Tag 2.3.0-rc4

Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>

.github/workflows/codeql.yml

@@ -11,7 +11,7 @@ concurrency:
 jobs:
   analyze:
     name: Analyze
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-22.04
     permissions:
       actions: read
       contents: read

.github/workflows/scripts/qemu-1-setup.sh

@@ -18,19 +18,21 @@ ssh-keygen -t ed25519 -f ~/.ssh/id_ed25519 -q -N ""
 
 # we expect RAM shortage
 cat << EOF | sudo tee /etc/ksmtuned.conf > /dev/null
+# /etc/ksmtuned.conf - Configuration file for ksmtuned
 # https://docs.redhat.com/en/documentation/red_hat_enterprise_linux/7/html/virtualization_tuning_and_optimization_guide/chap-ksm
 KSM_MONITOR_INTERVAL=60
 
 # Millisecond sleep between ksm scans for 16Gb server.
 # Smaller servers sleep more, bigger sleep less.
-KSM_SLEEP_MSEC=10
-KSM_NPAGES_BOOST=300
-KSM_NPAGES_DECAY=-50
-KSM_NPAGES_MIN=64
-KSM_NPAGES_MAX=2048
+KSM_SLEEP_MSEC=30
 
-KSM_THRES_COEF=25
-KSM_THRES_CONST=2048
+KSM_NPAGES_BOOST=0
+KSM_NPAGES_DECAY=0
+KSM_NPAGES_MIN=1000
+KSM_NPAGES_MAX=25000
+
+KSM_THRES_COEF=80
+KSM_THRES_CONST=8192
 
 LOGFILE=/var/log/ksmtuned.log
 DEBUG=1

.github/workflows/scripts/qemu-2-start.sh

@@ -14,7 +14,7 @@ OSv=$OS
 
 # compressed with .zst extension
 REPO="https://github.com/mcmilk/openzfs-freebsd-images"
-FREEBSD="$REPO/releases/download/v2024-09-16"
+FREEBSD="$REPO/releases/download/v2024-10-05"
 URLzs=""
 
 # Ubuntu mirrors
@@ -52,43 +52,55 @@ case "$OS" in
     OSNAME="Debian 12"
     URL="https://cloud.debian.org/images/cloud/bookworm/latest/debian-12-generic-amd64.qcow2"
     ;;
-  fedora39)
-    OSNAME="Fedora 39"
-    OSv="fedora39"
-    URL="https://download.fedoraproject.org/pub/fedora/linux/releases/39/Cloud/x86_64/images/Fedora-Cloud-Base-39-1.5.x86_64.qcow2"
-    ;;
   fedora40)
     OSNAME="Fedora 40"
-    OSv="fedora39"
+    OSv="fedora-unknown"
     URL="https://download.fedoraproject.org/pub/fedora/linux/releases/40/Cloud/x86_64/images/Fedora-Cloud-Base-Generic.x86_64-40-1.14.qcow2"
     ;;
-  freebsd13r)
+  fedora41)
+    OSNAME="Fedora 41"
+    OSv="fedora-unknown"
+    URL="https://download.fedoraproject.org/pub/fedora/linux/releases/41/Cloud/x86_64/images/Fedora-Cloud-Base-Generic-41-1.4.x86_64.qcow2"
+    ;;
+  freebsd13-3r)
+    OSNAME="FreeBSD 13.3-RELEASE"
+    OSv="freebsd13.0"
+    URLzs="$FREEBSD/amd64-freebsd-13.3-RELEASE.qcow2.zst"
+    BASH="/usr/local/bin/bash"
+    NIC="rtl8139"
+    ;;
+  freebsd13-4r)
     OSNAME="FreeBSD 13.4-RELEASE"
     OSv="freebsd13.0"
     URLzs="$FREEBSD/amd64-freebsd-13.4-RELEASE.qcow2.zst"
     BASH="/usr/local/bin/bash"
     NIC="rtl8139"
     ;;
-  freebsd13)
-    OSNAME="FreeBSD 13.4-STABLE"
-    OSv="freebsd13.0"
-    URLzs="$FREEBSD/amd64-freebsd-13.4-STABLE.qcow2.zst"
+  freebsd14-0r)
+    OSNAME="FreeBSD 14.0-RELEASE"
+    OSv="freebsd14.0"
+    URLzs="$FREEBSD/amd64-freebsd-14.0-RELEASE.qcow2.zst"
     BASH="/usr/local/bin/bash"
-    NIC="rtl8139"
     ;;
-  freebsd14r)
+  freebsd14-1r)
     OSNAME="FreeBSD 14.1-RELEASE"
     OSv="freebsd14.0"
     URLzs="$FREEBSD/amd64-freebsd-14.1-RELEASE.qcow2.zst"
     BASH="/usr/local/bin/bash"
     ;;
-  freebsd14)
+  freebsd13-4s)
+    OSNAME="FreeBSD 13.4-STABLE"
+    OSv="freebsd13.0"
+    URLzs="$FREEBSD/amd64-freebsd-13.4-STABLE.qcow2.zst"
+    BASH="/usr/local/bin/bash"
+    ;;
+  freebsd14-1s)
     OSNAME="FreeBSD 14.1-STABLE"
     OSv="freebsd14.0"
     URLzs="$FREEBSD/amd64-freebsd-14.1-STABLE.qcow2.zst"
     BASH="/usr/local/bin/bash"
     ;;
-  freebsd15)
+  freebsd15-0c)
     OSNAME="FreeBSD 15.0-CURRENT"
     OSv="freebsd14.0"
     URLzs="$FREEBSD/amd64-freebsd-15.0-CURRENT.qcow2.zst"

.github/workflows/scripts/qemu-3-deps.sh

@@ -13,10 +13,10 @@ function archlinux() {
   echo "##[endgroup]"
 
   echo "##[group]Install Development Tools"
-  sudo pacman -Sy --noconfirm base-devel bc cpio dhclient dkms fakeroot \
-    fio gdb inetutils jq less linux linux-headers lsscsi nfs-utils parted \
-    pax perf python-packaging python-setuptools qemu-guest-agent ksh samba \
-    sysstat rng-tools rsync wget xxhash
+  sudo pacman -Sy --noconfirm base-devel bc cpio cryptsetup dhclient dkms \
+    fakeroot fio gdb inetutils jq less linux linux-headers lsscsi nfs-utils \
+    parted pax perf python-packaging python-setuptools qemu-guest-agent ksh \
+    samba sysstat rng-tools rsync wget xxhash
   echo "##[endgroup]"
 }
 
@@ -30,11 +30,11 @@ function debian() {
 
   echo "##[group]Install Development Tools"
   sudo apt-get install -y \
-    acl alien attr autoconf bc cpio curl dbench dh-python dkms fakeroot \
-    fio gdb gdebi git ksh lcov isc-dhcp-client jq libacl1-dev libaio-dev \
-    libattr1-dev libblkid-dev libcurl4-openssl-dev libdevmapper-dev libelf-dev \
-    libffi-dev libmount-dev libpam0g-dev libselinux-dev libssl-dev libtool \
-    libtool-bin libudev-dev libunwind-dev linux-headers-$(uname -r) \
+    acl alien attr autoconf bc cpio cryptsetup curl dbench dh-python dkms \
+    fakeroot fio gdb gdebi git ksh lcov isc-dhcp-client jq libacl1-dev \
+    libaio-dev libattr1-dev libblkid-dev libcurl4-openssl-dev libdevmapper-dev \
+    libelf-dev libffi-dev libmount-dev libpam0g-dev libselinux-dev libssl-dev \
+    libtool libtool-bin libudev-dev libunwind-dev linux-headers-$(uname -r) \
     lsscsi nfs-kernel-server pamtester parted python3 python3-all-dev \
     python3-cffi python3-dev python3-distlib python3-packaging \
     python3-setuptools python3-sphinx qemu-guest-agent rng-tools rpm2cpio \
@@ -66,16 +66,23 @@ function rhel() {
   echo "##[endgroup]"
 
   echo "##[group]Install Development Tools"
-  sudo dnf group install -y "Development Tools"
+
+  # Alma wants "Development Tools", Fedora 41 wants "development-tools"
+  if ! sudo dnf group install -y "Development Tools" ; then
+    echo "Trying 'development-tools' instead of 'Development Tools'"
+    sudo dnf group install -y development-tools
+  fi
+
   sudo dnf install -y \
-    acl attr bc bzip2 curl dbench dkms elfutils-libelf-devel fio gdb git \
-    jq kernel-rpm-macros ksh libacl-devel libaio-devel libargon2-devel \
-    libattr-devel libblkid-devel libcurl-devel libffi-devel ncompress \
-    libselinux-devel libtirpc-devel libtool libudev-devel libuuid-devel \
-    lsscsi mdadm nfs-utils openssl-devel pam-devel pamtester parted perf \
-    python3 python3-cffi python3-devel python3-packaging kernel-devel \
-    python3-setuptools qemu-guest-agent rng-tools rpcgen rpm-build rsync \
-    samba sysstat systemd watchdog wget xfsprogs-devel xxhash zlib-devel
+    acl attr bc bzip2 cryptsetup curl dbench dkms elfutils-libelf-devel fio \
+    gdb git jq kernel-rpm-macros ksh libacl-devel libaio-devel \
+    libargon2-devel libattr-devel libblkid-devel libcurl-devel libffi-devel \
+    ncompress libselinux-devel libtirpc-devel libtool libudev-devel \
+    libuuid-devel lsscsi mdadm nfs-utils openssl-devel pam-devel pamtester \
+    parted perf python3 python3-cffi python3-devel python3-packaging \
+    kernel-devel python3-setuptools qemu-guest-agent rng-tools rpcgen \
+    rpm-build rsync samba sysstat systemd watchdog wget xfsprogs-devel xxhash \
+    zlib-devel
   echo "##[endgroup]"
 }
 
@@ -111,6 +118,7 @@ case "$1" in
     archlinux
     ;;
   debian*)
+    echo 'debconf debconf/frontend select Noninteractive' | sudo debconf-set-selections
     debian
     echo "##[group]Install Debian specific"
     sudo apt-get install -yq linux-perf dh-sequence-dkms

.github/workflows/scripts/qemu-4-build.sh

@@ -83,7 +83,7 @@ function rpm_build_and_install() {
   echo "##[endgroup]"
 
   echo "##[group]Install"
-  run sudo dnf -y --skip-broken localinstall $(ls *.rpm | grep -v src.rpm)
+  run sudo dnf -y --nobest install $(ls *.rpm | grep -v src.rpm)
   echo "##[endgroup]"
 
 }

.github/workflows/scripts/qemu-5-setup.sh

@@ -14,17 +14,21 @@ PID=$(pidof /usr/bin/qemu-system-x86_64)
 tail --pid=$PID -f /dev/null
 sudo virsh undefine openzfs
 
-# definitions of per operating system
+# default values per test vm:
+VMs=2
+CPU=2
+
+# cpu pinning
+CPUSET=("0,1" "2,3")
+
 case "$OS" in
   freebsd*)
-    VMs=2
-    CPU=3
+    # FreeBSD can't be optimized via ksmtuned
     RAM=6
     ;;
   *)
-    VMs=2
-    CPU=3
-    RAM=7
+    # Linux can be optimized via ksmtuned
+    RAM=8
     ;;
 esac
 
@@ -73,6 +77,7 @@ EOF
     --cpu host-passthrough \
     --virt-type=kvm --hvm \
     --vcpus=$CPU,sockets=1 \
+    --cpuset=${CPUSET[$((i-1))]} \
     --memory $((1024*RAM)) \
     --memballoon model=virtio \
     --graphics none \

.github/workflows/scripts/qemu-9-summary-page.sh

@@ -11,12 +11,10 @@ function output() {
 }
 
 function outfile() {
-  test -s "$1" || return
   cat "$1" >> "out-$logfile.md"
 }
 
 function outfile_plain() {
-  test -s "$1" || return
   output "<pre>"
   cat "$1" >> "out-$logfile.md"
   output "</pre>"
@@ -45,6 +43,8 @@ if [ ! -f out-1.md ]; then
     tar xf "$tarfile"
     test -s env.txt || continue
     source env.txt
+    # when uname.txt is there, the other files are also ok
+    test -s uname.txt || continue
     output "\n## Functional Tests: $OSNAME\n"
     outfile_plain uname.txt
     outfile_plain summary.txt

.github/workflows/zfs-qemu.yml

@@ -22,8 +22,8 @@ jobs:
       - name: Generate OS config and CI type
         id: os
         run: |
-          FULL_OS='["almalinux8", "almalinux9", "centos-stream9", "debian11", "debian12", "fedora39", "fedora40", "freebsd13", "freebsd13r", "freebsd14", "freebsd14r", "ubuntu20", "ubuntu22", "ubuntu24"]'
-          QUICK_OS='["almalinux8", "almalinux9", "debian12", "fedora40", "freebsd13", "freebsd14", "ubuntu24"]'
+          FULL_OS='["almalinux8", "almalinux9", "centos-stream9", "debian11", "debian12", "fedora40", "fedora41", "freebsd13-4r", "freebsd14-0r", "freebsd14-1s", "ubuntu20", "ubuntu22", "ubuntu24"]'
+          QUICK_OS='["almalinux8", "almalinux9", "debian12", "fedora41", "freebsd13-3r", "freebsd14-1r", "ubuntu24"]'
           # determine CI type when running on PR
           ci_type="full"
           if ${{ github.event_name == 'pull_request' }}; then
@@ -46,10 +46,12 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        # all:
-        # os:  [almalinux8, almalinux9, archlinux, centos-stream9, fedora39, fedora40, debian11, debian12, freebsd13, freebsd13r, freebsd14, freebsd14r, freebsd15, ubuntu20, ubuntu22, ubuntu24]
-        # openzfs:
-        # os: [almalinux8, almalinux9, centos-stream9, debian11, debian12, fedora39, fedora40, freebsd13, freebsd13r, freebsd14, freebsd14r, ubuntu20, ubuntu22, ubuntu24]
+        # rhl:     almalinux8, almalinux9, centos-stream9, fedora40, fedora41
+        # debian:  debian11, debian12, ubuntu20, ubuntu22, ubuntu24
+        # misc:    archlinux, tumbleweed
+        # FreeBSD Release: freebsd13-3r, freebsd13-4r, freebsd14-0r, freebsd14-1r
+        # FreeBSD Stable:  freebsd13-4s, freebsd14-1s
+        # FreeBSD Current: freebsd15-0c
         os: ${{ fromJson(needs.test-config.outputs.test_os) }}
     runs-on: ubuntu-24.04
     steps:

.mailmap

@@ -70,6 +70,7 @@ Rob Norris <robn@despairlabs.com>
 Rob Norris <rob.norris@klarasystems.com>
 Sam Lunt <samuel.j.lunt@gmail.com>
 Sanjeev Bagewadi <sanjeev.bagewadi@gmail.com>
+Sebastian Wuerl <s.wuerl@mailbox.org>
 Stoiko Ivanov <github@nomore.at>
 Tamas TEVESZ <ice@extreme.hu>
 WHR <msl0000023508@gmail.com>
@@ -78,6 +79,7 @@ Youzhong Yang <youzhong@gmail.com>
 
 # Signed-off-by: overriding Author:
 Ryan <errornointernet@envs.net> <error.nointernet@gmail.com>
+Sietse <sietse@wizdom.nu> <uglymotha@wizdom.nu>
 Qiuhao Chen <chenqiuhao1997@gmail.com> <haohao0924@126.com>
 Yuxin Wang <yuxinwang9999@gmail.com> <Bi11gates9999@gmail.com>
 Zhenlei Huang <zlei@FreeBSD.org> <zlei.huang@gmail.com>

AUTHORS

@@ -423,6 +423,7 @@ CONTRIBUTORS:
     Mathieu Velten <matmaul@gmail.com>
     Matt Fiddaman <github@m.fiddaman.uk>
     Matthew Ahrens <matt@delphix.com>
+    Matthew Heller <matthew.f.heller@gmail.com>
     Matthew Thode <mthode@mthode.org>
     Matthias Blankertz <matthias@blankertz.org>
     Matt Johnston <matt@fugro-fsi.com.au>
@@ -562,6 +563,7 @@ CONTRIBUTORS:
     Scot W. Stevenson <scot.stevenson@gmail.com>
     Sean Eric Fagan <sef@ixsystems.com>
     Sebastian Gottschall <s.gottschall@dd-wrt.com>
+    Sebastian Wuerl <s.wuerl@mailbox.org>
     Sebastien Roy <seb@delphix.com>
     Sen Haerens <sen@senhaerens.be>
     Serapheim Dimitropoulos <serapheim@delphix.com>
@@ -574,6 +576,7 @@ CONTRIBUTORS:
     Shawn Bayern <sbayern@law.fsu.edu>
     Shengqi Chen <harry-chen@outlook.com>
     Shen Yan <shenyanxxxy@qq.com>
+    Sietse <sietse@wizdom.nu>
     Simon Guest <simon.guest@tesujimath.org>
     Simon Klinkert <simon.klinkert@gmail.com>
     Sowrabha Gopal <sowrabha.gopal@delphix.com>
@@ -629,6 +632,7 @@ CONTRIBUTORS:
     Trevor Bautista <trevrb@trevrb.net>
     Trey Dockendorf <treydock@gmail.com>
     Troels Nørgaard <tnn@tradeshift.com>
+    tstabrawa <tstabrawa@users.noreply.github.com>
     Tulsi Jain <tulsi.jain@delphix.com>
     Turbo Fredriksson <turbo@bayour.com>
     Tyler J. Stachecki <stachecki.tyler@gmail.com>

META

@@ -2,9 +2,9 @@ Meta:          1
 Name:          zfs
 Branch:        1.0
 Version:       2.3.0
-Release:       rc1
+Release:       rc4
 Release-Tags:  relext
 License:       CDDL
 Author:        OpenZFS
-Linux-Maximum: 6.11
+Linux-Maximum: 6.12
 Linux-Minimum: 4.18

cmd/arc_summary

@@ -662,10 +662,7 @@ def section_arc(kstats_dict):
     print()
 
     print('ARC hash breakdown:')
-    prt_i1('Elements max:', f_hits(arc_stats['hash_elements_max']))
-    prt_i2('Elements current:',
-           f_perc(arc_stats['hash_elements'], arc_stats['hash_elements_max']),
-           f_hits(arc_stats['hash_elements']))
+    prt_i1('Elements:', f_hits(arc_stats['hash_elements']))
     prt_i1('Collisions:', f_hits(arc_stats['hash_collisions']))
 
     prt_i1('Chain max:', f_hits(arc_stats['hash_chain_max']))

cmd/zdb/zdb.c

@@ -1131,7 +1131,7 @@ dump_zap(objset_t *os, uint64_t object, void *data, size_t size)
 		    !!(zap_getflags(zc.zc_zap) & ZAP_FLAG_UINT64_KEY);
 
 		if (key64)
-			(void) printf("\t\t0x%010lx = ",
+			(void) printf("\t\t0x%010" PRIu64 "x = ",
 			    *(uint64_t *)attrp->za_name);
 		else
 			(void) printf("\t\t%s = ", attrp->za_name);
@@ -1967,17 +1967,53 @@ dump_dedup_ratio(const ddt_stat_t *dds)
 static void
 dump_ddt_log(ddt_t *ddt)
 {
+	if (ddt->ddt_version != DDT_VERSION_FDT ||
+	    !(ddt->ddt_flags & DDT_FLAG_LOG))
+		return;
+
 	for (int n = 0; n < 2; n++) {
 		ddt_log_t *ddl = &ddt->ddt_log[n];
 
+		char flagstr[64] = {0};
+		if (ddl->ddl_flags > 0) {
+			flagstr[0] = ' ';
+			int c = 1;
+			if (ddl->ddl_flags & DDL_FLAG_FLUSHING)
+				c += strlcpy(&flagstr[c], " FLUSHING",
+				    sizeof (flagstr) - c);
+			if (ddl->ddl_flags & DDL_FLAG_CHECKPOINT)
+				c += strlcpy(&flagstr[c], " CHECKPOINT",
+				    sizeof (flagstr) - c);
+			if (ddl->ddl_flags &
+			    ~(DDL_FLAG_FLUSHING|DDL_FLAG_CHECKPOINT))
+				c += strlcpy(&flagstr[c], " UNKNOWN",
+				    sizeof (flagstr) - c);
+			flagstr[1] = '[';
+			flagstr[c++] = ']';
+		}
+
 		uint64_t count = avl_numnodes(&ddl->ddl_tree);
-		if (count == 0)
-			continue;
 
-		printf(DMU_POOL_DDT_LOG ": %lu log entries\n",
-		    zio_checksum_table[ddt->ddt_checksum].ci_name, n, count);
+		printf(DMU_POOL_DDT_LOG ": flags=0x%02x%s; obj=%llu; "
+		    "len=%llu; txg=%llu; entries=%llu\n",
+		    zio_checksum_table[ddt->ddt_checksum].ci_name, n,
+		    ddl->ddl_flags, flagstr,
+		    (u_longlong_t)ddl->ddl_object,
+		    (u_longlong_t)ddl->ddl_length,
+		    (u_longlong_t)ddl->ddl_first_txg, (u_longlong_t)count);
 
-		if (dump_opt['D'] < 4)
+		if (ddl->ddl_flags & DDL_FLAG_CHECKPOINT) {
+			const ddt_key_t *ddk = &ddl->ddl_checkpoint;
+			printf("    checkpoint: "
+			    "%016llx:%016llx:%016llx:%016llx:%016llx\n",
+			    (u_longlong_t)ddk->ddk_cksum.zc_word[0],
+			    (u_longlong_t)ddk->ddk_cksum.zc_word[1],
+			    (u_longlong_t)ddk->ddk_cksum.zc_word[2],
+			    (u_longlong_t)ddk->ddk_cksum.zc_word[3],
+			    (u_longlong_t)ddk->ddk_prop);
+		}
+
+		if (count == 0 || dump_opt['D'] < 4)
 			continue;
 
 		ddt_lightweight_entry_t ddlwe;
@@ -1991,7 +2027,7 @@ dump_ddt_log(ddt_t *ddt)
 }
 
 static void
-dump_ddt(ddt_t *ddt, ddt_type_t type, ddt_class_t class)
+dump_ddt_object(ddt_t *ddt, ddt_type_t type, ddt_class_t class)
 {
 	char name[DDT_NAMELEN];
 	ddt_lightweight_entry_t ddlwe;
@@ -2016,11 +2052,8 @@ dump_ddt(ddt_t *ddt, ddt_type_t type, ddt_class_t class)
 
 	ddt_object_name(ddt, type, class, name);
 
-	(void) printf("%s: %llu entries, size %llu on disk, %llu in core\n",
-	    name,
-	    (u_longlong_t)count,
-	    (u_longlong_t)dspace,
-	    (u_longlong_t)mspace);
+	(void) printf("%s: dspace=%llu; mspace=%llu; entries=%llu\n", name,
+	    (u_longlong_t)dspace, (u_longlong_t)mspace, (u_longlong_t)count);
 
 	if (dump_opt['D'] < 3)
 		return;
@@ -2043,24 +2076,52 @@ dump_ddt(ddt_t *ddt, ddt_type_t type, ddt_class_t class)
 	(void) printf("\n");
 }
 
