if (sectors == 0)
		return -EINVAL;

	if (mddev->bitmap &&
2309
	    sectors < (mddev->bitmap->storage.bytes + 511) >> 9)
2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321
		return -EFBIG; /* Bitmap is too big for this small space */

	/* could make sure it isn't too big, but that isn't really
	 * needed - user-space should be careful.
	 */
	mddev->bitmap_info.space = sectors;
	return len;
}

static struct md_sysfs_entry bitmap_space =
__ATTR(space, S_IRUGO|S_IWUSR, space_show, space_store);

2322
static ssize_t
2323
timeout_show(struct mddev *mddev, char *page)
2324 2325 2326 2327
{
	ssize_t len;
	unsigned long secs = mddev->bitmap_info.daemon_sleep / HZ;
	unsigned long jifs = mddev->bitmap_info.daemon_sleep % HZ;
2328

2329 2330 2331 2332 2333 2334 2335 2336
	len = sprintf(page, "%lu", secs);
	if (jifs)
		len += sprintf(page+len, ".%03u", jiffies_to_msecs(jifs));
	len += sprintf(page+len, "\n");
	return len;
}

static ssize_t
2337
timeout_store(struct mddev *mddev, const char *buf, size_t len)
2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372
{
	/* timeout can be set at any time */
	unsigned long timeout;
	int rv = strict_strtoul_scaled(buf, &timeout, 4);
	if (rv)
		return rv;

	/* just to make sure we don't overflow... */
	if (timeout >= LONG_MAX / HZ)
		return -EINVAL;

	timeout = timeout * HZ / 10000;

	if (timeout >= MAX_SCHEDULE_TIMEOUT)
		timeout = MAX_SCHEDULE_TIMEOUT-1;
	if (timeout < 1)
		timeout = 1;
	mddev->bitmap_info.daemon_sleep = timeout;
	if (mddev->thread) {
		/* if thread->timeout is MAX_SCHEDULE_TIMEOUT, then
		 * the bitmap is all clean and we don't need to
		 * adjust the timeout right now
		 */
		if (mddev->thread->timeout < MAX_SCHEDULE_TIMEOUT) {
			mddev->thread->timeout = timeout;
			md_wakeup_thread(mddev->thread);
		}
	}
	return len;
}

static struct md_sysfs_entry bitmap_timeout =
__ATTR(time_base, S_IRUGO|S_IWUSR, timeout_show, timeout_store);

static ssize_t
2373
backlog_show(struct mddev *mddev, char *page)
2374 2375 2376 2377 2378
{
	return sprintf(page, "%lu\n", mddev->bitmap_info.max_write_behind);
}

static ssize_t
2379
backlog_store(struct mddev *mddev, const char *buf, size_t len)
2380 2381
{
	unsigned long backlog;
2382
	int rv = kstrtoul(buf, 10, &backlog);
2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394
	if (rv)
		return rv;
	if (backlog > COUNTER_MAX)
		return -EINVAL;
	mddev->bitmap_info.max_write_behind = backlog;
	return len;
}

static struct md_sysfs_entry bitmap_backlog =
__ATTR(backlog, S_IRUGO|S_IWUSR, backlog_show, backlog_store);

static ssize_t
2395
chunksize_show(struct mddev *mddev, char *page)
2396 2397 2398 2399 2400
{
	return sprintf(page, "%lu\n", mddev->bitmap_info.chunksize);
}

static ssize_t
2401
chunksize_store(struct mddev *mddev, const char *buf, size_t len)
2402 2403 2404 2405 2406 2407
{
	/* Can only be changed when no bitmap is active */
	int rv;
	unsigned long csize;
	if (mddev->bitmap)
		return -EBUSY;
2408
	rv = kstrtoul(buf, 10, &csize);
2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420
	if (rv)
		return rv;
	if (csize < 512 ||
	    !is_power_of_2(csize))
		return -EINVAL;
	mddev->bitmap_info.chunksize = csize;
	return len;
}

static struct md_sysfs_entry bitmap_chunksize =
__ATTR(chunksize, S_IRUGO|S_IWUSR, chunksize_show, chunksize_store);

