jbd2: Use tracepoints for history file
[safe/jmp/linux-2.6] / fs / jbd2 / commit.c
index 2856e11..d4cfd6d 100644 (file)
 #include <linux/slab.h>
 #include <linux/mm.h>
 #include <linux/pagemap.h>
+#include <linux/jiffies.h>
+#include <linux/crc32.h>
+#include <linux/writeback.h>
+#include <linux/backing-dev.h>
+#include <linux/bio.h>
+#include <linux/blkdev.h>
+#include <trace/events/jbd2.h>
 
 /*
  * Default IO end handler for temporary BJ_IO buffer_heads.
@@ -35,8 +42,8 @@ static void journal_end_buffer_io_sync(struct buffer_head *bh, int uptodate)
 }
 
 /*
- * When an ext3-ordered file is truncated, it is possible that many pages are
- * not sucessfully freed, because they are attached to a committing transaction.
+ * When an ext4 file is truncated, it is possible that some pages are not
+ * successfully freed, because they are attached to a committing transaction.
  * After the transaction commits, these pages are left on the LRU, with no
  * ->mapping, and with attached buffers.  These pages are trivially reclaimable
  * by the VM, but their apparent absence upsets the VM accounting, and it makes
@@ -63,7 +70,7 @@ static void release_buffer_page(struct buffer_head *bh)
                goto nope;
 
        /* OK, it's a truncated page */
-       if (TestSetPageLocked(page))
+       if (!trylock_page(page))
                goto nope;
 
        page_cache_get(page);
@@ -78,34 +85,24 @@ nope:
 }
 
 /*
- * Try to acquire jbd_lock_bh_state() against the buffer, when j_list_lock is
- * held.  For ranking reasons we must trylock.  If we lose, schedule away and
- * return 0.  j_list_lock is dropped in this case.
- */
-static int inverted_lock(journal_t *journal, struct buffer_head *bh)
-{
-       if (!jbd_trylock_bh_state(bh)) {
-               spin_unlock(&journal->j_list_lock);
-               schedule();
-               return 0;
-       }
-       return 1;
-}
-
-/* Done it all: now write the commit record.  We should have
+ * Done it all: now submit the commit record.  We should have
  * cleaned up our previous buffers by now, so if we are in abort
  * mode we can now just skip the rest of the journal write
  * entirely.
  *
  * Returns 1 if the journal needs to be aborted or 0 on success
  */
-static int journal_write_commit_record(journal_t *journal,
-                                       transaction_t *commit_transaction)
+static int journal_submit_commit_record(journal_t *journal,
+                                       transaction_t *commit_transaction,
+                                       struct buffer_head **cbh,
+                                       __u32 crc32_sum)
 {
        struct journal_head *descriptor;
+       struct commit_header *tmp;
        struct buffer_head *bh;
-       int i, ret;
+       int ret;
        int barrier_done = 0;
+       struct timespec now = current_kernel_time();
 
        if (is_journal_aborted(journal))
                return 0;
@@ -116,169 +113,230 @@ static int journal_write_commit_record(journal_t *journal,
 
        bh = jh2bh(descriptor);
 
-       /* AKPM: buglet - add `i' to tmp! */
-       for (i = 0; i < bh->b_size; i += 512) {
-               journal_header_t *tmp = (journal_header_t*)bh->b_data;
-               tmp->h_magic = cpu_to_be32(JBD2_MAGIC_NUMBER);
-               tmp->h_blocktype = cpu_to_be32(JBD2_COMMIT_BLOCK);
-               tmp->h_sequence = cpu_to_be32(commit_transaction->t_tid);
+       tmp = (struct commit_header *)bh->b_data;
+       tmp->h_magic = cpu_to_be32(JBD2_MAGIC_NUMBER);
+       tmp->h_blocktype = cpu_to_be32(JBD2_COMMIT_BLOCK);
+       tmp->h_sequence = cpu_to_be32(commit_transaction->t_tid);
+       tmp->h_commit_sec = cpu_to_be64(now.tv_sec);
+       tmp->h_commit_nsec = cpu_to_be32(now.tv_nsec);
+
+       if (JBD2_HAS_COMPAT_FEATURE(journal,
+                                   JBD2_FEATURE_COMPAT_CHECKSUM)) {
+               tmp->h_chksum_type      = JBD2_CRC32_CHKSUM;
+               tmp->h_chksum_size      = JBD2_CRC32_CHKSUM_SIZE;
+               tmp->h_chksum[0]        = cpu_to_be32(crc32_sum);
        }
 