+static void
+dump_ddt(ddt_t *ddt)
+{
+	if (!ddt || ddt->ddt_version == DDT_VERSION_UNCONFIGURED)
+		return;
+
+	char flagstr[64] = {0};
+	if (ddt->ddt_flags > 0) {
+		flagstr[0] = ' ';
+		int c = 1;
+		if (ddt->ddt_flags & DDT_FLAG_FLAT)
+			c += strlcpy(&flagstr[c], " FLAT",
+			    sizeof (flagstr) - c);
+		if (ddt->ddt_flags & DDT_FLAG_LOG)
+			c += strlcpy(&flagstr[c], " LOG",
+			    sizeof (flagstr) - c);
+		if (ddt->ddt_flags & ~DDT_FLAG_MASK)
+			c += strlcpy(&flagstr[c], " UNKNOWN",
+			    sizeof (flagstr) - c);
+		flagstr[1] = '[';
+		flagstr[c] = ']';
+	}
+
+	printf("DDT-%s: version=%llu [%s]; flags=0x%02llx%s; rootobj=%llu\n",
+	    zio_checksum_table[ddt->ddt_checksum].ci_name,
+	    (u_longlong_t)ddt->ddt_version,
+	    (ddt->ddt_version == 0) ? "LEGACY" :
+	    (ddt->ddt_version == 1) ? "FDT" : "UNKNOWN",
+	    (u_longlong_t)ddt->ddt_flags, flagstr,
+	    (u_longlong_t)ddt->ddt_dir_object);
+
+	for (ddt_type_t type = 0; type < DDT_TYPES; type++)
+		for (ddt_class_t class = 0; class < DDT_CLASSES; class++)
+			dump_ddt_object(ddt, type, class);
+
+	dump_ddt_log(ddt);
+}
+
 static void
 dump_all_ddts(spa_t *spa)
 {
 	ddt_histogram_t ddh_total = {{{0}}};
 	ddt_stat_t dds_total = {0};
 
-	for (enum zio_checksum c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++) {
-		ddt_t *ddt = spa->spa_ddt[c];
-		if (!ddt || ddt->ddt_version == DDT_VERSION_UNCONFIGURED)
-			continue;
-		for (ddt_type_t type = 0; type < DDT_TYPES; type++) {
-			for (ddt_class_t class = 0; class < DDT_CLASSES;
-			    class++) {
-				dump_ddt(ddt, type, class);
-			}
-		}
-		dump_ddt_log(ddt);
-	}
+	for (enum zio_checksum c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++)
+		dump_ddt(spa->spa_ddt[c]);
 
 	ddt_get_dedup_stats(spa, &dds_total);
 
@@ -2119,9 +2180,6 @@ dump_brt(spa_t *spa)
 		return;
 	}
 
-	brt_t *brt = spa->spa_brt;
-	VERIFY(brt);
-
 	char count[32], used[32], saved[32];
 	zdb_nicebytes(brt_get_used(spa), used, sizeof (used));
 	zdb_nicebytes(brt_get_saved(spa), saved, sizeof (saved));
@@ -2132,11 +2190,8 @@ dump_brt(spa_t *spa)
 	if (dump_opt['T'] < 2)
 		return;
 
-	for (uint64_t vdevid = 0; vdevid < brt->brt_nvdevs; vdevid++) {
-		brt_vdev_t *brtvd = &brt->brt_vdevs[vdevid];
-		if (brtvd == NULL)
-			continue;
-
+	for (uint64_t vdevid = 0; vdevid < spa->spa_brt_nvdevs; vdevid++) {
+		brt_vdev_t *brtvd = spa->spa_brt_vdevs[vdevid];
 		if (!brtvd->bv_initiated) {
 			printf("BRT: vdev %" PRIu64 ": empty\n", vdevid);
 			continue;
@@ -2152,34 +2207,54 @@ dump_brt(spa_t *spa)
 	if (dump_opt['T'] < 3)
 		return;
 
+	/* -TTT shows a per-vdev histograms; -TTTT shows all entries */
+	boolean_t do_histo = dump_opt['T'] == 3;
+
 	char dva[64];
+
+	if (!do_histo)
 		printf("\n%-16s %-10s\n", "DVA", "REFCNT");
 
-	for (uint64_t vdevid = 0; vdevid < brt->brt_nvdevs; vdevid++) {
-		brt_vdev_t *brtvd = &brt->brt_vdevs[vdevid];
-		if (brtvd == NULL || !brtvd->bv_initiated)
+	for (uint64_t vdevid = 0; vdevid < spa->spa_brt_nvdevs; vdevid++) {
+		brt_vdev_t *brtvd = spa->spa_brt_vdevs[vdevid];
+		if (!brtvd->bv_initiated)
 			continue;
 
+		uint64_t counts[64] = {};
+
 		zap_cursor_t zc;
 		zap_attribute_t *za = zap_attribute_alloc();
-		for (zap_cursor_init(&zc, brt->brt_mos, brtvd->bv_mos_entries);
+		for (zap_cursor_init(&zc, spa->spa_meta_objset,
+		    brtvd->bv_mos_entries);
 		    zap_cursor_retrieve(&zc, za) == 0;
 		    zap_cursor_advance(&zc)) {
 			uint64_t refcnt;
-			VERIFY0(zap_lookup_uint64(brt->brt_mos,
+			VERIFY0(zap_lookup_uint64(spa->spa_meta_objset,
 			    brtvd->bv_mos_entries,
 			    (const uint64_t *)za->za_name, 1,
 			    za->za_integer_length, za->za_num_integers,
 			    &refcnt));
 
-			uint64_t offset = *(const uint64_t *)za->za_name;
+			if (do_histo)
+				counts[highbit64(refcnt)]++;
+			else {
+				uint64_t offset =
+				    *(const uint64_t *)za->za_name;
 
-			snprintf(dva, sizeof (dva), "%" PRIu64 ":%llx", vdevid,
-			    (u_longlong_t)offset);
-			printf("%-16s %-10llu\n", dva, (u_longlong_t)refcnt);
+				snprintf(dva, sizeof (dva), "%" PRIu64 ":%llx",
+				    vdevid, (u_longlong_t)offset);
+				printf("%-16s %-10llu\n", dva,
+				    (u_longlong_t)refcnt);
+			}
 		}
 		zap_cursor_fini(&zc);
 		zap_attribute_free(za);
+
+		if (do_histo) {
+			printf("\nBRT: vdev %" PRIu64
+			    ": DVAs with 2^n refcnts:\n", vdevid);
+			dump_histogram(counts, 64, 0);
+		}
 	}
 }
 
@@ -4266,6 +4341,10 @@ dump_uberblock(uberblock_t *ub, const char *header, const char *footer)
 	(void) printf("\ttimestamp = %llu UTC = %s",
 	    (u_longlong_t)ub->ub_timestamp, ctime(&timestamp));
 
+	char blkbuf[BP_SPRINTF_LEN];
+	snprintf_blkptr(blkbuf, sizeof (blkbuf), &ub->ub_rootbp);
+	(void) printf("\tbp = %s\n", blkbuf);
+
 	(void) printf("\tmmp_magic = %016llx\n",
 	    (u_longlong_t)ub->ub_mmp_magic);
 	if (MMP_VALID(ub)) {
@@ -6874,7 +6953,7 @@ iterate_deleted_livelists(spa_t *spa, ll_iter_t func, void *arg)
 	for (zap_cursor_init(&zc, mos, zap_obj);
 	    zap_cursor_retrieve(&zc, attrp) == 0;
 	    (void) zap_cursor_advance(&zc)) {
-		dsl_deadlist_open(&ll, mos, attrp->za_first_integer);
+		VERIFY0(dsl_deadlist_open(&ll, mos, attrp->za_first_integer));
 		func(&ll, arg);
 		dsl_deadlist_close(&ll);
 	}
@@ -8204,16 +8283,13 @@ dump_mos_leaks(spa_t *spa)
 		}
 	}
 
-	if (spa->spa_brt != NULL) {
-		brt_t *brt = spa->spa_brt;
-		for (uint64_t vdevid = 0; vdevid < brt->brt_nvdevs; vdevid++) {
-			brt_vdev_t *brtvd = &brt->brt_vdevs[vdevid];
-			if (brtvd != NULL && brtvd->bv_initiated) {
+	for (uint64_t vdevid = 0; vdevid < spa->spa_brt_nvdevs; vdevid++) {
+		brt_vdev_t *brtvd = spa->spa_brt_vdevs[vdevid];
+		if (brtvd->bv_initiated) {
 			mos_obj_refd(brtvd->bv_mos_brtvdev);
 			mos_obj_refd(brtvd->bv_mos_entries);
 		}
 	}
-	}
 
 	/*
 	 * Visit all allocated objects and make sure they are referenced.

cmd/zdb/zdb_il.c

@@ -67,19 +67,19 @@ zil_prt_rec_create(zilog_t *zilog, int txtype, const void *arg)
 	const lr_create_t *lrc = arg;
 	const _lr_create_t *lr = &lrc->lr_create;
 	time_t crtime = lr->lr_crtime[0];
-	char *name, *link;
+	const char *name, *link;
 	lr_attr_t *lrattr;
 
-	name = (char *)(lr + 1);
+	name = (const char *)&lrc->lr_data[0];
 
 	if (lr->lr_common.lrc_txtype == TX_CREATE_ATTR ||
 	    lr->lr_common.lrc_txtype == TX_MKDIR_ATTR) {
-		lrattr = (lr_attr_t *)(lr + 1);
+		lrattr = (lr_attr_t *)&lrc->lr_data[0];
 		name += ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
 	}
 
 	if (txtype == TX_SYMLINK) {
-		link = name + strlen(name) + 1;
+		link = (const char *)&lrc->lr_data[strlen(name) + 1];
 		(void) printf("%s%s -> %s\n", tab_prefix, name, link);
 	} else if (txtype != TX_MKXATTR) {
 		(void) printf("%s%s\n", tab_prefix, name);
@@ -104,7 +104,7 @@ zil_prt_rec_remove(zilog_t *zilog, int txtype, const void *arg)
 	const lr_remove_t *lr = arg;
 
 	(void) printf("%sdoid %llu, name %s\n", tab_prefix,
-	    (u_longlong_t)lr->lr_doid, (char *)(lr + 1));
+	    (u_longlong_t)lr->lr_doid, (const char *)&lr->lr_data[0]);
 }
 
 static void
@@ -115,7 +115,7 @@ zil_prt_rec_link(zilog_t *zilog, int txtype, const void *arg)
 
 	(void) printf("%sdoid %llu, link_obj %llu, name %s\n", tab_prefix,
 	    (u_longlong_t)lr->lr_doid, (u_longlong_t)lr->lr_link_obj,
-	    (char *)(lr + 1));
+	    (const char *)&lr->lr_data[0]);
 }
 
 static void
@@ -124,8 +124,8 @@ zil_prt_rec_rename(zilog_t *zilog, int txtype, const void *arg)
 	(void) zilog, (void) txtype;
 	const lr_rename_t *lrr = arg;
 	const _lr_rename_t *lr = &lrr->lr_rename;
-	char *snm = (char *)(lr + 1);
-	char *tnm = snm + strlen(snm) + 1;
+	const char *snm = (const char *)&lrr->lr_data[0];
+	const char *tnm = (const char *)&lrr->lr_data[strlen(snm) + 1];
 
 	(void) printf("%ssdoid %llu, tdoid %llu\n", tab_prefix,
 	    (u_longlong_t)lr->lr_sdoid, (u_longlong_t)lr->lr_tdoid);
@@ -211,7 +211,7 @@ zil_prt_rec_write(zilog_t *zilog, int txtype, const void *arg)
 
 		/* data is stored after the end of the lr_write record */
 		data = abd_alloc(lr->lr_length, B_FALSE);
-		abd_copy_from_buf(data, lr + 1, lr->lr_length);
+		abd_copy_from_buf(data, &lr->lr_data[0], lr->lr_length);
 	}
 
 	(void) printf("%s", tab_prefix);
@@ -309,7 +309,7 @@ zil_prt_rec_setsaxattr(zilog_t *zilog, int txtype, const void *arg)
 	(void) zilog, (void) txtype;
 	const lr_setsaxattr_t *lr = arg;
 
-	char *name = (char *)(lr + 1);
+	const char *name = (const char *)&lr->lr_data[0];
 	(void) printf("%sfoid %llu\n", tab_prefix,
 	    (u_longlong_t)lr->lr_foid);
 
@@ -318,7 +318,7 @@ zil_prt_rec_setsaxattr(zilog_t *zilog, int txtype, const void *arg)
 		(void) printf("%sXAT_VALUE  NULL\n", tab_prefix);
 	} else {
 		(void) printf("%sXAT_VALUE  ", tab_prefix);
-		char *val = name + (strlen(name) + 1);
+		const char *val = (const char *)&lr->lr_data[strlen(name) + 1];
 		for (int i = 0; i < lr->lr_size; i++) {
 			(void) printf("%c", *val);
 			val++;

cmd/zed/agents/zfs_retire.c

@@ -445,8 +445,8 @@ zfs_retire_recv(fmd_hdl_t *hdl, fmd_event_t *ep, nvlist_t *nvl,
 		 * its a loopback event from spa_async_remove(). Just
 		 * ignore it.
 		 */
-		if (vs->vs_state == VDEV_STATE_REMOVED &&
-		    state == VDEV_STATE_REMOVED)
+		if ((vs->vs_state == VDEV_STATE_REMOVED && state ==
+		    VDEV_STATE_REMOVED) || vs->vs_state == VDEV_STATE_OFFLINE)
 			return;
 
 		/* Remove the vdev since device is unplugged */

cmd/zed/zed_disk_event.c

@@ -139,7 +139,8 @@ dev_event_nvlist(struct udev_device *dev)
 		 * is /dev/sda.
 		 */
 		struct udev_device *parent_dev = udev_device_get_parent(dev);
-		if ((value = udev_device_get_sysattr_value(parent_dev, "size"))
+		if (parent_dev != NULL &&
+		    (value = udev_device_get_sysattr_value(parent_dev, "size"))
 		    != NULL) {
 			uint64_t numval = DEV_BSIZE;
 

cmd/zfs/zfs_main.c

@@ -2162,6 +2162,7 @@ zfs_do_get(int argc, char **argv)
 	cb.cb_type = ZFS_TYPE_DATASET;
 
 	struct option long_options[] = {
+		{"json", no_argument, NULL, 'j'},
 		{"json-int", no_argument, NULL, ZFS_OPTION_JSON_NUMS_AS_INT},
 		{0, 0, 0, 0}
 	};
@@ -3760,8 +3761,13 @@ collect_dataset(zfs_handle_t *zhp, list_cbdata_t *cb)
 		if (cb->cb_json) {
 			if (pl->pl_prop == ZFS_PROP_NAME)
 				continue;
+			const char *prop_name;
+			if (pl->pl_prop != ZPROP_USERPROP)
+				prop_name = zfs_prop_to_name(pl->pl_prop);
+			else
+				prop_name = pl->pl_user_prop;
 			if (zprop_nvlist_one_property(
-			    zfs_prop_to_name(pl->pl_prop), propstr,
+			    prop_name, propstr,
 			    sourcetype, source, NULL, props,
 			    cb->cb_json_as_int) != 0)
 				nomem();
@@ -3852,6 +3858,7 @@ zfs_do_list(int argc, char **argv)
 	nvlist_t *data = NULL;
 
 	struct option long_options[] = {
+		{"json", no_argument, NULL, 'j'},
 		{"json-int", no_argument, NULL, ZFS_OPTION_JSON_NUMS_AS_INT},
 		{0, 0, 0, 0}
 	};
@@ -7436,9 +7443,15 @@ share_mount(int op, int argc, char **argv)
 	uint_t nthr;
 	jsobj = data = item = NULL;
 
+	struct option long_options[] = {
+		{"json", no_argument, NULL, 'j'},
+		{0, 0, 0, 0}
+	};
+
 	/* check options */
-	while ((c = getopt(argc, argv, op == OP_MOUNT ? ":ajRlvo:Of" : "al"))
-	    != -1) {
+	while ((c = getopt_long(argc, argv,
+	    op == OP_MOUNT ? ":ajRlvo:Of" : "al",
+	    op == OP_MOUNT ? long_options : NULL, NULL)) != -1) {
 		switch (c) {
 		case 'a':
 			do_all = 1;
@@ -8374,8 +8387,14 @@ zfs_do_channel_program(int argc, char **argv)
 	boolean_t sync_flag = B_TRUE, json_output = B_FALSE;
 	zpool_handle_t *zhp;
 
+	struct option long_options[] = {
+		{"json", no_argument, NULL, 'j'},
+		{0, 0, 0, 0}
+	};
+
 	/* check options */
-	while ((c = getopt(argc, argv, "nt:m:j")) != -1) {
+	while ((c = getopt_long(argc, argv, "nt:m:j", long_options,
+	    NULL)) != -1) {
 		switch (c) {
 		case 't':
 		case 'm': {
@@ -9083,7 +9102,13 @@ zfs_do_version(int argc, char **argv)
 	int c;
 	nvlist_t *jsobj = NULL, *zfs_ver = NULL;
 	boolean_t json = B_FALSE;
-	while ((c = getopt(argc, argv, "j")) != -1) {
+
+	struct option long_options[] = {
+		{"json", no_argument, NULL, 'j'},
+		{0, 0, 0, 0}
+	};
+
+	while ((c = getopt_long(argc, argv, "j", long_options, NULL)) != -1) {
 		switch (c) {
 		case 'j':
 			json = B_TRUE;
@@ -9187,7 +9212,7 @@ main(int argc, char **argv)
 	 * Special case '-V|--version'
 	 */
 	if ((strcmp(cmdname, "-V") == 0) || (strcmp(cmdname, "--version") == 0))
-		return (zfs_do_version(argc, argv));
+		return (zfs_version_print() != 0);
 
 	/*
 	 * Special case 'help'

cmd/zpool/zpool_main.c

@@ -512,7 +512,8 @@ get_usage(zpool_help_t idx)
 		return (gettext("\tinitialize [-c | -s | -u] [-w] <pool> "
 		    "[<device> ...]\n"));
 	case HELP_SCRUB:
-		return (gettext("\tscrub [-s | -p] [-w] [-e] <pool> ...\n"));
+		return (gettext("\tscrub [-e | -s | -p | -C] [-w] "
+		    "<pool> ...\n"));
 	case HELP_RESILVER:
 		return (gettext("\tresilver <pool> ...\n"));
 	case HELP_TRIM:
@@ -6882,8 +6883,13 @@ collect_pool(zpool_handle_t *zhp, list_cbdata_t *cb)
 		if (cb->cb_json) {
 			if (pl->pl_prop == ZPOOL_PROP_NAME)
 				continue;
+			const char *prop_name;
+			if (pl->pl_prop != ZPROP_USERPROP)
+				prop_name = zpool_prop_to_name(pl->pl_prop);
+			else
+				prop_name = pl->pl_user_prop;
 			(void) zprop_nvlist_one_property(
-			    zpool_prop_to_name(pl->pl_prop), propstr,
+			    prop_name, propstr,
 			    sourcetype, NULL, NULL, props, cb->cb_json_as_int);
 		} else {
 			/*
@@ -7340,6 +7346,7 @@ zpool_do_list(int argc, char **argv)
 	current_prop_type = ZFS_TYPE_POOL;
 
 	struct option long_options[] = {
+		{"json", no_argument, NULL, 'j'},
 		{"json-int", no_argument, NULL, ZPOOL_OPTION_JSON_NUMS_AS_INT},
 		{"json-pool-key-guid", no_argument, NULL,
 		    ZPOOL_OPTION_POOL_KEY_GUID},
@@ -7965,8 +7972,11 @@ zpool_do_online(int argc, char **argv)
 
 	poolname = argv[0];
 
-	if ((zhp = zpool_open(g_zfs, poolname)) == NULL)
+	if ((zhp = zpool_open(g_zfs, poolname)) == NULL) {
+		(void) fprintf(stderr, gettext("failed to open pool "
+		    "\"%s\""), poolname);
 		return (1);
+	}
 
 	for (i = 1; i < argc; i++) {
 		vdev_state_t oldstate;
@@ -7987,12 +7997,15 @@ zpool_do_online(int argc, char **argv)
 		    &l2cache, NULL);
 		if (tgt == NULL) {
 			ret = 1;
+			(void) fprintf(stderr, gettext("couldn't find device "
+			"\"%s\" in pool \"%s\"\n"), argv[i], poolname);
 			continue;
 		}
 		uint_t vsc;
 		oldstate = ((vdev_stat_t *)fnvlist_lookup_uint64_array(tgt,
 		    ZPOOL_CONFIG_VDEV_STATS, &vsc))->vs_state;
-		if (zpool_vdev_online(zhp, argv[i], flags, &newstate) == 0) {
+		if ((rc = zpool_vdev_online(zhp, argv[i], flags,
+		    &newstate)) == 0) {
 			if (newstate != VDEV_STATE_HEALTHY) {
 				(void) printf(gettext("warning: device '%s' "
 				    "onlined, but remains in faulted state\n"),
@@ -8018,6 +8031,9 @@ zpool_do_online(int argc, char **argv)
 				}
 			}
 		} else {
+			(void) fprintf(stderr, gettext("Failed to online "
+			    "\"%s\" in pool \"%s\": %d\n"),
+			    argv[i], poolname, rc);
 			ret = 1;
 		}
 	}
@@ -8102,8 +8118,11 @@ zpool_do_offline(int argc, char **argv)
 
 	poolname = argv[0];
 
-	if ((zhp = zpool_open(g_zfs, poolname)) == NULL)
+	if ((zhp = zpool_open(g_zfs, poolname)) == NULL) {
+		(void) fprintf(stderr, gettext("failed to open pool "
+		    "\"%s\""), poolname);
 		return (1);
+	}
 
 	for (i = 1; i < argc; i++) {
 		uint64_t guid = zpool_vdev_path_to_guid(zhp, argv[i]);
@@ -8411,12 +8430,13 @@ wait_callback(zpool_handle_t *zhp, void *data)
 }
 
 /*
- * zpool scrub [-s | -p] [-w] [-e] <pool> ...
+ * zpool scrub [-e | -s | -p | -C] [-w] <pool> ...
  *
  *	-e	Only scrub blocks in the error log.
  *	-s	Stop.  Stops any in-progress scrub.
  *	-p	Pause. Pause in-progress scrub.
  *	-w	Wait.  Blocks until scrub has completed.
+ *	-C	Scrub from last saved txg.
  */
 int
 zpool_do_scrub(int argc, char **argv)
@@ -8432,9 +8452,10 @@ zpool_do_scrub(int argc, char **argv)
 	boolean_t is_error_scrub = B_FALSE;
 	boolean_t is_pause = B_FALSE;
 	boolean_t is_stop = B_FALSE;
+	boolean_t is_txg_continue = B_FALSE;
 
 	/* check options */
-	while ((c = getopt(argc, argv, "spwe")) != -1) {
+	while ((c = getopt(argc, argv, "spweC")) != -1) {
 		switch (c) {
 		case 'e':
 			is_error_scrub = B_TRUE;
@@ -8448,6 +8469,9 @@ zpool_do_scrub(int argc, char **argv)
 		case 'w':
 			wait = B_TRUE;
 			break;
+		case 'C':
+			is_txg_continue = B_TRUE;
+			break;
 		case '?':
 			(void) fprintf(stderr, gettext("invalid option '%c'\n"),
 			    optopt);
@@ -8459,6 +8483,18 @@ zpool_do_scrub(int argc, char **argv)
 		(void) fprintf(stderr, gettext("invalid option "
 		    "combination :-s and -p are mutually exclusive\n"));
 		usage(B_FALSE);
+	} else if (is_pause && is_txg_continue) {
+		(void) fprintf(stderr, gettext("invalid option "
+		    "combination :-p and -C are mutually exclusive\n"));
+		usage(B_FALSE);
+	} else if (is_stop && is_txg_continue) {
+		(void) fprintf(stderr, gettext("invalid option "
+		    "combination :-s and -C are mutually exclusive\n"));
+		usage(B_FALSE);
+	} else if (is_error_scrub && is_txg_continue) {
+		(void) fprintf(stderr, gettext("invalid option "
+		    "combination :-e and -C are mutually exclusive\n"));
+		usage(B_FALSE);
 	} else {
 		if (is_error_scrub)
 			cb.cb_type = POOL_SCAN_ERRORSCRUB;
@@ -8467,6 +8503,8 @@ zpool_do_scrub(int argc, char **argv)
 			cb.cb_scrub_cmd = POOL_SCRUB_PAUSE;
 		} else if (is_stop) {
 			cb.cb_type = POOL_SCAN_NONE;
+		} else if (is_txg_continue) {
+			cb.cb_scrub_cmd = POOL_SCRUB_FROM_LAST_TXG;
 		} else {
 			cb.cb_scrub_cmd = POOL_SCRUB_NORMAL;
 		}
@@ -9224,6 +9262,12 @@ vdev_stats_nvlist(zpool_handle_t *zhp, status_cbdata_t *cb, nvlist_t *nv,
 		}
 	}
 
