This is the V6 of the patch series. This patch series contains fixes to a
bunch of kernel build warnings reported by Kernel Test Robot
([email protected]) and Dan Carpenter ([email protected]). Thank you!
V5 Cover Letter
---------------
This patch series supersedes the patch series "ext4: remove journal barrier
during fast commit" sent in July 2022[1]. Except following three patches
all the other patches are newly introduced in this series.
* ext4: convert i_fc_lock to spinlock
* ext4: for committing inode, make ext4_fc_track_inode wait
* ext4: drop i_fc_updates from inode fc info
This patchset does not introduce any on-disk format and / or user visible
API change. This patchset reworks fast-commit's commit path improve overall
performance of the commit path. Following optimizations have been added in
this series:
* Avoid having to lock the journal throughout the fast commit.
* Remove tracking of open handles per inode.
* Avoid issuing cache flushes when fast commits are contained within a
single FUA writes and there is no data that needs flushing.
* Temporarily lift committing thread's priority to match that of the
journal thread to reduce scheduling delays.
With the changes introduced in this patch series, now the commit path for
fast commits is as follows:
1. Lock the journal by calling jbd2_journal_lock_updates_no_rsv(). This
ensures that all the exsiting handles finish and no new handles can
start.
2. Mark all the fast commit eligible inodes as undergoing fast commit by
setting "EXT4_STATE_FC_COMMITTING" state.
3. Unlock the journal by calling jbd2_journal_unlock_updates. This allows
starting of new handles. If new handles try to start an update on any of
the inodes that are being committed, ext4_fc_track_inode() will block
until those inodes have finished the fast commit.
4. Submit data buffers of all the committing inodes.
5. Wait for [4] to complete.
6. Commit all the directory entry updates in the fast commit space.
7. Commit all the changed inodes in the fast commit space and clear
"EXT4_STATE_FC_COMMITTING" for all the inodes.
8. Write tail tag to ensure atomicity of commits.
(The above flow has been documented in the code as well)
I verified that the patch series introduces no regressions in "log" groups
when "fast_commit" feature is enabled.
Also, we have a paper on fast commits in USENIX ATC 2024 this year which
should become available on the website[2] in a few months.
[1] https://lwn.net/Articles/902022/
[2] https://www.usenix.org/conference/atc24
Harshad Shirwadkar (10):
ext4: convert i_fc_lock to spinlock
ext4: for committing inode, make ext4_fc_track_inode wait
ext4: mark inode dirty before grabbing i_data_sem in ext4_setattr
ext4: rework fast commit commit path
ext4: drop i_fc_updates from inode fc info
ext4: update code documentation
ext4: add nolock mode to ext4_map_blocks()
ext4: introduce selective flushing in fast commit
ext4: temporarily elevate commit thread priority
ext4: make fast commit ineligible on ext4_reserve_inode_write failure
fs/ext4/ext4.h | 29 ++--
fs/ext4/ext4_jbd2.h | 20 +--
fs/ext4/fast_commit.c | 279 ++++++++++++++++++++----------------
fs/ext4/fast_commit.h | 1 +
fs/ext4/inline.c | 3 +
fs/ext4/inode.c | 52 +++++--
fs/ext4/super.c | 7 +-
fs/jbd2/journal.c | 3 +-
fs/jbd2/transaction.c | 41 ++++--
include/linux/jbd2.h | 1 +
include/trace/events/ext4.h | 7 +-
11 files changed, 266 insertions(+), 177 deletions(-)
--
2.45.1.288.g0e0cd299f1-goog
Convert ext4_inode_info->i_fc_lock to spinlock to avoid sleeping
in invalid contexts.
Signed-off-by: Harshad Shirwadkar <[email protected]>
---
fs/ext4/ext4.h | 7 +++++--
fs/ext4/fast_commit.c | 24 +++++++++++-------------
fs/ext4/super.c | 2 +-
3 files changed, 17 insertions(+), 16 deletions(-)
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 983dad8c07ec..611b8c80d99c 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1062,8 +1062,11 @@ struct ext4_inode_info {
/* Fast commit wait queue for this inode */
wait_queue_head_t i_fc_wait;
- /* Protect concurrent accesses on i_fc_lblk_start, i_fc_lblk_len */
- struct mutex i_fc_lock;
+ /*
+ * Protect concurrent accesses on i_fc_lblk_start, i_fc_lblk_len
+ * and inode's EXT4_FC_STATE_COMMITTING state bit.
+ */
+ spinlock_t i_fc_lock;
/*
* i_disksize keeps track of what the inode size is ON DISK, not
diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index 87c009e0c59a..a1aadebfcd66 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -382,7 +382,7 @@ static int ext4_fc_track_template(
int ret;
tid = handle->h_transaction->t_tid;
- mutex_lock(&ei->i_fc_lock);
+ spin_lock(&ei->i_fc_lock);
if (tid == ei->i_sync_tid) {
update = true;
} else {
@@ -390,7 +390,7 @@ static int ext4_fc_track_template(
ei->i_sync_tid = tid;
}
ret = __fc_track_fn(inode, args, update);
- mutex_unlock(&ei->i_fc_lock);
+ spin_unlock(&ei->i_fc_lock);
if (!enqueue)
return ret;
@@ -424,19 +424,19 @@ static int __track_dentry_update(struct inode *inode, void *arg, bool update)
struct super_block *sb = inode->i_sb;
struct ext4_sb_info *sbi = EXT4_SB(sb);
- mutex_unlock(&ei->i_fc_lock);
+ spin_unlock(&ei->i_fc_lock);
if (IS_ENCRYPTED(dir)) {
ext4_fc_mark_ineligible(sb, EXT4_FC_REASON_ENCRYPTED_FILENAME,
NULL);
- mutex_lock(&ei->i_fc_lock);
+ spin_lock(&ei->i_fc_lock);
return -EOPNOTSUPP;
}
node = kmem_cache_alloc(ext4_fc_dentry_cachep, GFP_NOFS);
if (!node) {
