Tag zfs-0.8.2

META file and changelog updated. Signed-off-by: Tony Hutter <hutter2@llnl.gov>
Disabled resilver_defer feature leads to looping resilvers
2026-05-23 02:44:41 +03:00 · 2019-09-25 11:27:51 -07:00 · 2019-09-25 11:27:51 -07:00 · 2019-09-25 11:27:51 -07:00 · 2019-09-25 11:27:51 -07:00 · 2019-09-25 11:27:51 -07:00
208 changed files with 2831 additions and 1474 deletions
@@ -63,3 +63,14 @@ cscope.*
 *.log
 venv

+#
+# Module leftovers
+#
+/module/avl/zavl.mod
+/module/icp/icp.mod
+/module/lua/zlua.mod
+/module/nvpair/znvpair.mod
+/module/spl/spl.mod
+/module/unicode/zunicode.mod
+/module/zcommon/zcommon.mod
+/module/zfs/zfs.mod
@@ -1,10 +1,10 @@
 Meta:          1
 Name:          zfs
 Branch:        1.0
-Version:       0.8.1
+Version:       0.8.2
 Release:       1
 Release-Tags:  relext
 License:       CDDL
 Author:        OpenZFS on Linux
-Linux-Maximum: 5.1
+Linux-Maximum: 5.3
 Linux-Minimum: 2.6.32
@@ -52,7 +52,8 @@ distclean-local::
 		-type f -print | xargs $(RM)

 all-local:
-	-${top_srcdir}/scripts/zfs-tests.sh -c
+	-[ -x ${top_builddir}/scripts/zfs-tests.sh ] && \
+	    ${top_builddir}/scripts/zfs-tests.sh -c

 dist-hook: gitrev
 	cp ${top_srcdir}/include/zfs_gitrev.h $(distdir)/include; \
@@ -5,4 +5,4 @@ if USING_PYTHON
 SUBDIRS += arcstat arc_summary dbufstat
 endif

-SUBDIRS += mount_zfs zed zvol_id
+SUBDIRS += mount_zfs zed zvol_id zvol_wait
@@ -1,12 +1,11 @@
+SUBDIRS = zed.d
+
 include $(top_srcdir)/config/Rules.am

 DEFAULT_INCLUDES += \
 	-I$(top_srcdir)/include \
 	-I$(top_srcdir)/lib/libspl/include

-EXTRA_DIST = zed.d/README \
-	zed.d/history_event-zfs-list-cacher.sh.in
-
 sbin_PROGRAMS = zed

 ZED_SRC = \
@@ -47,55 +46,3 @@ zed_LDADD = \

 zed_LDADD += -lrt
 zed_LDFLAGS = -pthread
-
-zedconfdir = $(sysconfdir)/zfs/zed.d
-
-dist_zedconf_DATA = \
-	zed.d/zed-functions.sh \
-	zed.d/zed.rc
-
-zedexecdir = $(zfsexecdir)/zed.d
-
-dist_zedexec_SCRIPTS = \
-	zed.d/all-debug.sh \
-	zed.d/all-syslog.sh \
-	zed.d/data-notify.sh \
-	zed.d/generic-notify.sh \
-	zed.d/resilver_finish-notify.sh \
-	zed.d/scrub_finish-notify.sh \
-	zed.d/statechange-led.sh \
-	zed.d/statechange-notify.sh \
-	zed.d/vdev_clear-led.sh \
-	zed.d/vdev_attach-led.sh \
-	zed.d/pool_import-led.sh \
-	zed.d/resilver_finish-start-scrub.sh
-
-nodist_zedexec_SCRIPTS = zed.d/history_event-zfs-list-cacher.sh
-
-$(nodist_zedexec_SCRIPTS): %: %.in
-	-$(SED) -e 's,@bindir\@,$(bindir),g' \
-		-e 's,@runstatedir\@,$(runstatedir),g' \
-		-e 's,@sbindir\@,$(sbindir),g' \
-		-e 's,@sysconfdir\@,$(sysconfdir),g' \
-		$< >'$@'
-
-zedconfdefaults = \
-	all-syslog.sh \
-	data-notify.sh \
-	resilver_finish-notify.sh \
-	scrub_finish-notify.sh \
-	statechange-led.sh \
-	statechange-notify.sh \
-	vdev_clear-led.sh \
-	vdev_attach-led.sh \
-	pool_import-led.sh \
-	resilver_finish-start-scrub.sh
-
-install-data-hook:
-	$(MKDIR_P) "$(DESTDIR)$(zedconfdir)"
-	for f in $(zedconfdefaults); do \
-	  test -f "$(DESTDIR)$(zedconfdir)/$${f}" -o \
-	       -L "$(DESTDIR)$(zedconfdir)/$${f}" || \
-	    ln -s "$(zedexecdir)/$${f}" "$(DESTDIR)$(zedconfdir)"; \
-	done
-	chmod 0600 "$(DESTDIR)$(zedconfdir)/zed.rc"
@@ -116,7 +116,8 @@ zfs_agent_iter_vdev(zpool_handle_t *zhp, nvlist_t *nvl, void *arg)
 	/*
 	 * On a devid match, grab the vdev guid and expansion time, if any.
 	 */
-	if ((nvlist_lookup_string(nvl, ZPOOL_CONFIG_DEVID, &path) == 0) &&
+	if (gsp->gs_devid != NULL &&
+	    (nvlist_lookup_string(nvl, ZPOOL_CONFIG_DEVID, &path) == 0) &&
 	    (strcmp(gsp->gs_devid, path) == 0)) {
 		(void) nvlist_lookup_uint64(nvl, ZPOOL_CONFIG_GUID,
 		    &gsp->gs_vdev_guid);
@@ -0,0 +1,57 @@
+include $(top_srcdir)/config/Rules.am
+
+EXTRA_DIST = \
+	README \
+	history_event-zfs-list-cacher.sh.in
+
+zedconfdir = $(sysconfdir)/zfs/zed.d
+
+dist_zedconf_DATA = \
+	zed-functions.sh \
+	zed.rc
+
+zedexecdir = $(zfsexecdir)/zed.d
+
+dist_zedexec_SCRIPTS = \
+	all-debug.sh \
+	all-syslog.sh \
+	data-notify.sh \
+	generic-notify.sh \
+	resilver_finish-notify.sh \
+	scrub_finish-notify.sh \
+	statechange-led.sh \
+	statechange-notify.sh \
+	vdev_clear-led.sh \
+	vdev_attach-led.sh \
+	pool_import-led.sh \
+	resilver_finish-start-scrub.sh
+
+nodist_zedexec_SCRIPTS = history_event-zfs-list-cacher.sh
+
+$(nodist_zedexec_SCRIPTS): %: %.in
+	-$(SED) -e 's,@bindir\@,$(bindir),g' \
+		-e 's,@runstatedir\@,$(runstatedir),g' \
+		-e 's,@sbindir\@,$(sbindir),g' \
+		-e 's,@sysconfdir\@,$(sysconfdir),g' \
+		$< >'$@'
+
+zedconfdefaults = \
+	all-syslog.sh \
+	data-notify.sh \
+	resilver_finish-notify.sh \
+	scrub_finish-notify.sh \
+	statechange-led.sh \
+	statechange-notify.sh \
+	vdev_clear-led.sh \
+	vdev_attach-led.sh \
+	pool_import-led.sh \
+	resilver_finish-start-scrub.sh
+
+install-data-hook:
+	$(MKDIR_P) "$(DESTDIR)$(zedconfdir)"
+	for f in $(zedconfdefaults); do \
+	  test -f "$(DESTDIR)$(zedconfdir)/$${f}" -o \
+	       -L "$(DESTDIR)$(zedconfdir)/$${f}" || \
+	    ln -s "$(zedexecdir)/$${f}" "$(DESTDIR)$(zedconfdir)"; \
+	done
+	chmod 0600 "$(DESTDIR)$(zedconfdir)/zed.rc"
@@ -47,7 +47,7 @@ case "${ZEVENT_HISTORY_INTERNAL_NAME}" in
        # Only act if one of the tracked properties is altered.
        case "${ZEVENT_HISTORY_INTERNAL_STR%%=*}" in
            canmount|mountpoint|atime|relatime|devices|exec| \
-                readonly|setuid|nbmand) ;;
+                readonly|setuid|nbmand|encroot|keylocation) ;;
            *) exit 0 ;;
        esac
      ;;
@@ -62,7 +62,7 @@ zed_lock zfs-list
 trap abort_alter EXIT

 PROPS="name,mountpoint,canmount,atime,relatime,devices,exec,readonly"
-PROPS="${PROPS},setuid,nbmand"
+PROPS="${PROPS},setuid,nbmand,encroot,keylocation"

 "${ZFS}" list -H -t filesystem -o $PROPS -r "${ZEVENT_POOL}" > "${FSLIST_TMP}"

@@ -6446,8 +6446,25 @@ share_mount_one(zfs_handle_t *zhp, int op, int flags, char *protocol,
 			return (1);
 		}

-		if (zfs_mount(zhp, options, flags) != 0)
+		if (zfs_mount(zhp, options, flags) != 0) {
+			/*
+			 * Check if a mount sneaked in after we checked
+			 */
+			if (!explicit &&
+			    libzfs_errno(g_zfs) == EZFS_MOUNTFAILED) {
+				usleep(10 * MILLISEC);
+				libzfs_mnttab_cache(g_zfs, B_FALSE);
+
+				if (zfs_is_mounted(zhp, NULL)) {
+					(void) fprintf(stderr, gettext(
+					    "Ignoring previous 'already "
+					    "mounted' error for '%s'\n"),
+					    zfs_get_name(zhp));
+					return (0);
+				}
+			}
 			return (1);
+		}
 		break;
 	}

@@ -6733,8 +6750,8 @@ unshare_unmount_compare(const void *larg, const void *rarg, void *unused)

 /*
 * Convenience routine used by zfs_do_umount() and manual_unmount().  Given an
- * absolute path, find the entry /proc/self/mounts, verify that its a
- * ZFS filesystems, and unmount it appropriately.
+ * absolute path, find the entry /proc/self/mounts, verify that it's a
+ * ZFS filesystem, and unmount it appropriately.
 */
 static int
 unshare_unmount_path(int op, char *path, int flags, boolean_t is_manual)
@@ -785,7 +785,7 @@ add_prop_list_default(const char *propname, char *propval, nvlist_t **props,
 *	-P	Display full path for vdev name.
 *
 * Adds the given vdevs to 'pool'.  As with create, the bulk of this work is
- * handled by get_vdev_spec(), which constructs the nvlist needed to pass to
+ * handled by make_root_vdev(), which constructs the nvlist needed to pass to
 * libzfs.
 */
 int
@@ -883,7 +883,7 @@ zpool_do_add(int argc, char **argv)
 		}
 	}

-	/* pass off to get_vdev_spec for processing */
+	/* pass off to make_root_vdev for processing */
 	nvroot = make_root_vdev(zhp, props, force, !force, B_FALSE, dryrun,
 	    argc, argv);
 	if (nvroot == NULL) {
@@ -1232,9 +1232,9 @@ errout:
 *	-O	Set fsproperty=value in the pool's root file system
 *
 * Creates the named pool according to the given vdev specification.  The
- * bulk of the vdev processing is done in get_vdev_spec() in zpool_vdev.c.  Once
- * we get the nvlist back from get_vdev_spec(), we either print out the contents
- * (if '-n' was specified), or pass it to libzfs to do the creation.
+ * bulk of the vdev processing is done in make_root_vdev() in zpool_vdev.c.
+ * Once we get the nvlist back from make_root_vdev(), we either print out the
+ * contents (if '-n' was specified), or pass it to libzfs to do the creation.
 */
 int
 zpool_do_create(int argc, char **argv)
@@ -1388,7 +1388,7 @@ zpool_do_create(int argc, char **argv)
 		goto errout;
 	}

-	/* pass off to get_vdev_spec for bulk processing */
+	/* pass off to make_root_vdev for bulk processing */
 	nvroot = make_root_vdev(NULL, props, force, !force, B_FALSE, dryrun,
 	    argc - 1, argv + 1);
 	if (nvroot == NULL)
@@ -6111,9 +6111,8 @@ zpool_do_detach(int argc, char **argv)
 	int ret;

 	/* check options */
-	while ((c = getopt(argc, argv, "f")) != -1) {
+	while ((c = getopt(argc, argv, "")) != -1) {
 		switch (c) {
-		case 'f':
 		case '?':
 			(void) fprintf(stderr, gettext("invalid option '%c'\n"),
 			    optopt);
@@ -6342,12 +6341,11 @@ zpool_do_online(int argc, char **argv)
 	int flags = 0;

 	/* check options */
-	while ((c = getopt(argc, argv, "et")) != -1) {
+	while ((c = getopt(argc, argv, "e")) != -1) {
 		switch (c) {
 		case 'e':
 			flags |= ZFS_ONLINE_EXPAND;
 			break;
-		case 't':
 		case '?':
 			(void) fprintf(stderr, gettext("invalid option '%c'\n"),
 			    optopt);
@@ -433,6 +433,7 @@ check_disk(const char *path, blkid_cache cache, int force,
 		char *value = blkid_get_tag_value(cache, "TYPE", path);
 		(void) fprintf(stderr, gettext("%s is in use and contains "
 		    "a %s filesystem.\n"), path, value ? value : "unknown");
+		free(value);
 		return (-1);
 	}

@@ -53,7 +53,6 @@
 */
 #define	DUMP_GROUPING	4

-uint64_t total_write_size = 0;
 uint64_t total_stream_len = 0;
 FILE *send_stream = 0;
 boolean_t do_byteswap = B_FALSE;
@@ -219,6 +218,9 @@ main(int argc, char *argv[])
 {
 	char *buf = safe_malloc(SPA_MAXBLOCKSIZE);
 	uint64_t drr_record_count[DRR_NUMTYPES] = { 0 };
+	uint64_t total_payload_size = 0;
+	uint64_t total_overhead_size = 0;
+	uint64_t drr_byte_count[DRR_NUMTYPES] = { 0 };
 	char salt[ZIO_DATA_SALT_LEN * 2 + 1];
 	char iv[ZIO_DATA_IV_LEN * 2 + 1];
 	char mac[ZIO_DATA_MAC_LEN * 2 + 1];
@@ -336,7 +338,9 @@ main(int argc, char *argv[])
 		}

 		drr_record_count[drr->drr_type]++;
+		total_overhead_size += sizeof (*drr);
 		total_records++;
+		payload_size = 0;

 		switch (drr->drr_type) {
 		case DRR_BEGIN:
@@ -390,6 +394,7 @@ main(int argc, char *argv[])
 					nvlist_print(stdout, nv);
 					nvlist_free(nv);
 				}
+				payload_size = sz;
 			}
 			break;

@@ -554,7 +559,6 @@ main(int argc, char *argv[])
 			if (dump) {
 				print_block(buf, payload_size);
 			}
-			total_write_size += payload_size;
 			break;

 		case DRR_WRITE_BYREF:
@@ -683,6 +687,7 @@ main(int argc, char *argv[])
 				print_block(buf,
 				    P2ROUNDUP(drrwe->drr_psize, 8));
 			}
+			payload_size = P2ROUNDUP(drrwe->drr_psize, 8);
 			break;
 		case DRR_OBJECT_RANGE:
 			if (do_byteswap) {
@@ -723,6 +728,8 @@ main(int argc, char *argv[])
 			    (longlong_t)drrc->drr_checksum.zc_word[3]);
 		}
 		pcksum = zc;
+		drr_byte_count[drr->drr_type] += payload_size;
+		total_payload_size += payload_size;
 	}
 	free(buf);
 	fletcher_4_fini();
@@ -730,28 +737,40 @@ main(int argc, char *argv[])
 	/* Print final summary */

 	(void) printf("SUMMARY:\n");
-	(void) printf("\tTotal DRR_BEGIN records = %lld\n",
-	    (u_longlong_t)drr_record_count[DRR_BEGIN]);
-	(void) printf("\tTotal DRR_END records = %lld\n",
-	    (u_longlong_t)drr_record_count[DRR_END]);
-	(void) printf("\tTotal DRR_OBJECT records = %lld\n",
-	    (u_longlong_t)drr_record_count[DRR_OBJECT]);
-	(void) printf("\tTotal DRR_FREEOBJECTS records = %lld\n",
-	    (u_longlong_t)drr_record_count[DRR_FREEOBJECTS]);
-	(void) printf("\tTotal DRR_WRITE records = %lld\n",
-	    (u_longlong_t)drr_record_count[DRR_WRITE]);
-	(void) printf("\tTotal DRR_WRITE_BYREF records = %lld\n",
-	    (u_longlong_t)drr_record_count[DRR_WRITE_BYREF]);
-	(void) printf("\tTotal DRR_WRITE_EMBEDDED records = %lld\n",
-	    (u_longlong_t)drr_record_count[DRR_WRITE_EMBEDDED]);
-	(void) printf("\tTotal DRR_FREE records = %lld\n",
-	    (u_longlong_t)drr_record_count[DRR_FREE]);
-	(void) printf("\tTotal DRR_SPILL records = %lld\n",
-	    (u_longlong_t)drr_record_count[DRR_SPILL]);
+	(void) printf("\tTotal DRR_BEGIN records = %lld (%llu bytes)\n",
+	    (u_longlong_t)drr_record_count[DRR_BEGIN],
+	    (u_longlong_t)drr_byte_count[DRR_BEGIN]);
+	(void) printf("\tTotal DRR_END records = %lld (%llu bytes)\n",
+	    (u_longlong_t)drr_record_count[DRR_END],
+	    (u_longlong_t)drr_byte_count[DRR_END]);
+	(void) printf("\tTotal DRR_OBJECT records = %lld (%llu bytes)\n",
+	    (u_longlong_t)drr_record_count[DRR_OBJECT],
+	    (u_longlong_t)drr_byte_count[DRR_OBJECT]);
+	(void) printf("\tTotal DRR_FREEOBJECTS records = %lld (%llu bytes)\n",
+	    (u_longlong_t)drr_record_count[DRR_FREEOBJECTS],
+	    (u_longlong_t)drr_byte_count[DRR_FREEOBJECTS]);
+	(void) printf("\tTotal DRR_WRITE records = %lld (%llu bytes)\n",
+	    (u_longlong_t)drr_record_count[DRR_WRITE],
+	    (u_longlong_t)drr_byte_count[DRR_WRITE]);
+	(void) printf("\tTotal DRR_WRITE_BYREF records = %lld (%llu bytes)\n",
+	    (u_longlong_t)drr_record_count[DRR_WRITE_BYREF],
+	    (u_longlong_t)drr_byte_count[DRR_WRITE_BYREF]);
+	(void) printf("\tTotal DRR_WRITE_EMBEDDED records = %lld (%llu "
+	    "bytes)\n", (u_longlong_t)drr_record_count[DRR_WRITE_EMBEDDED],
+	    (u_longlong_t)drr_byte_count[DRR_WRITE_EMBEDDED]);
+	(void) printf("\tTotal DRR_FREE records = %lld (%llu bytes)\n",
+	    (u_longlong_t)drr_record_count[DRR_FREE],
+	    (u_longlong_t)drr_byte_count[DRR_FREE]);
+	(void) printf("\tTotal DRR_SPILL records = %lld (%llu bytes)\n",
+	    (u_longlong_t)drr_record_count[DRR_SPILL],
+	    (u_longlong_t)drr_byte_count[DRR_SPILL]);
 	(void) printf("\tTotal records = %lld\n",
 	    (u_longlong_t)total_records);
-	(void) printf("\tTotal write size = %lld (0x%llx)\n",
-	    (u_longlong_t)total_write_size, (u_longlong_t)total_write_size);
+	(void) printf("\tTotal payload size = %lld (0x%llx)\n",
+	    (u_longlong_t)total_payload_size, (u_longlong_t)total_payload_size);
+	(void) printf("\tTotal header overhead = %lld (0x%llx)\n",
+	    (u_longlong_t)total_overhead_size,
+	    (u_longlong_t)total_overhead_size);
 	(void) printf("\tTotal stream length = %lld (0x%llx)\n",
 	    (u_longlong_t)total_stream_len, (u_longlong_t)total_stream_len);
 	return (0);
@@ -2745,8 +2745,24 @@ ztest_spa_create_destroy(ztest_ds_t *zd, uint64_t id)
 	VERIFY3U(EEXIST, ==,
 	    spa_create(zo->zo_pool, nvroot, NULL, NULL, NULL));
 	nvlist_free(nvroot);
+
+	/*
+	 * We open a reference to the spa and then we try to export it
+	 * expecting one of the following errors:
+	 *
+	 * EBUSY
+	 *	Because of the reference we just opened.
+	 *
+	 * ZFS_ERR_EXPORT_IN_PROGRESS
+	 *	For the case that there is another ztest thread doing
+	 *	an export concurrently.
+	 */
 	VERIFY3U(0, ==, spa_open(zo->zo_pool, &spa, FTAG));
-	VERIFY3U(EBUSY, ==, spa_destroy(zo->zo_pool));
+	int error = spa_destroy(zo->zo_pool);
+	if (error != EBUSY && error != ZFS_ERR_EXPORT_IN_PROGRESS) {
+		fatal(0, "spa_destroy(%s) returned unexpected value %d",
+		    spa->spa_name, error);
+	}
 	spa_close(spa, FTAG);

