Tag zfs-2.3.1

META file and changelog updated.

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

.github/workflows/checkstyle.yaml

@@ -19,7 +19,7 @@ jobs:
       run: |
         # for x in lxd core20 snapd; do sudo snap remove $x; done
         sudo apt-get purge -y snapd google-chrome-stable firefox
-        ONLY_DEPS=1 .github/workflows/scripts/qemu-3-deps.sh ubuntu22
+        ONLY_DEPS=1 .github/workflows/scripts/qemu-3-deps-vm.sh ubuntu22
         sudo apt-get install -y cppcheck devscripts mandoc pax-utils shellcheck
         sudo python -m pipx install --quiet flake8
         # confirm that the tools are installed

.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/v2024-12-14"
 URLzs=""
 
 # Ubuntu mirrors
@@ -40,6 +40,12 @@ case "$OS" in
     # dns sometimes fails with that url  :/
     echo "89.187.191.12  geo.mirror.pkgbuild.com" | sudo tee /etc/hosts > /dev/null
     ;;
+  centos-stream10)
+    OSNAME="CentOS Stream 10"
+    # TODO: #16903 Overwrite OSv to stream9 for virt-install until it's added to osinfo
+    OSv="centos-stream9"
+    URL="https://cloud.centos.org/centos/10-stream/x86_64/images/CentOS-Stream-GenericCloud-10-latest.x86_64.qcow2"
+    ;;
   centos-stream9)
     OSNAME="CentOS Stream 9"
     URL="https://cloud.centos.org/centos/9-stream/x86_64/images/CentOS-Stream-GenericCloud-9-latest.x86_64.qcow2"
@@ -52,43 +58,56 @@ case "$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)
+  freebsd14-1r)
+    OSNAME="FreeBSD 14.1-RELEASE"
+    OSv="freebsd14.0"
+    URLzs="$FREEBSD/amd64-freebsd-14.1-RELEASE.qcow2.zst"
+    BASH="/usr/local/bin/bash"
+    ;;
+  freebsd14-2r)
+    OSNAME="FreeBSD 14.2-RELEASE"
+    OSv="freebsd14.0"
+    URLzs="$FREEBSD/amd64-freebsd-14.2-RELEASE.qcow2.zst"
+    BASH="/usr/local/bin/bash"
+    ;;
+  freebsd13-4s)
     OSNAME="FreeBSD 13.4-STABLE"
     OSv="freebsd13.0"
     URLzs="$FREEBSD/amd64-freebsd-13.4-STABLE.qcow2.zst"
     BASH="/usr/local/bin/bash"
     NIC="rtl8139"
     ;;
-  freebsd14r)
-    OSNAME="FreeBSD 14.1-RELEASE"
+  freebsd14-2s)
+    OSNAME="FreeBSD 14.2-STABLE"
     OSv="freebsd14.0"
-    URLzs="$FREEBSD/amd64-freebsd-14.1-RELEASE.qcow2.zst"
+    URLzs="$FREEBSD/amd64-freebsd-14.2-STABLE.qcow2.zst"
     BASH="/usr/local/bin/bash"
     ;;
-  freebsd14)
-    OSNAME="FreeBSD 14.1-STABLE"
-    OSv="freebsd14.0"
-    URLzs="$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"
@@ -201,6 +220,15 @@ sudo virt-install \
   --disk $DISK,bus=virtio,cache=none,format=$FORMAT,driver.discard=unmap \
   --import --noautoconsole >/dev/null
 
+# Give the VMs hostnames so we don't have to refer to them with
+# hardcoded IP addresses.
+#
+# vm0:          Initial VM we install dependencies and build ZFS on.
+# vm1..2        Testing VMs
+for i in {0..9} ; do
+  echo "192.168.122.1$i vm$i" | sudo tee -a /etc/hosts
+done
+
 # in case the directory isn't there already
 mkdir -p $HOME/.ssh
 

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

@@ -0,0 +1,232 @@
+#!/usr/bin/env bash
+
+######################################################################
+# 3) install dependencies for compiling and loading
+#
+# $1: OS name (like 'fedora41')
+######################################################################
+
+set -eu
+
+function archlinux() {
+  echo "##[group]Running pacman -Syu"
+  sudo btrfs filesystem resize max /
+  sudo pacman -Syu --noconfirm
+  echo "##[endgroup]"
+
+  echo "##[group]Install Development Tools"
+  sudo pacman -Sy --noconfirm base-devel bc cpio cryptsetup dhclient dkms \
+    fakeroot fio gdb inetutils jq less linux linux-headers lsscsi nfs-utils \
+    parted pax perf python-packaging python-setuptools qemu-guest-agent ksh \
+    samba sysstat rng-tools rsync wget xxhash
+  echo "##[endgroup]"
+}
+
+function debian() {
+  export DEBIAN_FRONTEND="noninteractive"
+
+  echo "##[group]Running apt-get update+upgrade"
+  sudo apt-get update -y
+  sudo apt-get upgrade -y
+  echo "##[endgroup]"
+
+  echo "##[group]Install Development Tools"
+  sudo apt-get install -y \
+    acl alien attr autoconf bc cpio cryptsetup curl dbench dh-python dkms \
+    fakeroot fio gdb gdebi git ksh lcov isc-dhcp-client jq libacl1-dev \
+    libaio-dev libattr1-dev libblkid-dev libcurl4-openssl-dev libdevmapper-dev \
+    libelf-dev libffi-dev libmount-dev libpam0g-dev libselinux-dev libssl-dev \
+    libtool libtool-bin libudev-dev libunwind-dev linux-headers-$(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 \
+    rsync samba sysstat uuid-dev watchdog wget xfslibs-dev  xxhash zlib1g-dev
+  echo "##[endgroup]"
+}
+
+function freebsd() {
+  export ASSUME_ALWAYS_YES="YES"
+
+  echo "##[group]Install Development Tools"
+  sudo pkg install -y autoconf automake autotools base64 checkbashisms fio \
+    gdb gettext gettext-runtime git gmake gsed jq ksh93 lcov libtool lscpu \
+    pkgconf python python3 pamtester pamtester qemu-guest-agent rsync xxhash
+  sudo pkg install -xy \
+    '^samba4[[:digit:]]+$' \
+    '^py3[[:digit:]]+-cffi$' \
+    '^py3[[:digit:]]+-sysctl$' \
+    '^py3[[:digit:]]+-packaging$'
+  echo "##[endgroup]"
+}
+
+# common packages for: almalinux, centos, redhat
+function rhel() {
+  echo "##[group]Running dnf update"
+  echo "max_parallel_downloads=10" | sudo -E tee -a /etc/dnf/dnf.conf
+  sudo dnf clean all
+  sudo dnf update -y --setopt=fastestmirror=1 --refresh
+  echo "##[endgroup]"
+
+  echo "##[group]Install 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 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]"
+}
+
+function tumbleweed() {
+  echo "##[group]Running zypper is TODO!"
+  sleep 23456
+  echo "##[endgroup]"
+}
+
+# Install dependencies
+case "$1" in
+  almalinux8)
+    echo "##[group]Enable epel and powertools repositories"
+    sudo dnf config-manager -y --set-enabled powertools
+    sudo dnf install -y epel-release
+    echo "##[endgroup]"
+    rhel
+    echo "##[group]Install kernel-abi-whitelists"
+    sudo dnf install -y kernel-abi-whitelists
+    echo "##[endgroup]"
+    ;;
+  almalinux9|centos-stream9|centos-stream10)
+    echo "##[group]Enable epel and crb repositories"
+    sudo dnf config-manager -y --set-enabled crb
+    sudo dnf install -y epel-release
+    echo "##[endgroup]"
+    rhel
+    echo "##[group]Install kernel-abi-stablelists"
+    sudo dnf install -y kernel-abi-stablelists
+    echo "##[endgroup]"
+    ;;
+  archlinux)
+    archlinux
+    ;;
+  debian*)
+    echo 'debconf debconf/frontend select Noninteractive' | sudo debconf-set-selections
+    debian
+    echo "##[group]Install Debian specific"
+    sudo apt-get install -yq linux-perf dh-sequence-dkms
+    echo "##[endgroup]"
+    ;;
+  fedora*)
+    rhel
+    sudo dnf install -y libunwind-devel
+    ;;
+  freebsd*)
+    freebsd
+    ;;
+  tumbleweed)
+    tumbleweed
+    ;;
+  ubuntu*)
+    debian
+    echo "##[group]Install Ubuntu specific"
+    sudo apt-get install -yq linux-tools-common libtirpc-dev \
+      linux-modules-extra-$(uname -r)
+    if [ "$1" != "ubuntu20" ]; then
+      sudo apt-get install -yq dh-sequence-dkms
+    fi
+    echo "##[endgroup]"
+    echo "##[group]Delete Ubuntu OpenZFS modules"
+    for i in $(find /lib/modules -name zfs -type d); do sudo rm -rvf $i; done
+    echo "##[endgroup]"
+    ;;
+esac
+
+# This script is used for checkstyle + zloop deps also.
+# Install only the needed packages and exit - when used this way.
+test -z "${ONLY_DEPS:-}" || exit 0
+
+# Start services
+echo "##[group]Enable services"
+case "$1" in
+  freebsd*)
+    # add virtio things
+    echo 'virtio_load="YES"' | sudo -E tee -a /boot/loader.conf
+    for i in balloon blk console random scsi; do
+      echo "virtio_${i}_load=\"YES\"" | sudo -E tee -a /boot/loader.conf
+    done
+    echo "fdescfs /dev/fd fdescfs rw 0 0" | sudo -E tee -a /etc/fstab
+    sudo -E mount /dev/fd
+    sudo -E touch /etc/zfs/exports
+    sudo -E sysrc mountd_flags="/etc/zfs/exports"
+    echo '[global]' | sudo -E tee /usr/local/etc/smb4.conf >/dev/null
+    sudo -E service nfsd enable
+    sudo -E service qemu-guest-agent enable
+    sudo -E service samba_server enable
+    ;;
+  debian*|ubuntu*)
+    sudo -E systemctl enable nfs-kernel-server
+    sudo -E systemctl enable qemu-guest-agent
+    sudo -E systemctl enable smbd
+    ;;
+  *)
+    # All other linux distros
+    sudo -E systemctl enable nfs-server
+    sudo -E systemctl enable qemu-guest-agent
+    sudo -E systemctl enable smb
+    ;;
+esac
+echo "##[endgroup]"
+
+# Setup Kernel cmdline
+CMDLINE="console=tty0 console=ttyS0,115200n8"
+CMDLINE="$CMDLINE selinux=0"
+CMDLINE="$CMDLINE random.trust_cpu=on"
+CMDLINE="$CMDLINE no_timer_check"
+case "$1" in
+  almalinux*|centos*|fedora*)
+    GRUB_CFG="/boot/grub2/grub.cfg"
+    GRUB_MKCONFIG="grub2-mkconfig"
+    CMDLINE="$CMDLINE biosdevname=0 net.ifnames=0"
+    echo 'GRUB_SERIAL_COMMAND="serial --speed=115200"' \
+      | sudo tee -a /etc/default/grub >/dev/null
+    ;;
+  ubuntu24)
+    GRUB_CFG="/boot/grub/grub.cfg"
+    GRUB_MKCONFIG="grub-mkconfig"
+    echo 'GRUB_DISABLE_OS_PROBER="false"' \
+      | sudo tee -a /etc/default/grub >/dev/null
+    ;;
+  *)
+    GRUB_CFG="/boot/grub/grub.cfg"
+    GRUB_MKCONFIG="grub-mkconfig"
+    ;;
+esac
+
+case "$1" in
+  archlinux|freebsd*)
+    true
+    ;;
+  *)
+    echo "##[group]Edit kernel cmdline"
+    sudo sed -i -e '/^GRUB_CMDLINE_LINUX/d' /etc/default/grub || true
+    echo "GRUB_CMDLINE_LINUX=\"$CMDLINE\"" \
+      | sudo tee -a /etc/default/grub >/dev/null
+    sudo $GRUB_MKCONFIG -o $GRUB_CFG
+    echo "##[endgroup]"
+    ;;
+esac
+
+# reset cloud-init configuration and poweroff
+sudo cloud-init clean --logs
+sleep 2 && sudo poweroff &
+exit 0

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

@@ -1,221 +1,15 @@
-#!/usr/bin/env bash
-
 ######################################################################
-# 3) install dependencies for compiling and loading
+# 3) Wait for VM to boot from previous step and launch dependencies
+#    script on it.
+#
+# $1: OS name (like 'fedora41')
 ######################################################################
 
-set -eu
-
-function archlinux() {
-  echo "##[group]Running pacman -Syu"
-  sudo btrfs filesystem resize max /
-  sudo pacman -Syu --noconfirm
-  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
-  echo "##[endgroup]"
-}
-
-function debian() {
-  export DEBIAN_FRONTEND="noninteractive"
-
-  echo "##[group]Running apt-get update+upgrade"
-  sudo apt-get update -y
-  sudo apt-get upgrade -y
-  echo "##[endgroup]"
-
-  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-$(uname -r) \
-    lsscsi nfs-kernel-server pamtester parted python3 python3-all-dev \
-    python3-cffi python3-dev python3-distlib python3-packaging \
-    python3-setuptools python3-sphinx qemu-guest-agent rng-tools rpm2cpio \
-    rsync samba sysstat uuid-dev watchdog wget xfslibs-dev  xxhash zlib1g-dev
-  echo "##[endgroup]"
-}
-
-function freebsd() {
-  export ASSUME_ALWAYS_YES="YES"
-
-  echo "##[group]Install Development Tools"
-  sudo pkg install -y autoconf automake autotools base64 checkbashisms fio \
-    gdb gettext gettext-runtime git gmake gsed jq ksh93 lcov libtool lscpu \
-    pkgconf python python3 pamtester pamtester qemu-guest-agent rsync xxhash
-  sudo pkg install -xy \
-    '^samba4[[:digit:]]+$' \
-    '^py3[[:digit:]]+-cffi$' \
-    '^py3[[:digit:]]+-sysctl$' \
-    '^py3[[:digit:]]+-packaging$'
-  echo "##[endgroup]"
-}
-
-# common packages for: almalinux, centos, redhat
-function rhel() {
-  echo "##[group]Running dnf update"
-  echo "max_parallel_downloads=10" | sudo -E tee -a /etc/dnf/dnf.conf
-  sudo dnf clean all
-  sudo dnf update -y --setopt=fastestmirror=1 --refresh
-  echo "##[endgroup]"
-
-  echo "##[group]Install Development Tools"
-  sudo dnf group install -y "Development Tools"
-  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
-  echo "##[endgroup]"
-}
-
-function tumbleweed() {
-  echo "##[group]Running zypper is TODO!"
-  sleep 23456
-  echo "##[endgroup]"
-}
-
-# Install dependencies
-case "$1" in
-  almalinux8)
-    echo "##[group]Enable epel and powertools repositories"
-    sudo dnf config-manager -y --set-enabled powertools
-    sudo dnf install -y epel-release
-    echo "##[endgroup]"
-    rhel
-    echo "##[group]Install kernel-abi-whitelists"
-    sudo dnf install -y kernel-abi-whitelists
-    echo "##[endgroup]"
-    ;;
-  almalinux9|centos-stream9)
-    echo "##[group]Enable epel and crb repositories"
-    sudo dnf config-manager -y --set-enabled crb
-    sudo dnf install -y epel-release
-    echo "##[endgroup]"
-    rhel
-    echo "##[group]Install kernel-abi-stablelists"
-    sudo dnf install -y kernel-abi-stablelists
-    echo "##[endgroup]"
-    ;;
-  archlinux)
-    archlinux
-    ;;
-  debian*)
-    debian
-    echo "##[group]Install Debian specific"
-    sudo apt-get install -yq linux-perf dh-sequence-dkms
-    echo "##[endgroup]"
-    ;;
-  fedora*)
-    rhel
-    ;;
-  freebsd*)
-    freebsd
-    ;;
-  tumbleweed)
-    tumbleweed
-    ;;
-  ubuntu*)
-    debian
-    echo "##[group]Install Ubuntu specific"
-    sudo apt-get install -yq linux-tools-common libtirpc-dev \
-      linux-modules-extra-$(uname -r)
-    if [ "$1" != "ubuntu20" ]; then
-      sudo apt-get install -yq dh-sequence-dkms
-    fi
-    echo "##[endgroup]"
-    echo "##[group]Delete Ubuntu OpenZFS modules"
-    for i in $(find /lib/modules -name zfs -type d); do sudo rm -rvf $i; done
-    echo "##[endgroup]"
-    ;;
-esac
-
-# This script is used for checkstyle + zloop deps also.
-# Install only the needed packages and exit - when used this way.
-test -z "${ONLY_DEPS:-}" || exit 0
-
-# Start services
-echo "##[group]Enable services"
-case "$1" in
-  freebsd*)
-    # add virtio things
-    echo 'virtio_load="YES"' | sudo -E tee -a /boot/loader.conf
-    for i in balloon blk console random scsi; do
-      echo "virtio_${i}_load=\"YES\"" | sudo -E tee -a /boot/loader.conf
-    done
-    echo "fdescfs /dev/fd fdescfs rw 0 0" | sudo -E tee -a /etc/fstab
-    sudo -E mount /dev/fd
-    sudo -E touch /etc/zfs/exports
-    sudo -E sysrc mountd_flags="/etc/zfs/exports"
-    echo '[global]' | sudo -E tee /usr/local/etc/smb4.conf >/dev/null
-    sudo -E service nfsd enable
-    sudo -E service qemu-guest-agent enable
-    sudo -E service samba_server enable
-    ;;
-  debian*|ubuntu*)
-    sudo -E systemctl enable nfs-kernel-server
-    sudo -E systemctl enable qemu-guest-agent
-    sudo -E systemctl enable smbd
-    ;;
-  *)
-    # All other linux distros
-    sudo -E systemctl enable nfs-server
-    sudo -E systemctl enable qemu-guest-agent
-    sudo -E systemctl enable smb
-    ;;
-esac
-echo "##[endgroup]"
-
-# Setup Kernel cmdline
-CMDLINE="console=tty0 console=ttyS0,115200n8"
-CMDLINE="$CMDLINE selinux=0"
-CMDLINE="$CMDLINE random.trust_cpu=on"
-CMDLINE="$CMDLINE no_timer_check"
-case "$1" in
-  almalinux*|centos*|fedora*)
-    GRUB_CFG="/boot/grub2/grub.cfg"
-    GRUB_MKCONFIG="grub2-mkconfig"
-    CMDLINE="$CMDLINE biosdevname=0 net.ifnames=0"
-    echo 'GRUB_SERIAL_COMMAND="serial --speed=115200"' \
-      | sudo tee -a /etc/default/grub >/dev/null
-    ;;
-  ubuntu24)
-    GRUB_CFG="/boot/grub/grub.cfg"
-    GRUB_MKCONFIG="grub-mkconfig"
-    echo 'GRUB_DISABLE_OS_PROBER="false"' \
-      | sudo tee -a /etc/default/grub >/dev/null
-    ;;
-  *)
-    GRUB_CFG="/boot/grub/grub.cfg"
-    GRUB_MKCONFIG="grub-mkconfig"
-    ;;
-esac
-
-case "$1" in
-  archlinux|freebsd*)
-    true
-    ;;
-  *)
-    echo "##[group]Edit kernel cmdline"
-    sudo sed -i -e '/^GRUB_CMDLINE_LINUX/d' /etc/default/grub || true
-    echo "GRUB_CMDLINE_LINUX=\"$CMDLINE\"" \
-      | sudo tee -a /etc/default/grub >/dev/null
-    sudo $GRUB_MKCONFIG -o $GRUB_CFG
-    echo "##[endgroup]"
-    ;;
-esac
-
-# reset cloud-init configuration and poweroff
-sudo cloud-init clean --logs
-sleep 2 && sudo poweroff &
-exit 0
+.github/workflows/scripts/qemu-wait-for-vm.sh vm0
+scp .github/workflows/scripts/qemu-3-deps-vm.sh zfs@vm0:qemu-3-deps-vm.sh
+PID=`pidof /usr/bin/qemu-system-x86_64`
+ssh zfs@vm0 '$HOME/qemu-3-deps-vm.sh' $1
+# wait for poweroff to succeed
+tail --pid=$PID -f /dev/null
+sleep 5 # avoid this: "error: Domain is already active"
+rm -f $HOME/.ssh/known_hosts

