commit.c

linux 内核源代码

/*
 * linux/fs/jbd2/commit.c
 *
 * Written by Stephen C. Tweedie <[e-mail]>, 1998
 *
 * Copyright 1998 Red Hat corp --- All Rights Reserved
 *
 * This file is part of the Linux kernel and is made available under
 * the terms of the GNU General Public License, version 2, or at your
 * option, any later version, incorporated herein by reference.
 *
 * Journal commit routines for the generic filesystem journaling code;
 * part of the ext2fs journaling system.
 */

#include <linux/time.h>
#include <linux/fs.h>
#include <linux/jbd2.h>
#include <linux/errno.h>
#include <linux/slab.h>
#include <linux/mm.h>
#include <linux/pagemap.h>

/*
 * Default IO end handler for temporary BJ_IO buffer_heads.
 */
static void journal_end_buffer_io_sync(struct buffer_head *bh, int uptodate)
{
	BUFFER_TRACE(bh, "");
	if (uptodate)
		set_buffer_uptodate(bh);
	else
		clear_buffer_uptodate(bh);
	unlock_buffer(bh);
}

/*
 * When an ext3-ordered file is truncated, it is possible that many pages are
 * not sucessfully freed, because they are attached to a committing transaction.
 * After the transaction commits, these pages are left on the LRU, with no
 * ->mapping, and with attached buffers.  These pages are trivially reclaimable
 * by the VM, but their apparent absence upsets the VM accounting, and it makes
 * the numbers in /proc/meminfo look odd.
 *
 * So here, we have a buffer which has just come off the forget list.  Look to
 * see if we can strip all buffers from the backing page.
 *
 * Called under lock_journal(), and possibly under journal_datalist_lock.  The
 * caller provided us with a ref against the buffer, and we drop that here.
 */
static void release_buffer_page(struct buffer_head *bh)
{
	struct page *page;

	if (buffer_dirty(bh))
		goto nope;
	if (atomic_read(&bh->b_count) != 1)
		goto nope;
	page = bh->b_page;
	if (!page)
		goto nope;
	if (page->mapping)
		goto nope;

	/* OK, it's a truncated page */
	if (TestSetPageLocked(page))
		goto nope;

	page_cache_get(page);
	__brelse(bh);
	try_to_free_buffers(page);
	unlock_page(page);
	page_cache_release(page);
	return;

nope:
	__brelse(bh);
}

/*
 * Try to acquire jbd_lock_bh_state() against the buffer, when j_list_lock is
 * held.  For ranking reasons we must trylock.  If we lose, schedule away and
 * return 0.  j_list_lock is dropped in this case.
 */
static int inverted_lock(journal_t *journal, struct buffer_head *bh)
{
	if (!jbd_trylock_bh_state(bh)) {
		spin_unlock(&journal->j_list_lock);
		schedule();
		return 0;
	}
	return 1;
}

/* Done it all: now write the commit record.  We should have
 * cleaned up our previous buffers by now, so if we are in abort
 * mode we can now just skip the rest of the journal write
 * entirely.
 *
 * Returns 1 if the journal needs to be aborted or 0 on success
 */
static int journal_write_commit_record(journal_t *journal,
					transaction_t *commit_transaction)
{
	struct journal_head *descriptor;
	struct buffer_head *bh;
	int i, ret;
	int barrier_done = 0;

	if (is_journal_aborted(journal))
		return 0;

	descriptor = jbd2_journal_get_descriptor_buffer(journal);
	if (!descriptor)
		return 1;

	bh = jh2bh(descriptor);

	/* AKPM: buglet - add `i' to tmp! */
	for (i = 0; i < bh->b_size; i += 512) {
		journal_header_t *tmp = (journal_header_t*)bh->b_data;
		tmp->h_magic = cpu_to_be32(JBD2_MAGIC_NUMBER);
		tmp->h_blocktype = cpu_to_be32(JBD2_COMMIT_BLOCK);
		tmp->h_sequence = cpu_to_be32(commit_transaction->t_tid);
	}

