aboutsummaryrefslogtreecommitdiff
path: root/sys/contrib/openzfs
diff options
context:
space:
mode:
Diffstat (limited to 'sys/contrib/openzfs')
-rwxr-xr-xsys/contrib/openzfs/.github/workflows/scripts/qemu-2-start.sh2
-rwxr-xr-xsys/contrib/openzfs/.github/workflows/scripts/qemu-4-build-vm.sh23
-rw-r--r--sys/contrib/openzfs/.github/workflows/zfs-qemu-packages.yml13
-rw-r--r--sys/contrib/openzfs/.mailmap4
-rw-r--r--sys/contrib/openzfs/AUTHORS13
-rw-r--r--sys/contrib/openzfs/META2
-rw-r--r--sys/contrib/openzfs/cmd/zdb/zdb.c91
-rw-r--r--sys/contrib/openzfs/cmd/zed/zed.d/zed-functions.sh3
-rw-r--r--sys/contrib/openzfs/cmd/zfs/zfs_main.c2
-rw-r--r--sys/contrib/openzfs/cmd/zhack.c10
-rwxr-xr-xsys/contrib/openzfs/cmd/zilstat.in1
-rw-r--r--sys/contrib/openzfs/cmd/zpool/zpool_main.c13
-rw-r--r--sys/contrib/openzfs/cmd/ztest.c29
-rw-r--r--sys/contrib/openzfs/config/kernel-mkdir.m42
-rw-r--r--sys/contrib/openzfs/config/toolchain-simd.m4113
-rw-r--r--sys/contrib/openzfs/contrib/debian/control8
-rw-r--r--sys/contrib/openzfs/contrib/icp/gcm-simd/boringssl/LICENSE253
-rw-r--r--sys/contrib/openzfs/contrib/icp/gcm-simd/boringssl/README11
-rw-r--r--sys/contrib/openzfs/contrib/icp/gcm-simd/boringssl/aes-gcm-avx2-x86_64-linux.S1328
-rw-r--r--sys/contrib/openzfs/contrib/initramfs/scripts/zfs3
-rw-r--r--sys/contrib/openzfs/include/os/freebsd/spl/sys/debug.h4
-rw-r--r--sys/contrib/openzfs/include/os/freebsd/spl/sys/proc.h4
-rw-r--r--sys/contrib/openzfs/include/os/linux/kernel/linux/simd_x86.h47
-rw-r--r--sys/contrib/openzfs/include/os/linux/spl/sys/debug.h4
-rw-r--r--sys/contrib/openzfs/include/os/linux/spl/sys/mutex.h2
-rw-r--r--sys/contrib/openzfs/include/os/linux/zfs/sys/trace_zil.h6
-rw-r--r--sys/contrib/openzfs/include/sys/dmu.h4
-rw-r--r--sys/contrib/openzfs/include/sys/spa.h2
-rw-r--r--sys/contrib/openzfs/include/sys/zfs_znode.h2
-rw-r--r--sys/contrib/openzfs/include/sys/zil.h33
-rw-r--r--sys/contrib/openzfs/include/sys/zil_impl.h22
-rw-r--r--sys/contrib/openzfs/include/sys/zio.h29
-rw-r--r--sys/contrib/openzfs/include/sys/zvol_impl.h5
-rw-r--r--sys/contrib/openzfs/lib/libicp/Makefile.am1
-rw-r--r--sys/contrib/openzfs/lib/libspl/include/sys/debug.h4
-rw-r--r--sys/contrib/openzfs/lib/libspl/include/sys/simd.h28
-rw-r--r--sys/contrib/openzfs/lib/libzfs/libzfs_diff.c4
-rw-r--r--sys/contrib/openzfs/lib/libzfs/libzfs_import.c2
-rw-r--r--sys/contrib/openzfs/lib/libzfs/libzfs_mount.c2
-rw-r--r--sys/contrib/openzfs/lib/libzfs/libzfs_sendrecv.c8
-rw-r--r--sys/contrib/openzfs/lib/libzpool/abd_os.c4
-rw-r--r--sys/contrib/openzfs/lib/libzpool/kernel.c77
-rw-r--r--sys/contrib/openzfs/lib/libzpool/util.c8
-rw-r--r--sys/contrib/openzfs/lib/libzutil/zutil_import.c43
-rw-r--r--sys/contrib/openzfs/man/man4/zfs.44
-rw-r--r--sys/contrib/openzfs/man/man8/zdb.814
-rw-r--r--sys/contrib/openzfs/man/man8/zfs-send.824
-rw-r--r--sys/contrib/openzfs/module/Kbuild.in1
-rw-r--r--sys/contrib/openzfs/module/avl/avl.c16
-rw-r--r--sys/contrib/openzfs/module/icp/algs/modes/gcm.c371
-rw-r--r--sys/contrib/openzfs/module/icp/algs/modes/modes.c2
-rw-r--r--sys/contrib/openzfs/module/icp/asm-x86_64/modes/THIRDPARTYLICENSE.boringssl253
-rw-r--r--sys/contrib/openzfs/module/icp/asm-x86_64/modes/THIRDPARTYLICENSE.boringssl.descrip1
-rw-r--r--sys/contrib/openzfs/module/icp/asm-x86_64/modes/aesni-gcm-avx2-vaes.S1323
-rw-r--r--sys/contrib/openzfs/module/icp/core/kcf_sched.c2
-rw-r--r--sys/contrib/openzfs/module/icp/include/modes/modes.h13
-rw-r--r--sys/contrib/openzfs/module/icp/io/aes.c10
-rw-r--r--sys/contrib/openzfs/module/nvpair/nvpair.c6
-rw-r--r--sys/contrib/openzfs/module/os/freebsd/spl/spl_kmem.c2
-rw-r--r--sys/contrib/openzfs/module/os/freebsd/spl/spl_sysevent.c2
-rw-r--r--sys/contrib/openzfs/module/os/freebsd/zfs/abd_os.c4
-rw-r--r--sys/contrib/openzfs/module/os/freebsd/zfs/dmu_os.c8
-rw-r--r--sys/contrib/openzfs/module/os/freebsd/zfs/kmod_core.c7
-rw-r--r--sys/contrib/openzfs/module/os/freebsd/zfs/vdev_geom.c2
-rw-r--r--sys/contrib/openzfs/module/os/freebsd/zfs/zfs_acl.c4
-rw-r--r--sys/contrib/openzfs/module/os/freebsd/zfs/zfs_ctldir.c48
-rw-r--r--sys/contrib/openzfs/module/os/freebsd/zfs/zfs_dir.c4
-rw-r--r--sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vfsops.c11
-rw-r--r--sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vnops_os.c98
-rw-r--r--sys/contrib/openzfs/module/os/freebsd/zfs/zfs_znode_os.c18
-rw-r--r--sys/contrib/openzfs/module/os/freebsd/zfs/zvol_os.c226
-rw-r--r--sys/contrib/openzfs/module/os/linux/spl/spl-condvar.c8
-rw-r--r--sys/contrib/openzfs/module/os/linux/spl/spl-generic.c2
-rw-r--r--sys/contrib/openzfs/module/os/linux/spl/spl-kmem-cache.c24
-rw-r--r--sys/contrib/openzfs/module/os/linux/spl/spl-kstat.c2
-rw-r--r--sys/contrib/openzfs/module/os/linux/spl/spl-thread.c2
-rw-r--r--sys/contrib/openzfs/module/os/linux/spl/spl-tsd.c2
-rw-r--r--sys/contrib/openzfs/module/os/linux/zfs/abd_os.c6
-rw-r--r--sys/contrib/openzfs/module/os/linux/zfs/vdev_disk.c6
-rw-r--r--sys/contrib/openzfs/module/os/linux/zfs/zfs_acl.c6
-rw-r--r--sys/contrib/openzfs/module/os/linux/zfs/zfs_ctldir.c8
-rw-r--r--sys/contrib/openzfs/module/os/linux/zfs/zfs_dir.c24
-rw-r--r--sys/contrib/openzfs/module/os/linux/zfs/zfs_sysfs.c2
-rw-r--r--sys/contrib/openzfs/module/os/linux/zfs/zfs_vfsops.c27
-rw-r--r--sys/contrib/openzfs/module/os/linux/zfs/zfs_vnops_os.c132
-rw-r--r--sys/contrib/openzfs/module/os/linux/zfs/zfs_znode_os.c58
-rw-r--r--sys/contrib/openzfs/module/os/linux/zfs/zpl_file.c63
-rw-r--r--sys/contrib/openzfs/module/os/linux/zfs/zpl_inode.c2
-rw-r--r--sys/contrib/openzfs/module/os/linux/zfs/zpl_super.c4
-rw-r--r--sys/contrib/openzfs/module/os/linux/zfs/zpl_xattr.c4
-rw-r--r--sys/contrib/openzfs/module/os/linux/zfs/zvol_os.c208
-rw-r--r--sys/contrib/openzfs/module/zcommon/simd_stat.c4
-rw-r--r--sys/contrib/openzfs/module/zfs/abd.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/arc.c68
-rw-r--r--sys/contrib/openzfs/module/zfs/bpobj.c6
-rw-r--r--sys/contrib/openzfs/module/zfs/btree.c8
-rw-r--r--sys/contrib/openzfs/module/zfs/dataset_kstats.c1
-rw-r--r--sys/contrib/openzfs/module/zfs/dbuf.c81
-rw-r--r--sys/contrib/openzfs/module/zfs/ddt.c8
-rw-r--r--sys/contrib/openzfs/module/zfs/ddt_log.c8
-rw-r--r--sys/contrib/openzfs/module/zfs/dmu.c12
-rw-r--r--sys/contrib/openzfs/module/zfs/dmu_direct.c6
-rw-r--r--sys/contrib/openzfs/module/zfs/dmu_object.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/dmu_objset.c4
-rw-r--r--sys/contrib/openzfs/module/zfs/dmu_recv.c12
-rw-r--r--sys/contrib/openzfs/module/zfs/dmu_redact.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/dmu_send.c6
-rw-r--r--sys/contrib/openzfs/module/zfs/dmu_tx.c18
-rw-r--r--sys/contrib/openzfs/module/zfs/dnode.c22
-rw-r--r--sys/contrib/openzfs/module/zfs/dnode_sync.c6
-rw-r--r--sys/contrib/openzfs/module/zfs/dsl_bookmark.c6
-rw-r--r--sys/contrib/openzfs/module/zfs/dsl_crypt.c4
-rw-r--r--sys/contrib/openzfs/module/zfs/dsl_dataset.c10
-rw-r--r--sys/contrib/openzfs/module/zfs/dsl_deadlist.c5
-rw-r--r--sys/contrib/openzfs/module/zfs/dsl_deleg.c20
-rw-r--r--sys/contrib/openzfs/module/zfs/dsl_destroy.c8
-rw-r--r--sys/contrib/openzfs/module/zfs/dsl_dir.c4
-rw-r--r--sys/contrib/openzfs/module/zfs/dsl_pool.c10
-rw-r--r--sys/contrib/openzfs/module/zfs/dsl_prop.c31
-rw-r--r--sys/contrib/openzfs/module/zfs/dsl_scan.c6
-rw-r--r--sys/contrib/openzfs/module/zfs/dsl_userhold.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/fm.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/metaslab.c34
-rw-r--r--sys/contrib/openzfs/module/zfs/mmp.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/multilist.c4
-rw-r--r--sys/contrib/openzfs/module/zfs/range_tree.c4
-rw-r--r--sys/contrib/openzfs/module/zfs/rrwlock.c8
-rw-r--r--sys/contrib/openzfs/module/zfs/sa.c24
-rw-r--r--sys/contrib/openzfs/module/zfs/spa.c66
-rw-r--r--sys/contrib/openzfs/module/zfs/spa_config.c115
-rw-r--r--sys/contrib/openzfs/module/zfs/spa_misc.c27
-rw-r--r--sys/contrib/openzfs/module/zfs/spa_stats.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/space_map.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/space_reftree.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/vdev.c46
-rw-r--r--sys/contrib/openzfs/module/zfs/vdev_draid.c16
-rw-r--r--sys/contrib/openzfs/module/zfs/vdev_indirect.c6
-rw-r--r--sys/contrib/openzfs/module/zfs/vdev_initialize.c6
-rw-r--r--sys/contrib/openzfs/module/zfs/vdev_label.c14
-rw-r--r--sys/contrib/openzfs/module/zfs/vdev_queue.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/vdev_raidz.c8
-rw-r--r--sys/contrib/openzfs/module/zfs/vdev_rebuild.c4
-rw-r--r--sys/contrib/openzfs/module/zfs/vdev_removal.c22
-rw-r--r--sys/contrib/openzfs/module/zfs/vdev_trim.c8
-rw-r--r--sys/contrib/openzfs/module/zfs/zap.c8
-rw-r--r--sys/contrib/openzfs/module/zfs/zap_micro.c6
-rw-r--r--sys/contrib/openzfs/module/zfs/zcp.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/zfeature.c4
-rw-r--r--sys/contrib/openzfs/module/zfs/zfs_fuid.c44
-rw-r--r--sys/contrib/openzfs/module/zfs/zfs_ioctl.c119
-rw-r--r--sys/contrib/openzfs/module/zfs/zfs_log.c4
-rw-r--r--sys/contrib/openzfs/module/zfs/zfs_quota.c4
-rw-r--r--sys/contrib/openzfs/module/zfs/zfs_rlock.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/zfs_sa.c15
-rw-r--r--sys/contrib/openzfs/module/zfs/zfs_vnops.c27
-rw-r--r--sys/contrib/openzfs/module/zfs/zil.c780
-rw-r--r--sys/contrib/openzfs/module/zfs/zio.c113
-rw-r--r--sys/contrib/openzfs/module/zfs/zio_checksum.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/zio_compress.c15
-rw-r--r--sys/contrib/openzfs/module/zfs/zio_inject.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/zrlock.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/zthr.c2
-rw-r--r--sys/contrib/openzfs/module/zfs/zvol.c396
-rw-r--r--sys/contrib/openzfs/module/zstd/zfs_zstd.c8
-rw-r--r--sys/contrib/openzfs/rpm/generic/zfs.spec.in1
-rwxr-xr-xsys/contrib/openzfs/scripts/spdxcheck.pl1
-rw-r--r--sys/contrib/openzfs/tests/runfiles/common.run8
-rw-r--r--sys/contrib/openzfs/tests/zfs-tests/cmd/.gitignore1
-rw-r--r--sys/contrib/openzfs/tests/zfs-tests/cmd/Makefile.am2
-rw-r--r--sys/contrib/openzfs/tests/zfs-tests/cmd/btree_test.c4
-rw-r--r--sys/contrib/openzfs/tests/zfs-tests/cmd/crypto_test.c2
-rw-r--r--sys/contrib/openzfs/tests/zfs-tests/cmd/mmap_write_sync.c84
-rw-r--r--sys/contrib/openzfs/tests/zfs-tests/include/commands.cfg1
-rw-r--r--sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am10
-rw-r--r--sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode.kshlib149
-rwxr-xr-xsys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_fsync_continue.ksh36
-rwxr-xr-xsys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_fsync_wait.ksh36
-rwxr-xr-xsys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_msync_continue.ksh36
-rwxr-xr-xsys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_msync_wait.ksh36
-rwxr-xr-xsys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_osync_continue.ksh36
-rwxr-xr-xsys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_osync_wait.ksh37
-rwxr-xr-xsys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_syncalways_continue.ksh37
-rwxr-xr-xsys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_syncalways_wait.ksh37
-rwxr-xr-xsys/contrib/openzfs/tests/zfs-tests/tests/functional/zvol/zvol_stress/zvol_stress_destroy.ksh66
184 files changed, 6559 insertions, 1755 deletions
diff --git a/sys/contrib/openzfs/.github/workflows/scripts/qemu-2-start.sh b/sys/contrib/openzfs/.github/workflows/scripts/qemu-2-start.sh
index 885a64037f89..70a2364f1fc6 100755
--- a/sys/contrib/openzfs/.github/workflows/scripts/qemu-2-start.sh
+++ b/sys/contrib/openzfs/.github/workflows/scripts/qemu-2-start.sh
@@ -109,7 +109,7 @@ case "$OS" in
KSRC="$FREEBSD_SNAP/../amd64/$FreeBSD/src.txz"
;;
freebsd15-0c)
- FreeBSD="15.0-CURRENT"
+ FreeBSD="15.0-PRERELEASE"
OSNAME="FreeBSD $FreeBSD"
OSv="freebsd14.0"
URLxz="$FREEBSD_SNAP/$FreeBSD/amd64/Latest/FreeBSD-$FreeBSD-amd64-BASIC-CI-ufs.raw.xz"
diff --git a/sys/contrib/openzfs/.github/workflows/scripts/qemu-4-build-vm.sh b/sys/contrib/openzfs/.github/workflows/scripts/qemu-4-build-vm.sh
index 17e976ebcc39..2807d9e77127 100755
--- a/sys/contrib/openzfs/.github/workflows/scripts/qemu-4-build-vm.sh
+++ b/sys/contrib/openzfs/.github/workflows/scripts/qemu-4-build-vm.sh
@@ -5,12 +5,13 @@
#
# Usage:
#
-# qemu-4-build-vm.sh OS [--enable-debug][--dkms][--poweroff]
-# [--release][--repo][--tarball]
+# qemu-4-build-vm.sh OS [--enable-debug][--dkms][--patch-level NUM]
+# [--poweroff][--release][--repo][--tarball]
#
# OS: OS name like 'fedora41'
# --enable-debug: Build RPMs with '--enable-debug' (for testing)
# --dkms: Build DKMS RPMs as well
+# --patch-level NUM: Use a custom patch level number for packages.
# --poweroff: Power-off the VM after building
# --release Build zfs-release*.rpm as well
# --repo After building everything, copy RPMs into /tmp/repo
@@ -21,6 +22,7 @@
ENABLE_DEBUG=""
DKMS=""
+PATCH_LEVEL=""
POWEROFF=""
RELEASE=""
REPO=""
@@ -35,6 +37,11 @@ while [[ $# -gt 0 ]]; do
DKMS=1
shift
;;
+ --patch-level)
+ PATCH_LEVEL=$2
+ shift
+ shift
+ ;;
--poweroff)
POWEROFF=1
shift
@@ -215,6 +222,10 @@ function rpm_build_and_install() {
run ./autogen.sh
echo "##[endgroup]"
+ if [ -n "$PATCH_LEVEL" ] ; then
+ sed -i -E 's/(Release:\s+)1/\1'$PATCH_LEVEL'/g' META
+ fi
+
echo "##[group]Configure"
run ./configure --enable-debuginfo $extra
echo "##[endgroup]"
@@ -328,7 +339,13 @@ fi
# almalinux9.5
# fedora42
source /etc/os-release
-sudo hostname "$ID$VERSION_ID"
+ if which hostnamectl &> /dev/null ; then
+ # Fedora 42+ use hostnamectl
+ sudo hostnamectl set-hostname "$ID$VERSION_ID"
+ sudo hostnamectl set-hostname --pretty "$ID$VERSION_ID"
+else
+ sudo hostname "$ID$VERSION_ID"
+fi
# save some sysinfo
uname -a > /var/tmp/uname.txt
diff --git a/sys/contrib/openzfs/.github/workflows/zfs-qemu-packages.yml b/sys/contrib/openzfs/.github/workflows/zfs-qemu-packages.yml
index 5b5afe746859..d8a95954fe1a 100644
--- a/sys/contrib/openzfs/.github/workflows/zfs-qemu-packages.yml
+++ b/sys/contrib/openzfs/.github/workflows/zfs-qemu-packages.yml
@@ -32,6 +32,11 @@ on:
options:
- "Build RPMs"
- "Test repo"
+ patch_level:
+ type: string
+ required: false
+ default: ""
+ description: "(optional) patch level number"
repo_url:
type: string
required: false
@@ -78,7 +83,13 @@ jobs:
mkdir -p /tmp/repo
ssh zfs@vm0 '$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 ${{ matrix.os }}
+ EXTRA=""
+ if [ -n "${{ github.event.inputs.patch_level }}" ] ; then
+ EXTRA="--patch-level ${{ github.event.inputs.patch_level }}"
+ fi
+
+ .github/workflows/scripts/qemu-4-build.sh $EXTRA \
+ --repo --release --dkms --tarball ${{ matrix.os }}
fi
- name: Prepare artifacts
diff --git a/sys/contrib/openzfs/.mailmap b/sys/contrib/openzfs/.mailmap
index b6d942c000b8..e6f09c6c9d43 100644
--- a/sys/contrib/openzfs/.mailmap
+++ b/sys/contrib/openzfs/.mailmap
@@ -23,6 +23,7 @@
# These maps are making names consistent where they have varied but the email
# address has never changed. In most cases, the full name is in the
# Signed-off-by of a commit with a matching author.
+Achill Gilgenast <achill@achill.org>
Ahelenia Ziemiańska <nabijaczleweli@gmail.com>
Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Alex John <alex@stty.io>
@@ -37,6 +38,7 @@ Crag Wang <crag0715@gmail.com>
Damian Szuberski <szuberskidamian@gmail.com>
Daniel Kolesa <daniel@octaforge.org>
Debabrata Banerjee <dbavatar@gmail.com>
+Diwakar Kristappagari <diwakar-k@hpe.com>
Finix Yan <yanchongwen@hotmail.com>
Gaurav Kumar <gauravk.18@gmail.com>
Gionatan Danti <g.danti@assyoma.it>
@@ -145,6 +147,7 @@ Gaurav Kumar <gauravk.18@gmail.com> <gaurkuma@users.noreply.github.com>
George Gaydarov <git@gg7.io> <gg7@users.noreply.github.com>
Georgy Yakovlev <gyakovlev@gentoo.org> <168902+gyakovlev@users.noreply.github.com>
Gerardwx <gerardw@alum.mit.edu> <Gerardwx@users.noreply.github.com>
+Germano Massullo <germano.massullo@gmail.com> <Germano0@users.noreply.github.com>
Gian-Carlo DeFazio <defazio1@llnl.gov> <defaziogiancarlo@users.noreply.github.com>
Giuseppe Di Natale <dinatale2@llnl.gov> <dinatale2@users.noreply.github.com>
Hajo Möller <dasjoe@gmail.com> <dasjoe@users.noreply.github.com>
@@ -164,6 +167,7 @@ John Ramsden <johnramsden@riseup.net> <johnramsden@users.noreply.github.com>
Jonathon Fernyhough <jonathon@m2x.dev> <559369+jonathonf@users.noreply.github.com>
Jose Luis Duran <jlduran@gmail.com> <jlduran@users.noreply.github.com>
Justin Hibbits <chmeeedalf@gmail.com> <chmeeedalf@users.noreply.github.com>
+Kaitlin Hoang <kthoang@amazon.com> <khoang98@users.noreply.github.com>
Kevin Greene <kevin.greene@delphix.com> <104801862+kxgreene@users.noreply.github.com>
Kevin Jin <lostking2008@hotmail.com> <33590050+jxdking@users.noreply.github.com>
Kevin P. Fleming <kevin@km6g.us> <kpfleming@users.noreply.github.com>
diff --git a/sys/contrib/openzfs/AUTHORS b/sys/contrib/openzfs/AUTHORS
index a9d249a66f1e..6c34c07f39ef 100644
--- a/sys/contrib/openzfs/AUTHORS
+++ b/sys/contrib/openzfs/AUTHORS
@@ -10,6 +10,7 @@ PAST MAINTAINERS:
CONTRIBUTORS:
Aaron Fineman <abyxcos@gmail.com>
+ Achill Gilgenast <achill@achill.org>
Adam D. Moss <c@yotes.com>
Adam Leventhal <ahl@delphix.com>
Adam Stevko <adam.stevko@gmail.com>
@@ -59,6 +60,7 @@ CONTRIBUTORS:
Andreas Buschmann <andreas.buschmann@tech.net.de>
Andreas Dilger <adilger@intel.com>
Andreas Vögele <andreas@andreasvoegele.com>
+ Andres <a-d-j-i@users.noreply.github.com>
Andrew Barnes <barnes333@gmail.com>
Andrew Hamilton <ahamilto@tjhsst.edu>
Andrew Innes <andrew.c12@gmail.com>
@@ -72,6 +74,7 @@ CONTRIBUTORS:
Andrey Prokopenko <job@terem.fr>
Andrey Vesnovaty <andrey.vesnovaty@gmail.com>
Andriy Gapon <avg@freebsd.org>
+ Andriy Tkachuk <andriy.tkachuk@seagate.com>
Andy Bakun <github@thwartedefforts.org>
Andy Fiddaman <omnios@citrus-it.co.uk>
Aniruddha Shankar <k@191a.net>
@@ -120,6 +123,7 @@ CONTRIBUTORS:
Caleb James DeLisle <calebdelisle@lavabit.com>
Cameron Harr <harr1@llnl.gov>
Cao Xuewen <cao.xuewen@zte.com.cn>
+ Carl George <carlwgeorge@gmail.com>
Carlo Landmeter <clandmeter@gmail.com>
Carlos Alberto Lopez Perez <clopez@igalia.com>
Cedric Maunoury <cedric.maunoury@gmail.com>
@@ -200,6 +204,7 @@ CONTRIBUTORS:
Dimitri John Ledkov <xnox@ubuntu.com>
Dimitry Andric <dimitry@andric.com>
Dirkjan Bussink <d.bussink@gmail.com>
+ Diwakar Kristappagari <diwakar-k@hpe.com>
Dmitry Khasanov <pik4ez@gmail.com>
Dominic Pearson <dsp@technoanimal.net>
Dominik Hassler <hadfl@omniosce.org>
@@ -250,6 +255,7 @@ CONTRIBUTORS:
George Wilson <gwilson@delphix.com>
Georgy Yakovlev <ya@sysdump.net>
Gerardwx <gerardw@alum.mit.edu>
+ Germano Massullo <germano.massullo@gmail.com>
Gian-Carlo DeFazio <defazio1@llnl.gov>
Gionatan Danti <g.danti@assyoma.it>
Giuseppe Di Natale <guss80@gmail.com>
@@ -287,6 +293,7 @@ CONTRIBUTORS:
Igor K <igor@dilos.org>
Igor Kozhukhov <ikozhukhov@gmail.com>
Igor Lvovsky <ilvovsky@gmail.com>
+ Igor Ostapenko <pm@igoro.pro>
ilbsmart <wgqimut@gmail.com>
Ilkka Sovanto <github@ilkka.kapsi.fi>
illiliti <illiliti@protonmail.com>
@@ -326,6 +333,7 @@ CONTRIBUTORS:
Jinshan Xiong <jinshan.xiong@intel.com>
Jitendra Patidar <jitendra.patidar@nutanix.com>
JK Dingwall <james@dingwall.me.uk>
+ Joel Low <joel@joelsplace.sg>
Joe Stein <joe.stein@delphix.com>
John-Mark Gurney <jmg@funkthat.com>
John Albietz <inthecloud247@gmail.com>
@@ -374,6 +382,7 @@ CONTRIBUTORS:
Kevin Jin <lostking2008@hotmail.com>
Kevin P. Fleming <kevin@km6g.us>
Kevin Tanguy <kevin.tanguy@ovh.net>
+ khoang98 <khoang98@users.noreply.github.com>
KireinaHoro <i@jsteward.moe>
Kjeld Schouten-Lebbing <kjeld@schouten-lebbing.nl>
Kleber Tarcísio <klebertarcisio@yahoo.com.br>
@@ -447,6 +456,7 @@ CONTRIBUTORS:
Max Zettlmeißl <max@zettlmeissl.de>
Md Islam <mdnahian@outlook.com>
megari <megari@iki.fi>
+ Meriel Luna Mittelbach <lunarlambda@gmail.com>
Michael D Labriola <michael.d.labriola@gmail.com>
Michael Franzl <michael@franzl.name>
Michael Gebetsroither <michael@mgeb.org>
@@ -494,6 +504,7 @@ CONTRIBUTORS:
Orivej Desh <orivej@gmx.fr>
Pablo Correa Gómez <ablocorrea@hotmail.com>
Palash Gandhi <pbg4930@rit.edu>
+ Patrick Fasano <patrick@patrickfasano.com>
Patrick Mooney <pmooney@pfmooney.com>
Patrik Greco <sikevux@sikevux.se>
Paul B. Henson <henson@acm.org>
@@ -535,6 +546,7 @@ CONTRIBUTORS:
Remy Blank <remy.blank@pobox.com>
renelson <bnelson@nelsonbe.com>
Reno Reckling <e-github@wthack.de>
+ René Wirnata <rene.wirnata@pandascience.net>
Ricardo M. Correia <ricardo.correia@oracle.com>
Riccardo Schirone <rschirone91@gmail.com>
Richard Allen <belperite@gmail.com>
@@ -640,6 +652,7 @@ CONTRIBUTORS:
tleydxdy <shironeko.github@tesaguri.club>
Tobin Harding <me@tobin.cc>
Todd Seidelmann <seidelma@users.noreply.github.com>
+ Todd Zullinger <tmz@pobox.com>
Tom Caputi <tcaputi@datto.com>
Tom Matthews <tom@axiom-partners.com>
Tomohiro Kusumi <kusumi.tomohiro@gmail.com>
diff --git a/sys/contrib/openzfs/META b/sys/contrib/openzfs/META
index 47f0795bfa11..1a9c671feac6 100644
--- a/sys/contrib/openzfs/META
+++ b/sys/contrib/openzfs/META
@@ -6,5 +6,5 @@ Release: 1
Release-Tags: relext
License: CDDL
Author: OpenZFS
-Linux-Maximum: 6.15
+Linux-Maximum: 6.16
Linux-Minimum: 4.18
diff --git a/sys/contrib/openzfs/cmd/zdb/zdb.c b/sys/contrib/openzfs/cmd/zdb/zdb.c
index bf44d9c322b4..957e6c67dd12 100644
--- a/sys/contrib/openzfs/cmd/zdb/zdb.c
+++ b/sys/contrib/openzfs/cmd/zdb/zdb.c
@@ -127,6 +127,7 @@ static zfs_range_tree_t *mos_refd_objs;
static spa_t *spa;
static objset_t *os;
static boolean_t kernel_init_done;
+static boolean_t corruption_found = B_FALSE;
static void snprintf_blkptr_compact(char *, size_t, const blkptr_t *,
boolean_t);
@@ -176,7 +177,7 @@ static int
sublivelist_verify_blkptr(void *arg, const blkptr_t *bp, boolean_t free,
dmu_tx_t *tx)
{
- ASSERT3P(tx, ==, NULL);
+ ASSERT0P(tx);
struct sublivelist_verify *sv = arg;
sublivelist_verify_block_refcnt_t current = {
.svbr_blk = *bp,
@@ -250,6 +251,7 @@ sublivelist_verify_func(void *args, dsl_deadlist_entry_t *dle)
&e->svbr_blk, B_TRUE);
(void) printf("\tERROR: %d unmatched FREE(s): %s\n",
e->svbr_refcnt, blkbuf);
+ corruption_found = B_TRUE;
}
zfs_btree_destroy(&sv->sv_pair);
@@ -405,6 +407,7 @@ verify_livelist_allocs(metaslab_verify_t *mv, uint64_t txg,
(u_longlong_t)DVA_GET_ASIZE(&found->svb_dva),
(u_longlong_t)found->svb_allocated_txg,
(u_longlong_t)txg);
+ corruption_found = B_TRUE;
}
}
}
@@ -426,6 +429,7 @@ metaslab_spacemap_validation_cb(space_map_entry_t *sme, void *arg)
(u_longlong_t)txg, (u_longlong_t)offset,
(u_longlong_t)size, (u_longlong_t)mv->mv_vdid,
(u_longlong_t)mv->mv_msid);
+ corruption_found = B_TRUE;
} else {
zfs_range_tree_add(mv->mv_allocated,
offset, size);
@@ -439,6 +443,7 @@ metaslab_spacemap_validation_cb(space_map_entry_t *sme, void *arg)
(u_longlong_t)txg, (u_longlong_t)offset,
(u_longlong_t)size, (u_longlong_t)mv->mv_vdid,
(u_longlong_t)mv->mv_msid);
+ corruption_found = B_TRUE;
} else {
zfs_range_tree_remove(mv->mv_allocated,
offset, size);
@@ -526,6 +531,7 @@ mv_populate_livelist_allocs(metaslab_verify_t *mv, sublivelist_verify_t *sv)
(u_longlong_t)DVA_GET_VDEV(&svb->svb_dva),
(u_longlong_t)DVA_GET_OFFSET(&svb->svb_dva),
(u_longlong_t)DVA_GET_ASIZE(&svb->svb_dva));
+ corruption_found = B_TRUE;
continue;
}
@@ -542,6 +548,7 @@ mv_populate_livelist_allocs(metaslab_verify_t *mv, sublivelist_verify_t *sv)
(u_longlong_t)DVA_GET_VDEV(&svb->svb_dva),
(u_longlong_t)DVA_GET_OFFSET(&svb->svb_dva),
(u_longlong_t)DVA_GET_ASIZE(&svb->svb_dva));
+ corruption_found = B_TRUE;
continue;
}
@@ -655,6 +662,7 @@ livelist_metaslab_validate(spa_t *spa)
}
(void) printf("ERROR: Found livelist blocks marked as allocated "
"for indirect vdevs:\n");
+ corruption_found = B_TRUE;
zfs_btree_index_t *where = NULL;
sublivelist_verify_block_t *svb;
@@ -827,7 +835,7 @@ usage(void)
(void) fprintf(stderr, "Specify an option more than once (e.g. -bb) "
"to make only that option verbose\n");
(void) fprintf(stderr, "Default is to dump everything non-verbosely\n");
- zdb_exit(1);
+ zdb_exit(2);
}
static void
@@ -892,9 +900,9 @@ dump_packed_nvlist(objset_t *os, uint64_t object, void *data, size_t size)
size_t nvsize = *(uint64_t *)data;
char *packed = umem_alloc(nvsize, UMEM_NOFAIL);
- VERIFY(0 == dmu_read(os, object, 0, nvsize, packed, DMU_READ_PREFETCH));
+ VERIFY0(dmu_read(os, object, 0, nvsize, packed, DMU_READ_PREFETCH));
- VERIFY(nvlist_unpack(packed, nvsize, &nv, 0) == 0);
+ VERIFY0(nvlist_unpack(packed, nvsize, &nv, 0));
umem_free(packed, nvsize);
@@ -1455,8 +1463,8 @@ get_obsolete_refcount(vdev_t *vd)
refcount++;
}
} else {
- ASSERT3P(vd->vdev_obsolete_sm, ==, NULL);
- ASSERT3U(obsolete_sm_object, ==, 0);
+ ASSERT0P(vd->vdev_obsolete_sm);
+ ASSERT0(obsolete_sm_object);
}
for (unsigned c = 0; c < vd->vdev_children; c++) {
refcount += get_obsolete_refcount(vd->vdev_child[c]);
@@ -1792,7 +1800,7 @@ print_vdev_indirect(vdev_t *vd)
vdev_indirect_births_t *vib = vd->vdev_indirect_births;
if (vim == NULL) {
- ASSERT3P(vib, ==, NULL);
+ ASSERT0P(vib);
return;
}
@@ -2043,10 +2051,10 @@ dump_ddt_object(ddt_t *ddt, ddt_type_t type, ddt_class_t class)
if (error == ENOENT)
return;
- ASSERT(error == 0);
+ ASSERT0(error);
error = ddt_object_count(ddt, type, class, &count);
- ASSERT(error == 0);
+ ASSERT0(error);
if (count == 0)
return;
@@ -2583,19 +2591,17 @@ snprintf_blkptr_compact(char *blkbuf, size_t buflen, const blkptr_t *bp,
}
}
-static void
+static u_longlong_t
print_indirect(spa_t *spa, blkptr_t *bp, const zbookmark_phys_t *zb,
const dnode_phys_t *dnp)
{
char blkbuf[BP_SPRINTF_LEN];
+ u_longlong_t offset;
int l;
- if (!BP_IS_EMBEDDED(bp)) {
- ASSERT3U(BP_GET_TYPE(bp), ==, dnp->dn_type);
- ASSERT3U(BP_GET_LEVEL(bp), ==, zb->zb_level);
- }
+ offset = (u_longlong_t)blkid2offset(dnp, bp, zb);
- (void) printf("%16llx ", (u_longlong_t)blkid2offset(dnp, bp, zb));
+ (void) printf("%16llx ", offset);
ASSERT(zb->zb_level >= 0);
@@ -2610,19 +2616,38 @@ print_indirect(spa_t *spa, blkptr_t *bp, const zbookmark_phys_t *zb,
snprintf_blkptr_compact(blkbuf, sizeof (blkbuf), bp, B_FALSE);
if (dump_opt['Z'] && BP_GET_COMPRESS(bp) == ZIO_COMPRESS_ZSTD)
snprintf_zstd_header(spa, blkbuf, sizeof (blkbuf), bp);
- (void) printf("%s\n", blkbuf);
+ (void) printf("%s", blkbuf);
+
+ if (!BP_IS_EMBEDDED(bp)) {
+ if (BP_GET_TYPE(bp) != dnp->dn_type) {
+ (void) printf(" (ERROR: Block pointer type "
+ "(%llu) does not match dnode type (%hhu))",
+ BP_GET_TYPE(bp), dnp->dn_type);
+ corruption_found = B_TRUE;
+ }
+ if (BP_GET_LEVEL(bp) != zb->zb_level) {
+ (void) printf(" (ERROR: Block pointer level "
+ "(%llu) does not match bookmark level (%ld))",
+ BP_GET_LEVEL(bp), zb->zb_level);
+ corruption_found = B_TRUE;
+ }
+ }
+ (void) printf("\n");
+
+ return (offset);
}
static int
visit_indirect(spa_t *spa, const dnode_phys_t *dnp,
blkptr_t *bp, const zbookmark_phys_t *zb)
{
+ u_longlong_t offset;
int err = 0;
if (BP_GET_BIRTH(bp) == 0)
return (0);
- print_indirect(spa, bp, zb, dnp);
+ offset = print_indirect(spa, bp, zb, dnp);
if (BP_GET_LEVEL(bp) > 0 && !BP_IS_HOLE(bp)) {
arc_flags_t flags = ARC_FLAG_WAIT;
@@ -2652,8 +2677,15 @@ visit_indirect(spa_t *spa, const dnode_phys_t *dnp,
break;
fill += BP_GET_FILL(cbp);
}
- if (!err)
- ASSERT3U(fill, ==, BP_GET_FILL(bp));
+ if (!err) {
+ if (fill != BP_GET_FILL(bp)) {
+ (void) printf("%16llx: Block pointer "
+ "fill (%llu) does not match calculated "
+ "value (%lu)\n", offset, BP_GET_FILL(bp),
+ fill);
+ corruption_found = B_TRUE;
+ }
+ }
arc_buf_destroy(buf, &buf);
}
@@ -2909,6 +2941,7 @@ dump_full_bpobj(bpobj_t *bpo, const char *name, int indent)
(void) printf("ERROR %u while trying to open "
"subobj id %llu\n",
error, (u_longlong_t)subobj);
+ corruption_found = B_TRUE;
continue;
}
dump_full_bpobj(&subbpo, "subobj", indent + 1);
@@ -3088,6 +3121,7 @@ bpobj_count_refd(bpobj_t *bpo)
(void) printf("ERROR %u while trying to open "
"subobj id %llu\n",
error, (u_longlong_t)subobj);
+ corruption_found = B_TRUE;
continue;
}
bpobj_count_refd(&subbpo);
@@ -3109,7 +3143,7 @@ dsl_deadlist_entry_count_refd(void *arg, dsl_deadlist_entry_t *dle)
static int
dsl_deadlist_entry_dump(void *arg, dsl_deadlist_entry_t *dle)
{
- ASSERT(arg == NULL);
+ ASSERT0P(arg);
if (dump_opt['d'] >= 5) {
char buf[128];
(void) snprintf(buf, sizeof (buf),
@@ -3347,7 +3381,7 @@ open_objset(const char *path, const void *tag, objset_t **osp)
uint64_t sa_attrs = 0;
uint64_t version = 0;
- VERIFY3P(sa_os, ==, NULL);
+ VERIFY0P(sa_os);
/*
* We can't own an objset if it's redacted. Therefore, we do this
@@ -3520,8 +3554,8 @@ dump_uidgid(objset_t *os, uint64_t uid, uint64_t gid)
uint64_t fuid_obj;
/* first find the fuid object. It lives in the master node */
- VERIFY(zap_lookup(os, MASTER_NODE_OBJ, ZFS_FUID_TABLES,
- 8, 1, &fuid_obj) == 0);
+ VERIFY0(zap_lookup(os, MASTER_NODE_OBJ, ZFS_FUID_TABLES,
+ 8, 1, &fuid_obj));
zfs_fuid_avl_tree_create(&idx_tree, &domain_tree);
(void) zfs_fuid_table_load(os, fuid_obj,
&idx_tree, &domain_tree);
@@ -7016,7 +7050,7 @@ deleted_livelists_count_blocks(spa_t *spa, zdb_cb_t *zbc)
static void
dump_livelist_cb(dsl_deadlist_t *ll, void *arg)
{
- ASSERT3P(arg, ==, NULL);
+ ASSERT0P(arg);
global_feature_count[SPA_FEATURE_LIVELIST]++;
dump_blkptr_list(ll, "Deleted Livelist");
dsl_deadlist_iterate(ll, sublivelist_verify_lightweight, NULL);
@@ -7913,7 +7947,7 @@ verify_checkpoint_vdev_spacemaps(spa_t *checkpoint, spa_t *current)
for (uint64_t c = ckpoint_rvd->vdev_children;
c < current_rvd->vdev_children; c++) {
vdev_t *current_vd = current_rvd->vdev_child[c];
- VERIFY3P(current_vd->vdev_checkpoint_sm, ==, NULL);
+ VERIFY0P(current_vd->vdev_checkpoint_sm);
}
}
@@ -9634,7 +9668,7 @@ main(int argc, char **argv)
} else if (objset_str && !zdb_numeric(objset_str + 1) &&
dump_opt['N']) {
printf("Supply a numeric objset ID with -N\n");
- error = 1;
+ error = 2;
goto fini;
}
} else {
@@ -9743,7 +9777,7 @@ main(int argc, char **argv)
if (error == 0) {
if (dump_opt['k'] && (target_is_spa || dump_opt['R'])) {
ASSERT(checkpoint_pool != NULL);
- ASSERT(checkpoint_target == NULL);
+ ASSERT0P(checkpoint_target);
error = spa_open(checkpoint_pool, &spa, FTAG);
if (error != 0) {
@@ -9936,5 +9970,8 @@ fini:
if (kernel_init_done)
kernel_fini();
+ if (corruption_found && error == 0)
+ error = 3;
+
return (error);
}
diff --git a/sys/contrib/openzfs/cmd/zed/zed.d/zed-functions.sh b/sys/contrib/openzfs/cmd/zed/zed.d/zed-functions.sh
index 6e00f153be1c..78d8f658ddd8 100644
--- a/sys/contrib/openzfs/cmd/zed/zed.d/zed-functions.sh
+++ b/sys/contrib/openzfs/cmd/zed/zed.d/zed-functions.sh
@@ -441,8 +441,9 @@ zed_notify_slack_webhook()
"${pathname}")"
# Construct the JSON message for posting.
+ # shellcheck disable=SC2016
#
- msg_json="$(printf '{"text": "*%s*\\n%s"}' "${subject}" "${msg_body}" )"
+ msg_json="$(printf '{"text": "*%s*\\n```%s```"}' "${subject}" "${msg_body}" )"
# Send the POST request and check for errors.
#
diff --git a/sys/contrib/openzfs/cmd/zfs/zfs_main.c b/sys/contrib/openzfs/cmd/zfs/zfs_main.c
index 363bb6da74ec..235f011af953 100644
--- a/sys/contrib/openzfs/cmd/zfs/zfs_main.c
+++ b/sys/contrib/openzfs/cmd/zfs/zfs_main.c
@@ -923,7 +923,7 @@ zfs_do_clone(int argc, char **argv)
return (!!ret);
usage:
- ASSERT3P(zhp, ==, NULL);
+ ASSERT0P(zhp);
nvlist_free(props);
usage(B_FALSE);
return (-1);
diff --git a/sys/contrib/openzfs/cmd/zhack.c b/sys/contrib/openzfs/cmd/zhack.c
index 8244bc83fa0d..2bd3051dce7b 100644
--- a/sys/contrib/openzfs/cmd/zhack.c
+++ b/sys/contrib/openzfs/cmd/zhack.c
@@ -162,9 +162,9 @@ zhack_import(char *target, boolean_t readonly)
props = NULL;
if (readonly) {
- VERIFY(nvlist_alloc(&props, NV_UNIQUE_NAME, 0) == 0);
- VERIFY(nvlist_add_uint64(props,
- zpool_prop_to_name(ZPOOL_PROP_READONLY), 1) == 0);
+ VERIFY0(nvlist_alloc(&props, NV_UNIQUE_NAME, 0));
+ VERIFY0(nvlist_add_uint64(props,
+ zpool_prop_to_name(ZPOOL_PROP_READONLY), 1));
}
zfeature_checks_disable = B_TRUE;
@@ -218,8 +218,8 @@ dump_obj(objset_t *os, uint64_t obj, const char *name)
} else {
ASSERT(za->za_integer_length == 1);
char val[1024];
- VERIFY(zap_lookup(os, obj, za->za_name,
- 1, sizeof (val), val) == 0);
+ VERIFY0(zap_lookup(os, obj, za->za_name,
+ 1, sizeof (val), val));
(void) printf("\t%s = %s\n", za->za_name, val);
}
}
diff --git a/sys/contrib/openzfs/cmd/zilstat.in b/sys/contrib/openzfs/cmd/zilstat.in
index 4140398bf4a3..d01db9b0914b 100755
--- a/sys/contrib/openzfs/cmd/zilstat.in
+++ b/sys/contrib/openzfs/cmd/zilstat.in
@@ -47,6 +47,7 @@ cols = {
"cec": [5, 1000, "zil_commit_error_count"],
"csc": [5, 1000, "zil_commit_stall_count"],
"cSc": [5, 1000, "zil_commit_suspend_count"],
+ "cCc": [5, 1000, "zil_commit_crash_count"],
"ic": [5, 1000, "zil_itx_count"],
"iic": [5, 1000, "zil_itx_indirect_count"],
"iib": [5, 1024, "zil_itx_indirect_bytes"],
diff --git a/sys/contrib/openzfs/cmd/zpool/zpool_main.c b/sys/contrib/openzfs/cmd/zpool/zpool_main.c
index d401e087916b..237e558da65b 100644
--- a/sys/contrib/openzfs/cmd/zpool/zpool_main.c
+++ b/sys/contrib/openzfs/cmd/zpool/zpool_main.c
@@ -752,10 +752,11 @@ usage(boolean_t requested)
}
/*
- * zpool initialize [-c | -s | -u] [-w] <pool> [<vdev> ...]
+ * zpool initialize [-c | -s | -u] [-w] <-a | pool> [<vdev> ...]
* Initialize all unused blocks in the specified vdevs, or all vdevs in the pool
* if none specified.
*
+ * -a Use all pools.
* -c Cancel. Ends active initializing.
* -s Suspend. Initializing can then be restarted with no flags.
* -u Uninitialize. Clears initialization state.
@@ -776,7 +777,7 @@ zpool_do_initialize(int argc, char **argv)
{"suspend", no_argument, NULL, 's'},
{"uninit", no_argument, NULL, 'u'},
{"wait", no_argument, NULL, 'w'},
- {"all", no_argument, NULL, 'a'},
+ {"all", no_argument, NULL, 'a'},
{0, 0, 0, 0}
};
@@ -8446,8 +8447,9 @@ date_string_to_sec(const char *timestr, boolean_t rounding)
}
/*
- * zpool scrub [-e | -s | -p | -C | -E | -S] [-w] <pool> ...
+ * zpool scrub [-e | -s | -p | -C | -E | -S] [-w] [-a | <pool> ...]
*
+ * -a Scrub all pools.
* -e Only scrub blocks in the error log.
* -E End date of scrub.
* -S Start date of scrub.
@@ -8621,8 +8623,9 @@ zpool_do_resilver(int argc, char **argv)
}
/*
- * zpool trim [-d] [-r <rate>] [-c | -s] <pool> [<device> ...]
+ * zpool trim [-d] [-r <rate>] [-c | -s] <-a | pool> [<device> ...]
*
+ * -a Trim all pools.
* -c Cancel. Ends any in-progress trim.
* -d Secure trim. Requires kernel and device support.
* -r <rate> Sets the TRIM rate in bytes (per second). Supports
@@ -12374,7 +12377,7 @@ zpool_do_events_next(ev_opts_t *opts)
nvlist_free(nvl);
}
- VERIFY(0 == close(zevent_fd));
+ VERIFY0(close(zevent_fd));
return (ret);
}
diff --git a/sys/contrib/openzfs/cmd/ztest.c b/sys/contrib/openzfs/cmd/ztest.c
index 2e88ae3e7994..89752dcb0f0f 100644
--- a/sys/contrib/openzfs/cmd/ztest.c
+++ b/sys/contrib/openzfs/cmd/ztest.c
@@ -273,7 +273,6 @@ extern int zfs_compressed_arc_enabled;
extern int zfs_abd_scatter_enabled;
extern uint_t dmu_object_alloc_chunk_shift;
extern boolean_t zfs_force_some_double_word_sm_entries;
-extern unsigned long zio_decompress_fail_fraction;
extern unsigned long zfs_reconstruct_indirect_damage_fraction;
extern uint64_t raidz_expand_max_reflow_bytes;
extern uint_t raidz_expand_pause_point;
@@ -829,8 +828,8 @@ static char *short_opts = NULL;
static void
init_options(void)
{
- ASSERT3P(long_opts, ==, NULL);
- ASSERT3P(short_opts, ==, NULL);
+ ASSERT0P(long_opts);
+ ASSERT0P(short_opts);
int count = sizeof (option_table) / sizeof (option_table[0]);
long_opts = umem_alloc(sizeof (struct option) * count, UMEM_NOFAIL);
@@ -1686,7 +1685,7 @@ ztest_rll_init(rll_t *rll)
static void
ztest_rll_destroy(rll_t *rll)
{
- ASSERT3P(rll->rll_writer, ==, NULL);
+ ASSERT0P(rll->rll_writer);
ASSERT0(rll->rll_readers);
mutex_destroy(&rll->rll_lock);
cv_destroy(&rll->rll_cv);
@@ -1720,7 +1719,7 @@ ztest_rll_unlock(rll_t *rll)
rll->rll_writer = NULL;
} else {
ASSERT3S(rll->rll_readers, >, 0);
- ASSERT3P(rll->rll_writer, ==, NULL);
+ ASSERT0P(rll->rll_writer);
rll->rll_readers--;
}
@@ -1996,7 +1995,7 @@ ztest_log_write(ztest_ds_t *zd, dmu_tx_t *tx, lr_write_t *lr)
dmu_read(zd->zd_os, lr->lr_foid, lr->lr_offset, lr->lr_length,
((lr_write_t *)&itx->itx_lr) + 1, DMU_READ_NO_PREFETCH |
DMU_KEEP_CACHING) != 0) {
- zil_itx_destroy(itx);
+ zil_itx_destroy(itx, 0);
itx = zil_itx_create(TX_WRITE, sizeof (*lr));
write_state = WR_NEED_COPY;
}
@@ -2278,8 +2277,8 @@ ztest_replay_write(void *arg1, void *arg2, boolean_t byteswap)
ztest_block_tag_t rbt;
- VERIFY(dmu_read(os, lr->lr_foid, offset,
- sizeof (rbt), &rbt, flags) == 0);
+ VERIFY0(dmu_read(os, lr->lr_foid, offset,
+ sizeof (rbt), &rbt, flags));
if (rbt.bt_magic == BT_MAGIC) {
ztest_bt_verify(&rbt, os, lr->lr_foid, 0,
offset, gen, txg, crtxg);
@@ -2966,7 +2965,7 @@ ztest_zil_commit(ztest_ds_t *zd, uint64_t id)
(void) pthread_rwlock_rdlock(&zd->zd_zilog_lock);
- zil_commit(zilog, ztest_random(ZTEST_OBJECTS));
+ VERIFY0(zil_commit(zilog, ztest_random(ZTEST_OBJECTS)));
/*
* Remember the committed values in zd, which is in parent/child
@@ -4007,7 +4006,7 @@ raidz_scratch_verify(void)
* requested by user, but scratch object was not created.
*/
case RRSS_SCRATCH_NOT_IN_USE:
- ASSERT3U(offset, ==, 0);
+ ASSERT0(offset);
break;
/*
@@ -5537,8 +5536,8 @@ ztest_dmu_read_write_zcopy(ztest_ds_t *zd, uint64_t id)
}
if (i == 1) {
- VERIFY(dmu_buf_hold(os, bigobj, off,
- FTAG, &dbt, DMU_READ_NO_PREFETCH) == 0);
+ VERIFY0(dmu_buf_hold(os, bigobj, off,
+ FTAG, &dbt, DMU_READ_NO_PREFETCH));
}
if (i != 5 || chunksize < (SPA_MINBLOCKSIZE * 2)) {
VERIFY0(dmu_assign_arcbuf_by_dbuf(bonus_db,
@@ -7937,7 +7936,7 @@ ztest_freeze(void)
*/
while (BP_IS_HOLE(&zd->zd_zilog->zl_header->zh_log)) {
ztest_dmu_object_alloc_free(zd, 0);
- zil_commit(zd->zd_zilog, 0);
+ VERIFY0(zil_commit(zd->zd_zilog, 0));
}
txg_wait_synced(spa_get_dsl(spa), 0);
@@ -7979,7 +7978,7 @@ ztest_freeze(void)
/*
* Commit all of the changes we just generated.
*/
- zil_commit(zd->zd_zilog, 0);
+ VERIFY0(zil_commit(zd->zd_zilog, 0));
txg_wait_synced(spa_get_dsl(spa), 0);
/*
@@ -8979,7 +8978,7 @@ main(int argc, char **argv)
exit(EXIT_FAILURE);
} else {
/* children should not be spawned if setting gvars fails */
- VERIFY3S(err, ==, 0);
+ VERIFY0(err);
}
/* Override location of zpool.cache */
diff --git a/sys/contrib/openzfs/config/kernel-mkdir.m4 b/sys/contrib/openzfs/config/kernel-mkdir.m4
index c1aebc387abe..78b32447c593 100644
--- a/sys/contrib/openzfs/config/kernel-mkdir.m4
+++ b/sys/contrib/openzfs/config/kernel-mkdir.m4
@@ -84,6 +84,8 @@ AC_DEFUN([ZFS_AC_KERNEL_MKDIR], [
AC_DEFINE(HAVE_IOPS_MKDIR_DENTRY, 1,
[iops->mkdir() returns struct dentry*])
],[
+ AC_MSG_RESULT(no)
+
dnl #
dnl # 6.3 API change
dnl # mkdir() takes struct mnt_idmap * as the first arg
diff --git a/sys/contrib/openzfs/config/toolchain-simd.m4 b/sys/contrib/openzfs/config/toolchain-simd.m4
index 061576fd94e3..f18c91007cde 100644
--- a/sys/contrib/openzfs/config/toolchain-simd.m4
+++ b/sys/contrib/openzfs/config/toolchain-simd.m4
@@ -24,6 +24,8 @@ AC_DEFUN([ZFS_AC_CONFIG_ALWAYS_TOOLCHAIN_SIMD], [
ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_AES
ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_PCLMULQDQ
ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_MOVBE
+ ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_VAES
+ ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_VPCLMULQDQ
ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_XSAVE
ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_XSAVEOPT
ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_XSAVES
@@ -38,9 +40,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_SSE], [
AC_MSG_CHECKING([whether host toolchain supports SSE])
AC_LINK_IFELSE([AC_LANG_SOURCE([[
- void main()
+ int main()
{
__asm__ __volatile__("xorps %xmm0, %xmm1");
+ return (0);
}
]])], [
AC_DEFINE([HAVE_SSE], 1, [Define if host toolchain supports SSE])
@@ -57,9 +60,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_SSE2], [
AC_MSG_CHECKING([whether host toolchain supports SSE2])
AC_LINK_IFELSE([AC_LANG_SOURCE([[
- void main()
+ int main()
{
__asm__ __volatile__("pxor %xmm0, %xmm1");
+ return (0);
}
]])], [
AC_DEFINE([HAVE_SSE2], 1, [Define if host toolchain supports SSE2])
@@ -76,10 +80,11 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_SSE3], [
AC_MSG_CHECKING([whether host toolchain supports SSE3])
AC_LINK_IFELSE([AC_LANG_SOURCE([[
- void main()
+ int main()
{
char v[16];
__asm__ __volatile__("lddqu %0,%%xmm0" :: "m"(v[0]));
+ return (0);
}
]])], [
AC_DEFINE([HAVE_SSE3], 1, [Define if host toolchain supports SSE3])
@@ -96,9 +101,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_SSSE3], [
AC_MSG_CHECKING([whether host toolchain supports SSSE3])
AC_LINK_IFELSE([AC_LANG_SOURCE([[
- void main()
+ int main()
{
__asm__ __volatile__("pshufb %xmm0,%xmm1");
+ return (0);
}
]])], [
AC_DEFINE([HAVE_SSSE3], 1, [Define if host toolchain supports SSSE3])
@@ -115,9 +121,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_SSE4_1], [
AC_MSG_CHECKING([whether host toolchain supports SSE4.1])
AC_LINK_IFELSE([AC_LANG_SOURCE([[
- void main()
+ int main()
{
__asm__ __volatile__("pmaxsb %xmm0,%xmm1");
+ return (0);
}
]])], [
AC_DEFINE([HAVE_SSE4_1], 1, [Define if host toolchain supports SSE4.1])
@@ -134,9 +141,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_SSE4_2], [
AC_MSG_CHECKING([whether host toolchain supports SSE4.2])
AC_LINK_IFELSE([AC_LANG_SOURCE([[
- void main()
+ int main()
{
__asm__ __volatile__("pcmpgtq %xmm0, %xmm1");
+ return (0);
}
]])], [
AC_DEFINE([HAVE_SSE4_2], 1, [Define if host toolchain supports SSE4.2])
@@ -153,10 +161,11 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_AVX], [
AC_MSG_CHECKING([whether host toolchain supports AVX])
AC_LINK_IFELSE([AC_LANG_SOURCE([[
- void main()
+ int main()
{
char v[32];
__asm__ __volatile__("vmovdqa %0,%%ymm0" :: "m"(v[0]));
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -174,9 +183,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_AVX2], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
__asm__ __volatile__("vpshufb %ymm0,%ymm1,%ymm2");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -194,9 +204,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_AVX512F], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
__asm__ __volatile__("vpandd %zmm0,%zmm1,%zmm2");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -214,9 +225,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_AVX512CD], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
__asm__ __volatile__("vplzcntd %zmm0,%zmm1");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -234,9 +246,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_AVX512DQ], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
__asm__ __volatile__("vandpd %zmm0,%zmm1,%zmm2");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -254,9 +267,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_AVX512BW], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
__asm__ __volatile__("vpshufb %zmm0,%zmm1,%zmm2");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -274,9 +288,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_AVX512IFMA], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
__asm__ __volatile__("vpmadd52luq %zmm0,%zmm1,%zmm2");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -294,9 +309,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_AVX512VBMI], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
__asm__ __volatile__("vpermb %zmm0,%zmm1,%zmm2");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -314,9 +330,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_AVX512PF], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
__asm__ __volatile__("vgatherpf0dps (%rsi,%zmm0,4){%k1}");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -334,9 +351,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_AVX512ER], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
__asm__ __volatile__("vexp2pd %zmm0,%zmm1");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -354,9 +372,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_AVX512VL], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
__asm__ __volatile__("vpabsq %zmm0,%zmm1");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -374,9 +393,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_AES], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
__asm__ __volatile__("aesenc %xmm0, %xmm1");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -394,9 +414,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_PCLMULQDQ], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
__asm__ __volatile__("pclmulqdq %0, %%xmm0, %%xmm1" :: "i"(0));
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -414,9 +435,10 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_MOVBE], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
__asm__ __volatile__("movbe 0(%eax), %eax");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -427,6 +449,48 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_MOVBE], [
])
dnl #
+dnl # ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_VAES
+dnl #
+AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_VAES], [
+ AC_MSG_CHECKING([whether host toolchain supports VAES])
+
+ AC_LINK_IFELSE([AC_LANG_SOURCE([
+ [
+ int main()
+ {
+ __asm__ __volatile__("vaesenc %ymm0, %ymm1, %ymm0");
+ return (0);
+ }
+ ]])], [
+ AC_MSG_RESULT([yes])
+ AC_DEFINE([HAVE_VAES], 1, [Define if host toolchain supports VAES])
+ ], [
+ AC_MSG_RESULT([no])
+ ])
+])
+
+dnl #
+dnl # ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_VPCLMULQDQ
+dnl #
+AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_VPCLMULQDQ], [
+ AC_MSG_CHECKING([whether host toolchain supports VPCLMULQDQ])
+
+ AC_LINK_IFELSE([AC_LANG_SOURCE([
+ [
+ int main()
+ {
+ __asm__ __volatile__("vpclmulqdq %0, %%ymm4, %%ymm3, %%ymm5" :: "i"(0));
+ return (0);
+ }
+ ]])], [
+ AC_MSG_RESULT([yes])
+ AC_DEFINE([HAVE_VPCLMULQDQ], 1, [Define if host toolchain supports VPCLMULQDQ])
+ ], [
+ AC_MSG_RESULT([no])
+ ])
+])
+
+dnl #
dnl # ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_XSAVE
dnl #
AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_XSAVE], [
@@ -434,10 +498,11 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_XSAVE], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
char b[4096] __attribute__ ((aligned (64)));
__asm__ __volatile__("xsave %[b]\n" : : [b] "m" (*b) : "memory");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -455,10 +520,11 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_XSAVEOPT], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
char b[4096] __attribute__ ((aligned (64)));
__asm__ __volatile__("xsaveopt %[b]\n" : : [b] "m" (*b) : "memory");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
@@ -476,10 +542,11 @@ AC_DEFUN([ZFS_AC_CONFIG_TOOLCHAIN_CAN_BUILD_XSAVES], [
AC_LINK_IFELSE([AC_LANG_SOURCE([
[
- void main()
+ int main()
{
char b[4096] __attribute__ ((aligned (64)));
__asm__ __volatile__("xsaves %[b]\n" : : [b] "m" (*b) : "memory");
+ return (0);
}
]])], [
AC_MSG_RESULT([yes])
diff --git a/sys/contrib/openzfs/contrib/debian/control b/sys/contrib/openzfs/contrib/debian/control
index 96a2bdd88665..c5358dedc0fd 100644
--- a/sys/contrib/openzfs/contrib/debian/control
+++ b/sys/contrib/openzfs/contrib/debian/control
@@ -100,8 +100,8 @@ 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, libzfs4linux, libzfs6linux
-Replaces: libzfs2, libzfs4, libzfs4linux, libzfs6linux
+Breaks: libzfs2, libzfs4, libzfs4linux, libzfs6linux, openzfs-libzfs4
+Replaces: libzfs2, libzfs4, libzfs4linux, libzfs6linux, openzfs-libzfs4
Conflicts: libzfs6linux
Description: OpenZFS filesystem library for Linux - general support
OpenZFS is a storage platform that encompasses the functionality of
@@ -128,8 +128,8 @@ Package: openzfs-libzpool6
Section: contrib/libs
Architecture: linux-any
Depends: ${misc:Depends}, ${shlibs:Depends}
-Breaks: libzpool2, libzpool5, libzpool5linux, libzpool6linux
-Replaces: libzpool2, libzpool5, libzpool5linux, libzpool6linux
+Breaks: libzpool2, libzpool5, libzpool6linux
+Replaces: libzpool2, libzpool5, libzpool6linux
Conflicts: libzpool6linux
Description: OpenZFS pool library for Linux
OpenZFS is a storage platform that encompasses the functionality of
diff --git a/sys/contrib/openzfs/contrib/icp/gcm-simd/boringssl/LICENSE b/sys/contrib/openzfs/contrib/icp/gcm-simd/boringssl/LICENSE
new file mode 100644
index 000000000000..04c03a37e0cb
--- /dev/null
+++ b/sys/contrib/openzfs/contrib/icp/gcm-simd/boringssl/LICENSE
@@ -0,0 +1,253 @@
+BoringSSL is a fork of OpenSSL. As such, large parts of it fall under OpenSSL
+licensing. Files that are completely new have a Google copyright and an ISC
+license. This license is reproduced at the bottom of this file.
+
+Contributors to BoringSSL are required to follow the CLA rules for Chromium:
+https://cla.developers.google.com/clas
+
+Files in third_party/ have their own licenses, as described therein. The MIT
+license, for third_party/fiat, which, unlike other third_party directories, is
+compiled into non-test libraries, is included below.
+
+The OpenSSL toolkit stays under a dual license, i.e. both the conditions of the
+OpenSSL License and the original SSLeay license apply to the toolkit. See below
+for the actual license texts. Actually both licenses are BSD-style Open Source
+licenses. In case of any license issues related to OpenSSL please contact
+openssl-core@openssl.org.
+
+The following are Google-internal bug numbers where explicit permission from
+some authors is recorded for use of their work. (This is purely for our own
+record keeping.)
+ 27287199
+ 27287880
+ 27287883
+ 263291445
+
+
+ OpenSSL License
+ ---------------
+
+/* ====================================================================
+ * Copyright (c) 1998-2011 The OpenSSL Project. All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in
+ * the documentation and/or other materials provided with the
+ * distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ * software must display the following acknowledgment:
+ * "This product includes software developed by the OpenSSL Project
+ * for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ * endorse or promote products derived from this software without
+ * prior written permission. For written permission, please contact
+ * openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ * nor may "OpenSSL" appear in their names without prior written
+ * permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ * acknowledgment:
+ * "This product includes software developed by the OpenSSL Project
+ * for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com). This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+ Original SSLeay License
+ -----------------------
+
+/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
+ * All rights reserved.
+ *
+ * This package is an SSL implementation written
+ * by Eric Young (eay@cryptsoft.com).
+ * The implementation was written so as to conform with Netscapes SSL.
+ *
+ * This library is free for commercial and non-commercial use as long as
+ * the following conditions are aheared to. The following conditions
+ * apply to all code found in this distribution, be it the RC4, RSA,
+ * lhash, DES, etc., code; not just the SSL code. The SSL documentation
+ * included with this distribution is covered by the same copyright terms
+ * except that the holder is Tim Hudson (tjh@cryptsoft.com).
+ *
+ * Copyright remains Eric Young's, and as such any Copyright notices in
+ * the code are not to be removed.
+ * If this package is used in a product, Eric Young should be given attribution
+ * as the author of the parts of the library used.
+ * This can be in the form of a textual message at program startup or
+ * in documentation (online or textual) provided with the package.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ * must display the following acknowledgement:
+ * "This product includes cryptographic software written by
+ * Eric Young (eay@cryptsoft.com)"
+ * The word 'cryptographic' can be left out if the rouines from the library
+ * being used are not cryptographic related :-).
+ * 4. If you include any Windows specific code (or a derivative thereof) from
+ * the apps directory (application code) you must include an acknowledgement:
+ * "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * The licence and distribution terms for any publically available version or
+ * derivative of this code cannot be changed. i.e. this code cannot simply be
+ * copied and put under another distribution licence
+ * [including the GNU Public Licence.]
+ */
+
+
+ISC license used for completely new code in BoringSSL:
+
+/* Copyright 2015 The BoringSSL Authors
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY
+ * SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION
+ * OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN
+ * CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */
+
+
+The code in third_party/fiat carries the MIT license:
+
+Copyright (c) 2015-2016 the fiat-crypto authors (see
+https://github.com/mit-plv/fiat-crypto/blob/master/AUTHORS).
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+
+Licenses for support code
+-------------------------
+
+Parts of the TLS test suite are under the Go license. This code is not included
+in BoringSSL (i.e. libcrypto and libssl) when compiled, however, so
+distributing code linked against BoringSSL does not trigger this license:
+
+Copyright (c) 2009 The Go Authors. All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+ * Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+ * Redistributions in binary form must reproduce the above
+copyright notice, this list of conditions and the following disclaimer
+in the documentation and/or other materials provided with the
+distribution.
+ * Neither the name of Google Inc. nor the names of its
+contributors may be used to endorse or promote products derived from
+this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+
+BoringSSL uses the Chromium test infrastructure to run a continuous build,
+trybots etc. The scripts which manage this, and the script for generating build
+metadata, are under the Chromium license. Distributing code linked against
+BoringSSL does not trigger this license.
+
+Copyright 2015 The Chromium Authors. All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+ * Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+ * Redistributions in binary form must reproduce the above
+copyright notice, this list of conditions and the following disclaimer
+in the documentation and/or other materials provided with the
+distribution.
+ * Neither the name of Google Inc. nor the names of its
+contributors may be used to endorse or promote products derived from
+this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/sys/contrib/openzfs/contrib/icp/gcm-simd/boringssl/README b/sys/contrib/openzfs/contrib/icp/gcm-simd/boringssl/README
new file mode 100644
index 000000000000..aa6fb6d477fa
--- /dev/null
+++ b/sys/contrib/openzfs/contrib/icp/gcm-simd/boringssl/README
@@ -0,0 +1,11 @@
+This directory contains the original BoringSSL [1] GCM x86-64 assembly
+files [2].
+
+The assembler files where then further modified to fit the ICP conventions.
+
+The main purpose to include these files (and the original ones) here, is to
+serve as a reference if upstream changes need to be applied to the files
+included and modified in the ICP.
+
+[1] https://github.com/google/boringssl
+[2] https://github.com/google/boringssl/blob/d5440dd2c2c500ac2d3bba4afec47a054b4d99ae/gen/bcm/aes-gcm-avx2-x86_64-linux.S
diff --git a/sys/contrib/openzfs/contrib/icp/gcm-simd/boringssl/aes-gcm-avx2-x86_64-linux.S b/sys/contrib/openzfs/contrib/icp/gcm-simd/boringssl/aes-gcm-avx2-x86_64-linux.S
new file mode 100644
index 000000000000..e7327c9de872
--- /dev/null
+++ b/sys/contrib/openzfs/contrib/icp/gcm-simd/boringssl/aes-gcm-avx2-x86_64-linux.S
@@ -0,0 +1,1328 @@
+// SPDX-License-Identifier: Apache-2.0
+// This file is generated from a similarly-named Perl script in the BoringSSL
+// source tree. Do not edit by hand.
