Merge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (40 commits)
  ext4: Adding error check after calling ext4_mb_regular_allocator()
  ext4: Fix dirtying of journalled buffers in data=journal mode
  ext4: re-inline ext4_rec_len_(to|from)_disk functions
  jbd2: Remove t_handle_lock from start_this_handle()
  jbd2: Change j_state_lock to be a rwlock_t
  jbd2: Use atomic variables to avoid taking t_handle_lock in jbd2_journal_stop
  ext4: Add mount options in superblock
  ext4: force block allocation on quota_off
  ext4: fix freeze deadlock under IO
  ext4: drop inode from orphan list if ext4_delete_inode() fails
  ext4: check to make make sure bd_dev is set before dereferencing it
  jbd2: Make barrier messages less scary
  ext4: don't print scary messages for allocation failures post-abort
  ext4: fix EFBIG edge case when writing to large non-extent file
  ext4: fix ext4_get_blocks references
  ext4: Always journal quota file modifications
  ext4: Fix potential memory leak in ext4_fill_super
  ext4: Don't error out the fs if the user tries to make a file too big
  ext4: allocate stripe-multiple IOs on stripe boundaries
  ext4: move aio completion after unwritten extent conversion
  ...

Fix up conflicts in fs/ext4/inode.c as per Ted.

Fix up xfs conflicts as per earlier xfs merge.

5 files changed, 236 insertions, 196 deletions

diff --git a/fs/jbd2/checkpoint.c b/fs/jbd2/checkpoint.c
index 076d1cc44f95..1c23a0f4e8a3 100644
--- a/fs/jbd2/checkpoint.c
+++ b/fs/jbd2/checkpoint.c
@@ -118,13 +118,13 @@ static int __try_to_free_cp_buf(struct journal_head *jh)
 void __jbd2_log_wait_for_space(journal_t *journal)
 {
         int nblocks, space_left;
-        assert_spin_locked(&journal->j_state_lock);
+        /* assert_spin_locked(&journal->j_state_lock); */
 
         nblocks = jbd_space_needed(journal);
         while (__jbd2_log_space_left(journal) < nblocks) {
                 if (journal->j_flags & JBD2_ABORT)
                         return;
-                spin_unlock(&journal->j_state_lock);
+                write_unlock(&journal->j_state_lock);
                 mutex_lock(&journal->j_checkpoint_mutex);
 
                 /*
@@ -138,7 +138,7 @@ void __jbd2_log_wait_for_space(journal_t *journal)
                  * filesystem, so abort the journal and leave a stack
                  * trace for forensic evidence.
                  */
-                spin_lock(&journal->j_state_lock);
+                write_lock(&journal->j_state_lock);
                 spin_lock(&journal->j_list_lock);
                 nblocks = jbd_space_needed(journal);
                 space_left = __jbd2_log_space_left(journal);
@@ -149,7 +149,7 @@ void __jbd2_log_wait_for_space(journal_t *journal)
                         if (journal->j_committing_transaction)
                                 tid = journal->j_committing_transaction->t_tid;
                         spin_unlock(&journal->j_list_lock);
-                        spin_unlock(&journal->j_state_lock);
+                        write_unlock(&journal->j_state_lock);
                         if (chkpt) {
                                 jbd2_log_do_checkpoint(journal);
                         } else if (jbd2_cleanup_journal_tail(journal) == 0) {
@@ -167,7 +167,7 @@ void __jbd2_log_wait_for_space(journal_t *journal)
                                 WARN_ON(1);
                                 jbd2_journal_abort(journal, 0);
                         }
-                        spin_lock(&journal->j_state_lock);
+                        write_lock(&journal->j_state_lock);
                 } else {
                         spin_unlock(&journal->j_list_lock);
                 }
@@ -474,7 +474,7 @@ int jbd2_cleanup_journal_tail(journal_t *journal)
          * next transaction ID we will write, and where it will
          * start. */
 
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         spin_lock(&journal->j_list_lock);
         transaction = journal->j_checkpoint_transactions;
         if (transaction) {
@@ -496,7 +496,7 @@ int jbd2_cleanup_journal_tail(journal_t *journal)
         /* If the oldest pinned transaction is at the tail of the log
            already then there's not much we can do right now. */
         if (journal->j_tail_sequence == first_tid) {
-                spin_unlock(&journal->j_state_lock);
+                write_unlock(&journal->j_state_lock);
                 return 1;
         }
 
@@ -516,7 +516,7 @@ int jbd2_cleanup_journal_tail(journal_t *journal)
         journal->j_free += freed;
         journal->j_tail_sequence = first_tid;
         journal->j_tail = blocknr;
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
 
         /*
          * If there is an external journal, we need to make sure that
@@ -775,7 +775,7 @@ void __jbd2_journal_drop_transaction(journal_t *journal, transaction_t *transact
         J_ASSERT(transaction->t_log_list == NULL);
         J_ASSERT(transaction->t_checkpoint_list == NULL);
         J_ASSERT(transaction->t_checkpoint_io_list == NULL);
-        J_ASSERT(transaction->t_updates == 0);
+        J_ASSERT(atomic_read(&transaction->t_updates) == 0);
         J_ASSERT(journal->j_committing_transaction != transaction);
         J_ASSERT(journal->j_running_transaction != transaction);
 
diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c
index 75716d3d2be0..f52e5e8049f1 100644
--- a/fs/jbd2/commit.c
+++ b/fs/jbd2/commit.c
@@ -150,11 +150,11 @@ static int journal_submit_commit_record(journal_t *journal,
          */
         if (ret == -EOPNOTSUPP && barrier_done) {
                 printk(KERN_WARNING
-                       "JBD: barrier-based sync failed on %s - "
-                       "disabling barriers\n", journal->j_devname);
-                spin_lock(&journal->j_state_lock);
+                       "JBD2: Disabling barriers on %s, "
+                       "not supported by device\n", journal->j_devname);
+                write_lock(&journal->j_state_lock);
                 journal->j_flags &= ~JBD2_BARRIER;
-                spin_unlock(&journal->j_state_lock);
+                write_unlock(&journal->j_state_lock);
 
