Merge branch 'for-3.10/drivers' of git://git.kernel.dk/linux-block

Pull block driver updates from Jens Axboe:
"It might look big in volume, but when categorized, not a lot of
drivers are touched. The pull request contains:

- mtip32xx fixes from Micron.

- A slew of drbd updates, this time in a nicer series.

- bcache, a flash/ssd caching framework from Kent.

- Fixes for cciss"

* 'for-3.10/drivers' of git://git.kernel.dk/linux-block: (66 commits)
bcache: Use bd_link_disk_holder()
bcache: Allocator cleanup/fixes
cciss: bug fix to prevent cciss from loading in kdump crash kernel
cciss: add cciss_allow_hpsa module parameter
drivers/block/mg_disk.c: add CONFIG_PM_SLEEP to suspend/resume functions
mtip32xx: Workaround for unaligned writes
bcache: Make sure blocksize isn't smaller than device blocksize
bcache: Fix merge_bvec_fn usage for when it modifies the bvm
bcache: Correctly check against BIO_MAX_PAGES
bcache: Hack around stuff that clones up to bi_max_vecs
bcache: Set ra_pages based on backing device's ra_pages
bcache: Take data offset from the bdev superblock.
mtip32xx: mtip32xx: Disable TRIM support
mtip32xx: fix a smatch warning
bcache: Disable broken btree fuzz tester
bcache: Fix a format string overflow
bcache: Fix a minor memory leak on device teardown
bcache: Documentation updates
bcache: Use WARN_ONCE() instead of __WARN()
bcache: Add missing #include <linux/prefetch.h>
...

Documentation/ABI/testing/sysfs-block-bcache

···
+What:		/sys/block/<disk>/bcache/unregister
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		A write to this file causes the backing device or cache to be
+		unregistered. If a backing device had dirty data in the cache,
+		writeback mode is automatically disabled and all dirty data is
+		flushed before the device is unregistered. Caches unregister
+		all associated backing devices before unregistering themselves.
+
+What:		/sys/block/<disk>/bcache/clear_stats
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		Writing to this file resets all the statistics for the device.
+
+What:		/sys/block/<disk>/bcache/cache
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For a backing device that has cache, a symlink to
+		the bcache/ dir of that cache.
+
+What:		/sys/block/<disk>/bcache/cache_hits
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For backing devices: integer number of full cache hits,
+		counted per bio. A partial cache hit counts as a miss.
+
+What:		/sys/block/<disk>/bcache/cache_misses
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For backing devices: integer number of cache misses.
+
+What:		/sys/block/<disk>/bcache/cache_hit_ratio
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For backing devices: cache hits as a percentage.
+
+What:		/sys/block/<disk>/bcache/sequential_cutoff
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For backing devices: Threshold past which sequential IO will
+		skip the cache. Read and written as bytes in human readable
+		units (i.e. echo 10M > sequntial_cutoff).
+
+What:		/sys/block/<disk>/bcache/bypassed
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		Sum of all reads and writes that have bypassed the cache (due
+		to the sequential cutoff).  Expressed as bytes in human
+		readable units.
+
+What:		/sys/block/<disk>/bcache/writeback
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For backing devices: When on, writeback caching is enabled and
+		writes will be buffered in the cache. When off, caching is in
+		writethrough mode; reads and writes will be added to the
+		cache but no write buffering will take place.
+
+What:		/sys/block/<disk>/bcache/writeback_running
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For backing devices: when off, dirty data will not be written
+		from the cache to the backing device. The cache will still be
+		used to buffer writes until it is mostly full, at which point
+		writes transparently revert to writethrough mode. Intended only
+		for benchmarking/testing.
+
+What:		/sys/block/<disk>/bcache/writeback_delay
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For backing devices: In writeback mode, when dirty data is
+		written to the cache and the cache held no dirty data for that
+		backing device, writeback from cache to backing device starts
+		after this delay, expressed as an integer number of seconds.
+
+What:		/sys/block/<disk>/bcache/writeback_percent
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For backing devices: If nonzero, writeback from cache to
+		backing device only takes place when more than this percentage
+		of the cache is used, allowing more write coalescing to take
+		place and reducing total number of writes sent to the backing
+		device. Integer between 0 and 40.
+
+What:		/sys/block/<disk>/bcache/synchronous
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For a cache, a boolean that allows synchronous mode to be
+		switched on and off. In synchronous mode all writes are ordered
+		such that the cache can reliably recover from unclean shutdown;
+		if disabled bcache will not generally wait for writes to
+		complete but if the cache is not shut down cleanly all data
+		will be discarded from the cache. Should not be turned off with
+		writeback caching enabled.
+
+What:		/sys/block/<disk>/bcache/discard
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For a cache, a boolean allowing discard/TRIM to be turned off
+		or back on if the device supports it.
+
+What:		/sys/block/<disk>/bcache/bucket_size
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For a cache, bucket size in human readable units, as set at
+		cache creation time; should match the erase block size of the
+		SSD for optimal performance.
+
+What:		/sys/block/<disk>/bcache/nbuckets
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For a cache, the number of usable buckets.
+
+What:		/sys/block/<disk>/bcache/tree_depth
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For a cache, height of the btree excluding leaf nodes (i.e. a
+		one node tree will have a depth of 0).
+
+What:		/sys/block/<disk>/bcache/btree_cache_size
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		Number of btree buckets/nodes that are currently cached in
+		memory; cache dynamically grows and shrinks in response to
+		memory pressure from the rest of the system.
+
+What:		/sys/block/<disk>/bcache/written
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For a cache, total amount of data in human readable units
+		written to the cache, excluding all metadata.
+
+What:		/sys/block/<disk>/bcache/btree_written
+Date:		November 2010
+Contact:	Kent Overstreet <[e-mail]>
+Description:
+		For a cache, sum of all btree writes in human readable units.

Documentation/bcache.txt

···
+Say you've got a big slow raid 6, and an X-25E or three. Wouldn't it be
+nice if you could use them as cache... Hence bcache.
+
+Wiki and git repositories are at:
+  http://bcache.evilpiepirate.org
+  http://evilpiepirate.org/git/linux-bcache.git
+  http://evilpiepirate.org/git/bcache-tools.git
+
+It's designed around the performance characteristics of SSDs - it only allocates
+in erase block sized buckets, and it uses a hybrid btree/log to track cached
+extants (which can be anywhere from a single sector to the bucket size). It's
+designed to avoid random writes at all costs; it fills up an erase block
+sequentially, then issues a discard before reusing it.
+
+Both writethrough and writeback caching are supported. Writeback defaults to
+off, but can be switched on and off arbitrarily at runtime. Bcache goes to
+great lengths to protect your data - it reliably handles unclean shutdown. (It
+doesn't even have a notion of a clean shutdown; bcache simply doesn't return
+writes as completed until they're on stable storage).
+
+Writeback caching can use most of the cache for buffering writes - writing
+dirty data to the backing device is always done sequentially, scanning from the
+start to the end of the index.
+
+Since random IO is what SSDs excel at, there generally won't be much benefit
+to caching large sequential IO. Bcache detects sequential IO and skips it;
+it also keeps a rolling average of the IO sizes per task, and as long as the
+average is above the cutoff it will skip all IO from that task - instead of
+caching the first 512k after every seek. Backups and large file copies should
+thus entirely bypass the cache.
+
+In the event of a data IO error on the flash it will try to recover by reading
+from disk or invalidating cache entries.  For unrecoverable errors (meta data
+or dirty data), caching is automatically disabled; if dirty data was present
+in the cache it first disables writeback caching and waits for all dirty data
+to be flushed.
+
+Getting started:
+You'll need make-bcache from the bcache-tools repository. Both the cache device
+and backing device must be formatted before use.
+  make-bcache -B /dev/sdb
+  make-bcache -C /dev/sdc
+
+make-bcache has the ability to format multiple devices at the same time - if
+you format your backing devices and cache device at the same time, you won't
+have to manually attach:
+  make-bcache -B /dev/sda /dev/sdb -C /dev/sdc
+
+To make bcache devices known to the kernel, echo them to /sys/fs/bcache/register:
+
+  echo /dev/sdb > /sys/fs/bcache/register
+  echo /dev/sdc > /sys/fs/bcache/register
+
+To register your bcache devices automatically, you could add something like
+this to an init script:
+
+  echo /dev/sd* > /sys/fs/bcache/register_quiet
+
+It'll look for bcache superblocks and ignore everything that doesn't have one.
+
+Registering the backing device makes the bcache show up in /dev; you can now
+format it and use it as normal. But the first time using a new bcache device,
+it'll be running in passthrough mode until you attach it to a cache. See the
+section on attaching.
+
+The devices show up at /dev/bcacheN, and can be controlled via sysfs from
+/sys/block/bcacheN/bcache:
+
+  mkfs.ext4 /dev/bcache0
+  mount /dev/bcache0 /mnt
+
+Cache devices are managed as sets; multiple caches per set isn't supported yet
+but will allow for mirroring of metadata and dirty data in the future. Your new
+cache set shows up as /sys/fs/bcache/<UUID>
+
+ATTACHING:
+
+After your cache device and backing device are registered, the backing device
+must be attached to your cache set to enable caching. Attaching a backing
+device to a cache set is done thusly, with the UUID of the cache set in
+/sys/fs/bcache:
+
+  echo <UUID> > /sys/block/bcache0/bcache/attach
+
+This only has to be done once. The next time you reboot, just reregister all
+your bcache devices. If a backing device has data in a cache somewhere, the
+/dev/bcache# device won't be created until the cache shows up - particularly
+important if you have writeback caching turned on.
+
+If you're booting up and your cache device is gone and never coming back, you
+can force run the backing device:
+
+  echo 1 > /sys/block/sdb/bcache/running
+
+(You need to use /sys/block/sdb (or whatever your backing device is called), not
+/sys/block/bcache0, because bcache0 doesn't exist yet. If you're using a
+partition, the bcache directory would be at /sys/block/sdb/sdb2/bcache)
+
+The backing device will still use that cache set if it shows up in the future,
+but all the cached data will be invalidated. If there was dirty data in the
+cache, don't expect the filesystem to be recoverable - you will have massive
+filesystem corruption, though ext4's fsck does work miracles.
+
+ERROR HANDLING:
+
+Bcache tries to transparently handle IO errors to/from the cache device without
+affecting normal operation; if it sees too many errors (the threshold is
+configurable, and defaults to 0) it shuts down the cache device and switches all
+the backing devices to passthrough mode.
+
+ - For reads from the cache, if they error we just retry the read from the
+   backing device.
+
+ - For writethrough writes, if the write to the cache errors we just switch to
+   invalidating the data at that lba in the cache (i.e. the same thing we do for
+   a write that bypasses the cache)
+
+ - For writeback writes, we currently pass that error back up to the
+   filesystem/userspace. This could be improved - we could retry it as a write
+   that skips the cache so we don't have to error the write.
+
+ - When we detach, we first try to flush any dirty data (if we were running in
+   writeback mode). It currently doesn't do anything intelligent if it fails to
+   read some of the dirty data, though.
+
+TROUBLESHOOTING PERFORMANCE:
+
+Bcache has a bunch of config options and tunables. The defaults are intended to
+be reasonable for typical desktop and server workloads, but they're not what you
+want for getting the best possible numbers when benchmarking.
+
+ - Bad write performance
+
+   If write performance is not what you expected, you probably wanted to be
+   running in writeback mode, which isn't the default (not due to a lack of
+   maturity, but simply because in writeback mode you'll lose data if something
+   happens to your SSD)
+
+   # echo writeback > /sys/block/bcache0/cache_mode
+
+ - Bad performance, or traffic not going to the SSD that you'd expect
+
+   By default, bcache doesn't cache everything. It tries to skip sequential IO -
+   because you really want to be caching the random IO, and if you copy a 10
+   gigabyte file you probably don't want that pushing 10 gigabytes of randomly
+   accessed data out of your cache.
+
+   But if you want to benchmark reads from cache, and you start out with fio
+   writing an 8 gigabyte test file - so you want to disable that.
+
+   # echo 0 > /sys/block/bcache0/bcache/sequential_cutoff
+
+   To set it back to the default (4 mb), do
+
+   # echo 4M > /sys/block/bcache0/bcache/sequential_cutoff
+
+ - Traffic's still going to the spindle/still getting cache misses
+
+   In the real world, SSDs don't always keep up with disks - particularly with
+   slower SSDs, many disks being cached by one SSD, or mostly sequential IO. So
+   you want to avoid being bottlenecked by the SSD and having it slow everything
+   down.
+
+   To avoid that bcache tracks latency to the cache device, and gradually
+   throttles traffic if the latency exceeds a threshold (it does this by
+   cranking down the sequential bypass).
+
+   You can disable this if you need to by setting the thresholds to 0:
+
+   # echo 0 > /sys/fs/bcache/<cache set>/congested_read_threshold_us
+   # echo 0 > /sys/fs/bcache/<cache set>/congested_write_threshold_us
+
+   The default is 2000 us (2 milliseconds) for reads, and 20000 for writes.
+
+ - Still getting cache misses, of the same data
+
+   One last issue that sometimes trips people up is actually an old bug, due to
+   the way cache coherency is handled for cache misses. If a btree node is full,
+   a cache miss won't be able to insert a key for the new data and the data
+   won't be written to the cache.
+
+   In practice this isn't an issue because as soon as a write comes along it'll
+   cause the btree node to be split, and you need almost no write traffic for
+   this to not show up enough to be noticable (especially since bcache's btree
+   nodes are huge and index large regions of the device). But when you're
+   benchmarking, if you're trying to warm the cache by reading a bunch of data
+   and there's no other traffic - that can be a problem.
+
+   Solution: warm the cache by doing writes, or use the testing branch (there's
+   a fix for the issue there).
+
+SYSFS - BACKING DEVICE:
+
+attach
+  Echo the UUID of a cache set to this file to enable caching.
+
+cache_mode
+  Can be one of either writethrough, writeback, writearound or none.
+
+clear_stats
+  Writing to this file resets the running total stats (not the day/hour/5 minute
+  decaying versions).
+
+detach
+  Write to this file to detach from a cache set. If there is dirty data in the
+  cache, it will be flushed first.
+
+dirty_data
+  Amount of dirty data for this backing device in the cache. Continuously
+  updated unlike the cache set's version, but may be slightly off.
+
+label
+  Name of underlying device.
+
+readahead
+  Size of readahead that should be performed.  Defaults to 0.  If set to e.g.
+  1M, it will round cache miss reads up to that size, but without overlapping
+  existing cache entries.
+
+running
+  1 if bcache is running (i.e. whether the /dev/bcache device exists, whether
+  it's in passthrough mode or caching).
+
+sequential_cutoff
+  A sequential IO will bypass the cache once it passes this threshhold; the
+  most recent 128 IOs are tracked so sequential IO can be detected even when
+  it isn't all done at once.
+
+sequential_merge
+  If non zero, bcache keeps a list of the last 128 requests submitted to compare
+  against all new requests to determine which new requests are sequential
+  continuations of previous requests for the purpose of determining sequential
+  cutoff. This is necessary if the sequential cutoff value is greater than the
+  maximum acceptable sequential size for any single request. 
+
+state
+  The backing device can be in one of four different states:
+
+  no cache: Has never been attached to a cache set.
+
+  clean: Part of a cache set, and there is no cached dirty data.
+
+  dirty: Part of a cache set, and there is cached dirty data.
+
+  inconsistent: The backing device was forcibly run by the user when there was
+  dirty data cached but the cache set was unavailable; whatever data was on the
+  backing device has likely been corrupted.
+
+stop
+  Write to this file to shut down the bcache device and close the backing
+  device.
+
+writeback_delay
+  When dirty data is written to the cache and it previously did not contain
+  any, waits some number of seconds before initiating writeback. Defaults to
+  30.
+
+writeback_percent
+  If nonzero, bcache tries to keep around this percentage of the cache dirty by
+  throttling background writeback and using a PD controller to smoothly adjust
+  the rate.
+
+writeback_rate
+  Rate in sectors per second - if writeback_percent is nonzero, background
+  writeback is throttled to this rate. Continuously adjusted by bcache but may
+  also be set by the user.
+
+writeback_running
+  If off, writeback of dirty data will not take place at all. Dirty data will
+  still be added to the cache until it is mostly full; only meant for
+  benchmarking. Defaults to on.
+
+SYSFS - BACKING DEVICE STATS:
+
+There are directories with these numbers for a running total, as well as
+versions that decay over the past day, hour and 5 minutes; they're also
+aggregated in the cache set directory as well.
+
+bypassed
+  Amount of IO (both reads and writes) that has bypassed the cache
+
+cache_hits
+cache_misses
+cache_hit_ratio
+  Hits and misses are counted per individual IO as bcache sees them; a
+  partial hit is counted as a miss.
+
+cache_bypass_hits
+cache_bypass_misses
+  Hits and misses for IO that is intended to skip the cache are still counted,
+  but broken out here.
+
+cache_miss_collisions
+  Counts instances where data was going to be inserted into the cache from a
+  cache miss, but raced with a write and data was already present (usually 0
+  since the synchronization for cache misses was rewritten)
+
+cache_readaheads
+  Count of times readahead occured.
+
+SYSFS - CACHE SET:
+
+average_key_size
+  Average data per key in the btree.
+
+bdev<0..n>
+  Symlink to each of the attached backing devices.
+
+block_size
+  Block size of the cache devices.
+
+btree_cache_size
+  Amount of memory currently used by the btree cache
+
+bucket_size
+  Size of buckets
+
+cache<0..n>
+  Symlink to each of the cache devices comprising this cache set. 
+
+cache_available_percent
+  Percentage of cache device free.
+
+clear_stats
+  Clears the statistics associated with this cache
+
+dirty_data
+  Amount of dirty data is in the cache (updated when garbage collection runs).
+
+flash_vol_create
+  Echoing a size to this file (in human readable units, k/M/G) creates a thinly
+  provisioned volume backed by the cache set.
+
+io_error_halflife
+io_error_limit
+  These determines how many errors we accept before disabling the cache.
+  Each error is decayed by the half life (in # ios).  If the decaying count
+  reaches io_error_limit dirty data is written out and the cache is disabled.
+
+journal_delay_ms
+  Journal writes will delay for up to this many milliseconds, unless a cache
+  flush happens sooner. Defaults to 100.
+
+root_usage_percent
+  Percentage of the root btree node in use.  If this gets too high the node
+  will split, increasing the tree depth.
+
+stop
+  Write to this file to shut down the cache set - waits until all attached
+  backing devices have been shut down.
+
+tree_depth
+  Depth of the btree (A single node btree has depth 0).
+
+unregister
+  Detaches all backing devices and closes the cache devices; if dirty data is
+  present it will disable writeback caching and wait for it to be flushed.
+
+SYSFS - CACHE SET INTERNAL:
+
+This directory also exposes timings for a number of internal operations, with
+separate files for average duration, average frequency, last occurence and max
+duration: garbage collection, btree read, btree node sorts and btree splits.
+
+active_journal_entries
+  Number of journal entries that are newer than the index.
+
+btree_nodes
+  Total nodes in the btree.
+
+btree_used_percent
+  Average fraction of btree in use.
+
+bset_tree_stats
+  Statistics about the auxiliary search trees
+
+btree_cache_max_chain
+  Longest chain in the btree node cache's hash table
+
+cache_read_races
+  Counts instances where while data was being read from the cache, the bucket
+  was reused and invalidated - i.e. where the pointer was stale after the read
+  completed. When this occurs the data is reread from the backing device.
+
+trigger_gc
+  Writing to this file forces garbage collection to run.
+
+SYSFS - CACHE DEVICE:
+
+block_size
+  Minimum granularity of writes - should match hardware sector size.
+
+btree_written
+  Sum of all btree writes, in (kilo/mega/giga) bytes
+
+bucket_size
+  Size of buckets
+
+cache_replacement_policy
+  One of either lru, fifo or random.
+
+discard
+  Boolean; if on a discard/TRIM will be issued to each bucket before it is
+  reused. Defaults to off, since SATA TRIM is an unqueued command (and thus
+  slow).
+
+freelist_percent
+  Size of the freelist as a percentage of nbuckets. Can be written to to
+  increase the number of buckets kept on the freelist, which lets you
+  artificially reduce the size of the cache at runtime. Mostly for testing
+  purposes (i.e. testing how different size caches affect your hit rate), but
+  since buckets are discarded when they move on to the freelist will also make
+  the SSD's garbage collection easier by effectively giving it more reserved
+  space.
+
+io_errors
+  Number of errors that have occured, decayed by io_error_halflife.
+
+metadata_written
+  Sum of all non data writes (btree writes and all other metadata).
+
+nbuckets
+  Total buckets in this cache
+
+priority_stats
+  Statistics about how recently data in the cache has been accessed.  This can
+  reveal your working set size.
+
+written
+  Sum of all data that has been written to the cache; comparison with
+  btree_written gives the amount of write inflation in bcache.

MAINTAINERS

···
 S:	Maintained
 F:	drivers/net/hamradio/baycom*
 
+BCACHE (BLOCK LAYER CACHE)
+M:	Kent Overstreet <[e-mail]>
+L:	[e-mail]
+W:	http://bcache.evilpiepirate.org
+S:	Maintained:
+F:	drivers/md/bcache/
+
 BEFS FILE SYSTEM
 S:	Orphan
 F:	Documentation/filesystems/befs.txt

drivers/block/aoe/aoecmd.c

···
 static void
 bufinit(struct buf *buf, struct request *rq, struct bio *bio)
 {
-	struct bio_vec *bv;
-
 	memset(buf, 0, sizeof(*buf));
 	buf->rq = rq;
 	buf->bio = bio;
 	buf->resid = bio->bi_size;
 	buf->sector = bio->bi_sector;
 	bio_pageinc(bio);
-	buf->bv = bv = bio_iovec(bio);
-	buf->bv_resid = bv->bv_len;
+	buf->bv = bio_iovec(bio);
+	buf->bv_resid = buf->bv->bv_len;
 	WARN_ON(buf->bv_resid == 0);
 }
 

drivers/block/cciss.c

···
 MODULE_PARM_DESC(cciss_simple_mode,
 	"Use 'simple mode' rather than 'performant mode'");
 
+static int cciss_allow_hpsa;
+module_param(cciss_allow_hpsa, int, S_IRUGO|S_IWUSR);
+MODULE_PARM_DESC(cciss_allow_hpsa,
+	"Prevent cciss driver from accessing hardware known to be "
+	" supported by the hpsa driver");
+
 static DEFINE_MUTEX(cciss_mutex);
 static struct proc_dir_entry *proc_cciss;
 
···
 	*board_id = ((subsystem_device_id << 16) & 0xffff0000) |
 			subsystem_vendor_id;
 
-	for (i = 0; i < ARRAY_SIZE(products); i++)
+	for (i = 0; i < ARRAY_SIZE(products); i++) {
+		/* Stand aside for hpsa driver on request */
+		if (cciss_allow_hpsa)
+			return -ENODEV;
 		if (*board_id == products[i].board_id)
 			return i;
+	}
 	dev_warn(&pdev->dev, "unrecognized board ID: 0x%08x, ignoring.\n",
 		*board_id);
 	return -ENODEV;
···
 	ctlr_info_t *h;
 	unsigned long flags;
 
+	/*
+	 * By default the cciss driver is used for all older HP Smart Array
+	 * controllers. There are module paramaters that allow a user to
+	 * override this behavior and instead use the hpsa SCSI driver. If
+	 * this is the case cciss may be loaded first from the kdump initrd
+	 * image and cause a kernel panic. So if reset_devices is true and
+	 * cciss_allow_hpsa is set just bail.
+	 */
+	if ((reset_devices) && (cciss_allow_hpsa == 1))
+		return -ENODEV;
 	rc = cciss_init_reset_devices(pdev);
 	if (rc) {
 		if (rc != -ENOTSUPP)

drivers/block/drbd/drbd_actlog.c

···
 	int err;
 };
 
-static int al_write_transaction(struct drbd_conf *mdev);
 
 void *drbd_md_get_buffer(struct drbd_conf *mdev)
 {
···
 	bio->bi_end_io = drbd_md_io_complete;
 	bio->bi_rw = rw;
 
-	if (!get_ldev_if_state(mdev, D_ATTACHING)) {  /* Corresponding put_ldev in drbd_md_io_complete() */
+	if (!(rw & WRITE) && mdev->state.disk == D_DISKLESS && mdev->ldev == NULL)
+		/* special case, drbd_md_read() during drbd_adm_attach(): no get_ldev */
+		;
+	else if (!get_ldev_if_state(mdev, D_ATTACHING)) {
+		/* Corresponding put_ldev in drbd_md_io_complete() */
 		dev_err(DEV, "ASSERT FAILED: get_ldev_if_state() == 1 in _drbd_md_sync_page_io()\n");
 		err = -ENODEV;
 		goto out;
···
 
 	BUG_ON(!bdev->md_bdev);
 
-	dev_dbg(DEV, "meta_data io: %s [%d]:%s(,%llus,%s)\n",
+	dev_dbg(DEV, "meta_data io: %s [%d]:%s(,%llus,%s) %pS\n",
 	     current->comm, current->pid, __func__,
-	     (unsigned long long)sector, (rw & WRITE) ? "WRITE" : "READ");
+	     (unsigned long long)sector, (rw & WRITE) ? "WRITE" : "READ",
+	     (void*)_RET_IP_ );
 
 	if (sector < drbd_md_first_sector(bdev) ||
 	    sector + 7 > drbd_md_last_sector(bdev))
···
 		     current->comm, current->pid, __func__,
 		     (unsigned long long)sector, (rw & WRITE) ? "WRITE" : "READ");
 
-	err = _drbd_md_sync_page_io(mdev, bdev, iop, sector, rw, MD_BLOCK_SIZE);
+	/* we do all our meta data IO in aligned 4k blocks. */
+	err = _drbd_md_sync_page_io(mdev, bdev, iop, sector, rw, 4096);
 	if (err) {
 		dev_err(DEV, "drbd_md_sync_page_io(,%llus,%s) failed with error %d\n",
 		    (unsigned long long)sector, (rw & WRITE) ? "WRITE" : "READ", err);
···
 	return err;
 }
 
-static struct lc_element *_al_get(struct drbd_conf *mdev, unsigned int enr)
+static struct bm_extent *find_active_resync_extent(struct drbd_conf *mdev, unsigned int enr)
 {
-	struct lc_element *al_ext;
 	struct lc_element *tmp;
-	int wake;
-
-	spin_lock_irq(&mdev->al_lock);
 	tmp = lc_find(mdev->resync, enr/AL_EXT_PER_BM_SECT);
 	if (unlikely(tmp != NULL)) {
 		struct bm_extent  *bm_ext = lc_entry(tmp, struct bm_extent, lce);
-		if (test_bit(BME_NO_WRITES, &bm_ext->flags)) {
-			wake = !test_and_set_bit(BME_PRIORITY, &bm_ext->flags);
-			spin_unlock_irq(&mdev->al_lock);
-			if (wake)
-				wake_up(&mdev->al_wait);
-			return NULL;
-		}
+		if (test_bit(BME_NO_WRITES, &bm_ext->flags))
+			return bm_ext;
 	}
-	al_ext = lc_get(mdev->act_log, enr);
+	return NULL;
+}
+
+static struct lc_element *_al_get(struct drbd_conf *mdev, unsigned int enr, bool nonblock)
+{
+	struct lc_element *al_ext;
+	struct bm_extent *bm_ext;
+	int wake;
+
+	spin_lock_irq(&mdev->al_lock);
+	bm_ext = find_active_resync_extent(mdev, enr);
+	if (bm_ext) {
+		wake = !test_and_set_bit(BME_PRIORITY, &bm_ext->flags);
+		spin_unlock_irq(&mdev->al_lock);
+		if (wake)
+			wake_up(&mdev->al_wait);
+		return NULL;
+	}
+	if (nonblock)
+		al_ext = lc_try_get(mdev->act_log, enr);
+	else
+		al_ext = lc_get(mdev->act_log, enr);
 	spin_unlock_irq(&mdev->al_lock);
 	return al_ext;
 }
 
-void drbd_al_begin_io(struct drbd_conf *mdev, struct drbd_interval *i)
+bool drbd_al_begin_io_fastpath(struct drbd_conf *mdev, struct drbd_interval *i)
+{
+	/* for bios crossing activity log extent boundaries,
+	 * we may need to activate two extents in one go */
+	unsigned first = i->sector >> (AL_EXTENT_SHIFT-9);
+	unsigned last = i->size == 0 ? first : (i->sector + (i->size >> 9) - 1) >> (AL_EXTENT_SHIFT-9);
+
+	D_ASSERT((unsigned)(last - first) <= 1);
+	D_ASSERT(atomic_read(&mdev->local_cnt) > 0);
+
+	/* FIXME figure out a fast path for bios crossing AL extent boundaries */
+	if (first != last)
+		return false;
+
+	return _al_get(mdev, first, true);
+}
+
+bool drbd_al_begin_io_prepare(struct drbd_conf *mdev, struct drbd_interval *i)
 {
 	/* for bios crossing activity log extent boundaries,
 	 * we may need to activate two extents in one go */
 	unsigned first = i->sector >> (AL_EXTENT_SHIFT-9);
 	unsigned last = i->size == 0 ? first : (i->sector + (i->size >> 9) - 1) >> (AL_EXTENT_SHIFT-9);
 	unsigned enr;
-	bool locked = false;
-
+	bool need_transaction = false;
 
 	D_ASSERT(first <= last);
 	D_ASSERT(atomic_read(&mdev->local_cnt) > 0);
 
-	for (enr = first; enr <= last; enr++)
-		wait_event(mdev->al_wait, _al_get(mdev, enr) != NULL);
+	for (enr = first; enr <= last; enr++) {
+		struct lc_element *al_ext;
+		wait_event(mdev->al_wait,
+				(al_ext = _al_get(mdev, enr, false)) != NULL);
+		if (al_ext->lc_number != enr)
+			need_transaction = true;
+	}
+	return need_transaction;
+}
+
+static int al_write_transaction(struct drbd_conf *mdev, bool delegate);
+
+/* When called through generic_make_request(), we must delegate
+ * activity log I/O to the worker thread: a further request
+ * submitted via generic_make_request() within the same task
+ * would be queued on current->bio_list, and would only start
+ * after this function returns (see generic_make_request()).
+ *
+ * However, if we *are* the worker, we must not delegate to ourselves.
+ */
+
+/*
+ * @delegate:   delegate activity log I/O to the worker thread
+ */
+void drbd_al_begin_io_commit(struct drbd_conf *mdev, bool delegate)
+{
+	bool locked = false;
+
+	BUG_ON(delegate && current == mdev->tconn->worker.task);
 
 	/* Serialize multiple transactions.
 	 * This uses test_and_set_bit, memory barrier is implicit.
···
 			(locked = lc_try_lock_for_transaction(mdev->act_log)));
 
 	if (locked) {
-		/* drbd_al_write_transaction(mdev,al_ext,enr);
-		 * recurses into generic_make_request(), which
-		 * disallows recursion, bios being serialized on the
-		 * current->bio_tail list now.
-		 * we have to delegate updates to the activity log
-		 * to the worker thread. */
-
 		/* Double check: it may have been committed by someone else,
 		 * while we have been waiting for the lock. */
 		if (mdev->act_log->pending_changes) {
···
 			write_al_updates = rcu_dereference(mdev->ldev->disk_conf)->al_updates;
 			rcu_read_unlock();
 
-			if (write_al_updates) {
-				al_write_transaction(mdev);
-				mdev->al_writ_cnt++;
-			}
-
+			if (write_al_updates)
+				al_write_transaction(mdev, delegate);
 			spin_lock_irq(&mdev->al_lock);
 			/* FIXME
 			if (err)
···
 		lc_unlock(mdev->act_log);
 		wake_up(&mdev->al_wait);
 	}
+}
+
+/*
+ * @delegate:   delegate activity log I/O to the worker thread
+ */
+void drbd_al_begin_io(struct drbd_conf *mdev, struct drbd_interval *i, bool delegate)
+{
+	BUG_ON(delegate && current == mdev->tconn->worker.task);
+
+	if (drbd_al_begin_io_prepare(mdev, i))
+		drbd_al_begin_io_commit(mdev, delegate);
+}
+
+int drbd_al_begin_io_nonblock(struct drbd_conf *mdev, struct drbd_interval *i)
+{
+	struct lru_cache *al = mdev->act_log;
+	/* for bios crossing activity log extent boundaries,
+	 * we may need to activate two extents in one go */
+	unsigned first = i->sector >> (AL_EXTENT_SHIFT-9);
+	unsigned last = i->size == 0 ? first : (i->sector + (i->size >> 9) - 1) >> (AL_EXTENT_SHIFT-9);
+	unsigned nr_al_extents;
+	unsigned available_update_slots;
+	unsigned enr;
+
+	D_ASSERT(first <= last);
+
+	nr_al_extents = 1 + last - first; /* worst case: all touched extends are cold. */
+	available_update_slots = min(al->nr_elements - al->used,
+				al->max_pending_changes - al->pending_changes);
+
+	/* We want all necessary updates for a given request within the same transaction
+	 * We could first check how many updates are *actually* needed,
+	 * and use that instead of the worst-case nr_al_extents */
+	if (available_update_slots < nr_al_extents)
+		return -EWOULDBLOCK;
+
+	/* Is resync active in this area? */
+	for (enr = first; enr <= last; enr++) {
+		struct lc_element *tmp;
+		tmp = lc_find(mdev->resync, enr/AL_EXT_PER_BM_SECT);
+		if (unlikely(tmp != NULL)) {
+			struct bm_extent  *bm_ext = lc_entry(tmp, struct bm_extent, lce);
+			if (test_bit(BME_NO_WRITES, &bm_ext->flags)) {
+				if (!test_and_set_bit(BME_PRIORITY, &bm_ext->flags))
+					return -EBUSY;
+				return -EWOULDBLOCK;
+			}
+		}
+	}
+
+	/* Checkout the refcounts.
+	 * Given that we checked for available elements and update slots above,
+	 * this has to be successful. */
+	for (enr = first; enr <= last; enr++) {
+		struct lc_element *al_ext;
+		al_ext = lc_get_cumulative(mdev->act_log, enr);
+		if (!al_ext)
+			dev_info(DEV, "LOGIC BUG for enr=%u\n", enr);
+	}
+	return 0;
 }
 
 void drbd_al_complete_io(struct drbd_conf *mdev, struct drbd_interval *i)
···
 		((PAGE_SHIFT + 3) -
 		/* resync extent number to bit */
 		 (BM_EXT_SHIFT - BM_BLOCK_SHIFT));
+}
+
+static sector_t al_tr_number_to_on_disk_sector(struct drbd_conf *mdev)
+{
+	const unsigned int stripes = mdev->ldev->md.al_stripes;
+	const unsigned int stripe_size_4kB = mdev->ldev->md.al_stripe_size_4k;
+
+	/* transaction number, modulo on-disk ring buffer wrap around */
+	unsigned int t = mdev->al_tr_number % (mdev->ldev->md.al_size_4k);
+
+	/* ... to aligned 4k on disk block */
+	t = ((t % stripes) * stripe_size_4kB) + t/stripes;
+
+	/* ... to 512 byte sector in activity log */
+	t *= 8;
+
+	/* ... plus offset to the on disk position */
+	return mdev->ldev->md.md_offset + mdev->ldev->md.al_offset + t;
 }
 
 static int
···
 	if (mdev->al_tr_cycle >= mdev->act_log->nr_elements)
 		mdev->al_tr_cycle = 0;
 
