From b21ebf143af219207214c79bc217beb39c43212a Mon Sep 17 00:00:00 2001 From: Harshad Shirwadkar Date: Thu, 5 Nov 2020 19:58:51 -0800 Subject: ext4: mark fc ineligible if inode gets evictied due to mem pressure If inode gets evicted due to memory pressure, we have to remove it from the fast commit list. However, that inode may have uncommitted changes that fast commits will lose. So, just fall back to full commits in this case. Also, rename the fast commit ineligiblity reason from "EXT4_FC_REASON_MEM" to "EXT4_FC_REASON_MEM_NOMEM" for better expression. Suggested-by: Jan Kara Signed-off-by: Harshad Shirwadkar Link: https://lore.kernel.org/r/20201106035911.1942128-3-harshadshirwadkar@gmail.com Signed-off-by: Theodore Ts'o --- fs/ext4/fast_commit.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs/ext4/fast_commit.c') diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c index 8d43058386c3..48b7bc129392 100644 --- a/fs/ext4/fast_commit.c +++ b/fs/ext4/fast_commit.c @@ -384,7 +384,7 @@ static int __track_dentry_update(struct inode *inode, void *arg, bool update) mutex_unlock(&ei->i_fc_lock); node = kmem_cache_alloc(ext4_fc_dentry_cachep, GFP_NOFS); if (!node) { - ext4_fc_mark_ineligible(inode->i_sb, EXT4_FC_REASON_MEM); + ext4_fc_mark_ineligible(inode->i_sb, EXT4_FC_REASON_NOMEM); mutex_lock(&ei->i_fc_lock); return -ENOMEM; } @@ -397,7 +397,7 @@ static int __track_dentry_update(struct inode *inode, void *arg, bool update) if (!node->fcd_name.name) { kmem_cache_free(ext4_fc_dentry_cachep, node); ext4_fc_mark_ineligible(inode->i_sb, - EXT4_FC_REASON_MEM); + EXT4_FC_REASON_NOMEM); mutex_lock(&ei->i_fc_lock); return -ENOMEM; } -- cgit v1.2.3 From a80f7fcf18672ae4971a6b713b58c0d389aa99fe Mon Sep 17 00:00:00 2001 From: Harshad Shirwadkar Date: Thu, 5 Nov 2020 19:58:53 -0800 Subject: ext4: fixup ext4_fc_track_* functions' signature Firstly, pass handle to all ext4_fc_track_* functions and use transaction id found in handle->h_transaction->h_tid for tracking fast commit updates. Secondly, don't pass inode to ext4_fc_track_link/create/unlink functions. inode can be found inside these functions as d_inode(dentry). However, rename path is an exeception. That's because in that case, we need inode that's not same as d_inode(dentry). To handle that, add a couple of low-level wrapper functions that take inode and dentry as arguments. Suggested-by: Jan Kara Signed-off-by: Harshad Shirwadkar Link: https://lore.kernel.org/r/20201106035911.1942128-5-harshadshirwadkar@gmail.com Signed-off-by: Theodore Ts'o --- fs/ext4/ext4.h | 16 +++++++++----- fs/ext4/extents.c | 2 +- fs/ext4/fast_commit.c | 48 +++++++++++++++++++++++++--------------- fs/ext4/inode.c | 10 ++++----- fs/ext4/namei.c | 61 +++++++++++++++++++++++---------------------------- 5 files changed, 74 insertions(+), 63 deletions(-) (limited to 'fs/ext4/fast_commit.c') diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index d513553ceafa..c6339bd75a93 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -2738,12 +2738,16 @@ extern void ext4_end_bitmap_read(struct buffer_head *bh, int uptodate); int ext4_fc_info_show(struct seq_file *seq, void *v); void ext4_fc_init(struct super_block *sb, journal_t *journal); void ext4_fc_init_inode(struct inode *inode); -void ext4_fc_track_range(struct inode *inode, ext4_lblk_t start, +void ext4_fc_track_range(handle_t *handle, struct inode *inode, ext4_lblk_t start, ext4_lblk_t end); -void ext4_fc_track_unlink(struct inode *inode, struct dentry *dentry); -void ext4_fc_track_link(struct inode *inode, struct dentry *dentry); -void ext4_fc_track_create(struct inode *inode, struct dentry *dentry); -void ext4_fc_track_inode(struct inode *inode); +void __ext4_fc_track_unlink(handle_t *handle, struct inode *inode, + struct dentry *dentry); +void __ext4_fc_track_link(handle_t *handle, struct inode *inode, + struct dentry *dentry); +void ext4_fc_track_unlink(handle_t *handle, struct dentry *dentry); +void ext4_fc_track_link(handle_t *handle, struct dentry *dentry); +void ext4_fc_track_create(handle_t *handle, struct dentry *dentry); +void ext4_fc_track_inode(handle_t *handle, struct inode *inode); void ext4_fc_mark_ineligible(struct super_block *sb, int reason); void ext4_fc_start_ineligible(struct super_block *sb, int reason); void ext4_fc_stop_ineligible(struct super_block *sb); @@ -3459,7 +3463,7 @@ extern int ext4_handle_dirty_dirblock(handle_t *handle, struct inode *inode, extern int ext4_ci_compare(const struct inode *parent, const struct qstr *fname, const struct qstr *entry, bool quick); -extern int __ext4_unlink(struct inode *dir, const struct qstr *d_name, +extern int __ext4_unlink(handle_t *handle, struct inode *dir, const struct qstr *d_name, struct inode *inode); extern int __ext4_link(struct inode *dir, struct inode *inode, struct dentry *dentry); diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index 94ba44785d28..17d7096b3212 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -4599,7 +4599,7 @@ static long ext4_zero_range(struct file *file, loff_t offset, ret = ext4_mark_inode_dirty(handle, inode); if (unlikely(ret)) goto out_handle; - ext4_fc_track_range(inode, offset >> inode->i_sb->s_blocksize_bits, + ext4_fc_track_range(handle, inode, offset >> inode->i_sb->s_blocksize_bits, (offset + len - 1) >> inode->i_sb->s_blocksize_bits); /* Zero out partial block at the edges of the range */ ret = ext4_zero_partial_blocks(handle, inode, offset, len); diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c index 48b7bc129392..9399e9cccb7e 100644 --- a/fs/ext4/fast_commit.c +++ b/fs/ext4/fast_commit.c @@ -323,13 +323,14 @@ static inline int ext4_fc_is_ineligible(struct super_block *sb) * If enqueue is set, this function enqueues the inode in fast commit list. */ static int ext4_fc_track_template( - struct inode *inode, int (*__fc_track_fn)(struct inode *, void *, bool), + handle_t *handle, struct inode *inode, + int (*__fc_track_fn)(struct inode *, void *, bool), void *args, int enqueue) { - tid_t running_txn_tid; bool update = false; struct ext4_inode_info *ei = EXT4_I(inode); struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); + tid_t tid = 0; int ret; if (!test_opt2(inode->i_sb, JOURNAL_FAST_COMMIT) || @@ -339,15 +340,13 @@ static int ext4_fc_track_template( if (ext4_fc_is_ineligible(inode->i_sb)) return -EINVAL; - running_txn_tid = sbi->s_journal ? - sbi->s_journal->j_commit_sequence + 1 : 0; - + tid = handle->h_transaction->t_tid; mutex_lock(&ei->i_fc_lock); - if (running_txn_tid == ei->i_sync_tid) { + if (tid == ei->i_sync_tid) { update = true; } else { ext4_fc_reset_inode(inode); - ei->i_sync_tid = running_txn_tid; + ei->i_sync_tid = tid; } ret = __fc_track_fn(inode, args, update); mutex_unlock(&ei->i_fc_lock); @@ -422,7 +421,8 @@ static int __track_dentry_update(struct inode *inode, void *arg, bool update) return 0; } -void ext4_fc_track_unlink(struct inode *inode, struct dentry *dentry) +void __ext4_fc_track_unlink(handle_t *handle, + struct inode *inode, struct dentry *dentry) { struct __track_dentry_update_args args; int ret; @@ -430,12 +430,18 @@ void ext4_fc_track_unlink(struct inode *inode, struct dentry *dentry) args.dentry = dentry; args.op = EXT4_FC_TAG_UNLINK; - ret = ext4_fc_track_template(inode, __track_dentry_update, + ret = ext4_fc_track_template(handle, inode, __track_dentry_update, (void *)&args, 0); trace_ext4_fc_track_unlink(inode, dentry, ret); } -void ext4_fc_track_link(struct inode *inode, struct dentry *dentry) +void ext4_fc_track_unlink(handle_t *handle, struct dentry *dentry) +{ + __ext4_fc_track_unlink(handle, d_inode(dentry), dentry); +} + +void __ext4_fc_track_link(handle_t *handle, + struct inode *inode, struct dentry *dentry) { struct __track_dentry_update_args args; int ret; @@ -443,20 +449,26 @@ void ext4_fc_track_link(struct inode *inode, struct dentry *dentry) args.dentry = dentry; args.op = EXT4_FC_TAG_LINK; - ret = ext4_fc_track_template(inode, __track_dentry_update, + ret = ext4_fc_track_template(handle, inode, __track_dentry_update, (void *)&args, 0); trace_ext4_fc_track_link(inode, dentry, ret); } -void ext4_fc_track_create(struct inode *inode, struct dentry *dentry) +void ext4_fc_track_link(handle_t *handle, struct dentry *dentry) +{ + __ext4_fc_track_link(handle, d_inode(dentry), dentry); +} + +void ext4_fc_track_create(handle_t *handle, struct dentry *dentry) { struct __track_dentry_update_args args; + struct inode *inode = d_inode(dentry); int ret; args.dentry = dentry; args.op = EXT4_FC_TAG_CREAT; - ret = ext4_fc_track_template(inode, __track_dentry_update, + ret = ext4_fc_track_template(handle, inode, __track_dentry_update, (void *)&args, 0); trace_ext4_fc_track_create(inode, dentry, ret); } @@ -472,14 +484,14 @@ static int __track_inode(struct inode *inode, void *arg, bool update) return 0; } -void ext4_fc_track_inode(struct inode *inode) +void ext4_fc_track_inode(handle_t *handle, struct inode *inode) { int ret; if (S_ISDIR(inode->i_mode)) return; - ret = ext4_fc_track_template(inode, __track_inode, NULL, 1); + ret = ext4_fc_track_template(handle, inode, __track_inode, NULL, 1); trace_ext4_fc_track_inode(inode, ret); } @@ -515,7 +527,7 @@ static int __track_range(struct inode *inode, void *arg, bool update) return 0; } -void ext4_fc_track_range(struct inode *inode, ext4_lblk_t start, +void ext4_fc_track_range(handle_t *handle, struct inode *inode, ext4_lblk_t start, ext4_lblk_t end) { struct __track_range_args args; @@ -527,7 +539,7 @@ void ext4_fc_track_range(struct inode *inode, ext4_lblk_t start, args.start = start; args.end = end; - ret = ext4_fc_track_template(inode, __track_range, &args, 1); + ret = ext4_fc_track_template(handle, inode, __track_range, &args, 1); trace_ext4_fc_track_range(inode, start, end, ret); } @@ -1263,7 +1275,7 @@ static int ext4_fc_replay_unlink(struct super_block *sb, struct ext4_fc_tl *tl) return 0; } - ret = __ext4_unlink(old_parent, &entry, inode); + ret = __ext4_unlink(NULL, old_parent, &entry, inode); /* -ENOENT ok coz it might not exist anymore. */ if (ret == -ENOENT) ret = 0; diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 081b6a86b5db..87120c4c44f3 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -732,7 +732,7 @@ out_sem: if (ret) return ret; } - ext4_fc_track_range(inode, map->m_lblk, + ext4_fc_track_range(handle, inode, map->m_lblk, map->m_lblk + map->m_len - 1); } @@ -4111,7 +4111,7 @@ int ext4_punch_hole(struct inode *inode, loff_t offset, loff_t length) up_write(&EXT4_I(inode)->i_data_sem); } - ext4_fc_track_range(inode, first_block, stop_block); + ext4_fc_track_range(handle, inode, first_block, stop_block); if (IS_SYNC(inode)) ext4_handle_sync(handle); @@ -5444,14 +5444,14 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr) } if (shrink) - ext4_fc_track_range(inode, + ext4_fc_track_range(handle, inode, (attr->ia_size > 0 ? attr->ia_size - 1 : 0) >> inode->i_sb->s_blocksize_bits, (oldsize > 0 ? oldsize - 1 : 0) >> inode->i_sb->s_blocksize_bits); else ext4_fc_track_range( - inode, + handle, inode, (oldsize > 0 ? oldsize - 1 : oldsize) >> inode->i_sb->s_blocksize_bits, (attr->ia_size > 0 ? attr->ia_size - 1 : 0) >> @@ -5701,7 +5701,7 @@ int ext4_mark_iloc_dirty(handle_t *handle, put_bh(iloc->bh); return -EIO; } - ext4_fc_track_inode(inode); + ext4_fc_track_inode(handle, inode); if (IS_I_VERSION(inode)) inode_inc_iversion(inode); diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c index f458d1d81d96..33509266f5a0 100644 --- a/fs/ext4/namei.c +++ b/fs/ext4/namei.c @@ -2606,7 +2606,7 @@ static int ext4_create(struct inode *dir, struct dentry *dentry, umode_t mode, bool excl) { handle_t *handle; - struct inode *inode, *inode_save; + struct inode *inode; int err, credits, retries = 0; err = dquot_initialize(dir); @@ -2624,11 +2624,9 @@ retry: inode->i_op = &ext4_file_inode_operations; inode->i_fop = &ext4_file_operations; ext4_set_aops(inode); - inode_save = inode; - ihold(inode_save); err = ext4_add_nondir(handle, dentry, &inode); - ext4_fc_track_create(inode_save, dentry); - iput(inode_save); + if (!err) + ext4_fc_track_create(handle, dentry); } if (handle) ext4_journal_stop(handle); @@ -2643,7 +2641,7 @@ static int ext4_mknod(struct inode *dir, struct dentry *dentry, umode_t mode, dev_t rdev) { handle_t *handle; - struct inode *inode, *inode_save; + struct inode *inode; int err, credits, retries = 0; err = dquot_initialize(dir); @@ -2660,12 +2658,9 @@ retry: if (!IS_ERR(inode)) { init_special_inode(inode, inode->i_mode, rdev); inode->i_op = &ext4_special_inode_operations; - inode_save = inode; - ihold(inode_save); err = ext4_add_nondir(handle, dentry, &inode); if (!err) - ext4_fc_track_create(inode_save, dentry); - iput(inode_save); + ext4_fc_track_create(handle, dentry); } if (handle) ext4_journal_stop(handle); @@ -2829,7 +2824,6 @@ out_clear_inode: iput(inode); goto out_retry; } - ext4_fc_track_create(inode, dentry); ext4_inc_count(dir); ext4_update_dx_flag(dir); @@ -2837,6 +2831,7 @@ out_clear_inode: if (err) goto out_clear_inode; d_instantiate_new(dentry, inode); + ext4_fc_track_create(handle, dentry); if (IS_DIRSYNC(dir)) ext4_handle_sync(handle); @@ -3171,7 +3166,7 @@ static int ext4_rmdir(struct inode *dir, struct dentry *dentry) goto end_rmdir; ext4_dec_count(dir); ext4_update_dx_flag(dir); - ext4_fc_track_unlink(inode, dentry); + ext4_fc_track_unlink(handle, dentry); retval = ext4_mark_inode_dirty(handle, dir); #ifdef CONFIG_UNICODE @@ -3192,13 +3187,12 @@ end_rmdir: return retval; } -int __ext4_unlink(struct inode *dir, const struct qstr *d_name, +int __ext4_unlink(handle_t *handle, struct inode *dir, const struct qstr *d_name, struct