From a0375156ca1041574b5d47cc7e32f10b891151b0 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Fri, 11 Jun 2010 23:14:04 -0400 Subject: [PATCH 01/40] ext4: Clean up s_dirt handling We don't need to set s_dirt in most of the ext4 code when journaling is enabled. In ext3/4 some of the summary statistics for # of free inodes, blocks, and directories are calculated from the per-block group statistics when the file system is mounted or unmounted. As a result the superblock doesn't have to be updated, either via the journal or by setting s_dirt. There are a few exceptions, most notably when resizing the file system, where the superblock needs to be modified --- and in that case it should be done as a journalled operation if possible, and s_dirt set only in no-journal mode. This patch will optimize out some unneeded disk writes when using ext4 with a journal. Signed-off-by: "Theodore Ts'o" --- fs/ext4/balloc.c | 6 +----- fs/ext4/ext4.h | 6 ++++++ fs/ext4/ext4_jbd2.c | 16 ++++++++++++++++ fs/ext4/ext4_jbd2.h | 5 +++++ fs/ext4/file.c | 2 +- fs/ext4/ialloc.c | 4 ++-- fs/ext4/mballoc.c | 4 ++-- fs/ext4/resize.c | 6 ++---- fs/ext4/xattr.c | 3 +-- 9 files changed, 36 insertions(+), 16 deletions(-) diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c index 95b7594c76f..bd30799a43e 100644 --- a/fs/ext4/balloc.c +++ b/fs/ext4/balloc.c @@ -377,14 +377,11 @@ void ext4_add_groupblocks(handle_t *handle, struct super_block *sb, ext4_grpblk_t bit; unsigned int i; struct ext4_group_desc *desc; - struct ext4_super_block *es; - struct ext4_sb_info *sbi; + struct ext4_sb_info *sbi = EXT4_SB(sb); int err = 0, ret, blk_free_count; ext4_grpblk_t blocks_freed; struct ext4_group_info *grp; - sbi = EXT4_SB(sb); - es = sbi->s_es; ext4_debug("Adding block(s) %llu-%llu\n", block, block + count - 1); ext4_get_group_no_and_offset(sb, block, &block_group, &bit); @@ -477,7 +474,6 @@ void ext4_add_groupblocks(handle_t *handle, struct super_block *sb, ret = ext4_handle_dirty_metadata(handle, NULL, gd_bh); if (!err) err = ret; - sb->s_dirt = 1; error_return: brelse(bitmap_bh); diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 19a4de57128..8b56b53328e 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1860,6 +1860,12 @@ static inline void ext4_unlock_group(struct super_block *sb, spin_unlock(ext4_group_lock_ptr(sb, group)); } +static inline void ext4_mark_super_dirty(struct super_block *sb) +{ + if (EXT4_SB(sb)->s_journal == NULL) + sb->s_dirt =1; +} + /* * Inodes and files operations */ diff --git a/fs/ext4/ext4_jbd2.c b/fs/ext4/ext4_jbd2.c index 53d2764d71c..cfd27b38fa1 100644 --- a/fs/ext4/ext4_jbd2.c +++ b/fs/ext4/ext4_jbd2.c @@ -143,3 +143,19 @@ int __ext4_handle_dirty_metadata(const char *where, handle_t *handle, } return err; } + +int __ext4_handle_dirty_super(const char *where, handle_t *handle, + struct super_block *sb) +{ + struct buffer_head *bh = EXT4_SB(sb)->s_sbh; + int err = 0; + + if (ext4_handle_valid(handle)) { + err = jbd2_journal_dirty_metadata(handle, bh); + if (err) + ext4_journal_abort_handle(where, __func__, bh, + handle, err); + } else + sb->s_dirt = 1; + return err; +} diff --git a/fs/ext4/ext4_jbd2.h b/fs/ext4/ext4_jbd2.h index dade0c02479..8ae8168900b 100644 --- a/fs/ext4/ext4_jbd2.h +++ b/fs/ext4/ext4_jbd2.h @@ -141,6 +141,9 @@ int __ext4_journal_get_create_access(const char *where, int __ext4_handle_dirty_metadata(const char *where, handle_t *handle, struct inode *inode, struct buffer_head *bh); +int __ext4_handle_dirty_super(const char *where, handle_t *handle, + struct super_block *sb); + #define ext4_journal_get_undo_access(handle, bh) \ __ext4_journal_get_undo_access(__func__, (handle), (bh)) #define ext4_journal_get_write_access(handle, bh) \ @@ -152,6 +155,8 @@ int __ext4_handle_dirty_metadata(const char *where, handle_t *handle, __ext4_journal_get_create_access(__func__, (handle), (bh)) #define ext4_handle_dirty_metadata(handle, inode, bh) \ __ext4_handle_dirty_metadata(__func__, (handle), (inode), (bh)) +#define ext4_handle_dirty_super(handle, sb) \ + __ext4_handle_dirty_super(__func__, (handle), (sb)) handle_t *ext4_journal_start_sb(struct super_block *sb, int nblocks); int __ext4_journal_stop(const char *where, handle_t *handle); diff --git a/fs/ext4/file.c b/fs/ext4/file.c index 5313ae4cda2..bd411c12d63 100644 --- a/fs/ext4/file.c +++ b/fs/ext4/file.c @@ -123,7 +123,7 @@ static int ext4_file_open(struct inode * inode, struct file * filp) if (!IS_ERR(cp)) { memcpy(sbi->s_es->s_last_mounted, cp, sizeof(sbi->s_es->s_last_mounted)); - sb->s_dirt = 1; + ext4_mark_super_dirty(sb); } } return dquot_file_open(inode, filp); diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c index 25c4b3173fd..ac377505ed5 100644 --- a/fs/ext4/ialloc.c +++ b/fs/ext4/ialloc.c @@ -279,7 +279,7 @@ out: err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh); if (!fatal) fatal = err; - sb->s_dirt = 1; + ext4_mark_super_dirty(sb); } else ext4_error(sb, "bit already cleared for inode %lu", ino); @@ -965,7 +965,7 @@ got: percpu_counter_dec(&sbi->s_freeinodes_counter); if (S_ISDIR(mode)) percpu_counter_inc(&sbi->s_dirs_counter); - sb->s_dirt = 1; + ext4_mark_super_dirty(sb); if (sbi->s_log_groups_per_flex) { flex_group = ext4_flex_group(sbi, group); diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index 12b3bc026a6..d9d267181dd 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -2812,7 +2812,7 @@ ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac, err = ext4_handle_dirty_metadata(handle, NULL, gdp_bh); out_err: - sb->s_dirt = 1; + ext4_mark_super_dirty(sb); brelse(bitmap_bh); return err; } @@ -4680,7 +4680,7 @@ do_more: put_bh(bitmap_bh); goto do_more; } - sb->s_dirt = 1; + ext4_mark_super_dirty(sb); error_return: if (freed) dquot_free_block(inode, freed); diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c index 6df797eb9ae..27527ae466e 100644 --- a/fs/ext4/resize.c +++ b/fs/ext4/resize.c @@ -921,8 +921,7 @@ int ext4_group_add(struct super_block *sb, struct ext4_new_group_data *input) &sbi->s_flex_groups[flex_group].free_inodes); } - ext4_handle_dirty_metadata(handle, NULL, sbi->s_sbh); - sb->s_dirt = 1; + ext4_handle_dirty_super(handle, sb); exit_journal: mutex_unlock(&sbi->s_resize_lock); @@ -1045,13 +1044,12 @@ int ext4_group_extend(struct super_block *sb, struct ext4_super_block *es, goto exit_put; } ext4_blocks_count_set(es, o_blocks_count + add); - ext4_handle_dirty_metadata(handle, NULL, EXT4_SB(sb)->s_sbh); - sb->s_dirt = 1; mutex_unlock(&EXT4_SB(sb)->s_resize_lock); ext4_debug("freeing blocks %llu through %llu\n", o_blocks_count, o_blocks_count + add); /* We add the blocks to the bitmap and set the group need init bit */ ext4_add_groupblocks(handle, sb, o_blocks_count, add); + ext4_handle_dirty_super(handle, sb); ext4_debug("freed blocks %llu through %llu\n", o_blocks_count, o_blocks_count + add); if ((err = ext4_journal_stop(handle))) diff --git a/fs/ext4/xattr.c b/fs/ext4/xattr.c index 04338009793..a6f31424957 100644 --- a/fs/ext4/xattr.c +++ b/fs/ext4/xattr.c @@ -458,8 +458,7 @@ static void ext4_xattr_update_super_block(handle_t *handle, if (ext4_journal_get_write_access(handle, EXT4_SB(sb)->s_sbh) == 0) { EXT4_SET_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_EXT_ATTR); - sb->s_dirt = 1; - ext4_handle_dirty_metadata(handle, NULL, EXT4_SB(sb)->s_sbh); + ext4_handle_dirty_super(handle, sb); } } From 07a038245b28df9196ffb2e8cc626e9b956a4e23 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Mon, 14 Jun 2010 09:54:48 -0400 Subject: [PATCH 02/40] ext4: Convert more i_flags references to use accessor functions These changes are not ones which are likely to result in races, but they should be fixed. Signed-off-by: "Theodore Ts'o" --- fs/ext4/dir.c | 3 ++- fs/ext4/inode.c | 4 ++-- fs/ext4/migrate.c | 2 +- 3 files changed, 5 insertions(+), 4 deletions(-) diff --git a/fs/ext4/dir.c b/fs/ext4/dir.c index ea5e6cb7e2a..2965c39d418 100644 --- a/fs/ext4/dir.c +++ b/fs/ext4/dir.c @@ -121,7 +121,8 @@ static int ext4_readdir(struct file *filp, * We don't set the inode dirty flag since it's not * critical that it get flushed back to the disk. */ - ext4_clear_inode_flag(filp->f_path.dentry->d_inode, EXT4_INODE_INDEX); + ext4_clear_inode_flag(filp->f_path.dentry->d_inode, + EXT4_INODE_INDEX); } stored = 0; offset = filp->f_pos & (sb->s_blocksize - 1); diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 42272d67955..6c6614cae9e 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -4976,7 +4976,7 @@ static blkcnt_t ext4_inode_blocks(struct ext4_inode *raw_inode, /* we are using combined 48 bit field */ i_blocks = ((u64)le16_to_cpu(raw_inode->i_blocks_high)) << 32 | le32_to_cpu(raw_inode->i_blocks_lo); - if (ei->i_flags & EXT4_HUGE_FILE_FL) { + if (ext4_test_inode_flag(inode, EXT4_INODE_HUGE_FILE)) { /* i_blocks represent file system block size */ return i_blocks << (inode->i_blkbits - 9); } else { @@ -5126,7 +5126,7 @@ struct inode *ext4_iget(struct super_block *sb, unsigned long ino) ei->i_file_acl); ret = -EIO; goto bad_inode; - } else if (ei->i_flags & EXT4_EXTENTS_FL) { + } else if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) { if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) || (S_ISLNK(inode->i_mode) && !ext4_inode_is_fast_symlink(inode))) diff --git a/fs/ext4/migrate.c b/fs/ext4/migrate.c index 6f3a27ec30b..1765c2c50a9 100644 --- a/fs/ext4/migrate.c +++ b/fs/ext4/migrate.c @@ -376,7 +376,7 @@ static int ext4_ext_swap_inode_data(handle_t *handle, struct inode *inode, * We have the extent map build with the tmp inode. * Now copy the i_data across */ - ei->i_flags |= EXT4_EXTENTS_FL; + ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS); memcpy(ei->i_data, tmp_ei->i_data, sizeof(ei->i_data)); /* From 5a0790c2c4a18435759a70e1562450035d778339 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Mon, 14 Jun 2010 13:28:03 -0400 Subject: [PATCH 03/40] ext4: remove initialized but not read variables No real bugs found, just removed some dead code. Found by gcc 4.6's new warnings. Signed-off-by: Andi Kleen Signed-off-by: "Theodore Ts'o" --- fs/ext4/extents.c | 8 ++------ fs/ext4/inode.c | 5 ----- fs/ext4/mballoc.c | 21 ++++++--------------- fs/ext4/namei.c | 2 -- fs/ext4/resize.c | 2 -- fs/jbd2/journal.c | 12 ------------ fs/jbd2/recovery.c | 10 ++-------- 7 files changed, 10 insertions(+), 50 deletions(-) diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index 377309c1af6..346de3daab7 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -1083,7 +1083,6 @@ static int ext4_ext_grow_indepth(handle_t *handle, struct inode *inode, { struct ext4_ext_path *curp = path; struct ext4_extent_header *neh; - struct ext4_extent_idx *fidx; struct buffer_head *bh; ext4_fsblk_t newblock; int err = 0; @@ -1144,10 +1143,10 @@ static int ext4_ext_grow_indepth(handle_t *handle, struct inode *inode, ext4_idx_store_pblock(curp->p_idx, newblock); neh = ext_inode_hdr(inode); - fidx = EXT_FIRST_INDEX(neh); ext_debug("new root: num %d(%d), lblock %d, ptr %llu\n", le16_to_cpu(neh->eh_entries), le16_to_cpu(neh->eh_max), - le32_to_cpu(fidx->ei_block), idx_pblock(fidx)); + le32_to_cpu(EXT_FIRST_INDEX(neh)->ei_block), + idx_pblock(EXT_FIRST_INDEX(neh))); neh->eh_depth = cpu_to_le16(path->p_depth + 1); err = ext4_ext_dirty(handle, inode, curp); @@ -2954,7 +2953,6 @@ static int ext4_split_unwritten_extents(handle_t *handle, struct ext4_extent *ex1 = NULL; struct ext4_extent *ex2 = NULL; struct ext4_extent *ex3 = NULL; - struct ext4_extent_header *eh; ext4_lblk_t ee_block, eof_block; unsigned int allocated, ee_len, depth; ext4_fsblk_t newblock; @@ -2971,7 +2969,6 @@ static int ext4_split_unwritten_extents(handle_t *handle, eof_block = map->m_lblk + map->m_len; depth = ext_depth(inode); - eh = path[depth].p_hdr; ex = path[depth].p_ext; ee_block = le32_to_cpu(ex->ee_block); ee_len = ext4_ext_get_actual_len(ex); @@ -3058,7 +3055,6 @@ static int ext4_split_unwritten_extents(handle_t *handle, err = PTR_ERR(path); goto out; } - eh = path[depth].p_hdr; ex = path[depth].p_ext; if (ex2 != &newex) ex2 = ex; diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 6c6614cae9e..15ff8d8d839 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -3146,13 +3146,10 @@ static int ext4_da_write_begin(struct file *file, struct address_space *mapping, int ret, retries = 0; struct page *page; pgoff_t index; - unsigned from, to; struct inode *inode = mapping->host; handle_t *handle; index = pos >> PAGE_CACHE_SHIFT; - from = pos & (PAGE_CACHE_SIZE - 1); - to = from + len; if (ext4_nonda_switch(inode->i_sb)) { *fsdata = (void *)FALL_BACK_TO_NONDELALLOC; @@ -5754,7 +5751,6 @@ static int ext4_expand_extra_isize(struct inode *inode, { struct ext4_inode *raw_inode; struct ext4_xattr_ibody_header *header; - struct ext4_xattr_entry *entry; if (EXT4_I(inode)->i_extra_isize >= new_extra_isize) return 0; @@ -5762,7 +5758,6 @@ static int ext4_expand_extra_isize(struct inode *inode, raw_inode = ext4_raw_inode(&iloc); header = IHDR(inode, raw_inode); - entry = IFIRST(header); /* No extended attributes present */ if (!ext4_test_inode_state(inode, EXT4_STATE_XATTR) || diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index d9d267181dd..b2948b04797 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -1999,7 +1999,6 @@ ext4_mb_regular_allocator(struct ext4_allocation_context *ac) ext4_group_t ngroups, group, i; int cr; int err = 0; - int bsbits; struct ext4_sb_info *sbi; struct super_block *sb; struct ext4_buddy e4b; @@ -2041,8 +2040,6 @@ ext4_mb_regular_allocator(struct ext4_allocation_context *ac) ac->ac_2order = i - 1; } - bsbits = ac->ac_sb->s_blocksize_bits; - /* if stream allocation is enabled, use global goal */ if (ac->ac_flags & EXT4_MB_STREAM_ALLOC) { /* TBD: may be hot point */ @@ -2712,7 +2709,6 @@ ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac, handle_t *handle, unsigned int reserv_blks) { struct buffer_head *bitmap_bh = NULL; - struct ext4_super_block *es; struct ext4_group_desc *gdp; struct buffer_head *gdp_bh; struct ext4_sb_info *sbi; @@ -2725,8 +2721,6 @@ ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac, sb = ac->ac_sb; sbi = EXT4_SB(sb); - es = sbi->s_es; - err = -EIO; bitmap_bh = ext4_read_block_bitmap(sb, ac->ac_b_ex.fe_group); @@ -2850,7 +2844,7 @@ ext4_mb_normalize_request(struct ext4_allocation_context *ac, int bsbits, max; ext4_lblk_t end; loff_t size, orig_size, start_off; - ext4_lblk_t start, orig_start; + ext4_lblk_t start; struct ext4_inode_info *ei = EXT4_I(ac->ac_inode); struct ext4_prealloc_space *pa; @@ -2881,6 +2875,7 @@ ext4_mb_normalize_request(struct ext4_allocation_context *ac, size = size << bsbits; if (size < i_size_read(ac->ac_inode)) size = i_size_read(ac->ac_inode); + orig_size = size; /* max size of free chunks */ max = 2 << bsbits; @@ -2922,8 +2917,8 @@ ext4_mb_normalize_request(struct ext4_allocation_context *ac, start_off = (loff_t)ac->ac_o_ex.fe_logical << bsbits; size = ac->ac_o_ex.fe_len << bsbits; } - orig_size = size = size >> bsbits; - orig_start = start = start_off >> bsbits; + size = size >> bsbits; + start = start_off >> bsbits; /* don't cover already allocated blocks in selected range */ if (ar->pleft && start <= ar->lleft) { @@ -3547,7 +3542,6 @@ ext4_mb_release_inode_pa(struct ext4_buddy *e4b, struct buffer_head *bitmap_bh, ext4_group_t group; ext4_grpblk_t bit; unsigned long long grp_blk_start; - sector_t start; int err = 0; int free = 0; @@ -3567,10 +3561,9 @@ ext4_mb_release_inode_pa(struct ext4_buddy *e4b, struct buffer_head *bitmap_bh, if (bit >= end) break; next = mb_find_next_bit(bitmap_bh->b_data, end, bit); - start = ext4_group_first_block_no(sb, group) + bit; mb_debug(1, " free preallocated %u/%u in group %u\n", - (unsigned) start, (unsigned) next - bit, - (unsigned) group); + (unsigned) ext4_group_first_block_no(sb, group) + bit, + (unsigned) next - bit, (unsigned) group); free += next - bit; if (ac) { @@ -4494,7 +4487,6 @@ void ext4_free_blocks(handle_t *handle, struct inode *inode, struct super_block *sb = inode->i_sb; struct ext4_allocation_context *ac = NULL; struct ext4_group_desc *gdp; - struct ext4_super_block *es; unsigned long freed = 0; unsigned int overflow; ext4_grpblk_t bit; @@ -4513,7 +4505,6 @@ void ext4_free_blocks(handle_t *handle, struct inode *inode, } sbi = EXT4_SB(sb); - es = EXT4_SB(sb)->s_es; if (!(flags & EXT4_FREE_BLOCKS_VALIDATED) && !ext4_data_block_valid(sbi, block, count)) { ext4_error(sb, "Freeing blocks not in datazone - " diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c index a43e6617b35..5a61f77e7d7 100644 --- a/fs/ext4/namei.c +++ b/fs/ext4/namei.c @@ -1088,7 +1088,6 @@ static struct dentry *ext4_lookup(struct inode *dir, struct dentry *dentry, stru struct dentry *ext4_get_parent(struct dentry *child) { __u32 ino; - struct inode *inode; static const struct qstr dotdot = { .name = "..", .len = 2, @@ -1097,7 +1096,6 @@ struct dentry *ext4_get_parent(struct dentry *child) struct buffer_head *bh; bh = ext4_find_entry(child->d_inode, &dotdot, &de); - inode = NULL; if (!bh) return ERR_PTR(-ENOENT); ino = le32_to_cpu(de->inode); diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c index 27527ae466e..ca5c8aa00a2 100644 --- a/fs/ext4/resize.c +++ b/fs/ext4/resize.c @@ -952,7 +952,6 @@ int ext4_group_extend(struct super_block *sb, struct ext4_super_block *es, ext4_fsblk_t n_blocks_count) { ext4_fsblk_t o_blocks_count; - ext4_group_t o_groups_count; ext4_grpblk_t last; ext4_grpblk_t add; struct buffer_head *bh; @@ -964,7 +963,6 @@ int ext4_group_extend(struct super_block *sb, struct ext4_super_block *es, * yet: we're going to revalidate es->s_blocks_count after * taking the s_resize_lock below. */ o_blocks_count = ext4_blocks_count(es); - o_groups_count = EXT4_SB(sb)->s_groups_count; if (test_opt(sb, DEBUG)) printk(KERN_DEBUG "EXT4-fs: extending last group from %llu uto %llu blocks\n", diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c index bc2ff593276..f7bf15787d6 100644 --- a/fs/jbd2/journal.c +++ b/fs/jbd2/journal.c @@ -1392,13 +1392,9 @@ int jbd2_journal_check_used_features (journal_t *journal, unsigned long compat, int jbd2_journal_check_available_features (journal_t *journal, unsigned long compat, unsigned long ro, unsigned long incompat) { - journal_superblock_t *sb; - if (!compat && !ro && !incompat) return 1; - sb = journal->j_superblock; - /* We can support any known requested features iff the * superblock is in version 2. Otherwise we fail to support any * extended sb features. */ @@ -1618,7 +1614,6 @@ int jbd2_journal_flush(journal_t *journal) int jbd2_journal_wipe(journal_t *journal, int write) { - journal_superblock_t *sb; int err = 0; J_ASSERT (!(journal->j_flags & JBD2_LOADED)); @@ -1627,8 +1622,6 @@ int jbd2_journal_wipe(journal_t *journal, int write) if (err) return err; - sb = journal->j_superblock; - if (!journal->j_tail) goto no_recovery; @@ -2202,8 +2195,6 @@ void jbd2_journal_init_jbd_inode(struct jbd2_inode *jinode, struct inode *inode) void jbd2_journal_release_jbd_inode(journal_t *journal, struct jbd2_inode *jinode) { - int writeout = 0; - if (!journal) return; restart: @@ -2220,9 +2211,6 @@ restart: goto restart; } - /* Do we need to wait for data writeback? */ - if (journal->j_committing_transaction == jinode->i_transaction) - writeout = 1; if (jinode->i_transaction) { list_del(&jinode->i_list); jinode->i_transaction = NULL; diff --git a/fs/jbd2/recovery.c b/fs/jbd2/recovery.c index 049281b7cb8..2bc4d5f116f 100644 --- a/fs/jbd2/recovery.c +++ b/fs/jbd2/recovery.c @@ -285,12 +285,10 @@ int jbd2_journal_recover(journal_t *journal) int jbd2_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); @@ -299,7 +297,8 @@ int jbd2_journal_skip_recovery(journal_t *journal) ++journal->j_transaction_sequence; } else { #ifdef CONFIG_JBD2_DEBUG - int dropped = info.end_transaction - be32_to_cpu(sb->s_sequence); + int dropped = info.end_transaction - + be32_to_cpu(journal->j_superblock->s_sequence); #endif jbd_debug(1, "JBD: ignoring %d transaction%s from the journal.\n", @@ -365,11 +364,6 @@ static int do_one_pass(journal_t *journal, 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); - /* * First thing is to establish what we expect to find in the log * (in terms of transaction IDs), and where (in terms of log From 206f7ab4f49a2021fcb8687f25395be77711ddee Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 14 Jun 2010 14:42:49 -0400 Subject: [PATCH 04/40] ext4: remove vestiges of nobh support The nobh option was only supported for writeback mode, but given that all write paths actually create buffer heads it effectively was a no-op already. Signed-off-by: Christoph Hellwig Signed-off-by: "Theodore Ts'o" --- fs/ext4/ext4.h | 1 - fs/ext4/ext4_jbd2.h | 10 ++++------ fs/ext4/inode.c | 27 ++++++--------------------- fs/ext4/super.c | 21 +++++---------------- 4 files changed, 15 insertions(+), 44 deletions(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 8b56b53328e..8b6d297c8c7 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -873,7 +873,6 @@ struct ext4_inode_info { #define EXT4_MOUNT_POSIX_ACL 0x08000 /* POSIX Access Control Lists */ #define EXT4_MOUNT_NO_AUTO_DA_ALLOC 0x10000 /* No auto delalloc mapping */ #define EXT4_MOUNT_BARRIER 0x20000 /* Use block barriers */ -#define EXT4_MOUNT_NOBH 0x40000 /* No bufferheads */ #define EXT4_MOUNT_QUOTA 0x80000 /* Some quota option set */ #define EXT4_MOUNT_USRQUOTA 0x100000 /* "old" user quota */ #define EXT4_MOUNT_GRPQUOTA 0x200000 /* "old" group quota */ diff --git a/fs/ext4/ext4_jbd2.h b/fs/ext4/ext4_jbd2.h index 8ae8168900b..38d1e66e584 100644 --- a/fs/ext4/ext4_jbd2.h +++ b/fs/ext4/ext4_jbd2.h @@ -313,17 +313,15 @@ static inline int ext4_should_writeback_data(struct inode *inode) * This function controls whether or not we should try to go down the * dioread_nolock code paths, which makes it safe to avoid taking * i_mutex for direct I/O reads. This only works for extent-based - * files, and it doesn't work for nobh or if data journaling is - * enabled, since the dioread_nolock code uses b_private to pass - * information back to the I/O completion handler, and this conflicts - * with the jbd's use of b_private. + * files, and it doesn't work if data journaling is enabled, since the + * dioread_nolock code uses b_private to pass information back to the + * I/O completion handler, and this conflicts with the jbd's use of + * b_private. */ static inline int ext4_should_dioread_nolock(struct inode *inode) { if (!test_opt(inode->i_sb, DIOREAD_NOLOCK)) return 0; - if (test_opt(inode->i_sb, NOBH)) - return 0; if (!S_ISREG(inode->i_mode)) return 0; if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 15ff8d8d839..b485987f014 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -2553,18 +2553,16 @@ static int ext4_da_get_block_prep(struct inode *inode, sector_t iblock, /* * This function is used as a standard get_block_t calback function * when there is no desire to allocate any blocks. It is used as a - * callback function for block_prepare_write(), nobh_writepage(), and - * block_write_full_page(). These functions should only try to map a - * single block at a time. + * callback function for block_prepare_write() and block_write_full_page(). + * These functions should only try to map a single block at a time. * * Since this function doesn't do block allocations even if the caller * requests it by passing in create=1, it is critically important that * any caller checks to make sure that any buffer heads are returned * by this function are either all already mapped or marked for - * delayed allocation before calling nobh_writepage() or - * block_write_full_page(). Otherwise, b_blocknr could be left - * unitialized, and the page write functions will be taken by - * surprise. + * delayed allocation before calling block_write_full_page(). Otherwise, + * b_blocknr could be left unitialized, and the page write functions will + * be taken by surprise. */ static int noalloc_get_block_write(struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create) @@ -2749,9 +2747,7 @@ static int ext4_writepage(struct page *page, return __ext4_journalled_writepage(page, len); } - if (test_opt(inode->i_sb, NOBH) && ext4_should_writeback_data(inode)) - ret = nobh_writepage(page, noalloc_get_block_write, wbc); - else if (page_bufs && buffer_uninit(page_bufs)) { + if (page_bufs && buffer_uninit(page_bufs)) { ext4_set_bh_endio(page_bufs, inode); ret = block_write_full_page_endio(page, noalloc_get_block_write, wbc, ext4_end_io_buffer_write); @@ -4125,17 +4121,6 @@ int ext4_block_truncate_page(handle_t *handle, length = blocksize - (offset & (blocksize - 1)); iblock = index << (PAGE_CACHE_SHIFT - inode->i_sb->s_blocksize_bits); - /* - * For "nobh" option, we can only work if we don't need to - * read-in the page - otherwise we create buffers to do the IO. - */ - if (!page_has_buffers(page) && test_opt(inode->i_sb, NOBH) && - ext4_should_writeback_data(inode) && PageUptodate(page)) { - zero_user(page, offset, length); - set_page_dirty(page); - goto unlock; - } - if (!page_has_buffers(page)) create_empty_buffers(page, blocksize, 0); diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 4e8983a9811..422a4ce6677 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -946,8 +946,6 @@ static int ext4_show_options(struct seq_file *seq, struct vfsmount *vfs) seq_puts(seq, ",journal_async_commit"); else if (test_opt(sb, JOURNAL_CHECKSUM)) seq_puts(seq, ",journal_checksum"); - if (test_opt(sb, NOBH)) - seq_puts(seq, ",nobh"); if (test_opt(sb, I_VERSION)) seq_puts(seq, ",i_version"); if (!test_opt(sb, DELALLOC)) @@ -1624,10 +1622,12 @@ set_qf_format: *n_blocks_count = option; break; case Opt_nobh: - set_opt(sbi->s_mount_opt, NOBH); + ext4_msg(sb, KERN_WARNING, + "Ignoring deprecated nobh option"); break; case Opt_bh: - clear_opt(sbi->s_mount_opt, NOBH); + ext4_msg(sb, KERN_WARNING, + "Ignoring deprecated bh option"); break; case Opt_i_version: set_opt(sbi->s_mount_opt, I_VERSION); @@ -2912,18 +2912,7 @@ no_journal: ext4_msg(sb, KERN_ERR, "insufficient memory"); goto failed_mount_wq; } - if (test_opt(sb, NOBH)) { - if (!