/*
- * linux/fs/jbd/recovery.c
+ * linux/fs/jbd2/recovery.c
*
* Written by Stephen C. Tweedie <sct@redhat.com>, 1999
*
#else
#include <linux/time.h>
#include <linux/fs.h>
-#include <linux/jbd.h>
+#include <linux/jbd2.h>
#include <linux/errno.h>
-#include <linux/slab.h>
+#include <linux/crc32.h>
+#include <linux/blkdev.h>
#endif
/*
{
int err;
unsigned int max, nbufs, next;
- unsigned long blocknr;
+ unsigned long long blocknr;
struct buffer_head *bh;
struct buffer_head * bufs[MAXBUF];
err = journal_bmap(journal, next, &blocknr);
if (err) {
- printk (KERN_ERR "JBD: bad block at offset %u\n",
+ printk(KERN_ERR "JBD2: bad block at offset %u\n",
next);
goto failed;
}
#endif /* __KERNEL__ */
+static inline __u32 get_be32(__be32 *p)
+{
+ unsigned char *cp = (unsigned char *) p;
+ __u32 ret;
+
+ ret = *cp++;
+ ret = (ret << 8) + *cp++;
+ ret = (ret << 8) + *cp++;
+ ret = (ret << 8) + *cp++;
+ return ret;
+}
+
+static inline __u16 get_be16(__be16 *p)
+{
+ unsigned char *cp = (unsigned char *) p;
+ __u16 ret;
+
+ ret = *cp++;
+ ret = (ret << 8) + *cp++;
+ return ret;
+}
/*
* Read a block from the journal
unsigned int offset)
{
int err;
- unsigned long blocknr;
+ unsigned long long blocknr;
struct buffer_head *bh;
*bhp = NULL;
if (offset >= journal->j_maxlen) {
- printk(KERN_ERR "JBD: corrupted journal superblock\n");
- return -EIO;
+ printk(KERN_ERR "JBD2: corrupted journal superblock\n");
+ return -EFSCORRUPTED;
}
err = journal_bmap(journal, offset, &blocknr);
if (err) {
- printk (KERN_ERR "JBD: bad block at offset %u\n",
+ printk(KERN_ERR "JBD2: bad block at offset %u\n",
offset);
return err;
}
}
if (!buffer_uptodate(bh)) {
- printk (KERN_ERR "JBD: Failed to read block at offset %u\n",
+ printk(KERN_ERR "JBD2: Failed to read block at offset %u\n",
offset);
brelse(bh);
return -EIO;
return 0;
}
+static int jbd2_descr_block_csum_verify(journal_t *j,
+ void *buf)
+{
+ struct journal_block_tail *tail;
+ __u32 provided;
+ __u32 calculated;
+
+ if (!journal_has_csum_v2or3(j))
+ return 1;
+
+ tail = (struct journal_block_tail *)((char *)buf + j->j_blocksize -
+ sizeof(struct journal_block_tail));
+ provided = tail->t_checksum;
+ tail->t_checksum = 0;
+ calculated = jbd2_chksum(j, j->j_csum_seed, buf, j->j_blocksize);
+ tail->t_checksum = provided;
+
+ return provided == ext2fs_cpu_to_be32(calculated);
+}
/*
* Count the number of in-use tags in a journal descriptor block.