-	sector =  mdev->ldev->md.md_offset
-		+ mdev->ldev->md.al_offset
-		+ mdev->al_tr_pos * (MD_BLOCK_SIZE>>9);
+	sector = al_tr_number_to_on_disk_sector(mdev);
 
 	crc = crc32c(0, buffer, 4096);
 	buffer->crc32c = cpu_to_be32(crc);
 
 	if (drbd_bm_write_hinted(mdev))
 		err = -EIO;
-		/* drbd_chk_io_error done already */
-	else if (drbd_md_sync_page_io(mdev, mdev->ldev, sector, WRITE)) {
-		err = -EIO;
-		drbd_chk_io_error(mdev, 1, DRBD_META_IO_ERROR);
-	} else {
-		/* advance ringbuffer position and transaction counter */
-		mdev->al_tr_pos = (mdev->al_tr_pos + 1) % (MD_AL_SECTORS*512/MD_BLOCK_SIZE);
-		mdev->al_tr_number++;
+	else {
+		bool write_al_updates;
+		rcu_read_lock();
+		write_al_updates = rcu_dereference(mdev->ldev->disk_conf)->al_updates;
+		rcu_read_unlock();
+		if (write_al_updates) {
+			if (drbd_md_sync_page_io(mdev, mdev->ldev, sector, WRITE)) {
+				err = -EIO;
+				drbd_chk_io_error(mdev, 1, DRBD_META_IO_ERROR);
+			} else {
+				mdev->al_tr_number++;
+				mdev->al_writ_cnt++;
+			}
+		}
 	}
 
 	drbd_md_put_buffer(mdev);
···
 /* Calls from worker context (see w_restart_disk_io()) need to write the
    transaction directly. Others came through generic_make_request(),
    those need to delegate it to the worker. */
-static int al_write_transaction(struct drbd_conf *mdev)
+static int al_write_transaction(struct drbd_conf *mdev, bool delegate)
 {
-	struct update_al_work al_work;
-
-	if (current == mdev->tconn->worker.task)
+	if (delegate) {
+		struct update_al_work al_work;
+		init_completion(&al_work.event);
+		al_work.w.cb = w_al_write_transaction;
+		al_work.w.mdev = mdev;
+		drbd_queue_work_front(&mdev->tconn->sender_work, &al_work.w);
+		wait_for_completion(&al_work.event);
+		return al_work.err;
+	} else
 		return _al_write_transaction(mdev);
-
-	init_completion(&al_work.event);
-	al_work.w.cb = w_al_write_transaction;
-	al_work.w.mdev = mdev;
-	drbd_queue_work_front(&mdev->tconn->sender_work, &al_work.w);
-	wait_for_completion(&al_work.event);
-
-	return al_work.err;
 }
 
 static int _try_lc_del(struct drbd_conf *mdev, struct lc_element *al_ext)

drivers/block/drbd/drbd_bitmap.c

···
 	}
 }
 
+/* For the layout, see comment above drbd_md_set_sector_offsets(). */
+static u64 drbd_md_on_disk_bits(struct drbd_backing_dev *ldev)
+{
+	u64 bitmap_sectors;
+	if (ldev->md.al_offset == 8)
+		bitmap_sectors = ldev->md.md_size_sect - ldev->md.bm_offset;
+	else
+		bitmap_sectors = ldev->md.al_offset - ldev->md.bm_offset;
+	return bitmap_sectors << (9 + 3);
+}
+
 /*
  * make sure the bitmap has enough room for the attached storage,
  * if necessary, resize.
···
 	words = ALIGN(bits, 64) >> LN2_BPL;
 
 	if (get_ldev(mdev)) {
-		u64 bits_on_disk = ((u64)mdev->ldev->md.md_size_sect-MD_BM_OFFSET) << 12;
+		u64 bits_on_disk = drbd_md_on_disk_bits(mdev->ldev);
 		put_ldev(mdev);
 		if (bits > bits_on_disk) {
 			dev_info(DEV, "bits = %lu\n", bits);

drivers/block/drbd/drbd_int.h

···
 	u32 flags;
 	u32 md_size_sect;
 
-	s32 al_offset;	/* signed relative sector offset to al area */
+	s32 al_offset;	/* signed relative sector offset to activity log */
 	s32 bm_offset;	/* signed relative sector offset to bitmap */
 
-	/* u32 al_nr_extents;	   important for restoring the AL
-	 * is stored into  ldev->dc.al_extents, which in turn
-	 * gets applied to act_log->nr_elements
-	 */
+	/* cached value of bdev->disk_conf->meta_dev_idx (see below) */
+	s32 meta_dev_idx;
+
+	/* see al_tr_number_to_on_disk_sector() */
+	u32 al_stripes;
+	u32 al_stripe_size_4k;
+	u32 al_size_4k; /* cached product of the above */
 };
 
 struct drbd_backing_dev {
···
 	} send;
 };
 
+struct submit_worker {
+	struct workqueue_struct *wq;
+	struct work_struct worker;
+
+	spinlock_t lock;
+	struct list_head writes;
+};
+
 struct drbd_conf {
 	struct drbd_tconn *tconn;
 	int vnr;			/* volume number within the connection */
···
 	struct lru_cache *act_log;	/* activity log */
 	unsigned int al_tr_number;
 	int al_tr_cycle;
-	int al_tr_pos;   /* position of the next transaction in the journal */
 	wait_queue_head_t seq_wait;
 	atomic_t packet_seq;
 	unsigned int peer_seq;
···
 	atomic_t ap_in_flight; /* App sectors in flight (waiting for ack) */
 	unsigned int peer_max_bio_size;
 	unsigned int local_max_bio_size;
+
+	/* any requests that would block in drbd_make_request()
+	 * are deferred to this single-threaded work queue */
+	struct submit_worker submit;
 };
 
 static inline struct drbd_conf *minor_to_mdev(unsigned int minor)
···
 		char *why, enum bm_flag flags);
 extern int drbd_bmio_set_n_write(struct drbd_conf *mdev);
 extern int drbd_bmio_clear_n_write(struct drbd_conf *mdev);
-extern void drbd_go_diskless(struct drbd_conf *mdev);
 extern void drbd_ldev_destroy(struct drbd_conf *mdev);
 
 /* Meta data layout
-   We reserve a 128MB Block (4k aligned)
-   * either at the end of the backing device
-   * or on a separate meta data device. */
+ *
+ * We currently have two possible layouts.
+ * Offsets in (512 byte) sectors.
+ * external:
+ *   |----------- md_size_sect ------------------|
+ *   [ 4k superblock ][ activity log ][  Bitmap  ]
+ *   | al_offset == 8 |
+ *   | bm_offset = al_offset + X      |
+ *  ==> bitmap sectors = md_size_sect - bm_offset
+ *
+ *  Variants:
+ *     old, indexed fixed size meta data:
+ *
+ * internal:
+ *            |----------- md_size_sect ------------------|
+ * [data.....][  Bitmap  ][ activity log ][ 4k superblock ][padding*]
+ *                        | al_offset < 0 |
+ *            | bm_offset = al_offset - Y |
+ *  ==> bitmap sectors = Y = al_offset - bm_offset
+ *
+ *  [padding*] are zero or up to 7 unused 512 Byte sectors to the
+ *  end of the device, so that the [4k superblock] will be 4k aligned.
+ *
+ *  The activity log consists of 4k transaction blocks,
+ *  which are written in a ring-buffer, or striped ring-buffer like fashion,
+ *  which are writtensize used to be fixed 32kB,
+ *  but is about to become configurable.
+ */
 
-/* The following numbers are sectors */
-/* Allows up to about 3.8TB, so if you want more,
+/* Our old fixed size meta data layout
+ * allows up to about 3.8TB, so if you want more,
  * you need to use the "flexible" meta data format. */
-#define MD_RESERVED_SECT (128LU << 11)  /* 128 MB, unit sectors */
-#define MD_AL_OFFSET	8    /* 8 Sectors after start of meta area */
-#define MD_AL_SECTORS	64   /* = 32 kB on disk activity log ring buffer */
-#define MD_BM_OFFSET (MD_AL_OFFSET + MD_AL_SECTORS)
-
-/* we do all meta data IO in 4k blocks */
-#define MD_BLOCK_SHIFT	12
-#define MD_BLOCK_SIZE	(1<<MD_BLOCK_SHIFT)
+#define MD_128MB_SECT (128LLU << 11)  /* 128 MB, unit sectors */
+#define MD_4kB_SECT	 8
+#define MD_32kB_SECT	64
 
 /* One activity log extent represents 4M of storage */
 #define AL_EXTENT_SHIFT 22
···
 
 /* in one sector of the bitmap, we have this many activity_log extents. */
 #define AL_EXT_PER_BM_SECT  (1 << (BM_EXT_SHIFT - AL_EXTENT_SHIFT))
-#define BM_WORDS_PER_AL_EXT (1 << (AL_EXTENT_SHIFT-BM_BLOCK_SHIFT-LN2_BPL))
 
 #define BM_BLOCKS_PER_BM_EXT_B (BM_EXT_SHIFT - BM_BLOCK_SHIFT)
 #define BM_BLOCKS_PER_BM_EXT_MASK  ((1<<BM_BLOCKS_PER_BM_EXT_B) - 1)
···
  */
 
 #define DRBD_MAX_SECTORS_32 (0xffffffffLU)
-#define DRBD_MAX_SECTORS_BM \
-	  ((MD_RESERVED_SECT - MD_BM_OFFSET) * (1LL<<(BM_EXT_SHIFT-9)))
-#if DRBD_MAX_SECTORS_BM < DRBD_MAX_SECTORS_32
-#define DRBD_MAX_SECTORS      DRBD_MAX_SECTORS_BM
-#define DRBD_MAX_SECTORS_FLEX DRBD_MAX_SECTORS_BM
-#elif !defined(CONFIG_LBDAF) && BITS_PER_LONG == 32
+/* we have a certain meta data variant that has a fixed on-disk size of 128
+ * MiB, of which 4k are our "superblock", and 32k are the fixed size activity
+ * log, leaving this many sectors for the bitmap.
+ */
+
+#define DRBD_MAX_SECTORS_FIXED_BM \
+	  ((MD_128MB_SECT - MD_32kB_SECT - MD_4kB_SECT) * (1LL<<(BM_EXT_SHIFT-9)))
+#if !defined(CONFIG_LBDAF) && BITS_PER_LONG == 32
 #define DRBD_MAX_SECTORS      DRBD_MAX_SECTORS_32
 #define DRBD_MAX_SECTORS_FLEX DRBD_MAX_SECTORS_32
 #else
-#define DRBD_MAX_SECTORS      DRBD_MAX_SECTORS_BM
+#define DRBD_MAX_SECTORS      DRBD_MAX_SECTORS_FIXED_BM
 /* 16 TB in units of sectors */
 #if BITS_PER_LONG == 32
 /* adjust by one page worth of bitmap,
···
 extern int proc_details;
 
 /* drbd_req */
+extern void do_submit(struct work_struct *ws);
 extern void __drbd_make_request(struct drbd_conf *, struct bio *, unsigned long);
 extern void drbd_make_request(struct request_queue *q, struct bio *bio);
 extern int drbd_read_remote(struct drbd_conf *mdev, struct drbd_request *req);
···
 extern const char *drbd_role_str(enum drbd_role s);
 
 /* drbd_actlog.c */
-extern void drbd_al_begin_io(struct drbd_conf *mdev, struct drbd_interval *i);
+extern int drbd_al_begin_io_nonblock(struct drbd_conf *mdev, struct drbd_interval *i);
+extern void drbd_al_begin_io_commit(struct drbd_conf *mdev, bool delegate);
+extern bool drbd_al_begin_io_fastpath(struct drbd_conf *mdev, struct drbd_interval *i);
+extern void drbd_al_begin_io(struct drbd_conf *mdev, struct drbd_interval *i, bool delegate);
 extern void drbd_al_complete_io(struct drbd_conf *mdev, struct drbd_interval *i);
 extern void drbd_rs_complete_io(struct drbd_conf *mdev, sector_t sector);
 extern int drbd_rs_begin_io(struct drbd_conf *mdev, sector_t sector);
···
  * BTW, for internal meta data, this happens to be the maximum capacity
  * we could agree upon with our peer node.
  */
-static inline sector_t _drbd_md_first_sector(int meta_dev_idx, struct drbd_backing_dev *bdev)
+static inline sector_t drbd_md_first_sector(struct drbd_backing_dev *bdev)
 {
-	switch (meta_dev_idx) {
+	switch (bdev->md.meta_dev_idx) {
 	case DRBD_MD_INDEX_INTERNAL:
 	case DRBD_MD_INDEX_FLEX_INT:
 		return bdev->md.md_offset + bdev->md.bm_offset;
···
 	}
 }
 
-static inline sector_t drbd_md_first_sector(struct drbd_backing_dev *bdev)
-{
-	int meta_dev_idx;
-
-	rcu_read_lock();
-	meta_dev_idx = rcu_dereference(bdev->disk_conf)->meta_dev_idx;
-	rcu_read_unlock();
-
-	return _drbd_md_first_sector(meta_dev_idx, bdev);
-}
-
 /**
  * drbd_md_last_sector() - Return the last sector number of the meta data area
  * @bdev:	Meta data block device.
  */
 static inline sector_t drbd_md_last_sector(struct drbd_backing_dev *bdev)
 {
-	int meta_dev_idx;
-
-	rcu_read_lock();
-	meta_dev_idx = rcu_dereference(bdev->disk_conf)->meta_dev_idx;
-	rcu_read_unlock();
-
-	switch (meta_dev_idx) {
+	switch (bdev->md.meta_dev_idx) {
 	case DRBD_MD_INDEX_INTERNAL:
 	case DRBD_MD_INDEX_FLEX_INT:
-		return bdev->md.md_offset + MD_AL_OFFSET - 1;
+		return bdev->md.md_offset + MD_4kB_SECT -1;
 	case DRBD_MD_INDEX_FLEX_EXT:
 	default:
-		return bdev->md.md_offset + bdev->md.md_size_sect;
+		return bdev->md.md_offset + bdev->md.md_size_sect -1;
 	}
 }
 
···
 static inline sector_t drbd_get_max_capacity(struct drbd_backing_dev *bdev)
 {
 	sector_t s;
-	int meta_dev_idx;
 
-	rcu_read_lock();
-	meta_dev_idx = rcu_dereference(bdev->disk_conf)->meta_dev_idx;
-	rcu_read_unlock();
-
-	switch (meta_dev_idx) {
+	switch (bdev->md.meta_dev_idx) {
 	case DRBD_MD_INDEX_INTERNAL:
 	case DRBD_MD_INDEX_FLEX_INT:
 		s = drbd_get_capacity(bdev->backing_bdev)
 			? min_t(sector_t, DRBD_MAX_SECTORS_FLEX,
-				_drbd_md_first_sector(meta_dev_idx, bdev))
+				drbd_md_first_sector(bdev))
 			: 0;
 		break;
 	case DRBD_MD_INDEX_FLEX_EXT:
···
 }
 
 /**
- * drbd_md_ss__() - Return the sector number of our meta data super block
- * @mdev:	DRBD device.
+ * drbd_md_ss() - Return the sector number of our meta data super block
  * @bdev:	Meta data block device.
  */
-static inline sector_t drbd_md_ss__(struct drbd_conf *mdev,
-				    struct drbd_backing_dev *bdev)
+static inline sector_t drbd_md_ss(struct drbd_backing_dev *bdev)
 {
-	int meta_dev_idx;
+	const int meta_dev_idx = bdev->md.meta_dev_idx;
 
-	rcu_read_lock();
-	meta_dev_idx = rcu_dereference(bdev->disk_conf)->meta_dev_idx;
-	rcu_read_unlock();
-
-	switch (meta_dev_idx) {
-	default: /* external, some index */
-		return MD_RESERVED_SECT * meta_dev_idx;
-	case DRBD_MD_INDEX_INTERNAL:
-		/* with drbd08, internal meta data is always "flexible" */
-	case DRBD_MD_INDEX_FLEX_INT:
-		/* sizeof(struct md_on_disk_07) == 4k
-		 * position: last 4k aligned block of 4k size */
-		if (!bdev->backing_bdev) {
-			if (__ratelimit(&drbd_ratelimit_state)) {
-				dev_err(DEV, "bdev->backing_bdev==NULL\n");
-				dump_stack();
-			}
-			return 0;
-		}
-		return (drbd_get_capacity(bdev->backing_bdev) & ~7ULL)
-			- MD_AL_OFFSET;
-	case DRBD_MD_INDEX_FLEX_EXT:
+	if (meta_dev_idx == DRBD_MD_INDEX_FLEX_EXT)
 		return 0;
-	}
+
+	/* Since drbd08, internal meta data is always "flexible".
+	 * position: last 4k aligned block of 4k size */
+	if (meta_dev_idx == DRBD_MD_INDEX_INTERNAL ||
+	    meta_dev_idx == DRBD_MD_INDEX_FLEX_INT)
+		return (drbd_get_capacity(bdev->backing_bdev) & ~7ULL) - 8;
+
+	/* external, some index; this is the old fixed size layout */
+	return MD_128MB_SECT * bdev->md.meta_dev_idx;
 }
 
 static inline void
···
 		if (mdev->state.disk == D_DISKLESS)
 			/* even internal references gone, safe to destroy */
 			drbd_ldev_destroy(mdev);
-		if (mdev->state.disk == D_FAILED)
+		if (mdev->state.disk == D_FAILED) {
 			/* all application IO references gone. */
-			drbd_go_diskless(mdev);
+			if (!test_and_set_bit(GO_DISKLESS, &mdev->flags))
+				drbd_queue_work(&mdev->tconn->sender_work, &mdev->go_diskless);
+		}
 		wake_up(&mdev->misc_wait);
 	}
 }

drivers/block/drbd/drbd_main.c

···
 #include <linux/reboot.h>
 #include <linux/notifier.h>
 #include <linux/kthread.h>
-
+#include <linux/workqueue.h>
 #define __KERNEL_SYSCALLS__
 #include <linux/unistd.h>
 #include <linux/vmalloc.h>
···
 	idr_for_each_entry(&minors, mdev, i) {
 		idr_remove(&minors, mdev_to_minor(mdev));
 		idr_remove(&mdev->tconn->volumes, mdev->vnr);
+		destroy_workqueue(mdev->submit.wq);
 		del_gendisk(mdev->vdisk);
 		/* synchronize_rcu(); No other threads running at this point */
 		kref_put(&mdev->kref, &drbd_minor_destroy);
···
 	kfree(tconn);
 }
 
+int init_submitter(struct drbd_conf *mdev)
+{
+	/* opencoded create_singlethread_workqueue(),
+	 * to be able to say "drbd%d", ..., minor */
+	mdev->submit.wq = alloc_workqueue("drbd%u_submit",
+			WQ_UNBOUND | WQ_MEM_RECLAIM, 1, mdev->minor);
+	if (!mdev->submit.wq)
+		return -ENOMEM;
+
+	INIT_WORK(&mdev->submit.worker, do_submit);
+	spin_lock_init(&mdev->submit.lock);
+	INIT_LIST_HEAD(&mdev->submit.writes);
+	return 0;
+}
+
 enum drbd_ret_code conn_new_minor(struct drbd_tconn *tconn, unsigned int minor, int vnr)
 {
 	struct drbd_conf *mdev;
···
 		goto out_idr_remove_minor;
 	}
 
+	if (init_submitter(mdev)) {
+		err = ERR_NOMEM;
+		drbd_msg_put_info("unable to create submit workqueue");
+		goto out_idr_remove_vol;
+	}
+
 	add_disk(disk);
 	kref_init(&mdev->kref); /* one ref for both idrs and the the add_disk */
 
···
 
 	return NO_ERROR;
 
+out_idr_remove_vol:
+	idr_remove(&tconn->volumes, vnr_got);
 out_idr_remove_minor:
 	idr_remove(&minors, minor_got);
 	synchronize_rcu();
···
 	blkdev_put(ldev->backing_bdev, FMODE_READ | FMODE_WRITE | FMODE_EXCL);
 	blkdev_put(ldev->md_bdev, FMODE_READ | FMODE_WRITE | FMODE_EXCL);
 
+	kfree(ldev->disk_conf);
 	kfree(ldev);
 }
 
···
 	rcu_read_unlock();
 }
 
+/* aligned 4kByte */
 struct meta_data_on_disk {
-	u64 la_size;           /* last agreed size. */
+	u64 la_size_sect;      /* last agreed size. */
 	u64 uuid[UI_SIZE];   /* UUIDs. */
 	u64 device_uuid;
 	u64 reserved_u64_1;
···
 	u32 magic;
 	u32 md_size_sect;
 	u32 al_offset;         /* offset to this block */
-	u32 al_nr_extents;     /* important for restoring the AL */
+	u32 al_nr_extents;     /* important for restoring the AL (userspace) */
 	      /* `-- act_log->nr_elements <-- ldev->dc.al_extents */
 	u32 bm_offset;         /* offset to the bitmap, from here */
 	u32 bm_bytes_per_bit;  /* BM_BLOCK_SIZE */
 	u32 la_peer_max_bio_size;   /* last peer max_bio_size */
-	u32 reserved_u32[3];
 
+	/* see al_tr_number_to_on_disk_sector() */
+	u32 al_stripes;
+	u32 al_stripe_size_4k;
+
+	u8 reserved_u8[4096 - (7*8 + 10*4)];
 } __packed;
 
 /**
···
 	struct meta_data_on_disk *buffer;
 	sector_t sector;
 	int i;
+
+	/* Don't accidentally change the DRBD meta data layout. */
+	BUILD_BUG_ON(UI_SIZE != 4);
+	BUILD_BUG_ON(sizeof(struct meta_data_on_disk) != 4096);
 
 	del_timer(&mdev->md_sync_timer);
 	/* timer may be rearmed by drbd_md_mark_dirty() now. */
···
 	if (!buffer)
 		goto out;
 
-	memset(buffer, 0, 512);
+	memset(buffer, 0, sizeof(*buffer));
 
-	buffer->la_size = cpu_to_be64(drbd_get_capacity(mdev->this_bdev));
+	buffer->la_size_sect = cpu_to_be64(drbd_get_capacity(mdev->this_bdev));
 	for (i = UI_CURRENT; i < UI_SIZE; i++)
 		buffer->uuid[i] = cpu_to_be64(mdev->ldev->md.uuid[i]);
 	buffer->flags = cpu_to_be32(mdev->ldev->md.flags);
···
 	buffer->bm_offset = cpu_to_be32(mdev->ldev->md.bm_offset);
 	buffer->la_peer_max_bio_size = cpu_to_be32(mdev->peer_max_bio_size);
 
-	D_ASSERT(drbd_md_ss__(mdev, mdev->ldev) == mdev->ldev->md.md_offset);
+	buffer->al_stripes = cpu_to_be32(mdev->ldev->md.al_stripes);
+	buffer->al_stripe_size_4k = cpu_to_be32(mdev->ldev->md.al_stripe_size_4k);
+
+	D_ASSERT(drbd_md_ss(mdev->ldev) == mdev->ldev->md.md_offset);
 	sector = mdev->ldev->md.md_offset;
 
 	if (drbd_md_sync_page_io(mdev, mdev->ldev, sector, WRITE)) {
···
 	put_ldev(mdev);
 }
 
+static int check_activity_log_stripe_size(struct drbd_conf *mdev,
+		struct meta_data_on_disk *on_disk,
+		struct drbd_md *in_core)
+{
+	u32 al_stripes = be32_to_cpu(on_disk->al_stripes);
+	u32 al_stripe_size_4k = be32_to_cpu(on_disk->al_stripe_size_4k);
+	u64 al_size_4k;
+
+	/* both not set: default to old fixed size activity log */
+	if (al_stripes == 0 && al_stripe_size_4k == 0) {
+		al_stripes = 1;
+		al_stripe_size_4k = MD_32kB_SECT/8;
+	}
+
+	/* some paranoia plausibility checks */
+
+	/* we need both values to be set */
+	if (al_stripes == 0 || al_stripe_size_4k == 0)
+		goto err;
+
+	al_size_4k = (u64)al_stripes * al_stripe_size_4k;
+
+	/* Upper limit of activity log area, to avoid potential overflow
+	 * problems in al_tr_number_to_on_disk_sector(). As right now, more
+	 * than 72 * 4k blocks total only increases the amount of history,
+	 * limiting this arbitrarily to 16 GB is not a real limitation ;-)  */
+	if (al_size_4k > (16 * 1024 * 1024/4))
+		goto err;
+
+	/* Lower limit: we need at least 8 transaction slots (32kB)
+	 * to not break existing setups */
+	if (al_size_4k < MD_32kB_SECT/8)
+		goto err;
+
+	in_core->al_stripe_size_4k = al_stripe_size_4k;
+	in_core->al_stripes = al_stripes;
+	in_core->al_size_4k = al_size_4k;
+
+	return 0;
+err:
+	dev_err(DEV, "invalid activity log striping: al_stripes=%u, al_stripe_size_4k=%u\n",
+			al_stripes, al_stripe_size_4k);
+	return -EINVAL;
+}
+
+static int check_offsets_and_sizes(struct drbd_conf *mdev, struct drbd_backing_dev *bdev)
+{
+	sector_t capacity = drbd_get_capacity(bdev->md_bdev);
+	struct drbd_md *in_core = &bdev->md;
+	s32 on_disk_al_sect;
+	s32 on_disk_bm_sect;
+
+	/* The on-disk size of the activity log, calculated from offsets, and
+	 * the size of the activity log calculated from the stripe settings,
+	 * should match.
+	 * Though we could relax this a bit: it is ok, if the striped activity log
+	 * fits in the available on-disk activity log size.
+	 * Right now, that would break how resize is implemented.
+	 * TODO: make drbd_determine_dev_size() (and the drbdmeta tool) aware
+	 * of possible unused padding space in the on disk layout. */
+	if (in_core->al_offset < 0) {
+		if (in_core->bm_offset > in_core->al_offset)
+			goto err;
+		on_disk_al_sect = -in_core->al_offset;
+		on_disk_bm_sect = in_core->al_offset - in_core->bm_offset;
+	} else {
+		if (in_core->al_offset != MD_4kB_SECT)
+			goto err;
+		if (in_core->bm_offset < in_core->al_offset + in_core->al_size_4k * MD_4kB_SECT)
+			goto err;
+
+		on_disk_al_sect = in_core->bm_offset - MD_4kB_SECT;
+		on_disk_bm_sect = in_core->md_size_sect - in_core->bm_offset;
+	}
+
+	/* old fixed size meta data is exactly that: fixed. */
+	if (in_core->meta_dev_idx >= 0) {
+		if (in_core->md_size_sect != MD_128MB_SECT
+		||  in_core->al_offset != MD_4kB_SECT
+		||  in_core->bm_offset != MD_4kB_SECT + MD_32kB_SECT
+		||  in_core->al_stripes != 1
+		||  in_core->al_stripe_size_4k != MD_32kB_SECT/8)
+			goto err;
+	}
+
+	if (capacity < in_core->md_size_sect)
+		goto err;
+	if (capacity - in_core->md_size_sect < drbd_md_first_sector(bdev))
+		goto err;
+
+	/* should be aligned, and at least 32k */
+	if ((on_disk_al_sect & 7) || (on_disk_al_sect < MD_32kB_SECT))
+		goto err;
+
+	/* should fit (for now: exactly) into the available on-disk space;
+	 * overflow prevention is in check_activity_log_stripe_size() above. */
+	if (on_disk_al_sect != in_core->al_size_4k * MD_4kB_SECT)
+		goto err;
+
+	/* again, should be aligned */
+	if (in_core->bm_offset & 7)
+		goto err;
+
+	/* FIXME check for device grow with flex external meta data? */
+
+	/* can the available bitmap space cover the last agreed device size? */
+	if (on_disk_bm_sect < (in_core->la_size_sect+7)/MD_4kB_SECT/8/512)
+		goto err;
+
+	return 0;
+
+err:
+	dev_err(DEV, "meta data offsets don't make sense: idx=%d "
+			"al_s=%u, al_sz4k=%u, al_offset=%d, bm_offset=%d, "
+			"md_size_sect=%u, la_size=%llu, md_capacity=%llu\n",
+			in_core->meta_dev_idx,
+			in_core->al_stripes, in_core->al_stripe_size_4k,
+			in_core->al_offset, in_core->bm_offset, in_core->md_size_sect,
+			(unsigned long long)in_core->la_size_sect,
+			(unsigned long long)capacity);
+
+	return -EINVAL;
+}
+
+
 /**
  * drbd_md_read() - Reads in the meta data super block
  * @mdev:	DRBD device.
  * @bdev:	Device from which the meta data should be read in.
  *
- * Return 0 (NO_ERROR) on success, and an enum drbd_ret_code in case
+ * Return NO_ERROR on success, and an enum drbd_ret_code in case
  * something goes wrong.
+ *
+ * Called exactly once during drbd_adm_attach(), while still being D_DISKLESS,
+ * even before @bdev is assigned to @mdev->ldev.
  */
 int drbd_md_read(struct drbd_conf *mdev, struct drbd_backing_dev *bdev)
 {
···
 	u32 magic, flags;
 	int i, rv = NO_ERROR;
 
-	if (!get_ldev_if_state(mdev, D_ATTACHING))
-		return ERR_IO_MD_DISK;
+	if (mdev->state.disk != D_DISKLESS)
+		return ERR_DISK_CONFIGURED;
 
 	buffer = drbd_md_get_buffer(mdev);
 	if (!buffer)
-		goto out;
+		return ERR_NOMEM;
+
+	/* First, figure out where our meta data superblock is located,
+	 * and read it. */
+	bdev->md.meta_dev_idx = bdev->disk_conf->meta_dev_idx;
+	bdev->md.md_offset = drbd_md_ss(bdev);
 
 	if (drbd_md_sync_page_io(mdev, bdev, bdev->md.md_offset, READ)) {
 		/* NOTE: can't do normal error processing here as this is
···
 		rv = ERR_MD_UNCLEAN;
 		goto err;
 	}
+
+	rv = ERR_MD_INVALID;
 	if (magic != DRBD_MD_MAGIC_08) {
 		if (magic == DRBD_MD_MAGIC_07)
 			dev_err(DEV, "Found old (0.7) meta data magic. Did you \"drbdadm create-md\"?\n");
 		else
 			dev_err(DEV, "Meta data magic not found. Did you \"drbdadm create-md\"?\n");
-		rv = ERR_MD_INVALID;
-		goto err;
-	}
-	if (be32_to_cpu(buffer->al_offset) != bdev->md.al_offset) {
-		dev_err(DEV, "unexpected al_offset: %d (expected %d)\n",
-		    be32_to_cpu(buffer->al_offset), bdev->md.al_offset);
-		rv = ERR_MD_INVALID;
-		goto err;
-	}
-	if (be32_to_cpu(buffer->bm_offset) != bdev->md.bm_offset) {
-		dev_err(DEV, "unexpected bm_offset: %d (expected %d)\n",
-		    be32_to_cpu(buffer->bm_offset), bdev->md.bm_offset);
-		rv = ERR_MD_INVALID;
-		goto err;
-	}
-	if (be32_to_cpu(buffer->md_size_sect) != bdev->md.md_size_sect) {
-		dev_err(DEV, "unexpected md_size: %u (expected %u)\n",
-		    be32_to_cpu(buffer->md_size_sect), bdev->md.md_size_sect);
-		rv = ERR_MD_INVALID;
 		goto err;
 	}
 
 	if (be32_to_cpu(buffer->bm_bytes_per_bit) != BM_BLOCK_SIZE) {
 		dev_err(DEV, "unexpected bm_bytes_per_bit: %u (expected %u)\n",
 		    be32_to_cpu(buffer->bm_bytes_per_bit), BM_BLOCK_SIZE);
-		rv = ERR_MD_INVALID;
 		goto err;
 	}
 
-	bdev->md.la_size_sect = be64_to_cpu(buffer->la_size);
+
+	/* convert to in_core endian */
+	bdev->md.la_size_sect = be64_to_cpu(buffer->la_size_sect);
 	for (i = UI_CURRENT; i < UI_SIZE; i++)
 		bdev->md.uuid[i] = be64_to_cpu(buffer->uuid[i]);
 	bdev->md.flags = be32_to_cpu(buffer->flags);
 	bdev->md.device_uuid = be64_to_cpu(buffer->device_uuid);
+
+	bdev->md.md_size_sect = be32_to_cpu(buffer->md_size_sect);
+	bdev->md.al_offset = be32_to_cpu(buffer->al_offset);
+	bdev->md.bm_offset = be32_to_cpu(buffer->bm_offset);
+
+	if (check_activity_log_stripe_size(mdev, buffer, &bdev->md))
+		goto err;
+	if (check_offsets_and_sizes(mdev, bdev))
+		goto err;
+
+	if (be32_to_cpu(buffer->bm_offset) != bdev->md.bm_offset) {
+		dev_err(DEV, "unexpected bm_offset: %d (expected %d)\n",
+		    be32_to_cpu(buffer->bm_offset), bdev->md.bm_offset);
+		goto err;
+	}
+	if (be32_to_cpu(buffer->md_size_sect) != bdev->md.md_size_sect) {
+		dev_err(DEV, "unexpected md_size: %u (expected %u)\n",
+		    be32_to_cpu(buffer->md_size_sect), bdev->md.md_size_sect);
+		goto err;
+	}
+
+	rv = NO_ERROR;
 
 	spin_lock_irq(&mdev->tconn->req_lock);
 	if (mdev->state.conn < C_CONNECTED) {
···
 
  err:
 	drbd_md_put_buffer(mdev);
- out:
-	put_ldev(mdev);
 
 	return rv;
 }
···
 	 * end up here after a failed attach, before ldev was even assigned.
 	 */
 	if (mdev->bitmap && mdev->ldev) {
+		/* An interrupted resync or similar is allowed to recounts bits
+		 * while we detach.
+		 * Any modifications would not be expected anymore, though.
+		 */
 		if (drbd_bitmap_io_from_worker(mdev, drbd_bm_write,
-					"detach", BM_LOCKED_MASK)) {
+					"detach", BM_LOCKED_TEST_ALLOWED)) {
 			if (test_bit(WAS_READ_ERROR, &mdev->flags)) {
 				drbd_md_set_flag(mdev, MDF_FULL_SYNC);
 				drbd_md_sync(mdev);
···
 
 	drbd_force_state(mdev, NS(disk, D_DISKLESS));
 	return 0;
-}
-
-void drbd_go_diskless(struct drbd_conf *mdev)
-{
-	D_ASSERT(mdev->state.disk == D_FAILED);
-	if (!test_and_set_bit(GO_DISKLESS, &mdev->flags))
-		drbd_queue_work(&mdev->tconn->sender_work, &mdev->go_diskless);
 }
 
 /**

drivers/block/drbd/drbd_nl.c

···
 	return 0;
 }
 
-/* initializes the md.*_offset members, so we are able to find
- * the on disk meta data */
+/* Initializes the md.*_offset members, so we are able to find
+ * the on disk meta data.
+ *
+ * We currently have two possible layouts:
+ * external:
+ *   |----------- md_size_sect ------------------|
+ *   [ 4k superblock ][ activity log ][  Bitmap  ]
+ *   | al_offset == 8 |
+ *   | bm_offset = al_offset + X      |
+ *  ==> bitmap sectors = md_size_sect - bm_offset
+ *
+ * internal:
+ *            |----------- md_size_sect ------------------|
+ * [data.....][  Bitmap  ][ activity log ][ 4k superblock ]
+ *                        | al_offset < 0 |
+ *            | bm_offset = al_offset - Y |
+ *  ==> bitmap sectors = Y = al_offset - bm_offset
+ *
+ *  Activity log size used to be fixed 32kB,
+ *  but is about to become configurable.
+ */
 static void drbd_md_set_sector_offsets(struct drbd_conf *mdev,
 				       struct drbd_backing_dev *bdev)
 {
 	sector_t md_size_sect = 0;
-	int meta_dev_idx;
+	unsigned int al_size_sect = bdev->md.al_size_4k * 8;
 