(test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_WRITEBACK_DATA)) { - ext4_msg(sb, KERN_WARNING, "Ignoring nobh option - " - "its supported only with writeback mode"); - clear_opt(sbi->s_mount_opt, NOBH); - } - if (test_opt(sb, DIOREAD_NOLOCK)) { - ext4_msg(sb, KERN_WARNING, "dioread_nolock option is " - "not supported with nobh mode"); - goto failed_mount_wq; - } - } + EXT4_SB(sb)->dio_unwritten_wq = create_workqueue("ext4-dio-unwritten"); if (!EXT4_SB(sb)->dio_unwritten_wq) { printk(KERN_ERR "EXT4-fs: failed to create DIO workqueue\n"); From c6ac12a6159c802ae8b757dd13563564e64333df Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 15 Jun 2010 12:19:59 -0400 Subject: [PATCH 05/40] ext4: update ctime when changing the file's permission by setfacl ext4 didn't update the ctime of the file when its permission was changed. Steps to reproduce: # touch aaa # stat -c %Z aaa 1275289822 # setfacl -m 'u::x,g::x,o::x' aaa # stat -c %Z aaa 1275289822 <- unchanged But, according to the spec of the ctime, ext4 must update it. Port of ext3 patch by Miao Xie . CC: linux-ext4@vger.kernel.org Signed-off-by: Jan Kara Signed-off-by: "Theodore Ts'o" --- fs/ext4/acl.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/ext4/acl.c b/fs/ext4/acl.c index feaf498feaa..5e2ed4504ea 100644 --- a/fs/ext4/acl.c +++ b/fs/ext4/acl.c @@ -204,6 +204,7 @@ ext4_set_acl(handle_t *handle, struct inode *inode, int type, return error; else { inode->i_mode = mode; + inode->i_ctime = ext4_current_time(inode); ext4_mark_inode_dirty(handle, inode); if (error == 0) acl = NULL; From 4a9cdec73f79b2858e9ecf0b6cfac7f6b200bf3a Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 29 Jun 2010 11:00:23 -0400 Subject: [PATCH 06/40] ext4: Add new superblock fields reserved for the Next3 snapshot feature Signed-off-by: "Theodore Ts'o" --- fs/ext4/ext4.h | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 8b6d297c8c7..3864a277545 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1002,7 +1002,13 @@ struct ext4_super_block { __u8 s_reserved_char_pad2; __le16 s_reserved_pad; __le64 s_kbytes_written; /* nr of lifetime kilobytes written */ - __u32 s_reserved[160]; /* Padding to the end of the block */ + __le32 s_snapshot_inum; /* Inode number of active snapshot */ + __le32 s_snapshot_id; /* sequential ID of active snapshot */ + __le64 s_snapshot_r_blocks_count; /* reserved blocks for active + snapshot's future use */ + __le32 s_snapshot_list; /* inode number of the head of the + on-disk snapshot list */ + __u32 s_reserved[155]; /* Padding to the end of the block */ }; #ifdef __KERNEL__ From c67d859e39896e4286249da89c4ca0ef8bd949cb Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 29 Jun 2010 11:07:07 -0400 Subject: [PATCH 07/40] ext4: clean up ext4_abort() so __func__ is now implicit Use a macro definition for ext4_abort() to clean up the .c files a wee bit. Signed-off-by: "Theodore Ts'o" --- fs/ext4/ext4.h | 4 +++- fs/ext4/ext4_jbd2.c | 4 ++-- fs/ext4/super.c | 11 +++++------ 3 files changed, 10 insertions(+), 9 deletions(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 3864a277545..d8ec824f5a2 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1629,8 +1629,10 @@ extern void ext4_error_inode(const char *, struct inode *, const char *, ...) extern void ext4_error_file(const char *, struct file *, const char *, ...) __attribute__ ((format (printf, 3, 4))); extern void __ext4_std_error(struct super_block *, const char *, int); -extern void ext4_abort(struct super_block *, const char *, const char *, ...) +extern void __ext4_abort(struct super_block *, const char *, const char *, ...) __attribute__ ((format (printf, 3, 4))); +#define ext4_abort(sb, message...) __ext4_abort(sb, __func__, \ + ## message) extern void __ext4_warning(struct super_block *, const char *, const char *, ...) __attribute__ ((format (printf, 3, 4))); diff --git a/fs/ext4/ext4_jbd2.c b/fs/ext4/ext4_jbd2.c index cfd27b38fa1..df26a015d04 100644 --- a/fs/ext4/ext4_jbd2.c +++ b/fs/ext4/ext4_jbd2.c @@ -93,8 +93,8 @@ int __ext4_forget(const char *where, handle_t *handle, int is_metadata, err = jbd2_journal_revoke(handle, blocknr, bh); if (err) { ext4_journal_abort_handle(where, __func__, bh, handle, err); - ext4_abort(inode->i_sb, __func__, - "error %d when attempting revoke", err); + __ext4_abort(inode->i_sb, where, + "error %d when attempting revoke", err); } BUFFER_TRACE(bh, "exit"); return err; diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 422a4ce6677..11441e483b6 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -248,7 +248,7 @@ handle_t *ext4_journal_start_sb(struct super_block *sb, int nblocks) journal = EXT4_SB(sb)->s_journal; if (journal) { if (is_journal_aborted(journal)) { - ext4_abort(sb, __func__, "Detected aborted journal"); + ext4_abort(sb, "Detected aborted journal"); return ERR_PTR(-EROFS); } return jbd2_journal_start(journal, nblocks); @@ -464,8 +464,8 @@ void __ext4_std_error(struct super_block *sb, const char *function, int errno) * case we take the easy way out and panic immediately. */ -void ext4_abort(struct super_block *sb, const char *function, - const char *fmt, ...) +void __ext4_abort(struct super_block *sb, const char *function, + const char *fmt, ...) { va_list args; @@ -660,8 +660,7 @@ static void ext4_put_super(struct super_block *sb) err = jbd2_journal_destroy(sbi->s_journal); sbi->s_journal = NULL; if (err < 0) - ext4_abort(sb, __func__, - "Couldn't clean up the journal"); + ext4_abort(sb, "Couldn't clean up the journal"); } ext4_release_system_zone(sb); @@ -3605,7 +3604,7 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data) } if (sbi->s_mount_flags & EXT4_MF_FS_ABORTED) - ext4_abort(sb, __func__, "Abort forced by user"); + ext4_abort(sb, "Abort forced by user"); sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | (test_opt(sb, POSIX_ACL) ? MS_POSIXACL : 0); From e29136f80e775b0310273932b4297a62f5574a29 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 29 Jun 2010 12:54:28 -0400 Subject: [PATCH 08/40] ext4: Enhance ext4_grp_locked_error() to take block and function numbers Also use a macro definition so that __func__ and __LINE__ is implicit. Signed-off-by: "Theodore Ts'o" --- fs/ext4/ext4.h | 10 +++++++--- fs/ext4/inode.c | 13 +++++++------ fs/ext4/mballoc.c | 41 +++++++++++++++++++++-------------------- fs/ext4/super.c | 13 ++++++++++--- 4 files changed, 45 insertions(+), 32 deletions(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index d8ec824f5a2..5a41881cafc 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1639,9 +1639,13 @@ extern void __ext4_warning(struct super_block *, const char *, #define ext4_warning(sb, message...) __ext4_warning(sb, __func__, ## message) extern void ext4_msg(struct super_block *, const char *, const char *, ...) __attribute__ ((format (printf, 3, 4))); -extern void ext4_grp_locked_error(struct super_block *, ext4_group_t, - const char *, const char *, ...) - __attribute__ ((format (printf, 4, 5))); +extern void __ext4_grp_locked_error(const char *, unsigned int, \ + struct super_block *, ext4_group_t, \ + unsigned long, ext4_fsblk_t, \ + const char *, ...) + __attribute__ ((format (printf, 7, 8))); +#define ext4_grp_locked_error(sb, grp, message...) \ + __ext4_grp_locked_error(__func__, __LINE__, (sb), (grp), ## message) extern void ext4_update_dynamic_rev(struct super_block *sb); extern int ext4_update_compat_feature(handle_t *handle, struct super_block *sb, __u32 compat); diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index b485987f014..64baadb4956 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -1128,8 +1128,8 @@ void ext4_da_update_reserve_space(struct inode *inode, ext4_discard_preallocations(inode); } -static int check_block_validity(struct inode *inode, const char *func, - struct ext4_map_blocks *map) +static int __check_block_validity(struct inode *inode, const char *func, + struct ext4_map_blocks *map) { if (!ext4_data_block_valid(EXT4_SB(inode->i_sb), map->m_pblk, map->m_len)) { @@ -1142,6 +1142,9 @@ static int check_block_validity(struct inode *inode, const char *func, return 0; } +#define check_block_validity(inode, map) \ + __check_block_validity((inode), __func__, (map)) + /* * Return the number of contiguous dirty pages in a given inode * starting at page frame idx. @@ -1244,7 +1247,7 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode, up_read((&EXT4_I(inode)->i_data_sem)); if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED) { - int ret = check_block_validity(inode, __func__, map); + int ret = check_block_validity(inode, map); if (ret != 0) return ret; } @@ -1324,9 +1327,7 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode, up_write((&EXT4_I(inode)->i_data_sem)); if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED) { - int ret = check_block_validity(inode, - "ext4_map_blocks_after_alloc", - map); + int ret = check_block_validity(inode, map); if (ret != 0) return ret; } diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index b2948b04797..3dfad95f0f9 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -446,10 +446,11 @@ static void mb_free_blocks_double(struct inode *inode, struct ext4_buddy *e4b, blocknr = ext4_group_first_block_no(sb, e4b->bd_group); blocknr += first + i; ext4_grp_locked_error(sb, e4b->bd_group, - __func__, "double-free of inode" - " %lu's block %llu(bit %u in group %u)", - inode ? inode->i_ino : 0, blocknr, - first + i, e4b->bd_group); + inode ? inode->i_ino : 0, + blocknr, + "freeing block already freed " + "(bit %u)", + first + i); } mb_clear_bit(first + i, e4b->bd_info->bb_bitmap); } @@ -712,9 +713,9 @@ void ext4_mb_generate_buddy(struct super_block *sb, grp->bb_fragments = fragments; if (free != grp->bb_free) { - ext4_grp_locked_error(sb, group, __func__, - "EXT4-fs: group %u: %u blocks in bitmap, %u in gd", - group, free, grp->bb_free); + ext4_grp_locked_error(sb, group, 0, 0, + "%u blocks in bitmap, %u in gd", + free, grp->bb_free); /* * If we intent to continue, we consider group descritor * corrupt and update bb_free using bitmap value @@ -1296,10 +1297,10 @@ static void mb_free_blocks(struct inode *inode, struct ext4_buddy *e4b, blocknr = ext4_group_first_block_no(sb, e4b->bd_group); blocknr += block; ext4_grp_locked_error(sb, e4b->bd_group, - __func__, "double-free of inode" - " %lu's block %llu(bit %u in group %u)", - inode ? inode->i_ino : 0, blocknr, block, - e4b->bd_group); + inode ? inode->i_ino : 0, + blocknr, + "freeing already freed block " + "(bit %u)", block); } mb_clear_bit(block, EXT4_MB_BITMAP(e4b)); e4b->bd_info->bb_counters[order]++; @@ -1788,8 +1789,8 @@ void ext4_mb_complex_scan_group(struct ext4_allocation_context *ac, * free blocks even though group info says we * we have free blocks */ - ext4_grp_locked_error(sb, e4b->bd_group, - __func__, "%d free blocks as per " + ext4_grp_locked_error(sb, e4b->bd_group, 0, 0, + "%d free blocks as per " "group info. But bitmap says 0", free); break; @@ -1798,8 +1799,8 @@ void ext4_mb_complex_scan_group(struct ext4_allocation_context *ac, mb_find_extent(e4b, 0, i, ac->ac_g_ex.fe_len, &ex); BUG_ON(ex.fe_len <= 0); if (free < ex.fe_len) { - ext4_grp_locked_error(sb, e4b->bd_group, - __func__, "%d free blocks as per " + ext4_grp_locked_error(sb, e4b->bd_group, 0, 0, + "%d free blocks as per " "group info. But got %d blocks", free, ex.fe_len); /* @@ -3584,8 +3585,7 @@ ext4_mb_release_inode_pa(struct ext4_buddy *e4b, struct buffer_head *bitmap_bh, pa, (unsigned long) pa->pa_lstart, (unsigned long) pa->pa_pstart, (unsigned long) pa->pa_len); - ext4_grp_locked_error(sb, group, - __func__, "free %u, pa_free %u", + ext4_grp_locked_error(sb, group, 0, 0, "free %u, pa_free %u", free, pa->pa_free); /* * pa is already deleted so we use the value obtained @@ -4395,6 +4395,7 @@ static noinline_for_stack int ext4_mb_free_metadata(handle_t *handle, struct ext4_buddy *e4b, struct ext4_free_data *new_entry) { + ext4_group_t group = e4b->bd_group; ext4_grpblk_t block; struct ext4_free_data *entry; struct ext4_group_info *db = e4b->bd_info; @@ -4427,9 +4428,9 @@ ext4_mb_free_metadata(handle_t *handle, struct ext4_buddy *e4b, else if (block >= (entry->start_blk + entry->count)) n = &(*n)->rb_right; else { - ext4_grp_locked_error(sb, e4b->bd_group, __func__, - "Double free of blocks %d (%d %d)", - block, entry->start_blk, entry->count); + ext4_grp_locked_error(sb, group, 0, + ext4_group_first_block_no(sb, group) + block, + "Block already on to-be-freed list"); return 0; } } diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 11441e483b6..39aeb454bf2 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -514,8 +514,10 @@ void __ext4_warning(struct super_block *sb, const char *function, va_end(args); } -void ext4_grp_locked_error(struct super_block *sb, ext4_group_t grp, - const char *function, const char *fmt, ...) +void __ext4_grp_locked_error(const char *function, unsigned int line, + struct super_block *sb, ext4_group_t grp, + unsigned long ino, ext4_fsblk_t block, + const char *fmt, ...) __releases(bitlock) __acquires(bitlock) { @@ -523,7 +525,12 @@ __acquires(bitlock) struct ext4_super_block *es = EXT4_SB(sb)->s_es; va_start(args, fmt); - printk(KERN_CRIT "EXT4-fs error (device %s): %s: ", sb->s_id, function); + printk(KERN_CRIT "EXT4-fs error (device %s): %s:%d: group %u", + sb->s_id, function, line, grp); + if (ino) + printk("inode %lu: ", ino); + if (block) + printk("block %llu:", (unsigned long long) block); vprintk(fmt, args); printk("\n"); va_end(args); From 90c7201b97bb7ac5a4e2605abc0efb5fdfb957f0 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 29 Jun 2010 14:53:24 -0400 Subject: [PATCH 09/40] ext4: Pass line number to ext4_journal_abort_handle() This allows the error messages to include the line number Signed-off-by: "Theodore Ts'o" --- fs/ext4/ext4_jbd2.c | 48 +++++++++++++++++++++++---------------------- fs/ext4/ext4_jbd2.h | 43 +++++++++++++++++++++------------------- fs/ext4/super.c | 9 +++++---- 3 files changed, 53 insertions(+), 47 deletions(-) diff --git a/fs/ext4/ext4_jbd2.c b/fs/ext4/ext4_jbd2.c index df26a015d04..9de37b9e177 100644 --- a/fs/ext4/ext4_jbd2.c +++ b/fs/ext4/ext4_jbd2.c @@ -6,29 +6,29 @@ #include -int __ext4_journal_get_undo_access(const char *where, handle_t *handle, - struct buffer_head *bh) +int __ext4_journal_get_undo_access(const char *where, unsigned int line, + handle_t *handle, struct buffer_head *bh) { int err = 0; if (ext4_handle_valid(handle)) { err = jbd2_journal_get_undo_access(handle, bh); if (err) - ext4_journal_abort_handle(where, __func__, bh, + ext4_journal_abort_handle(where, line, __func__, bh, handle, err); } return err; } -int __ext4_journal_get_write_access(const char *where, handle_t *handle, - struct buffer_head *bh) +int __ext4_journal_get_write_access(const char *where, unsigned int line, + handle_t *handle, struct buffer_head *bh) { int err = 0; if (ext4_handle_valid(handle)) { err = jbd2_journal_get_write_access(handle, bh); if (err) - ext4_journal_abort_handle(where, __func__, bh, + ext4_journal_abort_handle(where, line, __func__, bh, handle, err); } return err; @@ -46,9 +46,9 @@ int __ext4_journal_get_write_access(const char *where, handle_t *handle, * If the handle isn't valid we're not journaling, but we still need to * call into ext4_journal_revoke() to put the buffer head. */ -int __ext4_forget(const char *where, handle_t *handle, int is_metadata, - struct inode *inode, struct buffer_head *bh, - ext4_fsblk_t blocknr) +int __ext4_forget(const char *where, unsigned int line, handle_t *handle, + int is_metadata, struct inode *inode, + struct buffer_head *bh, ext4_fsblk_t blocknr) { int err; @@ -79,8 +79,8 @@ int __ext4_forget(const char *where, handle_t *handle, int is_metadata, BUFFER_TRACE(bh, "call jbd2_journal_forget"); err = jbd2_journal_forget(handle, bh); if (err) - ext4_journal_abort_handle(where, __func__, bh, - handle, err); + ext4_journal_abort_handle(where, line, __func__, + bh, handle, err); return err; } return 0; @@ -92,7 +92,8 @@ int __ext4_forget(const char *where, handle_t *handle, int is_metadata, BUFFER_TRACE(bh, "call jbd2_journal_revoke"); err = jbd2_journal_revoke(handle, blocknr, bh); if (err) { - ext4_journal_abort_handle(where, __func__, bh, handle, err); + ext4_journal_abort_handle(where, line, __func__, + bh, handle, err); __ext4_abort(inode->i_sb, where, "error %d when attempting revoke", err); } @@ -100,7 +101,7 @@ int __ext4_forget(const char *where, handle_t *handle, int is_metadata, return err; } -int __ext4_journal_get_create_access(const char *where, +int __ext4_journal_get_create_access(const char *where, unsigned int line, handle_t *handle, struct buffer_head *bh) { int err = 0; @@ -108,22 +109,23 @@ int __ext4_journal_get_create_access(const char *where, if (ext4_handle_valid(handle)) { err = jbd2_journal_get_create_access(handle, bh); if (err) - ext4_journal_abort_handle(where, __func__, bh, - handle, err); + ext4_journal_abort_handle(where, line, __func__, + bh, handle, err); } return err; } -int __ext4_handle_dirty_metadata(const char *where, handle_t *handle, - struct inode *inode, struct buffer_head *bh) +int __ext4_handle_dirty_metadata(const char *where, unsigned int line, + handle_t *handle, struct inode *inode, + struct buffer_head *bh) { int err = 0; if (ext4_handle_valid(handle)) { err = jbd2_journal_dirty_metadata(handle, bh); if (err) - ext4_journal_abort_handle(where, __func__, bh, - handle, err); + ext4_journal_abort_handle(where, line, __func__, + bh, handle, err); } else { if (inode) mark_buffer_dirty_inode(bh, inode); @@ -144,8 +146,8 @@ int __ext4_handle_dirty_metadata(const char *where, handle_t *handle, return err; } -int __ext4_handle_dirty_super(const char *where, handle_t *handle, - struct super_block *sb) +int __ext4_handle_dirty_super(const char *where, unsigned int line, + handle_t *handle, struct super_block *sb) { struct buffer_head *bh = EXT4_SB(sb)->s_sbh; int err = 0; @@ -153,8 +155,8 @@ int __ext4_handle_dirty_super(const char *where, handle_t *handle, if (ext4_handle_valid(handle)) { err = jbd2_journal_dirty_metadata(handle, bh); if (err) - ext4_journal_abort_handle(where, __func__, bh, - handle, err); + ext4_journal_abort_handle(where, line, __func__, + bh, handle, err); } else sb->s_dirt = 1; return err; diff --git a/fs/ext4/ext4_jbd2.h b/fs/ext4/ext4_jbd2.h index 38d1e66e584..6883c6be5b8 100644 --- a/fs/ext4/ext4_jbd2.h +++ b/fs/ext4/ext4_jbd2.h @@ -122,41 +122,44 @@ int ext4_mark_inode_dirty(handle_t *handle, struct inode *inode); /* * Wrapper functions with which ext4 calls into JBD. */ -void ext4_journal_abort_handle(const char *caller, const char *err_fn, +void ext4_journal_abort_handle(const char *caller, unsigned int line, + const char *err_fn, struct buffer_head *bh, handle_t *handle, int err); -int __ext4_journal_get_undo_access(const char *where, handle_t *handle, - struct buffer_head *bh); +int __ext4_journal_get_undo_access(const char *where, unsigned int line, + handle_t *handle, struct buffer_head *bh); -int __ext4_journal_get_write_access(const char *where, handle_t *handle, - struct buffer_head *bh); +int __ext4_journal_get_write_access(const char *where, unsigned int line, + handle_t *handle, struct buffer_head *bh); -int __ext4_forget(const char *where, handle_t *handle, int is_metadata, - struct inode *inode, struct buffer_head *bh, - ext4_fsblk_t blocknr); +int __ext4_forget(const char *where, unsigned int line, handle_t *handle, + int is_metadata, struct inode *inode, + struct buffer_head *bh, ext4_fsblk_t blocknr); -int __ext4_journal_get_create_access(const char *where, +int __ext4_journal_get_create_access(const char *where, unsigned int line, handle_t *handle, struct buffer_head *bh); -int __ext4_handle_dirty_metadata(const char *where, handle_t *handle, - struct inode *inode, struct buffer_head *bh); +int __ext4_handle_dirty_metadata(const char *where, unsigned int line, + handle_t *handle, struct inode *inode, + struct buffer_head *bh); -int __ext4_handle_dirty_super(const char *where, handle_t *handle, - struct super_block *sb); +int __ext4_handle_dirty_super(const char *where, unsigned int line, + handle_t *handle, struct super_block *sb); #define ext4_journal_get_undo_access(handle, bh) \ - __ext4_journal_get_undo_access(__func__, (handle), (bh)) + __ext4_journal_get_undo_access(__func__, __LINE__, (handle), (bh)) #define ext4_journal_get_write_access(handle, bh) \ - __ext4_journal_get_write_access(__func__, (handle), (bh)) + __ext4_journal_get_write_access(__func__, __LINE__, (handle), (bh)) #define ext4_forget(handle, is_metadata, inode, bh, block_nr) \ - __ext4_forget(__func__, (handle), (is_metadata), (inode), (bh),\ - (block_nr)) + __ext4_forget(__func__, __LINE__, (handle), (is_metadata), (inode), \ + (bh), (block_nr)) #define ext4_journal_get_create_access(handle, bh) \ - __ext4_journal_get_create_access(__func__, (handle), (bh)) + __ext4_journal_get_create_access(__func__, __LINE__, (handle), (bh)) #define ext4_handle_dirty_metadata(handle, inode, bh) \ - __ext4_handle_dirty_metadata(__func__, (handle), (inode), (bh)) + __ext4_handle_dirty_metadata(__func__, __LINE__, (handle), (inode), \ + (bh)) #define ext4_handle_dirty_super(handle, sb) \ - __ext4_handle_dirty_super(__func__, (handle), (sb)) + __ext4_handle_dirty_super(__func__, __LINE__, (handle), (sb)) handle_t *ext4_journal_start_sb(struct super_block *sb, int nblocks); int __ext4_journal_stop(const char *where, handle_t *handle); diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 39aeb454bf2..87db5ecfccb 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -283,8 +283,9 @@ int __ext4_journal_stop(const char *where, handle_t *handle) return err; } -void ext4_journal_abort_handle(const char *caller, const char *err_fn, - struct buffer_head *bh, handle_t *handle, int err) +void ext4_journal_abort_handle(const char *caller, unsigned int line, + const char *err_fn, struct buffer_head *bh, + handle_t *handle, int err) { char nbuf[16]; const char *errstr = ext4_decode_error(NULL, err, nbuf); @@ -300,8 +301,8 @@ void ext4_journal_abort_handle(const char *caller, const char *err_fn, if (is_handle_aborted(handle)) return; - printk(KERN_ERR "%s: aborting transaction: %s in %s\n", - caller, errstr, err_fn); + printk(KERN_ERR "%s:%d: aborting transaction: %s in %s\n", + caller, line, errstr, err_fn); jbd2_journal_abort_handle(handle); } From 60fd4da34d55a9cc0d857fc76dc12cf8cab4ed02 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 27 Jul 2010 11:54:40 -0400 Subject: [PATCH 10/40] ext4: Cleanup ext4_check_dir_entry so __func__ is now implicit Also start passing the line number to ext4_check_dir since we're going to need it in upcoming patch. Signed-off-by: "Theodore Ts'o" --- fs/ext4/dir.c | 11 ++++++----- fs/ext4/ext4.h | 8 +++++--- fs/ext4/namei.c | 14 ++++++-------- 3 files changed, 17 insertions(+), 16 deletions(-) diff --git a/fs/ext4/dir.c b/fs/ext4/dir.c index 2965c39d418..af581f08fe3 100644 --- a/fs/ext4/dir.c +++ b/fs/ext4/dir.c @@ -61,10 +61,11 @@ static unsigned char get_dtype(struct super_block *sb, int filetype) } -int ext4_check_dir_entry(const char *function, struct inode *dir, - struct ext4_dir_entry_2 *de, - struct buffer_head *bh, - unsigned int offset) +int __ext4_check_dir_entry(const char *function, unsigned int line, + struct inode *dir, + struct ext4_dir_entry_2 *de, + struct buffer_head *bh, + unsigned int offset) { const char *error_msg = NULL; const int rlen = ext4_rec_len_from_disk(de->rec_len, @@ -194,7 +195,7 @@ revalidate: while (!error && filp->f_pos < inode->i_size && offset < sb->s_blocksize) { de = (struct ext4_dir_entry_2 *) (bh->b_data + offset); - if (!ext4_check_dir_entry("ext4_readdir", inode, de, + if (!ext4_check_dir_entry(inode, de, bh, offset)) { /* * On error, skip the f_pos to the next block diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 5a41881cafc..73465b26976 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1515,9 +1515,11 @@ extern unsigned ext4_init_block_bitmap(struct super_block *sb, ext4_init_block_bitmap(sb, NULL, group, desc) /* dir.c */ -extern int ext4_check_dir_entry(const char *, struct inode *, - struct ext4_dir_entry_2 *, - struct buffer_head *, unsigned int); +extern int __ext4_check_dir_entry(const char *, unsigned int, struct inode *, + struct ext4_dir_entry_2 *, + struct buffer_head *, unsigned int); +#define ext4_check_dir_entry(dir, de, bh, offset) \ + __ext4_check_dir_entry(__func__, __LINE__, (dir), (de), (bh), (offset)) extern int ext4_htree_store_dirent(struct file *dir_file, __u32 hash, __u32 minor_hash, struct ext4_dir_entry_2 *dirent); diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c index 5a61f77e7d7..ea8b59d9621 100644 --- a/fs/ext4/namei.c +++ b/fs/ext4/namei.c @@ -605,7 +605,7 @@ static int htree_dirblock_to_tree(struct file *dir_file, dir->i_sb->s_blocksize - EXT4_DIR_REC_LEN(0)); for (; de < top; de = ext4_next_entry(de, dir->i_sb->s_blocksize)) { - if (!ext4_check_dir_entry("htree_dirblock_to_tree", dir, de, bh, + if (!ext4_check_dir_entry(dir, de, bh, (block<i_sb)) +((char *)de - bh->b_data))) { /* On error, skip the f_pos to the next block. */ @@ -844,8 +844,7 @@ static inline int search_dirblock(struct buffer_head *bh, if ((char *) de + namelen <= dlimit && ext4_match (namelen, name, de)) { /* found a match - just to be sure, do a full check */ - if (!ext4_check_dir_entry("ext4_find_entry", - dir, de, bh, offset)) + if (!ext4_check_dir_entry(dir, de, bh, offset)) return -1; *res_dir = de; return 1; @@ -1019,7 +1018,7 @@ static struct buffer_head * ext4_dx_find_entry(struct inode *dir, const struct q int off = (block << EXT4_BLOCK_SIZE_BITS(sb)) + ((char *) de - bh->b_data); - if (!ext4_check_dir_entry(__func__, dir, de, bh, off)) { + if (!ext4_check_dir_entry(dir, de, bh, off)) { brelse(bh); *err = ERR_BAD_DX_DIR; goto errout; @@ -1303,8 +1302,7 @@ static int add_dirent_to_buf(handle_t *handle, struct dentry *dentry, de = (struct ext4_dir_entry_2 *)bh->b_data; top = bh->b_data + blocksize - reclen; while ((char *) de <= top) { - if (!ext4_check_dir_entry("ext4_add_entry", dir, de, - bh, offset)) + if (!ext4_check_dir_entry(dir, de, bh, offset)) return -EIO; if (ext4_match(namelen, name, de)) return -EEXIST; @@ -1671,7 +1669,7 @@ static int ext4_delete_entry(handle_t *handle, pde = NULL; de = (struct ext4_dir_entry_2 *) bh->b_data; while (i < bh->b_size) { - if (!ext4_check_dir_entry("ext4_delete_entry", dir, de, bh, i)) + if (!ext4_check_dir_entry(dir, de, bh, i)) return -EIO; if (de == de_del) { BUFFER_TRACE(bh, "get_write_access"); @@ -1954,7 +1952,7 @@ static int empty_dir(struct inode *inode) } de = (struct ext4_dir_entry_2 *) bh->b_data; } - if (!ext4_check_dir_entry("empty_dir", inode, de, bh, offset)) { + if (!ext4_check_dir_entry(inode, de, bh, offset)) { de = (struct ext4_dir_entry_2 *)(bh->b_data + sb->s_blocksize); offset = (offset | (sb->s_blocksize - 1)) + 1; From c398eda0e43a791be0fca6f197a1e2bbb9f16070 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 27 Jul 2010 11:56:40 -0400 Subject: [PATCH 11/40] ext4: Pass line numbers to ext4_error() and friends Signed-off-by: "Theodore Ts'o" --- fs/ext4/dir.c | 7 +++---- fs/ext4/ext4.h | 44 ++++++++++++++++++++++++---------------- fs/ext4/ext4_jbd2.c | 12 +++++------ fs/ext4/ext4_jbd2.h | 4 ++-- fs/ext4/extents.c | 10 ++++----- fs/ext4/inode.c | 44 +++++++++++++++++++++------------------- fs/ext4/move_extent.c | 10 ++++----- fs/ext4/super.c | 47 ++++++++++++++++++++++++++----------------- 8 files changed, 98 insertions(+), 80 deletions(-) diff --git a/fs/ext4/dir.c b/fs/ext4/dir.c index af581f08fe3..62e8af04ed1 100644 --- a/fs/ext4/dir.c +++ b/fs/ext4/dir.c @@ -84,11 +84,10 @@ int __ext4_check_dir_entry(const char *function, unsigned int line, error_msg = "inode out of bounds"; if (error_msg != NULL) - ext4_error_inode(function, dir, - "bad entry in directory: %s - block=%llu" + ext4_error_inode(dir, function, line, bh->b_blocknr, + "bad entry in directory: %s - " "offset=%u(%u), inode=%u, rec_len=%d, name_len=%d", - error_msg, (unsigned long long) bh->b_blocknr, - (unsigned) (offset%bh->b_size), offset, + error_msg, (unsigned) (offset%bh->b_size), offset, le32_to_cpu(de->inode), rlen, de->name_len); return error_msg == NULL ? 1 : 0; diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 73465b26976..088938148f5 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -57,10 +57,13 @@ #endif #define EXT4_ERROR_INODE(inode, fmt, a...) \ - ext4_error_inode(__func__, (inode), (fmt), ## a) + ext4_error_inode((inode), __func__, __LINE__, 0, (fmt), ## a) + +#define EXT4_ERROR_INODE_BLOCK(inode, block, fmt, a...) \ + ext4_error_inode((inode), __func__, __LINE__, (block), (fmt), ## a) #define EXT4_ERROR_FILE(file, fmt, a...) \ - ext4_error_file(__func__, (file), (fmt), ## a) + ext4_error_file(__func__, __LINE__, (file), (fmt), ## a) /* data type for block offset of block group */ typedef int ext4_grpblk_t; @@ -1623,22 +1626,29 @@ extern int ext4_group_extend(struct super_block *sb, ext4_fsblk_t n_blocks_count); /* super.c */ -extern void __ext4_error(struct super_block *, const char *, const char *, ...) - __attribute__ ((format (printf, 3, 4))); -#define ext4_error(sb, message...) __ext4_error(sb, __func__, ## message) -extern void ext4_error_inode(const char *, struct inode *, const char *, ...) - __attribute__ ((format (printf, 3, 4))); -extern void ext4_error_file(const char *, struct file *, const char *, ...) - __attribute__ ((format (printf, 3, 4))); -extern void __ext4_std_error(struct super_block *, const char *, int); -extern void __ext4_abort(struct super_block *, const char *, const char *, ...) - __attribute__ ((format (printf, 3, 4))); +extern void __ext4_error(struct super_block *, const char *, unsigned int, + const char *, ...) + __attribute__ ((format (printf, 4, 5))); +#define ext4_error(sb, message...) __ext4_error(sb, __func__, \ + __LINE__, ## message) +extern void ext4_error_inode(struct inode *, const char *, unsigned int, + ext4_fsblk_t, const char *, ...) + __attribute__ ((format (printf, 5, 6))); +extern void ext4_error_file(struct file *, const char *, unsigned int, + const char *, ...) + __attribute__ ((format (printf, 4, 5))); +extern void __ext4_std_error(struct super_block *, const char *, + unsigned int, int); +extern void __ext4_abort(struct super_block *, const char *, unsigned int, + const char *, ...) + __attribute__ ((format (printf, 4, 5))); #define ext4_abort(sb, message...) __ext4_abort(sb, __func__, \ - ## message) -extern void __ext4_warning(struct super_block *, const char *, + __LINE__, ## message) +extern void __ext4_warning(struct super_block *, const char *, unsigned int, const char *, ...) - __attribute__ ((format (printf, 3, 4))); -#define ext4_warning(sb, message...) __ext4_warning(sb, __func__, ## message) + __attribute__ ((format (printf, 4, 5))); +#define ext4_warning(sb, message...) __ext4_warning(sb, __func__, \ + __LINE__, ## message) extern void ext4_msg(struct super_block *, const char *, const char *, ...) __attribute__ ((format (printf, 3, 4))); extern void __ext4_grp_locked_error(const char *, unsigned int, \ @@ -1781,7 +1791,7 @@ static inline unsigned int ext4_flex_bg_size(struct ext4_sb_info *sbi) #define ext4_std_error(sb, errno) \ do { \ if ((errno)) \ - __ext4_std_error((sb), __func__, (errno)); \ + __ext4_std_error((sb), __func__, __LINE__, (errno)); \ } while (0) #ifdef CONFIG_SMP diff --git a/fs/ext4/ext4_jbd2.c b/fs/ext4/ext4_jbd2.c index 9de37b9e177..23425cd68da 100644 --- a/fs/ext4/ext4_jbd2.c +++ b/fs/ext4/ext4_jbd2.c @@ -94,8 +94,8 @@ int __ext4_forget(const char *where, unsigned int line, handle_t *handle, if (err) { ext4_journal_abort_handle(where, line, __func__, bh, handle, err); - __ext4_abort(inode->i_sb, where, - "error %d when attempting revoke", err); + __ext4_abort(inode->i_sb, where, line, + "error %d when attempting revoke", err); } BUFFER_TRACE(bh, "exit"); return err; @@ -134,11 +134,9 @@ int __ext4_handle_dirty_metadata(const char *where, unsigned int line, if (inode && inode_needs_sync(inode)) { sync_dirty_buffer(bh); if (buffer_req(bh) && !buffer_uptodate(bh)) { - ext4_error(inode->i_sb, - "IO error syncing inode, " - "inode=%lu, block=%llu", - inode->i_ino, - (unsigned long long) bh->b_blocknr); + ext4_error_inode(inode, where, line, + bh->b_blocknr, + "IO error syncing itable block"); err = -EIO; } } diff --git a/fs/ext4/ext4_jbd2.h b/fs/ext4/ext4_jbd2.h index 6883c6be5b8..b0bd792c58c 100644 --- a/fs/ext4/ext4_jbd2.h +++ b/fs/ext4/ext4_jbd2.h @@ -162,7 +162,7 @@ int __ext4_handle_dirty_super(const char *where, unsigned int line, __ext4_handle_dirty_super(__func__, __LINE__, (handle), (sb)) handle_t *ext4_journal_start_sb(struct super_block *sb, int nblocks); -int __ext4_journal_stop(const char *where, handle_t *handle); +int __ext4_journal_stop(const char *where, unsigned int line, handle_t *handle); #define EXT4_NOJOURNAL_MAX_REF_COUNT ((unsigned long) 4096) @@ -215,7 +215,7 @@ static inline handle_t *ext4_journal_start(struct inode *inode, int nblocks) } #define ext4_journal_stop(handle) \ - __ext4_journal_stop(__func__, (handle)) + __ext4_journal_stop(__func__, __LINE__, (handle)) static inline handle_t *ext4_journal_current_handle(void) { diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index 346de3daab7..2c01d7391f1 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -401,9 +401,9 @@ static int ext4_valid_extent_entries(struct inode *inode, return 1; } -static int __ext4_ext_check(const char *function, struct inode *inode, - struct ext4_extent_header *eh, - int depth) +static int __ext4_ext_check(const char *function, unsigned int line, + struct inode *inode, struct ext4_extent_header *eh, + int depth) { const char *error_msg; int max = 0; @@ -436,7 +436,7 @@ static int __ext4_ext_check(const char *function, struct inode *inode, return 0; corrupted: - ext4_error_inode(function, inode, + ext4_error_inode(inode, function, line, 0, "bad header/extent: %s - magic %x, " "entries %u, max %u(%u), depth %u(%u)", error_msg, le16_to_cpu(eh->eh_magic), @@ -447,7 +447,7 @@ corrupted: } #define ext4_ext_check(inode, eh, depth) \ - __ext4_ext_check(__func__, inode, eh, depth) + __ext4_ext_check(__func__, __LINE__, inode, eh, depth) int ext4_ext_check_inode(struct inode *inode) { diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 64baadb4956..69ea663ef03 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -337,7 +337,8 @@ static int ext4_block_to_path(struct inode *inode, return n; } -static int __ext4_check_blockref(const char *function, struct inode *inode, +static int __ext4_check_blockref(const char *function, unsigned int line, + struct inode *inode, __le32 *p, unsigned int max) { __le32 *bref = p; @@ -348,8 +349,8 @@ static int __ext4_check_blockref(const char *function, struct inode *inode, if (blk && unlikely(!ext4_data_block_valid(EXT4_SB(inode->i_sb), blk, 1))) { - ext4_error_inode(function, inode, - "invalid block reference %u", blk); + ext4_error_inode(inode, function, line, blk, + "invalid block"); return -EIO; } } @@ -358,11 +359,13 @@ static int __ext4_check_blockref(const char *function, struct inode *inode, #define ext4_check_indirect_blockref(inode, bh) \ - __ext4_check_blockref(__func__, inode, (__le32 *)(bh)->b_data, \ + __ext4_check_blockref(__func__, __LINE__, inode, \ + (__le32 *)(bh)->b_data, \ EXT4_ADDR_PER_BLOCK((inode)->i_sb)) #define ext4_check_inode_blockref(inode) \ - __ext4_check_blockref(__func__, inode, EXT4_I(inode)->i_data, \ + __ext4_check_blockref(__func__, __LINE__, inode, \ + EXT4_I(inode)->i_data, \ EXT4_NDIR_BLOCKS) /** @@ -1129,21 +1132,22 @@ void ext4_da_update_reserve_space(struct inode *inode, } static int __check_block_validity(struct inode *inode, const char *func, - struct ext4_map_blocks *map) + unsigned int line, + struct ext4_map_blocks *map) { if (!ext4_data_block_valid(EXT4_SB(inode->i_sb), map->m_pblk, map->m_len)) { - ext4_error_inode(func, inode, - "lblock %lu mapped to illegal pblock %llu " - "(length %d)", (unsigned long) map->m_lblk, - map->m_pblk, map->m_len); + ext4_error_inode(inode, func, line, map->m_pblk, + "lblock %lu mapped to illegal pblock " + "(length %d)", (unsigned long) map->m_lblk, + map->m_len); return -EIO; } return 0; } #define check_block_validity(inode, map) \ - __check_block_validity((inode), __func__, (map)) + __check_block_validity((inode), __func__, __LINE__, (map)) /* * Return the number of contiguous dirty pages in a given inode @@ -4471,9 +4475,8 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode, * (should be rare). */ if (!bh) { - EXT4_ERROR_INODE(inode, - "Read failure block=%llu", - (unsigned long long) nr); + EXT4_ERROR_INODE_BLOCK(inode, nr, + "Read failure"); continue; } @@ -4788,8 +4791,8 @@ static int __ext4_get_inode_loc(struct inode *inode, bh = sb_getblk(sb, block); if (!bh) { - EXT4_ERROR_INODE(inode, "unable to read inode block - " - "block %llu", block); + EXT4_ERROR_INODE_BLOCK(inode, block, + "unable to read itable block"); return -EIO; } if (!buffer_uptodate(bh)) { @@ -4887,8 +4890,8 @@ make_io: submit_bh(READ_META, bh); wait_on_buffer(bh); if (!buffer_uptodate(bh)) { - EXT4_ERROR_INODE(inode, "unable to read inode " - "block %llu", block); + EXT4_ERROR_INODE_BLOCK(inode, block, + "unable to read itable block"); brelse(bh); return -EIO; } @@ -5389,9 +5392,8 @@ int ext4_write_inode(struct inode *inode, struct writeback_control *wbc) if (wbc->sync_mode == WB_SYNC_ALL) sync_dirty_buffer(iloc.bh); if (buffer_req(iloc.bh) && !buffer_uptodate(iloc.bh)) { - EXT4_ERROR_INODE(inode, - "IO error syncing inode (block=%llu)", - (unsigned long long) iloc.bh->b_blocknr); + EXT4_ERROR_INODE_BLOCK(inode, iloc.bh->b_blocknr, + "IO error syncing inode"); err = -EIO; } brelse(iloc.bh); diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c index 52abfa12762..5f1ed9fc913 100644 --- a/fs/ext4/move_extent.c +++ b/fs/ext4/move_extent.c @@ -148,17 +148,17 @@ mext_next_extent(struct inode *inode, struct ext4_ext_path *path, */ static int mext_check_null_inode(struct inode *inode1, struct inode *inode2, - const char *function) + const char *function, unsigned int line) { int ret = 0; if (inode1 == NULL) { - __ext4_error(inode2->i_sb, function, + __ext4_error(inode2->i_sb, function, line, "Both inodes should not be NULL: " "inode1 NULL inode2 %lu", inode2->i_ino); ret = -EIO; } else if (inode2 == NULL) { - __ext4_error(inode1->i_sb, function, + __ext4_error(inode1->i_sb, function, line, "Both inodes should not be NULL: " "inode1 %lu inode2 NULL", inode1->i_ino); ret = -EIO; @@ -1084,7 +1084,7 @@ mext_inode_double_lock(struct inode *inode1, struct inode *inode2) BUG_ON(inode1 == NULL && inode2 == NULL); - ret = mext_check_null_inode(inode1, inode2, __func__); + ret = mext_check_null_inode(inode1, inode2, __func__, __LINE__); if (ret < 0) goto out; @@ -1121,7 +1121,7 @@ mext_inode_double_unlock(struct inode *inode1, struct inode *inode2) BUG_ON(inode1 == NULL && inode2 == NULL); - ret = mext_check_null_inode(inode1, inode2, __func__); + ret = mext_check_null_inode(inode1, inode2, __func__, __LINE__); if (ret < 0) goto out; diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 87db5ecfccb..bcf74b31d01 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -262,7 +262,7 @@ handle_t *ext4_journal_start_sb(struct super_block *sb, int nblocks) * that sync() will call the filesystem's write_super callback if * appropriate. */ -int __ext4_journal_stop(const char *where, handle_t *handle) +int __ext4_journal_stop(const char *where, unsigned int line, handle_t *handle) { struct super_block *sb; int err; @@ -279,7 +279,7 @@ int __ext4_journal_stop(const char *where, handle_t *handle) if (!err) err = rc; if (err) - __ext4_std_error(sb, where, err); + __ext4_std_error(sb, where, line, err); return err; } @@ -350,12 +350,13 @@ static void ext4_handle_error(struct super_block *sb) } void __ext4_error(struct super_block *sb, const char *function, - const char *fmt, ...) + unsigned int line, const char *fmt, ...) { va_list args; va_start(args, fmt); - printk(KERN_CRIT "EXT4-fs error (device %s): %s: ", sb->s_id, function); + printk(KERN_CRIT "EXT4-fs error (device %s): %s:%d: comm %s: ", + sb->s_id, function, line, current->comm); vprintk(fmt, args); printk("\n"); va_end(args); @@ -363,14 +364,18 @@ void __ext4_error(struct super_block *sb, const char *function, ext4_handle_error(sb); } -void ext4_error_inode(const char *function, struct inode *inode, +void ext4_error_inode(struct inode *inode, const char *function, + unsigned int line, ext4_fsblk_t block, const char *fmt, ...) { va_list args; va_start(args, fmt); - printk(KERN_CRIT "EXT4-fs error (device %s): %s: inode #%lu: (comm %s) ", - inode->i_sb->s_id, function, inode->i_ino, current->comm); + printk(KERN_CRIT "EXT4-fs error (device %s): %s:%d: inode #%lu: ", + inode->i_sb->s_id, function, line, inode->i_ino); + if (block) + printk("block %llu: ", block); + printk("comm %s: ", current->comm); vprintk(fmt, args); printk("\n"); va_end(args); @@ -378,8 +383,8 @@ void ext4_error_inode(const char *function, struct inode *inode, ext4_handle_error(inode->i_sb); } -void ext4_error_file(const char *function, struct file *file, - const char *fmt, ...) +void ext4_error_file(struct file *file, const char *function, + unsigned int line, const char *fmt, ...) { va_list args; struct inode *inode = file->f_dentry->d_inode; @@ -390,8 +395,10 @@ void ext4_error_file(const char *function, struct file *file, if (!path) path = "(unknown)"; printk(KERN_CRIT - "EXT4-fs error (device %s): %s: inode #%lu (comm %s path %s): ", - inode->i_sb->s_id, function, inode->i_ino, current->comm, path); + "EXT4-fs error (device %s): %s:%d: inode #%lu " + "(comm %s path %s): ", + inode->i_sb->s_id, function, line, inode->i_ino, + current->comm, path); vprintk(fmt, args); printk("\n"); va_end(args); @@ -436,7 +443,8 @@ static const char *ext4_decode_error(struct super_block *sb, int errno, /* __ext4_std_error decodes expected errors from journaling functions * automatically and invokes the appropriate error response. */ -void __ext4_std_error(struct super_block *sb, const char *function, int errno) +void __ext4_std_error(struct super_block *sb, const char *function, + unsigned int line, int errno) { char nbuf[16]; const char *errstr; @@ -449,8 +457,8 @@ void __ext4_std_error(struct super_block *sb, const char *function, int errno) return; errstr = ext4_decode_error(sb, errno, nbuf); - printk(KERN_CRIT "EXT4-fs error (device %s) in %s: %s\n", - sb->s_id, function, errstr); + printk(KERN_CRIT "EXT4-fs error (device %s) in %s:%d: %s\n", + sb->s_id, function, line, errstr); ext4_handle_error(sb); } @@ -466,12 +474,13 @@ void __ext4_std_error(struct super_block *sb, const char *function, int errno) */ void __ext4_abort(struct super_block *sb, const char *function, - const char *fmt, ...) + unsigned int line, const char *fmt, ...) { va_list args; va_start(args, fmt); - printk(KERN_CRIT "EXT4-fs error (device %s): %s: ", sb->s_id, function); + printk(KERN_CRIT "EXT4-fs error (device %s): %s:%d: ", sb->s_id, + function, line); vprintk(fmt, args); printk("\n"); va_end(args); @@ -503,13 +512,13 @@ void ext4_msg (struct super_block * sb, const char *prefix, } void __ext4_warning(struct