Illumos #2882, #2883, #2900

2882 implement libzfs_core
2883 changing "canmount" property to "on" should not always remount dataset
2900 "zfs snapshot" should be able to create multiple, arbitrary snapshots at once

Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Chris Siden <christopher.siden@delphix.com>
Reviewed by: Garrett D'Amore <garrett@damore.org>
Reviewed by: Bill Pijewski <wdp@joyent.com>
Reviewed by: Dan Kruchinin <dan.kruchinin@gmail.com>
Approved by: Eric Schrock <Eric.Schrock@delphix.com>

References:
  https://www.illumos.org/issues/2882
  https://www.illumos.org/issues/2883
  https://www.illumos.org/issues/2900
  illumos/illumos-gate@4445fff

Ported-by: Tim Chase <tim@chase2k.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #1293

Porting notes:

WARNING: This patch changes the user/kernel ABI.  That means that
the zfs/zpool utilities built from master are NOT compatible with
the 0.6.2 kernel modules.  Ensure you load the matching kernel
modules from master after updating the utilities.  Otherwise the
zfs/zpool commands will be unable to interact with your pool and
you will see errors similar to the following:

  $ zpool list
  failed to read pool configuration: bad address
  no pools available

  $ zfs list
  no datasets available

Add zvol minor device creation to the new zfs_snapshot_nvl function.

Remove the logging of the "release" operation in
dsl_dataset_user_release_sync().  The logging caused a null dereference
because ds->ds_dir is zeroed in dsl_dataset_destroy_sync() and the
logging functions try to get the ds name via the dsl_dataset_name()
function. I've got no idea why this particular code would have worked
in Illumos.  This code has subsequently been completely reworked in
Illumos commit 3b2aab1 (3464 zfs synctask code needs restructuring).

Squash some "may be used uninitialized" warning/erorrs.

Fix some printf format warnings for %lld and %llu.

Apply a few spa_writeable() changes that were made to Illumos in
illumos/illumos-gate.git@cd1c8b8 as part of the 3112, 3113, 3114 and
3115 fixes.

Add a missing call to fnvlist_free(nvl) in log_internal() that was added
in Illumos to fix issue 3085 but couldn't be ported to ZoL at the time
(9e11c73) because it depended on future work.

cmd/zdb/zdb.c

@@ -57,6 +57,7 @@
 #include <sys/arc.h>
 #include <sys/ddt.h>
 #include <sys/zfeature.h>
+#include <zfs_comutil.h>
 #undef ZFS_MAXNAMELEN
 #include <libzfs.h>
 
@@ -209,6 +210,27 @@ dump_packed_nvlist(objset_t *os, uint64_t object, void *data, size_t size)
 	nvlist_free(nv);
 }
 
+/* ARGSUSED */
+static void
+dump_history_offsets(objset_t *os, uint64_t object, void *data, size_t size)
+{
+	spa_history_phys_t *shp = data;
+
+	if (shp == NULL)
+		return;
+
+	(void) printf("\t\tpool_create_len = %llu\n",
+	    (u_longlong_t)shp->sh_pool_create_len);
+	(void) printf("\t\tphys_max_off = %llu\n",
+	    (u_longlong_t)shp->sh_phys_max_off);
+	(void) printf("\t\tbof = %llu\n",
+	    (u_longlong_t)shp->sh_bof);
+	(void) printf("\t\teof = %llu\n",
+	    (u_longlong_t)shp->sh_eof);
+	(void) printf("\t\trecords_lost = %llu\n",
+	    (u_longlong_t)shp->sh_records_lost);
+}
+
 static void
 zdb_nicenum(uint64_t num, char *buf)
 {
@@ -869,21 +891,22 @@ dump_history(spa_t *spa)
 	for (i = 0; i < num; i++) {
 		uint64_t time, txg, ievent;
 		char *cmd, *intstr;
+		boolean_t printed = B_FALSE;
 
 		if (nvlist_lookup_uint64(events[i], ZPOOL_HIST_TIME,
 		    &time) != 0)
-			continue;
+			goto next;
 		if (nvlist_lookup_string(events[i], ZPOOL_HIST_CMD,
 		    &cmd) != 0) {
 			if (nvlist_lookup_uint64(events[i],
 			    ZPOOL_HIST_INT_EVENT, &ievent) != 0)
-				continue;
+				goto next;
 			verify(nvlist_lookup_uint64(events[i],
 			    ZPOOL_HIST_TXG, &txg) == 0);
 			verify(nvlist_lookup_string(events[i],
 			    ZPOOL_HIST_INT_STR, &intstr) == 0);
-			if (ievent >= LOG_END)
-				continue;
+			if (ievent >= ZFS_NUM_LEGACY_HISTORY_EVENTS)
+				goto next;
 