-	rcu_read_lock();
-	meta_dev_idx = rcu_dereference(bdev->disk_conf)->meta_dev_idx;
+	bdev->md.md_offset = drbd_md_ss(bdev);
 
-	switch (meta_dev_idx) {
+	switch (bdev->md.meta_dev_idx) {
 	default:
 		/* v07 style fixed size indexed meta data */
-		bdev->md.md_size_sect = MD_RESERVED_SECT;
-		bdev->md.md_offset = drbd_md_ss__(mdev, bdev);
-		bdev->md.al_offset = MD_AL_OFFSET;
-		bdev->md.bm_offset = MD_BM_OFFSET;
+		bdev->md.md_size_sect = MD_128MB_SECT;
+		bdev->md.al_offset = MD_4kB_SECT;
+		bdev->md.bm_offset = MD_4kB_SECT + al_size_sect;
 		break;
 	case DRBD_MD_INDEX_FLEX_EXT:
 		/* just occupy the full device; unit: sectors */
 		bdev->md.md_size_sect = drbd_get_capacity(bdev->md_bdev);
-		bdev->md.md_offset = 0;
-		bdev->md.al_offset = MD_AL_OFFSET;
-		bdev->md.bm_offset = MD_BM_OFFSET;
+		bdev->md.al_offset = MD_4kB_SECT;
+		bdev->md.bm_offset = MD_4kB_SECT + al_size_sect;
 		break;
 	case DRBD_MD_INDEX_INTERNAL:
 	case DRBD_MD_INDEX_FLEX_INT:
-		bdev->md.md_offset = drbd_md_ss__(mdev, bdev);
 		/* al size is still fixed */
-		bdev->md.al_offset = -MD_AL_SECTORS;
+		bdev->md.al_offset = -al_size_sect;
 		/* we need (slightly less than) ~ this much bitmap sectors: */
 		md_size_sect = drbd_get_capacity(bdev->backing_bdev);
 		md_size_sect = ALIGN(md_size_sect, BM_SECT_PER_EXT);
···
 
 		/* plus the "drbd meta data super block",
 		 * and the activity log; */
-		md_size_sect += MD_BM_OFFSET;
+		md_size_sect += MD_4kB_SECT + al_size_sect;
 
 		bdev->md.md_size_sect = md_size_sect;
 		/* bitmap offset is adjusted by 'super' block size */
-		bdev->md.bm_offset   = -md_size_sect + MD_AL_OFFSET;
+		bdev->md.bm_offset   = -md_size_sect + MD_4kB_SECT;
 		break;
 	}
-	rcu_read_unlock();
 }
 
 /* input size is expected to be in KB */
···
 enum determine_dev_size drbd_determine_dev_size(struct drbd_conf *mdev, enum dds_flags flags) __must_hold(local)
 {
 	sector_t prev_first_sect, prev_size; /* previous meta location */
-	sector_t la_size, u_size;
+	sector_t la_size_sect, u_size;
 	sector_t size;
 	char ppb[10];
 
···
 
 	prev_first_sect = drbd_md_first_sector(mdev->ldev);
 	prev_size = mdev->ldev->md.md_size_sect;
-	la_size = mdev->ldev->md.la_size_sect;
+	la_size_sect = mdev->ldev->md.la_size_sect;
 
 	/* TODO: should only be some assert here, not (re)init... */
 	drbd_md_set_sector_offsets(mdev, mdev->ldev);
···
 	if (rv == dev_size_error)
 		goto out;
 
-	la_size_changed = (la_size != mdev->ldev->md.la_size_sect);
+	la_size_changed = (la_size_sect != mdev->ldev->md.la_size_sect);
 
 	md_moved = prev_first_sect != drbd_md_first_sector(mdev->ldev)
 		|| prev_size	   != mdev->ldev->md.md_size_sect;
···
 		drbd_md_mark_dirty(mdev);
 	}
 
-	if (size > la_size)
+	if (size > la_size_sect)
 		rv = grew;
-	if (size < la_size)
+	if (size < la_size_sect)
 		rv = shrunk;
 out:
 	lc_unlock(mdev->act_log);
···
 		  sector_t u_size, int assume_peer_has_space)
 {
 	sector_t p_size = mdev->p_size;   /* partner's disk size. */
-	sector_t la_size = bdev->md.la_size_sect; /* last agreed size. */
+	sector_t la_size_sect = bdev->md.la_size_sect; /* last agreed size. */
 	sector_t m_size; /* my size */
 	sector_t size = 0;
 
···
 	if (p_size && m_size) {
 		size = min_t(sector_t, p_size, m_size);
 	} else {
-		if (la_size) {
-			size = la_size;
+		if (la_size_sect) {
+			size = la_size_sect;
 			if (m_size && m_size < size)
 				size = m_size;
 			if (p_size && p_size < size)
···
 	return 0 != (flags & DRBD_GENL_F_SET_DEFAULTS);
 }
 
-static void enforce_disk_conf_limits(struct disk_conf *dc)
+static unsigned int drbd_al_extents_max(struct drbd_backing_dev *bdev)
 {
-	if (dc->al_extents < DRBD_AL_EXTENTS_MIN)
-		dc->al_extents = DRBD_AL_EXTENTS_MIN;
-	if (dc->al_extents > DRBD_AL_EXTENTS_MAX)
-		dc->al_extents = DRBD_AL_EXTENTS_MAX;
+	/* This is limited by 16 bit "slot" numbers,
+	 * and by available on-disk context storage.
+	 *
+	 * Also (u16)~0 is special (denotes a "free" extent).
+	 *
+	 * One transaction occupies one 4kB on-disk block,
+	 * we have n such blocks in the on disk ring buffer,
+	 * the "current" transaction may fail (n-1),
+	 * and there is 919 slot numbers context information per transaction.
+	 *
+	 * 72 transaction blocks amounts to more than 2**16 context slots,
+	 * so cap there first.
+	 */
+	const unsigned int max_al_nr = DRBD_AL_EXTENTS_MAX;
+	const unsigned int sufficient_on_disk =
+		(max_al_nr + AL_CONTEXT_PER_TRANSACTION -1)
+		/AL_CONTEXT_PER_TRANSACTION;
 
-	if (dc->c_plan_ahead > DRBD_C_PLAN_AHEAD_MAX)
-		dc->c_plan_ahead = DRBD_C_PLAN_AHEAD_MAX;
+	unsigned int al_size_4k = bdev->md.al_size_4k;
+
+	if (al_size_4k > sufficient_on_disk)
+		return max_al_nr;
+
+	return (al_size_4k - 1) * AL_CONTEXT_PER_TRANSACTION;
 }
 
 int drbd_adm_disk_opts(struct sk_buff *skb, struct genl_info *info)
···
 	if (!expect(new_disk_conf->resync_rate >= 1))
 		new_disk_conf->resync_rate = 1;
 
-	enforce_disk_conf_limits(new_disk_conf);
+	if (new_disk_conf->al_extents < DRBD_AL_EXTENTS_MIN)
+		new_disk_conf->al_extents = DRBD_AL_EXTENTS_MIN;
+	if (new_disk_conf->al_extents > drbd_al_extents_max(mdev->ldev))
+		new_disk_conf->al_extents = drbd_al_extents_max(mdev->ldev);
+
+	if (new_disk_conf->c_plan_ahead > DRBD_C_PLAN_AHEAD_MAX)
+		new_disk_conf->c_plan_ahead = DRBD_C_PLAN_AHEAD_MAX;
 
 	fifo_size = (new_disk_conf->c_plan_ahead * 10 * SLEEP_TIME) / HZ;
 	if (fifo_size != mdev->rs_plan_s->size) {
···
 		goto fail;
 	}
 
-	enforce_disk_conf_limits(new_disk_conf);
+	if (new_disk_conf->c_plan_ahead > DRBD_C_PLAN_AHEAD_MAX)
+		new_disk_conf->c_plan_ahead = DRBD_C_PLAN_AHEAD_MAX;
 
 	new_plan = fifo_alloc((new_disk_conf->c_plan_ahead * 10 * SLEEP_TIME) / HZ);
 	if (!new_plan) {
···
 		retcode = ERR_MD_IDX_INVALID;
 		goto fail;
 	}
+
+	write_lock_irq(&global_state_lock);
+	retcode = drbd_resync_after_valid(mdev, new_disk_conf->resync_after);
+	write_unlock_irq(&global_state_lock);
+	if (retcode != NO_ERROR)
+		goto fail;
 
 	rcu_read_lock();
 	nc = rcu_dereference(mdev->tconn->net_conf);
···
 		goto fail;
 	}
 
-	/* RT - for drbd_get_max_capacity() DRBD_MD_INDEX_FLEX_INT */
-	drbd_md_set_sector_offsets(mdev, nbc);
+	/* Read our meta data super block early.
+	 * This also sets other on-disk offsets. */
+	retcode = drbd_md_read(mdev, nbc);
+	if (retcode != NO_ERROR)
+		goto fail;
+
+	if (new_disk_conf->al_extents < DRBD_AL_EXTENTS_MIN)
+		new_disk_conf->al_extents = DRBD_AL_EXTENTS_MIN;
+	if (new_disk_conf->al_extents > drbd_al_extents_max(nbc))
+		new_disk_conf->al_extents = drbd_al_extents_max(nbc);
 
 	if (drbd_get_max_capacity(nbc) < new_disk_conf->disk_size) {
 		dev_err(DEV, "max capacity %llu smaller than disk size %llu\n",
···
 		min_md_device_sectors = (2<<10);
 	} else {
 		max_possible_sectors = DRBD_MAX_SECTORS;
-		min_md_device_sectors = MD_RESERVED_SECT * (new_disk_conf->meta_dev_idx + 1);
+		min_md_device_sectors = MD_128MB_SECT * (new_disk_conf->meta_dev_idx + 1);
 	}
 
 	if (drbd_get_capacity(nbc->md_bdev) < min_md_device_sectors) {
···
 	if (!get_ldev_if_state(mdev, D_ATTACHING))
 		goto force_diskless;
 
-	drbd_md_set_sector_offsets(mdev, nbc);
-
 	if (!mdev->bitmap) {
 		if (drbd_bm_init(mdev)) {
 			retcode = ERR_NOMEM;
 			goto force_diskless_dec;
 		}
 	}
-
-	retcode = drbd_md_read(mdev, nbc);
-	if (retcode != NO_ERROR)
-		goto force_diskless_dec;
 
 	if (mdev->state.conn < C_CONNECTED &&
 	    mdev->state.role == R_PRIMARY &&
···
 		return SS_SUCCESS;
 	case SS_PRIMARY_NOP:
 		/* Our state checking code wants to see the peer outdated. */
-		rv = conn_request_state(tconn, NS2(conn, C_DISCONNECTING,
-						pdsk, D_OUTDATED), CS_VERBOSE);
+		rv = conn_request_state(tconn, NS2(conn, C_DISCONNECTING, pdsk, D_OUTDATED), 0);
+
+		if (rv == SS_OUTDATE_WO_CONN) /* lost connection before graceful disconnect succeeded */
+			rv = conn_request_state(tconn, NS(conn, C_DISCONNECTING), CS_VERBOSE);
+
 		break;
 	case SS_CW_FAILED_BY_PEER:
 		/* The peer probably wants to see us outdated. */
···
 	wait_event(mdev->misc_wait, !test_bit(BITMAP_IO, &mdev->flags));
 	drbd_flush_workqueue(mdev);
 
-	retcode = _drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_T), CS_ORDERED);
-
-	if (retcode < SS_SUCCESS && retcode != SS_NEED_CONNECTION)
+	/* If we happen to be C_STANDALONE R_SECONDARY, just change to
+	 * D_INCONSISTENT, and set all bits in the bitmap.  Otherwise,
+	 * try to start a resync handshake as sync target for full sync.
+	 */
+	if (mdev->state.conn == C_STANDALONE && mdev->state.role == R_SECONDARY) {
+		retcode = drbd_request_state(mdev, NS(disk, D_INCONSISTENT));
+		if (retcode >= SS_SUCCESS) {
+			if (drbd_bitmap_io(mdev, &drbd_bmio_set_n_write,
+				"set_n_write from invalidate", BM_LOCKED_MASK))
+				retcode = ERR_IO_MD_DISK;
+		}
+	} else
 		retcode = drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_T));
-
-	while (retcode == SS_NEED_CONNECTION) {
-		spin_lock_irq(&mdev->tconn->req_lock);
-		if (mdev->state.conn < C_CONNECTED)
-			retcode = _drbd_set_state(_NS(mdev, disk, D_INCONSISTENT), CS_VERBOSE, NULL);
-		spin_unlock_irq(&mdev->tconn->req_lock);
-
-		if (retcode != SS_NEED_CONNECTION)
-			break;
-
-		retcode = drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_T));
-	}
 	drbd_resume_io(mdev);
 
 out:
···
 	wait_event(mdev->misc_wait, !test_bit(BITMAP_IO, &mdev->flags));
 	drbd_flush_workqueue(mdev);
 
-	retcode = _drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_S), CS_ORDERED);
-	if (retcode < SS_SUCCESS) {
-		if (retcode == SS_NEED_CONNECTION && mdev->state.role == R_PRIMARY) {
-			/* The peer will get a resync upon connect anyways.
-			 * Just make that into a full resync. */
-			retcode = drbd_request_state(mdev, NS(pdsk, D_INCONSISTENT));
-			if (retcode >= SS_SUCCESS) {
-				if (drbd_bitmap_io(mdev, &drbd_bmio_set_susp_al,
-						   "set_n_write from invalidate_peer",
-						   BM_LOCKED_SET_ALLOWED))
-					retcode = ERR_IO_MD_DISK;
-			}
-		} else
-			retcode = drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_S));
-	}
+	/* If we happen to be C_STANDALONE R_PRIMARY, just set all bits
+	 * in the bitmap.  Otherwise, try to start a resync handshake
+	 * as sync source for full sync.
+	 */
+	if (mdev->state.conn == C_STANDALONE && mdev->state.role == R_PRIMARY) {
+		/* The peer will get a resync upon connect anyways. Just make that
+		   into a full resync. */
+		retcode = drbd_request_state(mdev, NS(pdsk, D_INCONSISTENT));
+		if (retcode >= SS_SUCCESS) {
+			if (drbd_bitmap_io(mdev, &drbd_bmio_set_susp_al,
+				"set_n_write from invalidate_peer",
+				BM_LOCKED_SET_ALLOWED))
+				retcode = ERR_IO_MD_DISK;
+		}
+	} else
+		retcode = drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_S));
 	drbd_resume_io(mdev);
 
 out:
···
 				    CS_VERBOSE + CS_WAIT_COMPLETE);
 		idr_remove(&mdev->tconn->volumes, mdev->vnr);
 		idr_remove(&minors, mdev_to_minor(mdev));
+		destroy_workqueue(mdev->submit.wq);
 		del_gendisk(mdev->vdisk);
 		synchronize_rcu();
 		kref_put(&mdev->kref, &drbd_minor_destroy);

drivers/block/drbd/drbd_proc.c

···
 
 static int drbd_proc_open(struct inode *inode, struct file *file)
 {
-	if (try_module_get(THIS_MODULE))
-		return single_open(file, drbd_seq_show, PDE_DATA(inode));
+	int err;
+
+	if (try_module_get(THIS_MODULE)) {
+		err = single_open(file, drbd_seq_show, PDE_DATA(inode));
+		if (err)
+			module_put(THIS_MODULE);
+		return err;
+	}
 	return -ENODEV;
 }
 

drivers/block/drbd/drbd_receiver.c

···
 		err = drbd_send_current_state(mdev);
 	clear_bit(USE_DEGR_WFC_T, &mdev->flags);
 	clear_bit(RESIZE_PENDING, &mdev->flags);
+	atomic_set(&mdev->ap_in_flight, 0);
 	mod_timer(&mdev->request_timer, jiffies + HZ); /* just start it here. */
 	return err;
 }
···
 		drbd_set_out_of_sync(mdev, peer_req->i.sector, peer_req->i.size);
 		peer_req->flags |= EE_CALL_AL_COMPLETE_IO;
 		peer_req->flags &= ~EE_MAY_SET_IN_SYNC;
-		drbd_al_begin_io(mdev, &peer_req->i);
+		drbd_al_begin_io(mdev, &peer_req->i, true);
 	}
 
 	err = drbd_submit_peer_request(mdev, peer_req, rw, DRBD_FAULT_DT_WR);
···
 		if (hg == -1 && mdev->state.role == R_PRIMARY) {
 			enum drbd_state_rv rv2;
 
-			drbd_set_role(mdev, R_SECONDARY, 0);
 			 /* drbd_change_state() does not sleep while in SS_IN_TRANSIENT_STATE,
 			  * we might be here in C_WF_REPORT_PARAMS which is transient.
 			  * we do not need to wait for the after state change work either. */
···
 
 	clear_bit(DISCARD_MY_DATA, &mdev->flags);
 
-	drbd_md_sync(mdev); /* update connected indicator, la_size, ... */
+	drbd_md_sync(mdev); /* update connected indicator, la_size_sect, ... */
 
 	return 0;
 }
···
 #if !defined(CONFIG_CRYPTO_HMAC) && !defined(CONFIG_CRYPTO_HMAC_MODULE)
 static int drbd_do_auth(struct drbd_tconn *tconn)
 {
-	dev_err(DEV, "This kernel was build without CONFIG_CRYPTO_HMAC.\n");
-	dev_err(DEV, "You need to disable 'cram-hmac-alg' in drbd.conf.\n");
+	conn_err(tconn, "This kernel was build without CONFIG_CRYPTO_HMAC.\n");
+	conn_err(tconn, "You need to disable 'cram-hmac-alg' in drbd.conf.\n");
 	return -1;
 }
 #else
···
 	bool ping_timeout_active = false;
 	struct net_conf *nc;
 	int ping_timeo, tcp_cork, ping_int;
+	struct sched_param param = { .sched_priority = 2 };
 
-	current->policy = SCHED_RR;  /* Make this a realtime task! */
-	current->rt_priority = 2;    /* more important than all other tasks */
+	rv = sched_setscheduler(current, SCHED_RR, &param);
+	if (rv < 0)
+		conn_err(tconn, "drbd_asender: ERROR set priority, ret=%d\n", rv);
 
 	while (get_t_state(thi) == RUNNING) {
 		drbd_thread_current_set_cpu(thi);

drivers/block/drbd/drbd_req.c

···
 static bool drbd_may_do_local_read(struct drbd_conf *mdev, sector_t sector, int size);
 
 /* Update disk stats at start of I/O request */
-static void _drbd_start_io_acct(struct drbd_conf *mdev, struct drbd_request *req, struct bio *bio)
+static void _drbd_start_io_acct(struct drbd_conf *mdev, struct drbd_request *req)
 {
-	const int rw = bio_data_dir(bio);
+	const int rw = bio_data_dir(req->master_bio);
 	int cpu;
 	cpu = part_stat_lock();
 	part_round_stats(cpu, &mdev->vdisk->part0);
 	part_stat_inc(cpu, &mdev->vdisk->part0, ios[rw]);
-	part_stat_add(cpu, &mdev->vdisk->part0, sectors[rw], bio_sectors(bio));
+	part_stat_add(cpu, &mdev->vdisk->part0, sectors[rw], req->i.size >> 9);
 	(void) cpu; /* The macro invocations above want the cpu argument, I do not like
 		       the compiler warning about cpu only assigned but never used... */
 	part_inc_in_flight(&mdev->vdisk->part0, rw);
···
 		else
 			root = &mdev->read_requests;
 		drbd_remove_request_interval(root, req);
-	} else if (!(s & RQ_POSTPONED))
-		D_ASSERT((s & (RQ_NET_MASK & ~RQ_NET_DONE)) == 0);
+	}
 
 	/* Before we can signal completion to the upper layers,
 	 * we may need to close the current transfer log epoch.
···
 		D_ASSERT(req->rq_state & RQ_NET_PENDING);
 		mod_rq_state(req, m, RQ_NET_PENDING, RQ_NET_OK|RQ_NET_DONE);
 		break;
+
+	case QUEUE_AS_DRBD_BARRIER:
+		start_new_tl_epoch(mdev->tconn);
+		mod_rq_state(req, m, 0, RQ_NET_OK|RQ_NET_DONE);
+		break;
 	};
 
 	return rv;
···
 	bool congested = false;
 	enum drbd_on_congestion on_congestion;
 
+	rcu_read_lock();
 	nc = rcu_dereference(tconn->net_conf);
 	on_congestion = nc ? nc->on_congestion : OC_BLOCK;
+	rcu_read_unlock();
 	if (on_congestion == OC_BLOCK ||
 	    tconn->agreed_pro_version < 96)
 		return;
···
 	struct drbd_conf *mdev = req->w.mdev;
 	int remote, send_oos;
 
-	rcu_read_lock();
 	remote = drbd_should_do_remote(mdev->state);
-	if (remote) {
-		maybe_pull_ahead(mdev);
-		remote = drbd_should_do_remote(mdev->state);
-	}
 	send_oos = drbd_should_send_out_of_sync(mdev->state);
-	rcu_read_unlock();
 
 	/* Need to replicate writes.  Unless it is an empty flush,
 	 * which is better mapped to a DRBD P_BARRIER packet,
···
 		/* The only size==0 bios we expect are empty flushes. */
 		D_ASSERT(req->master_bio->bi_rw & REQ_FLUSH);
 		if (remote)
-			start_new_tl_epoch(mdev->tconn);
-		return 0;
+			_req_mod(req, QUEUE_AS_DRBD_BARRIER);
+		return remote;
 	}
 
 	if (!remote && !send_oos)
···
 		bio_endio(bio, -EIO);
 }
 
-void __drbd_make_request(struct drbd_conf *mdev, struct bio *bio, unsigned long start_time)
+static void drbd_queue_write(struct drbd_conf *mdev, struct drbd_request *req)
 {
-	const int rw = bio_rw(bio);
-	struct bio_and_error m = { NULL, };
+	spin_lock(&mdev->submit.lock);
+	list_add_tail(&req->tl_requests, &mdev->submit.writes);
+	spin_unlock(&mdev->submit.lock);
+	queue_work(mdev->submit.wq, &mdev->submit.worker);
+}
+
+/* returns the new drbd_request pointer, if the caller is expected to
+ * drbd_send_and_submit() it (to save latency), or NULL if we queued the
+ * request on the submitter thread.
+ * Returns ERR_PTR(-ENOMEM) if we cannot allocate a drbd_request.
+ */
+struct drbd_request *
+drbd_request_prepare(struct drbd_conf *mdev, struct bio *bio, unsigned long start_time)
+{
+	const int rw = bio_data_dir(bio);
 	struct drbd_request *req;
-	bool no_remote = false;
 
 	/* allocate outside of all locks; */
 	req = drbd_req_new(mdev, bio);
···
 		 * if user cannot handle io errors, that's not our business. */
 		dev_err(DEV, "could not kmalloc() req\n");
 		bio_endio(bio, -ENOMEM);
-		return;
+		return ERR_PTR(-ENOMEM);
 	}
 	req->start_time = start_time;
 
···
 		req->private_bio = NULL;
 	}
 
-	/* For WRITES going to the local disk, grab a reference on the target
-	 * extent.  This waits for any resync activity in the corresponding
-	 * resync extent to finish, and, if necessary, pulls in the target
-	 * extent into the activity log, which involves further disk io because
-	 * of transactional on-disk meta data updates.
-	 * Empty flushes don't need to go into the activity log, they can only
-	 * flush data for pending writes which are already in there. */
+	/* Update disk stats */
+	_drbd_start_io_acct(mdev, req);
+
 	if (rw == WRITE && req->private_bio && req->i.size
 	&& !test_bit(AL_SUSPENDED, &mdev->flags)) {
+		if (!drbd_al_begin_io_fastpath(mdev, &req->i)) {
+			drbd_queue_write(mdev, req);
+			return NULL;
+		}
 		req->rq_state |= RQ_IN_ACT_LOG;
-		drbd_al_begin_io(mdev, &req->i);
 	}
+
+	return req;
+}
+
+static void drbd_send_and_submit(struct drbd_conf *mdev, struct drbd_request *req)
+{
+	const int rw = bio_rw(req->master_bio);
+	struct bio_and_error m = { NULL, };
+	bool no_remote = false;
 
 	spin_lock_irq(&mdev->tconn->req_lock);
 	if (rw == WRITE) {
···
 		 * but will re-aquire it before it returns here.
 		 * Needs to be before the check on drbd_suspended() */
 		complete_conflicting_writes(req);
+		/* no more giving up req_lock from now on! */
+
+		/* check for congestion, and potentially stop sending
+		 * full data updates, but start sending "dirty bits" only. */
+		maybe_pull_ahead(mdev);
 	}
 
-	/* no more giving up req_lock from now on! */
 
 	if (drbd_suspended(mdev)) {
 		/* push back and retry: */
···
 		}
 		goto out;
 	}
-
-	/* Update disk stats */
-	_drbd_start_io_acct(mdev, req, bio);
 
 	/* We fail READ/READA early, if we can not serve it.
 	 * We must do this before req is registered on any lists.
···
 
 	if (m.bio)
 		complete_master_bio(mdev, &m);
-	return;
+}
+
+void __drbd_make_request(struct drbd_conf *mdev, struct bio *bio, unsigned long start_time)
+{
+	struct drbd_request *req = drbd_request_prepare(mdev, bio, start_time);
+	if (IS_ERR_OR_NULL(req))
+		return;
+	drbd_send_and_submit(mdev, req);
+}
+
+static void submit_fast_path(struct drbd_conf *mdev, struct list_head *incoming)
+{
+	struct drbd_request *req, *tmp;
+	list_for_each_entry_safe(req, tmp, incoming, tl_requests) {
+		const int rw = bio_data_dir(req->master_bio);
+
+		if (rw == WRITE /* rw != WRITE should not even end up here! */
+		&& req->private_bio && req->i.size
+		&& !test_bit(AL_SUSPENDED, &mdev->flags)) {
+			if (!drbd_al_begin_io_fastpath(mdev, &req->i))
+				continue;
+
+			req->rq_state |= RQ_IN_ACT_LOG;
+		}
+
+		list_del_init(&req->tl_requests);
+		drbd_send_and_submit(mdev, req);
+	}
+}
+
+static bool prepare_al_transaction_nonblock(struct drbd_conf *mdev,
+					    struct list_head *incoming,
+					    struct list_head *pending)
+{
+	struct drbd_request *req, *tmp;
+	int wake = 0;
+	int err;
+
+	spin_lock_irq(&mdev->al_lock);
+	list_for_each_entry_safe(req, tmp, incoming, tl_requests) {
+		err = drbd_al_begin_io_nonblock(mdev, &req->i);
+		if (err == -EBUSY)
+			wake = 1;
+		if (err)
+			continue;
+		req->rq_state |= RQ_IN_ACT_LOG;
+		list_move_tail(&req->tl_requests, pending);
+	}
+	spin_unlock_irq(&mdev->al_lock);
+	if (wake)
+		wake_up(&mdev->al_wait);
+
+	return !list_empty(pending);
+}
+
+void do_submit(struct work_struct *ws)
+{
+	struct drbd_conf *mdev = container_of(ws, struct drbd_conf, submit.worker);
+	LIST_HEAD(incoming);
+	LIST_HEAD(pending);
+	struct drbd_request *req, *tmp;
+
+	for (;;) {
+		spin_lock(&mdev->submit.lock);
+		list_splice_tail_init(&mdev->submit.writes, &incoming);
+		spin_unlock(&mdev->submit.lock);
+
+		submit_fast_path(mdev, &incoming);
+		if (list_empty(&incoming))
+			break;
+
+		wait_event(mdev->al_wait, prepare_al_transaction_nonblock(mdev, &incoming, &pending));
+		/* Maybe more was queued, while we prepared the transaction?
+		 * Try to stuff them into this transaction as well.
+		 * Be strictly non-blocking here, no wait_event, we already
+		 * have something to commit.
+		 * Stop if we don't make any more progres.
+		 */
+		for (;;) {
+			LIST_HEAD(more_pending);
+			LIST_HEAD(more_incoming);
+			bool made_progress;
+
+			/* It is ok to look outside the lock,
+			 * it's only an optimization anyways */
+			if (list_empty(&mdev->submit.writes))
+				break;
+
+			spin_lock(&mdev->submit.lock);
+			list_splice_tail_init(&mdev->submit.writes, &more_incoming);
+			spin_unlock(&mdev->submit.lock);
+
+			if (list_empty(&more_incoming))
+				break;
+
+			made_progress = prepare_al_transaction_nonblock(mdev, &more_incoming, &more_pending);
+
+			list_splice_tail_init(&more_pending, &pending);
+			list_splice_tail_init(&more_incoming, &incoming);
+
+			if (!made_progress)
+				break;
+		}
+		drbd_al_begin_io_commit(mdev, false);
+
+		list_for_each_entry_safe(req, tmp, &pending, tl_requests) {
+			list_del_init(&req->tl_requests);
+			drbd_send_and_submit(mdev, req);
+		}
+	}
 }
 
 void drbd_make_request(struct request_queue *q, struct bio *bio)

drivers/block/drbd/drbd_req.h

···
 	QUEUE_FOR_NET_READ,
 	QUEUE_FOR_SEND_OOS,
 
+	/* An empty flush is queued as P_BARRIER,
+	 * which will cause it to complete "successfully",
+	 * even if the local disk flush failed.
+	 *
+	 * Just like "real" requests, empty flushes (blkdev_issue_flush()) will
+	 * only see an error if neither local nor remote data is reachable. */
+	QUEUE_AS_DRBD_BARRIER,
+
 	SEND_CANCELED,
 	SEND_FAILED,
 	HANDED_OVER_TO_NETWORK,

drivers/block/drbd/drbd_state.c

···
 		  mdev->tconn->agreed_pro_version < 88)
 		rv = SS_NOT_SUPPORTED;
 
+	else if (ns.role == R_PRIMARY && ns.disk < D_UP_TO_DATE && ns.pdsk < D_UP_TO_DATE)
+		rv = SS_NO_UP_TO_DATE_DISK;
+
+	else if ((ns.conn == C_STARTING_SYNC_S || ns.conn == C_STARTING_SYNC_T) &&
+                 ns.pdsk == D_UNKNOWN)
+		rv = SS_NEED_CONNECTION;
+
 	else if (ns.conn >= C_CONNECTED && ns.pdsk == D_UNKNOWN)
 		rv = SS_CONNECTED_OUTDATES;
 
···
 	if ((ns.conn == C_SYNC_TARGET || ns.conn == C_SYNC_SOURCE)
 	    && os.conn < C_WF_REPORT_PARAMS)
 		rv = SS_NEED_CONNECTION; /* No NetworkFailure -> SyncTarget etc... */
+
+	if (ns.conn == C_DISCONNECTING && ns.pdsk == D_OUTDATED &&
+	    os.conn < C_CONNECTED && os.pdsk > D_OUTDATED)
+		rv = SS_OUTDATE_WO_CONN;
 
 	return rv;
 }
···
 			&drbd_bmio_set_n_write, &abw_start_sync,
 			"set_n_write from StartingSync", BM_LOCKED_TEST_ALLOWED);
 
-	/* We are invalidating our self... */
-	if (os.conn < C_CONNECTED && ns.conn < C_CONNECTED &&
-	    os.disk > D_INCONSISTENT && ns.disk == D_INCONSISTENT)
-		/* other bitmap operation expected during this phase */
-		drbd_queue_bitmap_io(mdev, &drbd_bmio_set_n_write, NULL,
-			"set_n_write from invalidate", BM_LOCKED_MASK);
-
 	/* first half of local IO error, failure to attach,
 	 * or administrative detach */
 	if (os.disk != D_FAILED && ns.disk == D_FAILED) {
···
 	if (test_and_clear_bit(CONN_WD_ST_CHG_FAIL, &tconn->flags))
 		return SS_CW_FAILED_BY_PEER;
 
-	rv = tconn->cstate != C_WF_REPORT_PARAMS ? SS_CW_NO_NEED : SS_UNKNOWN_ERROR;
-
-	if (rv == SS_UNKNOWN_ERROR)
-		rv = conn_is_valid_transition(tconn, mask, val, 0);
-
-	if (rv == SS_SUCCESS)
-		rv = SS_UNKNOWN_ERROR; /* cont waiting, otherwise fail. */
+	rv = conn_is_valid_transition(tconn, mask, val, 0);
+	if (rv == SS_SUCCESS && tconn->cstate == C_WF_REPORT_PARAMS)
+		rv = SS_UNKNOWN_ERROR; /* continue waiting */
 
 	return rv;
 }

drivers/block/drbd/drbd_strings.c

···
 	[-SS_LOWER_THAN_OUTDATED] = "Disk state is lower than outdated",
 	[-SS_IN_TRANSIENT_STATE] = "In transient state, retry after next state change",
 	[-SS_CONCURRENT_ST_CHG] = "Concurrent state changes detected and aborted",
+	[-SS_OUTDATE_WO_CONN] = "Need a connection for a graceful disconnect/outdate peer",
 	[-SS_O_VOL_PEER_PRI] = "Other vol primary on peer not allowed by config",
 };
 

drivers/block/drbd/drbd_worker.c

···
 	md_io->done = 1;
 	wake_up(&mdev->misc_wait);
 	bio_put(bio);
-	put_ldev(mdev);
+	if (mdev->ldev) /* special case: drbd_md_read() during drbd_adm_attach() */
+		put_ldev(mdev);
 }
 
 /* reads on behalf of the partner,
···
 	struct drbd_conf *mdev = w->mdev;
 
 	if (bio_data_dir(req->master_bio) == WRITE && req->rq_state & RQ_IN_ACT_LOG)
-		drbd_al_begin_io(mdev, &req->i);
+		drbd_al_begin_io(mdev, &req->i, false);
 
 	drbd_req_make_private_bio(req, req->master_bio);
 	req->private_bio->bi_bdev = mdev->ldev->backing_bdev;
···
 	int resync_after;
 
 	while (1) {
-		if (!odev->ldev)
+		if (!odev->ldev || odev->state.disk == D_DISKLESS)
 			return 1;
 		rcu_read_lock();
 		resync_after = rcu_dereference(odev->ldev->disk_conf)->resync_after;
···
 		if (resync_after == -1)
 			return 1;
 		odev = minor_to_mdev(resync_after);
-		if (!expect(odev))
+		if (!odev)
 			return 1;
 		if ((odev->state.conn >= C_SYNC_SOURCE &&
 		     odev->state.conn <= C_PAUSED_SYNC_T) ||
···
 
 	if (o_minor == -1)
 		return NO_ERROR;
-	if (o_minor < -1 || minor_to_mdev(o_minor) == NULL)
+	if (o_minor < -1 || o_minor > MINORMASK)
 		return ERR_RESYNC_AFTER;
 
 	/* check for loops */
···
 	while (1) {
 		if (odev == mdev)
 			return ERR_RESYNC_AFTER_CYCLE;
+
+		/* You are free to depend on diskless, non-existing,
+		 * or not yet/no longer existing minors.
+		 * We only reject dependency loops.
+		 * We cannot follow the dependency chain beyond a detached or
+		 * missing minor.
+		 */
+		if (!odev || !odev->ldev || odev->state.disk == D_DISKLESS)
+			return NO_ERROR;
 
 		rcu_read_lock();
 		resync_after = rcu_dereference(odev->ldev->disk_conf)->resync_after;
···
 	clear_bit(B_RS_H_DONE, &mdev->flags);
 
 	write_lock_irq(&global_state_lock);
-	if (!get_ldev_if_state(mdev, D_NEGOTIATING)) {
+	/* Did some connection breakage or IO error race with us? */
+	if (mdev->state.conn < C_CONNECTED
+	|| !get_ldev_if_state(mdev, D_NEGOTIATING)) {
 		write_unlock_irq(&global_state_lock);
 		mutex_unlock(mdev->state_mutex);
 		return;

drivers/block/mg_disk.c

···
 	.getgeo = mg_getgeo
 };
 