-       JBUFFER_TRACE(descriptor, "write commit block");
-       set_buffer_dirty(bh);
-       if (journal->j_flags & JBD2_BARRIER) {
+       JBUFFER_TRACE(descriptor, "submit commit block");
+       lock_buffer(bh);
+       clear_buffer_dirty(bh);
+       set_buffer_uptodate(bh);
+       bh->b_end_io = journal_end_buffer_io_sync;
+
+       if (journal->j_flags & JBD2_BARRIER &&
+           !JBD2_HAS_INCOMPAT_FEATURE(journal,
+                                      JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT)) {
                set_buffer_ordered(bh);
                barrier_done = 1;
        }
-       ret = sync_dirty_buffer(bh);
+       ret = submit_bh(WRITE_SYNC_PLUG, bh);
+       if (barrier_done)
+               clear_buffer_ordered(bh);
+
        /* is it possible for another commit to fail at roughly
         * the same time as this one?  If so, we don't want to
         * trust the barrier flag in the super, but instead want
         * to remember if we sent a barrier request
         */
        if (ret == -EOPNOTSUPP && barrier_done) {
-               char b[BDEVNAME_SIZE];
-
                printk(KERN_WARNING
-                       "JBD: barrier-based sync failed on %s - "
-                       "disabling barriers\n",
-                       bdevname(journal->j_dev, b));
+                      "JBD: barrier-based sync failed on %s - "
+                      "disabling barriers\n", journal->j_devname);
                spin_lock(&journal->j_state_lock);
                journal->j_flags &= ~JBD2_BARRIER;
                spin_unlock(&journal->j_state_lock);
 
                /* And try again, without the barrier */
-               clear_buffer_ordered(bh);
+               lock_buffer(bh);
                set_buffer_uptodate(bh);
-               set_buffer_dirty(bh);
-               ret = sync_dirty_buffer(bh);
+               clear_buffer_dirty(bh);
+               ret = submit_bh(WRITE_SYNC_PLUG, bh);
        }
-       put_bh(bh);             /* One for getblk() */
-       jbd2_journal_put_journal_head(descriptor);
-
-       return (ret == -EIO);
+       *cbh = bh;
+       return ret;
 }
 
-static void journal_do_submit_data(struct buffer_head **wbuf, int bufs)
+/*
+ * This function along with journal_submit_commit_record
+ * allows to write the commit record asynchronously.
+ */
+static int journal_wait_on_commit_record(journal_t *journal,
+                                        struct buffer_head *bh)
 {
-       int i;
+       int ret = 0;
+
+retry:
+       clear_buffer_dirty(bh);
+       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);
+               journal->j_flags &= ~JBD2_BARRIER;
+               spin_unlock(&journal->j_state_lock);
+
+               lock_buffer(bh);
+               clear_buffer_dirty(bh);
+               set_buffer_uptodate(bh);
+               bh->b_end_io = journal_end_buffer_io_sync;
 
-       for (i = 0; i < bufs; i++) {
-               wbuf[i]->b_end_io = end_buffer_write_sync;
-               /* We use-up our safety reference in submit_bh() */
-               submit_bh(WRITE, wbuf[i]);
+               ret = submit_bh(WRITE_SYNC_PLUG, bh);
+               if (ret) {
+                       unlock_buffer(bh);
+                       return ret;
+               }
+               goto retry;
        }
+
+       if (unlikely(!buffer_uptodate(bh)))
+               ret = -EIO;
+       put_bh(bh);            /* One for getblk() */
+       jbd2_journal_put_journal_head(bh2jh(bh));
+
+       return ret;
 }
 
 /*
- *  Submit all the data buffers to disk
+ * write the filemap data using writepage() address_space_operations.
+ * We don't do block allocation here even for delalloc. We don't
+ * use writepages() because with dealyed allocation we may be doing
+ * block allocation in writepages().
  */
-static void journal_submit_data_buffers(journal_t *journal,
-                               transaction_t *commit_transaction)
+static int journal_submit_inode_data_buffers(struct address_space *mapping)
 {
-       struct journal_head *jh;
-       struct buffer_head *bh;
-       int locked;
-       int bufs = 0;
-       struct buffer_head **wbuf = journal->j_wbuf;
+       int ret;
+       struct writeback_control wbc = {
+               .sync_mode =  WB_SYNC_ALL,
+               .nr_to_write = mapping->nrpages * 2,
+               .range_start = 0,
+               .range_end = i_size_read(mapping->host),
+       };
+
+       ret = generic_writepages(mapping, &wbc);
+       return ret;
+}
 