 			(void) snprintf(internalstr,
 			    sizeof (internalstr),
@@ -896,6 +919,14 @@ dump_history(spa_t *spa)
 		(void) localtime_r(&tsec, &t);
 		(void) strftime(tbuf, sizeof (tbuf), "%F.%T", &t);
 		(void) printf("%s %s\n", tbuf, cmd);
+		printed = B_TRUE;
+
+next:
+		if (dump_opt['h'] > 1) {
+			if (!printed)
+				(void) printf("unrecognized record:\n");
+			dump_nvlist(events[i], 2);
+		}
 	}
 }
 
@@ -1529,7 +1560,7 @@ static object_viewer_t *object_viewer[DMU_OT_NUMTYPES + 1] = {
 	dump_zap,		/* other ZAP			*/
 	dump_zap,		/* persistent error log		*/
 	dump_uint8,		/* SPA history			*/
-	dump_uint64,		/* SPA history offsets		*/
+	dump_history_offsets,	/* SPA history offsets		*/
 	dump_zap,		/* Pool properties		*/
 	dump_zap,		/* DSL permissions		*/
 	dump_acl,		/* ZFS ACL			*/

cmd/zfs/Makefile.am

@@ -16,6 +16,7 @@ zfs_LDADD = \
 	$(top_builddir)/lib/libnvpair/libnvpair.la \
 	$(top_builddir)/lib/libuutil/libuutil.la \
 	$(top_builddir)/lib/libzpool/libzpool.la \
-	$(top_builddir)/lib/libzfs/libzfs.la
+	$(top_builddir)/lib/libzfs/libzfs.la \
+	$(top_builddir)/lib/libzfs_core/libzfs_core.la
 
 zfs_LDFLAGS = -pthread -lm $(ZLIB) -lrt -ldl $(LIBUUID) $(LIBBLKID)

cmd/zfs/zfs_main.c

@@ -55,6 +55,7 @@
 #include <time.h>
 
 #include <libzfs.h>
+#include <libzfs_core.h>
 #include <zfs_prop.h>
 #include <zfs_deleg.h>
 #include <libuutil.h>
@@ -72,6 +73,7 @@ libzfs_handle_t *g_zfs;
 
 static FILE *mnttab_file;
 static char history_str[HIS_MAX_RECORD_LEN];
+static boolean_t log_history = B_TRUE;
 
 static int zfs_do_clone(int argc, char **argv);
 static int zfs_do_create(int argc, char **argv);
@@ -261,7 +263,7 @@ get_usage(zfs_help_t idx)
 		return (gettext("\tshare <-a | filesystem>\n"));
 	case HELP_SNAPSHOT:
 		return (gettext("\tsnapshot|snap [-r] [-o property=value] ... "
-		    "<filesystem@snapname|volume@snapname>\n"));
+		    "<filesystem@snapname|volume@snapname> ...\n"));
 	case HELP_UNMOUNT:
 		return (gettext("\tunmount [-f] "
 		    "<-a | filesystem|mountpoint>\n"));
@@ -892,9 +894,9 @@ typedef struct destroy_cbdata {
 	nvlist_t	*cb_nvl;
 
 	/* first snap in contiguous run */
-	zfs_handle_t	*cb_firstsnap;
+	char		*cb_firstsnap;
 	/* previous snap in contiguous run */
-	zfs_handle_t	*cb_prevsnap;
+	char		*cb_prevsnap;
 	int64_t		cb_snapused;
 	char		*cb_snapspec;
 } destroy_cbdata_t;
@@ -1008,11 +1010,13 @@ destroy_print_cb(zfs_handle_t *zhp, void *arg)
 
