* any later version.
*/
+#include "config.h"
#ifdef HAVE_SYS_MOUNT_H
#include <sys/param.h>
#include <sys/mount.h>
*/
#undef USE_INODE_IO
+/* Checksumming functions */
+static int e2fsck_journal_verify_csum_type(journal_t *j,
+ journal_superblock_t *jsb)
+{
+ if (!JFS_HAS_INCOMPAT_FEATURE(j, JFS_FEATURE_INCOMPAT_CSUM_V2))
+ return 1;
+
+ return jsb->s_checksum_type == JBD2_CRC32C_CHKSUM;
+}
+
+static __u32 e2fsck_journal_sb_csum(journal_superblock_t *jsb)
+{
+ __u32 crc, old_crc;
+
+ old_crc = jsb->s_checksum;
+ jsb->s_checksum = 0;
+ crc = ext2fs_crc32c_le(~0, (unsigned char *)jsb,
+ sizeof(journal_superblock_t));
+ jsb->s_checksum = old_crc;
+
+ return crc;
+}
+
+static int e2fsck_journal_sb_csum_verify(journal_t *j,
+ journal_superblock_t *jsb)
+{
+ __u32 provided, calculated;
+
+ if (!JFS_HAS_INCOMPAT_FEATURE(j, JFS_FEATURE_INCOMPAT_CSUM_V2))
+ return 1;
+
+ provided = ext2fs_be32_to_cpu(jsb->s_checksum);
+ calculated = e2fsck_journal_sb_csum(jsb);
+
+ return provided == calculated;
+}
+
+static errcode_t e2fsck_journal_sb_csum_set(journal_t *j,
+ journal_superblock_t *jsb)
+{
+ __u32 crc;
+
+ if (!JFS_HAS_INCOMPAT_FEATURE(j, JFS_FEATURE_INCOMPAT_CSUM_V2))
+ return 0;
+
+ crc = e2fsck_journal_sb_csum(jsb);
+ jsb->s_checksum = ext2fs_cpu_to_be32(crc);
+ return 0;
+}
+
/* Kernel compatibility functions for handling the journal. These allow us
* to use the recovery.c file virtually unchanged from the kernel, so we
* don't have to do much to keep kernel and user recovery in sync.
*/
-int journal_bmap(journal_t *journal, blk_t block, unsigned long *phys)
+int journal_bmap(journal_t *journal, blk64_t block, unsigned long long *phys)
{
#ifdef USE_INODE_IO
*phys = block;
#else
struct inode *inode = journal->j_inode;
errcode_t retval;
- blk_t pblk;
+ blk64_t pblk;
if (!inode) {
*phys = block;
return 0;
}
- retval= ext2fs_bmap(inode->i_ctx->fs, inode->i_ino,
- &inode->i_ext2, NULL, 0, block, &pblk);
+ retval= ext2fs_bmap2(inode->i_ctx->fs, inode->i_ino,
+ &inode->i_ext2, NULL, 0, block, 0, &pblk);
*phys = pblk;
- return (retval);
+ return (int) retval;
#endif
}
-struct buffer_head *getblk(kdev_t kdev, blk_t blocknr, int blocksize)
+struct buffer_head *getblk(kdev_t kdev, blk64_t blocknr, int blocksize)
{
struct buffer_head *bh;
+ int bufsize = sizeof(*bh) + kdev->k_ctx->fs->blocksize -
+ sizeof(bh->b_data);
- bh = e2fsck_allocate_memory(kdev->k_ctx, sizeof(*bh), "block buffer");
+ bh = e2fsck_allocate_memory(kdev->k_ctx, bufsize, "block buffer");
if (!bh)
return NULL;
- jfs_debug(4, "getblk for block %lu (%d bytes)(total %d)\n",
- (unsigned long) blocknr, blocksize, ++bh_count);
+#ifdef CONFIG_JBD_DEBUG
+ if (journal_enable_debug >= 3)
+ bh_count++;
+#endif
