The CIL push lock is highly contended on larger machines, becoming a
hard bottleneck that about 700,000 transaction commits/s on >16p
machines. To address this, start moving the CIL tracking
infrastructure to utilise per-CPU structures.

We need to track the space used, the amount of log reservation space
reserved to write the CIL, the log items in the CIL and the busy
extents that need to be completed by the CIL commit.  This requires
a couple of per-cpu counters, an unordered per-cpu list and a
globally ordered per-cpu list.

Create a per-cpu structure to hold these and all the management
interfaces needed, as well as the hooks to handle hotplug CPUs.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>

The current blocking mechanism for pushing the inodegc queue out to
disk can result in systems becoming unusable when there is a long
running inodegc operation. This is because the statfs()
implementation currently issues a blocking flush of the inodegc
queue and a significant number of common system utilities will call
statfs() to discover something about the underlying filesystem.

This can result in userspace operations getting stuck on inodegc
progress, and when trying to remove a heavily reflinked file on slow
storage with a full journal, this can result in delays measuring in
hours.

Avoid this problem by adding "push" function that expedites the
flushing of the inodegc queue, but doesn't wait for it to complete.

Convert xfs_fs_statfs() and xfs_qm_scall_getquota() to use this
mechanism so they don't block but still ensure that queued
operations are expedited.

Fixes: ab23a776 ("xfs: per-cpu deferred inode inactivation queues")
Reported-by: NChris Dunlop <chris@onthe.net.au>
Signed-off-by: NDave Chinner <dchinner@redhat.com>
[djwong: fix _getquota_next to use _inodegc_push too]
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>

Currently inodegc work can sit queued on the per-cpu queue until
the workqueue is either flushed of the queue reaches a depth that
triggers work queuing (and later throttling). This means that we
could queue work that waits for a long time for some other event to
trigger flushing.

Hence instead of just queueing work at a specific depth, use a
delayed work that queues the work at a bound time. We can still
schedule the work immediately at a given depth, but we no long need
to worry about leaving a number of items on the list that won't get
processed until external events prevail.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>

The LARP patchset added an awkward coupling point between libxfs and
what would be libxlog, if the XFS log were actually its own library.
Move the code that enables logged xattr updates out of "lib"xlog and into
xfs_xattr.c so that it no longer has to know about xlog_* functions.

While we're at it, give xfs_xattr.c its own header file.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

Initialize and destroy the xattr log item caches in the same places that
we do all the other log item caches.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NAllison Henderson <allison.henderson@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

Just use a non-zero max_discard_sectors as an indicator for discard
support, similar to what is done for write zeroes.

The only places where needs special attention is the RAID5 driver,
which must clear discard support for security reasons by default,
even if the default stacking rules would allow for it.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Acked-by: Christoph Böhmwalder <christoph.boehmwalder@linbit.com> [drbd]
Acked-by: Jan Höppner <hoeppner@linux.ibm.com> [s390]
Acked-by: Coly Li <colyli@suse.de> [bcache]
Acked-by: David Sterba <dsterba@suse.com> [btrfs]
Reviewed-by: NChaitanya Kulkarni <kch@nvidia.com>
Link: https://lore.kernel.org/r/20220415045258.199825-25-hch@lst.deSigned-off-by: NJens Axboe <axboe@kernel.dk>

This commit enables XFS module to work with fs instances having 64-bit
per-inode extent counters by adding XFS_SB_FEAT_INCOMPAT_NREXT64 flag to the
list of supported incompat feature flags.
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NChandan Babu R <chandan.babu@oracle.com>

As mentioned in the previous commit, the kernel misuses sb_frextents in
the incore mount to reflect both incore reservations made by running
transactions as well as the actual count of free rt extents on disk.
This results in the superblock being written to the log with an
underestimate of the number of rt extents that are marked free in the
rtbitmap.

Teaching XFS to recompute frextents after log recovery avoids
operational problems in the current mount, but it doesn't solve the
problem of us writing undercounted frextents which are then recovered by
an older kernel that doesn't have that fix.

