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

// SPDX-License-Identifier: GPL-2.0-only
/*
 * Copyright (C) Sistina Software, Inc.  1997-2003 All rights reserved.
 * Copyright (C) 2004-2008 Red Hat, Inc.  All rights reserved.
 */

#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;
	blk_opf_t 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 = {
	.dirty_folio	= block_dirty_folio,
	.invalidate_folio = block_invalidate_folio,
	.writepage = gfs2_aspace_writepage,
	.release_folio = gfs2_release_folio,
};

const struct address_space_operations gfs2_rgrp_aops = {
	.dirty_folio	= block_dirty_folio,
	.invalidate_folio = block_invalidate_folio,
	.writepage = gfs2_aspace_writepage,
	.release_folio = gfs2_release_folio,
};

/**
 * 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();
		}
		if (!page_has_buffers(page))
			create_empty_buffers(page, sdp->sd_sb.sb_bsize, 0);
	} else {
		page = find_get_page_flags(mapping, index,
						FGP_LOCK|FGP_ACCESSED);
		if (!page)
			return NULL;
		if (!page_has_buffers(page)) {
			bh = NULL;
			goto out_unlock;
		}
	}

	/* 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);

out_unlock:
	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;
	struct bvec_iter_all iter_all;

	bio_for_each_segment_all(bvec, bio, 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(blk_opf_t opf, struct buffer_head *bhs[], int num)
{
	while (num > 0) {
		struct buffer_head *bh = *bhs;
		struct bio *bio;

		bio = bio_alloc(bh->b_bdev, num, opf, GFP_NOIO);
		bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9);
		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;
		submit_bio(bio);
	}
}

/**
 * gfs2_meta_read - Read a block from disk
 * @gl: The glock covering the block
 * @blkno: The block number
 * @flags: flags
 * @rahead: Do read-ahead
 * @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(gfs2_withdrawn(sdp)) && !gfs2_withdraw_in_prog(sdp)) {
		*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(gfs2_withdrawn(sdp)) && !gfs2_withdraw_in_prog(sdp))
		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(gfs2_withdrawn(sdp)) && !gfs2_withdraw_in_prog(sdp))
		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) {
		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);
		} else if (!list_empty(&bd->bd_ail_st_list) &&
					!list_empty(&bd->bd_ail_gl_list)) {
			gfs2_remove_from_ail(bd);
		}
	}
	clear_buffer_dirty(bh);
	clear_buffer_uptodate(bh);
}

/**
 * gfs2_ail1_wipe - remove deleted/freed buffers from the ail1 list
 * @sdp: superblock
 * @bstart: starting block address of buffers to remove
 * @blen: length of buffers to be removed
 *
 * This function is called from gfs2_journal wipe, whose job is to remove
 * buffers, corresponding to deleted blocks, from the journal. If we find any
 * bufdata elements on the system ail1 list, they haven't been written to
 * the journal yet. So we remove them.
 */
static void gfs2_ail1_wipe(struct gfs2_sbd *sdp, u64 bstart, u32 blen)
{
	struct gfs2_trans *tr, *s;
	struct gfs2_bufdata *bd, *bs;
	struct buffer_head *bh;
	u64 end = bstart + blen;

	gfs2_log_lock(sdp);
	spin_lock(&sdp->sd_ail_lock);
	list_for_each_entry_safe(tr, s, &sdp->sd_ail1_list, tr_list) {
		list_for_each_entry_safe(bd, bs, &tr->tr_ail1_list,
					 bd_ail_st_list) {
			bh = bd->bd_bh;
			if (bh->b_blocknr < bstart || bh->b_blocknr >= end)
				continue;

			gfs2_remove_from_journal(bh, REMOVE_JDATA);
		}
	}
	spin_unlock(&sdp->sd_ail_lock);
	gfs2_log_unlock(sdp);
}

static struct buffer_head *gfs2_getjdatabuf(struct gfs2_inode *ip, u64 blkno)
{
	struct address_space *mapping = ip->i_inode.i_mapping;
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
	struct page *page;
	struct buffer_head *bh;
	unsigned int shift = PAGE_SHIFT - sdp->sd_sb.sb_bsize_shift;
	unsigned long index = blkno >> shift; /* convert block to page */
	unsigned int bufnum = blkno - (index << shift);

	page = find_get_page_flags(mapping, index, FGP_LOCK|FGP_ACCESSED);
	if (!page)
		return NULL;
	if (!page_has_buffers(page)) {
		unlock_page(page);
		put_page(page);
		return NULL;
	}
	/* Locate header for our buffer within our page */
	for (bh = page_buffers(page); bufnum--; bh = bh->b_this_page)
		/* Do nothing */;
	get_bh(bh);
	unlock_page(page);
	put_page(page);
	return bh;
}

