dio.c
No OneTemporary
Actions

Size

30 KB

Referenced Files

None

Subscribers

None

dio.c
View Options

	/******************************************************************************
	*******************************************************************************
	**
	** Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved.
	** Copyright (C) 2004-2006 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 v.2.
	**
	*******************************************************************************
	******************************************************************************/

	#include <linux/sched.h>
	#include <linux/slab.h>
	#include <linux/smp_lock.h>
	#include <linux/spinlock.h>
	#include <asm/semaphore.h>
	#include <linux/completion.h>
	#include <linux/buffer_head.h>
	#include <linux/mm.h>
	#include <linux/pagemap.h>
	#include <linux/writeback.h>

	#include "gfs.h"
	#include "dio.h"
	#include "glock.h"
	#include "glops.h"
	#include "inode.h"
	#include "log.h"
	#include "lops.h"
	#include "rgrp.h"
	#include "trans.h"

	#define buffer_busy(bh) ((bh)->b_state & ((1ul << BH_Dirty) \| (1ul << BH_Lock)))

	/**
	* aspace_get_block -
	* @inode:
	* @lblock:
	* @bh_result:
	* @create:
	*
	* Returns: errno
	*/

	static int
	aspace_get_block(struct inode *inode, sector_t lblock,
	struct buffer_head *bh_result, int create)
	{
	gfs_assert_warn(get_v2sdp(inode->i_sb), FALSE);
	return -ENOSYS;
	}

	/**
	* gfs_aspace_writepage - write an aspace page
	* @page: the page
	* @wbc:
	*
	* Returns: errno
	*/

	static int
	gfs_aspace_writepage(struct page page, struct writeback_control wbc)
	{
	return block_write_full_page(page, aspace_get_block, wbc);
	}

	/**
	* stuck_releasepage - We're stuck in gfs_releasepage(). Print stuff out.
	* @bh: the buffer we're stuck on
	*
	*/

	static void
	stuck_releasepage(struct buffer_head *bh)
	{
	struct gfs_sbd *sdp = get_v2sdp(bh->b_page->mapping->host->i_sb);
	struct gfs_bufdata *bd = get_v2bd(bh);

	printk("GFS: fsid=%s: stuck in gfs_releasepage()...\n", sdp->sd_fsname);
	printk("GFS: fsid=%s: blkno = %"PRIu64", bh->b_count = %d\n",
	sdp->sd_fsname,
	(uint64_t)bh->b_blocknr,
	atomic_read(&bh->b_count));
	printk("GFS: fsid=%s: get_v2bd(bh) = %s\n",
	sdp->sd_fsname,
	(bd) ? "!NULL" : "NULL");

	if (bd) {
	struct gfs_glock *gl = bd->bd_gl;

	printk("GFS: fsid=%s: gl = (%u, %"PRIu64")\n",
	sdp->sd_fsname,
	gl->gl_name.ln_type,
	gl->gl_name.ln_number);

	printk("GFS: fsid=%s: bd_new_le.le_trans = %s\n",
	sdp->sd_fsname,
	(bd->bd_new_le.le_trans) ? "!NULL" : "NULL");
	printk("GFS: fsid=%s: bd_incore_le.le_trans = %s\n",
	sdp->sd_fsname,
	(bd->bd_incore_le.le_trans) ? "!NULL" : "NULL");
	printk("GFS: fsid=%s: bd_frozen = %s\n",
	sdp->sd_fsname,
	(bd->bd_frozen) ? "!NULL" : "NULL");
	printk("GFS: fsid=%s: bd_pinned = %u\n",
	sdp->sd_fsname, bd->bd_pinned);
	printk("GFS: fsid=%s: bd_ail_tr_list = %s\n",
	sdp->sd_fsname,
	(list_empty(&bd->bd_ail_tr_list)) ? "Empty" : "!Empty");

	if (gl->gl_ops == &gfs_inode_glops) {
	struct gfs_inode *ip = get_gl2ip(gl);

	if (ip) {
	unsigned int x;

	printk("GFS: fsid=%s: ip = %"PRIu64"/%"PRIu64"\n",
	sdp->sd_fsname,
	ip->i_num.no_formal_ino,
	ip->i_num.no_addr);
	printk("GFS: fsid=%s: ip->i_count = %d, ip->i_vnode = %s\n",
	sdp->sd_fsname,
	atomic_read(&ip->i_count),
	(ip->i_vnode) ? "!NULL" : "NULL");
	for (x = 0; x < GFS_MAX_META_HEIGHT; x++)
	printk("GFS: fsid=%s: ip->i_cache[%u] = %s\n",
	sdp->sd_fsname, x,
	(ip->i_cache[x]) ? "!NULL" : "NULL");
	}
	}
	}
	}