                 /* And try again, without the barrier */
                 lock_buffer(bh);
@@ -180,11 +180,11 @@ retry:
         wait_on_buffer(bh);
         if (buffer_eopnotsupp(bh) && (journal->j_flags & JBD2_BARRIER)) {
                 printk(KERN_WARNING
-                       "JBD2: wait_on_commit_record: sync failed on %s - "
-                       "disabling barriers\n", journal->j_devname);
-                spin_lock(&journal->j_state_lock);
+                       "JBD2: %s: disabling barries on %s - not supported "
+                       "by device\n", __func__, journal->j_devname);
+                write_lock(&journal->j_state_lock);
                 journal->j_flags &= ~JBD2_BARRIER;
-                spin_unlock(&journal->j_state_lock);
+                write_unlock(&journal->j_state_lock);
 
                 lock_buffer(bh);
                 clear_buffer_dirty(bh);
@@ -400,7 +400,7 @@ void jbd2_journal_commit_transaction(journal_t *journal)
         jbd_debug(1, "JBD: starting commit of transaction %d\n",
                         commit_transaction->t_tid);
 
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         commit_transaction->t_state = T_LOCKED;
 
         /*
@@ -417,23 +417,23 @@ void jbd2_journal_commit_transaction(journal_t *journal)
                                               stats.run.rs_locked);
 
         spin_lock(&commit_transaction->t_handle_lock);
-        while (commit_transaction->t_updates) {
+        while (atomic_read(&commit_transaction->t_updates)) {
                 DEFINE_WAIT(wait);
 
                 prepare_to_wait(&journal->j_wait_updates, &wait,
                                         TASK_UNINTERRUPTIBLE);
-                if (commit_transaction->t_updates) {
+                if (atomic_read(&commit_transaction->t_updates)) {
                         spin_unlock(&commit_transaction->t_handle_lock);
-                        spin_unlock(&journal->j_state_lock);
+                        write_unlock(&journal->j_state_lock);
                         schedule();
-                        spin_lock(&journal->j_state_lock);
+                        write_lock(&journal->j_state_lock);
                         spin_lock(&commit_transaction->t_handle_lock);
                 }
                 finish_wait(&journal->j_wait_updates, &wait);
         }
         spin_unlock(&commit_transaction->t_handle_lock);
 
-        J_ASSERT (commit_transaction->t_outstanding_credits <=
+        J_ASSERT (atomic_read(&commit_transaction->t_outstanding_credits) <=
                         journal->j_max_transaction_buffers);
 
         /*
@@ -497,7 +497,7 @@ void jbd2_journal_commit_transaction(journal_t *journal)
         start_time = ktime_get();
         commit_transaction->t_log_start = journal->j_head;
         wake_up(&journal->j_wait_transaction_locked);
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
 
         jbd_debug (3, "JBD: commit phase 2\n");
 
@@ -519,19 +519,20 @@ void jbd2_journal_commit_transaction(journal_t *journal)
          * transaction!  Now comes the tricky part: we need to write out
          * metadata.  Loop over the transaction's entire buffer list:
          */
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         commit_transaction->t_state = T_COMMIT;
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
 
         trace_jbd2_commit_logging(journal, commit_transaction);
         stats.run.rs_logging = jiffies;
         stats.run.rs_flushing = jbd2_time_diff(stats.run.rs_flushing,
                                                stats.run.rs_logging);
-        stats.run.rs_blocks = commit_transaction->t_outstanding_credits;
+        stats.run.rs_blocks =
+                atomic_read(&commit_transaction->t_outstanding_credits);
         stats.run.rs_blocks_logged = 0;
 
         J_ASSERT(commit_transaction->t_nr_buffers <=
-                 commit_transaction->t_outstanding_credits);
+                 atomic_read(&commit_transaction->t_outstanding_credits));
 
         err = 0;
         descriptor = NULL;
@@ -616,7 +617,7 @@ void jbd2_journal_commit_transaction(journal_t *journal)
                  * the free space in the log, but this counter is changed
                  * by jbd2_journal_next_log_block() also.
                  */
-                commit_transaction->t_outstanding_credits--;
+                atomic_dec(&commit_transaction->t_outstanding_credits);
 
                 /* Bump b_count to prevent truncate from stumbling over
                    the shadowed buffer!  @@@ This can go if we ever get
@@ -977,7 +978,7 @@ restart_loop:
          * __jbd2_journal_drop_transaction(). Otherwise we could race with
          * other checkpointing code processing the transaction...
          */
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         spin_lock(&journal->j_list_lock);
         /*
          * Now recheck if some buffers did not get attached to the transaction
@@ -985,7 +986,7 @@ restart_loop:
          */
         if (commit_transaction->t_forget) {
                 spin_unlock(&journal->j_list_lock);
-                spin_unlock(&journal->j_state_lock);
+                write_unlock(&journal->j_state_lock);
                 goto restart_loop;
         }
 
@@ -1003,7 +1004,8 @@ restart_loop:
          * File the transaction statistics
          */
         stats.ts_tid = commit_transaction->t_tid;
-        stats.run.rs_handle_count = commit_transaction->t_handle_count;
+        stats.run.rs_handle_count =
+                atomic_read(&commit_transaction->t_handle_count);
         trace_jbd2_run_stats(journal->j_fs_dev->bd_dev,
                              commit_transaction->t_tid, &stats.run);
 
@@ -1037,7 +1039,7 @@ restart_loop:
                                 journal->j_average_commit_time*3) / 4;
         else
                 journal->j_average_commit_time = commit_time;
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
 
         if (commit_transaction->t_checkpoint_list == NULL &&
             commit_transaction->t_checkpoint_io_list == NULL) {
diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index 036880895bfc..ad5866aaf0f9 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -41,6 +41,7 @@
 #include <linux/hash.h>
 #include <linux/log2.h>
 #include <linux/vmalloc.h>
+#include <linux/backing-dev.h>
 
 #define CREATE_TRACE_POINTS
 #include <trace/events/jbd2.h>
@@ -48,8 +49,6 @@
 #include <asm/uaccess.h>
 #include <asm/page.h>
 
