OpenZFS 6513 - partially filled holes lose birth time

pcd1193182 · behlendorf · commit bc77ba73fec8 · 2016-06-21T10:55:13.000-07:00
Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Boris Protopopov <bprotopopov@hotmail.com> Approved by: Richard Lowe <richlowe@richlowe.net>a Ported by: Boris Protopopov <bprotopopov@actifio.com> Signed-off-by: Boris Protopopov <bprotopopov@actifio.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> OpenZFS-issue: https://www.illumos.org/issues/6513 OpenZFS-commit: openzfs/openzfs@8df0bcf0 If a ZFS object contains a hole at level one, and then a data block is created at level 0 underneath that l1 block, l0 holes will be created. However, these l0 holes do not have the birth time property set; as a result, incremental sends will not send those holes. Fix is to modify the dbuf_read code to fill in birth time data.
diff --git a/include/sys/arc.h b/include/sys/arc.h
@@ -20,7 +20,7 @@
  */
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2012, 2014 by Delphix. All rights reserved.
+ * Copyright (c) 2012, 2016 by Delphix. All rights reserved.
  * Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
  */
 
@@ -193,9 +193,11 @@ int arc_read(zio_t *pio, spa_t *spa, const blkptr_t *bp,
     arc_flags_t *arc_flags, const zbookmark_phys_t *zb);
 zio_t *arc_write(zio_t *pio, spa_t *spa, uint64_t txg,
     blkptr_t *bp, arc_buf_t *buf, boolean_t l2arc, boolean_t l2arc_compress,
-    const zio_prop_t *zp, arc_done_func_t *ready, arc_done_func_t *physdone,
-    arc_done_func_t *done, void *private, zio_priority_t priority,
-    int zio_flags, const zbookmark_phys_t *zb);
+    const zio_prop_t *zp,
+    arc_done_func_t *ready, arc_done_func_t *child_ready,
+    arc_done_func_t *physdone, arc_done_func_t *done,
+    void *private, zio_priority_t priority, int zio_flags,
+    const zbookmark_phys_t *zb);
 
 arc_prune_t *arc_add_prune_callback(arc_prune_func_t *func, void *private);
 void arc_remove_prune_callback(arc_prune_t *p);
diff --git a/include/sys/arc_impl.h b/include/sys/arc_impl.h
@@ -101,6 +101,7 @@ typedef struct arc_write_callback arc_write_callback_t;
 struct arc_write_callback {
 	void		*awcb_private;
 	arc_done_func_t	*awcb_ready;
+	arc_done_func_t	*awcb_children_ready;
 	arc_done_func_t	*awcb_physdone;
 	arc_done_func_t	*awcb_done;
 	arc_buf_t	*awcb_buf;
diff --git a/include/sys/zio.h b/include/sys/zio.h
@@ -22,7 +22,7 @@
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
  * Copyright 2011 Nexenta Systems, Inc. All rights reserved.
- * Copyright (c) 2012, 2014 by Delphix. All rights reserved.
+ * Copyright (c) 2012, 2016 by Delphix. All rights reserved.
  * Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
  */
 
@@ -400,7 +400,8 @@ struct zio {
 	zio_transform_t *io_transform_stack;
 
 	/* Callback info */
-	zio_done_func_t *io_ready;
+	zio_done_func_t	*io_ready;
+	zio_done_func_t	*io_children_ready;
 	zio_done_func_t	*io_physdone;
 	zio_done_func_t	*io_done;
 	void		*io_private;
@@ -468,9 +469,10 @@ extern zio_t *zio_read(zio_t *pio, spa_t *spa, const blkptr_t *bp, void *data,
 
 extern zio_t *zio_write(zio_t *pio, spa_t *spa, uint64_t txg, blkptr_t *bp,
     void *data, uint64_t size, const zio_prop_t *zp,
-    zio_done_func_t *ready, zio_done_func_t *physdone, zio_done_func_t *done,
-    void *private,
-    zio_priority_t priority, enum zio_flag flags, const zbookmark_phys_t *zb);
+    zio_done_func_t *ready, zio_done_func_t *children_ready,
+    zio_done_func_t *physdone, zio_done_func_t *done,
+    void *private, zio_priority_t priority, enum zio_flag flags,
+    const zbookmark_phys_t *zb);
 
