Since r323578 we may remove the last reference to a covered vnode with
vrele() instead of vput(). So, v_usecount may be decremented before
the vnode is locked and zfsctl_snapdir_lookup may "catch" the vnode
with v_usecount of zero and v_holdcnt of one.
PR: 225795
Reported by: asomers
MFC after: 1 week
ZFS caches blocks it reads in its ARC, so in general the optional
pages are not as useful as with filesystems that read the data
directly into the target pages. But still the optional pages
are useful to reduce the number of page faults and associated
VM / VFS / ZFS calls.
Another case that gets optimized (as a side effect) is paging in
from a hole. ZFS DMU does not currently provide a convenient
API to check for a hole. Instead it creates a temporary zero-filled
block and allows accessing it as if it were a normal data block.
Getting multiple pages one by one from a hole results in repeated
creation and destruction of the temporary block (and an associated
ARC header).
Tested with fsx using various supported blocks sizes from 512 bytes
to 128 KB and additionally 1 MB.
Please note that in illumos and ZoL they do not do the range-locking in
the page-in path. This is because ZFS has a double-caching problem
between ARC and page cache and that requires zfs_read() and zfs_write()
to consult pages in the page cache. So, in those functions they first
lock a range and then lock pages corresponding to the range. While in
the page-in (and maybe page-out) path they first lock the pages and then
would lock the range. So, they would have a deadlock.
I believe that FreeBSD does not have that problem, because the page-in
deals only with invalid pages while zfs_read() and zfs_write() need to
access only valid pages. They do not wait on a busy page unless it's
already valid.
Reviewed by: kib
MFC after: 3 weeks
Differential Revision: https://reviews.freebsd.org/D14263
illumos/illumos-gate@d6e1c446d7d6e1c446d7https://www.illumos.org/issues/8857
I had an OS panic on one of our servers:
ffffff01809128c0 vpanic()
ffffff01809128e0 mutex_panic+0x58(fffffffffb94c904, ffffff597dde7f80)
ffffff0180912950 mutex_vector_enter+0x347(ffffff597dde7f80)
ffffff01809129b0 zio_remove_child+0x50(ffffff597dde7c58, ffffff32bd901ac0,
ffffff3373370908)
ffffff0180912a40 zio_done+0x390(ffffff32bd901ac0)
ffffff0180912a70 zio_execute+0x78(ffffff32bd901ac0)
ffffff0180912b30 taskq_thread+0x2d0(ffffff33bae44140)
ffffff0180912b40 thread_start+8()
It panicked here:
http://src.illumos.org/source/xref/illumos-gate/usr/src/uts/common/fs/zfs/
zio.c#430
pio->io_lock is DEAD, thus a panic. Further analysis shows the "pio"
(parent zio of "cio") has already been destroyed.
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Andriy Gapon <avg@FreeBSD.org>
Reviewed by: Youzhong Yang <youzhong@gmail.com>
Approved by: Dan McDonald <danmcd@omniti.com>
Author: George Wilson <george.wilson@delphix.com>
PR: 223803
Tested by: shiva.bhanujan@quorum.com
MFC after: 2 weeks
zfsctl_common_pathconf will report all the same variables that regular ZFS
volumes report. zfsctl_common_getacl will report an ACL equivalent to 555,
except that you can't read xattrs or edit attributes.
Fixes a bug where "ls .zfs" will occasionally print something like:
ls: .zfs/.: Operation not supported
PR: 225793
Reviewed by: avg
MFC after: 3 weeks
Sponsored by: Spectra Logic Corp
Differential Revision: https://reviews.freebsd.org/D14365
global to per-domain state. Protect reservations with the free lock
from the domain that they belong to. Refactor to make vm domains more
of a first class object.
Reviewed by: markj, kib, gallatin
Tested by: pho
Sponsored by: Netflix, Dell/EMC Isilon
Differential Revision: https://reviews.freebsd.org/D14000
We did that in the case of success to prevent the use of stale cached
data, but it makes even less sense to keep the cached data when we fail.