+#ifdef CONFIG_PM_SLEEP
 static int mg_suspend(struct device *dev)
 {
 	struct mg_drv_data *prv_data = dev->platform_data;
···
 
 	return 0;
 }
+#endif
 
 static SIMPLE_DEV_PM_OPS(mg_pm, mg_suspend, mg_resume);
 

drivers/block/mtip32xx/mtip32xx.c

···
 	atomic_set(&port->commands[tag].active, 0);
 	release_slot(port, tag);
 
-	up(&port->cmd_slot);
+	if (unlikely(command->unaligned))
+		up(&port->cmd_slot_unal);
+	else
+		up(&port->cmd_slot);
 }
 
 /*
···
 	}
 #endif
 
+#ifdef MTIP_TRIM /* Disabling TRIM support temporarily */
 	/* Demux ID.DRAT & ID.RZAT to determine trim support */
 	if (port->identify[69] & (1 << 14) && port->identify[69] & (1 << 5))
 		port->dd->trim_supp = true;
 	else
+#endif
 		port->dd->trim_supp = false;
 
 	/* Set the identify buffer as valid. */
···
  */
 static void mtip_hw_submit_io(struct driver_data *dd, sector_t sector,
 			      int nsect, int nents, int tag, void *callback,
-			      void *data, int dir)
+			      void *data, int dir, int unaligned)
 {
 	struct host_to_dev_fis	*fis;
 	struct mtip_port *port = dd->port;
···
 
 	command->scatter_ents = nents;
 
+	command->unaligned = unaligned;
 	/*
 	 * The number of retries for this command before it is
 	 * reported as a failure to the upper layers.
···
 	fis->res2        = 0;
 	fis->res3        = 0;
 	fill_command_sg(dd, command, nents);
+
+	if (unaligned)
+		fis->device |= 1 << 7;
 
 	/* Populate the command header */
 	command->command_header->opts =
···
  * return value
  *      None
  */
-static void mtip_hw_release_scatterlist(struct driver_data *dd, int tag)
+static void mtip_hw_release_scatterlist(struct driver_data *dd, int tag,
+								int unaligned)
 {
+	struct semaphore *sem = unaligned ? &dd->port->cmd_slot_unal :
+							&dd->port->cmd_slot;
 	release_slot(dd->port, tag);
+	up(sem);
 }
 
 /*
···
  *	or NULL if no command slots are available.
  */
 static struct scatterlist *mtip_hw_get_scatterlist(struct driver_data *dd,
-						   int *tag)
+						   int *tag, int unaligned)
 {
+	struct semaphore *sem = unaligned ? &dd->port->cmd_slot_unal :
+							&dd->port->cmd_slot;
+
 	/*
 	 * It is possible that, even with this semaphore, a thread
 	 * may think that no command slots are available. Therefore, we
 	 * need to make an attempt to get_slot().
 	 */
-	down(&dd->port->cmd_slot);
+	down(sem);
 	*tag = get_slot(dd->port);
 
 	if (unlikely(test_bit(MTIP_DDF_REMOVE_PENDING_BIT, &dd->dd_flag))) {
-		up(&dd->port->cmd_slot);
+		up(sem);
 		return NULL;
 	}
 	if (unlikely(*tag < 0)) {
-		up(&dd->port->cmd_slot);
+		up(sem);
 		return NULL;
 	}
 
···
 		dd->mmio + HOST_HSORG);
 }
 
+static int mtip_device_unaligned_constrained(struct driver_data *dd)
+{
+	return (dd->pdev->device == P420M_DEVICE_ID ? 1 : 0);
+}
+
 /*
  * Detect the details of the product, and store anything needed
  * into the driver data structure.  This includes product type and
···
 	for (i = 0; i < MTIP_MAX_SLOT_GROUPS; i++)
 		dd->work[i].port = dd->port;
 
+	/* Enable unaligned IO constraints for some devices */
+	if (mtip_device_unaligned_constrained(dd))
+		dd->unal_qdepth = MTIP_MAX_UNALIGNED_SLOTS;
+	else
+		dd->unal_qdepth = 0;
+
 	/* Counting semaphore to track command slot usage */
-	sema_init(&dd->port->cmd_slot, num_command_slots - 1);
+	sema_init(&dd->port->cmd_slot, num_command_slots - 1 - dd->unal_qdepth);
+	sema_init(&dd->port->cmd_slot_unal, dd->unal_qdepth);
 
 	/* Spinlock to prevent concurrent issue */
 	for (i = 0; i < MTIP_MAX_SLOT_GROUPS; i++)
···
 	struct scatterlist *sg;
 	struct bio_vec *bvec;
 	int nents = 0;
-	int tag = 0;
+	int tag = 0, unaligned = 0;
 
 	if (unlikely(dd->dd_flag & MTIP_DDF_STOP_IO)) {
 		if (unlikely(test_bit(MTIP_DDF_REMOVE_PENDING_BIT,
···
 		return;
 	}
 
-	sg = mtip_hw_get_scatterlist(dd, &tag);
+	if (bio_data_dir(bio) == WRITE && bio_sectors(bio) <= 64 &&
+							dd->unal_qdepth) {
+		if (bio->bi_sector % 8 != 0) /* Unaligned on 4k boundaries */
+			unaligned = 1;
+		else if (bio_sectors(bio) % 8 != 0) /* Aligned but not 4k/8k */
+			unaligned = 1;
+	}
+
+	sg = mtip_hw_get_scatterlist(dd, &tag, unaligned);
 	if (likely(sg != NULL)) {
 		blk_queue_bounce(queue, &bio);
 
···
 			dev_warn(&dd->pdev->dev,
 				"Maximum number of SGL entries exceeded\n");
 			bio_io_error(bio);
-			mtip_hw_release_scatterlist(dd, tag);
+			mtip_hw_release_scatterlist(dd, tag, unaligned);
 			return;
 		}
 
···
 				tag,
 				bio_endio,
 				bio,
-				bio_data_dir(bio));
+				bio_data_dir(bio),
+				unaligned);
 	} else
 		bio_io_error(bio);
 }
···
  */
 static int mtip_block_shutdown(struct driver_data *dd)
 {
-	dev_info(&dd->pdev->dev,
-		"Shutting down %s ...\n", dd->disk->disk_name);
-
 	/* Delete our gendisk structure, and cleanup the blk queue. */
 	if (dd->disk) {
-		if (dd->disk->queue)
-			del_gendisk(dd->disk);
-		else
-			put_disk(dd->disk);
-	}
+		dev_info(&dd->pdev->dev,
+			"Shutting down %s ...\n", dd->disk->disk_name);
 
+		if (dd->disk->queue) {
+			del_gendisk(dd->disk);
+			blk_cleanup_queue(dd->queue);
+		} else
+			put_disk(dd->disk);
+		dd->disk  = NULL;
+		dd->queue = NULL;
+	}
 
 	spin_lock(&rssd_index_lock);
 	ida_remove(&rssd_index_ida, dd->index);
 	spin_unlock(&rssd_index_lock);
-
-	blk_cleanup_queue(dd->queue);
-	dd->disk  = NULL;
-	dd->queue = NULL;
 
 	mtip_hw_shutdown(dd);
 	return 0;

drivers/block/mtip32xx/mtip32xx.h

···
 #define MTIP_FTL_REBUILD_MAGIC		0xED51
 #define MTIP_FTL_REBUILD_TIMEOUT_MS	2400000
 
+/* unaligned IO handling */
+#define MTIP_MAX_UNALIGNED_SLOTS	8
+
 /* Macro to extract the tag bit number from a tag value. */
 #define MTIP_TAG_BIT(tag)	(tag & 0x1F)
 
···
 
 	int scatter_ents; /* Number of scatter list entries used */
 
+	int unaligned; /* command is unaligned on 4k boundary */
+
 	struct scatterlist sg[MTIP_MAX_SG]; /* Scatter list entries */
 
 	int retries; /* The number of retries left for this command. */
···
 	 * command slots available.
 	 */
 	struct semaphore cmd_slot;
+
+	/* Semaphore to control queue depth of unaligned IOs */
+	struct semaphore cmd_slot_unal;
+
 	/* Spinlock for working around command-issue bug. */
 	spinlock_t cmd_issue_lock[MTIP_MAX_SLOT_GROUPS];
 };
···
 	atomic_t irq_workers_active;
 
 	int isr_binding;
+
+	int unal_qdepth; /* qdepth of unaligned IO queue */
 
 	struct list_head online_list; /* linkage for online list */
 

drivers/md/Kconfig

···
 
 	  In unsure, say N.
 
+source "drivers/md/bcache/Kconfig"
+
 config BLK_DEV_DM
 	tristate "Device mapper support"
 	---help---

drivers/md/Makefile

···
 obj-$(CONFIG_MD_RAID456)	+= raid456.o
 obj-$(CONFIG_MD_MULTIPATH)	+= multipath.o
 obj-$(CONFIG_MD_FAULTY)		+= faulty.o
+obj-$(CONFIG_BCACHE)		+= bcache/
 obj-$(CONFIG_BLK_DEV_MD)	+= md-mod.o
 obj-$(CONFIG_BLK_DEV_DM)	+= dm-mod.o
 obj-$(CONFIG_DM_BUFIO)		+= dm-bufio.o

drivers/md/bcache/Kconfig

···
+
+config BCACHE
+	tristate "Block device as cache"
+	select CLOSURES
+	---help---
+	Allows a block device to be used as cache for other devices; uses
+	a btree for indexing and the layout is optimized for SSDs.
+
+	See Documentation/bcache.txt for details.
+
+config BCACHE_DEBUG
+	bool "Bcache debugging"
+	depends on BCACHE
+	---help---
+	Don't select this option unless you're a developer
+
+	Enables extra debugging tools (primarily a fuzz tester)
+
+config BCACHE_EDEBUG
+	bool "Extended runtime checks"
+	depends on BCACHE
+	---help---
+	Don't select this option unless you're a developer
+
+	Enables extra runtime checks which significantly affect performance
+
+config BCACHE_CLOSURES_DEBUG
+	bool "Debug closures"
+	depends on BCACHE
+	select DEBUG_FS
+	---help---
+	Keeps all active closures in a linked list and provides a debugfs
+	interface to list them, which makes it possible to see asynchronous
+	operations that get stuck.
+
+# cgroup code needs to be updated:
+#
+#config CGROUP_BCACHE
+#	bool "Cgroup controls for bcache"
+#	depends on BCACHE && BLK_CGROUP
+#	---help---
+#	TODO

drivers/md/bcache/Makefile

···
+
+obj-$(CONFIG_BCACHE)	+= bcache.o
+
+bcache-y		:= alloc.o btree.o bset.o io.o journal.o writeback.o\
+	movinggc.o request.o super.o sysfs.o debug.o util.o trace.o stats.o closure.o
+
+CFLAGS_request.o	+= -Iblock

drivers/md/bcache/alloc.c

···
+/*
+ * Primary bucket allocation code
+ *
+ * Copyright 2012 Google, Inc.
+ *
+ * Allocation in bcache is done in terms of buckets:
+ *
+ * Each bucket has associated an 8 bit gen; this gen corresponds to the gen in
+ * btree pointers - they must match for the pointer to be considered valid.
+ *
+ * Thus (assuming a bucket has no dirty data or metadata in it) we can reuse a
+ * bucket simply by incrementing its gen.
+ *
+ * The gens (along with the priorities; it's really the gens are important but
+ * the code is named as if it's the priorities) are written in an arbitrary list
+ * of buckets on disk, with a pointer to them in the journal header.
+ *
+ * When we invalidate a bucket, we have to write its new gen to disk and wait
+ * for that write to complete before we use it - otherwise after a crash we
+ * could have pointers that appeared to be good but pointed to data that had
+ * been overwritten.
+ *
+ * Since the gens and priorities are all stored contiguously on disk, we can
+ * batch this up: We fill up the free_inc list with freshly invalidated buckets,
+ * call prio_write(), and when prio_write() finishes we pull buckets off the
+ * free_inc list and optionally discard them.
+ *
+ * free_inc isn't the only freelist - if it was, we'd often to sleep while
+ * priorities and gens were being written before we could allocate. c->free is a
+ * smaller freelist, and buckets on that list are always ready to be used.
+ *
+ * If we've got discards enabled, that happens when a bucket moves from the
+ * free_inc list to the free list.
+ *
+ * There is another freelist, because sometimes we have buckets that we know
+ * have nothing pointing into them - these we can reuse without waiting for
+ * priorities to be rewritten. These come from freed btree nodes and buckets
+ * that garbage collection discovered no longer had valid keys pointing into
+ * them (because they were overwritten). That's the unused list - buckets on the
+ * unused list move to the free list, optionally being discarded in the process.
+ *
+ * It's also important to ensure that gens don't wrap around - with respect to
+ * either the oldest gen in the btree or the gen on disk. This is quite
+ * difficult to do in practice, but we explicitly guard against it anyways - if
+ * a bucket is in danger of wrapping around we simply skip invalidating it that
+ * time around, and we garbage collect or rewrite the priorities sooner than we
+ * would have otherwise.
+ *
+ * bch_bucket_alloc() allocates a single bucket from a specific cache.
+ *
+ * bch_bucket_alloc_set() allocates one or more buckets from different caches
+ * out of a cache set.
+ *
+ * free_some_buckets() drives all the processes described above. It's called
+ * from bch_bucket_alloc() and a few other places that need to make sure free
+ * buckets are ready.
+ *
+ * invalidate_buckets_(lru|fifo)() find buckets that are available to be
+ * invalidated, and then invalidate them and stick them on the free_inc list -
+ * in either lru or fifo order.
+ */
+
+#include "bcache.h"
+#include "btree.h"
+
+#include <linux/random.h>
+
+#define MAX_IN_FLIGHT_DISCARDS		8U
+
+/* Bucket heap / gen */
+
+uint8_t bch_inc_gen(struct cache *ca, struct bucket *b)
+{
+	uint8_t ret = ++b->gen;
+
+	ca->set->need_gc = max(ca->set->need_gc, bucket_gc_gen(b));
+	WARN_ON_ONCE(ca->set->need_gc > BUCKET_GC_GEN_MAX);
+
+	if (CACHE_SYNC(&ca->set->sb)) {
+		ca->need_save_prio = max(ca->need_save_prio,
+					 bucket_disk_gen(b));
+		WARN_ON_ONCE(ca->need_save_prio > BUCKET_DISK_GEN_MAX);
+	}
+
+	return ret;
+}
+
+void bch_rescale_priorities(struct cache_set *c, int sectors)
+{
+	struct cache *ca;
+	struct bucket *b;
+	unsigned next = c->nbuckets * c->sb.bucket_size / 1024;
+	unsigned i;
+	int r;
+
+	atomic_sub(sectors, &c->rescale);
+
+	do {
+		r = atomic_read(&c->rescale);
+
+		if (r >= 0)
+			return;
+	} while (atomic_cmpxchg(&c->rescale, r, r + next) != r);
+
+	mutex_lock(&c->bucket_lock);
+
+	c->min_prio = USHRT_MAX;
+
+	for_each_cache(ca, c, i)
+		for_each_bucket(b, ca)
+			if (b->prio &&
+			    b->prio != BTREE_PRIO &&
+			    !atomic_read(&b->pin)) {
+				b->prio--;
+				c->min_prio = min(c->min_prio, b->prio);
+			}
+
+	mutex_unlock(&c->bucket_lock);
+}
+
+/* Discard/TRIM */
+
+struct discard {
+	struct list_head	list;
+	struct work_struct	work;
+	struct cache		*ca;
+	long			bucket;
+
+	struct bio		bio;
+	struct bio_vec		bv;
+};
+
+static void discard_finish(struct work_struct *w)
+{
+	struct discard *d = container_of(w, struct discard, work);
+	struct cache *ca = d->ca;
+	char buf[BDEVNAME_SIZE];
+
+	if (!test_bit(BIO_UPTODATE, &d->bio.bi_flags)) {
+		pr_notice("discard error on %s, disabling",
+			 bdevname(ca->bdev, buf));
+		d->ca->discard = 0;
+	}
+
+	mutex_lock(&ca->set->bucket_lock);
+
+	fifo_push(&ca->free, d->bucket);
+	list_add(&d->list, &ca->discards);
+	atomic_dec(&ca->discards_in_flight);
+
+	mutex_unlock(&ca->set->bucket_lock);
+
+	closure_wake_up(&ca->set->bucket_wait);
+	wake_up(&ca->set->alloc_wait);
+
+	closure_put(&ca->set->cl);
+}
+
+static void discard_endio(struct bio *bio, int error)
+{
+	struct discard *d = container_of(bio, struct discard, bio);
+	schedule_work(&d->work);
+}
+
+static void do_discard(struct cache *ca, long bucket)
+{
+	struct discard *d = list_first_entry(&ca->discards,
+					     struct discard, list);
+
+	list_del(&d->list);
+	d->bucket = bucket;
+
+	atomic_inc(&ca->discards_in_flight);
+	closure_get(&ca->set->cl);
+
+	bio_init(&d->bio);
+
+	d->bio.bi_sector	= bucket_to_sector(ca->set, d->bucket);
+	d->bio.bi_bdev		= ca->bdev;
+	d->bio.bi_rw		= REQ_WRITE|REQ_DISCARD;
+	d->bio.bi_max_vecs	= 1;
+	d->bio.bi_io_vec	= d->bio.bi_inline_vecs;
+	d->bio.bi_size		= bucket_bytes(ca);
+	d->bio.bi_end_io	= discard_endio;
+	bio_set_prio(&d->bio, IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0));
+
+	submit_bio(0, &d->bio);
+}
+
+/* Allocation */
+
+static inline bool can_inc_bucket_gen(struct bucket *b)
+{
+	return bucket_gc_gen(b) < BUCKET_GC_GEN_MAX &&
+		bucket_disk_gen(b) < BUCKET_DISK_GEN_MAX;
+}
+
+bool bch_bucket_add_unused(struct cache *ca, struct bucket *b)
+{
+	BUG_ON(GC_MARK(b) || GC_SECTORS_USED(b));
+
+	if (fifo_used(&ca->free) > ca->watermark[WATERMARK_MOVINGGC] &&
+	    CACHE_REPLACEMENT(&ca->sb) == CACHE_REPLACEMENT_FIFO)
+		return false;
+
+	b->prio = 0;
+
+	if (can_inc_bucket_gen(b) &&
+	    fifo_push(&ca->unused, b - ca->buckets)) {
+		atomic_inc(&b->pin);
+		return true;
+	}
+
+	return false;
+}
+
+static bool can_invalidate_bucket(struct cache *ca, struct bucket *b)
+{
+	return GC_MARK(b) == GC_MARK_RECLAIMABLE &&
+		!atomic_read(&b->pin) &&
+		can_inc_bucket_gen(b);
+}
+
+static void invalidate_one_bucket(struct cache *ca, struct bucket *b)
+{
+	bch_inc_gen(ca, b);
+	b->prio = INITIAL_PRIO;
+	atomic_inc(&b->pin);
+	fifo_push(&ca->free_inc, b - ca->buckets);
+}
+
+#define bucket_prio(b)				\
+	(((unsigned) (b->prio - ca->set->min_prio)) * GC_SECTORS_USED(b))
+
+#define bucket_max_cmp(l, r)	(bucket_prio(l) < bucket_prio(r))
+#define bucket_min_cmp(l, r)	(bucket_prio(l) > bucket_prio(r))
+
+static void invalidate_buckets_lru(struct cache *ca)
+{
+	struct bucket *b;
+	ssize_t i;
+
+	ca->heap.used = 0;
+
+	for_each_bucket(b, ca) {
+		/*
+		 * If we fill up the unused list, if we then return before
+		 * adding anything to the free_inc list we'll skip writing
+		 * prios/gens and just go back to allocating from the unused
+		 * list:
+		 */
+		if (fifo_full(&ca->unused))
+			return;
+
+		if (!can_invalidate_bucket(ca, b))
+			continue;
+
+		if (!GC_SECTORS_USED(b) &&
+		    bch_bucket_add_unused(ca, b))
+			continue;
+
+		if (!heap_full(&ca->heap))
+			heap_add(&ca->heap, b, bucket_max_cmp);
+		else if (bucket_max_cmp(b, heap_peek(&ca->heap))) {
+			ca->heap.data[0] = b;
+			heap_sift(&ca->heap, 0, bucket_max_cmp);
+		}
+	}
+
+	for (i = ca->heap.used / 2 - 1; i >= 0; --i)
+		heap_sift(&ca->heap, i, bucket_min_cmp);
+
+	while (!fifo_full(&ca->free_inc)) {
+		if (!heap_pop(&ca->heap, b, bucket_min_cmp)) {
+			/*
+			 * We don't want to be calling invalidate_buckets()
+			 * multiple times when it can't do anything
+			 */
+			ca->invalidate_needs_gc = 1;
+			bch_queue_gc(ca->set);
+			return;
+		}
+
+		invalidate_one_bucket(ca, b);
+	}
+}
+
+static void invalidate_buckets_fifo(struct cache *ca)
+{
+	struct bucket *b;
+	size_t checked = 0;
+
+	while (!fifo_full(&ca->free_inc)) {
+		if (ca->fifo_last_bucket <  ca->sb.first_bucket ||
+		    ca->fifo_last_bucket >= ca->sb.nbuckets)
+			ca->fifo_last_bucket = ca->sb.first_bucket;
+
+		b = ca->buckets + ca->fifo_last_bucket++;
+
+		if (can_invalidate_bucket(ca, b))
+			invalidate_one_bucket(ca, b);
+
+		if (++checked >= ca->sb.nbuckets) {
+			ca->invalidate_needs_gc = 1;
+			bch_queue_gc(ca->set);
+			return;
+		}
+	}
+}
+
+static void invalidate_buckets_random(struct cache *ca)
+{
+	struct bucket *b;
+	size_t checked = 0;
+
+	while (!fifo_full(&ca->free_inc)) {
+		size_t n;
+		get_random_bytes(&n, sizeof(n));
+
+		n %= (size_t) (ca->sb.nbuckets - ca->sb.first_bucket);
+		n += ca->sb.first_bucket;
+
+		b = ca->buckets + n;
+
+		if (can_invalidate_bucket(ca, b))
+			invalidate_one_bucket(ca, b);
+
+		if (++checked >= ca->sb.nbuckets / 2) {
+			ca->invalidate_needs_gc = 1;
+			bch_queue_gc(ca->set);
+			return;
+		}
+	}
+}
+
+static void invalidate_buckets(struct cache *ca)
+{
+	if (ca->invalidate_needs_gc)
+		return;
+
+	switch (CACHE_REPLACEMENT(&ca->sb)) {
+	case CACHE_REPLACEMENT_LRU:
+		invalidate_buckets_lru(ca);
+		break;
+	case CACHE_REPLACEMENT_FIFO:
+		invalidate_buckets_fifo(ca);
+		break;
+	case CACHE_REPLACEMENT_RANDOM:
+		invalidate_buckets_random(ca);
+		break;
+	}
+
+	pr_debug("free %zu/%zu free_inc %zu/%zu unused %zu/%zu",
+		 fifo_used(&ca->free), ca->free.size,
+		 fifo_used(&ca->free_inc), ca->free_inc.size,
+		 fifo_used(&ca->unused), ca->unused.size);
+}
+
+#define allocator_wait(ca, cond)					\
+do {									\
+	DEFINE_WAIT(__wait);						\
+									\
+	while (1) {							\
+		prepare_to_wait(&ca->set->alloc_wait,			\
+				&__wait, TASK_INTERRUPTIBLE);		\
+		if (cond)						\
+			break;						\
+									\
+		mutex_unlock(&(ca)->set->bucket_lock);			\
+		if (test_bit(CACHE_SET_STOPPING_2, &ca->set->flags)) {	\
+			finish_wait(&ca->set->alloc_wait, &__wait);	\
+			closure_return(cl);				\
+		}							\
+									\
+		schedule();						\
+		mutex_lock(&(ca)->set->bucket_lock);			\
+	}								\
+									\
+	finish_wait(&ca->set->alloc_wait, &__wait);			\
+} while (0)
+
+void bch_allocator_thread(struct closure *cl)
+{
+	struct cache *ca = container_of(cl, struct cache, alloc);
+
+	mutex_lock(&ca->set->bucket_lock);
+
+	while (1) {
+		/*
+		 * First, we pull buckets off of the unused and free_inc lists,
+		 * possibly issue discards to them, then we add the bucket to
+		 * the free list:
+		 */
+		while (1) {
+			long bucket;
+
+			if ((!atomic_read(&ca->set->prio_blocked) ||
+			     !CACHE_SYNC(&ca->set->sb)) &&
+			    !fifo_empty(&ca->unused))
+				fifo_pop(&ca->unused, bucket);
+			else if (!fifo_empty(&ca->free_inc))
+				fifo_pop(&ca->free_inc, bucket);
+			else
+				break;
+
+			allocator_wait(ca, (int) fifo_free(&ca->free) >
+				       atomic_read(&ca->discards_in_flight));
+
+			if (ca->discard) {
+				allocator_wait(ca, !list_empty(&ca->discards));
+				do_discard(ca, bucket);
+			} else {
+				fifo_push(&ca->free, bucket);
+				closure_wake_up(&ca->set->bucket_wait);
+			}
+		}
+
+		/*
+		 * We've run out of free buckets, we need to find some buckets
+		 * we can invalidate. First, invalidate them in memory and add
+		 * them to the free_inc list:
+		 */
+
+		allocator_wait(ca, ca->set->gc_mark_valid &&
+			       (ca->need_save_prio > 64 ||
+				!ca->invalidate_needs_gc));
+		invalidate_buckets(ca);
+
+		/*
+		 * Now, we write their new gens to disk so we can start writing
+		 * new stuff to them:
+		 */
+		allocator_wait(ca, !atomic_read(&ca->set->prio_blocked));
+		if (CACHE_SYNC(&ca->set->sb) &&
+		    (!fifo_empty(&ca->free_inc) ||
+		     ca->need_save_prio > 64))
+			bch_prio_write(ca);
+	}
+}
+
+long bch_bucket_alloc(struct cache *ca, unsigned watermark, struct closure *cl)
+{
+	long r = -1;
+again:
+	wake_up(&ca->set->alloc_wait);
+
+	if (fifo_used(&ca->free) > ca->watermark[watermark] &&
+	    fifo_pop(&ca->free, r)) {
+		struct bucket *b = ca->buckets + r;
+#ifdef CONFIG_BCACHE_EDEBUG
+		size_t iter;
+		long i;
+
+		for (iter = 0; iter < prio_buckets(ca) * 2; iter++)
+			BUG_ON(ca->prio_buckets[iter] == (uint64_t) r);
+
+		fifo_for_each(i, &ca->free, iter)
+			BUG_ON(i == r);
+		fifo_for_each(i, &ca->free_inc, iter)
+			BUG_ON(i == r);
+		fifo_for_each(i, &ca->unused, iter)
+			BUG_ON(i == r);
+#endif
+		BUG_ON(atomic_read(&b->pin) != 1);
+
+		SET_GC_SECTORS_USED(b, ca->sb.bucket_size);
+
+		if (watermark <= WATERMARK_METADATA) {
+			SET_GC_MARK(b, GC_MARK_METADATA);
+			b->prio = BTREE_PRIO;
+		} else {
+			SET_GC_MARK(b, GC_MARK_RECLAIMABLE);
+			b->prio = INITIAL_PRIO;
+		}
+
+		return r;
+	}
+
+	pr_debug("alloc failure: blocked %i free %zu free_inc %zu unused %zu",
+		 atomic_read(&ca->set->prio_blocked), fifo_used(&ca->free),
+		 fifo_used(&ca->free_inc), fifo_used(&ca->unused));
+
+	if (cl) {
+		closure_wait(&ca->set->bucket_wait, cl);
+
+		if (closure_blocking(cl)) {
+			mutex_unlock(&ca->set->bucket_lock);
+			closure_sync(cl);
+			mutex_lock(&ca->set->bucket_lock);
+			goto again;
+		}
+	}
+
+	return -1;
+}
+
+void bch_bucket_free(struct cache_set *c, struct bkey *k)
+{
+	unsigned i;
+
+	for (i = 0; i < KEY_PTRS(k); i++) {
+		struct bucket *b = PTR_BUCKET(c, k, i);
+
+		SET_GC_MARK(b, GC_MARK_RECLAIMABLE);
+		SET_GC_SECTORS_USED(b, 0);
+		bch_bucket_add_unused(PTR_CACHE(c, k, i), b);
+	}
+}
+
+int __bch_bucket_alloc_set(struct cache_set *c, unsigned watermark,
+			   struct bkey *k, int n, struct closure *cl)
+{
+	int i;
+
+	lockdep_assert_held(&c->bucket_lock);
+	BUG_ON(!n || n > c->caches_loaded || n > 8);
+
+	bkey_init(k);
+
+	/* sort by free space/prio of oldest data in caches */
+
+	for (i = 0; i < n; i++) {
+		struct cache *ca = c->cache_by_alloc[i];
+		long b = bch_bucket_alloc(ca, watermark, cl);
+
+		if (b == -1)
+			goto err;
+
+		k->ptr[i] = PTR(ca->buckets[b].gen,
+				bucket_to_sector(c, b),
+				ca->sb.nr_this_dev);
+
+		SET_KEY_PTRS(k, i + 1);
+	}
+
+	return 0;
+err:
+	bch_bucket_free(c, k);
+	__bkey_put(c, k);
+	return -1;
+}
+
+int bch_bucket_alloc_set(struct cache_set *c, unsigned watermark,
+			 struct bkey *k, int n, struct closure *cl)
+{
+	int ret;
+	mutex_lock(&c->bucket_lock);
+	ret = __bch_bucket_alloc_set(c, watermark, k, n, cl);
+	mutex_unlock(&c->bucket_lock);
+	return ret;
+}
+
+/* Init */
+
+void bch_cache_allocator_exit(struct cache *ca)
+{
+	struct discard *d;
+
+	while (!list_empty(&ca->discards)) {
+		d = list_first_entry(&ca->discards, struct discard, list);
+		cancel_work_sync(&d->work);
+		list_del(&d->list);
+		kfree(d);
+	}
+}
+
+int bch_cache_allocator_init(struct cache *ca)
+{
+	unsigned i;
+
+	/*
+	 * Reserve:
+	 * Prio/gen writes first
+	 * Then 8 for btree allocations
+	 * Then half for the moving garbage collector
+	 */
+
+	ca->watermark[WATERMARK_PRIO] = 0;
+
+	ca->watermark[WATERMARK_METADATA] = prio_buckets(ca);
+
+	ca->watermark[WATERMARK_MOVINGGC] = 8 +
+		ca->watermark[WATERMARK_METADATA];
+
+	ca->watermark[WATERMARK_NONE] = ca->free.size / 2 +
+		ca->watermark[WATERMARK_MOVINGGC];
+
+	for (i = 0; i < MAX_IN_FLIGHT_DISCARDS; i++) {
+		struct discard *d = kzalloc(sizeof(*d), GFP_KERNEL);
+		if (!d)
+			return -ENOMEM;
+
+		d->ca = ca;
+		INIT_WORK(&d->work, discard_finish);
+		list_add(&d->list, &ca->discards);
+	}
+
+	return 0;
+}