 extern zio_t *zio_rewrite(zio_t *pio, spa_t *spa, uint64_t txg, blkptr_t *bp,
     void *data, uint64_t size, zio_done_func_t *done, void *private,
diff --git a/module/zfs/arc.c b/module/zfs/arc.c
@@ -21,7 +21,7 @@
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
  * Copyright (c) 2012, Joyent, Inc. All rights reserved.
- * Copyright (c) 2011, 2015 by Delphix. All rights reserved.
+ * Copyright (c) 2011, 2016 by Delphix. All rights reserved.
  * Copyright (c) 2014 by Saso Kiselkov. All rights reserved.
  * Copyright 2014 Nexenta Systems, Inc.  All rights reserved.
  */
@@ -4981,6 +4981,15 @@ arc_write_ready(zio_t *zio)
 	hdr->b_flags |= ARC_FLAG_IO_IN_PROGRESS;
 }
 
+static void
+arc_write_children_ready(zio_t *zio)
+{
+	arc_write_callback_t *callback = zio->io_private;
+	arc_buf_t *buf = callback->awcb_buf;
+
+	callback->awcb_children_ready(zio, buf, callback->awcb_private);
+}
+
 /*
  * The SPA calls this callback for each physical write that happens on behalf
  * of a logical write.  See the comment in dbuf_write_physdone() for details.
@@ -5077,7 +5086,8 @@ arc_write_done(zio_t *zio)
 zio_t *
 arc_write(zio_t *pio, spa_t *spa, uint64_t txg,
     blkptr_t *bp, arc_buf_t *buf, boolean_t l2arc, boolean_t l2arc_compress,
-    const zio_prop_t *zp, arc_done_func_t *ready, arc_done_func_t *physdone,
+    const zio_prop_t *zp, arc_done_func_t *ready,
+    arc_done_func_t *children_ready, arc_done_func_t *physdone,
     arc_done_func_t *done, void *private, zio_priority_t priority,
     int zio_flags, const zbookmark_phys_t *zb)
 {
@@ -5097,13 +5107,16 @@ arc_write(zio_t *pio, spa_t *spa, uint64_t txg,
 		hdr->b_flags |= ARC_FLAG_L2COMPRESS;
 	callback = kmem_zalloc(sizeof (arc_write_callback_t), KM_SLEEP);
 	callback->awcb_ready = ready;
+	callback->awcb_children_ready = children_ready;
 	callback->awcb_physdone = physdone;
 	callback->awcb_done = done;
 	callback->awcb_private = private;
 	callback->awcb_buf = buf;
 
 	zio = zio_write(pio, spa, txg, bp, buf->b_data, hdr->b_size, zp,
-	    arc_write_ready, arc_write_physdone, arc_write_done, callback,
+	    arc_write_ready,
+	    (children_ready != NULL) ? arc_write_children_ready : NULL,
+	    arc_write_physdone, arc_write_done, callback,
 	    priority, zio_flags, zb);
 
 	return (zio);
diff --git a/module/zfs/dbuf.c b/module/zfs/dbuf.c
@@ -543,13 +543,50 @@ dbuf_verify(dmu_buf_impl_t *db)
 		 * If the blkptr isn't set but they have nonzero data,
 		 * it had better be dirty, otherwise we'll lose that
 		 * data when we evict this buffer.
+		 *
+		 * There is an exception to this rule for indirect blocks; in
+		 * this case, if the indirect block is a hole, we fill in a few
+		 * fields on each of the child blocks (importantly, birth time)
+		 * to prevent hole birth times from being lost when you
+		 * partially fill in a hole.
 		 */
 		if (db->db_dirtycnt == 0) {
-			ASSERTV(uint64_t *buf = db->db.db_data);
-			int i;
+			if (db->db_level == 0) {
+				uint64_t *buf = db->db.db_data;
+				int i;
 