+
+#include <openssl/asm_base.h>
+
+#if !defined(OPENSSL_NO_ASM) && defined(OPENSSL_X86_64) && defined(__ELF__)
+.section .rodata
+.align 16
+
+
+.Lbswap_mask:
+.quad 0x08090a0b0c0d0e0f, 0x0001020304050607
+
+
+
+
+
+
+
+
+.Lgfpoly:
+.quad 1, 0xc200000000000000
+
+
+.Lgfpoly_and_internal_carrybit:
+.quad 1, 0xc200000000000001
+
+.align 32
+
+.Lctr_pattern:
+.quad 0, 0
+.quad 1, 0
+.Linc_2blocks:
+.quad 2, 0
+.quad 2, 0
+
+.text
+.globl gcm_init_vpclmulqdq_avx2
+.hidden gcm_init_vpclmulqdq_avx2
+.type gcm_init_vpclmulqdq_avx2,@function
+.align 32
+gcm_init_vpclmulqdq_avx2:
+.cfi_startproc
+
+_CET_ENDBR
+
+
+
+
+
+ vpshufd $0x4e,(%rsi),%xmm3
+
+
+
+
+
+ vpshufd $0xd3,%xmm3,%xmm0
+ vpsrad $31,%xmm0,%xmm0
+ vpaddq %xmm3,%xmm3,%xmm3
+ vpand .Lgfpoly_and_internal_carrybit(%rip),%xmm0,%xmm0
+ vpxor %xmm0,%xmm3,%xmm3
+
+ vbroadcasti128 .Lgfpoly(%rip),%ymm6
+
+
+ vpclmulqdq $0x00,%xmm3,%xmm3,%xmm0
+ vpclmulqdq $0x01,%xmm3,%xmm3,%xmm1
+ vpclmulqdq $0x10,%xmm3,%xmm3,%xmm2
+ vpxor %xmm2,%xmm1,%xmm1
+ vpclmulqdq $0x01,%xmm0,%xmm6,%xmm2
+ vpshufd $0x4e,%xmm0,%xmm0
+ vpxor %xmm0,%xmm1,%xmm1
+ vpxor %xmm2,%xmm1,%xmm1
+ vpclmulqdq $0x11,%xmm3,%xmm3,%xmm5
+ vpclmulqdq $0x01,%xmm1,%xmm6,%xmm0
+ vpshufd $0x4e,%xmm1,%xmm1
+ vpxor %xmm1,%xmm5,%xmm5
+ vpxor %xmm0,%xmm5,%xmm5
+
+
+
+ vinserti128 $1,%xmm3,%ymm5,%ymm3
+ vinserti128 $1,%xmm5,%ymm5,%ymm5
+
+
+ vpclmulqdq $0x00,%ymm5,%ymm3,%ymm0
+ vpclmulqdq $0x01,%ymm5,%ymm3,%ymm1
+ vpclmulqdq $0x10,%ymm5,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpclmulqdq $0x01,%ymm0,%ymm6,%ymm2
+ vpshufd $0x4e,%ymm0,%ymm0
+ vpxor %ymm0,%ymm1,%ymm1
+ vpxor %ymm2,%ymm1,%ymm1
+ vpclmulqdq $0x11,%ymm5,%ymm3,%ymm4
+ vpclmulqdq $0x01,%ymm1,%ymm6,%ymm0
+ vpshufd $0x4e,%ymm1,%ymm1
+ vpxor %ymm1,%ymm4,%ymm4
+ vpxor %ymm0,%ymm4,%ymm4
+
+
+
+ vmovdqu %ymm3,96(%rdi)
+ vmovdqu %ymm4,64(%rdi)
+
+
+
+ vpunpcklqdq %ymm3,%ymm4,%ymm0
+ vpunpckhqdq %ymm3,%ymm4,%ymm1
+ vpxor %ymm1,%ymm0,%ymm0
+ vmovdqu %ymm0,128+32(%rdi)
+
+
+ vpclmulqdq $0x00,%ymm5,%ymm4,%ymm0
+ vpclmulqdq $0x01,%ymm5,%ymm4,%ymm1
+ vpclmulqdq $0x10,%ymm5,%ymm4,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpclmulqdq $0x01,%ymm0,%ymm6,%ymm2
+ vpshufd $0x4e,%ymm0,%ymm0
+ vpxor %ymm0,%ymm1,%ymm1
+ vpxor %ymm2,%ymm1,%ymm1
+ vpclmulqdq $0x11,%ymm5,%ymm4,%ymm3
+ vpclmulqdq $0x01,%ymm1,%ymm6,%ymm0
+ vpshufd $0x4e,%ymm1,%ymm1
+ vpxor %ymm1,%ymm3,%ymm3
+ vpxor %ymm0,%ymm3,%ymm3
+
+ vpclmulqdq $0x00,%ymm5,%ymm3,%ymm0
+ vpclmulqdq $0x01,%ymm5,%ymm3,%ymm1
+ vpclmulqdq $0x10,%ymm5,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpclmulqdq $0x01,%ymm0,%ymm6,%ymm2
+ vpshufd $0x4e,%ymm0,%ymm0
+ vpxor %ymm0,%ymm1,%ymm1
+ vpxor %ymm2,%ymm1,%ymm1
+ vpclmulqdq $0x11,%ymm5,%ymm3,%ymm4
+ vpclmulqdq $0x01,%ymm1,%ymm6,%ymm0
+ vpshufd $0x4e,%ymm1,%ymm1
+ vpxor %ymm1,%ymm4,%ymm4
+ vpxor %ymm0,%ymm4,%ymm4
+
+ vmovdqu %ymm3,32(%rdi)
+ vmovdqu %ymm4,0(%rdi)
+
+
+
+ vpunpcklqdq %ymm3,%ymm4,%ymm0
+ vpunpckhqdq %ymm3,%ymm4,%ymm1
+ vpxor %ymm1,%ymm0,%ymm0
+ vmovdqu %ymm0,128(%rdi)
+
+ vzeroupper
+ ret
+
+.cfi_endproc
+.size gcm_init_vpclmulqdq_avx2, . - gcm_init_vpclmulqdq_avx2
+.globl gcm_gmult_vpclmulqdq_avx2
+.hidden gcm_gmult_vpclmulqdq_avx2
+.type gcm_gmult_vpclmulqdq_avx2,@function
+.align 32
+gcm_gmult_vpclmulqdq_avx2:
+.cfi_startproc
+
+_CET_ENDBR
+
+
+
+ vmovdqu (%rdi),%xmm0
+ vmovdqu .Lbswap_mask(%rip),%xmm1
+ vmovdqu 128-16(%rsi),%xmm2
+ vmovdqu .Lgfpoly(%rip),%xmm3
+ vpshufb %xmm1,%xmm0,%xmm0
+
+ vpclmulqdq $0x00,%xmm2,%xmm0,%xmm4
+ vpclmulqdq $0x01,%xmm2,%xmm0,%xmm5
+ vpclmulqdq $0x10,%xmm2,%xmm0,%xmm6
+ vpxor %xmm6,%xmm5,%xmm5
+ vpclmulqdq $0x01,%xmm4,%xmm3,%xmm6
+ vpshufd $0x4e,%xmm4,%xmm4
+ vpxor %xmm4,%xmm5,%xmm5
+ vpxor %xmm6,%xmm5,%xmm5
+ vpclmulqdq $0x11,%xmm2,%xmm0,%xmm0
+ vpclmulqdq $0x01,%xmm5,%xmm3,%xmm4
+ vpshufd $0x4e,%xmm5,%xmm5
+ vpxor %xmm5,%xmm0,%xmm0
+ vpxor %xmm4,%xmm0,%xmm0
+
+
+ vpshufb %xmm1,%xmm0,%xmm0
+ vmovdqu %xmm0,(%rdi)
+ ret
+
+.cfi_endproc
+.size gcm_gmult_vpclmulqdq_avx2, . - gcm_gmult_vpclmulqdq_avx2
+.globl gcm_ghash_vpclmulqdq_avx2
+.hidden gcm_ghash_vpclmulqdq_avx2
+.type gcm_ghash_vpclmulqdq_avx2,@function
+.align 32
+gcm_ghash_vpclmulqdq_avx2:
+.cfi_startproc
+
+_CET_ENDBR
+
+
+
+
+
+
+ vmovdqu .Lbswap_mask(%rip),%xmm6
+ vmovdqu .Lgfpoly(%rip),%xmm7
+
+
+ vmovdqu (%rdi),%xmm5
+ vpshufb %xmm6,%xmm5,%xmm5
+
+
+ cmpq $32,%rcx
+ jb .Lghash_lastblock
+
+
+
+ vinserti128 $1,%xmm6,%ymm6,%ymm6
+ vinserti128 $1,%xmm7,%ymm7,%ymm7
+
+ cmpq $127,%rcx
+ jbe .Lghash_loop_1x
+
+
+ vmovdqu 128(%rsi),%ymm8
+ vmovdqu 128+32(%rsi),%ymm9
+.Lghash_loop_4x:
+
+ vmovdqu 0(%rdx),%ymm1
+ vpshufb %ymm6,%ymm1,%ymm1
+ vmovdqu 0(%rsi),%ymm2
+ vpxor %ymm5,%ymm1,%ymm1
+ vpclmulqdq $0x00,%ymm2,%ymm1,%ymm3
+ vpclmulqdq $0x11,%ymm2,%ymm1,%ymm5
+ vpunpckhqdq %ymm1,%ymm1,%ymm0
+ vpxor %ymm1,%ymm0,%ymm0
+ vpclmulqdq $0x00,%ymm8,%ymm0,%ymm4
+
+ vmovdqu 32(%rdx),%ymm1
+ vpshufb %ymm6,%ymm1,%ymm1
+ vmovdqu 32(%rsi),%ymm2
+ vpclmulqdq $0x00,%ymm2,%ymm1,%ymm0
+ vpxor %ymm0,%ymm3,%ymm3
+ vpclmulqdq $0x11,%ymm2,%ymm1,%ymm0
+ vpxor %ymm0,%ymm5,%ymm5
+ vpunpckhqdq %ymm1,%ymm1,%ymm0
+ vpxor %ymm1,%ymm0,%ymm0
+ vpclmulqdq $0x10,%ymm8,%ymm0,%ymm0
+ vpxor %ymm0,%ymm4,%ymm4
+
+ vmovdqu 64(%rdx),%ymm1
+ vpshufb %ymm6,%ymm1,%ymm1
+ vmovdqu 64(%rsi),%ymm2
+ vpclmulqdq $0x00,%ymm2,%ymm1,%ymm0
+ vpxor %ymm0,%ymm3,%ymm3
+ vpclmulqdq $0x11,%ymm2,%ymm1,%ymm0
+ vpxor %ymm0,%ymm5,%ymm5
+ vpunpckhqdq %ymm1,%ymm1,%ymm0
+ vpxor %ymm1,%ymm0,%ymm0
+ vpclmulqdq $0x00,%ymm9,%ymm0,%ymm0
+ vpxor %ymm0,%ymm4,%ymm4
+
+
+ vmovdqu 96(%rdx),%ymm1
+ vpshufb %ymm6,%ymm1,%ymm1
+ vmovdqu 96(%rsi),%ymm2
+ vpclmulqdq $0x00,%ymm2,%ymm1,%ymm0
+ vpxor %ymm0,%ymm3,%ymm3
+ vpclmulqdq $0x11,%ymm2,%ymm1,%ymm0
+ vpxor %ymm0,%ymm5,%ymm5
+ vpunpckhqdq %ymm1,%ymm1,%ymm0
+ vpxor %ymm1,%ymm0,%ymm0
+ vpclmulqdq $0x10,%ymm9,%ymm0,%ymm0
+ vpxor %ymm0,%ymm4,%ymm4
+
+ vpxor %ymm3,%ymm4,%ymm4
+ vpxor %ymm5,%ymm4,%ymm4
+
+
+ vbroadcasti128 .Lgfpoly(%rip),%ymm2
+ vpclmulqdq $0x01,%ymm3,%ymm2,%ymm0
+ vpshufd $0x4e,%ymm3,%ymm3
+ vpxor %ymm3,%ymm4,%ymm4
+ vpxor %ymm0,%ymm4,%ymm4
+
+ vpclmulqdq $0x01,%ymm4,%ymm2,%ymm0
+ vpshufd $0x4e,%ymm4,%ymm4
+ vpxor %ymm4,%ymm5,%ymm5
+ vpxor %ymm0,%ymm5,%ymm5
+ vextracti128 $1,%ymm5,%xmm0
+ vpxor %xmm0,%xmm5,%xmm5
+
+ subq $-128,%rdx
+ addq $-128,%rcx
+ cmpq $127,%rcx
+ ja .Lghash_loop_4x
+
+
+ cmpq $32,%rcx
+ jb .Lghash_loop_1x_done
+.Lghash_loop_1x:
+ vmovdqu (%rdx),%ymm0
+ vpshufb %ymm6,%ymm0,%ymm0
+ vpxor %ymm0,%ymm5,%ymm5
+ vmovdqu 128-32(%rsi),%ymm0
+ vpclmulqdq $0x00,%ymm0,%ymm5,%ymm1
+ vpclmulqdq $0x01,%ymm0,%ymm5,%ymm2
+ vpclmulqdq $0x10,%ymm0,%ymm5,%ymm3
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x01,%ymm1,%ymm7,%ymm3
+ vpshufd $0x4e,%ymm1,%ymm1
+ vpxor %ymm1,%ymm2,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x11,%ymm0,%ymm5,%ymm5
+ vpclmulqdq $0x01,%ymm2,%ymm7,%ymm1
+ vpshufd $0x4e,%ymm2,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpxor %ymm1,%ymm5,%ymm5
+
+ vextracti128 $1,%ymm5,%xmm0
+ vpxor %xmm0,%xmm5,%xmm5
+ addq $32,%rdx
+ subq $32,%rcx
+ cmpq $32,%rcx
+ jae .Lghash_loop_1x
+.Lghash_loop_1x_done:
+
+
+.Lghash_lastblock:
+ testq %rcx,%rcx
+ jz .Lghash_done
+ vmovdqu (%rdx),%xmm0
+ vpshufb %xmm6,%xmm0,%xmm0
+ vpxor %xmm0,%xmm5,%xmm5
+ vmovdqu 128-16(%rsi),%xmm0
+ vpclmulqdq $0x00,%xmm0,%xmm5,%xmm1
+ vpclmulqdq $0x01,%xmm0,%xmm5,%xmm2
+ vpclmulqdq $0x10,%xmm0,%xmm5,%xmm3
+ vpxor %xmm3,%xmm2,%xmm2
+ vpclmulqdq $0x01,%xmm1,%xmm7,%xmm3
+ vpshufd $0x4e,%xmm1,%xmm1
+ vpxor %xmm1,%xmm2,%xmm2
+ vpxor %xmm3,%xmm2,%xmm2
+ vpclmulqdq $0x11,%xmm0,%xmm5,%xmm5
+ vpclmulqdq $0x01,%xmm2,%xmm7,%xmm1
+ vpshufd $0x4e,%xmm2,%xmm2
+ vpxor %xmm2,%xmm5,%xmm5
+ vpxor %xmm1,%xmm5,%xmm5
+
+
+.Lghash_done:
+
+ vpshufb %xmm6,%xmm5,%xmm5
+ vmovdqu %xmm5,(%rdi)
+
+ vzeroupper
+ ret
+
+.cfi_endproc
+.size gcm_ghash_vpclmulqdq_avx2, . - gcm_ghash_vpclmulqdq_avx2
+.globl aes_gcm_enc_update_vaes_avx2
+.hidden aes_gcm_enc_update_vaes_avx2
+.type aes_gcm_enc_update_vaes_avx2,@function
+.align 32
+aes_gcm_enc_update_vaes_avx2:
+.cfi_startproc
+
+_CET_ENDBR
+ pushq %r12
+.cfi_adjust_cfa_offset 8
+.cfi_offset %r12,-16
+
+ movq 16(%rsp),%r12
+#ifdef BORINGSSL_DISPATCH_TEST
+.extern BORINGSSL_function_hit
+.hidden BORINGSSL_function_hit
+ movb $1,BORINGSSL_function_hit+8(%rip)
+#endif
+ vbroadcasti128 .Lbswap_mask(%rip),%ymm0
+
+
+
+ vmovdqu (%r12),%xmm1
+ vpshufb %xmm0,%xmm1,%xmm1
+ vbroadcasti128 (%r8),%ymm11
+ vpshufb %ymm0,%ymm11,%ymm11
+
+
+
+ movl 240(%rcx),%r10d
+ leal -20(,%r10,4),%r10d
+
+
+
+
+ leaq 96(%rcx,%r10,4),%r11
+ vbroadcasti128 (%rcx),%ymm9
+ vbroadcasti128 (%r11),%ymm10
+
+
+ vpaddd .Lctr_pattern(%rip),%ymm11,%ymm11
+
+
+
+ cmpq $127,%rdx
+ jbe .Lcrypt_loop_4x_done__func1
+
+ vmovdqu 128(%r9),%ymm7
+ vmovdqu 128+32(%r9),%ymm8
+
+
+
+ vmovdqu .Linc_2blocks(%rip),%ymm2
+ vpshufb %ymm0,%ymm11,%ymm12
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm13
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm14
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm15
+ vpaddd %ymm2,%ymm11,%ymm11
+
+
+ vpxor %ymm9,%ymm12,%ymm12
+ vpxor %ymm9,%ymm13,%ymm13
+ vpxor %ymm9,%ymm14,%ymm14
+ vpxor %ymm9,%ymm15,%ymm15
+
+ leaq 16(%rcx),%rax
+.Lvaesenc_loop_first_4_vecs__func1:
+ vbroadcasti128 (%rax),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ addq $16,%rax
+ cmpq %rax,%r11
+ jne .Lvaesenc_loop_first_4_vecs__func1
+ vpxor 0(%rdi),%ymm10,%ymm2
+ vpxor 32(%rdi),%ymm10,%ymm3
+ vpxor 64(%rdi),%ymm10,%ymm5
+ vpxor 96(%rdi),%ymm10,%ymm6
+ vaesenclast %ymm2,%ymm12,%ymm12
+ vaesenclast %ymm3,%ymm13,%ymm13
+ vaesenclast %ymm5,%ymm14,%ymm14
+ vaesenclast %ymm6,%ymm15,%ymm15
+ vmovdqu %ymm12,0(%rsi)
+ vmovdqu %ymm13,32(%rsi)
+ vmovdqu %ymm14,64(%rsi)
+ vmovdqu %ymm15,96(%rsi)
+
+ subq $-128,%rdi
+ addq $-128,%rdx
+ cmpq $127,%rdx
+ jbe .Lghash_last_ciphertext_4x__func1
+.align 16
+.Lcrypt_loop_4x__func1:
+
+
+
+
+ vmovdqu .Linc_2blocks(%rip),%ymm2
+ vpshufb %ymm0,%ymm11,%ymm12
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm13
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm14
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm15
+ vpaddd %ymm2,%ymm11,%ymm11
+
+
+ vpxor %ymm9,%ymm12,%ymm12
+ vpxor %ymm9,%ymm13,%ymm13
+ vpxor %ymm9,%ymm14,%ymm14
+ vpxor %ymm9,%ymm15,%ymm15
+
+ cmpl $24,%r10d
+ jl .Laes128__func1
+ je .Laes192__func1
+
+ vbroadcasti128 -208(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vbroadcasti128 -192(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+.Laes192__func1:
+ vbroadcasti128 -176(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vbroadcasti128 -160(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+.Laes128__func1:
+ prefetcht0 512(%rdi)
+ prefetcht0 512+64(%rdi)
+
+ vmovdqu 0(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 0(%r9),%ymm4
+ vpxor %ymm1,%ymm3,%ymm3
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x00,%ymm7,%ymm2,%ymm6
+
+ vbroadcasti128 -144(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vbroadcasti128 -128(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vmovdqu 32(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 32(%r9),%ymm4
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x10,%ymm7,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vbroadcasti128 -112(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vmovdqu 64(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 64(%r9),%ymm4
+
+ vbroadcasti128 -96(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+
+ vbroadcasti128 -80(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x00,%ymm8,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+
+ vmovdqu 96(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+
+ vbroadcasti128 -64(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vmovdqu 96(%r9),%ymm4
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x10,%ymm8,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vbroadcasti128 -48(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm1,%ymm6,%ymm6
+
+
+ vbroadcasti128 .Lgfpoly(%rip),%ymm4
+ vpclmulqdq $0x01,%ymm5,%ymm4,%ymm2
+ vpshufd $0x4e,%ymm5,%ymm5
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vbroadcasti128 -32(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vpclmulqdq $0x01,%ymm6,%ymm4,%ymm2
+ vpshufd $0x4e,%ymm6,%ymm6
+ vpxor %ymm6,%ymm1,%ymm1
+ vpxor %ymm2,%ymm1,%ymm1
+
+ vbroadcasti128 -16(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vextracti128 $1,%ymm1,%xmm2
+ vpxor %xmm2,%xmm1,%xmm1
+
+
+ subq $-128,%rsi
+ vpxor 0(%rdi),%ymm10,%ymm2
+ vpxor 32(%rdi),%ymm10,%ymm3
+ vpxor 64(%rdi),%ymm10,%ymm5
+ vpxor 96(%rdi),%ymm10,%ymm6
+ vaesenclast %ymm2,%ymm12,%ymm12
+ vaesenclast %ymm3,%ymm13,%ymm13
+ vaesenclast %ymm5,%ymm14,%ymm14
+ vaesenclast %ymm6,%ymm15,%ymm15
+ vmovdqu %ymm12,0(%rsi)
+ vmovdqu %ymm13,32(%rsi)
+ vmovdqu %ymm14,64(%rsi)
+ vmovdqu %ymm15,96(%rsi)
+
+ subq $-128,%rdi
+
+ addq $-128,%rdx
+ cmpq $127,%rdx
+ ja .Lcrypt_loop_4x__func1
+.Lghash_last_ciphertext_4x__func1:
+
+ vmovdqu 0(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 0(%r9),%ymm4
+ vpxor %ymm1,%ymm3,%ymm3
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x00,%ymm7,%ymm2,%ymm6
+
+ vmovdqu 32(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 32(%r9),%ymm4
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x10,%ymm7,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vmovdqu 64(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 64(%r9),%ymm4
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x00,%ymm8,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+
+ vmovdqu 96(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 96(%r9),%ymm4
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x10,%ymm8,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm1,%ymm6,%ymm6
+
+
+ vbroadcasti128 .Lgfpoly(%rip),%ymm4
+ vpclmulqdq $0x01,%ymm5,%ymm4,%ymm2
+ vpshufd $0x4e,%ymm5,%ymm5
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vpclmulqdq $0x01,%ymm6,%ymm4,%ymm2
+ vpshufd $0x4e,%ymm6,%ymm6
+ vpxor %ymm6,%ymm1,%ymm1
+ vpxor %ymm2,%ymm1,%ymm1
+ vextracti128 $1,%ymm1,%xmm2
+ vpxor %xmm2,%xmm1,%xmm1
+
+ subq $-128,%rsi
+.Lcrypt_loop_4x_done__func1:
+
+ testq %rdx,%rdx
+ jz .Ldone__func1
+
+
+
+
+
+ leaq 128(%r9),%r8
+ subq %rdx,%r8
+
+
+ vpxor %xmm5,%xmm5,%xmm5
+ vpxor %xmm6,%xmm6,%xmm6
+ vpxor %xmm7,%xmm7,%xmm7
+
+ cmpq $64,%rdx
+ jb .Llessthan64bytes__func1
+
+
+ vpshufb %ymm0,%ymm11,%ymm12
+ vpaddd .Linc_2blocks(%rip),%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm13
+ vpaddd .Linc_2blocks(%rip),%ymm11,%ymm11
+ vpxor %ymm9,%ymm12,%ymm12
+ vpxor %ymm9,%ymm13,%ymm13
+ leaq 16(%rcx),%rax
+.Lvaesenc_loop_tail_1__func1:
+ vbroadcasti128 (%rax),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ addq $16,%rax
+ cmpq %rax,%r11
+ jne .Lvaesenc_loop_tail_1__func1
+ vaesenclast %ymm10,%ymm12,%ymm12
+ vaesenclast %ymm10,%ymm13,%ymm13
+
+
+ vmovdqu 0(%rdi),%ymm2
+ vmovdqu 32(%rdi),%ymm3
+ vpxor %ymm2,%ymm12,%ymm12
+ vpxor %ymm3,%ymm13,%ymm13
+ vmovdqu %ymm12,0(%rsi)
+ vmovdqu %ymm13,32(%rsi)
+
+
+ vpshufb %ymm0,%ymm12,%ymm12
+ vpshufb %ymm0,%ymm13,%ymm13
+ vpxor %ymm1,%ymm12,%ymm12
+ vmovdqu (%r8),%ymm2
+ vmovdqu 32(%r8),%ymm3
+ vpclmulqdq $0x00,%ymm2,%ymm12,%ymm5
+ vpclmulqdq $0x01,%ymm2,%ymm12,%ymm6
+ vpclmulqdq $0x10,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%ymm2,%ymm12,%ymm7
+ vpclmulqdq $0x00,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm5,%ymm5
+ vpclmulqdq $0x01,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x10,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm7,%ymm7
+
+ addq $64,%r8
+ addq $64,%rdi
+ addq $64,%rsi
+ subq $64,%rdx
+ jz .Lreduce__func1
+
+ vpxor %xmm1,%xmm1,%xmm1
+
+
+.Llessthan64bytes__func1:
+ vpshufb %ymm0,%ymm11,%ymm12
+ vpaddd .Linc_2blocks(%rip),%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm13
+ vpxor %ymm9,%ymm12,%ymm12
+ vpxor %ymm9,%ymm13,%ymm13
+ leaq 16(%rcx),%rax
+.Lvaesenc_loop_tail_2__func1:
+ vbroadcasti128 (%rax),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ addq $16,%rax
+ cmpq %rax,%r11
+ jne .Lvaesenc_loop_tail_2__func1
+ vaesenclast %ymm10,%ymm12,%ymm12
+ vaesenclast %ymm10,%ymm13,%ymm13
+
+
+
+
+ cmpq $32,%rdx
+ jb .Lxor_one_block__func1
+ je .Lxor_two_blocks__func1
+
+.Lxor_three_blocks__func1:
+ vmovdqu 0(%rdi),%ymm2
+ vmovdqu 32(%rdi),%xmm3
+ vpxor %ymm2,%ymm12,%ymm12
+ vpxor %xmm3,%xmm13,%xmm13
+ vmovdqu %ymm12,0(%rsi)
+ vmovdqu %xmm13,32(%rsi)
+
+ vpshufb %ymm0,%ymm12,%ymm12
+ vpshufb %xmm0,%xmm13,%xmm13
+ vpxor %ymm1,%ymm12,%ymm12
+ vmovdqu (%r8),%ymm2
+ vmovdqu 32(%r8),%xmm3
+ vpclmulqdq $0x00,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm5,%ymm5
+ vpclmulqdq $0x01,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x10,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm7,%ymm7
+ jmp .Lghash_mul_one_vec_unreduced__func1
+
+.Lxor_two_blocks__func1:
+ vmovdqu (%rdi),%ymm2
+ vpxor %ymm2,%ymm12,%ymm12
+ vmovdqu %ymm12,(%rsi)
+ vpshufb %ymm0,%ymm12,%ymm12
+ vpxor %ymm1,%ymm12,%ymm12
+ vmovdqu (%r8),%ymm2
+ jmp .Lghash_mul_one_vec_unreduced__func1
+
+.Lxor_one_block__func1:
+ vmovdqu (%rdi),%xmm2
+ vpxor %xmm2,%xmm12,%xmm12
+ vmovdqu %xmm12,(%rsi)
+ vpshufb %xmm0,%xmm12,%xmm12
+ vpxor %xmm1,%xmm12,%xmm12
+ vmovdqu (%r8),%xmm2
+
+.Lghash_mul_one_vec_unreduced__func1:
+ vpclmulqdq $0x00,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm5,%ymm5
+ vpclmulqdq $0x01,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x10,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm7,%ymm7
+
+.Lreduce__func1:
+
+ vbroadcasti128 .Lgfpoly(%rip),%ymm2
+ vpclmulqdq $0x01,%ymm5,%ymm2,%ymm3
+ vpshufd $0x4e,%ymm5,%ymm5
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm3,%ymm6,%ymm6
+ vpclmulqdq $0x01,%ymm6,%ymm2,%ymm3
+ vpshufd $0x4e,%ymm6,%ymm6
+ vpxor %ymm6,%ymm7,%ymm7
+ vpxor %ymm3,%ymm7,%ymm7
+ vextracti128 $1,%ymm7,%xmm1
+ vpxor %xmm7,%xmm1,%xmm1
+
+.Ldone__func1:
+
+ vpshufb %xmm0,%xmm1,%xmm1
+ vmovdqu %xmm1,(%r12)
+
+ vzeroupper
+ popq %r12
+.cfi_adjust_cfa_offset -8
+.cfi_restore %r12
+ ret
+
+.cfi_endproc
+.size aes_gcm_enc_update_vaes_avx2, . - aes_gcm_enc_update_vaes_avx2
+.globl aes_gcm_dec_update_vaes_avx2
+.hidden aes_gcm_dec_update_vaes_avx2
+.type aes_gcm_dec_update_vaes_avx2,@function
+.align 32
+aes_gcm_dec_update_vaes_avx2:
+.cfi_startproc
+
+_CET_ENDBR
+ pushq %r12
+.cfi_adjust_cfa_offset 8
+.cfi_offset %r12,-16
+
+ movq 16(%rsp),%r12
+ vbroadcasti128 .Lbswap_mask(%rip),%ymm0
+
+
+
+ vmovdqu (%r12),%xmm1
+ vpshufb %xmm0,%xmm1,%xmm1
+ vbroadcasti128 (%r8),%ymm11
+ vpshufb %ymm0,%ymm11,%ymm11
+
+
+
+ movl 240(%rcx),%r10d
+ leal -20(,%r10,4),%r10d
+
+
+
+
+ leaq 96(%rcx,%r10,4),%r11
+ vbroadcasti128 (%rcx),%ymm9
+ vbroadcasti128 (%r11),%ymm10
+
+
+ vpaddd .Lctr_pattern(%rip),%ymm11,%ymm11
+
+
+
+ cmpq $127,%rdx
+ jbe .Lcrypt_loop_4x_done__func2
+
+ vmovdqu 128(%r9),%ymm7
+ vmovdqu 128+32(%r9),%ymm8
+.align 16
+.Lcrypt_loop_4x__func2:
+
+
+
+
+ vmovdqu .Linc_2blocks(%rip),%ymm2
+ vpshufb %ymm0,%ymm11,%ymm12
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm13
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm14
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm15
+ vpaddd %ymm2,%ymm11,%ymm11
+
+
+ vpxor %ymm9,%ymm12,%ymm12
+ vpxor %ymm9,%ymm13,%ymm13
+ vpxor %ymm9,%ymm14,%ymm14
+ vpxor %ymm9,%ymm15,%ymm15
+
+ cmpl $24,%r10d
+ jl .Laes128__func2
+ je .Laes192__func2
+
+ vbroadcasti128 -208(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vbroadcasti128 -192(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+.Laes192__func2:
+ vbroadcasti128 -176(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vbroadcasti128 -160(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+.Laes128__func2:
+ prefetcht0 512(%rdi)
+ prefetcht0 512+64(%rdi)
+
+ vmovdqu 0(%rdi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 0(%r9),%ymm4
+ vpxor %ymm1,%ymm3,%ymm3
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x00,%ymm7,%ymm2,%ymm6
+
+ vbroadcasti128 -144(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vbroadcasti128 -128(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vmovdqu 32(%rdi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 32(%r9),%ymm4
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x10,%ymm7,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vbroadcasti128 -112(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vmovdqu 64(%rdi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 64(%r9),%ymm4
+
+ vbroadcasti128 -96(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+
+ vbroadcasti128 -80(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x00,%ymm8,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+
+ vmovdqu 96(%rdi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+
+ vbroadcasti128 -64(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vmovdqu 96(%r9),%ymm4
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x10,%ymm8,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vbroadcasti128 -48(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm1,%ymm6,%ymm6
+
+
+ vbroadcasti128 .Lgfpoly(%rip),%ymm4
+ vpclmulqdq $0x01,%ymm5,%ymm4,%ymm2
+ vpshufd $0x4e,%ymm5,%ymm5
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vbroadcasti128 -32(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vpclmulqdq $0x01,%ymm6,%ymm4,%ymm2
+ vpshufd $0x4e,%ymm6,%ymm6
+ vpxor %ymm6,%ymm1,%ymm1
+ vpxor %ymm2,%ymm1,%ymm1
+
+ vbroadcasti128 -16(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vextracti128 $1,%ymm1,%xmm2
+ vpxor %xmm2,%xmm1,%xmm1
+
+
+
+ vpxor 0(%rdi),%ymm10,%ymm2
+ vpxor 32(%rdi),%ymm10,%ymm3
+ vpxor 64(%rdi),%ymm10,%ymm5
+ vpxor 96(%rdi),%ymm10,%ymm6
+ vaesenclast %ymm2,%ymm12,%ymm12
+ vaesenclast %ymm3,%ymm13,%ymm13
+ vaesenclast %ymm5,%ymm14,%ymm14
+ vaesenclast %ymm6,%ymm15,%ymm15
+ vmovdqu %ymm12,0(%rsi)
+ vmovdqu %ymm13,32(%rsi)
+ vmovdqu %ymm14,64(%rsi)
+ vmovdqu %ymm15,96(%rsi)
+
+ subq $-128,%rdi
+ subq $-128,%rsi
+ addq $-128,%rdx
+ cmpq $127,%rdx
+ ja .Lcrypt_loop_4x__func2
+.Lcrypt_loop_4x_done__func2:
+
+ testq %rdx,%rdx
+ jz .Ldone__func2
+
+
+
+
+
+ leaq 128(%r9),%r8
+ subq %rdx,%r8
+
+
+ vpxor %xmm5,%xmm5,%xmm5
+ vpxor %xmm6,%xmm6,%xmm6
+ vpxor %xmm7,%xmm7,%xmm7
+
+ cmpq $64,%rdx
+ jb .Llessthan64bytes__func2
+
+
+ vpshufb %ymm0,%ymm11,%ymm12
+ vpaddd .Linc_2blocks(%rip),%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm13
+ vpaddd .Linc_2blocks(%rip),%ymm11,%ymm11
+ vpxor %ymm9,%ymm12,%ymm12
+ vpxor %ymm9,%ymm13,%ymm13
+ leaq 16(%rcx),%rax
+.Lvaesenc_loop_tail_1__func2:
+ vbroadcasti128 (%rax),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ addq $16,%rax
+ cmpq %rax,%r11
+ jne .Lvaesenc_loop_tail_1__func2
+ vaesenclast %ymm10,%ymm12,%ymm12
+ vaesenclast %ymm10,%ymm13,%ymm13
+
+
+ vmovdqu 0(%rdi),%ymm2
+ vmovdqu 32(%rdi),%ymm3
+ vpxor %ymm2,%ymm12,%ymm12
+ vpxor %ymm3,%ymm13,%ymm13
+ vmovdqu %ymm12,0(%rsi)
+ vmovdqu %ymm13,32(%rsi)
+
+
+ vpshufb %ymm0,%ymm2,%ymm12
+ vpshufb %ymm0,%ymm3,%ymm13
+ vpxor %ymm1,%ymm12,%ymm12
+ vmovdqu (%r8),%ymm2
+ vmovdqu 32(%r8),%ymm3
+ vpclmulqdq $0x00,%ymm2,%ymm12,%ymm5
+ vpclmulqdq $0x01,%ymm2,%ymm12,%ymm6
+ vpclmulqdq $0x10,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%ymm2,%ymm12,%ymm7
+ vpclmulqdq $0x00,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm5,%ymm5
+ vpclmulqdq $0x01,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x10,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm7,%ymm7
+
+ addq $64,%r8
+ addq $64,%rdi
+ addq $64,%rsi
+ subq $64,%rdx
+ jz .Lreduce__func2
+
+ vpxor %xmm1,%xmm1,%xmm1
+
+
+.Llessthan64bytes__func2:
+ vpshufb %ymm0,%ymm11,%ymm12
+ vpaddd .Linc_2blocks(%rip),%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm13
+ vpxor %ymm9,%ymm12,%ymm12
+ vpxor %ymm9,%ymm13,%ymm13
+ leaq 16(%rcx),%rax
+.Lvaesenc_loop_tail_2__func2:
+ vbroadcasti128 (%rax),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ addq $16,%rax
+ cmpq %rax,%r11
+ jne .Lvaesenc_loop_tail_2__func2
+ vaesenclast %ymm10,%ymm12,%ymm12
+ vaesenclast %ymm10,%ymm13,%ymm13
+
+
+
+
+ cmpq $32,%rdx
+ jb .Lxor_one_block__func2
+ je .Lxor_two_blocks__func2
+
+.Lxor_three_blocks__func2:
+ vmovdqu 0(%rdi),%ymm2
+ vmovdqu 32(%rdi),%xmm3
+ vpxor %ymm2,%ymm12,%ymm12
+ vpxor %xmm3,%xmm13,%xmm13
+ vmovdqu %ymm12,0(%rsi)
+ vmovdqu %xmm13,32(%rsi)
+
+ vpshufb %ymm0,%ymm2,%ymm12
+ vpshufb %xmm0,%xmm3,%xmm13
+ vpxor %ymm1,%ymm12,%ymm12
+ vmovdqu (%r8),%ymm2
+ vmovdqu 32(%r8),%xmm3
+ vpclmulqdq $0x00,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm5,%ymm5
+ vpclmulqdq $0x01,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x10,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm7,%ymm7
+ jmp .Lghash_mul_one_vec_unreduced__func2
+
+.Lxor_two_blocks__func2:
+ vmovdqu (%rdi),%ymm2
+ vpxor %ymm2,%ymm12,%ymm12
+ vmovdqu %ymm12,(%rsi)
+ vpshufb %ymm0,%ymm2,%ymm12
+ vpxor %ymm1,%ymm12,%ymm12
+ vmovdqu (%r8),%ymm2
+ jmp .Lghash_mul_one_vec_unreduced__func2
+
+.Lxor_one_block__func2:
+ vmovdqu (%rdi),%xmm2
+ vpxor %xmm2,%xmm12,%xmm12
+ vmovdqu %xmm12,(%rsi)
+ vpshufb %xmm0,%xmm2,%xmm12
+ vpxor %xmm1,%xmm12,%xmm12
+ vmovdqu (%r8),%xmm2
+
+.Lghash_mul_one_vec_unreduced__func2:
+ vpclmulqdq $0x00,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm5,%ymm5
+ vpclmulqdq $0x01,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x10,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm7,%ymm7
+
+.Lreduce__func2:
+
+ vbroadcasti128 .Lgfpoly(%rip),%ymm2
+ vpclmulqdq $0x01,%ymm5,%ymm2,%ymm3
+ vpshufd $0x4e,%ymm5,%ymm5
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm3,%ymm6,%ymm6
+ vpclmulqdq $0x01,%ymm6,%ymm2,%ymm3
+ vpshufd $0x4e,%ymm6,%ymm6
+ vpxor %ymm6,%ymm7,%ymm7
+ vpxor %ymm3,%ymm7,%ymm7
+ vextracti128 $1,%ymm7,%xmm1
+ vpxor %xmm7,%xmm1,%xmm1
+
+.Ldone__func2:
+
+ vpshufb %xmm0,%xmm1,%xmm1
+ vmovdqu %xmm1,(%r12)
+
+ vzeroupper
+ popq %r12
+.cfi_adjust_cfa_offset -8
+.cfi_restore %r12
+ ret
+
+.cfi_endproc
+.size aes_gcm_dec_update_vaes_avx2, . - aes_gcm_dec_update_vaes_avx2
+#endif
diff --git a/sys/contrib/openzfs/contrib/initramfs/scripts/zfs b/sys/contrib/openzfs/contrib/initramfs/scripts/zfs
index c569b2528368..67707e9d80f4 100644
--- a/sys/contrib/openzfs/contrib/initramfs/scripts/zfs
+++ b/sys/contrib/openzfs/contrib/initramfs/scripts/zfs
@@ -979,7 +979,8 @@ mountroot()
touch /run/zfs_unlock_complete
if [ -e /run/zfs_unlock_complete_notify ]; then
- read -r < /run/zfs_unlock_complete_notify
+ # shellcheck disable=SC2034
+ read -r zfs_unlock_complete_notify < /run/zfs_unlock_complete_notify
fi
# ------------
diff --git a/sys/contrib/openzfs/include/os/freebsd/spl/sys/debug.h b/sys/contrib/openzfs/include/os/freebsd/spl/sys/debug.h
index 974704e92bbd..32bc02f3dc86 100644
--- a/sys/contrib/openzfs/include/os/freebsd/spl/sys/debug.h
+++ b/sys/contrib/openzfs/include/os/freebsd/spl/sys/debug.h
@@ -69,6 +69,10 @@
#define __maybe_unused __attribute__((unused))
#endif
+#ifndef __must_check
+#define __must_check __attribute__((__warn_unused_result__))
+#endif
+
/*
* Without this, we see warnings from objtool during normal Linux builds when
* the kernel is built with CONFIG_STACK_VALIDATION=y:
diff --git a/sys/contrib/openzfs/include/os/freebsd/spl/sys/proc.h b/sys/contrib/openzfs/include/os/freebsd/spl/sys/proc.h
index c6bc10d6babe..1cbd79ec893f 100644
--- a/sys/contrib/openzfs/include/os/freebsd/spl/sys/proc.h
+++ b/sys/contrib/openzfs/include/os/freebsd/spl/sys/proc.h
@@ -77,8 +77,8 @@ do_thread_create(caddr_t stk, size_t stksize, void (*proc)(void *), void *arg,
/*
* Be sure there are no surprises.
*/
- ASSERT(stk == NULL);
- ASSERT(len == 0);
+ ASSERT0P(stk);
+ ASSERT0(len);
ASSERT(state == TS_RUN);
if (pp == &p0)
diff --git a/sys/contrib/openzfs/include/os/linux/kernel/linux/simd_x86.h b/sys/contrib/openzfs/include/os/linux/kernel/linux/simd_x86.h
index cd245a5f0135..326f471d7c9b 100644
--- a/sys/contrib/openzfs/include/os/linux/kernel/linux/simd_x86.h
+++ b/sys/contrib/openzfs/include/os/linux/kernel/linux/simd_x86.h
@@ -139,15 +139,6 @@
*/
#if defined(HAVE_KERNEL_FPU_INTERNAL)
-/*
- * For kernels not exporting *kfpu_{begin,end} we have to use inline assembly
- * with the XSAVE{,OPT,S} instructions, so we need the toolchain to support at
- * least XSAVE.
- */
-#if !defined(HAVE_XSAVE)
-#error "Toolchain needs to support the XSAVE assembler instruction"
-#endif
-
#ifndef XFEATURE_MASK_XTILE
/*
* For kernels where this doesn't exist yet, we still don't want to break
@@ -335,9 +326,13 @@ kfpu_begin(void)
return;
}
#endif
+#if defined(HAVE_XSAVE)
if (static_cpu_has(X86_FEATURE_XSAVE)) {
kfpu_do_xsave("xsave", state, ~XFEATURE_MASK_XTILE);
- } else if (static_cpu_has(X86_FEATURE_FXSR)) {
+ return;
+ }
+#endif
+ if (static_cpu_has(X86_FEATURE_FXSR)) {
kfpu_save_fxsr(state);
} else {
kfpu_save_fsave(state);
@@ -390,9 +385,13 @@ kfpu_end(void)
goto out;
}
#endif
+#if defined(HAVE_XSAVE)
if (static_cpu_has(X86_FEATURE_XSAVE)) {
kfpu_do_xrstor("xrstor", state, ~XFEATURE_MASK_XTILE);
- } else if (static_cpu_has(X86_FEATURE_FXSR)) {
+ goto out;
+ }
+#endif
+ if (static_cpu_has(X86_FEATURE_FXSR)) {
kfpu_restore_fxsr(state);
} else {
kfpu_restore_fsave(state);
@@ -599,6 +598,32 @@ zfs_movbe_available(void)
}
/*
+ * Check if VAES instruction set is available
+ */
+static inline boolean_t
+zfs_vaes_available(void)
+{
+#if defined(X86_FEATURE_VAES)
+ return (!!boot_cpu_has(X86_FEATURE_VAES));
+#else
+ return (B_FALSE);
+#endif
+}
+
+/*
+ * Check if VPCLMULQDQ instruction set is available
+ */
+static inline boolean_t
+zfs_vpclmulqdq_available(void)
+{
+#if defined(X86_FEATURE_VPCLMULQDQ)
+ return (!!boot_cpu_has(X86_FEATURE_VPCLMULQDQ));
+#else
+ return (B_FALSE);
+#endif
+}
+
+/*
* Check if SHA_NI instruction set is available
*/
static inline boolean_t
diff --git a/sys/contrib/openzfs/include/os/linux/spl/sys/debug.h b/sys/contrib/openzfs/include/os/linux/spl/sys/debug.h
index 1671ba4074da..85b96e1e23a7 100644
--- a/sys/contrib/openzfs/include/os/linux/spl/sys/debug.h
+++ b/sys/contrib/openzfs/include/os/linux/spl/sys/debug.h
@@ -69,6 +69,10 @@
#define __maybe_unused __attribute__((unused))
#endif
+#ifndef __must_check
+#define __must_check __attribute__((__warn_unused_result__))
+#endif
+
/*
* Without this, we see warnings from objtool during normal Linux builds when
* the kernel is built with CONFIG_STACK_VALIDATION=y:
diff --git a/sys/contrib/openzfs/include/os/linux/spl/sys/mutex.h b/sys/contrib/openzfs/include/os/linux/spl/sys/mutex.h
index f000f53ab9b6..4eca2414fc5b 100644
--- a/sys/contrib/openzfs/include/os/linux/spl/sys/mutex.h
+++ b/sys/contrib/openzfs/include/os/linux/spl/sys/mutex.h
@@ -111,7 +111,7 @@ spl_mutex_lockdep_on_maybe(kmutex_t *mp) \
#undef mutex_destroy
#define mutex_destroy(mp) \
{ \
- VERIFY3P(mutex_owner(mp), ==, NULL); \
+ VERIFY0P(mutex_owner(mp)); \
}
#define mutex_tryenter(mp) \
diff --git a/sys/contrib/openzfs/include/os/linux/zfs/sys/trace_zil.h b/sys/contrib/openzfs/include/os/linux/zfs/sys/trace_zil.h
index 955462c85d10..e34ea46b3fe8 100644
--- a/sys/contrib/openzfs/include/os/linux/zfs/sys/trace_zil.h
+++ b/sys/contrib/openzfs/include/os/linux/zfs/sys/trace_zil.h
@@ -139,18 +139,18 @@
#define ZCW_TP_STRUCT_ENTRY \
__field(lwb_t *, zcw_lwb) \
__field(boolean_t, zcw_done) \
- __field(int, zcw_zio_error) \
+ __field(int, zcw_error) \
#define ZCW_TP_FAST_ASSIGN \
__entry->zcw_lwb = zcw->zcw_lwb; \
__entry->zcw_done = zcw->zcw_done; \
- __entry->zcw_zio_error = zcw->zcw_zio_error;
+ __entry->zcw_error = zcw->zcw_error;
#define ZCW_TP_PRINTK_FMT \
"zcw { lwb %p done %u error %u }"
#define ZCW_TP_PRINTK_ARGS \
- __entry->zcw_lwb, __entry->zcw_done, __entry->zcw_zio_error
+ __entry->zcw_lwb, __entry->zcw_done, __entry->zcw_error
/*
* Generic support for two argument tracepoints of the form:
diff --git a/sys/contrib/openzfs/include/sys/dmu.h b/sys/contrib/openzfs/include/sys/dmu.h
index 7c2024a16d8f..b18961be1282 100644
--- a/sys/contrib/openzfs/include/sys/dmu.h
+++ b/sys/contrib/openzfs/include/sys/dmu.h
@@ -742,8 +742,8 @@ dmu_buf_init_user(dmu_buf_user_t *dbu, dmu_buf_evict_func_t *evict_func_sync,
dmu_buf_evict_func_t *evict_func_async,
dmu_buf_t **clear_on_evict_dbufp __maybe_unused)
{
- ASSERT(dbu->dbu_evict_func_sync == NULL);
- ASSERT(dbu->dbu_evict_func_async == NULL);
+ ASSERT0P(dbu->dbu_evict_func_sync);
+ ASSERT0P(dbu->dbu_evict_func_async);
/* must have at least one evict func */
IMPLY(evict_func_sync == NULL, evict_func_async != NULL);
diff --git a/sys/contrib/openzfs/include/sys/spa.h b/sys/contrib/openzfs/include/sys/spa.h
index db6de332ae67..66db16b33c51 100644
--- a/sys/contrib/openzfs/include/sys/spa.h
+++ b/sys/contrib/openzfs/include/sys/spa.h
@@ -880,7 +880,6 @@ extern kcondvar_t spa_namespace_cv;
#define SPA_CONFIG_UPDATE_VDEVS 1
extern void spa_write_cachefile(spa_t *, boolean_t, boolean_t, boolean_t);
-extern void spa_config_load(void);
extern int spa_all_configs(uint64_t *generation, nvlist_t **pools);
extern void spa_config_set(spa_t *spa, nvlist_t *config);
extern nvlist_t *spa_config_generate(spa_t *spa, vdev_t *vd, uint64_t txg,
@@ -1244,7 +1243,6 @@ 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 *);
/* properties */
extern int spa_prop_set(spa_t *spa, nvlist_t *nvp);
diff --git a/sys/contrib/openzfs/include/sys/zfs_znode.h b/sys/contrib/openzfs/include/sys/zfs_znode.h
index ba577b80c98f..79b845a672a8 100644
--- a/sys/contrib/openzfs/include/sys/zfs_znode.h
+++ b/sys/contrib/openzfs/include/sys/zfs_znode.h
@@ -73,7 +73,7 @@ extern "C" {
pflags |= attr; \
else \
pflags &= ~attr; \
- VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_FLAGS(ZTOZSB(zp)), \
+ VERIFY0(sa_update(zp->z_sa_hdl, SA_ZPL_FLAGS(ZTOZSB(zp)), \
&pflags, sizeof (pflags), tx)); \
}
diff --git a/sys/contrib/openzfs/include/sys/zil.h b/sys/contrib/openzfs/include/sys/zil.h
index 9d1fb47e2dfc..da085998879b 100644
--- a/sys/contrib/openzfs/include/sys/zil.h
+++ b/sys/contrib/openzfs/include/sys/zil.h
@@ -456,7 +456,7 @@ typedef enum {
WR_NUM_STATES /* number of states */
} itx_wr_state_t;
-typedef void (*zil_callback_t)(void *data);
+typedef void (*zil_callback_t)(void *data, int err);
typedef struct itx {
list_node_t itx_node; /* linkage on zl_itx_list */
@@ -498,10 +498,13 @@ typedef struct zil_stats {
* (see zil_commit_writer_stall())
* - suspend: ZIL suspended
* (see zil_commit(), zil_get_commit_list())
+ * - crash: ZIL crashed
+ * (see zil_crash(), zil_commit(), ...)
*/
kstat_named_t zil_commit_error_count;
kstat_named_t zil_commit_stall_count;
kstat_named_t zil_commit_suspend_count;
+ kstat_named_t zil_commit_crash_count;
/*
* Number of transactions (reads, writes, renames, etc.)
@@ -549,6 +552,7 @@ typedef struct zil_sums {
wmsum_t zil_commit_error_count;
wmsum_t zil_commit_stall_count;
wmsum_t zil_commit_suspend_count;
+ wmsum_t zil_commit_crash_count;
wmsum_t zil_itx_count;
wmsum_t zil_itx_indirect_count;
wmsum_t zil_itx_indirect_bytes;
@@ -577,6 +581,25 @@ typedef struct zil_sums {
#define ZIL_STAT_BUMP(zil, stat) \
ZIL_STAT_INCR(zil, stat, 1);
+/*
+ * Flags for zil_commit_flags(). zil_commit() is a shortcut for
+ * zil_commit_flags(ZIL_COMMIT_FAILMODE), which is the most common use.
+ */
+typedef enum {
+ /*
+ * Try to commit the ZIL. If it fails, fall back to txg_wait_synced().
+ * If that fails, return EIO.
+ */
+ ZIL_COMMIT_NOW = 0,
+
+ /*
+ * Like ZIL_COMMIT_NOW, but if the ZIL commit fails because the pool
+ * suspended, act according to the pool's failmode= setting (wait for
+ * the pool to resume, or return EIO).
+ */
+ ZIL_COMMIT_FAILMODE = (1 << 1),
+} zil_commit_flag_t;
+
typedef int zil_parse_blk_func_t(zilog_t *zilog, const blkptr_t *bp, void *arg,
uint64_t txg);
typedef int zil_parse_lr_func_t(zilog_t *zilog, const lr_t *lr, void *arg,
@@ -606,14 +629,16 @@ extern boolean_t zil_destroy(zilog_t *zilog, boolean_t keep_first);
extern void zil_destroy_sync(zilog_t *zilog, dmu_tx_t *tx);
extern itx_t *zil_itx_create(uint64_t txtype, size_t lrsize);
-extern void zil_itx_destroy(itx_t *itx);
+extern void zil_itx_destroy(itx_t *itx, int err);
extern void zil_itx_assign(zilog_t *zilog, itx_t *itx, dmu_tx_t *tx);
extern void zil_async_to_sync(zilog_t *zilog, uint64_t oid);
-extern void zil_commit(zilog_t *zilog, uint64_t oid);
-extern void zil_commit_impl(zilog_t *zilog, uint64_t oid);
extern void zil_remove_async(zilog_t *zilog, uint64_t oid);
+extern int zil_commit_flags(zilog_t *zilog, uint64_t oid,
+ zil_commit_flag_t flags);
+extern int __must_check zil_commit(zilog_t *zilog, uint64_t oid);
+
extern int zil_reset(const char *osname, void *txarg);
extern int zil_claim(struct dsl_pool *dp,
struct dsl_dataset *ds, void *txarg);
diff --git a/sys/contrib/openzfs/include/sys/zil_impl.h b/sys/contrib/openzfs/include/sys/zil_impl.h
index 252264b9eae9..ea1364a7e35a 100644
--- a/sys/contrib/openzfs/include/sys/zil_impl.h
+++ b/sys/contrib/openzfs/include/sys/zil_impl.h
@@ -41,8 +41,8 @@ extern "C" {
*
* An lwb will start out in the "new" state, and transition to the "opened"
* state via a call to zil_lwb_write_open() on first itx assignment. When
- * transitioning from "new" to "opened" the zilog's "zl_issuer_lock" must be
- * held.
+ * transitioning from "new" to "opened" the zilog's "zl_issuer_lock" and
+ * LWB's "lwb_lock" must be held.
*
* After the lwb is "opened", it can be assigned number of itxs and transition
* into the "closed" state via zil_lwb_write_close() when full or on timeout.
@@ -100,16 +100,22 @@ typedef enum {
* holding the "zl_issuer_lock". After the lwb is issued, the zilog's
* "zl_lock" is used to protect the lwb against concurrent access.
*/
+typedef enum {
+ LWB_FLAG_SLIM = (1<<0), /* log block has slim format */
+ LWB_FLAG_SLOG = (1<<1), /* lwb_blk is on SLOG device */
+ LWB_FLAG_CRASHED = (1<<2), /* lwb is on the crash list */
+} lwb_flag_t;
+
typedef struct lwb {
zilog_t *lwb_zilog; /* back pointer to log struct */
blkptr_t lwb_blk; /* on disk address of this log blk */
- boolean_t lwb_slim; /* log block has slim format */
- boolean_t lwb_slog; /* lwb_blk is on SLOG device */
+ lwb_flag_t lwb_flags; /* extra info about this lwb */
int lwb_error; /* log block allocation error */
int lwb_nmax; /* max bytes in the buffer */
int lwb_nused; /* # used bytes in buffer */
int lwb_nfilled; /* # filled bytes in buffer */
int lwb_sz; /* size of block and buffer */
+ int lwb_min_sz; /* min size for range allocation */
lwb_state_t lwb_state; /* the state of this lwb */
char *lwb_buf; /* log write buffer */
zio_t *lwb_child_zio; /* parent zio for children */
@@ -124,7 +130,7 @@ typedef struct lwb {
list_t lwb_itxs; /* list of itx's */
list_t lwb_waiters; /* list of zil_commit_waiter's */
avl_tree_t lwb_vdev_tree; /* vdevs to flush after lwb write */
- kmutex_t lwb_vdev_lock; /* protects lwb_vdev_tree */
+ kmutex_t lwb_lock; /* protects lwb_vdev_tree and size */
} lwb_t;
/*
@@ -149,7 +155,7 @@ typedef struct zil_commit_waiter {
list_node_t zcw_node; /* linkage in lwb_t:lwb_waiter list */
lwb_t *zcw_lwb; /* back pointer to lwb when linked */
boolean_t zcw_done; /* B_TRUE when "done", else B_FALSE */
- int zcw_zio_error; /* contains the zio io_error value */
+ int zcw_error; /* result to return from zil_commit() */
} zil_commit_waiter_t;
/*
@@ -221,6 +227,7 @@ struct zilog {
uint64_t zl_cur_left; /* current burst remaining size */
uint64_t zl_cur_max; /* biggest record in current burst */
list_t zl_lwb_list; /* in-flight log write list */
+ list_t zl_lwb_crash_list; /* log writes in-flight at crash */
avl_tree_t zl_bp_tree; /* track bps during log parse */
clock_t zl_replay_time; /* lbolt of when replay started */
uint64_t zl_replay_blks; /* number of log blocks replayed */
@@ -245,6 +252,9 @@ struct zilog {
*/
uint64_t zl_max_block_size;
+ /* After crash, txg to restart zil */
+ uint64_t zl_restart_txg;
+
/* Pointer for per dataset zil sums */
zil_sums_t *zl_sums;
};
diff --git a/sys/contrib/openzfs/include/sys/zio.h b/sys/contrib/openzfs/include/sys/zio.h
index 4f46eab3db89..353805fcb969 100644
--- a/sys/contrib/openzfs/include/sys/zio.h
+++ b/sys/contrib/openzfs/include/sys/zio.h
@@ -360,26 +360,26 @@ struct zbookmark_err_phys {
(zb)->zb_blkid == ZB_ROOT_BLKID)
typedef struct zio_prop {
- enum zio_checksum zp_checksum;
- enum zio_compress zp_compress;
+ enum zio_checksum zp_checksum:8;
+ enum zio_compress zp_compress:8;
uint8_t zp_complevel;
uint8_t zp_level;
uint8_t zp_copies;
uint8_t zp_gang_copies;
- dmu_object_type_t zp_type;
- boolean_t zp_dedup;
- boolean_t zp_dedup_verify;
- boolean_t zp_nopwrite;
- boolean_t zp_brtwrite;
- boolean_t zp_encrypt;
- boolean_t zp_byteorder;
- boolean_t zp_direct_write;
- boolean_t zp_rewrite;
+ dmu_object_type_t zp_type:8;
+ dmu_object_type_t zp_storage_type:8;
+ boolean_t zp_dedup:1;
+ boolean_t zp_dedup_verify:1;
+ boolean_t zp_nopwrite:1;
+ boolean_t zp_brtwrite:1;
+ boolean_t zp_encrypt:1;
+ boolean_t zp_byteorder:1;
+ boolean_t zp_direct_write:1;
+ boolean_t zp_rewrite:1;
+ uint32_t zp_zpl_smallblk;
uint8_t zp_salt[ZIO_DATA_SALT_LEN];
uint8_t zp_iv[ZIO_DATA_IV_LEN];
uint8_t zp_mac[ZIO_DATA_MAC_LEN];
- uint32_t zp_zpl_smallblk;
- dmu_object_type_t zp_storage_type;
} zio_prop_t;
typedef struct zio_cksum_report zio_cksum_report_t;
@@ -622,7 +622,8 @@ extern zio_t *zio_free_sync(zio_t *pio, spa_t *spa, uint64_t txg,
const blkptr_t *bp, zio_flag_t flags);
extern int zio_alloc_zil(spa_t *spa, objset_t *os, uint64_t txg,
- blkptr_t *new_bp, uint64_t size, boolean_t *slog);
+ blkptr_t *new_bp, uint64_t min_size, uint64_t max_size, boolean_t *slog,
+ boolean_t allow_larger);
extern void zio_flush(zio_t *zio, vdev_t *vd);
extern void zio_shrink(zio_t *zio, uint64_t size);
diff --git a/sys/contrib/openzfs/include/sys/zvol_impl.h b/sys/contrib/openzfs/include/sys/zvol_impl.h
index f3dd9f26f23c..5422e66832c0 100644
--- a/sys/contrib/openzfs/include/sys/zvol_impl.h
+++ b/sys/contrib/openzfs/include/sys/zvol_impl.h
@@ -20,7 +20,7 @@
* CDDL HEADER END
*/
/*
- * Copyright (c) 2024, Klara, Inc.
+ * Copyright (c) 2024, 2025, Klara, Inc.
*/
#ifndef _SYS_ZVOL_IMPL_H
@@ -56,6 +56,7 @@ typedef struct zvol_state {
atomic_t zv_suspend_ref; /* refcount for suspend */
krwlock_t zv_suspend_lock; /* suspend lock */
kcondvar_t zv_removing_cv; /* ready to remove minor */
+ list_node_t zv_remove_node; /* node on removal list */
struct zvol_state_os *zv_zso; /* private platform state */
boolean_t zv_threading; /* volthreading property */
} zvol_state_t;
@@ -135,7 +136,7 @@ int zvol_os_rename_minor(zvol_state_t *zv, const char *newname);
int zvol_os_create_minor(const char *name);
int zvol_os_update_volsize(zvol_state_t *zv, uint64_t volsize);
boolean_t zvol_os_is_zvol(const char *path);
-void zvol_os_clear_private(zvol_state_t *zv);
+void zvol_os_remove_minor(zvol_state_t *zv);
void zvol_os_set_disk_ro(zvol_state_t *zv, int flags);
void zvol_os_set_capacity(zvol_state_t *zv, uint64_t capacity);
diff --git a/sys/contrib/openzfs/lib/libicp/Makefile.am b/sys/contrib/openzfs/lib/libicp/Makefile.am
index ce24d13a760f..23adba10bc44 100644
--- a/sys/contrib/openzfs/lib/libicp/Makefile.am
+++ b/sys/contrib/openzfs/lib/libicp/Makefile.am
@@ -69,6 +69,7 @@ nodist_libicp_la_SOURCES += \
module/icp/asm-x86_64/aes/aes_aesni.S \
module/icp/asm-x86_64/modes/gcm_pclmulqdq.S \
module/icp/asm-x86_64/modes/aesni-gcm-x86_64.S \
+ module/icp/asm-x86_64/modes/aesni-gcm-avx2-vaes.S \
module/icp/asm-x86_64/modes/ghash-x86_64.S \
module/icp/asm-x86_64/sha2/sha256-x86_64.S \
module/icp/asm-x86_64/sha2/sha512-x86_64.S \
diff --git a/sys/contrib/openzfs/lib/libspl/include/sys/debug.h b/sys/contrib/openzfs/lib/libspl/include/sys/debug.h
index cced309bd1bb..02f33a68b75b 100644
--- a/sys/contrib/openzfs/lib/libspl/include/sys/debug.h
+++ b/sys/contrib/openzfs/lib/libspl/include/sys/debug.h
@@ -38,4 +38,8 @@
#define __maybe_unused __attribute__((unused))
#endif
+#ifndef __must_check
+#define __must_check __attribute__((warn_unused_result))
+#endif
+
#endif
diff --git a/sys/contrib/openzfs/lib/libspl/include/sys/simd.h b/sys/contrib/openzfs/lib/libspl/include/sys/simd.h
index 1ef24f5a7d39..4772a5416b2e 100644
--- a/sys/contrib/openzfs/lib/libspl/include/sys/simd.h
+++ b/sys/contrib/openzfs/lib/libspl/include/sys/simd.h
@@ -102,7 +102,9 @@ typedef enum cpuid_inst_sets {
AES,
PCLMULQDQ,
MOVBE,
- SHA_NI
+ SHA_NI,
+ VAES,
+ VPCLMULQDQ
} cpuid_inst_sets_t;
/*
@@ -127,6 +129,8 @@ typedef struct cpuid_feature_desc {
#define _AES_BIT (1U << 25)
#define _PCLMULQDQ_BIT (1U << 1)
#define _MOVBE_BIT (1U << 22)
+#define _VAES_BIT (1U << 9)
+#define _VPCLMULQDQ_BIT (1U << 10)
#define _SHA_NI_BIT (1U << 29)
/*
@@ -157,6 +161,8 @@ static const cpuid_feature_desc_t cpuid_features[] = {
[PCLMULQDQ] = {1U, 0U, _PCLMULQDQ_BIT, ECX },
[MOVBE] = {1U, 0U, _MOVBE_BIT, ECX },
[SHA_NI] = {7U, 0U, _SHA_NI_BIT, EBX },
+ [VAES] = {7U, 0U, _VAES_BIT, ECX },
+ [VPCLMULQDQ] = {7U, 0U, _VPCLMULQDQ_BIT, ECX },
};
/*
@@ -231,6 +237,8 @@ CPUID_FEATURE_CHECK(aes, AES);
CPUID_FEATURE_CHECK(pclmulqdq, PCLMULQDQ);
CPUID_FEATURE_CHECK(movbe, MOVBE);
CPUID_FEATURE_CHECK(shani, SHA_NI);
+CPUID_FEATURE_CHECK(vaes, VAES);
+CPUID_FEATURE_CHECK(vpclmulqdq, VPCLMULQDQ);
/*
* Detect register set support
@@ -382,6 +390,24 @@ zfs_shani_available(void)
}
/*
+ * Check if VAES instruction is available
+ */
+static inline boolean_t
+zfs_vaes_available(void)
+{
+ return (__cpuid_has_vaes());
+}
+
+/*
+ * Check if VPCLMULQDQ instruction is available
+ */
+static inline boolean_t
+zfs_vpclmulqdq_available(void)
+{
+ return (__cpuid_has_vpclmulqdq());
+}
+
+/*
* AVX-512 family of instruction sets:
*
* AVX512F Foundation
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_diff.c b/sys/contrib/openzfs/lib/libzfs/libzfs_diff.c
index 6aa0375f98d7..5f50bce531f7 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs_diff.c
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs_diff.c
@@ -81,7 +81,7 @@ get_stats_for_obj(differ_info_t *di, const char *dsname, uint64_t obj,
/* we can get stats even if we failed to get a path */
(void) memcpy(sb, &zc.zc_stat, sizeof (zfs_stat_t));
if (error == 0) {
- ASSERT(di->zerr == 0);
+ ASSERT0(di->zerr);
(void) strlcpy(pn, zc.zc_value, maxlen);
return (0);
}
@@ -404,7 +404,7 @@ write_free_diffs(FILE *fp, differ_info_t *di, dmu_diff_record_t *dr)
(void) strlcpy(zc.zc_name, di->fromsnap, sizeof (zc.zc_name));
zc.zc_obj = dr->ddr_first - 1;
- ASSERT(di->zerr == 0);
+ ASSERT0(di->zerr);
while (zc.zc_obj < dr->ddr_last) {
int err;
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_import.c b/sys/contrib/openzfs/lib/libzfs/libzfs_import.c
index 599e8e6f7819..7f276e9592c9 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs_import.c
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs_import.c
@@ -122,7 +122,7 @@ const pool_config_ops_t libzfs_config_ops = {
static uint64_t
label_offset(uint64_t size, int l)
{
- ASSERT(P2PHASE_TYPED(size, sizeof (vdev_label_t), uint64_t) == 0);
+ ASSERT0(P2PHASE_TYPED(size, sizeof (vdev_label_t), uint64_t));
return (l * sizeof (vdev_label_t) + (l < VDEV_LABELS / 2 ?
0 : size - VDEV_LABELS * sizeof (vdev_label_t)));
}
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_mount.c b/sys/contrib/openzfs/lib/libzfs/libzfs_mount.c
index 2a81b658d342..5c9e2199eed4 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs_mount.c
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs_mount.c
@@ -516,7 +516,7 @@ zfs_mount_at(zfs_handle_t *zhp, const char *options, int flags,
} else if (rc == ENOTSUP) {
int spa_version;
- VERIFY(zfs_spa_version(zhp, &spa_version) == 0);
+ VERIFY0(zfs_spa_version(zhp, &spa_version));
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Can't mount a version %llu "
"file system on a version %d pool. Pool must be"
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_sendrecv.c b/sys/contrib/openzfs/lib/libzfs/libzfs_sendrecv.c
index 1ad10ebb3c15..77134d197904 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs_sendrecv.c
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs_sendrecv.c
@@ -2505,7 +2505,7 @@ zfs_send_cb_impl(zfs_handle_t *zhp, const char *fromsnap, const char *tosnap,
err = ENOENT;
if (sdd.cleanup_fd != -1) {
- VERIFY(0 == close(sdd.cleanup_fd));
+ VERIFY0(close(sdd.cleanup_fd));
sdd.cleanup_fd = -1;
}
@@ -2531,7 +2531,7 @@ err_out:
fnvlist_free(sdd.snapholds);
if (sdd.cleanup_fd != -1)
- VERIFY(0 == close(sdd.cleanup_fd));
+ VERIFY0(close(sdd.cleanup_fd));
return (err);
}
@@ -5108,7 +5108,7 @@ zfs_receive_one(libzfs_handle_t *hdl, int infd, const char *tosnap,
nvlist_t *holds, *errors = NULL;
int cleanup_fd = -1;
- VERIFY(0 == nvlist_alloc(&holds, 0, KM_SLEEP));
+ VERIFY0(nvlist_alloc(&holds, 0, KM_SLEEP));
for (pair = nvlist_next_nvpair(snapholds_nvlist, NULL);
pair != NULL;
pair = nvlist_next_nvpair(snapholds_nvlist, pair)) {
@@ -5560,7 +5560,7 @@ zfs_receive_impl(libzfs_handle_t *hdl, const char *tosnap,
if ((cp = strchr(nonpackage_sendfs, '@')) != NULL)
*cp = '\0';
sendfs = nonpackage_sendfs;
- VERIFY(finalsnap == NULL);
+ VERIFY0P(finalsnap);
}
return (zfs_receive_one(hdl, infd, tosnap, originsnap, flags,
&drr, &drr_noswap, sendfs, stream_nv, stream_avl, top_zfs,
diff --git a/sys/contrib/openzfs/lib/libzpool/abd_os.c b/sys/contrib/openzfs/lib/libzpool/abd_os.c
index 0d5795de143a..8bd7a64ab24a 100644
--- a/sys/contrib/openzfs/lib/libzpool/abd_os.c
+++ b/sys/contrib/openzfs/lib/libzpool/abd_os.c
@@ -302,7 +302,7 @@ abd_iter_at_end(struct abd_iter *aiter)
void
abd_iter_advance(struct abd_iter *aiter, size_t amount)
{
- ASSERT3P(aiter->iter_mapaddr, ==, NULL);
+ ASSERT0P(aiter->iter_mapaddr);
ASSERT0(aiter->iter_mapsize);
if (abd_iter_at_end(aiter))
@@ -315,7 +315,7 @@ abd_iter_advance(struct abd_iter *aiter, size_t amount)
void
abd_iter_map(struct abd_iter *aiter)
{
- ASSERT3P(aiter->iter_mapaddr, ==, NULL);
+ ASSERT0P(aiter->iter_mapaddr);
ASSERT0(aiter->iter_mapsize);
if (abd_iter_at_end(aiter))
diff --git a/sys/contrib/openzfs/lib/libzpool/kernel.c b/sys/contrib/openzfs/lib/libzpool/kernel.c
index 2e8bf160465a..fea2f81458f9 100644
--- a/sys/contrib/openzfs/lib/libzpool/kernel.c
+++ b/sys/contrib/openzfs/lib/libzpool/kernel.c
@@ -38,6 +38,7 @@
#include <sys/processor.h>
#include <sys/rrwlock.h>
#include <sys/spa.h>
+#include <sys/spa_impl.h>
#include <sys/stat.h>
#include <sys/systeminfo.h>
#include <sys/time.h>
@@ -369,7 +370,7 @@ cv_timedwait(kcondvar_t *cv, kmutex_t *mp, clock_t abstime)
if (delta <= 0)
return (-1);
- VERIFY(gettimeofday(&tv, NULL) == 0);
+ VERIFY0(gettimeofday(&tv, NULL));
ts.tv_sec = tv.tv_sec + delta / hz;
ts.tv_nsec = tv.tv_usec * NSEC_PER_USEC + (delta % hz) * (NANOSEC / hz);
@@ -811,6 +812,79 @@ umem_out_of_memory(void)
return (0);
}
+static void
+spa_config_load(void)
+{
+ void *buf = NULL;
+ nvlist_t *nvlist, *child;
+ nvpair_t *nvpair;
+ char *pathname;
+ zfs_file_t *fp;
+ zfs_file_attr_t zfa;
+ uint64_t fsize;
+ int err;
+
+ /*
+ * Open the configuration file.
+ */
+ pathname = kmem_alloc(MAXPATHLEN, KM_SLEEP);
+
+ (void) snprintf(pathname, MAXPATHLEN, "%s", spa_config_path);
+
+ err = zfs_file_open(pathname, O_RDONLY, 0, &fp);
+ if (err)
+ err = zfs_file_open(ZPOOL_CACHE_BOOT, O_RDONLY, 0, &fp);
+
+ kmem_free(pathname, MAXPATHLEN);
+
+ if (err)
+ return;
+
+ if (zfs_file_getattr(fp, &zfa))
+ goto out;
+
+ fsize = zfa.zfa_size;
+ buf = kmem_alloc(fsize, KM_SLEEP);
+
+ /*
+ * Read the nvlist from the file.
+ */
+ if (zfs_file_read(fp, buf, fsize, NULL) < 0)
+ goto out;
+
+ /*
+ * Unpack the nvlist.
+ */
+ if (nvlist_unpack(buf, fsize, &nvlist, KM_SLEEP) != 0)
+ goto out;
+
+ /*
+ * Iterate over all elements in the nvlist, creating a new spa_t for
+ * each one with the specified configuration.
+ */
+ mutex_enter(&spa_namespace_lock);
+ nvpair = NULL;
+ while ((nvpair = nvlist_next_nvpair(nvlist, nvpair)) != NULL) {
+ if (nvpair_type(nvpair) != DATA_TYPE_NVLIST)
+ continue;
+
+ child = fnvpair_value_nvlist(nvpair);
+
+ if (spa_lookup(nvpair_name(nvpair)) != NULL)
+ continue;
+ (void) spa_add(nvpair_name(nvpair), child, NULL);
+ }
+ mutex_exit(&spa_namespace_lock);
+
+ nvlist_free(nvlist);
+
+out:
+ if (buf != NULL)
+ kmem_free(buf, fsize);
+
+ zfs_file_close(fp);
+}
+
void
kernel_init(int mode)
{
@@ -835,6 +909,7 @@ kernel_init(int mode)
zstd_init();
spa_init((spa_mode_t)mode);
+ spa_config_load();
fletcher_4_init();
diff --git a/sys/contrib/openzfs/lib/libzpool/util.c b/sys/contrib/openzfs/lib/libzpool/util.c
index 1d0d1a1e56d9..66d6f43967d5 100644
--- a/sys/contrib/openzfs/lib/libzpool/util.c
+++ b/sys/contrib/openzfs/lib/libzpool/util.c
@@ -137,12 +137,10 @@ show_pool_stats(spa_t *spa)
nvlist_t *config, *nvroot;
const char *name;
- VERIFY(spa_get_stats(spa_name(spa), &config, NULL, 0) == 0);
+ VERIFY0(spa_get_stats(spa_name(spa), &config, NULL, 0));
- VERIFY(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
- &nvroot) == 0);
- VERIFY(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
- &name) == 0);
+ VERIFY0(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot));
+ VERIFY0(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME, &name));
show_vdev_stats(name, ZPOOL_CONFIG_CHILDREN, nvroot, 0);
show_vdev_stats(NULL, ZPOOL_CONFIG_L2CACHE, nvroot, 0);
diff --git a/sys/contrib/openzfs/lib/libzutil/zutil_import.c b/sys/contrib/openzfs/lib/libzutil/zutil_import.c
index ccdc874076c3..08367f4c064d 100644
--- a/sys/contrib/openzfs/lib/libzutil/zutil_import.c
+++ b/sys/contrib/openzfs/lib/libzutil/zutil_import.c
@@ -917,7 +917,7 @@ error:
static uint64_t
label_offset(uint64_t size, int l)
{
- ASSERT(P2PHASE_TYPED(size, sizeof (vdev_label_t), uint64_t) == 0);
+ ASSERT0(P2PHASE_TYPED(size, sizeof (vdev_label_t), uint64_t));
return (l * sizeof (vdev_label_t) + (l < VDEV_LABELS / 2 ?
0 : size - VDEV_LABELS * sizeof (vdev_label_t)));
}
@@ -1769,7 +1769,7 @@ zpool_find_import_cached(libpc_handle_t *hdl, importargs_t *iarg)
fnvlist_add_nvlist(pools, nvpair_name(pair),
fnvpair_value_nvlist(pair));
- VERIFY3P(nvlist_next_nvpair(nv, pair), ==, NULL);
+ VERIFY0P(nvlist_next_nvpair(nv, pair));
iarg->guid = saved_guid;
iarg->poolname = saved_poolname;
@@ -1903,30 +1903,43 @@ zpool_find_config(libpc_handle_t *hdl, const char *target, nvlist_t **configp,
*sepp = '\0';
pools = zpool_search_import(hdl, args);
+ if (pools == NULL) {
+ zutil_error_aux(hdl, dgettext(TEXT_DOMAIN, "no pools found"));
+ (void) zutil_error_fmt(hdl, LPC_UNKNOWN, dgettext(TEXT_DOMAIN,
+ "failed to find config for pool '%s'"), targetdup);
+ free(targetdup);
+ return (ENOENT);
+ }
- if (pools != NULL) {
- nvpair_t *elem = NULL;
- while ((elem = nvlist_next_nvpair(pools, elem)) != NULL) {
- VERIFY0(nvpair_value_nvlist(elem, &config));
- if (pool_match(config, targetdup)) {
- count++;
- if (match != NULL) {
- /* multiple matches found */
- continue;
- } else {
- match = fnvlist_dup(config);
- }
+ nvpair_t *elem = NULL;
+ while ((elem = nvlist_next_nvpair(pools, elem)) != NULL) {
+ VERIFY0(nvpair_value_nvlist(elem, &config));
+ if (pool_match(config, targetdup)) {
+ count++;
+ if (match != NULL) {
+ /* multiple matches found */
+ continue;
+ } else {
+ match = fnvlist_dup(config);
}
}
- fnvlist_free(pools);
}
+ fnvlist_free(pools);
if (count == 0) {
+ zutil_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "no matching pools"));
+ (void) zutil_error_fmt(hdl, LPC_UNKNOWN, dgettext(TEXT_DOMAIN,
+ "failed to find config for pool '%s'"), targetdup);
free(targetdup);
return (ENOENT);
}
if (count > 1) {
+ zutil_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "more than one matching pool"));
+ (void) zutil_error_fmt(hdl, LPC_UNKNOWN, dgettext(TEXT_DOMAIN,
+ "failed to find config for pool '%s'"), targetdup);
free(targetdup);
fnvlist_free(match);
return (EINVAL);
diff --git a/sys/contrib/openzfs/man/man4/zfs.4 b/sys/contrib/openzfs/man/man4/zfs.4
index 4a5f9fd93f4f..5c7958667f92 100644
--- a/sys/contrib/openzfs/man/man4/zfs.4
+++ b/sys/contrib/openzfs/man/man4/zfs.4
@@ -941,10 +941,6 @@ The target number of bytes the ARC should leave as free memory on the system.
If zero, equivalent to the bigger of
.Sy 512 KiB No and Sy all_system_memory/64 .
.
-.It Sy zfs_autoimport_disable Ns = Ns Sy 1 Ns | Ns 0 Pq int
-Disable pool import at module load by ignoring the cache file
-.Pq Sy spa_config_path .
-.
.It Sy zfs_checksum_events_per_second Ns = Ns Sy 20 Ns /s Pq uint
Rate limit checksum events to this many per second.
Note that this should not be set below the ZED thresholds
diff --git a/sys/contrib/openzfs/man/man8/zdb.8 b/sys/contrib/openzfs/man/man8/zdb.8
index 3984aaac5866..0a5b6af73fdb 100644
--- a/sys/contrib/openzfs/man/man8/zdb.8
+++ b/sys/contrib/openzfs/man/man8/zdb.8
@@ -15,7 +15,7 @@
.\" Copyright (c) 2017 Lawrence Livermore National Security, LLC.
.\" Copyright (c) 2017 Intel Corporation.
.\"
-.Dd October 27, 2024
+.Dd April 23, 2025
.Dt ZDB 8
.Os
.
@@ -531,6 +531,18 @@ option, with more occurrences enabling more verbosity.
If no options are specified, all information about the named pool will be
displayed at default verbosity.
.
+.Sh EXIT STATUS
+The
+.Nm
+utility exits
+.Sy 0
+on success,
+.Sy 1
+if a fatal error occurs,
+.Sy 2
+if invalid command line options were specified, or
+.Sy 3
+if on-disk corruption was detected, but was not fatal.
.Sh EXAMPLES
.Ss Example 1 : No Display the configuration of imported pool Ar rpool
.Bd -literal
diff --git a/sys/contrib/openzfs/man/man8/zfs-send.8 b/sys/contrib/openzfs/man/man8/zfs-send.8
index c920a5a48798..f7c6b840303c 100644
--- a/sys/contrib/openzfs/man/man8/zfs-send.8
+++ b/sys/contrib/openzfs/man/man8/zfs-send.8
@@ -173,8 +173,10 @@ The receiving system must have the
feature enabled.
If the
.Sy lz4_compress
-feature is active on the sending system, then the receiving system must have
-that feature enabled as well.
+or
+.Sy zstd_compress
+features are active on the sending system, then the receiving system must have
+the corresponding features enabled as well.
Datasets that are sent with this flag may not be
received as an encrypted dataset, since encrypted datasets cannot use the
.Sy embedded_data
@@ -201,8 +203,10 @@ property for details
.Pc .
If the
.Sy lz4_compress
-feature is active on the sending system, then the receiving system must have
-that feature enabled as well.
+or
+.Sy zstd_compress
+features are active on the sending system, then the receiving system must have
+the corresponding features enabled as well.
If the
.Sy large_blocks
feature is enabled on the sending system but the
@@ -357,8 +361,10 @@ property for details
.Pc .
If the
.Sy lz4_compress
-feature is active on the sending system, then the receiving system must have
-that feature enabled as well.
+or
+.Sy zstd_compress
+features are active on the sending system, then the receiving system must have
+the corresponding features enabled as well.
If the
.Sy large_blocks
feature is enabled on the sending system but the
@@ -400,8 +406,10 @@ The receiving system must have the
feature enabled.
If the
.Sy lz4_compress
-feature is active on the sending system, then the receiving system must have
-that feature enabled as well.
+or
+.Sy zstd_compress
+features are active on the sending system, then the receiving system must have
+the corresponding features enabled as well.
Datasets that are sent with this flag may not be received as an encrypted
dataset,
since encrypted datasets cannot use the
diff --git a/sys/contrib/openzfs/module/Kbuild.in b/sys/contrib/openzfs/module/Kbuild.in
index 3d6f288fa5da..362d2295e091 100644
--- a/sys/contrib/openzfs/module/Kbuild.in
+++ b/sys/contrib/openzfs/module/Kbuild.in
@@ -135,6 +135,7 @@ ICP_OBJS_X86_64 := \
asm-x86_64/sha2/sha256-x86_64.o \
asm-x86_64/sha2/sha512-x86_64.o \
asm-x86_64/modes/aesni-gcm-x86_64.o \
+ asm-x86_64/modes/aesni-gcm-avx2-vaes.o \
asm-x86_64/modes/gcm_pclmulqdq.o \
asm-x86_64/modes/ghash-x86_64.o
diff --git a/sys/contrib/openzfs/module/avl/avl.c b/sys/contrib/openzfs/module/avl/avl.c
index b6c1c02bc3f2..67cbcd3adeec 100644
--- a/sys/contrib/openzfs/module/avl/avl.c
+++ b/sys/contrib/openzfs/module/avl/avl.c
@@ -225,7 +225,7 @@ avl_nearest(avl_tree_t *tree, avl_index_t where, int direction)
size_t off = tree->avl_offset;
if (node == NULL) {
- ASSERT(tree->avl_root == NULL);
+ ASSERT0P(tree->avl_root);
return (NULL);
}
data = AVL_NODE2DATA(node, off);
@@ -478,7 +478,7 @@ avl_insert(avl_tree_t *tree, void *new_data, avl_index_t where)
size_t off = tree->avl_offset;
#ifdef _LP64
- ASSERT(((uintptr_t)new_data & 0x7) == 0);
+ ASSERT0(((uintptr_t)new_data & 0x7));
#endif
node = AVL_DATA2NODE(new_data, off);
@@ -495,10 +495,10 @@ avl_insert(avl_tree_t *tree, void *new_data, avl_index_t where)
AVL_SETBALANCE(node, 0);
AVL_SETPARENT(node, parent);
if (parent != NULL) {
- ASSERT(parent->avl_child[which_child] == NULL);
+ ASSERT0P(parent->avl_child[which_child]);
parent->avl_child[which_child] = node;
} else {
- ASSERT(tree->avl_root == NULL);
+ ASSERT0P(tree->avl_root);
tree->avl_root = node;
}
/*
@@ -608,7 +608,7 @@ avl_insert_here(
ASSERT(diff > 0 ? child == 1 : child == 0);
#endif
}
- ASSERT(node->avl_child[child] == NULL);
+ ASSERT0P(node->avl_child[child]);
avl_insert(tree, new_data, AVL_MKINDEX(node, child));
}
@@ -881,7 +881,7 @@ avl_create(avl_tree_t *tree, int (*compar) (const void *, const void *),
ASSERT(size > 0);
ASSERT(size >= offset + sizeof (avl_node_t));
#ifdef _LP64
- ASSERT((offset & 0x7) == 0);
+ ASSERT0((offset & 0x7));
#endif
tree->avl_compar = compar;
@@ -897,8 +897,8 @@ void
avl_destroy(avl_tree_t *tree)
{
ASSERT(tree);
- ASSERT(tree->avl_numnodes == 0);
- ASSERT(tree->avl_root == NULL);
+ ASSERT0(tree->avl_numnodes);
+ ASSERT0P(tree->avl_root);
}
diff --git a/sys/contrib/openzfs/module/icp/algs/modes/gcm.c b/sys/contrib/openzfs/module/icp/algs/modes/gcm.c
index c2a982b5a376..3cfa5b8165ce 100644
--- a/sys/contrib/openzfs/module/icp/algs/modes/gcm.c
+++ b/sys/contrib/openzfs/module/icp/algs/modes/gcm.c
@@ -46,6 +46,9 @@
#define IMPL_CYCLE (UINT32_MAX-1)
#ifdef CAN_USE_GCM_ASM
#define IMPL_AVX (UINT32_MAX-2)
+#if CAN_USE_GCM_ASM >= 2
+#define IMPL_AVX2 (UINT32_MAX-3)
+#endif
#endif
#define GCM_IMPL_READ(i) (*(volatile uint32_t *) &(i))
static uint32_t icp_gcm_impl = IMPL_FASTEST;
@@ -56,17 +59,16 @@ static uint32_t user_sel_impl = IMPL_FASTEST;
boolean_t gcm_avx_can_use_movbe = B_FALSE;
/*
* Whether to use the optimized openssl gcm and ghash implementations.
- * Set to true if module parameter icp_gcm_impl == "avx".
*/
-static boolean_t gcm_use_avx = B_FALSE;
-#define GCM_IMPL_USE_AVX (*(volatile boolean_t *)&gcm_use_avx)
+static gcm_impl gcm_impl_used = GCM_IMPL_GENERIC;
+#define GCM_IMPL_USED (*(volatile gcm_impl *)&gcm_impl_used)
extern boolean_t ASMABI atomic_toggle_boolean_nv(volatile boolean_t *);
static inline boolean_t gcm_avx_will_work(void);
-static inline void gcm_set_avx(boolean_t);
-static inline boolean_t gcm_toggle_avx(void);
-static inline size_t gcm_simd_get_htab_size(boolean_t);
+static inline boolean_t gcm_avx2_will_work(void);
+static inline void gcm_use_impl(gcm_impl impl);
+static inline gcm_impl gcm_toggle_impl(void);
static int gcm_mode_encrypt_contiguous_blocks_avx(gcm_ctx_t *, char *, size_t,
crypto_data_t *, size_t);
@@ -89,7 +91,7 @@ gcm_mode_encrypt_contiguous_blocks(gcm_ctx_t *ctx, char *data, size_t length,
void (*xor_block)(uint8_t *, uint8_t *))
{
#ifdef CAN_USE_GCM_ASM
- if (ctx->gcm_use_avx == B_TRUE)
+ if (ctx->impl != GCM_IMPL_GENERIC)
return (gcm_mode_encrypt_contiguous_blocks_avx(
ctx, data, length, out, block_size));
#endif
@@ -208,7 +210,7 @@ gcm_encrypt_final(gcm_ctx_t *ctx, crypto_data_t *out, size_t block_size,
{
(void) copy_block;
#ifdef CAN_USE_GCM_ASM
- if (ctx->gcm_use_avx == B_TRUE)
+ if (ctx->impl != GCM_IMPL_GENERIC)
return (gcm_encrypt_final_avx(ctx, out, block_size));
#endif
@@ -374,7 +376,7 @@ gcm_decrypt_final(gcm_ctx_t *ctx, crypto_data_t *out, size_t block_size,
void (*xor_block)(uint8_t *, uint8_t *))
{
#ifdef CAN_USE_GCM_ASM
- if (ctx->gcm_use_avx == B_TRUE)
+ if (ctx->impl != GCM_IMPL_GENERIC)
return (gcm_decrypt_final_avx(ctx, out, block_size));
#endif
@@ -631,23 +633,23 @@ gcm_init_ctx(gcm_ctx_t *gcm_ctx, char *param,
((aes_key_t *)gcm_ctx->gcm_keysched)->ops->needs_byteswap;
if (GCM_IMPL_READ(icp_gcm_impl) != IMPL_CYCLE) {
- gcm_ctx->gcm_use_avx = GCM_IMPL_USE_AVX;
+ gcm_ctx->impl = GCM_IMPL_USED;
} else {
/*
- * Handle the "cycle" implementation by creating avx and
- * non-avx contexts alternately.
+ * Handle the "cycle" implementation by creating different
+ * contexts, one per implementation.
*/
- gcm_ctx->gcm_use_avx = gcm_toggle_avx();
+ gcm_ctx->impl = gcm_toggle_impl();
- /* The avx impl. doesn't handle byte swapped key schedules. */
- if (gcm_ctx->gcm_use_avx == B_TRUE && needs_bswap == B_TRUE) {
- gcm_ctx->gcm_use_avx = B_FALSE;
+ /* The AVX impl. doesn't handle byte swapped key schedules. */
+ if (needs_bswap == B_TRUE) {
+ gcm_ctx->impl = GCM_IMPL_GENERIC;
}
/*
- * If this is a GCM context, use the MOVBE and the BSWAP
+ * If this is an AVX context, use the MOVBE and the BSWAP
* variants alternately.
*/
- if (gcm_ctx->gcm_use_avx == B_TRUE &&
+ if (gcm_ctx->impl == GCM_IMPL_AVX &&
zfs_movbe_available() == B_TRUE) {
(void) atomic_toggle_boolean_nv(
(volatile boolean_t *)&gcm_avx_can_use_movbe);
@@ -658,12 +660,13 @@ gcm_init_ctx(gcm_ctx_t *gcm_ctx, char *param,
* still they could be created by the aes generic implementation.
* Make sure not to use them since we'll corrupt data if we do.
*/
- if (gcm_ctx->gcm_use_avx == B_TRUE && needs_bswap == B_TRUE) {
- gcm_ctx->gcm_use_avx = B_FALSE;
+ if (gcm_ctx->impl != GCM_IMPL_GENERIC && needs_bswap == B_TRUE) {
+ gcm_ctx->impl = GCM_IMPL_GENERIC;
cmn_err_once(CE_WARN,
"ICP: Can't use the aes generic or cycle implementations "
- "in combination with the gcm avx implementation!");
+ "in combination with the gcm avx or avx2-vaes "
+ "implementation!");
cmn_err_once(CE_WARN,
"ICP: Falling back to a compatible implementation, "
"aes-gcm performance will likely be degraded.");
@@ -672,36 +675,20 @@ gcm_init_ctx(gcm_ctx_t *gcm_ctx, char *param,
"restore performance.");
}
- /* Allocate Htab memory as needed. */
- if (gcm_ctx->gcm_use_avx == B_TRUE) {
- size_t htab_len = gcm_simd_get_htab_size(gcm_ctx->gcm_use_avx);
-
- if (htab_len == 0) {
- return (CRYPTO_MECHANISM_PARAM_INVALID);
- }
- gcm_ctx->gcm_htab_len = htab_len;
- gcm_ctx->gcm_Htable =
- kmem_alloc(htab_len, KM_SLEEP);
-
- if (gcm_ctx->gcm_Htable == NULL) {
- return (CRYPTO_HOST_MEMORY);
- }
+ /*
+ * AVX implementations use Htable with sizes depending on
+ * implementation.
+ */
+ if (gcm_ctx->impl != GCM_IMPL_GENERIC) {
+ rv = gcm_init_avx(gcm_ctx, iv, iv_len, aad, aad_len,
+ block_size);
}
- /* Avx and non avx context initialization differs from here on. */
- if (gcm_ctx->gcm_use_avx == B_FALSE) {
+ else
#endif /* ifdef CAN_USE_GCM_ASM */
- if (gcm_init(gcm_ctx, iv, iv_len, aad, aad_len, block_size,
- encrypt_block, copy_block, xor_block) != CRYPTO_SUCCESS) {
- rv = CRYPTO_MECHANISM_PARAM_INVALID;
- }
-#ifdef CAN_USE_GCM_ASM
- } else {
- if (gcm_init_avx(gcm_ctx, iv, iv_len, aad, aad_len,
- block_size) != CRYPTO_SUCCESS) {
- rv = CRYPTO_MECHANISM_PARAM_INVALID;
- }
+ if (gcm_init(gcm_ctx, iv, iv_len, aad, aad_len, block_size,
+ encrypt_block, copy_block, xor_block) != CRYPTO_SUCCESS) {
+ rv = CRYPTO_MECHANISM_PARAM_INVALID;
}
-#endif /* ifdef CAN_USE_GCM_ASM */
return (rv);
}
@@ -767,6 +754,9 @@ gcm_impl_get_ops(void)
break;
#ifdef CAN_USE_GCM_ASM
case IMPL_AVX:
+#if CAN_USE_GCM_ASM >= 2
+ case IMPL_AVX2:
+#endif
/*
* Make sure that we return a valid implementation while
* switching to the avx implementation since there still
@@ -828,6 +818,13 @@ gcm_impl_init(void)
* Use the avx implementation if it's available and the implementation
* hasn't changed from its default value of fastest on module load.
*/
+#if CAN_USE_GCM_ASM >= 2
+ if (gcm_avx2_will_work()) {
+ if (GCM_IMPL_READ(user_sel_impl) == IMPL_FASTEST) {
+ gcm_use_impl(GCM_IMPL_AVX2);
+ }
+ } else
+#endif
if (gcm_avx_will_work()) {
#ifdef HAVE_MOVBE
if (zfs_movbe_available() == B_TRUE) {
@@ -835,7 +832,7 @@ gcm_impl_init(void)
}
#endif
if (GCM_IMPL_READ(user_sel_impl) == IMPL_FASTEST) {
- gcm_set_avx(B_TRUE);
+ gcm_use_impl(GCM_IMPL_AVX);
}
}
#endif
@@ -852,6 +849,7 @@ static const struct {
{ "fastest", IMPL_FASTEST },
#ifdef CAN_USE_GCM_ASM
{ "avx", IMPL_AVX },
+ { "avx2-vaes", IMPL_AVX2 },
#endif
};
@@ -887,7 +885,13 @@ gcm_impl_set(const char *val)
/* Check mandatory options */
for (i = 0; i < ARRAY_SIZE(gcm_impl_opts); i++) {
#ifdef CAN_USE_GCM_ASM
+#if CAN_USE_GCM_ASM >= 2
/* Ignore avx implementation if it won't work. */
+ if (gcm_impl_opts[i].sel == IMPL_AVX2 &&
+ !gcm_avx2_will_work()) {
+ continue;
+ }
+#endif
if (gcm_impl_opts[i].sel == IMPL_AVX && !gcm_avx_will_work()) {
continue;
}
@@ -915,11 +919,17 @@ gcm_impl_set(const char *val)
* Use the avx implementation if available and the requested one is
* avx or fastest.
*/
+#if CAN_USE_GCM_ASM >= 2
+ if (gcm_avx2_will_work() == B_TRUE &&
+ (impl == IMPL_AVX2 || impl == IMPL_FASTEST)) {
+ gcm_use_impl(GCM_IMPL_AVX2);
+ } else
+#endif
if (gcm_avx_will_work() == B_TRUE &&
(impl == IMPL_AVX || impl == IMPL_FASTEST)) {
- gcm_set_avx(B_TRUE);
+ gcm_use_impl(GCM_IMPL_AVX);
} else {
- gcm_set_avx(B_FALSE);
+ gcm_use_impl(GCM_IMPL_GENERIC);
}
#endif
@@ -952,6 +962,12 @@ icp_gcm_impl_get(char *buffer, zfs_kernel_param_t *kp)
for (i = 0; i < ARRAY_SIZE(gcm_impl_opts); i++) {
#ifdef CAN_USE_GCM_ASM
/* Ignore avx implementation if it won't work. */
+#if CAN_USE_GCM_ASM >= 2
+ if (gcm_impl_opts[i].sel == IMPL_AVX2 &&
+ !gcm_avx2_will_work()) {
+ continue;
+ }
+#endif
if (gcm_impl_opts[i].sel == IMPL_AVX && !gcm_avx_will_work()) {
continue;
}
@@ -993,9 +1009,6 @@ MODULE_PARM_DESC(icp_gcm_impl, "Select gcm implementation.");
/* Clear the FPU registers since they hold sensitive internal state. */
#define clear_fpu_regs() clear_fpu_regs_avx()
-#define GHASH_AVX(ctx, in, len) \
- gcm_ghash_avx((ctx)->gcm_ghash, (const uint64_t *)(ctx)->gcm_Htable, \
- in, len)
#define gcm_incr_counter_block(ctx) gcm_incr_counter_block_by(ctx, 1)
@@ -1010,20 +1023,77 @@ MODULE_PARM_DESC(icp_gcm_impl, "Select gcm implementation.");
static uint32_t gcm_avx_chunk_size =
((32 * 1024) / GCM_AVX_MIN_DECRYPT_BYTES) * GCM_AVX_MIN_DECRYPT_BYTES;
+/*
+ * GCM definitions: uint128_t is copied from include/crypto/modes.h
+ * Avoiding u128 because it is already defined in kernel sources.