Ideally, we should call vgone() on the vnode in the case of zfs_rezget
failure, but the current lock order prevents us from doing that.
The change also rearranges the order of unlinked check and the size
change check.
While there, add missing SET_ERROR in one of the error paths.
MFC after: 2 weeks
illumos/illumos-gate@5cb8d943bchttps://www.illumos.org/issues/8835:
Sequential reads not aligned to block size are not detected by ZFS
prefetcher as sequential, killing prefetch and severely hurting
performance. It is caused by dmu_zfetch() in case of misaligned
sequential accesses being called with overlap of one block.
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Allan Jude <allanjude@freebsd.org>
Approved by: Gordon Ross <gwr@nexenta.com>
Author: Alexander Motin <mav@FreeBSD.org>
illumos/illumos-gate@4ae5f5f06chttps://www.illumos.org/issues/8652:
Clang and GCC prefer to use unsigned ints to store enums. With Clang, that
causes tautological comparison warnings when comparing a zfs_prop_t or
zpool_prop_t variable to the macro ZPROP_INVAL. It's likely that error
handling code is being silently removed as a result.
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Igor Kozhukhov <igor@dilos.org>
Approved by: Gordon Ross <gwr@nexenta.com>
Author: Alan Somers <asomers@gmail.com>
illumos/illumos-gate@301fd1d6f2
Reviewed by: Alek Pinchuk <pinchuk.alek@gmail.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Approved by: Gordon Ross <gwr@nexenta.com>
Author: Sean Eric Fagan <sef@ixsystems.com>
illumos/illumos-gate@01a059ee0chttps://www.illumos.org/issues/8856:
arc_cksum_is_equal() calls zio_push_transform() that requires abd_t*
(second arg), but a void* is passed.
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Approved by: Gordon Ross <gwr@nexenta.com>
Author: Roman Strashkin <roman.strashkin@nexenta.com>
8930 zfs_zinactive: do not remove the node if the filesystem is readonly
illumos/illumos-gate@93c618e0f4https://www.illumos.org/issues/8930:
We normally remove an unlinked node when its last user goes away and the
node becomes inactive. However, we should not do that if the filesystem
is mounted read-only including the case where it has its readonly
property set. The node will remain on the unlinked queue, so it will
not be leaked.
One particular scenario is when we receive an incremental stream into a
mounted read-only filesystem and that stream contains an unlinked file
(still on the unlinked queue). If that file is opened before the
receive and some time later after the receive it becomes inactive we
would remove it and, thus, modify the read-only filesystem. As a
result, the filesystem would diverge from its source and further
incremental receives would not be possible (without forcing a rollback).
Another related scenario, that may or may not be possible depending on an
OS / VFS policy, is when an open file is unlinked, then the filesystem is
remounted read-only, and then the file is closed.
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Approved by: Gordon Ross <gwr@nexenta.com>
Author: Andriy Gapon <avg@FreeBSD.org>
illumos/illumos-gate@94ddd0900ahttps://www.illumos.org/issues/8909:
There's a race condition that exists if `zil_free_lwb` races with either
`zil_commit_waiter_timeout` and/or `zil_lwb_flush_vdevs_done`.
Here's an example panic due to this bug:
> ::status
debugging crash dump vmcore.0 (64-bit) from ip-10-110-205-40
operating system: 5.11 dlpx-5.2.2.0_2017-12-04-17-28-32b6ba51fb (i86pc)
image uuid: 4af0edfb-e58e-6ed8-cafc-d3e9167c7513
panic message:
BAD TRAP: type=e (#pf Page fault) rp=ffffff0010555970 addr=60 occurred in mo
dule "zfs" due to a NULL pointer dereference
dump content: kernel pages only
> $c
zio_shrink+0x12()
zil_lwb_write_issue+0x30d(ffffff03dcd15cc0, ffffff03e0730e20)
zil_commit_waiter_timeout+0xa2(ffffff03dcd15cc0, ffffff03d97ffcf8)
zil_commit_waiter+0xf3(ffffff03dcd15cc0, ffffff03d97ffcf8)
zil_commit+0x80(ffffff03dcd15cc0, 9a9)
zfs_write+0xc34(ffffff03dc38b140, ffffff0010555e60, 40, ffffff03e00fb758, 0)
fop_write+0x5b(ffffff03dc38b140, ffffff0010555e60, 40, ffffff03e00fb758, 0)
write+0x250(42, fffffd7ff4832000, 2000)
sys_syscall+0x177()
If there's an outstanding lwb that's in `zil_commit_waiter_timeout`
waiting to timeout, waiting on it's waiter's CV, we must be sure not to
call `zil_free_lwb`. If we end up calling `zil_free_lwb`, then that LWB
may be freed and can result in a use-after-free situation where the
stale lwb pointer stored in the `zil_commit_waiter_t` structure of the
thread waiting on the waiter's CV is used.