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

@@ -0,0 +1,370 @@
+#!/usr/bin/env bash
+
+######################################################################
+# 4) configure and build openzfs modules.  This is run on the VMs.
+#
+# Usage:
+#
+#       qemu-4-build-vm.sh OS [--enable-debug][--dkms][--poweroff]
+#           [--release][--repo][--tarball]
+#
+# OS:           OS name like 'fedora41'
+# --enable-debug:  Build RPMs with '--enable-debug' (for testing)
+# --dkms:       Build DKMS RPMs as well
+# --poweroff:   Power-off the VM after building
+# --release     Build zfs-release*.rpm as well
+# --repo        After building everything, copy RPMs into /tmp/repo
+#               in the ZFS RPM repository file structure.  Also
+#               copy tarballs if they were built.
+# --tarball:    Also build a tarball of ZFS source
+######################################################################
+
+ENABLE_DEBUG=""
+DKMS=""
+POWEROFF=""
+RELEASE=""
+REPO=""
+TARBALL=""
+while [[ $# -gt 0 ]]; do
+  case $1 in
+    --enable-debug)
+      ENABLE_DEBUG=1
+      shift
+      ;;
+    --dkms)
+      DKMS=1
+      shift
+      ;;
+    --poweroff)
+      POWEROFF=1
+      shift
+      ;;
+    --release)
+      RELEASE=1
+      shift
+      ;;
+    --repo)
+      REPO=1
+      shift
+      ;;
+    --tarball)
+      TARBALL=1
+      shift
+      ;;
+    *)
+      OS=$1
+      shift
+      ;;
+  esac
+done
+
+set -eu
+
+function run() {
+  LOG="/var/tmp/build-stderr.txt"
+  echo "****************************************************"
+  echo "$(date) ($*)"
+  echo "****************************************************"
+  ($@ || echo $? > /tmp/rv) 3>&1 1>&2 2>&3 | stdbuf -eL -oL tee -a $LOG
+  if [ -f /tmp/rv ]; then
+    RV=$(cat /tmp/rv)
+    echo "****************************************************"
+    echo "exit with value=$RV ($*)"
+    echo "****************************************************"
+    echo 1 > /var/tmp/build-exitcode.txt
+    exit $RV
+  fi
+}
+
+# Look at the RPMs in the current directory and copy/move them to
+# /tmp/repo, using the directory structure we use for the ZFS RPM repos.
+#
+# For example:
+# /tmp/repo/epel-testing/9.5
+# /tmp/repo/epel-testing/9.5/SRPMS
+# /tmp/repo/epel-testing/9.5/SRPMS/zfs-2.3.99-1.el9.src.rpm
+# /tmp/repo/epel-testing/9.5/SRPMS/zfs-kmod-2.3.99-1.el9.src.rpm
+# /tmp/repo/epel-testing/9.5/kmod
+# /tmp/repo/epel-testing/9.5/kmod/x86_64
+# /tmp/repo/epel-testing/9.5/kmod/x86_64/debug
+# /tmp/repo/epel-testing/9.5/kmod/x86_64/debug/kmod-zfs-debuginfo-2.3.99-1.el9.x86_64.rpm
+# /tmp/repo/epel-testing/9.5/kmod/x86_64/debug/libnvpair3-debuginfo-2.3.99-1.el9.x86_64.rpm
+# /tmp/repo/epel-testing/9.5/kmod/x86_64/debug/libuutil3-debuginfo-2.3.99-1.el9.x86_64.rpm
+# ...
+function copy_rpms_to_repo {
+  # Pick a RPM to query. It doesn't matter which one - we just want to extract
+  # the 'Build Host' value from it.
+  rpm=$(ls zfs-*.rpm | head -n 1)
+
+  # Get zfs version '2.2.99'
+  zfs_ver=$(rpm -qpi $rpm | awk '/Version/{print $3}')
+
+  # Get "2.1" or "2.2"
+  zfs_major=$(echo $zfs_ver | grep -Eo [0-9]+\.[0-9]+)
+
+  # Get 'almalinux9.5' or 'fedora41' type string
+  build_host=$(rpm -qpi $rpm | awk '/Build Host/{print $4}')
+
+  # Get '9.5' or '41' OS version
+  os_ver=$(echo $build_host | grep -Eo '[0-9\.]+$')
+
+  # Our ZFS version and OS name will determine which repo the RPMs
+  # will go in (regular or testing).  Fedora always gets the newest
+  # releases, and Alma gets the older releases.
+  case $build_host in
+  almalinux*)
+    case $zfs_major in
+    2.2)
+      d="epel"
+      ;;
+    *)
+      d="epel-testing"
+      ;;
+    esac
+    ;;
+  fedora*)
+    d="fedora"
+    ;;
+  esac
+
+  prefix=/tmp/repo
+  dst="$prefix/$d/$os_ver"
+
+  # Special case: move zfs-release*.rpm out of the way first (if we built them).
+  # This will make filtering the other RPMs easier.
+  mkdir -p $dst
+  mv zfs-release*.rpm $dst || true
+
+  # Copy source RPMs
+  mkdir -p $dst/SRPMS
+  cp $(ls *.src.rpm) $dst/SRPMS/
+
+  if [[ "$build_host" =~ "almalinux" ]] ; then
+    # Copy kmods+userspace
+    mkdir -p $dst/kmod/x86_64/debug
+    cp $(ls *.rpm | grep -Ev 'src.rpm|dkms|debuginfo') $dst/kmod/x86_64
+    cp *debuginfo*.rpm $dst/kmod/x86_64/debug
+  fi
+
+  if [ -n "$DKMS" ] ; then
+    # Copy dkms+userspace
+    mkdir -p $dst/x86_64
+    cp $(ls *.rpm | grep -Ev 'src.rpm|kmod|debuginfo') $dst/x86_64
+  fi
+
+  # Copy debug
+  mkdir -p $dst/x86_64/debug
+  cp $(ls *debuginfo*.rpm | grep -v kmod) $dst/x86_64/debug
+}
+
+function freebsd() {
+  extra="${1:-}"
+
+  export MAKE="gmake"
+  echo "##[group]Autogen.sh"
+  run ./autogen.sh
+  echo "##[endgroup]"
+
+  echo "##[group]Configure"
+  run ./configure \
+    --prefix=/usr/local \
+    --with-libintl-prefix=/usr/local \
+    --enable-pyzfs \
+    --enable-debuginfo $extra
+  echo "##[endgroup]"
+
+  echo "##[group]Build"
+  run gmake -j$(sysctl -n hw.ncpu)
+  echo "##[endgroup]"
+
+  echo "##[group]Install"
+  run sudo gmake install
+  echo "##[endgroup]"
+}
+
+function linux() {
+  extra="${1:-}"
+
+  echo "##[group]Autogen.sh"
+  run ./autogen.sh
+  echo "##[endgroup]"
+
+  echo "##[group]Configure"
+  run ./configure \
+    --prefix=/usr \
+    --enable-pyzfs \
+    --enable-debuginfo $extra
+  echo "##[endgroup]"
+
+  echo "##[group]Build"
+  run make -j$(nproc)
+  echo "##[endgroup]"
+
+  echo "##[group]Install"
+  run sudo make install
+  echo "##[endgroup]"
+}
+
+function rpm_build_and_install() {
+  extra="${1:-}"
+
+  # Build RPMs with XZ compression by default (since gzip decompression is slow)
+  echo "%_binary_payload w7.xzdio" >> ~/.rpmmacros
+
+  echo "##[group]Autogen.sh"
+  run ./autogen.sh
+  echo "##[endgroup]"
+
+  echo "##[group]Configure"
+  run ./configure --enable-debuginfo $extra
+  echo "##[endgroup]"
+
+  echo "##[group]Build"
+  run make pkg-kmod pkg-utils
+  echo "##[endgroup]"
+
+  if [ -n "$DKMS" ] ; then
+    echo "##[group]DKMS"
+    make rpm-dkms
+    echo "##[endgroup]"
+  fi
+
+  if [ -n "$REPO" ] ; then
+    echo "Skipping install since we're only building RPMs and nothing else"
+  else
+    echo "##[group]Install"
+    run sudo dnf -y --nobest install $(ls *.rpm | grep -Ev 'dkms|src.rpm')
+    echo "##[endgroup]"
+  fi
+
+  # Optionally build the zfs-release.*.rpm
+  if [ -n "$RELEASE" ] ; then
+    echo "##[group]Release"
+    pushd ~
+    sudo dnf -y install rpm-build || true
+    # Check out a sparse copy of zfsonlinux.github.com.git so we don't get
+    # all the binaries.  We just need a few kilobytes of files to build RPMs.
+    git clone --depth 1 --no-checkout \
+      https://github.com/zfsonlinux/zfsonlinux.github.com.git
+
+    cd zfsonlinux.github.com
+    git sparse-checkout set zfs-release
+    git checkout
+    cd zfs-release
+
+    mkdir -p ~/rpmbuild/{BUILDROOT,SPECS,RPMS,SRPMS,SOURCES,BUILD}
+    cp RPM-GPG-KEY-openzfs* *.repo ~/rpmbuild/SOURCES
+    cp zfs-release.spec ~/rpmbuild/SPECS/
+    rpmbuild -ba ~/rpmbuild/SPECS/zfs-release.spec
+
+    # ZFS release RPMs are built.  Copy them to the ~/zfs directory just to
+    # keep all the RPMs in the same place.
+    cp ~/rpmbuild/RPMS/noarch/*.rpm .
+    cp ~/rpmbuild/SRPMS/*.rpm .
+    popd
+    rm -fr ~/rpmbuild
+    echo "##[endgroup]"
+  fi
+
+  if [ -n "$REPO" ] ; then
+    echo "##[group]Repo"
+    copy_rpms_to_repo
+    echo "##[endgroup]"
+  fi
+}
+
+function deb_build_and_install() {
+  extra="${1:-}"
+
+  echo "##[group]Autogen.sh"
+  run ./autogen.sh
+  echo "##[endgroup]"
+
+  echo "##[group]Configure"
+  run ./configure \
+    --prefix=/usr \
+    --enable-pyzfs \
+    --enable-debuginfo $extra
+  echo "##[endgroup]"
+
+  echo "##[group]Build"
+  run make native-deb-kmod native-deb-utils
+  echo "##[endgroup]"
+
+  echo "##[group]Install"
+  # Do kmod install.  Note that when you build the native debs, the
+  # packages themselves are placed in parent directory '../' rather than
+  # in the source directory like the rpms are.
+  run sudo apt-get -y install $(find ../ | grep -E '\.deb$' \
+    | grep -Ev 'dkms|dracut')
+  echo "##[endgroup]"
+}
+
+# Debug: show kernel cmdline
+if [ -f /proc/cmdline ] ; then
+  cat /proc/cmdline || true
+fi
+
+# Set our hostname to our OS name and version number.  Specifically, we set the
+# major and minor number so that when we query the Build Host field in the RPMs
+# we build, we can see what specific version of Fedora/Almalinux we were using
+# to build them.  This is helpful for matching up KMOD versions.
+#
+# Examples:
+#
+# rhel8.10
+# almalinux9.5
+# fedora40
+source /etc/os-release
+sudo hostname "$ID$VERSION_ID"
+
+# save some sysinfo
+uname -a > /var/tmp/uname.txt
+
+cd $HOME/zfs
+export PATH="$PATH:/sbin:/usr/sbin:/usr/local/sbin"
+
+extra=""
+if [ -n "$ENABLE_DEBUG" ] ; then
+  extra="--enable-debug"
+fi
+
+# build
+case "$OS" in
+  freebsd*)
+    freebsd "$extra"
+    ;;
+  alma*|centos*)
+    rpm_build_and_install "--with-spec=redhat $extra"
+    ;;
+  fedora*)
+    rpm_build_and_install "$extra"
+    ;;
+  debian*|ubuntu*)
+    deb_build_and_install "$extra"
+    ;;
+  *)
+    linux "$extra"
+    ;;
+esac
+
+# Optionally build tarballs.  The tarball's root directory name will be named
+# after the current tag, like 'zfs-2.3.0' or 'master'.
+if [ -n "$TARBALL" ] ; then
+  tag="$(git symbolic-ref -q --short HEAD || git describe --tags --exact-match)"
+  git archive --format=tar.gz -o $tag.tar.gz $tag
+  if [ -n "$REPO" ] ; then
+    mkdir -p /tmp/repo/releases
+    cp $tag.tar.gz /tmp/repo/releases
+  fi
+fi
+
+# building the zfs module was ok
+echo 0 > /var/tmp/build-exitcode.txt
+
+# reset cloud-init configuration and poweroff
+if [ -n "$POWEROFF" ] ; then
+        sudo cloud-init clean --logs
+        sync && sleep 2 && sudo poweroff &
+fi
+exit 0

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

@@ -3,151 +3,9 @@
 ######################################################################
 # 4) configure and build openzfs modules
 ######################################################################
+echo "Build modules in QEMU machine"
 
-set -eu
+# Bring our VM back up and copy over ZFS source
+.github/workflows/scripts/qemu-prepare-for-build.sh
 
-function run() {
-  LOG="/var/tmp/build-stderr.txt"
-  echo "****************************************************"
-  echo "$(date) ($*)"
-  echo "****************************************************"
-  ($@ || echo $? > /tmp/rv) 3>&1 1>&2 2>&3 | stdbuf -eL -oL tee -a $LOG
-  if [ -f /tmp/rv ]; then
-    RV=$(cat /tmp/rv)
-    echo "****************************************************"
-    echo "exit with value=$RV ($*)"
-    echo "****************************************************"
-    echo 1 > /var/tmp/build-exitcode.txt
-    exit $RV
-  fi
-}
-
-function freebsd() {
-  export MAKE="gmake"
-  echo "##[group]Autogen.sh"
-  run ./autogen.sh
-  echo "##[endgroup]"
-
-  echo "##[group]Configure"
-  run ./configure \
-    --prefix=/usr/local \
-    --with-libintl-prefix=/usr/local \
-    --enable-pyzfs \
-    --enable-debug \
-    --enable-debuginfo
-  echo "##[endgroup]"
-
-  echo "##[group]Build"
-  run gmake -j$(sysctl -n hw.ncpu)
-  echo "##[endgroup]"
-
-  echo "##[group]Install"
-  run sudo gmake install
-  echo "##[endgroup]"
-}
-
-function linux() {
-  echo "##[group]Autogen.sh"
-  run ./autogen.sh
-  echo "##[endgroup]"
-
-  echo "##[group]Configure"
-  run ./configure \
-    --prefix=/usr \
-    --enable-pyzfs \
-    --enable-debug \
-    --enable-debuginfo
-  echo "##[endgroup]"
-
-  echo "##[group]Build"
-  run make -j$(nproc)
-  echo "##[endgroup]"
-
-  echo "##[group]Install"
-  run sudo make install
-  echo "##[endgroup]"
-}
-
-function rpm_build_and_install() {
-  EXTRA_CONFIG="${1:-}"
-  echo "##[group]Autogen.sh"
-  run ./autogen.sh
-  echo "##[endgroup]"
-
-  echo "##[group]Configure"
-  run ./configure --enable-debug --enable-debuginfo $EXTRA_CONFIG
-  echo "##[endgroup]"
-
-  echo "##[group]Build"
-  run make pkg-kmod pkg-utils
-  echo "##[endgroup]"
-
-  echo "##[group]Install"
-  run sudo dnf -y --skip-broken localinstall $(ls *.rpm | grep -v src.rpm)
-  echo "##[endgroup]"
-
-}
-
-function deb_build_and_install() {
-echo "##[group]Autogen.sh"
-  run ./autogen.sh
-  echo "##[endgroup]"
-
-  echo "##[group]Configure"
-  run ./configure \
-    --prefix=/usr \
-    --enable-pyzfs \
-    --enable-debug \
-    --enable-debuginfo
-  echo "##[endgroup]"
-
-  echo "##[group]Build"
-  run make native-deb-kmod native-deb-utils
-  echo "##[endgroup]"
-
-  echo "##[group]Install"
-  # Do kmod install.  Note that when you build the native debs, the
-  # packages themselves are placed in parent directory '../' rather than
-  # in the source directory like the rpms are.
-  run sudo apt-get -y install $(find ../ | grep -E '\.deb$' \
-    | grep -Ev 'dkms|dracut')
-  echo "##[endgroup]"
-}
-
-# Debug: show kernel cmdline
-if [ -f /proc/cmdline ] ; then
-  cat /proc/cmdline || true
-fi
-
-# save some sysinfo
-uname -a > /var/tmp/uname.txt
-
-cd $HOME/zfs
-export PATH="$PATH:/sbin:/usr/sbin:/usr/local/sbin"
-
-# build
-case "$1" in
-  freebsd*)
-    freebsd
-    ;;
-  alma*|centos*)
-    rpm_build_and_install "--with-spec=redhat"
-    ;;
-  fedora*)
-    rpm_build_and_install
-    ;;
-  debian*|ubuntu*)
-    deb_build_and_install
-    ;;
-  *)
-    linux
-    ;;
-esac
-
-# building the zfs module was ok
-echo 0 > /var/tmp/build-exitcode.txt
-
-# reset cloud-init configuration and poweroff
-sudo cloud-init clean --logs
-sync && sleep 2 && sudo poweroff &
-exit 0
+ssh zfs@vm0 '$HOME/zfs/.github/workflows/scripts/qemu-4-build-vm.sh' $@

.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 $((1024*RAM)) \
     --memballoon model=virtio \
     --graphics none \
@@ -104,9 +109,7 @@ rm crontab.txt
 # check if the machines are okay
 echo "Waiting for vm's to come up...  (${VMs}x CPU=$CPU RAM=$RAM)"
 for i in $(seq 1 $VMs); do
-  while true; do
-    ssh 2>/dev/null zfs@192.168.122.1$i "uname -a" && break
-  done
+  .github/workflows/scripts/qemu-wait-for-vm.sh vm$i
 done
 echo "All $VMs VMs are up now."
 

.github/workflows/scripts/qemu-7-prepare.sh

@@ -29,8 +29,9 @@ MERGE="$BASE/.github/workflows/scripts/merge_summary.awk"
 
 # catch result files of testings (vm's should be there)
 for i in $(seq 1 $VMs); do
-  rsync -arL zfs@192.168.122.1$i:$RESPATH/current $RESPATH/vm$i || true
-  scp zfs@192.168.122.1$i:"/var/tmp/*.txt" $RESPATH/vm$i || true
+  rsync -arL zfs@vm$i:$RESPATH/current $RESPATH/vm$i || true
+  scp zfs@vm$i:"/var/tmp/*.txt" $RESPATH/vm$i || true
+  scp zfs@vm$i:"/var/tmp/*.rpm" $RESPATH/vm$i || true
 done
 cp -f /var/tmp/*.txt $RESPATH || true
 cd $RESPATH

.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/scripts/qemu-prepare-for-build.sh

@@ -0,0 +1,8 @@
+#!/usr/bin/env bash
+
+# Helper script to run after installing dependencies.  This brings the VM back
+# up and copies over the zfs source directory.
+echo "Build modules in QEMU machine"
+sudo virsh start openzfs
+.github/workflows/scripts/qemu-wait-for-vm.sh vm0
+rsync -ar $HOME/work/zfs/zfs zfs@vm0:./

.github/workflows/scripts/qemu-test-repo-vm.sh

@@ -0,0 +1,90 @@
+#!/bin/bash
+#
+# Do a test install of ZFS from an external repository.
+#
+# USAGE:
+#
+# 	./qemu-test-repo-vm [URL]
+#
+# URL:		URL to use instead of http://download.zfsonlinux.org
+#		If blank, use the default repo from zfs-release RPM.
+
+set -e
+
+source /etc/os-release
+OS="$ID"
+VERSION="$VERSION_ID"
+
+ALTHOST=""
+if [ -n "$1" ] ; then
+	ALTHOST="$1"
+fi
+
+# Write summary to /tmp/repo so our artifacts scripts pick it up
+mkdir /tmp/repo
+SUMMARY=/tmp/repo/$OS-$VERSION-summary.txt
+
+# $1: Repo 'zfs' 'zfs-kmod' 'zfs-testing' 'zfs-testing-kmod'
+# $2: (optional) Alternate host than 'http://download.zfsonlinux.org' to
+#     install from.  Blank means use default from zfs-release RPM.
+function test_install {
+	repo=$1
+	host=""
+	if [ -n "$2" ] ; then
+		host=$2
+	fi
+
+	args="--disablerepo=zfs --enablerepo=$repo"
+
+	# If we supplied an alternate repo URL, and have not already edited
+	# zfs.repo, then update the repo file.
+	if [ -n "$host" ] && ! grep -q $host /etc/yum.repos.d/zfs.repo ; then
+		sudo sed -i "s;baseurl=http://download.zfsonlinux.org;baseurl=$host;g" /etc/yum.repos.d/zfs.repo
+	fi
+
+	sudo dnf -y install $args zfs zfs-test
+
+	# Load modules and create a simple pool as a sanity test.
+	sudo /usr/share/zfs/zfs.sh -r
+	truncate -s 100M /tmp/file
+	sudo zpool create tank /tmp/file
+	sudo zpool status
+
+	# Print out repo name, rpm installed (kmod or dkms), and repo URL
+	baseurl=$(grep -A 5 "\[$repo\]" /etc/yum.repos.d/zfs.repo  | awk -F'=' '/baseurl=/{print $2; exit}')
+	package=$(sudo rpm -qa | grep zfs | grep -E 'kmod|dkms')
+
+	echo "$repo $package $baseurl" >> $SUMMARY
+
+	sudo zpool destroy tank
+	sudo rm /tmp/file
+	sudo dnf -y remove zfs
+}
+
+echo "##[group]Installing from repo"
+# The openzfs docs are the authoritative instructions for the install.  Use
+# the specific version of zfs-release RPM it recommends.
+case $OS in
+almalinux*)
+	url='https://raw.githubusercontent.com/openzfs/openzfs-docs/refs/heads/master/docs/Getting%20Started/RHEL-based%20distro/index.rst'
+	name=$(curl -Ls $url | grep 'dnf install' | grep -Eo 'zfs-release-[0-9]+-[0-9]+')
+	sudo dnf -y install https://zfsonlinux.org/epel/$name$(rpm --eval "%{dist}").noarch.rpm 2>&1
+	sudo rpm -qi zfs-release
+	test_install zfs $ALTHOST
+	test_install zfs-kmod $ALTHOST
+	test_install zfs-testing $ALTHOST
+	test_install zfs-testing-kmod $ALTHOST
+	;;
+fedora*)
+	url='https://raw.githubusercontent.com/openzfs/openzfs-docs/refs/heads/master/docs/Getting%20Started/Fedora/index.rst'
+	name=$(curl -Ls $url | grep 'dnf install' | grep -Eo 'zfs-release-[0-9]+-[0-9]+')
+	sudo dnf -y install -y https://zfsonlinux.org/fedora/$name$(rpm --eval "%{dist}").noarch.rpm
+	test_install zfs $ALTHOST
+	;;
+esac
+echo "##[endgroup]"
+
+# Write out a simple version of the summary here. Later on we will collate all
+# the summaries and put them into a nice table in the workflow Summary page.
+echo "Summary: "
+cat $SUMMARY

.github/workflows/scripts/qemu-wait-for-vm.sh

@@ -0,0 +1,10 @@
+#!/bin/bash
+#
+# Wait for a VM to boot up and become active.  This is used in a number of our
+# scripts.
+#
+# $1: VM hostname or IP address
+
+while pidof /usr/bin/qemu-system-x86_64 >/dev/null; do
+  ssh 2>/dev/null zfs@$1 "uname -a" && break
+done

