[linux-2.6-block.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_to_write_flags(wbc);

	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;
	struct bvec_iter_all iter_all;

	bio_for_each_segment_all(bvec, bio, i, iter_all) {
		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_status);
			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 op, int op_flags, struct buffer_head *bhs[],
			    int num)
{
	while (num > 0) {
		struct buffer_head *bh = *bhs;
		struct bio *bio;

		bio = bio_alloc(GFP_NOIO, num);
		bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9);
		bio_set_dev(bio, bh->b_bdev);
		while (num > 0) {
			bh = *bhs;
			if (!bio_add_page(bio, bh->b_page, bh->b_size, bh_offset(bh))) {
				BUG_ON(bio->bi_iter.bi_size == 0);
				break;
			}
			bhs++;
			num--;
		}
		bio->bi_end_io = gfs2_meta_read_endio;
		bio_set_op_attrs(bio, op, op_flags);
		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(REQ_OP_READ, 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 && test_bit(TR_TOUCHED, &tr->tr_flags))
			gfs2_io_error_bh_wd(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 && test_bit(TR_TOUCHED, &tr->tr_flags))
			gfs2_io_error_bh_wd(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++;
		set_bit(TR_TOUCHED, &tr->tr_flags);
		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;
	} else {
		*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, REQ_META | REQ_PRIO, 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,
				    REQ_RAHEAD | REQ_META | REQ_PRIO,
				    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;
7637241e	40	int write_flags = REQ_META \| REQ_PRIO \| wbc_to_write_flags(wbc);
b3b94faa	41
4a0f9a32 SW	42	BUG_ON(!PageLocked(page));
	43	BUG_ON(!page_has_buffers(page));
	44
	45	head = page_buffers(page);
	46	bh = head;
	47
	48	do {
	49	if (!buffer_mapped(bh))
	50	continue;
	51	/*
	52	* If it's a fully non-blocking write attempt and we cannot
	53	* lock the buffer then redirty the page. Note that this can
e76e0ec9	54	* potentially cause a busy-wait loop from flusher thread and kswapd
4a0f9a32 SW	55	* activity, but those code paths have their own higher-level
	56	* throttling.
	57	*/
1b430bee	58	if (wbc->sync_mode != WB_SYNC_NONE) {
4a0f9a32 SW	59	lock_buffer(bh);
	60	} else if (!trylock_buffer(bh)) {
	61	redirty_page_for_writepage(wbc, page);
	62	continue;
	63	}
	64	if (test_clear_buffer_dirty(bh)) {
	65	mark_buffer_async_write(bh);
	66	} else {
	67	unlock_buffer(bh);
	68	}
	69	} while ((bh = bh->b_this_page) != head);
	70
	71	/*
	72	* The page and its buffers are protected by PageWriteback(), so we can
	73	* drop the bh refcounts early.
