Tag zfs-2.1.11

META file and changelog updated.

Signed-off-by: Tony Hutter <hutter2@llnl.gov>

.github/workflows/build-dependencies.txt

@@ -0,0 +1,57 @@
+acl
+alien
+attr
+autoconf
+bc
+build-essential
+curl
+dbench
+debhelper-compat
+dh-python
+dkms
+fakeroot
+fio
+gdb
+gdebi
+git
+ksh
+lcov
+libacl1-dev
+libaio-dev
+libattr1-dev
+libblkid-dev
+libcurl4-openssl-dev
+libdevmapper-dev
+libelf-dev
+libffi-dev
+libmount-dev
+libpam0g-dev
+libselinux1-dev
+libssl-dev
+libtool
+libudev-dev
+linux-headers-generic
+lsscsi
+mdadm
+nfs-kernel-server
+pamtester
+parted
+po-debconf
+python3
+python3-all-dev
+python3-cffi
+python3-dev
+python3-packaging
+python3-pip
+python3-setuptools
+python3-sphinx
+rng-tools-debian
+rsync
+samba
+sysstat
+uuid-dev
+watchdog
+wget
+xfslibs-dev
+xz-utils
+zlib1g-dev

.github/workflows/checkstyle-dependencies.txt

@@ -0,0 +1,2 @@
+pax-utils
+shellcheck

.github/workflows/checkstyle.yaml

@@ -6,20 +6,27 @@ on:
 
 jobs:
   checkstyle:
-    runs-on: ubuntu-20.04
+    runs-on: ubuntu-22.04
     steps:
     - uses: actions/checkout@v3
       with:
         ref: ${{ github.event.pull_request.head.sha }}
     - name: Install dependencies
       run: |
-        sudo apt-get update
+        # https://github.com/orgs/community/discussions/47863
-        sudo apt-get install --yes -qq build-essential autoconf libtool gawk alien fakeroot linux-headers-$(uname -r)
+        sudo apt-mark hold grub-efi-amd64-signed
-        sudo apt-get install --yes -qq zlib1g-dev uuid-dev libattr1-dev libblkid-dev libselinux-dev libudev-dev libssl-dev python-dev python-setuptools python-cffi python3 python3-dev python3-setuptools python3-cffi
+        sudo apt-get update --fix-missing
-        # packages for tests
+        sudo apt-get upgrade
-        sudo apt-get install --yes -qq parted lsscsi ksh attr acl nfs-kernel-server fio
+        sudo xargs --arg-file=${{ github.workspace }}/.github/workflows/build-dependencies.txt apt-get install -qq
-        sudo apt-get install --yes -qq mandoc cppcheck pax-utils devscripts
+        sudo xargs --arg-file=${{ github.workspace }}/.github/workflows/checkstyle-dependencies.txt apt-get install -qq
-        sudo -E pip --quiet install flake8
+        sudo python3 -m pip install --quiet flake8
+        sudo apt-get clean
+
+        # confirm that the tools are installed
+        # the build system doesn't fail when they are not
+        flake8 --version
+        scanelf --version
+        shellcheck --version
     - name: Prepare
       run: |
         sh ./autogen.sh

.github/workflows/scripts/reclaim_disk_space.sh

@@ -0,0 +1,23 @@
+#!/bin/sh
+
+set -eu
+
+# remove 4GiB of images
+sudo systemd-run docker system prune --force --all --volumes
+
+# remove unused software
+sudo systemd-run --wait rm -rf \
+  "$AGENT_TOOLSDIRECTORY" \
+  /opt/* \
+  /usr/local/* \
+  /usr/share/az* \
+  /usr/share/dotnet \
+  /usr/share/gradle* \
+  /usr/share/miniconda \
+  /usr/share/swift \
+  /var/lib/gems \
+  /var/lib/mysql \
+  /var/lib/snapd
+
+# trim the cleaned space
+sudo fstrim /

.github/workflows/zfs-tests-functional.yml

@@ -9,7 +9,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        os: [18.04, 20.04]
+        os: [20.04, 22.04]
     runs-on: ubuntu-${{ matrix.os }}
     steps:
     - uses: actions/checkout@v3
@@ -17,17 +17,12 @@ jobs:
         ref: ${{ github.event.pull_request.head.sha }}
     - name: Install dependencies
       run: |
-        sudo apt-get update
+        # https://github.com/orgs/community/discussions/47863
-        sudo apt-get install --yes -qq build-essential autoconf libtool gdb lcov \
+        sudo apt-mark hold grub-efi-amd64-signed
-          git alien fakeroot wget curl bc fio acl \
+        sudo apt-get update --fix-missing
-          sysstat mdadm lsscsi parted gdebi attr dbench watchdog ksh \
+        sudo apt-get upgrade
-          nfs-kernel-server samba rng-tools xz-utils \
+        sudo xargs --arg-file=${{ github.workspace }}/.github/workflows/build-dependencies.txt apt-get install -qq
-          zlib1g-dev uuid-dev libblkid-dev libselinux-dev \
+        sudo apt-get clean
-          xfslibs-dev libattr1-dev libacl1-dev libudev-dev libdevmapper-dev \
-          libssl-dev libffi-dev libaio-dev libelf-dev libmount-dev \
-          libpam0g-dev pamtester python-dev python-setuptools python-cffi \
-          python3 python3-dev python3-setuptools python3-cffi python3-packaging \
-          libcurl4-openssl-dev
     - name: Autogen.sh
       run: |
         sh autogen.sh
@@ -56,12 +51,13 @@ jobs:
             sudo udevadm control --reload-rules
           fi
         fi
-        # Workaround to provide additional free space for testing.
+    - name: Clear the kernel ring buffer
-        #   https://github.com/actions/virtual-environments/issues/2840
+      run: |
-        sudo rm -rf /usr/share/dotnet
+        sudo dmesg -c >/var/tmp/dmesg-prerun
-        sudo rm -rf /opt/ghc
+    - name: Reclaim and report disk space
-        sudo rm -rf "/usr/local/share/boost"
+      run: |
-        sudo rm -rf "$AGENT_TOOLSDIRECTORY"
+        ${{ github.workspace }}/.github/workflows/scripts/reclaim_disk_space.sh
+        df -h /
     - name: Tests
       run: |
         /usr/share/zfs/zfs-tests.sh -vR -s 3G
@@ -71,7 +67,7 @@ jobs:
       run: |
         RESULTS_PATH=$(readlink -f /var/tmp/test_results/current)
         sudo dmesg > $RESULTS_PATH/dmesg
-        sudo cp /var/log/syslog $RESULTS_PATH/
+        sudo cp /var/log/syslog /var/tmp/dmesg-prerun $RESULTS_PATH/
         sudo chmod +r $RESULTS_PATH/*
         # Replace ':' in dir names, actions/upload-artifact doesn't support it
         for f in $(find /var/tmp/test_results -name '*:*'); do mv "$f" "${f//:/__}"; done
@@ -79,5 +75,7 @@ jobs:
       if: failure()
       with:
         name: Test logs Ubuntu-${{ matrix.os }}
-        path: /var/tmp/test_results/20*/
+        path: |
+          /var/tmp/test_results/*
+          !/var/tmp/test_results/current
         if-no-files-found: ignore

.github/workflows/zfs-tests-sanity.yml

@@ -6,24 +6,19 @@ on:
 
 jobs:
   tests:
-    runs-on: ubuntu-20.04
+    runs-on: ubuntu-22.04
     steps:
     - uses: actions/checkout@v3
       with:
         ref: ${{ github.event.pull_request.head.sha }}
     - name: Install dependencies
       run: |
-        sudo apt-get update
+        # https://github.com/orgs/community/discussions/47863
-        sudo apt-get install --yes -qq build-essential autoconf libtool gdb lcov \
+        sudo apt-mark hold grub-efi-amd64-signed
-          git alien fakeroot wget curl bc fio acl \
+        sudo apt-get update --fix-missing
-          sysstat mdadm lsscsi parted gdebi attr dbench watchdog ksh \
+        sudo apt-get upgrade
-          nfs-kernel-server samba rng-tools xz-utils \
+        sudo xargs --arg-file=${{ github.workspace }}/.github/workflows/build-dependencies.txt apt-get install -qq
-          zlib1g-dev uuid-dev libblkid-dev libselinux-dev \
+        sudo apt-get clean
-          xfslibs-dev libattr1-dev libacl1-dev libudev-dev libdevmapper-dev \
-          libssl-dev libffi-dev libaio-dev libelf-dev libmount-dev \
-          libpam0g-dev pamtester python-dev python-setuptools python-cffi \
-          python3 python3-dev python3-setuptools python3-cffi python3-packaging \
-          libcurl4-openssl-dev
     - name: Autogen.sh
       run: |
         sh autogen.sh
@@ -52,12 +47,13 @@ jobs:
             sudo udevadm control --reload-rules
           fi
         fi
-        # Workaround to provide additional free space for testing.
+    - name: Clear the kernel ring buffer
-        #   https://github.com/actions/virtual-environments/issues/2840
+      run: |
-        sudo rm -rf /usr/share/dotnet
+        sudo dmesg -c >/var/tmp/dmesg-prerun
-        sudo rm -rf /opt/ghc
+    - name: Reclaim and report disk space
-        sudo rm -rf "/usr/local/share/boost"
+      run: |
-        sudo rm -rf "$AGENT_TOOLSDIRECTORY"
+        ${{ github.workspace }}/.github/workflows/scripts/reclaim_disk_space.sh
+        df -h /
     - name: Tests
       run: |
         /usr/share/zfs/zfs-tests.sh -vR -s 3G -r sanity
@@ -67,13 +63,15 @@ jobs:
       run: |
         RESULTS_PATH=$(readlink -f /var/tmp/test_results/current)
         sudo dmesg > $RESULTS_PATH/dmesg
-        sudo cp /var/log/syslog $RESULTS_PATH/
+        sudo cp /var/log/syslog /var/tmp/dmesg-prerun $RESULTS_PATH/
         sudo chmod +r $RESULTS_PATH/*
         # Replace ':' in dir names, actions/upload-artifact doesn't support it
         for f in $(find /var/tmp/test_results -name '*:*'); do mv "$f" "${f//:/__}"; done
     - uses: actions/upload-artifact@v3
       if: failure()
       with:
-        name: Test logs
+        name: Test logs Ubuntu-${{ matrix.os }}
-        path: /var/tmp/test_results/20*/
+        path: |
+          /var/tmp/test_results/*
+          !/var/tmp/test_results/current
         if-no-files-found: ignore

.github/workflows/zloop.yml

@@ -6,7 +6,7 @@ on:
 
 jobs:
   tests:
-    runs-on: ubuntu-20.04
+    runs-on: ubuntu-22.04
     env:
       TEST_DIR: /var/tmp/zloop
     steps:
@@ -15,15 +15,12 @@ jobs:
         ref: ${{ github.event.pull_request.head.sha }}
     - name: Install dependencies
       run: |
-        sudo apt-get update
+        # https://github.com/orgs/community/discussions/47863
-        sudo apt-get install --yes -qq build-essential autoconf libtool gdb \
+        sudo apt-mark hold grub-efi-amd64-signed
-          git alien fakeroot \
+        sudo apt-get update --fix-missing
-          zlib1g-dev uuid-dev libblkid-dev libselinux-dev \
+        sudo apt-get upgrade
-          xfslibs-dev libattr1-dev libacl1-dev libudev-dev libdevmapper-dev \
+        sudo xargs --arg-file=${{ github.workspace }}/.github/workflows/build-dependencies.txt apt-get install -qq
-          libssl-dev libffi-dev libaio-dev libelf-dev libmount-dev \
+        sudo apt-get clean
-          libpam0g-dev \
-          python-dev python-setuptools python-cffi python-packaging \
-          python3 python3-dev python3-setuptools python3-cffi python3-packaging
     - name: Autogen.sh
       run: |
         sh autogen.sh
@@ -44,8 +41,9 @@ jobs:
     - name: Tests
       run: |
         sudo mkdir -p $TEST_DIR
-        # run for 20 minutes to have a total runner time of 30 minutes
+        # run for 10 minutes or at most 2 iterations for a maximum runner
-        sudo /usr/share/zfs/zloop.sh -t 1200 -l -m1 -- -T 120 -P 60
+        #   time of 20 minutes.
+        sudo /usr/share/zfs/zloop.sh -t 600 -I 2 -l -m1 -- -T 120 -P 60
     - name: Prepare artifacts
       if: failure()
       run: |

CODE_OF_CONDUCT.md

@@ -1,2 +1,2 @@
-The [OpenZFS Code of Conduct](http://www.open-zfs.org/wiki/Code_of_Conduct)
+The [OpenZFS Code of Conduct](https://openzfs.org/wiki/Code_of_Conduct)
 applies to spaces associated with the OpenZFS project, including GitHub.

META

@@ -1,10 +1,10 @@
 Meta:          1
 Name:          zfs
 Branch:        1.0
-Version:       2.1.7
+Version:       2.1.11
 Release:       1
 Release-Tags:  relext
 License:       CDDL
 Author:        OpenZFS
-Linux-Maximum: 6.0
+Linux-Maximum: 6.2
 Linux-Minimum: 3.10

cmd/arc_summary/Makefile.am

@@ -1,13 +1,8 @@
 bin_SCRIPTS = arc_summary
 
 CLEANFILES = arc_summary
-EXTRA_DIST = arc_summary2 arc_summary3
+EXTRA_DIST = arc_summary3
-
-if USING_PYTHON_2
-SCRIPT = arc_summary2
-else
 SCRIPT = arc_summary3
-endif
 
 arc_summary: $(SCRIPT)
 	cp $< $@

cmd/arc_summary/arc_summary3

@@ -191,21 +191,13 @@ elif sys.platform.startswith('linux'):
         # there, so we fall back on modinfo
         command = ["/sbin/modinfo", request, "-0"]
 
-        # The recommended way to do this is with subprocess.run(). However,
-        # some installed versions of Python are < 3.5, so we offer them
-        # the option of doing it the old way (for now)
         info = ''
 
         try:
 
-            if 'run' in dir(subprocess):
+            info = subprocess.run(command, stdout=subprocess.PIPE,
-                info = subprocess.run(command, stdout=subprocess.PIPE,
+                                  check=True, universal_newlines=True)
-                                      universal_newlines=True)
+            raw_output = info.stdout.split('\0')
-                raw_output = info.stdout.split('\0')
-            else:
-                info = subprocess.check_output(command,
-                                               universal_newlines=True)
-                raw_output = info.split('\0')
 
         except subprocess.CalledProcessError:
             print("Error: Descriptions not available",

cmd/arcstat/arcstat.in

@@ -47,7 +47,7 @@
 # @hdr is the array of fields that needs to be printed, so we
 # just iterate over this array and print the values using our pretty printer.
 #
-# This script must remain compatible with Python 2.6+ and Python 3.4+.
+# This script must remain compatible with Python 3.6+.
 #
 
 import sys

cmd/dbufstat/dbufstat.in

@@ -27,7 +27,7 @@
 # Copyright (C) 2013 Lawrence Livermore National Security, LLC.
 # Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
 #
-# This script must remain compatible with Python 2.6+ and Python 3.4+.
+# This script must remain compatible with and Python 3.6+.
 #
 
 import sys

cmd/zed/agents/zfs_agents.c

@@ -80,6 +80,7 @@ zfs_agent_iter_vdev(zpool_handle_t *zhp, nvlist_t *nvl, void *arg)
 	char *path = NULL;
 	uint_t c, children;
 	nvlist_t **child;
+	uint64_t vdev_guid;
 
 	/*
 	 * First iterate over any children.
@@ -100,7 +101,7 @@ zfs_agent_iter_vdev(zpool_handle_t *zhp, nvlist_t *nvl, void *arg)
 	    &child, &children) == 0) {
 		for (c = 0; c < children; c++) {
 			if (zfs_agent_iter_vdev(zhp, child[c], gsp)) {
-				gsp->gs_vdev_type = DEVICE_TYPE_L2ARC;
+				gsp->gs_vdev_type = DEVICE_TYPE_SPARE;
 				return (B_TRUE);
 			}
 		}
@@ -109,7 +110,7 @@ zfs_agent_iter_vdev(zpool_handle_t *zhp, nvlist_t *nvl, void *arg)
 	    &child, &children) == 0) {
 		for (c = 0; c < children; c++) {
 			if (zfs_agent_iter_vdev(zhp, child[c], gsp)) {
-				gsp->gs_vdev_type = DEVICE_TYPE_SPARE;
+				gsp->gs_vdev_type = DEVICE_TYPE_L2ARC;
 				return (B_TRUE);
 			}
 		}
@@ -126,6 +127,21 @@ zfs_agent_iter_vdev(zpool_handle_t *zhp, nvlist_t *nvl, void *arg)
 		    &gsp->gs_vdev_expandtime);
 		return (B_TRUE);
 	}
+	/*
+	 * Otherwise, on a vdev guid match, grab the devid and expansion
+	 * time. The devid might be missing on removal since its not part
+	 * of blkid cache and L2ARC VDEV does not contain pool guid in its
+	 * blkid, so this is a special case for L2ARC VDEV.
+	 */
+	else if (gsp->gs_vdev_guid != 0 && gsp->gs_devid == NULL &&
+	    nvlist_lookup_uint64(nvl, ZPOOL_CONFIG_GUID, &vdev_guid) == 0 &&
+	    gsp->gs_vdev_guid == vdev_guid) {
+		(void) nvlist_lookup_string(nvl, ZPOOL_CONFIG_DEVID,
+		    &gsp->gs_devid);
+		(void) nvlist_lookup_uint64(nvl, ZPOOL_CONFIG_EXPANSION_TIME,
+		    &gsp->gs_vdev_expandtime);
+		return (B_TRUE);
+	}
 
 	return (B_FALSE);
 }
@@ -148,13 +164,13 @@ zfs_agent_iter_pool(zpool_handle_t *zhp, void *arg)
 	/*
 	 * if a match was found then grab the pool guid
 	 */
-	if (gsp->gs_vdev_guid) {
+	if (gsp->gs_vdev_guid && gsp->gs_devid) {
 		(void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
 		    &gsp->gs_pool_guid);
 	}
 
 	zpool_close(zhp);
-	return (gsp->gs_vdev_guid != 0);
+	return (gsp->gs_devid != NULL && gsp->gs_vdev_guid != 0);
 }
 
 void
@@ -195,11 +211,13 @@ zfs_agent_post_event(const char *class, const char *subclass, nvlist_t *nvl)
 		uint64_t pool_guid = 0, vdev_guid = 0;
 		guid_search_t search = { 0 };
 		device_type_t devtype = DEVICE_TYPE_PRIMARY;
+		char *devid = NULL;
 
 		class = "resource.fs.zfs.removed";
 		subclass = "";
 
 		(void) nvlist_add_string(payload, FM_CLASS, class);
+		(void) nvlist_lookup_string(nvl, DEV_IDENTIFIER, &devid);
 		(void) nvlist_lookup_uint64(nvl, ZFS_EV_POOL_GUID, &pool_guid);
 		(void) nvlist_lookup_uint64(nvl, ZFS_EV_VDEV_GUID, &vdev_guid);
 
@@ -209,20 +227,24 @@ zfs_agent_post_event(const char *class, const char *subclass, nvlist_t *nvl)
 		(void) nvlist_add_int64_array(payload, FM_EREPORT_TIME, tod, 2);
 
 		/*
+		 * If devid is missing but vdev_guid is available, find devid
+		 * and pool_guid from vdev_guid.
 		 * For multipath, spare and l2arc devices ZFS_EV_VDEV_GUID or
 		 * ZFS_EV_POOL_GUID may be missing so find them.
 		 */
-		if (pool_guid == 0 || vdev_guid == 0) {
+		if (devid == NULL || pool_guid == 0 || vdev_guid == 0) {
-			if ((nvlist_lookup_string(nvl, DEV_IDENTIFIER,
+			if (devid == NULL)
-			    &search.gs_devid) == 0) &&
+				search.gs_vdev_guid = vdev_guid;
-			    (zpool_iter(g_zfs_hdl, zfs_agent_iter_pool, &search)
+			else
-			    == 1)) {
+				search.gs_devid = devid;
-				if (pool_guid == 0)
+			zpool_iter(g_zfs_hdl, zfs_agent_iter_pool, &search);
-					pool_guid = search.gs_pool_guid;
+			if (devid == NULL)
-				if (vdev_guid == 0)
+				devid = search.gs_devid;
-					vdev_guid = search.gs_vdev_guid;
+			if (pool_guid == 0)
-				devtype = search.gs_vdev_type;
+				pool_guid = search.gs_pool_guid;
-			}
+			if (vdev_guid == 0)
+				vdev_guid = search.gs_vdev_guid;
+			devtype = search.gs_vdev_type;
 		}
 
 		/*
@@ -235,7 +257,9 @@ zfs_agent_post_event(const char *class, const char *subclass, nvlist_t *nvl)
 		    search.gs_vdev_expandtime + 10 > tv.tv_sec) {
 			zed_log_msg(LOG_INFO, "agent post event: ignoring '%s' "
 			    "for recently expanded device '%s'", EC_DEV_REMOVE,
-			    search.gs_devid);
+			    devid);
+			fnvlist_free(payload);
+			free(event);
 			goto out;
 		}
 

cmd/zed/agents/zfs_mod.c

@@ -185,10 +185,12 @@ zfs_process_add(zpool_handle_t *zhp, nvlist_t *vdev, boolean_t labeled)
 	uint64_t wholedisk = 0ULL;
 	uint64_t offline = 0ULL, faulted = 0ULL;
 	uint64_t guid = 0ULL;
+	uint64_t is_spare = 0;
 	char *physpath = NULL, *new_devid = NULL, *enc_sysfs_path = NULL;
 	char rawpath[PATH_MAX], fullpath[PATH_MAX];
 	char devpath[PATH_MAX];
 	int ret;
+	int online_flag = ZFS_ONLINE_CHECKREMOVE | ZFS_ONLINE_UNSPARE;
 	boolean_t is_sd = B_FALSE;
 	boolean_t is_mpath_wholedisk = B_FALSE;
 	uint_t c;
@@ -214,6 +216,7 @@ zfs_process_add(zpool_handle_t *zhp, nvlist_t *vdev, boolean_t labeled)
 	(void) nvlist_lookup_uint64(vdev, ZPOOL_CONFIG_FAULTED, &faulted);
 
 	(void) nvlist_lookup_uint64(vdev, ZPOOL_CONFIG_GUID, &guid);
+	(void) nvlist_lookup_uint64(vdev, ZPOOL_CONFIG_IS_SPARE, &is_spare);
 
 	/*
 	 * Special case:
@@ -304,11 +307,13 @@ zfs_process_add(zpool_handle_t *zhp, nvlist_t *vdev, boolean_t labeled)
 		}
 	}
 
+	if (is_spare)
+		online_flag |= ZFS_ONLINE_SPARE;
+
 	/*
 	 * Attempt to online the device.
 	 */
-	if (zpool_vdev_online(zhp, fullpath,
+	if (zpool_vdev_online(zhp, fullpath, online_flag, &newstate) == 0 &&
-	    ZFS_ONLINE_CHECKREMOVE | ZFS_ONLINE_UNSPARE, &newstate) == 0 &&
 	    (newstate == VDEV_STATE_HEALTHY ||
 	    newstate == VDEV_STATE_DEGRADED)) {
 		zed_log_msg(LOG_INFO,
@@ -525,7 +530,9 @@ typedef struct dev_data {
 	boolean_t		dd_islabeled;
 	uint64_t		dd_pool_guid;
 	uint64_t		dd_vdev_guid;
+	uint64_t		dd_new_vdev_guid;
 	const char		*dd_new_devid;
+	uint64_t		dd_num_spares;
 } dev_data_t;
 
 static void
@@ -535,6 +542,8 @@ zfs_iter_vdev(zpool_handle_t *zhp, nvlist_t *nvl, void *data)
 	char *path = NULL;
 	uint_t c, children;
 	nvlist_t **child;
+	uint64_t guid = 0;
+	uint64_t isspare = 0;
 
 	/*
 	 * First iterate over any children.
@@ -560,19 +569,16 @@ zfs_iter_vdev(zpool_handle_t *zhp, nvlist_t *nvl, void *data)
 	}
 
 	/* once a vdev was matched and processed there is nothing left to do */
-	if (dp->dd_found)
+	if (dp->dd_found && dp->dd_num_spares == 0)
 		return;
+	(void) nvlist_lookup_uint64(nvl, ZPOOL_CONFIG_GUID, &guid);
 
 	/*
 	 * Match by GUID if available otherwise fallback to devid or physical
 	 */
 	if (dp->dd_vdev_guid != 0) {
-		uint64_t guid;
+		if (guid != dp->dd_vdev_guid)
-
-		if (nvlist_lookup_uint64(nvl, ZPOOL_CONFIG_GUID,
-		    &guid) != 0 || guid != dp->dd_vdev_guid) {
 			return;
-		}
 		zed_log_msg(LOG_INFO, "  zfs_iter_vdev: matched on %llu", guid);
 		dp->dd_found = B_TRUE;
 
@@ -582,6 +588,12 @@ zfs_iter_vdev(zpool_handle_t *zhp, nvlist_t *nvl, void *data)
 		 * illumos, substring matching is not required to accommodate
 		 * the partition suffix. An exact match will be present in
 		 * the dp->dd_compare value.
+		 * If the attached disk already contains a vdev GUID, it means
+		 * the disk is not clean. In such a scenario, the physical path
+		 * would be a match that makes the disk faulted when trying to
+		 * online it. So, we would only want to proceed if either GUID
+		 * matches with the last attached disk or the disk is in clean
+		 * state.
 		 */
 		if (nvlist_lookup_string(nvl, dp->dd_prop, &path) != 0 ||
 		    strcmp(dp->dd_compare, path) != 0) {
@@ -589,6 +601,12 @@ zfs_iter_vdev(zpool_handle_t *zhp, nvlist_t *nvl, void *data)
 			    __func__, dp->dd_compare, path);
 			return;
 		}
+		if (dp->dd_new_vdev_guid != 0 && dp->dd_new_vdev_guid != guid) {
+			zed_log_msg(LOG_INFO, "  %s: no match (GUID:%llu"
+			    " != vdev GUID:%llu)", __func__,
+			    dp->dd_new_vdev_guid, guid);
+			return;
+		}
 
 		zed_log_msg(LOG_INFO, "  zfs_iter_vdev: matched %s on %s",
 		    dp->dd_prop, path);
@@ -601,6 +619,10 @@ zfs_iter_vdev(zpool_handle_t *zhp, nvlist_t *nvl, void *data)
 		}
 	}
 
+	if (dp->dd_found == B_TRUE && nvlist_lookup_uint64(nvl,
+	    ZPOOL_CONFIG_IS_SPARE, &isspare) == 0 && isspare)
+		dp->dd_num_spares++;
+
 	(dp->dd_func)(zhp, nvl, dp->dd_islabeled);
 }
 
@@ -661,7 +683,9 @@ zfs_iter_pool(zpool_handle_t *zhp, void *data)
 	}
 
 	zpool_close(zhp);
-	return (dp->dd_found);	/* cease iteration after a match */
+
+	/* cease iteration after a match */
+	return (dp->dd_found && dp->dd_num_spares == 0);
 }
 
