bitmap.c 55.8 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 49
static int bitmap_checkpage(struct bitmap *bitmap,
			    unsigned long page, int create)
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
	mappage = kzalloc(PAGE_SIZE, GFP_NOIO);
76 77 78
	spin_lock_irq(&bitmap->lock);

	if (mappage == NULL) {
79 80
		pr_debug("%s: bitmap map page allocation failed, hijacking\n",
			 bmname(bitmap));
81 82 83 84
		/* 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;
85 86
	} else if (bitmap->bp[page].map ||
		   bitmap->bp[page].hijacked) {
87
		/* somebody beat us to getting the page */
88
		kfree(mappage);
89
		return 0;
90
	} else {
91

92
		/* no page was in place and we have one, so install it */
93

94 95 96
		bitmap->bp[page].map = mappage;
		bitmap->missing_pages--;
	}
97 98 99 100 101 102
	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 */
103
static void bitmap_checkfree(struct bitmap *bitmap, unsigned long page)
104 105 106 107 108 109 110 111 112 113 114
{
	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;
115 116 117 118 119
	} else {
		/* normal case, free the page */
		ptr = bitmap->bp[page].map;
		bitmap->bp[page].map = NULL;
		bitmap->missing_pages++;
120
		kfree(ptr);
121 122 123 124 125 126 127 128 129 130 131
	}
}

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

/*
 * basic page I/O operations
 */

132
/* IO operations when bitmap is stored near all superblocks */
133 134 135
static int read_sb_page(struct mddev *mddev, loff_t offset,
			struct page *page,
			unsigned long index, int size)
136 137 138
{
	/* choose a good rdev and read the page from there */

139
	struct md_rdev *rdev;
140 141
	sector_t target;

N
NeilBrown 已提交
142
	rdev_for_each(rdev, mddev) {
143 144
		if (! test_bit(In_sync, &rdev->flags)
		    || test_bit(Faulty, &rdev->flags))
145 146
			continue;

J
Jonathan Brassow 已提交
147
		target = offset + index * (PAGE_SIZE/512);
148

149
		if (sync_page_io(rdev, target,
150
				 roundup(size, bdev_logical_block_size(rdev->bdev)),
J
Jonathan Brassow 已提交
151
				 page, READ, true)) {
152
			page->index = index;
153
			return 0;
154 155
		}
	}
156
	return -EIO;
157 158
}

159
static struct md_rdev *next_active_rdev(struct md_rdev *rdev, struct mddev *mddev)
160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179
{
	/* 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
	 * list_for_each_continue_rcu.
	 */
	struct list_head *pos;
	rcu_read_lock();
	if (rdev == NULL)
		/* start at the beginning */
		pos = &mddev->disks;
	else {
		/* release the previous rdev and start from there. */
		rdev_dec_pending(rdev, mddev);
		pos = &rdev->same_set;
	}
	list_for_each_continue_rcu(pos, &mddev->disks) {
180
		rdev = list_entry(pos, struct md_rdev, same_set);
181 182 183 184 185 186 187 188 189 190 191 192
		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;
}

193
static int write_sb_page(struct bitmap *bitmap, struct page *page, int wait)
194
{
195
	struct md_rdev *rdev = NULL;
196
	struct block_device *bdev;
197
	struct mddev *mddev = bitmap->mddev;
198
	struct bitmap_storage *store = &bitmap->storage;
199

200
	while ((rdev = next_active_rdev(rdev, mddev)) != NULL) {
201 202
		int size = PAGE_SIZE;
		loff_t offset = mddev->bitmap_info.offset;
203 204 205

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

206 207 208 209 210
		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,
211
				       bdev_logical_block_size(bdev));
212
		}
213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252
		/* 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);
253
	}
254 255

	if (wait)
256
		md_super_wait(mddev);
257
	return 0;
258 259 260

 bad_alignment:
	return -EINVAL;
261 262
}

263
static void bitmap_file_kick(struct bitmap *bitmap);
264
/*
265
 * write out a page to a file
266
 */
267
static void write_page(struct bitmap *bitmap, struct page *page, int wait)
268
{
269
	struct buffer_head *bh;
270

271
	if (bitmap->storage.file == NULL) {
272 273
		switch (write_sb_page(bitmap, page, wait)) {
		case -EINVAL:
274
			set_bit(BITMAP_WRITE_ERROR, &bitmap->flags);
275
		}
276
	} else {
277

278
		bh = page_buffers(page);
279

280 281 282 283
		while (bh && bh->b_blocknr) {
			atomic_inc(&bitmap->pending_writes);
			set_buffer_locked(bh);
			set_buffer_mapped(bh);
J
Jens Axboe 已提交
284
			submit_bh(WRITE | REQ_SYNC, bh);
285 286
			bh = bh->b_this_page;
		}
287

288
		if (wait)
289 290
			wait_event(bitmap->write_wait,
				   atomic_read(&bitmap->pending_writes)==0);
291
	}
292
	if (test_bit(BITMAP_WRITE_ERROR, &bitmap->flags))
293
		bitmap_file_kick(bitmap);
294 295 296 297 298
}

static void end_bitmap_write(struct buffer_head *bh, int uptodate)
{
	struct bitmap *bitmap = bh->b_private;
299

300 301
	if (!uptodate)
		set_bit(BITMAP_WRITE_ERROR, &bitmap->flags);
302 303 304
	if (atomic_dec_and_test(&bitmap->pending_writes))
		wake_up(&bitmap->write_wait);
}
305

306 307 308 309 310 311 312 313 314 315
/* copied from buffer.c */
static void
__clear_page_buffers(struct page *page)
{
	ClearPagePrivate(page);
	set_page_private(page, 0);
	page_cache_release(page);
}
static void free_buffers(struct page *page)
{
316
	struct buffer_head *bh;
317

318 319 320 321
	if (!PagePrivate(page))
		return;

	bh = page_buffers(page);
322 323 324 325
	while (bh) {
		struct buffer_head *next = bh->b_this_page;
		free_buffer_head(bh);
		bh = next;
326
	}
327 328
	__clear_page_buffers(page);
	put_page(page);
329 330
}

331 332 333 334 335 336 337
/* 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.
 */
338 339 340 341
static int read_page(struct file *file, unsigned long index,
		     struct bitmap *bitmap,
		     unsigned long count,
		     struct page *page)
342
{
343
	int ret = 0;
J
Josef Sipek 已提交
344
	struct inode *inode = file->f_path.dentry->d_inode;
345 346
	struct buffer_head *bh;
	sector_t block;
347

348 349
	pr_debug("read bitmap file (%dB @ %llu)\n", (int)PAGE_SIZE,
		 (unsigned long long)index << PAGE_SHIFT);
350

351 352
	bh = alloc_page_buffers(page, 1<<inode->i_blkbits, 0);
	if (!bh) {
353
		ret = -ENOMEM;
354 355
		goto out;
	}
356 357 358 359 360 361 362 363 364
	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! */
365
				ret = -EINVAL;
366 367 368 369 370 371 372 373 374 375
				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;
376 377 378 379
			atomic_inc(&bitmap->pending_writes);
			set_buffer_locked(bh);
			set_buffer_mapped(bh);
			submit_bh(READ, bh);
380 381 382 383 384
		}
		block++;
		bh = bh->b_this_page;
	}
	page->index = index;
385 386 387

	wait_event(bitmap->write_wait,
		   atomic_read(&bitmap->pending_writes)==0);
388
	if (test_bit(BITMAP_WRITE_ERROR, &bitmap->flags))
389
		ret = -EIO;
390
out:
391 392
	if (ret)
		printk(KERN_ALERT "md: bitmap read error: (%dB @ %llu): %d\n",
393 394
			(int)PAGE_SIZE,
			(unsigned long long)index << PAGE_SHIFT,
395 396
			ret);
	return ret;
397 398 399 400 401 402 403
}