-       /*
-        * Whenever we unlock the journal and sleep, things can get added
-        * onto ->t_sync_datalist, so we have to keep looping back to
-        * write_out_data until we *know* that the list is empty.
-        *
-        * Cleanup any flushed data buffers from the data list.  Even in
-        * abort mode, we want to flush this out as soon as possible.
-        */
-write_out_data:
-       cond_resched();
-       spin_lock(&journal->j_list_lock);
+/*
+ * Submit all the data buffers of inode associated with the transaction to
+ * disk.
+ *
+ * We are in a committing transaction. Therefore no new inode can be added to
+ * our inode list. We use JI_COMMIT_RUNNING flag to protect inode we currently
+ * operate on from being released while we write out pages.
+ */
+static int journal_submit_data_buffers(journal_t *journal,
+               transaction_t *commit_transaction)
+{
+       struct jbd2_inode *jinode;
+       int err, ret = 0;
+       struct address_space *mapping;
 
-       while (commit_transaction->t_sync_datalist) {
-               jh = commit_transaction->t_sync_datalist;
-               bh = jh2bh(jh);
-               locked = 0;
-
-               /* Get reference just to make sure buffer does not disappear
-                * when we are forced to drop various locks */
-               get_bh(bh);
-               /* If the buffer is dirty, we need to submit IO and hence
-                * we need the buffer lock. We try to lock the buffer without
-                * blocking. If we fail, we need to drop j_list_lock and do
-                * blocking lock_buffer().
+       spin_lock(&journal->j_list_lock);
+       list_for_each_entry(jinode, &commit_transaction->t_inode_list, i_list) {
+               mapping = jinode->i_vfs_inode->i_mapping;
+               jinode->i_flags |= JI_COMMIT_RUNNING;
+               spin_unlock(&journal->j_list_lock);
+               /*
+                * submit the inode data buffers. We use writepage
+                * instead of writepages. Because writepages can do
+                * block allocation  with delalloc. We need to write
+                * only allocated blocks here.
                 */
-               if (buffer_dirty(bh)) {
-                       if (test_set_buffer_locked(bh)) {
-                               BUFFER_TRACE(bh, "needs blocking lock");
-                               spin_unlock(&journal->j_list_lock);
-                               /* Write out all data to prevent deadlocks */
-                               journal_do_submit_data(wbuf, bufs);
-                               bufs = 0;
-                               lock_buffer(bh);
-                               spin_lock(&journal->j_list_lock);
-                       }
-                       locked = 1;
-               }
-               /* We have to get bh_state lock. Again out of order, sigh. */
-               if (!inverted_lock(journal, bh)) {
-                       jbd_lock_bh_state(bh);
-                       spin_lock(&journal->j_list_lock);
-               }
-               /* Someone already cleaned up the buffer? */
-               if (!buffer_jbd(bh)
-                       || jh->b_transaction != commit_transaction
-                       || jh->b_jlist != BJ_SyncData) {
-                       jbd_unlock_bh_state(bh);
-                       if (locked)
-                               unlock_buffer(bh);
-                       BUFFER_TRACE(bh, "already cleaned up");
-                       put_bh(bh);
-                       continue;
-               }
-               if (locked && test_clear_buffer_dirty(bh)) {
-                       BUFFER_TRACE(bh, "needs writeout, adding to array");
-                       wbuf[bufs++] = bh;
-                       __jbd2_journal_file_buffer(jh, commit_transaction,
-                                               BJ_Locked);
-                       jbd_unlock_bh_state(bh);
-                       if (bufs == journal->j_wbufsize) {
-                               spin_unlock(&journal->j_list_lock);
-                               journal_do_submit_data(wbuf, bufs);
-                               bufs = 0;
-                               goto write_out_data;
-                       }
-               } else if (!locked && buffer_locked(bh)) {
-                       __jbd2_journal_file_buffer(jh, commit_transaction,
-                                               BJ_Locked);
-                       jbd_unlock_bh_state(bh);
-                       put_bh(bh);
-               } else {
-                       BUFFER_TRACE(bh, "writeout complete: unfile");
-                       __jbd2_journal_unfile_buffer(jh);
-                       jbd_unlock_bh_state(bh);
-                       if (locked)
-                               unlock_buffer(bh);
-                       jbd2_journal_remove_journal_head(bh);
-                       /* Once for our safety reference, once for
-                        * jbd2_journal_remove_journal_head() */
-                       put_bh(bh);
-                       put_bh(bh);
+               trace_jbd2_submit_inode_data(jinode->i_vfs_inode);
+               err = journal_submit_inode_data_buffers(mapping);
+               if (!ret)
+                       ret = err;
+               spin_lock(&journal->j_list_lock);
+               J_ASSERT(jinode->i_transaction == commit_transaction);
+               jinode->i_flags &= ~JI_COMMIT_RUNNING;
+               wake_up_bit(&jinode->i_flags, __JI_COMMIT_RUNNING);
+       }
+       spin_unlock(&journal->j_list_lock);
+       return ret;
+}
+
+/*
+ * Wait for data submitted for writeout, refile inodes to proper
+ * transaction if needed.
+ *
+ */
+static int journal_finish_inode_data_buffers(journal_t *journal,
+               transaction_t *commit_transaction)
+{
+       struct jbd2_inode *jinode, *next_i;
+       int err, ret = 0;
+
+       /* For locking, see the comment in journal_submit_data_buffers() */
+       spin_lock(&journal->j_list_lock);
+       list_for_each_entry(jinode, &commit_transaction->t_inode_list, i_list) {
+               jinode->i_flags |= JI_COMMIT_RUNNING;
+               spin_unlock(&journal->j_list_lock);
+               err = filemap_fdatawait(jinode->i_vfs_inode->i_mapping);
+               if (err) {
+                       /*
+                        * Because AS_EIO is cleared by
+                        * wait_on_page_writeback_range(), set it again so
+                        * that user process can get -EIO from fsync().
+                        */
+                       set_bit(AS_EIO,
+                               &jinode->i_vfs_inode->i_mapping->flags);
+
+                       if (!ret)
+                               ret = err;
                }
+               spin_lock(&journal->j_list_lock);
+               jinode->i_flags &= ~JI_COMMIT_RUNNING;
+               wake_up_bit(&jinode->i_flags, __JI_COMMIT_RUNNING);
+       }
 