 /*
@@ -670,7 +694,7 @@ zfs_iter_pool(zpool_handle_t *zhp, void *data)
  */
 static boolean_t
 devphys_iter(const char *physical, const char *devid, zfs_process_func_t func,
-    boolean_t is_slice)
+    boolean_t is_slice, uint64_t new_vdev_guid)
 {
 	dev_data_t data = { 0 };
 
@@ -680,6 +704,7 @@ devphys_iter(const char *physical, const char *devid, zfs_process_func_t func,
 	data.dd_found = B_FALSE;
 	data.dd_islabeled = is_slice;
 	data.dd_new_devid = devid;	/* used by auto replace code */
+	data.dd_new_vdev_guid = new_vdev_guid;
 
 	(void) zpool_iter(g_zfshdl, zfs_iter_pool, &data);
 
@@ -848,7 +873,7 @@ zfs_deliver_add(nvlist_t *nvl, boolean_t is_lofi)
 	if (devid_iter(devid, zfs_process_add, is_slice))
 		return (0);
 	if (devpath != NULL && devphys_iter(devpath, devid, zfs_process_add,
-	    is_slice))
+	    is_slice, vdev_guid))
 		return (0);
 	if (vdev_guid != 0)
 		(void) guid_iter(pool_guid, vdev_guid, devid, zfs_process_add,

cmd/zed/agents/zfs_retire.c

@@ -75,6 +75,8 @@ typedef struct find_cbdata {
 	uint64_t	cb_guid;
 	zpool_handle_t	*cb_zhp;
 	nvlist_t	*cb_vdev;
+	uint64_t	cb_vdev_guid;
+	uint64_t	cb_num_spares;
 } find_cbdata_t;
 
 static int
@@ -140,6 +142,64 @@ find_vdev(libzfs_handle_t *zhdl, nvlist_t *nv, uint64_t search_guid)
 	return (NULL);
 }
 
+static int
+remove_spares(zpool_handle_t *zhp, void *data)
+{
+	nvlist_t *config, *nvroot;
+	nvlist_t **spares;
+	uint_t nspares;
+	char *devname;
+	find_cbdata_t *cbp = data;
+	uint64_t spareguid = 0;
+	vdev_stat_t *vs;
+	unsigned int c;
+
+	config = zpool_get_config(zhp, NULL);
+	if (nvlist_lookup_nvlist(config,
+	    ZPOOL_CONFIG_VDEV_TREE, &nvroot) != 0) {
+		zpool_close(zhp);
+		return (0);
+	}
+
+	if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
+	    &spares, &nspares) != 0) {
+		zpool_close(zhp);
+		return (0);
+	}
+
+	for (int i = 0; i < nspares; i++) {
+		if (nvlist_lookup_uint64(spares[i], ZPOOL_CONFIG_GUID,
+		    &spareguid) == 0 && spareguid == cbp->cb_vdev_guid) {
+			devname = zpool_vdev_name(NULL, zhp, spares[i],
+			    B_FALSE);
+			nvlist_lookup_uint64_array(spares[i],
+			    ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &c);
+			if (vs->vs_state != VDEV_STATE_REMOVED &&
+			    zpool_vdev_remove_wanted(zhp, devname) == 0)
+				cbp->cb_num_spares++;
+			break;
+		}
+	}
+
+	zpool_close(zhp);
+	return (0);
+}
+
+/*
+ * Given a vdev guid, find and remove all spares associated with it.
+ */
+static int
+find_and_remove_spares(libzfs_handle_t *zhdl, uint64_t vdev_guid)
+{
+	find_cbdata_t cb;
+
+	cb.cb_num_spares = 0;
+	cb.cb_vdev_guid = vdev_guid;
+	zpool_iter(zhdl, remove_spares, &cb);
+
+	return (cb.cb_num_spares);
+}
+
 /*
  * Given a (pool, vdev) GUID pair, find the matching pool and vdev.
  */
@@ -315,6 +375,8 @@ zfs_retire_recv(fmd_hdl_t *hdl, fmd_event_t *ep, nvlist_t *nvl,
 	libzfs_handle_t *zhdl = zdp->zrd_hdl;
 	boolean_t fault_device, degrade_device;
 	boolean_t is_repair;
+	boolean_t l2arc = B_FALSE;
+	boolean_t spare = B_FALSE;
 	char *scheme;
 	nvlist_t *vdev = NULL;
 	char *uuid;
@@ -323,6 +385,8 @@ zfs_retire_recv(fmd_hdl_t *hdl, fmd_event_t *ep, nvlist_t *nvl,
 	boolean_t is_disk;
 	vdev_aux_t aux;
 	uint64_t state = 0;
+	vdev_stat_t *vs;
+	unsigned int c;
 
 	fmd_hdl_debug(hdl, "zfs_retire_recv: '%s'", class);
 
@@ -339,10 +403,26 @@ zfs_retire_recv(fmd_hdl_t *hdl, fmd_event_t *ep, nvlist_t *nvl,
 		char *devtype;
 		char *devname;
 
+		if (nvlist_lookup_string(nvl, FM_EREPORT_PAYLOAD_ZFS_VDEV_TYPE,
+		    &devtype) == 0) {
+			if (strcmp(devtype, VDEV_TYPE_SPARE) == 0)
+				spare = B_TRUE;
+			else if (strcmp(devtype, VDEV_TYPE_L2CACHE) == 0)
+				l2arc = B_TRUE;
+		}
+
+		if (nvlist_lookup_uint64(nvl,
+		    FM_EREPORT_PAYLOAD_ZFS_VDEV_GUID, &vdev_guid) != 0)
+			return;
+
+		if (spare) {
+			int nspares = find_and_remove_spares(zhdl, vdev_guid);
+			fmd_hdl_debug(hdl, "%d spares removed", nspares);
+			return;
+		}
+
 		if (nvlist_lookup_uint64(nvl, FM_EREPORT_PAYLOAD_ZFS_POOL_GUID,
-		    &pool_guid) != 0 ||
+		    &pool_guid) != 0)
-		    nvlist_lookup_uint64(nvl, FM_EREPORT_PAYLOAD_ZFS_VDEV_GUID,
-		    &vdev_guid) != 0)
 			return;
 
 		if ((zhp = find_by_guid(zhdl, pool_guid, vdev_guid,
@@ -351,13 +431,28 @@ zfs_retire_recv(fmd_hdl_t *hdl, fmd_event_t *ep, nvlist_t *nvl,
 
 		devname = zpool_vdev_name(NULL, zhp, vdev, B_FALSE);
 
-		/* Can't replace l2arc with a spare: offline the device */
+		nvlist_lookup_uint64_array(vdev, ZPOOL_CONFIG_VDEV_STATS,
-		if (nvlist_lookup_string(nvl, FM_EREPORT_PAYLOAD_ZFS_VDEV_TYPE,
+		    (uint64_t **)&vs, &c);
-		    &devtype) == 0 && strcmp(devtype, VDEV_TYPE_L2CACHE) == 0) {
+
-			fmd_hdl_debug(hdl, "zpool_vdev_offline '%s'", devname);
+		/*
-			zpool_vdev_offline(zhp, devname, B_TRUE);
+		 * If state removed is requested for already removed vdev,
-		} else if (!fmd_prop_get_int32(hdl, "spare_on_remove") ||
+		 * its a loopback event from spa_async_remove(). Just
-		    replace_with_spare(hdl, zhp, vdev) == B_FALSE) {
+		 * ignore it.
+		 */
+		if (vs->vs_state == VDEV_STATE_REMOVED &&
+		    state == VDEV_STATE_REMOVED)
+			return;
+
+		/* Remove the vdev since device is unplugged */
+		if (l2arc || (strcmp(class, "resource.fs.zfs.removed") == 0)) {
+			int status = zpool_vdev_remove_wanted(zhp, devname);
+			fmd_hdl_debug(hdl, "zpool_vdev_remove_wanted '%s'"
+			    ", ret:%d", devname, status);
+		}
+
+		/* Replace the vdev with a spare if its not a l2arc */
+		if (!l2arc && (!fmd_prop_get_int32(hdl, "spare_on_remove") ||
+		    replace_with_spare(hdl, zhp, vdev) == B_FALSE)) {
 			/* Could not handle with spare */
 			fmd_hdl_debug(hdl, "no spare for '%s'", devname);
 		}

cmd/zed/zed.d/statechange-notify.sh

@@ -37,7 +37,7 @@ if [ "${ZEVENT_VDEV_STATE_STR}" != "FAULTED" ] \
 fi
 
 umask 077
-note_subject="ZFS device fault for pool ${ZEVENT_POOL_GUID} on $(hostname)"
+note_subject="ZFS device fault for pool ${ZEVENT_POOL} on $(hostname)"
 note_pathname="$(mktemp)"
 {
     if [ "${ZEVENT_VDEV_STATE_STR}" = "FAULTED" ] ; then
@@ -65,7 +65,7 @@ note_pathname="$(mktemp)"
     [ -n "${ZEVENT_VDEV_GUID}" ] && echo "  vguid: ${ZEVENT_VDEV_GUID}"
     [ -n "${ZEVENT_VDEV_DEVID}" ] && echo "  devid: ${ZEVENT_VDEV_DEVID}"
 
-    echo "   pool: ${ZEVENT_POOL_GUID}"
+    echo "   pool: ${ZEVENT_POOL} (${ZEVENT_POOL_GUID})"
 
 } > "${note_pathname}"
 

cmd/zfs/zfs_main.c

@@ -315,14 +315,14 @@ get_usage(zfs_help_t idx)
 	case HELP_ROLLBACK:
 		return (gettext("\trollback [-rRf] <snapshot>\n"));
 	case HELP_SEND:
-		return (gettext("\tsend [-DnPpRvLecwhb] [-[i|I] snapshot] "
+		return (gettext("\tsend [-DnPpRVvLecwhb] [-[i|I] snapshot] "
 		    "<snapshot>\n"
-		    "\tsend [-DnvPLecw] [-i snapshot|bookmark] "
+		    "\tsend [-DnVvPLecw] [-i snapshot|bookmark] "
 		    "<filesystem|volume|snapshot>\n"
-		    "\tsend [-DnPpvLec] [-i bookmark|snapshot] "
+		    "\tsend [-DnPpVvLec] [-i bookmark|snapshot] "
 		    "--redact <bookmark> <snapshot>\n"
-		    "\tsend [-nvPe] -t <receive_resume_token>\n"
+		    "\tsend [-nVvPe] -t <receive_resume_token>\n"
-		    "\tsend [-Pnv] --saved filesystem\n"));
+		    "\tsend [-PnVv] --saved filesystem\n"));
 	case HELP_SET:
 		return (gettext("\tset <property=value> ... "
 		    "<filesystem|volume|snapshot> ...\n"));
@@ -3474,6 +3474,8 @@ print_header(list_cbdata_t *cb)
 	boolean_t first = B_TRUE;
 	boolean_t right_justify;
 
+	color_start(ANSI_BOLD);
+
 	for (; pl != NULL; pl = pl->pl_next) {
 		if (!first) {
 			(void) printf("  ");
@@ -3500,9 +3502,31 @@ print_header(list_cbdata_t *cb)
 			(void) printf("%-*s", (int)pl->pl_width, header);
 	}
 
+	color_end();
+
 	(void) printf("\n");
 }
 
+/*
+ * Decides on the color that the avail value should be printed in.
+ * > 80% used = yellow
+ * > 90% used = red
+ */
+static const char *
+zfs_list_avail_color(zfs_handle_t *zhp)
+{
+	uint64_t used = zfs_prop_get_int(zhp, ZFS_PROP_USED);
+	uint64_t avail = zfs_prop_get_int(zhp, ZFS_PROP_AVAILABLE);
+	int percentage = (int)((double)avail / MAX(avail + used, 1) * 100);
+
+	if (percentage > 20)
+		return (NULL);
+	else if (percentage > 10)
+		return (ANSI_YELLOW);
+	else
+		return (ANSI_RED);
+}
+
 /*
  * Given a dataset and a list of fields, print out all the properties according
  * to the described layout.
@@ -3565,6 +3589,22 @@ print_dataset(zfs_handle_t *zhp, list_cbdata_t *cb)
 			right_justify = B_FALSE;
 		}
 
+		/*
+		 * zfs_list_avail_color() needs ZFS_PROP_AVAILABLE + USED
+		 * - so we need another for() search for the USED part
+		 * - when no colors wanted, we can skip the whole thing
+		 */
+		if (use_color() && pl->pl_prop == ZFS_PROP_AVAILABLE) {
+			zprop_list_t *pl2 = cb->cb_proplist;
+			for (; pl2 != NULL; pl2 = pl2->pl_next) {
+				if (pl2->pl_prop == ZFS_PROP_USED) {
+					color_start(zfs_list_avail_color(zhp));
+					/* found it, no need for more loops */
+					break;
+				}
+			}
+		}
+
 		/*
 		 * If this is being called in scripted mode, or if this is the
 		 * last column and it is left-justified, don't include a width
@@ -3576,6 +3616,9 @@ print_dataset(zfs_handle_t *zhp, list_cbdata_t *cb)
 			(void) printf("%*s", (int)pl->pl_width, propstr);
 		else
 			(void) printf("%-*s", (int)pl->pl_width, propstr);
+
+		if (pl->pl_prop == ZFS_PROP_AVAILABLE)
+			color_end();
 	}
 
 	(void) printf("\n");
@@ -4407,6 +4450,7 @@ zfs_do_send(int argc, char **argv)
 		{"props",	no_argument,		NULL, 'p'},
 		{"parsable",	no_argument,		NULL, 'P'},
 		{"dedup",	no_argument,		NULL, 'D'},
+		{"proctitle",	no_argument,		NULL, 'V'},
 		{"verbose",	no_argument,		NULL, 'v'},
 		{"dryrun",	no_argument,		NULL, 'n'},
 		{"large-block",	no_argument,		NULL, 'L'},
@@ -4421,7 +4465,7 @@ zfs_do_send(int argc, char **argv)
 	};
 
 	/* check options */
-	while ((c = getopt_long(argc, argv, ":i:I:RsDpvnPLeht:cwbd:S",
+	while ((c = getopt_long(argc, argv, ":i:I:RsDpVvnPLeht:cwbd:S",
 	    long_options, NULL)) != -1) {
 		switch (c) {
 		case 'i':
@@ -4456,6 +4500,9 @@ zfs_do_send(int argc, char **argv)
 		case 'P':
 			flags.parsable = B_TRUE;
 			break;
+		case 'V':
+			flags.progressastitle = B_TRUE;
+			break;
 		case 'v':
 			flags.verbosity++;
 			flags.progress = B_TRUE;
@@ -4531,7 +4578,7 @@ zfs_do_send(int argc, char **argv)
 		}
 	}
 
-	if (flags.parsable && flags.verbosity == 0)
+	if ((flags.parsable || flags.progressastitle) && flags.verbosity == 0)
 		flags.verbosity = 1;
 
 	argc -= optind;
@@ -8693,6 +8740,8 @@ main(int argc, char **argv)
 
 	libzfs_print_on_error(g_zfs, B_TRUE);
 
+	zfs_setproctitle_init(argc, argv, environ);
+
 	/*
 	 * Many commands modify input strings for string parsing reasons.
 	 * We create a copy to protect the original argv.

cmd/zpool/zpool_main.c

@@ -4205,6 +4205,8 @@ print_iostat_header_impl(iostat_cbdata_t *cb, unsigned int force_column_width,
 	unsigned int namewidth;
 	const char *title;
 
+	color_start(ANSI_BOLD);
+
 	if (cb->cb_flags & IOS_ANYHISTO_M) {
 		title = histo_to_title[IOS_HISTO_IDX(cb->cb_flags)];
 	} else if (cb->cb_vdev_names_count) {
@@ -4238,6 +4240,8 @@ print_iostat_header_impl(iostat_cbdata_t *cb, unsigned int force_column_width,
 	if (cb->vcdl != NULL)
 		print_cmd_columns(cb->vcdl, 1);
 
+	color_end();
+
 	printf("\n");
 }
 
@@ -4247,6 +4251,37 @@ print_iostat_header(iostat_cbdata_t *cb)
 	print_iostat_header_impl(cb, 0, NULL);
 }
 
+/*
+ * Prints a size string (i.e. 120M) with the suffix ("M") colored
+ * by order of magnitude. Uses column_size to add padding.
+ */
+static void
+print_stat_color(const char *statbuf, unsigned int column_size)
+{
+	fputs("  ", stdout);
+	size_t len = strlen(statbuf);
+	while (len < column_size) {
+		fputc(' ', stdout);
+		column_size--;
+	}
+	if (*statbuf == '0') {
+		color_start(ANSI_GRAY);
+		fputc('0', stdout);
+	} else {
+		for (; *statbuf; statbuf++) {
+			if (*statbuf == 'K') color_start(ANSI_GREEN);
+			else if (*statbuf == 'M') color_start(ANSI_YELLOW);
+			else if (*statbuf == 'G') color_start(ANSI_RED);
+			else if (*statbuf == 'T') color_start(ANSI_BOLD_BLUE);
+			else if (*statbuf == 'P') color_start(ANSI_MAGENTA);
+			else if (*statbuf == 'E') color_start(ANSI_CYAN);
+			fputc(*statbuf, stdout);
+			if (--column_size <= 0)
+				break;
+		}
+	}
+	color_end();
+}
 
 /*
  * Display a single statistic.
@@ -4262,7 +4297,7 @@ print_one_stat(uint64_t value, enum zfs_nicenum_format format,
 	if (scripted)
 		printf("\t%s", buf);
 	else
-		printf("  %*s", column_size, buf);
+		print_stat_color(buf, column_size);
 }
 
 /*
@@ -5414,7 +5449,13 @@ print_zpool_dir_scripts(char *dirpath)
 	if ((dir = opendir(dirpath)) != NULL) {
 		/* print all the files and directories within directory */
 		while ((ent = readdir(dir)) != NULL) {
-			sprintf(fullpath, "%s/%s", dirpath, ent->d_name);
+			if (snprintf(fullpath, sizeof (fullpath), "%s/%s",
+			    dirpath, ent->d_name) >= sizeof (fullpath)) {
+				(void) fprintf(stderr,
+				    gettext("internal error: "
+				    "ZPOOL_SCRIPTS_PATH too large.\n"));
+				exit(1);
+			}
 
 			/* Print the scripts */
 			if (stat(fullpath, &dir_stat) == 0)

cmd/ztest/ztest.c

@@ -1184,7 +1184,7 @@ ztest_kill(ztest_shared_t *zs)
 	 * See comment above spa_write_cachefile().
 	 */
 	mutex_enter(&spa_namespace_lock);
-	spa_write_cachefile(ztest_spa, B_FALSE, B_FALSE);
+	spa_write_cachefile(ztest_spa, B_FALSE, B_FALSE, B_FALSE);
 	mutex_exit(&spa_namespace_lock);
 
 	(void) kill(getpid(), SIGKILL);

config/always-python.m4

@@ -1,7 +1,6 @@
 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
+dnl # with Python 3.6. This option is intended to
-dnl # and used with either interpreter.  This option is intended to
 dnl # to provide a method to specify the default system version, and
 dnl # set the PYTHON environment variable accordingly.
 dnl #
@@ -13,9 +12,7 @@ AC_DEFUN([ZFS_AC_CONFIG_ALWAYS_PYTHON], [
 		[with_python=check])
 
 	AS_CASE([$with_python],
-		[check], [AC_CHECK_PROGS([PYTHON], [python3 python2], [:])],
+		[check], [AC_CHECK_PROGS([PYTHON], [python3], [:])],
-		[2*], [PYTHON="python${with_python}"],
-		[*python2*], [PYTHON="${with_python}"],
 		[3*], [PYTHON="python${with_python}"],
 		[*python3*], [PYTHON="${with_python}"],
 		[no], [PYTHON=":"],
@@ -23,8 +20,7 @@ AC_DEFUN([ZFS_AC_CONFIG_ALWAYS_PYTHON], [
 	)
 
 	dnl #
-	dnl # Minimum supported Python versions for utilities:
+	dnl # Minimum supported Python versions for utilities: Python 3.6
-	dnl # Python 2.6 or Python 3.4
 	dnl #
 	AM_PATH_PYTHON([], [], [:])
 	AS_IF([test -z "$PYTHON_VERSION"], [
@@ -33,25 +29,16 @@ AC_DEFUN([ZFS_AC_CONFIG_ALWAYS_PYTHON], [
 	PYTHON_MINOR=${PYTHON_VERSION#*\.}
 
 	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],
+			AS_IF([test $PYTHON_MINOR -lt 6],
-				[AC_MSG_ERROR("Python >= 3.4 is required")])
+				[AC_MSG_ERROR("Python >= 3.6 is required")])
 		],
 		[:|2|3], [],
 		[PYTHON_VERSION=3]
 	)
 
 	AM_CONDITIONAL([USING_PYTHON], [test "$PYTHON" != :])
-	AM_CONDITIONAL([USING_PYTHON_2], [test "x${PYTHON_VERSION%%\.*}" = x2])
+	AC_SUBST([PYTHON_SHEBANG], [python3])
-	AM_CONDITIONAL([USING_PYTHON_3], [test "x${PYTHON_VERSION%%\.*}" = x3])
-
-	AM_COND_IF([USING_PYTHON_2],
-		[AC_SUBST([PYTHON_SHEBANG], [python2])],
-		[AC_SUBST([PYTHON_SHEBANG], [python3])])
 
 	dnl #
 	dnl # Request that packages be built for a specific Python version.

config/always-pyzfs.m4

@@ -18,7 +18,7 @@ AC_DEFUN([ZFS_AC_PYTHON_MODULE], [
 ])
 
 dnl #
-dnl # Determines if pyzfs can be built, requires Python 2.7 or later.
+dnl # Determines if pyzfs can be built, requires Python 3.6 or later.
 dnl #
 AC_DEFUN([ZFS_AC_CONFIG_ALWAYS_PYZFS], [
 	AC_ARG_ENABLE([pyzfs],
@@ -72,12 +72,11 @@ AC_DEFUN([ZFS_AC_CONFIG_ALWAYS_PYZFS], [
 	])
 
 	dnl #
-	dnl # Require python-devel libraries
+	dnl # Require python3-devel libraries
 	dnl #
 	AS_IF([test "x$enable_pyzfs" = xcheck  -o "x$enable_pyzfs" = xyes], [
 		AS_CASE([$PYTHON_VERSION],
-			[3.*], [PYTHON_REQUIRED_VERSION=">= '3.4.0'"],
+			[3.*], [PYTHON_REQUIRED_VERSION=">= '3.6.0'"],
-			[2.*], [PYTHON_REQUIRED_VERSION=">= '2.7.0'"],
 			[AC_MSG_ERROR("Python $PYTHON_VERSION unknown")]
 		)
 

config/ax_python_devel.m4

@@ -97,23 +97,13 @@ AC_DEFUN([AX_PYTHON_DEVEL],[
 	# Check for a version of Python >= 2.1.0
 	#
 	AC_MSG_CHECKING([for a version of Python >= '2.1.0'])
-	ac_supports_python_ver=`cat<<EOD | $PYTHON -
+	ac_supports_python_ver=`$PYTHON -c "import sys; \
-from __future__ import print_function;
+		ver = sys.version.split ()[[0]]; \
-import sys;
+		print (ver >= '2.1.0')"`
-try:
-	from packaging import version;
-except ImportError:
-	from distlib import version;
-ver = sys.version.split ()[[0]];
-(tst_cmp, tst_ver) = ">= '2.1.0'".split ();
-tst_ver = tst_ver.strip ("'");
-eval ("print (version.LegacyVersion (ver)"+ tst_cmp +"version.LegacyVersion (tst_ver))")
-EOD`
 	if test "$ac_supports_python_ver" != "True"; then
 		if test -z "$PYTHON_NOVERSIONCHECK"; then
 			AC_MSG_RESULT([no])
-			m4_ifvaln([$2],[$2],[
+			AC_MSG_FAILURE([
-				AC_MSG_FAILURE([
 This version of the AC@&t@_PYTHON_DEVEL macro
 doesn't work properly with versions of Python before
 2.1.0. You may need to re-run configure, setting the
@@ -122,7 +112,6 @@ PYTHON_EXTRA_LIBS and PYTHON_EXTRA_LDFLAGS by hand.
 Moreover, to disable this check, set PYTHON_NOVERSIONCHECK
 to something else than an empty string.
 ])
-			])
 		else
 			AC_MSG_RESULT([skip at user request])
 		fi
@@ -131,37 +120,47 @@ to something else than an empty string.
 	fi
 
 	#
-	# if the macro parameter ``version'' is set, honour it
+	# If the macro parameter ``version'' is set, honour it.
+	# A Python shim class, VPy, is used to implement correct version comparisons via
+	# string expressions, since e.g. a naive textual ">= 2.7.3" won't work for
+	# Python 2.7.10 (the ".1" being evaluated as less than ".3").
 	#
 	if test -n "$1"; then
 		AC_MSG_CHECKING([for a version of Python $1])
-		# Why the strip ()?  Because if we don't, version.parse
+                cat << EOF > ax_python_devel_vpy.py
-		# will, for example, report 3.10.0 >= '3.11.0'
+class VPy:
-		ac_supports_python_ver=`cat<<EOD | $PYTHON -
+    def vtup(self, s):
-
+        return tuple(map(int, s.strip().replace("rc", ".").split(".")))
-from __future__ import print_function;
+    def __init__(self):
-import sys;
+        import sys
-try:
+        self.vpy = tuple(sys.version_info)
-	from packaging import version;
+    def __eq__(self, s):
-except ImportError:
+        return self.vpy == self.vtup(s)
-	from distlib import version;
+    def __ne__(self, s):
-ver = sys.version.split ()[[0]];
+        return self.vpy != self.vtup(s)
-(tst_cmp, tst_ver) = "$1".split ();
+    def __lt__(self, s):
-tst_ver = tst_ver.strip ("'");
+        return self.vpy < self.vtup(s)
-eval ("print (version.LegacyVersion (ver)"+ tst_cmp +"version.LegacyVersion (tst_ver))")
+    def __gt__(self, s):
-EOD`
+        return self.vpy > self.vtup(s)
+    def __le__(self, s):
+        return self.vpy <= self.vtup(s)
+    def __ge__(self, s):
+        return self.vpy >= self.vtup(s)
+EOF
+		ac_supports_python_ver=`$PYTHON -c "import ax_python_devel_vpy; \
+                        ver = ax_python_devel_vpy.VPy(); \
+			print (ver $1)"`
+                rm -rf ax_python_devel_vpy*.py* __pycache__/ax_python_devel_vpy*.py*
 		if test "$ac_supports_python_ver" = "True"; then
-		   AC_MSG_RESULT([yes])
+			AC_MSG_RESULT([yes])
 		else
 			AC_MSG_RESULT([no])
-			m4_ifvaln([$2],[$2],[
+			AC_MSG_ERROR([this package requires Python $1.
-				AC_MSG_ERROR([this package requires Python $1.
 If you have it installed, but it isn't the default Python
 interpreter in your system path, please pass the PYTHON_VERSION
 variable to configure. See ``configure --help'' for reference.
 ])
-				PYTHON_VERSION=""
+			PYTHON_VERSION=""
-			])
 		fi
 	fi
 

config/deb.am

@@ -66,7 +66,7 @@ deb-utils: deb-local rpm-utils-initramfs
 ## to do this, so we install a shim onto the path which calls the real
 ## dh_shlibdeps with the required arguments.
 	path_prepend=`mktemp -d /tmp/intercept.XXXXXX`; \
-	echo "#$(SHELL)" > $${path_prepend}/dh_shlibdeps; \
+	echo "#!$(SHELL)" > $${path_prepend}/dh_shlibdeps; \
 	echo "`which dh_shlibdeps` -- \
 	 -xlibuutil3linux -xlibnvpair3linux -xlibzfs5linux -xlibzpool5linux" \
 	 >> $${path_prepend}/dh_shlibdeps; \
@@ -74,7 +74,7 @@ deb-utils: deb-local rpm-utils-initramfs
 ## Debianized packages from the auto-generated dependencies of the new debs,
 ## which should NOT be mixed with the alien-generated debs created here
 	chmod +x $${path_prepend}/dh_shlibdeps; \
-	env PATH=$${path_prepend}:$${PATH} \
+	env "PATH=$${path_prepend}:$${PATH}" \
 	fakeroot $(ALIEN) --bump=0 --scripts --to-deb --target=$$debarch \
 	    $$pkg1 $$pkg2 $$pkg3 $$pkg4 $$pkg5 $$pkg6 $$pkg7 \
 	    $$pkg8 $$pkg9 $$pkg10 || exit 1; \

config/kernel-acl.m4

@@ -165,6 +165,9 @@ dnl #
 dnl # 5.15 API change,
 dnl # Added the bool rcu argument to get_acl for rcu path walk.
 dnl #
+dnl # 6.2 API change,
+dnl # get_acl() was renamed to get_inode_acl()
+dnl #
 AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_OPERATIONS_GET_ACL], [
 	ZFS_LINUX_TEST_SRC([inode_operations_get_acl], [
 		#include <linux/fs.h>
@@ -189,6 +192,18 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_OPERATIONS_GET_ACL], [
 			.get_acl = get_acl_fn,
 		};
 	],[])
+
+	ZFS_LINUX_TEST_SRC([inode_operations_get_inode_acl], [
+		#include <linux/fs.h>
+
+		struct posix_acl *get_inode_acl_fn(struct inode *inode, int type,
+		    bool rcu) { return NULL; }
+
+		static const struct inode_operations
+		    iops __attribute__ ((unused)) = {
+			.get_inode_acl = get_inode_acl_fn,
+		};
+	],[])
 ])
 