+ */
+typedef struct {
+ uint64_t hi, lo;
+} uint128_t;
+
extern void ASMABI clear_fpu_regs_avx(void);
extern void ASMABI gcm_xor_avx(const uint8_t *src, uint8_t *dst);
extern void ASMABI aes_encrypt_intel(const uint32_t rk[], int nr,
const uint32_t pt[4], uint32_t ct[4]);
extern void ASMABI gcm_init_htab_avx(uint64_t *Htable, const uint64_t H[2]);
+#if CAN_USE_GCM_ASM >= 2
+extern void ASMABI gcm_init_vpclmulqdq_avx2(uint128_t Htable[16],
+ const uint64_t H[2]);
+#endif
extern void ASMABI gcm_ghash_avx(uint64_t ghash[2], const uint64_t *Htable,
const uint8_t *in, size_t len);
+#if CAN_USE_GCM_ASM >= 2
+extern void ASMABI gcm_ghash_vpclmulqdq_avx2(uint64_t ghash[2],
+ const uint64_t *Htable, const uint8_t *in, size_t len);
+#endif
+static inline void GHASH_AVX(gcm_ctx_t *ctx, const uint8_t *in, size_t len)
+{
+ switch (ctx->impl) {
+#if CAN_USE_GCM_ASM >= 2
+ case GCM_IMPL_AVX2:
+ gcm_ghash_vpclmulqdq_avx2(ctx->gcm_ghash,
+ (const uint64_t *)ctx->gcm_Htable, in, len);
+ break;
+#endif
+
+ case GCM_IMPL_AVX:
+ gcm_ghash_avx(ctx->gcm_ghash,
+ (const uint64_t *)ctx->gcm_Htable, in, len);
+ break;
+
+ default:
+ VERIFY(B_FALSE);
+ }
+}
+typedef size_t ASMABI aesni_gcm_encrypt_impl(const uint8_t *, uint8_t *,
+ size_t, const void *, uint64_t *, const uint64_t *Htable, uint64_t *);
extern size_t ASMABI aesni_gcm_encrypt(const uint8_t *, uint8_t *, size_t,
const void *, uint64_t *, uint64_t *);
+#if CAN_USE_GCM_ASM >= 2
+extern void ASMABI aes_gcm_enc_update_vaes_avx2(const uint8_t *in,
+ uint8_t *out, size_t len, const void *key, const uint8_t ivec[16],
+ const uint128_t Htable[16], uint8_t Xi[16]);
+#endif
+typedef size_t ASMABI aesni_gcm_decrypt_impl(const uint8_t *, uint8_t *,
+ size_t, const void *, uint64_t *, const uint64_t *Htable, uint64_t *);
extern size_t ASMABI aesni_gcm_decrypt(const uint8_t *, uint8_t *, size_t,
const void *, uint64_t *, uint64_t *);
+#if CAN_USE_GCM_ASM >= 2
+extern void ASMABI aes_gcm_dec_update_vaes_avx2(const uint8_t *in,
+ uint8_t *out, size_t len, const void *key, const uint8_t ivec[16],
+ const uint128_t Htable[16], uint8_t Xi[16]);
+#endif
+
+static inline boolean_t
+gcm_avx2_will_work(void)
+{
+ return (kfpu_allowed() &&
+ zfs_avx2_available() && zfs_vaes_available() &&
+ zfs_vpclmulqdq_available());
+}
static inline boolean_t
gcm_avx_will_work(void)
@@ -1035,33 +1105,67 @@ gcm_avx_will_work(void)
}
static inline void
-gcm_set_avx(boolean_t val)
+gcm_use_impl(gcm_impl impl)
{
- if (gcm_avx_will_work() == B_TRUE) {
- atomic_swap_32(&gcm_use_avx, val);
+ switch (impl) {
+#if CAN_USE_GCM_ASM >= 2
+ case GCM_IMPL_AVX2:
+ if (gcm_avx2_will_work() == B_TRUE) {
+ atomic_swap_32(&gcm_impl_used, impl);
+ return;
+ }
+
+ zfs_fallthrough;
+#endif
+
+ case GCM_IMPL_AVX:
+ if (gcm_avx_will_work() == B_TRUE) {
+ atomic_swap_32(&gcm_impl_used, impl);
+ return;
+ }
+
+ zfs_fallthrough;
+
+ default:
+ atomic_swap_32(&gcm_impl_used, GCM_IMPL_GENERIC);
}
}
static inline boolean_t
-gcm_toggle_avx(void)
+gcm_impl_will_work(gcm_impl impl)
{
- if (gcm_avx_will_work() == B_TRUE) {
- return (atomic_toggle_boolean_nv(&GCM_IMPL_USE_AVX));
- } else {
- return (B_FALSE);
+ switch (impl) {
+#if CAN_USE_GCM_ASM >= 2
+ case GCM_IMPL_AVX2:
+ return (gcm_avx2_will_work());
+#endif
+
+ case GCM_IMPL_AVX:
+ return (gcm_avx_will_work());
+
+ default:
+ return (B_TRUE);
}
}
-static inline size_t
-gcm_simd_get_htab_size(boolean_t simd_mode)
+static inline gcm_impl
+gcm_toggle_impl(void)
{
- switch (simd_mode) {
- case B_TRUE:
- return (2 * 6 * 2 * sizeof (uint64_t));
+ gcm_impl current_impl, new_impl;
+ do { /* handle races */
+ current_impl = atomic_load_32(&gcm_impl_used);
+ new_impl = current_impl;
+ while (B_TRUE) { /* handle incompatble implementations */
+ new_impl = (new_impl + 1) % GCM_IMPL_MAX;
+ if (gcm_impl_will_work(new_impl)) {
+ break;
+ }
+ }
- default:
- return (0);
- }
+ } while (atomic_cas_32(&gcm_impl_used, current_impl, new_impl) !=
+ current_impl);
+
+ return (new_impl);
}
@@ -1077,6 +1181,50 @@ gcm_incr_counter_block_by(gcm_ctx_t *ctx, int n)
ctx->gcm_cb[1] = (ctx->gcm_cb[1] & ~counter_mask) | counter;
}
+static size_t aesni_gcm_encrypt_avx(const uint8_t *in, uint8_t *out,
+ size_t len, const void *key, uint64_t *iv, const uint64_t *Htable,
+ uint64_t *Xip)
+{
+ (void) Htable;
+ return (aesni_gcm_encrypt(in, out, len, key, iv, Xip));
+}
+
+#if CAN_USE_GCM_ASM >= 2
+// kSizeTWithoutLower4Bits is a mask that can be used to zero the lower four
+// bits of a |size_t|.
+// This is from boringssl/crypto/fipsmodule/aes/gcm.cc.inc
+static const size_t kSizeTWithoutLower4Bits = (size_t)-16;
+
+/* The following CRYPTO methods are from boringssl/crypto/internal.h */
+static inline uint32_t CRYPTO_bswap4(uint32_t x) {
+ return (__builtin_bswap32(x));
+}
+
+static inline uint32_t CRYPTO_load_u32_be(const void *in) {
+ uint32_t v;
+ memcpy(&v, in, sizeof (v));
+ return (CRYPTO_bswap4(v));
+}
+
+static inline void CRYPTO_store_u32_be(void *out, uint32_t v) {
+ v = CRYPTO_bswap4(v);
+ memcpy(out, &v, sizeof (v));
+}
+
+static size_t aesni_gcm_encrypt_avx2(const uint8_t *in, uint8_t *out,
+ size_t len, const void *key, uint64_t *iv, const uint64_t *Htable,
+ uint64_t *Xip)
+{
+ uint8_t *ivec = (uint8_t *)iv;
+ len &= kSizeTWithoutLower4Bits;
+ aes_gcm_enc_update_vaes_avx2(in, out, len, key, ivec,
+ (const uint128_t *)Htable, (uint8_t *)Xip);
+ CRYPTO_store_u32_be(&ivec[12],
+ CRYPTO_load_u32_be(&ivec[12]) + len / 16);
+ return (len);
+}
+#endif /* if CAN_USE_GCM_ASM >= 2 */
+
/*
* Encrypt multiple blocks of data in GCM mode.
* This is done in gcm_avx_chunk_size chunks, utilizing AVX assembler routines
@@ -1091,8 +1239,15 @@ gcm_mode_encrypt_contiguous_blocks_avx(gcm_ctx_t *ctx, char *data,
size_t done = 0;
uint8_t *datap = (uint8_t *)data;
size_t chunk_size = (size_t)GCM_CHUNK_SIZE_READ;
+ aesni_gcm_encrypt_impl *encrypt_blocks =
+#if CAN_USE_GCM_ASM >= 2
+ ctx->impl == GCM_IMPL_AVX2 ?
+ aesni_gcm_encrypt_avx2 :
+#endif
+ aesni_gcm_encrypt_avx;
const aes_key_t *key = ((aes_key_t *)ctx->gcm_keysched);
uint64_t *ghash = ctx->gcm_ghash;
+ uint64_t *htable = ctx->gcm_Htable;
uint64_t *cb = ctx->gcm_cb;
uint8_t *ct_buf = NULL;
uint8_t *tmp = (uint8_t *)ctx->gcm_tmp;
@@ -1156,8 +1311,8 @@ gcm_mode_encrypt_contiguous_blocks_avx(gcm_ctx_t *ctx, char *data,
/* Do the bulk encryption in chunk_size blocks. */
for (; bleft >= chunk_size; bleft -= chunk_size) {
kfpu_begin();
- done = aesni_gcm_encrypt(
- datap, ct_buf, chunk_size, key, cb, ghash);
+ done = encrypt_blocks(
+ datap, ct_buf, chunk_size, key, cb, htable, ghash);
clear_fpu_regs();
kfpu_end();
@@ -1180,7 +1335,8 @@ gcm_mode_encrypt_contiguous_blocks_avx(gcm_ctx_t *ctx, char *data,
/* Bulk encrypt the remaining data. */
kfpu_begin();
if (bleft >= GCM_AVX_MIN_ENCRYPT_BYTES) {
- done = aesni_gcm_encrypt(datap, ct_buf, bleft, key, cb, ghash);
+ done = encrypt_blocks(datap, ct_buf, bleft, key, cb, htable,
+ ghash);
if (done == 0) {
rv = CRYPTO_FAILED;
goto out;
@@ -1293,6 +1449,29 @@ gcm_encrypt_final_avx(gcm_ctx_t *ctx, crypto_data_t *out, size_t block_size)
return (CRYPTO_SUCCESS);
}
+static size_t aesni_gcm_decrypt_avx(const uint8_t *in, uint8_t *out,
+ size_t len, const void *key, uint64_t *iv, const uint64_t *Htable,
+ uint64_t *Xip)
+{
+ (void) Htable;
+ return (aesni_gcm_decrypt(in, out, len, key, iv, Xip));
+}
+
+#if CAN_USE_GCM_ASM >= 2
+static size_t aesni_gcm_decrypt_avx2(const uint8_t *in, uint8_t *out,
+ size_t len, const void *key, uint64_t *iv, const uint64_t *Htable,
+ uint64_t *Xip)
+{
+ uint8_t *ivec = (uint8_t *)iv;
+ len &= kSizeTWithoutLower4Bits;
+ aes_gcm_dec_update_vaes_avx2(in, out, len, key, ivec,
+ (const uint128_t *)Htable, (uint8_t *)Xip);
+ CRYPTO_store_u32_be(&ivec[12],
+ CRYPTO_load_u32_be(&ivec[12]) + len / 16);
+ return (len);
+}
+#endif /* if CAN_USE_GCM_ASM >= 2 */
+
/*
* Finalize decryption: We just have accumulated crypto text, so now we
* decrypt it here inplace.
@@ -1306,10 +1485,17 @@ gcm_decrypt_final_avx(gcm_ctx_t *ctx, crypto_data_t *out, size_t block_size)
B_FALSE);
size_t chunk_size = (size_t)GCM_CHUNK_SIZE_READ;
+ aesni_gcm_decrypt_impl *decrypt_blocks =
+#if CAN_USE_GCM_ASM >= 2
+ ctx->impl == GCM_IMPL_AVX2 ?
+ aesni_gcm_decrypt_avx2 :
+#endif
+ aesni_gcm_decrypt_avx;
size_t pt_len = ctx->gcm_processed_data_len - ctx->gcm_tag_len;
uint8_t *datap = ctx->gcm_pt_buf;
const aes_key_t *key = ((aes_key_t *)ctx->gcm_keysched);
uint32_t *cb = (uint32_t *)ctx->gcm_cb;
+ uint64_t *htable = ctx->gcm_Htable;
uint64_t *ghash = ctx->gcm_ghash;
uint32_t *tmp = (uint32_t *)ctx->gcm_tmp;
int rv = CRYPTO_SUCCESS;
@@ -1322,8 +1508,8 @@ gcm_decrypt_final_avx(gcm_ctx_t *ctx, crypto_data_t *out, size_t block_size)
*/
for (bleft = pt_len; bleft >= chunk_size; bleft -= chunk_size) {
kfpu_begin();
- done = aesni_gcm_decrypt(datap, datap, chunk_size,
- (const void *)key, ctx->gcm_cb, ghash);
+ done = decrypt_blocks(datap, datap, chunk_size,
+ (const void *)key, ctx->gcm_cb, htable, ghash);
clear_fpu_regs();
kfpu_end();
if (done != chunk_size) {
@@ -1334,8 +1520,8 @@ gcm_decrypt_final_avx(gcm_ctx_t *ctx, crypto_data_t *out, size_t block_size)
/* Decrypt remainder, which is less than chunk size, in one go. */
kfpu_begin();
if (bleft >= GCM_AVX_MIN_DECRYPT_BYTES) {
- done = aesni_gcm_decrypt(datap, datap, bleft,
- (const void *)key, ctx->gcm_cb, ghash);
+ done = decrypt_blocks(datap, datap, bleft,
+ (const void *)key, ctx->gcm_cb, htable, ghash);
if (done == 0) {
clear_fpu_regs();
kfpu_end();
@@ -1424,13 +1610,42 @@ gcm_init_avx(gcm_ctx_t *ctx, const uint8_t *iv, size_t iv_len,
ASSERT3S(((aes_key_t *)ctx->gcm_keysched)->ops->needs_byteswap, ==,
B_FALSE);
+ size_t htab_len = 0;
+#if CAN_USE_GCM_ASM >= 2
+ if (ctx->impl == GCM_IMPL_AVX2) {
+ /*
+ * BoringSSL's API specifies uint128_t[16] for htab; but only
+ * uint128_t[12] are used.
+ * See https://github.com/google/boringssl/blob/
+ * 813840dd094f9e9c1b00a7368aa25e656554221f1/crypto/fipsmodule/
+ * modes/asm/aes-gcm-avx2-x86_64.pl#L198-L200
+ */
+ htab_len = (2 * 8 * sizeof (uint128_t));
+ } else
+#endif /* CAN_USE_GCM_ASM >= 2 */
+ {
+ htab_len = (2 * 6 * sizeof (uint128_t));
+ }
+
+ ctx->gcm_Htable = kmem_alloc(htab_len, KM_SLEEP);
+ if (ctx->gcm_Htable == NULL) {
+ return (CRYPTO_HOST_MEMORY);
+ }
+
/* Init H (encrypt zero block) and create the initial counter block. */
memset(H, 0, sizeof (ctx->gcm_H));
kfpu_begin();
aes_encrypt_intel(keysched, aes_rounds,
(const uint32_t *)H, (uint32_t *)H);
- gcm_init_htab_avx(ctx->gcm_Htable, H);
+#if CAN_USE_GCM_ASM >= 2
+ if (ctx->impl == GCM_IMPL_AVX2) {
+ gcm_init_vpclmulqdq_avx2((uint128_t *)ctx->gcm_Htable, H);
+ } else
+#endif /* if CAN_USE_GCM_ASM >= 2 */
+ {
+ gcm_init_htab_avx(ctx->gcm_Htable, H);
+ }
if (iv_len == 12) {
memcpy(cb, iv, 12);
diff --git a/sys/contrib/openzfs/module/icp/algs/modes/modes.c b/sys/contrib/openzfs/module/icp/algs/modes/modes.c
index 343591cd9691..ef3c1806e4b6 100644
--- a/sys/contrib/openzfs/module/icp/algs/modes/modes.c
+++ b/sys/contrib/openzfs/module/icp/algs/modes/modes.c
@@ -171,7 +171,7 @@ gcm_clear_ctx(gcm_ctx_t *ctx)
explicit_memset(ctx->gcm_remainder, 0, sizeof (ctx->gcm_remainder));
explicit_memset(ctx->gcm_H, 0, sizeof (ctx->gcm_H));
#if defined(CAN_USE_GCM_ASM)
- if (ctx->gcm_use_avx == B_TRUE) {
+ if (ctx->impl != GCM_IMPL_GENERIC) {
ASSERT3P(ctx->gcm_Htable, !=, NULL);
explicit_memset(ctx->gcm_Htable, 0, ctx->gcm_htab_len);
kmem_free(ctx->gcm_Htable, ctx->gcm_htab_len);
diff --git a/sys/contrib/openzfs/module/icp/asm-x86_64/modes/THIRDPARTYLICENSE.boringssl b/sys/contrib/openzfs/module/icp/asm-x86_64/modes/THIRDPARTYLICENSE.boringssl
new file mode 100644
index 000000000000..04c03a37e0cb
--- /dev/null
+++ b/sys/contrib/openzfs/module/icp/asm-x86_64/modes/THIRDPARTYLICENSE.boringssl
@@ -0,0 +1,253 @@
+BoringSSL is a fork of OpenSSL. As such, large parts of it fall under OpenSSL
+licensing. Files that are completely new have a Google copyright and an ISC
+license. This license is reproduced at the bottom of this file.
+
+Contributors to BoringSSL are required to follow the CLA rules for Chromium:
+https://cla.developers.google.com/clas
+
+Files in third_party/ have their own licenses, as described therein. The MIT
+license, for third_party/fiat, which, unlike other third_party directories, is
+compiled into non-test libraries, is included below.
+
+The OpenSSL toolkit stays under a dual license, i.e. both the conditions of the
+OpenSSL License and the original SSLeay license apply to the toolkit. See below
+for the actual license texts. Actually both licenses are BSD-style Open Source
+licenses. In case of any license issues related to OpenSSL please contact
+openssl-core@openssl.org.
+
+The following are Google-internal bug numbers where explicit permission from
+some authors is recorded for use of their work. (This is purely for our own
+record keeping.)
+ 27287199
+ 27287880
+ 27287883
+ 263291445
+
+
+ OpenSSL License
+ ---------------
+
+/* ====================================================================
+ * Copyright (c) 1998-2011 The OpenSSL Project. All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in
+ * the documentation and/or other materials provided with the
+ * distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ * software must display the following acknowledgment:
+ * "This product includes software developed by the OpenSSL Project
+ * for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ * endorse or promote products derived from this software without
+ * prior written permission. For written permission, please contact
+ * openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ * nor may "OpenSSL" appear in their names without prior written
+ * permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ * acknowledgment:
+ * "This product includes software developed by the OpenSSL Project
+ * for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com). This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+ Original SSLeay License
+ -----------------------
+
+/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
+ * All rights reserved.
+ *
+ * This package is an SSL implementation written
+ * by Eric Young (eay@cryptsoft.com).
+ * The implementation was written so as to conform with Netscapes SSL.
+ *
+ * This library is free for commercial and non-commercial use as long as
+ * the following conditions are aheared to. The following conditions
+ * apply to all code found in this distribution, be it the RC4, RSA,
+ * lhash, DES, etc., code; not just the SSL code. The SSL documentation
+ * included with this distribution is covered by the same copyright terms
+ * except that the holder is Tim Hudson (tjh@cryptsoft.com).
+ *
+ * Copyright remains Eric Young's, and as such any Copyright notices in
+ * the code are not to be removed.
+ * If this package is used in a product, Eric Young should be given attribution
+ * as the author of the parts of the library used.
+ * This can be in the form of a textual message at program startup or
+ * in documentation (online or textual) provided with the package.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ * must display the following acknowledgement:
+ * "This product includes cryptographic software written by
+ * Eric Young (eay@cryptsoft.com)"
+ * The word 'cryptographic' can be left out if the rouines from the library
+ * being used are not cryptographic related :-).
+ * 4. If you include any Windows specific code (or a derivative thereof) from
+ * the apps directory (application code) you must include an acknowledgement:
+ * "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * The licence and distribution terms for any publically available version or
+ * derivative of this code cannot be changed. i.e. this code cannot simply be
+ * copied and put under another distribution licence
+ * [including the GNU Public Licence.]
+ */
+
+
+ISC license used for completely new code in BoringSSL:
+
+/* Copyright 2015 The BoringSSL Authors
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY
+ * SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION
+ * OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN
+ * CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */
+
+
+The code in third_party/fiat carries the MIT license:
+
+Copyright (c) 2015-2016 the fiat-crypto authors (see
+https://github.com/mit-plv/fiat-crypto/blob/master/AUTHORS).
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+
+Licenses for support code
+-------------------------
+
+Parts of the TLS test suite are under the Go license. This code is not included
+in BoringSSL (i.e. libcrypto and libssl) when compiled, however, so
+distributing code linked against BoringSSL does not trigger this license:
+
+Copyright (c) 2009 The Go Authors. All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+ * Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+ * Redistributions in binary form must reproduce the above
+copyright notice, this list of conditions and the following disclaimer
+in the documentation and/or other materials provided with the
+distribution.
+ * Neither the name of Google Inc. nor the names of its
+contributors may be used to endorse or promote products derived from
+this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+
+BoringSSL uses the Chromium test infrastructure to run a continuous build,
+trybots etc. The scripts which manage this, and the script for generating build
+metadata, are under the Chromium license. Distributing code linked against
+BoringSSL does not trigger this license.
+
+Copyright 2015 The Chromium Authors. All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+ * Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+ * Redistributions in binary form must reproduce the above
+copyright notice, this list of conditions and the following disclaimer
+in the documentation and/or other materials provided with the
+distribution.
+ * Neither the name of Google Inc. nor the names of its
+contributors may be used to endorse or promote products derived from
+this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/sys/contrib/openzfs/module/icp/asm-x86_64/modes/THIRDPARTYLICENSE.boringssl.descrip b/sys/contrib/openzfs/module/icp/asm-x86_64/modes/THIRDPARTYLICENSE.boringssl.descrip
new file mode 100644
index 000000000000..f63a67a4d2ae
--- /dev/null
+++ b/sys/contrib/openzfs/module/icp/asm-x86_64/modes/THIRDPARTYLICENSE.boringssl.descrip
@@ -0,0 +1 @@
+PORTIONS OF AES GCM and GHASH FUNCTIONALITY
diff --git a/sys/contrib/openzfs/module/icp/asm-x86_64/modes/aesni-gcm-avx2-vaes.S b/sys/contrib/openzfs/module/icp/asm-x86_64/modes/aesni-gcm-avx2-vaes.S
new file mode 100644
index 000000000000..3d1b045127e2
--- /dev/null
+++ b/sys/contrib/openzfs/module/icp/asm-x86_64/modes/aesni-gcm-avx2-vaes.S
@@ -0,0 +1,1323 @@
+// SPDX-License-Identifier: Apache-2.0
+// This file is generated from a similarly-named Perl script in the BoringSSL
+// source tree. Do not edit by hand.
+
+#if defined(__x86_64__) && defined(HAVE_AVX) && \
+ defined(HAVE_VAES) && defined(HAVE_VPCLMULQDQ)
+
+#define _ASM
+#include <sys/asm_linkage.h>
+
+/* Windows userland links with OpenSSL */
+#if !defined (_WIN32) || defined (_KERNEL)
+
+.section .rodata
+.balign 16
+
+
+.Lbswap_mask:
+.quad 0x08090a0b0c0d0e0f, 0x0001020304050607
+
+
+
+
+
+
+
+
+.Lgfpoly:
+.quad 1, 0xc200000000000000
+
+
+.Lgfpoly_and_internal_carrybit:
+.quad 1, 0xc200000000000001
+
+.balign 32
+
+.Lctr_pattern:
+.quad 0, 0
+.quad 1, 0
+.Linc_2blocks:
+.quad 2, 0
+.quad 2, 0
+
+ENTRY_ALIGN(gcm_init_vpclmulqdq_avx2, 32)
+.cfi_startproc
+
+ENDBR
+
+
+
+
+
+ vmovdqu (%rsi),%xmm3
+ // KCF/ICP stores H in network byte order with the hi qword first
+ // so we need to swap all bytes, not the 2 qwords.
+ vmovdqu .Lbswap_mask(%rip),%xmm4
+ vpshufb %xmm4,%xmm3,%xmm3
+
+
+
+
+
+ vpshufd $0xd3,%xmm3,%xmm0
+ vpsrad $31,%xmm0,%xmm0
+ vpaddq %xmm3,%xmm3,%xmm3
+ vpand .Lgfpoly_and_internal_carrybit(%rip),%xmm0,%xmm0
+ vpxor %xmm0,%xmm3,%xmm3
+
+ vbroadcasti128 .Lgfpoly(%rip),%ymm6
+
+
+ vpclmulqdq $0x00,%xmm3,%xmm3,%xmm0
+ vpclmulqdq $0x11,%xmm3,%xmm3,%xmm5
+ vpclmulqdq $0x01,%xmm0,%xmm6,%xmm1
+ vpshufd $0x4e,%xmm0,%xmm0
+ vpxor %xmm0,%xmm1,%xmm1
+ vpclmulqdq $0x01,%xmm1,%xmm6,%xmm0
+ vpshufd $0x4e,%xmm1,%xmm1
+ vpxor %xmm1,%xmm5,%xmm5
+ vpxor %xmm0,%xmm5,%xmm5
+
+
+
+ vinserti128 $1,%xmm3,%ymm5,%ymm3
+ vinserti128 $1,%xmm5,%ymm5,%ymm5
+
+
+ vpclmulqdq $0x00,%ymm5,%ymm3,%ymm0
+ vpclmulqdq $0x01,%ymm5,%ymm3,%ymm1
+ vpclmulqdq $0x10,%ymm5,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpclmulqdq $0x01,%ymm0,%ymm6,%ymm2
+ vpshufd $0x4e,%ymm0,%ymm0
+ vpxor %ymm0,%ymm1,%ymm1
+ vpxor %ymm2,%ymm1,%ymm1
+ vpclmulqdq $0x11,%ymm5,%ymm3,%ymm4
+ vpclmulqdq $0x01,%ymm1,%ymm6,%ymm0
+ vpshufd $0x4e,%ymm1,%ymm1
+ vpxor %ymm1,%ymm4,%ymm4
+ vpxor %ymm0,%ymm4,%ymm4
+
+
+
+ vmovdqu %ymm3,96(%rdi)
+ vmovdqu %ymm4,64(%rdi)
+
+
+
+ vpunpcklqdq %ymm3,%ymm4,%ymm0
+ vpunpckhqdq %ymm3,%ymm4,%ymm1
+ vpxor %ymm1,%ymm0,%ymm0
+ vmovdqu %ymm0,128+32(%rdi)
+
+
+ vpclmulqdq $0x00,%ymm5,%ymm4,%ymm0
+ vpclmulqdq $0x01,%ymm5,%ymm4,%ymm1
+ vpclmulqdq $0x10,%ymm5,%ymm4,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpclmulqdq $0x01,%ymm0,%ymm6,%ymm2
+ vpshufd $0x4e,%ymm0,%ymm0
+ vpxor %ymm0,%ymm1,%ymm1
+ vpxor %ymm2,%ymm1,%ymm1
+ vpclmulqdq $0x11,%ymm5,%ymm4,%ymm3
+ vpclmulqdq $0x01,%ymm1,%ymm6,%ymm0
+ vpshufd $0x4e,%ymm1,%ymm1
+ vpxor %ymm1,%ymm3,%ymm3
+ vpxor %ymm0,%ymm3,%ymm3
+
+ vpclmulqdq $0x00,%ymm5,%ymm3,%ymm0
+ vpclmulqdq $0x01,%ymm5,%ymm3,%ymm1
+ vpclmulqdq $0x10,%ymm5,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpclmulqdq $0x01,%ymm0,%ymm6,%ymm2
+ vpshufd $0x4e,%ymm0,%ymm0
+ vpxor %ymm0,%ymm1,%ymm1
+ vpxor %ymm2,%ymm1,%ymm1
+ vpclmulqdq $0x11,%ymm5,%ymm3,%ymm4
+ vpclmulqdq $0x01,%ymm1,%ymm6,%ymm0
+ vpshufd $0x4e,%ymm1,%ymm1
+ vpxor %ymm1,%ymm4,%ymm4
+ vpxor %ymm0,%ymm4,%ymm4
+
+ vmovdqu %ymm3,32(%rdi)
+ vmovdqu %ymm4,0(%rdi)
+
+
+
+ vpunpcklqdq %ymm3,%ymm4,%ymm0
+ vpunpckhqdq %ymm3,%ymm4,%ymm1
+ vpxor %ymm1,%ymm0,%ymm0
+ vmovdqu %ymm0,128(%rdi)
+
+ vzeroupper
+ RET
+
+.cfi_endproc
+SET_SIZE(gcm_init_vpclmulqdq_avx2)
+ENTRY_ALIGN(gcm_gmult_vpclmulqdq_avx2, 32)
+.cfi_startproc
+
+ENDBR
+
+
+
+ vmovdqu (%rdi),%xmm0
+ vmovdqu .Lbswap_mask(%rip),%xmm1
+ vmovdqu 128-16(%rsi),%xmm2
+ vmovdqu .Lgfpoly(%rip),%xmm3
+ vpshufb %xmm1,%xmm0,%xmm0
+
+ vpclmulqdq $0x00,%xmm2,%xmm0,%xmm4
+ vpclmulqdq $0x01,%xmm2,%xmm0,%xmm5
+ vpclmulqdq $0x10,%xmm2,%xmm0,%xmm6
+ vpxor %xmm6,%xmm5,%xmm5
+ vpclmulqdq $0x01,%xmm4,%xmm3,%xmm6
+ vpshufd $0x4e,%xmm4,%xmm4
+ vpxor %xmm4,%xmm5,%xmm5
+ vpxor %xmm6,%xmm5,%xmm5
+ vpclmulqdq $0x11,%xmm2,%xmm0,%xmm0
+ vpclmulqdq $0x01,%xmm5,%xmm3,%xmm4
+ vpshufd $0x4e,%xmm5,%xmm5
+ vpxor %xmm5,%xmm0,%xmm0
+ vpxor %xmm4,%xmm0,%xmm0
+
+
+ vpshufb %xmm1,%xmm0,%xmm0
+ vmovdqu %xmm0,(%rdi)
+
+
+ RET
+
+.cfi_endproc
+SET_SIZE(gcm_gmult_vpclmulqdq_avx2)
+ENTRY_ALIGN(gcm_ghash_vpclmulqdq_avx2, 32)
+.cfi_startproc
+
+ENDBR
+
+
+
+
+
+
+ vmovdqu .Lbswap_mask(%rip),%xmm6
+ vmovdqu .Lgfpoly(%rip),%xmm7
+
+
+ vmovdqu (%rdi),%xmm5
+ vpshufb %xmm6,%xmm5,%xmm5
+
+
+ cmpq $32,%rcx
+ jb .Lghash_lastblock
+
+
+
+ vinserti128 $1,%xmm6,%ymm6,%ymm6
+ vinserti128 $1,%xmm7,%ymm7,%ymm7
+
+ cmpq $127,%rcx
+ jbe .Lghash_loop_1x
+
+
+ vmovdqu 128(%rsi),%ymm8
+ vmovdqu 128+32(%rsi),%ymm9
+.Lghash_loop_4x:
+
+ vmovdqu 0(%rdx),%ymm1
+ vpshufb %ymm6,%ymm1,%ymm1
+ vmovdqu 0(%rsi),%ymm2
+ vpxor %ymm5,%ymm1,%ymm1
+ vpclmulqdq $0x00,%ymm2,%ymm1,%ymm3
+ vpclmulqdq $0x11,%ymm2,%ymm1,%ymm5
+ vpunpckhqdq %ymm1,%ymm1,%ymm0
+ vpxor %ymm1,%ymm0,%ymm0
+ vpclmulqdq $0x00,%ymm8,%ymm0,%ymm4
+
+ vmovdqu 32(%rdx),%ymm1
+ vpshufb %ymm6,%ymm1,%ymm1
+ vmovdqu 32(%rsi),%ymm2
+ vpclmulqdq $0x00,%ymm2,%ymm1,%ymm0
+ vpxor %ymm0,%ymm3,%ymm3
+ vpclmulqdq $0x11,%ymm2,%ymm1,%ymm0
+ vpxor %ymm0,%ymm5,%ymm5
+ vpunpckhqdq %ymm1,%ymm1,%ymm0
+ vpxor %ymm1,%ymm0,%ymm0
+ vpclmulqdq $0x10,%ymm8,%ymm0,%ymm0
+ vpxor %ymm0,%ymm4,%ymm4
+
+ vmovdqu 64(%rdx),%ymm1
+ vpshufb %ymm6,%ymm1,%ymm1
+ vmovdqu 64(%rsi),%ymm2
+ vpclmulqdq $0x00,%ymm2,%ymm1,%ymm0
+ vpxor %ymm0,%ymm3,%ymm3
+ vpclmulqdq $0x11,%ymm2,%ymm1,%ymm0
+ vpxor %ymm0,%ymm5,%ymm5
+ vpunpckhqdq %ymm1,%ymm1,%ymm0
+ vpxor %ymm1,%ymm0,%ymm0
+ vpclmulqdq $0x00,%ymm9,%ymm0,%ymm0
+ vpxor %ymm0,%ymm4,%ymm4
+
+
+ vmovdqu 96(%rdx),%ymm1
+ vpshufb %ymm6,%ymm1,%ymm1
+ vmovdqu 96(%rsi),%ymm2
+ vpclmulqdq $0x00,%ymm2,%ymm1,%ymm0
+ vpxor %ymm0,%ymm3,%ymm3
+ vpclmulqdq $0x11,%ymm2,%ymm1,%ymm0
+ vpxor %ymm0,%ymm5,%ymm5
+ vpunpckhqdq %ymm1,%ymm1,%ymm0
+ vpxor %ymm1,%ymm0,%ymm0
+ vpclmulqdq $0x10,%ymm9,%ymm0,%ymm0
+ vpxor %ymm0,%ymm4,%ymm4
+
+ vpxor %ymm3,%ymm4,%ymm4
+ vpxor %ymm5,%ymm4,%ymm4
+
+
+ vbroadcasti128 .Lgfpoly(%rip),%ymm2
+ vpclmulqdq $0x01,%ymm3,%ymm2,%ymm0
+ vpshufd $0x4e,%ymm3,%ymm3
+ vpxor %ymm3,%ymm4,%ymm4
+ vpxor %ymm0,%ymm4,%ymm4
+
+ vpclmulqdq $0x01,%ymm4,%ymm2,%ymm0
+ vpshufd $0x4e,%ymm4,%ymm4
+ vpxor %ymm4,%ymm5,%ymm5
+ vpxor %ymm0,%ymm5,%ymm5
+ vextracti128 $1,%ymm5,%xmm0
+ vpxor %xmm0,%xmm5,%xmm5
+
+ subq $-128,%rdx
+ addq $-128,%rcx
+ cmpq $127,%rcx
+ ja .Lghash_loop_4x
+
+
+ cmpq $32,%rcx
+ jb .Lghash_loop_1x_done
+.Lghash_loop_1x:
+ vmovdqu (%rdx),%ymm0
+ vpshufb %ymm6,%ymm0,%ymm0
+ vpxor %ymm0,%ymm5,%ymm5
+ vmovdqu 128-32(%rsi),%ymm0
+ vpclmulqdq $0x00,%ymm0,%ymm5,%ymm1
+ vpclmulqdq $0x01,%ymm0,%ymm5,%ymm2
+ vpclmulqdq $0x10,%ymm0,%ymm5,%ymm3
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x01,%ymm1,%ymm7,%ymm3
+ vpshufd $0x4e,%ymm1,%ymm1
+ vpxor %ymm1,%ymm2,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x11,%ymm0,%ymm5,%ymm5
+ vpclmulqdq $0x01,%ymm2,%ymm7,%ymm1
+ vpshufd $0x4e,%ymm2,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpxor %ymm1,%ymm5,%ymm5
+
+ vextracti128 $1,%ymm5,%xmm0
+ vpxor %xmm0,%xmm5,%xmm5
+ addq $32,%rdx
+ subq $32,%rcx
+ cmpq $32,%rcx
+ jae .Lghash_loop_1x
+.Lghash_loop_1x_done:
+
+
+.Lghash_lastblock:
+ testq %rcx,%rcx
+ jz .Lghash_done
+ vmovdqu (%rdx),%xmm0
+ vpshufb %xmm6,%xmm0,%xmm0
+ vpxor %xmm0,%xmm5,%xmm5
+ vmovdqu 128-16(%rsi),%xmm0
+ vpclmulqdq $0x00,%xmm0,%xmm5,%xmm1
+ vpclmulqdq $0x01,%xmm0,%xmm5,%xmm2
+ vpclmulqdq $0x10,%xmm0,%xmm5,%xmm3
+ vpxor %xmm3,%xmm2,%xmm2
+ vpclmulqdq $0x01,%xmm1,%xmm7,%xmm3
+ vpshufd $0x4e,%xmm1,%xmm1
+ vpxor %xmm1,%xmm2,%xmm2
+ vpxor %xmm3,%xmm2,%xmm2
+ vpclmulqdq $0x11,%xmm0,%xmm5,%xmm5
+ vpclmulqdq $0x01,%xmm2,%xmm7,%xmm1
+ vpshufd $0x4e,%xmm2,%xmm2
+ vpxor %xmm2,%xmm5,%xmm5
+ vpxor %xmm1,%xmm5,%xmm5
+
+
+.Lghash_done:
+
+ vpshufb %xmm6,%xmm5,%xmm5
+ vmovdqu %xmm5,(%rdi)
+
+ vzeroupper
+ RET
+
+.cfi_endproc
+SET_SIZE(gcm_ghash_vpclmulqdq_avx2)
+ENTRY_ALIGN(aes_gcm_enc_update_vaes_avx2, 32)
+.cfi_startproc
+
+ENDBR
+ pushq %r12
+.cfi_adjust_cfa_offset 8
+.cfi_offset %r12,-16
+
+ movq 16(%rsp),%r12
+#ifdef BORINGSSL_DISPATCH_TEST
+.extern BORINGSSL_function_hit
+.hidden BORINGSSL_function_hit
+ movb $1,BORINGSSL_function_hit+6(%rip)
+#endif
+ vbroadcasti128 .Lbswap_mask(%rip),%ymm0
+
+
+
+ vmovdqu (%r12),%xmm1
+ vpshufb %xmm0,%xmm1,%xmm1
+ vbroadcasti128 (%r8),%ymm11
+ vpshufb %ymm0,%ymm11,%ymm11
+
+
+
+ movl 504(%rcx),%r10d // ICP has a larger offset for rounds.
+ leal -24(,%r10,4),%r10d // ICP uses 10,12,14 not 9,11,13 for rounds.
+
+
+
+
+ leaq 96(%rcx,%r10,4),%r11
+ vbroadcasti128 (%rcx),%ymm9
+ vbroadcasti128 (%r11),%ymm10
+
+
+ vpaddd .Lctr_pattern(%rip),%ymm11,%ymm11
+
+
+
+ cmpq $127,%rdx
+ jbe .Lcrypt_loop_4x_done__func1
+
+ vmovdqu 128(%r9),%ymm7
+ vmovdqu 128+32(%r9),%ymm8
+
+
+
+ vmovdqu .Linc_2blocks(%rip),%ymm2
+ vpshufb %ymm0,%ymm11,%ymm12
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm13
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm14
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm15
+ vpaddd %ymm2,%ymm11,%ymm11
+
+
+ vpxor %ymm9,%ymm12,%ymm12
+ vpxor %ymm9,%ymm13,%ymm13
+ vpxor %ymm9,%ymm14,%ymm14
+ vpxor %ymm9,%ymm15,%ymm15
+
+ leaq 16(%rcx),%rax
+.Lvaesenc_loop_first_4_vecs__func1:
+ vbroadcasti128 (%rax),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ addq $16,%rax
+ cmpq %rax,%r11
+ jne .Lvaesenc_loop_first_4_vecs__func1
+ vpxor 0(%rdi),%ymm10,%ymm2
+ vpxor 32(%rdi),%ymm10,%ymm3
+ vpxor 64(%rdi),%ymm10,%ymm5
+ vpxor 96(%rdi),%ymm10,%ymm6
+ vaesenclast %ymm2,%ymm12,%ymm12
+ vaesenclast %ymm3,%ymm13,%ymm13
+ vaesenclast %ymm5,%ymm14,%ymm14
+ vaesenclast %ymm6,%ymm15,%ymm15
+ vmovdqu %ymm12,0(%rsi)
+ vmovdqu %ymm13,32(%rsi)
+ vmovdqu %ymm14,64(%rsi)
+ vmovdqu %ymm15,96(%rsi)
+
+ subq $-128,%rdi
+ addq $-128,%rdx
+ cmpq $127,%rdx
+ jbe .Lghash_last_ciphertext_4x__func1
+.balign 16
+.Lcrypt_loop_4x__func1:
+
+
+
+
+ vmovdqu .Linc_2blocks(%rip),%ymm2
+ vpshufb %ymm0,%ymm11,%ymm12
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm13
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm14
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm15
+ vpaddd %ymm2,%ymm11,%ymm11
+
+
+ vpxor %ymm9,%ymm12,%ymm12
+ vpxor %ymm9,%ymm13,%ymm13
+ vpxor %ymm9,%ymm14,%ymm14
+ vpxor %ymm9,%ymm15,%ymm15
+
+ cmpl $24,%r10d
+ jl .Laes128__func1
+ je .Laes192__func1
+
+ vbroadcasti128 -208(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vbroadcasti128 -192(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+.Laes192__func1:
+ vbroadcasti128 -176(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vbroadcasti128 -160(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+.Laes128__func1:
+ prefetcht0 512(%rdi)
+ prefetcht0 512+64(%rdi)
+
+ vmovdqu 0(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 0(%r9),%ymm4
+ vpxor %ymm1,%ymm3,%ymm3
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x00,%ymm7,%ymm2,%ymm6
+
+ vbroadcasti128 -144(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vbroadcasti128 -128(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vmovdqu 32(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 32(%r9),%ymm4
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x10,%ymm7,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vbroadcasti128 -112(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vmovdqu 64(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 64(%r9),%ymm4
+
+ vbroadcasti128 -96(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+
+ vbroadcasti128 -80(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x00,%ymm8,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+
+ vmovdqu 96(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+
+ vbroadcasti128 -64(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vmovdqu 96(%r9),%ymm4
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x10,%ymm8,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vbroadcasti128 -48(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm1,%ymm6,%ymm6
+
+
+ vbroadcasti128 .Lgfpoly(%rip),%ymm4
+ vpclmulqdq $0x01,%ymm5,%ymm4,%ymm2
+ vpshufd $0x4e,%ymm5,%ymm5
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vbroadcasti128 -32(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vpclmulqdq $0x01,%ymm6,%ymm4,%ymm2
+ vpshufd $0x4e,%ymm6,%ymm6
+ vpxor %ymm6,%ymm1,%ymm1
+ vpxor %ymm2,%ymm1,%ymm1
+
+ vbroadcasti128 -16(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vextracti128 $1,%ymm1,%xmm2
+ vpxor %xmm2,%xmm1,%xmm1
+
+
+ subq $-128,%rsi
+ vpxor 0(%rdi),%ymm10,%ymm2
+ vpxor 32(%rdi),%ymm10,%ymm3
+ vpxor 64(%rdi),%ymm10,%ymm5
+ vpxor 96(%rdi),%ymm10,%ymm6
+ vaesenclast %ymm2,%ymm12,%ymm12
+ vaesenclast %ymm3,%ymm13,%ymm13
+ vaesenclast %ymm5,%ymm14,%ymm14
+ vaesenclast %ymm6,%ymm15,%ymm15
+ vmovdqu %ymm12,0(%rsi)
+ vmovdqu %ymm13,32(%rsi)
+ vmovdqu %ymm14,64(%rsi)
+ vmovdqu %ymm15,96(%rsi)
+
+ subq $-128,%rdi
+
+ addq $-128,%rdx
+ cmpq $127,%rdx
+ ja .Lcrypt_loop_4x__func1
+.Lghash_last_ciphertext_4x__func1:
+
+ vmovdqu 0(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 0(%r9),%ymm4
+ vpxor %ymm1,%ymm3,%ymm3
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x00,%ymm7,%ymm2,%ymm6
+
+ vmovdqu 32(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 32(%r9),%ymm4
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x10,%ymm7,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vmovdqu 64(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 64(%r9),%ymm4
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x00,%ymm8,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+
+ vmovdqu 96(%rsi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 96(%r9),%ymm4
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x10,%ymm8,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm1,%ymm6,%ymm6
+
+
+ vbroadcasti128 .Lgfpoly(%rip),%ymm4
+ vpclmulqdq $0x01,%ymm5,%ymm4,%ymm2
+ vpshufd $0x4e,%ymm5,%ymm5
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vpclmulqdq $0x01,%ymm6,%ymm4,%ymm2
+ vpshufd $0x4e,%ymm6,%ymm6
+ vpxor %ymm6,%ymm1,%ymm1
+ vpxor %ymm2,%ymm1,%ymm1
+ vextracti128 $1,%ymm1,%xmm2
+ vpxor %xmm2,%xmm1,%xmm1
+
+ subq $-128,%rsi
+.Lcrypt_loop_4x_done__func1:
+
+ testq %rdx,%rdx
+ jz .Ldone__func1
+
+
+
+
+
+ leaq 128(%r9),%r8
+ subq %rdx,%r8
+
+
+ vpxor %xmm5,%xmm5,%xmm5
+ vpxor %xmm6,%xmm6,%xmm6
+ vpxor %xmm7,%xmm7,%xmm7
+
+ cmpq $64,%rdx
+ jb .Llessthan64bytes__func1
+
+
+ vpshufb %ymm0,%ymm11,%ymm12
+ vpaddd .Linc_2blocks(%rip),%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm13
+ vpaddd .Linc_2blocks(%rip),%ymm11,%ymm11
+ vpxor %ymm9,%ymm12,%ymm12
+ vpxor %ymm9,%ymm13,%ymm13
+ leaq 16(%rcx),%rax
+.Lvaesenc_loop_tail_1__func1:
+ vbroadcasti128 (%rax),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ addq $16,%rax
+ cmpq %rax,%r11
+ jne .Lvaesenc_loop_tail_1__func1
+ vaesenclast %ymm10,%ymm12,%ymm12
+ vaesenclast %ymm10,%ymm13,%ymm13
+
+
+ vmovdqu 0(%rdi),%ymm2
+ vmovdqu 32(%rdi),%ymm3
+ vpxor %ymm2,%ymm12,%ymm12
+ vpxor %ymm3,%ymm13,%ymm13
+ vmovdqu %ymm12,0(%rsi)
+ vmovdqu %ymm13,32(%rsi)
+
+
+ vpshufb %ymm0,%ymm12,%ymm12
+ vpshufb %ymm0,%ymm13,%ymm13
+ vpxor %ymm1,%ymm12,%ymm12
+ vmovdqu (%r8),%ymm2
+ vmovdqu 32(%r8),%ymm3
+ vpclmulqdq $0x00,%ymm2,%ymm12,%ymm5
+ vpclmulqdq $0x01,%ymm2,%ymm12,%ymm6
+ vpclmulqdq $0x10,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%ymm2,%ymm12,%ymm7
+ vpclmulqdq $0x00,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm5,%ymm5
+ vpclmulqdq $0x01,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x10,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm7,%ymm7
+
+ addq $64,%r8
+ addq $64,%rdi
+ addq $64,%rsi
+ subq $64,%rdx
+ jz .Lreduce__func1
+
+ vpxor %xmm1,%xmm1,%xmm1
+
+
+.Llessthan64bytes__func1:
+ vpshufb %ymm0,%ymm11,%ymm12
+ vpaddd .Linc_2blocks(%rip),%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm13
+ vpxor %ymm9,%ymm12,%ymm12
+ vpxor %ymm9,%ymm13,%ymm13
+ leaq 16(%rcx),%rax
+.Lvaesenc_loop_tail_2__func1:
+ vbroadcasti128 (%rax),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ addq $16,%rax
+ cmpq %rax,%r11
+ jne .Lvaesenc_loop_tail_2__func1
+ vaesenclast %ymm10,%ymm12,%ymm12
+ vaesenclast %ymm10,%ymm13,%ymm13
+
+
+
+
+ cmpq $32,%rdx
+ jb .Lxor_one_block__func1
+ je .Lxor_two_blocks__func1
+
+.Lxor_three_blocks__func1:
+ vmovdqu 0(%rdi),%ymm2
+ vmovdqu 32(%rdi),%xmm3
+ vpxor %ymm2,%ymm12,%ymm12
+ vpxor %xmm3,%xmm13,%xmm13
+ vmovdqu %ymm12,0(%rsi)
+ vmovdqu %xmm13,32(%rsi)
+
+ vpshufb %ymm0,%ymm12,%ymm12
+ vpshufb %xmm0,%xmm13,%xmm13
+ vpxor %ymm1,%ymm12,%ymm12
+ vmovdqu (%r8),%ymm2
+ vmovdqu 32(%r8),%xmm3
+ vpclmulqdq $0x00,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm5,%ymm5
+ vpclmulqdq $0x01,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x10,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm7,%ymm7
+ jmp .Lghash_mul_one_vec_unreduced__func1
+
+.Lxor_two_blocks__func1:
+ vmovdqu (%rdi),%ymm2
+ vpxor %ymm2,%ymm12,%ymm12
+ vmovdqu %ymm12,(%rsi)
+ vpshufb %ymm0,%ymm12,%ymm12
+ vpxor %ymm1,%ymm12,%ymm12
+ vmovdqu (%r8),%ymm2
+ jmp .Lghash_mul_one_vec_unreduced__func1
+
+.Lxor_one_block__func1:
+ vmovdqu (%rdi),%xmm2
+ vpxor %xmm2,%xmm12,%xmm12
+ vmovdqu %xmm12,(%rsi)
+ vpshufb %xmm0,%xmm12,%xmm12
+ vpxor %xmm1,%xmm12,%xmm12
+ vmovdqu (%r8),%xmm2
+
+.Lghash_mul_one_vec_unreduced__func1:
+ vpclmulqdq $0x00,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm5,%ymm5
+ vpclmulqdq $0x01,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x10,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm7,%ymm7
+
+.Lreduce__func1:
+
+ vbroadcasti128 .Lgfpoly(%rip),%ymm2
+ vpclmulqdq $0x01,%ymm5,%ymm2,%ymm3
+ vpshufd $0x4e,%ymm5,%ymm5
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm3,%ymm6,%ymm6
+ vpclmulqdq $0x01,%ymm6,%ymm2,%ymm3
+ vpshufd $0x4e,%ymm6,%ymm6
+ vpxor %ymm6,%ymm7,%ymm7
+ vpxor %ymm3,%ymm7,%ymm7
+ vextracti128 $1,%ymm7,%xmm1
+ vpxor %xmm7,%xmm1,%xmm1
+
+.Ldone__func1:
+
+ vpshufb %xmm0,%xmm1,%xmm1
+ vmovdqu %xmm1,(%r12)
+
+ vzeroupper
+ popq %r12
+.cfi_adjust_cfa_offset -8
+.cfi_restore %r12
+ RET
+
+.cfi_endproc
+SET_SIZE(aes_gcm_enc_update_vaes_avx2)
+ENTRY_ALIGN(aes_gcm_dec_update_vaes_avx2, 32)
+.cfi_startproc
+
+ENDBR
+ pushq %r12
+.cfi_adjust_cfa_offset 8
+.cfi_offset %r12,-16
+
+ movq 16(%rsp),%r12
+ vbroadcasti128 .Lbswap_mask(%rip),%ymm0
+
+
+
+ vmovdqu (%r12),%xmm1
+ vpshufb %xmm0,%xmm1,%xmm1
+ vbroadcasti128 (%r8),%ymm11
+ vpshufb %ymm0,%ymm11,%ymm11
+
+
+
+ movl 504(%rcx),%r10d // ICP has a larger offset for rounds.
+ leal -24(,%r10,4),%r10d // ICP uses 10,12,14 not 9,11,13 for rounds.
+
+
+
+
+ leaq 96(%rcx,%r10,4),%r11
+ vbroadcasti128 (%rcx),%ymm9
+ vbroadcasti128 (%r11),%ymm10
+
+
+ vpaddd .Lctr_pattern(%rip),%ymm11,%ymm11
+
+
+
+ cmpq $127,%rdx
+ jbe .Lcrypt_loop_4x_done__func2
+
+ vmovdqu 128(%r9),%ymm7
+ vmovdqu 128+32(%r9),%ymm8
+.balign 16
+.Lcrypt_loop_4x__func2:
+
+
+
+
+ vmovdqu .Linc_2blocks(%rip),%ymm2
+ vpshufb %ymm0,%ymm11,%ymm12
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm13
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm14
+ vpaddd %ymm2,%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm15
+ vpaddd %ymm2,%ymm11,%ymm11
+
+
+ vpxor %ymm9,%ymm12,%ymm12
+ vpxor %ymm9,%ymm13,%ymm13
+ vpxor %ymm9,%ymm14,%ymm14
+ vpxor %ymm9,%ymm15,%ymm15
+
+ cmpl $24,%r10d
+ jl .Laes128__func2
+ je .Laes192__func2
+
+ vbroadcasti128 -208(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vbroadcasti128 -192(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+.Laes192__func2:
+ vbroadcasti128 -176(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vbroadcasti128 -160(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+.Laes128__func2:
+ prefetcht0 512(%rdi)
+ prefetcht0 512+64(%rdi)
+
+ vmovdqu 0(%rdi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 0(%r9),%ymm4
+ vpxor %ymm1,%ymm3,%ymm3
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x00,%ymm7,%ymm2,%ymm6
+
+ vbroadcasti128 -144(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vbroadcasti128 -128(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vmovdqu 32(%rdi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 32(%r9),%ymm4
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x10,%ymm7,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vbroadcasti128 -112(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vmovdqu 64(%rdi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+ vmovdqu 64(%r9),%ymm4
+
+ vbroadcasti128 -96(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+
+ vbroadcasti128 -80(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x00,%ymm8,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+
+ vmovdqu 96(%rdi),%ymm3
+ vpshufb %ymm0,%ymm3,%ymm3
+
+ vbroadcasti128 -64(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vmovdqu 96(%r9),%ymm4
+ vpclmulqdq $0x00,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm5,%ymm5
+ vpclmulqdq $0x11,%ymm4,%ymm3,%ymm2
+ vpxor %ymm2,%ymm1,%ymm1
+ vpunpckhqdq %ymm3,%ymm3,%ymm2
+ vpxor %ymm3,%ymm2,%ymm2
+ vpclmulqdq $0x10,%ymm8,%ymm2,%ymm2
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vbroadcasti128 -48(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm1,%ymm6,%ymm6
+
+
+ vbroadcasti128 .Lgfpoly(%rip),%ymm4
+ vpclmulqdq $0x01,%ymm5,%ymm4,%ymm2
+ vpshufd $0x4e,%ymm5,%ymm5
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm2,%ymm6,%ymm6
+
+ vbroadcasti128 -32(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+
+ vpclmulqdq $0x01,%ymm6,%ymm4,%ymm2
+ vpshufd $0x4e,%ymm6,%ymm6
+ vpxor %ymm6,%ymm1,%ymm1
+ vpxor %ymm2,%ymm1,%ymm1
+
+ vbroadcasti128 -16(%r11),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ vaesenc %ymm2,%ymm14,%ymm14
+ vaesenc %ymm2,%ymm15,%ymm15
+
+ vextracti128 $1,%ymm1,%xmm2
+ vpxor %xmm2,%xmm1,%xmm1
+
+
+
+ vpxor 0(%rdi),%ymm10,%ymm2
+ vpxor 32(%rdi),%ymm10,%ymm3
+ vpxor 64(%rdi),%ymm10,%ymm5
+ vpxor 96(%rdi),%ymm10,%ymm6
+ vaesenclast %ymm2,%ymm12,%ymm12
+ vaesenclast %ymm3,%ymm13,%ymm13
+ vaesenclast %ymm5,%ymm14,%ymm14
+ vaesenclast %ymm6,%ymm15,%ymm15
+ vmovdqu %ymm12,0(%rsi)
+ vmovdqu %ymm13,32(%rsi)
+ vmovdqu %ymm14,64(%rsi)
+ vmovdqu %ymm15,96(%rsi)
+
+ subq $-128,%rdi
+ subq $-128,%rsi
+ addq $-128,%rdx
+ cmpq $127,%rdx
+ ja .Lcrypt_loop_4x__func2
+.Lcrypt_loop_4x_done__func2:
+
+ testq %rdx,%rdx
+ jz .Ldone__func2
+
+
+
+
+
+ leaq 128(%r9),%r8
+ subq %rdx,%r8
+
+
+ vpxor %xmm5,%xmm5,%xmm5
+ vpxor %xmm6,%xmm6,%xmm6
+ vpxor %xmm7,%xmm7,%xmm7
+
+ cmpq $64,%rdx
+ jb .Llessthan64bytes__func2
+
+
+ vpshufb %ymm0,%ymm11,%ymm12
+ vpaddd .Linc_2blocks(%rip),%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm13
+ vpaddd .Linc_2blocks(%rip),%ymm11,%ymm11
+ vpxor %ymm9,%ymm12,%ymm12
+ vpxor %ymm9,%ymm13,%ymm13
+ leaq 16(%rcx),%rax
+.Lvaesenc_loop_tail_1__func2:
+ vbroadcasti128 (%rax),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ addq $16,%rax
+ cmpq %rax,%r11
+ jne .Lvaesenc_loop_tail_1__func2
+ vaesenclast %ymm10,%ymm12,%ymm12
+ vaesenclast %ymm10,%ymm13,%ymm13
+
+
+ vmovdqu 0(%rdi),%ymm2
+ vmovdqu 32(%rdi),%ymm3
+ vpxor %ymm2,%ymm12,%ymm12
+ vpxor %ymm3,%ymm13,%ymm13
+ vmovdqu %ymm12,0(%rsi)
+ vmovdqu %ymm13,32(%rsi)
+
+
+ vpshufb %ymm0,%ymm2,%ymm12
+ vpshufb %ymm0,%ymm3,%ymm13
+ vpxor %ymm1,%ymm12,%ymm12
+ vmovdqu (%r8),%ymm2
+ vmovdqu 32(%r8),%ymm3
+ vpclmulqdq $0x00,%ymm2,%ymm12,%ymm5
+ vpclmulqdq $0x01,%ymm2,%ymm12,%ymm6
+ vpclmulqdq $0x10,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%ymm2,%ymm12,%ymm7
+ vpclmulqdq $0x00,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm5,%ymm5
+ vpclmulqdq $0x01,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x10,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%ymm3,%ymm13,%ymm4
+ vpxor %ymm4,%ymm7,%ymm7
+
+ addq $64,%r8
+ addq $64,%rdi
+ addq $64,%rsi
+ subq $64,%rdx
+ jz .Lreduce__func2
+
+ vpxor %xmm1,%xmm1,%xmm1
+
+
+.Llessthan64bytes__func2:
+ vpshufb %ymm0,%ymm11,%ymm12
+ vpaddd .Linc_2blocks(%rip),%ymm11,%ymm11
+ vpshufb %ymm0,%ymm11,%ymm13
+ vpxor %ymm9,%ymm12,%ymm12
+ vpxor %ymm9,%ymm13,%ymm13
+ leaq 16(%rcx),%rax
+.Lvaesenc_loop_tail_2__func2:
+ vbroadcasti128 (%rax),%ymm2
+ vaesenc %ymm2,%ymm12,%ymm12
+ vaesenc %ymm2,%ymm13,%ymm13
+ addq $16,%rax
+ cmpq %rax,%r11
+ jne .Lvaesenc_loop_tail_2__func2
+ vaesenclast %ymm10,%ymm12,%ymm12
+ vaesenclast %ymm10,%ymm13,%ymm13
+
+
+
+
+ cmpq $32,%rdx
+ jb .Lxor_one_block__func2
+ je .Lxor_two_blocks__func2
+
+.Lxor_three_blocks__func2:
+ vmovdqu 0(%rdi),%ymm2
+ vmovdqu 32(%rdi),%xmm3
+ vpxor %ymm2,%ymm12,%ymm12
+ vpxor %xmm3,%xmm13,%xmm13
+ vmovdqu %ymm12,0(%rsi)
+ vmovdqu %xmm13,32(%rsi)
+
+ vpshufb %ymm0,%ymm2,%ymm12
+ vpshufb %xmm0,%xmm3,%xmm13
+ vpxor %ymm1,%ymm12,%ymm12
+ vmovdqu (%r8),%ymm2
+ vmovdqu 32(%r8),%xmm3
+ vpclmulqdq $0x00,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm5,%ymm5
+ vpclmulqdq $0x01,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x10,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%xmm3,%xmm13,%xmm4
+ vpxor %ymm4,%ymm7,%ymm7
+ jmp .Lghash_mul_one_vec_unreduced__func2
+
+.Lxor_two_blocks__func2:
+ vmovdqu (%rdi),%ymm2
+ vpxor %ymm2,%ymm12,%ymm12
+ vmovdqu %ymm12,(%rsi)
+ vpshufb %ymm0,%ymm2,%ymm12
+ vpxor %ymm1,%ymm12,%ymm12
+ vmovdqu (%r8),%ymm2
+ jmp .Lghash_mul_one_vec_unreduced__func2
+
+.Lxor_one_block__func2:
+ vmovdqu (%rdi),%xmm2
+ vpxor %xmm2,%xmm12,%xmm12
+ vmovdqu %xmm12,(%rsi)
+ vpshufb %xmm0,%xmm2,%xmm12
+ vpxor %xmm1,%xmm12,%xmm12
+ vmovdqu (%r8),%xmm2
+
+.Lghash_mul_one_vec_unreduced__func2:
+ vpclmulqdq $0x00,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm5,%ymm5
+ vpclmulqdq $0x01,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x10,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm6,%ymm6
+ vpclmulqdq $0x11,%ymm2,%ymm12,%ymm4
+ vpxor %ymm4,%ymm7,%ymm7
+
+.Lreduce__func2:
+
+ vbroadcasti128 .Lgfpoly(%rip),%ymm2
+ vpclmulqdq $0x01,%ymm5,%ymm2,%ymm3
+ vpshufd $0x4e,%ymm5,%ymm5
+ vpxor %ymm5,%ymm6,%ymm6
+ vpxor %ymm3,%ymm6,%ymm6
+ vpclmulqdq $0x01,%ymm6,%ymm2,%ymm3
+ vpshufd $0x4e,%ymm6,%ymm6
+ vpxor %ymm6,%ymm7,%ymm7
+ vpxor %ymm3,%ymm7,%ymm7
+ vextracti128 $1,%ymm7,%xmm1
+ vpxor %xmm7,%xmm1,%xmm1
+
+.Ldone__func2:
+
+ vpshufb %xmm0,%xmm1,%xmm1
+ vmovdqu %xmm1,(%r12)
+
+ vzeroupper
+ popq %r12
+.cfi_adjust_cfa_offset -8
+.cfi_restore %r12
+ RET
+
+.cfi_endproc
+SET_SIZE(aes_gcm_dec_update_vaes_avx2)
+
+#endif /* !_WIN32 || _KERNEL */
+
+/* Mark the stack non-executable. */
+#if defined(__linux__) && defined(__ELF__)
+.section .note.GNU-stack,"",%progbits
+#endif
+
+#endif /* defined(__x86_64__) && defined(HAVE_AVX) && defined(HAVE_AES) ... */
diff --git a/sys/contrib/openzfs/module/icp/core/kcf_sched.c b/sys/contrib/openzfs/module/icp/core/kcf_sched.c
index 759f0d81d521..75e1052a4ed4 100644
--- a/sys/contrib/openzfs/module/icp/core/kcf_sched.c
+++ b/sys/contrib/openzfs/module/icp/core/kcf_sched.c
@@ -124,7 +124,7 @@ kcf_context_cache_destructor(void *buf, void *cdrarg)
(void) cdrarg;
kcf_context_t *kctx = (kcf_context_t *)buf;
- ASSERT(kctx->kc_refcnt == 0);
+ ASSERT0(kctx->kc_refcnt);
}
void
diff --git a/sys/contrib/openzfs/module/icp/include/modes/modes.h b/sys/contrib/openzfs/module/icp/include/modes/modes.h
index ca734cf4f045..de11d9eafafb 100644
--- a/sys/contrib/openzfs/module/icp/include/modes/modes.h
+++ b/sys/contrib/openzfs/module/icp/include/modes/modes.h
@@ -42,7 +42,7 @@ extern "C" {
*/
#if defined(__x86_64__) && defined(HAVE_AVX) && \
defined(HAVE_AES) && defined(HAVE_PCLMULQDQ)
-#define CAN_USE_GCM_ASM
+#define CAN_USE_GCM_ASM (HAVE_VAES && HAVE_VPCLMULQDQ ? 2 : 1)
extern boolean_t gcm_avx_can_use_movbe;
#endif
@@ -129,6 +129,15 @@ typedef struct ccm_ctx {
#define ccm_copy_to ccm_common.cc_copy_to
#define ccm_flags ccm_common.cc_flags
+#ifdef CAN_USE_GCM_ASM
+typedef enum gcm_impl {
+ GCM_IMPL_GENERIC = 0,
+ GCM_IMPL_AVX,
+ GCM_IMPL_AVX2,
+ GCM_IMPL_MAX,
+} gcm_impl;
+#endif
+
/*
* gcm_tag_len: Length of authentication tag.
*
@@ -174,7 +183,7 @@ typedef struct gcm_ctx {
uint64_t gcm_len_a_len_c[2];
uint8_t *gcm_pt_buf;
#ifdef CAN_USE_GCM_ASM
- boolean_t gcm_use_avx;
+ enum gcm_impl impl;
#endif
} gcm_ctx_t;
diff --git a/sys/contrib/openzfs/module/icp/io/aes.c b/sys/contrib/openzfs/module/icp/io/aes.c
index ba703efa71fc..ca586eaf97ef 100644
--- a/sys/contrib/openzfs/module/icp/io/aes.c
+++ b/sys/contrib/openzfs/module/icp/io/aes.c
@@ -236,16 +236,16 @@ aes_encrypt_atomic(crypto_mechanism_t *mechanism,
aes_xor_block);
if (ret != CRYPTO_SUCCESS)
goto out;
- ASSERT(aes_ctx.ac_remainder_len == 0);
+ ASSERT0(aes_ctx.ac_remainder_len);
} else if (mechanism->cm_type == AES_GCM_MECH_INFO_TYPE) {
ret = gcm_encrypt_final((gcm_ctx_t *)&aes_ctx,
ciphertext, AES_BLOCK_LEN, aes_encrypt_block,
aes_copy_block, aes_xor_block);
if (ret != CRYPTO_SUCCESS)
goto out;
- ASSERT(aes_ctx.ac_remainder_len == 0);
+ ASSERT0(aes_ctx.ac_remainder_len);
} else {
- ASSERT(aes_ctx.ac_remainder_len == 0);
+ ASSERT0(aes_ctx.ac_remainder_len);
}
if (plaintext != ciphertext) {
@@ -337,7 +337,7 @@ aes_decrypt_atomic(crypto_mechanism_t *mechanism,
ret = ccm_decrypt_final((ccm_ctx_t *)&aes_ctx,
plaintext, AES_BLOCK_LEN, aes_encrypt_block,
aes_copy_block, aes_xor_block);
- ASSERT(aes_ctx.ac_remainder_len == 0);
+ ASSERT0(aes_ctx.ac_remainder_len);
if ((ret == CRYPTO_SUCCESS) &&
(ciphertext != plaintext)) {
plaintext->cd_length =
@@ -349,7 +349,7 @@ aes_decrypt_atomic(crypto_mechanism_t *mechanism,
ret = gcm_decrypt_final((gcm_ctx_t *)&aes_ctx,
plaintext, AES_BLOCK_LEN, aes_encrypt_block,
aes_xor_block);
- ASSERT(aes_ctx.ac_remainder_len == 0);
+ ASSERT0(aes_ctx.ac_remainder_len);
if ((ret == CRYPTO_SUCCESS) &&
(ciphertext != plaintext)) {
plaintext->cd_length =
diff --git a/sys/contrib/openzfs/module/nvpair/nvpair.c b/sys/contrib/openzfs/module/nvpair/nvpair.c
index 811cfc87d7a4..eb8c14b4a783 100644
--- a/sys/contrib/openzfs/module/nvpair/nvpair.c
+++ b/sys/contrib/openzfs/module/nvpair/nvpair.c
@@ -265,7 +265,7 @@ nv_priv_alloc_embedded(nvpriv_t *priv)
static int
nvt_tab_alloc(nvpriv_t *priv, uint64_t buckets)
{
- ASSERT3P(priv->nvp_hashtable, ==, NULL);
+ ASSERT0P(priv->nvp_hashtable);
ASSERT0(priv->nvp_nbuckets);
ASSERT0(priv->nvp_nentries);
@@ -334,7 +334,7 @@ nvt_lookup_name_type(const nvlist_t *nvl, const char *name, data_type_t type)
i_nvp_t **tab = priv->nvp_hashtable;
if (tab == NULL) {
- ASSERT3P(priv->nvp_list, ==, NULL);
+ ASSERT0P(priv->nvp_list);
ASSERT0(priv->nvp_nbuckets);
ASSERT0(priv->nvp_nentries);
return (NULL);
@@ -540,7 +540,7 @@ nvt_add_nvpair(nvlist_t *nvl, nvpair_t *nvp)
/* insert link at the beginning of the bucket */
i_nvp_t *new_entry = NVPAIR2I_NVP(nvp);
- ASSERT3P(new_entry->nvi_hashtable_next, ==, NULL);
+ ASSERT0P(new_entry->nvi_hashtable_next);
new_entry->nvi_hashtable_next = bucket;
// cppcheck-suppress nullPointerRedundantCheck
tab[index] = new_entry;
diff --git a/sys/contrib/openzfs/module/os/freebsd/spl/spl_kmem.c b/sys/contrib/openzfs/module/os/freebsd/spl/spl_kmem.c
index 6d198fad5203..ae6e36d988c2 100644
--- a/sys/contrib/openzfs/module/os/freebsd/spl/spl_kmem.c
+++ b/sys/contrib/openzfs/module/os/freebsd/spl/spl_kmem.c
@@ -160,7 +160,7 @@ kmem_cache_create(const char *name, size_t bufsize, size_t align,
{
kmem_cache_t *cache;
- ASSERT3P(vmp, ==, NULL);
+ ASSERT0P(vmp);
cache = kmem_alloc(sizeof (*cache), KM_SLEEP);
strlcpy(cache->kc_name, name, sizeof (cache->kc_name));
diff --git a/sys/contrib/openzfs/module/os/freebsd/spl/spl_sysevent.c b/sys/contrib/openzfs/module/os/freebsd/spl/spl_sysevent.c
index 9da633c2b1be..3c2d39b20c09 100644
--- a/sys/contrib/openzfs/module/os/freebsd/spl/spl_sysevent.c
+++ b/sys/contrib/openzfs/module/os/freebsd/spl/spl_sysevent.c
@@ -256,7 +256,7 @@ sysevent_worker(void *arg __unused)
* free `ze`, so just inline the free() here -- events have already
* been drained.
*/
- VERIFY3P(ze->ze_zevent, ==, NULL);
+ VERIFY0P(ze->ze_zevent);
kmem_free(ze, sizeof (zfs_zevent_t));
kthread_exit();
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/abd_os.c b/sys/contrib/openzfs/module/os/freebsd/zfs/abd_os.c
index fbf67f6a14a8..4bf487cdc469 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/abd_os.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/abd_os.c
@@ -507,7 +507,7 @@ abd_iter_at_end(struct abd_iter *aiter)
void
abd_iter_advance(struct abd_iter *aiter, size_t amount)
{
- ASSERT3P(aiter->iter_mapaddr, ==, NULL);
+ ASSERT0P(aiter->iter_mapaddr);
ASSERT0(aiter->iter_mapsize);
/* There's nothing left to advance to, so do nothing */
@@ -526,7 +526,7 @@ abd_iter_map(struct abd_iter *aiter)
{
void *paddr;
- ASSERT3P(aiter->iter_mapaddr, ==, NULL);
+ ASSERT0P(aiter->iter_mapaddr);
ASSERT0(aiter->iter_mapsize);
/* There's nothing left to iterate over, so do nothing */
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/dmu_os.c b/sys/contrib/openzfs/module/os/freebsd/zfs/dmu_os.c
index 364bbfc60abd..26cc7981bfcd 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/dmu_os.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/dmu_os.c
@@ -156,7 +156,7 @@ dmu_read_pages(objset_t *os, uint64_t object, vm_page_t *ma, int count,
if (dbp[0]->db_offset != 0 || numbufs > 1) {
for (i = 0; i < numbufs; i++) {
ASSERT(ISP2(dbp[i]->db_size));
- ASSERT3U((dbp[i]->db_offset % dbp[i]->db_size), ==, 0);
+ ASSERT0((dbp[i]->db_offset % dbp[i]->db_size));
ASSERT3U(dbp[i]->db_size, ==, dbp[0]->db_size);
}
}
@@ -175,7 +175,7 @@ dmu_read_pages(objset_t *os, uint64_t object, vm_page_t *ma, int count,
vm_page_sunbusy(m);
break;
}
- ASSERT3U(m->dirty, ==, 0);
+ ASSERT0(m->dirty);
ASSERT(!pmap_page_is_write_mapped(m));
ASSERT3U(db->db_size, >, PAGE_SIZE);
@@ -201,7 +201,7 @@ dmu_read_pages(objset_t *os, uint64_t object, vm_page_t *ma, int count,
if (m != bogus_page) {
vm_page_assert_xbusied(m);
ASSERT(vm_page_none_valid(m));
- ASSERT3U(m->dirty, ==, 0);
+ ASSERT0(m->dirty);
ASSERT(!pmap_page_is_write_mapped(m));
va = zfs_map_page(m, &sf);
}
@@ -295,7 +295,7 @@ dmu_read_pages(objset_t *os, uint64_t object, vm_page_t *ma, int count,
vm_page_sunbusy(m);
break;
}
- ASSERT3U(m->dirty, ==, 0);
+ ASSERT0(m->dirty);
ASSERT(!pmap_page_is_write_mapped(m));
ASSERT3U(db->db_size, >, PAGE_SIZE);
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/kmod_core.c b/sys/contrib/openzfs/module/os/freebsd/zfs/kmod_core.c
index c114db14a916..b218c0da8125 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/kmod_core.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/kmod_core.c
@@ -112,7 +112,6 @@ static int zfs__fini(void);
static void zfs_shutdown(void *, int);
static eventhandler_tag zfs_shutdown_event_tag;
-static eventhandler_tag zfs_mountroot_event_tag;
#define ZFS_MIN_KSTACK_PAGES 4
@@ -311,9 +310,6 @@ zfs_modevent(module_t mod, int type, void *unused __unused)
zfs_shutdown_event_tag = EVENTHANDLER_REGISTER(
shutdown_post_sync, zfs_shutdown, NULL,
SHUTDOWN_PRI_FIRST);
- zfs_mountroot_event_tag = EVENTHANDLER_REGISTER(
- mountroot, spa_boot_init, NULL,
- SI_ORDER_ANY);
}
return (err);
case MOD_UNLOAD:
@@ -322,9 +318,6 @@ zfs_modevent(module_t mod, int type, void *unused __unused)
if (zfs_shutdown_event_tag != NULL)
EVENTHANDLER_DEREGISTER(shutdown_post_sync,
zfs_shutdown_event_tag);
- if (zfs_mountroot_event_tag != NULL)
- EVENTHANDLER_DEREGISTER(mountroot,
- zfs_mountroot_event_tag);
}
return (err);
case MOD_SHUTDOWN:
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/vdev_geom.c b/sys/contrib/openzfs/module/os/freebsd/zfs/vdev_geom.c
index c8ab7cc7cf8e..bbd1dafc69be 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/vdev_geom.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/vdev_geom.c
@@ -1236,7 +1236,7 @@ vdev_geom_io_done(zio_t *zio)
struct bio *bp = zio->io_bio;
if (zio->io_type != ZIO_TYPE_READ && zio->io_type != ZIO_TYPE_WRITE) {
- ASSERT3P(bp, ==, NULL);
+ ASSERT0P(bp);
return;
}
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_acl.c b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_acl.c
index 5c5adc6cc12b..b15a3e6e38c0 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_acl.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_acl.c
@@ -1632,7 +1632,7 @@ zfs_acl_ids_create(znode_t *dzp, int flag, vattr_t *vap, cred_t *cr,
if (zfsvfs->z_replay == B_FALSE)
ASSERT_VOP_ELOCKED(ZTOV(dzp), __func__);
} else
- ASSERT3P(dzp->z_vnode, ==, NULL);
+ ASSERT0P(dzp->z_vnode);
memset(acl_ids, 0, sizeof (zfs_acl_ids_t));
acl_ids->z_mode = MAKEIMODE(vap->va_type, vap->va_mode);
@@ -2014,7 +2014,7 @@ top:
error = zfs_aclset_common(zp, aclp, cr, tx);
ASSERT0(error);
- ASSERT3P(zp->z_acl_cached, ==, NULL);
+ ASSERT0P(zp->z_acl_cached);
zp->z_acl_cached = aclp;
if (fuid_dirtied)
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_ctldir.c b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_ctldir.c
index 8d0ff9b25e30..a222c5de4a2a 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_ctldir.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_ctldir.c
@@ -357,7 +357,7 @@ zfsctl_create(zfsvfs_t *zfsvfs)
vnode_t *rvp;
uint64_t crtime[2];
- ASSERT3P(zfsvfs->z_ctldir, ==, NULL);
+ ASSERT0P(zfsvfs->z_ctldir);
snapdir = sfs_alloc_node(sizeof (*snapdir), "snapshot", ZFSCTL_INO_ROOT,
ZFSCTL_INO_SNAPDIR);
@@ -494,7 +494,7 @@ zfsctl_common_getattr(vnode_t *vp, vattr_t *vap)
vap->va_uid = 0;
vap->va_gid = 0;
- vap->va_rdev = 0;
+ vap->va_rdev = NODEV;
/*
* We are a purely virtual object, so we have no
* blocksize or allocated blocks.
@@ -674,6 +674,7 @@ zfsctl_root_readdir(struct vop_readdir_args *ap)
zfs_uio_t uio;
int *eofp = ap->a_eofflag;
off_t dots_offset;
+ ssize_t orig_resid;
int error;
zfs_uio_init(&uio, ap->a_uio);
@@ -688,16 +689,16 @@ zfsctl_root_readdir(struct vop_readdir_args *ap)
* count to return is 0.
*/
if (zfs_uio_offset(&uio) == 3 * sizeof (entry)) {
+ if (eofp != NULL)
+ *eofp = 1;
return (0);
}
+ orig_resid = zfs_uio_resid(&uio);
error = sfs_readdir_common(zfsvfs->z_root, ZFSCTL_INO_ROOT, ap, &uio,
&dots_offset);
- if (error != 0) {
- if (error == ENAMETOOLONG) /* ran out of destination space */
- error = 0;
- return (error);
- }
+ if (error != 0)
+ goto err;
if (zfs_uio_offset(&uio) != dots_offset)
return (SET_ERROR(EINVAL));
@@ -710,8 +711,11 @@ zfsctl_root_readdir(struct vop_readdir_args *ap)
entry.d_reclen = sizeof (entry);
error = vfs_read_dirent(ap, &entry, zfs_uio_offset(&uio));
if (error != 0) {
- if (error == ENAMETOOLONG)
- error = 0;
+err:
+ if (error == ENAMETOOLONG) {
+ error = orig_resid == zfs_uio_resid(&uio) ?
+ EINVAL : 0;
+ }
return (SET_ERROR(error));
}
if (eofp != NULL)
@@ -1056,17 +1060,21 @@ zfsctl_snapdir_readdir(struct vop_readdir_args *ap)
zfs_uio_t uio;
int *eofp = ap->a_eofflag;
off_t dots_offset;
+ ssize_t orig_resid;
int error;
zfs_uio_init(&uio, ap->a_uio);
+ orig_resid = zfs_uio_resid(&uio);
ASSERT3S(vp->v_type, ==, VDIR);
error = sfs_readdir_common(ZFSCTL_INO_ROOT, ZFSCTL_INO_SNAPDIR, ap,
&uio, &dots_offset);
if (error != 0) {
- if (error == ENAMETOOLONG) /* ran out of destination space */
- error = 0;
+ if (error == ENAMETOOLONG) { /* ran out of destination space */
+ error = orig_resid == zfs_uio_resid(&uio) ?
+ EINVAL : 0;
+ }
return (error);
}
@@ -1084,9 +1092,13 @@ zfsctl_snapdir_readdir(struct vop_readdir_args *ap)
dsl_pool_config_exit(dmu_objset_pool(zfsvfs->z_os), FTAG);
if (error != 0) {
if (error == ENOENT) {
- if (eofp != NULL)
- *eofp = 1;
- error = 0;
+ if (orig_resid == zfs_uio_resid(&uio)) {
+ error = EINVAL;
+ } else {
+ error = 0;
+ if (eofp != NULL)
+ *eofp = 1;
+ }
}
zfs_exit(zfsvfs, FTAG);
return (error);
@@ -1099,8 +1111,10 @@ zfsctl_snapdir_readdir(struct vop_readdir_args *ap)
entry.d_reclen = sizeof (entry);
error = vfs_read_dirent(ap, &entry, zfs_uio_offset(&uio));
if (error != 0) {
- if (error == ENAMETOOLONG)
- error = 0;
+ if (error == ENAMETOOLONG) {
+ error = orig_resid == zfs_uio_resid(&uio) ?