	JBUFFER_TRACE(descriptor, "write commit block");
	set_buffer_dirty(bh);
	if (journal->j_flags & JBD2_BARRIER) {
		set_buffer_ordered(bh);
		barrier_done = 1;
	}
	ret = sync_dirty_buffer(bh);
	/* is it possible for another commit to fail at roughly
	 * the same time as this one?  If so, we don't want to
	 * trust the barrier flag in the super, but instead want
	 * to remember if we sent a barrier request
	 */
	if (ret == -EOPNOTSUPP && barrier_done) {
		char b[BDEVNAME_SIZE];

		printk(KERN_WARNING
			"JBD: barrier-based sync failed on %s - "
			"disabling barriers\n",
			bdevname(journal->j_dev, b));
		spin_lock(&journal->j_state_lock);
		journal->j_flags &= ~JBD2_BARRIER;
		spin_unlock(&journal->j_state_lock);

		/* And try again, without the barrier */
		clear_buffer_ordered(bh);
		set_buffer_uptodate(bh);
		set_buffer_dirty(bh);
		ret = sync_dirty_buffer(bh);
	}
	put_bh(bh);		/* One for getblk() */
	jbd2_journal_put_journal_head(descriptor);

	return (ret == -EIO);
}

static void journal_do_submit_data(struct buffer_head **wbuf, int bufs)
{
	int i;

	for (i = 0; i < bufs; i++) {
		wbuf[i]->b_end_io = end_buffer_write_sync;
		/* We use-up our safety reference in submit_bh() */
		submit_bh(WRITE, wbuf[i]);
	}
}

/*
 *  Submit all the data buffers to disk
 */
static void journal_submit_data_buffers(journal_t *journal,
				transaction_t *commit_transaction)
{
	struct journal_head *jh;
	struct buffer_head *bh;
	int locked;
	int bufs = 0;
	struct buffer_head **wbuf = journal->j_wbuf;

	/*
	 * Whenever we unlock the journal and sleep, things can get added
	 * onto ->t_sync_datalist, so we have to keep looping back to
	 * write_out_data until we *know* that the list is empty.
	 *
	 * Cleanup any flushed data buffers from the data list.  Even in
	 * abort mode, we want to flush this out as soon as possible.
	 */
write_out_data:
	cond_resched();
	spin_lock(&journal->j_list_lock);

	while (commit_transaction->t_sync_datalist) {
		jh = commit_transaction->t_sync_datalist;
		bh = jh2bh(jh);
		locked = 0;

		/* Get reference just to make sure buffer does not disappear
		 * when we are forced to drop various locks */
		get_bh(bh);
		/* If the buffer is dirty, we need to submit IO and hence
		 * we need the buffer lock. We try to lock the buffer without
		 * blocking. If we fail, we need to drop j_list_lock and do
		 * blocking lock_buffer().
		 */
		if (buffer_dirty(bh)) {
			if (test_set_buffer_locked(bh)) {
				BUFFER_TRACE(bh, "needs blocking lock");
				spin_unlock(&journal->j_list_lock);
				/* Write out all data to prevent deadlocks */
				journal_do_submit_data(wbuf, bufs);
				bufs = 0;
				lock_buffer(bh);
				spin_lock(&journal->j_list_lock);
			}
			locked = 1;
		}
		/* We have to get bh_state lock. Again out of order, sigh. */
		if (!inverted_lock(journal, bh)) {
			jbd_lock_bh_state(bh);
			spin_lock(&journal->j_list_lock);
		}
		/* Someone already cleaned up the buffer? */
		if (!buffer_jbd(bh)
			|| jh->b_transaction != commit_transaction
			|| jh->b_jlist != BJ_SyncData) {
			jbd_unlock_bh_state(bh);
			if (locked)
				unlock_buffer(bh);
			BUFFER_TRACE(bh, "already cleaned up");
			put_bh(bh);
			continue;
		}
		if (locked && test_clear_buffer_dirty(bh)) {
			BUFFER_TRACE(bh, "needs writeout, adding to array");
			wbuf[bufs++] = bh;
			__jbd2_journal_file_buffer(jh, commit_transaction,
						BJ_Locked);
			jbd_unlock_bh_state(bh);
			if (bufs == journal->j_wbufsize) {
				spin_unlock(&journal->j_list_lock);
				journal_do_submit_data(wbuf, bufs);
				bufs = 0;
				goto write_out_data;
			}
		} else if (!locked && buffer_locked(bh)) {
			__jbd2_journal_file_buffer(jh, commit_transaction,
						BJ_Locked);
			jbd_unlock_bh_state(bh);
			put_bh(bh);
		} else {
			BUFFER_TRACE(bh, "writeout complete: unfile");
			__jbd2_journal_unfile_buffer(jh);
			jbd_unlock_bh_state(bh);
			if (locked)
				unlock_buffer(bh);
			jbd2_journal_remove_journal_head(bh);
			/* Once for our safety reference, once for
			 * jbd2_journal_remove_journal_head() */
			put_bh(bh);
			put_bh(bh);
		}