 AC_DEFUN([ZFS_AC_KERNEL_INODE_OPERATIONS_GET_ACL], [
@@ -201,7 +216,12 @@ AC_DEFUN([ZFS_AC_KERNEL_INODE_OPERATIONS_GET_ACL], [
 			AC_MSG_RESULT(yes)
 			AC_DEFINE(HAVE_GET_ACL_RCU, 1, [iops->get_acl() takes rcu])
 		],[
-			ZFS_LINUX_TEST_ERROR([iops->get_acl()])
+			ZFS_LINUX_TEST_RESULT([inode_operations_get_inode_acl], [
+				AC_MSG_RESULT(yes)
+				AC_DEFINE(HAVE_GET_INODE_ACL, 1, [has iops->get_inode_acl()])
+			],[
+				ZFS_LINUX_TEST_ERROR([iops->get_acl() or iops->get_inode_acl()])
+			])
 		])
 	])
 ])
@@ -213,7 +233,22 @@ dnl #
 dnl # 5.12 API change,
 dnl # set_acl() added a user_namespace* parameter first
 dnl #
+dnl # 6.2 API change,
+dnl # set_acl() second paramter changed to a struct dentry *
+dnl #
 AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_OPERATIONS_SET_ACL], [
+	ZFS_LINUX_TEST_SRC([inode_operations_set_acl_userns_dentry], [
+		#include <linux/fs.h>
+
+		int set_acl_fn(struct user_namespace *userns,
+		    struct dentry *dent, struct posix_acl *acl,
+		    int type) { return 0; }
+
+		static const struct inode_operations
+		    iops __attribute__ ((unused)) = {
+			.set_acl = set_acl_fn,
+		};
+	],[])
 	ZFS_LINUX_TEST_SRC([inode_operations_set_acl_userns], [
 		#include <linux/fs.h>
 
@@ -246,11 +281,18 @@ AC_DEFUN([ZFS_AC_KERNEL_INODE_OPERATIONS_SET_ACL], [
 		AC_DEFINE(HAVE_SET_ACL, 1, [iops->set_acl() exists])
 		AC_DEFINE(HAVE_SET_ACL_USERNS, 1, [iops->set_acl() takes 4 args])
 	],[
-		ZFS_LINUX_TEST_RESULT([inode_operations_set_acl], [
+		ZFS_LINUX_TEST_RESULT([inode_operations_set_acl_userns_dentry], [
 			AC_MSG_RESULT(yes)
-			AC_DEFINE(HAVE_SET_ACL, 1, [iops->set_acl() exists, takes 3 args])
+			AC_DEFINE(HAVE_SET_ACL, 1, [iops->set_acl() exists])
+			AC_DEFINE(HAVE_SET_ACL_USERNS_DENTRY_ARG2, 1,
+			    [iops->set_acl() takes 4 args, arg2 is struct dentry *])
 		],[
-			AC_MSG_RESULT(no)
+			ZFS_LINUX_TEST_RESULT([inode_operations_set_acl], [
+				AC_MSG_RESULT(yes)
+				AC_DEFINE(HAVE_SET_ACL, 1, [iops->set_acl() exists, takes 3 args])
+			],[
+				ZFS_LINUX_REQUIRE_API([i_op->set_acl()], [3.14])
+			])
 		])
 	])
 ])

config/kernel-blkdev.m4

@@ -103,6 +103,57 @@ AC_DEFUN([ZFS_AC_KERNEL_BLKDEV_CHECK_DISK_CHANGE], [
 	])
 ])
 
+dnl #
+dnl # bdev_kobj() is introduced from 5.12
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_SRC_BLKDEV_BDEV_KOBJ], [
+	ZFS_LINUX_TEST_SRC([bdev_kobj], [
+		#include <linux/fs.h>
+		#include <linux/blkdev.h>
+		#include <linux/kobject.h>
+	], [
+		struct block_device *bdev = NULL;
+		struct kobject *disk_kobj;
+		disk_kobj = bdev_kobj(bdev);
+	])
+])
+
+AC_DEFUN([ZFS_AC_KERNEL_BLKDEV_BDEV_KOBJ], [
+	AC_MSG_CHECKING([whether bdev_kobj() exists])
+	ZFS_LINUX_TEST_RESULT([bdev_kobj], [
+		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_BDEV_KOBJ, 1,
+		    [bdev_kobj() exists])
+	], [
+		AC_MSG_RESULT(no)
+	])
+])
+
+dnl #
+dnl # part_to_dev() was removed in 5.12
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_SRC_BLKDEV_PART_TO_DEV], [
+	ZFS_LINUX_TEST_SRC([part_to_dev], [
+		#include <linux/fs.h>
+		#include <linux/blkdev.h>
+	], [
+		struct hd_struct *p = NULL;
+		struct device *pdev;
+		pdev = part_to_dev(p);
+	])
+])
+
+AC_DEFUN([ZFS_AC_KERNEL_BLKDEV_PART_TO_DEV], [
+	AC_MSG_CHECKING([whether part_to_dev() exists])
+	ZFS_LINUX_TEST_RESULT([part_to_dev], [
+		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_PART_TO_DEV, 1,
+		    [part_to_dev() exists])
+	], [
+		AC_MSG_RESULT(no)
+	])
+])
+
 dnl #
 dnl # 5.10 API, check_disk_change() is removed, in favor of
 dnl # bdev_check_media_change(), which doesn't force revalidation
@@ -405,6 +456,8 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_BLKDEV], [
 	ZFS_AC_KERNEL_SRC_BLKDEV_BDEV_WHOLE
 	ZFS_AC_KERNEL_SRC_BLKDEV_BDEVNAME
 	ZFS_AC_KERNEL_SRC_BLKDEV_ISSUE_SECURE_ERASE
+	ZFS_AC_KERNEL_SRC_BLKDEV_BDEV_KOBJ
+	ZFS_AC_KERNEL_SRC_BLKDEV_PART_TO_DEV
 ])
 
 AC_DEFUN([ZFS_AC_KERNEL_BLKDEV], [
@@ -421,4 +474,6 @@ AC_DEFUN([ZFS_AC_KERNEL_BLKDEV], [
 	ZFS_AC_KERNEL_BLKDEV_BDEVNAME
 	ZFS_AC_KERNEL_BLKDEV_GET_ERESTARTSYS
 	ZFS_AC_KERNEL_BLKDEV_ISSUE_SECURE_ERASE
+	ZFS_AC_KERNEL_BLKDEV_BDEV_KOBJ
+	ZFS_AC_KERNEL_BLKDEV_PART_TO_DEV
 ])

config/kernel-generic_io_acct.m4

@@ -2,7 +2,20 @@ dnl #
 dnl # Check for generic io accounting interface.
 dnl #
 AC_DEFUN([ZFS_AC_KERNEL_SRC_GENERIC_IO_ACCT], [
-	ZFS_LINUX_TEST_SRC([bdev_io_acct], [
+	ZFS_LINUX_TEST_SRC([bdev_io_acct_63], [
+		#include <linux/blkdev.h>
+	], [
+		struct block_device *bdev = NULL;
+		struct bio *bio = NULL;
+		unsigned long passed_time = 0;
+		unsigned long start_time;
+
+		start_time = bdev_start_io_acct(bdev, bio_op(bio),
+		    passed_time);
+		bdev_end_io_acct(bdev, bio_op(bio), bio_sectors(bio), start_time);
+	])
+
+	ZFS_LINUX_TEST_SRC([bdev_io_acct_old], [
 		#include <linux/blkdev.h>
 	], [
 		struct block_device *bdev = NULL;
@@ -63,74 +76,85 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_GENERIC_IO_ACCT], [
 
 AC_DEFUN([ZFS_AC_KERNEL_GENERIC_IO_ACCT], [
 	dnl #
-	dnl # 5.19 API,
+	dnl # Linux 6.3, and then backports thereof, changed
+	dnl # the signatures on bdev_start_io_acct/bdev_end_io_acct
 	dnl #
-	dnl # disk_start_io_acct() and disk_end_io_acct() have been replaced by
+	AC_MSG_CHECKING([whether 6.3+ bdev_*_io_acct() are available])
-	dnl # bdev_start_io_acct() and bdev_end_io_acct().
+	ZFS_LINUX_TEST_RESULT([bdev_io_acct_63], [
-	dnl #
-	AC_MSG_CHECKING([whether generic bdev_*_io_acct() are available])
-	ZFS_LINUX_TEST_RESULT([bdev_io_acct], [
 		AC_MSG_RESULT(yes)
-		AC_DEFINE(HAVE_BDEV_IO_ACCT, 1, [bdev_*_io_acct() available])
+		AC_DEFINE(HAVE_BDEV_IO_ACCT_63, 1, [bdev_*_io_acct() available])
 	], [
 		AC_MSG_RESULT(no)
 
 		dnl #
-		dnl # 5.12 API,
+		dnl # 5.19 API,
 		dnl #
-		dnl # bio_start_io_acct() and bio_end_io_acct() became GPL-exported
+		dnl # disk_start_io_acct() and disk_end_io_acct() have been replaced by
-		dnl # so use disk_start_io_acct() and disk_end_io_acct() instead
+		dnl # bdev_start_io_acct() and bdev_end_io_acct().
 		dnl #
-		AC_MSG_CHECKING([whether generic disk_*_io_acct() are available])
+		AC_MSG_CHECKING([whether pre-6.3 bdev_*_io_acct() are available])
-		ZFS_LINUX_TEST_RESULT([disk_io_acct], [
+		ZFS_LINUX_TEST_RESULT([bdev_io_acct_old], [
 			AC_MSG_RESULT(yes)
-			AC_DEFINE(HAVE_DISK_IO_ACCT, 1, [disk_*_io_acct() available])
+			AC_DEFINE(HAVE_BDEV_IO_ACCT_OLD, 1, [bdev_*_io_acct() available])
 		], [
 			AC_MSG_RESULT(no)
-
 			dnl #
-			dnl # 5.7 API,
+			dnl # 5.12 API,
 			dnl #
-			dnl # Added bio_start_io_acct() and bio_end_io_acct() helpers.
+			dnl # bio_start_io_acct() and bio_end_io_acct() became GPL-exported
+			dnl # so use disk_start_io_acct() and disk_end_io_acct() instead
 			dnl #
-			AC_MSG_CHECKING([whether generic bio_*_io_acct() are available])
+			AC_MSG_CHECKING([whether generic disk_*_io_acct() are available])
-			ZFS_LINUX_TEST_RESULT([bio_io_acct], [
+			ZFS_LINUX_TEST_RESULT([disk_io_acct], [
 				AC_MSG_RESULT(yes)
-				AC_DEFINE(HAVE_BIO_IO_ACCT, 1, [bio_*_io_acct() available])
+				AC_DEFINE(HAVE_DISK_IO_ACCT, 1, [disk_*_io_acct() available])
 			], [
 				AC_MSG_RESULT(no)
 
 				dnl #
-				dnl # 4.14 API,
+				dnl # 5.7 API,
 				dnl #
-				dnl # generic_start_io_acct/generic_end_io_acct now require
+				dnl # Added bio_start_io_acct() and bio_end_io_acct() helpers.
-				dnl # request_queue to be provided. No functional changes,
-				dnl # but preparation for inflight accounting.
 				dnl #
-				AC_MSG_CHECKING([whether generic_*_io_acct wants 4 args])
+				AC_MSG_CHECKING([whether generic bio_*_io_acct() are available])
-				ZFS_LINUX_TEST_RESULT_SYMBOL([generic_acct_4args],
+				ZFS_LINUX_TEST_RESULT([bio_io_acct], [
-				    [generic_start_io_acct], [block/bio.c], [
 					AC_MSG_RESULT(yes)
-					AC_DEFINE(HAVE_GENERIC_IO_ACCT_4ARG, 1,
+					AC_DEFINE(HAVE_BIO_IO_ACCT, 1, [bio_*_io_acct() available])
-					    [generic_*_io_acct() 4 arg available])
 				], [
 					AC_MSG_RESULT(no)
 
 					dnl #
-					dnl # 3.19 API addition
+					dnl # 4.14 API,
 					dnl #
-					dnl # torvalds/linux@394ffa50 allows us to increment
+					dnl # generic_start_io_acct/generic_end_io_acct now require
-					dnl # iostat counters without generic_make_request().
+					dnl # request_queue to be provided. No functional changes,
+					dnl # but preparation for inflight accounting.
 					dnl #
-					AC_MSG_CHECKING(
+					AC_MSG_CHECKING([whether generic_*_io_acct wants 4 args])
-					    [whether generic_*_io_acct wants 3 args])
+					ZFS_LINUX_TEST_RESULT_SYMBOL([generic_acct_4args],
-					ZFS_LINUX_TEST_RESULT_SYMBOL([generic_acct_3args],
 					    [generic_start_io_acct], [block/bio.c], [
 						AC_MSG_RESULT(yes)
-						AC_DEFINE(HAVE_GENERIC_IO_ACCT_3ARG, 1,
+						AC_DEFINE(HAVE_GENERIC_IO_ACCT_4ARG, 1,
-						    [generic_*_io_acct() 3 arg available])
+						    [generic_*_io_acct() 4 arg available])
 					], [
 						AC_MSG_RESULT(no)
+
+						dnl #
+						dnl # 3.19 API addition
+						dnl #
+						dnl # torvalds/linux@394ffa50 allows us to increment
+						dnl # iostat counters without generic_make_request().
+						dnl #
+						AC_MSG_CHECKING(
+						    [whether generic_*_io_acct wants 3 args])
+						ZFS_LINUX_TEST_RESULT_SYMBOL([generic_acct_3args],
+						    [generic_start_io_acct], [block/bio.c], [
+							AC_MSG_RESULT(yes)
+							AC_DEFINE(HAVE_GENERIC_IO_ACCT_3ARG, 1,
+							    [generic_*_io_acct() 3 arg available])
+						], [
+							AC_MSG_RESULT(no)
+						])
 					])
 				])
 			])

config/kernel-global_page_state.m4

@@ -55,7 +55,7 @@ dnl #
 AC_DEFUN([ZFS_AC_KERNEL_ENUM_MEMBER], [
 	AC_MSG_CHECKING([whether enum $2 contains $1])
 	AS_IF([AC_TRY_COMMAND(
-	    "${srcdir}/scripts/enum-extract.pl" "$2" "$3" | egrep -qx $1)],[
+	    "${srcdir}/scripts/enum-extract.pl" "$2" "$3" | grep -Eqx $1)],[
 		AC_MSG_RESULT([yes])
 		AC_DEFINE(m4_join([_], [ZFS_ENUM], m4_toupper($2), $1), 1,
 		    [enum $2 contains $1])

config/kernel-tmpfile.m4

@@ -3,11 +3,25 @@ dnl # 3.11 API change
 dnl # Add support for i_op->tmpfile
 dnl #
 AC_DEFUN([ZFS_AC_KERNEL_SRC_TMPFILE], [
+	dnl #
+	dnl # 6.1 API change
+	dnl # use struct file instead of struct dentry
+	dnl #
+	ZFS_LINUX_TEST_SRC([inode_operations_tmpfile], [
+		#include <linux/fs.h>
+		int tmpfile(struct user_namespace *userns,
+		    struct inode *inode, struct file *file,
+		    umode_t mode) { return 0; }
+		static struct inode_operations
+		    iops __attribute__ ((unused)) = {
+			.tmpfile = tmpfile,
+		};
+	],[])
 	dnl #
 	dnl # 5.11 API change
 	dnl # add support for userns parameter to tmpfile
 	dnl #
-	ZFS_LINUX_TEST_SRC([inode_operations_tmpfile_userns], [
+	ZFS_LINUX_TEST_SRC([inode_operations_tmpfile_dentry_userns], [
 		#include <linux/fs.h>
 		int tmpfile(struct user_namespace *userns,
 		    struct inode *inode, struct dentry *dentry,
@@ -17,7 +31,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_TMPFILE], [
 			.tmpfile = tmpfile,
 		};
 	],[])
-	ZFS_LINUX_TEST_SRC([inode_operations_tmpfile], [
+	ZFS_LINUX_TEST_SRC([inode_operations_tmpfile_dentry], [
 			#include <linux/fs.h>
 			int tmpfile(struct inode *inode, struct dentry *dentry,
 			    umode_t mode) { return 0; }
@@ -30,16 +44,24 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_TMPFILE], [
 
 AC_DEFUN([ZFS_AC_KERNEL_TMPFILE], [
 	AC_MSG_CHECKING([whether i_op->tmpfile() exists])
-	ZFS_LINUX_TEST_RESULT([inode_operations_tmpfile_userns], [
+	ZFS_LINUX_TEST_RESULT([inode_operations_tmpfile], [
 		AC_MSG_RESULT(yes)
 		AC_DEFINE(HAVE_TMPFILE, 1, [i_op->tmpfile() exists])
 		AC_DEFINE(HAVE_TMPFILE_USERNS, 1, [i_op->tmpfile() has userns])
 	],[
-		ZFS_LINUX_TEST_RESULT([inode_operations_tmpfile], [
+		ZFS_LINUX_TEST_RESULT([inode_operations_tmpfile_dentry_userns], [
 			AC_MSG_RESULT(yes)
 			AC_DEFINE(HAVE_TMPFILE, 1, [i_op->tmpfile() exists])
+			AC_DEFINE(HAVE_TMPFILE_USERNS, 1, [i_op->tmpfile() has userns])
+			AC_DEFINE(HAVE_TMPFILE_DENTRY, 1, [i_op->tmpfile() uses old dentry signature])
 		],[
-			AC_MSG_RESULT(no)
+			ZFS_LINUX_TEST_RESULT([inode_operations_tmpfile_dentry], [
+				AC_MSG_RESULT(yes)
+				AC_DEFINE(HAVE_TMPFILE, 1, [i_op->tmpfile() exists])
+				AC_DEFINE(HAVE_TMPFILE_DENTRY, 1, [i_op->tmpfile() uses old dentry signature])
+			],[
+				ZFS_LINUX_REQUIRE_API([i_op->tmpfile()], [3.11])
+			])
 		])
 	])
 ])

config/kernel.m4

@@ -404,11 +404,11 @@ AC_DEFUN([ZFS_AC_KERNEL], [
 	utsrelease1=$kernelbuild/include/linux/version.h
 	utsrelease2=$kernelbuild/include/linux/utsrelease.h
 	utsrelease3=$kernelbuild/include/generated/utsrelease.h
-	AS_IF([test -r $utsrelease1 && fgrep -q UTS_RELEASE $utsrelease1], [
+	AS_IF([test -r $utsrelease1 && grep -qF UTS_RELEASE $utsrelease1], [
 		utsrelease=$utsrelease1
-	], [test -r $utsrelease2 && fgrep -q UTS_RELEASE $utsrelease2], [
+	], [test -r $utsrelease2 && grep -qF UTS_RELEASE $utsrelease2], [
 		utsrelease=$utsrelease2
-	], [test -r $utsrelease3 && fgrep -q UTS_RELEASE $utsrelease3], [
+	], [test -r $utsrelease3 && grep -qF UTS_RELEASE $utsrelease3], [
 		utsrelease=$utsrelease3
 	])
 
@@ -946,3 +946,35 @@ AC_DEFUN([ZFS_LINUX_TRY_COMPILE_HEADER], [
 		    [test -f build/conftest/conftest.ko], [$3], [$4], [$5])
 	])
 ])
+
+dnl #
+dnl # AS_VERSION_COMPARE_LE
+dnl # like AS_VERSION_COMPARE_LE, but runs $3 if (and only if) $1 <= $2
+dnl # AS_VERSION_COMPARE_LE (version-1, version-2, [action-if-less-or-equal], [action-if-greater])
+dnl #
+AC_DEFUN([AS_VERSION_COMPARE_LE], [
+	AS_VERSION_COMPARE([$1], [$2], [$3], [$3], [$4])
+])
+
+dnl #
+dnl # ZFS_LINUX_REQUIRE_API
+dnl # like ZFS_LINUX_TEST_ERROR, except only fails if the kernel is
+dnl # at least some specified version.
+dnl #
+AC_DEFUN([ZFS_LINUX_REQUIRE_API], [
+	AS_VERSION_COMPARE_LE([$2], [$kernsrcver], [
+		AC_MSG_ERROR([
+		*** None of the expected "$1" interfaces were detected. This
+		*** interface is expected for kernels version "$2" and above.
+		*** This may be because your kernel version is newer than what is
+		*** supported, or you are using a patched custom kernel with
+		*** incompatible modifications.  Newer kernels may have incompatible
+		*** APIs.
+		***
+		*** ZFS Version: $ZFS_META_ALIAS
+		*** Compatible Kernels: $ZFS_META_KVER_MIN - $ZFS_META_KVER_MAX
+		])
+	], [
+		AC_MSG_RESULT(no)
+	])
+])

config/zfs-build.m4

@@ -173,7 +173,7 @@ AC_DEFUN([ZFS_AC_DEBUG_KMEM_TRACKING], [
 ])
 
 AC_DEFUN([ZFS_AC_DEBUG_INVARIANTS_DETECT_FREEBSD], [
-	AS_IF([sysctl -n kern.conftxt | fgrep -qx $'options\tINVARIANTS'],
+	AS_IF([sysctl -n kern.conftxt | grep -Fqx $'options\tINVARIANTS'],
 		[enable_invariants="yes"],
 		[enable_invariants="no"])
 ])
@@ -518,6 +518,8 @@ AC_DEFUN([ZFS_AC_DEFAULT_PACKAGE], [
 			VENDOR=alpine ;
 		elif test -f /bin/freebsd-version ; then
 			VENDOR=freebsd ;
+		elif test -f /etc/openEuler-release ; then
+			VENDOR=openeuler ;
 		else
 			VENDOR= ;
 		fi],
@@ -542,6 +544,7 @@ AC_DEFUN([ZFS_AC_DEFAULT_PACKAGE], [
 		ubuntu)     DEFAULT_PACKAGE=deb  ;;
 		debian)     DEFAULT_PACKAGE=deb  ;;
 		freebsd)    DEFAULT_PACKAGE=pkg  ;;
+		openeuler)  DEFAULT_PACKAGE=rpm  ;;
 		*)          DEFAULT_PACKAGE=rpm  ;;
 	esac
 	AC_MSG_RESULT([$DEFAULT_PACKAGE])
@@ -555,35 +558,14 @@ AC_DEFUN([ZFS_AC_DEFAULT_PACKAGE], [
 	AC_MSG_RESULT([$initdir])
 	AC_SUBST(initdir)
 
-	AC_MSG_CHECKING([default init script type and shell])
+	AC_MSG_CHECKING([default shell])
 	case "$VENDOR" in
-		toss)       DEFAULT_INIT_SCRIPT=redhat ;;
+		gentoo)     DEFAULT_INIT_SHELL="/sbin/openrc-run";;
-		redhat)     DEFAULT_INIT_SCRIPT=redhat ;;
+		alpine)     DEFAULT_INIT_SHELL="/sbin/openrc-run";;
-		fedora)     DEFAULT_INIT_SCRIPT=fedora ;;
+		*)          DEFAULT_INIT_SHELL="/bin/sh"         ;;
-		gentoo)     DEFAULT_INIT_SCRIPT=openrc ;;
-		alpine)     DEFAULT_INIT_SCRIPT=openrc ;;
-		arch)       DEFAULT_INIT_SCRIPT=lsb    ;;
-		sles)       DEFAULT_INIT_SCRIPT=lsb    ;;
-		slackware)  DEFAULT_INIT_SCRIPT=lsb    ;;
-		lunar)      DEFAULT_INIT_SCRIPT=lunar  ;;
-		ubuntu)     DEFAULT_INIT_SCRIPT=lsb    ;;
-		debian)     DEFAULT_INIT_SCRIPT=lsb    ;;
-		freebsd)    DEFAULT_INIT_SCRIPT=freebsd;;
-		*)          DEFAULT_INIT_SCRIPT=lsb    ;;
 	esac
 
-	# On gentoo, it's possible that OpenRC isn't installed.  Check if
+	AC_MSG_RESULT([$DEFAULT_INIT_SHELL])
-	# /sbin/openrc-run exists, and if not, fall back to generic defaults.
-
-	DEFAULT_INIT_SHELL="/bin/sh"
-	AS_IF([test "$DEFAULT_INIT_SCRIPT" = "openrc"], [
-		AS_IF([test -x "/sbin/openrc-run"],
-			[DEFAULT_INIT_SHELL="/sbin/openrc-run"],
-			[DEFAULT_INIT_SCRIPT=lsb])
-	])
-
-	AC_MSG_RESULT([$DEFAULT_INIT_SCRIPT:$DEFAULT_INIT_SHELL])
-	AC_SUBST(DEFAULT_INIT_SCRIPT)
 	AC_SUBST(DEFAULT_INIT_SHELL)
 