int nr = 0, size = journal->j_blocksize;
int tag_bytes = journal_tag_bytes(journal);
+ if (journal_has_csum_v2or3(journal))
+ size -= sizeof(struct journal_block_tail);
+
tagp = &bh->b_data[sizeof(journal_header_t)];
while ((tagp - bh->b_data + tag_bytes) <= size) {
nr++;
tagp += tag_bytes;
- if (!(tag->t_flags & cpu_to_be32(JFS_FLAG_SAME_UUID)))
+ if (!(get_be16(&tag->t_flags) & JFS_FLAG_SAME_UUID))
tagp += 16;
- if (tag->t_flags & cpu_to_be32(JFS_FLAG_LAST_TAG))
+ if (get_be16(&tag->t_flags) & JFS_FLAG_LAST_TAG)
break;
}
*/
int journal_recover(journal_t *journal)
{
- int err;
+ int err, err2;
journal_superblock_t * sb;
struct recovery_info info;
if (!sb->s_start) {
jbd_debug(1, "No recovery required, last transaction %d\n",
- be32_to_cpu(sb->s_sequence));
- journal->j_transaction_sequence = be32_to_cpu(sb->s_sequence) + 1;
+ ext2fs_be32_to_cpu(sb->s_sequence));
+ journal->j_transaction_sequence = ext2fs_be32_to_cpu(sb->s_sequence) + 1;
return 0;
}
if (!err)
err = do_one_pass(journal, &info, PASS_REPLAY);
- jbd_debug(1, "JBD: recovery, exit status %d, "
+ jbd_debug(1, "JBD2: recovery, exit status %d, "
"recovered transactions %u to %u\n",
err, info.start_transaction, info.end_transaction);
- jbd_debug(1, "JBD: Replayed %d and revoked %d/%d blocks\n",
+ jbd_debug(1, "JBD2: Replayed %d and revoked %d/%d blocks\n",
info.nr_replays, info.nr_revoke_hits, info.nr_revokes);
/* Restart the log at the next transaction ID, thus invalidating
journal->j_transaction_sequence = ++info.end_transaction;
journal_clear_revoke(journal);
- sync_blockdev(journal->j_fs_dev);
+ err2 = sync_blockdev(journal->j_fs_dev);
+ if (!err)
+ err = err2;
+ /* Make sure all replayed data is on permanent storage */
+ if (journal->j_flags & JFS_BARRIER) {
+ err2 = blkdev_issue_flush(journal->j_fs_dev, GFP_KERNEL, NULL);
+ if (!err)
+ err = err2;
+ }
return err;
}
* Locate any valid recovery information from the journal and set up the
* journal structures in memory to ignore it (presumably because the
* caller has evidence that it is out of date).
- * This function does'nt appear to be exorted..
+ * This function doesn't appear to be exported..
*
* We perform one pass over the journal to allow us to tell the user how
* much recovery information is being erased, and to let us initialise
int journal_skip_recovery(journal_t *journal)
{
int err;
- journal_superblock_t * sb;
struct recovery_info info;
memset (&info, 0, sizeof(info));
- sb = journal->j_superblock;
err = do_one_pass(journal, &info, PASS_SCAN);
if (err) {
- printk(KERN_ERR "JBD: error %d scanning journal\n", err);
+ printk(KERN_ERR "JBD2: error %d scanning journal\n", err);
++journal->j_transaction_sequence;
} else {
-#ifdef CONFIG_JBD_DEBUG
- int dropped = info.end_transaction - be32_to_cpu(sb->s_sequence);
-#endif
+#ifdef CONFIG_JFS_DEBUG
+ int dropped = info.end_transaction -
+ ext2fs_be32_to_cpu(journal->j_superblock->s_sequence);
jbd_debug(1,
- "JBD: ignoring %d transaction%s from the journal.\n",
+ "JBD2: ignoring %d transaction%s from the journal.\n",
dropped, (dropped == 1) ? "" : "s");
+#endif
journal->j_transaction_sequence = ++info.end_transaction;