-               if (lock_need_resched(&journal->j_list_lock)) {
-                       spin_unlock(&journal->j_list_lock);
-                       goto write_out_data;
+       /* Now refile inode to proper lists */
+       list_for_each_entry_safe(jinode, next_i,
+                                &commit_transaction->t_inode_list, i_list) {
+               list_del(&jinode->i_list);
+               if (jinode->i_next_transaction) {
+                       jinode->i_transaction = jinode->i_next_transaction;
+                       jinode->i_next_transaction = NULL;
+                       list_add(&jinode->i_list,
+                               &jinode->i_transaction->t_inode_list);
+               } else {
+                       jinode->i_transaction = NULL;
                }
        }
        spin_unlock(&journal->j_list_lock);
-       journal_do_submit_data(wbuf, bufs);
+
+       return ret;
+}
+
+static __u32 jbd2_checksum_data(__u32 crc32_sum, struct buffer_head *bh)
+{
+       struct page *page = bh->b_page;
+       char *addr;
+       __u32 checksum;
+
+       addr = kmap_atomic(page, KM_USER0);
+       checksum = crc32_be(crc32_sum,
+               (void *)(addr + offset_in_page(bh->b_data)), bh->b_size);
+       kunmap_atomic(addr, KM_USER0);
+
+       return checksum;
 }
 
-static inline void write_tag_block(int tag_bytes, journal_block_tag_t *tag,
+static void write_tag_block(int tag_bytes, journal_block_tag_t *tag,
                                   unsigned long long block)
 {
        tag->t_blocknr = cpu_to_be32(block & (u32)~0);
-       if (tag_bytes > JBD_TAG_SIZE32)
+       if (tag_bytes > JBD2_TAG_SIZE32)
                tag->t_blocknr_high = cpu_to_be32((block >> 31) >> 1);
 }
 
@@ -290,6 +348,7 @@ static inline void write_tag_block(int tag_bytes, journal_block_tag_t *tag,
  */
 void jbd2_journal_commit_transaction(journal_t *journal)
 {
+       struct transaction_stats_s stats;
        transaction_t *commit_transaction;
        struct journal_head *jh, *new_jh, *descriptor;
        struct buffer_head **wbuf = journal->j_wbuf;
@@ -297,14 +356,19 @@ void jbd2_journal_commit_transaction(journal_t *journal)
        int flags;
        int err;
        unsigned long long blocknr;
+       ktime_t start_time;
+       u64 commit_time;
        char *tagp = NULL;
        journal_header_t *header;
        journal_block_tag_t *tag = NULL;
        int space_left = 0;
        int first_tag = 0;
        int tag_flag;
-       int i;
+       int i, to_free = 0;
        int tag_bytes = journal_tag_bytes(journal);
+       struct buffer_head *cbh = NULL; /* For transactional checksums */
+       __u32 crc32_sum = ~0;
+       int write_op = WRITE;
 
