[deliverable/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_flags = REQ_META | REQ_PRIO |
		(wbc->sync_mode == WB_SYNC_ALL ? WRITE_SYNC : 0);

	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 flusher thread 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(REQ_OP_WRITE, write_flags, 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,
};

const struct address_space_operations gfs2_rgrp_aops = {
	.writepage = gfs2_aspace_writepage,
	.releasepage = gfs2_releasepage,
};

/**
 * 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_name.ln_sbd;
	struct page *page;
	struct buffer_head *bh;
	unsigned int shift;
	unsigned long index;
	unsigned int bufnum;

	if (mapping == NULL)
		mapping = &sdp->sd_aspace;

	shift = PAGE_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_get_page_flags(mapping, index,
						FGP_LOCK|FGP_ACCESSED);
		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);
	put_page(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;
}

static void gfs2_meta_read_endio(struct bio *bio)
{
	struct bio_vec *bvec;
	int i;

	bio_for_each_segment_all(bvec, bio, i) {
		struct page *page = bvec->bv_page;
		struct buffer_head *bh = page_buffers(page);
		unsigned int len = bvec->bv_len;

		while (bh_offset(bh) < bvec->bv_offset)
			bh = bh->b_this_page;
		do {
			struct buffer_head *next = bh->b_this_page;
			len -= bh->b_size;
			bh->b_end_io(bh, !bio->bi_error);
			bh = next;
		} while (bh && len);
	}
	bio_put(bio);
}

/*
 * Submit several consecutive buffer head I/O requests as a single bio I/O
 * request.  (See submit_bh_wbc.)
 */
static void gfs2_submit_bhs(int rw, struct buffer_head *bhs[], int num)
{
	struct buffer_head *bh = bhs[0];
	struct bio *bio;
	int i;

	if (!num)
		return;

	bio = bio_alloc(GFP_NOIO, num);
	bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9);
	bio->bi_bdev = bh->b_bdev;
	for (i = 0; i < num; i++) {
		bh = bhs[i];
		bio_add_page(bio, bh->b_page, bh->b_size, bh_offset(bh));
	}
	bio->bi_end_io = gfs2_meta_read_endio;
	bio->bi_rw = rw;
	submit_bio(bio);
}

/**
 * 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,
		   int rahead, struct buffer_head **bhp)
{
	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
	struct buffer_head *bh, *bhs[2];
	int num = 0;

	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) {
		*bhp = NULL;
		return -EIO;
	}

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

	lock_buffer(bh);
	if (buffer_uptodate(bh)) {
		unlock_buffer(bh);
		flags &= ~DIO_WAIT;
	} else {
		bh->b_end_io = end_buffer_read_sync;
		get_bh(bh);
		bhs[num++] = bh;
	}

	if (rahead) {
		bh = gfs2_getbuf(gl, blkno + 1, CREATE);

		lock_buffer(bh);
		if (buffer_uptodate(bh)) {
			unlock_buffer(bh);
			brelse(bh);
		} else {
			bh->b_end_io = end_buffer_read_sync;
			bhs[num++] = bh;
		}
	}

	gfs2_submit_bhs(READ_SYNC | REQ_META | REQ_PRIO, bhs, num);
	if (!(flags & DIO_WAIT))
		return 0;

	bh = *bhp;
	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);
		*bhp = NULL;
		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;
}

void gfs2_remove_from_journal(struct buffer_head *bh, int meta)
{
	struct address_space *mapping = bh->b_page->mapping;
	struct gfs2_sbd *sdp = gfs2_mapping2sbd(mapping);
	struct gfs2_bufdata *bd = bh->b_private;
	struct gfs2_trans *tr = current->journal_info;
	int was_pinned = 0;