}
return err;
}
-#if 0
-static inline unsigned long long read_tag_block(int tag_bytes, journal_block_tag_t *tag)
+static inline unsigned long long read_tag_block(journal_t *journal,
+ journal_block_tag_t *tag)
{
- unsigned long long block = be32_to_cpu(tag->t_blocknr);
- if (tag_bytes > JBD_TAG_SIZE32)
- block |= (__u64)be32_to_cpu(tag->t_blocknr_high) << 32;
+ unsigned long long block = get_be32(&tag->t_blocknr);
+ if (jfs_has_feature_64bit(journal))
+ block |= (u64)get_be32(&tag->t_blocknr_high) << 32;
return block;
}
-#endif
/*
* calc_chksums calculates the checksums for the blocks described in the
wrap(journal, *next_log_block);
err = jread(&obh, journal, io_block);
if (err) {
- printk(KERN_ERR "JBD: IO error %d recovering block "
+ printk(KERN_ERR "JBD2: IO error %d recovering block "
"%lu in log\n", err, io_block);
return 1;
} else {
*crc32_sum = crc32_be(*crc32_sum, (void *)obh->b_data,
obh->b_size);
}
- brelse(obh);
+ put_bh(obh);
}
return 0;
}
+static int jbd2_commit_block_csum_verify(journal_t *j, void *buf)
+{
+ struct commit_header *h;
+ __u32 provided;
+ __u32 calculated;
+
+ if (!journal_has_csum_v2or3(j))
+ return 1;
+
+ h = buf;
+ provided = h->h_chksum[0];
+ h->h_chksum[0] = 0;
+ calculated = jbd2_chksum(j, j->j_csum_seed, buf, j->j_blocksize);
+ h->h_chksum[0] = provided;
+
+ return provided == ext2fs_cpu_to_be32(calculated);
+}
+
+static int jbd2_block_tag_csum_verify(journal_t *j, journal_block_tag_t *tag,
+ void *buf, __u32 sequence)
+{
+ journal_block_tag3_t *tag3 = (journal_block_tag3_t *)tag;
+ __u32 csum32;
+ __u32 seq;
+
+ if (!journal_has_csum_v2or3(j))
+ return 1;
+
+ seq = ext2fs_cpu_to_be32(sequence);
+ csum32 = jbd2_chksum(j, j->j_csum_seed, (__u8 *)&seq, sizeof(seq));
+ csum32 = jbd2_chksum(j, csum32, buf, j->j_blocksize);
+
+ if (jfs_has_feature_csum3(j))
+ return get_be32(&tag3->t_checksum) == csum32;
+
+ return get_be16(&tag->t_checksum) == (csum32 & 0xFFFF);
+}
+
static int do_one_pass(journal_t *journal,
struct recovery_info *info, enum passtype pass)
{
int blocktype;
int tag_bytes = journal_tag_bytes(journal);
__u32 crc32_sum = ~0; /* Transactional Checksums */
-
- /* Precompute the maximum metadata descriptors in a descriptor block */
- int MAX_BLOCKS_PER_DESC;
- MAX_BLOCKS_PER_DESC = ((journal->j_blocksize-sizeof(journal_header_t))
- / tag_bytes);
+ int descr_csum_size = 0;
+ int block_error = 0;
/*
* First thing is to establish what we expect to find in the log
*/
sb = journal->j_superblock;
- next_commit_ID = be32_to_cpu(sb->s_sequence);
- next_log_block = be32_to_cpu(sb->s_start);
+ next_commit_ID = ext2fs_be32_to_cpu(sb->s_sequence);
+ next_log_block = ext2fs_be32_to_cpu(sb->s_start);
first_commit_ID = next_commit_ID;
if (pass == PASS_SCAN)
* either the next descriptor block or the final commit
* record. */
- jbd_debug(3, "JBD: checking block %ld\n", next_log_block);
+ jbd_debug(3, "JBD2: checking block %ld\n", next_log_block);
err = jread(&bh, journal, next_log_block);
if (err)
goto failed;
tmp = (journal_header_t *)bh->b_data;
- if (tmp->h_magic != cpu_to_be32(JFS_MAGIC_NUMBER)) {
