Index: stable/10/cddl/contrib/opensolaris/cmd/zfs/zfs.8 =================================================================== --- stable/10/cddl/contrib/opensolaris/cmd/zfs/zfs.8 (revision 297101) +++ stable/10/cddl/contrib/opensolaris/cmd/zfs/zfs.8 (revision 297102) @@ -1,3728 +1,3731 @@ '\" te .\" Copyright (c) 2013, Martin Matuska . .\" All Rights Reserved. .\" .\" 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 http://www.opensolaris.org/os/licensing. .\" 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] .\" .\" Copyright (c) 2010, Sun Microsystems, Inc. All Rights Reserved. .\" Copyright (c) 2011, 2014 by Delphix. All rights reserved. .\" Copyright (c) 2011, Pawel Jakub Dawidek .\" Copyright (c) 2012, Glen Barber .\" Copyright (c) 2012, Bryan Drewery .\" Copyright (c) 2013 by Saso Kiselkov. All rights reserved. .\" Copyright (c) 2014, Joyent, Inc. All rights reserved. .\" Copyright (c) 2013, Steven Hartland .\" Copyright (c) 2014 Nexenta Systems, Inc. All Rights Reserved. .\" Copyright (c) 2014, Xin LI .\" Copyright (c) 2014-2015, The FreeBSD Foundation, All Rights Reserved. .\" .\" $FreeBSD$ .\" .Dd September 14, 2015 .Dt ZFS 8 .Os .Sh NAME .Nm zfs .Nd configures ZFS file systems .Sh SYNOPSIS .Nm .Op Fl \&? .Nm .Cm create .Op Fl pu .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... Ar filesystem .Nm .Cm create .Op Fl ps .Op Fl b Ar blocksize .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Fl V .Ar size volume .Nm .Cm destroy .Op Fl fnpRrv .Ar filesystem Ns | Ns Ar volume .Nm .Cm destroy .Op Fl dnpRrv .Sm off .Ar filesystem Ns | Ns volume .Ns @snap .Op % Ns Ar snap .Op , Ns Ar snap Op % Ns Ar snap .Op , Ns ... .Sm on .Nm .Cm destroy .Ar filesystem Ns | Ns Ar volume Ns # Ns Ar bookmark .Nm .Cm snapshot Ns | Ns Cm snap .Op Fl r .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Ar filesystem@snapname Ns | Ns Ar volume@snapname .Ar filesystem@snapname Ns | Ns Ar volume@snapname Ns ... .Nm .Cm rollback .Op Fl rRf .Ar snapshot .Nm .Cm clone .Op Fl p .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Ar snapshot filesystem Ns | Ns Ar volume .Nm .Cm promote .Ar clone-filesystem .Nm .Cm rename .Op Fl f .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot .Nm .Cm rename .Op Fl f .Fl p .Ar filesystem Ns | Ns Ar volume .Ar filesystem Ns | Ns Ar volume .Nm .Cm rename .Fl r .Ar snapshot snapshot .Nm .Cm rename .Fl u .Op Fl p .Ar filesystem filesystem .Nm .Cm list .Op Fl r Ns | Ns Fl d Ar depth .Op Fl Hp .Op Fl o Ar property Ns Oo , Ns property Ns Oc Ns ... .Op Fl t Ar type Ns Oo , Ns type Ns Oc Ns ... .Oo Fl s Ar property Oc Ns ... .Oo Fl S Ar property Oc Ns ... .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot .Nm .Cm set .Ar property Ns = Ns Ar value Oo Ar property Ns = Ns Ar value Oc Ns ... .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot Ns ... .Nm .Cm get .Op Fl r Ns | Ns Fl d Ar depth .Op Fl Hp .Op Fl o Ar all | field Ns Oo , Ns Ar field Oc Ns ... .Op Fl t Ar type Ns Oo Ns , Ar type Oc Ns ... .Op Fl s Ar source Ns Oo Ns , Ns Ar source Oc Ns ... .Ar all | property Ns Oo Ns , Ns Ar property Oc Ns ... .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot Ns ... .Nm .Cm inherit .Op Fl rS .Ar property .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot Ns ... .Nm .Cm upgrade .Op Fl v .Nm .Cm upgrade .Op Fl r .Op Fl V Ar version .Fl a | Ar filesystem .Nm .Cm userspace .Op Fl Hinp .Op Fl o Ar field Ns Oo , Ns Ar field Oc Ns ... .Oo Fl s Ar field Oc Ns ... .Oo Fl S Ar field Oc Ns ... .Op Fl t Ar type Ns Oo Ns , Ns Ar type Oc Ns ... .Ar filesystem Ns | Ns Ar snapshot .Nm .Cm groupspace .Op Fl Hinp .Op Fl o Ar field Ns Oo , Ns field Oc Ns ... .Oo Fl s Ar field Oc Ns ... .Oo Fl S Ar field Oc Ns ... .Op Fl t Ar type Ns Oo Ns , Ns Ar type Oc Ns ... .Ar filesystem Ns | Ns Ar snapshot .Nm .Cm mount .Nm .Cm mount .Op Fl vO .Op Fl o Ar property Ns Oo , Ns Ar property Oc Ns ... .Fl a | Ar filesystem .Nm .Cm unmount Ns | Ns Cm umount .Op Fl f .Fl a | Ar filesystem Ns | Ns Ar mountpoint .Nm .Cm share .Fl a | Ar filesystem .Nm .Cm unshare .Fl a | Ar filesystem Ns | Ns Ar mountpoint .Nm .Cm bookmark .Ar snapshot .Ar bookmark .Nm .Cm send .Op Fl DnPpRveL .Op Fl i Ar snapshot | Fl I Ar snapshot .Ar snapshot .Nm .Cm send .Op Fl eL .Op Fl i Ar snapshot Ns | Ns bookmark .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot .Nm .Cm send .Op Fl Penv .Fl t Ar receive_resume_token .Nm .Cm receive Ns | Ns Cm recv .Op Fl vnsFu .Op Fl o Sy origin Ns = Ns Ar snapshot .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot .Nm .Cm receive Ns | Ns Cm recv .Op Fl vnsFu .Op Fl d | e .Op Fl o Sy origin Ns = Ns Ar snapshot .Ar filesystem .Nm .Cm receive Ns | Ns Cm recv .Fl A .Ar filesystem Ns | Ns Ar volume .Nm .Cm allow .Ar filesystem Ns | Ns Ar volume .Nm .Cm allow .Op Fl ldug .Ar user Ns | Ns Ar group Ns Oo Ns , Ns Ar user Ns | Ns Ar group Oc Ns ... .Ar perm Ns | Ns Ar @setname Ns .Oo Ns , Ns Ar perm Ns | Ns Ar @setname Oc Ns ... .Ar filesystem Ns | Ns Ar volume .Nm .Cm allow .Op Fl ld .Fl e Ns | Ns Cm everyone .Ar perm Ns | Ns Ar @setname Ns Op Ns , Ns Ar perm Ns | Ns Ar @setname Ns .Ns ... .Ar filesystem Ns | Ns Ar volume .Nm .Cm allow .Fl c .Ar perm Ns | Ns Ar @setname Ns Op Ns , Ns Ar perm Ns | Ns Ar @setname Ns .Ns ... .Ar filesystem Ns | Ns Ar volume .Nm .Cm allow .Fl s .Ar @setname .Ar perm Ns | Ns Ar @setname Ns Op Ns , Ns Ar perm Ns | Ns Ar @setname Ns .Ns ... .Ar filesystem Ns | Ns Ar volume .Nm .Cm unallow .Op Fl rldug .Ar user Ns | Ns Ar group Ns Oo Ns , Ns Ar user Ns | Ns Ar group Oc Ns ... .Oo Ar perm Ns | Ns Ar @setname Ns Op , Ns Ar perm Ns | Ns Ar @setname Ns .Ns ... Oc .Ar filesystem Ns | Ns Ar volume .Nm .Cm unallow .Op Fl rld .Fl e Ns | Ns Cm everyone .Oo Ar perm Ns | Ns Ar @setname Ns Op , Ns Ar perm Ns | Ns Ar @setname Ns .Ns ... Oc .Ar filesystem Ns | Ns Ar volume .Nm .Cm unallow .Op Fl r .Fl c .Oo Ar perm Ns | Ns Ar @setname Ns Op , Ns Ar perm Ns | Ns Ar @setname Ns .Ns ... Oc .Ar filesystem Ns | Ns Ar volume .Nm .Cm unallow .Op Fl r .Fl s .Ar @setname .Oo Ar perm Ns | Ns Ar @setname Ns Op , Ns Ar perm Ns | Ns Ar @setname Ns .Ns ... Oc .Ar filesystem Ns | Ns Ar volume .Nm .Cm hold .Op Fl r .Ar tag snapshot Ns ... .Nm .Cm holds .Op Fl r .Ar snapshot Ns ... .Nm .Cm release .Op Fl r .Ar tag snapshot Ns ... .Nm .Cm diff .Op Fl FHt .Ar snapshot .Op Ar snapshot Ns | Ns Ar filesystem .Nm .Cm jail .Ar jailid Ns | Ns Ar jailname filesystem .Nm .Cm unjail .Ar jailid Ns | Ns Ar jailname filesystem .Sh DESCRIPTION The .Nm command configures .Tn ZFS datasets within a .Tn ZFS storage pool, as described in .Xr zpool 8 . A dataset is identified by a unique path within the .Tn ZFS namespace. For example: .Bd -ragged -offset 4n .No pool/ Ns Brq filesystem,volume,snapshot .Ed .Pp where the maximum length of a dataset name is .Dv MAXNAMELEN (256 bytes). .Pp A dataset can be one of the following: .Bl -hang -width 12n .It Sy file system A .Tn ZFS dataset of type .Em filesystem can be mounted within the standard system namespace and behaves like other file systems. While .Tn ZFS file systems are designed to be .Tn POSIX compliant, known issues exist that prevent compliance in some cases. Applications that depend on standards conformance might fail due to nonstandard behavior when checking file system free space. .It Sy volume A logical volume exported as a raw or block device. This type of dataset should only be used under special circumstances. File systems are typically used in most environments. .It Sy snapshot A read-only version of a file system or volume at a given point in time. It is specified as .Em filesystem@name or .Em volume@name . .El .Ss ZFS File System Hierarchy A .Tn ZFS storage pool is a logical collection of devices that provide space for datasets. A storage pool is also the root of the .Tn ZFS file system hierarchy. .Pp The root of the pool can be accessed as a file system, such as mounting and unmounting, taking snapshots, and setting properties. The physical storage characteristics, however, are managed by the .Xr zpool 8 command. .Pp See .Xr zpool 8 for more information on creating and administering pools. .Ss Snapshots A snapshot is a read-only copy of a file system or volume. Snapshots can be created extremely quickly, and initially consume no additional space within the pool. As data within the active dataset changes, the snapshot consumes more data than would otherwise be shared with the active dataset. .Pp Snapshots can have arbitrary names. Snapshots of volumes can be cloned or rolled back, but cannot be accessed independently. .Pp File system snapshots can be accessed under the .Pa \&.zfs/snapshot directory in the root of the file system. Snapshots are automatically mounted on demand and may be unmounted at regular intervals. The visibility of the .Pa \&.zfs directory can be controlled by the .Sy snapdir property. .Ss Clones A clone is a writable volume or file system whose initial contents are the same as another dataset. As with snapshots, creating a clone is nearly instantaneous, and initially consumes no additional space. .Pp Clones can only be created from a snapshot. When a snapshot is cloned, it creates an implicit dependency between the parent and child. Even though the clone is created somewhere else in the dataset hierarchy, the original snapshot cannot be destroyed as long as a clone exists. The .Sy origin property exposes this dependency, and the .Cm destroy command lists any such dependencies, if they exist. .Pp The clone parent-child dependency relationship can be reversed by using the .Cm promote subcommand. This causes the "origin" file system to become a clone of the specified file system, which makes it possible to destroy the file system that the clone was created from. .Ss Mount Points Creating a .Tn ZFS file system is a simple operation, so the number of file systems per system is likely to be numerous. To cope with this, .Tn ZFS automatically manages mounting and unmounting file systems without the need to edit the .Pa /etc/fstab file. All automatically managed file systems are mounted by .Tn ZFS at boot time. .Pp By default, file systems are mounted under .Pa /path , where .Ar path is the name of the file system in the .Tn ZFS namespace. Directories are created and destroyed as needed. .Pp A file system can also have a mount point set in the .Sy mountpoint property. This directory is created as needed, and .Tn ZFS automatically mounts the file system when the .Qq Nm Cm mount Fl a command is invoked (without editing .Pa /etc/fstab ) . The .Sy mountpoint property can be inherited, so if .Em pool/home has a mount point of .Pa /home , then .Em pool/home/user automatically inherits a mount point of .Pa /home/user . .Pp A file system .Sy mountpoint property of .Cm none prevents the file system from being mounted. .Pp If needed, .Tn ZFS file systems can also be managed with traditional tools .Pq Xr mount 8 , Xr umount 8 , Xr fstab 5 . If a file system's mount point is set to .Cm legacy , .Tn ZFS makes no attempt to manage the file system, and the administrator is responsible for mounting and unmounting the file system. .Ss Jails .No A Tn ZFS dataset can be attached to a jail by using the .Qq Nm Cm jail subcommand. You cannot attach a dataset to one jail and the children of the same dataset to another jails. To allow management of the dataset from within a jail, the .Sy jailed property has to be set and the jail needs access to the .Pa /dev/zfs device. The .Sy quota property cannot be changed from within a jail. See .Xr jail 8 for information on how to allow mounting .Tn ZFS datasets from within a jail. .Pp .No A Tn ZFS dataset can be detached from a jail using the .Qq Nm Cm unjail subcommand. .Pp After a dataset is attached to a jail and the jailed property is set, a jailed file system cannot be mounted outside the jail, since the jail administrator might have set the mount point to an unacceptable value. .Ss Deduplication Deduplication is the process for removing redundant data at the block-level, reducing the total amount of data stored. If a file system has the .Cm dedup property enabled, duplicate data blocks are removed synchronously. The result is that only unique data is stored and common components are shared among files. .Ss Native Properties Properties are divided into two types, native properties and user-defined (or "user") properties. Native properties either export internal statistics or control .Tn ZFS behavior. In addition, native properties are either editable or read-only. User properties have no effect on .Tn ZFS behavior, but you can use them to annotate datasets in a way that is meaningful in your environment. For more information about user properties, see the .Qq Sx User Properties section, below. .Pp Every dataset has a set of properties that export statistics about the dataset as well as control various behaviors. Properties are inherited from the parent unless overridden by the child. Some properties apply only to certain types of datasets (file systems, volumes, or snapshots). .Pp The values of numeric properties can be specified using human-readable suffixes (for example, .Sy k , KB , M , Gb , and so forth, up to .Sy Z for zettabyte). The following are all valid (and equal) specifications: .Bd -ragged -offset 4n 1536M, 1.5g, 1.50GB .Ed .Pp The values of non-numeric properties are case sensitive and must be lowercase, except for .Sy mountpoint , sharenfs , No and Sy sharesmb . .Pp The following native properties consist of read-only statistics about the dataset. These properties can be neither set, nor inherited. Native properties apply to all dataset types unless otherwise noted. .Bl -tag -width 2n .It Sy available The amount of space available to the dataset and all its children, assuming that there is no other activity in the pool. Because space is shared within a pool, availability can be limited by any number of factors, including physical pool size, quotas, reservations, or other datasets within the pool. .Pp This property can also be referred to by its shortened column name, .Sy avail . .It Sy compressratio For non-snapshots, the compression ratio achieved for the .Sy used space of this dataset, expressed as a multiplier. The .Sy used property includes descendant datasets, and, for clones, does not include the space shared with the origin snapshot. For snapshots, the .Sy compressratio is the same as the .Sy refcompressratio property. Compression can be turned on by running: .Qq Nm Cm set compression=on Ar dataset The default value is .Cm off . .It Sy creation The time this dataset was created. .It Sy clones For snapshots, this property is a comma-separated list of filesystems or volumes which are clones of this snapshot. The clones' .Sy origin property is this snapshot. If the .Sy clones property is not empty, then this snapshot can not be destroyed (even with the .Fl r or .Fl f options). .It Sy defer_destroy This property is .Cm on if the snapshot has been marked for deferred destroy by using the .Qq Nm Cm destroy -d command. Otherwise, the property is .Cm off . .It Sy filesystem_count The total number of filesystems and volumes that exist under this location in the dataset tree. This value is only available when a .Sy filesystem_limit has been set somewhere in the tree under which the dataset resides. .It Sy logicalreferenced The amount of space that is .Qq logically accessible by this dataset. See the .Sy referenced property. The logical space ignores the effect of the .Sy compression and .Sy copies properties, giving a quantity closer to the amount of data that applications see. However, it does include space consumed by metadata. .Pp This property can also be referred to by its shortened column name, .Sy lrefer . .It Sy logicalused The amount of space that is .Qq logically consumed by this dataset and all its descendents. See the .Sy used property. The logical space ignores the effect of the .Sy compression and .Sy copies properties, giving a quantity closer to the amount of data that applications see. .Pp This property can also be referred to by its shortened column name, .Sy lused . .It Sy mounted For file systems, indicates whether the file system is currently mounted. This property can be either .Cm yes or .Cm no . .It Sy origin For cloned file systems or volumes, the snapshot from which the clone was created. See also the .Sy clones property. .It Sy receive_resume_token For filesystems or volumes which have saved partially-completed state from .Sy zfs receive -s , this opaque token can be provided to .Sy zfs send -t to resume and complete the .Sy zfs receive . .It Sy referenced The amount of data that is accessible by this dataset, which may or may not be shared with other datasets in the pool. When a snapshot or clone is created, it initially references the same amount of space as the file system or snapshot it was created from, since its contents are identical. .Pp This property can also be referred to by its shortened column name, .Sy refer . .It Sy refcompressratio The compression ratio achieved for the .Sy referenced space of this dataset, expressed as a multiplier. See also the .Sy compressratio property. .It Sy snapshot_count The total number of snapshots that exist under this location in the dataset tree. This value is only available when a .Sy snapshot_limit has been set somewhere in the tree under which the dataset resides. .It Sy type The type of dataset: .Sy filesystem , volume , No or Sy snapshot . .It Sy used The amount of space consumed by this dataset and all its descendents. This is the value that is checked against this dataset's quota and reservation. The space used does not include this dataset's reservation, but does take into account the reservations of any descendent datasets. The amount of space that a dataset consumes from its parent, as well as the amount of space that are freed if this dataset is recursively destroyed, is the greater of its space used and its reservation. .Pp When snapshots (see the .Qq Sx Snapshots section) are created, their space is initially shared between the snapshot and the file system, and possibly with previous snapshots. As the file system changes, space that was previously shared becomes unique to the snapshot, and counted in the snapshot's space used. Additionally, deleting snapshots can increase the amount of space unique to (and used by) other snapshots. .Pp The amount of space used, available, or referenced does not take into account pending changes. Pending changes are generally accounted for within a few seconds. Committing a change to a disk using .Xr fsync 2 or .Sy O_SYNC does not necessarily guarantee that the space usage information is updated immediately. .It Sy usedby* The .Sy usedby* properties decompose the .Sy used properties into the various reasons that space is used. Specifically, .Sy used No = .Sy usedbysnapshots + usedbydataset + usedbychildren + usedbyrefreservation . These properties are only available for datasets created with .Tn ZFS pool version 13 pools and higher. .It Sy usedbysnapshots The amount of space consumed by snapshots of this dataset. In particular, it is the amount of space that would be freed if all of this dataset's snapshots were destroyed. Note that this is not simply the sum of the snapshots' .Sy used properties because space can be shared by multiple snapshots. .It Sy usedbydataset The amount of space used by this dataset itself, which would be freed if the dataset were destroyed (after first removing any .Sy refreservation and destroying any necessary snapshots or descendents). .It Sy usedbychildren The amount of space used by children of this dataset, which would be freed if all the dataset's children were destroyed. .It Sy usedbyrefreservation The amount of space used by a .Sy refreservation set on this dataset, which would be freed if the .Sy refreservation was removed. .It Sy userused@ Ns Ar user The amount of space consumed by the specified user in this dataset. Space is charged to the owner of each file, as displayed by .Qq Nm ls Fl l . The amount of space charged is displayed by .Qq Nm du and .Qq Nm ls Fl s . See the .Qq Nm Cm userspace subcommand for more information. .Pp Unprivileged users can access only their own space usage. The root user, or a user who has been granted the .Sy userused privilege with .Qq Nm Cm allow , can access everyone's usage. .Pp The .Sy userused@ Ns ... properties are not displayed by .Qq Nm Cm get all . The user's name must be appended after the .Sy @ symbol, using one of the following forms: .Bl -bullet -offset 2n .It POSIX name (for example, .Em joe ) .It POSIX numeric ID (for example, .Em 1001 ) .El .It Sy userrefs This property is set to the number of user holds on this snapshot. User holds are set by using the .Qq Nm Cm hold command. .It Sy groupused@ Ns Ar group The amount of space consumed by the specified group in this dataset. Space is charged to the group of each file, as displayed by .Nm ls Fl l . See the .Sy userused@ Ns Ar user property for more information. .Pp Unprivileged users can only access their own groups' space usage. The root user, or a user who has been granted the .Sy groupused privilege with .Qq Nm Cm allow , can access all groups' usage. .It Sy volblocksize Ns = Ns Ar blocksize For volumes, specifies the block size of the volume. The .Ar blocksize cannot be changed once the volume has been written, so it should be set at volume creation time. The default .Ar blocksize for volumes is 8 Kbytes. Any power of 2 from 512 bytes to 128 Kbytes is valid. .Pp This property can also be referred to by its shortened column name, .Sy volblock . .It Sy written The amount of .Sy referenced space written to this dataset since the previous snapshot. .It Sy written@ Ns Ar snapshot The amount of .Sy referenced space written to this dataset since the specified snapshot. This is the space that is referenced by this dataset but was not referenced by the specified snapshot. .Pp The .Ar snapshot may be specified as a short snapshot name (just the part after the .Sy @ ) , in which case it will be interpreted as a snapshot in the same filesystem as this dataset. The .Ar snapshot may be a full snapshot name .Pq Em filesystem@snapshot , which for clones may be a snapshot in the origin's filesystem (or the origin of the origin's filesystem, etc). .El .Pp The following native properties can be used to change the behavior of a .Tn ZFS dataset. .Bl -tag -width 2n .It Xo .Sy aclinherit Ns = Ns Cm discard | .Cm noallow | .Cm restricted | .Cm passthrough | .Cm passthrough-x .Xc Controls how .Tn ACL entries are inherited when files and directories are created. A file system with an .Sy aclinherit property of .Cm discard does not inherit any .Tn ACL entries. A file system with an .Sy aclinherit property value of .Cm noallow only inherits inheritable .Tn ACL entries that specify "deny" permissions. The property value .Cm restricted (the default) removes the .Em write_acl and .Em write_owner permissions when the .Tn ACL entry is inherited. A file system with an .Sy aclinherit property value of .Cm passthrough inherits all inheritable .Tn ACL entries without any modifications made to the .Tn ACL entries when they are inherited. A file system with an .Sy aclinherit property value of .Cm passthrough-x has the same meaning as .Cm passthrough , except that the .Em owner@ , group@ , No and Em everyone@ Tn ACE Ns s inherit the execute permission only if the file creation mode also requests the execute bit. .Pp When the property value is set to .Cm passthrough , files are created with a mode determined by the inheritable .Tn ACE Ns s. If no inheritable .Tn ACE Ns s exist that affect the mode, then the mode is set in accordance to the requested mode from the application. .It Sy aclmode Ns = Ns Cm discard | groupmask | passthrough | restricted Controls how an .Tn ACL is modified during .Xr chmod 2 . A file system with an .Sy aclmode property of .Cm discard (the default) deletes all .Tn ACL entries that do not represent the mode of the file. An .Sy aclmode property of .Cm groupmask reduces permissions granted in all .Em ALLOW entries found in the .Tn ACL such that they are no greater than the group permissions specified by .Xr chmod 2 . A file system with an .Sy aclmode property of .Cm passthrough indicates that no changes are made to the .Tn ACL other than creating or updating the necessary .Tn ACL entries to represent the new mode of the file or directory. An .Sy aclmode property of .Cm restricted will cause the .Xr chmod 2 operation to return an error when used on any file or directory which has a non-trivial .Tn ACL whose entries can not be represented by a mode. .Xr chmod 2 is required to change the set user ID, set group ID, or sticky bits on a file or directory, as they do not have equivalent .Tn ACL entries. In order to use .Xr chmod 2 on a file or directory with a non-trivial .Tn ACL when .Sy aclmode is set to .Cm restricted , you must first remove all .Tn ACL entries which do not represent the current mode. .It Sy atime Ns = Ns Cm on | off Controls whether the access time for files is updated when they are read. Turning this property off avoids producing write traffic when reading files and can result in significant performance gains, though it might confuse mailers and other similar utilities. The default value is .Cm on . .It Sy canmount Ns = Ns Cm on | off | noauto If this property is set to .Cm off , the file system cannot be mounted, and is ignored by .Qq Nm Cm mount Fl a . Setting this property to .Cm off is similar to setting the .Sy mountpoint property to .Cm none , except that the dataset still has a normal .Sy mountpoint property, which can be inherited. Setting this property to .Cm off allows datasets to be used solely as a mechanism to inherit properties. One example of setting .Sy canmount Ns = Ns Cm off is to have two datasets with the same .Sy mountpoint , so that the children of both datasets appear in the same directory, but might have different inherited characteristics. .Pp When the .Cm noauto value is set, a dataset can only be mounted and unmounted explicitly. The dataset is not mounted automatically when the dataset is created or imported, nor is it mounted by the .Qq Nm Cm mount Fl a command or unmounted by the .Qq Nm Cm umount Fl a command. .Pp This property is not inherited. .It Sy checksum Ns = Ns Cm on | off | fletcher2 | fletcher4 | sha256 | noparity Controls the checksum used to verify data integrity. The default value is .Cm on , which automatically selects an appropriate algorithm (currently, .Cm fletcher4 , but this may change in future