[linux.git] / fs / gfs2 / meta_io.c

/*
 * Copyright (C) Sistina Software, Inc.  1997-2003 All rights reserved.
 * Copyright (C) 2004-2008 Red Hat, Inc.  All rights reserved.
 *
 * This copyrighted material is made available to anyone wishing to use,
 * modify, copy, or redistribute it subject to the terms and conditions
 * of the GNU General Public License version 2.
 */

#include <linux/sched.h>
#include <linux/slab.h>
#include <linux/spinlock.h>
#include <linux/completion.h>
#include <linux/buffer_head.h>
#include <linux/mm.h>
#include <linux/pagemap.h>
#include <linux/writeback.h>
#include <linux/swap.h>
#include <linux/delay.h>
#include <linux/bio.h>
#include <linux/gfs2_ondisk.h>

#include "gfs2.h"
#include "incore.h"
#include "glock.h"
#include "glops.h"
#include "inode.h"
#include "log.h"
#include "lops.h"
#include "meta_io.h"
#include "rgrp.h"
#include "trans.h"
#include "util.h"
#include "trace_gfs2.h"

static int gfs2_aspace_writepage(struct page *page, struct writeback_control *wbc)
{
	struct buffer_head *bh, *head;
	int nr_underway = 0;
	int write_op = REQ_META | REQ_PRIO |
		(wbc->sync_mode == WB_SYNC_ALL ? WRITE_SYNC : WRITE);

	BUG_ON(!PageLocked(page));
	BUG_ON(!page_has_buffers(page));

	head = page_buffers(page);
	bh = head;

	do {
		if (!buffer_mapped(bh))
			continue;
		/*
		 * If it's a fully non-blocking write attempt and we cannot
		 * lock the buffer then redirty the page.  Note that this can
		 * potentially cause a busy-wait loop from pdflush and kswapd
		 * activity, but those code paths have their own higher-level
		 * throttling.
		 */
		if (wbc->sync_mode != WB_SYNC_NONE) {
			lock_buffer(bh);
		} else if (!trylock_buffer(bh)) {
			redirty_page_for_writepage(wbc, page);
			continue;
		}
		if (test_clear_buffer_dirty(bh)) {
			mark_buffer_async_write(bh);
		} else {
			unlock_buffer(bh);
		}
	} while ((bh = bh->b_this_page) != head);

	/*
	 * The page and its buffers are protected by PageWriteback(), so we can
	 * drop the bh refcounts early.
	 */
	BUG_ON(PageWriteback(page));
	set_page_writeback(page);

	do {
		struct buffer_head *next = bh->b_this_page;
		if (buffer_async_write(bh)) {
			submit_bh(write_op, bh);
			nr_underway++;
		}
		bh = next;
	} while (bh != head);
	unlock_page(page);

	if (nr_underway == 0)
		end_page_writeback(page);

	return 0;
}

const struct address_space_operations gfs2_meta_aops = {
	.writepage = gfs2_aspace_writepage,
	.releasepage = gfs2_releasepage,
};

/**
 * gfs2_meta_sync - Sync all buffers associated with a glock
 * @gl: The glock
 *
 */

void gfs2_meta_sync(struct gfs2_glock *gl)
{
	struct address_space *mapping = gfs2_glock2aspace(gl);
	int error;

	filemap_fdatawrite(mapping);
	error = filemap_fdatawait(mapping);

	if (error)
		gfs2_io_error(gl->gl_sbd);
}

/**
 * gfs2_getbuf - Get a buffer with a given address space
 * @gl: the glock
 * @blkno: the block number (filesystem scope)
 * @create: 1 if the buffer should be created
 *
 * Returns: the buffer
 */

struct buffer_head *gfs2_getbuf(struct gfs2_glock *gl, u64 blkno, int create)
{
	struct address_space *mapping = gfs2_glock2aspace(gl);
	struct gfs2_sbd *sdp = gl->gl_sbd;
	struct page *page;
	struct buffer_head *bh;
	unsigned int shift;
	unsigned long index;
	unsigned int bufnum;

	shift = PAGE_CACHE_SHIFT - sdp->sd_sb.sb_bsize_shift;
	index = blkno >> shift;             /* convert block to page */
	bufnum = blkno - (index << shift);  /* block buf index within page */

	if (create) {
		for (;;) {
			page = grab_cache_page(mapping, index);
			if (page)
				break;
			yield();
		}
	} else {
		page = find_lock_page(mapping, index);
		if (!page)
			return NULL;
	}

	if (!page_has_buffers(page))
		create_empty_buffers(page, sdp->sd_sb.sb_bsize, 0);