/**
 * gfs2_journal_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_journal_wipe(struct gfs2_inode *ip, u64 bstart, u32 blen)
{
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
	struct buffer_head *bh;
	int ty;

	gfs2_ail1_wipe(sdp, bstart, blen);
	while (blen) {
		ty = REMOVE_META;
		bh = gfs2_getbuf(ip->i_gl, bstart, NO_CREATE);
		if (!bh && gfs2_is_jdata(ip)) {
			bh = gfs2_getjdatabuf(ip, bstart);
			ty = REMOVE_JDATA;
		}
		if (bh) {
			lock_buffer(bh);
			gfs2_log_lock(sdp);
			spin_lock(&sdp->sd_ail_lock);
			gfs2_remove_from_journal(bh, ty);
			spin_unlock(&sdp->sd_ail_lock);
			gfs2_log_unlock(sdp);
			unlock_buffer(bh);
			brelse(bh);
		}

		bstart++;
		blen--;
	}
}

/**
 * gfs2_meta_buffer - Get a metadata buffer
 * @ip: The GFS2 inode
 * @mtype: The block type (GFS2_METATYPE_*)
 * @num: The block number (device relative) of the buffer
 * @bhp: the buffer is returned here
 *
 * Returns: errno
 */

int gfs2_meta_buffer(struct gfs2_inode *ip, u32 mtype, 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;
	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
7336d0e6	1	// SPDX-License-Identifier: GPL-2.0-only
b3b94faa DT	2	/*
b3b94faa DT	3	* Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved.
091806ed	4	* Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved.
b3b94faa DT	5	*/
	6
	7	#include <linux/sched.h>
	8	#include <linux/slab.h>
	9	#include <linux/spinlock.h>
	10	#include <linux/completion.h>
	11	#include <linux/buffer_head.h>
	12	#include <linux/mm.h>
	13	#include <linux/pagemap.h>
	14	#include <linux/writeback.h>
	15	#include <linux/swap.h>
	16	#include <linux/delay.h>
2e565bb6	17	#include <linux/bio.h>
5c676f6d	18	#include <linux/gfs2_ondisk.h>
b3b94faa DT	19
b3b94faa DT	20	#include "gfs2.h"
5c676f6d	21	#include "incore.h"
b3b94faa DT	22	#include "glock.h"
	23	#include "glops.h"
	24	#include "inode.h"
	25	#include "log.h"
	26	#include "lops.h"
	27	#include "meta_io.h"
	28	#include "rgrp.h"
	29	#include "trans.h"
5c676f6d	30	#include "util.h"
627c10b7	31	#include "trace_gfs2.h"
b3b94faa	32
4a0f9a32	33	static int gfs2_aspace_writepage(struct page page, struct writeback_control wbc)
b3b94faa	34	{
4a0f9a32 SW	35	struct buffer_head bh, head;
4a0f9a32 SW	36	int nr_underway = 0;
67688c08	37	blk_opf_t write_flags = REQ_META \| REQ_PRIO \| wbc_to_write_flags(wbc);
b3b94faa	38
4a0f9a32 SW	39	BUG_ON(!PageLocked(page));
	40	BUG_ON(!page_has_buffers(page));
	41
	42	head = page_buffers(page);
	43	bh = head;
	44
	45	do {
	46	if (!buffer_mapped(bh))
	47	continue;
	48	/*
	49	* If it's a fully non-blocking write attempt and we cannot
	50	* lock the buffer then redirty the page. Note that this can
e76e0ec9	51	* potentially cause a busy-wait loop from flusher thread and kswapd
4a0f9a32 SW	52	* activity, but those code paths have their own higher-level
	53	* throttling.
	54	*/
1b430bee	55	if (wbc->sync_mode != WB_SYNC_NONE) {