 	AC_MSG_CHECKING([default nfs server init script])
@@ -602,6 +584,7 @@ AC_DEFUN([ZFS_AC_DEFAULT_PACKAGE], [
 		redhat)     initconfdir=/etc/sysconfig ;;
 		fedora)     initconfdir=/etc/sysconfig ;;
 		sles)       initconfdir=/etc/sysconfig ;;
+		openeuler)  initconfdir=/etc/sysconfig ;;
 		ubuntu)     initconfdir=/etc/default   ;;
 		debian)     initconfdir=/etc/default   ;;
 		freebsd)    initconfdir=$sysconfdir/rc.conf.d;;

contrib/dracut/02zfsexpandknowledge/module-setup.sh.in

@@ -57,6 +57,12 @@ array_contains () {
 }
 
 check() {
+    # https://github.com/dracutdevs/dracut/pull/1711 provides a zfs_devs
+    # function to detect the physical devices backing zfs pools. If this
+    # function exists in the version of dracut this module is being called
+    # from, then it does not need to run.
+    type zfs_devs >/dev/null 2>&1 && return 1
+
     local mp
     local dev
     local blockdevs

contrib/dracut/90zfs/Makefile.am

@@ -15,6 +15,7 @@ pkgdracut_SCRIPTS = \
 
 pkgdracut_DATA = \
 	zfs-env-bootfs.service \
+	zfs-nonroot-necessities.service \
 	zfs-snapshot-bootfs.service \
 	zfs-rollback-bootfs.service
 

contrib/dracut/90zfs/module-setup.sh.in

@@ -84,11 +84,24 @@ install() {
 		inst_simple "${moddir}/zfs-env-bootfs.service" "${systemdsystemunitdir}/zfs-env-bootfs.service"
 		systemctl -q --root "${initdir}" add-wants zfs-import.target zfs-env-bootfs.service
 
+		inst_simple "${moddir}/zfs-nonroot-necessities.service" "${systemdsystemunitdir}/zfs-nonroot-necessities.service"
+		systemctl -q --root "${initdir}" add-requires initrd-root-fs.target zfs-nonroot-necessities.service
+
 		for _service in \
 			"zfs-import-scan.service" \
 			"zfs-import-cache.service"; do
 			inst_simple "${systemdsystemunitdir}/${_service}"
 			systemctl -q --root "${initdir}" add-wants zfs-import.target "${_service}"
+
+			# Add user-provided unit overrides
+			# - /etc/systemd/system/zfs-import-{scan,cache}.service
+			# - /etc/systemd/system/zfs-import-{scan,cache}.service.d/overrides.conf
+			# -H ensures they are marked host-only
+			# -o ensures there is no error upon absence of these files
+			inst_multiple -o -H \
+				"${systemdsystemconfdir}/${_service}" \
+				"${systemdsystemconfdir}/${_service}.d/"*.conf
+
 		done
 
 		for _service in \

contrib/dracut/90zfs/mount-zfs.sh.in

@@ -82,7 +82,7 @@ ZFS_DATASET="${ZFS_DATASET:-${root}}"
 ZFS_POOL="${ZFS_DATASET%%/*}"
 
 
-if ! zpool get -Ho name "${ZFS_POOL}" > /dev/null 2>&1; then
+if ! zpool get -Ho value name "${ZFS_POOL}" > /dev/null 2>&1; then
     info "ZFS: Importing pool ${ZFS_POOL}..."
     # shellcheck disable=SC2086
     if ! zpool import -N ${ZPOOL_IMPORT_OPTS} "${ZFS_POOL}"; then

contrib/dracut/90zfs/zfs-env-bootfs.service.in

@@ -1,6 +1,5 @@
 [Unit]
-Description=Set BOOTFS environment for dracut
+Description=Set BOOTFS and BOOTFSFLAGS environment variables for dracut
-Documentation=man:zpool(8)
 DefaultDependencies=no
 After=zfs-import-cache.service
 After=zfs-import-scan.service
@@ -8,7 +7,17 @@ Before=zfs-import.target
 
 [Service]
 Type=oneshot
-ExecStart=/bin/sh -c "exec systemctl set-environment BOOTFS=$(@sbindir@/zpool list -H -o bootfs | grep -m1 -vFx -)"
+ExecStart=/bin/sh -c '                                                                         \
+    . /lib/dracut-zfs-lib.sh;                                                                  \
+    decode_root_args || exit 0;                                                                \
+    [ "$root" = "zfs:AUTO" ] && root="$(@sbindir@/zpool list -H -o bootfs | grep -m1 -vFx -)"; \
+    rootflags="$(getarg rootflags=)";                                                          \
+    case ",$rootflags," in                                                                     \
+        *,zfsutil,*) ;;                                                                        \
+        ,,) rootflags=zfsutil ;;                                                               \
+        *)  rootflags="zfsutil,$rootflags" ;;                                                  \
+    esac;                                                                                      \
+    exec systemctl set-environment BOOTFS="$root" BOOTFSFLAGS="$rootflags"'
 
 [Install]
 WantedBy=zfs-import.target

contrib/dracut/90zfs/zfs-generator.sh.in

@@ -14,81 +14,24 @@ GENERATOR_DIR="$1"
 . /lib/dracut-zfs-lib.sh
 decode_root_args || exit 0
 
-[ -z "${rootflags}" ] && rootflags=$(getarg rootflags=)
-case ",${rootflags}," in
-	*,zfsutil,*) ;;
-	,,)	rootflags=zfsutil ;;
-	*)	rootflags="zfsutil,${rootflags}" ;;
-esac
-
 [ -n "$debug" ] && echo "zfs-generator: writing extension for sysroot.mount to $GENERATOR_DIR/sysroot.mount.d/zfs-enhancement.conf" >> /dev/kmsg
 
 
-mkdir -p "$GENERATOR_DIR"/sysroot.mount.d "$GENERATOR_DIR"/initrd-root-fs.target.requires "$GENERATOR_DIR"/dracut-pre-mount.service.d
+mkdir -p "$GENERATOR_DIR"/sysroot.mount.d "$GENERATOR_DIR"/dracut-pre-mount.service.d
+
 {
     echo "[Unit]"
     echo "Before=initrd-root-fs.target"
     echo "After=zfs-import.target"
     echo
     echo "[Mount]"
-    if [ "${root}" = "zfs:AUTO" ]; then
+    echo "PassEnvironment=BOOTFS BOOTFSFLAGS"
-      echo "PassEnvironment=BOOTFS"
+    echo 'What=${BOOTFS}'
-      echo 'What=${BOOTFS}'
-    else
-      echo "What=${root}"
-    fi
     echo "Type=zfs"
-    echo "Options=${rootflags}"
+    echo 'Options=${BOOTFSFLAGS}'
 } > "$GENERATOR_DIR"/sysroot.mount.d/zfs-enhancement.conf
 ln -fs ../sysroot.mount "$GENERATOR_DIR"/initrd-root-fs.target.requires/sysroot.mount
 
-
-if [ "${root}" = "zfs:AUTO" ]; then
-  {
-      echo "[Unit]"
-      echo "Before=initrd-root-fs.target"
-      echo "After=sysroot.mount"
-      echo "DefaultDependencies=no"
-      echo
-      echo "[Service]"
-      echo "Type=oneshot"
-      echo "PassEnvironment=BOOTFS"
-      echo "ExecStart=/bin/sh -c '" '                                        \
-        . /lib/dracut-zfs-lib.sh;                                            \
-        _zfs_nonroot_necessities_cb() {                                      \
-            zfs mount | grep -m1 -q "^$1 " && return 0;                      \
-            echo "Mounting $1 on /sysroot$2";                                \
-            mount -o zfsutil -t zfs "$1" "/sysroot$2";                       \
-        };                                                                   \
-        for_relevant_root_children "${BOOTFS}" _zfs_nonroot_necessities_cb;' \
-      "'"
-  } > "$GENERATOR_DIR"/zfs-nonroot-necessities.service
-  ln -fs ../zfs-nonroot-necessities.service "$GENERATOR_DIR"/initrd-root-fs.target.requires/zfs-nonroot-necessities.service
-else
-  # We can solve this statically at generation time, so do!
-  _zfs_generator_cb() {
-      dset="${1}"
-      mpnt="${2}"
-      unit="$(systemd-escape --suffix=mount -p "/sysroot${mpnt}")"
-
-      {
-          echo "[Unit]"
-          echo "Before=initrd-root-fs.target"
-          echo "After=sysroot.mount"
-          echo
-          echo "[Mount]"
-          echo "Where=/sysroot${mpnt}"
-          echo "What=${dset}"
-          echo "Type=zfs"
-          echo "Options=zfsutil"
-      } > "$GENERATOR_DIR/${unit}"
-      ln -fs ../"${unit}" "$GENERATOR_DIR"/initrd-root-fs.target.requires/"${unit}"
-  }
-
-  for_relevant_root_children "${root}" _zfs_generator_cb
-fi
-
-
 {
     echo "[Unit]"
     echo "After=zfs-import.target"

contrib/dracut/90zfs/zfs-lib.sh.in

@@ -38,7 +38,7 @@ mount_dataset() {
 
 # for_relevant_root_children DATASET EXEC
 #   Runs "EXEC dataset mountpoint" for all children of DATASET that are needed for system bringup
-#   Used by zfs-generator.sh and friends, too!
+#   Used by zfs-nonroot-necessities.service and friends, too!
 for_relevant_root_children() {
     dataset="${1}"
     exec="${2}"

contrib/dracut/90zfs/zfs-nonroot-necessities.service.in

@@ -0,0 +1,20 @@
+[Unit]
+Before=initrd-root-fs.target
+After=sysroot.mount
+DefaultDependencies=no
+ConditionEnvironment=BOOTFS
+
+[Service]
+Type=oneshot
+PassEnvironment=BOOTFS
+ExecStart=/bin/sh -c '                                                \
+    . /lib/dracut-zfs-lib.sh;                                         \
+    _zfs_nonroot_necessities_cb() {                                   \
+        @sbindir@/zfs mount | grep -m1 -q "^$1 " && return 0;         \
+        echo "Mounting $1 on /sysroot$2";                             \
+        mount -o zfsutil -t zfs "$1" "/sysroot$2";                    \
+    };                                                                \
+    for_relevant_root_children "${BOOTFS}" _zfs_nonroot_necessities_cb'
+
+[Install]
+RequiredBy=initrd-root-fs.target

contrib/dracut/90zfs/zfs-rollback-bootfs.service.in

@@ -5,8 +5,9 @@ After=zfs-import.target dracut-pre-mount.service zfs-snapshot-bootfs.service
 Before=dracut-mount.service
 DefaultDependencies=no
 ConditionKernelCommandLine=bootfs.rollback
+ConditionEnvironment=BOOTFS
 
 [Service]
 Type=oneshot
-ExecStart=/bin/sh -c '. /lib/dracut-zfs-lib.sh; decode_root_args || exit; [ "$root" = "zfs:AUTO" ] && root="$BOOTFS"; SNAPNAME="$(getarg bootfs.rollback)"; exec @sbindir@/zfs rollback -Rf "$root@${SNAPNAME:-%v}"'
+ExecStart=/bin/sh -c '. /lib/dracut-lib.sh; SNAPNAME="$(getarg bootfs.rollback)"; exec @sbindir@/zfs rollback -Rf "$BOOTFS@${SNAPNAME:-%v}"'
 RemainAfterExit=yes

contrib/dracut/90zfs/zfs-snapshot-bootfs.service.in

@@ -5,8 +5,9 @@ After=zfs-import.target dracut-pre-mount.service
 Before=dracut-mount.service
 DefaultDependencies=no
 ConditionKernelCommandLine=bootfs.snapshot
+ConditionEnvironment=BOOTFS
 
 [Service]
 Type=oneshot
-ExecStart=-/bin/sh -c '. /lib/dracut-zfs-lib.sh; decode_root_args || exit; [ "$root" = "zfs:AUTO" ] && root="$BOOTFS"; SNAPNAME="$(getarg bootfs.snapshot)"; exec @sbindir@/zfs snapshot "$root@${SNAPNAME:-%v}"'
+ExecStart=-/bin/sh -c '. /lib/dracut-lib.sh; SNAPNAME="$(getarg bootfs.snapshot)"; exec @sbindir@/zfs snapshot "$BOOTFS@${SNAPNAME:-%v}"'
 RemainAfterExit=yes

contrib/initramfs/scripts/Makefile.am

@@ -7,5 +7,6 @@ dist_scripts_SCRIPTS = \
 
 SUBDIRS = local-top
 
+SHELLCHECK_IGNORE = ,SC2295
 SHELLCHECKDIRS = $(SUBDIRS)
 SHELLCHECK_SHELL = sh

contrib/initramfs/scripts/zfs

@@ -192,7 +192,7 @@ import_pool()
 
 	# Verify that the pool isn't already imported
 	# Make as sure as we can to not require '-f' to import.
-	"${ZPOOL}" get name,guid -o value -H 2>/dev/null | grep -Fxq "$pool" && return 0
+	"${ZPOOL}" get -H -o value name,guid 2>/dev/null | grep -Fxq "$pool" && return 0
 
 	# For backwards compatibility, make sure that ZPOOL_IMPORT_PATH is set
 	# to something we can use later with the real import(s). We want to
@@ -331,25 +331,24 @@ mount_fs()
 		# Can't use the mountpoint property. Might be one of our
 		# clones. Check the 'org.zol:mountpoint' property set in
 		# clone_snap() if that's usable.
-		mountpoint=$(get_fs_value "$fs" org.zol:mountpoint)
+		mountpoint1=$(get_fs_value "$fs" org.zol:mountpoint)
-		if [ "$mountpoint" = "legacy" ] ||
+		if [ "$mountpoint1" = "legacy" ] ||
-		   [ "$mountpoint" = "none" ] ||
+		   [ "$mountpoint1" = "none" ] ||
-		   [ "$mountpoint" = "-" ]
+		   [ "$mountpoint1" = "-" ]
 		then
 			if [ "$fs" != "${ZFS_BOOTFS}" ]; then
 				# We don't have a proper mountpoint and this
 				# isn't the root fs.
 				return 0
-			else
-				# Last hail-mary: Hope 'rootmnt' is set!
-				mountpoint=""
 			fi
-		fi
+			# Don't use mount.zfs -o zfsutils for legacy mountpoint
-
+			if [ "$mountpoint" = "legacy" ]; then
-		# If it's not a legacy filesystem, it can only be a
+				ZFS_CMD="mount.zfs"
-		# native one...
+			fi
-		if [ "$mountpoint" = "legacy" ]; then
+			# Last hail-mary: Hope 'rootmnt' is set!
-			ZFS_CMD="mount.zfs"
+			mountpoint=""
+		else
+			mountpoint="$mountpoint1"
 		fi
 	fi
 
@@ -503,7 +502,7 @@ clone_snap()
 		echo "Error: $ZFS_ERROR"
 		echo ""
 		echo "Failed to clone snapshot."
-		echo "Make sure that the any problems are corrected and then make sure"
+		echo "Make sure that any problems are corrected and then make sure"
 		echo "that the dataset '$destfs' exists and is bootable."
 		shell
 	else
@@ -884,12 +883,12 @@ mountroot()
 	fi
 
 	# In case the pool was specified as guid, resolve guid to name
-	pool="$("${ZPOOL}" get name,guid -o name,value -H | \
+	pool="$("${ZPOOL}" get -H -o name,value name,guid | \
 	    awk -v pool="${ZFS_RPOOL}" '$2 == pool { print $1 }')"
 	if [ -n "$pool" ]; then
 		# If $ZFS_BOOTFS contains guid, replace the guid portion with $pool
 		ZFS_BOOTFS=$(echo "$ZFS_BOOTFS" | \
-			sed -e "s/$("${ZPOOL}" get guid -o value "$pool" -H)/$pool/g")
+			sed -e "s/$("${ZPOOL}" get -H -o value guid "$pool")/$pool/g")
 		ZFS_RPOOL="${pool}"
 	fi
 

contrib/pyzfs/setup.py.in

@@ -29,13 +29,12 @@ setup(
         "Development Status :: 4 - Beta",
         "Intended Audience :: Developers",
         "License :: OSI Approved :: Apache Software License",
-        "Programming Language :: Python :: 2",
-        "Programming Language :: Python :: 2.7",
         "Programming Language :: Python :: 3",
-        "Programming Language :: Python :: 3.4",
-        "Programming Language :: Python :: 3.5",
         "Programming Language :: Python :: 3.6",
         "Programming Language :: Python :: 3.7",
+        "Programming Language :: Python :: 3.8",
+        "Programming Language :: Python :: 3.9",
+        "Programming Language :: Python :: 3.10",
         "Topic :: System :: Filesystems",
         "Topic :: Software Development :: Libraries",
     ],
@@ -53,7 +52,7 @@ setup(
     setup_requires=[
         "cffi",
     ],
-    python_requires='>=2.7,!=3.0.*,!=3.1.*,!=3.2.*,!=3.3.*,<4',
+    python_requires='>=3.6,<4',
     zip_safe=False,
     test_suite="libzfs_core.test",
 )

etc/systemd/system/zfs-zed.service.in

@@ -5,7 +5,7 @@ ConditionPathIsDirectory=/sys/module/zfs
 
 [Service]
 ExecStart=@sbindir@/zed -F
-Restart=on-abort
+Restart=always
 
 [Install]
 Alias=zed.service

include/libzfs.h

@@ -307,6 +307,7 @@ extern int zpool_vdev_remove_cancel(zpool_handle_t *);
 extern int zpool_vdev_indirect_size(zpool_handle_t *, const char *, uint64_t *);
 extern int zpool_vdev_split(zpool_handle_t *, char *, nvlist_t **, nvlist_t *,
     splitflags_t);
+_LIBZFS_H int zpool_vdev_remove_wanted(zpool_handle_t *, const char *);
 
 extern int zpool_vdev_fault(zpool_handle_t *, uint64_t, vdev_aux_t);
 extern int zpool_vdev_degrade(zpool_handle_t *, uint64_t, vdev_aux_t);
@@ -692,6 +693,9 @@ typedef struct sendflags {
 	/* show progress (ie. -v) */
 	boolean_t progress;
 
+	/* show progress as process title (ie. -V) */
+	boolean_t progressastitle;
+
 	/* large blocks (>128K) are permitted */
 	boolean_t largeblock;
 

include/libzutil.h

@@ -150,14 +150,32 @@ int zfs_ioctl_fd(int fd, unsigned long request, struct zfs_cmd *zc);
 /*
  * List of colors to use
  */
+#define	ANSI_BLACK	"\033[0;30m"
 #define	ANSI_RED	"\033[0;31m"
+#define	ANSI_GREEN	"\033[0;32m"
 #define	ANSI_YELLOW	"\033[0;33m"
+#define	ANSI_BLUE	"\033[0;34m"
+#define	ANSI_BOLD_BLUE	"\033[1;34m" /* light blue */
+#define	ANSI_MAGENTA	"\033[0;35m"
+#define	ANSI_CYAN	"\033[0;36m"
+#define	ANSI_GRAY	"\033[0;37m"
+
 #define	ANSI_RESET	"\033[0m"
 #define	ANSI_BOLD	"\033[1m"
 
-void color_start(char *color);
+int use_color(void);
+void color_start(const char *color);
 void color_end(void);
-int printf_color(char *color, char *format, ...);
+int printf_color(const char *color, char *format, ...);
+
+#ifdef __linux__
+extern char **environ;
+_LIBZUTIL_H void zfs_setproctitle_init(int argc, char *argv[], char *envp[]);
+_LIBZUTIL_H void zfs_setproctitle(const char *fmt, ...);
+#else
+#define	zfs_setproctitle(fmt, ...)	setproctitle(fmt, ##__VA_ARGS__)
+#define	zfs_setproctitle_init(x, y, z)	((void)0)
+#endif
 
 /*
  * These functions are used by the ZFS libraries and cmd/zpool code, but are

include/os/linux/kernel/linux/blkdev_compat.h

@@ -256,6 +256,32 @@ bio_set_bi_error(struct bio *bio, int error)
 #define	BIO_END_IO(bio, error)		bio_endio(bio, error);
 #endif /* HAVE_1ARG_BIO_END_IO_T */
 
+/*
+ * 5.15 MACRO,
+ *   GD_DEAD
+ *
+ * 2.6.36 - 5.14 MACRO,
+ *   GENHD_FL_UP
+ *
+ * Check the disk status and return B_TRUE if alive
+ * otherwise B_FALSE
+ */
+static inline boolean_t
+zfs_check_disk_status(struct block_device *bdev)
+{
+#if defined(GENHD_FL_UP)
+	return (!!(bdev->bd_disk->flags & GENHD_FL_UP));
+#elif defined(GD_DEAD)
+	return (!test_bit(GD_DEAD, &bdev->bd_disk->state));
+#else
+/*
+ * This is encountered if neither GENHD_FL_UP nor GD_DEAD is available in
+ * the kernel - likely due to an MACRO change that needs to be chased down.
+ */
+#error "Unsupported kernel: no usable disk status check"
+#endif
+}
+
 /*
  * 4.1 API,
  * 3.10.0 CentOS 7.x API,
@@ -357,7 +383,11 @@ vdev_lookup_bdev(const char *path, dev_t *dev)
 static inline void
 bio_set_op_attrs(struct bio *bio, unsigned rw, unsigned flags)
 {
+#if defined(HAVE_BIO_BI_OPF)
+	bio->bi_opf = rw | flags;
+#else
 	bio->bi_rw |= rw | flags;
+#endif /* HAVE_BIO_BI_OPF */
 }
 #endif
 
@@ -385,7 +415,7 @@ static inline void
 bio_set_flush(struct bio *bio)
 {
 #if defined(HAVE_REQ_PREFLUSH)	/* >= 4.10 */
-	bio_set_op_attrs(bio, 0, REQ_PREFLUSH);
+	bio_set_op_attrs(bio, 0, REQ_PREFLUSH | REQ_OP_WRITE);
 #elif defined(WRITE_FLUSH_FUA)	/* >= 2.6.37 and <= 4.9 */
 	bio_set_op_attrs(bio, 0, WRITE_FLUSH_FUA);
 #else
@@ -551,7 +581,10 @@ blk_generic_start_io_acct(struct request_queue *q __attribute__((unused)),
     struct gendisk *disk __attribute__((unused)),
     int rw __attribute__((unused)), struct bio *bio)
 {
-#if defined(HAVE_BDEV_IO_ACCT)
+#if defined(HAVE_BDEV_IO_ACCT_63)
+	return (bdev_start_io_acct(bio->bi_bdev, bio_op(bio),
+	    jiffies));
+#elif defined(HAVE_BDEV_IO_ACCT_OLD)
 	return (bdev_start_io_acct(bio->bi_bdev, bio_sectors(bio),
 	    bio_op(bio), jiffies));
 #elif defined(HAVE_DISK_IO_ACCT)
@@ -577,7 +610,10 @@ blk_generic_end_io_acct(struct request_queue *q __attribute__((unused)),
     struct gendisk *disk __attribute__((unused)),
     int rw __attribute__((unused)), struct bio *bio, unsigned long start_time)
 {
-#if defined(HAVE_BDEV_IO_ACCT)
+#if defined(HAVE_BDEV_IO_ACCT_63)
+	bdev_end_io_acct(bio->bi_bdev, bio_op(bio), bio_sectors(bio),
+	    start_time);
+#elif defined(HAVE_BDEV_IO_ACCT_OLD)
 	bdev_end_io_acct(bio->bi_bdev, bio_op(bio), start_time);
 #elif defined(HAVE_DISK_IO_ACCT)
 	disk_end_io_acct(disk, bio_op(bio), start_time);

include/os/linux/kernel/linux/dcache_compat.h

@@ -35,6 +35,10 @@
 #define	d_make_root(inode)	d_alloc_root(inode)
 #endif /* HAVE_D_MAKE_ROOT */
 
+#ifdef HAVE_DENTRY_D_U_ALIASES
+#define	d_alias			d_u.d_alias
+#endif
+
 /*
  * 2.6.30 API change,
  * The const keyword was added to the 'struct dentry_operations' in
@@ -70,11 +74,7 @@ zpl_d_drop_aliases(struct inode *inode)
 {
 	struct dentry *dentry;
 	spin_lock(&inode->i_lock);
-#ifdef HAVE_DENTRY_D_U_ALIASES
-	hlist_for_each_entry(dentry, &inode->i_dentry, d_u.d_alias) {
-#else
 	hlist_for_each_entry(dentry, &inode->i_dentry, d_alias) {
-#endif
 		if (!IS_ROOT(dentry) && !d_mountpoint(dentry) &&
 		    (dentry->d_inode == inode)) {
 			d_drop(dentry);

include/os/linux/spl/sys/Makefile.am

@@ -20,6 +20,7 @@ KERNEL_H = \
 	kmem.h \
 	kstat.h \
 	list.h \
+	misc.h \
 	mod_os.h \
 	mutex.h \
 	param.h \

include/os/linux/spl/sys/misc.h

@@ -0,0 +1,29 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License (the "License").
+ * You may not use this file except in compliance with the License.
+ *
+ * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+ * or https://opensource.org/licenses/CDDL-1.0.
+ * See the License for the specific language governing permissions
+ * and limitations under the License.
+ *
+ * When distributing Covered Code, include this CDDL HEADER in each
+ * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+ * If applicable, add the following below this CDDL HEADER, with the
+ * fields enclosed by brackets "[]" replaced with your own identifying
+ * information: Portions Copyright [yyyy] [name of copyright owner]
+ *
+ * CDDL HEADER END
+ */
+
+#ifndef _OS_LINUX_SPL_MISC_H
+#define	_OS_LINUX_SPL_MISC_H
+
+#include <linux/kobject.h>
+
+extern void spl_signal_kobj_evt(struct block_device *bdev);
+
+#endif

include/os/linux/zfs/sys/zpl.h

@@ -67,11 +67,14 @@ extern int zpl_xattr_security_init(struct inode *ip, struct inode *dip,
 #if defined(HAVE_SET_ACL_USERNS)
 extern int zpl_set_acl(struct user_namespace *userns, struct inode *ip,
     struct posix_acl *acl, int type);
+#elif defined(HAVE_SET_ACL_USERNS_DENTRY_ARG2)
+extern int zpl_set_acl(struct user_namespace *userns, struct dentry *dentry,
+    struct posix_acl *acl, int type);
 #else
 extern int zpl_set_acl(struct inode *ip, struct posix_acl *acl, int type);
 #endif /* HAVE_SET_ACL_USERNS */
 #endif /* HAVE_SET_ACL */
-#if defined(HAVE_GET_ACL_RCU)
+#if defined(HAVE_GET_ACL_RCU) || defined(HAVE_GET_INODE_ACL)
 extern struct posix_acl *zpl_get_acl(struct inode *ip, int type, bool rcu);
 #elif defined(HAVE_GET_ACL)
 extern struct posix_acl *zpl_get_acl(struct inode *ip, int type);

include/sys/arc_impl.h

@@ -30,6 +30,7 @@
 #define	_SYS_ARC_IMPL_H
 
 #include <sys/arc.h>
+#include <sys/multilist.h>
 #include <sys/zio_crypt.h>
 #include <sys/zthr.h>
 #include <sys/aggsum.h>

include/sys/bpobj.h

@@ -87,6 +87,7 @@ int livelist_bpobj_iterate_from_nofree(bpobj_t *bpo, bpobj_itor_t func,
     void *arg, int64_t start);
 
 void bpobj_enqueue_subobj(bpobj_t *bpo, uint64_t subobj, dmu_tx_t *tx);
+void bpobj_prefetch_subobj(bpobj_t *bpo, uint64_t subobj);
 void bpobj_enqueue(bpobj_t *bpo, const blkptr_t *bp, boolean_t bp_freed,
     dmu_tx_t *tx);
 

include/sys/btree.h

@@ -65,7 +65,7 @@ extern "C" {
  * them, and increased memory overhead. Increasing these values results in
  * higher variance in operation time, and reduces memory overhead.
  */
-#define	BTREE_CORE_ELEMS	128
+#define	BTREE_CORE_ELEMS	126
 #define	BTREE_LEAF_SIZE		4096
 
 extern kmem_cache_t *zfs_btree_leaf_cache;
@@ -95,9 +95,6 @@ typedef struct zfs_btree_leaf {
 	uint8_t		btl_elems[];
 } zfs_btree_leaf_t;
 