+ if (tmp->h_magic != ext2fs_cpu_to_be32(JFS_MAGIC_NUMBER)) {
brelse(bh);
break;
}
- blocktype = be32_to_cpu(tmp->h_blocktype);
- sequence = be32_to_cpu(tmp->h_sequence);
+ blocktype = ext2fs_be32_to_cpu(tmp->h_blocktype);
+ sequence = ext2fs_be32_to_cpu(tmp->h_sequence);
jbd_debug(3, "Found magic %d, sequence %d\n",
blocktype, sequence);
switch(blocktype) {
case JFS_DESCRIPTOR_BLOCK:
+ /* Verify checksum first */
+ if (journal_has_csum_v2or3(journal))
+ descr_csum_size =
+ sizeof(struct journal_block_tail);
+ if (descr_csum_size > 0 &&
+ !jbd2_descr_block_csum_verify(journal,
+ bh->b_data)) {
+ err = -EFSBADCRC;
+ brelse(bh);
+ goto failed;
+ }
+
/* If it is a valid descriptor block, replay it
* in pass REPLAY; if journal_checksums enabled, then
* calculate checksums in PASS_SCAN, otherwise,
* just skip over the blocks it describes. */
if (pass != PASS_REPLAY) {
if (pass == PASS_SCAN &&
- JFS_HAS_COMPAT_FEATURE(journal,
- JFS_FEATURE_COMPAT_CHECKSUM) &&
+ jfs_has_feature_checksum(journal) &&
!info->end_transaction) {
if (calc_chksums(journal, bh,
&next_log_block,
&crc32_sum)) {
- brelse(bh);
+ put_bh(bh);
break;
}
- brelse(bh);
+ put_bh(bh);
continue;
}
next_log_block += count_tags(journal, bh);
wrap(journal, next_log_block);
- brelse(bh);
+ put_bh(bh);
continue;
}
tagp = &bh->b_data[sizeof(journal_header_t)];
while ((tagp - bh->b_data + tag_bytes)
- <= journal->j_blocksize) {
+ <= journal->j_blocksize - descr_csum_size) {
unsigned long io_block;
tag = (journal_block_tag_t *) tagp;
- flags = be32_to_cpu(tag->t_flags);
+ flags = get_be16(&tag->t_flags);
io_block = next_log_block++;
wrap(journal, next_log_block);
/* Recover what we can, but
* report failure at the end. */
success = err;
- printk (KERN_ERR
- "JBD: IO error %d recovering "
+ printk(KERN_ERR
+ "JBD2: IO error %d recovering "
"block %ld in log\n",
err, io_block);
} else {
- unsigned long blocknr;
+ unsigned long long blocknr;
J_ASSERT(obh != NULL);
- blocknr = be32_to_cpu(tag->t_blocknr);
+ blocknr = read_tag_block(journal,
+ tag);
/* If the block has been
* revoked, then we're all done
goto skip_write;
}
+ /* Look for block corruption */
+ if (!jbd2_block_tag_csum_verify(
+ journal, tag, obh->b_data,
+ ext2fs_be32_to_cpu(tmp->h_sequence))) {
+ brelse(obh);
+ success = -EFSBADCRC;
+ printk(KERN_ERR "JBD2: Invalid "
+ "checksum recovering "
+ "block %llu in log\n",
+ blocknr);
+ block_error = 1;
+ goto skip_write;
+ }
+
/* Find a buffer for the new
* data being restored */
nbh = __getblk(journal->j_fs_dev,
journal->j_blocksize);
if (nbh == NULL) {
printk(KERN_ERR
- "JBD: Out of memory "
+ "JBD2: Out of memory "
"during recovery.\n");
err = -ENOMEM;
brelse(bh);
memcpy(nbh->b_data, obh->b_data,
journal->j_blocksize);
if (flags & JFS_FLAG_ESCAPE) {
- *((__be32 *)nbh->b_data) =
- cpu_to_be32(JFS_MAGIC_NUMBER);
+ __u32 magic = ext2fs_cpu_to_be32(JFS_MAGIC_NUMBER);
+ memcpy(nbh->b_data, &magic,
+ sizeof(magic));
}
BUFFER_TRACE(nbh, "marking dirty");
continue;
case JFS_COMMIT_BLOCK:
- jbd_debug(3, "Commit block for #%u found\n",
- next_commit_ID);
/* How to differentiate between interrupted commit
* and journal corruption ?