inode *inode) { int retval = -ENOENT; struct buffer_head *bh; struct ext4_dir_entry_2 *de; - handle_t *handle = NULL; int skip_remove_dentry = 0; bh = ext4_find_entry(dir, d_name, &de, NULL); @@ -3217,14 +3211,7 @@ int __ext4_unlink(struct inode *dir, const struct qstr *d_name, if (EXT4_SB(inode->i_sb)->s_mount_state & EXT4_FC_REPLAY) skip_remove_dentry = 1; else - goto out_bh; - } - - handle = ext4_journal_start(dir, EXT4_HT_DIR, - EXT4_DATA_TRANS_BLOCKS(dir->i_sb)); - if (IS_ERR(handle)) { - retval = PTR_ERR(handle); - goto out_bh; + goto out; } if (IS_DIRSYNC(dir)) @@ -3233,12 +3220,12 @@ int __ext4_unlink(struct inode *dir, const struct qstr *d_name, if (!skip_remove_dentry) { retval = ext4_delete_entry(handle, dir, de, bh); if (retval) - goto out_handle; + goto out; dir->i_ctime = dir->i_mtime = current_time(dir); ext4_update_dx_flag(dir); retval = ext4_mark_inode_dirty(handle, dir); if (retval) - goto out_handle; + goto out; } else { retval = 0; } @@ -3252,15 +3239,14 @@ int __ext4_unlink(struct inode *dir, const struct qstr *d_name, inode->i_ctime = current_time(inode); retval = ext4_mark_inode_dirty(handle, inode); -out_handle: - ext4_journal_stop(handle); -out_bh: +out: brelse(bh); return retval; } static int ext4_unlink(struct inode *dir, struct dentry *dentry) { + handle_t *handle; int retval; if (unlikely(ext4_forced_shutdown(EXT4_SB(dir->i_sb)))) @@ -3278,9 +3264,16 @@ static int ext4_unlink(struct inode *dir, struct dentry *dentry) if (retval) goto out_trace; - retval = __ext4_unlink(dir, &dentry->d_name, d_inode(dentry)); + handle = ext4_journal_start(dir, EXT4_HT_DIR, + EXT4_DATA_TRANS_BLOCKS(dir->i_sb)); + if (IS_ERR(handle)) { + retval = PTR_ERR(handle); + goto out_trace; + } + + retval = __ext4_unlink(handle, dir, &dentry->d_name, d_inode(dentry)); if (!retval) - ext4_fc_track_unlink(d_inode(dentry), dentry); + ext4_fc_track_unlink(handle, dentry); #ifdef CONFIG_UNICODE /* VFS negative dentries are incompatible with Encoding and * Case-insensitiveness. Eventually we'll want avoid @@ -3291,6 +3284,8 @@ static int ext4_unlink(struct inode *dir, struct dentry *dentry) if (IS_CASEFOLDED(dir)) d_invalidate(dentry); #endif + if (handle) + ext4_journal_stop(handle); out_trace: trace_ext4_unlink_exit(dentry, retval); @@ -3447,7 +3442,6 @@ retry: err = ext4_add_entry(handle, dentry, inode); if (!err) { - ext4_fc_track_link(inode, dentry); err = ext4_mark_inode_dirty(handle, inode); /* this can happen only for tmpfile being * linked the first time @@ -3455,6 +3449,7 @@ retry: if (inode->i_nlink == 1) ext4_orphan_del(handle, inode); d_instantiate(dentry, inode); + ext4_fc_track_link(handle, dentry); } else { drop_nlink(inode); iput(inode); @@ -3915,9 +3910,9 @@ static int ext4_rename(struct inode *old_dir, struct dentry *old_dentry, EXT4_FC_REASON_RENAME_DIR); } else { if (new.inode) - ext4_fc_track_unlink(new.inode, new.dentry); - ext4_fc_track_link(old.inode, new.dentry); - ext4_fc_track_unlink(old.inode, old.dentry); + ext4_fc_track_unlink(handle, new.dentry); + __ext4_fc_track_link(handle, old.inode, new.dentry); + __ext4_fc_track_unlink(handle, old.inode, old.dentry); } if (new.inode) { -- cgit v1.2.3 From a1e5e465b31d6015fccb359d99053b39e5180466 Mon Sep 17 00:00:00 2001 From: Harshad Shirwadkar Date: Thu, 5 Nov 2020 19:58:55 -0800 Subject: ext4: clean up the JBD2 API that initializes fast commits This patch removes jbd2_fc_init() API and its related functions to simplify enabling fast commits. With this change, the number of fast commit blocks to use is solely determined by the JBD2 layer. So, we move the default value for minimum number of fast commit blocks from ext4/fast_commit.h to include/linux/jbd2.h. However, whether or not to use fast commits is determined by the file system. The file system just sets the fast commit feature using jbd2_journal_set_features(). JBD2 layer then determines how many blocks to use for fast commits (based on the value found in the JBD2 superblock). Note that the JBD2 feature flag of fast commits is just an indication that there are fast commit blocks present on disk. It doesn't tell JBD2 layer about the intent of the file system of whether to it wants to use fast commit or not. That's why, we blindly clear the fast commit flag in journal_reset() after the recovery is done. Suggested-by: Jan Kara Signed-off-by: Harshad Shirwadkar Link: https://lore.kernel.org/r/20201106035911.1942128-7-harshadshirwadkar@gmail.com Signed-off-by: Theodore Ts'o --- fs/ext4/fast_commit.c | 14 -------- fs/ext4/fast_commit.h | 3 -- fs/ext4/super.c | 8 +++++ fs/jbd2/journal.c | 96 ++++++++++++++++++++++++++++++--------------------- include/linux/jbd2.h | 2 +- 5 files changed, 65 insertions(+), 58 deletions(-) (limited to 'fs/ext4/fast_commit.c') diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c index 9399e9cccb7e..bab60c5d5095 100644 --- a/fs/ext4/fast_commit.c +++ b/fs/ext4/fast_commit.c @@ -2091,8 +2091,6 @@ static int ext4_fc_replay(journal_t *journal, struct buffer_head *bh, void ext4_fc_init(struct super_block *sb, journal_t *journal) { - int num_fc_blocks; - /* * We set replay callback even if fast commit disabled because we may * could still have fast commit blocks that need to be replayed even if @@ -2102,18 +2100,6 @@ void ext4_fc_init(struct super_block *sb, journal_t *journal) if (!test_opt2(sb, JOURNAL_FAST_COMMIT)) return; journal->j_fc_cleanup_callback = ext4_fc_cleanup; - if (!buffer_uptodate(journal->j_sb_buffer) - && ext4_read_bh_lock(journal->j_sb_buffer, REQ_META | REQ_PRIO, - true)) { - ext4_msg(sb, KERN_ERR, "I/O error on journal"); - return; - } - num_fc_blocks = be32_to_cpu(journal->j_superblock->s_num_fc_blks); - if (jbd2_fc_init(journal, num_fc_blocks ? num_fc_blocks : - EXT4_NUM_FC_BLKS)) { - pr_warn("Error while enabling fast commits, turning off."); - ext4_clear_feature_fast_commit(sb); - } } const char *fc_ineligible_reasons[] = { diff --git a/fs/ext4/fast_commit.h b/fs/ext4/fast_commit.h index 140fbb6af19e..1d96e0ac8138 100644 --- a/fs/ext4/fast_commit.h +++ b/fs/ext4/fast_commit.h @@ -3,9 +3,6 @@ #ifndef __FAST_COMMIT_H__ #define __FAST_COMMIT_H__ -/* Number of blocks in journal area to allocate for fast commits */ -#define EXT4_NUM_FC_BLKS 256 - /* Fast commit tags */ #define EXT4_FC_TAG_ADD_RANGE 0x0001 #define EXT4_FC_TAG_DEL_RANGE 0x0002 diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 8a6dd433bb70..ba02d7c86fb3 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -4857,6 +4857,14 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) goto failed_mount_wq; } + if (test_opt2(sb, JOURNAL_FAST_COMMIT) && + !jbd2_journal_set_features(EXT4_SB(sb)->s_journal, 0, 0, + JBD2_FEATURE_INCOMPAT_FAST_COMMIT)) { + ext4_msg(sb, KERN_ERR, + "Failed to set fast commit journal feature"); + goto failed_mount_wq; + } + /* We