Create an incore percpu counter to mirror the ondisk frextents.  This
new counter will track transaction reservations and the only time we
will touch the incore super counter (i.e the one that gets logged) is
when those transactions commit updates to the rt bitmap.  This is in
contrast to the lazysbcount counters (e.g. fdblocks), where we know that
log recovery will always fix any incorrect counter that we log.
As a bonus, we only take m_sb_lock at transaction commit time.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NDave Chinner <david@fromorbit.com>

On a modern filesystem, we don't allow userspace to allocate blocks for
data storage from the per-AG space reservations, the user-controlled
reservation pool that prevents ENOSPC in the middle of internal
operations, or the internal per-AG set-aside that prevents unwanted
filesystem shutdowns due to ENOSPC during a bmap btree split.

Since we now consider freespace btree blocks as unavailable for
allocation for data storage, we shouldn't report those blocks via statfs
either.  This makes the numbers that we return via the statfs f_bavail
and f_bfree fields a more conservative estimate of actual free space.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

In commit 02b9984d, we pushed a sync_filesystem() call from the VFS
into xfs_fs_remount.  The only time that we ever need to push dirty file
data or metadata to disk for a remount is if we're remounting the
filesystem read only, so this really could be moved to xfs_remount_ro.

Once we've moved the call site, actually check the return value from
sync_filesystem.

Fixes: 02b9984d ("fs: push sync_filesystem() down to the file system's remount_fs()")
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Now that the VFS will do something with the return values from
->sync_fs, make ours pass on error codes.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NJan Kara <jack@suse.cz>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NChristian Brauner <brauner@kernel.org>

As part of multiple customer escalations due to file data corruption
after copy on write operations, I wrote some fstests that use fsstress
to hammer on COW to shake things loose.  Regrettably, I caught some
filesystem shutdowns due to incorrect rmap operations with the following
loop:

mount <filesystem>				# (0)
fsstress <run only readonly ops> &		# (1)
while true; do
	fsstress <run all ops>
	mount -o remount,ro			# (2)
	fsstress <run only readonly ops>
	mount -o remount,rw			# (3)
done

When (2) happens, notice that (1) is still running.  xfs_remount_ro will
call xfs_blockgc_stop to walk the inode cache to free all the COW
extents, but the blockgc mechanism races with (1)'s reader threads to
take IOLOCKs and loses, which means that it doesn't clean them all out.
Call such a file (A).

When (3) happens, xfs_remount_rw calls xfs_reflink_recover_cow, which
walks the ondisk refcount btree and frees any COW extent that it finds.
This function does not check the inode cache, which means that incore
COW forks of inode (A) is now inconsistent with the ondisk metadata.  If
one of those former COW extents are allocated and mapped into another
file (B) and someone triggers a COW to the stale reservation in (A), A's
dirty data will be written into (B) and once that's done, those blocks
will be transferred to (A)'s data fork without bumping the refcount.

The results are catastrophic -- file (B) and the refcount btree are now
corrupt.  In the first patch, we fixed the race condition in (2) so that
(A) will always flush the COW fork.  In this second patch, we move the
_recover_cow call to the initial mount call in (0) for safety.

As mentioned previously, xfs_reflink_recover_cow walks the refcount
btree looking for COW staging extents, and frees them.  This was
intended to be run at mount time (when we know there are no live inodes)
to clean up any leftover staging events that may have been left behind
during an unclean shutdown.  As a time "optimization" for readonly
mounts, we deferred this to the ro->rw transition, not realizing that
any failure to clean all COW forks during a rw->ro transition would
result in catastrophic corruption.

Therefore, remove this optimization and only run the recovery routine
when we're guaranteed not to have any COW staging extents anywhere,
which means we always run this at mount time.  While we're at it, move
the callsite to xfs_log_mount_finish because any refcount btree
expansion (however unlikely given that we're removing records from the
right side of the index) must be fed by a per-AG reservation, which
doesn't exist in its current location.