/*
 * bitmap file superblock operations
 */

/* update the event counter and sync the superblock to disk */
404
void bitmap_update_sb(struct bitmap *bitmap)
405 406 407 408
{
	bitmap_super_t *sb;

	if (!bitmap || !bitmap->mddev) /* no bitmap for this array */
409
		return;
410 411
	if (bitmap->mddev->bitmap_info.external)
		return;
412
	if (!bitmap->storage.sb_page) /* no superblock */
413
		return;
414
	sb = kmap_atomic(bitmap->storage.sb_page);
415
	sb->events = cpu_to_le64(bitmap->mddev->events);
416
	if (bitmap->mddev->events < bitmap->events_cleared)
417 418
		/* rocking back to read-only */
		bitmap->events_cleared = bitmap->mddev->events;
419 420
	sb->events_cleared = cpu_to_le64(bitmap->events_cleared);
	sb->state = cpu_to_le32(bitmap->flags);
421 422 423
	/* 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);
424
	kunmap_atomic(sb);
425
	write_page(bitmap, bitmap->storage.sb_page, 1);
426 427 428 429 430 431 432
}

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

433
	if (!bitmap || !bitmap->storage.sb_page)
434
		return;
435
	sb = kmap_atomic(bitmap->storage.sb_page);
436
	printk(KERN_DEBUG "%s: bitmap file superblock:\n", bmname(bitmap));
437 438 439
	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",
440 441 442 443
					*(__u32 *)(sb->uuid+0),
					*(__u32 *)(sb->uuid+4),
					*(__u32 *)(sb->uuid+8),
					*(__u32 *)(sb->uuid+12));
444
	printk(KERN_DEBUG "        events: %llu\n",
445
			(unsigned long long) le64_to_cpu(sb->events));
446
	printk(KERN_DEBUG "events cleared: %llu\n",
447
			(unsigned long long) le64_to_cpu(sb->events_cleared));
448 449 450 451 452
	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);
453
	printk(KERN_DEBUG "max write behind: %d\n", le32_to_cpu(sb->write_behind));
454
	kunmap_atomic(sb);
455 456
}

457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473
/*
 * 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;
	int err = -EINVAL;

474 475 476 477
	bitmap->storage.sb_page = alloc_page(GFP_KERNEL);
	if (IS_ERR(bitmap->storage.sb_page)) {
		err = PTR_ERR(bitmap->storage.sb_page);
		bitmap->storage.sb_page = NULL;
478 479
		return err;
	}
480
	bitmap->storage.sb_page->index = 0;
481

482
	sb = kmap_atomic(bitmap->storage.sb_page);
483 484 485 486 487 488 489

	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)) {
490
		kunmap_atomic(sb);
491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519
		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;
	if (!daemon_sleep ||
	    (daemon_sleep < 1) || (daemon_sleep > MAX_SCHEDULE_TIMEOUT)) {
		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);

520
	set_bit(BITMAP_STALE, &bitmap->flags);
521
	sb->state = cpu_to_le32(bitmap->flags);
522 523 524
	bitmap->events_cleared = bitmap->mddev->events;
	sb->events_cleared = cpu_to_le64(bitmap->mddev->events);

525
	kunmap_atomic(sb);
526 527 528 529

	return 0;
}

530 531 532 533 534
/* 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;
535
	unsigned long chunksize, daemon_sleep, write_behind;
536 537
	unsigned long long events;
	int err = -EINVAL;
538
	struct page *sb_page;
539

540
	if (!bitmap->storage.file && !bitmap->mddev->bitmap_info.offset) {
541 542 543
		chunksize = 128 * 1024 * 1024;
		daemon_sleep = 5 * HZ;
		write_behind = 0;
544
		set_bit(BITMAP_STALE, &bitmap->flags);
545 546 547
		err = 0;
		goto out_no_sb;
	}
548
	/* page 0 is the superblock, read it... */
549 550 551
	sb_page = alloc_page(GFP_KERNEL);
	if (!sb_page)
		return -ENOMEM;
552
	bitmap->storage.sb_page = sb_page;
553

554 555
	if (bitmap->storage.file) {
		loff_t isize = i_size_read(bitmap->storage.file->f_mapping->host);
556 557
		int bytes = isize > PAGE_SIZE ? PAGE_SIZE : isize;

558
		err = read_page(bitmap->storage.file, 0,
559
				bitmap, bytes, sb_page);
560
	} else {
561 562 563 564
		err = read_sb_page(bitmap->mddev,
				   bitmap->mddev->bitmap_info.offset,
				   sb_page,
				   0, sizeof(bitmap_super_t));
565
	}
566
	if (err)
567 568
		return err;

569
	sb = kmap_atomic(sb_page);
570 571

	chunksize = le32_to_cpu(sb->chunksize);
572
	daemon_sleep = le32_to_cpu(sb->daemon_sleep) * HZ;
573
	write_behind = le32_to_cpu(sb->write_behind);
574 575 576 577

	/* verify that the bitmap-specific fields are valid */
	if (sb->magic != cpu_to_le32(BITMAP_MAGIC))
		reason = "bad magic";
578 579
	else if (le32_to_cpu(sb->version) < BITMAP_MAJOR_LO ||
		 le32_to_cpu(sb->version) > BITMAP_MAJOR_HI)
580
		reason = "unrecognized superblock version";
581
	else if (chunksize < 512)
582
		reason = "bitmap chunksize too small";
J
Jonathan Brassow 已提交
583
	else if (!is_power_of_2(chunksize))
584
		reason = "bitmap chunksize not a power of 2";
585
	else if (daemon_sleep < 1 || daemon_sleep > MAX_SCHEDULE_TIMEOUT)
586
		reason = "daemon sleep period out of range";
587 588
	else if (write_behind > COUNTER_MAX)
		reason = "write-behind limit out of range (0 - 16383)";
589 590 591 592 593 594 595 596 597
	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);

598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615
	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);
		if (events < bitmap->mddev->events) {
			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);
616
			set_bit(BITMAP_STALE, &bitmap->flags);
617
		}
618
	}
619

620
	/* assign fields using values from superblock */
621
	bitmap->flags |= le32_to_cpu(sb->state);
622
	if (le32_to_cpu(sb->version) == BITMAP_MAJOR_HOSTENDIAN)
623
		set_bit(BITMAP_HOSTENDIAN, &bitmap->flags);
624 625 626
	bitmap->events_cleared = le64_to_cpu(sb->events_cleared);
	err = 0;
out:
627
	kunmap_atomic(sb);
628
out_no_sb:
629
	if (test_bit(BITMAP_STALE, &bitmap->flags))
630 631 632 633
		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;
634 635 636 637 638 639 640 641 642
	if (err)
		bitmap_print_sb(bitmap);
	return err;
}

/*
 * general bitmap file operations
 */

643 644 645 646 647 648
/*
 * 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.
 */
649
/* calculate the index of the page that contains this bit */
650 651
static inline unsigned long file_page_index(struct bitmap_storage *store,
					    unsigned long chunk)
652
{
653
	if (store->sb_page)
654 655
		chunk += sizeof(bitmap_super_t) << 3;
	return chunk >> PAGE_BIT_SHIFT;
656 657 658
}

/* calculate the (bit) offset of this bit within a page */
659 660
static inline unsigned long file_page_offset(struct bitmap_storage *store,
					     unsigned long chunk)
661
{
662
	if (store->sb_page)
663 664
		chunk += sizeof(bitmap_super_t) << 3;
	return chunk & (PAGE_BITS - 1);
665 666 667 668 669 670 671 672 673
}