	/* Locate header for our buffer within our page */
	for (bh = page_buffers(page); bufnum--; bh = bh->b_this_page)
		/* Do nothing */;
	get_bh(bh);

	if (!buffer_mapped(bh))
		map_bh(bh, sdp->sd_vfs, blkno);

	unlock_page(page);
	mark_page_accessed(page);
	page_cache_release(page);

	return bh;
}

static void meta_prep_new(struct buffer_head *bh)
{
	struct gfs2_meta_header *mh = (struct gfs2_meta_header *)bh->b_data;

	lock_buffer(bh);
	clear_buffer_dirty(bh);
	set_buffer_uptodate(bh);
	unlock_buffer(bh);

	mh->mh_magic = cpu_to_be32(GFS2_MAGIC);
}

/**
 * gfs2_meta_new - Get a block
 * @gl: The glock associated with this block
 * @blkno: The block number
 *
 * Returns: The buffer
 */

struct buffer_head *gfs2_meta_new(struct gfs2_glock *gl, u64 blkno)
{
	struct buffer_head *bh;
	bh = gfs2_getbuf(gl, blkno, CREATE);
	meta_prep_new(bh);
	return bh;
}

/**
 * gfs2_meta_read - Read a block from disk
 * @gl: The glock covering the block
 * @blkno: The block number
 * @flags: flags
 * @bhp: the place where the buffer is returned (NULL on failure)
 *
 * Returns: errno
 */

int gfs2_meta_read(struct gfs2_glock *gl, u64 blkno, int flags,
		   struct buffer_head **bhp)
{
	struct gfs2_sbd *sdp = gl->gl_sbd;
	struct buffer_head *bh;

	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
		return -EIO;

	*bhp = bh = gfs2_getbuf(gl, blkno, CREATE);

	lock_buffer(bh);
	if (buffer_uptodate(bh)) {
		unlock_buffer(bh);
		return 0;
	}
	bh->b_end_io = end_buffer_read_sync;
	get_bh(bh);
	submit_bh(READ_SYNC | REQ_META | REQ_PRIO, bh);
	if (!(flags & DIO_WAIT))
		return 0;

	wait_on_buffer(bh);
	if (unlikely(!buffer_uptodate(bh))) {
		struct gfs2_trans *tr = current->journal_info;
		if (tr && tr->tr_touched)
			gfs2_io_error_bh(sdp, bh);
		brelse(bh);
		return -EIO;
	}

	return 0;
}

/**
 * gfs2_meta_wait - Reread a block from disk
 * @sdp: the filesystem
 * @bh: The block to wait for
 *
 * Returns: errno
 */

int gfs2_meta_wait(struct gfs2_sbd *sdp, struct buffer_head *bh)
{
	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
		return -EIO;

	wait_on_buffer(bh);

	if (!buffer_uptodate(bh)) {
		struct gfs2_trans *tr = current->journal_info;
		if (tr && tr->tr_touched)
			gfs2_io_error_bh(sdp, bh);
		return -EIO;
	}
	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
		return -EIO;

	return 0;
}

/**
 * gfs2_attach_bufdata - attach a struct gfs2_bufdata structure to a buffer
 * @gl: the glock the buffer belongs to
 * @bh: The buffer to be attached to
 * @meta: Flag to indicate whether its metadata or not
 */

void gfs2_attach_bufdata(struct gfs2_glock *gl, struct buffer_head *bh,
			 int meta)
{
	struct gfs2_bufdata *bd;

	if (meta)
		lock_page(bh->b_page);