	/**
	* gfs_aspace_releasepage - free the metadata associated with a page
	* @page: the page that's being released
	* @gfp_mask: passed from Linux VFS, ignored by us
	*
	* Call try_to_free_buffers() if the buffers in this page can be
	* released.
	*
	* Returns: 0
	*/

	static int
	gfs_aspace_releasepage(struct page *page, gfp_t gfp_mask)
	{
	struct inode *aspace = page->mapping->host;
	struct gfs_sbd *sdp = get_v2sdp(aspace->i_sb);
	struct buffer_head bh, head;
	struct gfs_bufdata *bd;
	unsigned long t;

	if (!page_has_buffers(page))
	goto out;

	head = bh = page_buffers(page);
	do {
	t = jiffies;

	while (atomic_read(&bh->b_count)) {
	if (atomic_read(&aspace->i_writecount)) {
	if (time_after_eq(jiffies,
	t +
	gfs_tune_get(sdp, gt_stall_secs) * HZ)) {
	stuck_releasepage(bh);
	t = jiffies;
	}

	yield();
	continue;
	}

	return 0;
	}

	bd = get_v2bd(bh);
	if (bd) {
	gfs_assert_warn(sdp, bd->bd_bh == bh);
	gfs_assert_warn(sdp, !bd->bd_new_le.le_trans);
	gfs_assert_warn(sdp, !bd->bd_incore_le.le_trans);
	gfs_assert_warn(sdp, !bd->bd_frozen);
	gfs_assert_warn(sdp, !bd->bd_pinned);
	gfs_assert_warn(sdp, list_empty(&bd->bd_ail_tr_list));
	kmem_cache_free(gfs_bufdata_cachep, bd);
	atomic_dec(&sdp->sd_bufdata_count);
	set_v2bd(bh, NULL);
	}

	bh = bh->b_this_page;
	}
	while (bh != head);

	out:
	return try_to_free_buffers(page);
	}

	static struct address_space_operations aspace_aops = {
	.writepage = gfs_aspace_writepage,
	.releasepage = gfs_aspace_releasepage,
	};

	/**
	* gfs_aspace_get - Create and initialize a struct inode structure
	* @sdp: the filesystem the aspace is in
	*
	* Right now a struct inode is just a struct inode. Maybe Linux
	* will supply a more lightweight address space construct (that works)
	* in the future.
	*
	* Make sure pages/buffers in this aspace aren't in high memory.
	*
	* Returns: the aspace
	*/

	struct inode *
	gfs_aspace_get(struct gfs_sbd *sdp)
	{
	struct inode *aspace;

	aspace = new_inode(sdp->sd_vfs);
	if (aspace) {
	mapping_set_gfp_mask(aspace->i_mapping, GFP_KERNEL);
	aspace->i_mapping->a_ops = &aspace_aops;
	aspace->i_size = ~0ULL;
	set_v2ip(aspace, NULL);
	insert_inode_hash(aspace);
	}

	return aspace;
	}

	/**
	* gfs_aspace_put - get rid of an aspace
	* @aspace:
	*
	*/

	void
	gfs_aspace_put(struct inode *aspace)
	{
	remove_inode_hash(aspace);
	iput(aspace);
	}

	/**
	* gfs_ail_start_trans - Start I/O on a part of the AIL
	* @sdp: the filesystem
	* @tr: the part of the AIL
	*
	*/

	void
	gfs_ail_start_trans(struct gfs_sbd sdp, struct gfs_trans tr)
	{
	struct list_head head, tmp, *prev;
	struct gfs_bufdata *bd;
	struct buffer_head *bh;
	int retry;

	do {
	retry = FALSE;

	spin_lock(&sdp->sd_ail_lock);

	for (head = &tr->tr_ail_bufs, tmp = head->prev, prev = tmp->prev;
	tmp != head;
	tmp = prev, prev = tmp->prev) {
	bd = list_entry(tmp, struct gfs_bufdata, bd_ail_tr_list);
	bh = bd->bd_bh;

	if (gfs_trylock_buffer(bh))
	continue;

	if (bd->bd_pinned) {
	gfs_unlock_buffer(bh);
	continue;
	}

	if (!buffer_busy(bh)) {
	if (!buffer_uptodate(bh))
	gfs_io_error_bh(sdp, bh);

	list_del_init(&bd->bd_ail_tr_list);
	list_del(&bd->bd_ail_gl_list);

	gfs_unlock_buffer(bh);
	brelse(bh);
	continue;
	}

	if (buffer_dirty(bh)) {
	list_move(&bd->bd_ail_tr_list, head);

	spin_unlock(&sdp->sd_ail_lock);
	wait_on_buffer(bh);
	ll_rw_block(WRITE, 1, &bh);
	spin_lock(&sdp->sd_ail_lock);

	gfs_unlock_buffer(bh);
	retry = TRUE;
	break;
	}

	gfs_unlock_buffer(bh);
	}

	spin_unlock(&sdp->sd_ail_lock);
	} while (retry);
	}

	/**
	* gfs_ail_empty_trans - Check whether or not a trans in the AIL has been synced
	* @sdp: the filesystem
	* @tr: the transaction
	*
	*/

	int
	gfs_ail_empty_trans(struct gfs_sbd sdp, struct gfs_trans tr)
	{
	struct list_head head, tmp, *prev;
	struct gfs_bufdata *bd;
	struct buffer_head *bh;
	int ret;

