Copy stable/10@r272459 to releng/10.1 as part of
the 10.1-RELEASE process.

Approved by: re (implicit)
Sponsored by: The FreeBSD Foundation

MFC Alexander Motin's GEOM direct dispatch work:

r256603:
Introduce new function devstat_end_transaction_bio_bt(), adding new argument
to specify present time. Use this function to move binuptime() out of lock,
substantially reducing lock congestion when slow timecounter is used.

r256606:
Move g_io_deliver() out of the lock, as required for direct dispatch.
Move g_destroy_bio() out too to reduce lock scope even more.

r256607:
Fix passing uninitialized bio_resid argument to g_trace().

r256610:
Add unmapped I/O support to GEOM RAID.

r256830:
Restore BIO_UNMAPPED and BIO_TRANSIENT_MAPPING in biodonne() when unmapping
temporary mapped buffer. That fixes double unmap if biodone() called twice
for the same BIO (but with different done methods).

r256880:
Merge GEOM direct dispatch changes from the projects/camlock branch.

When safety requirements are met, it allows to avoid passing I/O requests
to GEOM g_up/g_down thread, executing them directly in the caller context.
That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid
several context switches per I/O.

r259247:
Fix bug introduced at r256607. We have to recalculate bp_resid here since
sizes of original and completed requests may differ due to end of media.

Testing of the stable/10 merge was done by Netflix, but all of the credit
goes to Alexander and iX Systems.

Submitted by: mav
Sponsored by: iX Systems

Copy head (r256279) to stable/10 as part of the 10.0-RELEASE cycle.

Approved by: re (implicit)
Sponsored by: The FreeBSD Foundation

- Add a general purpose resource allocator, vmem, from NetBSD. It was
originally inspired by the Solaris vmem detailed in the proceedings
of usenix 2001. The NetBSD version was heavily refactored for bugs
and simplicity.
- Use this resource allocator to allocate the buffer and transient maps.
Buffer cache defrags are reduced by 25% when used by filesystems with
mixed block sizes. Ultimately this may permit dynamic buffer cache
sizing on low KVA machines.

Discussed with: alc, kib, attilio
Tested by: pho
Sponsored by: EMC / Isilon Storage Division

Correct the page count when excess length is trimmed from the bio.

Reported and tested by: Ivan Klymenko <fidaj@ukr.net

Assert that transient mapping of the bio is only done when unmapped
buffers are allowed.

Sponsored by: The FreeBSD Foundation

Implement the concept of the unmapped VMIO buffers, i.e. buffers which
do not map the b_pages pages into buffer_map KVA. The use of the
unmapped buffers eliminate the need to perform TLB shootdown for
mapping on the buffer creation and reuse, greatly reducing the amount
of IPIs for shootdown on big-SMP machines and eliminating up to 25-30%
of the system time on i/o intensive workloads.

The unmapped buffer should be explicitely requested by the GB_UNMAPPED
flag by the consumer. For unmapped buffer, no KVA reservation is
performed at all. The consumer might request unmapped buffer which
does have a KVA reserve, to manually map it without recursing into
buffer cache and blocking, with the GB_KVAALLOC flag.

When the mapped buffer is requested and unmapped buffer already
exists, the cache performs an upgrade, possibly reusing the KVA
reservation.

Unmapped buffer is translated into unmapped bio in g_vfs_strategy().
Unmapped bio carry a pointer to the vm_page_t array, offset and length
instead of the data pointer. The provider which processes the bio
should explicitely specify a readiness to accept unmapped bio,
otherwise g_down geom thread performs the transient upgrade of the bio
request by mapping the pages into the new bio_transient_map KVA
submap.

The bio_transient_map submap claims up to 10% of the buffer map, and
the total buffer_map + bio_transient_map KVA usage stays the
same. Still, it could be manually tuned by kern.bio_transient_maxcnt
tunable, in the units of the transient mappings. Eventually, the
bio_transient_map could be removed after all geom classes and drivers
can accept unmapped i/o requests.