	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 == REMOVE_META)
			tr->tr_num_buf_rm++;
		else
			tr->tr_num_databuf_rm++;
		tr->tr_touched = 1;
		was_pinned = 1;
		brelse(bh);
	}
	if (bd) {
		spin_lock(&sdp->sd_ail_lock);
		if (bd->bd_tr) {
			gfs2_trans_add_revoke(sdp, bd);
		} else if (was_pinned) {
			bh->b_private = NULL;
			kmem_cache_free(gfs2_bufdata_cachep, 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, REMOVE_META);
			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;
	int rahead = 0;

	if (num == ip->i_no_addr)
		rahead = ip->i_rahead;

	ret = gfs2_meta_read(gl, num, DIO_WAIT, rahead, &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_name.ln_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(REQ_OP_READ, 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(REQ_OP_READ, 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;
2a222ca9 MC	40	int write_flags = REQ_META \| REQ_PRIO \|
2a222ca9 MC	41	(wbc->sync_mode == WB_SYNC_ALL ? WRITE_SYNC : 0);
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
e76e0ec9	55	* potentially cause a busy-wait loop from flusher thread and kswapd
4a0f9a32 SW	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)) {
2a222ca9	82	submit_bh(REQ_OP_WRITE, write_flags, bh);
4a0f9a32 SW	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
1b2ad412 SW	100	const struct address_space_operations gfs2_rgrp_aops = {
	101	.writepage = gfs2_aspace_writepage,
	102	.releasepage = gfs2_releasepage,
	103	};
	104
b3b94faa	105	/**
6802e340	106	* gfs2_getbuf - Get a buffer with a given address space
cb4c0313	107	* @gl: the glock
b3b94faa DT	108	* @blkno: the block number (filesystem scope)
	109	* @create: 1 if the buffer should be created
	110	*
	111	* Returns: the buffer
	112	*/
	113
6802e340	114	struct buffer_head gfs2_getbuf(struct gfs2_glock gl, u64 blkno, int create)
b3b94faa	115	{
009d8518	116	struct address_space *mapping = gfs2_glock2aspace(gl);
15562c43	117	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
b3b94faa DT	118	struct page *page;
	119	struct buffer_head *bh;
	120	unsigned int shift;
	121	unsigned long index;
	122	unsigned int bufnum;
	123
70d4ee94 SW	124	if (mapping == NULL)
	125	mapping = &sdp->sd_aspace;
	126
09cbfeaf	127	shift = PAGE_SHIFT - sdp->sd_sb.sb_bsize_shift;
b3b94faa DT	128	index = blkno >> shift; /* convert block to page */
	129	bufnum = blkno - (index << shift); /* block buf index within page */
	130
	131	if (create) {
	132	for (;;) {
cb4c0313	133	page = grab_cache_page(mapping, index);
b3b94faa DT	134	if (page)
	135	break;
	136	yield();
	137	}
	138	} else {
2457aec6 MG	139	page = find_get_page_flags(mapping, index,
2457aec6 MG	140	FGP_LOCK\|FGP_ACCESSED);
b3b94faa DT	141	if (!page)
	142	return NULL;
	143	}
	144
	145	if (!page_has_buffers(page))
	146	create_empty_buffers(page, sdp->sd_sb.sb_bsize, 0);
	147
	148	/* Locate header for our buffer within our page */
	149	for (bh = page_buffers(page); bufnum--; bh = bh->b_this_page)
	150	/* Do nothing */;
	151	get_bh(bh);
	152
	153	if (!buffer_mapped(bh))
	154	map_bh(bh, sdp->sd_vfs, blkno);
	155
	156	unlock_page(page);
09cbfeaf	157	put_page(page);
b3b94faa DT	158