 	(void) pthread_rwlock_unlock(&ztest_name_lock);
@@ -0,0 +1 @@
+dist_bin_SCRIPTS = zvol_wait
@@ -0,0 +1,112 @@
+#!/bin/sh
+
+count_zvols() {
+	if [ -z "$zvols" ]; then
+		echo 0
+	else
+		echo "$zvols" | wc -l
+	fi
+}
+
+filter_out_zvols_with_links() {
+	while read -r zvol; do
+		if [ ! -L "/dev/zvol/$zvol" ]; then
+			echo "$zvol"
+		fi
+	done
+}
+
+filter_out_deleted_zvols() {
+	while read -r zvol; do
+		if zfs list "$zvol" >/dev/null 2>&1; then
+			echo "$zvol"
+		fi
+	done
+}
+
+list_zvols() {
+	zfs list -t volume -H -o name,volmode,receive_resume_token |
+		while read -r zvol_line; do
+		name=$(echo "$zvol_line" | awk '{print $1}')
+		volmode=$(echo "$zvol_line" | awk '{print $2}')
+		token=$(echo "$zvol_line" | awk '{print $3}')
+		#
+		# /dev links are not created for zvols with volmode = "none".
+		#
+		[ "$volmode" = "none" ] && continue
+		#
+		# We also also ignore partially received zvols if it is
+		# not an incremental receive, as those won't even have a block
+		# device minor node created yet.
+		#
+		if [ "$token" != "-" ]; then
+			#
+			# Incremental receives create an invisible clone that
+			# is not automatically displayed by zfs list.
+			#
+			if ! zfs list "$name/%recv" >/dev/null 2>&1; then
+				continue
+			fi
+		fi
+		echo "$name"
+	done
+}
+
+zvols=$(list_zvols)
+zvols_count=$(count_zvols)
+if [ "$zvols_count" -eq 0 ]; then
+	echo "No zvols found, nothing to do."
+	exit 0
+fi
+
+echo "Testing $zvols_count zvol links"
+
+outer_loop=0
+while [ "$outer_loop" -lt 20 ]; do
+	outer_loop=$((outer_loop + 1))
+
+	old_zvols_count=$(count_zvols)
+
+	inner_loop=0
+	while [ "$inner_loop" -lt 30 ]; do
+		inner_loop=$((inner_loop + 1))
+
+		zvols="$(echo "$zvols" | filter_out_zvols_with_links)"
+
+		zvols_count=$(count_zvols)
+		if [ "$zvols_count" -eq 0 ]; then
+			echo "All zvol links are now present."
+			exit 0
+		fi
+		sleep 1
+	done
+
+	echo "Still waiting on $zvols_count zvol links ..."
+	#
+	# Although zvols should normally not be deleted at boot time,
+	# if that is the case then their links will be missing and
+	# we would stall.
+	#
+	if [ "$old_zvols_count" -eq "$zvols_count" ]; then
+		echo "No progress since last loop."
+		echo "Checking if any zvols were deleted."
+
+		zvols=$(echo "$zvols" | filter_out_deleted_zvols)
+		zvols_count=$(count_zvols)
+
+		if [ "$old_zvols_count" -ne "$zvols_count" ]; then
+			echo "$((old_zvols_count - zvols_count)) zvol(s) deleted."
+		fi
+
+		if [ "$zvols_count" -ne 0 ]; then
+			echo "Remaining zvols:"
+			echo "$zvols"
+		else
+			echo "All zvol links are now present."
+			exit 0
+		fi
+	fi
+done
+
+echo "Timed out waiting on zvol links"
+exit 1
@@ -1,47 +1,3 @@
-dnl #
-dnl # ZFS_AC_PYTHON_VERSION(version, [action-if-true], [action-if-false])
-dnl #
-dnl # Verify Python version
-dnl #
-AC_DEFUN([ZFS_AC_PYTHON_VERSION], [
-	ver_check=`$PYTHON -c "import sys; print (sys.version.split()[[0]] $1)"`
-	AS_IF([test "$ver_check" = "True"], [
-		m4_ifvaln([$2], [$2])
-	], [
-		m4_ifvaln([$3], [$3])
-	])
-])
-
-dnl #
-dnl # ZFS_AC_PYTHON_VERSION_IS_2
-dnl # ZFS_AC_PYTHON_VERSION_IS_3
-dnl #
-dnl # Tests if the $PYTHON_VERSION matches 2.x or 3.x.
-dnl #
-AC_DEFUN([ZFS_AC_PYTHON_VERSION_IS_2],
-	[test "${PYTHON_VERSION%%\.*}" = "2"])
-AC_DEFUN([ZFS_AC_PYTHON_VERSION_IS_3],
-	[test "${PYTHON_VERSION%%\.*}" = "3"])
-
-dnl #
-dnl # ZFS_AC_PYTHON_MODULE(module_name, [action-if-true], [action-if-false])
-dnl #
-dnl # Checks for Python module. Freely inspired by AX_PYTHON_MODULE
-dnl # https://www.gnu.org/software/autoconf-archive/ax_python_module.html
-dnl # Required by ZFS_AC_CONFIG_ALWAYS_PYZFS.
-dnl #
-AC_DEFUN([ZFS_AC_PYTHON_MODULE], [
-	PYTHON_NAME=`basename $PYTHON`
-	AC_MSG_CHECKING([for $PYTHON_NAME module: $1])
-	AS_IF([$PYTHON -c "import $1" 2>/dev/null], [
-		AC_MSG_RESULT(yes)
-		m4_ifvaln([$2], [$2])
-	], [
-		AC_MSG_RESULT(no)
-		m4_ifvaln([$3], [$3])
-	])
-])
-
 dnl #
 dnl # The majority of the python scripts are written to be compatible
 dnl # with Python 2.6 and Python 3.4.  Therefore, they may be installed
@@ -66,35 +22,38 @@ AC_DEFUN([ZFS_AC_CONFIG_ALWAYS_PYTHON], [
 		[AC_MSG_ERROR([Unknown --with-python value '$with_python'])]
 	)

-	AS_IF([test $PYTHON != :], [
-		AS_IF([$PYTHON --version >/dev/null 2>&1],
-			[AM_PATH_PYTHON([2.6], [], [:])],
-			[AC_MSG_ERROR([Cannot find $PYTHON in your system path])]
-		)
-	])
-	AM_CONDITIONAL([USING_PYTHON], [test "$PYTHON" != :])
-	AM_CONDITIONAL([USING_PYTHON_2], [ZFS_AC_PYTHON_VERSION_IS_2])
-	AM_CONDITIONAL([USING_PYTHON_3], [ZFS_AC_PYTHON_VERSION_IS_3])
-
 	dnl #
 	dnl # Minimum supported Python versions for utilities:
-	dnl # Python 2.6.x, or Python 3.4.x
+	dnl # Python 2.6 or Python 3.4
 	dnl #
-	AS_IF([ZFS_AC_PYTHON_VERSION_IS_2], [
-		ZFS_AC_PYTHON_VERSION([>= '2.6'], [ true ],
-			[AC_MSG_ERROR("Python >= 2.6.x is not available")])
+	AM_PATH_PYTHON([], [], [:])
+	AS_IF([test -z "$PYTHON_VERSION"], [
+		PYTHON_VERSION=$(basename $PYTHON | tr -cd 0-9.)
 	])
+	PYTHON_MINOR=${PYTHON_VERSION#*\.}

-	AS_IF([ZFS_AC_PYTHON_VERSION_IS_3], [
-		ZFS_AC_PYTHON_VERSION([>= '3.4'], [ true ],
-			[AC_MSG_ERROR("Python >= 3.4.x is not available")])
-	])
+	AS_CASE([$PYTHON_VERSION],
+		[2.*], [
+			AS_IF([test $PYTHON_MINOR -lt 6],
+				[AC_MSG_ERROR("Python >= 2.6 is required")])
+		],
+		[3.*], [
+			AS_IF([test $PYTHON_MINOR -lt 4],
+				[AC_MSG_ERROR("Python >= 3.4 is required")])
+		],
+		[:|2|3], [],
+		[PYTHON_VERSION=3]
+	)
+
+	AM_CONDITIONAL([USING_PYTHON], [test "$PYTHON" != :])
+	AM_CONDITIONAL([USING_PYTHON_2], [test "x${PYTHON_VERSION%%\.*}" = x2])
+	AM_CONDITIONAL([USING_PYTHON_3], [test "x${PYTHON_VERSION%%\.*}" = x3])

 	dnl #
 	dnl # Request that packages be built for a specific Python version.
 	dnl #
-	AS_IF([test $with_python != check], [
-		PYTHON_PKG_VERSION=`echo ${PYTHON} | tr -d 'a-zA-Z.'`
+	AS_IF([test "x$with_python" != xcheck], [
+		PYTHON_PKG_VERSION=$(echo $PYTHON_VERSION | tr -d .)
 		DEFINE_PYTHON_PKG_VERSION='--define "__use_python_pkg_version '${PYTHON_PKG_VERSION}'"'
 		DEFINE_PYTHON_VERSION='--define "__use_python '${PYTHON}'"'
 	], [
@@ -1,5 +1,24 @@
 dnl #
-dnl # Determines if pyzfs can be built, requires Python 2.7 or latter.
+dnl # ZFS_AC_PYTHON_MODULE(module_name, [action-if-true], [action-if-false])
+dnl #
+dnl # Checks for Python module. Freely inspired by AX_PYTHON_MODULE
+dnl # https://www.gnu.org/software/autoconf-archive/ax_python_module.html
+dnl # Required by ZFS_AC_CONFIG_ALWAYS_PYZFS.
+dnl #
+AC_DEFUN([ZFS_AC_PYTHON_MODULE], [
+	PYTHON_NAME=$(basename $PYTHON)
+	AC_MSG_CHECKING([for $PYTHON_NAME module: $1])
+	AS_IF([$PYTHON -c "import $1" 2>/dev/null], [
+		AC_MSG_RESULT(yes)
+		m4_ifvaln([$2], [$2])
+	], [
+		AC_MSG_RESULT(no)
+		m4_ifvaln([$3], [$3])
+	])
+])
+
+dnl #
+dnl # Determines if pyzfs can be built, requires Python 2.7 or later.
 dnl #
 AC_DEFUN([ZFS_AC_CONFIG_ALWAYS_PYZFS], [
 	AC_ARG_ENABLE([pyzfs],
@@ -18,7 +37,7 @@ AC_DEFUN([ZFS_AC_CONFIG_ALWAYS_PYZFS], [
 			DEFINE_PYZFS='--without pyzfs'
 		])
 	], [
-		AS_IF([test $PYTHON != :], [
+		AS_IF([test "$PYTHON" != :], [
 			DEFINE_PYZFS=''
 		], [
 			enable_pyzfs=no
@@ -31,20 +50,16 @@ AC_DEFUN([ZFS_AC_CONFIG_ALWAYS_PYZFS], [
 	dnl # Require python-devel libraries
 	dnl #
 	AS_IF([test "x$enable_pyzfs" = xcheck  -o "x$enable_pyzfs" = xyes], [
-		AS_IF([ZFS_AC_PYTHON_VERSION_IS_2], [
-			PYTHON_REQUIRED_VERSION=">= '2.7.0'"
-		], [
-			AS_IF([ZFS_AC_PYTHON_VERSION_IS_3], [
-				PYTHON_REQUIRED_VERSION=">= '3.4.0'"
-			], [
-				AC_MSG_ERROR("Python $PYTHON_VERSION unknown")
-			])
-		])
+		AS_CASE([$PYTHON_VERSION],
+			[3.*], [PYTHON_REQUIRED_VERSION=">= '3.4.0'"],
+			[2.*], [PYTHON_REQUIRED_VERSION=">= '2.7.0'"],
+			[AC_MSG_ERROR("Python $PYTHON_VERSION unknown")]
+		)

 		AX_PYTHON_DEVEL([$PYTHON_REQUIRED_VERSION], [
 			AS_IF([test "x$enable_pyzfs" = xyes], [
 				AC_MSG_ERROR("Python $PYTHON_REQUIRED_VERSION development library is not installed")
-			], [test ! "x$enable_pyzfs" = xno], [
+			], [test "x$enable_pyzfs" != xno], [
 				enable_pyzfs=no
 			])
 		])
@@ -57,7 +72,7 @@ AC_DEFUN([ZFS_AC_CONFIG_ALWAYS_PYZFS], [
 		ZFS_AC_PYTHON_MODULE([setuptools], [], [
 			AS_IF([test "x$enable_pyzfs" = xyes], [
 				AC_MSG_ERROR("Python $PYTHON_VERSION setuptools is not installed")
-			], [test ! "x$enable_pyzfs" = xno], [
+			], [test "x$enable_pyzfs" != xno], [
 				enable_pyzfs=no
 			])
 		])
@@ -70,7 +85,7 @@ AC_DEFUN([ZFS_AC_CONFIG_ALWAYS_PYZFS], [
 		ZFS_AC_PYTHON_MODULE([cffi], [], [
 			AS_IF([test "x$enable_pyzfs" = xyes], [
 				AC_MSG_ERROR("Python $PYTHON_VERSION cffi is not installed")
-			], [test ! "x$enable_pyzfs" = xno], [
+			], [test "x$enable_pyzfs" != xno], [
 				enable_pyzfs=no
 			])
 		])
@@ -81,7 +96,7 @@ AC_DEFUN([ZFS_AC_CONFIG_ALWAYS_PYZFS], [
 	dnl #
 	AS_IF([test "x$enable_pyzfs" = xcheck], [enable_pyzfs=yes])

-	AM_CONDITIONAL([PYZFS_ENABLED], [test x$enable_pyzfs = xyes])
+	AM_CONDITIONAL([PYZFS_ENABLED], [test "x$enable_pyzfs" = xyes])
 	AC_SUBST([PYZFS_ENABLED], [$enable_pyzfs])
 	AC_SUBST(pythonsitedir, [$PYTHON_SITE_PKG])

@@ -20,7 +20,7 @@ deb-kmod: deb-local rpm-kmod
 	arch=`$(RPM) -qp $${name}-kmod-$${version}.src.rpm --qf %{arch} | tail -1`; \
 	debarch=`$(DPKG) --print-architecture`; \
 	pkg1=kmod-$${name}*$${version}.$${arch}.rpm; \
-	fakeroot $(ALIEN) --bump=0 --scripts --to-deb --target=$$debarch $$pkg1; \
+	fakeroot $(ALIEN) --bump=0 --scripts --to-deb --target=$$debarch $$pkg1 || exit 1; \
 	$(RM) $$pkg1


@@ -30,7 +30,7 @@ deb-dkms: deb-local rpm-dkms
 	arch=`$(RPM) -qp $${name}-dkms-$${version}.src.rpm --qf %{arch} | tail -1`; \
 	debarch=`$(DPKG) --print-architecture`; \
 	pkg1=$${name}-dkms-$${version}.$${arch}.rpm; \
-	fakeroot $(ALIEN) --bump=0 --scripts --to-deb --target=$$debarch $$pkg1; \
+	fakeroot $(ALIEN) --bump=0 --scripts --to-deb --target=$$debarch $$pkg1 || exit 1; \
 	$(RM) $$pkg1

 deb-utils: deb-local rpm-utils
@@ -45,7 +45,7 @@ deb-utils: deb-local rpm-utils
 	pkg5=libzpool2-$${version}.$${arch}.rpm; \
 	pkg6=libzfs2-devel-$${version}.$${arch}.rpm; \
 	pkg7=$${name}-test-$${version}.$${arch}.rpm; \
-	pkg8=$${name}-dracut-$${version}.$${arch}.rpm; \
+	pkg8=$${name}-dracut-$${version}.noarch.rpm; \
 	pkg9=$${name}-initramfs-$${version}.$${arch}.rpm; \
 	pkg10=`ls python*-pyzfs-$${version}* | tail -1`; \
 ## Arguments need to be passed to dh_shlibdeps. Alien provides no mechanism
@@ -63,7 +63,7 @@ deb-utils: deb-local rpm-utils
 	env PATH=$${path_prepend}:$${PATH} \
 	fakeroot $(ALIEN) --bump=0 --scripts --to-deb --target=$$debarch \
 	    $$pkg1 $$pkg2 $$pkg3 $$pkg4 $$pkg5 $$pkg6 $$pkg7 \
-	    $$pkg8 $$pkg9 $$pkg10; \
+	    $$pkg8 $$pkg9 $$pkg10 || exit 1; \
 	$(RM) $${path_prepend}/dh_shlibdeps; \
 	rmdir $${path_prepend}; \
 	$(RM) $$pkg1 $$pkg2 $$pkg3 $$pkg4 $$pkg5 $$pkg6 $$pkg7 \
@@ -18,7 +18,8 @@ AC_DEFUN([ZFS_AC_KERNEL_FPU], [
 		#include <asm/fpu/api.h>
 	],[
 	],[
-		AC_DEFINE(HAVE_KERNEL_FPU_API_HEADER, 1, [kernel has asm/fpu/api.h])
+		AC_DEFINE(HAVE_KERNEL_FPU_API_HEADER, 1,
+		    [kernel has asm/fpu/api.h])
 		AC_MSG_RESULT(asm/fpu/api.h)
 	],[
 		AC_MSG_RESULT(i387.h & xcr.h)
@@ -39,8 +40,10 @@ AC_DEFUN([ZFS_AC_KERNEL_FPU], [
 		kernel_fpu_end();
 	], [kernel_fpu_begin], [arch/x86/kernel/fpu/core.c], [
 		AC_MSG_RESULT(kernel_fpu_*)
-		AC_DEFINE(HAVE_KERNEL_FPU, 1, [kernel has kernel_fpu_* functions])
-		AC_DEFINE(KERNEL_EXPORTS_X86_FPU, 1, [kernel exports FPU functions])
+		AC_DEFINE(HAVE_KERNEL_FPU, 1,
+		    [kernel has kernel_fpu_* functions])
+		AC_DEFINE(KERNEL_EXPORTS_X86_FPU, 1,
+		    [kernel exports FPU functions])
 	],[
 		ZFS_LINUX_TRY_COMPILE_SYMBOL([
 			#include <linux/module.h>
@@ -1,24 +0,0 @@
-dnl #
-dnl # 2.6.36 API change,
-dnl # The 'struct fs_struct->lock' was changed from a rwlock_t to
-dnl # a spinlock_t to improve the fastpath performance.
-dnl #
-AC_DEFUN([ZFS_AC_KERNEL_FS_STRUCT_SPINLOCK], [
-	AC_MSG_CHECKING([whether struct fs_struct uses spinlock_t])
-	tmp_flags="$EXTRA_KCFLAGS"
-	EXTRA_KCFLAGS="-Werror"
-	ZFS_LINUX_TRY_COMPILE([
-		#include <linux/sched.h>
-		#include <linux/fs_struct.h>
-	],[
-		static struct fs_struct fs;
-		spin_lock_init(&fs.lock);
-	],[
-		AC_MSG_RESULT(yes)
-		AC_DEFINE(HAVE_FS_STRUCT_SPINLOCK, 1,
-		          [struct fs_struct uses spinlock_t])
-	],[
-		AC_MSG_RESULT(no)
-	])
-	EXTRA_KCFLAGS="$tmp_flags"
-])
@@ -1,26 +1,51 @@
+dnl # 4.14-rc3 API change
+dnl # https://lwn.net/Articles/735887/
 dnl #
-dnl # 4.15 API change
-dnl # https://lkml.org/lkml/2017/11/25/90
 dnl # Check if timer_list.func get passed a timer_list or an unsigned long
 dnl # (older kernels).  Also sanity check the from_timer() and timer_setup()
 dnl # macros are available as well, since they will be used in the same newer
 dnl # kernels that support the new timer_list.func signature.
 dnl #
-AC_DEFUN([ZFS_AC_KERNEL_TIMER_FUNCTION_TIMER_LIST], [
-	AC_MSG_CHECKING([whether timer_list.function gets a timer_list])
+dnl # Also check for the existance of flags in struct timer_list, they were
+dnl # added in 4.1-rc8 via 0eeda71bc30d.
+
+AC_DEFUN([ZFS_AC_KERNEL_TIMER_SETUP], [
+	AC_MSG_CHECKING([whether timer_setup() is available])
 	tmp_flags="$EXTRA_KCFLAGS"
 	EXTRA_KCFLAGS="-Werror"
+
+	ZFS_LINUX_TRY_COMPILE([
+		#include <linux/timer.h>
+
+		struct my_task_timer {
+			struct timer_list timer;
+			int data;
+		};
+
+		void task_expire(struct timer_list *tl)
+		{
+			struct my_task_timer *task_timer = from_timer(task_timer, tl, timer);
+			task_timer->data = 42;
+		}
+	],[
+		struct my_task_timer task_timer;
+		timer_setup(&task_timer.timer, task_expire, 0);
+	],[
+		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_KERNEL_TIMER_SETUP, 1,
+		    [timer_setup() is available])
+	],[
+		AC_MSG_RESULT(no)
+	])
+
+	AC_MSG_CHECKING([whether timer function expects timer_list])
+
 	ZFS_LINUX_TRY_COMPILE([
 		#include <linux/timer.h>
 		void task_expire(struct timer_list *tl) {}
 	],[
-		#ifndef from_timer
-		#error "No from_timer() macro"
-		#endif
-
-		struct timer_list timer;
-		timer.function = task_expire;
-		timer_setup(&timer, NULL, 0);
+		struct timer_list tl;
+		tl.function = task_expire;
 	],[
 		AC_MSG_RESULT(yes)
 		AC_DEFINE(HAVE_KERNEL_TIMER_FUNCTION_TIMER_LIST, 1,
@@ -28,5 +53,21 @@ AC_DEFUN([ZFS_AC_KERNEL_TIMER_FUNCTION_TIMER_LIST], [
 	],[
 		AC_MSG_RESULT(no)
 	])
+
+	AC_MSG_CHECKING([whether struct timer_list has flags])
+
+	ZFS_LINUX_TRY_COMPILE([
+		#include <linux/timer.h>
+	],[
+		struct timer_list tl;
+		tl.flags = 2;
+	],[
+		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_KERNEL_TIMER_LIST_FLAGS, 1,
+		    [struct timer_list has a flags member])
+	],[
+		AC_MSG_RESULT(no)
+	])
+
 	EXTRA_KCFLAGS="$tmp_flags"
 ])
@@ -12,7 +12,6 @@ AC_DEFUN([ZFS_AC_CONFIG_KERNEL], [
 	ZFS_AC_KERNEL_CTL_NAME
 	ZFS_AC_KERNEL_PDE_DATA
 	ZFS_AC_KERNEL_2ARGS_VFS_FSYNC
-	ZFS_AC_KERNEL_FS_STRUCT_SPINLOCK
 	ZFS_AC_KERNEL_KUIDGID_T
 	ZFS_AC_KERNEL_FALLOCATE
 	ZFS_AC_KERNEL_2ARGS_ZLIB_DEFLATE_WORKSPACESIZE
@@ -36,7 +35,7 @@ AC_DEFUN([ZFS_AC_CONFIG_KERNEL], [
 	ZFS_AC_KERNEL_GROUP_INFO_GID
 	ZFS_AC_KERNEL_WRITE
 	ZFS_AC_KERNEL_READ
-	ZFS_AC_KERNEL_TIMER_FUNCTION_TIMER_LIST
+	ZFS_AC_KERNEL_TIMER_SETUP
 	ZFS_AC_KERNEL_DECLARE_EVENT_CLASS
 	ZFS_AC_KERNEL_CURRENT_BIO_TAIL
 	ZFS_AC_KERNEL_SUPER_USER_NS
@@ -120,8 +120,10 @@ AC_CONFIG_FILES([
 	cmd/dbufstat/Makefile
 	cmd/arc_summary/Makefile
 	cmd/zed/Makefile
+	cmd/zed/zed.d/Makefile
 	cmd/raidz_test/Makefile
 	cmd/zgenhostid/Makefile
+	cmd/zvol_wait/Makefile
 	contrib/Makefile
 	contrib/bash_completion.d/Makefile
 	contrib/dracut/Makefile
@@ -271,6 +273,7 @@ AC_CONFIG_FILES([
 	tests/zfs-tests/tests/functional/cli_user/zfs_list/Makefile
 	tests/zfs-tests/tests/functional/cli_user/zpool_iostat/Makefile
 	tests/zfs-tests/tests/functional/cli_user/zpool_list/Makefile
+	tests/zfs-tests/tests/functional/cli_user/zpool_status/Makefile
 	tests/zfs-tests/tests/functional/compression/Makefile
 	tests/zfs-tests/tests/functional/cp_files/Makefile
 	tests/zfs-tests/tests/functional/ctime/Makefile
@@ -326,6 +329,7 @@ AC_CONFIG_FILES([
 	tests/zfs-tests/tests/functional/snapshot/Makefile
 	tests/zfs-tests/tests/functional/snapused/Makefile
 	tests/zfs-tests/tests/functional/sparse/Makefile
+	tests/zfs-tests/tests/functional/suid/Makefile
 	tests/zfs-tests/tests/functional/alloc_class/Makefile
 	tests/zfs-tests/tests/functional/threadsappend/Makefile
 	tests/zfs-tests/tests/functional/tmpfile/Makefile
@@ -144,7 +144,7 @@ ask_for_password() {