+	if (cb->cb_print_dio_verify) {
+		nice_num_str_nvlist(vds, "dio_verify_errors",
+		    vs->vs_dio_verify_errors, cb->cb_literal,
+		    cb->cb_json_as_int, ZFS_NICENUM_1024);
+	}
+
 	if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT,
 	    &notpresent) == 0) {
 		nice_num_str_nvlist(vds, ZPOOL_CONFIG_NOT_PRESENT,
@@ -10010,9 +10054,8 @@ print_removal_status(zpool_handle_t *zhp, pool_removal_stat_t *prs)
 		(void) printf(gettext("Removal of %s canceled on %s"),
 		    vdev_name, ctime(&end));
 	} else {
-		uint64_t copied, total, elapsed, mins_left, hours_left;
+		uint64_t copied, total, elapsed, rate, mins_left, hours_left;
 		double fraction_done;
-		uint_t rate;
 
 		assert(prs->prs_state == DSS_SCANNING);
 
@@ -10108,9 +10151,8 @@ print_raidz_expand_status(zpool_handle_t *zhp, pool_raidz_expand_stat_t *pres)
 		    copied_buf, time_buf, ctime((time_t *)&end));
 	} else {
 		char examined_buf[7], total_buf[7], rate_buf[7];
-		uint64_t copied, total, elapsed, secs_left;
+		uint64_t copied, total, elapsed, rate, secs_left;
 		double fraction_done;
-		uint_t rate;
 
 		assert(pres->pres_state == DSS_SCANNING);
 
@@ -10975,6 +11017,7 @@ zpool_do_status(int argc, char **argv)
 
 	struct option long_options[] = {
 		{"power", no_argument, NULL, ZPOOL_OPTION_POWER},
+		{"json", no_argument, NULL, 'j'},
 		{"json-int", no_argument, NULL, ZPOOL_OPTION_JSON_NUMS_AS_INT},
 		{"json-flat-vdevs", no_argument, NULL,
 		    ZPOOL_OPTION_JSON_FLAT_VDEVS},
@@ -12583,6 +12626,7 @@ zpool_do_get(int argc, char **argv)
 	current_prop_type = cb.cb_type;
 
 	struct option long_options[] = {
+		{"json", no_argument, NULL, 'j'},
 		{"json-int", no_argument, NULL, ZPOOL_OPTION_JSON_NUMS_AS_INT},
 		{"json-pool-key-guid", no_argument, NULL,
 		    ZPOOL_OPTION_POOL_KEY_GUID},
@@ -13497,7 +13541,12 @@ zpool_do_version(int argc, char **argv)
 	int c;
 	nvlist_t *jsobj = NULL, *zfs_ver = NULL;
 	boolean_t json = B_FALSE;
-	while ((c = getopt(argc, argv, "j")) != -1) {
+
+	struct option long_options[] = {
+		{"json", no_argument, NULL, 'j'},
+	};
+
+	while ((c = getopt_long(argc, argv, "j", long_options, NULL)) != -1) {
 		switch (c) {
 		case 'j':
 			json = B_TRUE;
@@ -13613,7 +13662,7 @@ main(int argc, char **argv)
 	 * Special case '-V|--version'
 	 */
 	if ((strcmp(cmdname, "-V") == 0) || (strcmp(cmdname, "--version") == 0))
-		return (zpool_do_version(argc, argv));
+		return (zfs_version_print() != 0);
 
 	/*
 	 * Special case 'help'

cmd/ztest.c

@@ -6717,6 +6717,17 @@ out:
  *
  * Only after a full scrub has been completed is it safe to start injecting
  * data corruption.  See the comment in zfs_fault_inject().
+ *
+ * EBUSY may be returned for the following six cases.  It's the callers
+ * responsibility to handle them accordingly.
+ *
+ * Current state                Requested
+ * 1. Normal Scrub Running      Normal Scrub or Error Scrub
+ * 2. Normal Scrub Paused       Error Scrub
+ * 3. Normal Scrub Paused       Pause Normal Scrub
+ * 4. Error Scrub Running       Normal Scrub or Error Scrub
+ * 5. Error Scrub Paused        Pause Error Scrub
+ * 6. Resilvering               Anything else
  */
 static int
 ztest_scrub_impl(spa_t *spa)
@@ -8082,8 +8093,14 @@ ztest_raidz_expand_check(spa_t *spa)
 		(void) printf("verifying an interrupted raidz "
 		    "expansion using a pool scrub ...\n");
 	}
+
 	/* Will fail here if there is non-recoverable corruption detected */
-	VERIFY0(ztest_scrub_impl(spa));
+	int error = ztest_scrub_impl(spa);
+	if (error == EBUSY)
+		error = 0;
+
+	VERIFY0(error);
+
 	if (ztest_opts.zo_verbose >= 1) {
 		(void) printf("raidz expansion scrub check complete\n");
 	}

config/deb.am

@@ -58,9 +58,9 @@ deb-utils: deb-local rpm-utils-initramfs
 	pkg1=$${name}-$${version}.$${arch}.rpm; \
 	pkg2=libnvpair3-$${version}.$${arch}.rpm; \
 	pkg3=libuutil3-$${version}.$${arch}.rpm; \
-	pkg4=libzfs5-$${version}.$${arch}.rpm; \
-	pkg5=libzpool5-$${version}.$${arch}.rpm; \
-	pkg6=libzfs5-devel-$${version}.$${arch}.rpm; \
+	pkg4=libzfs6-$${version}.$${arch}.rpm; \
+	pkg5=libzpool6-$${version}.$${arch}.rpm; \
+	pkg6=libzfs6-devel-$${version}.$${arch}.rpm; \
 	pkg7=$${name}-test-$${version}.$${arch}.rpm; \
 	pkg8=$${name}-dracut-$${version}.noarch.rpm; \
 	pkg9=$${name}-initramfs-$${version}.$${arch}.rpm; \
@@ -72,7 +72,7 @@ deb-utils: deb-local rpm-utils-initramfs
 	path_prepend=`mktemp -d /tmp/intercept.XXXXXX`; \
 	echo "#!$(SHELL)" > $${path_prepend}/dh_shlibdeps; \
 	echo "`which dh_shlibdeps` -- \
-	 -xlibuutil3linux -xlibnvpair3linux -xlibzfs5linux -xlibzpool5linux" \
+	 -xlibuutil3linux -xlibnvpair3linux -xlibzfs6linux -xlibzpool6linux" \
 	 >> $${path_prepend}/dh_shlibdeps; \
 ## These -x arguments are passed to dpkg-shlibdeps, which exclude the
 ## Debianized packages from the auto-generated dependencies of the new debs,
@@ -93,13 +93,17 @@ debian:
 	cp -r contrib/debian debian; chmod +x debian/rules;
 
 native-deb-utils: native-deb-local debian
+	while [ -f debian/deb-build.lock ]; do sleep 1; done; \
+	echo "native-deb-utils" > debian/deb-build.lock; \
 	cp contrib/debian/control debian/control; \
-	$(DPKGBUILD) -b -rfakeroot -us -uc;
+	$(DPKGBUILD) -b -rfakeroot -us -uc; \
+	$(RM) -f debian/deb-build.lock
 
 native-deb-kmod: native-deb-local debian
+	while [ -f debian/deb-build.lock ]; do sleep 1; done; \
+	echo "native-deb-kmod" > debian/deb-build.lock; \
 	sh scripts/make_gitrev.sh; \
-	fakeroot debian/rules override_dh_binary-modules;
+	fakeroot debian/rules override_dh_binary-modules; \
+	$(RM) -f debian/deb-build.lock
 
 native-deb: native-deb-utils native-deb-kmod
-
-.NOTPARALLEL: native-deb native-deb-utils native-deb-kmod

config/kernel-kthread.m4

@@ -17,14 +17,21 @@ AC_DEFUN([ZFS_AC_KERNEL_KTHREAD_COMPLETE_AND_EXIT], [
 
 AC_DEFUN([ZFS_AC_KERNEL_KTHREAD_DEQUEUE_SIGNAL], [
 	dnl #
-	dnl # 5.17 API: enum pid_type * as new 4th dequeue_signal() argument,
-	dnl # 5768d8906bc23d512b1a736c1e198aa833a6daa4 ("signal: Requeue signals in the appropriate queue")
+	dnl # prehistory:
+	dnl #     int dequeue_signal(struct task_struct *task, sigset_t *mask,
+	dnl #         siginfo_t *info)
 	dnl #
-	dnl # int dequeue_signal(struct task_struct *task, sigset_t *mask, kernel_siginfo_t *info);
-	dnl # int dequeue_signal(struct task_struct *task, sigset_t *mask, kernel_siginfo_t *info, enum pid_type *type);
+	dnl # 4.20: kernel_siginfo_t introduced, replaces siginfo_t
+	dnl #     int dequeue_signal(struct task_struct *task, sigset_t *mask,
+	dnl           kernel_siginfo_t *info)
 	dnl #
-	dnl # 6.12 API: first arg struct_task* removed
-	dnl # int dequeue_signal(sigset_t *mask, kernel_siginfo_t *info, enum pid_type *type);
+	dnl # 5.17: enum pid_type introduced as 4th arg
+	dnl #     int dequeue_signal(struct task_struct *task, sigset_t *mask,
+	dnl #         kernel_siginfo_t *info, enum pid_type *type)
+	dnl #
+	dnl # 6.12: first arg struct_task* removed
+	dnl #     int dequeue_signal(sigset_t *mask, kernel_siginfo_t *info,
+	dnl #         enum pid_type *type)
 	dnl #
 	AC_MSG_CHECKING([whether dequeue_signal() takes 4 arguments])
 	ZFS_LINUX_TEST_RESULT([kthread_dequeue_signal_4arg], [
@@ -33,11 +40,11 @@ AC_DEFUN([ZFS_AC_KERNEL_KTHREAD_DEQUEUE_SIGNAL], [
 		    [dequeue_signal() takes 4 arguments])
 	], [
 		AC_MSG_RESULT(no)
-		AC_MSG_CHECKING([whether dequeue_signal() a task argument])
-		ZFS_LINUX_TEST_RESULT([kthread_dequeue_signal_3arg_task], [
+		AC_MSG_CHECKING([whether 3-arg dequeue_signal() takes a type argument])
+		ZFS_LINUX_TEST_RESULT([kthread_dequeue_signal_3arg_type], [
 			AC_MSG_RESULT(yes)
-			AC_DEFINE(HAVE_DEQUEUE_SIGNAL_3ARG_TASK, 1,
-			    [dequeue_signal() takes a task argument])
+			AC_DEFINE(HAVE_DEQUEUE_SIGNAL_3ARG_TYPE, 1,
+			    [3-arg dequeue_signal() takes a type argument])
 		], [
 			AC_MSG_RESULT(no)
 		])
@@ -56,17 +63,6 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_KTHREAD_COMPLETE_AND_EXIT], [
 ])
 
 AC_DEFUN([ZFS_AC_KERNEL_SRC_KTHREAD_DEQUEUE_SIGNAL], [
-	ZFS_LINUX_TEST_SRC([kthread_dequeue_signal_3arg_task], [
-		#include <linux/sched/signal.h>
-	], [
-		struct task_struct *task = NULL;
-		sigset_t *mask = NULL;
-		kernel_siginfo_t *info = NULL;
-		int error __attribute__ ((unused));
-
-		error = dequeue_signal(task, mask, info);
-	])
-
 	ZFS_LINUX_TEST_SRC([kthread_dequeue_signal_4arg], [
 		#include <linux/sched/signal.h>
 	], [
@@ -78,6 +74,17 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_KTHREAD_DEQUEUE_SIGNAL], [
 
 		error = dequeue_signal(task, mask, info, type);
 	])
+
+	ZFS_LINUX_TEST_SRC([kthread_dequeue_signal_3arg_type], [
+		#include <linux/sched/signal.h>
+	], [
+		sigset_t *mask = NULL;
+		kernel_siginfo_t *info = NULL;
+		enum pid_type *type = NULL;
+		int error __attribute__ ((unused));
+
+		error = dequeue_signal(mask, info, type);
+	])
 ])
 
 AC_DEFUN([ZFS_AC_KERNEL_KTHREAD], [

config/kernel-register_sysctl_table.m4

@@ -36,7 +36,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_REGISTER_SYSCTL_SZ], [
 	ZFS_LINUX_TEST_SRC([has_register_sysctl_sz], [
 		#include <linux/sysctl.h>
 	],[
-		struct ctl_table test_table[] __attribute__((unused)) = {0};
+		struct ctl_table test_table[] __attribute__((unused)) = {{}};
 		register_sysctl_sz("", test_table, 0);
 	])
 ])

config/kernel-vfs-invalidate_folio.m4

@@ -1,33 +0,0 @@
-dnl #
-dnl # Linux 5.18 uses invalidate_folio in lieu of invalidate_page
-dnl #
-AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_INVALIDATE_FOLIO], [
-	ZFS_LINUX_TEST_SRC([vfs_has_invalidate_folio], [
-		#include <linux/fs.h>
-
-		static void
-		test_invalidate_folio(struct folio *folio, size_t offset,
-		                      size_t len) {
-			(void) folio; (void) offset; (void) len;
-			return;
-		}
-
-		static const struct address_space_operations
-		    aops __attribute__ ((unused)) = {
-			.invalidate_folio	= test_invalidate_folio,
-		};
-	],[])
-])
-
-AC_DEFUN([ZFS_AC_KERNEL_VFS_INVALIDATE_FOLIO], [
-	dnl #
-	dnl # Linux 5.18 uses invalidate_folio in lieu of invalidate_page
-	dnl #
-	AC_MSG_CHECKING([whether invalidate_folio exists])
-	ZFS_LINUX_TEST_RESULT([vfs_has_invalidate_folio], [
-		AC_MSG_RESULT([yes])
-		AC_DEFINE(HAVE_VFS_INVALIDATE_FOLIO, 1, [invalidate_folio exists])
-	],[
-		AC_MSG_RESULT([no])
-	])
-])

config/kernel-vfs-migrate_folio.m4

@@ -0,0 +1,27 @@
+dnl #
+dnl # Linux 6.0 uses migrate_folio in lieu of migrate_page
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_MIGRATE_FOLIO], [
+	ZFS_LINUX_TEST_SRC([vfs_has_migrate_folio], [
+		#include <linux/fs.h>
+		#include <linux/migrate.h>
+
+		static const struct address_space_operations
+		    aops __attribute__ ((unused)) = {
+			.migrate_folio	= migrate_folio,
+		};
+	],[])
+])
+
+AC_DEFUN([ZFS_AC_KERNEL_VFS_MIGRATE_FOLIO], [
+	dnl #
+	dnl # Linux 6.0 uses migrate_folio in lieu of migrate_page
+	dnl #
+	AC_MSG_CHECKING([whether migrate_folio exists])
+	ZFS_LINUX_TEST_RESULT([vfs_has_migrate_folio], [
+		AC_MSG_RESULT([yes])
+		AC_DEFINE(HAVE_VFS_MIGRATE_FOLIO, 1, [migrate_folio exists])
+	],[
+		AC_MSG_RESULT([no])
+	])
+])

config/kernel-vfs-release_folio.m4

@@ -1,32 +0,0 @@
-dnl #
-dnl # Linux 5.19 uses release_folio in lieu of releasepage
-dnl #
-AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_RELEASE_FOLIO], [
-	ZFS_LINUX_TEST_SRC([vfs_has_release_folio], [
-		#include <linux/fs.h>
-
-		static bool
-		test_release_folio(struct folio *folio, gfp_t gfp) {
-			(void) folio; (void) gfp;
-			return (0);
-		}
-
-		static const struct address_space_operations
-		    aops __attribute__ ((unused)) = {
-			.release_folio	= test_release_folio,
-		};
-	],[])
-])
-
-AC_DEFUN([ZFS_AC_KERNEL_VFS_RELEASE_FOLIO], [
-	dnl #
-	dnl # Linux 5.19 uses release_folio in lieu of releasepage
-	dnl #
-	AC_MSG_CHECKING([whether release_folio exists])
-	ZFS_LINUX_TEST_RESULT([vfs_has_release_folio], [
-		AC_MSG_RESULT([yes])
-		AC_DEFINE(HAVE_VFS_RELEASE_FOLIO, 1, [release_folio exists])
-	],[
-		AC_MSG_RESULT([no])
-	])
-])

config/kernel.m4

@@ -77,8 +77,7 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_SRC], [
 	ZFS_AC_KERNEL_SRC_SGET
 	ZFS_AC_KERNEL_SRC_VFS_FILEMAP_DIRTY_FOLIO
 	ZFS_AC_KERNEL_SRC_VFS_READ_FOLIO
-	ZFS_AC_KERNEL_SRC_VFS_RELEASE_FOLIO
-	ZFS_AC_KERNEL_SRC_VFS_INVALIDATE_FOLIO
+	ZFS_AC_KERNEL_SRC_VFS_MIGRATE_FOLIO
 	ZFS_AC_KERNEL_SRC_VFS_FSYNC_2ARGS
 	ZFS_AC_KERNEL_SRC_VFS_DIRECT_IO
 	ZFS_AC_KERNEL_SRC_VFS_READPAGES
@@ -189,8 +188,7 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_RESULT], [
 	ZFS_AC_KERNEL_SGET
 	ZFS_AC_KERNEL_VFS_FILEMAP_DIRTY_FOLIO
 	ZFS_AC_KERNEL_VFS_READ_FOLIO
-	ZFS_AC_KERNEL_VFS_RELEASE_FOLIO
-	ZFS_AC_KERNEL_VFS_INVALIDATE_FOLIO
+	ZFS_AC_KERNEL_VFS_MIGRATE_FOLIO
 	ZFS_AC_KERNEL_VFS_FSYNC_2ARGS
 	ZFS_AC_KERNEL_VFS_DIRECT_IO
 	ZFS_AC_KERNEL_VFS_READPAGES

config/user.m4

@@ -33,7 +33,7 @@ AC_DEFUN([ZFS_AC_CONFIG_USER], [
 	ZFS_AC_CONFIG_USER_MAKEDEV_IN_MKDEV
 	ZFS_AC_CONFIG_USER_ZFSEXEC
 
-	AC_CHECK_FUNCS([execvpe issetugid mlockall strlcat strlcpy gettid])
+	AC_CHECK_FUNCS([execvpe issetugid mlockall strerror_l strlcat strlcpy gettid])
 
 	AC_SUBST(RM)
 ])

contrib/debian/Makefile.am

@@ -12,14 +12,14 @@ dist_noinst_DATA += %D%/openzfs-libpam-zfs.postinst
 dist_noinst_DATA += %D%/openzfs-libpam-zfs.prerm
 dist_noinst_DATA += %D%/openzfs-libuutil3.docs
 dist_noinst_DATA += %D%/openzfs-libuutil3.install.in
-dist_noinst_DATA += %D%/openzfs-libzfs4.docs
-dist_noinst_DATA += %D%/openzfs-libzfs4.install.in
+dist_noinst_DATA += %D%/openzfs-libzfs6.docs
+dist_noinst_DATA += %D%/openzfs-libzfs6.install.in
 dist_noinst_DATA += %D%/openzfs-libzfsbootenv1.docs
 dist_noinst_DATA += %D%/openzfs-libzfsbootenv1.install.in
 dist_noinst_DATA += %D%/openzfs-libzfs-dev.docs
 dist_noinst_DATA += %D%/openzfs-libzfs-dev.install.in
-dist_noinst_DATA += %D%/openzfs-libzpool5.docs
-dist_noinst_DATA += %D%/openzfs-libzpool5.install.in
+dist_noinst_DATA += %D%/openzfs-libzpool6.docs
+dist_noinst_DATA += %D%/openzfs-libzpool6.install.in
 dist_noinst_DATA += %D%/openzfs-python3-pyzfs.install
 dist_noinst_DATA += %D%/openzfs-zfs-dkms.config
 dist_noinst_DATA += %D%/openzfs-zfs-dkms.dkms

contrib/debian/clean

@@ -6,6 +6,6 @@ contrib/pyzfs/libzfs_core/bindings/__pycache__/
 contrib/pyzfs/pyzfs.egg-info/
 debian/openzfs-libnvpair3.install
 debian/openzfs-libuutil3.install
-debian/openzfs-libzfs4.install
+debian/openzfs-libzfs6.install
 debian/openzfs-libzfs-dev.install
-debian/openzfs-libzpool5.install
+debian/openzfs-libzpool6.install

contrib/debian/control

@@ -78,9 +78,9 @@ Architecture: linux-any
 Depends: libssl-dev | libssl1.0-dev,
          openzfs-libnvpair3 (= ${binary:Version}),
          openzfs-libuutil3 (= ${binary:Version}),
-         openzfs-libzfs4 (= ${binary:Version}),
+         openzfs-libzfs6 (= ${binary:Version}),
          openzfs-libzfsbootenv1 (= ${binary:Version}),
-         openzfs-libzpool5 (= ${binary:Version}),
+         openzfs-libzpool6 (= ${binary:Version}),
          ${misc:Depends}
 Replaces: libzfslinux-dev
 Conflicts: libzfslinux-dev
@@ -90,18 +90,18 @@ Description: OpenZFS filesystem development files for Linux
  libraries of OpenZFS filesystem.
  .
  This package includes the development files of libnvpair3, libuutil3,
- libzpool5 and libzfs4.
+ libzpool6 and libzfs6.
 
-Package: openzfs-libzfs4
+Package: openzfs-libzfs6
 Section: contrib/libs
 Architecture: linux-any
 Depends: ${misc:Depends}, ${shlibs:Depends}
 # The libcurl4 is loaded through dlopen("libcurl.so.4").
 # https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=988521
 Recommends: libcurl4
-Breaks: libzfs2, libzfs4
-Replaces: libzfs2, libzfs4, libzfs4linux
-Conflicts: libzfs4linux
+Breaks: libzfs2, libzfs4, libzfs4linux, libzfs6linux
+Replaces: libzfs2, libzfs4, libzfs4linux, libzfs6linux
+Conflicts: libzfs6linux
 Description: OpenZFS filesystem library for Linux - general support
  OpenZFS is a storage platform that encompasses the functionality of
  traditional filesystems and volume managers. It supports data checksums,
@@ -123,13 +123,13 @@ Description: OpenZFS filesystem library for Linux - label info support
  .
  The zfsbootenv library provides support for modifying ZFS label information.
 