-EXPORT_SYMBOL(jbd2_journal_start);
-EXPORT_SYMBOL(jbd2_journal_restart);
 EXPORT_SYMBOL(jbd2_journal_extend);
 EXPORT_SYMBOL(jbd2_journal_stop);
 EXPORT_SYMBOL(jbd2_journal_lock_updates);
@@ -143,7 +142,7 @@ static int kjournald2(void *arg)
         /*
          * And now, wait forever for commit wakeup events.
          */
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
 
 loop:
         if (journal->j_flags & JBD2_UNMOUNT)
@@ -154,10 +153,10 @@ loop:
 
         if (journal->j_commit_sequence != journal->j_commit_request) {
                 jbd_debug(1, "OK, requests differ\n");
-                spin_unlock(&journal->j_state_lock);
+                write_unlock(&journal->j_state_lock);
                 del_timer_sync(&journal->j_commit_timer);
                 jbd2_journal_commit_transaction(journal);
-                spin_lock(&journal->j_state_lock);
+                write_lock(&journal->j_state_lock);
                 goto loop;
         }
 
@@ -169,9 +168,9 @@ loop:
                  * be already stopped.
                  */
                 jbd_debug(1, "Now suspending kjournald2\n");
-                spin_unlock(&journal->j_state_lock);
+                write_unlock(&journal->j_state_lock);
                 refrigerator();
-                spin_lock(&journal->j_state_lock);
+                write_lock(&journal->j_state_lock);
         } else {
                 /*
                  * We assume on resume that commits are already there,
@@ -191,9 +190,9 @@ loop:
                 if (journal->j_flags & JBD2_UNMOUNT)
                         should_sleep = 0;
                 if (should_sleep) {
-                        spin_unlock(&journal->j_state_lock);
+                        write_unlock(&journal->j_state_lock);
                         schedule();
-                        spin_lock(&journal->j_state_lock);
+                        write_lock(&journal->j_state_lock);
                 }
                 finish_wait(&journal->j_wait_commit, &wait);
         }
@@ -211,7 +210,7 @@ loop:
         goto loop;
 
 end_loop:
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
         del_timer_sync(&journal->j_commit_timer);
         journal->j_task = NULL;
         wake_up(&journal->j_wait_done_commit);
@@ -234,16 +233,16 @@ static int jbd2_journal_start_thread(journal_t *journal)
 
 static void journal_kill_thread(journal_t *journal)
 {
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         journal->j_flags |= JBD2_UNMOUNT;
 
         while (journal->j_task) {
                 wake_up(&journal->j_wait_commit);
-                spin_unlock(&journal->j_state_lock);
+                write_unlock(&journal->j_state_lock);
                 wait_event(journal->j_wait_done_commit, journal->j_task == NULL);
-                spin_lock(&journal->j_state_lock);
+                write_lock(&journal->j_state_lock);
         }
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
 }
 
 /*
@@ -310,7 +309,17 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
          */
         J_ASSERT_BH(bh_in, buffer_jbddirty(bh_in));
 
-        new_bh = alloc_buffer_head(GFP_NOFS|__GFP_NOFAIL);
+retry_alloc:
+        new_bh = alloc_buffer_head(GFP_NOFS);
+        if (!new_bh) {
+                /*
+                 * Failure is not an option, but __GFP_NOFAIL is going
+                 * away; so we retry ourselves here.
+                 */
+                congestion_wait(BLK_RW_ASYNC, HZ/50);
+                goto retry_alloc;
+        }
+
         /* keep subsequent assertions sane */
         new_bh->b_state = 0;
         init_buffer(new_bh, NULL, NULL);
@@ -442,7 +451,7 @@ int __jbd2_log_space_left(journal_t *journal)
 {
         int left = journal->j_free;
 
-        assert_spin_locked(&journal->j_state_lock);
+        /* assert_spin_locked(&journal->j_state_lock); */
 
         /*
          * Be pessimistic here about the number of those free blocks which
@@ -487,9 +496,9 @@ int jbd2_log_start_commit(journal_t *journal, tid_t tid)
 {
         int ret;
 
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         ret = __jbd2_log_start_commit(journal, tid);
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
         return ret;
 }
 
@@ -508,7 +517,7 @@ int jbd2_journal_force_commit_nested(journal_t *journal)
         transaction_t *transaction = NULL;
         tid_t tid;
 
-        spin_lock(&journal->j_state_lock);
+        read_lock(&journal->j_state_lock);
         if (journal->j_running_transaction && !current->journal_info) {
                 transaction = journal->j_running_transaction;
                 __jbd2_log_start_commit(journal, transaction->t_tid);
@@ -516,12 +525,12 @@ int jbd2_journal_force_commit_nested(journal_t *journal)
                 transaction = journal->j_committing_transaction;
 
         if (!transaction) {
-                spin_unlock(&journal->j_state_lock);
+                read_unlock(&journal->j_state_lock);
                 return 0;       /* Nothing to retry */
         }
 
         tid = transaction->t_tid;
-        spin_unlock(&journal->j_state_lock);
+        read_unlock(&journal->j_state_lock);
         jbd2_log_wait_commit(journal, tid);
         return 1;
 }
@@ -535,7 +544,7 @@ int jbd2_journal_start_commit(journal_t *journal, tid_t *ptid)
 {
         int ret = 0;
 
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         if (journal->j_running_transaction) {
                 tid_t tid = journal->j_running_transaction->t_tid;
 
@@ -554,7 +563,7 @@ int jbd2_journal_start_commit(journal_t *journal, tid_t *ptid)
                         *ptid = journal->j_committing_transaction->t_tid;
                 ret = 1;
         }
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
         return ret;
 }
 
@@ -566,26 +575,24 @@ int jbd2_log_wait_commit(journal_t *journal, tid_t tid)
 {
         int err = 0;
 
+        read_lock(&journal->j_state_lock);
 #ifdef CONFIG_JBD2_DEBUG
-        spin_lock(&journal->j_state_lock);
         if (!tid_geq(journal->j_commit_request, tid)) {
                 printk(KERN_EMERG
                        "%s: error: j_commit_request=%d, tid=%d\n",
                        __func__, journal->j_commit_request, tid);
         }
-        spin_unlock(&journal->j_state_lock);
 #endif
-        spin_lock(&journal->j_state_lock);
         while (tid_gt(tid, journal->j_commit_sequence)) {
                 jbd_debug(1, "JBD: want %d, j_commit_sequence=%d\n",
                                   tid, journal->j_commit_sequence);
                 wake_up(&journal->j_wait_commit);
-                spin_unlock(&journal->j_state_lock);
+                read_unlock(&journal->j_state_lock);
                 wait_event(journal->j_wait_done_commit,
                                 !tid_gt(tid, journal->j_commit_sequence));
-                spin_lock(&journal->j_state_lock);
+                read_lock(&journal->j_state_lock);
         }
-        spin_unlock(&journal->j_state_lock);
+        read_unlock(&journal->j_state_lock);
 
         if (unlikely(is_journal_aborted(journal))) {
                 printk(KERN_EMERG "journal commit I/O error\n");
@@ -602,7 +609,7 @@ int jbd2_journal_next_log_block(journal_t *journal, unsigned long long *retp)
 {
         unsigned long blocknr;
 