have now updated the journal if required, so we can * validate the data journaling mode. */ switch (test_opt(sb, DATA_FLAGS)) { diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c index c3c768248527..500152f0421a 100644 --- a/fs/jbd2/journal.c +++ b/fs/jbd2/journal.c @@ -1352,19 +1352,12 @@ static journal_t *journal_init_common(struct block_device *bdev, /* We need enough buffers to write out full descriptor block. */ n = journal->j_blocksize / jbd2_min_tag_size(); journal->j_wbufsize = n; + journal->j_fc_wbuf = NULL; journal->j_wbuf = kmalloc_array(n, sizeof(struct buffer_head *), GFP_KERNEL); if (!journal->j_wbuf) goto err_cleanup; - if (journal->j_fc_wbufsize > 0) { - journal->j_fc_wbuf = kmalloc_array(journal->j_fc_wbufsize, - sizeof(struct buffer_head *), - GFP_KERNEL); - if (!journal->j_fc_wbuf) - goto err_cleanup; - } - bh = getblk_unmovable(journal->j_dev, start, journal->j_blocksize); if (!bh) { pr_err("%s: Cannot get buffer for journal superblock\n", @@ -1378,23 +1371,11 @@ static journal_t *journal_init_common(struct block_device *bdev, err_cleanup: kfree(journal->j_wbuf); - kfree(journal->j_fc_wbuf); jbd2_journal_destroy_revoke(journal); kfree(journal); return NULL; } -int jbd2_fc_init(journal_t *journal, int num_fc_blks) -{ - journal->j_fc_wbufsize = num_fc_blks; - journal->j_fc_wbuf = kmalloc_array(journal->j_fc_wbufsize, - sizeof(struct buffer_head *), GFP_KERNEL); - if (!journal->j_fc_wbuf) - return -ENOMEM; - return 0; -} -EXPORT_SYMBOL(jbd2_fc_init); - /* jbd2_journal_init_dev and jbd2_journal_init_inode: * * Create a journal structure assigned some fixed set of disk blocks to @@ -1512,16 +1493,7 @@ static int journal_reset(journal_t *journal) } journal->j_first = first; - - if (jbd2_has_feature_fast_commit(journal) && - journal->j_fc_wbufsize > 0) { - journal->j_fc_last = last; - journal->j_last = last - journal->j_fc_wbufsize; - journal->j_fc_first = journal->j_last + 1; - journal->j_fc_off = 0; - } else { - journal->j_last = last; - } + journal->j_last = last; journal->j_head = journal->j_first; journal->j_tail = journal->j_first; @@ -1533,6 +1505,13 @@ static int journal_reset(journal_t *journal) journal->j_max_transaction_buffers = jbd2_journal_get_max_txn_bufs(journal); + /* + * Now that journal recovery is done, turn fast commits off here. This + * way, if fast commit was enabled before the crash but if now FS has + * disabled it, we don't enable fast commits. + */ + jbd2_clear_feature_fast_commit(journal); + /* * As a special case, if the on-disk copy is already marked as needing * no recovery (s_start == 0), then we can safely defer the superblock @@ -1872,6 +1851,7 @@ static int load_superblock(journal_t *journal) { int err; journal_superblock_t *sb; + int num_fc_blocks; err = journal_get_superblock(journal); if (err) @@ -1883,15 +1863,17 @@ static int load_superblock(journal_t *journal) journal->j_tail = be32_to_cpu(sb->s_start); journal->j_first = be32_to_cpu(sb->s_first); journal->j_errno = be32_to_cpu(sb->s_errno); + journal->j_last = be32_to_cpu(sb->s_maxlen); - if (jbd2_has_feature_fast_commit(journal) && - journal->j_fc_wbufsize > 0) { + if (jbd2_has_feature_fast_commit(journal)) { journal->j_fc_last = be32_to_cpu(sb->s_maxlen); - journal->j_last = journal->j_fc_last - journal->j_fc_wbufsize; + num_fc_blocks = be32_to_cpu(sb->s_num_fc_blks); + if (!num_fc_blocks) + num_fc_blocks = JBD2_MIN_FC_BLOCKS; + if (journal->j_last - num_fc_blocks >= JBD2_MIN_JOURNAL_BLOCKS) + journal->j_last = journal->j_fc_last - num_fc_blocks; journal->j_fc_first = journal->j_last + 1; journal->j_fc_off = 0; - } else { - journal->j_last = be32_to_cpu(sb->s_maxlen); } return 0; @@ -1954,9 +1936,6 @@ int jbd2_journal_load(journal_t *journal) */ journal->j_flags &= ~JBD2_ABORT; - if (journal->j_fc_wbufsize > 0) - jbd2_journal_set_features(journal, 0, 0, - JBD2_FEATURE_INCOMPAT_FAST_COMMIT); /* OK, we've finished with the dynamic journal bits: * reinitialise the dynamic contents of the superblock in memory * and reset them on disk. */ @@ -2040,8 +2019,7 @@ int jbd2_journal_destroy(journal_t *journal) jbd2_journal_destroy_revoke(journal); if (journal->j_chksum_driver) crypto_free_shash(journal->j_chksum_driver); - if (journal->j_fc_wbufsize > 0) - kfree(journal->j_fc_wbuf); + kfree(journal->j_fc_wbuf); kfree(journal->j_wbuf); kfree(journal); @@ -2116,6 +2094,37 @@ int jbd2_journal_check_available_features(journal_t *journal, unsigned long comp return 0; } +static int +jbd2_journal_initialize_fast_commit(journal_t *journal) +{ + journal_superblock_t *sb = journal->j_superblock; + unsigned long long num_fc_blks; + + num_fc_blks = be32_to_cpu(sb->s_num_fc_blks); + if (num_fc_blks == 0) + num_fc_blks = JBD2_MIN_FC_BLOCKS; + if (journal->j_last - num_fc_blks < JBD2_MIN_JOURNAL_BLOCKS) + return -ENOSPC; + + /* Are we called twice? */ + WARN_ON(journal->j_fc_wbuf != NULL); + journal->j_fc_wbuf = kmalloc_array(num_fc_blks, + sizeof(struct buffer_head *), GFP_KERNEL); + if (!journal->j_fc_wbuf) + return -ENOMEM; + + journal->j_fc_wbufsize = num_fc_blks; + journal->j_fc_last = journal->j_last; + journal->j_last = journal->j_fc_last - num_fc_blks; + journal->j_fc_first = journal->j_last + 1; + journal->j_fc_off = 0; + journal->j_free = journal->j_last - journal->j_first; + journal->j_max_transaction_buffers = + jbd2_journal_get_max_txn_bufs(journal); + + return 0; +} + /** * int jbd2_journal_set_features() - Mark a given journal feature in the superblock * @journal: Journal to act on. @@ -2159,6 +2168,13 @@ int jbd2_journal_set_features(journal_t *journal, unsigned long compat, sb = journal->j_superblock; + if (incompat & JBD2_FEATURE_INCOMPAT_FAST_COMMIT) { + if (jbd2_journal_initialize_fast_commit(journal)) { + pr_err("JBD2: Cannot enable fast commits.\n"); + return 0; + } + } + /* Load the checksum driver if necessary */ if ((journal->j_chksum_driver == NULL) && INCOMPAT_FEATURE_ON(JBD2_FEATURE_INCOMPAT_CSUM_V3)) { diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h index e0b6b53eae64..b2caf7bbd8e5 100644 --- a/include/linux/jbd2.h +++ b/include/linux/jbd2.h @@ -68,6 +68,7 @@ extern void *jbd2_alloc(size_t size, gfp_t flags); extern void jbd2_free(void *ptr, size_t size); #define JBD2_MIN_JOURNAL_BLOCKS 1024 +#define JBD2_MIN_FC_BLOCKS 256 #ifdef __KERNEL__ @@ -1614,7 +1615,6 @@ extern void __jbd2_journal_drop_transaction(journal_t *, transaction_t *); extern int jbd2_cleanup_journal_tail(journal_t *); /* Fast commit related APIs */ -int jbd2_fc_init(journal_t *journal, int num_fc_blks); int jbd2_fc_begin_commit(journal_t *journal, tid_t tid); int jbd2_fc_end_commit(journal_t *journal); int jbd2_fc_end_commit_fallback(journal_t *journal, tid_t tid); -- cgit v1.2.3 From 0bce577bf9cae13ae32d391432d0030e3f67fc1d Mon Sep 17 00:00:00 2001 From: Harshad