revoke.c

linux 内核源代码

/*
 * linux/fs/jbd2/revoke.c
 *
 * Written by Stephen C. Tweedie <[e-mail]>, 2000
 *
 * Copyright 2000 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 revoke routines for the generic filesystem journaling code;
 * part of the ext2fs journaling system.
 *
 * Revoke is the mechanism used to prevent old log records for deleted
 * metadata from being replayed on top of newer data using the same
 * blocks.  The revoke mechanism is used in two separate places:
 *
 * + Commit: during commit we write the entire list of the current
 *   transaction's revoked blocks to the journal
 *
 * + Recovery: during recovery we record the transaction ID of all
 *   revoked blocks.  If there are multiple revoke records in the log
 *   for a single block, only the last one counts, and if there is a log
 *   entry for a block beyond the last revoke, then that log entry still
 *   gets replayed.
 *
 * We can get interactions between revokes and new log data within a
 * single transaction:
 *
 * Block is revoked and then journaled:
 *   The desired end result is the journaling of the new block, so we
 *   cancel the revoke before the transaction commits.
 *
 * Block is journaled and then revoked:
 *   The revoke must take precedence over the write of the block, so we
 *   need either to cancel the journal entry or to write the revoke
 *   later in the log than the log block.  In this case, we choose the
 *   latter: journaling a block cancels any revoke record for that block
 *   in the current transaction, so any revoke for that block in the
 *   transaction must have happened after the block was journaled and so
 *   the revoke must take precedence.
 *
 * Block is revoked and then written as data:
 *   The data write is allowed to succeed, but the revoke is _not_
 *   cancelled.  We still need to prevent old log records from
 *   overwriting the new data.  We don't even need to clear the revoke
 *   bit here.
 *
 * Revoke information on buffers is a tri-state value:
 *
 * RevokeValid clear:	no cached revoke status, need to look it up
 * RevokeValid set, Revoked clear:
 *			buffer has not been revoked, and cancel_revoke
 *			need do nothing.
 * RevokeValid set, Revoked set:
 *			buffer has been revoked.
 */

#ifndef __KERNEL__
#include "jfs_user.h"
#else
#include <linux/time.h>
#include <linux/fs.h>
#include <linux/jbd2.h>
#include <linux/errno.h>
#include <linux/slab.h>
#include <linux/list.h>
#include <linux/init.h>
#endif
#include <linux/log2.h>

static struct kmem_cache *jbd2_revoke_record_cache;
static struct kmem_cache *jbd2_revoke_table_cache;

/* Each revoke record represents one single revoked block.  During
   journal replay, this involves recording the transaction ID of the
   last transaction to revoke this block. */

struct jbd2_revoke_record_s
{
	struct list_head  hash;
	tid_t		  sequence;	/* Used for recovery only */
	unsigned long long	  blocknr;
};


/* The revoke table is just a simple hash table of revoke records. */
struct jbd2_revoke_table_s
{
	/* It is conceivable that we might want a larger hash table
	 * for recovery.  Must be a power of two. */
	int		  hash_size;
	int		  hash_shift;
	struct list_head *hash_table;
};


#ifdef __KERNEL__
static void write_one_revoke_record(journal_t *, transaction_t *,
				    struct journal_head **, int *,
				    struct jbd2_revoke_record_s *);
static void flush_descriptor(journal_t *, struct journal_head *, int);
#endif

/* Utility functions to maintain the revoke table */

/* Borrowed from buffer.c: this is a tried and tested block hash function */
static inline int hash(journal_t *journal, unsigned long long block)
{
	struct jbd2_revoke_table_s *table = journal->j_revoke;
	int hash_shift = table->hash_shift;
	int hash = (int)block ^ (int)((block >> 31) >> 1);

	return ((hash << (hash_shift - 6)) ^
		(hash >> 13) ^
		(hash << (hash_shift - 12))) & (table->hash_size - 1);
}

static int insert_revoke_hash(journal_t *journal, unsigned long long blocknr,
			      tid_t seq)
{
	struct list_head *hash_list;
	struct jbd2_revoke_record_s *record;