+ EINVAL : 0;
+ }
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(error));
}
@@ -1367,7 +1381,7 @@ zfsctl_snapshot_unmount(const char *snapname, int flags __unused)
int err = getzfsvfs(snapname, &zfsvfs);
if (err != 0) {
- ASSERT3P(zfsvfs, ==, NULL);
+ ASSERT0P(zfsvfs);
return (0);
}
vfsp = zfsvfs->z_vfs;
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_dir.c b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_dir.c
index 191df832d726..75ba2ea0cb9e 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_dir.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_dir.c
@@ -273,7 +273,7 @@ zfs_unlinked_add(znode_t *zp, dmu_tx_t *tx)
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
ASSERT(zp->z_unlinked);
- ASSERT3U(zp->z_links, ==, 0);
+ ASSERT0(zp->z_links);
VERIFY0(zap_add_int(zfsvfs->z_os, zfsvfs->z_unlinkedobj, zp->z_id, tx));
@@ -437,7 +437,7 @@ zfs_rmnode(znode_t *zp)
uint64_t count;
int error;
- ASSERT3U(zp->z_links, ==, 0);
+ ASSERT0(zp->z_links);
if (zfsvfs->z_replay == B_FALSE)
ASSERT_VOP_ELOCKED(ZTOV(zp), __func__);
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vfsops.c b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vfsops.c
index 0456552ed07e..79b784288911 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vfsops.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vfsops.c
@@ -455,8 +455,13 @@ zfs_sync(vfs_t *vfsp, int waitfor)
return (0);
}
- if (zfsvfs->z_log != NULL)
- zil_commit(zfsvfs->z_log, 0);
+ if (zfsvfs->z_log != NULL) {
+ error = zil_commit(zfsvfs->z_log, 0);
+ if (error != 0) {
+ zfs_exit(zfsvfs, FTAG);
+ return (error);
+ }
+ }
zfs_exit(zfsvfs, FTAG);
} else {
@@ -1091,7 +1096,7 @@ zfsvfs_setup(zfsvfs_t *zfsvfs, boolean_t mounting)
if (mounting) {
boolean_t readonly;
- ASSERT3P(zfsvfs->z_kstat.dk_kstats, ==, NULL);
+ ASSERT0P(zfsvfs->z_kstat.dk_kstats);
error = dataset_kstats_create(&zfsvfs->z_kstat, zfsvfs->z_os);
if (error)
return (error);
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vnops_os.c b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vnops_os.c
index c4270d8b5d5c..174141a5deab 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vnops_os.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vnops_os.c
@@ -1101,7 +1101,7 @@ zfs_create(znode_t *dzp, const char *name, vattr_t *vap, int excl, int mode,
zfs_exit(zfsvfs, FTAG);
return (error);
}
- ASSERT3P(zp, ==, NULL);
+ ASSERT0P(zp);
/*
* Create a new file object and update the directory
@@ -1193,8 +1193,8 @@ out:
*zpp = zp;
}
- if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ if (error == 0 && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
+ error = zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
@@ -1323,9 +1323,8 @@ out:
if (xzp)
vrele(ZTOV(xzp));
- if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
-
+ if (error == 0 && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
+ error = zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
@@ -1482,7 +1481,7 @@ zfs_mkdir(znode_t *dzp, const char *dirname, vattr_t *vap, znode_t **zpp,
zfs_exit(zfsvfs, FTAG);
return (error);
}
- ASSERT3P(zp, ==, NULL);
+ ASSERT0P(zp);
if ((error = zfs_zaccess(dzp, ACE_ADD_SUBDIRECTORY, 0, B_FALSE, cr,
mnt_ns))) {
@@ -1556,8 +1555,8 @@ out:
getnewvnode_drop_reserve();
- if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ if (error == 0 && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
+ error = zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
@@ -1637,8 +1636,8 @@ zfs_rmdir_(vnode_t *dvp, vnode_t *vp, const char *name, cred_t *cr)
if (zfsvfs->z_use_namecache)
cache_vop_rmdir(dvp, vp);
out:
- if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ if (error == 0 && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
+ error = zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
@@ -1696,6 +1695,7 @@ zfs_readdir(vnode_t *vp, zfs_uio_t *uio, cred_t *cr, int *eofp,
objset_t *os;
caddr_t outbuf;
size_t bufsize;
+ ssize_t orig_resid;
zap_cursor_t zc;
zap_attribute_t *zap;
uint_t bytes_wanted;
@@ -1736,7 +1736,7 @@ zfs_readdir(vnode_t *vp, zfs_uio_t *uio, cred_t *cr, int *eofp,
/*
* Quit if directory has been removed (posix)
*/
- if ((*eofp = zp->z_unlinked) != 0) {
+ if ((*eofp = (zp->z_unlinked != 0)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (0);
}
@@ -1744,6 +1744,7 @@ zfs_readdir(vnode_t *vp, zfs_uio_t *uio, cred_t *cr, int *eofp,
error = 0;
os = zfsvfs->z_os;
offset = zfs_uio_offset(uio);
+ orig_resid = zfs_uio_resid(uio);
prefetch = zp->z_zn_prefetch;
zap = zap_attribute_long_alloc();
@@ -1923,7 +1924,7 @@ update:
kmem_free(outbuf, bufsize);
if (error == ENOENT)
- error = 0;
+ error = orig_resid == zfs_uio_resid(uio) ? EINVAL : 0;
ZFS_ACCESSTIME_STAMP(zfsvfs, zp);
@@ -2014,7 +2015,7 @@ zfs_getattr(vnode_t *vp, vattr_t *vap, int flags, cred_t *cr)
if (vp->v_type == VBLK || vp->v_type == VCHR)
vap->va_rdev = zfs_cmpldev(rdev);
else
- vap->va_rdev = 0;
+ vap->va_rdev = NODEV;
vap->va_gen = zp->z_gen;
vap->va_flags = 0; /* FreeBSD: Reset chflags(2) flags. */
vap->va_filerev = zp->z_seq;
@@ -3009,8 +3010,8 @@ out:
}
out2:
- if (os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ if (err == 0 && os->os_sync == ZFS_SYNC_ALWAYS)
+ err = zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (err);
@@ -3539,7 +3540,7 @@ out_seq:
out:
if (error == 0 && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ error = zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
@@ -3731,7 +3732,7 @@ zfs_symlink(znode_t *dzp, const char *name, vattr_t *vap,
*zpp = zp;
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ error = zil_commit(zilog, 0);
}
zfs_exit(zfsvfs, FTAG);
@@ -3921,8 +3922,8 @@ zfs_link(znode_t *tdzp, znode_t *szp, const char *name, cred_t *cr,
vnevent_link(ZTOV(szp), ct);
}
- if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ if (error == 0 && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
+ error = zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
@@ -4313,7 +4314,7 @@ typedef struct {
} putpage_commit_arg_t;
static void
-zfs_putpage_commit_cb(void *arg)
+zfs_putpage_commit_cb(void *arg, int err)
{
putpage_commit_arg_t *pca = arg;
vm_object_t object = pca->pca_pages[0]->object;
@@ -4322,7 +4323,17 @@ zfs_putpage_commit_cb(void *arg)
for (uint_t i = 0; i < pca->pca_npages; i++) {
vm_page_t pp = pca->pca_pages[i];
- vm_page_undirty(pp);
+
+ if (err == 0) {
+ /*
+ * Writeback succeeded, so undirty the page. If it
+ * fails, we leave it in the same state it was. That's
+ * most likely dirty, so it will get tried again some
+ * other time.
+ */
+ vm_page_undirty(pp);
+ }
+
vm_page_sunbusy(pp);
}
@@ -4510,8 +4521,13 @@ zfs_putpages(struct vnode *vp, vm_page_t *ma, size_t len, int flags,
out:
zfs_rangelock_exit(lr);
- if (commit)
- zil_commit(zfsvfs->z_log, zp->z_id);
+ if (commit) {
+ err = zil_commit(zfsvfs->z_log, zp->z_id);
+ if (err != 0) {
+ zfs_exit(zfsvfs, FTAG);
+ return (err);
+ }
+ }
dataset_kstats_update_write_kstats(&zfsvfs->z_kstat, len);
@@ -5223,8 +5239,32 @@ struct vop_fsync_args {
static int
zfs_freebsd_fsync(struct vop_fsync_args *ap)
{
+ vnode_t *vp = ap->a_vp;
+ int err = 0;
+
+ /*
+ * Push any dirty mmap()'d data out to the DMU and ZIL, ready for
+ * zil_commit() to be called in zfs_fsync().
+ */
+ if (vm_object_mightbedirty(vp->v_object)) {
+ zfs_vmobject_wlock(vp->v_object);
+ if (!vm_object_page_clean(vp->v_object, 0, 0, 0))
+ err = SET_ERROR(EIO);
+ zfs_vmobject_wunlock(vp->v_object);
+ if (err) {
+ /*
+ * Unclear what state things are in. zfs_putpages()
+ * will ensure the pages remain dirty if they haven't
+ * been written down to the DMU, but because there may
+ * be nothing logged, we can't assume that zfs_sync()
+ * -> zil_commit() will give us a useful error. It's
+ * safest if we just error out here.
+ */
+ return (err);
+ }
+ }
- return (zfs_fsync(VTOZ(ap->a_vp), 0, ap->a_td->td_ucred));
+ return (zfs_fsync(VTOZ(vp), 0, ap->a_td->td_ucred));
}
#ifndef _SYS_SYSPROTO_H_
@@ -6773,9 +6813,11 @@ zfs_deallocate(struct vop_deallocate_args *ap)
if (error == 0) {
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS ||
(ap->a_ioflag & IO_SYNC) != 0)
- zil_commit(zilog, zp->z_id);
- *ap->a_offset = off + len;
- *ap->a_len = 0;
+ error = zil_commit(zilog, zp->z_id);
+ if (error == 0) {
+ *ap->a_offset = off + len;
+ *ap->a_len = 0;
+ }
}
zfs_exit(zfsvfs, FTAG);
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_znode_os.c b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_znode_os.c
index 775f54a65f7d..7cd0a153577c 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_znode_os.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_znode_os.c
@@ -161,15 +161,15 @@ zfs_znode_cache_destructor(void *buf, void *arg)
znode_t *zp = buf;
ASSERT(!POINTER_IS_VALID(zp->z_zfsvfs));
- ASSERT3P(zp->z_vnode, ==, NULL);
+ ASSERT0P(zp->z_vnode);
ASSERT(!list_link_active(&zp->z_link_node));
mutex_destroy(&zp->z_lock);
mutex_destroy(&zp->z_acl_lock);
rw_destroy(&zp->z_xattr_lock);
zfs_rangelock_fini(&zp->z_rangelock);
- ASSERT3P(zp->z_acl_cached, ==, NULL);
- ASSERT3P(zp->z_xattr_cached, ==, NULL);
+ ASSERT0P(zp->z_acl_cached);
+ ASSERT0P(zp->z_xattr_cached);
}
@@ -195,7 +195,7 @@ zfs_znode_init(void)
/*
* Initialize zcache
*/
- ASSERT3P(znode_uma_zone, ==, NULL);
+ ASSERT0P(znode_uma_zone);
znode_uma_zone = uma_zcreate("zfs_znode_cache",
sizeof (znode_t), zfs_znode_cache_constructor_smr,
zfs_znode_cache_destructor_smr, NULL, NULL, 0, 0);
@@ -224,7 +224,7 @@ zfs_znode_init(void)
/*
* Initialize zcache
*/
- ASSERT3P(znode_cache, ==, NULL);
+ ASSERT0P(znode_cache);
znode_cache = kmem_cache_create("zfs_znode_cache",
sizeof (znode_t), 0, zfs_znode_cache_constructor,
zfs_znode_cache_destructor, NULL, NULL, NULL, KMC_RECLAIMABLE);
@@ -353,8 +353,8 @@ zfs_znode_sa_init(zfsvfs_t *zfsvfs, znode_t *zp,
ASSERT(!POINTER_IS_VALID(zp->z_zfsvfs) || (zfsvfs == zp->z_zfsvfs));
ASSERT(MUTEX_HELD(ZFS_OBJ_MUTEX(zfsvfs, zp->z_id)));
- ASSERT3P(zp->z_sa_hdl, ==, NULL);
- ASSERT3P(zp->z_acl_cached, ==, NULL);
+ ASSERT0P(zp->z_sa_hdl);
+ ASSERT0P(zp->z_acl_cached);
if (sa_hdl == NULL) {
VERIFY0(sa_handle_get_from_db(zfsvfs->z_os, db, zp,
SA_HDL_SHARED, &zp->z_sa_hdl));
@@ -1127,7 +1127,7 @@ zfs_rezget(znode_t *zp)
}
rw_exit(&zp->z_xattr_lock);
- ASSERT3P(zp->z_sa_hdl, ==, NULL);
+ ASSERT0P(zp->z_sa_hdl);
err = sa_buf_hold(zfsvfs->z_os, obj_num, NULL, &db);
if (err) {
ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num);
@@ -1298,7 +1298,7 @@ zfs_znode_free(znode_t *zp)
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
char *symlink;
- ASSERT3P(zp->z_sa_hdl, ==, NULL);
+ ASSERT0P(zp->z_sa_hdl);
zp->z_vnode = NULL;
mutex_enter(&zfsvfs->z_znodes_lock);
POINTER_INVALIDATE(&zp->z_zfsvfs);
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/zvol_os.c b/sys/contrib/openzfs/module/os/freebsd/zfs/zvol_os.c
index 72a7c4ea082a..0dd2ecd7fd8d 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/zvol_os.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/zvol_os.c
@@ -31,7 +31,7 @@
* Copyright (c) 2012, 2017 by Delphix. All rights reserved.
* Copyright (c) 2013, Joyent, Inc. All rights reserved.
* Copyright (c) 2014 Integros [integros.com]
- * Copyright (c) 2024, Klara, Inc.
+ * Copyright (c) 2024, 2025, Klara, Inc.
*/
/* Portions Copyright 2011 Martin Matuska <mm@FreeBSD.org> */
@@ -196,7 +196,6 @@ DECLARE_GEOM_CLASS(zfs_zvol_class, zfs_zvol);
static int zvol_geom_open(struct g_provider *pp, int flag, int count);
static int zvol_geom_close(struct g_provider *pp, int flag, int count);
-static void zvol_geom_destroy(zvol_state_t *zv);
static int zvol_geom_access(struct g_provider *pp, int acr, int acw, int ace);
static void zvol_geom_bio_start(struct bio *bp);
static int zvol_geom_bio_getattr(struct bio *bp);
@@ -226,25 +225,14 @@ zvol_geom_open(struct g_provider *pp, int flag, int count)
}
retry:
- rw_enter(&zvol_state_lock, ZVOL_RW_READER);
- /*
- * Obtain a copy of private under zvol_state_lock to make sure either
- * the result of zvol free code setting private to NULL is observed,
- * or the zv is protected from being freed because of the positive
- * zv_open_count.
- */
- zv = pp->private;
- if (zv == NULL) {
- rw_exit(&zvol_state_lock);
- err = SET_ERROR(ENXIO);
- goto out_locked;
- }
+ zv = atomic_load_ptr(&pp->private);
+ if (zv == NULL)
+ return (SET_ERROR(ENXIO));
mutex_enter(&zv->zv_state_lock);
if (zv->zv_zso->zso_dying || zv->zv_flags & ZVOL_REMOVING) {
- rw_exit(&zvol_state_lock);
err = SET_ERROR(ENXIO);
- goto out_zv_locked;
+ goto out_locked;
}
ASSERT3S(zv->zv_volmode, ==, ZFS_VOLMODE_GEOM);
@@ -257,8 +245,24 @@ retry:
drop_suspend = B_TRUE;
if (!rw_tryenter(&zv->zv_suspend_lock, ZVOL_RW_READER)) {
mutex_exit(&zv->zv_state_lock);
+
+ /*
+ * Removal may happen while the locks are down, so
+ * we can't trust zv any longer; we have to start over.
+ */
+ zv = atomic_load_ptr(&pp->private);
+ if (zv == NULL)
+ return (SET_ERROR(ENXIO));
+
rw_enter(&zv->zv_suspend_lock, ZVOL_RW_READER);
mutex_enter(&zv->zv_state_lock);
+
+ if (zv->zv_zso->zso_dying ||
+ zv->zv_flags & ZVOL_REMOVING) {
+ err = SET_ERROR(ENXIO);
+ goto out_locked;
+ }
+
/* Check to see if zv_suspend_lock is needed. */
if (zv->zv_open_count != 0) {
rw_exit(&zv->zv_suspend_lock);
@@ -266,7 +270,6 @@ retry:
}
}
}
- rw_exit(&zvol_state_lock);
ASSERT(MUTEX_HELD(&zv->zv_state_lock));
@@ -294,7 +297,7 @@ retry:
if (drop_namespace)
mutex_exit(&spa_namespace_lock);
if (err)
- goto out_zv_locked;
+ goto out_locked;
pp->mediasize = zv->zv_volsize;
pp->stripeoffset = 0;
pp->stripesize = zv->zv_volblocksize;
@@ -329,9 +332,8 @@ out_opened:
zvol_last_close(zv);
wakeup(zv);
}
-out_zv_locked:
- mutex_exit(&zv->zv_state_lock);
out_locked:
+ mutex_exit(&zv->zv_state_lock);
if (drop_suspend)
rw_exit(&zv->zv_suspend_lock);
return (err);
@@ -345,12 +347,9 @@ zvol_geom_close(struct g_provider *pp, int flag, int count)
boolean_t drop_suspend = B_TRUE;
int new_open_count;
- rw_enter(&zvol_state_lock, ZVOL_RW_READER);
- zv = pp->private;
- if (zv == NULL) {
- rw_exit(&zvol_state_lock);
+ zv = atomic_load_ptr(&pp->private);
+ if (zv == NULL)
return (SET_ERROR(ENXIO));
- }
mutex_enter(&zv->zv_state_lock);
if (zv->zv_flags & ZVOL_EXCL) {
@@ -377,6 +376,15 @@ zvol_geom_close(struct g_provider *pp, int flag, int count)
mutex_exit(&zv->zv_state_lock);
rw_enter(&zv->zv_suspend_lock, ZVOL_RW_READER);
mutex_enter(&zv->zv_state_lock);
+
+ /*
+ * Unlike in zvol_geom_open(), we don't check if
+ * removal started here, because we might be one of the
+ * openers that needs to be thrown out! If we're the
+ * last, we need to call zvol_last_close() below to
+ * finish cleanup. So, no special treatment for us.
+ */
+
/* Check to see if zv_suspend_lock is needed. */
new_open_count = zv->zv_open_count - count;
if (new_open_count != 0) {
@@ -387,7 +395,6 @@ zvol_geom_close(struct g_provider *pp, int flag, int count)
} else {
drop_suspend = B_FALSE;
}
- rw_exit(&zvol_state_lock);
ASSERT(MUTEX_HELD(&zv->zv_state_lock));
@@ -408,20 +415,6 @@ zvol_geom_close(struct g_provider *pp, int flag, int count)
return (0);
}
-static void
-zvol_geom_destroy(zvol_state_t *zv)
-{
- struct zvol_state_geom *zsg = &zv->zv_zso->zso_geom;
- struct g_provider *pp = zsg->zsg_provider;
-
- ASSERT3S(zv->zv_volmode, ==, ZFS_VOLMODE_GEOM);
-
- g_topology_assert();
-
- zsg->zsg_provider = NULL;
- g_wither_geom(pp->geom, ENXIO);
-}
-
void
zvol_wait_close(zvol_state_t *zv)
{
@@ -454,7 +447,7 @@ zvol_geom_access(struct g_provider *pp, int acr, int acw, int ace)
("Unsupported access request to %s (acr=%d, acw=%d, ace=%d).",
pp->name, acr, acw, ace));
- if (pp->private == NULL) {
+ if (atomic_load_ptr(&pp->private) == NULL) {
if (acr <= 0 && acw <= 0 && ace <= 0)
return (0);
return (pp->error);
@@ -727,9 +720,9 @@ unlock:
break;
}
- if (commit) {
+ if (error == 0 && commit) {
commit:
- zil_commit(zv->zv_zilog, ZVOL_OBJ);
+ error = zil_commit(zv->zv_zilog, ZVOL_OBJ);
}
resume:
rw_exit(&zv->zv_suspend_lock);
@@ -906,8 +899,8 @@ zvol_cdev_write(struct cdev *dev, struct uio *uio_s, int ioflag)
zfs_rangelock_exit(lr);
int64_t nwritten = start_resid - zfs_uio_resid(&uio);
dataset_kstats_update_write_kstats(&zv->zv_kstat, nwritten);
- if (commit)
- zil_commit(zv->zv_zilog, ZVOL_OBJ);
+ if (error == 0 && commit)
+ error = zil_commit(zv->zv_zilog, ZVOL_OBJ);
rw_exit(&zv->zv_suspend_lock);
return (error);
@@ -921,25 +914,14 @@ zvol_cdev_open(struct cdev *dev, int flags, int fmt, struct thread *td)
boolean_t drop_suspend = B_FALSE;
retry:
- rw_enter(&zvol_state_lock, ZVOL_RW_READER);
- /*
- * Obtain a copy of si_drv2 under zvol_state_lock to make sure either
- * the result of zvol free code setting si_drv2 to NULL is observed,
- * or the zv is protected from being freed because of the positive
- * zv_open_count.
- */
- zv = dev->si_drv2;
- if (zv == NULL) {
- rw_exit(&zvol_state_lock);
- err = SET_ERROR(ENXIO);
- goto out_locked;
- }
+ zv = atomic_load_ptr(&dev->si_drv2);
+ if (zv == NULL)
+ return (SET_ERROR(ENXIO));
mutex_enter(&zv->zv_state_lock);
- if (zv->zv_zso->zso_dying) {
- rw_exit(&zvol_state_lock);
+ if (zv->zv_zso->zso_dying || zv->zv_flags & ZVOL_REMOVING) {
err = SET_ERROR(ENXIO);
- goto out_zv_locked;
+ goto out_locked;
}
ASSERT3S(zv->zv_volmode, ==, ZFS_VOLMODE_DEV);
@@ -954,6 +936,13 @@ retry:
mutex_exit(&zv->zv_state_lock);
rw_enter(&zv->zv_suspend_lock, ZVOL_RW_READER);
mutex_enter(&zv->zv_state_lock);
+
+ if (unlikely(zv->zv_flags & ZVOL_REMOVING)) {
+ /* Removal started while locks were down. */
+ err = SET_ERROR(ENXIO);
+ goto out_locked;
+ }
+
/* Check to see if zv_suspend_lock is needed. */
if (zv->zv_open_count != 0) {
rw_exit(&zv->zv_suspend_lock);
@@ -961,7 +950,6 @@ retry:
}
}
}
- rw_exit(&zvol_state_lock);
ASSERT(MUTEX_HELD(&zv->zv_state_lock));
@@ -989,7 +977,7 @@ retry:
if (drop_namespace)
mutex_exit(&spa_namespace_lock);
if (err)
- goto out_zv_locked;
+ goto out_locked;
}
ASSERT(MUTEX_HELD(&zv->zv_state_lock));
@@ -1016,9 +1004,8 @@ out_opened:
zvol_last_close(zv);
wakeup(zv);
}
-out_zv_locked:
- mutex_exit(&zv->zv_state_lock);
out_locked:
+ mutex_exit(&zv->zv_state_lock);
if (drop_suspend)
rw_exit(&zv->zv_suspend_lock);
return (err);
@@ -1030,12 +1017,9 @@ zvol_cdev_close(struct cdev *dev, int flags, int fmt, struct thread *td)
zvol_state_t *zv;
boolean_t drop_suspend = B_TRUE;
- rw_enter(&zvol_state_lock, ZVOL_RW_READER);
- zv = dev->si_drv2;
- if (zv == NULL) {
- rw_exit(&zvol_state_lock);
+ zv = atomic_load_ptr(&dev->si_drv2);
+ if (zv == NULL)
return (SET_ERROR(ENXIO));
- }
mutex_enter(&zv->zv_state_lock);
if (zv->zv_flags & ZVOL_EXCL) {
@@ -1060,6 +1044,15 @@ zvol_cdev_close(struct cdev *dev, int flags, int fmt, struct thread *td)
mutex_exit(&zv->zv_state_lock);
rw_enter(&zv->zv_suspend_lock, ZVOL_RW_READER);
mutex_enter(&zv->zv_state_lock);
+
+ /*
+ * Unlike in zvol_cdev_open(), we don't check if
+ * removal started here, because we might be one of the
+ * openers that needs to be thrown out! If we're the
+ * last, we need to call zvol_last_close() below to
+ * finish cleanup. So, no special treatment for us.
+ */
+
/* Check to see if zv_suspend_lock is needed. */
if (zv->zv_open_count != 1) {
rw_exit(&zv->zv_suspend_lock);
@@ -1069,7 +1062,6 @@ zvol_cdev_close(struct cdev *dev, int flags, int fmt, struct thread *td)
} else {
drop_suspend = B_FALSE;
}
- rw_exit(&zvol_state_lock);
ASSERT(MUTEX_HELD(&zv->zv_state_lock));
@@ -1101,7 +1093,8 @@ zvol_cdev_ioctl(struct cdev *dev, ulong_t cmd, caddr_t data,
int error;
boolean_t sync;
- zv = dev->si_drv2;
+ zv = atomic_load_ptr(&dev->si_drv2);
+ ASSERT3P(zv, !=, NULL);
error = 0;
KASSERT(zv->zv_open_count > 0,
@@ -1117,7 +1110,7 @@ zvol_cdev_ioctl(struct cdev *dev, ulong_t cmd, caddr_t data,
case DIOCGFLUSH:
rw_enter(&zv->zv_suspend_lock, ZVOL_RW_READER);
if (zv->zv_zilog != NULL)
- zil_commit(zv->zv_zilog, ZVOL_OBJ);
+ error = zil_commit(zv->zv_zilog, ZVOL_OBJ);
rw_exit(&zv->zv_suspend_lock);
break;
case DIOCGDELETE:
@@ -1152,7 +1145,7 @@ zvol_cdev_ioctl(struct cdev *dev, ulong_t cmd, caddr_t data,
}
zfs_rangelock_exit(lr);
if (sync)
- zil_commit(zv->zv_zilog, ZVOL_OBJ);
+ error = zil_commit(zv->zv_zilog, ZVOL_OBJ);
rw_exit(&zv->zv_suspend_lock);
break;
case DIOCGSTRIPESIZE:
@@ -1162,6 +1155,7 @@ zvol_cdev_ioctl(struct cdev *dev, ulong_t cmd, caddr_t data,
*(off_t *)data = 0;
break;
case DIOCGATTR: {
+ rw_enter(&zv->zv_suspend_lock, ZVOL_RW_READER);
spa_t *spa = dmu_objset_spa(zv->zv_objset);
struct diocgattr_arg *arg = (struct diocgattr_arg *)data;
uint64_t refd, avail, usedobjs, availobjs;
@@ -1186,6 +1180,7 @@ zvol_cdev_ioctl(struct cdev *dev, ulong_t cmd, caddr_t data,
arg->value.off = refd / DEV_BSIZE;
} else
error = SET_ERROR(ENOIOCTL);
+ rw_exit(&zv->zv_suspend_lock);
break;
}
case FIOSEEKHOLE:
@@ -1196,10 +1191,12 @@ zvol_cdev_ioctl(struct cdev *dev, ulong_t cmd, caddr_t data,
hole = (cmd == FIOSEEKHOLE);
noff = *off;
+ rw_enter(&zv->zv_suspend_lock, ZVOL_RW_READER);
lr = zfs_rangelock_enter(&zv->zv_rangelock, 0, UINT64_MAX,
RL_READER);
error = dmu_offset_next(zv->zv_objset, ZVOL_OBJ, hole, &noff);
zfs_rangelock_exit(lr);
+ rw_exit(&zv->zv_suspend_lock);
*off = noff;
break;
}
@@ -1400,42 +1397,65 @@ zvol_alloc(const char *name, uint64_t volsize, uint64_t volblocksize,
* Remove minor node for the specified volume.
*/
void
-zvol_os_free(zvol_state_t *zv)
+zvol_os_remove_minor(zvol_state_t *zv)
{
- ASSERT(!RW_LOCK_HELD(&zv->zv_suspend_lock));
- ASSERT(!MUTEX_HELD(&zv->zv_state_lock));
+ ASSERT(MUTEX_HELD(&zv->zv_state_lock));
ASSERT0(zv->zv_open_count);
+ ASSERT0(atomic_read(&zv->zv_suspend_ref));
+ ASSERT(zv->zv_flags & ZVOL_REMOVING);
- ZFS_LOG(1, "ZVOL %s destroyed.", zv->zv_name);
-
- rw_destroy(&zv->zv_suspend_lock);
- zfs_rangelock_fini(&zv->zv_rangelock);
+ struct zvol_state_os *zso = zv->zv_zso;
+ zv->zv_zso = NULL;
if (zv->zv_volmode == ZFS_VOLMODE_GEOM) {
- struct zvol_state_geom *zsg = &zv->zv_zso->zso_geom;
- struct g_provider *pp __maybe_unused = zsg->zsg_provider;
-
- ASSERT3P(pp->private, ==, NULL);
+ struct zvol_state_geom *zsg = &zso->zso_geom;
+ struct g_provider *pp = zsg->zsg_provider;
+ atomic_store_ptr(&pp->private, NULL);
+ mutex_exit(&zv->zv_state_lock);
g_topology_lock();
- zvol_geom_destroy(zv);
+ g_wither_geom(pp->geom, ENXIO);
g_topology_unlock();
} else if (zv->zv_volmode == ZFS_VOLMODE_DEV) {
- struct zvol_state_dev *zsd = &zv->zv_zso->zso_dev;
+ struct zvol_state_dev *zsd = &zso->zso_dev;
struct cdev *dev = zsd->zsd_cdev;
+ if (dev != NULL)
+ atomic_store_ptr(&dev->si_drv2, NULL);
+ mutex_exit(&zv->zv_state_lock);
+
if (dev != NULL) {
- ASSERT3P(dev->si_drv2, ==, NULL);
destroy_dev(dev);
knlist_clear(&zsd->zsd_selinfo.si_note, 0);
knlist_destroy(&zsd->zsd_selinfo.si_note);
}
}
+ kmem_free(zso, sizeof (struct zvol_state_os));
+
+ mutex_enter(&zv->zv_state_lock);
+}
+
+void
+zvol_os_free(zvol_state_t *zv)
+{
+ ASSERT(!RW_LOCK_HELD(&zv->zv_suspend_lock));
+ ASSERT(!MUTEX_HELD(&zv->zv_state_lock));
+ ASSERT0(zv->zv_open_count);
+ ASSERT0P(zv->zv_zso);
+
+ ASSERT0P(zv->zv_objset);
+ ASSERT0P(zv->zv_zilog);
+ ASSERT0P(zv->zv_dn);
+
+ ZFS_LOG(1, "ZVOL %s destroyed.", zv->zv_name);
+
+ rw_destroy(&zv->zv_suspend_lock);
+ zfs_rangelock_fini(&zv->zv_rangelock);
+
mutex_destroy(&zv->zv_state_lock);
cv_destroy(&zv->zv_removing_cv);
dataset_kstats_destroy(&zv->zv_kstat);
- kmem_free(zv->zv_zso, sizeof (struct zvol_state_os));
kmem_free(zv, sizeof (zvol_state_t));
zvol_minors--;
}
@@ -1493,11 +1513,11 @@ zvol_os_create_minor(const char *name)
zv->zv_objset = os;
- ASSERT3P(zv->zv_kstat.dk_kstats, ==, NULL);
+ ASSERT0P(zv->zv_kstat.dk_kstats);
error = dataset_kstats_create(&zv->zv_kstat, zv->zv_objset);
if (error)
goto out_dmu_objset_disown;
- ASSERT3P(zv->zv_zilog, ==, NULL);
+ ASSERT0P(zv->zv_zilog);
zv->zv_zilog = zil_open(os, zvol_get_data, &zv->zv_kstat.dk_zil_sums);
if (spa_writeable(dmu_objset_spa(os))) {
if (zil_replay_disable)
@@ -1538,28 +1558,6 @@ out_doi:
return (error);
}
-void
-zvol_os_clear_private(zvol_state_t *zv)
-{
- ASSERT(RW_LOCK_HELD(&zvol_state_lock));
- if (zv->zv_volmode == ZFS_VOLMODE_GEOM) {
- struct zvol_state_geom *zsg = &zv->zv_zso->zso_geom;
- struct g_provider *pp = zsg->zsg_provider;
-
- if (pp->private == NULL) /* already cleared */
- return;
-
- pp->private = NULL;
- ASSERT(!RW_LOCK_HELD(&zv->zv_suspend_lock));
- } else if (zv->zv_volmode == ZFS_VOLMODE_DEV) {
- struct zvol_state_dev *zsd = &zv->zv_zso->zso_dev;
- struct cdev *dev = zsd->zsd_cdev;
-
- if (dev != NULL)
- dev->si_drv2 = NULL;
- }
-}
-
int
zvol_os_update_volsize(zvol_state_t *zv, uint64_t volsize)
{
diff --git a/sys/contrib/openzfs/module/os/linux/spl/spl-condvar.c b/sys/contrib/openzfs/module/os/linux/spl/spl-condvar.c
index ce9c9e39e60c..aac5f2ebbfd2 100644
--- a/sys/contrib/openzfs/module/os/linux/spl/spl-condvar.c
+++ b/sys/contrib/openzfs/module/os/linux/spl/spl-condvar.c
@@ -66,9 +66,9 @@ void
__cv_init(kcondvar_t *cvp, char *name, kcv_type_t type, void *arg)
{
ASSERT(cvp);
- ASSERT(name == NULL);
+ ASSERT0P(name);
ASSERT(type == CV_DEFAULT);
- ASSERT(arg == NULL);
+ ASSERT0P(arg);
cvp->cv_magic = CV_MAGIC;
init_waitqueue_head(&cvp->cv_event);
@@ -83,7 +83,7 @@ static int
cv_destroy_wakeup(kcondvar_t *cvp)
{
if (!atomic_read(&cvp->cv_waiters) && !atomic_read(&cvp->cv_refs)) {
- ASSERT(cvp->cv_mutex == NULL);
+ ASSERT0P(cvp->cv_mutex);
ASSERT(!waitqueue_active(&cvp->cv_event));
return (1);
}
@@ -104,7 +104,7 @@ __cv_destroy(kcondvar_t *cvp)
while (cv_destroy_wakeup(cvp) == 0)
wait_event_timeout(cvp->cv_destroy, cv_destroy_wakeup(cvp), 1);
- ASSERT3P(cvp->cv_mutex, ==, NULL);
+ ASSERT0P(cvp->cv_mutex);
ASSERT3S(atomic_read(&cvp->cv_refs), ==, 0);
ASSERT3S(atomic_read(&cvp->cv_waiters), ==, 0);
ASSERT3S(waitqueue_active(&cvp->cv_event), ==, 0);
diff --git a/sys/contrib/openzfs/module/os/linux/spl/spl-generic.c b/sys/contrib/openzfs/module/os/linux/spl/spl-generic.c
index f37699b4347e..89ca4a648b2f 100644
--- a/sys/contrib/openzfs/module/os/linux/spl/spl-generic.c
+++ b/sys/contrib/openzfs/module/os/linux/spl/spl-generic.c
@@ -709,7 +709,7 @@ zone_get_hostid(void *zone)
{
uint32_t hostid;
- ASSERT3P(zone, ==, NULL);
+ ASSERT0P(zone);
if (spl_hostid != 0)
return ((uint32_t)(spl_hostid & HW_HOSTID_MASK));
diff --git a/sys/contrib/openzfs/module/os/linux/spl/spl-kmem-cache.c b/sys/contrib/openzfs/module/os/linux/spl/spl-kmem-cache.c
index fab80289b278..22e4ed169d03 100644
--- a/sys/contrib/openzfs/module/os/linux/spl/spl-kmem-cache.c
+++ b/sys/contrib/openzfs/module/os/linux/spl/spl-kmem-cache.c
@@ -296,7 +296,7 @@ spl_slab_free(spl_kmem_slab_t *sks,
spl_kmem_cache_t *skc;
ASSERT(sks->sks_magic == SKS_MAGIC);
- ASSERT(sks->sks_ref == 0);
+ ASSERT0(sks->sks_ref);
skc = sks->sks_cache;
ASSERT(skc->skc_magic == SKC_MAGIC);
@@ -598,7 +598,7 @@ static void
spl_magazine_free(spl_kmem_magazine_t *skm)
{
ASSERT(skm->skm_magic == SKM_MAGIC);
- ASSERT(skm->skm_avail == 0);
+ ASSERT0(skm->skm_avail);
kfree(skm);
}
@@ -610,7 +610,7 @@ spl_magazine_create(spl_kmem_cache_t *skc)
{
int i = 0;
- ASSERT((skc->skc_flags & KMC_SLAB) == 0);
+ ASSERT0((skc->skc_flags & KMC_SLAB));
skc->skc_mag = kzalloc(sizeof (spl_kmem_magazine_t *) *
num_possible_cpus(), kmem_flags_convert(KM_SLEEP));
@@ -640,7 +640,7 @@ spl_magazine_destroy(spl_kmem_cache_t *skc)
spl_kmem_magazine_t *skm;
int i = 0;
- ASSERT((skc->skc_flags & KMC_SLAB) == 0);
+ ASSERT0((skc->skc_flags & KMC_SLAB));
for_each_possible_cpu(i) {
skm = skc->skc_mag[i];
@@ -679,8 +679,8 @@ spl_kmem_cache_create(const char *name, size_t size, size_t align,
/*
* Unsupported flags
*/
- ASSERT(vmp == NULL);
- ASSERT(reclaim == NULL);
+ ASSERT0P(vmp);
+ ASSERT0P(reclaim);
might_sleep();
@@ -863,11 +863,11 @@ spl_kmem_cache_destroy(spl_kmem_cache_t *skc)
* Validate there are no objects in use and free all the
* spl_kmem_slab_t, spl_kmem_obj_t, and object buffers.
*/
- ASSERT3U(skc->skc_slab_alloc, ==, 0);
- ASSERT3U(skc->skc_obj_alloc, ==, 0);
- ASSERT3U(skc->skc_slab_total, ==, 0);
- ASSERT3U(skc->skc_obj_total, ==, 0);
- ASSERT3U(skc->skc_obj_emergency, ==, 0);
+ ASSERT0(skc->skc_slab_alloc);
+ ASSERT0(skc->skc_obj_alloc);
+ ASSERT0(skc->skc_slab_total);
+ ASSERT0(skc->skc_obj_total);
+ ASSERT0(skc->skc_obj_emergency);
ASSERT(list_empty(&skc->skc_complete_list));
ASSERT3U(percpu_counter_sum(&skc->skc_linux_alloc), ==, 0);
@@ -986,7 +986,7 @@ spl_cache_grow(spl_kmem_cache_t *skc, int flags, void **obj)
ASSERT0(flags & ~KM_PUBLIC_MASK);
ASSERT(skc->skc_magic == SKC_MAGIC);
- ASSERT((skc->skc_flags & KMC_SLAB) == 0);
+ ASSERT0((skc->skc_flags & KMC_SLAB));
*obj = NULL;
diff --git a/sys/contrib/openzfs/module/os/linux/spl/spl-kstat.c b/sys/contrib/openzfs/module/os/linux/spl/spl-kstat.c
index 48f70b00c96b..02c5b42bc4a0 100644
--- a/sys/contrib/openzfs/module/os/linux/spl/spl-kstat.c
+++ b/sys/contrib/openzfs/module/os/linux/spl/spl-kstat.c
@@ -541,7 +541,7 @@ __kstat_create(const char *ks_module, int ks_instance, const char *ks_name,
kstat_t *ksp;
ASSERT(ks_module);
- ASSERT(ks_instance == 0);
+ ASSERT0(ks_instance);
ASSERT(ks_name);
if ((ks_type == KSTAT_TYPE_INTR) || (ks_type == KSTAT_TYPE_IO))
diff --git a/sys/contrib/openzfs/module/os/linux/spl/spl-thread.c b/sys/contrib/openzfs/module/os/linux/spl/spl-thread.c
index f42f455222de..8f5c73b13df5 100644
--- a/sys/contrib/openzfs/module/os/linux/spl/spl-thread.c
+++ b/sys/contrib/openzfs/module/os/linux/spl/spl-thread.c
@@ -80,7 +80,7 @@ __thread_create(caddr_t stk, size_t stksize, thread_func_t func,
/* Option pp is simply ignored */
/* Variable stack size unsupported */
- ASSERT(stk == NULL);
+ ASSERT0P(stk);
tp = kmem_alloc(sizeof (thread_priv_t), KM_PUSHPAGE);
if (tp == NULL)
diff --git a/sys/contrib/openzfs/module/os/linux/spl/spl-tsd.c b/sys/contrib/openzfs/module/os/linux/spl/spl-tsd.c
index 34a61bef7d4f..2e8cedf0dc87 100644
--- a/sys/contrib/openzfs/module/os/linux/spl/spl-tsd.c
+++ b/sys/contrib/openzfs/module/os/linux/spl/spl-tsd.c
@@ -161,7 +161,7 @@ tsd_hash_add(tsd_hash_table_t *table, uint_t key, pid_t pid, void *value)
ulong_t hash;
int rc = 0;
- ASSERT3P(tsd_hash_search(table, key, pid), ==, NULL);
+ ASSERT0P(tsd_hash_search(table, key, pid));
/* New entry allocate structure, set value, and add to hash */
entry = kmem_alloc(sizeof (tsd_hash_entry_t), KM_PUSHPAGE);
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/abd_os.c b/sys/contrib/openzfs/module/os/linux/zfs/abd_os.c
index 248c9b7a6d3b..8a8316f63c48 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/abd_os.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/abd_os.c
@@ -863,9 +863,9 @@ abd_iter_advance(struct abd_iter *aiter, size_t amount)
* Ensure that last chunk is not in use. abd_iterate_*() must clear
* this state (directly or abd_iter_unmap()) before advancing.
*/
- ASSERT3P(aiter->iter_mapaddr, ==, NULL);
+ ASSERT0P(aiter->iter_mapaddr);
ASSERT0(aiter->iter_mapsize);
- ASSERT3P(aiter->iter_page, ==, NULL);
+ ASSERT0P(aiter->iter_page);
ASSERT0(aiter->iter_page_doff);
ASSERT0(aiter->iter_page_dsize);
@@ -897,7 +897,7 @@ abd_iter_map(struct abd_iter *aiter)
void *paddr;
size_t offset = 0;
- ASSERT3P(aiter->iter_mapaddr, ==, NULL);
+ ASSERT0P(aiter->iter_mapaddr);
ASSERT0(aiter->iter_mapsize);
/* There's nothing left to iterate over, so do nothing */
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/vdev_disk.c b/sys/contrib/openzfs/module/os/linux/zfs/vdev_disk.c
index 154ca22d9513..830fad7fe793 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/vdev_disk.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/vdev_disk.c
@@ -552,7 +552,7 @@ vdev_bio_associate_blkg(struct bio *bio)
#endif
ASSERT3P(q, !=, NULL);
- ASSERT3P(bio->bi_blkg, ==, NULL);
+ ASSERT0P(bio->bi_blkg);
if (q->root_blkg && vdev_blkg_tryget(q->root_blkg))
bio->bi_blkg = q->root_blkg;
@@ -574,7 +574,7 @@ vdev_bio_set_dev(struct bio *bio, struct block_device *bdev)
bio->bi_bdev = bdev;
ASSERT3P(q, !=, NULL);
- ASSERT3P(bio->bi_blkg, ==, NULL);
+ ASSERT0P(bio->bi_blkg);
if (q->root_blkg && vdev_blkg_tryget(q->root_blkg))
bio->bi_blkg = q->root_blkg;
@@ -806,7 +806,7 @@ vbio_completion(struct bio *bio)
* here; instead we stash vbio on the zio and take care of it in the
* done callback.
*/
- ASSERT3P(zio->io_bio, ==, NULL);
+ ASSERT0P(zio->io_bio);
zio->io_bio = vbio;
zio_delay_interrupt(zio);
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_acl.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_acl.c
index 1b169122f25b..daa4b5776837 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_acl.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_acl.c
@@ -1900,7 +1900,7 @@ zfs_acl_ids_create(znode_t *dzp, int flag, vattr_t *vap, cred_t *cr,
if (!(flag & IS_ROOT_NODE) &&
(dzp->z_pflags & ZFS_INHERIT_ACE) &&
!(dzp->z_pflags & ZFS_XATTR)) {
- VERIFY(0 == zfs_acl_node_read(dzp, B_TRUE,
+ VERIFY0(zfs_acl_node_read(dzp, B_TRUE,
&paclp, B_FALSE));
acl_ids->z_aclp = zfs_acl_inherit(zfsvfs,
vap->va_mode, paclp, acl_ids->z_mode, &need_chmod);
@@ -2204,8 +2204,8 @@ top:
}
error = zfs_aclset_common(zp, aclp, cr, tx);
- ASSERT(error == 0);
- ASSERT(zp->z_acl_cached == NULL);
+ ASSERT0(error);
+ ASSERT0P(zp->z_acl_cached);
zp->z_acl_cached = aclp;
if (fuid_dirtied)
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_ctldir.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_ctldir.c
index 6552a933ce0a..fb4de50480a3 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_ctldir.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_ctldir.c
@@ -494,9 +494,9 @@ zfsctl_inode_alloc(zfsvfs_t *zfsvfs, uint64_t id,
if (!creation)
now = current_time(ip);
zp = ITOZ(ip);
- ASSERT3P(zp->z_dirlocks, ==, NULL);
- ASSERT3P(zp->z_acl_cached, ==, NULL);
- ASSERT3P(zp->z_xattr_cached, ==, NULL);
+ ASSERT0P(zp->z_dirlocks);
+ ASSERT0P(zp->z_acl_cached);
+ ASSERT0P(zp->z_xattr_cached);
zp->z_id = id;
zp->z_unlinked = B_FALSE;
zp->z_atime_dirty = B_FALSE;
@@ -590,7 +590,7 @@ zfsctl_inode_lookup(zfsvfs_t *zfsvfs, uint64_t id,
int
zfsctl_create(zfsvfs_t *zfsvfs)
{
- ASSERT(zfsvfs->z_ctldir == NULL);
+ ASSERT0P(zfsvfs->z_ctldir);
zfsvfs->z_ctldir = zfsctl_inode_alloc(zfsvfs, ZFSCTL_INO_ROOT,
&zpl_fops_root, &zpl_ops_root, 0);
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_dir.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_dir.c
index 2f935bb3fc8c..e8de536606e2 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_dir.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_dir.c
@@ -463,7 +463,7 @@ zfs_unlinked_add(znode_t *zp, dmu_tx_t *tx)
zfsvfs_t *zfsvfs = ZTOZSB(zp);
ASSERT(zp->z_unlinked);
- ASSERT(ZTOI(zp)->i_nlink == 0);
+ ASSERT0(ZTOI(zp)->i_nlink);
VERIFY3U(0, ==,
zap_add_int(zfsvfs->z_os, zfsvfs->z_unlinkedobj, zp->z_id, tx));
@@ -662,8 +662,8 @@ zfs_rmnode(znode_t *zp)
uint64_t links;
int error;
- ASSERT(ZTOI(zp)->i_nlink == 0);
- ASSERT(atomic_read(&ZTOI(zp)->i_count) == 0);
+ ASSERT0(ZTOI(zp)->i_nlink);
+ ASSERT0(atomic_read(&ZTOI(zp)->i_count));
/*
* If this is an attribute directory, purge its contents.
@@ -710,7 +710,7 @@ zfs_rmnode(znode_t *zp)
&xattr_obj, sizeof (xattr_obj));
if (error == 0 && xattr_obj) {
error = zfs_zget(zfsvfs, xattr_obj, &xzp);
- ASSERT(error == 0);
+ ASSERT0(error);
}
acl_obj = zfs_external_acl(zp);
@@ -744,12 +744,12 @@ zfs_rmnode(znode_t *zp)
}
if (xzp) {
- ASSERT(error == 0);
+ ASSERT0(error);
mutex_enter(&xzp->z_lock);
xzp->z_unlinked = B_TRUE; /* mark xzp for deletion */
clear_nlink(ZTOI(xzp)); /* no more links to it */
links = 0;
- VERIFY(0 == sa_update(xzp->z_sa_hdl, SA_ZPL_LINKS(zfsvfs),
+ VERIFY0(sa_update(xzp->z_sa_hdl, SA_ZPL_LINKS(zfsvfs),
&links, sizeof (links), tx));
mutex_exit(&xzp->z_lock);
zfs_unlinked_add(xzp, tx);
@@ -872,7 +872,7 @@ zfs_link_create(zfs_dirlock_t *dl, znode_t *zp, dmu_tx_t *tx, int flag)
ctime);
}
error = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
- ASSERT(error == 0);
+ ASSERT0(error);
mutex_exit(&zp->z_lock);
@@ -894,7 +894,7 @@ zfs_link_create(zfs_dirlock_t *dl, znode_t *zp, dmu_tx_t *tx, int flag)
&dzp->z_pflags, sizeof (dzp->z_pflags));
zfs_tstamp_update_setup(dzp, CONTENT_MODIFIED, mtime, ctime);
error = sa_bulk_update(dzp->z_sa_hdl, bulk, count, tx);
- ASSERT(error == 0);
+ ASSERT0(error);
mutex_exit(&dzp->z_lock);
return (0);
@@ -986,7 +986,7 @@ zfs_drop_nlink_locked(znode_t *zp, dmu_tx_t *tx, boolean_t *unlinkedp)
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs),
NULL, &links, sizeof (links));
error = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
- ASSERT3U(error, ==, 0);
+ ASSERT0(error);
if (unlinkedp != NULL)
*unlinkedp = unlinked;
@@ -1058,7 +1058,7 @@ zfs_link_destroy(zfs_dirlock_t *dl, znode_t *zp, dmu_tx_t *tx, int flag,
/* The only error is !zfs_dirempty() and we checked earlier. */
error = zfs_drop_nlink_locked(zp, tx, &unlinked);
- ASSERT3U(error, ==, 0);
+ ASSERT0(error);
mutex_exit(&zp->z_lock);
} else {
error = zfs_dropname(dl, zp, dzp, tx, flag);
@@ -1083,7 +1083,7 @@ zfs_link_destroy(zfs_dirlock_t *dl, znode_t *zp, dmu_tx_t *tx, int flag,
NULL, &dzp->z_pflags, sizeof (dzp->z_pflags));
zfs_tstamp_update_setup(dzp, CONTENT_MODIFIED, mtime, ctime);
error = sa_bulk_update(dzp->z_sa_hdl, bulk, count, tx);
- ASSERT(error == 0);
+ ASSERT0(error);
mutex_exit(&dzp->z_lock);
if (unlinkedp != NULL)
@@ -1167,7 +1167,7 @@ zfs_make_xattrdir(znode_t *zp, vattr_t *vap, znode_t **xzpp, cred_t *cr)
ASSERT(error == 0 && parent == zp->z_id);
#endif
- VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs), &xzp->z_id,
+ VERIFY0(sa_update(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs), &xzp->z_id,
sizeof (xzp->z_id), tx));
if (!zp->z_unlinked)
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_sysfs.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_sysfs.c
index 1c187d7b9cab..895d80b2d79e 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_sysfs.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_sysfs.c
@@ -223,7 +223,7 @@ zfs_kobj_add(zfs_mod_kobj_t *zkobj, struct kobject *parent, const char *name)
{
/* zko_default_group.attrs must be NULL terminated */
ASSERT(zkobj->zko_default_group.attrs != NULL);
- ASSERT(zkobj->zko_default_group.attrs[zkobj->zko_attr_count] == NULL);
+ ASSERT0P(zkobj->zko_default_group.attrs[zkobj->zko_attr_count]);
kobject_init(&zkobj->zko_kobj, &zkobj->zko_kobj_type);
return (kobject_add(&zkobj->zko_kobj, parent, name));
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_vfsops.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_vfsops.c
index 396faef8f646..cd606e667bff 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_vfsops.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_vfsops.c
@@ -279,19 +279,14 @@ zfs_sync(struct super_block *sb, int wait, cred_t *cr)
return (err);
/*
- * If the pool is suspended, just return an error. This is to help
- * with shutting down with pools suspended, as we don't want to block
- * in that case.
+ * Sync any pending writes, but do not block if the pool is suspended.
+ * This is to help with shutting down with pools suspended, as we don't
+ * want to block in that case.
*/
- if (spa_suspended(zfsvfs->z_os->os_spa)) {
- zfs_exit(zfsvfs, FTAG);
- return (SET_ERROR(EIO));
- }
-
- zil_commit(zfsvfs->z_log, 0);
+ err = zil_commit_flags(zfsvfs->z_log, 0, ZIL_COMMIT_NOW);
zfs_exit(zfsvfs, FTAG);
- return (0);
+ return (err);
}
static void
@@ -883,7 +878,7 @@ zfsvfs_setup(zfsvfs_t *zfsvfs, boolean_t mounting)
* operations out since we closed the ZIL.
*/
if (mounting) {
- ASSERT3P(zfsvfs->z_kstat.dk_kstats, ==, NULL);
+ ASSERT0P(zfsvfs->z_kstat.dk_kstats);
error = dataset_kstats_create(&zfsvfs->z_kstat, zfsvfs->z_os);
if (error)
return (error);
@@ -1676,7 +1671,7 @@ zfs_umount(struct super_block *sb)
if (zfsvfs->z_arc_prune != NULL)
arc_remove_prune_callback(zfsvfs->z_arc_prune);
- VERIFY(zfsvfs_teardown(zfsvfs, B_TRUE) == 0);
+ VERIFY0(zfsvfs_teardown(zfsvfs, B_TRUE));
os = zfsvfs->z_os;
/*
@@ -1802,8 +1797,8 @@ zfs_vget(struct super_block *sb, struct inode **ipp, fid_t *fidp)
ASSERT(*ipp != NULL);
if (object == ZFSCTL_INO_SNAPDIR) {
- VERIFY(zfsctl_root_lookup(*ipp, "snapshot", ipp,
- 0, kcred, NULL, NULL) == 0);
+ VERIFY0(zfsctl_root_lookup(*ipp, "snapshot", ipp,
+ 0, kcred, NULL, NULL));
} else {
/*
* Must have an existing ref, so igrab()
@@ -1905,7 +1900,7 @@ zfs_resume_fs(zfsvfs_t *zfsvfs, dsl_dataset_t *ds)
goto bail;
ds->ds_dir->dd_activity_cancelled = B_FALSE;
- VERIFY(zfsvfs_setup(zfsvfs, B_FALSE) == 0);
+ VERIFY0(zfsvfs_setup(zfsvfs, B_FALSE));
zfs_set_fuid_feature(zfsvfs);
zfsvfs->z_rollback_time = jiffies;
@@ -2078,7 +2073,7 @@ zfs_set_version(zfsvfs_t *zfsvfs, uint64_t newvers)
ZFS_SA_ATTRS, 8, 1, &sa_obj, tx);
ASSERT0(error);
- VERIFY(0 == sa_set_sa_object(os, sa_obj));
+ VERIFY0(sa_set_sa_object(os, sa_obj));
sa_register_update_callback(os, zfs_sa_upgrade);
}
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_vnops_os.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_vnops_os.c
index 6a2fc5ad7935..6106726651a3 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_vnops_os.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_vnops_os.c
@@ -841,8 +841,8 @@ out:
*zpp = zp;
}
- if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ if (error == 0 && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
+ error = zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
@@ -1203,8 +1203,8 @@ out:
zfs_zrele_async(xzp);
}
- if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ if (error == 0 && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
+ error = zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
@@ -1392,14 +1392,15 @@ out:
zfs_dirent_unlock(dl);
- if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
-
if (error != 0) {
zrele(zp);
} else {
zfs_znode_update_vfs(dzp);
zfs_znode_update_vfs(zp);
+
+ if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
+ error = zil_commit(zilog, 0);
+
}
zfs_exit(zfsvfs, FTAG);
return (error);
@@ -1528,8 +1529,8 @@ out:
zfs_znode_update_vfs(zp);
zrele(zp);
- if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ if (error == 0 && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
+ error = zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
@@ -2483,10 +2484,10 @@ top:
new_mode = zp->z_mode;
}
err = zfs_acl_chown_setattr(zp);
- ASSERT(err == 0);
+ ASSERT0(err);
if (attrzp) {
err = zfs_acl_chown_setattr(attrzp);
- ASSERT(err == 0);
+ ASSERT0(err);
}
}
@@ -2600,7 +2601,7 @@ out:
if (err == 0 && xattr_count > 0) {
err2 = sa_bulk_update(attrzp->z_sa_hdl, xattr_bulk,
xattr_count, tx);
- ASSERT(err2 == 0);
+ ASSERT0(err2);
}
if (aclp)
@@ -2630,8 +2631,8 @@ out:
}
out2:
- if (os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ if (err == 0 && os->os_sync == ZFS_SYNC_ALWAYS)
+ err = zil_commit(zilog, 0);
out3:
kmem_free(xattr_bulk, sizeof (sa_bulk_attr_t) * bulks);
@@ -3157,7 +3158,7 @@ top:
* zfs_link_create() to add back the same entry, but with a new
* dnode (szp), should not fail.
*/
- ASSERT3P(tzp, ==, NULL);
+ ASSERT0P(tzp);
goto commit_link_tzp;
}
@@ -3235,8 +3236,8 @@ out:
zfs_dirent_unlock(sdl);
zfs_dirent_unlock(tdl);
- if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ if (error == 0 && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
+ error = zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
@@ -3436,7 +3437,7 @@ top:
*zpp = zp;
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ error = zil_commit(zilog, 0);
} else {
zrele(zp);
}
@@ -3654,8 +3655,8 @@ top:
* operation are sync safe.
*/
if (is_tmpfile) {
- VERIFY(zap_remove_int(zfsvfs->z_os,
- zfsvfs->z_unlinkedobj, szp->z_id, tx) == 0);
+ VERIFY0(zap_remove_int(zfsvfs->z_os,
+ zfsvfs->z_unlinkedobj, szp->z_id, tx));
} else {
if (flags & FIGNORECASE)
txtype |= TX_CI;
@@ -3670,18 +3671,20 @@ top:
zfs_dirent_unlock(dl);
- if (!is_tmpfile && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
-
- if (is_tmpfile && zfsvfs->z_os->os_sync != ZFS_SYNC_DISABLED) {
- txg_wait_flag_t wait_flags =
- spa_get_failmode(dmu_objset_spa(zfsvfs->z_os)) ==
- ZIO_FAILURE_MODE_CONTINUE ? TXG_WAIT_SUSPEND : 0;
- error = txg_wait_synced_flags(dmu_objset_pool(zfsvfs->z_os),
- txg, wait_flags);
- if (error != 0) {
- ASSERT3U(error, ==, ESHUTDOWN);
- error = SET_ERROR(EIO);
+ if (error == 0) {
+ if (!is_tmpfile && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
+ error = zil_commit(zilog, 0);
+
+ if (is_tmpfile && zfsvfs->z_os->os_sync != ZFS_SYNC_DISABLED) {
+ txg_wait_flag_t wait_flags =
+ spa_get_failmode(dmu_objset_spa(zfsvfs->z_os)) ==
+ ZIO_FAILURE_MODE_CONTINUE ? TXG_WAIT_SUSPEND : 0;
+ error = txg_wait_synced_flags(
+ dmu_objset_pool(zfsvfs->z_os), txg, wait_flags);
+ if (error != 0) {
+ ASSERT3U(error, ==, ESHUTDOWN);
+ error = SET_ERROR(EIO);
+ }
}
}
@@ -3691,16 +3694,42 @@ top:
return (error);
}
-static void
-zfs_putpage_commit_cb(void *arg)
+/* Finish page writeback. */
+static inline void
+zfs_page_writeback_done(struct page *pp, int err)
{
- struct page *pp = arg;
+ if (err != 0) {
+ /*
+ * Writeback failed. Re-dirty the page. It was undirtied before
+ * the IO was issued (in zfs_putpage() or write_cache_pages()).
+ * The kernel only considers writeback for dirty pages; if we
+ * don't do this, it is eligible for eviction without being
+ * written out, which we definitely don't want.
+ */
+#ifdef HAVE_VFS_FILEMAP_DIRTY_FOLIO
+ filemap_dirty_folio(page_mapping(pp), page_folio(pp));
+#else
+ __set_page_dirty_nobuffers(pp);
+#endif
+ }
ClearPageError(pp);
end_page_writeback(pp);
}
/*
+ * ZIL callback for page writeback. Passes to zfs_log_write() in zfs_putpage()
+ * for syncing writes. Called when the ZIL itx has been written to the log or
+ * the whole txg syncs, or if the ZIL crashes or the pool suspends. Any failure
+ * is passed as `err`.
+ */
+static void
+zfs_putpage_commit_cb(void *arg, int err)
+{
+ zfs_page_writeback_done(arg, err);
+}
+
+/*
* Push a page out to disk, once the page is on stable storage the
* registered commit callback will be run as notification of completion.
*
@@ -3853,16 +3882,15 @@ zfs_putpage(struct inode *ip, struct page *pp, struct writeback_control *wbc,
err = dmu_tx_assign(tx, DMU_TX_WAIT);
if (err != 0) {
dmu_tx_abort(tx);
-#ifdef HAVE_VFS_FILEMAP_DIRTY_FOLIO
- filemap_dirty_folio(page_mapping(pp), page_folio(pp));
-#else
- __set_page_dirty_nobuffers(pp);
-#endif
- ClearPageError(pp);
- end_page_writeback(pp);
+ zfs_page_writeback_done(pp, err);
zfs_rangelock_exit(lr);
zfs_exit(zfsvfs, FTAG);
- return (err);
+
+ /*
+ * Don't return error for an async writeback; we've re-dirtied
+ * the page so it will be tried again some other time.
+ */
+ return (for_sync ? err : 0);
}
va = kmap(pp);
@@ -3916,7 +3944,7 @@ zfs_putpage(struct inode *ip, struct page *pp, struct writeback_control *wbc,
* ALL, zfs_putpage should do it.
*
* Summary:
- * for_sync: 0=unlock immediately; 1 unlock once on disk
+ * for_sync: 0=unlock immediately; 1=unlock once on disk
* sync_mode: NONE=caller will commit; ALL=we will commit
*/
boolean_t need_commit = (wbc->sync_mode != WB_SYNC_NONE);
@@ -3931,16 +3959,24 @@ zfs_putpage(struct inode *ip, struct page *pp, struct writeback_control *wbc,
B_FALSE, for_sync ? zfs_putpage_commit_cb : NULL, pp);
if (!for_sync) {
- ClearPageError(pp);
- end_page_writeback(pp);
+ /*
+ * Async writeback is logged and written to the DMU, so page
+ * can now be unlocked.
+ */
+ zfs_page_writeback_done(pp, 0);
}
dmu_tx_commit(tx);
zfs_rangelock_exit(lr);
- if (need_commit)
- zil_commit(zfsvfs->z_log, zp->z_id);
+ if (need_commit) {
+ err = zil_commit_flags(zfsvfs->z_log, zp->z_id, ZIL_COMMIT_NOW);
+ if (err != 0) {
+ zfs_exit(zfsvfs, FTAG);
+ return (err);
+ }
+ }
dataset_kstats_update_write_kstats(&zfsvfs->z_kstat, pglen);
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_znode_os.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_znode_os.c
index 7683eeb3cf9f..bcaabeb32b8a 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_znode_os.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_znode_os.c
@@ -144,9 +144,9 @@ zfs_znode_cache_destructor(void *buf, void *arg)
rw_destroy(&zp->z_xattr_lock);
zfs_rangelock_fini(&zp->z_rangelock);
- ASSERT3P(zp->z_dirlocks, ==, NULL);
- ASSERT3P(zp->z_acl_cached, ==, NULL);
- ASSERT3P(zp->z_xattr_cached, ==, NULL);
+ ASSERT0P(zp->z_dirlocks);
+ ASSERT0P(zp->z_acl_cached);
+ ASSERT0P(zp->z_xattr_cached);
}
static int
@@ -178,13 +178,13 @@ zfs_znode_init(void)
* backed by kmalloc() when on the Linux slab in order that any
* wait_on_bit() operations on the related inode operate properly.
*/
- ASSERT(znode_cache == NULL);
+ ASSERT0P(znode_cache);
znode_cache = kmem_cache_create("zfs_znode_cache",
sizeof (znode_t), 0, zfs_znode_cache_constructor,
zfs_znode_cache_destructor, NULL, NULL, NULL,
KMC_SLAB | KMC_RECLAIMABLE);
- ASSERT(znode_hold_cache == NULL);
+ ASSERT0P(znode_hold_cache);
znode_hold_cache = kmem_cache_create("zfs_znode_hold_cache",
sizeof (znode_hold_t), 0, zfs_znode_hold_cache_constructor,
zfs_znode_hold_cache_destructor, NULL, NULL, NULL, 0);
@@ -327,10 +327,10 @@ zfs_znode_sa_init(zfsvfs_t *zfsvfs, znode_t *zp,
mutex_enter(&zp->z_lock);
- ASSERT(zp->z_sa_hdl == NULL);
- ASSERT(zp->z_acl_cached == NULL);
+ ASSERT0P(zp->z_sa_hdl);
+ ASSERT0P(zp->z_acl_cached);
if (sa_hdl == NULL) {
- VERIFY(0 == sa_handle_get_from_db(zfsvfs->z_os, db, zp,
+ VERIFY0(sa_handle_get_from_db(zfsvfs->z_os, db, zp,
SA_HDL_SHARED, &zp->z_sa_hdl));
} else {
zp->z_sa_hdl = sa_hdl;
@@ -530,9 +530,9 @@ zfs_znode_alloc(zfsvfs_t *zfsvfs, dmu_buf_t *db, int blksz,
return (NULL);
zp = ITOZ(ip);
- ASSERT(zp->z_dirlocks == NULL);
- ASSERT3P(zp->z_acl_cached, ==, NULL);
- ASSERT3P(zp->z_xattr_cached, ==, NULL);
+ ASSERT0P(zp->z_dirlocks);
+ ASSERT0P(zp->z_acl_cached);
+ ASSERT0P(zp->z_xattr_cached);
zp->z_unlinked = B_FALSE;
zp->z_atime_dirty = B_FALSE;
zp->z_is_ctldir = B_FALSE;
@@ -611,7 +611,7 @@ zfs_znode_alloc(zfsvfs_t *zfsvfs, dmu_buf_t *db, int blksz,
* processing so do not hash unlinked znodes.
*/
if (links > 0)
- VERIFY3S(insert_inode_locked(ip), ==, 0);
+ VERIFY0(insert_inode_locked(ip));
mutex_enter(&zfsvfs->z_znodes_lock);
list_insert_tail(&zfsvfs->z_all_znodes, zp);
@@ -811,7 +811,7 @@ zfs_mknode(znode_t *dzp, vattr_t *vap, dmu_tx_t *tx, cred_t *cr,
}
/* Now add in all of the "SA" attributes */
- VERIFY(0 == sa_handle_get_from_db(zfsvfs->z_os, db, NULL, SA_HDL_SHARED,
+ VERIFY0(sa_handle_get_from_db(zfsvfs->z_os, db, NULL, SA_HDL_SHARED,
&sa_hdl));
/*
@@ -901,7 +901,7 @@ zfs_mknode(znode_t *dzp, vattr_t *vap, dmu_tx_t *tx, cred_t *cr,
acl_ids->z_fuid, acl_ids->z_fgid);
}
- VERIFY(sa_replace_all_by_template(sa_hdl, sa_attrs, cnt, tx) == 0);
+ VERIFY0(sa_replace_all_by_template(sa_hdl, sa_attrs, cnt, tx));
if (!(flag & IS_ROOT_NODE)) {
/*
@@ -1200,7 +1200,7 @@ zfs_rezget(znode_t *zp)
}
rw_exit(&zp->z_xattr_lock);
- ASSERT(zp->z_sa_hdl == NULL);
+ ASSERT0P(zp->z_sa_hdl);
err = sa_buf_hold(zfsvfs->z_os, obj_num, NULL, &db);
if (err) {
zfs_znode_hold_exit(zfsvfs, zh);
@@ -1314,9 +1314,9 @@ zfs_znode_delete(znode_t *zp, dmu_tx_t *tx)
zh = zfs_znode_hold_enter(zfsvfs, obj);
if (acl_obj) {
VERIFY(!zp->z_is_sa);
- VERIFY(0 == dmu_object_free(os, acl_obj, tx));
+ VERIFY0(dmu_object_free(os, acl_obj, tx));
}
- VERIFY(0 == dmu_object_free(os, obj, tx));
+ VERIFY0(dmu_object_free(os, obj, tx));
zfs_znode_dmu_fini(zp);
zfs_znode_hold_exit(zfsvfs, zh);
}
@@ -1536,7 +1536,7 @@ zfs_extend(znode_t *zp, uint64_t end)
zp->z_size = end;
- VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_SIZE(ZTOZSB(zp)),
+ VERIFY0(sa_update(zp->z_sa_hdl, SA_ZPL_SIZE(ZTOZSB(zp)),
&zp->z_size, sizeof (zp->z_size), tx));
zfs_rangelock_exit(lr);
@@ -1726,7 +1726,7 @@ zfs_trunc(znode_t *zp, uint64_t end)
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs),
NULL, &zp->z_pflags, 8);
}
- VERIFY(sa_bulk_update(zp->z_sa_hdl, bulk, count, tx) == 0);
+ VERIFY0(sa_bulk_update(zp->z_sa_hdl, bulk, count, tx));
dmu_tx_commit(tx);
zfs_rangelock_exit(lr);
@@ -1793,7 +1793,7 @@ log:
NULL, &zp->z_pflags, 8);
zfs_tstamp_update_setup(zp, CONTENT_MODIFIED, mtime, ctime);
error = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
- ASSERT(error == 0);
+ ASSERT0(error);
zfs_log_truncate(zilog, tx, TX_TRUNCATE, zp, off, len);
@@ -1840,7 +1840,7 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx)
moid = MASTER_NODE_OBJ;
error = zap_create_claim(os, moid, DMU_OT_MASTER_NODE,
DMU_OT_NONE, 0, tx);
- ASSERT(error == 0);
+ ASSERT0(error);
/*
* Set starting attributes.
@@ -1853,7 +1853,7 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx)
const char *name;
ASSERT(nvpair_type(elem) == DATA_TYPE_UINT64);
- VERIFY(nvpair_value_uint64(elem, &val) == 0);
+ VERIFY0(nvpair_value_uint64(elem, &val));
name = nvpair_name(elem);
if (strcmp(name, zfs_prop_to_name(ZFS_PROP_VERSION)) == 0) {
if (val < version)
@@ -1861,7 +1861,7 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx)
} else {
error = zap_update(os, moid, name, 8, 1, &val, tx);
}
- ASSERT(error == 0);
+ ASSERT0(error);
if (strcmp(name, zfs_prop_to_name(ZFS_PROP_NORMALIZE)) == 0)
norm = val;
else if (strcmp(name, zfs_prop_to_name(ZFS_PROP_CASE)) == 0)
@@ -1869,7 +1869,7 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx)
}
ASSERT(version != 0);
error = zap_update(os, moid, ZPL_VERSION_STR, 8, 1, &version, tx);
- ASSERT(error == 0);
+ ASSERT0(error);
/*
* Create zap object used for SA attribute registration
@@ -1879,7 +1879,7 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx)
sa_obj = zap_create(os, DMU_OT_SA_MASTER_NODE,
DMU_OT_NONE, 0, tx);
error = zap_add(os, moid, ZFS_SA_ATTRS, 8, 1, &sa_obj, tx);
- ASSERT(error == 0);
+ ASSERT0(error);
} else {
sa_obj = 0;
}
@@ -1889,7 +1889,7 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx)
obj = zap_create(os, DMU_OT_UNLINKED_SET, DMU_OT_NONE, 0, tx);
error = zap_add(os, moid, ZFS_UNLINKED_SET, 8, 1, &obj, tx);
- ASSERT(error == 0);
+ ASSERT0(error);
/*
* Create root znode. Create minimal znode/inode/zfsvfs/sb
@@ -1922,7 +1922,7 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx)
error = sa_setup(os, sa_obj, zfs_attr_table, ZPL_END,
&zfsvfs->z_attr_table);
- ASSERT(error == 0);
+ ASSERT0(error);
/*
* Fold case on file systems that are always or sometimes case
@@ -1946,12 +1946,12 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx)
mutex_init(&zfsvfs->z_hold_locks[i], NULL, MUTEX_DEFAULT, NULL);
}
- VERIFY(0 == zfs_acl_ids_create(rootzp, IS_ROOT_NODE, &vattr,
+ VERIFY0(zfs_acl_ids_create(rootzp, IS_ROOT_NODE, &vattr,
cr, NULL, &acl_ids, zfs_init_idmap));
zfs_mknode(rootzp, &vattr, tx, cr, IS_ROOT_NODE, &zp, &acl_ids);
ASSERT3P(zp, ==, rootzp);
error = zap_add(os, moid, ZFS_ROOT_OBJ, 8, 1, &rootzp->z_id, tx);
- ASSERT(error == 0);
+ ASSERT0(error);
zfs_acl_ids_free(&acl_ids);
atomic_set(&ZTOI(rootzp)->i_count, 0);
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zpl_file.c b/sys/contrib/openzfs/module/os/linux/zfs/zpl_file.c
index ef7bd7352084..d07317b0d910 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zpl_file.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zpl_file.c
@@ -22,6 +22,7 @@
/*
* Copyright (c) 2011, Lawrence Livermore National Security, LLC.
* Copyright (c) 2015 by Chunwei Chen. All rights reserved.
+ * Copyright (c) 2025, Klara, Inc.
*/
@@ -106,6 +107,10 @@ zpl_iterate(struct file *filp, struct dir_context *ctx)
return (error);
}
+static inline int
+zpl_write_cache_pages(struct address_space *mapping,
+ struct writeback_control *wbc, void *data);
+
static int
zpl_fsync(struct file *filp, loff_t start, loff_t end, int datasync)
{
@@ -115,9 +120,38 @@ zpl_fsync(struct file *filp, loff_t start, loff_t end, int datasync)
int error;
fstrans_cookie_t cookie;
- error = filemap_write_and_wait_range(inode->i_mapping, start, end);
- if (error)
- return (error);
+ /*
+ * Force dirty pages in the range out to the DMU and the log, ready
+ * for zil_commit() to write down.
+ *
+ * We call write_cache_pages() directly to ensure that zpl_putpage() is
+ * called with the flags we need. We need WB_SYNC_NONE to avoid a call
+ * to zil_commit() (since we're doing this as a kind of pre-sync); but
+ * we do need for_sync so that the pages remain in writeback until
+ * they're on disk, and so that we get an error if the DMU write fails.
+ */
+ if (filemap_range_has_page(inode->i_mapping, start, end)) {
+ int for_sync = 1;
+ struct writeback_control wbc = {
+ .sync_mode = WB_SYNC_NONE,
+ .nr_to_write = LONG_MAX,
+ .range_start = start,
+ .range_end = end,
+ };
+ error =
+ zpl_write_cache_pages(inode->i_mapping, &wbc, &for_sync);
+ if (error != 0) {
+ /*
+ * Unclear what state things are in. zfs_putpage() will
+ * ensure the pages remain dirty if they haven't been
+ * written down to the DMU, but because there may be
+ * nothing logged, we can't assume that zfs_sync() ->
+ * zil_commit() will give us a useful error. It's
+ * safest if we just error out here.
+ */
+ return (error);
+ }
+ }
crhold(cr);
cookie = spl_fstrans_mark();
@@ -494,11 +528,30 @@ zpl_writepages(struct address_space *mapping, struct writeback_control *wbc)
if (sync_mode != wbc->sync_mode) {
if ((result = zpl_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (result);
- if (zfsvfs->z_log != NULL)
- zil_commit(zfsvfs->z_log, zp->z_id);
+
+ if (zfsvfs->z_log != NULL) {
+ /*
+ * We don't want to block here if the pool suspends,
+ * because this is not a syncing op by itself, but
+ * might be part of one that the caller will
+ * coordinate.
+ */
+ result = -zil_commit_flags(zfsvfs->z_log, zp->z_id,
+ ZIL_COMMIT_NOW);
+ }
+
zpl_exit(zfsvfs, FTAG);
/*
+ * If zil_commit_flags() failed, it's unclear what state things
+ * are currently in. putpage() has written back out what it can
+ * to the DMU, but it may not be on disk. We have little choice
+ * but to escape.
+ */
+ if (result != 0)
+ return (result);
+
+ /*
* We need to call write_cache_pages() again (we can't just
* return after the commit) because the previous call in
* non-SYNC mode does not guarantee that we got all the dirty
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zpl_inode.c b/sys/contrib/openzfs/module/os/linux/zfs/zpl_inode.c
index f9f6406f8b47..f97662d052c7 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zpl_inode.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zpl_inode.c
@@ -247,7 +247,7 @@ zpl_mknod(struct inode *dir, struct dentry *dentry, umode_t mode,
* and fifos, but we want to know if this behavior ever changes.
*/
if (S_ISSOCK(mode) || S_ISFIFO(mode))
- ASSERT(rdev == 0);
+ ASSERT0(rdev);
crhold(cr);
vap = kmem_zalloc(sizeof (vattr_t), KM_SLEEP);
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zpl_super.c b/sys/contrib/openzfs/module/os/linux/zfs/zpl_super.c
index 94dcdd0b887d..53819628627d 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zpl_super.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zpl_super.c
@@ -49,7 +49,7 @@ zpl_inode_alloc(struct super_block *sb)
static void
zpl_inode_free(struct inode *ip)
{
- ASSERT(atomic_read(&ip->i_count) == 0);
+ ASSERT0(atomic_read(&ip->i_count));
zfs_inode_free(ip);
}
#endif
@@ -57,7 +57,7 @@ zpl_inode_free(struct inode *ip)
static void
zpl_inode_destroy(struct inode *ip)
{
- ASSERT(atomic_read(&ip->i_count) == 0);
+ ASSERT0(atomic_read(&ip->i_count));
zfs_inode_destroy(ip);
}
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zpl_xattr.c b/sys/contrib/openzfs/module/os/linux/zfs/zpl_xattr.c
index a098197e7448..d93282db815a 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zpl_xattr.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zpl_xattr.c
@@ -1494,7 +1494,7 @@ zpl_posix_acl_free(void *arg)
acl_rel_head = NULL;
if (cmpxchg(&acl_rel_tail, &a->next,
&acl_rel_head) == &a->next) {
- ASSERT3P(a->next, ==, NULL);
+ ASSERT0P(a->next);
a->next = freelist;
freelist = a;
break;
@@ -1544,7 +1544,7 @@ zpl_posix_acl_release_impl(struct posix_acl *acl)
a->time = ddi_get_lbolt();
/* atomically points tail to us and get the previous tail */
prev = xchg(&acl_rel_tail, &a->next);
- ASSERT3P(*prev, ==, NULL);
+ ASSERT0P(*prev);
*prev = a;
/* if it was empty before, schedule the free task */
if (prev == &acl_rel_head)
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zvol_os.c b/sys/contrib/openzfs/module/os/linux/zfs/zvol_os.c
index a7431cc4da9d..bac166fcd89e 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zvol_os.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zvol_os.c
@@ -22,7 +22,7 @@
/*
* Copyright (c) 2012, 2020 by Delphix. All rights reserved.
* Copyright (c) 2024, Rob Norris <robn@despairlabs.com>
- * Copyright (c) 2024, Klara, Inc.
+ * Copyright (c) 2024, 2025, Klara, Inc.
*/
#include <sys/dataset_kstats.h>
@@ -84,8 +84,9 @@ static unsigned int zvol_blk_mq_blocks_per_thread = 8;
static inline void
zvol_end_io(struct bio *bio, struct request *rq, int error)
{
+ ASSERT3U(error, >=, 0);
if (bio) {
- bio->bi_status = errno_to_bi_status(-error);
+ bio->bi_status = errno_to_bi_status(error);
bio_endio(bio);
} else {
blk_mq_end_request(rq, errno_to_bi_status(error));
@@ -208,8 +209,14 @@ zvol_write(zv_request_t *zvr)
disk = zv->zv_zso->zvo_disk;
/* bio marked as FLUSH need to flush before write */
- if (io_is_flush(bio, rq))
- zil_commit(zv->zv_zilog, ZVOL_OBJ);
+ if (io_is_flush(bio, rq)) {
+ error = zil_commit(zv->zv_zilog, ZVOL_OBJ);
+ if (error != 0) {
+ rw_exit(&zv->zv_suspend_lock);
+ zvol_end_io(bio, rq, -error);
+ return;
+ }
+ }
/* Some requests are just for flush and nothing else. */
if (io_size(bio, rq) == 0) {
@@ -273,8 +280,8 @@ zvol_write(zv_request_t *zvr)
dataset_kstats_update_write_kstats(&zv->zv_kstat, nwritten);
task_io_account_write(nwritten);
- if (sync)
- zil_commit(zv->zv_zilog, ZVOL_OBJ);
+ if (error == 0 && sync)
+ error = zil_commit(zv->zv_zilog, ZVOL_OBJ);
rw_exit(&zv->zv_suspend_lock);
@@ -282,7 +289,7 @@ zvol_write(zv_request_t *zvr)
blk_generic_end_io_acct(q, disk, WRITE, bio, start_time);
}
- zvol_end_io(bio, rq, -error);
+ zvol_end_io(bio, rq, error);
}
static void
@@ -361,7 +368,7 @@ zvol_discard(zv_request_t *zvr)
zfs_rangelock_exit(lr);
if (error == 0 && sync)
- zil_commit(zv->zv_zilog, ZVOL_OBJ);
+ error = zil_commit(zv->zv_zilog, ZVOL_OBJ);
unlock:
rw_exit(&zv->zv_suspend_lock);
@@ -371,7 +378,7 @@ unlock:
start_time);
}
- zvol_end_io(bio, rq, -error);
+ zvol_end_io(bio, rq, error);
}
static void
@@ -449,7 +456,7 @@ zvol_read(zv_request_t *zvr)
blk_generic_end_io_acct(q, disk, READ, bio, start_time);
}
- zvol_end_io(bio, rq, -error);
+ zvol_end_io(bio, rq, error);
}
static void
@@ -480,7 +487,7 @@ zvol_request_impl(zvol_state_t *zv, struct bio *bio, struct request *rq,
int rw = io_data_dir(bio, rq);
if (unlikely(zv->zv_flags & ZVOL_REMOVING)) {
- zvol_end_io(bio, rq, -SET_ERROR(ENXIO));
+ zvol_end_io(bio, rq, SET_ERROR(ENXIO));
goto out;
}
@@ -499,7 +506,7 @@ zvol_request_impl(zvol_state_t *zv, struct bio *bio, struct request *rq,
(long long unsigned)offset,
(long unsigned)size);
- zvol_end_io(bio, rq, -SET_ERROR(EIO));
+ zvol_end_io(bio, rq, SET_ERROR(EIO));
goto out;
}
@@ -512,8 +519,8 @@ zvol_request_impl(zvol_state_t *zv, struct bio *bio, struct request *rq,
#ifdef HAVE_BLK_MQ_RQ_HCTX
blk_mq_hw_queue = rq->mq_hctx->queue_num;
#else
- blk_mq_hw_queue =
- rq->q->queue_hw_ctx[rq->q->mq_map[rq->cpu]]->queue_num;
+ blk_mq_hw_queue = rq->q->queue_hw_ctx[
+ rq->q->mq_map[raw_smp_processor_id()]]->queue_num;
#endif
taskq_hash = cityhash3((uintptr_t)zv, offset >> ZVOL_TASKQ_OFFSET_SHIFT,
blk_mq_hw_queue);
@@ -521,7 +528,7 @@ zvol_request_impl(zvol_state_t *zv, struct bio *bio, struct request *rq,
if (rw == WRITE) {
if (unlikely(zv->zv_flags & ZVOL_RDONLY)) {
- zvol_end_io(bio, rq, -SET_ERROR(EROFS));
+ zvol_end_io(bio, rq, SET_ERROR(EROFS));
goto out;
}
@@ -672,28 +679,19 @@ zvol_open(struct block_device *bdev, fmode_t flag)
retry:
#endif
- rw_enter(&zvol_state_lock, RW_READER);
- /*
- * Obtain a copy of private_data under the zvol_state_lock to make
- * sure that either the result of zvol free code path setting
- * disk->private_data to NULL is observed, or zvol_os_free()
- * is not called on this zv because of the positive zv_open_count.
- */
+
#ifdef HAVE_BLK_MODE_T
- zv = disk->private_data;
+ zv = atomic_load_ptr(&disk->private_data);
#else
- zv = bdev->bd_disk->private_data;
+ zv = atomic_load_ptr(&bdev->bd_disk->private_data);
#endif
if (zv == NULL) {
- rw_exit(&zvol_state_lock);
return (-SET_ERROR(ENXIO));
}
mutex_enter(&zv->zv_state_lock);
-
if (unlikely(zv->zv_flags & ZVOL_REMOVING)) {
mutex_exit(&zv->zv_state_lock);
- rw_exit(&zvol_state_lock);
return (-SET_ERROR(ENXIO));
}
@@ -705,8 +703,28 @@ retry:
if (zv->zv_open_count == 0) {
if (!rw_tryenter(&zv->zv_suspend_lock, RW_READER)) {
mutex_exit(&zv->zv_state_lock);
+
+ /*
+ * Removal may happen while the locks are down, so
+ * we can't trust zv any longer; we have to start over.
+ */
+#ifdef HAVE_BLK_MODE_T
+ zv = atomic_load_ptr(&disk->private_data);
+#else
+ zv = atomic_load_ptr(&bdev->bd_disk->private_data);
+#endif
+ if (zv == NULL)
+ return (-SET_ERROR(ENXIO));
+
rw_enter(&zv->zv_suspend_lock, RW_READER);
mutex_enter(&zv->zv_state_lock);
+
+ if (unlikely(zv->zv_flags & ZVOL_REMOVING)) {
+ mutex_exit(&zv->zv_state_lock);
+ rw_exit(&zv->zv_suspend_lock);
+ return (-SET_ERROR(ENXIO));
+ }
+
/* check to see if zv_suspend_lock is needed */
if (zv->zv_open_count != 0) {
rw_exit(&zv->zv_suspend_lock);
@@ -717,7 +735,6 @@ retry:
drop_suspend = B_TRUE;
}
}
- rw_exit(&zvol_state_lock);
ASSERT(MUTEX_HELD(&zv->zv_state_lock));
@@ -814,11 +831,11 @@ zvol_release(struct gendisk *disk, fmode_t unused)
#if !defined(HAVE_BLOCK_DEVICE_OPERATIONS_RELEASE_1ARG)
(void) unused;
#endif
- zvol_state_t *zv;
boolean_t drop_suspend = B_TRUE;
- rw_enter(&zvol_state_lock, RW_READER);
- zv = disk->private_data;
+ zvol_state_t *zv = atomic_load_ptr(&disk->private_data);
+ if (zv == NULL)
+ return;
mutex_enter(&zv->zv_state_lock);
ASSERT3U(zv->zv_open_count, >, 0);
@@ -832,6 +849,15 @@ zvol_release(struct gendisk *disk, fmode_t unused)
mutex_exit(&zv->zv_state_lock);
rw_enter(&zv->zv_suspend_lock, RW_READER);
mutex_enter(&zv->zv_state_lock);
+
+ /*
+ * Unlike in zvol_open(), we don't check if removal
+ * started here, because we might be one of the openers
+ * that needs to be thrown out! If we're the last, we
+ * need to call zvol_last_close() below to finish
+ * cleanup. So, no special treatment for us.