A similar situation can occur if an lwb is issued to disk, and thus in
the `LWB_STATE_ISSUED` state, and `zil_free_lwb` is called while the
disk is servicing that lwb. In this situation, the lwb will be freed by
`zil_free_lwb`, which will result in a use-after-free situation when the
lwb's zio completes, and `zil_lwb_flush_vdevs_done` is called.
This race condition is prevented in `zil_close` by calling `zil_commit`
before `zil_free_lwb` is called, which will ensure all outstanding (i.e.
all lwb's in the `LWB_STATE_OPEN` and/or `LWB_STATE_ISSUED` states)
reach the `LWB_STATE_DONE` state before the lwb's are freed
(`zil_commit` will not return untill all the lwb's are
`LWB_STATE_DONE`).
Further, this race condition is prevented in `zil_sync` by only calling
`zil_free_lwb` for lwb's that do not have their `lwb_buf` pointer set.
All lwb's not in the `LWB_STATE_DONE` state will have a non-null value
for this pointer; the pointer is only cleared in
`zil_lwb_flush_vdevs_done`, at which point the lwb's state will be
changed to `LWB_STATE_DONE`.
This race is present in `zil_suspend`, leading to this bug.
At first glance, it would appear as though this would not be true
because `zil_suspend` will call `zil_commit`, just like `zil_close`, but
the problem is that `zil_suspend` will set the zilog's `zl_suspend`
field prior to calling `zil_commit`. Further, in `zil_commit`, if
`zl_suspend` is set, `zil_commit` will take a special branch of logic
and use `txg_wait_synced` instead of performing the normal `zil_commit`
logic.
This call to `txg_wait_synced` might be good enough for the data to
reach disk safely before it returns, but it does not ensure that all
outstanding lwb's reach the `LWB_STATE_DONE` state before it returns.
This is because, if there's an lwb "stuck" in
`zil_commit_waiter_timeout`, waiting for it's lwb to timeout, it will
maintain a non-null value for it's `lwb_buf` field and thus `zil_sync`
will not free that lwb. Thus, even though the lwb's data is already on
disk, the lwb will be left lingering, waiting on the CV, and will
eventually timeout and be issued to disk even though the write is
unnesseary.
So, after `zil_commit` is called from `zil_suspend`, we incorrectly
assume that there are not outstanding lwb's, and proceed to free all
lwb's found on the zilog's lwb list. As a result, we free the lwb that
will later be used `zil_commit_waiter_timeout`.
Reviewed by: John Kennedy <jwk404@gmail.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Brad Lewis <brad.lewis@delphix.com>
Reviewed by: Igor Kozhukhov <igor@dilos.org>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: Prakash Surya <prakash.surya@delphix.com>
illumos/illumos-gate@cf07d3da99https://www.illumos.org/issues/8603:
To help make the ZIL's code more understandable, it was suggested that
the zilog_t's "zl_writer_lock" field should be renamed to "zl_issuer_lock".
Reviewed by: C Fraire <cfraire@me.com>
Approved by: Dan McDonald <danmcd@joyent.com>
Author: Prakash Surya <prakash.surya@delphix.com>
illumos/illumos-gate@a3b2868063https://www.illumos.org/issues/8677
We want to be able to run channel programs outside of synching context.
