buffer.c

/*
 *  linux/fs/buffer.c
 *
 *  Copyright (C) 1991, 1992, 2002  Linus Torvalds
 */

/*
 * Start bdflush() with kernel_thread not syscall - Paul Gortmaker, 12/95
 *
 * Removed a lot of unnecessary code and simplified things now that
 * the buffer cache isn't our primary cache - Andrew Tridgell 12/96
 *
 * Speed up hash, lru, and free list operations.  Use gfp() for allocating
 * hash table, use SLAB cache for buffer heads. SMP threading.  -DaveM
 *
 * Added 32k buffer block sizes - these are required older ARM systems. - RMK
 *
 * async buffer flushing, 1999 Andrea Arcangeli <andrea@suse.de>
 */

#include <linux/kernel.h>
#include <linux/syscalls.h>
#include <linux/fs.h>
#include <linux/mm.h>
#include <linux/percpu.h>
#include <linux/slab.h>
#include <linux/capability.h>
#include <linux/blkdev.h>
#include <linux/file.h>
#include <linux/quotaops.h>
#include <linux/highmem.h>
#include <linux/module.h>
#include <linux/writeback.h>
#include <linux/hash.h>
#include <linux/suspend.h>
#include <linux/buffer_head.h>
#include <linux/task_io_accounting_ops.h>
#include <linux/bio.h>
#include <linux/notifier.h>
#include <linux/cpu.h>
#include <linux/bitops.h>
#include <linux/mpage.h>
#include <linux/bit_spinlock.h>

static int fsync_buffers_list(spinlock_t *lock, struct list_head *list);

#define BH_ENTRY(list) list_entry((list), struct buffer_head, b_assoc_buffers)

inline void
init_buffer(struct buffer_head *bh, bh_end_io_t *handler, void *private)
{
	bh->b_end_io = handler;
	bh->b_private = private;
}

static int sync_buffer(void *word)
{
	struct block_device *bd;
	struct buffer_head *bh
		= container_of(word, struct buffer_head, b_state);

	smp_mb();
	bd = bh->b_bdev;
	if (bd)
		blk_run_address_space(bd->bd_inode->i_mapping);
	io_schedule();
	return 0;
}

void fastcall __lock_buffer(struct buffer_head *bh)
{
	wait_on_bit_lock(&bh->b_state, BH_Lock, sync_buffer,
							TASK_UNINTERRUPTIBLE);
}
EXPORT_SYMBOL(__lock_buffer);

void fastcall unlock_buffer(struct buffer_head *bh)
{
	smp_mb__before_clear_bit();
	clear_buffer_locked(bh);
	smp_mb__after_clear_bit();
	wake_up_bit(&bh->b_state, BH_Lock);
}

/*
 * Block until a buffer comes unlocked.  This doesn't stop it
 * from becoming locked again - you have to lock it yourself
 * if you want to preserve its state.
 */
void __wait_on_buffer(struct buffer_head * bh)
{
	wait_on_bit(&bh->b_state, BH_Lock, sync_buffer, TASK_UNINTERRUPTIBLE);
}

static void
__clear_page_buffers(struct page *page)
{
	ClearPagePrivate(page);
	set_page_private(page, 0);
	page_cache_release(page);
}

static void buffer_io_error(struct buffer_head *bh)
{
	char b[BDEVNAME_SIZE];

	printk(KERN_ERR "Buffer I/O error on device %s, logical block %Lu\n",
			bdevname(bh->b_bdev, b),
			(unsigned long long)bh->b_blocknr);
}

/*
 * Default synchronous end-of-IO handler..  Just mark it up-to-date and
 * unlock the buffer. This is what ll_rw_block uses too.
 */
void end_buffer_read_sync(struct buffer_head *bh, int uptodate)
{
	if (uptodate) {
		set_buffer_uptodate(bh);
	} else {
		/* This happens, due to failed READA attempts. */
		clear_buffer_uptodate(bh);
	}
	unlock_buffer(bh);
	put_bh(bh);
}

void end_buffer_write_sync(struct buffer_head *bh, int uptodate)
{
	char b[BDEVNAME_SIZE];

	if (uptodate) {
		set_buffer_uptodate(bh);
	} else {
		if (!buffer_eopnotsupp(bh) && printk_ratelimit()) {
			buffer_io_error(bh);
			printk(KERN_WARNING "lost page write due to "
					"I/O error on %s\n",
				       bdevname(bh->b_bdev, b));
		}
		set_buffer_write_io_error(bh);
		clear_buffer_uptodate(bh);
	}
	unlock_buffer(bh);
	put_bh(bh);
}