super_block *sb, const char *function, - const char *fmt, ...) + unsigned int line, const char *fmt, ...) { va_list args; va_start(args, fmt); - printk(KERN_WARNING "EXT4-fs warning (device %s): %s: ", - sb->s_id, function); + printk(KERN_WARNING "EXT4-fs warning (device %s): %s:%d: ", + sb->s_id, function, line); vprintk(fmt, args); printk("\n"); va_end(args); From 1c13d5c0872870cca3e612aa045d492ead9ab004 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 27 Jul 2010 11:56:03 -0400 Subject: [PATCH 12/40] ext4: Save error information to the superblock for analysis Save number of file system errors, and the time function name, line number, block number, and inode number of the first and most recent errors reported on the file system in the superblock. Signed-off-by: "Theodore Ts'o" --- fs/ext4/block_validity.c | 8 +++- fs/ext4/ext4.h | 17 ++++++++- fs/ext4/ext4_jbd2.c | 5 +++ fs/ext4/inode.c | 2 + fs/ext4/super.c | 80 ++++++++++++++++++++++++++++++---------- 5 files changed, 90 insertions(+), 22 deletions(-) diff --git a/fs/ext4/block_validity.c b/fs/ext4/block_validity.c index 5b6973fbf1b..3db5084db9b 100644 --- a/fs/ext4/block_validity.c +++ b/fs/ext4/block_validity.c @@ -229,16 +229,20 @@ int ext4_data_block_valid(struct ext4_sb_info *sbi, ext4_fsblk_t start_blk, if ((start_blk <= le32_to_cpu(sbi->s_es->s_first_data_block)) || (start_blk + count < start_blk) || - (start_blk + count > ext4_blocks_count(sbi->s_es))) + (start_blk + count > ext4_blocks_count(sbi->s_es))) { + sbi->s_es->s_last_error_block = cpu_to_le64(start_blk); return 0; + } while (n) { entry = rb_entry(n, struct ext4_system_zone, node); if (start_blk + count - 1 < entry->start_blk) n = n->rb_left; else if (start_blk >= (entry->start_blk + entry->count)) n = n->rb_right; - else + else { + sbi->s_es->s_last_error_block = cpu_to_le64(start_blk); return 0; + } } return 1; } diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 088938148f5..6b96125e725 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1011,9 +1011,24 @@ struct ext4_super_block { snapshot's future use */ __le32 s_snapshot_list; /* inode number of the head of the on-disk snapshot list */ - __u32 s_reserved[155]; /* Padding to the end of the block */ +#define EXT4_S_ERR_START offsetof(struct ext4_super_block, s_error_count) + __le32 s_error_count; /* number of fs errors */ + __le32 s_first_error_time; /* first time an error happened */ + __le32 s_first_error_ino; /* inode involved in first error */ + __le64 s_first_error_block; /* block involved of first error */ + __u8 s_first_error_func[32]; /* function where the error happened */ + __le32 s_first_error_line; /* line number where error happened */ + __le32 s_last_error_time; /* most recent time of an error */ + __le32 s_last_error_ino; /* inode involved in last error */ + __le32 s_last_error_line; /* line number where error happened */ + __le64 s_last_error_block; /* block involved of last error */ + __u8 s_last_error_func[32]; /* function where the error happened */ +#define EXT4_S_ERR_END offsetof(struct ext4_super_block, s_reserved) + __le32 s_reserved[128]; /* Padding to the end of the block */ }; +#define EXT4_S_ERR_LEN (EXT4_S_ERR_END - EXT4_S_ERR_START) + #ifdef __KERNEL__ /* diff --git a/fs/ext4/ext4_jbd2.c b/fs/ext4/ext4_jbd2.c index 23425cd68da..6e272ef6ba9 100644 --- a/fs/ext4/ext4_jbd2.c +++ b/fs/ext4/ext4_jbd2.c @@ -134,6 +134,11 @@ int __ext4_handle_dirty_metadata(const char *where, unsigned int line, if (inode && inode_needs_sync(inode)) { sync_dirty_buffer(bh); if (buffer_req(bh) && !buffer_uptodate(bh)) { + struct ext4_super_block *es; + + es = EXT4_SB(inode->i_sb)->s_es; + es->s_last_error_block = + cpu_to_le64(bh->b_blocknr); ext4_error_inode(inode, where, line, bh->b_blocknr, "IO error syncing itable block"); diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 69ea663ef03..755ba868223 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -341,6 +341,7 @@ static int __ext4_check_blockref(const char *function, unsigned int line, struct inode *inode, __le32 *p, unsigned int max) { + struct ext4_super_block *es = EXT4_SB(inode->i_sb)->s_es; __le32 *bref = p; unsigned int blk; @@ -349,6 +350,7 @@ static int __ext4_check_blockref(const char *function, unsigned int line, if (blk && unlikely(!ext4_data_block_valid(EXT4_SB(inode->i_sb), blk, 1))) { + es->s_last_error_block = cpu_to_le64(blk); ext4_error_inode(inode, function, line, blk, "invalid block"); return -EIO; diff --git a/fs/ext4/super.c b/fs/ext4/super.c index bcf74b31d01..a94d3f56898 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -307,6 +307,35 @@ void ext4_journal_abort_handle(const char *caller, unsigned int line, jbd2_journal_abort_handle(handle); } +static void __save_error_info(struct super_block *sb, const char *func, + unsigned int line) +{ + struct ext4_super_block *es = EXT4_SB(sb)->s_es; + + EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS; + es->s_state |= cpu_to_le16(EXT4_ERROR_FS); + es->s_last_error_time = cpu_to_le32(get_seconds()); + strncpy(es->s_last_error_func, func, sizeof(es->s_last_error_func)); + es->s_last_error_line = cpu_to_le32(line); + if (!es->s_first_error_time) { + es->s_first_error_time = es->s_last_error_time; + strncpy(es->s_first_error_func, func, + sizeof(es->s_first_error_func)); + es->s_first_error_line = cpu_to_le32(line); + es->s_first_error_ino = es->s_last_error_ino; + es->s_first_error_block = es->s_last_error_block; + } + es->s_error_count = cpu_to_le32(le32_to_cpu(es->s_error_count) + 1); +} + +static void save_error_info(struct super_block *sb, const char *func, + unsigned int line) +{ + __save_error_info(sb, func, line); + ext4_commit_super(sb, 1); +} + + /* Deal with the reporting of failure conditions on a filesystem such as * inconsistencies detected or read IO failures. * @@ -324,11 +353,6 @@ void ext4_journal_abort_handle(const char *caller, unsigned int line, static void ext4_handle_error(struct super_block *sb) { - struct ext4_super_block *es = EXT4_SB(sb)->s_es; - - EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS; - es->s_state |= cpu_to_le16(EXT4_ERROR_FS); - if (sb->s_flags & MS_RDONLY) return; @@ -343,7 +367,6 @@ static void ext4_handle_error(struct super_block *sb) ext4_msg(sb, KERN_CRIT, "Remounting filesystem read-only"); sb->s_flags |= MS_RDONLY; } - ext4_commit_super(sb, 1); if (test_opt(sb, ERRORS_PANIC)) panic("EXT4-fs (device %s): panic forced after error\n", sb->s_id); @@ -369,7 +392,11 @@ void ext4_error_inode(struct inode *inode, const char *function, const char *fmt, ...) { va_list args; + struct ext4_super_block *es = EXT4_SB(inode->i_sb)->s_es; + es->s_last_error_ino = cpu_to_le32(inode->i_ino); + es->s_last_error_block = cpu_to_le64(block); + save_error_info(inode->i_sb, function, line); va_start(args, fmt); printk(KERN_CRIT "EXT4-fs error (device %s): %s:%d: inode #%lu: ", inode->i_sb->s_id, function, line, inode->i_ino); @@ -387,9 +414,13 @@ void ext4_error_file(struct file *file, const char *function, unsigned int line, const char *fmt, ...) { va_list args; + struct ext4_super_block *es; struct inode *inode = file->f_dentry->d_inode; char pathname[80], *path; + es = EXT4_SB(inode->i_sb)->s_es; + es->s_last_error_ino = cpu_to_le32(inode->i_ino); + save_error_info(inode->i_sb, function, line); va_start(args, fmt); path = d_path(&(file->f_path), pathname, sizeof(pathname)); if (!path) @@ -459,6 +490,7 @@ void __ext4_std_error(struct super_block *sb, const char *function, errstr = ext4_decode_error(sb, errno, nbuf); printk(KERN_CRIT "EXT4-fs error (device %s) in %s:%d: %s\n", sb->s_id, function, line, errstr); + save_error_info(sb, function, line); ext4_handle_error(sb); } @@ -478,6 +510,7 @@ void __ext4_abort(struct super_block *sb, const char *function, { va_list args; + save_error_info(sb, function, line); va_start(args, fmt); printk(KERN_CRIT "EXT4-fs error (device %s): %s:%d: ", sb->s_id, function, line); @@ -485,18 +518,16 @@ void __ext4_abort(struct super_block *sb, const char *function, printk("\n"); va_end(args); + if ((sb->s_flags & MS_RDONLY) == 0) { + ext4_msg(sb, KERN_CRIT, "Remounting filesystem read-only"); + sb->s_flags |= MS_RDONLY; + EXT4_SB(sb)->s_mount_flags |= EXT4_MF_FS_ABORTED; + if (EXT4_SB(sb)->s_journal) + jbd2_journal_abort(EXT4_SB(sb)->s_journal, -EIO); + save_error_info(sb, function, line); + } if (test_opt(sb, ERRORS_PANIC)) panic("EXT4-fs panic from previous error\n"); - - if (sb->s_flags & MS_RDONLY) - return; - - ext4_msg(sb, KERN_CRIT, "Remounting filesystem read-only"); - EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS; - sb->s_flags |= MS_RDONLY; - EXT4_SB(sb)->s_mount_flags |= EXT4_MF_FS_ABORTED; - if (EXT4_SB(sb)->s_journal) - jbd2_journal_abort(EXT4_SB(sb)->s_journal, -EIO); } void ext4_msg (struct super_block * sb, const char *prefix, @@ -534,6 +565,9 @@ __acquires(bitlock) va_list args; struct ext4_super_block *es = EXT4_SB(sb)->s_es; + es->s_last_error_ino = cpu_to_le32(ino); + es->s_last_error_block = cpu_to_le64(block); + __save_error_info(sb, function, line); va_start(args, fmt); printk(KERN_CRIT "EXT4-fs error (device %s): %s:%d: group %u", sb->s_id, function, line, grp); @@ -546,11 +580,10 @@ __acquires(bitlock) va_end(args); if (test_opt(sb, ERRORS_CONT)) { - EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS; - es->s_state |= cpu_to_le16(EXT4_ERROR_FS); ext4_commit_super(sb, 0); return; } + ext4_unlock_group(sb, grp); ext4_handle_error(sb); /* @@ -3332,8 +3365,17 @@ static int ext4_load_journal(struct super_block *sb, if (!EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER)) err = jbd2_journal_wipe(journal, !really_read_only); - if (!err) + if (!err) { + char *save = kmalloc(EXT4_S_ERR_LEN, GFP_KERNEL); + if (save) + memcpy(save, ((char *) es) + + EXT4_S_ERR_START, EXT4_S_ERR_LEN); err = jbd2_journal_load(journal); + if (save) + memcpy(((char *) es) + EXT4_S_ERR_START, + save, EXT4_S_ERR_LEN); + kfree(save); + } if (err) { ext4_msg(sb, KERN_ERR, "error loading journal"); From 66e61a9e9504f61b9a928c9055368c81da613a50 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 27 Jul 2010 11:56:04 -0400 Subject: [PATCH 13/40] ext4: Once a day, printk file system error information to dmesg This allows us to grab any file system error messages by scraping /var/log/messages. This will make it easy for us to do error analysis across the very large number of machines as we deploy ext4 across the fleet. Signed-off-by: "Theodore Ts'o" --- fs/ext4/ext4.h | 3 +++ fs/ext4/super.c | 59 +++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 62 insertions(+) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 6b96125e725..5d3d768d950 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1166,6 +1166,9 @@ struct ext4_sb_info { /* workqueue for dio unwritten */ struct workqueue_struct *dio_unwritten_wq; + + /* timer for periodic error stats printing */ + struct timer_list s_err_report; }; static inline struct ext4_sb_info *EXT4_SB(struct super_block *sb) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index a94d3f56898..ed00c14d708 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -325,6 +325,12 @@ static void __save_error_info(struct super_block *sb, const char *func, es->s_first_error_ino = es->s_last_error_ino; es->s_first_error_block = es->s_last_error_block; } + /* + * Start the daily error reporting function if it hasn't been + * started already + */ + if (!es->s_error_count) + mod_timer(&EXT4_SB(sb)->s_err_report, jiffies + 24*60*60*HZ); es->s_error_count = cpu_to_le32(le32_to_cpu(es->s_error_count) + 1); } @@ -2480,6 +2486,53 @@ static int ext4_feature_set_ok(struct super_block *sb, int readonly) return 1; } +/* + * This function is called once a day if we have errors logged + * on the file system + */ +static void print_daily_error_info(unsigned long arg) +{ + struct super_block *sb = (struct super_block *) arg; + struct ext4_sb_info *sbi; + struct ext4_super_block *es; + + sbi = EXT4_SB(sb); + es = sbi->s_es; + + if (es->s_error_count) + ext4_msg(sb, KERN_NOTICE, "error count: %u", + le32_to_cpu(es->s_error_count)); + if (es->s_first_error_time) { + printk(KERN_NOTICE "EXT4-fs (%s): initial error at %u: %.*s:%d", + sb->s_id, le32_to_cpu(es->s_first_error_time), + (int) sizeof(es->s_first_error_func), + es->s_first_error_func, + le32_to_cpu(es->s_first_error_line)); + if (es->s_first_error_ino) + printk(": inode %u", + le32_to_cpu(es->s_first_error_ino)); + if (es->s_first_error_block) + printk(": block %llu", (unsigned long long) + le64_to_cpu(es->s_first_error_block)); + printk("\n"); + } + if (es->s_last_error_time) { + printk(KERN_NOTICE "EXT4-fs (%s): last error at %u: %.*s:%d", + sb->s_id, le32_to_cpu(es->s_last_error_time), + (int) sizeof(es->s_last_error_func), + es->s_last_error_func, + le32_to_cpu(es->s_last_error_line)); + if (es->s_last_error_ino) + printk(": inode %u", + le32_to_cpu(es->s_last_error_ino)); + if (es->s_last_error_block) + printk(": block %llu", (unsigned long long) + le64_to_cpu(es->s_last_error_block)); + printk("\n"); + } + mod_timer(&sbi->s_err_report, jiffies + 24*60*60*HZ); /* Once a day */ +} + static int ext4_fill_super(struct super_block *sb, void *data, int silent) __releases(kernel_lock) __acquires(kernel_lock) @@ -3083,6 +3136,12 @@ no_journal: ext4_msg(sb, KERN_INFO, "mounted filesystem with%s. " "Opts: %s", descr, orig_data); + init_timer(&sbi->s_err_report); + sbi->s_err_report.function = print_daily_error_info; + sbi->s_err_report.data = (unsigned long) sb; + if (es->s_error_count) + mod_timer(&sbi->s_err_report, jiffies + 300*HZ); /* 5 minutes */ + lock_kernel(); kfree(orig_data); return 0; From 89eeddf03327e19cfcbb18efa98e5470e2f5c563 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 27 Jul 2010 11:56:04 -0400 Subject: [PATCH 14/40] ext4: Define s_jnl_backup_type in superblock This has been in use by e2fsprogs for a while; define it to keep the super block fields in sync. Signed-off-by: "Theodore Ts'o" --- fs/ext4/ext4.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 5d3d768d950..4c7d4727d6b 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -984,7 +984,7 @@ struct ext4_super_block { __le32 s_last_orphan; /* start of list of inodes to delete */ __le32 s_hash_seed[4]; /* HTREE hash seed */ __u8 s_def_hash_version; /* Default hash version to use */ - __u8 s_reserved_char_pad; + __u8 s_jnl_backup_type; __le16 s_desc_size; /* size of group descriptor */ /*100*/ __le32 s_default_mount_opts; __le32 s_first_meta_bg; /* First metablock block group */ @@ -1002,7 +1002,7 @@ struct ext4_super_block { __le64 s_mmp_block; /* Block for multi-mount protection */ __le32 s_raid_stripe_width; /* blocks on all data disks (N*stride)*/ __u8 s_log_groups_per_flex; /* FLEX_BG group size */ - __u8 s_reserved_char_pad2; + __u8 s_reserved_char_pad; __le16 s_reserved_pad; __le64 s_kbytes_written; /* nr of lifetime kilobytes written */ __le32 s_snapshot_inum; /* Inode number of active snapshot */ From e5880d76aea443b04e07da19830da0f6f7494eef Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 27 Jul 2010 11:56:04 -0400 Subject: [PATCH 15/40] ext4: fix potential NULL dereference while tracing The allocation_context pointer can be NULL. Signed-off-by: "Theodore Ts'o" --- fs/ext4/mballoc.c | 4 ++-- include/trace/events/ext4.h | 20 ++++++++++++-------- 2 files changed, 14 insertions(+), 10 deletions(-) diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index 3dfad95f0f9..8b3b9344a59 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -3575,7 +3575,7 @@ ext4_mb_release_inode_pa(struct ext4_buddy *e4b, struct buffer_head *bitmap_bh, trace_ext4_mballoc_discard(ac); } - trace_ext4_mb_release_inode_pa(ac, pa, grp_blk_start + bit, + trace_ext4_mb_release_inode_pa(sb, ac, pa, grp_blk_start + bit, next - bit); mb_free_blocks(pa->pa_inode, e4b, bit, next - bit); bit = next + 1; @@ -3606,7 +3606,7 @@ ext4_mb_release_group_pa(struct ext4_buddy *e4b, ext4_group_t group; ext4_grpblk_t bit; - trace_ext4_mb_release_group_pa(ac, pa); + trace_ext4_mb_release_group_pa(sb, ac, pa); BUG_ON(pa->pa_deleted == 0); ext4_get_group_no_and_offset(sb, pa->pa_pstart, &group, &bit); BUG_ON(group != e4b->bd_group && pa->pa_len != 0); diff --git a/include/trace/events/ext4.h b/include/trace/events/ext4.h index f3865c7b416..01e9e0076a9 100644 --- a/include/trace/events/ext4.h +++ b/include/trace/events/ext4.h @@ -395,11 +395,12 @@ DEFINE_EVENT(ext4__mb_new_pa, ext4_mb_new_group_pa, ); TRACE_EVENT(ext4_mb_release_inode_pa, - TP_PROTO(struct ext4_allocation_context *ac, + TP_PROTO(struct super_block *sb, + struct ext4_allocation_context *ac, struct ext4_prealloc_space *pa, unsigned long long block, unsigned int count), - TP_ARGS(ac, pa, block, count), + TP_ARGS(sb, ac, pa, block, count), TP_STRUCT__entry( __field( dev_t, dev ) @@ -410,8 +411,9 @@ TRACE_EVENT(ext4_mb_release_inode_pa, ), TP_fast_assign( - __entry->dev = ac->ac_sb->s_dev; - __entry->ino = ac->ac_inode->i_ino; + __entry->dev = sb->s_dev; + __entry->ino = (ac && ac->ac_inode) ? + ac->ac_inode->i_ino : 0; __entry->block = block; __entry->count = count; ), @@ -422,10 +424,11 @@ TRACE_EVENT(ext4_mb_release_inode_pa, ); TRACE_EVENT(ext4_mb_release_group_pa, - TP_PROTO(struct ext4_allocation_context *ac, + TP_PROTO(struct super_block *sb, + struct ext4_allocation_context *ac, struct ext4_prealloc_space *pa), - TP_ARGS(ac, pa), + TP_ARGS(sb, ac, pa), TP_STRUCT__entry( __field( dev_t, dev ) @@ -436,8 +439,9 @@ TRACE_EVENT(ext4_mb_release_group_pa, ), TP_fast_assign( - __entry->dev = ac->ac_sb->s_dev; - __entry->ino = ac->ac_inode->i_ino; + __entry->dev = sb->s_dev; + __entry->ino = (ac && ac->ac_inode) ? + ac->ac_inode->i_ino : 0; __entry->pa_pstart = pa->pa_pstart; __entry->pa_len = pa->pa_len; ), From a271fe8527fe9637bdd82c97123b1356940dd84b Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Tue, 27 Jul 2010 11:56:04 -0400 Subject: [PATCH 16/40] ext4: Remove unnecessary casts of private_data Signed-off-by: Joe Perches Signed-off-by: "Theodore Ts'o" --- fs/ext4/dir.c | 2 +- fs/ext4/mballoc.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/ext4/dir.c b/fs/ext4/dir.c index 62e8af04ed1..374510f72ba 100644 --- a/fs/ext4/dir.c +++ b/fs/ext4/dir.c @@ -344,7 +344,7 @@ int ext4_htree_store_dirent(struct file *dir_file, __u32 hash, struct dir_private_info *info; int len; - info = (struct dir_private_info *) dir_file->private_data; + info = dir_file->private_data; p = &info->root.rb_node; /* Create and allocate the fname structure */ diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index 8b3b9344a59..84185d21500 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -2219,7 +2219,7 @@ static int ext4_mb_seq_groups_open(struct inode *inode, struct file *file) rc = seq_open(file, &ext4_mb_seq_groups_ops); if (rc == 0) { - struct seq_file *m = (struct seq_file *)file->private_data; + struct seq_file *m = file->private_data; m->private = sb; } return rc; From 40389687382bf0ae71458e7c0f828137a438a956 Mon Sep 17 00:00:00 2001 From: Amir G Date: Tue, 27 Jul 2010 11:56:05 -0400 Subject: [PATCH 17/40] ext4: Fix block bitmap inconsistencies after a crash when deleting files We have experienced bitmap inconsistencies after crash during file delete under heavy load. The crash is not file system related and I the following patch in ext4_free_branches() fixes the recovery problem. If the transaction is restarted and there is a crash before the new transaction is committed, then after recovery, the blocks that this indirect block points to have been freed, but the indirect block itself has not been freed and may still point to some of the free blocks (because of the ext4_forget()). So ext4_forget() should be called inside ext4_free_blocks() to avoid this problem. Signed-off-by: Amir Goldstein Signed-off-by: "Theodore Ts'o" --- fs/ext4/inode.c | 35 +++++++++++++---------------------- 1 file changed, 13 insertions(+), 22 deletions(-) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 755ba868223..699d1d01c5d 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -4489,27 +4489,6 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode, (__le32 *) bh->b_data + addr_per_block, depth); - /* - * We've probably journalled the indirect block several - * times during the truncate. But it's no longer - * needed and we now drop it from the transaction via - * jbd2_journal_revoke(). - * - * That's easy if it's exclusively part of this - * transaction. But if it's part of the committing - * transaction then jbd2_journal_forget() will simply - * brelse() it. That means that if the underlying - * block is reallocated in ext4_get_block(), - * unmap_underlying_metadata() will find this block - * and will try to get rid of it. damn, damn. - * - * If this block has already been committed to the - * journal, a revoke record will be written. And - * revoke records must be emitted *before* clearing - * this block's bit in the bitmaps. - */ - ext4_forget(handle, 1, inode, bh, bh->b_blocknr); - /* * Everything below this this pointer has been * released. Now let this top-of-subtree go. @@ -4534,8 +4513,20 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode, blocks_for_truncate(inode)); } + /* + * The forget flag here is critical because if + * we are journaling (and not doing data + * journaling), we have to make sure a revoke + * record is written to prevent the journal + * replay from overwriting the (former) + * indirect block if it gets reallocated as a + * data block. This must happen in the same + * transaction where the data blocks are + * actually freed. + */ ext4_free_blocks(handle, inode, 0, nr, 1, - EXT4_FREE_BLOCKS_METADATA); + EXT4_FREE_BLOCKS_METADATA| + EXT4_FREE_BLOCKS_FORGET); if (parent_bh) { /* From 47def82672b3ba4e7c5e9a4fe48a556f8684d0d6 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 27 Jul 2010 11:56:05 -0400 Subject: [PATCH 18/40] jbd2: Remove __GFP_NOFAIL from jbd2 layer __GFP_NOFAIL is going away, so add our own retry loop. Also add jbd2__journal_start() and jbd2__journal_restart() which take a gfp mask, so that file systems can optionally (re)start transaction handles using GFP_KERNEL. If they do this, then they need to be prepared to handle receiving an PTR_ERR(-ENOMEM) error, and be ready to reflect that error up to userspace. Signed-off-by: "Theodore Ts'o" --- fs/jbd2/journal.c | 15 ++++++++--- fs/jbd2/transaction.c | 61 +++++++++++++++++++++++++++++-------------- include/linux/jbd2.h | 4 ++- 3 files changed, 57 insertions(+), 23 deletions(-) diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c index f7bf15787d6..a79d3345b55 100644 --- a/fs/jbd2/journal.c +++ b/fs/jbd2/journal.c @@ -41,6 +41,7 @@ #include #include #include +#include #define CREATE_TRACE_POINTS #include @@ -48,8 +49,6 @@ #include #include -EXPORT_SYMBOL(jbd2_journal_start); -EXPORT_SYMBOL(jbd2_journal_restart); EXPORT_SYMBOL(jbd2_journal_extend); EXPORT_SYMBOL(jbd2_journal_stop); EXPORT_SYMBOL(jbd2_journal_lock_updates); @@ -311,7 +310,17 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction, */ J_ASSERT_BH(bh_in, buffer_jbddirty(bh_in)); - new_bh = alloc_buffer_head(GFP_NOFS|__GFP_NOFAIL); +retry_alloc: + new_bh = alloc_buffer_head(GFP_NOFS); + if (!new_bh) { + /* + * Failure is not an option, but __GFP_NOFAIL is going + * away; so we retry ourselves here. + */ + congestion_wait(BLK_RW_ASYNC, HZ/50); + goto retry_alloc; + } + /* keep subsequent assertions sane */ new_bh->b_state = 0; init_buffer(new_bh, NULL, NULL); diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c index e214d68620a..001e95fb0fe 100644 --- a/fs/jbd2/transaction.c +++ b/fs/jbd2/transaction.c @@ -26,6 +26,8 @@ #include #include #include +#include +#include static void __jbd2_journal_temp_unlink_buffer(struct journal_head *jh); @@ -83,30 +85,38 @@ jbd2_get_transaction(journal_t *journal, transaction_t *transaction) * transaction's buffer credits. */ -static int start_this_handle(journal_t *journal, handle_t *handle) +static int start_this_handle(journal_t *journal, handle_t *handle, + int gfp_mask) { transaction_t *transaction; int needed; int nblocks = handle->h_buffer_credits; transaction_t *new_transaction = NULL; - int ret = 0; unsigned long ts = jiffies; if (nblocks > journal->j_max_transaction_buffers) { printk(KERN_ERR "JBD: %s wants too many credits (%d > %d)\n", current->comm, nblocks, journal->j_max_transaction_buffers); - ret = -ENOSPC; - goto out; + return -ENOSPC; } alloc_transaction: if (!journal->j_running_transaction) { - new_transaction = kzalloc(sizeof(*new_transaction), - GFP_NOFS|__GFP_NOFAIL); + new_transaction = kzalloc(sizeof(*new_transaction), gfp_mask); if (!new_transaction) { - ret = -ENOMEM; - goto out; + /* + * If __GFP_FS is not present, then we may be + * being called from inside the fs writeback + * layer, so we MUST NOT fail. Since + * __GFP_NOFAIL is going away, we will arrange + * to retry the allocation ourselves. + */ + if ((gfp_mask & __GFP_FS) == 0) { + congestion_wait(BLK_RW_ASYNC, HZ/50); + goto alloc_transaction; + } + return -ENOMEM; } } @@ -123,8 +133,8 @@ repeat_locked: if (is_journal_aborted(journal) || (journal->j_errno != 0 && !