ext4_fc_mark_ineligible(sb, EXT4_FC_REASON_NOMEM, NULL);
- mutex_lock(&ei->i_fc_lock);
+ spin_lock(&ei->i_fc_lock);
return -ENOMEM;
}
@@ -448,7 +448,7 @@ 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(sb, EXT4_FC_REASON_NOMEM, NULL);
- mutex_lock(&ei->i_fc_lock);
+ spin_lock(&ei->i_fc_lock);
return -ENOMEM;
}
memcpy((u8 *)node->fcd_name.name, dentry->d_name.name,
@@ -482,7 +482,7 @@ static int __track_dentry_update(struct inode *inode, void *arg, bool update)
list_add_tail(&node->fcd_dilist, &ei->i_fc_dilist);
}
spin_unlock(&sbi->s_fc_lock);
- mutex_lock(&ei->i_fc_lock);
+ spin_lock(&ei->i_fc_lock);
return 0;
}
@@ -614,10 +614,8 @@ static int __track_range(struct inode *inode, void *arg, bool update)
struct __track_range_args *__arg =
(struct __track_range_args *)arg;
- if (inode->i_ino < EXT4_FIRST_INO(inode->i_sb)) {
- ext4_debug("Special inode %ld being modified\n", inode->i_ino);
+ if (inode->i_ino < EXT4_FIRST_INO(inode->i_sb))
return -ECANCELED;
- }
oldstart = ei->i_fc_lblk_start;
@@ -896,15 +894,15 @@ static int ext4_fc_write_inode_data(struct inode *inode, u32 *crc)
struct ext4_extent *ex;
int ret;
- mutex_lock(&ei->i_fc_lock);
+ spin_lock(&ei->i_fc_lock);
if (ei->i_fc_lblk_len == 0) {
- mutex_unlock(&ei->i_fc_lock);
+ spin_unlock(&ei->i_fc_lock);
return 0;
}
old_blk_size = ei->i_fc_lblk_start;
new_blk_size = ei->i_fc_lblk_start + ei->i_fc_lblk_len - 1;
ei->i_fc_lblk_len = 0;
- mutex_unlock(&ei->i_fc_lock);
+ spin_unlock(&ei->i_fc_lock);
cur_lblk_off = old_blk_size;
ext4_debug("will try writing %d to %d for inode %ld\n",
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index f9a4a4e89dac..77173ec91e49 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -1436,7 +1436,7 @@ static struct inode *ext4_alloc_inode(struct super_block *sb)
atomic_set(&ei->i_unwritten, 0);
INIT_WORK(&ei->i_rsv_conversion_work, ext4_end_io_rsv_work);
ext4_fc_init_inode(&ei->vfs_inode);
- mutex_init(&ei->i_fc_lock);
+ spin_lock_init(&ei->i_fc_lock);
return &ei->vfs_inode;
}
--
2.45.1.288.g0e0cd299f1-goog
This patch reworks fast commit's commit path to remove locking the
journal for the entire duration of a fast commit. Instead, we only lock
the journal while marking all the eligible inodes as "committing". This
allows handles to make progress in parallel with the fast commit.
Signed-off-by: Harshad Shirwadkar <[email protected]>
---
fs/ext4/fast_commit.c | 75 ++++++++++++++++++++++++++++---------------
fs/jbd2/journal.c | 3 +-
fs/jbd2/transaction.c | 41 +++++++++++++++--------
include/linux/jbd2.h | 1 +
4 files changed, 80 insertions(+), 40 deletions(-)
diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index fa93ce399440..3aca5b20aac5 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -290,20 +290,30 @@ void ext4_fc_del(struct inode *inode)
if (ext4_fc_disabled(inode->i_sb))
return;
-restart:
spin_lock(&EXT4_SB(inode->i_sb)->s_fc_lock);
if (list_empty(&ei->i_fc_list) && list_empty(&ei->i_fc_dilist)) {
spin_unlock(&EXT4_SB(inode->i_sb)->s_fc_lock);
return;
}
- if (ext4_test_inode_state(inode, EXT4_STATE_FC_COMMITTING)) {
- ext4_fc_wait_committing_inode(inode);
- goto restart;
- }
-
- if (!list_empty(&ei->i_fc_list))
- list_del_init(&ei->i_fc_list);
+ /*
+ * Since ext4_fc_del is called from ext4_evict_inode while having a
+ * handle open, there is no need for us to wait here even if a fast
+ * commit is going on. That is because, if this inode is being
+ * committed, ext4_mark_inode_dirty would have waited for inode commit
+ * operation to finish before we come here. So, by the time we come
+ * here, inode's EXT4_STATE_FC_COMMITTING would have been cleared. So,
+ * we shouldn't see EXT4_STATE_FC_COMMITTING to be set on this inode
+ * here.
+ *
+ * We may come here without any handles open in the "no_delete" case of
+ * ext4_evict_inode as well. However, if that happens, we first mark the
+ * file system as fast commit ineligible anyway. So, even in that case,
+ * it is okay to remove the inode from the fc list.
+ */
+ WARN_ON(ext4_test_inode_state(inode, EXT4_STATE_FC_COMMITTING)
+ && !ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_INELIGIBLE));
+ list_del_init(&ei->i_fc_list);
/*
* Since this inode is getting removed, let's also remove all FC
@@ -326,8 +336,6 @@ void ext4_fc_del(struct inode *inode)
fc_dentry->fcd_name.len > DNAME_INLINE_LEN)
kfree(fc_dentry->fcd_name.name);
kmem_cache_free(ext4_fc_dentry_cachep, fc_dentry);
-
- return;
}
/*
@@ -999,19 +1007,6 @@ static int ext4_fc_submit_inode_data_all(journal_t *journal)
spin_lock(&sbi->s_fc_lock);
list_for_each_entry(ei, &sbi->s_fc_q[FC_Q_MAIN], i_fc_list) {
- ext4_set_inode_state(&ei->vfs_inode, EXT4_STATE_FC_COMMITTING);
- while (atomic_read(&ei->i_fc_updates)) {
- DEFINE_WAIT(wait);
-
- prepare_to_wait(&ei->i_fc_wait, &wait,
- TASK_UNINTERRUPTIBLE);
- if (atomic_read(&ei->i_fc_updates)) {
- spin_unlock(&sbi->s_fc_lock);
- schedule();
- spin_lock(&sbi->s_fc_lock);
- }
- finish_wait(&ei->i_fc_wait, &wait);
- }
spin_unlock(&sbi->s_fc_lock);
ret = jbd2_submit_inode_data(journal, ei->jinode);
if (ret)
@@ -1124,6 +1119,20 @@ static int ext4_fc_perform_commit(journal_t *journal)
int ret = 0;
u32 crc = 0;
+ /*
+ * Wait for all the handles of the current transaction to complete
+ * and then lock the journal. Since this is essentially the commit
+ * path, we don't need to wait for reserved handles.