/*
 * Write out and wait upon all the dirty data associated with a block
 * device via its mapping.  Does not take the superblock lock.
 */
int sync_blockdev(struct block_device *bdev)
{
	int ret = 0;

	if (bdev)
		ret = filemap_write_and_wait(bdev->bd_inode->i_mapping);
	return ret;
}
EXPORT_SYMBOL(sync_blockdev);

/*
 * Write out and wait upon all dirty data associated with this
 * device.   Filesystem data as well as the underlying block
 * device.  Takes the superblock lock.
 */
int fsync_bdev(struct block_device *bdev)
{
	struct super_block *sb = get_super(bdev);
	if (sb) {
		int res = fsync_super(sb);
		drop_super(sb);
		return res;
	}
	return sync_blockdev(bdev);
}

/**
 * freeze_bdev  --  lock a filesystem and force it into a consistent state
 * @bdev:	blockdevice to lock
 *
 * This takes the block device bd_mount_sem to make sure no new mounts
 * happen on bdev until thaw_bdev() is called.
 * If a superblock is found on this device, we take the s_umount semaphore
 * on it to make sure nobody unmounts until the snapshot creation is done.
 */
struct super_block *freeze_bdev(struct block_device *bdev)
{
	struct super_block *sb;

	down(&bdev->bd_mount_sem);
	sb = get_super(bdev);
	if (sb && !(sb->s_flags & MS_RDONLY)) {
		sb->s_frozen = SB_FREEZE_WRITE;
		smp_wmb();

		__fsync_super(sb);

		sb->s_frozen = SB_FREEZE_TRANS;
		smp_wmb();

		sync_blockdev(sb->s_bdev);

		if (sb->s_op->write_super_lockfs)
			sb->s_op->write_super_lockfs(sb);
	}

	sync_blockdev(bdev);
	return sb;	/* thaw_bdev releases s->s_umount and bd_mount_sem */
}
EXPORT_SYMBOL(freeze_bdev);

/**
 * thaw_bdev  -- unlock filesystem
 * @bdev:	blockdevice to unlock
 * @sb:		associated superblock
 *
 * Unlocks the filesystem and marks it writeable again after freeze_bdev().
 */
void thaw_bdev(struct block_device *bdev, struct super_block *sb)
{
	if (sb) {
		BUG_ON(sb->s_bdev != bdev);

		if (sb->s_op->unlockfs)
			sb->s_op->unlockfs(sb);
		sb->s_frozen = SB_UNFROZEN;
		smp_wmb();
		wake_up(&sb->s_wait_unfrozen);
		drop_super(sb);
	}

	up(&bdev->bd_mount_sem);
}
EXPORT_SYMBOL(thaw_bdev);

/*
 * Various filesystems appear to want __find_get_block to be non-blocking.
 * But it's the page lock which protects the buffers.  To get around this,
 * we get exclusion from try_to_free_buffers with the blockdev mapping's
 * private_lock.
 *
 * Hack idea: for the blockdev mapping, i_bufferlist_lock contention
 * may be quite high.  This code could TryLock the page, and if that
 * succeeds, there is no need to take private_lock. (But if
 * private_lock is contended then so is mapping->tree_lock).
 */
static struct buffer_head *
__find_get_block_slow(struct block_device *bdev, sector_t block)
{
	struct inode *bd_inode = bdev->bd_inode;
	struct address_space *bd_mapping = bd_inode->i_mapping;
	struct buffer_head *ret = NULL;
	pgoff_t index;
	struct buffer_head *bh;
	struct buffer_head *head;
	struct page *page;
	int all_mapped = 1;

	index = block >> (PAGE_CACHE_SHIFT - bd_inode->i_blkbits);
	page = find_get_page(bd_mapping, index);
	if (!page)
		goto out;

	spin_lock(&bd_mapping->private_lock);
	if (!page_has_buffers(page))
		goto out_unlock;
	head = page_buffers(page);
	bh = head;
	do {
		if (bh->b_blocknr == block) {
			ret = bh;
			get_bh(bh);
			goto out_unlock;
		}
		if (!buffer_mapped(bh))
			all_mapped = 0;
		bh = bh->b_this_page;
	} while (bh != head);