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         J_ASSERT(journal->j_free > 1);
 
         blocknr = journal->j_head;
@@ -610,7 +617,7 @@ int jbd2_journal_next_log_block(journal_t *journal, unsigned long long *retp)
         journal->j_free--;
         if (journal->j_head == journal->j_last)
                 journal->j_head = journal->j_first;
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
         return jbd2_journal_bmap(journal, blocknr, retp);
 }
 
@@ -830,7 +837,7 @@ static journal_t * journal_init_common (void)
         mutex_init(&journal->j_checkpoint_mutex);
         spin_lock_init(&journal->j_revoke_lock);
         spin_lock_init(&journal->j_list_lock);
-        spin_lock_init(&journal->j_state_lock);
+        rwlock_init(&journal->j_state_lock);
 
         journal->j_commit_interval = (HZ * JBD2_DEFAULT_MAX_COMMIT_AGE);
         journal->j_min_batch_time = 0;
@@ -1096,14 +1103,14 @@ void jbd2_journal_update_superblock(journal_t *journal, int wait)
                 set_buffer_uptodate(bh);
         }
 
-        spin_lock(&journal->j_state_lock);
+        read_lock(&journal->j_state_lock);
         jbd_debug(1,"JBD: updating superblock (start %ld, seq %d, errno %d)\n",
                   journal->j_tail, journal->j_tail_sequence, journal->j_errno);
 
         sb->s_sequence = cpu_to_be32(journal->j_tail_sequence);
         sb->s_start    = cpu_to_be32(journal->j_tail);
         sb->s_errno    = cpu_to_be32(journal->j_errno);
-        spin_unlock(&journal->j_state_lock);
+        read_unlock(&journal->j_state_lock);
 
         BUFFER_TRACE(bh, "marking dirty");
         mark_buffer_dirty(bh);
@@ -1124,12 +1131,12 @@ out:
          * any future commit will have to be careful to update the
          * superblock again to re-record the true start of the log. */
 
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         if (sb->s_start)
                 journal->j_flags &= ~JBD2_FLUSHED;
         else
                 journal->j_flags |= JBD2_FLUSHED;
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
 }
 
 /*
@@ -1391,13 +1398,9 @@ int jbd2_journal_check_used_features (journal_t *journal, unsigned long compat,
 int jbd2_journal_check_available_features (journal_t *journal, unsigned long compat,
                                       unsigned long ro, unsigned long incompat)
 {
-        journal_superblock_t *sb;
-
         if (!compat && !ro && !incompat)
                 return 1;
 
-        sb = journal->j_superblock;
-
         /* We can support any known requested features iff the
          * superblock is in version 2.  Otherwise we fail to support any
          * extended sb features. */
@@ -1545,7 +1548,7 @@ int jbd2_journal_flush(journal_t *journal)
         transaction_t *transaction = NULL;
         unsigned long old_tail;
 
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
 
         /* Force everything buffered to the log... */
         if (journal->j_running_transaction) {
@@ -1558,10 +1561,10 @@ int jbd2_journal_flush(journal_t *journal)
         if (transaction) {
                 tid_t tid = transaction->t_tid;
 
-                spin_unlock(&journal->j_state_lock);
+                write_unlock(&journal->j_state_lock);
                 jbd2_log_wait_commit(journal, tid);
         } else {
-                spin_unlock(&journal->j_state_lock);
+                write_unlock(&journal->j_state_lock);
         }
 
         /* ...and flush everything in the log out to disk. */
@@ -1585,12 +1588,12 @@ int jbd2_journal_flush(journal_t *journal)
          * the magic code for a fully-recovered superblock.  Any future
          * commits of data to the journal will restore the current
          * s_start value. */
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         old_tail = journal->j_tail;
         journal->j_tail = 0;
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
         jbd2_journal_update_superblock(journal, 1);
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         journal->j_tail = old_tail;
 
         J_ASSERT(!journal->j_running_transaction);
@@ -1598,7 +1601,7 @@ int jbd2_journal_flush(journal_t *journal)
         J_ASSERT(!journal->j_checkpoint_transactions);
         J_ASSERT(journal->j_head == journal->j_tail);
         J_ASSERT(journal->j_tail_sequence == journal->j_transaction_sequence);
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
         return 0;
 }
 
@@ -1617,7 +1620,6 @@ int jbd2_journal_flush(journal_t *journal)
 
 int jbd2_journal_wipe(journal_t *journal, int write)
 {
-        journal_superblock_t *sb;
         int err = 0;
 
         J_ASSERT (!(journal->j_flags & JBD2_LOADED));
@@ -1626,8 +1628,6 @@ int jbd2_journal_wipe(journal_t *journal, int write)
         if (err)
                 return err;
 
-        sb = journal->j_superblock;
-
         if (!journal->j_tail)
                 goto no_recovery;
 
@@ -1665,12 +1665,12 @@ void __jbd2_journal_abort_hard(journal_t *journal)
         printk(KERN_ERR "Aborting journal on device %s.\n",
                journal->j_devname);
 
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         journal->j_flags |= JBD2_ABORT;
         transaction = journal->j_running_transaction;
         if (transaction)
                 __jbd2_log_start_commit(journal, transaction->t_tid);
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
 }
 
 /* Soft abort: record the abort error status in the journal superblock,
@@ -1755,12 +1755,12 @@ int jbd2_journal_errno(journal_t *journal)
 {
         int err;
 
-        spin_lock(&journal->j_state_lock);
+        read_lock(&journal->j_state_lock);
         if (journal->j_flags & JBD2_ABORT)
                 err = -EROFS;
         else
                 err = journal->j_errno;
-        spin_unlock(&journal->j_state_lock);
+        read_unlock(&journal->j_state_lock);
         return err;
 }
 
@@ -1775,12 +1775,12 @@ int jbd2_journal_clear_err(journal_t *journal)
 {
         int err = 0;
 