	if (bh->b_private) {
		if (meta)
			unlock_page(bh->b_page);
		return;
	}

	bd = kmem_cache_zalloc(gfs2_bufdata_cachep, GFP_NOFS | __GFP_NOFAIL);
	bd->bd_bh = bh;
	bd->bd_gl = gl;

	if (meta)
		lops_init_le(bd, &gfs2_buf_lops);
	else
		lops_init_le(bd, &gfs2_databuf_lops);
	bh->b_private = bd;

	if (meta)
		unlock_page(bh->b_page);
}

void gfs2_remove_from_journal(struct buffer_head *bh, struct gfs2_trans *tr, int meta)
{
	struct address_space *mapping = bh->b_page->mapping;
	struct gfs2_sbd *sdp = gfs2_mapping2sbd(mapping);
	struct gfs2_bufdata *bd = bh->b_private;

	if (test_clear_buffer_pinned(bh)) {
		trace_gfs2_pin(bd, 0);
		atomic_dec(&sdp->sd_log_pinned);
		list_del_init(&bd->bd_list);
		if (meta) {
			gfs2_assert_warn(sdp, sdp->sd_log_num_buf);
			sdp->sd_log_num_buf--;
			tr->tr_num_buf_rm++;
		} else {
			gfs2_assert_warn(sdp, sdp->sd_log_num_databuf);
			sdp->sd_log_num_databuf--;
			tr->tr_num_databuf_rm++;
		}
		tr->tr_touched = 1;
		brelse(bh);
	}
	if (bd) {
		spin_lock(&sdp->sd_ail_lock);
		if (bd->bd_ail) {
			gfs2_remove_from_ail(bd);
			bh->b_private = NULL;
			bd->bd_bh = NULL;
			bd->bd_blkno = bh->b_blocknr;
			gfs2_trans_add_revoke(sdp, bd);
		}
		spin_unlock(&sdp->sd_ail_lock);
	}
	clear_buffer_dirty(bh);
	clear_buffer_uptodate(bh);
}

/**
 * gfs2_meta_wipe - make inode's buffers so they aren't dirty/pinned anymore
 * @ip: the inode who owns the buffers
 * @bstart: the first buffer in the run
 * @blen: the number of buffers in the run
 *
 */

void gfs2_meta_wipe(struct gfs2_inode *ip, u64 bstart, u32 blen)
{
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
	struct buffer_head *bh;

	while (blen) {
		bh = gfs2_getbuf(ip->i_gl, bstart, NO_CREATE);
		if (bh) {
			lock_buffer(bh);
			gfs2_log_lock(sdp);
			gfs2_remove_from_journal(bh, current->journal_info, 1);
			gfs2_log_unlock(sdp);
			unlock_buffer(bh);
			brelse(bh);
		}

		bstart++;
		blen--;
	}
}

/**
 * gfs2_meta_indirect_buffer - Get a metadata buffer
 * @ip: The GFS2 inode
 * @height: The level of this buf in the metadata (indir addr) tree (if any)
 * @num: The block number (device relative) of the buffer
 * @bhp: the buffer is returned here
 *
 * Returns: errno
 */

int gfs2_meta_indirect_buffer(struct gfs2_inode *ip, int height, u64 num,
			      struct buffer_head **bhp)
{
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
	struct gfs2_glock *gl = ip->i_gl;
	struct buffer_head *bh;
	int ret = 0;
	u32 mtype = height ? GFS2_METATYPE_IN : GFS2_METATYPE_DI;

	ret = gfs2_meta_read(gl, num, DIO_WAIT, &bh);
	if (ret == 0 && gfs2_metatype_check(sdp, bh, mtype)) {
		brelse(bh);
		ret = -EIO;
	}
	*bhp = bh;
	return ret;
}

/**
 * gfs2_meta_ra - start readahead on an extent of a file
 * @gl: the glock the blocks belong to
 * @dblock: the starting disk block
 * @extlen: the number of blocks in the extent
 *
 * returns: the first buffer in the extent
 */

struct buffer_head *gfs2_meta_ra(struct gfs2_glock *gl, u64 dblock, u32 extlen)
{
	struct gfs2_sbd *sdp = gl->gl_sbd;
	struct buffer_head *first_bh, *bh;
	u32 max_ra = gfs2_tune_get(sdp, gt_max_readahead) >>
			  sdp->sd_sb.sb_bsize_shift;

