MFC r337229: Reduce taskq and context-switch cost of zio pipe

When doing a read from disk, ZFS creates 3 ZIO's: a zio_null(), the
logical zio_read(), and then a physical zio. Currently, each of these
results in a separate taskq_dispatch(zio_execute).

On high-read-iops workloads, this causes a significant performance
impact. By processing all 3 ZIO's in a single taskq entry, we reduce the
overhead on taskq locking and context switching. We accomplish this by
allowing zio_done() to return a "next zio to execute" to zio_execute().

This results in a ~12% performance increase for random reads, from
96,000 iops to 108,000 iops (with recordsize=8k, on SSD's).

Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed by: George Wilson <george.wilson@delphix.com>
Signed-off-by: Matthew Ahrens <mahrens@delphix.com>
External-issue: DLPX-59292
Closes #7736

zfsonlinux/zfs@62840030a7dceaee013ddbcc1eebcfc7922edf7c

MFC r336951: MFV r336950: 9290 device removal reduces redundancy of mirrors

Mirrors are supposed to provide redundancy in the face of whole-disk failure
and silent damage (e.g. some data on disk is not right, but ZFS hasn't
detected the whole device as being broken). However, the current device
removal implementation bypasses some of the mirror's redundancy.

illumos/illumos-gate@3a4b1be953ee5601bab748afa07c26ed4996cde6

Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Prashanth Sreenivasa <pks@delphix.com>
Reviewed by: Sara Hartse <sara.hartse@delphix.com>
Reviewed by: Serapheim Dimitropoulos <serapheim@delphix.com>
Reviewed by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed by: Tim Chase <tim@chase2k.com>
Approved by: Richard Lowe <richlowe@richlowe.net>
Author: Matthew Ahrens <mahrens@delphix.com>

MFC r336949:
MFV r336948: 9112 Improve allocation performance on high-end systems

On high-end systems running async sequential write workloads, especially
NUMA systems with flash or NVMe storage, one significant performance
bottleneck is selecting a metaslab to do allocations from. This process
can be parallelized, providing significant performance increases for
these workloads.

illumos/illumos-gate@f78cdc34af236a6199dd9e21376f4a46348c0d56

Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Serapheim Dimitropoulos <serapheim.dimitro@delphix.com>
Reviewed by: Alexander Motin <mav@FreeBSD.org>
Approved by: Gordon Ross <gwr@nexenta.com>
Author: Paul Dagnelie <pcd@delphix.com>

MFC r334844, r336180, r336458

r334844

This originated from ZFS On Linux, as
https://github.com/zfsonlinux/zfs/commit/d4a72f23863382bdf6d0ae33196f5b5decbc48fd

During scans (scrubs or resilvers), it sorts the blocks in each transaction
group by block offset; the result can be a significant improvement. (On my
test system just now, which I put some effort to introduce fragmentation into
the pool since I set it up yesterday, a scrub went from 1h2m to 33.5m with the
changes.) I've seen similar rations on production systems.

r336180

Fix up some missed and mis-merges from the sequential scan code
(r334844). Most of the changes involve moving some code around to
reduce conflicts with future merges. One of the missing changes
included a notification on scrub cancellation.

r336458

Fix a couple of typos in r334844 noticed by Richard Kojedzinszky

Approved by: mav
Sponsored by: iXsystems, Inc

MFC r331701: MFV r331695, 331700: 9166 zfs storage pool checkpoint

illumos/illumos-gate@8671400134a11c848244896ca51a7db4d0f69da4

The idea of Storage Pool Checkpoint (aka zpool checkpoint) deals with
exactly that. It can be thought of as a “pool-wide snapshot” (or a
variation of extreme rewind that doesn’t corrupt your data). It remembers
the entire state of the pool at the point that it was taken and the user
can revert back to it later or discard it. Its generic use case is an
administrator that is about to perform a set of destructive actions to ZFS
as part of a critical procedure. She takes a checkpoint of the pool before
performing the actions, then rewinds back to it if one of them fails or puts
the pool into an unexpected state. Otherwise, she discards it. With the
assumption that no one else is making modifications to ZFS, she basically
wraps all these actions into a “high-level transaction”.

Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: John Kennedy <john.kennedy@delphix.com>
Reviewed by: Dan Kimmel <dan.kimmel@delphix.com>
Approved by: Richard Lowe <richlowe@richlowe.net>
Author: Serapheim Dimitropoulos <serapheim.dimitro@delphix.com>

MFC r329732: MFV r329502: 7614 zfs device evacuation/removal

illumos/illumos-gate@5cabbc6b49070407fb9610cfe73d4c0e0dea3e77

https://www.illumos.org/issues/7614:
This project allows top-level vdevs to be removed from the storage pool with
“zpool remove”, reducing the total amount of storage in the pool. This
operation copies all allocated regions of the device to be removed onto other
devices, recording the mapping from old to new location. After the removal is
complete, read and free operations to the removed (now “indirect”) vdev must
be remapped and performed at the new location on disk. The indirect mapping
table is kept in memory whenever the pool is loaded, so there is minimal
performance overhead when doing operations on the indirect vdev.

The size of the in-memory mapping table will be reduced when its entries
become “obsolete” because they are no longer used by any block pointers in
the pool. An entry becomes obsolete when all the blocks that use it are
freed. An entry can also become obsolete when all the snapshots that
reference it are deleted, and the block pointers that reference it have been
“remapped” in all filesystems/zvols (and clones). Whenever an indirect block
is written, all the block pointers in it will be “remapped” to their new
(concrete) locations if possible. This process can be accelerated by using
the “zfs remap” command to proactively rewrite all indirect blocks that
reference indirect (removed) vdevs.

Note that when a device is removed, we do not verify the checksum of the data
that is copied. This makes the process much faster, but if it were used on
redundant vdevs (i.e. mirror or raidz vdevs), it would be possible to copy
the wrong data, when we have the correct data on e.g. the other side of the
mirror. Therefore, mirror and raidz devices can not be removed.

Reviewed by: Alex Reece <alex@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: John Kennedy <john.kennedy@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Richard Laager <rlaager@wiktel.com>
Reviewed by: Tim Chase <tim@chase2k.com>
Approved by: Garrett D'Amore <garrett@damore.org>
Author: Prashanth Sreenivasa <pks@delphix.com>

MFC r307317: MFV r307313:
5120 zfs should allow large block/gzip/raidz boot pool (loader project)

Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Yuri Pankov <yuri.pankov@nexenta.com>
Reviewed by: Andrew Stormont <andyjstormont@gmail.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: Toomas Soome <tsoome@me.com>

openzfs/openzfs@c8811bd3e2427dddbac6c05a59cfe117d8fea370

FreeBSD still does not support booting from gzip-compressed datasets,
so keep one chunk of this commit out.

MFC r329314: MFV r329313: 8857 zio_remove_child() panic due to already destroyed parent zio

PR: 223803

MFC r328228: MFV r328227: 8909 8585 can cause a use-after-free kernel panic

illumos/illumos-gate@94ddd0900a8838f62bba15e270649a42f4ef9f81

https://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>

MFC r324011, r324016: MFV r323535: 8585 improve batching done in zil_commit()

FreeBSD notes:
- this MFV reverts FreeBSD commit r314549 to make the merge easier
- at present our emulation of cv_timedwait_hires is rather poor,
so I elected to use cv_timedwait_sbt directly
Please see the differential revision for details.
Unfortunately, I did not get any positive reviews, so there could be
bugs in the FreeBSD-specific piece of the merge.
Hence, the long MFC timeout.

illumos/illumos-gate@1271e4b10dfaaed576c08a812f466f6e81370e5e
https://github.com/illumos/illumos-gate/commit/1271e4b10dfaaed576c08a812f466f6e81370e5e

https://www.illumos.org/issues/8585
The current implementation of zil_commit() can introduce significant
latency, beyond what is inherent due to the latency of the underlying
storage. The additional latency comes from two main problems:
1. When there's outstanding ZIL blocks being written (i.e. there's
already a "writer thread" in progress), then any new calls to
zil_commit() will block waiting for the currently oustanding ZIL
blocks to complete. The blocks written for each "writer thread" is
coined a "batch", and there can only ever be a single "batch" being
written at a time. When a batch is being written, any new ZIL
transactions will have to wait for the next batch to be written,
which won't occur until the current batch finishes.
As a result, the underlying storage may not be used as efficiently
as possible. While "new" threads enter zil_commit() and are blocked
waiting for the next batch, it's possible that the underlying
storage isn't fully utilized by the current batch of ZIL blocks. In
that case, it'd be better to allow these new threads to generate
(and issue) a new ZIL block, such that it could be serviced by the
underlying storage concurrently with the other ZIL blocks that are
being serviced.
2. Any call to zil_commit() must wait for all ZIL blocks in its "batch"
to complete, prior to zil_commit() returning. The size of any given
batch is proportional to the number of ZIL transaction in the queue
at the time that the batch starts processing the queue; which
doesn't occur until the previous batch completes. Thus, if there's a
lot of transactions in the queue, the batch could be composed of
many ZIL blocks, and each call to zil_commit() will have to wait for
all of these writes to complete (even if the thread calling
zil_commit() only cared about one of the transactions in the batch).