Fixes: 174edb0e ("xfs: store in-progress CoW allocations in the refcount btree")
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NChandan Babu R <chandan.babu@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

As part of multiple customer escalations due to file data corruption
after copy on write operations, I wrote some fstests that use fsstress
to hammer on COW to shake things loose.  Regrettably, I caught some
filesystem shutdowns due to incorrect rmap operations with the following
loop:

mount <filesystem>				# (0)
fsstress <run only readonly ops> &		# (1)
while true; do
	fsstress <run all ops>
	mount -o remount,ro			# (2)
	fsstress <run only readonly ops>
	mount -o remount,rw			# (3)
done

When (2) happens, notice that (1) is still running.  xfs_remount_ro will
call xfs_blockgc_stop to walk the inode cache to free all the COW
extents, but the blockgc mechanism races with (1)'s reader threads to
take IOLOCKs and loses, which means that it doesn't clean them all out.
Call such a file (A).

When (3) happens, xfs_remount_rw calls xfs_reflink_recover_cow, which
walks the ondisk refcount btree and frees any COW extent that it finds.
This function does not check the inode cache, which means that incore
COW forks of inode (A) is now inconsistent with the ondisk metadata.  If
one of those former COW extents are allocated and mapped into another
file (B) and someone triggers a COW to the stale reservation in (A), A's
dirty data will be written into (B) and once that's done, those blocks
will be transferred to (A)'s data fork without bumping the refcount.

The results are catastrophic -- file (B) and the refcount btree are now
corrupt.  Solve this race by forcing the xfs_blockgc_free_space to run
synchronously, which causes xfs_icwalk to return to inodes that were
skipped because the blockgc code couldn't take the IOLOCK.  This is safe
to do here because the VFS has already prohibited new writer threads.

Fixes: 10ddf64e ("xfs: remove leftover CoW reservations when remounting ro")
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChandan Babu R <chandan.babu@oracle.com>

Hide the DAX device lookup from the xfs_super.c code.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NDan Williams <dan.j.williams@intel.com>
Link: https://lore.kernel.org/r/20211129102203.2243509-22-hch@lst.deSigned-off-by: NDan Williams <dan.j.williams@intel.com>

Just open code the block size and dax_dev == NULL checks in the callers.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NMike Snitzer <snitzer@redhat.com>
Reviewed-by: Gao Xiang <hsiangkao@linux.alibaba.com> [erofs]
Reviewed-by: NDan Williams <dan.j.williams@intel.com>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Link: https://lore.kernel.org/r/20211129102203.2243509-9-hch@lst.deSigned-off-by: NDan Williams <dan.j.williams@intel.com>

Factor out another DAX setup helper to simplify future changes.  Also
move the experimental warning after the checks to not clutter the log
too much if the setup failed.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDan Williams <dan.j.williams@intel.com>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Link: https://lore.kernel.org/r/20211129102203.2243509-8-hch@lst.deSigned-off-by: NDan Williams <dan.j.williams@intel.com>

Fix following checkincludes.pl warning:
./fs/xfs/xfs_super.c: xfs_btree.h is included more than once.

The include is in line 15. Remove the duplicated here.
Signed-off-by: NWan Jiabing <wanjiabing@vivo.com>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>

xfs_bmap_add_free isn't a block mapping function; it schedules deferred
freeing operations for a later point in a compound transaction chain.
While it's primarily used by bunmapi, its use has expanded beyond that.
Move it to xfs_alloc.c and rename the function since it's now general
freeing functionality.  Bring the slab cache bits in line with the
way we handle the other intent items.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NChandan Babu R <chandan.babu@oracle.com>

Create slab caches for the high-level structures that coordinate
deferred intent items, since they're used fairly heavily.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NChandan Babu R <chandan.babu@oracle.com>

Now that we've gotten rid of the kmem_zone_t typedef, rename the
variables to _cache since that's what they are.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NChandan Babu R <chandan.babu@oracle.com>