+ */
+
/* check to see if zv_suspend_lock is needed */
if (zv->zv_open_count != 1) {
rw_exit(&zv->zv_suspend_lock);
@@ -841,7 +867,6 @@ zvol_release(struct gendisk *disk, fmode_t unused)
} else {
drop_suspend = B_FALSE;
}
- rw_exit(&zvol_state_lock);
ASSERT(MUTEX_HELD(&zv->zv_state_lock));
@@ -861,9 +886,10 @@ static int
zvol_ioctl(struct block_device *bdev, fmode_t mode,
unsigned int cmd, unsigned long arg)
{
- zvol_state_t *zv = bdev->bd_disk->private_data;
int error = 0;
+ zvol_state_t *zv = atomic_load_ptr(&bdev->bd_disk->private_data);
+ ASSERT3P(zv, !=, NULL);
ASSERT3U(zv->zv_open_count, >, 0);
switch (cmd) {
@@ -886,16 +912,18 @@ zvol_ioctl(struct block_device *bdev, fmode_t mode,
case BLKZNAME:
mutex_enter(&zv->zv_state_lock);
- error = copy_to_user((void *)arg, zv->zv_name, MAXNAMELEN);
+ error = -copy_to_user((void *)arg, zv->zv_name, MAXNAMELEN);
mutex_exit(&zv->zv_state_lock);
+ if (error)
+ error = SET_ERROR(error);
break;
default:
- error = -ENOTTY;
+ error = SET_ERROR(ENOTTY);
break;
}
- return (SET_ERROR(error));
+ return (-error);
}
#ifdef CONFIG_COMPAT
@@ -914,9 +942,8 @@ zvol_check_events(struct gendisk *disk, unsigned int clearing)
{
unsigned int mask = 0;
- rw_enter(&zvol_state_lock, RW_READER);
+ zvol_state_t *zv = atomic_load_ptr(&disk->private_data);
- zvol_state_t *zv = disk->private_data;
if (zv != NULL) {
mutex_enter(&zv->zv_state_lock);
mask = zv->zv_changed ? DISK_EVENT_MEDIA_CHANGE : 0;
@@ -924,17 +951,14 @@ zvol_check_events(struct gendisk *disk, unsigned int clearing)
mutex_exit(&zv->zv_state_lock);
}
- rw_exit(&zvol_state_lock);
-
return (mask);
}
static int
zvol_revalidate_disk(struct gendisk *disk)
{
- rw_enter(&zvol_state_lock, RW_READER);
+ zvol_state_t *zv = atomic_load_ptr(&disk->private_data);
- zvol_state_t *zv = disk->private_data;
if (zv != NULL) {
mutex_enter(&zv->zv_state_lock);
set_capacity(zv->zv_zso->zvo_disk,
@@ -942,8 +966,6 @@ zvol_revalidate_disk(struct gendisk *disk)
mutex_exit(&zv->zv_state_lock);
}
- rw_exit(&zvol_state_lock);
-
return (0);
}
@@ -962,16 +984,6 @@ zvol_os_update_volsize(zvol_state_t *zv, uint64_t volsize)
return (0);
}
-void
-zvol_os_clear_private(zvol_state_t *zv)
-{
- /*
- * Cleared while holding zvol_state_lock as a writer
- * which will prevent zvol_open() from opening it.
- */
- zv->zv_zso->zvo_disk->private_data = NULL;
-}
-
/*
* Provide a simple virtual geometry for legacy compatibility. For devices
* smaller than 1 MiB a small head and sector count is used to allow very
@@ -981,9 +993,10 @@ zvol_os_clear_private(zvol_state_t *zv)
static int
zvol_getgeo(struct block_device *bdev, struct hd_geometry *geo)
{
- zvol_state_t *zv = bdev->bd_disk->private_data;
sector_t sectors;
+ zvol_state_t *zv = atomic_load_ptr(&bdev->bd_disk->private_data);
+ ASSERT3P(zv, !=, NULL);
ASSERT3U(zv->zv_open_count, >, 0);
sectors = get_capacity(zv->zv_zso->zvo_disk);
@@ -1408,53 +1421,70 @@ out_kmem:
return (ret);
}
-/*
- * Cleanup then free a zvol_state_t which was created by zvol_alloc().
- * At this time, the structure is not opened by anyone, is taken off
- * the zvol_state_list, and has its private data set to NULL.
- * The zvol_state_lock is dropped.
- *
- * This function may take many milliseconds to complete (e.g. we've seen
- * it take over 256ms), due to the calls to "blk_cleanup_queue" and
- * "del_gendisk". Thus, consumers need to be careful to account for this
- * latency when calling this function.
- */
void
-zvol_os_free(zvol_state_t *zv)
+zvol_os_remove_minor(zvol_state_t *zv)
{
-
- ASSERT(!RW_LOCK_HELD(&zv->zv_suspend_lock));
- ASSERT(!MUTEX_HELD(&zv->zv_state_lock));
+ ASSERT(MUTEX_HELD(&zv->zv_state_lock));
ASSERT0(zv->zv_open_count);
- ASSERT3P(zv->zv_zso->zvo_disk->private_data, ==, NULL);
+ ASSERT0(atomic_read(&zv->zv_suspend_ref));
+ ASSERT(zv->zv_flags & ZVOL_REMOVING);
- rw_destroy(&zv->zv_suspend_lock);
- zfs_rangelock_fini(&zv->zv_rangelock);
+ struct zvol_state_os *zso = zv->zv_zso;
+ zv->zv_zso = NULL;
+
+ /* Clearing private_data will make new callers return immediately. */
+ atomic_store_ptr(&zso->zvo_disk->private_data, NULL);
- del_gendisk(zv->zv_zso->zvo_disk);
+ /*
+ * Drop the state lock before calling del_gendisk(). There may be
+ * callers waiting to acquire it, but del_gendisk() will block until
+ * they exit, which would deadlock.
+ */
+ mutex_exit(&zv->zv_state_lock);
+
+ del_gendisk(zso->zvo_disk);
#if defined(HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS) && \
(defined(HAVE_BLK_ALLOC_DISK) || defined(HAVE_BLK_ALLOC_DISK_2ARG))
#if defined(HAVE_BLK_CLEANUP_DISK)
- blk_cleanup_disk(zv->zv_zso->zvo_disk);
+ blk_cleanup_disk(zso->zvo_disk);
#else
- put_disk(zv->zv_zso->zvo_disk);
+ put_disk(zso->zvo_disk);
#endif
#else
- blk_cleanup_queue(zv->zv_zso->zvo_queue);
- put_disk(zv->zv_zso->zvo_disk);
+ blk_cleanup_queue(zso->zvo_queue);
+ put_disk(zso->zvo_disk);
#endif
- if (zv->zv_zso->use_blk_mq)
- blk_mq_free_tag_set(&zv->zv_zso->tag_set);
+ if (zso->use_blk_mq)
+ blk_mq_free_tag_set(&zso->tag_set);
- ida_simple_remove(&zvol_ida,
- MINOR(zv->zv_zso->zvo_dev) >> ZVOL_MINOR_BITS);
+ ida_simple_remove(&zvol_ida, MINOR(zso->zvo_dev) >> ZVOL_MINOR_BITS);
+
+ kmem_free(zso, sizeof (struct zvol_state_os));
+
+ mutex_enter(&zv->zv_state_lock);
+}
+
+void
+zvol_os_free(zvol_state_t *zv)
+{
+
+ ASSERT(!RW_LOCK_HELD(&zv->zv_suspend_lock));
+ ASSERT(!MUTEX_HELD(&zv->zv_state_lock));
+ ASSERT0(zv->zv_open_count);
+ ASSERT0P(zv->zv_zso);
+
+ ASSERT0P(zv->zv_objset);
+ ASSERT0P(zv->zv_zilog);
+ ASSERT0P(zv->zv_dn);
+
+ rw_destroy(&zv->zv_suspend_lock);
+ zfs_rangelock_fini(&zv->zv_rangelock);
cv_destroy(&zv->zv_removing_cv);
mutex_destroy(&zv->zv_state_lock);
dataset_kstats_destroy(&zv->zv_kstat);
- kmem_free(zv->zv_zso, sizeof (struct zvol_state_os));
kmem_free(zv, sizeof (zvol_state_t));
}
@@ -1474,7 +1504,9 @@ __zvol_os_add_disk(struct gendisk *disk)
{
int error = 0;
#ifdef HAVE_ADD_DISK_RET
- error = add_disk(disk);
+ error = -add_disk(disk);
+ if (error)
+ error = SET_ERROR(error);
#else
add_disk(disk);
#endif
@@ -1649,11 +1681,11 @@ zvol_os_create_minor(const char *name)
blk_queue_flag_set(QUEUE_FLAG_SCSI_PASSTHROUGH, zv->zv_zso->zvo_queue);
#endif
- ASSERT3P(zv->zv_kstat.dk_kstats, ==, NULL);
+ ASSERT0P(zv->zv_kstat.dk_kstats);
error = dataset_kstats_create(&zv->zv_kstat, zv->zv_objset);
if (error)
goto out_dmu_objset_disown;
- ASSERT3P(zv->zv_zilog, ==, NULL);
+ ASSERT0P(zv->zv_zilog);
zv->zv_zilog = zil_open(os, zvol_get_data, &zv->zv_kstat.dk_zil_sums);
if (spa_writeable(dmu_objset_spa(os))) {
if (zil_replay_disable)
@@ -1759,10 +1791,10 @@ zvol_init(void)
return (error);
}
- error = register_blkdev(zvol_major, ZVOL_DRIVER);
+ error = -register_blkdev(zvol_major, ZVOL_DRIVER);
if (error) {
printk(KERN_INFO "ZFS: register_blkdev() failed %d\n", error);
- return (error);
+ return (SET_ERROR(error));
}
if (zvol_blk_mq_queue_depth == 0) {
diff --git a/sys/contrib/openzfs/module/zcommon/simd_stat.c b/sys/contrib/openzfs/module/zcommon/simd_stat.c
index 11e2080ff9f2..007ae9e4fbbc 100644
--- a/sys/contrib/openzfs/module/zcommon/simd_stat.c
+++ b/sys/contrib/openzfs/module/zcommon/simd_stat.c
@@ -118,6 +118,10 @@ simd_stat_kstat_data(char *buf, size_t size, void *data)
"pclmulqdq", zfs_pclmulqdq_available());
off += SIMD_STAT_PRINT(simd_stat_kstat_payload,
"movbe", zfs_movbe_available());
+ off += SIMD_STAT_PRINT(simd_stat_kstat_payload,
+ "vaes", zfs_vaes_available());
+ off += SIMD_STAT_PRINT(simd_stat_kstat_payload,
+ "vpclmulqdq", zfs_vpclmulqdq_available());
off += SIMD_STAT_PRINT(simd_stat_kstat_payload,
"osxsave", boot_cpu_has(X86_FEATURE_OSXSAVE));
diff --git a/sys/contrib/openzfs/module/zfs/abd.c b/sys/contrib/openzfs/module/zfs/abd.c
index 826928e67350..bf9b13c30509 100644
--- a/sys/contrib/openzfs/module/zfs/abd.c
+++ b/sys/contrib/openzfs/module/zfs/abd.c
@@ -563,7 +563,7 @@ abd_get_offset_impl(abd_t *abd, abd_t *sabd, size_t off, size_t size)
left -= csize;
off = 0;
}
- ASSERT3U(left, ==, 0);
+ ASSERT0(left);
} else {
abd = abd_get_offset_scatter(abd, sabd, off, size);
}
diff --git a/sys/contrib/openzfs/module/zfs/arc.c b/sys/contrib/openzfs/module/zfs/arc.c
index 3483be64ec57..df41e3b49204 100644
--- a/sys/contrib/openzfs/module/zfs/arc.c
+++ b/sys/contrib/openzfs/module/zfs/arc.c
@@ -2239,8 +2239,8 @@ arc_evictable_space_increment(arc_buf_hdr_t *hdr, arc_state_t *state)
ASSERT(HDR_HAS_L1HDR(hdr));
if (GHOST_STATE(state)) {
- ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
- ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_buf);
+ ASSERT0P(hdr->b_l1hdr.b_pabd);
ASSERT(!HDR_HAS_RABD(hdr));
(void) zfs_refcount_add_many(&state->arcs_esize[type],
HDR_GET_LSIZE(hdr), hdr);
@@ -2278,8 +2278,8 @@ arc_evictable_space_decrement(arc_buf_hdr_t *hdr, arc_state_t *state)
ASSERT(HDR_HAS_L1HDR(hdr));
if (GHOST_STATE(state)) {
- ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
- ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_buf);
+ ASSERT0P(hdr->b_l1hdr.b_pabd);
ASSERT(!HDR_HAS_RABD(hdr));
(void) zfs_refcount_remove_many(&state->arcs_esize[type],
HDR_GET_LSIZE(hdr), hdr);
@@ -2319,7 +2319,7 @@ add_reference(arc_buf_hdr_t *hdr, const void *tag)
if (!HDR_EMPTY(hdr) && !MUTEX_HELD(HDR_LOCK(hdr))) {
ASSERT(state == arc_anon);
ASSERT(zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt));
- ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_buf);
}
if ((zfs_refcount_add(&hdr->b_l1hdr.b_refcnt, tag) == 1) &&
@@ -2503,7 +2503,7 @@ arc_change_state(arc_state_t *new_state, arc_buf_hdr_t *hdr)
(void) zfs_refcount_add_many(
&new_state->arcs_size[type],
HDR_GET_LSIZE(hdr), hdr);
- ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_pabd);
ASSERT(!HDR_HAS_RABD(hdr));
} else {
@@ -2547,7 +2547,7 @@ arc_change_state(arc_state_t *new_state, arc_buf_hdr_t *hdr)
if (update_old && old_state != arc_l2c_only) {
ASSERT(HDR_HAS_L1HDR(hdr));
if (GHOST_STATE(old_state)) {
- ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_pabd);
ASSERT(!HDR_HAS_RABD(hdr));
/*
@@ -2758,7 +2758,7 @@ arc_buf_alloc_impl(arc_buf_hdr_t *hdr, spa_t *spa, const zbookmark_phys_t *zb,
VERIFY(hdr->b_type == ARC_BUFC_DATA ||
hdr->b_type == ARC_BUFC_METADATA);
ASSERT3P(ret, !=, NULL);
- ASSERT3P(*ret, ==, NULL);
+ ASSERT0P(*ret);
IMPLY(encrypted, compressed);
buf = *ret = kmem_cache_alloc(buf_cache, KM_PUSHPAGE);
@@ -2982,7 +2982,7 @@ static void
arc_share_buf(arc_buf_hdr_t *hdr, arc_buf_t *buf)
{
ASSERT(arc_can_share(hdr, buf));
- ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_pabd);
ASSERT(!ARC_BUF_ENCRYPTED(buf));
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
@@ -3201,14 +3201,14 @@ arc_hdr_alloc_abd(arc_buf_hdr_t *hdr, int alloc_flags)
if (alloc_rdata) {
size = HDR_GET_PSIZE(hdr);
- ASSERT3P(hdr->b_crypt_hdr.b_rabd, ==, NULL);
+ ASSERT0P(hdr->b_crypt_hdr.b_rabd);
hdr->b_crypt_hdr.b_rabd = arc_get_data_abd(hdr, size, hdr,
alloc_flags);
ASSERT3P(hdr->b_crypt_hdr.b_rabd, !=, NULL);
ARCSTAT_INCR(arcstat_raw_size, size);
} else {
size = arc_hdr_size(hdr);
- ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_pabd);
hdr->b_l1hdr.b_pabd = arc_get_data_abd(hdr, size, hdr,
alloc_flags);
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
@@ -3290,7 +3290,7 @@ arc_hdr_alloc(uint64_t spa, int32_t psize, int32_t lsize,
ASSERT(HDR_EMPTY(hdr));
#ifdef ZFS_DEBUG
- ASSERT3P(hdr->b_l1hdr.b_freeze_cksum, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_freeze_cksum);
#endif
HDR_SET_PSIZE(hdr, psize);
HDR_SET_LSIZE(hdr, lsize);
@@ -3351,12 +3351,12 @@ arc_hdr_realloc(arc_buf_hdr_t *hdr, kmem_cache_t *old, kmem_cache_t *new)
nhdr->b_l1hdr.b_state = arc_l2c_only;
/* Verify previous threads set to NULL before freeing */
- ASSERT3P(nhdr->b_l1hdr.b_pabd, ==, NULL);
+ ASSERT0P(nhdr->b_l1hdr.b_pabd);
ASSERT(!HDR_HAS_RABD(hdr));
} else {
- ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_buf);
#ifdef ZFS_DEBUG
- ASSERT3P(hdr->b_l1hdr.b_freeze_cksum, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_freeze_cksum);
#endif
/*
@@ -3375,7 +3375,7 @@ arc_hdr_realloc(arc_buf_hdr_t *hdr, kmem_cache_t *old, kmem_cache_t *new)
* might try to be accessed, even though it was removed.
*/
VERIFY(!HDR_L2_WRITING(hdr));
- VERIFY3P(hdr->b_l1hdr.b_pabd, ==, NULL);
+ VERIFY0P(hdr->b_l1hdr.b_pabd);
ASSERT(!HDR_HAS_RABD(hdr));
arc_hdr_clear_flags(nhdr, ARC_FLAG_HAS_L1HDR);
@@ -3698,12 +3698,12 @@ arc_hdr_destroy(arc_buf_hdr_t *hdr)
arc_hdr_free_abd(hdr, B_TRUE);
}
- ASSERT3P(hdr->b_hash_next, ==, NULL);
+ ASSERT0P(hdr->b_hash_next);
if (HDR_HAS_L1HDR(hdr)) {
ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node));
- ASSERT3P(hdr->b_l1hdr.b_acb, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_acb);
#ifdef ZFS_DEBUG
- ASSERT3P(hdr->b_l1hdr.b_freeze_cksum, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_freeze_cksum);
#endif
kmem_cache_free(hdr_full_cache, hdr);
} else {
@@ -3771,7 +3771,7 @@ arc_evict_hdr(arc_buf_hdr_t *hdr, uint64_t *real_evicted)
ASSERT(MUTEX_HELD(HDR_LOCK(hdr)));
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
- ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_buf);
ASSERT0(zfs_refcount_count(&hdr->b_l1hdr.b_refcnt));
*real_evicted = 0;
@@ -3796,7 +3796,7 @@ arc_evict_hdr(arc_buf_hdr_t *hdr, uint64_t *real_evicted)
DTRACE_PROBE1(arc__delete, arc_buf_hdr_t *, hdr);
if (HDR_HAS_L2HDR(hdr)) {
- ASSERT(hdr->b_l1hdr.b_pabd == NULL);
+ ASSERT0P(hdr->b_l1hdr.b_pabd);
ASSERT(!HDR_HAS_RABD(hdr));
/*
* This buffer is cached on the 2nd Level ARC;
@@ -5554,7 +5554,7 @@ static void
arc_hdr_verify(arc_buf_hdr_t *hdr, blkptr_t *bp)
{
if (BP_IS_HOLE(bp) || BP_IS_EMBEDDED(bp)) {
- ASSERT3U(HDR_GET_PSIZE(hdr), ==, 0);
+ ASSERT0(HDR_GET_PSIZE(hdr));
ASSERT3U(arc_hdr_get_compress(hdr), ==, ZIO_COMPRESS_OFF);
} else {
if (HDR_COMPRESSION_ENABLED(hdr)) {
@@ -6132,14 +6132,14 @@ top:
}
if (GHOST_STATE(hdr->b_l1hdr.b_state)) {
- ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_pabd);
ASSERT(!HDR_HAS_RABD(hdr));
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
ASSERT0(zfs_refcount_count(
&hdr->b_l1hdr.b_refcnt));
- ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_buf);
#ifdef ZFS_DEBUG
- ASSERT3P(hdr->b_l1hdr.b_freeze_cksum, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_freeze_cksum);
#endif
} else if (HDR_IO_IN_PROGRESS(hdr)) {
/*
@@ -6233,7 +6233,7 @@ top:
acb->acb_nobuf = no_buf;
acb->acb_zb = *zb;
- ASSERT3P(hdr->b_l1hdr.b_acb, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_acb);
hdr->b_l1hdr.b_acb = acb;
if (HDR_HAS_L2HDR(hdr) &&
@@ -6717,7 +6717,7 @@ arc_release(arc_buf_t *buf, const void *tag)
nhdr = arc_hdr_alloc(spa, psize, lsize, protected,
compress, hdr->b_complevel, type);
- ASSERT3P(nhdr->b_l1hdr.b_buf, ==, NULL);
+ ASSERT0P(nhdr->b_l1hdr.b_buf);
ASSERT0(zfs_refcount_count(&nhdr->b_l1hdr.b_refcnt));
VERIFY3U(nhdr->b_type, ==, type);
ASSERT(!HDR_SHARED_DATA(nhdr));
@@ -6804,7 +6804,7 @@ arc_write_ready(zio_t *zio)
if (HDR_HAS_RABD(hdr))
arc_hdr_free_abd(hdr, B_TRUE);
}
- ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_pabd);
ASSERT(!HDR_HAS_RABD(hdr));
ASSERT(!HDR_SHARED_DATA(hdr));
ASSERT(!arc_buf_is_shared(buf));
@@ -6948,7 +6948,7 @@ arc_write_done(zio_t *zio)
arc_buf_t *buf = callback->awcb_buf;
arc_buf_hdr_t *hdr = buf->b_hdr;
- ASSERT3P(hdr->b_l1hdr.b_acb, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_acb);
if (zio->io_error == 0) {
arc_hdr_verify(hdr, zio->io_bp);
@@ -6973,7 +6973,7 @@ arc_write_done(zio_t *zio)
arc_buf_hdr_t *exists;
kmutex_t *hash_lock;
- ASSERT3U(zio->io_error, ==, 0);
+ ASSERT0(zio->io_error);
arc_cksum_verify(buf);
@@ -6994,7 +6994,7 @@ arc_write_done(zio_t *zio)
arc_hdr_destroy(exists);
mutex_exit(hash_lock);
exists = buf_hash_insert(hdr, &hash_lock);
- ASSERT3P(exists, ==, NULL);
+ ASSERT0P(exists);
} else if (zio->io_flags & ZIO_FLAG_NOPWRITE) {
/* nopwrite */
ASSERT(zio->io_prop.zp_nopwrite);
@@ -7007,7 +7007,7 @@ arc_write_done(zio_t *zio)
ASSERT(ARC_BUF_LAST(hdr->b_l1hdr.b_buf));
ASSERT(hdr->b_l1hdr.b_state == arc_anon);
ASSERT(BP_GET_DEDUP(zio->io_bp));
- ASSERT(BP_GET_LEVEL(zio->io_bp) == 0);
+ ASSERT0(BP_GET_LEVEL(zio->io_bp));
}
}
arc_hdr_clear_flags(hdr, ARC_FLAG_IO_IN_PROGRESS);
@@ -7044,7 +7044,7 @@ arc_write(zio_t *pio, spa_t *spa, uint64_t txg,
ASSERT3P(done, !=, NULL);
ASSERT(!HDR_IO_ERROR(hdr));
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
- ASSERT3P(hdr->b_l1hdr.b_acb, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_acb);
ASSERT3P(hdr->b_l1hdr.b_buf, !=, NULL);
if (uncached)
arc_hdr_set_flags(hdr, ARC_FLAG_UNCACHED);
@@ -7113,7 +7113,7 @@ arc_write(zio_t *pio, spa_t *spa, uint64_t txg,
arc_hdr_set_compress(hdr, ZIO_COMPRESS_OFF);
ASSERT(!arc_buf_is_shared(buf));
- ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
+ ASSERT0P(hdr->b_l1hdr.b_pabd);
zio = zio_write(pio, spa, txg, bp,
abd_get_from_buf(buf->b_data, HDR_GET_LSIZE(hdr)),
diff --git a/sys/contrib/openzfs/module/zfs/bpobj.c b/sys/contrib/openzfs/module/zfs/bpobj.c
index 0a8a077edf63..ea9fbd036c6e 100644
--- a/sys/contrib/openzfs/module/zfs/bpobj.c
+++ b/sys/contrib/openzfs/module/zfs/bpobj.c
@@ -160,8 +160,8 @@ bpobj_open(bpobj_t *bpo, objset_t *os, uint64_t object)
memset(bpo, 0, sizeof (*bpo));
mutex_init(&bpo->bpo_lock, NULL, MUTEX_DEFAULT, NULL);
- ASSERT(bpo->bpo_dbuf == NULL);
- ASSERT(bpo->bpo_phys == NULL);
+ ASSERT0P(bpo->bpo_dbuf);
+ ASSERT0P(bpo->bpo_phys);
ASSERT(object != 0);
ASSERT3U(doi.doi_type, ==, DMU_OT_BPOBJ);
ASSERT3U(doi.doi_bonus_type, ==, DMU_OT_BPOBJ_HDR);
@@ -478,7 +478,7 @@ bpobj_iterate_impl(bpobj_t *initial_bpo, bpobj_itor_t func, void *arg,
* We have unprocessed subobjs. Process the next one.
*/
ASSERT(bpo->bpo_havecomp);
- ASSERT3P(bpobj_size, ==, NULL);
+ ASSERT0P(bpobj_size);
/* Add the last subobj to stack. */
int64_t i = bpi->bpi_unprocessed_subobjs - 1;
diff --git a/sys/contrib/openzfs/module/zfs/btree.c b/sys/contrib/openzfs/module/zfs/btree.c
index aa282f711bc3..725b96a3b2c7 100644
--- a/sys/contrib/openzfs/module/zfs/btree.c
+++ b/sys/contrib/openzfs/module/zfs/btree.c
@@ -1110,7 +1110,7 @@ zfs_btree_add_idx(zfs_btree_t *tree, const void *value,
if (where->bti_node == NULL) {
ASSERT3U(tree->bt_num_elems, ==, 1);
ASSERT3S(tree->bt_height, ==, -1);
- ASSERT3P(tree->bt_root, ==, NULL);
+ ASSERT0P(tree->bt_root);
ASSERT0(where->bti_offset);
tree->bt_num_nodes++;
@@ -1947,7 +1947,7 @@ void
zfs_btree_destroy(zfs_btree_t *tree)
{
ASSERT0(tree->bt_num_elems);
- ASSERT3P(tree->bt_root, ==, NULL);
+ ASSERT0P(tree->bt_root);
}
/* Verify that every child of this node has the correct parent pointer. */
@@ -1969,10 +1969,10 @@ static void
zfs_btree_verify_pointers(zfs_btree_t *tree)
{
if (tree->bt_height == -1) {
- VERIFY3P(tree->bt_root, ==, NULL);
+ VERIFY0P(tree->bt_root);
return;
}
- VERIFY3P(tree->bt_root->bth_parent, ==, NULL);
+ VERIFY0P(tree->bt_root->bth_parent);
zfs_btree_verify_pointers_helper(tree, tree->bt_root);
}
diff --git a/sys/contrib/openzfs/module/zfs/dataset_kstats.c b/sys/contrib/openzfs/module/zfs/dataset_kstats.c
index d3baabd6169f..e5abcd2044cf 100644
--- a/sys/contrib/openzfs/module/zfs/dataset_kstats.c
+++ b/sys/contrib/openzfs/module/zfs/dataset_kstats.c
@@ -44,6 +44,7 @@ static dataset_kstat_values_t empty_dataset_kstats = {
{ "zil_commit_error_count", KSTAT_DATA_UINT64 },
{ "zil_commit_stall_count", KSTAT_DATA_UINT64 },
{ "zil_commit_suspend_count", KSTAT_DATA_UINT64 },
+ { "zil_commit_crash_count", KSTAT_DATA_UINT64 },
{ "zil_itx_count", KSTAT_DATA_UINT64 },
{ "zil_itx_indirect_count", KSTAT_DATA_UINT64 },
{ "zil_itx_indirect_bytes", KSTAT_DATA_UINT64 },
diff --git a/sys/contrib/openzfs/module/zfs/dbuf.c b/sys/contrib/openzfs/module/zfs/dbuf.c
index 432c99cec960..7403f10d91b7 100644
--- a/sys/contrib/openzfs/module/zfs/dbuf.c
+++ b/sys/contrib/openzfs/module/zfs/dbuf.c
@@ -523,7 +523,7 @@ dbuf_verify_user(dmu_buf_impl_t *db, dbvu_verify_type_t verify_type)
return;
/* Only data blocks support the attachment of user data. */
- ASSERT(db->db_level == 0);
+ ASSERT0(db->db_level);
/* Clients must resolve a dbuf before attaching user data. */
ASSERT(db->db.db_data != NULL);
@@ -1128,8 +1128,8 @@ dbuf_verify(dmu_buf_impl_t *db)
DB_DNODE_ENTER(db);
dn = DB_DNODE(db);
if (dn == NULL) {
- ASSERT(db->db_parent == NULL);
- ASSERT(db->db_blkptr == NULL);
+ ASSERT0P(db->db_parent);
+ ASSERT0P(db->db_blkptr);
} else {
ASSERT3U(db->db.db_object, ==, dn->dn_object);
ASSERT3P(db->db_objset, ==, dn->dn_objset);
@@ -1180,7 +1180,7 @@ dbuf_verify(dmu_buf_impl_t *db)
/* db is pointed to by the dnode */
/* ASSERT3U(db->db_blkid, <, dn->dn_nblkptr); */
if (DMU_OBJECT_IS_SPECIAL(db->db.db_object))
- ASSERT(db->db_parent == NULL);
+ ASSERT0P(db->db_parent);
else
ASSERT(db->db_parent != NULL);
if (db->db_blkid != DMU_SPILL_BLKID)
@@ -1219,7 +1219,7 @@ dbuf_verify(dmu_buf_impl_t *db)
int i;
for (i = 0; i < db->db.db_size >> 3; i++) {
- ASSERT(buf[i] == 0);
+ ASSERT0(buf[i]);
}
} else {
blkptr_t *bps = db->db.db_data;
@@ -1259,7 +1259,7 @@ dbuf_clear_data(dmu_buf_impl_t *db)
{
ASSERT(MUTEX_HELD(&db->db_mtx));
dbuf_evict_user(db);
- ASSERT3P(db->db_buf, ==, NULL);
+ ASSERT0P(db->db_buf);
db->db.db_data = NULL;
if (db->db_state != DB_NOFILL) {
db->db_state = DB_UNCACHED;
@@ -1384,13 +1384,13 @@ dbuf_read_done(zio_t *zio, const zbookmark_phys_t *zb, const blkptr_t *bp,
* All reads are synchronous, so we must have a hold on the dbuf
*/
ASSERT(zfs_refcount_count(&db->db_holds) > 0);
- ASSERT(db->db_buf == NULL);
- ASSERT(db->db.db_data == NULL);
+ ASSERT0P(db->db_buf);
+ ASSERT0P(db->db.db_data);
if (buf == NULL) {
/* i/o error */
ASSERT(zio == NULL || zio->io_error != 0);
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
- ASSERT3P(db->db_buf, ==, NULL);
+ ASSERT0P(db->db_buf);
db->db_state = DB_UNCACHED;
DTRACE_SET_STATE(db, "i/o error");
} else if (db->db_level == 0 && db->db_freed_in_flight) {
@@ -1584,7 +1584,7 @@ dbuf_read_impl(dmu_buf_impl_t *db, dnode_t *dn, zio_t *zio, dmu_flags_t flags,
ASSERT(!zfs_refcount_is_zero(&db->db_holds));
ASSERT(MUTEX_HELD(&db->db_mtx));
ASSERT(db->db_state == DB_UNCACHED || db->db_state == DB_NOFILL);
- ASSERT(db->db_buf == NULL);
+ ASSERT0P(db->db_buf);
ASSERT(db->db_parent == NULL ||
RW_LOCK_HELD(&db->db_parent->db_rwlock));
@@ -1682,7 +1682,7 @@ dbuf_fix_old_data(dmu_buf_impl_t *db, uint64_t txg)
ASSERT(MUTEX_HELD(&db->db_mtx));
ASSERT(db->db.db_data != NULL);
- ASSERT(db->db_level == 0);
+ ASSERT0(db->db_level);
ASSERT(db->db.db_object != DMU_META_DNODE_OBJECT);
if (dr == NULL ||
@@ -1901,8 +1901,8 @@ dbuf_noread(dmu_buf_impl_t *db, dmu_flags_t flags)
while (db->db_state == DB_READ || db->db_state == DB_FILL)
cv_wait(&db->db_changed, &db->db_mtx);
if (db->db_state == DB_UNCACHED) {
- ASSERT(db->db_buf == NULL);
- ASSERT(db->db.db_data == NULL);
+ ASSERT0P(db->db_buf);
+ ASSERT0P(db->db.db_data);
dbuf_set_data(db, dbuf_alloc_arcbuf(db));
db->db_state = DB_FILL;
DTRACE_SET_STATE(db, "assigning filled buffer");
@@ -1929,7 +1929,7 @@ dbuf_unoverride(dbuf_dirty_record_t *dr)
* comes from dbuf_dirty() callers who must also hold a range lock.
*/
ASSERT(dr->dt.dl.dr_override_state != DR_IN_DMU_SYNC);
- ASSERT(db->db_level == 0);
+ ASSERT0(db->db_level);
if (db->db_blkid == DMU_BONUS_BLKID ||
dr->dt.dl.dr_override_state == DR_NOT_OVERRIDDEN)
@@ -1994,7 +1994,7 @@ dbuf_free_range(dnode_t *dn, uint64_t start_blkid, uint64_t end_blkid,
mutex_enter(&dn->dn_dbufs_mtx);
db = avl_find(&dn->dn_dbufs, db_search, &where);
- ASSERT3P(db, ==, NULL);
+ ASSERT0P(db);
db = avl_nearest(&dn->dn_dbufs, where, AVL_AFTER);
@@ -2017,7 +2017,7 @@ dbuf_free_range(dnode_t *dn, uint64_t start_blkid, uint64_t end_blkid,
if (db->db_state == DB_UNCACHED ||
db->db_state == DB_NOFILL ||
db->db_state == DB_EVICTING) {
- ASSERT(db->db.db_data == NULL);
+ ASSERT0P(db->db.db_data);
mutex_exit(&db->db_mtx);
continue;
}
@@ -2557,12 +2557,13 @@ dbuf_undirty(dmu_buf_impl_t *db, dmu_tx_t *tx)
/*
* Due to our use of dn_nlevels below, this can only be called
- * in open context, unless we are operating on the MOS.
- * From syncing context, dn_nlevels may be different from the
- * dn_nlevels used when dbuf was dirtied.
+ * in open context, unless we are operating on the MOS or it's
+ * a special object. From syncing context, dn_nlevels may be
+ * different from the dn_nlevels used when dbuf was dirtied.
*/
ASSERT(db->db_objset ==
dmu_objset_pool(db->db_objset)->dp_meta_objset ||
+ DMU_OBJECT_IS_SPECIAL(db->db.db_object) ||
txg != spa_syncing_txg(dmu_objset_spa(db->db_objset)));
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
ASSERT0(db->db_level);
@@ -2896,8 +2897,8 @@ dmu_buf_will_clone_or_dio(dmu_buf_t *db_fake, dmu_tx_t *tx)
dbuf_clear_data(db);
}
- ASSERT3P(db->db_buf, ==, NULL);
- ASSERT3P(db->db.db_data, ==, NULL);
+ ASSERT0P(db->db_buf);
+ ASSERT0P(db->db.db_data);
db->db_state = DB_NOFILL;
DTRACE_SET_STATE(db,
@@ -2932,7 +2933,7 @@ dmu_buf_will_fill_flags(dmu_buf_t *db_fake, dmu_tx_t *tx, boolean_t canfail,
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
ASSERT(tx->tx_txg != 0);
- ASSERT(db->db_level == 0);
+ ASSERT0(db->db_level);
ASSERT(!zfs_refcount_is_zero(&db->db_holds));
ASSERT(db->db.db_object != DMU_META_DNODE_OBJECT ||
@@ -3144,7 +3145,7 @@ dbuf_assign_arcbuf(dmu_buf_impl_t *db, arc_buf_t *buf, dmu_tx_t *tx,
{
ASSERT(!zfs_refcount_is_zero(&db->db_holds));
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
- ASSERT(db->db_level == 0);
+ ASSERT0(db->db_level);
ASSERT3U(dbuf_is_metadata(db), ==, arc_is_metadata(buf));
ASSERT(buf != NULL);
ASSERT3U(arc_buf_lsize(buf), ==, db->db.db_size);
@@ -3209,7 +3210,7 @@ dbuf_assign_arcbuf(dmu_buf_impl_t *db, arc_buf_t *buf, dmu_tx_t *tx,
VERIFY(!dbuf_undirty(db, tx));
db->db_state = DB_UNCACHED;
}
- ASSERT(db->db_buf == NULL);
+ ASSERT0P(db->db_buf);
dbuf_set_data(db, buf);
db->db_state = DB_FILL;
DTRACE_SET_STATE(db, "filling assigned arcbuf");
@@ -3269,7 +3270,7 @@ dbuf_destroy(dmu_buf_impl_t *db)
}
ASSERT(db->db_state == DB_UNCACHED || db->db_state == DB_NOFILL);
- ASSERT(db->db_data_pending == NULL);
+ ASSERT0P(db->db_data_pending);
ASSERT(list_is_empty(&db->db_dirty_records));
db->db_state = DB_EVICTING;
@@ -3321,11 +3322,11 @@ dbuf_destroy(dmu_buf_impl_t *db)
db->db_parent = NULL;
- ASSERT(db->db_buf == NULL);
- ASSERT(db->db.db_data == NULL);
- ASSERT(db->db_hash_next == NULL);
- ASSERT(db->db_blkptr == NULL);
- ASSERT(db->db_data_pending == NULL);
+ ASSERT0P(db->db_buf);
+ ASSERT0P(db->db.db_data);
+ ASSERT0P(db->db_hash_next);
+ ASSERT0P(db->db_blkptr);
+ ASSERT0P(db->db_data_pending);
ASSERT3U(db->db_caching_status, ==, DB_NO_CACHE);
ASSERT(!multilist_link_active(&db->db_cache_link));
@@ -3960,7 +3961,7 @@ dbuf_hold_impl(dnode_t *dn, uint8_t level, uint64_t blkid,
if (fail_uncached)
return (SET_ERROR(ENOENT));
- ASSERT3P(parent, ==, NULL);
+ ASSERT0P(parent);
err = dbuf_findbp(dn, level, blkid, fail_sparse, &parent, &bp);
if (fail_sparse) {
if (err == 0 && bp && BP_IS_HOLE(bp))
@@ -4064,7 +4065,7 @@ dbuf_create_bonus(dnode_t *dn)
{
ASSERT(RW_WRITE_HELD(&dn->dn_struct_rwlock));
- ASSERT(dn->dn_bonus == NULL);
+ ASSERT0P(dn->dn_bonus);
dn->dn_bonus = dbuf_create(dn, 0, DMU_BONUS_BLKID, dn->dn_dbuf, NULL,
dbuf_hash(dn->dn_objset, dn->dn_object, 0, DMU_BONUS_BLKID));
dn->dn_bonus->db_pending_evict = FALSE;
@@ -4416,7 +4417,7 @@ dbuf_check_blkptr(dnode_t *dn, dmu_buf_impl_t *db)
* inappropriate to hook it in (i.e., nlevels mismatch).
*/
ASSERT(db->db_blkid < dn->dn_phys->dn_nblkptr);
- ASSERT(db->db_parent == NULL);
+ ASSERT0P(db->db_parent);
db->db_parent = dn->dn_dbuf;
db->db_blkptr = &dn->dn_phys->dn_blkptr[db->db_blkid];
DBUF_VERIFY(db);
@@ -4477,7 +4478,7 @@ dbuf_prepare_encrypted_dnode_leaf(dbuf_dirty_record_t *dr)
ASSERT(MUTEX_HELD(&db->db_mtx));
ASSERT3U(db->db.db_object, ==, DMU_META_DNODE_OBJECT);
- ASSERT3U(db->db_level, ==, 0);
+ ASSERT0(db->db_level);
if (!db->db_objset->os_raw_receive && arc_is_encrypted(db->db_buf)) {
zbookmark_phys_t zb;
@@ -4588,7 +4589,7 @@ dbuf_sync_leaf_verify_bonus_dnode(dbuf_dirty_record_t *dr)
/* ensure that everything is zero after our data */
for (; datap_end < datap_max; datap_end++)
- ASSERT(*datap_end == 0);
+ ASSERT0(*datap_end);
#endif
}
@@ -4596,7 +4597,7 @@ static blkptr_t *
dbuf_lightweight_bp(dbuf_dirty_record_t *dr)
{
/* This must be a lightweight dirty record. */
- ASSERT3P(dr->dr_dbuf, ==, NULL);
+ ASSERT0P(dr->dr_dbuf);
dnode_t *dn = dr->dr_dnode;
if (dn->dn_phys->dn_nlevels == 1) {
@@ -4739,7 +4740,7 @@ dbuf_sync_leaf(dbuf_dirty_record_t *dr, dmu_tx_t *tx)
*/
if (db->db_state == DB_UNCACHED) {
/* This buffer has been freed since it was dirtied */
- ASSERT3P(db->db.db_data, ==, NULL);
+ ASSERT0P(db->db.db_data);
} else if (db->db_state == DB_FILL) {
/* This buffer was freed and is now being re-filled */
ASSERT(db->db.db_data != dr->dt.dl.dr_data);
@@ -4756,9 +4757,9 @@ dbuf_sync_leaf(dbuf_dirty_record_t *dr, dmu_tx_t *tx)
*/
dbuf_dirty_record_t *dr_head =
list_head(&db->db_dirty_records);
- ASSERT3P(db->db_buf, ==, NULL);
- ASSERT3P(db->db.db_data, ==, NULL);
- ASSERT3P(dr_head->dt.dl.dr_data, ==, NULL);
+ ASSERT0P(db->db_buf);
+ ASSERT0P(db->db.db_data);
+ ASSERT0P(dr_head->dt.dl.dr_data);
ASSERT3U(dr_head->dt.dl.dr_override_state, ==, DR_OVERRIDDEN);
} else {
ASSERT(db->db_state == DB_CACHED || db->db_state == DB_NOFILL);
diff --git a/sys/contrib/openzfs/module/zfs/ddt.c b/sys/contrib/openzfs/module/zfs/ddt.c
index e0b9fc3951ff..d6658375f810 100644
--- a/sys/contrib/openzfs/module/zfs/ddt.c
+++ b/sys/contrib/openzfs/module/zfs/ddt.c
@@ -397,7 +397,7 @@ ddt_object_create(ddt_t *ddt, ddt_type_t type, ddt_class_t class,
ddt_object_name(ddt, type, class, name);
- ASSERT3U(*objectp, ==, 0);
+ ASSERT0(*objectp);
VERIFY0(ddt_ops[type]->ddt_op_create(os, objectp, tx, prehash));
ASSERT3U(*objectp, !=, 0);
@@ -1011,7 +1011,7 @@ ddt_free(const ddt_t *ddt, ddt_entry_t *dde)
{
if (dde->dde_io != NULL) {
for (int p = 0; p < DDT_NPHYS(ddt); p++)
- ASSERT3P(dde->dde_io->dde_lead_zio[p], ==, NULL);
+ ASSERT0P(dde->dde_io->dde_lead_zio[p]);
if (dde->dde_io->dde_repair_abd != NULL)
abd_free(dde->dde_io->dde_repair_abd);
@@ -1421,7 +1421,7 @@ ddt_key_compare(const void *x1, const void *x2)
static void
ddt_create_dir(ddt_t *ddt, dmu_tx_t *tx)
{
- ASSERT3U(ddt->ddt_dir_object, ==, 0);
+ ASSERT0(ddt->ddt_dir_object);
ASSERT3U(ddt->ddt_version, ==, DDT_VERSION_FDT);
char name[DDT_NAMELEN];
@@ -2395,7 +2395,7 @@ ddt_sync(spa_t *spa, uint64_t txg)
* scan's root zio here so that we can wait for any scan IOs in
* addition to the regular ddt IOs.
*/
- ASSERT3P(scn->scn_zio_root, ==, NULL);
+ ASSERT0P(scn->scn_zio_root);
scn->scn_zio_root = rio;
for (enum zio_checksum c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++) {
diff --git a/sys/contrib/openzfs/module/zfs/ddt_log.c b/sys/contrib/openzfs/module/zfs/ddt_log.c
index dbd381aa9609..3d30e244c1f7 100644
--- a/sys/contrib/openzfs/module/zfs/ddt_log.c
+++ b/sys/contrib/openzfs/module/zfs/ddt_log.c
@@ -116,7 +116,7 @@ static void
ddt_log_create_one(ddt_t *ddt, ddt_log_t *ddl, uint_t n, dmu_tx_t *tx)
{
ASSERT3U(ddt->ddt_dir_object, >, 0);
- ASSERT3U(ddl->ddl_object, ==, 0);
+ ASSERT0(ddl->ddl_object);
char name[DDT_NAMELEN];
ddt_log_name(ddt, name, n);
@@ -194,7 +194,7 @@ void
ddt_log_begin(ddt_t *ddt, size_t nentries, dmu_tx_t *tx, ddt_log_update_t *dlu)
{
ASSERT3U(nentries, >, 0);
- ASSERT3P(dlu->dlu_dbp, ==, NULL);
+ ASSERT0P(dlu->dlu_dbp);
if (ddt->ddt_log_active->ddl_object == 0)
ddt_log_create(ddt, tx);
@@ -748,8 +748,8 @@ ddt_log_load(ddt_t *ddt)
void
ddt_log_alloc(ddt_t *ddt)
{
- ASSERT3P(ddt->ddt_log_active, ==, NULL);
- ASSERT3P(ddt->ddt_log_flushing, ==, NULL);
+ ASSERT0P(ddt->ddt_log_active);
+ ASSERT0P(ddt->ddt_log_flushing);
avl_create(&ddt->ddt_log[0].ddl_tree, ddt_key_compare,
sizeof (ddt_log_entry_t), offsetof(ddt_log_entry_t, ddle_node));
diff --git a/sys/contrib/openzfs/module/zfs/dmu.c b/sys/contrib/openzfs/module/zfs/dmu.c
index 296e58ef9cd8..f7f808d5b8f7 100644
--- a/sys/contrib/openzfs/module/zfs/dmu.c
+++ b/sys/contrib/openzfs/module/zfs/dmu.c
@@ -1343,7 +1343,7 @@ dmu_prealloc(objset_t *os, uint64_t object, uint64_t offset, uint64_t size,
if (size == 0)
return;
- VERIFY(0 == dmu_buf_hold_array(os, object, offset, size,
+ VERIFY0(dmu_buf_hold_array(os, object, offset, size,
FALSE, FTAG, &numbufs, &dbp));
for (i = 0; i < numbufs; i++) {
@@ -1872,7 +1872,7 @@ dmu_sync_ready(zio_t *zio, arc_buf_t *buf, void *varg)
*/
BP_SET_LSIZE(bp, db->db_size);
} else if (!BP_IS_EMBEDDED(bp)) {
- ASSERT(BP_GET_LEVEL(bp) == 0);
+ ASSERT0(BP_GET_LEVEL(bp));
BP_SET_FILL(bp, 1);
}
}
@@ -2405,7 +2405,7 @@ dmu_write_policy(objset_t *os, dnode_t *dn, int level, int wp, zio_prop_t *zp)
}
}
} else if (wp & WP_NOFILL) {
- ASSERT(level == 0);
+ ASSERT0(level);
/*
* If we're writing preallocated blocks, we aren't actually
@@ -2865,7 +2865,7 @@ byteswap_uint64_array(void *vbuf, size_t size)
size_t count = size >> 3;
int i;
- ASSERT((size & 7) == 0);
+ ASSERT0((size & 7));
for (i = 0; i < count; i++)
buf[i] = BSWAP_64(buf[i]);
@@ -2878,7 +2878,7 @@ byteswap_uint32_array(void *vbuf, size_t size)
size_t count = size >> 2;
int i;
- ASSERT((size & 3) == 0);
+ ASSERT0((size & 3));
for (i = 0; i < count; i++)
buf[i] = BSWAP_32(buf[i]);
@@ -2891,7 +2891,7 @@ byteswap_uint16_array(void *vbuf, size_t size)
size_t count = size >> 1;
int i;
- ASSERT((size & 1) == 0);
+ ASSERT0((size & 1));
for (i = 0; i < count; i++)
buf[i] = BSWAP_16(buf[i]);
diff --git a/sys/contrib/openzfs/module/zfs/dmu_direct.c b/sys/contrib/openzfs/module/zfs/dmu_direct.c
index 930ff101eca3..d44c686088fc 100644
--- a/sys/contrib/openzfs/module/zfs/dmu_direct.c
+++ b/sys/contrib/openzfs/module/zfs/dmu_direct.c
@@ -95,9 +95,9 @@ dmu_write_direct_done(zio_t *zio)
abd_free(zio->io_abd);
mutex_enter(&db->db_mtx);
- ASSERT3P(db->db_buf, ==, NULL);
- ASSERT3P(dr->dt.dl.dr_data, ==, NULL);
- ASSERT3P(db->db.db_data, ==, NULL);
+ ASSERT0P(db->db_buf);
+ ASSERT0P(dr->dt.dl.dr_data);
+ ASSERT0P(db->db.db_data);
db->db_state = DB_UNCACHED;
mutex_exit(&db->db_mtx);
diff --git a/sys/contrib/openzfs/module/zfs/dmu_object.c b/sys/contrib/openzfs/module/zfs/dmu_object.c
index b4ff7d224cc9..207cc6d0e713 100644
--- a/sys/contrib/openzfs/module/zfs/dmu_object.c
+++ b/sys/contrib/openzfs/module/zfs/dmu_object.c
@@ -90,7 +90,7 @@ dmu_object_alloc_impl(objset_t *os, dmu_object_type_t ot, int blocksize,
if (allocated_dnode != NULL) {
ASSERT3P(tag, !=, NULL);
} else {
- ASSERT3P(tag, ==, NULL);
+ ASSERT0P(tag);
tag = FTAG;
}
diff --git a/sys/contrib/openzfs/module/zfs/dmu_objset.c b/sys/contrib/openzfs/module/zfs/dmu_objset.c
index c135f620800f..a77f338bdfd3 100644
--- a/sys/contrib/openzfs/module/zfs/dmu_objset.c
+++ b/sys/contrib/openzfs/module/zfs/dmu_objset.c
@@ -724,7 +724,7 @@ dmu_objset_from_ds(dsl_dataset_t *ds, objset_t **osp)
if (err == 0) {
mutex_enter(&ds->ds_lock);
- ASSERT(ds->ds_objset == NULL);
+ ASSERT0P(ds->ds_objset);
ds->ds_objset = os;
mutex_exit(&ds->ds_lock);
}
@@ -2226,7 +2226,7 @@ dmu_objset_userquota_get_ids(dnode_t *dn, boolean_t before, dmu_tx_t *tx)
rf |= DB_RF_HAVESTRUCT;
error = dmu_spill_hold_by_dnode(dn, rf,
FTAG, (dmu_buf_t **)&db);
- ASSERT(error == 0);
+ ASSERT0(error);
mutex_enter(&db->db_mtx);
data = (before) ? db->db.db_data :
dmu_objset_userquota_find_data(db, tx);
diff --git a/sys/contrib/openzfs/module/zfs/dmu_recv.c b/sys/contrib/openzfs/module/zfs/dmu_recv.c
index 73227b58c140..45c7af2bdcd2 100644
--- a/sys/contrib/openzfs/module/zfs/dmu_recv.c
+++ b/sys/contrib/openzfs/module/zfs/dmu_recv.c
@@ -866,7 +866,7 @@ dmu_recv_begin_sync(void *arg, dmu_tx_t *tx)
*/
if (dcp == NULL && drrb->drr_fromguid == 0 &&
drba->drba_origin == NULL) {
- ASSERT3P(dcp, ==, NULL);
+ ASSERT0P(dcp);
dcp = &dummy_dcp;
if (featureflags & DMU_BACKUP_FEATURE_RAW)
@@ -881,7 +881,7 @@ dmu_recv_begin_sync(void *arg, dmu_tx_t *tx)
if (drba->drba_cookie->drc_fromsnapobj != 0) {
VERIFY0(dsl_dataset_hold_obj(dp,
drba->drba_cookie->drc_fromsnapobj, FTAG, &snap));
- ASSERT3P(dcp, ==, NULL);
+ ASSERT0P(dcp);
}
if (drc->drc_heal) {
/* When healing we want to use the provided snapshot */
@@ -905,7 +905,7 @@ dmu_recv_begin_sync(void *arg, dmu_tx_t *tx)
if (drba->drba_origin != NULL) {
VERIFY0(dsl_dataset_hold(dp, drba->drba_origin,
FTAG, &origin));
- ASSERT3P(dcp, ==, NULL);
+ ASSERT0P(dcp);
}
/* Create new dataset. */
@@ -2792,7 +2792,7 @@ receive_read_payload_and_next_header(dmu_recv_cookie_t *drc, int len, void *buf)
drc->drc_rrd->bytes_read = drc->drc_bytes_read;
}
} else {
- ASSERT3P(buf, ==, NULL);
+ ASSERT0P(buf);
}
drc->drc_prev_cksum = drc->drc_cksum;
@@ -3450,7 +3450,7 @@ dmu_recv_stream(dmu_recv_cookie_t *drc, offset_t *voffp)
break;
}
- ASSERT3P(drc->drc_rrd, ==, NULL);
+ ASSERT0P(drc->drc_rrd);
drc->drc_rrd = drc->drc_next_rrd;
drc->drc_next_rrd = NULL;
/* Allocates and loads header into drc->drc_next_rrd */
@@ -3468,7 +3468,7 @@ dmu_recv_stream(dmu_recv_cookie_t *drc, offset_t *voffp)
drc->drc_rrd = NULL;
}
- ASSERT3P(drc->drc_rrd, ==, NULL);
+ ASSERT0P(drc->drc_rrd);
drc->drc_rrd = kmem_zalloc(sizeof (*drc->drc_rrd), KM_SLEEP);
drc->drc_rrd->eos_marker = B_TRUE;
bqueue_enqueue_flush(&rwa->q, drc->drc_rrd, 1);
diff --git a/sys/contrib/openzfs/module/zfs/dmu_redact.c b/sys/contrib/openzfs/module/zfs/dmu_redact.c
index 9226ac9e4b80..5a22ed71a5fe 100644
--- a/sys/contrib/openzfs/module/zfs/dmu_redact.c
+++ b/sys/contrib/openzfs/module/zfs/dmu_redact.c
@@ -1067,7 +1067,7 @@ dmu_redact_snap(const char *snapname, nvlist_t *redactnvl,
}
if (err != 0)
goto out;
- VERIFY3P(nvlist_next_nvpair(redactnvl, pair), ==, NULL);
+ VERIFY0P(nvlist_next_nvpair(redactnvl, pair));
boolean_t resuming = B_FALSE;
zfs_bookmark_phys_t bookmark;
diff --git a/sys/contrib/openzfs/module/zfs/dmu_send.c b/sys/contrib/openzfs/module/zfs/dmu_send.c
index deeba29e159a..8ecb99d5f57c 100644
--- a/sys/contrib/openzfs/module/zfs/dmu_send.c
+++ b/sys/contrib/openzfs/module/zfs/dmu_send.c
@@ -962,7 +962,7 @@ do_dump(dmu_send_cookie_t *dscp, struct send_range *range)
char *data = NULL;
if (srdp->abd != NULL) {
data = abd_to_buf(srdp->abd);
- ASSERT3P(srdp->abuf, ==, NULL);
+ ASSERT0P(srdp->abuf);
} else if (srdp->abuf != NULL) {
data = srdp->abuf->b_data;
}
@@ -2514,7 +2514,7 @@ dmu_send_impl(struct dmu_send_params *dspp)
* list in the stream.
*/
if (dspp->numfromredactsnaps != NUM_SNAPS_NOT_REDACTED) {
- ASSERT3P(from_rl, ==, NULL);
+ ASSERT0P(from_rl);
fnvlist_add_uint64_array(nvl, BEGINNV_REDACT_FROM_SNAPS,
dspp->fromredactsnaps, (uint_t)dspp->numfromredactsnaps);
if (dspp->numfromredactsnaps > 0) {
@@ -2891,7 +2891,7 @@ dmu_send(const char *tosnap, const char *fromsnap, boolean_t embedok,
&fromds);
if (err != 0) {
- ASSERT3P(fromds, ==, NULL);
+ ASSERT0P(fromds);
} else {
/*
* We need to make a deep copy of the redact
diff --git a/sys/contrib/openzfs/module/zfs/dmu_tx.c b/sys/contrib/openzfs/module/zfs/dmu_tx.c
index d85d8b89423e..40c0b3402a05 100644
--- a/sys/contrib/openzfs/module/zfs/dmu_tx.c
+++ b/sys/contrib/openzfs/module/zfs/dmu_tx.c
@@ -126,7 +126,7 @@ dmu_tx_hold_dnode_impl(dmu_tx_t *tx, dnode_t *dn, enum dmu_tx_hold_type type,
* problem, but there's no way for it to happen (for
* now, at least).
*/
- ASSERT(dn->dn_assigned_txg == 0);
+ ASSERT0(dn->dn_assigned_txg);
dn->dn_assigned_txg = tx->tx_txg;
(void) zfs_refcount_add(&dn->dn_tx_holds, tx);
mutex_exit(&dn->dn_mtx);
@@ -443,7 +443,7 @@ dmu_tx_count_free(dmu_tx_hold_t *txh, uint64_t off, uint64_t len)
dnode_t *dn = txh->txh_dnode;
int err;
- ASSERT(tx->tx_txg == 0);
+ ASSERT0(tx->tx_txg);
if (off >= (dn->dn_maxblkid + 1) * dn->dn_datablksz)
return;
@@ -607,7 +607,7 @@ dmu_tx_hold_zap_impl(dmu_tx_hold_t *txh, const char *name)
dnode_t *dn = txh->txh_dnode;
int err;
- ASSERT(tx->tx_txg == 0);
+ ASSERT0(tx->tx_txg);
dmu_tx_count_dnode(txh);
@@ -681,7 +681,7 @@ dmu_tx_hold_bonus(dmu_tx_t *tx, uint64_t object)
{
dmu_tx_hold_t *txh;
- ASSERT(tx->tx_txg == 0);
+ ASSERT0(tx->tx_txg);
txh = dmu_tx_hold_object_impl(tx, tx->tx_objset,
object, THT_BONUS, 0, 0);
@@ -706,7 +706,7 @@ dmu_tx_hold_space(dmu_tx_t *tx, uint64_t space)
{
dmu_tx_hold_t *txh;
- ASSERT(tx->tx_txg == 0);
+ ASSERT0(tx->tx_txg);
txh = dmu_tx_hold_object_impl(tx, tx->tx_objset,
DMU_NEW_OBJECT, THT_SPACE, space, 0);
@@ -1232,7 +1232,7 @@ dmu_tx_assign(dmu_tx_t *tx, dmu_tx_flag_t flags)
{
int err;
- ASSERT(tx->tx_txg == 0);
+ ASSERT0(tx->tx_txg);
ASSERT0(flags & ~(DMU_TX_WAIT | DMU_TX_NOTHROTTLE | DMU_TX_SUSPEND));
IMPLY(flags & DMU_TX_SUSPEND, flags & DMU_TX_WAIT);
ASSERT(!dsl_pool_sync_context(tx->tx_pool));
@@ -1328,7 +1328,7 @@ dmu_tx_wait(dmu_tx_t *tx)
dsl_pool_t *dp = tx->tx_pool;
hrtime_t before;
- ASSERT(tx->tx_txg == 0);
+ ASSERT0(tx->tx_txg);
ASSERT(!dsl_pool_config_held(tx->tx_pool));
/*
@@ -1644,12 +1644,12 @@ dmu_tx_hold_sa(dmu_tx_t *tx, sa_handle_t *hdl, boolean_t may_grow)
dmu_tx_hold_zap(tx, sa->sa_layout_attr_obj, B_TRUE, NULL);
if (sa->sa_force_spill || may_grow || hdl->sa_spill) {
- ASSERT(tx->tx_txg == 0);
+ ASSERT0(tx->tx_txg);
dmu_tx_hold_spill(tx, object);
} else {
DB_DNODE_ENTER(db);
if (DB_DNODE(db)->dn_have_spill) {
- ASSERT(tx->tx_txg == 0);
+ ASSERT0(tx->tx_txg);
dmu_tx_hold_spill(tx, object);
}
DB_DNODE_EXIT(db);
diff --git a/sys/contrib/openzfs/module/zfs/dnode.c b/sys/contrib/openzfs/module/zfs/dnode.c
index 451e1533efa0..963ff41232a3 100644
--- a/sys/contrib/openzfs/module/zfs/dnode.c
+++ b/sys/contrib/openzfs/module/zfs/dnode.c
@@ -214,7 +214,7 @@ dnode_dest(void *arg, void *unused)
for (int i = 0; i < TXG_SIZE; i++) {
ASSERT(!multilist_link_active(&dn->dn_dirty_link[i]));
- ASSERT3P(dn->dn_free_ranges[i], ==, NULL);
+ ASSERT0P(dn->dn_free_ranges[i]);
list_destroy(&dn->dn_dirty_records[i]);
ASSERT0(dn->dn_next_nblkptr[i]);
ASSERT0(dn->dn_next_nlevels[i]);
@@ -231,10 +231,10 @@ dnode_dest(void *arg, void *unused)
ASSERT0(dn->dn_assigned_txg);
ASSERT0(dn->dn_dirty_txg);
ASSERT0(dn->dn_dirtyctx);
- ASSERT3P(dn->dn_dirtyctx_firstset, ==, NULL);
- ASSERT3P(dn->dn_bonus, ==, NULL);
+ ASSERT0P(dn->dn_dirtyctx_firstset);
+ ASSERT0P(dn->dn_bonus);
ASSERT(!dn->dn_have_spill);
- ASSERT3P(dn->dn_zio, ==, NULL);
+ ASSERT0P(dn->dn_zio);
ASSERT0(dn->dn_oldused);
ASSERT0(dn->dn_oldflags);
ASSERT0(dn->dn_olduid);
@@ -318,7 +318,7 @@ dnode_kstats_update(kstat_t *ksp, int rw)
void
dnode_init(void)
{
- ASSERT(dnode_cache == NULL);
+ ASSERT0P(dnode_cache);
dnode_cache = kmem_cache_create("dnode_t", sizeof (dnode_t),
0, dnode_cons, dnode_dest, NULL, NULL, NULL, KMC_RECLAIMABLE);
kmem_cache_set_move(dnode_cache, dnode_move);
@@ -509,7 +509,7 @@ dnode_buf_byteswap(void *vbuf, size_t size)
int i = 0;
ASSERT3U(sizeof (dnode_phys_t), ==, (1<<DNODE_SHIFT));
- ASSERT((size & (sizeof (dnode_phys_t)-1)) == 0);
+ ASSERT0((size & (sizeof (dnode_phys_t)-1)));
while (i < size) {
dnode_phys_t *dnp = (void *)(((char *)vbuf) + i);
@@ -673,7 +673,7 @@ dnode_destroy(dnode_t *dn)
objset_t *os = dn->dn_objset;
boolean_t complete_os_eviction = B_FALSE;
- ASSERT((dn->dn_id_flags & DN_ID_NEW_EXIST) == 0);
+ ASSERT0((dn->dn_id_flags & DN_ID_NEW_EXIST));
mutex_enter(&os->os_lock);
POINTER_INVALIDATE(&dn->dn_objset);
@@ -780,7 +780,7 @@ dnode_allocate(dnode_t *dn, dmu_object_type_t ot, int blocksize, int ibs,
ASSERT0(dn->dn_next_maxblkid[i]);
ASSERT(!multilist_link_active(&dn->dn_dirty_link[i]));
ASSERT3P(list_head(&dn->dn_dirty_records[i]), ==, NULL);
- ASSERT3P(dn->dn_free_ranges[i], ==, NULL);
+ ASSERT0P(dn->dn_free_ranges[i]);
}
dn->dn_type = ot;
@@ -958,7 +958,7 @@ dnode_move_impl(dnode_t *odn, dnode_t *ndn)
ndn->dn_dirty_txg = odn->dn_dirty_txg;
ndn->dn_dirtyctx = odn->dn_dirtyctx;
ndn->dn_dirtyctx_firstset = odn->dn_dirtyctx_firstset;
- ASSERT(zfs_refcount_count(&odn->dn_tx_holds) == 0);
+ ASSERT0(zfs_refcount_count(&odn->dn_tx_holds));
zfs_refcount_transfer(&ndn->dn_holds, &odn->dn_holds);
ASSERT(avl_is_empty(&ndn->dn_dbufs));
avl_swap(&ndn->dn_dbufs, &odn->dn_dbufs);
@@ -2304,7 +2304,7 @@ dnode_free_range(dnode_t *dn, uint64_t off, uint64_t len, dmu_tx_t *tx)
if ((off >> blkshift) > dn->dn_maxblkid)
return;
} else {
- ASSERT(dn->dn_maxblkid == 0);
+ ASSERT0(dn->dn_maxblkid);
if (off == 0 && len >= blksz) {
/*
* Freeing the whole block; fast-track this request.
@@ -2524,7 +2524,7 @@ dnode_diduse_space(dnode_t *dn, int64_t delta)
}
space += delta;
if (spa_version(dn->dn_objset->os_spa) < SPA_VERSION_DNODE_BYTES) {
- ASSERT((dn->dn_phys->dn_flags & DNODE_FLAG_USED_BYTES) == 0);
+ ASSERT0((dn->dn_phys->dn_flags & DNODE_FLAG_USED_BYTES));
ASSERT0(P2PHASE(space, 1<<DEV_BSHIFT));
dn->dn_phys->dn_used = space >> DEV_BSHIFT;
} else {
diff --git a/sys/contrib/openzfs/module/zfs/dnode_sync.c b/sys/contrib/openzfs/module/zfs/dnode_sync.c
index 4067f221f1bf..046ceddb3609 100644
--- a/sys/contrib/openzfs/module/zfs/dnode_sync.c
+++ b/sys/contrib/openzfs/module/zfs/dnode_sync.c
@@ -209,8 +209,8 @@ free_verify(dmu_buf_impl_t *db, uint64_t start, uint64_t end, dmu_tx_t *tx)
rw_exit(&dn->dn_struct_rwlock);
if (err == ENOENT)
continue;
- ASSERT(err == 0);
- ASSERT(child->db_level == 0);
+ ASSERT0(err);
+ ASSERT0(child->db_level);
dr = dbuf_find_dirty_eq(child, txg);
/* data_old better be zeroed */
@@ -868,7 +868,7 @@ dnode_sync(dnode_t *dn, dmu_tx_t *tx)
dbuf_sync_list(list, dn->dn_phys->dn_nlevels - 1, tx);
if (!DMU_OBJECT_IS_SPECIAL(dn->dn_object)) {
- ASSERT3P(list_head(list), ==, NULL);
+ ASSERT0P(list_head(list));
dnode_rele(dn, (void *)(uintptr_t)tx->tx_txg);
}
diff --git a/sys/contrib/openzfs/module/zfs/dsl_bookmark.c b/sys/contrib/openzfs/module/zfs/dsl_bookmark.c
index fdc8b7b198f0..ee574c499f9f 100644
--- a/sys/contrib/openzfs/module/zfs/dsl_bookmark.c
+++ b/sys/contrib/openzfs/module/zfs/dsl_bookmark.c
@@ -243,7 +243,7 @@ dsl_bookmark_create_check_impl(dsl_pool_t *dp,
/* error is retval of the following if-cascade */
if (strchr(source, '@') != NULL) {
dsl_dataset_t *source_snap_ds;
- ASSERT3S(snapshot_namecheck(source, NULL, NULL), ==, 0);
+ ASSERT0(snapshot_namecheck(source, NULL, NULL));
error = dsl_dataset_hold(dp, source, FTAG, &source_snap_ds);
if (error == 0) {
VERIFY(source_snap_ds->ds_is_snapshot);
@@ -258,7 +258,7 @@ dsl_bookmark_create_check_impl(dsl_pool_t *dp,
}
} else if (strchr(source, '#') != NULL) {
zfs_bookmark_phys_t source_phys;
- ASSERT3S(bookmark_namecheck(source, NULL, NULL), ==, 0);
+ ASSERT0(bookmark_namecheck(source, NULL, NULL));
/*
* Source must exists and be an earlier point in newbm_ds's
* timeline (newbm_ds's origin may be a snap of source's ds)
@@ -501,7 +501,7 @@ dsl_bookmark_create_sync_impl_snap(const char *bookmark, const char *snapshot,
sizeof (uint64_t) * num_redact_snaps);
local_rl->rl_phys->rlp_num_snaps = num_redact_snaps;
if (bookmark_redacted) {
- ASSERT3P(redaction_list, ==, NULL);
+ ASSERT0P(redaction_list);
local_rl->rl_phys->rlp_last_blkid = UINT64_MAX;
local_rl->rl_phys->rlp_last_object = UINT64_MAX;
dsl_redaction_list_long_rele(local_rl, tag);
diff --git a/sys/contrib/openzfs/module/zfs/dsl_crypt.c b/sys/contrib/openzfs/module/zfs/dsl_crypt.c
index 6b6bb8d45b6b..f519b937edc0 100644
--- a/sys/contrib/openzfs/module/zfs/dsl_crypt.c
+++ b/sys/contrib/openzfs/module/zfs/dsl_crypt.c
@@ -534,7 +534,7 @@ out:
static void
dsl_crypto_key_free(dsl_crypto_key_t *dck)
{
- ASSERT(zfs_refcount_count(&dck->dck_holds) == 0);
+ ASSERT0(zfs_refcount_count(&dck->dck_holds));
/* destroy the zio_crypt_key_t */
zio_crypt_key_destroy(&dck->dck_key);
@@ -1912,7 +1912,7 @@ dsl_dataset_create_crypt_sync(uint64_t dsobj, dsl_dir_t *dd,
/* clones always use their origin's wrapping key */
if (dsl_dir_is_clone(dd)) {
- ASSERT3P(dcp, ==, NULL);
+ ASSERT0P(dcp);
/*
* If this is an encrypted clone we just need to clone the
diff --git a/sys/contrib/openzfs/module/zfs/dsl_dataset.c b/sys/contrib/openzfs/module/zfs/dsl_dataset.c
index b767c9641419..420687480a76 100644
--- a/sys/contrib/openzfs/module/zfs/dsl_dataset.c
+++ b/sys/contrib/openzfs/module/zfs/dsl_dataset.c
@@ -450,7 +450,7 @@ dsl_dataset_evict_sync(void *dbu)
{
dsl_dataset_t *ds = dbu;
- ASSERT(ds->ds_owner == NULL);
+ ASSERT0P(ds->ds_owner);
unique_remove(ds->ds_fsid_guid);
}
@@ -460,7 +460,7 @@ dsl_dataset_evict_async(void *dbu)
{
dsl_dataset_t *ds = dbu;
- ASSERT(ds->ds_owner == NULL);
+ ASSERT0P(ds->ds_owner);
ds->ds_dbuf = NULL;
@@ -1187,7 +1187,7 @@ dsl_dataset_create_sync_dd(dsl_dir_t *dd, dsl_dataset_t *origin,
ASSERT(origin == NULL || origin->ds_dir->dd_pool == dp);
ASSERT(origin == NULL || dsl_dataset_phys(origin)->ds_num_children > 0);
ASSERT(dmu_tx_is_syncing(tx));
- ASSERT(dsl_dir_phys(dd)->dd_head_dataset_obj == 0);
+ ASSERT0(dsl_dir_phys(dd)->dd_head_dataset_obj);
dsobj = dmu_object_alloc(mos, DMU_OT_DSL_DATASET, 0,
DMU_OT_DSL_DATASET, sizeof (dsl_dataset_phys_t), tx);
@@ -2112,7 +2112,7 @@ dsl_dataset_sync(dsl_dataset_t *ds, zio_t *rio, dmu_tx_t *tx)
{
ASSERT(dmu_tx_is_syncing(tx));
ASSERT(ds->ds_objset != NULL);
- ASSERT(dsl_dataset_phys(ds)->ds_next_snap_obj == 0);
+ ASSERT0(dsl_dataset_phys(ds)->ds_next_snap_obj);
/*
* in case we had to change ds_fsid_guid when we opened it,
@@ -4180,7 +4180,7 @@ dsl_dataset_clone_swap_sync_impl(dsl_dataset_t *clone,
dsl_pool_t *dp = dmu_tx_pool(tx);
int64_t unused_refres_delta;
- ASSERT(clone->ds_reserved == 0);
+ ASSERT0(clone->ds_reserved);
/*
* NOTE: On DEBUG kernels there could be a race between this and
* the check function if spa_asize_inflation is adjusted...