(journal->j_flags & JBD2_ACK_ERR))) { spin_unlock(&journal->j_state_lock); - ret = -EROFS; - goto out; + kfree(new_transaction); + return -EROFS; } /* Wait on the journal's transaction barrier if necessary */ @@ -240,10 +250,8 @@ repeat_locked: spin_unlock(&journal->j_state_lock); lock_map_acquire(&handle->h_lockdep_map); -out: - if (unlikely(new_transaction)) /* It's usually NULL */ - kfree(new_transaction); - return ret; + kfree(new_transaction); + return 0; } static struct lock_class_key jbd2_handle_key; @@ -278,7 +286,7 @@ static handle_t *new_handle(int nblocks) * * Return a pointer to a newly allocated handle, or NULL on failure */ -handle_t *jbd2_journal_start(journal_t *journal, int nblocks) +handle_t *jbd2__journal_start(journal_t *journal, int nblocks, int gfp_mask) { handle_t *handle = journal_current_handle(); int err; @@ -298,7 +306,7 @@ handle_t *jbd2_journal_start(journal_t *journal, int nblocks) current->journal_info = handle; - err = start_this_handle(journal, handle); + err = start_this_handle(journal, handle, gfp_mask); if (err < 0) { jbd2_free_handle(handle); current->journal_info = NULL; @@ -308,6 +316,15 @@ handle_t *jbd2_journal_start(journal_t *journal, int nblocks) out: return handle; } +EXPORT_SYMBOL(jbd2__journal_start); + + +handle_t *jbd2_journal_start(journal_t *journal, int nblocks) +{ + return jbd2__journal_start(journal, nblocks, GFP_NOFS); +} +EXPORT_SYMBOL(jbd2_journal_start); + /** * int jbd2_journal_extend() - extend buffer credits. @@ -394,8 +411,7 @@ out: * transaction capabable of guaranteeing the requested number of * credits. */ - -int jbd2_journal_restart(handle_t *handle, int nblocks) +int jbd2__journal_restart(handle_t *handle, int nblocks, int gfp_mask) { transaction_t *transaction = handle->h_transaction; journal_t *journal = transaction->t_journal; @@ -428,11 +444,18 @@ int jbd2_journal_restart(handle_t *handle, int nblocks) lock_map_release(&handle->h_lockdep_map); handle->h_buffer_credits = nblocks; - ret = start_this_handle(journal, handle); + ret = start_this_handle(journal, handle, gfp_mask); return ret; } +EXPORT_SYMBOL(jbd2__journal_restart); +int jbd2_journal_restart(handle_t *handle, int nblocks) +{ + return jbd2__journal_restart(handle, nblocks, GFP_NOFS); +} +EXPORT_SYMBOL(jbd2_journal_restart); + /** * void jbd2_journal_lock_updates () - establish a transaction barrier. * @journal: Journal to establish a barrier on. diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h index a4d2e9f7088..5a72bc75b27 100644 --- a/include/linux/jbd2.h +++ b/include/linux/jbd2.h @@ -1081,7 +1081,9 @@ static inline handle_t *journal_current_handle(void) */ extern handle_t *jbd2_journal_start(journal_t *, int nblocks); -extern int jbd2_journal_restart (handle_t *, int nblocks); +extern handle_t *jbd2__journal_start(journal_t *, int nblocks, int gfp_mask); +extern int jbd2_journal_restart(handle_t *, int nblocks); +extern int jbd2__journal_restart(handle_t *, int nblocks, int gfp_mask); extern int jbd2_journal_extend (handle_t *, int nblocks); extern int jbd2_journal_get_write_access(handle_t *, struct buffer_head *); extern int jbd2_journal_get_create_access (handle_t *, struct buffer_head *); From 5c521830cf3dfcf7638d409d8e02ed21020c064f Mon Sep 17 00:00:00 2001 From: Jiaying Zhang Date: Tue, 27 Jul 2010 11:56:05 -0400 Subject: [PATCH 19/40] ext4: Support discard requests when running in no-journal mode Issue discard request in ext4_free_blocks() when ext4 has no journal and is mounted with discard option. Signed-off-by: Jiaying Zhang Signed-off-by: "Theodore Ts'o" --- fs/ext4/mballoc.c | 37 +++++++++++++++++++++---------------- 1 file changed, 21 insertions(+), 16 deletions(-) diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index 84185d21500..5338b1ca64b 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -2558,6 +2558,22 @@ int ext4_mb_release(struct super_block *sb) return 0; } +static inline void ext4_issue_discard(struct super_block *sb, + ext4_group_t block_group, ext4_grpblk_t block, int count) +{ + int ret; + ext4_fsblk_t discard_block; + + discard_block = block + ext4_group_first_block_no(sb, block_group); + trace_ext4_discard_blocks(sb, + (unsigned long long) discard_block, count); + ret = sb_issue_discard(sb, discard_block, count); + if (ret == EOPNOTSUPP) { + ext4_warning(sb, "discard not supported, disabling"); + clear_opt(EXT4_SB(sb)->s_mount_opt, DISCARD); + } +} + /* * This function is called by the jbd2 layer once the commit has finished, * so we know we can free the blocks that were released with that commit. @@ -2577,22 +2593,9 @@ static void release_blocks_on_commit(journal_t *journal, transaction_t *txn) mb_debug(1, "gonna free %u blocks in group %u (0x%p):", entry->count, entry->group, entry); - if (test_opt(sb, DISCARD)) { - int ret; - ext4_fsblk_t discard_block; - - discard_block = entry->start_blk + - ext4_group_first_block_no(sb, entry->group); - trace_ext4_discard_blocks(sb, - (unsigned long long)discard_block, - entry->count); - ret = sb_issue_discard(sb, discard_block, entry->count); - if (ret == EOPNOTSUPP) { - ext4_warning(sb, - "discard not supported, disabling"); - clear_opt(EXT4_SB(sb)->s_mount_opt, DISCARD); - } - } + if (test_opt(sb, DISCARD)) + ext4_issue_discard(sb, entry->group, + entry->start_blk, entry->count); err = ext4_mb_load_buddy(sb, entry->group, &e4b); /* we expect to find existing buddy because it's pinned */ @@ -4639,6 +4642,8 @@ do_more: mb_clear_bits(bitmap_bh->b_data, bit, count); mb_free_blocks(inode, &e4b, bit, count); ext4_mb_return_to_preallocation(inode, &e4b, block, count); + if (test_opt(sb, DISCARD)) + ext4_issue_discard(sb, block_group, bit, count); } ret = ext4_free_blks_count(sb, gdp) + count; From 552ef8024f909d9b3a7442d0ab0d48a22de24e9e Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 27 Jul 2010 11:56:06 -0400 Subject: [PATCH 20/40] direct-io: move aio_complete into ->end_io Filesystems with unwritten extent support must not complete an AIO request until the transaction to convert the extent has been commited. That means the aio_complete calls needs to be moved into the ->end_io callback so that the filesystem can control when to call it exactly. This makes a bit of a mess out of dio_complete and the ->end_io callback prototype even more complicated. Signed-off-by: Christoph Hellwig Reviewed-by: Jan Kara Signed-off-by: "Theodore Ts'o" --- fs/direct-io.c | 26 ++++++++++++++------------ fs/ext4/inode.c | 10 +++++++--- fs/ocfs2/aops.c | 7 ++++++- fs/xfs/linux-2.6/xfs_aops.c | 7 ++++++- fs/xfs/linux-2.6/xfs_aops.h | 2 ++ include/linux/fs.h | 3 ++- 6 files changed, 37 insertions(+), 18 deletions(-) diff --git a/fs/direct-io.c b/fs/direct-io.c index 7600aacf531..a10cb91cade 100644 --- a/fs/direct-io.c +++ b/fs/direct-io.c @@ -218,7 +218,7 @@ static struct page *dio_get_page(struct dio *dio) * filesystems can use it to hold additional state between get_block calls and * dio_complete. */ -static int dio_complete(struct dio *dio, loff_t offset, int ret) +static int dio_complete(struct dio *dio, loff_t offset, int ret, bool is_async) { ssize_t transferred = 0; @@ -239,14 +239,6 @@ static int dio_complete(struct dio *dio, loff_t offset, int ret) transferred = dio->i_size - offset; } - if (dio->end_io && dio->result) - dio->end_io(dio->iocb, offset, transferred, - dio->map_bh.b_private); - - if (dio->flags & DIO_LOCKING) - /* lockdep: non-owner release */ - up_read_non_owner(&dio->inode->i_alloc_sem); - if (ret == 0) ret = dio->page_errors; if (ret == 0) @@ -254,6 +246,17 @@ static int dio_complete(struct dio *dio, loff_t offset, int ret) if (ret == 0) ret = transferred; + if (dio->end_io && dio->result) { + dio->end_io(dio->iocb, offset, transferred, + dio->map_bh.b_private, ret, is_async); + } else if (is_async) { + aio_complete(dio->iocb, ret, 0); + } + + if (dio->flags & DIO_LOCKING) + /* lockdep: non-owner release */ + up_read_non_owner(&dio->inode->i_alloc_sem); + return ret; } @@ -277,8 +280,7 @@ static void dio_bio_end_aio(struct bio *bio, int error) spin_unlock_irqrestore(&dio->bio_lock, flags); if (remaining == 0) { - int ret = dio_complete(dio, dio->iocb->ki_pos, 0); - aio_complete(dio->iocb, ret, 0); + dio_complete(dio, dio->iocb->ki_pos, 0, true); kfree(dio); } } @@ -1126,7 +1128,7 @@ direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode, spin_unlock_irqrestore(&dio->bio_lock, flags); if (ret2 == 0) { - ret = dio_complete(dio, offset, ret); + ret = dio_complete(dio, offset, ret, false); kfree(dio); } else BUG_ON(ret != -EIOCBQUEUED); diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 699d1d01c5d..609159e990d 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -3775,7 +3775,8 @@ static ext4_io_end_t *ext4_init_io_end (struct inode *inode, gfp_t flags) } static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset, - ssize_t size, void *private) + ssize_t size, void *private, int ret, + bool is_async) { ext4_io_end_t *io_end = iocb->private; struct workqueue_struct *wq; @@ -3784,7 +3785,7 @@ static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset, /* if not async direct IO or dio with 0 bytes write, just return */ if (!io_end || !size) - return; + goto out; ext_debug("ext4_end_io_dio(): io_end 0x%p" "for inode %lu, iocb 0x%p, offset %llu, size %llu\n", @@ -3795,7 +3796,7 @@ static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset, if (io_end->flag != EXT4_IO_UNWRITTEN){ ext4_free_io_end(io_end); iocb->private = NULL; - return; + goto out; } io_end->offset = offset; @@ -3812,6 +3813,9 @@ static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset, list_add_tail(&io_end->list, &ei->i_completed_io_list); spin_unlock_irqrestore(&ei->i_completed_io_lock, flags); iocb->private = NULL; +out: + if (is_async) + aio_complete(iocb, ret, 0); } static void ext4_end_io_buffer_write(struct buffer_head *bh, int uptodate) diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index 3623ca20cc1..1d2b1f156bc 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -609,7 +609,9 @@ bail: static void ocfs2_dio_end_io(struct kiocb *iocb, loff_t offset, ssize_t bytes, - void *private) + void *private, + int ret, + bool is_async) { struct inode *inode = iocb->ki_filp->f_path.dentry->d_inode; int level; @@ -623,6 +625,9 @@ static void ocfs2_dio_end_io(struct kiocb *iocb, if (!level) up_read(&inode->i_alloc_sem); ocfs2_rw_unlock(inode, level); + + if (is_async) + aio_complete(iocb, ret, 0); } /* diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c index 34640d6dbdc..5895aaf62ac 100644 --- a/fs/xfs/linux-2.6/xfs_aops.c +++ b/fs/xfs/linux-2.6/xfs_aops.c @@ -1599,7 +1599,9 @@ xfs_end_io_direct( struct kiocb *iocb, loff_t offset, ssize_t size, - void *private) + void *private, + int ret, + bool is_async) { xfs_ioend_t *ioend = iocb->private; @@ -1645,6 +1647,9 @@ xfs_end_io_direct( * against double-freeing. */ iocb->private = NULL; + + if (is_async) + aio_complete(iocb, ret, 0); } STATIC ssize_t diff --git a/fs/xfs/linux-2.6/xfs_aops.h b/fs/xfs/linux-2.6/xfs_aops.h index 4cfc6ea87df..9f566d92ae3 100644 --- a/fs/xfs/linux-2.6/xfs_aops.h +++ b/fs/xfs/linux-2.6/xfs_aops.h @@ -37,6 +37,8 @@ typedef struct xfs_ioend { size_t io_size; /* size of the extent */ xfs_off_t io_offset; /* offset in the file */ struct work_struct io_work; /* xfsdatad work queue */ + struct kiocb *io_iocb; + int io_result; } xfs_ioend_t; extern const struct address_space_operations xfs_address_space_operations; diff --git a/include/linux/fs.h b/include/linux/fs.h index 471e1ff5079..a0912f6075e 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -415,7 +415,8 @@ struct buffer_head; typedef int (get_block_t)(struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create); typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset, - ssize_t bytes, void *private); + ssize_t bytes, void *private, int ret, + bool is_async); /* * Attribute flags. These should be or-ed together to figure out what From 5b3ff237bef43b9e7fb7d1eb858e29b73fd664f9 Mon Sep 17 00:00:00 2001 From: "jiayingz@google.com (Jiaying Zhang)" <> Date: Tue, 27 Jul 2010 11:56:06 -0400 Subject: [PATCH 21/40] ext4: move aio completion after unwritten extent conversion This patch is to be applied upon Christoph's "direct-io: move aio_complete into ->end_io" patch. It adds iocb and result fields to struct ext4_io_end_t, so that we can call aio_complete from ext4_end_io_nolock() after the extent conversion has finished. I have verified with Christoph's aio-dio test that used to fail after a few runs on an original kernel but now succeeds on the patched kernel. See http://thread.gmane.org/gmane.comp.file-systems.ext4/19659 for details. Signed-off-by: Jiaying Zhang Signed-off-by: "Theodore Ts'o" --- fs/ext4/ext4.h | 4 +++- fs/ext4/inode.c | 17 ++++++++++++----- 2 files changed, 15 insertions(+), 6 deletions(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 4c7d4727d6b..fbb39478df2 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -170,13 +170,15 @@ struct mpage_da_data { }; #define EXT4_IO_UNWRITTEN 0x1 typedef struct ext4_io_end { - struct list_head list; /* per-file finished AIO list */ + struct list_head list; /* per-file finished IO list */ struct inode *inode; /* file being written to */ unsigned int flag; /* unwritten or not */ struct page *page; /* page struct for buffer write */ loff_t offset; /* offset in the file */ ssize_t size; /* size of the extent */ struct work_struct work; /* data work queue */ + struct kiocb *iocb; /* iocb struct for AIO */ + int result; /* error value for AIO */ } ext4_io_end_t; /* diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 609159e990d..46d2079373c 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -3668,6 +3668,8 @@ static int ext4_end_io_nolock(ext4_io_end_t *io) return ret; } + if (io->iocb) + aio_complete(io->iocb, io->result, 0); /* clear the DIO AIO unwritten flag */ io->flag = 0; return ret; @@ -3767,6 +3769,8 @@ static ext4_io_end_t *ext4_init_io_end (struct inode *inode, gfp_t flags) io->offset = 0; io->size = 0; io->page = NULL; + io->iocb = NULL; + io->result = 0; INIT_WORK(&io->work, ext4_end_io_work); INIT_LIST_HEAD(&io->list); } @@ -3796,12 +3800,18 @@ static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset, if (io_end->flag != EXT4_IO_UNWRITTEN){ ext4_free_io_end(io_end); iocb->private = NULL; - goto out; +out: + if (is_async) + aio_complete(iocb, ret, 0); + return; } io_end->offset = offset; io_end->size = size; - io_end->flag = EXT4_IO_UNWRITTEN; + if (is_async) { + io_end->iocb = iocb; + io_end->result = ret; + } wq = EXT4_SB(io_end->inode->i_sb)->dio_unwritten_wq; /* queue the work to convert unwritten extents to written */ @@ -3813,9 +3823,6 @@ static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset, list_add_tail(&io_end->list, &ei->i_completed_io_list); spin_unlock_irqrestore(&ei->i_completed_io_lock, flags); iocb->private = NULL; -out: - if (is_async) - aio_complete(iocb, ret, 0); } static void ext4_end_io_buffer_write(struct buffer_head *bh, int uptodate) From 506bf2d82165c09b179a5077e01037f6270a4db3 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Tue, 27 Jul 2010 11:56:06 -0400 Subject: [PATCH 22/40] ext4: allocate stripe-multiple IOs on stripe boundaries For some reason, today mballoc only allocates IOs which are exactly stripe-sized on a stripe boundary. If you have a multiple (say, a 128k IO on a 64k stripe) you may end up unaligned. It seems to me that a simple change to align stripe-multiple IOs on stripe boundaries would be a very good idea, unless this breaks some other mballoc heuristic for some reason... Reported-by: Mike Snitzer Signed-off-by: Eric Sandeen Signed-off-by: "Theodore Ts'o" --- fs/ext4/mballoc.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index 5338b1ca64b..a75de7d44dc 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -1822,8 +1822,7 @@ void ext4_mb_complex_scan_group(struct ext4_allocation_context *ac, /* * This is a special case for storages like raid5 - * we try to find stripe-aligned chunks for stripe-size requests - * XXX should do so at least for multiples of stripe size as well + * we try to find stripe-aligned chunks for stripe-size-multiple requests */ static noinline_for_stack void ext4_mb_scan_aligned(struct ext4_allocation_context *ac, @@ -2092,8 +2091,8 @@ repeat: ac->ac_groups_scanned++; if (cr == 0) ext4_mb_simple_scan_group(ac, &e4b); - else if (cr == 1 && - ac->ac_g_ex.fe_len == sbi->s_stripe) + else if (cr == 1 && sbi->s_stripe && + !(ac->ac_g_ex.fe_len % sbi->s_stripe)) ext4_mb_scan_aligned(ac, &e4b); else ext4_mb_complex_scan_group(ac, &e4b); From 0c095c7f113e9fd05913d6e1b2cccbe356be039e Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 27 Jul 2010 11:56:06 -0400 Subject: [PATCH 23/40] ext4: Don't error out the fs if the user tries to make a file too big If the user attempts to make a non-extent-mapped file to be too large, return EFBIG, but don't call ext4_std_err() which will end up marking the file system as containing an error. Thanks to Toshiyuki Okajima-san at Fujitsu for pointing this out. Signed-off-by: "Theodore Ts'o" --- fs/ext4/inode.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 46d2079373c..38ec77fc327 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -5472,10 +5472,8 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr) if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) { struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); - if (attr->ia_size > sbi->s_bitmap_maxbytes) { - error = -EFBIG; - goto err_out; - } + if (attr->ia_size > sbi->s_bitmap_maxbytes) + return -EFBIG; } } From dcc7dae3cb21184a317f10a12250bd8d6f458077 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Tue, 27 Jul 2010 11:56:07 -0400 Subject: [PATCH 24/40] ext4: Fix potential memory leak in ext4_fill_super Under heavy memory pressure we may hit out of memory situation and as result kstrdup'ed options will not be freed. Fix it. Signed-off-by: Cyrill Gorcunov Signed-off-by: "Theodore Ts'o" --- fs/ext4/super.