releases). The value .Cm off disables integrity checking on user data. The value .Cm noparity not only disables integrity but also disables maintaining parity for user data. This setting is used internally by a dump device residing on a RAID-Z pool and should not be used by any other dataset. Disabling checksums is .Em NOT a recommended practice. .It Sy compression Ns = Ns Cm on | off | lzjb | gzip | gzip- Ns Ar N | Cm zle | Cm lz4 Controls the compression algorithm used for this dataset. Setting compression to .Cm on indicates that the current default compression algorithm should be used. The default balances compression and decompression speed, with compression ratio and is expected to work well on a wide variety of workloads. Unlike all other settings for this property, on does not select a fixed compression type. As new compression algorithms are added to ZFS and enabled on a pool, the default compression algorithm may change. The current default compression algorthm is either .Cm lzjb or, if the .Sy lz4_compress feature is enabled, .Cm lz4 . The .Cm lzjb compression algorithm is optimized for performance while providing decent data compression. Setting compression to .Cm on uses the .Cm lzjb compression algorithm. The .Cm gzip compression algorithm uses the same compression as the .Xr gzip 1 command. You can specify the .Cm gzip level by using the value .Cm gzip- Ns Ar N where .Ar N is an integer from 1 (fastest) to 9 (best compression ratio). Currently, .Cm gzip is equivalent to .Cm gzip-6 (which is also the default for .Xr gzip 1 ) . The .Cm zle compression algorithm compresses runs of zeros. .Pp The .Sy lz4 compression algorithm is a high-performance replacement for the .Sy lzjb algorithm. It features significantly faster compression and decompression, as well as a moderately higher compression ratio than .Sy lzjb , but can only be used on pools with the .Sy lz4_compress feature set to .Sy enabled . See .Xr zpool-features 7 for details on ZFS feature flags and the .Sy lz4_compress feature. .Pp This property can also be referred to by its shortened column name .Cm compress . Changing this property affects only newly-written data. .It Sy copies Ns = Ns Cm 1 | 2 | 3 Controls the number of copies of data stored for this dataset. These copies are in addition to any redundancy provided by the pool, for example, mirroring or RAID-Z. The copies are stored on different disks, if possible. The space used by multiple copies is charged to the associated file and dataset, changing the .Sy used property and counting against quotas and reservations. .Pp Changing this property only affects newly-written data. Therefore, set this property at file system creation time by using the .Fl o Cm copies= Ns Ar N option. .It Sy dedup Ns = Ns Cm on | off | verify | sha256 Ns Op Cm ,verify Configures deduplication for a dataset. The default value is .Cm off . The default deduplication checksum is .Cm sha256 (this may change in the future). When .Sy dedup is enabled, the checksum defined here overrides the .Sy checksum property. Setting the value to .Cm verify has the same effect as the setting .Cm sha256,verify . .Pp If set to .Cm verify , .Tn ZFS will do a byte-to-byte comparsion in case of two blocks having the same signature to make sure the block contents are identical. .It Sy devices Ns = Ns Cm on | off The .Sy devices property is currently not supported on .Fx . .It Sy exec Ns = Ns Cm on | off Controls whether processes can be executed from within this file system. The default value is .Cm on . .It Sy mlslabel Ns = Ns Ar label | Cm none The .Sy mlslabel property is currently not supported on .Fx . .It Sy filesystem_limit Ns = Ns Ar count | Cm none Limits the number of filesystems and volumes that can exist under this point in the dataset tree. The limit is not enforced if the user is allowed to change the limit. Setting a .Sy filesystem_limit on a descendent of a filesystem that already has a .Sy filesystem_limit does not override the ancestor's .Sy filesystem_limit , but rather imposes an additional limit. This feature must be enabled to be used .Po see .Xr zpool-features 7 .Pc . .It Sy mountpoint Ns = Ns Ar path | Cm none | legacy Controls the mount point used for this file system. See the .Qq Sx Mount Points section for more information on how this property is used. .Pp When the .Sy mountpoint property is changed for a file system, the file system and any children that inherit the mount point are unmounted. If the new value is .Cm legacy , then they remain unmounted. Otherwise, they are automatically remounted in the new location if the property was previously .Cm legacy or .Cm none , or if they were mounted before the property was changed. In addition, any shared file systems are unshared and shared in the new location. .It Sy nbmand Ns = Ns Cm on | off The .Sy nbmand property is currently not supported on .Fx . .It Sy primarycache Ns = Ns Cm all | none | metadata Controls what is cached in the primary cache (ARC). If this property is set to .Cm all , then both user data and metadata is cached. If this property is set to .Cm none , then neither user data nor metadata is cached. If this property is set to .Cm metadata , then only metadata is cached. The default value is .Cm all . .It Sy quota Ns = Ns Ar size | Cm none Limits the amount of space a dataset and its descendents can consume. This property enforces a hard limit on the amount of space used. This includes all space consumed by descendents, including file systems and snapshots. Setting a quota on a descendent of a dataset that already has a quota does not override the ancestor's quota, but rather imposes an additional limit. .Pp Quotas cannot be set on volumes, as the .Sy volsize property acts as an implicit quota. .It Sy snapshot_limit Ns = Ns Ar count | Cm none Limits the number of snapshots that can be created on a dataset and its descendents. Setting a .Sy snapshot_limit on a descendent of a dataset that already has a .Sy snapshot_limit does not override the ancestor's .Sy snapshot_limit , but rather imposes an additional limit. The limit is not enforced if the user is allowed to change the limit. For example, this means that recursive snapshots taken from the global zone are counted against each delegated dataset within a jail. This feature must be enabled to be used .Po see .Xr zpool-features 7 .Pc . .It Sy userquota@ Ns Ar user Ns = Ns Ar size | Cm none Limits the amount of space consumed by the specified user. Similar to the .Sy refquota property, the .Sy userquota space calculation does not include space that is used by descendent datasets, such as snapshots and clones. User space consumption is identified by the .Sy userspace@ Ns Ar user property. .Pp Enforcement of user quotas may be delayed by several seconds. This delay means that a user might exceed their quota before the system notices that they are over quota and begins to refuse additional writes with the .Em EDQUOT error message. See the .Cm userspace subcommand for more information. .Pp Unprivileged users can only access their own groups' space usage. The root user, or a user who has been granted the .Sy userquota privilege with .Qq Nm Cm allow , can get and set everyone's quota. .Pp This property is not available on volumes, on file systems before version 4, or on pools before version 15. The .Sy userquota@ Ns ... properties are not displayed by .Qq Nm Cm get all . The user's name must be appended after the .Sy @ symbol, using one of the following forms: .Bl -bullet -offset 2n .It POSIX name (for example, .Em joe ) .It POSIX numeric ID (for example, .Em 1001 ) .El .It Sy groupquota@ Ns Ar group Ns = Ns Ar size | Cm none Limits the amount of space consumed by the specified group. Group space consumption is identified by the .Sy userquota@ Ns Ar user property. .Pp Unprivileged users can access only their own groups' space usage. The root user, or a user who has been granted the .Sy groupquota privilege with .Qq Nm Cm allow , can get and set all groups' quotas. .It Sy readonly Ns = Ns Cm on | off Controls whether this dataset can be modified. The default value is .Cm off . .It Sy recordsize Ns = Ns Ar size Specifies a suggested block size for files in the file system. This property is designed solely for use with database workloads that access files in fixed-size records. .Tn ZFS automatically tunes block sizes according to internal algorithms optimized for typical access patterns. .Pp For databases that create very large files but access them in small random chunks, these algorithms may be suboptimal. Specifying a .Sy recordsize greater than or equal to the record size of the database can result in significant performance gains. Use of this property for general purpose file systems is strongly discouraged, and may adversely affect performance. .Pp The size specified must be a power of two greater than or equal to 512 and less than or equal to 128 Kbytes. If the .Sy large_blocks feature is enabled on the pool, the size may be up to 1 Mbyte. See .Xr zpool-features 7 for details on ZFS feature flags. .Pp Changing the file system's .Sy recordsize affects only files created afterward; existing files are unaffected. .Pp This property can also be referred to by its shortened column name, .Sy recsize . .It Sy redundant_metadata Ns = Ns Cm all | most Controls what types of metadata are stored redundantly. ZFS stores an extra copy of metadata, so that if a single block is corrupted, the amount of user data lost is limited. This extra copy is in addition to any redundancy provided at the pool level .Pq e.g. by mirroring or RAID-Z , and is in addition to an extra copy specified by the .Sy copies property .Pq up to a total of 3 copies . For example if the pool is mirrored, .Cm copies Ns = Ns Ar 2 , and .Cm redundant_metadata Ns = Ns Ar most , then ZFS stores 6 copies of most metadata, and 4 copies of data and some metadata. .Pp When set to .Cm all , ZFS stores an extra copy of all metadata. If a single on-disk block is corrupt, at worst a single block of user data .Po which is .Cm recordsize bytes long can be lost. .Pc .Pp When set to .Cm most , ZFS stores an extra copy of most types of metadata. This can improve performance of random writes, because less metadata must be written. In practice, at worst about 100 blocks .Po of .Cm recordsize bytes each .Pc of user data can be lost if a single on-disk block is corrupt. The exact behavior of which metadata blocks are stored redundantly may change in future releases. .Pp The default value is .Cm all . .It Sy refquota Ns = Ns Ar size | Cm none Limits the amount of space a dataset can consume. This property enforces a hard limit on the amount of space used. This hard limit does not include space used by descendents, including file systems and snapshots. .It Sy refreservation Ns = Ns Ar size | Cm none The minimum amount of space guaranteed to a dataset, not including its descendents. When the amount of space used is below this value, the dataset is treated as if it were taking up the amount of space specified by .Sy refreservation . The .Sy refreservation reservation is accounted for in the parent datasets' space used, and counts against the parent datasets' quotas and reservations. .Pp If .Sy refreservation is set, a snapshot is only allowed if there is enough free pool space outside of this reservation to accommodate the current number of "referenced" bytes in the dataset. .Pp This property can also be referred to by its shortened column name, .Sy refreserv . .It Sy reservation Ns = Ns Ar size | Cm none The minimum amount of space guaranteed to a dataset and its descendents. When the amount of space used is below this value, the dataset is treated as if it were taking up the amount of space specified by its reservation. Reservations are accounted for in the parent datasets' space used, and count against the parent datasets' quotas and reservations. .Pp This property can also be referred to by its shortened column name, .Sy reserv . .It Sy secondarycache Ns = Ns Cm all | none | metadata Controls what is cached in the secondary cache (L2ARC). If this property is set to .Cm all , then both user data and metadata is cached. If this property is set to .Cm none , then neither user data nor metadata is cached. If this property is set to .Cm metadata , then only metadata is cached. The default value is .Cm all . .It Sy setuid Ns = Ns Cm on | off Controls whether the .No set- Ns Tn UID bit is respected for the file system. The default value is .Cm on . .It Sy sharesmb Ns = Ns Cm on | off | Ar opts The .Sy sharesmb property currently has no effect on .Fx . .It Sy sharenfs Ns = Ns Cm on | off | Ar opts Controls whether the file system is shared via .Tn NFS , and what options are used. A file system with a .Sy sharenfs property of .Cm off is managed the traditional way via .Xr exports 5 . Otherwise, the file system is automatically shared and unshared with the .Qq Nm Cm share and .Qq Nm Cm unshare commands. If the property is set to .Cm on no .Tn NFS export options are used. Otherwise, .Tn NFS export options are equivalent to the contents of this property. The export options may be comma-separated. See .Xr exports 5 for a list of valid options. .Pp When the .Sy sharenfs property is changed for a dataset, the .Xr mountd 8 daemon is reloaded. .It Sy logbias Ns = Ns Cm latency | throughput Provide a hint to .Tn ZFS about handling of synchronous requests in this dataset. If .Sy logbias is set to .Cm latency (the default), .Tn ZFS will use pool log devices (if configured) to handle the requests at low latency. If .Sy logbias is set to .Cm throughput , .Tn ZFS will not use configured pool log devices. .Tn ZFS will instead optimize synchronous operations for global pool throughput and efficient use of resources. .It Sy snapdir Ns = Ns Cm hidden | visible Controls whether the .Pa \&.zfs directory is hidden or visible in the root of the file system as discussed in the .Qq Sx Snapshots section. The default value is .Cm hidden . .It Sy sync Ns = Ns Cm standard | always | disabled Controls the behavior of synchronous requests (e.g. .Xr fsync 2 , O_DSYNC). This property accepts the following values: .Bl -tag -offset 4n -width 8n .It Sy standard This is the POSIX specified behavior of ensuring all synchronous requests are written to stable storage and all devices are flushed to ensure data is not cached by device controllers (this is the default). .It Sy always All file system transactions are written and flushed before their system calls return. This has a large performance penalty. .It Sy disabled Disables synchronous requests. File system transactions are only committed to stable storage periodically. This option will give the highest performance. However, it is very dangerous as .Tn ZFS would be ignoring the synchronous transaction demands of applications such as databases or .Tn NFS . Administrators should only use this option when the risks are understood. .El .It Sy volsize Ns = Ns Ar size For volumes, specifies the logical size of the volume. By default, creating a volume establishes a reservation of equal size. For storage pools with a version number of 9 or higher, a .Sy refreservation is set instead. Any changes to .Sy volsize are reflected in an equivalent change to the reservation (or .Sy refreservation ) . The .Sy volsize can only be set to a multiple of .Cm volblocksize , and cannot be zero. .Pp The reservation is kept equal to the volume's logical size to prevent unexpected behavior for consumers. Without the reservation, the volume could run out of space, resulting in undefined behavior or data corruption, depending on how the volume is used. These effects can also occur when the volume size is changed while it is in use (particularly when shrinking the size). Extreme care should be used when adjusting the volume size. .Pp Though not recommended, a "sparse volume" (also known as "thin provisioning") can be created by specifying the .Fl s option to the .Qq Nm Cm create Fl V command, or by changing the reservation after the volume has been created. A "sparse volume" is a volume where the reservation is less then the volume size. Consequently, writes to a sparse volume can fail with .Sy ENOSPC when the pool is low on space. For a sparse volume, changes to .Sy volsize are not reflected in the reservation. .It Sy volmode Ns = Ns Cm default | geom | dev | none This property specifies how volumes should be exposed to the OS. Setting it to .Sy geom exposes volumes as .Xr geom 4 providers, providing maximal functionality. Setting it to .Sy dev exposes volumes only as cdev device in devfs. Such volumes can be accessed only as raw disk device files, i.e. they can not be partitioned, mounted, participate in RAIDs, etc, but they are faster, and in some use scenarios with untrusted consumer, such as NAS or VM storage, can be more safe. Volumes with property set to .Sy none are not exposed outside ZFS, but can be snapshoted, cloned, replicated, etc, that can be suitable for backup purposes. Value .Sy default means that volumes exposition is controlled by system-wide sysctl/tunable .Va vfs.zfs.vol.mode , where .Sy geom , .Sy dev and .Sy none are encoded as 1, 2 and 3 respectively. The default values is .Sy geom . This property can be changed any time, but so far it is processed only during volume creation and pool import. .It Sy vscan Ns = Ns Cm off | on The .Sy vscan property is currently not supported on .Fx . .It Sy xattr Ns = Ns Cm off | on The .Sy xattr property is currently not supported on .Fx . .It Sy jailed Ns = Ns Cm off | on Controls whether the dataset is managed from a jail. See the .Qq Sx Jails section for more information. The default value is .Cm off . .El .Pp The following three properties cannot be changed after the file system is created, and therefore, should be set when the file system is created. If the properties are not set with the .Qq Nm Cm create or .Nm zpool Cm create commands, these properties are inherited from the parent dataset. If the parent dataset lacks these properties due to having been created prior to these features being supported, the new file system will have the default values for these properties. .Bl -tag -width 4n .It Sy casesensitivity Ns = Ns Cm sensitive | insensitive | mixed Indicates whether the file name matching algorithm used by the file system should be case-sensitive, case-insensitive, or allow a combination of both styles of matching. The default value for the .Sy casesensitivity property is .Cm sensitive . Traditionally, UNIX and POSIX file systems have case-sensitive file names. .Pp The .Cm mixed value for the .Sy casesensitivity property indicates that the file system can support requests for both case-sensitive and case-insensitive matching behavior. .It Sy normalization Ns = Ns Cm none | formC | formD | formKC | formKD Indicates whether the file system should perform a .Sy unicode normalization of file names whenever two file names are compared, and which normalization algorithm should be used. File names are always stored unmodified, names are normalized as part of any comparison process. If this property is set to a legal value other than .Cm none , and the .Sy utf8only property was left unspecified, the .Sy utf8only property is automatically set to .Cm on . The default value of the .Sy normalization property is .Cm none . This property cannot be changed after the file system is created. .It Sy utf8only Ns = Ns Cm on | off Indicates whether the file system should reject file names that include characters that are not present in the .Sy UTF-8 character code set. If this property is explicitly set to .Cm off , the normalization property must either not be explicitly set or be set to .Cm none . The default value for the .Sy utf8only property is .Cm off . This property cannot be changed after the file system is created. .El .Pp The .Sy casesensitivity , normalization , No and Sy utf8only properties are also new permissions that can be assigned to non-privileged users by using the .Tn ZFS delegated administration feature. .Ss Temporary Mount Point Properties When a file system is mounted, either through .Xr mount 8 for legacy mounts or the .Qq Nm Cm mount command for normal file systems, its mount options are set according to its properties. The correlation between properties and mount options is as follows: .Bl -column -offset 4n "PROPERTY" "MOUNT OPTION" .It "PROPERTY MOUNT OPTION" .It "atime atime/noatime" .It "exec exec/noexec" .It "readonly ro/rw" .It "setuid suid/nosuid" .El .Pp In addition, these options can be set on a per-mount basis using the .Fl o option, without affecting the property that is stored on disk. The values specified on the command line override the values stored in the dataset. These properties are reported as "temporary" by the .Qq Nm Cm get command. If the properties are changed while the dataset is mounted, the new setting overrides any temporary settings. .Ss User Properties In addition to the standard native properties, .Tn ZFS supports arbitrary user properties. User properties have no effect on .Tn ZFS behavior, but applications or administrators can use them to annotate datasets (file systems, volumes, and snapshots). .Pp User property names must contain a colon .Pq Sy \&: character to distinguish them from native properties. They may contain lowercase letters, numbers, and the following punctuation characters: colon .Pq Sy \&: , dash .Pq Sy \&- , period .Pq Sy \&. and underscore .Pq Sy \&_ . The expected convention is that the property name is divided into two portions such as .Em module Ns Sy \&: Ns Em property , but this namespace is not enforced by .Tn ZFS . User property names can be at most 256 characters, and cannot begin with a dash .Pq Sy \&- . .Pp When making programmatic use of user properties, it is strongly suggested to use a reversed .Tn DNS domain name for the .Ar module component of property names to reduce the chance that two independently-developed packages use the same property name for different purposes. Property names beginning with .Em com.sun are reserved for use by Sun Microsystems. .Pp The values of user properties are arbitrary strings, are always inherited, and are never validated. All of the commands that operate on properties .Po .Qq Nm Cm list , .Qq Nm Cm get , .Qq Nm Cm set and so forth .Pc can be used to manipulate both native properties and user properties. Use the .Qq Nm Cm inherit command to clear a user property. If the property is not defined in any parent dataset, it is removed entirely. Property values are limited to 1024 characters. .Sh SUBCOMMANDS All subcommands that modify state are logged persistently to the pool in their original form. .Bl -tag -width 2n .It Xo .Nm .Op Fl \&? .Xc .Pp Displays a help message. .It Xo .Nm .Cm create .Op Fl pu .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Ar filesystem .Xc .Pp Creates a new .Tn ZFS file system. The file system is automatically mounted according to the .Sy mountpoint property inherited from the parent. .Bl -tag -width indent .It Fl p Creates all the non-existing parent datasets. Datasets created in this manner are automatically mounted according to the .Sy mountpoint property inherited from their parent. Any property specified on the command line using the .Fl o option is ignored. If the target filesystem already exists, the operation completes successfully. .It Fl u Newly created file system is not mounted. .It Fl o Ar property Ns = Ns Ar value Sets the specified property as if the command .Qq Nm Cm set Ar property Ns = Ns Ar value was invoked at the same time the dataset was created. Any editable .Tn ZFS property can also be set at creation time. Multiple .Fl o options can be specified. An error results if the same property is specified in multiple .Fl o options. .El .It Xo .Nm .Cm create .Op Fl ps .Op Fl b Ar blocksize .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Fl V .Ar size volume .Xc .Pp Creates a volume of the given size. The volume is exported as a block device in .Pa /dev/zvol/path , where .Ar path is the name of the volume in the .Tn ZFS namespace. The size represents the logical size as exported by the device. By default, a reservation of equal size is created. .Pp .Ar size is automatically rounded up to the nearest 128 Kbytes to ensure that the volume has an integral number of blocks regardless of .Ar blocksize . .Bl -tag -width indent .It Fl p Creates all the non-existing parent datasets. Datasets created in this manner are automatically mounted according to the .Sy mountpoint property inherited from their parent. Any property specified on the command line using the .Fl o option is ignored. If the target filesystem already exists, the operation completes successfully. .It Fl s Creates a sparse volume with no reservation. See .Sy volsize in the .Qq Sx Native Properties section for more information about sparse volumes. .It Fl b Ar blocksize Equivalent to .Fl o Cm volblocksize Ns = Ns Ar blocksize . If this option is specified in conjunction with .Fl o Cm volblocksize , the resulting behavior is undefined. .It Fl o Ar property Ns = Ns Ar value Sets the specified property as if the .Qq Nm Cm set Ar property Ns = Ns Ar value command was invoked at the same time the dataset was created. Any editable .Tn ZFS property can also be set at creation time. Multiple .Fl o options can be specified. An error results if the same property is specified in multiple .Fl o options. .El .It Xo .Nm .Cm destroy .Op Fl fnpRrv .Ar filesystem Ns | Ns Ar volume .Xc .Pp Destroys the given dataset. By default, the command unshares any file systems that are currently shared, unmounts any file systems that are currently mounted, and refuses to destroy a dataset that has active dependents (children or clones). .Bl -tag -width indent .It Fl r Recursively destroy all children. .It Fl R Recursively destroy all dependents, including cloned file systems outside the target hierarchy. .It Fl f Force an unmount of any file systems using the .Qq Nm Cm unmount Fl f command. This option has no effect on non-file systems or unmounted file systems. .It Fl n Do a dry-run ("No-op") deletion. No data will be deleted. This is useful in conjunction with the .Fl v or .Fl p flags to determine what data would be deleted. .It Fl p Print machine-parsable verbose information about the deleted data. .It Fl v Print verbose information about the deleted data. .El .Pp Extreme care should be taken when applying either the .Fl r or the .Fl R options, as they can destroy large portions of a pool and cause unexpected behavior for mounted file systems in use. .It Xo .Nm .Cm destroy .Op Fl dnpRrv .Sm off .Ar snapshot .Op % Ns Ar snapname .Op , Ns ... .Sm on .Xc .Pp The given snapshots are destroyed immediately if and only if the .Qq Nm Cm destroy command without the .Fl d option would have destroyed it. Such immediate destruction would occur, for example, if the snapshot had no clones and the user-initiated reference count were zero. .Pp If a snapshot does not qualify for immediate destruction, it is marked for deferred deletion. In this state, it exists as a usable, visible snapshot until both of the preconditions listed above are met, at which point it is destroyed. .Pp An inclusive range of snapshots may be specified by separating the first and last snapshots with a percent sign .Pq Sy % . The first and/or last snapshots may be left blank, in which case the filesystem's oldest or newest snapshot will be implied. .Pp Multiple snapshots (or ranges of snapshots) of the same filesystem or volume may be specified in a comma-separated list of snapshots. Only the snapshot's short name (the part after the .Sy @ ) should be specified when using a range or comma-separated list to identify multiple snapshots. .Bl -tag -width indent .It Fl r Destroy (or mark for deferred deletion) all snapshots with this name in descendent file systems. .It Fl R Recursively destroy all clones of these snapshots, including the clones, snapshots, and children. If this flag is specified, the .Fl d flag will have no effect. .It Fl n Do a dry-run ("No-op") deletion. No data will be deleted. This is useful in conjunction with the .Fl v or .Fl p flags to determine what data would be deleted. .It Fl p Print machine-parsable verbose information about the deleted data. .It Fl v Print verbose information about the deleted data. .It Fl d Defer snapshot deletion. .El .Pp Extreme care should be taken when applying either the .Fl r or the .Fl R options, as they can destroy large portions of a pool and cause unexpected behavior for mounted file systems in use. .It Xo .Nm .Cm destroy .Ar filesystem Ns | Ns Ar volume Ns # Ns Ar bookmark .Xc .Pp The given bookmark is destroyed. .It Xo .Nm .Cm snapshot Ns | Ns Cm snap .Op Fl r .