	/* we might be here because some of the buffers on this page are
	 * not mapped.  This is due to various races between
	 * file io on the block device and getblk.  It gets dealt with
	 * elsewhere, don't buffer_error if we had some unmapped buffers
	 */
	if (all_mapped) {
		printk("__find_get_block_slow() failed. "
			"block=%llu, b_blocknr=%llu\n",
			(unsigned long long)block,
			(unsigned long long)bh->b_blocknr);
		printk("b_state=0x%08lx, b_size=%zu\n",
			bh->b_state, bh->b_size);
		printk("device blocksize: %d\n", 1 << bd_inode->i_blkbits);
	}
out_unlock:
	spin_unlock(&bd_mapping->private_lock);
	page_cache_release(page);
out:
	return ret;
}

/* If invalidate_buffers() will trash dirty buffers, it means some kind
   of fs corruption is going on. Trashing dirty data always imply losing
   information that was supposed to be just stored on the physical layer
   by the user.

   Thus invalidate_buffers in general usage is not allwowed to trash
   dirty buffers. For example ioctl(FLSBLKBUF) expects dirty data to
   be preserved.  These buffers are simply skipped.
  
   We also skip buffers which are still in use.  For example this can
   happen if a userspace program is reading the block device.

   NOTE: In the case where the user removed a removable-media-disk even if
   there's still dirty data not synced on disk (due a bug in the device driver
   or due an error of the user), by not destroying the dirty buffers we could
   generate corruption also on the next media inserted, thus a parameter is
   necessary to handle this case in the most safe way possible (trying
   to not corrupt also the new disk inserted with the data belonging to
   the old now corrupted disk). Also for the ramdisk the natural thing
   to do in order to release the ramdisk memory is to destroy dirty buffers.

   These are two special cases. Normal usage imply the device driver
   to issue a sync on the device (without waiting I/O completion) and
   then an invalidate_buffers call that doesn't trash dirty buffers.

   For handling cache coherency with the blkdev pagecache the 'update' case
   is been introduced. It is needed to re-read from disk any pinned
   buffer. NOTE: re-reading from disk is destructive so we can do it only
   when we assume nobody is changing the buffercache under our I/O and when
   we think the disk contains more recent information than the buffercache.
   The update == 1 pass marks the buffers we need to update, the update == 2
   pass does the actual I/O. */
void invalidate_bdev(struct block_device *bdev)
{
	struct address_space *mapping = bdev->bd_inode->i_mapping;

	if (mapping->nrpages == 0)
		return;

	invalidate_bh_lrus();
	invalidate_mapping_pages(mapping, 0, -1);
}

/*
 * Kick pdflush then try to free up some ZONE_NORMAL memory.
 */
static void free_more_memory(void)
{
	struct zone **zones;
	pg_data_t *pgdat;

	wakeup_pdflush(1024);
	yield();

	for_each_online_pgdat(pgdat) {
		zones = pgdat->node_zonelists[gfp_zone(GFP_NOFS)].zones;
		if (*zones)
			try_to_free_pages(zones, 0, GFP_NOFS);
	}
}

/*
 * I/O completion handler for block_read_full_page() - pages
 * which come unlocked at the end of I/O.
 */
static void end_buffer_async_read(struct buffer_head *bh, int uptodate)
{
	unsigned long flags;
	struct buffer_head *first;
	struct buffer_head *tmp;
	struct page *page;
	int page_uptodate = 1;