/*
 * return a pointer to the page in the filemap that contains the given bit
 *
 * this lookup is complicated by the fact that the bitmap sb might be exactly
 * 1 page (e.g., x86) or less than 1 page -- so the bitmap might start on page
 * 0 or page 1
 */
674
static inline struct page *filemap_get_page(struct bitmap_storage *store,
675
					    unsigned long chunk)
676
{
677
	if (file_page_index(store, chunk) >= store->file_pages)
678
		return NULL;
679 680
	return store->filemap[file_page_index(store, chunk)
			      - file_page_index(store, 0)];
681 682
}

683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734
static int bitmap_storage_alloc(struct bitmap_storage *store,
				unsigned long chunks, int with_super)
{
	int pnum;
	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);

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

	if (with_super && !store->sb_page) {
		store->sb_page = alloc_page(GFP_KERNEL);
		if (store->sb_page == NULL)
			return -ENOMEM;
		store->sb_page->index = 0;
	}
	pnum = 0;
	if (store->sb_page) {
		store->filemap[0] = store->sb_page;
		pnum = 1;
	}
	for ( ; pnum < num_pages; pnum++) {
		store->filemap[pnum] = alloc_page(GFP_KERNEL);
		if (!store->filemap[pnum]) {
			store->file_pages = pnum;
			return -ENOMEM;
		}
		store->filemap[pnum]->index = pnum;
	}
	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;
}

735 736 737 738 739 740
static void bitmap_file_unmap(struct bitmap *bitmap)
{
	struct page **map, *sb_page;
	unsigned long *attr;
	int pages;
	unsigned long flags;
741
	struct bitmap_storage *store = &bitmap->storage;
742 743

	spin_lock_irqsave(&bitmap->lock, flags);
744 745 746 747 748 749 750 751
	map = store->filemap;
	store->filemap = NULL;
	attr = store->filemap_attr;
	store->filemap_attr = NULL;
	pages = store->file_pages;
	store->file_pages = 0;
	sb_page = store->sb_page;
	store->sb_page = NULL;
752 753 754
	spin_unlock_irqrestore(&bitmap->lock, flags);

	while (pages--)
755
		if (map[pages] != sb_page) /* 0 is sb_page, release it below */
756
			free_buffers(map[pages]);
757 758 759
	kfree(map);
	kfree(attr);

760 761
	if (sb_page)
		free_buffers(sb_page);
762 763 764 765 766 767 768 769
}

static void bitmap_file_put(struct bitmap *bitmap)
{
	struct file *file;
	unsigned long flags;

	spin_lock_irqsave(&bitmap->lock, flags);
770 771
	file = bitmap->storage.file;
	bitmap->storage.file = NULL;
772 773
	spin_unlock_irqrestore(&bitmap->lock, flags);

774 775 776
	if (file)
		wait_event(bitmap->write_wait,
			   atomic_read(&bitmap->pending_writes)==0);
777 778
	bitmap_file_unmap(bitmap);

779
	if (file) {
J
Josef Sipek 已提交
780
		struct inode *inode = file->f_path.dentry->d_inode;
781
		invalidate_mapping_pages(inode->i_mapping, 0, -1);
782
		fput(file);
783
	}
784 785 786 787 788 789 790 791 792 793 794
}

/*
 * 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;

795
	if (!test_and_set_bit(BITMAP_STALE, &bitmap->flags)) {
796
		bitmap_update_sb(bitmap);
797

798
		if (bitmap->storage.file) {
799 800
			path = kmalloc(PAGE_SIZE, GFP_KERNEL);
			if (path)
801 802
				ptr = d_path(&bitmap->storage.file->f_path,
					     path, PAGE_SIZE);
C
Christoph Hellwig 已提交
803

804 805
			printk(KERN_ALERT
			      "%s: kicking failed bitmap file %s from array!\n",
C
Christoph Hellwig 已提交
806
			      bmname(bitmap), IS_ERR(ptr) ? "" : ptr);
807

808 809 810 811 812
			kfree(path);
		} else
			printk(KERN_ALERT
			       "%s: disabling internal bitmap due to errors\n",
			       bmname(bitmap));
813
	}
814 815 816 817 818 819 820

	bitmap_file_put(bitmap);

	return;
}

enum bitmap_page_attr {
821
	BITMAP_PAGE_DIRTY = 0,     /* there are set bits that need to be synced */
822 823
	BITMAP_PAGE_PENDING = 1,   /* there are bits that are being cleaned.
				    * i.e. counter is 1 or 2. */
824
	BITMAP_PAGE_NEEDWRITE = 2, /* there are cleared bits that need to be synced */
825 826
};

827 828
static inline void set_page_attr(struct bitmap *bitmap, int pnum,
				 enum bitmap_page_attr attr)
829
{
830
	__set_bit((pnum<<2) + attr, bitmap->storage.filemap_attr);
831 832
}

833 834
static inline void clear_page_attr(struct bitmap *bitmap, int pnum,
				   enum bitmap_page_attr attr)
835
{
836
	__clear_bit((pnum<<2) + attr, bitmap->storage.filemap_attr);
837 838
}

839
static inline unsigned long test_page_attr(struct bitmap *bitmap, int pnum,
840
					   enum bitmap_page_attr attr)
841
{
842
	return test_bit((pnum<<2) + attr, bitmap->storage.filemap_attr);
843 844 845 846 847 848 849 850 851 852 853 854
}

/*
 * 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;
855
	struct page *page;
856
	void *kaddr;
857
	unsigned long chunk = block >> bitmap->chunkshift;
858

859
	page = filemap_get_page(&bitmap->storage, chunk);
860 861
	if (!page)
		return;
862
	bit = file_page_offset(&bitmap->storage, chunk);
863

864
	/* set the bit */
865
	kaddr = kmap_atomic(page);
866
	if (test_bit(BITMAP_HOSTENDIAN, &bitmap->flags))
867 868 869
		set_bit(bit, kaddr);
	else
		__set_bit_le(bit, kaddr);
870
	kunmap_atomic(kaddr);
871
	pr_debug("set file bit %lu page %lu\n", bit, page->index);
872
	/* record page number so it gets flushed to disk when unplug occurs */
873
	set_page_attr(bitmap, page->index, BITMAP_PAGE_DIRTY);
874 875
}

876 877 878 879 880 881 882
static void bitmap_file_clear_bit(struct bitmap *bitmap, sector_t block)
{
	unsigned long bit;
	struct page *page;
	void *paddr;
	unsigned long chunk = block >> bitmap->chunkshift;

883
	page = filemap_get_page(&bitmap->storage, chunk);
884 885
	if (!page)
		return;
886
	bit = file_page_offset(&bitmap->storage, chunk);
887
	paddr = kmap_atomic(page);
888
	if (test_bit(BITMAP_HOSTENDIAN, &bitmap->flags))
889 890 891 892
		clear_bit(bit, paddr);
	else
		__clear_bit_le(bit, paddr);
	kunmap_atomic(paddr);
893 894
	if (!test_page_attr(bitmap, page->index, BITMAP_PAGE_NEEDWRITE)) {
		set_page_attr(bitmap, page->index, BITMAP_PAGE_PENDING);
895 896 897 898
		bitmap->allclean = 0;
	}
}

899 900 901
/* 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 */
902
void bitmap_unplug(struct bitmap *bitmap)
903
{
904 905
	unsigned long i, flags;
	int dirty, need_write;
906 907
	int wait = 0;

908
	if (!bitmap || !bitmap->storage.filemap)
909
		return;
910 911 912

	/* look at each page to see if there are any set bits that need to be
	 * flushed out to disk */
913
	for (i = 0; i < bitmap->storage.file_pages; i++) {
914
		spin_lock_irqsave(&bitmap->lock, flags);
915
		if (!bitmap->storage.filemap) {
916
			spin_unlock_irqrestore(&bitmap->lock, flags);
917
			return;
918
		}
919 920 921 922
		dirty = test_page_attr(bitmap, i, BITMAP_PAGE_DIRTY);
		need_write = test_page_attr(bitmap, i, BITMAP_PAGE_NEEDWRITE);
		clear_page_attr(bitmap, i, BITMAP_PAGE_DIRTY);
		clear_page_attr(bitmap, i, BITMAP_PAGE_NEEDWRITE);
923
		if (dirty || need_write)
924
			clear_page_attr(bitmap, i, BITMAP_PAGE_PENDING);
925
		if (dirty)
926 927 928
			wait = 1;
		spin_unlock_irqrestore(&bitmap->lock, flags);