+ jfs_debug(4, "getblk for block %llu (%d bytes)(total %d)\n",
+ (unsigned long long) blocknr, blocksize, bh_count);
bh->b_ctx = kdev->k_ctx;
if (kdev->k_dev == K_DEV_FS)
bh->b_io = kdev->k_ctx->fs->io;
- else
+ else
bh->b_io = kdev->k_ctx->journal_io;
bh->b_size = blocksize;
bh->b_blocknr = blocknr;
if (kdev->k_dev == K_DEV_FS)
io = kdev->k_ctx->fs->io;
- else
+ else
io = kdev->k_ctx->journal_io;
io_channel_flush(io);
void ll_rw_block(int rw, int nr, struct buffer_head *bhp[])
{
- int retval;
+ errcode_t retval;
struct buffer_head *bh;
for (; nr > 0; --nr) {
bh = *bhp++;
if (rw == READ && !bh->b_uptodate) {
- jfs_debug(3, "reading block %lu/%p\n",
- (unsigned long) bh->b_blocknr, (void *) bh);
- retval = io_channel_read_blk(bh->b_io,
+ jfs_debug(3, "reading block %llu/%p\n",
+ bh->b_blocknr, (void *) bh);
+ retval = io_channel_read_blk64(bh->b_io,
bh->b_blocknr,
1, bh->b_data);
if (retval) {
com_err(bh->b_ctx->device_name, retval,
- "while reading block %lu\n",
- (unsigned long) bh->b_blocknr);
- bh->b_err = retval;
+ "while reading block %llu\n",
+ bh->b_blocknr);
+ bh->b_err = (int) retval;
continue;
}
bh->b_uptodate = 1;
} else if (rw == WRITE && bh->b_dirty) {
- jfs_debug(3, "writing block %lu/%p\n",
- (unsigned long) bh->b_blocknr, (void *) bh);
- retval = io_channel_write_blk(bh->b_io,
+ jfs_debug(3, "writing block %llu/%p\n",
+ bh->b_blocknr,
+ (void *) bh);
+ retval = io_channel_write_blk64(bh->b_io,
bh->b_blocknr,
1, bh->b_data);
if (retval) {
com_err(bh->b_ctx->device_name, retval,
- "while writing block %lu\n",
- (unsigned long) bh->b_blocknr);
- bh->b_err = retval;
+ "while writing block %llu\n",
+ bh->b_blocknr);
+ bh->b_err = (int) retval;
continue;
}
bh->b_dirty = 0;
bh->b_uptodate = 1;
} else {
- jfs_debug(3, "no-op %s for block %lu\n",
- rw == READ ? "read" : "write",
- (unsigned long) bh->b_blocknr);
+ jfs_debug(3, "no-op %s for block %llu\n",
+ rw == READ ? "read" : "write",
+ bh->b_blocknr);
}
}
}
{
if (bh->b_dirty)
ll_rw_block(WRITE, 1, &bh);
- jfs_debug(3, "freeing block %lu/%p (total %d)\n",
- (unsigned long) bh->b_blocknr, (void *) bh, --bh_count);
+ jfs_debug(3, "freeing block %llu/%p (total %d)\n",
+ bh->b_blocknr, (void *) bh, --bh_count);
ext2fs_free_mem(&bh);
}
};
static int process_journal_block(ext2_filsys fs,
- blk_t *block_nr,
+ blk64_t *block_nr,
e2_blkcnt_t blockcnt,
- blk_t ref_block EXT2FS_ATTR((unused)),
+ blk64_t ref_block EXT2FS_ATTR((unused)),
int ref_offset EXT2FS_ATTR((unused)),
void *priv_data)
{
struct process_block_struct *p;
- blk_t blk = *block_nr;
+ blk64_t blk = *block_nr;
p = (struct process_block_struct *) priv_data;
- if (blk < fs->super->s_first_data_block ||
- blk >= fs->super->s_blocks_count)
+ if (!blk || blk < fs->super->s_first_data_block ||