+ */
+ jbd2_journal_lock_updates_no_rsv(journal);
+ spin_lock(&sbi->s_fc_lock);
+ list_for_each_entry(iter, &sbi->s_fc_q[FC_Q_MAIN], i_fc_list) {
+ ext4_set_inode_state(&iter->vfs_inode,
+ EXT4_STATE_FC_COMMITTING);
+ }
+ spin_unlock(&sbi->s_fc_lock);
+ jbd2_journal_unlock_updates(journal);
+
ret = ext4_fc_submit_inode_data_all(journal);
if (ret)
return ret;
@@ -1174,6 +1183,18 @@ static int ext4_fc_perform_commit(journal_t *journal)
ret = ext4_fc_write_inode(inode, &crc);
if (ret)
goto out;
+ ext4_clear_inode_state(inode, EXT4_STATE_FC_COMMITTING);
+ /*
+ * Make sure clearing of EXT4_STATE_FC_COMMITTING is
+ * visible before we send the wakeup. Pairs with implicit
+ * barrier in prepare_to_wait() in ext4_fc_track_inode().
+ */
+ smp_mb();
+#if (BITS_PER_LONG < 64)
+ wake_up_bit(&iter->i_state_flags, EXT4_STATE_FC_COMMITTING);
+#else
+ wake_up_bit(&iter->i_flags, EXT4_STATE_FC_COMMITTING);
+#endif
spin_lock(&sbi->s_fc_lock);
}
spin_unlock(&sbi->s_fc_lock);
@@ -1311,13 +1332,17 @@ static void ext4_fc_cleanup(journal_t *journal, int full, tid_t tid)
spin_lock(&sbi->s_fc_lock);
list_for_each_entry_safe(iter, iter_n, &sbi->s_fc_q[FC_Q_MAIN],
i_fc_list) {
- list_del_init(&iter->i_fc_list);
ext4_clear_inode_state(&iter->vfs_inode,
EXT4_STATE_FC_COMMITTING);
if (iter->i_sync_tid <= tid)
ext4_fc_reset_inode(&iter->vfs_inode);
- /* Make sure EXT4_STATE_FC_COMMITTING bit is clear */
+ /*
+ * Make sure clearing of EXT4_STATE_FC_COMMITTING is
+ * visible before we send the wakeup. Pairs with implicit
+ * barrier in prepare_to_wait() in ext4_fc_track_inode().
+ */
smp_mb();
+ list_del_init(&iter->i_fc_list);
#if (BITS_PER_LONG < 64)
wake_up_bit(&iter->i_state_flags, EXT4_STATE_FC_COMMITTING);
#else
diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index b6c114c11b97..7df1720f013b 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -58,6 +58,7 @@ MODULE_PARM_DESC(jbd2_debug, "Debugging level for jbd2");
EXPORT_SYMBOL(jbd2_journal_extend);
EXPORT_SYMBOL(jbd2_journal_stop);
EXPORT_SYMBOL(jbd2_journal_lock_updates);
+EXPORT_SYMBOL(jbd2_journal_lock_updates_no_rsv);
EXPORT_SYMBOL(jbd2_journal_unlock_updates);
EXPORT_SYMBOL(jbd2_journal_get_write_access);
EXPORT_SYMBOL(jbd2_journal_get_create_access);
@@ -742,7 +743,6 @@ int jbd2_fc_begin_commit(journal_t *journal, tid_t tid)
}
journal->j_flags |= JBD2_FAST_COMMIT_ONGOING;
write_unlock(&journal->j_state_lock);
- jbd2_journal_lock_updates(journal);
return 0;
}
@@ -754,7 +754,6 @@ EXPORT_SYMBOL(jbd2_fc_begin_commit);
*/
static int __jbd2_fc_end_commit(journal_t *journal, tid_t tid, bool fallback)
{
- jbd2_journal_unlock_updates(journal);
if (journal->j_fc_cleanup_callback)
journal->j_fc_cleanup_callback(journal, 0, tid);
write_lock(&journal->j_state_lock);
diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c
index cb0b8d6fc0c6..4361e5c56490 100644
--- a/fs/jbd2/transaction.c
+++ b/fs/jbd2/transaction.c
@@ -865,25 +865,15 @@ void jbd2_journal_wait_updates(journal_t *journal)
}
}
-/**
- * jbd2_journal_lock_updates () - establish a transaction barrier.
- * @journal: Journal to establish a barrier on.
- *
- * This locks out any further updates from being started, and blocks
- * until all existing updates have completed, returning only once the
- * journal is in a quiescent state with no updates running.
- *
- * The journal lock should not be held on entry.
- */
-void jbd2_journal_lock_updates(journal_t *journal)
+static void __jbd2_journal_lock_updates(journal_t *journal, bool wait_on_rsv)
{
jbd2_might_wait_for_commit(journal);
write_lock(&journal->j_state_lock);
++journal->j_barrier_count;
- /* Wait until there are no reserved handles */
- if (atomic_read(&journal->j_reserved_credits)) {
+ if (wait_on_rsv && atomic_read(&journal->j_reserved_credits)) {
+ /* Wait until there are no reserved handles */
write_unlock(&journal->j_state_lock);
wait_event(journal->j_wait_reserved,
atomic_read(&journal->j_reserved_credits) == 0);
@@ -904,6 +894,31 @@ void jbd2_journal_lock_updates(journal_t *journal)
mutex_lock(&journal->j_barrier);
}
+/**
+ * jbd2_journal_lock_updates () - establish a transaction barrier.
+ * @journal: Journal to establish a barrier on.
+ *
+ * This locks out any further updates from being started, and blocks
+ * until all existing updates have completed, returning only once the
+ * journal is in a quiescent state with no updates running.
+ *
+ * The journal lock should not be held on entry.
+ */
+void jbd2_journal_lock_updates(journal_t *journal)
+{
+ __jbd2_journal_lock_updates(journal, true);
+}
+
+/**
+ * jbd2_journal_lock_updates_no_rsv - same as above, but doesn't
+ * wait for reserved handles to finish.
+ * @journal: Journal to establish barrier on.
+ */
+void jbd2_journal_lock_updates_no_rsv(journal_t *journal)
+{
+ __jbd2_journal_lock_updates(journal, false);
+}
+
/**
* jbd2_journal_unlock_updates () - release barrier
* @journal: Journal to release the barrier on.
diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h
index 7479f64c0939..cfac287c0ad4 100644
--- a/include/linux/jbd2.h
+++ b/include/linux/jbd2.h
@@ -1531,6 +1531,7 @@ bool jbd2_journal_try_to_free_buffers(journal_t *journal, struct folio *folio);
extern int jbd2_journal_stop(handle_t *);
extern int jbd2_journal_flush(journal_t *journal, unsigned int flags);
extern void jbd2_journal_lock_updates (journal_t *);
+extern void jbd2_journal_lock_updates_no_rsv(journal_t *journal);
extern void jbd2_journal_unlock_updates (journal_t *);
void jbd2_journal_wait_updates(journal_t *);
--
2.45.1.288.g0e0cd299f1-goog
The new logic introduced in this series does not require tracking number
of active handles open on an inode. So, drop it.