    { flock -s 9;
        # Prompt for password with plymouth, if installed and running.
-        if whereis plymouth >/dev/null 2>&1 && plymouth --ping 2>/dev/null; then
+        if type plymouth >/dev/null 2>&1 && plymouth --ping 2>/dev/null; then
            plymouth ask-for-password \
                --prompt "$ply_prompt" --number-of-tries="$ply_tries" \
                --command="$ply_cmd"
@@ -11,13 +11,18 @@ EXTRA_DIST = \
 	$(top_srcdir)/contrib/initramfs/README.initramfs.markdown

 install-initrdSCRIPTS: $(EXTRA_DIST)
-	for d in conf.d conf-hooks.d hooks scripts scripts/local-top; do \
-	  $(MKDIR_P) $(DESTDIR)$(initrddir)/$$d; \
-	  cp $(top_srcdir)/contrib/initramfs/$$d/zfs \
-	    $(DESTDIR)$(initrddir)/$$d/; \
+	for d in conf.d conf-hooks.d scripts/local-top; do \
+		$(MKDIR_P) $(DESTDIR)$(initrddir)/$$d; \
+		cp $(top_srcdir)/contrib/initramfs/$$d/zfs \
+		    $(DESTDIR)$(initrddir)/$$d/; \
 	done
-	if [ -f etc/init.d/zfs ]; then \
-	  $(MKDIR_P) $(DESTDIR)$(DEFAULT_INITCONF_DIR); \
-	  cp $(top_srcdir)/etc/init.d/zfs \
-	    $(DESTDIR)$(DEFAULT_INITCONF_DIR)/; \
+	for d in hooks scripts; do \
+		$(MKDIR_P) $(DESTDIR)$(initrddir)/$$d; \
+		cp $(top_builddir)/contrib/initramfs/$$d/zfs \
+		    $(DESTDIR)$(initrddir)/$$d/; \
+	done
+	if [ -f $(top_builddir)/etc/init.d/zfs ]; then \
+		$(MKDIR_P) $(DESTDIR)$(DEFAULT_INITCONF_DIR); \
+		cp $(top_builddir)/etc/init.d/zfs \
+		    $(DESTDIR)$(DEFAULT_INITCONF_DIR)/; \
 	fi
@@ -411,29 +411,29 @@ decrypt_fs()

 		# Determine dataset that holds key for root dataset
 		ENCRYPTIONROOT=$(${ZFS} get -H -o value encryptionroot "${fs}")
-		DECRYPT_CMD="${ZFS} load-key '${ENCRYPTIONROOT}'"

 		# If root dataset is encrypted...
 		if ! [ "${ENCRYPTIONROOT}" = "-" ]; then
-
+			TRY_COUNT=3
 			# Prompt with plymouth, if active
 			if [ -e /bin/plymouth ] && /bin/plymouth --ping 2>/dev/null; then
-				plymouth ask-for-password --prompt "Encrypted ZFS password for ${ENCRYPTIONROOT}" \
-					--number-of-tries="3" \
-					--command="${DECRYPT_CMD}"
+				while [ $TRY_COUNT -gt 0 ]; do
+					plymouth ask-for-password --prompt "Encrypted ZFS password for ${ENCRYPTIONROOT}" | \
+						$ZFS load-key "${ENCRYPTIONROOT}" && break
+					TRY_COUNT=$((TRY_COUNT - 1))
+				done

 			# Prompt with systemd, if active 
 			elif [ -e /run/systemd/system ]; then
-				TRY_COUNT=3
 				while [ $TRY_COUNT -gt 0 ]; do
 					systemd-ask-password "Encrypted ZFS password for ${ENCRYPTIONROOT}" --no-tty | \
-						${DECRYPT_CMD} && break
+						$ZFS load-key "${ENCRYPTIONROOT}" && break
 					TRY_COUNT=$((TRY_COUNT - 1))
 				done

 			# Prompt with ZFS tty, otherwise
 			else
-				eval "${DECRYPT_CMD}"
+				$ZFS load-key "${ENCRYPTIONROOT}"
 			fi
 		fi
 	fi
@@ -24,7 +24,7 @@ all-local:
 # files are later created by manually loading the Python modules.
 #
 install-exec-local:
-	$(PYTHON) $(srcdir)/setup.py install \
+	$(PYTHON) $(builddir)/setup.py install \
 	--prefix $(prefix) \
 	--root $(DESTDIR)/ \
 	--install-lib $(pythonsitedir) \
@@ -294,13 +294,6 @@ checksystem()
 	# Just make sure that /dev/zfs is created.
 	udev_trigger

-	if ! [ "$(uname -m)" = "x86_64" ]; then
-		echo "Warning: You're not running 64bit. Currently native zfs in";
-		echo "         Linux is only supported and tested on 64bit.";
-		# should we break here? People doing this should know what they
-		# do, thus i'm not breaking here.
-	fi
-
 	return 0
 }

@@ -71,6 +71,8 @@ process_line() {
  p_readonly="${8}"
  p_setuid="${9}"
  p_nbmand="${10}"
+  p_encroot="${11}"
+  p_keyloc="${12}"

  # Check for canmount=off .
  if [ "${p_canmount}" = "off" ] ; then
@@ -168,6 +170,54 @@ process_line() {
      "${dataset}" >/dev/kmsg
  fi

+  # Minimal pre-requisites to mount a ZFS dataset
+  wants="zfs-import.target"
+  if [ -n "${p_encroot}" ] &&
+      [ "${p_encroot}" != "-" ] ; then
+    keyloadunit="zfs-load-key-$(systemd-escape "${p_encroot}").service"
+    if [ "${p_encroot}" = "${dataset}" ] ; then
+        pathdep=""
+      if [ "${p_keyloc%%://*}" = "file" ] ; then
+        pathdep="RequiresMountsFor='${p_keyloc#file://}'"
+        keyloadcmd="@sbindir@/zfs load-key '${dataset}'"
+      elif [ "${p_keyloc}" = "prompt" ] ; then
+        keyloadcmd="sh -c 'set -eu;"\
+"count=0;"\
+"while [ \$\$count -lt 3 ];do"\
+"  systemd-ask-password --id=\"zfs:${dataset}\""\
+"    \"Enter passphrase for ${dataset}:\"|"\
+"    @sbindir@/zfs load-key \"${dataset}\" && exit 0;"\
+"  count=\$\$((count + 1));"\
+"done;"\
+"exit 1'"
+      else
+        printf 'zfs-mount-generator: (%s) invalid keylocation\n' \
+          "${dataset}" >/dev/kmsg
+      fi
+      cat > "${dest_norm}/${keyloadunit}" << EOF
+# Automatically generated by zfs-mount-generator
+
+[Unit]
+Description=Load ZFS key for ${dataset}
+SourcePath=${cachefile}
+Documentation=man:zfs-mount-generator(8)
+DefaultDependencies=no
+Wants=${wants}
+After=${wants}
+${pathdep}
+
+[Service]
+Type=oneshot
+RemainAfterExit=yes
+ExecStart=${keyloadcmd}
+ExecStop=@sbindir@/zfs unload-key '${dataset}'
+EOF
+    fi
+    # Update the dependencies for the mount file to require the
+    # key-loading unit.
+    wants="${wants} ${keyloadunit}"
+  fi
+
  # If the mountpoint has already been created, give it precedence.
  if [ -e "${dest_norm}/${mountfile}" ] ; then
    printf 'zfs-mount-generator: %s already exists\n' "${mountfile}" \
@@ -183,8 +233,8 @@ process_line() {
 SourcePath=${cachefile}
 Documentation=man:zfs-mount-generator(8)
 Before=local-fs.target zfs-mount.service
-After=zfs-import.target
-Wants=zfs-import.target
+After=${wants}
+Wants=${wants}

 [Mount]
 Where=${p_mountpoint}
@@ -5,4 +5,5 @@ enable zfs-import.target
 enable zfs-mount.service
 enable zfs-share.service
 enable zfs-zed.service
+enable zfs-volume-wait.service
 enable zfs.target
@@ -7,7 +7,9 @@ systemdunit_DATA = \
 	zfs-import-scan.service \
 	zfs-mount.service \
 	zfs-share.service \
+	zfs-volume-wait.service \
 	zfs-import.target \
+	zfs-volumes.target \
 	zfs.target

 EXTRA_DIST = \
@@ -17,6 +19,8 @@ EXTRA_DIST = \
 	$(top_srcdir)/etc/systemd/system/zfs-mount.service.in \
 	$(top_srcdir)/etc/systemd/system/zfs-share.service.in \
 	$(top_srcdir)/etc/systemd/system/zfs-import.target.in \
+	$(top_srcdir)/etc/systemd/system/zfs-volume-wait.service.in \
+	$(top_srcdir)/etc/systemd/system/zfs-volumes.target.in \
 	$(top_srcdir)/etc/systemd/system/zfs.target.in \
 	$(top_srcdir)/etc/systemd/system/50-zfs.preset.in

@@ -5,6 +5,7 @@ After=nfs-server.service nfs-kernel-server.service
 After=smb.service
 Before=rpc-statd-notify.service
 Wants=zfs-mount.service
+After=zfs-mount.service
 PartOf=nfs-server.service nfs-kernel-server.service
 PartOf=smb.service

@@ -0,0 +1,13 @@
+[Unit]
+Description=Wait for ZFS Volume (zvol) links in /dev
+DefaultDependencies=no
+After=systemd-udev-settle.service
+After=zfs-import.target
+
+[Service]
+Type=oneshot
+RemainAfterExit=yes
+ExecStart=@bindir@/zvol_wait
+
+[Install]
+WantedBy=zfs-volumes.target
@@ -0,0 +1,7 @@
+[Unit]
+Description=ZFS volumes are ready
+After=zfs-volume-wait.service
+Requires=zfs-volume-wait.service
+
+[Install]
+WantedBy=zfs.target
@@ -147,6 +147,7 @@ typedef enum zfs_error {
 	EZFS_NO_TRIM,		/* no active trim */
 	EZFS_TRIM_NOTSUP,	/* device does not support trim */
 	EZFS_NO_RESILVER_DEFER,	/* pool doesn't support resilver_defer */
+	EZFS_EXPORT_IN_PROGRESS,	/* currently exporting the pool */
 	EZFS_UNKNOWN
 } zfs_error_t;

@@ -26,8 +26,10 @@
 * USER API:
 *
 * Kernel fpu methods:
- * 	kfpu_begin()
- * 	kfpu_end()
+ *	kfpu_allowed()
+ *	kfpu_initialize()
+ *	kfpu_begin()
+ *	kfpu_end()
 */

 #ifndef _SIMD_AARCH64_H
@@ -26,8 +26,10 @@
 * USER API:
 *
 * Kernel fpu methods:
- * 	kfpu_begin()
- * 	kfpu_end()
+ *	kfpu_allowed()
+ *	kfpu_initialize()
+ *	kfpu_begin()
+ *	kfpu_end()
 *
 * SIMD support:
 *
@@ -37,31 +39,31 @@
 * all relevant feature test functions should be called.
 *
 * Supported features:
- * 	zfs_sse_available()
- * 	zfs_sse2_available()
- * 	zfs_sse3_available()
- * 	zfs_ssse3_available()
- * 	zfs_sse4_1_available()
- * 	zfs_sse4_2_available()
+ *	zfs_sse_available()
+ *	zfs_sse2_available()
+ *	zfs_sse3_available()
+ *	zfs_ssse3_available()
+ *	zfs_sse4_1_available()
+ *	zfs_sse4_2_available()
 *
- * 	zfs_avx_available()
- * 	zfs_avx2_available()
+ *	zfs_avx_available()
+ *	zfs_avx2_available()
 *
- * 	zfs_bmi1_available()
- * 	zfs_bmi2_available()
+ *	zfs_bmi1_available()
+ *	zfs_bmi2_available()
 *
- * 	zfs_avx512f_available()
- * 	zfs_avx512cd_available()
- * 	zfs_avx512er_available()
- * 	zfs_avx512pf_available()
- * 	zfs_avx512bw_available()
- * 	zfs_avx512dq_available()
- * 	zfs_avx512vl_available()
- * 	zfs_avx512ifma_available()
- * 	zfs_avx512vbmi_available()
+ *	zfs_avx512f_available()
+ *	zfs_avx512cd_available()
+ *	zfs_avx512er_available()
+ *	zfs_avx512pf_available()
+ *	zfs_avx512bw_available()
+ *	zfs_avx512dq_available()
+ *	zfs_avx512vl_available()
+ *	zfs_avx512ifma_available()
+ *	zfs_avx512vbmi_available()
 *
 * NOTE(AVX-512VL):	If using AVX-512 instructions with 128Bit registers
- * 			also add zfs_avx512vl_available() to feature check.
+ *			also add zfs_avx512vl_available() to feature check.
 */

 #ifndef _SIMD_X86_H
@@ -190,7 +192,7 @@ typedef struct cpuid_feature_desc {
 * Descriptions of supported instruction sets
 */
 static const cpuid_feature_desc_t cpuid_features[] = {
-	[SSE]		= {1U, 0U,	1U << 25, 	EDX	},
+	[SSE]		= {1U, 0U,	1U << 25,	EDX	},
 	[SSE2]		= {1U, 0U,	1U << 26,	EDX	},
 	[SSE3]		= {1U, 0U,	1U << 0,	ECX	},
 	[SSSE3]		= {1U, 0U,	1U << 9,	ECX	},
@@ -127,6 +127,8 @@ spl_mutex_lockdep_on_maybe(kmutex_t *mp)			\
 })
 /* END CSTYLED */

+#define	NESTED_SINGLE 1
+
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
 #define	mutex_enter_nested(mp, subclass)			\
 {								\
@@ -179,7 +181,4 @@ spl_mutex_lockdep_on_maybe(kmutex_t *mp)			\
 	/* NOTE: do not dereference mp after this point */	\
 }

-int spl_mutex_init(void);
-void spl_mutex_fini(void);
-
 #endif /* _SPL_MUTEX_H */
@@ -29,43 +29,6 @@
 #include <linux/rwsem.h>
 #include <linux/sched.h>

-/* Linux kernel compatibility */
-#if defined(CONFIG_PREEMPT_RT_FULL)
-#define	SPL_RWSEM_SINGLE_READER_VALUE	(1)
-#define	SPL_RWSEM_SINGLE_WRITER_VALUE	(0)
-#elif defined(CONFIG_RWSEM_GENERIC_SPINLOCK)
-#define	SPL_RWSEM_SINGLE_READER_VALUE	(1)
-#define	SPL_RWSEM_SINGLE_WRITER_VALUE	(-1)
-#elif defined(RWSEM_ACTIVE_MASK)
-#define	SPL_RWSEM_SINGLE_READER_VALUE	(RWSEM_ACTIVE_READ_BIAS)
-#define	SPL_RWSEM_SINGLE_WRITER_VALUE	(RWSEM_ACTIVE_WRITE_BIAS)
-#endif
-
-/* Linux 3.16 changed activity to count for rwsem-spinlock */
-#if defined(CONFIG_PREEMPT_RT_FULL)
-#define	RWSEM_COUNT(sem)	sem->read_depth
-#elif defined(HAVE_RWSEM_ACTIVITY)
-#define	RWSEM_COUNT(sem)	sem->activity
-/* Linux 4.8 changed count to an atomic_long_t for !rwsem-spinlock */
-#elif defined(HAVE_RWSEM_ATOMIC_LONG_COUNT)
-#define	RWSEM_COUNT(sem)	atomic_long_read(&(sem)->count)
-#else
-#define	RWSEM_COUNT(sem)	sem->count
-#endif
-
-#if defined(RWSEM_SPINLOCK_IS_RAW)
-#define	spl_rwsem_lock_irqsave(lk, fl)		raw_spin_lock_irqsave(lk, fl)
-#define	spl_rwsem_unlock_irqrestore(lk, fl)	\
-    raw_spin_unlock_irqrestore(lk, fl)
-#define	spl_rwsem_trylock_irqsave(lk, fl)	raw_spin_trylock_irqsave(lk, fl)
-#else
-#define	spl_rwsem_lock_irqsave(lk, fl)		spin_lock_irqsave(lk, fl)
-#define	spl_rwsem_unlock_irqrestore(lk, fl)	spin_unlock_irqrestore(lk, fl)
-#define	spl_rwsem_trylock_irqsave(lk, fl)	spin_trylock_irqsave(lk, fl)
-#endif /* RWSEM_SPINLOCK_IS_RAW */
-
-#define	spl_rwsem_is_locked(rwsem)		rwsem_is_locked(rwsem)
-
 typedef enum {
 	RW_DRIVER	= 2,
 	RW_DEFAULT	= 4,
@@ -78,15 +41,9 @@ typedef enum {
 	RW_READER	= 2
 } krw_t;

-/*
- * If CONFIG_RWSEM_SPIN_ON_OWNER is defined, rw_semaphore will have an owner
- * field, so we don't need our own.
- */
 typedef struct {
 	struct rw_semaphore rw_rwlock;
-#ifndef CONFIG_RWSEM_SPIN_ON_OWNER
 	kthread_t *rw_owner;
-#endif
 #ifdef CONFIG_LOCKDEP
 	krw_type_t	rw_type;
 #endif /* CONFIG_LOCKDEP */
@@ -97,31 +54,19 @@ typedef struct {
 static inline void
 spl_rw_set_owner(krwlock_t *rwp)
 {
-/*
- * If CONFIG_RWSEM_SPIN_ON_OWNER is defined, down_write, up_write,
- * downgrade_write and __init_rwsem will set/clear owner for us.
- */
-#ifndef CONFIG_RWSEM_SPIN_ON_OWNER
 	rwp->rw_owner = current;
-#endif
 }

 static inline void
 spl_rw_clear_owner(krwlock_t *rwp)
 {
-#ifndef CONFIG_RWSEM_SPIN_ON_OWNER
 	rwp->rw_owner = NULL;
-#endif
 }

 static inline kthread_t *
 rw_owner(krwlock_t *rwp)
 {
-#ifdef CONFIG_RWSEM_SPIN_ON_OWNER
-	return (SEM(rwp)->owner);
-#else
 	return (rwp->rw_owner);
-#endif
 }