	BUG_ON(!buffer_async_read(bh));

	page = bh->b_page;
	if (uptodate) {
		set_buffer_uptodate(bh);
	} else {
		clear_buffer_uptodate(bh);
		if (printk_ratelimit())
			buffer_io_error(bh);
		SetPageError(page);
	}

	/*
	 * Be _very_ careful from here on. Bad things can happen if
	 * two buffer heads end IO at almost the same time and both
	 * decide that the page is now completely done.
	 */
	first = page_buffers(page);
	local_irq_save(flags);
	bit_spin_lock(BH_Uptodate_Lock, &first->b_state);
	clear_buffer_async_read(bh);
	unlock_buffer(bh);
	tmp = bh;
	do {
		if (!buffer_uptodate(tmp))
			page_uptodate = 0;
		if (buffer_async_read(tmp)) {
			BUG_ON(!buffer_locked(tmp));
			goto still_busy;
		}
		tmp = tmp->b_this_page;
	} while (tmp != bh);
	bit_spin_unlock(BH_Uptodate_Lock, &first->b_state);
	local_irq_restore(flags);

	/*
	 * If none of the buffers had errors and they are all
	 * uptodate then we can set the page uptodate.
	 */
	if (page_uptodate && !PageError(page))
		SetPageUptodate(page);
	unlock_page(page);
	return;

still_busy:
	bit_spin_unlock(BH_Uptodate_Lock, &first->b_state);
	local_irq_restore(flags);
	return;
}

/*
 * Completion handler for block_write_full_page() - pages which are unlocked
 * during I/O, and which have PageWriteback cleared upon I/O completion.
 */
static void end_buffer_async_write(struct buffer_head *bh, int uptodate)
{
	char b[BDEVNAME_SIZE];
	unsigned long flags;
	struct buffer_head *first;
	struct buffer_head *tmp;
	struct page *page;

	BUG_ON(!buffer_async_write(bh));

	page = bh->b_page;
	if (uptodate) {
		set_buffer_uptodate(bh);
	} else {
		if (printk_ratelimit()) {
			buffer_io_error(bh);
			printk(KERN_WARNING "lost page write due to "
					"I/O error on %s\n",
			       bdevname(bh->b_bdev, b));
		}
		set_bit(AS_EIO, &page->mapping->flags);
		set_buffer_write_io_error(bh);
		clear_buffer_uptodate(bh);
		SetPageError(page);
	}

	first = page_buffers(page);
	local_irq_save(flags);
	bit_spin_lock(BH_Uptodate_Lock, &first->b_state);

	clear_buffer_async_write(bh);
	unlock_buffer(bh);
	tmp = bh->b_this_page;
	while (tmp != bh) {
		if (buffer_async_write(tmp)) {
			BUG_ON(!buffer_locked(tmp));
			goto still_busy;
		}
		tmp = tmp->b_this_page;
	}
	bit_spin_unlock(BH_Uptodate_Lock, &first->b_state);
	local_irq_restore(flags);
	end_page_writeback(page);
	return;

still_busy:
	bit_spin_unlock(BH_Uptodate_Lock, &first->b_state);
	local_irq_restore(flags);
	return;
}

/*
 * If a page's buffers are under async readin (end_buffer_async_read
 * completion) then there is a possibility that another thread of
 * control could lock one of the buffers after it has completed
 * but while some of the other buffers have not completed.  This
 * locked buffer would confuse end_buffer_async_read() into not unlocking
 * the page.  So the absence of BH_Async_Read tells end_buffer_async_read()
 * that this buffer is not under async I/O.
 *
 * The page comes unlocked when it has no locked buffer_async buffers
 * left.
 *
 * PageLocked prevents anyone starting new async I/O reads any of
 * the buffers.
 *
 * PageWriteback is used to prevent simultaneous writeout of the same
 * page.
 *
 * PageLocked prevents anyone from starting writeback of a page which is
 * under read I/O (PageWriteback is only ever set against a locked page).
 */
static void mark_buffer_async_read(struct buffer_head *bh)
{
	bh->b_end_io = end_buffer_async_read;
	set_buffer_async_read(bh);
}

void mark_buffer_async_write(struct buffer_head *bh)
{
	bh->b_end_io = end_buffer_async_write;
	set_buffer_async_write(bh);
}
EXPORT_SYMBOL(mark_buffer_async_write);