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index ed00c14d708..d573f6c1a4d 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -2550,7 +2550,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) struct inode *root; char *cp; const char *descr; - int ret = -EINVAL; + int ret = -ENOMEM; int blocksize; unsigned int db_count; unsigned int i; @@ -2561,13 +2561,13 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) sbi = kzalloc(sizeof(*sbi), GFP_KERNEL); if (!sbi) - return -ENOMEM; + goto out_free_orig; sbi->s_blockgroup_lock = kzalloc(sizeof(struct blockgroup_lock), GFP_KERNEL); if (!sbi->s_blockgroup_lock) { kfree(sbi); - return -ENOMEM; + goto out_free_orig; } sb->s_fs_info = sbi; sbi->s_mount_opt = 0; @@ -2584,6 +2584,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) for (cp = sb->s_id; (cp = strchr(cp, '/'));) *cp = '!'; + ret = -EINVAL; blocksize = sb_min_blocksize(sb, EXT4_MIN_BLOCK_SIZE); if (!blocksize) { ext4_msg(sb, KERN_ERR, "unable to set blocksize"); @@ -3190,6 +3191,7 @@ out_fail: kfree(sbi->s_blockgroup_lock); kfree(sbi); lock_kernel(); +out_free_orig: kfree(orig_data); return ret; } From 62d2b5f2dcd3707b070efb16bbfdf6947c38c194 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 27 Jul 2010 11:56:07 -0400 Subject: [PATCH 25/40] ext4: Always journal quota file modifications When journaled quota options are not specified, we do writes to quota files just in data=ordered mode. This actually causes warnings from JBD2 about dirty journaled buffer because ext4_getblk unconditionally treats a block allocated by it as metadata. Since quota actually is filesystem metadata, the easiest way to get rid of the warning is to always treat quota writes as metadata... Signed-off-by: Jan Kara Signed-off-by: "Theodore Ts'o" --- fs/ext4/super.c | 19 +++++-------------- 1 file changed, 5 insertions(+), 14 deletions(-) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index d573f6c1a4d..c1036bc8a53 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -4138,7 +4138,6 @@ static ssize_t ext4_quota_write(struct super_block *sb, int type, ext4_lblk_t blk = off >> EXT4_BLOCK_SIZE_BITS(sb); int err = 0; int offset = off & (sb->s_blocksize - 1); - int journal_quota = EXT4_SB(sb)->s_qf_names[type] != NULL; struct buffer_head *bh; handle_t *handle = journal_current_handle(); @@ -4163,24 +4162,16 @@ static ssize_t ext4_quota_write(struct super_block *sb, int type, bh = ext4_bread(handle, inode, blk, 1, &err); if (!bh) goto out; - if (journal_quota) { - err = ext4_journal_get_write_access(handle, bh); - if (err) { - brelse(bh); - goto out; - } + err = ext4_journal_get_write_access(handle, bh); + if (err) { + brelse(bh); + goto out; } lock_buffer(bh); memcpy(bh->b_data+offset, data, len); flush_dcache_page(bh->b_page); unlock_buffer(bh); - if (journal_quota) - err = ext4_handle_dirty_metadata(handle, NULL, bh); - else { - /* Always do at least ordered writes for quotas */ - err = ext4_jbd2_file_inode(handle, inode); - mark_buffer_dirty(bh); - } + err = ext4_handle_dirty_metadata(handle, NULL, bh); brelse(bh); out: if (err) { From 79e8303677fc15f508b9877e0fea1925c4add6f3 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Tue, 27 Jul 2010 11:56:07 -0400 Subject: [PATCH 26/40] ext4: fix ext4_get_blocks references ext4_get_blocks got renamed to ext4_map_blocks, but left stale comments and a prototype littered around. Signed-off-by: Eric Sandeen Signed-off-by: "Theodore Ts'o" --- fs/ext4/ext4.h | 5 +---- fs/ext4/inode.c | 10 +++++----- 2 files changed, 6 insertions(+), 9 deletions(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index fbb39478df2..9ca3637eca5 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -465,7 +465,7 @@ struct ext4_new_group_data { }; /* - * Flags used by ext4_get_blocks() + * Flags used by ext4_map_blocks() */ /* Allocate any needed blocks and/or convert an unitialized extent to be an initialized ext4 */ @@ -1954,9 +1954,6 @@ extern int ext4_convert_unwritten_extents(struct inode *inode, loff_t offset, ssize_t len); extern int ext4_map_blocks(handle_t *handle, struct inode *inode, struct ext4_map_blocks *map, int flags); -extern int ext4_get_blocks(handle_t *handle, struct inode *inode, - sector_t block, unsigned int max_blocks, - struct buffer_head *bh, int flags); extern int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, __u64 start, __u64 len); /* move_extent.c */ diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 38ec77fc327..8279ad73b92 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -2201,7 +2201,7 @@ static int mpage_da_map_blocks(struct mpage_da_data *mpd) BUG_ON(!handle); /* - * Call ext4_get_blocks() to allocate any delayed allocation + * Call ext4_map_blocks() to allocate any delayed allocation * blocks, or to convert an uninitialized extent to be * initialized (in the case where we have written into * one or more preallocated blocks). @@ -2210,7 +2210,7 @@ static int mpage_da_map_blocks(struct mpage_da_data *mpd) * indicate that we are on the delayed allocation path. This * affects functions in many different parts of the allocation * call path. This flag exists primarily because we don't - * want to change *many* call functions, so ext4_get_blocks() + * want to change *many* call functions, so ext4_map_blocks() * will set the magic i_delalloc_reserved_flag once the * inode's allocation semaphore is taken. * @@ -2327,7 +2327,7 @@ static void mpage_add_bh_to_extent(struct mpage_da_data *mpd, * XXX Don't go larger than mballoc is willing to allocate * This is a stopgap solution. We eventually need to fold * mpage_da_submit_io() into this function and then call - * ext4_get_blocks() multiple times in a loop + * ext4_map_blocks() multiple times in a loop */ if (nrblocks >= 8*1024*1024/mpd->inode->i_sb->s_blocksize) goto flush_it; @@ -3948,7 +3948,7 @@ static ssize_t ext4_ext_direct_IO(int rw, struct kiocb *iocb, return -ENOMEM; /* * we save the io structure for current async - * direct IO, so that later ext4_get_blocks() + * direct IO, so that later ext4_map_blocks() * could flag the io structure whether there * is a unwritten extents needs to be converted * when IO is completed. @@ -5675,7 +5675,7 @@ int ext4_writepage_trans_blocks(struct inode *inode) * Calculate the journal credits for a chunk of data modification. * * This is called from DIO, fallocate or whoever calling - * ext4_get_blocks() to map/allocate a chunk of contiguous disk blocks. + * ext4_map_blocks() to map/allocate a chunk of contiguous disk blocks. * * journal buffers for data blocks are not included here, as DIO * and fallocate do no need to journal data buffers. From d889dc8382c4d71b6d538b7b13777bc1ec51df10 Mon Sep 17 00:00:00 2001 From: Toshiyuki Okajima Date: Tue, 27 Jul 2010 11:56:07 -0400 Subject: [PATCH 27/40] ext4: fix EFBIG edge case when writing to large non-extent file By running the following reproducer, we can confirm that the write system call returns with 0 when it should return the error EFBIG. #!/bin/sh /bin/dd if=/dev/zero of=./img bs=1k count=1 seek=1024k > /dev/null 2>&1 /sbin/mkfs.ext3 -Fq ./img /bin/mount -o loop -t ext4 ./img /mnt /bin/touch /mnt/file strace /bin/dd if=/dev/zero of=/mnt/file conv=notrunc bs=1k count=1 seek=$((2194719883264/1024)) 2>&1 | /bin/egrep "write.* 1024\) = " /bin/umount /mnt exit Signed-off-by: Toshiyuki Okajima Signed-off-by: "Theodore Ts'o" Cc: Eric Sandeen --- fs/ext4/file.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/ext4/file.c b/fs/ext4/file.c index bd411c12d63..ee92b66d455 100644 --- a/fs/ext4/file.c +++ b/fs/ext4/file.c @@ -70,7 +70,8 @@ ext4_file_write(struct kiocb *iocb, const struct iovec *iov, struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); size_t length = iov_length(iov, nr_segs); - if (pos > sbi->s_bitmap_maxbytes) + if ((pos > sbi->s_bitmap_maxbytes || + (pos == sbi->s_bitmap_maxbytes && length > 0))) return -EFBIG; if (pos + length > sbi->s_bitmap_maxbytes) { From e3570639c8b5f2c6a5018a2649c2b7c276af76d7 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Tue, 27 Jul 2010 11:56:08 -0400 Subject: [PATCH 28/40] ext4: don't print scary messages for allocation failures post-abort I often get emails containing the "This should not happen!!" message, conveniently trimmed to remove things like: sd 0:0:0:0: [sda] Unhandled error code sd 0:0:0:0: [sda] Result: hostbyte=DID_OK driverbyte=DRIVER_TIMEOUT sd 0:0:0:0: [sda] CDB: Write(10): 2a 00 03 13 c9 70 00 00 28 00 end_request: I/O error, dev sda, sector 51628400 Aborting journal on device dm-0-8. EXT4-fs error (device dm-0): ext4_journal_start_sb: Detected aborted journal EXT4-fs (dm-0): Remounting filesystem read-only I don't think there is any value to the verbosity if the reason is due to a filesystem abort; it just obfuscates the root cause. Signed-off-by: Eric Sandeen Signed-off-by: "Theodore Ts'o" --- fs/ext4/inode.c | 25 ++++++++++++++----------- fs/ext4/mballoc.c | 3 +++ 2 files changed, 17 insertions(+), 11 deletions(-) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 8279ad73b92..a52d5af9918 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -2228,6 +2228,8 @@ static int mpage_da_map_blocks(struct mpage_da_data *mpd) blks = ext4_map_blocks(handle, mpd->inode, &map, get_blocks_flags); if (blks < 0) { + struct super_block *sb = mpd->inode->i_sb; + err = blks; /* * If get block returns with error we simply @@ -2238,7 +2240,7 @@ static int mpage_da_map_blocks(struct mpage_da_data *mpd) return 0; if (err == -ENOSPC && - ext4_count_free_blocks(mpd->inode->i_sb)) { + ext4_count_free_blocks(sb)) { mpd->retval = err; return 0; } @@ -2250,16 +2252,17 @@ static int mpage_da_map_blocks(struct mpage_da_data *mpd) * writepage and writepages will again try to write * the same. */ - ext4_msg(mpd->inode->i_sb, KERN_CRIT, - "delayed block allocation failed for inode %lu at " - "logical offset %llu with max blocks %zd with " - "error %d", mpd->inode->i_ino, - (unsigned long long) next, - mpd->b_size >> mpd->inode->i_blkbits, err); - printk(KERN_CRIT "This should not happen!! " - "Data will be lost\n"); - if (err == -ENOSPC) { - ext4_print_free_blocks(mpd->inode); + if (!(EXT4_SB(sb)->s_mount_flags & EXT4_MF_FS_ABORTED)) { + ext4_msg(sb, KERN_CRIT, + "delayed block allocation failed for inode %lu " + "at logical offset %llu with max blocks %zd " + "with error %d", mpd->inode->i_ino, + (unsigned long long) next, + mpd->b_size >> mpd->inode->i_blkbits, err); + ext4_msg(sb, KERN_CRIT, + "This should not happen!! Data will be lost\n"); + if (err == -ENOSPC) + ext4_print_free_blocks(mpd->inode); } /* invalidate all the pages */ ext4_da_block_invalidatepages(mpd, next, diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index a75de7d44dc..3da28281bc5 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -3884,6 +3884,9 @@ static void ext4_mb_show_ac(struct ext4_allocation_context *ac) struct super_block *sb = ac->ac_sb; ext4_group_t ngroups, i; + if (EXT4_SB(sb)->s_mount_flags & EXT4_MF_FS_ABORTED) + return; + printk(KERN_ERR "EXT4-fs: Can't allocate:" " Allocation context details:\n"); printk(KERN_ERR "EXT4-fs: status %d flags %d\n", From cc937db74bd5fe75f4cfebbfc4a04c1da10f0695 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Tue, 27 Jul 2010 11:56:08 -0400 Subject: [PATCH 29/40] jbd2: Make barrier messages less scary Saying things like "sync failed" when a device does not support barriers makes users slightly more worried than they need to be; rather than talking about sync failures, let's just state the barrier-based facts. Signed-off-by: Eric Sandeen Signed-off-by: "Theodore Ts'o" --- fs/jbd2/commit.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index 75716d3d2be..af056810acb 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -150,8 +150,8 @@ static int journal_submit_commit_record(journal_t *journal, */ if (ret == -EOPNOTSUPP && barrier_done) { printk(KERN_WARNING - "JBD: barrier-based sync failed on %s - " - "disabling barriers\n", journal->j_devname); + "JBD2: Disabling barriers on %s, " + "not supported by device\n", journal->j_devname); spin_lock(&journal->j_state_lock); journal->j_flags &= ~JBD2_BARRIER; spin_unlock(&journal->j_state_lock); @@ -180,8 +180,8 @@ retry: wait_on_buffer(bh); if (buffer_eopnotsupp(bh) && (journal->j_flags & JBD2_BARRIER)) { printk(KERN_WARNING - "JBD2: wait_on_commit_record: sync failed on %s - " - "disabling barriers\n", journal->j_devname); + "JBD2: %s: disabling barries on %s - not supported " + "by device\n", __func__, journal->j_devname); spin_lock(&journal->j_state_lock); journal->j_flags &= ~JBD2_BARRIER; spin_unlock(&journal->j_state_lock); From f613dfcb3345dacb8cf99b7bb359acc1c18a1157 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 27 Jul 2010 11:56:08 -0400 Subject: [PATCH 30/40] ext4: check to make make sure bd_dev is set before dereferencing it There are some drivers which may not set bdev->bd_dev. So make sure it is non-NULL before dereferencing it. Google-Bug-Id: 1773557 Signed-off-by: "Theodore Ts'o" --- fs/ext4/super.c | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index c1036bc8a53..e046eba2478 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -2304,6 +2304,8 @@ static ssize_t session_write_kbytes_show(struct ext4_attr *a, { struct super_block *sb = sbi->s_buddy_cache->i_sb; + if (!sb->s_bdev->bd_part) + return snprintf(buf, PAGE_SIZE, "0\n"); return snprintf(buf, PAGE_SIZE, "%lu\n", (part_stat_read(sb->s_bdev->bd_part, sectors[1]) - sbi->s_sectors_written_start) >> 1); @@ -2314,6 +2316,8 @@ static ssize_t lifetime_write_kbytes_show(struct ext4_attr *a, { struct super_block *sb = sbi->s_buddy_cache->i_sb; + if (!sb->s_bdev->bd_part) + return snprintf(buf, PAGE_SIZE, "0\n"); return snprintf(buf, PAGE_SIZE, "%llu\n", (unsigned long long)(sbi->s_kbytes_written + ((part_stat_read(sb->s_bdev->bd_part, sectors[1]) - @@ -2575,8 +2579,9 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) sbi->s_resgid = EXT4_DEF_RESGID; sbi->s_inode_readahead_blks = EXT4_DEF_INODE_READAHEAD_BLKS; sbi->s_sb_block = sb_block; - sbi->s_sectors_written_start = part_stat_read(sb->s_bdev->bd_part, - sectors[1]); + if (sb->s_bdev->bd_part) + sbi->s_sectors_written_start = + part_stat_read(sb->s_bdev->bd_part, sectors[1]); unlock_kernel(); @@ -3492,10 +3497,14 @@ static int ext4_commit_super(struct super_block *sb, int sync) */ if (!(sb->s_flags & MS_RDONLY)) es->s_wtime = cpu_to_le32(get_seconds()); - es->s_kbytes_written = - cpu_to_le64(EXT4_SB(sb)->s_kbytes_written + + if (sb->s_bdev->bd_part) + es->s_kbytes_written = + cpu_to_le64(EXT4_SB(sb)->s_kbytes_written + ((part_stat_read(sb->s_bdev->bd_part, sectors[1]) - EXT4_SB(sb)->s_sectors_written_start) >> 1)); + else + es->s_kbytes_written = + cpu_to_le64(EXT4_SB(sb)->s_kbytes_written); ext4_free_blocks_count_set(es, percpu_counter_sum_positive( &EXT4_SB(sb)->s_freeblocks_counter)); es->s_free_inodes_count = cpu_to_le32(percpu_counter_sum_positive( From 4538821993f4486c76090dfb377c60c0a0e71ba3 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Thu, 29 Jul 2010 15:06:10 -0400 Subject: [PATCH 31/40] ext4: drop inode from orphan list if ext4_delete_inode() fails There were some error paths in ext4_delete_inode() which was not dropping the inode from the orphan list. This could lead to a BUG_ON on umount when the orphan list is discovered to be non-empty. Signed-off-by: "Theodore Ts'o" --- fs/ext4/inode.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index a52d5af9918..533b607f9cb 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -221,6 +221,7 @@ void ext4_delete_inode(struct inode *inode) "couldn't extend journal (err %d)", err); stop_handle: ext4_journal_stop(handle); + ext4_orphan_del(NULL, inode); goto no_delete; } } From 437f88cc031ffe7f37f3e705367f4fe1f4be8b0f Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Sun, 1 Aug 2010 17:33:29 -0400 Subject: [PATCH 32/40] ext4: fix freeze deadlock under IO Commit 6b0310fbf087ad6 caused a regression resulting in deadlocks when freezing a filesystem which had active IO; the vfs_check_frozen level (SB_FREEZE_WRITE) did not let the freeze-related IO syncing through. Duh. Changing the test to FREEZE_TRANS should let the normal freeze syncing get through the fs, but still block any transactions from starting once the fs is completely frozen. I tested this by running fsstress in the background while periodically snapshotting the fs and running fsck on the result. I ran into occasional deadlocks, but different ones. I think this is a fine fix for the problem at hand, and the other deadlocky things will need more investigation. Reported-by: Phillip Susi Signed-off-by: Eric Sandeen Signed-off-by: "Theodore Ts'o" --- fs/ext4/super.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index e046eba2478..282a2704be2 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -241,7 +241,7 @@ handle_t *ext4_journal_start_sb(struct super_block *sb, int nblocks) if (sb->s_flags & MS_RDONLY) return ERR_PTR(-EROFS); - vfs_check_frozen(sb, SB_FREEZE_WRITE); + vfs_check_frozen(sb, SB_FREEZE_TRANS); /* Special case here: if the journal has aborted behind our * backs (eg. EIO in the commit thread), then we still need to * take the FS itself readonly cleanly. */ @@ -3608,7 +3608,7 @@ int ext4_force_commit(struct super_block *sb) journal = EXT4_SB(sb)->s_journal; if (journal) { - vfs_check_frozen(sb, SB_FREEZE_WRITE); + vfs_check_frozen(sb, SB_FREEZE_TRANS); ret = ext4_journal_force_commit(journal); } From ca0e05e4b15193aeba72b995e90de990db7f8304 Mon Sep 17 00:00:00 2001 From: Dmitry Monakhov Date: Sun, 1 Aug 2010 17:48:36 -0400 Subject: [PATCH 33/40] ext4: force block allocation on quota_off Perform full sync procedure so that any delayed allocation blocks are allocated so quota will be consistent. Signed-off-by: Dmitry Monakhov Signed-off-by: "Theodore Ts'o" --- fs/ext4/super.c | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 282a2704be2..3e3f6484c22 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -1118,6 +1118,7 @@ static int ext4_mark_dquot_dirty(struct dquot *dquot); static int ext4_write_info(struct super_block *sb, int type); static int ext4_quota_on(struct super_block *sb, int type, int format_id, char *path); +static int ext4_quota_off(struct super_block *sb, int type); static int ext4_quota_on_mount(struct super_block *sb, int type); static ssize_t ext4_quota_read(struct super_block *sb, int type, char *data, size_t len, loff_t off); @@ -1139,7 +1140,7 @@ static const struct dquot_operations ext4_quota_operations = { static const struct quotactl_ops ext4_qctl_operations = { .quota_on = ext4_quota_on, - .quota_off = dquot_quota_off, + .quota_off = ext4_quota_off, .quota_sync = dquot_quota_sync, .get_info = dquot_get_dqinfo, .set_info = dquot_set_dqinfo, @@ -4098,6 +4099,18 @@ static int ext4_quota_on(struct super_block *sb, int type, int format_id, return err; } +static int ext4_quota_off(struct super_block *sb, int type) +{ + /* Force all delayed allocation blocks to be allocated */ + if (test_opt(sb, DELALLOC)) { + down_read(&sb->s_umount); + sync_filesystem(sb); + up_read(&sb->s_umount); + } + + return dquot_quota_off(sb, type); +} + /* Read data from quotafile - avoid pagecache and such because we cannot afford * acquiring the locks... As quota files are never truncated and quota code * itself serializes the operations (and noone else should touch the files) From 8b67f04ab9de5d8f3a71aef72bf02c995a506db5 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Sun, 1 Aug 2010 23:14:20 -0400 Subject: [PATCH 34/40] ext4: Add mount options in superblock Allow mount options to be stored in the superblock. Also add default mount option bits for nobarrier, block_validity, discard, and nodelalloc. Signed-off-by: "Theodore Ts'o" --- fs/ext4/ext4.h | 9 +++++++-- fs/ext4/super.c | 29 +++++++++++++++++++++++------ 2 files changed, 30 insertions(+), 8 deletions(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 9ca3637eca5..ed14e1db083 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1025,8 +1025,9 @@ struct ext4_super_block { __le32 s_last_error_line; /* line number where error happened */ __le64 s_last_error_block; /* block involved of last error */ __u8 s_last_error_func[32]; /* function where the error happened */ -#define EXT4_S_ERR_END offsetof(struct ext4_super_block, s_reserved) - __le32 s_reserved[128]; /* Padding to the end of the block */ +#define EXT4_S_ERR_END offsetof(struct ext4_super_block, s_mount_opts) + __u8 s_mount_opts[64]; + __le32 s_reserved[112]; /* Padding to the end of the block */ }; #define EXT4_S_ERR_LEN (EXT4_S_ERR_END - EXT4_S_ERR_START) @@ -1341,6 +1342,10 @@ EXT4_INODE_BIT_FNS(state, state_flags) #define EXT4_DEFM_JMODE_DATA 0x0020 #define EXT4_DEFM_JMODE_ORDERED 0x0040 #define EXT4_DEFM_JMODE_WBACK 0x0060 +#define EXT4_DEFM_NOBARRIER 0x0100 +#define EXT4_DEFM_BLOCK_VALIDITY 0x0200 +#define EXT4_DEFM_DISCARD 0x0400 +#define EXT4_DEFM_NODELALLOC 0x0800 /* * Default journal batch times diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 3e3f6484c22..3fd65eb66cc 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -1003,10 +1003,10 @@ static int ext4_show_options(struct seq_file *seq, struct vfsmount *vfs) seq_puts(seq, ",journal_checksum"); if (test_opt(sb, I_VERSION)) seq_puts(seq, ",i_version"); - if (!test_opt(sb, DELALLOC)) + if (!test_opt(sb, DELALLOC) && + !(def_mount_opts & EXT4_DEFM_NODELALLOC)) seq_puts(seq, ",nodelalloc"); - if (sbi->s_stripe) seq_printf(seq, ",stripe=%lu", sbi->s_stripe); /* @@ -1030,7 +1030,7 @@ static int ext4_show_options(struct seq_file *seq, struct vfsmount *vfs) if (test_opt(sb, NO_AUTO_DA_ALLOC)) seq_puts(seq, ",noauto_da_alloc"); - if (test_opt(sb, DISCARD)) + if (test_opt(sb, DISCARD) && !(def_mount_opts & EXT4_DEFM_DISCARD)) seq_puts(seq, ",discard"); if (test_opt(sb, NOLOAD)) @@ -1039,6 +1039,10 @@ static int ext4_show_options(struct seq_file *seq, struct vfsmount *vfs) if (test_opt(sb, DIOREAD_NOLOCK)) seq_puts(seq, ",dioread_nolock"); + if (test_opt(sb, BLOCK_VALIDITY) && + !(def_mount_opts & EXT4_DEFM_BLOCK_VALIDITY)) + seq_puts(seq, ",block_validity"); + ext4_show_quota_options(seq, sb); return 0; @@ -2655,6 +2659,10 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) set_opt(sbi->s_mount_opt, ERRORS_CONT); else set_opt(sbi->s_mount_opt, ERRORS_RO); + if (def_mount_opts & EXT4_DEFM_BLOCK_VALIDITY) + set_opt(sbi->s_mount_opt, BLOCK_VALIDITY); + if (def_mount_opts & EXT4_DEFM_DISCARD) + set_opt(sbi->s_mount_opt, DISCARD); sbi->s_resuid = le16_to_cpu(es->s_def_resuid); sbi->s_resgid = le16_to_cpu(es->s_def_resgid); @@ -2662,15 +2670,23 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) sbi->s_min_batch_time = EXT4_DEF_MIN_BATCH_TIME; sbi->s_max_batch_time = EXT4_DEF_MAX_BATCH_TIME; - set_opt(sbi->s_mount_opt, BARRIER); + if ((def_mount_opts & EXT4_DEFM_NOBARRIER) == 0) + set_opt(sbi->s_mount_opt, BARRIER); /* * enable delayed allocation by default * Use -o nodelalloc to turn it off */ - if (!IS_EXT3_SB(sb)) + if (!IS_EXT3_SB(sb) && + ((def_mount_opts & EXT4_DEFM_NODELALLOC) == 0)) set_opt(sbi->s_mount_opt, DELALLOC); + if (!parse_options((char *) sbi->s_es->s_mount_opts, sb, + &journal_devnum, &journal_ioprio, NULL, 0)) { + ext4_msg(sb, KERN_WARNING, + "failed to parse options in superblock: %s", + sbi->s_es->s_mount_opts); + } if (!parse_options((char *) data, sb, &journal_devnum, &journal_ioprio, NULL, 0)) goto failed_mount; @@ -3141,7 +3157,8 @@ no_journal: descr = "out journal"; ext4_msg(sb, KERN_INFO, "mounted filesystem with%s. " - "Opts: %s", descr, orig_data); + "Opts: %s%s%s", descr, sbi->s_es->s_mount_opts, + *sbi->s_es->s_mount_opts ? "; " : "", orig_data); init_timer(&sbi->s_err_report); sbi->s_err_report.function = print_daily_error_info; From a51dca9cd3bb4ec5a05bfb6feabf024a5c808a37 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Mon, 2 Aug 2010 08:43:25 -0400 Subject: [PATCH 35/40] jbd2: Use atomic variables to avoid taking t_handle_lock in jbd2_journal_stop By using an atomic_t for t_updates and t_outstanding credits, this should allow us to not need to take transaction t_handle_lock in jbd2_journal_stop(). Signed-off-by: "Theodore Ts'o" --- fs/jbd2/checkpoint.c | 2 +- fs/jbd2/commit.c | 13 ++++---- fs/jbd2/transaction.c | 69 ++++++++++++++++++++++++------------------- include/linux/jbd2.h | 8 ++--- 4 files changed, 51 insertions(+), 41 deletions(-) diff --git a/fs/jbd2/checkpoint.c b/fs/jbd2/checkpoint.c index 076d1cc44f9..f8cdc02520f 100644 --- a/fs/jbd2/checkpoint.c +++ b/fs/jbd2/checkpoint.c @@ -775,7 +775,7 @@ void __jbd2_journal_drop_transaction(journal_t *journal, transaction_t *transact J_ASSERT(transaction->t_log_list == NULL); J_ASSERT(transaction->t_checkpoint_list == NULL); J_ASSERT(transaction->t_checkpoint_io_list == NULL); - J_ASSERT(transaction->t_updates == 0); + J_ASSERT(atomic_read(&transaction->t_updates) == 0); J_ASSERT(journal->j_committing_transaction != transaction); J_ASSERT(journal->j_running_transaction != transaction); diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index af056810acb..fbd2c564e91 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -417,12 +417,12 @@ void jbd2_journal_commit_transaction(journal_t *journal) stats.run.rs_locked); spin_lock(&commit_transaction->t_handle_lock); - while (commit_transaction->t_updates) { + while (atomic_read(&commit_transaction->t_updates)) { DEFINE_WAIT(wait); prepare_to_wait(&journal->j_wait_updates, &wait, TASK_UNINTERRUPTIBLE); - if (commit_transaction->t_updates) { + if (atomic_read(&commit_transaction->t_updates)) { spin_unlock(&commit_transaction->t_handle_lock); spin_unlock(&journal->j_state_lock); schedule(); @@ -433,7 +433,7 @@ void jbd2_journal_commit_transaction(journal_t *journal) } spin_unlock(&commit_transaction->t_handle_lock); - J_ASSERT (commit_transaction->t_outstanding_credits <= + J_ASSERT (atomic_read(&commit_transaction->t_outstanding_credits) <= journal->j_max_transaction_buffers); /* @@ -527,11 +527,12 @@ void jbd2_journal_commit_transaction(journal_t *journal) stats.run.rs_logging = jiffies; stats.run.rs_flushing = jbd2_time_diff(stats.run.rs_flushing, stats.run.rs_logging); - stats.run.rs_blocks = commit_transaction->t_outstanding_credits; + stats.run.rs_blocks = + atomic_read(&commit_transaction->t_outstanding_credits); stats.run.rs_blocks_logged = 0; J_ASSERT(commit_transaction->t_nr_buffers <= - commit_transaction->t_outstanding_credits); + atomic_read(&commit_transaction->t_outstanding_credits)); err = 0; descriptor = NULL; @@ -616,7 +617,7 @@ void jbd2_journal_commit_transaction(journal_t *journal) * the free space in the log, but this counter is changed * by jbd2_journal_next_log_block() also. */ - commit_transaction->t_outstanding_credits--; + atomic_dec(&commit_transaction->t_outstanding_credits); /* Bump b_count to prevent truncate from stumbling over the shadowed buffer! @@@ This can go if we ever get diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c index 001e95fb0fe..9c64c7ec48d 100644 --- a/fs/jbd2/transaction.c +++ b/fs/jbd2/transaction.c @@ -55,6 +55,8 @@ jbd2_get_transaction(journal_t *journal, transaction_t *transaction) transaction->t_tid = journal->j_transaction_sequence++; transaction->t_expires = jiffies + journal->j_commit_interval; spin_lock_init(&transaction->t_handle_lock); + atomic_set(&transaction->t_updates, 0); + atomic_set(&transaction->t_outstanding_credits, 0); INIT_LIST_HEAD(&transaction->t_inode_list); INIT_LIST_HEAD(&transaction->t_private_list); @@ -177,7 +179,7 @@ repeat_locked: * checkpoint to free some more log space. */ spin_lock(&transaction->t_handle_lock); - needed = transaction->t_outstanding_credits + nblocks; + needed = atomic_read(&transaction->t_outstanding_credits) + nblocks; if (needed > journal->j_max_transaction_buffers) { /* @@ -240,11 +242,12 @@ repeat_locked: } handle->h_transaction = transaction; - transaction->t_outstanding_credits += nblocks; - transaction->t_updates++; + atomic_add(nblocks, &transaction->t_outstanding_credits); + atomic_inc(&transaction->t_updates); transaction->t_handle_count++; jbd_debug(4, "Handle %p given %d credits (total %d, free %d)\n", - handle, nblocks, transaction->t_outstanding_credits, + handle, nblocks, + atomic_read(&transaction->t_outstanding_credits), __jbd2_log_space_left(journal)); spin_unlock(&transaction->t_handle_lock); spin_unlock(&journal->j_state_lock); @@ -369,7 +372,7 @@ int jbd2_journal_extend(handle_t *handle, int nblocks) } spin_lock(&transaction->t_handle_lock); - wanted = transaction->t_outstanding_credits + nblocks; + wanted = atomic_read(&transaction->t_outstanding_credits) + nblocks; if (wanted > journal->j_max_transaction_buffers) { jbd_debug(3, "denied handle %p %d blocks: " @@ -384,7 +387,7 @@ int jbd2_journal_extend(handle_t *handle, int nblocks) } handle->h_buffer_credits += nblocks; - transaction->t_outstanding_credits += nblocks; + atomic_add(nblocks, &transaction->t_outstanding_credits); result = 0; jbd_debug(3, "extended handle %p by %d\n", handle, nblocks); @@ -426,15 +429,14 @@ int jbd2__journal_restart(handle_t *handle, int nblocks, int gfp_mask) * First unlink the handle from its current transaction, and start the * commit on that. */ - J_ASSERT(transaction->t_updates > 0); + J_ASSERT(atomic_read(&transaction->t_updates) > 0); J_ASSERT(journal_current_handle() == handle); spin_lock(&journal->j_state_lock); spin_lock(&transaction->t_handle_lock); - transaction->t_outstanding_credits -= handle->h_buffer_credits; - transaction->t_updates--; - - if (!transaction->t_updates) + atomic_sub(handle->h_buffer_credits, + &transaction->t_outstanding_credits); + if (atomic_dec_and_test(&transaction->t_updates)) wake_up(&journal->j_wait_updates); spin_unlock(&transaction->t_handle_lock); @@ -481,7 +483,7 @@ void jbd2_journal_lock_updates(journal_t *journal) break; spin_lock(&transaction->t_handle_lock); - if (!transaction->t_updates) { + if (!atomic_read(&transaction->t_updates)) { spin_unlock(&transaction->t_handle_lock); break; } @@ -1258,7 +1260,8 @@ int jbd2_journal_stop(handle_t *handle) { transaction_t *transaction = handle->h_transaction; journal_t *journal = transaction->t_journal; - int err; + int err, wait_for_commit = 0; + tid_t tid; pid_t pid; J_ASSERT(journal_current_handle() == handle); @@ -1266,7 +1269,7 @@ int jbd2_journal_stop(handle_t *handle) if (is_handle_aborted(handle)) err = -EIO; else { - J_ASSERT(transaction->t_updates > 0); + J_ASSERT(atomic_read(&transaction->t_updates) > 0); err = 0; } @@ -1334,14 +1337,8 @@ int jbd2_journal_stop(handle_t *handle) if (handle->h_sync) transaction->t_synchronous_commit = 1; current->journal_info = NULL; - spin_lock(&transaction->t_handle_lock); - transaction->t_outstanding_credits -= handle->h_buffer_credits; - transaction->t_updates--; - if (!transaction->t_updates) { - wake_up(&journal->j_wait_updates); - if (journal->j_barrier_count) - wake_up(&journal->j_wait_transaction_locked); - } + atomic_sub(handle->h_buffer_credits, + &transaction->t_outstanding_credits); /* * If the handle is marked SYNC, we need to set another commit @@ -1350,15 +1347,13 @@ int jbd2_journal_stop(handle_t *handle) * transaction is too old now. */ if (handle->h_sync || - transaction->t_outstanding_credits > - journal->j_max_transaction_buffers || - time_after_eq(jiffies, transaction->t_expires)) { + (atomic_read(&transaction->t_outstanding_credits) > + journal->j_max_transaction_buffers) || + time_after_eq(jiffies, transaction->t_expires)) { /* Do this even for aborted journals: an abort still * completes the commit thread, it just doesn't write * anything to disk. */ - tid_t tid = transaction->t_tid; - spin_unlock(&transaction->t_handle_lock); jbd_debug(2, "transaction too old, requesting commit for " "handle %p\n", handle); /* This is non-blocking */ @@ -1369,11 +1364,25 @@ int jbd2_journal_stop(handle_t *handle) * to wait for the commit to complete. */ if (handle->h_sync && !(current->flags & PF_MEMALLOC)) - err = jbd2_log_wait_commit(journal, tid); - } else { - spin_unlock(&transaction->t_handle_lock); + wait_for_commit = 1; } + /* + * Once we drop t_updates, if it goes to zero the transaction + * could start commiting on us and eventually disappear. So + * once we do this, we must not dereference transaction + * pointer again. + */ + tid = transaction->t_tid; + if (atomic_dec_and_test(&transaction->t_updates)) { + wake_up(&journal->j_wait_updates); + if (journal->j_barrier_count) + wake_up(&journal->j_wait_transaction_locked); + } + + if (wait_for_commit) + err = jbd2_log_wait_commit(journal, tid); + lock_map_release(&handle->h_lockdep_map); jbd2_free_handle(handle); diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h index 5a72bc75b27..a72ce21de0e 100644 --- a/include/linux/jbd2.h +++ b/include/linux/jbd2.h @@ -601,13 +601,13 @@ struct transaction_s * Number of outstanding updates running on this transaction * [t_handle_lock] */ - int t_updates; + atomic_t t_updates; /* * Number of buffers reserved for use by all handles in this transaction * handle but not yet modified. [t_handle_lock] */ - int t_outstanding_credits; + atomic_t t_outstanding_credits; /* * Forward and backward links for the circular list of all transactions @@ -1258,8 +1258,8 @@ static inline int jbd_space_needed(journal_t *journal) { int nblocks = journal->j_max_transaction_buffers; if (journal->j_committing_transaction) - nblocks += journal->j_committing_transaction-> - t_outstanding_credits; + nblocks += atomic_read(&journal->j_committing_transaction-> + t_outstanding_credits); return nblocks; } From a931da6ac9331a6c80dd91c199105806f2336188 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 3 Aug 2010 21:35:12 -0400 Subject: [PATCH 36/40] jbd2: Change j_state_lock to be a rwlock_t Lockstat reports have shown that j_state_lock is a major source of lock contention, especially on systems with more than 4 CPU cores. So change it to be a read/write spinlock. Signed-off-by: "Theodore Ts'o" --- fs/ext4/inode.c | 4 +- fs/ext4/super.c | 4 +- fs/jbd2/checkpoint.c | 16 ++++---- fs/jbd2/commit.c | 26 ++++++------ fs/jbd2/journal.c | 94 +++++++++++++++++++++---------------------- fs/jbd2/transaction.c | 74 ++++++++++++++++++---------------- fs/ocfs2/journal.c | 4 +- include/linux/jbd2.h | 2 +- 8 files changed, 114 insertions(+), 110 deletions(-) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 533b607f9cb..ab2247d642c 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -5066,7 +5066,7 @@ struct inode *ext4_iget(struct super_block *sb, unsigned long ino) transaction_t *transaction; tid_t tid; - spin_lock(&journal->j_state_lock); + read_lock(&journal->j_state_lock); if (journal->j_running_transaction) transaction = journal->j_running_transaction; else @@ -5075,7 +5075,7 @@ struct inode *ext4_iget(struct super_block *sb, unsigned long ino) tid = transaction->t_tid; else tid = journal->j_commit_sequence; - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); ei->i_sync_tid = tid; ei->i_datasync_tid = tid; } diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 3fd65eb66cc..81cb3fc1218 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -3232,7 +3232,7 @@ static void ext4_init_journal_params(struct super_block *sb, journal_t *journal) journal->j_min_batch_time = sbi->s_min_batch_time; journal->j_max_batch_time = sbi->s_max_batch_time; - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); if (test_opt(sb, BARRIER)) journal->j_flags |= JBD2_BARRIER; else @@ -3241,7 +3241,7 @@ static void ext4_init_journal_params(struct super_block *sb, journal_t *journal) journal->j_flags |= JBD2_ABORT_ON_SYNCDATA_ERR; else journal->j_flags &= ~JBD2_ABORT_ON_SYNCDATA_ERR; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); } static journal_t *ext4_get_journal(struct super_block *sb, diff --git a/fs/jbd2/checkpoint.c b/fs/jbd2/checkpoint.c index f8cdc02520f..1c23a0f4e8a 100644 --- a/fs/jbd2/checkpoint.c +++ b/fs/jbd2/checkpoint.c @@ -118,13 +118,13 @@ static int __try_to_free_cp_buf(struct journal_head *jh) void __jbd2_log_wait_for_space(journal_t *journal) { int nblocks, space_left; - assert_spin_locked(&journal->j_state_lock); + /* assert_spin_locked(&journal->j_state_lock); */ nblocks = jbd_space_needed(journal); while (__jbd2_log_space_left(journal) < nblocks) { if (journal->j_flags & JBD2_ABORT) return; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); mutex_lock(&journal->j_checkpoint_mutex); /* @@ -138,7 +138,7 @@ void __jbd2_log_wait_for_space(journal_t *journal) * filesystem, so abort the journal and leave a stack * trace for forensic evidence. */ - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); spin_lock(&journal->j_list_lock); nblocks = jbd_space_needed(journal); space_left = __jbd2_log_space_left(journal); @@ -149,7 +149,7 @@ void __jbd2_log_wait_for_space(journal_t *journal) if (journal->j_committing_transaction) tid = journal->j_committing_transaction->t_tid; spin_unlock(&journal->j_list_lock); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); if (chkpt) { jbd2_log_do_checkpoint(journal); } else if (jbd2_cleanup_journal_tail(journal) == 0) { @@ -167,7 +167,7 @@ void __jbd2_log_wait_for_space(journal_t *journal) WARN_ON(1); jbd2_journal_abort(journal, 0); } - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); } else { spin_unlock(&journal->j_list_lock); } @@ -474,7 +474,7 @@ int jbd2_cleanup_journal_tail(journal_t *journal) * next transaction ID we will write, and where it will * start. */ - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); spin_lock(&journal->j_list_lock); transaction = journal->j_checkpoint_transactions; if (transaction) { @@ -496,7 +496,7 @@ int jbd2_cleanup_journal_tail(journal_t *journal) /* If the oldest pinned transaction is at the tail of the log already then there's not much we can do right now. */ if (journal->j_tail_sequence == first_tid) { - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); return 1; } @@ -516,7 +516,7 @@ int jbd2_cleanup_journal_tail(journal_t *journal) journal->j_free += freed; journal->j_tail_sequence = first_tid; journal->j_tail = blocknr; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); /* * If there is an external journal, we need to make sure that diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index fbd2c564e91..67bb0a2f35e 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -152,9 +152,9 @@ static int journal_submit_commit_record(journal_t *journal, printk(KERN_WARNING "JBD2: Disabling barriers on %s, " "not supported by device\n", journal->j_devname); - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); journal->j_flags &= ~JBD2_BARRIER; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); /* And try again, without the barrier */ lock_buffer(bh); @@ -182,9 +182,9 @@ retry: printk(KERN_WARNING "JBD2: %s: disabling barries on %s - not supported " "by device\n", __func__, journal->j_devname); - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); journal->j_flags &= ~JBD2_BARRIER; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); lock_buffer(bh); clear_buffer_dirty(bh); @@ -400,7 +400,7 @@ void jbd2_journal_commit_transaction(journal_t *journal) jbd_debug(1, "JBD: starting commit of transaction %d\n", commit_transaction->t_tid); - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); commit_transaction->t_state = T_LOCKED; /* @@ -424,9 +424,9 @@ void jbd2_journal_commit_transaction(journal_t *journal) TASK_UNINTERRUPTIBLE); if (atomic_read(&commit_transaction->t_updates)) { spin_unlock(&commit_transaction->t_handle_lock); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); schedule(); - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); spin_lock(&commit_transaction->t_handle_lock); } finish_wait(&journal->j_wait_updates, &wait); @@ -497,7 +497,7 @@ void jbd2_journal_commit_transaction(journal_t *journal) start_time = ktime_get(); commit_transaction->t_log_start = journal->j_head; wake_up(&journal->j_wait_transaction_locked); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); jbd_debug (3, "JBD: commit phase 2\n"); @@ -519,9 +519,9 @@ void jbd2_journal_commit_transaction(journal_t *journal) * transaction! Now comes the tricky part: we need to write out * metadata. Loop over the transaction's entire buffer list: */ - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); commit_transaction->t_state = T_COMMIT; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); trace_jbd2_commit_logging(journal, commit_transaction); stats.run.rs_logging = jiffies; @@ -978,7 +978,7 @@ restart_loop: * __jbd2_journal_drop_transaction(). Otherwise we could race with * other checkpointing code processing the transaction... */ - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); spin_lock(&journal->j_list_lock); /* * Now recheck if some buffers did not get attached to the transaction @@ -986,7 +986,7 @@ restart_loop: */ if (commit_transaction->t_forget) { spin_unlock(&journal->j_list_lock); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); goto restart_loop; } @@ -1038,7 +1038,7 @@ restart_loop: journal->j_average_commit_time*3) / 4; else journal->j_average_commit_time = commit_time; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); if (commit_transaction->t_checkpoint_list == NULL && commit_transaction->t_checkpoint_io_list == NULL) { diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c index a79d3345b55..e7bf0fd9cec 100644 --- a/fs/jbd2/journal.c +++ b/fs/jbd2/journal.c @@ -142,7 +142,7 @@ static int kjournald2(void *arg) /* * And now, wait forever for commit wakeup events. */ - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); loop: if (journal->j_flags & JBD2_UNMOUNT) @@ -153,10 +153,10 @@ loop: if (journal->j_commit_sequence != journal->j_commit_request) { jbd_debug(1, "OK, requests differ\n"); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); del_timer_sync(&journal->j_commit_timer); jbd2_journal_commit_transaction(journal); - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); goto loop; } @@ -168,9 +168,9 @@ loop: * be already stopped. */ jbd_debug(1, "Now suspending kjournald2\n"); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); refrigerator(); - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); } else { /* * We assume on resume that commits are already there, @@ -190,9 +190,9 @@ loop: if (journal->j_flags & JBD2_UNMOUNT) should_sleep = 0; if (should_sleep) { - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); schedule(); - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); } finish_wait(&journal->j_wait_commit, &wait); } @@ -210,7 +210,7 @@ loop: goto loop; end_loop: - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); del_timer_sync(&journal->j_commit_timer); journal->j_task = NULL; wake_up(&journal->j_wait_done_commit); @@ -233,16 +233,16 @@ static int jbd2_journal_start_thread(journal_t *journal) static void journal_kill_thread(journal_t *journal) { - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); journal->j_flags |= JBD2_UNMOUNT; while (journal->j_task) { wake_up(&journal->j_wait_commit); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); wait_event(journal->j_wait_done_commit, journal->j_task == NULL); - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); } - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); } /* @@ -452,7 +452,7 @@ int __jbd2_log_space_left(journal_t *journal) { int left = journal->j_free; - assert_spin_locked(&journal->j_state_lock); + /* assert_spin_locked(&journal->j_state_lock); */ /* * Be pessimistic here about the number of those free blocks which @@ -497,9 +497,9 @@ int jbd2_log_start_commit(journal_t *journal, tid_t tid) { int ret; - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); ret = __jbd2_log_start_commit(journal, tid); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); return ret; } @@ -518,7 +518,7 @@ int jbd2_journal_force_commit_nested(journal_t *journal) transaction_t *transaction = NULL; tid_t tid; - spin_lock(&journal->j_state_lock); + read_lock(&journal->j_state_lock); if (journal->j_running_transaction && !current->journal_info) { transaction = journal->j_running_transaction; __jbd2_log_start_commit(journal, transaction->t_tid); @@ -526,12 +526,12 @@ int jbd2_journal_force_commit_nested(journal_t *journal) transaction = journal->j_committing_transaction; if (!transaction) { - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); return 0; /* Nothing to retry */ } tid = transaction->t_tid; - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); jbd2_log_wait_commit(journal, tid); return 1; } @@ -545,7 +545,7 @@ int jbd2_journal_start_commit(journal_t *journal, tid_t *ptid) { int ret = 0; - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); if (journal->j_running_transaction) { tid_t tid = journal->j_running_transaction->t_tid; @@ -564,7 +564,7 @@ int jbd2_journal_start_commit(journal_t *journal, tid_t *ptid) *ptid = journal->j_committing_transaction->t_tid; ret = 1; } - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); return ret; } @@ -576,26 +576,24 @@ int jbd2_log_wait_commit(journal_t *journal, tid_t tid) { int err = 0; + read_lock(&journal->j_state_lock); #ifdef CONFIG_JBD2_DEBUG - spin_lock(&journal->j_state_lock); if (!tid_geq(journal->j_commit_request, tid)) { printk(KERN_EMERG "%s: error: j_commit_request=%d, tid=%d\n", __func__, journal->j_commit_request, tid); } - spin_unlock(&journal->j_state_lock); #endif - spin_lock(&journal->j_state_lock); while (tid_gt(tid, journal->j_commit_sequence)) { jbd_debug(1, "JBD: want %d, j_commit_sequence=%d\n", tid, journal->j_commit_sequence); wake_up(&journal->j_wait_commit); - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); wait_event(journal->j_wait_done_commit, !tid_gt(tid, journal->j_commit_sequence)); - spin_lock(&journal->j_state_lock); + read_lock(&journal->j_state_lock); } - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); if (unlikely(is_journal_aborted(journal))) { printk(KERN_EMERG "journal commit I/O error\n"); @@ -612,7 +610,7 @@ int jbd2_journal_next_log_block(journal_t *journal, unsigned long long *retp) { unsigned long blocknr; - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); J_ASSERT(journal->j_free > 1); blocknr = journal->j_head; @@ -620,7 +618,7 @@ int jbd2_journal_next_log_block(journal_t *journal, unsigned long long *retp) journal->j_free--; if (journal->j_head == journal->j_last) journal->j_head = journal->j_first; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); return jbd2_journal_bmap(journal, blocknr, retp); } @@ -840,7 +838,7 @@ static journal_t * journal_init_common (void) mutex_init(&journal->j_checkpoint_mutex); spin_lock_init(&journal->j_revoke_lock); spin_lock_init(&journal->j_list_lock); - spin_lock_init(&journal->j_state_lock); + rwlock_init(&journal->j_state_lock); journal->j_commit_interval = (HZ * JBD2_DEFAULT_MAX_COMMIT_AGE); journal->j_min_batch_time = 0; @@ -1106,14 +1104,14 @@ void jbd2_journal_update_superblock(journal_t *journal, int wait) set_buffer_uptodate(bh); } - spin_lock(&journal->j_state_lock); + read_lock(&journal->j_state_lock); jbd_debug(1,"JBD: updating superblock (start %ld, seq %d, errno %d)\n", journal->j_tail, journal->j_tail_sequence, journal->j_errno); sb->s_sequence = cpu_to_be32(journal->j_tail_sequence); sb->s_start = cpu_to_be32(journal->j_tail); sb->s_errno = cpu_to_be32(journal->j_errno); - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); BUFFER_TRACE(bh, "marking dirty"); mark_buffer_dirty(bh); @@ -1134,12 +1132,12 @@ out: * any future commit will have to be careful to update the * superblock again to re-record the true start of the log. */ - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); if (sb->s_start) journal->j_flags &= ~JBD2_FLUSHED; else journal->j_flags |= JBD2_FLUSHED; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); } /* @@ -1551,7 +1549,7 @@ int jbd2_journal_flush(journal_t *journal) transaction_t *transaction = NULL; unsigned long old_tail; - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); /* Force everything buffered to the log... */ if (journal->j_running_transaction) { @@ -1564,10 +1562,10 @@ int jbd2_journal_flush(journal_t *journal) if (transaction) { tid_t tid = transaction->t_tid; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); jbd2_log_wait_commit(journal, tid); } else { - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); } /* ...and flush everything in the log out to disk. */ @@ -1591,12 +1589,12 @@ int jbd2_journal_flush(journal_t *journal) * the magic code for a fully-recovered superblock. Any future * commits of data to the journal will restore the current * s_start value. */ - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); old_tail = journal->j_tail; journal->j_tail = 0; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); jbd2_journal_update_superblock(journal, 1); - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); journal->j_tail = old_tail; J_ASSERT(!journal->j_running_transaction); @@ -1604,7 +1602,7 @@ int jbd2_journal_flush(journal_t *journal) J_ASSERT(!journal->j_checkpoint_transactions); J_ASSERT(journal->j_head == journal->j_tail); J_ASSERT(journal->j_tail_sequence == journal->j_transaction_sequence); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); return 0; } @@ -1668,12 +1666,12 @@ void __jbd2_journal_abort_hard(journal_t *journal) printk(KERN_ERR "Aborting journal on device %s.