4a0f9a32 SW	56	lock_buffer(bh);
	57	} else if (!trylock_buffer(bh)) {
	58	redirty_page_for_writepage(wbc, page);
	59	continue;
	60	}
	61	if (test_clear_buffer_dirty(bh)) {
	62	mark_buffer_async_write(bh);
	63	} else {
	64	unlock_buffer(bh);
	65	}
	66	} while ((bh = bh->b_this_page) != head);
	67
	68	/*
	69	* The page and its buffers are protected by PageWriteback(), so we can
	70	* drop the bh refcounts early.
	71	*/
	72	BUG_ON(PageWriteback(page));
	73	set_page_writeback(page);
	74
	75	do {
	76	struct buffer_head *next = bh->b_this_page;
	77	if (buffer_async_write(bh)) {
1420c4a5	78	submit_bh(REQ_OP_WRITE \| write_flags, bh);
4a0f9a32 SW	79	nr_underway++;
	80	}
	81	bh = next;
	82	} while (bh != head);
	83	unlock_page(page);
	84
4a0f9a32 SW	85	if (nr_underway == 0)
	86	end_page_writeback(page);
	87
eaefbf96	88	return 0;
b3b94faa DT	89	}
b3b94faa DT	90
009d8518	91	const struct address_space_operations gfs2_meta_aops = {
e621900a	92	.dirty_folio = block_dirty_folio,
7ba13abb	93	.invalidate_folio = block_invalidate_folio,
b3b94faa	94	.writepage = gfs2_aspace_writepage,
e45c20d1	95	.release_folio = gfs2_release_folio,
b3b94faa DT	96	};
b3b94faa DT	97
1b2ad412	98	const struct address_space_operations gfs2_rgrp_aops = {
e621900a	99	.dirty_folio = block_dirty_folio,
7ba13abb	100	.invalidate_folio = block_invalidate_folio,
1b2ad412	101	.writepage = gfs2_aspace_writepage,
e45c20d1	102	.release_folio = gfs2_release_folio,
1b2ad412 SW	103	};
1b2ad412 SW	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	}
e7dfab82 BP	138	if (!page_has_buffers(page))
e7dfab82 BP	139	create_empty_buffers(page, sdp->sd_sb.sb_bsize, 0);
b3b94faa	140	} else {
2457aec6 MG	141	page = find_get_page_flags(mapping, index,
2457aec6 MG	142	FGP_LOCK\|FGP_ACCESSED);
b3b94faa DT	143	if (!page)
b3b94faa DT	144	return NULL;
e7dfab82 BP	145	if (!page_has_buffers(page)) {
	146	bh = NULL;
	147	goto out_unlock;
	148	}
b3b94faa DT	149	}
b3b94faa DT	150
b3b94faa DT	151	/* Locate header for our buffer within our page */
	152	for (bh = page_buffers(page); bufnum--; bh = bh->b_this_page)
	153	/* Do nothing */;
	154	get_bh(bh);
	155
	156	if (!buffer_mapped(bh))
	157	map_bh(bh, sdp->sd_vfs, blkno);
	158
e7dfab82	159	out_unlock:
b3b94faa	160	unlock_page(page);
09cbfeaf	161	put_page(page);
b3b94faa DT	162
	163	return bh;
	164	}
	165
	166	static void meta_prep_new(struct buffer_head *bh)
	167	{
	168	struct gfs2_meta_header mh = (struct gfs2_meta_header )bh->b_data;
	169
	170	lock_buffer(bh);
	171	clear_buffer_dirty(bh);
	172	set_buffer_uptodate(bh);
	173	unlock_buffer(bh);
	174
	175	mh->mh_magic = cpu_to_be32(GFS2_MAGIC);
	176	}
	177
	178	/**
	179	* gfs2_meta_new - Get a block
	180	* @gl: The glock associated with this block
	181	* @blkno: The block number
	182	*
	183	* Returns: The buffer
	184	*/
	185
cd915493	186	struct buffer_head gfs2_meta_new(struct gfs2_glock gl, u64 blkno)
b3b94faa DT	187	{