	74	*/
	75	BUG_ON(PageWriteback(page));
	76	set_page_writeback(page);
	77
	78	do {
	79	struct buffer_head *next = bh->b_this_page;
	80	if (buffer_async_write(bh)) {
2a222ca9	81	submit_bh(REQ_OP_WRITE, write_flags, bh);
4a0f9a32 SW	82	nr_underway++;
	83	}
	84	bh = next;
	85	} while (bh != head);
	86	unlock_page(page);
	87
4a0f9a32 SW	88	if (nr_underway == 0)
	89	end_page_writeback(page);
	90
eaefbf96	91	return 0;
b3b94faa DT	92	}
b3b94faa DT	93
009d8518	94	const struct address_space_operations gfs2_meta_aops = {
b3b94faa	95	.writepage = gfs2_aspace_writepage,
4340fe62	96	.releasepage = gfs2_releasepage,
b3b94faa DT	97	};
b3b94faa DT	98
1b2ad412 SW	99	const struct address_space_operations gfs2_rgrp_aops = {
	100	.writepage = gfs2_aspace_writepage,
	101	.releasepage = gfs2_releasepage,
	102	};
	103
b3b94faa	104	/**
6802e340	105	* gfs2_getbuf - Get a buffer with a given address space
cb4c0313	106	* @gl: the glock
b3b94faa DT	107	* @blkno: the block number (filesystem scope)
	108	* @create: 1 if the buffer should be created
	109	*
	110	* Returns: the buffer
	111	*/
	112
6802e340	113	struct buffer_head gfs2_getbuf(struct gfs2_glock gl, u64 blkno, int create)
b3b94faa	114	{
009d8518	115	struct address_space *mapping = gfs2_glock2aspace(gl);
15562c43	116	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
b3b94faa DT	117	struct page *page;
	118	struct buffer_head *bh;
	119	unsigned int shift;
	120	unsigned long index;
	121	unsigned int bufnum;
	122
70d4ee94 SW	123	if (mapping == NULL)
	124	mapping = &sdp->sd_aspace;
	125
09cbfeaf	126	shift = PAGE_SHIFT - sdp->sd_sb.sb_bsize_shift;
b3b94faa DT	127	index = blkno >> shift; /* convert block to page */
	128	bufnum = blkno - (index << shift); /* block buf index within page */
	129
	130	if (create) {
	131	for (;;) {
cb4c0313	132	page = grab_cache_page(mapping, index);
b3b94faa DT	133	if (page)
	134	break;
	135	yield();
	136	}
	137	} else {
2457aec6 MG	138	page = find_get_page_flags(mapping, index,
2457aec6 MG	139	FGP_LOCK\|FGP_ACCESSED);
b3b94faa DT	140	if (!page)
	141	return NULL;
	142	}
	143
	144	if (!page_has_buffers(page))
	145	create_empty_buffers(page, sdp->sd_sb.sb_bsize, 0);
	146
	147	/* Locate header for our buffer within our page */
	148	for (bh = page_buffers(page); bufnum--; bh = bh->b_this_page)
	149	/* Do nothing */;
	150	get_bh(bh);
	151
	152	if (!buffer_mapped(bh))
	153	map_bh(bh, sdp->sd_vfs, blkno);
	154
	155	unlock_page(page);
09cbfeaf	156	put_page(page);
b3b94faa DT	157
	158	return bh;
	159	}
	160
	161	static void meta_prep_new(struct buffer_head *bh)
	162	{
	163	struct gfs2_meta_header mh = (struct gfs2_meta_header )bh->b_data;
	164
	165	lock_buffer(bh);
	166	clear_buffer_dirty(bh);
	167	set_buffer_uptodate(bh);
	168	unlock_buffer(bh);
	169
	170	mh->mh_magic = cpu_to_be32(GFS2_MAGIC);
	171	}
	172
	173	/**
	174	* gfs2_meta_new - Get a block
	175	* @gl: The glock associated with this block
	176	* @blkno: The block number
	177	*
	178	* Returns: The buffer
	179	*/
	180