Shirwadkar Date: Thu, 5 Nov 2020 19:58:58 -0800 Subject: jbd2: don't pass tid to jbd2_fc_end_commit_fallback() In jbd2_fc_end_commit_fallback(), we know which tid to commit. There's no need for caller to pass it. Suggested-by: Jan Kara Signed-off-by: Harshad Shirwadkar Link: https://lore.kernel.org/r/20201106035911.1942128-10-harshadshirwadkar@gmail.com Signed-off-by: Theodore Ts'o --- fs/ext4/fast_commit.c | 2 +- fs/jbd2/journal.c | 12 +++++++++--- include/linux/jbd2.h | 2 +- 3 files changed, 11 insertions(+), 5 deletions(-) (limited to 'fs/ext4/fast_commit.c') diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c index bab60c5d5095..e69c580fa91e 100644 --- a/fs/ext4/fast_commit.c +++ b/fs/ext4/fast_commit.c @@ -1143,7 +1143,7 @@ out: "Fast commit ended with blks = %d, reason = %d, subtid - %d", nblks, reason, subtid); if (reason == EXT4_FC_REASON_FC_FAILED) - return jbd2_fc_end_commit_fallback(journal, commit_tid); + return jbd2_fc_end_commit_fallback(journal); if (reason == EXT4_FC_REASON_FC_START_FAILED || reason == EXT4_FC_REASON_INELIGIBLE) return jbd2_complete_transaction(journal, commit_tid); diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c index 778ea50fc8d1..59166e299cde 100644 --- a/fs/jbd2/journal.c +++ b/fs/jbd2/journal.c @@ -777,13 +777,19 @@ static int __jbd2_fc_end_commit(journal_t *journal, tid_t tid, bool fallback) int jbd2_fc_end_commit(journal_t *journal) { - return __jbd2_fc_end_commit(journal, 0, 0); + return __jbd2_fc_end_commit(journal, 0, false); } EXPORT_SYMBOL(jbd2_fc_end_commit); -int jbd2_fc_end_commit_fallback(journal_t *journal, tid_t tid) +int jbd2_fc_end_commit_fallback(journal_t *journal) { - return __jbd2_fc_end_commit(journal, tid, 1); + tid_t tid; + + read_lock(&journal->j_state_lock); + tid = journal->j_running_transaction ? + journal->j_running_transaction->t_tid : 0; + read_unlock(&journal->j_state_lock); + return __jbd2_fc_end_commit(journal, tid, true); } EXPORT_SYMBOL(jbd2_fc_end_commit_fallback); diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h index 5f0ef6380b0c..1c49fd62ff2e 100644 --- a/include/linux/jbd2.h +++ b/include/linux/jbd2.h @@ -1619,7 +1619,7 @@ extern int jbd2_cleanup_journal_tail(journal_t *); /* Fast commit related APIs */ int jbd2_fc_begin_commit(journal_t *journal, tid_t tid); int jbd2_fc_end_commit(journal_t *journal); -int jbd2_fc_end_commit_fallback(journal_t *journal, tid_t tid); +int jbd2_fc_end_commit_fallback(journal_t *journal); int jbd2_fc_get_buf(journal_t *journal, struct buffer_head **bh_out); int jbd2_submit_inode_data(struct jbd2_inode *jinode); int jbd2_wait_inode_data(journal_t *journal, struct jbd2_inode *jinode); -- cgit v1.2.3 From f6634e2609d13d7aa8852734e16300845db915d5 Mon Sep 17 00:00:00 2001 From: Harshad Shirwadkar Date: Thu, 5 Nov 2020 19:59:02 -0800 Subject: ext4: dedpulicate the code to wait on inode that's being committed This patch removes the deduplicates the code that implements waiting on inode that's being committed. That code is moved into a new function. Suggested-by: Jan Kara Signed-off-by: Harshad Shirwadkar Link: https://lore.kernel.org/r/20201106035911.1942128-14-harshadshirwadkar@gmail.com Signed-off-by: Theodore Ts'o --- fs/ext4/fast_commit.c | 61 +++++++++++++++++++++++---------------------------- 1 file changed, 27 insertions(+), 34 deletions(-) (limited to 'fs/ext4/fast_commit.c') diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c index e69c580fa91e..fc5a5e6a581d 100644 --- a/fs/ext4/fast_commit.c +++ b/fs/ext4/fast_commit.c @@ -155,6 +155,30 @@ void ext4_fc_init_inode(struct inode *inode) ei->i_fc_committed_subtid = 0; } +/* This function must be called with sbi->s_fc_lock held. */ +static void ext4_fc_wait_committing_inode(struct inode *inode) +{ + wait_queue_head_t *wq; + struct ext4_inode_info *ei = EXT4_I(inode); + +#if (BITS_PER_LONG < 64) + DEFINE_WAIT_BIT(wait, &ei->i_state_flags, + EXT4_STATE_FC_COMMITTING); + wq = bit_waitqueue(&ei->i_state_flags, + EXT4_STATE_FC_COMMITTING); +#else + DEFINE_WAIT_BIT(wait, &ei->i_flags, + EXT4_STATE_FC_COMMITTING); + wq = bit_waitqueue(&ei->i_flags, + EXT4_STATE_FC_COMMITTING); +#endif + lockdep_assert_held(&EXT4_SB(inode->i_sb)->s_fc_lock); + prepare_to_wait(wq, &wait.wq_entry, TASK_UNINTERRUPTIBLE); + spin_unlock(&EXT4_SB(inode->i_sb)->s_fc_lock); + schedule(); + finish_wait(wq, &wait.wq_entry); +} + /* * Inform Ext4's fast about start of an inode update * @@ -176,22 +200,7 @@ restart: goto out; if (ext4_test_inode_state(inode, EXT4_STATE_FC_COMMITTING)) { - wait_queue_head_t *wq; -#if (BITS_PER_LONG < 64) - DEFINE_WAIT_BIT(wait, &ei->i_state_flags, - EXT4_STATE_FC_COMMITTING); - wq = bit_waitqueue(&ei->i_state_flags, - EXT4_STATE_FC_COMMITTING); -#else - DEFINE_WAIT_BIT(wait, &ei->i_flags, - EXT4_STATE_FC_COMMITTING); - wq = bit_waitqueue(&ei->i_flags, - EXT4_STATE_FC_COMMITTING); -#endif - prepare_to_wait(wq, &wait.wq_entry, TASK_UNINTERRUPTIBLE); - spin_unlock(&EXT4_SB(inode->i_sb)->s_fc_lock); - schedule(); - finish_wait(wq, &wait.wq_entry); + ext4_fc_wait_committing_inode(inode); goto restart; } out: @@ -234,26 +243,10 @@ restart: } if (ext4_test_inode_state(inode, EXT4_STATE_FC_COMMITTING)) { - wait_queue_head_t *wq; -#if (BITS_PER_LONG < 64) - DEFINE_WAIT_BIT(wait, &ei->i_state_flags, - EXT4_STATE_FC_COMMITTING); - wq = bit_waitqueue(&ei->i_state_flags, - EXT4_STATE_FC_COMMITTING); -#else - DEFINE_WAIT_BIT(wait, &ei->i_flags, - EXT4_STATE_FC_COMMITTING); - wq = bit_waitqueue(&ei->i_flags, - EXT4_STATE_FC_COMMITTING); -#endif - prepare_to_wait(wq, &wait.wq_entry, TASK_UNINTERRUPTIBLE); - spin_unlock(&EXT4_SB(inode->i_sb)->s_fc_lock); - schedule(); - finish_wait(wq, &wait.wq_entry); + ext4_fc_wait_committing_inode(inode); goto restart; } - if (!list_empty(&ei->i_fc_list)) - list_del_init(&ei->i_fc_list); + list_del_init(&ei->i_fc_list); spin_unlock(&EXT4_SB(inode->i_sb)->s_fc_lock); } -- cgit v1.2.3 From a740762fb3b36dbdddb63ebe65b71cea3014f1c3 Mon Sep 17 00:00:00 2001 From: Harshad Shirwadkar Date: Thu, 5 Nov 2020 19:59:03 -0800 Subject: ext4: fix code documentatioon Add a TODO to remember fixing REQ_FUA | REQ_PREFLUSH for fast commit buffers. Also, fix a typo in top level comment in fast_commit.c Signed-off-by: Harshad Shirwadkar Link: https://lore.kernel.org/r/20201106035911.1942128-15-harshadshirwadkar@gmail.com Signed-off-by: Theodore Ts'o --- fs/ext4/fast_commit.