repeat:
	record = kmem_cache_alloc(jbd2_revoke_record_cache, GFP_NOFS);
	if (!record)
		goto oom;

	record->sequence = seq;
	record->blocknr = blocknr;
	hash_list = &journal->j_revoke->hash_table[hash(journal, blocknr)];
	spin_lock(&journal->j_revoke_lock);
	list_add(&record->hash, hash_list);
	spin_unlock(&journal->j_revoke_lock);
	return 0;

oom:
	if (!journal_oom_retry)
		return -ENOMEM;
	jbd_debug(1, "ENOMEM in %s, retrying\n", __FUNCTION__);
	yield();
	goto repeat;
}

/* Find a revoke record in the journal's hash table. */

static struct jbd2_revoke_record_s *find_revoke_record(journal_t *journal,
						      unsigned long long blocknr)
{
	struct list_head *hash_list;
	struct jbd2_revoke_record_s *record;

	hash_list = &journal->j_revoke->hash_table[hash(journal, blocknr)];

	spin_lock(&journal->j_revoke_lock);
	record = (struct jbd2_revoke_record_s *) hash_list->next;
	while (&(record->hash) != hash_list) {
		if (record->blocknr == blocknr) {
			spin_unlock(&journal->j_revoke_lock);
			return record;
		}
		record = (struct jbd2_revoke_record_s *) record->hash.next;
	}
	spin_unlock(&journal->j_revoke_lock);
	return NULL;
}

int __init jbd2_journal_init_revoke_caches(void)
{
	jbd2_revoke_record_cache = kmem_cache_create("jbd2_revoke_record",
					   sizeof(struct jbd2_revoke_record_s),
					   0, SLAB_HWCACHE_ALIGN, NULL);
	if (jbd2_revoke_record_cache == 0)
		return -ENOMEM;

	jbd2_revoke_table_cache = kmem_cache_create("jbd2_revoke_table",
					   sizeof(struct jbd2_revoke_table_s),
					   0, 0, NULL);
	if (jbd2_revoke_table_cache == 0) {
		kmem_cache_destroy(jbd2_revoke_record_cache);
		jbd2_revoke_record_cache = NULL;
		return -ENOMEM;
	}
	return 0;
}

void jbd2_journal_destroy_revoke_caches(void)
{
	kmem_cache_destroy(jbd2_revoke_record_cache);
	jbd2_revoke_record_cache = NULL;
	kmem_cache_destroy(jbd2_revoke_table_cache);
	jbd2_revoke_table_cache = NULL;
}

/* Initialise the revoke table for a given journal to a given size. */

int jbd2_journal_init_revoke(journal_t *journal, int hash_size)
{
	int shift, tmp;

	J_ASSERT (journal->j_revoke_table[0] == NULL);

	shift = 0;
	tmp = hash_size;
	while((tmp >>= 1UL) != 0UL)
		shift++;

	journal->j_revoke_table[0] = kmem_cache_alloc(jbd2_revoke_table_cache, GFP_KERNEL);
	if (!journal->j_revoke_table[0])
		return -ENOMEM;
	journal->j_revoke = journal->j_revoke_table[0];

	/* Check that the hash_size is a power of two */
	J_ASSERT(is_power_of_2(hash_size));

	journal->j_revoke->hash_size = hash_size;

	journal->j_revoke->hash_shift = shift;

	journal->j_revoke->hash_table =
		kmalloc(hash_size * sizeof(struct list_head), GFP_KERNEL);
	if (!journal->j_revoke->hash_table) {
		kmem_cache_free(jbd2_revoke_table_cache, journal->j_revoke_table[0]);
		journal->j_revoke = NULL;
		return -ENOMEM;
	}

	for (tmp = 0; tmp < hash_size; tmp++)
		INIT_LIST_HEAD(&journal->j_revoke->hash_table[tmp]);

	journal->j_revoke_table[1] = kmem_cache_alloc(jbd2_revoke_table_cache, GFP_KERNEL);
	if (!journal->j_revoke_table[1]) {
		kfree(journal->j_revoke_table[0]->hash_table);
		kmem_cache_free(jbd2_revoke_table_cache, journal->j_revoke_table[0]);
		return -ENOMEM;
	}

	journal->j_revoke = journal->j_revoke_table[1];