-			for (i = 0; i < db->db.db_size >> 3; i++) {
-				ASSERT(buf[i] == 0);
+				for (i = 0; i < db->db.db_size >> 3; i++) {
+					ASSERT(buf[i] == 0);
+				}
+			} else {
+				int i;
+				blkptr_t *bps = db->db.db_data;
+				ASSERT3U(1 << DB_DNODE(db)->dn_indblkshift, ==,
+				    db->db.db_size);
+				/*
+				 * We want to verify that all the blkptrs in the
+				 * indirect block are holes, but we may have
+				 * automatically set up a few fields for them.
+				 * We iterate through each blkptr and verify
+				 * they only have those fields set.
+				 */
+				for (i = 0;
+				    i < db->db.db_size / sizeof (blkptr_t);
+				    i++) {
+					blkptr_t *bp = &bps[i];
+					ASSERT(ZIO_CHECKSUM_IS_ZERO(
+					    &bp->blk_cksum));
+					ASSERT(
+					    DVA_IS_EMPTY(&bp->blk_dva[0]) &&
+					    DVA_IS_EMPTY(&bp->blk_dva[1]) &&
+					    DVA_IS_EMPTY(&bp->blk_dva[2]));
+					ASSERT0(bp->blk_fill);
+					ASSERT0(bp->blk_pad[0]);
+					ASSERT0(bp->blk_pad[1]);
+					ASSERT(!BP_IS_EMBEDDED(bp));
+					ASSERT(BP_IS_HOLE(bp));
+					ASSERT0(bp->blk_phys_birth);
+				}
 			}
 		}
 	}
@@ -718,10 +755,32 @@ dbuf_read_impl(dmu_buf_impl_t *db, zio_t *zio, uint32_t flags)
 	    BP_IS_HOLE(db->db_blkptr)))) {
 		arc_buf_contents_t type = DBUF_GET_BUFC_TYPE(db);
 
-		DB_DNODE_EXIT(db);
 		dbuf_set_data(db, arc_buf_alloc(db->db_objset->os_spa,
 		    db->db.db_size, db, type));
 		bzero(db->db.db_data, db->db.db_size);
+
+		if (db->db_blkptr != NULL && db->db_level > 0 &&
+		    BP_IS_HOLE(db->db_blkptr) &&
+		    db->db_blkptr->blk_birth != 0) {
+			blkptr_t *bps = db->db.db_data;
+			int i;
+			for (i = 0; i < ((1 <<
+			    DB_DNODE(db)->dn_indblkshift) / sizeof (blkptr_t));
+			    i++) {
+				blkptr_t *bp = &bps[i];
+				ASSERT3U(BP_GET_LSIZE(db->db_blkptr), ==,
+				    1 << dn->dn_indblkshift);
+				BP_SET_LSIZE(bp,
+				    BP_GET_LEVEL(db->db_blkptr) == 1 ?
+				    dn->dn_datablksz :
+				    BP_GET_LSIZE(db->db_blkptr));
+				BP_SET_TYPE(bp, BP_GET_TYPE(db->db_blkptr));
+				BP_SET_LEVEL(bp,
+				    BP_GET_LEVEL(db->db_blkptr) - 1);
+				BP_SET_BIRTH(bp, db->db_blkptr->blk_birth, 0);
+			}
+		}
+		DB_DNODE_EXIT(db);
 		db->db_state = DB_CACHED;
 		mutex_exit(&db->db_mtx);
 		return (0);
@@ -3094,6 +3153,45 @@ dbuf_write_ready(zio_t *zio, arc_buf_t *buf, void *vdb)
 	rw_exit(&dn->dn_struct_rwlock);
 }
 
+/* ARGSUSED */
+/*
+ * This function gets called just prior to running through the compression
+ * stage of the zio pipeline. If we're an indirect block comprised of only
+ * holes, then we want this indirect to be compressed away to a hole. In
+ * order to do that we must zero out any information about the holes that
+ * this indirect points to prior to before we try to compress it.
+ */
+static void
+dbuf_write_children_ready(zio_t *zio, arc_buf_t *buf, void *vdb)
+{
+	dmu_buf_impl_t *db = vdb;
+	dnode_t *dn;
+	blkptr_t *bp;
+	uint64_t i;
+	int epbs;
+
+	ASSERT3U(db->db_level, >, 0);
+	DB_DNODE_ENTER(db);
+	dn = DB_DNODE(db);
+	epbs = dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT;
+
+	/* Determine if all our children are holes */
+	for (i = 0, bp = db->db.db_data; i < 1 << epbs; i++, bp++) {
+		if (!BP_IS_HOLE(bp))
+			break;
+	}
+
+	/*
+	 * If all the children are holes, then zero them all out so that
+	 * we may get compressed away.
+	 */
+	if (i == 1 << epbs) {
+		/* didn't find any non-holes */
+		bzero(db->db.db_data, db->db.db_size);
+	}
+	DB_DNODE_EXIT(db);
+}
+
 /*
  * The SPA will call this callback several times for each zio - once
  * for every physical child i/o (zio->io_phys_children times).  This
@@ -3348,7 +3446,8 @@ dbuf_write(dbuf_dirty_record_t *dr, arc_buf_t *data, dmu_tx_t *tx)
 