        /*
         * First job: lock down the current transaction and wait for
@@ -331,12 +395,26 @@ void jbd2_journal_commit_transaction(journal_t *journal)
        commit_transaction = journal->j_running_transaction;
        J_ASSERT(commit_transaction->t_state == T_RUNNING);
 
+       trace_jbd2_start_commit(journal, commit_transaction);
        jbd_debug(1, "JBD: starting commit of transaction %d\n",
                        commit_transaction->t_tid);
 
        spin_lock(&journal->j_state_lock);
        commit_transaction->t_state = T_LOCKED;
 
+       /*
+        * Use plugged writes here, since we want to submit several before
+        * we unplug the device. We don't do explicit unplugging in here,
+        * instead we rely on sync_buffer() doing the unplug for us.
+        */
+       if (commit_transaction->t_synchronous_commit)
+               write_op = WRITE_SYNC_PLUG;
+       trace_jbd2_commit_locking(journal, commit_transaction);
+       stats.run.rs_wait = commit_transaction->t_max_wait;
+       stats.run.rs_locked = jiffies;
+       stats.run.rs_running = jbd2_time_diff(commit_transaction->t_start,
+                                             stats.run.rs_locked);
+
        spin_lock(&commit_transaction->t_handle_lock);
        while (commit_transaction->t_updates) {
                DEFINE_WAIT(wait);
@@ -384,7 +462,7 @@ void jbd2_journal_commit_transaction(journal_t *journal)
                        struct buffer_head *bh = jh2bh(jh);
 
                        jbd_lock_bh_state(bh);
-                       jbd2_slab_free(jh->b_committed_data, bh->b_size);
+                       jbd2_free(jh->b_committed_data, bh->b_size);
                        jh->b_committed_data = NULL;
                        jbd_unlock_bh_state(bh);
                }
@@ -407,9 +485,15 @@ void jbd2_journal_commit_transaction(journal_t *journal)
         */
        jbd2_journal_switch_revoke_table(journal);
 
+       trace_jbd2_commit_flushing(journal, commit_transaction);
+       stats.run.rs_flushing = jiffies;
+       stats.run.rs_locked = jbd2_time_diff(stats.run.rs_locked,
+                                            stats.run.rs_flushing);
+
        commit_transaction->t_state = T_FLUSH;
        journal->j_committing_transaction = commit_transaction;
        journal->j_running_transaction = NULL;
+       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);
@@ -417,87 +501,38 @@ void jbd2_journal_commit_transaction(journal_t *journal)
        jbd_debug (3, "JBD: commit phase 2\n");
 
        /*
-        * First, drop modified flag: all accesses to the buffers
-        * will be tracked for a new trasaction only -bzzz
-        */
-       spin_lock(&journal->j_list_lock);
-       if (commit_transaction->t_buffers) {
-               new_jh = jh = commit_transaction->t_buffers->b_tnext;
-               do {
-                       J_ASSERT_JH(new_jh, new_jh->b_modified == 1 ||
-                                       new_jh->b_modified == 0);
-                       new_jh->b_modified = 0;
-                       new_jh = new_jh->b_tnext;
-               } while (new_jh != jh);
-       }
-       spin_unlock(&journal->j_list_lock);
-
-       /*
         * Now start flushing things to disk, in the order they appear
         * on the transaction lists.  Data blocks go first.
         */
-       err = 0;
-       journal_submit_data_buffers(journal, commit_transaction);
-
-       /*
-        * Wait for all previously submitted IO to complete.
-        */
-       spin_lock(&journal->j_list_lock);
-       while (commit_transaction->t_locked_list) {
-               struct buffer_head *bh;
-
-               jh = commit_transaction->t_locked_list->b_tprev;
-               bh = jh2bh(jh);
-               get_bh(bh);
-               if (buffer_locked(bh)) {
-                       spin_unlock(&journal->j_list_lock);
-                       wait_on_buffer(bh);
-                       if (unlikely(!buffer_uptodate(bh)))
-                               err = -EIO;
-                       spin_lock(&journal->j_list_lock);
-               }
-               if (!inverted_lock(journal, bh)) {
-                       put_bh(bh);
-                       spin_lock(&journal->j_list_lock);
-                       continue;
-               }
-               if (buffer_jbd(bh) && jh->b_jlist == BJ_Locked) {
-                       __jbd2_journal_unfile_buffer(jh);
-                       jbd_unlock_bh_state(bh);
-                       jbd2_journal_remove_journal_head(bh);
-                       put_bh(bh);
-               } else {
-                       jbd_unlock_bh_state(bh);
-               }
-               put_bh(bh);
-               cond_resched_lock(&journal->j_list_lock);
-       }
-       spin_unlock(&journal->j_list_lock);
-
+       err = journal_submit_data_buffers(journal, commit_transaction);
        if (err)
-               __jbd2_journal_abort_hard(journal);
+               jbd2_journal_abort(journal, err);
 