	BUG_ON(!extlen);

	if (max_ra < 1)
		max_ra = 1;
	if (extlen > max_ra)
		extlen = max_ra;

	first_bh = gfs2_getbuf(gl, dblock, CREATE);

	if (buffer_uptodate(first_bh))
		goto out;
	if (!buffer_locked(first_bh))
		ll_rw_block(READ_SYNC | REQ_META, 1, &first_bh);

	dblock++;
	extlen--;

	while (extlen) {
		bh = gfs2_getbuf(gl, dblock, CREATE);

		if (!buffer_uptodate(bh) && !buffer_locked(bh))
			ll_rw_block(READA | REQ_META, 1, &bh);
		brelse(bh);
		dblock++;
		extlen--;
		if (!buffer_locked(first_bh) && buffer_uptodate(first_bh))
			goto out;
	}

	wait_on_buffer(first_bh);
out:
	return first_bh;
}
Commit	Line	Data
b3b94faa DT	1	/*
b3b94faa DT	2	* Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved.
091806ed	3	* Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved.
b3b94faa DT	4	*
	5	* This copyrighted material is made available to anyone wishing to use,
	6	* modify, copy, or redistribute it subject to the terms and conditions
e9fc2aa0	7	* of the GNU General Public License version 2.
b3b94faa DT	8	*/
	9
	10	#include <linux/sched.h>
	11	#include <linux/slab.h>
	12	#include <linux/spinlock.h>
	13	#include <linux/completion.h>
	14	#include <linux/buffer_head.h>
	15	#include <linux/mm.h>
	16	#include <linux/pagemap.h>
	17	#include <linux/writeback.h>
	18	#include <linux/swap.h>
	19	#include <linux/delay.h>
2e565bb6	20	#include <linux/bio.h>
5c676f6d	21	#include <linux/gfs2_ondisk.h>
b3b94faa DT	22
b3b94faa DT	23	#include "gfs2.h"
5c676f6d	24	#include "incore.h"
b3b94faa DT	25	#include "glock.h"
	26	#include "glops.h"
	27	#include "inode.h"
	28	#include "log.h"
	29	#include "lops.h"
	30	#include "meta_io.h"
	31	#include "rgrp.h"
	32	#include "trans.h"
5c676f6d	33	#include "util.h"
627c10b7	34	#include "trace_gfs2.h"
b3b94faa	35
4a0f9a32	36	static int gfs2_aspace_writepage(struct page page, struct writeback_control wbc)
b3b94faa	37	{
4a0f9a32 SW	38	struct buffer_head bh, head;
4a0f9a32 SW	39	int nr_underway = 0;
65299a3b	40	int write_op = REQ_META \| REQ_PRIO \|
721a9602	41	(wbc->sync_mode == WB_SYNC_ALL ? WRITE_SYNC : WRITE);
b3b94faa	42
4a0f9a32 SW	43	BUG_ON(!PageLocked(page));
	44	BUG_ON(!page_has_buffers(page));
	45
	46	head = page_buffers(page);
	47	bh = head;
	48
	49	do {
	50	if (!buffer_mapped(bh))
	51	continue;
	52	/*
	53	* If it's a fully non-blocking write attempt and we cannot
	54	* lock the buffer then redirty the page. Note that this can
	55	* potentially cause a busy-wait loop from pdflush and kswapd
	56	* activity, but those code paths have their own higher-level
	57	* throttling.
	58	*/
1b430bee	59	if (wbc->sync_mode != WB_SYNC_NONE) {
4a0f9a32 SW	60	lock_buffer(bh);
	61	} else if (!trylock_buffer(bh)) {
	62	redirty_page_for_writepage(wbc, page);
	63	continue;
	64	}
	65	if (test_clear_buffer_dirty(bh)) {
	66	mark_buffer_async_write(bh);
	67	} else {