+ blk >= ext2fs_blocks_count(fs->super))
return BLOCK_ABORT;
if (blockcnt >= 0)
journal_t *journal = NULL;
errcode_t retval = 0;
io_manager io_ptr = 0;
- unsigned long start = 0;
+ unsigned long long start = 0;
int ext_journal = 0;
int tried_backup_jnl = 0;
tried_backup_jnl)
goto errout;
memset(&j_inode->i_ext2, 0, sizeof(struct ext2_inode));
- memcpy(&j_inode->i_ext2.i_block[0], sb->s_jnl_blocks,
+ memcpy(&j_inode->i_ext2.i_block[0], sb->s_jnl_blocks,
EXT2_N_BLOCKS*4);
+ j_inode->i_ext2.i_size_high = sb->s_jnl_blocks[15];
j_inode->i_ext2.i_size = sb->s_jnl_blocks[16];
j_inode->i_ext2.i_links_count = 1;
j_inode->i_ext2.i_mode = LINUX_S_IFREG | 0600;
retval = EXT2_ET_NO_JOURNAL;
goto try_backup_journal;
}
- if (j_inode->i_ext2.i_size / journal->j_blocksize <
+ if (EXT2_I_SIZE(&j_inode->i_ext2) / journal->j_blocksize <
JFS_MIN_JOURNAL_BLOCKS) {
retval = EXT2_ET_JOURNAL_TOO_SMALL;
goto try_backup_journal;
}
pb.last_block = -1;
- retval = ext2fs_block_iterate2(ctx->fs, j_inode->i_ino,
- BLOCK_FLAG_HOLE, 0,
+ retval = ext2fs_block_iterate3(ctx->fs, j_inode->i_ino,
+ BLOCK_FLAG_HOLE, 0,
process_journal_block, &pb);
- if ((pb.last_block+1) * ctx->fs->blocksize <
- j_inode->i_ext2.i_size) {
+ if ((pb.last_block + 1) * ctx->fs->blocksize <
+ (int) EXT2_I_SIZE(&j_inode->i_ext2)) {
retval = EXT2_ET_JOURNAL_TOO_SMALL;
goto try_backup_journal;
}
goto errout;
}
- journal->j_maxlen = j_inode->i_ext2.i_size / journal->j_blocksize;
+ journal->j_maxlen = EXT2_I_SIZE(&j_inode->i_ext2) /
+ journal->j_blocksize;
#ifdef USE_INODE_IO
retval = ext2fs_inode_io_intern2(ctx->fs, sb->s_journal_inum,
#else
journal->j_inode = j_inode;
ctx->journal_io = ctx->fs->io;
- if ((retval = journal_bmap(journal, 0, &start)) != 0)
+ if ((retval = (errcode_t) journal_bmap(journal, 0, &start)) != 0)
goto errout;
#endif
} else {
#ifndef USE_INODE_IO
if (ext_journal)
#endif
- retval = io_ptr->open(journal_name, IO_FLAG_RW,
+ {
+ int flags = IO_FLAG_RW;
+ if (!(ctx->mount_flags & EXT2_MF_ISROOT &&
+ ctx->mount_flags & EXT2_MF_READONLY))
+ flags |= IO_FLAG_EXCLUSIVE;
+ if ((ctx->mount_flags & EXT2_MF_READONLY) &&
+ (ctx->options & E2F_OPT_FORCE))
+ flags &= ~IO_FLAG_EXCLUSIVE;
+
+
+ retval = io_ptr->open(journal_name, flags,
&ctx->journal_io);
+ }
if (retval)
goto errout;
io_channel_set_blksize(ctx->journal_io, ctx->fs->blocksize);
if (ext_journal) {
+ blk64_t maxlen;
+
if (ctx->fs->blocksize == 1024)
start = 1;
bh = getblk(dev_journal, start, ctx->fs->blocksize);
sizeof(jsuper));
brelse(bh);
#ifdef WORDS_BIGENDIAN
- if (jsuper.s_magic == ext2fs_swab16(EXT2_SUPER_MAGIC))
+ if (jsuper.s_magic == ext2fs_swab16(EXT2_SUPER_MAGIC))
ext2fs_swap_super(&jsuper);
#endif
if (jsuper.s_magic != EXT2_SUPER_MAGIC ||