.github/workflows/scripts/replace-dupes-with-symlinks.sh

@@ -0,0 +1,32 @@
+#!/bin/bash
+#
+# Recursively go though a directory structure and replace duplicate files with
+# symlinks.  This cuts down our RPM repo size by ~25%.
+#
+#	replace-dupes-with-symlinks.sh [DIR]
+#
+# DIR: Directory to traverse.  Defaults to current directory if not specified.
+#
+
+src="$1"
+if [ -z "$src" ] ; then
+	src="."
+fi
+
+declare -A db
+
+pushd "$src"
+while read line ; do
+	bn="$(basename $line)"
+	if [ -z "${db[$bn]}" ] ; then
+		# First time this file has been seen
+		db[$bn]="$line"
+	else
+		if diff -b "$line" "${db[$bn]}" &>/dev/null ; then
+			# Files are the same, make a symlink
+			rm "$line"
+			ln -sr "${db[$bn]}" "$line"
+		fi
+	fi
+done <<< "$(find . -type f)"
+popd

.github/workflows/zfs-qemu-packages.yml

@@ -0,0 +1,140 @@
+# This workflow is used to build and test RPM packages.  It is a
+# 'workflow_dispatch' workflow, which means it gets run manually.
+#
+# The workflow has a dropdown menu with two options:
+#
+# Build RPMs - Build release RPMs and tarballs and put them into an artifact
+#              ZIP file.  The directory structure used in the ZIP file mirrors
+#              the ZFS yum repo.
+#
+# Test repo -  Test install the ZFS RPMs from the ZFS repo.  On EL distos, this
+#              will do a DKMS and KMOD test install from both the regular and
+#              testing repos.  On Fedora, it will do a DKMS install from the
+#              regular repo.  All test install results will be displayed in the
+#              Summary page.  Note that the workflow provides an optional text
+#              text box where you can specify the full URL to an alternate repo.
+#              If left blank, it will install from the default repo from the
+#              zfs-release RPM (http://download.zfsonlinux.org).
+#
+# Most users will never need to use this workflow.  It will be used primary by
+# ZFS admins for building and testing releases.
+#
+name: zfs-qemu-packages
+
+on:
+  workflow_dispatch:
+    inputs:
+      test_type:
+        type: choice
+        required: false
+        default: "Build RPMs"
+        description: "Build RPMs or test the repo?"
+        options:
+        - "Build RPMs"
+        - "Test repo"
+      repo_url:
+        type: string
+        required: false
+        default: ""
+        description: "(optional) repo URL (blank: use http://download.zfsonlinux.org)"
+concurrency:
+  group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
+  cancel-in-progress: true
+
+jobs:
+  zfs-qemu-packages-jobs:
+    name: qemu-VMs
+    strategy:
+      fail-fast: false
+      matrix:
+        os: ['almalinux8', 'almalinux9', 'fedora40', 'fedora41']
+    runs-on: ubuntu-24.04
+    steps:
+    - uses: actions/checkout@v4
+      with:
+        ref: ${{ github.event.pull_request.head.sha }}
+
+    - name: Setup QEMU
+      timeout-minutes: 10
+      run: .github/workflows/scripts/qemu-1-setup.sh
+
+    - name: Start build machine
+      timeout-minutes: 10
+      run: .github/workflows/scripts/qemu-2-start.sh ${{ matrix.os }}
+
+    - name: Install dependencies
+      timeout-minutes: 20
+      run: |
+        .github/workflows/scripts/qemu-3-deps.sh ${{ matrix.os }}
+
+    - name: Build modules or Test repo
+      timeout-minutes: 30
+      run: |
+        set -e
+        if [ "${{ github.event.inputs.test_type }}" == "Test repo" ] ; then
+                # Bring VM back up and copy over zfs source
+                .github/workflows/scripts/qemu-prepare-for-build.sh
+
+                mkdir -p /tmp/repo
+                ssh zfs@vm0 '$HOME/zfs/.github/workflows/scripts/qemu-test-repo-vm.sh' ${{ github.event.inputs.repo_url }}
+        else
+                .github/workflows/scripts/qemu-4-build.sh --repo --release --dkms --tarball ${{ matrix.os }}
+        fi
+
+    - name: Prepare artifacts
+      if: always()
+      timeout-minutes: 10
+      run: |
+        rsync -a zfs@vm0:/tmp/repo /tmp || true
+        .github/workflows/scripts/replace-dupes-with-symlinks.sh /tmp/repo
+        tar -cf ${{ matrix.os }}-repo.tar -C /tmp repo
+
+    - uses: actions/upload-artifact@v4
+      id: artifact-upload
+      if: always()
+      with:
+        name: ${{ matrix.os }}-repo
+        path: ${{ matrix.os }}-repo.tar
+        compression-level: 0
+        retention-days: 2
+        if-no-files-found: ignore
+
+  combine_repos:
+    if: always()
+    needs: [zfs-qemu-packages-jobs]
+    name: "Results"
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/download-artifact@v4
+      id: artifact-download
+      if: always()
+    - name: Test Summary
+      if: always()
+      run: |
+        for i in $(find . -type f -iname "*.tar") ; do
+          tar -xf $i -C /tmp
+        done
+        tar -cf all-repo.tar -C /tmp repo
+
+        # If we're installing from a repo, print out the summary of the versions
+        # that got installed using Markdown.
+        if [ "${{ github.event.inputs.test_type }}" == "Test repo" ] ; then
+          cd /tmp/repo
+          for i in $(ls *.txt) ; do
+            nicename="$(echo $i | sed 's/.txt//g; s/-/ /g')"
+            echo "### $nicename" >> $GITHUB_STEP_SUMMARY
+            echo "|repo|RPM|URL|" >> $GITHUB_STEP_SUMMARY
+            echo "|:---|:---|:---|" >> $GITHUB_STEP_SUMMARY
+            awk '{print "|"$1"|"$2"|"$3"|"}' $i >> $GITHUB_STEP_SUMMARY
+          done
+        fi
+
+    - uses: actions/upload-artifact@v4
+      id: artifact-upload2
+      if: always()
+      with:
+        name: all-repo
+        path: all-repo.tar
+        compression-level: 0
+        retention-days: 5
+        if-no-files-found: ignore

.github/workflows/zfs-qemu.yml

@@ -3,6 +3,18 @@ name: zfs-qemu
 on:
   push:
   pull_request:
+  workflow_dispatch:
+    inputs:
+      include_stream9:
+        type: boolean
+        required: false
+        default: false
+        description: 'Test on CentOS 9 stream'
+      include_stream10:
+        type: boolean
+        required: false
+        default: false
+        description: 'Test on CentOS 10 stream'
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
@@ -22,8 +34,8 @@ 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", "debian11", "debian12", "fedora40", "fedora41", "freebsd13-3r", "freebsd13-4s", "freebsd14-1r", "freebsd14-2s", "freebsd15-0c", "ubuntu20", "ubuntu22", "ubuntu24"]'
+          QUICK_OS='["almalinux8", "almalinux9", "debian12", "fedora41", "freebsd13-3r", "freebsd14-2r", "ubuntu24"]'
           # determine CI type when running on PR
           ci_type="full"
           if ${{ github.event_name == 'pull_request' }}; then
@@ -37,6 +49,16 @@ jobs:
             os_selection="$FULL_OS"
           fi
           os_json=$(echo ${os_selection} | jq -c)
+
+          # Add optional runners
+          if [ "${{ github.event.inputs.include_stream9 }}" == 'true' ]; then
+            os_json=$(echo $os_json | jq -c '. += ["centos-stream9"]')
+          fi
+          if [ "${{ github.event.inputs.include_stream10 }}" == 'true' ]; then
+            os_json=$(echo $os_json | jq -c '. += ["centos-stream10"]')
+          fi
+
+          echo $os_json
           echo "os=$os_json" >> $GITHUB_OUTPUT
           echo "ci_type=$ci_type" >> $GITHUB_OUTPUT
 
@@ -46,10 +68,13 @@ 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 variants of 2024-12:
+        # FreeBSD Release: freebsd13-3r, freebsd13-4r, freebsd14-1r, freebsd14-2r
+        # FreeBSD Stable:  freebsd13-4s, freebsd14-2s
+        # FreeBSD Current: freebsd15-0c
         os: ${{ fromJson(needs.test-config.outputs.test_os) }}
     runs-on: ubuntu-24.04
     steps:
@@ -67,31 +92,11 @@ jobs:
 
     - name: Install dependencies
       timeout-minutes: 20
-      run: |
-        echo "Install dependencies in QEMU machine"
-        IP=192.168.122.10
-        while pidof /usr/bin/qemu-system-x86_64 >/dev/null; do
-          ssh 2>/dev/null zfs@$IP "uname -a" && break
-        done
-        scp .github/workflows/scripts/qemu-3-deps.sh zfs@$IP:qemu-3-deps.sh
-        PID=`pidof /usr/bin/qemu-system-x86_64`
-        ssh zfs@$IP '$HOME/qemu-3-deps.sh' ${{ matrix.os }}
-        # wait for poweroff to succeed
-        tail --pid=$PID -f /dev/null
-        sleep 5 # avoid this: "error: Domain is already active"
-        rm -f $HOME/.ssh/known_hosts
+      run: .github/workflows/scripts/qemu-3-deps.sh ${{ matrix.os }}
 
     - name: Build modules
       timeout-minutes: 30
-      run: |
-        echo "Build modules in QEMU machine"
-        sudo virsh start openzfs
-        IP=192.168.122.10
-        while pidof /usr/bin/qemu-system-x86_64 >/dev/null; do
-          ssh 2>/dev/null zfs@$IP "uname -a" && break
-        done
-        rsync -ar $HOME/work/zfs/zfs zfs@$IP:./
-        ssh zfs@$IP '$HOME/zfs/.github/workflows/scripts/qemu-4-build.sh' ${{ matrix.os }}
+      run: .github/workflows/scripts/qemu-4-build.sh --poweroff --enable-debug ${{ matrix.os }}
 
     - name: Setup testing machines
       timeout-minutes: 5

.github/workflows/zloop.yml

@@ -20,7 +20,7 @@ jobs:
     - name: Install dependencies
       run: |
         sudo apt-get purge -y snapd google-chrome-stable firefox
-        ONLY_DEPS=1 .github/workflows/scripts/qemu-3-deps.sh ubuntu24
+        ONLY_DEPS=1 .github/workflows/scripts/qemu-3-deps-vm.sh ubuntu24
     - name: Autogen.sh
       run: |
         sed -i '/DEBUG_CFLAGS="-Werror"/s/^/#/' config/zfs-build.m4

.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

@@ -1,10 +1,10 @@
 Meta:          1
 Name:          zfs
 Branch:        1.0
-Version:       2.3.0
-Release:       rc1
+Version:       2.3.1
+Release:       1
 Release-Tags:  relext
 License:       CDDL
 Author:        OpenZFS
-Linux-Maximum: 6.11
+Linux-Maximum: 6.13
 Linux-Minimum: 4.18

RELEASES.md

@@ -28,7 +28,7 @@ Two release branches are maintained for OpenZFS, they are:
     Minor changes to support these distribution kernels will be applied as
     needed. New kernel versions released after the OpenZFS LTS release are
     not supported. LTS releases will receive patches for at least 2 years.
-    The current LTS release is OpenZFS 2.1.
+    The current LTS release is OpenZFS 2.2.
 
   * OpenZFS current - Tracks the newest MAJOR.MINOR release. This branch
     includes support for the latest OpenZFS features and recently releases

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

@@ -122,7 +122,7 @@ static int flagbits[256];
 
 static uint64_t max_inflight_bytes = 256 * 1024 * 1024; /* 256MB */
 static int leaked_objects = 0;
-static range_tree_t *mos_refd_objs;
+static zfs_range_tree_t *mos_refd_objs;
 static spa_t *spa;
 static objset_t *os;
 static boolean_t kernel_init_done;
@@ -325,7 +325,7 @@ typedef struct metaslab_verify {
 	/*
 	 * What's currently allocated for this metaslab.
 	 */
-	range_tree_t *mv_allocated;
+	zfs_range_tree_t *mv_allocated;
 } metaslab_verify_t;
 
 typedef void ll_iter_t(dsl_deadlist_t *ll, void *arg);
@@ -417,7 +417,7 @@ metaslab_spacemap_validation_cb(space_map_entry_t *sme, void *arg)
 	uint64_t txg = sme->sme_txg;
 
 	if (sme->sme_type == SM_ALLOC) {
-		if (range_tree_contains(mv->mv_allocated,
+		if (zfs_range_tree_contains(mv->mv_allocated,
 		    offset, size)) {
 			(void) printf("ERROR: DOUBLE ALLOC: "
 			    "%llu [%llx:%llx] "
@@ -426,11 +426,11 @@ metaslab_spacemap_validation_cb(space_map_entry_t *sme, void *arg)
 			    (u_longlong_t)size, (u_longlong_t)mv->mv_vdid,
 			    (u_longlong_t)mv->mv_msid);
 		} else {
-			range_tree_add(mv->mv_allocated,
+			zfs_range_tree_add(mv->mv_allocated,
 			    offset, size);
 		}
 	} else {
-		if (!range_tree_contains(mv->mv_allocated,
+		if (!zfs_range_tree_contains(mv->mv_allocated,
 		    offset, size)) {
 			(void) printf("ERROR: DOUBLE FREE: "
 			    "%llu [%llx:%llx] "
@@ -439,7 +439,7 @@ metaslab_spacemap_validation_cb(space_map_entry_t *sme, void *arg)
 			    (u_longlong_t)size, (u_longlong_t)mv->mv_vdid,
 			    (u_longlong_t)mv->mv_msid);
 		} else {
-			range_tree_remove(mv->mv_allocated,
+			zfs_range_tree_remove(mv->mv_allocated,
 			    offset, size);
 		}
 	}
@@ -614,11 +614,11 @@ livelist_metaslab_validate(spa_t *spa)
 			    (longlong_t)vd->vdev_ms_count);
 
 			uint64_t shift, start;
-			range_seg_type_t type =
+			zfs_range_seg_type_t type =
 			    metaslab_calculate_range_tree_type(vd, m,
 			    &start, &shift);
 			metaslab_verify_t mv;
-			mv.mv_allocated = range_tree_create(NULL,
+			mv.mv_allocated = zfs_range_tree_create(NULL,
 			    type, NULL, start, shift);
 			mv.mv_vdid = vd->vdev_id;
 			mv.mv_msid = m->ms_id;
@@ -633,8 +633,8 @@ livelist_metaslab_validate(spa_t *spa)
 			spacemap_check_ms_sm(m->ms_sm, &mv);
 			spacemap_check_sm_log(spa, &mv);
 
-			range_tree_vacate(mv.mv_allocated, NULL, NULL);
-			range_tree_destroy(mv.mv_allocated);
+			zfs_range_tree_vacate(mv.mv_allocated, NULL, NULL);
+			zfs_range_tree_destroy(mv.mv_allocated);
 			zfs_btree_clear(&mv.mv_livelist_allocs);
 			zfs_btree_destroy(&mv.mv_livelist_allocs);
 		}
@@ -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);
@@ -1633,9 +1633,9 @@ static void
 dump_metaslab_stats(metaslab_t *msp)
 {
 	char maxbuf[32];
-	range_tree_t *rt = msp->ms_allocatable;
+	zfs_range_tree_t *rt = msp->ms_allocatable;
 	zfs_btree_t *t = &msp->ms_allocatable_by_size;
-	int free_pct = range_tree_space(rt) * 100 / msp->ms_size;
+	int free_pct = zfs_range_tree_space(rt) * 100 / msp->ms_size;
 
 	/* max sure nicenum has enough space */
 	_Static_assert(sizeof (maxbuf) >= NN_NUMBUF_SZ, "maxbuf truncated");
@@ -1646,7 +1646,7 @@ dump_metaslab_stats(metaslab_t *msp)
 	    "segments", zfs_btree_numnodes(t), "maxsize", maxbuf,
 	    "freepct", free_pct);
 	(void) printf("\tIn-memory histogram:\n");
-	dump_histogram(rt->rt_histogram, RANGE_TREE_HISTOGRAM_SIZE, 0);
+	dump_histogram(rt->rt_histogram, ZFS_RANGE_TREE_HISTOGRAM_SIZE, 0);
 }
 
 static void
@@ -1668,7 +1668,7 @@ dump_metaslab(metaslab_t *msp)
 	if (dump_opt['m'] > 2 && !dump_opt['L']) {
 		mutex_enter(&msp->ms_lock);
 		VERIFY0(metaslab_load(msp));
-		range_tree_stat_verify(msp->ms_allocatable);
+		zfs_range_tree_stat_verify(msp->ms_allocatable);
 		dump_metaslab_stats(msp);
 		metaslab_unload(msp);
 		mutex_exit(&msp->ms_lock);
@@ -1769,7 +1769,8 @@ dump_metaslab_groups(spa_t *spa, boolean_t show_special)
 			(void) printf("%3llu%%\n",
 			    (u_longlong_t)mg->mg_fragmentation);
 		}
-		dump_histogram(mg->mg_histogram, RANGE_TREE_HISTOGRAM_SIZE, 0);
+		dump_histogram(mg->mg_histogram,
+		    ZFS_RANGE_TREE_HISTOGRAM_SIZE, 0);
 	}
 
 	(void) printf("\tpool %s\tfragmentation", spa_name(spa));
@@ -1778,7 +1779,7 @@ dump_metaslab_groups(spa_t *spa, boolean_t show_special)
 		(void) printf("\t%3s\n", "-");
 	else
 		(void) printf("\t%3llu%%\n", (u_longlong_t)fragmentation);
-	dump_histogram(mc->mc_histogram, RANGE_TREE_HISTOGRAM_SIZE, 0);
+	dump_histogram(mc->mc_histogram, ZFS_RANGE_TREE_HISTOGRAM_SIZE, 0);
 }
 
 static void
@@ -1967,17 +1968,53 @@ 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 +2028,7 @@ 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 +2053,8 @@ 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 +2077,52 @@ 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 +2181,6 @@ 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 +2191,8 @@ 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 +2208,54 @@ dump_brt(spa_t *spa)
 	if (dump_opt['T'] < 3)
 		return;
 
-	char dva[64];
-	printf("\n%-16s %-10s\n", "DVA", "REFCNT");
+	/* -TTT shows a per-vdev histograms; -TTTT shows all entries */
+	boolean_t do_histo = dump_opt['T'] == 3;
 
-	for (uint64_t vdevid = 0; vdevid < brt->brt_nvdevs; vdevid++) {
-		brt_vdev_t *brtvd = &brt->brt_vdevs[vdevid];
-		if (brtvd == NULL || !brtvd->bv_initiated)
+	char dva[64];
+
+	if (!do_histo)
+		printf("\n%-16s %-10s\n", "DVA", "REFCNT");
+
+	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);
+		}
 	}
 }
 
@@ -2217,12 +2293,12 @@ dump_dtl(vdev_t *vd, int indent)
 	    required ? "DTL-required" : "DTL-expendable");
 
 	for (int t = 0; t < DTL_TYPES; t++) {
-		range_tree_t *rt = vd->vdev_dtl[t];
-		if (range_tree_space(rt) == 0)
+		zfs_range_tree_t *rt = vd->vdev_dtl[t];
+		if (zfs_range_tree_space(rt) == 0)
 			continue;
 		(void) snprintf(prefix, sizeof (prefix), "\t%*s%s",
 		    indent + 2, "", name[t]);
-		range_tree_walk(rt, dump_dtl_seg, prefix);
+		zfs_range_tree_walk(rt, dump_dtl_seg, prefix);
 		if (dump_opt['d'] > 5 && vd->vdev_children == 0)
 			dump_spacemap(spa->spa_meta_objset,
 			    vd->vdev_dtl_sm);
@@ -4266,6 +4342,10 @@ 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)) {
@@ -6179,9 +6259,9 @@ load_unflushed_svr_segs_cb(spa_t *spa, space_map_entry_t *sme,
 		return (0);
 
 	if (sme->sme_type == SM_ALLOC)
-		range_tree_add(svr->svr_allocd_segs, offset, size);
+		zfs_range_tree_add(svr->svr_allocd_segs, offset, size);
 	else
-		range_tree_remove(svr->svr_allocd_segs, offset, size);
+		zfs_range_tree_remove(svr->svr_allocd_segs, offset, size);
 
 	return (0);
 }
@@ -6235,18 +6315,20 @@ zdb_claim_removing(spa_t *spa, zdb_cb_t *zcb)
 	vdev_t *vd = vdev_lookup_top(spa, svr->svr_vdev_id);
 	vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
 
-	ASSERT0(range_tree_space(svr->svr_allocd_segs));
+	ASSERT0(zfs_range_tree_space(svr->svr_allocd_segs));
 
-	range_tree_t *allocs = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0);
+	zfs_range_tree_t *allocs = zfs_range_tree_create(NULL, ZFS_RANGE_SEG64,
+	    NULL, 0, 0);
 	for (uint64_t msi = 0; msi < vd->vdev_ms_count; msi++) {
 		metaslab_t *msp = vd->vdev_ms[msi];
 
-		ASSERT0(range_tree_space(allocs));
+		ASSERT0(zfs_range_tree_space(allocs));
 		if (msp->ms_sm != NULL)
 			VERIFY0(space_map_load(msp->ms_sm, allocs, SM_ALLOC));
-		range_tree_vacate(allocs, range_tree_add, svr->svr_allocd_segs);
+		zfs_range_tree_vacate(allocs, zfs_range_tree_add,
+		    svr->svr_allocd_segs);
 	}
-	range_tree_destroy(allocs);
+	zfs_range_tree_destroy(allocs);
 
 	iterate_through_spacemap_logs(spa, load_unflushed_svr_segs_cb, svr);
 
@@ -6255,12 +6337,12 @@ zdb_claim_removing(spa_t *spa, zdb_cb_t *zcb)
 	 * because we have not allocated mappings for
 	 * it yet.
 	 */
-	range_tree_clear(svr->svr_allocd_segs,
+	zfs_range_tree_clear(svr->svr_allocd_segs,
 	    vdev_indirect_mapping_max_offset(vim),
 	    vd->vdev_asize - vdev_indirect_mapping_max_offset(vim));
 