This would greatly improve performance of channel program that just gather
information, as we won't have to wait for synching context anymore.
This feature should introduce the following:
- A new command line flag in "zfs program" to specify our intention to
run in open context.
- A new flag/option within the channel program ioctl which selects the
context.
- Appropriate error handling whenever we try a channel program in
open-context that contains zfs.sync* expressions.
- Documentation for the new feature in the manual pages.
Reviewed by: Matt Ahrens <mahrens@delphix.com>
Reviewed by: Chris Williamson <chris.williamson@delphix.com>
Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: Serapheim Dimitropoulos <serapheim@delphix.com>
Nowadays we do not pass zfs_cmd_t directly through the ioctl interface.
Instead a small zfs_iocparm_t object is passed and the command is
explicitly copied in and out. So, the check has become irrelevant.
MFC after: 3 weeks
Sponsored by: Panzura
These return the jail ID and jail name for the traced process,
respectively, and are analogous to "zonename" on Solaris/illumos.
"zonename" is now aliased to "jailname".
Also add some stress tests for the new variables.
Submitted by: Domagoj Stolfa <domagoj.stolfa@gmail.com>
Reviewed by: dteske (previous version)
MFC after: 2 weeks
Sponsored by: DARPA, AFRL
Differential Revision: https://reviews.freebsd.org/D13877
rather than kmem arena size to determine available memory.
Initialize the UMA limit to LONG_MAX to avoid spurious wakeups on boot before
the real limit is set.
PR: 224330 (partial), 224080
Reviewed by: markj, avg
Sponsored by: Netflix / Dell EMC Isilon
Differential Revision: https://reviews.freebsd.org/D13494
r326987 enabled two #if 0'd-out EMLINK checks in zfs_link_create() for
link overflow. However, one of the checks (when the vnode adding a link
is a directory such as for mkdir) always returned even if the link did not
overflow. Change this to only return early if it needs to report an
EMLINK error.
Reported by: db, shurd
Sponsored by: Chelsio Communications
On the one hand, FIFOs should respect other variables not supported by
the fifofs vnode operation (such as _PC_NAME_MAX, _PC_LINK_MAX, etc.).
These values are fs-specific and must come from a fs-specific method.
On the other hand, filesystems that support FIFOs are required to
support _PC_PIPE_BUF on directory vnodes that can contain FIFOs.
Given this latter requirement, once the fs-specific VOP_PATHCONF
method supports _PC_PIPE_BUF for directories, it is also suitable for
FIFOs permitting a single VOP_PATHCONF method to be used for both
FIFOs and non-FIFOs.
To that end, retire all of the FIFO-specific pathconf methods from
filesystems and change FIFO-specific vnode operation switches to use
the existing fs-specific VOP_PATHCONF method. For fifofs, set it's
VOP_PATHCONF to VOP_PANIC since it should no longer be used.
While here, move _PC_PIPE_BUF handling out of vop_stdpathconf() so that
only filesystems supporting FIFOs will report a value. In addition,
only report a valid _PC_PIPE_BUF for directories and FIFOs.
Discussed with: bde
Reviewed by: kib (part of a larger patch)
MFC after: 1 month
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D12572
Having all filesystems fall through to default values isn't always correct
and these values can vary for different filesystem implementations. Most
of these changes just use the existing default values with a few exceptions:
- Don't report CHOWN_RESTRICTED for ZFS since it doesn't do the exact
permissions check this claims for chown().
- Use NANDFS_NAME_LEN for NAME_MAX for nandfs.
- Don't report a LINK_MAX of 0 on smbfs. Now fail with EINVAL to
indicate hard links aren't supported.
Requested by: bde (though perhaps not this exact implementation)
Reviewed by: kib (earlier version)
MFC after: 1 month
Sponsored by: Chelsio Communications
- Define a ZFS_LINK_MAX as the ZFS version of LINK_MAX which is set to
UINT64_MAX to match the on-disk format.
- Enable the currently #if 0'd code to check for link overflows and
return EMLINK.