cd915493	181	struct buffer_head gfs2_meta_new(struct gfs2_glock gl, u64 blkno)
b3b94faa DT	182	{
b3b94faa DT	183	struct buffer_head *bh;
6802e340	184	bh = gfs2_getbuf(gl, blkno, CREATE);
b3b94faa DT	185	meta_prep_new(bh);
	186	return bh;
	187	}
	188
39b0555f	189	static void gfs2_meta_read_endio(struct bio *bio)
c8d57703	190	{
39b0555f AG	191	struct bio_vec *bvec;
39b0555f AG	192	int i;
6dc4f100	193	struct bvec_iter_all iter_all;
39b0555f	194
6dc4f100	195	bio_for_each_segment_all(bvec, bio, i, iter_all) {
39b0555f AG	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;
4e4cbee9	205	bh->b_end_io(bh, !bio->bi_status);
39b0555f AG	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	*/
e1b1afa6 MC	216	static void gfs2_submit_bhs(int op, int op_flags, struct buffer_head *bhs[],
e1b1afa6 MC	217	int num)
39b0555f	218	{
23e5671a AG	219	while (num > 0) {
	220	struct buffer_head bh = bhs;
	221	struct bio *bio;
	222
	223	bio = bio_alloc(GFP_NOIO, num);
	224	bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9);
74d46992	225	bio_set_dev(bio, bh->b_bdev);
23e5671a AG	226	while (num > 0) {
	227	bh = *bhs;
	228	if (!bio_add_page(bio, bh->b_page, bh->b_size, bh_offset(bh))) {
	229	BUG_ON(bio->bi_iter.bi_size == 0);
	230	break;
	231	}
	232	bhs++;
	233	num--;
	234	}
	235	bio->bi_end_io = gfs2_meta_read_endio;
	236	bio_set_op_attrs(bio, op, op_flags);
	237	submit_bio(bio);
c8d57703	238	}
c8d57703 AG	239	}
c8d57703 AG	240
b3b94faa DT	241	/**
	242	* gfs2_meta_read - Read a block from disk
	243	* @gl: The glock covering the block
	244	* @blkno: The block number
7276b3b0	245	* @flags: flags
b3b94faa DT	246	* @bhp: the place where the buffer is returned (NULL on failure)
	247	*
	248	* Returns: errno
	249	*/
	250
cd915493	251	int gfs2_meta_read(struct gfs2_glock *gl, u64 blkno, int flags,
c8d57703	252	int rahead, struct buffer_head **bhp)
b3b94faa	253	{
15562c43	254	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
39b0555f AG	255	struct buffer_head bh, bhs[2];
39b0555f AG	256	int num = 0;
c969f58c	257
44b8db13 MY	258	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) {
44b8db13 MY	259	*bhp = NULL;
c969f58c	260	return -EIO;
44b8db13	261	}
c969f58c SW	262
	263	*bhp = bh = gfs2_getbuf(gl, blkno, CREATE);
	264
	265	lock_buffer(bh);
	266	if (buffer_uptodate(bh)) {
	267	unlock_buffer(bh);
39b0555f AG	268	flags &= ~DIO_WAIT;
	269	} else {
	270	bh->b_end_io = end_buffer_read_sync;
	271	get_bh(bh);
	272	bhs[num++] = bh;
c969f58c	273	}
39b0555f AG	274
	275	if (rahead) {
	276	bh = gfs2_getbuf(gl, blkno + 1, CREATE);
	277
	278	lock_buffer(bh);
	279	if (buffer_uptodate(bh)) {
	280	unlock_buffer(bh);
	281	brelse(bh);
	282	} else {
	283	bh->b_end_io = end_buffer_read_sync;
	284	bhs[num++] = bh;
	285	}
	286	}
	287
70fd7614	288	gfs2_submit_bhs(REQ_OP_READ, REQ_META \| REQ_PRIO, bhs, num);
c969f58c SW	289	if (!(flags & DIO_WAIT))
	290	return 0;
	291
39b0555f	292	bh = *bhp;
c969f58c SW	293	wait_on_buffer(bh);
	294	if (unlikely(!buffer_uptodate(bh))) {
