xfs_aops.c

LINUX 2.6.17.4的源码

	xfs_iomap_t		*mp,
	xfs_off_t		offset,
	uint			block_bits)
{
	sector_t		bn;

	ASSERT(mp->iomap_bn != IOMAP_DADDR_NULL);

	bn = (mp->iomap_bn >> (block_bits - BBSHIFT)) +
	      ((offset - mp->iomap_offset) >> block_bits);

	ASSERT(bn || (mp->iomap_flags & IOMAP_REALTIME));

	bh->b_blocknr = bn;
	set_buffer_mapped(bh);
}

STATIC void
xfs_map_at_offset(
	struct buffer_head	*bh,
	loff_t			offset,
	int			block_bits,
	xfs_iomap_t		*iomapp)
{
	ASSERT(!(iomapp->iomap_flags & IOMAP_HOLE));
	ASSERT(!(iomapp->iomap_flags & IOMAP_DELAY));

	lock_buffer(bh);
	xfs_map_buffer(bh, iomapp, offset, block_bits);
	bh->b_bdev = iomapp->iomap_target->bt_bdev;
	set_buffer_mapped(bh);
	clear_buffer_delay(bh);
	clear_buffer_unwritten(bh);
}

/*
 * Look for a page at index that is suitable for clustering.
 */
STATIC unsigned int
xfs_probe_page(
	struct page		*page,
	unsigned int		pg_offset,
	int			mapped)
{
	int			ret = 0;

	if (PageWriteback(page))
		return 0;

	if (page->mapping && PageDirty(page)) {
		if (page_has_buffers(page)) {
			struct buffer_head	*bh, *head;

			bh = head = page_buffers(page);
			do {
				if (!buffer_uptodate(bh))
					break;
				if (mapped != buffer_mapped(bh))
					break;
				ret += bh->b_size;
				if (ret >= pg_offset)
					break;
			} while ((bh = bh->b_this_page) != head);
		} else
			ret = mapped ? 0 : PAGE_CACHE_SIZE;
	}

	return ret;
}

STATIC size_t
xfs_probe_cluster(
	struct inode		*inode,
	struct page		*startpage,
	struct buffer_head	*bh,
	struct buffer_head	*head,
	int			mapped)
{
	struct pagevec		pvec;
	pgoff_t			tindex, tlast, tloff;
	size_t			total = 0;
	int			done = 0, i;

	/* First sum forwards in this page */
	do {
		if (!buffer_uptodate(bh) || (mapped != buffer_mapped(bh)))
			return total;
		total += bh->b_size;
	} while ((bh = bh->b_this_page) != head);

	/* if we reached the end of the page, sum forwards in following pages */
	tlast = i_size_read(inode) >> PAGE_CACHE_SHIFT;
	tindex = startpage->index + 1;

	/* Prune this back to avoid pathological behavior */
	tloff = min(tlast, startpage->index + 64);

	pagevec_init(&pvec, 0);
	while (!done && tindex <= tloff) {
		unsigned len = min_t(pgoff_t, PAGEVEC_SIZE, tlast - tindex + 1);

		if (!pagevec_lookup(&pvec, inode->i_mapping, tindex, len))
			break;

		for (i = 0; i < pagevec_count(&pvec); i++) {
			struct page *page = pvec.pages[i];
			size_t pg_offset, len = 0;

			if (tindex == tlast) {
				pg_offset =
				    i_size_read(inode) & (PAGE_CACHE_SIZE - 1);
				if (!pg_offset) {
					done = 1;
					break;
				}
			} else
				pg_offset = PAGE_CACHE_SIZE;

			if (page->index == tindex && !TestSetPageLocked(page)) {
				len = xfs_probe_page(page, pg_offset, mapped);
				unlock_page(page);
			}

			if (!len) {
				done = 1;
				break;
			}

			total += len;
			tindex++;
		}

		pagevec_release(&pvec);
		cond_resched();
	}

	return total;
}

/*
 * Test if a given page is suitable for writing as part of an unwritten
 * or delayed allocate extent.
 */
STATIC int
xfs_is_delayed_page(
	struct page		*page,
	unsigned int		type)
{
	if (PageWriteback(page))
		return 0;

	if (page->mapping && page_has_buffers(page)) {
		struct buffer_head	*bh, *head;
		int			acceptable = 0;

		bh = head = page_buffers(page);
		do {
			if (buffer_unwritten(bh))
				acceptable = (type == IOMAP_UNWRITTEN);
			else if (buffer_delay(bh))
				acceptable = (type == IOMAP_DELAY);
			else if (buffer_dirty(bh) && buffer_mapped(bh))
				acceptable = (type == 0);
			else
				break;
		} while ((bh = bh->b_this_page) != head);