-       jbd2_journal_write_revoke_records(journal, commit_transaction);
+       jbd2_journal_write_revoke_records(journal, commit_transaction,
+                                         write_op);
 
        jbd_debug(3, "JBD: commit phase 2\n");
 
        /*
-        * If we found any dirty or locked buffers, then we should have
-        * looped back up to the write_out_data label.  If there weren't
-        * any then journal_clean_data_list should have wiped the list
-        * clean by now, so check that it is in fact empty.
-        */
-       J_ASSERT (commit_transaction->t_sync_datalist == NULL);
-
-       jbd_debug (3, "JBD: commit phase 3\n");
-
-       /*
         * Way to go: we have now written out all of the data for a
         * 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);
        commit_transaction->t_state = T_COMMIT;
+       spin_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_logged = 0;
 
+       J_ASSERT(commit_transaction->t_nr_buffers <=
+                commit_transaction->t_outstanding_credits);
+
+       err = 0;
        descriptor = NULL;
        bufs = 0;
        while (commit_transaction->t_buffers) {
@@ -507,10 +542,15 @@ void jbd2_journal_commit_transaction(journal_t *journal)
                jh = commit_transaction->t_buffers;
 
                /* If we're in abort mode, we just un-journal the buffer and
-                  release it for background writing. */
+                  release it. */
 
                if (is_journal_aborted(journal)) {
+                       clear_buffer_jbddirty(jh2bh(jh));
                        JBUFFER_TRACE(jh, "journal is aborting: refile");
+                       jbd2_buffer_abort_trigger(jh,
+                                                 jh->b_frozen_data ?
+                                                 jh->b_frozen_triggers :
+                                                 jh->b_triggers);
                        jbd2_journal_refile_buffer(journal, jh);
                        /* If that was the last one, we need to clean up
                         * any descriptor buffers which may have been
@@ -533,7 +573,7 @@ void jbd2_journal_commit_transaction(journal_t *journal)
 
                        descriptor = jbd2_journal_get_descriptor_buffer(journal);
                        if (!descriptor) {
-                               __jbd2_journal_abort_hard(journal);
+                               jbd2_journal_abort(journal, -EIO);
                                continue;
                        }
 
@@ -566,7 +606,7 @@ void jbd2_journal_commit_transaction(journal_t *journal)
                   and repeat this loop: we'll fall into the
                   refile-on-abort condition above. */
                if (err) {
-                       __jbd2_journal_abort_hard(journal);
+                       jbd2_journal_abort(journal, err);
                        continue;
                }
 
@@ -639,13 +679,23 @@ void jbd2_journal_commit_transaction(journal_t *journal)
 start_journal_io:
                        for (i = 0; i < bufs; i++) {
                                struct buffer_head *bh = wbuf[i];
+                               /*
+                                * Compute checksum.
+                                */
+                               if (JBD2_HAS_COMPAT_FEATURE(journal,
+                                       JBD2_FEATURE_COMPAT_CHECKSUM)) {
+                                       crc32_sum =
+                                           jbd2_checksum_data(crc32_sum, bh);
+                               }
+
                                lock_buffer(bh);
                                clear_buffer_dirty(bh);
                                set_buffer_uptodate(bh);
                                bh->b_end_io = journal_end_buffer_io_sync;
-                               submit_bh(WRITE, bh);
+                               submit_bh(write_op, bh);
                        }
                        cond_resched();
+                       stats.run.rs_blocks_logged += bufs;
 
                        /* Force a new descriptor to be generated next
                            time round the loop. */
@@ -654,6 +704,35 @@ start_journal_io:
                }
        }
 