	spin_lock(&sdp->sd_ail_lock);

	for (head = &tr->tr_ail_bufs, tmp = head->prev, prev = tmp->prev;
	tmp != head;
	tmp = prev, prev = tmp->prev) {
	bd = list_entry(tmp, struct gfs_bufdata, bd_ail_tr_list);
	bh = bd->bd_bh;

	if (gfs_trylock_buffer(bh))
	continue;

	if (bd->bd_pinned \|\| buffer_busy(bh)) {
	gfs_unlock_buffer(bh);
	continue;
	}

	if (!buffer_uptodate(bh))
	gfs_io_error_bh(sdp, bh);

	list_del_init(&bd->bd_ail_tr_list);
	list_del(&bd->bd_ail_gl_list);

	gfs_unlock_buffer(bh);
	brelse(bh);
	}

	ret = list_empty(head);

	spin_unlock(&sdp->sd_ail_lock);

	return ret;
	}

	/**
	* ail_empty_gl - remove all buffers for a given lock from the AIL
	* @gl: the glock
	*
	* None of the buffers should be dirty, locked, or pinned.
	*/

	static void
	ail_empty_gl(struct gfs_glock *gl)
	{
	struct gfs_sbd *sdp = gl->gl_sbd;
	struct gfs_bufdata *bd;
	struct buffer_head *bh;

	spin_lock(&sdp->sd_ail_lock);

	while (!list_empty(&gl->gl_ail_bufs)) {
	bd = list_entry(gl->gl_ail_bufs.next,
	struct gfs_bufdata, bd_ail_gl_list);
	bh = bd->bd_bh;

	gfs_assert_withdraw(sdp, !bd->bd_pinned && !buffer_busy(bh));
	if (!buffer_uptodate(bh))
	gfs_io_error_bh(sdp, bh);

	list_del_init(&bd->bd_ail_tr_list);
	list_del(&bd->bd_ail_gl_list);

	brelse(bh);
	}

	spin_unlock(&sdp->sd_ail_lock);
	}

	/**
	* gfs_inval_buf - Invalidate all buffers associated with a glock
	* @gl: the glock
	*
	*/

	void
	gfs_inval_buf(struct gfs_glock *gl)
	{
	struct inode *aspace = gl->gl_aspace;
	struct address_space *mapping = gl->gl_aspace->i_mapping;

	ail_empty_gl(gl);

	atomic_inc(&aspace->i_writecount);
	truncate_inode_pages(mapping, 0);
	atomic_dec(&aspace->i_writecount);

	gfs_assert_withdraw(gl->gl_sbd, !mapping->nrpages);
	}

	/**
	* gfs_sync_buf - Sync all buffers associated with a glock
	* @gl: The glock
	* @flags: DIO_START \| DIO_WAIT \| DIO_CHECK
	*
	*/

	void
	gfs_sync_buf(struct gfs_glock *gl, int flags)
	{
	struct address_space *mapping = gl->gl_aspace->i_mapping;
	int error = 0;

	if (flags & DIO_START)
	error = filemap_fdatawrite(mapping);
	if (!error && (flags & DIO_WAIT))
	error = filemap_fdatawait(mapping);
	if (!error && (flags & (DIO_INVISIBLE \| DIO_CHECK)) == DIO_CHECK)
	ail_empty_gl(gl);

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

	/**
	* getbuf - Get a buffer with a given address space
	* @sdp: the filesystem
	* @aspace: the address space
	* @blkno: the block number (filesystem scope)
	* @create: TRUE if the buffer should be created
	*
	* Returns: the buffer
	*/

	static struct buffer_head *
	getbuf(struct gfs_sbd sdp, struct inode aspace, uint64_t blkno, int create)
	{
	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) {
	RETRY_MALLOC(page = grab_cache_page(aspace->i_mapping, index), page);
	} else {
	page = find_lock_page(aspace->i_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);
	else if (gfs_assert_warn(sdp, bh->b_bdev == sdp->sd_vfs->s_bdev &&
	bh->b_blocknr == blkno))
	map_bh(bh, sdp->sd_vfs, blkno);

	unlock_page(page);
	page_cache_release(page);

	return bh;
	}

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

	struct buffer_head *
	gfs_dgetblk(struct gfs_glock *gl, uint64_t blkno)
	{
	return getbuf(gl->gl_sbd, gl->gl_aspace, blkno, CREATE);
	}

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

	int
	gfs_dread(struct gfs_glock *gl, uint64_t blkno,
	int flags, struct buffer_head **bhp)
	{
	int error;

	*bhp = gfs_dgetblk(gl, blkno);
	error = gfs_dreread(gl->gl_sbd, *bhp, flags);
	if (error)
	brelse(*bhp);

	return error;
	}

	/**
	* gfs_prep_new_buffer - Mark a new buffer we just gfs_dgetblk()ed uptodate
	* @bh: the buffer
	*
	*/

	void
	gfs_prep_new_buffer(struct buffer_head *bh)
	{
	wait_on_buffer(bh);
	clear_buffer_dirty(bh);
	set_buffer_uptodate(bh);
	}