		if (acceptable)
			return 1;
	}

	return 0;
}

/*
 * Allocate & map buffers for page given the extent map. Write it out.
 * except for the original page of a writepage, this is called on
 * delalloc/unwritten pages only, for the original page it is possible
 * that the page has no mapping at all.
 */
STATIC int
xfs_convert_page(
	struct inode		*inode,
	struct page		*page,
	loff_t			tindex,
	xfs_iomap_t		*mp,
	xfs_ioend_t		**ioendp,
	struct writeback_control *wbc,
	int			startio,
	int			all_bh)
{
	struct buffer_head	*bh, *head;
	xfs_off_t		end_offset;
	unsigned long		p_offset;
	unsigned int		type;
	int			bbits = inode->i_blkbits;
	int			len, page_dirty;
	int			count = 0, done = 0, uptodate = 1;
 	xfs_off_t		offset = page_offset(page);

	if (page->index != tindex)
		goto fail;
	if (TestSetPageLocked(page))
		goto fail;
	if (PageWriteback(page))
		goto fail_unlock_page;
	if (page->mapping != inode->i_mapping)
		goto fail_unlock_page;
	if (!xfs_is_delayed_page(page, (*ioendp)->io_type))
		goto fail_unlock_page;

	/*
	 * page_dirty is initially a count of buffers on the page before
	 * EOF and is decremented as we move each into a cleanable state.
	 *
	 * Derivation:
	 *
	 * End offset is the highest offset that this page should represent.
	 * If we are on the last page, (end_offset & (PAGE_CACHE_SIZE - 1))
	 * will evaluate non-zero and be less than PAGE_CACHE_SIZE and
	 * hence give us the correct page_dirty count. On any other page,
	 * it will be zero and in that case we need page_dirty to be the
	 * count of buffers on the page.
	 */
	end_offset = min_t(unsigned long long,
			(xfs_off_t)(page->index + 1) << PAGE_CACHE_SHIFT,
			i_size_read(inode));

	len = 1 << inode->i_blkbits;
	p_offset = min_t(unsigned long, end_offset & (PAGE_CACHE_SIZE - 1),
					PAGE_CACHE_SIZE);
	p_offset = p_offset ? roundup(p_offset, len) : PAGE_CACHE_SIZE;
	page_dirty = p_offset / len;

	bh = head = page_buffers(page);
	do {
		if (offset >= end_offset)
			break;
		if (!buffer_uptodate(bh))
			uptodate = 0;
		if (!(PageUptodate(page) || buffer_uptodate(bh))) {
			done = 1;
			continue;
		}

		if (buffer_unwritten(bh) || buffer_delay(bh)) {
			if (buffer_unwritten(bh))
				type = IOMAP_UNWRITTEN;
			else
				type = IOMAP_DELAY;

			if (!xfs_iomap_valid(mp, offset)) {
				done = 1;
				continue;
			}

			ASSERT(!(mp->iomap_flags & IOMAP_HOLE));
			ASSERT(!(mp->iomap_flags & IOMAP_DELAY));

			xfs_map_at_offset(bh, offset, bbits, mp);
			if (startio) {
				xfs_add_to_ioend(inode, bh, offset,
						type, ioendp, done);
			} else {
				set_buffer_dirty(bh);
				unlock_buffer(bh);
				mark_buffer_dirty(bh);
			}
			page_dirty--;
			count++;
		} else {
			type = 0;
			if (buffer_mapped(bh) && all_bh && startio) {
				lock_buffer(bh);
				xfs_add_to_ioend(inode, bh, offset,
						type, ioendp, done);
				count++;
				page_dirty--;
			} else {
				done = 1;
			}
		}
	} while (offset += len, (bh = bh->b_this_page) != head);

	if (uptodate && bh == head)
		SetPageUptodate(page);

	if (startio) {
		if (count) {
			struct backing_dev_info *bdi;

			bdi = inode->i_mapping->backing_dev_info;
			wbc->nr_to_write--;
			if (bdi_write_congested(bdi)) {
				wbc->encountered_congestion = 1;
				done = 1;
			} else if (wbc->nr_to_write <= 0) {
				done = 1;
			}
		}
		xfs_start_page_writeback(page, wbc, !page_dirty, count);
	}

	return done;
 fail_unlock_page:
	unlock_page(page);
 fail:
	return 1;
}

/*
 * Convert & write out a cluster of pages in the same extent as defined
 * by mp and following the start page.
 */
STATIC void
xfs_cluster_write(
	struct inode		*inode,
	pgoff_t			tindex,
	xfs_iomap_t		*iomapp,
	xfs_ioend_t		**ioendp,
	struct writeback_control *wbc,
	int			startio,
	int			all_bh,
	pgoff_t			tlast)
{
	struct pagevec		pvec;
	int			done = 0, i;

	pagevec_init(&pvec, 0);
	while (!done && tindex <= tlast) {
		unsigned len = min_t(pgoff_t, PAGEVEC_SIZE, tlast - tindex + 1);

		if (!pagevec_lookup(&pvec, inode->i_mapping, tindex, len))
			break;

		for (i = 0; i < pagevec_count(&pvec); i++) {
			done = xfs_convert_page(inode, pvec.pages[i], tindex++,
					iomapp, ioendp, wbc, startio, all_bh);
			if (done)
				break;
		}

		pagevec_release(&pvec);
		cond_resched();
	}
}

/*
 * Calling this without startio set means we are being asked to make a dirty
 * page ready for freeing it's buffers.  When called with startio set then
 * we are coming from writepage.
 *
 * When called with startio set it is important that we write the WHOLE
 * page if possible.
 * The bh->b_state's cannot know if any of the blocks or which block for
 * that matter are dirty due to mmap writes, and therefore bh uptodate is
 * only valid if the page itself isn't completely uptodate.  Some layers
 * may clear the page dirty flag prior to calling write page, under the
 * assumption the entire page will be written out; by not writing out the
 * whole page the page can be reused before all valid dirty data is
 * written out.  Note: in the case of a page that has been dirty'd by
 * mapwrite and but partially setup by block_prepare_write the
 * bh->b_states's will not agree and only ones setup by BPW/BCW will have
 * valid state, thus the whole page must be written out thing.
 */