Unmapped support can be turned off by the vfs.unmapped_buf_allowed
tunable, disabling which makes the buffer (or cluster) creation
requests to ignore GB_UNMAPPED and GB_KVAALLOC flags. Unmapped
buffers are only enabled by default on the architectures where
pmap_copy_page() was implemented and tested.

In the rework, filesystem metadata is not the subject to maxbufspace
limit anymore. Since the metadata buffers are always mapped, the
buffers still have to fit into the buffer map, which provides a
reasonable (but practically unreachable) upper bound on it. The
non-metadata buffer allocations, both mapped and unmapped, is
accounted against maxbufspace, as before. Effectively, this means that
the maxbufspace is forced on mapped and unmapped buffers separately.
The pre-patch bufspace limiting code did not worked, because
buffer_map fragmentation does not allow the limit to be reached.

By Jeff Roberson request, the getnewbuf() function was split into
smaller single-purpose functions.

Sponsored by: The FreeBSD Foundation
Discussed with: jeff (previous version)
Tested by: pho, scottl (previous version), jhb, bf
MFC after: 2 weeks

Reset provider-specific fields when resending I/O request in low memory
conditions. This fixes assertion which checks those fields when kernel is
compiled with DIAGNOSTIC.

Reported by: kib, pho
MFC after: 1 week

Clone BIO_ORDERED flag, for disk drivers (namely CAM) that try to
consume it.

Sponsored by: Intel
Discussed with: gibbs, scottl

Implement media change notification for DA and CD removable media devices.
It includes three parts:
1) Modifications to CAM to detect media media changes and report them to
disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes
Asynchronous Notification mechanism to receive events from hardware.
Active polling with TEST UNIT READY commands with 3 seconds period is used
for incapable hardware. After that both CD and DA drivers work the same way,
detecting two conditions: "NOT READY: Medium not present" after medium was
detected previously, and "UNIT ATTENTION: Not ready to ready change, medium
may have changed". First one reported to disk(9) as media removal, second
as media insert/change. To reliably receive second event new
AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by
generic error handling code in cam_periph_error().
2) Modifications to GEOM core to handle media remove and change events.
Media removal handled by spoiling all consumers attached to the provider.
Media change event also schedules provider retaste after spoiling to probe
new media. New flag G_CF_ORPHAN was added to consumers to reflect that
consumer is in process of destruction. It allows retaste to create new
geom instance of the same class, while previous one is still dying.
3) Modifications to some GEOM classes: DEV -- to report media change
events to devd; VFS -- to handle spoiling same as orphan to prevent
accessing replaced media. PART class already handles spoiling alike to
orphan.

Reviewed by: silence on geom@ and scsi@
Tested by: avg
Sponsored by: iXsystems, Inc. / PC-BSD
MFC after: 2 months

Correct bioq_disksort so that bioq_insert_tail() offers barrier semantic.
Add the BIO_ORDERED flag for struct bio and update bio clients to use it.

The barrier semantics of bioq_insert_tail() were broken in two ways:

o In bioq_disksort(), an added bio could be inserted at the head of
the queue, even when a barrier was present, if the sort key for
the new entry was less than that of the last queued barrier bio.

o The last_offset used to generate the sort key for newly queued bios
did not stay at the position of the barrier until either the
barrier was de-queued, or a new barrier (which updates last_offset)
was queued. When a barrier is in effect, we know that the disk
will pass through the barrier position just before the
"blocked bios" are released, so using the barrier's offset for
last_offset is the optimal choice.

sys/geom/sched/subr_disk.c:
sys/kern/subr_disk.c:
o Update last_offset in bioq_insert_tail().

o Only update last_offset in bioq_remove() if the removed bio is
at the head of the queue (typically due to a call via
bioq_takefirst()) and no barrier is active.