		if (lock_need_resched(&journal->j_list_lock)) {
			spin_unlock(&journal->j_list_lock);
			goto write_out_data;
		}
	}
	spin_unlock(&journal->j_list_lock);
	journal_do_submit_data(wbuf, bufs);
}

static inline void write_tag_block(int tag_bytes, journal_block_tag_t *tag,
				   unsigned long long block)
{
	tag->t_blocknr = cpu_to_be32(block & (u32)~0);
	if (tag_bytes > JBD2_TAG_SIZE32)
		tag->t_blocknr_high = cpu_to_be32((block >> 31) >> 1);
}

/*
 * jbd2_journal_commit_transaction
 *
 * The primary function for committing a transaction to the log.  This
 * function is called by the journal thread to begin a complete commit.
 */
void jbd2_journal_commit_transaction(journal_t *journal)
{
	transaction_t *commit_transaction;
	struct journal_head *jh, *new_jh, *descriptor;
	struct buffer_head **wbuf = journal->j_wbuf;
	int bufs;
	int flags;
	int err;
	unsigned long long blocknr;
	char *tagp = NULL;
	journal_header_t *header;
	journal_block_tag_t *tag = NULL;
	int space_left = 0;
	int first_tag = 0;
	int tag_flag;
	int i;
	int tag_bytes = journal_tag_bytes(journal);

	/*
	 * First job: lock down the current transaction and wait for
	 * all outstanding updates to complete.
	 */

#ifdef COMMIT_STATS
	spin_lock(&journal->j_list_lock);
	summarise_journal_usage(journal);
	spin_unlock(&journal->j_list_lock);
#endif

	/* Do we need to erase the effects of a prior jbd2_journal_flush? */
	if (journal->j_flags & JBD2_FLUSHED) {
		jbd_debug(3, "super block updated\n");
		jbd2_journal_update_superblock(journal, 1);
	} else {
		jbd_debug(3, "superblock not updated\n");
	}

	J_ASSERT(journal->j_running_transaction != NULL);
	J_ASSERT(journal->j_committing_transaction == NULL);

	commit_transaction = journal->j_running_transaction;
	J_ASSERT(commit_transaction->t_state == T_RUNNING);

	jbd_debug(1, "JBD: starting commit of transaction %d\n",
			commit_transaction->t_tid);

	spin_lock(&journal->j_state_lock);
	commit_transaction->t_state = T_LOCKED;

	spin_lock(&commit_transaction->t_handle_lock);
	while (commit_transaction->t_updates) {
		DEFINE_WAIT(wait);

		prepare_to_wait(&journal->j_wait_updates, &wait,
					TASK_UNINTERRUPTIBLE);
		if (commit_transaction->t_updates) {
			spin_unlock(&commit_transaction->t_handle_lock);
			spin_unlock(&journal->j_state_lock);
			schedule();
			spin_lock(&journal->j_state_lock);
			spin_lock(&commit_transaction->t_handle_lock);
		}
		finish_wait(&journal->j_wait_updates, &wait);
	}
	spin_unlock(&commit_transaction->t_handle_lock);