	/**
	* gfs_dreread - Reread a block from disk
	* @sdp: the filesystem
	* @bh: The block to read
	* @flags: Flags that control the read
	*
	* Returns: errno
	*/

	int
	gfs_dreread(struct gfs_sbd sdp, struct buffer_head bh, int flags)
	{
	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
	return -EIO;

	/* Fill in meta-header if we have a cached copy, else read from disk */
	if (flags & DIO_NEW) {
	if (gfs_mhc_fish(sdp, bh))
	return 0;
	clear_buffer_uptodate(bh);
	}

	if (flags & DIO_FORCE)
	clear_buffer_uptodate(bh);

	if ((flags & DIO_START) && !buffer_uptodate(bh))
	ll_rw_block(READ, 1, &bh);

	if (flags & DIO_WAIT) {
	wait_on_buffer(bh);

	if (!buffer_uptodate(bh)) {
	gfs_io_error_bh(sdp, bh);
	return -EIO;
	}
	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
	return -EIO;
	}

	return 0;
	}

	/**
	* gfs_dwrite - Write a buffer to disk (and/or wait for write to complete)
	* @sdp: the filesystem
	* @bh: The buffer to write
	* @flags: DIO_XXX The type of write/wait operation to do
	*
	* Returns: errno
	*/

	int
	gfs_dwrite(struct gfs_sbd sdp, struct buffer_head bh, int flags)
	{
	if (gfs_assert_warn(sdp, !test_bit(SDF_ROFS, &sdp->sd_flags)))
	return -EIO;
	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
	return -EIO;

	if (flags & DIO_CLEAN) {
	lock_buffer(bh);
	clear_buffer_dirty(bh);
	unlock_buffer(bh);
	}

	if (flags & DIO_DIRTY) {
	if (gfs_assert_warn(sdp, buffer_uptodate(bh)))
	return -EIO;
	mark_buffer_dirty(bh);
	}

	if ((flags & DIO_START) && buffer_dirty(bh)) {
	wait_on_buffer(bh);
	ll_rw_block(WRITE, 1, &bh);
	}

	if (flags & DIO_WAIT) {
	wait_on_buffer(bh);

	if (!buffer_uptodate(bh) \|\| buffer_dirty(bh)) {
	gfs_io_error_bh(sdp, bh);
	return -EIO;
	}
	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
	return -EIO;
	}

	return 0;
	}

	/**
	* gfs_attach_bufdata - attach a struct gfs_bufdata structure to a buffer
	* @bh: The buffer to be attached to
	* @gl: the glock the buffer belongs to
	*
	*/

	void
	gfs_attach_bufdata(struct buffer_head bh, struct gfs_glock gl)
	{
	struct gfs_bufdata *bd;

	lock_page(bh->b_page);

	/* If there's one attached already, we're done */
	if (get_v2bd(bh)) {
	unlock_page(bh->b_page);
	return;
	}

	RETRY_MALLOC(bd = kmem_cache_alloc(gfs_bufdata_cachep, GFP_KERNEL), bd);
	atomic_inc(&gl->gl_sbd->sd_bufdata_count);

	memset(bd, 0, sizeof(struct gfs_bufdata));

	bd->bd_bh = bh;
	bd->bd_gl = gl;

	INIT_LE(&bd->bd_new_le, &gfs_buf_lops);
	INIT_LE(&bd->bd_incore_le, &gfs_buf_lops);

	init_MUTEX(&bd->bd_lock);

	INIT_LIST_HEAD(&bd->bd_ail_tr_list);

	set_v2bd(bh, bd);

	unlock_page(bh->b_page);
	}

	/**
	* gfs_is_pinned - Figure out if a buffer is pinned or not
	* @sdp: the filesystem the buffer belongs to
	* @bh: The buffer to be pinned
	*
	* Returns: TRUE if the buffer is pinned, FALSE otherwise
	*/

	int
	gfs_is_pinned(struct gfs_sbd sdp, struct buffer_head bh)
	{
	struct gfs_bufdata *bd = get_v2bd(bh);
	int ret = FALSE;

	if (bd) {
	gfs_lock_buffer(bh);
	if (bd->bd_pinned)
	ret = TRUE;
	gfs_unlock_buffer(bh);
	}

	return ret;
	}

	/**
	* gfs_dpin - Pin a metadata buffer in memory
	* @sdp: the filesystem the buffer belongs to
	* @bh: The buffer to be pinned
	*
	* "Pinning" means keeping buffer from being written to its in-place location.
	* A buffer should be pinned from the time it is added to a new transaction,
	* until after it has been written to the log.
	* If an earlier change to this buffer is still pinned, waiting to be written
	* to on-disk log, we need to keep a "frozen" copy of the old data while this
	* transaction is modifying the real data. We keep the frozen copy until
	* this transaction's incore_commit(), i.e. until the transaction has
	* finished modifying the real data, at which point we can use the real
	* buffer for logging, even if the frozen copy didn't get written to the log.
	*
	*/

	void
	gfs_dpin(struct gfs_sbd sdp, struct buffer_head bh)
	{
	struct gfs_bufdata *bd = get_v2bd(bh);
	char *data;

	gfs_assert_withdraw(sdp, !test_bit(SDF_ROFS, &sdp->sd_flags));

	gfs_lock_buffer(bh);

	gfs_assert_warn(sdp, !bd->bd_frozen);

	if (!bd->bd_pinned++) {
	wait_on_buffer(bh);