*
* | GO TO NEXT "Journal Corruption"
* | TRANSACTION
* |
- * {(n+1)th transanction}
+ * {(n+1)th transaction}
* |
* _______|______________
* | |
* much to do other than move on to the next sequence
* number. */
if (pass == PASS_SCAN &&
- JFS_HAS_COMPAT_FEATURE(journal,
- JFS_FEATURE_COMPAT_CHECKSUM)) {
+ jfs_has_feature_checksum(journal)) {
int chksum_err, chksum_seen;
struct commit_header *cbh =
(struct commit_header *)bh->b_data;
unsigned found_chksum =
- be32_to_cpu(cbh->h_chksum[0]);
+ ext2fs_be32_to_cpu(cbh->h_chksum[0]);
chksum_err = chksum_seen = 0;
- jbd_debug(3, "Checksums %x %x\n",
- crc32_sum, found_chksum);
if (info->end_transaction) {
journal->j_failed_commit =
info->end_transaction;
}
if (crc32_sum == found_chksum &&
- cbh->h_chksum_type == JBD2_CRC32_CHKSUM &&
+ cbh->h_chksum_type == JFS_CRC32_CHKSUM &&
cbh->h_chksum_size ==
- JBD2_CRC32_CHKSUM_SIZE)
+ JFS_CRC32_CHKSUM_SIZE)
chksum_seen = 1;
else if (!(cbh->h_chksum_type == 0 &&
cbh->h_chksum_size == 0 &&
if (chksum_err) {
info->end_transaction = next_commit_ID;
- jbd_debug(1, "Checksum_err %x %x\n",
- crc32_sum, found_chksum);
- if (!JFS_HAS_INCOMPAT_FEATURE(journal,
- JFS_FEATURE_INCOMPAT_ASYNC_COMMIT)){
+
+ if (!jfs_has_feature_async_commit(journal)){
journal->j_failed_commit =
next_commit_ID;
brelse(bh);
}
crc32_sum = ~0;
}
+ if (pass == PASS_SCAN &&
+ !jbd2_commit_block_csum_verify(journal,
+ bh->b_data)) {
+ info->end_transaction = next_commit_ID;
+
+ if (!jfs_has_feature_async_commit(journal)) {
+ journal->j_failed_commit =
+ next_commit_ID;
+ brelse(bh);
+ break;
+ }
+ }
brelse(bh);
next_commit_ID++;
continue;
/* It's really bad news if different passes end up at
* different places (but possible due to IO errors). */
if (info->end_transaction != next_commit_ID) {
- printk (KERN_ERR "JBD: recovery pass %d ended at "
+ printk(KERN_ERR "JBD2: recovery pass %d ended at "
"transaction %u, expected %u\n",
pass, next_commit_ID, info->end_transaction);
if (!success)
success = -EIO;
}
}
-
+ if (block_error && success == 0)
+ success = -EIO;
return success;
failed:
return err;
}
+static int jbd2_revoke_block_csum_verify(journal_t *j,
+ void *buf)
+{
+ struct journal_revoke_tail *tail;
+ __u32 provided;
+ __u32 calculated;
+
+ if (!journal_has_csum_v2or3(j))
+ return 1;
+
+ tail = (struct journal_revoke_tail *)((char *)buf + j->j_blocksize -
+ sizeof(struct journal_revoke_tail));
+ provided = tail->r_checksum;
+ tail->r_checksum = 0;
+ calculated = jbd2_chksum(j, j->j_csum_seed, buf, j->j_blocksize);
+ tail->r_checksum = provided;
+
+ return provided == ext2fs_cpu_to_be32(calculated);
+}
/* Scan a revoke record, marking all blocks mentioned as revoked. */
{
journal_revoke_header_t *header;
int offset, max;
+ unsigned csum_size = 0;
+ __u32 rcount;
+ int record_len = 4;
header = (journal_revoke_header_t *) bh->b_data;
offset = sizeof(journal_revoke_header_t);
- max = be32_to_cpu(header->r_count);
+ rcount = ext2fs_be32_to_cpu(header->r_count);
+
+ if (!jbd2_revoke_block_csum_verify(journal, header))
+ return -EFSBADCRC;
+
+ if (journal_has_csum_v2or3(journal))
+ csum_size = sizeof(struct journal_revoke_tail);
+ if (rcount > journal->j_blocksize - csum_size)
+ return -EINVAL;
+ max = rcount;
+
+ if (jfs_has_feature_64bit(journal))
+ record_len = 8;
- while (offset < max) {
- unsigned long blocknr;
+ while (offset + record_len <= max) {
+ unsigned long long blocknr;
int err;
- blocknr = be32_to_cpu(* ((__be32 *) (bh->b_data+offset)));
- offset += 4;
+ if (record_len == 4)
+ blocknr = ext2fs_be32_to_cpu(* ((__u32 *) (bh->b_data+offset)));
+ else
+ blocknr = ext2fs_be64_to_cpu(* ((__u64 *) (bh->b_data+offset)));
+ offset += record_len;
err = journal_set_revoke(journal, blocknr, sequence);
if (err)
return err;
git://git.whamcloud.com / tools / e2fsprogs.git / blobdiff