- Don't clamp the link count reported in stat() to LINK_MAX as that is
still the 16-bit limit, but report the full link counts. Also,
avoid possibly overflowing the reported link count to 0 when adjusting
the link count to account for ".snapshot".
- Update the LINK_MAX reported by pathconf() to report ZFS_LINK_MAX
rather than LINK_MAX (but clamped to LONG_MAX for 32-bit systems).
Reviewed by: avg (earlier version)
Sponsored by: Chelsio Communications
Otherwise a poorly timed lowmem event may attempt to acquire a destroyed
lock. Unregister the handler before destroying the ARC reclaim thread.
Reported by: gjb
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D13480
The DTrace fasttrap entry points expect a struct reg containing the
register values of the calling thread. Perform the conversion in
fasttrap rather than in the trap handler: this reduces the number of
ifdefs and avoids wasting stack space for traps that don't involve
DTrace.
MFC after: 2 weeks
"panic: vdev_geom_close_locked: cp->private is NULL"
This panic will result if ZFS fails to open a device due to either of the
following reasons:
1) The device's sector size is greater than 8KB.
2) ZFS wants to open the device RW, but it can't be opened for writing.
The solution is to change the initialization order to ensure that the
assertion will be satisfied.
PR: 221066
Reported by: David NewHamlet <wheelcomplex@gmail.com>
Reviewed by: avg
MFC after: 3 weeks
Sponsored by: Spectra Logic Corp
Differential Revision: https://reviews.freebsd.org/D13278
"panic: vdev_geom_close_locked: cp->private is NULL"
This panic will result if ZFS fails to open a device due to either of the
following reasons:
1) The device's sector size is greater than 8KB.
2) ZFS wants to open the device RW, but it can't be opened for writing.
The solution is to change the initialization order to ensure that the
assertion will be satisfied.
PR: 221066
Reported by: David NewHamlet <wheelcomplex@gmail.com>
Reviewed by: avg
MFC after: 3 weeks
Sponsored by: Spectra Logic Corp
Differential Revision: https://reviews.freebsd.org/D13278
We may create probes in the nascent child process, so we first need to
ensure that any inherited tracepoints are first removed. Otherwise the
probe sites will not be in the state expected by fasttrap, and it won't
be able to enable the probes.
MFC after: 2 weeks
The problem happens when the writes have offsets and sizes aligned with
a filesystem's recordsize (maximum block size). In this scenario
dmu_tx_assign() would fail because of being over the quota, but the uio
would already be modified in the code path where we copy data from the
uio into a borrowed ARC buffer. That makes an appearance of a partial
write, so zfs_write() would return success and the uio would be modified
consistently with writing a single block.
That bug can result in a data loss because the writes over the quota
would appear to succeed while the actual data is being discarded.
This commit fixes the bug by ensuring that the uio is not changed until
after all error checks are done. To achieve that the code now uses
uiocopy() + uioskip() as in the original illumos design. We can do that
now that uiocopy() has been updated in r326067 to use
vn_io_fault_uiomove().
Reported by: mav
Analyzed by: mav
Reviewed by: mav
Pointyhat to: avg (myself)
MFC after: 1 week
X-MFC after: r326067
X-Erratum: wanted
In general, higher-level code will atomically verify that the process
is not exiting and hold the process. In one case, we were using uwrite()
to copy a probed instruction to a per-thread scratch space block, but
copyout() can be used for this purpose instead; this change effectively
reverts r227291.
MFC after: 1 week
the ARC reclaim thread running longer than needed.
Update the arc::needfree dtrace probe triggered in arc_lowmem() to also report
the value we may want to free.
Submitted by: Nikita Kozlov <nikita.kozlov at blade-group.com>
Reviewed by: avg
Approved by: avg
MFC after: 3 weeks
Sponsored by: blade
Differential Revision: https://reviews.freebsd.org/D12163
illumos/illumos-gate@27295216542729521654https://www.illumos.org/issues/7531
I found that some buffers that could be L2ARC eligible are not flagged
such, leading to some performance impact. As a test I ran the same IO
workload 10 times in a raw. It is a metadata only workload (files
listing). l2arc_noprefetch=0.