b3b94faa DT	188	struct buffer_head *bh;
6802e340	189	bh = gfs2_getbuf(gl, blkno, CREATE);
b3b94faa DT	190	meta_prep_new(bh);
	191	return bh;
	192	}
	193
39b0555f	194	static void gfs2_meta_read_endio(struct bio *bio)
c8d57703	195	{
39b0555f	196	struct bio_vec *bvec;
6dc4f100	197	struct bvec_iter_all iter_all;
39b0555f	198
2b070cfe	199	bio_for_each_segment_all(bvec, bio, iter_all) {
39b0555f AG	200	struct page *page = bvec->bv_page;
	201	struct buffer_head *bh = page_buffers(page);
	202	unsigned int len = bvec->bv_len;
	203
	204	while (bh_offset(bh) < bvec->bv_offset)
	205	bh = bh->b_this_page;
	206	do {
	207	struct buffer_head *next = bh->b_this_page;
	208	len -= bh->b_size;
4e4cbee9	209	bh->b_end_io(bh, !bio->bi_status);
39b0555f AG	210	bh = next;
	211	} while (bh && len);
	212	}
	213	bio_put(bio);
	214	}
c8d57703	215
39b0555f AG	216	/*
	217	* Submit several consecutive buffer head I/O requests as a single bio I/O
	218	* request. (See submit_bh_wbc.)
	219	*/
67688c08	220	static void gfs2_submit_bhs(blk_opf_t opf, struct buffer_head *bhs[], int num)
39b0555f	221	{
23e5671a AG	222	while (num > 0) {
	223	struct buffer_head bh = bhs;
	224	struct bio *bio;
	225
67688c08	226	bio = bio_alloc(bh->b_bdev, num, opf, GFP_NOIO);
23e5671a	227	bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9);
23e5671a AG	228	while (num > 0) {
	229	bh = *bhs;
	230	if (!bio_add_page(bio, bh->b_page, bh->b_size, bh_offset(bh))) {
	231	BUG_ON(bio->bi_iter.bi_size == 0);
	232	break;
	233	}
	234	bhs++;
	235	num--;
	236	}
	237	bio->bi_end_io = gfs2_meta_read_endio;
23e5671a	238	submit_bio(bio);
c8d57703	239	}
c8d57703 AG	240	}
c8d57703 AG	241
b3b94faa DT	242	/**
	243	* gfs2_meta_read - Read a block from disk
	244	* @gl: The glock covering the block
	245	* @blkno: The block number
7276b3b0	246	* @flags: flags
c551f66c	247	* @rahead: Do read-ahead
b3b94faa DT	248	* @bhp: the place where the buffer is returned (NULL on failure)
	249	*
	250	* Returns: errno
	251	*/
	252
cd915493	253	int gfs2_meta_read(struct gfs2_glock *gl, u64 blkno, int flags,
c8d57703	254	int rahead, struct buffer_head **bhp)
b3b94faa	255	{
15562c43	256	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
39b0555f AG	257	struct buffer_head bh, bhs[2];
39b0555f AG	258	int num = 0;
c969f58c	259
8cc67f70	260	if (unlikely(gfs2_withdrawn(sdp)) && !gfs2_withdraw_in_prog(sdp)) {
44b8db13	261	*bhp = NULL;
c969f58c	262	return -EIO;
44b8db13	263	}
c969f58c SW	264
	265	*bhp = bh = gfs2_getbuf(gl, blkno, CREATE);
	266
	267	lock_buffer(bh);
	268	if (buffer_uptodate(bh)) {
	269	unlock_buffer(bh);
39b0555f AG	270	flags &= ~DIO_WAIT;
	271	} else {
	272	bh->b_end_io = end_buffer_read_sync;
	273	get_bh(bh);
	274	bhs[num++] = bh;
c969f58c	275	}
39b0555f AG	276
	277	if (rahead) {
	278	bh = gfs2_getbuf(gl, blkno + 1, CREATE);
	279
	280	lock_buffer(bh);
	281	if (buffer_uptodate(bh)) {
	282	unlock_buffer(bh);
	283	brelse(bh);
	284	} else {