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         if (journal->j_flags & JBD2_ABORT)
                 err = -EROFS;
         else
                 journal->j_errno = 0;
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
         return err;
 }
 
@@ -1793,10 +1793,10 @@ int jbd2_journal_clear_err(journal_t *journal)
  */
 void jbd2_journal_ack_err(journal_t *journal)
 {
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         if (journal->j_errno)
                 journal->j_flags |= JBD2_ACK_ERR;
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
 }
 
 int jbd2_journal_blocks_per_page(struct inode *inode)
@@ -2201,8 +2201,6 @@ void jbd2_journal_init_jbd_inode(struct jbd2_inode *jinode, struct inode *inode)
 void jbd2_journal_release_jbd_inode(journal_t *journal,
                                     struct jbd2_inode *jinode)
 {
-        int writeout = 0;
-
         if (!journal)
                 return;
 restart:
@@ -2219,9 +2217,6 @@ restart:
                 goto restart;
         }
 
-        /* Do we need to wait for data writeback? */
-        if (journal->j_committing_transaction == jinode->i_transaction)
-                writeout = 1;
         if (jinode->i_transaction) {
                 list_del(&jinode->i_list);
                 jinode->i_transaction = NULL;
diff --git a/fs/jbd2/recovery.c b/fs/jbd2/recovery.c
index 049281b7cb89..2bc4d5f116f1 100644
--- a/fs/jbd2/recovery.c
+++ b/fs/jbd2/recovery.c
@@ -285,12 +285,10 @@ int jbd2_journal_recover(journal_t *journal)
 int jbd2_journal_skip_recovery(journal_t *journal)
 {
         int                     err;
-        journal_superblock_t *  sb;
 
         struct recovery_info    info;
 
         memset (&info, 0, sizeof(info));
-        sb = journal->j_superblock;
 
         err = do_one_pass(journal, &info, PASS_SCAN);
 
@@ -299,7 +297,8 @@ int jbd2_journal_skip_recovery(journal_t *journal)
                 ++journal->j_transaction_sequence;
         } else {
 #ifdef CONFIG_JBD2_DEBUG
-                int dropped = info.end_transaction - be32_to_cpu(sb->s_sequence);
+                int dropped = info.end_transaction - 
+                        be32_to_cpu(journal->j_superblock->s_sequence);
 #endif
                 jbd_debug(1,
                           "JBD: ignoring %d transaction%s from the journal.\n",
@@ -365,11 +364,6 @@ static int do_one_pass(journal_t *journal,
         int                     tag_bytes = journal_tag_bytes(journal);
         __u32                   crc32_sum = ~0; /* Transactional Checksums */
 
-        /* Precompute the maximum metadata descriptors in a descriptor block */
-        int                     MAX_BLOCKS_PER_DESC;
-        MAX_BLOCKS_PER_DESC = ((journal->j_blocksize-sizeof(journal_header_t))
-                               / tag_bytes);
-
         /*
          * First thing is to establish what we expect to find in the log
          * (in terms of transaction IDs), and where (in terms of log
diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c
index b8e0806681bb..d95cc9d0401d 100644
--- a/fs/jbd2/transaction.c
+++ b/fs/jbd2/transaction.c
@@ -26,6 +26,8 @@
 #include <linux/mm.h>
 #include <linux/highmem.h>
 #include <linux/hrtimer.h>
+#include <linux/backing-dev.h>
+#include <linux/module.h>
 
 static void __jbd2_journal_temp_unlink_buffer(struct journal_head *jh);
 
@@ -53,6 +55,9 @@ jbd2_get_transaction(journal_t *journal, transaction_t *transaction)
         transaction->t_tid = journal->j_transaction_sequence++;
         transaction->t_expires = jiffies + journal->j_commit_interval;
         spin_lock_init(&transaction->t_handle_lock);
+        atomic_set(&transaction->t_updates, 0);
+        atomic_set(&transaction->t_outstanding_credits, 0);
+        atomic_set(&transaction->t_handle_count, 0);
         INIT_LIST_HEAD(&transaction->t_inode_list);
         INIT_LIST_HEAD(&transaction->t_private_list);
 
@@ -83,65 +88,75 @@ jbd2_get_transaction(journal_t *journal, transaction_t *transaction)
  * transaction's buffer credits.
  */
 
-static int start_this_handle(journal_t *journal, handle_t *handle)
+static int start_this_handle(journal_t *journal, handle_t *handle,
+                             int gfp_mask)
 {
         transaction_t *transaction;
         int needed;
         int nblocks = handle->h_buffer_credits;
         transaction_t *new_transaction = NULL;
-        int ret = 0;
         unsigned long ts = jiffies;
 
         if (nblocks > journal->j_max_transaction_buffers) {
                 printk(KERN_ERR "JBD: %s wants too many credits (%d > %d)\n",
                        current->comm, nblocks,
                        journal->j_max_transaction_buffers);
-                ret = -ENOSPC;
-                goto out;
+                return -ENOSPC;
         }
 
 alloc_transaction:
         if (!journal->j_running_transaction) {
-                new_transaction = kzalloc(sizeof(*new_transaction),
-                                                GFP_NOFS|__GFP_NOFAIL);
+                new_transaction = kzalloc(sizeof(*new_transaction), gfp_mask);
                 if (!new_transaction) {
-                        ret = -ENOMEM;
-                        goto out;
+                        /*
+                         * If __GFP_FS is not present, then we may be
+                         * being called from inside the fs writeback
+                         * layer, so we MUST NOT fail.  Since
+                         * __GFP_NOFAIL is going away, we will arrange
+                         * to retry the allocation ourselves.
+                         */
+                        if ((gfp_mask & __GFP_FS) == 0) {
+                                congestion_wait(BLK_RW_ASYNC, HZ/50);
+                                goto alloc_transaction;
+                        }
+                        return -ENOMEM;
                 }
         }
 
         jbd_debug(3, "New handle %p going live.\n", handle);
 
-repeat:
-
         /*
          * We need to hold j_state_lock until t_updates has been incremented,
          * for proper journal barrier handling
          */
-        spin_lock(&journal->j_state_lock);
-repeat_locked:
+repeat:
+        read_lock(&journal->j_state_lock);
         if (is_journal_aborted(journal) ||
             (journal->j_errno != 0 && !(journal->j_flags & JBD2_ACK_ERR))) {
-                spin_unlock(&journal->j_state_lock);
-                ret = -EROFS;
-                goto out;
+                read_unlock(&journal->j_state_lock);
+                kfree(new_transaction);
+                return -EROFS;
         }
 