Now that we have the infrastructure to track the max possible height of
each btree type, we can create a separate slab cache for cursors of each
type of btree.  For smaller indices like the free space btrees, this
means that we can pack more cursors into a slab page, improving slab
utilization.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

To support future btree code, we need to be able to size btree cursors
dynamically for very large btrees.  Switch the maxlevels computation to
use the precomputed values in the superblock, and create cursors that
can handle a certain height.  For now, we retain the btree cursor cache
that can handle up to 9-level btrees, though a subsequent patch
introduces separate caches for each btree type, where each cache's
objects will be exactly tall enough to handle the specific btree type.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Split out the btree level information into a separate struct and put it
at the end of the cursor structure as a VLA.  Files with huge data forks
(and in the future, the realtime rmap btree) will require the ability to
support many more levels than a per-AG btree cursor, which means that
we're going to create per-btree type cursor caches to conserve memory
for the more common case.

Note that a subsequent patch actually introduces dynamic cursor heights.
This one merely rearranges the structure to prepare for that.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NChandan Babu R <chandan.babu@oracle.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

All callers already have a dax_device obtained from fs_dax_get_by_bdev
at hand, so just pass that to dax_supported() insted of doing another
lookup.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDan Williams <dan.j.williams@intel.com>
Link: https://lore.kernel.org/r/20210826135510.6293-10-hch@lst.deSigned-off-by: NDan Williams <dan.j.williams@intel.com>

Refactor the DAX setup code in preparation of removing
bdev_dax_supported.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDan Williams <dan.j.williams@intel.com>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Link: https://lore.kernel.org/r/20210826135510.6293-9-hch@lst.deSigned-off-by: NDan Williams <dan.j.williams@intel.com>

Rather than open coding XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5
checks everywhere, add a simple wrapper to encapsulate this and make
the code easier to read.

This allows us to remove the xfs_sb_version_has_v3inode() wrapper
which is only used in xfs_format.h now and is just a version number
check.

There are a couple of places where we should be checking the mount
feature bits rather than the superblock version (e.g. remount), so
those are converted to use xfs_has_crc(mp) instead.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>

This is a conversion of the remaining xfs_sb_version_has..(sbp)
checks to use xfs_has_..(mp) feature checks.

This was largely done with a vim replacement macro that did:

:0,$s/xfs_sb_version_has\(.*\)&\(.*\)->m_sb/xfs_has_\1\2/g<CR>

A couple of other variants were also used, and the rest touched up
by hand.

$ size -t fs/xfs/built-in.a
	   text    data     bss     dec     hex filename
before	1127533  311352     484 1439369  15f689 (TOTALS)
after	1125360  311352     484 1437196  15ee0c (TOTALS)
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>

Remove the shouty macro and instead use the inline function that
matches other state/feature check wrapper naming. This conversion
was done with sed.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>

The remaining mount flags kept in m_flags are actually runtime state
flags. These change dynamically, so they really should be updated
atomically so we don't potentially lose an update due to racing
modifications.

Convert these remaining flags to be stored in m_opstate and use
atomic bitops to set and clear the flags. This also adds a couple of
simple wrappers for common state checks - read only and shutdown.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>

Replace m_flags feature checks with xfs_has_<feature>() calls and
rework the setup code to set flags in m_features.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>

Convert the xfs_sb_version_hasfoo() to checks against
mp->m_features. Checks of the superblock itself during disk
operations (e.g. in the read/write verifiers and the to/from disk
formatters) are not converted - they operate purely on the
superblock state. Everything else should use the mount features.

Large parts of this conversion were done with sed with commands like
this:

for f in `git grep -l xfs_sb_version_has fs/xfs/*.c`; do
	sed -i -e 's/xfs_sb_version_has\(.*\)(&\(.*\)->m_sb)/xfs_has_\1(\2)/' $f
done

With manual cleanups for things like "xfs_has_extflgbit" and other
little inconsistencies in naming.