goto errout;
}
- journal->j_maxlen = jsuper.s_blocks_count;
+ maxlen = ext2fs_blocks_count(&jsuper);
+ journal->j_maxlen = (maxlen < 1ULL << 32) ? maxlen : (1ULL << 32) - 1;
start++;
}
sb->s_feature_compat &= ~EXT3_FEATURE_COMPAT_HAS_JOURNAL;
sb->s_journal_inum = 0;
ctx->flags |= E2F_FLAG_JOURNAL_INODE;
+ ctx->fs->flags &= ~EXT2_FLAG_MASTER_SB_ONLY;
e2fsck_clear_recover(ctx, 1);
return 0;
}
jsb->s_nr_users)
clear_v2_journal_fields(journal);
break;
-
+
case JFS_SUPERBLOCK_V2:
journal->j_format_version = 2;
if (ntohl(jsb->s_nr_users) > 1 &&
case JFS_COMMIT_BLOCK:
case JFS_REVOKE_BLOCK:
return EXT2_ET_CORRUPT_SUPERBLOCK;
-
+
/* If we don't understand the superblock major type, but there
* is a magic number, then it is likely to be a new format we
* just don't understand, so leave it alone. */
if (JFS_HAS_INCOMPAT_FEATURE(journal, ~JFS_KNOWN_INCOMPAT_FEATURES))
return EXT2_ET_UNSUPP_FEATURE;
-
+
if (JFS_HAS_RO_COMPAT_FEATURE(journal, ~JFS_KNOWN_ROCOMPAT_FEATURES))
return EXT2_ET_RO_UNSUPP_FEATURE;
+ /* Checksum v1 and v2 are mutually exclusive features. */
+ if (JFS_HAS_INCOMPAT_FEATURE(journal, JFS_FEATURE_INCOMPAT_CSUM_V2) &&
+ JFS_HAS_COMPAT_FEATURE(journal, JFS_FEATURE_COMPAT_CHECKSUM))
+ return EXT2_ET_CORRUPT_SUPERBLOCK;
+
+ if (!e2fsck_journal_verify_csum_type(journal, jsb) ||
+ !e2fsck_journal_sb_csum_verify(journal, jsb))
+ return EXT2_ET_CORRUPT_SUPERBLOCK;
+
/* We have now checked whether we know enough about the journal
* format to be able to proceed safely, so any other checks that
* fail we should attempt to recover from. */
/* Leave a valid existing V1 superblock signature alone.
* Anything unrecognisable we overwrite with a new V2
* signature. */
-
+
if (jsb->s_header.h_magic != htonl(JFS_MAGIC_NUMBER) ||
jsb->s_header.h_blocktype != htonl(JFS_SUPERBLOCK_V1)) {
jsb->s_header.h_magic = htonl(JFS_MAGIC_NUMBER);
}
/* Zero out everything else beyond the superblock header */
-
+
p = ((char *) jsb) + sizeof(journal_header_t);
memset (p, 0, ctx->fs->blocksize-sizeof(journal_header_t));
for (i = 0; i < 4; i ++)
new_seq ^= u.val[i];
jsb->s_sequence = htonl(new_seq);
+ e2fsck_journal_sb_csum_set(journal, jsb);
mark_buffer_dirty(journal->j_sb_buffer);
ll_rw_block(WRITE, 1, &journal->j_sb_buffer);
jsb->s_sequence = htonl(journal->j_transaction_sequence);
if (reset)
jsb->s_start = 0; /* this marks the journal as empty */
+ e2fsck_journal_sb_csum_set(journal, jsb);
mark_buffer_dirty(journal->j_sb_buffer);
}
brelse(journal->j_sb_buffer);
io_channel_close(ctx->journal_io);
ctx->journal_io = 0;
}
-
+
#ifndef USE_INODE_IO
if (journal->j_inode)
ext2fs_free_mem(&journal->j_inode);
* This function makes sure that the superblock fields regarding the
* journal are consistent.