diff --git a/sys/contrib/openzfs/module/zfs/dsl_deadlist.c b/sys/contrib/openzfs/module/zfs/dsl_deadlist.c
index 9ffc998ac173..41ac72bf1c16 100644
--- a/sys/contrib/openzfs/module/zfs/dsl_deadlist.c
+++ b/sys/contrib/openzfs/module/zfs/dsl_deadlist.c
@@ -1037,7 +1037,7 @@ dsl_livelist_iterate(void *arg, const blkptr_t *bp, boolean_t bp_freed,
avl_tree_t *avl = lia->avl;
bplist_t *to_free = lia->to_free;
zthr_t *t = lia->t;
- ASSERT(tx == NULL);
+ ASSERT0P(tx);
if ((t != NULL) && (zthr_has_waiters(t) || zthr_iscancelled(t)))
return (SET_ERROR(EINTR));
@@ -1049,7 +1049,8 @@ dsl_livelist_iterate(void *arg, const blkptr_t *bp, boolean_t bp_freed,
ASSERT3U(BP_GET_PSIZE(bp), ==, BP_GET_PSIZE(&found->le_bp));
ASSERT3U(BP_GET_CHECKSUM(bp), ==,
BP_GET_CHECKSUM(&found->le_bp));
- ASSERT3U(BP_GET_BIRTH(bp), ==, BP_GET_BIRTH(&found->le_bp));
+ ASSERT3U(BP_GET_PHYSICAL_BIRTH(bp), ==,
+ BP_GET_PHYSICAL_BIRTH(&found->le_bp));
}
if (bp_freed) {
if (found == NULL) {
diff --git a/sys/contrib/openzfs/module/zfs/dsl_deleg.c b/sys/contrib/openzfs/module/zfs/dsl_deleg.c
index c01a06e98340..fdd37b36e280 100644
--- a/sys/contrib/openzfs/module/zfs/dsl_deleg.c
+++ b/sys/contrib/openzfs/module/zfs/dsl_deleg.c
@@ -102,7 +102,7 @@ dsl_deleg_can_allow(char *ddname, nvlist_t *nvp, cred_t *cr)
nvlist_t *perms;
nvpair_t *permpair = NULL;
- VERIFY(nvpair_value_nvlist(whopair, &perms) == 0);
+ VERIFY0(nvpair_value_nvlist(whopair, &perms));
while ((permpair = nvlist_next_nvpair(perms, permpair))) {
const char *perm = nvpair_name(permpair);
@@ -189,8 +189,7 @@ dsl_deleg_set_sync(void *arg, dmu_tx_t *tx)
const char *perm = nvpair_name(permpair);
uint64_t n = 0;
- VERIFY(zap_update(mos, jumpobj,
- perm, 8, 1, &n, tx) == 0);
+ VERIFY0(zap_update(mos, jumpobj, perm, 8, 1, &n, tx));
spa_history_log_internal_dd(dd, "permission update", tx,
"%s %s", whokey, perm);
}
@@ -225,7 +224,7 @@ dsl_deleg_unset_sync(void *arg, dmu_tx_t *tx)
if (zap_lookup(mos, zapobj, whokey, 8,
1, &jumpobj) == 0) {
(void) zap_remove(mos, zapobj, whokey, tx);
- VERIFY(0 == zap_destroy(mos, jumpobj, tx));
+ VERIFY0(zap_destroy(mos, jumpobj, tx));
}
spa_history_log_internal_dd(dd, "permission who remove",
tx, "%s", whokey);
@@ -243,7 +242,7 @@ dsl_deleg_unset_sync(void *arg, dmu_tx_t *tx)
if (zap_count(mos, jumpobj, &n) == 0 && n == 0) {
(void) zap_remove(mos, zapobj,
whokey, tx);
- VERIFY(0 == zap_destroy(mos,
+ VERIFY0(zap_destroy(mos,
jumpobj, tx));
}
spa_history_log_internal_dd(dd, "permission remove", tx,
@@ -332,7 +331,7 @@ dsl_deleg_get(const char *ddname, nvlist_t **nvp)
basezc = kmem_alloc(sizeof (zap_cursor_t), KM_SLEEP);
baseza = zap_attribute_alloc();
source = kmem_alloc(ZFS_MAX_DATASET_NAME_LEN, KM_SLEEP);
- VERIFY(nvlist_alloc(nvp, NV_UNIQUE_NAME, KM_SLEEP) == 0);
+ VERIFY0(nvlist_alloc(nvp, NV_UNIQUE_NAME, KM_SLEEP));
for (dd = startdd; dd != NULL; dd = dd->dd_parent) {
nvlist_t *sp_nvp;
@@ -706,7 +705,7 @@ copy_create_perms(dsl_dir_t *dd, uint64_t pzapobj,
ZFS_DELEG_LOCAL, &uid);
if (zap_lookup(mos, zapobj, whokey, 8, 1, &jumpobj) == ENOENT) {
jumpobj = zap_create(mos, DMU_OT_DSL_PERMS, DMU_OT_NONE, 0, tx);
- VERIFY(zap_add(mos, zapobj, whokey, 8, 1, &jumpobj, tx) == 0);
+ VERIFY0(zap_add(mos, zapobj, whokey, 8, 1, &jumpobj, tx));
}
za = zap_attribute_alloc();
@@ -716,8 +715,7 @@ copy_create_perms(dsl_dir_t *dd, uint64_t pzapobj,
uint64_t zero = 0;
ASSERT(za->za_integer_length == 8 && za->za_num_integers == 1);
- VERIFY(zap_update(mos, jumpobj, za->za_name,
- 8, 1, &zero, tx) == 0);
+ VERIFY0(zap_update(mos, jumpobj, za->za_name, 8, 1, &zero, tx));
}
zap_cursor_fini(&zc);
zap_attribute_free(za);
@@ -761,10 +759,10 @@ dsl_deleg_destroy(objset_t *mos, uint64_t zapobj, dmu_tx_t *tx)
zap_cursor_retrieve(&zc, za) == 0;
zap_cursor_advance(&zc)) {
ASSERT(za->za_integer_length == 8 && za->za_num_integers == 1);
- VERIFY(0 == zap_destroy(mos, za->za_first_integer, tx));
+ VERIFY0(zap_destroy(mos, za->za_first_integer, tx));
}
zap_cursor_fini(&zc);
- VERIFY(0 == zap_destroy(mos, zapobj, tx));
+ VERIFY0(zap_destroy(mos, zapobj, tx));
zap_attribute_free(za);
return (0);
}
diff --git a/sys/contrib/openzfs/module/zfs/dsl_destroy.c b/sys/contrib/openzfs/module/zfs/dsl_destroy.c
index fff49c97f4d2..ea01ee586f8b 100644
--- a/sys/contrib/openzfs/module/zfs/dsl_destroy.c
+++ b/sys/contrib/openzfs/module/zfs/dsl_destroy.c
@@ -350,7 +350,7 @@ dsl_destroy_snapshot_sync_impl(dsl_dataset_t *ds, boolean_t defer, dmu_tx_t *tx)
dsl_dataset_deactivate_feature(ds, f, tx);
}
if (dsl_dataset_phys(ds)->ds_prev_snap_obj != 0) {
- ASSERT3P(ds->ds_prev, ==, NULL);
+ ASSERT0P(ds->ds_prev);
VERIFY0(dsl_dataset_hold_obj(dp,
dsl_dataset_phys(ds)->ds_prev_snap_obj, FTAG, &ds_prev));
after_branch_point =
@@ -465,7 +465,7 @@ dsl_destroy_snapshot_sync_impl(dsl_dataset_t *ds, boolean_t defer, dmu_tx_t *tx)
&used, &comp, &uncomp);
dsl_dataset_phys(ds_next)->ds_unique_bytes += used;
dsl_dataset_rele(ds_nextnext, FTAG);
- ASSERT3P(ds_next->ds_prev, ==, NULL);
+ ASSERT0P(ds_next->ds_prev);
/* Collapse range in this head. */
dsl_dataset_t *hds;
@@ -525,7 +525,7 @@ dsl_destroy_snapshot_sync_impl(dsl_dataset_t *ds, boolean_t defer, dmu_tx_t *tx)
/* remove from snapshot namespace */
dsl_dataset_t *ds_head;
- ASSERT(dsl_dataset_phys(ds)->ds_snapnames_zapobj == 0);
+ ASSERT0(dsl_dataset_phys(ds)->ds_snapnames_zapobj);
VERIFY0(dsl_dataset_hold_obj(dp,
dsl_dir_phys(ds->ds_dir)->dd_head_dataset_obj, FTAG, &ds_head));
VERIFY0(dsl_dataset_get_snapname(ds));
@@ -728,7 +728,7 @@ kill_blkptr(spa_t *spa, zilog_t *zilog, const blkptr_t *bp,
*/
dsl_free(ka->tx->tx_pool, ka->tx->tx_txg, bp);
} else {
- ASSERT(zilog == NULL);
+ ASSERT0P(zilog);
ASSERT3U(BP_GET_BIRTH(bp), >,
dsl_dataset_phys(ka->ds)->ds_prev_snap_txg);
(void) dsl_dataset_block_kill(ka->ds, bp, tx, B_FALSE);
diff --git a/sys/contrib/openzfs/module/zfs/dsl_dir.c b/sys/contrib/openzfs/module/zfs/dsl_dir.c
index f24cd2049533..6ce1890cfea1 100644
--- a/sys/contrib/openzfs/module/zfs/dsl_dir.c
+++ b/sys/contrib/openzfs/module/zfs/dsl_dir.c
@@ -151,8 +151,8 @@ dsl_dir_evict_async(void *dbu)
for (t = 0; t < TXG_SIZE; t++) {
ASSERT(!txg_list_member(&dp->dp_dirty_dirs, dd, t));
- ASSERT(dd->dd_tempreserved[t] == 0);
- ASSERT(dd->dd_space_towrite[t] == 0);
+ ASSERT0(dd->dd_tempreserved[t]);
+ ASSERT0(dd->dd_space_towrite[t]);
}
if (dd->dd_parent)
diff --git a/sys/contrib/openzfs/module/zfs/dsl_pool.c b/sys/contrib/openzfs/module/zfs/dsl_pool.c
index 4f1f66b835f2..f47822df8b53 100644
--- a/sys/contrib/openzfs/module/zfs/dsl_pool.c
+++ b/sys/contrib/openzfs/module/zfs/dsl_pool.c
@@ -522,8 +522,8 @@ dsl_pool_create(spa_t *spa, nvlist_t *zplprops __attribute__((unused)),
/* create and open the free_bplist */
obj = bpobj_alloc(dp->dp_meta_objset, SPA_OLD_MAXBLOCKSIZE, tx);
- VERIFY(zap_add(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
- DMU_POOL_FREE_BPOBJ, sizeof (uint64_t), 1, &obj, tx) == 0);
+ VERIFY0(zap_add(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
+ DMU_POOL_FREE_BPOBJ, sizeof (uint64_t), 1, &obj, tx));
VERIFY0(bpobj_open(&dp->dp_free_bpobj,
dp->dp_meta_objset, obj));
}
@@ -1077,7 +1077,7 @@ upgrade_clones_cb(dsl_pool_t *dp, dsl_dataset_t *hds, void *arg)
dsl_dataset_phys(prev)->ds_num_children++;
if (dsl_dataset_phys(ds)->ds_next_snap_obj == 0) {
- ASSERT(ds->ds_prev == NULL);
+ ASSERT0P(ds->ds_prev);
VERIFY0(dsl_dataset_hold_obj(dp,
dsl_dataset_phys(ds)->ds_prev_snap_obj,
ds, &ds->ds_prev));
@@ -1173,7 +1173,7 @@ dsl_pool_create_origin(dsl_pool_t *dp, dmu_tx_t *tx)
dsl_dataset_t *ds;
ASSERT(dmu_tx_is_syncing(tx));
- ASSERT(dp->dp_origin_snap == NULL);
+ ASSERT0P(dp->dp_origin_snap);
ASSERT(rrw_held(&dp->dp_config_rwlock, RW_WRITER));
/* create the origin dir, ds, & snap-ds */
@@ -1250,7 +1250,7 @@ dsl_pool_user_hold_create_obj(dsl_pool_t *dp, dmu_tx_t *tx)
{
objset_t *mos = dp->dp_meta_objset;
- ASSERT(dp->dp_tmp_userrefs_obj == 0);
+ ASSERT0(dp->dp_tmp_userrefs_obj);
ASSERT(dmu_tx_is_syncing(tx));
dp->dp_tmp_userrefs_obj = zap_create_link(mos, DMU_OT_USERREFS,
diff --git a/sys/contrib/openzfs/module/zfs/dsl_prop.c b/sys/contrib/openzfs/module/zfs/dsl_prop.c
index b76f22df61e2..51f624da5689 100644
--- a/sys/contrib/openzfs/module/zfs/dsl_prop.c
+++ b/sys/contrib/openzfs/module/zfs/dsl_prop.c
@@ -815,7 +815,7 @@ dsl_prop_set_sync_impl(dsl_dataset_t *ds, const char *propname,
*/
err = zap_update(mos, zapobj, recvdstr,
intsz, numints, value, tx);
- ASSERT(err == 0);
+ ASSERT0(err);
break;
case (ZPROP_SRC_NONE | ZPROP_SRC_LOCAL | ZPROP_SRC_RECEIVED):
/*
@@ -1166,7 +1166,7 @@ dsl_prop_get_all_impl(objset_t *mos, uint64_t propobj,
if (nvlist_exists(nv, propname))
continue;
- VERIFY(nvlist_alloc(&propval, NV_UNIQUE_NAME, KM_SLEEP) == 0);
+ VERIFY0(nvlist_alloc(&propval, NV_UNIQUE_NAME, KM_SLEEP));
if (za->za_integer_length == 1) {
/*
* String property
@@ -1179,8 +1179,7 @@ dsl_prop_get_all_impl(objset_t *mos, uint64_t propobj,
kmem_free(tmp, za->za_num_integers);
break;
}
- VERIFY(nvlist_add_string(propval, ZPROP_VALUE,
- tmp) == 0);
+ VERIFY0(nvlist_add_string(propval, ZPROP_VALUE, tmp));
kmem_free(tmp, za->za_num_integers);
} else {
/*
@@ -1191,8 +1190,8 @@ dsl_prop_get_all_impl(objset_t *mos, uint64_t propobj,
za->za_first_integer);
}
- VERIFY(nvlist_add_string(propval, ZPROP_SOURCE, source) == 0);
- VERIFY(nvlist_add_nvlist(nv, propname, propval) == 0);
+ VERIFY0(nvlist_add_string(propval, ZPROP_SOURCE, source));
+ VERIFY0(nvlist_add_nvlist(nv, propname, propval));
nvlist_free(propval);
}
zap_cursor_fini(&zc);
@@ -1215,7 +1214,7 @@ dsl_prop_get_all_ds(dsl_dataset_t *ds, nvlist_t **nvp,
int err = 0;
char setpoint[ZFS_MAX_DATASET_NAME_LEN];
- VERIFY(nvlist_alloc(nvp, NV_UNIQUE_NAME, KM_SLEEP) == 0);
+ VERIFY0(nvlist_alloc(nvp, NV_UNIQUE_NAME, KM_SLEEP));
if (ds->ds_is_snapshot)
flags |= DSL_PROP_GET_SNAPSHOT;
@@ -1333,18 +1332,18 @@ dsl_prop_nvlist_add_uint64(nvlist_t *nv, zfs_prop_t prop, uint64_t value)
uint64_t default_value;
if (nvlist_lookup_nvlist(nv, propname, &propval) == 0) {
- VERIFY(nvlist_add_uint64(propval, ZPROP_VALUE, value) == 0);
+ VERIFY0(nvlist_add_uint64(propval, ZPROP_VALUE, value));
return;
}
- VERIFY(nvlist_alloc(&propval, NV_UNIQUE_NAME, KM_SLEEP) == 0);
- VERIFY(nvlist_add_uint64(propval, ZPROP_VALUE, value) == 0);
+ VERIFY0(nvlist_alloc(&propval, NV_UNIQUE_NAME, KM_SLEEP));
+ VERIFY0(nvlist_add_uint64(propval, ZPROP_VALUE, value));
/* Indicate the default source if we can. */
if (dodefault(prop, 8, 1, &default_value) == 0 &&
value == default_value) {
- VERIFY(nvlist_add_string(propval, ZPROP_SOURCE, "") == 0);
+ VERIFY0(nvlist_add_string(propval, ZPROP_SOURCE, ""));
}
- VERIFY(nvlist_add_nvlist(nv, propname, propval) == 0);
+ VERIFY0(nvlist_add_nvlist(nv, propname, propval));
nvlist_free(propval);
}
@@ -1355,13 +1354,13 @@ dsl_prop_nvlist_add_string(nvlist_t *nv, zfs_prop_t prop, const char *value)
const char *propname = zfs_prop_to_name(prop);
if (nvlist_lookup_nvlist(nv, propname, &propval) == 0) {
- VERIFY(nvlist_add_string(propval, ZPROP_VALUE, value) == 0);
+ VERIFY0(nvlist_add_string(propval, ZPROP_VALUE, value));
return;
}
- VERIFY(nvlist_alloc(&propval, NV_UNIQUE_NAME, KM_SLEEP) == 0);
- VERIFY(nvlist_add_string(propval, ZPROP_VALUE, value) == 0);
- VERIFY(nvlist_add_nvlist(nv, propname, propval) == 0);
+ VERIFY0(nvlist_alloc(&propval, NV_UNIQUE_NAME, KM_SLEEP));
+ VERIFY0(nvlist_add_string(propval, ZPROP_VALUE, value));
+ VERIFY0(nvlist_add_nvlist(nv, propname, propval));
nvlist_free(propval);
}
diff --git a/sys/contrib/openzfs/module/zfs/dsl_scan.c b/sys/contrib/openzfs/module/zfs/dsl_scan.c
index 5052992d775c..fcd50c459d07 100644
--- a/sys/contrib/openzfs/module/zfs/dsl_scan.c
+++ b/sys/contrib/openzfs/module/zfs/dsl_scan.c
@@ -1784,7 +1784,7 @@ dsl_scan_zil_block(zilog_t *zilog, const blkptr_t *bp, void *arg,
SET_BOOKMARK(&zb, zh->zh_log.blk_cksum.zc_word[ZIL_ZC_OBJSET],
ZB_ZIL_OBJECT, ZB_ZIL_LEVEL, bp->blk_cksum.zc_word[ZIL_ZC_SEQ]);
- VERIFY(0 == scan_funcs[scn->scn_phys.scn_func](dp, bp, &zb));
+ VERIFY0(scan_funcs[scn->scn_phys.scn_func](dp, bp, &zb));
return (0);
}
@@ -1820,7 +1820,7 @@ dsl_scan_zil_record(zilog_t *zilog, const lr_t *lrc, void *arg,
lr->lr_foid, ZB_ZIL_LEVEL,
lr->lr_offset / BP_GET_LSIZE(bp));
- VERIFY(0 == scan_funcs[scn->scn_phys.scn_func](dp, bp, &zb));
+ VERIFY0(scan_funcs[scn->scn_phys.scn_func](dp, bp, &zb));
}
return (0);
}
@@ -5141,7 +5141,7 @@ dsl_scan_io_queue_vdev_xfer(vdev_t *svd, vdev_t *tvd)
mutex_enter(&svd->vdev_scan_io_queue_lock);
mutex_enter(&tvd->vdev_scan_io_queue_lock);
- VERIFY3P(tvd->vdev_scan_io_queue, ==, NULL);
+ VERIFY0P(tvd->vdev_scan_io_queue);
tvd->vdev_scan_io_queue = svd->vdev_scan_io_queue;
svd->vdev_scan_io_queue = NULL;
if (tvd->vdev_scan_io_queue != NULL)
diff --git a/sys/contrib/openzfs/module/zfs/dsl_userhold.c b/sys/contrib/openzfs/module/zfs/dsl_userhold.c
index 57c70e4ce3d2..f91b7a1eb69a 100644
--- a/sys/contrib/openzfs/module/zfs/dsl_userhold.c
+++ b/sys/contrib/openzfs/module/zfs/dsl_userhold.c
@@ -335,7 +335,7 @@ dsl_dataset_user_hold(nvlist_t *holds, minor_t cleanup_minor, nvlist_t *errlist)
dduha.dduha_holds = holds;
/* chkholds can have non-unique name */
- VERIFY(0 == nvlist_alloc(&dduha.dduha_chkholds, 0, KM_SLEEP));
+ VERIFY0(nvlist_alloc(&dduha.dduha_chkholds, 0, KM_SLEEP));
dduha.dduha_errlist = errlist;
dduha.dduha_minor = cleanup_minor;
diff --git a/sys/contrib/openzfs/module/zfs/fm.c b/sys/contrib/openzfs/module/zfs/fm.c
index a092817efedd..ae788b2310d8 100644
--- a/sys/contrib/openzfs/module/zfs/fm.c
+++ b/sys/contrib/openzfs/module/zfs/fm.c
@@ -337,7 +337,7 @@ zfs_zevent_next(zfs_zevent_t *ze, nvlist_t **event, uint64_t *event_size,
}
}
- VERIFY(nvlist_size(ev->ev_nvl, &size, NV_ENCODE_NATIVE) == 0);
+ VERIFY0(nvlist_size(ev->ev_nvl, &size, NV_ENCODE_NATIVE));
if (size > *event_size) {
*event_size = size;
error = ENOMEM;
diff --git a/sys/contrib/openzfs/module/zfs/metaslab.c b/sys/contrib/openzfs/module/zfs/metaslab.c
index 0e5f09b2724c..9f4399af56bd 100644
--- a/sys/contrib/openzfs/module/zfs/metaslab.c
+++ b/sys/contrib/openzfs/module/zfs/metaslab.c
@@ -391,7 +391,7 @@ static kstat_t *metaslab_ksp;
void
metaslab_stat_init(void)
{
- ASSERT(metaslab_alloc_trace_cache == NULL);
+ ASSERT0P(metaslab_alloc_trace_cache);
metaslab_alloc_trace_cache = kmem_cache_create(
"metaslab_alloc_trace_cache", sizeof (metaslab_alloc_trace_t),
0, NULL, NULL, NULL, NULL, NULL, 0);
@@ -456,16 +456,16 @@ metaslab_class_destroy(metaslab_class_t *mc)
{
spa_t *spa = mc->mc_spa;
- ASSERT(mc->mc_alloc == 0);
- ASSERT(mc->mc_deferred == 0);
- ASSERT(mc->mc_space == 0);
- ASSERT(mc->mc_dspace == 0);
+ ASSERT0(mc->mc_alloc);
+ ASSERT0(mc->mc_deferred);
+ ASSERT0(mc->mc_space);
+ ASSERT0(mc->mc_dspace);
for (int i = 0; i < spa->spa_alloc_count; i++) {
metaslab_class_allocator_t *mca = &mc->mc_allocator[i];
avl_destroy(&mca->mca_tree);
mutex_destroy(&mca->mca_lock);
- ASSERT(mca->mca_rotor == NULL);
+ ASSERT0P(mca->mca_rotor);
ASSERT0(mca->mca_reserved);
}
mutex_destroy(&mc->mc_lock);
@@ -1087,8 +1087,8 @@ metaslab_group_destroy(metaslab_group_t *mg)
{
spa_t *spa = mg->mg_class->mc_spa;
- ASSERT(mg->mg_prev == NULL);
- ASSERT(mg->mg_next == NULL);
+ ASSERT0P(mg->mg_prev);
+ ASSERT0P(mg->mg_next);
/*
* We may have gone below zero with the activation count
* either because we never activated in the first place or
@@ -1118,8 +1118,8 @@ metaslab_group_activate(metaslab_group_t *mg)
ASSERT3U(spa_config_held(spa, SCL_ALLOC, RW_WRITER), !=, 0);
- ASSERT(mg->mg_prev == NULL);
- ASSERT(mg->mg_next == NULL);
+ ASSERT0P(mg->mg_prev);
+ ASSERT0P(mg->mg_next);
ASSERT(mg->mg_activation_count <= 0);
if (++mg->mg_activation_count <= 0)
@@ -1164,8 +1164,8 @@ metaslab_group_passivate(metaslab_group_t *mg)
if (--mg->mg_activation_count != 0) {
for (int i = 0; i < spa->spa_alloc_count; i++)
ASSERT(mc->mc_allocator[i].mca_rotor != mg);
- ASSERT(mg->mg_prev == NULL);
- ASSERT(mg->mg_next == NULL);
+ ASSERT0P(mg->mg_prev);
+ ASSERT0P(mg->mg_next);
ASSERT(mg->mg_activation_count < 0);
return;
}
@@ -1345,7 +1345,7 @@ metaslab_group_histogram_remove(metaslab_group_t *mg, metaslab_t *msp)
static void
metaslab_group_add(metaslab_group_t *mg, metaslab_t *msp)
{
- ASSERT(msp->ms_group == NULL);
+ ASSERT0P(msp->ms_group);
mutex_enter(&mg->mg_lock);
msp->ms_group = mg;
msp->ms_weight = 0;
@@ -3017,7 +3017,7 @@ metaslab_fini(metaslab_t *msp)
metaslab_group_remove(mg, msp);
mutex_enter(&msp->ms_lock);
- VERIFY(msp->ms_group == NULL);
+ VERIFY0P(msp->ms_group);
/*
* If this metaslab hasn't been through metaslab_sync_done() yet its
@@ -5739,7 +5739,7 @@ metaslab_unalloc_dva(spa_t *spa, const dva_t *dva, uint64_t txg)
ASSERT(!vd->vdev_removing);
ASSERT(vdev_is_concrete(vd));
ASSERT0(vd->vdev_indirect_config.vic_mapping_object);
- ASSERT3P(vd->vdev_indirect_mapping, ==, NULL);
+ ASSERT0P(vd->vdev_indirect_mapping);
if (DVA_GET_GANG(dva))
size = vdev_gang_header_asize(vd);
@@ -5997,7 +5997,7 @@ metaslab_alloc_range(spa_t *spa, metaslab_class_t *mc, uint64_t psize,
}
ASSERT(ndvas > 0 && ndvas <= spa_max_replication(spa));
- ASSERT(BP_GET_NDVAS(bp) == 0);
+ ASSERT0(BP_GET_NDVAS(bp));
ASSERT(hintbp == NULL || ndvas <= BP_GET_NDVAS(hintbp));
ASSERT3P(zal, !=, NULL);
@@ -6029,7 +6029,7 @@ metaslab_alloc_range(spa_t *spa, metaslab_class_t *mc, uint64_t psize,
smallest_psize = MIN(cur_psize, smallest_psize);
}
}
- ASSERT(error == 0);
+ ASSERT0(error);
ASSERT(BP_GET_NDVAS(bp) == ndvas);
if (actual_psize)
*actual_psize = smallest_psize;
diff --git a/sys/contrib/openzfs/module/zfs/mmp.c b/sys/contrib/openzfs/module/zfs/mmp.c
index f3665d29b8b4..7db72b9b04b0 100644
--- a/sys/contrib/openzfs/module/zfs/mmp.c
+++ b/sys/contrib/openzfs/module/zfs/mmp.c
@@ -260,7 +260,7 @@ mmp_thread_stop(spa_t *spa)
zfs_dbgmsg("MMP thread stopped pool '%s' gethrtime %llu",
spa_name(spa), gethrtime());
- ASSERT(mmp->mmp_thread == NULL);
+ ASSERT0P(mmp->mmp_thread);
mmp->mmp_thread_exiting = 0;
}
diff --git a/sys/contrib/openzfs/module/zfs/multilist.c b/sys/contrib/openzfs/module/zfs/multilist.c
index 7b85d19e19ee..46fb79269310 100644
--- a/sys/contrib/openzfs/module/zfs/multilist.c
+++ b/sys/contrib/openzfs/module/zfs/multilist.c
@@ -81,7 +81,7 @@ multilist_create_impl(multilist_t *ml, size_t size, size_t offset,
ml->ml_num_sublists = num;
ml->ml_index_func = index_func;
- ml->ml_sublists = kmem_zalloc(sizeof (multilist_sublist_t) *
+ ml->ml_sublists = vmem_zalloc(sizeof (multilist_sublist_t) *
ml->ml_num_sublists, KM_SLEEP);
ASSERT3P(ml->ml_sublists, !=, NULL);
@@ -134,7 +134,7 @@ multilist_destroy(multilist_t *ml)
}
ASSERT3P(ml->ml_sublists, !=, NULL);
- kmem_free(ml->ml_sublists,
+ vmem_free(ml->ml_sublists,
sizeof (multilist_sublist_t) * ml->ml_num_sublists);
ml->ml_num_sublists = 0;
diff --git a/sys/contrib/openzfs/module/zfs/range_tree.c b/sys/contrib/openzfs/module/zfs/range_tree.c
index fc2b17606bd2..ea2d2c7227c8 100644
--- a/sys/contrib/openzfs/module/zfs/range_tree.c
+++ b/sys/contrib/openzfs/module/zfs/range_tree.c
@@ -377,7 +377,7 @@ zfs_range_tree_add_impl(void *arg, uint64_t start, uint64_t size, uint64_t fill)
return;
}
- ASSERT3P(rs, ==, NULL);
+ ASSERT0P(rs);
/*
* Determine whether or not we will have to merge with our neighbors.
@@ -867,7 +867,7 @@ zfs_range_tree_remove_xor_add_segment(uint64_t start, uint64_t end,
next = zfs_btree_next(&removefrom->rt_root, &where, &where);
}
- VERIFY3P(curr, ==, NULL);
+ VERIFY0P(curr);
if (start != end) {
VERIFY3U(start, <, end);
diff --git a/sys/contrib/openzfs/module/zfs/rrwlock.c b/sys/contrib/openzfs/module/zfs/rrwlock.c
index 8ee784619839..d0df39b93560 100644
--- a/sys/contrib/openzfs/module/zfs/rrwlock.c
+++ b/sys/contrib/openzfs/module/zfs/rrwlock.c
@@ -108,7 +108,7 @@ rrn_add(rrwlock_t *rrl, const void *tag)
rn->rn_rrl = rrl;
rn->rn_next = tsd_get(rrw_tsd_key);
rn->rn_tag = tag;
- VERIFY(tsd_set(rrw_tsd_key, rn) == 0);
+ VERIFY0(tsd_set(rrw_tsd_key, rn));
}
/*
@@ -129,7 +129,7 @@ rrn_find_and_remove(rrwlock_t *rrl, const void *tag)
if (prev)
prev->rn_next = rn->rn_next;
else
- VERIFY(tsd_set(rrw_tsd_key, rn->rn_next) == 0);
+ VERIFY0(tsd_set(rrw_tsd_key, rn->rn_next));
kmem_free(rn, sizeof (*rn));
return (B_TRUE);
}
@@ -155,7 +155,7 @@ rrw_destroy(rrwlock_t *rrl)
{
mutex_destroy(&rrl->rr_lock);
cv_destroy(&rrl->rr_cv);
- ASSERT(rrl->rr_writer == NULL);
+ ASSERT0P(rrl->rr_writer);
zfs_refcount_destroy(&rrl->rr_anon_rcount);
zfs_refcount_destroy(&rrl->rr_linked_rcount);
}
@@ -188,7 +188,7 @@ rrw_enter_read_impl(rrwlock_t *rrl, boolean_t prio, const void *tag)
} else {
(void) zfs_refcount_add(&rrl->rr_anon_rcount, tag);
}
- ASSERT(rrl->rr_writer == NULL);
+ ASSERT0P(rrl->rr_writer);
mutex_exit(&rrl->rr_lock);
}
diff --git a/sys/contrib/openzfs/module/zfs/sa.c b/sys/contrib/openzfs/module/zfs/sa.c
index 5db470ce6242..7ad25d4d85ba 100644
--- a/sys/contrib/openzfs/module/zfs/sa.c
+++ b/sys/contrib/openzfs/module/zfs/sa.c
@@ -304,7 +304,7 @@ sa_get_spill(sa_handle_t *hdl)
if (hdl->sa_spill == NULL) {
if ((rc = dmu_spill_hold_existing(hdl->sa_bonus, NULL,
&hdl->sa_spill)) == 0)
- VERIFY(0 == sa_build_index(hdl, SA_SPILL));
+ VERIFY0(sa_build_index(hdl, SA_SPILL));
} else {
rc = 0;
}
@@ -432,7 +432,7 @@ sa_add_layout_entry(objset_t *os, const sa_attr_type_t *attrs, int attr_count,
(void) snprintf(attr_name, sizeof (attr_name),
"%d", (int)lot_num);
- VERIFY(0 == zap_update(os, os->os_sa->sa_layout_attr_obj,
+ VERIFY0(zap_update(os, os->os_sa->sa_layout_attr_obj,
attr_name, 2, attr_count, attrs, tx));
}
@@ -505,7 +505,7 @@ sa_resize_spill(sa_handle_t *hdl, uint32_t size, dmu_tx_t *tx)
}
error = dbuf_spill_set_blksz(hdl->sa_spill, blocksize, tx);
- ASSERT(error == 0);
+ ASSERT0(error);
return (error);
}
@@ -717,7 +717,7 @@ sa_build_layouts(sa_handle_t *hdl, sa_bulk_attr_t *attr_desc, int attr_count,
if (BUF_SPACE_NEEDED(spill_used, spillhdrsize) >
hdl->sa_spill->db_size)
- VERIFY(0 == sa_resize_spill(hdl,
+ VERIFY0(sa_resize_spill(hdl,
BUF_SPACE_NEEDED(spill_used, spillhdrsize), tx));
}
@@ -791,7 +791,7 @@ sa_build_layouts(sa_handle_t *hdl, sa_bulk_attr_t *attr_desc, int attr_count,
hdl->sa_bonus_tab = NULL;
}
if (!sa->sa_force_spill)
- VERIFY(0 == sa_build_index(hdl, SA_BONUS));
+ VERIFY0(sa_build_index(hdl, SA_BONUS));
if (hdl->sa_spill) {
sa_idx_tab_rele(hdl->sa_os, hdl->sa_spill_tab);
if (!spilling) {
@@ -801,10 +801,10 @@ sa_build_layouts(sa_handle_t *hdl, sa_bulk_attr_t *attr_desc, int attr_count,
dmu_buf_rele(hdl->sa_spill, NULL);
hdl->sa_spill = NULL;
hdl->sa_spill_tab = NULL;
- VERIFY(0 == dmu_rm_spill(hdl->sa_os,
+ VERIFY0(dmu_rm_spill(hdl->sa_os,
sa_handle_object(hdl), tx));
} else {
- VERIFY(0 == sa_build_index(hdl, SA_SPILL));
+ VERIFY0(sa_build_index(hdl, SA_SPILL));
}
}
@@ -1733,10 +1733,10 @@ sa_add_projid(sa_handle_t *hdl, dmu_tx_t *tx, uint64_t projid)
NULL, dxattr_obj, dxattr_size);
}
- VERIFY(dmu_set_bonustype(db, DMU_OT_SA, tx) == 0);
- VERIFY(sa_replace_all_by_template_locked(hdl, attrs, count, tx) == 0);
+ VERIFY0(dmu_set_bonustype(db, DMU_OT_SA, tx));
+ VERIFY0(sa_replace_all_by_template_locked(hdl, attrs, count, tx));
if (znode_acl.z_acl_extern_obj) {
- VERIFY(0 == dmu_object_free(zfsvfs->z_os,
+ VERIFY0(dmu_object_free(zfsvfs->z_os,
znode_acl.z_acl_extern_obj, tx));
}
@@ -1858,7 +1858,7 @@ sa_attr_register_sync(sa_handle_t *hdl, dmu_tx_t *tx)
continue;
ATTR_ENCODE(attr_value, tb[i].sa_attr, tb[i].sa_length,
tb[i].sa_byteswap);
- VERIFY(0 == zap_update(hdl->sa_os, sa->sa_reg_attr_obj,
+ VERIFY0(zap_update(hdl->sa_os, sa->sa_reg_attr_obj,
tb[i].sa_name, 8, 1, &attr_value, tx));
tb[i].sa_registered = B_TRUE;
}
@@ -2013,7 +2013,7 @@ sa_modify_attrs(sa_handle_t *hdl, sa_attr_type_t newattr,
* Only a variable-sized attribute can be
* replaced here, and its size must be changing.
*/
- ASSERT3U(reg_length, ==, 0);
+ ASSERT0(reg_length);
ASSERT3U(length, !=, buflen);
SA_ADD_BULK_ATTR(attr_desc, j, attr,
locator, datastart, buflen);
diff --git a/sys/contrib/openzfs/module/zfs/spa.c b/sys/contrib/openzfs/module/zfs/spa.c
index 5ecb175fbd63..b3bb46da263b 100644
--- a/sys/contrib/openzfs/module/zfs/spa.c
+++ b/sys/contrib/openzfs/module/zfs/spa.c
@@ -426,10 +426,10 @@ spa_prop_add_user(nvlist_t *nvl, const char *propname, char *strval,
{
nvlist_t *propval;
- VERIFY(nvlist_alloc(&propval, NV_UNIQUE_NAME, KM_SLEEP) == 0);
- VERIFY(nvlist_add_uint64(propval, ZPROP_SOURCE, src) == 0);
- VERIFY(nvlist_add_string(propval, ZPROP_VALUE, strval) == 0);
- VERIFY(nvlist_add_nvlist(nvl, propname, propval) == 0);
+ VERIFY0(nvlist_alloc(&propval, NV_UNIQUE_NAME, KM_SLEEP));
+ VERIFY0(nvlist_add_uint64(propval, ZPROP_SOURCE, src));
+ VERIFY0(nvlist_add_string(propval, ZPROP_VALUE, strval));
+ VERIFY0(nvlist_add_nvlist(nvl, propname, propval));
nvlist_free(propval);
}
@@ -965,7 +965,7 @@ spa_prop_set(spa_t *spa, nvlist_t *nvp)
uint64_t ver = 0;
if (prop == ZPOOL_PROP_VERSION) {
- VERIFY(nvpair_value_uint64(elem, &ver) == 0);
+ VERIFY0(nvpair_value_uint64(elem, &ver));
} else {
ASSERT(zpool_prop_feature(nvpair_name(elem)));
ver = SPA_VERSION_FEATURES;
@@ -1295,7 +1295,7 @@ spa_taskqs_fini(spa_t *spa, zio_type_t t, zio_taskq_type_t q)
spa_taskqs_t *tqs = &spa->spa_zio_taskq[t][q];
if (tqs->stqs_taskq == NULL) {
- ASSERT3U(tqs->stqs_count, ==, 0);
+ ASSERT0(tqs->stqs_count);
return;
}
@@ -1836,9 +1836,9 @@ static void
spa_deactivate(spa_t *spa)
{
ASSERT(spa->spa_sync_on == B_FALSE);
- ASSERT(spa->spa_dsl_pool == NULL);
- ASSERT(spa->spa_root_vdev == NULL);
- ASSERT(spa->spa_async_zio_root == NULL);
+ ASSERT0P(spa->spa_dsl_pool);
+ ASSERT0P(spa->spa_root_vdev);
+ ASSERT0P(spa->spa_async_zio_root);
ASSERT(spa->spa_state != POOL_STATE_UNINITIALIZED);
spa_evicting_os_wait(spa);
@@ -2021,7 +2021,7 @@ spa_unload_log_sm_flush_all(spa_t *spa)
dmu_tx_t *tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir);
VERIFY0(dmu_tx_assign(tx, DMU_TX_WAIT | DMU_TX_SUSPEND));
- ASSERT3U(spa->spa_log_flushall_txg, ==, 0);
+ ASSERT0(spa->spa_log_flushall_txg);
spa->spa_log_flushall_txg = dmu_tx_get_txg(tx);
dmu_tx_commit(tx);
@@ -2280,7 +2280,7 @@ spa_unload(spa_t *spa)
*/
if (spa->spa_root_vdev)
vdev_free(spa->spa_root_vdev);
- ASSERT(spa->spa_root_vdev == NULL);
+ ASSERT0P(spa->spa_root_vdev);
/*
* Close the dsl pool.
@@ -2418,8 +2418,8 @@ spa_load_spares(spa_t *spa)
spa->spa_spares.sav_vdevs = kmem_zalloc(nspares * sizeof (void *),
KM_SLEEP);
for (i = 0; i < spa->spa_spares.sav_count; i++) {
- VERIFY(spa_config_parse(spa, &vd, spares[i], NULL, 0,
- VDEV_ALLOC_SPARE) == 0);
+ VERIFY0(spa_config_parse(spa, &vd, spares[i], NULL, 0,
+ VDEV_ALLOC_SPARE));
ASSERT(vd != NULL);
spa->spa_spares.sav_vdevs[i] = vd;
@@ -2546,8 +2546,8 @@ spa_load_l2cache(spa_t *spa)
/*
* Create new vdev
*/
- VERIFY(spa_config_parse(spa, &vd, l2cache[i], NULL, 0,
- VDEV_ALLOC_L2CACHE) == 0);
+ VERIFY0(spa_config_parse(spa, &vd, l2cache[i], NULL, 0,
+ VDEV_ALLOC_L2CACHE));
ASSERT(vd != NULL);
newvdevs[i] = vd;
@@ -2799,7 +2799,7 @@ spa_passivate_log(spa_t *spa)
vdev_t *tvd = rvd->vdev_child[c];
if (tvd->vdev_islog) {
- ASSERT3P(tvd->vdev_log_mg, ==, NULL);
+ ASSERT0P(tvd->vdev_log_mg);
metaslab_group_passivate(tvd->vdev_mg);
slog_found = B_TRUE;
}
@@ -2822,7 +2822,7 @@ spa_activate_log(spa_t *spa)
vdev_t *tvd = rvd->vdev_child[c];
if (tvd->vdev_islog) {
- ASSERT3P(tvd->vdev_log_mg, ==, NULL);
+ ASSERT0P(tvd->vdev_log_mg);
metaslab_group_activate(tvd->vdev_mg);
}
}
@@ -3259,7 +3259,7 @@ spa_livelist_delete_cb(void *arg, zthr_t *z)
static void
spa_start_livelist_destroy_thread(spa_t *spa)
{
- ASSERT3P(spa->spa_livelist_delete_zthr, ==, NULL);
+ ASSERT0P(spa->spa_livelist_delete_zthr);
spa->spa_livelist_delete_zthr =
zthr_create("z_livelist_destroy",
spa_livelist_delete_cb_check, spa_livelist_delete_cb, spa,
@@ -3275,7 +3275,7 @@ static int
livelist_track_new_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed,
dmu_tx_t *tx)
{
- ASSERT(tx == NULL);
+ ASSERT0P(tx);
livelist_new_arg_t *lna = arg;
if (bp_freed) {
bplist_append(lna->frees, bp);
@@ -3469,7 +3469,7 @@ spa_start_livelist_condensing_thread(spa_t *spa)
spa->spa_to_condense.syncing = B_FALSE;
spa->spa_to_condense.cancelled = B_FALSE;
- ASSERT3P(spa->spa_livelist_condense_zthr, ==, NULL);
+ ASSERT0P(spa->spa_livelist_condense_zthr);
spa->spa_livelist_condense_zthr =
zthr_create("z_livelist_condense",
spa_livelist_condense_cb_check,
@@ -3486,7 +3486,7 @@ spa_spawn_aux_threads(spa_t *spa)
spa_start_livelist_destroy_thread(spa);
spa_start_livelist_condensing_thread(spa);
- ASSERT3P(spa->spa_checkpoint_discard_zthr, ==, NULL);
+ ASSERT0P(spa->spa_checkpoint_discard_zthr);
spa->spa_checkpoint_discard_zthr =
zthr_create("z_checkpoint_discard",
spa_checkpoint_discard_thread_check,
@@ -4091,11 +4091,11 @@ spa_ld_parse_config(spa_t *spa, spa_import_type_t type)
nvlist_free(spa->spa_load_info);
spa->spa_load_info = fnvlist_alloc();
- ASSERT(spa->spa_comment == NULL);
+ ASSERT0P(spa->spa_comment);
if (nvlist_lookup_string(config, ZPOOL_CONFIG_COMMENT, &comment) == 0)
spa->spa_comment = spa_strdup(comment);
- ASSERT(spa->spa_compatibility == NULL);
+ ASSERT0P(spa->spa_compatibility);
if (nvlist_lookup_string(config, ZPOOL_CONFIG_COMPATIBILITY,
&compatibility) == 0)
spa->spa_compatibility = spa_strdup(compatibility);
@@ -5913,7 +5913,7 @@ spa_load_best(spa_t *spa, spa_load_state_t state, uint64_t max_request,
nvlist_free(config);
if (state == SPA_LOAD_RECOVER) {
- ASSERT3P(loadinfo, ==, NULL);
+ ASSERT0P(loadinfo);
spa_import_progress_remove(spa_guid(spa));
return (rewind_error);
} else {
@@ -9091,7 +9091,7 @@ spa_vdev_setfru(spa_t *spa, uint64_t guid, const char *newfru)
int
spa_scrub_pause_resume(spa_t *spa, pool_scrub_cmd_t cmd)
{
- ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == 0);
+ ASSERT0(spa_config_held(spa, SCL_ALL, RW_WRITER));
if (dsl_scan_resilvering(spa->spa_dsl_pool))
return (SET_ERROR(EBUSY));
@@ -9102,7 +9102,7 @@ spa_scrub_pause_resume(spa_t *spa, pool_scrub_cmd_t cmd)
int
spa_scan_stop(spa_t *spa)
{
- ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == 0);
+ ASSERT0(spa_config_held(spa, SCL_ALL, RW_WRITER));
if (dsl_scan_resilvering(spa->spa_dsl_pool))
return (SET_ERROR(EBUSY));
@@ -9119,7 +9119,7 @@ int
spa_scan_range(spa_t *spa, pool_scan_func_t func, uint64_t txgstart,
uint64_t txgend)
{
- ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == 0);
+ ASSERT0(spa_config_held(spa, SCL_ALL, RW_WRITER));
if (func >= POOL_SCAN_FUNCS || func == POOL_SCAN_NONE)
return (SET_ERROR(ENOTSUP));
@@ -9548,7 +9548,7 @@ spa_sync_frees(spa_t *spa, bplist_t *bpl, dmu_tx_t *tx)
{
zio_t *zio = zio_root(spa, NULL, NULL, 0);
bplist_iterate(bpl, spa_free_sync_cb, zio, tx);
- VERIFY(zio_wait(zio) == 0);
+ VERIFY0(zio_wait(zio));
}
/*
@@ -9587,7 +9587,7 @@ spa_sync_nvlist(spa_t *spa, uint64_t obj, nvlist_t *nv, dmu_tx_t *tx)
size_t nvsize = 0;
dmu_buf_t *db;
- VERIFY(nvlist_size(nv, &nvsize, NV_ENCODE_XDR) == 0);
+ VERIFY0(nvlist_size(nv, &nvsize, NV_ENCODE_XDR));
/*
* Write full (SPA_CONFIG_BLOCKSIZE) blocks of configuration
@@ -9597,15 +9597,15 @@ spa_sync_nvlist(spa_t *spa, uint64_t obj, nvlist_t *nv, dmu_tx_t *tx)
bufsize = P2ROUNDUP((uint64_t)nvsize, SPA_CONFIG_BLOCKSIZE);
packed = vmem_alloc(bufsize, KM_SLEEP);
- VERIFY(nvlist_pack(nv, &packed, &nvsize, NV_ENCODE_XDR,
- KM_SLEEP) == 0);
+ VERIFY0(nvlist_pack(nv, &packed, &nvsize, NV_ENCODE_XDR,
+ KM_SLEEP));
memset(packed + nvsize, 0, bufsize - nvsize);
dmu_write(spa->spa_meta_objset, obj, 0, bufsize, packed, tx);
vmem_free(packed, bufsize);
- VERIFY(0 == dmu_bonus_hold(spa->spa_meta_objset, obj, FTAG, &db));
+ VERIFY0(dmu_bonus_hold(spa->spa_meta_objset, obj, FTAG, &db));
dmu_buf_will_dirty(db, tx);
*(uint64_t *)db->db_data = nvsize;
dmu_buf_rele(db, FTAG);
@@ -10541,7 +10541,7 @@ spa_sync_tq_create(spa_t *spa, const char *name)
{
kthread_t **kthreads;
- ASSERT(spa->spa_sync_tq == NULL);
+ ASSERT0P(spa->spa_sync_tq);
ASSERT3S(spa->spa_alloc_count, <=, boot_ncpus);
/*
diff --git a/sys/contrib/openzfs/module/zfs/spa_config.c b/sys/contrib/openzfs/module/zfs/spa_config.c
index 7d4d06659146..cf28955b0c50 100644
--- a/sys/contrib/openzfs/module/zfs/spa_config.c
+++ b/sys/contrib/openzfs/module/zfs/spa_config.c
@@ -48,18 +48,17 @@
/*
* Pool configuration repository.
*
- * Pool configuration is stored as a packed nvlist on the filesystem. By
- * default, all pools are stored in /etc/zfs/zpool.cache and loaded on boot
- * (when the ZFS module is loaded). Pools can also have the 'cachefile'
- * property set that allows them to be stored in an alternate location until
- * the control of external software.
+ * Pool configuration is stored as a packed nvlist on the filesystem. When
+ * pools are imported they are added to the /etc/zfs/zpool.cache file and
+ * removed from it when exported. For each cache file, we have a single nvlist
+ * which holds all the configuration information. Pools can also have the
+ * 'cachefile' property set which allows this config to be stored in an
+ * alternate location under the control of external software.
*
- * For each cache file, we have a single nvlist which holds all the
- * configuration information. When the module loads, we read this information
- * from /etc/zfs/zpool.cache and populate the SPA namespace. This namespace is
- * maintained independently in spa.c. Whenever the namespace is modified, or
- * the configuration of a pool is changed, we call spa_write_cachefile(), which
- * walks through all the active pools and writes the configuration to disk.
+ * The kernel independantly maintains an AVL tree of imported pools. See the
+ * "SPA locking" comment in spa.c. Whenever a pool configuration is modified
+ * we call spa_write_cachefile() which walks through all the active pools and
+ * writes the updated configuration to to /etc/zfs/zpool.cache file.
*/
static uint64_t spa_config_generation = 1;
@@ -69,94 +68,6 @@ static uint64_t spa_config_generation = 1;
* userland pools when doing testing.
*/
char *spa_config_path = (char *)ZPOOL_CACHE;
-#ifdef _KERNEL
-static int zfs_autoimport_disable = B_TRUE;
-#endif
-
-/*
- * Called when the module is first loaded, this routine loads the configuration
- * file into the SPA namespace. It does not actually open or load the pools; it
- * only populates the namespace.
- */
-void
-spa_config_load(void)
-{
- void *buf = NULL;
- nvlist_t *nvlist, *child;
- nvpair_t *nvpair;
- char *pathname;
- zfs_file_t *fp;
- zfs_file_attr_t zfa;
- uint64_t fsize;
- int err;
-
-#ifdef _KERNEL
- if (zfs_autoimport_disable)
- return;
-#endif
-
- /*
- * Open the configuration file.
- */
- pathname = kmem_alloc(MAXPATHLEN, KM_SLEEP);
-
- (void) snprintf(pathname, MAXPATHLEN, "%s", spa_config_path);
-
- err = zfs_file_open(pathname, O_RDONLY, 0, &fp);
-
-#ifdef __FreeBSD__
- if (err)
- err = zfs_file_open(ZPOOL_CACHE_BOOT, O_RDONLY, 0, &fp);
-#endif
- kmem_free(pathname, MAXPATHLEN);
-
- if (err)
- return;
-
- if (zfs_file_getattr(fp, &zfa))
- goto out;
-
- fsize = zfa.zfa_size;
- buf = kmem_alloc(fsize, KM_SLEEP);
-
- /*
- * Read the nvlist from the file.
- */
- if (zfs_file_read(fp, buf, fsize, NULL) < 0)
- goto out;
-
- /*
- * Unpack the nvlist.
- */
- if (nvlist_unpack(buf, fsize, &nvlist, KM_SLEEP) != 0)
- goto out;
-
- /*
- * Iterate over all elements in the nvlist, creating a new spa_t for
- * each one with the specified configuration.
- */
- mutex_enter(&spa_namespace_lock);
- nvpair = NULL;
- while ((nvpair = nvlist_next_nvpair(nvlist, nvpair)) != NULL) {
- if (nvpair_type(nvpair) != DATA_TYPE_NVLIST)
- continue;
-
- child = fnvpair_value_nvlist(nvpair);
-
- if (spa_lookup(nvpair_name(nvpair)) != NULL)
- continue;
- (void) spa_add(nvpair_name(nvpair), child, NULL);
- }
- mutex_exit(&spa_namespace_lock);
-
- nvlist_free(nvlist);
-
-out:
- if (buf != NULL)
- kmem_free(buf, fsize);
-
- zfs_file_close(fp);
-}
static int
spa_config_remove(spa_config_dirent_t *dp)
@@ -623,7 +534,6 @@ spa_config_update(spa_t *spa, int what)
spa_config_update(spa, SPA_CONFIG_UPDATE_VDEVS);
}
-EXPORT_SYMBOL(spa_config_load);
EXPORT_SYMBOL(spa_all_configs);
EXPORT_SYMBOL(spa_config_set);
EXPORT_SYMBOL(spa_config_generate);
@@ -634,8 +544,3 @@ EXPORT_SYMBOL(spa_config_update);
ZFS_MODULE_PARAM(zfs_spa, spa_, config_path, STRING, ZMOD_RD,
"SPA config file (/etc/zfs/zpool.cache)");
#endif
-
-#ifdef _KERNEL
-ZFS_MODULE_PARAM(zfs, zfs_, autoimport_disable, INT, ZMOD_RW,
- "Disable pool import at module load");
-#endif
diff --git a/sys/contrib/openzfs/module/zfs/spa_misc.c b/sys/contrib/openzfs/module/zfs/spa_misc.c
index 2eba8362a166..dceafbc27556 100644
--- a/sys/contrib/openzfs/module/zfs/spa_misc.c
+++ b/sys/contrib/openzfs/module/zfs/spa_misc.c
@@ -471,9 +471,9 @@ spa_config_lock_destroy(spa_t *spa)
spa_config_lock_t *scl = &spa->spa_config_lock[i];
mutex_destroy(&scl->scl_lock);
cv_destroy(&scl->scl_cv);
- ASSERT(scl->scl_writer == NULL);
- ASSERT(scl->scl_write_wanted == 0);
- ASSERT(scl->scl_count == 0);
+ ASSERT0P(scl->scl_writer);
+ ASSERT0(scl->scl_write_wanted);
+ ASSERT0(scl->scl_count);
}
}
@@ -784,24 +784,23 @@ spa_add(const char *name, nvlist_t *config, const char *altroot)
dp->scd_path = altroot ? NULL : spa_strdup(spa_config_path);
list_insert_head(&spa->spa_config_list, dp);
- VERIFY(nvlist_alloc(&spa->spa_load_info, NV_UNIQUE_NAME,
- KM_SLEEP) == 0);
+ VERIFY0(nvlist_alloc(&spa->spa_load_info, NV_UNIQUE_NAME, KM_SLEEP));
if (config != NULL) {
nvlist_t *features;
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_FEATURES_FOR_READ,
&features) == 0) {
- VERIFY(nvlist_dup(features, &spa->spa_label_features,
- 0) == 0);
+ VERIFY0(nvlist_dup(features,
+ &spa->spa_label_features, 0));
}
- VERIFY(nvlist_dup(config, &spa->spa_config, 0) == 0);
+ VERIFY0(nvlist_dup(config, &spa->spa_config, 0));
}
if (spa->spa_label_features == NULL) {
- VERIFY(nvlist_alloc(&spa->spa_label_features, NV_UNIQUE_NAME,
- KM_SLEEP) == 0);
+ VERIFY0(nvlist_alloc(&spa->spa_label_features, NV_UNIQUE_NAME,
+ KM_SLEEP));
}
spa->spa_min_ashift = INT_MAX;
@@ -2549,13 +2548,6 @@ spa_name_compare(const void *a1, const void *a2)
}
void
-spa_boot_init(void *unused)
-{
- (void) unused;
- spa_config_load();
-}
-
-void
spa_init(spa_mode_t mode)
{
mutex_init(&spa_namespace_lock, NULL, MUTEX_DEFAULT, NULL);
@@ -2608,7 +2600,6 @@ spa_init(spa_mode_t mode)
chksum_init();
zpool_prop_init();
zpool_feature_init();
- spa_config_load();
vdev_prop_init();
l2arc_start();
scan_init();
diff --git a/sys/contrib/openzfs/module/zfs/spa_stats.c b/sys/contrib/openzfs/module/zfs/spa_stats.c
index 6d7cabcf766d..2c87122a0aa9 100644
--- a/sys/contrib/openzfs/module/zfs/spa_stats.c
+++ b/sys/contrib/openzfs/module/zfs/spa_stats.c
@@ -718,7 +718,7 @@ spa_mmp_history_set(spa_t *spa, uint64_t mmp_node_id, int io_error,
for (smh = list_tail(&shl->procfs_list.pl_list); smh != NULL;
smh = list_prev(&shl->procfs_list.pl_list, smh)) {
if (smh->mmp_node_id == mmp_node_id) {
- ASSERT(smh->io_error == 0);
+ ASSERT0(smh->io_error);
smh->io_error = io_error;
smh->duration = duration;
error = 0;
diff --git a/sys/contrib/openzfs/module/zfs/space_map.c b/sys/contrib/openzfs/module/zfs/space_map.c
index c429e0edd168..5f24963f2291 100644
--- a/sys/contrib/openzfs/module/zfs/space_map.c
+++ b/sys/contrib/openzfs/module/zfs/space_map.c
@@ -817,7 +817,7 @@ space_map_open(space_map_t **smp, objset_t *os, uint64_t object,
space_map_t *sm;
int error;
- ASSERT(*smp == NULL);
+ ASSERT0P(*smp);
ASSERT(os != NULL);
ASSERT(object != 0);
diff --git a/sys/contrib/openzfs/module/zfs/space_reftree.c b/sys/contrib/openzfs/module/zfs/space_reftree.c
index 9b2d5ed31dc9..889980e08c06 100644
--- a/sys/contrib/openzfs/module/zfs/space_reftree.c
+++ b/sys/contrib/openzfs/module/zfs/space_reftree.c
@@ -149,6 +149,6 @@ space_reftree_generate_map(avl_tree_t *t, zfs_range_tree_t *rt, int64_t minref)
}
}
}
- ASSERT(refcnt == 0);
+ ASSERT0(refcnt);
ASSERT(start == -1ULL);
}
diff --git a/sys/contrib/openzfs/module/zfs/vdev.c b/sys/contrib/openzfs/module/zfs/vdev.c
index 70b14fb9b2c8..9cf35e379000 100644
--- a/sys/contrib/openzfs/module/zfs/vdev.c
+++ b/sys/contrib/openzfs/module/zfs/vdev.c
@@ -554,7 +554,7 @@ vdev_add_child(vdev_t *pvd, vdev_t *cvd)
vdev_t **newchild;
ASSERT(spa_config_held(cvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL);
- ASSERT(cvd->vdev_parent == NULL);
+ ASSERT0P(cvd->vdev_parent);
cvd->vdev_parent = pvd;
@@ -578,7 +578,7 @@ vdev_add_child(vdev_t *pvd, vdev_t *cvd)
pvd->vdev_nonrot &= cvd->vdev_nonrot;
cvd->vdev_top = (pvd->vdev_top ? pvd->vdev_top: cvd);
- ASSERT(cvd->vdev_top->vdev_parent->vdev_parent == NULL);
+ ASSERT0P(cvd->vdev_top->vdev_parent->vdev_parent);
/*
* Walk up all ancestors to update guid sum.
@@ -1101,10 +1101,10 @@ vdev_free(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
- ASSERT3P(vd->vdev_initialize_thread, ==, NULL);
- ASSERT3P(vd->vdev_trim_thread, ==, NULL);
- ASSERT3P(vd->vdev_autotrim_thread, ==, NULL);
- ASSERT3P(vd->vdev_rebuild_thread, ==, NULL);
+ ASSERT0P(vd->vdev_initialize_thread);
+ ASSERT0P(vd->vdev_trim_thread);
+ ASSERT0P(vd->vdev_autotrim_thread);
+ ASSERT0P(vd->vdev_rebuild_thread);
/*
* Scan queues are normally destroyed at the end of a scan. If the
@@ -1133,7 +1133,7 @@ vdev_free(vdev_t *vd)
for (int c = 0; c < vd->vdev_children; c++)
vdev_free(vd->vdev_child[c]);
- ASSERT(vd->vdev_child == NULL);
+ ASSERT0P(vd->vdev_child);
ASSERT(vd->vdev_guid_sum == vd->vdev_guid);
if (vd->vdev_ops->vdev_op_fini != NULL)
@@ -1162,7 +1162,7 @@ vdev_free(vdev_t *vd)
*/
vdev_remove_child(vd->vdev_parent, vd);
- ASSERT(vd->vdev_parent == NULL);
+ ASSERT0P(vd->vdev_parent);
ASSERT(!list_link_active(&vd->vdev_leaf_node));
/*
@@ -1309,9 +1309,9 @@ vdev_top_transfer(vdev_t *svd, vdev_t *tvd)
ASSERT0(tvd->vdev_indirect_config.vic_births_object);
ASSERT0(tvd->vdev_indirect_config.vic_mapping_object);
ASSERT3U(tvd->vdev_indirect_config.vic_prev_indirect_vdev, ==, -1ULL);
- ASSERT3P(tvd->vdev_indirect_mapping, ==, NULL);
- ASSERT3P(tvd->vdev_indirect_births, ==, NULL);
- ASSERT3P(tvd->vdev_obsolete_sm, ==, NULL);
+ ASSERT0P(tvd->vdev_indirect_mapping);
+ ASSERT0P(tvd->vdev_indirect_births);
+ ASSERT0P(tvd->vdev_obsolete_sm);
ASSERT0(tvd->vdev_noalloc);
ASSERT0(tvd->vdev_removing);
ASSERT0(tvd->vdev_rebuilding);
@@ -1464,7 +1464,7 @@ vdev_remove_parent(vdev_t *cvd)
if (cvd == cvd->vdev_top)
vdev_top_transfer(mvd, cvd);
- ASSERT(mvd->vdev_children == 0);
+ ASSERT0(mvd->vdev_children);
vdev_free(mvd);
}
@@ -2134,14 +2134,14 @@ vdev_open(vdev_t *vd)
* faulted, bail out of the open.
*/
if (!vd->vdev_removed && vd->vdev_faulted) {
- ASSERT(vd->vdev_children == 0);
+ ASSERT0(vd->vdev_children);
ASSERT(vd->vdev_label_aux == VDEV_AUX_ERR_EXCEEDED ||
vd->vdev_label_aux == VDEV_AUX_EXTERNAL);
vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED,
vd->vdev_label_aux);
return (SET_ERROR(ENXIO));
} else if (vd->vdev_offline) {
- ASSERT(vd->vdev_children == 0);
+ ASSERT0(vd->vdev_children);
vdev_set_state(vd, B_TRUE, VDEV_STATE_OFFLINE, VDEV_AUX_NONE);
return (SET_ERROR(ENXIO));
}
@@ -2197,7 +2197,7 @@ vdev_open(vdev_t *vd)
* the vdev is accessible. If we're faulted, bail.
*/
if (vd->vdev_faulted) {
- ASSERT(vd->vdev_children == 0);
+ ASSERT0(vd->vdev_children);
ASSERT(vd->vdev_label_aux == VDEV_AUX_ERR_EXCEEDED ||
vd->vdev_label_aux == VDEV_AUX_EXTERNAL);
vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED,
@@ -2206,7 +2206,7 @@ vdev_open(vdev_t *vd)
}
if (vd->vdev_degraded) {
- ASSERT(vd->vdev_children == 0);
+ ASSERT0(vd->vdev_children);
vdev_set_state(vd, B_TRUE, VDEV_STATE_DEGRADED,
VDEV_AUX_ERR_EXCEEDED);
} else {
@@ -3945,7 +3945,7 @@ vdev_load(vdev_t *vd)
if (error == 0 && checkpoint_sm_obj != 0) {
objset_t *mos = spa_meta_objset(vd->vdev_spa);
ASSERT(vd->vdev_asize != 0);
- ASSERT3P(vd->vdev_checkpoint_sm, ==, NULL);
+ ASSERT0P(vd->vdev_checkpoint_sm);
error = space_map_open(&vd->vdev_checkpoint_sm,
mos, checkpoint_sm_obj, 0, vd->vdev_asize,
@@ -3993,7 +3993,7 @@ vdev_load(vdev_t *vd)
if (error == 0 && obsolete_sm_object != 0) {
objset_t *mos = vd->vdev_spa->spa_meta_objset;
ASSERT(vd->vdev_asize != 0);
- ASSERT3P(vd->vdev_obsolete_sm, ==, NULL);
+ ASSERT0P(vd->vdev_obsolete_sm);
if ((error = space_map_open(&vd->vdev_obsolete_sm, mos,
obsolete_sm_object, 0, vd->vdev_asize, 0))) {
@@ -4521,7 +4521,7 @@ top:
/*
* Prevent any future allocations.
*/
- ASSERT3P(tvd->vdev_log_mg, ==, NULL);
+ ASSERT0P(tvd->vdev_log_mg);
metaslab_group_passivate(mg);
(void) spa_vdev_state_exit(spa, vd, 0);
@@ -5194,7 +5194,7 @@ vdev_stat_update(zio_t *zio, uint64_t psize)
int64_t
vdev_deflated_space(vdev_t *vd, int64_t space)
{
- ASSERT((space & (SPA_MINBLOCKSIZE-1)) == 0);
+ ASSERT0((space & (SPA_MINBLOCKSIZE-1)));
ASSERT(vd->vdev_deflate_ratio != 0 || vd->vdev_isl2cache);
return ((space >> SPA_MINBLOCKSHIFT) * vd->vdev_deflate_ratio);
@@ -5286,8 +5286,8 @@ vdev_config_dirty(vdev_t *vd)
if (nvlist_lookup_nvlist_array(sav->sav_config,
ZPOOL_CONFIG_L2CACHE, &aux, &naux) != 0) {
- VERIFY(nvlist_lookup_nvlist_array(sav->sav_config,
- ZPOOL_CONFIG_SPARES, &aux, &naux) == 0);
+ VERIFY0(nvlist_lookup_nvlist_array(sav->sav_config,
+ ZPOOL_CONFIG_SPARES, &aux, &naux));
}
ASSERT(c < naux);
@@ -5675,7 +5675,7 @@ vdev_expand(vdev_t *vd, uint64_t txg)
(vd->vdev_asize >> vd->vdev_ms_shift) > vd->vdev_ms_count &&
vdev_is_concrete(vd)) {
vdev_metaslab_group_create(vd);
- VERIFY(vdev_metaslab_init(vd, txg) == 0);
+ VERIFY0(vdev_metaslab_init(vd, txg));
vdev_config_dirty(vd);
}
}
diff --git a/sys/contrib/openzfs/module/zfs/vdev_draid.c b/sys/contrib/openzfs/module/zfs/vdev_draid.c
index feec5fd3ce17..a05289102af2 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_draid.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_draid.c
@@ -477,7 +477,7 @@ vdev_draid_generate_perms(const draid_map_t *map, uint8_t **permsp)
VERIFY3U(map->dm_children, <=, VDEV_DRAID_MAX_CHILDREN);
VERIFY3U(map->dm_seed, !=, 0);
VERIFY3U(map->dm_nperms, !=, 0);
- VERIFY3P(map->dm_perms, ==, NULL);
+ VERIFY0P(map->dm_perms);
#ifdef _KERNEL
/*
@@ -590,7 +590,7 @@ vdev_draid_psize_to_asize(vdev_t *vd, uint64_t psize, uint64_t txg)
uint64_t asize = (rows * vdc->vdc_groupwidth) << ashift;
ASSERT3U(asize, !=, 0);
- ASSERT3U(asize % (vdc->vdc_groupwidth), ==, 0);
+ ASSERT0(asize % (vdc->vdc_groupwidth));
return (asize);
}
@@ -704,7 +704,7 @@ vdev_draid_map_alloc_scrub(zio_t *zio, uint64_t abd_offset, raidz_row_t *rr)
uint64_t skip_off = 0;
ASSERT3U(zio->io_type, ==, ZIO_TYPE_READ);
- ASSERT3P(rr->rr_abd_empty, ==, NULL);
+ ASSERT0P(rr->rr_abd_empty);
if (rr->rr_nempty > 0) {
rr->rr_abd_empty = abd_alloc_linear(rr->rr_nempty * skip_size,
@@ -793,7 +793,7 @@ vdev_draid_map_alloc_empty(zio_t *zio, raidz_row_t *rr)
uint64_t skip_off = 0;
ASSERT3U(zio->io_type, ==, ZIO_TYPE_READ);
- ASSERT3P(rr->rr_abd_empty, ==, NULL);
+ ASSERT0P(rr->rr_abd_empty);
if (rr->rr_nempty > 0) {
rr->rr_abd_empty = abd_alloc_linear(rr->rr_nempty * skip_size,
@@ -807,7 +807,7 @@ vdev_draid_map_alloc_empty(zio_t *zio, raidz_row_t *rr)
/* empty data column (small read), add a skip sector */
ASSERT3U(skip_size, ==, parity_size);
ASSERT3U(rr->rr_nempty, !=, 0);
- ASSERT3P(rc->rc_abd, ==, NULL);
+ ASSERT0P(rc->rc_abd);
rc->rc_abd = abd_get_offset_size(rr->rr_abd_empty,
skip_off, skip_size);
skip_off += skip_size;
@@ -1623,7 +1623,7 @@ vdev_draid_rebuild_asize(vdev_t *vd, uint64_t start, uint64_t asize,
SPA_MAXBLOCKSIZE);
ASSERT3U(vdev_draid_get_astart(vd, start), ==, start);
- ASSERT3U(asize % (vdc->vdc_groupwidth << ashift), ==, 0);
+ ASSERT0(asize % (vdc->vdc_groupwidth << ashift));
/* Chunks must evenly span all data columns in the group. */
psize = (((psize >> ashift) / ndata) * ndata) << ashift;
@@ -1634,7 +1634,7 @@ vdev_draid_rebuild_asize(vdev_t *vd, uint64_t start, uint64_t asize,
uint64_t left = vdev_draid_group_to_offset(vd, group + 1) - start;
chunk_size = MIN(chunk_size, left);
- ASSERT3U(chunk_size % (vdc->vdc_groupwidth << ashift), ==, 0);
+ ASSERT0(chunk_size % (vdc->vdc_groupwidth << ashift));
ASSERT3U(vdev_draid_offset_to_group(vd, start), ==,
vdev_draid_offset_to_group(vd, start + chunk_size - 1));
@@ -2272,7 +2272,7 @@ vdev_draid_init(spa_t *spa, nvlist_t *nv, void **tsd)
ASSERT3U(vdc->vdc_groupwidth, <=, vdc->vdc_ndisks);
ASSERT3U(vdc->vdc_groupsz, >=, 2 * VDEV_DRAID_ROWHEIGHT);
ASSERT3U(vdc->vdc_devslicesz, >=, VDEV_DRAID_ROWHEIGHT);
- ASSERT3U(vdc->vdc_devslicesz % VDEV_DRAID_ROWHEIGHT, ==, 0);
+ ASSERT0(vdc->vdc_devslicesz % VDEV_DRAID_ROWHEIGHT);
ASSERT3U((vdc->vdc_groupwidth * vdc->vdc_ngroups) %
vdc->vdc_ndisks, ==, 0);
diff --git a/sys/contrib/openzfs/module/zfs/vdev_indirect.c b/sys/contrib/openzfs/module/zfs/vdev_indirect.c
index 9fc71fa0e03e..7538f471e63c 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_indirect.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_indirect.c
@@ -792,7 +792,7 @@ spa_condense_indirect_start_sync(vdev_t *vd, dmu_tx_t *tx)
DMU_POOL_CONDENSING_INDIRECT, sizeof (uint64_t),
sizeof (*scip) / sizeof (uint64_t), scip, tx));
- ASSERT3P(spa->spa_condensing_indirect, ==, NULL);
+ ASSERT0P(spa->spa_condensing_indirect);
spa->spa_condensing_indirect = spa_condensing_indirect_create(spa);
zfs_dbgmsg("starting condense of vdev %llu in txg %llu: "
@@ -882,7 +882,7 @@ spa_condense_fini(spa_t *spa)
void
spa_start_indirect_condensing_thread(spa_t *spa)
{
- ASSERT3P(spa->spa_condense_zthr, ==, NULL);
+ ASSERT0P(spa->spa_condense_zthr);
spa->spa_condense_zthr = zthr_create("z_indirect_condense",
spa_condense_indirect_thread_check,
spa_condense_indirect_thread, spa, minclsyspri);
@@ -1504,7 +1504,7 @@ vdev_indirect_splits_checksum_validate(indirect_vsd_t *iv, zio_t *zio)
is != NULL; is = list_next(&iv->iv_splits, is)) {
ASSERT3P(is->is_good_child->ic_data, !=, NULL);
- ASSERT3P(is->is_good_child->ic_duplicate, ==, NULL);
+ ASSERT0P(is->is_good_child->ic_duplicate);
abd_copy_off(zio->io_abd, is->is_good_child->ic_data,
is->is_split_offset, 0, is->is_size);
diff --git a/sys/contrib/openzfs/module/zfs/vdev_initialize.c b/sys/contrib/openzfs/module/zfs/vdev_initialize.c
index 9243c76e810d..27188c46e561 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_initialize.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_initialize.c
@@ -632,7 +632,7 @@ vdev_initialize(vdev_t *vd)
ASSERT(MUTEX_HELD(&vd->vdev_initialize_lock));
ASSERT(vd->vdev_ops->vdev_op_leaf);
ASSERT(vdev_is_concrete(vd));
- ASSERT3P(vd->vdev_initialize_thread, ==, NULL);
+ ASSERT0P(vd->vdev_initialize_thread);
ASSERT(!vd->vdev_detached);
ASSERT(!vd->vdev_initialize_exit_wanted);
ASSERT(!vd->vdev_top->vdev_removing);
@@ -653,7 +653,7 @@ vdev_uninitialize(vdev_t *vd)
ASSERT(MUTEX_HELD(&vd->vdev_initialize_lock));
ASSERT(vd->vdev_ops->vdev_op_leaf);
ASSERT(vdev_is_concrete(vd));
- ASSERT3P(vd->vdev_initialize_thread, ==, NULL);
+ ASSERT0P(vd->vdev_initialize_thread);
ASSERT(!vd->vdev_detached);
ASSERT(!vd->vdev_initialize_exit_wanted);
ASSERT(!vd->vdev_top->vdev_removing);
@@ -672,7 +672,7 @@ vdev_initialize_stop_wait_impl(vdev_t *vd)
while (vd->vdev_initialize_thread != NULL)
cv_wait(&vd->vdev_initialize_cv, &vd->vdev_initialize_lock);
- ASSERT3P(vd->vdev_initialize_thread, ==, NULL);
+ ASSERT0P(vd->vdev_initialize_thread);
vd->vdev_initialize_exit_wanted = B_FALSE;
}
diff --git a/sys/contrib/openzfs/module/zfs/vdev_label.c b/sys/contrib/openzfs/module/zfs/vdev_label.c
index 6baa6236aac2..c44f654b0261 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_label.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_label.c
@@ -163,7 +163,7 @@ uint64_t
vdev_label_offset(uint64_t psize, int l, uint64_t offset)
{
ASSERT(offset < sizeof (vdev_label_t));
- ASSERT(P2PHASE_TYPED(psize, sizeof (vdev_label_t), uint64_t) == 0);
+ ASSERT0(P2PHASE_TYPED(psize, sizeof (vdev_label_t), uint64_t));
return (offset + l * sizeof (vdev_label_t) + (l < VDEV_LABELS / 2 ?
0 : psize - VDEV_LABELS * sizeof (vdev_label_t)));
@@ -768,12 +768,12 @@ vdev_top_config_generate(spa_t *spa, nvlist_t *config)
}
if (idx) {
- VERIFY(nvlist_add_uint64_array(config, ZPOOL_CONFIG_HOLE_ARRAY,
- array, idx) == 0);
+ VERIFY0(nvlist_add_uint64_array(config,
+ ZPOOL_CONFIG_HOLE_ARRAY, array, idx));
}
- VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_VDEV_CHILDREN,
- rvd->vdev_children) == 0);
+ VERIFY0(nvlist_add_uint64(config, ZPOOL_CONFIG_VDEV_CHILDREN,
+ rvd->vdev_children));
kmem_free(array, rvd->vdev_children * sizeof (uint64_t));
}
@@ -1189,8 +1189,8 @@ vdev_label_init(vdev_t *vd, uint64_t crtxg, vdev_labeltype_t reason)
* vdev uses as described above, and automatically expires if we
* fail.
*/
- VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_CREATE_TXG,
- crtxg) == 0);
+ VERIFY0(nvlist_add_uint64(label, ZPOOL_CONFIG_CREATE_TXG,
+ crtxg));
}
buf = vp->vp_nvlist;
diff --git a/sys/contrib/openzfs/module/zfs/vdev_queue.c b/sys/contrib/openzfs/module/zfs/vdev_queue.c
index aa41f7066036..c12713b107bf 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_queue.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_queue.c
@@ -780,7 +780,7 @@ vdev_queue_aggregate(vdev_queue_t *vq, zio_t *zio)
if (dio->io_flags & ZIO_FLAG_NODATA) {
/* allocate a buffer for a write gap */
ASSERT3U(dio->io_type, ==, ZIO_TYPE_WRITE);
- ASSERT3P(dio->io_abd, ==, NULL);
+ ASSERT0P(dio->io_abd);
abd_gang_add(aio->io_abd,
abd_get_zeros(dio->io_size), B_TRUE);
} else {
diff --git a/sys/contrib/openzfs/module/zfs/vdev_raidz.c b/sys/contrib/openzfs/module/zfs/vdev_raidz.c
index 210cdcab1ecc..b597d6daefde 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_raidz.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_raidz.c
@@ -412,7 +412,7 @@ vdev_raidz_map_free(raidz_map_t *rm)
rm->rm_nphys_cols);
}
- ASSERT3P(rm->rm_lr, ==, NULL);
+ ASSERT0P(rm->rm_lr);
kmem_free(rm, offsetof(raidz_map_t, rm_row[rm->rm_nrows]));
}
@@ -2431,7 +2431,7 @@ raidz_start_skip_writes(zio_t *zio)
vdev_t *cvd = vd->vdev_child[rc->rc_devidx];
if (rc->rc_size != 0)
continue;
- ASSERT3P(rc->rc_abd, ==, NULL);
+ ASSERT0P(rc->rc_abd);
ASSERT3U(rc->rc_offset, <,
cvd->vdev_psize - VDEV_LABEL_END_SIZE);
@@ -3363,7 +3363,7 @@ vdev_raidz_io_done_reconstruct_known_missing(zio_t *zio, raidz_map_t *rm,
* also have been fewer parity errors than parity
* columns or, again, we wouldn't be in this code path.
*/
- ASSERT(parity_untried == 0);
+ ASSERT0(parity_untried);
ASSERT(parity_errors < rr->rr_firstdatacol);
/*
@@ -4743,7 +4743,7 @@ spa_raidz_expand_thread(void *arg, zthr_t *zthr)
void
spa_start_raidz_expansion_thread(spa_t *spa)
{
- ASSERT3P(spa->spa_raidz_expand_zthr, ==, NULL);
+ ASSERT0P(spa->spa_raidz_expand_zthr);
spa->spa_raidz_expand_zthr = zthr_create("raidz_expand",
spa_raidz_expand_thread_check, spa_raidz_expand_thread,
spa, defclsyspri);
diff --git a/sys/contrib/openzfs/module/zfs/vdev_rebuild.c b/sys/contrib/openzfs/module/zfs/vdev_rebuild.c
index cf259788ccf4..47b3b9921abe 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_rebuild.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_rebuild.c
@@ -256,7 +256,7 @@ vdev_rebuild_initiate_sync(void *arg, dmu_tx_t *tx)
"vdev_id=%llu vdev_guid=%llu started",
(u_longlong_t)vd->vdev_id, (u_longlong_t)vd->vdev_guid);
- ASSERT3P(vd->vdev_rebuild_thread, ==, NULL);
+ ASSERT0P(vd->vdev_rebuild_thread);
vd->vdev_rebuild_thread = thread_create(NULL, 0,
vdev_rebuild_thread, vd, 0, &p0, TS_RUN, maxclsyspri);
@@ -413,7 +413,7 @@ vdev_rebuild_reset_sync(void *arg, dmu_tx_t *tx)
mutex_enter(&vd->vdev_rebuild_lock);
ASSERT(vrp->vrp_rebuild_state == VDEV_REBUILD_ACTIVE);
- ASSERT3P(vd->vdev_rebuild_thread, ==, NULL);
+ ASSERT0P(vd->vdev_rebuild_thread);
vrp->vrp_last_offset = 0;
vrp->vrp_min_txg = 0;
diff --git a/sys/contrib/openzfs/module/zfs/vdev_removal.c b/sys/contrib/openzfs/module/zfs/vdev_removal.c
index 3887be4bd548..2f7a739da241 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_removal.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_removal.c
@@ -344,10 +344,10 @@ spa_vdev_remove_aux(nvlist_t *config, const char *name, nvlist_t **dev,
for (int i = 0, j = 0; i < count; i++) {
if (dev[i] == dev_to_remove)
continue;
- VERIFY(nvlist_dup(dev[i], &newdev[j++], KM_SLEEP) == 0);
+ VERIFY0(nvlist_dup(dev[i], &newdev[j++], KM_SLEEP));
}
- VERIFY(nvlist_remove(config, name, DATA_TYPE_NVLIST_ARRAY) == 0);
+ VERIFY0(nvlist_remove(config, name, DATA_TYPE_NVLIST_ARRAY));
fnvlist_add_nvlist_array(config, name, (const nvlist_t * const *)newdev,
count - 1);
@@ -423,7 +423,7 @@ vdev_remove_initiate_sync(void *arg, dmu_tx_t *tx)
svr = spa_vdev_removal_create(vd);
ASSERT(vd->vdev_removing);
- ASSERT3P(vd->vdev_indirect_mapping, ==, NULL);
+ ASSERT0P(vd->vdev_indirect_mapping);
spa_feature_incr(spa, SPA_FEATURE_DEVICE_REMOVAL, tx);
if (spa_feature_is_enabled(spa, SPA_FEATURE_OBSOLETE_COUNTS)) {
@@ -529,7 +529,7 @@ vdev_remove_initiate_sync(void *arg, dmu_tx_t *tx)
* but in any case only when there are outstanding free i/os, which
* there are not).
*/
- ASSERT3P(spa->spa_vdev_removal, ==, NULL);
+ ASSERT0P(spa->spa_vdev_removal);
spa->spa_vdev_removal = svr;
svr->svr_thread = thread_create(NULL, 0,
spa_vdev_remove_thread, spa, 0, &p0, TS_RUN, minclsyspri);
@@ -1362,11 +1362,11 @@ vdev_remove_complete(spa_t *spa)
txg_wait_synced(spa->spa_dsl_pool, 0);
txg = spa_vdev_enter(spa);
vdev_t *vd = vdev_lookup_top(spa, spa->spa_vdev_removal->svr_vdev_id);
- ASSERT3P(vd->vdev_initialize_thread, ==, NULL);
- ASSERT3P(vd->vdev_trim_thread, ==, NULL);
- ASSERT3P(vd->vdev_autotrim_thread, ==, NULL);
+ ASSERT0P(vd->vdev_initialize_thread);
+ ASSERT0P(vd->vdev_trim_thread);
+ ASSERT0P(vd->vdev_autotrim_thread);
vdev_rebuild_stop_wait(vd);
- ASSERT3P(vd->vdev_rebuild_thread, ==, NULL);
+ ASSERT0P(vd->vdev_rebuild_thread);
sysevent_t *ev = spa_event_create(spa, vd, NULL,
ESC_ZFS_VDEV_REMOVE_DEV);
@@ -1868,7 +1868,7 @@ spa_vdev_remove_cancel_sync(void *arg, dmu_tx_t *tx)
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
objset_t *mos = spa->spa_meta_objset;
- ASSERT3P(svr->svr_thread, ==, NULL);
+ ASSERT0P(svr->svr_thread);
spa_feature_decr(spa, SPA_FEATURE_DEVICE_REMOVAL, tx);
@@ -2076,7 +2076,7 @@ spa_vdev_remove_log(vdev_t *vd, uint64_t *txg)
ASSERT(vd->vdev_islog);
ASSERT(vd == vd->vdev_top);
- ASSERT3P(vd->vdev_log_mg, ==, NULL);
+ ASSERT0P(vd->vdev_log_mg);
ASSERT(MUTEX_HELD(&spa_namespace_lock));
/*
@@ -2112,7 +2112,7 @@ spa_vdev_remove_log(vdev_t *vd, uint64_t *txg)
if (error != 0) {
metaslab_group_activate(mg);
- ASSERT3P(vd->vdev_log_mg, ==, NULL);
+ ASSERT0P(vd->vdev_log_mg);
return (error);
}
ASSERT0(vd->vdev_stat.vs_alloc);
diff --git a/sys/contrib/openzfs/module/zfs/vdev_trim.c b/sys/contrib/openzfs/module/zfs/vdev_trim.c
index fc8d5b8e9a8a..eee18b367909 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_trim.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_trim.c
@@ -1010,7 +1010,7 @@ vdev_trim(vdev_t *vd, uint64_t rate, boolean_t partial, boolean_t secure)
ASSERT(MUTEX_HELD(&vd->vdev_trim_lock));
ASSERT(vd->vdev_ops->vdev_op_leaf);
ASSERT(vdev_is_concrete(vd));
- ASSERT3P(vd->vdev_trim_thread, ==, NULL);
+ ASSERT0P(vd->vdev_trim_thread);
ASSERT(!vd->vdev_detached);
ASSERT(!vd->vdev_trim_exit_wanted);
ASSERT(!vd->vdev_top->vdev_removing);
@@ -1032,7 +1032,7 @@ vdev_trim_stop_wait_impl(vdev_t *vd)
while (vd->vdev_trim_thread != NULL)
cv_wait(&vd->vdev_trim_cv, &vd->vdev_trim_lock);
- ASSERT3P(vd->vdev_trim_thread, ==, NULL);
+ ASSERT0P(vd->vdev_trim_thread);
vd->vdev_trim_exit_wanted = B_FALSE;
}
@@ -1539,7 +1539,7 @@ vdev_autotrim_stop_wait(vdev_t *tvd)
cv_wait(&tvd->vdev_autotrim_cv,
&tvd->vdev_autotrim_lock);
- ASSERT3P(tvd->vdev_autotrim_thread, ==, NULL);
+ ASSERT0P(tvd->vdev_autotrim_thread);
tvd->vdev_autotrim_exit_wanted = B_FALSE;
}
mutex_exit(&tvd->vdev_autotrim_lock);
@@ -1712,7 +1712,7 @@ vdev_trim_l2arc(spa_t *spa)
mutex_enter(&vd->vdev_trim_lock);
ASSERT(vd->vdev_ops->vdev_op_leaf);
ASSERT(vdev_is_concrete(vd));
- ASSERT3P(vd->vdev_trim_thread, ==, NULL);
+ ASSERT0P(vd->vdev_trim_thread);
ASSERT(!vd->vdev_detached);
ASSERT(!vd->vdev_trim_exit_wanted);
ASSERT(!vd->vdev_top->vdev_removing);
diff --git a/sys/contrib/openzfs/module/zfs/zap.c b/sys/contrib/openzfs/module/zfs/zap.c
index 0896690c97e3..3e4e997798a3 100644
--- a/sys/contrib/openzfs/module/zfs/zap.c
+++ b/sys/contrib/openzfs/module/zfs/zap.c
@@ -921,7 +921,7 @@ fzap_add_cd(zap_name_t *zn,
ASSERT(RW_LOCK_HELD(&zap->zap_rwlock));
ASSERT(!zap->zap_ismicro);
- ASSERT(fzap_check(zn, integer_size, num_integers) == 0);
+ ASSERT0(fzap_check(zn, integer_size, num_integers));
err = zap_deref_leaf(zap, zn->zn_hash, tx, RW_WRITER, &l);
if (err != 0)
@@ -1386,7 +1386,7 @@ again:
}
err = zap_entry_read_name(zap, &zeh,
za->za_name_len, za->za_name);
- ASSERT(err == 0);
+ ASSERT0(err);
za->za_normalization_conflict =
zap_entry_normalization_conflict(&zeh,
@@ -1546,7 +1546,7 @@ zap_shrink(zap_name_t *zn, zap_leaf_t *l, dmu_tx_t *tx)
boolean_t trunc = B_FALSE;
int err = 0;
- ASSERT3U(zap_leaf_phys(l)->l_hdr.lh_nentries, ==, 0);
+ ASSERT0(zap_leaf_phys(l)->l_hdr.lh_nentries);
ASSERT3U(prefix_len, <=, zap_f_phys(zap)->zap_ptrtbl.zt_shift);
ASSERT(RW_LOCK_HELD(&zap->zap_rwlock));
ASSERT3U(ZAP_HASH_IDX(hash, prefix_len), ==, prefix);
@@ -1564,7 +1564,7 @@ zap_shrink(zap_name_t *zn, zap_leaf_t *l, dmu_tx_t *tx)
uint64_t sl_hash = ZAP_PREFIX_HASH(sl_prefix, prefix_len);
int slbit = prefix & 1;
- ASSERT3U(zap_leaf_phys(l)->l_hdr.lh_nentries, ==, 0);
+ ASSERT0(zap_leaf_phys(l)->l_hdr.lh_nentries);
/*
* Check if there is a sibling by reading ptrtbl ptrs.
diff --git a/sys/contrib/openzfs/module/zfs/zap_micro.c b/sys/contrib/openzfs/module/zfs/zap_micro.c
index 411b1a9db5ab..ea4e3117a8b9 100644
--- a/sys/contrib/openzfs/module/zfs/zap_micro.c
+++ b/sys/contrib/openzfs/module/zfs/zap_micro.c
@@ -346,7 +346,7 @@ zap_name_alloc_uint64(zap_t *zap, const uint64_t *key, int numints)
{
zap_name_t *zn = kmem_cache_alloc(zap_name_cache, KM_SLEEP);
- ASSERT(zap->zap_normflags == 0);
+ ASSERT0(zap->zap_normflags);
zn->zn_zap = zap;
zn->zn_key_intlen = sizeof (*key);
zn->zn_key_orig = zn->zn_key_norm = key;
@@ -1876,7 +1876,7 @@ zap_cursor_serialize(zap_cursor_t *zc)
return (-1ULL);
if (zc->zc_zap == NULL)
return (zc->zc_serialized);
- ASSERT((zc->zc_hash & zap_maxcd(zc->zc_zap)) == 0);
+ ASSERT0((zc->zc_hash & zap_maxcd(zc->zc_zap)));
ASSERT(zc->zc_cd < zap_maxcd(zc->zc_zap));
/*
@@ -1911,7 +1911,7 @@ zap_cursor_retrieve(zap_cursor_t *zc, zap_attribute_t *za)
* we must add to the existing zc_cd, which may already
* be 1 due to the zap_cursor_advance.
*/
- ASSERT(zc->zc_hash == 0);
+ ASSERT0(zc->zc_hash);
hb = zap_hashbits(zc->zc_zap);
zc->zc_hash = zc->zc_serialized << (64 - hb);
zc->zc_cd += zc->zc_serialized >> hb;
diff --git a/sys/contrib/openzfs/module/zfs/zcp.c b/sys/contrib/openzfs/module/zfs/zcp.c
index 9aecf67fd256..c6684f453e95 100644
--- a/sys/contrib/openzfs/module/zfs/zcp.c
+++ b/sys/contrib/openzfs/module/zfs/zcp.c
@@ -765,7 +765,7 @@ zcp_lua_alloc(void *ud, void *ptr, size_t osize, size_t nsize)
return (NULL);
}
(void) memcpy(luabuf, ptr, osize);
- VERIFY3P(zcp_lua_alloc(ud, ptr, osize, 0), ==, NULL);
+ VERIFY0P(zcp_lua_alloc(ud, ptr, osize, 0));
return (luabuf);
}
}
diff --git a/sys/contrib/openzfs/module/zfs/zfeature.c b/sys/contrib/openzfs/module/zfs/zfeature.c
index 7dfe00d42a08..0816ea134bf3 100644
--- a/sys/contrib/openzfs/module/zfs/zfeature.c
+++ b/sys/contrib/openzfs/module/zfs/zfeature.c
@@ -210,8 +210,8 @@ spa_features_check(spa_t *spa, boolean_t for_write,
za->za_name, 1, MAXPATHLEN, buf) == 0)
desc = buf;
- VERIFY(nvlist_add_string(unsup_feat,
- za->za_name, desc) == 0);
+ VERIFY0(nvlist_add_string(unsup_feat,
+ za->za_name, desc));
}
}
}
diff --git a/sys/contrib/openzfs/module/zfs/zfs_fuid.c b/sys/contrib/openzfs/module/zfs/zfs_fuid.c
index 10a6d289fbf8..2af1efe82e62 100644
--- a/sys/contrib/openzfs/module/zfs/zfs_fuid.c
+++ b/sys/contrib/openzfs/module/zfs/zfs_fuid.c
@@ -112,8 +112,7 @@ zfs_fuid_table_load(objset_t *os, uint64_t fuid_obj, avl_tree_t *idx_tree,
uint64_t fuid_size;
ASSERT(fuid_obj != 0);
- VERIFY(0 == dmu_bonus_hold(os, fuid_obj,
- FTAG, &db));
+ VERIFY0(dmu_bonus_hold(os, fuid_obj, FTAG, &db));
fuid_size = *(uint64_t *)db->db_data;
dmu_buf_rele(db, FTAG);
@@ -125,22 +124,21 @@ zfs_fuid_table_load(objset_t *os, uint64_t fuid_obj, avl_tree_t *idx_tree,
int i;
packed = kmem_alloc(fuid_size, KM_SLEEP);
- VERIFY(dmu_read(os, fuid_obj, 0,
- fuid_size, packed, DMU_READ_PREFETCH) == 0);
- VERIFY(nvlist_unpack(packed, fuid_size,
- &nvp, 0) == 0);
- VERIFY(nvlist_lookup_nvlist_array(nvp, FUID_NVP_ARRAY,
- &fuidnvp, &count) == 0);
+ VERIFY0(dmu_read(os, fuid_obj, 0,
+ fuid_size, packed, DMU_READ_PREFETCH));
+ VERIFY0(nvlist_unpack(packed, fuid_size, &nvp, 0));
+ VERIFY0(nvlist_lookup_nvlist_array(nvp, FUID_NVP_ARRAY,
+ &fuidnvp, &count));
for (i = 0; i != count; i++) {
fuid_domain_t *domnode;
const char *domain;
uint64_t idx;
- VERIFY(nvlist_lookup_string(fuidnvp[i], FUID_DOMAIN,
- &domain) == 0);
- VERIFY(nvlist_lookup_uint64(fuidnvp[i], FUID_IDX,
- &idx) == 0);
+ VERIFY0(nvlist_lookup_string(fuidnvp[i], FUID_DOMAIN,
+ &domain));
+ VERIFY0(nvlist_lookup_uint64(fuidnvp[i], FUID_IDX,
+ &idx));
domnode = kmem_alloc(sizeof (fuid_domain_t), KM_SLEEP);
@@ -246,35 +244,33 @@ zfs_fuid_sync(zfsvfs_t *zfsvfs, dmu_tx_t *tx)
&zfsvfs->z_fuid_obj, tx) == 0);
}
- VERIFY(nvlist_alloc(&nvp, NV_UNIQUE_NAME, KM_SLEEP) == 0);
+ VERIFY0(nvlist_alloc(&nvp, NV_UNIQUE_NAME, KM_SLEEP));
numnodes = avl_numnodes(&zfsvfs->z_fuid_idx);
fuids = kmem_alloc(numnodes * sizeof (void *), KM_SLEEP);
for (i = 0, domnode = avl_first(&zfsvfs->z_fuid_domain); domnode; i++,
domnode = AVL_NEXT(&zfsvfs->z_fuid_domain, domnode)) {
- VERIFY(nvlist_alloc(&fuids[i], NV_UNIQUE_NAME, KM_SLEEP) == 0);
- VERIFY(nvlist_add_uint64(fuids[i], FUID_IDX,
- domnode->f_idx) == 0);
- VERIFY(nvlist_add_uint64(fuids[i], FUID_OFFSET, 0) == 0);
- VERIFY(nvlist_add_string(fuids[i], FUID_DOMAIN,
- domnode->f_ksid->kd_name) == 0);
+ VERIFY0(nvlist_alloc(&fuids[i], NV_UNIQUE_NAME, KM_SLEEP));
+ VERIFY0(nvlist_add_uint64(fuids[i], FUID_IDX,
+ domnode->f_idx));
+ VERIFY0(nvlist_add_uint64(fuids[i], FUID_OFFSET, 0));
+ VERIFY0(nvlist_add_string(fuids[i], FUID_DOMAIN,
+ domnode->f_ksid->kd_name));
}
fnvlist_add_nvlist_array(nvp, FUID_NVP_ARRAY,
(const nvlist_t * const *)fuids, numnodes);
for (i = 0; i != numnodes; i++)
nvlist_free(fuids[i]);
kmem_free(fuids, numnodes * sizeof (void *));
- VERIFY(nvlist_size(nvp, &nvsize, NV_ENCODE_XDR) == 0);
+ VERIFY0(nvlist_size(nvp, &nvsize, NV_ENCODE_XDR));
packed = kmem_alloc(nvsize, KM_SLEEP);
- VERIFY(nvlist_pack(nvp, &packed, &nvsize,
- NV_ENCODE_XDR, KM_SLEEP) == 0);
+ VERIFY0(nvlist_pack(nvp, &packed, &nvsize, NV_ENCODE_XDR, KM_SLEEP));
nvlist_free(nvp);
zfsvfs->z_fuid_size = nvsize;
dmu_write(zfsvfs->z_os, zfsvfs->z_fuid_obj, 0,
zfsvfs->z_fuid_size, packed, tx);
kmem_free(packed, zfsvfs->z_fuid_size);
- VERIFY(0 == dmu_bonus_hold(zfsvfs->z_os, zfsvfs->z_fuid_obj,
- FTAG, &db));
+ VERIFY0(dmu_bonus_hold(zfsvfs->z_os, zfsvfs->z_fuid_obj, FTAG, &db));
dmu_buf_will_dirty(db, tx);
*(uint64_t *)db->db_data = zfsvfs->z_fuid_size;
dmu_buf_rele(db, FTAG);
diff --git a/sys/contrib/openzfs/module/zfs/zfs_ioctl.c b/sys/contrib/openzfs/module/zfs/zfs_ioctl.c
index dcb71229f96a..121b966b9864 100644
--- a/sys/contrib/openzfs/module/zfs/zfs_ioctl.c
+++ b/sys/contrib/openzfs/module/zfs/zfs_ioctl.c
@@ -1493,7 +1493,7 @@ zfs_ioc_pool_create(zfs_cmd_t *zc)
goto pool_props_bad;
(void) nvlist_remove_all(props, ZPOOL_HIDDEN_ARGS);
- VERIFY(nvlist_alloc(&zplprops, NV_UNIQUE_NAME, KM_SLEEP) == 0);
+ VERIFY0(nvlist_alloc(&zplprops, NV_UNIQUE_NAME, KM_SLEEP));
error = zfs_fill_zplprops_root(version, rootprops,
zplprops, NULL);
if (error != 0)
@@ -2245,7 +2245,7 @@ nvl_add_zplprop(objset_t *os, nvlist_t *props, zfs_prop_t prop)
*/
if ((error = zfs_get_zplprop(os, prop, &value)) != 0)
return (error);
- VERIFY(nvlist_add_uint64(props, zfs_prop_to_name(prop), value) == 0);
+ VERIFY0(nvlist_add_uint64(props, zfs_prop_to_name(prop), value));
return (0);
}
@@ -2280,7 +2280,7 @@ zfs_ioc_objset_zplprops(zfs_cmd_t *zc)
dmu_objset_type(os) == DMU_OST_ZFS) {
nvlist_t *nv;
- VERIFY(nvlist_alloc(&nv, NV_UNIQUE_NAME, KM_SLEEP) == 0);
+ VERIFY0(nvlist_alloc(&nv, NV_UNIQUE_NAME, KM_SLEEP));
if ((err = nvl_add_zplprop(os, nv, ZFS_PROP_VERSION)) == 0 &&
(err = nvl_add_zplprop(os, nv, ZFS_PROP_NORMALIZE)) == 0 &&
(err = nvl_add_zplprop(os, nv, ZFS_PROP_UTF8ONLY)) == 0 &&
@@ -2483,7 +2483,7 @@ zfs_prop_set_userquota(const char *dsname, nvpair_t *pair)
if (nvpair_type(pair) == DATA_TYPE_NVLIST) {
nvlist_t *attrs;
- VERIFY(nvpair_value_nvlist(pair, &attrs) == 0);
+ VERIFY0(nvpair_value_nvlist(pair, &attrs));
if (nvlist_lookup_nvpair(attrs, ZPROP_VALUE,
&pair) != 0)
return (SET_ERROR(EINVAL));
@@ -2538,9 +2538,8 @@ zfs_prop_set_special(const char *dsname, zprop_source_t source,
if (nvpair_type(pair) == DATA_TYPE_NVLIST) {
nvlist_t *attrs;
- VERIFY(nvpair_value_nvlist(pair, &attrs) == 0);
- VERIFY(nvlist_lookup_nvpair(attrs, ZPROP_VALUE,
- &pair) == 0);
+ VERIFY0(nvpair_value_nvlist(pair, &attrs));
+ VERIFY0(nvlist_lookup_nvpair(attrs, ZPROP_VALUE, &pair));
}
/* all special properties are numeric except for keylocation */
@@ -2932,14 +2931,14 @@ props_skip(nvlist_t *props, nvlist_t *skipped, nvlist_t **newprops)
{
nvpair_t *pair;
- VERIFY(nvlist_alloc(newprops, NV_UNIQUE_NAME, KM_SLEEP) == 0);
+ VERIFY0(nvlist_alloc(newprops, NV_UNIQUE_NAME, KM_SLEEP));
pair = NULL;
while ((pair = nvlist_next_nvpair(props, pair)) != NULL) {
if (nvlist_exists(skipped, nvpair_name(pair)))
continue;
- VERIFY(nvlist_add_nvpair(*newprops, pair) == 0);
+ VERIFY0(nvlist_add_nvpair(*newprops, pair));
}
}
@@ -3064,11 +3063,11 @@ zfs_ioc_inherit_prop(zfs_cmd_t *zc)
switch (type) {
case PROP_TYPE_STRING:
- VERIFY(0 == nvlist_add_string(dummy, propname, ""));
+ VERIFY0(nvlist_add_string(dummy, propname, ""));
break;
case PROP_TYPE_NUMBER:
case PROP_TYPE_INDEX:
- VERIFY(0 == nvlist_add_uint64(dummy, propname, 0));
+ VERIFY0(nvlist_add_uint64(dummy, propname, 0));
break;
default:
err = SET_ERROR(EINVAL);
@@ -3454,14 +3453,14 @@ zfs_fill_zplprops_impl(objset_t *os, uint64_t zplver,
/*
* Put the version in the zplprops
*/
- VERIFY(nvlist_add_uint64(zplprops,
- zfs_prop_to_name(ZFS_PROP_VERSION), zplver) == 0);
+ VERIFY0(nvlist_add_uint64(zplprops,
+ zfs_prop_to_name(ZFS_PROP_VERSION), zplver));
if (norm == ZFS_PROP_UNDEFINED &&
(error = zfs_get_zplprop(os, ZFS_PROP_NORMALIZE, &norm)) != 0)
return (error);
- VERIFY(nvlist_add_uint64(zplprops,
- zfs_prop_to_name(ZFS_PROP_NORMALIZE), norm) == 0);
+ VERIFY0(nvlist_add_uint64(zplprops,
+ zfs_prop_to_name(ZFS_PROP_NORMALIZE), norm));
/*
* If we're normalizing, names must always be valid UTF-8 strings.