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Approved by: Dan McDonald <danmcd@joyent.com>
Author: benrubson <ben.rubson@gmail.com>
MFC after: 8 days
illumos/illumos-gate@f37ae9a714f37ae9a714https://www.illumos.org/issues/8713
If we're creating a pool with version >= SPA_VERSION_DSL_SCRUB (v11) we need to
account for additional space needed by the origin dataset which will also be
snapshotted: "poolname"+"/"+"$ORIGIN"+"@"+"$ORIGIN".
Enforce this limit in pool_namecheck().
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Approved by: Dan McDonald <danmcd@joyent.com>
Author: loli10K <ezomori.nozomu@gmail.com>
MFC after: 1 week
The generic (naive) implementation of posix_fallocate cannot provide the
standard mandated guarantee that overwrites would never fail due to the lack
of free space. The fundamental reason is the copy-on-write architecture
of ZFS. Other features like compression and deduplication can also
increase the size difference between the (pre-)allocated dummy content
and the future content.
So, until ZFS can properly implement the feature it's better to report
that it is unsupported rather than providing an ersatz implementation.
Please note that EINVAL is used to report that the underlying file system
does not support the operation (POSIX.1-2008).
illumos and ZoL seem to do the same.
MFC after: 3 weeks
Sponsored by: Panzura
If vdev_geom_close doesn't close the consumer, then the subsequent call
to vdev_geom_open() would be just a NOP and would always return success.
Thus, at present vdev_reopen() would always succeed for vdev_geom devices
even if the underlying provider is in error state.
The problem was introduced as a result of an optimization in rS308055.
The most significant manifistation of the problem is that
zio_vdev_io_done() --> vdev_probe() --> SPA_ASYNC_PROBE -->
spa_async_probe() --> vdev_reopen()
chain of calls and events becomes a NOP as well.
This chain is invoked when zio_vdev_io_done() detects an "unexpected"
error from the lower level I/O.
Additionally, that call path may race with SPA_ASYNC_REMOVE path because
of the asynchronous nature of them both. So, the SPA_ASYNC_PROBE may
erroneously mark a vdev as being healthy after SPA_ASYNC_REMOVE marked
it as removed.
Reviewed by: asomers, mav
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D12731
Don't check for SPA_MINDEVSIZE in vdev_geom_attach when opening by path.
It's redundant with the check in vdev_open, and failing to attach here
results in the wrong error message being printed. However, still check for
it in some other situations:
* When opening by guids, so we don't get bogged down reading from slow
devices like floppy drives.
* In vdev_geom_read_pool_label for the same reason, because we iterate over
all providers.
* If the caller requests that we verify the guid, because then we'll have to
read from the device before vdev_open verifies the size.
PR: 222227
Reported by: Marie Helene Kvello-Aune <marieheleneka@gmail.com>
Reviewed by: avg, mav
MFC after: 3 weeks
Sponsored by: Spectra Logic Corp
Differential Revision: https://reviews.freebsd.org/D12531
Unlike spa_async_thread that can get started only from spa_sync()
spa_async_thread_vd can get started from other contexts.
Additionally, spa_async_thread_vd does not really depend on
spa sync being enabled.
The incorrect assert could be triggered by importing a pool in the
read-only mode and then disconnecting one of its disks.
In this case spa_sync_on was false because the pool was read-only
and spa_async_thread_vd was started to handle SPA_ASYNC_REMOVE event.
Note: spa_async_thread_vd() currently exists only in FreeBSD, it was
split out of spa_async_thread() in r253990.
Discussed with: mav
MFC after: 2 weeks
illumos/illumos-gate@4923c69fdd4923c69fdd
FreeBSD note: the manual page is to be updated separately.
https://www.illumos.org/issues/8067
Add an option to zdb to print a literal embedded block pointer supplied on the
command line:
zdb -E [-A] word0:word1:...:word15
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Alex Reece <alex@delphix.com>
Reviewed by: Yuri Pankov <yuri.pankov@gmail.com>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: Matthew Ahrens <mahrens@delphix.com>
MFC after: 3 weeks
