Refcounted DSL Crypto Key Mappings

Since native ZFS encryption was merged, we have been fighting
against a series of bugs that come down to the same problem: Key
mappings (which must be present during all I/O operations) are
created and destroyed based on dataset ownership, but I/Os can
have traditionally been allowed to "leak" into the next txg after
the dataset is disowned.

In the past we have attempted to solve this problem by trying to
ensure that datasets are disowned ater all I/O is finished by
calling txg_wait_synced(), but we have repeatedly found edge cases
that need to be squashed and code paths that might incur a high
number of txg syncs. This patch attempts to resolve this issue
differently, by adding a reference to the key mapping for each txg
it is dirtied in. By doing so, we can remove many of the
unnecessary calls to txg_wait_synced() we have added in the past
and ensure we don't need to deal with this problem in the future.

Reviewed-by: Jorgen Lundman <[email protected]>
Reviewed by: Matthew Ahrens <[email protected]>
Reviewed-by: Brian Behlendorf <[email protected]>
Signed-off-by: Tom Caputi <[email protected]>
Closes #7949 

1 files changed, 25 insertions, 2 deletions

diff --git a/module/zfs/dsl_pool.c b/module/zfs/dsl_pool.c
index e8f519b18..bd2055710 100644
--- a/module/zfs/dsl_pool.c
+++ b/module/zfs/dsl_pool.c
@@ -516,7 +516,8 @@ dsl_pool_create(spa_t *spa, nvlist_t *zplprops, dsl_crypto_params_t *dcp,
 	obj = dsl_dataset_create_sync_dd(dp->dp_root_dir, NULL, dcp, 0, tx);
 
 	/* create the root objset */
-	VERIFY0(dsl_dataset_hold_obj(dp, obj, FTAG, &ds));
+	VERIFY0(dsl_dataset_hold_obj_flags(dp, obj,
+	    DS_HOLD_FLAG_DECRYPT, FTAG, &ds));
 #ifdef _KERNEL
 	{
 		objset_t *os;
@@ -527,7 +528,7 @@ dsl_pool_create(spa_t *spa, nvlist_t *zplprops, dsl_crypto_params_t *dcp,
 		zfs_create_fs(os, kcred, zplprops, tx);
 	}
 #endif
-	dsl_dataset_rele(ds, FTAG);
+	dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG);
 
 	dmu_tx_commit(tx);
 
@@ -690,9 +691,22 @@ dsl_pool_sync(dsl_pool_t *dp, uint64_t txg)
 	 */
 	zio = zio_root(dp->dp_spa, NULL, NULL, ZIO_FLAG_MUSTSUCCEED);
 	while ((ds = txg_list_remove(&dp->dp_dirty_datasets, txg)) != NULL) {
+		objset_t *os = ds->ds_objset;
+
 		ASSERT(list_link_active(&ds->ds_synced_link));
 		dmu_buf_rele(ds->ds_dbuf, ds);
 		dsl_dataset_sync(ds, zio, tx);
+
+		/*
+		 * Release any key mappings created by calls to
+		 * dsl_dataset_dirty() from the userquota accounting
+		 * code paths.
+		 */
+		if (os->os_encrypted && !os->os_raw_receive &&
+		    !os->os_next_write_raw[txg & TXG_MASK]) {
+			ASSERT3P(ds->ds_key_mapping, !=, NULL);
+			key_mapping_rele(dp->dp_spa, ds->ds_key_mapping, ds);
+		}
 	}
 	VERIFY0(zio_wait(zio));
 
@@ -702,8 +716,17 @@ dsl_pool_sync(dsl_pool_t *dp, uint64_t txg)
 	 *
 	 *  - move dead blocks from the pending deadlist to the on-disk deadlist
 	 *  - release hold from dsl_dataset_dirty()
+	 *  - release key mapping hold from dsl_dataset_dirty()
 	 */
 	while ((ds = list_remove_head(&synced_datasets)) != NULL) {
+		objset_t *os = ds->ds_objset;
+
+		if (os->os_encrypted && !os->os_raw_receive &&
+		    !os->os_next_write_raw[txg & TXG_MASK]) {
+			ASSERT3P(ds->ds_key_mapping, !=, NULL);
+			key_mapping_rele(dp->dp_spa, ds->ds_key_mapping, ds);
+		}
+
 		dsl_dataset_sync_done(ds, tx);
 	}