1 files changed, 93 insertions, 14 deletions

diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c
index 39b7805a599a..46b4e347ed7d 100644
--- a/fs/jbd2/transaction.c
+++ b/fs/jbd2/transaction.c
@@ -25,6 +25,7 @@
 #include <linux/timer.h>
 #include <linux/mm.h>
 #include <linux/highmem.h>
+#include <linux/hrtimer.h>
 
 static void __jbd2_journal_temp_unlink_buffer(struct journal_head *jh);
 
@@ -48,6 +49,7 @@ jbd2_get_transaction(journal_t *journal, transaction_t *transaction)
 {
         transaction->t_journal = journal;
         transaction->t_state = T_RUNNING;
+        transaction->t_start_time = ktime_get();
         transaction->t_tid = journal->j_transaction_sequence++;
         transaction->t_expires = jiffies + journal->j_commit_interval;
         spin_lock_init(&transaction->t_handle_lock);
@@ -741,6 +743,12 @@ done:
                 source = kmap_atomic(page, KM_USER0);
                 memcpy(jh->b_frozen_data, source+offset, jh2bh(jh)->b_size);
                 kunmap_atomic(source, KM_USER0);
+
+                /*
+                 * Now that the frozen data is saved off, we need to store
+                 * any matching triggers.
+                 */
+                jh->b_frozen_triggers = jh->b_triggers;
         }
         jbd_unlock_bh_state(bh);
 
@@ -944,6 +952,47 @@ out:
 }
 
 /**
+ * void jbd2_journal_set_triggers() - Add triggers for commit writeout
+ * @bh: buffer to trigger on
+ * @type: struct jbd2_buffer_trigger_type containing the trigger(s).
+ *
+ * Set any triggers on this journal_head.  This is always safe, because
+ * triggers for a committing buffer will be saved off, and triggers for
+ * a running transaction will match the buffer in that transaction.
+ *
+ * Call with NULL to clear the triggers.
+ */
+void jbd2_journal_set_triggers(struct buffer_head *bh,
+                               struct jbd2_buffer_trigger_type *type)
+{
+        struct journal_head *jh = bh2jh(bh);
+
+        jh->b_triggers = type;
+}
+
+void jbd2_buffer_commit_trigger(struct journal_head *jh, void *mapped_data,
+                                struct jbd2_buffer_trigger_type *triggers)
+{
+        struct buffer_head *bh = jh2bh(jh);
+
+        if (!triggers || !triggers->t_commit)
+                return;
+
+        triggers->t_commit(triggers, bh, mapped_data, bh->b_size);
+}
+
+void jbd2_buffer_abort_trigger(struct journal_head *jh,
+                               struct jbd2_buffer_trigger_type *triggers)
+{
+        if (!triggers || !triggers->t_abort)
+                return;
+
+        triggers->t_abort(triggers, jh2bh(jh));
+}
+
+
+
+/**
  * int jbd2_journal_dirty_metadata() -  mark a buffer as containing dirty metadata
  * @handle: transaction to add buffer to.
  * @bh: buffer to mark
@@ -1193,7 +1242,7 @@ int jbd2_journal_stop(handle_t *handle)
 {
         transaction_t *transaction = handle->h_transaction;
         journal_t *journal = transaction->t_journal;
-        int old_handle_count, err;
+        int err;
         pid_t pid;
 
         J_ASSERT(journal_current_handle() == handle);
@@ -1216,24 +1265,54 @@ int jbd2_journal_stop(handle_t *handle)
         /*
          * Implement synchronous transaction batching.  If the handle
          * was synchronous, don't force a commit immediately.  Let's
-         * yield and let another thread piggyback onto this transaction.
-         * Keep doing that while new threads continue to arrive.
-         * It doesn't cost much - we're about to run a commit and sleep
-         * on IO anyway.  Speeds up many-threaded, many-dir operations
-         * by 30x or more...
+         * yield and let another thread piggyback onto this
+         * transaction.  Keep doing that while new threads continue to
+         * arrive.  It doesn't cost much - we're about to run a commit
+         * and sleep on IO anyway.  Speeds up many-threaded, many-dir
+         * operations by 30x or more...
          *
-         * But don't do this if this process was the most recent one to
-         * perform a synchronous write.  We do this to detect the case where a
-         * single process is doing a stream of sync writes.  No point in waiting
-         * for joiners in that case.
+         * We try and optimize the sleep time against what the
+         * underlying disk can do, instead of having a static sleep
+         * time.  This is useful for the case where our storage is so
+         * fast that it is more optimal to go ahead and force a flush
+         * and wait for the transaction to be committed than it is to
+         * wait for an arbitrary amount of time for new writers to
+         * join the transaction.  We achieve this by measuring how
+         * long it takes to commit a transaction, and compare it with
+         * how long this transaction has been running, and if run time
+         * < commit time then we sleep for the delta and commit.  This
+         * greatly helps super fast disks that would see slowdowns as
+         * more threads started doing fsyncs.
+         *
+         * But don't do this if this process was the most recent one
+         * to perform a synchronous write.  We do this to detect the
+         * case where a single process is doing a stream of sync
+         * writes.  No point in waiting for joiners in that case.
          */
         pid = current->pid;
         if (handle->h_sync && journal->j_last_sync_writer != pid) {
+                u64 commit_time, trans_time;
+
                 journal->j_last_sync_writer = pid;
-                do {
-                        old_handle_count = transaction->t_handle_count;
-                        schedule_timeout_uninterruptible(1);
-                } while (old_handle_count != transaction->t_handle_count);
+
+                spin_lock(&journal->j_state_lock);
+                commit_time = journal->j_average_commit_time;
+                spin_unlock(&journal->j_state_lock);
+
+                trans_time = ktime_to_ns(ktime_sub(ktime_get(),
+                                                   transaction->t_start_time));
+
+                commit_time = max_t(u64, commit_time,
+                                    1000*journal->j_min_batch_time);
+                commit_time = min_t(u64, commit_time,
+                                    1000*journal->j_max_batch_time);
+
+                if (trans_time < commit_time) {
+                        ktime_t expires = ktime_add_ns(ktime_get(),
+                                                       commit_time);
+                        set_current_state(TASK_UNINTERRUPTIBLE);
+                        schedule_hrtimeout(&expires, HRTIMER_MODE_ABS);
+                }
         }
 
         current->journal_info = NULL;