	68	unlock_buffer(bh);
	69	}
	70	} while ((bh = bh->b_this_page) != head);
	71
	72	/*
	73	* The page and its buffers are protected by PageWriteback(), so we can
	74	* drop the bh refcounts early.
	75	*/
	76	BUG_ON(PageWriteback(page));
	77	set_page_writeback(page);
	78
	79	do {
	80	struct buffer_head *next = bh->b_this_page;
	81	if (buffer_async_write(bh)) {
	82	submit_bh(write_op, bh);
	83	nr_underway++;
	84	}
	85	bh = next;
	86	} while (bh != head);
	87	unlock_page(page);
	88
4a0f9a32 SW	89	if (nr_underway == 0)
	90	end_page_writeback(page);
	91
eaefbf96	92	return 0;
b3b94faa DT	93	}
b3b94faa DT	94
009d8518	95	const struct address_space_operations gfs2_meta_aops = {
b3b94faa	96	.writepage = gfs2_aspace_writepage,
4340fe62	97	.releasepage = gfs2_releasepage,
b3b94faa DT	98	};
b3b94faa DT	99
b3b94faa DT	100	/**
	101	* gfs2_meta_sync - Sync all buffers associated with a glock
	102	* @gl: The glock
b3b94faa DT	103	*
	104	*/
	105
7276b3b0	106	void gfs2_meta_sync(struct gfs2_glock *gl)
b3b94faa	107	{
009d8518	108	struct address_space *mapping = gfs2_glock2aspace(gl);
7276b3b0	109	int error;
b3b94faa	110
7276b3b0 SW	111	filemap_fdatawrite(mapping);
7276b3b0 SW	112	error = filemap_fdatawait(mapping);
b3b94faa DT	113
	114	if (error)
	115	gfs2_io_error(gl->gl_sbd);
	116	}
	117
	118	/**
6802e340	119	* gfs2_getbuf - Get a buffer with a given address space
cb4c0313	120	* @gl: the glock
b3b94faa DT	121	* @blkno: the block number (filesystem scope)
	122	* @create: 1 if the buffer should be created
	123	*
	124	* Returns: the buffer
	125	*/
	126
6802e340	127	struct buffer_head gfs2_getbuf(struct gfs2_glock gl, u64 blkno, int create)
b3b94faa	128	{
009d8518	129	struct address_space *mapping = gfs2_glock2aspace(gl);
cb4c0313	130	struct gfs2_sbd *sdp = gl->gl_sbd;
b3b94faa DT	131	struct page *page;
	132	struct buffer_head *bh;
	133	unsigned int shift;
	134	unsigned long index;
	135	unsigned int bufnum;
	136
	137	shift = PAGE_CACHE_SHIFT - sdp->sd_sb.sb_bsize_shift;
	138	index = blkno >> shift; /* convert block to page */
	139	bufnum = blkno - (index << shift); /* block buf index within page */
	140
	141	if (create) {
	142	for (;;) {
cb4c0313	143	page = grab_cache_page(mapping, index);
b3b94faa DT	144	if (page)
	145	break;
	146	yield();
	147	}
	148	} else {
cb4c0313	149	page = find_lock_page(mapping, index);
b3b94faa DT	150	if (!page)
	151	return NULL;
	152	}
	153
	154	if (!page_has_buffers(page))
	155	create_empty_buffers(page, sdp->sd_sb.sb_bsize, 0);
	156
	157	/* Locate header for our buffer within our page */
	158	for (bh = page_buffers(page); bufnum--; bh = bh->b_this_page)
	159	/* Do nothing */;
	160	get_bh(bh);
	161
	162	if (!buffer_mapped(bh))
	163	map_bh(bh, sdp->sd_vfs, blkno);
	164
	165	unlock_page(page);
	166	mark_page_accessed(page);
	167	page_cache_release(page);
	168
	169	return bh;
	170	}