         /* Wait on the journal's transaction barrier if necessary */
         if (journal->j_barrier_count) {
-                spin_unlock(&journal->j_state_lock);
+                read_unlock(&journal->j_state_lock);
                 wait_event(journal->j_wait_transaction_locked,
                                 journal->j_barrier_count == 0);
                 goto repeat;
         }
 
         if (!journal->j_running_transaction) {
-                if (!new_transaction) {
-                        spin_unlock(&journal->j_state_lock);
+                read_unlock(&journal->j_state_lock);
+                if (!new_transaction)
                         goto alloc_transaction;
+                write_lock(&journal->j_state_lock);
+                if (!journal->j_running_transaction) {
+                        jbd2_get_transaction(journal, new_transaction);
+                        new_transaction = NULL;
                 }
-                jbd2_get_transaction(journal, new_transaction);
-                new_transaction = NULL;
+                write_unlock(&journal->j_state_lock);
+                goto repeat;
         }
 
         transaction = journal->j_running_transaction;
@@ -155,7 +170,7 @@ repeat_locked:
 
                 prepare_to_wait(&journal->j_wait_transaction_locked,
                                         &wait, TASK_UNINTERRUPTIBLE);
-                spin_unlock(&journal->j_state_lock);
+                read_unlock(&journal->j_state_lock);
                 schedule();
                 finish_wait(&journal->j_wait_transaction_locked, &wait);
                 goto repeat;
@@ -166,8 +181,8 @@ repeat_locked:
          * buffers requested by this operation, we need to stall pending a log
          * checkpoint to free some more log space.
          */
-        spin_lock(&transaction->t_handle_lock);
-        needed = transaction->t_outstanding_credits + nblocks;
+        needed = atomic_add_return(nblocks,
+                                   &transaction->t_outstanding_credits);
 
         if (needed > journal->j_max_transaction_buffers) {
                 /*
@@ -178,11 +193,11 @@ repeat_locked:
                 DEFINE_WAIT(wait);
 
                 jbd_debug(2, "Handle %p starting new commit...\n", handle);
-                spin_unlock(&transaction->t_handle_lock);
+                atomic_sub(nblocks, &transaction->t_outstanding_credits);
                 prepare_to_wait(&journal->j_wait_transaction_locked, &wait,
                                 TASK_UNINTERRUPTIBLE);
                 __jbd2_log_start_commit(journal, transaction->t_tid);
-                spin_unlock(&journal->j_state_lock);
+                read_unlock(&journal->j_state_lock);
                 schedule();
                 finish_wait(&journal->j_wait_transaction_locked, &wait);
                 goto repeat;
@@ -215,35 +230,48 @@ repeat_locked:
          */
         if (__jbd2_log_space_left(journal) < jbd_space_needed(journal)) {
                 jbd_debug(2, "Handle %p waiting for checkpoint...\n", handle);
-                spin_unlock(&transaction->t_handle_lock);
-                __jbd2_log_wait_for_space(journal);
-                goto repeat_locked;
+                atomic_sub(nblocks, &transaction->t_outstanding_credits);
+                read_unlock(&journal->j_state_lock);
+                write_lock(&journal->j_state_lock);
+                if (__jbd2_log_space_left(journal) < jbd_space_needed(journal))
+                        __jbd2_log_wait_for_space(journal);
+                write_unlock(&journal->j_state_lock);
+                goto repeat;
         }
 
         /* OK, account for the buffers that this operation expects to
-         * use and add the handle to the running transaction. */
-
-        if (time_after(transaction->t_start, ts)) {
+         * use and add the handle to the running transaction. 
+         *
+         * In order for t_max_wait to be reliable, it must be
+         * protected by a lock.  But doing so will mean that
+         * start_this_handle() can not be run in parallel on SMP
+         * systems, which limits our scalability.  So we only enable
+         * it when debugging is enabled.  We may want to use a
+         * separate flag, eventually, so we can enable this
+         * independently of debugging.
+         */
+#ifdef CONFIG_JBD2_DEBUG
+        if (jbd2_journal_enable_debug &&
+            time_after(transaction->t_start, ts)) {
                 ts = jbd2_time_diff(ts, transaction->t_start);
+                spin_lock(&transaction->t_handle_lock);
                 if (ts > transaction->t_max_wait)
                         transaction->t_max_wait = ts;
+                spin_unlock(&transaction->t_handle_lock);
         }
-
+#endif
         handle->h_transaction = transaction;
-        transaction->t_outstanding_credits += nblocks;
-        transaction->t_updates++;
-        transaction->t_handle_count++;
+        atomic_inc(&transaction->t_updates);
+        atomic_inc(&transaction->t_handle_count);
         jbd_debug(4, "Handle %p given %d credits (total %d, free %d)\n",
-                  handle, nblocks, transaction->t_outstanding_credits,
+                  handle, nblocks,
+                  atomic_read(&transaction->t_outstanding_credits),
                   __jbd2_log_space_left(journal));
-        spin_unlock(&transaction->t_handle_lock);
-        spin_unlock(&journal->j_state_lock);
+        read_unlock(&journal->j_state_lock);
 
         lock_map_acquire(&handle->h_lockdep_map);
-out:
-        if (unlikely(new_transaction))          /* It's usually NULL */
-                kfree(new_transaction);
-        return ret;
+        kfree(new_transaction);
+        return 0;
 }
 
 static struct lock_class_key jbd2_handle_key;
@@ -278,7 +306,7 @@ static handle_t *new_handle(int nblocks)
  *
  * Return a pointer to a newly allocated handle, or NULL on failure
  */
-handle_t *jbd2_journal_start(journal_t *journal, int nblocks)
+handle_t *jbd2__journal_start(journal_t *journal, int nblocks, int gfp_mask)
 {
         handle_t *handle = journal_current_handle();
         int err;
@@ -298,7 +326,7 @@ handle_t *jbd2_journal_start(journal_t *journal, int nblocks)
 
         current->journal_info = handle;
 