\n", journal->j_devname); - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); journal->j_flags |= JBD2_ABORT; transaction = journal->j_running_transaction; if (transaction) __jbd2_log_start_commit(journal, transaction->t_tid); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); } /* Soft abort: record the abort error status in the journal superblock, @@ -1758,12 +1756,12 @@ int jbd2_journal_errno(journal_t *journal) { int err; - spin_lock(&journal->j_state_lock); + read_lock(&journal->j_state_lock); if (journal->j_flags & JBD2_ABORT) err = -EROFS; else err = journal->j_errno; - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); return err; } @@ -1778,12 +1776,12 @@ int jbd2_journal_clear_err(journal_t *journal) { int err = 0; - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); if (journal->j_flags & JBD2_ABORT) err = -EROFS; else journal->j_errno = 0; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); return err; } @@ -1796,10 +1794,10 @@ int jbd2_journal_clear_err(journal_t *journal) */ void jbd2_journal_ack_err(journal_t *journal) { - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); if (journal->j_errno) journal->j_flags |= JBD2_ACK_ERR; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); } int jbd2_journal_blocks_per_page(struct inode *inode) diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c index 9c64c7ec48d..663065142b4 100644 --- a/fs/jbd2/transaction.c +++ b/fs/jbd2/transaction.c @@ -124,36 +124,38 @@ alloc_transaction: jbd_debug(3, "New handle %p going live.\n", handle); -repeat: - /* * We need to hold j_state_lock until t_updates has been incremented, * for proper journal barrier handling */ - spin_lock(&journal->j_state_lock); -repeat_locked: +repeat: + read_lock(&journal->j_state_lock); if (is_journal_aborted(journal) || (journal->j_errno != 0 && !(journal->j_flags & JBD2_ACK_ERR))) { - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); kfree(new_transaction); return -EROFS; } /* Wait on the journal's transaction barrier if necessary */ if (journal->j_barrier_count) { - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); wait_event(journal->j_wait_transaction_locked, journal->j_barrier_count == 0); goto repeat; } if (!journal->j_running_transaction) { - if (!new_transaction) { - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); + if (!new_transaction) goto alloc_transaction; + write_lock(&journal->j_state_lock); + if (!journal->j_running_transaction) { + jbd2_get_transaction(journal, new_transaction); + new_transaction = NULL; } - jbd2_get_transaction(journal, new_transaction); - new_transaction = NULL; + write_unlock(&journal->j_state_lock); + goto repeat; } transaction = journal->j_running_transaction; @@ -167,7 +169,7 @@ repeat_locked: prepare_to_wait(&journal->j_wait_transaction_locked, &wait, TASK_UNINTERRUPTIBLE); - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); schedule(); finish_wait(&journal->j_wait_transaction_locked, &wait); goto repeat; @@ -194,7 +196,7 @@ repeat_locked: prepare_to_wait(&journal->j_wait_transaction_locked, &wait, TASK_UNINTERRUPTIBLE); __jbd2_log_start_commit(journal, transaction->t_tid); - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); schedule(); finish_wait(&journal->j_wait_transaction_locked, &wait); goto repeat; @@ -228,8 +230,12 @@ repeat_locked: if (__jbd2_log_space_left(journal) < jbd_space_needed(journal)) { jbd_debug(2, "Handle %p waiting for checkpoint...\n", handle); spin_unlock(&transaction->t_handle_lock); - __jbd2_log_wait_for_space(journal); - goto repeat_locked; + read_unlock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); + if (__jbd2_log_space_left(journal) < jbd_space_needed(journal)) + __jbd2_log_wait_for_space(journal); + write_unlock(&journal->j_state_lock); + goto repeat; } /* OK, account for the buffers that this operation expects to @@ -250,7 +256,7 @@ repeat_locked: atomic_read(&transaction->t_outstanding_credits), __jbd2_log_space_left(journal)); spin_unlock(&transaction->t_handle_lock); - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); lock_map_acquire(&handle->h_lockdep_map); kfree(new_transaction); @@ -362,7 +368,7 @@ int jbd2_journal_extend(handle_t *handle, int nblocks) result = 1; - spin_lock(&journal->j_state_lock); + read_lock(&journal->j_state_lock); /* Don't extend a locked-down transaction! */ if (handle->h_transaction->t_state != T_RUNNING) { @@ -394,7 +400,7 @@ int jbd2_journal_extend(handle_t *handle, int nblocks) unlock: spin_unlock(&transaction->t_handle_lock); error_out: - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); out: return result; } @@ -432,7 +438,7 @@ int jbd2__journal_restart(handle_t *handle, int nblocks, int gfp_mask) J_ASSERT(atomic_read(&transaction->t_updates) > 0); J_ASSERT(journal_current_handle() == handle); - spin_lock(&journal->j_state_lock); + read_lock(&journal->j_state_lock); spin_lock(&transaction->t_handle_lock); atomic_sub(handle->h_buffer_credits, &transaction->t_outstanding_credits); @@ -442,7 +448,7 @@ int jbd2__journal_restart(handle_t *handle, int nblocks, int gfp_mask) jbd_debug(2, "restarting handle %p\n", handle); __jbd2_log_start_commit(journal, transaction->t_tid); - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); lock_map_release(&handle->h_lockdep_map); handle->h_buffer_credits = nblocks; @@ -472,7 +478,7 @@ void jbd2_journal_lock_updates(journal_t *journal) { DEFINE_WAIT(wait); - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); ++journal->j_barrier_count; /* Wait until there are no running updates */ @@ -490,12 +496,12 @@ void jbd2_journal_lock_updates(journal_t *journal) prepare_to_wait(&journal->j_wait_updates, &wait, TASK_UNINTERRUPTIBLE); spin_unlock(&transaction->t_handle_lock); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); schedule(); finish_wait(&journal->j_wait_updates, &wait); - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); } - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); /* * We have now established a barrier against other normal updates, but @@ -519,9 +525,9 @@ void jbd2_journal_unlock_updates (journal_t *journal) J_ASSERT(journal->j_barrier_count != 0); mutex_unlock(&journal->j_barrier); - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); --journal->j_barrier_count; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); wake_up(&journal->j_wait_transaction_locked); } @@ -1314,9 +1320,9 @@ int jbd2_journal_stop(handle_t *handle) journal->j_last_sync_writer = pid; - spin_lock(&journal->j_state_lock); + read_lock(&journal->j_state_lock); commit_time = journal->j_average_commit_time; - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); trans_time = ktime_to_ns(ktime_sub(ktime_get(), transaction->t_start_time)); @@ -1748,7 +1754,7 @@ static int journal_unmap_buffer(journal_t *journal, struct buffer_head *bh) goto zap_buffer_unlocked; /* OK, we have data buffer in journaled mode */ - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); jbd_lock_bh_state(bh); spin_lock(&journal->j_list_lock); @@ -1801,7 +1807,7 @@ static int journal_unmap_buffer(journal_t *journal, struct buffer_head *bh) jbd2_journal_put_journal_head(jh); spin_unlock(&journal->j_list_lock); jbd_unlock_bh_state(bh); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); return ret; } else { /* There is no currently-running transaction. So the @@ -1815,7 +1821,7 @@ static int journal_unmap_buffer(journal_t *journal, struct buffer_head *bh) jbd2_journal_put_journal_head(jh); spin_unlock(&journal->j_list_lock); jbd_unlock_bh_state(bh); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); return ret; } else { /* The orphan record's transaction has @@ -1839,7 +1845,7 @@ static int journal_unmap_buffer(journal_t *journal, struct buffer_head *bh) jbd2_journal_put_journal_head(jh); spin_unlock(&journal->j_list_lock); jbd_unlock_bh_state(bh); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); return 0; } else { /* Good, the buffer belongs to the running transaction. @@ -1858,7 +1864,7 @@ zap_buffer: zap_buffer_no_jh: spin_unlock(&journal->j_list_lock); jbd_unlock_bh_state(bh); - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); zap_buffer_unlocked: clear_buffer_dirty(bh); J_ASSERT_BH(bh, !buffer_jbddirty(bh)); @@ -2165,9 +2171,9 @@ int jbd2_journal_begin_ordered_truncate(journal_t *journal, /* Locks are here just to force reading of recent values, it is * enough that the transaction was not committing before we started * a transaction adding the inode to orphan list */ - spin_lock(&journal->j_state_lock); + read_lock(&journal->j_state_lock); commit_trans = journal->j_committing_transaction; - spin_unlock(&journal->j_state_lock); + read_unlock(&journal->j_state_lock); spin_lock(&journal->j_list_lock); inode_trans = jinode->i_transaction; spin_unlock(&journal->j_list_lock); diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c index 47878cf1641..9c1b92ebeb9 100644 --- a/fs/ocfs2/journal.c +++ b/fs/ocfs2/journal.c @@ -760,13 +760,13 @@ void ocfs2_set_journal_params(struct ocfs2_super *osb) if (osb->osb_commit_interval) commit_interval = osb->osb_commit_interval; - spin_lock(&journal->j_state_lock); + write_lock(&journal->j_state_lock); journal->j_commit_interval = commit_interval; if (osb->s_mount_opt & OCFS2_MOUNT_BARRIER) journal->j_flags |= JBD2_BARRIER; else journal->j_flags &= ~JBD2_BARRIER; - spin_unlock(&journal->j_state_lock); + write_unlock(&journal->j_state_lock); } int ocfs2_journal_init(struct ocfs2_journal *journal, int *dirty) diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h index a72ce21de0e..15d5743ccfb 100644 --- a/include/linux/jbd2.h +++ b/include/linux/jbd2.h @@ -764,7 +764,7 @@ struct journal_s /* * Protect the various scalars in the journal */ - spinlock_t j_state_lock; + rwlock_t j_state_lock; /* * Number of processes waiting to create a barrier lock [j_state_lock] From 8dd420466c7bfc459fa04680bd5690bfc41a4553 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 3 Aug 2010 21:38:29 -0400 Subject: [PATCH 37/40] jbd2: Remove t_handle_lock from start_this_handle() This should remove the last exclusive lock from start_this_handle(), so that we should now be able to start multiple transactions at the same time on large SMP systems. Signed-off-by: "Theodore Ts'o" --- fs/jbd2/commit.c | 3 ++- fs/jbd2/transaction.c | 33 ++++++++++++++++++++++----------- include/linux/jbd2.h | 2 +- 3 files changed, 25 insertions(+), 13 deletions(-) diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index 67bb0a2f35e..f52e5e8049f 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -1004,7 +1004,8 @@ restart_loop: * File the transaction statistics */ stats.ts_tid = commit_transaction->t_tid; - stats.run.rs_handle_count = commit_transaction->t_handle_count; + stats.run.rs_handle_count = + atomic_read(&commit_transaction->t_handle_count); trace_jbd2_run_stats(journal->j_fs_dev->bd_dev, commit_transaction->t_tid, &stats.run); diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c index 663065142b4..0752bcda535 100644 --- a/fs/jbd2/transaction.c +++ b/fs/jbd2/transaction.c @@ -57,6 +57,7 @@ jbd2_get_transaction(journal_t *journal, transaction_t *transaction) spin_lock_init(&transaction->t_handle_lock); atomic_set(&transaction->t_updates, 0); atomic_set(&transaction->t_outstanding_credits, 0); + atomic_set(&transaction->t_handle_count, 0); INIT_LIST_HEAD(&transaction->t_inode_list); INIT_LIST_HEAD(&transaction->t_private_list); @@ -180,8 +181,8 @@ repeat: * buffers requested by this operation, we need to stall pending a log * checkpoint to free some more log space. */ - spin_lock(&transaction->t_handle_lock); - needed = atomic_read(&transaction->t_outstanding_credits) + nblocks; + needed = atomic_add_return(nblocks, + &transaction->t_outstanding_credits); if (needed > journal->j_max_transaction_buffers) { /* @@ -192,7 +193,7 @@ repeat: DEFINE_WAIT(wait); jbd_debug(2, "Handle %p starting new commit...\n", handle); - spin_unlock(&transaction->t_handle_lock); + atomic_sub(nblocks, &transaction->t_outstanding_credits); prepare_to_wait(&journal->j_wait_transaction_locked, &wait, TASK_UNINTERRUPTIBLE); __jbd2_log_start_commit(journal, transaction->t_tid); @@ -229,7 +230,7 @@ repeat: */ if (__jbd2_log_space_left(journal) < jbd_space_needed(journal)) { jbd_debug(2, "Handle %p waiting for checkpoint...\n", handle); - spin_unlock(&transaction->t_handle_lock); + atomic_sub(nblocks, &transaction->t_outstanding_credits); read_unlock(&journal->j_state_lock); write_lock(&journal->j_state_lock); if (__jbd2_log_space_left(journal) < jbd_space_needed(journal)) @@ -239,23 +240,33 @@ repeat: } /* OK, account for the buffers that this operation expects to - * use and add the handle to the running transaction. */ - - if (time_after(transaction->t_start, ts)) { + * use and add the handle to the running transaction. + * + * In order for t_max_wait to be reliable, it must be + * protected by a lock. But doing so will mean that + * start_this_handle() can not be run in parallel on SMP + * systems, which limits our scalability. So we only enable + * it when debugging is enabled. We may want to use a + * separate flag, eventually, so we can enable this + * independently of debugging. + */ +#ifdef CONFIG_JBD2_DEBUG + if (jbd2_journal_enable_debug && + time_after(transaction->t_start, ts)) { ts = jbd2_time_diff(ts, transaction->t_start); + spin_lock(&transaction->t_handle_lock); if (ts > transaction->t_max_wait) transaction->t_max_wait = ts; + spin_unlock(&transaction->t_handle_lock); } - +#endif handle->h_transaction = transaction; - atomic_add(nblocks, &transaction->t_outstanding_credits); atomic_inc(&transaction->t_updates); - transaction->t_handle_count++; + atomic_inc(&transaction->t_handle_count); jbd_debug(4, "Handle %p given %d credits (total %d, free %d)\n", handle, nblocks, atomic_read(&transaction->t_outstanding_credits), __jbd2_log_space_left(journal)); - spin_unlock(&transaction->t_handle_lock); read_unlock(&journal->j_state_lock); lock_map_acquire(&handle->h_lockdep_map); diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h index 15d5743ccfb..01743b5446f 100644 --- a/include/linux/jbd2.h +++ b/include/linux/jbd2.h @@ -629,7 +629,7 @@ struct transaction_s /* * How many handles used this transaction? [t_handle_lock] */ - int t_handle_count; + atomic_t t_handle_count; /* * This transaction is being forced and some process is From 0cfc9255a1efb0467de2162950197750570ecec0 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Thu, 5 Aug 2010 01:46:37 -0400 Subject: [PATCH 38/40] ext4: re-inline ext4_rec_len_(to|from)_disk functions commit 3d0518f4, "ext4: New rec_len encoding for very large blocksizes" made several changes to this path, but from a perf perspective, un-inlining ext4_rec_len_from_disk() seems most significant. This function is called from ext4_check_dir_entry(), which on a file-creation workload is called extremely often. I tested this with bonnie: # bonnie++ -u root -s 0 -f -x 200 -d /mnt/test -n 32 (this does 200 iterations) and got this for the file creations: ext4 stock: Average = 21206.8 files/s ext4 inlined: Average = 22346.7 files/s (+5%) Signed-off-by: Eric Sandeen Signed-off-by: "Theodore Ts'o" --- fs/ext4/ext4.h | 39 +++++++++++++++++++++++++++++++++++++-- fs/ext4/namei.c | 24 ------------------------ 2 files changed, 37 insertions(+), 26 deletions(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index ed14e1db083..e03841d9f30 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1411,6 +1411,43 @@ struct ext4_dir_entry_2 { ~EXT4_DIR_ROUND) #define EXT4_MAX_REC_LEN ((1<<16)-1) +/* + * If we ever get support for fs block sizes > page_size, we'll need + * to remove the #if statements in the next two functions... + */ +static inline unsigned int +ext4_rec_len_from_disk(__le16 dlen, unsigned blocksize) +{ + unsigned len = le16_to_cpu(dlen); + +#if (PAGE_CACHE_SIZE >= 65536) + if (len == EXT4_MAX_REC_LEN || len == 0) + return blocksize; + return (len & 65532) | ((len & 3) << 16); +#else + return len; +#endif +} + +static inline __le16 ext4_rec_len_to_disk(unsigned len, unsigned blocksize) +{ + if ((len > blocksize) || (blocksize > (1 << 18)) || (len & 3)) + BUG(); +#if (PAGE_CACHE_SIZE >= 65536) + if (len < 65536) + return cpu_to_le16(len); + if (len == blocksize) { + if (blocksize == 65536) + return cpu_to_le16(EXT4_MAX_REC_LEN); + else + return cpu_to_le16(0); + } + return cpu_to_le16((len & 65532) | ((len >> 16) & 3)); +#else + return cpu_to_le16(len); +#endif +} + /* * Hash Tree Directory indexing * (c) Daniel Phillips, 2001 @@ -1636,8 +1673,6 @@ extern long ext4_compat_ioctl(struct file *, unsigned int, unsigned long); extern int ext4_ext_migrate(struct inode *); /* namei.c */ -extern unsigned int ext4_rec_len_from_disk(__le16 dlen, unsigned blocksize); -extern __le16 ext4_rec_len_to_disk(unsigned len, unsigned blocksize); extern int ext4_orphan_add(handle_t *, struct inode *); extern int ext4_orphan_del(handle_t *, struct inode *); extern int ext4_htree_fill_tree(struct file *dir_file, __u32 start_hash, diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c index ea8b59d9621..314c0d3b3fa 100644 --- a/fs/ext4/namei.c +++ b/fs/ext4/namei.c @@ -179,30 +179,6 @@ static struct buffer_head * ext4_dx_find_entry(struct inode *dir, static int ext4_dx_add_entry(handle_t *handle, struct dentry *dentry, struct inode *inode); -unsigned int ext4_rec_len_from_disk(__le16 dlen, unsigned blocksize) -{ - unsigned len = le16_to_cpu(dlen); - - if (len == EXT4_MAX_REC_LEN || len == 0) - return blocksize; - return (len & 65532) | ((len & 3) << 16); -} - -__le16 ext4_rec_len_to_disk(unsigned len, unsigned blocksize) -{ - if ((len > blocksize) || (blocksize > (1 << 18)) || (len & 3)) - BUG(); - if (len < 65536) - return cpu_to_le16(len); - if (len == blocksize) { - if (blocksize == 65536) - return cpu_to_le16(EXT4_MAX_REC_LEN); - else - return cpu_to_le16(0); - } - return cpu_to_le16((len & 65532) | ((len >> 16) & 3)); -} - /* * p is at least 6 bytes before the end of page */ From 56d35a4cd13e7bc5eca5b2dba5a41794afb17e11 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 5 Aug 2010 14:41:42 -0400 Subject: [PATCH 39/40] ext4: Fix dirtying of journalled buffers in data=journal mode In data=journal mode, we still use block_write_begin() to prepare page for writing. This function can occasionally mark buffer dirty which violates journalling assumptions - when a buffer is part of a transaction, it should be dirty and a buffer can be already part of a forget list of some transaction when block_write_begin() gets called. This violation of journalling assumptions then results in "JBD: Spotted dirty metadata buffer..." warnings. In fact, temporary dirtying the buffer while the page is still locked does not really cause problems to the journalling because we won't write the buffer until the page gets unlocked. So we just have to make sure to clear dirty bits before unlocking the page. Signed-off-by: Jan Kara --- fs/ext4/inode.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index ab2247d642c..a0ab3754d0d 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -1527,9 +1527,25 @@ static int walk_page_buffers(handle_t *handle, static int do_journal_get_write_access(handle_t *handle, struct buffer_head *bh) { + int dirty = buffer_dirty(bh); + int ret; + if (!buffer_mapped(bh) || buffer_freed(bh)) return 0; - return ext4_journal_get_write_access(handle, bh); + /* + * __block_prepare_write() could have dirtied some buffers. Clean + * the dirty bit as jbd2_journal_get_write_access() could complain + * otherwise about fs integrity issues. Setting of the dirty bit + * by __block_prepare_write() isn't a real problem here as we clear + * the bit before releasing a page lock and thus writeback cannot + * ever write the buffer. + */ + if (dirty) + clear_buffer_dirty(bh); + ret = ext4_journal_get_write_access(handle, bh); + if (!ret && dirty) + ret = ext4_handle_dirty_metadata(handle, NULL, bh); + return ret; } /* From 6c7a120ac6c62316ab1fc78dfc0a7b13f3bfcbff Mon Sep 17 00:00:00 2001 From: Aditya Kali Date: Thu, 5 Aug 2010 16:22:24 -0400 Subject: [PATCH 40/40] ext4: Adding error check after calling ext4_mb_regular_allocator() If the bitmap block on disk is bad, ext4_mb_load_buddy() returns an error. This error is returned to the caller, ext4_mb_regular_allocator() and then to ext4_mb_new_blocks(). But ext4_mb_new_blocks() did not check for the return value of ext4_mb_regular_allocator() and would repeatedly try to load the bitmap block. The fix simply catches the return value and exits out of the 'repeat' loop after cleanup. We also take the opportunity to clean up the error handling in ext4_mb_new_blocks(). Google-Bug-Id: 2853530 Signed-off-by: Aditya Kali Signed-off-by: "Theodore Ts'o" --- fs/ext4/mballoc.c | 34 +++++++++++++++++----------------- 1 file changed, 17 insertions(+), 17 deletions(-) diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index 3da28281bc5..1f360f07cb4 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -4253,7 +4253,7 @@ static int ext4_mb_discard_preallocations(struct super_block *sb, int needed) * to usual allocation */ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle, - struct ext4_allocation_request *ar, int *errp) + struct ext4_allocation_request *ar, int *errp) { int freed; struct ext4_allocation_context *ac = NULL; @@ -4297,7 +4297,7 @@ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle, inquota = ar->len; if (ar->len == 0) { *errp = -EDQUOT; - goto out3; + goto out; } } @@ -4305,13 +4305,13 @@ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle, if (!ac) { ar->len = 0; *errp = -ENOMEM; - goto out1; + goto out; } *errp = ext4_mb_initialize_context(ac, ar); if (*errp) { ar->len = 0; - goto out2; + goto out; } ac->ac_op = EXT4_MB_HISTORY_PREALLOC; @@ -4320,7 +4320,9 @@ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle, ext4_mb_normalize_request(ac, ar); repeat: /* allocate space in core */ - ext4_mb_regular_allocator(ac); + *errp = ext4_mb_regular_allocator(ac); + if (*errp) + goto errout; /* as we've just preallocated more space than * user requested orinally, we store allocated @@ -4331,7 +4333,7 @@ repeat: } if (likely(ac->ac_status == AC_STATUS_FOUND)) { *errp = ext4_mb_mark_diskspace_used(ac, handle, reserv_blks); - if (*errp == -EAGAIN) { + if (*errp == -EAGAIN) { /* * drop the reference that we took * in ext4_mb_use_best_found @@ -4342,12 +4344,10 @@ repeat: ac->ac_b_ex.fe_len = 0; ac->ac_status = AC_STATUS_CONTINUE; goto repeat; - } else if (*errp) { + } else if (*errp) + errout: ext4_discard_allocated_blocks(ac); - ac->ac_b_ex.fe_len = 0; - ar->len = 0; - ext4_mb_show_ac(ac); - } else { + else { block = ext4_grp_offs_to_block(sb, &ac->ac_b_ex); ar->len = ac->ac_b_ex.fe_len; } @@ -4356,19 +4356,19 @@ repeat: if (freed) goto repeat; *errp = -ENOSPC; + } + + if (*errp) { ac->ac_b_ex.fe_len = 0; ar->len = 0; ext4_mb_show_ac(ac); } - ext4_mb_release_context(ac); - -out2: - kmem_cache_free(ext4_ac_cachep, ac); -out1: +out: + if (ac) + kmem_cache_free(ext4_ac_cachep, ac); if (inquota && ar->len < inquota) dquot_free_block(ar->inode, inquota - ar->len); -out3: if (!ar->len) { if (!EXT4_I(ar->inode)->i_delalloc_reserved_flag) /* release all the reserved blocks if non delalloc */