Signed-off-by: Harshad Shirwadkar <[email protected]>
Reviewed-by: Jan Kara <[email protected]>
---
fs/ext4/ext4.h | 5 ----
fs/ext4/fast_commit.c | 68 -------------------------------------------
2 files changed, 73 deletions(-)
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 611b8c80d99c..d802040e94df 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1056,9 +1056,6 @@ struct ext4_inode_info {
/* End of lblk range that needs to be committed in this fast commit */
ext4_lblk_t i_fc_lblk_len;
- /* Number of ongoing updates on this inode */
- atomic_t i_fc_updates;
-
/* Fast commit wait queue for this inode */
wait_queue_head_t i_fc_wait;
@@ -2903,8 +2900,6 @@ void __ext4_fc_track_create(handle_t *handle, struct inode *inode,
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, handle_t *handle);
-void ext4_fc_start_update(struct inode *inode);
-void ext4_fc_stop_update(struct inode *inode);
void ext4_fc_del(struct inode *inode);
bool ext4_fc_replay_check_excluded(struct super_block *sb, ext4_fsblk_t block);
void ext4_fc_replay_cleanup(struct super_block *sb);
diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index 3aca5b20aac5..ecbbcaf78598 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -201,32 +201,6 @@ void ext4_fc_init_inode(struct inode *inode)
INIT_LIST_HEAD(&ei->i_fc_list);
INIT_LIST_HEAD(&ei->i_fc_dilist);
init_waitqueue_head(&ei->i_fc_wait);
- atomic_set(&ei->i_fc_updates, 0);
-}
-
-/* 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);
-
-#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);
}
static bool ext4_fc_disabled(struct super_block *sb)
@@ -235,48 +209,6 @@ static bool ext4_fc_disabled(struct super_block *sb)
(EXT4_SB(sb)->s_mount_state & EXT4_FC_REPLAY));
}
-/*
- * Inform Ext4's fast about start of an inode update
- *
- * This function is called by the high level call VFS callbacks before
- * performing any inode update. This function blocks if there's an ongoing
- * fast commit on the inode in question.
- */
-void ext4_fc_start_update(struct inode *inode)
-{
- struct ext4_inode_info *ei = EXT4_I(inode);
-
- if (ext4_fc_disabled(inode->i_sb))
- return;
-
-restart:
- spin_lock(&EXT4_SB(inode->i_sb)->s_fc_lock);
- if (list_empty(&ei->i_fc_list))
- goto out;
-
- if (ext4_test_inode_state(inode, EXT4_STATE_FC_COMMITTING)) {
- ext4_fc_wait_committing_inode(inode);
- goto restart;
- }
-out:
- atomic_inc(&ei->i_fc_updates);
- spin_unlock(&EXT4_SB(inode->i_sb)->s_fc_lock);
-}
-
-/*
- * Stop inode update and wake up waiting fast commits if any.
- */
-void ext4_fc_stop_update(struct inode *inode)
-{
- struct ext4_inode_info *ei = EXT4_I(inode);
-
- if (ext4_fc_disabled(inode->i_sb))
- return;
-
- if (atomic_dec_and_test(&ei->i_fc_updates))
- wake_up_all(&ei->i_fc_wait);
-}
-
/*
* Remove inode from fast commit list. If the inode is being committed
* we wait until inode commit is done.
--
2.45.1.288.g0e0cd299f1-goog
Add nolock flag to ext4_map_blocks() which skips grabbing
i_data_sem in ext4_map_blocks. In FC commit path, we first
mark the inode as committing and thereby prevent any mutations
on it. Thus, it should be safe to call ext4_map_blocks()
without i_data_sem in this case. This is a workaround to
the problem mentioned in RFC V4 version cover letter[1] of this
patch series which pointed out that there is in incosistency between
ext4_map_blocks() behavior when EXT4_GET_BLOCKS_CACHED_NOWAIT is
passed. This patch gets rid of the need to call ext4_map_blocks()
with EXT4_GET_BLOCKS_CACHED_NOWAIT and instead call it with
EXT4_GET_BLOCKS_NOLOCK. I verified that generic/311 which failed
in cached_nowait mode passes with nolock mode.
[1] https://lwn.net/Articles/902022/
Signed-off-by: Harshad Shirwadkar <[email protected]>
---
fs/ext4/ext4.h | 1 +
fs/ext4/fast_commit.c | 16 ++++++++--------
fs/ext4/inode.c | 14 ++++++++++++--
3 files changed, 21 insertions(+), 10 deletions(-)
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index d802040e94df..196c513f82dd 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -720,6 +720,7 @@ enum {
#define EXT4_GET_BLOCKS_IO_SUBMIT 0x0400
/* Caller is in the atomic contex, find extent if it has been cached */
#define EXT4_GET_BLOCKS_CACHED_NOWAIT 0x0800
+#define EXT4_GET_BLOCKS_NOLOCK 0x1000
/*
* The bit position of these flags must not overlap with any of the
diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index b81b0292aa59..0b7064f8dfa5 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -559,13 +559,6 @@ void ext4_fc_track_inode(handle_t *handle, struct inode *inode)
!list_empty(&ei->i_fc_list))
return;
- /*
- * If we come here, we may sleep while waiting for the inode to
- * commit. We shouldn't be holding i_data_sem in write mode when we go
- * to sleep since the commit path needs to grab the lock while
- * committing the inode.
- */
- WARN_ON(lockdep_is_held_type(&ei->i_data_sem, 1));
while (ext4_test_inode_state(inode, EXT4_STATE_FC_COMMITTING)) {
#if (BITS_PER_LONG < 64)
@@ -898,7 +891,14 @@ static int ext4_fc_write_inode_data(struct inode *inode, u32 *crc)
while (cur_lblk_off <= new_blk_size) {
map.m_lblk = cur_lblk_off;
map.m_len = new_blk_size - cur_lblk_off + 1;
- ret = ext4_map_blocks(NULL, inode, &map, 0);
+ /*
+ * Given that this inode is being committed,
+ * EXT4_STATE_FC_COMMITTING is already set on this inode.
+ * Which means all the mutations on the inode are paused
+ * until the commit operation is complete. Thus it is safe
+ * call ext4_map_blocks() in no lock mode.
+ */
+ ret = ext4_map_blocks(NULL, inode, &map, EXT4_GET_BLOCKS_NOLOCK);
if (ret < 0)
return -ECANCELED;
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 61ffbdc2fb16..f00408017c7a 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -546,7 +546,8 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode,
* Try to see if we can get the block without requesting a new
* file system block.