929
		if (dirty || need_write)
930
			write_page(bitmap, bitmap->storage.filemap[i], 0);
931 932
	}
	if (wait) { /* if any writes were performed, we need to wait on them */
933
		if (bitmap->storage.file)
934 935
			wait_event(bitmap->write_wait,
				   atomic_read(&bitmap->pending_writes)==0);
936
		else
937
			md_super_wait(bitmap->mddev);
938
	}
939
	if (test_bit(BITMAP_WRITE_ERROR, &bitmap->flags))
940
		bitmap_file_kick(bitmap);
941
}
942
EXPORT_SYMBOL(bitmap_unplug);
943

944
static void bitmap_set_memory_bits(struct bitmap *bitmap, sector_t offset, int needed);
945 946 947 948 949 950 951
/* * 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.
952 953 954
 *
 * We ignore all bits for sectors that end earlier than 'start'.
 * This is used when reading an out-of-date bitmap...
955
 */
956
static int bitmap_init_from_disk(struct bitmap *bitmap, sector_t start)
957 958
{
	unsigned long i, chunks, index, oldindex, bit;
959
	struct page *page = NULL;
960
	unsigned long bit_cnt = 0;
961
	struct file *file;
962
	unsigned long offset;
963 964
	int outofdate;
	int ret = -ENOSPC;
965
	void *paddr;
966
	struct bitmap_storage *store = &bitmap->storage;
967 968

	chunks = bitmap->chunks;
969
	file = store->file;
970

971 972
	if (!file && !bitmap->mddev->bitmap_info.offset) {
		/* No permanent bitmap - fill with '1s'. */
973 974
		store->filemap = NULL;
		store->file_pages = 0;
975 976 977 978 979 980 981 982 983 984
		for (i = 0; i < chunks ; i++) {
			/* if the disk bit is set, set the memory bit */
			int needed = ((sector_t)(i+1) << (bitmap->chunkshift)
				      >= start);
			bitmap_set_memory_bits(bitmap,
					       (sector_t)i << bitmap->chunkshift,
					       needed);
		}
		return 0;
	}
985

986
	outofdate = test_bit(BITMAP_STALE, &bitmap->flags);
987 988 989 990
	if (outofdate)
		printk(KERN_INFO "%s: bitmap file is out of date, doing full "
			"recovery\n", bmname(bitmap));

991
	if (file && i_size_read(file->f_mapping->host) < store->bytes) {
992
		printk(KERN_INFO "%s: bitmap file too short %lu < %lu\n",
993 994 995
		       bmname(bitmap),
		       (unsigned long) i_size_read(file->f_mapping->host),
		       store->bytes);
996
		goto err;
997
	}
998

999
	oldindex = ~0L;
1000
	offset = 0;
1001
	if (!bitmap->mddev->bitmap_info.external)
1002
		offset = sizeof(bitmap_super_t);
1003 1004

	for (i = 0; i < chunks; i++) {
1005
		int b;
1006 1007
		index = file_page_index(&bitmap->storage, i);
		bit = file_page_offset(&bitmap->storage, i);
1008
		if (index != oldindex) { /* this is a new page, read it in */
1009
			int count;
1010
			/* unmap the old page, we're done with it */
1011 1012
			if (index == store->file_pages-1)
				count = store->bytes - index * PAGE_SIZE;
1013 1014
			else
				count = PAGE_SIZE;
1015
			page = store->filemap[index];
1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026
			if (file)
				ret = read_page(file, index, bitmap,
						count, page);
			else
				ret = read_sb_page(
					bitmap->mddev,
					bitmap->mddev->bitmap_info.offset,
					page,
					index, count);

			if (ret)
1027
				goto err;
1028

1029 1030 1031 1032 1033
			oldindex = index;

			if (outofdate) {
				/*
				 * if bitmap is out of date, dirty the
1034
				 * whole page and write it out
1035
				 */
1036
				paddr = kmap_atomic(page);
1037
				memset(paddr + offset, 0xff,
1038
				       PAGE_SIZE - offset);
1039
				kunmap_atomic(paddr);
1040 1041 1042
				write_page(bitmap, page, 1);

				ret = -EIO;
1043 1044
				if (test_bit(BITMAP_WRITE_ERROR,
					     &bitmap->flags))
1045
					goto err;
1046 1047
			}
		}
1048
		paddr = kmap_atomic(page);
1049
		if (test_bit(BITMAP_HOSTENDIAN, &bitmap->flags))
1050
			b = test_bit(bit, paddr);
1051
		else
A
Akinobu Mita 已提交
1052
			b = test_bit_le(bit, paddr);
1053
		kunmap_atomic(paddr);
1054
		if (b) {
1055
			/* if the disk bit is set, set the memory bit */
1056
			int needed = ((sector_t)(i+1) << bitmap->chunkshift
1057 1058
				      >= start);
			bitmap_set_memory_bits(bitmap,
1059
					       (sector_t)i << bitmap->chunkshift,
1060
					       needed);
1061 1062
			bit_cnt++;
		}
1063
		offset = 0;
1064 1065 1066
	}

	printk(KERN_INFO "%s: bitmap initialized from disk: "
1067
	       "read %lu pages, set %lu of %lu bits\n",
1068
	       bmname(bitmap), store->file_pages,
1069
	       bit_cnt, chunks);
1070 1071

	return 0;
1072

1073 1074 1075
 err:
	printk(KERN_INFO "%s: bitmap initialisation failed: %d\n",
	       bmname(bitmap), ret);
1076 1077 1078
	return ret;
}

1079 1080 1081 1082 1083
void bitmap_write_all(struct bitmap *bitmap)
{
	/* We don't actually write all bitmap blocks here,
	 * just flag them as needing to be written
	 */
1084
	int i;
1085

1086
	if (!bitmap || !bitmap->storage.filemap)
1087
		return;
1088
	if (bitmap->storage.file)
1089 1090 1091
		/* Only one copy, so nothing needed */
		return;

1092
	spin_lock_irq(&bitmap->lock);
1093
	for (i = 0; i < bitmap->storage.file_pages; i++)
1094
		set_page_attr(bitmap, i,
1095
			      BITMAP_PAGE_NEEDWRITE);
1096
	bitmap->allclean = 0;
1097
	spin_unlock_irq(&bitmap->lock);
1098 1099
}

1100 1101
static void bitmap_count_page(struct bitmap *bitmap, sector_t offset, int inc)
{
1102
	sector_t chunk = offset >> bitmap->chunkshift;
1103 1104 1105 1106
	unsigned long page = chunk >> PAGE_COUNTER_SHIFT;
	bitmap->bp[page].count += inc;
	bitmap_checkfree(bitmap, page);
}
1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117

static void bitmap_set_pending(struct bitmap *bitmap, sector_t offset)
{
	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;
}

1118
static bitmap_counter_t *bitmap_get_counter(struct bitmap *bitmap,
N
NeilBrown 已提交
1119
					    sector_t offset, sector_t *blocks,
1120 1121 1122 1123 1124 1125 1126
					    int create);

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

1127
void bitmap_daemon_work(struct mddev *mddev)
1128
{
1129
	struct bitmap *bitmap;
1130
	unsigned long j;
1131
	unsigned long nextpage;
1132
	unsigned long flags;
N
NeilBrown 已提交
1133
	sector_t blocks;
1134