	/* If this buffer is in the AIL and it has already been written
	to in-place disk block, remove it from the AIL. */

	spin_lock(&sdp->sd_ail_lock);
	if (!list_empty(&bd->bd_ail_tr_list) && !buffer_busy(bh)) {
	list_del_init(&bd->bd_ail_tr_list);
	list_del(&bd->bd_ail_gl_list);
	brelse(bh);
	}
	spin_unlock(&sdp->sd_ail_lock);

	clear_buffer_dirty(bh);
	wait_on_buffer(bh);

	if (!buffer_uptodate(bh))
	gfs_io_error_bh(sdp, bh);
	} else {
	gfs_unlock_buffer(bh);

	gfs_assert_withdraw(sdp, buffer_uptodate(bh));

	data = gmalloc(sdp->sd_sb.sb_bsize);

	gfs_lock_buffer(bh);

	/* Create frozen copy, if needed. */
	if (bd->bd_pinned > 1) {
	memcpy(data, bh->b_data, sdp->sd_sb.sb_bsize);
	bd->bd_frozen = data;
	} else
	kfree(data);
	}

	gfs_unlock_buffer(bh);

	get_bh(bh);
	}

	/**
	* gfs_dunpin - Unpin a buffer
	* @sdp: the filesystem the buffer belongs to
	* @bh: The buffer to unpin
	* @tr: The transaction in the AIL that contains this buffer
	* If NULL, don't attach buffer to any AIL list
	* (i.e. when dropping a pin reference when merging a new transaction
	* with an already existing incore transaction)
	*
	* Called for (meta) buffers, after they've been logged to on-disk journal.
	* Make a (meta) buffer writeable to in-place location on-disk, if recursive
	* pin count is 1 (i.e. no other, later transaction is modifying this buffer).
	* Add buffer to AIL lists of 1) the latest transaction that's modified and
	* logged (on-disk) the buffer, and of 2) the glock that protects the buffer.
	* A single buffer might have been modified by more than one transaction
	* since the buffer's previous write to disk (in-place location). We keep
	* the buffer on only one transaction's AIL list, i.e. that of the latest
	* transaction that's completed logging this buffer (no need to write it to
	* in-place block multiple times for multiple transactions, only once with
	* the most up-to-date data).
	* A single buffer will be protected by one and only one glock. If buffer is
	* already on a (previous) transaction's AIL, we know that we're already
	* on buffer's glock's AIL.
	*
	*/

	void
	gfs_dunpin(struct gfs_sbd sdp, struct buffer_head bh, struct gfs_trans *tr)
	{
	struct gfs_bufdata *bd = get_v2bd(bh);

	gfs_assert_withdraw(sdp, buffer_uptodate(bh));

	gfs_lock_buffer(bh);

	if (gfs_assert_warn(sdp, bd->bd_pinned)) {
	gfs_unlock_buffer(bh);
	return;
	}

	/* No other (later) transaction is modifying buffer; ready to write */
	if (bd->bd_pinned == 1)
	mark_buffer_dirty(bh);

	bd->bd_pinned--;

	gfs_unlock_buffer(bh);

	if (tr) {
	spin_lock(&sdp->sd_ail_lock);

	if (list_empty(&bd->bd_ail_tr_list)) {
	/* Buffer not attached to any earlier transaction. Add
	it to glock's AIL, and this trans' AIL (below). */
	list_add(&bd->bd_ail_gl_list, &bd->bd_gl->gl_ail_bufs);
	} else {
	/* Was part of earlier transaction.
	Move from that trans' AIL to this newer one's AIL.
	Buf is already on glock's AIL. */
	list_del_init(&bd->bd_ail_tr_list);
	brelse(bh);
	}
	list_add(&bd->bd_ail_tr_list, &tr->tr_ail_bufs);

	spin_unlock(&sdp->sd_ail_lock);
	} else
	brelse(bh);
	}

	/**
	* logbh_end_io - Called by OS at the end of a logbh ("fake" bh) write to log
	* @bh: the buffer
	* @uptodate: whether or not the write succeeded
	*
	* Interrupt context, no ENTER/RETURN
	*
	*/

	static void
	logbh_end_io(struct buffer_head *bh, int uptodate)
	{
	if (uptodate)
	set_buffer_uptodate(bh);
	else
	clear_buffer_uptodate(bh);
	unlock_buffer(bh);
	}