+       /* Done it all: now write the commit record asynchronously. */
+
+       if (JBD2_HAS_INCOMPAT_FEATURE(journal,
+                                     JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT)) {
+               err = journal_submit_commit_record(journal, commit_transaction,
+                                                &cbh, crc32_sum);
+               if (err)
+                       __jbd2_journal_abort_hard(journal);
+               if (journal->j_flags & JBD2_BARRIER)
+                       blkdev_issue_flush(journal->j_dev, NULL);
+       }
+
+       /*
+        * This is the right place to wait for data buffers both for ASYNC
+        * and !ASYNC commit. If commit is ASYNC, we need to wait only after
+        * the commit block went to disk (which happens above). If commit is
+        * SYNC, we need to wait for data buffers before we start writing
+        * commit block, which happens below in such setting.
+        */
+       err = journal_finish_inode_data_buffers(journal, commit_transaction);
+       if (err) {
+               printk(KERN_WARNING
+                       "JBD2: Detected IO errors while flushing file data "
+                      "on %s\n", journal->j_devname);
+               if (journal->j_flags & JBD2_ABORT_ON_SYNCDATA_ERR)
+                       jbd2_journal_abort(journal, err);
+               err = 0;
+       }
+
        /* Lo and behold: we have just managed to send a transaction to
            the log.  Before we can commit it, wait for the IO so far to
            complete.  Control buffers being written are on the
@@ -665,7 +744,7 @@ start_journal_io:
           so we incur less scheduling load.
        */
 
-       jbd_debug(3, "JBD: commit phase 4\n");
+       jbd_debug(3, "JBD: commit phase 3\n");
 
        /*
         * akpm: these are BJ_IO, and j_list_lock is not needed.
@@ -724,7 +803,7 @@ wait_for_iobuf:
 
        J_ASSERT (commit_transaction->t_shadow_list == NULL);
 
-       jbd_debug(3, "JBD: commit phase 5\n");
+       jbd_debug(3, "JBD: commit phase 4\n");
 
        /* Here we wait for the revoke record and descriptor record buffers */
  wait_for_ctlbuf:
@@ -751,22 +830,32 @@ wait_for_iobuf:
                /* AKPM: bforget here */
        }
 
-       jbd_debug(3, "JBD: commit phase 6\n");
+       if (err)
+               jbd2_journal_abort(journal, err);
+
+       jbd_debug(3, "JBD: commit phase 5\n");
 
-       if (journal_write_commit_record(journal, commit_transaction))
-               err = -EIO;
+       if (!JBD2_HAS_INCOMPAT_FEATURE(journal,
+                                      JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT)) {
+               err = journal_submit_commit_record(journal, commit_transaction,
+                                               &cbh, crc32_sum);
+               if (err)
+                       __jbd2_journal_abort_hard(journal);
+       }
+       if (!err && !is_journal_aborted(journal))
+               err = journal_wait_on_commit_record(journal, cbh);
 
        if (err)
-               __jbd2_journal_abort_hard(journal);
+               jbd2_journal_abort(journal, err);
 
        /* End of a transaction!  Finally, we can do checkpoint
            processing: any buffers committed as a result of this
            transaction can be removed from any checkpoint list it was on
            before. */
 
-       jbd_debug(3, "JBD: commit phase 7\n");
+       jbd_debug(3, "JBD: commit phase 6\n");
 
-       J_ASSERT(commit_transaction->t_sync_datalist == NULL);
+       J_ASSERT(list_empty(&commit_transaction->t_inode_list));
        J_ASSERT(commit_transaction->t_buffers == NULL);
        J_ASSERT(commit_transaction->t_checkpoint_list == NULL);
        J_ASSERT(commit_transaction->t_iobuf_list == NULL);
@@ -799,23 +888,29 @@ restart_loop:
                 * data.
                 *
                 * Otherwise, we can just throw away the frozen data now.
+                *
+                * We also know that the frozen data has already fired
+                * its triggers if they exist, so we can clear that too.
                 */
                if (jh->b_committed_data) {
-                       jbd2_slab_free(jh->b_committed_data, bh->b_size);
+                       jbd2_free(jh->b_committed_data, bh->b_size);
                        jh->b_committed_data = NULL;
                        if (jh->b_frozen_data) {
                                jh->b_committed_data = jh->b_frozen_data;
                                jh->b_frozen_data = NULL;
+                               jh->b_frozen_triggers = NULL;
                        }
                } else if (jh->b_frozen_data) {
-                       jbd2_slab_free(jh->b_frozen_data, bh->b_size);
+                       jbd2_free(jh->b_frozen_data, bh->b_size);
                        jh->b_frozen_data = NULL;
+                       jh->b_frozen_triggers = NULL;
                }
 
                spin_lock(&journal->j_list_lock);
                cp_transaction = jh->b_cp_transaction;
                if (cp_transaction) {
                        JBUFFER_TRACE(jh, "remove from old cp transaction");
+                       cp_transaction->t_chp_stats.cs_dropped++;
                        __jbd2_journal_remove_checkpoint(jh);
                }
 