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Ar filesystem@snapname Ns | Ns volume@snapname .Ar filesystem@snapname Ns | Ns volume@snapname Ns ... .Xc .Pp Creates snapshots with the given names. All previous modifications by successful system calls to the file system are part of the snapshots. Snapshots are taken atomically, so that all snapshots correspond to the same moment in time. See the .Qq Sx Snapshots section for details. .Bl -tag -width indent .It Fl r Recursively create snapshots of all descendent datasets .It Fl o Ar property Ns = Ns Ar value Sets the specified property; see .Qq Nm Cm create for details. .El .It Xo .Nm .Cm rollback .Op Fl rRf .Ar snapshot .Xc .Pp Roll back the given dataset to a previous snapshot. When a dataset is rolled back, all data that has changed since the snapshot is discarded, and the dataset reverts to the state at the time of the snapshot. By default, the command refuses to roll back to a snapshot other than the most recent one. In order to do so, all intermediate snapshots and bookmarks must be destroyed by specifying the .Fl r option. .Pp The .Fl rR options do not recursively destroy the child snapshots of a recursive snapshot. Only direct snapshots of the specified filesystem are destroyed by either of these options. To completely roll back a recursive snapshot, you must rollback the individual child snapshots. .Bl -tag -width indent .It Fl r Destroy any snapshots and bookmarks more recent than the one specified. .It Fl R Destroy any more recent snapshots and bookmarks, as well as any clones of those snapshots. .It Fl f Used with the .Fl R option to force an unmount of any clone file systems that are to be destroyed. .El .It Xo .Nm .Cm clone .Op Fl p .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Ar snapshot filesystem Ns | Ns Ar volume .Xc .Pp Creates a clone of the given snapshot. See the .Qq Sx Clones section for details. The target dataset can be located anywhere in the .Tn ZFS hierarchy, and is created as the same type as the original. .Bl -tag -width indent .It Fl p Creates all the non-existing parent datasets. Datasets created in this manner are automatically mounted according to the .Sy mountpoint property inherited from their parent. If the target filesystem or volume already exists, the operation completes successfully. .It Fl o Ar property Ns = Ns Ar value Sets the specified property; see .Qq Nm Cm create for details. .El .It Xo .Nm .Cm promote .Ar clone-filesystem .Xc .Pp Promotes a clone file system to no longer be dependent on its "origin" snapshot. This makes it possible to destroy the file system that the clone was created from. The clone parent-child dependency relationship is reversed, so that the origin file system becomes a clone of the specified file system. .Pp The snapshot that was cloned, and any snapshots previous to this snapshot, are now owned by the promoted clone. The space they use moves from the origin file system to the promoted clone, so enough space must be available to accommodate these snapshots. No new space is consumed by this operation, but the space accounting is adjusted. The promoted clone must not have any conflicting snapshot names of its own. The .Cm rename subcommand can be used to rename any conflicting snapshots. .It Xo .Nm .Cm rename .Op Fl f .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot .Xc .It Xo .Nm .Cm rename .Op Fl f .Fl p .Ar filesystem Ns | Ns Ar volume .Ar filesystem Ns | Ns Ar volume .Xc .It Xo .Nm .Cm rename .Fl u .Op Fl p .Ar filesystem filesystem .Xc .Pp Renames the given dataset. The new target can be located anywhere in the .Tn ZFS hierarchy, with the exception of snapshots. Snapshots can only be renamed within the parent file system or volume. When renaming a snapshot, the parent file system of the snapshot does not need to be specified as part of the second argument. Renamed file systems can inherit new mount points, in which case they are unmounted and remounted at the new mount point. .Bl -tag -width indent .It Fl p Creates all the nonexistent parent datasets. Datasets created in this manner are automatically mounted according to the .Sy mountpoint property inherited from their parent. .It Fl u Do not remount file systems during rename. If a file system's .Sy mountpoint property is set to .Cm legacy or .Cm none , file system is not unmounted even if this option is not given. .It Fl f Force unmount any filesystems that need to be unmounted in the process. This flag has no effect if used together with the .Fl u flag. .El .It Xo .Nm .Cm rename .Fl r .Ar snapshot snapshot .Xc .Pp Recursively rename the snapshots of all descendent datasets. Snapshots are the only dataset that can be renamed recursively. .It Xo .Nm .Cm list .Op Fl r Ns | Ns Fl d Ar depth .Op Fl Hp .Op Fl o Ar property Ns Oo , Ns Ar property Oc Ns ... .Op Fl t Ar type Ns Oo , Ns Ar type Oc Ns ... .Oo Fl s Ar property Oc Ns ... .Oo Fl S Ar property Oc Ns ... .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot Ns ... .Xc .Pp Lists the property information for the given datasets in tabular form. If specified, you can list property information by the absolute pathname or the relative pathname. By default, all file systems and volumes are displayed. Snapshots are displayed if the .Sy listsnaps property is .Cm on (the default is .Cm off ) . The following fields are displayed, .Sy name , used , available , referenced , mountpoint . .Bl -tag -width indent .It Fl r Recursively display any children of the dataset on the command line. .It Fl d Ar depth Recursively display any children of the dataset, limiting the recursion to .Ar depth . A depth of .Sy 1 will display only the dataset and its direct children. .It Fl H Used for scripting mode. Do not print headers and separate fields by a single tab instead of arbitrary white space. .It Fl p Display numbers in parsable (exact) values. .It Fl o Ar property Ns Oo , Ns Ar property Oc Ns ... A comma-separated list of properties to display. The property must be: .Bl -bullet -offset 2n .It One of the properties described in the .Qq Sx Native Properties section .It A user property .It The value .Cm name to display the dataset name .It The value .Cm space to display space usage properties on file systems and volumes. This is a shortcut for specifying .Fl o .Sy name,avail,used,usedsnap,usedds,usedrefreserv,usedchild .Fl t .Sy filesystem,volume syntax. .El .It Fl t Ar type Ns Oo , Ns Ar type Oc Ns ... A comma-separated list of types to display, where .Ar type is one of .Sy filesystem , snapshot , snap , volume , bookmark , No or Sy all . For example, specifying .Fl t Cm snapshot displays only snapshots. .It Fl s Ar property A property for sorting the output by column in ascending order based on the value of the property. The property must be one of the properties described in the .Qq Sx Properties section, or the special value .Cm name to sort by the dataset name. Multiple properties can be specified at one time using multiple .Fl s property options. Multiple .Fl s options are evaluated from left to right in decreasing order of importance. .Pp The following is a list of sorting criteria: .Bl -bullet -offset 2n .It Numeric types sort in numeric order. .It String types sort in alphabetical order. .It Types inappropriate for a row sort that row to the literal bottom, regardless of the specified ordering. .It If no sorting options are specified the existing behavior of .Qq Nm Cm list is preserved. .El .It Fl S Ar property Same as the .Fl s option, but sorts by property in descending order. .El .It Xo .Nm .Cm set .Ar property Ns = Ns Ar value Oo Ar property Ns = Ns Ar value Oc Ns ... .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot .Xc .Pp Sets the property or list of properties to the given value(s) for each dataset. Only some properties can be edited. See the "Properties" section for more information on what properties can be set and acceptable values. Numeric values can be specified as exact values, or in a human-readable form with a suffix of .Sy B , K , M , G , T , P , E , Z (for bytes, kilobytes, megabytes, gigabytes, terabytes, petabytes, exabytes, or zettabytes, respectively). User properties can be set on snapshots. For more information, see the .Qq Sx User Properties section. .It Xo .Nm .Cm get .Op Fl r Ns | Ns Fl d Ar depth .Op Fl Hp .Op Fl o Ar all | field Ns Oo , Ns Ar field Oc Ns ... .Op Fl t Ar type Ns Oo , Ns Ar type Oc Ns ... .Op Fl s Ar source Ns Oo , Ns Ar source Oc Ns ... .Ar all | property Ns Oo , Ns Ar property Oc Ns ... .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot Ns ... .Xc .Pp Displays properties for the given datasets. If no datasets are specified, then the command displays properties for all datasets on the system. For each property, the following columns are displayed: .Pp .Bl -hang -width "property" -offset indent -compact .It name Dataset name .It property Property name .It value Property value .It source Property source. Can either be local, default, temporary, inherited, received, or none (\&-). .El .Pp All columns except the .Sy RECEIVED column are displayed by default. The columns to display can be specified by using the .Fl o option. This command takes a comma-separated list of properties as described in the .Qq Sx Native Properties and .Qq Sx User Properties sections. .Pp The special value .Cm all can be used to display all properties that apply to the given dataset's type (filesystem, volume, snapshot, or bookmark). .Bl -tag -width indent .It Fl r Recursively display properties for any children. .It Fl d Ar depth Recursively display any children of the dataset, limiting the recursion to .Ar depth . A depth of .Sy 1 will display only the dataset and its direct children. .It Fl H Display output in a form more easily parsed by scripts. Any headers are omitted, and fields are explicitly separated by a single tab instead of an arbitrary amount of space. .It Fl p Display numbers in parsable (exact) values. .It Fl o Cm all | Ar field Ns Oo , Ns Ar field Oc Ns ... A comma-separated list of columns to display. Supported values are .Sy name,property,value,received,source . Default values are .Sy name,property,value,source . The keyword .Cm all specifies all columns. .It Fl t Ar type Ns Oo , Ns Ar type Oc Ns ... A comma-separated list of types to display, where .Ar type is one of .Sy filesystem , snapshot , volume , No or Sy all . For example, specifying .Fl t Cm snapshot displays only snapshots. .It Fl s Ar source Ns Oo , Ns Ar source Oc Ns ... A comma-separated list of sources to display. Those properties coming from a source other than those in this list are ignored. Each source must be one of the following: .Sy local,default,inherited,temporary,received,none . The default value is all sources. .El .It Xo .Nm .Cm inherit .Op Fl rS .Ar property .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot Ns ... .Xc .Pp Clears the specified property, causing it to be inherited from an ancestor, restored to default if no ancestor has the property set, or with the .Fl S option reverted to the received value if one exists. See the .Qq Sx Properties section for a listing of default values, and details on which properties can be inherited. .Bl -tag -width indent .It Fl r Recursively inherit the given property for all children. .It Fl S Revert the property to the received value if one exists; otherwise operate as if the .Fl S option was not specified. .El .It Xo .Nm .Cm upgrade .Op Fl v .Xc .Pp Displays a list of file systems that are not the most recent version. .Bl -tag -width indent .It Fl v Displays .Tn ZFS filesystem versions supported by the current software. The current .Tn ZFS filesystem version and all previous supported versions are displayed, along with an explanation of the features provided with each version. .El .It Xo .Nm .Cm upgrade .Op Fl r .Op Fl V Ar version .Fl a | Ar filesystem .Xc .Pp Upgrades file systems to a new on-disk version. Once this is done, the file systems will no longer be accessible on systems running older versions of the software. .Qq Nm Cm send streams generated from new snapshots of these file systems cannot be accessed on systems running older versions of the software. .Pp In general, the file system version is independent of the pool version. See .Xr zpool 8 for information on the .Nm zpool Cm upgrade command. .Pp In some cases, the file system version and the pool version are interrelated and the pool version must be upgraded before the file system version can be upgraded. .Bl -tag -width indent .It Fl r Upgrade the specified file system and all descendent file systems. .It Fl V Ar version Upgrade to the specified .Ar version . If the .Fl V flag is not specified, this command upgrades to the most recent version. This option can only be used to increase the version number, and only up to the most recent version supported by this software. .It Fl a Upgrade all file systems on all imported pools. .It Ar filesystem Upgrade the specified file system. .El .It Xo .Nm .Cm userspace .Op Fl Hinp .Op Fl o Ar field Ns Oo , Ns Ar field Oc Ns ... .Oo Fl s Ar field Oc Ns ... .Oo Fl S Ar field Oc Ns ... .Op Fl t Ar type Ns Oo , Ns Ar type Oc Ns ... .Ar filesystem Ns | Ns Ar snapshot .Xc .Pp Displays space consumed by, and quotas on, each user in the specified filesystem or snapshot. This corresponds to the .Sy userused@ Ns Ar user and .Sy userquota@ Ns Ar user properties. .Bl -tag -width indent .It Fl n Print numeric ID instead of user/group name. .It Fl H Do not print headers, use tab-delimited output. .It Fl p Use exact (parsable) numeric output. .It Fl o Ar field Ns Oo , Ns Ar field Oc Ns ... Display only the specified fields from the following set: .Sy type,name,used,quota . The default is to display all fields. .It Fl s Ar field Sort output by this field. The .Fl s and .Fl S flags may be specified multiple times to sort first by one field, then by another. The default is .Fl s Cm type Fl s Cm name . .It Fl S Ar field Sort by this field in reverse order. See .Fl s . .It Fl t Ar type Ns Oo , Ns Ar type Oc Ns ... Print only the specified types from the following set: .Sy all,posixuser,smbuser,posixgroup,smbgroup . .Pp The default is .Fl t Cm posixuser,smbuser . .Pp The default can be changed to include group types. .It Fl i Translate SID to POSIX ID. This flag currently has no effect on .Fx . .El .It Xo .Nm .Cm groupspace .Op Fl Hinp .Op Fl o Ar field Ns Oo , Ns Ar field Oc Ns ... .Oo Fl s Ar field Oc Ns ... .Oo Fl S Ar field Oc Ns ... .Op Fl t Ar type Ns Oo , Ns Ar type Oc Ns ... .Ar filesystem Ns | Ns Ar snapshot .Xc .Pp Displays space consumed by, and quotas on, each group in the specified filesystem or snapshot. This subcommand is identical to .Qq Nm Cm userspace , except that the default types to display are .Fl t Sy posixgroup,smbgroup . .It Xo .Nm .Cm mount .Xc .Pp Displays all .Tn ZFS file systems currently mounted. .Bl -tag -width indent .It Fl f .El .It Xo .Nm .Cm mount .Op Fl vO .Op Fl o Ar property Ns Oo , Ns Ar property Oc Ns ... .Fl a | Ar filesystem .Xc .Pp Mounts .Tn ZFS file systems. .Bl -tag -width indent .It Fl v Report mount progress. .It Fl O Perform an overlay mount. Overlay mounts are not supported on .Fx . .It Fl o Ar property Ns Oo , Ns Ar property Oc Ns ... An optional, comma-separated list of mount options to use temporarily for the duration of the mount. See the .Qq Sx Temporary Mount Point Properties section for details. .It Fl a Mount all available .Tn ZFS file systems. This command may be executed on .Fx system startup by .Pa /etc/rc.d/zfs . For more information, see variable .Va zfs_enable in .Xr rc.conf 5 . .It Ar filesystem Mount the specified filesystem. .El .It Xo .Nm .Cm unmount Ns | Ns Cm umount .Op Fl f .Fl a | Ar filesystem Ns | Ns Ar mountpoint .Xc .Pp Unmounts currently mounted .Tn ZFS file systems. .Bl -tag -width indent .It Fl f Forcefully unmount the file system, even if it is currently in use. .It Fl a Unmount all available .Tn ZFS file systems. .It Ar filesystem | mountpoint Unmount the specified filesystem. The command can also be given a path to a .Tn ZFS file system mount point on the system. .El .It Xo .Nm .Cm share .Fl a | Ar filesystem .Xc .Pp Shares .Tn ZFS file systems that have the .Sy sharenfs property set. .Bl -tag -width indent .It Fl a Share all .Tn ZFS file systems that have the .Sy sharenfs property set. This command may be executed on .Fx system startup by .Pa /etc/rc.d/zfs . For more information, see variable .Va zfs_enable in .Xr rc.conf 5 . .It Ar filesystem Share the specified filesystem according to the .Tn sharenfs property. File systems are shared when the .Tn sharenfs property is set. .El .It Xo .Nm .Cm unshare .Fl a | Ar filesystem Ns | Ns Ar mountpoint .Xc .Pp Unshares .Tn ZFS file systems that have the .Tn sharenfs property set. .Bl -tag -width indent .It Fl a Unshares .Tn ZFS file systems that have the .Sy sharenfs property set. This command may be executed on .Fx system shutdown by .Pa /etc/rc.d/zfs . For more information, see variable .Va zfs_enable in .Xr rc.conf 5 . .It Ar filesystem | mountpoint Unshare the specified filesystem. The command can also be given a path to a .Tn ZFS file system shared on the system. .El .It Xo .Nm .Cm bookmark .Ar snapshot .Ar bookmark .Xc .Pp Creates a bookmark of the given snapshot. Bookmarks mark the point in time when the snapshot was created, and can be used as the incremental source for a .Qq Nm Cm send command. .Pp This feature must be enabled to be used. See .Xr zpool-features 7 for details on ZFS feature flags and the .Sy bookmark feature. .It Xo .Nm .Cm send .Op Fl DnPpRveL .Op Fl i Ar snapshot | Fl I Ar snapshot .Ar snapshot .Xc .Pp Creates a stream representation of the last .Ar snapshot argument (not part of .Fl i or .Fl I ) which is written to standard output. The output can be redirected to a file or to a different system (for example, using .Xr ssh 1 ) . By default, a full stream is generated. .Bl -tag -width indent .It Fl i Ar snapshot Generate an incremental stream from the first .Ar snapshot Pq the incremental source to the second .Ar snapshot Pq the incremental target . The incremental source can be specified as the last component of the snapshot name .Pq the Em @ No character and following and it is assumed to be from the same file system as the incremental target. .Pp If the destination is a clone, the source may be the origin snapshot, which must be fully specified (for example, .Cm pool/fs@origin , not just .Cm @origin ) . .It Fl I Ar snapshot Generate a stream package that sends all intermediary snapshots from the first .Ar snapshot to the second .Ar snapshot . For example, .Ic -I @a fs@d is similar to .Ic -i @a fs@b; -i @b fs@c; -i @c fs@d . The incremental source may be specified as with the .Fl i option. .It Fl R Generate a replication stream package, which will replicate the specified filesystem, and all descendent file systems, up to the named snapshot. When received, all properties, snapshots, descendent file systems, and clones are preserved. .Pp If the .Fl i or .Fl I flags are used in conjunction with the .Fl R flag, an incremental replication stream is generated. The current values of properties, and current snapshot and file system names are set when the stream is received. If the .Fl F flag is specified when this stream is received, snapshots and file systems that do not exist on the sending side are destroyed. .It Fl D Generate a deduplicated stream. Blocks which would have been sent multiple times in the send stream will only be sent once. The receiving system must also support this feature to receive a deduplicated stream. This flag can be used regardless of the dataset's .Sy dedup property, but performance will be much better if the filesystem uses a dedup-capable checksum (eg. .Sy sha256 ) . .It Fl L Generate a stream which may contain blocks larger than 128KB. This flag has no effect if the .Sy large_blocks pool feature is disabled, or if the .Sy recordsize property of this filesystem has never been set above 128KB. The receiving system must have the .Sy large_blocks pool feature enabled as well. See .Xr zpool-features 7 for details on ZFS feature flags and the .Sy large_blocks feature. .It Fl e Generate a more compact stream by using WRITE_EMBEDDED records for blocks which are stored more compactly on disk by the .Sy embedded_data pool feature. This flag has no effect if the .Sy embedded_data feature is disabled. The receiving system must have the .Sy embedded_data 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. See .Xr zpool-features 7 for details on ZFS feature flags and the .Sy embedded_data feature. .It Fl p Include the dataset's properties in the stream. This flag is implicit when .Fl R is specified. The receiving system must also support this feature. .It Fl n Do a dry-run ("No-op") send. Do not generate any actual send data. This is useful in conjunction with the .Fl v or .Fl P flags to determine what data will be sent. In this case, the verbose output will be written to standard output (contrast with a non-dry-run, where the stream is written to standard output and the verbose output goes to standard error). .It Fl P Print machine-parsable verbose information about the stream package generated. .It Fl v Print verbose information about the stream package generated. This information includes a per-second report of how much data has been sent. .El .Pp The format of the stream is committed. You will be able to receive your streams on future versions of .Tn ZFS . .It Xo .Nm .Cm send .Op Fl eL .Op Fl i Ar snapshot Ns | Ns Ar bookmark .