	159	return bh;
	160	}
	161
	162	static void meta_prep_new(struct buffer_head *bh)
	163	{
	164	struct gfs2_meta_header mh = (struct gfs2_meta_header )bh->b_data;
	165
	166	lock_buffer(bh);
	167	clear_buffer_dirty(bh);
	168	set_buffer_uptodate(bh);
	169	unlock_buffer(bh);
	170
	171	mh->mh_magic = cpu_to_be32(GFS2_MAGIC);
	172	}
	173
	174	/**
	175	* gfs2_meta_new - Get a block
	176	* @gl: The glock associated with this block
	177	* @blkno: The block number
	178	*
	179	* Returns: The buffer
	180	*/
	181
cd915493	182	struct buffer_head gfs2_meta_new(struct gfs2_glock gl, u64 blkno)
b3b94faa DT	183	{
b3b94faa DT	184	struct buffer_head *bh;
6802e340	185	bh = gfs2_getbuf(gl, blkno, CREATE);
b3b94faa DT	186	meta_prep_new(bh);
	187	return bh;
	188	}
	189
39b0555f	190	static void gfs2_meta_read_endio(struct bio *bio)
c8d57703	191	{
39b0555f AG	192	struct bio_vec *bvec;
	193	int i;
	194
	195	bio_for_each_segment_all(bvec, bio, i) {
	196	struct page *page = bvec->bv_page;
	197	struct buffer_head *bh = page_buffers(page);
	198	unsigned int len = bvec->bv_len;
	199
	200	while (bh_offset(bh) < bvec->bv_offset)
	201	bh = bh->b_this_page;
	202	do {
	203	struct buffer_head *next = bh->b_this_page;
	204	len -= bh->b_size;
	205	bh->b_end_io(bh, !bio->bi_error);
	206	bh = next;
	207	} while (bh && len);
	208	}
	209	bio_put(bio);
	210	}
c8d57703	211
39b0555f AG	212	/*
	213	* Submit several consecutive buffer head I/O requests as a single bio I/O
	214	* request. (See submit_bh_wbc.)
	215	*/
	216	static void gfs2_submit_bhs(int rw, struct buffer_head *bhs[], int num)
	217	{
	218	struct buffer_head *bh = bhs[0];
	219	struct bio *bio;
	220	int i;
	221
	222	if (!num)
c8d57703	223	return;
39b0555f AG	224
	225	bio = bio_alloc(GFP_NOIO, num);
	226	bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9);
	227	bio->bi_bdev = bh->b_bdev;
	228	for (i = 0; i < num; i++) {
	229	bh = bhs[i];
	230	bio_add_page(bio, bh->b_page, bh->b_size, bh_offset(bh));
c8d57703	231	}
39b0555f	232	bio->bi_end_io = gfs2_meta_read_endio;
4e49ea4a MC	233	bio->bi_rw = rw;
4e49ea4a MC	234	submit_bio(bio);
c8d57703 AG	235	}
c8d57703 AG	236
b3b94faa DT	237	/**
	238	* gfs2_meta_read - Read a block from disk
	239	* @gl: The glock covering the block
	240	* @blkno: The block number
7276b3b0	241	* @flags: flags
b3b94faa DT	242	* @bhp: the place where the buffer is returned (NULL on failure)
	243	*
	244	* Returns: errno
	245	*/
	246
cd915493	247	int gfs2_meta_read(struct gfs2_glock *gl, u64 blkno, int flags,
c8d57703	248	int rahead, struct buffer_head **bhp)
b3b94faa	249	{
15562c43	250	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
39b0555f AG	251	struct buffer_head bh, bhs[2];
39b0555f AG	252	int num = 0;
c969f58c	253
44b8db13 MY	254	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) {
44b8db13 MY	255	*bhp = NULL;
c969f58c	256	return -EIO;
44b8db13	257	}
c969f58c SW	258
	259	*bhp = bh = gfs2_getbuf(gl, blkno, CREATE);
	260
	261	lock_buffer(bh);
	262	if (buffer_uptodate(bh)) {
	263	unlock_buffer(bh);
39b0555f AG	264	flags &= ~DIO_WAIT;
	265	} else {
	266	bh->b_end_io = end_buffer_read_sync;
	267	get_bh(bh);
	268	bhs[num++] = bh;
c969f58c	269	}