	171
	172	static void meta_prep_new(struct buffer_head *bh)
	173	{
	174	struct gfs2_meta_header mh = (struct gfs2_meta_header )bh->b_data;
	175
	176	lock_buffer(bh);
	177	clear_buffer_dirty(bh);
	178	set_buffer_uptodate(bh);
	179	unlock_buffer(bh);
	180
	181	mh->mh_magic = cpu_to_be32(GFS2_MAGIC);
	182	}
	183
	184	/**
	185	* gfs2_meta_new - Get a block
	186	* @gl: The glock associated with this block
	187	* @blkno: The block number
	188	*
	189	* Returns: The buffer
	190	*/
	191
cd915493	192	struct buffer_head gfs2_meta_new(struct gfs2_glock gl, u64 blkno)
b3b94faa DT	193	{
b3b94faa DT	194	struct buffer_head *bh;
6802e340	195	bh = gfs2_getbuf(gl, blkno, CREATE);
b3b94faa DT	196	meta_prep_new(bh);
	197	return bh;
	198	}
	199
	200	/**
	201	* gfs2_meta_read - Read a block from disk
	202	* @gl: The glock covering the block
	203	* @blkno: The block number
7276b3b0	204	* @flags: flags
b3b94faa DT	205	* @bhp: the place where the buffer is returned (NULL on failure)
	206	*
	207	* Returns: errno
	208	*/
	209
cd915493	210	int gfs2_meta_read(struct gfs2_glock *gl, u64 blkno, int flags,
b3b94faa DT	211	struct buffer_head **bhp)
b3b94faa DT	212	{
c969f58c SW	213	struct gfs2_sbd *sdp = gl->gl_sbd;
	214	struct buffer_head *bh;
	215
	216	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
	217	return -EIO;
	218
	219	*bhp = bh = gfs2_getbuf(gl, blkno, CREATE);
	220
	221	lock_buffer(bh);
	222	if (buffer_uptodate(bh)) {
	223	unlock_buffer(bh);
	224	return 0;
	225	}
	226	bh->b_end_io = end_buffer_read_sync;
	227	get_bh(bh);
65299a3b	228	submit_bh(READ_SYNC \| REQ_META \| REQ_PRIO, bh);
c969f58c SW	229	if (!(flags & DIO_WAIT))
	230	return 0;
	231
	232	wait_on_buffer(bh);
	233	if (unlikely(!buffer_uptodate(bh))) {
	234	struct gfs2_trans *tr = current->journal_info;
	235	if (tr && tr->tr_touched)
	236	gfs2_io_error_bh(sdp, bh);
	237	brelse(bh);
	238	return -EIO;
7276b3b0	239	}
b3b94faa	240
7276b3b0	241	return 0;
b3b94faa DT	242	}
	243
	244	/**
7276b3b0	245	* gfs2_meta_wait - Reread a block from disk
b3b94faa	246	* @sdp: the filesystem
7276b3b0	247	* @bh: The block to wait for
b3b94faa DT	248	*
	249	* Returns: errno
	250	*/
	251
7276b3b0	252	int gfs2_meta_wait(struct gfs2_sbd sdp, struct buffer_head bh)
b3b94faa DT	253	{
	254	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
	255	return -EIO;
	256
7276b3b0	257	wait_on_buffer(bh);
b3b94faa	258
7276b3b0 SW	259	if (!buffer_uptodate(bh)) {
	260	struct gfs2_trans *tr = current->journal_info;
	261	if (tr && tr->tr_touched)
	262	gfs2_io_error_bh(sdp, bh);
	263	return -EIO;
b3b94faa	264	}
7276b3b0 SW	265	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
7276b3b0 SW	266	return -EIO;
b3b94faa DT	267
	268	return 0;
	269	}
	270
	271	/**
586dfdaa	272	* gfs2_attach_bufdata - attach a struct gfs2_bufdata structure to a buffer
b3b94faa DT	273	* @gl: the glock the buffer belongs to