-Package: openzfs-libzpool5
+Package: openzfs-libzpool6
 Section: contrib/libs
 Architecture: linux-any
 Depends: ${misc:Depends}, ${shlibs:Depends}
-Breaks: libzpool2, libzpool5
-Replaces: libzpool2, libzpool5, libzpool5linux
-Conflicts: libzpool5linux
+Breaks: libzpool2, libzpool5, libzpool5linux, libzpool6linux
+Replaces: libzpool2, libzpool5, libzpool5linux, libzpool6linux
+Conflicts: libzpool6linux
 Description: OpenZFS pool library for Linux
  OpenZFS is a storage platform that encompasses the functionality of
  traditional filesystems and volume managers. It supports data checksums,
@@ -246,8 +246,8 @@ Architecture: linux-any
 Pre-Depends: ${misc:Pre-Depends}
 Depends: openzfs-libnvpair3 (= ${binary:Version}),
          openzfs-libuutil3 (= ${binary:Version}),
-         openzfs-libzfs4 (= ${binary:Version}),
-         openzfs-libzpool5 (= ${binary:Version}),
+         openzfs-libzfs6 (= ${binary:Version}),
+         openzfs-libzpool6 (= ${binary:Version}),
          python3,
          ${misc:Depends},
          ${shlibs:Depends}

contrib/debian/openzfs-zfsutils.install

@@ -98,6 +98,7 @@ usr/share/man/man8/zpool-attach.8
 usr/share/man/man8/zpool-checkpoint.8
 usr/share/man/man8/zpool-clear.8
 usr/share/man/man8/zpool-create.8
+usr/share/man/man8/zpool-ddtprune.8
 usr/share/man/man8/zpool-destroy.8
 usr/share/man/man8/zpool-detach.8
 usr/share/man/man8/zpool-ddtprune.8
@@ -113,6 +114,7 @@ usr/share/man/man8/zpool-list.8
 usr/share/man/man8/zpool-offline.8
 usr/share/man/man8/zpool-online.8
 usr/share/man/man8/zpool-prefetch.8
+usr/share/man/man8/zpool-prefetch.8
 usr/share/man/man8/zpool-reguid.8
 usr/share/man/man8/zpool-remove.8
 usr/share/man/man8/zpool-reopen.8

contrib/initramfs/scripts/zfs

@@ -344,7 +344,7 @@ mount_fs()
 
 	# Need the _original_ datasets mountpoint!
 	mountpoint=$(get_fs_value "$fs" mountpoint)
-	ZFS_CMD="mount -o zfsutil -t zfs"
+	ZFS_CMD="mount.zfs -o zfsutil"
 	if [ "$mountpoint" = "legacy" ] || [ "$mountpoint" = "none" ]; then
 		# Can't use the mountpoint property. Might be one of our
 		# clones. Check the 'org.zol:mountpoint' property set in
@@ -359,9 +359,8 @@ mount_fs()
 				# isn't the root fs.
 				return 0
 			fi
-			# Don't use mount.zfs -o zfsutils for legacy mountpoint
 			if [ "$mountpoint" = "legacy" ]; then
-				ZFS_CMD="mount -t zfs"
+				ZFS_CMD="mount.zfs"
 			fi
 			# Last hail-mary: Hope 'rootmnt' is set!
 			mountpoint=""

include/libzutil.h

@@ -276,7 +276,11 @@ _LIBZUTIL_H void update_vdev_config_dev_sysfs_path(nvlist_t *nv,
  * Thread-safe strerror() for use in ZFS libraries
  */
 static inline char *zfs_strerror(int errnum) {
+#ifdef HAVE_STRERROR_L
 	return (strerror_l(errnum, uselocale(0)));
+#else
+	return (strerror(errnum));
+#endif
 }
 
 #ifdef	__cplusplus

include/os/freebsd/linux/compiler.h

@@ -1,10 +1,5 @@
 /*
- * Copyright (c) 2010 Isilon Systems, Inc.
- * Copyright (c) 2010 iXsystems, Inc.
- * Copyright (c) 2010 Panasas, Inc.
- * Copyright (c) 2013-2016 Mellanox Technologies, Ltd.
- * Copyright (c) 2015 François Tigeot
- * All rights reserved.
+ * Copyright (c) 2024 Warner Losh.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -26,76 +21,14 @@
  * 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.
- *
- * $FreeBSD$
  */
-#ifndef	_LINUX_COMPILER_H_
-#define	_LINUX_COMPILER_H_
 
-#include <sys/cdefs.h>
+/*
+ * FreeBSD's LinuxKPI compiler.h as far back as FreeBSD 12 has what we need,
+ * except zfs_fallthrough.
+ */
+#pragma once
+
+#include <compat/linuxkpi/common/include/linux/compiler.h>
 
-#define	__user
-#define	__kernel
-#define	__safe
-#define	__force
-#define	__nocast
-#define	__iomem
-#define	__chk_user_ptr(x)		((void)0)
-#define	__chk_io_ptr(x)			((void)0)
-#define	__builtin_warning(x, y...)	(1)
-#define	__acquires(x)
-#define	__releases(x)
-#define	__acquire(x)			do { } while (0)
-#define	__release(x)			do { } while (0)
-#define	__cond_lock(x, c)		(c)
-#define	__bitwise
-#define	__devinitdata
-#define	__deprecated
-#define	__init
-#define	__initconst
-#define	__devinit
-#define	__devexit
-#define	__exit
-#define	__rcu
-#define	__percpu
-#define	__weak __weak_symbol
-#define	__malloc
-#define	___stringify(...)		#__VA_ARGS__
-#define	__stringify(...)		___stringify(__VA_ARGS__)
-#define	__attribute_const__		__attribute__((__const__))
-#undef __always_inline
-#define	__always_inline			inline
-#define	noinline			__noinline
-#define	____cacheline_aligned		__aligned(CACHE_LINE_SIZE)
 #define	zfs_fallthrough			__attribute__((__fallthrough__))
-
-#if !defined(_KERNEL) && !defined(_STANDALONE)
-#define	likely(x)			__builtin_expect(!!(x), 1)
-#define	unlikely(x)			__builtin_expect(!!(x), 0)
-#endif
-#define	typeof(x)			__typeof(x)
-
-#define	uninitialized_var(x)		x = x
-#define	__maybe_unused			__unused
-#define	__always_unused			__unused
-#define	__must_check			__result_use_check
-
-#define	__printf(a, b)			__printflike(a, b)
-
-#define	barrier()			__asm__ __volatile__("": : :"memory")
-#define	___PASTE(a, b) a##b
-#define	__PASTE(a, b) ___PASTE(a, b)
-
-#define	ACCESS_ONCE(x)			(*(volatile __typeof(x) *)&(x))
-
-#define	WRITE_ONCE(x, v) do {		\
-	barrier();			\
-	ACCESS_ONCE(x) = (v);		\
-	barrier();			\
-} while (0)
-
-#define	lockless_dereference(p) READ_ONCE(p)
-
-#define	_AT(T, X)	((T)(X))
-
-#endif	/* _LINUX_COMPILER_H_ */

include/os/freebsd/spl/sys/ccompat.h

@@ -70,15 +70,6 @@ hlist_del(struct hlist_node *n)
 		n->next->pprev = n->pprev;
 }
 	/* BEGIN CSTYLED */
-#define	READ_ONCE(x) ({			\
-	__typeof(x) __var = ({		\
-		barrier();		\
-		ACCESS_ONCE(x);		\
-	});				\
-	barrier();			\
-	__var;				\
-})
-
 #define	HLIST_HEAD_INIT { }
 #define	HLIST_HEAD(name) struct hlist_head name = HLIST_HEAD_INIT
 #define	INIT_HLIST_HEAD(head) (head)->first = NULL

include/os/freebsd/spl/sys/debug.h

@@ -95,10 +95,6 @@ spl_assert(const char *buf, const char *file, const char *func, int line)
 #ifndef expect
 #define	expect(expr, value) (__builtin_expect((expr), (value)))
 #endif
-#ifndef __linux__
-#define	likely(expr)   expect((expr) != 0, 1)
-#define	unlikely(expr) expect((expr) != 0, 0)
-#endif
 
 #define	PANIC(fmt, a...)						\
 	spl_panic(__FILE__, __FUNCTION__, __LINE__, fmt, ## a)
@@ -109,7 +105,7 @@ spl_assert(const char *buf, const char *file, const char *func, int line)
 	    __FILE__, __FUNCTION__, __LINE__))
 
 #define	VERIFYF(cond, str, ...)		do {				\
-		if (unlikely(!cond))					\
+		if (unlikely(!(cond)))					\
 		    spl_panic(__FILE__, __FUNCTION__, __LINE__,		\
 		    "VERIFY(" #cond ") failed " str "\n", __VA_ARGS__);\
 	} while (0)
@@ -205,7 +201,7 @@ spl_assert(const char *buf, const char *file, const char *func, int line)
 		    "failed (%lld " #OP " %lld) " STR "\n",		\
 		    (long long)(_verify3_left),				\
 		    (long long)(_verify3_right),			\
-		    __VA_ARGS);						\
+		    __VA_ARGS__);					\
 	} while (0)
 
 #define	VERIFY3UF(LEFT, OP, RIGHT, STR, ...)	do {			\
@@ -217,7 +213,7 @@ spl_assert(const char *buf, const char *file, const char *func, int line)
 		    "failed (%llu " #OP " %llu) " STR "\n",		\
 		    (unsigned long long)(_verify3_left),		\
 		    (unsigned long long)(_verify3_right),		\
-		    __VA_ARGS);						\
+		    __VA_ARGS__);					\
 	} while (0)
 
 #define	VERIFY3PF(LEFT, OP, RIGHT, STR, ...)	do {			\

include/os/freebsd/spl/sys/sdt.h

@@ -31,9 +31,9 @@
 
 #include_next <sys/sdt.h>
 #ifdef KDTRACE_HOOKS
-/* BEGIN CSTYLED */
 SDT_PROBE_DECLARE(sdt, , , set__error);
 
+/* BEGIN CSTYLED */
 #define	SET_ERROR(err)	({ 					\
 	SDT_PROBE1(sdt, , , set__error, (uintptr_t)err);	\
 	err;							\

include/os/freebsd/spl/sys/simd.h

@@ -50,7 +50,7 @@
 #define	kfpu_fini()		do {} while (0)
 #endif
 
-#define	simd_stat_init()	0
-#define	simd_stat_fini()	0
+#define	simd_stat_init()	do {} while (0)
+#define	simd_stat_fini()	do {} while (0)
 
 #endif

include/os/freebsd/spl/sys/vnode.h

@@ -68,47 +68,30 @@ enum symfollow { NO_FOLLOW = NOFOLLOW };
 #include <vm/vm_object.h>
 
 typedef	struct vop_vector	vnodeops_t;
-#define	VOP_FID		VOP_VPTOFH
 #define	vop_fid		vop_vptofh
 #define	vop_fid_args	vop_vptofh_args
 #define	a_fid		a_fhp
 
-#define	rootvfs		(rootvnode == NULL ? NULL : rootvnode->v_mount)
-
-#ifndef IN_BASE
-static __inline int
-vn_is_readonly(vnode_t *vp)
-{
-	return (vp->v_mount->mnt_flag & MNT_RDONLY);
-}
-#endif
 #define	vn_vfswlock(vp)		(0)
 #define	vn_vfsunlock(vp)	do { } while (0)
-#define	vn_ismntpt(vp)	   \
-	((vp)->v_type == VDIR && (vp)->v_mountedhere != NULL)
-#define	vn_mountedvfs(vp)	((vp)->v_mountedhere)
+
+#ifndef IN_BASE
 #define	vn_has_cached_data(vp)	\
 	((vp)->v_object != NULL && \
 	(vp)->v_object->resident_page_count > 0)
 
-#ifndef IN_BASE
 static __inline void
 vn_flush_cached_data(vnode_t *vp, boolean_t sync)
 {
 	if (vm_object_mightbedirty(vp->v_object)) {
 		int flags = sync ? OBJPC_SYNC : 0;
-		vn_lock(vp, LK_SHARED | LK_RETRY);
 		zfs_vmobject_wlock(vp->v_object);
 		vm_object_page_clean(vp->v_object, 0, 0, flags);
 		zfs_vmobject_wunlock(vp->v_object);
-		VOP_UNLOCK(vp);
 	}
 }
 #endif
 
-#define	vn_exists(vp)		do { } while (0)
-#define	vn_invalid(vp)		do { } while (0)
-#define	vn_free(vp)		do { } while (0)
 #define	vn_matchops(vp, vops)	((vp)->v_op == &(vops))
 
 #define	VN_HOLD(v)	vref(v)
@@ -123,9 +106,6 @@ vn_flush_cached_data(vnode_t *vp, boolean_t sync)
 #define	vnevent_rename_dest(vp, dvp, name, ct)	do { } while (0)
 #define	vnevent_rename_dest_dir(vp, ct)		do { } while (0)
 
-#define	specvp(vp, rdev, type, cr)	(VN_HOLD(vp), (vp))
-#define	MANDLOCK(vp, mode)	(0)
-
 /*
  * We will use va_spare is place of Solaris' va_mask.
  * This field is initialized in zfs_setattr().

include/os/freebsd/zfs/sys/abd_os.h

@@ -26,8 +26,10 @@
 #ifndef _ABD_OS_H
 #define	_ABD_OS_H
 
+#ifdef _KERNEL
 #include <sys/vm.h>
 #include <vm/vm_page.h>
+#endif
 
 #ifdef __cplusplus
 extern "C" {
@@ -47,8 +49,10 @@ struct abd_linear {
 #endif
 };
 
+#ifdef _KERNEL
 __attribute__((malloc))
 struct abd *abd_alloc_from_pages(vm_page_t *, unsigned long, uint64_t);
+#endif
 
 #ifdef __cplusplus
 }

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

@@ -109,7 +109,7 @@ spl_assert(const char *buf, const char *file, const char *func, int line)
 	    __FILE__, __FUNCTION__, __LINE__))
 
 #define	VERIFYF(cond, str, ...)		do {				\
-		if (unlikely(!cond))					\
+		if (unlikely(!(cond)))					\
 		    spl_panic(__FILE__, __FUNCTION__, __LINE__,		\
 		    "VERIFY(" #cond ") failed " str "\n", __VA_ARGS__);\
 	} while (0)
@@ -205,7 +205,7 @@ spl_assert(const char *buf, const char *file, const char *func, int line)
 		    "failed (%lld " #OP " %lld) " STR "\n",		\
 		    (long long)(_verify3_left),				\
 		    (long long)(_verify3_right),			\
-		    __VA_ARGS);						\
+		    __VA_ARGS__);					\
 	} while (0)
 
 #define	VERIFY3UF(LEFT, OP, RIGHT, STR, ...)	do {			\
@@ -217,7 +217,7 @@ spl_assert(const char *buf, const char *file, const char *func, int line)
 		    "failed (%llu " #OP " %llu) " STR "\n",		\
 		    (unsigned long long)(_verify3_left),		\
 		    (unsigned long long)(_verify3_right),		\
-		    __VA_ARGS);						\
+		    __VA_ARGS__);					\
 	} while (0)
 
 #define	VERIFY3PF(LEFT, OP, RIGHT, STR, ...)	do {			\

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

@@ -38,8 +38,7 @@
 #include <sys/rwlock.h>
 #include <sys/wait.h>
 #include <sys/wmsum.h>
-
-typedef struct kstat_s kstat_t;
+#include <sys/kstat.h>
 
 #define	TASKQ_NAMELEN		31
 

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

@@ -42,7 +42,7 @@
 #define	TS_ZOMB				EXIT_ZOMBIE
 #define	TS_STOPPED			TASK_STOPPED
 
-typedef void (*thread_func_t)(void *);
+typedef void (*thread_func_t)(void *) __attribute__((noreturn));
 
 #define	thread_create_named(name, stk, stksize, func, arg, len,	\
     pp, state, pri)	\

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

@@ -30,6 +30,8 @@
 extern "C" {
 #endif
 
+struct abd;
+
 struct abd_scatter {
 	uint_t		abd_offset;
 	uint_t		abd_nents;
@@ -41,10 +43,8 @@ struct abd_linear {
 	struct scatterlist *abd_sgl; /* for LINEAR_PAGE */
 };
 
-typedef struct abd abd_t;
-
 typedef int abd_iter_page_func_t(struct page *, size_t, size_t, void *);
-int abd_iterate_page_func(abd_t *, size_t, size_t, abd_iter_page_func_t *,
+int abd_iterate_page_func(struct abd *, size_t, size_t, abd_iter_page_func_t *,
     void *);
 
 /*
@@ -52,11 +52,11 @@ int abd_iterate_page_func(abd_t *, size_t, size_t, abd_iter_page_func_t *,
  * Note: these are only needed to support vdev_classic. See comment in
  * vdev_disk.c.
  */
-unsigned int abd_bio_map_off(struct bio *, abd_t *, unsigned int, size_t);
-unsigned long abd_nr_pages_off(abd_t *, unsigned int, size_t);
+unsigned int abd_bio_map_off(struct bio *, struct abd *, unsigned int, size_t);
+unsigned long abd_nr_pages_off(struct abd *, unsigned int, size_t);
 
 __attribute__((malloc))
-abd_t *abd_alloc_from_pages(struct page **, unsigned long, uint64_t);
+struct abd *abd_alloc_from_pages(struct page **, unsigned long, uint64_t);
 
 #ifdef __cplusplus
 }

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

@@ -69,6 +69,7 @@ typedef struct vfs {
 	boolean_t	vfs_do_relatime;
 	boolean_t	vfs_nbmand;
 	boolean_t	vfs_do_nbmand;
+	kmutex_t	vfs_mntpt_lock;
 } vfs_t;
 
 typedef struct zfs_mnt {

include/sys/arc.h

@@ -347,6 +347,7 @@ void l2arc_fini(void);
 void l2arc_start(void);
 void l2arc_stop(void);
 void l2arc_spa_rebuild_start(spa_t *spa);
+void l2arc_spa_rebuild_stop(spa_t *spa);
 
 #ifndef _KERNEL
 extern boolean_t arc_watch;

include/sys/arc_impl.h

@@ -942,6 +942,7 @@ typedef struct arc_sums {
 	wmsum_t arcstat_evict_l2_eligible_mru;
 	wmsum_t arcstat_evict_l2_ineligible;
 	wmsum_t arcstat_evict_l2_skip;
+	wmsum_t arcstat_hash_elements;
 	wmsum_t arcstat_hash_collisions;
 	wmsum_t arcstat_hash_chains;
 	aggsum_t arcstat_size;

include/sys/brt_impl.h

@@ -86,28 +86,38 @@ typedef struct brt_vdev_phys {
 	uint64_t	bvp_savedspace;
 } brt_vdev_phys_t;
 
-typedef struct brt_vdev {
+struct brt_vdev {
+	/*
+	 * Pending changes from open contexts.
+	 */
+	kmutex_t	bv_pending_lock;
+	avl_tree_t	bv_pending_tree[TXG_SIZE];
+	/*
+	 * Protects bv_mos_*.
+	 */
+	krwlock_t	bv_mos_entries_lock ____cacheline_aligned;
+	/*
+	 * Protects all the fields starting from bv_initiated.
+	 */
+	krwlock_t	bv_lock ____cacheline_aligned;
 	/*
 	 * VDEV id.
 	 */
-	uint64_t	bv_vdevid;
-	/*
-	 * Is the structure initiated?
-	 * (bv_entcount and bv_bitmap are allocated?)
-	 */
-	boolean_t	bv_initiated;
+	uint64_t	bv_vdevid ____cacheline_aligned;
 	/*
 	 * Object number in the MOS for the entcount array and brt_vdev_phys.
 	 */
 	uint64_t	bv_mos_brtvdev;
 	/*
-	 * Object number in the MOS for the entries table.
+	 * Object number in the MOS and dnode for the entries table.
 	 */
 	uint64_t	bv_mos_entries;
+	dnode_t		*bv_mos_entries_dnode;
 	/*
-	 * Entries to sync.
+	 * Is the structure initiated?
+	 * (bv_entcount and bv_bitmap are allocated?)
 	 */
-	avl_tree_t	bv_tree;
+	boolean_t	bv_initiated;
 	/*
 	 * Does the bv_entcount[] array needs byte swapping?
 	 */
@@ -120,6 +130,26 @@ typedef struct brt_vdev {
 	 * This is the array with BRT entry count per BRT_RANGESIZE.
 	 */
 	uint16_t	*bv_entcount;
+	/*
+	 * bv_entcount[] potentially can be a bit too big to sychronize it all
+	 * when we just changed few entcounts. The fields below allow us to
+	 * track updates to bv_entcount[] array since the last sync.
+	 * A single bit in the bv_bitmap represents as many entcounts as can
+	 * fit into a single BRT_BLOCKSIZE.
+	 * For example we have 65536 entcounts in the bv_entcount array
+	 * (so the whole array is 128kB). We updated bv_entcount[2] and
+	 * bv_entcount[5]. In that case only first bit in the bv_bitmap will
+	 * be set and we will write only first BRT_BLOCKSIZE out of 128kB.
+	 */
+	ulong_t		*bv_bitmap;
+	/*
+	 * bv_entcount[] needs updating on disk.
+	 */
+	boolean_t	bv_entcount_dirty;
+	/*
+	 * brt_vdev_phys needs updating on disk.
+	 */
+	boolean_t	bv_meta_dirty;
 	/*
 	 * Sum of all bv_entcount[]s.
 	 */
@@ -133,65 +163,27 @@ typedef struct brt_vdev {
 	 */
 	uint64_t	bv_savedspace;
 	/*
-	 * brt_vdev_phys needs updating on disk.
+	 * Entries to sync.
 	 */
-	boolean_t	bv_meta_dirty;
-	/*
-	 * bv_entcount[] needs updating on disk.
-	 */
-	boolean_t	bv_entcount_dirty;
-	/*
-	 * bv_entcount[] potentially can be a bit too big to sychronize it all
-	 * when we just changed few entcounts. The fields below allow us to
-	 * track updates to bv_entcount[] array since the last sync.
-	 * A single bit in the bv_bitmap represents as many entcounts as can
-	 * fit into a single BRT_BLOCKSIZE.
-	 * For example we have 65536 entcounts in the bv_entcount array
-	 * (so the whole array is 128kB). We updated bv_entcount[2] and
-	 * bv_entcount[5]. In that case only first bit in the bv_bitmap will
-	 * be set and we will write only first BRT_BLOCKSIZE out of 128kB.
-	 */
-	ulong_t		*bv_bitmap;
-	uint64_t	bv_nblocks;
-} brt_vdev_t;
+	avl_tree_t	bv_tree;
+};
 
-/*
- * In-core brt
- */
-typedef struct brt {
-	krwlock_t	brt_lock;
-	spa_t		*brt_spa;
-#define	brt_mos		brt_spa->spa_meta_objset
-	uint64_t	brt_rangesize;
-	uint64_t	brt_usedspace;
-	uint64_t	brt_savedspace;
-	avl_tree_t	brt_pending_tree[TXG_SIZE];
-	kmutex_t	brt_pending_lock[TXG_SIZE];
-	/* Sum of all entries across all bv_trees. */
-	uint64_t	brt_nentries;
-	brt_vdev_t	*brt_vdevs;
-	uint64_t	brt_nvdevs;
-} brt_t;
-
-/* Size of bre_offset / sizeof (uint64_t). */
+/* Size of offset / sizeof (uint64_t). */
 #define	BRT_KEY_WORDS	(1)
 
+#define	BRE_OFFSET(bre)	(DVA_GET_OFFSET(&(bre)->bre_bp.blk_dva[0]))
+
 /*
  * In-core brt entry.
- * On-disk we use bre_offset as the key and bre_refcount as the value.
+ * On-disk we use ZAP with offset as the key and count as the value.
  */
 typedef struct brt_entry {
-	uint64_t	bre_offset;
-	uint64_t	bre_refcount;
 	avl_node_t	bre_node;
+	blkptr_t	bre_bp;
+	uint64_t	bre_count;
+	uint64_t	bre_pcount;
 } brt_entry_t;
 