1135 1136 1137
	/* Use a mutex to guard daemon_work against
	 * bitmap_destroy.
	 */
1138
	mutex_lock(&mddev->bitmap_info.mutex);
1139 1140
	bitmap = mddev->bitmap;
	if (bitmap == NULL) {
1141
		mutex_unlock(&mddev->bitmap_info.mutex);
1142
		return;
1143
	}
1144
	if (time_before(jiffies, bitmap->daemon_lastrun
N
NeilBrown 已提交
1145
			+ mddev->bitmap_info.daemon_sleep))
1146 1147
		goto done;

1148
	bitmap->daemon_lastrun = jiffies;
1149
	if (bitmap->allclean) {
N
NeilBrown 已提交
1150
		mddev->thread->timeout = MAX_SCHEDULE_TIMEOUT;
1151
		goto done;
1152 1153
	}
	bitmap->allclean = 1;
1154

1155 1156 1157 1158
	/* 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.
	 */
1159
	spin_lock_irqsave(&bitmap->lock, flags);
1160
	for (j = 0; j < bitmap->storage.file_pages; j++)
1161
		if (test_page_attr(bitmap, j,
1162
				   BITMAP_PAGE_PENDING)) {
1163
			set_page_attr(bitmap, j,
1164
				      BITMAP_PAGE_NEEDWRITE);
1165
			clear_page_attr(bitmap, j,
1166 1167 1168 1169 1170 1171 1172 1173 1174
					BITMAP_PAGE_PENDING);
		}

	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;
1175 1176
		if (bitmap->storage.filemap) {
			sb = kmap_atomic(bitmap->storage.sb_page);
1177 1178 1179
			sb->events_cleared =
				cpu_to_le64(bitmap->events_cleared);
			kunmap_atomic(sb);
1180
			set_page_attr(bitmap, 0,
1181 1182
				      BITMAP_PAGE_NEEDWRITE);
		}
1183 1184 1185 1186 1187
	}
	/* Now look at the bitmap counters and if any are '2' or '1',
	 * decrement and handle accordingly.
	 */
	nextpage = 0;
1188 1189
	for (j = 0; j < bitmap->chunks; j++) {
		bitmap_counter_t *bmc;
1190
		sector_t  block = (sector_t)j << bitmap->chunkshift;
1191

1192 1193 1194 1195
		if (j == nextpage) {
			nextpage += PAGE_COUNTER_RATIO;
			if (!bitmap->bp[j >> PAGE_COUNTER_SHIFT].pending) {
				j |= PAGE_COUNTER_MASK;
1196 1197
				continue;
			}
1198
			bitmap->bp[j >> PAGE_COUNTER_SHIFT].pending = 0;
1199
		}
1200
		bmc = bitmap_get_counter(bitmap,
1201
					 block,
1202
					 &blocks, 0);
1203 1204

		if (!bmc) {
1205
			j |= PAGE_COUNTER_MASK;
1206 1207 1208 1209 1210
			continue;
		}
		if (*bmc == 1 && !bitmap->need_sync) {
			/* We can clear the bit */
			*bmc = 0;
1211 1212
			bitmap_count_page(bitmap, block, -1);
			bitmap_file_clear_bit(bitmap, block);
1213 1214
		} else if (*bmc && *bmc <= 2) {
			*bmc = 1;
1215
			bitmap_set_pending(bitmap, block);
1216
			bitmap->allclean = 0;
1217
		}
1218 1219
	}

1220 1221 1222 1223 1224 1225 1226 1227
	/* 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.
	 */
1228
	for (j = 0; j < bitmap->storage.file_pages; j++) {
1229

1230
		if (test_page_attr(bitmap, j,
1231 1232 1233
				   BITMAP_PAGE_DIRTY))
			/* bitmap_unplug will handle the rest */
			break;
1234
		if (test_page_attr(bitmap, j,
1235
				   BITMAP_PAGE_NEEDWRITE)) {
1236
			clear_page_attr(bitmap, j,
1237
					BITMAP_PAGE_NEEDWRITE);
1238
			spin_unlock_irqrestore(&bitmap->lock, flags);
1239
			write_page(bitmap, bitmap->storage.filemap[j], 0);
1240
			spin_lock_irqsave(&bitmap->lock, flags);
1241
			if (!bitmap->storage.filemap)
1242
				break;
1243 1244
		}
	}
1245
	spin_unlock_irqrestore(&bitmap->lock, flags);
1246

1247
 done:
1248
	if (bitmap->allclean == 0)
N
NeilBrown 已提交
1249 1250
		mddev->thread->timeout =
			mddev->bitmap_info.daemon_sleep;
1251
	mutex_unlock(&mddev->bitmap_info.mutex);
1252 1253 1254
}

static bitmap_counter_t *bitmap_get_counter(struct bitmap *bitmap,
N
NeilBrown 已提交
1255
					    sector_t offset, sector_t *blocks,
1256
					    int create)
1257 1258
__releases(bitmap->lock)
__acquires(bitmap->lock)
1259 1260 1261 1262 1263
{
	/* 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.
	 */
1264
	sector_t chunk = offset >> bitmap->chunkshift;
1265 1266 1267
	unsigned long page = chunk >> PAGE_COUNTER_SHIFT;
	unsigned long pageoff = (chunk & PAGE_COUNTER_MASK) << COUNTER_BYTE_SHIFT;
	sector_t csize;
1268
	int err;
1269

1270 1271 1272 1273
	err = bitmap_checkpage(bitmap, page, create);

	if (bitmap->bp[page].hijacked ||
	    bitmap->bp[page].map == NULL)
1274
		csize = ((sector_t)1) << (bitmap->chunkshift +
1275 1276
					  PAGE_COUNTER_SHIFT - 1);
	else
1277
		csize = ((sector_t)1) << bitmap->chunkshift;
1278 1279 1280
	*blocks = csize - (offset & (csize - 1));

	if (err < 0)
1281
		return NULL;
1282

1283 1284 1285 1286 1287 1288 1289 1290
	/* 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];
1291
	} else /* page is allocated */
1292 1293 1294 1295
		return (bitmap_counter_t *)
			&(bitmap->bp[page].map[pageoff]);
}

1296
int bitmap_startwrite(struct bitmap *bitmap, sector_t offset, unsigned long sectors, int behind)
1297
{
1298 1299
	if (!bitmap)
		return 0;
1300 1301

	if (behind) {
1302
		int bw;
1303
		atomic_inc(&bitmap->behind_writes);
1304 1305 1306 1307
		bw = atomic_read(&bitmap->behind_writes);
		if (bw > bitmap->behind_writes_used)
			bitmap->behind_writes_used = bw;

1308 1309
		pr_debug("inc write-behind count %d/%lu\n",
			 bw, bitmap->mddev->bitmap_info.max_write_behind);
1310 1311
	}

1312
	while (sectors) {
N
NeilBrown 已提交
1313
		sector_t blocks;
1314 1315 1316 1317 1318 1319 1320 1321 1322
		bitmap_counter_t *bmc;

		spin_lock_irq(&bitmap->lock);
		bmc = bitmap_get_counter(bitmap, offset, &blocks, 1);
		if (!bmc) {
			spin_unlock_irq(&bitmap->lock);
			return 0;
		}

1323
		if (unlikely(COUNTER(*bmc) == COUNTER_MAX)) {
1324 1325 1326 1327 1328 1329 1330 1331
			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);
			spin_unlock_irq(&bitmap->lock);
J
Jens Axboe 已提交
1332
			io_schedule();
1333 1334 1335 1336
			finish_wait(&bitmap->overflow_wait, &__wait);
			continue;
		}

1337
		switch (*bmc) {
1338 1339
		case 0:
			bitmap_file_set_bit(bitmap, offset);
1340
			bitmap_count_page(bitmap, offset, 1);
1341 1342 1343 1344
			/* fall through */
		case 1:
			*bmc = 2;
		}
1345