	/* Check that the hash_size is a power of two */
	J_ASSERT(is_power_of_2(hash_size));

	journal->j_revoke->hash_size = hash_size;

	journal->j_revoke->hash_shift = shift;

	journal->j_revoke->hash_table =
		kmalloc(hash_size * sizeof(struct list_head), GFP_KERNEL);
	if (!journal->j_revoke->hash_table) {
		kfree(journal->j_revoke_table[0]->hash_table);
		kmem_cache_free(jbd2_revoke_table_cache, journal->j_revoke_table[0]);
		kmem_cache_free(jbd2_revoke_table_cache, journal->j_revoke_table[1]);
		journal->j_revoke = NULL;
		return -ENOMEM;
	}

	for (tmp = 0; tmp < hash_size; tmp++)
		INIT_LIST_HEAD(&journal->j_revoke->hash_table[tmp]);

	spin_lock_init(&journal->j_revoke_lock);

	return 0;
}

/* Destoy a journal's revoke table.  The table must already be empty! */

void jbd2_journal_destroy_revoke(journal_t *journal)
{
	struct jbd2_revoke_table_s *table;
	struct list_head *hash_list;
	int i;

	table = journal->j_revoke_table[0];
	if (!table)
		return;

	for (i=0; i<table->hash_size; i++) {
		hash_list = &table->hash_table[i];
		J_ASSERT (list_empty(hash_list));
	}

	kfree(table->hash_table);
	kmem_cache_free(jbd2_revoke_table_cache, table);
	journal->j_revoke = NULL;

	table = journal->j_revoke_table[1];
	if (!table)
		return;

	for (i=0; i<table->hash_size; i++) {
		hash_list = &table->hash_table[i];
		J_ASSERT (list_empty(hash_list));
	}

	kfree(table->hash_table);
	kmem_cache_free(jbd2_revoke_table_cache, table);
	journal->j_revoke = NULL;
}


#ifdef __KERNEL__

/*
 * jbd2_journal_revoke: revoke a given buffer_head from the journal.  This
 * prevents the block from being replayed during recovery if we take a
 * crash after this current transaction commits.  Any subsequent
 * metadata writes of the buffer in this transaction cancel the
 * revoke.
 *
 * Note that this call may block --- it is up to the caller to make
 * sure that there are no further calls to journal_write_metadata
 * before the revoke is complete.  In ext3, this implies calling the
 * revoke before clearing the block bitmap when we are deleting
 * metadata.
 *
 * Revoke performs a jbd2_journal_forget on any buffer_head passed in as a
 * parameter, but does _not_ forget the buffer_head if the bh was only
 * found implicitly.
 *
 * bh_in may not be a journalled buffer - it may have come off
 * the hash tables without an attached journal_head.
 *
 * If bh_in is non-zero, jbd2_journal_revoke() will decrement its b_count
 * by one.
 */

int jbd2_journal_revoke(handle_t *handle, unsigned long long blocknr,
		   struct buffer_head *bh_in)
{
	struct buffer_head *bh = NULL;
	journal_t *journal;
	struct block_device *bdev;
	int err;

	might_sleep();
	if (bh_in)
		BUFFER_TRACE(bh_in, "enter");

	journal = handle->h_transaction->t_journal;
	if (!jbd2_journal_set_features(journal, 0, 0, JBD2_FEATURE_INCOMPAT_REVOKE)){
		J_ASSERT (!"Cannot set revoke feature!");
		return -EINVAL;
	}

	bdev = journal->j_fs_dev;
	bh = bh_in;

	if (!bh) {
		bh = __find_get_block(bdev, blocknr, journal->j_blocksize);
		if (bh)
			BUFFER_TRACE(bh, "found on hash");
	}
#ifdef JBD2_EXPENSIVE_CHECKING
	else {
		struct buffer_head *bh2;