 	if (nvlist_exists(cb->cb_nvl, name)) {
 		if (cb->cb_firstsnap == NULL)
-			cb->cb_firstsnap = zfs_handle_dup(zhp);
+			cb->cb_firstsnap = strdup(name);
 		if (cb->cb_prevsnap != NULL)
-			zfs_close(cb->cb_prevsnap);
+			free(cb->cb_prevsnap);
 		/* this snap continues the current range */
-		cb->cb_prevsnap = zfs_handle_dup(zhp);
+		cb->cb_prevsnap = strdup(name);
+		if (cb->cb_firstsnap == NULL || cb->cb_prevsnap == NULL)
+			nomem();
 		if (cb->cb_verbose) {
 			if (cb->cb_parsable) {
 				(void) printf("destroy\t%s\n", name);
@@ -1027,12 +1031,12 @@ destroy_print_cb(zfs_handle_t *zhp, void *arg)
 	} else if (cb->cb_firstsnap != NULL) {
 		/* end of this range */
 		uint64_t used = 0;
-		err = zfs_get_snapused_int(cb->cb_firstsnap,
+		err = lzc_snaprange_space(cb->cb_firstsnap,
 		    cb->cb_prevsnap, &used);
 		cb->cb_snapused += used;
-		zfs_close(cb->cb_firstsnap);
+		free(cb->cb_firstsnap);
 		cb->cb_firstsnap = NULL;
-		zfs_close(cb->cb_prevsnap);
+		free(cb->cb_prevsnap);
 		cb->cb_prevsnap = NULL;
 	}
 	zfs_close(zhp);
@@ -1049,13 +1053,13 @@ destroy_print_snapshots(zfs_handle_t *fs_zhp, destroy_cbdata_t *cb)
 	if (cb->cb_firstsnap != NULL) {
 		uint64_t used = 0;
 		if (err == 0) {
-			err = zfs_get_snapused_int(cb->cb_firstsnap,
+			err = lzc_snaprange_space(cb->cb_firstsnap,
 			    cb->cb_prevsnap, &used);
 		}
 		cb->cb_snapused += used;
-		zfs_close(cb->cb_firstsnap);
+		free(cb->cb_firstsnap);
 		cb->cb_firstsnap = NULL;
-		zfs_close(cb->cb_prevsnap);
+		free(cb->cb_prevsnap);
 		cb->cb_prevsnap = NULL;
 	}
 	return (err);
@@ -1908,9 +1912,11 @@ upgrade_set_callback(zfs_handle_t *zhp, void *data)
 			/*
 			 * If they did "zfs upgrade -a", then we could
 			 * be doing ioctls to different pools.  We need
-			 * to log this history once to each pool.
+			 * to log this history once to each pool, and bypass
+			 * the normal history logging that happens in main().
 			 */
-			verify(zpool_stage_history(g_zfs, history_str) == 0);
+			(void) zpool_log_history(g_zfs, history_str);
+			log_history = B_FALSE;
 		}
 		if (zfs_prop_set(zhp, "version", verstr) == 0)
 			cb->cb_numupgraded++;
@@ -3422,6 +3428,32 @@ zfs_do_set(int argc, char **argv)
 	return (ret);
 }
 
+typedef struct snap_cbdata {
+	nvlist_t *sd_nvl;
+	boolean_t sd_recursive;
+	const char *sd_snapname;
+} snap_cbdata_t;
+
+static int
+zfs_snapshot_cb(zfs_handle_t *zhp, void *arg)
+{
+	snap_cbdata_t *sd = arg;
+	char *name;
+	int rv = 0;
+	int error;
+
+	error = asprintf(&name, "%s@%s", zfs_get_name(zhp), sd->sd_snapname);
+	if (error == -1)
+		nomem();
+	fnvlist_add_boolean(sd->sd_nvl, name);
+	free(name);
+
+	if (sd->sd_recursive)
+		rv = zfs_iter_filesystems(zhp, zfs_snapshot_cb, sd);
+	zfs_close(zhp);
+	return (rv);
+}
+
 /*
  * zfs snapshot [-r] [-o prop=value] ... <fs@snap>
  *
@@ -3431,13 +3463,16 @@ zfs_do_set(int argc, char **argv)
 static int
 zfs_do_snapshot(int argc, char **argv)
 {
-	boolean_t recursive = B_FALSE;
 	int ret = 0;
 	signed char c;
 	nvlist_t *props;
+	snap_cbdata_t sd = { 0 };
+	boolean_t multiple_snaps = B_FALSE;
 
 	if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0)
 		nomem();
+	if (nvlist_alloc(&sd.sd_nvl, NV_UNIQUE_NAME, 0) != 0)
+		nomem();
 