 		dr->dr_zio = zio_write(zio, os->os_spa, txg,
 		    &dr->dr_bp_copy, contents, db->db.db_size, &zp,
-		    dbuf_write_override_ready, NULL, dbuf_write_override_done,
+		    dbuf_write_override_ready, NULL, NULL,
+		    dbuf_write_override_done,
 		    dr, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED, &zb);
 		mutex_enter(&db->db_mtx);
 		dr->dt.dl.dr_override_state = DR_NOT_OVERRIDDEN;
@@ -3359,14 +3458,26 @@ dbuf_write(dbuf_dirty_record_t *dr, arc_buf_t *data, dmu_tx_t *tx)
 		ASSERT(zp.zp_checksum == ZIO_CHECKSUM_OFF);
 		dr->dr_zio = zio_write(zio, os->os_spa, txg,
 		    &dr->dr_bp_copy, NULL, db->db.db_size, &zp,
-		    dbuf_write_nofill_ready, NULL, dbuf_write_nofill_done, db,
+		    dbuf_write_nofill_ready, NULL, NULL,
+		    dbuf_write_nofill_done, db,
 		    ZIO_PRIORITY_ASYNC_WRITE,
 		    ZIO_FLAG_MUSTSUCCEED | ZIO_FLAG_NODATA, &zb);
 	} else {
+		arc_done_func_t *children_ready_cb = NULL;
 		ASSERT(arc_released(data));
+
+		/*
+		 * For indirect blocks, we want to setup the children
+		 * ready callback so that we can properly handle an indirect
+		 * block that only contains holes.
+		 */
+		if (db->db_level != 0)
+			children_ready_cb = dbuf_write_children_ready;
+
 		dr->dr_zio = arc_write(zio, os->os_spa, txg,
 		    &dr->dr_bp_copy, data, DBUF_IS_L2CACHEABLE(db),
 		    DBUF_IS_L2COMPRESSIBLE(db), &zp, dbuf_write_ready,
+		    children_ready_cb,
 		    dbuf_write_physdone, dbuf_write_done, db,
 		    ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED, &zb);
 	}
diff --git a/module/zfs/dmu.c b/module/zfs/dmu.c
@@ -20,8 +20,9 @@
  */
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2011, 2015 by Delphix. All rights reserved.
+ * Copyright (c) 2011, 2016 by Delphix. All rights reserved.
  * Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
+ * Copyright (c) 2013, Joyent, Inc. All rights reserved.
  * Copyright (c) 2014, Nexenta Systems, Inc. All rights reserved.
  * Copyright (c) 2015 by Chunwei Chen. All rights reserved.
  */
@@ -1480,10 +1481,11 @@ dmu_sync_late_arrival(zio_t *pio, objset_t *os, dmu_sync_cb_t *done, zgd_t *zgd,
 	dsa->dsa_zgd = zgd;
 	dsa->dsa_tx = tx;
 
-	zio_nowait(zio_write(pio, os->os_spa, dmu_tx_get_txg(tx), zgd->zgd_bp,
-	    zgd->zgd_db->db_data, zgd->zgd_db->db_size, zp,
-	    dmu_sync_late_arrival_ready, NULL, dmu_sync_late_arrival_done, dsa,
-	    ZIO_PRIORITY_SYNC_WRITE, ZIO_FLAG_CANFAIL|ZIO_FLAG_FASTWRITE, zb));
+	zio_nowait(zio_write(pio, os->os_spa, dmu_tx_get_txg(tx),
+	    zgd->zgd_bp, zgd->zgd_db->db_data, zgd->zgd_db->db_size,
+	    zp, dmu_sync_late_arrival_ready, NULL,
+	    NULL, dmu_sync_late_arrival_done, dsa, ZIO_PRIORITY_SYNC_WRITE,
+	    ZIO_FLAG_CANFAIL, zb));
 