*/
- down_read(&EXT4_I(inode)->i_data_sem);
+ if (!(flags & EXT4_GET_BLOCKS_NOLOCK))
+ down_read(&EXT4_I(inode)->i_data_sem);
if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) {
retval = ext4_ext_map_blocks(handle, inode, map, 0);
} else {
@@ -573,7 +574,15 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode,
ext4_es_insert_extent(inode, map->m_lblk, map->m_len,
map->m_pblk, status);
}
- up_read((&EXT4_I(inode)->i_data_sem));
+ /*
+ * We should never call ext4_map_blocks() in nolock mode outside
+ * of fast commit path.
+ */
+ WARN_ON((flags & EXT4_GET_BLOCKS_NOLOCK) &&
+ !ext4_test_inode_state(inode,
+ EXT4_STATE_FC_COMMITTING));
+ if (!(flags & EXT4_GET_BLOCKS_NOLOCK))
+ up_read((&EXT4_I(inode)->i_data_sem));
found:
if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED) {
@@ -614,6 +623,7 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode,
* the write lock of i_data_sem, and call get_block()
* with create == 1 flag.
*/
+ WARN_ON((flags & EXT4_GET_BLOCKS_NOLOCK) != 0);
down_write(&EXT4_I(inode)->i_data_sem);
/*
--
2.45.1.288.g0e0cd299f1-goog
Fast commit by default makes every inode on which
ext4_reserve_inode_write() is called. Thus, if that function
fails for some reason, make the next fast commit ineligible.
Signed-off-by: Harshad Shirwadkar <[email protected]>
---
fs/ext4/fast_commit.c | 1 +
fs/ext4/fast_commit.h | 1 +
fs/ext4/inode.c | 29 ++++++++++++++++++-----------
include/trace/events/ext4.h | 7 +++++--
4 files changed, 25 insertions(+), 13 deletions(-)
diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index 55a13d3ff681..e7cac190527c 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -2291,6 +2291,7 @@ static const char * const fc_ineligible_reasons[] = {
[EXT4_FC_REASON_FALLOC_RANGE] = "Falloc range op",
[EXT4_FC_REASON_INODE_JOURNAL_DATA] = "Data journalling",
[EXT4_FC_REASON_ENCRYPTED_FILENAME] = "Encrypted filename",
+ [EXT4_FC_REASON_INODE_RSV_WRITE_FAIL] = "Inode reserve write failure"
};
int ext4_fc_info_show(struct seq_file *seq, void *v)
diff --git a/fs/ext4/fast_commit.h b/fs/ext4/fast_commit.h
index 2fadb2c4780c..f7f85c3dd3af 100644
--- a/fs/ext4/fast_commit.h
+++ b/fs/ext4/fast_commit.h
@@ -97,6 +97,7 @@ enum {
EXT4_FC_REASON_FALLOC_RANGE,
EXT4_FC_REASON_INODE_JOURNAL_DATA,
EXT4_FC_REASON_ENCRYPTED_FILENAME,
+ EXT4_FC_REASON_INODE_RSV_WRITE_FAIL,
EXT4_FC_REASON_MAX
};
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index f00408017c7a..8fd6e5637542 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -5729,20 +5729,27 @@ ext4_reserve_inode_write(handle_t *handle, struct inode *inode,
{
int err;
- if (unlikely(ext4_forced_shutdown(inode->i_sb)))
- return -EIO;
+ if (unlikely(ext4_forced_shutdown(inode->i_sb))) {
+ err = -EIO;
+ goto out;
+ }
err = ext4_get_inode_loc(inode, iloc);
- if (!err) {
- BUFFER_TRACE(iloc->bh, "get_write_access");
- err = ext4_journal_get_write_access(handle, inode->i_sb,
- iloc->bh, EXT4_JTR_NONE);
- if (err) {
- brelse(iloc->bh);
- iloc->bh = NULL;
- }
- ext4_fc_track_inode(handle, inode);
+ if (err)
+ goto out;
+
+ BUFFER_TRACE(iloc->bh, "get_write_access");
+ err = ext4_journal_get_write_access(handle, inode->i_sb,
+ iloc->bh, EXT4_JTR_NONE);
+ if (err) {
+ brelse(iloc->bh);
+ iloc->bh = NULL;
}
+ ext4_fc_track_inode(handle, inode);
+out:
+ if (err)
+ ext4_fc_mark_ineligible(inode->i_sb,
+ EXT4_FC_REASON_INODE_RSV_WRITE_FAIL, handle);
ext4_std_error(inode->i_sb, err);
return err;
}
diff --git a/include/trace/events/ext4.h b/include/trace/events/ext4.h
index a697f4b77162..597845d5c1e8 100644
--- a/include/trace/events/ext4.h
+++ b/include/trace/events/ext4.h
@@ -105,6 +105,7 @@ TRACE_DEFINE_ENUM(EXT4_FC_REASON_RENAME_DIR);
TRACE_DEFINE_ENUM(EXT4_FC_REASON_FALLOC_RANGE);
TRACE_DEFINE_ENUM(EXT4_FC_REASON_INODE_JOURNAL_DATA);
TRACE_DEFINE_ENUM(EXT4_FC_REASON_ENCRYPTED_FILENAME);
+TRACE_DEFINE_ENUM(EXT4_FC_REASON_INODE_RSV_WRITE_FAIL);
TRACE_DEFINE_ENUM(EXT4_FC_REASON_MAX);
#define show_fc_reason(reason) \
@@ -118,7 +119,8 @@ TRACE_DEFINE_ENUM(EXT4_FC_REASON_MAX);
{ EXT4_FC_REASON_RENAME_DIR, "RENAME_DIR"}, \
{ EXT4_FC_REASON_FALLOC_RANGE, "FALLOC_RANGE"}, \
{ EXT4_FC_REASON_INODE_JOURNAL_DATA, "INODE_JOURNAL_DATA"}, \
- { EXT4_FC_REASON_ENCRYPTED_FILENAME, "ENCRYPTED_FILENAME"})
+ { EXT4_FC_REASON_ENCRYPTED_FILENAME, "ENCRYPTED_FILENAME"}, \
+ { EXT4_FC_REASON_INODE_RSV_WRITE_FAIL, "INODE_RSV_WRITE_FAIL"})
TRACE_DEFINE_ENUM(CR_POWER2_ALIGNED);
TRACE_DEFINE_ENUM(CR_GOAL_LEN_FAST);
@@ -2805,7 +2807,7 @@ TRACE_EVENT(ext4_fc_stats,
),
TP_printk("dev %d,%d fc ineligible reasons:\n"
- "%s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u"
+ "%s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u"
"num_commits:%lu, ineligible: %lu, numblks: %lu",
MAJOR(__entry->dev), MINOR(__entry->dev),
FC_REASON_NAME_STAT(EXT4_FC_REASON_XATTR),
@@ -2818,6 +2820,7 @@ TRACE_EVENT(ext4_fc_stats,
FC_REASON_NAME_STAT(EXT4_FC_REASON_FALLOC_RANGE),
FC_REASON_NAME_STAT(EXT4_FC_REASON_INODE_JOURNAL_DATA),
FC_REASON_NAME_STAT(EXT4_FC_REASON_ENCRYPTED_FILENAME),
+ FC_REASON_NAME_STAT(EXT4_FC_REASON_INODE_RSV_WRITE_FAIL),
__entry->fc_commits, __entry->fc_ineligible_commits,
__entry->fc_numblks)
);
--
2.45.1.288.g0e0cd299f1-goog
Unlike JBD2 based full commits, there is no dedicated journal thread
for fast commits. Thus to reduce scheduling delays between IO
submission and completion, temporarily elevate the committer thread's
priority to match the configured priority of the JBD2 journal
thread.