39b0555f AG	270
	271	if (rahead) {
	272	bh = gfs2_getbuf(gl, blkno + 1, CREATE);
	273
	274	lock_buffer(bh);
	275	if (buffer_uptodate(bh)) {
	276	unlock_buffer(bh);
	277	brelse(bh);
	278	} else {
	279	bh->b_end_io = end_buffer_read_sync;
	280	bhs[num++] = bh;
	281	}
	282	}
	283
	284	gfs2_submit_bhs(READ_SYNC \| REQ_META \| REQ_PRIO, bhs, num);
c969f58c SW	285	if (!(flags & DIO_WAIT))
	286	return 0;
	287
39b0555f	288	bh = *bhp;
c969f58c SW	289	wait_on_buffer(bh);
	290	if (unlikely(!buffer_uptodate(bh))) {
	291	struct gfs2_trans *tr = current->journal_info;
	292	if (tr && tr->tr_touched)
	293	gfs2_io_error_bh(sdp, bh);
	294	brelse(bh);
44b8db13	295	*bhp = NULL;
c969f58c	296	return -EIO;
7276b3b0	297	}
b3b94faa	298
7276b3b0	299	return 0;
b3b94faa DT	300	}
	301
	302	/**
7276b3b0	303	* gfs2_meta_wait - Reread a block from disk
b3b94faa	304	* @sdp: the filesystem
7276b3b0	305	* @bh: The block to wait for
b3b94faa DT	306	*
	307	* Returns: errno
	308	*/
	309
7276b3b0	310	int gfs2_meta_wait(struct gfs2_sbd sdp, struct buffer_head bh)
b3b94faa DT	311	{
	312	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
	313	return -EIO;
	314
7276b3b0	315	wait_on_buffer(bh);
b3b94faa	316
7276b3b0 SW	317	if (!buffer_uptodate(bh)) {
	318	struct gfs2_trans *tr = current->journal_info;
	319	if (tr && tr->tr_touched)
	320	gfs2_io_error_bh(sdp, bh);
	321	return -EIO;
b3b94faa	322	}
7276b3b0 SW	323	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
7276b3b0 SW	324	return -EIO;
b3b94faa DT	325
	326	return 0;
	327	}
	328
68cd4ce2	329	void gfs2_remove_from_journal(struct buffer_head *bh, int meta)
16615be1	330	{
009d8518 SW	331	struct address_space *mapping = bh->b_page->mapping;
009d8518 SW	332	struct gfs2_sbd *sdp = gfs2_mapping2sbd(mapping);
16615be1	333	struct gfs2_bufdata *bd = bh->b_private;
68cd4ce2	334	struct gfs2_trans *tr = current->journal_info;
502be2a3	335	int was_pinned = 0;
009d8518	336
16615be1	337	if (test_clear_buffer_pinned(bh)) {
627c10b7	338	trace_gfs2_pin(bd, 0);
5e687eac	339	atomic_dec(&sdp->sd_log_pinned);
c0752aa7	340	list_del_init(&bd->bd_list);
68cd4ce2	341	if (meta == REMOVE_META)
16615be1	342	tr->tr_num_buf_rm++;
022ef4fe	343	else
16615be1	344	tr->tr_num_databuf_rm++;
16615be1	345	tr->tr_touched = 1;
502be2a3	346	was_pinned = 1;
16615be1 SW	347	brelse(bh);
	348	}
	349	if (bd) {
c618e87a	350	spin_lock(&sdp->sd_ail_lock);
16ca9412	351	if (bd->bd_tr) {
16615be1	352	gfs2_trans_add_revoke(sdp, bd);
502be2a3 BP	353	} else if (was_pinned) {
	354	bh->b_private = NULL;
	355	kmem_cache_free(gfs2_bufdata_cachep, bd);
16615be1	356	}
c618e87a	357	spin_unlock(&sdp->sd_ail_lock);
16615be1 SW	358	}
	359	clear_buffer_dirty(bh);
	360	clear_buffer_uptodate(bh);
	361	}
	362
b3b94faa DT	363	/**
	364	* gfs2_meta_wipe - make inode's buffers so they aren't dirty/pinned anymore
	365	* @ip: the inode who owns the buffers
	366	* @bstart: the first buffer in the run
	367	* @blen: the number of buffers in the run
	368	*
	369	*/
	370
cd915493	371	void gfs2_meta_wipe(struct gfs2_inode *ip, u64 bstart, u32 blen)
b3b94faa	372	{
feaa7bba	373	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
b3b94faa DT	374	struct buffer_head *bh;