The result is ia lot less typing to check features and an XFS binary
size reduced by a bit over 3kB:

$ size -t fs/xfs/built-in.a
	text	   data	    bss	    dec	    hex	filenam
before	1130866  311352     484 1442702  16038e (TOTALS)
after	1127727  311352     484 1439563  15f74b (TOTALS)
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>

The attr2 feature is somewhat unique in that it has both a superblock
feature bit to enable it and mount options to enable and disable it.

Back when it was first introduced in 2005, attr2 was disabled unless
either the attr2 superblock feature bit was set, or the attr2 mount
option was set. If the superblock feature bit was not set but the
mount option was set, then when the first attr2 format inode fork
was created, it would set the superblock feature bit. This is as it
should be - the superblock feature bit indicated the presence of the
attr2 on disk format.

The noattr2 mount option, however, did not affect the superblock
feature bit. If noattr2 was specified, the on-disk superblock
feature bit was ignored and the code always just created attr1
format inode forks.  If neither of the attr2 or noattr2 mounts
option were specified, then the behaviour was determined by the
superblock feature bit.

This was all pretty sane.

Fast foward 3 years, and we are dealing with fallout from the
botched sb_features2 addition and having to deal with feature
mismatches between the sb_features2 and sb_bad_features2 fields. The
attr2 feature bit was one of these flags. The reconciliation was
done well after mount option parsing and, unfortunately, the feature
reconciliation had a bug where it ignored the noattr2 mount option.

For reasons lost to the mists of time, it was decided that resolving
this issue in commit 7c12f296 ("[XFS] Fix up noattr2 so that it
will properly update the versionnum and features2 fields.") required
noattr2 to clear the superblock attr2 feature bit.  This greatly
complicated the attr2 behaviour and broke rules about feature bits
needing to be set when those specific features are present in the
filesystem.

By complicated, I mean that it introduced problems due to feature
bit interactions with log recovery. All of the superblock feature
bit checks are done prior to log recovery, but if we crash after
removing a feature bit, then on the next mount we see the feature
bit in the unrecovered superblock, only to have it go away after the
log has been replayed.  This means our mount time feature processing
could be all wrong.

Hence you can mount with noattr2, crash shortly afterwards, and
mount again without attr2 or noattr2 and still have attr2 enabled
because the second mount sees attr2 still enabled in the superblock
before recovery runs and removes the feature bit. It's just a mess.

Further, this is all legacy code as the v5 format requires attr2 to
be enabled at all times and it cannot be disabled.  i.e. the noattr2
mount option returns an error when used on v5 format filesystems.

To straighten this all out, this patch reverts the attr2/noattr2
mount option behaviour back to the original behaviour. There is no
reason for disabling attr2 these days, so we will only do this when
the noattr2 mount option is set. This will not remove the superblock
feature bit. The superblock bit will provide the default behaviour
and only track whether attr2 is present on disk or not. The attr2
mount option will enable the creation of attr2 format inode forks,
and if the superblock feature bit is not set it will be added when
the first attr2 inode fork is created.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>

We only use the CIL workqueue in the CIL, so it makes no sense to
hang it off the xfs_mount and have to walk multiple pointers back up
to the mount when we have the CIL structures right there.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>

Because we use a single work structure attached to the CIL rather
than the CIL context, we can only queue a single work item at a
time. This results in the CIL being single threaded and limits
performance when it becomes CPU bound.

The design of the CIL is that it is pipelined and multiple commits
can be running concurrently, but the way the work is currently
implemented means that it is not pipelining as it was intended. The
critical work to switch the CIL context can take a few milliseconds
to run, but the rest of the CIL context flush can take hundreds of
milliseconds to complete. The context switching is the serialisation
point of the CIL, once the context has been switched the rest of the
context push can run asynchrnously with all other context pushes.

Hence we can move the work to the CIL context so that we can run
multiple CIL pushes at the same time and spread the majority of
the work out over multiple CPUs. We can keep the per-cpu CIL commit
state on the CIL rather than the context, because the context is
pinned to the CIL until the switch is done and we aggregate and
drain the per-cpu state held on the CIL during the context switch.