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs/ext4/fast_commit.c') diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c index fc5a5e6a581d..639b2a308c7b 100644 --- a/fs/ext4/fast_commit.c +++ b/fs/ext4/fast_commit.c @@ -83,7 +83,7 @@ * * Atomicity of commits * -------------------- - * In order to gaurantee atomicity during the commit operation, fast commit + * In order to guarantee atomicity during the commit operation, fast commit * uses "EXT4_FC_TAG_TAIL" tag that marks a fast commit as complete. Tail * tag contains CRC of the contents and TID of the transaction after which * this fast commit should be applied. Recovery code replays fast commit @@ -542,6 +542,7 @@ static void ext4_fc_submit_bh(struct super_block *sb) int write_flags = REQ_SYNC; struct buffer_head *bh = EXT4_SB(sb)->s_fc_bh; + /* TODO: REQ_FUA | REQ_PREFLUSH is unnecessarily expensive. */ if (test_opt(sb, BARRIER)) write_flags |= REQ_FUA | REQ_PREFLUSH; lock_buffer(bh); -- cgit v1.2.3 From 764b3fd31d131c4b8b5fa064aa94382091923aec Mon Sep 17 00:00:00 2001 From: Harshad Shirwadkar Date: Thu, 5 Nov 2020 19:59:04 -0800 Subject: ext4: mark buf dirty before submitting fast commit buffer Mark the fast commit buffer as dirty before submission. Suggested-by: Jan Kara Signed-off-by: Harshad Shirwadkar Link: https://lore.kernel.org/r/20201106035911.1942128-16-harshadshirwadkar@gmail.com Signed-off-by: Theodore Ts'o --- fs/ext4/fast_commit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/ext4/fast_commit.c') diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c index 639b2a308c7b..05e6e76a7663 100644 --- a/fs/ext4/fast_commit.c +++ b/fs/ext4/fast_commit.c @@ -546,7 +546,7 @@ static void ext4_fc_submit_bh(struct super_block *sb) if (test_opt(sb, BARRIER)) write_flags |= REQ_FUA | REQ_PREFLUSH; lock_buffer(bh); - clear_buffer_dirty(bh); + set_buffer_dirty(bh); set_buffer_uptodate(bh); bh->b_end_io = ext4_end_buffer_io_sync; submit_bh(REQ_OP_WRITE, write_flags, bh); -- cgit v1.2.3 From 1ceecb537f72734e4315638e7a1bb62e56c86fbf Mon Sep 17 00:00:00 2001 From: Harshad Shirwadkar Date: Thu, 5 Nov 2020 19:59:06 -0800 Subject: ext4: fix inode dirty check in case of fast commits In case of fast commits, determine if the inode is dirty by checking if the inode is on fast commit list. This also helps us get rid of ext4_inode_info.i_fc_committed_subtid field. Reported-by: Andrea Righi Tested-by: Andrea Righi Reviewed-by: Jan Kara Signed-off-by: Harshad Shirwadkar Link: https://lore.kernel.org/r/20201106035911.1942128-18-harshadshirwadkar@gmail.com Signed-off-by: Theodore Ts'o --- fs/ext4/ext4.h | 3 --- fs/ext4/fast_commit.c | 3 --- fs/ext4/inode.c | 3 +-- 3 files changed, 1 insertion(+), 8 deletions(-) (limited to 'fs/ext4/fast_commit.c') diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index c6339bd75a93..2a89c5bbb61b 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1028,9 +1028,6 @@ struct ext4_inode_info { * protected by sbi->s_fc_lock. */ - /* Fast commit subtid when this inode was committed */ - unsigned int i_fc_committed_subtid; - /* Start of lblk range that needs to be committed in this fast commit */ ext4_lblk_t i_fc_lblk_start; diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c index 05e6e76a7663..6b963e09af2c 100644 --- a/fs/ext4/fast_commit.c +++ b/fs/ext4/fast_commit.c @@ -152,7 +152,6 @@ void ext4_fc_init_inode(struct inode *inode) INIT_LIST_HEAD(&ei->i_fc_list); init_waitqueue_head(&ei->i_fc_wait); atomic_set(&ei->i_fc_updates, 0); - ei->i_fc_committed_subtid = 0; } /* This function must be called with sbi->s_fc_lock held. */ @@ -1037,8 +1036,6 @@ static int ext4_fc_perform_commit(journal_t *journal) if (ret) goto out; spin_lock(&sbi->s_fc_lock); - EXT4_I(inode)->i_fc_committed_subtid = - atomic_read(&sbi->s_fc_subtid); } spin_unlock(&sbi->s_fc_lock); diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 87120c4c44f3..000bf70e88ed 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -3312,8 +3312,7 @@ static bool ext4_inode_datasync_dirty(struct inode *inode) EXT4_I(inode)->i_datasync_tid)) return false; if (test_opt2(inode->i_sb, JOURNAL_FAST_COMMIT)) - return atomic_read(&EXT4_SB(inode->i_sb)->s_fc_subtid) < - EXT4_I(inode)->i_fc_committed_subtid; + return !list_empty(&EXT4_I(inode)->i_fc_list); return true; } -- cgit v1.2.3 From 556e0319fbb8eee3fa19fdcc27c8bcb4af1c7211 Mon Sep 17 00:00:00 2001 From: Harshad Shirwadkar Date: Thu, 5 Nov 2020 19:59:07 -0800 Subject: ext4: disable fast commit with data journalling Fast commits don't work with data journalling. This patch disables the fast commit support when data journalling is turned on. Suggested-by: Jan Kara Signed-off-by: Harshad Shirwadkar Reviewed-by: Jan Kara Link: https://lore.kernel.org/r/20201106035911.1942128-19-harshadshirwadkar@gmail.com Signed-off-by: Theodore Ts'o --- fs/ext4/fast_commit.c | 7 +++++++ fs/ext4/fast_commit.h | 1 + fs/ext4/super.c | 3 ++- include/trace/events/ext4.h | 6 ++++-- 4 files changed, 14 insertions(+), 3 deletions(-) (limited to 'fs/ext4/fast_commit.c') diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c index 6b963e09af2c..fb9b4e9d82b2 100644 --- a/fs/ext4/fast_commit.c +++ b/fs/ext4/fast_commit.c @@ -483,6 +483,12 @@ void ext4_fc_track_inode(handle_t *handle, struct inode *inode) if (S_ISDIR(inode->i_mode)) return; + if (ext4_should_journal_data(inode)) { + ext4_fc_mark_ineligible(inode->i_sb, + EXT4_FC_REASON_INODE_JOURNAL_DATA); + return; + } + ret = ext4_fc_track_template(handle, inode, __track_inode, NULL, 1); trace_ext4_fc_track_inode(inode, ret); } @@ -2102,6 +2108,7 @@ const char *fc_ineligible_reasons[] = { "Resize", "Dir renamed", "Falloc range op", + "Data journalling", "FC Commit Failed" }; diff --git a/fs/ext4/fast_commit.h b/fs/ext4/fast_commit.h index 1d96e0ac8138..3a6e5a1fa1b8 100644 --- a/fs/ext4/fast_commit.h +++ b/fs/ext4/fast_commit.h @@ -102,6 +102,7 @@ enum { EXT4_FC_REASON_RESIZE, EXT4_FC_REASON_RENAME_DIR, EXT4_FC_REASON_FALLOC_RANGE, + EXT4_FC_REASON_INODE_JOURNAL_DATA, EXT4_FC_COMMIT_FAILED, EXT4_FC_REASON_MAX }; diff --git a/fs/ext4/super.c b/fs/ext4/super.c index ba02d7c86fb3..9e2e06ed2c45 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -4340,9 +4340,10 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) #endif if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA) { - printk_once(KERN_WARNING "EXT4-fs: Warning: mounting with data=journal disables delayed allocation, dioread_nolock, and O_DIRECT support!\n"); + printk_once(KERN_WARNING "EXT4-fs: Warning: mounting with data=journal disables delayed allocation, dioread_nolock, O_DIRECT and fast_commit support!