-	zcb->zcb_removing_size += range_tree_space(svr->svr_allocd_segs);
-	range_tree_vacate(svr->svr_allocd_segs, claim_segment_cb, vd);
+	zcb->zcb_removing_size += zfs_range_tree_space(svr->svr_allocd_segs);
+	zfs_range_tree_vacate(svr->svr_allocd_segs, claim_segment_cb, vd);
 
 	spa_config_exit(spa, SCL_CONFIG, FTAG);
 }
@@ -6363,7 +6445,8 @@ checkpoint_sm_exclude_entry_cb(space_map_entry_t *sme, void *arg)
 	 * also verify that the entry is there to begin with.
 	 */
 	mutex_enter(&ms->ms_lock);
-	range_tree_remove(ms->ms_allocatable, sme->sme_offset, sme->sme_run);
+	zfs_range_tree_remove(ms->ms_allocatable, sme->sme_offset,
+	    sme->sme_run);
 	mutex_exit(&ms->ms_lock);
 
 	cseea->cseea_checkpoint_size += sme->sme_run;
@@ -6494,9 +6577,9 @@ load_unflushed_cb(spa_t *spa, space_map_entry_t *sme, uint64_t txg, void *arg)
 		return (0);
 
 	if (*uic_maptype == sme->sme_type)
-		range_tree_add(ms->ms_allocatable, offset, size);
+		zfs_range_tree_add(ms->ms_allocatable, offset, size);
 	else
-		range_tree_remove(ms->ms_allocatable, offset, size);
+		zfs_range_tree_remove(ms->ms_allocatable, offset, size);
 
 	return (0);
 }
@@ -6530,7 +6613,7 @@ load_concrete_ms_allocatable_trees(spa_t *spa, maptype_t maptype)
 			    (longlong_t)vd->vdev_ms_count);
 
 			mutex_enter(&msp->ms_lock);
-			range_tree_vacate(msp->ms_allocatable, NULL, NULL);
+			zfs_range_tree_vacate(msp->ms_allocatable, NULL, NULL);
 
 			/*
 			 * We don't want to spend the CPU manipulating the
@@ -6563,7 +6646,7 @@ load_indirect_ms_allocatable_tree(vdev_t *vd, metaslab_t *msp,
 	vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
 
 	mutex_enter(&msp->ms_lock);
-	range_tree_vacate(msp->ms_allocatable, NULL, NULL);
+	zfs_range_tree_vacate(msp->ms_allocatable, NULL, NULL);
 
 	/*
 	 * We don't want to spend the CPU manipulating the
@@ -6587,7 +6670,7 @@ load_indirect_ms_allocatable_tree(vdev_t *vd, metaslab_t *msp,
 		 */
 		ASSERT3U(ent_offset + ent_len, <=,
 		    msp->ms_start + msp->ms_size);
-		range_tree_add(msp->ms_allocatable, ent_offset, ent_len);
+		zfs_range_tree_add(msp->ms_allocatable, ent_offset, ent_len);
 	}
 
 	if (!msp->ms_loaded)
@@ -6733,7 +6816,7 @@ zdb_check_for_obsolete_leaks(vdev_t *vd, zdb_cb_t *zcb)
 		for (uint64_t inner_offset = 0;
 		    inner_offset < DVA_GET_ASIZE(&vimep->vimep_dst);
 		    inner_offset += 1ULL << vd->vdev_ashift) {
-			if (range_tree_contains(msp->ms_allocatable,
+			if (zfs_range_tree_contains(msp->ms_allocatable,
 			    offset + inner_offset, 1ULL << vd->vdev_ashift)) {
 				obsolete_bytes += 1ULL << vd->vdev_ashift;
 			}
@@ -6816,10 +6899,10 @@ zdb_leak_fini(spa_t *spa, zdb_cb_t *zcb)
 			 * not referenced, which is not a bug.
 			 */
 			if (vd->vdev_ops == &vdev_indirect_ops) {
-				range_tree_vacate(msp->ms_allocatable,
+				zfs_range_tree_vacate(msp->ms_allocatable,
 				    NULL, NULL);
 			} else {
-				range_tree_vacate(msp->ms_allocatable,
+				zfs_range_tree_vacate(msp->ms_allocatable,
 				    zdb_leak, vd);
 			}
 			if (msp->ms_loaded) {
@@ -6874,7 +6957,7 @@ 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);
 	}
@@ -7717,7 +7800,7 @@ verify_checkpoint_sm_entry_cb(space_map_entry_t *sme, void *arg)
 	 * their respective ms_allocateable trees should not contain them.
 	 */
 	mutex_enter(&ms->ms_lock);
-	range_tree_verify_not_present(ms->ms_allocatable,
+	zfs_range_tree_verify_not_present(ms->ms_allocatable,
 	    sme->sme_offset, sme->sme_run);
 	mutex_exit(&ms->ms_lock);
 
@@ -7868,8 +7951,9 @@ verify_checkpoint_ms_spacemaps(spa_t *checkpoint, spa_t *current)
 			 * This way we ensure that none of the blocks that
 			 * are part of the checkpoint were freed by mistake.
 			 */
-			range_tree_walk(ckpoint_msp->ms_allocatable,
-			    (range_tree_func_t *)range_tree_verify_not_present,
+			zfs_range_tree_walk(ckpoint_msp->ms_allocatable,
+			    (zfs_range_tree_func_t *)
+			    zfs_range_tree_verify_not_present,
 			    current_msp->ms_allocatable);
 		}
 	}
@@ -8009,7 +8093,7 @@ static void
 mos_obj_refd(uint64_t obj)
 {
 	if (obj != 0 && mos_refd_objs != NULL)
-		range_tree_add(mos_refd_objs, obj, 1);
+		zfs_range_tree_add(mos_refd_objs, obj, 1);
 }
 
 /*
@@ -8019,8 +8103,8 @@ static void
 mos_obj_refd_multiple(uint64_t obj)
 {
 	if (obj != 0 && mos_refd_objs != NULL &&
-	    !range_tree_contains(mos_refd_objs, obj, 1))
-		range_tree_add(mos_refd_objs, obj, 1);
+	    !zfs_range_tree_contains(mos_refd_objs, obj, 1))
+		zfs_range_tree_add(mos_refd_objs, obj, 1);
 }
 
 static void
@@ -8204,14 +8288,11 @@ 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) {
-				mos_obj_refd(brtvd->bv_mos_brtvdev);
-				mos_obj_refd(brtvd->bv_mos_entries);
-			}
+	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);
 		}
 	}
 
@@ -8220,8 +8301,8 @@ dump_mos_leaks(spa_t *spa)
 	 */
 	uint64_t object = 0;
 	while (dmu_object_next(mos, &object, B_FALSE, 0) == 0) {
-		if (range_tree_contains(mos_refd_objs, object, 1)) {
-			range_tree_remove(mos_refd_objs, object, 1);
+		if (zfs_range_tree_contains(mos_refd_objs, object, 1)) {
+			zfs_range_tree_remove(mos_refd_objs, object, 1);
 		} else {
 			dmu_object_info_t doi;
 			const char *name;
@@ -8239,11 +8320,11 @@ dump_mos_leaks(spa_t *spa)
 			rv = 2;
 		}
 	}
-	(void) range_tree_walk(mos_refd_objs, mos_leaks_cb, NULL);
-	if (!range_tree_is_empty(mos_refd_objs))
+	(void) zfs_range_tree_walk(mos_refd_objs, mos_leaks_cb, NULL);
+	if (!zfs_range_tree_is_empty(mos_refd_objs))
 		rv = 2;
-	range_tree_vacate(mos_refd_objs, NULL, NULL);
-	range_tree_destroy(mos_refd_objs);
+	zfs_range_tree_vacate(mos_refd_objs, NULL, NULL);
+	zfs_range_tree_destroy(mos_refd_objs);
 	return (rv);
 }
 
@@ -8365,8 +8446,8 @@ dump_zpool(spa_t *spa)
 
 	if (dump_opt['d'] || dump_opt['i']) {
 		spa_feature_t f;
-		mos_refd_objs = range_tree_create(NULL, RANGE_SEG64, NULL, 0,
-		    0);
+		mos_refd_objs = zfs_range_tree_create(NULL, ZFS_RANGE_SEG64,
+		    NULL, 0, 0);
 		dump_objset(dp->dp_meta_objset);
 
 		if (dump_opt['d'] >= 3) {
@@ -8962,7 +9043,7 @@ zdb_read_block(char *thing, spa_t *spa)
 		const blkptr_t *b = (const blkptr_t *)(void *)
 		    ((uintptr_t)buf + (uintptr_t)blkptr_offset);
 		if (zfs_blkptr_verify(spa, b,
-		    BLK_CONFIG_NEEDED, BLK_VERIFY_ONLY) == B_FALSE) {
+		    BLK_CONFIG_NEEDED, BLK_VERIFY_ONLY)) {
 			abd_return_buf_copy(pabd, buf, lsize);
 			borrowed = B_FALSE;
 			buf = lbuf;
@@ -8971,7 +9052,7 @@ zdb_read_block(char *thing, spa_t *spa)
 			b = (const blkptr_t *)(void *)
 			    ((uintptr_t)buf + (uintptr_t)blkptr_offset);
 			if (lsize == -1 || zfs_blkptr_verify(spa, b,
-			    BLK_CONFIG_NEEDED, BLK_VERIFY_LOG) == B_FALSE) {
+			    BLK_CONFIG_NEEDED, BLK_VERIFY_LOG)) {
 				printf("invalid block pointer at this DVA\n");
 				goto out;
 			}

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

@@ -500,7 +500,7 @@ usage_prop_cb(int prop, void *cb)
 {
 	FILE *fp = cb;
 
-	(void) fprintf(fp, "\t%-15s ", zfs_prop_to_name(prop));
+	(void) fprintf(fp, "\t%-22s ", zfs_prop_to_name(prop));
 
 	if (zfs_prop_readonly(prop))
 		(void) fprintf(fp, " NO    ");
@@ -561,40 +561,40 @@ usage(boolean_t requested)
 		(void) fprintf(fp, "%s",
 		    gettext("\nThe following properties are supported:\n"));
 
-		(void) fprintf(fp, "\n\t%-14s %s  %s   %s\n\n",
+		(void) fprintf(fp, "\n\t%-21s %s  %s   %s\n\n",
 		    "PROPERTY", "EDIT", "INHERIT", "VALUES");
 
 		/* Iterate over all properties */
 		(void) zprop_iter(usage_prop_cb, fp, B_FALSE, B_TRUE,
 		    ZFS_TYPE_DATASET);
 
-		(void) fprintf(fp, "\t%-15s ", "userused@...");
+		(void) fprintf(fp, "\t%-22s ", "userused@...");
 		(void) fprintf(fp, " NO       NO   <size>\n");
-		(void) fprintf(fp, "\t%-15s ", "groupused@...");
+		(void) fprintf(fp, "\t%-22s ", "groupused@...");
 		(void) fprintf(fp, " NO       NO   <size>\n");
-		(void) fprintf(fp, "\t%-15s ", "projectused@...");
+		(void) fprintf(fp, "\t%-22s ", "projectused@...");
 		(void) fprintf(fp, " NO       NO   <size>\n");
-		(void) fprintf(fp, "\t%-15s ", "userobjused@...");
+		(void) fprintf(fp, "\t%-22s ", "userobjused@...");
 		(void) fprintf(fp, " NO       NO   <size>\n");
-		(void) fprintf(fp, "\t%-15s ", "groupobjused@...");
+		(void) fprintf(fp, "\t%-22s ", "groupobjused@...");
 		(void) fprintf(fp, " NO       NO   <size>\n");
-		(void) fprintf(fp, "\t%-15s ", "projectobjused@...");
+		(void) fprintf(fp, "\t%-22s ", "projectobjused@...");
 		(void) fprintf(fp, " NO       NO   <size>\n");
-		(void) fprintf(fp, "\t%-15s ", "userquota@...");
+		(void) fprintf(fp, "\t%-22s ", "userquota@...");
 		(void) fprintf(fp, "YES       NO   <size> | none\n");
-		(void) fprintf(fp, "\t%-15s ", "groupquota@...");
+		(void) fprintf(fp, "\t%-22s ", "groupquota@...");
 		(void) fprintf(fp, "YES       NO   <size> | none\n");
-		(void) fprintf(fp, "\t%-15s ", "projectquota@...");
+		(void) fprintf(fp, "\t%-22s ", "projectquota@...");
 		(void) fprintf(fp, "YES       NO   <size> | none\n");
-		(void) fprintf(fp, "\t%-15s ", "userobjquota@...");
+		(void) fprintf(fp, "\t%-22s ", "userobjquota@...");
 		(void) fprintf(fp, "YES       NO   <size> | none\n");
-		(void) fprintf(fp, "\t%-15s ", "groupobjquota@...");
+		(void) fprintf(fp, "\t%-22s ", "groupobjquota@...");
 		(void) fprintf(fp, "YES       NO   <size> | none\n");
-		(void) fprintf(fp, "\t%-15s ", "projectobjquota@...");
+		(void) fprintf(fp, "\t%-22s ", "projectobjquota@...");
 		(void) fprintf(fp, "YES       NO   <size> | none\n");
-		(void) fprintf(fp, "\t%-15s ", "written@<snap>");
+		(void) fprintf(fp, "\t%-22s ", "written@<snap>");
 		(void) fprintf(fp, " NO       NO   <size>\n");
-		(void) fprintf(fp, "\t%-15s ", "written#<bookmark>");
+		(void) fprintf(fp, "\t%-22s ", "written#<bookmark>");
 		(void) fprintf(fp, " NO       NO   <size>\n");
 
 		(void) fprintf(fp, gettext("\nSizes are specified in bytes "
@@ -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/zinject/zinject.c

@@ -22,7 +22,7 @@
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
  * Copyright (c) 2012, 2015 by Delphix. All rights reserved.
  * Copyright (c) 2017, Intel Corporation.
- * Copyright (c) 2023-2024, Klara Inc.
+ * Copyright (c) 2023-2025, Klara, Inc.
  */
 
 /*
@@ -242,6 +242,36 @@ err_to_str(int err)
 	return ("[unknown]");
 }
 
+static const char *const iotypestrtable[ZINJECT_IOTYPES] = {
+	[ZINJECT_IOTYPE_NULL]	= "null",
+	[ZINJECT_IOTYPE_READ]	= "read",
+	[ZINJECT_IOTYPE_WRITE]	= "write",
+	[ZINJECT_IOTYPE_FREE]	= "free",
+	[ZINJECT_IOTYPE_CLAIM]	= "claim",
+	[ZINJECT_IOTYPE_FLUSH]	= "flush",
+	[ZINJECT_IOTYPE_TRIM]	= "trim",
+	[ZINJECT_IOTYPE_ALL]	= "all",
+	[ZINJECT_IOTYPE_PROBE]	= "probe",
+};
+
+static zinject_iotype_t
+str_to_iotype(const char *arg)
+{
+	for (uint_t iotype = 0; iotype < ZINJECT_IOTYPES; iotype++)
+		if (iotypestrtable[iotype] != NULL &&
+		    strcasecmp(iotypestrtable[iotype], arg) == 0)
+			return (iotype);
+	return (ZINJECT_IOTYPES);
+}
+
+static const char *
+iotype_to_str(zinject_iotype_t iotype)
+{
+	if (iotype >= ZINJECT_IOTYPES || iotypestrtable[iotype] == NULL)
+		return ("[unknown]");
+	return (iotypestrtable[iotype]);
+}
+
 /*
  * Print usage message.
  */
@@ -404,27 +434,30 @@ print_data_handler(int id, const char *pool, zinject_record_t *record,
 
 	if (*count == 0) {
 		(void) printf("%3s  %-15s  %-6s  %-6s  %-8s  %3s  %-4s  "
-		    "%-15s\n", "ID", "POOL", "OBJSET", "OBJECT", "TYPE",
-		    "LVL", "DVAs", "RANGE");
+		    "%-15s  %-6s  %-15s\n", "ID", "POOL", "OBJSET", "OBJECT",
+		    "TYPE", "LVL", "DVAs", "RANGE", "MATCH", "INJECT");
 		(void) printf("---  ---------------  ------  "
-		    "------  --------  ---  ----  ---------------\n");
+		    "------  --------  ---  ----  ---------------  "
+		    "------  ------\n");
 	}
 
 	*count += 1;
 
-	(void) printf("%3d  %-15s  %-6llu  %-6llu  %-8s  %-3d  0x%02x  ",
-	    id, pool, (u_longlong_t)record->zi_objset,
-	    (u_longlong_t)record->zi_object, type_to_name(record->zi_type),
-	    record->zi_level, record->zi_dvas);
-
-
-	if (record->zi_start == 0 &&
-	    record->zi_end == -1ULL)
-		(void) printf("all\n");
+	char rangebuf[32];
+	if (record->zi_start == 0 && record->zi_end == -1ULL)
+		snprintf(rangebuf, sizeof (rangebuf), "all");
 	else
-		(void) printf("[%llu, %llu]\n", (u_longlong_t)record->zi_start,
+		snprintf(rangebuf, sizeof (rangebuf), "[%llu, %llu]",
+		    (u_longlong_t)record->zi_start,
 		    (u_longlong_t)record->zi_end);
 
+
+	(void) printf("%3d  %-15s  %-6llu  %-6llu  %-8s  %-3d  0x%02x  %-15s  "
+	    "%6lu  %6lu\n", id, pool, (u_longlong_t)record->zi_objset,
+	    (u_longlong_t)record->zi_object, type_to_name(record->zi_type),
+	    record->zi_level, record->zi_dvas, rangebuf,
+	    record->zi_match_count, record->zi_inject_count);
+
 	return (0);
 }
 
@@ -432,10 +465,6 @@ static int
 print_device_handler(int id, const char *pool, zinject_record_t *record,
     void *data)
 {
-	static const char *iotypestr[] = {
-	    "null", "read", "write", "free", "claim", "flush", "trim", "all",
-	};
-
 	int *count = data;
 
 	if (record->zi_guid == 0 || record->zi_func[0] != '\0')
@@ -445,11 +474,14 @@ print_device_handler(int id, const char *pool, zinject_record_t *record,
 		return (0);
 
 	if (*count == 0) {
-		(void) printf("%3s  %-15s  %-16s  %-5s  %-10s  %-9s\n",
-		    "ID", "POOL", "GUID", "TYPE", "ERROR", "FREQ");
+		(void) printf("%3s  %-15s  %-16s  %-5s  %-10s  %-9s  "
+		    "%-6s  %-6s\n",
+		    "ID", "POOL", "GUID", "TYPE", "ERROR", "FREQ",
+		    "MATCH", "INJECT");
 		(void) printf(
 		    "---  ---------------  ----------------  "
-		    "-----  ----------  ---------\n");
+		    "-----  ----------  ---------  "
+		    "------  ------\n");
 	}
 
 	*count += 1;
@@ -457,9 +489,10 @@ print_device_handler(int id, const char *pool, zinject_record_t *record,
 	double freq = record->zi_freq == 0 ? 100.0f :
 	    (((double)record->zi_freq) / ZI_PERCENTAGE_MAX) * 100.0f;
 
-	(void) printf("%3d  %-15s  %llx  %-5s  %-10s  %8.4f%%\n", id, pool,
-	    (u_longlong_t)record->zi_guid, iotypestr[record->zi_iotype],
-	    err_to_str(record->zi_error), freq);
+	(void) printf("%3d  %-15s  %llx  %-5s  %-10s  %8.4f%%  "
+	    "%6lu  %6lu\n", id, pool, (u_longlong_t)record->zi_guid,
+	    iotype_to_str(record->zi_iotype), err_to_str(record->zi_error),
+	    freq, record->zi_match_count, record->zi_inject_count);
 
 	return (0);
 }
@@ -477,18 +510,25 @@ print_delay_handler(int id, const char *pool, zinject_record_t *record,
 		return (0);
 
 	if (*count == 0) {
-		(void) printf("%3s  %-15s  %-15s  %-15s  %s\n",
-		    "ID", "POOL", "DELAY (ms)", "LANES", "GUID");
-		(void) printf("---  ---------------  ---------------  "
-		    "---------------  ----------------\n");
+		(void) printf("%3s  %-15s  %-16s  %-10s  %-5s  %-9s  "
+		    "%-6s  %-6s\n",
+		    "ID", "POOL", "GUID", "DELAY (ms)", "LANES", "FREQ",
+		    "MATCH", "INJECT");
+		(void) printf("---  ---------------  ----------------  "
+		    "----------  -----  ---------  "
+		    "------  ------\n");
 	}
 
 	*count += 1;
 
-	(void) printf("%3d  %-15s  %-15llu  %-15llu  %llx\n", id, pool,
+	double freq = record->zi_freq == 0 ? 100.0f :
+	    (((double)record->zi_freq) / ZI_PERCENTAGE_MAX) * 100.0f;
+
+	(void) printf("%3d  %-15s  %llx  %10llu  %5llu  %8.4f%%  "
+	    "%6lu  %6lu\n", id, pool, (u_longlong_t)record->zi_guid,
 	    (u_longlong_t)NSEC2MSEC(record->zi_timer),
 	    (u_longlong_t)record->zi_nlanes,
-	    (u_longlong_t)record->zi_guid);
+	    freq, record->zi_match_count, record->zi_inject_count);
 
 	return (0);
 }
@@ -852,7 +892,7 @@ main(int argc, char **argv)
 	int quiet = 0;
 	int error = 0;
 	int domount = 0;
-	int io_type = ZIO_TYPES;
+	int io_type = ZINJECT_IOTYPE_ALL;
 	int action = VDEV_STATE_UNKNOWN;
 	err_type_t type = TYPE_INVAL;
 	err_type_t label = TYPE_INVAL;
@@ -1046,19 +1086,8 @@ main(int argc, char **argv)
 			}
 			break;
 		case 'T':
-			if (strcasecmp(optarg, "read") == 0) {
-				io_type = ZIO_TYPE_READ;
-			} else if (strcasecmp(optarg, "write") == 0) {
-				io_type = ZIO_TYPE_WRITE;
-			} else if (strcasecmp(optarg, "free") == 0) {
-				io_type = ZIO_TYPE_FREE;
-			} else if (strcasecmp(optarg, "claim") == 0) {
-				io_type = ZIO_TYPE_CLAIM;
-			} else if (strcasecmp(optarg, "flush") == 0) {
-				io_type = ZIO_TYPE_FLUSH;
-			} else if (strcasecmp(optarg, "all") == 0) {
-				io_type = ZIO_TYPES;
-			} else {
+			io_type = str_to_iotype(optarg);
+			if (io_type == ZINJECT_IOTYPES) {
 				(void) fprintf(stderr, "invalid I/O type "
 				    "'%s': must be 'read', 'write', 'free', "
 				    "'claim', 'flush' or 'all'\n", optarg);
@@ -1180,7 +1209,7 @@ main(int argc, char **argv)
 		}
 
 		if (error == EILSEQ &&
-		    (record.zi_freq == 0 || io_type != ZIO_TYPE_READ)) {
+		    (record.zi_freq == 0 || io_type != ZINJECT_IOTYPE_READ)) {
 			(void) fprintf(stderr, "device corrupt errors require "
 			    "io type read and a frequency value\n");
 			libzfs_fini(g_zfs);
@@ -1195,9 +1224,9 @@ main(int argc, char **argv)
 
 		if (record.zi_nlanes) {
 			switch (io_type) {
-			case ZIO_TYPE_READ:
-			case ZIO_TYPE_WRITE:
-			case ZIO_TYPES:
+			case ZINJECT_IOTYPE_READ:
+			case ZINJECT_IOTYPE_WRITE:
+			case ZINJECT_IOTYPE_ALL:
 				break;
 			default:
 				(void) fprintf(stderr, "I/O type for a delay "