-#define	BTREE_LEAF_ESIZE	(BTREE_LEAF_SIZE - \
-    offsetof(zfs_btree_leaf_t, btl_elems))
-
 typedef struct zfs_btree_index {
 	zfs_btree_hdr_t	*bti_node;
 	uint32_t	bti_offset;
@@ -109,14 +106,15 @@ typedef struct zfs_btree_index {
 } zfs_btree_index_t;
 
 typedef struct btree {
-	zfs_btree_hdr_t		*bt_root;
+	int (*bt_compar) (const void *, const void *);
-	int64_t			bt_height;
 	size_t			bt_elem_size;
+	size_t			bt_leaf_size;
 	uint32_t		bt_leaf_cap;
+	int32_t			bt_height;
 	uint64_t		bt_num_elems;
 	uint64_t		bt_num_nodes;
+	zfs_btree_hdr_t		*bt_root;
 	zfs_btree_leaf_t	*bt_bulk; // non-null if bulk loading
-	int (*bt_compar) (const void *, const void *);
 } zfs_btree_t;
 
 /*
@@ -132,9 +130,12 @@ void zfs_btree_fini(void);
  * compar - function to compare two nodes, it must return exactly: -1, 0, or +1
  *          -1 for <, 0 for ==, and +1 for >
  * size   - the value of sizeof(struct my_type)
+ * lsize  - custom leaf size
  */
 void zfs_btree_create(zfs_btree_t *, int (*) (const void *, const void *),
     size_t);
+void zfs_btree_create_custom(zfs_btree_t *, int (*)(const void *, const void *),
+    size_t, size_t);
 
 /*
  * Find a node with a matching value in the tree. Returns the matching node

include/sys/dnode.h

@@ -36,6 +36,7 @@
 #include <sys/dmu_zfetch.h>
 #include <sys/zrlock.h>
 #include <sys/multilist.h>
+#include <sys/wmsum.h>
 
 #ifdef	__cplusplus
 extern "C" {
@@ -587,10 +588,42 @@ typedef struct dnode_stats {
 	kstat_named_t dnode_move_active;
 } dnode_stats_t;
 
+typedef struct dnode_sums {
+	wmsum_t dnode_hold_dbuf_hold;
+	wmsum_t dnode_hold_dbuf_read;
+	wmsum_t dnode_hold_alloc_hits;
+	wmsum_t dnode_hold_alloc_misses;
+	wmsum_t dnode_hold_alloc_interior;
+	wmsum_t dnode_hold_alloc_lock_retry;
+	wmsum_t dnode_hold_alloc_lock_misses;
+	wmsum_t dnode_hold_alloc_type_none;
+	wmsum_t dnode_hold_free_hits;
+	wmsum_t dnode_hold_free_misses;
+	wmsum_t dnode_hold_free_lock_misses;
+	wmsum_t dnode_hold_free_lock_retry;
+	wmsum_t dnode_hold_free_refcount;
+	wmsum_t dnode_hold_free_overflow;
+	wmsum_t dnode_free_interior_lock_retry;
+	wmsum_t dnode_allocate;
+	wmsum_t dnode_reallocate;
+	wmsum_t dnode_buf_evict;
+	wmsum_t dnode_alloc_next_chunk;
+	wmsum_t dnode_alloc_race;
+	wmsum_t dnode_alloc_next_block;
+	wmsum_t dnode_move_invalid;
+	wmsum_t dnode_move_recheck1;
+	wmsum_t dnode_move_recheck2;
+	wmsum_t dnode_move_special;
+	wmsum_t dnode_move_handle;
+	wmsum_t dnode_move_rwlock;
+	wmsum_t dnode_move_active;
+} dnode_sums_t;
+
 extern dnode_stats_t dnode_stats;
+extern dnode_sums_t dnode_sums;
 
 #define	DNODE_STAT_INCR(stat, val) \
-    atomic_add_64(&dnode_stats.stat.value.ui64, (val));
+    wmsum_add(&dnode_sums.stat, (val))
 #define	DNODE_STAT_BUMP(stat) \
     DNODE_STAT_INCR(stat, 1);
 

include/sys/fs/zfs.h

@@ -1518,6 +1518,7 @@ typedef enum {
 #define	ZFS_ONLINE_UNSPARE	0x2
 #define	ZFS_ONLINE_FORCEFAULT	0x4
 #define	ZFS_ONLINE_EXPAND	0x8
+#define	ZFS_ONLINE_SPARE	0x10
 #define	ZFS_OFFLINE_TEMPORARY	0x1
 
 /*

include/sys/spa.h

@@ -837,7 +837,7 @@ extern kmutex_t spa_namespace_lock;
 #define	SPA_CONFIG_UPDATE_POOL	0
 #define	SPA_CONFIG_UPDATE_VDEVS	1
 
-extern void spa_write_cachefile(spa_t *, boolean_t, boolean_t);
+extern void spa_write_cachefile(spa_t *, boolean_t, boolean_t, boolean_t);
 extern void spa_config_load(void);
 extern nvlist_t *spa_all_configs(uint64_t *);
 extern void spa_config_set(spa_t *spa, nvlist_t *config);

include/sys/vdev.h

@@ -147,6 +147,7 @@ extern int vdev_degrade(spa_t *spa, uint64_t guid, vdev_aux_t aux);
 extern int vdev_online(spa_t *spa, uint64_t guid, uint64_t flags,
     vdev_state_t *);
 extern int vdev_offline(spa_t *spa, uint64_t guid, uint64_t flags);
+extern int vdev_remove_wanted(spa_t *spa, uint64_t guid);
 extern void vdev_clear(spa_t *spa, vdev_t *vd);
 
 extern boolean_t vdev_is_dead(vdev_t *vd);
@@ -189,6 +190,8 @@ typedef enum vdev_config_flag {
 	VDEV_CONFIG_MISSING = 1 << 4
 } vdev_config_flag_t;
 
+extern void vdev_post_kobj_evt(vdev_t *vd);
+extern void vdev_clear_kobj_evt(vdev_t *vd);
 extern void vdev_top_config_generate(spa_t *spa, nvlist_t *config);
 extern nvlist_t *vdev_config_generate(spa_t *spa, vdev_t *vd,
     boolean_t getstats, vdev_config_flag_t flags);

include/sys/vdev_impl.h

@@ -69,6 +69,7 @@ extern uint32_t zfs_vdev_async_write_max_active;
  * Virtual device operations
  */
 typedef int	vdev_init_func_t(spa_t *spa, nvlist_t *nv, void **tsd);
+typedef void	vdev_kobj_post_evt_func_t(vdev_t *vd);
 typedef void	vdev_fini_func_t(vdev_t *vd);
 typedef int	vdev_open_func_t(vdev_t *vd, uint64_t *size, uint64_t *max_size,
     uint64_t *ashift, uint64_t *pshift);
@@ -123,6 +124,7 @@ typedef const struct vdev_ops {
 	vdev_config_generate_func_t	*vdev_op_config_generate;
 	vdev_nparity_func_t		*vdev_op_nparity;
 	vdev_ndisks_func_t		*vdev_op_ndisks;
+	vdev_kobj_post_evt_func_t	*vdev_op_kobj_evt_post;
 	char				vdev_op_type[16];
 	boolean_t			vdev_op_leaf;
 } vdev_ops_t;
@@ -435,6 +437,7 @@ struct vdev {
 	boolean_t	vdev_isl2cache;	/* was a l2cache device		*/
 	boolean_t	vdev_copy_uberblocks;  /* post expand copy uberblocks */
 	boolean_t	vdev_resilver_deferred;  /* resilver deferred */
+	boolean_t	vdev_kobj_flag; /* kobj event record */
 	vdev_queue_t	vdev_queue;	/* I/O deadline schedule queue	*/
 	vdev_cache_t	vdev_cache;	/* physical block cache		*/
 	spa_aux_vdev_t	*vdev_aux;	/* for l2cache and spares vdevs	*/

include/sys/zap_impl.h

@@ -66,10 +66,9 @@ typedef struct mzap_phys {
 } mzap_phys_t;
 
 typedef struct mzap_ent {
-	avl_node_t mze_node;
+	uint32_t mze_hash;
-	int mze_chunkid;
+	uint16_t mze_cd; /* copy from mze_phys->mze_cd */
-	uint64_t mze_hash;
+	uint16_t mze_chunkid;
-	uint32_t mze_cd; /* copy from mze_phys->mze_cd */
 } mzap_ent_t;
 
 #define	MZE_PHYS(zap, mze) \
@@ -164,7 +163,7 @@ typedef struct zap {
 			int16_t zap_num_entries;
 			int16_t zap_num_chunks;
 			int16_t zap_alloc_next;
-			avl_tree_t zap_avl;
+			zfs_btree_t zap_tree;
 		} zap_micro;
 	} zap_u;
 } zap_t;
@@ -202,7 +201,7 @@ int zap_lockdir(objset_t *os, uint64_t obj, dmu_tx_t *tx,
     krw_t lti, boolean_t fatreader, boolean_t adding, void *tag, zap_t **zapp);
 void zap_unlockdir(zap_t *zap, void *tag);
 void zap_evict_sync(void *dbu);
-zap_name_t *zap_name_alloc(zap_t *zap, const char *key, matchtype_t mt);
+zap_name_t *zap_name_alloc_str(zap_t *zap, const char *key, matchtype_t mt);
 void zap_name_free(zap_name_t *zn);
 int zap_hashbits(zap_t *zap);
 uint32_t zap_maxcd(zap_t *zap);

include/sys/zfs_context.h

@@ -51,6 +51,7 @@ extern "C" {
 #include <sys/kmem.h>
 #include <sys/kmem_cache.h>
 #include <sys/vmem.h>
+#include <sys/misc.h>
 #include <sys/taskq.h>
 #include <sys/param.h>
 #include <sys/disp.h>

include/sys/zfs_znode.h

@@ -217,9 +217,9 @@ typedef struct znode {
 
 typedef struct znode_hold {
 	uint64_t	zh_obj;		/* object id */
-	kmutex_t	zh_lock;	/* lock serializing object access */
 	avl_node_t	zh_node;	/* avl tree linkage */
-	zfs_refcount_t	zh_refcount;	/* active consumer reference count */
+	kmutex_t	zh_lock;	/* lock serializing object access */
+	int		zh_refcount;	/* active consumer reference count */
 } znode_hold_t;
 
 static inline uint64_t

include/sys/zrlock.h

@@ -34,9 +34,8 @@ extern "C" {
 
 typedef struct zrlock {
 	kmutex_t zr_mtx;
-	volatile int32_t zr_refcount;
 	kcondvar_t zr_cv;
-	uint16_t zr_pad;
+	volatile int32_t zr_refcount;
 #ifdef	ZFS_DEBUG
 	kthread_t *zr_owner;
 	const char *zr_caller;

lib/libzfs/libzfs.abi

@@ -99,6 +99,7 @@
     <elf-symbol name='sa_validate_shareopts' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='snapshot_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='unshare_one' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
+    <elf-symbol name='use_color' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='zcmd_alloc_dst_nvlist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='zcmd_expand_dst_nvlist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='zcmd_free_nvlists' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
@@ -378,6 +379,7 @@
     <elf-symbol name='zpool_vdev_path_to_guid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='zpool_vdev_remove' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='zpool_vdev_remove_cancel' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
+    <elf-symbol name='zpool_vdev_remove_wanted' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='zpool_vdev_split' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='zpool_wait' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='zpool_wait_status' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
@@ -5069,6 +5071,9 @@
       <parameter type-id='9cf59a50'/>
       <return type-id='48b5725f'/>
     </function-decl>
+    <function-decl name='use_color' mangled-name='use_color' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='use_color'>
+      <return type-id='95e97e5e'/>
+    </function-decl>
     <function-decl name='mkdirp' visibility='default' binding='global' size-in-bits='64'>
       <parameter type-id='80f4b756'/>
       <parameter type-id='d50d396c'/>
@@ -5858,6 +5863,11 @@
       <parameter type-id='c19b74c3' name='istmp'/>
       <return type-id='95e97e5e'/>
     </function-decl>
+    <function-decl name='zpool_vdev_remove_wanted' mangled-name='zpool_vdev_remove_wanted' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_remove_wanted'>
+      <parameter type-id='4c81de99' name='zhp'/>
+      <parameter type-id='80f4b756' name='path'/>
+      <return type-id='95e97e5e'/>
+    </function-decl>
     <function-decl name='zpool_vdev_fault' mangled-name='zpool_vdev_fault' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_fault'>
       <parameter type-id='4c81de99' name='zhp'/>
       <parameter type-id='9c313c2d' name='guid'/>
@@ -6098,7 +6108,7 @@
     <array-type-def dimensions='1' type-id='b96825af' size-in-bits='64' id='13339fda'>
       <subrange length='8' type-id='7359adad' id='56e0c0b1'/>
     </array-type-def>
-    <class-decl name='sendflags' size-in-bits='544' is-struct='yes' visibility='default' id='f6aa15be'>
+    <class-decl name='sendflags' size-in-bits='576' is-struct='yes' visibility='default' id='f6aa15be'>
       <data-member access='public' layout-offset-in-bits='0'>
         <var-decl name='verbosity' type-id='95e97e5e' visibility='default'/>
       </data-member>
@@ -6130,24 +6140,27 @@
         <var-decl name='progress' type-id='c19b74c3' visibility='default'/>
       </data-member>
       <data-member access='public' layout-offset-in-bits='320'>
-        <var-decl name='largeblock' type-id='c19b74c3' visibility='default'/>
+        <var-decl name='progressastitle' type-id='c19b74c3' visibility='default'/>
       </data-member>
       <data-member access='public' layout-offset-in-bits='352'>
-        <var-decl name='embed_data' type-id='c19b74c3' visibility='default'/>
+        <var-decl name='largeblock' type-id='c19b74c3' visibility='default'/>
       </data-member>
       <data-member access='public' layout-offset-in-bits='384'>
-        <var-decl name='compress' type-id='c19b74c3' visibility='default'/>
+        <var-decl name='embed_data' type-id='c19b74c3' visibility='default'/>
       </data-member>
       <data-member access='public' layout-offset-in-bits='416'>
-        <var-decl name='raw' type-id='c19b74c3' visibility='default'/>
+        <var-decl name='compress' type-id='c19b74c3' visibility='default'/>
       </data-member>
       <data-member access='public' layout-offset-in-bits='448'>
-        <var-decl name='backup' type-id='c19b74c3' visibility='default'/>
+        <var-decl name='raw' type-id='c19b74c3' visibility='default'/>
       </data-member>
       <data-member access='public' layout-offset-in-bits='480'>
-        <var-decl name='holds' type-id='c19b74c3' visibility='default'/>
+        <var-decl name='backup' type-id='c19b74c3' visibility='default'/>
       </data-member>
       <data-member access='public' layout-offset-in-bits='512'>
+        <var-decl name='holds' type-id='c19b74c3' visibility='default'/>
+      </data-member>
+      <data-member access='public' layout-offset-in-bits='544'>
         <var-decl name='saved' type-id='c19b74c3' visibility='default'/>
       </data-member>
     </class-decl>
@@ -6728,6 +6741,11 @@
       <parameter type-id='95e97e5e'/>
       <return type-id='48b5725f'/>
     </function-decl>
+    <function-decl name='zfs_setproctitle' visibility='default' binding='global' size-in-bits='64'>
+      <parameter type-id='80f4b756'/>
+      <parameter is-variadic='yes'/>
+      <return type-id='48b5725f'/>
+    </function-decl>
     <function-decl name='avl_insert' visibility='default' binding='global' size-in-bits='64'>
       <parameter type-id='a3681dea'/>
       <parameter type-id='eaa32e2f'/>

lib/libzfs/libzfs_diff.c

@@ -45,16 +45,21 @@
 #include <pthread.h>
 #include <sys/zfs_ioctl.h>
 #include <libzfs.h>
+#include <libzutil.h>
 #include "libzfs_impl.h"
 
 #define	ZDIFF_SNAPDIR		"/.zfs/snapshot/"
 #define	ZDIFF_PREFIX		"zfs-diff-%d"
 
 #define	ZDIFF_ADDED	'+'
-#define	ZDIFF_MODIFIED	'M'
+#define	ZDIFF_MODIFIED	"M"
 #define	ZDIFF_REMOVED	'-'
-#define	ZDIFF_RENAMED	'R'
+#define	ZDIFF_RENAMED	"R"
 
+#define	ZDIFF_ADDED_COLOR    ANSI_GREEN
+#define	ZDIFF_MODIFIED_COLOR ANSI_YELLOW
+#define	ZDIFF_REMOVED_COLOR  ANSI_RED
+#define	ZDIFF_RENAMED_COLOR  ANSI_BOLD_BLUE
 
 /*
  * Given a {dsname, object id}, get the object path
@@ -129,48 +134,54 @@ stream_bytes(FILE *fp, const char *string)
 	}
 }
 
-static void
+/*
-print_what(FILE *fp, mode_t what)
+ * Takes the type of change (like `print_file`), outputs the appropriate color
+ */
+static const char *
+type_to_color(char type)
 {
-	char symbol;
+	if (type == '+')
+		return (ZDIFF_ADDED_COLOR);
+	else if (type == '-')
+		return (ZDIFF_REMOVED_COLOR);
+	else if (type == 'M')
+		return (ZDIFF_MODIFIED_COLOR);
+	else if (type == 'R')
+		return (ZDIFF_RENAMED_COLOR);
+	else
+		return (NULL);
+}
 
+
+static char
+get_what(mode_t what)
+{
 	switch (what & S_IFMT) {
 	case S_IFBLK:
-		symbol = 'B';
+		return ('B');
-		break;
 	case S_IFCHR:
-		symbol = 'C';
+		return ('C');
-		break;
 	case S_IFDIR:
-		symbol = '/';
+		return ('/');
-		break;
 #ifdef S_IFDOOR
 	case S_IFDOOR:
-		symbol = '>';
+		return ('>');
-		break;
 #endif
 	case S_IFIFO:
-		symbol = '|';
+		return ('|');
-		break;
 	case S_IFLNK:
-		symbol = '@';
+		return ('@');
-		break;
 #ifdef S_IFPORT
 	case S_IFPORT:
-		symbol = 'P';
+		return ('P');
-		break;
 #endif
 	case S_IFSOCK:
-		symbol = '=';
+		return ('=');
-		break;
 	case S_IFREG:
-		symbol = 'F';
+		return ('F');
-		break;
 	default:
-		symbol = '?';
+		return ('?');
-		break;
 	}
-	(void) fprintf(fp, "%c", symbol);
 }
 
 static void
@@ -189,57 +200,63 @@ static void
 print_rename(FILE *fp, differ_info_t *di, const char *old, const char *new,
     zfs_stat_t *isb)
 {
+	if (isatty(fileno(fp)))
+		color_start(ZDIFF_RENAMED_COLOR);
 	if (di->timestamped)
 		(void) fprintf(fp, "%10lld.%09lld\t",
 		    (longlong_t)isb->zs_ctime[0],
 		    (longlong_t)isb->zs_ctime[1]);
-	(void) fprintf(fp, "%c\t", ZDIFF_RENAMED);
+	(void) fputs(ZDIFF_RENAMED "\t", fp);
-	if (di->classify) {
+	if (di->classify)
-		print_what(fp, isb->zs_mode);
+		(void) fprintf(fp, "%c\t", get_what(isb->zs_mode));
-		(void) fprintf(fp, "\t");
-	}
 	print_cmn(fp, di, old);
-	if (di->scripted)
+	(void) fputs(di->scripted ? "\t" : " -> ", fp);
-		(void) fprintf(fp, "\t");
-	else
-		(void) fprintf(fp, " -> ");
 	print_cmn(fp, di, new);
-	(void) fprintf(fp, "\n");
+	(void) fputc('\n', fp);
+
+	if (isatty(fileno(fp)))
+		color_end();
 }
 
 static void
 print_link_change(FILE *fp, differ_info_t *di, int delta, const char *file,
     zfs_stat_t *isb)
 {
+	if (isatty(fileno(fp)))
+		color_start(ZDIFF_MODIFIED_COLOR);
+
 	if (di->timestamped)
 		(void) fprintf(fp, "%10lld.%09lld\t",
 		    (longlong_t)isb->zs_ctime[0],
 		    (longlong_t)isb->zs_ctime[1]);
-	(void) fprintf(fp, "%c\t", ZDIFF_MODIFIED);
+	(void) fputs(ZDIFF_MODIFIED "\t", fp);
-	if (di->classify) {
+	if (di->classify)
-		print_what(fp, isb->zs_mode);
+		(void) fprintf(fp, "%c\t", get_what(isb->zs_mode));
-		(void) fprintf(fp, "\t");
-	}
 	print_cmn(fp, di, file);
-	(void) fprintf(fp, "\t(%+d)", delta);
+	(void) fprintf(fp, "\t(%+d)\n", delta);
-	(void) fprintf(fp, "\n");
+	if (isatty(fileno(fp)))
+		color_end();
 }
 
 static void
 print_file(FILE *fp, differ_info_t *di, char type, const char *file,
     zfs_stat_t *isb)
 {
+	if (isatty(fileno(fp)))
+		color_start(type_to_color(type));
+
 	if (di->timestamped)
 		(void) fprintf(fp, "%10lld.%09lld\t",
 		    (longlong_t)isb->zs_ctime[0],
 		    (longlong_t)isb->zs_ctime[1]);
 	(void) fprintf(fp, "%c\t", type);
-	if (di->classify) {
+	if (di->classify)
-		print_what(fp, isb->zs_mode);
+		(void) fprintf(fp, "%c\t", get_what(isb->zs_mode));
-		(void) fprintf(fp, "\t");
-	}
 	print_cmn(fp, di, file);
-	(void) fprintf(fp, "\n");
+	(void) fputc('\n', fp);
+
+	if (isatty(fileno(fp)))
+		color_end();
 }
 
 static int
@@ -332,7 +349,7 @@ write_inuse_diffs_one(FILE *fp, differ_info_t *di, uint64_t dobj)
 			print_link_change(fp, di, change,
 			    change > 0 ? fobjname : tobjname, &tsb);
 		} else if (strcmp(fobjname, tobjname) == 0) {
-			print_file(fp, di, ZDIFF_MODIFIED, fobjname, &tsb);
+			print_file(fp, di, *ZDIFF_MODIFIED, fobjname, &tsb);
 		} else {
 			print_rename(fp, di, fobjname, tobjname, &tsb);
 		}

lib/libzfs/libzfs_pool.c

@@ -3051,7 +3051,7 @@ zpool_vdev_online(zpool_handle_t *zhp, const char *path, int flags,
 
 	verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0);
 
-	if (avail_spare)
+	if (!(flags & ZFS_ONLINE_SPARE) && avail_spare)
 		return (zfs_error(hdl, EZFS_ISSPARE, msg));
 
 	if ((flags & ZFS_ONLINE_EXPAND ||
@@ -3158,6 +3158,40 @@ zpool_vdev_offline(zpool_handle_t *zhp, const char *path, boolean_t istmp)
 	}
 }
 
+/*
+ * Remove the specified vdev asynchronously from the configuration, so
+ * that it may come ONLINE if reinserted. This is called from zed on
+ * Udev remove event.
+ * Note: We also have a similar function zpool_vdev_remove() that
+ * removes the vdev from the pool.
+ */
+int
+zpool_vdev_remove_wanted(zpool_handle_t *zhp, const char *path)
+{
+	zfs_cmd_t zc = {"\0"};
+	char errbuf[1024];
+	nvlist_t *tgt;
+	boolean_t avail_spare, l2cache;
+	libzfs_handle_t *hdl = zhp->zpool_hdl;
+
+	(void) snprintf(errbuf, sizeof (errbuf),
+	    dgettext(TEXT_DOMAIN, "cannot remove %s"), path);
+
+	(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+	if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache,
+	    NULL)) == NULL)
+		return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
+
+	zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
+
+	zc.zc_cookie = VDEV_STATE_REMOVED;
+
+	if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) == 0)
+		return (0);
+
+	return (zpool_standard_error(hdl, errno, errbuf));
+}
+
 /*
  * Mark the given vdev faulted.
  */

lib/libzfs/libzfs_sendrecv.c

@@ -83,6 +83,9 @@ typedef struct progress_arg {
 	boolean_t pa_parsable;
 	boolean_t pa_estimate;
 	int pa_verbosity;
+	boolean_t pa_astitle;
+	boolean_t pa_progress;
+	uint64_t pa_size;
 } progress_arg_t;
 
 static int
@@ -712,6 +715,7 @@ typedef struct send_dump_data {
 	boolean_t seenfrom, seento, replicate, doall, fromorigin;
 	boolean_t dryrun, parsable, progress, embed_data, std_out;
 	boolean_t large_block, compress, raw, holds;
+	boolean_t progressastitle;
 	int outfd;
 	boolean_t err;
 	nvlist_t *fss;
@@ -904,6 +908,7 @@ send_progress_thread(void *arg)
 	zfs_handle_t *zhp = pa->pa_zhp;
 	uint64_t bytes;
 	uint64_t blocks;
+	uint64_t total = pa->pa_size / 100;
 	char buf[16];
 	time_t t;
 	struct tm *tm;
@@ -922,7 +927,7 @@ send_progress_thread(void *arg)
 			return ((void *)(uintptr_t)err);
 		}
 
-		if (firstloop && !pa->pa_parsable) {
+		if (firstloop && !pa->pa_parsable && pa->pa_progress) {
 			(void) fprintf(stderr,
 			    "TIME       %s   %sSNAPSHOT %s\n",
 			    pa->pa_estimate ? "BYTES" : " SENT",
@@ -934,6 +939,17 @@ send_progress_thread(void *arg)
 		(void) time(&t);
 		tm = localtime(&t);
 
+		if (pa->pa_astitle) {
+			char buf_bytes[16];
+			char buf_size[16];
+			int pct;
+			zfs_nicenum(bytes, buf_bytes, sizeof (buf_bytes));
+			zfs_nicenum(pa->pa_size, buf_size, sizeof (buf_size));
+			pct = (total > 0) ? bytes / total : 100;
+			zfs_setproctitle("sending %s (%d%%: %s/%s)",
+			    zhp->zfs_name, MIN(pct, 100), buf_bytes, buf_size);
+		}
+
 		if (pa->pa_verbosity >= 2 && pa->pa_parsable) {
 			(void) fprintf(stderr,
 			    "%02d:%02d:%02d\t%llu\t%llu\t%s\n",
@@ -950,7 +966,7 @@ send_progress_thread(void *arg)
 			(void) fprintf(stderr, "%02d:%02d:%02d\t%llu\t%s\n",
 			    tm->tm_hour, tm->tm_min, tm->tm_sec,
 			    (u_longlong_t)bytes, zhp->zfs_name);
-		} else {
+		} else if (pa->pa_progress) {
 			zfs_nicebytes(bytes, buf, sizeof (buf));
 			(void) fprintf(stderr, "%02d:%02d:%02d   %5s   %s\n",
 			    tm->tm_hour, tm->tm_min, tm->tm_sec,
@@ -1114,12 +1130,15 @@ dump_snapshot(zfs_handle_t *zhp, void *arg)
 		 * If progress reporting is requested, spawn a new thread to
 		 * poll ZFS_IOC_SEND_PROGRESS at a regular interval.
 		 */
-		if (sdd->progress) {
+		if (sdd->progress || sdd->progressastitle) {
 			pa.pa_zhp = zhp;
 			pa.pa_fd = sdd->outfd;
 			pa.pa_parsable = sdd->parsable;
 			pa.pa_estimate = B_FALSE;
 			pa.pa_verbosity = sdd->verbosity;
+			pa.pa_size = sdd->size;
+			pa.pa_astitle = sdd->progressastitle;
+			pa.pa_progress = sdd->progress;
 
 			if ((err = pthread_create(&tid, NULL,
 			    send_progress_thread, &pa)) != 0) {
@@ -1131,7 +1150,7 @@ dump_snapshot(zfs_handle_t *zhp, void *arg)
 		err = dump_ioctl(zhp, sdd->prevsnap, sdd->prevsnap_obj,
 		    fromorigin, sdd->outfd, flags, sdd->debugnv);
 