	285	bh->b_end_io = end_buffer_read_sync;
	286	bhs[num++] = bh;
	287	}
	288	}
	289
67688c08	290	gfs2_submit_bhs(REQ_OP_READ \| REQ_META \| REQ_PRIO, bhs, num);
c969f58c SW	291	if (!(flags & DIO_WAIT))
	292	return 0;
	293
39b0555f	294	bh = *bhp;
c969f58c SW	295	wait_on_buffer(bh);
	296	if (unlikely(!buffer_uptodate(bh))) {
	297	struct gfs2_trans *tr = current->journal_info;
9862ca05	298	if (tr && test_bit(TR_TOUCHED, &tr->tr_flags))
9e1a9ecd	299	gfs2_io_error_bh_wd(sdp, bh);
c969f58c	300	brelse(bh);
44b8db13	301	*bhp = NULL;
c969f58c	302	return -EIO;
7276b3b0	303	}
b3b94faa	304
7276b3b0	305	return 0;
b3b94faa DT	306	}
	307
	308	/**
7276b3b0	309	* gfs2_meta_wait - Reread a block from disk
b3b94faa	310	* @sdp: the filesystem
7276b3b0	311	* @bh: The block to wait for
b3b94faa DT	312	*
	313	* Returns: errno
	314	*/
	315
7276b3b0	316	int gfs2_meta_wait(struct gfs2_sbd sdp, struct buffer_head bh)
b3b94faa	317	{
8cc67f70	318	if (unlikely(gfs2_withdrawn(sdp)) && !gfs2_withdraw_in_prog(sdp))
b3b94faa DT	319	return -EIO;
b3b94faa DT	320
7276b3b0	321	wait_on_buffer(bh);
b3b94faa	322
7276b3b0 SW	323	if (!buffer_uptodate(bh)) {
7276b3b0 SW	324	struct gfs2_trans *tr = current->journal_info;
9862ca05	325	if (tr && test_bit(TR_TOUCHED, &tr->tr_flags))
9e1a9ecd	326	gfs2_io_error_bh_wd(sdp, bh);
7276b3b0	327	return -EIO;
b3b94faa	328	}
8cc67f70	329	if (unlikely(gfs2_withdrawn(sdp)) && !gfs2_withdraw_in_prog(sdp))
7276b3b0	330	return -EIO;
b3b94faa DT	331
	332	return 0;
	333	}
	334
68cd4ce2	335	void gfs2_remove_from_journal(struct buffer_head *bh, int meta)
16615be1	336	{
009d8518 SW	337	struct address_space *mapping = bh->b_page->mapping;
009d8518 SW	338	struct gfs2_sbd *sdp = gfs2_mapping2sbd(mapping);
16615be1	339	struct gfs2_bufdata *bd = bh->b_private;
68cd4ce2	340	struct gfs2_trans *tr = current->journal_info;
502be2a3	341	int was_pinned = 0;
009d8518	342
16615be1	343	if (test_clear_buffer_pinned(bh)) {
627c10b7	344	trace_gfs2_pin(bd, 0);
5e687eac	345	atomic_dec(&sdp->sd_log_pinned);
c0752aa7	346	list_del_init(&bd->bd_list);
68cd4ce2	347	if (meta == REMOVE_META)
16615be1	348	tr->tr_num_buf_rm++;
022ef4fe	349	else
16615be1	350	tr->tr_num_databuf_rm++;
9862ca05	351	set_bit(TR_TOUCHED, &tr->tr_flags);
502be2a3	352	was_pinned = 1;
16615be1 SW	353	brelse(bh);
	354	}
	355	if (bd) {
16ca9412	356	if (bd->bd_tr) {
16615be1	357	gfs2_trans_add_revoke(sdp, bd);
502be2a3 BP	358	} else if (was_pinned) {
	359	bh->b_private = NULL;
	360	kmem_cache_free(gfs2_bufdata_cachep, bd);
68942870 BP	361	} else if (!list_empty(&bd->bd_ail_st_list) &&
	362	!list_empty(&bd->bd_ail_gl_list)) {
	363	gfs2_remove_from_ail(bd);
16615be1 SW	364	}
	365	}
	366	clear_buffer_dirty(bh);
	367	clear_buffer_uptodate(bh);
	368	}
	369
b3b94faa	370	/**
68942870 BP	371	* gfs2_ail1_wipe - remove deleted/freed buffers from the ail1 list
	372	* @sdp: superblock
	373	* @bstart: starting block address of buffers to remove