	/**
	* gfs_logbh_init - Initialize a fake buffer head
	* @sdp: the filesystem
	* @bh: the buffer to initialize
	* @blkno: the block address of the buffer
	* @data: the data to be written
	*
	*/

	void
	gfs_logbh_init(struct gfs_sbd sdp, struct buffer_head bh,
	uint64_t blkno, char *data)
	{
	memset(bh, 0, sizeof(struct buffer_head));
	bh->b_state = (1 << BH_Mapped) \| (1 << BH_Uptodate) \| (1 << BH_Lock);
	atomic_set(&bh->b_count, 1);
	set_bh_page(bh, virt_to_page(data), ((unsigned long)data) & (PAGE_SIZE - 1));
	bh->b_blocknr = blkno;
	bh->b_size = sdp->sd_sb.sb_bsize;
	bh->b_bdev = sdp->sd_vfs->s_bdev;
	init_buffer(bh, logbh_end_io, NULL);
	INIT_LIST_HEAD(&bh->b_assoc_buffers);
	}

	/**
	* gfs_logbh_uninit - Clean up a fake buffer head
	* @sdp: the filesystem
	* @bh: the buffer to clean
	*
	*/

	void
	gfs_logbh_uninit(struct gfs_sbd sdp, struct buffer_head bh)
	{
	gfs_assert_warn(sdp, test_bit(SDF_SHUTDOWN, &sdp->sd_flags) \|\|
	!buffer_busy(bh));
	gfs_assert_warn(sdp, atomic_read(&bh->b_count) == 1);
	}

	/**
	* gfs_logbh_start - Start writing a fake buffer head
	* @sdp: the filesystem
	* @bh: the buffer to write
	*
	* This starts a block write to our journal.
	*/

	void
	gfs_logbh_start(struct gfs_sbd sdp, struct buffer_head bh)
	{
	submit_bh(WRITE, bh);
	}

	/**
	* gfs_logbh_wait - Wait for the write of a fake buffer head to complete
	* @sdp: the filesystem
	* @bh: the buffer to write
	*
	* This waits for a block write to our journal to complete.
	*
	* Returns: errno
	*/

	int
	gfs_logbh_wait(struct gfs_sbd sdp, struct buffer_head bh)
	{
	wait_on_buffer(bh);

	if (!buffer_uptodate(bh) \|\| buffer_dirty(bh)) {
	gfs_io_error_bh(sdp, bh);
	return -EIO;
	}
	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
	return -EIO;

	return 0;
	}

	/**
	* gfs_replay_buf - write a log buffer to its inplace location
	* @gl: the journal's glock
	* @bh: the buffer
	*
	* Returns: errno
	*/

	int
	gfs_replay_buf(struct gfs_glock gl, struct buffer_head bh)
	{
	struct gfs_sbd *sdp = gl->gl_sbd;
	struct gfs_bufdata *bd;

	bd = get_v2bd(bh);
	if (!bd) {
	gfs_attach_bufdata(bh, gl);
	bd = get_v2bd(bh);
	}

	mark_buffer_dirty(bh);

	if (list_empty(&bd->bd_ail_tr_list)) {
	get_bh(bh);
	list_add(&bd->bd_ail_tr_list, &sdp->sd_recovery_bufs);
	}

	return 0;
	}

	/**
	* gfs_replay_check - Check up on journal replay
	* @sdp: the filesystem
	*
	*/

	void
	gfs_replay_check(struct gfs_sbd *sdp)
	{
	struct buffer_head *bh;
	struct gfs_bufdata *bd;

	while (!list_empty(&sdp->sd_recovery_bufs)) {
	bd = list_entry(sdp->sd_recovery_bufs.prev,
	struct gfs_bufdata, bd_ail_tr_list);
	bh = bd->bd_bh;

	if (buffer_busy(bh)) {
	list_move(&bd->bd_ail_tr_list,
	&sdp->sd_recovery_bufs);
	break;
	} else {
	list_del_init(&bd->bd_ail_tr_list);
	if (!buffer_uptodate(bh))
	gfs_io_error_bh(sdp, bh);
	brelse(bh);
	}
	}
	}

	/**
	* gfs_replay_wait - Wait for all replayed buffers to hit the disk
	* @sdp: the filesystem
	*
	*/

	void
	gfs_replay_wait(struct gfs_sbd *sdp)
	{
	struct list_head head, tmp, *prev;
	struct buffer_head *bh;
	struct gfs_bufdata *bd;

	for (head = &sdp->sd_recovery_bufs, tmp = head->prev, prev = tmp->prev;
	tmp != head;
	tmp = prev, prev = tmp->prev) {
	bd = list_entry(tmp, struct gfs_bufdata, bd_ail_tr_list);
	bh = bd->bd_bh;

	if (!buffer_busy(bh)) {
	list_del_init(&bd->bd_ail_tr_list);
	if (!buffer_uptodate(bh))
	gfs_io_error_bh(sdp, bh);
	brelse(bh);
	continue;
	}

	if (buffer_dirty(bh)) {
	wait_on_buffer(bh);
	ll_rw_block(WRITE, 1, &bh);
	}
	}

	while (!list_empty(head)) {
	bd = list_entry(head->prev, struct gfs_bufdata, bd_ail_tr_list);
	bh = bd->bd_bh;

	wait_on_buffer(bh);

	gfs_assert_withdraw(sdp, !buffer_busy(bh));

	list_del_init(&bd->bd_ail_tr_list);
	if (!buffer_uptodate(bh))
	gfs_io_error_bh(sdp, bh);
	brelse(bh);
	}
	}