	375
	376	while (blen) {
6802e340	377	bh = gfs2_getbuf(ip->i_gl, bstart, NO_CREATE);
b3b94faa	378	if (bh) {
1ad38c43 SW	379	lock_buffer(bh);
1ad38c43 SW	380	gfs2_log_lock(sdp);
68cd4ce2	381	gfs2_remove_from_journal(bh, REMOVE_META);
1ad38c43	382	gfs2_log_unlock(sdp);
b3b94faa	383	unlock_buffer(bh);
b3b94faa DT	384	brelse(bh);
	385	}
	386
	387	bstart++;
	388	blen--;
	389	}
	390	}
	391
b3b94faa DT	392	/**
	393	* gfs2_meta_indirect_buffer - Get a metadata buffer
	394	* @ip: The GFS2 inode
	395	* @height: The level of this buf in the metadata (indir addr) tree (if any)
	396	* @num: The block number (device relative) of the buffer
b3b94faa DT	397	* @bhp: the buffer is returned here
b3b94faa DT	398	*
b3b94faa DT	399	* Returns: errno
	400	*/
	401
cd915493	402	int gfs2_meta_indirect_buffer(struct gfs2_inode *ip, int height, u64 num,
f2f9c812	403	struct buffer_head **bhp)
b3b94faa	404	{
7276b3b0 SW	405	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
7276b3b0 SW	406	struct gfs2_glock *gl = ip->i_gl;
f91a0d3e SW	407	struct buffer_head *bh;
f91a0d3e SW	408	int ret = 0;
f2f9c812	409	u32 mtype = height ? GFS2_METATYPE_IN : GFS2_METATYPE_DI;
c8d57703 AG	410	int rahead = 0;
	411
	412	if (num == ip->i_no_addr)
	413	rahead = ip->i_rahead;
b3b94faa	414
c8d57703	415	ret = gfs2_meta_read(gl, num, DIO_WAIT, rahead, &bh);
f2f9c812 BP	416	if (ret == 0 && gfs2_metatype_check(sdp, bh, mtype)) {
	417	brelse(bh);
	418	ret = -EIO;
b3b94faa	419	}
b3b94faa	420	*bhp = bh;
f91a0d3e	421	return ret;
b3b94faa DT	422	}
	423
	424	/**
	425	* gfs2_meta_ra - start readahead on an extent of a file
	426	* @gl: the glock the blocks belong to
	427	* @dblock: the starting disk block
	428	* @extlen: the number of blocks in the extent
	429	*
7276b3b0	430	* returns: the first buffer in the extent
b3b94faa DT	431	*/
b3b94faa DT	432
7276b3b0	433	struct buffer_head gfs2_meta_ra(struct gfs2_glock gl, u64 dblock, u32 extlen)
b3b94faa	434	{
15562c43	435	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
b3b94faa	436	struct buffer_head first_bh, bh;
cd915493	437	u32 max_ra = gfs2_tune_get(sdp, gt_max_readahead) >>
568f4c96	438	sdp->sd_sb.sb_bsize_shift;
b3b94faa	439
7276b3b0 SW	440	BUG_ON(!extlen);
	441
	442	if (max_ra < 1)
	443	max_ra = 1;
b3b94faa DT	444	if (extlen > max_ra)
	445	extlen = max_ra;
	446
6802e340	447	first_bh = gfs2_getbuf(gl, dblock, CREATE);
b3b94faa DT	448
	449	if (buffer_uptodate(first_bh))
	450	goto out;
7276b3b0	451	if (!buffer_locked(first_bh))
dfec8a14	452	ll_rw_block(REQ_OP_READ, READ_SYNC \| REQ_META, 1, &first_bh);
b3b94faa DT	453
	454	dblock++;
	455	extlen--;
	456
	457	while (extlen) {
6802e340	458	bh = gfs2_getbuf(gl, dblock, CREATE);
b3b94faa	459
7276b3b0	460	if (!buffer_uptodate(bh) && !buffer_locked(bh))
dfec8a14	461	ll_rw_block(REQ_OP_READ, READA \| REQ_META, 1, &bh);
7276b3b0	462	brelse(bh);
b3b94faa DT	463	dblock++;
b3b94faa DT	464	extlen--;
7276b3b0 SW	465	if (!buffer_locked(first_bh) && buffer_uptodate(first_bh))
7276b3b0 SW	466	goto out;
b3b94faa DT	467	}
b3b94faa DT	468
7276b3b0	469	wait_on_buffer(first_bh);
a91ea69f	470	out:
7276b3b0	471	return first_bh;
b3b94faa DT	472	}
b3b94faa DT	473