	295	struct gfs2_trans *tr = current->journal_info;
9862ca05	296	if (tr && test_bit(TR_TOUCHED, &tr->tr_flags))
9e1a9ecd	297	gfs2_io_error_bh_wd(sdp, bh);
c969f58c	298	brelse(bh);
44b8db13	299	*bhp = NULL;
c969f58c	300	return -EIO;
7276b3b0	301	}
b3b94faa	302
7276b3b0	303	return 0;
b3b94faa DT	304	}
	305
	306	/**
7276b3b0	307	* gfs2_meta_wait - Reread a block from disk
b3b94faa	308	* @sdp: the filesystem
7276b3b0	309	* @bh: The block to wait for
b3b94faa DT	310	*
	311	* Returns: errno
	312	*/
	313
7276b3b0	314	int gfs2_meta_wait(struct gfs2_sbd sdp, struct buffer_head bh)
b3b94faa DT	315	{
	316	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
	317	return -EIO;
	318
7276b3b0	319	wait_on_buffer(bh);
b3b94faa	320
7276b3b0 SW	321	if (!buffer_uptodate(bh)) {
7276b3b0 SW	322	struct gfs2_trans *tr = current->journal_info;
9862ca05	323	if (tr && test_bit(TR_TOUCHED, &tr->tr_flags))
9e1a9ecd	324	gfs2_io_error_bh_wd(sdp, bh);
7276b3b0	325	return -EIO;
b3b94faa	326	}
7276b3b0 SW	327	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
7276b3b0 SW	328	return -EIO;
b3b94faa DT	329
	330	return 0;
	331	}
	332
68cd4ce2	333	void gfs2_remove_from_journal(struct buffer_head *bh, int meta)
16615be1	334	{
009d8518 SW	335	struct address_space *mapping = bh->b_page->mapping;
009d8518 SW	336	struct gfs2_sbd *sdp = gfs2_mapping2sbd(mapping);
16615be1	337	struct gfs2_bufdata *bd = bh->b_private;
68cd4ce2	338	struct gfs2_trans *tr = current->journal_info;
502be2a3	339	int was_pinned = 0;
009d8518	340
16615be1	341	if (test_clear_buffer_pinned(bh)) {
627c10b7	342	trace_gfs2_pin(bd, 0);
5e687eac	343	atomic_dec(&sdp->sd_log_pinned);
c0752aa7	344	list_del_init(&bd->bd_list);
68cd4ce2	345	if (meta == REMOVE_META)
16615be1	346	tr->tr_num_buf_rm++;
022ef4fe	347	else
16615be1	348	tr->tr_num_databuf_rm++;
9862ca05	349	set_bit(TR_TOUCHED, &tr->tr_flags);
502be2a3	350	was_pinned = 1;
16615be1 SW	351	brelse(bh);
	352	}
	353	if (bd) {
c618e87a	354	spin_lock(&sdp->sd_ail_lock);
16ca9412	355	if (bd->bd_tr) {
16615be1	356	gfs2_trans_add_revoke(sdp, bd);
502be2a3 BP	357	} else if (was_pinned) {
	358	bh->b_private = NULL;
	359	kmem_cache_free(gfs2_bufdata_cachep, bd);
16615be1	360	}
c618e87a	361	spin_unlock(&sdp->sd_ail_lock);
16615be1 SW	362	}
	363	clear_buffer_dirty(bh);
	364	clear_buffer_uptodate(bh);
	365	}
	366
b3b94faa DT	367	/**
	368	* gfs2_meta_wipe - make inode's buffers so they aren't dirty/pinned anymore
	369	* @ip: the inode who owns the buffers
	370	* @bstart: the first buffer in the run
	371	* @blen: the number of buffers in the run
	372	*
	373	*/
	374
cd915493	375	void gfs2_meta_wipe(struct gfs2_inode *ip, u64 bstart, u32 blen)
b3b94faa	376	{
feaa7bba	377	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
b3b94faa DT	378	struct buffer_head *bh;
	379
	380	while (blen) {
6802e340	381	bh = gfs2_getbuf(ip->i_gl, bstart, NO_CREATE);
b3b94faa	382	if (bh) {
1ad38c43 SW	383	lock_buffer(bh);
1ad38c43 SW	384	gfs2_log_lock(sdp);
68cd4ce2	385	gfs2_remove_from_journal(bh, REMOVE_META);