Signed-off-by: Harshad Shirwadkar <[email protected]>
---
fs/ext4/ext4.h | 4 +++-
fs/ext4/fast_commit.c | 13 +++++++++++++
fs/ext4/super.c | 5 ++---
3 files changed, 18 insertions(+), 4 deletions(-)
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 3721daea2890..d52df8a85271 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -2287,10 +2287,12 @@ static inline int ext4_forced_shutdown(struct super_block *sb)
#define EXT4_DEFM_NODELALLOC 0x0800
/*
- * Default journal batch times
+ * Default journal batch times and ioprio.
*/
#define EXT4_DEF_MIN_BATCH_TIME 0
#define EXT4_DEF_MAX_BATCH_TIME 15000 /* 15ms */
+#define EXT4_DEF_JOURNAL_IOPRIO (IOPRIO_PRIO_VALUE(IOPRIO_CLASS_BE, 3))
+
/*
* Minimum number of groups in a flexgroup before we separate out
diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index 35c89bee452c..55a13d3ff681 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -1205,6 +1205,7 @@ int ext4_fc_commit(journal_t *journal, tid_t commit_tid)
int subtid = atomic_read(&sbi->s_fc_subtid);
int status = EXT4_FC_STATUS_OK, fc_bufs_before = 0;
ktime_t start_time, commit_time;
+ int old_ioprio, journal_ioprio;
if (!test_opt2(sb, JOURNAL_FAST_COMMIT))
return jbd2_complete_transaction(journal, commit_tid);
@@ -1212,6 +1213,7 @@ int ext4_fc_commit(journal_t *journal, tid_t commit_tid)
trace_ext4_fc_commit_start(sb, commit_tid);
start_time = ktime_get();
+ old_ioprio = get_current_ioprio();
restart_fc:
ret = jbd2_fc_begin_commit(journal, commit_tid);
@@ -1242,6 +1244,15 @@ int ext4_fc_commit(journal_t *journal, tid_t commit_tid)
goto fallback;
}
+ /*
+ * Now that we know that this thread is going to do a fast commit,
+ * elevate the priority to match that of the journal thread.
+ */
+ if (journal->j_task->io_context)
+ journal_ioprio = sbi->s_journal->j_task->io_context->ioprio;
+ else
+ journal_ioprio = EXT4_DEF_JOURNAL_IOPRIO;
+ set_task_ioprio(current, journal_ioprio);
fc_bufs_before = (sbi->s_fc_bytes + bsize - 1) / bsize;
ret = ext4_fc_perform_commit(journal);
if (ret < 0) {
@@ -1256,6 +1267,7 @@ int ext4_fc_commit(journal_t *journal, tid_t commit_tid)
}
atomic_inc(&sbi->s_fc_subtid);
ret = jbd2_fc_end_commit(journal);
+ set_task_ioprio(current, old_ioprio);
/*
* weight the commit time higher than the average time so we
* don't react too strongly to vast changes in the commit time
@@ -1265,6 +1277,7 @@ int ext4_fc_commit(journal_t *journal, tid_t commit_tid)
return ret;
fallback:
+ set_task_ioprio(current, old_ioprio);
ret = jbd2_fc_end_commit_fallback(journal);
ext4_fc_update_stats(sb, status, 0, 0, commit_tid);
return ret;
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 77173ec91e49..18d9d2631559 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -1833,7 +1833,6 @@ static const struct fs_parameter_spec ext4_param_specs[] = {
{}
};
-#define DEFAULT_JOURNAL_IOPRIO (IOPRIO_PRIO_VALUE(IOPRIO_CLASS_BE, 3))
#define MOPT_SET 0x0001
#define MOPT_CLEAR 0x0002
@@ -5211,7 +5210,7 @@ static int __ext4_fill_super(struct fs_context *fc, struct super_block *sb)
/* Set defaults for the variables that will be set during parsing */
if (!(ctx->spec & EXT4_SPEC_JOURNAL_IOPRIO))
- ctx->journal_ioprio = DEFAULT_JOURNAL_IOPRIO;
+ ctx->journal_ioprio = EXT4_DEF_JOURNAL_IOPRIO;
sbi->s_inode_readahead_blks = EXT4_DEF_INODE_READAHEAD_BLKS;
sbi->s_sectors_written_start =
@@ -6471,7 +6470,7 @@ static int __ext4_remount(struct fs_context *fc, struct super_block *sb)
ctx->journal_ioprio =
sbi->s_journal->j_task->io_context->ioprio;
else
- ctx->journal_ioprio = DEFAULT_JOURNAL_IOPRIO;
+ ctx->journal_ioprio = EXT4_DEF_JOURNAL_IOPRIO;
}
--
2.45.1.288.g0e0cd299f1-goog
With fast commits, if the entire commit is contained within a single
block and there isn't any data that needs a flush, we can avoid sending
expensive cache flush to disk. Single block metadata only fast commits
can be written using FUA to guarantee consistency.
Signed-off-by: Harshad Shirwadkar <[email protected]>
---
fs/ext4/ext4.h | 12 ++++++++++++
fs/ext4/ext4_jbd2.h | 20 ++++++++++++--------
fs/ext4/fast_commit.c | 23 ++++++++++++++++++-----
3 files changed, 42 insertions(+), 13 deletions(-)
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 196c513f82dd..3721daea2890 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1744,6 +1744,13 @@ struct ext4_sb_info {
*/
struct list_head s_fc_dentry_q[2]; /* directory entry updates */
unsigned int s_fc_bytes;
+
+ /*
+ * This flag indicates whether a full flush is needed on
+ * next fast commit.