-        err = start_this_handle(journal, handle);
+        err = start_this_handle(journal, handle, gfp_mask);
         if (err < 0) {
                 jbd2_free_handle(handle);
                 current->journal_info = NULL;
@@ -308,6 +336,15 @@ handle_t *jbd2_journal_start(journal_t *journal, int nblocks)
 out:
         return handle;
 }
+EXPORT_SYMBOL(jbd2__journal_start);
+
+
+handle_t *jbd2_journal_start(journal_t *journal, int nblocks)
+{
+        return jbd2__journal_start(journal, nblocks, GFP_NOFS);
+}
+EXPORT_SYMBOL(jbd2_journal_start);
+
 
 /**
  * int jbd2_journal_extend() - extend buffer credits.
@@ -342,7 +379,7 @@ int jbd2_journal_extend(handle_t *handle, int nblocks)
 
         result = 1;
 
-        spin_lock(&journal->j_state_lock);
+        read_lock(&journal->j_state_lock);
 
         /* Don't extend a locked-down transaction! */
         if (handle->h_transaction->t_state != T_RUNNING) {
@@ -352,7 +389,7 @@ int jbd2_journal_extend(handle_t *handle, int nblocks)
         }
 
         spin_lock(&transaction->t_handle_lock);
-        wanted = transaction->t_outstanding_credits + nblocks;
+        wanted = atomic_read(&transaction->t_outstanding_credits) + nblocks;
 
         if (wanted > journal->j_max_transaction_buffers) {
                 jbd_debug(3, "denied handle %p %d blocks: "
@@ -367,14 +404,14 @@ int jbd2_journal_extend(handle_t *handle, int nblocks)
         }
 
         handle->h_buffer_credits += nblocks;
-        transaction->t_outstanding_credits += nblocks;
+        atomic_add(nblocks, &transaction->t_outstanding_credits);
         result = 0;
 
         jbd_debug(3, "extended handle %p by %d\n", handle, nblocks);
 unlock:
         spin_unlock(&transaction->t_handle_lock);
 error_out:
-        spin_unlock(&journal->j_state_lock);
+        read_unlock(&journal->j_state_lock);
 out:
         return result;
 }
@@ -394,8 +431,7 @@ out:
  * transaction capabable of guaranteeing the requested number of
  * credits.
  */
-
-int jbd2_journal_restart(handle_t *handle, int nblocks)
+int jbd2__journal_restart(handle_t *handle, int nblocks, int gfp_mask)
 {
         transaction_t *transaction = handle->h_transaction;
         journal_t *journal = transaction->t_journal;
@@ -410,29 +446,35 @@ int jbd2_journal_restart(handle_t *handle, int nblocks)
          * First unlink the handle from its current transaction, and start the
          * commit on that.
          */
-        J_ASSERT(transaction->t_updates > 0);
+        J_ASSERT(atomic_read(&transaction->t_updates) > 0);
         J_ASSERT(journal_current_handle() == handle);
 
-        spin_lock(&journal->j_state_lock);
+        read_lock(&journal->j_state_lock);
         spin_lock(&transaction->t_handle_lock);
-        transaction->t_outstanding_credits -= handle->h_buffer_credits;
-        transaction->t_updates--;
-
-        if (!transaction->t_updates)
+        atomic_sub(handle->h_buffer_credits,
+                   &transaction->t_outstanding_credits);
+        if (atomic_dec_and_test(&transaction->t_updates))
                 wake_up(&journal->j_wait_updates);
         spin_unlock(&transaction->t_handle_lock);
 
         jbd_debug(2, "restarting handle %p\n", handle);
         __jbd2_log_start_commit(journal, transaction->t_tid);
-        spin_unlock(&journal->j_state_lock);
+        read_unlock(&journal->j_state_lock);
 
         lock_map_release(&handle->h_lockdep_map);
         handle->h_buffer_credits = nblocks;
-        ret = start_this_handle(journal, handle);
+        ret = start_this_handle(journal, handle, gfp_mask);
         return ret;
 }
+EXPORT_SYMBOL(jbd2__journal_restart);
 
 
+int jbd2_journal_restart(handle_t *handle, int nblocks)
+{
+        return jbd2__journal_restart(handle, nblocks, GFP_NOFS);
+}
+EXPORT_SYMBOL(jbd2_journal_restart);
+
 /**
  * void jbd2_journal_lock_updates () - establish a transaction barrier.
  * @journal:  Journal to establish a barrier on.
@@ -447,7 +489,7 @@ void jbd2_journal_lock_updates(journal_t *journal)
 {
         DEFINE_WAIT(wait);
 
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         ++journal->j_barrier_count;
 
         /* Wait until there are no running updates */
@@ -458,19 +500,19 @@ void jbd2_journal_lock_updates(journal_t *journal)
                         break;
 
                 spin_lock(&transaction->t_handle_lock);
-                if (!transaction->t_updates) {
+                if (!atomic_read(&transaction->t_updates)) {
                         spin_unlock(&transaction->t_handle_lock);
                         break;
                 }
                 prepare_to_wait(&journal->j_wait_updates, &wait,
                                 TASK_UNINTERRUPTIBLE);
                 spin_unlock(&transaction->t_handle_lock);
-                spin_unlock(&journal->j_state_lock);
+                write_unlock(&journal->j_state_lock);
                 schedule();
                 finish_wait(&journal->j_wait_updates, &wait);
-                spin_lock(&journal->j_state_lock);
+                write_lock(&journal->j_state_lock);
         }
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
 
         /*
          * We have now established a barrier against other normal updates, but
@@ -494,9 +536,9 @@ void jbd2_journal_unlock_updates (journal_t *journal)
         J_ASSERT(journal->j_barrier_count != 0);
 
         mutex_unlock(&journal->j_barrier);
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         --journal->j_barrier_count;
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
         wake_up(&journal->j_wait_transaction_locked);
 }
 
@@ -1238,7 +1280,8 @@ int jbd2_journal_stop(handle_t *handle)
 {
         transaction_t *transaction = handle->h_transaction;
         journal_t *journal = transaction->t_journal;
-        int err;
+        int err, wait_for_commit = 0;
+        tid_t tid;
         pid_t pid;
 
         J_ASSERT(journal_current_handle() == handle);
@@ -1246,7 +1289,7 @@ int jbd2_journal_stop(handle_t *handle)
         if (is_handle_aborted(handle))
                 err = -EIO;
         else {
-                J_ASSERT(transaction->t_updates > 0);
+                J_ASSERT(atomic_read(&transaction->t_updates) > 0);
                 err = 0;
         }
 