@@ -841,6 +936,8 @@ restart_loop:
                if (buffer_jbddirty(bh)) {
                        JBUFFER_TRACE(jh, "add to new checkpointing trans");
                        __jbd2_journal_insert_checkpoint(jh, commit_transaction);
+                       if (is_journal_aborted(journal))
+                               clear_buffer_jbddirty(bh);
                        JBUFFER_TRACE(jh, "refile for checkpoint writeback");
                        __jbd2_journal_refile_buffer(jh);
                        jbd_unlock_bh_state(bh);
@@ -867,10 +964,10 @@ restart_loop:
        }
        spin_unlock(&journal->j_list_lock);
        /*
-        * This is a bit sleazy.  We borrow j_list_lock to protect
-        * journal->j_committing_transaction in __jbd2_journal_remove_checkpoint.
-        * Really, __jbd2_journal_remove_checkpoint should be using j_state_lock but
-        * it's a bit hassle to hold that across __jbd2_journal_remove_checkpoint
+        * This is a bit sleazy.  We use j_list_lock to protect transition
+        * of a transaction into T_FINISHED state and calling
+        * __jbd2_journal_drop_transaction(). Otherwise we could race with
+        * other checkpointing code processing the transaction...
         */
        spin_lock(&journal->j_state_lock);
        spin_lock(&journal->j_list_lock);
@@ -886,18 +983,58 @@ restart_loop:
 
        /* Done with this transaction! */
 
-       jbd_debug(3, "JBD: commit phase 8\n");
+       jbd_debug(3, "JBD: commit phase 7\n");
 
        J_ASSERT(commit_transaction->t_state == T_COMMIT);
 
+       commit_transaction->t_start = jiffies;
+       stats.run.rs_logging = jbd2_time_diff(stats.run.rs_logging,
+                                             commit_transaction->t_start);
+
+       /*
+        * File the transaction statistics
+        */
+       stats.ts_tid = commit_transaction->t_tid;
+       stats.run.rs_handle_count = commit_transaction->t_handle_count;
+       trace_jbd2_run_stats(journal->j_fs_dev->bd_dev,
+                            commit_transaction->t_tid, &stats.run);
+
+       /*
+        * Calculate overall stats
+        */
+       spin_lock(&journal->j_history_lock);
+       journal->j_stats.ts_tid++;
+       journal->j_stats.run.rs_wait += stats.run.rs_wait;
+       journal->j_stats.run.rs_running += stats.run.rs_running;
+       journal->j_stats.run.rs_locked += stats.run.rs_locked;
+       journal->j_stats.run.rs_flushing += stats.run.rs_flushing;
+       journal->j_stats.run.rs_logging += stats.run.rs_logging;
+       journal->j_stats.run.rs_handle_count += stats.run.rs_handle_count;
+       journal->j_stats.run.rs_blocks += stats.run.rs_blocks;
+       journal->j_stats.run.rs_blocks_logged += stats.run.rs_blocks_logged;
+       spin_unlock(&journal->j_history_lock);
+
        commit_transaction->t_state = T_FINISHED;
        J_ASSERT(commit_transaction == journal->j_committing_transaction);
        journal->j_commit_sequence = commit_transaction->t_tid;
        journal->j_committing_transaction = NULL;
+       commit_time = ktime_to_ns(ktime_sub(ktime_get(), start_time));
+
+       /*
+        * weight the commit time higher than the average time so we don't
+        * react too strongly to vast changes in the commit time
+        */
+       if (likely(journal->j_average_commit_time))
+               journal->j_average_commit_time = (commit_time +
+                               journal->j_average_commit_time*3) / 4;
+       else
+               journal->j_average_commit_time = commit_time;
        spin_unlock(&journal->j_state_lock);
 
-       if (commit_transaction->t_checkpoint_list == NULL) {
+       if (commit_transaction->t_checkpoint_list == NULL &&
+           commit_transaction->t_checkpoint_io_list == NULL) {
                __jbd2_journal_drop_transaction(journal, commit_transaction);
+               to_free = 1;
        } else {
                if (journal->j_checkpoint_transactions == NULL) {
                        journal->j_checkpoint_transactions = commit_transaction;
@@ -916,8 +1053,14 @@ restart_loop:
        }
        spin_unlock(&journal->j_list_lock);
 
+       if (journal->j_commit_callback)
+               journal->j_commit_callback(journal, commit_transaction);
+
+       trace_jbd2_end_commit(journal, commit_transaction);
        jbd_debug(1, "JBD: commit %d complete, head %d\n",
                  journal->j_commit_sequence, journal->j_tail_sequence);
+       if (to_free)
+               kfree(commit_transaction);
 
        wake_up(&journal->j_wait_done_commit);
 }