/*
 * fs/buffer.c contains helper functions for buffer-backed address space's
 * fsync functions.  A common requirement for buffer-based filesystems is
 * that certain data from the backing blockdev needs to be written out for
 * a successful fsync().  For example, ext2 indirect blocks need to be
 * written back and waited upon before fsync() returns.
 *
 * The functions mark_buffer_inode_dirty(), fsync_inode_buffers(),
 * inode_has_buffers() and invalidate_inode_buffers() are provided for the
 * management of a list of dependent buffers at ->i_mapping->private_list.
 *
 * Locking is a little subtle: try_to_free_buffers() will remove buffers
 * from their controlling inode's queue when they are being freed.  But
 * try_to_free_buffers() will be operating against the *blockdev* mapping
 * at the time, not against the S_ISREG file which depends on those buffers.
 * So the locking for private_list is via the private_lock in the address_space
 * which backs the buffers.  Which is different from the address_space 
 * against which the buffers are listed.  So for a particular address_space,
 * mapping->private_lock does *not* protect mapping->private_list!  In fact,
 * mapping->private_list will always be protected by the backing blockdev's
 * ->private_lock.
 *
 * Which introduces a requirement: all buffers on an address_space's
 * ->private_list must be from the same address_space: the blockdev's.
 *
 * address_spaces which do not place buffers at ->private_list via these
 * utility functions are free to use private_lock and private_list for
 * whatever they want.  The only requirement is that list_empty(private_list)
 * be true at clear_inode() time.
 *
 * FIXME: clear_inode should not call invalidate_inode_buffers().  The
 * filesystems should do that.  invalidate_inode_buffers() should just go
 * BUG_ON(!list_empty).
 *
 * FIXME: mark_buffer_dirty_inode() is a data-plane operation.  It should
 * take an address_space, not an inode.  And it should be called
 * mark_buffer_dirty_fsync() to clearly define why those buffers are being
 * queued up.
 *
 * FIXME: mark_buffer_dirty_inode() doesn't need to add the buffer to the
 * list if it is already on a list.  Because if the buffer is on a list,
 * it *must* already be on the right one.  If not, the filesystem is being
 * silly.  This will save a ton of locking.  But first we have to ensure
 * that buffers are taken *off* the old inode's list when they are freed
 * (presumably in truncate).  That requires careful auditing of all
 * filesystems (do it inside bforget()).  It could also be done by bringing
 * b_inode back.
 */

/*
 * The buffer's backing address_space's private_lock must be held
 */
static inline void __remove_assoc_queue(struct buffer_head *bh)
{
	list_del_init(&bh->b_assoc_buffers);
	WARN_ON(!bh->b_assoc_map);
	if (buffer_write_io_error(bh))
		set_bit(AS_EIO, &bh->b_assoc_map->flags);
	bh->b_assoc_map = NULL;
}

int inode_has_buffers(struct inode *inode)
{
	return !list_empty(&inode->i_data.private_list);
}

/*
 * osync is designed to support O_SYNC io.  It waits synchronously for
 * all already-submitted IO to complete, but does not queue any new
 * writes to the disk.
 *
 * To do O_SYNC writes, just queue the buffer writes with ll_rw_block as
 * you dirty the buffers, and then use osync_inode_buffers to wait for
 * completion.  Any other dirty buffers which are not yet queued for
 * write will not be flushed to disk by the osync.
 */
static int osync_buffers_list(spinlock_t *lock, struct list_head *list)
{
	struct buffer_head *bh;
	struct list_head *p;
	int err = 0;

	spin_lock(lock);
repeat:
	list_for_each_prev(p, list) {
		bh = BH_ENTRY(p);
		if (buffer_locked(bh)) {
			get_bh(bh);
			spin_unlock(lock);
			wait_on_buffer(bh);
			if (!buffer_uptodate(bh))
				err = -EIO;
			brelse(bh);
			spin_lock(lock);
			goto repeat;
		}
	}
	spin_unlock(lock);
	return err;
}

