Illumos 6288 - dmu_buf_will_dirty could be faster

6288 dmu_buf_will_dirty could be faster
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Paul Dagnelie <pcd@delphix.com>
Reviewed by: Justin Gibbs <gibbs@scsiguy.com>
Reviewed by: Richard Elling <Richard.Elling@RichardElling.com>
Approved by: Robert Mustacchi <rm@joyent.com>

References:
  https://www.illumos.org/issues/6288
  https://github.com/illumos/illumos-gate/commit/0f2e7d0

Porting notes:
- [module/zfs/dbuf.c]
  - Fix 'warning: ISO C90 forbids mixed declarations and code'
    by moving 'dbuf_dirty_record_t *dr' to start of code block.

Ported-by: kernelOfTruth kerneloftruth@gmail.com
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>

diff --git a/module/zfs/dbuf.c b/module/zfs/dbuf.c
index d53d8607aba93748674bdc65043edd29c82d312b..e65f573e280ee49331f3c42ba1034ea158aea5a2 100644 (file)
--- a/module/zfs/dbuf.c
+++ b/module/zfs/dbuf.c
@@ -1182,6 +1182,32 @@ dbuf_release_bp(dmu_buf_impl_t *db)
        (void) arc_release(db->db_buf, db);
 }
 
+/*
+ * We already have a dirty record for this TXG, and we are being
+ * dirtied again.
+ */
+static void
+dbuf_redirty(dbuf_dirty_record_t *dr)
+{
+       dmu_buf_impl_t *db = dr->dr_dbuf;
+
+       ASSERT(MUTEX_HELD(&db->db_mtx));
+
+       if (db->db_level == 0 && db->db_blkid != DMU_BONUS_BLKID) {
+               /*
+                * If this buffer has already been written out,
+                * we now need to reset its state.
+                */
+               dbuf_unoverride(dr);
+               if (db->db.db_object != DMU_META_DNODE_OBJECT &&
+                   db->db_state != DB_NOFILL) {
+                       /* Already released on initial dirty, so just thaw. */
+                       ASSERT(arc_released(db->db_buf));
+                       arc_buf_thaw(db->db_buf);
+               }
+       }
+}
+
 dbuf_dirty_record_t *
 dbuf_dirty(dmu_buf_impl_t *db, dmu_tx_t *tx)
 {
@@ -1254,16 +1280,7 @@ dbuf_dirty(dmu_buf_impl_t *db, dmu_tx_t *tx)
        if (dr && dr->dr_txg == tx->tx_txg) {
                DB_DNODE_EXIT(db);
 
-               if (db->db_level == 0 && db->db_blkid != DMU_BONUS_BLKID) {
-                       /*
-                        * If this buffer has already been written out,
-                        * we now need to reset its state.
-                        */
-                       dbuf_unoverride(dr);
-                       if (db->db.db_object != DMU_META_DNODE_OBJECT &&
-                           db->db_state != DB_NOFILL)
-                               arc_buf_thaw(db->db_buf);
-               }
+               dbuf_redirty(dr);
                mutex_exit(&db->db_mtx);
                return (dr);
        }
@@ -1564,10 +1581,35 @@ dmu_buf_will_dirty(dmu_buf_t *db_fake, dmu_tx_t *tx)
 {
        dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake;
        int rf = DB_RF_MUST_SUCCEED | DB_RF_NOPREFETCH;
+       dbuf_dirty_record_t *dr;
 
        ASSERT(tx->tx_txg != 0);
        ASSERT(!refcount_is_zero(&db->db_holds));
 
+       /*
+        * Quick check for dirtyness.  For already dirty blocks, this
+        * reduces runtime of this function by >90%, and overall performance
+        * by 50% for some workloads (e.g. file deletion with indirect blocks
+        * cached).
+        */
+       mutex_enter(&db->db_mtx);
+
+       for (dr = db->db_last_dirty;
+           dr != NULL && dr->dr_txg >= tx->tx_txg; dr = dr->dr_next) {
+               /*
+                * It's possible that it is already dirty but not cached,
+                * because there are some calls to dbuf_dirty() that don't
+                * go through dmu_buf_will_dirty().
+                */
+               if (dr->dr_txg == tx->tx_txg && db->db_state == DB_CACHED) {
+                       /* This dbuf is already dirty and cached. */
+                       dbuf_redirty(dr);
+                       mutex_exit(&db->db_mtx);
+                       return;
+               }
+       }
+       mutex_exit(&db->db_mtx);
+
        DB_DNODE_ENTER(db);
        if (RW_WRITE_HELD(&DB_DNODE(db)->dn_struct_rwlock))
                rf |= DB_RF_HAVESTRUCT;