2421
static ssize_t metadata_show(struct mddev *mddev, char *page)
2422
{
G
Goldwyn Rodrigues 已提交
2423 2424
	if (mddev_is_clustered(mddev))
		return sprintf(page, "clustered\n");
2425 2426 2427 2428
	return sprintf(page, "%s\n", (mddev->bitmap_info.external
				      ? "external" : "internal"));
}

2429
static ssize_t metadata_store(struct mddev *mddev, const char *buf, size_t len)
2430 2431 2432 2433 2434 2435 2436
{
	if (mddev->bitmap ||
	    mddev->bitmap_info.file ||
	    mddev->bitmap_info.offset)
		return -EBUSY;
	if (strncmp(buf, "external", 8) == 0)
		mddev->bitmap_info.external = 1;
G
Goldwyn Rodrigues 已提交
2437 2438
	else if ((strncmp(buf, "internal", 8) == 0) ||
			(strncmp(buf, "clustered", 9) == 0))
2439 2440 2441 2442 2443 2444 2445 2446 2447
		mddev->bitmap_info.external = 0;
	else
		return -EINVAL;
	return len;
}

static struct md_sysfs_entry bitmap_metadata =
__ATTR(metadata, S_IRUGO|S_IWUSR, metadata_show, metadata_store);

2448
static ssize_t can_clear_show(struct mddev *mddev, char *page)
2449 2450
{
	int len;
2451
	spin_lock(&mddev->lock);
2452 2453 2454 2455 2456
	if (mddev->bitmap)
		len = sprintf(page, "%s\n", (mddev->bitmap->need_sync ?
					     "false" : "true"));
	else
		len = sprintf(page, "\n");
2457
	spin_unlock(&mddev->lock);
2458 2459 2460
	return len;
}

2461
static ssize_t can_clear_store(struct mddev *mddev, const char *buf, size_t len)
2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478
{
	if (mddev->bitmap == NULL)
		return -ENOENT;
	if (strncmp(buf, "false", 5) == 0)
		mddev->bitmap->need_sync = 1;
	else if (strncmp(buf, "true", 4) == 0) {
		if (mddev->degraded)
			return -EBUSY;
		mddev->bitmap->need_sync = 0;
	} else
		return -EINVAL;
	return len;
}

static struct md_sysfs_entry bitmap_can_clear =
__ATTR(can_clear, S_IRUGO|S_IWUSR, can_clear_show, can_clear_store);

2479
static ssize_t
2480
behind_writes_used_show(struct mddev *mddev, char *page)
2481
{
2482 2483
	ssize_t ret;
	spin_lock(&mddev->lock);
2484
	if (mddev->bitmap == NULL)
2485 2486 2487 2488 2489 2490
		ret = sprintf(page, "0\n");
	else
		ret = sprintf(page, "%lu\n",
			      mddev->bitmap->behind_writes_used);
	spin_unlock(&mddev->lock);
	return ret;
2491 2492 2493
}

static ssize_t
2494
behind_writes_used_reset(struct mddev *mddev, const char *buf, size_t len)
2495 2496 2497 2498 2499 2500 2501 2502 2503 2504
{
	if (mddev->bitmap)
		mddev->bitmap->behind_writes_used = 0;
	return len;
}

static struct md_sysfs_entry max_backlog_used =
__ATTR(max_backlog_used, S_IRUGO | S_IWUSR,
       behind_writes_used_show, behind_writes_used_reset);

2505 2506
static struct attribute *md_bitmap_attrs[] = {
	&bitmap_location.attr,
2507
	&bitmap_space.attr,
2508 2509 2510
	&bitmap_timeout.attr,
	&bitmap_backlog.attr,
	&bitmap_chunksize.attr,
2511 2512
	&bitmap_metadata.attr,
	&bitmap_can_clear.attr,
2513
	&max_backlog_used.attr,
2514 2515 2516 2517 2518 2519 2520
	NULL
};
struct attribute_group md_bitmap_group = {
	.name = "bitmap",
	.attrs = md_bitmap_attrs,
};