o In bioq_disksort(), if we have a barrier (insert_point is non-NULL),
set prev to the barrier and cur to it's next element. Now that
last_offset is kept at the barrier position, this change isn't
strictly necessary, but since we have to take a decision branch
anyway, it does avoid one, no-op, loop iteration in the while
loop that immediately follows.

o In bioq_disksort(), bypass the normal sort for bios with the
BIO_ORDERED attribute and instead insert them into the queue
with bioq_insert_tail(). bioq_insert_tail() not only gives
the desired command order during insertion, but also provides
barrier semantics so that commands disksorted in the future
cannot pass the just enqueued transaction.

sys/sys/bio.h:
Add BIO_ORDERED as bit 4 of the bio_flags field in struct bio.

sys/cam/ata/ata_da.c:
sys/cam/scsi/scsi_da.c
Use an ordered command for SCSI/ATA-NCQ commands issued in
response to bios with the BIO_ORDERED flag set.

sys/cam/scsi/scsi_da.c
Use an ordered tag when issuing a synchronize cache command.

Wrap some lines to 80 columns.

sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_geom.c
sys/geom/geom_io.c
Mark bios with the BIO_FLUSH command as BIO_ORDERED.

Sponsored by: Spectra Logic Corporation
MFC after: 1 month

Untangle g_print_bio(), silencing Coverity.

Found with: Coverity Prevent
CID: 3566, 3567

g_io_check: respond to zero pp->mediasize with ENXIO

Previsouly this condition was reported with EIO by bio_offset > mediasize
check.
Perhaps that check should be extended to bio_offset+bio_length > mediasize.

MFC after: 1 week

Do not fetch precise time of request start when stats collection disabled.

Reviewed by: pjd, phk

Call wakeup() only for the first request on the queue.

MFp4:
Remove msleep() timeout from g_io_schedule_up/down(). It works fine
without it, saving few percents of CPU on high request rates without
need to rearm callout twice per request.

Make gjournal work with kernel compiled with "options DIAGNOSTIC".
Previously, it would panic immediately.

Reviewed by: pjd
Approved by: re (kib)

As discussed in the devsummit, introduce two fields in the
struct bio to store classification information, and a hook
for classifier functions that can be called by g_io_request().

This code is from Fabio Checconi as part of his GSOC work.

Just a fixup for a KTRACE message I stumbled upon many moons ago.

Reviewed by: Scott Long
MFC after: 2 days

Don't limit BIO_DELETE requests to MAXPHYS, they perform no data
transfers, so they are not subject to the VM system limitation.

Save stack only when KTR_GEOM is both compiled into the kernel and enabled
in debug.ktr.mask. Because saving stack is very expensive, it's better only
to do it when one really wants to.

Reported by: Dan Nelson

Implement g_delete_data() similar to g_read_data() and g_write_data().

OK'ed by: phk

Use pause() rather than tsleep() on stack variables and function pointers.

Use tsleep() rather than msleep() with a NULL mtx parameter.

We expect 'bio_data != NULL' for BIO_{READ,WRITE,GETATTR}, but for
BIO_{DELETE,FLUSH} we expect 'bio_data == NULL'.

Reviewed by: phk

Add a new I/O request - BIO_FLUSH, which basically tells providers below to
flush their caches. For now will mostly be used by disks to flush their
write cache.

Sponsored by: home.pl

Add g_duplicate_bio() function which does the same thing what g_clone_bio()
is doing, but g_duplicate_bio() allocates new bio with M_WAITOK flag.

Fix a typo.

Assert proper use of bio_caller1, bio_caller2, bio_cflags, bio_driver1,
bio_driver2 and bio_pflags fields.

Reviewed by: phk

- Add a new simple facility for marking the current thread as being in a
state where sleeping on a sleep queue is not allowed. The facility
doesn't support recursion but uses a simple private per-thread flag
(TDP_NOSLEEPING). The sleepq_add() function will panic if the flag is
set and INVARIANTS is enabled.
- Use this new facility to replace the g_xup and g_xdown mutexes that were
(ab)used to achieve similar behavior.
- Disallow sleeping in interrupt threads when invoking interrupt handlers.