\n"); /* can't mount with both data=journal and dioread_nolock. */ clear_opt(sb, DIOREAD_NOLOCK); + clear_opt2(sb, JOURNAL_FAST_COMMIT); if (test_opt2(sb, EXPLICIT_DELALLOC)) { ext4_msg(sb, KERN_ERR, "can't mount with " "both data=journal and delalloc"); diff --git a/include/trace/events/ext4.h b/include/trace/events/ext4.h index 0f899d3b09d3..70ae5497b73a 100644 --- a/include/trace/events/ext4.h +++ b/include/trace/events/ext4.h @@ -104,7 +104,8 @@ TRACE_DEFINE_ENUM(ES_REFERENCED_B); { EXT4_FC_REASON_SWAP_BOOT, "SWAP_BOOT"}, \ { EXT4_FC_REASON_RESIZE, "RESIZE"}, \ { EXT4_FC_REASON_RENAME_DIR, "RENAME_DIR"}, \ - { EXT4_FC_REASON_FALLOC_RANGE, "FALLOC_RANGE"}) + { EXT4_FC_REASON_FALLOC_RANGE, "FALLOC_RANGE"}, \ + { EXT4_FC_REASON_INODE_JOURNAL_DATA, "INODE_JOURNAL_DATA"}) TRACE_EVENT(ext4_other_inode_update_time, TP_PROTO(struct inode *inode, ino_t orig_ino), @@ -2917,7 +2918,7 @@ TRACE_EVENT(ext4_fc_stats, ), TP_printk("dev %d:%d fc ineligible reasons:\n" - "%s:%d, %s:%d, %s:%d, %s:%d, %s:%d, %s:%d, %s:%d, %s:%d; " + "%s:%d, %s:%d, %s:%d, %s:%d, %s:%d, %s:%d, %s:%d, %s:%d, %s:%d; " "num_commits:%ld, ineligible: %ld, numblks: %ld", MAJOR(__entry->dev), MINOR(__entry->dev), FC_REASON_NAME_STAT(EXT4_FC_REASON_XATTR), @@ -2928,6 +2929,7 @@ TRACE_EVENT(ext4_fc_stats, FC_REASON_NAME_STAT(EXT4_FC_REASON_RESIZE), FC_REASON_NAME_STAT(EXT4_FC_REASON_RENAME_DIR), FC_REASON_NAME_STAT(EXT4_FC_REASON_FALLOC_RANGE), + FC_REASON_NAME_STAT(EXT4_FC_REASON_INODE_JOURNAL_DATA), __entry->sbi->s_fc_stats.fc_num_commits, __entry->sbi->s_fc_stats.fc_ineligible_commits, __entry->sbi->s_fc_stats.fc_numblks) -- cgit v1.2.3 From da0c5d2695265962f20099737348fcb3ff524d0f Mon Sep 17 00:00:00 2001 From: Harshad Shirwadkar Date: Thu, 5 Nov 2020 19:59:08 -0800 Subject: ext4: issue fsdev cache flush before starting fast commit If the journal dev is different from fsdev, issue a cache flush before committing fast commit blocks to disk. Suggested-by: Jan Kara Signed-off-by: Harshad Shirwadkar Reviewed-by: Jan Kara Link: https://lore.kernel.org/r/20201106035911.1942128-20-harshadshirwadkar@gmail.com Signed-off-by: Theodore Ts'o --- fs/ext4/fast_commit.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'fs/ext4/fast_commit.c') diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c index fb9b4e9d82b2..ebe5f423f8f2 100644 --- a/fs/ext4/fast_commit.c +++ b/fs/ext4/fast_commit.c @@ -1007,6 +1007,13 @@ static int ext4_fc_perform_commit(journal_t *journal) if (ret) return ret; + /* + * If file system device is different from journal device, issue a cache + * flush before we start writing fast commit blocks. + */ + if (journal->j_fs_dev != journal->j_dev) + blkdev_issue_flush(journal->j_fs_dev, GFP_NOFS); + blk_start_plug(&plug); if (sbi->s_fc_bytes == 0) { /* -- cgit v1.2.3 From 9b5f6c9b83d912c63ef9fb486a052be79b06f8b0 Mon Sep 17 00:00:00 2001 From: Harshad Shirwadkar Date: Thu, 5 Nov 2020 19:59:09 -0800 Subject: ext4: make s_mount_flags modifications atomic Fast commit file system states are recorded in sbi->s_mount_flags. Fast commit expects these bit manipulations to be atomic. This patch adds helpers to make those modifications atomic. Suggested-by: Jan Kara Signed-off-by: Harshad Shirwadkar Link: https://lore.kernel.org/r/20201106035911.1942128-21-harshadshirwadkar@gmail.com Signed-off-by: Theodore Ts'o --- fs/ext4/ext4.h | 40 +++++++++++++++++++++++++++++----------- fs/ext4/fast_commit.c | 18 +++++++++--------- fs/ext4/file.c | 4 ++-- fs/ext4/fsync.c | 2 +- fs/ext4/inode.c | 4 ++-- fs/ext4/mballoc.c | 4 ++-- fs/ext4/super.c | 14 +++++++------- 7 files changed, 52 insertions(+), 34 deletions(-) (limited to 'fs/ext4/fast_commit.c') diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 2a89c5bbb61b..1b399cafb15a 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1419,16 +1419,6 @@ struct ext4_super_block { #ifdef __KERNEL__ -/* - * run-time mount flags - */ -#define EXT4_MF_MNTDIR_SAMPLED 0x0001 -#define EXT4_MF_FS_ABORTED 0x0002 /* Fatal error detected */ -#define EXT4_MF_FC_INELIGIBLE 0x0004 /* Fast commit ineligible */ -#define EXT4_MF_FC_COMMITTING 0x0008 /* File system underoing a fast - * commit. - */ - #ifdef CONFIG_FS_ENCRYPTION #define DUMMY_ENCRYPTION_ENABLED(sbi) ((sbi)->s_dummy_enc_policy.policy != NULL) #else @@ -1463,7 +1453,7 @@ struct ext4_sb_info { struct buffer_head * __rcu *s_group_desc; unsigned int s_mount_opt; unsigned int s_mount_opt2; - unsigned int s_mount_flags; + unsigned long s_mount_flags; unsigned int s_def_mount_opt; ext4_fsblk_t s_sb_block; atomic64_t s_resv_clusters; @@ -1691,6 +1681,34 @@ static inline int ext4_valid_inum(struct super_block *sb, unsigned long ino) _v; \ }) +/* + * run-time mount flags + */ +enum { + EXT4_MF_MNTDIR_SAMPLED, + EXT4_MF_FS_ABORTED, /* Fatal error detected */ + EXT4_MF_FC_INELIGIBLE, /* Fast commit ineligible */ + EXT4_MF_FC_COMMITTING /* File system underoing a fast + * commit. + */ +}; + +static inline void ext4_set_mount_flag(struct super_block *sb, int bit) +{ + set_bit(bit, &EXT4_SB(sb)->s_mount_flags); +} + +static inline void ext4_clear_mount_flag(struct super_block *sb, int bit) +{ + clear_bit(bit, &EXT4_SB(sb)->s_mount_flags); +} + +static inline int ext4_test_mount_flag(struct super_block *sb, int bit) +{ + return test_bit(bit, &EXT4_SB(sb)->s_mount_flags); +} + + /* * Simulate_fail codes */ diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c index ebe5f423f8f2..5cd6630ab1b9 100644 --- a/fs/ext4/fast_commit.c +++ b/fs/ext4/fast_commit.c @@ -261,7 +261,7 @@ void ext4_fc_mark_ineligible(struct super_block *sb, int reason) (EXT4_SB(sb)->s_mount_state & EXT4_FC_REPLAY)) return; - sbi->s_mount_flags |= EXT4_MF_FC_INELIGIBLE; + ext4_set_mount_flag(sb, EXT4_MF_FC_INELIGIBLE); WARN_ON(reason >= EXT4_FC_REASON_MAX); sbi->s_fc_stats.fc_ineligible_reason_count[reason]++; } @@ -294,14 +294,14 @@ void ext4_fc_stop_ineligible(struct super_block *sb) (EXT4_SB(sb)->s_mount_state & EXT4_FC_REPLAY)) return; - EXT4_SB(sb)->s_mount_flags |= EXT4_MF_FC_INELIGIBLE; + ext4_set_mount_flag(sb, EXT4_MF_FC_INELIGIBLE); atomic_dec(&EXT4_SB(sb)->s_fc_ineligible_updates); } static inline int ext4_fc_is_ineligible(struct super_block *sb) { - return (EXT4_SB(sb)->s_mount_flags & EXT4_MF_FC_INELIGIBLE) || - atomic_read(&EXT4_SB(sb)->s_fc_ineligible_updates); + return (ext4_test_mount_flag(sb, EXT4_MF_FC_INELIGIBLE) || + atomic_read(&EXT4_SB(sb)->s_fc_ineligible_updates)); } /* @@ -349,7 +349,7 @@ static int ext4_fc_track_template( spin_lock(&sbi->s_fc_lock); if (list_empty(&EXT4_I(inode)->i_fc_list)) list_add_tail(&EXT4_I(inode)->i_fc_list, - (sbi->s_mount_flags & EXT4_MF_FC_COMMITTING) ? + (ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_COMMITTING)) ? &sbi->s_fc_q[FC_Q_STAGING] : &sbi->s_fc_q[FC_Q_MAIN]); spin_unlock(&sbi->s_fc_lock); @@ -402,7 +402,7 @@ static int __track_dentry_update(struct inode *inode, void *arg, bool update) node->fcd_name.len = dentry->d_name.len; spin_lock(&sbi->s_fc_lock); - if (sbi->s_mount_flags & EXT4_MF_FC_COMMITTING) + if (ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_COMMITTING)) list_add_tail(&node->fcd_list, &sbi->s_fc_dentry_q[FC_Q_STAGING]); else @@ -857,7 +857,7 @@ static int ext4_fc_submit_inode_data_all(journal_t *journal) int ret = 0; spin_lock(&sbi->s_fc_lock); - sbi->s_mount_flags |= EXT4_MF_FC_COMMITTING; + ext4_set_mount_flag(sb, EXT4_MF_FC_COMMITTING); list_for_each(pos, &sbi->s_fc_q[FC_Q_MAIN]) { ei = list_entry(pos, struct