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);
@@ -8457,7 +8481,19 @@ zpool_do_scrub(int argc, char **argv)
 
 	if (is_pause && is_stop) {
 		(void) fprintf(stderr, gettext("invalid option "
-		    "combination :-s and -p are mutually exclusive\n"));
+		    "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)
@@ -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-automount.m4

@@ -5,7 +5,7 @@ dnl # solution to handling automounts.  Prior to this cifs/nfs clients
 dnl # which required automount support would abuse the follow_link()
 dnl # operation on directories for this purpose.
 dnl #
-AC_DEFUN([ZFS_AC_KERNEL_SRC_AUTOMOUNT], [
+AC_DEFUN([ZFS_AC_KERNEL_SRC_D_AUTOMOUNT], [
 	ZFS_LINUX_TEST_SRC([dentry_operations_d_automount], [
 		#include <linux/dcache.h>
 		static struct vfsmount *d_automount(struct path *p) { return NULL; }
@@ -15,7 +15,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_AUTOMOUNT], [
 	])
 ])
 
-AC_DEFUN([ZFS_AC_KERNEL_AUTOMOUNT], [
+AC_DEFUN([ZFS_AC_KERNEL_D_AUTOMOUNT], [
 	AC_MSG_CHECKING([whether dops->d_automount() exists])
 	ZFS_LINUX_TEST_RESULT([dentry_operations_d_automount], [
 		AC_MSG_RESULT(yes)
@@ -23,3 +23,40 @@ AC_DEFUN([ZFS_AC_KERNEL_AUTOMOUNT], [
 		ZFS_LINUX_TEST_ERROR([dops->d_automount()])
 	])
 ])
+
+dnl #
+dnl # 6.14 API change
+dnl # dops->d_revalidate now has four args.
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_SRC_D_REVALIDATE_4ARGS], [
+	ZFS_LINUX_TEST_SRC([dentry_operations_d_revalidate_4args], [
+		#include <linux/dcache.h>
+		static int d_revalidate(struct inode *dir,
+		    const struct qstr *name, struct dentry *dentry,
+		    unsigned int fl) { return 0; }
+		struct dentry_operations dops __attribute__ ((unused)) = {
+			.d_revalidate = d_revalidate,
+		};
+	])
+])
+
+AC_DEFUN([ZFS_AC_KERNEL_D_REVALIDATE_4ARGS], [
+	AC_MSG_CHECKING([whether dops->d_revalidate() takes 4 args])
+	ZFS_LINUX_TEST_RESULT([dentry_operations_d_revalidate_4args], [
+		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_D_REVALIDATE_4ARGS, 1,
+		    [dops->d_revalidate() takes 4 args])
+	],[
+		AC_MSG_RESULT(no)
+	])
+])
+
+AC_DEFUN([ZFS_AC_KERNEL_SRC_AUTOMOUNT], [
+	ZFS_AC_KERNEL_SRC_D_AUTOMOUNT
+	ZFS_AC_KERNEL_SRC_D_REVALIDATE_4ARGS
+])
+
+AC_DEFUN([ZFS_AC_KERNEL_AUTOMOUNT], [
+	ZFS_AC_KERNEL_D_AUTOMOUNT
+	ZFS_AC_KERNEL_D_REVALIDATE_4ARGS
+])

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-pin-user-pages.m4

@@ -0,0 +1,33 @@
+dnl #
+dnl # Check for pin_user_pages_unlocked().
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_SRC_PIN_USER_PAGES], [
+	ZFS_LINUX_TEST_SRC([pin_user_pages_unlocked], [
+		#include <linux/mm.h>
+	],[
+		unsigned long start = 0;
+		unsigned long nr_pages = 1;
+		struct page **pages = NULL;
+		unsigned int gup_flags = 0;
+		long ret __attribute__ ((unused));
+
+		ret = pin_user_pages_unlocked(start, nr_pages, pages,
+		    gup_flags);
+	])
+])
+
+AC_DEFUN([ZFS_AC_KERNEL_PIN_USER_PAGES], [
+
+	dnl #
+	dnl # Kernal 5.8 introduced the pin_user_pages* interfaces which should
+	dnl # be used for Direct I/O requests.
+	dnl #
+	AC_MSG_CHECKING([whether pin_user_pages_unlocked() is available])
+	ZFS_LINUX_TEST_RESULT([pin_user_pages_unlocked], [
+		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_PIN_USER_PAGES_UNLOCKED, 1,
+		    [pin_user_pages_unlocked() is available])
+	],[
+		AC_MSG_RESULT(no)
+	])
+])

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-direct_IO.m4

@@ -1,57 +0,0 @@
-dnl #
-dnl # Check for Direct I/O interfaces.
-dnl #
-AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_DIRECT_IO], [
-	ZFS_LINUX_TEST_SRC([direct_io_iter], [
-		#include <linux/fs.h>
-
-		static ssize_t test_direct_IO(struct kiocb *kiocb,
-		    struct iov_iter *iter) { return 0; }
-
-		static const struct address_space_operations
-		    aops __attribute__ ((unused)) = {
-			.direct_IO = test_direct_IO,
-		};
-	],[])
-
-	ZFS_LINUX_TEST_SRC([direct_io_iter_offset], [
-		#include <linux/fs.h>
-
-		static ssize_t test_direct_IO(struct kiocb *kiocb,
-		    struct iov_iter *iter, loff_t offset) { return 0; }
-
-		static const struct address_space_operations
-		    aops __attribute__ ((unused)) = {
-			.direct_IO = test_direct_IO,
-		};
-	],[])
-])
-
-AC_DEFUN([ZFS_AC_KERNEL_VFS_DIRECT_IO], [
-	dnl #
-	dnl # Linux 4.6.x API change
-	dnl #
-	AC_MSG_CHECKING([whether aops->direct_IO() uses iov_iter])
-	ZFS_LINUX_TEST_RESULT([direct_io_iter], [
-		AC_MSG_RESULT([yes])
-		AC_DEFINE(HAVE_VFS_DIRECT_IO_ITER, 1,
-		    [aops->direct_IO() uses iov_iter without rw])
-	],[
-		AC_MSG_RESULT([no])
-
-		dnl #
-		dnl # Linux 4.1.x API change
-		dnl #
-		AC_MSG_CHECKING(
-		    [whether aops->direct_IO() uses offset])
-		ZFS_LINUX_TEST_RESULT([direct_io_iter_offset], [
-			AC_MSG_RESULT([yes])
-			AC_DEFINE(HAVE_VFS_DIRECT_IO_ITER_OFFSET, 1,
-			    [aops->direct_IO() uses iov_iter with offset])
-
-		],[
-			AC_MSG_RESULT([no])
-			ZFS_LINUX_TEST_ERROR([Direct I/O])
-		])
-	])
-])

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-iov_iter.m4

@@ -13,9 +13,17 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_IOV_ITER], [
 		error = fault_in_iov_iter_readable(&iter, size);
 	])
 
+	ZFS_LINUX_TEST_SRC([iov_iter_type], [
+		#include <linux/fs.h>
+		#include <linux/uio.h>
+	],[
+		struct iov_iter iter = { 0 };
+		__attribute__((unused)) enum iter_type i = iov_iter_type(&iter);
+	])
+
 	ZFS_LINUX_TEST_SRC([iov_iter_get_pages2], [
 		#include <linux/uio.h>
-	], [
+	],[
 		struct iov_iter iter = { 0 };
 		struct page **pages = NULL;
 		size_t maxsize = 4096;
@@ -27,26 +35,13 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_IOV_ITER], [
 		    &start);
 	])
 
-	ZFS_LINUX_TEST_SRC([iov_iter_get_pages], [
-		#include <linux/uio.h>
-	], [
-		struct iov_iter iter = { 0 };
-		struct page **pages = NULL;
-		size_t maxsize = 4096;
-		unsigned maxpages = 1;
-		size_t start;
-		size_t ret __attribute__ ((unused));
-
-		ret = iov_iter_get_pages(&iter, pages, maxsize, maxpages,
-		    &start);
-	])
-
-	ZFS_LINUX_TEST_SRC([iov_iter_type], [
-		#include <linux/fs.h>
+	ZFS_LINUX_TEST_SRC([iter_is_ubuf], [
 		#include <linux/uio.h>
 	],[
 		struct iov_iter iter = { 0 };
-		__attribute__((unused)) enum iter_type i = iov_iter_type(&iter);
+		bool ret __attribute__((unused));
+
+		ret = iter_is_ubuf(&iter);
 	])
 
 	ZFS_LINUX_TEST_SRC([iter_iov], [
@@ -59,7 +54,6 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_IOV_ITER], [
 ])
 
 AC_DEFUN([ZFS_AC_KERNEL_VFS_IOV_ITER], [
-	enable_vfs_iov_iter="yes"
 
 	AC_MSG_CHECKING([whether fault_in_iov_iter_readable() is available])
 	ZFS_LINUX_TEST_RESULT([fault_in_iov_iter_readable], [
@@ -70,27 +64,6 @@ AC_DEFUN([ZFS_AC_KERNEL_VFS_IOV_ITER], [
 		AC_MSG_RESULT(no)
 	])
 
-	dnl #
-	dnl # Kernel 6.0 changed iov_iter_get_pages() to iov_iter_page_pages2().
-	dnl #
-	AC_MSG_CHECKING([whether iov_iter_get_pages2() is available])
-	ZFS_LINUX_TEST_RESULT([iov_iter_get_pages2], [
-		AC_MSG_RESULT(yes)
-		AC_DEFINE(HAVE_IOV_ITER_GET_PAGES2, 1,
-		    [iov_iter_get_pages2() is available])
-	], [
-		AC_MSG_RESULT(no)
-		AC_MSG_CHECKING([whether iov_iter_get_pages() is available])
-			ZFS_LINUX_TEST_RESULT([iov_iter_get_pages], [
-			AC_MSG_RESULT(yes)
-			AC_DEFINE(HAVE_IOV_ITER_GET_PAGES, 1,
-			    [iov_iter_get_pages() is available])
-		], [
-			AC_MSG_RESULT(no)
-			enable_vfs_iov_iter="no"
-		])
-	])
-
 	dnl #
 	dnl # This checks for iov_iter_type() in linux/uio.h. It is not
 	dnl # required, however, and the module will compiled without it
@@ -105,15 +78,29 @@ AC_DEFUN([ZFS_AC_KERNEL_VFS_IOV_ITER], [
 		AC_MSG_RESULT(no)
 	])
 
+
 	dnl #
-	dnl # As of the 4.9 kernel support is provided for iovecs, kvecs,
-	dnl # bvecs and pipes in the iov_iter structure.  As long as the
-	dnl # other support interfaces are all available the iov_iter can
-	dnl # be correctly used in the uio structure.
+	dnl # Kernel 6.0 changed iov_iter_get_pages() to iov_iter_get_pages2().
 	dnl #
-	AS_IF([test "x$enable_vfs_iov_iter" = "xyes"], [
-		AC_DEFINE(HAVE_VFS_IOV_ITER, 1,
-		    [All required iov_iter interfaces are available])
+	AC_MSG_CHECKING([whether iov_iter_get_pages2() is available])
+	ZFS_LINUX_TEST_RESULT([iov_iter_get_pages2], [
+		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_IOV_ITER_GET_PAGES2, 1,
+		    [iov_iter_get_pages2() is available])
+	],[
+		AC_MSG_RESULT(no)
+	])
+
+	dnl #
+	dnl # Kernel 6.0 introduced the ITER_UBUF iov_iter type. iter_is_ubuf()
+	dnl # was also added to determine if the iov_iter is an ITER_UBUF.
+	dnl #
+	AC_MSG_CHECKING([whether iter_is_ubuf() is available])
+	ZFS_LINUX_TEST_RESULT([iter_is_ubuf], [
+		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_ITER_IS_UBUF, 1, [iter_is_ubuf() is available])
+	],[
+		AC_MSG_RESULT(no)
 	])
 
 	dnl #

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-xattr-handler.m4

@@ -54,7 +54,6 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_GET_DENTRY_INODE_FLAGS], [
 ])
 
 AC_DEFUN([ZFS_AC_KERNEL_XATTR_HANDLER_GET_DENTRY_INODE_FLAGS], [
-	AC_MSG_RESULT(no)
 	AC_MSG_CHECKING(
 	    [whether xattr_handler->get() wants dentry and inode and flags])
 	ZFS_LINUX_TEST_RESULT([xattr_handler_get_dentry_inode_flags], [

config/kernel.m4

@@ -77,10 +77,8 @@ 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
 	ZFS_AC_KERNEL_SRC_VFS_SET_PAGE_DIRTY_NOBUFFERS
 	ZFS_AC_KERNEL_SRC_VFS_IOV_ITER
@@ -129,6 +127,7 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_SRC], [
 	ZFS_AC_KERNEL_SRC_MM_PAGE_SIZE
 	ZFS_AC_KERNEL_SRC_MM_PAGE_MAPPING
 	ZFS_AC_KERNEL_SRC_FILE
+	ZFS_AC_KERNEL_SRC_PIN_USER_PAGES
 	case "$host_cpu" in
 		powerpc*)
 			ZFS_AC_KERNEL_SRC_CPU_HAS_FEATURE
@@ -189,10 +188,8 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_RESULT], [
 	ZFS_AC_KERNEL_SGET
 	ZFS_AC_KERNEL_VFS_FILEMAP_DIRTY_FOLIO
 	ZFS_AC_KERNEL_VFS_READ_FOLIO
-	ZFS_AC_KERNEL_VFS_RELEASE_FOLIO
-	ZFS_AC_KERNEL_VFS_INVALIDATE_FOLIO
+	ZFS_AC_KERNEL_VFS_MIGRATE_FOLIO
 	ZFS_AC_KERNEL_VFS_FSYNC_2ARGS
-	ZFS_AC_KERNEL_VFS_DIRECT_IO
 	ZFS_AC_KERNEL_VFS_READPAGES
 	ZFS_AC_KERNEL_VFS_SET_PAGE_DIRTY_NOBUFFERS
 	ZFS_AC_KERNEL_VFS_IOV_ITER
@@ -242,6 +239,7 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_RESULT], [
 	ZFS_AC_KERNEL_MM_PAGE_MAPPING
 	ZFS_AC_KERNEL_1ARG_ASSIGN_STR
 	ZFS_AC_KERNEL_FILE
+	ZFS_AC_KERNEL_PIN_USER_PAGES
 	case "$host_cpu" in
 		powerpc*)
 			ZFS_AC_KERNEL_CPU_HAS_FEATURE
@@ -683,11 +681,16 @@ AC_DEFUN([ZFS_LINUX_COMPILE], [
 		building kernel modules])
 	AC_ARG_VAR([KERNEL_LLVM], [Binary option to
 		build kernel modules with LLVM/CLANG toolchain])
+	AC_ARG_VAR([KERNEL_CROSS_COMPILE], [Cross compile prefix
+		for kernel module builds])
+	AC_ARG_VAR([KERNEL_ARCH], [Architecture to build kernel modules for])
 	AC_TRY_COMMAND([
 	    KBUILD_MODPOST_NOFINAL="$5" KBUILD_MODPOST_WARN="$6"
 	    make modules -k -j$TEST_JOBS ${KERNEL_CC:+CC=$KERNEL_CC}
 	    ${KERNEL_LD:+LD=$KERNEL_LD} ${KERNEL_LLVM:+LLVM=$KERNEL_LLVM}
 	    CONFIG_MODULES=y CFLAGS_MODULE=-DCONFIG_MODULES
+	    ${KERNEL_CROSS_COMPILE:+CROSS_COMPILE=$KERNEL_CROSS_COMPILE}
+	    ${KERNEL_ARCH:+ARCH=$KERNEL_ARCH}
 	    -C $LINUX_OBJ $ARCH_UM M=$PWD/$1 >$1/build.log 2>&1])
 	AS_IF([AC_TRY_COMMAND([$2])], [$3], [$4])
 ])

config/zfs-build.m4

@@ -393,6 +393,8 @@ AC_DEFUN([ZFS_AC_RPM], [
 		RPM_DEFINE_KMOD=${RPM_DEFINE_KMOD}' --define "kernel_cc KERNEL_CC=$(KERNEL_CC)"'
 		RPM_DEFINE_KMOD=${RPM_DEFINE_KMOD}' --define "kernel_ld KERNEL_LD=$(KERNEL_LD)"'
 		RPM_DEFINE_KMOD=${RPM_DEFINE_KMOD}' --define "kernel_llvm KERNEL_LLVM=$(KERNEL_LLVM)"'
+		RPM_DEFINE_KMOD=${RPM_DEFINE_KMOD}' --define "kernel_cross_compile KERNEL_CROSS_COMPILE=$(KERNEL_CROSS_COMPILE)"'
+		RPM_DEFINE_KMOD=${RPM_DEFINE_KMOD}' --define "kernel_arch KERNEL_ARCH=$(KERNEL_ARCH)"'
 	])
 
 	RPM_DEFINE_DKMS=''
@@ -627,7 +629,7 @@ AC_DEFUN([ZFS_AC_DEFAULT_PACKAGE], [
 
 	AC_MSG_CHECKING([default bash completion directory])
 	case "$VENDOR" in
-		alpine|artix|debian|gentoo|ubuntu)
+		alpine|arch|artix|debian|gentoo|ubuntu)
 			bashcompletiondir=/usr/share/bash-completion/completions
 			;;
 		freebsd)

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=""

contrib/pam_zfs_key/pam_zfs_key.c

@@ -63,6 +63,7 @@ pam_syslog(pam_handle_t *pamh, int loglevel, const char *fmt, ...)
 #include <sys/file.h>
 #include <sys/wait.h>
 #include <pwd.h>
+#include <lib/libzfs/libzfs_impl.h>
 
 #include <sys/mman.h>
 
@@ -370,67 +371,6 @@ change_key(pam_handle_t *pamh, const char *ds_name,
 	return (0);
 }
 
-static int
-decrypt_mount(pam_handle_t *pamh, const char *ds_name,
-    const char *passphrase, boolean_t noop)
-{
-	zfs_handle_t *ds = zfs_open(g_zfs, ds_name, ZFS_TYPE_FILESYSTEM);
-	if (ds == NULL) {
-		pam_syslog(pamh, LOG_ERR, "dataset %s not found", ds_name);
-		return (-1);
-	}
-	pw_password_t *key = prepare_passphrase(pamh, ds, passphrase, NULL);
-	if (key == NULL) {
-		zfs_close(ds);
-		return (-1);
-	}
-	int ret = lzc_load_key(ds_name, noop, (uint8_t *)key->value,
-	    WRAPPING_KEY_LEN);
-	pw_free(key);
-	if (ret && ret != EEXIST) {
-		pam_syslog(pamh, LOG_ERR, "load_key failed: %d", ret);
-		zfs_close(ds);
-		return (-1);
-	}
-	if (noop) {
-		goto out;
-	}
-	ret = zfs_mount(ds, NULL, 0);
-	if (ret) {
-		pam_syslog(pamh, LOG_ERR, "mount failed: %d", ret);
-		zfs_close(ds);
-		return (-1);
-	}
-out:
-	zfs_close(ds);
-	return (0);
-}
-
-static int
-unmount_unload(pam_handle_t *pamh, const char *ds_name, boolean_t force)
-{
-	zfs_handle_t *ds = zfs_open(g_zfs, ds_name, ZFS_TYPE_FILESYSTEM);
-	if (ds == NULL) {
-		pam_syslog(pamh, LOG_ERR, "dataset %s not found", ds_name);
-		return (-1);
-	}
-	int ret = zfs_unmount(ds, NULL, force ? MS_FORCE : 0);
-	if (ret) {
-		pam_syslog(pamh, LOG_ERR, "zfs_unmount failed with: %d", ret);
-		zfs_close(ds);
-		return (-1);
-	}
-
-	ret = lzc_unload_key(ds_name);
-	if (ret) {
-		pam_syslog(pamh, LOG_ERR, "unload_key failed with: %d", ret);
-		zfs_close(ds);
-		return (-1);
-	}
-	zfs_close(ds);
-	return (0);
-}
-
 typedef struct {
 	char *homes_prefix;
 	char *runstatedir;
@@ -443,6 +383,7 @@ typedef struct {
 	boolean_t unmount_and_unload;
 	boolean_t force_unmount;
 	boolean_t recursive_homes;
+	boolean_t mount_recursively;
 } zfs_key_config_t;
 
 static int
@@ -481,6 +422,7 @@ zfs_key_config_load(pam_handle_t *pamh, zfs_key_config_t *config,
 	config->unmount_and_unload = B_TRUE;
 	config->force_unmount = B_FALSE;
 	config->recursive_homes = B_FALSE;
+	config->mount_recursively = B_FALSE;
 	config->dsname = NULL;
 	config->homedir = NULL;
 	for (int c = 0; c < argc; c++) {
@@ -500,6 +442,8 @@ zfs_key_config_load(pam_handle_t *pamh, zfs_key_config_t *config,
 			config->force_unmount = B_TRUE;
 		} else if (strcmp(argv[c], "recursive_homes") == 0) {
 			config->recursive_homes = B_TRUE;
+		} else if (strcmp(argv[c], "mount_recursively") == 0) {
+			config->mount_recursively = B_TRUE;
 		} else if (strcmp(argv[c], "prop_mountpoint") == 0) {
 			if (config->homedir == NULL)
 				config->homedir = strdup(entry->pw_dir);
@@ -508,6 +452,217 @@ zfs_key_config_load(pam_handle_t *pamh, zfs_key_config_t *config,
 	return (PAM_SUCCESS);
 }
 