MFC after: 1 week
Reviewed by: phk

Use KTR to log allocations and destructions of bios.
This should hopefully allow to track down "duplicate free of g_bio" panics.

By design I left a tiny race in updating the I/O statistics based on
the assumption that performance was more important that beancounter
quality statistics.

As it transpires the microoptimization is not measurable in the
real world and the inconsistent statistics confuse users, so revert
the decision.

MT6 candidate: possibly
MT5 candidate: possibly

Add KTR_GEOM, which allows tracing of basic GEOM I/O events occuring
in the g_up and g_down threads. Each time a bio is propelled up and
down the stack, an event is generating showing the provider, offset,
and length, as well as thread wakeup and work status information.

Trace information about a buffer while we still control it.

Reviewed by: phk
Approved by: sam (mentor)

Don't set the BIO_ONQUEUE debugging flag until we actually put the bio
onto a queue. This made the ENOMEM handling an instant panic.

Protect the start/end counts on consumers and providers with the up/down
mutexes.

Make it possible to also protect the disk statistics (at a minor cost in
performance) by setting bit 2 of kern.geom.collectstats.

- Set maximum request size to MAXPHYS (128kB), instead of DFLPHYS (64kB).
- Set minimum request size to sectorsize, instead of 512 bytes.

Approved by: phk (some time ago)

Add more KASSERTS and checks.

Introduce g_alloc_bio() as a waiting variant of g_new_bio().

Use in places where we can sleep and where we previously failed to check
for a NULL pointer.

MT5 candidate.

When sending request once again because of ENOMEM, reset bio_children
and bio_inbed fields to 0. Without this change we can end up with
I/O leakage in some rare situations.
I tested this change by putting failure probability mechanism simlar
to this used in NOP class into g_clone_bio(9) function, so it was
able to return NULL with the given probability.

Discussed with: phk

The g_up and g_down threads use a local 'mymutex' mutex to allow WITNESS
to warn about attempts to sleep in the I/O path. This change pushes the
definition and use of 'mymutex' behind #ifdef WITNESS to avoid the cost
in non-debugging cases. This results in a clear .22% performance win for
512 byte and 1k I/O tests on my SMP test box. Not much, but every bit
counts.

Make the sysctl kern.geom.collectstats more granular.

Bit 0 controls statistics collection on GEOM providers.
Bit 1 controls statistics collection on GEOM consumers.

Default value is 1.

Prodded by: scottl

Calculate bio_completed properly or die!

Approved by: phk

Added g_print_bio() function to print informations about given bio.

Approved by: phk, scottl (mentor)

Bring back the geom_bioqueues, they _are_ a good idea.

ATA will uses these RSN.

Correct usage of mtx_init() API. This is not a functional change since
the code happened to work because MTX_DEF and NULL are both defined as 0.

Reviewed by: phk

Forgotten commit: If a provider has zero sectorsize, it is an
indication of lack of media.

Tripped up: peter

Remove KASSERT check for negative bio_offsets, add "normal" EIO
error return for same.

Allow our bio tools to be used for local bio-chopping by not mandating
a bio_from value. bio_to is still mandated (mostly for debuggign) and
shall be copied from the parent bio.

Add more KASSERTS().

Reorder a couple of KASSERTS to give more sensible messages.

Found by: GEOM 101 class of '03

In case we encounter a zero sectorsize provider in g_io_check(), fail
the request with a printf rather than a divide by zero error.

Sleep on "-" in our normal state to simplify debugging.

Use __FBSDID().

Approved by: phk

Hide the "ENOMEM" notice messages behind bootverbose. They are still
a valuable debugging tool for certain kinds of problems.