*/
-int e2fsck_check_ext3_journal(e2fsck_t ctx)
+errcode_t e2fsck_check_ext3_journal(e2fsck_t ctx)
{
struct ext2_super_block *sb = ctx->fs->super;
journal_t *journal;
struct problem_context pctx;
problem_t problem;
int reset = 0, force_fsck = 0;
- int retval;
+ errcode_t retval;
/* If we don't have any journal features, don't do anything more */
if (!(sb->s_feature_compat & EXT3_FEATURE_COMPAT_HAS_JOURNAL) &&
e2fsck_clear_recover(ctx, force_fsck);
} else if (!(ctx->options & E2F_OPT_READONLY)) {
sb->s_feature_compat |= EXT3_FEATURE_COMPAT_HAS_JOURNAL;
+ ctx->fs->flags &= ~EXT2_FLAG_MASTER_SB_ONLY;
ext2fs_mark_super_dirty(ctx->fs);
}
}
* ignore the fact that journal apparently has data;
* accidentally replaying over valid data would be far
* worse than skipping a questionable recovery.
- *
+ *
* XXX should we abort with a fatal error here? What
* will the ext3 kernel code do if a filesystem with
* !NEEDS_RECOVERY but with a non-zero
*/
}
+ /*
+ * If we don't need to do replay the journal, check to see if
+ * the journal's errno is set; if so, we need to mark the file
+ * system as being corrupt and clear the journal's s_errno.
+ */
+ if (!(sb->s_feature_incompat & EXT3_FEATURE_INCOMPAT_RECOVER) &&
+ journal->j_superblock->s_errno) {
+ ctx->fs->super->s_state |= EXT2_ERROR_FS;
+ ext2fs_mark_super_dirty(ctx->fs);
+ journal->j_superblock->s_errno = 0;
+ e2fsck_journal_sb_csum_set(journal, journal->j_superblock);
+ mark_buffer_dirty(journal->j_sb_buffer);
+ }
+
e2fsck_journal_release(ctx, journal, reset, 0);
return retval;
}
static errcode_t recover_ext3_journal(e2fsck_t ctx)
{
+ struct problem_context pctx;
journal_t *journal;
- int retval;
+ errcode_t retval;
+
+ clear_problem_context(&pctx);
journal_init_revoke_caches();
retval = e2fsck_get_journal(ctx, &journal);
retval = journal_init_revoke(journal, 1024);
if (retval)
goto errout;
-
+
retval = -journal_recover(journal);
if (retval)
goto errout;
-
- if (journal->j_superblock->s_errno) {
- ctx->fs->super->s_state |= EXT2_ERROR_FS;
- ext2fs_mark_super_dirty(ctx->fs);
- journal->j_superblock->s_errno = 0;
+
+ if (journal->j_failed_commit) {
+ pctx.ino = journal->j_failed_commit;
+ fix_problem(ctx, PR_0_JNL_TXN_CORRUPT, &pctx);
+ journal->j_superblock->s_errno = -EINVAL;
mark_buffer_dirty(journal->j_sb_buffer);
}
-
+
errout:
journal_destroy_revoke(journal);
journal_destroy_revoke_caches();
return retval;
}
-int e2fsck_run_ext3_journal(e2fsck_t ctx)
+errcode_t e2fsck_run_ext3_journal(e2fsck_t ctx)
{
io_manager io_ptr = ctx->fs->io->manager;
int blocksize = ctx->fs->blocksize;
errcode_t retval, recover_retval;
+ io_stats stats = 0;
+ unsigned long long kbytes_written = 0;
printf(_("%s: recovering journal\n"), ctx->device_name);
if (ctx->options & E2F_OPT_READONLY) {
ext2fs_flush(ctx->fs); /* Force out any modifications */
recover_retval = recover_ext3_journal(ctx);
-
+
/*
* Reload the filesystem context to get up-to-date data from disk
* because journal recovery will change the filesystem under us.