	/**
	* gfs_wipe_buffers - make inode's buffers so they aren't dirty/AILed anymore
	* @ip: the inode who owns the buffers
	* @rgd: the resource group
	* @bstart: the first buffer in the run
	* @blen: the number of buffers in the run
	*
	* Called when de-allocating a contiguous run of meta blocks within an rgrp.
	* Make sure all buffers for de-alloc'd blocks are removed from the AIL, if
	* they can be. Dirty or pinned blocks are left alone. Add relevant
	* meta-headers to meta-header cache, so we don't need to read disk
	* if we re-allocate blocks.
	*/

	void
	gfs_wipe_buffers(struct gfs_inode ip, struct gfs_rgrpd rgd,
	uint64_t bstart, uint32_t blen)
	{
	struct gfs_sbd *sdp = ip->i_sbd;
	struct inode *aspace = ip->i_gl->gl_aspace;
	struct buffer_head *bh;
	struct gfs_bufdata *bd;
	int busy;
	int add = FALSE;

	while (blen) {
	bh = getbuf(sdp, aspace, bstart, NO_CREATE);
	if (bh) {

	bd = get_v2bd(bh);

	if (buffer_uptodate(bh)) {
	if (bd) {
	gfs_lock_buffer(bh);
	gfs_mhc_add(rgd, &bh, 1);
	busy = bd->bd_pinned \|\| buffer_busy(bh);
	gfs_unlock_buffer(bh);

	if (busy)
	add = TRUE;
	else {
	spin_lock(&sdp->sd_ail_lock);
	if (!list_empty(&bd->bd_ail_tr_list)) {
	list_del_init(&bd->bd_ail_tr_list);
	list_del(&bd->bd_ail_gl_list);
	brelse(bh);
	}
	spin_unlock(&sdp->sd_ail_lock);
	}
	} else {
	gfs_assert_withdraw(sdp, !buffer_dirty(bh));
	wait_on_buffer(bh);
	gfs_assert_withdraw(sdp, !buffer_busy(bh));
	gfs_mhc_add(rgd, &bh, 1);
	}
	} else {
	gfs_assert_withdraw(sdp, !bd \|\| !bd->bd_pinned);
	gfs_assert_withdraw(sdp, !buffer_dirty(bh));
	wait_on_buffer(bh);
	gfs_assert_withdraw(sdp, !buffer_busy(bh));
	}

	brelse(bh);
	}

	bstart++;
	blen--;
	}

	if (add)
	gfs_depend_add(rgd, ip->i_num.no_formal_ino);
	}

	/**
	* gfs_sync_meta - sync all the buffers in a filesystem
	* @sdp: the filesystem
	*
	* Flush metadata blocks to on-disk journal, then
	* Flush metadata blocks (now in AIL) to on-disk in-place locations
	* Periodically keep checking until done (AIL empty)
	*/

	void
	gfs_sync_meta(struct gfs_sbd *sdp)
	{
	gfs_log_flush(sdp);
	for (;;) {
	gfs_ail_start(sdp, DIO_ALL);
	if (gfs_ail_empty(sdp))
	break;
	set_current_state(TASK_UNINTERRUPTIBLE);
	schedule_timeout(HZ / 10);
	}
	}

	/**
	* gfs_flush_meta_cache - get rid of any references on buffers for this inode
	* @ip: The GFS inode
	*
	* This releases buffers that are in the most-recently-used array of
	* blocks used for indirect block addressing for this inode.
	* Don't confuse this with the meta-HEADER cache (mhc)!
	*/

	void
	gfs_flush_meta_cache(struct gfs_inode *ip)
	{
	struct buffer_head **bh_slot;
	unsigned int x;

	spin_lock(&ip->i_spin);

	for (x = 0; x < GFS_MAX_META_HEIGHT; x++) {
	bh_slot = &ip->i_cache[x];
	if (*bh_slot) {
	brelse(*bh_slot);
	*bh_slot = NULL;
	}
	}

	spin_unlock(&ip->i_spin);
	}

	/**
	* gfs_get_meta_buffer - Get a metadata buffer
	* @ip: The GFS inode
	* @height: The level of this buf in the metadata (indir addr) tree (if any)
	* @num: The block number (device relative) of the buffer
	* @new: Non-zero if we may create a new buffer
	* @bhp: the buffer is returned here
	*
	* Returns: errno
	*/

	int
	gfs_get_meta_buffer(struct gfs_inode *ip, int height, uint64_t num, int new,
	struct buffer_head **bhp)
	{
	struct buffer_head bh, *bh_slot = &ip->i_cache[height];
	int flags = ((new) ? DIO_NEW : 0) \| DIO_START \| DIO_WAIT;
	int error;