Approved by: re/scottl

Use an uma-zone for allocation bio requests.

Back out all the stuff that didn't belong in the last commit.

Use g_slice_spoiled() rather than g_std_spoiled().

Remember to free the buffer we got from g_read_data().

Time has run from the "run GEOM in userland" harness, and the new regression
test is built to test GEOM as running in the kernel.

This commit is basically "unifdef -D_KERNEL" to remove the mainly #include
related code to support the userland-harness.

Retire the experimental bio_taskqueue(), it was not quite as usable as
hoped. It can be revived from here, should other drivers be able to
use it.

Remove all references to BIO_SETATTR. We will not be using it.

Fix a bug in the ENOMEM pacing code which probably made it panic systems
after a lot of ENOMEM errors.

Retire the GEOM private statistics code and use devstat instead.

Including <sys/stdint.h> is (almost?) universally only to be able to use
%j in printfs, so put a newsted include in <sys/systm.h> where the printf
prototype lives and save everybody else the trouble.

Don't abuse the statistics counters for detecting if we have outstanding
I/O requests, instead use the new dedicated fields in the consumer and
provider to track this.

Back out M_* changes, per decision of the TRB.

Approved by: trb

Implement a bio-taskqueue to reduce number of context switches in
disk I/O processing.

The intent is that the disk driver in its hardware interrupt
routine will simply schedule the bio on the task queue with
a routine to finish off whatever needs done.

The g_up thread will then schedule this routine, the likely
outcome of which is a biodone() which queues the bio on
g_up's regular queue where it will be picked up and processed.

Compared to the using the regular taskqueue, this saves one
contextswitch.

Change our scheduling of the g_up and g_down queues to be water-tight,
at the cost of breaking the userland regression test-shims.

Input and ideas from: scottl

Don't short-circuit zero-length requests of they are BIO_[SG]ETATTR.

Turn the "updating" flag (back) into two sequence number fields at
either ends of the structure so we have a way to determine if a
snapshot is consistent.

Update the statistics collection code to track busy time instead of
idle time.

Statistics now default to "on" and can be turned off with
sysctl kern.geom.collectstats=0

Performance impact of statistics collection is on the order of
800 nsec per consumer/provider set on a 700MHz Athlon.

Move the g_stat struct to its own .h file, we will export it to other code.

Insted of embedding a struct g_stat in consumers and providers, merely
include a pointer.

Remove a couple of <sys/time.h> includes now unneeded.

Add a special allocator for struct g_stat. This allocator will allocate
entire pages and hand out g_stat functions from there. The "id" field
indicates free/used status.

Add "/dev/geom.stats" device driver whic exports the pages from the
allocator to userland with mmap(2) in read-only mode.

This mmap(2) interface should be considered a non-public interface and
the functions in libgeom (not yet committed) should be used to access
the statistics data.

Commit the correct copy of the g_stat structure.

Add debug.sizeof.g_stat sysctl.

Set the id field of the g_stat when we create consumers and providers.

Remove biocount from consumer, we will use the counters in the g_stat
structure instead. Replace one field which will need to be atomically
manipulated with two fields which will not (stat.nop and stat.nend).

Change add companion field to bio_children: bio_inbed for the exact
same reason.

Don't output the biocount in the confdot output.

Fix KASSERT in g_io_request().

Add sysctl kern.geom.collectstats defaulting to off.

Collect the following raw statistics conditioned on this sysctl:

for each consumer and provider {
total number of operations started.
total number of operations completed.
time last operation completed.
sum of idle-time.
for each of BIO_READ, BIO_WRITE and BIO_DELETE {
number of operations completed.
number of bytes completed.
number of ENOMEM errors.
number of other errors.
sum of transaction time.
}
}

API for getting hold of these statistics data not included yet.

Rename bio_linkage to the more obvious bio_parent.
Add bio_t0 timestamp, and include <sys/time.h> where needed

Put the checks we perform on a bio before calling ::start in their
own function, handle all validation and truncation at the time we
process the bio instead of when it gets scheduled.