Reviewed by: Brad Lewis <brad.lewis@delphix.com>
Reviewed by: Matt Ahrens <mahrens@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Approved by: Dan McDonald <danmcd@joyent.com>
Author: Prakash Surya <prakash.surya@delphix.com>

MFC r320156, r320185, r320186, r320262, r320452, r321111:
MFV r318946: 8021 ARC buf data scatter-ization

illumos/illumos-gate@770499e185d15678ccb0be57ebc626ad18d93383
https://github.com/illumos/illumos-gate/commit/770499e185d15678ccb0be57ebc626ad1
8d93383

https://www.illumos.org/issues/8021
The ARC buf data project (known simply as "ABD" since its genesis in the ZoL
community) changes the way the ARC allocates `b_pdata` memory from using linea
r
`void *` buffers to using scatter/gather lists of fixed-size 1KB chunks. This
improves ZFS's performance by helping to defragment the address space occupied
by the ARC, in particular for cases where compressed ARC is enabled. It could
also ease future work to allocate pages directly from `segkpm` for minimal-
overhead memory allocations, bypassing the `kmem` subsystem.
This is essentially the same change as the one which recently landed in ZFS on
Linux, although they made some platform-specific changes while adapting this
work to their codebase:
1. Implemented the equivalent of the `segkpm` suggestion for future work
mentioned above to bypass issues that they've had with the Linux kernel memory
allocator.
2. Changed the internal representation of the ABD's scatter/gather list so it
could be used to pass I/O directly into Linux block device drivers. (This
feature is not available in the illumos block device interface yet.)

FreeBSD notes:
- the actual (default) chunk size is 4KB (despite the text above saying 1KB)
- we can try to reimplement ABDs, so that they are not permanently
mapped into the KVA unless explicitly requested, especially on
platforms with scarce KVA
- we can try to use unmapped I/O and avoid intermediate allocation of a
linear, virtual memory mapped buffer
- we can try to avoid extra data copying by referring to chunks / pages
in the original ABD

Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Paul Dagnelie <pcd@delphix.com>
Reviewed by: John Kennedy <john.kennedy@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Prashanth Sreenivasa <pks@delphix.com>
Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com>
Reviewed by: Chris Williamson <chris.williamson@delphix.com>
Approved by: Richard Lowe <richlowe@richlowe.net>
Author: Dan Kimmel <dan.kimmel@delphix.com>

MFC r320153: revert r315852 which introduced zio_buf_alloc_nowait for use
in vdev_queue_aggregate

I think that the change is still good, but reconciling it with a planned
merge of the ARC buf data scatter-ization is a bit more tedious
than I can handle.

MFC r318828: MFV r316917: 7968 multi-threaded spa_sync()

illumos/illumos-gate@94c2d0eb22e9624151ee84a7edbf7178e1bf4087
https://github.com/illumos/illumos-gate/commit/94c2d0eb22e9624151ee84a7edbf7178e1bf4087

https://www.illumos.org/issues/7968
spa_sync() iterates over all the dirty dnodes and processes each of them by
calling dnode_sync(). If there are many dirty dnodes (e.g. because we created
or removed a lot of files), the single thread of spa_sync() calling
dnode_sync() can become a bottleneck. Additionally, if many dnodes are dirtied
concurrently in open context (e.g. due to concurrent file creation), the
os_lock will experience lock contention via dnode_setdirty().
The solution is to track dirty dnodes on a multilist_t, and for spa_sync() to
use separate threads to process each of the sublists in the multilist.
On the concurrent file creation microbenchmark, the performance improvement
from dnode_setdirty() is up to 7%. Additionally, the wall clock time spent in
spa_sync() is reduced to 15%-40% of the single-threaded case. In terms of cost/
reward, once the other bottlenecks are addressed, fixing this bug will provide
a medium-large performance gain and require a medium amount of effort to
implement.

Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com>
Reviewed by: Brad Lewis <brad.lewis@delphix.com>
Reviewed by: Saso Kiselkov <saso.kiselkov@nexenta.com>
Reviewed by: Brian Behlendorf <behlendorf1@llnl.gov>
Approved by: Dan McDonald <danmcd@omniti.com>
Author: Matthew Ahrens <mahrens@delphix.com>

MFC r317414: MFV 316894

7252 7628 compressed zfs send / receive

illumos/illumos-gate@5602294fda888d923d57a78bafdaf48ae6223dea
https://github.com/illumos/illumos-gate/commit/5602294fda888d923d57a78bafdaf48ae6223dea

https://www.illumos.org/issues/7252
This feature includes code to allow a system with compressed ARC enabled to
send data in its compressed form straight out of the ARC, and receive data in
its compressed form directly into the ARC.

https://www.illumos.org/issues/7628
We should have longer, more readable versions of the ZFS send / recv options.

7628 create long versions of ZFS send / receive options

Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: John Kennedy <john.kennedy@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Paul Dagnelie <pcd@delphix.com>
Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com>
Reviewed by: Sebastien Roy <sebastien.roy@delphix.com>
Reviewed by: David Quigley <dpquigl@davequigley.com>
Reviewed by: Thomas Caputi <tcaputi@datto.com>
Approved by: Dan McDonald <danmcd@omniti.com>
Author: Dan Kimmel <dan.kimmel@delphix.com>

MFC r315896: MFV r315290, r315291: 7303 dynamic metaslab selection

illumos/illumos-gate@8363e80ae72609660f6090766ca8c2c18aa53f0c
https://github.com/illumos/illumos-gate/commit/8363e80ae72609660f6090766ca8c2c18

https://www.illumos.org/issues/7303

This change introduces a new weighting algorithm to improve metaslab selection
.
The new weighting algorithm relies on the SPACEMAP_HISTOGRAM feature. As a res
ult,
the metaslab weight now encodes the type of weighting algorithm used
(size-based vs segment-based).

This also introduce a new allocation tracing facility and two new dcmds to hel
p
debug allocation problems. Each zio now contains a zio_alloc_list_t structure
that is populated as the zio goes through the allocations stage. Here's an
example of how to use the tracing facility:

> c5ec000::print zio_t io_alloc_list | ::walk list | ::metaslab_trace
MSID DVA ASIZE WEIGHT RESULT VDEV
- 0 400 0 NOT_ALLOCATABLE ztest.0a
- 0 400 0 NOT_ALLOCATABLE ztest.0a
- 0 400 0 ENOSPC ztest.0a
- 0 200 0 NOT_ALLOCATABLE ztest.0a
- 0 200 0 NOT_ALLOCATABLE ztest.0a
- 0 200 0 ENOSPC ztest.0a
1 0 400 1 x 8M 17b1a00 ztest.0a

> 1ff2400::print zio_t io_alloc_list | ::walk list | ::metaslab_trace
MSID DVA ASIZE WEIGHT RESULT VDEV
- 0 200 0 NOT_ALLOCATABLE mirror-2
- 0 200 0 NOT_ALLOCATABLE mirror-0
1 0 200 1 x 4M 112ae00 mirror-1
- 1 200 0 NOT_ALLOCATABLE mirror-2
- 1 200 0 NOT_ALLOCATABLE mirror-0
1 1 200 1 x 4M 112b000 mirror-1
- 2 200 0 NOT_ALLOCATABLE mirror-2

If the metaslab is using segment-based weighting then the WEIGHT column will
display the number of segments available in the bucket where the allocation
attempt was made.

Author: George Wilson <george.wilson@delphix.com>
Reviewed by: Alex Reece <alex@delphix.com>
Reviewed by: Chris Siden <christopher.siden@delphix.com>
Reviewed by: Dan Kimmel <dan.kimmel@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Paul Dagnelie <paul.dagnelie@delphix.com>
Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Don Brady <don.brady@intel.com>
Approved by: Richard Lowe <richlowe@richlowe.net>

MFC r315852: zfs: add zio_buf_alloc_nowait and use it in vdev_queue_aggregate

MFC r308782:
After some ZIL changes 6 years ago zil_slog_limit got partially broken
due to zl_itx_list_sz not updated when async itx'es upgraded to sync.
Actually because of other changes about that time zl_itx_list_sz is not
really required to implement the functionality, so this patch removes
some unneeded broken code and variables.