 #ifdef CONFIG_LOCKDEP
@@ -148,6 +93,11 @@ spl_rw_lockdep_on_maybe(krwlock_t *rwp)			\
 #define	spl_rw_lockdep_on_maybe(rwp)
 #endif /* CONFIG_LOCKDEP */

+static inline int
+RW_LOCK_HELD(krwlock_t *rwp)
+{
+	return (rwsem_is_locked(SEM(rwp)));
+}

 static inline int
 RW_WRITE_HELD(krwlock_t *rwp)
@@ -155,55 +105,10 @@ RW_WRITE_HELD(krwlock_t *rwp)
 	return (rw_owner(rwp) == current);
 }

-static inline int
-RW_LOCK_HELD(krwlock_t *rwp)
-{
-	return (spl_rwsem_is_locked(SEM(rwp)));
-}
-
 static inline int
 RW_READ_HELD(krwlock_t *rwp)
 {
-	if (!RW_LOCK_HELD(rwp))
-		return (0);
-
-	/*
-	 * rw_semaphore cheat sheet:
-	 *
-	 * < 3.16:
-	 * There's no rw_semaphore.owner, so use rwp.owner instead.
-	 * If rwp.owner == NULL then it's a reader
-	 *
-	 * 3.16 - 4.7:
-	 * rw_semaphore.owner added (https://lwn.net/Articles/596656/)
-	 * and CONFIG_RWSEM_SPIN_ON_OWNER introduced.
-	 * If rw_semaphore.owner == NULL then it's a reader
-	 *
-	 * 4.8 - 4.16.16:
-	 * RWSEM_READER_OWNED added as an internal #define.
-	 * (https://lore.kernel.org/patchwork/patch/678590/)
-	 * If rw_semaphore.owner == 1 then it's a reader
-	 *
-	 * 4.16.17 - 4.19:
-	 * RWSEM_OWNER_UNKNOWN introduced as ((struct task_struct *)-1L)
-	 * (https://do-db2.lkml.org/lkml/2018/5/15/985)
-	 * If rw_semaphore.owner == 1 then it's a reader.
-	 *
-	 * 4.20+:
-	 * RWSEM_OWNER_UNKNOWN changed to ((struct task_struct *)-2L)
-	 * (https://lkml.org/lkml/2018/9/6/986)
-	 * If rw_semaphore.owner & 1 then it's a reader, and also the reader's
-	 * task_struct may be embedded in rw_semaphore->owner.
-	 */
-#if	defined(CONFIG_RWSEM_SPIN_ON_OWNER) && defined(RWSEM_OWNER_UNKNOWN)
-	if (RWSEM_OWNER_UNKNOWN == (struct task_struct *)-2L) {
-		/* 4.20+ kernels with CONFIG_RWSEM_SPIN_ON_OWNER */
-		return ((unsigned long) SEM(rwp)->owner & 1);
-	}
-#endif
-
-	/* < 4.20 kernel or !CONFIG_RWSEM_SPIN_ON_OWNER */
-	return (rw_owner(rwp) == NULL || (unsigned long) rw_owner(rwp) == 1);
+	return (RW_LOCK_HELD(rwp) && rw_owner(rwp) == NULL);
 }

 /*
@@ -228,6 +133,12 @@ RW_READ_HELD(krwlock_t *rwp)
 */
 #define	rw_destroy(rwp)		((void) 0)

+/*
+ * Upgrading a rwsem from a reader to a writer is not supported by the
+ * Linux kernel.  The lock must be dropped and reacquired as a writer.
+ */
+#define	rw_tryupgrade(rwp)	RW_WRITE_HELD(rwp)
+
 #define	rw_tryenter(rwp, rw)						\
 ({									\
 	int _rc_ = 0;							\
@@ -285,25 +196,6 @@ RW_READ_HELD(krwlock_t *rwp)
 	downgrade_write(SEM(rwp));					\
 	spl_rw_lockdep_on_maybe(rwp);					\
 })
-
-#define	rw_tryupgrade(rwp)						\
-({									\
-	int _rc_ = 0;							\
-									\
-	if (RW_WRITE_HELD(rwp)) {					\
-		_rc_ = 1;						\
-	} else {							\
-		spl_rw_lockdep_off_maybe(rwp);				\
-		if ((_rc_ = rwsem_tryupgrade(SEM(rwp))))		\
-			spl_rw_set_owner(rwp);				\
-		spl_rw_lockdep_on_maybe(rwp);				\
-	}								\
-	_rc_;								\
-})
 /* END CSTYLED */

-int spl_rw_init(void);
-void spl_rw_fini(void);
-int rwsem_tryupgrade(struct rw_semaphore *rwsem);
-
 #endif /* _SPL_RWLOCK_H */
@@ -72,4 +72,29 @@ usleep_range(unsigned long min, unsigned long max)
 #define	USEC_TO_TICK(us)		usecs_to_jiffies(us)
 #define	NSEC_TO_TICK(ns)		usecs_to_jiffies(ns / NSEC_PER_USEC)

+#ifndef from_timer
+#define	from_timer(var, timer, timer_field) \
+	container_of(timer, typeof(*var), timer_field)
+#endif
+
+#ifdef HAVE_KERNEL_TIMER_FUNCTION_TIMER_LIST
+typedef struct timer_list *spl_timer_list_t;
+#else
+typedef unsigned long spl_timer_list_t;
+#endif
+
+#ifndef HAVE_KERNEL_TIMER_SETUP
+
+static inline void
+timer_setup(struct timer_list *timer, void (*func)(spl_timer_list_t), u32 fl)
+{
+#ifdef HAVE_KERNEL_TIMER_LIST_FLAGS
+	(timer)->flags = fl;
+#endif
+	init_timer(timer);
+	setup_timer(timer, func, (spl_timer_list_t)(timer));
+}
+
+#endif /* HAVE_KERNEL_TIMER_SETUP */
+
 #endif  /* _SPL_TIMER_H */
@@ -182,7 +182,6 @@ extern int vn_space(vnode_t *vp, int cmd, struct flock *bfp, int flag,
 extern file_t *vn_getf(int fd);
 extern void vn_releasef(int fd);
 extern void vn_areleasef(int fd, uf_info_t *fip);
-extern int vn_set_pwd(const char *filename);

 int spl_vn_init(void);
 void spl_vn_fini(void);
@@ -46,6 +46,7 @@ extern "C" {
 */
 #define	DNODE_MUST_BE_ALLOCATED	1
 #define	DNODE_MUST_BE_FREE	2
+#define	DNODE_DRY_RUN		4

 /*
 * dnode_next_offset() flags.
@@ -415,6 +416,7 @@ int dnode_hold_impl(struct objset *dd, uint64_t object, int flag, int dn_slots,
 boolean_t dnode_add_ref(dnode_t *dn, void *ref);
 void dnode_rele(dnode_t *dn, void *ref);
 void dnode_rele_and_unlock(dnode_t *dn, void *tag, boolean_t evicting);
+int dnode_try_claim(objset_t *os, uint64_t object, int slots);
 void dnode_setdirty(dnode_t *dn, dmu_tx_t *tx);
 void dnode_sync(dnode_t *dn, dmu_tx_t *tx);
 void dnode_allocate(dnode_t *dn, dmu_object_type_t ot, int blocksize, int ibs,
@@ -532,11 +534,6 @@ typedef struct dnode_stats {
 	 * a range of dnode slots which would overflow the dnode_phys_t.
 	 */
 	kstat_named_t dnode_hold_free_overflow;
-	/*
-	 * Number of times a dnode_hold(...) was attempted on a dnode
-	 * which had already been unlinked in an earlier txg.
-	 */
-	kstat_named_t dnode_hold_free_txg;
 	/*
 	 * Number of times dnode_free_interior_slots() needed to retry
 	 * acquiring a slot zrl lock due to contention.
@@ -37,9 +37,11 @@ typedef struct zfs_bookmark_phys {
 	uint64_t zbm_creation_txg;	/* birth transaction group */
 	uint64_t zbm_creation_time;	/* bookmark creation time */

-	/* the following fields are reserved for redacted send / recv */
+	/* fields used for redacted send / recv */
 	uint64_t zbm_redaction_obj;	/* redaction list object */
 	uint64_t zbm_flags;		/* ZBM_FLAG_* */
+
+	/* fields used for bookmark written size */
 	uint64_t zbm_referenced_bytes_refd;
 	uint64_t zbm_compressed_bytes_refd;
 	uint64_t zbm_uncompressed_bytes_refd;
@@ -209,7 +209,6 @@ void dsl_dataset_create_crypt_sync(uint64_t dsobj, dsl_dir_t *dd,
    struct dsl_dataset *origin, dsl_crypto_params_t *dcp, dmu_tx_t *tx);
 uint64_t dsl_crypto_key_create_sync(uint64_t crypt, dsl_wrapping_key_t *wkey,
    dmu_tx_t *tx);
-int dmu_objset_clone_crypt_check(dsl_dir_t *parentdd, dsl_dir_t *origindd);
 uint64_t dsl_crypto_key_clone_sync(dsl_dir_t *origindd, dmu_tx_t *tx);
 void dsl_crypto_key_destroy_sync(uint64_t dckobj, dmu_tx_t *tx);

@@ -1318,6 +1318,7 @@ typedef enum {
 	ZFS_ERR_FROM_IVSET_GUID_MISSING,
 	ZFS_ERR_FROM_IVSET_GUID_MISMATCH,
 	ZFS_ERR_SPILL_BLOCK_FLAG_MISSING,
+	ZFS_ERR_EXPORT_IN_PROGRESS,
 } zfs_errno_t;

 /*
@@ -50,6 +50,7 @@ int metaslab_init(metaslab_group_t *, uint64_t, uint64_t, uint64_t,
 void metaslab_fini(metaslab_t *);

 int metaslab_load(metaslab_t *);
+void metaslab_potentially_unload(metaslab_t *, uint64_t);
 void metaslab_unload(metaslab_t *);

 uint64_t metaslab_allocated_space(metaslab_t *);
@@ -89,6 +89,8 @@ void multilist_sublist_insert_head(multilist_sublist_t *, void *);
 void multilist_sublist_insert_tail(multilist_sublist_t *, void *);
 void multilist_sublist_move_forward(multilist_sublist_t *mls, void *obj);
 void multilist_sublist_remove(multilist_sublist_t *, void *);
+int  multilist_sublist_is_empty(multilist_sublist_t *);
+int  multilist_sublist_is_empty_idx(multilist_t *, unsigned int);

 void *multilist_sublist_head(multilist_sublist_t *);
 void *multilist_sublist_tail(multilist_sublist_t *);
@@ -54,8 +54,10 @@ extern "C" {
 */
 typedef struct pathname {
 	char	*pn_buf;		/* underlying storage */
+#if 0 /* unused in ZoL */
 	char	*pn_path;		/* remaining pathname */
 	size_t	pn_pathlen;		/* remaining length */
+#endif
 	size_t	pn_bufsize;		/* total size of pn_buf */
 } pathname_t;

@@ -1104,7 +1104,7 @@ extern uint64_t spa_missing_tvds_allowed(spa_t *spa);
 extern void spa_set_missing_tvds(spa_t *spa, uint64_t missing);
 extern boolean_t spa_top_vdevs_spacemap_addressable(spa_t *spa);
 extern boolean_t spa_multihost(spa_t *spa);
-extern unsigned long spa_get_hostid(void);
+extern uint32_t spa_get_hostid(spa_t *spa);
 extern void spa_activate_allocation_classes(spa_t *, dmu_tx_t *);

 extern int spa_mode(spa_t *spa);
@@ -219,6 +219,7 @@ struct spa {
 	spa_taskqs_t	spa_zio_taskq[ZIO_TYPES][ZIO_TASKQ_TYPES];
 	dsl_pool_t	*spa_dsl_pool;
 	boolean_t	spa_is_initializing;	/* true while opening pool */
+	boolean_t	spa_is_exporting;	/* true while exporting pool */
 	metaslab_class_t *spa_normal_class;	/* normal data class */
 	metaslab_class_t *spa_log_class;	/* intent log data class */
 	metaslab_class_t *spa_special_class;	/* special allocation class */
@@ -394,6 +395,7 @@ struct spa {
 	mmp_thread_t	spa_mmp;		/* multihost mmp thread */
 	list_t		spa_leaf_list;		/* list of leaf vdevs */
 	uint64_t	spa_leaf_list_gen;	/* track leaf_list changes */
+	uint32_t	spa_hostid;		/* cached system hostid */

 	/*
 	 * spa_refcount & spa_config_lock must be the last elements
@@ -14,7 +14,7 @@
 */

 /*
- * Copyright (c) 2014, 2017 by Delphix. All rights reserved.
+ * Copyright (c) 2014, 2019 by Delphix. All rights reserved.
 */

 #ifndef _SYS_VDEV_REMOVAL_H
@@ -81,13 +81,13 @@ extern void spa_vdev_condense_suspend(spa_t *);
 extern int spa_vdev_remove(spa_t *, uint64_t, boolean_t);
 extern void free_from_removing_vdev(vdev_t *, uint64_t, uint64_t);
 extern int spa_removal_get_stats(spa_t *, pool_removal_stat_t *);
-extern void svr_sync(spa_t *spa, dmu_tx_t *tx);
+extern void svr_sync(spa_t *, dmu_tx_t *);
 extern void spa_vdev_remove_suspend(spa_t *);
 extern int spa_vdev_remove_cancel(spa_t *);
-extern void spa_vdev_removal_destroy(spa_vdev_removal_t *svr);
+extern void spa_vdev_removal_destroy(spa_vdev_removal_t *);
+extern uint64_t spa_remove_max_segment(spa_t *);

 extern int vdev_removal_max_span;
-extern int zfs_remove_max_segment;

 #ifdef	__cplusplus
 }
@@ -21,7 +21,7 @@

 /*
 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2013 by Delphix. All rights reserved.
+ * Copyright (c) 2012, 2018 by Delphix. All rights reserved.
 * Copyright 2017 Nexenta Systems, Inc.
 */

@@ -350,6 +350,7 @@ typedef struct zap_cursor {
 	uint64_t zc_serialized;
 	uint64_t zc_hash;
 	uint32_t zc_cd;
+	boolean_t zc_prefetch;
 } zap_cursor_t;

 typedef struct {
@@ -375,7 +376,9 @@ typedef struct {
 * Initialize a zap cursor, pointing to the "first" attribute of the
 * zapobj.  You must _fini the cursor when you are done with it.
 */
-void zap_cursor_init(zap_cursor_t *zc, objset_t *ds, uint64_t zapobj);
+void zap_cursor_init(zap_cursor_t *zc, objset_t *os, uint64_t zapobj);
+void zap_cursor_init_noprefetch(zap_cursor_t *zc, objset_t *os,
+    uint64_t zapobj);
 void zap_cursor_fini(zap_cursor_t *zc);

 /*
@@ -257,6 +257,8 @@ extern void mutex_enter(kmutex_t *mp);
 extern void mutex_exit(kmutex_t *mp);
 extern int mutex_tryenter(kmutex_t *mp);

+#define	NESTED_SINGLE 1
+#define	mutex_enter_nested(mp, class) mutex_enter(mp)
 /*
 * RW locks
 */
@@ -196,6 +196,7 @@ typedef struct znode {
 	uint8_t		z_atime_dirty;	/* atime needs to be synced */
 	uint8_t		z_zn_prefetch;	/* Prefetch znodes? */
 	uint8_t		z_moved;	/* Has this znode been moved? */
+	boolean_t   z_suspended;    /* extra ref from a suspend? */
 	uint_t		z_blksz;	/* block size in bytes */
 	uint_t		z_seq;		/* modification sequence number */
 	uint64_t	z_mapcnt;	/* number of pages mapped to file */
@@ -371,7 +372,7 @@ extern void zfs_log_create(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 extern int zfs_log_create_txtype(zil_create_t, vsecattr_t *vsecp,
    vattr_t *vap);
 extern void zfs_log_remove(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
-    znode_t *dzp, char *name, uint64_t foid);
+    znode_t *dzp, char *name, uint64_t foid, boolean_t unlinked);
 #define	ZFS_NO_OBJECT	0	/* no object id */
 extern void zfs_log_link(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
    znode_t *dzp, znode_t *zp, char *name);
@@ -105,8 +105,7 @@ extern size_t lz4_compress_zfs(void *src, void *dst, size_t s_len, size_t d_len,
    int level);
 extern int lz4_decompress_zfs(void *src, void *dst, size_t s_len, size_t d_len,
    int level);
-extern int lz4_decompress_abd(abd_t *src, void *dst, size_t s_len, size_t d_len,
-    int level);
+
 /*
 * Compress and decompress data if necessary.
 */
@@ -43,6 +43,8 @@ typedef enum {
 	NAME_ERR_RESERVED,		/* entire name is reserved */
 	NAME_ERR_DISKLIKE,		/* reserved disk name (c[0-9].*) */
 	NAME_ERR_TOOLONG,		/* name is too long */
+	NAME_ERR_SELF_REF,		/* reserved self path name ('.') */
+	NAME_ERR_PARENT_REF,		/* reserved parent path name ('..') */
 	NAME_ERR_NO_AT,			/* permission set is missing '@' */
 } namecheck_err_t;

@@ -9,4 +9,4 @@ Version: @VERSION@
 URL: http://zfsonlinux.org
 Requires: libzfs_core
 Cflags: -I${includedir}/libzfs -I${includedir}/libspl
-Libs: -L${libdir} -lzfs
+Libs: -L${libdir} -lzfs -lnvpair
@@ -475,9 +475,10 @@ change_one(zfs_handle_t *zhp, void *data)
 	prop_changelist_t *clp = data;
 	char property[ZFS_MAXPROPLEN];
 	char where[64];
-	prop_changenode_t *cn;
+	prop_changenode_t *cn = NULL;
 	zprop_source_t sourcetype = ZPROP_SRC_NONE;
 	zprop_source_t share_sourcetype = ZPROP_SRC_NONE;
+	int ret = 0;

 	/*
 	 * We only want to unmount/unshare those filesystems that may inherit
@@ -493,8 +494,7 @@ change_one(zfs_handle_t *zhp, void *data)
 	    zfs_prop_get(zhp, clp->cl_prop, property,
 	    sizeof (property), &sourcetype, where, sizeof (where),
 	    B_FALSE) != 0) {
-		zfs_close(zhp);
-		return (0);
+		goto out;
 	}

 	/*
@@ -506,8 +506,7 @@ change_one(zfs_handle_t *zhp, void *data)
 	    zfs_prop_get(zhp, clp->cl_shareprop, property,
 	    sizeof (property), &share_sourcetype, where, sizeof (where),
 	    B_FALSE) != 0) {
-		zfs_close(zhp);
-		return (0);
+		goto out;
 	}

 	if (clp->cl_alldependents || clp->cl_allchildren ||
@@ -518,8 +517,8 @@ change_one(zfs_handle_t *zhp, void *data)
 	    share_sourcetype == ZPROP_SRC_INHERITED))) {
 		if ((cn = zfs_alloc(zfs_get_handle(zhp),
 		    sizeof (prop_changenode_t))) == NULL) {
-			zfs_close(zhp);
-			return (-1);
+			ret = -1;
+			goto out;
 		}

 		cn->cn_handle = zhp;
@@ -541,16 +540,23 @@ change_one(zfs_handle_t *zhp, void *data)
 			uu_avl_insert(clp->cl_tree, cn, idx);
 		} else {
 			free(cn);
-			zfs_close(zhp);
+			cn = NULL;
 		}

 		if (!clp->cl_alldependents)
-			return (zfs_iter_children(zhp, change_one, data));
-	} else {
-		zfs_close(zhp);
+			ret = zfs_iter_children(zhp, change_one, data);
+
+		/*
+		 * If we added the handle to the changelist, we will re-use it
+		 * later so return without closing it.
+		 */
+		if (cn != NULL)
+			return (ret);
 	}

-	return (0);
+out:
+	zfs_close(zhp);
+	return (ret);
 }

 static int
@@ -740,14 +740,6 @@ zfs_crypto_create(libzfs_handle_t *hdl, char *parent_name, nvlist_t *props,
 		pcrypt = ZIO_CRYPT_OFF;
 	}

-	/* Check for encryption being explicitly truned off */
-	if (crypt == ZIO_CRYPT_OFF && pcrypt != ZIO_CRYPT_OFF) {
-		ret = EINVAL;
-		zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
-		    "Invalid encryption value. Dataset must be encrypted."));
-		goto out;
-	}
-
 	/* Get the inherited encryption property if we don't have it locally */
 	if (!local_crypt)
 		crypt = pcrypt;
@@ -849,10 +841,7 @@ int
 zfs_crypto_clone_check(libzfs_handle_t *hdl, zfs_handle_t *origin_zhp,
    char *parent_name, nvlist_t *props)
 {
-	int ret;
 	char errbuf[1024];
-	zfs_handle_t *pzhp = NULL;
-	uint64_t pcrypt, ocrypt;