Remove M_TRYWAIT/M_WAITOK/M_WAIT. Callers should use 0.
Merge M_NOWAIT/M_DONTWAIT into a single flag M_NOWAIT.

white-space changes

Balk at unaligned requests.

MFC candidate.

malloc(9) with M_NOWAIT seems to return NULL a lot more than I would have
expected under -current. This is a problem for GEOM because the up/down
threads cannot sleep waiting for memory to become free. The reason they
cannot sleep is that paging things out to disk may be the only way we can
clear up some RAM. Nice catch-22 there.

Implement a rudimentary ENOMEM recovery strategy: If an I/O request
fails with an error code of ENOMEM, schedule it for a retry, and
tell the down-thread to sleep hz/10 to get other parts of the system
a chance to free up some memory, in particular the up-path in GEOM.

All caches should probably start to monitor malloc(9) failures using the new
malloc_last_fail() function, and release when it indicates congestion.

Sponsored by: DARPA & NAI Labs.

Don't track bio allocation in debug output.

Sponsored by: DARPA & NAI Labs.

Add more KASSERTS.

Sponsored by: DARPA & NAI Labs.

Add support g_clone_bio() and g_std_done() to spawn multiple children
of a bio and correctly gather status when done.

Sponsored by: DARPA & NAI Labs.

For now, don't wait for drives to stop returning EBUSY. There is too
much broken harware around it seems.

Sponsored by: DARPA & NAI Labs.

Put a printf under #ifdef DIAGNOSTIC.

Sponsored by: DARPA & NAI Labs.

Retire g_io_fail() and let g_io_deliver() take an error argument instead.

Sponsored by: DARPA & NAI Labs.

Introduce g_write_data() function.

Sponsored by: DARPA & NAI Labs

Void functions cannot use return(foo) even if foo is also returning void.

Sponsored by: DARPA & NAI Labs.

Setattr should not retry on EBUSY, we could get EBUSY back because
a disklabel modification tries to change an open device, and no
counter-examples exists.

Be less facist about when we can do Setattr, the openmodes of devices
are so loosely managed that the "exclusive" count is almost useless.

Sponsored by: DARPA & NAI Labs.

Allocate bio's with M_NOWAIT and let the caller deal with the problems.

Sponsored by: DARPA & NAI Labs.

Use biowait() rather than DIY.

Sponsored by: DARPA & NAI Labs

Don't grab Giant around malloc(9) and free(9).
Don't grab Giant around wakeup(9).
Don't print verbose messages about each device found in geom_dev.
Various cleanups.

Sponsored by: DARPA & NAI Labs.

Constifixation of attribute argument to g_io_[gs]etattr()

Sponsored by: DARPA & NAI Labs

Change callers of mtx_init() to pass in an appropriate lock type name. In
most cases NULL is passed, but in some cases such as network driver locks
(which use the MTX_NETWORK_LOCK macro) and UMA zone locks, a name is used.

Tested on: i386, alpha, sparc64

Centralize EOF handling and improve access controls for bio scheduling.

Sponsored by: DARPA & NAI Labs

Eliminate some thread pointers which do not make sense anymore.

Split private parts of geom.h into geom_int.h. The latter should
never be included in class implemtations.

Push BIO_FORMAT into a local hack inside the floppy drivers where
it belongs.

First commit of the GEOM subsystem to make it easier for people to
test and play with this.

This is not yet production quality and should be run only on dedicated
test boxes.

For people who want to develop transformations for GEOM there exist a
set of shims to run geom in userland (ask phk@freebsd.org).

Reports of all kinds to: phk@freebsd.org
Please include in report:
dmesg
sysctl debug.geomdot
sysctl debug.geomconf

Known significant limitations:
no kernel dump facility.
ioctls severely restricted.

Sponsored by: DARPA, NAI Labs