@@ -3471,55 +3470,55 @@ zfs_fill_zplprops_impl(objset_t *os, uint64_t zplver,
if (u8 == ZFS_PROP_UNDEFINED &&
(error = zfs_get_zplprop(os, ZFS_PROP_UTF8ONLY, &u8)) != 0)
return (error);
- VERIFY(nvlist_add_uint64(zplprops,
- zfs_prop_to_name(ZFS_PROP_UTF8ONLY), u8) == 0);
+ VERIFY0(nvlist_add_uint64(zplprops,
+ zfs_prop_to_name(ZFS_PROP_UTF8ONLY), u8));
if (sense == ZFS_PROP_UNDEFINED &&
(error = zfs_get_zplprop(os, ZFS_PROP_CASE, &sense)) != 0)
return (error);
- VERIFY(nvlist_add_uint64(zplprops,
- zfs_prop_to_name(ZFS_PROP_CASE), sense) == 0);
+ VERIFY0(nvlist_add_uint64(zplprops,
+ zfs_prop_to_name(ZFS_PROP_CASE), sense));
if (duq == ZFS_PROP_UNDEFINED &&
(error = zfs_get_zplprop(os, ZFS_PROP_DEFAULTUSERQUOTA, &duq)) != 0)
return (error);
- VERIFY(nvlist_add_uint64(zplprops,
- zfs_prop_to_name(ZFS_PROP_DEFAULTUSERQUOTA), duq) == 0);
+ VERIFY0(nvlist_add_uint64(zplprops,
+ zfs_prop_to_name(ZFS_PROP_DEFAULTUSERQUOTA), duq));
if (dgq == ZFS_PROP_UNDEFINED &&
(error = zfs_get_zplprop(os, ZFS_PROP_DEFAULTGROUPQUOTA,
&dgq)) != 0)
return (error);
- VERIFY(nvlist_add_uint64(zplprops,
- zfs_prop_to_name(ZFS_PROP_DEFAULTGROUPQUOTA), dgq) == 0);
+ VERIFY0(nvlist_add_uint64(zplprops,
+ zfs_prop_to_name(ZFS_PROP_DEFAULTGROUPQUOTA), dgq));
if (dpq == ZFS_PROP_UNDEFINED &&
(error = zfs_get_zplprop(os, ZFS_PROP_DEFAULTPROJECTQUOTA,
&dpq)) != 0)
return (error);
- VERIFY(nvlist_add_uint64(zplprops,
- zfs_prop_to_name(ZFS_PROP_DEFAULTPROJECTQUOTA), dpq) == 0);
+ VERIFY0(nvlist_add_uint64(zplprops,
+ zfs_prop_to_name(ZFS_PROP_DEFAULTPROJECTQUOTA), dpq));
if (duoq == ZFS_PROP_UNDEFINED &&
(error = zfs_get_zplprop(os, ZFS_PROP_DEFAULTUSEROBJQUOTA,
&duoq)) != 0)
return (error);
- VERIFY(nvlist_add_uint64(zplprops,
- zfs_prop_to_name(ZFS_PROP_DEFAULTUSEROBJQUOTA), duoq) == 0);
+ VERIFY0(nvlist_add_uint64(zplprops,
+ zfs_prop_to_name(ZFS_PROP_DEFAULTUSEROBJQUOTA), duoq));
if (dgoq == ZFS_PROP_UNDEFINED &&
(error = zfs_get_zplprop(os, ZFS_PROP_DEFAULTGROUPOBJQUOTA,
&dgoq)) != 0)
return (error);
- VERIFY(nvlist_add_uint64(zplprops,
- zfs_prop_to_name(ZFS_PROP_DEFAULTGROUPOBJQUOTA), dgoq) == 0);
+ VERIFY0(nvlist_add_uint64(zplprops,
+ zfs_prop_to_name(ZFS_PROP_DEFAULTGROUPOBJQUOTA), dgoq));
if (dpoq == ZFS_PROP_UNDEFINED &&
(error = zfs_get_zplprop(os, ZFS_PROP_DEFAULTPROJECTOBJQUOTA,
&dpoq)) != 0)
return (error);
- VERIFY(nvlist_add_uint64(zplprops,
- zfs_prop_to_name(ZFS_PROP_DEFAULTPROJECTOBJQUOTA), dpoq) == 0);
+ VERIFY0(nvlist_add_uint64(zplprops,
+ zfs_prop_to_name(ZFS_PROP_DEFAULTPROJECTOBJQUOTA), dpoq));
if (is_ci)
*is_ci = (sense == ZFS_CASE_INSENSITIVE);
@@ -3668,8 +3667,8 @@ zfs_ioc_create(const char *fsname, nvlist_t *innvl, nvlist_t *outnvl)
* file system creation, so go figure them out
* now.
*/
- VERIFY(nvlist_alloc(&zct.zct_zplprops,
- NV_UNIQUE_NAME, KM_SLEEP) == 0);
+ VERIFY0(nvlist_alloc(&zct.zct_zplprops,
+ NV_UNIQUE_NAME, KM_SLEEP));
error = zfs_fill_zplprops(fsname, nvprops,
zct.zct_zplprops, &is_insensitive);
if (error != 0) {
@@ -4916,9 +4915,8 @@ zfs_check_settable(const char *dsname, nvpair_t *pair, cred_t *cr)
* format.
*/
nvlist_t *attrs;
- VERIFY(nvpair_value_nvlist(pair, &attrs) == 0);
- VERIFY(nvlist_lookup_nvpair(attrs, ZPROP_VALUE,
- &pair) == 0);
+ VERIFY0(nvpair_value_nvlist(pair, &attrs));
+ VERIFY0(nvlist_lookup_nvpair(attrs, ZPROP_VALUE, &pair));
}
/*
@@ -5103,7 +5101,7 @@ zfs_check_clearable(const char *dataset, nvlist_t *props, nvlist_t **errlist)
if (props == NULL)
return (0);
- VERIFY(nvlist_alloc(&errors, NV_UNIQUE_NAME, KM_SLEEP) == 0);
+ VERIFY0(nvlist_alloc(&errors, NV_UNIQUE_NAME, KM_SLEEP));
zc = kmem_alloc(sizeof (zfs_cmd_t), KM_SLEEP);
(void) strlcpy(zc->zc_name, dataset, sizeof (zc->zc_name));
@@ -5115,9 +5113,8 @@ zfs_check_clearable(const char *dataset, nvlist_t *props, nvlist_t **errlist)
sizeof (zc->zc_value));
if ((err = zfs_check_settable(dataset, pair, CRED())) != 0 ||
(err = zfs_secpolicy_inherit_prop(zc, NULL, CRED())) != 0) {
- VERIFY(nvlist_remove_nvpair(props, pair) == 0);
- VERIFY(nvlist_add_int32(errors,
- zc->zc_value, err) == 0);
+ VERIFY0(nvlist_remove_nvpair(props, pair));
+ VERIFY0(nvlist_add_int32(errors, zc->zc_value, err));
}
pair = next_pair;
}
@@ -5127,7 +5124,7 @@ zfs_check_clearable(const char *dataset, nvlist_t *props, nvlist_t **errlist)
nvlist_free(errors);
errors = NULL;
} else {
- VERIFY(nvpair_value_int32(pair, &rv) == 0);
+ VERIFY0(nvpair_value_int32(pair, &rv));
}
if (errlist == NULL)
@@ -5144,16 +5141,14 @@ propval_equals(nvpair_t *p1, nvpair_t *p2)
if (nvpair_type(p1) == DATA_TYPE_NVLIST) {
/* dsl_prop_get_all_impl() format */
nvlist_t *attrs;
- VERIFY(nvpair_value_nvlist(p1, &attrs) == 0);
- VERIFY(nvlist_lookup_nvpair(attrs, ZPROP_VALUE,
- &p1) == 0);
+ VERIFY0(nvpair_value_nvlist(p1, &attrs));
+ VERIFY0(nvlist_lookup_nvpair(attrs, ZPROP_VALUE, &p1));
}
if (nvpair_type(p2) == DATA_TYPE_NVLIST) {
nvlist_t *attrs;
- VERIFY(nvpair_value_nvlist(p2, &attrs) == 0);
- VERIFY(nvlist_lookup_nvpair(attrs, ZPROP_VALUE,
- &p2) == 0);
+ VERIFY0(nvpair_value_nvlist(p2, &attrs));
+ VERIFY0(nvlist_lookup_nvpair(attrs, ZPROP_VALUE, &p2));
}
if (nvpair_type(p1) != nvpair_type(p2))
@@ -5162,14 +5157,14 @@ propval_equals(nvpair_t *p1, nvpair_t *p2)
if (nvpair_type(p1) == DATA_TYPE_STRING) {
const char *valstr1, *valstr2;
- VERIFY(nvpair_value_string(p1, &valstr1) == 0);
- VERIFY(nvpair_value_string(p2, &valstr2) == 0);
+ VERIFY0(nvpair_value_string(p1, &valstr1));
+ VERIFY0(nvpair_value_string(p2, &valstr2));
return (strcmp(valstr1, valstr2) == 0);
} else {
uint64_t intval1, intval2;
- VERIFY(nvpair_value_uint64(p1, &intval1) == 0);
- VERIFY(nvpair_value_uint64(p2, &intval2) == 0);
+ VERIFY0(nvpair_value_uint64(p1, &intval1));
+ VERIFY0(nvpair_value_uint64(p2, &intval2));
return (intval1 == intval2);
}
}
@@ -5237,7 +5232,7 @@ extract_delay_props(nvlist_t *props)
};
int i;
- VERIFY(nvlist_alloc(&delayprops, NV_UNIQUE_NAME, KM_SLEEP) == 0);
+ VERIFY0(nvlist_alloc(&delayprops, NV_UNIQUE_NAME, KM_SLEEP));
for (nvp = nvlist_next_nvpair(props, NULL); nvp != NULL;
nvp = nvlist_next_nvpair(props, nvp)) {
@@ -5253,8 +5248,8 @@ extract_delay_props(nvlist_t *props)
}
if (delayable[i] != 0) {
tmp = nvlist_prev_nvpair(props, nvp);
- VERIFY(nvlist_add_nvpair(delayprops, nvp) == 0);
- VERIFY(nvlist_remove_nvpair(props, nvp) == 0);
+ VERIFY0(nvlist_add_nvpair(delayprops, nvp));
+ VERIFY0(nvlist_remove_nvpair(props, nvp));
nvp = tmp;
}
}
@@ -5485,15 +5480,15 @@ zfs_ioc_recv_impl(char *tofs, char *tosnap, const char *origin,
* using ASSERT() will be just like a VERIFY.
*/
if (recv_delayprops != NULL) {
- ASSERT(nvlist_merge(recvprops, recv_delayprops, 0) == 0);
+ ASSERT0(nvlist_merge(recvprops, recv_delayprops, 0));
nvlist_free(recv_delayprops);
}
if (local_delayprops != NULL) {
- ASSERT(nvlist_merge(localprops, local_delayprops, 0) == 0);
+ ASSERT0(nvlist_merge(localprops, local_delayprops, 0));
nvlist_free(local_delayprops);
}
if (inherited_delayprops != NULL) {
- ASSERT(nvlist_merge(localprops, inherited_delayprops, 0) == 0);
+ ASSERT0(nvlist_merge(localprops, inherited_delayprops, 0));
nvlist_free(inherited_delayprops);
}
*read_bytes = off - noff;
@@ -7342,8 +7337,8 @@ zfs_ioctl_register_legacy(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func,
ASSERT3U(ioc, >=, ZFS_IOC_FIRST);
ASSERT3U(ioc, <, ZFS_IOC_LAST);
- ASSERT3P(vec->zvec_legacy_func, ==, NULL);
- ASSERT3P(vec->zvec_func, ==, NULL);
+ ASSERT0P(vec->zvec_legacy_func);
+ ASSERT0P(vec->zvec_func);
vec->zvec_legacy_func = func;
vec->zvec_secpolicy = secpolicy;
@@ -7366,8 +7361,8 @@ zfs_ioctl_register(const char *name, zfs_ioc_t ioc, zfs_ioc_func_t *func,
ASSERT3U(ioc, >=, ZFS_IOC_FIRST);
ASSERT3U(ioc, <, ZFS_IOC_LAST);
- ASSERT3P(vec->zvec_legacy_func, ==, NULL);
- ASSERT3P(vec->zvec_func, ==, NULL);
+ ASSERT0P(vec->zvec_legacy_func);
+ ASSERT0P(vec->zvec_func);
/* if we are logging, the name must be valid */
ASSERT(!allow_log || namecheck != NO_NAME);
@@ -8148,7 +8143,7 @@ zfsdev_ioctl_common(uint_t vecnum, zfs_cmd_t *zc, int flag)
spa_t *spa;
nvlist_t *lognv = NULL;
- ASSERT(vec->zvec_legacy_func == NULL);
+ ASSERT0P(vec->zvec_legacy_func);
/*
* Add the innvl to the lognv before calling the func,
diff --git a/sys/contrib/openzfs/module/zfs/zfs_log.c b/sys/contrib/openzfs/module/zfs/zfs_log.c
index 2f61ecfd9b3b..ea17e049279f 100644
--- a/sys/contrib/openzfs/module/zfs/zfs_log.c
+++ b/sys/contrib/openzfs/module/zfs/zfs_log.c
@@ -620,7 +620,7 @@ zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype,
if (zil_replaying(zilog, tx) || zp->z_unlinked ||
zfs_xattr_owner_unlinked(zp)) {
if (callback != NULL)
- callback(callback_data);
+ callback(callback_data, 0);
return;
}
@@ -663,7 +663,7 @@ zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype,
DMU_KEEP_CACHING);
DB_DNODE_EXIT(db);
if (err != 0) {
- zil_itx_destroy(itx);
+ zil_itx_destroy(itx, 0);
itx = zil_itx_create(txtype, sizeof (*lr));
lr = (lr_write_t *)&itx->itx_lr;
wr_state = WR_NEED_COPY;
diff --git a/sys/contrib/openzfs/module/zfs/zfs_quota.c b/sys/contrib/openzfs/module/zfs/zfs_quota.c
index b8fe512d4f09..2e91ccc27d6d 100644
--- a/sys/contrib/openzfs/module/zfs/zfs_quota.c
+++ b/sys/contrib/openzfs/module/zfs/zfs_quota.c
@@ -374,7 +374,7 @@ zfs_set_userquota(zfsvfs_t *zfsvfs, zfs_userquota_prop_t type,
if (*objp == 0) {
*objp = zap_create(zfsvfs->z_os, DMU_OT_USERGROUP_QUOTA,
DMU_OT_NONE, 0, tx);
- VERIFY(0 == zap_add(zfsvfs->z_os, MASTER_NODE_OBJ,
+ VERIFY0(zap_add(zfsvfs->z_os, MASTER_NODE_OBJ,
zfs_userquota_prop_prefixes[type], 8, 1, objp, tx));
}
mutex_exit(&zfsvfs->z_lock);
@@ -386,7 +386,7 @@ zfs_set_userquota(zfsvfs_t *zfsvfs, zfs_userquota_prop_t type,
} else {
err = zap_update(zfsvfs->z_os, *objp, buf, 8, 1, &quota, tx);
}
- ASSERT(err == 0);
+ ASSERT0(err);
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
dmu_tx_commit(tx);
diff --git a/sys/contrib/openzfs/module/zfs/zfs_rlock.c b/sys/contrib/openzfs/module/zfs/zfs_rlock.c
index 53eb3ef1b66e..4035baff77d6 100644
--- a/sys/contrib/openzfs/module/zfs/zfs_rlock.c
+++ b/sys/contrib/openzfs/module/zfs/zfs_rlock.c
@@ -666,7 +666,7 @@ zfs_rangelock_reduce(zfs_locked_range_t *lr, uint64_t off, uint64_t len)
/* Ensure there are no other locks */
ASSERT3U(avl_numnodes(&rl->rl_tree), ==, 1);
- ASSERT3U(lr->lr_offset, ==, 0);
+ ASSERT0(lr->lr_offset);
ASSERT3U(lr->lr_type, ==, RL_WRITER);
ASSERT(!lr->lr_proxy);
ASSERT3U(lr->lr_length, ==, UINT64_MAX);
diff --git a/sys/contrib/openzfs/module/zfs/zfs_sa.c b/sys/contrib/openzfs/module/zfs/zfs_sa.c
index 59b6ae4e4203..8b4fc6fd7fbd 100644
--- a/sys/contrib/openzfs/module/zfs/zfs_sa.c
+++ b/sys/contrib/openzfs/module/zfs/zfs_sa.c
@@ -169,7 +169,7 @@ zfs_sa_set_scanstamp(znode_t *zp, xvattr_t *xvap, dmu_tx_t *tx)
ASSERT(MUTEX_HELD(&zp->z_lock));
VERIFY((xoap = xva_getxoptattr(xvap)) != NULL);
if (zp->z_is_sa)
- VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_SCANSTAMP(zfsvfs),
+ VERIFY0(sa_update(zp->z_sa_hdl, SA_ZPL_SCANSTAMP(zfsvfs),
&xoap->xoa_av_scanstamp,
sizeof (xoap->xoa_av_scanstamp), tx));
else {
@@ -181,12 +181,12 @@ zfs_sa_set_scanstamp(znode_t *zp, xvattr_t *xvap, dmu_tx_t *tx)
len = sizeof (xoap->xoa_av_scanstamp) +
ZFS_OLD_ZNODE_PHYS_SIZE;
if (len > doi.doi_bonus_size)
- VERIFY(dmu_set_bonus(db, len, tx) == 0);
+ VERIFY0(dmu_set_bonus(db, len, tx));
(void) memcpy((caddr_t)db->db_data + ZFS_OLD_ZNODE_PHYS_SIZE,
xoap->xoa_av_scanstamp, sizeof (xoap->xoa_av_scanstamp));
zp->z_pflags |= ZFS_BONUS_SCANSTAMP;
- VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_FLAGS(zfsvfs),
+ VERIFY0(sa_update(zp->z_sa_hdl, SA_ZPL_FLAGS(zfsvfs),
&zp->z_pflags, sizeof (uint64_t), tx));
}
}
@@ -286,7 +286,7 @@ zfs_sa_set_xattr(znode_t *zp, const char *name, const void *value, size_t vsize)
dmu_tx_commit(tx);
if (logsaxattr && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ error = zil_commit(zilog, 0);
}
out_free:
vmem_free(obj, size);
@@ -427,11 +427,10 @@ zfs_sa_upgrade(sa_handle_t *hdl, dmu_tx_t *tx)
zp->z_pflags &= ~ZFS_BONUS_SCANSTAMP;
}
- VERIFY(dmu_set_bonustype(db, DMU_OT_SA, tx) == 0);
- VERIFY(sa_replace_all_by_template_locked(hdl, sa_attrs,
- count, tx) == 0);
+ VERIFY0(dmu_set_bonustype(db, DMU_OT_SA, tx));
+ VERIFY0(sa_replace_all_by_template_locked(hdl, sa_attrs, count, tx));
if (znode_acl.z_acl_extern_obj)
- VERIFY(0 == dmu_object_free(zfsvfs->z_os,
+ VERIFY0(dmu_object_free(zfsvfs->z_os,
znode_acl.z_acl_extern_obj, tx));
zp->z_is_sa = B_TRUE;
diff --git a/sys/contrib/openzfs/module/zfs/zfs_vnops.c b/sys/contrib/openzfs/module/zfs/zfs_vnops.c
index 74aa91a4f2eb..7bb9ba57c69e 100644
--- a/sys/contrib/openzfs/module/zfs/zfs_vnops.c
+++ b/sys/contrib/openzfs/module/zfs/zfs_vnops.c
@@ -27,6 +27,7 @@
* Copyright 2017 Nexenta Systems, Inc.
* Copyright (c) 2021, 2022 by Pawel Jakub Dawidek
* Copyright (c) 2025, Rob Norris <robn@despairlabs.com>
+ * Copyright (c) 2025, Klara, Inc.
*/
/* Portions Copyright 2007 Jeremy Teo */
@@ -116,7 +117,7 @@ zfs_fsync(znode_t *zp, int syncflag, cred_t *cr)
if (zfsvfs->z_os->os_sync != ZFS_SYNC_DISABLED) {
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
- zil_commit(zfsvfs->z_log, zp->z_id);
+ error = zil_commit(zfsvfs->z_log, zp->z_id);
zfs_exit(zfsvfs, FTAG);
}
return (error);
@@ -375,8 +376,13 @@ zfs_read(struct znode *zp, zfs_uio_t *uio, int ioflag, cred_t *cr)
frsync = !!(ioflag & FRSYNC);
#endif
if (zfsvfs->z_log &&
- (frsync || zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS))
- zil_commit(zfsvfs->z_log, zp->z_id);
+ (frsync || zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)) {
+ error = zil_commit(zfsvfs->z_log, zp->z_id);
+ if (error != 0) {
+ zfs_exit(zfsvfs, FTAG);
+ return (error);
+ }
+ }
/*
* Lock the range against changes.
@@ -1074,8 +1080,13 @@ zfs_write(znode_t *zp, zfs_uio_t *uio, int ioflag, cred_t *cr)
return (error);
}
- if (commit)
- zil_commit(zilog, zp->z_id);
+ if (commit) {
+ error = zil_commit(zilog, zp->z_id);
+ if (error != 0) {
+ zfs_exit(zfsvfs, FTAG);
+ return (error);
+ }
+ }
int64_t nwritten = start_resid - zfs_uio_resid(uio);
dataset_kstats_update_write_kstats(&zfsvfs->z_kstat, nwritten);
@@ -1260,8 +1271,8 @@ zfs_setsecattr(znode_t *zp, vsecattr_t *vsecp, int flag, cred_t *cr)
zilog = zfsvfs->z_log;
error = zfs_setacl(zp, vsecp, skipaclchk, cr);
- if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
- zil_commit(zilog, 0);
+ if (error == 0 && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
+ error = zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
@@ -1946,7 +1957,7 @@ unlock:
ZFS_ACCESSTIME_STAMP(inzfsvfs, inzp);
if (outos->os_sync == ZFS_SYNC_ALWAYS) {
- zil_commit(zilog, outzp->z_id);
+ error = zil_commit(zilog, outzp->z_id);
}
*inoffp += done;
diff --git a/sys/contrib/openzfs/module/zfs/zil.c b/sys/contrib/openzfs/module/zfs/zil.c
index 6e4f84257407..0307df55aa21 100644
--- a/sys/contrib/openzfs/module/zfs/zil.c
+++ b/sys/contrib/openzfs/module/zfs/zil.c
@@ -24,6 +24,7 @@
* Copyright (c) 2011, 2018 by Delphix. All rights reserved.
* Copyright (c) 2014 Integros [integros.com]
* Copyright (c) 2018 Datto Inc.
+ * Copyright (c) 2025, Klara, Inc.
*/
/* Portions Copyright 2010 Robert Milkowski */
@@ -103,6 +104,7 @@ static zil_kstat_values_t zil_stats = {
{ "zil_commit_error_count", KSTAT_DATA_UINT64 },
{ "zil_commit_stall_count", KSTAT_DATA_UINT64 },
{ "zil_commit_suspend_count", KSTAT_DATA_UINT64 },
+ { "zil_commit_crash_count", KSTAT_DATA_UINT64 },
{ "zil_itx_count", KSTAT_DATA_UINT64 },
{ "zil_itx_indirect_count", KSTAT_DATA_UINT64 },
{ "zil_itx_indirect_bytes", KSTAT_DATA_UINT64 },
@@ -145,7 +147,7 @@ static uint64_t zil_slog_bulk = 64 * 1024 * 1024;
static kmem_cache_t *zil_lwb_cache;
static kmem_cache_t *zil_zcw_cache;
-static void zil_lwb_commit(zilog_t *zilog, lwb_t *lwb, itx_t *itx);
+static int zil_lwb_commit(zilog_t *zilog, lwb_t *lwb, itx_t *itx);
static itx_t *zil_itx_clone(itx_t *oitx);
static uint64_t zil_max_waste_space(zilog_t *zilog);
@@ -367,6 +369,7 @@ zil_sums_init(zil_sums_t *zs)
wmsum_init(&zs->zil_commit_error_count, 0);
wmsum_init(&zs->zil_commit_stall_count, 0);
wmsum_init(&zs->zil_commit_suspend_count, 0);
+ wmsum_init(&zs->zil_commit_crash_count, 0);
wmsum_init(&zs->zil_itx_count, 0);
wmsum_init(&zs->zil_itx_indirect_count, 0);
wmsum_init(&zs->zil_itx_indirect_bytes, 0);
@@ -392,6 +395,7 @@ zil_sums_fini(zil_sums_t *zs)
wmsum_fini(&zs->zil_commit_error_count);
wmsum_fini(&zs->zil_commit_stall_count);
wmsum_fini(&zs->zil_commit_suspend_count);
+ wmsum_fini(&zs->zil_commit_crash_count);
wmsum_fini(&zs->zil_itx_count);
wmsum_fini(&zs->zil_itx_indirect_count);
wmsum_fini(&zs->zil_itx_indirect_bytes);
@@ -422,6 +426,8 @@ zil_kstat_values_update(zil_kstat_values_t *zs, zil_sums_t *zil_sums)
wmsum_value(&zil_sums->zil_commit_stall_count);
zs->zil_commit_suspend_count.value.ui64 =
wmsum_value(&zil_sums->zil_commit_suspend_count);
+ zs->zil_commit_crash_count.value.ui64 =
+ wmsum_value(&zil_sums->zil_commit_crash_count);
zs->zil_itx_count.value.ui64 =
wmsum_value(&zil_sums->zil_itx_count);
zs->zil_itx_indirect_count.value.ui64 =
@@ -813,34 +819,37 @@ zil_lwb_vdev_compare(const void *x1, const void *x2)
* we choose them here and later make the block allocation match.
*/
static lwb_t *
-zil_alloc_lwb(zilog_t *zilog, int sz, blkptr_t *bp, boolean_t slog,
- uint64_t txg, lwb_state_t state)
+zil_alloc_lwb(zilog_t *zilog, blkptr_t *bp, int min_sz, int sz,
+ boolean_t slog, uint64_t txg)
{
lwb_t *lwb;
lwb = kmem_cache_alloc(zil_lwb_cache, KM_SLEEP);
+ lwb->lwb_flags = 0;
lwb->lwb_zilog = zilog;
if (bp) {
lwb->lwb_blk = *bp;
- lwb->lwb_slim = (BP_GET_CHECKSUM(bp) == ZIO_CHECKSUM_ZILOG2);
+ if (BP_GET_CHECKSUM(bp) == ZIO_CHECKSUM_ZILOG2)
+ lwb->lwb_flags |= LWB_FLAG_SLIM;
sz = BP_GET_LSIZE(bp);
+ lwb->lwb_min_sz = sz;
} else {
BP_ZERO(&lwb->lwb_blk);
- lwb->lwb_slim = (spa_version(zilog->zl_spa) >=
- SPA_VERSION_SLIM_ZIL);
+ if (spa_version(zilog->zl_spa) >= SPA_VERSION_SLIM_ZIL)
+ lwb->lwb_flags |= LWB_FLAG_SLIM;
+ lwb->lwb_min_sz = min_sz;
}
- lwb->lwb_slog = slog;
+ if (slog)
+ lwb->lwb_flags |= LWB_FLAG_SLOG;
lwb->lwb_error = 0;
- if (lwb->lwb_slim) {
- lwb->lwb_nmax = sz;
- lwb->lwb_nused = lwb->lwb_nfilled = sizeof (zil_chain_t);
- } else {
- lwb->lwb_nmax = sz - sizeof (zil_chain_t);
- lwb->lwb_nused = lwb->lwb_nfilled = 0;
- }
+ /*
+ * Buffer allocation and capacity setup will be done in
+ * zil_lwb_write_open() when the LWB is opened for ITX assignment.
+ */
+ lwb->lwb_nmax = lwb->lwb_nused = lwb->lwb_nfilled = 0;
lwb->lwb_sz = sz;
- lwb->lwb_state = state;
- lwb->lwb_buf = zio_buf_alloc(sz);
+ lwb->lwb_buf = NULL;
+ lwb->lwb_state = LWB_STATE_NEW;
lwb->lwb_child_zio = NULL;
lwb->lwb_write_zio = NULL;
lwb->lwb_root_zio = NULL;
@@ -851,8 +860,6 @@ zil_alloc_lwb(zilog_t *zilog, int sz, blkptr_t *bp, boolean_t slog,
mutex_enter(&zilog->zl_lock);
list_insert_tail(&zilog->zl_lwb_list, lwb);
- if (state != LWB_STATE_NEW)
- zilog->zl_last_lwb_opened = lwb;
mutex_exit(&zilog->zl_lock);
return (lwb);
@@ -864,15 +871,15 @@ zil_free_lwb(zilog_t *zilog, lwb_t *lwb)
ASSERT(MUTEX_HELD(&zilog->zl_lock));
ASSERT(lwb->lwb_state == LWB_STATE_NEW ||
lwb->lwb_state == LWB_STATE_FLUSH_DONE);
- ASSERT3P(lwb->lwb_child_zio, ==, NULL);
- ASSERT3P(lwb->lwb_write_zio, ==, NULL);
- ASSERT3P(lwb->lwb_root_zio, ==, NULL);
+ ASSERT0P(lwb->lwb_child_zio);
+ ASSERT0P(lwb->lwb_write_zio);
+ ASSERT0P(lwb->lwb_root_zio);
ASSERT3U(lwb->lwb_alloc_txg, <=, spa_syncing_txg(zilog->zl_spa));
ASSERT3U(lwb->lwb_max_txg, <=, spa_syncing_txg(zilog->zl_spa));
VERIFY(list_is_empty(&lwb->lwb_itxs));
VERIFY(list_is_empty(&lwb->lwb_waiters));
ASSERT(avl_is_empty(&lwb->lwb_vdev_tree));
- ASSERT(!MUTEX_HELD(&lwb->lwb_vdev_lock));
+ ASSERT(!MUTEX_HELD(&lwb->lwb_lock));
/*
* Clear the zilog's field to indicate this lwb is no longer
@@ -991,8 +998,8 @@ zil_create(zilog_t *zilog)
*/
txg_wait_synced(zilog->zl_dmu_pool, zilog->zl_destroy_txg);
- ASSERT(zh->zh_claim_txg == 0);
- ASSERT(zh->zh_replay_seq == 0);
+ ASSERT0(zh->zh_claim_txg);
+ ASSERT0(zh->zh_replay_seq);
blk = zh->zh_log;
@@ -1013,7 +1020,7 @@ zil_create(zilog_t *zilog)
}
error = zio_alloc_zil(zilog->zl_spa, zilog->zl_os, txg, &blk,
- ZIL_MIN_BLKSZ, &slog);
+ ZIL_MIN_BLKSZ, ZIL_MIN_BLKSZ, &slog, B_TRUE);
if (error == 0)
zil_init_log_chain(zilog, &blk);
}
@@ -1022,7 +1029,7 @@ zil_create(zilog_t *zilog)
* Allocate a log write block (lwb) for the first log block.
*/
if (error == 0)
- lwb = zil_alloc_lwb(zilog, 0, &blk, slog, txg, LWB_STATE_NEW);
+ lwb = zil_alloc_lwb(zilog, &blk, 0, 0, slog, txg);
/*
* If we just allocated the first log block, commit our transaction
@@ -1104,7 +1111,7 @@ zil_destroy(zilog_t *zilog, boolean_t keep_first)
zilog->zl_keep_first = keep_first;
if (!list_is_empty(&zilog->zl_lwb_list)) {
- ASSERT(zh->zh_claim_txg == 0);
+ ASSERT0(zh->zh_claim_txg);
VERIFY(!keep_first);
while ((lwb = list_remove_head(&zilog->zl_lwb_list)) != NULL) {
if (lwb->lwb_buf != NULL)
@@ -1250,7 +1257,7 @@ zil_check_log_chain(dsl_pool_t *dp, dsl_dataset_t *ds, void *tx)
blkptr_t *bp;
int error;
- ASSERT(tx == NULL);
+ ASSERT0P(tx);
error = dmu_objset_from_ds(ds, &os);
if (error != 0) {
@@ -1318,10 +1325,12 @@ zil_check_log_chain(dsl_pool_t *dp, dsl_dataset_t *ds, void *tx)
* zil_commit() is racing with spa_sync().
*/
static void
-zil_commit_waiter_skip(zil_commit_waiter_t *zcw)
+zil_commit_waiter_done(zil_commit_waiter_t *zcw, int err)
{
mutex_enter(&zcw->zcw_lock);
ASSERT3B(zcw->zcw_done, ==, B_FALSE);
+ zcw->zcw_lwb = NULL;
+ zcw->zcw_error = err;
zcw->zcw_done = B_TRUE;
cv_broadcast(&zcw->zcw_cv);
mutex_exit(&zcw->zcw_lock);
@@ -1351,7 +1360,7 @@ zil_commit_waiter_link_lwb(zil_commit_waiter_t *zcw, lwb_t *lwb)
ASSERT(!list_link_active(&zcw->zcw_node));
list_insert_tail(&lwb->lwb_waiters, zcw);
- ASSERT3P(zcw->zcw_lwb, ==, NULL);
+ ASSERT0P(zcw->zcw_lwb);
zcw->zcw_lwb = lwb;
}
@@ -1365,7 +1374,7 @@ zil_commit_waiter_link_nolwb(zil_commit_waiter_t *zcw, list_t *nolwb)
{
ASSERT(!list_link_active(&zcw->zcw_node));
list_insert_tail(nolwb, zcw);
- ASSERT3P(zcw->zcw_lwb, ==, NULL);
+ ASSERT0P(zcw->zcw_lwb);
}
void
@@ -1383,7 +1392,7 @@ zil_lwb_add_block(lwb_t *lwb, const blkptr_t *bp)
if (zil_nocacheflush)
return;
- mutex_enter(&lwb->lwb_vdev_lock);
+ mutex_enter(&lwb->lwb_lock);
for (i = 0; i < ndvas; i++) {
zvsearch.zv_vdev = DVA_GET_VDEV(&bp->blk_dva[i]);
if (avl_find(t, &zvsearch, &where) == NULL) {
@@ -1392,7 +1401,7 @@ zil_lwb_add_block(lwb_t *lwb, const blkptr_t *bp)
avl_insert(t, zv, where);
}
}
- mutex_exit(&lwb->lwb_vdev_lock);
+ mutex_exit(&lwb->lwb_lock);
}
static void
@@ -1409,12 +1418,12 @@ zil_lwb_flush_defer(lwb_t *lwb, lwb_t *nlwb)
/*
* While 'lwb' is at a point in its lifetime where lwb_vdev_tree does
- * not need the protection of lwb_vdev_lock (it will only be modified
+ * not need the protection of lwb_lock (it will only be modified
* while holding zilog->zl_lock) as its writes and those of its
* children have all completed. The younger 'nlwb' may be waiting on
* future writes to additional vdevs.
*/
- mutex_enter(&nlwb->lwb_vdev_lock);
+ mutex_enter(&nlwb->lwb_lock);
/*
* Tear down the 'lwb' vdev tree, ensuring that entries which do not
* exist in 'nlwb' are moved to it, freeing any would-be duplicates.
@@ -1428,7 +1437,7 @@ zil_lwb_flush_defer(lwb_t *lwb, lwb_t *nlwb)
kmem_free(zv, sizeof (*zv));
}
}
- mutex_exit(&nlwb->lwb_vdev_lock);
+ mutex_exit(&nlwb->lwb_lock);
}
void
@@ -1482,13 +1491,9 @@ zil_lwb_flush_vdevs_done(zio_t *zio)
}
while ((itx = list_remove_head(&lwb->lwb_itxs)) != NULL)
- zil_itx_destroy(itx);
+ zil_itx_destroy(itx, 0);
while ((zcw = list_remove_head(&lwb->lwb_waiters)) != NULL) {
- mutex_enter(&zcw->zcw_lock);
-
- ASSERT3P(zcw->zcw_lwb, ==, lwb);
- zcw->zcw_lwb = NULL;
/*
* We expect any ZIO errors from child ZIOs to have been
* propagated "up" to this specific LWB's root ZIO, in
@@ -1503,14 +1508,7 @@ zil_lwb_flush_vdevs_done(zio_t *zio)
* errors not being handled correctly here. See the
* comment above the call to "zio_flush" for details.
*/
-
- zcw->zcw_zio_error = zio->io_error;
-
- ASSERT3B(zcw->zcw_done, ==, B_FALSE);
- zcw->zcw_done = B_TRUE;
- cv_broadcast(&zcw->zcw_cv);
-
- mutex_exit(&zcw->zcw_lock);
+ zil_commit_waiter_done(zcw, zio->io_error);
}
uint64_t txg = lwb->lwb_issued_txg;
@@ -1582,7 +1580,7 @@ zil_lwb_write_done(zio_t *zio)
avl_tree_t *t = &lwb->lwb_vdev_tree;
void *cookie = NULL;
zil_vdev_node_t *zv;
- lwb_t *nlwb;
+ lwb_t *nlwb = NULL;
ASSERT3S(spa_config_held(spa, SCL_STATE, RW_READER), !=, 0);
@@ -1602,9 +1600,11 @@ zil_lwb_write_done(zio_t *zio)
* its write ZIO a parent this ZIO. In such case we can not defer
* our flushes or below may be a race between the done callbacks.
*/
- nlwb = list_next(&zilog->zl_lwb_list, lwb);
- if (nlwb && nlwb->lwb_state != LWB_STATE_ISSUED)
- nlwb = NULL;
+ if (!(lwb->lwb_flags & LWB_FLAG_CRASHED)) {
+ nlwb = list_next(&zilog->zl_lwb_list, lwb);
+ if (nlwb && nlwb->lwb_state != LWB_STATE_ISSUED)
+ nlwb = NULL;
+ }
mutex_exit(&zilog->zl_lock);
if (avl_numnodes(t) == 0)
@@ -1618,12 +1618,17 @@ zil_lwb_write_done(zio_t *zio)
* written out.
*
* Additionally, we don't perform any further error handling at
- * this point (e.g. setting "zcw_zio_error" appropriately), as
- * we expect that to occur in "zil_lwb_flush_vdevs_done" (thus,
- * we expect any error seen here, to have been propagated to
- * that function).
+ * this point (e.g. setting "zcw_error" appropriately), as we
+ * expect that to occur in "zil_lwb_flush_vdevs_done" (thus, we
+ * expect any error seen here, to have been propagated to that
+ * function).
+ *
+ * Note that we treat a "crashed" LWB as though it was in error,
+ * even if it did appear to succeed, because we've already
+ * signaled error and cleaned up waiters and committers in
+ * zil_crash(); we just want to clean up and get out of here.
*/
- if (zio->io_error != 0) {
+ if (zio->io_error != 0 || (lwb->lwb_flags & LWB_FLAG_CRASHED)) {
while ((zv = avl_destroy_nodes(t, &cookie)) != NULL)
kmem_free(zv, sizeof (*zv));
return;
@@ -1736,10 +1741,26 @@ zil_lwb_write_open(zilog_t *zilog, lwb_t *lwb)
return;
}
+ mutex_enter(&lwb->lwb_lock);
mutex_enter(&zilog->zl_lock);
lwb->lwb_state = LWB_STATE_OPENED;
zilog->zl_last_lwb_opened = lwb;
mutex_exit(&zilog->zl_lock);
+ mutex_exit(&lwb->lwb_lock);
+
+ /*
+ * Allocate buffer and set up LWB capacities.
+ */
+ ASSERT0P(lwb->lwb_buf);
+ ASSERT3U(lwb->lwb_sz, >, 0);
+ lwb->lwb_buf = zio_buf_alloc(lwb->lwb_sz);
+ if (lwb->lwb_flags & LWB_FLAG_SLIM) {
+ lwb->lwb_nmax = lwb->lwb_sz;
+ lwb->lwb_nused = lwb->lwb_nfilled = sizeof (zil_chain_t);
+ } else {
+ lwb->lwb_nmax = lwb->lwb_sz - sizeof (zil_chain_t);
+ lwb->lwb_nused = lwb->lwb_nfilled = 0;
+ }
}
/*
@@ -1756,6 +1777,8 @@ static uint_t
zil_lwb_plan(zilog_t *zilog, uint64_t size, uint_t *minsize)
{
uint_t md = zilog->zl_max_block_size - sizeof (zil_chain_t);
+ uint_t waste = zil_max_waste_space(zilog);
+ waste = MAX(waste, zilog->zl_cur_max);
if (size <= md) {
/*
@@ -1766,9 +1789,10 @@ zil_lwb_plan(zilog_t *zilog, uint64_t size, uint_t *minsize)
} else if (size > 8 * md) {
/*
* Big bursts use maximum blocks. The first block size
- * is hard to predict, but it does not really matter.
+ * is hard to predict, but we need at least enough space
+ * to make reasonable progress.
*/
- *minsize = 0;
+ *minsize = waste;
return (md);
}
@@ -1781,57 +1805,52 @@ zil_lwb_plan(zilog_t *zilog, uint64_t size, uint_t *minsize)
uint_t s = size;
uint_t n = DIV_ROUND_UP(s, md - sizeof (lr_write_t));
uint_t chunk = DIV_ROUND_UP(s, n);
- uint_t waste = zil_max_waste_space(zilog);
- waste = MAX(waste, zilog->zl_cur_max);
if (chunk <= md - waste) {
*minsize = MAX(s - (md - waste) * (n - 1), waste);
return (chunk);
} else {
- *minsize = 0;
+ *minsize = waste;
return (md);
}
}
/*
* Try to predict next block size based on previous history. Make prediction
- * sufficient for 7 of 8 previous bursts. Don't try to save if the saving is
- * less then 50%, extra writes may cost more, but we don't want single spike
- * to badly affect our predictions.
+ * sufficient for 7 of 8 previous bursts, but don't try to save if the saving
+ * is less then 50%. Extra writes may cost more, but we don't want single
+ * spike to badly affect our predictions.
*/
-static uint_t
-zil_lwb_predict(zilog_t *zilog)
+static void
+zil_lwb_predict(zilog_t *zilog, uint64_t *min_predict, uint64_t *max_predict)
{
- uint_t m, o;
+ uint_t m1 = 0, m2 = 0, o;
- /* If we are in the middle of a burst, take it into account also. */
- if (zilog->zl_cur_size > 0) {
- o = zil_lwb_plan(zilog, zilog->zl_cur_size, &m);
- } else {
+ /* If we are in the middle of a burst, take it as another data point. */
+ if (zilog->zl_cur_size > 0)
+ o = zil_lwb_plan(zilog, zilog->zl_cur_size, &m1);
+ else
o = UINT_MAX;
- m = 0;
- }
- /* Find minimum optimal size. We don't need to go below that. */
- for (int i = 0; i < ZIL_BURSTS; i++)
- o = MIN(o, zilog->zl_prev_opt[i]);
-
- /* Find two biggest minimal first block sizes above the optimal. */
- uint_t m1 = MAX(m, o), m2 = o;
+ /* Find two largest minimal first block sizes. */
for (int i = 0; i < ZIL_BURSTS; i++) {
- m = zilog->zl_prev_min[i];
- if (m >= m1) {
+ uint_t cur = zilog->zl_prev_min[i];
+ if (cur >= m1) {
m2 = m1;
- m1 = m;
- } else if (m > m2) {
- m2 = m;
+ m1 = cur;
+ } else if (cur > m2) {
+ m2 = cur;
}
}
- /*
- * If second minimum size gives 50% saving -- use it. It may cost us
- * one additional write later, but the space saving is just too big.
- */
- return ((m1 < m2 * 2) ? m1 : m2);
+ /* Minimum should guarantee progress in most cases. */
+ *min_predict = (m1 < m2 * 2) ? m1 : m2;
+
+ /* Maximum doesn't need to go below the minimum optimal size. */
+ for (int i = 0; i < ZIL_BURSTS; i++)
+ o = MIN(o, zilog->zl_prev_opt[i]);
+ m1 = MAX(m1, o);
+ m2 = MAX(m2, o);
+ *max_predict = (m1 < m2 * 2) ? m1 : m2;
}
/*
@@ -1839,12 +1858,13 @@ zil_lwb_predict(zilog_t *zilog)
* Has to be called under zl_issuer_lock to chain more lwbs.
*/
static lwb_t *
-zil_lwb_write_close(zilog_t *zilog, lwb_t *lwb, lwb_state_t state)
+zil_lwb_write_close(zilog_t *zilog, lwb_t *lwb)
{
- uint64_t blksz, plan, plan2;
+ uint64_t minbs, maxbs;
ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock));
ASSERT3S(lwb->lwb_state, ==, LWB_STATE_OPENED);
+ membar_producer();
lwb->lwb_state = LWB_STATE_CLOSED;
/*
@@ -1869,33 +1889,40 @@ zil_lwb_write_close(zilog_t *zilog, lwb_t *lwb, lwb_state_t state)
* Try to predict what can it be and plan for the worst case.
*/
uint_t m;
- plan = zil_lwb_plan(zilog, zilog->zl_cur_left, &m);
+ maxbs = zil_lwb_plan(zilog, zilog->zl_cur_left, &m);
+ minbs = m;
if (zilog->zl_parallel) {
- plan2 = zil_lwb_plan(zilog, zilog->zl_cur_left +
- zil_lwb_predict(zilog), &m);
- if (plan < plan2)
- plan = plan2;
+ uint64_t minp, maxp;
+ zil_lwb_predict(zilog, &minp, &maxp);
+ maxp = zil_lwb_plan(zilog, zilog->zl_cur_left + maxp,
+ &m);
+ if (maxbs < maxp)
+ maxbs = maxp;
}
} else {
/*
* The previous burst is done and we can only predict what
* will come next.
*/
- plan = zil_lwb_predict(zilog);
+ zil_lwb_predict(zilog, &minbs, &maxbs);
}
- blksz = plan + sizeof (zil_chain_t);
- blksz = P2ROUNDUP_TYPED(blksz, ZIL_MIN_BLKSZ, uint64_t);
- blksz = MIN(blksz, zilog->zl_max_block_size);
- DTRACE_PROBE3(zil__block__size, zilog_t *, zilog, uint64_t, blksz,
- uint64_t, plan);
- return (zil_alloc_lwb(zilog, blksz, NULL, 0, 0, state));
+ minbs += sizeof (zil_chain_t);
+ maxbs += sizeof (zil_chain_t);
+ minbs = P2ROUNDUP_TYPED(minbs, ZIL_MIN_BLKSZ, uint64_t);
+ maxbs = P2ROUNDUP_TYPED(maxbs, ZIL_MIN_BLKSZ, uint64_t);
+ maxbs = MIN(maxbs, zilog->zl_max_block_size);
+ minbs = MIN(minbs, maxbs);
+ DTRACE_PROBE3(zil__block__size, zilog_t *, zilog, uint64_t, minbs,
+ uint64_t, maxbs);
+
+ return (zil_alloc_lwb(zilog, NULL, minbs, maxbs, 0, 0));
}
/*
* Finalize previously closed block and issue the write zio.
*/
-static void
+static int
zil_lwb_write_issue(zilog_t *zilog, lwb_t *lwb)
{
spa_t *spa = zilog->zl_spa;
@@ -1909,8 +1936,13 @@ zil_lwb_write_issue(zilog_t *zilog, lwb_t *lwb)
/* Actually fill the lwb with the data. */
for (itx_t *itx = list_head(&lwb->lwb_itxs); itx;
- itx = list_next(&lwb->lwb_itxs, itx))
- zil_lwb_commit(zilog, lwb, itx);
+ itx = list_next(&lwb->lwb_itxs, itx)) {
+ error = zil_lwb_commit(zilog, lwb, itx);
+ if (error != 0) {
+ ASSERT3U(error, ==, ESHUTDOWN);
+ return (error);
+ }
+ }
lwb->lwb_nused = lwb->lwb_nfilled;
ASSERT3U(lwb->lwb_nused, <=, lwb->lwb_nmax);
@@ -1928,19 +1960,21 @@ zil_lwb_write_issue(zilog_t *zilog, lwb_t *lwb)
lwb->lwb_state = LWB_STATE_READY;
if (BP_IS_HOLE(&lwb->lwb_blk) && lwb->lwb_error == 0) {
mutex_exit(&zilog->zl_lock);
- return;
+ return (0);
}
mutex_exit(&zilog->zl_lock);
next_lwb:
- if (lwb->lwb_slim)
+ if (lwb->lwb_flags & LWB_FLAG_SLIM)
zilc = (zil_chain_t *)lwb->lwb_buf;
else
zilc = (zil_chain_t *)(lwb->lwb_buf + lwb->lwb_nmax);
- int wsz = lwb->lwb_sz;
+ uint64_t alloc_size = BP_GET_LSIZE(&lwb->lwb_blk);
+ int wsz = alloc_size;
if (lwb->lwb_error == 0) {
abd_t *lwb_abd = abd_get_from_buf(lwb->lwb_buf, lwb->lwb_sz);
- if (!lwb->lwb_slog || zilog->zl_cur_size <= zil_slog_bulk)
+ if (!(lwb->lwb_flags & LWB_FLAG_SLOG) ||
+ zilog->zl_cur_size <= zil_slog_bulk)
prio = ZIO_PRIORITY_SYNC_WRITE;
else
prio = ZIO_PRIORITY_ASYNC_WRITE;
@@ -1948,16 +1982,17 @@ next_lwb:
ZB_ZIL_OBJECT, ZB_ZIL_LEVEL,
lwb->lwb_blk.blk_cksum.zc_word[ZIL_ZC_SEQ]);
lwb->lwb_write_zio = zio_rewrite(lwb->lwb_root_zio, spa, 0,
- &lwb->lwb_blk, lwb_abd, lwb->lwb_sz, zil_lwb_write_done,
+ &lwb->lwb_blk, lwb_abd, alloc_size, zil_lwb_write_done,
lwb, prio, ZIO_FLAG_CANFAIL, &zb);
zil_lwb_add_block(lwb, &lwb->lwb_blk);
- if (lwb->lwb_slim) {
+ if (lwb->lwb_flags & LWB_FLAG_SLIM) {
/* For Slim ZIL only write what is used. */
wsz = P2ROUNDUP_TYPED(lwb->lwb_nused, ZIL_MIN_BLKSZ,
int);
- ASSERT3S(wsz, <=, lwb->lwb_sz);
- zio_shrink(lwb->lwb_write_zio, wsz);
+ ASSERT3S(wsz, <=, alloc_size);
+ if (wsz < alloc_size)
+ zio_shrink(lwb->lwb_write_zio, wsz);
wsz = lwb->lwb_write_zio->io_size;
}
memset(lwb->lwb_buf + lwb->lwb_nused, 0, wsz - lwb->lwb_nused);
@@ -1993,13 +2028,53 @@ next_lwb:
BP_ZERO(bp);
error = lwb->lwb_error;
if (error == 0) {
- error = zio_alloc_zil(spa, zilog->zl_os, txg, bp, nlwb->lwb_sz,
- &slog);
+ /*
+ * Allocation flexibility depends on LWB state:
+ * if NEW: allow range allocation and larger sizes;
+ * if OPENED: use fixed predetermined allocation size;
+ * if CLOSED + Slim: allocate precisely for actual usage.
+ */
+ boolean_t flexible = (nlwb->lwb_state == LWB_STATE_NEW);
+ if (flexible) {
+ /* We need to prevent opening till we update lwb_sz. */
+ mutex_enter(&nlwb->lwb_lock);
+ flexible = (nlwb->lwb_state == LWB_STATE_NEW);
+ if (!flexible)
+ mutex_exit(&nlwb->lwb_lock); /* We lost. */
+ }
+ boolean_t closed_slim = (nlwb->lwb_state == LWB_STATE_CLOSED &&
+ (lwb->lwb_flags & LWB_FLAG_SLIM));
+
+ uint64_t min_size, max_size;
+ if (closed_slim) {
+ /* This transition is racy, but only one way. */
+ membar_consumer();
+ min_size = max_size = P2ROUNDUP_TYPED(nlwb->lwb_nused,
+ ZIL_MIN_BLKSZ, uint64_t);
+ } else if (flexible) {
+ min_size = nlwb->lwb_min_sz;
+ max_size = nlwb->lwb_sz;
+ } else {
+ min_size = max_size = nlwb->lwb_sz;
+ }
+
+ error = zio_alloc_zil(spa, zilog->zl_os, txg, bp,
+ min_size, max_size, &slog, flexible);
+ if (error == 0) {
+ if (closed_slim)
+ ASSERT3U(BP_GET_LSIZE(bp), ==, max_size);
+ else if (flexible)
+ nlwb->lwb_sz = BP_GET_LSIZE(bp);
+ else
+ ASSERT3U(BP_GET_LSIZE(bp), ==, nlwb->lwb_sz);
+ }
+ if (flexible)
+ mutex_exit(&nlwb->lwb_lock);
}
if (error == 0) {
ASSERT3U(BP_GET_BIRTH(bp), ==, txg);
- BP_SET_CHECKSUM(bp, nlwb->lwb_slim ? ZIO_CHECKSUM_ZILOG2 :
- ZIO_CHECKSUM_ZILOG);
+ BP_SET_CHECKSUM(bp, (nlwb->lwb_flags & LWB_FLAG_SLIM) ?
+ ZIO_CHECKSUM_ZILOG2 : ZIO_CHECKSUM_ZILOG);
bp->blk_cksum = lwb->lwb_blk.blk_cksum;
bp->blk_cksum.zc_word[ZIL_ZC_SEQ]++;
}
@@ -2028,14 +2103,15 @@ next_lwb:
if (nlwb) {
nlwb->lwb_blk = *bp;
nlwb->lwb_error = error;
- nlwb->lwb_slog = slog;
+ if (slog)
+ nlwb->lwb_flags |= LWB_FLAG_SLOG;
nlwb->lwb_alloc_txg = txg;
if (nlwb->lwb_state != LWB_STATE_READY)
nlwb = NULL;
}
mutex_exit(&zilog->zl_lock);
- if (lwb->lwb_slog) {
+ if (lwb->lwb_flags & LWB_FLAG_SLOG) {
ZIL_STAT_BUMP(zilog, zil_itx_metaslab_slog_count);
ZIL_STAT_INCR(zilog, zil_itx_metaslab_slog_bytes,
lwb->lwb_nused);
@@ -2065,6 +2141,8 @@ next_lwb:
lwb = nlwb;
if (lwb)
goto next_lwb;
+
+ return (0);
}
/*
@@ -2207,7 +2285,6 @@ zil_lwb_assign(zilog_t *zilog, lwb_t *lwb, itx_t *itx, list_t *ilwbs)
ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock));
ASSERT3P(lwb, !=, NULL);
- ASSERT3P(lwb->lwb_buf, !=, NULL);
zil_lwb_write_open(zilog, lwb);
@@ -2249,9 +2326,10 @@ cont:
(dlen % max_log_data == 0 ||
lwb_sp < reclen + dlen % max_log_data))) {
list_insert_tail(ilwbs, lwb);
- lwb = zil_lwb_write_close(zilog, lwb, LWB_STATE_OPENED);
+ lwb = zil_lwb_write_close(zilog, lwb);
if (lwb == NULL)
return (NULL);
+ zil_lwb_write_open(zilog, lwb);
lwb_sp = lwb->lwb_nmax - lwb->lwb_nused;
}
@@ -2308,11 +2386,13 @@ cont:
return (lwb);
}
+static void zil_crash(zilog_t *zilog);
+
/*
* Fill the actual transaction data into the lwb, following zil_lwb_assign().
* Does not require locking.
*/
-static void
+static int
zil_lwb_commit(zilog_t *zilog, lwb_t *lwb, itx_t *itx)
{
lr_t *lr, *lrb;
@@ -2324,7 +2404,7 @@ zil_lwb_commit(zilog_t *zilog, lwb_t *lwb, itx_t *itx)
lrw = (lr_write_t *)lr;
if (lr->lrc_txtype == TX_COMMIT)
- return;
+ return (0);
reclen = lr->lrc_reclen;
dlen = zil_itx_data_size(itx);
@@ -2410,16 +2490,35 @@ zil_lwb_commit(zilog_t *zilog, lwb_t *lwb, itx_t *itx)
". Falling back to txg_wait_synced().",
error);
zfs_fallthrough;
- case EIO:
- txg_wait_synced(zilog->zl_dmu_pool,
- lr->lrc_txg);
+ case EIO: {
+ int error = txg_wait_synced_flags(
+ zilog->zl_dmu_pool,
+ lr->lrc_txg, TXG_WAIT_SUSPEND);
+ if (error != 0) {
+ ASSERT3U(error, ==, ESHUTDOWN);
+ /*
+ * zil_lwb_commit() is called from a
+ * loop over a list of itxs at the
+ * top of zil_lwb_write_issue(), which
+ * itself is called from a loop over a
+ * list of lwbs in various places.
+ * zil_crash() will free those itxs
+ * and sometimes the lwbs, so they
+ * are invalid when zil_crash() returns.
+ * Callers must pretty much abort
+ * immediately.
+ */
+ zil_crash(zilog);
+ return (error);
+ }
zfs_fallthrough;
+ }
case ENOENT:
zfs_fallthrough;
case EEXIST:
zfs_fallthrough;
case EALREADY:
- return;
+ return (0);
}
}
}
@@ -2427,6 +2526,8 @@ zil_lwb_commit(zilog_t *zilog, lwb_t *lwb, itx_t *itx)
lwb->lwb_nfilled += reclen + dlen;
ASSERT3S(lwb->lwb_nfilled, <=, lwb->lwb_nused);
ASSERT0(P2PHASE(lwb->lwb_nfilled, sizeof (uint64_t)));
+
+ return (0);
}
itx_t *
@@ -2468,7 +2569,7 @@ zil_itx_clone(itx_t *oitx)
}
void
-zil_itx_destroy(itx_t *itx)
+zil_itx_destroy(itx_t *itx, int err)
{
ASSERT3U(itx->itx_size, >=, sizeof (itx_t));
ASSERT3U(itx->itx_lr.lrc_reclen, ==,
@@ -2477,7 +2578,7 @@ zil_itx_destroy(itx_t *itx)
IMPLY(itx->itx_callback != NULL, itx->itx_lr.lrc_txtype != TX_COMMIT);
if (itx->itx_callback != NULL)
- itx->itx_callback(itx->itx_callback_data);
+ itx->itx_callback(itx->itx_callback_data, err);
zio_data_buf_free(itx, itx->itx_size);
}
@@ -2518,9 +2619,9 @@ zil_itxg_clean(void *arg)
* called) we will hit this case.
*/
if (itx->itx_lr.lrc_txtype == TX_COMMIT)
- zil_commit_waiter_skip(itx->itx_private);
+ zil_commit_waiter_done(itx->itx_private, 0);
- zil_itx_destroy(itx);
+ zil_itx_destroy(itx, 0);
}
cookie = NULL;
@@ -2530,7 +2631,7 @@ zil_itxg_clean(void *arg)
while ((itx = list_remove_head(list)) != NULL) {
/* commit itxs should never be on the async lists. */
ASSERT3U(itx->itx_lr.lrc_txtype, !=, TX_COMMIT);
- zil_itx_destroy(itx);
+ zil_itx_destroy(itx, 0);
}
list_destroy(list);
kmem_free(ian, sizeof (itx_async_node_t));
@@ -2592,7 +2693,7 @@ zil_remove_async(zilog_t *zilog, uint64_t oid)
while ((itx = list_remove_head(&clean_list)) != NULL) {
/* commit itxs should never be on the async lists. */
ASSERT3U(itx->itx_lr.lrc_txtype, !=, TX_COMMIT);
- zil_itx_destroy(itx);
+ zil_itx_destroy(itx, 0);
}
list_destroy(&clean_list);
}
@@ -2677,6 +2778,68 @@ zil_itx_assign(zilog_t *zilog, itx_t *itx, dmu_tx_t *tx)
}
/*
+ * Post-crash cleanup. This is called from zil_clean() because it needs to
+ * do cleanup after every txg until the ZIL is restarted, and zilog_dirty()
+ * can arrange that easily, unlike zil_sync() which is more complicated to
+ * get a call to without actual dirty data.
+ */
+static void
+zil_crash_clean(zilog_t *zilog, uint64_t synced_txg)
+{
+ ASSERT(MUTEX_HELD(&zilog->zl_lock));
+ ASSERT3U(zilog->zl_restart_txg, >, 0);
+
+ /* Clean up anything on the crash list from earlier txgs */
+ lwb_t *lwb;
+ while ((lwb = list_head(&zilog->zl_lwb_crash_list)) != NULL) {
+ if (lwb->lwb_alloc_txg >= synced_txg ||
+ lwb->lwb_max_txg >= synced_txg) {
+ /*
+ * This lwb was allocated or updated on this txg, or
+ * in the future. We stop processing here, to avoid
+ * the strange situation of freeing a ZIL block on
+ * on the same or earlier txg than what it was
+ * allocated for.
+ *
+ * We'll take care of it on the next txg.
+ */
+ break;
+ }
+
+ /* This LWB is from the past, so we can clean it up now. */
+ ASSERT(lwb->lwb_flags & LWB_FLAG_CRASHED);
+ list_remove(&zilog->zl_lwb_crash_list, lwb);
+ if (lwb->lwb_buf != NULL)
+ zio_buf_free(lwb->lwb_buf, lwb->lwb_sz);
+ if (!BP_IS_HOLE(&lwb->lwb_blk))
+ /*
+ * Free on the next txg, since zil_clean() is called
+ * once synced_txg has already been completed.
+ */
+ zio_free(zilog->zl_spa, synced_txg+1, &lwb->lwb_blk);
+ zil_free_lwb(zilog, lwb);
+ }
+
+ if (zilog->zl_restart_txg > synced_txg) {
+ /*
+ * Not reached the restart txg yet, so mark the ZIL dirty for
+ * the next txg and we'll consider it all again then.
+ */
+ zilog_dirty(zilog, synced_txg+1);
+ return;
+ }
+
+ /*
+ * Reached the restart txg, so we can allow new calls to zil_commit().
+ * All ZIL txgs have long past so there should be no IO waiting.
+ */
+ ASSERT(list_is_empty(&zilog->zl_lwb_list));
+ ASSERT(list_is_empty(&zilog->zl_lwb_crash_list));
+
+ zilog->zl_restart_txg = 0;
+}
+
+/*
* If there are any in-memory intent log transactions which have now been
* synced then start up a taskq to free them. We should only do this after we
* have written out the uberblocks (i.e. txg has been committed) so that
@@ -2691,6 +2854,15 @@ zil_clean(zilog_t *zilog, uint64_t synced_txg)
ASSERT3U(synced_txg, <, ZILTEST_TXG);
+ /* Do cleanup and restart after crash. */
+ if (zilog->zl_restart_txg > 0) {
+ mutex_enter(&zilog->zl_lock);
+ /* Make sure we didn't lose a race. */
+ if (zilog->zl_restart_txg > 0)
+ zil_crash_clean(zilog, synced_txg);
+ mutex_exit(&zilog->zl_lock);
+ }
+
mutex_enter(&itxg->itxg_lock);
if (itxg->itxg_itxs == NULL || itxg->itxg_txg == ZILTEST_TXG) {
mutex_exit(&itxg->itxg_lock);
@@ -2875,7 +3047,7 @@ zil_prune_commit_list(zilog_t *zilog)
* never any itx's for it to wait on), so it's
* safe to skip this waiter and mark it done.
*/
- zil_commit_waiter_skip(itx->itx_private);
+ zil_commit_waiter_done(itx->itx_private, 0);
} else {
zil_commit_waiter_link_lwb(itx->itx_private, last_lwb);
}
@@ -2883,13 +3055,13 @@ zil_prune_commit_list(zilog_t *zilog)
mutex_exit(&zilog->zl_lock);
list_remove(&zilog->zl_itx_commit_list, itx);
- zil_itx_destroy(itx);
+ zil_itx_destroy(itx, 0);
}
IMPLY(itx != NULL, itx->itx_lr.lrc_txtype != TX_COMMIT);
}
-static void
+static int
zil_commit_writer_stall(zilog_t *zilog)
{
/*
@@ -2914,8 +3086,22 @@ zil_commit_writer_stall(zilog_t *zilog)
*/
ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock));
ZIL_STAT_BUMP(zilog, zil_commit_stall_count);
- txg_wait_synced(zilog->zl_dmu_pool, 0);
+
+ int err = txg_wait_synced_flags(zilog->zl_dmu_pool, 0,
+ TXG_WAIT_SUSPEND);
+ if (err != 0) {
+ ASSERT3U(err, ==, ESHUTDOWN);
+ zil_crash(zilog);
+ }
+
+ /*
+ * Either zil_sync() has been called to wait for and clean up any
+ * in-flight LWBs, or zil_crash() has emptied out the list and arranged
+ * for them to be cleaned up later.
+ */
ASSERT(list_is_empty(&zilog->zl_lwb_list));
+
+ return (err);
}
static void
@@ -3082,7 +3268,7 @@ zil_process_commit_list(zilog_t *zilog, zil_commit_waiter_t *zcw, list_t *ilwbs)
} else {
ASSERT3S(lrc->lrc_txtype, !=, TX_COMMIT);
zilog->zl_cur_left -= zil_itx_full_size(itx);
- zil_itx_destroy(itx);
+ zil_itx_destroy(itx, 0);
}
}
@@ -3092,10 +3278,21 @@ zil_process_commit_list(zilog_t *zilog, zil_commit_waiter_t *zcw, list_t *ilwbs)
* "next" lwb on-disk. When this happens, we must stall
* the ZIL write pipeline; see the comment within
* zil_commit_writer_stall() for more details.
+ *
+ * ESHUTDOWN has to be handled carefully here. If we get it,
+ * then the pool suspended and zil_crash() was called, so we
+ * need to stop trying and just get an error back to the
+ * callers.
*/
- while ((lwb = list_remove_head(ilwbs)) != NULL)
- zil_lwb_write_issue(zilog, lwb);
- zil_commit_writer_stall(zilog);
+ int err = 0;
+ while ((lwb = list_remove_head(ilwbs)) != NULL) {
+ if (err == 0)
+ err = zil_lwb_write_issue(zilog, lwb);
+ }
+ if (err != ESHUTDOWN)
+ err = zil_commit_writer_stall(zilog);
+ if (err == ESHUTDOWN)
+ err = SET_ERROR(EIO);
/*
* Additionally, we have to signal and mark the "nolwb"
@@ -3105,7 +3302,7 @@ zil_process_commit_list(zilog_t *zilog, zil_commit_waiter_t *zcw, list_t *ilwbs)
*/
zil_commit_waiter_t *zcw;
while ((zcw = list_remove_head(&nolwb_waiters)) != NULL)
- zil_commit_waiter_skip(zcw);
+ zil_commit_waiter_done(zcw, err);
/*
* And finally, we have to destroy the itx's that
@@ -3113,7 +3310,7 @@ zil_process_commit_list(zilog_t *zilog, zil_commit_waiter_t *zcw, list_t *ilwbs)
* the itx's callback if one exists for the itx.
*/
while ((itx = list_remove_head(&nolwb_itxs)) != NULL)
- zil_itx_destroy(itx);
+ zil_itx_destroy(itx, err);
} else {
ASSERT(list_is_empty(&nolwb_waiters));
ASSERT3P(lwb, !=, NULL);
@@ -3167,11 +3364,17 @@ zil_process_commit_list(zilog_t *zilog, zil_commit_waiter_t *zcw, list_t *ilwbs)
(!zilog->zl_parallel || zilog->zl_suspend > 0)) {
zil_burst_done(zilog);
list_insert_tail(ilwbs, lwb);
- lwb = zil_lwb_write_close(zilog, lwb, LWB_STATE_NEW);
+ lwb = zil_lwb_write_close(zilog, lwb);
if (lwb == NULL) {
- while ((lwb = list_remove_head(ilwbs)) != NULL)
- zil_lwb_write_issue(zilog, lwb);
- zil_commit_writer_stall(zilog);
+ int err = 0;
+ while ((lwb =
+ list_remove_head(ilwbs)) != NULL) {
+ if (err == 0)
+ err = zil_lwb_write_issue(
+ zilog, lwb);
+ }
+ if (err != ESHUTDOWN)
+ (void) zil_commit_writer_stall(zilog);
}
}
}
@@ -3230,10 +3433,23 @@ zil_commit_writer(zilog_t *zilog, zil_commit_waiter_t *zcw)
zil_prune_commit_list(zilog);
zil_process_commit_list(zilog, zcw, &ilwbs);
+ /*
+ * If the ZIL failed somewhere inside zil_process_commit_list(), it's
+ * will be because a fallback to txg_wait_sync_flags() happened at some
+ * point (eg zil_commit_writer_stall()). All cases should issue and
+ * empty ilwbs, so there will be nothing to in the issue loop below.
+ * That's why we don't have to plumb the error value back from
+ * zil_process_commit_list(), and don't have to skip it.
+ */
+ IMPLY(zilog->zl_restart_txg > 0, list_is_empty(&ilwbs));
+
out:
mutex_exit(&zilog->zl_issuer_lock);
- while ((lwb = list_remove_head(&ilwbs)) != NULL)
- zil_lwb_write_issue(zilog, lwb);
+ int err = 0;
+ while ((lwb = list_remove_head(&ilwbs)) != NULL) {
+ if (err == 0)
+ err = zil_lwb_write_issue(zilog, lwb);
+ }
list_destroy(&ilwbs);
return (wtxg);
}
@@ -3326,7 +3542,7 @@ zil_commit_waiter_timeout(zilog_t *zilog, zil_commit_waiter_t *zcw)
* hasn't been issued.
*/
zil_burst_done(zilog);
- lwb_t *nlwb = zil_lwb_write_close(zilog, lwb, LWB_STATE_NEW);
+ lwb_t *nlwb = zil_lwb_write_close(zilog, lwb);
ASSERT3S(lwb->lwb_state, ==, LWB_STATE_CLOSED);
@@ -3402,7 +3618,7 @@ zil_commit_waiter(zilog_t *zilog, zil_commit_waiter_t *zcw)
* commit itxs. When this occurs, the commit waiters linked
* off of these commit itxs will not be committed to an
* lwb. Additionally, these commit waiters will not be
- * marked done until zil_commit_waiter_skip() is called via
+ * marked done until zil_commit_waiter_done() is called via
* zil_itxg_clean().
*
* Thus, it's possible for this commit waiter (i.e. the
@@ -3480,7 +3696,7 @@ zil_alloc_commit_waiter(void)
list_link_init(&zcw->zcw_node);
zcw->zcw_lwb = NULL;
zcw->zcw_done = B_FALSE;
- zcw->zcw_zio_error = 0;
+ zcw->zcw_error = 0;
return (zcw);
}
@@ -3489,7 +3705,7 @@ static void
zil_free_commit_waiter(zil_commit_waiter_t *zcw)
{
ASSERT(!list_link_active(&zcw->zcw_node));
- ASSERT3P(zcw->zcw_lwb, ==, NULL);
+ ASSERT0P(zcw->zcw_lwb);
ASSERT3B(zcw->zcw_done, ==, B_TRUE);
mutex_destroy(&zcw->zcw_lock);
cv_destroy(&zcw->zcw_cv);
@@ -3526,6 +3742,99 @@ zil_commit_itx_assign(zilog_t *zilog, zil_commit_waiter_t *zcw)
}
/*
+ * Crash the ZIL. This is something like suspending, but abandons the ZIL
+ * without further IO until the wanted txg completes. No effort is made to
+ * close the on-disk chain or do any other on-disk work, as the pool may
+ * have suspended. zil_sync() will handle cleanup as normal and restart the
+ * ZIL once enough txgs have passed.
+ */
+static void
+zil_crash(zilog_t *zilog)
+{
+ mutex_enter(&zilog->zl_lock);
+
+ uint64_t txg = spa_syncing_txg(zilog->zl_spa);
+ uint64_t restart_txg =
+ spa_syncing_txg(zilog->zl_spa) + TXG_CONCURRENT_STATES;
+
+ if (zilog->zl_restart_txg > 0) {
+ /*
+ * If the ZIL is already crashed, it's almost certainly because
+ * we lost a race involving multiple callers from
+ * zil_commit_impl().
+ */
+
+ /*
+ * This sanity check is to support my understanding that in the
+ * event of multiple callers to zil_crash(), only one of them
+ * can possibly be in the codepath to issue lwbs; the rest
+ * should be calling from zil_commit_impl() after their waiters
+ * have completed. As I understand it, a second thread trying
+ * to issue will eventually wait on zl_issuer_lock, and then
+ * have no work to do and leave.
+ *
+ * If more lwbs had been created an issued between zil_crash()
+ * calls, then we probably just need to take those too, add
+ * them to the crash list and clean them up, but it complicates
+ * this function and I don't think it can happend.
+ */
+ ASSERT(list_is_empty(&zilog->zl_lwb_list));
+
+ mutex_exit(&zilog->zl_lock);
+ return;
+ }
+
+ zilog->zl_restart_txg = restart_txg;
+
+ /*
+ * Capture any live LWBs. Depending on the state of the pool they may
+ * represent in-flight IO that won't return for some time, and we want
+ * to make sure they don't get in the way of normal ZIL operation.
+ */
+ ASSERT(list_is_empty(&zilog->zl_lwb_crash_list));
+ list_move_tail(&zilog->zl_lwb_crash_list, &zilog->zl_lwb_list);
+
+ /*
+ * Run through the LWB list; erroring all itxes and signalling error
+ * to all waiters.
+ */
+ for (lwb_t *lwb = list_head(&zilog->zl_lwb_crash_list); lwb != NULL;
+ lwb = list_next(&zilog->zl_lwb_crash_list, lwb)) {
+ ASSERT(!(lwb->lwb_flags & LWB_FLAG_CRASHED));
+ lwb->lwb_flags |= LWB_FLAG_CRASHED;
+
+ itx_t *itx;
+ while ((itx = list_remove_head(&lwb->lwb_itxs)) != NULL)
+ zil_itx_destroy(itx, EIO);
+
+ zil_commit_waiter_t *zcw;
+ while ((zcw = list_remove_head(&lwb->lwb_waiters)) != NULL) {
+ mutex_enter(&zcw->zcw_lock);
+ zcw->zcw_lwb = NULL;
+ zcw->zcw_error = EIO;
+ zcw->zcw_done = B_TRUE;
+ cv_broadcast(&zcw->zcw_cv);
+ mutex_exit(&zcw->zcw_lock);
+ }
+ }
+
+ /*
+ * Zero the ZIL header bp after the ZIL restarts. We'll free it in
+ * zil_clean() when we clean up the lwbs.
+ */
+ zil_header_t *zh = zil_header_in_syncing_context(zilog);
+ BP_ZERO(&zh->zh_log);
+
+ /*
+ * Mark this ZIL dirty on the next txg, so that zil_clean() will be
+ * called for cleanup.
+ */
+ zilog_dirty(zilog, txg+1);
+
+ mutex_exit(&zilog->zl_lock);
+}
+
+/*
* Commit ZFS Intent Log transactions (itxs) to stable storage.
*
* When writing ZIL transactions to the on-disk representation of the
@@ -3640,9 +3949,17 @@ zil_commit_itx_assign(zilog_t *zilog, zil_commit_waiter_t *zcw)
* but the order in which they complete will be the same order in
* which they were created.
*/
-void
+static int zil_commit_impl(zilog_t *zilog, uint64_t foid);
+
+int
zil_commit(zilog_t *zilog, uint64_t foid)
{
+ return (zil_commit_flags(zilog, foid, ZIL_COMMIT_FAILMODE));
+}
+
+int
+zil_commit_flags(zilog_t *zilog, uint64_t foid, zil_commit_flag_t flags)
+{
/*
* We should never attempt to call zil_commit on a snapshot for
* a couple of reasons:
@@ -3659,7 +3976,7 @@ zil_commit(zilog_t *zilog, uint64_t foid)
ASSERT3B(dmu_objset_is_snapshot(zilog->zl_os), ==, B_FALSE);
if (zilog->zl_sync == ZFS_SYNC_DISABLED)
- return;
+ return (0);
if (!spa_writeable(zilog->zl_spa)) {
/*
@@ -3670,10 +3987,23 @@ zil_commit(zilog_t *zilog, uint64_t foid)
* verifying that truth before we return to the caller.
*/
ASSERT(list_is_empty(&zilog->zl_lwb_list));
- ASSERT3P(zilog->zl_last_lwb_opened, ==, NULL);
+ ASSERT0P(zilog->zl_last_lwb_opened);
for (int i = 0; i < TXG_SIZE; i++)
- ASSERT3P(zilog->zl_itxg[i].itxg_itxs, ==, NULL);
- return;
+ ASSERT0P(zilog->zl_itxg[i].itxg_itxs);
+ return (0);
+ }
+
+ int err = 0;
+
+ /*
+ * If the ZIL crashed, bypass it entirely, and rely on txg_wait_sync()
+ * to get the data out to disk.
+ */
+ if (zilog->zl_restart_txg > 0) {
+ ZIL_STAT_BUMP(zilog, zil_commit_crash_count);
+ err = txg_wait_synced_flags(zilog->zl_dmu_pool, 0,
+ TXG_WAIT_SUSPEND);
+ goto out;
}
/*
@@ -3685,14 +4015,43 @@ zil_commit(zilog_t *zilog, uint64_t foid)
*/
if (zilog->zl_suspend > 0) {
ZIL_STAT_BUMP(zilog, zil_commit_suspend_count);
- txg_wait_synced(zilog->zl_dmu_pool, 0);
- return;
+ err = txg_wait_synced_flags(zilog->zl_dmu_pool, 0,
+ TXG_WAIT_SUSPEND);
+ if (err != 0) {
+ ASSERT3U(err, ==, ESHUTDOWN);
+ zil_crash(zilog);
+ }
+ goto out;
}
- zil_commit_impl(zilog, foid);
+ err = zil_commit_impl(zilog, foid);
+
+out:
+ if (err == 0)
+ return (0);
+
+ /*
+ * The ZIL write failed and the pool is suspended. There's nothing else
+ * we can do except return or block.
+ */
+ ASSERT3U(err, ==, ESHUTDOWN);
+
+ /*
+ * Return error if failmode=continue or caller will handle directly.
+ */
+ if (!(flags & ZIL_COMMIT_FAILMODE) ||
+ spa_get_failmode(zilog->zl_spa) == ZIO_FAILURE_MODE_CONTINUE)
+ return (SET_ERROR(EIO));
+
+ /*
+ * Block until the pool returns. We assume that the data will make
+ * it out to disk in the end, and so return success.