 	(void) snprintf(errbuf, sizeof (errbuf),
 	    dgettext(TEXT_DOMAIN, "Encryption clone error"));
@@ -865,40 +854,12 @@ zfs_crypto_clone_check(libzfs_handle_t *hdl, zfs_handle_t *origin_zhp,
 	    nvlist_exists(props, zfs_prop_to_name(ZFS_PROP_KEYLOCATION)) ||
 	    nvlist_exists(props, zfs_prop_to_name(ZFS_PROP_ENCRYPTION)) ||
 	    nvlist_exists(props, zfs_prop_to_name(ZFS_PROP_PBKDF2_ITERS))) {
-		ret = EINVAL;
 		zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
 		    "Encryption properties must inherit from origin dataset."));
-		goto out;
+		return (EINVAL);
 	}

-	/* get a reference to parent dataset, should never be NULL */
-	pzhp = make_dataset_handle(hdl, parent_name);
-	if (pzhp == NULL) {
-		zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
-		    "Failed to lookup parent."));
-		return (ENOENT);
-	}
-
-	/* Lookup parent's crypt */
-	pcrypt = zfs_prop_get_int(pzhp, ZFS_PROP_ENCRYPTION);
-	ocrypt = zfs_prop_get_int(origin_zhp, ZFS_PROP_ENCRYPTION);
-
-	/* all children of encrypted parents must be encrypted */
-	if (pcrypt != ZIO_CRYPT_OFF && ocrypt == ZIO_CRYPT_OFF) {
-		ret = EINVAL;
-		zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
-		    "Cannot create unencrypted clone as a child "
-		    "of encrypted parent."));
-		goto out;
-	}
-
-	zfs_close(pzhp);
 	return (0);
-
-out:
-	if (pzhp != NULL)
-		zfs_close(pzhp);
-	return (ret);
 }

 typedef struct loadkeys_cbdata {
@@ -197,6 +197,16 @@ zfs_validate_name(libzfs_handle_t *hdl, const char *path, int type,
 				    "reserved disk name"));
 				break;

+			case NAME_ERR_SELF_REF:
+				zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+				    "self reference, '.' is found in name"));
+				break;
+
+			case NAME_ERR_PARENT_REF:
+				zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+				    "parent reference, '..' is found in name"));
+				break;
+
 			default:
 				zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
 				    "(%d) not defined"), why);
@@ -4107,6 +4117,16 @@ zfs_promote(zfs_handle_t *zhp)

 	if (ret != 0) {
 		switch (ret) {
+		case EACCES:
+			/*
+			 * Promoting encrypted dataset outside its
+			 * encryption root.
+			 */
+			zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+			    "cannot promote dataset outside its "
+			    "encryption root"));
+			return (zfs_error(hdl, EZFS_EXISTS, errbuf));
+
 		case EEXIST:
 			/* There is a conflicting snapshot name. */
 			zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
@@ -4470,8 +4490,6 @@ zfs_rename(zfs_handle_t *zhp, const char *target, boolean_t recursive,
 	zfs_cmd_t zc = {"\0"};
 	char *delim;
 	prop_changelist_t *cl = NULL;
-	zfs_handle_t *zhrp = NULL;
-	char *parentname = NULL;
 	char parent[ZFS_MAX_DATASET_NAME_LEN];
 	libzfs_handle_t *hdl = zhp->zfs_hdl;
 	char errbuf[1024];
@@ -4566,7 +4584,8 @@ zfs_rename(zfs_handle_t *zhp, const char *target, boolean_t recursive,
 	}

 	if (recursive) {
-		parentname = zfs_strdup(zhp->zfs_hdl, zhp->zfs_name);
+		zfs_handle_t *zhrp;
+		char *parentname = zfs_strdup(zhp->zfs_hdl, zhp->zfs_name);
 		if (parentname == NULL) {
 			ret = -1;
 			goto error;
@@ -4574,10 +4593,12 @@ zfs_rename(zfs_handle_t *zhp, const char *target, boolean_t recursive,
 		delim = strchr(parentname, '@');
 		*delim = '\0';
 		zhrp = zfs_open(zhp->zfs_hdl, parentname, ZFS_TYPE_DATASET);
+		free(parentname);
 		if (zhrp == NULL) {
 			ret = -1;
 			goto error;
 		}
+		zfs_close(zhrp);
 	} else if (zhp->zfs_type != ZFS_TYPE_SNAPSHOT) {
 		if ((cl = changelist_gather(zhp, ZFS_PROP_NAME,
 		    CL_GATHER_ITER_MOUNTED,
@@ -4621,16 +4642,9 @@ zfs_rename(zfs_handle_t *zhp, const char *target, boolean_t recursive,
 			    "with the new name"));
 			(void) zfs_error(hdl, EZFS_EXISTS, errbuf);
 		} else if (errno == EACCES) {
-			if (zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION) ==
-			    ZIO_CRYPT_OFF) {
-				zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
-				    "cannot rename an unencrypted dataset to "
-				    "be a decendent of an encrypted one"));
-			} else {
-				zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
-				    "cannot move encryption child outside of "
-				    "its encryption root"));
-			}
+			zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+			    "cannot move encrypted child outside of "
+			    "its encryption root"));
 			(void) zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf);
 		} else {
 			(void) zfs_standard_error(zhp->zfs_hdl, errno, errbuf);
@@ -4650,12 +4664,6 @@ zfs_rename(zfs_handle_t *zhp, const char *target, boolean_t recursive,
 	}

 error:
-	if (parentname != NULL) {
-		free(parentname);
-	}
-	if (zhrp != NULL) {
-		zfs_close(zhrp);
-	}
 	if (cl != NULL) {
 		changelist_free(cl);
 	}
@@ -1302,12 +1302,14 @@ mountpoint_cmp(const void *arga, const void *argb)
 }

 /*
- * Return true if path2 is a child of path1.
+ * Return true if path2 is a child of path1 or path2 equals path1 or
+ * path1 is "/" (path2 is always a child of "/").
 */
 static boolean_t
 libzfs_path_contains(const char *path1, const char *path2)
 {
-	return (strstr(path2, path1) == path2 && path2[strlen(path1)] == '/');
+	return (strcmp(path1, path2) == 0 || strcmp(path1, "/") == 0 ||
+	    (strstr(path2, path1) == path2 && path2[strlen(path1)] == '/'));
 }

 /*
@@ -2827,7 +2827,7 @@ recv_fix_encryption_hierarchy(libzfs_handle_t *hdl, const char *destname,
 		is_clone = zhp->zfs_dmustats.dds_origin[0] != '\0';
 		(void) zfs_crypto_get_encryption_root(zhp, &is_encroot, NULL);

-		/* we don't need to do anything for unencrypted filesystems */
+		/* we don't need to do anything for unencrypted datasets */
 		if (crypt == ZIO_CRYPT_OFF) {
 			zfs_close(zhp);
 			continue;
@@ -3992,11 +3992,18 @@ zfs_receive_one(libzfs_handle_t *hdl, int infd, const char *tosnap,
 		}
 	} else {
 		/*
-		 * if the fs does not exist, look for it based on the
-		 * fromsnap GUID
+		 * If the fs does not exist, look for it based on the
+		 * fromsnap GUID.
 		 */
-		(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
-		    "cannot receive incremental stream"));
+		if (resuming) {
+			(void) snprintf(errbuf, sizeof (errbuf),
+			    dgettext(TEXT_DOMAIN,
+			    "cannot receive resume stream"));
+		} else {
+			(void) snprintf(errbuf, sizeof (errbuf),
+			    dgettext(TEXT_DOMAIN,
+			    "cannot receive incremental stream"));
+		}

 		(void) strcpy(name, destsnap);
 		*strchr(name, '@') = '\0';
@@ -4210,34 +4217,6 @@ zfs_receive_one(libzfs_handle_t *hdl, int infd, const char *tosnap,
 			goto out;
 		}

-		/*
-		 * It is invalid to receive a properties stream that was
-		 * unencrypted on the send side as a child of an encrypted
-		 * parent. Technically there is nothing preventing this, but
-		 * it would mean that the encryption=off property which is
-		 * locally set on the send side would not be received correctly.
-		 * We can infer encryption=off if the stream is not raw and
-		 * properties were included since the send side will only ever
-		 * send the encryption property in a raw nvlist header. This
-		 * check will be avoided if the user specifically overrides
-		 * the encryption property on the command line.
-		 */
-		if (!raw && rcvprops != NULL &&
-		    !nvlist_exists(cmdprops,
-		    zfs_prop_to_name(ZFS_PROP_ENCRYPTION))) {
-			uint64_t crypt;
-
-			crypt = zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION);
-
-			if (crypt != ZIO_CRYPT_OFF) {
-				zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
-				    "parent '%s' must not be encrypted to "
-				    "receive unenecrypted property"), name);
-				err = zfs_error(hdl, EZFS_BADPROP, errbuf);
-				zfs_close(zhp);
-				goto out;
-			}
-		}
 		zfs_close(zhp);

 		newfs = B_TRUE;
@@ -4274,6 +4253,24 @@ zfs_receive_one(libzfs_handle_t *hdl, int infd, const char *tosnap,
 	    &oxprops, &wkeydata, &wkeylen, errbuf)) != 0)
 		goto out;

+	/*
+	 * When sending with properties (zfs send -p), the encryption property
+	 * is not included because it is a SETONCE property and therefore
+	 * treated as read only. However, we are always able to determine its
+	 * value because raw sends will include it in the DRR_BDEGIN payload
+	 * and non-raw sends with properties are not allowed for encrypted
+	 * datasets. Therefore, if this is a non-raw properties stream, we can
+	 * infer that the value should be ZIO_CRYPT_OFF and manually add that
+	 * to the received properties.
+	 */
+	if (stream_wantsnewfs && !raw && rcvprops != NULL &&
+	    !nvlist_exists(cmdprops, zfs_prop_to_name(ZFS_PROP_ENCRYPTION))) {
+		if (oxprops == NULL)
+			oxprops = fnvlist_alloc();
+		fnvlist_add_uint64(oxprops,
+		    zfs_prop_to_name(ZFS_PROP_ENCRYPTION), ZIO_CRYPT_OFF);
+	}
+
 	err = ioctl_err = lzc_receive_with_cmdprops(destsnap, rcvprops,
 	    oxprops, wkeydata, wkeylen, origin, flags->force, flags->resumable,
 	    raw, infd, drr_noswap, cleanup_fd, &read_bytes, &errflags,
@@ -4428,14 +4425,15 @@ zfs_receive_one(libzfs_handle_t *hdl, int infd, const char *tosnap,
 			*cp = '@';
 			break;
 		case EINVAL:
-			if (flags->resumable)
+			if (flags->resumable) {
 				zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
 				    "kernel modules must be upgraded to "
 				    "receive this stream."));
-			if (embedded && !raw)
+			} else if (embedded && !raw) {
 				zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
 				    "incompatible embedded data stream "
 				    "feature with encrypted receive."));
+			}
 			(void) zfs_error(hdl, EZFS_BADSTREAM, errbuf);
 			break;
 		case ECKSUM:
@@ -303,6 +303,8 @@ libzfs_error_description(libzfs_handle_t *hdl)
 	case EZFS_NO_RESILVER_DEFER:
 		return (dgettext(TEXT_DOMAIN, "this action requires the "
 		    "resilver_defer feature"));
+	case EZFS_EXPORT_IN_PROGRESS:
+		return (dgettext(TEXT_DOMAIN, "pool export in progress"));
 	case EZFS_UNKNOWN:
 		return (dgettext(TEXT_DOMAIN, "unknown error"));
 	default:
@@ -598,6 +600,9 @@ zpool_standard_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...)
 	case ZFS_ERR_VDEV_TOO_BIG:
 		zfs_verror(hdl, EZFS_VDEV_TOO_BIG, fmt, ap);
 		break;
+	case ZFS_ERR_EXPORT_IN_PROGRESS:
+		zfs_verror(hdl, EZFS_EXPORT_IN_PROGRESS, fmt, ap);
+		break;
 	case ZFS_ERR_IOC_CMD_UNAVAIL:
 		zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "the loaded zfs "
 		    "module does not support this operation. A reboot may "
@@ -1134,7 +1139,7 @@ int
 zcmd_alloc_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, size_t len)
 {
 	if (len == 0)
-		len = 16 * 1024;
+		len = 256 * 1024;
 	zc->zc_nvlist_dst_size = len;
 	zc->zc_nvlist_dst =
 	    (uint64_t)(uintptr_t)zfs_alloc(hdl, zc->zc_nvlist_dst_size);
@@ -52,7 +52,7 @@
 *
 *  - Thin Layer.  libzfs_core is a thin layer, marshaling arguments
 *  to/from the kernel ioctls.  There is generally a 1:1 correspondence
- *  between libzfs_core functions and ioctls to /dev/zfs.
+ *  between libzfs_core functions and ioctls to ZFS_DEV.
 *
 *  - Clear Atomicity.  Because libzfs_core functions are generally 1:1
 *  with kernel ioctls, and kernel ioctls are general atomic, each
@@ -135,7 +135,7 @@ libzfs_core_init(void)
 {
 	(void) pthread_mutex_lock(&g_lock);
 	if (g_refcount == 0) {
-		g_fd = open("/dev/zfs", O_RDWR);
+		g_fd = open(ZFS_DEV, O_RDWR);
 		if (g_fd < 0) {
 			(void) pthread_mutex_unlock(&g_lock);
 			return (errno);
@@ -499,7 +499,7 @@ lzc_sync(const char *pool_name, nvlist_t *innvl, nvlist_t **outnvl)
 * The snapshots must all be in the same pool.
 * The value is the name of the hold (string type).
 *
- * If cleanup_fd is not -1, it must be the result of open("/dev/zfs", O_EXCL).
+ * If cleanup_fd is not -1, it must be the result of open(ZFS_DEV, O_EXCL).
 * In this case, when the cleanup_fd is closed (including on process
 * termination), the holds will be released.  If the system is shut down
 * uncleanly, the holds will be released when the pool is next opened
@@ -223,7 +223,7 @@ pool_active(void *unused, const char *name, uint64_t guid,
 	 * Use ZFS_IOC_POOL_SYNC to confirm if a pool is active
 	 */

-	fd = open("/dev/zfs", O_RDWR);
+	fd = open(ZFS_DEV, O_RDWR);
 	if (fd < 0)
 		return (-1);

@@ -1,4 +1,4 @@
-dist_man_MANS = zhack.1 ztest.1 raidz_test.1
+dist_man_MANS = zhack.1 ztest.1 raidz_test.1 zvol_wait.1
 EXTRA_DIST = cstyle.1

 install-data-local:
@@ -0,0 +1,21 @@
+.Dd July 5, 2019
+.Dt ZVOL_WAIT 1 SMM
+.Os Linux
+.Sh NAME
+.Nm zvol_wait
+.Nd Wait for ZFS volume links in
+.Em /dev
+to be created.
+.Sh SYNOPSIS
+.Nm
+.Sh DESCRIPTION
+When a ZFS pool is imported, ZFS will register each ZFS volume
+(zvol) as a disk device with the system. As the disks are registered,
+.Xr \fBudev 7\fR
+will asynchronously create symlinks under
+.Em /dev/zvol
+using the zvol's name.
+.Nm
+will wait for all those symlinks to be created before returning.
+.Sh SEE ALSO
+.Xr \fBudev 7\fR
@@ -104,6 +104,18 @@ to a log2 fraction of the target arc size.
 Default value: \fB6\fR.
 .RE

+.sp
+.ne 2
+.na
+\fBdmu_prefetch_max\fR (int)
+.ad
+.RS 12n
+Limit the amount we can prefetch with one call to this amount (in bytes).
+This helps to limit the amount of memory that can be used by prefetching.
+.sp
+Default value: \fB134,217,728\fR (128MB).
+.RE
+
 .sp
 .ne 2
 .na
@@ -502,6 +514,19 @@ regular reads (but there's no reason it has to be the same).
 Default value: \fB32,768\fR.
 .RE

+.sp
+.ne 2
+.na
+\fBzap_iterate_prefetch\fR (int)
+.ad
+.RS 12n
+If this is set, when we start iterating over a ZAP object, zfs will prefetch
+the entire object (all leaf blocks).  However, this is limited by
+\fBdmu_prefetch_max\fR.
+.sp
+Use \fB1\fR for on (default) and \fB0\fR for off.
+.RE
+
 .sp
 .ne 2
 .na
@@ -1817,7 +1842,7 @@ this value. If a metaslab group exceeds this threshold then it will be
 skipped unless all metaslab groups within the metaslab class have also
 crossed this threshold.
 .sp
-Default value: \fB85\fR.
+Default value: \fB95\fR.
 .RE

 .sp
@@ -2169,6 +2194,33 @@ pool cannot be returned to a healthy state prior to removing the device.
 Default value: \fB0\fR.
 .RE

+.sp
+.ne 2
+.na
+\fBzfs_removal_suspend_progress\fR (int)
+.ad
+.RS 12n
+.sp
+This is used by the test suite so that it can ensure that certain actions
+happen while in the middle of a removal.
+.sp
+Default value: \fB0\fR.
+.RE
+
+.sp
+.ne 2
+.na
+\fBzfs_remove_max_segment\fR (int)
+.ad
+.RS 12n
+.sp
+The largest contiguous segment that we will attempt to allocate when removing
+a device.  This can be no larger than 16MB.  If there is a performance
+problem with attempting to allocate large blocks, consider decreasing this.
+.sp
+Default value: \fB16,777,216\fR (16MB).
+.RE
+
 .sp
 .ne 2
 .na
@@ -2419,9 +2471,21 @@ Default value: \fB25\fR.
 \fBzfs_sync_pass_dont_compress\fR (int)
 .ad
 .RS 12n
-Don't compress starting in this pass
+Starting in this sync pass, we disable compression (including of metadata). 
+With the default setting, in practice, we don't have this many sync passes,
+so this has no effect.
 .sp
-Default value: \fB5\fR.
+The original intent was that disabling compression would help the sync passes
+to converge. However, in practice disabling compression increases the average
+number of sync passes, because when we turn compression off, a lot of block's
+size will change and thus we have to re-allocate (not overwrite) them. It
+also increases the number of 128KB allocations (e.g. for indirect blocks and
+spacemaps) because these will not be compressed. The 128K allocations are
+especially detrimental to performance on highly fragmented systems, which may
+have very few free segments of this size, and may need to load new metaslabs
+to satisfy 128K allocations.
+.sp
+Default value: \fB8\fR.
 .RE

 .sp
@@ -26,7 +26,7 @@ information on ZFS mountpoints must be stored separately. The output
 of the command
 .PP
 .RS 4
-zfs list -H -o name,mountpoint,canmount,atime,relatime,devices,exec,readonly,setuid,nbmand
+zfs list -H -o name,mountpoint,canmount,atime,relatime,devices,exec,readonly,setuid,nbmand,encroot,keylocation
 .RE
 .PP
 for datasets that should be mounted by systemd, should be kept
@@ -1,11 +1,11 @@
-subdir-m += avl
-subdir-m += icp
-subdir-m += lua
-subdir-m += nvpair
-subdir-m += spl
-subdir-m += unicode
-subdir-m += zcommon
-subdir-m += zfs
+obj-m += avl/
+obj-m += icp/
+obj-m += lua/
+obj-m += nvpair/
+obj-m += spl/
+obj-m += unicode/
+obj-m += zcommon/
+obj-m += zfs/

 INSTALL_MOD_DIR ?= extra

@@ -60,14 +60,15 @@ modules_install:
 modules_uninstall:
 	@# Uninstall the kernel modules
 	kmoddir=$(DESTDIR)$(INSTALL_MOD_PATH)/lib/modules/@LINUX_VERSION@
-	list='$(subdir-m)'; for subdir in $$list; do \
-		$(RM) -R $$kmoddir/$(INSTALL_MOD_DIR)/$$subdir; \
+	list='$(obj-m)'; for objdir in $$list; do \
+		$(RM) -R $$kmoddir/$(INSTALL_MOD_DIR)/$$objdir; \
 	done

 distdir:
-	list='$(subdir-m)'; for subdir in $$list; do \
-		(cd @top_srcdir@/module && find $$subdir -name '*.c' -o -name '*.h' -o -name '*.S' |\
-		xargs cp --parents -t $$distdir); \
+	list='$(obj-m)'; for objdir in $$list; do \
+		(cd @top_srcdir@/module && find $$objdir \
+		-name '*.c' -o -name '*.h' -o -name '*.S' | \
+		xargs cp --parents -t @abs_top_builddir@/module/$$distdir); \
 	done

 distclean maintainer-clean: clean
@@ -303,16 +303,21 @@ aes_impl_init(void)
 	}
 	aes_supp_impl_cnt = c;

-	/* set fastest implementation. assume hardware accelerated is fastest */
+	/*
+	 * Set the fastest implementation given the assumption that the
+	 * hardware accelerated version is the fastest.
+	 */
 #if defined(__x86_64)
 #if defined(HAVE_AES)
-	if (aes_aesni_impl.is_supported())
+	if (aes_aesni_impl.is_supported()) {
 		memcpy(&aes_fastest_impl, &aes_aesni_impl,
 		    sizeof (aes_fastest_impl));
-	else
+	} else
 #endif
+	{
 		memcpy(&aes_fastest_impl, &aes_x86_64_impl,
 		    sizeof (aes_fastest_impl));
+	}
 #else
 	memcpy(&aes_fastest_impl, &aes_generic_impl,
 	    sizeof (aes_fastest_impl));
@@ -646,7 +646,7 @@ const gcm_impl_ops_t *gcm_all_impl[] = {
 /* Indicate that benchmark has been completed */
 static boolean_t gcm_impl_initialized = B_FALSE;