+typedef struct {
+	pam_handle_t *pamh;
+	zfs_key_config_t *target;
+} mount_umount_dataset_data_t;
+
+static int
+mount_dataset(zfs_handle_t *zhp, void *data)
+{
+	mount_umount_dataset_data_t *mount_umount_dataset_data = data;
+
+	zfs_key_config_t *target = mount_umount_dataset_data->target;
+	pam_handle_t *pamh = mount_umount_dataset_data->pamh;
+
+	/* Refresh properties to get the latest key status */
+	zfs_refresh_properties(zhp);
+
+	int ret = 0;
+
+	/* Check if dataset type is filesystem */
+	if (zhp->zfs_type != ZFS_TYPE_FILESYSTEM) {
+		pam_syslog(pamh, LOG_DEBUG,
+		    "dataset is not filesystem: %s, skipping.",
+		    zfs_get_name(zhp));
+		return (0);
+	}
+
+	/* Check if encryption key is available */
+	if (zfs_prop_get_int(zhp, ZFS_PROP_KEYSTATUS) ==
+	    ZFS_KEYSTATUS_UNAVAILABLE) {
+		pam_syslog(pamh, LOG_WARNING,
+		    "key unavailable for: %s, skipping",
+		    zfs_get_name(zhp));
+		return (0);
+	}
+
+	/* Check if prop canmount is on */
+	if (zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT) != ZFS_CANMOUNT_ON) {
+		pam_syslog(pamh, LOG_INFO,
+		    "canmount is not on for: %s, skipping",
+		    zfs_get_name(zhp));
+		return (0);
+	}
+
+	/* Get mountpoint prop for check */
+	char mountpoint[ZFS_MAXPROPLEN];
+	if ((ret = zfs_prop_get(zhp, ZFS_PROP_MOUNTPOINT, mountpoint,
+	    sizeof (mountpoint), NULL, NULL, 0, 1)) != 0) {
+		pam_syslog(pamh, LOG_ERR,
+		    "failed to get mountpoint prop: %d", ret);
+		return (-1);
+	}
+
+	/* Check if mountpoint isn't none or legacy */
+	if (strcmp(mountpoint, ZFS_MOUNTPOINT_NONE) == 0 ||
+	    strcmp(mountpoint, ZFS_MOUNTPOINT_LEGACY) == 0) {
+		pam_syslog(pamh, LOG_INFO,
+		    "mountpoint is none or legacy for: %s, skipping",
+		    zfs_get_name(zhp));
+		return (0);
+	}
+
+	/* Don't mount the dataset if already mounted */
+	if (zfs_is_mounted(zhp, NULL)) {
+		pam_syslog(pamh, LOG_INFO, "already mounted: %s",
+		    zfs_get_name(zhp));
+		return (0);
+	}
+
+	/* Mount the dataset */
+	ret = zfs_mount(zhp, NULL, 0);
+	if (ret) {
+		pam_syslog(pamh, LOG_ERR,
+		    "zfs_mount failed for %s with: %d", zfs_get_name(zhp),
+		    ret);
+		return (ret);
+	}
+
+	/* Recursively mount children if the recursive flag is set */
+	if (target->mount_recursively) {
+		ret = zfs_iter_filesystems_v2(zhp, 0, mount_dataset, data);
+		if (ret != 0) {
+			pam_syslog(pamh, LOG_ERR,
+			    "child iteration failed: %d", ret);
+			return (-1);
+		}
+	}
+
+	return (ret);
+}
+
+static int
+umount_dataset(zfs_handle_t *zhp, void *data)
+{
+	mount_umount_dataset_data_t *mount_umount_dataset_data = data;
+
+	zfs_key_config_t *target = mount_umount_dataset_data->target;
+	pam_handle_t *pamh = mount_umount_dataset_data->pamh;
+
+	int ret = 0;
+	/* Recursively umount children if the recursive flag is set */
+	if (target->mount_recursively) {
+		ret = zfs_iter_filesystems_v2(zhp, 0, umount_dataset, data);
+		if (ret != 0) {
+			pam_syslog(pamh, LOG_ERR,
+			    "child iteration failed: %d", ret);
+			return (-1);
+		}
+	}
+
+	/* Check if dataset type is filesystem */
+	if (zhp->zfs_type != ZFS_TYPE_FILESYSTEM) {
+		pam_syslog(pamh, LOG_DEBUG,
+		    "dataset is not filesystem: %s, skipping",
+		    zfs_get_name(zhp));
+		return (0);
+	}
+
+	/* Don't umount the dataset if already unmounted */
+	if (zfs_is_mounted(zhp, NULL) == 0) {
+		pam_syslog(pamh, LOG_INFO, "already unmounted: %s",
+		    zfs_get_name(zhp));
+		return (0);
+	}
+
+	/* Unmount the dataset */
+	ret = zfs_unmount(zhp, NULL, target->force_unmount ? MS_FORCE : 0);
+	if (ret) {
+		pam_syslog(pamh, LOG_ERR,
+		    "zfs_unmount failed for %s with: %d", zfs_get_name(zhp),
+		    ret);
+		return (ret);
+	}
+
+	return (ret);
+}
+
+static int
+decrypt_mount(pam_handle_t *pamh, zfs_key_config_t *config, const char *ds_name,
+	const char *passphrase, boolean_t noop)
+{
+	zfs_handle_t *ds = zfs_open(g_zfs, ds_name, ZFS_TYPE_FILESYSTEM);
+	if (ds == NULL) {
+		pam_syslog(pamh, LOG_ERR, "dataset %s not found", ds_name);
+		return (-1);
+	}
+	pw_password_t *key = prepare_passphrase(pamh, ds, passphrase, NULL);
+	if (key == NULL) {
+		zfs_close(ds);
+		return (-1);
+	}
+	int ret = lzc_load_key(ds_name, noop, (uint8_t *)key->value,
+	    WRAPPING_KEY_LEN);
+	pw_free(key);
+	if (ret && ret != EEXIST) {
+		pam_syslog(pamh, LOG_ERR, "load_key failed: %d", ret);
+		zfs_close(ds);
+		return (-1);
+	}
+
+	if (noop) {
+		zfs_close(ds);
+		return (0);
+	}
+
+	mount_umount_dataset_data_t data;
+	data.pamh = pamh;
+	data.target = config;
+
+	ret = mount_dataset(ds, &data);
+	if (ret != 0) {
+		pam_syslog(pamh, LOG_ERR, "mount failed: %d", ret);
+		zfs_close(ds);
+		return (-1);
+	}
+
+	zfs_close(ds);
+	return (0);
+}
+
+static int
+unmount_unload(pam_handle_t *pamh, const char *ds_name,
+    zfs_key_config_t *target)
+{
+	zfs_handle_t *ds = zfs_open(g_zfs, ds_name, ZFS_TYPE_FILESYSTEM);
+	if (ds == NULL) {
+		pam_syslog(pamh, LOG_ERR, "dataset %s not found", ds_name);
+		return (-1);
+	}
+
+	mount_umount_dataset_data_t data;
+	data.pamh = pamh;
+	data.target = target;
+
+	int ret = umount_dataset(ds, &data);
+	if (ret) {
+		pam_syslog(pamh, LOG_ERR,
+		    "unmount_dataset failed with: %d", ret);
+		zfs_close(ds);
+		return (-1);
+	}
+
+	ret = lzc_unload_key(ds_name);
+	if (ret) {
+		pam_syslog(pamh, LOG_ERR, "unload_key failed with: %d", ret);
+		zfs_close(ds);
+		return (-1);
+	}
+	zfs_close(ds);
+	return (0);
+}
+
 static void
 zfs_key_config_free(zfs_key_config_t *config)
 {
@@ -548,7 +703,7 @@ find_dsname_by_prop_value(zfs_handle_t *zhp, void *data)
 }
 
 static char *
-zfs_key_config_get_dataset(zfs_key_config_t *config)
+zfs_key_config_get_dataset(pam_handle_t *pamh, zfs_key_config_t *config)
 {
 	if (config->homedir != NULL &&
 	    config->homes_prefix != NULL) {
@@ -559,7 +714,7 @@ zfs_key_config_get_dataset(zfs_key_config_t *config)
 			zfs_handle_t *zhp = zfs_open(g_zfs,
 			    config->homes_prefix, ZFS_TYPE_FILESYSTEM);
 			if (zhp == NULL) {
-				pam_syslog(NULL, LOG_ERR,
+				pam_syslog(pamh, LOG_ERR,
 				    "dataset %s not found",
 				    config->homes_prefix);
 				return (NULL);
@@ -697,13 +852,13 @@ pam_sm_authenticate(pam_handle_t *pamh, int flags,
 		zfs_key_config_free(&config);
 		return (PAM_SERVICE_ERR);
 	}
-	char *dataset = zfs_key_config_get_dataset(&config);
+	char *dataset = zfs_key_config_get_dataset(pamh, &config);
 	if (!dataset) {
 		pam_zfs_free();
 		zfs_key_config_free(&config);
 		return (PAM_SERVICE_ERR);
 	}
-	if (decrypt_mount(pamh, dataset, token->value, B_TRUE) == -1) {
+	if (decrypt_mount(pamh, &config, dataset, token->value, B_TRUE) == -1) {
 		free(dataset);
 		pam_zfs_free();
 		zfs_key_config_free(&config);
@@ -749,7 +904,7 @@ pam_sm_chauthtok(pam_handle_t *pamh, int flags,
 			zfs_key_config_free(&config);
 			return (PAM_SERVICE_ERR);
 		}
-		char *dataset = zfs_key_config_get_dataset(&config);
+		char *dataset = zfs_key_config_get_dataset(pamh, &config);
 		if (!dataset) {
 			pam_zfs_free();
 			zfs_key_config_free(&config);
@@ -763,7 +918,7 @@ pam_sm_chauthtok(pam_handle_t *pamh, int flags,
 			zfs_key_config_free(&config);
 			return (PAM_SERVICE_ERR);
 		}
-		if (decrypt_mount(pamh, dataset,
+		if (decrypt_mount(pamh, &config, dataset,
 		    old_token->value, B_TRUE) == -1) {
 			pam_syslog(pamh, LOG_ERR,
 			    "old token mismatch");
@@ -784,7 +939,7 @@ pam_sm_chauthtok(pam_handle_t *pamh, int flags,
 			pw_clear(pamh, OLD_PASSWORD_VAR_NAME);
 			return (PAM_SERVICE_ERR);
 		}
-		char *dataset = zfs_key_config_get_dataset(&config);
+		char *dataset = zfs_key_config_get_dataset(pamh, &config);
 		if (!dataset) {
 			pam_zfs_free();
 			zfs_key_config_free(&config);
@@ -793,7 +948,7 @@ pam_sm_chauthtok(pam_handle_t *pamh, int flags,
 			return (PAM_SERVICE_ERR);
 		}
 		int was_loaded = is_key_loaded(pamh, dataset);
-		if (!was_loaded && decrypt_mount(pamh, dataset,
+		if (!was_loaded && decrypt_mount(pamh, &config, dataset,
 		    old_token->value, B_FALSE) == -1) {
 			free(dataset);
 			pam_zfs_free();
@@ -804,7 +959,7 @@ pam_sm_chauthtok(pam_handle_t *pamh, int flags,
 		}
 		int changed = change_key(pamh, dataset, token->value);
 		if (!was_loaded) {
-			unmount_unload(pamh, dataset, config.force_unmount);
+			unmount_unload(pamh, dataset, &config);
 		}
 		free(dataset);
 		pam_zfs_free();
@@ -856,13 +1011,14 @@ pam_sm_open_session(pam_handle_t *pamh, int flags,
 		zfs_key_config_free(&config);
 		return (PAM_SERVICE_ERR);
 	}
-	char *dataset = zfs_key_config_get_dataset(&config);
+	char *dataset = zfs_key_config_get_dataset(pamh, &config);
 	if (!dataset) {
 		pam_zfs_free();
 		zfs_key_config_free(&config);
 		return (PAM_SERVICE_ERR);
 	}
-	if (decrypt_mount(pamh, dataset, token->value, B_FALSE) == -1) {
+	if (decrypt_mount(pamh, &config, dataset,
+	    token->value, B_FALSE) == -1) {
 		free(dataset);
 		pam_zfs_free();
 		zfs_key_config_free(&config);
@@ -910,13 +1066,13 @@ pam_sm_close_session(pam_handle_t *pamh, int flags,
 			zfs_key_config_free(&config);
 			return (PAM_SERVICE_ERR);
 		}
-		char *dataset = zfs_key_config_get_dataset(&config);
+		char *dataset = zfs_key_config_get_dataset(pamh, &config);
 		if (!dataset) {
 			pam_zfs_free();
 			zfs_key_config_free(&config);
 			return (PAM_SESSION_ERR);
 		}
-		if (unmount_unload(pamh, dataset, config.force_unmount) == -1) {
+		if (unmount_unload(pamh, dataset, &config) == -1) {
 			free(dataset);
 			pam_zfs_free();
 			zfs_key_config_free(&config);

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

@@ -8,6 +8,13 @@ After=systemd-remount-fs.service
 Before=local-fs.target
 ConditionPathIsDirectory=/sys/module/zfs
 
+# This merely tells the service manager
+# that unmounting everything undoes the
+# effect of this service. No extra logic
+# is ran as a result of these settings.
+Conflicts=umount.target
+Before=umount.target
+
 [Service]
 Type=oneshot
 RemainAfterExit=yes

include/Makefile.am

@@ -150,7 +150,6 @@ COMMON_H = \
 	sys/zio_compress.h \
 	sys/zio_crypt.h \
 	sys/zio_impl.h \
-	sys/zio_priority.h \
 	sys/zrlock.h \
 	sys/zthr.h \
 	\

include/libzutil.h

@@ -27,7 +27,7 @@
 #define	_LIBZUTIL_H extern __attribute__((visibility("default")))
 
 #include <string.h>
-#include <locale.h>
+#include <pthread.h>
 #include <sys/nvpair.h>
 #include <sys/fs/zfs.h>
 
@@ -276,7 +276,14 @@ _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) {
-	return (strerror_l(errnum, uselocale(0)));
+	static __thread char errbuf[512];
+	static pthread_mutex_t zfs_strerror_lock = PTHREAD_MUTEX_INITIALIZER;
+
+	(void) pthread_mutex_lock(&zfs_strerror_lock);
+	(void) strlcpy(errbuf, strerror(errnum), sizeof (errbuf));
+	(void) pthread_mutex_unlock(&zfs_strerror_lock);
+
+	return (errbuf);
 }
 
 #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/mod_os.h

@@ -94,6 +94,9 @@
 #define	param_set_max_auto_ashift_args(var) \
     CTLTYPE_UINT, NULL, 0, param_set_max_auto_ashift, "IU"
 
+#define	param_set_raidz_impl_args(var) \
+    CTLTYPE_STRING, NULL, 0, param_set_raidz_impl, "A"
+
 #define	spa_taskq_read_param_set_args(var) \
     CTLTYPE_STRING, NULL, 0, spa_taskq_read_param, "A"
 

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/spl/sys/uio.h

@@ -40,7 +40,7 @@
  */
 #define	UIO_DIRECT	0x0001 /* Direct I/O request */
 
-#if defined(HAVE_VFS_IOV_ITER) && defined(HAVE_FAULT_IN_IOV_ITER_READABLE)
+#if defined(HAVE_FAULT_IN_IOV_ITER_READABLE)
 #define	iov_iter_fault_in_readable(a, b)	fault_in_iov_iter_readable(a, b)
 #endif
 
@@ -52,12 +52,9 @@ typedef enum zfs_uio_rw {
 } zfs_uio_rw_t;
 
 typedef enum zfs_uio_seg {
-	UIO_USERSPACE =		0,
-	UIO_SYSSPACE =		1,
-	UIO_BVEC =		2,
-#if defined(HAVE_VFS_IOV_ITER)
-	UIO_ITER =		3,
-#endif
+	UIO_SYSSPACE =		0,
+	UIO_BVEC =		1,
+	UIO_ITER =		2,
 } zfs_uio_seg_t;
 
 /*
@@ -66,15 +63,14 @@ typedef enum zfs_uio_seg {
 typedef struct {
 	struct page	**pages;	/* Mapped pages */
 	long 		npages;		/* Number of mapped pages */
+	boolean_t	pinned;		/* Whether FOLL_PIN was used */
 } zfs_uio_dio_t;
 
 typedef struct zfs_uio {
 	union {
 		const struct iovec	*uio_iov;
 		const struct bio_vec	*uio_bvec;
-#if defined(HAVE_VFS_IOV_ITER)
 		struct iov_iter		*uio_iter;
-#endif
 	};
 	int		uio_iovcnt;	/* Number of iovecs */
 	offset_t	uio_soffset;	/* Starting logical offset */
@@ -129,7 +125,7 @@ zfs_uio_iovec_init(zfs_uio_t *uio, const struct iovec *iov,
     unsigned long nr_segs, offset_t offset, zfs_uio_seg_t seg, ssize_t resid,
     size_t skip)
 {
-	ASSERT(seg == UIO_USERSPACE || seg == UIO_SYSSPACE);
+	ASSERT(seg == UIO_SYSSPACE);
 
 	uio->uio_iov = iov;
 	uio->uio_iovcnt = nr_segs;
@@ -175,7 +171,6 @@ zfs_uio_bvec_init(zfs_uio_t *uio, struct bio *bio, struct request *rq)
 	memset(&uio->uio_dio, 0, sizeof (zfs_uio_dio_t));
 }
 
-#if defined(HAVE_VFS_IOV_ITER)
 static inline void
 zfs_uio_iov_iter_init(zfs_uio_t *uio, struct iov_iter *iter, offset_t offset,
     ssize_t resid, size_t skip)
@@ -192,7 +187,6 @@ zfs_uio_iov_iter_init(zfs_uio_t *uio, struct iov_iter *iter, offset_t offset,
 	uio->uio_soffset = uio->uio_loffset;
 	memset(&uio->uio_dio, 0, sizeof (zfs_uio_dio_t));
 }
-#endif /* HAVE_VFS_IOV_ITER */
 
 #if defined(HAVE_ITER_IOV)
 #define	zfs_uio_iter_iov(iter)	iter_iov((iter))
@@ -206,4 +200,13 @@ zfs_uio_iov_iter_init(zfs_uio_t *uio, struct iov_iter *iter, offset_t offset,
 #define	zfs_uio_iov_iter_type(iter)	(iter)->type
 #endif
 
+#if defined(HAVE_ITER_IS_UBUF)
+#define	zfs_user_backed_iov_iter(iter)	\
+	(iter_is_ubuf((iter)) || \
+	(zfs_uio_iov_iter_type((iter)) == ITER_IOVEC))
+#else
+#define	zfs_user_backed_iov_iter(iter) \
+	(zfs_uio_iov_iter_type((iter)) == ITER_IOVEC)
+#endif
+
 #endif /* SPL_UIO_H */

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;
 
-			/*
-			 * 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];
+			/* 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.
+					 */
+					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

@@ -48,7 +48,6 @@
 #include <sys/cred.h>
 #include <sys/fs/zfs.h>
 #include <sys/zio_compress.h>
-#include <sys/zio_priority.h>
 #include <sys/uio.h>
 #include <sys/zfs_file.h>
 
@@ -381,6 +380,7 @@ 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/dnode.h

@@ -335,7 +335,7 @@ struct dnode {
 	/* protected by dn_mtx: */
 	kmutex_t dn_mtx;
 	list_t dn_dirty_records[TXG_SIZE];
-	struct range_tree *dn_free_ranges[TXG_SIZE];
+	struct zfs_range_tree *dn_free_ranges[TXG_SIZE];
 	uint64_t dn_allocated_txg;
 	uint64_t dn_free_txg;
 	uint64_t dn_assigned_txg;

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

@@ -21,7 +21,7 @@
 
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2011, 2024 by Delphix. All rights reserved.
+ * Copyright (c) 2011, 2014, 2016, 2024 by Delphix. All rights reserved.
  * Copyright 2011 Nexenta Systems, Inc.  All rights reserved.
  * Copyright (c) 2013, 2017 Joyent, Inc. All rights reserved.
  * Copyright (c) 2014 Integros [integros.com]
@@ -37,7 +37,6 @@
 #define	_SYS_FS_ZFS_H extern __attribute__((visibility("default")))
 
 #include <sys/time.h>
-#include <sys/zio_priority.h>
 
 #ifdef	__cplusplus
 extern "C" {
@@ -265,6 +264,7 @@ 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 +1088,7 @@ 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;
 
@@ -1124,6 +1125,26 @@ typedef enum zio_type {
  */
 #define	ZIO_TYPE_IOCTL	ZIO_TYPE_FLUSH
 
+/*
+ * ZIO priority types. Needed to interpret vdev statistics below.
+ *
+ * NOTE: PLEASE UPDATE THE ENUM STRINGS IN zfs_valstr.c IF YOU ADD ANOTHER
+ * VALUE.
+ */
+typedef enum zio_priority {
+	ZIO_PRIORITY_SYNC_READ,
+	ZIO_PRIORITY_SYNC_WRITE,	/* ZIL */
+	ZIO_PRIORITY_ASYNC_READ,	/* prefetch */
+	ZIO_PRIORITY_ASYNC_WRITE,	/* spa_sync() */
+	ZIO_PRIORITY_SCRUB,		/* asynchronous scrub/resilver reads */
+	ZIO_PRIORITY_REMOVAL,		/* reads/writes for vdev removal */
+	ZIO_PRIORITY_INITIALIZING,	/* initializing I/O */
+	ZIO_PRIORITY_TRIM,		/* trim I/O (discard) */
+	ZIO_PRIORITY_REBUILD,		/* reads/writes for vdev rebuild */
+	ZIO_PRIORITY_NUM_QUEUEABLE,
+	ZIO_PRIORITY_NOW,		/* non-queued i/os (e.g. free) */
+} zio_priority_t;
+
 /*
  * Pool statistics.  Note: all fields should be 64-bit because this
  * is passed between kernel and userland as an nvlist uint64 array.