+ */
+ txg_wait_synced(zilog->zl_dmu_pool, 0);
+ return (0);
}
-void
+static int
zil_commit_impl(zilog_t *zilog, uint64_t foid)
{
ZIL_STAT_BUMP(zilog, zil_commit_count);
@@ -3729,7 +4088,8 @@ zil_commit_impl(zilog_t *zilog, uint64_t foid)
uint64_t wtxg = zil_commit_writer(zilog, zcw);
zil_commit_waiter(zilog, zcw);
- if (zcw->zcw_zio_error != 0) {
+ int err = 0;
+ if (zcw->zcw_error != 0) {
/*
* If there was an error writing out the ZIL blocks that
* this thread is waiting on, then we fallback to
@@ -3741,13 +4101,29 @@ zil_commit_impl(zilog_t *zilog, uint64_t foid)
ZIL_STAT_BUMP(zilog, zil_commit_error_count);
DTRACE_PROBE2(zil__commit__io__error,
zilog_t *, zilog, zil_commit_waiter_t *, zcw);
- txg_wait_synced(zilog->zl_dmu_pool, 0);
+ err = txg_wait_synced_flags(zilog->zl_dmu_pool, 0,
+ TXG_WAIT_SUSPEND);
} else if (wtxg != 0) {
ZIL_STAT_BUMP(zilog, zil_commit_suspend_count);
- txg_wait_synced(zilog->zl_dmu_pool, wtxg);
+ err = txg_wait_synced_flags(zilog->zl_dmu_pool, wtxg,
+ TXG_WAIT_SUSPEND);
}
zil_free_commit_waiter(zcw);
+
+ if (err == 0)
+ return (0);
+
+ /*
+ * ZIL write failed and pool failed in the fallback to
+ * txg_wait_synced_flags(). Right now we have no idea if the data is on
+ * disk and the pool is probably suspended so we have no idea when it's
+ * coming back. All we can do is shut down and return error to the
+ * caller.
+ */
+ ASSERT3U(err, ==, ESHUTDOWN);
+ zil_crash(zilog);
+ return (err);
}
/*
@@ -3773,7 +4149,7 @@ zil_sync(zilog_t *zilog, dmu_tx_t *tx)
mutex_enter(&zilog->zl_lock);
- ASSERT(zilog->zl_stop_sync == 0);
+ ASSERT0(zilog->zl_stop_sync);
if (*replayed_seq != 0) {
ASSERT(zh->zh_replay_seq < *replayed_seq);
@@ -3848,7 +4224,7 @@ zil_lwb_cons(void *vbuf, void *unused, int kmflag)
offsetof(zil_commit_waiter_t, zcw_node));
avl_create(&lwb->lwb_vdev_tree, zil_lwb_vdev_compare,
sizeof (zil_vdev_node_t), offsetof(zil_vdev_node_t, zv_node));
- mutex_init(&lwb->lwb_vdev_lock, NULL, MUTEX_DEFAULT, NULL);
+ mutex_init(&lwb->lwb_lock, NULL, MUTEX_DEFAULT, NULL);
return (0);
}
@@ -3857,7 +4233,7 @@ zil_lwb_dest(void *vbuf, void *unused)
{
(void) unused;
lwb_t *lwb = vbuf;
- mutex_destroy(&lwb->lwb_vdev_lock);
+ mutex_destroy(&lwb->lwb_lock);
avl_destroy(&lwb->lwb_vdev_tree);
list_destroy(&lwb->lwb_waiters);
list_destroy(&lwb->lwb_itxs);
@@ -3943,6 +4319,8 @@ zil_alloc(objset_t *os, zil_header_t *zh_phys)
list_create(&zilog->zl_lwb_list, sizeof (lwb_t),
offsetof(lwb_t, lwb_node));
+ list_create(&zilog->zl_lwb_crash_list, sizeof (lwb_t),
+ offsetof(lwb_t, lwb_node));
list_create(&zilog->zl_itx_commit_list, sizeof (itx_t),
offsetof(itx_t, itx_node));
@@ -3967,9 +4345,12 @@ zil_free(zilog_t *zilog)
ASSERT0(zilog->zl_suspend);
ASSERT0(zilog->zl_suspending);
+ ASSERT0(zilog->zl_restart_txg);
ASSERT(list_is_empty(&zilog->zl_lwb_list));
list_destroy(&zilog->zl_lwb_list);
+ ASSERT(list_is_empty(&zilog->zl_lwb_crash_list));
+ list_destroy(&zilog->zl_lwb_crash_list);
ASSERT(list_is_empty(&zilog->zl_itx_commit_list));
list_destroy(&zilog->zl_itx_commit_list);
@@ -4005,8 +4386,8 @@ zil_open(objset_t *os, zil_get_data_t *get_data, zil_sums_t *zil_sums)
{
zilog_t *zilog = dmu_objset_zil(os);
- ASSERT3P(zilog->zl_get_data, ==, NULL);
- ASSERT3P(zilog->zl_last_lwb_opened, ==, NULL);
+ ASSERT0P(zilog->zl_get_data);
+ ASSERT0P(zilog->zl_last_lwb_opened);
ASSERT(list_is_empty(&zilog->zl_lwb_list));
zilog->zl_get_data = get_data;
@@ -4025,7 +4406,8 @@ zil_close(zilog_t *zilog)
uint64_t txg;
if (!dmu_objset_is_snapshot(zilog->zl_os)) {
- zil_commit(zilog, 0);
+ if (zil_commit_flags(zilog, 0, ZIL_COMMIT_NOW) != 0)
+ txg_wait_synced(zilog->zl_dmu_pool, 0);
} else {
ASSERT(list_is_empty(&zilog->zl_lwb_list));
ASSERT0(zilog->zl_dirty_max_txg);
@@ -4074,7 +4456,7 @@ zil_close(zilog_t *zilog)
if (lwb != NULL) {
ASSERT(list_is_empty(&zilog->zl_lwb_list));
ASSERT3S(lwb->lwb_state, ==, LWB_STATE_NEW);
- zio_buf_free(lwb->lwb_buf, lwb->lwb_sz);
+ ASSERT0P(lwb->lwb_buf);
zil_free_lwb(zilog, lwb);
}
mutex_exit(&zilog->zl_lock);
@@ -4126,6 +4508,17 @@ zil_suspend(const char *osname, void **cookiep)
return (SET_ERROR(EBUSY));
}
+ if (zilog->zl_restart_txg > 0) {
+ /*
+ * ZIL crashed. It effectively _is_ suspended, but callers
+ * are usually trying to make sure it's empty on-disk, which
+ * we can't guarantee right now.
+ */
+ mutex_exit(&zilog->zl_lock);
+ dmu_objset_rele(os, suspend_tag);
+ return (SET_ERROR(EBUSY));
+ }
+
/*
* Don't put a long hold in the cases where we can avoid it. This
* is when there is no cookie so we are doing a suspend & resume
@@ -4154,11 +4547,16 @@ zil_suspend(const char *osname, void **cookiep)
cv_wait(&zilog->zl_cv_suspend, &zilog->zl_lock);
mutex_exit(&zilog->zl_lock);
- if (cookiep == NULL)
+ if (zilog->zl_restart_txg > 0) {
+ /* ZIL crashed while we were waiting. */
+ zil_resume(os);
+ error = SET_ERROR(EBUSY);
+ } else if (cookiep == NULL)
zil_resume(os);
else
*cookiep = os;
- return (0);
+
+ return (error);
}
/*
@@ -4199,17 +4597,34 @@ zil_suspend(const char *osname, void **cookiep)
* would just call txg_wait_synced(), because zl_suspend is set.
* txg_wait_synced() doesn't wait for these lwb's to be
* LWB_STATE_FLUSH_DONE before returning.
+ *
+ * However, zil_commit_impl() itself can return an error if any of the
+ * lwbs fail, or the pool suspends in the fallback
+ * txg_wait_sync_flushed(), which affects what we do next, so we
+ * capture that error.
*/
- zil_commit_impl(zilog, 0);
+ error = zil_commit_impl(zilog, 0);
+ if (error == ESHUTDOWN)
+ /* zil_commit_impl() has called zil_crash() already */
+ error = SET_ERROR(EBUSY);
/*
* Now that we've ensured all lwb's are LWB_STATE_FLUSH_DONE, we
* use txg_wait_synced() to ensure the data from the zilog has
* migrated to the main pool before calling zil_destroy().
*/
- txg_wait_synced(zilog->zl_dmu_pool, 0);
+ if (error == 0) {
+ error = txg_wait_synced_flags(zilog->zl_dmu_pool, 0,
+ TXG_WAIT_SUSPEND);
+ if (error != 0) {
+ ASSERT3U(error, ==, ESHUTDOWN);
+ zil_crash(zilog);
+ error = SET_ERROR(EBUSY);
+ }
+ }
- zil_destroy(zilog, B_FALSE);
+ if (error == 0)
+ zil_destroy(zilog, B_FALSE);
mutex_enter(&zilog->zl_lock);
zilog->zl_suspending = B_FALSE;
@@ -4223,7 +4638,8 @@ zil_suspend(const char *osname, void **cookiep)
zil_resume(os);
else
*cookiep = os;
- return (0);
+
+ return (error);
}
void
@@ -4386,7 +4802,7 @@ zil_replay(objset_t *os, void *arg,
zilog->zl_replay = B_TRUE;
zilog->zl_replay_time = ddi_get_lbolt();
- ASSERT(zilog->zl_replay_blks == 0);
+ ASSERT0(zilog->zl_replay_blks);
(void) zil_parse(zilog, zil_incr_blks, zil_replay_log_record, &zr,
zh->zh_claim_txg, B_TRUE);
vmem_free(zr.zr_lr, 2 * SPA_MAXBLOCKSIZE);
diff --git a/sys/contrib/openzfs/module/zfs/zio.c b/sys/contrib/openzfs/module/zfs/zio.c
index 218aec6093e2..4cf8912d4269 100644
--- a/sys/contrib/openzfs/module/zfs/zio.c
+++ b/sys/contrib/openzfs/module/zfs/zio.c
@@ -339,8 +339,8 @@ zio_fini(void)
}
for (size_t i = 0; i < n; i++) {
- VERIFY3P(zio_buf_cache[i], ==, NULL);
- VERIFY3P(zio_data_buf_cache[i], ==, NULL);
+ VERIFY0P(zio_buf_cache[i]);
+ VERIFY0P(zio_data_buf_cache[i]);
}
if (zio_ksp != NULL) {
@@ -771,7 +771,7 @@ zio_add_child_impl(zio_t *pio, zio_t *cio, boolean_t first)
else
mutex_enter(&cio->io_lock);
- ASSERT(pio->io_state[ZIO_WAIT_DONE] == 0);
+ ASSERT0(pio->io_state[ZIO_WAIT_DONE]);
uint64_t *countp = pio->io_children[cio->io_child_type];
for (int w = 0; w < ZIO_WAIT_TYPES; w++)
@@ -821,7 +821,7 @@ zio_wait_for_children(zio_t *zio, uint8_t childbits, enum zio_wait_type wait)
boolean_t waiting = B_FALSE;
mutex_enter(&zio->io_lock);
- ASSERT(zio->io_stall == NULL);
+ ASSERT0P(zio->io_stall);
for (int c = 0; c < ZIO_CHILD_TYPES; c++) {
if (!(ZIO_CHILD_BIT_IS_SET(childbits, c)))
continue;
@@ -955,8 +955,8 @@ zio_create(zio_t *pio, spa_t *spa, uint64_t txg, const blkptr_t *bp,
zio_t *zio;
IMPLY(type != ZIO_TYPE_TRIM, psize <= SPA_MAXBLOCKSIZE);
- ASSERT(P2PHASE(psize, SPA_MINBLOCKSIZE) == 0);
- ASSERT(P2PHASE(offset, SPA_MINBLOCKSIZE) == 0);
+ ASSERT0(P2PHASE(psize, SPA_MINBLOCKSIZE));
+ ASSERT0(P2PHASE(offset, SPA_MINBLOCKSIZE));
ASSERT(!vd || spa_config_held(spa, SCL_STATE_ALL, RW_READER));
ASSERT(!bp || !(flags & ZIO_FLAG_CONFIG_WRITER));
@@ -1451,7 +1451,7 @@ zio_free(spa_t *spa, uint64_t txg, const blkptr_t *bp)
metaslab_check_free(spa, bp);
bplist_append(&spa->spa_free_bplist[txg & TXG_MASK], bp);
} else {
- VERIFY3P(zio_free_sync(NULL, spa, txg, bp, 0), ==, NULL);
+ VERIFY0P(zio_free_sync(NULL, spa, txg, bp, 0));
}
}
@@ -1559,7 +1559,7 @@ zio_read_phys(zio_t *pio, vdev_t *vd, uint64_t offset, uint64_t size,
{
zio_t *zio;
- ASSERT(vd->vdev_children == 0);
+ ASSERT0(vd->vdev_children);
ASSERT(!labels || offset + size <= VDEV_LABEL_START_SIZE ||
offset >= vd->vdev_psize - VDEV_LABEL_END_SIZE);
ASSERT3U(offset + size, <=, vd->vdev_psize);
@@ -1580,7 +1580,7 @@ zio_write_phys(zio_t *pio, vdev_t *vd, uint64_t offset, uint64_t size,
{
zio_t *zio;
- ASSERT(vd->vdev_children == 0);
+ ASSERT0(vd->vdev_children);
ASSERT(!labels || offset + size <= VDEV_LABEL_START_SIZE ||
offset >= vd->vdev_psize - VDEV_LABEL_END_SIZE);
ASSERT3U(offset + size, <=, vd->vdev_psize);
@@ -1747,7 +1747,7 @@ zio_flush(zio_t *pio, vdev_t *vd)
void
zio_shrink(zio_t *zio, uint64_t size)
{
- ASSERT3P(zio->io_executor, ==, NULL);
+ ASSERT0P(zio->io_executor);
ASSERT3U(zio->io_orig_size, ==, zio->io_size);
ASSERT3U(size, <=, zio->io_size);
@@ -1941,7 +1941,7 @@ zio_write_compress(zio_t *zio)
}
ASSERT(zio->io_child_type != ZIO_CHILD_DDT);
- ASSERT(zio->io_bp_override == NULL);
+ ASSERT0P(zio->io_bp_override);
if (!BP_IS_HOLE(bp) && BP_GET_BIRTH(bp) == zio->io_txg) {
/*
@@ -2436,7 +2436,7 @@ __zio_execute(zio_t *zio)
ASSERT(!MUTEX_HELD(&zio->io_lock));
ASSERT(ISP2(stage));
- ASSERT(zio->io_stall == NULL);
+ ASSERT0P(zio->io_stall);
do {
stage <<= 1;
@@ -2509,7 +2509,7 @@ zio_wait(zio_t *zio)
int error;
ASSERT3S(zio->io_stage, ==, ZIO_STAGE_OPEN);
- ASSERT3P(zio->io_executor, ==, NULL);
+ ASSERT0P(zio->io_executor);
zio->io_waiter = curthread;
ASSERT0(zio->io_queued_timestamp);
@@ -2551,7 +2551,7 @@ zio_nowait(zio_t *zio)
if (zio == NULL)
return;
- ASSERT3P(zio->io_executor, ==, NULL);
+ ASSERT0P(zio->io_executor);
if (zio->io_child_type == ZIO_CHILD_LOGICAL &&
list_is_empty(&zio->io_parent_list)) {
@@ -2590,8 +2590,8 @@ zio_reexecute(void *arg)
ASSERT(pio->io_child_type == ZIO_CHILD_LOGICAL);
ASSERT(pio->io_orig_stage == ZIO_STAGE_OPEN);
- ASSERT(pio->io_gang_leader == NULL);
- ASSERT(pio->io_gang_tree == NULL);
+ ASSERT0P(pio->io_gang_leader);
+ ASSERT0P(pio->io_gang_tree);
mutex_enter(&pio->io_lock);
pio->io_flags = pio->io_orig_flags;
@@ -2689,7 +2689,7 @@ zio_suspend(spa_t *spa, zio_t *zio, zio_suspend_reason_t reason)
ASSERT(!(zio->io_flags & ZIO_FLAG_GODFATHER));
ASSERT(zio != spa->spa_suspend_zio_root);
ASSERT(zio->io_child_type == ZIO_CHILD_LOGICAL);
- ASSERT(zio_unique_parent(zio) == NULL);
+ ASSERT0P(zio_unique_parent(zio));
ASSERT(zio->io_stage == ZIO_STAGE_DONE);
zio_add_child(spa->spa_suspend_zio_root, zio);
}
@@ -2908,7 +2908,7 @@ zio_gang_node_alloc(zio_gang_node_t **gnpp, uint64_t gangblocksize)
{
zio_gang_node_t *gn;
- ASSERT(*gnpp == NULL);
+ ASSERT0P(*gnpp);
gn = kmem_zalloc(sizeof (*gn) +
(gbh_nblkptrs(gangblocksize) * sizeof (gn)), KM_SLEEP);
@@ -2925,7 +2925,7 @@ zio_gang_node_free(zio_gang_node_t **gnpp)
zio_gang_node_t *gn = *gnpp;
for (int g = 0; g < gbh_nblkptrs(gn->gn_allocsize); g++)
- ASSERT(gn->gn_child[g] == NULL);
+ ASSERT0P(gn->gn_child[g]);
zio_buf_free(gn->gn_gbh, gn->gn_allocsize);
kmem_free(gn, sizeof (*gn) +
@@ -3362,11 +3362,11 @@ zio_nop_write(zio_t *zio)
zio_prop_t *zp = &zio->io_prop;
ASSERT(BP_IS_HOLE(bp));
- ASSERT(BP_GET_LEVEL(bp) == 0);
+ ASSERT0(BP_GET_LEVEL(bp));
ASSERT(!(zio->io_flags & ZIO_FLAG_IO_REWRITE));
ASSERT(zp->zp_nopwrite);
ASSERT(!zp->zp_dedup);
- ASSERT(zio->io_bp_override == NULL);
+ ASSERT0P(zio->io_bp_override);
ASSERT(IO_IS_ALLOCATING(zio));
/*
@@ -3495,7 +3495,7 @@ zio_ddt_read_start(zio_t *zio)
ddt_univ_phys_t *ddp = dde->dde_phys;
blkptr_t blk;
- ASSERT(zio->io_vsd == NULL);
+ ASSERT0P(zio->io_vsd);
zio->io_vsd = dde;
if (v_self == DDT_PHYS_NONE)
@@ -3560,7 +3560,7 @@ zio_ddt_read_done(zio_t *zio)
zio->io_vsd = NULL;
}
- ASSERT(zio->io_vsd == NULL);
+ ASSERT0P(zio->io_vsd);
return (zio);
}
@@ -4415,7 +4415,7 @@ static void
zio_dva_unallocate(zio_t *zio, zio_gang_node_t *gn, blkptr_t *bp)
{
ASSERT(BP_GET_BIRTH(bp) == zio->io_txg || BP_IS_HOLE(bp));
- ASSERT(zio->io_bp_override == NULL);
+ ASSERT0P(zio->io_bp_override);
if (!BP_IS_HOLE(bp)) {
metaslab_free(zio->io_spa, bp, BP_GET_BIRTH(bp), B_TRUE);
@@ -4434,12 +4434,15 @@ zio_dva_unallocate(zio_t *zio, zio_gang_node_t *gn, blkptr_t *bp)
*/
int
zio_alloc_zil(spa_t *spa, objset_t *os, uint64_t txg, blkptr_t *new_bp,
- uint64_t size, boolean_t *slog)
+ uint64_t min_size, uint64_t max_size, boolean_t *slog,
+ boolean_t allow_larger)
{
int error;
zio_alloc_list_t io_alloc_list;
+ uint64_t alloc_size = 0;
ASSERT(txg > spa_syncing_txg(spa));
+ ASSERT3U(min_size, <=, max_size);
metaslab_trace_init(&io_alloc_list);
@@ -4448,7 +4451,7 @@ zio_alloc_zil(spa_t *spa, objset_t *os, uint64_t txg, blkptr_t *new_bp,
* Fill in the obvious ones before calling into metaslab_alloc().
*/
BP_SET_TYPE(new_bp, DMU_OT_INTENT_LOG);
- BP_SET_PSIZE(new_bp, size);
+ BP_SET_PSIZE(new_bp, max_size);
BP_SET_LEVEL(new_bp, 0);
/*
@@ -4463,43 +4466,51 @@ zio_alloc_zil(spa_t *spa, objset_t *os, uint64_t txg, blkptr_t *new_bp,
ZIOSTAT_BUMP(ziostat_total_allocations);
/* Try log class (dedicated slog devices) first */
- error = metaslab_alloc(spa, spa_log_class(spa), size, new_bp, 1,
- txg, NULL, flags, &io_alloc_list, allocator, NULL);
+ error = metaslab_alloc_range(spa, spa_log_class(spa), min_size,
+ max_size, new_bp, 1, txg, NULL, flags, &io_alloc_list, allocator,
+ NULL, &alloc_size);
*slog = (error == 0);
/* Try special_embedded_log class (reserved on special vdevs) */
if (error != 0) {
- error = metaslab_alloc(spa, spa_special_embedded_log_class(spa),
- size, new_bp, 1, txg, NULL, flags, &io_alloc_list,
- allocator, NULL);
+ error = metaslab_alloc_range(spa,
+ spa_special_embedded_log_class(spa), min_size, max_size,
+ new_bp, 1, txg, NULL, flags, &io_alloc_list, allocator,
+ NULL, &alloc_size);
}
/* Try special class (general special vdev allocation) */
if (error != 0) {
- error = metaslab_alloc(spa, spa_special_class(spa), size,
- new_bp, 1, txg, NULL, flags, &io_alloc_list, allocator,
- NULL);
+ error = metaslab_alloc_range(spa, spa_special_class(spa),
+ min_size, max_size, new_bp, 1, txg, NULL, flags,
+ &io_alloc_list, allocator, NULL, &alloc_size);
}
/* Try embedded_log class (reserved on normal vdevs) */
if (error != 0) {
- error = metaslab_alloc(spa, spa_embedded_log_class(spa), size,
- new_bp, 1, txg, NULL, flags, &io_alloc_list, allocator,
- NULL);
+ error = metaslab_alloc_range(spa, spa_embedded_log_class(spa),
+ min_size, max_size, new_bp, 1, txg, NULL, flags,
+ &io_alloc_list, allocator, NULL, &alloc_size);
}
/* Finally fall back to normal class */
if (error != 0) {
ZIOSTAT_BUMP(ziostat_alloc_class_fallbacks);
- error = metaslab_alloc(spa, spa_normal_class(spa), size,
- new_bp, 1, txg, NULL, flags, &io_alloc_list, allocator,
- NULL);
+ error = metaslab_alloc_range(spa, spa_normal_class(spa),
+ min_size, max_size, new_bp, 1, txg, NULL, flags,
+ &io_alloc_list, allocator, NULL, &alloc_size);
}
metaslab_trace_fini(&io_alloc_list);
if (error == 0) {
- BP_SET_LSIZE(new_bp, size);
- BP_SET_PSIZE(new_bp, size);
+ if (!allow_larger)
+ alloc_size = MIN(alloc_size, max_size);
+ else if (max_size <= SPA_OLD_MAXBLOCKSIZE)
+ alloc_size = MIN(alloc_size, SPA_OLD_MAXBLOCKSIZE);
+ alloc_size = P2ALIGN_TYPED(alloc_size, ZIL_MIN_BLKSZ, uint64_t);
+
+ BP_SET_LSIZE(new_bp, alloc_size);
+ BP_SET_PSIZE(new_bp, alloc_size);
BP_SET_COMPRESS(new_bp, ZIO_COMPRESS_OFF);
BP_SET_CHECKSUM(new_bp,
spa_version(spa) >= SPA_VERSION_SLIM_ZIL
@@ -4527,8 +4538,8 @@ zio_alloc_zil(spa_t *spa, objset_t *os, uint64_t txg, blkptr_t *new_bp,
}
} else {
zfs_dbgmsg("%s: zil block allocation failure: "
- "size %llu, error %d", spa_name(spa), (u_longlong_t)size,
- error);
+ "min_size %llu, max_size %llu, error %d", spa_name(spa),
+ (u_longlong_t)min_size, (u_longlong_t)max_size, error);
}
return (error);
@@ -4559,8 +4570,8 @@ zio_vdev_io_start(zio_t *zio)
zio->io_delay = 0;
- ASSERT(zio->io_error == 0);
- ASSERT(zio->io_child_error[ZIO_CHILD_VDEV] == 0);
+ ASSERT0(zio->io_error);
+ ASSERT0(zio->io_child_error[ZIO_CHILD_VDEV]);
if (vd == NULL) {
if (!(zio->io_flags & ZIO_FLAG_CONFIG_WRITER))
@@ -4751,7 +4762,7 @@ zio_vdev_io_done(zio_t *zio)
ops->vdev_op_io_done(zio);
if (unexpected_error && vd->vdev_remove_wanted == B_FALSE)
- VERIFY(vdev_probe(vd, zio) == NULL);
+ VERIFY0P(vdev_probe(vd, zio));
return (zio);
}
@@ -4903,7 +4914,7 @@ void
zio_vdev_io_reissue(zio_t *zio)
{
ASSERT(zio->io_stage == ZIO_STAGE_VDEV_IO_START);
- ASSERT(zio->io_error == 0);
+ ASSERT0(zio->io_error);
zio->io_stage >>= 1;
}
@@ -4920,7 +4931,7 @@ void
zio_vdev_io_bypass(zio_t *zio)
{
ASSERT(zio->io_stage == ZIO_STAGE_VDEV_IO_START);
- ASSERT(zio->io_error == 0);
+ ASSERT0(zio->io_error);
zio->io_flags |= ZIO_FLAG_IO_BYPASS;
zio->io_stage = ZIO_STAGE_VDEV_IO_ASSESS >> 1;
@@ -5298,7 +5309,7 @@ zio_ready(zio_t *zio)
ASSERT(IO_IS_ALLOCATING(zio));
ASSERT(BP_GET_BIRTH(bp) == zio->io_txg ||
BP_IS_HOLE(bp) || (zio->io_flags & ZIO_FLAG_NOPWRITE));
- ASSERT(zio->io_children[ZIO_CHILD_GANG][ZIO_WAIT_READY] == 0);
+ ASSERT0(zio->io_children[ZIO_CHILD_GANG][ZIO_WAIT_READY]);
zio->io_ready(zio);
}
@@ -5448,7 +5459,7 @@ zio_done(zio_t *zio)
for (int c = 0; c < ZIO_CHILD_TYPES; c++)
for (int w = 0; w < ZIO_WAIT_TYPES; w++)
- ASSERT(zio->io_children[c][w] == 0);
+ ASSERT0(zio->io_children[c][w]);
if (zio->io_bp != NULL && !BP_IS_EMBEDDED(zio->io_bp)) {
ASSERT(memcmp(zio->io_bp, &zio->io_bp_copy,
diff --git a/sys/contrib/openzfs/module/zfs/zio_checksum.c b/sys/contrib/openzfs/module/zfs/zio_checksum.c
index 63d0c6dadd46..1d0646a61185 100644
--- a/sys/contrib/openzfs/module/zfs/zio_checksum.c
+++ b/sys/contrib/openzfs/module/zfs/zio_checksum.c
@@ -215,7 +215,7 @@ zio_checksum_info_t zio_checksum_table[ZIO_CHECKSUM_FUNCTIONS] = {
spa_feature_t
zio_checksum_to_feature(enum zio_checksum cksum)
{
- VERIFY((cksum & ~ZIO_CHECKSUM_MASK) == 0);
+ VERIFY0((cksum & ~ZIO_CHECKSUM_MASK));
switch (cksum) {
case ZIO_CHECKSUM_BLAKE3:
diff --git a/sys/contrib/openzfs/module/zfs/zio_compress.c b/sys/contrib/openzfs/module/zfs/zio_compress.c
index 9f0ac1b63146..89ceeb58ad91 100644
--- a/sys/contrib/openzfs/module/zfs/zio_compress.c
+++ b/sys/contrib/openzfs/module/zfs/zio_compress.c
@@ -38,12 +38,6 @@
#include <sys/zstd/zstd.h>
/*
- * If nonzero, every 1/X decompression attempts will fail, simulating
- * an undetected memory error.
- */
-static unsigned long zio_decompress_fail_fraction = 0;
-
-/*
* Compression vectors.
*/
zio_compress_info_t zio_compress_table[ZIO_COMPRESS_FUNCTIONS] = {
@@ -171,15 +165,6 @@ zio_decompress_data(enum zio_compress c, abd_t *src, abd_t *dst,
else
err = ci->ci_decompress(src, dst, s_len, d_len, ci->ci_level);
- /*
- * Decompression shouldn't fail, because we've already verified
- * the checksum. However, for extra protection (e.g. against bitflips
- * in non-ECC RAM), we handle this error (and test it).
- */
- if (zio_decompress_fail_fraction != 0 &&
- random_in_range(zio_decompress_fail_fraction) == 0)
- err = SET_ERROR(EINVAL);
-
return (err);
}
diff --git a/sys/contrib/openzfs/module/zfs/zio_inject.c b/sys/contrib/openzfs/module/zfs/zio_inject.c
index df7b01ba879e..981a1be4847c 100644
--- a/sys/contrib/openzfs/module/zfs/zio_inject.c
+++ b/sys/contrib/openzfs/module/zfs/zio_inject.c
@@ -1119,7 +1119,7 @@ zio_clear_fault(int id)
kmem_free(handler->zi_lanes, sizeof (*handler->zi_lanes) *
handler->zi_record.zi_nlanes);
} else {
- ASSERT3P(handler->zi_lanes, ==, NULL);
+ ASSERT0P(handler->zi_lanes);
}
if (handler->zi_spa_name != NULL)
diff --git a/sys/contrib/openzfs/module/zfs/zrlock.c b/sys/contrib/openzfs/module/zfs/zrlock.c
index 3c0f1b7bbbc1..09c110945c97 100644
--- a/sys/contrib/openzfs/module/zfs/zrlock.c
+++ b/sys/contrib/openzfs/module/zfs/zrlock.c
@@ -129,7 +129,7 @@ zrl_tryenter(zrlock_t *zrl)
(uint32_t *)&zrl->zr_refcount, 0, ZRL_LOCKED);
if (cas == 0) {
#ifdef ZFS_DEBUG
- ASSERT3P(zrl->zr_owner, ==, NULL);
+ ASSERT0P(zrl->zr_owner);
zrl->zr_owner = curthread;
#endif
return (1);
diff --git a/sys/contrib/openzfs/module/zfs/zthr.c b/sys/contrib/openzfs/module/zfs/zthr.c
index 597a510528ea..d245ce4946e0 100644
--- a/sys/contrib/openzfs/module/zfs/zthr.c
+++ b/sys/contrib/openzfs/module/zfs/zthr.c
@@ -316,7 +316,7 @@ zthr_destroy(zthr_t *t)
{
ASSERT(!MUTEX_HELD(&t->zthr_state_lock));
ASSERT(!MUTEX_HELD(&t->zthr_request_lock));
- VERIFY3P(t->zthr_thread, ==, NULL);
+ VERIFY0P(t->zthr_thread);
mutex_destroy(&t->zthr_request_lock);
mutex_destroy(&t->zthr_state_lock);
cv_destroy(&t->zthr_cv);
diff --git a/sys/contrib/openzfs/module/zfs/zvol.c b/sys/contrib/openzfs/module/zfs/zvol.c
index 7e264f308cf2..2fd3e1c37045 100644
--- a/sys/contrib/openzfs/module/zfs/zvol.c
+++ b/sys/contrib/openzfs/module/zfs/zvol.c
@@ -38,25 +38,36 @@
* Copyright 2014 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2016 Actifio, Inc. All rights reserved.
* Copyright (c) 2012, 2019 by Delphix. All rights reserved.
- * Copyright (c) 2024, Klara, Inc.
+ * Copyright (c) 2024, 2025, Klara, Inc.
*/
/*
* Note on locking of zvol state structures.
*
- * These structures are used to maintain internal state used to emulate block
- * devices on top of zvols. In particular, management of device minor number
- * operations - create, remove, rename, and set_snapdev - involves access to
- * these structures. The zvol_state_lock is primarily used to protect the
- * zvol_state_list. The zv->zv_state_lock is used to protect the contents
- * of the zvol_state_t structures, as well as to make sure that when the
- * time comes to remove the structure from the list, it is not in use, and
- * therefore, it can be taken off zvol_state_list and freed.
+ * zvol_state_t represents the connection between a single dataset
+ * (DMU_OST_ZVOL) and the device "minor" (some OS-specific representation of a
+ * "disk" or "device" or "volume", eg, a /dev/zdXX node, a GEOM object, etc).
*
- * The zv_suspend_lock was introduced to allow for suspending I/O to a zvol,
- * e.g. for the duration of receive and rollback operations. This lock can be
- * held for significant periods of time. Given that it is undesirable to hold
- * mutexes for long periods of time, the following lock ordering applies:
+ * The global zvol_state_lock is used to protect access to zvol_state_list and
+ * zvol_htable, which are the primary way to obtain a zvol_state_t from a name.
+ * It should not be used for anything not name-relateds, and you should avoid
+ * sleeping or waiting while its held. See zvol_find_by_name(), zvol_insert(),
+ * zvol_remove().
+ *
+ * The zv_state_lock is used to protect the contents of the associated
+ * zvol_state_t. Most of the zvol_state_t is dedicated to control and
+ * configuration; almost none of it is needed for data operations (that is,
+ * read, write, flush) so this lock is rarely taken during general IO. It
+ * should be released quickly; you should avoid sleeping or waiting while its
+ * held.
+ *
+ * zv_suspend_lock is used to suspend IO/data operations to a zvol. The read
+ * half should held for the duration of an IO operation. The write half should
+ * be taken when something to wait for IO to complete and the block further IO,
+ * eg for the duration of receive and rollback operations. This lock can be
+ * held for long periods of time.
+ *
+ * Thus, the following lock ordering appies.
* - take zvol_state_lock if necessary, to protect zvol_state_list
* - take zv_suspend_lock if necessary, by the code path in question
* - take zv_state_lock to protect zvol_state_t
@@ -67,9 +78,8 @@
* these operations are serialized per pool. Consequently, we can be certain
* that for a given zvol, there is only one operation at a time in progress.
* That is why one can be sure that first, zvol_state_t for a given zvol is
- * allocated and placed on zvol_state_list, and then other minor operations
- * for this zvol are going to proceed in the order of issue.
- *
+ * allocated and placed on zvol_state_list, and then other minor operations for
+ * this zvol are going to proceed in the order of issue.
*/
#include <sys/dataset_kstats.h>
@@ -215,8 +225,8 @@ zvol_create_cb(objset_t *os, void *arg, cred_t *cr, dmu_tx_t *tx)
int error;
uint64_t volblocksize, volsize;
- VERIFY(nvlist_lookup_uint64(nvprops,
- zfs_prop_to_name(ZFS_PROP_VOLSIZE), &volsize) == 0);
+ VERIFY0(nvlist_lookup_uint64(nvprops,
+ zfs_prop_to_name(ZFS_PROP_VOLSIZE), &volsize));
if (nvlist_lookup_uint64(nvprops,
zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE), &volblocksize) != 0)
volblocksize = zfs_prop_default_numeric(ZFS_PROP_VOLBLOCKSIZE);
@@ -225,21 +235,20 @@ zvol_create_cb(objset_t *os, void *arg, cred_t *cr, dmu_tx_t *tx)
* These properties must be removed from the list so the generic
* property setting step won't apply to them.
*/
- VERIFY(nvlist_remove_all(nvprops,
- zfs_prop_to_name(ZFS_PROP_VOLSIZE)) == 0);
+ VERIFY0(nvlist_remove_all(nvprops, zfs_prop_to_name(ZFS_PROP_VOLSIZE)));
(void) nvlist_remove_all(nvprops,
zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE));
error = dmu_object_claim(os, ZVOL_OBJ, DMU_OT_ZVOL, volblocksize,
DMU_OT_NONE, 0, tx);
- ASSERT(error == 0);
+ ASSERT0(error);
error = zap_create_claim(os, ZVOL_ZAP_OBJ, DMU_OT_ZVOL_PROP,
DMU_OT_NONE, 0, tx);
- ASSERT(error == 0);
+ ASSERT0(error);
error = zap_update(os, ZVOL_ZAP_OBJ, "size", 8, 1, &volsize, tx);
- ASSERT(error == 0);
+ ASSERT0(error);
}
/*
@@ -254,7 +263,7 @@ zvol_get_stats(objset_t *os, nvlist_t *nv)
error = zap_lookup(os, ZVOL_ZAP_OBJ, "size", 8, 1, &val);
if (error)
- return (SET_ERROR(error));
+ return (error);
dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_VOLSIZE, val);
doi = kmem_alloc(sizeof (dmu_object_info_t), KM_SLEEP);
@@ -267,7 +276,7 @@ zvol_get_stats(objset_t *os, nvlist_t *nv)
kmem_free(doi, sizeof (dmu_object_info_t));
- return (SET_ERROR(error));
+ return (error);
}
/*
@@ -305,7 +314,7 @@ zvol_update_volsize(uint64_t volsize, objset_t *os)
error = dmu_tx_assign(tx, DMU_TX_WAIT);
if (error) {
dmu_tx_abort(tx);
- return (SET_ERROR(error));
+ return (error);
}
txg = dmu_tx_get_txg(tx);
@@ -337,7 +346,7 @@ zvol_set_volsize(const char *name, uint64_t volsize)
error = dsl_prop_get_integer(name,
zfs_prop_to_name(ZFS_PROP_READONLY), &readonly, NULL);
if (error != 0)
- return (SET_ERROR(error));
+ return (error);
if (readonly)
return (SET_ERROR(EROFS));
@@ -353,7 +362,7 @@ zvol_set_volsize(const char *name, uint64_t volsize)
FTAG, &os)) != 0) {
if (zv != NULL)
mutex_exit(&zv->zv_state_lock);
- return (SET_ERROR(error));
+ return (error);
}
owned = B_TRUE;
if (zv != NULL)
@@ -390,7 +399,7 @@ out:
if (error == 0 && zv != NULL)
zvol_os_update_volsize(zv, volsize);
- return (SET_ERROR(error));
+ return (error);
}
/*
@@ -401,7 +410,7 @@ zvol_set_volthreading(const char *name, boolean_t value)
{
zvol_state_t *zv = zvol_find_by_name(name, RW_NONE);
if (zv == NULL)
- return (ENOENT);
+ return (SET_ERROR(ENOENT));
zv->zv_threading = value;
mutex_exit(&zv->zv_state_lock);
return (0);
@@ -450,8 +459,10 @@ zvol_check_volblocksize(const char *name, uint64_t volblocksize)
* We don't allow setting the property above 1MB,
* unless the tunable has been changed.
*/
- if (volblocksize > zfs_max_recordsize)
+ if (volblocksize > zfs_max_recordsize) {
+ spa_close(spa, FTAG);
return (SET_ERROR(EDOM));
+ }
spa_close(spa, FTAG);
}
@@ -618,7 +629,7 @@ zvol_clone_range(zvol_state_t *zv_src, uint64_t inoff, zvol_state_t *zv_dst,
dmu_tx_t *tx;
blkptr_t *bps;
size_t maxblocks;
- int error = EINVAL;
+ int error = 0;
rw_enter(&zv_dst->zv_suspend_lock, RW_READER);
if (zv_dst->zv_zilog == NULL) {
@@ -644,23 +655,22 @@ zvol_clone_range(zvol_state_t *zv_src, uint64_t inoff, zvol_state_t *zv_dst,
*/
if (!spa_feature_is_enabled(dmu_objset_spa(outos),
SPA_FEATURE_BLOCK_CLONING)) {
- error = EOPNOTSUPP;
+ error = SET_ERROR(EOPNOTSUPP);
goto out;
}
if (dmu_objset_spa(inos) != dmu_objset_spa(outos)) {
- error = EXDEV;
+ error = SET_ERROR(EXDEV);
goto out;
}
if (inos->os_encrypted != outos->os_encrypted) {
- error = EXDEV;
+ error = SET_ERROR(EXDEV);
goto out;
}
if (zv_src->zv_volblocksize != zv_dst->zv_volblocksize) {
- error = EINVAL;
+ error = SET_ERROR(EINVAL);
goto out;
}
if (inoff >= zv_src->zv_volsize || outoff >= zv_dst->zv_volsize) {
- error = 0;
goto out;
}
@@ -671,17 +681,15 @@ zvol_clone_range(zvol_state_t *zv_src, uint64_t inoff, zvol_state_t *zv_dst,
len = zv_src->zv_volsize - inoff;
if (len > zv_dst->zv_volsize - outoff)
len = zv_dst->zv_volsize - outoff;
- if (len == 0) {
- error = 0;
+ if (len == 0)
goto out;
- }
/*
* No overlapping if we are cloning within the same file
*/
if (zv_src == zv_dst) {
if (inoff < outoff + len && outoff < inoff + len) {
- error = EINVAL;
+ error = SET_ERROR(EINVAL);
goto out;
}
}
@@ -691,7 +699,7 @@ zvol_clone_range(zvol_state_t *zv_src, uint64_t inoff, zvol_state_t *zv_dst,
*/
if ((inoff % zv_src->zv_volblocksize) != 0 ||
(outoff % zv_dst->zv_volblocksize) != 0) {
- error = EINVAL;
+ error = SET_ERROR(EINVAL);
goto out;
}
@@ -699,7 +707,7 @@ zvol_clone_range(zvol_state_t *zv_src, uint64_t inoff, zvol_state_t *zv_dst,
* Length must be multiple of block size
*/
if ((len % zv_src->zv_volblocksize) != 0) {
- error = EINVAL;
+ error = SET_ERROR(EINVAL);
goto out;
}
@@ -771,13 +779,13 @@ zvol_clone_range(zvol_state_t *zv_src, uint64_t inoff, zvol_state_t *zv_dst,
zfs_rangelock_exit(outlr);
zfs_rangelock_exit(inlr);
if (error == 0 && zv_dst->zv_objset->os_sync == ZFS_SYNC_ALWAYS) {
- zil_commit(zilog_dst, ZVOL_OBJ);
+ error = zil_commit(zilog_dst, ZVOL_OBJ);
}
out:
if (zv_src != zv_dst)
rw_exit(&zv_src->zv_suspend_lock);
rw_exit(&zv_dst->zv_suspend_lock);
- return (SET_ERROR(error));
+ return (error);
}
/*
@@ -897,7 +905,7 @@ zvol_log_write(zvol_state_t *zv, dmu_tx_t *tx, uint64_t offset,
if (wr_state == WR_COPIED &&
dmu_read_by_dnode(zv->zv_dn, offset, len, lr + 1,
DMU_READ_NO_PREFETCH | DMU_KEEP_CACHING) != 0) {
- zil_itx_destroy(itx);
+ zil_itx_destroy(itx, 0);
itx = zil_itx_create(TX_WRITE, sizeof (*lr));
lr = (lr_write_t *)&itx->itx_lr;
wr_state = WR_NEED_COPY;
@@ -916,7 +924,7 @@ zvol_log_write(zvol_state_t *zv, dmu_tx_t *tx, uint64_t offset,
itx->itx_private = zv;
- (void) zil_itx_assign(zilog, itx, tx);
+ zil_itx_assign(zilog, itx, tx);
offset += len;
size -= len;
@@ -1026,7 +1034,7 @@ zvol_get_data(void *arg, uint64_t arg2, lr_write_t *lr, char *buf,
zvol_get_done(zgd, error);
- return (SET_ERROR(error));
+ return (error);
}
/*
@@ -1071,15 +1079,15 @@ zvol_setup_zv(zvol_state_t *zv)
error = dsl_prop_get_integer(zv->zv_name, "readonly", &ro, NULL);
if (error)
- return (SET_ERROR(error));
+ return (error);
error = zap_lookup(os, ZVOL_ZAP_OBJ, "size", 8, 1, &volsize);
if (error)
- return (SET_ERROR(error));
+ return (error);
error = dnode_hold(os, ZVOL_OBJ, zv, &zv->zv_dn);
if (error)
- return (SET_ERROR(error));
+ return (error);
zvol_os_set_capacity(zv, volsize >> 9);
zv->zv_volsize = volsize;
@@ -1121,7 +1129,7 @@ zvol_shutdown_zv(zvol_state_t *zv)
*/
if (zv->zv_flags & ZVOL_WRITTEN_TO)
txg_wait_synced(dmu_objset_pool(zv->zv_objset), 0);
- (void) dmu_objset_evict_dbufs(zv->zv_objset);
+ dmu_objset_evict_dbufs(zv->zv_objset);
}
/*
@@ -1198,7 +1206,7 @@ zvol_resume(zvol_state_t *zv)
if (zv->zv_flags & ZVOL_REMOVING)
cv_broadcast(&zv->zv_removing_cv);
- return (SET_ERROR(error));
+ return (error);
}
int
@@ -1214,7 +1222,7 @@ zvol_first_open(zvol_state_t *zv, boolean_t readonly)
boolean_t ro = (readonly || (strchr(zv->zv_name, '@') != NULL));
error = dmu_objset_own(zv->zv_name, DMU_OST_ZVOL, ro, B_TRUE, zv, &os);
if (error)
- return (SET_ERROR(error));
+ return (error);
zv->zv_objset = os;
@@ -1440,41 +1448,32 @@ zvol_task_update_status(zvol_task_t *task, uint64_t total, uint64_t done,
}
}
-static const char *
-zvol_task_op_msg(zvol_async_op_t op)
-{
- switch (op) {
- case ZVOL_ASYNC_CREATE_MINORS:
- return ("create");
- case ZVOL_ASYNC_REMOVE_MINORS:
- return ("remove");
- case ZVOL_ASYNC_RENAME_MINORS:
- return ("rename");
- case ZVOL_ASYNC_SET_SNAPDEV:
- case ZVOL_ASYNC_SET_VOLMODE:
- return ("set property");
- default:
- return ("unknown");
- }
-
- __builtin_unreachable();
- return (NULL);
-}
-
static void
zvol_task_report_status(zvol_task_t *task)
{
+#ifdef ZFS_DEBUG
+ static const char *const msg[] = {
+ "create",
+ "remove",
+ "rename",
+ "set snapdev",
+ "set volmode",
+ "unknown",
+ };
if (task->zt_status == 0)
return;
+ zvol_async_op_t op = MIN(task->zt_op, ZVOL_ASYNC_MAX);
if (task->zt_error) {
dprintf("The %s minors zvol task was not ok, last error %d\n",
- zvol_task_op_msg(task->zt_op), task->zt_error);
+ msg[op], task->zt_error);
} else {
- dprintf("The %s minors zvol task was not ok\n",
- zvol_task_op_msg(task->zt_op));
+ dprintf("The %s minors zvol task was not ok\n", msg[op]);
}
+#else
+ (void) task;
+#endif
}
/*
@@ -1581,184 +1580,156 @@ zvol_create_minors_impl(zvol_task_t *task)
}
/*
- * Remove minors for specified dataset including children and snapshots.
- */
-
-/*
- * Remove the minor for a given zvol. This will do it all:
- * - flag the zvol for removal, so new requests are rejected
- * - wait until outstanding requests are completed
- * - remove it from lists
- * - free it
- * It's also usable as a taskq task, and smells nice too.
+ * Remove minors for specified dataset and, optionally, its children and
+ * snapshots.
*/
static void
-zvol_remove_minor_task(void *arg)
-{
- zvol_state_t *zv = (zvol_state_t *)arg;
-
- ASSERT(!RW_LOCK_HELD(&zvol_state_lock));
- ASSERT(!MUTEX_HELD(&zv->zv_state_lock));
-
- mutex_enter(&zv->zv_state_lock);
- while (zv->zv_open_count > 0 || atomic_read(&zv->zv_suspend_ref)) {
- zv->zv_flags |= ZVOL_REMOVING;
- cv_wait(&zv->zv_removing_cv, &zv->zv_state_lock);
- }
- mutex_exit(&zv->zv_state_lock);
-
- rw_enter(&zvol_state_lock, RW_WRITER);
- mutex_enter(&zv->zv_state_lock);
-
- zvol_remove(zv);
- zvol_os_clear_private(zv);
-
- mutex_exit(&zv->zv_state_lock);
- rw_exit(&zvol_state_lock);
-
- zvol_os_free(zv);
-}
-
-static void
-zvol_free_task(void *arg)
-{
- zvol_os_free(arg);
-}
-
-static void
zvol_remove_minors_impl(zvol_task_t *task)
{
zvol_state_t *zv, *zv_next;
const char *name = task ? task->zt_name1 : NULL;
int namelen = ((name) ? strlen(name) : 0);
- taskqid_t t;
- list_t delay_list, free_list;
+ boolean_t children = task ? !!task->zt_value : B_TRUE;
if (zvol_inhibit_dev)
return;
- list_create(&delay_list, sizeof (zvol_state_t),
- offsetof(zvol_state_t, zv_next));
- list_create(&free_list, sizeof (zvol_state_t),
- offsetof(zvol_state_t, zv_next));
+ /*
+ * We collect up zvols that we want to remove on a separate list, so
+ * that we don't have to hold zvol_state_lock for the whole time.
+ *
+ * We can't remove them from the global lists until we're completely
+ * done with them, because that would make them appear to ZFS-side ops
+ * that they don't exist, and the name might be reused, which can't be
+ * good.
+ */
+ list_t remove_list;
+ list_create(&remove_list, sizeof (zvol_state_t),
+ offsetof(zvol_state_t, zv_remove_node));
- rw_enter(&zvol_state_lock, RW_WRITER);
+ rw_enter(&zvol_state_lock, RW_READER);
for (zv = list_head(&zvol_state_list); zv != NULL; zv = zv_next) {
zv_next = list_next(&zvol_state_list, zv);
mutex_enter(&zv->zv_state_lock);
+ if (zv->zv_flags & ZVOL_REMOVING) {
+ /* Another thread is handling shutdown, skip it. */
+ mutex_exit(&zv->zv_state_lock);
+ continue;
+ }
+
+ /*
+ * This zvol should be removed if:
+ * - no name was offered (ie removing all at shutdown); or
+ * - name matches exactly; or
+ * - we were asked to remove children, and
+ * - the start of the name matches, and
+ * - there is a '/' immediately after the matched name; or
+ * - there is a '@' immediately after the matched name
+ */
if (name == NULL || strcmp(zv->zv_name, name) == 0 ||
- (strncmp(zv->zv_name, name, namelen) == 0 &&
+ (children && strncmp(zv->zv_name, name, namelen) == 0 &&
(zv->zv_name[namelen] == '/' ||
zv->zv_name[namelen] == '@'))) {
- /*
- * By holding zv_state_lock here, we guarantee that no
- * one is currently using this zv
- */
/*
- * If in use, try to throw everyone off and try again
- * later.
+ * Matched, so mark it removal. We want to take the
+ * write half of the suspend lock to make sure that
+ * the zvol is not suspended, and give any data ops
+ * chance to finish.
*/
- if (zv->zv_open_count > 0 ||
- atomic_read(&zv->zv_suspend_ref)) {
- zv->zv_flags |= ZVOL_REMOVING;
- t = taskq_dispatch(
- zv->zv_objset->os_spa->spa_zvol_taskq,
- zvol_remove_minor_task, zv, TQ_SLEEP);
- if (t == TASKQID_INVALID) {
- /*
- * Couldn't create the task, so we'll
- * do it in place once the loop is
- * finished.
- */
- list_insert_head(&delay_list, zv);
- }
+ mutex_exit(&zv->zv_state_lock);
+ rw_enter(&zv->zv_suspend_lock, RW_WRITER);
+ mutex_enter(&zv->zv_state_lock);
+
+ if (zv->zv_flags & ZVOL_REMOVING) {
+ /* Another thread has taken it, let them. */
mutex_exit(&zv->zv_state_lock);
+ rw_exit(&zv->zv_suspend_lock);
continue;
}
- zvol_remove(zv);
-
/*
- * Cleared while holding zvol_state_lock as a writer
- * which will prevent zvol_open() from opening it.
+ * Mark it and unlock. New entries will see the flag
+ * and return ENXIO.
*/
- zvol_os_clear_private(zv);
-
- /* Drop zv_state_lock before zvol_free() */
+ zv->zv_flags |= ZVOL_REMOVING;
mutex_exit(&zv->zv_state_lock);
+ rw_exit(&zv->zv_suspend_lock);
- /* Try parallel zv_free, if failed do it in place */
- t = taskq_dispatch(system_taskq, zvol_free_task, zv,
- TQ_SLEEP);
- if (t == TASKQID_INVALID)
- list_insert_head(&free_list, zv);
- } else {
+ /* Put it on the list for the next stage. */
+ list_insert_head(&remove_list, zv);
+ } else
mutex_exit(&zv->zv_state_lock);
- }
}
- rw_exit(&zvol_state_lock);
- /* Wait for zvols that we couldn't create a remove task for */
- while ((zv = list_remove_head(&delay_list)) != NULL)
- zvol_remove_minor_task(zv);
-
- /* Free any that we couldn't free in parallel earlier */
- while ((zv = list_remove_head(&free_list)) != NULL)
- zvol_os_free(zv);
-}
-
-/* Remove minor for this specific volume only */
-static int
-zvol_remove_minor_impl(const char *name)
-{
- zvol_state_t *zv = NULL, *zv_next;
-
- if (zvol_inhibit_dev)
- return (0);
+ rw_exit(&zvol_state_lock);
- rw_enter(&zvol_state_lock, RW_WRITER);
+ /* Didn't match any, nothing to do! */
+ if (list_is_empty(&remove_list)) {
+ if (task)
+ task->zt_error = SET_ERROR(ENOENT);
+ return;
+ }
- for (zv = list_head(&zvol_state_list); zv != NULL; zv = zv_next) {
- zv_next = list_next(&zvol_state_list, zv);
+ /* Actually shut them all down. */
+ for (zv = list_head(&remove_list); zv != NULL; zv = zv_next) {
+ zv_next = list_next(&remove_list, zv);
mutex_enter(&zv->zv_state_lock);
- if (strcmp(zv->zv_name, name) == 0)
- /* Found, leave the the loop with zv_lock held */
- break;
- mutex_exit(&zv->zv_state_lock);
- }
- if (zv == NULL) {
- rw_exit(&zvol_state_lock);
- return (ENOENT);
- }
-
- ASSERT(MUTEX_HELD(&zv->zv_state_lock));
+ /*
+ * Still open or suspended, just wait. This can happen if, for
+ * example, we managed to acquire zv_state_lock in the moments
+ * where zvol_open() or zvol_release() are trading locks to
+ * call zvol_first_open() or zvol_last_close().
+ */
+ while (zv->zv_open_count > 0 ||
+ atomic_read(&zv->zv_suspend_ref))
+ cv_wait(&zv->zv_removing_cv, &zv->zv_state_lock);
- if (zv->zv_open_count > 0 || atomic_read(&zv->zv_suspend_ref)) {
/*
- * In use, so try to throw everyone off, then wait
- * until finished.
+ * No users, shut down the OS side. This may not remove the
+ * minor from view immediately, depending on the kernel
+ * specifics, but it will ensure that it is unusable and that
+ * this zvol_state_t can never again be reached from an OS-side
+ * operation.
*/
- zv->zv_flags |= ZVOL_REMOVING;
+ zvol_os_remove_minor(zv);
mutex_exit(&zv->zv_state_lock);
+
+ /* Remove it from the name lookup lists */
+ rw_enter(&zvol_state_lock, RW_WRITER);
+ zvol_remove(zv);
rw_exit(&zvol_state_lock);
- zvol_remove_minor_task(zv);
- return (0);
}
- zvol_remove(zv);
- zvol_os_clear_private(zv);
+ /*
+ * Our own references on remove_list is the last one, free them and
+ * we're done.
+ */
+ while ((zv = list_remove_head(&remove_list)) != NULL)
+ zvol_os_free(zv);
- mutex_exit(&zv->zv_state_lock);
- rw_exit(&zvol_state_lock);
+ list_destroy(&remove_list);
+}
- zvol_os_free(zv);
+/* Remove minor for this specific volume only */
+static int
+zvol_remove_minor_impl(const char *name)
+{
+ if (zvol_inhibit_dev)
+ return (0);
- return (0);
+ zvol_task_t task;
+ memset(&task, 0, sizeof (zvol_task_t));
+ strlcpy(task.zt_name1, name, sizeof (task.zt_name1));
+ task.zt_value = B_FALSE;
+
+ zvol_remove_minors_impl(&task);
+
+ return (task.zt_error);
}
/*
@@ -2078,6 +2049,7 @@ zvol_remove_minors(spa_t *spa, const char *name, boolean_t async)
task = kmem_zalloc(sizeof (zvol_task_t), KM_SLEEP);
task->zt_op = ZVOL_ASYNC_REMOVE_MINORS;
strlcpy(task->zt_name1, name, sizeof (task->zt_name1));
+ task->zt_value = B_TRUE;
id = taskq_dispatch(spa->spa_zvol_taskq, zvol_task_cb, task, TQ_SLEEP);
if ((async == B_FALSE) && (id != TASKQID_INVALID))
taskq_wait_id(spa->spa_zvol_taskq, id);
@@ -2199,20 +2171,12 @@ zvol_fini_impl(void)
zvol_remove_minors_impl(NULL);
- /*
- * The call to "zvol_remove_minors_impl" may dispatch entries to
- * the system_taskq, but it doesn't wait for those entries to
- * complete before it returns. Thus, we must wait for all of the
- * removals to finish, before we can continue.
- */
- taskq_wait_outstanding(system_taskq, 0);
-
kmem_free(zvol_htable, ZVOL_HT_SIZE * sizeof (struct hlist_head));
list_destroy(&zvol_state_list);
rw_destroy(&zvol_state_lock);
if (ztqs->tqs_taskq == NULL) {
- ASSERT3U(ztqs->tqs_cnt, ==, 0);
+ ASSERT0(ztqs->tqs_cnt);
} else {
for (uint_t i = 0; i < ztqs->tqs_cnt; i++) {
ASSERT3P(ztqs->tqs_taskq[i], !=, NULL);
diff --git a/sys/contrib/openzfs/module/zstd/zfs_zstd.c b/sys/contrib/openzfs/module/zstd/zfs_zstd.c
index 391216d6e263..3db196953f74 100644
--- a/sys/contrib/openzfs/module/zstd/zfs_zstd.c
+++ b/sys/contrib/openzfs/module/zstd/zfs_zstd.c
@@ -876,9 +876,9 @@ static void __init
zstd_mempool_init(void)
{
zstd_mempool_cctx =
- kmem_zalloc(ZSTD_POOL_MAX * sizeof (struct zstd_pool), KM_SLEEP);
+ vmem_zalloc(ZSTD_POOL_MAX * sizeof (struct zstd_pool), KM_SLEEP);
zstd_mempool_dctx =
- kmem_zalloc(ZSTD_POOL_MAX * sizeof (struct zstd_pool), KM_SLEEP);
+ vmem_zalloc(ZSTD_POOL_MAX * sizeof (struct zstd_pool), KM_SLEEP);
for (int i = 0; i < ZSTD_POOL_MAX; i++) {
mutex_init(&zstd_mempool_cctx[i].barrier, NULL,
@@ -924,8 +924,8 @@ zstd_mempool_deinit(void)
release_pool(&zstd_mempool_dctx[i]);
}
- kmem_free(zstd_mempool_dctx, ZSTD_POOL_MAX * sizeof (struct zstd_pool));
- kmem_free(zstd_mempool_cctx, ZSTD_POOL_MAX * sizeof (struct zstd_pool));
+ vmem_free(zstd_mempool_dctx, ZSTD_POOL_MAX * sizeof (struct zstd_pool));
+ vmem_free(zstd_mempool_cctx, ZSTD_POOL_MAX * sizeof (struct zstd_pool));
zstd_mempool_dctx = NULL;
zstd_mempool_cctx = NULL;
}
diff --git a/sys/contrib/openzfs/rpm/generic/zfs.spec.in b/sys/contrib/openzfs/rpm/generic/zfs.spec.in
index dddc0a6c8f02..1ce668e7b86d 100644
--- a/sys/contrib/openzfs/rpm/generic/zfs.spec.in
+++ b/sys/contrib/openzfs/rpm/generic/zfs.spec.in
@@ -506,7 +506,6 @@ systemctl --system daemon-reload >/dev/null || true
# Core utilities
%{_sbindir}/*
%{_bindir}/raidz_test
-%{_sbindir}/zgenhostid
%{_bindir}/zvol_wait
# Optional Python 3 scripts
%{_bindir}/arc_summary
diff --git a/sys/contrib/openzfs/scripts/spdxcheck.pl b/sys/contrib/openzfs/scripts/spdxcheck.pl
index 88f5a235d70c..cdab5368f19c 100755
--- a/sys/contrib/openzfs/scripts/spdxcheck.pl
+++ b/sys/contrib/openzfs/scripts/spdxcheck.pl
@@ -190,6 +190,7 @@ my @path_license_tags = (
['BSD-2-Clause OR GPL-2.0-only', 'CDDL-1.0'],
'module/icp' => ['Apache-2.0', 'CDDL-1.0'],
+ 'contrib/icp' => ['Apache-2.0', 'CDDL-1.0'],
# Python bindings are always Apache-2.0
'contrib/pyzfs' => ['Apache-2.0'],
diff --git a/sys/contrib/openzfs/tests/runfiles/common.run b/sys/contrib/openzfs/tests/runfiles/common.run
index 9fad8946f4f3..2da46458289a 100644
--- a/sys/contrib/openzfs/tests/runfiles/common.run
+++ b/sys/contrib/openzfs/tests/runfiles/common.run
@@ -725,7 +725,11 @@ tests = ['fadvise_willneed']
tags = ['functional', 'fadvise']
[tests/functional/failmode]
-tests = ['failmode_dmu_tx_wait', 'failmode_dmu_tx_continue']
+tests = ['failmode_dmu_tx_wait', 'failmode_dmu_tx_continue',
+ 'failmode_fsync_wait', 'failmode_fsync_continue',
+ 'failmode_msync_wait', 'failmode_msync_continue',
+ 'failmode_osync_wait', 'failmode_osync_continue',
+ 'failmode_syncalways_wait', 'failmode_syncalways_continue']
tags = ['functional', 'failmode']
[tests/functional/fallocate]
@@ -1089,7 +1093,7 @@ tests = ['zvol_misc_002_pos', 'zvol_misc_hierarchy', 'zvol_misc_rename_inuse',
tags = ['functional', 'zvol', 'zvol_misc']
[tests/functional/zvol/zvol_stress]
-tests = ['zvol_stress']
+tests = ['zvol_stress', 'zvol_stress_destroy']
tags = ['functional', 'zvol', 'zvol_stress']
[tests/functional/zvol/zvol_swap]
diff --git a/sys/contrib/openzfs/tests/zfs-tests/cmd/.gitignore b/sys/contrib/openzfs/tests/zfs-tests/cmd/.gitignore
index 1cd90024e94d..62f1684acfb4 100644
--- a/sys/contrib/openzfs/tests/zfs-tests/cmd/.gitignore
+++ b/sys/contrib/openzfs/tests/zfs-tests/cmd/.gitignore
@@ -28,6 +28,7 @@
/mmap_seek
/mmap_sync
/mmapwrite
+/mmap_write_sync
/nvlist_to_lua
/randfree_file
/randwritecomp
diff --git a/sys/contrib/openzfs/tests/zfs-tests/cmd/Makefile.am b/sys/contrib/openzfs/tests/zfs-tests/cmd/Makefile.am
index d5448055a1e1..85c3cf3c35a8 100644
--- a/sys/contrib/openzfs/tests/zfs-tests/cmd/Makefile.am
+++ b/sys/contrib/openzfs/tests/zfs-tests/cmd/Makefile.am
@@ -74,7 +74,7 @@ scripts_zfs_tests_bin_PROGRAMS += %D%/mkbusy %D%/mkfile %D%/mkfiles %D%/mktree
scripts_zfs_tests_bin_PROGRAMS += \
%D%/mmap_exec %D%/mmap_ftruncate %D%/mmap_seek \
- %D%/mmap_sync %D%/mmapwrite %D%/readmmap
+ %D%/mmap_sync %D%/mmapwrite %D%/readmmap %D%/mmap_write_sync
%C%_mmapwrite_LDADD = -lpthread
if WANT_MMAP_LIBAIO
diff --git a/sys/contrib/openzfs/tests/zfs-tests/cmd/btree_test.c b/sys/contrib/openzfs/tests/zfs-tests/cmd/btree_test.c
index e7b80d01efaa..f8948a61833d 100644
--- a/sys/contrib/openzfs/tests/zfs-tests/cmd/btree_test.c
+++ b/sys/contrib/openzfs/tests/zfs-tests/cmd/btree_test.c
@@ -207,7 +207,7 @@ insert_find_remove(zfs_btree_t *bt, char *why)
"Found removed value (%llu)\n", *p);
return (1);
}
- ASSERT3S(zfs_btree_numnodes(bt), ==, 0);
+ ASSERT0(zfs_btree_numnodes(bt));
zfs_btree_verify(bt);
return (0);
@@ -279,7 +279,7 @@ drain_tree(zfs_btree_t *bt, char *why)
node = avl_last(&avl);
ASSERT3U(node->data, ==, *(uint64_t *)zfs_btree_last(bt, NULL));
}
- ASSERT3S(zfs_btree_numnodes(bt), ==, 0);
+ ASSERT0(zfs_btree_numnodes(bt));
void *avl_cookie = NULL;
while ((node = avl_destroy_nodes(&avl, &avl_cookie)) != NULL)
diff --git a/sys/contrib/openzfs/tests/zfs-tests/cmd/crypto_test.c b/sys/contrib/openzfs/tests/zfs-tests/cmd/crypto_test.c
index e08003f80464..cbebd33e0bf6 100644
--- a/sys/contrib/openzfs/tests/zfs-tests/cmd/crypto_test.c
+++ b/sys/contrib/openzfs/tests/zfs-tests/cmd/crypto_test.c
@@ -529,6 +529,8 @@ static const char *aes_gcm_impl[][2] = {
{ "aesni", "pclmulqdq" },
{ "x86_64", "avx" },
{ "aesni", "avx" },
+ { "x86_64", "avx2" },
+ { "aesni", "avx2" },
};
/* signature of function to call after setting implementation params */
diff --git a/sys/contrib/openzfs/tests/zfs-tests/cmd/mmap_write_sync.c b/sys/contrib/openzfs/tests/zfs-tests/cmd/mmap_write_sync.c
new file mode 100644
index 000000000000..ad5e37f24960
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/cmd/mmap_write_sync.c
@@ -0,0 +1,84 @@
+// SPDX-License-Identifier: CDDL-1.0
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License (the "License").
+ * You may not use this file except in compliance with the License.
+ *
+ * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+ * or https://opensource.org/licenses/CDDL-1.0.
+ * See the License for the specific language governing permissions
+ * and limitations under the License.
+ *
+ * When distributing Covered Code, include this CDDL HEADER in each
+ * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+ * If applicable, add the following below this CDDL HEADER, with the
+ * fields enclosed by brackets "[]" replaced with your own identifying
+ * information: Portions Copyright [yyyy] [name of copyright owner]
+ *
+ * CDDL HEADER END
+ */
+
+/*
+ * Copyright (c) 2025, Klara, Inc.
+ */
+
+#include <stdio.h>
+#include <unistd.h>
+#include <stdlib.h>
+#include <fcntl.h>
+#include <sys/stat.h>
+#include <sys/mman.h>
+
+#define PAGES (8)
+
+int
+main(int argc, char **argv)
+{
+ if (argc != 2) {
+ fprintf(stderr, "usage: %s <filename>\n", argv[0]);
+ exit(1);
+ }
+
+ long page_size = sysconf(_SC_PAGESIZE);
+ if (page_size < 0) {
+ perror("sysconf");
+ exit(2);
+ }
+ size_t map_size = page_size * PAGES;
+
+ int fd = open(argv[1], O_CREAT|O_RDWR, S_IRWXU|S_IRWXG|S_IRWXO);
+ if (fd < 0) {
+ perror("open");
+ exit(2);
+ }
+
+ if (ftruncate(fd, map_size) < 0) {
+ perror("ftruncate");
+ close(fd);
+ exit(2);
+ }
+
+ uint64_t *p =
+ mmap(NULL, map_size, PROT_READ|PROT_WRITE, MAP_SHARED, fd, 0);
+ if (p == MAP_FAILED) {
+ perror("mmap");
+ close(fd);
+ exit(2);
+ }
+
+ for (int i = 0; i < (map_size / sizeof (uint64_t)); i++)
+ p[i] = 0x0123456789abcdef;
+
+ if (msync(p, map_size, MS_SYNC) < 0) {
+ perror("msync");
+ munmap(p, map_size);
+ close(fd);
+ exit(3);
+ }
+
+ munmap(p, map_size);
+ close(fd);
+ exit(0);
+}
diff --git a/sys/contrib/openzfs/tests/zfs-tests/include/commands.cfg b/sys/contrib/openzfs/tests/zfs-tests/include/commands.cfg
index bbaa8665ecc8..884a99d785bc 100644
--- a/sys/contrib/openzfs/tests/zfs-tests/include/commands.cfg
+++ b/sys/contrib/openzfs/tests/zfs-tests/include/commands.cfg
@@ -210,6 +210,7 @@ export ZFSTEST_FILES='badsend
mmap_seek
mmap_sync
mmapwrite
+ mmap_write_sync
nvlist_to_lua
randfree_file
randwritecomp
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am b/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am
index c2542287c1d7..41e7b45ef4ec 100644
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am
@@ -276,6 +276,7 @@ nobase_dist_datadir_zfs_tests_tests_DATA += \
functional/direct/dio.kshlib \
functional/events/events.cfg \
functional/events/events_common.kshlib \
+ functional/failmode/failmode.kshlib \
functional/fault/fault.cfg \
functional/gang_blocks/gang_blocks.kshlib \
functional/grow/grow.cfg \
@@ -1541,6 +1542,14 @@ nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \
functional/failmode/cleanup.ksh \
functional/failmode/failmode_dmu_tx_wait.ksh \
functional/failmode/failmode_dmu_tx_continue.ksh \
+ functional/failmode/failmode_fsync_wait.ksh \
+ functional/failmode/failmode_fsync_continue.ksh \
+ functional/failmode/failmode_msync_wait.ksh \
+ functional/failmode/failmode_msync_continue.ksh \
+ functional/failmode/failmode_osync_wait.ksh \
+ functional/failmode/failmode_osync_continue.ksh \
+ functional/failmode/failmode_syncalways_wait.ksh \
+ functional/failmode/failmode_syncalways_continue.ksh \
functional/failmode/setup.ksh \
functional/fallocate/cleanup.ksh \
functional/fallocate/fallocate_prealloc.ksh \
@@ -2235,6 +2244,7 @@ nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \
functional/zvol/zvol_stress/cleanup.ksh \
functional/zvol/zvol_stress/setup.ksh \
functional/zvol/zvol_stress/zvol_stress.ksh \
+ functional/zvol/zvol_stress/zvol_stress_destroy.ksh \
functional/zvol/zvol_swap/cleanup.ksh \
functional/zvol/zvol_swap/setup.ksh \
functional/zvol/zvol_swap/zvol_swap_001_pos.ksh \
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode.kshlib b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode.kshlib
new file mode 100644
index 000000000000..d0b7404557ab
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode.kshlib
@@ -0,0 +1,149 @@
+#!/bin/ksh -p
+# SPDX-License-Identifier: CDDL-1.0
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or https://opensource.org/licenses/CDDL-1.0.
+# See the License for the specific language governing permissions
+# and limitations under the License.
+#
+# When distributing Covered Code, include this CDDL HEADER in each
+# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+# If applicable, add the following below this CDDL HEADER, with the
+# fields enclosed by brackets "[]" replaced with your own identifying
+# information: Portions Copyright [yyyy] [name of copyright owner]
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2025, Klara, Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+
+typeset -A failmode_sync_helper_cmd=(
+ ["fsync"]='dd if=/dev/urandom of=DATAFILE bs=128k count=1 conv=fsync'
+ ["msync"]='mmap_write_sync DATAFILE'
+ ["osync"]='dd if=/dev/urandom of=DATAFILE bs=128k count=1 oflag=sync'
+ ["syncalways"]='dd if=/dev/urandom of=DATAFILE bs=128k count=1'
+)
+
+typeset -A failmode_sync_helper_dsopts=(
+ ["syncalways"]="-o sync=always"
+)
+
+function failmode_sync_cleanup
+{
+ zinject -c all || true
+ zpool clear $TESTPOOL || true
+ destroy_pool $TESTPOOL
+}
+
+#
+# failmode_sync_test <failmode> <helper>
+#
+# run a failmode sync test:
+# - failmode: wait|continue
+# - helper: fsync|msync|osync|syncalways
+#
+function failmode_sync_test
+{
+ typeset failmode=$1
+ typeset helper=$2
+
+ # we'll need two disks, one for the main pool, one for the log
+ read -r DISK1 DISK2 _ <<<"$DISKS"
+
+ # file to write to the pool
+ typeset datafile="/$TESTPOOL/$TESTFS/datafile"
+
+ # create a single-disk pool with a separate log and the wanted failmode
+ log_must zpool create \
+ -f -o failmode=$failmode $TESTPOOL $DISK1 log $DISK2
+
+ # create the test dataset. we bias the ZIL towards the log device to
+ # try to ensure that the sync write never involves the main device
+ log_must zfs create \
+ -o recordsize=128k -o logbias=latency \
+ ${failmode_sync_helper_dsopts[$helper]} \
+ $TESTPOOL/$TESTFS
+
+ # create the target file. the ZIL head structure is created on first
+ # use, and does a full txg wait to finish, which we want to avoid
+ log_must dd if=/dev/zero of=$datafile bs=128k count=1 conv=fsync
+ log_must zpool sync
+
+ # inject errors. writes will fail, as will the followup probes
+ zinject -d $DISK1 -e io -T write $TESTPOOL
+ zinject -d $DISK1 -e nxio -T probe $TESTPOOL
+ zinject -d $DISK2 -e io -T write $TESTPOOL
+ zinject -d $DISK2 -e nxio -T probe $TESTPOOL
+
+ # run the helper program in the background. the pool should immediately
+ # suspend, and the sync op block or fail based on the failmode
+ typeset helper_cmd=${failmode_sync_helper_cmd[$helper]/DATAFILE/$datafile}
+ log_note "running failmode sync helper: $helper_cmd"
+ $helper_cmd &
+ typeset -i pid=$!
+
+ # should only take a moment, but give it a chance
+ log_note "waiting for pool to suspend"
+ typeset -i tries=10
+ until [[ $(kstat_pool $TESTPOOL state) == "SUSPENDED" ]] ; do
+ if ((tries-- == 0)); then
+ log_fail "pool didn't suspend"
+ fi
+ sleep 1
+ done
+
+ # zil_commit() should have noticed the suspend by now
+ typeset -i zilerr=$(kstat zil.zil_commit_error_count)
+
+ # see if the helper program blocked
+ typeset -i blocked
+ if kill -0 $pid ; then
+ blocked=1
+ log_note "$helper: blocked in the kernel"
+ else
+ blocked=0
+ log_note "$helper: exited while pool suspended"
+ fi
+
+ # bring the pool back online
+ zinject -c all
+ zpool clear $TESTPOOL
+
+ # program definitely exited now, get its return code
+ wait $pid
+ typeset -i rc=$?
+
+ failmode_sync_cleanup
+
+ log_note "$helper: zilerr=$zilerr blocked=$blocked rc=$rc"
+
+ # confirm expected results for the failmode
+ if [[ $failmode = "wait" ]] ; then
+ # - the ZIL saw an error, and fell back to a txg sync
+ # - sync op blocked when the pool suspended
+ # - after resume, sync op succeeded, helper returned success
+ log_must test $zilerr -ne 0
+ log_must test $blocked -eq 1
+ log_must test $rc -eq 0
+ elif [[ $failmode = "continue" ]] ; then
+ # confirm expected results:
+ # - the ZIL saw an error, and fell back to a txg sync
+ # - helper exited when the pool suspended
+ # - sync op returned an error, so helper returned failure
+ log_must test $zilerr -ne 0
+ log_must test $blocked -eq 0
+ log_must test $rc -ne 0
+ else
+ log_fail "impossible failmode: $failmode"
+ fi
+}
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_fsync_continue.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_fsync_continue.ksh
new file mode 100755
index 000000000000..7b145d3a2b4c
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_fsync_continue.ksh
@@ -0,0 +1,36 @@
+#!/bin/ksh -p
+# SPDX-License-Identifier: CDDL-1.0
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or https://opensource.org/licenses/CDDL-1.0.
+# See the License for the specific language governing permissions
+# and limitations under the License.
+#
+# When distributing Covered Code, include this CDDL HEADER in each
+# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+# If applicable, add the following below this CDDL HEADER, with the
+# fields enclosed by brackets "[]" replaced with your own identifying
+# information: Portions Copyright [yyyy] [name of copyright owner]
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2025, Klara, Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+. $STF_SUITE/tests/functional/failmode/failmode.kshlib
+
+typeset desc="fsync() returns when pool suspends with failmode=continue"
+
+log_assert $desc
+log_onexit failmode_sync_cleanup
+log_must failmode_sync_test continue fsync
+log_pass $desc
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_fsync_wait.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_fsync_wait.ksh
new file mode 100755
index 000000000000..677d226b5481
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_fsync_wait.ksh
@@ -0,0 +1,36 @@
+#!/bin/ksh -p
+# SPDX-License-Identifier: CDDL-1.0
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or https://opensource.org/licenses/CDDL-1.0.
+# See the License for the specific language governing permissions
+# and limitations under the License.
+#
+# When distributing Covered Code, include this CDDL HEADER in each
+# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+# If applicable, add the following below this CDDL HEADER, with the
+# fields enclosed by brackets "[]" replaced with your own identifying
+# information: Portions Copyright [yyyy] [name of copyright owner]
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2025, Klara, Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+. $STF_SUITE/tests/functional/failmode/failmode.kshlib
+
+typeset desc="fsync() blocks when pool suspends with failmode=wait"
+
+log_assert $desc
+log_onexit failmode_sync_cleanup
+log_must failmode_sync_test wait fsync
+log_pass $desc
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_msync_continue.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_msync_continue.ksh
new file mode 100755
index 000000000000..0c79ee15a1ba
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_msync_continue.ksh
@@ -0,0 +1,36 @@
+#!/bin/ksh -p
+# SPDX-License-Identifier: CDDL-1.0
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or https://opensource.org/licenses/CDDL-1.0.
+# See the License for the specific language governing permissions
+# and limitations under the License.
+#
+# When distributing Covered Code, include this CDDL HEADER in each
+# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+# If applicable, add the following below this CDDL HEADER, with the
+# fields enclosed by brackets "[]" replaced with your own identifying
+# information: Portions Copyright [yyyy] [name of copyright owner]
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2025, Klara, Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+. $STF_SUITE/tests/functional/failmode/failmode.kshlib
+
+typeset desc="msync() returns when pool suspends with failmode=continue"
+
+log_assert $desc
+log_onexit failmode_sync_cleanup
+log_must failmode_sync_test continue msync
+log_pass $desc
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_msync_wait.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_msync_wait.ksh
new file mode 100755
index 000000000000..a59d8cc50d61
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_msync_wait.ksh
@@ -0,0 +1,36 @@
+#!/bin/ksh -p
+# SPDX-License-Identifier: CDDL-1.0
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or https://opensource.org/licenses/CDDL-1.0.
+# See the License for the specific language governing permissions
+# and limitations under the License.
+#
+# When distributing Covered Code, include this CDDL HEADER in each
+# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+# If applicable, add the following below this CDDL HEADER, with the
+# fields enclosed by brackets "[]" replaced with your own identifying
+# information: Portions Copyright [yyyy] [name of copyright owner]
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2025, Klara, Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+. $STF_SUITE/tests/functional/failmode/failmode.kshlib
+
+typeset desc="msync() blocks when pool suspends with failmode=wait"
+
+log_assert $desc
+log_onexit failmode_sync_cleanup
+log_must failmode_sync_test wait msync
+log_pass $desc
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_osync_continue.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_osync_continue.ksh
new file mode 100755
index 000000000000..c4fa0c8f042c
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_osync_continue.ksh
@@ -0,0 +1,36 @@
+#!/bin/ksh -p
+# SPDX-License-Identifier: CDDL-1.0
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or https://opensource.org/licenses/CDDL-1.0.
+# See the License for the specific language governing permissions
+# and limitations under the License.
+#
+# When distributing Covered Code, include this CDDL HEADER in each
+# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+# If applicable, add the following below this CDDL HEADER, with the
+# fields enclosed by brackets "[]" replaced with your own identifying
+# information: Portions Copyright [yyyy] [name of copyright owner]
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2025, Klara, Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+. $STF_SUITE/tests/functional/failmode/failmode.kshlib
+
+typeset desc="O_SYNC returns when pool suspends with failmode=continue"
+
+log_assert $desc
+log_onexit failmode_sync_cleanup
+log_must failmode_sync_test continue osync
+log_pass $desc
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_osync_wait.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_osync_wait.ksh
new file mode 100755
index 000000000000..5f65cf92ad33
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_osync_wait.ksh
@@ -0,0 +1,37 @@
+#!/bin/ksh -p
+# SPDX-License-Identifier: CDDL-1.0
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or https://opensource.org/licenses/CDDL-1.0.
+# See the License for the specific language governing permissions
+# and limitations under the License.
+#
+# When distributing Covered Code, include this CDDL HEADER in each
+# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+# If applicable, add the following below this CDDL HEADER, with the
+# fields enclosed by brackets "[]" replaced with your own identifying
+# information: Portions Copyright [yyyy] [name of copyright owner]
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2025, Klara, Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+. $STF_SUITE/tests/functional/failmode/failmode.kshlib
+
+typeset desc="O_SYNC blocks when pool suspends with failmode=wait"
+
+log_assert $desc
+log_onexit failmode_sync_cleanup
+log_must failmode_sync_test wait osync
+log_pass $desc
+
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_syncalways_continue.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_syncalways_continue.ksh
new file mode 100755
index 000000000000..b80d776224a0
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_syncalways_continue.ksh
@@ -0,0 +1,37 @@
+#!/bin/ksh -p
+# SPDX-License-Identifier: CDDL-1.0
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or https://opensource.org/licenses/CDDL-1.0.
+# See the License for the specific language governing permissions
+# and limitations under the License.
+#
+# When distributing Covered Code, include this CDDL HEADER in each
+# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+# If applicable, add the following below this CDDL HEADER, with the
+# fields enclosed by brackets "[]" replaced with your own identifying
+# information: Portions Copyright [yyyy] [name of copyright owner]
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2025, Klara, Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+. $STF_SUITE/tests/functional/failmode/failmode.kshlib
+
+typeset desc="write()+sync=always returns when pool suspends with failmode=continue"
+
+log_assert $desc
+log_onexit failmode_sync_cleanup
+log_must failmode_sync_test continue syncalways
+log_pass $desc
+
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_syncalways_wait.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_syncalways_wait.ksh
new file mode 100755
index 000000000000..4fcb167b5c77
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/failmode/failmode_syncalways_wait.ksh
@@ -0,0 +1,37 @@
+#!/bin/ksh -p
+# SPDX-License-Identifier: CDDL-1.0
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or https://opensource.org/licenses/CDDL-1.0.
+# See the License for the specific language governing permissions
+# and limitations under the License.
+#
+# When distributing Covered Code, include this CDDL HEADER in each
+# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+# If applicable, add the following below this CDDL HEADER, with the
+# fields enclosed by brackets "[]" replaced with your own identifying
+# information: Portions Copyright [yyyy] [name of copyright owner]
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2025, Klara, Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+. $STF_SUITE/tests/functional/failmode/failmode.kshlib
+
+typeset desc="write()+sync=always blocks when pool suspends with failmode=wait"
+
+log_assert $desc
+log_onexit failmode_sync_cleanup
+log_must failmode_sync_test wait syncalways
+log_pass $desc
+
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/zvol/zvol_stress/zvol_stress_destroy.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/zvol/zvol_stress/zvol_stress_destroy.ksh
new file mode 100755
index 000000000000..669b59fac01f
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/zvol/zvol_stress/zvol_stress_destroy.ksh
@@ -0,0 +1,66 @@
+#!/bin/ksh -p
+# SPDX-License-Identifier: CDDL-1.0
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or https://opensource.org/licenses/CDDL-1.0.
+# See the License for the specific language governing permissions
+# and limitations under the License.
+#
+# When distributing Covered Code, include this CDDL HEADER in each
+# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+# If applicable, add the following below this CDDL HEADER, with the
+# fields enclosed by brackets "[]" replaced with your own identifying
+# information: Portions Copyright [yyyy] [name of copyright owner]
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2025, Klara, Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+
+verify_runnable "global"
+
+typeset -i nzvols=1000
+typeset -i parallel=$(( $(get_num_cpus) * 2 ))
+
+function cleanup {
+ for zvol in $(zfs list -Ho name -t vol) ; do
+ log_must_busy zfs destroy $zvol
+ done
+}
+
+log_onexit cleanup
+
+log_assert "stress test concurrent zvol create/destroy"
+
+function destroy_zvols_until {
+ typeset cond=$1
+ while true ; do
+ IFS='' zfs list -Ho name -t vol | read -r -d '' zvols
+ if [[ -n $zvols ]] ; then
+ echo $zvols | xargs -n 1 -P $parallel zfs destroy
+ fi
+ if ! $cond ; then
+ break
+ fi
+ done
+}
+
+( seq $nzvols | \
+ xargs -P $parallel -I % zfs create -s -V 1G $TESTPOOL/testvol% ) &
+cpid=$!
+sleep 1
+
+destroy_zvols_until "kill -0 $cpid"
+destroy_zvols_until "false"
+
+log_pass "stress test done"