-		if (sdd->progress) {
+		if (sdd->progress || sdd->progressastitle) {
 			void *status = NULL;
 			(void) pthread_cancel(tid);
 			(void) pthread_join(tid, &status);
@@ -1462,7 +1481,7 @@ lzc_flags_from_sendflags(const sendflags_t *flags)
 static int
 estimate_size(zfs_handle_t *zhp, const char *from, int fd, sendflags_t *flags,
     uint64_t resumeobj, uint64_t resumeoff, uint64_t bytes,
-    const char *redactbook, char *errbuf)
+    const char *redactbook, char *errbuf, uint64_t *sizep)
 {
 	uint64_t size;
 	FILE *fout = flags->dryrun ? stdout : stderr;
@@ -1470,7 +1489,7 @@ estimate_size(zfs_handle_t *zhp, const char *from, int fd, sendflags_t *flags,
 	int err = 0;
 	pthread_t ptid;
 
-	if (flags->progress) {
+	if (flags->progress || flags->progressastitle) {
 		pa.pa_zhp = zhp;
 		pa.pa_fd = fd;
 		pa.pa_parsable = flags->parsable;
@@ -1489,8 +1508,9 @@ estimate_size(zfs_handle_t *zhp, const char *from, int fd, sendflags_t *flags,
 	err = lzc_send_space_resume_redacted(zhp->zfs_name, from,
 	    lzc_flags_from_sendflags(flags), resumeobj, resumeoff, bytes,
 	    redactbook, fd, &size);
+	*sizep = size;
 
-	if (flags->progress) {
+	if (flags->progress || flags->progressastitle) {
 		void *status = NULL;
 		(void) pthread_cancel(ptid);
 		(void) pthread_join(ptid, &status);
@@ -1505,6 +1525,9 @@ estimate_size(zfs_handle_t *zhp, const char *from, int fd, sendflags_t *flags,
 		}
 	}
 
+	if (!flags->progress && !flags->parsable)
+		return (err);
+
 	if (err != 0) {
 		zfs_error_aux(zhp->zfs_hdl, "%s", strerror(err));
 		return (zfs_error(zhp->zfs_hdl, EZFS_BADBACKUP,
@@ -1638,6 +1661,7 @@ zfs_send_resume_impl(libzfs_handle_t *hdl, sendflags_t *flags, int outfd,
 	uint64_t *redact_snap_guids = NULL;
 	int num_redact_snaps = 0;
 	char *redact_book = NULL;
+	uint64_t size = 0;
 
 	(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
 	    "cannot resume send"));
@@ -1731,7 +1755,7 @@ zfs_send_resume_impl(libzfs_handle_t *hdl, sendflags_t *flags, int outfd,
 		}
 	}
 
-	if (flags->verbosity != 0) {
+	if (flags->verbosity != 0 || flags->progressastitle) {
 		/*
 		 * Some of these may have come from the resume token, set them
 		 * here for size estimate purposes.
@@ -1748,7 +1772,7 @@ zfs_send_resume_impl(libzfs_handle_t *hdl, sendflags_t *flags, int outfd,
 		if (lzc_flags & LZC_SEND_FLAG_SAVED)
 			tmpflags.saved = B_TRUE;
 		error = estimate_size(zhp, fromname, outfd, &tmpflags,
-		    resumeobj, resumeoff, bytes, redact_book, errbuf);
+		    resumeobj, resumeoff, bytes, redact_book, errbuf, &size);
 	}
 
 	if (!flags->dryrun) {
@@ -1758,12 +1782,15 @@ zfs_send_resume_impl(libzfs_handle_t *hdl, sendflags_t *flags, int outfd,
 		 * If progress reporting is requested, spawn a new thread to
 		 * poll ZFS_IOC_SEND_PROGRESS at a regular interval.
 		 */
-		if (flags->progress) {
+		if (flags->progress || flags->progressastitle) {
 			pa.pa_zhp = zhp;
 			pa.pa_fd = outfd;
 			pa.pa_parsable = flags->parsable;
 			pa.pa_estimate = B_FALSE;
 			pa.pa_verbosity = flags->verbosity;
+			pa.pa_size = size;
+			pa.pa_astitle = flags->progressastitle;
+			pa.pa_progress = flags->progress;
 
 			error = pthread_create(&tid, NULL,
 			    send_progress_thread, &pa);
@@ -1780,7 +1807,7 @@ zfs_send_resume_impl(libzfs_handle_t *hdl, sendflags_t *flags, int outfd,
 		if (redact_book != NULL)
 			free(redact_book);
 
-		if (flags->progress) {
+		if (flags->progress || flags->progress) {
 			void *status = NULL;
 			(void) pthread_cancel(tid);
 			(void) pthread_join(tid, &status);
@@ -1790,6 +1817,7 @@ zfs_send_resume_impl(libzfs_handle_t *hdl, sendflags_t *flags, int outfd,
 				(void) snprintf(errbuf, sizeof (errbuf),
 				    dgettext(TEXT_DOMAIN,
 				    "progress thread exited nonzero"));
+				zfs_close(zhp);
 				return (zfs_standard_error(hdl, error, errbuf));
 			}
 		}
@@ -2199,6 +2227,7 @@ zfs_send(zfs_handle_t *zhp, const char *fromsnap, const char *tosnap,
 	sdd.verbosity = flags->verbosity;
 	sdd.parsable = flags->parsable;
 	sdd.progress = flags->progress;
+	sdd.progressastitle = flags->progressastitle;
 	sdd.dryrun = flags->dryrun;
 	sdd.large_block = flags->largeblock;
 	sdd.embed_data = flags->embed_data;
@@ -2410,6 +2439,7 @@ zfs_send_one(zfs_handle_t *zhp, const char *from, int fd, sendflags_t *flags,
 	char *name = zhp->zfs_name;
 	pthread_t ptid;
 	progress_arg_t pa = { 0 };
+	uint64_t size = 0;
 
 	char errbuf[1024];
 	(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
@@ -2492,9 +2522,9 @@ zfs_send_one(zfs_handle_t *zhp, const char *from, int fd, sendflags_t *flags,
 	/*
 	 * Perform size estimate if verbose was specified.
 	 */
-	if (flags->verbosity != 0) {
+	if (flags->verbosity != 0 || flags->progressastitle) {
 		err = estimate_size(zhp, from, fd, flags, 0, 0, 0, redactbook,
-		    errbuf);
+		    errbuf, &size);
 		if (err != 0)
 			return (err);
 	}
@@ -2506,12 +2536,15 @@ zfs_send_one(zfs_handle_t *zhp, const char *from, int fd, sendflags_t *flags,
 	 * If progress reporting is requested, spawn a new thread to poll
 	 * ZFS_IOC_SEND_PROGRESS at a regular interval.
 	 */
-	if (flags->progress) {
+	if (flags->progress || flags->progressastitle) {
 		pa.pa_zhp = zhp;
 		pa.pa_fd = fd;
 		pa.pa_parsable = flags->parsable;
 		pa.pa_estimate = B_FALSE;
 		pa.pa_verbosity = flags->verbosity;
+		pa.pa_size = size;
+		pa.pa_astitle = flags->progressastitle;
+		pa.pa_progress = flags->progress;
 
 		err = pthread_create(&ptid, NULL,
 		    send_progress_thread, &pa);
@@ -2525,7 +2558,7 @@ zfs_send_one(zfs_handle_t *zhp, const char *from, int fd, sendflags_t *flags,
 	err = lzc_send_redacted(name, from, fd,
 	    lzc_flags_from_sendflags(flags), redactbook);
 
-	if (flags->progress) {
+	if (flags->progress || flags->progressastitle) {
 		void *status = NULL;
 		(void) pthread_cancel(ptid);
 		(void) pthread_join(ptid, &status);
@@ -3966,6 +3999,15 @@ zfs_setup_cmdline_props(libzfs_handle_t *hdl, zfs_type_t type,
 			goto error;
 		}
 
+		/*
+		 * For plain replicated send, we can ignore encryption
+		 * properties other than first stream
+		 */
+		if ((zfs_prop_encryption_key_param(prop) || prop ==
+		    ZFS_PROP_ENCRYPTION) && !newfs && recursive && !raw) {
+			continue;
+		}
+
 		/* incremental streams can only exclude encryption properties */
 		if ((zfs_prop_encryption_key_param(prop) ||
 		    prop == ZFS_PROP_ENCRYPTION) && !newfs &&
@@ -4065,7 +4107,8 @@ zfs_setup_cmdline_props(libzfs_handle_t *hdl, zfs_type_t type,
 		if (cp != NULL)
 			*cp = '\0';
 
-		if (!raw && zfs_crypto_create(hdl, namebuf, voprops, NULL,
+		if (!raw && !(!newfs && recursive) &&
+		    zfs_crypto_create(hdl, namebuf, voprops, NULL,
 		    B_FALSE, wkeydata_out, wkeylen_out) != 0) {
 			fnvlist_free(voprops);
 			ret = zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf);

lib/libzfs/libzfs_util.c

@@ -2027,7 +2027,7 @@ zfs_version_print(void)
  * Return 1 if the user requested ANSI color output, and our terminal supports
  * it.  Return 0 for no color.
  */
-static int
+int
 use_color(void)
 {
 	static int use_color = -1;
@@ -2073,31 +2073,38 @@ use_color(void)
 }
 
 /*
- * color_start() and color_end() are used for when you want to colorize a block
+ * The functions color_start() and color_end() are used for when you want
- * of text.  For example:
+ * to colorize a block of text.
  *
- * color_start(ANSI_RED_FG)
+ * For example:
+ * color_start(ANSI_RED)
  * printf("hello");
  * printf("world");
  * color_end();
  */
 void
-color_start(char *color)
+color_start(const char *color)
 {
-	if (use_color())
+	if (color && use_color()) {
-		printf("%s", color);
+		fputs(color, stdout);
+		fflush(stdout);
+	}
 }
 
 void
 color_end(void)
 {
-	if (use_color())
+	if (use_color()) {
-		printf(ANSI_RESET);
+		fputs(ANSI_RESET, stdout);
+		fflush(stdout);
+	}
 }
 
-/* printf() with a color.  If color is NULL, then do a normal printf. */
+/*
+ * printf() with a color. If color is NULL, then do a normal printf.
+ */
 int
-printf_color(char *color, char *format, ...)
+printf_color(const char *color, char *format, ...)
 {
 	va_list aptr;
 	int rc;

lib/libzfs/os/linux/libzfs_sendrecv_os.c

@@ -35,6 +35,22 @@
 void
 libzfs_set_pipe_max(int infd)
 {
+#if __linux__
+	/*
+	 * Sadly, Linux has an unfixed deadlock if you do SETPIPE_SZ on a pipe
+	 * with data in it.
+	 * cf. #13232, https://bugzilla.kernel.org/show_bug.cgi?id=212295
+	 *
+	 * And since the problem is in waking up the writer, there's nothing
+	 * we can do about it from here.
+	 *
+	 * So if people want to, they can set this, but they
+	 * may regret it...
+	 */
+	if (getenv("ZFS_SET_PIPE_MAX") == NULL)
+		return;
+#endif
+
 	FILE *procf = fopen("/proc/sys/fs/pipe-max-size", "re");
 
 	if (procf != NULL) {

lib/libzfs_core/libzfs_core.abi

@@ -272,6 +272,8 @@
     <elf-symbol name='zfs_niceraw' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='zfs_nicetime' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='zfs_resolve_shortname' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
+    <elf-symbol name='zfs_setproctitle' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
+    <elf-symbol name='zfs_setproctitle_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='zfs_strcmp_pathname' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='zfs_strip_partition' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
     <elf-symbol name='zfs_strip_path' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
@@ -3340,6 +3342,30 @@
       <return type-id='95e97e5e'/>
     </function-decl>
   </abi-instr>
+  <abi-instr address-size='64' path='os/linux/zutil_setproctitle.c' language='LANG_C99'>
+    <function-decl name='warnx' visibility='default' binding='global' size-in-bits='64'>
+      <parameter type-id='80f4b756'/>
+      <parameter is-variadic='yes'/>
+      <return type-id='48b5725f'/>
+    </function-decl>
+    <function-decl name='setenv' visibility='default' binding='global' size-in-bits='64'>
+      <parameter type-id='80f4b756'/>
+      <parameter type-id='80f4b756'/>
+      <parameter type-id='95e97e5e'/>
+      <return type-id='95e97e5e'/>
+    </function-decl>
+    <function-decl name='zfs_setproctitle_init' mangled-name='zfs_setproctitle_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_setproctitle_init'>
+      <parameter type-id='95e97e5e' name='argc'/>
+      <parameter type-id='9b23c9ad' name='argv'/>
+      <parameter type-id='9b23c9ad' name='envp'/>
+      <return type-id='48b5725f'/>
+    </function-decl>
+    <function-decl name='zfs_setproctitle' mangled-name='zfs_setproctitle' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_setproctitle'>
+      <parameter type-id='80f4b756' name='fmt'/>
+      <parameter is-variadic='yes'/>
+      <return type-id='48b5725f'/>
+    </function-decl>
+  </abi-instr>
   <abi-instr address-size='64' path='zutil_device_path.c' language='LANG_C99'>
     <qualified-type-def type-id='26a90f95' restrict='yes' id='266fe297'/>
     <qualified-type-def type-id='80f4b756' const='yes' id='b99c00c9'/>

lib/libzutil/Makefile.am

@@ -18,6 +18,7 @@ USER_C = \
 
 if BUILD_LINUX
 USER_C += \
+	os/linux/zutil_setproctitle.c \
 	os/linux/zutil_device_path_os.c \
 	os/linux/zutil_import_os.c \
 	os/linux/zutil_compat.c

lib/libzutil/os/linux/zutil_setproctitle.c

@@ -0,0 +1,299 @@
+/*
+ * Copyright © 2013 Guillem Jover <guillem@hadrons.org>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. The name of the author may not be used to endorse or promote products
+ *    derived from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
+ * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY
+ * AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL
+ * THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+ * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+ * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
+ * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
+ * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
+ * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
+ * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <errno.h>
+#include <stddef.h>
+#include <stdarg.h>
+#include <stdlib.h>
+#include <stdio.h>
+#include <err.h>
+#include <unistd.h>
+#include <string.h>
+#include <sys/param.h>
+#include <libzutil.h>
+
+static struct {
+	/* Original value. */
+	const char *arg0;
+
+	/* Title space available. */
+	char *base, *end;
+
+	/* Pointer to original nul character within base. */
+	char *nul;
+
+	boolean_t warned;
+	boolean_t reset;
+	int error;
+} SPT;
+
+#define	LIBBSD_IS_PATHNAME_SEPARATOR(c) ((c) == '/')
+#define	SPT_MAXTITLE 255
+
+extern const char *__progname;
+
+static const char *
+getprogname(void)
+{
+	return (__progname);
+}
+
+static void
+setprogname(const char *progname)
+{
+	size_t i;
+
+	for (i = strlen(progname); i > 0; i--) {
+		if (LIBBSD_IS_PATHNAME_SEPARATOR(progname[i - 1])) {
+			__progname = progname + i;
+			return;
+		}
+	}
+	__progname = progname;
+}
+
+
+static inline size_t
+spt_min(size_t a, size_t b)
+{
+	return ((a < b) ? a : b);
+}
+
+/*
+ * For discussion on the portability of the various methods, see
+ * https://lists.freebsd.org/pipermail/freebsd-stable/2008-June/043136.html
+ */
+static int
+spt_clearenv(void)
+{
+	char **tmp;
+
+	tmp = malloc(sizeof (*tmp));
+	if (tmp == NULL)
+		return (errno);
+
+	tmp[0] = NULL;
+	environ = tmp;
+
+	return (0);
+}
+
+static int
+spt_copyenv(int envc, char *envp[])
+{
+	char **envcopy;
+	char *eq;
+	int envsize;
+	int i, error;
+
+	if (environ != envp)
+		return (0);
+
+	/*
+	 * Make a copy of the old environ array of pointers, in case
+	 * clearenv() or setenv() is implemented to free the internal
+	 * environ array, because we will need to access the old environ
+	 * contents to make the new copy.
+	 */
+	envsize = (envc + 1) * sizeof (char *);
+	envcopy = malloc(envsize);
+	if (envcopy == NULL)
+		return (errno);
+	memcpy(envcopy, envp, envsize);
+
+	error = spt_clearenv();
+	if (error) {
+		environ = envp;
+		free(envcopy);
+		return (error);
+	}
+
+	for (i = 0; envcopy[i]; i++) {
+		eq = strchr(envcopy[i], '=');
+		if (eq == NULL)
+			continue;
+
+		*eq = '\0';
+		if (setenv(envcopy[i], eq + 1, 1) < 0)
+			error = errno;
+		*eq = '=';
+
+		if (error) {
+			environ = envp;
+			free(envcopy);
+			return (error);
+		}
+	}
+
+	/*
+	 * Dispose of the shallow copy, now that we've finished transfering
+	 * the old environment.
+	 */
+	free(envcopy);
+
+	return (0);
+}
+
+static int
+spt_copyargs(int argc, char *argv[])
+{
+	char *tmp;
+	int i;
+
+	for (i = 1; i < argc || (i >= argc && argv[i]); i++) {
+		if (argv[i] == NULL)
+			continue;
+
+		tmp = strdup(argv[i]);
+		if (tmp == NULL)
+			return (errno);
+
+		argv[i] = tmp;
+	}
+
+	return (0);
+}
+
+void
+zfs_setproctitle_init(int argc, char *argv[], char *envp[])
+{
+	char *base, *end, *nul, *tmp;
+	int i, envc, error;
+
+	/* Try to make sure we got called with main() arguments. */
+	if (argc < 0)
+		return;
+
+	base = argv[0];
+	if (base == NULL)
+		return;
+
+	nul = base + strlen(base);
+	end = nul + 1;
+
+	for (i = 0; i < argc || (i >= argc && argv[i]); i++) {
+		if (argv[i] == NULL || argv[i] != end)
+			continue;
+
+		end = argv[i] + strlen(argv[i]) + 1;
+	}
+
+	for (i = 0; envp[i]; i++) {
+		if (envp[i] != end)
+			continue;
+
+		end = envp[i] + strlen(envp[i]) + 1;
+	}
+	envc = i;
+
+	SPT.arg0 = strdup(argv[0]);
+	if (SPT.arg0 == NULL) {
+		SPT.error = errno;
+		return;
+	}
+
+	tmp = strdup(getprogname());
+	if (tmp == NULL) {
+		SPT.error = errno;
+		return;
+	}
+	setprogname(tmp);
+
+	error = spt_copyenv(envc, envp);
+	if (error) {
+		SPT.error = error;
+		return;
+	}
+
+	error = spt_copyargs(argc, argv);
+	if (error) {
+		SPT.error = error;
+		return;
+	}
+
+	SPT.nul  = nul;
+	SPT.base = base;
+	SPT.end  = end;
+}
+
+void
+zfs_setproctitle(const char *fmt, ...)
+{
+	/* Use buffer in case argv[0] is passed. */
+	char buf[SPT_MAXTITLE + 1];
+	va_list ap;
+	char *nul;
+	int len;
+	if (SPT.base == NULL) {
+		if (!SPT.warned) {
+			warnx("setproctitle not initialized, please"
+			    "call zfs_setproctitle_init()");
+			SPT.warned = B_TRUE;
+		}
+		return;
+	}
+
+	if (fmt) {
+		if (fmt[0] == '-') {
+			/* Skip program name prefix. */
+			fmt++;
+			len = 0;
+		} else {
+			/* Print program name heading for grep. */
+			snprintf(buf, sizeof (buf), "%s: ", getprogname());
+			len = strlen(buf);
+		}
+
+		va_start(ap, fmt);
+		len += vsnprintf(buf + len, sizeof (buf) - len, fmt, ap);
+		va_end(ap);
+	} else {
+		len = snprintf(buf, sizeof (buf), "%s", SPT.arg0);
+	}
+
+	if (len <= 0) {
+		SPT.error = errno;
+		return;
+	}
+
+	if (!SPT.reset) {
+		memset(SPT.base, 0, SPT.end - SPT.base);
+		SPT.reset = B_TRUE;
+	} else {
+		memset(SPT.base, 0, spt_min(sizeof (buf), SPT.end - SPT.base));
+	}
+
+	len = spt_min(len, spt_min(sizeof (buf), SPT.end - SPT.base) - 1);
+	memcpy(SPT.base, buf, len);
+	nul = SPT.base + len;
+
+	if (nul < SPT.nul) {
+		*SPT.nul = '.';
+	} else if (nul == SPT.nul && nul + 1 < SPT.end) {
+		*SPT.nul = ' ';
+		*++nul = '\0';
+	}
+}

man/man4/zfs.4

@@ -15,7 +15,7 @@
 .\" own identifying information:
 .\" Portions Copyright [yyyy] [name of copyright owner]
 .\"
-.Dd June 1, 2021
+.Dd January 10, 2023
 .Dt ZFS 4
 .Os
 .
@@ -233,6 +233,12 @@ relative to the pool.
 Make some blocks above a certain size be gang blocks.
 This option is used by the test suite to facilitate testing.
 .
+.It Sy zfs_default_bs Ns = Ns Sy 9 Po 512 B Pc Pq int
+Default dnode block size as a power of 2.
+.
+.It Sy zfs_default_ibs Ns = Ns Sy 17 Po 128 KiB Pc Pq int
+Default dnode indirect block size as a power of 2.
+.
 .It Sy zfs_history_output_max Ns = Ns Sy 1048576 Ns B Ns B Po 1MB Pc Pq int
 When attempting to log an output nvlist of an ioctl in the on-disk history,
 the output will not be stored if it is larger than this size (in bytes).
@@ -2126,6 +2132,13 @@ On very fragmented pools, lowering this
 .Pq typically to Sy 36kB
 can improve performance.
 .
+.It Sy zil_min_commit_timeout Ns = Ns Sy 5000 Pq u64
+This sets the minimum delay in nanoseconds ZIL care to delay block commit,
+waiting for more records.
+If ZIL writes are too fast, kernel may not be able sleep for so short interval,
+increasing log latency above allowed by
+.Sy zfs_commit_timeout_pct .
+.
 .It Sy zil_nocacheflush Ns = Ns Sy 0 Ns | Ns 1 Pq int
 Disable the cache flush commands that are normally sent to disk by
 the ZIL after an LWB write has completed.

man/man7/dracut.zfs.7

@@ -1,6 +1,6 @@
 .\" SPDX-License-Identifier: 0BSD
 .\"
-.Dd April 4, 2022
+.Dd March 28, 2023
 .Dt DRACUT.ZFS 7
 .Os
 .
@@ -28,13 +28,13 @@ zfs-import-scan.service \(da                       \(da           | zfs-import-c
                  zfs-import.target \(-> dracut-pre-mount.service
                         |          \(ua            |
                         | dracut-zfs-generator  |
-                        |  ____________________/|
+                        | _____________________/|
                         |/                      \(da
-                        |                   sysroot.mount \(<-\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em dracut-zfs-generator
+                        |                   sysroot.mount \(<-\(em\(em\(em dracut-zfs-generator
-                        |                       |                                        \(da   |
+                        |                       |
-                        |                       \(da            sysroot-{usr,etc,lib,&c.}.mount |
+                        |                       \(da
-                        |             initrd-root-fs.target \(<-\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em or                 \(da
+                        |             initrd-root-fs.target \(<-\(em zfs-nonroot-necessities.service
-                        |                       |              zfs-nonroot-necessities.service
+                        |                       |                                 |
                         |                       \(da                                 |
                         \(da             dracut-mount.service                        |
        zfs-snapshot-bootfs.service              |                                 |
@@ -42,7 +42,7 @@ zfs-import-scan.service \(da                       \(da           | zfs-import-c
                         \(da                       …                                 |
        zfs-rollback-bootfs.service              |                                 |
                         |                       \(da                                 |
-                        |               sysroot-usr.mount \(<-\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em/
+                        |          /sysroot/{usr,etc,lib,&c.} \(<-\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em/
                         |                       |
                         |                       \(da
                         |                initrd-fs.target

man/man8/zfs-send.8

@@ -29,7 +29,7 @@
 .\" Copyright 2018 Nexenta Systems, Inc.
 .\" Copyright 2019 Joyent, Inc.
 .\"
-.Dd April 15, 2021
+.Dd January 12, 2023
 .Dt ZFS-SEND 8
 .Os
 .
@@ -39,28 +39,28 @@
 .Sh SYNOPSIS
 .Nm zfs
 .Cm send
-.Op Fl DLPRbcehnpsvw
+.Op Fl DLPVRbcehnpsvw
 .Op Oo Fl I Ns | Ns Fl i Oc Ar snapshot
 .Ar snapshot
 .Nm zfs
 .Cm send
-.Op Fl DLPcensvw
+.Op Fl DLPVcensvw
 .Op Fl i Ar snapshot Ns | Ns Ar bookmark
 .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot
 .Nm zfs
 .Cm send
 .Fl -redact Ar redaction_bookmark
-.Op Fl DLPcenpv
+.Op Fl DLPVcenpv
 .Op Fl i Ar snapshot Ns | Ns Ar bookmark
 .Ar snapshot
 .Nm zfs
 .Cm send
-.Op Fl Penv
+.Op Fl PVenv
 .Fl t
 .Ar receive_resume_token
 .Nm zfs
 .Cm send
-.Op Fl Pnv
+.Op Fl PVnv
 .Fl S Ar filesystem
 .Nm zfs
 .Cm redact
@@ -72,7 +72,7 @@
 .It Xo
 .Nm zfs
 .Cm send
-.Op Fl DLPRbcehnpvw
+.Op Fl DLPVRbcehnpvw
 .Op Oo Fl I Ns | Ns Fl i Oc Ar snapshot
 .Ar snapshot
 .Xc
@@ -140,6 +140,8 @@ If the
 flag is used to send encrypted datasets, then
 .Fl w
 must also be specified.
+.It Fl V , -proctitle
+Set the process title to a per-second report of how much data has been sent.
 .It Fl e , -embed
 Generate a more compact stream by using
 .Sy WRITE_EMBEDDED
@@ -285,7 +287,7 @@ You will be able to receive your streams on future versions of ZFS.
 .It Xo
 .Nm zfs
 .Cm send
-.Op Fl DLPcenvw
+.Op Fl DLPVcenvw
 .Op Fl i Ar snapshot Ns | Ns Ar bookmark
 .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot
 .Xc
@@ -417,7 +419,7 @@ This information includes a per-second report of how much data has been sent.
 .Nm zfs
 .Cm send
 .Fl -redact Ar redaction_bookmark
-.Op Fl DLPcenpv
+.Op Fl DLPVcenpv
 .Op Fl i Ar snapshot Ns | Ns Ar bookmark
 .Ar snapshot
 .Xc
@@ -511,7 +513,7 @@ raw sends and redacted sends cannot be combined at this time.
 .It Xo
 .Nm zfs
 .Cm send
-.Op Fl Penv
+.Op Fl PVenv
 .Fl t
 .Ar receive_resume_token
 .Xc
@@ -526,7 +528,7 @@ for more details.
 .It Xo
 .Nm zfs
 .Cm send
-.Op Fl Pnv
+.Op Fl PVnv
 .Op Fl i Ar snapshot Ns | Ns Ar bookmark
 .Fl S
 .Ar filesystem

man/man8/zfs.8

@@ -704,6 +704,14 @@ command will be undone if the share is ever unshared (like via a reboot).
 .El
 .
 .Sh ENVIRONMENT VARIABLES
+.Bl -tag -width "ZFS_COLOR"
+.It Sy ZFS_COLOR
+Use ANSI color in
+.Nm zfs Cm diff
+and
+.Nm zfs Cm list
+output.
+.El
 .Bl -tag -width "ZFS_MOUNT_HELPER"
 .It Sy ZFS_MOUNT_HELPER
 Cause
@@ -713,6 +721,14 @@ to use
 to mount ZFS datasets.
 This option is provided for backwards compatibility with older ZFS versions.
 .El
+.Bl -tag -width "ZFS_SET_PIPE_MAX"
+.It Sy ZFS_SET_PIPE_MAX
+Tells
+.Nm zfs
+to set the maximum pipe size for sends/recieves.
+Disabled by default on Linux
+due to an unfixed deadlock in Linux's pipe size handling code.
+.El
 .
 .Sh INTERFACE STABILITY
 .Sy Committed .