include/sys/metaslab.h

@@ -139,7 +139,7 @@ void metaslab_set_selected_txg(metaslab_t *, uint64_t);
 
 extern int metaslab_debug_load;
 
-range_seg_type_t metaslab_calculate_range_tree_type(vdev_t *vdev,
+zfs_range_seg_type_t metaslab_calculate_range_tree_type(vdev_t *vdev,
     metaslab_t *msp, uint64_t *start, uint64_t *shift);
 
 #ifdef	__cplusplus

include/sys/metaslab_impl.h

@@ -200,7 +200,7 @@ struct metaslab_class {
 	uint64_t		mc_deferred;	/* total deferred frees */
 	uint64_t		mc_space;	/* total space (alloc + free) */
 	uint64_t		mc_dspace;	/* total deflated space */
-	uint64_t		mc_histogram[RANGE_TREE_HISTOGRAM_SIZE];
+	uint64_t		mc_histogram[ZFS_RANGE_TREE_HISTOGRAM_SIZE];
 
 	/*
 	 * List of all loaded metaslabs in the class, sorted in order of most
@@ -290,7 +290,7 @@ struct metaslab_group {
 	uint64_t		mg_allocations;
 	uint64_t		mg_failed_allocations;
 	uint64_t		mg_fragmentation;
-	uint64_t		mg_histogram[RANGE_TREE_HISTOGRAM_SIZE];
+	uint64_t		mg_histogram[ZFS_RANGE_TREE_HISTOGRAM_SIZE];
 
 	int			mg_ms_disabled;
 	boolean_t		mg_disabled_updating;
@@ -398,8 +398,8 @@ struct metaslab {
 	uint64_t	ms_size;
 	uint64_t	ms_fragmentation;
 
-	range_tree_t	*ms_allocating[TXG_SIZE];
-	range_tree_t	*ms_allocatable;
+	zfs_range_tree_t	*ms_allocating[TXG_SIZE];
+	zfs_range_tree_t	*ms_allocatable;
 	uint64_t	ms_allocated_this_txg;
 	uint64_t	ms_allocating_total;
 
@@ -408,10 +408,12 @@ struct metaslab {
 	 * ms_free*tree only have entries while syncing, and are empty
 	 * between syncs.
 	 */
-	range_tree_t	*ms_freeing;	/* to free this syncing txg */
-	range_tree_t	*ms_freed;	/* already freed this syncing txg */
-	range_tree_t	*ms_defer[TXG_DEFER_SIZE];
-	range_tree_t	*ms_checkpointing; /* to add to the checkpoint */
+	zfs_range_tree_t	*ms_freeing;	/* to free this syncing txg */
+	/* already freed this syncing txg */
+	zfs_range_tree_t	*ms_freed;
+	zfs_range_tree_t	*ms_defer[TXG_DEFER_SIZE];
+	/* to add to the checkpoint */
+	zfs_range_tree_t	*ms_checkpointing;
 
 	/*
 	 * The ms_trim tree is the set of allocatable segments which are
@@ -421,7 +423,7 @@ struct metaslab {
 	 * is unloaded.  Its purpose is to aggregate freed ranges to
 	 * facilitate efficient trimming.
 	 */
-	range_tree_t	*ms_trim;
+	zfs_range_tree_t	*ms_trim;
 
 	boolean_t	ms_condensing;	/* condensing? */
 	boolean_t	ms_condense_wanted;
@@ -542,8 +544,8 @@ struct metaslab {
 	 * Allocs and frees that are committed to the vdev log spacemap but
 	 * not yet to this metaslab's spacemap.
 	 */
-	range_tree_t	*ms_unflushed_allocs;
-	range_tree_t	*ms_unflushed_frees;
+	zfs_range_tree_t	*ms_unflushed_allocs;
+	zfs_range_tree_t	*ms_unflushed_frees;
 
 	/*
 	 * We have flushed entries up to but not including this TXG. In

include/sys/range_tree.h

@@ -37,25 +37,25 @@
 extern "C" {
 #endif
 
-#define	RANGE_TREE_HISTOGRAM_SIZE	64
+#define	ZFS_RANGE_TREE_HISTOGRAM_SIZE	64
 
-typedef struct range_tree_ops range_tree_ops_t;
+typedef struct zfs_range_tree_ops zfs_range_tree_ops_t;
 
-typedef enum range_seg_type {
-	RANGE_SEG32,
-	RANGE_SEG64,
-	RANGE_SEG_GAP,
-	RANGE_SEG_NUM_TYPES,
-} range_seg_type_t;
+typedef enum zfs_range_seg_type {
+	ZFS_RANGE_SEG32,
+	ZFS_RANGE_SEG64,
+	ZFS_RANGE_SEG_GAP,
+	ZFS_RANGE_SEG_NUM_TYPES,
+} zfs_range_seg_type_t;
 
 /*
  * Note: the range_tree may not be accessed concurrently; consumers
  * must provide external locking if required.
  */
-typedef struct range_tree {
+typedef struct zfs_range_tree {
 	zfs_btree_t	rt_root;	/* offset-ordered segment b-tree */
 	uint64_t	rt_space;	/* sum of all segments in the map */
-	range_seg_type_t rt_type;	/* type of range_seg_t in use */
+	zfs_range_seg_type_t rt_type;	/* type of zfs_range_seg_t in use */
 	/*
 	 * All data that is stored in the range tree must have a start higher
 	 * than or equal to rt_start, and all sizes and offsets must be
@@ -63,7 +63,7 @@ typedef struct range_tree {
 	 */
 	uint8_t		rt_shift;
 	uint64_t	rt_start;
-	const range_tree_ops_t *rt_ops;
+	const zfs_range_tree_ops_t *rt_ops;
 	void		*rt_arg;
 	uint64_t	rt_gap;		/* allowable inter-segment gap */
 
@@ -72,61 +72,61 @@ typedef struct range_tree {
 	 * rt_histogram[i], contains the number of ranges whose size is:
 	 * 2^i <= size of range in bytes < 2^(i+1)
 	 */
-	uint64_t	rt_histogram[RANGE_TREE_HISTOGRAM_SIZE];
-} range_tree_t;
+	uint64_t	rt_histogram[ZFS_RANGE_TREE_HISTOGRAM_SIZE];
+} zfs_range_tree_t;
 
-typedef struct range_seg32 {
+typedef struct zfs_range_seg32 {
 	uint32_t	rs_start;	/* starting offset of this segment */
 	uint32_t	rs_end;		/* ending offset (non-inclusive) */
-} range_seg32_t;
+} zfs_range_seg32_t;
 
 /*
  * Extremely large metaslabs, vdev-wide trees, and dnode-wide trees may
  * require 64-bit integers for ranges.
  */
-typedef struct range_seg64 {
+typedef struct zfs_range_seg64 {
 	uint64_t	rs_start;	/* starting offset of this segment */
 	uint64_t	rs_end;		/* ending offset (non-inclusive) */
-} range_seg64_t;
+} zfs_range_seg64_t;
 
-typedef struct range_seg_gap {
+typedef struct zfs_range_seg_gap {
 	uint64_t	rs_start;	/* starting offset of this segment */
 	uint64_t	rs_end;		/* ending offset (non-inclusive) */
 	uint64_t	rs_fill;	/* actual fill if gap mode is on */
-} range_seg_gap_t;
+} zfs_range_seg_gap_t;
 
 /*
  * This type needs to be the largest of the range segs, since it will be stack
  * allocated and then cast the actual type to do tree operations.
  */
-typedef range_seg_gap_t range_seg_max_t;
+typedef zfs_range_seg_gap_t zfs_range_seg_max_t;
 
 /*
  * This is just for clarity of code purposes, so we can make it clear that a
  * pointer is to a range seg of some type; when we need to do the actual math,
  * we'll figure out the real type.
  */
-typedef void range_seg_t;
+typedef void zfs_range_seg_t;
 
-struct range_tree_ops {
-	void    (*rtop_create)(range_tree_t *rt, void *arg);
-	void    (*rtop_destroy)(range_tree_t *rt, void *arg);
-	void	(*rtop_add)(range_tree_t *rt, void *rs, void *arg);
-	void    (*rtop_remove)(range_tree_t *rt, void *rs, void *arg);
-	void	(*rtop_vacate)(range_tree_t *rt, void *arg);
+struct zfs_range_tree_ops {
+	void    (*rtop_create)(zfs_range_tree_t *rt, void *arg);
+	void    (*rtop_destroy)(zfs_range_tree_t *rt, void *arg);
+	void	(*rtop_add)(zfs_range_tree_t *rt, void *rs, void *arg);
+	void    (*rtop_remove)(zfs_range_tree_t *rt, void *rs, void *arg);
+	void	(*rtop_vacate)(zfs_range_tree_t *rt, void *arg);
 };
 
 static inline uint64_t
-rs_get_start_raw(const range_seg_t *rs, const range_tree_t *rt)
+zfs_rs_get_start_raw(const zfs_range_seg_t *rs, const zfs_range_tree_t *rt)
 {
-	ASSERT3U(rt->rt_type, <=, RANGE_SEG_NUM_TYPES);
+	ASSERT3U(rt->rt_type, <=, ZFS_RANGE_SEG_NUM_TYPES);
 	switch (rt->rt_type) {
-	case RANGE_SEG32:
-		return (((const range_seg32_t *)rs)->rs_start);
-	case RANGE_SEG64:
-		return (((const range_seg64_t *)rs)->rs_start);
-	case RANGE_SEG_GAP:
-		return (((const range_seg_gap_t *)rs)->rs_start);
+	case ZFS_RANGE_SEG32:
+		return (((const zfs_range_seg32_t *)rs)->rs_start);
+	case ZFS_RANGE_SEG64:
+		return (((const zfs_range_seg64_t *)rs)->rs_start);
+	case ZFS_RANGE_SEG_GAP:
+		return (((const zfs_range_seg_gap_t *)rs)->rs_start);
 	default:
 		VERIFY(0);
 		return (0);
@@ -134,16 +134,16 @@ rs_get_start_raw(const range_seg_t *rs, const range_tree_t *rt)
 }
 
 static inline uint64_t
-rs_get_end_raw(const range_seg_t *rs, const range_tree_t *rt)
+zfs_rs_get_end_raw(const zfs_range_seg_t *rs, const zfs_range_tree_t *rt)
 {
-	ASSERT3U(rt->rt_type, <=, RANGE_SEG_NUM_TYPES);
+	ASSERT3U(rt->rt_type, <=, ZFS_RANGE_SEG_NUM_TYPES);
 	switch (rt->rt_type) {
-	case RANGE_SEG32:
-		return (((const range_seg32_t *)rs)->rs_end);
-	case RANGE_SEG64:
-		return (((const range_seg64_t *)rs)->rs_end);
-	case RANGE_SEG_GAP:
-		return (((const range_seg_gap_t *)rs)->rs_end);
+	case ZFS_RANGE_SEG32:
+		return (((const zfs_range_seg32_t *)rs)->rs_end);
+	case ZFS_RANGE_SEG64:
+		return (((const zfs_range_seg64_t *)rs)->rs_end);
+	case ZFS_RANGE_SEG_GAP:
+		return (((const zfs_range_seg_gap_t *)rs)->rs_end);
 	default:
 		VERIFY(0);
 		return (0);
@@ -151,20 +151,20 @@ rs_get_end_raw(const range_seg_t *rs, const range_tree_t *rt)
 }
 
 static inline uint64_t
-rs_get_fill_raw(const range_seg_t *rs, const range_tree_t *rt)
+zfs_rs_get_fill_raw(const zfs_range_seg_t *rs, const zfs_range_tree_t *rt)
 {
-	ASSERT3U(rt->rt_type, <=, RANGE_SEG_NUM_TYPES);
+	ASSERT3U(rt->rt_type, <=, ZFS_RANGE_SEG_NUM_TYPES);
 	switch (rt->rt_type) {
-	case RANGE_SEG32: {
-		const range_seg32_t *r32 = (const range_seg32_t *)rs;
+	case ZFS_RANGE_SEG32: {
+		const zfs_range_seg32_t *r32 = (const zfs_range_seg32_t *)rs;
 		return (r32->rs_end - r32->rs_start);
 	}
-	case RANGE_SEG64: {
-		const range_seg64_t *r64 = (const range_seg64_t *)rs;
+	case ZFS_RANGE_SEG64: {
+		const zfs_range_seg64_t *r64 = (const zfs_range_seg64_t *)rs;
 		return (r64->rs_end - r64->rs_start);
 	}
-	case RANGE_SEG_GAP:
-		return (((const range_seg_gap_t *)rs)->rs_fill);
+	case ZFS_RANGE_SEG_GAP:
+		return (((const zfs_range_seg_gap_t *)rs)->rs_fill);
 	default:
 		VERIFY(0);
 		return (0);
@@ -173,37 +173,37 @@ rs_get_fill_raw(const range_seg_t *rs, const range_tree_t *rt)
 }
 
 static inline uint64_t
-rs_get_start(const range_seg_t *rs, const range_tree_t *rt)
+zfs_rs_get_start(const zfs_range_seg_t *rs, const zfs_range_tree_t *rt)
 {
-	return ((rs_get_start_raw(rs, rt) << rt->rt_shift) + rt->rt_start);
+	return ((zfs_rs_get_start_raw(rs, rt) << rt->rt_shift) + rt->rt_start);
 }
 
 static inline uint64_t
-rs_get_end(const range_seg_t *rs, const range_tree_t *rt)
+zfs_rs_get_end(const zfs_range_seg_t *rs, const zfs_range_tree_t *rt)
 {
-	return ((rs_get_end_raw(rs, rt) << rt->rt_shift) + rt->rt_start);
+	return ((zfs_rs_get_end_raw(rs, rt) << rt->rt_shift) + rt->rt_start);
 }
 
 static inline uint64_t
-rs_get_fill(const range_seg_t *rs, const range_tree_t *rt)
+zfs_rs_get_fill(const zfs_range_seg_t *rs, const zfs_range_tree_t *rt)
 {
-	return (rs_get_fill_raw(rs, rt) << rt->rt_shift);
+	return (zfs_rs_get_fill_raw(rs, rt) << rt->rt_shift);
 }
 
 static inline void
-rs_set_start_raw(range_seg_t *rs, range_tree_t *rt, uint64_t start)
+zfs_rs_set_start_raw(zfs_range_seg_t *rs, zfs_range_tree_t *rt, uint64_t start)
 {
-	ASSERT3U(rt->rt_type, <=, RANGE_SEG_NUM_TYPES);
+	ASSERT3U(rt->rt_type, <=, ZFS_RANGE_SEG_NUM_TYPES);
 	switch (rt->rt_type) {
-	case RANGE_SEG32:
+	case ZFS_RANGE_SEG32:
 		ASSERT3U(start, <=, UINT32_MAX);
-		((range_seg32_t *)rs)->rs_start = (uint32_t)start;
+		((zfs_range_seg32_t *)rs)->rs_start = (uint32_t)start;
 		break;
-	case RANGE_SEG64:
-		((range_seg64_t *)rs)->rs_start = start;
+	case ZFS_RANGE_SEG64:
+		((zfs_range_seg64_t *)rs)->rs_start = start;
 		break;
-	case RANGE_SEG_GAP:
-		((range_seg_gap_t *)rs)->rs_start = start;
+	case ZFS_RANGE_SEG_GAP:
+		((zfs_range_seg_gap_t *)rs)->rs_start = start;
 		break;
 	default:
 		VERIFY(0);
@@ -211,19 +211,19 @@ rs_set_start_raw(range_seg_t *rs, range_tree_t *rt, uint64_t start)
 }
 
 static inline void
-rs_set_end_raw(range_seg_t *rs, range_tree_t *rt, uint64_t end)
+zfs_rs_set_end_raw(zfs_range_seg_t *rs, zfs_range_tree_t *rt, uint64_t end)
 {
-	ASSERT3U(rt->rt_type, <=, RANGE_SEG_NUM_TYPES);
+	ASSERT3U(rt->rt_type, <=, ZFS_RANGE_SEG_NUM_TYPES);
 	switch (rt->rt_type) {
-	case RANGE_SEG32:
+	case ZFS_RANGE_SEG32:
 		ASSERT3U(end, <=, UINT32_MAX);
-		((range_seg32_t *)rs)->rs_end = (uint32_t)end;
+		((zfs_range_seg32_t *)rs)->rs_end = (uint32_t)end;
 		break;
-	case RANGE_SEG64:
-		((range_seg64_t *)rs)->rs_end = end;
+	case ZFS_RANGE_SEG64:
+		((zfs_range_seg64_t *)rs)->rs_end = end;
 		break;
-	case RANGE_SEG_GAP:
-		((range_seg_gap_t *)rs)->rs_end = end;
+	case ZFS_RANGE_SEG_GAP:
+		((zfs_range_seg_gap_t *)rs)->rs_end = end;
 		break;
 	default:
 		VERIFY(0);
@@ -231,18 +231,19 @@ rs_set_end_raw(range_seg_t *rs, range_tree_t *rt, uint64_t end)
 }
 
 static inline void
-rs_set_fill_raw(range_seg_t *rs, range_tree_t *rt, uint64_t fill)
+zfs_zfs_rs_set_fill_raw(zfs_range_seg_t *rs, zfs_range_tree_t *rt,
+    uint64_t fill)
 {
-	ASSERT3U(rt->rt_type, <=, RANGE_SEG_NUM_TYPES);
+	ASSERT3U(rt->rt_type, <=, ZFS_RANGE_SEG_NUM_TYPES);
 	switch (rt->rt_type) {
-	case RANGE_SEG32:
+	case ZFS_RANGE_SEG32:
 		/* fall through */
-	case RANGE_SEG64:
-		ASSERT3U(fill, ==, rs_get_end_raw(rs, rt) - rs_get_start_raw(rs,
-		    rt));
+	case ZFS_RANGE_SEG64:
+		ASSERT3U(fill, ==, zfs_rs_get_end_raw(rs, rt) -
+		    zfs_rs_get_start_raw(rs, rt));
 		break;
-	case RANGE_SEG_GAP:
-		((range_seg_gap_t *)rs)->rs_fill = fill;
+	case ZFS_RANGE_SEG_GAP:
+		((zfs_range_seg_gap_t *)rs)->rs_fill = fill;
 		break;
 	default:
 		VERIFY(0);
@@ -250,67 +251,73 @@ rs_set_fill_raw(range_seg_t *rs, range_tree_t *rt, uint64_t fill)
 }
 
 static inline void
-rs_set_start(range_seg_t *rs, range_tree_t *rt, uint64_t start)
+zfs_rs_set_start(zfs_range_seg_t *rs, zfs_range_tree_t *rt, uint64_t start)
 {
 	ASSERT3U(start, >=, rt->rt_start);
 	ASSERT(IS_P2ALIGNED(start, 1ULL << rt->rt_shift));
-	rs_set_start_raw(rs, rt, (start - rt->rt_start) >> rt->rt_shift);
+	zfs_rs_set_start_raw(rs, rt, (start - rt->rt_start) >> rt->rt_shift);
 }
 
 static inline void
-rs_set_end(range_seg_t *rs, range_tree_t *rt, uint64_t end)
+zfs_rs_set_end(zfs_range_seg_t *rs, zfs_range_tree_t *rt, uint64_t end)
 {
 	ASSERT3U(end, >=, rt->rt_start);
 	ASSERT(IS_P2ALIGNED(end, 1ULL << rt->rt_shift));
-	rs_set_end_raw(rs, rt, (end - rt->rt_start) >> rt->rt_shift);
+	zfs_rs_set_end_raw(rs, rt, (end - rt->rt_start) >> rt->rt_shift);
 }
 
 static inline void
-rs_set_fill(range_seg_t *rs, range_tree_t *rt, uint64_t fill)
+zfs_rs_set_fill(zfs_range_seg_t *rs, zfs_range_tree_t *rt, uint64_t fill)
 {
 	ASSERT(IS_P2ALIGNED(fill, 1ULL << rt->rt_shift));
-	rs_set_fill_raw(rs, rt, fill >> rt->rt_shift);
+	zfs_zfs_rs_set_fill_raw(rs, rt, fill >> rt->rt_shift);
 }
 
-typedef void range_tree_func_t(void *arg, uint64_t start, uint64_t size);
+typedef void zfs_range_tree_func_t(void *arg, uint64_t start, uint64_t size);
 
-range_tree_t *range_tree_create_gap(const range_tree_ops_t *ops,
-    range_seg_type_t type, void *arg, uint64_t start, uint64_t shift,
+zfs_range_tree_t *zfs_range_tree_create_gap(const zfs_range_tree_ops_t *ops,
+    zfs_range_seg_type_t type, void *arg, uint64_t start, uint64_t shift,
     uint64_t gap);
-range_tree_t *range_tree_create(const range_tree_ops_t *ops,
-    range_seg_type_t type, void *arg, uint64_t start, uint64_t shift);
-void range_tree_destroy(range_tree_t *rt);
-boolean_t range_tree_contains(range_tree_t *rt, uint64_t start, uint64_t size);
-range_seg_t *range_tree_find(range_tree_t *rt, uint64_t start, uint64_t size);
-boolean_t range_tree_find_in(range_tree_t *rt, uint64_t start, uint64_t size,
-    uint64_t *ostart, uint64_t *osize);
-void range_tree_verify_not_present(range_tree_t *rt,
+zfs_range_tree_t *zfs_range_tree_create(const zfs_range_tree_ops_t *ops,
+    zfs_range_seg_type_t type, void *arg, uint64_t start, uint64_t shift);
+void zfs_range_tree_destroy(zfs_range_tree_t *rt);
+boolean_t zfs_range_tree_contains(zfs_range_tree_t *rt, uint64_t start,
+    uint64_t size);
+zfs_range_seg_t *zfs_range_tree_find(zfs_range_tree_t *rt, uint64_t start,
+    uint64_t size);
+boolean_t zfs_range_tree_find_in(zfs_range_tree_t *rt, uint64_t start,
+    uint64_t size, uint64_t *ostart, uint64_t *osize);
+void zfs_range_tree_verify_not_present(zfs_range_tree_t *rt,
     uint64_t start, uint64_t size);
-void range_tree_resize_segment(range_tree_t *rt, range_seg_t *rs,
+void zfs_range_tree_resize_segment(zfs_range_tree_t *rt, zfs_range_seg_t *rs,
     uint64_t newstart, uint64_t newsize);
-uint64_t range_tree_space(range_tree_t *rt);
-uint64_t range_tree_numsegs(range_tree_t *rt);
-boolean_t range_tree_is_empty(range_tree_t *rt);
-void range_tree_swap(range_tree_t **rtsrc, range_tree_t **rtdst);
-void range_tree_stat_verify(range_tree_t *rt);
-uint64_t range_tree_min(range_tree_t *rt);
-uint64_t range_tree_max(range_tree_t *rt);
-uint64_t range_tree_span(range_tree_t *rt);
+uint64_t zfs_range_tree_space(zfs_range_tree_t *rt);
+uint64_t zfs_range_tree_numsegs(zfs_range_tree_t *rt);
+boolean_t zfs_range_tree_is_empty(zfs_range_tree_t *rt);
+void zfs_range_tree_swap(zfs_range_tree_t **rtsrc, zfs_range_tree_t **rtdst);
+void zfs_range_tree_stat_verify(zfs_range_tree_t *rt);
+uint64_t zfs_range_tree_min(zfs_range_tree_t *rt);
+uint64_t zfs_range_tree_max(zfs_range_tree_t *rt);
+uint64_t zfs_range_tree_span(zfs_range_tree_t *rt);
 