drivers/md/bcache/bcache.h

···
+#ifndef _BCACHE_H
+#define _BCACHE_H
+
+/*
+ * SOME HIGH LEVEL CODE DOCUMENTATION:
+ *
+ * Bcache mostly works with cache sets, cache devices, and backing devices.
+ *
+ * Support for multiple cache devices hasn't quite been finished off yet, but
+ * it's about 95% plumbed through. A cache set and its cache devices is sort of
+ * like a md raid array and its component devices. Most of the code doesn't care
+ * about individual cache devices, the main abstraction is the cache set.
+ *
+ * Multiple cache devices is intended to give us the ability to mirror dirty
+ * cached data and metadata, without mirroring clean cached data.
+ *
+ * Backing devices are different, in that they have a lifetime independent of a
+ * cache set. When you register a newly formatted backing device it'll come up
+ * in passthrough mode, and then you can attach and detach a backing device from
+ * a cache set at runtime - while it's mounted and in use. Detaching implicitly
+ * invalidates any cached data for that backing device.
+ *
+ * A cache set can have multiple (many) backing devices attached to it.
+ *
+ * There's also flash only volumes - this is the reason for the distinction
+ * between struct cached_dev and struct bcache_device. A flash only volume
+ * works much like a bcache device that has a backing device, except the
+ * "cached" data is always dirty. The end result is that we get thin
+ * provisioning with very little additional code.
+ *
+ * Flash only volumes work but they're not production ready because the moving
+ * garbage collector needs more work. More on that later.
+ *
+ * BUCKETS/ALLOCATION:
+ *
+ * Bcache is primarily designed for caching, which means that in normal
+ * operation all of our available space will be allocated. Thus, we need an
+ * efficient way of deleting things from the cache so we can write new things to
+ * it.
+ *
+ * To do this, we first divide the cache device up into buckets. A bucket is the
+ * unit of allocation; they're typically around 1 mb - anywhere from 128k to 2M+
+ * works efficiently.
+ *
+ * Each bucket has a 16 bit priority, and an 8 bit generation associated with
+ * it. The gens and priorities for all the buckets are stored contiguously and
+ * packed on disk (in a linked list of buckets - aside from the superblock, all
+ * of bcache's metadata is stored in buckets).
+ *
+ * The priority is used to implement an LRU. We reset a bucket's priority when
+ * we allocate it or on cache it, and every so often we decrement the priority
+ * of each bucket. It could be used to implement something more sophisticated,
+ * if anyone ever gets around to it.
+ *
+ * The generation is used for invalidating buckets. Each pointer also has an 8
+ * bit generation embedded in it; for a pointer to be considered valid, its gen
+ * must match the gen of the bucket it points into.  Thus, to reuse a bucket all
+ * we have to do is increment its gen (and write its new gen to disk; we batch
+ * this up).
+ *
+ * Bcache is entirely COW - we never write twice to a bucket, even buckets that
+ * contain metadata (including btree nodes).
+ *
+ * THE BTREE:
+ *
+ * Bcache is in large part design around the btree.
+ *
+ * At a high level, the btree is just an index of key -> ptr tuples.
+ *
+ * Keys represent extents, and thus have a size field. Keys also have a variable
+ * number of pointers attached to them (potentially zero, which is handy for
+ * invalidating the cache).
+ *
+ * The key itself is an inode:offset pair. The inode number corresponds to a
+ * backing device or a flash only volume. The offset is the ending offset of the
+ * extent within the inode - not the starting offset; this makes lookups
+ * slightly more convenient.
+ *
+ * Pointers contain the cache device id, the offset on that device, and an 8 bit
+ * generation number. More on the gen later.
+ *
+ * Index lookups are not fully abstracted - cache lookups in particular are
+ * still somewhat mixed in with the btree code, but things are headed in that
+ * direction.
+ *
+ * Updates are fairly well abstracted, though. There are two different ways of
+ * updating the btree; insert and replace.
+ *
+ * BTREE_INSERT will just take a list of keys and insert them into the btree -
+ * overwriting (possibly only partially) any extents they overlap with. This is
+ * used to update the index after a write.
+ *
+ * BTREE_REPLACE is really cmpxchg(); it inserts a key into the btree iff it is
+ * overwriting a key that matches another given key. This is used for inserting
+ * data into the cache after a cache miss, and for background writeback, and for
+ * the moving garbage collector.
+ *
+ * There is no "delete" operation; deleting things from the index is
+ * accomplished by either by invalidating pointers (by incrementing a bucket's
+ * gen) or by inserting a key with 0 pointers - which will overwrite anything
+ * previously present at that location in the index.
+ *
+ * This means that there are always stale/invalid keys in the btree. They're
+ * filtered out by the code that iterates through a btree node, and removed when
+ * a btree node is rewritten.
+ *
+ * BTREE NODES:
+ *
+ * Our unit of allocation is a bucket, and we we can't arbitrarily allocate and
+ * free smaller than a bucket - so, that's how big our btree nodes are.
+ *
+ * (If buckets are really big we'll only use part of the bucket for a btree node
+ * - no less than 1/4th - but a bucket still contains no more than a single
+ * btree node. I'd actually like to change this, but for now we rely on the
+ * bucket's gen for deleting btree nodes when we rewrite/split a node.)
+ *
+ * Anyways, btree nodes are big - big enough to be inefficient with a textbook
+ * btree implementation.
+ *
+ * The way this is solved is that btree nodes are internally log structured; we
+ * can append new keys to an existing btree node without rewriting it. This
+ * means each set of keys we write is sorted, but the node is not.
+ *
+ * We maintain this log structure in memory - keeping 1Mb of keys sorted would
+ * be expensive, and we have to distinguish between the keys we have written and
+ * the keys we haven't. So to do a lookup in a btree node, we have to search
+ * each sorted set. But we do merge written sets together lazily, so the cost of
+ * these extra searches is quite low (normally most of the keys in a btree node
+ * will be in one big set, and then there'll be one or two sets that are much
+ * smaller).
+ *
+ * This log structure makes bcache's btree more of a hybrid between a
+ * conventional btree and a compacting data structure, with some of the
+ * advantages of both.
+ *
+ * GARBAGE COLLECTION:
+ *
+ * We can't just invalidate any bucket - it might contain dirty data or
+ * metadata. If it once contained dirty data, other writes might overwrite it
+ * later, leaving no valid pointers into that bucket in the index.
+ *
+ * Thus, the primary purpose of garbage collection is to find buckets to reuse.
+ * It also counts how much valid data it each bucket currently contains, so that
+ * allocation can reuse buckets sooner when they've been mostly overwritten.
+ *
+ * It also does some things that are really internal to the btree
+ * implementation. If a btree node contains pointers that are stale by more than
+ * some threshold, it rewrites the btree node to avoid the bucket's generation
+ * wrapping around. It also merges adjacent btree nodes if they're empty enough.
+ *
+ * THE JOURNAL:
+ *
+ * Bcache's journal is not necessary for consistency; we always strictly
+ * order metadata writes so that the btree and everything else is consistent on
+ * disk in the event of an unclean shutdown, and in fact bcache had writeback
+ * caching (with recovery from unclean shutdown) before journalling was
+ * implemented.
+ *
+ * Rather, the journal is purely a performance optimization; we can't complete a
+ * write until we've updated the index on disk, otherwise the cache would be
+ * inconsistent in the event of an unclean shutdown. This means that without the
+ * journal, on random write workloads we constantly have to update all the leaf
+ * nodes in the btree, and those writes will be mostly empty (appending at most
+ * a few keys each) - highly inefficient in terms of amount of metadata writes,
+ * and it puts more strain on the various btree resorting/compacting code.
+ *
+ * The journal is just a log of keys we've inserted; on startup we just reinsert
+ * all the keys in the open journal entries. That means that when we're updating
+ * a node in the btree, we can wait until a 4k block of keys fills up before
+ * writing them out.
+ *
+ * For simplicity, we only journal updates to leaf nodes; updates to parent
+ * nodes are rare enough (since our leaf nodes are huge) that it wasn't worth
+ * the complexity to deal with journalling them (in particular, journal replay)
+ * - updates to non leaf nodes just happen synchronously (see btree_split()).
+ */
+
+#define pr_fmt(fmt) "bcache: %s() " fmt "\n", __func__
+
+#include <linux/bio.h>
+#include <linux/blktrace_api.h>
+#include <linux/kobject.h>
+#include <linux/list.h>
+#include <linux/mutex.h>
+#include <linux/rbtree.h>
+#include <linux/rwsem.h>
+#include <linux/types.h>
+#include <linux/workqueue.h>
+
+#include "util.h"
+#include "closure.h"
+
+struct bucket {
+	atomic_t	pin;
+	uint16_t	prio;
+	uint8_t		gen;
+	uint8_t		disk_gen;
+	uint8_t		last_gc; /* Most out of date gen in the btree */
+	uint8_t		gc_gen;
+	uint16_t	gc_mark;
+};
+
+/*
+ * I'd use bitfields for these, but I don't trust the compiler not to screw me
+ * as multiple threads touch struct bucket without locking
+ */
+
+BITMASK(GC_MARK,	 struct bucket, gc_mark, 0, 2);
+#define GC_MARK_RECLAIMABLE	0
+#define GC_MARK_DIRTY		1
+#define GC_MARK_METADATA	2
+BITMASK(GC_SECTORS_USED, struct bucket, gc_mark, 2, 14);
+
+struct bkey {
+	uint64_t	high;
+	uint64_t	low;
+	uint64_t	ptr[];
+};
+
+/* Enough for a key with 6 pointers */
+#define BKEY_PAD		8
+
+#define BKEY_PADDED(key)					\
+	union { struct bkey key; uint64_t key ## _pad[BKEY_PAD]; }
+
+/* Version 0: Cache device
+ * Version 1: Backing device
+ * Version 2: Seed pointer into btree node checksum
+ * Version 3: Cache device with new UUID format
+ * Version 4: Backing device with data offset
+ */
+#define BCACHE_SB_VERSION_CDEV			0
+#define BCACHE_SB_VERSION_BDEV			1
+#define BCACHE_SB_VERSION_CDEV_WITH_UUID	3
+#define BCACHE_SB_VERSION_BDEV_WITH_OFFSET	4
+#define BCACHE_SB_MAX_VERSION			4
+
+#define SB_SECTOR		8
+#define SB_SIZE			4096
+#define SB_LABEL_SIZE		32
+#define SB_JOURNAL_BUCKETS	256U
+/* SB_JOURNAL_BUCKETS must be divisible by BITS_PER_LONG */
+#define MAX_CACHES_PER_SET	8
+
+#define BDEV_DATA_START_DEFAULT	16	/* sectors */
+
+struct cache_sb {
+	uint64_t		csum;
+	uint64_t		offset;	/* sector where this sb was written */
+	uint64_t		version;
+
+	uint8_t			magic[16];
+
+	uint8_t			uuid[16];
+	union {
+		uint8_t		set_uuid[16];
+		uint64_t	set_magic;
+	};
+	uint8_t			label[SB_LABEL_SIZE];
+
+	uint64_t		flags;
+	uint64_t		seq;
+	uint64_t		pad[8];
+
+	union {
+	struct {
+		/* Cache devices */
+		uint64_t	nbuckets;	/* device size */
+
+		uint16_t	block_size;	/* sectors */
+		uint16_t	bucket_size;	/* sectors */
+
+		uint16_t	nr_in_set;
+		uint16_t	nr_this_dev;
+	};
+	struct {
+		/* Backing devices */
+		uint64_t	data_offset;
+
+		/*
+		 * block_size from the cache device section is still used by
+		 * backing devices, so don't add anything here until we fix
+		 * things to not need it for backing devices anymore
+		 */
+	};
+	};
+
+	uint32_t		last_mount;	/* time_t */
+
+	uint16_t		first_bucket;
+	union {
+		uint16_t	njournal_buckets;
+		uint16_t	keys;
+	};
+	uint64_t		d[SB_JOURNAL_BUCKETS];	/* journal buckets */
+};
+
+BITMASK(CACHE_SYNC,		struct cache_sb, flags, 0, 1);
+BITMASK(CACHE_DISCARD,		struct cache_sb, flags, 1, 1);
+BITMASK(CACHE_REPLACEMENT,	struct cache_sb, flags, 2, 3);
+#define CACHE_REPLACEMENT_LRU	0U
+#define CACHE_REPLACEMENT_FIFO	1U
+#define CACHE_REPLACEMENT_RANDOM 2U
+
+BITMASK(BDEV_CACHE_MODE,	struct cache_sb, flags, 0, 4);
+#define CACHE_MODE_WRITETHROUGH	0U
+#define CACHE_MODE_WRITEBACK	1U
+#define CACHE_MODE_WRITEAROUND	2U
+#define CACHE_MODE_NONE		3U
+BITMASK(BDEV_STATE,		struct cache_sb, flags, 61, 2);
+#define BDEV_STATE_NONE		0U
+#define BDEV_STATE_CLEAN	1U
+#define BDEV_STATE_DIRTY	2U
+#define BDEV_STATE_STALE	3U
+
+/* Version 1: Seed pointer into btree node checksum
+ */
+#define BCACHE_BSET_VERSION	1
+
+/*
+ * This is the on disk format for btree nodes - a btree node on disk is a list
+ * of these; within each set the keys are sorted
+ */
+struct bset {
+	uint64_t		csum;
+	uint64_t		magic;
+	uint64_t		seq;
+	uint32_t		version;
+	uint32_t		keys;
+
+	union {
+		struct bkey	start[0];
+		uint64_t	d[0];
+	};
+};
+
+/*
+ * On disk format for priorities and gens - see super.c near prio_write() for
+ * more.
+ */
+struct prio_set {
+	uint64_t		csum;
+	uint64_t		magic;
+	uint64_t		seq;
+	uint32_t		version;
+	uint32_t		pad;
+
+	uint64_t		next_bucket;
+
+	struct bucket_disk {
+		uint16_t	prio;
+		uint8_t		gen;
+	} __attribute((packed)) data[];
+};
+
+struct uuid_entry {
+	union {
+		struct {
+			uint8_t		uuid[16];
+			uint8_t		label[32];
+			uint32_t	first_reg;
+			uint32_t	last_reg;
+			uint32_t	invalidated;
+
+			uint32_t	flags;
+			/* Size of flash only volumes */
+			uint64_t	sectors;
+		};
+
+		uint8_t	pad[128];
+	};
+};
+
+BITMASK(UUID_FLASH_ONLY,	struct uuid_entry, flags, 0, 1);
+
+#include "journal.h"
+#include "stats.h"
+struct search;
+struct btree;
+struct keybuf;
+
+struct keybuf_key {
+	struct rb_node		node;
+	BKEY_PADDED(key);
+	void			*private;
+};
+
+typedef bool (keybuf_pred_fn)(struct keybuf *, struct bkey *);
+
+struct keybuf {
+	keybuf_pred_fn		*key_predicate;
+
+	struct bkey		last_scanned;
+	spinlock_t		lock;
+
+	/*
+	 * Beginning and end of range in rb tree - so that we can skip taking
+	 * lock and checking the rb tree when we need to check for overlapping
+	 * keys.
+	 */
+	struct bkey		start;
+	struct bkey		end;
+
+	struct rb_root		keys;
+
+#define KEYBUF_NR		100
+	DECLARE_ARRAY_ALLOCATOR(struct keybuf_key, freelist, KEYBUF_NR);
+};
+
+struct bio_split_pool {
+	struct bio_set		*bio_split;
+	mempool_t		*bio_split_hook;
+};
+
+struct bio_split_hook {
+	struct closure		cl;
+	struct bio_split_pool	*p;
+	struct bio		*bio;
+	bio_end_io_t		*bi_end_io;
+	void			*bi_private;
+};
+
+struct bcache_device {
+	struct closure		cl;
+
+	struct kobject		kobj;
+
+	struct cache_set	*c;
+	unsigned		id;
+#define BCACHEDEVNAME_SIZE	12
+	char			name[BCACHEDEVNAME_SIZE];
+
+	struct gendisk		*disk;
+
+	/* If nonzero, we're closing */
+	atomic_t		closing;
+
+	/* If nonzero, we're detaching/unregistering from cache set */
+	atomic_t		detaching;
+
+	atomic_long_t		sectors_dirty;
+	unsigned long		sectors_dirty_gc;
+	unsigned long		sectors_dirty_last;
+	long			sectors_dirty_derivative;
+
+	mempool_t		*unaligned_bvec;
+	struct bio_set		*bio_split;
+
+	unsigned		data_csum:1;
+
+	int (*cache_miss)(struct btree *, struct search *,
+			  struct bio *, unsigned);
+	int (*ioctl) (struct bcache_device *, fmode_t, unsigned, unsigned long);
+
+	struct bio_split_pool	bio_split_hook;
+};
+
+struct io {
+	/* Used to track sequential IO so it can be skipped */
+	struct hlist_node	hash;
+	struct list_head	lru;
+
+	unsigned long		jiffies;
+	unsigned		sequential;
+	sector_t		last;
+};
+
+struct cached_dev {
+	struct list_head	list;
+	struct bcache_device	disk;
+	struct block_device	*bdev;
+
+	struct cache_sb		sb;
+	struct bio		sb_bio;
+	struct bio_vec		sb_bv[1];
+	struct closure_with_waitlist sb_write;
+
+	/* Refcount on the cache set. Always nonzero when we're caching. */
+	atomic_t		count;
+	struct work_struct	detach;
+
+	/*
+	 * Device might not be running if it's dirty and the cache set hasn't
+	 * showed up yet.
+	 */
+	atomic_t		running;
+
+	/*
+	 * Writes take a shared lock from start to finish; scanning for dirty
+	 * data to refill the rb tree requires an exclusive lock.
+	 */
+	struct rw_semaphore	writeback_lock;
+
+	/*
+	 * Nonzero, and writeback has a refcount (d->count), iff there is dirty
+	 * data in the cache. Protected by writeback_lock; must have an
+	 * shared lock to set and exclusive lock to clear.
+	 */
+	atomic_t		has_dirty;
+
+	struct ratelimit	writeback_rate;
+	struct delayed_work	writeback_rate_update;
+
+	/*
+	 * Internal to the writeback code, so read_dirty() can keep track of
+	 * where it's at.
+	 */
+	sector_t		last_read;
+
+	/* Number of writeback bios in flight */
+	atomic_t		in_flight;
+	struct closure_with_timer writeback;
+	struct closure_waitlist	writeback_wait;
+
+	struct keybuf		writeback_keys;
+
+	/* For tracking sequential IO */
+#define RECENT_IO_BITS	7
+#define RECENT_IO	(1 << RECENT_IO_BITS)
+	struct io		io[RECENT_IO];
+	struct hlist_head	io_hash[RECENT_IO + 1];
+	struct list_head	io_lru;
+	spinlock_t		io_lock;
+
+	struct cache_accounting	accounting;
+
+	/* The rest of this all shows up in sysfs */
+	unsigned		sequential_cutoff;
+	unsigned		readahead;
+
+	unsigned		sequential_merge:1;
+	unsigned		verify:1;
+
+	unsigned		writeback_metadata:1;
+	unsigned		writeback_running:1;
+	unsigned char		writeback_percent;
+	unsigned		writeback_delay;
+
+	int			writeback_rate_change;
+	int64_t			writeback_rate_derivative;
+	uint64_t		writeback_rate_target;
+
+	unsigned		writeback_rate_update_seconds;
+	unsigned		writeback_rate_d_term;
+	unsigned		writeback_rate_p_term_inverse;
+	unsigned		writeback_rate_d_smooth;
+};
+
+enum alloc_watermarks {
+	WATERMARK_PRIO,
+	WATERMARK_METADATA,
+	WATERMARK_MOVINGGC,
+	WATERMARK_NONE,
+	WATERMARK_MAX
+};
+
+struct cache {
+	struct cache_set	*set;
+	struct cache_sb		sb;
+	struct bio		sb_bio;
+	struct bio_vec		sb_bv[1];
+
+	struct kobject		kobj;
+	struct block_device	*bdev;
+
+	unsigned		watermark[WATERMARK_MAX];
+
+	struct closure		alloc;
+	struct workqueue_struct	*alloc_workqueue;
+
+	struct closure		prio;
+	struct prio_set		*disk_buckets;
+
+	/*
+	 * When allocating new buckets, prio_write() gets first dibs - since we
+	 * may not be allocate at all without writing priorities and gens.
+	 * prio_buckets[] contains the last buckets we wrote priorities to (so
+	 * gc can mark them as metadata), prio_next[] contains the buckets
+	 * allocated for the next prio write.
+	 */
+	uint64_t		*prio_buckets;
+	uint64_t		*prio_last_buckets;
+
+	/*
+	 * free: Buckets that are ready to be used
+	 *
+	 * free_inc: Incoming buckets - these are buckets that currently have
+	 * cached data in them, and we can't reuse them until after we write
+	 * their new gen to disk. After prio_write() finishes writing the new
+	 * gens/prios, they'll be moved to the free list (and possibly discarded
+	 * in the process)
+	 *
+	 * unused: GC found nothing pointing into these buckets (possibly
+	 * because all the data they contained was overwritten), so we only
+	 * need to discard them before they can be moved to the free list.
+	 */
+	DECLARE_FIFO(long, free);
+	DECLARE_FIFO(long, free_inc);
+	DECLARE_FIFO(long, unused);
+
+	size_t			fifo_last_bucket;
+
+	/* Allocation stuff: */
+	struct bucket		*buckets;
+
+	DECLARE_HEAP(struct bucket *, heap);
+
+	/*
+	 * max(gen - disk_gen) for all buckets. When it gets too big we have to
+	 * call prio_write() to keep gens from wrapping.
+	 */
+	uint8_t			need_save_prio;
+	unsigned		gc_move_threshold;
+
+	/*
+	 * If nonzero, we know we aren't going to find any buckets to invalidate
+	 * until a gc finishes - otherwise we could pointlessly burn a ton of
+	 * cpu
+	 */
+	unsigned		invalidate_needs_gc:1;
+
+	bool			discard; /* Get rid of? */
+
+	/*
+	 * We preallocate structs for issuing discards to buckets, and keep them
+	 * on this list when they're not in use; do_discard() issues discards
+	 * whenever there's work to do and is called by free_some_buckets() and
+	 * when a discard finishes.
+	 */
+	atomic_t		discards_in_flight;
+	struct list_head	discards;
+
+	struct journal_device	journal;
+
+	/* The rest of this all shows up in sysfs */
+#define IO_ERROR_SHIFT		20
+	atomic_t		io_errors;
+	atomic_t		io_count;
+
+	atomic_long_t		meta_sectors_written;
+	atomic_long_t		btree_sectors_written;
+	atomic_long_t		sectors_written;
+
+	struct bio_split_pool	bio_split_hook;
+};
+
+struct gc_stat {
+	size_t			nodes;
+	size_t			key_bytes;
+
+	size_t			nkeys;
+	uint64_t		data;	/* sectors */
+	uint64_t		dirty;	/* sectors */
+	unsigned		in_use; /* percent */
+};
+
+/*
+ * Flag bits, for how the cache set is shutting down, and what phase it's at:
+ *
+ * CACHE_SET_UNREGISTERING means we're not just shutting down, we're detaching
+ * all the backing devices first (their cached data gets invalidated, and they
+ * won't automatically reattach).
+ *
+ * CACHE_SET_STOPPING always gets set first when we're closing down a cache set;
+ * we'll continue to run normally for awhile with CACHE_SET_STOPPING set (i.e.
+ * flushing dirty data).
+ *
+ * CACHE_SET_STOPPING_2 gets set at the last phase, when it's time to shut down
+ * the allocation thread.
+ */
+#define CACHE_SET_UNREGISTERING		0
+#define	CACHE_SET_STOPPING		1
+#define	CACHE_SET_STOPPING_2		2
+
+struct cache_set {
+	struct closure		cl;
+
+	struct list_head	list;
+	struct kobject		kobj;
+	struct kobject		internal;
+	struct dentry		*debug;
+	struct cache_accounting accounting;
+
+	unsigned long		flags;
+
+	struct cache_sb		sb;
+
+	struct cache		*cache[MAX_CACHES_PER_SET];
+	struct cache		*cache_by_alloc[MAX_CACHES_PER_SET];
+	int			caches_loaded;
+
+	struct bcache_device	**devices;
+	struct list_head	cached_devs;
+	uint64_t		cached_dev_sectors;
+	struct closure		caching;
+
+	struct closure_with_waitlist sb_write;
+
+	mempool_t		*search;
+	mempool_t		*bio_meta;
+	struct bio_set		*bio_split;
+
+	/* For the btree cache */
+	struct shrinker		shrink;
+
+	/* For the allocator itself */
+	wait_queue_head_t	alloc_wait;
+
+	/* For the btree cache and anything allocation related */
+	struct mutex		bucket_lock;
+
+	/* log2(bucket_size), in sectors */
+	unsigned short		bucket_bits;
+
+	/* log2(block_size), in sectors */
+	unsigned short		block_bits;
+
+	/*
+	 * Default number of pages for a new btree node - may be less than a
+	 * full bucket
+	 */
+	unsigned		btree_pages;
+
+	/*
+	 * Lists of struct btrees; lru is the list for structs that have memory
+	 * allocated for actual btree node, freed is for structs that do not.
+	 *
+	 * We never free a struct btree, except on shutdown - we just put it on
+	 * the btree_cache_freed list and reuse it later. This simplifies the
+	 * code, and it doesn't cost us much memory as the memory usage is
+	 * dominated by buffers that hold the actual btree node data and those
+	 * can be freed - and the number of struct btrees allocated is
+	 * effectively bounded.
+	 *
+	 * btree_cache_freeable effectively is a small cache - we use it because
+	 * high order page allocations can be rather expensive, and it's quite
+	 * common to delete and allocate btree nodes in quick succession. It
+	 * should never grow past ~2-3 nodes in practice.
+	 */
+	struct list_head	btree_cache;
+	struct list_head	btree_cache_freeable;
+	struct list_head	btree_cache_freed;
+
+	/* Number of elements in btree_cache + btree_cache_freeable lists */
+	unsigned		bucket_cache_used;
+
+	/*
+	 * If we need to allocate memory for a new btree node and that
+	 * allocation fails, we can cannibalize another node in the btree cache
+	 * to satisfy the allocation. However, only one thread can be doing this
+	 * at a time, for obvious reasons - try_harder and try_wait are
+	 * basically a lock for this that we can wait on asynchronously. The
+	 * btree_root() macro releases the lock when it returns.
+	 */
+	struct closure		*try_harder;
+	struct closure_waitlist	try_wait;
+	uint64_t		try_harder_start;
+
+	/*
+	 * When we free a btree node, we increment the gen of the bucket the
+	 * node is in - but we can't rewrite the prios and gens until we
+	 * finished whatever it is we were doing, otherwise after a crash the
+	 * btree node would be freed but for say a split, we might not have the
+	 * pointers to the new nodes inserted into the btree yet.
+	 *
+	 * This is a refcount that blocks prio_write() until the new keys are
+	 * written.
+	 */
+	atomic_t		prio_blocked;
+	struct closure_waitlist	bucket_wait;
+
+	/*
+	 * For any bio we don't skip we subtract the number of sectors from
+	 * rescale; when it hits 0 we rescale all the bucket priorities.
+	 */
+	atomic_t		rescale;
+	/*
+	 * When we invalidate buckets, we use both the priority and the amount
+	 * of good data to determine which buckets to reuse first - to weight
+	 * those together consistently we keep track of the smallest nonzero
+	 * priority of any bucket.
+	 */
+	uint16_t		min_prio;
+
+	/*
+	 * max(gen - gc_gen) for all buckets. When it gets too big we have to gc
+	 * to keep gens from wrapping around.
+	 */
+	uint8_t			need_gc;
+	struct gc_stat		gc_stats;
+	size_t			nbuckets;
+
+	struct closure_with_waitlist gc;
+	/* Where in the btree gc currently is */
+	struct bkey		gc_done;
+
+	/*
+	 * The allocation code needs gc_mark in struct bucket to be correct, but
+	 * it's not while a gc is in progress. Protected by bucket_lock.
+	 */
+	int			gc_mark_valid;
+
+	/* Counts how many sectors bio_insert has added to the cache */
+	atomic_t		sectors_to_gc;
+
+	struct closure		moving_gc;
+	struct closure_waitlist	moving_gc_wait;
+	struct keybuf		moving_gc_keys;
+	/* Number of moving GC bios in flight */
+	atomic_t		in_flight;
+
+	struct btree		*root;
+
+#ifdef CONFIG_BCACHE_DEBUG
+	struct btree		*verify_data;
+	struct mutex		verify_lock;
+#endif
+
+	unsigned		nr_uuids;
+	struct uuid_entry	*uuids;
+	BKEY_PADDED(uuid_bucket);
+	struct closure_with_waitlist uuid_write;
+
+	/*
+	 * A btree node on disk could have too many bsets for an iterator to fit
+	 * on the stack - this is a single element mempool for btree_read_work()
+	 */
+	struct mutex		fill_lock;
+	struct btree_iter	*fill_iter;
+
+	/*
+	 * btree_sort() is a merge sort and requires temporary space - single
+	 * element mempool
+	 */
+	struct mutex		sort_lock;
+	struct bset		*sort;
+
+	/* List of buckets we're currently writing data to */
+	struct list_head	data_buckets;
+	spinlock_t		data_bucket_lock;
+
+	struct journal		journal;
+
+#define CONGESTED_MAX		1024
+	unsigned		congested_last_us;
+	atomic_t		congested;
+
+	/* The rest of this all shows up in sysfs */
+	unsigned		congested_read_threshold_us;
+	unsigned		congested_write_threshold_us;
+
+	spinlock_t		sort_time_lock;
+	struct time_stats	sort_time;
+	struct time_stats	btree_gc_time;
+	struct time_stats	btree_split_time;
+	spinlock_t		btree_read_time_lock;
+	struct time_stats	btree_read_time;
+	struct time_stats	try_harder_time;
+
+	atomic_long_t		cache_read_races;
+	atomic_long_t		writeback_keys_done;
+	atomic_long_t		writeback_keys_failed;
+	unsigned		error_limit;
+	unsigned		error_decay;
+	unsigned short		journal_delay_ms;
+	unsigned		verify:1;
+	unsigned		key_merging_disabled:1;
+	unsigned		gc_always_rewrite:1;
+	unsigned		shrinker_disabled:1;
+	unsigned		copy_gc_enabled:1;
+
+#define BUCKET_HASH_BITS	12
+	struct hlist_head	bucket_hash[1 << BUCKET_HASH_BITS];
+};
+
+static inline bool key_merging_disabled(struct cache_set *c)
+{
+#ifdef CONFIG_BCACHE_DEBUG
+	return c->key_merging_disabled;
+#else
+	return 0;
+#endif
+}
+
+static inline bool SB_IS_BDEV(const struct cache_sb *sb)
+{
+	return sb->version == BCACHE_SB_VERSION_BDEV
+		|| sb->version == BCACHE_SB_VERSION_BDEV_WITH_OFFSET;
+}
+
+struct bbio {
+	unsigned		submit_time_us;
+	union {
+		struct bkey	key;
+		uint64_t	_pad[3];
+		/*
+		 * We only need pad = 3 here because we only ever carry around a
+		 * single pointer - i.e. the pointer we're doing io to/from.
+		 */
+	};
+	struct bio		bio;
+};
+
+static inline unsigned local_clock_us(void)
+{
+	return local_clock() >> 10;
+}
+
+#define MAX_BSETS		4U
+
+#define BTREE_PRIO		USHRT_MAX
+#define INITIAL_PRIO		32768
+
+#define btree_bytes(c)		((c)->btree_pages * PAGE_SIZE)
+#define btree_blocks(b)							\
+	((unsigned) (KEY_SIZE(&b->key) >> (b)->c->block_bits))
+
+#define btree_default_blocks(c)						\
+	((unsigned) ((PAGE_SECTORS * (c)->btree_pages) >> (c)->block_bits))
+
+#define bucket_pages(c)		((c)->sb.bucket_size / PAGE_SECTORS)
+#define bucket_bytes(c)		((c)->sb.bucket_size << 9)
+#define block_bytes(c)		((c)->sb.block_size << 9)
+
+#define __set_bytes(i, k)	(sizeof(*(i)) + (k) * sizeof(uint64_t))
+#define set_bytes(i)		__set_bytes(i, i->keys)
+
+#define __set_blocks(i, k, c)	DIV_ROUND_UP(__set_bytes(i, k), block_bytes(c))
+#define set_blocks(i, c)	__set_blocks(i, (i)->keys, c)
+
+#define node(i, j)		((struct bkey *) ((i)->d + (j)))
+#define end(i)			node(i, (i)->keys)
+
+#define index(i, b)							\
+	((size_t) (((void *) i - (void *) (b)->sets[0].data) /		\
+		   block_bytes(b->c)))
+
+#define btree_data_space(b)	(PAGE_SIZE << (b)->page_order)
+
+#define prios_per_bucket(c)				\
+	((bucket_bytes(c) - sizeof(struct prio_set)) /	\
+	 sizeof(struct bucket_disk))
+#define prio_buckets(c)					\
+	DIV_ROUND_UP((size_t) (c)->sb.nbuckets, prios_per_bucket(c))
+
+#define JSET_MAGIC		0x245235c1a3625032ULL
+#define PSET_MAGIC		0x6750e15f87337f91ULL
+#define BSET_MAGIC		0x90135c78b99e07f5ULL
+
+#define jset_magic(c)		((c)->sb.set_magic ^ JSET_MAGIC)
+#define pset_magic(c)		((c)->sb.set_magic ^ PSET_MAGIC)
+#define bset_magic(c)		((c)->sb.set_magic ^ BSET_MAGIC)
+
+/* Bkey fields: all units are in sectors */
+
+#define KEY_FIELD(name, field, offset, size)				\
+	BITMASK(name, struct bkey, field, offset, size)
+
+#define PTR_FIELD(name, offset, size)					\
+	static inline uint64_t name(const struct bkey *k, unsigned i)	\
+	{ return (k->ptr[i] >> offset) & ~(((uint64_t) ~0) << size); }	\
+									\
+	static inline void SET_##name(struct bkey *k, unsigned i, uint64_t v)\
+	{								\
+		k->ptr[i] &= ~(~((uint64_t) ~0 << size) << offset);	\
+		k->ptr[i] |= v << offset;				\
+	}
+
+KEY_FIELD(KEY_PTRS,	high, 60, 3)
+KEY_FIELD(HEADER_SIZE,	high, 58, 2)
+KEY_FIELD(KEY_CSUM,	high, 56, 2)
+KEY_FIELD(KEY_PINNED,	high, 55, 1)
+KEY_FIELD(KEY_DIRTY,	high, 36, 1)
+
+KEY_FIELD(KEY_SIZE,	high, 20, 16)
+KEY_FIELD(KEY_INODE,	high, 0,  20)
+
+/* Next time I change the on disk format, KEY_OFFSET() won't be 64 bits */
+
+static inline uint64_t KEY_OFFSET(const struct bkey *k)
+{
+	return k->low;
+}
+
+static inline void SET_KEY_OFFSET(struct bkey *k, uint64_t v)
+{
+	k->low = v;
+}
+
+PTR_FIELD(PTR_DEV,		51, 12)
+PTR_FIELD(PTR_OFFSET,		8,  43)
+PTR_FIELD(PTR_GEN,		0,  8)
+
+#define PTR_CHECK_DEV		((1 << 12) - 1)
+
+#define PTR(gen, offset, dev)						\
+	((((uint64_t) dev) << 51) | ((uint64_t) offset) << 8 | gen)
+
+static inline size_t sector_to_bucket(struct cache_set *c, sector_t s)
+{
+	return s >> c->bucket_bits;
+}
+
+static inline sector_t bucket_to_sector(struct cache_set *c, size_t b)
+{
+	return ((sector_t) b) << c->bucket_bits;
+}
+
+static inline sector_t bucket_remainder(struct cache_set *c, sector_t s)
+{
+	return s & (c->sb.bucket_size - 1);
+}
+
+static inline struct cache *PTR_CACHE(struct cache_set *c,
+				      const struct bkey *k,
+				      unsigned ptr)
+{
+	return c->cache[PTR_DEV(k, ptr)];
+}
+
+static inline size_t PTR_BUCKET_NR(struct cache_set *c,
+				   const struct bkey *k,
+				   unsigned ptr)
+{
+	return sector_to_bucket(c, PTR_OFFSET(k, ptr));
+}
+
+static inline struct bucket *PTR_BUCKET(struct cache_set *c,
+					const struct bkey *k,
+					unsigned ptr)
+{
+	return PTR_CACHE(c, k, ptr)->buckets + PTR_BUCKET_NR(c, k, ptr);
+}
+
+/* Btree key macros */
+
+/*
+ * The high bit being set is a relic from when we used it to do binary
+ * searches - it told you where a key started. It's not used anymore,
+ * and can probably be safely dropped.
+ */
+#define KEY(dev, sector, len)						\
+((struct bkey) {							\
+	.high = (1ULL << 63) | ((uint64_t) (len) << 20) | (dev),	\
+	.low = (sector)							\
+})
+
+static inline void bkey_init(struct bkey *k)
+{
+	*k = KEY(0, 0, 0);
+}
+
+#define KEY_START(k)		(KEY_OFFSET(k) - KEY_SIZE(k))
+#define START_KEY(k)		KEY(KEY_INODE(k), KEY_START(k), 0)
+#define MAX_KEY			KEY(~(~0 << 20), ((uint64_t) ~0) >> 1, 0)
+#define ZERO_KEY		KEY(0, 0, 0)
+
+/*
+ * This is used for various on disk data structures - cache_sb, prio_set, bset,
+ * jset: The checksum is _always_ the first 8 bytes of these structs
+ */
+#define csum_set(i)							\
+	bch_crc64(((void *) (i)) + sizeof(uint64_t),			\
+	      ((void *) end(i)) - (((void *) (i)) + sizeof(uint64_t)))
+
+/* Error handling macros */
+
+#define btree_bug(b, ...)						\
+do {									\
+	if (bch_cache_set_error((b)->c, __VA_ARGS__))			\
+		dump_stack();						\
+} while (0)
+
+#define cache_bug(c, ...)						\
+do {									\
+	if (bch_cache_set_error(c, __VA_ARGS__))			\
+		dump_stack();						\
+} while (0)
+
+#define btree_bug_on(cond, b, ...)					\
+do {									\
+	if (cond)							\
+		btree_bug(b, __VA_ARGS__);				\
+} while (0)
+
+#define cache_bug_on(cond, c, ...)					\
+do {									\
+	if (cond)							\
+		cache_bug(c, __VA_ARGS__);				\
+} while (0)
+
+#define cache_set_err_on(cond, c, ...)					\
+do {									\
+	if (cond)							\
+		bch_cache_set_error(c, __VA_ARGS__);			\
+} while (0)
+
+/* Looping macros */
+
+#define for_each_cache(ca, cs, iter)					\
+	for (iter = 0; ca = cs->cache[iter], iter < (cs)->sb.nr_in_set; iter++)
+
+#define for_each_bucket(b, ca)						\
+	for (b = (ca)->buckets + (ca)->sb.first_bucket;			\
+	     b < (ca)->buckets + (ca)->sb.nbuckets; b++)
+
+static inline void __bkey_put(struct cache_set *c, struct bkey *k)
+{
+	unsigned i;
+
+	for (i = 0; i < KEY_PTRS(k); i++)
+		atomic_dec_bug(&PTR_BUCKET(c, k, i)->pin);
+}
+
+/* Blktrace macros */
+
+#define blktrace_msg(c, fmt, ...)					\
+do {									\
+	struct request_queue *q = bdev_get_queue(c->bdev);		\
+	if (q)								\
+		blk_add_trace_msg(q, fmt, ##__VA_ARGS__);		\
+} while (0)
+
+#define blktrace_msg_all(s, fmt, ...)					\
+do {									\
+	struct cache *_c;						\
+	unsigned i;							\
+	for_each_cache(_c, (s), i)					\
+		blktrace_msg(_c, fmt, ##__VA_ARGS__);			\
+} while (0)
+
+static inline void cached_dev_put(struct cached_dev *dc)
+{
+	if (atomic_dec_and_test(&dc->count))
+		schedule_work(&dc->detach);
+}
+
+static inline bool cached_dev_get(struct cached_dev *dc)
+{
+	if (!atomic_inc_not_zero(&dc->count))
+		return false;
+
+	/* Paired with the mb in cached_dev_attach */
+	smp_mb__after_atomic_inc();
+	return true;
+}
+
+/*
+ * bucket_gc_gen() returns the difference between the bucket's current gen and
+ * the oldest gen of any pointer into that bucket in the btree (last_gc).
+ *
+ * bucket_disk_gen() returns the difference between the current gen and the gen
+ * on disk; they're both used to make sure gens don't wrap around.
+ */
+
+static inline uint8_t bucket_gc_gen(struct bucket *b)
+{
+	return b->gen - b->last_gc;
+}
+
+static inline uint8_t bucket_disk_gen(struct bucket *b)
+{
+	return b->gen - b->disk_gen;
+}
+
+#define BUCKET_GC_GEN_MAX	96U
+#define BUCKET_DISK_GEN_MAX	64U
+
+#define kobj_attribute_write(n, fn)					\
+	static struct kobj_attribute ksysfs_##n = __ATTR(n, S_IWUSR, NULL, fn)
+
+#define kobj_attribute_rw(n, show, store)				\
+	static struct kobj_attribute ksysfs_##n =			\
+		__ATTR(n, S_IWUSR|S_IRUSR, show, store)
+
+/* Forward declarations */
+
+void bch_writeback_queue(struct cached_dev *);
+void bch_writeback_add(struct cached_dev *, unsigned);
+
+void bch_count_io_errors(struct cache *, int, const char *);
+void bch_bbio_count_io_errors(struct cache_set *, struct bio *,
+			      int, const char *);
+void bch_bbio_endio(struct cache_set *, struct bio *, int, const char *);
+void bch_bbio_free(struct bio *, struct cache_set *);
+struct bio *bch_bbio_alloc(struct cache_set *);
+
+struct bio *bch_bio_split(struct bio *, int, gfp_t, struct bio_set *);
+void bch_generic_make_request(struct bio *, struct bio_split_pool *);
+void __bch_submit_bbio(struct bio *, struct cache_set *);
+void bch_submit_bbio(struct bio *, struct cache_set *, struct bkey *, unsigned);
+
+uint8_t bch_inc_gen(struct cache *, struct bucket *);
+void bch_rescale_priorities(struct cache_set *, int);
+bool bch_bucket_add_unused(struct cache *, struct bucket *);
+void bch_allocator_thread(struct closure *);
+
+long bch_bucket_alloc(struct cache *, unsigned, struct closure *);
+void bch_bucket_free(struct cache_set *, struct bkey *);
+
+int __bch_bucket_alloc_set(struct cache_set *, unsigned,
+			   struct bkey *, int, struct closure *);
+int bch_bucket_alloc_set(struct cache_set *, unsigned,
+			 struct bkey *, int, struct closure *);
+
+__printf(2, 3)
+bool bch_cache_set_error(struct cache_set *, const char *, ...);
+
+void bch_prio_write(struct cache *);
+void bch_write_bdev_super(struct cached_dev *, struct closure *);
+
+extern struct workqueue_struct *bcache_wq, *bch_gc_wq;
+extern const char * const bch_cache_modes[];
+extern struct mutex bch_register_lock;
+extern struct list_head bch_cache_sets;
+
+extern struct kobj_type bch_cached_dev_ktype;
+extern struct kobj_type bch_flash_dev_ktype;
+extern struct kobj_type bch_cache_set_ktype;
+extern struct kobj_type bch_cache_set_internal_ktype;
+extern struct kobj_type bch_cache_ktype;
+
+void bch_cached_dev_release(struct kobject *);
+void bch_flash_dev_release(struct kobject *);
+void bch_cache_set_release(struct kobject *);
+void bch_cache_release(struct kobject *);
+
+int bch_uuid_write(struct cache_set *);
+void bcache_write_super(struct cache_set *);
+
+int bch_flash_dev_create(struct cache_set *c, uint64_t size);
+
+int bch_cached_dev_attach(struct cached_dev *, struct cache_set *);
+void bch_cached_dev_detach(struct cached_dev *);
+void bch_cached_dev_run(struct cached_dev *);
+void bcache_device_stop(struct bcache_device *);
+
+void bch_cache_set_unregister(struct cache_set *);
+void bch_cache_set_stop(struct cache_set *);
+
+struct cache_set *bch_cache_set_alloc(struct cache_sb *);
+void bch_btree_cache_free(struct cache_set *);
+int bch_btree_cache_alloc(struct cache_set *);
+void bch_writeback_init_cached_dev(struct cached_dev *);
+void bch_moving_init_cache_set(struct cache_set *);
+
+void bch_cache_allocator_exit(struct cache *ca);
+int bch_cache_allocator_init(struct cache *ca);
+
+void bch_debug_exit(void);
+int bch_debug_init(struct kobject *);
+void bch_writeback_exit(void);
+int bch_writeback_init(void);
+void bch_request_exit(void);
+int bch_request_init(void);
+void bch_btree_exit(void);
+int bch_btree_init(void);
+
+#endif /* _BCACHE_H */