-/* Select aes implementation */
+/* Select GCM implementation */
 #define	IMPL_FASTEST	(UINT32_MAX)
 #define	IMPL_CYCLE	(UINT32_MAX-1)

@@ -713,13 +713,15 @@ gcm_impl_init(void)

 	/* set fastest implementation. assume hardware accelerated is fastest */
 #if defined(__x86_64) && defined(HAVE_PCLMULQDQ)
-	if (gcm_pclmulqdq_impl.is_supported())
+	if (gcm_pclmulqdq_impl.is_supported()) {
 		memcpy(&gcm_fastest_impl, &gcm_pclmulqdq_impl,
 		    sizeof (gcm_fastest_impl));
-	else
+	} else
 #endif
+	{
 		memcpy(&gcm_fastest_impl, &gcm_generic_impl,
 		    sizeof (gcm_fastest_impl));
+	}

 	strcpy(gcm_fastest_impl.name, "fastest");

@@ -742,7 +744,7 @@ static const struct {
 * If we are called before init(), user preference will be saved in
 * user_sel_impl, and applied in later init() call. This occurs when module
 * parameter is specified on module load. Otherwise, directly update
- * icp_aes_impl.
+ * icp_gcm_impl.
 *
 * @val		Name of gcm implementation to use
 * @param	Unused.
@@ -162,7 +162,7 @@ typedef enum aes_mech_type {
 #endif /* _AES_IMPL */

 /*
- * Methods used to define aes implementation
+ * Methods used to define AES implementation
 *
 * @aes_gen_f Key generation
 * @aes_enc_f Function encrypts one block
@@ -37,12 +37,12 @@ extern "C" {
 #include <sys/crypto/common.h>

 /*
- * Methods used to define gcm implementation
+ * Methods used to define GCM implementation
 *
 * @gcm_mul_f Perform carry-less multiplication
 * @gcm_will_work_f Function tests whether implementation will function
 */
-typedef void 		(*gcm_mul_f)(uint64_t *, uint64_t *, uint64_t *);
+typedef void		(*gcm_mul_f)(uint64_t *, uint64_t *, uint64_t *);
 typedef boolean_t	(*gcm_will_work_f)(void);

 #define	GCM_IMPL_NAME_MAX (16)
@@ -61,7 +61,7 @@
 #elif defined(__mips__)
 #define JMP_BUF_CNT	12
 #elif defined(__s390x__)
-#define JMP_BUF_CNT	9
+#define JMP_BUF_CNT	18
 #else
 #define	JMP_BUF_CNT	1
 #endif
@@ -431,9 +431,12 @@ static int llex (LexState *ls, SemInfo *seminfo) {
        if (sep >= 0) {
          read_long_string(ls, seminfo, sep);
          return TK_STRING;
-        }
-        else if (sep == -1) return '[';
-        else lexerror(ls, "invalid long string delimiter", TK_STRING);
+        } else if (sep == -1) {
+		return '[';
+        } else {
+		lexerror(ls, "invalid long string delimiter", TK_STRING);
+		break;
+	}
      }
      case '=': {
        next(ls);
@@ -16,10 +16,8 @@ $(MODULE)-objs += spl-kmem.o
 $(MODULE)-objs += spl-kmem-cache.o
 $(MODULE)-objs += spl-kobj.o
 $(MODULE)-objs += spl-kstat.o
-$(MODULE)-objs += spl-mutex.o
 $(MODULE)-objs += spl-proc.o
 $(MODULE)-objs += spl-procfs-list.o
-$(MODULE)-objs += spl-rwlock.o
 $(MODULE)-objs += spl-taskq.o
 $(MODULE)-objs += spl-thread.o
 $(MODULE)-objs += spl-tsd.o
@@ -154,26 +154,39 @@ EXPORT_SYMBOL(__cv_wait_sig);
 #if defined(HAVE_IO_SCHEDULE_TIMEOUT)
 #define	spl_io_schedule_timeout(t)	io_schedule_timeout(t)
 #else
+
+struct spl_task_timer {
+	struct timer_list timer;
+	struct task_struct *task;
+};
+
 static void
-__cv_wakeup(unsigned long data)
+__cv_wakeup(spl_timer_list_t t)
 {
-	wake_up_process((struct task_struct *)data);
+	struct timer_list *tmr = (struct timer_list *)t;
+	struct spl_task_timer *task_timer = from_timer(task_timer, tmr, timer);
+
+	wake_up_process(task_timer->task);
 }

 static long
 spl_io_schedule_timeout(long time_left)
 {
 	long expire_time = jiffies + time_left;
-	struct timer_list timer;
+	struct spl_task_timer task_timer;
+	struct timer_list *timer = &task_timer.timer;

-	init_timer(&timer);
-	setup_timer(&timer, __cv_wakeup, (unsigned long)current);
-	timer.expires = expire_time;
-	add_timer(&timer);
+	task_timer.task = current;
+
+	timer_setup(timer, __cv_wakeup, 0);
+
+	timer->expires = expire_time;
+	add_timer(timer);

 	io_schedule();

-	del_timer_sync(&timer);
+	del_timer_sync(timer);
+
 	time_left = expire_time - jiffies;

 	return (time_left < 0 ? 0 : time_left);
@@ -694,51 +694,41 @@ spl_init(void)
 	if ((rc = spl_kvmem_init()))
 		goto out1;

-	if ((rc = spl_mutex_init()))
+	if ((rc = spl_tsd_init()))
 		goto out2;

-	if ((rc = spl_rw_init()))
+	if ((rc = spl_taskq_init()))
 		goto out3;

-	if ((rc = spl_tsd_init()))
+	if ((rc = spl_kmem_cache_init()))
 		goto out4;

-	if ((rc = spl_taskq_init()))
+	if ((rc = spl_vn_init()))
 		goto out5;

-	if ((rc = spl_kmem_cache_init()))
+	if ((rc = spl_proc_init()))
 		goto out6;

-	if ((rc = spl_vn_init()))
+	if ((rc = spl_kstat_init()))
 		goto out7;

-	if ((rc = spl_proc_init()))
-		goto out8;
-
-	if ((rc = spl_kstat_init()))
-		goto out9;
-
 	if ((rc = spl_zlib_init()))
-		goto out10;
+		goto out8;

 	return (rc);

-out10:
-	spl_kstat_fini();
-out9:
-	spl_proc_fini();
 out8:
-	spl_vn_fini();
+	spl_kstat_fini();
 out7:
-	spl_kmem_cache_fini();
+	spl_proc_fini();
 out6:
-	spl_taskq_fini();
+	spl_vn_fini();
 out5:
-	spl_tsd_fini();
+	spl_kmem_cache_fini();
 out4:
-	spl_rw_fini();
+	spl_taskq_fini();
 out3:
-	spl_mutex_fini();
+	spl_tsd_fini();
 out2:
 	spl_kvmem_fini();
 out1:
@@ -755,8 +745,6 @@ spl_fini(void)
 	spl_kmem_cache_fini();
 	spl_taskq_fini();
 	spl_tsd_fini();
-	spl_rw_fini();
-	spl_mutex_fini();
 	spl_kvmem_fini();
 }

@@ -180,7 +180,8 @@ spl_kmem_alloc_impl(size_t size, int flags, int node)
 		 */
 		if ((size > spl_kmem_alloc_max) || use_vmem) {
 			if (flags & KM_VMEM) {
-				ptr = __vmalloc(size, lflags, PAGE_KERNEL);
+				ptr = __vmalloc(size, lflags | __GFP_HIGHMEM,
+				    PAGE_KERNEL);
 			} else {
 				return (NULL);
 			}
@@ -1,30 +0,0 @@
-/*
- *  Copyright (C) 2007-2010 Lawrence Livermore National Security, LLC.
- *  Copyright (C) 2007 The Regents of the University of California.
- *  Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
- *  Written by Brian Behlendorf <behlendorf1@llnl.gov>.
- *  UCRL-CODE-235197
- *
- *  This file is part of the SPL, Solaris Porting Layer.
- *  For details, see <http://zfsonlinux.org/>.
- *
- *  The SPL is free software; you can redistribute it and/or modify it
- *  under the terms of the GNU General Public License as published by the
- *  Free Software Foundation; either version 2 of the License, or (at your
- *  option) any later version.
- *
- *  The SPL is distributed in the hope that it will be useful, but WITHOUT
- *  ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- *  FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
- *  for more details.
- *
- *  You should have received a copy of the GNU General Public License along
- *  with the SPL.  If not, see <http://www.gnu.org/licenses/>.
- *
- *  Solaris Porting Layer (SPL) Mutex Implementation.
- */
-
-#include <sys/mutex.h>
-
-int spl_mutex_init(void) { return 0; }
-void spl_mutex_fini(void) { }
@@ -1,132 +0,0 @@
-/*
- *  Copyright (C) 2007-2010 Lawrence Livermore National Security, LLC.
- *  Copyright (C) 2007 The Regents of the University of California.
- *  Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
- *  Written by Brian Behlendorf <behlendorf1@llnl.gov>.
- *  UCRL-CODE-235197
- *
- *  This file is part of the SPL, Solaris Porting Layer.
- *  For details, see <http://zfsonlinux.org/>.
- *
- *  The SPL is free software; you can redistribute it and/or modify it
- *  under the terms of the GNU General Public License as published by the
- *  Free Software Foundation; either version 2 of the License, or (at your
- *  option) any later version.
- *
- *  The SPL is distributed in the hope that it will be useful, but WITHOUT
- *  ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- *  FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
- *  for more details.
- *
- *  You should have received a copy of the GNU General Public License along
- *  with the SPL.  If not, see <http://www.gnu.org/licenses/>.
- *
- *  Solaris Porting Layer (SPL) Reader/Writer Lock Implementation.
- */
-
-#include <sys/rwlock.h>
-#include <linux/module.h>
-
-#if defined(CONFIG_PREEMPT_RT_FULL)
-
-#include <linux/rtmutex.h>
-#define	RT_MUTEX_OWNER_MASKALL	1UL
-
-static int
-__rwsem_tryupgrade(struct rw_semaphore *rwsem)
-{
-#if defined(READER_BIAS) && defined(WRITER_BIAS)
-	/*
-	 * After the 4.9.20-rt16 kernel the realtime patch series lifted the
-	 * single reader restriction.  While this could be accommodated by
-	 * adding additional compatibility code assume the rwsem can never
-	 * be upgraded.  All caller must already cleanly handle this case.
-	 */
-	return (0);
-#else
-	ASSERT((struct task_struct *)
-	    ((unsigned long)rwsem->lock.owner & ~RT_MUTEX_OWNER_MASKALL) ==
-	    current);
-
-	/*
-	 * Prior to 4.9.20-rt16 kernel the realtime patch series, rwsem is
-	 * implemented as a single mutex held by readers and writers alike.
-	 * However, this implementation would prevent a thread from taking
-	 * a read lock twice, as the mutex would already be locked on
-	 * the second attempt. Therefore the implementation allows a
-	 * single thread to take a rwsem as read lock multiple times
-	 * tracking that nesting as read_depth counter.
-	 */
-	if (rwsem->read_depth <= 1) {
-		/*
-		 * In case, the current thread has not taken the lock
-		 * more than once as read lock, we can allow an
-		 * upgrade to a write lock. rwsem_rt.h implements
-		 * write locks as read_depth == 0.
-		 */
-		rwsem->read_depth = 0;
-		return (1);
-	}
-	return (0);
-#endif
-}
-#elif defined(CONFIG_RWSEM_GENERIC_SPINLOCK)
-static int
-__rwsem_tryupgrade(struct rw_semaphore *rwsem)
-{
-	int ret = 0;
-	unsigned long flags;
-	spl_rwsem_lock_irqsave(&rwsem->wait_lock, flags);
-	if (RWSEM_COUNT(rwsem) == SPL_RWSEM_SINGLE_READER_VALUE &&
-	    list_empty(&rwsem->wait_list)) {
-		ret = 1;
-		RWSEM_COUNT(rwsem) = SPL_RWSEM_SINGLE_WRITER_VALUE;
-	}
-	spl_rwsem_unlock_irqrestore(&rwsem->wait_lock, flags);
-	return (ret);
-}
-#elif defined(RWSEM_ACTIVE_MASK)
-#if defined(HAVE_RWSEM_ATOMIC_LONG_COUNT)
-static int
-__rwsem_tryupgrade(struct rw_semaphore *rwsem)
-{
-	long val;
-	val = atomic_long_cmpxchg(&rwsem->count, SPL_RWSEM_SINGLE_READER_VALUE,
-	    SPL_RWSEM_SINGLE_WRITER_VALUE);
-	return (val == SPL_RWSEM_SINGLE_READER_VALUE);
-}
-#else
-static int
-__rwsem_tryupgrade(struct rw_semaphore *rwsem)
-{
-	typeof(rwsem->count) val;
-	val = cmpxchg(&rwsem->count, SPL_RWSEM_SINGLE_READER_VALUE,
-	    SPL_RWSEM_SINGLE_WRITER_VALUE);
-	return (val == SPL_RWSEM_SINGLE_READER_VALUE);
-}
-#endif
-#else
-static int
-__rwsem_tryupgrade(struct rw_semaphore *rwsem)
-{
-	return (0);
-}
-#endif
-
-int
-rwsem_tryupgrade(struct rw_semaphore *rwsem)
-{
-	if (__rwsem_tryupgrade(rwsem)) {
-		rwsem_release(&rwsem->dep_map, 1, _RET_IP_);
-		rwsem_acquire(&rwsem->dep_map, 0, 1, _RET_IP_);
-#ifdef CONFIG_RWSEM_SPIN_ON_OWNER
-		rwsem->owner = current;
-#endif
-		return (1);
-	}
-	return (0);
-}
-EXPORT_SYMBOL(rwsem_tryupgrade);
-
-int spl_rw_init(void) { return 0; }
-void spl_rw_fini(void) { }
@@ -24,6 +24,7 @@
 *  Solaris Porting Layer (SPL) Task Queue Implementation.
 */

+#include <sys/timer.h>
 #include <sys/taskq.h>
 #include <sys/kmem.h>
 #include <sys/tsd.h>
@@ -242,20 +243,13 @@ task_expire_impl(taskq_ent_t *t)
 	wake_up(&tq->tq_work_waitq);
 }

-#ifdef HAVE_KERNEL_TIMER_FUNCTION_TIMER_LIST
 static void
-task_expire(struct timer_list *tl)
+task_expire(spl_timer_list_t tl)
 {
-	taskq_ent_t *t = from_timer(t, tl, tqent_timer);
+	struct timer_list *tmr = (struct timer_list *)tl;
+	taskq_ent_t *t = from_timer(t, tmr, tqent_timer);
 	task_expire_impl(t);
 }
-#else
-static void
-task_expire(unsigned long data)
-{
-	task_expire_impl((taskq_ent_t *)data);
-}
-#endif

 /*
 * Returns the lowest incomplete taskqid_t.  The taskqid_t may
@@ -597,9 +591,6 @@ taskq_dispatch(taskq_t *tq, task_func_t func, void *arg, uint_t flags)
 	t->tqent_func = func;
 	t->tqent_arg = arg;
 	t->tqent_taskq = tq;
-#ifndef HAVE_KERNEL_TIMER_FUNCTION_TIMER_LIST
-	t->tqent_timer.data = 0;
-#endif
 	t->tqent_timer.function = NULL;
 	t->tqent_timer.expires = 0;
 	t->tqent_birth = jiffies;
@@ -649,9 +640,6 @@ taskq_dispatch_delay(taskq_t *tq, task_func_t func, void *arg,
 	t->tqent_func = func;
 	t->tqent_arg = arg;
 	t->tqent_taskq = tq;
-#ifndef HAVE_KERNEL_TIMER_FUNCTION_TIMER_LIST
-	t->tqent_timer.data = (unsigned long)t;
-#endif
 	t->tqent_timer.function = task_expire;
 	t->tqent_timer.expires = (unsigned long)expire_time;
 	add_timer(&t->tqent_timer);
@@ -744,11 +732,7 @@ taskq_init_ent(taskq_ent_t *t)
 {
 	spin_lock_init(&t->tqent_lock);
 	init_waitqueue_head(&t->tqent_waitq);
-#ifdef HAVE_KERNEL_TIMER_FUNCTION_TIMER_LIST
 	timer_setup(&t->tqent_timer, NULL, 0);
-#else
-	init_timer(&t->tqent_timer);
-#endif
 	INIT_LIST_HEAD(&t->tqent_list);
 	t->tqent_id = 0;
 	t->tqent_func = NULL;
@@ -153,8 +153,9 @@ spl_kthread_create(int (*func)(void *), void *data, const char namefmt[], ...)
 			if (PTR_ERR(tsk) == -ENOMEM)
 				continue;
 			return (NULL);
-		} else
+		} else {
 			return (tsk);
+		}
 	} while (1);
 }
 EXPORT_SYMBOL(spl_kthread_create);
@@ -641,64 +641,6 @@ vn_areleasef(int fd, uf_info_t *fip)
 } /* releasef() */
 EXPORT_SYMBOL(areleasef);

-
-static void
-vn_set_fs_pwd(struct fs_struct *fs, struct path *path)
-{
-	struct path old_pwd;
-
-#ifdef HAVE_FS_STRUCT_SPINLOCK
-	spin_lock(&fs->lock);
-	old_pwd = fs->pwd;
-	fs->pwd = *path;
-	path_get(path);
-	spin_unlock(&fs->lock);
-#else
-	write_lock(&fs->lock);
-	old_pwd = fs->pwd;
-	fs->pwd = *path;
-	path_get(path);
-	write_unlock(&fs->lock);
-#endif /* HAVE_FS_STRUCT_SPINLOCK */
-
-	if (old_pwd.dentry)
-		path_put(&old_pwd);
-}
-
-int
-vn_set_pwd(const char *filename)
-{
-	struct path path;
-	mm_segment_t saved_fs;
-	int rc;
-
-	/*
-	 * user_path_dir() and __user_walk() both expect 'filename' to be
-	 * a user space address so we must briefly increase the data segment
-	 * size to ensure strncpy_from_user() does not fail with -EFAULT.
-	 */
-	saved_fs = get_fs();
-	set_fs(KERNEL_DS);
-
-	rc = user_path_dir(filename, &path);
-	if (rc)
-		goto out;
-
-	rc = inode_permission(path.dentry->d_inode, MAY_EXEC | MAY_ACCESS);
-	if (rc)
-		goto dput_and_out;
-
-	vn_set_fs_pwd(current->fs, &path);
-
-dput_and_out:
-	path_put(&path);
-out:
-	set_fs(saved_fs);
-
-	return (-rc);
-} /* vn_set_pwd() */
-EXPORT_SYMBOL(vn_set_pwd);
-
 static int
 vn_cache_constructor(void *buf, void *cdrarg, int kmflags)
 {
@@ -592,8 +592,9 @@ fletcher_4_incremental_byteswap(void *buf, size_t size, void *data)
 }

 #if defined(_KERNEL)
-/* Fletcher 4 kstats */
-
+/*
+ * Fletcher 4 kstats
+ */
 static int
 fletcher_4_kstat_headers(char *buf, size_t size)
 {
@@ -669,7 +670,6 @@ fletcher_4_benchmark_impl(boolean_t native, char *data, uint64_t data_size)
 	zio_cksum_t zc;
 	uint32_t i, l, sel_save = IMPL_READ(fletcher_4_impl_chosen);

-
 	fletcher_checksum_func_t *fletcher_4_test = native ?
 	    fletcher_4_native : fletcher_4_byteswap;

@@ -232,6 +232,27 @@ entity_namecheck(const char *path, namecheck_err_t *why, char *what)
 			}
 		}