 	return (0);
 }
@@ -1636,8 +1638,8 @@ dmu_sync(zio_t *pio, uint64_t txg, dmu_sync_cb_t *done, zgd_t *zgd)
 	zio_nowait(arc_write(pio, os->os_spa, txg,
 	    bp, dr->dt.dl.dr_data, DBUF_IS_L2CACHEABLE(db),
 	    DBUF_IS_L2COMPRESSIBLE(db), &zp, dmu_sync_ready,
-	    NULL, dmu_sync_done, dsa, ZIO_PRIORITY_SYNC_WRITE,
-	    ZIO_FLAG_CANFAIL, &zb));
+	    NULL, NULL, dmu_sync_done, dsa,
+	    ZIO_PRIORITY_SYNC_WRITE, ZIO_FLAG_CANFAIL, &zb));
 
 	return (0);
 }
diff --git a/module/zfs/dmu_objset.c b/module/zfs/dmu_objset.c
@@ -20,7 +20,7 @@
  */
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2012, 2014 by Delphix. All rights reserved.
+ * Copyright (c) 2012, 2016 by Delphix. All rights reserved.
  * Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
  * Copyright (c) 2013, Joyent, Inc. All rights reserved.
  * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved.
@@ -1110,9 +1110,9 @@ dmu_objset_sync(objset_t *os, zio_t *pio, dmu_tx_t *tx)
 
 	zio = arc_write(pio, os->os_spa, tx->tx_txg,
 	    os->os_rootbp, os->os_phys_buf, DMU_OS_IS_L2CACHEABLE(os),
-	    DMU_OS_IS_L2COMPRESSIBLE(os), &zp, dmu_objset_write_ready,
-	    NULL, dmu_objset_write_done, os, ZIO_PRIORITY_ASYNC_WRITE,
-	    ZIO_FLAG_MUSTSUCCEED, &zb);
+	    DMU_OS_IS_L2COMPRESSIBLE(os),
+	    &zp, dmu_objset_write_ready, NULL, NULL, dmu_objset_write_done,
+	    os, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED, &zb);
 
 	/*
 	 * Sync special dnodes - the parent IO for the sync is the root block
diff --git a/module/zfs/dnode_sync.c b/module/zfs/dnode_sync.c
@@ -21,7 +21,7 @@
 
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2012, 2015 by Delphix. All rights reserved.
+ * Copyright (c) 2012, 2016 by Delphix. All rights reserved.
  * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved.
  */
 
@@ -60,20 +60,14 @@ dnode_increase_indirection(dnode_t *dn, dmu_tx_t *tx)
 	dprintf("os=%p obj=%llu, increase to %d\n", dn->dn_objset,
 	    dn->dn_object, dn->dn_phys->dn_nlevels);
 
-	/* check for existing blkptrs in the dnode */
-	for (i = 0; i < nblkptr; i++)
-		if (!BP_IS_HOLE(&dn->dn_phys->dn_blkptr[i]))
-			break;
-	if (i != nblkptr) {
-		/* transfer dnode's block pointers to new indirect block */
-		(void) dbuf_read(db, NULL, DB_RF_MUST_SUCCEED|DB_RF_HAVESTRUCT);
-		ASSERT(db->db.db_data);
-		ASSERT(arc_released(db->db_buf));
-		ASSERT3U(sizeof (blkptr_t) * nblkptr, <=, db->db.db_size);
-		bcopy(dn->dn_phys->dn_blkptr, db->db.db_data,
-		    sizeof (blkptr_t) * nblkptr);
-		arc_buf_freeze(db->db_buf);
-	}
+	/* transfer dnode's block pointers to new indirect block */
+	(void) dbuf_read(db, NULL, DB_RF_MUST_SUCCEED|DB_RF_HAVESTRUCT);
+	ASSERT(db->db.db_data);
+	ASSERT(arc_released(db->db_buf));
+	ASSERT3U(sizeof (blkptr_t) * nblkptr, <=, db->db.db_size);
+	bcopy(dn->dn_phys->dn_blkptr, db->db.db_data,
+	    sizeof (blkptr_t) * nblkptr);
+	arc_buf_freeze(db->db_buf);
 
 	/* set dbuf's parent pointers to new indirect buf */
 	for (i = 0; i < nblkptr; i++) {
diff --git a/module/zfs/zio.c b/module/zfs/zio.c