drivers/md/bcache/bset.c

···
+/*
+ * Code for working with individual keys, and sorted sets of keys with in a
+ * btree node
+ *
+ * Copyright 2012 Google, Inc.
+ */
+
+#include "bcache.h"
+#include "btree.h"
+#include "debug.h"
+
+#include <linux/random.h>
+#include <linux/prefetch.h>
+
+/* Keylists */
+
+void bch_keylist_copy(struct keylist *dest, struct keylist *src)
+{
+	*dest = *src;
+
+	if (src->list == src->d) {
+		size_t n = (uint64_t *) src->top - src->d;
+		dest->top = (struct bkey *) &dest->d[n];
+		dest->list = dest->d;
+	}
+}
+
+int bch_keylist_realloc(struct keylist *l, int nptrs, struct cache_set *c)
+{
+	unsigned oldsize = (uint64_t *) l->top - l->list;
+	unsigned newsize = oldsize + 2 + nptrs;
+	uint64_t *new;
+
+	/* The journalling code doesn't handle the case where the keys to insert
+	 * is bigger than an empty write: If we just return -ENOMEM here,
+	 * bio_insert() and bio_invalidate() will insert the keys created so far
+	 * and finish the rest when the keylist is empty.
+	 */
+	if (newsize * sizeof(uint64_t) > block_bytes(c) - sizeof(struct jset))
+		return -ENOMEM;
+
+	newsize = roundup_pow_of_two(newsize);
+
+	if (newsize <= KEYLIST_INLINE ||
+	    roundup_pow_of_two(oldsize) == newsize)
+		return 0;
+
+	new = krealloc(l->list == l->d ? NULL : l->list,
+		       sizeof(uint64_t) * newsize, GFP_NOIO);
+
+	if (!new)
+		return -ENOMEM;
+
+	if (l->list == l->d)
+		memcpy(new, l->list, sizeof(uint64_t) * KEYLIST_INLINE);
+
+	l->list = new;
+	l->top = (struct bkey *) (&l->list[oldsize]);
+
+	return 0;
+}
+
+struct bkey *bch_keylist_pop(struct keylist *l)
+{
+	struct bkey *k = l->bottom;
+
+	if (k == l->top)
+		return NULL;
+
+	while (bkey_next(k) != l->top)
+		k = bkey_next(k);
+
+	return l->top = k;
+}
+
+/* Pointer validation */
+
+bool __bch_ptr_invalid(struct cache_set *c, int level, const struct bkey *k)
+{
+	unsigned i;
+
+	if (level && (!KEY_PTRS(k) || !KEY_SIZE(k) || KEY_DIRTY(k)))
+		goto bad;
+
+	if (!level && KEY_SIZE(k) > KEY_OFFSET(k))
+		goto bad;
+
+	if (!KEY_SIZE(k))
+		return true;
+
+	for (i = 0; i < KEY_PTRS(k); i++)
+		if (ptr_available(c, k, i)) {
+			struct cache *ca = PTR_CACHE(c, k, i);
+			size_t bucket = PTR_BUCKET_NR(c, k, i);
+			size_t r = bucket_remainder(c, PTR_OFFSET(k, i));
+
+			if (KEY_SIZE(k) + r > c->sb.bucket_size ||
+			    bucket <  ca->sb.first_bucket ||
+			    bucket >= ca->sb.nbuckets)
+				goto bad;
+		}
+
+	return false;
+bad:
+	cache_bug(c, "spotted bad key %s: %s", pkey(k), bch_ptr_status(c, k));
+	return true;
+}
+
+bool bch_ptr_bad(struct btree *b, const struct bkey *k)
+{
+	struct bucket *g;
+	unsigned i, stale;
+
+	if (!bkey_cmp(k, &ZERO_KEY) ||
+	    !KEY_PTRS(k) ||
+	    bch_ptr_invalid(b, k))
+		return true;
+
+	if (KEY_PTRS(k) && PTR_DEV(k, 0) == PTR_CHECK_DEV)
+		return true;
+
+	for (i = 0; i < KEY_PTRS(k); i++)
+		if (ptr_available(b->c, k, i)) {
+			g = PTR_BUCKET(b->c, k, i);
+			stale = ptr_stale(b->c, k, i);
+
+			btree_bug_on(stale > 96, b,
+				     "key too stale: %i, need_gc %u",
+				     stale, b->c->need_gc);
+
+			btree_bug_on(stale && KEY_DIRTY(k) && KEY_SIZE(k),
+				     b, "stale dirty pointer");
+
+			if (stale)
+				return true;
+
+#ifdef CONFIG_BCACHE_EDEBUG
+			if (!mutex_trylock(&b->c->bucket_lock))
+				continue;
+
+			if (b->level) {
+				if (KEY_DIRTY(k) ||
+				    g->prio != BTREE_PRIO ||
+				    (b->c->gc_mark_valid &&
+				     GC_MARK(g) != GC_MARK_METADATA))
+					goto bug;
+
+			} else {
+				if (g->prio == BTREE_PRIO)
+					goto bug;
+
+				if (KEY_DIRTY(k) &&
+				    b->c->gc_mark_valid &&
+				    GC_MARK(g) != GC_MARK_DIRTY)
+					goto bug;
+			}
+			mutex_unlock(&b->c->bucket_lock);
+#endif
+		}
+
+	return false;
+#ifdef CONFIG_BCACHE_EDEBUG
+bug:
+	mutex_unlock(&b->c->bucket_lock);
+	btree_bug(b,
+"inconsistent pointer %s: bucket %zu pin %i prio %i gen %i last_gc %i mark %llu gc_gen %i",
+		  pkey(k), PTR_BUCKET_NR(b->c, k, i), atomic_read(&g->pin),
+		  g->prio, g->gen, g->last_gc, GC_MARK(g), g->gc_gen);
+	return true;
+#endif
+}
+
+/* Key/pointer manipulation */
+
+void bch_bkey_copy_single_ptr(struct bkey *dest, const struct bkey *src,
+			      unsigned i)
+{
+	BUG_ON(i > KEY_PTRS(src));
+
+	/* Only copy the header, key, and one pointer. */
+	memcpy(dest, src, 2 * sizeof(uint64_t));
+	dest->ptr[0] = src->ptr[i];
+	SET_KEY_PTRS(dest, 1);
+	/* We didn't copy the checksum so clear that bit. */
+	SET_KEY_CSUM(dest, 0);
+}
+
+bool __bch_cut_front(const struct bkey *where, struct bkey *k)
+{
+	unsigned i, len = 0;
+
+	if (bkey_cmp(where, &START_KEY(k)) <= 0)
+		return false;
+
+	if (bkey_cmp(where, k) < 0)
+		len = KEY_OFFSET(k) - KEY_OFFSET(where);
+	else
+		bkey_copy_key(k, where);
+
+	for (i = 0; i < KEY_PTRS(k); i++)
+		SET_PTR_OFFSET(k, i, PTR_OFFSET(k, i) + KEY_SIZE(k) - len);
+
+	BUG_ON(len > KEY_SIZE(k));
+	SET_KEY_SIZE(k, len);
+	return true;
+}
+
+bool __bch_cut_back(const struct bkey *where, struct bkey *k)
+{
+	unsigned len = 0;
+
+	if (bkey_cmp(where, k) >= 0)
+		return false;
+
+	BUG_ON(KEY_INODE(where) != KEY_INODE(k));
+
+	if (bkey_cmp(where, &START_KEY(k)) > 0)
+		len = KEY_OFFSET(where) - KEY_START(k);
+
+	bkey_copy_key(k, where);
+
+	BUG_ON(len > KEY_SIZE(k));
+	SET_KEY_SIZE(k, len);
+	return true;
+}
+
+static uint64_t merge_chksums(struct bkey *l, struct bkey *r)
+{
+	return (l->ptr[KEY_PTRS(l)] + r->ptr[KEY_PTRS(r)]) &
+		~((uint64_t)1 << 63);
+}
+
+/* Tries to merge l and r: l should be lower than r
+ * Returns true if we were able to merge. If we did merge, l will be the merged
+ * key, r will be untouched.
+ */
+bool bch_bkey_try_merge(struct btree *b, struct bkey *l, struct bkey *r)
+{
+	unsigned i;
+
+	if (key_merging_disabled(b->c))
+		return false;
+
+	if (KEY_PTRS(l) != KEY_PTRS(r) ||
+	    KEY_DIRTY(l) != KEY_DIRTY(r) ||
+	    bkey_cmp(l, &START_KEY(r)))
+		return false;
+
+	for (i = 0; i < KEY_PTRS(l); i++)
+		if (l->ptr[i] + PTR(0, KEY_SIZE(l), 0) != r->ptr[i] ||
+		    PTR_BUCKET_NR(b->c, l, i) != PTR_BUCKET_NR(b->c, r, i))
+			return false;
+
+	/* Keys with no pointers aren't restricted to one bucket and could
+	 * overflow KEY_SIZE
+	 */
+	if (KEY_SIZE(l) + KEY_SIZE(r) > USHRT_MAX) {
+		SET_KEY_OFFSET(l, KEY_OFFSET(l) + USHRT_MAX - KEY_SIZE(l));
+		SET_KEY_SIZE(l, USHRT_MAX);
+
+		bch_cut_front(l, r);
+		return false;
+	}
+
+	if (KEY_CSUM(l)) {
+		if (KEY_CSUM(r))
+			l->ptr[KEY_PTRS(l)] = merge_chksums(l, r);
+		else
+			SET_KEY_CSUM(l, 0);
+	}
+
+	SET_KEY_OFFSET(l, KEY_OFFSET(l) + KEY_SIZE(r));
+	SET_KEY_SIZE(l, KEY_SIZE(l) + KEY_SIZE(r));
+
+	return true;
+}
+
+/* Binary tree stuff for auxiliary search trees */
+
+static unsigned inorder_next(unsigned j, unsigned size)
+{
+	if (j * 2 + 1 < size) {
+		j = j * 2 + 1;
+
+		while (j * 2 < size)
+			j *= 2;
+	} else
+		j >>= ffz(j) + 1;
+
+	return j;
+}
+
+static unsigned inorder_prev(unsigned j, unsigned size)
+{
+	if (j * 2 < size) {
+		j = j * 2;
+
+		while (j * 2 + 1 < size)
+			j = j * 2 + 1;
+	} else
+		j >>= ffs(j);
+
+	return j;
+}
+
+/* I have no idea why this code works... and I'm the one who wrote it
+ *
+ * However, I do know what it does:
+ * Given a binary tree constructed in an array (i.e. how you normally implement
+ * a heap), it converts a node in the tree - referenced by array index - to the
+ * index it would have if you did an inorder traversal.
+ *
+ * Also tested for every j, size up to size somewhere around 6 million.
+ *
+ * The binary tree starts at array index 1, not 0
+ * extra is a function of size:
+ *   extra = (size - rounddown_pow_of_two(size - 1)) << 1;
+ */
+static unsigned __to_inorder(unsigned j, unsigned size, unsigned extra)
+{
+	unsigned b = fls(j);
+	unsigned shift = fls(size - 1) - b;
+
+	j  ^= 1U << (b - 1);
+	j <<= 1;
+	j  |= 1;
+	j <<= shift;
+
+	if (j > extra)
+		j -= (j - extra) >> 1;
+
+	return j;
+}
+
+static unsigned to_inorder(unsigned j, struct bset_tree *t)
+{
+	return __to_inorder(j, t->size, t->extra);
+}
+
+static unsigned __inorder_to_tree(unsigned j, unsigned size, unsigned extra)
+{
+	unsigned shift;
+
+	if (j > extra)
+		j += j - extra;
+
+	shift = ffs(j);
+
+	j >>= shift;
+	j  |= roundup_pow_of_two(size) >> shift;
+
+	return j;
+}
+
+static unsigned inorder_to_tree(unsigned j, struct bset_tree *t)
+{
+	return __inorder_to_tree(j, t->size, t->extra);
+}
+
+#if 0
+void inorder_test(void)
+{
+	unsigned long done = 0;
+	ktime_t start = ktime_get();
+
+	for (unsigned size = 2;
+	     size < 65536000;
+	     size++) {
+		unsigned extra = (size - rounddown_pow_of_two(size - 1)) << 1;
+		unsigned i = 1, j = rounddown_pow_of_two(size - 1);
+
+		if (!(size % 4096))
+			printk(KERN_NOTICE "loop %u, %llu per us\n", size,
+			       done / ktime_us_delta(ktime_get(), start));
+
+		while (1) {
+			if (__inorder_to_tree(i, size, extra) != j)
+				panic("size %10u j %10u i %10u", size, j, i);
+
+			if (__to_inorder(j, size, extra) != i)
+				panic("size %10u j %10u i %10u", size, j, i);
+
+			if (j == rounddown_pow_of_two(size) - 1)
+				break;
+
+			BUG_ON(inorder_prev(inorder_next(j, size), size) != j);
+
+			j = inorder_next(j, size);
+			i++;
+		}
+
+		done += size - 1;
+	}
+}
+#endif
+
+/*
+ * Cacheline/offset <-> bkey pointer arithmatic:
+ *
+ * t->tree is a binary search tree in an array; each node corresponds to a key
+ * in one cacheline in t->set (BSET_CACHELINE bytes).
+ *
+ * This means we don't have to store the full index of the key that a node in
+ * the binary tree points to; to_inorder() gives us the cacheline, and then
+ * bkey_float->m gives us the offset within that cacheline, in units of 8 bytes.
+ *
+ * cacheline_to_bkey() and friends abstract out all the pointer arithmatic to
+ * make this work.
+ *
+ * To construct the bfloat for an arbitrary key we need to know what the key
+ * immediately preceding it is: we have to check if the two keys differ in the
+ * bits we're going to store in bkey_float->mantissa. t->prev[j] stores the size
+ * of the previous key so we can walk backwards to it from t->tree[j]'s key.
+ */
+
+static struct bkey *cacheline_to_bkey(struct bset_tree *t, unsigned cacheline,
+				      unsigned offset)
+{
+	return ((void *) t->data) + cacheline * BSET_CACHELINE + offset * 8;
+}
+
+static unsigned bkey_to_cacheline(struct bset_tree *t, struct bkey *k)
+{
+	return ((void *) k - (void *) t->data) / BSET_CACHELINE;
+}
+
+static unsigned bkey_to_cacheline_offset(struct bkey *k)
+{
+	return ((size_t) k & (BSET_CACHELINE - 1)) / sizeof(uint64_t);
+}
+
+static struct bkey *tree_to_bkey(struct bset_tree *t, unsigned j)
+{
+	return cacheline_to_bkey(t, to_inorder(j, t), t->tree[j].m);
+}
+
+static struct bkey *tree_to_prev_bkey(struct bset_tree *t, unsigned j)
+{
+	return (void *) (((uint64_t *) tree_to_bkey(t, j)) - t->prev[j]);
+}
+
+/*
+ * For the write set - the one we're currently inserting keys into - we don't
+ * maintain a full search tree, we just keep a simple lookup table in t->prev.
+ */
+static struct bkey *table_to_bkey(struct bset_tree *t, unsigned cacheline)
+{
+	return cacheline_to_bkey(t, cacheline, t->prev[cacheline]);
+}
+
+static inline uint64_t shrd128(uint64_t high, uint64_t low, uint8_t shift)
+{
+#ifdef CONFIG_X86_64
+	asm("shrd %[shift],%[high],%[low]"
+	    : [low] "+Rm" (low)
+	    : [high] "R" (high),
+	    [shift] "ci" (shift)
+	    : "cc");
+#else
+	low >>= shift;
+	low  |= (high << 1) << (63U - shift);
+#endif
+	return low;
+}
+
+static inline unsigned bfloat_mantissa(const struct bkey *k,
+				       struct bkey_float *f)
+{
+	const uint64_t *p = &k->low - (f->exponent >> 6);
+	return shrd128(p[-1], p[0], f->exponent & 63) & BKEY_MANTISSA_MASK;
+}
+
+static void make_bfloat(struct bset_tree *t, unsigned j)
+{
+	struct bkey_float *f = &t->tree[j];
+	struct bkey *m = tree_to_bkey(t, j);
+	struct bkey *p = tree_to_prev_bkey(t, j);
+
+	struct bkey *l = is_power_of_2(j)
+		? t->data->start
+		: tree_to_prev_bkey(t, j >> ffs(j));
+
+	struct bkey *r = is_power_of_2(j + 1)
+		? node(t->data, t->data->keys - bkey_u64s(&t->end))
+		: tree_to_bkey(t, j >> (ffz(j) + 1));
+
+	BUG_ON(m < l || m > r);
+	BUG_ON(bkey_next(p) != m);
+
+	if (KEY_INODE(l) != KEY_INODE(r))
+		f->exponent = fls64(KEY_INODE(r) ^ KEY_INODE(l)) + 64;
+	else
+		f->exponent = fls64(r->low ^ l->low);
+
+	f->exponent = max_t(int, f->exponent - BKEY_MANTISSA_BITS, 0);
+
+	/*
+	 * Setting f->exponent = 127 flags this node as failed, and causes the
+	 * lookup code to fall back to comparing against the original key.
+	 */
+
+	if (bfloat_mantissa(m, f) != bfloat_mantissa(p, f))
+		f->mantissa = bfloat_mantissa(m, f) - 1;
+	else
+		f->exponent = 127;
+}
+
+static void bset_alloc_tree(struct btree *b, struct bset_tree *t)
+{
+	if (t != b->sets) {
+		unsigned j = roundup(t[-1].size,
+				     64 / sizeof(struct bkey_float));
+
+		t->tree = t[-1].tree + j;
+		t->prev = t[-1].prev + j;
+	}
+
+	while (t < b->sets + MAX_BSETS)
+		t++->size = 0;
+}
+
+static void bset_build_unwritten_tree(struct btree *b)
+{
+	struct bset_tree *t = b->sets + b->nsets;
+
+	bset_alloc_tree(b, t);
+
+	if (t->tree != b->sets->tree + bset_tree_space(b)) {
+		t->prev[0] = bkey_to_cacheline_offset(t->data->start);
+		t->size = 1;
+	}
+}
+
+static void bset_build_written_tree(struct btree *b)
+{
+	struct bset_tree *t = b->sets + b->nsets;
+	struct bkey *k = t->data->start;
+	unsigned j, cacheline = 1;
+
+	bset_alloc_tree(b, t);
+
+	t->size = min_t(unsigned,
+			bkey_to_cacheline(t, end(t->data)),
+			b->sets->tree + bset_tree_space(b) - t->tree);
+
+	if (t->size < 2) {
+		t->size = 0;
+		return;
+	}
+
+	t->extra = (t->size - rounddown_pow_of_two(t->size - 1)) << 1;
+
+	/* First we figure out where the first key in each cacheline is */
+	for (j = inorder_next(0, t->size);
+	     j;
+	     j = inorder_next(j, t->size)) {
+		while (bkey_to_cacheline(t, k) != cacheline)
+			k = bkey_next(k);
+
+		t->prev[j] = bkey_u64s(k);
+		k = bkey_next(k);
+		cacheline++;
+		t->tree[j].m = bkey_to_cacheline_offset(k);
+	}
+
+	while (bkey_next(k) != end(t->data))
+		k = bkey_next(k);
+
+	t->end = *k;
+
+	/* Then we build the tree */
+	for (j = inorder_next(0, t->size);
+	     j;
+	     j = inorder_next(j, t->size))
+		make_bfloat(t, j);
+}
+
+void bch_bset_fix_invalidated_key(struct btree *b, struct bkey *k)
+{
+	struct bset_tree *t;
+	unsigned inorder, j = 1;
+
+	for (t = b->sets; t <= &b->sets[b->nsets]; t++)
+		if (k < end(t->data))
+			goto found_set;
+
+	BUG();
+found_set:
+	if (!t->size || !bset_written(b, t))
+		return;
+
+	inorder = bkey_to_cacheline(t, k);
+
+	if (k == t->data->start)
+		goto fix_left;
+
+	if (bkey_next(k) == end(t->data)) {
+		t->end = *k;
+		goto fix_right;
+	}
+
+	j = inorder_to_tree(inorder, t);
+
+	if (j &&
+	    j < t->size &&
+	    k == tree_to_bkey(t, j))
+fix_left:	do {
+			make_bfloat(t, j);
+			j = j * 2;
+		} while (j < t->size);
+
+	j = inorder_to_tree(inorder + 1, t);
+
+	if (j &&
+	    j < t->size &&
+	    k == tree_to_prev_bkey(t, j))
+fix_right:	do {
+			make_bfloat(t, j);
+			j = j * 2 + 1;
+		} while (j < t->size);
+}
+
+void bch_bset_fix_lookup_table(struct btree *b, struct bkey *k)
+{
+	struct bset_tree *t = &b->sets[b->nsets];
+	unsigned shift = bkey_u64s(k);
+	unsigned j = bkey_to_cacheline(t, k);
+
+	/* We're getting called from btree_split() or btree_gc, just bail out */
+	if (!t->size)
+		return;
+
+	/* k is the key we just inserted; we need to find the entry in the
+	 * lookup table for the first key that is strictly greater than k:
+	 * it's either k's cacheline or the next one
+	 */
+	if (j < t->size &&
+	    table_to_bkey(t, j) <= k)
+		j++;
+
+	/* Adjust all the lookup table entries, and find a new key for any that
+	 * have gotten too big
+	 */
+	for (; j < t->size; j++) {
+		t->prev[j] += shift;
+
+		if (t->prev[j] > 7) {
+			k = table_to_bkey(t, j - 1);
+
+			while (k < cacheline_to_bkey(t, j, 0))
+				k = bkey_next(k);
+
+			t->prev[j] = bkey_to_cacheline_offset(k);
+		}
+	}
+
+	if (t->size == b->sets->tree + bset_tree_space(b) - t->tree)
+		return;
+
+	/* Possibly add a new entry to the end of the lookup table */
+
+	for (k = table_to_bkey(t, t->size - 1);
+	     k != end(t->data);
+	     k = bkey_next(k))
+		if (t->size == bkey_to_cacheline(t, k)) {
+			t->prev[t->size] = bkey_to_cacheline_offset(k);
+			t->size++;
+		}
+}
+
+void bch_bset_init_next(struct btree *b)
+{
+	struct bset *i = write_block(b);
+
+	if (i != b->sets[0].data) {
+		b->sets[++b->nsets].data = i;
+		i->seq = b->sets[0].data->seq;
+	} else
+		get_random_bytes(&i->seq, sizeof(uint64_t));
+
+	i->magic	= bset_magic(b->c);
+	i->version	= 0;
+	i->keys		= 0;
+
+	bset_build_unwritten_tree(b);
+}
+
+struct bset_search_iter {
+	struct bkey *l, *r;
+};
+
+static struct bset_search_iter bset_search_write_set(struct btree *b,
+						     struct bset_tree *t,
+						     const struct bkey *search)
+{
+	unsigned li = 0, ri = t->size;
+
+	BUG_ON(!b->nsets &&
+	       t->size < bkey_to_cacheline(t, end(t->data)));
+
+	while (li + 1 != ri) {
+		unsigned m = (li + ri) >> 1;
+
+		if (bkey_cmp(table_to_bkey(t, m), search) > 0)
+			ri = m;
+		else
+			li = m;
+	}
+
+	return (struct bset_search_iter) {
+		table_to_bkey(t, li),
+		ri < t->size ? table_to_bkey(t, ri) : end(t->data)
+	};
+}
+
+static struct bset_search_iter bset_search_tree(struct btree *b,
+						struct bset_tree *t,
+						const struct bkey *search)
+{
+	struct bkey *l, *r;
+	struct bkey_float *f;
+	unsigned inorder, j, n = 1;
+
+	do {
+		unsigned p = n << 4;
+		p &= ((int) (p - t->size)) >> 31;
+
+		prefetch(&t->tree[p]);
+
+		j = n;
+		f = &t->tree[j];
+
+		/*
+		 * n = (f->mantissa > bfloat_mantissa())
+		 *	? j * 2
+		 *	: j * 2 + 1;
+		 *
+		 * We need to subtract 1 from f->mantissa for the sign bit trick
+		 * to work  - that's done in make_bfloat()
+		 */
+		if (likely(f->exponent != 127))
+			n = j * 2 + (((unsigned)
+				      (f->mantissa -
+				       bfloat_mantissa(search, f))) >> 31);
+		else
+			n = (bkey_cmp(tree_to_bkey(t, j), search) > 0)
+				? j * 2
+				: j * 2 + 1;
+	} while (n < t->size);
+
+	inorder = to_inorder(j, t);
+
+	/*
+	 * n would have been the node we recursed to - the low bit tells us if
+	 * we recursed left or recursed right.
+	 */
+	if (n & 1) {
+		l = cacheline_to_bkey(t, inorder, f->m);
+
+		if (++inorder != t->size) {
+			f = &t->tree[inorder_next(j, t->size)];
+			r = cacheline_to_bkey(t, inorder, f->m);
+		} else
+			r = end(t->data);
+	} else {
+		r = cacheline_to_bkey(t, inorder, f->m);
+
+		if (--inorder) {
+			f = &t->tree[inorder_prev(j, t->size)];
+			l = cacheline_to_bkey(t, inorder, f->m);
+		} else
+			l = t->data->start;
+	}
+
+	return (struct bset_search_iter) {l, r};
+}
+
+struct bkey *__bch_bset_search(struct btree *b, struct bset_tree *t,
+			       const struct bkey *search)
+{
+	struct bset_search_iter i;
+
+	/*
+	 * First, we search for a cacheline, then lastly we do a linear search
+	 * within that cacheline.
+	 *
+	 * To search for the cacheline, there's three different possibilities:
+	 *  * The set is too small to have a search tree, so we just do a linear
+	 *    search over the whole set.
+	 *  * The set is the one we're currently inserting into; keeping a full
+	 *    auxiliary search tree up to date would be too expensive, so we
+	 *    use a much simpler lookup table to do a binary search -
+	 *    bset_search_write_set().
+	 *  * Or we use the auxiliary search tree we constructed earlier -
+	 *    bset_search_tree()
+	 */
+
+	if (unlikely(!t->size)) {
+		i.l = t->data->start;
+		i.r = end(t->data);
+	} else if (bset_written(b, t)) {
+		/*
+		 * Each node in the auxiliary search tree covers a certain range
+		 * of bits, and keys above and below the set it covers might
+		 * differ outside those bits - so we have to special case the
+		 * start and end - handle that here:
+		 */
+
+		if (unlikely(bkey_cmp(search, &t->end) >= 0))
+			return end(t->data);
+
+		if (unlikely(bkey_cmp(search, t->data->start) < 0))
+			return t->data->start;
+
+		i = bset_search_tree(b, t, search);
+	} else
+		i = bset_search_write_set(b, t, search);
+
+#ifdef CONFIG_BCACHE_EDEBUG
+	BUG_ON(bset_written(b, t) &&
+	       i.l != t->data->start &&
+	       bkey_cmp(tree_to_prev_bkey(t,
+		  inorder_to_tree(bkey_to_cacheline(t, i.l), t)),
+			search) > 0);
+
+	BUG_ON(i.r != end(t->data) &&
+	       bkey_cmp(i.r, search) <= 0);
+#endif
+
+	while (likely(i.l != i.r) &&
+	       bkey_cmp(i.l, search) <= 0)
+		i.l = bkey_next(i.l);
+
+	return i.l;
+}
+
+/* Btree iterator */
+
+static inline bool btree_iter_cmp(struct btree_iter_set l,
+				  struct btree_iter_set r)
+{
+	int64_t c = bkey_cmp(&START_KEY(l.k), &START_KEY(r.k));
+
+	return c ? c > 0 : l.k < r.k;
+}
+
+static inline bool btree_iter_end(struct btree_iter *iter)
+{
+	return !iter->used;
+}
+
+void bch_btree_iter_push(struct btree_iter *iter, struct bkey *k,
+			 struct bkey *end)
+{
+	if (k != end)
+		BUG_ON(!heap_add(iter,
+				 ((struct btree_iter_set) { k, end }),
+				 btree_iter_cmp));
+}
+
+struct bkey *__bch_btree_iter_init(struct btree *b, struct btree_iter *iter,
+			       struct bkey *search, struct bset_tree *start)
+{
+	struct bkey *ret = NULL;
+	iter->size = ARRAY_SIZE(iter->data);
+	iter->used = 0;
+
+	for (; start <= &b->sets[b->nsets]; start++) {
+		ret = bch_bset_search(b, start, search);
+		bch_btree_iter_push(iter, ret, end(start->data));
+	}
+
+	return ret;
+}
+
+struct bkey *bch_btree_iter_next(struct btree_iter *iter)
+{
+	struct btree_iter_set unused;
+	struct bkey *ret = NULL;
+
+	if (!btree_iter_end(iter)) {
+		ret = iter->data->k;
+		iter->data->k = bkey_next(iter->data->k);
+
+		if (iter->data->k > iter->data->end) {
+			WARN_ONCE(1, "bset was corrupt!\n");
+			iter->data->k = iter->data->end;
+		}
+
+		if (iter->data->k == iter->data->end)
+			heap_pop(iter, unused, btree_iter_cmp);
+		else
+			heap_sift(iter, 0, btree_iter_cmp);
+	}
+
+	return ret;
+}
+
+struct bkey *bch_btree_iter_next_filter(struct btree_iter *iter,
+					struct btree *b, ptr_filter_fn fn)
+{
+	struct bkey *ret;
+
+	do {
+		ret = bch_btree_iter_next(iter);
+	} while (ret && fn(b, ret));
+
+	return ret;
+}
+
+struct bkey *bch_next_recurse_key(struct btree *b, struct bkey *search)
+{
+	struct btree_iter iter;
+
+	bch_btree_iter_init(b, &iter, search);
+	return bch_btree_iter_next_filter(&iter, b, bch_ptr_bad);
+}
+
+/* Mergesort */
+
+static void btree_sort_fixup(struct btree_iter *iter)
+{
+	while (iter->used > 1) {
+		struct btree_iter_set *top = iter->data, *i = top + 1;
+		struct bkey *k;
+
+		if (iter->used > 2 &&
+		    btree_iter_cmp(i[0], i[1]))
+			i++;
+
+		for (k = i->k;
+		     k != i->end && bkey_cmp(top->k, &START_KEY(k)) > 0;
+		     k = bkey_next(k))
+			if (top->k > i->k)
+				__bch_cut_front(top->k, k);
+			else if (KEY_SIZE(k))
+				bch_cut_back(&START_KEY(k), top->k);
+
+		if (top->k < i->k || k == i->k)
+			break;
+
+		heap_sift(iter, i - top, btree_iter_cmp);
+	}
+}
+
+static void btree_mergesort(struct btree *b, struct bset *out,
+			    struct btree_iter *iter,
+			    bool fixup, bool remove_stale)
+{
+	struct bkey *k, *last = NULL;
+	bool (*bad)(struct btree *, const struct bkey *) = remove_stale
+		? bch_ptr_bad
+		: bch_ptr_invalid;
+
+	while (!btree_iter_end(iter)) {
+		if (fixup && !b->level)
+			btree_sort_fixup(iter);
+
+		k = bch_btree_iter_next(iter);
+		if (bad(b, k))
+			continue;
+
+		if (!last) {
+			last = out->start;
+			bkey_copy(last, k);
+		} else if (b->level ||
+			   !bch_bkey_try_merge(b, last, k)) {
+			last = bkey_next(last);
+			bkey_copy(last, k);
+		}
+	}
+
+	out->keys = last ? (uint64_t *) bkey_next(last) - out->d : 0;
+
+	pr_debug("sorted %i keys", out->keys);
+	bch_check_key_order(b, out);
+}
+
+static void __btree_sort(struct btree *b, struct btree_iter *iter,
+			 unsigned start, unsigned order, bool fixup)
+{
+	uint64_t start_time;
+	bool remove_stale = !b->written;
+	struct bset *out = (void *) __get_free_pages(__GFP_NOWARN|GFP_NOIO,
+						     order);
+	if (!out) {
+		mutex_lock(&b->c->sort_lock);
+		out = b->c->sort;
+		order = ilog2(bucket_pages(b->c));
+	}
+
+	start_time = local_clock();
+
+	btree_mergesort(b, out, iter, fixup, remove_stale);
+	b->nsets = start;
+
+	if (!fixup && !start && b->written)
+		bch_btree_verify(b, out);
+
+	if (!start && order == b->page_order) {
+		/*
+		 * Our temporary buffer is the same size as the btree node's
+		 * buffer, we can just swap buffers instead of doing a big
+		 * memcpy()
+		 */
+
+		out->magic	= bset_magic(b->c);
+		out->seq	= b->sets[0].data->seq;
+		out->version	= b->sets[0].data->version;
+		swap(out, b->sets[0].data);
+
+		if (b->c->sort == b->sets[0].data)
+			b->c->sort = out;
+	} else {
+		b->sets[start].data->keys = out->keys;
+		memcpy(b->sets[start].data->start, out->start,
+		       (void *) end(out) - (void *) out->start);
+	}
+
+	if (out == b->c->sort)
+		mutex_unlock(&b->c->sort_lock);
+	else
+		free_pages((unsigned long) out, order);
+
+	if (b->written)
+		bset_build_written_tree(b);
+
+	if (!start) {
+		spin_lock(&b->c->sort_time_lock);
+		bch_time_stats_update(&b->c->sort_time, start_time);
+		spin_unlock(&b->c->sort_time_lock);
+	}
+}
+
+void bch_btree_sort_partial(struct btree *b, unsigned start)
+{
+	size_t oldsize = 0, order = b->page_order, keys = 0;
+	struct btree_iter iter;
+	__bch_btree_iter_init(b, &iter, NULL, &b->sets[start]);
+
+	BUG_ON(b->sets[b->nsets].data == write_block(b) &&
+	       (b->sets[b->nsets].size || b->nsets));
+
+	if (b->written)
+		oldsize = bch_count_data(b);
+
+	if (start) {
+		unsigned i;
+
+		for (i = start; i <= b->nsets; i++)
+			keys += b->sets[i].data->keys;
+
+		order = roundup_pow_of_two(__set_bytes(b->sets->data,
+						       keys)) / PAGE_SIZE;
+		if (order)
+			order = ilog2(order);
+	}
+
+	__btree_sort(b, &iter, start, order, false);
+
+	EBUG_ON(b->written && bch_count_data(b) != oldsize);
+}
+
+void bch_btree_sort_and_fix_extents(struct btree *b, struct btree_iter *iter)
+{
+	BUG_ON(!b->written);
+	__btree_sort(b, iter, 0, b->page_order, true);
+}
+
+void bch_btree_sort_into(struct btree *b, struct btree *new)
+{
+	uint64_t start_time = local_clock();
+
+	struct btree_iter iter;
+	bch_btree_iter_init(b, &iter, NULL);
+
+	btree_mergesort(b, new->sets->data, &iter, false, true);
+
+	spin_lock(&b->c->sort_time_lock);
+	bch_time_stats_update(&b->c->sort_time, start_time);
+	spin_unlock(&b->c->sort_time_lock);
+
+	bkey_copy_key(&new->key, &b->key);
+	new->sets->size = 0;
+}
+
+void bch_btree_sort_lazy(struct btree *b)
+{
+	if (b->nsets) {
+		unsigned i, j, keys = 0, total;
+
+		for (i = 0; i <= b->nsets; i++)
+			keys += b->sets[i].data->keys;
+
+		total = keys;
+
+		for (j = 0; j < b->nsets; j++) {
+			if (keys * 2 < total ||
+			    keys < 1000) {
+				bch_btree_sort_partial(b, j);
+				return;
+			}
+
+			keys -= b->sets[j].data->keys;
+		}
+
+		/* Must sort if b->nsets == 3 or we'll overflow */
+		if (b->nsets >= (MAX_BSETS - 1) - b->level) {
+			bch_btree_sort(b);
+			return;
+		}
+	}
+
+	bset_build_written_tree(b);
+}
+
+/* Sysfs stuff */
+
+struct bset_stats {
+	size_t nodes;
+	size_t sets_written, sets_unwritten;
+	size_t bytes_written, bytes_unwritten;
+	size_t floats, failed;
+};
+
+static int bch_btree_bset_stats(struct btree *b, struct btree_op *op,
+			    struct bset_stats *stats)
+{
+	struct bkey *k;
+	unsigned i;
+
+	stats->nodes++;
+
+	for (i = 0; i <= b->nsets; i++) {
+		struct bset_tree *t = &b->sets[i];
+		size_t bytes = t->data->keys * sizeof(uint64_t);
+		size_t j;
+
+		if (bset_written(b, t)) {
+			stats->sets_written++;
+			stats->bytes_written += bytes;
+
+			stats->floats += t->size - 1;
+
+			for (j = 1; j < t->size; j++)
+				if (t->tree[j].exponent == 127)
+					stats->failed++;
+		} else {
+			stats->sets_unwritten++;
+			stats->bytes_unwritten += bytes;
+		}
+	}
+
+	if (b->level) {
+		struct btree_iter iter;
+
+		for_each_key_filter(b, k, &iter, bch_ptr_bad) {
+			int ret = btree(bset_stats, k, b, op, stats);
+			if (ret)
+				return ret;
+		}
+	}
+
+	return 0;
+}
+
+int bch_bset_print_stats(struct cache_set *c, char *buf)
+{
+	struct btree_op op;
+	struct bset_stats t;
+	int ret;
+
+	bch_btree_op_init_stack(&op);
+	memset(&t, 0, sizeof(struct bset_stats));
+
+	ret = btree_root(bset_stats, c, &op, &t);
+	if (ret)
+		return ret;
+
+	return snprintf(buf, PAGE_SIZE,
+			"btree nodes:		%zu\n"
+			"written sets:		%zu\n"
+			"unwritten sets:		%zu\n"
+			"written key bytes:	%zu\n"
+			"unwritten key bytes:	%zu\n"
+			"floats:			%zu\n"
+			"failed:			%zu\n",
+			t.nodes,
+			t.sets_written, t.sets_unwritten,
+			t.bytes_written, t.bytes_unwritten,
+			t.floats, t.failed);
+}