1ad38c43	386	gfs2_log_unlock(sdp);
b3b94faa	387	unlock_buffer(bh);
b3b94faa DT	388	brelse(bh);
	389	}
	390
	391	bstart++;
	392	blen--;
	393	}
	394	}
	395
b3b94faa DT	396	/**
	397	* gfs2_meta_indirect_buffer - Get a metadata buffer
	398	* @ip: The GFS2 inode
	399	* @height: The level of this buf in the metadata (indir addr) tree (if any)
	400	* @num: The block number (device relative) of the buffer
b3b94faa DT	401	* @bhp: the buffer is returned here
b3b94faa DT	402	*
b3b94faa DT	403	* Returns: errno
	404	*/
	405
cd915493	406	int gfs2_meta_indirect_buffer(struct gfs2_inode *ip, int height, u64 num,
f2f9c812	407	struct buffer_head **bhp)
b3b94faa	408	{
7276b3b0 SW	409	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
7276b3b0 SW	410	struct gfs2_glock *gl = ip->i_gl;
f91a0d3e SW	411	struct buffer_head *bh;
f91a0d3e SW	412	int ret = 0;
f2f9c812	413	u32 mtype = height ? GFS2_METATYPE_IN : GFS2_METATYPE_DI;
c8d57703 AG	414	int rahead = 0;
	415
	416	if (num == ip->i_no_addr)
	417	rahead = ip->i_rahead;
b3b94faa	418
c8d57703	419	ret = gfs2_meta_read(gl, num, DIO_WAIT, rahead, &bh);
f2f9c812 BP	420	if (ret == 0 && gfs2_metatype_check(sdp, bh, mtype)) {
	421	brelse(bh);
	422	ret = -EIO;
61eaadcd BP	423	} else {
61eaadcd BP	424	*bhp = bh;
b3b94faa	425	}
f91a0d3e	426	return ret;
b3b94faa DT	427	}
	428
	429	/**
	430	* gfs2_meta_ra - start readahead on an extent of a file
	431	* @gl: the glock the blocks belong to
	432	* @dblock: the starting disk block
	433	* @extlen: the number of blocks in the extent
	434	*
7276b3b0	435	* returns: the first buffer in the extent
b3b94faa DT	436	*/
b3b94faa DT	437
7276b3b0	438	struct buffer_head gfs2_meta_ra(struct gfs2_glock gl, u64 dblock, u32 extlen)
b3b94faa	439	{
15562c43	440	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
b3b94faa	441	struct buffer_head first_bh, bh;
cd915493	442	u32 max_ra = gfs2_tune_get(sdp, gt_max_readahead) >>
568f4c96	443	sdp->sd_sb.sb_bsize_shift;
b3b94faa	444
7276b3b0 SW	445	BUG_ON(!extlen);
	446
	447	if (max_ra < 1)
	448	max_ra = 1;
b3b94faa DT	449	if (extlen > max_ra)
	450	extlen = max_ra;
	451
6802e340	452	first_bh = gfs2_getbuf(gl, dblock, CREATE);
b3b94faa DT	453
	454	if (buffer_uptodate(first_bh))
	455	goto out;
7276b3b0	456	if (!buffer_locked(first_bh))
e477b24b	457	ll_rw_block(REQ_OP_READ, REQ_META \| REQ_PRIO, 1, &first_bh);
b3b94faa DT	458
	459	dblock++;
	460	extlen--;
	461
	462	while (extlen) {
6802e340	463	bh = gfs2_getbuf(gl, dblock, CREATE);
b3b94faa	464
7276b3b0	465	if (!buffer_uptodate(bh) && !buffer_locked(bh))
e477b24b CL	466	ll_rw_block(REQ_OP_READ,
	467	REQ_RAHEAD \| REQ_META \| REQ_PRIO,
	468	1, &bh);
7276b3b0	469	brelse(bh);
b3b94faa DT	470	dblock++;
b3b94faa DT	471	extlen--;
7276b3b0 SW	472	if (!buffer_locked(first_bh) && buffer_uptodate(first_bh))
7276b3b0 SW	473	goto out;
b3b94faa DT	474	}
b3b94faa DT	475
7276b3b0	476	wait_on_buffer(first_bh);
a91ea69f	477	out:
7276b3b0	478	return first_bh;
b3b94faa DT	479	}
b3b94faa DT	480