b3b94faa DT	274	* @bh: The buffer to be attached to
586dfdaa	275	* @meta: Flag to indicate whether its metadata or not
b3b94faa DT	276	*/
b3b94faa DT	277
568f4c96 SW	278	void gfs2_attach_bufdata(struct gfs2_glock gl, struct buffer_head bh,
568f4c96 SW	279	int meta)
b3b94faa DT	280	{
	281	struct gfs2_bufdata *bd;
	282
18ec7d5c SW	283	if (meta)
18ec7d5c SW	284	lock_page(bh->b_page);
b3b94faa	285
5c676f6d	286	if (bh->b_private) {
18ec7d5c SW	287	if (meta)
18ec7d5c SW	288	unlock_page(bh->b_page);
b3b94faa DT	289	return;
	290	}
	291
3e5cd087	292	bd = kmem_cache_zalloc(gfs2_bufdata_cachep, GFP_NOFS \| __GFP_NOFAIL);
b3b94faa DT	293	bd->bd_bh = bh;
	294	bd->bd_gl = gl;
	295
82ffa516	296	if (meta)
c0752aa7	297	lops_init_le(bd, &gfs2_buf_lops);
82ffa516	298	else
c0752aa7	299	lops_init_le(bd, &gfs2_databuf_lops);
5c676f6d	300	bh->b_private = bd;
b3b94faa	301
18ec7d5c SW	302	if (meta)
18ec7d5c SW	303	unlock_page(bh->b_page);
b3b94faa DT	304	}
b3b94faa DT	305
16615be1 SW	306	void gfs2_remove_from_journal(struct buffer_head bh, struct gfs2_trans tr, int meta)
16615be1 SW	307	{
009d8518 SW	308	struct address_space *mapping = bh->b_page->mapping;
009d8518 SW	309	struct gfs2_sbd *sdp = gfs2_mapping2sbd(mapping);
16615be1	310	struct gfs2_bufdata *bd = bh->b_private;
009d8518	311
16615be1	312	if (test_clear_buffer_pinned(bh)) {
627c10b7	313	trace_gfs2_pin(bd, 0);
5e687eac	314	atomic_dec(&sdp->sd_log_pinned);
c0752aa7	315	list_del_init(&bd->bd_list);
16615be1 SW	316	if (meta) {
	317	gfs2_assert_warn(sdp, sdp->sd_log_num_buf);
	318	sdp->sd_log_num_buf--;
	319	tr->tr_num_buf_rm++;
	320	} else {
	321	gfs2_assert_warn(sdp, sdp->sd_log_num_databuf);
	322	sdp->sd_log_num_databuf--;
	323	tr->tr_num_databuf_rm++;
	324	}
	325	tr->tr_touched = 1;
	326	brelse(bh);
	327	}
	328	if (bd) {
c618e87a	329	spin_lock(&sdp->sd_ail_lock);
16615be1	330	if (bd->bd_ail) {
f91a0d3e	331	gfs2_remove_from_ail(bd);
16615be1 SW	332	bh->b_private = NULL;
	333	bd->bd_bh = NULL;
	334	bd->bd_blkno = bh->b_blocknr;
	335	gfs2_trans_add_revoke(sdp, bd);
	336	}
c618e87a	337	spin_unlock(&sdp->sd_ail_lock);
16615be1 SW	338	}
	339	clear_buffer_dirty(bh);
	340	clear_buffer_uptodate(bh);
	341	}
	342
b3b94faa DT	343	/**
	344	* gfs2_meta_wipe - make inode's buffers so they aren't dirty/pinned anymore
	345	* @ip: the inode who owns the buffers
	346	* @bstart: the first buffer in the run
	347	* @blen: the number of buffers in the run
	348	*
	349	*/
	350
cd915493	351	void gfs2_meta_wipe(struct gfs2_inode *ip, u64 bstart, u32 blen)
b3b94faa	352	{
feaa7bba	353	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
b3b94faa DT	354	struct buffer_head *bh;
	355
	356	while (blen) {
6802e340	357	bh = gfs2_getbuf(ip->i_gl, bstart, NO_CREATE);
b3b94faa	358	if (bh) {
1ad38c43 SW	359	lock_buffer(bh);
1ad38c43 SW	360	gfs2_log_lock(sdp);