+		if (*end == '\0' || *end == '/') {
+			int component_length = end - start;
+			/* Validate the contents of this component is not '.' */
+			if (component_length == 1) {
+				if (start[0] == '.') {
+					if (why)
+						*why = NAME_ERR_SELF_REF;
+					return (-1);
+				}
+			}
+
+			/* Validate the content of this component is not '..' */
+			if (component_length == 2) {
+				if (start[0] == '.' && start[1] == '.') {
+					if (why)
+						*why = NAME_ERR_PARENT_REF;
+					return (-1);
+				}
+			}
+		}
+
 		/* Snapshot or bookmark delimiter found */
 		if (*end == '@' || *end == '#') {
 			/* Multiple delimiters are not allowed */
@@ -1370,8 +1370,10 @@ abd_raidz_gen_iterate(abd_t **cabds, abd_t *dabd,
 		switch (parity) {
 			case 3:
 				len = MIN(caiters[2].iter_mapsize, len);
+				/* falls through */
 			case 2:
 				len = MIN(caiters[1].iter_mapsize, len);
+				/* falls through */
 			case 1:
 				len = MIN(caiters[0].iter_mapsize, len);
 		}
@@ -1461,9 +1463,11 @@ abd_raidz_rec_iterate(abd_t **cabds, abd_t **tabds,
 			case 3:
 				len = MIN(xiters[2].iter_mapsize, len);
 				len = MIN(citers[2].iter_mapsize, len);
+				/* falls through */
 			case 2:
 				len = MIN(xiters[1].iter_mapsize, len);
 				len = MIN(citers[1].iter_mapsize, len);
+				/* falls through */
 			case 1:
 				len = MIN(xiters[0].iter_mapsize, len);
 				len = MIN(citers[0].iter_mapsize, len);
@@ -21,7 +21,7 @@
 /*
 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
 * Copyright (c) 2018, Joyent, Inc.
- * Copyright (c) 2011, 2018 by Delphix. All rights reserved.
+ * Copyright (c) 2011, 2019 by Delphix. All rights reserved.
 * Copyright (c) 2014 by Saso Kiselkov. All rights reserved.
 * Copyright 2017 Nexenta Systems, Inc.  All rights reserved.
 */
@@ -1872,7 +1872,8 @@ arc_buf_try_copy_decompressed_data(arc_buf_t *buf)
 	 * There were no decompressed bufs, so there should not be a
 	 * checksum on the hdr either.
 	 */
-	EQUIV(!copied, hdr->b_l1hdr.b_freeze_cksum == NULL);
+	if (zfs_flags & ZFS_DEBUG_MODIFY)
+		EQUIV(!copied, hdr->b_l1hdr.b_freeze_cksum == NULL);

 	return (copied);
 }
@@ -2253,7 +2254,6 @@ arc_buf_fill(arc_buf_t *buf, spa_t *spa, const zbookmark_phys_t *zb,
 		 */
 		if (arc_buf_try_copy_decompressed_data(buf)) {
 			/* Skip byteswapping and checksumming (already done) */
-			ASSERT3P(hdr->b_l1hdr.b_freeze_cksum, !=, NULL);
 			return (0);
 		} else {
 			error = zio_decompress_data(HDR_GET_COMPRESS(hdr),
@@ -4801,8 +4801,6 @@ arc_reduce_target_size(int64_t to_free)
 	if (c > to_free && c - to_free > arc_c_min) {
 		arc_c = c - to_free;
 		atomic_add_64(&arc_p, -(arc_p >> arc_shrink_shift));
-		if (asize < arc_c)
-			arc_c = MAX(asize, arc_c_min);
 		if (arc_p > arc_c)
 			arc_p = (arc_c >> 1);
 		ASSERT(arc_c >= arc_c_min);
@@ -5081,6 +5079,9 @@ arc_kmem_reap_soon(void)
 static boolean_t
 arc_adjust_cb_check(void *arg, zthr_t *zthr)
 {
+	if (!arc_initialized)
+		return (B_FALSE);
+
 	/*
 	 * This is necessary so that any changes which may have been made to
 	 * many of the zfs_arc_* module parameters will be propagated to
@@ -5168,6 +5169,9 @@ arc_adjust_cb(void *arg, zthr_t *zthr)
 static boolean_t
 arc_reap_cb_check(void *arg, zthr_t *zthr)
 {
+	if (!arc_initialized)
+		return (B_FALSE);
+
 	int64_t free_memory = arc_available_memory();

 	/*
@@ -5608,7 +5612,7 @@ arc_get_data_impl(arc_buf_hdr_t *hdr, uint64_t size, void *tag)
 		 * If we are growing the cache, and we are adding anonymous
 		 * data, and we have outgrown arc_p, update arc_p
 		 */
-		if (aggsum_compare(&arc_size, arc_c) < 0 &&
+		if (aggsum_upper_bound(&arc_size) < arc_c &&
 		    hdr->b_l1hdr.b_state == arc_anon &&
 		    (zfs_refcount_count(&arc_anon->arcs_size) +
 		    zfs_refcount_count(&arc_mru->arcs_size) > arc_p))
@@ -7926,11 +7930,9 @@ arc_fini(void)

 	list_destroy(&arc_prune_list);
 	mutex_destroy(&arc_prune_mtx);
-	(void) zthr_cancel(arc_adjust_zthr);
-	zthr_destroy(arc_adjust_zthr);

+	(void) zthr_cancel(arc_adjust_zthr);
 	(void) zthr_cancel(arc_reap_zthr);
-	zthr_destroy(arc_reap_zthr);

 	mutex_destroy(&arc_adjust_lock);
 	cv_destroy(&arc_adjust_waiters_cv);
@@ -7943,6 +7945,14 @@ arc_fini(void)
 	buf_fini();
 	arc_state_fini();

+	/*
+	 * We destroy the zthrs after all the ARC state has been
+	 * torn down to avoid the case of them receiving any
+	 * wakeup() signals after they are destroyed.
+	 */
+	zthr_destroy(arc_adjust_zthr);
+	zthr_destroy(arc_reap_zthr);
+
 	ASSERT0(arc_loaned_bytes);
 }

@@ -8760,7 +8770,7 @@ l2arc_apply_transforms(spa_t *spa, arc_buf_hdr_t *hdr, uint64_t asize,

 	/*
 	 * If this data simply needs its own buffer, we simply allocate it
-	 * and copy the data. This may be done to elimiate a depedency on a
+	 * and copy the data. This may be done to eliminate a dependency on a
 	 * shared buffer or to reallocate the buffer to match asize.
 	 */
 	if (HDR_HAS_RABD(hdr) && asize != psize) {
@@ -2591,7 +2591,8 @@ dbuf_destroy(dmu_buf_impl_t *db)
 	if (db->db_blkid != DMU_BONUS_BLKID) {
 		boolean_t needlock = !MUTEX_HELD(&dn->dn_dbufs_mtx);
 		if (needlock)
-			mutex_enter(&dn->dn_dbufs_mtx);
+			mutex_enter_nested(&dn->dn_dbufs_mtx,
+			    NESTED_SINGLE);
 		avl_remove(&dn->dn_dbufs, db);
 		atomic_dec_32(&dn->dn_dbufs_count);
 		membar_producer();
@@ -21,6 +21,7 @@

 /*
 * Copyright (c) 2009, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2018 by Delphix. All rights reserved.
 */

 #include <sys/zfs_context.h>
@@ -117,7 +118,18 @@ ddt_zap_walk(objset_t *os, uint64_t object, ddt_entry_t *dde, uint64_t *walk)
 	zap_attribute_t za;
 	int error;

-	zap_cursor_init_serialized(&zc, os, object, *walk);
+	if (*walk == 0) {
+		/*
+		 * We don't want to prefetch the entire ZAP object, because
+		 * it can be enormous.  Also the primary use of DDT iteration
+		 * is for scrubbing, in which case we will be issuing many
+		 * scrub I/Os for each ZAP block that we read in, so
+		 * reading the ZAP is unlikely to be the bottleneck.
+		 */
+		zap_cursor_init_noprefetch(&zc, os, object);
+	} else {
+		zap_cursor_init_serialized(&zc, os, object, *walk);
+	}
 	if ((error = zap_cursor_retrieve(&zc, &za)) == 0) {
 		uchar_t cbuf[sizeof (dde->dde_phys) + 1];
 		uint64_t csize = za.za_num_integers;
@@ -81,6 +81,13 @@ int zfs_dmu_offset_next_sync = 0;
 */
 int zfs_object_remap_one_indirect_delay_ms = 0;

+/*
+ * Limit the amount we can prefetch with one call to this amount.  This
+ * helps to limit the amount of memory that can be used by prefetching.
+ * Larger objects should be prefetched a bit at a time.
+ */
+int dmu_prefetch_max = 8 * SPA_MAXBLOCKSIZE;
+
 const dmu_object_type_info_t dmu_ot[DMU_OT_NUMTYPES] = {
 	{DMU_BSWAP_UINT8,  TRUE,  FALSE, FALSE, "unallocated"		},
 	{DMU_BSWAP_ZAP,    TRUE,  TRUE,  FALSE, "object directory"	},
@@ -667,6 +674,11 @@ dmu_prefetch(objset_t *os, uint64_t object, int64_t level, uint64_t offset,
 		return;
 	}

+	/*
+	 * See comment before the definition of dmu_prefetch_max.
+	 */
+	len = MIN(len, dmu_prefetch_max);
+
 	/*
 	 * XXX - Note, if the dnode for the requested object is not
 	 * already cached, we will do a *synchronous* read in the
@@ -2629,6 +2641,10 @@ module_param(zfs_dmu_offset_next_sync, int, 0644);
 MODULE_PARM_DESC(zfs_dmu_offset_next_sync,
 	"Enable forcing txg sync to find holes");

+module_param(dmu_prefetch_max, int, 0644);
+MODULE_PARM_DESC(dmu_prefetch_max,
+	"Limit one prefetch call to this size");
+
 /* END CSTYLED */

 #endif
@@ -1348,13 +1348,6 @@ dmu_objset_clone_check(void *arg, dmu_tx_t *tx)
 		return (SET_ERROR(EINVAL));
 	}

-	error = dmu_objset_clone_crypt_check(pdd, origin->ds_dir);
-	if (error != 0) {
-		dsl_dataset_rele(origin, FTAG);
-		dsl_dir_rele(pdd, FTAG);
-		return (error);
-	}
-
 	dsl_dataset_rele(origin, FTAG);
 	dsl_dir_rele(pdd, FTAG);

@@ -1699,6 +1692,8 @@ dmu_objset_sync(objset_t *os, zio_t *pio, dmu_tx_t *tx)
 	zio_t *zio;
 	list_t *list;
 	dbuf_dirty_record_t *dr;
+	int num_sublists;
+	multilist_t *ml;
 	blkptr_t *blkptr_copy = kmem_alloc(sizeof (*os->os_rootbp), KM_SLEEP);
 	*blkptr_copy = *os->os_rootbp;

@@ -1787,10 +1782,13 @@ dmu_objset_sync(objset_t *os, zio_t *pio, dmu_tx_t *tx)
 		}
 	}

-	for (int i = 0;
-	    i < multilist_get_num_sublists(os->os_dirty_dnodes[txgoff]); i++) {
+	ml = os->os_dirty_dnodes[txgoff];
+	num_sublists = multilist_get_num_sublists(ml);
+	for (int i = 0; i < num_sublists; i++) {
+		if (multilist_sublist_is_empty_idx(ml, i))
+			continue;
 		sync_dnodes_arg_t *sda = kmem_alloc(sizeof (*sda), KM_SLEEP);
-		sda->sda_list = os->os_dirty_dnodes[txgoff];
+		sda->sda_list = ml;
 		sda->sda_sublist_idx = i;
 		sda->sda_tx = tx;
 		(void) taskq_dispatch(dmu_objset_pool(os)->dp_sync_taskq,
@@ -2093,6 +2091,8 @@ userquota_updates_task(void *arg)
 void
 dmu_objset_do_userquota_updates(objset_t *os, dmu_tx_t *tx)
 {
+	int num_sublists;
+
 	if (!dmu_objset_userused_enabled(os))
 		return;

@@ -2125,8 +2125,10 @@ dmu_objset_do_userquota_updates(objset_t *os, dmu_tx_t *tx)
 		    DMU_OT_USERGROUP_USED, DMU_OT_NONE, 0, tx));
 	}

-	for (int i = 0;
-	    i < multilist_get_num_sublists(os->os_synced_dnodes); i++) {
+	num_sublists = multilist_get_num_sublists(os->os_synced_dnodes);
+	for (int i = 0; i < num_sublists; i++) {
+		if (multilist_sublist_is_empty_idx(os->os_synced_dnodes, i))
+			continue;
 		userquota_updates_arg_t *uua =
 		    kmem_alloc(sizeof (*uua), KM_SLEEP);
 		uua->uua_os = os;
@@ -86,21 +86,25 @@ recv_begin_check_existing_impl(dmu_recv_begin_arg_t *drba, dsl_dataset_t *ds,
 	boolean_t raw = (featureflags & DMU_BACKUP_FEATURE_RAW) != 0;
 	boolean_t embed = (featureflags & DMU_BACKUP_FEATURE_EMBED_DATA) != 0;

-	/* temporary clone name must not exist */
+	/* Temporary clone name must not exist. */
 	error = zap_lookup(dp->dp_meta_objset,
 	    dsl_dir_phys(ds->ds_dir)->dd_child_dir_zapobj, recv_clone_name,
 	    8, 1, &val);
 	if (error != ENOENT)
 		return (error == 0 ? EBUSY : error);

-	/* new snapshot name must not exist */
+	/* Resume state must not be set. */
+	if (dsl_dataset_has_resume_receive_state(ds))
+		return (SET_ERROR(EBUSY));
+
+	/* New snapshot name must not exist. */
 	error = zap_lookup(dp->dp_meta_objset,
 	    dsl_dataset_phys(ds)->ds_snapnames_zapobj,
 	    drba->drba_cookie->drc_tosnap, 8, 1, &val);
 	if (error != ENOENT)
 		return (error == 0 ? EEXIST : error);

-	/* must not have children if receiving a ZVOL */
+	/* Must not have children if receiving a ZVOL. */
 	error = zap_count(dp->dp_meta_objset,
 	    dsl_dir_phys(ds->ds_dir)->dd_child_dir_zapobj, &children);
 	if (error != 0)
@@ -327,7 +331,7 @@ dmu_recv_begin_check(void *arg, dmu_tx_t *tx)
 		/* Open the parent of tofs */
 		ASSERT3U(strlen(tofs), <, sizeof (buf));
 		(void) strlcpy(buf, tofs, strrchr(tofs, '/') - tofs + 1);
-		error = dsl_dataset_hold_flags(dp, buf, dsflags, FTAG, &ds);
+		error = dsl_dataset_hold(dp, buf, FTAG, &ds);
 		if (error != 0)
 			return (error);

@@ -345,13 +349,13 @@ dmu_recv_begin_check(void *arg, dmu_tx_t *tx)
 			error = dmu_objset_create_crypt_check(ds->ds_dir,
 			    drba->drba_dcp, &will_encrypt);
 			if (error != 0) {
-				dsl_dataset_rele_flags(ds, dsflags, FTAG);
+				dsl_dataset_rele(ds, FTAG);
 				return (error);
 			}

 			if (will_encrypt &&
 			    (featureflags & DMU_BACKUP_FEATURE_EMBED_DATA)) {
-				dsl_dataset_rele_flags(ds, dsflags, FTAG);
+				dsl_dataset_rele(ds, FTAG);
 				return (SET_ERROR(EINVAL));
 			}
 		}
@@ -364,25 +368,25 @@ dmu_recv_begin_check(void *arg, dmu_tx_t *tx)
 		error = dsl_fs_ss_limit_check(ds->ds_dir, 1,
 		    ZFS_PROP_FILESYSTEM_LIMIT, NULL, drba->drba_cred);
 		if (error != 0) {
-			dsl_dataset_rele_flags(ds, dsflags, FTAG);
+			dsl_dataset_rele(ds, FTAG);
 			return (error);
 		}

 		error = dsl_fs_ss_limit_check(ds->ds_dir, 1,
 		    ZFS_PROP_SNAPSHOT_LIMIT, NULL, drba->drba_cred);
 		if (error != 0) {
-			dsl_dataset_rele_flags(ds, dsflags, FTAG);
+			dsl_dataset_rele(ds, FTAG);
 			return (error);
 		}

 		/* can't recv below anything but filesystems (eg. no ZVOLs) */
 		error = dmu_objset_from_ds(ds, &os);
 		if (error != 0) {
-			dsl_dataset_rele_flags(ds, dsflags, FTAG);
+			dsl_dataset_rele(ds, FTAG);
 			return (error);
 		}
 		if (dmu_objset_type(os) != DMU_OST_ZFS) {
-			dsl_dataset_rele_flags(ds, dsflags, FTAG);
+			dsl_dataset_rele(ds, FTAG);
 			return (SET_ERROR(ZFS_ERR_WRONG_PARENT));
 		}

@@ -392,31 +396,31 @@ dmu_recv_begin_check(void *arg, dmu_tx_t *tx)
 			error = dsl_dataset_hold_flags(dp, drba->drba_origin,
 			    dsflags, FTAG, &origin);
 			if (error != 0) {
-				dsl_dataset_rele_flags(ds, dsflags, FTAG);
+				dsl_dataset_rele(ds, FTAG);
 				return (error);
 			}
 			if (!origin->ds_is_snapshot) {
 				dsl_dataset_rele_flags(origin, dsflags, FTAG);
-				dsl_dataset_rele_flags(ds, dsflags, FTAG);
+				dsl_dataset_rele(ds, FTAG);
 				return (SET_ERROR(EINVAL));
 			}
 			if (dsl_dataset_phys(origin)->ds_guid != fromguid &&
 			    fromguid != 0) {
 				dsl_dataset_rele_flags(origin, dsflags, FTAG);
-				dsl_dataset_rele_flags(ds, dsflags, FTAG);
+				dsl_dataset_rele(ds, FTAG);
 				return (SET_ERROR(ENODEV));
 			}
 			if (origin->ds_dir->dd_crypto_obj != 0 &&
 			    (featureflags & DMU_BACKUP_FEATURE_EMBED_DATA)) {
 				dsl_dataset_rele_flags(origin, dsflags, FTAG);
-				dsl_dataset_rele_flags(ds, dsflags, FTAG);
+				dsl_dataset_rele(ds, FTAG);
 				return (SET_ERROR(EINVAL));
 			}
 			dsl_dataset_rele_flags(origin,
 			    dsflags, FTAG);
 		}

-		dsl_dataset_rele_flags(ds, dsflags, FTAG);
+		dsl_dataset_rele(ds, FTAG);
 		error = 0;
 	}
 	return (error);
@@ -925,6 +925,25 @@ dmu_tx_try_assign(dmu_tx_t *tx, uint64_t txg_how)
 	    txh = list_next(&tx->tx_holds, txh)) {
 		dnode_t *dn = txh->txh_dnode;
 		if (dn != NULL) {
+			/*
+			 * This thread can't hold the dn_struct_rwlock
+			 * while assigning the tx, because this can lead to
+			 * deadlock. Specifically, if this dnode is already
+			 * assigned to an earlier txg, this thread may need
+			 * to wait for that txg to sync (the ERESTART case
+			 * below).  The other thread that has assigned this
+			 * dnode to an earlier txg prevents this txg from
+			 * syncing until its tx can complete (calling
+			 * dmu_tx_commit()), but it may need to acquire the
+			 * dn_struct_rwlock to do so (e.g. via
+			 * dmu_buf_hold*()).
+			 *
+			 * Note that this thread can't hold the lock for
+			 * read either, but the rwlock doesn't record
+			 * enough information to make that assertion.
+			 */
+			ASSERT(!RW_WRITE_HELD(&dn->dn_struct_rwlock));
+
 			mutex_enter(&dn->dn_mtx);
 			if (dn->dn_assigned_txg == tx->tx_txg - 1) {
 				mutex_exit(&dn->dn_mtx);
@@ -1319,7 +1338,10 @@ dmu_tx_hold_sa(dmu_tx_t *tx, sa_handle_t *hdl, boolean_t may_grow)

 	object = sa_handle_object(hdl);

-	dmu_tx_hold_bonus(tx, object);
+	dmu_buf_impl_t *db = (dmu_buf_impl_t *)hdl->sa_bonus;
+	DB_DNODE_ENTER(db);
+	dmu_tx_hold_bonus_by_dnode(tx, DB_DNODE(db));
+	DB_DNODE_EXIT(db);

 	if (tx->tx_objset->os_sa->sa_master_obj == 0)
 		return;
@@ -1341,7 +1363,6 @@ dmu_tx_hold_sa(dmu_tx_t *tx, sa_handle_t *hdl, boolean_t may_grow)
 		ASSERT(tx->tx_txg == 0);
 		dmu_tx_hold_spill(tx, object);
 	} else {
-		dmu_buf_impl_t *db = (dmu_buf_impl_t *)hdl->sa_bonus;
 		dnode_t *dn;

 		DB_DNODE_ENTER(db);
@@ -55,7 +55,6 @@ dnode_stats_t dnode_stats = {
 	{ "dnode_hold_free_lock_retry",		KSTAT_DATA_UINT64 },
 	{ "dnode_hold_free_overflow",		KSTAT_DATA_UINT64 },
 	{ "dnode_hold_free_refcount",		KSTAT_DATA_UINT64 },
-	{ "dnode_hold_free_txg",		KSTAT_DATA_UINT64 },
 	{ "dnode_free_interior_lock_retry",	KSTAT_DATA_UINT64 },
 	{ "dnode_allocate",			KSTAT_DATA_UINT64 },
 	{ "dnode_reallocate",			KSTAT_DATA_UINT64 },
@@ -1255,6 +1254,10 @@ dnode_buf_evict_async(void *dbu)
 * as an extra dnode slot by an large dnode, in which case it returns
 * ENOENT.
 *
+ * If the DNODE_DRY_RUN flag is set, we don't actually hold the dnode, just
+ * return whether the hold would succeed or not. tag and dnp should set to
+ * NULL in this case.
+ *
 * errors:
 * EINVAL - Invalid object number or flags.
 * ENOSPC - Hole too small to fulfill "slots" request (DNODE_MUST_BE_FREE)
@@ -1283,6 +1286,7 @@ dnode_hold_impl(objset_t *os, uint64_t object, int flag, int slots,

 	ASSERT(!(flag & DNODE_MUST_BE_ALLOCATED) || (slots == 0));
 	ASSERT(!(flag & DNODE_MUST_BE_FREE) || (slots > 0));
+	IMPLY(flag & DNODE_DRY_RUN, (tag == NULL) && (dnp == NULL));