-typedef struct brt_pending_entry {
-	blkptr_t	bpe_bp;
-	int		bpe_count;
-	avl_node_t	bpe_node;
-} brt_pending_entry_t;
-
 #ifdef	__cplusplus
 }
 #endif

include/sys/dbuf.h

@@ -171,7 +171,6 @@ typedef struct dbuf_dirty_record {
 			 * gets COW'd in a subsequent transaction group.
 			 */
 			arc_buf_t *dr_data;
-			blkptr_t dr_overridden_by;
 			override_states_t dr_override_state;
 			uint8_t dr_copies;
 			boolean_t dr_nopwrite;
@@ -179,14 +178,21 @@ typedef struct dbuf_dirty_record {
 			boolean_t dr_diowrite;
 			boolean_t dr_has_raw_params;
 
+			/* Override and raw params are mutually exclusive. */
+			union {
+				blkptr_t dr_overridden_by;
+				struct {
 					/*
-			 * If dr_has_raw_params is set, the following crypt
-			 * params will be set on the BP that's written.
+					 * If dr_has_raw_params is set, the
+					 * following crypt params will be set
+					 * on the BP that's written.
 					 */
 					boolean_t dr_byteorder;
 					uint8_t	dr_salt[ZIO_DATA_SALT_LEN];
 					uint8_t	dr_iv[ZIO_DATA_IV_LEN];
 					uint8_t	dr_mac[ZIO_DATA_MAC_LEN];
+				};
+			};
 		} dl;
 		struct dirty_lightweight_leaf {
 			/*
@@ -264,6 +270,27 @@ typedef struct dmu_buf_impl {
 	 */
 	uint8_t db_level;
 
+	/* This block was freed while a read or write was active. */
+	uint8_t db_freed_in_flight;
+
+	/*
+	 * Evict user data as soon as the dirty and reference counts are equal.
+	 */
+	uint8_t db_user_immediate_evict;
+
+	/*
+	 * dnode_evict_dbufs() or dnode_evict_bonus() tried to evict this dbuf,
+	 * but couldn't due to outstanding references.  Evict once the refcount
+	 * drops to 0.
+	 */
+	uint8_t db_pending_evict;
+
+	/* Number of TXGs in which this buffer is dirty. */
+	uint8_t db_dirtycnt;
+
+	/* The buffer was partially read.  More reads may follow. */
+	uint8_t db_partial_read;
+
 	/*
 	 * Protects db_buf's contents if they contain an indirect block or data
 	 * block of the meta-dnode. We use this lock to protect the structure of
@@ -288,6 +315,9 @@ typedef struct dmu_buf_impl {
 	 */
 	dbuf_states_t db_state;
 
+	/* In which dbuf cache this dbuf is, if any. */
+	dbuf_cached_state_t db_caching_status;
+
 	/*
 	 * Refcount accessed by dmu_buf_{hold,rele}.
 	 * If nonzero, the buffer can't be destroyed.
@@ -304,39 +334,10 @@ typedef struct dmu_buf_impl {
 	/* Link in dbuf_cache or dbuf_metadata_cache */
 	multilist_node_t db_cache_link;
 
-	/* Tells us which dbuf cache this dbuf is in, if any */
-	dbuf_cached_state_t db_caching_status;
-
 	uint64_t db_hash;
 
-	/* Data which is unique to data (leaf) blocks: */
-
 	/* User callback information. */
 	dmu_buf_user_t *db_user;
-
-	/*
-	 * Evict user data as soon as the dirty and reference
-	 * counts are equal.
-	 */
-	uint8_t db_user_immediate_evict;
-
-	/*
-	 * This block was freed while a read or write was
-	 * active.
-	 */
-	uint8_t db_freed_in_flight;
-
-	/*
-	 * dnode_evict_dbufs() or dnode_evict_bonus() tried to
-	 * evict this dbuf, but couldn't due to outstanding
-	 * references.  Evict once the refcount drops to 0.
-	 */
-	uint8_t db_pending_evict;
-
-	uint8_t db_dirtycnt;
-
-	/* The buffer was partially read.  More reads may follow. */
-	uint8_t db_partial_read;
 } dmu_buf_impl_t;
 
 #define	DBUF_HASH_MUTEX(h, idx) \
@@ -351,6 +352,8 @@ typedef struct dbuf_hash_table {
 
 typedef void (*dbuf_prefetch_fn)(void *, uint64_t, uint64_t, boolean_t);
 
+extern kmem_cache_t *dbuf_dirty_kmem_cache;
+
 uint64_t dbuf_whichblock(const struct dnode *di, const int64_t level,
     const uint64_t offset);
 

include/sys/dmu.h

@@ -381,6 +381,7 @@ typedef struct dmu_buf {
 #define	DMU_POOL_CREATION_VERSION	"creation_version"
 #define	DMU_POOL_SCAN			"scan"
 #define	DMU_POOL_ERRORSCRUB		"error_scrub"
+#define	DMU_POOL_LAST_SCRUBBED_TXG	"last_scrubbed_txg"
 #define	DMU_POOL_FREE_BPOBJ		"free_bpobj"
 #define	DMU_POOL_BPTREE_OBJ		"bptree_obj"
 #define	DMU_POOL_EMPTY_BPOBJ		"empty_bpobj"

include/sys/dsl_deadlist.h

@@ -89,7 +89,7 @@ extern int zfs_livelist_min_percent_shared;
 
 typedef int deadlist_iter_t(void *args, dsl_deadlist_entry_t *dle);
 
-void dsl_deadlist_open(dsl_deadlist_t *dl, objset_t *os, uint64_t object);
+int dsl_deadlist_open(dsl_deadlist_t *dl, objset_t *os, uint64_t object);
 void dsl_deadlist_close(dsl_deadlist_t *dl);
 void dsl_deadlist_iterate(dsl_deadlist_t *dl, deadlist_iter_t func, void *arg);
 uint64_t dsl_deadlist_alloc(objset_t *os, dmu_tx_t *tx);

include/sys/dsl_dir.h

@@ -198,7 +198,7 @@ void dsl_dir_set_reservation_sync_impl(dsl_dir_t *dd, uint64_t value,
     dmu_tx_t *tx);
 void dsl_dir_zapify(dsl_dir_t *dd, dmu_tx_t *tx);
 boolean_t dsl_dir_is_zapified(dsl_dir_t *dd);
-void dsl_dir_livelist_open(dsl_dir_t *dd, uint64_t obj);
+int dsl_dir_livelist_open(dsl_dir_t *dd, uint64_t obj);
 void dsl_dir_livelist_close(dsl_dir_t *dd);
 void dsl_dir_remove_livelist(dsl_dir_t *dd, dmu_tx_t *tx, boolean_t total);
 int dsl_dir_wait(dsl_dir_t *dd, dsl_dataset_t *ds, zfs_wait_activity_t activity,

include/sys/dsl_scan.h

@@ -179,6 +179,12 @@ typedef struct dsl_scan {
 	dsl_errorscrub_phys_t errorscrub_phys;
 } dsl_scan_t;
 
+typedef struct {
+	pool_scan_func_t func;
+	uint64_t	 txgstart;
+	uint64_t	 txgend;
+} setup_sync_arg_t;
+
 typedef struct dsl_scan_io_queue dsl_scan_io_queue_t;
 
 void scan_init(void);
@@ -189,7 +195,8 @@ void dsl_scan_setup_sync(void *, dmu_tx_t *);
 void dsl_scan_fini(struct dsl_pool *dp);
 void dsl_scan_sync(struct dsl_pool *, dmu_tx_t *);
 int dsl_scan_cancel(struct dsl_pool *);
-int dsl_scan(struct dsl_pool *, pool_scan_func_t);
+int dsl_scan(struct dsl_pool *, pool_scan_func_t, uint64_t starttxg,
+    uint64_t txgend);
 void dsl_scan_assess_vdev(struct dsl_pool *dp, vdev_t *vd);
 boolean_t dsl_scan_scrubbing(const struct dsl_pool *dp);
 boolean_t dsl_errorscrubbing(const struct dsl_pool *dp);

include/sys/fm/fs/zfs.h

@@ -42,7 +42,8 @@ extern "C" {
 #define	FM_EREPORT_ZFS_DATA			"data"
 #define	FM_EREPORT_ZFS_DELAY			"delay"
 #define	FM_EREPORT_ZFS_DEADMAN			"deadman"
-#define	FM_EREPORT_ZFS_DIO_VERIFY		"dio_verify"
+#define	FM_EREPORT_ZFS_DIO_VERIFY_WR		"dio_verify_wr"
+#define	FM_EREPORT_ZFS_DIO_VERIFY_RD		"dio_verify_rd"
 #define	FM_EREPORT_ZFS_POOL			"zpool"
 #define	FM_EREPORT_ZFS_DEVICE_UNKNOWN		"vdev.unknown"
 #define	FM_EREPORT_ZFS_DEVICE_OPEN_FAILED	"vdev.open_failed"

include/sys/fs/zfs.h

@@ -265,6 +265,7 @@ typedef enum {
 	ZPOOL_PROP_DEDUP_TABLE_SIZE,
 	ZPOOL_PROP_DEDUP_TABLE_QUOTA,
 	ZPOOL_PROP_DEDUPCACHED,
+	ZPOOL_PROP_LAST_SCRUBBED_TXG,
 	ZPOOL_NUM_PROPS
 } zpool_prop_t;
 
@@ -1088,6 +1089,7 @@ typedef enum pool_scan_func {
 typedef enum pool_scrub_cmd {
 	POOL_SCRUB_NORMAL = 0,
 	POOL_SCRUB_PAUSE,
+	POOL_SCRUB_FROM_LAST_TXG,
 	POOL_SCRUB_FLAGS_END
 } pool_scrub_cmd_t;
 

include/sys/spa.h

@@ -53,6 +53,7 @@ extern "C" {
 /*
  * Forward references that lots of things need.
  */
+typedef struct brt_vdev brt_vdev_t;
 typedef struct spa spa_t;
 typedef struct vdev vdev_t;
 typedef struct metaslab metaslab_t;
@@ -821,6 +822,8 @@ extern void spa_l2cache_drop(spa_t *spa);
 
 /* scanning */
 extern int spa_scan(spa_t *spa, pool_scan_func_t func);
+extern int spa_scan_range(spa_t *spa, pool_scan_func_t func, uint64_t txgstart,
+    uint64_t txgend);
 extern int spa_scan_stop(spa_t *spa);
 extern int spa_scrub_pause_resume(spa_t *spa, pool_scrub_cmd_t flag);
 
@@ -1079,6 +1082,7 @@ extern uint64_t spa_get_deadman_failmode(spa_t *spa);
 extern void spa_set_deadman_failmode(spa_t *spa, const char *failmode);
 extern boolean_t spa_suspended(spa_t *spa);
 extern uint64_t spa_bootfs(spa_t *spa);
+extern uint64_t spa_get_last_scrubbed_txg(spa_t *spa);
 extern uint64_t spa_delegation(spa_t *spa);
 extern objset_t *spa_meta_objset(spa_t *spa);
 extern space_map_t *spa_syncing_log_sm(spa_t *spa);

include/sys/spa_impl.h

@@ -318,6 +318,7 @@ struct spa {
 	uint64_t	spa_scan_pass_scrub_spent_paused; /* total paused */
 	uint64_t	spa_scan_pass_exam;	/* examined bytes per pass */
 	uint64_t	spa_scan_pass_issued;	/* issued bytes per pass */
+	uint64_t	spa_scrubbed_last_txg;	/* last txg scrubbed */
 
 	/* error scrub pause time in milliseconds */
 	uint64_t	spa_scan_pass_errorscrub_pause;
@@ -412,8 +413,12 @@ struct spa {
 	uint64_t	spa_dedup_dspace;	/* Cache get_dedup_dspace() */
 	uint64_t	spa_dedup_checksum;	/* default dedup checksum */
 	uint64_t	spa_dspace;		/* dspace in normal class */
+	uint64_t	spa_rdspace;		/* raw (non-dedup) --//-- */
 	boolean_t	spa_active_ddt_prune;	/* ddt prune process active */
-	struct brt	*spa_brt;		/* in-core BRT */
+	brt_vdev_t	**spa_brt_vdevs;	/* array of per-vdev BRTs */
+	uint64_t	spa_brt_nvdevs;		/* number of vdevs in BRT */
+	uint64_t	spa_brt_rangesize;	/* pool's BRT range size */
+	krwlock_t	spa_brt_lock;		/* Protects brt_vdevs/nvdevs */
 	kmutex_t	spa_vdev_top_lock;	/* dueling offline/remove */
 	kmutex_t	spa_proc_lock;		/* protects spa_proc* */
 	kcondvar_t	spa_proc_cv;		/* spa_proc_state transitions */

include/sys/vdev_raidz.h

@@ -57,7 +57,7 @@ void vdev_raidz_reconstruct(struct raidz_map *, const int *, int);
 void vdev_raidz_child_done(zio_t *);
 void vdev_raidz_io_done(zio_t *);
 void vdev_raidz_checksum_error(zio_t *, struct raidz_col *, abd_t *);
-struct raidz_row *vdev_raidz_row_alloc(int);
+struct raidz_row *vdev_raidz_row_alloc(int, zio_t *);
 void vdev_raidz_reflow_copy_scratch(spa_t *);
 void raidz_dtl_reassessed(vdev_t *);
 

include/sys/zap.h

@@ -223,11 +223,15 @@ int zap_lookup_norm(objset_t *ds, uint64_t zapobj, const char *name,
     boolean_t *normalization_conflictp);
 int zap_lookup_uint64(objset_t *os, uint64_t zapobj, const uint64_t *key,
     int key_numints, uint64_t integer_size, uint64_t num_integers, void *buf);
+int zap_lookup_uint64_by_dnode(dnode_t *dn, const uint64_t *key,
+    int key_numints, uint64_t integer_size, uint64_t num_integers, void *buf);
 int zap_contains(objset_t *ds, uint64_t zapobj, const char *name);
 int zap_prefetch(objset_t *os, uint64_t zapobj, const char *name);
 int zap_prefetch_object(objset_t *os, uint64_t zapobj);
 int zap_prefetch_uint64(objset_t *os, uint64_t zapobj, const uint64_t *key,
     int key_numints);
+int zap_prefetch_uint64_by_dnode(dnode_t *dn, const uint64_t *key,
+    int key_numints);
 
 int zap_lookup_by_dnode(dnode_t *dn, const char *name,
     uint64_t integer_size, uint64_t num_integers, void *buf);
@@ -236,9 +240,6 @@ int zap_lookup_norm_by_dnode(dnode_t *dn, const char *name,
     matchtype_t mt, char *realname, int rn_len,
     boolean_t *ncp);
 
-int zap_count_write_by_dnode(dnode_t *dn, const char *name,
-    int add, zfs_refcount_t *towrite, zfs_refcount_t *tooverwrite);
-
 /*
  * Create an attribute with the given name and value.
  *

include/sys/zio.h

@@ -208,25 +208,25 @@ typedef uint64_t zio_flag_t;
 #define	ZIO_FLAG_PROBE		(1ULL << 16)
 #define	ZIO_FLAG_TRYHARD	(1ULL << 17)
 #define	ZIO_FLAG_OPTIONAL	(1ULL << 18)
-
+#define	ZIO_FLAG_DIO_READ	(1ULL << 19)
 #define	ZIO_FLAG_VDEV_INHERIT	(ZIO_FLAG_DONT_QUEUE - 1)
 
 	/*
 	 * Flags not inherited by any children.
 	 */
-#define	ZIO_FLAG_DONT_QUEUE	(1ULL << 19)	/* must be first for INHERIT */
-#define	ZIO_FLAG_DONT_PROPAGATE	(1ULL << 20)
-#define	ZIO_FLAG_IO_BYPASS	(1ULL << 21)
-#define	ZIO_FLAG_IO_REWRITE	(1ULL << 22)
-#define	ZIO_FLAG_RAW_COMPRESS	(1ULL << 23)
-#define	ZIO_FLAG_RAW_ENCRYPT	(1ULL << 24)
-#define	ZIO_FLAG_GANG_CHILD	(1ULL << 25)
-#define	ZIO_FLAG_DDT_CHILD	(1ULL << 26)
-#define	ZIO_FLAG_GODFATHER	(1ULL << 27)
-#define	ZIO_FLAG_NOPWRITE	(1ULL << 28)
-#define	ZIO_FLAG_REEXECUTED	(1ULL << 29)
-#define	ZIO_FLAG_DELEGATED	(1ULL << 30)
-#define	ZIO_FLAG_DIO_CHKSUM_ERR	(1ULL << 31)
+#define	ZIO_FLAG_DONT_QUEUE	(1ULL << 20)	/* must be first for INHERIT */
+#define	ZIO_FLAG_DONT_PROPAGATE	(1ULL << 21)
+#define	ZIO_FLAG_IO_BYPASS	(1ULL << 22)
+#define	ZIO_FLAG_IO_REWRITE	(1ULL << 23)
+#define	ZIO_FLAG_RAW_COMPRESS	(1ULL << 24)
+#define	ZIO_FLAG_RAW_ENCRYPT	(1ULL << 25)
+#define	ZIO_FLAG_GANG_CHILD	(1ULL << 26)
+#define	ZIO_FLAG_DDT_CHILD	(1ULL << 27)
+#define	ZIO_FLAG_GODFATHER	(1ULL << 28)
+#define	ZIO_FLAG_NOPWRITE	(1ULL << 29)
+#define	ZIO_FLAG_REEXECUTED	(1ULL << 30)
+#define	ZIO_FLAG_DELEGATED	(1ULL << 31)
+#define	ZIO_FLAG_DIO_CHKSUM_ERR	(1ULL << 32)
 
 #define	ZIO_ALLOCATOR_NONE	(-1)
 #define	ZIO_HAS_ALLOCATOR(zio)	((zio)->io_allocator != ZIO_ALLOCATOR_NONE)
@@ -647,6 +647,7 @@ extern void zio_vdev_io_redone(zio_t *zio);
 extern void zio_change_priority(zio_t *pio, zio_priority_t priority);
 
 extern void zio_checksum_verified(zio_t *zio);
+extern void zio_dio_chksum_verify_error_report(zio_t *zio);
 extern int zio_worst_error(int e1, int e2);
 
 extern enum zio_checksum zio_checksum_select(enum zio_checksum child,

lib/libspl/atomic.c

@@ -35,7 +35,6 @@
 		(void) __atomic_add_fetch(target, 1, __ATOMIC_SEQ_CST);	\
 	}
 
-/* BEGIN CSTYLED */
 ATOMIC_INC(8, uint8_t)
 ATOMIC_INC(16, uint16_t)
 ATOMIC_INC(32, uint32_t)
@@ -44,7 +43,6 @@ ATOMIC_INC(uchar, uchar_t)
 ATOMIC_INC(ushort, ushort_t)
 ATOMIC_INC(uint, uint_t)
 ATOMIC_INC(ulong, ulong_t)
-/* END CSTYLED */
 
 
 #define	ATOMIC_DEC(name, type) \
@@ -53,7 +51,6 @@ ATOMIC_INC(ulong, ulong_t)
 		(void) __atomic_sub_fetch(target, 1, __ATOMIC_SEQ_CST);	\
 	}
 
-/* BEGIN CSTYLED */
 ATOMIC_DEC(8, uint8_t)
 ATOMIC_DEC(16, uint16_t)
 ATOMIC_DEC(32, uint32_t)
@@ -62,7 +59,6 @@ ATOMIC_DEC(uchar, uchar_t)
 ATOMIC_DEC(ushort, ushort_t)
 ATOMIC_DEC(uint, uint_t)
 ATOMIC_DEC(ulong, ulong_t)
-/* END CSTYLED */
 
 
 #define	ATOMIC_ADD(name, type1, type2) \
@@ -77,7 +73,6 @@ atomic_add_ptr(volatile void *target, ssize_t bits)
 	(void) __atomic_add_fetch((void **)target, bits, __ATOMIC_SEQ_CST);
 }
 
-/* BEGIN CSTYLED */
 ATOMIC_ADD(8, uint8_t, int8_t)
 ATOMIC_ADD(16, uint16_t, int16_t)
 ATOMIC_ADD(32, uint32_t, int32_t)
@@ -86,7 +81,6 @@ ATOMIC_ADD(char, uchar_t, signed char)
 ATOMIC_ADD(short, ushort_t, short)
 ATOMIC_ADD(int, uint_t, int)
 ATOMIC_ADD(long, ulong_t, long)
-/* END CSTYLED */
 
 
 #define	ATOMIC_SUB(name, type1, type2) \
@@ -101,7 +95,6 @@ atomic_sub_ptr(volatile void *target, ssize_t bits)
 	(void) __atomic_sub_fetch((void **)target, bits, __ATOMIC_SEQ_CST);
 }
 
-/* BEGIN CSTYLED */
 ATOMIC_SUB(8, uint8_t, int8_t)
 ATOMIC_SUB(16, uint16_t, int16_t)
 ATOMIC_SUB(32, uint32_t, int32_t)
@@ -110,7 +103,6 @@ ATOMIC_SUB(char, uchar_t, signed char)
 ATOMIC_SUB(short, ushort_t, short)
 ATOMIC_SUB(int, uint_t, int)
 ATOMIC_SUB(long, ulong_t, long)
-/* END CSTYLED */
 
 
 #define	ATOMIC_OR(name, type) \
@@ -119,7 +111,6 @@ ATOMIC_SUB(long, ulong_t, long)
 		(void) __atomic_or_fetch(target, bits, __ATOMIC_SEQ_CST); \
 	}
 
-/* BEGIN CSTYLED */
 ATOMIC_OR(8, uint8_t)
 ATOMIC_OR(16, uint16_t)
 ATOMIC_OR(32, uint32_t)
@@ -128,7 +119,6 @@ ATOMIC_OR(uchar, uchar_t)
 ATOMIC_OR(ushort, ushort_t)
 ATOMIC_OR(uint, uint_t)
 ATOMIC_OR(ulong, ulong_t)
-/* END CSTYLED */
 
 
 #define	ATOMIC_AND(name, type) \
@@ -137,7 +127,6 @@ ATOMIC_OR(ulong, ulong_t)
 		(void) __atomic_and_fetch(target, bits, __ATOMIC_SEQ_CST); \
 	}
 
-/* BEGIN CSTYLED */
 ATOMIC_AND(8, uint8_t)
 ATOMIC_AND(16, uint16_t)
 ATOMIC_AND(32, uint32_t)
@@ -146,7 +135,6 @@ ATOMIC_AND(uchar, uchar_t)
 ATOMIC_AND(ushort, ushort_t)
 ATOMIC_AND(uint, uint_t)
 ATOMIC_AND(ulong, ulong_t)
-/* END CSTYLED */
 
 
 /*
@@ -159,7 +147,6 @@ ATOMIC_AND(ulong, ulong_t)
 		return (__atomic_add_fetch(target, 1, __ATOMIC_SEQ_CST)); \
 	}
 
-/* BEGIN CSTYLED */
 ATOMIC_INC_NV(8, uint8_t)
 ATOMIC_INC_NV(16, uint16_t)
 ATOMIC_INC_NV(32, uint32_t)
@@ -168,7 +155,6 @@ ATOMIC_INC_NV(uchar, uchar_t)
 ATOMIC_INC_NV(ushort, ushort_t)
 ATOMIC_INC_NV(uint, uint_t)
 ATOMIC_INC_NV(ulong, ulong_t)
-/* END CSTYLED */
 
 
 #define	ATOMIC_DEC_NV(name, type) \
@@ -177,7 +163,6 @@ ATOMIC_INC_NV(ulong, ulong_t)
 		return (__atomic_sub_fetch(target, 1, __ATOMIC_SEQ_CST)); \
 	}
 
-/* BEGIN CSTYLED */
 ATOMIC_DEC_NV(8, uint8_t)
 ATOMIC_DEC_NV(16, uint16_t)
 ATOMIC_DEC_NV(32, uint32_t)
@@ -186,7 +171,6 @@ ATOMIC_DEC_NV(uchar, uchar_t)
 ATOMIC_DEC_NV(ushort, ushort_t)
 ATOMIC_DEC_NV(uint, uint_t)
 ATOMIC_DEC_NV(ulong, ulong_t)
-/* END CSTYLED */
 