16615be1	361	gfs2_remove_from_journal(bh, current->journal_info, 1);
1ad38c43	362	gfs2_log_unlock(sdp);
b3b94faa	363	unlock_buffer(bh);
b3b94faa DT	364	brelse(bh);
	365	}
	366
	367	bstart++;
	368	blen--;
	369	}
	370	}
	371
b3b94faa DT	372	/**
	373	* gfs2_meta_indirect_buffer - Get a metadata buffer
	374	* @ip: The GFS2 inode
	375	* @height: The level of this buf in the metadata (indir addr) tree (if any)
	376	* @num: The block number (device relative) of the buffer
b3b94faa DT	377	* @bhp: the buffer is returned here
b3b94faa DT	378	*
b3b94faa DT	379	* Returns: errno
	380	*/
	381
cd915493	382	int gfs2_meta_indirect_buffer(struct gfs2_inode *ip, int height, u64 num,
f2f9c812	383	struct buffer_head **bhp)
b3b94faa	384	{
7276b3b0 SW	385	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
7276b3b0 SW	386	struct gfs2_glock *gl = ip->i_gl;
f91a0d3e SW	387	struct buffer_head *bh;
f91a0d3e SW	388	int ret = 0;
f2f9c812	389	u32 mtype = height ? GFS2_METATYPE_IN : GFS2_METATYPE_DI;
b3b94faa	390
f2f9c812 BP	391	ret = gfs2_meta_read(gl, num, DIO_WAIT, &bh);
	392	if (ret == 0 && gfs2_metatype_check(sdp, bh, mtype)) {
	393	brelse(bh);
	394	ret = -EIO;
b3b94faa	395	}
b3b94faa	396	*bhp = bh;
f91a0d3e	397	return ret;
b3b94faa DT	398	}
	399
	400	/**
	401	* gfs2_meta_ra - start readahead on an extent of a file
	402	* @gl: the glock the blocks belong to
	403	* @dblock: the starting disk block
	404	* @extlen: the number of blocks in the extent
	405	*
7276b3b0	406	* returns: the first buffer in the extent
b3b94faa DT	407	*/
b3b94faa DT	408
7276b3b0	409	struct buffer_head gfs2_meta_ra(struct gfs2_glock gl, u64 dblock, u32 extlen)
b3b94faa DT	410	{
b3b94faa DT	411	struct gfs2_sbd *sdp = gl->gl_sbd;
b3b94faa	412	struct buffer_head first_bh, bh;
cd915493	413	u32 max_ra = gfs2_tune_get(sdp, gt_max_readahead) >>
568f4c96	414	sdp->sd_sb.sb_bsize_shift;
b3b94faa	415
7276b3b0 SW	416	BUG_ON(!extlen);
	417
	418	if (max_ra < 1)
	419	max_ra = 1;
b3b94faa DT	420	if (extlen > max_ra)
	421	extlen = max_ra;
	422
6802e340	423	first_bh = gfs2_getbuf(gl, dblock, CREATE);
b3b94faa DT	424
	425	if (buffer_uptodate(first_bh))
	426	goto out;
7276b3b0	427	if (!buffer_locked(first_bh))
20ed0535	428	ll_rw_block(READ_SYNC \| REQ_META, 1, &first_bh);
b3b94faa DT	429
	430	dblock++;
	431	extlen--;
	432
	433	while (extlen) {
6802e340	434	bh = gfs2_getbuf(gl, dblock, CREATE);
b3b94faa	435
7276b3b0	436	if (!buffer_uptodate(bh) && !buffer_locked(bh))
20ed0535	437	ll_rw_block(READA \| REQ_META, 1, &bh);
7276b3b0	438	brelse(bh);
b3b94faa DT	439	dblock++;
b3b94faa DT	440	extlen--;
7276b3b0 SW	441	if (!buffer_locked(first_bh) && buffer_uptodate(first_bh))
7276b3b0 SW	442	goto out;
b3b94faa DT	443	}
b3b94faa DT	444
7276b3b0	445	wait_on_buffer(first_bh);
a91ea69f	446	out:
7276b3b0	447	return first_bh;
b3b94faa DT	448	}
b3b94faa DT	449