 	/*
 	 * If you are holding the spa config lock as writer, you shouldn't
@@ -1312,8 +1316,11 @@ dnode_hold_impl(objset_t *os, uint64_t object, int flag, int slots,
 		if ((flag & DNODE_MUST_BE_FREE) && type != DMU_OT_NONE)
 			return (SET_ERROR(EEXIST));
 		DNODE_VERIFY(dn);
-		(void) zfs_refcount_add(&dn->dn_holds, tag);
-		*dnp = dn;
+		/* Don't actually hold if dry run, just return 0 */
+		if (!(flag & DNODE_DRY_RUN)) {
+			(void) zfs_refcount_add(&dn->dn_holds, tag);
+			*dnp = dn;
+		}
 		return (0);
 	}

@@ -1455,6 +1462,14 @@ dnode_hold_impl(objset_t *os, uint64_t object, int flag, int slots,
 			return (SET_ERROR(ENOENT));
 		}

+		/* Don't actually hold if dry run, just return 0 */
+		if (flag & DNODE_DRY_RUN) {
+			mutex_exit(&dn->dn_mtx);
+			dnode_slots_rele(dnc, idx, slots);
+			dbuf_rele(db, FTAG);
+			return (0);
+		}
+
 		DNODE_STAT_BUMP(dnode_hold_alloc_hits);
 	} else if (flag & DNODE_MUST_BE_FREE) {

@@ -1512,6 +1527,14 @@ dnode_hold_impl(objset_t *os, uint64_t object, int flag, int slots,
 			return (SET_ERROR(EEXIST));
 		}

+		/* Don't actually hold if dry run, just return 0 */
+		if (flag & DNODE_DRY_RUN) {
+			mutex_exit(&dn->dn_mtx);
+			dnode_slots_rele(dnc, idx, slots);
+			dbuf_rele(db, FTAG);
+			return (0);
+		}
+
 		dnode_set_slots(dnc, idx + 1, slots - 1, DN_SLOT_INTERIOR);
 		DNODE_STAT_BUMP(dnode_hold_free_hits);
 	} else {
@@ -1519,15 +1542,7 @@ dnode_hold_impl(objset_t *os, uint64_t object, int flag, int slots,
 		return (SET_ERROR(EINVAL));
 	}

-	if (dn->dn_free_txg) {
-		DNODE_STAT_BUMP(dnode_hold_free_txg);
-		type = dn->dn_type;
-		mutex_exit(&dn->dn_mtx);
-		dnode_slots_rele(dnc, idx, slots);
-		dbuf_rele(db, FTAG);
-		return (SET_ERROR((flag & DNODE_MUST_BE_ALLOCATED) ?
-		    ENOENT : EEXIST));
-	}
+	ASSERT0(dn->dn_free_txg);

 	if (zfs_refcount_add(&dn->dn_holds, tag) == 1)
 		dbuf_add_ref(db, dnh);
@@ -1618,6 +1633,16 @@ dnode_rele_and_unlock(dnode_t *dn, void *tag, boolean_t evicting)
 	}
 }

+/*
+ * Test whether we can create a dnode at the specified location.
+ */
+int
+dnode_try_claim(objset_t *os, uint64_t object, int slots)
+{
+	return (dnode_hold_impl(os, object, DNODE_MUST_BE_FREE | DNODE_DRY_RUN,
+	    slots, NULL, NULL));
+}
+
 void
 dnode_setdirty(dnode_t *dn, dmu_tx_t *tx)
 {
@@ -2483,3 +2508,13 @@ out:

 	return (error);
 }
+
+#if defined(_KERNEL)
+EXPORT_SYMBOL(dnode_hold);
+EXPORT_SYMBOL(dnode_rele);
+EXPORT_SYMBOL(dnode_set_nlevels);
+EXPORT_SYMBOL(dnode_set_blksz);
+EXPORT_SYMBOL(dnode_free_range);
+EXPORT_SYMBOL(dnode_evict_dbufs);
+EXPORT_SYMBOL(dnode_evict_bonus);
+#endif
@@ -1610,15 +1610,8 @@ dsl_dir_rename_crypt_check(dsl_dir_t *dd, dsl_dir_t *newparent)
 	int ret;
 	uint64_t curr_rddobj, parent_rddobj;

-	if (dd->dd_crypto_obj == 0) {
-		/* children of encrypted parents must be encrypted */
-		if (newparent->dd_crypto_obj != 0) {
-			ret = SET_ERROR(EACCES);
-			goto error;
-		}
-
+	if (dd->dd_crypto_obj == 0)
 		return (0);
-	}

 	ret = dsl_dir_get_encryption_root_ddobj(dd, &curr_rddobj);
 	if (ret != 0)
@@ -1683,11 +1676,15 @@ dsl_dataset_promote_crypt_check(dsl_dir_t *target, dsl_dir_t *origin)
 	 * Check that the parent of the target has the same encryption root.
 	 */
 	ret = dsl_dir_get_encryption_root_ddobj(origin->dd_parent, &op_rddobj);
-	if (ret != 0)
+	if (ret == ENOENT)
+		return (SET_ERROR(EACCES));
+	else if (ret != 0)
 		return (ret);

 	ret = dsl_dir_get_encryption_root_ddobj(target->dd_parent, &tp_rddobj);
-	if (ret != 0)
+	if (ret == ENOENT)
+		return (SET_ERROR(EACCES));
+	else if (ret != 0)
 		return (ret);

 	if (op_rddobj != tp_rddobj)
@@ -1747,34 +1744,6 @@ dsl_dataset_promote_crypt_sync(dsl_dir_t *target, dsl_dir_t *origin,
 	kmem_free(keylocation, ZAP_MAXVALUELEN);
 }

-int
-dmu_objset_clone_crypt_check(dsl_dir_t *parentdd, dsl_dir_t *origindd)
-{
-	int ret;
-	uint64_t pcrypt, crypt;
-
-	/*
-	 * Check that we are not making an unencrypted child of an
-	 * encrypted parent.
-	 */
-	ret = dsl_dir_get_crypt(parentdd, &pcrypt);
-	if (ret != 0)
-		return (ret);
-
-	ret = dsl_dir_get_crypt(origindd, &crypt);
-	if (ret != 0)
-		return (ret);
-
-	ASSERT3U(pcrypt, !=, ZIO_CRYPT_INHERIT);
-	ASSERT3U(crypt, !=, ZIO_CRYPT_INHERIT);
-
-	if (crypt == ZIO_CRYPT_OFF && pcrypt != ZIO_CRYPT_OFF)
-		return (SET_ERROR(EINVAL));
-
-	return (0);
-}
-
-
 int
 dmu_objset_create_crypt_check(dsl_dir_t *parentdd, dsl_crypto_params_t *dcp,
    boolean_t *will_encrypt)
@@ -1805,13 +1774,6 @@ dmu_objset_create_crypt_check(dsl_dir_t *parentdd, dsl_crypto_params_t *dcp,
 	ASSERT3U(pcrypt, !=, ZIO_CRYPT_INHERIT);
 	ASSERT3U(crypt, !=, ZIO_CRYPT_INHERIT);

-	/*
-	 * We can't create an unencrypted child of an encrypted parent
-	 * under any circumstances.
-	 */
-	if (crypt == ZIO_CRYPT_OFF && pcrypt != ZIO_CRYPT_OFF)
-		return (SET_ERROR(EINVAL));
-
 	/* check for valid dcp with no encryption (inherited or local) */
 	if (crypt == ZIO_CRYPT_OFF) {
 		/* Must not specify encryption params */
@@ -2662,11 +2624,13 @@ dsl_dataset_crypt_stats(dsl_dataset_t *ds, nvlist_t *nv)
 	}

 	if (dsl_dir_get_encryption_root_ddobj(dd, &intval) == 0) {
-		VERIFY0(dsl_dir_hold_obj(dd->dd_pool, intval, NULL, FTAG,
-		    &enc_root));
-		dsl_dir_name(enc_root, buf);
-		dsl_dir_rele(enc_root, FTAG);
-		dsl_prop_nvlist_add_string(nv, ZFS_PROP_ENCRYPTION_ROOT, buf);
+		if (dsl_dir_hold_obj(dd->dd_pool, intval, NULL, FTAG,
+		    &enc_root) == 0) {
+			dsl_dir_name(enc_root, buf);
+			dsl_dir_rele(enc_root, FTAG);
+			dsl_prop_nvlist_add_string(nv,
+			    ZFS_PROP_ENCRYPTION_ROOT, buf);
+		}
 	}
 }

@@ -1059,9 +1059,10 @@ dsl_destroy_head(const char *name)
 		/*
 		 * Head deletion is processed in one txg on old pools;
 		 * remove the objects from open context so that the txg sync
-		 * is not too long.
+		 * is not too long. This optimization can only work for
+		 * encrypted datasets if the wrapping key is loaded.
 		 */
-		error = dmu_objset_own(name, DMU_OST_ANY, B_FALSE, B_FALSE,
+		error = dmu_objset_own(name, DMU_OST_ANY, B_FALSE, B_TRUE,
 		    FTAG, &os);
 		if (error == 0) {
 			uint64_t prev_snap_txg =
@@ -1073,7 +1074,7 @@ dsl_destroy_head(const char *name)
 				(void) dmu_free_long_object(os, obj);
 			/* sync out all frees */
 			txg_wait_synced(dmu_objset_pool(os), 0);
-			dmu_objset_disown(os, B_FALSE, FTAG);
+			dmu_objset_disown(os, B_TRUE, FTAG);
 		}
 	}

@@ -23,7 +23,7 @@
 * Copyright (c) 2011, 2017 by Delphix. All rights reserved.
 * Copyright 2016 Gary Mills
 * Copyright (c) 2017 Datto Inc.
- * Copyright 2017 Joyent, Inc.
+ * Copyright 2019 Joyent, Inc.
 */

 #include <sys/dsl_scan.h>
@@ -952,13 +952,16 @@ dsl_scan_done(dsl_scan_t *scn, boolean_t complete, dmu_tx_t *tx)
 		 * will find the drives that need to be resilvered
 		 * when the machine reboots and start the resilver then.
 		 */
-		boolean_t resilver_needed =
-		    dsl_scan_clear_deferred(spa->spa_root_vdev, tx);
-		if (resilver_needed) {
-			spa_history_log_internal(spa,
-			    "starting deferred resilver", tx,
-			    "errors=%llu", spa_get_errlog_size(spa));
-			spa_async_request(spa, SPA_ASYNC_RESILVER);
+		if (spa_feature_is_enabled(spa, SPA_FEATURE_RESILVER_DEFER)) {
+			boolean_t resilver_needed =
+			    dsl_scan_clear_deferred(spa->spa_root_vdev, tx);
+			if (resilver_needed) {
+				spa_history_log_internal(spa,
+				    "starting deferred resilver", tx,
+				    "errors=%llu",
+				    (u_longlong_t)spa_get_errlog_size(spa));
+				spa_async_request(spa, SPA_ASYNC_RESILVER);
+			}
 		}
 	}

@@ -2165,16 +2168,17 @@ ds_clone_swapped_bookmark(dsl_dataset_t *ds1, dsl_dataset_t *ds2,
 }

 /*
- * Called when a parent dataset and its clone are swapped. If we were
+ * Called when an origin dataset and its clone are swapped.  If we were
 * currently traversing the dataset, we need to switch to traversing the
- * newly promoted parent.
+ * newly promoted clone.
 */
 void
 dsl_scan_ds_clone_swapped(dsl_dataset_t *ds1, dsl_dataset_t *ds2, dmu_tx_t *tx)
 {
 	dsl_pool_t *dp = ds1->ds_dir->dd_pool;
 	dsl_scan_t *scn = dp->dp_scan;
-	uint64_t mintxg;
+	uint64_t mintxg1, mintxg2;
+	boolean_t ds1_queued, ds2_queued;

 	if (!dsl_scan_is_running(scn))
 		return;
@@ -2182,44 +2186,81 @@ dsl_scan_ds_clone_swapped(dsl_dataset_t *ds1, dsl_dataset_t *ds2, dmu_tx_t *tx)
 	ds_clone_swapped_bookmark(ds1, ds2, &scn->scn_phys.scn_bookmark);
 	ds_clone_swapped_bookmark(ds1, ds2, &scn->scn_phys_cached.scn_bookmark);

-	if (scan_ds_queue_contains(scn, ds1->ds_object, &mintxg)) {
-		scan_ds_queue_remove(scn, ds1->ds_object);
-		scan_ds_queue_insert(scn, ds2->ds_object, mintxg);
+	/*
+	 * Handle the in-memory scan queue.
+	 */
+	ds1_queued = scan_ds_queue_contains(scn, ds1->ds_object, &mintxg1);
+	ds2_queued = scan_ds_queue_contains(scn, ds2->ds_object, &mintxg2);
+
+	/* Sanity checking. */
+	if (ds1_queued) {
+		ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg);
+		ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg);
 	}
-	if (scan_ds_queue_contains(scn, ds2->ds_object, &mintxg)) {
-		scan_ds_queue_remove(scn, ds2->ds_object);
-		scan_ds_queue_insert(scn, ds1->ds_object, mintxg);
+	if (ds2_queued) {
+		ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg);
+		ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg);
 	}

-	if (zap_lookup_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj,
-	    ds1->ds_object, &mintxg) == 0) {
-		int err;
-		ASSERT3U(mintxg, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg);
-		ASSERT3U(mintxg, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg);
-		VERIFY3U(0, ==, zap_remove_int(dp->dp_meta_objset,
+	if (ds1_queued && ds2_queued) {
+		/*
+		 * If both are queued, we don't need to do anything.
+		 * The swapping code below would not handle this case correctly,
+		 * since we can't insert ds2 if it is already there. That's
+		 * because scan_ds_queue_insert() prohibits a duplicate insert
+		 * and panics.
+		 */
+	} else if (ds1_queued) {
+		scan_ds_queue_remove(scn, ds1->ds_object);
+		scan_ds_queue_insert(scn, ds2->ds_object, mintxg1);
+	} else if (ds2_queued) {
+		scan_ds_queue_remove(scn, ds2->ds_object);
+		scan_ds_queue_insert(scn, ds1->ds_object, mintxg2);
+	}
+
+	/*
+	 * Handle the on-disk scan queue.
+	 * The on-disk state is an out-of-date version of the in-memory state,
+	 * so the in-memory and on-disk values for ds1_queued and ds2_queued may
+	 * be different. Therefore we need to apply the swap logic to the
+	 * on-disk state independently of the in-memory state.
+	 */
+	ds1_queued = zap_lookup_int_key(dp->dp_meta_objset,
+	    scn->scn_phys.scn_queue_obj, ds1->ds_object, &mintxg1) == 0;
+	ds2_queued = zap_lookup_int_key(dp->dp_meta_objset,
+	    scn->scn_phys.scn_queue_obj, ds2->ds_object, &mintxg2) == 0;
+
+	/* Sanity checking. */
+	if (ds1_queued) {
+		ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg);
+		ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg);
+	}
+	if (ds2_queued) {
+		ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg);
+		ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg);
+	}
+
+	if (ds1_queued && ds2_queued) {
+		/*
+		 * If both are queued, we don't need to do anything.
+		 * Alternatively, we could check for EEXIST from
+		 * zap_add_int_key() and back out to the original state, but
+		 * that would be more work than checking for this case upfront.
+		 */
+	} else if (ds1_queued) {
+		VERIFY3S(0, ==, zap_remove_int(dp->dp_meta_objset,
 		    scn->scn_phys.scn_queue_obj, ds1->ds_object, tx));
-		err = zap_add_int_key(dp->dp_meta_objset,
-		    scn->scn_phys.scn_queue_obj, ds2->ds_object, mintxg, tx);
-		VERIFY(err == 0 || err == EEXIST);
-		if (err == EEXIST) {
-			/* Both were there to begin with */
-			VERIFY(0 == zap_add_int_key(dp->dp_meta_objset,
-			    scn->scn_phys.scn_queue_obj,
-			    ds1->ds_object, mintxg, tx));
-		}
+		VERIFY3S(0, ==, zap_add_int_key(dp->dp_meta_objset,
+		    scn->scn_phys.scn_queue_obj, ds2->ds_object, mintxg1, tx));
 		zfs_dbgmsg("clone_swap ds %llu; in queue; "
 		    "replacing with %llu",
 		    (u_longlong_t)ds1->ds_object,
 		    (u_longlong_t)ds2->ds_object);
-	}
-	if (zap_lookup_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj,
-	    ds2->ds_object, &mintxg) == 0) {
-		ASSERT3U(mintxg, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg);
-		ASSERT3U(mintxg, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg);
-		VERIFY3U(0, ==, zap_remove_int(dp->dp_meta_objset,
+	} else if (ds2_queued) {
+		VERIFY3S(0, ==, zap_remove_int(dp->dp_meta_objset,
 		    scn->scn_phys.scn_queue_obj, ds2->ds_object, tx));
-		VERIFY(0 == zap_add_int_key(dp->dp_meta_objset,
-		    scn->scn_phys.scn_queue_obj, ds1->ds_object, mintxg, tx));
+		VERIFY3S(0, ==, zap_add_int_key(dp->dp_meta_objset,
+		    scn->scn_phys.scn_queue_obj, ds1->ds_object, mintxg2, tx));
 		zfs_dbgmsg("clone_swap ds %llu; in queue; "
 		    "replacing with %llu",
 		    (u_longlong_t)ds2->ds_object,
@@ -20,7 +20,7 @@
 */
 /*
 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2011, 2018 by Delphix. All rights reserved.
+ * Copyright (c) 2011, 2019 by Delphix. All rights reserved.
 * Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
 * Copyright (c) 2017, Intel Corporation.
 */
@@ -103,12 +103,27 @@ int zfs_mg_noalloc_threshold = 0;

 /*
 * Metaslab groups are considered eligible for allocations if their
- * fragmenation metric (measured as a percentage) is less than or equal to
- * zfs_mg_fragmentation_threshold. If a metaslab group exceeds this threshold
- * then it will be skipped unless all metaslab groups within the metaslab
- * class have also crossed this threshold.
+ * fragmenation metric (measured as a percentage) is less than or
+ * equal to zfs_mg_fragmentation_threshold. If a metaslab group
+ * exceeds this threshold then it will be skipped unless all metaslab
+ * groups within the metaslab class have also crossed this threshold.
+ *
+ * This tunable was introduced to avoid edge cases where we continue
+ * allocating from very fragmented disks in our pool while other, less
+ * fragmented disks, exists. On the other hand, if all disks in the
+ * pool are uniformly approaching the threshold, the threshold can
+ * be a speed bump in performance, where we keep switching the disks
+ * that we allocate from (e.g. we allocate some segments from disk A
+ * making it bypassing the threshold while freeing segments from disk
+ * B getting its fragmentation below the threshold).
+ *
+ * Empirically, we've seen that our vdev selection for allocations is
+ * good enough that fragmentation increases uniformly across all vdevs
+ * the majority of the time. Thus we set the threshold percentage high
+ * enough to avoid hitting the speed bump on pools that are being pushed
+ * to the edge.
 */
-int zfs_mg_fragmentation_threshold = 85;
+int zfs_mg_fragmentation_threshold = 95;

 /*
 * Allow metaslabs to keep their active state as long as their fragmentation
@@ -2934,6 +2949,30 @@ metaslab_sync(metaslab_t *msp, uint64_t txg)
 	dmu_tx_commit(tx);
 }

+void
+metaslab_potentially_unload(metaslab_t *msp, uint64_t txg)
+{
+	/*
+	 * If the metaslab is loaded and we've not tried to load or allocate
+	 * from it in 'metaslab_unload_delay' txgs, then unload it.
+	 */
+	if (msp->ms_loaded &&
+	    msp->ms_disabled == 0 &&
+	    msp->ms_selected_txg + metaslab_unload_delay < txg) {
+		for (int t = 1; t < TXG_CONCURRENT_STATES; t++) {
+			VERIFY0(range_tree_space(
+			    msp->ms_allocating[(txg + t) & TXG_MASK]));
+		}
+		if (msp->ms_allocator != -1) {
+			metaslab_passivate(msp, msp->ms_weight &
+			    ~METASLAB_ACTIVE_MASK);
+		}
+
+		if (!metaslab_debug_unload)
+			metaslab_unload(msp);
+	}
+}
+
 /*
 * Called after a transaction group has completely synced to mark
 * all of the metaslab's free space as usable.
@@ -3071,27 +3110,6 @@ metaslab_sync_done(metaslab_t *msp, uint64_t txg)
 	 */
 	metaslab_recalculate_weight_and_sort(msp);

-	/*
-	 * If the metaslab is loaded and we've not tried to load or allocate
-	 * from it in 'metaslab_unload_delay' txgs, then unload it.
-	 */
-	if (msp->ms_loaded &&
-	    msp->ms_disabled == 0 &&
-	    msp->ms_selected_txg + metaslab_unload_delay < txg) {
-
-		for (int t = 1; t < TXG_CONCURRENT_STATES; t++) {
-			VERIFY0(range_tree_space(
-			    msp->ms_allocating[(txg + t) & TXG_MASK]));
-		}
-		if (msp->ms_allocator != -1) {
-			metaslab_passivate(msp, msp->ms_weight &
-			    ~METASLAB_ACTIVE_MASK);
-		}
-
-		if (!metaslab_debug_unload)
-			metaslab_unload(msp);
-	}
-
 	ASSERT0(range_tree_space(msp->ms_allocating[txg & TXG_MASK]));
 	ASSERT0(range_tree_space(msp->ms_freeing));
 	ASSERT0(range_tree_space(msp->ms_freed));
--- a/Show More
+++ b/Show More