STATIC int
xfs_page_state_convert(
	struct inode	*inode,
	struct page	*page,
	struct writeback_control *wbc,
	int		startio,
	int		unmapped) /* also implies page uptodate */
{
	struct buffer_head	*bh, *head;
	xfs_iomap_t		iomap;
	xfs_ioend_t		*ioend = NULL, *iohead = NULL;
	loff_t			offset;
	unsigned long           p_offset = 0;
	unsigned int		type;
	__uint64_t              end_offset;
	pgoff_t                 end_index, last_index, tlast;
	ssize_t			size, len;
	int			flags, err, iomap_valid = 0, uptodate = 1;
	int			page_dirty, count = 0;
	int			trylock = 0;
	int			all_bh = unmapped;

	if (startio) {
		if (wbc->sync_mode == WB_SYNC_NONE && wbc->nonblocking)
			trylock |= BMAPI_TRYLOCK;
	}

	/* Is this page beyond the end of the file? */
	offset = i_size_read(inode);
	end_index = offset >> PAGE_CACHE_SHIFT;
	last_index = (offset - 1) >> PAGE_CACHE_SHIFT;
	if (page->index >= end_index) {
		if ((page->index >= end_index + 1) ||
		    !(i_size_read(inode) & (PAGE_CACHE_SIZE - 1))) {
			if (startio)
				unlock_page(page);
			return 0;
		}
	}

	/*
	 * page_dirty is initially a count of buffers on the page before
	 * EOF and is decremented as we move each into a cleanable state.
	 *
	 * Derivation:
	 *
	 * End offset is the highest offset that this page should represent.
	 * If we are on the last page, (end_offset & (PAGE_CACHE_SIZE - 1))
	 * will evaluate non-zero and be less than PAGE_CACHE_SIZE and
	 * hence give us the correct page_dirty count. On any other page,
	 * it will be zero and in that case we need page_dirty to be the
	 * count of buffers on the page.
 	 */
	end_offset = min_t(unsigned long long,
			(xfs_off_t)(page->index + 1) << PAGE_CACHE_SHIFT, offset);
	len = 1 << inode->i_blkbits;
	p_offset = min_t(unsigned long, end_offset & (PAGE_CACHE_SIZE - 1),
					PAGE_CACHE_SIZE);
	p_offset = p_offset ? roundup(p_offset, len) : PAGE_CACHE_SIZE;
	page_dirty = p_offset / len;

	bh = head = page_buffers(page);
	offset = page_offset(page);
	flags = -1;
	type = 0;

	/* TODO: cleanup count and page_dirty */

	do {
		if (offset >= end_offset)
			break;
		if (!buffer_uptodate(bh))
			uptodate = 0;
		if (!(PageUptodate(page) || buffer_uptodate(bh)) && !startio) {
			/*
			 * the iomap is actually still valid, but the ioend
			 * isn't.  shouldn't happen too often.
			 */
			iomap_valid = 0;
			continue;
		}

		if (iomap_valid)
			iomap_valid = xfs_iomap_valid(&iomap, offset);

		/*
		 * First case, map an unwritten extent and prepare for
		 * extent state conversion transaction on completion.
		 *
		 * Second case, allocate space for a delalloc buffer.
		 * We can return EAGAIN here in the release page case.
		 *
		 * Third case, an unmapped buffer was found, and we are
		 * in a path where we need to write the whole page out.
 		 */
		if (buffer_unwritten(bh) || buffer_delay(bh) ||
		    ((buffer_uptodate(bh) || PageUptodate(page)) &&
		     !buffer_mapped(bh) && (unmapped || startio))) {
		     	/*
			 * Make sure we don't use a read-only iomap
			 */
		     	if (flags == BMAPI_READ)
				iomap_valid = 0;

			if (buffer_unwritten(bh)) {
				type = IOMAP_UNWRITTEN;
				flags = BMAPI_WRITE | BMAPI_IGNSTATE;
			} else if (buffer_delay(bh)) {
				type = IOMAP_DELAY;
				flags = BMAPI_ALLOCATE | trylock;
			} else {
				type = IOMAP_NEW;
				flags = BMAPI_WRITE | BMAPI_MMAP;
			}

			if (!iomap_valid) {
				if (type == IOMAP_NEW) {
					size = xfs_probe_cluster(inode,
							page, bh, head, 0);
				} else {
					size = len;
				}