 
 #define	ATOMIC_ADD_NV(name, type1, type2) \
@@ -201,7 +185,6 @@ atomic_add_ptr_nv(volatile void *target, ssize_t bits)
 	return (__atomic_add_fetch((void **)target, bits, __ATOMIC_SEQ_CST));
 }
 
-/* BEGIN CSTYLED */
 ATOMIC_ADD_NV(8, uint8_t, int8_t)
 ATOMIC_ADD_NV(16, uint16_t, int16_t)
 ATOMIC_ADD_NV(32, uint32_t, int32_t)
@@ -210,7 +193,6 @@ ATOMIC_ADD_NV(char, uchar_t, signed char)
 ATOMIC_ADD_NV(short, ushort_t, short)
 ATOMIC_ADD_NV(int, uint_t, int)
 ATOMIC_ADD_NV(long, ulong_t, long)
-/* END CSTYLED */
 
 
 #define	ATOMIC_SUB_NV(name, type1, type2) \
@@ -225,7 +207,6 @@ atomic_sub_ptr_nv(volatile void *target, ssize_t bits)
 	return (__atomic_sub_fetch((void **)target, bits, __ATOMIC_SEQ_CST));
 }
 
-/* BEGIN CSTYLED */
 ATOMIC_SUB_NV(8, uint8_t, int8_t)
 ATOMIC_SUB_NV(char, uchar_t, signed char)
 ATOMIC_SUB_NV(16, uint16_t, int16_t)
@@ -234,7 +215,6 @@ ATOMIC_SUB_NV(32, uint32_t, int32_t)
 ATOMIC_SUB_NV(int, uint_t, int)
 ATOMIC_SUB_NV(long, ulong_t, long)
 ATOMIC_SUB_NV(64, uint64_t, int64_t)
-/* END CSTYLED */
 
 
 #define	ATOMIC_OR_NV(name, type) \
@@ -243,7 +223,6 @@ ATOMIC_SUB_NV(64, uint64_t, int64_t)
 		return (__atomic_or_fetch(target, bits, __ATOMIC_SEQ_CST)); \
 	}
 
-/* BEGIN CSTYLED */
 ATOMIC_OR_NV(8, uint8_t)
 ATOMIC_OR_NV(16, uint16_t)
 ATOMIC_OR_NV(32, uint32_t)
@@ -252,7 +231,6 @@ ATOMIC_OR_NV(uchar, uchar_t)
 ATOMIC_OR_NV(ushort, ushort_t)
 ATOMIC_OR_NV(uint, uint_t)
 ATOMIC_OR_NV(ulong, ulong_t)
-/* END CSTYLED */
 
 
 #define	ATOMIC_AND_NV(name, type) \
@@ -261,7 +239,6 @@ ATOMIC_OR_NV(ulong, ulong_t)
 		return (__atomic_and_fetch(target, bits, __ATOMIC_SEQ_CST)); \
 	}
 
-/* BEGIN CSTYLED */
 ATOMIC_AND_NV(8, uint8_t)
 ATOMIC_AND_NV(16, uint16_t)
 ATOMIC_AND_NV(32, uint32_t)
@@ -270,7 +247,6 @@ ATOMIC_AND_NV(uchar, uchar_t)
 ATOMIC_AND_NV(ushort, ushort_t)
 ATOMIC_AND_NV(uint, uint_t)
 ATOMIC_AND_NV(ulong, ulong_t)
-/* END CSTYLED */
 
 
 /*
@@ -300,7 +276,6 @@ atomic_cas_ptr(volatile void *target, void *exp, void *des)
 	return (exp);
 }
 
-/* BEGIN CSTYLED */
 ATOMIC_CAS(8, uint8_t)
 ATOMIC_CAS(16, uint16_t)
 ATOMIC_CAS(32, uint32_t)
@@ -309,7 +284,6 @@ ATOMIC_CAS(uchar, uchar_t)
 ATOMIC_CAS(ushort, ushort_t)
 ATOMIC_CAS(uint, uint_t)
 ATOMIC_CAS(ulong, ulong_t)
-/* END CSTYLED */
 
 
 /*
@@ -322,7 +296,6 @@ ATOMIC_CAS(ulong, ulong_t)
 		return (__atomic_exchange_n(target, bits, __ATOMIC_SEQ_CST)); \
 	}
 
-/* BEGIN CSTYLED */
 ATOMIC_SWAP(8, uint8_t)
 ATOMIC_SWAP(16, uint16_t)
 ATOMIC_SWAP(32, uint32_t)
@@ -331,7 +304,6 @@ ATOMIC_SWAP(uchar, uchar_t)
 ATOMIC_SWAP(ushort, ushort_t)
 ATOMIC_SWAP(uint, uint_t)
 ATOMIC_SWAP(ulong, ulong_t)
-/* END CSTYLED */
 
 void *
 atomic_swap_ptr(volatile void *target, void *bits)

lib/libspl/backtrace.c

@@ -25,19 +25,32 @@
 
 #include <sys/backtrace.h>
 #include <sys/types.h>
+#include <sys/debug.h>
 #include <unistd.h>
 
 /*
- * libspl_backtrace() must be safe to call from inside a signal hander. This
- * mostly means it must not allocate, and so we can't use things like printf.
+ * Output helpers. libspl_backtrace() must not block, must be thread-safe and
+ * must be safe to call from a signal handler. At least, that means not having
+ * printf, so we end up having to call write() directly on the fd. That's
+ * awkward, as we always have to pass through a length, and some systems will
+ * complain if we don't consume the return. So we have some macros to make
+ * things a little more palatable.
  */
+#define	spl_bt_write_n(fd, s, n) \
+	do { ssize_t r __maybe_unused = write(fd, s, n); } while (0)
+#define	spl_bt_write(fd, s)		spl_bt_write_n(fd, s, sizeof (s))
 
 #if defined(HAVE_LIBUNWIND)
 #define	UNW_LOCAL_ONLY
 #include <libunwind.h>
 
+/*
+ * Convert `v` to ASCII hex characters. The bottom `n` nybbles (4-bits ie one
+ * hex digit) will be written, up to `buflen`. The buffer will not be
+ * null-terminated. Returns the number of digits written.
+ */
 static size_t
-libspl_u64_to_hex_str(uint64_t v, size_t digits, char *buf, size_t buflen)
+spl_bt_u64_to_hex_str(uint64_t v, size_t n, char *buf, size_t buflen)
 {
 	static const char hexdigits[] = {
 	    '0', '1', '2', '3', '4', '5', '6', '7',
@@ -45,10 +58,10 @@ libspl_u64_to_hex_str(uint64_t v, size_t digits, char *buf, size_t buflen)
 	};
 
 	size_t pos = 0;
-	boolean_t want = (digits == 0);
+	boolean_t want = (n == 0);
 	for (int i = 15; i >= 0; i--) {
 		const uint64_t d = v >> (i * 4) & 0xf;
-		if (!want && (d != 0 || digits > i))
+		if (!want && (d != 0 || n > i))
 			want = B_TRUE;
 		if (want) {
 			buf[pos++] = hexdigits[d];
@@ -62,40 +75,181 @@ libspl_u64_to_hex_str(uint64_t v, size_t digits, char *buf, size_t buflen)
 void
 libspl_backtrace(int fd)
 {
-	ssize_t ret __attribute__((unused));
 	unw_context_t uc;
 	unw_cursor_t cp;
-	unw_word_t loc;
+	unw_word_t v;
 	char buf[128];
 	size_t n;
+	int err;
 
-	ret = write(fd, "Call trace:\n", 12);
+	/* Snapshot the current frame and state. */
 	unw_getcontext(&uc);
+
+	/*
+	 * TODO: walk back to the frame that tripped the assertion / the place
+	 *       where the signal was recieved.
+	 */
+
+	/*
+	 * Register dump. We're going to loop over all the registers in the
+	 * top frame, and show them, with names, in a nice three-column
+	 * layout, which keeps us within 80 columns.
+	 */
+	spl_bt_write(fd, "Registers:\n");
+
+	/* Initialise a frame cursor, starting at the current frame */
 	unw_init_local(&cp, &uc);
-	while (unw_step(&cp) > 0) {
-		unw_get_reg(&cp, UNW_REG_IP, &loc);
-		ret = write(fd, "  [0x", 5);
-		n = libspl_u64_to_hex_str(loc, 10, buf, sizeof (buf));
-		ret = write(fd, buf, n);
-		ret = write(fd, "] ", 2);
-		unw_get_proc_name(&cp, buf, sizeof (buf), &loc);
-		for (n = 0; n < sizeof (buf) && buf[n] != '\0'; n++) {}
-		ret = write(fd, buf, n);
-		ret = write(fd, "+0x", 3);
-		n = libspl_u64_to_hex_str(loc, 2, buf, sizeof (buf));
-		ret = write(fd, buf, n);
-#ifdef HAVE_LIBUNWIND_ELF
-		ret = write(fd, " (in ", 5);
-		unw_get_elf_filename(&cp, buf, sizeof (buf), &loc);
-		for (n = 0; n < sizeof (buf) && buf[n] != '\0'; n++) {}
-		ret = write(fd, buf, n);
-		ret = write(fd, " +0x", 4);
-		n = libspl_u64_to_hex_str(loc, 2, buf, sizeof (buf));
-		ret = write(fd, buf, n);
-		ret = write(fd, ")", 1);
-#endif
-		ret = write(fd, "\n", 1);
+
+	/*
+	 * libunwind's list of possible registers for this architecture is an
+	 * enum, unw_regnum_t. UNW_TDEP_LAST_REG is the highest-numbered
+	 * register in that list, however, not all register numbers in this
+	 * range are defined by the architecture, and not all defined registers
+	 * will be present on every implementation of that architecture.
+	 * Moreover, libunwind provides nice names for most, but not all
+	 * registers, but these are hardcoded; a name being available does not
+	 * mean that register is available.
+	 *
+	 * So, we have to pull this all together here. We try to get the value
+	 * of every possible register. If we get a value for it, then the
+	 * register must exist, and so we get its name. If libunwind has no
+	 * name for it, we synthesize something. These cases should be rare,
+	 * and they're usually for uninteresting or niche registers, so it
+	 * shouldn't really matter. We can see the value, and that's the main
+	 * thing.
+	 */
+	uint_t cols = 0;
+	for (uint_t regnum = 0; regnum <= UNW_TDEP_LAST_REG; regnum++) {
+		/*
+		 * Get the value. Any error probably means the register
+		 * doesn't exist, and we skip it.
+		 */
+		if (unw_get_reg(&cp, regnum, &v) < 0)
+			continue;
+
+		/*
+		 * Register name. If libunwind doesn't have a name for it,
+		 * it will return "???". As a shortcut, we just treat '?'
+		 * is an alternate end-of-string character.
+		 */
+		const char *name = unw_regname(regnum);
+		for (n = 0; name[n] != '\0' && name[n] != '?'; n++) {}
+		if (n == 0) {
+			/*
+			 * No valid name, so make one of the form "?xx", where
+			 * "xx" is the two-char hex of libunwind's register
+			 * number.
+			 */
+			buf[0] = '?';
+			n = spl_bt_u64_to_hex_str(regnum, 2,
+			    &buf[1], sizeof (buf)-1) + 1;
+			name = buf;
 		}
+
+		/*
+		 * Two spaces of padding before each column, plus extra
+		 * spaces to align register names shorter than three chars.
+		 */
+		spl_bt_write_n(fd, "      ", 5-MIN(n, 3));
+
+		/* Register name and column punctuation */
+		spl_bt_write_n(fd, name, n);
+		spl_bt_write(fd, ": 0x");
+
+		/*
+		 * Convert register value (from unw_get_reg()) to hex. We're
+		 * assuming that all registers are 64-bits wide, which is
+		 * probably fine for any general-purpose registers on any
+		 * machine currently in use. A more generic way would be to
+		 * look at the width of unw_word_t, but that would also
+		 * complicate the column code a bit. This is fine.
+		 */
+		n = spl_bt_u64_to_hex_str(v, 16, buf, sizeof (buf));
+		spl_bt_write_n(fd, buf, n);
+
+		/* Every third column, emit a newline */
+		if (!(++cols % 3))
+			spl_bt_write(fd, "\n");
+	}
+
+	/* If we finished before the third column, emit a newline. */
+	if (cols % 3)
+		spl_bt_write(fd, "\n");
+
+	/* Now the main event, the backtrace. */
+	spl_bt_write(fd, "Call trace:\n");
+
+	/* Reset the cursor to the top again. */
+	unw_init_local(&cp, &uc);
+
+	do {
+		/*
+		 * Getting the IP should never fail; libunwind handles it
+		 * specially, because its used a lot internally. Still, no
+		 * point being silly about it, as the last thing we want is
+		 * our crash handler to crash. So if it ever does fail, we'll
+		 * show an error line, but keep going to the next frame.
+		 */
+		if (unw_get_reg(&cp, UNW_REG_IP, &v) < 0) {
+			spl_bt_write(fd, "  [couldn't get IP register; "
+			    "corrupt frame?]");
+			continue;
+		}
+
+		/* IP & punctuation */
+		n = spl_bt_u64_to_hex_str(v, 16, buf, sizeof (buf));
+		spl_bt_write(fd, "  [0x");
+		spl_bt_write_n(fd, buf, n);
+		spl_bt_write(fd, "] ");
+
+		/*
+		 * Function ("procedure") name for the current frame. `v`
+		 * receives the offset from the named function to the IP, which
+		 * we show as a "+offset" suffix.
+		 *
+		 * If libunwind can't determine the name, we just show "???"
+		 * instead. We've already displayed the IP above; that will
+		 * have to do.
+		 *
+		 * unw_get_proc_name() will return ENOMEM if the buffer is too
+		 * small, instead truncating the name. So we treat that as a
+		 * success and use whatever is in the buffer.
+		 */
+		err = unw_get_proc_name(&cp, buf, sizeof (buf), &v);
+		if (err == 0 || err == -UNW_ENOMEM) {
+			for (n = 0; n < sizeof (buf) && buf[n] != '\0'; n++) {}
+			spl_bt_write_n(fd, buf, n);
+
+			/* Offset from proc name */
+			spl_bt_write(fd, "+0x");
+			n = spl_bt_u64_to_hex_str(v, 2, buf, sizeof (buf));
+			spl_bt_write_n(fd, buf, n);
+		} else
+			spl_bt_write(fd, "???");
+
+#ifdef HAVE_LIBUNWIND_ELF
+		/*
+		 * Newer libunwind has unw_get_elf_filename(), which gets
+		 * the name of the ELF object that the frame was executing in.
+		 * Like `unw_get_proc_name()`, `v` recieves the offset within
+		 * the file, and UNW_ENOMEM indicates that a truncate filename
+		 * was left in the buffer.
+		 */
+		err = unw_get_elf_filename(&cp, buf, sizeof (buf), &v);
+		if (err == 0 || err == -UNW_ENOMEM) {
+			for (n = 0; n < sizeof (buf) && buf[n] != '\0'; n++) {}
+			spl_bt_write(fd, " (in ");
+			spl_bt_write_n(fd, buf, n);
+
+			/* Offset within file */
+			spl_bt_write(fd, " +0x");
+			n = spl_bt_u64_to_hex_str(v, 2, buf, sizeof (buf));
+			spl_bt_write_n(fd, buf, n);
+			spl_bt_write(fd, ")");
+		}
+#endif
+		spl_bt_write(fd, "\n");
+	} while (unw_step(&cp) > 0);
 }
 #elif defined(HAVE_BACKTRACE)
 #include <execinfo.h>
@@ -103,15 +257,12 @@ libspl_backtrace(int fd)
 void
 libspl_backtrace(int fd)
 {
-	ssize_t ret __attribute__((unused));
 	void *btptrs[64];
 	size_t nptrs = backtrace(btptrs, 64);
-	ret = write(fd, "Call trace:\n", 12);
+	spl_bt_write(fd, "Call trace:\n");
 	backtrace_symbols_fd(btptrs, nptrs, fd);
 }
 #else
-#include <sys/debug.h>
-
 void
 libspl_backtrace(int fd __maybe_unused)
 {

lib/libspl/include/sys/uio.h

@@ -92,20 +92,20 @@ zfs_dio_page_aligned(void *buf)
 static inline boolean_t
 zfs_dio_offset_aligned(uint64_t offset, uint64_t blksz)
 {
-	return (IS_P2ALIGNED(offset, blksz));
+	return ((IS_P2ALIGNED(offset, blksz)) ? B_TRUE : B_FALSE);
 }
 
 static inline boolean_t
 zfs_dio_size_aligned(uint64_t size, uint64_t blksz)
 {
-	return ((size % blksz) == 0);
+	return (((size % blksz) == 0) ? B_TRUE : B_FALSE);
 }
 
 static inline boolean_t
 zfs_dio_aligned(uint64_t offset, uint64_t size, uint64_t blksz)
 {
-	return (zfs_dio_offset_aligned(offset, blksz) &&
-	    zfs_dio_size_aligned(size, blksz));
+	return ((zfs_dio_offset_aligned(offset, blksz) &&
+	    zfs_dio_size_aligned(size, blksz)) ? B_TRUE : B_FALSE);
 }
 
 static inline void

lib/libzfs/Makefile.am

@@ -70,7 +70,7 @@ if BUILD_FREEBSD
 libzfs_la_LIBADD += -lutil -lgeom
 endif
 
-libzfs_la_LDFLAGS += -version-info 5:0:1
+libzfs_la_LDFLAGS += -version-info 6:0:0
 
 pkgconfig_DATA += %D%/libzfs.pc
 

lib/libzfs/libzfs.abi

@@ -1,4 +1,4 @@
-<abi-corpus version='2.0' architecture='elf-amd-x86_64' soname='libzfs.so.4'>
+<abi-corpus version='2.0' architecture='elf-amd-x86_64' soname='libzfs.so.6'>
   <elf-needed>
     <dependency name='libzfs_core.so.3'/>
     <dependency name='libnvpair.so.3'/>
@@ -3132,7 +3132,8 @@
       <enumerator name='ZPOOL_PROP_DEDUP_TABLE_SIZE' value='36'/>
       <enumerator name='ZPOOL_PROP_DEDUP_TABLE_QUOTA' value='37'/>
       <enumerator name='ZPOOL_PROP_DEDUPCACHED' value='38'/>
-      <enumerator name='ZPOOL_NUM_PROPS' value='39'/>
+      <enumerator name='ZPOOL_PROP_LAST_SCRUBBED_TXG' value='39'/>
+      <enumerator name='ZPOOL_NUM_PROPS' value='40'/>
     </enum-decl>
     <typedef-decl name='zpool_prop_t' type-id='af1ba157' id='5d0c23fb'/>
     <typedef-decl name='regoff_t' type-id='95e97e5e' id='54a2a2a8'/>
@@ -5984,7 +5985,8 @@
       <underlying-type type-id='9cac1fee'/>
       <enumerator name='POOL_SCRUB_NORMAL' value='0'/>
       <enumerator name='POOL_SCRUB_PAUSE' value='1'/>
-      <enumerator name='POOL_SCRUB_FLAGS_END' value='2'/>
+      <enumerator name='POOL_SCRUB_FROM_LAST_TXG' value='2'/>
+      <enumerator name='POOL_SCRUB_FLAGS_END' value='3'/>
     </enum-decl>
     <typedef-decl name='pool_scrub_cmd_t' type-id='a1474cbd' id='b51cf3c2'/>
     <enum-decl name='zpool_errata' id='d9abbf54'>

lib/libzfs/libzfs_pool.c

@@ -471,13 +471,15 @@ int
 zpool_get_userprop(zpool_handle_t *zhp, const char *propname, char *buf,
     size_t len, zprop_source_t *srctype)
 {
-	nvlist_t *nv, *nvl;
+	nvlist_t *nv;
 	uint64_t ival;
 	const char *value;
 	zprop_source_t source = ZPROP_SRC_LOCAL;
 
-	nvl = zhp->zpool_props;
-	if (nvlist_lookup_nvlist(nvl, propname, &nv) == 0) {
+	if (zhp->zpool_props == NULL)
+		zpool_get_all_props(zhp);
+
+	if (nvlist_lookup_nvlist(zhp->zpool_props, propname, &nv) == 0) {
 		if (nvlist_lookup_uint64(nv, ZPROP_SOURCE, &ival) == 0)
 			source = ival;
 		verify(nvlist_lookup_string(nv, ZPROP_VALUE, &value) == 0);
@@ -2796,7 +2798,7 @@ zpool_scan(zpool_handle_t *zhp, pool_scan_func_t func, pool_scrub_cmd_t cmd)
 	}
 
 	/*
-	 * With EBUSY, five cases are possible:
+	 * With EBUSY, six cases are possible:
 	 *
 	 * Current state		Requested
 	 * 1. Normal Scrub Running	Normal Scrub or Error Scrub

lib/libzfs/libzfs_util.c

@@ -932,6 +932,7 @@ libzfs_run_process_impl(const char *path, char *argv[], char *env[], int flags,
 	pid = fork();
 	if (pid == 0) {
 		/* Child process */
+		setpgid(0, 0);
 		devnull_fd = open("/dev/null", O_WRONLY | O_CLOEXEC);
 
 		if (devnull_fd < 0)

lib/libzpool/Makefile.am

@@ -212,7 +212,7 @@ if BUILD_FREEBSD
 libzpool_la_LIBADD += -lgeom
 endif
 
-libzpool_la_LDFLAGS += -version-info 5:0:0
+libzpool_la_LDFLAGS += -version-info 6:0:0
 
 if TARGET_CPU_POWERPC
 module/zfs/libzpool_la-vdev_raidz_math_powerpc_altivec.$(OBJEXT) : CFLAGS += -maltivec

lib/libzpool/zfs_debug.c

@@ -35,9 +35,25 @@ typedef struct zfs_dbgmsg {
 
 static list_t zfs_dbgmsgs;
 static kmutex_t zfs_dbgmsgs_lock;
+static uint_t zfs_dbgmsg_size = 0;
+static uint_t zfs_dbgmsg_maxsize = 4<<20; /* 4MB */
 
 int zfs_dbgmsg_enable = B_TRUE;
 