 	/* check options */
 	while ((c = getopt(argc, argv, "ro:")) != -1) {
@@ -3447,7 +3482,8 @@ zfs_do_snapshot(int argc, char **argv)
 				return (1);
 			break;
 		case 'r':
-			recursive = B_TRUE;
+			sd.sd_recursive = B_TRUE;
+			multiple_snaps = B_TRUE;
 			break;
 		case '?':
 			(void) fprintf(stderr, gettext("invalid option '%c'\n"),
@@ -3464,18 +3500,35 @@ zfs_do_snapshot(int argc, char **argv)
 		(void) fprintf(stderr, gettext("missing snapshot argument\n"));
 		goto usage;
 	}
-	if (argc > 1) {
-		(void) fprintf(stderr, gettext("too many arguments\n"));
-		goto usage;
+
+	if (argc > 1)
+		multiple_snaps = B_TRUE;
+	for (; argc > 0; argc--, argv++) {
+		char *atp;
+		zfs_handle_t *zhp;
+
+		atp = strchr(argv[0], '@');
+		if (atp == NULL)
+			goto usage;
+		*atp = '\0';
+		sd.sd_snapname = atp + 1;
+		zhp = zfs_open(g_zfs, argv[0],
+		    ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
+		if (zhp == NULL)
+			goto usage;
+		if (zfs_snapshot_cb(zhp, &sd) != 0)
+			goto usage;
 	}
 
-	ret = zfs_snapshot(g_zfs, argv[0], recursive, props);
+	ret = zfs_snapshot_nvl(g_zfs, sd.sd_nvl, props);
+	nvlist_free(sd.sd_nvl);
 	nvlist_free(props);
-	if (ret && recursive)
+	if (ret != 0 && multiple_snaps)
 		(void) fprintf(stderr, gettext("no snapshots were created\n"));
 	return (ret != 0);
 
 usage:
+	nvlist_free(sd.sd_nvl);
 	nvlist_free(props);
 	usage(B_FALSE);
 	return (-1);
@@ -6384,8 +6437,7 @@ main(int argc, char **argv)
 
 	mnttab_file = g_zfs->libzfs_mnttab;
 
-	zpool_set_history_str("zfs", argc, argv, history_str);
-	verify(zpool_stage_history(g_zfs, history_str) == 0);
+	zfs_save_arguments(argc, argv, history_str, sizeof (history_str));
 
 	libzfs_print_on_error(g_zfs, B_TRUE);
 
@@ -6408,6 +6460,9 @@ main(int argc, char **argv)
 	}
 	libzfs_fini(g_zfs);
 
+	if (ret == 0 && log_history)
+		(void) zpool_log_history(g_zfs, history_str);
+
 	/*
 	 * The 'ZFS_ABORT' environment variable causes us to dump core on exit
 	 * for the purposes of running ::findleaks.

cmd/zhack/zhack.c

@@ -279,6 +279,9 @@ feature_enable_sync(void *arg1, void *arg2, dmu_tx_t *tx)
 	zfeature_info_t *feature = arg2;
 
 	spa_feature_enable(spa, feature, tx);
+	spa_history_log_internal(spa, "zhack enable feature", tx,
+	    "name=%s can_readonly=%u",
+	    feature->fi_guid, feature->fi_can_readonly);
 }
 
 static void
@@ -356,6 +359,8 @@ feature_incr_sync(void *arg1, void *arg2, dmu_tx_t *tx)
 	zfeature_info_t *feature = arg2;
 
 	spa_feature_incr(spa, feature, tx);
+	spa_history_log_internal(spa, "zhack feature incr", tx,
+	    "name=%s", feature->fi_guid);
 }
 
 static void
@@ -365,6 +370,8 @@ feature_decr_sync(void *arg1, void *arg2, dmu_tx_t *tx)
 	zfeature_info_t *feature = arg2;
 
 	spa_feature_decr(spa, feature, tx);
+	spa_history_log_internal(spa, "zhack feature decr", tx,
+	    "name=%s", feature->fi_guid);
 }
 
 static void