man/man8/zpool.8

@@ -433,6 +433,8 @@ to dump core on exit for the purposes of running
 .It Sy ZFS_COLOR
 Use ANSI color in
 .Nm zpool status
+and
+.Nm zpool iostat
 output.
 .It Sy ZPOOL_IMPORT_PATH
 The search path for devices or files to use with the pool.

module/os/freebsd/spl/spl_kstat.c

@@ -39,6 +39,7 @@ __FBSDID("$FreeBSD$");
 #include <sys/sysctl.h>
 #include <sys/kstat.h>
 #include <sys/sbuf.h>
+#include <sys/zone.h>
 
 static MALLOC_DEFINE(M_KSTAT, "kstat_data", "Kernel statistics");
 
@@ -134,6 +135,55 @@ kstat_sysctl_string(SYSCTL_HANDLER_ARGS)
 	return (sysctl_handle_string(oidp, val, len, req));
 }
 
+static int
+kstat_sysctl_dataset(SYSCTL_HANDLER_ARGS)
+{
+	kstat_t *ksp = arg1;
+	kstat_named_t *ksent;
+	kstat_named_t *ksent_ds;
+	uint64_t val;
+	char *ds_name;
+	uint32_t ds_len = 0;
+
+	ksent_ds = ksent = ksp->ks_data;
+	ds_name = KSTAT_NAMED_STR_PTR(ksent_ds);
+	ds_len = KSTAT_NAMED_STR_BUFLEN(ksent_ds);
+	ds_name[ds_len-1] = '\0';
+
+	if (!zone_dataset_visible(ds_name, NULL)) {
+		return (EPERM);
+	}
+
+	/* Select the correct element */
+	ksent += arg2;
+	/* Update the aggsums before reading */
+	(void) ksp->ks_update(ksp, KSTAT_READ);
+	val = ksent->value.ui64;
+
+	return (sysctl_handle_64(oidp, &val, 0, req));
+}
+
+static int
+kstat_sysctl_dataset_string(SYSCTL_HANDLER_ARGS)
+{
+	kstat_t *ksp = arg1;
+	kstat_named_t *ksent = ksp->ks_data;
+	char *val;
+	uint32_t len = 0;
+
+	/* Select the correct element */
+	ksent += arg2;
+	val = KSTAT_NAMED_STR_PTR(ksent);
+	len = KSTAT_NAMED_STR_BUFLEN(ksent);
+	val[len-1] = '\0';
+
+	if (!zone_dataset_visible(val, NULL)) {
+		return (EPERM);
+	}
+
+	return (sysctl_handle_string(oidp, val, len, req));
+}
+
 static int
 kstat_sysctl_io(SYSCTL_HANDLER_ARGS)
 {
@@ -422,11 +472,20 @@ kstat_install_named(kstat_t *ksp)
 			    ksp, i, kstat_sysctl, "Q", namelast);
 			break;
 		case KSTAT_DATA_UINT64:
-			SYSCTL_ADD_PROC(&ksp->ks_sysctl_ctx,
+			if (strcmp(ksp->ks_class, "dataset") == 0) {
-			    SYSCTL_CHILDREN(ksp->ks_sysctl_root),
+				SYSCTL_ADD_PROC(&ksp->ks_sysctl_ctx,
-			    OID_AUTO, namelast,
+				    SYSCTL_CHILDREN(ksp->ks_sysctl_root),
-			    CTLTYPE_U64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
+				    OID_AUTO, namelast,
-			    ksp, i, kstat_sysctl, "QU", namelast);
+				    CTLTYPE_U64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
+				    ksp, i, kstat_sysctl_dataset, "QU",
+				    namelast);
+			} else {
+				SYSCTL_ADD_PROC(&ksp->ks_sysctl_ctx,
+				    SYSCTL_CHILDREN(ksp->ks_sysctl_root),
+				    OID_AUTO, namelast,
+				    CTLTYPE_U64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
+				    ksp, i, kstat_sysctl, "QU", namelast);
+			}
 			break;
 		case KSTAT_DATA_LONG:
 			SYSCTL_ADD_PROC(&ksp->ks_sysctl_ctx,
@@ -443,11 +502,21 @@ kstat_install_named(kstat_t *ksp)
 			    ksp, i, kstat_sysctl, "LU", namelast);
 			break;
 		case KSTAT_DATA_STRING:
-			SYSCTL_ADD_PROC(&ksp->ks_sysctl_ctx,
+			if (strcmp(ksp->ks_class, "dataset") == 0) {
-			    SYSCTL_CHILDREN(ksp->ks_sysctl_root),
+				SYSCTL_ADD_PROC(&ksp->ks_sysctl_ctx,
-			    OID_AUTO, namelast,
+				    SYSCTL_CHILDREN(ksp->ks_sysctl_root),
-			    CTLTYPE_STRING | CTLFLAG_RD | CTLFLAG_MPSAFE,
+				    OID_AUTO, namelast, CTLTYPE_STRING |
-			    ksp, i, kstat_sysctl_string, "A", namelast);
+				    CTLFLAG_RD | CTLFLAG_MPSAFE,
+				    ksp, i, kstat_sysctl_dataset_string, "A",
+				    namelast);
+			} else {
+				SYSCTL_ADD_PROC(&ksp->ks_sysctl_ctx,
+				    SYSCTL_CHILDREN(ksp->ks_sysctl_root),
+				    OID_AUTO, namelast, CTLTYPE_STRING |
+				    CTLFLAG_RD | CTLFLAG_MPSAFE,
+				    ksp, i, kstat_sysctl_string, "A",
+				    namelast);
+			}
 			break;
 		default:
 			panic("unsupported type: %d", typelast);

module/os/freebsd/spl/spl_vfs.c

@@ -125,7 +125,6 @@ mount_snapshot(kthread_t *td, vnode_t **vpp, const char *fstype, char *fspath,
 	struct vfsconf *vfsp;
 	struct mount *mp;
 	vnode_t *vp, *mvp;
-	struct ucred *cr;
 	int error;
 
 	ASSERT_VOP_ELOCKED(*vpp, "mount_snapshot");
@@ -194,15 +193,8 @@ mount_snapshot(kthread_t *td, vnode_t **vpp, const char *fstype, char *fspath,
 	 * mount(8) and df(1) output.
 	 */
 	mp->mnt_flag |= MNT_IGNORE;
-	/*
-	 * XXX: This is evil, but we can't mount a snapshot as a regular user.
-	 * XXX: Is is safe when snapshot is mounted from within a jail?
-	 */
-	cr = td->td_ucred;
-	td->td_ucred = kcred;
-	error = VFS_MOUNT(mp);
-	td->td_ucred = cr;
 
+	error = VFS_MOUNT(mp);
 	if (error != 0) {
 		/*
 		 * Clear VI_MOUNT and decrement the use count "atomically",

module/os/freebsd/zfs/arc_os.c

@@ -160,8 +160,6 @@ arc_prune_task(void *arg)
 {
 	int64_t nr_scan = (intptr_t)arg;
 
-	arc_reduce_target_size(ptob(nr_scan));
-
 #ifndef __ILP32__
 	if (nr_scan > INT_MAX)
 		nr_scan = INT_MAX;

module/os/freebsd/zfs/kmod_core.c

@@ -132,16 +132,13 @@ zfsdev_ioctl(struct cdev *dev, ulong_t zcmd, caddr_t arg, int flag,
 	len = IOCPARM_LEN(zcmd);
 	vecnum = zcmd & 0xff;
 	zp = (void *)arg;
-	uaddr = (void *)zp->zfs_cmd;
 	error = 0;
 	zcl = NULL;
 
-	if (len != sizeof (zfs_iocparm_t)) {
+	if (len != sizeof (zfs_iocparm_t))
-		printf("len %d vecnum: %d sizeof (zfs_cmd_t) %ju\n",
-		    len, vecnum, (uintmax_t)sizeof (zfs_cmd_t));
 		return (EINVAL);
-	}
 
+	uaddr = (void *)zp->zfs_cmd;
 	zc = kmem_zalloc(sizeof (zfs_cmd_t), KM_SLEEP);
 	/*
 	 * Remap ioctl code for legacy user binaries

module/os/freebsd/zfs/spa_os.c

@@ -95,6 +95,8 @@ spa_generate_rootconf(const char *name)
 	for (i = 0; i < count; i++) {
 		uint64_t txg;
 
+		if (configs[i] == NULL)
+			continue;
 		txg = fnvlist_lookup_uint64(configs[i], ZPOOL_CONFIG_POOL_TXG);
 		if (txg > best_txg) {
 			best_txg = txg;

module/os/freebsd/zfs/sysctl_os.c

@@ -356,14 +356,6 @@ SYSCTL_UINT(_vfs_zfs_zfetch, OID_AUTO, max_idistance, CTLFLAG_RWTUN,
 /* dsl_pool.c */
 
 /* dnode.c */
-extern int zfs_default_bs;
-SYSCTL_INT(_vfs_zfs, OID_AUTO, default_bs, CTLFLAG_RWTUN,
-    &zfs_default_bs, 0, "Default dnode block shift");
-
-extern int zfs_default_ibs;
-SYSCTL_INT(_vfs_zfs, OID_AUTO, default_ibs, CTLFLAG_RWTUN,
-    &zfs_default_ibs, 0, "Default dnode indirect block shift");
-
 
 /* dsl_scan.c */
 

module/os/freebsd/zfs/zfs_vfsops.c

@@ -64,6 +64,7 @@
 #include <sys/dsl_dir.h>
 #include <sys/spa_boot.h>
 #include <sys/jail.h>
+#include <sys/osd.h>
 #include <ufs/ufs/quota.h>
 #include <sys/zfs_quota.h>
 
@@ -90,6 +91,20 @@ int zfs_debug_level;
 SYSCTL_INT(_vfs_zfs, OID_AUTO, debug, CTLFLAG_RWTUN, &zfs_debug_level, 0,
 	"Debug level");
 
+struct zfs_jailparam {
+	int mount_snapshot;
+};
+
+static struct zfs_jailparam zfs_jailparam0 = {
+	.mount_snapshot = 0,
+};
+
+static int zfs_jailparam_slot;
+
+SYSCTL_JAIL_PARAM_SYS_NODE(zfs, CTLFLAG_RW, "Jail ZFS parameters");
+SYSCTL_JAIL_PARAM(_zfs, mount_snapshot, CTLTYPE_INT | CTLFLAG_RW, "I",
+	"Allow mounting snapshots in the .zfs directory for unjailed datasets");
+
 SYSCTL_NODE(_vfs_zfs, OID_AUTO, version, CTLFLAG_RD, 0, "ZFS versions");
 static int zfs_version_acl = ZFS_ACL_VERSION;
 SYSCTL_INT(_vfs_zfs_version, OID_AUTO, acl, CTLFLAG_RD, &zfs_version_acl, 0,
@@ -720,7 +735,7 @@ zfs_register_callbacks(vfs_t *vfsp)
 		nbmand = B_FALSE;
 	} else if (vfs_optionisset(vfsp, MNTOPT_NBMAND, NULL)) {
 		nbmand = B_TRUE;
-	} else if ((error = dsl_prop_get_int_ds(ds, "nbmand", &nbmand) != 0)) {
+	} else if ((error = dsl_prop_get_int_ds(ds, "nbmand", &nbmand)) != 0) {
 		dsl_pool_config_exit(dmu_objset_pool(os), FTAG);
 		return (error);
 	}
@@ -1332,7 +1347,7 @@ zfs_mount(vfs_t *vfsp)
 	char		*osname;
 	int		error = 0;
 	int		canwrite;
-	bool		checkpointrewind;
+	bool		checkpointrewind, isctlsnap = false;
 
 	if (vfs_getopt(vfsp->mnt_optnew, "from", (void **)&osname, NULL))
 		return (SET_ERROR(EINVAL));
@@ -1347,6 +1362,8 @@ zfs_mount(vfs_t *vfsp)
 	}
 
 	fetch_osname_options(osname, &checkpointrewind);
+	isctlsnap = (mvp != NULL && zfsctl_is_node(mvp) &&
+	    strchr(osname, '@') != NULL);
 
 	/*
 	 * Check for mount privilege?
@@ -1355,7 +1372,9 @@ zfs_mount(vfs_t *vfsp)
 	 * we have local permission to allow it
 	 */
 	error = secpolicy_fs_mount(cr, mvp, vfsp);
-	if (error) {
+	if (error && isctlsnap) {
+		secpolicy_fs_mount_clearopts(cr, vfsp);
+	} else if (error) {
 		if (dsl_deleg_access(osname, ZFS_DELEG_PERM_MOUNT, cr) != 0)
 			goto out;
 
@@ -1392,8 +1411,27 @@ zfs_mount(vfs_t *vfsp)
 	 */
 	if (!INGLOBALZONE(curproc) &&
 	    (!zone_dataset_visible(osname, &canwrite) || !canwrite)) {
-		error = SET_ERROR(EPERM);
+		boolean_t mount_snapshot = B_FALSE;
-		goto out;
+
+		/*
+		 * Snapshots may be mounted in .zfs for unjailed datasets
+		 * if allowed by the jail param zfs.mount_snapshot.
+		 */
+		if (isctlsnap) {
+			struct prison *pr;
+			struct zfs_jailparam *zjp;
+
+			pr = curthread->td_ucred->cr_prison;
+			mtx_lock(&pr->pr_mtx);
+			zjp = osd_jail_get(pr, zfs_jailparam_slot);
+			mtx_unlock(&pr->pr_mtx);
+			if (zjp && zjp->mount_snapshot)
+				mount_snapshot = B_TRUE;
+		}
+		if (!mount_snapshot) {
+			error = SET_ERROR(EPERM);
+			goto out;
+		}
 	}
 
 	vfsp->vfs_flag |= MNT_NFS4ACLS;
@@ -2343,3 +2381,236 @@ zfsvfs_update_fromname(const char *oldname, const char *newname)
 	mtx_unlock(&mountlist_mtx);
 }
 #endif
+
+/*
+ * Find a prison with ZFS info.
+ * Return the ZFS info and the (locked) prison.
+ */
+static struct zfs_jailparam *
+zfs_jailparam_find(struct prison *spr, struct prison **prp)
+{
+	struct prison *pr;
+	struct zfs_jailparam *zjp;
+
+	for (pr = spr; ; pr = pr->pr_parent) {
+		mtx_lock(&pr->pr_mtx);
+		if (pr == &prison0) {
+			zjp = &zfs_jailparam0;
+			break;
+		}
+		zjp = osd_jail_get(pr, zfs_jailparam_slot);
+		if (zjp != NULL)
+			break;
+		mtx_unlock(&pr->pr_mtx);
+	}
+	*prp = pr;
+
+	return (zjp);
+}
+
+/*
+ * Ensure a prison has its own ZFS info.  If zjpp is non-null, point it to the
+ * ZFS info and lock the prison.
+ */
+static void
+zfs_jailparam_alloc(struct prison *pr, struct zfs_jailparam **zjpp)
+{
+	struct prison *ppr;
+	struct zfs_jailparam *zjp, *nzjp;
+	void **rsv;
+
+	/* If this prison already has ZFS info, return that. */
+	zjp = zfs_jailparam_find(pr, &ppr);
+	if (ppr == pr)
+		goto done;
+
+	/*
+	 * Allocate a new info record.  Then check again, in case something
+	 * changed during the allocation.
+	 */
+	mtx_unlock(&ppr->pr_mtx);
+	nzjp = malloc(sizeof (struct zfs_jailparam), M_PRISON, M_WAITOK);
+	rsv = osd_reserve(zfs_jailparam_slot);
+	zjp = zfs_jailparam_find(pr, &ppr);
+	if (ppr == pr) {
+		free(nzjp, M_PRISON);
+		osd_free_reserved(rsv);
+		goto done;
+	}
+	/* Inherit the initial values from the ancestor. */
+	mtx_lock(&pr->pr_mtx);
+	(void) osd_jail_set_reserved(pr, zfs_jailparam_slot, rsv, nzjp);
+	(void) memcpy(nzjp, zjp, sizeof (*zjp));
+	zjp = nzjp;
+	mtx_unlock(&ppr->pr_mtx);
+done:
+	if (zjpp != NULL)
+		*zjpp = zjp;
+	else
+		mtx_unlock(&pr->pr_mtx);
+}
+
+/*
+ * Jail OSD methods for ZFS VFS info.
+ */
+static int
+zfs_jailparam_create(void *obj, void *data)
+{
+	struct prison *pr = obj;
+	struct vfsoptlist *opts = data;
+	int jsys;
+
+	if (vfs_copyopt(opts, "zfs", &jsys, sizeof (jsys)) == 0 &&
+	    jsys == JAIL_SYS_INHERIT)
+		return (0);
+	/*
+	 * Inherit a prison's initial values from its parent
+	 * (different from JAIL_SYS_INHERIT which also inherits changes).
+	 */
+	zfs_jailparam_alloc(pr, NULL);
+	return (0);
+}
+
+static int
+zfs_jailparam_get(void *obj, void *data)
+{
+	struct prison *ppr, *pr = obj;
+	struct vfsoptlist *opts = data;
+	struct zfs_jailparam *zjp;
+	int jsys, error;
+
+	zjp = zfs_jailparam_find(pr, &ppr);
+	jsys = (ppr == pr) ? JAIL_SYS_NEW : JAIL_SYS_INHERIT;
+	error = vfs_setopt(opts, "zfs", &jsys, sizeof (jsys));
+	if (error != 0 && error != ENOENT)
+		goto done;
+	if (jsys == JAIL_SYS_NEW) {
+		error = vfs_setopt(opts, "zfs.mount_snapshot",
+		    &zjp->mount_snapshot, sizeof (zjp->mount_snapshot));
+		if (error != 0 && error != ENOENT)
+			goto done;
+	} else {
+		/*
+		 * If this prison is inheriting its ZFS info, report
+		 * empty/zero parameters.
+		 */
+		static int mount_snapshot = 0;
+
+		error = vfs_setopt(opts, "zfs.mount_snapshot",
+		    &mount_snapshot, sizeof (mount_snapshot));
+		if (error != 0 && error != ENOENT)
+			goto done;
+	}
+	error = 0;
+done:
+	mtx_unlock(&ppr->pr_mtx);
+	return (error);
+}
+
+static int
+zfs_jailparam_set(void *obj, void *data)
+{
+	struct prison *pr = obj;
+	struct prison *ppr;
+	struct vfsoptlist *opts = data;
+	int error, jsys, mount_snapshot;
+
+	/* Set the parameters, which should be correct. */
+	error = vfs_copyopt(opts, "zfs", &jsys, sizeof (jsys));
+	if (error == ENOENT)
+		jsys = -1;
+	error = vfs_copyopt(opts, "zfs.mount_snapshot", &mount_snapshot,
+	    sizeof (mount_snapshot));
+	if (error == ENOENT)
+		mount_snapshot = -1;
+	else
+		jsys = JAIL_SYS_NEW;
+	if (jsys == JAIL_SYS_NEW) {
+		/* "zfs=new" or "zfs.*": the prison gets its own ZFS info. */
+		struct zfs_jailparam *zjp;
+
+		/*
+		 * A child jail cannot have more permissions than its parent
+		 */
+		if (pr->pr_parent != &prison0) {
+			zjp = zfs_jailparam_find(pr->pr_parent, &ppr);
+			mtx_unlock(&ppr->pr_mtx);
+			if (zjp->mount_snapshot < mount_snapshot) {
+				return (EPERM);
+			}
+		}
+		zfs_jailparam_alloc(pr, &zjp);
+		if (mount_snapshot != -1)
+			zjp->mount_snapshot = mount_snapshot;
+		mtx_unlock(&pr->pr_mtx);
+	} else {
+		/* "zfs=inherit": inherit the parent's ZFS info. */
+		mtx_lock(&pr->pr_mtx);
+		osd_jail_del(pr, zfs_jailparam_slot);
+		mtx_unlock(&pr->pr_mtx);
+	}
+	return (0);
+}
+
+static int
+zfs_jailparam_check(void *obj __unused, void *data)
+{
+	struct vfsoptlist *opts = data;
+	int error, jsys, mount_snapshot;
+
+	/* Check that the parameters are correct. */
+	error = vfs_copyopt(opts, "zfs", &jsys, sizeof (jsys));
+	if (error != ENOENT) {
+		if (error != 0)
+			return (error);
+		if (jsys != JAIL_SYS_NEW && jsys != JAIL_SYS_INHERIT)
+			return (EINVAL);
+	}
+	error = vfs_copyopt(opts, "zfs.mount_snapshot", &mount_snapshot,
+	    sizeof (mount_snapshot));
+	if (error != ENOENT) {
+		if (error != 0)
+			return (error);
+		if (mount_snapshot != 0 && mount_snapshot != 1)
+			return (EINVAL);
+	}
+	return (0);
+}
+
+static void
+zfs_jailparam_destroy(void *data)
+{
+
+	free(data, M_PRISON);
+}
+
+static void
+zfs_jailparam_sysinit(void *arg __unused)
+{
+	struct prison *pr;
+	osd_method_t  methods[PR_MAXMETHOD] = {
+		[PR_METHOD_CREATE] = zfs_jailparam_create,
+		[PR_METHOD_GET] = zfs_jailparam_get,
+		[PR_METHOD_SET] = zfs_jailparam_set,
+		[PR_METHOD_CHECK] = zfs_jailparam_check,
+	};
+
+	zfs_jailparam_slot = osd_jail_register(zfs_jailparam_destroy, methods);
+	/* Copy the defaults to any existing prisons. */
+	sx_slock(&allprison_lock);
+	TAILQ_FOREACH(pr, &allprison, pr_list)
+		zfs_jailparam_alloc(pr, NULL);
+	sx_sunlock(&allprison_lock);
+}
+
+static void
+zfs_jailparam_sysuninit(void *arg __unused)
+{
+
+	osd_jail_deregister(zfs_jailparam_slot);
+}
+
+SYSINIT(zfs_jailparam_sysinit, SI_SUB_DRIVERS, SI_ORDER_ANY,
+	zfs_jailparam_sysinit, NULL);
+SYSUNINIT(zfs_jailparam_sysuninit, SI_SUB_DRIVERS, SI_ORDER_ANY,
+	zfs_jailparam_sysuninit, NULL);

module/os/freebsd/zfs/zfs_znode.c

@@ -536,6 +536,9 @@ zfs_znode_alloc(zfsvfs_t *zfsvfs, dmu_buf_t *db, int blksz,
 	 * Acquire vnode lock before making it available to the world.
 	 */
 	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
+#if __FreeBSD_version >= 1400077
+	vn_set_state(vp, VSTATE_CONSTRUCTED);
+#endif
 	VN_LOCK_AREC(vp);
 	if (vp->v_type != VFIFO)
 		VN_LOCK_ASHARE(vp);

module/os/freebsd/zfs/zvol_os.c

@@ -1161,7 +1161,10 @@ zvol_cdev_ioctl(struct cdev *dev, ulong_t cmd, caddr_t data,
 
 		hole = (cmd == FIOSEEKHOLE);
 		noff = *off;
+		lr = zfs_rangelock_enter(&zv->zv_rangelock, 0, UINT64_MAX,
+		    RL_READER);
 		error = dmu_offset_next(zv->zv_objset, ZVOL_OBJ, hole, &noff);
+		zfs_rangelock_exit(lr);
 		*off = noff;
 		break;
 	}

module/os/linux/spl/spl-generic.c

@@ -48,6 +48,7 @@
 #include <linux/mod_compat.h>
 #include <sys/cred.h>
 #include <sys/vnode.h>
+#include <sys/misc.h>
 
 char spl_gitrev[64] = ZFS_META_GITREV;
 
@@ -540,6 +541,38 @@ ddi_copyin(const void *from, void *to, size_t len, int flags)
 }
 EXPORT_SYMBOL(ddi_copyin);
 