/**
 * sync_mapping_buffers - write out and wait upon a mapping's "associated"
 *                        buffers
 * @mapping: the mapping which wants those buffers written
 *
 * Starts I/O against the buffers at mapping->private_list, and waits upon
 * that I/O.
 *
 * Basically, this is a convenience function for fsync().
 * @mapping is a file or directory which needs those buffers to be written for
 * a successful fsync().
 */
int sync_mapping_buffers(struct address_space *mapping)
{
	struct address_space *buffer_mapping = mapping->assoc_mapping;

	if (buffer_mapping == NULL || list_empty(&mapping->private_list))
		return 0;

	return fsync_buffers_list(&buffer_mapping->private_lock,
					&mapping->private_list);
}
EXPORT_SYMBOL(sync_mapping_buffers);

/*
 * Called when we've recently written block `bblock', and it is known that
 * `bblock' was for a buffer_boundary() buffer.  This means that the block at
 * `bblock + 1' is probably a dirty indirect block.  Hunt it down and, if it's
 * dirty, schedule it for IO.  So that indirects merge nicely with their data.
 */
void write_boundary_block(struct block_device *bdev,
			sector_t bblock, unsigned blocksize)
{
	struct buffer_head *bh = __find_get_block(bdev, bblock + 1, blocksize);
	if (bh) {
		if (buffer_dirty(bh))
			ll_rw_block(WRITE, 1, &bh);
		put_bh(bh);
	}
}

void mark_buffer_dirty_inode(struct buffer_head *bh, struct inode *inode)
{
	struct address_space *mapping = inode->i_mapping;
	struct address_space *buffer_mapping = bh->b_page->mapping;

	mark_buffer_dirty(bh);
	if (!mapping->assoc_mapping) {
		mapping->assoc_mapping = buffer_mapping;
	} else {
		BUG_ON(mapping->assoc_mapping != buffer_mapping);
	}
	if (list_empty(&bh->b_assoc_buffers)) {
		spin_lock(&buffer_mapping->private_lock);
		list_move_tail(&bh->b_assoc_buffers,
				&mapping->private_list);
		bh->b_assoc_map = mapping;
		spin_unlock(&buffer_mapping->private_lock);
	}
}
EXPORT_SYMBOL(mark_buffer_dirty_inode);

/*
 * Mark the page dirty, and set it dirty in the radix tree, and mark the inode
 * dirty.
 *
 * If warn is true, then emit a warning if the page is not uptodate and has
 * not been truncated.
 */
static int __set_page_dirty(struct page *page,
		struct address_space *mapping, int warn)
{
	if (unlikely(!mapping))
		return !TestSetPageDirty(page);

	if (TestSetPageDirty(page))
		return 0;

	write_lock_irq(&mapping->tree_lock);
	if (page->mapping) {	/* Race with truncate? */
		WARN_ON_ONCE(warn && !PageUptodate(page));

		if (mapping_cap_account_dirty(mapping)) {
			__inc_zone_page_state(page, NR_FILE_DIRTY);
			task_io_account_write(PAGE_CACHE_SIZE);
		}
		radix_tree_tag_set(&mapping->page_tree,
				page_index(page), PAGECACHE_TAG_DIRTY);
	}
	write_unlock_irq(&mapping->tree_lock);
	__mark_inode_dirty(mapping->host, I_DIRTY_PAGES);

	return 1;
}

/*
 * Add a page to the dirty page list.
 *
 * It is a sad fact of life that this function is called from several places
 * deeply under spinlocking.  It may not sleep.
 *
 * If the page has buffers, the uptodate buffers are set dirty, to preserve
 * dirty-state coherency between the page and the buffers.  It the page does
 * not have buffers then when they are later attached they will all be set
 * dirty.
 *
 * The buffers are dirtied before the page is dirtied.  There's a small race
 * window in which a writepage caller may see the page cleanness but not the
 * buffer dirtiness.  That's fine.  If this code were to set the page dirty
 * before the buffers, a concurrent writepage caller could clear the page dirty
 * bit, see a bunch of clean buffers and we'd end up with dirty buffers/clean
 * page on the dirty page list.
 *
 * We use private_lock to lock against try_to_free_buffers while using the
 * page's buffer list.  Also use this to protect against clean buffers being
 * added to the page after it was set dirty.
 *
 * FIXME: may need to call ->reservepage here as well.  That's rather up to the
 * address_space though.
 */
int __set_page_dirty_buffers(struct page *page)
{
	struct address_space *mapping = page_mapping(page);