	J_ASSERT (commit_transaction->t_outstanding_credits <=
			journal->j_max_transaction_buffers);

	/*
	 * First thing we are allowed to do is to discard any remaining
	 * BJ_Reserved buffers.  Note, it is _not_ permissible to assume
	 * that there are no such buffers: if a large filesystem
	 * operation like a truncate needs to split itself over multiple
	 * transactions, then it may try to do a jbd2_journal_restart() while
	 * there are still BJ_Reserved buffers outstanding.  These must
	 * be released cleanly from the current transaction.
	 *
	 * In this case, the filesystem must still reserve write access
	 * again before modifying the buffer in the new transaction, but
	 * we do not require it to remember exactly which old buffers it
	 * has reserved.  This is consistent with the existing behaviour
	 * that multiple jbd2_journal_get_write_access() calls to the same
	 * buffer are perfectly permissable.
	 */
	while (commit_transaction->t_reserved_list) {
		jh = commit_transaction->t_reserved_list;
		JBUFFER_TRACE(jh, "reserved, unused: refile");
		/*
		 * A jbd2_journal_get_undo_access()+jbd2_journal_release_buffer() may
		 * leave undo-committed data.
		 */
		if (jh->b_committed_data) {
			struct buffer_head *bh = jh2bh(jh);

			jbd_lock_bh_state(bh);
			jbd2_free(jh->b_committed_data, bh->b_size);
			jh->b_committed_data = NULL;
			jbd_unlock_bh_state(bh);
		}
		jbd2_journal_refile_buffer(journal, jh);
	}

	/*
	 * Now try to drop any written-back buffers from the journal's
	 * checkpoint lists.  We do this *before* commit because it potentially
	 * frees some memory
	 */
	spin_lock(&journal->j_list_lock);
	__jbd2_journal_clean_checkpoint_list(journal);
	spin_unlock(&journal->j_list_lock);

	jbd_debug (3, "JBD: commit phase 1\n");

	/*
	 * Switch to a new revoke table.
	 */
	jbd2_journal_switch_revoke_table(journal);

	commit_transaction->t_state = T_FLUSH;
	journal->j_committing_transaction = commit_transaction;
	journal->j_running_transaction = NULL;
	commit_transaction->t_log_start = journal->j_head;
	wake_up(&journal->j_wait_transaction_locked);
	spin_unlock(&journal->j_state_lock);

	jbd_debug (3, "JBD: commit phase 2\n");

	/*
	 * First, drop modified flag: all accesses to the buffers
	 * will be tracked for a new trasaction only -bzzz
	 */
	spin_lock(&journal->j_list_lock);
	if (commit_transaction->t_buffers) {
		new_jh = jh = commit_transaction->t_buffers->b_tnext;
		do {
			J_ASSERT_JH(new_jh, new_jh->b_modified == 1 ||
					new_jh->b_modified == 0);
			new_jh->b_modified = 0;
			new_jh = new_jh->b_tnext;
		} while (new_jh != jh);
	}
	spin_unlock(&journal->j_list_lock);

	/*
	 * Now start flushing things to disk, in the order they appear
	 * on the transaction lists.  Data blocks go first.
	 */
	err = 0;
	journal_submit_data_buffers(journal, commit_transaction);

	/*
	 * Wait for all previously submitted IO to complete.
	 */
	spin_lock(&journal->j_list_lock);
	while (commit_transaction->t_locked_list) {
		struct buffer_head *bh;

		jh = commit_transaction->t_locked_list->b_tprev;
		bh = jh2bh(jh);
		get_bh(bh);
		if (buffer_locked(bh)) {
			spin_unlock(&journal->j_list_lock);
			wait_on_buffer(bh);
			if (unlikely(!buffer_uptodate(bh)))
				err = -EIO;
			spin_lock(&journal->j_list_lock);
		}
		if (!inverted_lock(journal, bh)) {
			put_bh(bh);
			spin_lock(&journal->j_list_lock);
			continue;
		}