However, because we no longer serialise the CIL work, we can have
effectively unlimited CIL pushes in progress. We don't want to do
this - not only does it create contention on the iclogs and the
state machine locks, we can run the log right out of space with
outstanding pushes. Instead, limit the work concurrency to 4
concurrent works being processed at a time. This is enough
concurrency to remove the CIL from being a CPU bound bottleneck but
not enough to create new contention points or unbound concurrency
issues.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>

log->l_flags doesn't actually contain "flags" as such, it contains
operational state information that can change at runtime. For the
shutdown state, this at least should be an atomic bit because
it is read without holding locks in many places and so using atomic
bitops for the state field modifications makes sense.

This allows us to use things like test_and_set_bit() on state
changes (e.g. setting XLOG_TAIL_WARN) to avoid races in setting the
state when we aren't holding locks.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>

These two features were merged a year ago, userspace tooling have been
merged, and no serious errors have been reported by the developers.
Drop the experimental tag to encourage wider testing.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NChandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: NCarlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: NBill O'Donnell <billodo@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

Now that we have the infrastructure to switch background workers on and
off at will, fix the block gc worker code so that we don't actually run
the worker when the filesystem is frozen, same as we do for deferred
inactivation.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Users have come to expect that the space accounting information in
statfs and getquota reports are fairly accurate.  Now that we inactivate
inodes from a background queue, these numbers can be thrown off by
whatever resources are singly-owned by the inodes in the queue.  Flush
the pending inactivations when userspace asks for a space usage report.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Move inode inactivation to background work contexts so that it no
longer runs in the context that releases the final reference to an
inode. This will allow process work that ends up blocking on
inactivation to continue doing work while the filesytem processes
the inactivation in the background.

A typical demonstration of this is unlinking an inode with lots of
extents. The extents are removed during inactivation, so this blocks
the process that unlinked the inode from the directory structure. By
moving the inactivation to the background process, the userspace
applicaiton can keep working (e.g. unlinking the next inode in the
directory) while the inactivation work on the previous inode is
done by a different CPU.

The implementation of the queue is relatively simple. We use a
per-cpu lockless linked list (llist) to queue inodes for
inactivation without requiring serialisation mechanisms, and a work
item to allow the queue to be processed by a CPU bound worker
thread. We also keep a count of the queue depth so that we can
trigger work after a number of deferred inactivations have been
queued.

The use of a bound workqueue with a single work depth allows the
workqueue to run one work item per CPU. We queue the work item on
the CPU we are currently running on, and so this essentially gives
us affine per-cpu worker threads for the per-cpu queues. THis
maintains the effective CPU affinity that occurs within XFS at the
AG level due to all objects in a directory being local to an AG.
Hence inactivation work tends to run on the same CPU that last
accessed all the objects that inactivation accesses and this
maintains hot CPU caches for unlink workloads.

A depth of 32 inodes was chosen to match the number of inodes in an
inode cluster buffer. This hopefully allows sequential
allocation/unlink behaviours to defering inactivation of all the
inodes in a single cluster buffer at a time, further helping
maintain hot CPU and buffer cache accesses while running
inactivations.

A hard per-cpu queue throttle of 256 inode has been set to avoid
runaway queuing when inodes that take a long to time inactivate are
being processed. For example, when unlinking inodes with large
numbers of extents that can take a lot of processing to free.
Signed-off-by: NDave Chinner <dchinner@redhat.com>
[djwong: tweak comments and tracepoints, convert opflags to state bits]
Reviewed-by: NDarrick J. Wong <djwong@kernel.org>
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>

Move the xfs_inactive call and all the other debugging checks and stats
updates into xfs_inode_mark_reclaimable because most of that are
implementation details about the inode cache.  This is preparation for
deferred inactivation that is coming up.  We also move it around
xfs_icache.c in preparation for deferred inactivation.
Signed-off-by: NDarrick J. Wong <djwong@kernel.org>
Reviewed-by: NDave Chinner <dchinner@redhat.com>