ext4_inode_info, i_fc_list); ext4_set_inode_state(&ei->vfs_inode, EXT4_STATE_FC_COMMITTING); @@ -1206,8 +1206,8 @@ static void ext4_fc_cleanup(journal_t *journal, int full) list_splice_init(&sbi->s_fc_q[FC_Q_STAGING], &sbi->s_fc_q[FC_Q_STAGING]); - sbi->s_mount_flags &= ~EXT4_MF_FC_COMMITTING; - sbi->s_mount_flags &= ~EXT4_MF_FC_INELIGIBLE; + ext4_clear_mount_flag(sb, EXT4_MF_FC_COMMITTING); + ext4_clear_mount_flag(sb, EXT4_MF_FC_INELIGIBLE); if (full) sbi->s_fc_bytes = 0; diff --git a/fs/ext4/file.c b/fs/ext4/file.c index 80ad5ccc0288..3ed8c048fb12 100644 --- a/fs/ext4/file.c +++ b/fs/ext4/file.c @@ -780,13 +780,13 @@ static int ext4_sample_last_mounted(struct super_block *sb, handle_t *handle; int err; - if (likely(sbi->s_mount_flags & EXT4_MF_MNTDIR_SAMPLED)) + if (likely(ext4_test_mount_flag(sb, EXT4_MF_MNTDIR_SAMPLED))) return 0; if (sb_rdonly(sb) || !sb_start_intwrite_trylock(sb)) return 0; - sbi->s_mount_flags |= EXT4_MF_MNTDIR_SAMPLED; + ext4_set_mount_flag(sb, EXT4_MF_MNTDIR_SAMPLED); /* * Sample where the filesystem has been mounted and * store it in the superblock for sysadmin convenience diff --git a/fs/ext4/fsync.c b/fs/ext4/fsync.c index 81a545fd14a3..a42ca95840f2 100644 --- a/fs/ext4/fsync.c +++ b/fs/ext4/fsync.c @@ -143,7 +143,7 @@ int ext4_sync_file(struct file *file, loff_t start, loff_t end, int datasync) if (sb_rdonly(inode->i_sb)) { /* Make sure that we read updated s_mount_flags value */ smp_rmb(); - if (sbi->s_mount_flags & EXT4_MF_FS_ABORTED) + if (ext4_test_mount_flag(inode->i_sb, EXT4_MF_FS_ABORTED)) ret = -EROFS; goto out; } diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 000bf70e88ed..0d8385aea898 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -2442,7 +2442,7 @@ static int mpage_map_and_submit_extent(handle_t *handle, struct super_block *sb = inode->i_sb; if (ext4_forced_shutdown(EXT4_SB(sb)) || - EXT4_SB(sb)->s_mount_flags & EXT4_MF_FS_ABORTED) + ext4_test_mount_flag(sb, EXT4_MF_FS_ABORTED)) goto invalidate_dirty_pages; /* * Let the uper layers retry transient errors. @@ -2676,7 +2676,7 @@ static int ext4_writepages(struct address_space *mapping, * the stack trace. */ if (unlikely(ext4_forced_shutdown(EXT4_SB(mapping->host->i_sb)) || - sbi->s_mount_flags & EXT4_MF_FS_ABORTED)) { + ext4_test_mount_flag(inode->i_sb, EXT4_MF_FS_ABORTED))) { ret = -EROFS; goto out_writepages; } diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index f067e83f149e..24af9ed5c3e5 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -4477,7 +4477,7 @@ static inline void ext4_mb_show_pa(struct super_block *sb) { ext4_group_t i, ngroups; - if (EXT4_SB(sb)->s_mount_flags & EXT4_MF_FS_ABORTED) + if (ext4_test_mount_flag(sb, EXT4_MF_FS_ABORTED)) return; ngroups = ext4_get_groups_count(sb); @@ -4508,7 +4508,7 @@ static void ext4_mb_show_ac(struct ext4_allocation_context *ac) { struct super_block *sb = ac->ac_sb; - if (EXT4_SB(sb)->s_mount_flags & EXT4_MF_FS_ABORTED) + if (ext4_test_mount_flag(sb, EXT4_MF_FS_ABORTED)) return; mb_debug(sb, "Can't allocate:" diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 9e2e06ed2c45..7bd02c252f7f 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -686,7 +686,7 @@ static void ext4_handle_error(struct super_block *sb) if (!test_opt(sb, ERRORS_CONT)) { journal_t *journal = EXT4_SB(sb)->s_journal; - EXT4_SB(sb)->s_mount_flags |= EXT4_MF_FS_ABORTED; + ext4_set_mount_flag(sb, EXT4_MF_FS_ABORTED); if (journal) jbd2_journal_abort(journal, -EIO); } @@ -904,7 +904,7 @@ void __ext4_abort(struct super_block *sb, const char *function, va_end(args); if (sb_rdonly(sb) == 0) { - EXT4_SB(sb)->s_mount_flags |= EXT4_MF_FS_ABORTED; + ext4_set_mount_flag(sb, EXT4_MF_FS_ABORTED); if (EXT4_SB(sb)->s_journal) jbd2_journal_abort(EXT4_SB(sb)->s_journal, -EIO); @@ -2153,7 +2153,7 @@ static int handle_mount_opt(struct super_block *sb, char *opt, int token, ext4_msg(sb, KERN_WARNING, "Ignoring removed %s option", opt); return 1; case Opt_abort: - sbi->s_mount_flags |= EXT4_MF_FS_ABORTED; + ext4_set_mount_flag(sb, EXT4_MF_FS_ABORTED); return 1; case Opt_i_version: sb->s_flags |= SB_I_VERSION; @@ -4778,8 +4778,8 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) INIT_LIST_HEAD(&sbi->s_fc_dentry_q[FC_Q_MAIN]); INIT_LIST_HEAD(&sbi->s_fc_dentry_q[FC_Q_STAGING]); sbi->s_fc_bytes = 0; - sbi->s_mount_flags &= ~EXT4_MF_FC_INELIGIBLE; - sbi->s_mount_flags &= ~EXT4_MF_FC_COMMITTING; + ext4_clear_mount_flag(sb, EXT4_MF_FC_INELIGIBLE); + ext4_clear_mount_flag(sb, EXT4_MF_FC_COMMITTING); spin_lock_init(&sbi->s_fc_lock); memset(&sbi->s_fc_stats, 0, sizeof(sbi->s_fc_stats)); sbi->s_fc_replay_state.fc_regions = NULL; @@ -5881,7 +5881,7 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data) goto restore_opts; } - if (sbi->s_mount_flags & EXT4_MF_FS_ABORTED) + if (ext4_test_mount_flag(sb, EXT4_MF_FS_ABORTED)) ext4_abort(sb, EXT4_ERR_ESHUTDOWN, "Abort forced by user"); sb->s_flags = (sb->s_flags & ~SB_POSIXACL) | @@ -5895,7 +5895,7 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data) } if ((bool)(*flags & SB_RDONLY) != sb_rdonly(sb)) { - if (sbi->s_mount_flags & EXT4_MF_FS_ABORTED) { + if (ext4_test_mount_flag(sb, EXT4_MF_FS_ABORTED)) { err = -EROFS; goto restore_opts; } -- cgit v1.2.3 From fa329e27317f7f0762001b9fb1e76c387a9db25d Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Fri, 6 Nov 2020 23:59:42 -0500 Subject: ext4: fix sparse warnings in fast_commit code Add missing __acquire() and __releases() annotations, and make fc_ineligible_reasons[] static, as it is not used outside of fs/ext4/fast_commit.c. Signed-off-by: Theodore Ts'o --- fs/ext4/fast_commit.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'fs/ext4/fast_commit.c') diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c index 5cd6630ab1b9..f2033e13a273 100644 --- a/fs/ext4/fast_commit.c +++ b/fs/ext4/fast_commit.c @@ -156,6 +156,7 @@ void ext4_fc_init_inode(struct inode *inode) /* This function must be called with sbi->s_fc_lock held. */ static void ext4_fc_wait_committing_inode(struct inode *inode) +__releases(&EXT4_SB(inode->i_sb)->s_fc_lock) { wait_queue_head_t *wq; struct ext4_inode_info *ei = EXT4_I(inode); @@ -911,6 +912,8 @@ static int ext4_fc_wait_inode_data_all(journal_t *journal) /* Commit all the directory entry updates */ static int ext4_fc_commit_dentry_updates(journal_t *journal, u32 *crc) +__acquires(&sbi->s_fc_lock) +__releases(&sbi->s_fc_lock) { struct super_block *sb = (struct super_block *)(journal->j_private); struct ext4_sb_info *sbi = EXT4_SB(sb); @@ -2106,7 +2109,7 @@ void ext4_fc_init(struct super_block *sb, journal_t *journal) journal->j_fc_cleanup_callback = ext4_fc_cleanup; } -const char *fc_ineligible_reasons[] = { +static const char *fc_ineligible_reasons[] = { "Extended attributes changed", "Cross rename", "Journal flag changed", -- cgit v1.2.3