	/* Try to use the gfs_inode's MRU metadata tree cache */
	spin_lock(&ip->i_spin);
	bh = *bh_slot;
	if (bh) {
	if (bh->b_blocknr == num)
	get_bh(bh);
	else
	bh = NULL;
	}
	spin_unlock(&ip->i_spin);

	if (bh) {
	error = gfs_dreread(ip->i_sbd, bh, flags);
	if (error) {
	brelse(bh);
	return error;
	}
	} else {
	error = gfs_dread(ip->i_gl, num, flags, &bh);
	if (error)
	return error;

	spin_lock(&ip->i_spin);
	if (*bh_slot != bh) {
	if (*bh_slot)
	brelse(*bh_slot);
	*bh_slot = bh;
	get_bh(bh);
	}
	spin_unlock(&ip->i_spin);
	}

	if (new) {
	if (gfs_assert_warn(ip->i_sbd, height)) {
	brelse(bh);
	return -EIO;
	}
	gfs_trans_add_bh(ip->i_gl, bh);
	gfs_metatype_set(bh, GFS_METATYPE_IN, GFS_FORMAT_IN);
	gfs_buffer_clear_tail(bh, sizeof(struct gfs_meta_header));
	} else if (gfs_metatype_check(ip->i_sbd, bh,
	(height) ? GFS_METATYPE_IN : GFS_METATYPE_DI)) {
	brelse(bh);
	return -EIO;
	}

	*bhp = bh;

	return 0;
	}

	/**
	* gfs_get_data_buffer - Get a data buffer
	* @ip: The GFS inode
	* @num: The block number (device relative) of the data block
	* @new: Non-zero if this is a new allocation
	* @bhp: the buffer is returned here
	*
	* Returns: errno
	*/

	int
	gfs_get_data_buffer(struct gfs_inode *ip, uint64_t block, int new,
	struct buffer_head **bhp)
	{
	struct buffer_head *bh;
	int error = 0;

	if (block == ip->i_num.no_addr) {
	if (gfs_assert_warn(ip->i_sbd, !new))
	return -EIO;
	error = gfs_dread(ip->i_gl, block, DIO_START \| DIO_WAIT, &bh);
	if (error)
	return error;
	if (gfs_metatype_check(ip->i_sbd, bh, GFS_METATYPE_DI)) {
	brelse(bh);
	return -EIO;
	}
	} else if (gfs_is_jdata(ip)) {
	if (new) {
	error = gfs_dread(ip->i_gl, block,
	DIO_NEW \| DIO_START \| DIO_WAIT, &bh);
	if (error)
	return error;
	gfs_trans_add_bh(ip->i_gl, bh);
	gfs_metatype_set(bh, GFS_METATYPE_JD, GFS_FORMAT_JD);
	gfs_buffer_clear_tail(bh, sizeof(struct gfs_meta_header));
	} else {
	error = gfs_dread(ip->i_gl, block,
	DIO_START \| DIO_WAIT, &bh);
	if (error)
	return error;
	if (gfs_metatype_check(ip->i_sbd, bh, GFS_METATYPE_JD)) {
	brelse(bh);
	return -EIO;
	}
	}
	} else {
	if (new) {
	bh = gfs_dgetblk(ip->i_gl, block);
	gfs_prep_new_buffer(bh);
	} else {
	error = gfs_dread(ip->i_gl, block,
	DIO_START \| DIO_WAIT, &bh);
	if (error)
	return error;
	}
	}

	*bhp = bh;

	return 0;
	}

	/**
	* gfs_start_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
	*
	*/

	void
	gfs_start_ra(struct gfs_glock *gl, uint64_t dblock, uint32_t extlen)
	{
	struct gfs_sbd *sdp = gl->gl_sbd;
	struct inode *aspace = gl->gl_aspace;
	struct buffer_head first_bh, bh;
	uint32_t max_ra = gfs_tune_get(sdp, gt_max_readahead) >> sdp->sd_sb.sb_bsize_shift;
	int error;

	if (!extlen)
	return;
	if (!max_ra)
	return;
	if (extlen > max_ra)
	extlen = max_ra;

	first_bh = getbuf(sdp, aspace, dblock, CREATE);

	if (buffer_uptodate(first_bh))
	goto out;
	if (!buffer_locked(first_bh)) {
	error = gfs_dreread(sdp, first_bh, DIO_START);
	if (error)
	goto out;
	}

	dblock++;
	extlen--;

	while (extlen) {
	bh = getbuf(sdp, aspace, dblock, CREATE);

	if (!buffer_uptodate(bh) && !buffer_locked(bh)) {
	error = gfs_dreread(sdp, bh, DIO_START);
	brelse(bh);
	if (error)
	goto out;
	} else
	brelse(bh);

	dblock++;
	extlen--;

	if (buffer_uptodate(first_bh))
	break;
	}

	out:
	brelse(first_bh);
	}

File Metadata

Mime Type: text/x-c
Expires: Thu, Feb 27, 4:50 AM (1 d, 13 h ago)
Storage Engine: blob
Storage Format: Raw Data
Storage Handle: 1466179
Default Alt Text: dio.c (30 KB)

dio.cNo OneTemporaryActions

dio.cView Options

File Metadata

Event Timeline

dio.c
No OneTemporary
Actions

dio.c
View Options