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot .Xc .Pp Generate a send stream, which may be of a filesystem, and may be incremental from a bookmark. If the destination is a filesystem or volume, the pool must be read-only, or the filesystem must not be mounted. When the stream generated from a filesystem or volume is received, the default snapshot name will be .Pq --head-- . .Bl -tag -width indent .It Fl i Ar snapshot Ns | Ns bookmark Generate an incremental send stream. The incremental source must be an earlier snapshot in the destination's history. It will commonly be an earlier snapshot in the destination's filesystem, in which case it can be specified as the last component of the name .Pq the Em # No or Em @ No character and following . .Pp If the incremental target is a clone, the incremental source can be the origin snapshot, or an earlier snapshot in the origin's filesystem, or the origin's origin, etc. .It Fl L Generate a stream which may contain blocks larger than 128KB. This flag has no effect if the .Sy large_blocks pool feature is disabled, or if the .Sy recordsize property of this filesystem has never been set above 128KB. The receiving system must have the .Sy large_blocks pool feature enabled as well. See .Xr zpool-features 7 for details on ZFS feature flags and the .Sy large_blocks feature. .It Fl e Generate a more compact stream by using WRITE_EMBEDDED records for blocks which are stored more compactly on disk by the .Sy embedded_data pool feature. This flag has no effect if the .Sy embedded_data feature is disabled. The receiving system must have the .Sy embedded_data 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. See .Xr zpool-features 7 for details on ZFS feature flags and the .Sy embedded_data feature. .El .It Xo .Nm .Cm send .Op Fl Penv .Fl t .Ar receive_resume_token .Xc Creates a send stream which resumes an interrupted receive. The .Ar receive_resume_token is the value of this property on the filesystem or volume that was being received into. See the documentation for .Sy zfs receive -s for more details. .It Xo .Nm .Cm receive Ns | Ns Cm recv .Op Fl vnsFu .Op Fl o Sy origin Ns = Ns Ar snapshot .Ar filesystem Ns | Ns Ar volume Ns | Ns Ar snapshot .Xc .It Xo .Nm .Cm receive Ns | Ns Cm recv .Op Fl vnsFu .Op Fl d | e .Op Fl o Sy origin Ns = Ns Ar snapshot .Ar filesystem .Xc .Pp Creates a snapshot whose contents are as specified in the stream provided on standard input. If a full stream is received, then a new file system is created as well. Streams are created using the .Qq Nm Cm send subcommand, which by default creates a full stream. .Qq Nm Cm recv can be used as an alias for .Qq Nm Cm receive . .Pp If an incremental stream is received, then the destination file system must already exist, and its most recent snapshot must match the incremental stream's source. For .Sy zvol Ns s, the destination device link is destroyed and recreated, which means the .Sy zvol cannot be accessed during the .Sy receive operation. .Pp When a snapshot replication package stream that is generated by using the .Qq Nm Cm send Fl R command is received, any snapshots that do not exist on the sending location are destroyed by using the .Qq Nm Cm destroy Fl d command. .Pp The name of the snapshot (and file system, if a full stream is received) that this subcommand creates depends on the argument type and the .Fl d or .Fl e option. .Pp If the argument is a snapshot name, the specified .Ar snapshot is created. If the argument is a file system or volume name, a snapshot with the same name as the sent snapshot is created within the specified .Ar filesystem or .Ar volume . If the .Fl d or .Fl e option is specified, the snapshot name is determined by appending the sent snapshot's name to the specified .Ar filesystem . If the .Fl d option is specified, all but the pool name of the sent snapshot path is appended (for example, .Sy b/c@1 appended from sent snapshot .Sy a/b/c@1 ) , and if the .Fl e option is specified, only the tail of the sent snapshot path is appended (for example, .Sy c@1 appended from sent snapshot .Sy a/b/c@1 ) . In the case of .Fl d , any file systems needed to replicate the path of the sent snapshot are created within the specified file system. .Bl -tag -width indent .It Fl d Use the full sent snapshot path without the first element (without pool name) to determine the name of the new snapshot as described in the paragraph above. .It Fl e Use only the last element of the sent snapshot path to determine the name of the new snapshot as described in the paragraph above. .It Fl u File system that is associated with the received stream is not mounted. .It Fl v Print verbose information about the stream and the time required to perform the receive operation. .It Fl n Do not actually receive the stream. This can be useful in conjunction with the .Fl v option to verify the name the receive operation would use. .It Fl o Sy origin Ns = Ns Ar snapshot Forces the stream to be received as a clone of the given snapshot. -This is only valid if the stream is an incremental stream whose source -is the same as the provided origin. +If the stream is a full send stream, this will create the filesystem +described by the stream as a clone of the specified snapshot. Which +snapshot was specified will not affect the success or failure of the +receive, as long as the snapshot does exist. If the stream is an +incremental send stream, all the normal verification will be performed. .It Fl F Force a rollback of the file system to the most recent snapshot before performing the receive operation. If receiving an incremental replication stream (for example, one generated by .Qq Nm Cm send Fl R Bro Fl i | Fl I Brc ) , destroy snapshots and file systems that do not exist on the sending side. .It Fl s If the receive is interrupted, save the partially received state, rather than deleting it. Interruption may be due to premature termination of the stream .Po e.g. due to network failure or failure of the remote system if the stream is being read over a network connection .Pc , a checksum error in the stream, termination of the .Nm zfs Cm receive process, or unclean shutdown of the system. .Pp The receive can be resumed with a stream generated by .Nm zfs Cm send Fl t Ar token , where the .Ar token is the value of the .Sy receive_resume_token property of the filesystem or volume which is received into. .Pp To use this flag, the storage pool must have the .Sy extensible_dataset feature enabled. See .Xr zpool-features 5 for details on ZFS feature flags. .El .It Xo .Nm .Cm receive Ns | Ns Cm recv .Fl A .Ar filesystem Ns | Ns Ar volume .Xc Abort an interrupted .Nm zfs Cm receive Fl s , deleting its saved partially received state. .It Xo .Nm .Cm allow .Ar filesystem Ns | Ns Ar volume .Xc .Pp Displays permissions that have been delegated on the specified filesystem or volume. See the other forms of .Qq Nm Cm allow for more information. .It Xo .Nm .Cm allow .Op Fl ldug .Ar user Ns | Ns Ar group Ns Oo Ns , Ns Ar user Ns | Ns Ar group Oc Ns ... .Ar perm Ns | Ns Ar @setname Ns .Oo Ns , Ns Ar perm Ns | Ns Ar @setname Oc Ns ... .Ar filesystem Ns | Ns Ar volume .Xc .It Xo .Nm .Cm allow .Op Fl ld .Fl e Ns | Ns Cm everyone .Ar perm Ns | Ns Ar @setname Ns Op Ns , Ns Ar perm Ns | Ns Ar @setname Ns .Ns ... .Ar filesystem Ns | Ns Ar volume .Xc .Pp Delegates .Tn ZFS administration permission for the file systems to non-privileged users. .Bl -tag -width indent .It Xo .Op Fl ug .Ar user Ns | Ns Ar group Ns Oo , Ar user Ns | Ns Ar group Oc Ns ... .Xc Specifies to whom the permissions are delegated. Multiple entities can be specified as a comma-separated list. If neither of the .Fl ug options are specified, then the argument is interpreted preferentially as the keyword .Cm everyone , then as a user name, and lastly as a group name. To specify a user or group named .Qq everyone , use the .Fl u or .Fl g options. To specify a group with the same name as a user, use the .Fl g option. .It Op Fl e Ns | Ns Cm everyone Specifies that the permissions be delegated to .Qq everyone . .It Xo .Ar perm Ns | Ns Ar @setname Ns Oo , Ns Ar perm Ns | Ns Ar @setname Oc Ns ... .Xc The permissions to delegate. Multiple permissions may be specified as a comma-separated list. Permission names are the same as .Tn ZFS subcommand and property names. See the property list below. Property set names, which begin with an at sign .Pq Sy @ , may be specified. See the .Fl s form below for details. .It Xo .Op Fl ld .Ar filesystem Ns | Ns Ar volume .Xc Specifies where the permissions are delegated. If neither of the .Fl ld options are specified, or both are, then the permissions are allowed for the file system or volume, and all of its descendents. If only the .Fl l option is used, then is allowed "locally" only for the specified file system. If only the .Fl d option is used, then is allowed only for the descendent file systems. .El .Pp Permissions are generally the ability to use a .Tn ZFS subcommand or change a .Tn ZFS property. The following permissions are available: .Bl -column -offset 4n "secondarycache" "subcommand" .It NAME Ta TYPE Ta NOTES .It allow Ta subcommand Ta Must Xo also have the permission that is being allowed .Xc .It clone Ta subcommand Ta Must Xo also have the 'create' ability and 'mount' ability in the origin file system .Xc .It create Ta subcommand Ta Must also have the 'mount' ability .It destroy Ta subcommand Ta Must also have the 'mount' ability .It diff Ta subcommand Ta Allows lookup of paths within a dataset given an object number, and the ability to create snapshots necessary to 'zfs diff' .It hold Ta subcommand Ta Allows adding a user hold to a snapshot .It mount Ta subcommand Ta Allows mount/umount of Tn ZFS No datasets .It promote Ta subcommand Ta Must Xo also have the 'mount' and 'promote' ability in the origin file system .Xc .It receive Ta subcommand Ta Must also have the 'mount' and 'create' ability .It release Ta subcommand Ta Allows Xo releasing a user hold which might destroy the snapshot .Xc .It rename Ta subcommand Ta Must Xo also have the 'mount' and 'create' ability in the new parent .Xc .It rollback Ta subcommand Ta Must also have the 'mount' ability .It send Ta subcommand .It share Ta subcommand Ta Allows Xo sharing file systems over the .Tn NFS protocol .Xc .It snapshot Ta subcommand Ta Must also have the 'mount' ability .It groupquota Ta other Ta Allows accessing any groupquota@... property .It groupused Ta other Ta Allows reading any groupused@... property .It userprop Ta other Ta Allows changing any user property .It userquota Ta other Ta Allows accessing any userquota@... property .It userused Ta other Ta Allows reading any userused@... property .It aclinherit Ta property .It aclmode Ta property .It atime Ta property .It canmount Ta property .It casesensitivity Ta property .It checksum Ta property .It compression Ta property .It copies Ta property .It dedup Ta property .It devices Ta property .It exec Ta property .It filesystem_limit Ta property .It logbias Ta property .It jailed Ta property .It mlslabel Ta property .It mountpoint Ta property .It nbmand Ta property .It normalization Ta property .It primarycache Ta property .It quota Ta property .It readonly Ta property .It recordsize Ta property .It refquota Ta property .It refreservation Ta property .It reservation Ta property .It secondarycache Ta property .It setuid Ta property .It sharenfs Ta property .It sharesmb Ta property .It snapdir Ta property .It snapshot_limit Ta property .It sync Ta property .It utf8only Ta property .It version Ta property .It volblocksize Ta property .It volsize Ta property .It vscan Ta property .It xattr Ta property .El .It Xo .Nm .Cm allow .Fl c .Ar perm Ns | Ns Ar @setname Ns Op Ns , Ns Ar perm Ns | Ns Ar @setname Ns .Ns ... .Ar filesystem Ns | Ns Ar volume .Xc .Pp Sets "create time" permissions. These permissions are granted (locally) to the creator of any newly-created descendent file system. .It Xo .Nm .Cm allow .Fl s .Ar @setname .Ar perm Ns | Ns Ar @setname Ns Op Ns , Ns Ar perm Ns | Ns Ar @setname Ns .Ns ... .Ar filesystem Ns | Ns Ar volume .Xc .Pp Defines or adds permissions to a permission set. The set can be used by other .Qq Nm Cm allow commands for the specified file system and its descendents. Sets are evaluated dynamically, so changes to a set are immediately reflected. Permission sets follow the same naming restrictions as ZFS file systems, but the name must begin with an "at sign" .Pq Sy @ , and can be no more than 64 characters long. .It Xo .Nm .Cm unallow .Op Fl rldug .Ar user Ns | Ns Ar group Ns Oo Ns , Ns Ar user Ns | Ns Ar group Oc Ns ... .Oo Ar perm Ns | Ns Ar @setname Ns Op , Ns Ar perm Ns | Ns Ar @setname Ns .Ns ... Oc .Ar filesystem Ns | Ns Ar volume .Xc .It Xo .Nm .Cm unallow .Op Fl rld .Fl e Ns | Ns Cm everyone .Oo Ar perm Ns | Ns Ar @setname Ns Op , Ns Ar perm Ns | Ns Ar @setname Ns .Ns ... Oc .Ar filesystem Ns | Ns Ar volume .Xc .It Xo .Nm .Cm unallow .Op Fl r .Fl c .Oo Ar perm Ns | Ns Ar @setname Ns Op , Ns Ar perm Ns | Ns Ar @setname Ns .Ns ... Oc .Ar filesystem Ns | Ns Ar volume .Xc .Pp Removes permissions that were granted with the .Qq Nm Cm allow command. No permissions are explicitly denied, so other permissions granted are still in effect. For example, if the permission is granted by an ancestor. If no permissions are specified, then all permissions for the specified .Ar user , group , No or everyone are removed. Specifying .Cm everyone .Po or using the Fl e option .Pc only removes the permissions that were granted to everyone , not all permissions for every user and group. See the .Qq Nm Cm allow command for a description of the .Fl ldugec options. .Bl -tag -width indent .It Fl r Recursively remove the permissions from this file system and all descendents. .El .It Xo .Nm .Cm unallow .Op Fl r .Fl s .Ar @setname .Oo Ar perm Ns | Ns Ar @setname Ns Op , Ns Ar perm Ns | Ns Ar @setname Ns .Ns ... Oc .Ar filesystem Ns | Ns Ar volume .Xc .Pp Removes permissions from a permission set. If no permissions are specified, then all permissions are removed, thus removing the set entirely. .It Xo .Nm .Cm hold .Op Fl r .Ar tag snapshot Ns ... .Xc .Pp Adds a single reference, named with the .Ar tag argument, to the specified snapshot or snapshots. Each snapshot has its own tag namespace, and tags must be unique within that space. .Pp If a hold exists on a snapshot, attempts to destroy that snapshot by using the .Qq Nm Cm destroy command returns .Em EBUSY . .Bl -tag -width indent .It Fl r Specifies that a hold with the given tag is applied recursively to the snapshots of all descendent file systems. .El .It Xo .Nm .Cm holds .Op Fl r .Ar snapshot Ns ... .Xc .Pp Lists all existing user references for the given snapshot or snapshots. .Bl -tag -width indent .It Fl r Lists the holds that are set on the named descendent snapshots, in addition to listing the holds on the named snapshot. .El .It Xo .Nm .Cm release .Op Fl r .Ar tag snapshot Ns ... .Xc .Pp Removes a single reference, named with the .Ar tag argument, from the specified snapshot or snapshots. The tag must already exist for each snapshot. .Bl -tag -width indent .It Fl r Recursively releases a hold with the given tag on the snapshots of all descendent file systems. .El .It Xo .Nm .Cm diff .Op Fl FHt .Ar snapshot .Op Ar snapshot Ns | Ns Ar filesystem .Xc .Pp Display the difference between a snapshot of a given filesystem and another snapshot of that filesystem from a later time or the current contents of the filesystem. The first column is a character indicating the type of change, the other columns indicate pathname, new pathname .Pq in case of rename , change in link count, and optionally file type and/or change time. .Pp The types of change are: .Bl -column -offset 2n indent .It \&- Ta path was removed .It \&+ Ta path was added .It \&M Ta path was modified .It \&R Ta path was renamed .El .Bl -tag -width indent .It Fl F Display an indication of the type of file, in a manner similar to the .Fl F option of .Xr ls 1 . .Bl -column -offset 2n indent .It \&B Ta block device .It \&C Ta character device .It \&F Ta regular file .It \&/ Ta directory .It \&@ Ta symbolic link .It \&= Ta socket .It \&> Ta door (not supported on Fx ) .It \&| Ta named pipe (not supported on Fx ) .It \&P Ta event port (not supported on Fx ) .El .It Fl H Give more parsable tab-separated output, without header lines and without arrows. .It Fl t Display the path's inode change time as the first column of output. .El .It Xo .Nm .Cm jail .Ar jailid filesystem .Xc .Pp Attaches the specified .Ar filesystem to the jail identified by JID .Ar jailid . From now on this file system tree can be managed from within a jail if the .Sy jailed property has been set. To use this functionality, the jail needs the .Va allow.mount and .Va allow.mount.zfs parameters set to 1 and the .Va enforce_statfs parameter set to a value lower than 2. .Pp See .Xr jail 8 for more information on managing jails and configuring the parameters above. .It Xo .Nm .Cm unjail .Ar jailid filesystem .Xc .Pp Detaches the specified .Ar filesystem from the jail identified by JID .Ar jailid . .El .Sh EXIT STATUS The following exit values are returned: .Bl -tag -offset 2n -width 2n .It 0 Successful completion. .It 1 An error occurred. .It 2 Invalid command line options were specified. .El .Sh EXAMPLES .Bl -tag -width 0n .It Sy Example 1 No Creating a Tn ZFS No File System Hierarchy .Pp The following commands create a file system named .Em pool/home and a file system named .Em pool/home/bob . The mount point .Pa /home is set for the parent file system, and is automatically inherited by the child file system. .Bd -literal -offset 2n .Li # Ic zfs create pool/home .Li # Ic zfs set mountpoint=/home pool/home .Li # Ic zfs create pool/home/bob .Ed .It Sy Example 2 No Creating a Tn ZFS No Snapshot .Pp The following command creates a snapshot named .Sy yesterday . This snapshot is mounted on demand in the .Pa \&.zfs/snapshot directory at the root of the .Em pool/home/bob file system. .Bd -literal -offset 2n .Li # Ic zfs snapshot pool/home/bob@yesterday .Ed .It Sy Example 3 No Creating and Destroying Multiple Snapshots .Pp The following command creates snapshots named .Em yesterday of .Em pool/home and all of its descendent file systems. Each snapshot is mounted on demand in the .Pa \&.zfs/snapshot directory at the root of its file system. The second command destroys the newly created snapshots. .Bd -literal -offset 2n .Li # Ic zfs snapshot -r pool/home@yesterday .Li # Ic zfs destroy -r pool/home@yesterday .Ed .It Sy Example 4 No Disabling and Enabling File System Compression .Pp The following command disables the .Sy compression property for all file systems under .Em pool/home . The next command explicitly enables .Sy compression for .Em pool/home/anne . .Bd -literal -offset 2n .Li # Ic zfs set compression=off pool/home .Li # Ic zfs set compression=on pool/home/anne .Ed .It Sy Example 5 No Listing Tn ZFS No Datasets .Pp The following command lists all active file systems and volumes in the system. Snapshots are displayed if the .Sy listsnaps property is .Cm on . The default is .Cm off . See .Xr zpool 8 for more information on pool properties. .Bd -literal -offset 2n .Li # Ic zfs list NAME USED AVAIL REFER MOUNTPOINT pool 450K 457G 18K /pool pool/home 315K 457G 21K /home pool/home/anne 18K 457G 18K /home/anne pool/home/bob 276K 457G 276K /home/bob .Ed .It Sy Example 6 No Setting a Quota on a Tn ZFS No File System .Pp The following command sets a quota of 50 Gbytes for .Em pool/home/bob . .Bd -literal -offset 2n .Li # Ic zfs set quota=50G pool/home/bob .Ed .It Sy Example 7 No Listing Tn ZFS No Properties .Pp The following command lists all properties for .Em pool/home/bob . .Bd -literal -offset 2n .Li # Ic zfs get all pool/home/bob NAME PROPERTY VALUE SOURCE pool/home/bob type filesystem - pool/home/bob creation Tue Jul 21 15:53 2009 - pool/home/bob used 21K - pool/home/bob available 20.0G - pool/home/bob referenced 21K - pool/home/bob compressratio 1.00x - pool/home/bob mounted yes - pool/home/bob quota 20G local pool/home/bob reservation none default pool/home/bob recordsize 128K default pool/home/bob mountpoint /home/bob default pool/home/bob sharenfs off default pool/home/bob checksum on default pool/home/bob compression on local pool/home/bob atime on default pool/home/bob devices on default pool/home/bob exec on default pool/home/bob filesystem_limit none default pool/home/bob setuid on default pool/home/bob readonly off default pool/home/bob jailed off default pool/home/bob snapdir hidden default pool/home/bob snapshot_limit none default pool/home/bob aclmode discard default pool/home/bob aclinherit restricted default pool/home/bob canmount on default pool/home/bob xattr on default pool/home/bob copies 1 default pool/home/bob version 5 - pool/home/bob utf8only off - pool/home/bob normalization none - pool/home/bob casesensitivity sensitive - pool/home/bob vscan off default pool/home/bob nbmand off default pool/home/bob sharesmb off default pool/home/bob refquota none default pool/home/bob refreservation none default pool/home/bob primarycache all default pool/home/bob secondarycache all default pool/home/bob usedbysnapshots 0 - pool/home/bob usedbydataset 21K - pool/home/bob usedbychildren 0 - pool/home/bob usedbyrefreservation 0 - pool/home/bob logbias latency default pool/home/bob dedup off default pool/home/bob mlslabel - pool/home/bob sync standard default pool/home/bob refcompressratio 1.00x - .Ed .Pp The following command gets a single property value. .Bd -literal -offset 2n .Li # Ic zfs get -H -o value compression pool/home/bob on .Ed .Pp The following command lists all properties with local settings for .Em pool/home/bob . .Bd -literal -offset 2n .Li # Ic zfs get -s local -o name,property,value all pool/home/bob NAME PROPERTY VALUE pool/home/bob quota 20G pool/home/bob compression on .Ed .It Sy Example 8 No Rolling Back a Tn ZFS No File System .Pp The following command reverts the contents of .Em pool/home/anne to the snapshot named .Em yesterday , deleting all intermediate snapshots. .Bd -literal -offset 2n .Li # Ic zfs rollback -r pool/home/anne@yesterday .Ed .It Sy Example 9 No Creating a Tn ZFS No Clone .Pp The following command creates a writable file system whose initial contents are the same as .Em pool/home/bob@yesterday . .Bd -literal -offset 2n .Li # Ic zfs clone pool/home/bob@yesterday pool/clone .Ed .It Sy Example 10 No Promoting a Tn ZFS No Clone .Pp The following commands illustrate how to test out changes to a file system, and then replace the original file system with the changed one, using clones, clone promotion, and renaming: .Bd -literal -offset 2n .Li # Ic zfs create pool/project/production .Ed .Pp Populate .Pa /pool/project/production with data and continue with the following commands: .Bd -literal -offset 2n .Li # Ic zfs snapshot pool/project/production@today .Li # Ic zfs clone pool/project/production@today pool/project/beta .Ed .Pp Now make changes to .Pa /pool/project/beta and continue with the following commands: .Bd -literal -offset 2n .Li # Ic zfs promote pool/project/beta .Li # Ic zfs rename pool/project/production pool/project/legacy .Li # Ic zfs rename pool/project/beta pool/project/production .Ed .Pp Once the legacy version is no longer needed, it can be destroyed. .Bd -literal -offset 2n .Li # Ic zfs destroy pool/project/legacy .Ed .It Sy Example 11 No Inheriting Tn ZFS No Properties .Pp The following command causes .Em pool/home/bob and .Em pool/home/anne to inherit the .Sy checksum property from their parent. .Bd -literal -offset 2n .Li # Ic zfs inherit checksum pool/home/bob pool/home/anne .Ed .It Sy Example 12 No Remotely Replicating Tn ZFS No Data .Pp The following commands send a full stream and then an incremental stream to a remote machine, restoring them into .Sy poolB/received/fs@a and .Sy poolB/received/fs@b , respectively. .Sy poolB must contain the file system .Sy poolB/received , and must not initially contain .Sy poolB/received/fs . .Bd -literal -offset 2n .Li # Ic zfs send pool/fs@a | ssh host zfs receive poolB/received/fs@a .Li # Ic zfs send -i a pool/fs@b | ssh host zfs receive poolB/received/fs .Ed .It Xo .Sy Example 13 Using the .Qq zfs receive -d Option .Xc .Pp The following command sends a full stream of .Sy poolA/fsA/fsB@snap to a remote machine, receiving it into .Sy poolB/received/fsA/fsB@snap . The .Sy fsA/fsB@snap portion of the received snapshot's name is determined from the name of the sent snapshot. .Sy poolB must contain the file system .Sy poolB/received . If .Sy poolB/received/fsA does not exist, it is created as an empty file system. .Bd -literal -offset 2n .Li # Ic zfs send poolA/fsA/fsB@snap | ssh host zfs receive -d poolB/received .Ed .It Sy Example 14 No Setting User Properties .Pp The following example sets the user-defined .Sy com.example:department property for a dataset. .Bd -literal -offset 2n .Li # Ic zfs set com.example:department=12345 tank/accounting .Ed .It Sy Example 15 No Performing a Rolling Snapshot .Pp The following example shows how to maintain a history of snapshots with a consistent naming scheme. To keep a week's worth of snapshots, the user destroys the oldest snapshot, renames the remaining snapshots, and then creates a new snapshot, as follows: .Bd -literal -offset 2n .Li # Ic zfs destroy -r pool/users@7daysago .Li # Ic zfs rename -r pool/users@6daysago @7daysago .Li # Ic zfs rename -r pool/users@5daysago @6daysago .Li # Ic zfs rename -r pool/users@4daysago @5daysago .Li # Ic zfs rename -r pool/users@3daysago @4daysago .Li # Ic zfs rename -r pool/users@2daysago @3daysago .Li # Ic zfs rename -r pool/users@yesterday @2daysago .Li # Ic zfs rename -r pool/users@today @yesterday .Li # Ic zfs snapshot -r pool/users@today .Ed .It Xo .Sy Example 16 Setting .Qq sharenfs Property Options on a ZFS File System .Xc .Pp The following command shows how to set .Sy sharenfs property options to enable root access for a specific network on the .Em tank/home file system. The contents of the .Sy sharenfs property are valid .Xr exports 5 options. .Bd -literal -offset 2n .Li # Ic zfs set sharenfs="maproot=root,network 192.168.0.0/24" tank/home .Ed .Pp Another way to write this command with the same result is: .Bd -literal -offset 2n .Li # Ic set zfs sharenfs="-maproot=root -network 192.168.0.0/24" tank/home .Ed .It Xo .Sy Example 17 Delegating .Tn ZFS Administration Permissions on a .Tn ZFS Dataset .Xc .Pp The following example shows how to set permissions so that user .Em cindys can create, destroy, mount, and take snapshots on .Em tank/cindys . The permissions on .Em tank/cindys are also displayed. .Bd -literal -offset 2n .Li # Ic zfs allow cindys create,destroy,mount,snapshot tank/cindys .Li # Ic zfs allow tank/cindys ---- Permissions on tank/cindys -------------------------------------- Local+Descendent permissions: user cindys create,destroy,mount,snapshot .Ed .It Sy Example 18 No Delegating Create Time Permissions on a Tn ZFS No Dataset .Pp The following example shows how to grant anyone in the group .Em staff to create file systems in .Em tank/users . This syntax also allows staff members to destroy their own file systems, but not destroy anyone else's file system. The permissions on .Em tank/users are also displayed. .Bd -literal -offset 2n .Li # Ic zfs allow staff create,mount tank/users .Li # Ic zfs allow -c destroy tank/users .Li # Ic zfs allow tank/users ---- Permissions on tank/users --------------------------------------- Permission sets: destroy Local+Descendent permissions: group staff create,mount .Ed .It Xo .Sy Example 19 Defining and Granting a Permission Set on a .Tn ZFS Dataset .Xc .Pp The following example shows how to define and grant a permission set on the .Em tank/users file system. The permissions on .Em tank/users are also displayed. .Bd -literal -offset 2n .Li # Ic zfs allow -s @pset create,destroy,snapshot,mount tank/users .Li # Ic zfs allow staff @pset tank/users .Li # Ic zfs allow tank/users ---- Permissions on tank/users --------------------------------------- Permission sets: @pset create,destroy,mount,snapshot Local+Descendent permissions: group staff @pset .Ed .It Sy Example 20 No Delegating Property Permissions on a Tn ZFS No Dataset .Pp The following example shows to grant the ability to set quotas and reservations on the .Sy users/home file system. The permissions on .Sy users/home are also displayed. .Bd -literal -offset 2n .Li # Ic zfs allow cindys quota,reservation users/home .Li # Ic zfs allow users/home ---- Permissions on users/home --------------------------------------- Local+Descendent permissions: user cindys quota,reservation .Li # Ic su - cindys .Li cindys% Ic zfs set quota=10G users/home/marks .Li cindys% Ic zfs get quota users/home/marks NAME PROPERTY VALUE SOURCE users/home/marks quota 10G local .Ed .It Sy Example 21 No Removing ZFS Delegated Permissions on a Tn ZFS No Dataset .Pp The following example shows how to remove the snapshot permission from the .Em staff group on the .Em tank/users file system. The permissions on .Em tank/users are also displayed. .Bd -literal -offset 2n .Li # Ic zfs unallow staff snapshot tank/users .Li # Ic zfs allow tank/users ---- Permissions on tank/users --------------------------------------- Permission sets: @pset create,destroy,mount,snapshot Local+Descendent permissions: group staff @pset .Ed .It Sy Example 22 Showing the differences between a snapshot and a ZFS Dataset .Pp The following example shows how to see what has changed between a prior snapshot of a ZFS Dataset and its current state. The .Fl F option is used to indicate type information for the files affected. .Bd -literal -offset 2n .Li # Ic zfs diff tank/test@before tank/test M / /tank/test/ M F /tank/test/linked (+1) R F /tank/test/oldname -> /tank/test/newname - F /tank/test/deleted + F /tank/test/created M F /tank/test/modified .Ed .El .Sh SEE ALSO .Xr chmod 2 , .Xr fsync 2 , .Xr exports 5 , .Xr fstab 5 , .Xr rc.conf 5 , .Xr jail 8 , .Xr mount 8 , .Xr umount 8 , .Xr zpool 8 .Sh AUTHORS This manual page is a .Xr mdoc 7 reimplementation of the .Tn OpenSolaris manual page .Em zfs(1M) , modified and customized for .Fx and licensed under the Common Development and Distribution License .Pq Tn CDDL . .Pp The .Xr mdoc 7 implementation of this manual page was initially written by .An Martin Matuska Aq mm@FreeBSD.org . Index: stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_send.c =================================================================== --- stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_send.c (revision 297101) +++ stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_send.c (revision 297102) @@ -1,3167 +1,3219 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2011, 2015 by Delphix. All rights reserved. * Copyright (c) 2014, Joyent, Inc. All rights reserved. * Copyright (c) 2012, Martin Matuska . All rights reserved. * Copyright 2014 HybridCluster. All rights reserved. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef __FreeBSD__ #undef dump_write #define dump_write dmu_dump_write #endif /* Set this tunable to TRUE to replace corrupt data with 0x2f5baddb10c */ int zfs_send_corrupt_data = B_FALSE; int zfs_send_queue_length = 16 * 1024 * 1024; int zfs_recv_queue_length = 16 * 1024 * 1024; static char *dmu_recv_tag = "dmu_recv_tag"; const char *recv_clone_name = "%recv"; #define BP_SPAN(datablkszsec, indblkshift, level) \ (((uint64_t)datablkszsec) << (SPA_MINBLOCKSHIFT + \ (level) * (indblkshift - SPA_BLKPTRSHIFT))) static void byteswap_record(dmu_replay_record_t *drr); struct send_thread_arg { bqueue_t q; dsl_dataset_t *ds; /* Dataset to traverse */ uint64_t fromtxg; /* Traverse from this txg */ int flags; /* flags to pass to traverse_dataset */ int error_code; boolean_t cancel; zbookmark_phys_t resume; }; struct send_block_record { boolean_t eos_marker; /* Marks the end of the stream */ blkptr_t bp; zbookmark_phys_t zb; uint8_t indblkshift; uint16_t datablkszsec; bqueue_node_t ln; }; static int dump_bytes(dmu_sendarg_t *dsp, void *buf, int len) { dsl_dataset_t *ds = dmu_objset_ds(dsp->dsa_os); struct uio auio; struct iovec aiov; ASSERT0(len % 8); aiov.iov_base = buf; aiov.iov_len = len; auio.uio_iov = &aiov; auio.uio_iovcnt = 1; auio.uio_resid = len; auio.uio_segflg = UIO_SYSSPACE; auio.uio_rw = UIO_WRITE; auio.uio_offset = (off_t)-1; auio.uio_td = dsp->dsa_td; #ifdef _KERNEL if (dsp->dsa_fp->f_type == DTYPE_VNODE) bwillwrite(); dsp->dsa_err = fo_write(dsp->dsa_fp, &auio, dsp->dsa_td->td_ucred, 0, dsp->dsa_td); #else fprintf(stderr, "%s: returning EOPNOTSUPP\n", __func__); dsp->dsa_err = EOPNOTSUPP; #endif mutex_enter(&ds->ds_sendstream_lock); *dsp->dsa_off += len; mutex_exit(&ds->ds_sendstream_lock); return (dsp->dsa_err); } /* * For all record types except BEGIN, fill in the checksum (overlaid in * drr_u.drr_checksum.drr_checksum). The checksum verifies everything * up to the start of the checksum itself. */ static int dump_record(dmu_sendarg_t *dsp, void *payload, int payload_len) { ASSERT3U(offsetof(dmu_replay_record_t, drr_u.drr_checksum.drr_checksum), ==, sizeof (dmu_replay_record_t) - sizeof (zio_cksum_t)); fletcher_4_incremental_native(dsp->dsa_drr, offsetof(dmu_replay_record_t, drr_u.drr_checksum.drr_checksum), &dsp->dsa_zc); if (dsp->dsa_drr->drr_type != DRR_BEGIN) { ASSERT(ZIO_CHECKSUM_IS_ZERO(&dsp->dsa_drr->drr_u. drr_checksum.drr_checksum)); dsp->dsa_drr->drr_u.drr_checksum.drr_checksum = dsp->dsa_zc; } fletcher_4_incremental_native(&dsp->dsa_drr-> drr_u.drr_checksum.drr_checksum, sizeof (zio_cksum_t), &dsp->dsa_zc); if (dump_bytes(dsp, dsp->dsa_drr, sizeof (dmu_replay_record_t)) != 0) return (SET_ERROR(EINTR)); if (payload_len != 0) { fletcher_4_incremental_native(payload, payload_len, &dsp->dsa_zc); if (dump_bytes(dsp, payload, payload_len) != 0) return (SET_ERROR(EINTR)); } return (0); } +/* + * Fill in the drr_free struct, or perform aggregation if the previous record is + * also a free record, and the two are adjacent. + * + * Note that we send free records even for a full send, because we want to be + * able to receive a full send as a clone, which requires a list of all the free + * and freeobject records that were generated on the source. + */ static int dump_free(dmu_sendarg_t *dsp, uint64_t object, uint64_t offset, uint64_t length) { struct drr_free *drrf = &(dsp->dsa_drr->drr_u.drr_free); /* * When we receive a free record, dbuf_free_range() assumes * that the receiving system doesn't have any dbufs in the range * being freed. This is always true because there is a one-record * constraint: we only send one WRITE record for any given * object,offset. We know that the one-record constraint is * true because we always send data in increasing order by * object,offset. * * If the increasing-order constraint ever changes, we should find * another way to assert that the one-record constraint is still * satisfied. */ ASSERT(object > dsp->dsa_last_data_object || (object == dsp->dsa_last_data_object && offset > dsp->dsa_last_data_offset)); - /* - * If we are doing a non-incremental send, then there can't - * be any data in the dataset we're receiving into. Therefore - * a free record would simply be a no-op. Save space by not - * sending it to begin with. - */ - if (!dsp->dsa_incremental) - return (0); - if (length != -1ULL && offset + length < offset) length = -1ULL; /* * If there is a pending op, but it's not PENDING_FREE, push it out, * since free block aggregation can only be done for blocks of the * same type (i.e., DRR_FREE records can only be aggregated with * other DRR_FREE records. DRR_FREEOBJECTS records can only be * aggregated with other DRR_FREEOBJECTS records. */ if (dsp->dsa_pending_op != PENDING_NONE && dsp->dsa_pending_op != PENDING_FREE) { if (dump_record(dsp, NULL, 0) != 0) return (SET_ERROR(EINTR)); dsp->dsa_pending_op = PENDING_NONE; } if (dsp->dsa_pending_op == PENDING_FREE) { /* * There should never be a PENDING_FREE if length is -1 * (because dump_dnode is the only place where this * function is called with a -1, and only after flushing * any pending record). */ ASSERT(length != -1ULL); /* * Check to see whether this free block can be aggregated * with pending one. */ if (drrf->drr_object == object && drrf->drr_offset + drrf->drr_length == offset) { drrf->drr_length += length; return (0); } else { /* not a continuation. Push out pending record */ if (dump_record(dsp, NULL, 0) != 0) return (SET_ERROR(EINTR)); dsp->dsa_pending_op = PENDING_NONE; } } /* create a FREE record and make it pending */ bzero(dsp->dsa_drr, sizeof (dmu_replay_record_t)); dsp->dsa_drr->drr_type = DRR_FREE; drrf->drr_object = object; drrf->drr_offset = offset; drrf->drr_length = length; drrf->drr_toguid = dsp->dsa_toguid; if (length == -1ULL) { if (dump_record(dsp, NULL, 0) != 0) return (SET_ERROR(EINTR)); } else { dsp->dsa_pending_op = PENDING_FREE; } return (0); } static int dump_write(dmu_sendarg_t *dsp, dmu_object_type_t type, uint64_t object, uint64_t offset, int blksz, const blkptr_t *bp, void *data) { struct drr_write *drrw = &(dsp->dsa_drr->drr_u.drr_write); /* * We send data in increasing object, offset order. * See comment in dump_free() for details. */ ASSERT(object > dsp->dsa_last_data_object || (object == dsp->dsa_last_data_object && offset > dsp->dsa_last_data_offset)); dsp->dsa_last_data_object = object; dsp->dsa_last_data_offset = offset + blksz - 1; /* * If there is any kind of pending aggregation (currently either * a grouping of free objects or free blocks), push it out to * the stream, since aggregation can't be done across operations * of different types. */ if (dsp->dsa_pending_op != PENDING_NONE) { if (dump_record(dsp, NULL, 0) != 0) return (SET_ERROR(EINTR)); dsp->dsa_pending_op = PENDING_NONE; } /* write a WRITE record */ bzero(dsp->dsa_drr, sizeof (dmu_replay_record_t)); dsp->dsa_drr->drr_type = DRR_WRITE; drrw->drr_object = object; drrw->drr_type = type; drrw->drr_offset = offset; drrw->drr_length = blksz; drrw->drr_toguid = dsp->dsa_toguid; if (bp == NULL || BP_IS_EMBEDDED(bp)) { /* * There's no pre-computed checksum for partial-block * writes or embedded BP's, so (like * fletcher4-checkummed blocks) userland will have to * compute a dedup-capable checksum itself. */ drrw->drr_checksumtype = ZIO_CHECKSUM_OFF; } else { drrw->drr_checksumtype = BP_GET_CHECKSUM(bp); if (zio_checksum_table[drrw->drr_checksumtype].ci_flags & ZCHECKSUM_FLAG_DEDUP) drrw->drr_checksumflags |= DRR_CHECKSUM_DEDUP; DDK_SET_LSIZE(&drrw->drr_key, BP_GET_LSIZE(bp)); DDK_SET_PSIZE(&drrw->drr_key, BP_GET_PSIZE(bp)); DDK_SET_COMPRESS(&drrw->drr_key, BP_GET_COMPRESS(bp)); drrw->drr_key.ddk_cksum = bp->blk_cksum; } if (dump_record(dsp, data, blksz) != 0) return (SET_ERROR(EINTR)); return (0); } static int dump_write_embedded(dmu_sendarg_t *dsp, uint64_t object, uint64_t offset, int blksz, const blkptr_t *bp) { char buf[BPE_PAYLOAD_SIZE]; struct drr_write_embedded *drrw = &(dsp->dsa_drr->drr_u.drr_write_embedded); if (dsp->dsa_pending_op != PENDING_NONE) { if (dump_record(dsp, NULL, 0) != 0) return (EINTR); dsp->dsa_pending_op = PENDING_NONE; } ASSERT(BP_IS_EMBEDDED(bp)); bzero(dsp->dsa_drr, sizeof (dmu_replay_record_t)); dsp->dsa_drr->drr_type = DRR_WRITE_EMBEDDED; drrw->drr_object = object; drrw->drr_offset = offset; drrw->drr_length = blksz; drrw->drr_toguid = dsp->dsa_toguid; drrw->drr_compression = BP_GET_COMPRESS(bp); drrw->drr_etype = BPE_GET_ETYPE(bp); drrw->drr_lsize = BPE_GET_LSIZE(bp); drrw->drr_psize = BPE_GET_PSIZE(bp); decode_embedded_bp_compressed(bp, buf); if (dump_record(dsp, buf, P2ROUNDUP(drrw->drr_psize, 8)) != 0) return (EINTR); return (0); } static int dump_spill(dmu_sendarg_t *dsp, uint64_t object, int blksz, void *data) { struct drr_spill *drrs = &(dsp->dsa_drr->drr_u.drr_spill); if (dsp->dsa_pending_op != PENDING_NONE) { if (dump_record(dsp, NULL, 0) != 0) return (SET_ERROR(EINTR)); dsp->dsa_pending_op = PENDING_NONE; } /* write a SPILL record */ bzero(dsp->dsa_drr, sizeof (dmu_replay_record_t)); dsp->dsa_drr->drr_type = DRR_SPILL; drrs->drr_object = object; drrs->drr_length = blksz; drrs->drr_toguid = dsp->dsa_toguid; if (dump_record(dsp, data, blksz) != 0) return (SET_ERROR(EINTR)); return (0); } static int dump_freeobjects(dmu_sendarg_t *dsp, uint64_t firstobj, uint64_t numobjs) { struct drr_freeobjects *drrfo = &(dsp->dsa_drr->drr_u.drr_freeobjects); - /* See comment in dump_free(). */ - if (!dsp->dsa_incremental) - return (0); - /* * If there is a pending op, but it's not PENDING_FREEOBJECTS, * push it out, since free block aggregation can only be done for * blocks of the same type (i.e., DRR_FREE records can only be * aggregated with other DRR_FREE records. DRR_FREEOBJECTS records * can only be aggregated with other DRR_FREEOBJECTS records. */ if (dsp->dsa_pending_op != PENDING_NONE && dsp->dsa_pending_op != PENDING_FREEOBJECTS) { if (dump_record(dsp, NULL, 0) != 0) return (SET_ERROR(EINTR)); dsp->dsa_pending_op = PENDING_NONE; } if (dsp->dsa_pending_op == PENDING_FREEOBJECTS) { /* * See whether this free object array can be aggregated * with pending one */ if (drrfo->drr_firstobj + drrfo->drr_numobjs == firstobj) { drrfo->drr_numobjs += numobjs; return (0); } else { /* can't be aggregated. Push out pending record */ if (dump_record(dsp, NULL, 0) != 0) return (SET_ERROR(EINTR)); dsp->dsa_pending_op = PENDING_NONE; } } /* write a FREEOBJECTS record */ bzero(dsp->dsa_drr, sizeof (dmu_replay_record_t)); dsp->dsa_drr->drr_type = DRR_FREEOBJECTS; drrfo->drr_firstobj = firstobj; drrfo->drr_numobjs = numobjs; drrfo->drr_toguid = dsp->dsa_toguid; dsp->dsa_pending_op = PENDING_FREEOBJECTS; return (0); } static int dump_dnode(dmu_sendarg_t *dsp, uint64_t object, dnode_phys_t *dnp) { struct drr_object *drro = &(dsp->dsa_drr->drr_u.drr_object); if (object < dsp->dsa_resume_object) { /* * Note: when resuming, we will visit all the dnodes in * the block of dnodes that we are resuming from. In * this case it's unnecessary to send the dnodes prior to * the one we are resuming from. We should be at most one * block's worth of dnodes behind the resume point. */ ASSERT3U(dsp->dsa_resume_object - object, <, 1 << (DNODE_BLOCK_SHIFT - DNODE_SHIFT)); return (0); } if (dnp == NULL || dnp->dn_type == DMU_OT_NONE) return (dump_freeobjects(dsp, object, 1)); if (dsp->dsa_pending_op != PENDING_NONE) { if (dump_record(dsp, NULL, 0) != 0) return (SET_ERROR(EINTR)); dsp->dsa_pending_op = PENDING_NONE; } /* write an OBJECT record */ bzero(dsp->dsa_drr, sizeof (dmu_replay_record_t)); dsp->dsa_drr->drr_type = DRR_OBJECT; drro->drr_object = object; drro->drr_type = dnp->dn_type; drro->drr_bonustype = dnp->dn_bonustype; drro->drr_blksz = dnp->dn_datablkszsec << SPA_MINBLOCKSHIFT; drro->drr_bonuslen = dnp->dn_bonuslen; drro->drr_checksumtype = dnp->dn_checksum; drro->drr_compress = dnp->dn_compress; drro->drr_toguid = dsp->dsa_toguid; if (!(dsp->dsa_featureflags & DMU_BACKUP_FEATURE_LARGE_BLOCKS) && drro->drr_blksz > SPA_OLD_MAXBLOCKSIZE) drro->drr_blksz = SPA_OLD_MAXBLOCKSIZE; if (dump_record(dsp, DN_BONUS(dnp), P2ROUNDUP(dnp->dn_bonuslen, 8)) != 0) { return (SET_ERROR(EINTR)); } /* Free anything past the end of the file. */ if (dump_free(dsp, object, (dnp->dn_maxblkid + 1) * (dnp->dn_datablkszsec << SPA_MINBLOCKSHIFT), -1ULL) != 0) return (SET_ERROR(EINTR)); if (dsp->dsa_err != 0) return (SET_ERROR(EINTR)); return (0); } static boolean_t backup_do_embed(dmu_sendarg_t *dsp, const blkptr_t *bp) { if (!BP_IS_EMBEDDED(bp)) return (B_FALSE); /* * Compression function must be legacy, or explicitly enabled. */ if ((BP_GET_COMPRESS(bp) >= ZIO_COMPRESS_LEGACY_FUNCTIONS && !(dsp->dsa_featureflags & DMU_BACKUP_FEATURE_EMBED_DATA_LZ4))) return (B_FALSE); /* * Embed type must be explicitly enabled. */ switch (BPE_GET_ETYPE(bp)) { case BP_EMBEDDED_TYPE_DATA: if (dsp->dsa_featureflags & DMU_BACKUP_FEATURE_EMBED_DATA) return (B_TRUE); break; default: return (B_FALSE); } return (B_FALSE); } /* * This is the callback function to traverse_dataset that acts as the worker * thread for dmu_send_impl. */ /*ARGSUSED*/ static int send_cb(spa_t *spa, zilog_t *zilog, const blkptr_t *bp, const zbookmark_phys_t *zb, const struct dnode_phys *dnp, void *arg) { struct send_thread_arg *sta = arg; struct send_block_record *record; uint64_t record_size; int err = 0; ASSERT(zb->zb_object == DMU_META_DNODE_OBJECT || zb->zb_object >= sta->resume.zb_object); if (sta->cancel) return (SET_ERROR(EINTR)); if (bp == NULL) { ASSERT3U(zb->zb_level, ==, ZB_DNODE_LEVEL); return (0); } else if (zb->zb_level < 0) { return (0); } record = kmem_zalloc(sizeof (struct send_block_record), KM_SLEEP); record->eos_marker = B_FALSE; record->bp = *bp; record->zb = *zb; record->indblkshift = dnp->dn_indblkshift; record->datablkszsec = dnp->dn_datablkszsec; record_size = dnp->dn_datablkszsec << SPA_MINBLOCKSHIFT; bqueue_enqueue(&sta->q, record, record_size); return (err); } /* * This function kicks off the traverse_dataset. It also handles setting the * error code of the thread in case something goes wrong, and pushes the End of * Stream record when the traverse_dataset call has finished. If there is no * dataset to traverse, the thread immediately pushes End of Stream marker. */ static void send_traverse_thread(void *arg) { struct send_thread_arg *st_arg = arg; int err; struct send_block_record *data; if (st_arg->ds != NULL) { err = traverse_dataset_resume(st_arg->ds, st_arg->fromtxg, &st_arg->resume, st_arg->flags, send_cb, st_arg); if (err != EINTR) st_arg->error_code = err; } data = kmem_zalloc(sizeof (*data), KM_SLEEP); data->eos_marker = B_TRUE; bqueue_enqueue(&st_arg->q, data, 1); thread_exit(); } /* * This function actually handles figuring out what kind of record needs to be * dumped, reading the data (which has hopefully been prefetched), and calling * the appropriate helper function. */ static int do_dump(dmu_sendarg_t *dsa, struct send_block_record *data) { dsl_dataset_t *ds = dmu_objset_ds(dsa->dsa_os); const blkptr_t *bp = &data->bp; const zbookmark_phys_t *zb = &data->zb; uint8_t indblkshift = data->indblkshift; uint16_t dblkszsec = data->datablkszsec; spa_t *spa = ds->ds_dir->dd_pool->dp_spa; dmu_object_type_t type = bp ? BP_GET_TYPE(bp) : DMU_OT_NONE; int err = 0; ASSERT3U(zb->zb_level, >=, 0); ASSERT(zb->zb_object == DMU_META_DNODE_OBJECT || zb->zb_object >= dsa->dsa_resume_object); if (zb->zb_object != DMU_META_DNODE_OBJECT && DMU_OBJECT_IS_SPECIAL(zb->zb_object)) { return (0); } else if (BP_IS_HOLE(bp) && zb->zb_object == DMU_META_DNODE_OBJECT) { uint64_t span = BP_SPAN(dblkszsec, indblkshift, zb->zb_level); uint64_t dnobj = (zb->zb_blkid * span) >> DNODE_SHIFT; err = dump_freeobjects(dsa, dnobj, span >> DNODE_SHIFT); } else if (BP_IS_HOLE(bp)) { uint64_t span = BP_SPAN(dblkszsec, indblkshift, zb->zb_level); uint64_t offset = zb->zb_blkid * span; err = dump_free(dsa, zb->zb_object, offset, span); } else if (zb->zb_level > 0 || type == DMU_OT_OBJSET) { return (0); } else if (type == DMU_OT_DNODE) { int blksz = BP_GET_LSIZE(bp); arc_flags_t aflags = ARC_FLAG_WAIT; arc_buf_t *abuf; ASSERT0(zb->zb_level); if (arc_read(NULL, spa, bp, arc_getbuf_func, &abuf, ZIO_PRIORITY_ASYNC_READ, ZIO_FLAG_CANFAIL, &aflags, zb) != 0) return (SET_ERROR(EIO)); dnode_phys_t *blk = abuf->b_data; uint64_t dnobj = zb->zb_blkid * (blksz >> DNODE_SHIFT); for (int i = 0; i < blksz >> DNODE_SHIFT; i++) { err = dump_dnode(dsa, dnobj + i, blk + i); if (err != 0) break; } (void) arc_buf_remove_ref(abuf, &abuf); } else if (type == DMU_OT_SA) { arc_flags_t aflags = ARC_FLAG_WAIT; arc_buf_t *abuf; int blksz = BP_GET_LSIZE(bp); if (arc_read(NULL, spa, bp, arc_getbuf_func, &abuf, ZIO_PRIORITY_ASYNC_READ, ZIO_FLAG_CANFAIL, &aflags, zb) != 0) return (SET_ERROR(EIO)); err = dump_spill(dsa, zb->zb_object, blksz, abuf->b_data); (void) arc_buf_remove_ref(abuf, &abuf); } else if (backup_do_embed(dsa, bp)) { /* it's an embedded level-0 block of a regular object */ int blksz = dblkszsec << SPA_MINBLOCKSHIFT; ASSERT0(zb->zb_level); err = dump_write_embedded(dsa, zb->zb_object, zb->zb_blkid * blksz, blksz, bp); } else { /* it's a level-0 block of a regular object */ arc_flags_t aflags = ARC_FLAG_WAIT; arc_buf_t *abuf; int blksz = dblkszsec << SPA_MINBLOCKSHIFT; uint64_t offset; ASSERT0(zb->zb_level); ASSERT(zb->zb_object > dsa->dsa_resume_object || (zb->zb_object == dsa->dsa_resume_object && zb->zb_blkid * blksz >= dsa->dsa_resume_offset)); if (arc_read(NULL, spa, bp, arc_getbuf_func, &abuf, ZIO_PRIORITY_ASYNC_READ, ZIO_FLAG_CANFAIL, &aflags, zb) != 0) { if (zfs_send_corrupt_data) { /* Send a block filled with 0x"zfs badd bloc" */ abuf = arc_buf_alloc(spa, blksz, &abuf, ARC_BUFC_DATA); uint64_t *ptr; for (ptr = abuf->b_data; (char *)ptr < (char *)abuf->b_data + blksz; ptr++) *ptr = 0x2f5baddb10cULL; } else { return (SET_ERROR(EIO)); } } offset = zb->zb_blkid * blksz; if (!(dsa->dsa_featureflags & DMU_BACKUP_FEATURE_LARGE_BLOCKS) && blksz > SPA_OLD_MAXBLOCKSIZE) { char *buf = abuf->b_data; while (blksz > 0 && err == 0) { int n = MIN(blksz, SPA_OLD_MAXBLOCKSIZE); err = dump_write(dsa, type, zb->zb_object, offset, n, NULL, buf); offset += n; buf += n; blksz -= n; } } else { err = dump_write(dsa, type, zb->zb_object, offset, blksz, bp, abuf->b_data); } (void) arc_buf_remove_ref(abuf, &abuf); } ASSERT(err == 0 || err == EINTR); return (err); } /* * Pop the new data off the queue, and free the old data. */ static struct send_block_record * get_next_record(bqueue_t *bq, struct send_block_record *data) { struct send_block_record *tmp = bqueue_dequeue(bq); kmem_free(data, sizeof (*data)); return (tmp); } /* * Actually do the bulk of the work in a zfs send. * * Note: Releases dp using the specified tag. */ static int dmu_send_impl(void *tag, dsl_pool_t *dp, dsl_dataset_t *to_ds, zfs_bookmark_phys_t *ancestor_zb, boolean_t is_clone, boolean_t embedok, boolean_t large_block_ok, int outfd, uint64_t resumeobj, uint64_t resumeoff, #ifdef illumos vnode_t *vp, offset_t *off) #else struct file *fp, offset_t *off) #endif { objset_t *os; dmu_replay_record_t *drr; dmu_sendarg_t *dsp; int err; uint64_t fromtxg = 0; uint64_t featureflags = 0; struct send_thread_arg to_arg = { 0 }; err = dmu_objset_from_ds(to_ds, &os); if (err != 0) { dsl_pool_rele(dp, tag); return (err); } drr = kmem_zalloc(sizeof (dmu_replay_record_t), KM_SLEEP); drr->drr_type = DRR_BEGIN; drr->drr_u.drr_begin.drr_magic = DMU_BACKUP_MAGIC; DMU_SET_STREAM_HDRTYPE(drr->drr_u.drr_begin.drr_versioninfo, DMU_SUBSTREAM); #ifdef _KERNEL if (dmu_objset_type(os) == DMU_OST_ZFS) { uint64_t version; if (zfs_get_zplprop(os, ZFS_PROP_VERSION, &version) != 0) { kmem_free(drr, sizeof (dmu_replay_record_t)); dsl_pool_rele(dp, tag); return (SET_ERROR(EINVAL)); } if (version >= ZPL_VERSION_SA) { featureflags |= DMU_BACKUP_FEATURE_SA_SPILL; } } #endif if (large_block_ok && to_ds->ds_feature_inuse[SPA_FEATURE_LARGE_BLOCKS]) featureflags |= DMU_BACKUP_FEATURE_LARGE_BLOCKS; if (embedok && spa_feature_is_active(dp->dp_spa, SPA_FEATURE_EMBEDDED_DATA)) { featureflags |= DMU_BACKUP_FEATURE_EMBED_DATA; if (spa_feature_is_active(dp->dp_spa, SPA_FEATURE_LZ4_COMPRESS)) featureflags |= DMU_BACKUP_FEATURE_EMBED_DATA_LZ4; } if (resumeobj != 0 || resumeoff != 0) { featureflags |= DMU_BACKUP_FEATURE_RESUMING; } DMU_SET_FEATUREFLAGS(drr->drr_u.drr_begin.drr_versioninfo, featureflags); drr->drr_u.drr_begin.drr_creation_time = dsl_dataset_phys(to_ds)->ds_creation_time; drr->drr_u.drr_begin.drr_type = dmu_objset_type(os); if (is_clone) drr->drr_u.drr_begin.drr_flags |= DRR_FLAG_CLONE; drr->drr_u.drr_begin.drr_toguid = dsl_dataset_phys(to_ds)->ds_guid; if (dsl_dataset_phys(to_ds)->ds_flags & DS_FLAG_CI_DATASET) drr->drr_u.drr_begin.drr_flags |= DRR_FLAG_CI_DATA; + drr->drr_u.drr_begin.drr_flags |= DRR_FLAG_FREERECORDS; if (ancestor_zb != NULL) { drr->drr_u.drr_begin.drr_fromguid = ancestor_zb->zbm_guid; fromtxg = ancestor_zb->zbm_creation_txg; } dsl_dataset_name(to_ds, drr->drr_u.drr_begin.drr_toname); if (!to_ds->ds_is_snapshot) { (void) strlcat(drr->drr_u.drr_begin.drr_toname, "@--head--", sizeof (drr->drr_u.drr_begin.drr_toname)); } dsp = kmem_zalloc(sizeof (dmu_sendarg_t), KM_SLEEP); dsp->dsa_drr = drr; dsp->dsa_outfd = outfd; dsp->dsa_proc = curproc; dsp->dsa_td = curthread; dsp->dsa_fp = fp; dsp->dsa_os = os; dsp->dsa_off = off; dsp->dsa_toguid = dsl_dataset_phys(to_ds)->ds_guid; dsp->dsa_pending_op = PENDING_NONE; - dsp->dsa_incremental = (ancestor_zb != NULL); dsp->dsa_featureflags = featureflags; dsp->dsa_resume_object = resumeobj; dsp->dsa_resume_offset = resumeoff; mutex_enter(&to_ds->ds_sendstream_lock); list_insert_head(&to_ds->ds_sendstreams, dsp); mutex_exit(&to_ds->ds_sendstream_lock); dsl_dataset_long_hold(to_ds, FTAG); dsl_pool_rele(dp, tag); void *payload = NULL; size_t payload_len = 0; if (resumeobj != 0 || resumeoff != 0) { dmu_object_info_t to_doi; err = dmu_object_info(os, resumeobj, &to_doi); if (err != 0) goto out; SET_BOOKMARK(&to_arg.resume, to_ds->ds_object, resumeobj, 0, resumeoff / to_doi.doi_data_block_size); nvlist_t *nvl = fnvlist_alloc(); fnvlist_add_uint64(nvl, "resume_object", resumeobj); fnvlist_add_uint64(nvl, "resume_offset", resumeoff); payload = fnvlist_pack(nvl, &payload_len); drr->drr_payloadlen = payload_len; fnvlist_free(nvl); } err = dump_record(dsp, payload, payload_len); fnvlist_pack_free(payload, payload_len); if (err != 0) { err = dsp->dsa_err; goto out; } err = bqueue_init(&to_arg.q, zfs_send_queue_length, offsetof(struct send_block_record, ln)); to_arg.error_code = 0; to_arg.cancel = B_FALSE; to_arg.ds = to_ds; to_arg.fromtxg = fromtxg; to_arg.flags = TRAVERSE_PRE | TRAVERSE_PREFETCH; (void) thread_create(NULL, 0, send_traverse_thread, &to_arg, 0, &p0, TS_RUN, minclsyspri); struct send_block_record *to_data; to_data = bqueue_dequeue(&to_arg.q); while (!to_data->eos_marker && err == 0) { err = do_dump(dsp, to_data); to_data = get_next_record(&to_arg.q, to_data); if (issig(JUSTLOOKING) && issig(FORREAL)) err = EINTR; } if (err != 0) { to_arg.cancel = B_TRUE; while (!to_data->eos_marker) { to_data = get_next_record(&to_arg.q, to_data); } } kmem_free(to_data, sizeof (*to_data)); bqueue_destroy(&to_arg.q); if (err == 0 && to_arg.error_code != 0) err = to_arg.error_code; if (err != 0) goto out; if (dsp->dsa_pending_op != PENDING_NONE) if (dump_record(dsp, NULL, 0) != 0) err = SET_ERROR(EINTR); if (err != 0) { if (err == EINTR && dsp->dsa_err != 0) err = dsp->dsa_err; goto out; } bzero(drr, sizeof (dmu_replay_record_t)); drr->drr_type = DRR_END; drr->drr_u.drr_end.drr_checksum = dsp->dsa_zc; drr->drr_u.drr_end.drr_toguid = dsp->dsa_toguid; if (dump_record(dsp, NULL, 0) != 0) err = dsp->dsa_err; out: mutex_enter(&to_ds->ds_sendstream_lock); list_remove(&to_ds->ds_sendstreams, dsp); mutex_exit(&to_ds->ds_sendstream_lock); kmem_free(drr, sizeof (dmu_replay_record_t)); kmem_free(dsp, sizeof (dmu_sendarg_t)); dsl_dataset_long_rele(to_ds, FTAG); return (err); } int dmu_send_obj(const char *pool, uint64_t tosnap, uint64_t fromsnap, boolean_t embedok, boolean_t large_block_ok, #ifdef illumos int outfd, vnode_t *vp, offset_t *off) #else int outfd, struct file *fp, offset_t *off) #endif { dsl_pool_t *dp; dsl_dataset_t *ds; dsl_dataset_t *fromds = NULL; int err; err = dsl_pool_hold(pool, FTAG, &dp); if (err != 0) return (err); err = dsl_dataset_hold_obj(dp, tosnap, FTAG, &ds); if (err != 0) { dsl_pool_rele(dp, FTAG); return (err); } if (fromsnap != 0) { zfs_bookmark_phys_t zb; boolean_t is_clone; err = dsl_dataset_hold_obj(dp, fromsnap, FTAG, &fromds); if (err != 0) { dsl_dataset_rele(ds, FTAG); dsl_pool_rele(dp, FTAG); return (err); } if (!dsl_dataset_is_before(ds, fromds, 0)) err = SET_ERROR(EXDEV); zb.zbm_creation_time = dsl_dataset_phys(fromds)->ds_creation_time; zb.zbm_creation_txg = dsl_dataset_phys(fromds)->ds_creation_txg; zb.zbm_guid = dsl_dataset_phys(fromds)->ds_guid; is_clone = (fromds->ds_dir != ds->ds_dir); dsl_dataset_rele(fromds, FTAG); err = dmu_send_impl(FTAG, dp, ds, &zb, is_clone, embedok, large_block_ok, outfd, 0, 0, fp, off); } else { err = dmu_send_impl(FTAG, dp, ds, NULL, B_FALSE, embedok, large_block_ok, outfd, 0, 0, fp, off); } dsl_dataset_rele(ds, FTAG); return (err); } int dmu_send(const char *tosnap, const char *fromsnap, boolean_t embedok, boolean_t large_block_ok, int outfd, uint64_t resumeobj, uint64_t resumeoff, #ifdef illumos vnode_t *vp, offset_t *off) #else struct file *fp, offset_t *off) #endif { dsl_pool_t *dp; dsl_dataset_t *ds; int err; boolean_t owned = B_FALSE; if (fromsnap != NULL && strpbrk(fromsnap, "@#") == NULL) return (SET_ERROR(EINVAL)); err = dsl_pool_hold(tosnap, FTAG, &dp); if (err != 0) return (err); if (strchr(tosnap, '@') == NULL && spa_writeable(dp->dp_spa)) { /* * We are sending a filesystem or volume. Ensure * that it doesn't change by owning the dataset. */ err = dsl_dataset_own(dp, tosnap, FTAG, &ds); owned = B_TRUE; } else { err = dsl_dataset_hold(dp, tosnap, FTAG, &ds); } if (err != 0) { dsl_pool_rele(dp, FTAG); return (err); } if (fromsnap != NULL) { zfs_bookmark_phys_t zb; boolean_t is_clone = B_FALSE; int fsnamelen = strchr(tosnap, '@') - tosnap; /* * If the fromsnap is in a different filesystem, then * mark the send stream as a clone. */ if (strncmp(tosnap, fromsnap, fsnamelen) != 0 || (fromsnap[fsnamelen] != '@' && fromsnap[fsnamelen] != '#')) { is_clone = B_TRUE; } if (strchr(fromsnap, '@')) { dsl_dataset_t *fromds; err = dsl_dataset_hold(dp, fromsnap, FTAG, &fromds); if (err == 0) { if (!dsl_dataset_is_before(ds, fromds, 0)) err = SET_ERROR(EXDEV); zb.zbm_creation_time = dsl_dataset_phys(fromds)->ds_creation_time; zb.zbm_creation_txg = dsl_dataset_phys(fromds)->ds_creation_txg; zb.zbm_guid = dsl_dataset_phys(fromds)->ds_guid; is_clone = (ds->ds_dir != fromds->ds_dir); dsl_dataset_rele(fromds, FTAG); } } else { err = dsl_bookmark_lookup(dp, fromsnap, ds, &zb); } if (err != 0) { dsl_dataset_rele(ds, FTAG); dsl_pool_rele(dp, FTAG); return (err); } err = dmu_send_impl(FTAG, dp, ds, &zb, is_clone, embedok, large_block_ok, outfd, resumeobj, resumeoff, fp, off); } else { err = dmu_send_impl(FTAG, dp, ds, NULL, B_FALSE, embedok, large_block_ok, outfd, resumeobj, resumeoff, fp, off); } if (owned) dsl_dataset_disown(ds, FTAG); else dsl_dataset_rele(ds, FTAG); return (err); } static int dmu_adjust_send_estimate_for_indirects(dsl_dataset_t *ds, uint64_t size, uint64_t *sizep) { int err; /* * Assume that space (both on-disk and in-stream) is dominated by * data. We will adjust for indirect blocks and the copies property, * but ignore per-object space used (eg, dnodes and DRR_OBJECT records). */ /* * Subtract out approximate space used by indirect blocks. * Assume most space is used by data blocks (non-indirect, non-dnode). * Assume all blocks are recordsize. Assume ditto blocks and * internal fragmentation counter out compression. * * Therefore, space used by indirect blocks is sizeof(blkptr_t) per * block, which we observe in practice. */ uint64_t recordsize; err = dsl_prop_get_int_ds(ds, "recordsize", &recordsize); if (err != 0) return (err); size -= size / recordsize * sizeof (blkptr_t); /* Add in the space for the record associated with each block. */ size += size / recordsize * sizeof (dmu_replay_record_t); *sizep = size; return (0); } int dmu_send_estimate(dsl_dataset_t *ds, dsl_dataset_t *fromds, uint64_t *sizep) { dsl_pool_t *dp = ds->ds_dir->dd_pool; int err; uint64_t size; ASSERT(dsl_pool_config_held(dp)); /* tosnap must be a snapshot */ if (!ds->ds_is_snapshot) return (SET_ERROR(EINVAL)); /* fromsnap, if provided, must be a snapshot */ if (fromds != NULL && !fromds->ds_is_snapshot) return (SET_ERROR(EINVAL)); /* * fromsnap must be an earlier snapshot from the same fs as tosnap, * or the origin's fs. */ if (fromds != NULL && !dsl_dataset_is_before(ds, fromds, 0)) return (SET_ERROR(EXDEV)); /* Get uncompressed size estimate of changed data. */ if (fromds == NULL) { size = dsl_dataset_phys(ds)->ds_uncompressed_bytes; } else { uint64_t used, comp; err = dsl_dataset_space_written(fromds, ds, &used, &comp, &size); if (err != 0) return (err); } err = dmu_adjust_send_estimate_for_indirects(ds, size, sizep); return (err); } /* * Simple callback used to traverse the blocks of a snapshot and sum their * uncompressed size */ /* ARGSUSED */ static int dmu_calculate_send_traversal(spa_t *spa, zilog_t *zilog, const blkptr_t *bp, const zbookmark_phys_t *zb, const dnode_phys_t *dnp, void *arg) { uint64_t *spaceptr = arg; if (bp != NULL && !BP_IS_HOLE(bp)) { *spaceptr += BP_GET_UCSIZE(bp); } return (0); } /* * Given a desination snapshot and a TXG, calculate the approximate size of a * send stream sent from that TXG. from_txg may be zero, indicating that the * whole snapshot will be sent. */ int dmu_send_estimate_from_txg(dsl_dataset_t *ds, uint64_t from_txg, uint64_t *sizep) { dsl_pool_t *dp = ds->ds_dir->dd_pool; int err; uint64_t size = 0; ASSERT(dsl_pool_config_held(dp)); /* tosnap must be a snapshot */ if (!dsl_dataset_is_snapshot(ds)) return (SET_ERROR(EINVAL)); /* verify that from_txg is before the provided snapshot was taken */ if (from_txg >= dsl_dataset_phys(ds)->ds_creation_txg) { return (SET_ERROR(EXDEV)); } /* * traverse the blocks of the snapshot with birth times after * from_txg, summing their uncompressed size */ err = traverse_dataset(ds, from_txg, TRAVERSE_POST, dmu_calculate_send_traversal, &size); if (err) return (err); err = dmu_adjust_send_estimate_for_indirects(ds, size, sizep); return (err); } typedef struct dmu_recv_begin_arg { const char *drba_origin; dmu_recv_cookie_t *drba_cookie; cred_t *drba_cred; uint64_t drba_snapobj; } dmu_recv_begin_arg_t; static int recv_begin_check_existing_impl(dmu_recv_begin_arg_t *drba, dsl_dataset_t *ds, uint64_t fromguid) { uint64_t val; int error; dsl_pool_t *dp = ds->ds_dir->dd_pool; /* temporary clone name must not exist */ error = zap_lookup(dp->dp_meta_objset, dsl_dir_phys(ds->ds_dir)->dd_child_dir_zapobj, recv_clone_name, 8, 1, &val); if (error != ENOENT) return (error == 0 ? EBUSY : error); /* new snapshot name must not exist */ error = zap_lookup(dp->dp_meta_objset, dsl_dataset_phys(ds)->ds_snapnames_zapobj, drba->drba_cookie->drc_tosnap, 8, 1, &val); if (error != ENOENT) return (error == 0 ? EEXIST : error); /* * Check snapshot limit before receiving. We'll recheck again at the * end, but might as well abort before receiving if we're already over * the limit. * * Note that we do not check the file system limit with * dsl_dir_fscount_check because the temporary %clones don't count * against that limit. */ error = dsl_fs_ss_limit_check(ds->ds_dir, 1, ZFS_PROP_SNAPSHOT_LIMIT, NULL, drba->drba_cred); if (error != 0) return (error); if (fromguid != 0) { dsl_dataset_t *snap; uint64_t obj = dsl_dataset_phys(ds)->ds_prev_snap_obj; /* Find snapshot in this dir that matches fromguid. */ while (obj != 0) { error = dsl_dataset_hold_obj(dp, obj, FTAG, &snap); if (error != 0) return (SET_ERROR(ENODEV)); if (snap->ds_dir != ds->ds_dir) { dsl_dataset_rele(snap, FTAG); return (SET_ERROR(ENODEV)); } if (dsl_dataset_phys(snap)->ds_guid == fromguid) break; obj = dsl_dataset_phys(snap)->ds_prev_snap_obj; dsl_dataset_rele(snap, FTAG); } if (obj == 0) return (SET_ERROR(ENODEV)); if (drba->drba_cookie->drc_force) { drba->drba_snapobj = obj; } else { /* * If we are not forcing, there must be no * changes since fromsnap. */ if (dsl_dataset_modified_since_snap(ds, snap)) { dsl_dataset_rele(snap, FTAG); return (SET_ERROR(ETXTBSY)); } drba->drba_snapobj = ds->ds_prev->ds_object; } dsl_dataset_rele(snap, FTAG); } else { /* if full, then must be forced */ if (!drba->drba_cookie->drc_force) return (SET_ERROR(EEXIST)); /* start from $ORIGIN@$ORIGIN, if supported */ drba->drba_snapobj = dp->dp_origin_snap != NULL ? dp->dp_origin_snap->ds_object : 0; } return (0); } static int dmu_recv_begin_check(void *arg, dmu_tx_t *tx) { dmu_recv_begin_arg_t *drba = arg; dsl_pool_t *dp = dmu_tx_pool(tx); struct drr_begin *drrb = drba->drba_cookie->drc_drrb; uint64_t fromguid = drrb->drr_fromguid; int flags = drrb->drr_flags; int error; uint64_t featureflags = DMU_GET_FEATUREFLAGS(drrb->drr_versioninfo); dsl_dataset_t *ds; const char *tofs = drba->drba_cookie->drc_tofs; /* already checked */ ASSERT3U(drrb->drr_magic, ==, DMU_BACKUP_MAGIC); ASSERT(!(featureflags & DMU_BACKUP_FEATURE_RESUMING)); if (DMU_GET_STREAM_HDRTYPE(drrb->drr_versioninfo) == DMU_COMPOUNDSTREAM || drrb->drr_type >= DMU_OST_NUMTYPES || ((flags & DRR_FLAG_CLONE) && drba->drba_origin == NULL)) return (SET_ERROR(EINVAL)); /* Verify pool version supports SA if SA_SPILL feature set */ if ((featureflags & DMU_BACKUP_FEATURE_SA_SPILL) && spa_version(dp->dp_spa) < SPA_VERSION_SA) return (SET_ERROR(ENOTSUP)); if (drba->drba_cookie->drc_resumable && !spa_feature_is_enabled(dp->dp_spa, SPA_FEATURE_EXTENSIBLE_DATASET)) return (SET_ERROR(ENOTSUP)); /* * The receiving code doesn't know how to translate a WRITE_EMBEDDED * record to a plan WRITE record, so the pool must have the * EMBEDDED_DATA feature enabled if the stream has WRITE_EMBEDDED * records. Same with WRITE_EMBEDDED records that use LZ4 compression. */ if ((featureflags & DMU_BACKUP_FEATURE_EMBED_DATA) && !spa_feature_is_enabled(dp->dp_spa, SPA_FEATURE_EMBEDDED_DATA)) return (SET_ERROR(ENOTSUP)); if ((featureflags & DMU_BACKUP_FEATURE_EMBED_DATA_LZ4) && !spa_feature_is_enabled(dp->dp_spa, SPA_FEATURE_LZ4_COMPRESS)) return (SET_ERROR(ENOTSUP)); /* * The receiving code doesn't know how to translate large blocks * to smaller ones, so the pool must have the LARGE_BLOCKS * feature enabled if the stream has LARGE_BLOCKS. */ if ((featureflags & DMU_BACKUP_FEATURE_LARGE_BLOCKS) && !spa_feature_is_enabled(dp->dp_spa, SPA_FEATURE_LARGE_BLOCKS)) return (SET_ERROR(ENOTSUP)); error = dsl_dataset_hold(dp, tofs, FTAG, &ds); if (error == 0) { /* target fs already exists; recv into temp clone */ /* Can't recv a clone into an existing fs */ - if (flags & DRR_FLAG_CLONE) { + if (flags & DRR_FLAG_CLONE || drba->drba_origin) { dsl_dataset_rele(ds, FTAG); return (SET_ERROR(EINVAL)); } error = recv_begin_check_existing_impl(drba, ds, fromguid); dsl_dataset_rele(ds, FTAG); } else if (error == ENOENT) { /* target fs does not exist; must be a full backup or clone */ char buf[MAXNAMELEN]; /* * If it's a non-clone incremental, we are missing the * target fs, so fail the recv. */ if (fromguid != 0 && !(flags & DRR_FLAG_CLONE || drba->drba_origin)) return (SET_ERROR(ENOENT)); + /* + * If we're receiving a full send as a clone, and it doesn't + * contain all the necessary free records and freeobject + * records, reject it. + */ + if (fromguid == 0 && drba->drba_origin && + !(flags & DRR_FLAG_FREERECORDS)) + return (SET_ERROR(EINVAL)); + /* Open the parent of tofs */ ASSERT3U(strlen(tofs), <, MAXNAMELEN); (void) strlcpy(buf, tofs, strrchr(tofs, '/') - tofs + 1); error = dsl_dataset_hold(dp, buf, FTAG, &ds); if (error != 0) return (error); /* * Check filesystem and snapshot limits before receiving. We'll * recheck snapshot limits again at the end (we create the * filesystems and increment those counts during begin_sync). */ error = dsl_fs_ss_limit_check(ds->ds_dir, 1, ZFS_PROP_FILESYSTEM_LIMIT, NULL, drba->drba_cred); if (error != 0) { dsl_dataset_rele(ds, FTAG); return (error); } error = dsl_fs_ss_limit_check(ds->ds_dir, 1, ZFS_PROP_SNAPSHOT_LIMIT, NULL, drba->drba_cred); if (error != 0) { dsl_dataset_rele(ds, FTAG); return (error); } if (drba->drba_origin != NULL) { dsl_dataset_t *origin; error = dsl_dataset_hold(dp, drba->drba_origin, FTAG, &origin); if (error != 0) { dsl_dataset_rele(ds, FTAG); return (error); } if (!origin->ds_is_snapshot) { dsl_dataset_rele(origin, FTAG); dsl_dataset_rele(ds, FTAG); return (SET_ERROR(EINVAL)); } - if (dsl_dataset_phys(origin)->ds_guid != fromguid) { + if (dsl_dataset_phys(origin)->ds_guid != fromguid && + fromguid != 0) { dsl_dataset_rele(origin, FTAG); dsl_dataset_rele(ds, FTAG); return (SET_ERROR(ENODEV)); } dsl_dataset_rele(origin, FTAG); } dsl_dataset_rele(ds, FTAG); error = 0; } return (error); } static void dmu_recv_begin_sync(void *arg, dmu_tx_t *tx) { dmu_recv_begin_arg_t *drba = arg; dsl_pool_t *dp = dmu_tx_pool(tx); objset_t *mos = dp->dp_meta_objset; struct drr_begin *drrb = drba->drba_cookie->drc_drrb; const char *tofs = drba->drba_cookie->drc_tofs; dsl_dataset_t *ds, *newds; uint64_t dsobj; int error; uint64_t crflags = 0; if (drrb->drr_flags & DRR_FLAG_CI_DATA) crflags |= DS_FLAG_CI_DATASET; error = dsl_dataset_hold(dp, tofs, FTAG, &ds); if (error == 0) { /* create temporary clone */ dsl_dataset_t *snap = NULL; if (drba->drba_snapobj != 0) { VERIFY0(dsl_dataset_hold_obj(dp, drba->drba_snapobj, FTAG, &snap)); } dsobj = dsl_dataset_create_sync(ds->ds_dir, recv_clone_name, snap, crflags, drba->drba_cred, tx); if (drba->drba_snapobj != 0) dsl_dataset_rele(snap, FTAG); dsl_dataset_rele(ds, FTAG); } else { dsl_dir_t *dd; const char *tail; dsl_dataset_t *origin = NULL; VERIFY0(dsl_dir_hold(dp, tofs, FTAG, &dd, &tail)); if (drba->drba_origin != NULL) { VERIFY0(dsl_dataset_hold(dp, drba->drba_origin, FTAG, &origin)); } /* Create new dataset. */ dsobj = dsl_dataset_create_sync(dd, strrchr(tofs, '/') + 1, origin, crflags, drba->drba_cred, tx); if (origin != NULL) dsl_dataset_rele(origin, FTAG); dsl_dir_rele(dd, FTAG); drba->drba_cookie->drc_newfs = B_TRUE; } VERIFY0(dsl_dataset_own_obj(dp, dsobj, dmu_recv_tag, &newds)); if (drba->drba_cookie->drc_resumable) { dsl_dataset_zapify(newds, tx); if (drrb->drr_fromguid != 0) { VERIFY0(zap_add(mos, dsobj, DS_FIELD_RESUME_FROMGUID, 8, 1, &drrb->drr_fromguid, tx)); } VERIFY0(zap_add(mos, dsobj, DS_FIELD_RESUME_TOGUID, 8, 1, &drrb->drr_toguid, tx)); VERIFY0(zap_add(mos, dsobj, DS_FIELD_RESUME_TONAME, 1, strlen(drrb->drr_toname) + 1, drrb->drr_toname, tx)); uint64_t one = 1; uint64_t zero = 0; VERIFY0(zap_add(mos, dsobj, DS_FIELD_RESUME_OBJECT, 8, 1, &one, tx)); VERIFY0(zap_add(mos, dsobj, DS_FIELD_RESUME_OFFSET, 8, 1, &zero, tx)); VERIFY0(zap_add(mos, dsobj, DS_FIELD_RESUME_BYTES, 8, 1, &zero, tx)); if (DMU_GET_FEATUREFLAGS(drrb->drr_versioninfo) & DMU_BACKUP_FEATURE_EMBED_DATA) { VERIFY0(zap_add(mos, dsobj, DS_FIELD_RESUME_EMBEDOK, 8, 1, &one, tx)); } } dmu_buf_will_dirty(newds->ds_dbuf, tx); dsl_dataset_phys(newds)->ds_flags |= DS_FLAG_INCONSISTENT; /* * If we actually created a non-clone, we need to create the * objset in our new dataset. */ if (BP_IS_HOLE(dsl_dataset_get_blkptr(newds))) { (void) dmu_objset_create_impl(dp->dp_spa, newds, dsl_dataset_get_blkptr(newds), drrb->drr_type, tx); } drba->drba_cookie->drc_ds = newds; spa_history_log_internal_ds(newds, "receive", tx, ""); } static int dmu_recv_resume_begin_check(void *arg, dmu_tx_t *tx) { dmu_recv_begin_arg_t *drba = arg; dsl_pool_t *dp = dmu_tx_pool(tx); struct drr_begin *drrb = drba->drba_cookie->drc_drrb; int error; uint64_t featureflags = DMU_GET_FEATUREFLAGS(drrb->drr_versioninfo); dsl_dataset_t *ds; const char *tofs = drba->drba_cookie->drc_tofs; /* already checked */ ASSERT3U(drrb->drr_magic, ==, DMU_BACKUP_MAGIC); ASSERT(featureflags & DMU_BACKUP_FEATURE_RESUMING); if (DMU_GET_STREAM_HDRTYPE(drrb->drr_versioninfo) == DMU_COMPOUNDSTREAM || drrb->drr_type >= DMU_OST_NUMTYPES) return (SET_ERROR(EINVAL)); /* Verify pool version supports SA if SA_SPILL feature set */ if ((featureflags & DMU_BACKUP_FEATURE_SA_SPILL) && spa_version(dp->dp_spa) < SPA_VERSION_SA) return (SET_ERROR(ENOTSUP)); /* * The receiving code doesn't know how to translate a WRITE_EMBEDDED * record to a plain WRITE record, so the pool must have the * EMBEDDED_DATA feature enabled if the stream has WRITE_EMBEDDED * records. Same with WRITE_EMBEDDED records that use LZ4 compression. */ if ((featureflags & DMU_BACKUP_FEATURE_EMBED_DATA) && !spa_feature_is_enabled(dp->dp_spa, SPA_FEATURE_EMBEDDED_DATA)) return (SET_ERROR(ENOTSUP)); if ((featureflags & DMU_BACKUP_FEATURE_EMBED_DATA_LZ4) && !spa_feature_is_enabled(dp->dp_spa, SPA_FEATURE_LZ4_COMPRESS)) return (SET_ERROR(ENOTSUP)); char recvname[ZFS_MAXNAMELEN]; (void) snprintf(recvname, sizeof (recvname), "%s/%s", tofs, recv_clone_name); if (dsl_dataset_hold(dp, recvname, FTAG, &ds) != 0) { /* %recv does not exist; continue in tofs */ error = dsl_dataset_hold(dp, tofs, FTAG, &ds); if (error != 0) return (error); } /* check that ds is marked inconsistent */ if (!DS_IS_INCONSISTENT(ds)) { dsl_dataset_rele(ds, FTAG); return (SET_ERROR(EINVAL)); } /* check that there is resuming data, and that the toguid matches */ if (!dsl_dataset_is_zapified(ds)) { dsl_dataset_rele(ds, FTAG); return (SET_ERROR(EINVAL)); } uint64_t val; error = zap_lookup(dp->dp_meta_objset, ds->ds_object, DS_FIELD_RESUME_TOGUID, sizeof (val), 1, &val); if (error != 0 || drrb->drr_toguid != val) { dsl_dataset_rele(ds, FTAG); return (SET_ERROR(EINVAL)); } /* * Check if the receive is still running. If so, it will be owned. * Note that nothing else can own the dataset (e.g. after the receive * fails) because it will be marked inconsistent. */ if (dsl_dataset_has_owner(ds)) { dsl_dataset_rele(ds, FTAG); return (SET_ERROR(EBUSY)); } /* There should not be any snapshots of this fs yet. */ if (ds->ds_prev != NULL && ds->ds_prev->ds_dir == ds->ds_dir) { dsl_dataset_rele(ds, FTAG); return (SET_ERROR(EINVAL)); } /* * Note: resume point will be checked when we process the first WRITE * record. */ /* check that the origin matches */ val = 0; (void) zap_lookup(dp->dp_meta_objset, ds->ds_object, DS_FIELD_RESUME_FROMGUID, sizeof (val), 1, &val); if (drrb->drr_fromguid != val) { dsl_dataset_rele(ds, FTAG); return (SET_ERROR(EINVAL)); } dsl_dataset_rele(ds, FTAG); return (0); } static void dmu_recv_resume_begin_sync(void *arg, dmu_tx_t *tx) { dmu_recv_begin_arg_t *drba = arg; dsl_pool_t *dp = dmu_tx_pool(tx); const char *tofs = drba->drba_cookie->drc_tofs; dsl_dataset_t *ds; uint64_t dsobj; char recvname[ZFS_MAXNAMELEN]; (void) snprintf(recvname, sizeof (recvname), "%s/%s", tofs, recv_clone_name); if (dsl_dataset_hold(dp, recvname, FTAG, &ds) != 0) { /* %recv does not exist; continue in tofs */ VERIFY0(dsl_dataset_hold(dp, tofs, FTAG, &ds)); drba->drba_cookie->drc_newfs = B_TRUE; } /* clear the inconsistent flag so that we can own it */ ASSERT(DS_IS_INCONSISTENT(ds)); dmu_buf_will_dirty(ds->ds_dbuf, tx); dsl_dataset_phys(ds)->ds_flags &= ~DS_FLAG_INCONSISTENT; dsobj = ds->ds_object; dsl_dataset_rele(ds, FTAG); VERIFY0(dsl_dataset_own_obj(dp, dsobj, dmu_recv_tag, &ds)); dmu_buf_will_dirty(ds->ds_dbuf, tx); dsl_dataset_phys(ds)->ds_flags |= DS_FLAG_INCONSISTENT; ASSERT(!BP_IS_HOLE(dsl_dataset_get_blkptr(ds))); drba->drba_cookie->drc_ds = ds; spa_history_log_internal_ds(ds, "resume receive", tx, ""); } /* * NB: callers *MUST* call dmu_recv_stream() if dmu_recv_begin() * succeeds; otherwise we will leak the holds on the datasets. */ int dmu_recv_begin(char *tofs, char *tosnap, dmu_replay_record_t *drr_begin, boolean_t force, boolean_t resumable, char *origin, dmu_recv_cookie_t *drc) { dmu_recv_begin_arg_t drba = { 0 }; bzero(drc, sizeof (dmu_recv_cookie_t)); drc->drc_drr_begin = drr_begin; drc->drc_drrb = &drr_begin->drr_u.drr_begin; drc->drc_tosnap = tosnap; drc->drc_tofs = tofs; drc->drc_force = force; drc->drc_resumable = resumable; drc->drc_cred = CRED(); if (drc->drc_drrb->drr_magic == BSWAP_64(DMU_BACKUP_MAGIC)) { drc->drc_byteswap = B_TRUE; fletcher_4_incremental_byteswap(drr_begin, sizeof (dmu_replay_record_t), &drc->drc_cksum); byteswap_record(drr_begin); } else if (drc->drc_drrb->drr_magic == DMU_BACKUP_MAGIC) { fletcher_4_incremental_native(drr_begin, sizeof (dmu_replay_record_t), &drc->drc_cksum); } else { return (SET_ERROR(EINVAL)); } drba.drba_origin = origin; drba.drba_cookie = drc; drba.drba_cred = CRED(); if (DMU_GET_FEATUREFLAGS(drc->drc_drrb->drr_versioninfo) & DMU_BACKUP_FEATURE_RESUMING) { return (dsl_sync_task(tofs, dmu_recv_resume_begin_check, dmu_recv_resume_begin_sync, &drba, 5, ZFS_SPACE_CHECK_NORMAL)); } else { return (dsl_sync_task(tofs, dmu_recv_begin_check, dmu_recv_begin_sync, &drba, 5, ZFS_SPACE_CHECK_NORMAL)); } } struct receive_record_arg { dmu_replay_record_t header; void *payload; /* Pointer to a buffer containing the payload */ /* * If the record is a write, pointer to the arc_buf_t containing the * payload. */ arc_buf_t *write_buf; int payload_size; uint64_t bytes_read; /* bytes read from stream when record created */ boolean_t eos_marker; /* Marks the end of the stream */ bqueue_node_t node; }; struct receive_writer_arg { objset_t *os; boolean_t byteswap; bqueue_t q; /* * These three args are used to signal to the main thread that we're * done. */ kmutex_t mutex; kcondvar_t cv; boolean_t done; int err; /* A map from guid to dataset to help handle dedup'd streams. */ avl_tree_t *guid_to_ds_map; boolean_t resumable; uint64_t last_object, last_offset; uint64_t bytes_read; /* bytes read when current record created */ }; +struct objlist { + list_t list; /* List of struct receive_objnode. */ + /* + * Last object looked up. Used to assert that objects are being looked + * up in ascending order. + */ + uint64_t last_lookup; +}; + +struct receive_objnode { + list_node_t node; + uint64_t object; +}; + struct receive_arg { objset_t *os; kthread_t *td; struct file *fp; uint64_t voff; /* The current offset in the stream */ uint64_t bytes_read; /* * A record that has had its payload read in, but hasn't yet been handed * off to the worker thread. */ struct receive_record_arg *rrd; /* A record that has had its header read in, but not its payload. */ struct receive_record_arg *next_rrd; zio_cksum_t cksum; zio_cksum_t prev_cksum; int err; boolean_t byteswap; /* Sorted list of objects not to issue prefetches for. */ - list_t ignore_obj_list; + struct objlist ignore_objlist; }; -struct receive_ign_obj_node { - list_node_t node; - uint64_t object; -}; - typedef struct guid_map_entry { uint64_t guid; dsl_dataset_t *gme_ds; avl_node_t avlnode; } guid_map_entry_t; static int guid_compare(const void *arg1, const void *arg2) { const guid_map_entry_t *gmep1 = arg1; const guid_map_entry_t *gmep2 = arg2; if (gmep1->guid < gmep2->guid) return (-1); else if (gmep1->guid > gmep2->guid) return (1); return (0); } static void free_guid_map_onexit(void *arg) { avl_tree_t *ca = arg; void *cookie = NULL; guid_map_entry_t *gmep; while ((gmep = avl_destroy_nodes(ca, &cookie)) != NULL) { dsl_dataset_long_rele(gmep->gme_ds, gmep); dsl_dataset_rele(gmep->gme_ds, gmep); kmem_free(gmep, sizeof (guid_map_entry_t)); } avl_destroy(ca); kmem_free(ca, sizeof (avl_tree_t)); } static int restore_bytes(struct receive_arg *ra, void *buf, int len, off_t off, ssize_t *resid) { struct uio auio; struct iovec aiov; int error; aiov.iov_base = buf; aiov.iov_len = len; auio.uio_iov = &aiov; auio.uio_iovcnt = 1; auio.uio_resid = len; auio.uio_segflg = UIO_SYSSPACE; auio.uio_rw = UIO_READ; auio.uio_offset = off; auio.uio_td = ra->td; #ifdef _KERNEL error = fo_read(ra->fp, &auio, ra->td->td_ucred, FOF_OFFSET, ra->td); #else fprintf(stderr, "%s: returning EOPNOTSUPP\n", __func__); error = EOPNOTSUPP; #endif *resid = auio.uio_resid; return (error); } static int receive_read(struct receive_arg *ra, int len, void *buf) { int done = 0; /* some things will require 8-byte alignment, so everything must */ ASSERT0(len % 8); while (done < len) { ssize_t resid; ra->err = restore_bytes(ra, buf + done, len - done, ra->voff, &resid); if (resid == len - done) { /* * Note: ECKSUM indicates that the receive * was interrupted and can potentially be resumed. */ ra->err = SET_ERROR(ECKSUM); } ra->voff += len - done - resid; done = len - resid; if (ra->err != 0) return (ra->err); } ra->bytes_read += len; ASSERT3U(done, ==, len); return (0); } static void byteswap_record(dmu_replay_record_t *drr) { #define DO64(X) (drr->drr_u.X = BSWAP_64(drr->drr_u.X)) #define DO32(X) (drr->drr_u.X = BSWAP_32(drr->drr_u.X)) drr->drr_type = BSWAP_32(drr->drr_type); drr->drr_payloadlen = BSWAP_32(drr->drr_payloadlen); switch (drr->drr_type) { case DRR_BEGIN: DO64(drr_begin.drr_magic); DO64(drr_begin.drr_versioninfo); DO64(drr_begin.drr_creation_time); DO32(drr_begin.drr_type); DO32(drr_begin.drr_flags); DO64(drr_begin.drr_toguid); DO64(drr_begin.drr_fromguid); break; case DRR_OBJECT: DO64(drr_object.drr_object); DO32(drr_object.drr_type); DO32(drr_object.drr_bonustype); DO32(drr_object.drr_blksz); DO32(drr_object.drr_bonuslen); DO64(drr_object.drr_toguid); break; case DRR_FREEOBJECTS: DO64(drr_freeobjects.drr_firstobj); DO64(drr_freeobjects.drr_numobjs); DO64(drr_freeobjects.drr_toguid); break; case DRR_WRITE: DO64(drr_write.drr_object); DO32(drr_write.drr_type); DO64(drr_write.drr_offset); DO64(drr_write.drr_length); DO64(drr_write.drr_toguid); ZIO_CHECKSUM_BSWAP(&drr->drr_u.drr_write.drr_key.ddk_cksum); DO64(drr_write.drr_key.ddk_prop); break; case DRR_WRITE_BYREF: DO64(drr_write_byref.drr_object); DO64(drr_write_byref.drr_offset); DO64(drr_write_byref.drr_length); DO64(drr_write_byref.drr_toguid); DO64(drr_write_byref.drr_refguid); DO64(drr_write_byref.drr_refobject); DO64(drr_write_byref.drr_refoffset); ZIO_CHECKSUM_BSWAP(&drr->drr_u.drr_write_byref. drr_key.ddk_cksum); DO64(drr_write_byref.drr_key.ddk_prop); break; case DRR_WRITE_EMBEDDED: DO64(drr_write_embedded.drr_object); DO64(drr_write_embedded.drr_offset); DO64(drr_write_embedded.drr_length); DO64(drr_write_embedded.drr_toguid); DO32(drr_write_embedded.drr_lsize); DO32(drr_write_embedded.drr_psize); break; case DRR_FREE: DO64(drr_free.drr_object); DO64(drr_free.drr_offset); DO64(drr_free.drr_length); DO64(drr_free.drr_toguid); break; case DRR_SPILL: DO64(drr_spill.drr_object); DO64(drr_spill.drr_length); DO64(drr_spill.drr_toguid); break; case DRR_END: DO64(drr_end.drr_toguid); ZIO_CHECKSUM_BSWAP(&drr->drr_u.drr_end.drr_checksum); break; } if (drr->drr_type != DRR_BEGIN) { ZIO_CHECKSUM_BSWAP(&drr->drr_u.drr_checksum.drr_checksum); } #undef DO64 #undef DO32 } static inline uint8_t deduce_nblkptr(dmu_object_type_t bonus_type, uint64_t bonus_size) { if (bonus_type == DMU_OT_SA) { return (1); } else { return (1 + ((DN_MAX_BONUSLEN - bonus_size) >> SPA_BLKPTRSHIFT)); } } static void save_resume_state(struct receive_writer_arg *rwa, uint64_t object, uint64_t offset, dmu_tx_t *tx) { int txgoff = dmu_tx_get_txg(tx) & TXG_MASK; if (!rwa->resumable) return; /* * We use ds_resume_bytes[] != 0 to indicate that we need to * update this on disk, so it must not be 0. */ ASSERT(rwa->bytes_read != 0); /* * We only resume from write records, which have a valid * (non-meta-dnode) object number. */ ASSERT(object != 0); /* * For resuming to work correctly, we must receive records in order, * sorted by object,offset. This is checked by the callers, but * assert it here for good measure. */ ASSERT3U(object, >=, rwa->os->os_dsl_dataset->ds_resume_object[txgoff]); ASSERT(object != rwa->os->os_dsl_dataset->ds_resume_object[txgoff] || offset >= rwa->os->os_dsl_dataset->ds_resume_offset[txgoff]); ASSERT3U(rwa->bytes_read, >=, rwa->os->os_dsl_dataset->ds_resume_bytes[txgoff]); rwa->os->os_dsl_dataset->ds_resume_object[txgoff] = object; rwa->os->os_dsl_dataset->ds_resume_offset[txgoff] = offset; rwa->os->os_dsl_dataset->ds_resume_bytes[txgoff] = rwa->bytes_read; } static int receive_object(struct receive_writer_arg *rwa, struct drr_object *drro, void *data) { dmu_object_info_t doi; dmu_tx_t *tx; uint64_t object; int err; if (drro->drr_type == DMU_OT_NONE || !DMU_OT_IS_VALID(drro->drr_type) || !DMU_OT_IS_VALID(drro->drr_bonustype) || drro->drr_checksumtype >= ZIO_CHECKSUM_FUNCTIONS || drro->drr_compress >= ZIO_COMPRESS_FUNCTIONS || P2PHASE(drro->drr_blksz, SPA_MINBLOCKSIZE) || drro->drr_blksz < SPA_MINBLOCKSIZE || drro->drr_blksz > spa_maxblocksize(dmu_objset_spa(rwa->os)) || drro->drr_bonuslen > DN_MAX_BONUSLEN) { return (SET_ERROR(EINVAL)); } err = dmu_object_info(rwa->os, drro->drr_object, &doi); if (err != 0 && err != ENOENT) return (SET_ERROR(EINVAL)); object = err == 0 ? drro->drr_object : DMU_NEW_OBJECT; /* * If we are losing blkptrs or changing the block size this must * be a new file instance. We must clear out the previous file * contents before we can change this type of metadata in the dnode. */ if (err == 0) { int nblkptr; nblkptr = deduce_nblkptr(drro->drr_bonustype, drro->drr_bonuslen); if (drro->drr_blksz != doi.doi_data_block_size || nblkptr < doi.doi_nblkptr) { err = dmu_free_long_range(rwa->os, drro->drr_object, 0, DMU_OBJECT_END); if (err != 0) return (SET_ERROR(EINVAL)); } } tx = dmu_tx_create(rwa->os); dmu_tx_hold_bonus(tx, object); err = dmu_tx_assign(tx, TXG_WAIT); if (err != 0) { dmu_tx_abort(tx); return (err); } if (object == DMU_NEW_OBJECT) { /* currently free, want to be allocated */ err = dmu_object_claim(rwa->os, drro->drr_object, drro->drr_type, drro->drr_blksz, drro->drr_bonustype, drro->drr_bonuslen, tx); } else if (drro->drr_type != doi.doi_type || drro->drr_blksz != doi.doi_data_block_size || drro->drr_bonustype != doi.doi_bonus_type || drro->drr_bonuslen != doi.doi_bonus_size) { /* currently allocated, but with different properties */ err = dmu_object_reclaim(rwa->os, drro->drr_object, drro->drr_type, drro->drr_blksz, drro->drr_bonustype, drro->drr_bonuslen, tx); } if (err != 0) { dmu_tx_commit(tx); return (SET_ERROR(EINVAL)); } dmu_object_set_checksum(rwa->os, drro->drr_object, drro->drr_checksumtype, tx); dmu_object_set_compress(rwa->os, drro->drr_object, drro->drr_compress, tx); if (data != NULL) { dmu_buf_t *db; VERIFY0(dmu_bonus_hold(rwa->os, drro->drr_object, FTAG, &db)); dmu_buf_will_dirty(db, tx); ASSERT3U(db->db_size, >=, drro->drr_bonuslen); bcopy(data, db->db_data, drro->drr_bonuslen); if (rwa->byteswap) { dmu_object_byteswap_t byteswap = DMU_OT_BYTESWAP(drro->drr_bonustype); dmu_ot_byteswap[byteswap].ob_func(db->db_data, drro->drr_bonuslen); } dmu_buf_rele(db, FTAG); } dmu_tx_commit(tx); return (0); } /* ARGSUSED */ static int receive_freeobjects(struct receive_writer_arg *rwa, struct drr_freeobjects *drrfo) { uint64_t obj; + int next_err = 0; if (drrfo->drr_firstobj + drrfo->drr_numobjs < drrfo->drr_firstobj) return (SET_ERROR(EINVAL)); for (obj = drrfo->drr_firstobj; - obj < drrfo->drr_firstobj + drrfo->drr_numobjs; - (void) dmu_object_next(rwa->os, &obj, FALSE, 0)) { + obj < drrfo->drr_firstobj + drrfo->drr_numobjs && next_err == 0; + next_err = dmu_object_next(rwa->os, &obj, FALSE, 0)) { int err; if (dmu_object_info(rwa->os, obj, NULL) != 0) continue; err = dmu_free_long_object(rwa->os, obj); if (err != 0) return (err); } - + if (next_err != ESRCH) + return (next_err); return (0); } static int receive_write(struct receive_writer_arg *rwa, struct drr_write *drrw, arc_buf_t *abuf) { dmu_tx_t *tx; int err; if (drrw->drr_offset + drrw->drr_length < drrw->drr_offset || !DMU_OT_IS_VALID(drrw->drr_type)) return (SET_ERROR(EINVAL)); /* * For resuming to work, records must be in increasing order * by (object, offset). */ if (drrw->drr_object < rwa->last_object || (drrw->drr_object == rwa->last_object && drrw->drr_offset < rwa->last_offset)) { return (SET_ERROR(EINVAL)); } rwa->last_object = drrw->drr_object; rwa->last_offset = drrw->drr_offset; if (dmu_object_info(rwa->os, drrw->drr_object, NULL) != 0) return (SET_ERROR(EINVAL)); tx = dmu_tx_create(rwa->os); dmu_tx_hold_write(tx, drrw->drr_object, drrw->drr_offset, drrw->drr_length); err = dmu_tx_assign(tx, TXG_WAIT); if (err != 0) { dmu_tx_abort(tx); return (err); } if (rwa->byteswap) { dmu_object_byteswap_t byteswap = DMU_OT_BYTESWAP(drrw->drr_type); dmu_ot_byteswap[byteswap].ob_func(abuf->b_data, drrw->drr_length); } dmu_buf_t *bonus; if (dmu_bonus_hold(rwa->os, drrw->drr_object, FTAG, &bonus) != 0) return (SET_ERROR(EINVAL)); dmu_assign_arcbuf(bonus, drrw->drr_offset, abuf, tx); /* * Note: If the receive fails, we want the resume stream to start * with the same record that we last successfully received (as opposed * to the next record), so that we can verify that we are * resuming from the correct location. */ save_resume_state(rwa, drrw->drr_object, drrw->drr_offset, tx); dmu_tx_commit(tx); dmu_buf_rele(bonus, FTAG); return (0); } /* * Handle a DRR_WRITE_BYREF record. This record is used in dedup'ed * streams to refer to a copy of the data that is already on the * system because it came in earlier in the stream. This function * finds the earlier copy of the data, and uses that copy instead of * data from the stream to fulfill this write. */ static int receive_write_byref(struct receive_writer_arg *rwa, struct drr_write_byref *drrwbr) { dmu_tx_t *tx; int err; guid_map_entry_t gmesrch; guid_map_entry_t *gmep; avl_index_t where; objset_t *ref_os = NULL; dmu_buf_t *dbp; if (drrwbr->drr_offset + drrwbr->drr_length < drrwbr->drr_offset) return (SET_ERROR(EINVAL)); /* * If the GUID of the referenced dataset is different from the * GUID of the target dataset, find the referenced dataset. */ if (drrwbr->drr_toguid != drrwbr->drr_refguid) { gmesrch.guid = drrwbr->drr_refguid; if ((gmep = avl_find(rwa->guid_to_ds_map, &gmesrch, &where)) == NULL) { return (SET_ERROR(EINVAL)); } if (dmu_objset_from_ds(gmep->gme_ds, &ref_os)) return (SET_ERROR(EINVAL)); } else { ref_os = rwa->os; } err = dmu_buf_hold(ref_os, drrwbr->drr_refobject, drrwbr->drr_refoffset, FTAG, &dbp, DMU_READ_PREFETCH); if (err != 0) return (err); tx = dmu_tx_create(rwa->os); dmu_tx_hold_write(tx, drrwbr->drr_object, drrwbr->drr_offset, drrwbr->drr_length); err = dmu_tx_assign(tx, TXG_WAIT); if (err != 0) { dmu_tx_abort(tx); return (err); } dmu_write(rwa->os, drrwbr->drr_object, drrwbr->drr_offset, drrwbr->drr_length, dbp->db_data, tx); dmu_buf_rele(dbp, FTAG); /* See comment in restore_write. */ save_resume_state(rwa, drrwbr->drr_object, drrwbr->drr_offset, tx); dmu_tx_commit(tx); return (0); } static int receive_write_embedded(struct receive_writer_arg *rwa, struct drr_write_embedded *drrwe, void *data) { dmu_tx_t *tx; int err; if (drrwe->drr_offset + drrwe->drr_length < drrwe->drr_offset) return (EINVAL); if (drrwe->drr_psize > BPE_PAYLOAD_SIZE) return (EINVAL); if (drrwe->drr_etype >= NUM_BP_EMBEDDED_TYPES) return (EINVAL); if (drrwe->drr_compression >= ZIO_COMPRESS_FUNCTIONS) return (EINVAL); tx = dmu_tx_create(rwa->os); dmu_tx_hold_write(tx, drrwe->drr_object, drrwe->drr_offset, drrwe->drr_length); err = dmu_tx_assign(tx, TXG_WAIT); if (err != 0) { dmu_tx_abort(tx); return (err); } dmu_write_embedded(rwa->os, drrwe->drr_object, drrwe->drr_offset, data, drrwe->drr_etype, drrwe->drr_compression, drrwe->drr_lsize, drrwe->drr_psize, rwa->byteswap ^ ZFS_HOST_BYTEORDER, tx); /* See comment in restore_write. */ save_resume_state(rwa, drrwe->drr_object, drrwe->drr_offset, tx); dmu_tx_commit(tx); return (0); } static int receive_spill(struct receive_writer_arg *rwa, struct drr_spill *drrs, void *data) { dmu_tx_t *tx; dmu_buf_t *db, *db_spill; int err; if (drrs->drr_length < SPA_MINBLOCKSIZE || drrs->drr_length > spa_maxblocksize(dmu_objset_spa(rwa->os))) return (SET_ERROR(EINVAL)); if (dmu_object_info(rwa->os, drrs->drr_object, NULL) != 0) return (SET_ERROR(EINVAL)); VERIFY0(dmu_bonus_hold(rwa->os, drrs->drr_object, FTAG, &db)); if ((err = dmu_spill_hold_by_bonus(db, FTAG, &db_spill)) != 0) { dmu_buf_rele(db, FTAG); return (err); } tx = dmu_tx_create(rwa->os); dmu_tx_hold_spill(tx, db->db_object); err = dmu_tx_assign(tx, TXG_WAIT); if (err != 0) { dmu_buf_rele(db, FTAG); dmu_buf_rele(db_spill, FTAG); dmu_tx_abort(tx); return (err); } dmu_buf_will_dirty(db_spill, tx); if (db_spill->db_size < drrs->drr_length) VERIFY(0 == dbuf_spill_set_blksz(db_spill, drrs->drr_length, tx)); bcopy(data, db_spill->db_data, drrs->drr_length); dmu_buf_rele(db, FTAG); dmu_buf_rele(db_spill, FTAG); dmu_tx_commit(tx); return (0); } /* ARGSUSED */ static int receive_free(struct receive_writer_arg *rwa, struct drr_free *drrf) { int err; if (drrf->drr_length != -1ULL && drrf->drr_offset + drrf->drr_length < drrf->drr_offset) return (SET_ERROR(EINVAL)); if (dmu_object_info(rwa->os, drrf->drr_object, NULL) != 0) return (SET_ERROR(EINVAL)); err = dmu_free_long_range(rwa->os, drrf->drr_object, drrf->drr_offset, drrf->drr_length); return (err); } /* used to destroy the drc_ds on error */ static void dmu_recv_cleanup_ds(dmu_recv_cookie_t *drc) { if (drc->drc_resumable) { /* wait for our resume state to be written to disk */ txg_wait_synced(drc->drc_ds->ds_dir->dd_pool, 0); dsl_dataset_disown(drc->drc_ds, dmu_recv_tag); } else { char name[MAXNAMELEN]; dsl_dataset_name(drc->drc_ds, name); dsl_dataset_disown(drc->drc_ds, dmu_recv_tag); (void) dsl_destroy_head(name); } } static void receive_cksum(struct receive_arg *ra, int len, void *buf) { if (ra->byteswap) { fletcher_4_incremental_byteswap(buf, len, &ra->cksum); } else { fletcher_4_incremental_native(buf, len, &ra->cksum); } } /* * Read the payload into a buffer of size len, and update the current record's * payload field. * Allocate ra->next_rrd and read the next record's header into * ra->next_rrd->header. * Verify checksum of payload and next record. */ static int receive_read_payload_and_next_header(struct receive_arg *ra, int len, void *buf) { int err; if (len != 0) { ASSERT3U(len, <=, SPA_MAXBLOCKSIZE); err = receive_read(ra, len, buf); if (err != 0) return (err); receive_cksum(ra, len, buf); /* note: rrd is NULL when reading the begin record's payload */ if (ra->rrd != NULL) { ra->rrd->payload = buf; ra->rrd->payload_size = len; ra->rrd->bytes_read = ra->bytes_read; } } ra->prev_cksum = ra->cksum; ra->next_rrd = kmem_zalloc(sizeof (*ra->next_rrd), KM_SLEEP); err = receive_read(ra, sizeof (ra->next_rrd->header), &ra->next_rrd->header); ra->next_rrd->bytes_read = ra->bytes_read; if (err != 0) { kmem_free(ra->next_rrd, sizeof (*ra->next_rrd)); ra->next_rrd = NULL; return (err); } if (ra->next_rrd->header.drr_type == DRR_BEGIN) { kmem_free(ra->next_rrd, sizeof (*ra->next_rrd)); ra->next_rrd = NULL; return (SET_ERROR(EINVAL)); } /* * Note: checksum is of everything up to but not including the * checksum itself. */ ASSERT3U(offsetof(dmu_replay_record_t, drr_u.drr_checksum.drr_checksum), ==, sizeof (dmu_replay_record_t) - sizeof (zio_cksum_t)); receive_cksum(ra, offsetof(dmu_replay_record_t, drr_u.drr_checksum.drr_checksum), &ra->next_rrd->header); zio_cksum_t cksum_orig = ra->next_rrd->header.drr_u.drr_checksum.drr_checksum; zio_cksum_t *cksump = &ra->next_rrd->header.drr_u.drr_checksum.drr_checksum; if (ra->byteswap) byteswap_record(&ra->next_rrd->header); if ((!ZIO_CHECKSUM_IS_ZERO(cksump)) && !ZIO_CHECKSUM_EQUAL(ra->cksum, *cksump)) { kmem_free(ra->next_rrd, sizeof (*ra->next_rrd)); ra->next_rrd = NULL; return (SET_ERROR(ECKSUM)); } receive_cksum(ra, sizeof (cksum_orig), &cksum_orig); return (0); } +static void +objlist_create(struct objlist *list) +{ + list_create(&list->list, sizeof (struct receive_objnode), + offsetof(struct receive_objnode, node)); + list->last_lookup = 0; +} + +static void +objlist_destroy(struct objlist *list) +{ + for (struct receive_objnode *n = list_remove_head(&list->list); + n != NULL; n = list_remove_head(&list->list)) { + kmem_free(n, sizeof (*n)); + } + list_destroy(&list->list); +} + /* + * This function looks through the objlist to see if the specified object number + * is contained in the objlist. In the process, it will remove all object + * numbers in the list that are smaller than the specified object number. Thus, + * any lookup of an object number smaller than a previously looked up object + * number will always return false; therefore, all lookups should be done in + * ascending order. + */ +static boolean_t +objlist_exists(struct objlist *list, uint64_t object) +{ + struct receive_objnode *node = list_head(&list->list); + ASSERT3U(object, >=, list->last_lookup); + list->last_lookup = object; + while (node != NULL && node->object < object) { + VERIFY3P(node, ==, list_remove_head(&list->list)); + kmem_free(node, sizeof (*node)); + node = list_head(&list->list); + } + return (node != NULL && node->object == object); +} + +/* + * The objlist is a list of object numbers stored in ascending order. However, + * the insertion of new object numbers does not seek out the correct location to + * store a new object number; instead, it appends it to the list for simplicity. + * Thus, any users must take care to only insert new object numbers in ascending + * order. + */ +static void +objlist_insert(struct objlist *list, uint64_t object) +{ + struct receive_objnode *node = kmem_zalloc(sizeof (*node), KM_SLEEP); + node->object = object; +#ifdef ZFS_DEBUG + struct receive_objnode *last_object = list_tail(&list->list); + uint64_t last_objnum = (last_object != NULL ? last_object->object : 0); + ASSERT3U(node->object, >, last_objnum); +#endif + list_insert_tail(&list->list, node); +} + +/* * Issue the prefetch reads for any necessary indirect blocks. * * We use the object ignore list to tell us whether or not to issue prefetches * for a given object. We do this for both correctness (in case the blocksize * of an object has changed) and performance (if the object doesn't exist, don't * needlessly try to issue prefetches). We also trim the list as we go through * the stream to prevent it from growing to an unbounded size. * * The object numbers within will always be in sorted order, and any write * records we see will also be in sorted order, but they're not sorted with * respect to each other (i.e. we can get several object records before * receiving each object's write records). As a result, once we've reached a * given object number, we can safely remove any reference to lower object * numbers in the ignore list. In practice, we receive up to 32 object records * before receiving write records, so the list can have up to 32 nodes in it. */ /* ARGSUSED */ static void receive_read_prefetch(struct receive_arg *ra, uint64_t object, uint64_t offset, uint64_t length) { - struct receive_ign_obj_node *node = list_head(&ra->ignore_obj_list); - while (node != NULL && node->object < object) { - VERIFY3P(node, ==, list_remove_head(&ra->ignore_obj_list)); - kmem_free(node, sizeof (*node)); - node = list_head(&ra->ignore_obj_list); - } - if (node == NULL || node->object > object) { + if (!objlist_exists(&ra->ignore_objlist, object)) { dmu_prefetch(ra->os, object, 1, offset, length, ZIO_PRIORITY_SYNC_READ); } } /* * Read records off the stream, issuing any necessary prefetches. */ static int receive_read_record(struct receive_arg *ra) { int err; switch (ra->rrd->header.drr_type) { case DRR_OBJECT: { struct drr_object *drro = &ra->rrd->header.drr_u.drr_object; uint32_t size = P2ROUNDUP(drro->drr_bonuslen, 8); void *buf = kmem_zalloc(size, KM_SLEEP); dmu_object_info_t doi; err = receive_read_payload_and_next_header(ra, size, buf); if (err != 0) { kmem_free(buf, size); return (err); } err = dmu_object_info(ra->os, drro->drr_object, &doi); /* * See receive_read_prefetch for an explanation why we're * storing this object in the ignore_obj_list. */ if (err == ENOENT || (err == 0 && doi.doi_data_block_size != drro->drr_blksz)) { - struct receive_ign_obj_node *node = - kmem_zalloc(sizeof (*node), - KM_SLEEP); - node->object = drro->drr_object; -#ifdef ZFS_DEBUG - struct receive_ign_obj_node *last_object = - list_tail(&ra->ignore_obj_list); - uint64_t last_objnum = (last_object != NULL ? - last_object->object : 0); - ASSERT3U(node->object, >, last_objnum); -#endif - list_insert_tail(&ra->ignore_obj_list, node); + objlist_insert(&ra->ignore_objlist, drro->drr_object); err = 0; } return (err); } case DRR_FREEOBJECTS: { err = receive_read_payload_and_next_header(ra, 0, NULL); return (err); } case DRR_WRITE: { struct drr_write *drrw = &ra->rrd->header.drr_u.drr_write; arc_buf_t *abuf = arc_loan_buf(dmu_objset_spa(ra->os), drrw->drr_length); err = receive_read_payload_and_next_header(ra, drrw->drr_length, abuf->b_data); if (err != 0) { dmu_return_arcbuf(abuf); return (err); } ra->rrd->write_buf = abuf; receive_read_prefetch(ra, drrw->drr_object, drrw->drr_offset, drrw->drr_length); return (err); } case DRR_WRITE_BYREF: { struct drr_write_byref *drrwb = &ra->rrd->header.drr_u.drr_write_byref; err = receive_read_payload_and_next_header(ra, 0, NULL); receive_read_prefetch(ra, drrwb->drr_object, drrwb->drr_offset, drrwb->drr_length); return (err); } case DRR_WRITE_EMBEDDED: { struct drr_write_embedded *drrwe = &ra->rrd->header.drr_u.drr_write_embedded; uint32_t size = P2ROUNDUP(drrwe->drr_psize, 8); void *buf = kmem_zalloc(size, KM_SLEEP); err = receive_read_payload_and_next_header(ra, size, buf); if (err != 0) { kmem_free(buf, size); return (err); } receive_read_prefetch(ra, drrwe->drr_object, drrwe->drr_offset, drrwe->drr_length); return (err); } case DRR_FREE: { /* * It might be beneficial to prefetch indirect blocks here, but * we don't really have the data to decide for sure. */ err = receive_read_payload_and_next_header(ra, 0, NULL); return (err); } case DRR_END: { struct drr_end *drre = &ra->rrd->header.drr_u.drr_end; if (!ZIO_CHECKSUM_EQUAL(ra->prev_cksum, drre->drr_checksum)) return (SET_ERROR(ECKSUM)); return (0); } case DRR_SPILL: { struct drr_spill *drrs = &ra->rrd->header.drr_u.drr_spill; void *buf = kmem_zalloc(drrs->drr_length, KM_SLEEP); err = receive_read_payload_and_next_header(ra, drrs->drr_length, buf); if (err != 0) kmem_free(buf, drrs->drr_length); return (err); } default: return (SET_ERROR(EINVAL)); } } /* * Commit the records to the pool. */ static int receive_process_record(struct receive_writer_arg *rwa, struct receive_record_arg *rrd) { int err; /* Processing in order, therefore bytes_read should be increasing. */ ASSERT3U(rrd->bytes_read, >=, rwa->bytes_read); rwa->bytes_read = rrd->bytes_read; switch (rrd->header.drr_type) { case DRR_OBJECT: { struct drr_object *drro = &rrd->header.drr_u.drr_object; err = receive_object(rwa, drro, rrd->payload); kmem_free(rrd->payload, rrd->payload_size); rrd->payload = NULL; return (err); } case DRR_FREEOBJECTS: { struct drr_freeobjects *drrfo = &rrd->header.drr_u.drr_freeobjects; return (receive_freeobjects(rwa, drrfo)); } case DRR_WRITE: { struct drr_write *drrw = &rrd->header.drr_u.drr_write; err = receive_write(rwa, drrw, rrd->write_buf); /* if receive_write() is successful, it consumes the arc_buf */ if (err != 0) dmu_return_arcbuf(rrd->write_buf); rrd->write_buf = NULL; rrd->payload = NULL; return (err); } case DRR_WRITE_BYREF: { struct drr_write_byref *drrwbr = &rrd->header.drr_u.drr_write_byref; return (receive_write_byref(rwa, drrwbr)); } case DRR_WRITE_EMBEDDED: { struct drr_write_embedded *drrwe = &rrd->header.drr_u.drr_write_embedded; err = receive_write_embedded(rwa, drrwe, rrd->payload); kmem_free(rrd->payload, rrd->payload_size); rrd->payload = NULL; return (err); } case DRR_FREE: { struct drr_free *drrf = &rrd->header.drr_u.drr_free; return (receive_free(rwa, drrf)); } case DRR_SPILL: { struct drr_spill *drrs = &rrd->header.drr_u.drr_spill; err = receive_spill(rwa, drrs, rrd->payload); kmem_free(rrd->payload, rrd->payload_size); rrd->payload = NULL; return (err); } default: return (SET_ERROR(EINVAL)); } } /* * dmu_recv_stream's worker thread; pull records off the queue, and then call * receive_process_record When we're done, signal the main thread and exit. */ static void receive_writer_thread(void *arg) { struct receive_writer_arg *rwa = arg; struct receive_record_arg *rrd; for (rrd = bqueue_dequeue(&rwa->q); !rrd->eos_marker; rrd = bqueue_dequeue(&rwa->q)) { /* * If there's an error, the main thread will stop putting things * on the queue, but we need to clear everything in it before we * can exit. */ if (rwa->err == 0) { rwa->err = receive_process_record(rwa, rrd); } else if (rrd->write_buf != NULL) { dmu_return_arcbuf(rrd->write_buf); rrd->write_buf = NULL; rrd->payload = NULL; } else if (rrd->payload != NULL) { kmem_free(rrd->payload, rrd->payload_size); rrd->payload = NULL; } kmem_free(rrd, sizeof (*rrd)); } kmem_free(rrd, sizeof (*rrd)); mutex_enter(&rwa->mutex); rwa->done = B_TRUE; cv_signal(&rwa->cv); mutex_exit(&rwa->mutex); thread_exit(); } static int resume_check(struct receive_arg *ra, nvlist_t *begin_nvl) { uint64_t val; objset_t *mos = dmu_objset_pool(ra->os)->dp_meta_objset; uint64_t dsobj = dmu_objset_id(ra->os); uint64_t resume_obj, resume_off; if (nvlist_lookup_uint64(begin_nvl, "resume_object", &resume_obj) != 0 || nvlist_lookup_uint64(begin_nvl, "resume_offset", &resume_off) != 0) { return (SET_ERROR(EINVAL)); } VERIFY0(zap_lookup(mos, dsobj, DS_FIELD_RESUME_OBJECT, sizeof (val), 1, &val)); if (resume_obj != val) return (SET_ERROR(EINVAL)); VERIFY0(zap_lookup(mos, dsobj, DS_FIELD_RESUME_OFFSET, sizeof (val), 1, &val)); if (resume_off != val) return (SET_ERROR(EINVAL)); return (0); } - /* * Read in the stream's records, one by one, and apply them to the pool. There * are two threads involved; the thread that calls this function will spin up a * worker thread, read the records off the stream one by one, and issue * prefetches for any necessary indirect blocks. It will then push the records * onto an internal blocking queue. The worker thread will pull the records off * the queue, and actually write the data into the DMU. This way, the worker * thread doesn't have to wait for reads to complete, since everything it needs * (the indirect blocks) will be prefetched. * * NB: callers *must* call dmu_recv_end() if this succeeds. */ int dmu_recv_stream(dmu_recv_cookie_t *drc, struct file *fp, offset_t *voffp, int cleanup_fd, uint64_t *action_handlep) { int err = 0; struct receive_arg ra = { 0 }; struct receive_writer_arg rwa = { 0 }; int featureflags; nvlist_t *begin_nvl = NULL; ra.byteswap = drc->drc_byteswap; ra.cksum = drc->drc_cksum; ra.td = curthread; ra.fp = fp; ra.voff = *voffp; if (dsl_dataset_is_zapified(drc->drc_ds)) { (void) zap_lookup(drc->drc_ds->ds_dir->dd_pool->dp_meta_objset, drc->drc_ds->ds_object, DS_FIELD_RESUME_BYTES, sizeof (ra.bytes_read), 1, &ra.bytes_read); } - list_create(&ra.ignore_obj_list, sizeof (struct receive_ign_obj_node), - offsetof(struct receive_ign_obj_node, node)); + objlist_create(&ra.ignore_objlist); /* these were verified in dmu_recv_begin */ ASSERT3U(DMU_GET_STREAM_HDRTYPE(drc->drc_drrb->drr_versioninfo), ==, DMU_SUBSTREAM); ASSERT3U(drc->drc_drrb->drr_type, <, DMU_OST_NUMTYPES); /* * Open the objset we are modifying. */ VERIFY0(dmu_objset_from_ds(drc->drc_ds, &ra.os)); ASSERT(dsl_dataset_phys(drc->drc_ds)->ds_flags & DS_FLAG_INCONSISTENT); featureflags = DMU_GET_FEATUREFLAGS(drc->drc_drrb->drr_versioninfo); /* if this stream is dedup'ed, set up the avl tree for guid mapping */ if (featureflags & DMU_BACKUP_FEATURE_DEDUP) { minor_t minor; if (cleanup_fd == -1) { ra.err = SET_ERROR(EBADF); goto out; } ra.err = zfs_onexit_fd_hold(cleanup_fd, &minor); if (ra.err != 0) { cleanup_fd = -1; goto out; } if (*action_handlep == 0) { rwa.guid_to_ds_map = kmem_alloc(sizeof (avl_tree_t), KM_SLEEP); avl_create(rwa.guid_to_ds_map, guid_compare, sizeof (guid_map_entry_t), offsetof(guid_map_entry_t, avlnode)); err = zfs_onexit_add_cb(minor, free_guid_map_onexit, rwa.guid_to_ds_map, action_handlep); if (ra.err != 0) goto out; } else { err = zfs_onexit_cb_data(minor, *action_handlep, (void **)&rwa.guid_to_ds_map); if (ra.err != 0) goto out; } drc->drc_guid_to_ds_map = rwa.guid_to_ds_map; } uint32_t payloadlen = drc->drc_drr_begin->drr_payloadlen; void *payload = NULL; if (payloadlen != 0) payload = kmem_alloc(payloadlen, KM_SLEEP); err = receive_read_payload_and_next_header(&ra, payloadlen, payload); if (err != 0) { if (payloadlen != 0) kmem_free(payload, payloadlen); goto out; } if (payloadlen != 0) { err = nvlist_unpack(payload, payloadlen, &begin_nvl, KM_SLEEP); kmem_free(payload, payloadlen); if (err != 0) goto out; } if (featureflags & DMU_BACKUP_FEATURE_RESUMING) { err = resume_check(&ra, begin_nvl); if (err != 0) goto out; } (void) bqueue_init(&rwa.q, zfs_recv_queue_length, offsetof(struct receive_record_arg, node)); cv_init(&rwa.cv, NULL, CV_DEFAULT, NULL); mutex_init(&rwa.mutex, NULL, MUTEX_DEFAULT, NULL); rwa.os = ra.os; rwa.byteswap = drc->drc_byteswap; rwa.resumable = drc->drc_resumable; (void) thread_create(NULL, 0, receive_writer_thread, &rwa, 0, &p0, TS_RUN, minclsyspri); /* * We're reading rwa.err without locks, which is safe since we are the * only reader, and the worker thread is the only writer. It's ok if we * miss a write for an iteration or two of the loop, since the writer * thread will keep freeing records we send it until we send it an eos * marker. * * We can leave this loop in 3 ways: First, if rwa.err is * non-zero. In that case, the writer thread will free the rrd we just * pushed. Second, if we're interrupted; in that case, either it's the * first loop and ra.rrd was never allocated, or it's later, and ra.rrd * has been handed off to the writer thread who will free it. Finally, * if receive_read_record fails or we're at the end of the stream, then * we free ra.rrd and exit. */ while (rwa.err == 0) { if (issig(JUSTLOOKING) && issig(FORREAL)) { err = SET_ERROR(EINTR); break; } ASSERT3P(ra.rrd, ==, NULL); ra.rrd = ra.next_rrd; ra.next_rrd = NULL; /* Allocates and loads header into ra.next_rrd */ err = receive_read_record(&ra); if (ra.rrd->header.drr_type == DRR_END || err != 0) { kmem_free(ra.rrd, sizeof (*ra.rrd)); ra.rrd = NULL; break; } bqueue_enqueue(&rwa.q, ra.rrd, sizeof (struct receive_record_arg) + ra.rrd->payload_size); ra.rrd = NULL; } if (ra.next_rrd == NULL) ra.next_rrd = kmem_zalloc(sizeof (*ra.next_rrd), KM_SLEEP); ra.next_rrd->eos_marker = B_TRUE; bqueue_enqueue(&rwa.q, ra.next_rrd, 1); mutex_enter(&rwa.mutex); while (!rwa.done) { cv_wait(&rwa.cv, &rwa.mutex); } mutex_exit(&rwa.mutex); cv_destroy(&rwa.cv); mutex_destroy(&rwa.mutex); bqueue_destroy(&rwa.q); if (err == 0) err = rwa.err; out: nvlist_free(begin_nvl); if ((featureflags & DMU_BACKUP_FEATURE_DEDUP) && (cleanup_fd != -1)) zfs_onexit_fd_rele(cleanup_fd); if (err != 0) { /* * Clean up references. If receive is not resumable, * destroy what we created, so we don't leave it in * the inconsistent state. */ dmu_recv_cleanup_ds(drc); } *voffp = ra.voff; - for (struct receive_ign_obj_node *n = - list_remove_head(&ra.ignore_obj_list); n != NULL; - n = list_remove_head(&ra.ignore_obj_list)) { - kmem_free(n, sizeof (*n)); - } - list_destroy(&ra.ignore_obj_list); + objlist_destroy(&ra.ignore_objlist); return (err); } static int dmu_recv_end_check(void *arg, dmu_tx_t *tx) { dmu_recv_cookie_t *drc = arg; dsl_pool_t *dp = dmu_tx_pool(tx); int error; ASSERT3P(drc->drc_ds->ds_owner, ==, dmu_recv_tag); if (!drc->drc_newfs) { dsl_dataset_t *origin_head; error = dsl_dataset_hold(dp, drc->drc_tofs, FTAG, &origin_head); if (error != 0) return (error); if (drc->drc_force) { /* * We will destroy any snapshots in tofs (i.e. before * origin_head) that are after the origin (which is * the snap before drc_ds, because drc_ds can not * have any snaps of its own). */ uint64_t obj; obj = dsl_dataset_phys(origin_head)->ds_prev_snap_obj; while (obj != dsl_dataset_phys(drc->drc_ds)->ds_prev_snap_obj) { dsl_dataset_t *snap; error = dsl_dataset_hold_obj(dp, obj, FTAG, &snap); if (error != 0) break; if (snap->ds_dir != origin_head->ds_dir) error = SET_ERROR(EINVAL); if (error == 0) { error = dsl_destroy_snapshot_check_impl( snap, B_FALSE); } obj = dsl_dataset_phys(snap)->ds_prev_snap_obj; dsl_dataset_rele(snap, FTAG); if (error != 0) break; } if (error != 0) { dsl_dataset_rele(origin_head, FTAG); return (error); } } error = dsl_dataset_clone_swap_check_impl(drc->drc_ds, origin_head, drc->drc_force, drc->drc_owner, tx); if (error != 0) { dsl_dataset_rele(origin_head, FTAG); return (error); } error = dsl_dataset_snapshot_check_impl(origin_head, drc->drc_tosnap, tx, B_TRUE, 1, drc->drc_cred); dsl_dataset_rele(origin_head, FTAG); if (error != 0) return (error); error = dsl_destroy_head_check_impl(drc->drc_ds, 1); } else { error = dsl_dataset_snapshot_check_impl(drc->drc_ds, drc->drc_tosnap, tx, B_TRUE, 1, drc->drc_cred); } return (error); } static void dmu_recv_end_sync(void *arg, dmu_tx_t *tx) { dmu_recv_cookie_t *drc = arg; dsl_pool_t *dp = dmu_tx_pool(tx); spa_history_log_internal_ds(drc->drc_ds, "finish receiving", tx, "snap=%s", drc->drc_tosnap); if (!drc->drc_newfs) { dsl_dataset_t *origin_head; VERIFY0(dsl_dataset_hold(dp, drc->drc_tofs, FTAG, &origin_head)); if (drc->drc_force) { /* * Destroy any snapshots of drc_tofs (origin_head) * after the origin (the snap before drc_ds). */ uint64_t obj; obj = dsl_dataset_phys(origin_head)->ds_prev_snap_obj; while (obj != dsl_dataset_phys(drc->drc_ds)->ds_prev_snap_obj) { dsl_dataset_t *snap; VERIFY0(dsl_dataset_hold_obj(dp, obj, FTAG, &snap)); ASSERT3P(snap->ds_dir, ==, origin_head->ds_dir); obj = dsl_dataset_phys(snap)->ds_prev_snap_obj; dsl_destroy_snapshot_sync_impl(snap, B_FALSE, tx); dsl_dataset_rele(snap, FTAG); } } VERIFY3P(drc->drc_ds->ds_prev, ==, origin_head->ds_prev); dsl_dataset_clone_swap_sync_impl(drc->drc_ds, origin_head, tx); dsl_dataset_snapshot_sync_impl(origin_head, drc->drc_tosnap, tx); /* set snapshot's creation time and guid */ dmu_buf_will_dirty(origin_head->ds_prev->ds_dbuf, tx); dsl_dataset_phys(origin_head->ds_prev)->ds_creation_time = drc->drc_drrb->drr_creation_time; dsl_dataset_phys(origin_head->ds_prev)->ds_guid = drc->drc_drrb->drr_toguid; dsl_dataset_phys(origin_head->ds_prev)->ds_flags &= ~DS_FLAG_INCONSISTENT; dmu_buf_will_dirty(origin_head->ds_dbuf, tx); dsl_dataset_phys(origin_head)->ds_flags &= ~DS_FLAG_INCONSISTENT; dsl_dataset_rele(origin_head, FTAG); dsl_destroy_head_sync_impl(drc->drc_ds, tx); if (drc->drc_owner != NULL) VERIFY3P(origin_head->ds_owner, ==, drc->drc_owner); } else { dsl_dataset_t *ds = drc->drc_ds; dsl_dataset_snapshot_sync_impl(ds, drc->drc_tosnap, tx); /* set snapshot's creation time and guid */ dmu_buf_will_dirty(ds->ds_prev->ds_dbuf, tx); dsl_dataset_phys(ds->ds_prev)->ds_creation_time = drc->drc_drrb->drr_creation_time; dsl_dataset_phys(ds->ds_prev)->ds_guid = drc->drc_drrb->drr_toguid; dsl_dataset_phys(ds->ds_prev)->ds_flags &= ~DS_FLAG_INCONSISTENT; dmu_buf_will_dirty(ds->ds_dbuf, tx); dsl_dataset_phys(ds)->ds_flags &= ~DS_FLAG_INCONSISTENT; if (dsl_dataset_has_resume_receive_state(ds)) { (void) zap_remove(dp->dp_meta_objset, ds->ds_object, DS_FIELD_RESUME_FROMGUID, tx); (void) zap_remove(dp->dp_meta_objset, ds->ds_object, DS_FIELD_RESUME_OBJECT, tx); (void) zap_remove(dp->dp_meta_objset, ds->ds_object, DS_FIELD_RESUME_OFFSET, tx); (void) zap_remove(dp->dp_meta_objset, ds->ds_object, DS_FIELD_RESUME_BYTES, tx); (void) zap_remove(dp->dp_meta_objset, ds->ds_object, DS_FIELD_RESUME_TOGUID, tx); (void) zap_remove(dp->dp_meta_objset, ds->ds_object, DS_FIELD_RESUME_TONAME, tx); } } drc->drc_newsnapobj = dsl_dataset_phys(drc->drc_ds)->ds_prev_snap_obj; /* * Release the hold from dmu_recv_begin. This must be done before * we return to open context, so that when we free the dataset's dnode, * we can evict its bonus buffer. */ dsl_dataset_disown(drc->drc_ds, dmu_recv_tag); drc->drc_ds = NULL; } static int add_ds_to_guidmap(const char *name, avl_tree_t *guid_map, uint64_t snapobj) { dsl_pool_t *dp; dsl_dataset_t *snapds; guid_map_entry_t *gmep; int err; ASSERT(guid_map != NULL); err = dsl_pool_hold(name, FTAG, &dp); if (err != 0) return (err); gmep = kmem_alloc(sizeof (*gmep), KM_SLEEP); err = dsl_dataset_hold_obj(dp, snapobj, gmep, &snapds); if (err == 0) { gmep->guid = dsl_dataset_phys(snapds)->ds_guid; gmep->gme_ds = snapds; avl_add(guid_map, gmep); dsl_dataset_long_hold(snapds, gmep); } else kmem_free(gmep, sizeof (*gmep)); dsl_pool_rele(dp, FTAG); return (err); } static int dmu_recv_end_modified_blocks = 3; static int dmu_recv_existing_end(dmu_recv_cookie_t *drc) { int error; char name[MAXNAMELEN]; #ifdef _KERNEL /* * We will be destroying the ds; make sure its origin is unmounted if * necessary. */ dsl_dataset_name(drc->drc_ds, name); zfs_destroy_unmount_origin(name); #endif error = dsl_sync_task(drc->drc_tofs, dmu_recv_end_check, dmu_recv_end_sync, drc, dmu_recv_end_modified_blocks, ZFS_SPACE_CHECK_NORMAL); if (error != 0) dmu_recv_cleanup_ds(drc); return (error); } static int dmu_recv_new_end(dmu_recv_cookie_t *drc) { int error; error = dsl_sync_task(drc->drc_tofs, dmu_recv_end_check, dmu_recv_end_sync, drc, dmu_recv_end_modified_blocks, ZFS_SPACE_CHECK_NORMAL); if (error != 0) { dmu_recv_cleanup_ds(drc); } else if (drc->drc_guid_to_ds_map != NULL) { (void) add_ds_to_guidmap(drc->drc_tofs, drc->drc_guid_to_ds_map, drc->drc_newsnapobj); } return (error); } int dmu_recv_end(dmu_recv_cookie_t *drc, void *owner) { drc->drc_owner = owner; if (drc->drc_newfs) return (dmu_recv_new_end(drc)); else return (dmu_recv_existing_end(drc)); } /* * Return TRUE if this objset is currently being received into. */ boolean_t dmu_objset_is_receiving(objset_t *os) { return (os->os_dsl_dataset != NULL && os->os_dsl_dataset->ds_owner == dmu_recv_tag); } Index: stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_impl.h =================================================================== --- stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_impl.h (revision 297101) +++ stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_impl.h (revision 297102) @@ -1,316 +1,315 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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 2010 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ /* * Copyright (c) 2012, Joyent, Inc. All rights reserved. * Copyright (c) 2012, Martin Matuska . All rights reserved. - * Copyright (c) 2013, 2014 by Delphix. All rights reserved. + * Copyright (c) 2013, 2015 by Delphix. All rights reserved. */ #ifndef _SYS_DMU_IMPL_H #define _SYS_DMU_IMPL_H #include #include #include #include #include #include #ifdef __cplusplus extern "C" { #endif /* * This is the locking strategy for the DMU. Numbers in parenthesis are * cases that use that lock order, referenced below: * * ARC is self-contained * bplist is self-contained * refcount is self-contained * txg is self-contained (hopefully!) * zst_lock * zf_rwlock * * XXX try to improve evicting path? * * dp_config_rwlock > os_obj_lock > dn_struct_rwlock > * dn_dbufs_mtx > hash_mutexes > db_mtx > dd_lock > leafs * * dp_config_rwlock * must be held before: everything * protects dd namespace changes * protects property changes globally * held from: * dsl_dir_open/r: * dsl_dir_create_sync/w: * dsl_dir_sync_destroy/w: * dsl_dir_rename_sync/w: * dsl_prop_changed_notify/r: * * os_obj_lock * must be held before: * everything except dp_config_rwlock * protects os_obj_next * held from: * dmu_object_alloc: dn_dbufs_mtx, db_mtx, hash_mutexes, dn_struct_rwlock * * dn_struct_rwlock * must be held before: * everything except dp_config_rwlock and os_obj_lock * protects structure of dnode (eg. nlevels) * db_blkptr can change when syncing out change to nlevels * dn_maxblkid * dn_nlevels * dn_*blksz* * phys nlevels, maxblkid, physical blkptr_t's (?) * held from: * callers of dbuf_read_impl, dbuf_hold[_impl], dbuf_prefetch * dmu_object_info_from_dnode: dn_dirty_mtx (dn_datablksz) * dmu_tx_count_free: * dbuf_read_impl: db_mtx, dmu_zfetch() * dmu_zfetch: zf_rwlock/r, zst_lock, dbuf_prefetch() * dbuf_new_size: db_mtx * dbuf_dirty: db_mtx * dbuf_findbp: (callers, phys? - the real need) * dbuf_create: dn_dbufs_mtx, hash_mutexes, db_mtx (phys?) * dbuf_prefetch: dn_dirty_mtx, hash_mutexes, db_mtx, dn_dbufs_mtx * dbuf_hold_impl: hash_mutexes, db_mtx, dn_dbufs_mtx, dbuf_findbp() * dnode_sync/w (increase_indirection): db_mtx (phys) * dnode_set_blksz/w: dn_dbufs_mtx (dn_*blksz*) * dnode_new_blkid/w: (dn_maxblkid) * dnode_free_range/w: dn_dirty_mtx (dn_maxblkid) * dnode_next_offset: (phys) * * dn_dbufs_mtx * must be held before: * db_mtx, hash_mutexes * protects: * dn_dbufs * dn_evicted * held from: * dmu_evict_user: db_mtx (dn_dbufs) * dbuf_free_range: db_mtx (dn_dbufs) * dbuf_remove_ref: db_mtx, callees: * dbuf_hash_remove: hash_mutexes, db_mtx * dbuf_create: hash_mutexes, db_mtx (dn_dbufs) * dnode_set_blksz: (dn_dbufs) * * hash_mutexes (global) * must be held before: * db_mtx * protects dbuf_hash_table (global) and db_hash_next * held from: * dbuf_find: db_mtx * dbuf_hash_insert: db_mtx * dbuf_hash_remove: db_mtx * * db_mtx (meta-leaf) * must be held before: * dn_mtx, dn_dirty_mtx, dd_lock (leaf mutexes) * protects: * db_state * db_holds * db_buf * db_changed * db_data_pending * db_dirtied * db_link * db_dirty_node (??) * db_dirtycnt * db_d.* * db.* * held from: * dbuf_dirty: dn_mtx, dn_dirty_mtx * dbuf_dirty->dsl_dir_willuse_space: dd_lock * dbuf_dirty->dbuf_new_block->dsl_dataset_block_freeable: dd_lock * dbuf_undirty: dn_dirty_mtx (db_d) * dbuf_write_done: dn_dirty_mtx (db_state) * dbuf_* * dmu_buf_update_user: none (db_d) * dmu_evict_user: none (db_d) (maybe can eliminate) * dbuf_find: none (db_holds) * dbuf_hash_insert: none (db_holds) * dmu_buf_read_array_impl: none (db_state, db_changed) * dmu_sync: none (db_dirty_node, db_d) * dnode_reallocate: none (db) * * dn_mtx (leaf) * protects: * dn_dirty_dbufs * dn_ranges * phys accounting * dn_allocated_txg * dn_free_txg * dn_assigned_txg * dd_assigned_tx * dn_notxholds * dn_dirtyctx * dn_dirtyctx_firstset * (dn_phys copy fields?) * (dn_phys contents?) * held from: * dnode_* * dbuf_dirty: none * dbuf_sync: none (phys accounting) * dbuf_undirty: none (dn_ranges, dn_dirty_dbufs) * dbuf_write_done: none (phys accounting) * dmu_object_info_from_dnode: none (accounting) * dmu_tx_commit: none * dmu_tx_hold_object_impl: none * dmu_tx_try_assign: dn_notxholds(cv) * dmu_tx_unassign: none * * dd_lock * must be held before: * ds_lock * ancestors' dd_lock * protects: * dd_prop_cbs * dd_sync_* * dd_used_bytes * dd_tempreserved * dd_space_towrite * dd_myname * dd_phys accounting? * held from: * dsl_dir_* * dsl_prop_changed_notify: none (dd_prop_cbs) * dsl_prop_register: none (dd_prop_cbs) * dsl_prop_unregister: none (dd_prop_cbs) * dsl_dataset_block_freeable: none (dd_sync_*) * * os_lock (leaf) * protects: * os_dirty_dnodes * os_free_dnodes * os_dnodes * os_downgraded_dbufs * dn_dirtyblksz * dn_dirty_link * held from: * dnode_create: none (os_dnodes) * dnode_destroy: none (os_dnodes) * dnode_setdirty: none (dn_dirtyblksz, os_*_dnodes) * dnode_free: none (dn_dirtyblksz, os_*_dnodes) * * ds_lock * protects: * ds_objset * ds_open_refcount * ds_snapname * ds_phys accounting * ds_phys userrefs zapobj * ds_reserved * held from: * dsl_dataset_* * * dr_mtx (leaf) * protects: * dr_children * held from: * dbuf_dirty * dbuf_undirty * dbuf_sync_indirect * dnode_new_blkid */ struct objset; struct dmu_pool; typedef struct dmu_xuio { int next; int cnt; struct arc_buf **bufs; iovec_t *iovp; } dmu_xuio_t; typedef struct xuio_stats { /* loaned yet not returned arc_buf */ kstat_named_t xuiostat_onloan_rbuf; kstat_named_t xuiostat_onloan_wbuf; /* whether a copy is made when loaning out a read buffer */ kstat_named_t xuiostat_rbuf_copied; kstat_named_t xuiostat_rbuf_nocopy; /* whether a copy is made when assigning a write buffer */ kstat_named_t xuiostat_wbuf_copied; kstat_named_t xuiostat_wbuf_nocopy; } xuio_stats_t; static xuio_stats_t xuio_stats = { { "onloan_read_buf", KSTAT_DATA_UINT64 }, { "onloan_write_buf", KSTAT_DATA_UINT64 }, { "read_buf_copied", KSTAT_DATA_UINT64 }, { "read_buf_nocopy", KSTAT_DATA_UINT64 }, { "write_buf_copied", KSTAT_DATA_UINT64 }, { "write_buf_nocopy", KSTAT_DATA_UINT64 } }; #define XUIOSTAT_INCR(stat, val) \ atomic_add_64(&xuio_stats.stat.value.ui64, (val)) #define XUIOSTAT_BUMP(stat) XUIOSTAT_INCR(stat, 1) /* * The list of data whose inclusion in a send stream can be pending from * one call to backup_cb to another. Multiple calls to dump_free() and * dump_freeobjects() can be aggregated into a single DRR_FREE or * DRR_FREEOBJECTS replay record. */ typedef enum { PENDING_NONE, PENDING_FREE, PENDING_FREEOBJECTS } dmu_pendop_t; typedef struct dmu_sendarg { list_node_t dsa_link; dmu_replay_record_t *dsa_drr; kthread_t *dsa_td; struct file *dsa_fp; int dsa_outfd; struct proc *dsa_proc; offset_t *dsa_off; objset_t *dsa_os; zio_cksum_t dsa_zc; uint64_t dsa_toguid; int dsa_err; dmu_pendop_t dsa_pending_op; - boolean_t dsa_incremental; uint64_t dsa_featureflags; uint64_t dsa_last_data_object; uint64_t dsa_last_data_offset; uint64_t dsa_resume_object; uint64_t dsa_resume_offset; } dmu_sendarg_t; void dmu_object_zapify(objset_t *, uint64_t, dmu_object_type_t, dmu_tx_t *); void dmu_object_free_zapified(objset_t *, uint64_t, dmu_tx_t *); int dmu_buf_hold_noread(objset_t *, uint64_t, uint64_t, void *, dmu_buf_t **); #ifdef __cplusplus } #endif #endif /* _SYS_DMU_IMPL_H */ Index: stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zfs_ioctl.h =================================================================== --- stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zfs_ioctl.h (revision 297101) +++ stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zfs_ioctl.h (revision 297102) @@ -1,435 +1,445 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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) 2005, 2010, Oracle and/or its affiliates. All rights reserved. - * Copyright (c) 2012, 2014 by Delphix. All rights reserved. + * Copyright (c) 2012, 2015 by Delphix. All rights reserved. */ #ifndef _SYS_ZFS_IOCTL_H #define _SYS_ZFS_IOCTL_H #include #include #include #include #include #include #ifdef _KERNEL #include #endif /* _KERNEL */ #ifdef __cplusplus extern "C" { #endif /* * The structures in this file are passed between userland and the * kernel. Userland may be running a 32-bit process, while the kernel * is 64-bit. Therefore, these structures need to compile the same in * 32-bit and 64-bit. This means not using type "long", and adding * explicit padding so that the 32-bit structure will not be packed more * tightly than the 64-bit structure (which requires 64-bit alignment). */ /* * Property values for snapdir */ #define ZFS_SNAPDIR_HIDDEN 0 #define ZFS_SNAPDIR_VISIBLE 1 /* * Field manipulation macros for the drr_versioninfo field of the * send stream header. */ /* * Header types for zfs send streams. */ typedef enum drr_headertype { DMU_SUBSTREAM = 0x1, DMU_COMPOUNDSTREAM = 0x2 } drr_headertype_t; #define DMU_GET_STREAM_HDRTYPE(vi) BF64_GET((vi), 0, 2) #define DMU_SET_STREAM_HDRTYPE(vi, x) BF64_SET((vi), 0, 2, x) #define DMU_GET_FEATUREFLAGS(vi) BF64_GET((vi), 2, 30) #define DMU_SET_FEATUREFLAGS(vi, x) BF64_SET((vi), 2, 30, x) /* * Feature flags for zfs send streams (flags in drr_versioninfo) */ #define DMU_BACKUP_FEATURE_DEDUP (1 << 0) #define DMU_BACKUP_FEATURE_DEDUPPROPS (1 << 1) #define DMU_BACKUP_FEATURE_SA_SPILL (1 << 2) /* flags #3 - #15 are reserved for incompatible closed-source implementations */ #define DMU_BACKUP_FEATURE_EMBED_DATA (1 << 16) #define DMU_BACKUP_FEATURE_EMBED_DATA_LZ4 (1 << 17) /* flag #18 is reserved for a Delphix feature */ #define DMU_BACKUP_FEATURE_LARGE_BLOCKS (1 << 19) #define DMU_BACKUP_FEATURE_RESUMING (1 << 20) /* * Mask of all supported backup features */ #define DMU_BACKUP_FEATURE_MASK (DMU_BACKUP_FEATURE_DEDUP | \ DMU_BACKUP_FEATURE_DEDUPPROPS | DMU_BACKUP_FEATURE_SA_SPILL | \ DMU_BACKUP_FEATURE_EMBED_DATA | DMU_BACKUP_FEATURE_EMBED_DATA_LZ4 | \ DMU_BACKUP_FEATURE_RESUMING | \ DMU_BACKUP_FEATURE_LARGE_BLOCKS) /* Are all features in the given flag word currently supported? */ #define DMU_STREAM_SUPPORTED(x) (!((x) & ~DMU_BACKUP_FEATURE_MASK)) typedef enum dmu_send_resume_token_version { ZFS_SEND_RESUME_TOKEN_VERSION = 1 } dmu_send_resume_token_version_t; /* * The drr_versioninfo field of the dmu_replay_record has the * following layout: * * 64 56 48 40 32 24 16 8 0 * +-------+-------+-------+-------+-------+-------+-------+-------+ * | reserved | feature-flags |C|S| * +-------+-------+-------+-------+-------+-------+-------+-------+ * * The low order two bits indicate the header type: SUBSTREAM (0x1) * or COMPOUNDSTREAM (0x2). Using two bits for this is historical: * this field used to be a version number, where the two version types * were 1 and 2. Using two bits for this allows earlier versions of * the code to be able to recognize send streams that don't use any * of the features indicated by feature flags. */ #define DMU_BACKUP_MAGIC 0x2F5bacbacULL #define DRR_FLAG_CLONE (1<<0) #define DRR_FLAG_CI_DATA (1<<1) +/* + * This send stream, if it is a full send, includes the FREE and FREEOBJECT + * records that are created by the sending process. This means that the send + * stream can be received as a clone, even though it is not an incremental. + * This is not implemented as a feature flag, because the receiving side does + * not need to have implemented it to receive this stream; it is fully backwards + * compatible. We need a flag, though, because full send streams without it + * cannot necessarily be received as a clone correctly. + */ +#define DRR_FLAG_FREERECORDS (1<<2) /* * flags in the drr_checksumflags field in the DRR_WRITE and * DRR_WRITE_BYREF blocks */ #define DRR_CHECKSUM_DEDUP (1<<0) #define DRR_IS_DEDUP_CAPABLE(flags) ((flags) & DRR_CHECKSUM_DEDUP) /* * zfs ioctl command structure */ struct drr_begin { uint64_t drr_magic; uint64_t drr_versioninfo; /* was drr_version */ uint64_t drr_creation_time; dmu_objset_type_t drr_type; uint32_t drr_flags; uint64_t drr_toguid; uint64_t drr_fromguid; char drr_toname[MAXNAMELEN]; }; struct drr_end { zio_cksum_t drr_checksum; uint64_t drr_toguid; }; struct drr_object { uint64_t drr_object; dmu_object_type_t drr_type; dmu_object_type_t drr_bonustype; uint32_t drr_blksz; uint32_t drr_bonuslen; uint8_t drr_checksumtype; uint8_t drr_compress; uint8_t drr_pad[6]; uint64_t drr_toguid; /* bonus content follows */ }; struct drr_freeobjects { uint64_t drr_firstobj; uint64_t drr_numobjs; uint64_t drr_toguid; }; struct drr_write { uint64_t drr_object; dmu_object_type_t drr_type; uint32_t drr_pad; uint64_t drr_offset; uint64_t drr_length; uint64_t drr_toguid; uint8_t drr_checksumtype; uint8_t drr_checksumflags; uint8_t drr_pad2[6]; ddt_key_t drr_key; /* deduplication key */ /* content follows */ }; struct drr_free { uint64_t drr_object; uint64_t drr_offset; uint64_t drr_length; uint64_t drr_toguid; }; struct drr_write_byref { /* where to put the data */ uint64_t drr_object; uint64_t drr_offset; uint64_t drr_length; uint64_t drr_toguid; /* where to find the prior copy of the data */ uint64_t drr_refguid; uint64_t drr_refobject; uint64_t drr_refoffset; /* properties of the data */ uint8_t drr_checksumtype; uint8_t drr_checksumflags; uint8_t drr_pad2[6]; ddt_key_t drr_key; /* deduplication key */ }; struct drr_spill { uint64_t drr_object; uint64_t drr_length; uint64_t drr_toguid; uint64_t drr_pad[4]; /* needed for crypto */ /* spill data follows */ }; typedef struct dmu_replay_record { enum { DRR_BEGIN, DRR_OBJECT, DRR_FREEOBJECTS, DRR_WRITE, DRR_FREE, DRR_END, DRR_WRITE_BYREF, DRR_SPILL, DRR_WRITE_EMBEDDED, DRR_NUMTYPES } drr_type; uint32_t drr_payloadlen; union { struct drr_begin drr_begin; struct drr_end drr_end; struct drr_object drr_object; struct drr_freeobjects drr_freeobjects; struct drr_write drr_write; struct drr_free drr_free; struct drr_write_byref drr_write_byref; struct drr_spill drr_spill; struct drr_write_embedded { uint64_t drr_object; uint64_t drr_offset; /* logical length, should equal blocksize */ uint64_t drr_length; uint64_t drr_toguid; uint8_t drr_compression; uint8_t drr_etype; uint8_t drr_pad[6]; uint32_t drr_lsize; /* uncompressed size of payload */ uint32_t drr_psize; /* compr. (real) size of payload */ /* (possibly compressed) content follows */ } drr_write_embedded; /* * Nore: drr_checksum is overlaid with all record types * except DRR_BEGIN. Therefore its (non-pad) members * must not overlap with members from the other structs. * We accomplish this by putting its members at the very * end of the struct. */ struct drr_checksum { uint64_t drr_pad[34]; /* * fletcher-4 checksum of everything preceding the * checksum. */ zio_cksum_t drr_checksum; } drr_checksum; } drr_u; } dmu_replay_record_t; /* diff record range types */ typedef enum diff_type { DDR_NONE = 0x1, DDR_INUSE = 0x2, DDR_FREE = 0x4 } diff_type_t; /* * The diff reports back ranges of free or in-use objects. */ typedef struct dmu_diff_record { uint64_t ddr_type; uint64_t ddr_first; uint64_t ddr_last; } dmu_diff_record_t; typedef struct zinject_record { uint64_t zi_objset; uint64_t zi_object; uint64_t zi_start; uint64_t zi_end; uint64_t zi_guid; uint32_t zi_level; uint32_t zi_error; uint64_t zi_type; uint32_t zi_freq; uint32_t zi_failfast; char zi_func[MAXNAMELEN]; uint32_t zi_iotype; int32_t zi_duration; uint64_t zi_timer; uint32_t zi_cmd; uint32_t zi_pad; } zinject_record_t; #define ZINJECT_NULL 0x1 #define ZINJECT_FLUSH_ARC 0x2 #define ZINJECT_UNLOAD_SPA 0x4 typedef enum zinject_type { ZINJECT_UNINITIALIZED, ZINJECT_DATA_FAULT, ZINJECT_DEVICE_FAULT, ZINJECT_LABEL_FAULT, ZINJECT_IGNORED_WRITES, ZINJECT_PANIC, ZINJECT_DELAY_IO, } zinject_type_t; typedef struct zfs_share { uint64_t z_exportdata; uint64_t z_sharedata; uint64_t z_sharetype; /* 0 = share, 1 = unshare */ uint64_t z_sharemax; /* max length of share string */ } zfs_share_t; /* * ZFS file systems may behave the usual, POSIX-compliant way, where * name lookups are case-sensitive. They may also be set up so that * all the name lookups are case-insensitive, or so that only some * lookups, the ones that set an FIGNORECASE flag, are case-insensitive. */ typedef enum zfs_case { ZFS_CASE_SENSITIVE, ZFS_CASE_INSENSITIVE, ZFS_CASE_MIXED } zfs_case_t; typedef struct zfs_cmd { char zc_name[MAXPATHLEN]; /* name of pool or dataset */ uint64_t zc_nvlist_src; /* really (char *) */ uint64_t zc_nvlist_src_size; uint64_t zc_nvlist_dst; /* really (char *) */ uint64_t zc_nvlist_dst_size; boolean_t zc_nvlist_dst_filled; /* put an nvlist in dst? */ int zc_pad2; /* * The following members are for legacy ioctls which haven't been * converted to the new method. */ uint64_t zc_history; /* really (char *) */ char zc_value[MAXPATHLEN * 2]; char zc_string[MAXNAMELEN]; uint64_t zc_guid; uint64_t zc_nvlist_conf; /* really (char *) */ uint64_t zc_nvlist_conf_size; uint64_t zc_cookie; uint64_t zc_objset_type; uint64_t zc_perm_action; uint64_t zc_history_len; uint64_t zc_history_offset; uint64_t zc_obj; uint64_t zc_iflags; /* internal to zfs(7fs) */ zfs_share_t zc_share; uint64_t zc_jailid; dmu_objset_stats_t zc_objset_stats; dmu_replay_record_t zc_begin_record; zinject_record_t zc_inject_record; uint32_t zc_defer_destroy; uint32_t zc_flags; uint64_t zc_action_handle; int zc_cleanup_fd; uint8_t zc_simple; boolean_t zc_resumable; uint64_t zc_sendobj; uint64_t zc_fromobj; uint64_t zc_createtxg; zfs_stat_t zc_stat; } zfs_cmd_t; typedef struct zfs_useracct { char zu_domain[256]; uid_t zu_rid; uint32_t zu_pad; uint64_t zu_space; } zfs_useracct_t; #define ZFSDEV_MAX_MINOR (1 << 16) #define ZFS_MIN_MINOR (ZFSDEV_MAX_MINOR + 1) #define ZPOOL_EXPORT_AFTER_SPLIT 0x1 #ifdef _KERNEL typedef struct zfs_creat { nvlist_t *zct_zplprops; nvlist_t *zct_props; } zfs_creat_t; extern int zfs_secpolicy_snapshot_perms(const char *name, cred_t *cr); extern int zfs_secpolicy_rename_perms(const char *from, const char *to, cred_t *cr); extern int zfs_secpolicy_destroy_perms(const char *name, cred_t *cr); extern int zfs_busy(void); extern int zfs_unmount_snap(const char *); extern void zfs_destroy_unmount_origin(const char *); /* * ZFS minor numbers can refer to either a control device instance or * a zvol. Depending on the value of zss_type, zss_data points to either * a zvol_state_t or a zfs_onexit_t. */ enum zfs_soft_state_type { ZSST_ZVOL, ZSST_CTLDEV }; typedef struct zfs_soft_state { enum zfs_soft_state_type zss_type; void *zss_data; } zfs_soft_state_t; extern void *zfsdev_get_soft_state(minor_t minor, enum zfs_soft_state_type which); extern minor_t zfsdev_minor_alloc(void); extern void *zfsdev_state; #endif /* _KERNEL */ #ifdef __cplusplus } #endif #endif /* _SYS_ZFS_IOCTL_H */ Index: stable/10 =================================================================== --- stable/10 (revision 297101) +++ stable/10 (revision 297102) Property changes on: stable/10 ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,1 ## Merged /head:r294815