+static void
+zfs_dbgmsg_purge(uint_t max_size)
+{
+	while (zfs_dbgmsg_size > max_size) {
+		zfs_dbgmsg_t *zdm = list_remove_head(&zfs_dbgmsgs);
+		if (zdm == NULL)
+			return;
+
+		uint_t size = zdm->zdm_size;
+		kmem_free(zdm, size);
+		zfs_dbgmsg_size -= size;
+	}
+}
+
 void
 zfs_dbgmsg_init(void)
 {
@@ -74,6 +90,8 @@ __zfs_dbgmsg(char *buf)
 
 	mutex_enter(&zfs_dbgmsgs_lock);
 	list_insert_tail(&zfs_dbgmsgs, zdm);
+	zfs_dbgmsg_size += size;
+	zfs_dbgmsg_purge(zfs_dbgmsg_maxsize);
 	mutex_exit(&zfs_dbgmsgs_lock);
 }
 

man/man4/zfs.4

@@ -18,7 +18,7 @@
 .\"
 .\" Copyright (c) 2024, Klara, Inc.
 .\"
-.Dd October 2, 2024
+.Dd November 1, 2024
 .Dt ZFS 4
 .Os
 .
@@ -436,7 +436,7 @@ write.
 It can also help to identify if reported checksum errors are tied to Direct I/O
 writes.
 Each verify error causes a
-.Sy dio_verify
+.Sy dio_verify_wr
 zevent.
 Direct Write I/O checkum verify errors can be seen with
 .Nm zpool Cm status Fl d .
@@ -1333,9 +1333,10 @@ results in vector instructions
 from the respective CPU instruction set being used.
 .
 .It Sy zfs_bclone_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
-Enable the experimental block cloning feature.
+Enables access to the block cloning feature.
 If this setting is 0, then even if feature@block_cloning is enabled,
-attempts to clone blocks will act as though the feature is disabled.
+using functions and system calls that attempt to clone blocks will act as
+though the feature is disabled.
 .
 .It Sy zfs_bclone_wait_dirty Ns = Ns Sy 0 Ns | Ns 1 Pq int
 When set to 1 the FICLONE and FICLONERANGE ioctls wait for dirty data to be

man/man5/vdev_id.conf.5

@@ -92,6 +92,11 @@ before a generic mapping for the same slot.
 In this way a custom mapping may be applied to a particular channel
 and a default mapping applied to the others.
 .
+.It Sy zpad_slot Ar digits
+Pad slot numbers with zeros to make them
+.Ar digits
+long, which can help to make disk names a consistent length and easier to sort.
+.
 .It Sy multipath Sy yes Ns | Ns Sy no
 Specifies whether
 .Xr vdev_id 8
@@ -122,7 +127,7 @@ device is connected to.
 The default is
 .Sy 4 .
 .
-.It Sy slot Sy bay Ns | Ns Sy phy Ns | Ns Sy port Ns | Ns Sy id Ns | Ns Sy lun Ns | Ns Sy ses
+.It Sy slot Sy bay Ns | Ns Sy phy Ns | Ns Sy port Ns | Ns Sy id Ns | Ns Sy lun Ns | Ns Sy bay_lun Ns | Ns Sy ses
 Specifies from which element of a SAS identifier the slot number is
 taken.
 The default is
@@ -138,6 +143,9 @@ use the SAS port as the slot number.
 use the scsi id as the slot number.
 .It Sy lun
 use the scsi lun as the slot number.
+.It Sy bay_lun
+read the slot number from the bay identifier and append the lun number.
+Useful for multi-lun multi-actuator hard drives.
 .It Sy ses
 use the SCSI Enclosure Services (SES) enclosure device slot number,
 as reported by

man/man7/zpoolprops.7

@@ -28,7 +28,7 @@
 .\" Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
 .\" Copyright (c) 2023, Klara Inc.
 .\"
-.Dd July 29, 2024
+.Dd November 18, 2024
 .Dt ZPOOLPROPS 7
 .Os
 .
@@ -135,6 +135,19 @@ A unique identifier for the pool.
 The current health of the pool.
 Health can be one of
 .Sy ONLINE , DEGRADED , FAULTED , OFFLINE, REMOVED , UNAVAIL .
+.It Sy last_scrubbed_txg
+Indicates the transaction group (TXG) up to which the most recent scrub
+operation has checked and repaired the dataset.
+This provides insight into the data integrity status of their pool at
+a specific point in time.
+.Xr zpool-scrub 8
+can utilize this property to scan only data that has changed since the last
+scrub completed, when given the
+.Fl C
+flag.
+This property is not updated when performing an error scrub with the
+.Fl e
+flag.
 .It Sy leaked
 Space not released while
 .Sy freeing

man/man8/zdb.8

@@ -14,7 +14,7 @@
 .\" Copyright (c) 2017 Lawrence Livermore National Security, LLC.
 .\" Copyright (c) 2017 Intel Corporation.
 .\"
-.Dd November 18, 2023
+.Dd October 27, 2024
 .Dt ZDB 8
 .Os
 .
@@ -408,6 +408,8 @@ blocks cloned, the space saving as a result of cloning, and the saving ratio.
 .It Fl TT
 Display the per-vdev BRT statistics, including total references.
 .It Fl TTT
+Display histograms of per-vdev BRT refcounts.
+.It Fl TTTT
 Dump the contents of the block reference tables.
 .It Fl u , -uberblock
 Display the current uberblock.

man/man8/zfs-list.8

@@ -71,7 +71,7 @@ The following fields are displayed:
 Used for scripting mode.
 Do not print headers and separate fields by a single tab instead of arbitrary
 white space.
-.It Fl j Op Ar --json-int
+.It Fl j , -json Op Ar --json-int
 Print the output in JSON format.
 Specify
 .Sy --json-int

man/man8/zfs-mount.8

@@ -59,7 +59,7 @@
 .Xc
 Displays all ZFS file systems currently mounted.
 .Bl -tag -width "-j"
-.It Fl j
+.It Fl j , -json
 Displays all mounted file systems in JSON format.
 .El
 .It Xo

man/man8/zfs-program.8

@@ -50,7 +50,7 @@ and any attempts to access or modify other pools will cause an error.
 .
 .Sh OPTIONS
 .Bl -tag -width "-t"
-.It Fl j
+.It Fl j , -json
 Display channel program output in JSON format.
 When this flag is specified and standard output is empty -
 channel program encountered an error.

man/man8/zfs-set.8

@@ -130,7 +130,7 @@ The value
 can be used to display all properties that apply to the given dataset's type
 .Pq Sy filesystem , volume , snapshot , No or Sy bookmark .
 .Bl -tag -width "-s source"
-.It Fl j Op Ar --json-int
+.It Fl j , -json Op Ar --json-int
 Display the output in JSON format.
 Specify
 .Sy --json-int

man/man8/zinject.8

@@ -23,7 +23,7 @@
 .\"
 .\" lint-ok: WARNING: sections out of conventional order: Sh SYNOPSIS
 .\"
-.Dd April 4, 2024
+.Dd December 2, 2024
 .Dt ZINJECT 8
 .Os
 .
@@ -268,7 +268,7 @@ Run for this many seconds before reporting failure.
 .It Fl T Ar failure
 Set the failure type to one of
 .Sy all ,
-.Sy ioctl ,
+.Sy flush ,
 .Sy claim ,
 .Sy free ,
 .Sy read ,

man/man8/zpool-events.8

@@ -98,7 +98,10 @@ This can be an indicator of problems with the underlying storage device.
 The number of delay events is ratelimited by the
 .Sy zfs_slow_io_events_per_second
 module parameter.
-.It Sy dio_verify
+.It Sy dio_verify_rd
+Issued when there was a checksum verify error after a Direct I/O read has been
+issued.
+.It Sy dio_verify_wr
 Issued when there was a checksum verify error after a Direct I/O write has been
 issued.
 This event can only take place if the module parameter

man/man8/zpool-get.8

@@ -98,7 +98,7 @@ See the
 .Xr zpoolprops 7
 manual page for more information on the available pool properties.
 .Bl -tag -compact -offset Ds -width "-o field"
-.It Fl j Op Ar --json-int, --json-pool-key-guid
+.It Fl j , -json Op Ar --json-int, --json-pool-key-guid
 Display the list of properties in JSON format.
 Specify
 .Sy --json-int
@@ -157,7 +157,7 @@ See the
 .Xr vdevprops 7
 manual page for more information on the available pool properties.
 .Bl -tag -compact -offset Ds -width "-o field"
-.It Fl j Op Ar --json-int
+.It Fl j , -json Op Ar --json-int
 Display the list of properties in JSON format.
 Specify
 .Sy --json-int

man/man8/zpool-list.8

@@ -59,7 +59,7 @@ is specified, the command exits after
 .Ar count
 reports are printed.
 .Bl -tag -width Ds
-.It Fl j Op Ar --json-int, --json-pool-key-guid
+.It Fl j , -json Op Ar --json-int, --json-pool-key-guid
 Display the list of pools in JSON format.
 Specify
 .Sy --json-int

man/man8/zpool-remove.8

@@ -109,7 +109,7 @@ Stops and cancels an in-progress removal of a top-level vdev.
 .El
 .
 .Sh EXAMPLES
-.\" These are, respectively, examples 14 from zpool.8
+.\" These are, respectively, examples 15 from zpool.8
 .\" Make sure to update them bidirectionally
 .Ss Example 1 : No Removing a Mirrored top-level (Log or Data) Device
 The following commands remove the mirrored log device
@@ -142,9 +142,43 @@ The command to remove the mirrored log
 .Ar mirror-2 No is :
 .Dl # Nm zpool Cm remove Ar tank mirror-2
 .Pp
+At this point, the log device no longer exists
+(both sides of the mirror have been removed):
+.Bd -literal -compact -offset Ds
+  pool: tank
+ state: ONLINE
+  scan: none requested
+config:
+
+        NAME        STATE     READ WRITE CKSUM
+        tank        ONLINE       0     0     0
+          mirror-0  ONLINE       0     0     0
+            sda     ONLINE       0     0     0
+            sdb     ONLINE       0     0     0
+          mirror-1  ONLINE       0     0     0
+            sdc     ONLINE       0     0     0
+            sdd     ONLINE       0     0     0
+.Ed
+.Pp
 The command to remove the mirrored data
 .Ar mirror-1 No is :
 .Dl # Nm zpool Cm remove Ar tank mirror-1
+.Pp
+After
+.Ar mirror-1 No has been evacuated, the pool remains redundant, but
+the total amount of space is reduced:
+.Bd -literal -compact -offset Ds
+  pool: tank
+ state: ONLINE
+  scan: none requested
+config:
+
+        NAME        STATE     READ WRITE CKSUM
+        tank        ONLINE       0     0     0
+          mirror-0  ONLINE       0     0     0
+            sda     ONLINE       0     0     0
+            sdb     ONLINE       0     0     0
+.Ed
 .
 .Sh SEE ALSO
 .Xr zpool-add 8 ,

man/man8/zpool-scrub.8

@@ -26,7 +26,7 @@
 .\" Copyright 2017 Nexenta Systems, Inc.
 .\" Copyright (c) 2017 Open-E, Inc. All Rights Reserved.
 .\"
-.Dd June 22, 2023
+.Dd November 18, 2024
 .Dt ZPOOL-SCRUB 8
 .Os
 .
@@ -36,9 +36,8 @@
 .Sh SYNOPSIS
 .Nm zpool
 .Cm scrub
-.Op Fl s Ns | Ns Fl p
+.Op Ns Fl e | Ns Fl p | Fl s Ns | Fl C Ns
 .Op Fl w
-.Op Fl e
 .Ar pool Ns …
 .
 .Sh DESCRIPTION
@@ -114,6 +113,10 @@ The pool must have been scrubbed at least once with the
 feature enabled to use this option.
 Error scrubbing cannot be run simultaneously with regular scrubbing or
 resilvering, nor can it be run when a regular scrub is paused.
+.It Fl C
+Continue scrub from last saved txg (see zpool
+.Sy last_scrubbed_txg
+property).
 .El
 .Sh EXAMPLES
 .Ss Example 1

man/man8/zpool-status.8

@@ -70,7 +70,7 @@ See the
 option of
 .Nm zpool Cm iostat
 for complete details.
-.It Fl j Op Ar --json-int, --json-flat-vdevs, --json-pool-key-guid
+.It Fl j , -json Op Ar --json-int, --json-flat-vdevs, --json-pool-key-guid
 Display the status for ZFS pools in JSON format.
 Specify
 .Sy --json-int
@@ -82,14 +82,18 @@ Specify
 .Sy --json-pool-key-guid
 to set pool GUID as key for pool objects instead of pool names.
 .It Fl d
-Display the number of Direct I/O write checksum verify errors that have occured
-on a top-level VDEV.
+Display the number of Direct I/O read/write checksum verify errors that have
+occured on a top-level VDEV.
 See
 .Sx zfs_vdev_direct_write_verify
 in
 .Xr zfs 4
 for details about the conditions that can cause Direct I/O write checksum
 verify failures to occur.
+Direct I/O reads checksum verify errors can also occur if the contents of the
+buffer are being manipulated after the I/O has been issued and is in flight.
+In the case of Direct I/O read checksum verify errors, the I/O will be reissued
+through the ARC.
 .It Fl D
 Display a histogram of deduplication statistics, showing the allocated
 .Pq physically present on disk

man/man8/zpool.8

@@ -405,9 +405,43 @@ The command to remove the mirrored log
 .Ar mirror-2 No is :
 .Dl # Nm zpool Cm remove Ar tank mirror-2
 .Pp
+At this point, the log device no longer exists
+(both sides of the mirror have been removed):
+.Bd -literal -compact -offset Ds
+  pool: tank
+ state: ONLINE
+  scan: none requested
+config:
+
+        NAME        STATE     READ WRITE CKSUM
+        tank        ONLINE       0     0     0
+          mirror-0  ONLINE       0     0     0
+            sda     ONLINE       0     0     0
+            sdb     ONLINE       0     0     0
+          mirror-1  ONLINE       0     0     0
+            sdc     ONLINE       0     0     0
+            sdd     ONLINE       0     0     0
+.Ed
+.Pp
 The command to remove the mirrored data
 .Ar mirror-1 No is :
 .Dl # Nm zpool Cm remove Ar tank mirror-1
+.Pp
+After
+.Ar mirror-1 No has been evacuated, the pool remains redundant, but
+the total amount of space is reduced:
+.Bd -literal -compact -offset Ds
+  pool: tank
+ state: ONLINE
+  scan: none requested
+config:
+
+        NAME        STATE     READ WRITE CKSUM
+        tank        ONLINE       0     0     0
+          mirror-0  ONLINE       0     0     0
+            sda     ONLINE       0     0     0
+            sdb     ONLINE       0     0     0
+.Ed
 .
 .Ss Example 16 : No Displaying expanded space on a device
 The following command displays the detailed information for the pool

module/nvpair/nvpair.c

@@ -3281,7 +3281,6 @@ nvs_xdr_nvp_##type(XDR *xdrs, void *ptr, ...)	\
 
 #endif
 
-/* BEGIN CSTYLED */
 NVS_BUILD_XDRPROC_T(char);
 NVS_BUILD_XDRPROC_T(short);
 NVS_BUILD_XDRPROC_T(u_short);
@@ -3289,7 +3288,6 @@ NVS_BUILD_XDRPROC_T(int);
 NVS_BUILD_XDRPROC_T(u_int);
 NVS_BUILD_XDRPROC_T(longlong_t);
 NVS_BUILD_XDRPROC_T(u_longlong_t);
-/* END CSTYLED */
 
 /*
  * The format of xdr encoded nvpair is:

module/os/freebsd/spl/spl_dtrace.c

@@ -31,5 +31,4 @@
 #include <sys/queue.h>
 #include <sys/sdt.h>
 
-/* CSTYLED */
 SDT_PROBE_DEFINE1(sdt, , , set__error, "int");

module/os/freebsd/zfs/abd_os.c

@@ -620,9 +620,16 @@ abd_borrow_buf_copy(abd_t *abd, size_t n)
 
 /*
  * Return a borrowed raw buffer to an ABD. If the ABD is scattered, this will
- * no change the contents of the ABD and will ASSERT that you didn't modify
- * the buffer since it was borrowed. If you want any changes you made to buf to
- * be copied back to abd, use abd_return_buf_copy() instead.
+ * not change the contents of the ABD. If you want any changes you made to
+ * buf to be copied back to abd, use abd_return_buf_copy() instead. If the
+ * ABD is not constructed from user pages from Direct I/O then an ASSERT
+ * checks to make sure the contents of the buffer have not changed since it was
+ * borrowed. We can not ASSERT the contents of the buffer have not changed if
+ * it is composed of user pages. While Direct I/O write pages are placed under
+ * write protection and can not be changed, this is not the case for Direct I/O
+ * reads. The pages of a Direct I/O read could be manipulated at any time.
+ * Checksum verifications in the ZIO pipeline check for this issue and handle
+ * it by returning an error on checksum verification failure.
  */
 void
 abd_return_buf(abd_t *abd, void *buf, size_t n)
@@ -632,8 +639,34 @@ abd_return_buf(abd_t *abd, void *buf, size_t n)
 #ifdef ZFS_DEBUG
 	(void) zfs_refcount_remove_many(&abd->abd_children, n, buf);
 #endif
-	if (abd_is_linear(abd)) {
+	if (abd_is_from_pages(abd)) {
+		if (!abd_is_linear_page(abd))
+			zio_buf_free(buf, n);
+	} else if (abd_is_linear(abd)) {
 		ASSERT3P(buf, ==, abd_to_buf(abd));
+	} else if (abd_is_gang(abd)) {
+#ifdef ZFS_DEBUG
+		/*
+		 * We have to be careful with gang ABD's that we do not ASSERT
+		 * for any ABD's that contain user pages from Direct I/O. See
+		 * the comment above about Direct I/O read buffers possibly
+		 * being manipulated. In order to handle this, we jsut iterate
+		 * through the gang ABD and only verify ABD's that are not from
+		 * user pages.
+		 */
+		void *cmp_buf = buf;
+
+		for (abd_t *cabd = list_head(&ABD_GANG(abd).abd_gang_chain);
+		    cabd != NULL;
+		    cabd = list_next(&ABD_GANG(abd).abd_gang_chain, cabd)) {
+			if (!abd_is_from_pages(cabd)) {
+				ASSERT0(abd_cmp_buf(cabd, cmp_buf,
+				    cabd->abd_size));
+			}
+			cmp_buf = (char *)cmp_buf + cabd->abd_size;
+		}
+#endif
+		zio_buf_free(buf, n);
 	} else {
 		ASSERT0(abd_cmp_buf(abd, buf, n));
 		zio_buf_free(buf, n);

module/os/freebsd/zfs/dmu_os.c

@@ -103,6 +103,7 @@ dmu_write_pages(objset_t *os, uint64_t object, uint64_t offset, uint64_t size,
 			    db->db_offset + bufoff);
 			thiscpy = MIN(PAGESIZE, tocpy - copied);
 			va = zfs_map_page(*ma, &sf);
+			ASSERT(db->db_data != NULL);
 			memcpy((char *)db->db_data + bufoff, va, thiscpy);
 			zfs_unmap_page(sf);
 			ma += 1;
@@ -172,6 +173,7 @@ dmu_read_pages(objset_t *os, uint64_t object, vm_page_t *ma, int count,
 		ASSERT3U(db->db_size, >, PAGE_SIZE);
 		bufoff = IDX_TO_OFF(m->pindex) % db->db_size;
 		va = zfs_map_page(m, &sf);
+		ASSERT(db->db_data != NULL);
 		memcpy(va, (char *)db->db_data + bufoff, PAGESIZE);
 		zfs_unmap_page(sf);
 		vm_page_valid(m);
@@ -211,8 +213,10 @@ dmu_read_pages(objset_t *os, uint64_t object, vm_page_t *ma, int count,
 		 */
 		tocpy = MIN(db->db_size - bufoff, PAGESIZE - pgoff);
 		ASSERT3S(tocpy, >=, 0);
-		if (m != bogus_page)
+		if (m != bogus_page) {
+			ASSERT(db->db_data != NULL);
 			memcpy(va + pgoff, (char *)db->db_data + bufoff, tocpy);
+		}
 
 		pgoff += tocpy;
 		ASSERT3S(pgoff, >=, 0);
@@ -290,6 +294,7 @@ dmu_read_pages(objset_t *os, uint64_t object, vm_page_t *ma, int count,
 		bufoff = IDX_TO_OFF(m->pindex) % db->db_size;
 		tocpy = MIN(db->db_size - bufoff, PAGESIZE);
 		va = zfs_map_page(m, &sf);
+		ASSERT(db->db_data != NULL);
 		memcpy(va, (char *)db->db_data + bufoff, tocpy);
 		if (tocpy < PAGESIZE) {
 			ASSERT3S(i, ==, *rahead - 1);

module/os/freebsd/zfs/sysctl_os.c

@@ -187,12 +187,10 @@ param_set_arc_max(SYSCTL_HANDLER_ARGS)
 	return (0);
 }
 
-/* BEGIN CSTYLED */
 SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_max,
 	CTLTYPE_ULONG | CTLFLAG_RWTUN | CTLFLAG_MPSAFE,
 	NULL, 0, param_set_arc_max, "LU",
 	"Maximum ARC size in bytes (LEGACY)");
-/* END CSTYLED */
 
 int
 param_set_arc_min(SYSCTL_HANDLER_ARGS)
@@ -218,12 +216,10 @@ param_set_arc_min(SYSCTL_HANDLER_ARGS)
 	return (0);
 }
 
-/* BEGIN CSTYLED */
 SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_min,
 	CTLTYPE_ULONG | CTLFLAG_RWTUN | CTLFLAG_MPSAFE,
 	NULL, 0, param_set_arc_min, "LU",
 	"Minimum ARC size in bytes (LEGACY)");
-/* END CSTYLED */
 
 extern uint_t zfs_arc_free_target;
 
@@ -252,13 +248,11 @@ param_set_arc_free_target(SYSCTL_HANDLER_ARGS)
  * NOTE: This sysctl is CTLFLAG_RW not CTLFLAG_RWTUN due to its dependency on
  * pagedaemon initialization.
  */
-/* BEGIN CSTYLED */
 SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_free_target,
 	CTLTYPE_UINT | CTLFLAG_RW | CTLFLAG_MPSAFE,
 	NULL, 0, param_set_arc_free_target, "IU",
 	"Desired number of free pages below which ARC triggers reclaim"
 	" (LEGACY)");
-/* END CSTYLED */
 
 int
 param_set_arc_no_grow_shift(SYSCTL_HANDLER_ARGS)
@@ -278,84 +272,64 @@ param_set_arc_no_grow_shift(SYSCTL_HANDLER_ARGS)
 	return (0);
 }
 
-/* BEGIN CSTYLED */
 SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_no_grow_shift,
 	CTLTYPE_INT | CTLFLAG_RWTUN | CTLFLAG_MPSAFE,
 	NULL, 0, param_set_arc_no_grow_shift, "I",
 	"log2(fraction of ARC which must be free to allow growing) (LEGACY)");
-/* END CSTYLED */
 
 extern uint64_t l2arc_write_max;
 
-/* BEGIN CSTYLED */
 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_write_max,
 	CTLFLAG_RWTUN, &l2arc_write_max, 0,
 	"Max write bytes per interval (LEGACY)");
-/* END CSTYLED */
 
 extern uint64_t l2arc_write_boost;
 
-/* BEGIN CSTYLED */
 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_write_boost,
 	CTLFLAG_RWTUN, &l2arc_write_boost, 0,
 	"Extra write bytes during device warmup (LEGACY)");
-/* END CSTYLED */
 
 extern uint64_t l2arc_headroom;
 
-/* BEGIN CSTYLED */
 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_headroom,
 	CTLFLAG_RWTUN, &l2arc_headroom, 0,
 	"Number of max device writes to precache (LEGACY)");
-/* END CSTYLED */
 
 extern uint64_t l2arc_headroom_boost;
 
-/* BEGIN CSTYLED */
 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_headroom_boost,
 	CTLFLAG_RWTUN, &l2arc_headroom_boost, 0,
 	"Compressed l2arc_headroom multiplier (LEGACY)");
-/* END CSTYLED */
 
 extern uint64_t l2arc_feed_secs;
 
-/* BEGIN CSTYLED */
 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_feed_secs,
 	CTLFLAG_RWTUN, &l2arc_feed_secs, 0,
 	"Seconds between L2ARC writing (LEGACY)");
-/* END CSTYLED */
 
 extern uint64_t l2arc_feed_min_ms;
 
-/* BEGIN CSTYLED */
 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_feed_min_ms,
 	CTLFLAG_RWTUN, &l2arc_feed_min_ms, 0,
 	"Min feed interval in milliseconds (LEGACY)");
-/* END CSTYLED */
 
 extern int l2arc_noprefetch;
 
-/* BEGIN CSTYLED */
 SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_noprefetch,
 	CTLFLAG_RWTUN, &l2arc_noprefetch, 0,
 	"Skip caching prefetched buffers (LEGACY)");
-/* END CSTYLED */
 
 extern int l2arc_feed_again;
 