drivers/md/bcache/bset.h

···
+#ifndef _BCACHE_BSET_H
+#define _BCACHE_BSET_H
+
+/*
+ * BKEYS:
+ *
+ * A bkey contains a key, a size field, a variable number of pointers, and some
+ * ancillary flag bits.
+ *
+ * We use two different functions for validating bkeys, bch_ptr_invalid and
+ * bch_ptr_bad().
+ *
+ * bch_ptr_invalid() primarily filters out keys and pointers that would be
+ * invalid due to some sort of bug, whereas bch_ptr_bad() filters out keys and
+ * pointer that occur in normal practice but don't point to real data.
+ *
+ * The one exception to the rule that ptr_invalid() filters out invalid keys is
+ * that it also filters out keys of size 0 - these are keys that have been
+ * completely overwritten. It'd be safe to delete these in memory while leaving
+ * them on disk, just unnecessary work - so we filter them out when resorting
+ * instead.
+ *
+ * We can't filter out stale keys when we're resorting, because garbage
+ * collection needs to find them to ensure bucket gens don't wrap around -
+ * unless we're rewriting the btree node those stale keys still exist on disk.
+ *
+ * We also implement functions here for removing some number of sectors from the
+ * front or the back of a bkey - this is mainly used for fixing overlapping
+ * extents, by removing the overlapping sectors from the older key.
+ *
+ * BSETS:
+ *
+ * A bset is an array of bkeys laid out contiguously in memory in sorted order,
+ * along with a header. A btree node is made up of a number of these, written at
+ * different times.
+ *
+ * There could be many of them on disk, but we never allow there to be more than
+ * 4 in memory - we lazily resort as needed.
+ *
+ * We implement code here for creating and maintaining auxiliary search trees
+ * (described below) for searching an individial bset, and on top of that we
+ * implement a btree iterator.
+ *
+ * BTREE ITERATOR:
+ *
+ * Most of the code in bcache doesn't care about an individual bset - it needs
+ * to search entire btree nodes and iterate over them in sorted order.
+ *
+ * The btree iterator code serves both functions; it iterates through the keys
+ * in a btree node in sorted order, starting from either keys after a specific
+ * point (if you pass it a search key) or the start of the btree node.
+ *
+ * AUXILIARY SEARCH TREES:
+ *
+ * Since keys are variable length, we can't use a binary search on a bset - we
+ * wouldn't be able to find the start of the next key. But binary searches are
+ * slow anyways, due to terrible cache behaviour; bcache originally used binary
+ * searches and that code topped out at under 50k lookups/second.
+ *
+ * So we need to construct some sort of lookup table. Since we only insert keys
+ * into the last (unwritten) set, most of the keys within a given btree node are
+ * usually in sets that are mostly constant. We use two different types of
+ * lookup tables to take advantage of this.
+ *
+ * Both lookup tables share in common that they don't index every key in the
+ * set; they index one key every BSET_CACHELINE bytes, and then a linear search
+ * is used for the rest.
+ *
+ * For sets that have been written to disk and are no longer being inserted
+ * into, we construct a binary search tree in an array - traversing a binary
+ * search tree in an array gives excellent locality of reference and is very
+ * fast, since both children of any node are adjacent to each other in memory
+ * (and their grandchildren, and great grandchildren...) - this means
+ * prefetching can be used to great effect.
+ *
+ * It's quite useful performance wise to keep these nodes small - not just
+ * because they're more likely to be in L2, but also because we can prefetch
+ * more nodes on a single cacheline and thus prefetch more iterations in advance
+ * when traversing this tree.
+ *
+ * Nodes in the auxiliary search tree must contain both a key to compare against
+ * (we don't want to fetch the key from the set, that would defeat the purpose),
+ * and a pointer to the key. We use a few tricks to compress both of these.
+ *
+ * To compress the pointer, we take advantage of the fact that one node in the
+ * search tree corresponds to precisely BSET_CACHELINE bytes in the set. We have
+ * a function (to_inorder()) that takes the index of a node in a binary tree and
+ * returns what its index would be in an inorder traversal, so we only have to
+ * store the low bits of the offset.
+ *
+ * The key is 84 bits (KEY_DEV + key->key, the offset on the device). To
+ * compress that,  we take advantage of the fact that when we're traversing the
+ * search tree at every iteration we know that both our search key and the key
+ * we're looking for lie within some range - bounded by our previous
+ * comparisons. (We special case the start of a search so that this is true even
+ * at the root of the tree).
+ *
+ * So we know the key we're looking for is between a and b, and a and b don't
+ * differ higher than bit 50, we don't need to check anything higher than bit
+ * 50.
+ *
+ * We don't usually need the rest of the bits, either; we only need enough bits
+ * to partition the key range we're currently checking.  Consider key n - the
+ * key our auxiliary search tree node corresponds to, and key p, the key
+ * immediately preceding n.  The lowest bit we need to store in the auxiliary
+ * search tree is the highest bit that differs between n and p.
+ *
+ * Note that this could be bit 0 - we might sometimes need all 80 bits to do the
+ * comparison. But we'd really like our nodes in the auxiliary search tree to be
+ * of fixed size.
+ *
+ * The solution is to make them fixed size, and when we're constructing a node
+ * check if p and n differed in the bits we needed them to. If they don't we
+ * flag that node, and when doing lookups we fallback to comparing against the
+ * real key. As long as this doesn't happen to often (and it seems to reliably
+ * happen a bit less than 1% of the time), we win - even on failures, that key
+ * is then more likely to be in cache than if we were doing binary searches all
+ * the way, since we're touching so much less memory.
+ *
+ * The keys in the auxiliary search tree are stored in (software) floating
+ * point, with an exponent and a mantissa. The exponent needs to be big enough
+ * to address all the bits in the original key, but the number of bits in the
+ * mantissa is somewhat arbitrary; more bits just gets us fewer failures.
+ *
+ * We need 7 bits for the exponent and 3 bits for the key's offset (since keys
+ * are 8 byte aligned); using 22 bits for the mantissa means a node is 4 bytes.
+ * We need one node per 128 bytes in the btree node, which means the auxiliary
+ * search trees take up 3% as much memory as the btree itself.
+ *
+ * Constructing these auxiliary search trees is moderately expensive, and we
+ * don't want to be constantly rebuilding the search tree for the last set
+ * whenever we insert another key into it. For the unwritten set, we use a much
+ * simpler lookup table - it's just a flat array, so index i in the lookup table
+ * corresponds to the i range of BSET_CACHELINE bytes in the set. Indexing
+ * within each byte range works the same as with the auxiliary search trees.
+ *
+ * These are much easier to keep up to date when we insert a key - we do it
+ * somewhat lazily; when we shift a key up we usually just increment the pointer
+ * to it, only when it would overflow do we go to the trouble of finding the
+ * first key in that range of bytes again.
+ */
+
+/* Btree key comparison/iteration */
+
+struct btree_iter {
+	size_t size, used;
+	struct btree_iter_set {
+		struct bkey *k, *end;
+	} data[MAX_BSETS];
+};
+
+struct bset_tree {
+	/*
+	 * We construct a binary tree in an array as if the array
+	 * started at 1, so that things line up on the same cachelines
+	 * better: see comments in bset.c at cacheline_to_bkey() for
+	 * details
+	 */
+
+	/* size of the binary tree and prev array */
+	unsigned	size;
+
+	/* function of size - precalculated for to_inorder() */
+	unsigned	extra;
+
+	/* copy of the last key in the set */
+	struct bkey	end;
+	struct bkey_float *tree;
+
+	/*
+	 * The nodes in the bset tree point to specific keys - this
+	 * array holds the sizes of the previous key.
+	 *
+	 * Conceptually it's a member of struct bkey_float, but we want
+	 * to keep bkey_float to 4 bytes and prev isn't used in the fast
+	 * path.
+	 */
+	uint8_t		*prev;
+
+	/* The actual btree node, with pointers to each sorted set */
+	struct bset	*data;
+};
+
+static __always_inline int64_t bkey_cmp(const struct bkey *l,
+					const struct bkey *r)
+{
+	return unlikely(KEY_INODE(l) != KEY_INODE(r))
+		? (int64_t) KEY_INODE(l) - (int64_t) KEY_INODE(r)
+		: (int64_t) KEY_OFFSET(l) - (int64_t) KEY_OFFSET(r);
+}
+
+static inline size_t bkey_u64s(const struct bkey *k)
+{
+	BUG_ON(KEY_CSUM(k) > 1);
+	return 2 + KEY_PTRS(k) + (KEY_CSUM(k) ? 1 : 0);
+}
+
+static inline size_t bkey_bytes(const struct bkey *k)
+{
+	return bkey_u64s(k) * sizeof(uint64_t);
+}
+
+static inline void bkey_copy(struct bkey *dest, const struct bkey *src)
+{
+	memcpy(dest, src, bkey_bytes(src));
+}
+
+static inline void bkey_copy_key(struct bkey *dest, const struct bkey *src)
+{
+	if (!src)
+		src = &KEY(0, 0, 0);
+
+	SET_KEY_INODE(dest, KEY_INODE(src));
+	SET_KEY_OFFSET(dest, KEY_OFFSET(src));
+}
+
+static inline struct bkey *bkey_next(const struct bkey *k)
+{
+	uint64_t *d = (void *) k;
+	return (struct bkey *) (d + bkey_u64s(k));
+}
+
+/* Keylists */
+
+struct keylist {
+	struct bkey		*top;
+	union {
+		uint64_t		*list;
+		struct bkey		*bottom;
+	};
+
+	/* Enough room for btree_split's keys without realloc */
+#define KEYLIST_INLINE		16
+	uint64_t		d[KEYLIST_INLINE];
+};
+
+static inline void bch_keylist_init(struct keylist *l)
+{
+	l->top = (void *) (l->list = l->d);
+}
+
+static inline void bch_keylist_push(struct keylist *l)
+{
+	l->top = bkey_next(l->top);
+}
+
+static inline void bch_keylist_add(struct keylist *l, struct bkey *k)
+{
+	bkey_copy(l->top, k);
+	bch_keylist_push(l);
+}
+
+static inline bool bch_keylist_empty(struct keylist *l)
+{
+	return l->top == (void *) l->list;
+}
+
+static inline void bch_keylist_free(struct keylist *l)
+{
+	if (l->list != l->d)
+		kfree(l->list);
+}
+
+void bch_keylist_copy(struct keylist *, struct keylist *);
+struct bkey *bch_keylist_pop(struct keylist *);
+int bch_keylist_realloc(struct keylist *, int, struct cache_set *);
+
+void bch_bkey_copy_single_ptr(struct bkey *, const struct bkey *,
+			      unsigned);
+bool __bch_cut_front(const struct bkey *, struct bkey *);
+bool __bch_cut_back(const struct bkey *, struct bkey *);
+
+static inline bool bch_cut_front(const struct bkey *where, struct bkey *k)
+{
+	BUG_ON(bkey_cmp(where, k) > 0);
+	return __bch_cut_front(where, k);
+}
+
+static inline bool bch_cut_back(const struct bkey *where, struct bkey *k)
+{
+	BUG_ON(bkey_cmp(where, &START_KEY(k)) < 0);
+	return __bch_cut_back(where, k);
+}
+
+const char *bch_ptr_status(struct cache_set *, const struct bkey *);
+bool __bch_ptr_invalid(struct cache_set *, int level, const struct bkey *);
+bool bch_ptr_bad(struct btree *, const struct bkey *);
+
+static inline uint8_t gen_after(uint8_t a, uint8_t b)
+{
+	uint8_t r = a - b;
+	return r > 128U ? 0 : r;
+}
+
+static inline uint8_t ptr_stale(struct cache_set *c, const struct bkey *k,
+				unsigned i)
+{
+	return gen_after(PTR_BUCKET(c, k, i)->gen, PTR_GEN(k, i));
+}
+
+static inline bool ptr_available(struct cache_set *c, const struct bkey *k,
+				 unsigned i)
+{
+	return (PTR_DEV(k, i) < MAX_CACHES_PER_SET) && PTR_CACHE(c, k, i);
+}
+
+
+typedef bool (*ptr_filter_fn)(struct btree *, const struct bkey *);
+
+struct bkey *bch_next_recurse_key(struct btree *, struct bkey *);
+struct bkey *bch_btree_iter_next(struct btree_iter *);
+struct bkey *bch_btree_iter_next_filter(struct btree_iter *,
+					struct btree *, ptr_filter_fn);
+
+void bch_btree_iter_push(struct btree_iter *, struct bkey *, struct bkey *);
+struct bkey *__bch_btree_iter_init(struct btree *, struct btree_iter *,
+				   struct bkey *, struct bset_tree *);
+
+/* 32 bits total: */
+#define BKEY_MID_BITS		3
+#define BKEY_EXPONENT_BITS	7
+#define BKEY_MANTISSA_BITS	22
+#define BKEY_MANTISSA_MASK	((1 << BKEY_MANTISSA_BITS) - 1)
+
+struct bkey_float {
+	unsigned	exponent:BKEY_EXPONENT_BITS;
+	unsigned	m:BKEY_MID_BITS;
+	unsigned	mantissa:BKEY_MANTISSA_BITS;
+} __packed;
+
+/*
+ * BSET_CACHELINE was originally intended to match the hardware cacheline size -
+ * it used to be 64, but I realized the lookup code would touch slightly less
+ * memory if it was 128.
+ *
+ * It definites the number of bytes (in struct bset) per struct bkey_float in
+ * the auxiliar search tree - when we're done searching the bset_float tree we
+ * have this many bytes left that we do a linear search over.
+ *
+ * Since (after level 5) every level of the bset_tree is on a new cacheline,
+ * we're touching one fewer cacheline in the bset tree in exchange for one more
+ * cacheline in the linear search - but the linear search might stop before it
+ * gets to the second cacheline.
+ */
+
+#define BSET_CACHELINE		128
+#define bset_tree_space(b)	(btree_data_space(b) / BSET_CACHELINE)
+
+#define bset_tree_bytes(b)	(bset_tree_space(b) * sizeof(struct bkey_float))
+#define bset_prev_bytes(b)	(bset_tree_space(b) * sizeof(uint8_t))
+
+void bch_bset_init_next(struct btree *);
+
+void bch_bset_fix_invalidated_key(struct btree *, struct bkey *);
+void bch_bset_fix_lookup_table(struct btree *, struct bkey *);
+
+struct bkey *__bch_bset_search(struct btree *, struct bset_tree *,
+			   const struct bkey *);
+
+static inline struct bkey *bch_bset_search(struct btree *b, struct bset_tree *t,
+					   const struct bkey *search)
+{
+	return search ? __bch_bset_search(b, t, search) : t->data->start;
+}
+
+bool bch_bkey_try_merge(struct btree *, struct bkey *, struct bkey *);
+void bch_btree_sort_lazy(struct btree *);
+void bch_btree_sort_into(struct btree *, struct btree *);
+void bch_btree_sort_and_fix_extents(struct btree *, struct btree_iter *);
+void bch_btree_sort_partial(struct btree *, unsigned);
+
+static inline void bch_btree_sort(struct btree *b)
+{
+	bch_btree_sort_partial(b, 0);
+}
+
+int bch_bset_print_stats(struct cache_set *, char *);
+
+#endif