1346 1347 1348 1349 1350 1351 1352
		(*bmc)++;

		spin_unlock_irq(&bitmap->lock);

		offset += blocks;
		if (sectors > blocks)
			sectors -= blocks;
1353 1354
		else
			sectors = 0;
1355 1356 1357
	}
	return 0;
}
1358
EXPORT_SYMBOL(bitmap_startwrite);
1359 1360

void bitmap_endwrite(struct bitmap *bitmap, sector_t offset, unsigned long sectors,
1361
		     int success, int behind)
1362
{
1363 1364
	if (!bitmap)
		return;
1365
	if (behind) {
1366 1367
		if (atomic_dec_and_test(&bitmap->behind_writes))
			wake_up(&bitmap->behind_wait);
1368 1369 1370
		pr_debug("dec write-behind count %d/%lu\n",
			 atomic_read(&bitmap->behind_writes),
			 bitmap->mddev->bitmap_info.max_write_behind);
1371 1372
	}

1373
	while (sectors) {
N
NeilBrown 已提交
1374
		sector_t blocks;
1375 1376 1377 1378 1379 1380 1381 1382 1383 1384
		unsigned long flags;
		bitmap_counter_t *bmc;

		spin_lock_irqsave(&bitmap->lock, flags);
		bmc = bitmap_get_counter(bitmap, offset, &blocks, 0);
		if (!bmc) {
			spin_unlock_irqrestore(&bitmap->lock, flags);
			return;
		}

1385
		if (success && !bitmap->mddev->degraded &&
1386 1387 1388
		    bitmap->events_cleared < bitmap->mddev->events) {
			bitmap->events_cleared = bitmap->mddev->events;
			bitmap->need_sync = 1;
1389
			sysfs_notify_dirent_safe(bitmap->sysfs_can_clear);
1390 1391
		}

1392
		if (!success && !NEEDED(*bmc))
1393 1394
			*bmc |= NEEDED_MASK;

1395
		if (COUNTER(*bmc) == COUNTER_MAX)
1396 1397
			wake_up(&bitmap->overflow_wait);

1398
		(*bmc)--;
1399
		if (*bmc <= 2) {
1400
			bitmap_set_pending(bitmap, offset);
1401 1402
			bitmap->allclean = 0;
		}
1403 1404 1405 1406
		spin_unlock_irqrestore(&bitmap->lock, flags);
		offset += blocks;
		if (sectors > blocks)
			sectors -= blocks;
1407 1408
		else
			sectors = 0;
1409 1410
	}
}
1411
EXPORT_SYMBOL(bitmap_endwrite);
1412

N
NeilBrown 已提交
1413
static int __bitmap_start_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks,
1414
			       int degraded)
1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430
{
	bitmap_counter_t *bmc;
	int rv;
	if (bitmap == NULL) {/* FIXME or bitmap set as 'failed' */
		*blocks = 1024;
		return 1; /* always resync if no bitmap */
	}
	spin_lock_irq(&bitmap->lock);
	bmc = bitmap_get_counter(bitmap, offset, blocks, 0);
	rv = 0;
	if (bmc) {
		/* locked */
		if (RESYNC(*bmc))
			rv = 1;
		else if (NEEDED(*bmc)) {
			rv = 1;
1431 1432 1433 1434
			if (!degraded) { /* don't set/clear bits if degraded */
				*bmc |= RESYNC_MASK;
				*bmc &= ~NEEDED_MASK;
			}
1435 1436 1437 1438 1439 1440
		}
	}
	spin_unlock_irq(&bitmap->lock);
	return rv;
}

N
NeilBrown 已提交
1441
int bitmap_start_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks,
1442 1443 1444 1445 1446 1447 1448 1449 1450 1451
		      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
1452
	sector_t blocks1;
1453 1454 1455 1456 1457 1458 1459 1460 1461 1462

	*blocks = 0;
	while (*blocks < (PAGE_SIZE>>9)) {
		rv |= __bitmap_start_sync(bitmap, offset,
					  &blocks1, degraded);
		offset += blocks1;
		*blocks += blocks1;
	}
	return rv;
}
1463
EXPORT_SYMBOL(bitmap_start_sync);
1464

N
NeilBrown 已提交
1465
void bitmap_end_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks, int aborted)
1466 1467 1468
{
	bitmap_counter_t *bmc;
	unsigned long flags;
1469 1470

	if (bitmap == NULL) {
1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484
		*blocks = 1024;
		return;
	}
	spin_lock_irqsave(&bitmap->lock, flags);
	bmc = bitmap_get_counter(bitmap, offset, blocks, 0);
	if (bmc == NULL)
		goto unlock;
	/* locked */
	if (RESYNC(*bmc)) {
		*bmc &= ~RESYNC_MASK;

		if (!NEEDED(*bmc) && aborted)
			*bmc |= NEEDED_MASK;
		else {
1485
			if (*bmc <= 2) {
1486
				bitmap_set_pending(bitmap, offset);
1487 1488
				bitmap->allclean = 0;
			}
1489 1490 1491 1492 1493
		}
	}
 unlock:
	spin_unlock_irqrestore(&bitmap->lock, flags);
}
1494
EXPORT_SYMBOL(bitmap_end_sync);
1495 1496 1497 1498 1499 1500 1501 1502

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
1503
	sector_t blocks;
N
NeilBrown 已提交
1504 1505
	if (!bitmap)
		return;
1506 1507
	while (sector < bitmap->mddev->resync_max_sectors) {
		bitmap_end_sync(bitmap, sector, &blocks, 0);
N
NeilBrown 已提交
1508 1509 1510
		sector += blocks;
	}
}
1511
EXPORT_SYMBOL(bitmap_close_sync);
N
NeilBrown 已提交
1512 1513 1514 1515

void bitmap_cond_end_sync(struct bitmap *bitmap, sector_t sector)
{
	sector_t s = 0;
N
NeilBrown 已提交
1516
	sector_t blocks;
N
NeilBrown 已提交
1517 1518 1519 1520 1521 1522 1523 1524

	if (!bitmap)
		return;
	if (sector == 0) {
		bitmap->last_end_sync = jiffies;
		return;
	}
	if (time_before(jiffies, (bitmap->last_end_sync
1525
				  + bitmap->mddev->bitmap_info.daemon_sleep)))
N
NeilBrown 已提交
1526 1527 1528 1529
		return;
	wait_event(bitmap->mddev->recovery_wait,
		   atomic_read(&bitmap->mddev->recovery_active) == 0);

1530
	bitmap->mddev->curr_resync_completed = sector;
1531
	set_bit(MD_CHANGE_CLEAN, &bitmap->mddev->flags);
1532
	sector &= ~((1ULL << bitmap->chunkshift) - 1);
N
NeilBrown 已提交
1533 1534 1535 1536
	s = 0;
	while (s < sector && s < bitmap->mddev->resync_max_sectors) {
		bitmap_end_sync(bitmap, s, &blocks, 0);
		s += blocks;
1537
	}
N
NeilBrown 已提交
1538
	bitmap->last_end_sync = jiffies;
1539
	sysfs_notify(&bitmap->mddev->kobj, NULL, "sync_completed");
1540
}
1541
EXPORT_SYMBOL(bitmap_cond_end_sync);
1542

1543
static void bitmap_set_memory_bits(struct bitmap *bitmap, sector_t offset, int needed)
1544 1545
{
	/* For each chunk covered by any of these sectors, set the
1546
	 * counter to 2 and possibly set resync_needed.  They should all
1547 1548
	 * be 0 at this point
	 */
1549