*/
- ext2fs_close(ctx->fs);
+ if (ctx->fs->super->s_kbytes_written &&
+ ctx->fs->io->manager->get_stats)
+ ctx->fs->io->manager->get_stats(ctx->fs->io, &stats);
+ if (stats && stats->bytes_written)
+ kbytes_written = stats->bytes_written >> 10;
+
+ ext2fs_mmp_stop(ctx->fs);
+ ext2fs_free(ctx->fs);
retval = ext2fs_open(ctx->filesystem_name, EXT2_FLAG_RW,
ctx->superblock, blocksize, io_ptr,
&ctx->fs);
-
if (retval) {
com_err(ctx->program_name, retval,
_("while trying to re-open %s"),
ctx->fs->priv_data = ctx;
ctx->fs->now = ctx->now;
ctx->fs->flags |= EXT2_FLAG_MASTER_SB_ONLY;
+ ctx->fs->super->s_kbytes_written += kbytes_written;
/* Set the superblock flags */
- e2fsck_clear_recover(ctx, recover_retval);
- return recover_retval;
+ e2fsck_clear_recover(ctx, recover_retval != 0);
+
+ /*
+ * Do one last sanity check, and propagate journal->s_errno to
+ * the EXT2_ERROR_FS flag in the fs superblock if needed.
+ */
+ retval = e2fsck_check_ext3_journal(ctx);
+ return retval ? retval : recover_retval;
}
/*
ext2_ino_t ino;
errcode_t retval;
const char * const * cpp;
- int group, mount_flags;
-
+ dgrp_t group;
+ int mount_flags;
+
clear_problem_context(&pctx);
/*
if (fix_problem(ctx, PR_0_BACKUP_JNL, &pctx)) {
memcpy(sb->s_jnl_blocks, inode.i_block,
EXT2_N_BLOCKS*4);
+ sb->s_jnl_blocks[15] = inode.i_size_high;
sb->s_jnl_blocks[16] = inode.i_size;
sb->s_jnl_backup_type = EXT3_JNL_BACKUP_BLOCKS;
ext2fs_mark_super_dirty(fs);
*/
if (sb->s_journal_inum == EXT2_JOURNAL_INO)
return;
-
+
/*
* The journal inode had better have only one link and not be readable.
*/
pctx.str = *cpp;
if (!fix_problem(ctx, PR_0_MOVE_JOURNAL, &pctx))
return;
-
+
/*
* OK, we've done all the checks, let's actually move the
* journal inode. Errors at this point mean we need to force
goto err_out;
group = ext2fs_group_of_ino(fs, ino);
- ext2fs_unmark_inode_bitmap(fs->inode_map, ino);
+ ext2fs_unmark_inode_bitmap2(fs->inode_map, ino);
ext2fs_mark_ib_dirty(fs);
- fs->group_desc[group].bg_free_inodes_count++;
+ ext2fs_bg_free_inodes_count_set(fs, group, ext2fs_bg_free_inodes_count(fs, group) + 1);
+ ext2fs_group_desc_csum_set(fs, group);
fs->super->s_free_inodes_count++;
return;
if (!journal_name)
return 0;
- if (stat(journal_name, &st) < 0)
+ if (stat(journal_name, &st) < 0) {
+ free(journal_name);
return 0;
+ }
if (st.st_rdev != sb->s_journal_dev) {
clear_problem_context(&pctx);