	374	* @blen: length of buffers to be removed
	375	*
	376	* This function is called from gfs2_journal wipe, whose job is to remove
	377	* buffers, corresponding to deleted blocks, from the journal. If we find any
	378	* bufdata elements on the system ail1 list, they haven't been written to
	379	* the journal yet. So we remove them.
	380	*/
	381	static void gfs2_ail1_wipe(struct gfs2_sbd *sdp, u64 bstart, u32 blen)
	382	{
	383	struct gfs2_trans tr, s;
	384	struct gfs2_bufdata bd, bs;
	385	struct buffer_head *bh;
	386	u64 end = bstart + blen;
	387
	388	gfs2_log_lock(sdp);
	389	spin_lock(&sdp->sd_ail_lock);
	390	list_for_each_entry_safe(tr, s, &sdp->sd_ail1_list, tr_list) {
	391	list_for_each_entry_safe(bd, bs, &tr->tr_ail1_list,
	392	bd_ail_st_list) {
	393	bh = bd->bd_bh;
	394	if (bh->b_blocknr < bstart \|\| bh->b_blocknr >= end)
	395	continue;
	396
	397	gfs2_remove_from_journal(bh, REMOVE_JDATA);
	398	}
	399	}
	400	spin_unlock(&sdp->sd_ail_lock);
	401	gfs2_log_unlock(sdp);
	402	}
	403
	404	static struct buffer_head gfs2_getjdatabuf(struct gfs2_inode ip, u64 blkno)
	405	{
	406	struct address_space *mapping = ip->i_inode.i_mapping;
	407	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
	408	struct page *page;
	409	struct buffer_head *bh;
	410	unsigned int shift = PAGE_SHIFT - sdp->sd_sb.sb_bsize_shift;
	411	unsigned long index = blkno >> shift; /* convert block to page */
	412	unsigned int bufnum = blkno - (index << shift);
	413
	414	page = find_get_page_flags(mapping, index, FGP_LOCK\|FGP_ACCESSED);
	415	if (!page)
	416	return NULL;
	417	if (!page_has_buffers(page)) {
	418	unlock_page(page);
	419	put_page(page);
	420	return NULL;
	421	}
	422	/* Locate header for our buffer within our page */
	423	for (bh = page_buffers(page); bufnum--; bh = bh->b_this_page)
	424	/* Do nothing */;
	425	get_bh(bh);
	426	unlock_page(page);
	427	put_page(page);
	428	return bh;
	429	}
	430
	431	/**
	432	* gfs2_journal_wipe - make inode's buffers so they aren't dirty/pinned anymore
b3b94faa DT	433	* @ip: the inode who owns the buffers
	434	* @bstart: the first buffer in the run
	435	* @blen: the number of buffers in the run
	436	*
	437	*/
	438
68942870	439	void gfs2_journal_wipe(struct gfs2_inode *ip, u64 bstart, u32 blen)
b3b94faa	440	{
feaa7bba	441	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
b3b94faa	442	struct buffer_head *bh;
68942870	443	int ty;
b3b94faa	444
68942870	445	gfs2_ail1_wipe(sdp, bstart, blen);
b3b94faa	446	while (blen) {
68942870	447	ty = REMOVE_META;
6802e340	448	bh = gfs2_getbuf(ip->i_gl, bstart, NO_CREATE);
68942870 BP	449	if (!bh && gfs2_is_jdata(ip)) {
	450	bh = gfs2_getjdatabuf(ip, bstart);
	451	ty = REMOVE_JDATA;
	452	}
b3b94faa	453	if (bh) {
1ad38c43 SW	454	lock_buffer(bh);
1ad38c43 SW	455	gfs2_log_lock(sdp);
68942870 BP	456	spin_lock(&sdp->sd_ail_lock);
	457	gfs2_remove_from_journal(bh, ty);
	458	spin_unlock(&sdp->sd_ail_lock);