+ */
+ int fc_flush_required;
+
/*
* Main fast commit lock. This lock protects accesses to the
* following fields:
@@ -2905,6 +2912,11 @@ void ext4_fc_del(struct inode *inode);
bool ext4_fc_replay_check_excluded(struct super_block *sb, ext4_fsblk_t block);
void ext4_fc_replay_cleanup(struct super_block *sb);
int ext4_fc_commit(journal_t *journal, tid_t commit_tid);
+static inline void ext4_fc_mark_needs_flush(struct super_block *sb)
+{
+ EXT4_SB(sb)->fc_flush_required = 1;
+}
+
int __init ext4_fc_init_dentry_cache(void);
void ext4_fc_destroy_dentry_cache(void);
int ext4_fc_record_regions(struct super_block *sb, int ino,
diff --git a/fs/ext4/ext4_jbd2.h b/fs/ext4/ext4_jbd2.h
index 0c77697d5e90..e3a4f5c49b6e 100644
--- a/fs/ext4/ext4_jbd2.h
+++ b/fs/ext4/ext4_jbd2.h
@@ -420,19 +420,23 @@ static inline int ext4_journal_force_commit(journal_t *journal)
static inline int ext4_jbd2_inode_add_write(handle_t *handle,
struct inode *inode, loff_t start_byte, loff_t length)
{
- if (ext4_handle_valid(handle))
- return jbd2_journal_inode_ranged_write(handle,
- EXT4_I(inode)->jinode, start_byte, length);
- return 0;
+ if (!ext4_handle_valid(handle))
+ return 0;
+
+ ext4_fc_mark_needs_flush(inode->i_sb);
+ return jbd2_journal_inode_ranged_write(handle,
+ EXT4_I(inode)->jinode, start_byte, length);
}
static inline int ext4_jbd2_inode_add_wait(handle_t *handle,
struct inode *inode, loff_t start_byte, loff_t length)
{
- if (ext4_handle_valid(handle))
- return jbd2_journal_inode_ranged_wait(handle,
- EXT4_I(inode)->jinode, start_byte, length);
- return 0;
+ if (!ext4_handle_valid(handle))
+ return 0;
+
+ ext4_fc_mark_needs_flush(inode->i_sb);
+ return jbd2_journal_inode_ranged_wait(handle,
+ EXT4_I(inode)->jinode, start_byte, length);
}
static inline void ext4_update_inode_fsync_trans(handle_t *handle,
diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index 0b7064f8dfa5..35c89bee452c 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -638,11 +638,24 @@ void ext4_fc_track_range(handle_t *handle, struct inode *inode, ext4_lblk_t star
static void ext4_fc_submit_bh(struct super_block *sb, bool is_tail)
{
blk_opf_t write_flags = REQ_SYNC;
- struct buffer_head *bh = EXT4_SB(sb)->s_fc_bh;
+ struct ext4_sb_info *sbi = EXT4_SB(sb);
+ struct buffer_head *bh = sbi->s_fc_bh;
+ int old = 1, new = 0;
+
+ if (!is_tail) {
+ /*
+ * This commit has at least 1 non-tail block,
+ * thus FLUSH is required.
+ */
+ ext4_fc_mark_needs_flush(sb);
+ } else {
+ /* Use cmpxchg to ensure that no flush requrest is lost. */
+ if (cmpxchg(&sbi->fc_flush_required, old, new))
+ /* Old value was 1, so request a flush. */
+ write_flags |= REQ_PREFLUSH;
+ write_flags |= REQ_FUA;
+ }
- /* Add REQ_FUA | REQ_PREFLUSH only its tail */
- if (test_opt(sb, BARRIER) && is_tail)
- write_flags |= REQ_FUA | REQ_PREFLUSH;
lock_buffer(bh);
set_buffer_dirty(bh);
set_buffer_uptodate(bh);
@@ -1090,7 +1103,7 @@ static int ext4_fc_perform_commit(journal_t *journal)
* 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)
+ if (sbi->fc_flush_required && journal->j_fs_dev != journal->j_dev)
blkdev_issue_flush(journal->j_fs_dev);
blk_start_plug(&plug);
--
2.45.1.288.g0e0cd299f1-goog
Mark inode dirty first and then grab i_data_sem in ext4_setattr().
Signed-off-by: Harshad Shirwadkar <[email protected]>
---
fs/ext4/inode.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index aa6440992a55..61ffbdc2fb16 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -5410,12 +5410,13 @@ int ext4_setattr(struct mnt_idmap *idmap, struct dentry *dentry,
(attr->ia_size > 0 ? attr->ia_size - 1 : 0) >>
inode->i_sb->s_blocksize_bits);
- down_write(&EXT4_I(inode)->i_data_sem);
- old_disksize = EXT4_I(inode)->i_disksize;
- EXT4_I(inode)->i_disksize = attr->ia_size;
rc = ext4_mark_inode_dirty(handle, inode);
if (!error)
error = rc;
+ down_write(&EXT4_I(inode)->i_data_sem);
+ old_disksize = EXT4_I(inode)->i_disksize;
+ EXT4_I(inode)->i_disksize = attr->ia_size;
+
/*
* We have to update i_size under i_data_sem together
* with i_disksize to avoid races with writeback code
--
2.45.1.288.g0e0cd299f1-goog
This patch updates code documentation to reflect the commit path changes
made in this series.
Signed-off-by: Harshad Shirwadkar <[email protected]>
Reviewed-by: Jan Kara <[email protected]>
---
fs/ext4/fast_commit.c | 39 ++++++++++++++++++++++++++-------------
1 file changed, 26 insertions(+), 13 deletions(-)
diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index ecbbcaf78598..b81b0292aa59 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -49,14 +49,21 @@
* that need to be committed during a fast commit in another in memory queue of
* inodes. During the commit operation, we commit in the following order:
*
- * [1] Lock inodes for any further data updates by setting COMMITTING state
- * [2] Submit data buffers of all the inodes
- * [3] Wait for [2] to complete
- * [4] Commit all the directory entry updates in the fast commit space
- * [5] Commit all the changed inode structures
- * [6] Write tail tag (this tag ensures the atomicity, please read the following
+ * [1] Lock the journal by calling jbd2_journal_lock_updates. This ensures that
+ * all the exsiting handles finish and no new handles can start.
+ * [2] Mark all the fast commit eligible inodes as undergoing fast commit
+ * by setting "EXT4_STATE_FC_COMMITTING" state.
+ * [3] Unlock the journal by calling jbd2_journal_unlock_updates. This allows
+ * starting of new handles. If new handles try to start an update on
+ * any of the inodes that are being committed, ext4_fc_track_inode()
+ * will block until those inodes have finished the fast commit.
+ * [4] Submit data buffers of all the committing inodes.
+ * [5] Wait for [4] to complete.