@@ -1291,9 +1334,9 @@ int jbd2_journal_stop(handle_t *handle)
 
                 journal->j_last_sync_writer = pid;
 
-                spin_lock(&journal->j_state_lock);
+                read_lock(&journal->j_state_lock);
                 commit_time = journal->j_average_commit_time;
-                spin_unlock(&journal->j_state_lock);
+                read_unlock(&journal->j_state_lock);
 
                 trans_time = ktime_to_ns(ktime_sub(ktime_get(),
                                                    transaction->t_start_time));
@@ -1314,14 +1357,8 @@ int jbd2_journal_stop(handle_t *handle)
         if (handle->h_sync)
                 transaction->t_synchronous_commit = 1;
         current->journal_info = NULL;
-        spin_lock(&transaction->t_handle_lock);
-        transaction->t_outstanding_credits -= handle->h_buffer_credits;
-        transaction->t_updates--;
-        if (!transaction->t_updates) {
-                wake_up(&journal->j_wait_updates);
-                if (journal->j_barrier_count)
-                        wake_up(&journal->j_wait_transaction_locked);
-        }
+        atomic_sub(handle->h_buffer_credits,
+                   &transaction->t_outstanding_credits);
 
         /*
          * If the handle is marked SYNC, we need to set another commit
@@ -1330,15 +1367,13 @@ int jbd2_journal_stop(handle_t *handle)
          * transaction is too old now.
          */
         if (handle->h_sync ||
-                        transaction->t_outstanding_credits >
-                                journal->j_max_transaction_buffers ||
-                        time_after_eq(jiffies, transaction->t_expires)) {
+            (atomic_read(&transaction->t_outstanding_credits) >
+             journal->j_max_transaction_buffers) ||
+            time_after_eq(jiffies, transaction->t_expires)) {
                 /* Do this even for aborted journals: an abort still
                  * completes the commit thread, it just doesn't write
                  * anything to disk. */
-                tid_t tid = transaction->t_tid;
 
-                spin_unlock(&transaction->t_handle_lock);
                 jbd_debug(2, "transaction too old, requesting commit for "
                                         "handle %p\n", handle);
                 /* This is non-blocking */
@@ -1349,11 +1384,25 @@ int jbd2_journal_stop(handle_t *handle)
                  * to wait for the commit to complete.
                  */
                 if (handle->h_sync && !(current->flags & PF_MEMALLOC))
-                        err = jbd2_log_wait_commit(journal, tid);
-        } else {
-                spin_unlock(&transaction->t_handle_lock);
+                        wait_for_commit = 1;
         }
 
+        /*
+         * Once we drop t_updates, if it goes to zero the transaction
+         * could start commiting on us and eventually disappear.  So
+         * once we do this, we must not dereference transaction
+         * pointer again.
+         */
+        tid = transaction->t_tid;
+        if (atomic_dec_and_test(&transaction->t_updates)) {
+                wake_up(&journal->j_wait_updates);
+                if (journal->j_barrier_count)
+                        wake_up(&journal->j_wait_transaction_locked);
+        }
+
+        if (wait_for_commit)
+                err = jbd2_log_wait_commit(journal, tid);
+
         lock_map_release(&handle->h_lockdep_map);
 
         jbd2_free_handle(handle);
@@ -1719,7 +1768,7 @@ static int journal_unmap_buffer(journal_t *journal, struct buffer_head *bh)
                 goto zap_buffer_unlocked;
 
         /* OK, we have data buffer in journaled mode */
-        spin_lock(&journal->j_state_lock);
+        write_lock(&journal->j_state_lock);
         jbd_lock_bh_state(bh);
         spin_lock(&journal->j_list_lock);
 
@@ -1772,7 +1821,7 @@ static int journal_unmap_buffer(journal_t *journal, struct buffer_head *bh)
                         jbd2_journal_put_journal_head(jh);
                         spin_unlock(&journal->j_list_lock);
                         jbd_unlock_bh_state(bh);
-                        spin_unlock(&journal->j_state_lock);
+                        write_unlock(&journal->j_state_lock);
                         return ret;
                 } else {
                         /* There is no currently-running transaction. So the
@@ -1786,7 +1835,7 @@ static int journal_unmap_buffer(journal_t *journal, struct buffer_head *bh)
                                 jbd2_journal_put_journal_head(jh);
                                 spin_unlock(&journal->j_list_lock);
                                 jbd_unlock_bh_state(bh);
-                                spin_unlock(&journal->j_state_lock);
+                                write_unlock(&journal->j_state_lock);
                                 return ret;
                         } else {
                                 /* The orphan record's transaction has
@@ -1810,7 +1859,7 @@ static int journal_unmap_buffer(journal_t *journal, struct buffer_head *bh)
                 jbd2_journal_put_journal_head(jh);
                 spin_unlock(&journal->j_list_lock);
                 jbd_unlock_bh_state(bh);
-                spin_unlock(&journal->j_state_lock);
+                write_unlock(&journal->j_state_lock);
                 return 0;
         } else {
                 /* Good, the buffer belongs to the running transaction.
@@ -1829,7 +1878,7 @@ zap_buffer:
 zap_buffer_no_jh:
         spin_unlock(&journal->j_list_lock);
         jbd_unlock_bh_state(bh);
-        spin_unlock(&journal->j_state_lock);
+        write_unlock(&journal->j_state_lock);
 zap_buffer_unlocked:
         clear_buffer_dirty(bh);
         J_ASSERT_BH(bh, !buffer_jbddirty(bh));
@@ -2136,9 +2185,9 @@ int jbd2_journal_begin_ordered_truncate(journal_t *journal,
         /* Locks are here just to force reading of recent values, it is
          * enough that the transaction was not committing before we started
          * a transaction adding the inode to orphan list */
-        spin_lock(&journal->j_state_lock);
+        read_lock(&journal->j_state_lock);
         commit_trans = journal->j_committing_transaction;
-        spin_unlock(&journal->j_state_lock);
+        read_unlock(&journal->j_state_lock);
         spin_lock(&journal->j_list_lock);
         inode_trans = jinode->i_transaction;
         spin_unlock(&journal->j_list_lock);