1ad38c43	459	gfs2_log_unlock(sdp);
b3b94faa	460	unlock_buffer(bh);
b3b94faa DT	461	brelse(bh);
	462	}
	463
	464	bstart++;
	465	blen--;
	466	}
	467	}
	468
b3b94faa	469	/**
6d8da302	470	* gfs2_meta_buffer - Get a metadata buffer
b3b94faa	471	* @ip: The GFS2 inode
6d8da302	472	* @mtype: The block type (GFS2_METATYPE_*)
b3b94faa	473	* @num: The block number (device relative) of the buffer
b3b94faa DT	474	* @bhp: the buffer is returned here
b3b94faa DT	475	*
b3b94faa DT	476	* Returns: errno
	477	*/
	478
6d8da302 AG	479	int gfs2_meta_buffer(struct gfs2_inode *ip, u32 mtype, u64 num,
6d8da302 AG	480	struct buffer_head **bhp)
b3b94faa	481	{
7276b3b0 SW	482	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
7276b3b0 SW	483	struct gfs2_glock *gl = ip->i_gl;
f91a0d3e SW	484	struct buffer_head *bh;
f91a0d3e SW	485	int ret = 0;
c8d57703 AG	486	int rahead = 0;
	487
	488	if (num == ip->i_no_addr)
	489	rahead = ip->i_rahead;
b3b94faa	490
c8d57703	491	ret = gfs2_meta_read(gl, num, DIO_WAIT, rahead, &bh);
f2f9c812 BP	492	if (ret == 0 && gfs2_metatype_check(sdp, bh, mtype)) {
	493	brelse(bh);
	494	ret = -EIO;
61eaadcd BP	495	} else {
61eaadcd BP	496	*bhp = bh;
b3b94faa	497	}
f91a0d3e	498	return ret;
b3b94faa DT	499	}
	500
	501	/**
	502	* gfs2_meta_ra - start readahead on an extent of a file
	503	* @gl: the glock the blocks belong to
	504	* @dblock: the starting disk block
	505	* @extlen: the number of blocks in the extent
	506	*
7276b3b0	507	* returns: the first buffer in the extent
b3b94faa DT	508	*/
b3b94faa DT	509
7276b3b0	510	struct buffer_head gfs2_meta_ra(struct gfs2_glock gl, u64 dblock, u32 extlen)
b3b94faa	511	{
15562c43	512	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
b3b94faa	513	struct buffer_head first_bh, bh;
cd915493	514	u32 max_ra = gfs2_tune_get(sdp, gt_max_readahead) >>
568f4c96	515	sdp->sd_sb.sb_bsize_shift;
b3b94faa	516
7276b3b0 SW	517	BUG_ON(!extlen);
	518
	519	if (max_ra < 1)
	520	max_ra = 1;
b3b94faa DT	521	if (extlen > max_ra)
	522	extlen = max_ra;
	523
6802e340	524	first_bh = gfs2_getbuf(gl, dblock, CREATE);
b3b94faa DT	525
	526	if (buffer_uptodate(first_bh))
	527	goto out;
7276b3b0	528	if (!buffer_locked(first_bh))
1420c4a5	529	ll_rw_block(REQ_OP_READ \| REQ_META \| REQ_PRIO, 1, &first_bh);
b3b94faa DT	530
	531	dblock++;
	532	extlen--;
	533
	534	while (extlen) {
6802e340	535	bh = gfs2_getbuf(gl, dblock, CREATE);
b3b94faa	536
7276b3b0	537	if (!buffer_uptodate(bh) && !buffer_locked(bh))
1420c4a5 BVA	538	ll_rw_block(REQ_OP_READ \| REQ_RAHEAD \| REQ_META \|
1420c4a5 BVA	539	REQ_PRIO, 1, &bh);
7276b3b0	540	brelse(bh);
b3b94faa DT	541	dblock++;
b3b94faa DT	542	extlen--;
7276b3b0 SW	543	if (!buffer_locked(first_bh) && buffer_uptodate(first_bh))
7276b3b0 SW	544	goto out;
b3b94faa DT	545	}
b3b94faa DT	546
7276b3b0	547	wait_on_buffer(first_bh);
a91ea69f	548	out:
7276b3b0	549	return first_bh;
b3b94faa DT	550	}
b3b94faa DT	551