Original idea of zil_slog_limit was to reduce chance of SLOG abuse by
single heavy logger, that increased latency for other (more latency critical)
loggers, by pushing heavy log out into the main pool instead of SLOG. Beside
huge latency increase for heavy writers, this implementation caused double
write of all data, since the log records were explicitly prepared for SLOG.
Since we now have I/O scheduler, I've found it can be much more efficient
to reduce priority of heavy logger SLOG writes from ZIO_PRIORITY_SYNC_WRITE
to ZIO_PRIORITY_ASYNC_WRITE, while still leave them on SLOG.

Existing ZIL implementation had problem with space efficiency when it
has to write large chunks of data into log blocks of limited size. In some
cases efficiency stopped to almost as low as 50%. In case of ZIL stored on
spinning rust, that also reduced log write speed in half, since head had to
uselessly fly over allocated but not written areas. This change improves
the situation by offloading problematic operations from z*_log_write() to
zil_lwb_commit(), which knows real situation of log blocks allocation and
can split large requests into pieces much more efficiently. Also as side
effect it removes one of two data copy operations done by ZIL code WR_COPIED
case.

While there, untangle and unify code of z*_log_write() functions.
Also zfs_log_write() alike to zvol_log_write() can now handle writes crossing
block boundary, that may also improve efficiency if ZPL is made to do that.

Sponsored by: iXsystems, Inc.

MFC r305331: MFV r304155:
7090 zfs should improve allocation order and throttle allocations

illumos/illumos-gate@0f7643c7376dd69a08acbfc9d1d7d548b10c846a
https://github.com/illumos/illumos-gate/commit/0f7643c7376dd69a08acbfc9d1d7d548b
10c846a

https://www.illumos.org/issues/7090
When write I/Os are issued, they are issued in block order but the ZIO pipelin
e
will drive them asynchronously through the allocation stage which can result i
n
blocks being allocated out-of-order. It would be nice to preserve as much of
the logical order as possible.
In addition, the allocations are equally scattered across all top-level VDEVs
but not all top-level VDEVs are created equally. The pipeline should be able t
o
detect devices that are more capable of handling allocations and should
allocate more blocks to those devices. This allows for dynamic allocation
distribution when devices are imbalanced as fuller devices will tend to be
slower than empty devices.
The change includes a new pool-wide allocation queue which would throttle and
order allocations in the ZIO pipeline. The queue would be ordered by issued
time and offset and would provide an initial amount of allocation of work to
each top-level vdev. The allocation logic utilizes a reservation system to
reserve allocations that will be performed by the allocator. Once an allocatio
n
is successfully completed it's scheduled on a given top-level vdev. Each top-
level vdev maintains a maximum number of allocations that it can handle
(mg_alloc_queue_depth). The pool-wide reserved allocations (top-levels *
mg_alloc_queue_depth) are distributed across the top-level vdevs metaslab
groups and round robin across all eligible metaslab groups to distribute the
work. As top-levels complete their work, they receive additional work from the
pool-wide allocation queue until the allocation queue is emptied.

Reviewed by: Adam Leventhal <ahl@delphix.com>
Reviewed by: Alex Reece <alex@delphix.com>
Reviewed by: Christopher Siden <christopher.siden@delphix.com>
Reviewed by: Dan Kimmel <dan.kimmel@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Paul Dagnelie <paul.dagnelie@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Sebastien Roy <sebastien.roy@delphix.com>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: George Wilson <george.wilson@delphix.com>

MFC r305323: MFV r302991: 6950 ARC should cache compressed data

illumos/illumos-gate@dcbf3bd6a1f1360fc1afcee9e22c6dcff7844bf2
https://github.com/illumos/illumos-gate/commit/dcbf3bd6a1f1360fc1afcee9e22c6dcff7844bf2

https://www.illumos.org/issues/6950
When reading compressed data from disk, the ARC should keep the compressed
block cached and only decompress it when consumers access the block. The
uncompressed data should be short-lived allowing the ARC to cache a much larger
amount of data. The DMU would also maintain a smaller cache of uncompressed
blocks to minimize the impact of decompressing frequently accessed blocks.

Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Dan Kimmel <dan.kimmel@delphix.com>
Reviewed by: Matt Ahrens <mahrens@delphix.com>
Reviewed by: Paul Dagnelie <pcd@delphix.com>
Reviewed by: Don Brady <don.brady@intel.com>
Reviewed by: Richard Elling <Richard.Elling@RichardElling.com>
Approved by: Richard Lowe <richlowe@richlowe.net>
Author: George Wilson <george.wilson@delphix.com>

MFC r302838: 6513 partially filled holes lose birth time