	if (unlikely(!mapping))
		return !TestSetPageDirty(page);

	spin_lock(&mapping->private_lock);
	if (page_has_buffers(page)) {
		struct buffer_head *head = page_buffers(page);
		struct buffer_head *bh = head;

		do {
			set_buffer_dirty(bh);
			bh = bh->b_this_page;
		} while (bh != head);
	}
	spin_unlock(&mapping->private_lock);

	return __set_page_dirty(page, mapping, 1);
}
EXPORT_SYMBOL(__set_page_dirty_buffers);

/*
 * Write out and wait upon a list of buffers.
 *
 * We have conflicting pressures: we want to make sure that all
 * initially dirty buffers get waited on, but that any subsequently
 * dirtied buffers don't.  After all, we don't want fsync to last
 * forever if somebody is actively writing to the file.
 *
 * Do this in two main stages: first we copy dirty buffers to a
 * temporary inode list, queueing the writes as we go.  Then we clean
 * up, waiting for those writes to complete.
 * 
 * During this second stage, any subsequent updates to the file may end
 * up refiling the buffer on the original inode's dirty list again, so
 * there is a chance we will end up with a buffer queued for write but
 * not yet completed on that list.  So, as a final cleanup we go through
 * the osync code to catch these locked, dirty buffers without requeuing
 * any newly dirty buffers for write.
 */
static int fsync_buffers_list(spinlock_t *lock, struct list_head *list)
{
	struct buffer_head *bh;
	struct list_head tmp;
	int err = 0, err2;

	INIT_LIST_HEAD(&tmp);

	spin_lock(lock);
	while (!list_empty(list)) {
		bh = BH_ENTRY(list->next);
		__remove_assoc_queue(bh);
		if (buffer_dirty(bh) || buffer_locked(bh)) {
			list_add(&bh->b_assoc_buffers, &tmp);
			if (buffer_dirty(bh)) {
				get_bh(bh);
				spin_unlock(lock);
				/*
				 * Ensure any pending I/O completes so that
				 * ll_rw_block() actually writes the current
				 * contents - it is a noop if I/O is still in
				 * flight on potentially older contents.
				 */
				ll_rw_block(SWRITE, 1, &bh);
				brelse(bh);
				spin_lock(lock);
			}
		}
	}

	while (!list_empty(&tmp)) {
		bh = BH_ENTRY(tmp.prev);
		list_del_init(&bh->b_assoc_buffers);
		get_bh(bh);
		spin_unlock(lock);
		wait_on_buffer(bh);
		if (!buffer_uptodate(bh))
			err = -EIO;
		brelse(bh);
		spin_lock(lock);
	}
	
	spin_unlock(lock);
	err2 = osync_buffers_list(lock, list);
	if (err)
		return err;
	else
		return err2;
}

/*
 * Invalidate any and all dirty buffers on a given inode.  We are
 * probably unmounting the fs, but that doesn't mean we have already
 * done a sync().  Just drop the buffers from the inode list.
 *
 * NOTE: we take the inode's blockdev's mapping's private_lock.  Which
 * assumes that all the buffers are against the blockdev.  Not true
 * for reiserfs.
 */
void invalidate_inode_buffers(struct inode *inode)
{
	if (inode_has_buffers(inode)) {
		struct address_space *mapping = &inode->i_data;
		struct list_head *list = &mapping->private_list;
		struct address_space *buffer_mapping = mapping->assoc_mapping;

		spin_lock(&buffer_mapping->private_lock);
		while (!list_empty(list))
			__remove_assoc_queue(BH_ENTRY(list->next));
		spin_unlock(&buffer_mapping->private_lock);
	}
}