+ * [6] Commit all the directory entry updates in the fast commit space.
+ * [7] Commit all the changed inodes in the fast commit space and clear
+ * "EXT4_STATE_FC_COMMITTING" for these inodes.
+ * [8] Write tail tag (this tag ensures the atomicity, please read the following
* section for more details).
- * [7] Wait for [4], [5] and [6] to complete.
*
* All the inode updates must call ext4_fc_start_update() before starting an
* update. If such an ongoing update is present, fast commit waits for it to
@@ -142,6 +149,13 @@
* similarly. Thus, by converting a non-idempotent procedure into a series of
* idempotent outcomes, fast commits ensured idempotence during the replay.
*
+ * Locking
+ * -------
+ * sbi->s_fc_lock protects the fast commit inodes queue and the fast commit
+ * dentry queue. ei->i_fc_lock protects the fast commit related info in a given
+ * inode. Most of the code avoids acquiring both the locks, but if one must do
+ * that then sbi->s_fc_lock must be acquired before ei->i_fc_lock.
+ *
* TODOs
* -----
*
@@ -156,13 +170,12 @@
* fast commit recovery even if that area is invalidated by later full
* commits.
*
- * 1) Fast commit's commit path locks the entire file system during fast
- * commit. This has significant performance penalty. Instead of that, we
- * should use ext4_fc_start/stop_update functions to start inode level
- * updates from ext4_journal_start/stop. Once we do that we can drop file
- * system locking during commit path.
+ * 1) Handle more ineligible cases.
*
- * 2) Handle more ineligible cases.
+ * 2) Change ext4_fc_commit() to lookup logical to physical mapping using extent
+ * status tree. This would get rid of the need to call ext4_fc_track_inode()
+ * before acquiring i_data_sem. To do that we would need to ensure that
+ * modified extents from the extent status tree are not evicted from memory.
*/
#include <trace/events/ext4.h>
--
2.45.1.288.g0e0cd299f1-goog
If the inode that's being requested to track using ext4_fc_track_inode
is being committed, then wait until the inode finishes the
commit. Also, add calls to ext4_fc_track_inode at the right places.
With this patch, now calling ext4_reserve_inode_write() results in
inode being tracked for next fast commit. A subtle lock ordering
requirement with i_data_sem (which is documented in the code) requires
that ext4_fc_track_inode() be called before grabbing i_data_sem. So,
this patch also adds explicit ext4_fc_track_inode() calls in places
where i_data_sem grabbed.
Signed-off-by: Harshad Shirwadkar <[email protected]>
---
fs/ext4/fast_commit.c | 34 ++++++++++++++++++++++++++++++++++
fs/ext4/inline.c | 3 +++
fs/ext4/inode.c | 4 ++++
3 files changed, 41 insertions(+)
diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index a1aadebfcd66..fa93ce399440 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -581,6 +581,8 @@ static int __track_inode(struct inode *inode, void *arg, bool update)
void ext4_fc_track_inode(handle_t *handle, struct inode *inode)
{
+ struct ext4_inode_info *ei = EXT4_I(inode);
+ wait_queue_head_t *wq;
int ret;
if (S_ISDIR(inode->i_mode))
@@ -598,6 +600,38 @@ void ext4_fc_track_inode(handle_t *handle, struct inode *inode)
if (ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_INELIGIBLE))
return;
+ if (!test_opt2(inode->i_sb, JOURNAL_FAST_COMMIT) ||
+ (EXT4_SB(inode->i_sb)->s_mount_state & EXT4_FC_REPLAY) ||
+ ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_INELIGIBLE) ||
+ !list_empty(&ei->i_fc_list))
+ return;
+
+ /*
+ * If we come here, we may sleep while waiting for the inode to
+ * commit. We shouldn't be holding i_data_sem in write mode when we go
+ * to sleep since the commit path needs to grab the lock while
+ * committing the inode.
+ */
+ WARN_ON(lockdep_is_held_type(&ei->i_data_sem, 1));
+
+ while (ext4_test_inode_state(inode, EXT4_STATE_FC_COMMITTING)) {
+#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);
+ if (ext4_test_inode_state(inode, EXT4_STATE_FC_COMMITTING))
+ schedule();
+ finish_wait(wq, &wait.wq_entry);
+ }
+
ret = ext4_fc_track_template(handle, inode, __track_inode, NULL, 1);
trace_ext4_fc_track_inode(handle, inode, ret);
}
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index d5bd1e3a5d36..bd5778e680c0 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -596,6 +596,7 @@ static int ext4_convert_inline_data_to_extent(struct address_space *mapping,
goto out;
}
+ ext4_fc_track_inode(handle, inode);
ret = ext4_destroy_inline_data_nolock(handle, inode);
if (ret)
goto out;
@@ -696,6 +697,7 @@ int ext4_try_to_write_inline_data(struct address_space *mapping,
goto convert;
}
+ ext4_fc_track_inode(handle, inode);
ret = ext4_journal_get_write_access(handle, inode->i_sb, iloc.bh,
EXT4_JTR_NONE);
if (ret)
@@ -948,6 +950,7 @@ int ext4_da_write_inline_data_begin(struct address_space *mapping,
if (ret < 0)
goto out_release_page;
}
+ ext4_fc_track_inode(handle, inode);
ret = ext4_journal_get_write_access(handle, inode->i_sb, iloc.bh,
EXT4_JTR_NONE);
if (ret)
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 4bae9ccf5fe0..aa6440992a55 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -607,6 +607,7 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode,
*/
map->m_flags &= ~EXT4_MAP_FLAGS;
+ ext4_fc_track_inode(handle, inode);
/*
* New blocks allocate and/or writing to unwritten extent
* will possibly result in updating i_data, so we take
@@ -3978,6 +3979,7 @@ int ext4_punch_hole(struct file *file, loff_t offset, loff_t length)
if (stop_block > first_block) {
ext4_lblk_t hole_len = stop_block - first_block;
+ ext4_fc_track_inode(handle, inode);
down_write(&EXT4_I(inode)->i_data_sem);
ext4_discard_preallocations(inode);
@@ -4131,6 +4133,7 @@ int ext4_truncate(struct inode *inode)
if (err)
goto out_stop;
+ ext4_fc_track_inode(handle, inode);
down_write(&EXT4_I(inode)->i_data_sem);
ext4_discard_preallocations(inode);
@@ -5727,6 +5730,7 @@ ext4_reserve_inode_write(handle_t *handle, struct inode *inode,
brelse(iloc->bh);
iloc->bh = NULL;
}
+ ext4_fc_track_inode(handle, inode);
}
ext4_std_error(inode->i_sb, err);
return err;
--
2.45.1.288.g0e0cd299f1-goog