-void range_tree_add(void *arg, uint64_t start, uint64_t size);
-void range_tree_remove(void *arg, uint64_t start, uint64_t size);
-void range_tree_remove_fill(range_tree_t *rt, uint64_t start, uint64_t size);
-void range_tree_adjust_fill(range_tree_t *rt, range_seg_t *rs, int64_t delta);
-void range_tree_clear(range_tree_t *rt, uint64_t start, uint64_t size);
+void zfs_range_tree_add(void *arg, uint64_t start, uint64_t size);
+void zfs_range_tree_remove(void *arg, uint64_t start, uint64_t size);
+void zfs_range_tree_remove_fill(zfs_range_tree_t *rt, uint64_t start,
+    uint64_t size);
+void zfs_range_tree_adjust_fill(zfs_range_tree_t *rt, zfs_range_seg_t *rs,
+    int64_t delta);
+void zfs_range_tree_clear(zfs_range_tree_t *rt, uint64_t start, uint64_t size);
 
-void range_tree_vacate(range_tree_t *rt, range_tree_func_t *func, void *arg);
-void range_tree_walk(range_tree_t *rt, range_tree_func_t *func, void *arg);
-range_seg_t *range_tree_first(range_tree_t *rt);
+void zfs_range_tree_vacate(zfs_range_tree_t *rt, zfs_range_tree_func_t *func,
+    void *arg);
+void zfs_range_tree_walk(zfs_range_tree_t *rt, zfs_range_tree_func_t *func,
+    void *arg);
+zfs_range_seg_t *zfs_range_tree_first(zfs_range_tree_t *rt);
 
-void range_tree_remove_xor_add_segment(uint64_t start, uint64_t end,
-    range_tree_t *removefrom, range_tree_t *addto);
-void range_tree_remove_xor_add(range_tree_t *rt, range_tree_t *removefrom,
-    range_tree_t *addto);
+void zfs_range_tree_remove_xor_add_segment(uint64_t start, uint64_t end,
+    zfs_range_tree_t *removefrom, zfs_range_tree_t *addto);
+void zfs_range_tree_remove_xor_add(zfs_range_tree_t *rt,
+    zfs_range_tree_t *removefrom, zfs_range_tree_t *addto);
 
 #ifdef	__cplusplus
 }

include/sys/sa.h

@@ -49,10 +49,10 @@ typedef uint16_t	sa_attr_type_t;
  * Attribute to register support for.
  */
 typedef struct sa_attr_reg {
-	const char 		*sa_name;	/* attribute name */
-	uint16_t 		sa_length;
+	const char		*sa_name;	/* attribute name */
+	uint16_t		sa_length;
 	sa_bswap_type_t		sa_byteswap;	/* bswap function enum */
-	sa_attr_type_t 		sa_attr; /* filled in during registration */
+	sa_attr_type_t		sa_attr; /* filled in during registration */
 } sa_attr_reg_t;
 
 
@@ -77,7 +77,7 @@ typedef struct sa_bulk_attr {
 	uint16_t		sa_length;
 	sa_attr_type_t		sa_attr;
 	/* the following are private to the sa framework */
-	void 			*sa_addr;
+	void			*sa_addr;
 	uint16_t		sa_buftype;
 	uint16_t		sa_size;
 } sa_bulk_attr_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);
@@ -1209,7 +1213,7 @@ extern void vdev_mirror_stat_fini(void);
 /* Initialization and termination */
 extern void spa_init(spa_mode_t mode);
 extern void spa_fini(void);
-extern void spa_boot_init(void);
+extern void spa_boot_init(void *);
 
 /* properties */
 extern int spa_prop_set(spa_t *spa, nvlist_t *nvp);

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/space_map.h

@@ -207,28 +207,28 @@ boolean_t sm_entry_is_double_word(uint64_t e);
 
 typedef int (*sm_cb_t)(space_map_entry_t *sme, void *arg);
 
-int space_map_load(space_map_t *sm, range_tree_t *rt, maptype_t maptype);
-int space_map_load_length(space_map_t *sm, range_tree_t *rt, maptype_t maptype,
-    uint64_t length);
+int space_map_load(space_map_t *sm, zfs_range_tree_t *rt, maptype_t maptype);
+int space_map_load_length(space_map_t *sm, zfs_range_tree_t *rt,
+    maptype_t maptype, uint64_t length);
 int space_map_iterate(space_map_t *sm, uint64_t length,
     sm_cb_t callback, void *arg);
 int space_map_incremental_destroy(space_map_t *sm, sm_cb_t callback, void *arg,
     dmu_tx_t *tx);
 
-boolean_t space_map_histogram_verify(space_map_t *sm, range_tree_t *rt);
+boolean_t space_map_histogram_verify(space_map_t *sm, zfs_range_tree_t *rt);
 void space_map_histogram_clear(space_map_t *sm);
-void space_map_histogram_add(space_map_t *sm, range_tree_t *rt,
+void space_map_histogram_add(space_map_t *sm, zfs_range_tree_t *rt,
     dmu_tx_t *tx);
 
 uint64_t space_map_object(space_map_t *sm);
 int64_t space_map_allocated(space_map_t *sm);
 uint64_t space_map_length(space_map_t *sm);
-uint64_t space_map_entries(space_map_t *sm, range_tree_t *rt);
+uint64_t space_map_entries(space_map_t *sm, zfs_range_tree_t *rt);
 uint64_t space_map_nblocks(space_map_t *sm);
 
-void space_map_write(space_map_t *sm, range_tree_t *rt, maptype_t maptype,
+void space_map_write(space_map_t *sm, zfs_range_tree_t *rt, maptype_t maptype,
     uint64_t vdev_id, dmu_tx_t *tx);
-uint64_t space_map_estimate_optimal_size(space_map_t *sm, range_tree_t *rt,
+uint64_t space_map_estimate_optimal_size(space_map_t *sm, zfs_range_tree_t *rt,
     uint64_t vdev_id);
 void space_map_truncate(space_map_t *sm, int blocksize, dmu_tx_t *tx);
 uint64_t space_map_alloc(objset_t *os, int blocksize, dmu_tx_t *tx);

include/sys/space_reftree.h

@@ -46,8 +46,8 @@ void space_reftree_create(avl_tree_t *t);
 void space_reftree_destroy(avl_tree_t *t);
 void space_reftree_add_seg(avl_tree_t *t, uint64_t start, uint64_t end,
     int64_t refcnt);
-void space_reftree_add_map(avl_tree_t *t, range_tree_t *rt, int64_t refcnt);
-void space_reftree_generate_map(avl_tree_t *t, range_tree_t *rt,
+void space_reftree_add_map(avl_tree_t *t, zfs_range_tree_t *rt, int64_t refcnt);
+void space_reftree_generate_map(avl_tree_t *t, zfs_range_tree_t *rt,
     int64_t minref);
 
 #ifdef	__cplusplus

include/sys/vdev.h

@@ -106,12 +106,12 @@ extern void vdev_expand(vdev_t *vd, uint64_t txg);
 extern void vdev_split(vdev_t *vd);
 extern void vdev_deadman(vdev_t *vd, const char *tag);
 
-typedef void vdev_xlate_func_t(void *arg, range_seg64_t *physical_rs);
+typedef void vdev_xlate_func_t(void *arg, zfs_range_seg64_t *physical_rs);
 
-extern boolean_t vdev_xlate_is_empty(range_seg64_t *rs);
-extern void vdev_xlate(vdev_t *vd, const range_seg64_t *logical_rs,
-    range_seg64_t *physical_rs, range_seg64_t *remain_rs);
-extern void vdev_xlate_walk(vdev_t *vd, const range_seg64_t *logical_rs,
+extern boolean_t vdev_xlate_is_empty(zfs_range_seg64_t *rs);
+extern void vdev_xlate(vdev_t *vd, const zfs_range_seg64_t *logical_rs,
+    zfs_range_seg64_t *physical_rs, zfs_range_seg64_t *remain_rs);
+extern void vdev_xlate_walk(vdev_t *vd, const zfs_range_seg64_t *logical_rs,
     vdev_xlate_func_t *func, void *arg);
 
 extern void vdev_get_stats_ex(vdev_t *vd, vdev_stat_t *vs, vdev_stat_ex_t *vsx);

include/sys/vdev_impl.h

@@ -91,8 +91,8 @@ typedef void	vdev_remap_func_t(vdev_t *vd, uint64_t offset, uint64_t size,
  * Given a target vdev, translates the logical range "in" to the physical
  * range "res"
  */
-typedef void vdev_xlation_func_t(vdev_t *cvd, const range_seg64_t *logical,
-    range_seg64_t *physical, range_seg64_t *remain);
+typedef void vdev_xlation_func_t(vdev_t *cvd, const zfs_range_seg64_t *logical,
+    zfs_range_seg64_t *physical, zfs_range_seg64_t *remain);
 typedef uint64_t vdev_rebuild_asize_func_t(vdev_t *vd, uint64_t start,
     uint64_t size, uint64_t max_segment);
 typedef void vdev_metaslab_init_func_t(vdev_t *vd, uint64_t *startp,
@@ -299,7 +299,8 @@ struct vdev {
 	kcondvar_t	vdev_initialize_cv;
 	uint64_t	vdev_initialize_offset[TXG_SIZE];
 	uint64_t	vdev_initialize_last_offset;
-	range_tree_t	*vdev_initialize_tree;	/* valid while initializing */
+	/* valid while initializing */
+	zfs_range_tree_t	*vdev_initialize_tree;
 	uint64_t	vdev_initialize_bytes_est;
 	uint64_t	vdev_initialize_bytes_done;
 	uint64_t	vdev_initialize_action_time;	/* start and end time */
@@ -375,7 +376,7 @@ struct vdev {
 	 * from multiple zio threads.
 	 */
 	kmutex_t	vdev_obsolete_lock;
-	range_tree_t	*vdev_obsolete_segments;
+	zfs_range_tree_t	*vdev_obsolete_segments;
 	space_map_t	*vdev_obsolete_sm;
 
 	/*
@@ -388,7 +389,7 @@ struct vdev {
 	/*
 	 * Leaf vdev state.
 	 */
-	range_tree_t	*vdev_dtl[DTL_TYPES]; /* dirty time logs	*/
+	zfs_range_tree_t	*vdev_dtl[DTL_TYPES]; /* dirty time logs */
 	space_map_t	*vdev_dtl_sm;	/* dirty time log space map	*/
 	txg_node_t	vdev_dtl_node;	/* per-txg dirty DTL linkage	*/
 	uint64_t	vdev_dtl_object; /* DTL object			*/
@@ -615,8 +616,8 @@ extern vdev_ops_t vdev_indirect_ops;
 /*
  * Common size functions
  */
-extern void vdev_default_xlate(vdev_t *vd, const range_seg64_t *logical_rs,
-    range_seg64_t *physical_rs, range_seg64_t *remain_rs);
+extern void vdev_default_xlate(vdev_t *vd, const zfs_range_seg64_t *logical_rs,
+    zfs_range_seg64_t *physical_rs, zfs_range_seg64_t *remain_rs);
 extern uint64_t vdev_default_asize(vdev_t *vd, uint64_t psize, uint64_t txg);
 extern uint64_t vdev_default_min_asize(vdev_t *vd);
 extern uint64_t vdev_get_min_asize(vdev_t *vd);
@@ -645,6 +646,10 @@ extern int vdev_obsolete_counts_are_precise(vdev_t *vd, boolean_t *are_precise);
 int vdev_checkpoint_sm_object(vdev_t *vd, uint64_t *sm_obj);
 void vdev_metaslab_group_create(vdev_t *vd);
 uint64_t vdev_best_ashift(uint64_t logical, uint64_t a, uint64_t b);
+#if defined(__linux__)
+int param_get_raidz_impl(char *buf, zfs_kernel_param_t *kp);
+#endif
+int param_set_raidz_impl(ZFS_MODULE_PARAM_ARGS);
 
 /*
  * Vdev ashift optimization tunables

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 *);
 
@@ -66,6 +66,8 @@ extern const zio_vsd_ops_t vdev_raidz_vsd_ops;
 /*
  * vdev_raidz_math interface
  */
+/* Required, but not used, by ZFS_MODULE_PARAM_CALL */
+extern uint32_t zfs_vdev_raidz_impl;
 void vdev_raidz_math_init(void);
 void vdev_raidz_math_fini(void);
 const struct raidz_impl_ops *vdev_raidz_math_get_ops(void);
@@ -73,6 +75,7 @@ int vdev_raidz_math_generate(struct raidz_map *, struct raidz_row *);
 int vdev_raidz_math_reconstruct(struct raidz_map *, struct raidz_row *,
     const int *, const int *, const int);
 int vdev_raidz_impl_set(const char *);
+int vdev_raidz_impl_get(char *buffer, size_t size);
 
 typedef struct vdev_raidz_expand {
 	uint64_t vre_vdev_id;

include/sys/vdev_rebuild.h

@@ -65,7 +65,8 @@ typedef struct vdev_rebuild_phys {
 typedef struct vdev_rebuild {
 	vdev_t		*vr_top_vdev;		/* top-level vdev to rebuild */
 	metaslab_t	*vr_scan_msp;		/* scanning disabled metaslab */
-	range_tree_t	*vr_scan_tree;		/* scan ranges (in metaslab) */
+	/* scan ranges (in metaslab) */
+	zfs_range_tree_t	*vr_scan_tree;
 	kmutex_t	vr_io_lock;		/* inflight IO lock */
 	kcondvar_t	vr_io_cv;		/* inflight IO cv */
 

include/sys/vdev_removal.h

@@ -35,7 +35,7 @@ typedef struct spa_vdev_removal {
 	/* Thread performing a vdev removal. */
 	kthread_t	*svr_thread;
 	/* Segments left to copy from the current metaslab. */
-	range_tree_t	*svr_allocd_segs;
+	zfs_range_tree_t	*svr_allocd_segs;
 	kmutex_t	svr_lock;
 	kcondvar_t	svr_cv;
 	boolean_t	svr_thread_exit;
@@ -49,7 +49,7 @@ typedef struct spa_vdev_removal {
 	 * Ranges that were freed while a mapping was in flight.  This is
 	 * a subset of the ranges covered by vdev_im_new_segments.
 	 */
-	range_tree_t	*svr_frees[TXG_SIZE];
+	zfs_range_tree_t	*svr_frees[TXG_SIZE];
 
 	/*
 	 * Number of bytes which we have finished our work for

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/zfs_ioctl.h

@@ -23,7 +23,7 @@
  * Copyright (c) 2012, 2024 by Delphix. All rights reserved.
  * Copyright 2016 RackTop Systems.
  * Copyright (c) 2017, Intel Corporation.
- * Copyright (c) 2024, Klara, Inc.
+ * Copyright (c) 2024-2025, Klara, Inc.
  */
 
 #ifndef	_SYS_ZFS_IOCTL_H
@@ -421,6 +421,8 @@ typedef struct zinject_record {
 	uint64_t	zi_nlanes;
 	uint32_t	zi_cmd;
 	uint32_t	zi_dvas;
+	uint64_t	zi_match_count;		/* count of times matched */
+	uint64_t	zi_inject_count;	/* count of times injected */
 } zinject_record_t;
 
 #define	ZINJECT_NULL		0x1
@@ -454,6 +456,25 @@ typedef enum zinject_type {
 	ZINJECT_DELAY_EXPORT,
 } zinject_type_t;
 
+typedef enum zinject_iotype {
+	/*
+	 * Compatibility: zi_iotype used to be set to ZIO_TYPE_, so make sure
+	 * the corresponding ZINJECT_IOTYPE_ matches. Note that existing here
+	 * does not mean that injections are possible for all these types.
+	 */
+	ZINJECT_IOTYPE_NULL	= ZIO_TYPE_NULL,
+	ZINJECT_IOTYPE_READ	= ZIO_TYPE_READ,
+	ZINJECT_IOTYPE_WRITE	= ZIO_TYPE_WRITE,
+	ZINJECT_IOTYPE_FREE	= ZIO_TYPE_FREE,
+	ZINJECT_IOTYPE_CLAIM	= ZIO_TYPE_CLAIM,
+	ZINJECT_IOTYPE_FLUSH	= ZIO_TYPE_FLUSH,
+	ZINJECT_IOTYPE_TRIM	= ZIO_TYPE_TRIM,
+	ZINJECT_IOTYPE_ALL	= ZIO_TYPES,
+	/* Room for future expansion for ZIO_TYPE_* */
+	ZINJECT_IOTYPE_PROBE	= 16,
+	ZINJECT_IOTYPES,
+} zinject_iotype_t;
+
 typedef struct zfs_share {
 	uint64_t	z_exportdata;
 	uint64_t	z_sharedata;

include/sys/zio.h

@@ -35,7 +35,6 @@
 #ifndef _ZIO_H
 #define	_ZIO_H
 
-#include <sys/zio_priority.h>
 #include <sys/zfs_context.h>
 #include <sys/spa.h>
 #include <sys/txg.h>
@@ -208,25 +207,25 @@ 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)
@@ -546,6 +545,7 @@ enum blk_verify_flag {
 enum blk_config_flag {
 	BLK_CONFIG_HELD,   // SCL_VDEV held for writer
 	BLK_CONFIG_NEEDED, // SCL_VDEV should be obtained for reader
+	BLK_CONFIG_NEEDED_TRY, // Try with SCL_VDEV for reader
 	BLK_CONFIG_SKIP,   // skip checks which require SCL_VDEV
 };
 
@@ -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,
@@ -662,7 +663,7 @@ extern void zio_suspend(spa_t *spa, zio_t *zio, zio_suspend_reason_t);
 extern int zio_resume(spa_t *spa);
 extern void zio_resume_wait(spa_t *spa);
 
-extern boolean_t zfs_blkptr_verify(spa_t *spa, const blkptr_t *bp,
+extern int zfs_blkptr_verify(spa_t *spa, const blkptr_t *bp,
     enum blk_config_flag blk_config, enum blk_verify_flag blk_verify);
 
 /*

include/sys/zio_priority.h

@@ -1,47 +0,0 @@
-/*
- * CDDL HEADER START
- *
- * This file and its contents are supplied under the terms of the
- * Common Development and Distribution License ("CDDL"), version 1.0.
- * You may only use this file in accordance with the terms of version
- * 1.0 of the CDDL.
- *
- * A full copy of the text of the CDDL should have accompanied this
- * source.  A copy of the CDDL is also available via the Internet at
- * http://www.illumos.org/license/CDDL.
- *
- * CDDL HEADER END
- */
-/*
- * Copyright (c) 2014, 2016 by Delphix. All rights reserved.
- */
-#ifndef	_ZIO_PRIORITY_H
-#define	_ZIO_PRIORITY_H
-
-#ifdef	__cplusplus
-extern "C" {
-#endif
-
-/*
- * NOTE: PLEASE UPDATE THE ENUM STRINGS IN zfs_valstr.c IF YOU ADD ANOTHER
- * VALUE.
- */
-typedef enum zio_priority {
-	ZIO_PRIORITY_SYNC_READ,
-	ZIO_PRIORITY_SYNC_WRITE,	/* ZIL */
-	ZIO_PRIORITY_ASYNC_READ,	/* prefetch */
-	ZIO_PRIORITY_ASYNC_WRITE,	/* spa_sync() */
-	ZIO_PRIORITY_SCRUB,		/* asynchronous scrub/resilver reads */
-	ZIO_PRIORITY_REMOVAL,		/* reads/writes for vdev removal */
-	ZIO_PRIORITY_INITIALIZING,	/* initializing I/O */
-	ZIO_PRIORITY_TRIM,		/* trim I/O (discard) */
-	ZIO_PRIORITY_REBUILD,		/* reads/writes for vdev rebuild */
-	ZIO_PRIORITY_NUM_QUEUEABLE,
-	ZIO_PRIORITY_NOW,		/* non-queued i/os (e.g. free) */
-} zio_priority_t;
-
-#ifdef	__cplusplus
-}
-#endif
-
-#endif	/* _ZIO_PRIORITY_H */

include/sys/zvol_impl.h

@@ -88,6 +88,11 @@ int zvol_get_data(void *arg, uint64_t arg2, lr_write_t *lr, char *buf,
 int zvol_init_impl(void);
 void zvol_fini_impl(void);
 void zvol_wait_close(zvol_state_t *zv);
+int zvol_clone_range(zvol_state_handle_t *, uint64_t,
+    zvol_state_handle_t *, uint64_t, uint64_t);
+void zvol_log_clone_range(zilog_t *zilog, dmu_tx_t *tx, int txtype,
+    uint64_t off, uint64_t len, uint64_t blksz, const blkptr_t *bps,
+    size_t nbps);
 
 /*
  * platform dependent functions exported to platform independent code