/*
 * Remove any clean buffers from the inode's buffer list.  This is called
 * when we're trying to free the inode itself.  Those buffers can pin it.
 *
 * Returns true if all buffers were removed.
 */
int remove_inode_buffers(struct inode *inode)
{
	int ret = 1;

	if (inode_has_buffers(inode)) {
		struct address_space *mapping = &inode->i_data;
		struct list_head *list = &mapping->private_list;
		struct address_space *buffer_mapping = mapping->assoc_mapping;

		spin_lock(&buffer_mapping->private_lock);
		while (!list_empty(list)) {
			struct buffer_head *bh = BH_ENTRY(list->next);
			if (buffer_dirty(bh)) {
				ret = 0;
				break;
			}
			__remove_assoc_queue(bh);
		}
		spin_unlock(&buffer_mapping->private_lock);
	}
	return ret;
}

/*
 * Create the appropriate buffers when given a page for data area and
 * the size of each buffer.. Use the bh->b_this_page linked list to
 * follow the buffers created.  Return NULL if unable to create more
 * buffers.
 *
 * The retry flag is used to differentiate async IO (paging, swapping)
 * which may not fail from ordinary buffer allocations.
 */
struct buffer_head *alloc_page_buffers(struct page *page, unsigned long size,
		int retry)
{
	struct buffer_head *bh, *head;
	long offset;

try_again:
	head = NULL;
	offset = PAGE_SIZE;
	while ((offset -= size) >= 0) {
		bh = alloc_buffer_head(GFP_NOFS);
		if (!bh)
			goto no_grow;

		bh->b_bdev = NULL;
		bh->b_this_page = head;
		bh->b_blocknr = -1;
		head = bh;

		bh->b_state = 0;
		atomic_set(&bh->b_count, 0);
		bh->b_private = NULL;
		bh->b_size = size;

		/* Link the buffer to its page */
		set_bh_page(bh, page, offset);

		init_buffer(bh, NULL, NULL);
	}
	return head;
/*
 * In case anything failed, we just free everything we got.
 */
no_grow:
	if (head) {
		do {
			bh = head;
			head = head->b_this_page;
			free_buffer_head(bh);
		} while (head);
	}

	/*
	 * Return failure for non-async IO requests.  Async IO requests
	 * are not allowed to fail, so we have to wait until buffer heads
	 * become available.  But we don't want tasks sleeping with 
	 * partially complete buffers, so all were released above.
	 */
	if (!retry)
		return NULL;

	/* We're _really_ low on memory. Now we just
	 * wait for old buffer heads to become free due to
	 * finishing IO.  Since this is an async request and
	 * the reserve list is empty, we're sure there are 
	 * async buffer heads in use.
	 */
	free_more_memory();
	goto try_again;
}
EXPORT_SYMBOL_GPL(alloc_page_buffers);

static inline void
link_dev_buffers(struct page *page, struct buffer_head *head)
{
	struct buffer_head *bh, *tail;

	bh = head;
	do {
		tail = bh;
		bh = bh->b_this_page;
	} while (bh);
	tail->b_this_page = head;
	attach_page_buffers(page, head);
}

/*
 * Initialise the state of a blockdev page's buffers.
 */ 
static void
init_page_buffers(struct page *page, struct block_device *bdev,
			sector_t block, int size)
{
	struct buffer_head *head = page_buffers(page);
	struct buffer_head *bh = head;
	int uptodate = PageUptodate(page);

	do {
		if (!buffer_mapped(bh)) {
			init_buffer(bh, NULL, NULL);
			bh->b_bdev = bdev;
			bh->b_blocknr = block;
			if (uptodate)
				set_buffer_uptodate(bh);
			set_buffer_mapped(bh);
		}
		block++;
		bh = bh->b_this_page;
	} while (bh != head);
}

/*
 * Create the page-cache page that contains the requested block.
 *
 * This is user purely for blockdev mappings.
 */
static struct page *
grow_dev_page(struct block_device *bdev, sector_t block,
		pgoff_t index, int size)
{
	struct inode *inode = bdev->bd_inode;
	struct page *page;