N
NeilBrown 已提交
1550
	sector_t secs;
1551 1552 1553 1554
	bitmap_counter_t *bmc;
	spin_lock_irq(&bitmap->lock);
	bmc = bitmap_get_counter(bitmap, offset, &secs, 1);
	if (!bmc) {
1555
		spin_unlock_irq(&bitmap->lock);
1556
		return;
1557
	}
1558
	if (!*bmc) {
1559
		*bmc = 2 | (needed ? NEEDED_MASK : 0);
1560
		bitmap_count_page(bitmap, offset, 1);
1561
		bitmap_set_pending(bitmap, offset);
1562
		bitmap->allclean = 0;
1563 1564
	}
	spin_unlock_irq(&bitmap->lock);
1565 1566
}

1567 1568 1569 1570 1571 1572
/* 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++) {
1573
		sector_t sec = (sector_t)chunk << bitmap->chunkshift;
1574
		bitmap_set_memory_bits(bitmap, sec, 1);
1575
		spin_lock_irq(&bitmap->lock);
1576
		bitmap_file_set_bit(bitmap, sec);
1577
		spin_unlock_irq(&bitmap->lock);
1578 1579 1580 1581 1582 1583
		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;
1584 1585 1586
	}
}

1587 1588 1589
/*
 * flush out any pending updates
 */
1590
void bitmap_flush(struct mddev *mddev)
1591 1592
{
	struct bitmap *bitmap = mddev->bitmap;
1593
	long sleep;
1594 1595 1596 1597 1598 1599 1600

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

	/* run the daemon_work three time to ensure everything is flushed
	 * that can be
	 */
1601
	sleep = mddev->bitmap_info.daemon_sleep * 2;
1602
	bitmap->daemon_lastrun -= sleep;
1603
	bitmap_daemon_work(mddev);
1604
	bitmap->daemon_lastrun -= sleep;
1605
	bitmap_daemon_work(mddev);
1606
	bitmap->daemon_lastrun -= sleep;
1607
	bitmap_daemon_work(mddev);
1608 1609 1610
	bitmap_update_sb(bitmap);
}

1611 1612 1613
/*
 * free memory that was allocated
 */
1614
static void bitmap_free(struct bitmap *bitmap)
1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636
{
	unsigned long k, pages;
	struct bitmap_page *bp;

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

	/* release the bitmap file and kill the daemon */
	bitmap_file_put(bitmap);

	bp = bitmap->bp;
	pages = bitmap->pages;

	/* 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);
}
1637

1638
void bitmap_destroy(struct mddev *mddev)
1639 1640 1641 1642 1643 1644
{
	struct bitmap *bitmap = mddev->bitmap;

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

1645
	mutex_lock(&mddev->bitmap_info.mutex);
1646
	mddev->bitmap = NULL; /* disconnect from the md device */
1647
	mutex_unlock(&mddev->bitmap_info.mutex);
1648 1649
	if (mddev->thread)
		mddev->thread->timeout = MAX_SCHEDULE_TIMEOUT;
1650

1651 1652 1653
	if (bitmap->sysfs_can_clear)
		sysfs_put(bitmap->sysfs_can_clear);

1654 1655
	bitmap_free(bitmap);
}
1656 1657 1658 1659 1660

/*
 * initialize the bitmap structure
 * if this returns an error, bitmap_destroy must be called to do clean up
 */
1661
int bitmap_create(struct mddev *mddev)
1662 1663
{
	struct bitmap *bitmap;
1664
	sector_t blocks = mddev->resync_max_sectors;
1665 1666
	unsigned long chunks;
	unsigned long pages;
1667
	struct file *file = mddev->bitmap_info.file;
1668
	int err;
1669
	struct sysfs_dirent *bm = NULL;
1670

A
Alexey Dobriyan 已提交
1671
	BUILD_BUG_ON(sizeof(bitmap_super_t) != 256);
1672

1673
	BUG_ON(file && mddev->bitmap_info.offset);
1674

1675
	bitmap = kzalloc(sizeof(*bitmap), GFP_KERNEL);
1676 1677 1678 1679
	if (!bitmap)
		return -ENOMEM;

	spin_lock_init(&bitmap->lock);
1680 1681
	atomic_set(&bitmap->pending_writes, 0);
	init_waitqueue_head(&bitmap->write_wait);
1682
	init_waitqueue_head(&bitmap->overflow_wait);
1683
	init_waitqueue_head(&bitmap->behind_wait);
1684

1685 1686
	bitmap->mddev = mddev;

1687 1688
	if (mddev->kobj.sd)
		bm = sysfs_get_dirent(mddev->kobj.sd, NULL, "bitmap");
1689
	if (bm) {
1690
		bitmap->sysfs_can_clear = sysfs_get_dirent(bm, NULL, "can_clear");
1691 1692 1693 1694
		sysfs_put(bm);
	} else
		bitmap->sysfs_can_clear = NULL;

1695
	bitmap->storage.file = file;
1696 1697
	if (file) {
		get_file(file);
1698 1699 1700 1701
		/* As future accesses to this file will use bmap,
		 * and bypass the page cache, we must sync the file
		 * first.
		 */
1702
		vfs_fsync(file, 1);
1703
	}
1704
	/* read superblock from bitmap file (this sets mddev->bitmap_info.chunksize) */
1705 1706 1707 1708 1709 1710 1711 1712 1713 1714
	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 {
1715 1716 1717 1718 1719 1720 1721
		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;
	}
1722
	if (err)
1723
		goto error;
1724

1725
	bitmap->daemon_lastrun = jiffies;
1726 1727
	bitmap->chunkshift = (ffz(~mddev->bitmap_info.chunksize)
			      - BITMAP_BLOCK_SHIFT);
1728

1729
	chunks = (blocks + (1 << bitmap->chunkshift) - 1) >>
1730
			bitmap->chunkshift;
1731
	pages = (chunks + PAGE_COUNTER_RATIO - 1) / PAGE_COUNTER_RATIO;
1732 1733 1734 1735 1736 1737 1738

	BUG_ON(!pages);

	bitmap->chunks = chunks;
	bitmap->pages = pages;
	bitmap->missing_pages = pages;

1739
	bitmap->bp = kzalloc(pages * sizeof(*bitmap->bp), GFP_KERNEL);
1740

1741
	err = -ENOMEM;
1742
	if (!bitmap->bp)
1743
		goto error;
1744

1745 1746 1747 1748 1749 1750
	if (file || mddev->bitmap_info.offset) {
		err = bitmap_storage_alloc(&bitmap->storage, bitmap->chunks,
					   !mddev->bitmap_info.external);
		if (err)
			goto error;
	}
1751 1752 1753 1754 1755 1756
	printk(KERN_INFO "created bitmap (%lu pages) for device %s\n",
		pages, bmname(bitmap));

	mddev->bitmap = bitmap;


1757
	return test_bit(BITMAP_WRITE_ERROR, &bitmap->flags) ? -EIO : 0;
1758 1759 1760 1761 1762 1763

 error:
	bitmap_free(bitmap);
	return err;
}

1764
int bitmap_load(struct mddev *mddev)
1765 1766
{
	int err = 0;
1767
	sector_t start = 0;
1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779
	sector_t sector = 0;
	struct bitmap *bitmap = mddev->bitmap;

	if (!bitmap)
		goto out;

	/* 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
1780
		sector_t blocks;
1781 1782 1783 1784 1785
		bitmap_start_sync(bitmap, sector, &blocks, 0);
		sector += blocks;
	}
	bitmap_close_sync(bitmap);

1786 1787 1788 1789 1790 1791
	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;

1792
	mutex_lock(&mddev->bitmap_info.mutex);
1793
	err = bitmap_init_from_disk(bitmap, start);
1794
	mutex_unlock(&mddev->bitmap_info.mutex);
1795

1796
	if (err)
1797
		goto out;
1798
	clear_bit(BITMAP_STALE, &bitmap->flags);
1799 1800 1801

	/* Kick recovery in case any bits were set */
	set_bit(MD_RECOVERY_NEEDED, &bitmap->mddev->recovery);
1802