-/* BEGIN CSTYLED */
 SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_feed_again,
 	CTLFLAG_RWTUN, &l2arc_feed_again, 0,
 	"Turbo L2ARC warmup (LEGACY)");
-/* END CSTYLED */
 
 extern int l2arc_norw;
 
-/* BEGIN CSTYLED */
 SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_norw,
 	CTLFLAG_RWTUN, &l2arc_norw, 0,
 	"No reads during writes (LEGACY)");
-/* END CSTYLED */
 
 static int
 param_get_arc_state_size(SYSCTL_HANDLER_ARGS)
@@ -370,7 +344,6 @@ param_get_arc_state_size(SYSCTL_HANDLER_ARGS)
 
 extern arc_state_t ARC_anon;
 
-/* BEGIN CSTYLED */
 SYSCTL_PROC(_vfs_zfs, OID_AUTO, anon_size,
 	CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
 	&ARC_anon, 0, param_get_arc_state_size, "Q",
@@ -381,11 +354,9 @@ SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, anon_metadata_esize, CTLFLAG_RD,
 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, anon_data_esize, CTLFLAG_RD,
 	&ARC_anon.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
 	"size of evictable data in anonymous state");
-/* END CSTYLED */
 
 extern arc_state_t ARC_mru;
 
-/* BEGIN CSTYLED */
 SYSCTL_PROC(_vfs_zfs, OID_AUTO, mru_size,
 	CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
 	&ARC_mru, 0, param_get_arc_state_size, "Q",
@@ -396,11 +367,9 @@ SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_metadata_esize, CTLFLAG_RD,
 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_data_esize, CTLFLAG_RD,
 	&ARC_mru.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
 	"size of evictable data in mru state");
-/* END CSTYLED */
 
 extern arc_state_t ARC_mru_ghost;
 
-/* BEGIN CSTYLED */
 SYSCTL_PROC(_vfs_zfs, OID_AUTO, mru_ghost_size,
 	CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
 	&ARC_mru_ghost, 0, param_get_arc_state_size, "Q",
@@ -411,11 +380,9 @@ SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_ghost_metadata_esize, CTLFLAG_RD,
 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_ghost_data_esize, CTLFLAG_RD,
 	&ARC_mru_ghost.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
 	"size of evictable data in mru ghost state");
-/* END CSTYLED */
 
 extern arc_state_t ARC_mfu;
 
-/* BEGIN CSTYLED */
 SYSCTL_PROC(_vfs_zfs, OID_AUTO, mfu_size,
 	CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
 	&ARC_mfu, 0, param_get_arc_state_size, "Q",
@@ -426,11 +393,9 @@ SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_metadata_esize, CTLFLAG_RD,
 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_data_esize, CTLFLAG_RD,
 	&ARC_mfu.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
 	"size of evictable data in mfu state");
-/* END CSTYLED */
 
 extern arc_state_t ARC_mfu_ghost;
 
-/* BEGIN CSTYLED */
 SYSCTL_PROC(_vfs_zfs, OID_AUTO, mfu_ghost_size,
 	CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
 	&ARC_mfu_ghost, 0, param_get_arc_state_size, "Q",
@@ -441,11 +406,9 @@ SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_ghost_metadata_esize, CTLFLAG_RD,
 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_ghost_data_esize, CTLFLAG_RD,
 	&ARC_mfu_ghost.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
 	"size of evictable data in mfu ghost state");
-/* END CSTYLED */
 
 extern arc_state_t ARC_uncached;
 
-/* BEGIN CSTYLED */
 SYSCTL_PROC(_vfs_zfs, OID_AUTO, uncached_size,
 	CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
 	&ARC_uncached, 0, param_get_arc_state_size, "Q",
@@ -456,16 +419,13 @@ SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, uncached_metadata_esize, CTLFLAG_RD,
 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, uncached_data_esize, CTLFLAG_RD,
 	&ARC_uncached.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
 	"size of evictable data in uncached state");
-/* END CSTYLED */
 
 extern arc_state_t ARC_l2c_only;
 
-/* BEGIN CSTYLED */
 SYSCTL_PROC(_vfs_zfs, OID_AUTO, l2c_only_size,
 	CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
 	&ARC_l2c_only, 0, param_get_arc_state_size, "Q",
 	"size of l2c_only state");
-/* END CSTYLED */
 
 /* dbuf.c */
 
@@ -477,19 +437,15 @@ SYSCTL_NODE(_vfs_zfs, OID_AUTO, zfetch, CTLFLAG_RW, 0, "ZFS ZFETCH (LEGACY)");
 
 extern uint32_t	zfetch_max_distance;
 
-/* BEGIN CSTYLED */
 SYSCTL_UINT(_vfs_zfs_zfetch, OID_AUTO, max_distance,
 	CTLFLAG_RWTUN, &zfetch_max_distance, 0,
 	"Max bytes to prefetch per stream (LEGACY)");
-/* END CSTYLED */
 
 extern uint32_t	zfetch_max_idistance;
 
-/* BEGIN CSTYLED */
 SYSCTL_UINT(_vfs_zfs_zfetch, OID_AUTO, max_idistance,
 	CTLFLAG_RWTUN, &zfetch_max_idistance, 0,
 	"Max bytes to prefetch indirects for per stream (LEGACY)");
-/* END CSTYLED */
 
 /* dsl_pool.c */
 
@@ -527,12 +483,10 @@ param_set_active_allocator(SYSCTL_HANDLER_ARGS)
  */
 extern int zfs_metaslab_sm_blksz_no_log;
 
-/* BEGIN CSTYLED */
 SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, sm_blksz_no_log,
 	CTLFLAG_RDTUN, &zfs_metaslab_sm_blksz_no_log, 0,
 	"Block size for space map in pools with log space map disabled.  "
 	"Power of 2 greater than 4096.");
-/* END CSTYLED */
 
 /*
  * When the log space map feature is enabled, we accumulate a lot of
@@ -541,12 +495,10 @@ SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, sm_blksz_no_log,
  */
 extern int zfs_metaslab_sm_blksz_with_log;
 
-/* BEGIN CSTYLED */
 SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, sm_blksz_with_log,
 	CTLFLAG_RDTUN, &zfs_metaslab_sm_blksz_with_log, 0,
 	"Block size for space map in pools with log space map enabled.  "
 	"Power of 2 greater than 4096.");
-/* END CSTYLED */
 
 /*
  * The in-core space map representation is more compact than its on-disk form.
@@ -556,29 +508,23 @@ SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, sm_blksz_with_log,
  */
 extern uint_t zfs_condense_pct;
 
-/* BEGIN CSTYLED */
 SYSCTL_UINT(_vfs_zfs, OID_AUTO, condense_pct,
 	CTLFLAG_RWTUN, &zfs_condense_pct, 0,
 	"Condense on-disk spacemap when it is more than this many percents"
 	" of in-memory counterpart");
-/* END CSTYLED */
 
 extern uint_t zfs_remove_max_segment;
 
-/* BEGIN CSTYLED */
 SYSCTL_UINT(_vfs_zfs, OID_AUTO, remove_max_segment,
 	CTLFLAG_RWTUN, &zfs_remove_max_segment, 0,
 	"Largest contiguous segment ZFS will attempt to allocate when removing"
 	" a device");
-/* END CSTYLED */
 
 extern int zfs_removal_suspend_progress;
 
-/* BEGIN CSTYLED */
 SYSCTL_INT(_vfs_zfs, OID_AUTO, removal_suspend_progress,
 	CTLFLAG_RWTUN, &zfs_removal_suspend_progress, 0,
 	"Ensures certain actions can happen while in the middle of a removal");
-/* END CSTYLED */
 
 /*
  * Minimum size which forces the dynamic allocator to change
@@ -588,12 +534,10 @@ SYSCTL_INT(_vfs_zfs, OID_AUTO, removal_suspend_progress,
  */
 extern uint64_t metaslab_df_alloc_threshold;
 
-/* BEGIN CSTYLED */
 SYSCTL_QUAD(_vfs_zfs_metaslab, OID_AUTO, df_alloc_threshold,
 	CTLFLAG_RWTUN, &metaslab_df_alloc_threshold, 0,
 	"Minimum size which forces the dynamic allocator to change its"
 	" allocation strategy");
-/* END CSTYLED */
 
 /*
  * The minimum free space, in percent, which must be available
@@ -603,12 +547,10 @@ SYSCTL_QUAD(_vfs_zfs_metaslab, OID_AUTO, df_alloc_threshold,
  */
 extern uint_t metaslab_df_free_pct;
 
-/* BEGIN CSTYLED */
 SYSCTL_UINT(_vfs_zfs_metaslab, OID_AUTO, df_free_pct,
 	CTLFLAG_RWTUN, &metaslab_df_free_pct, 0,
 	"The minimum free space, in percent, which must be available in a"
 	" space map to continue allocations in a first-fit fashion");
-/* END CSTYLED */
 
 /* mmp.c */
 
@@ -631,28 +573,22 @@ param_set_multihost_interval(SYSCTL_HANDLER_ARGS)
 
 extern int zfs_ccw_retry_interval;
 
-/* BEGIN CSTYLED */
 SYSCTL_INT(_vfs_zfs, OID_AUTO, ccw_retry_interval,
 	CTLFLAG_RWTUN, &zfs_ccw_retry_interval, 0,
 	"Configuration cache file write, retry after failure, interval"
 	" (seconds)");
-/* END CSTYLED */
 
 extern uint64_t zfs_max_missing_tvds_cachefile;
 
-/* BEGIN CSTYLED */
 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, max_missing_tvds_cachefile,
 	CTLFLAG_RWTUN, &zfs_max_missing_tvds_cachefile, 0,
 	"Allow importing pools with missing top-level vdevs in cache file");
-/* END CSTYLED */
 
 extern uint64_t zfs_max_missing_tvds_scan;
 
-/* BEGIN CSTYLED */
 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, max_missing_tvds_scan,
 	CTLFLAG_RWTUN, &zfs_max_missing_tvds_scan, 0,
 	"Allow importing pools with missing top-level vdevs during scan");
-/* END CSTYLED */
 
 /* spa_misc.c */
 
@@ -681,11 +617,9 @@ sysctl_vfs_zfs_debug_flags(SYSCTL_HANDLER_ARGS)
 	return (0);
 }
 
-/* BEGIN CSTYLED */
 SYSCTL_PROC(_vfs_zfs, OID_AUTO, debugflags,
 	CTLTYPE_UINT | CTLFLAG_MPSAFE | CTLFLAG_RWTUN, NULL, 0,
 	sysctl_vfs_zfs_debug_flags, "IU", "Debug flags for ZFS testing.");
-/* END CSTYLED */
 
 int
 param_set_deadman_synctime(SYSCTL_HANDLER_ARGS)
@@ -768,10 +702,8 @@ param_set_slop_shift(SYSCTL_HANDLER_ARGS)
 
 extern int space_map_ibs;
 
-/* BEGIN CSTYLED */
 SYSCTL_INT(_vfs_zfs, OID_AUTO, space_map_ibs, CTLFLAG_RWTUN,
 	&space_map_ibs, 0, "Space map indirect block shift");
-/* END CSTYLED */
 
 
 /* vdev.c */
@@ -795,13 +727,11 @@ param_set_min_auto_ashift(SYSCTL_HANDLER_ARGS)
 	return (0);
 }
 
-/* BEGIN CSTYLED */
 SYSCTL_PROC(_vfs_zfs, OID_AUTO, min_auto_ashift,
 	CTLTYPE_UINT | CTLFLAG_RWTUN | CTLFLAG_MPSAFE,
 	&zfs_vdev_min_auto_ashift, sizeof (zfs_vdev_min_auto_ashift),
 	param_set_min_auto_ashift, "IU",
 	"Min ashift used when creating new top-level vdev. (LEGACY)");
-/* END CSTYLED */
 
 int
 param_set_max_auto_ashift(SYSCTL_HANDLER_ARGS)
@@ -822,14 +752,12 @@ param_set_max_auto_ashift(SYSCTL_HANDLER_ARGS)
 	return (0);
 }
 
-/* BEGIN CSTYLED */
 SYSCTL_PROC(_vfs_zfs, OID_AUTO, max_auto_ashift,
 	CTLTYPE_UINT | CTLFLAG_RWTUN | CTLFLAG_MPSAFE,
 	&zfs_vdev_max_auto_ashift, sizeof (zfs_vdev_max_auto_ashift),
 	param_set_max_auto_ashift, "IU",
 	"Max ashift used when optimizing for logical -> physical sector size on"
 	" new top-level vdevs. (LEGACY)");
-/* END CSTYLED */
 
 /*
  * Since the DTL space map of a vdev is not expected to have a lot of
@@ -837,11 +765,9 @@ SYSCTL_PROC(_vfs_zfs, OID_AUTO, max_auto_ashift,
  */
 extern int zfs_vdev_dtl_sm_blksz;
 
-/* BEGIN CSTYLED */
 SYSCTL_INT(_vfs_zfs, OID_AUTO, dtl_sm_blksz,
 	CTLFLAG_RDTUN, &zfs_vdev_dtl_sm_blksz, 0,
 	"Block size for DTL space map.  Power of 2 greater than 4096.");
-/* END CSTYLED */
 
 /*
  * vdev-wide space maps that have lots of entries written to them at
@@ -850,19 +776,15 @@ SYSCTL_INT(_vfs_zfs, OID_AUTO, dtl_sm_blksz,
  */
 extern int zfs_vdev_standard_sm_blksz;
 
-/* BEGIN CSTYLED */
 SYSCTL_INT(_vfs_zfs, OID_AUTO, standard_sm_blksz,
 	CTLFLAG_RDTUN, &zfs_vdev_standard_sm_blksz, 0,
 	"Block size for standard space map.  Power of 2 greater than 4096.");
-/* END CSTYLED */
 
 extern int vdev_validate_skip;
 
-/* BEGIN CSTYLED */
 SYSCTL_INT(_vfs_zfs, OID_AUTO, validate_skip,
 	CTLFLAG_RDTUN, &vdev_validate_skip, 0,
 	"Enable to bypass vdev_validate().");
-/* END CSTYLED */
 
 /* vdev_mirror.c */
 
@@ -870,17 +792,13 @@ SYSCTL_INT(_vfs_zfs, OID_AUTO, validate_skip,
 
 extern uint_t zfs_vdev_max_active;
 
-/* BEGIN CSTYLED */
 SYSCTL_UINT(_vfs_zfs, OID_AUTO, top_maxinflight,
 	CTLFLAG_RWTUN, &zfs_vdev_max_active, 0,
 	"The maximum number of I/Os of all types active for each device."
 	" (LEGACY)");
-/* END CSTYLED */
 
 /* zio.c */
 
-/* BEGIN CSTYLED */
 SYSCTL_INT(_vfs_zfs_zio, OID_AUTO, exclude_metadata,
 	CTLFLAG_RDTUN, &zio_exclude_metadata, 0,
 	"Exclude metadata buffers from dumps as well");
-/* END CSTYLED */

module/os/freebsd/zfs/zfs_vnops_os.c

@@ -291,8 +291,12 @@ zfs_ioctl(vnode_t *vp, ulong_t com, intptr_t data, int flag, cred_t *cred,
 	case F_SEEK_HOLE:
 	{
 		off = *(offset_t *)data;
+		error = vn_lock(vp, LK_SHARED);
+		if (error)
+			return (error);
 		/* offset parameter is in/out */
 		error = zfs_holey(VTOZ(vp), com, &off);
+		VOP_UNLOCK(vp);
 		if (error)
 			return (error);
 		*(offset_t *)data = off;
@@ -452,8 +456,10 @@ mappedread_sf(znode_t *zp, int nbytes, zfs_uio_t *uio)
 				if (!vm_page_wired(pp) && pp->valid == 0 &&
 				    vm_page_busy_tryupgrade(pp))
 					vm_page_free(pp);
-				else
+				else {
+					vm_page_deactivate_noreuse(pp);
 					vm_page_sunbusy(pp);
+				}
 				zfs_vmobject_wunlock(obj);
 			}
 		} else {
@@ -3928,6 +3934,7 @@ zfs_getpages(struct vnode *vp, vm_page_t *ma, int count, int *rbehind,
 	if (zfs_enter_verify_zp(zfsvfs, zp, FTAG) != 0)
 		return (zfs_vm_pagerret_error);
 
+	object = ma[0]->object;
 	start = IDX_TO_OFF(ma[0]->pindex);
 	end = IDX_TO_OFF(ma[count - 1]->pindex + 1);
 
@@ -3936,32 +3943,44 @@ zfs_getpages(struct vnode *vp, vm_page_t *ma, int count, int *rbehind,
 	 * Note that we need to handle the case of the block size growing.
 	 */
 	for (;;) {
+		uint64_t len;
+
 		blksz = zp->z_blksz;
+		len = roundup(end, blksz) - rounddown(start, blksz);
+
 		lr = zfs_rangelock_tryenter(&zp->z_rangelock,
-		    rounddown(start, blksz),
-		    roundup(end, blksz) - rounddown(start, blksz), RL_READER);
+		    rounddown(start, blksz), len, RL_READER);
 		if (lr == NULL) {
-			if (rahead != NULL) {
-				*rahead = 0;
-				rahead = NULL;
-			}
-			if (rbehind != NULL) {
-				*rbehind = 0;
-				rbehind = NULL;
-			}
-			break;
+			/*
+			 * Avoid a deadlock with update_pages().  We need to
+			 * hold the range lock when copying from the DMU, so
+			 * give up the busy lock to allow update_pages() to
+			 * proceed.  We might need to allocate new pages, which
+			 * isn't quite right since this allocation isn't subject
+			 * to the page fault handler's OOM logic, but this is
+			 * the best we can do for now.
+			 */
+			for (int i = 0; i < count; i++)
+				vm_page_xunbusy(ma[i]);
+
+			lr = zfs_rangelock_enter(&zp->z_rangelock,
+			    rounddown(start, blksz), len, RL_READER);
+
+			zfs_vmobject_wlock(object);
+			(void) vm_page_grab_pages(object, OFF_TO_IDX(start),
+			    VM_ALLOC_NORMAL | VM_ALLOC_WAITOK | VM_ALLOC_ZERO,
+			    ma, count);
+			zfs_vmobject_wunlock(object);
 		}
 		if (blksz == zp->z_blksz)
 			break;
 		zfs_rangelock_exit(lr);
 	}
 
-	object = ma[0]->object;
 	zfs_vmobject_wlock(object);
 	obj_size = object->un_pager.vnp.vnp_size;
 	zfs_vmobject_wunlock(object);
 	if (IDX_TO_OFF(ma[count - 1]->pindex) >= obj_size) {
-		if (lr != NULL)
 		zfs_rangelock_exit(lr);
 		zfs_exit(zfsvfs, FTAG);
 		return (zfs_vm_pagerret_bad);
@@ -3987,10 +4006,32 @@ zfs_getpages(struct vnode *vp, vm_page_t *ma, int count, int *rbehind,
 	 * ZFS will panic if we request DMU to read beyond the end of the last
 	 * allocated block.
 	 */
-	error = dmu_read_pages(zfsvfs->z_os, zp->z_id, ma, count, &pgsin_b,
-	    &pgsin_a, MIN(end, obj_size) - (end - PAGE_SIZE));
+	for (int i = 0; i < count; i++) {
+		int dummypgsin, count1, j, last_size;
+
+		if (vm_page_any_valid(ma[i])) {
+			ASSERT(vm_page_all_valid(ma[i]));
+			continue;
+		}
+		for (j = i + 1; j < count; j++) {
+			if (vm_page_any_valid(ma[j])) {
+				ASSERT(vm_page_all_valid(ma[j]));
+				break;
+			}
+		}
+		count1 = j - i;
+		dummypgsin = 0;
+		last_size = j == count ?
+		    MIN(end, obj_size) - (end - PAGE_SIZE) : PAGE_SIZE;
+		error = dmu_read_pages(zfsvfs->z_os, zp->z_id, &ma[i], count1,
+		    i == 0 ? &pgsin_b : &dummypgsin,
+		    j == count ? &pgsin_a : &dummypgsin,
+		    last_size);
+		if (error != 0)
+			break;
+		i += count1 - 1;
+	}
 
-	if (lr != NULL)
 	zfs_rangelock_exit(lr);
 	ZFS_ACCESSTIME_STAMP(zfsvfs, zp);
 
@@ -6159,7 +6200,7 @@ zfs_freebsd_copy_file_range(struct vop_copy_file_range_args *ap)
 	} else {
 #if (__FreeBSD_version >= 1302506 && __FreeBSD_version < 1400000) || \
 	__FreeBSD_version >= 1400086
-		vn_lock_pair(invp, false, LK_EXCLUSIVE, outvp, false,
+		vn_lock_pair(invp, false, LK_SHARED, outvp, false,
 		    LK_EXCLUSIVE);
 #else
 		vn_lock_pair(invp, false, outvp, false);

module/os/freebsd/zfs/zfs_znode_os.c

@@ -370,8 +370,6 @@ zfs_znode_sa_init(zfsvfs_t *zfsvfs, znode_t *zp,
 	 */
 	if (zp->z_id == zfsvfs->z_root && zfsvfs->z_parent == zfsvfs)
 		ZTOV(zp)->v_flag |= VROOT;
-
-	vn_exists(ZTOV(zp));
 }
 
 void

module/os/freebsd/zfs/zio_crypt.c

@@ -1686,11 +1686,10 @@ zio_do_crypt_data(boolean_t encrypt, zio_crypt_key_t *key,
 	freebsd_crypt_session_t *tmpl = NULL;
 	uint8_t *authbuf = NULL;
 
-
+	memset(&puio_s, 0, sizeof (puio_s));
+	memset(&cuio_s, 0, sizeof (cuio_s));
 	zfs_uio_init(&puio, &puio_s);
 	zfs_uio_init(&cuio, &cuio_s);
-	memset(GET_UIO_STRUCT(&puio), 0, sizeof (struct uio));
-	memset(GET_UIO_STRUCT(&cuio), 0, sizeof (struct uio));
 
 #ifdef FCRYPTO_DEBUG
 	printf("%s(%s, %p, %p, %d, %p, %p, %u, %s, %p, %p, %p)\n",
@@ -1824,7 +1823,6 @@ error:
 }
 
 #if defined(_KERNEL) && defined(HAVE_SPL)
-/* CSTYLED */
 module_param(zfs_key_max_salt_uses, ulong, 0644);
 MODULE_PARM_DESC(zfs_key_max_salt_uses, "Max number of times a salt value "
 	"can be used for generating encryption keys before it is rotated");

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

@@ -33,7 +33,6 @@
  * But we would still default to the current default of not to do that.
  */
 static unsigned int spl_panic_halt;
-/* CSTYLED */
 module_param(spl_panic_halt, uint, 0644);
 MODULE_PARM_DESC(spl_panic_halt, "Cause kernel panic on assertion failures");
 

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

@@ -54,7 +54,6 @@
 unsigned long spl_hostid = 0;
 EXPORT_SYMBOL(spl_hostid);
 
-/* CSTYLED */
 module_param(spl_hostid, ulong, 0644);
 MODULE_PARM_DESC(spl_hostid, "The system hostid.");
 

module/os/linux/spl/spl-kmem-cache.c

@@ -48,7 +48,6 @@
 #define	smp_mb__after_atomic(x) smp_mb__after_clear_bit(x)
 #endif
 
-/* BEGIN CSTYLED */
 /*
  * Cache magazines are an optimization designed to minimize the cost of
  * allocating memory.  They do this by keeping a per-cpu cache of recently
@@ -97,7 +96,6 @@ static unsigned int spl_kmem_cache_kmem_threads = 4;
 module_param(spl_kmem_cache_kmem_threads, uint, 0444);
 MODULE_PARM_DESC(spl_kmem_cache_kmem_threads,
 	"Number of spl_kmem_cache threads");
-/* END CSTYLED */
 
 /*
  * Slab allocation interfaces