+/*
+ * Post a uevent to userspace whenever a new vdev adds to the pool. It is
+ * necessary to sync blkid information with udev, which zed daemon uses
+ * during device hotplug to identify the vdev.
+ */
+void
+spl_signal_kobj_evt(struct block_device *bdev)
+{
+#if defined(HAVE_BDEV_KOBJ) || defined(HAVE_PART_TO_DEV)
+#ifdef HAVE_BDEV_KOBJ
+	struct kobject *disk_kobj = bdev_kobj(bdev);
+#else
+	struct kobject *disk_kobj = &part_to_dev(bdev->bd_part)->kobj;
+#endif
+	if (disk_kobj) {
+		int ret = kobject_uevent(disk_kobj, KOBJ_CHANGE);
+		if (ret) {
+			pr_warn("ZFS: Sending event '%d' to kobject: '%s'"
+			    " (%p): failed(ret:%d)\n", KOBJ_CHANGE,
+			    kobject_name(disk_kobj), disk_kobj, ret);
+		}
+	}
+#else
+/*
+ * This is encountered if neither bdev_kobj() nor part_to_dev() is available
+ * in the kernel - likely due to an API change that needs to be chased down.
+ */
+#error "Unsupported kernel: unable to get struct kobj from bdev"
+#endif
+}
+EXPORT_SYMBOL(spl_signal_kobj_evt);
+
 int
 ddi_copyout(const void *from, void *to, size_t len, int flags)
 {

module/os/linux/zfs/qat_compress.c

@@ -247,7 +247,7 @@ qat_compress_impl(qat_compress_dir_t dir, char *src, int src_len,
 	Cpa8U *buffer_meta_src = NULL;
 	Cpa8U *buffer_meta_dst = NULL;
 	Cpa32U buffer_meta_size = 0;
-	CpaDcRqResults dc_results;
+	CpaDcRqResults dc_results = {.checksum = 1};
 	CpaStatus status = CPA_STATUS_FAIL;
 	Cpa32U hdr_sz = 0;
 	Cpa32U compressed_sz;

module/os/linux/zfs/vdev_disk.c

@@ -179,6 +179,18 @@ vdev_disk_error(zio_t *zio)
 	    zio->io_flags);
 }
 
+static void
+vdev_disk_kobj_evt_post(vdev_t *v)
+{
+	vdev_disk_t *vd = v->vdev_tsd;
+	if (vd && vd->vd_bdev) {
+		spl_signal_kobj_evt(vd->vd_bdev);
+	} else {
+		vdev_dbgmsg(v, "vdev_disk_t is NULL for VDEV:%s\n",
+		    v->vdev_path);
+	}
+}
+
 static int
 vdev_disk_open(vdev_t *v, uint64_t *psize, uint64_t *max_psize,
     uint64_t *logical_ashift, uint64_t *physical_ashift)
@@ -290,6 +302,13 @@ vdev_disk_open(vdev_t *v, uint64_t *psize, uint64_t *max_psize,
 		bdev = blkdev_get_by_path(v->vdev_path, mode | FMODE_EXCL,
 		    zfs_vdev_holder);
 		if (unlikely(PTR_ERR(bdev) == -ENOENT)) {
+			/*
+			 * There is no point of waiting since device is removed
+			 * explicitly
+			 */
+			if (v->vdev_removed)
+				break;
+
 			schedule_timeout(MSEC_TO_TICK(10));
 		} else if (unlikely(PTR_ERR(bdev) == -ERESTARTSYS)) {
 			timeout = MSEC2NSEC(zfs_vdev_open_timeout_ms * 10);
@@ -899,7 +918,7 @@ vdev_disk_io_done(zio_t *zio)
 		vdev_t *v = zio->io_vd;
 		vdev_disk_t *vd = v->vdev_tsd;
 
-		if (zfs_check_media_change(vd->vd_bdev)) {
+		if (!zfs_check_disk_status(vd->vd_bdev)) {
 			invalidate_bdev(vd->vd_bdev);
 			v->vdev_remove_wanted = B_TRUE;
 			spa_async_request(zio->io_spa, SPA_ASYNC_REMOVE);
@@ -955,7 +974,8 @@ vdev_ops_t vdev_disk_ops = {
 	.vdev_op_nparity = NULL,
 	.vdev_op_ndisks = NULL,
 	.vdev_op_type = VDEV_TYPE_DISK,		/* name of this vdev type */
-	.vdev_op_leaf = B_TRUE			/* leaf vdev */
+	.vdev_op_leaf = B_TRUE,			/* leaf vdev */
+	.vdev_op_kobj_evt_post = vdev_disk_kobj_evt_post
 };
 
 /*

module/os/linux/zfs/zfs_dir.c

@@ -1066,9 +1066,6 @@ zfs_make_xattrdir(znode_t *zp, vattr_t *vap, znode_t **xzpp, cred_t *cr)
 
 	*xzpp = NULL;
 
-	if ((error = zfs_zaccess(zp, ACE_WRITE_NAMED_ATTRS, 0, B_FALSE, cr)))
-		return (error);
-
 	if ((error = zfs_acl_ids_create(zp, IS_XATTR, vap, cr, NULL,
 	    &acl_ids)) != 0)
 		return (error);

module/os/linux/zfs/zfs_vnops_os.c

@@ -579,6 +579,7 @@ zfs_create(znode_t *dzp, char *name, vattr_t *vap, int excl,
 	boolean_t	fuid_dirtied;
 	boolean_t	have_acl = B_FALSE;
 	boolean_t	waited = B_FALSE;
+	boolean_t	skip_acl = (flag & ATTR_NOACLCHECK) ? B_TRUE : B_FALSE;
 
 	/*
 	 * If we have an ephemeral id, ACL, or XVATTR then
@@ -651,7 +652,7 @@ top:
 		 * Create a new file object and update the directory
 		 * to reference it.
 		 */
-		if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, B_FALSE, cr))) {
+		if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, skip_acl, cr))) {
 			if (have_acl)
 				zfs_acl_ids_free(&acl_ids);
 			goto out;

module/os/linux/zfs/zfs_znode.c

@@ -162,8 +162,7 @@ zfs_znode_hold_cache_constructor(void *buf, void *arg, int kmflags)
 	znode_hold_t *zh = buf;
 
 	mutex_init(&zh->zh_lock, NULL, MUTEX_DEFAULT, NULL);
-	zfs_refcount_create(&zh->zh_refcount);
+	zh->zh_refcount = 0;
-	zh->zh_obj = ZFS_NO_OBJECT;
 
 	return (0);
 }
@@ -174,7 +173,6 @@ zfs_znode_hold_cache_destructor(void *buf, void *arg)
 	znode_hold_t *zh = buf;
 
 	mutex_destroy(&zh->zh_lock);
-	zfs_refcount_destroy(&zh->zh_refcount);
 }
 
 void
@@ -273,26 +271,26 @@ zfs_znode_hold_enter(zfsvfs_t *zfsvfs, uint64_t obj)
 	boolean_t found = B_FALSE;
 
 	zh_new = kmem_cache_alloc(znode_hold_cache, KM_SLEEP);
-	zh_new->zh_obj = obj;
 	search.zh_obj = obj;
 
 	mutex_enter(&zfsvfs->z_hold_locks[i]);
 	zh = avl_find(&zfsvfs->z_hold_trees[i], &search, NULL);
 	if (likely(zh == NULL)) {
 		zh = zh_new;
+		zh->zh_obj = obj;
 		avl_add(&zfsvfs->z_hold_trees[i], zh);
 	} else {
 		ASSERT3U(zh->zh_obj, ==, obj);
 		found = B_TRUE;
 	}
-	zfs_refcount_add(&zh->zh_refcount, NULL);
+	zh->zh_refcount++;
+	ASSERT3S(zh->zh_refcount, >, 0);
 	mutex_exit(&zfsvfs->z_hold_locks[i]);
 
 	if (found == B_TRUE)
 		kmem_cache_free(znode_hold_cache, zh_new);
 
 	ASSERT(MUTEX_NOT_HELD(&zh->zh_lock));
-	ASSERT3S(zfs_refcount_count(&zh->zh_refcount), >, 0);
 	mutex_enter(&zh->zh_lock);
 
 	return (zh);
@@ -305,11 +303,11 @@ zfs_znode_hold_exit(zfsvfs_t *zfsvfs, znode_hold_t *zh)
 	boolean_t remove = B_FALSE;
 
 	ASSERT(zfs_znode_held(zfsvfs, zh->zh_obj));
-	ASSERT3S(zfs_refcount_count(&zh->zh_refcount), >, 0);
 	mutex_exit(&zh->zh_lock);
 
 	mutex_enter(&zfsvfs->z_hold_locks[i]);
-	if (zfs_refcount_remove(&zh->zh_refcount, NULL) == 0) {
+	ASSERT3S(zh->zh_refcount, >, 0);
+	if (--zh->zh_refcount == 0) {
 		avl_remove(&zfsvfs->z_hold_trees[i], zh);
 		remove = B_TRUE;
 	}

module/os/linux/zfs/zpl_inode.c

@@ -224,12 +224,17 @@ zpl_mknod(struct inode *dir, struct dentry *dentry, umode_t mode,
 
 #ifdef HAVE_TMPFILE
 static int
+#ifndef HAVE_TMPFILE_DENTRY
+zpl_tmpfile(struct user_namespace *userns, struct inode *dir,
+    struct file *file, umode_t mode)
+#else
 #ifdef HAVE_TMPFILE_USERNS
 zpl_tmpfile(struct user_namespace *userns, struct inode *dir,
     struct dentry *dentry, umode_t mode)
 #else
 zpl_tmpfile(struct inode *dir, struct dentry *dentry, umode_t mode)
 #endif
+#endif
 {
 	cred_t *cr = CRED();
 	struct inode *ip;
@@ -252,11 +257,21 @@ zpl_tmpfile(struct inode *dir, struct dentry *dentry, umode_t mode)
 	if (error == 0) {
 		/* d_tmpfile will do drop_nlink, so we should set it first */
 		set_nlink(ip, 1);
+#ifndef HAVE_TMPFILE_DENTRY
+		d_tmpfile(file, ip);
+
+		error = zpl_xattr_security_init(ip, dir,
+		    &file->f_path.dentry->d_name);
+#else
 		d_tmpfile(dentry, ip);
 
 		error = zpl_xattr_security_init(ip, dir, &dentry->d_name);
+#endif
 		if (error == 0)
 			error = zpl_init_acl(ip, dir);
+#ifndef HAVE_TMPFILE_DENTRY
+		error = finish_open_simple(file, error);
+#endif
 		/*
 		 * don't need to handle error here, file is already in
 		 * unlinked set.
@@ -711,7 +726,11 @@ const struct inode_operations zpl_inode_operations = {
 #if defined(HAVE_SET_ACL)
 	.set_acl	= zpl_set_acl,
 #endif /* HAVE_SET_ACL */
+#if defined(HAVE_GET_INODE_ACL)
+	.get_inode_acl	= zpl_get_acl,
+#else
 	.get_acl	= zpl_get_acl,
+#endif /* HAVE_GET_INODE_ACL */
 #endif /* CONFIG_FS_POSIX_ACL */
 };
 
@@ -744,7 +763,11 @@ const struct inode_operations zpl_dir_inode_operations = {
 #if defined(HAVE_SET_ACL)
 	.set_acl	= zpl_set_acl,
 #endif /* HAVE_SET_ACL */
+#if defined(HAVE_GET_INODE_ACL)
+	.get_inode_acl	= zpl_get_acl,
+#else
 	.get_acl	= zpl_get_acl,
+#endif /* HAVE_GET_INODE_ACL */
 #endif /* CONFIG_FS_POSIX_ACL */
 };
 
@@ -783,6 +806,10 @@ const struct inode_operations zpl_special_inode_operations = {
 #if defined(HAVE_SET_ACL)
 	.set_acl	= zpl_set_acl,
 #endif /* HAVE_SET_ACL */
+#if defined(HAVE_GET_INODE_ACL)
+	.get_inode_acl	= zpl_get_acl,
+#else
 	.get_acl	= zpl_get_acl,
+#endif /* HAVE_GET_INODE_ACL */
 #endif /* CONFIG_FS_POSIX_ACL */
 };

module/os/linux/zfs/zpl_xattr.c

@@ -496,7 +496,7 @@ zpl_xattr_set_dir(struct inode *ip, const char *name, const void *value,
 		vap->va_gid = crgetgid(cr);
 
 		error = -zfs_create(dxzp, (char *)name, vap, 0, 0644, &xzp,
-		    cr, 0, NULL);
+		    cr, ATTR_NOACLCHECK, NULL);
 		if (error)
 			goto out;
 	}
@@ -1004,11 +1004,18 @@ int
 #ifdef HAVE_SET_ACL_USERNS
 zpl_set_acl(struct user_namespace *userns, struct inode *ip,
     struct posix_acl *acl, int type)
+#elif defined(HAVE_SET_ACL_USERNS_DENTRY_ARG2)
+zpl_set_acl(struct user_namespace *userns, struct dentry *dentry,
+    struct posix_acl *acl, int type)
 #else
 zpl_set_acl(struct inode *ip, struct posix_acl *acl, int type)
 #endif /* HAVE_SET_ACL_USERNS */
 {
+#ifdef HAVE_SET_ACL_USERNS_DENTRY_ARG2
+	return (zpl_set_acl_impl(d_inode(dentry), acl, type));
+#else
 	return (zpl_set_acl_impl(ip, acl, type));
+#endif /* HAVE_SET_ACL_USERNS_DENTRY_ARG2 */
 }
 #endif /* HAVE_SET_ACL */
 
@@ -1067,7 +1074,7 @@ zpl_get_acl_impl(struct inode *ip, int type)
 	return (acl);
 }
 
-#if defined(HAVE_GET_ACL_RCU)
+#if defined(HAVE_GET_ACL_RCU) || defined(HAVE_GET_INODE_ACL)
 struct posix_acl *
 zpl_get_acl(struct inode *ip, int type, bool rcu)
 {

module/unicode/u8_textprep.c

@@ -23,6 +23,9 @@
  * Use is subject to license terms.
  */
 
+/*
+ * Copyright 2022 MNX Cloud, Inc.
+ */
 
 
 
@@ -213,10 +216,10 @@ const int8_t u8_number_of_bytes[0x100] = {
 /*	80  81  82  83  84  85  86  87  88  89  8A  8B  8C  8D  8E  8F  */
 	I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_,
 
-/*  	90  91  92  93  94  95  96  97  98  99  9A  9B  9C  9D  9E  9F  */
+/*	90  91  92  93  94  95  96  97  98  99  9A  9B  9C  9D  9E  9F  */
 	I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_,
 
-/*  	A0  A1  A2  A3  A4  A5  A6  A7  A8  A9  AA  AB  AC  AD  AE  AF  */
+/*	A0  A1  A2  A3  A4  A5  A6  A7  A8  A9  AA  AB  AC  AD  AE  AF  */
 	I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_,
 
 /*	B0  B1  B2  B3  B4  B5  B6  B7  B8  B9  BA  BB  BC  BD  BE  BF  */
@@ -1286,8 +1289,12 @@ TRY_THE_NEXT_MARK:
 		saved_l = l - disp[last];
 
 		while (p < oslast) {
-			size = u8_number_of_bytes[*p];
+			int8_t number_of_bytes = u8_number_of_bytes[*p];
-			if (size <= 1 || (p + size) > oslast)
+
+			if (number_of_bytes <= 1)
+				break;
+			size = number_of_bytes;
+			if ((p + size) > oslast)
 				break;
 
 			saved_p = p;
@@ -1378,8 +1385,10 @@ SAFE_RETURN:
  */
 static size_t
 collect_a_seq(size_t uv, uchar_t *u8s, uchar_t **source, uchar_t *slast,
-    boolean_t is_it_toupper, boolean_t is_it_tolower,
+    boolean_t is_it_toupper,
-    boolean_t canonical_decomposition, boolean_t compatibility_decomposition,
+    boolean_t is_it_tolower,
+    boolean_t canonical_decomposition,
+    boolean_t compatibility_decomposition,
     boolean_t canonical_composition,
     int *errnum, u8_normalization_states_t *state)
 {

module/zfs/arc.c

@@ -5943,6 +5943,7 @@ arc_read(zio_t *pio, spa_t *spa, const blkptr_t *bp,
 	    (zio_flags & ZIO_FLAG_RAW_ENCRYPT) != 0;
 	boolean_t embedded_bp = !!BP_IS_EMBEDDED(bp);
 	boolean_t no_buf = *arc_flags & ARC_FLAG_NO_BUF;
+	arc_buf_t *buf = NULL;
 	int rc = 0;
 
 	ASSERT(!embedded_bp ||
@@ -5972,7 +5973,7 @@ top:
 	if (!zfs_blkptr_verify(spa, bp, zio_flags & ZIO_FLAG_CONFIG_WRITER,
 	    BLK_VERIFY_LOG)) {
 		rc = SET_ERROR(ECKSUM);
-		goto out;
+		goto done;
 	}
 
 	if (!embedded_bp) {
@@ -5992,7 +5993,6 @@ top:
 	 */
 	if (hdr != NULL && HDR_HAS_L1HDR(hdr) && (HDR_HAS_RABD(hdr) ||
 	    (hdr->b_l1hdr.b_pabd != NULL && !encrypted_read))) {
-		arc_buf_t *buf = NULL;
 		*arc_flags |= ARC_FLAG_CACHED;
 
 		if (HDR_IO_IN_PROGRESS(hdr)) {
@@ -6002,7 +6002,7 @@ top:
 				mutex_exit(hash_lock);
 				ARCSTAT_BUMP(arcstat_cached_only_in_progress);
 				rc = SET_ERROR(ENOENT);
-				goto out;
+				goto done;
 			}
 
 			ASSERT3P(head_zio, !=, NULL);
@@ -6130,9 +6130,7 @@ top:
 		ARCSTAT_CONDSTAT(!HDR_PREFETCH(hdr),
 		    demand, prefetch, !HDR_ISTYPE_METADATA(hdr),
 		    data, metadata, hits);
-
+		goto done;
-		if (done)
-			done(NULL, zb, bp, buf, private);
 	} else {
 		uint64_t lsize = BP_GET_LSIZE(bp);
 		uint64_t psize = BP_GET_PSIZE(bp);
@@ -6145,10 +6143,10 @@ top:
 		int alloc_flags = encrypted_read ? ARC_HDR_ALLOC_RDATA : 0;
 
 		if (*arc_flags & ARC_FLAG_CACHED_ONLY) {
-			rc = SET_ERROR(ENOENT);
 			if (hash_lock != NULL)
 				mutex_exit(hash_lock);
-			goto out;
+			rc = SET_ERROR(ENOENT);
+			goto done;
 		}
 
 		if (hdr == NULL) {
@@ -6474,6 +6472,16 @@ out:
 		spa_read_history_add(spa, zb, *arc_flags);
 	spl_fstrans_unmark(cookie);
 	return (rc);
+
+done:
+	if (done)
+		done(NULL, zb, bp, buf, private);
+	if (pio && rc != 0) {
+		zio_t *zio = zio_null(pio, spa, NULL, NULL, NULL, zio_flags);
+		zio->io_error = rc;
+		zio_nowait(zio);
+	}
+	goto out;
 }
 
 arc_prune_t *

module/zfs/bpobj.c

@@ -663,14 +663,13 @@ bpobj_enqueue_subobj(bpobj_t *bpo, uint64_t subobj, dmu_tx_t *tx)
 	}
 
 	VERIFY3U(0, ==, bpobj_open(&subbpo, bpo->bpo_os, subobj));
-	VERIFY3U(0, ==, bpobj_space(&subbpo, &used, &comp, &uncomp));
-
 	if (bpobj_is_empty(&subbpo)) {
 		/* No point in having an empty subobj. */
 		bpobj_close(&subbpo);
 		bpobj_free(bpo->bpo_os, subobj, tx);
 		return;
 	}
+	VERIFY3U(0, ==, bpobj_space(&subbpo, &used, &comp, &uncomp));
 
 	mutex_enter(&bpo->bpo_lock);
 	dmu_buf_will_dirty(bpo->bpo_dbuf, tx);
@@ -780,6 +779,68 @@ bpobj_enqueue_subobj(bpobj_t *bpo, uint64_t subobj, dmu_tx_t *tx)
 
 }
 
+/*
+ * Prefetch metadata required for bpobj_enqueue_subobj().
+ */
+void
+bpobj_prefetch_subobj(bpobj_t *bpo, uint64_t subobj)
+{
+	dmu_object_info_t doi;
+	bpobj_t subbpo;
+	uint64_t subsubobjs;
+	boolean_t copy_subsub = B_TRUE;
+	boolean_t copy_bps = B_TRUE;
+
+	ASSERT(bpobj_is_open(bpo));
+	ASSERT(subobj != 0);
+
+	if (subobj == dmu_objset_pool(bpo->bpo_os)->dp_empty_bpobj)
+		return;
+
+	if (bpobj_open(&subbpo, bpo->bpo_os, subobj) != 0)
+		return;
+	if (bpobj_is_empty(&subbpo)) {
+		bpobj_close(&subbpo);
+		return;
+	}
+	subsubobjs = subbpo.bpo_phys->bpo_subobjs;
+	bpobj_close(&subbpo);
+
+	if (subsubobjs != 0) {
+		if (dmu_object_info(bpo->bpo_os, subsubobjs, &doi) != 0)
+			return;
+		if (doi.doi_max_offset > doi.doi_data_block_size)
+			copy_subsub = B_FALSE;
+	}
+
+	if (dmu_object_info(bpo->bpo_os, subobj, &doi) != 0)
+		return;
+	if (doi.doi_max_offset > doi.doi_data_block_size || !copy_subsub)
+		copy_bps = B_FALSE;
+
+	if (copy_subsub && subsubobjs != 0) {
+		if (bpo->bpo_phys->bpo_subobjs) {
+			dmu_prefetch(bpo->bpo_os, bpo->bpo_phys->bpo_subobjs, 0,
+			    bpo->bpo_phys->bpo_num_subobjs * sizeof (subobj), 1,
+			    ZIO_PRIORITY_ASYNC_READ);
+		}
+		dmu_prefetch(bpo->bpo_os, subsubobjs, 0, 0, 1,
+		    ZIO_PRIORITY_ASYNC_READ);
+	}
+
+	if (copy_bps) {
+		dmu_prefetch(bpo->bpo_os, bpo->bpo_object, 0,
+		    bpo->bpo_phys->bpo_num_blkptrs * sizeof (blkptr_t), 1,
+		    ZIO_PRIORITY_ASYNC_READ);
+		dmu_prefetch(bpo->bpo_os, subobj, 0, 0, 1,
+		    ZIO_PRIORITY_ASYNC_READ);
+	} else if (bpo->bpo_phys->bpo_subobjs) {
+		dmu_prefetch(bpo->bpo_os, bpo->bpo_phys->bpo_subobjs, 0,
+		    bpo->bpo_phys->bpo_num_subobjs * sizeof (subobj), 1,
+		    ZIO_PRIORITY_ASYNC_READ);
+	}
+}
+
 void
 bpobj_enqueue(bpobj_t *bpo, const blkptr_t *bp, boolean_t bp_freed,
     dmu_tx_t *tx)

module/zfs/btree.c

@@ -102,7 +102,7 @@ zfs_btree_poison_node(zfs_btree_t *tree, zfs_btree_hdr_t *hdr)
 		(void) memset(leaf->btl_elems, 0x0f, hdr->bth_first * size);
 		(void) memset(leaf->btl_elems +
 		    (hdr->bth_first + hdr->bth_count) * size, 0x0f,
-		    BTREE_LEAF_ESIZE -
+		    tree->bt_leaf_size - offsetof(zfs_btree_leaf_t, btl_elems) -
 		    (hdr->bth_first + hdr->bth_count) * size);
 	}
 #endif
@@ -173,16 +173,44 @@ zfs_btree_fini(void)
 	kmem_cache_destroy(zfs_btree_leaf_cache);
 }
 
+static void *
+zfs_btree_leaf_alloc(zfs_btree_t *tree)
+{
+	if (tree->bt_leaf_size == BTREE_LEAF_SIZE)
+		return (kmem_cache_alloc(zfs_btree_leaf_cache, KM_SLEEP));
+	else
+		return (kmem_alloc(tree->bt_leaf_size, KM_SLEEP));
+}
+
+static void
+zfs_btree_leaf_free(zfs_btree_t *tree, void *ptr)
+{
+	if (tree->bt_leaf_size == BTREE_LEAF_SIZE)
+		return (kmem_cache_free(zfs_btree_leaf_cache, ptr));
+	else
+		return (kmem_free(ptr, tree->bt_leaf_size));
+}
+
 void
 zfs_btree_create(zfs_btree_t *tree, int (*compar) (const void *, const void *),
     size_t size)
 {
-	ASSERT3U(size, <=, BTREE_LEAF_ESIZE / 2);
+	zfs_btree_create_custom(tree, compar, size, BTREE_LEAF_SIZE);
+}
 
-	bzero(tree, sizeof (*tree));
+void
+zfs_btree_create_custom(zfs_btree_t *tree,
+    int (*compar) (const void *, const void *),
+    size_t size, size_t lsize)
+{
+	size_t esize = lsize - offsetof(zfs_btree_leaf_t, btl_elems);
+
+	ASSERT3U(size, <=, esize / 2);
+	memset(tree, 0, sizeof (*tree));
 	tree->bt_compar = compar;
 	tree->bt_elem_size = size;
-	tree->bt_leaf_cap = P2ALIGN(BTREE_LEAF_ESIZE / size, 2);
+	tree->bt_leaf_size = lsize;
+	tree->bt_leaf_cap = P2ALIGN(esize / size, 2);
 	tree->bt_height = -1;
 	tree->bt_bulk = NULL;
 }
@@ -290,7 +318,7 @@ zfs_btree_find(zfs_btree_t *tree, const void *value, zfs_btree_index_t *where)
 
 	zfs_btree_core_t *node = NULL;
 	uint32_t child = 0;
-	uint64_t depth = 0;
+	uint32_t depth = 0;
 
 	/*
 	 * Iterate down the tree, finding which child the value should be in
@@ -811,8 +839,7 @@ zfs_btree_insert_into_leaf(zfs_btree_t *tree, zfs_btree_leaf_t *leaf,
 		move_count++;
 	}
 	tree->bt_num_nodes++;
-	zfs_btree_leaf_t *new_leaf = kmem_cache_alloc(zfs_btree_leaf_cache,
+	zfs_btree_leaf_t *new_leaf = zfs_btree_leaf_alloc(tree);
-	    KM_SLEEP);
 	zfs_btree_hdr_t *new_hdr = &new_leaf->btl_hdr;
 	new_hdr->bth_parent = leaf->btl_hdr.bth_parent;
 	new_hdr->bth_first = (tree->bt_bulk ? 0 : capacity / 4) +
@@ -1078,8 +1105,7 @@ zfs_btree_add_idx(zfs_btree_t *tree, const void *value,
 		ASSERT0(where->bti_offset);
 
 		tree->bt_num_nodes++;
-		zfs_btree_leaf_t *leaf = kmem_cache_alloc(zfs_btree_leaf_cache,
+		zfs_btree_leaf_t *leaf = zfs_btree_leaf_alloc(tree);
-		    KM_SLEEP);
 		tree->bt_root = &leaf->btl_hdr;
 		tree->bt_height++;
 
@@ -1378,7 +1404,7 @@ zfs_btree_node_destroy(zfs_btree_t *tree, zfs_btree_hdr_t *node)
 {
 	tree->bt_num_nodes--;
 	if (!zfs_btree_is_core(node)) {
-		kmem_cache_free(zfs_btree_leaf_cache, node);
+		zfs_btree_leaf_free(tree, node);
 	} else {
 		kmem_free(node, sizeof (zfs_btree_core_t) +
 		    BTREE_CORE_ELEMS * tree->bt_elem_size);
@@ -1991,7 +2017,7 @@ zfs_btree_verify_counts(zfs_btree_t *tree)
  */
 static uint64_t
 zfs_btree_verify_height_helper(zfs_btree_t *tree, zfs_btree_hdr_t *hdr,
-    int64_t height)
+    int32_t height)
 {
 	if (!zfs_btree_is_core(hdr)) {
 		VERIFY0(height);
@@ -2117,8 +2143,10 @@ zfs_btree_verify_poison_helper(zfs_btree_t *tree, zfs_btree_hdr_t *hdr)
 		zfs_btree_leaf_t *leaf = (zfs_btree_leaf_t *)hdr;
 		for (size_t i = 0; i < hdr->bth_first * size; i++)
 			VERIFY3U(leaf->btl_elems[i], ==, 0x0f);
+		size_t esize = tree->bt_leaf_size -
+		    offsetof(zfs_btree_leaf_t, btl_elems);
 		for (size_t i = (hdr->bth_first + hdr->bth_count) * size;
-		    i < BTREE_LEAF_ESIZE; i++)
+		    i < esize; i++)
 			VERIFY3U(leaf->btl_elems[i], ==, 0x0f);
 	} else {
 		zfs_btree_core_t *node = (zfs_btree_core_t *)hdr;