1803
	mddev->thread->timeout = mddev->bitmap_info.daemon_sleep;
1804
	md_wakeup_thread(mddev->thread);
1805

1806 1807
	bitmap_update_sb(bitmap);

1808
	if (test_bit(BITMAP_WRITE_ERROR, &bitmap->flags))
1809 1810
		err = -EIO;
out:
1811
	return err;
1812
}
1813
EXPORT_SYMBOL_GPL(bitmap_load);
1814

1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832
void bitmap_status(struct seq_file *seq, struct bitmap *bitmap)
{
	unsigned long chunk_kb;
	unsigned long flags;

	if (!bitmap)
		return;

	spin_lock_irqsave(&bitmap->lock, flags);
	chunk_kb = bitmap->mddev->bitmap_info.chunksize >> 10;
	seq_printf(seq, "bitmap: %lu/%lu pages [%luKB], "
		   "%lu%s chunk",
		   bitmap->pages - bitmap->missing_pages,
		   bitmap->pages,
		   (bitmap->pages - bitmap->missing_pages)
		   << (PAGE_SHIFT - 10),
		   chunk_kb ? chunk_kb : bitmap->mddev->bitmap_info.chunksize,
		   chunk_kb ? "KB" : "B");
1833
	if (bitmap->storage.file) {
1834
		seq_printf(seq, ", file: ");
1835
		seq_path(seq, &bitmap->storage.file->f_path, " \t\n");
1836 1837 1838 1839 1840 1841
	}

	seq_printf(seq, "\n");
	spin_unlock_irqrestore(&bitmap->lock, flags);
}

1842
static ssize_t
1843
location_show(struct mddev *mddev, char *page)
1844 1845
{
	ssize_t len;
1846
	if (mddev->bitmap_info.file)
1847
		len = sprintf(page, "file");
1848
	else if (mddev->bitmap_info.offset)
1849
		len = sprintf(page, "%+lld", (long long)mddev->bitmap_info.offset);
1850
	else
1851 1852 1853 1854 1855 1856
		len = sprintf(page, "none");
	len += sprintf(page+len, "\n");
	return len;
}

static ssize_t
1857
location_store(struct mddev *mddev, const char *buf, size_t len)
1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901
{

	if (mddev->pers) {
		if (!mddev->pers->quiesce)
			return -EBUSY;
		if (mddev->recovery || mddev->sync_thread)
			return -EBUSY;
	}

	if (mddev->bitmap || mddev->bitmap_info.file ||
	    mddev->bitmap_info.offset) {
		/* bitmap already configured.  Only option is to clear it */
		if (strncmp(buf, "none", 4) != 0)
			return -EBUSY;
		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;
			restore_bitmap_write_access(f);
			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 */
			return -EINVAL;
		} else {
			int rv;
			if (buf[0] == '+')
				rv = strict_strtoll(buf+1, 10, &offset);
			else
				rv = strict_strtoll(buf, 10, &offset);
			if (rv)
				return rv;
			if (offset == 0)
				return -EINVAL;
1902 1903
			if (mddev->bitmap_info.external == 0 &&
			    mddev->major_version == 0 &&
1904 1905 1906 1907 1908 1909
			    offset != mddev->bitmap_info.default_offset)
				return -EINVAL;
			mddev->bitmap_info.offset = offset;
			if (mddev->pers) {
				mddev->pers->quiesce(mddev, 1);
				rv = bitmap_create(mddev);
1910 1911
				if (!rv)
					rv = bitmap_load(mddev);
1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934
				if (rv) {
					bitmap_destroy(mddev);
					mddev->bitmap_info.offset = 0;
				}
				mddev->pers->quiesce(mddev, 0);
				if (rv)
					return rv;
			}
		}
	}
	if (!mddev->external) {
		/* Ensure new bitmap info is stored in
		 * metadata promptly.
		 */
		set_bit(MD_CHANGE_DEVS, &mddev->flags);
		md_wakeup_thread(mddev->thread);
	}
	return len;
}

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

1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958
/* '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 &&
1959
	    sectors < (mddev->bitmap->storage.bytes + 511) >> 9)
1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971
		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);

1972
static ssize_t
1973
timeout_show(struct mddev *mddev, char *page)
1974 1975 1976 1977
{
	ssize_t len;
	unsigned long secs = mddev->bitmap_info.daemon_sleep / HZ;
	unsigned long jifs = mddev->bitmap_info.daemon_sleep % HZ;
1978

1979 1980 1981 1982 1983 1984 1985 1986
	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
1987
timeout_store(struct mddev *mddev, const char *buf, size_t len)
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
{
	/* 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
2023
backlog_show(struct mddev *mddev, char *page)
2024 2025 2026 2027 2028
{
	return sprintf(page, "%lu\n", mddev->bitmap_info.max_write_behind);
}

static ssize_t
2029
backlog_store(struct mddev *mddev, const char *buf, size_t len)
2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044
{
	unsigned long backlog;
	int rv = strict_strtoul(buf, 10, &backlog);
	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
2045
chunksize_show(struct mddev *mddev, char *page)
2046 2047 2048 2049 2050
{
	return sprintf(page, "%lu\n", mddev->bitmap_info.chunksize);
}

static ssize_t
2051
chunksize_store(struct mddev *mddev, const char *buf, size_t len)
2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070
{
	/* Can only be changed when no bitmap is active */
	int rv;
	unsigned long csize;
	if (mddev->bitmap)
		return -EBUSY;
	rv = strict_strtoul(buf, 10, &csize);
	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);

2071
static ssize_t metadata_show(struct mddev *mddev, char *page)
2072 2073 2074 2075 2076
{
	return sprintf(page, "%s\n", (mddev->bitmap_info.external
				      ? "external" : "internal"));
}

2077
static ssize_t metadata_store(struct mddev *mddev, const char *buf, size_t len)
2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094
{
	if (mddev->bitmap ||
	    mddev->bitmap_info.file ||
	    mddev->bitmap_info.offset)
		return -EBUSY;
	if (strncmp(buf, "external", 8) == 0)
		mddev->bitmap_info.external = 1;
	else if (strncmp(buf, "internal", 8) == 0)
		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);

2095
static ssize_t can_clear_show(struct mddev *mddev, char *page)
2096 2097 2098 2099 2100 2101 2102 2103 2104 2105
{
	int len;
	if (mddev->bitmap)
		len = sprintf(page, "%s\n", (mddev->bitmap->need_sync ?
					     "false" : "true"));
	else
		len = sprintf(page, "\n");
	return len;
}

2106
static ssize_t can_clear_store(struct mddev *mddev, const char *buf, size_t len)
2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123
{
	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);

2124
static ssize_t
2125
behind_writes_used_show(struct mddev *mddev, char *page)
2126 2127 2128 2129 2130 2131 2132 2133
{
	if (mddev->bitmap == NULL)
		return sprintf(page, "0\n");
	return sprintf(page, "%lu\n",
		       mddev->bitmap->behind_writes_used);
}

static ssize_t
2134
behind_writes_used_reset(struct mddev *mddev, const char *buf, size_t len)
2135 2136 2137 2138 2139 2140 2141 2142 2143 2144
{
	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);

2145 2146
static struct attribute *md_bitmap_attrs[] = {
	&bitmap_location.attr,
2147
	&bitmap_space.attr,
2148 2149 2150
	&bitmap_timeout.attr,
	&bitmap_backlog.attr,
	&bitmap_chunksize.attr,
2151 2152
	&bitmap_metadata.attr,
	&bitmap_can_clear.attr,
2153
	&max_backlog_used.attr,
2154 2155 2156 2157 2158 2159 2160
	NULL
};
struct attribute_group md_bitmap_group = {
	.name = "bitmap",
	.attrs = md_bitmap_attrs,
};