drivers/md/bcache/btree.c

···
+/*
+ * Copyright (C) 2010 Kent Overstreet <[e-mail]>
+ *
+ * Uses a block device as cache for other block devices; optimized for SSDs.
+ * All allocation is done in buckets, which should match the erase block size
+ * of the device.
+ *
+ * Buckets containing cached data are kept on a heap sorted by priority;
+ * bucket priority is increased on cache hit, and periodically all the buckets
+ * on the heap have their priority scaled down. This currently is just used as
+ * an LRU but in the future should allow for more intelligent heuristics.
+ *
+ * Buckets have an 8 bit counter; freeing is accomplished by incrementing the
+ * counter. Garbage collection is used to remove stale pointers.
+ *
+ * Indexing is done via a btree; nodes are not necessarily fully sorted, rather
+ * as keys are inserted we only sort the pages that have not yet been written.
+ * When garbage collection is run, we resort the entire node.
+ *
+ * All configuration is done via sysfs; see Documentation/bcache.txt.
+ */
+
+#include "bcache.h"
+#include "btree.h"
+#include "debug.h"
+#include "request.h"
+
+#include <linux/slab.h>
+#include <linux/bitops.h>
+#include <linux/hash.h>
+#include <linux/prefetch.h>
+#include <linux/random.h>
+#include <linux/rcupdate.h>
+#include <trace/events/bcache.h>
+
+/*
+ * Todo:
+ * register_bcache: Return errors out to userspace correctly
+ *
+ * Writeback: don't undirty key until after a cache flush
+ *
+ * Create an iterator for key pointers
+ *
+ * On btree write error, mark bucket such that it won't be freed from the cache
+ *
+ * Journalling:
+ *   Check for bad keys in replay
+ *   Propagate barriers
+ *   Refcount journal entries in journal_replay
+ *
+ * Garbage collection:
+ *   Finish incremental gc
+ *   Gc should free old UUIDs, data for invalid UUIDs
+ *
+ * Provide a way to list backing device UUIDs we have data cached for, and
+ * probably how long it's been since we've seen them, and a way to invalidate
+ * dirty data for devices that will never be attached again
+ *
+ * Keep 1 min/5 min/15 min statistics of how busy a block device has been, so
+ * that based on that and how much dirty data we have we can keep writeback
+ * from being starved
+ *
+ * Add a tracepoint or somesuch to watch for writeback starvation
+ *
+ * When btree depth > 1 and splitting an interior node, we have to make sure
+ * alloc_bucket() cannot fail. This should be true but is not completely
+ * obvious.
+ *
+ * Make sure all allocations get charged to the root cgroup
+ *
+ * Plugging?
+ *
+ * If data write is less than hard sector size of ssd, round up offset in open
+ * bucket to the next whole sector
+ *
+ * Also lookup by cgroup in get_open_bucket()
+ *
+ * Superblock needs to be fleshed out for multiple cache devices
+ *
+ * Add a sysfs tunable for the number of writeback IOs in flight
+ *
+ * Add a sysfs tunable for the number of open data buckets
+ *
+ * IO tracking: Can we track when one process is doing io on behalf of another?
+ * IO tracking: Don't use just an average, weigh more recent stuff higher
+ *
+ * Test module load/unload
+ */
+
+static const char * const op_types[] = {
+	"insert", "replace"
+};
+
+static const char *op_type(struct btree_op *op)
+{
+	return op_types[op->type];
+}
+
+#define MAX_NEED_GC		64
+#define MAX_SAVE_PRIO		72
+
+#define PTR_DIRTY_BIT		(((uint64_t) 1 << 36))
+
+#define PTR_HASH(c, k)							\
+	(((k)->ptr[0] >> c->bucket_bits) | PTR_GEN(k, 0))
+
+struct workqueue_struct *bch_gc_wq;
+static struct workqueue_struct *btree_io_wq;
+
+void bch_btree_op_init_stack(struct btree_op *op)
+{
+	memset(op, 0, sizeof(struct btree_op));
+	closure_init_stack(&op->cl);
+	op->lock = -1;
+	bch_keylist_init(&op->keys);
+}
+
+/* Btree key manipulation */
+
+static void bkey_put(struct cache_set *c, struct bkey *k, int level)
+{
+	if ((level && KEY_OFFSET(k)) || !level)
+		__bkey_put(c, k);
+}
+
+/* Btree IO */
+
+static uint64_t btree_csum_set(struct btree *b, struct bset *i)
+{
+	uint64_t crc = b->key.ptr[0];
+	void *data = (void *) i + 8, *end = end(i);
+
+	crc = bch_crc64_update(crc, data, end - data);
+	return crc ^ 0xffffffffffffffffULL;
+}
+
+static void btree_bio_endio(struct bio *bio, int error)
+{
+	struct closure *cl = bio->bi_private;
+	struct btree *b = container_of(cl, struct btree, io.cl);
+
+	if (error)
+		set_btree_node_io_error(b);
+
+	bch_bbio_count_io_errors(b->c, bio, error, (bio->bi_rw & WRITE)
+				 ? "writing btree" : "reading btree");
+	closure_put(cl);
+}
+
+static void btree_bio_init(struct btree *b)
+{
+	BUG_ON(b->bio);
+	b->bio = bch_bbio_alloc(b->c);
+
+	b->bio->bi_end_io	= btree_bio_endio;
+	b->bio->bi_private	= &b->io.cl;
+}
+
+void bch_btree_read_done(struct closure *cl)
+{
+	struct btree *b = container_of(cl, struct btree, io.cl);
+	struct bset *i = b->sets[0].data;
+	struct btree_iter *iter = b->c->fill_iter;
+	const char *err = "bad btree header";
+	BUG_ON(b->nsets || b->written);
+
+	bch_bbio_free(b->bio, b->c);
+	b->bio = NULL;
+
+	mutex_lock(&b->c->fill_lock);
+	iter->used = 0;
+
+	if (btree_node_io_error(b) ||
+	    !i->seq)
+		goto err;
+
+	for (;
+	     b->written < btree_blocks(b) && i->seq == b->sets[0].data->seq;
+	     i = write_block(b)) {
+		err = "unsupported bset version";
+		if (i->version > BCACHE_BSET_VERSION)
+			goto err;
+
+		err = "bad btree header";
+		if (b->written + set_blocks(i, b->c) > btree_blocks(b))
+			goto err;
+
+		err = "bad magic";
+		if (i->magic != bset_magic(b->c))
+			goto err;
+
+		err = "bad checksum";
+		switch (i->version) {
+		case 0:
+			if (i->csum != csum_set(i))
+				goto err;
+			break;
+		case BCACHE_BSET_VERSION:
+			if (i->csum != btree_csum_set(b, i))
+				goto err;
+			break;
+		}
+
+		err = "empty set";
+		if (i != b->sets[0].data && !i->keys)
+			goto err;
+
+		bch_btree_iter_push(iter, i->start, end(i));
+
+		b->written += set_blocks(i, b->c);
+	}
+
+	err = "corrupted btree";
+	for (i = write_block(b);
+	     index(i, b) < btree_blocks(b);
+	     i = ((void *) i) + block_bytes(b->c))
+		if (i->seq == b->sets[0].data->seq)
+			goto err;
+
+	bch_btree_sort_and_fix_extents(b, iter);
+
+	i = b->sets[0].data;
+	err = "short btree key";
+	if (b->sets[0].size &&
+	    bkey_cmp(&b->key, &b->sets[0].end) < 0)
+		goto err;
+
+	if (b->written < btree_blocks(b))
+		bch_bset_init_next(b);
+out:
+
+	mutex_unlock(&b->c->fill_lock);
+
+	spin_lock(&b->c->btree_read_time_lock);
+	bch_time_stats_update(&b->c->btree_read_time, b->io_start_time);
+	spin_unlock(&b->c->btree_read_time_lock);
+
+	smp_wmb(); /* read_done is our write lock */
+	set_btree_node_read_done(b);
+
+	closure_return(cl);
+err:
+	set_btree_node_io_error(b);
+	bch_cache_set_error(b->c, "%s at bucket %zu, block %zu, %u keys",
+			    err, PTR_BUCKET_NR(b->c, &b->key, 0),
+			    index(i, b), i->keys);
+	goto out;
+}
+
+void bch_btree_read(struct btree *b)
+{
+	BUG_ON(b->nsets || b->written);
+
+	if (!closure_trylock(&b->io.cl, &b->c->cl))
+		BUG();
+
+	b->io_start_time = local_clock();
+
+	btree_bio_init(b);
+	b->bio->bi_rw	= REQ_META|READ_SYNC;
+	b->bio->bi_size	= KEY_SIZE(&b->key) << 9;
+
+	bch_bio_map(b->bio, b->sets[0].data);
+
+	pr_debug("%s", pbtree(b));
+	trace_bcache_btree_read(b->bio);
+	bch_submit_bbio(b->bio, b->c, &b->key, 0);
+
+	continue_at(&b->io.cl, bch_btree_read_done, system_wq);
+}
+
+static void btree_complete_write(struct btree *b, struct btree_write *w)
+{
+	if (w->prio_blocked &&
+	    !atomic_sub_return(w->prio_blocked, &b->c->prio_blocked))
+		wake_up(&b->c->alloc_wait);
+
+	if (w->journal) {
+		atomic_dec_bug(w->journal);
+		__closure_wake_up(&b->c->journal.wait);
+	}
+
+	if (w->owner)
+		closure_put(w->owner);
+
+	w->prio_blocked	= 0;
+	w->journal	= NULL;
+	w->owner	= NULL;
+}
+
+static void __btree_write_done(struct closure *cl)
+{
+	struct btree *b = container_of(cl, struct btree, io.cl);
+	struct btree_write *w = btree_prev_write(b);
+
+	bch_bbio_free(b->bio, b->c);
+	b->bio = NULL;
+	btree_complete_write(b, w);
+
+	if (btree_node_dirty(b))
+		queue_delayed_work(btree_io_wq, &b->work,
+				   msecs_to_jiffies(30000));
+
+	closure_return(cl);
+}
+
+static void btree_write_done(struct closure *cl)
+{
+	struct btree *b = container_of(cl, struct btree, io.cl);
+	struct bio_vec *bv;
+	int n;
+
+	__bio_for_each_segment(bv, b->bio, n, 0)
+		__free_page(bv->bv_page);
+
+	__btree_write_done(cl);
+}
+
+static void do_btree_write(struct btree *b)
+{
+	struct closure *cl = &b->io.cl;
+	struct bset *i = b->sets[b->nsets].data;
+	BKEY_PADDED(key) k;
+
+	i->version	= BCACHE_BSET_VERSION;
+	i->csum		= btree_csum_set(b, i);
+
+	btree_bio_init(b);
+	b->bio->bi_rw	= REQ_META|WRITE_SYNC;
+	b->bio->bi_size	= set_blocks(i, b->c) * block_bytes(b->c);
+	bch_bio_map(b->bio, i);
+
+	bkey_copy(&k.key, &b->key);
+	SET_PTR_OFFSET(&k.key, 0, PTR_OFFSET(&k.key, 0) + bset_offset(b, i));
+
+	if (!bch_bio_alloc_pages(b->bio, GFP_NOIO)) {
+		int j;
+		struct bio_vec *bv;
+		void *base = (void *) ((unsigned long) i & ~(PAGE_SIZE - 1));
+
+		bio_for_each_segment(bv, b->bio, j)
+			memcpy(page_address(bv->bv_page),
+			       base + j * PAGE_SIZE, PAGE_SIZE);
+
+		trace_bcache_btree_write(b->bio);
+		bch_submit_bbio(b->bio, b->c, &k.key, 0);
+
+		continue_at(cl, btree_write_done, NULL);
+	} else {
+		b->bio->bi_vcnt = 0;
+		bch_bio_map(b->bio, i);
+
+		trace_bcache_btree_write(b->bio);
+		bch_submit_bbio(b->bio, b->c, &k.key, 0);
+
+		closure_sync(cl);
+		__btree_write_done(cl);
+	}
+}
+
+static void __btree_write(struct btree *b)
+{
+	struct bset *i = b->sets[b->nsets].data;
+
+	BUG_ON(current->bio_list);
+
+	closure_lock(&b->io, &b->c->cl);
+	cancel_delayed_work(&b->work);
+
+	clear_bit(BTREE_NODE_dirty,	 &b->flags);
+	change_bit(BTREE_NODE_write_idx, &b->flags);
+
+	bch_check_key_order(b, i);
+	BUG_ON(b->written && !i->keys);
+
+	do_btree_write(b);
+
+	pr_debug("%s block %i keys %i", pbtree(b), b->written, i->keys);
+
+	b->written += set_blocks(i, b->c);
+	atomic_long_add(set_blocks(i, b->c) * b->c->sb.block_size,
+			&PTR_CACHE(b->c, &b->key, 0)->btree_sectors_written);
+
+	bch_btree_sort_lazy(b);
+
+	if (b->written < btree_blocks(b))
+		bch_bset_init_next(b);
+}
+
+static void btree_write_work(struct work_struct *w)
+{
+	struct btree *b = container_of(to_delayed_work(w), struct btree, work);
+
+	down_write(&b->lock);
+
+	if (btree_node_dirty(b))
+		__btree_write(b);
+	up_write(&b->lock);
+}
+
+void bch_btree_write(struct btree *b, bool now, struct btree_op *op)
+{
+	struct bset *i = b->sets[b->nsets].data;
+	struct btree_write *w = btree_current_write(b);
+
+	BUG_ON(b->written &&
+	       (b->written >= btree_blocks(b) ||
+		i->seq != b->sets[0].data->seq ||
+		!i->keys));
+
+	if (!btree_node_dirty(b)) {
+		set_btree_node_dirty(b);
+		queue_delayed_work(btree_io_wq, &b->work,
+				   msecs_to_jiffies(30000));
+	}
+
+	w->prio_blocked += b->prio_blocked;
+	b->prio_blocked = 0;
+
+	if (op && op->journal && !b->level) {
+		if (w->journal &&
+		    journal_pin_cmp(b->c, w, op)) {
+			atomic_dec_bug(w->journal);
+			w->journal = NULL;
+		}
+
+		if (!w->journal) {
+			w->journal = op->journal;
+			atomic_inc(w->journal);
+		}
+	}
+
+	if (current->bio_list)
+		return;
+
+	/* Force write if set is too big */
+	if (now ||
+	    b->level ||
+	    set_bytes(i) > PAGE_SIZE - 48) {
+		if (op && now) {
+			/* Must wait on multiple writes */
+			BUG_ON(w->owner);
+			w->owner = &op->cl;
+			closure_get(&op->cl);
+		}
+
+		__btree_write(b);
+	}
+	BUG_ON(!b->written);
+}
+
+/*
+ * Btree in memory cache - allocation/freeing
+ * mca -> memory cache
+ */
+
+static void mca_reinit(struct btree *b)
+{
+	unsigned i;
+
+	b->flags	= 0;
+	b->written	= 0;
+	b->nsets	= 0;
+
+	for (i = 0; i < MAX_BSETS; i++)
+		b->sets[i].size = 0;
+	/*
+	 * Second loop starts at 1 because b->sets[0]->data is the memory we
+	 * allocated
+	 */
+	for (i = 1; i < MAX_BSETS; i++)
+		b->sets[i].data = NULL;
+}
+
+#define mca_reserve(c)	(((c->root && c->root->level)		\
+			  ? c->root->level : 1) * 8 + 16)
+#define mca_can_free(c)						\
+	max_t(int, 0, c->bucket_cache_used - mca_reserve(c))
+
+static void mca_data_free(struct btree *b)
+{
+	struct bset_tree *t = b->sets;
+	BUG_ON(!closure_is_unlocked(&b->io.cl));
+
+	if (bset_prev_bytes(b) < PAGE_SIZE)
+		kfree(t->prev);
+	else
+		free_pages((unsigned long) t->prev,
+			   get_order(bset_prev_bytes(b)));
+
+	if (bset_tree_bytes(b) < PAGE_SIZE)
+		kfree(t->tree);
+	else
+		free_pages((unsigned long) t->tree,
+			   get_order(bset_tree_bytes(b)));
+
+	free_pages((unsigned long) t->data, b->page_order);
+
+	t->prev = NULL;
+	t->tree = NULL;
+	t->data = NULL;
+	list_move(&b->list, &b->c->btree_cache_freed);
+	b->c->bucket_cache_used--;
+}
+
+static void mca_bucket_free(struct btree *b)
+{
+	BUG_ON(btree_node_dirty(b));
+
+	b->key.ptr[0] = 0;
+	hlist_del_init_rcu(&b->hash);
+	list_move(&b->list, &b->c->btree_cache_freeable);
+}
+
+static unsigned btree_order(struct bkey *k)
+{
+	return ilog2(KEY_SIZE(k) / PAGE_SECTORS ?: 1);
+}
+
+static void mca_data_alloc(struct btree *b, struct bkey *k, gfp_t gfp)
+{
+	struct bset_tree *t = b->sets;
+	BUG_ON(t->data);
+
+	b->page_order = max_t(unsigned,
+			      ilog2(b->c->btree_pages),
+			      btree_order(k));
+
+	t->data = (void *) __get_free_pages(gfp, b->page_order);
+	if (!t->data)
+		goto err;
+
+	t->tree = bset_tree_bytes(b) < PAGE_SIZE
+		? kmalloc(bset_tree_bytes(b), gfp)
+		: (void *) __get_free_pages(gfp, get_order(bset_tree_bytes(b)));
+	if (!t->tree)
+		goto err;
+
+	t->prev = bset_prev_bytes(b) < PAGE_SIZE
+		? kmalloc(bset_prev_bytes(b), gfp)
+		: (void *) __get_free_pages(gfp, get_order(bset_prev_bytes(b)));
+	if (!t->prev)
+		goto err;
+
+	list_move(&b->list, &b->c->btree_cache);
+	b->c->bucket_cache_used++;
+	return;
+err:
+	mca_data_free(b);
+}
+
+static struct btree *mca_bucket_alloc(struct cache_set *c,
+				      struct bkey *k, gfp_t gfp)
+{
+	struct btree *b = kzalloc(sizeof(struct btree), gfp);
+	if (!b)
+		return NULL;
+
+	init_rwsem(&b->lock);
+	lockdep_set_novalidate_class(&b->lock);
+	INIT_LIST_HEAD(&b->list);
+	INIT_DELAYED_WORK(&b->work, btree_write_work);
+	b->c = c;
+	closure_init_unlocked(&b->io);
+
+	mca_data_alloc(b, k, gfp);
+	return b;
+}
+
+static int mca_reap(struct btree *b, struct closure *cl, unsigned min_order)
+{
+	lockdep_assert_held(&b->c->bucket_lock);
+
+	if (!down_write_trylock(&b->lock))
+		return -ENOMEM;
+
+	if (b->page_order < min_order) {
+		rw_unlock(true, b);
+		return -ENOMEM;
+	}
+
+	BUG_ON(btree_node_dirty(b) && !b->sets[0].data);
+
+	if (cl && btree_node_dirty(b))
+		bch_btree_write(b, true, NULL);
+
+	if (cl)
+		closure_wait_event_async(&b->io.wait, cl,
+			 atomic_read(&b->io.cl.remaining) == -1);
+
+	if (btree_node_dirty(b) ||
+	    !closure_is_unlocked(&b->io.cl) ||
+	    work_pending(&b->work.work)) {
+		rw_unlock(true, b);
+		return -EAGAIN;
+	}
+
+	return 0;
+}
+
+static int bch_mca_shrink(struct shrinker *shrink, struct shrink_control *sc)
+{
+	struct cache_set *c = container_of(shrink, struct cache_set, shrink);
+	struct btree *b, *t;
+	unsigned long i, nr = sc->nr_to_scan;
+
+	if (c->shrinker_disabled)
+		return 0;
+
+	if (c->try_harder)
+		return 0;
+
+	/*
+	 * If nr == 0, we're supposed to return the number of items we have
+	 * cached. Not allowed to return -1.
+	 */
+	if (!nr)
+		return mca_can_free(c) * c->btree_pages;
+
+	/* Return -1 if we can't do anything right now */
+	if (sc->gfp_mask & __GFP_WAIT)
+		mutex_lock(&c->bucket_lock);
+	else if (!mutex_trylock(&c->bucket_lock))
+		return -1;
+
+	nr /= c->btree_pages;
+	nr = min_t(unsigned long, nr, mca_can_free(c));
+
+	i = 0;
+	list_for_each_entry_safe(b, t, &c->btree_cache_freeable, list) {
+		if (!nr)
+			break;
+
+		if (++i > 3 &&
+		    !mca_reap(b, NULL, 0)) {
+			mca_data_free(b);
+			rw_unlock(true, b);
+			--nr;
+		}
+	}
+
+	/*
+	 * Can happen right when we first start up, before we've read in any
+	 * btree nodes
+	 */
+	if (list_empty(&c->btree_cache))
+		goto out;
+
+	for (i = 0; nr && i < c->bucket_cache_used; i++) {
+		b = list_first_entry(&c->btree_cache, struct btree, list);
+		list_rotate_left(&c->btree_cache);
+
+		if (!b->accessed &&
+		    !mca_reap(b, NULL, 0)) {
+			mca_bucket_free(b);
+			mca_data_free(b);
+			rw_unlock(true, b);
+			--nr;
+		} else
+			b->accessed = 0;
+	}
+out:
+	nr = mca_can_free(c) * c->btree_pages;
+	mutex_unlock(&c->bucket_lock);
+	return nr;
+}
+
+void bch_btree_cache_free(struct cache_set *c)
+{
+	struct btree *b;
+	struct closure cl;
+	closure_init_stack(&cl);
+
+	if (c->shrink.list.next)
+		unregister_shrinker(&c->shrink);
+
+	mutex_lock(&c->bucket_lock);
+
+#ifdef CONFIG_BCACHE_DEBUG
+	if (c->verify_data)
+		list_move(&c->verify_data->list, &c->btree_cache);
+#endif
+
+	list_splice(&c->btree_cache_freeable,
+		    &c->btree_cache);
+
+	while (!list_empty(&c->btree_cache)) {
+		b = list_first_entry(&c->btree_cache, struct btree, list);
+
+		if (btree_node_dirty(b))
+			btree_complete_write(b, btree_current_write(b));
+		clear_bit(BTREE_NODE_dirty, &b->flags);
+
+		mca_data_free(b);
+	}
+
+	while (!list_empty(&c->btree_cache_freed)) {
+		b = list_first_entry(&c->btree_cache_freed,
+				     struct btree, list);
+		list_del(&b->list);
+		cancel_delayed_work_sync(&b->work);
+		kfree(b);
+	}
+
+	mutex_unlock(&c->bucket_lock);
+}
+
+int bch_btree_cache_alloc(struct cache_set *c)
+{
+	unsigned i;
+
+	/* XXX: doesn't check for errors */
+
+	closure_init_unlocked(&c->gc);
+
+	for (i = 0; i < mca_reserve(c); i++)
+		mca_bucket_alloc(c, &ZERO_KEY, GFP_KERNEL);
+
+	list_splice_init(&c->btree_cache,
+			 &c->btree_cache_freeable);
+
+#ifdef CONFIG_BCACHE_DEBUG
+	mutex_init(&c->verify_lock);
+
+	c->verify_data = mca_bucket_alloc(c, &ZERO_KEY, GFP_KERNEL);
+
+	if (c->verify_data &&
+	    c->verify_data->sets[0].data)
+		list_del_init(&c->verify_data->list);
+	else
+		c->verify_data = NULL;
+#endif
+
+	c->shrink.shrink = bch_mca_shrink;
+	c->shrink.seeks = 4;
+	c->shrink.batch = c->btree_pages * 2;
+	register_shrinker(&c->shrink);
+
+	return 0;
+}
+
+/* Btree in memory cache - hash table */
+
+static struct hlist_head *mca_hash(struct cache_set *c, struct bkey *k)
+{
+	return &c->bucket_hash[hash_32(PTR_HASH(c, k), BUCKET_HASH_BITS)];
+}
+
+static struct btree *mca_find(struct cache_set *c, struct bkey *k)
+{
+	struct btree *b;
+
+	rcu_read_lock();
+	hlist_for_each_entry_rcu(b, mca_hash(c, k), hash)
+		if (PTR_HASH(c, &b->key) == PTR_HASH(c, k))
+			goto out;
+	b = NULL;
+out:
+	rcu_read_unlock();
+	return b;
+}
+
+static struct btree *mca_cannibalize(struct cache_set *c, struct bkey *k,
+				     int level, struct closure *cl)
+{
+	int ret = -ENOMEM;
+	struct btree *i;
+
+	if (!cl)
+		return ERR_PTR(-ENOMEM);
+
+	/*
+	 * Trying to free up some memory - i.e. reuse some btree nodes - may
+	 * require initiating IO to flush the dirty part of the node. If we're
+	 * running under generic_make_request(), that IO will never finish and
+	 * we would deadlock. Returning -EAGAIN causes the cache lookup code to
+	 * punt to workqueue and retry.
+	 */
+	if (current->bio_list)
+		return ERR_PTR(-EAGAIN);
+
+	if (c->try_harder && c->try_harder != cl) {
+		closure_wait_event_async(&c->try_wait, cl, !c->try_harder);
+		return ERR_PTR(-EAGAIN);
+	}
+
+	/* XXX: tracepoint */
+	c->try_harder = cl;
+	c->try_harder_start = local_clock();
+retry:
+	list_for_each_entry_reverse(i, &c->btree_cache, list) {
+		int r = mca_reap(i, cl, btree_order(k));
+		if (!r)
+			return i;
+		if (r != -ENOMEM)
+			ret = r;
+	}
+
+	if (ret == -EAGAIN &&
+	    closure_blocking(cl)) {
+		mutex_unlock(&c->bucket_lock);
+		closure_sync(cl);
+		mutex_lock(&c->bucket_lock);
+		goto retry;
+	}
+
+	return ERR_PTR(ret);
+}
+
+/*
+ * We can only have one thread cannibalizing other cached btree nodes at a time,
+ * or we'll deadlock. We use an open coded mutex to ensure that, which a
+ * cannibalize_bucket() will take. This means every time we unlock the root of
+ * the btree, we need to release this lock if we have it held.
+ */
+void bch_cannibalize_unlock(struct cache_set *c, struct closure *cl)
+{
+	if (c->try_harder == cl) {
+		bch_time_stats_update(&c->try_harder_time, c->try_harder_start);
+		c->try_harder = NULL;
+		__closure_wake_up(&c->try_wait);
+	}
+}
+
+static struct btree *mca_alloc(struct cache_set *c, struct bkey *k,
+			       int level, struct closure *cl)
+{
+	struct btree *b;
+
+	lockdep_assert_held(&c->bucket_lock);
+
+	if (mca_find(c, k))
+		return NULL;
+
+	/* btree_free() doesn't free memory; it sticks the node on the end of
+	 * the list. Check if there's any freed nodes there:
+	 */
+	list_for_each_entry(b, &c->btree_cache_freeable, list)
+		if (!mca_reap(b, NULL, btree_order(k)))
+			goto out;
+
+	/* We never free struct btree itself, just the memory that holds the on
+	 * disk node. Check the freed list before allocating a new one:
+	 */
+	list_for_each_entry(b, &c->btree_cache_freed, list)
+		if (!mca_reap(b, NULL, 0)) {
+			mca_data_alloc(b, k, __GFP_NOWARN|GFP_NOIO);
+			if (!b->sets[0].data)
+				goto err;
+			else
+				goto out;
+		}
+
+	b = mca_bucket_alloc(c, k, __GFP_NOWARN|GFP_NOIO);
+	if (!b)
+		goto err;
+
+	BUG_ON(!down_write_trylock(&b->lock));
+	if (!b->sets->data)
+		goto err;
+out:
+	BUG_ON(!closure_is_unlocked(&b->io.cl));
+
+	bkey_copy(&b->key, k);
+	list_move(&b->list, &c->btree_cache);
+	hlist_del_init_rcu(&b->hash);
+	hlist_add_head_rcu(&b->hash, mca_hash(c, k));
+
+	lock_set_subclass(&b->lock.dep_map, level + 1, _THIS_IP_);
+	b->level	= level;
+
+	mca_reinit(b);
+
+	return b;
+err:
+	if (b)
+		rw_unlock(true, b);
+
+	b = mca_cannibalize(c, k, level, cl);
+	if (!IS_ERR(b))
+		goto out;
+
+	return b;
+}
+
+/**
+ * bch_btree_node_get - find a btree node in the cache and lock it, reading it
+ * in from disk if necessary.
+ *
+ * If IO is necessary, it uses the closure embedded in struct btree_op to wait;
+ * if that closure is in non blocking mode, will return -EAGAIN.
+ *
+ * The btree node will have either a read or a write lock held, depending on
+ * level and op->lock.
+ */
+struct btree *bch_btree_node_get(struct cache_set *c, struct bkey *k,
+				 int level, struct btree_op *op)
+{
+	int i = 0;
+	bool write = level <= op->lock;
+	struct btree *b;
+
+	BUG_ON(level < 0);
+retry:
+	b = mca_find(c, k);
+
+	if (!b) {
+		mutex_lock(&c->bucket_lock);
+		b = mca_alloc(c, k, level, &op->cl);
+		mutex_unlock(&c->bucket_lock);
+
+		if (!b)
+			goto retry;
+		if (IS_ERR(b))
+			return b;
+
+		bch_btree_read(b);
+
+		if (!write)
+			downgrade_write(&b->lock);
+	} else {
+		rw_lock(write, b, level);
+		if (PTR_HASH(c, &b->key) != PTR_HASH(c, k)) {
+			rw_unlock(write, b);
+			goto retry;
+		}
+		BUG_ON(b->level != level);
+	}
+
+	b->accessed = 1;
+
+	for (; i <= b->nsets && b->sets[i].size; i++) {
+		prefetch(b->sets[i].tree);
+		prefetch(b->sets[i].data);
+	}
+
+	for (; i <= b->nsets; i++)
+		prefetch(b->sets[i].data);
+
+	if (!closure_wait_event(&b->io.wait, &op->cl,
+				btree_node_read_done(b))) {
+		rw_unlock(write, b);
+		b = ERR_PTR(-EAGAIN);
+	} else if (btree_node_io_error(b)) {
+		rw_unlock(write, b);
+		b = ERR_PTR(-EIO);
+	} else
+		BUG_ON(!b->written);
+
+	return b;
+}
+
+static void btree_node_prefetch(struct cache_set *c, struct bkey *k, int level)
+{
+	struct btree *b;
+
+	mutex_lock(&c->bucket_lock);
+	b = mca_alloc(c, k, level, NULL);
+	mutex_unlock(&c->bucket_lock);
+
+	if (!IS_ERR_OR_NULL(b)) {
+		bch_btree_read(b);
+		rw_unlock(true, b);
+	}
+}
+
+/* Btree alloc */
+
+static void btree_node_free(struct btree *b, struct btree_op *op)
+{
+	unsigned i;
+
+	/*
+	 * The BUG_ON() in btree_node_get() implies that we must have a write
+	 * lock on parent to free or even invalidate a node
+	 */
+	BUG_ON(op->lock <= b->level);
+	BUG_ON(b == b->c->root);
+	pr_debug("bucket %s", pbtree(b));
+
+	if (btree_node_dirty(b))
+		btree_complete_write(b, btree_current_write(b));
+	clear_bit(BTREE_NODE_dirty, &b->flags);
+
+	if (b->prio_blocked &&
+	    !atomic_sub_return(b->prio_blocked, &b->c->prio_blocked))
+		wake_up(&b->c->alloc_wait);
+
+	b->prio_blocked = 0;
+
+	cancel_delayed_work(&b->work);
+
+	mutex_lock(&b->c->bucket_lock);
+
+	for (i = 0; i < KEY_PTRS(&b->key); i++) {
+		BUG_ON(atomic_read(&PTR_BUCKET(b->c, &b->key, i)->pin));
+
+		bch_inc_gen(PTR_CACHE(b->c, &b->key, i),
+			    PTR_BUCKET(b->c, &b->key, i));
+	}
+
+	bch_bucket_free(b->c, &b->key);
+	mca_bucket_free(b);
+	mutex_unlock(&b->c->bucket_lock);
+}
+
+struct btree *bch_btree_node_alloc(struct cache_set *c, int level,
+				   struct closure *cl)
+{
+	BKEY_PADDED(key) k;
+	struct btree *b = ERR_PTR(-EAGAIN);
+
+	mutex_lock(&c->bucket_lock);
+retry:
+	if (__bch_bucket_alloc_set(c, WATERMARK_METADATA, &k.key, 1, cl))
+		goto err;
+
+	SET_KEY_SIZE(&k.key, c->btree_pages * PAGE_SECTORS);
+
+	b = mca_alloc(c, &k.key, level, cl);
+	if (IS_ERR(b))
+		goto err_free;
+
+	if (!b) {
+		cache_bug(c,
+			"Tried to allocate bucket that was in btree cache");
+		__bkey_put(c, &k.key);
+		goto retry;
+	}
+
+	set_btree_node_read_done(b);
+	b->accessed = 1;
+	bch_bset_init_next(b);
+
+	mutex_unlock(&c->bucket_lock);
+	return b;
+err_free:
+	bch_bucket_free(c, &k.key);
+	__bkey_put(c, &k.key);
+err:
+	mutex_unlock(&c->bucket_lock);
+	return b;
+}
+
+static struct btree *btree_node_alloc_replacement(struct btree *b,
+						  struct closure *cl)
+{
+	struct btree *n = bch_btree_node_alloc(b->c, b->level, cl);
+	if (!IS_ERR_OR_NULL(n))
+		bch_btree_sort_into(b, n);
+
+	return n;
+}
+
+/* Garbage collection */
+
+uint8_t __bch_btree_mark_key(struct cache_set *c, int level, struct bkey *k)
+{
+	uint8_t stale = 0;
+	unsigned i;
+	struct bucket *g;
+
+	/*
+	 * ptr_invalid() can't return true for the keys that mark btree nodes as
+	 * freed, but since ptr_bad() returns true we'll never actually use them
+	 * for anything and thus we don't want mark their pointers here
+	 */
+	if (!bkey_cmp(k, &ZERO_KEY))
+		return stale;
+
+	for (i = 0; i < KEY_PTRS(k); i++) {
+		if (!ptr_available(c, k, i))
+			continue;
+
+		g = PTR_BUCKET(c, k, i);
+
+		if (gen_after(g->gc_gen, PTR_GEN(k, i)))
+			g->gc_gen = PTR_GEN(k, i);
+
+		if (ptr_stale(c, k, i)) {
+			stale = max(stale, ptr_stale(c, k, i));
+			continue;
+		}
+
+		cache_bug_on(GC_MARK(g) &&
+			     (GC_MARK(g) == GC_MARK_METADATA) != (level != 0),
+			     c, "inconsistent ptrs: mark = %llu, level = %i",
+			     GC_MARK(g), level);
+
+		if (level)
+			SET_GC_MARK(g, GC_MARK_METADATA);
+		else if (KEY_DIRTY(k))
+			SET_GC_MARK(g, GC_MARK_DIRTY);
+
+		/* guard against overflow */
+		SET_GC_SECTORS_USED(g, min_t(unsigned,
+					     GC_SECTORS_USED(g) + KEY_SIZE(k),
+					     (1 << 14) - 1));
+
+		BUG_ON(!GC_SECTORS_USED(g));
+	}
+
+	return stale;
+}
+
+#define btree_mark_key(b, k)	__bch_btree_mark_key(b->c, b->level, k)
+
+static int btree_gc_mark_node(struct btree *b, unsigned *keys,
+			      struct gc_stat *gc)
+{
+	uint8_t stale = 0;
+	unsigned last_dev = -1;
+	struct bcache_device *d = NULL;
+	struct bkey *k;
+	struct btree_iter iter;
+	struct bset_tree *t;
+
+	gc->nodes++;
+
+	for_each_key_filter(b, k, &iter, bch_ptr_invalid) {
+		if (last_dev != KEY_INODE(k)) {
+			last_dev = KEY_INODE(k);
+
+			d = KEY_INODE(k) < b->c->nr_uuids
+				? b->c->devices[last_dev]
+				: NULL;
+		}
+
+		stale = max(stale, btree_mark_key(b, k));
+
+		if (bch_ptr_bad(b, k))
+			continue;
+
+		*keys += bkey_u64s(k);
+
+		gc->key_bytes += bkey_u64s(k);
+		gc->nkeys++;
+
+		gc->data += KEY_SIZE(k);
+		if (KEY_DIRTY(k)) {
+			gc->dirty += KEY_SIZE(k);
+			if (d)
+				d->sectors_dirty_gc += KEY_SIZE(k);
+		}
+	}
+
+	for (t = b->sets; t <= &b->sets[b->nsets]; t++)
+		btree_bug_on(t->size &&
+			     bset_written(b, t) &&
+			     bkey_cmp(&b->key, &t->end) < 0,
+			     b, "found short btree key in gc");
+
+	return stale;
+}
+
+static struct btree *btree_gc_alloc(struct btree *b, struct bkey *k,
+				    struct btree_op *op)
+{
+	/*
+	 * We block priorities from being written for the duration of garbage
+	 * collection, so we can't sleep in btree_alloc() ->
+	 * bch_bucket_alloc_set(), or we'd risk deadlock - so we don't pass it
+	 * our closure.
+	 */
+	struct btree *n = btree_node_alloc_replacement(b, NULL);
+
+	if (!IS_ERR_OR_NULL(n)) {
+		swap(b, n);
+
+		memcpy(k->ptr, b->key.ptr,
+		       sizeof(uint64_t) * KEY_PTRS(&b->key));
+
+		__bkey_put(b->c, &b->key);
+		atomic_inc(&b->c->prio_blocked);
+		b->prio_blocked++;
+
+		btree_node_free(n, op);
+		up_write(&n->lock);
+	}
+
+	return b;
+}
+
+/*
+ * Leaving this at 2 until we've got incremental garbage collection done; it
+ * could be higher (and has been tested with 4) except that garbage collection
+ * could take much longer, adversely affecting latency.
+ */
+#define GC_MERGE_NODES	2U
+
+struct gc_merge_info {
+	struct btree	*b;
+	struct bkey	*k;
+	unsigned	keys;
+};
+
+static void btree_gc_coalesce(struct btree *b, struct btree_op *op,
+			      struct gc_stat *gc, struct gc_merge_info *r)
+{
+	unsigned nodes = 0, keys = 0, blocks;
+	int i;
+
+	while (nodes < GC_MERGE_NODES && r[nodes].b)
+		keys += r[nodes++].keys;
+
+	blocks = btree_default_blocks(b->c) * 2 / 3;
+
+	if (nodes < 2 ||
+	    __set_blocks(b->sets[0].data, keys, b->c) > blocks * (nodes - 1))
+		return;
+
+	for (i = nodes - 1; i >= 0; --i) {
+		if (r[i].b->written)
+			r[i].b = btree_gc_alloc(r[i].b, r[i].k, op);
+
+		if (r[i].b->written)
+			return;
+	}
+
+	for (i = nodes - 1; i > 0; --i) {
+		struct bset *n1 = r[i].b->sets->data;
+		struct bset *n2 = r[i - 1].b->sets->data;
+		struct bkey *k, *last = NULL;
+
+		keys = 0;
+
+		if (i == 1) {
+			/*
+			 * Last node we're not getting rid of - we're getting
+			 * rid of the node at r[0]. Have to try and fit all of
+			 * the remaining keys into this node; we can't ensure
+			 * they will always fit due to rounding and variable
+			 * length keys (shouldn't be possible in practice,
+			 * though)
+			 */
+			if (__set_blocks(n1, n1->keys + r->keys,
+					 b->c) > btree_blocks(r[i].b))
+				return;
+
+			keys = n2->keys;
+			last = &r->b->key;
+		} else
+			for (k = n2->start;
+			     k < end(n2);
+			     k = bkey_next(k)) {
+				if (__set_blocks(n1, n1->keys + keys +
+						 bkey_u64s(k), b->c) > blocks)
+					break;
+
+				last = k;
+				keys += bkey_u64s(k);
+			}
+
+		BUG_ON(__set_blocks(n1, n1->keys + keys,
+				    b->c) > btree_blocks(r[i].b));
+
+		if (last) {
+			bkey_copy_key(&r[i].b->key, last);
+			bkey_copy_key(r[i].k, last);
+		}
+
+		memcpy(end(n1),
+		       n2->start,
+		       (void *) node(n2, keys) - (void *) n2->start);
+
+		n1->keys += keys;
+
+		memmove(n2->start,
+			node(n2, keys),
+			(void *) end(n2) - (void *) node(n2, keys));
+
+		n2->keys -= keys;
+
+		r[i].keys	= n1->keys;
+		r[i - 1].keys	= n2->keys;
+	}
+
+	btree_node_free(r->b, op);
+	up_write(&r->b->lock);
+
+	pr_debug("coalesced %u nodes", nodes);
+
+	gc->nodes--;
+	nodes--;
+
+	memmove(&r[0], &r[1], sizeof(struct gc_merge_info) * nodes);
+	memset(&r[nodes], 0, sizeof(struct gc_merge_info));
+}
+
+static int btree_gc_recurse(struct btree *b, struct btree_op *op,
+			    struct closure *writes, struct gc_stat *gc)
+{
+	void write(struct btree *r)
+	{
+		if (!r->written)
+			bch_btree_write(r, true, op);
+		else if (btree_node_dirty(r)) {
+			BUG_ON(btree_current_write(r)->owner);
+			btree_current_write(r)->owner = writes;
+			closure_get(writes);
+
+			bch_btree_write(r, true, NULL);
+		}
+
+		up_write(&r->lock);
+	}
+
+	int ret = 0, stale;
+	unsigned i;
+	struct gc_merge_info r[GC_MERGE_NODES];
+
+	memset(r, 0, sizeof(r));
+
+	while ((r->k = bch_next_recurse_key(b, &b->c->gc_done))) {
+		r->b = bch_btree_node_get(b->c, r->k, b->level - 1, op);
+
+		if (IS_ERR(r->b)) {
+			ret = PTR_ERR(r->b);
+			break;
+		}
+
+		r->keys	= 0;
+		stale = btree_gc_mark_node(r->b, &r->keys, gc);
+
+		if (!b->written &&
+		    (r->b->level || stale > 10 ||
+		     b->c->gc_always_rewrite))
+			r->b = btree_gc_alloc(r->b, r->k, op);
+
+		if (r->b->level)
+			ret = btree_gc_recurse(r->b, op, writes, gc);
+
+		if (ret) {
+			write(r->b);
+			break;
+		}
+
+		bkey_copy_key(&b->c->gc_done, r->k);
+
+		if (!b->written)
+			btree_gc_coalesce(b, op, gc, r);
+
+		if (r[GC_MERGE_NODES - 1].b)
+			write(r[GC_MERGE_NODES - 1].b);
+
+		memmove(&r[1], &r[0],
+			sizeof(struct gc_merge_info) * (GC_MERGE_NODES - 1));
+
+		/* When we've got incremental GC working, we'll want to do
+		 * if (should_resched())
+		 *	return -EAGAIN;
+		 */
+		cond_resched();
+#if 0
+		if (need_resched()) {
+			ret = -EAGAIN;
+			break;
+		}
+#endif
+	}
+
+	for (i = 1; i < GC_MERGE_NODES && r[i].b; i++)
+		write(r[i].b);
+
+	/* Might have freed some children, must remove their keys */
+	if (!b->written)
+		bch_btree_sort(b);
+
+	return ret;
+}
+
+static int bch_btree_gc_root(struct btree *b, struct btree_op *op,
+			     struct closure *writes, struct gc_stat *gc)
+{
+	struct btree *n = NULL;
+	unsigned keys = 0;
+	int ret = 0, stale = btree_gc_mark_node(b, &keys, gc);
+
+	if (b->level || stale > 10)
+		n = btree_node_alloc_replacement(b, NULL);
+
+	if (!IS_ERR_OR_NULL(n))
+		swap(b, n);
+
+	if (b->level)
+		ret = btree_gc_recurse(b, op, writes, gc);
+
+	if (!b->written || btree_node_dirty(b)) {
+		atomic_inc(&b->c->prio_blocked);
+		b->prio_blocked++;
+		bch_btree_write(b, true, n ? op : NULL);
+	}
+
+	if (!IS_ERR_OR_NULL(n)) {
+		closure_sync(&op->cl);
+		bch_btree_set_root(b);
+		btree_node_free(n, op);
+		rw_unlock(true, b);
+	}
+
+	return ret;
+}
+
+static void btree_gc_start(struct cache_set *c)
+{
+	struct cache *ca;
+	struct bucket *b;
+	struct bcache_device **d;
+	unsigned i;
+
+	if (!c->gc_mark_valid)
+		return;
+
+	mutex_lock(&c->bucket_lock);
+
+	c->gc_mark_valid = 0;
+	c->gc_done = ZERO_KEY;
+
+	for_each_cache(ca, c, i)
+		for_each_bucket(b, ca) {
+			b->gc_gen = b->gen;
+			if (!atomic_read(&b->pin))
+				SET_GC_MARK(b, GC_MARK_RECLAIMABLE);
+		}
+
+	for (d = c->devices;
+	     d < c->devices + c->nr_uuids;
+	     d++)
+		if (*d)
+			(*d)->sectors_dirty_gc = 0;
+
+	mutex_unlock(&c->bucket_lock);
+}
+
+size_t bch_btree_gc_finish(struct cache_set *c)
+{
+	size_t available = 0;
+	struct bucket *b;
+	struct cache *ca;
+	struct bcache_device **d;
+	unsigned i;
+
+	mutex_lock(&c->bucket_lock);
+
+	set_gc_sectors(c);
+	c->gc_mark_valid = 1;
+	c->need_gc	= 0;
+
+	if (c->root)
+		for (i = 0; i < KEY_PTRS(&c->root->key); i++)
+			SET_GC_MARK(PTR_BUCKET(c, &c->root->key, i),
+				    GC_MARK_METADATA);
+
+	for (i = 0; i < KEY_PTRS(&c->uuid_bucket); i++)
+		SET_GC_MARK(PTR_BUCKET(c, &c->uuid_bucket, i),
+			    GC_MARK_METADATA);
+
+	for_each_cache(ca, c, i) {
+		uint64_t *i;
+
+		ca->invalidate_needs_gc = 0;
+
+		for (i = ca->sb.d; i < ca->sb.d + ca->sb.keys; i++)
+			SET_GC_MARK(ca->buckets + *i, GC_MARK_METADATA);
+
+		for (i = ca->prio_buckets;
+		     i < ca->prio_buckets + prio_buckets(ca) * 2; i++)
+			SET_GC_MARK(ca->buckets + *i, GC_MARK_METADATA);
+
+		for_each_bucket(b, ca) {
+			b->last_gc	= b->gc_gen;
+			c->need_gc	= max(c->need_gc, bucket_gc_gen(b));
+
+			if (!atomic_read(&b->pin) &&
+			    GC_MARK(b) == GC_MARK_RECLAIMABLE) {
+				available++;
+				if (!GC_SECTORS_USED(b))
+					bch_bucket_add_unused(ca, b);
+			}
+		}
+	}
+
+	for (d = c->devices;
+	     d < c->devices + c->nr_uuids;
+	     d++)
+		if (*d) {
+			unsigned long last =
+				atomic_long_read(&((*d)->sectors_dirty));
+			long difference = (*d)->sectors_dirty_gc - last;
+
+			pr_debug("sectors dirty off by %li", difference);
+
+			(*d)->sectors_dirty_last += difference;
+
+			atomic_long_set(&((*d)->sectors_dirty),
+					(*d)->sectors_dirty_gc);
+		}
+
+	mutex_unlock(&c->bucket_lock);
+	return available;
+}
+
+static void bch_btree_gc(struct closure *cl)
+{
+	struct cache_set *c = container_of(cl, struct cache_set, gc.cl);
+	int ret;
+	unsigned long available;
+	struct gc_stat stats;
+	struct closure writes;
+	struct btree_op op;
+
+	uint64_t start_time = local_clock();
+	trace_bcache_gc_start(c->sb.set_uuid);
+	blktrace_msg_all(c, "Starting gc");
+
+	memset(&stats, 0, sizeof(struct gc_stat));
+	closure_init_stack(&writes);
+	bch_btree_op_init_stack(&op);
+	op.lock = SHRT_MAX;
+
+	btree_gc_start(c);
+
+	ret = btree_root(gc_root, c, &op, &writes, &stats);
+	closure_sync(&op.cl);
+	closure_sync(&writes);
+
+	if (ret) {
+		blktrace_msg_all(c, "Stopped gc");
+		pr_warn("gc failed!");
+
+		continue_at(cl, bch_btree_gc, bch_gc_wq);
+	}
+
+	/* Possibly wait for new UUIDs or whatever to hit disk */
+	bch_journal_meta(c, &op.cl);
+	closure_sync(&op.cl);
+
+	available = bch_btree_gc_finish(c);
+
+	bch_time_stats_update(&c->btree_gc_time, start_time);
+
+	stats.key_bytes *= sizeof(uint64_t);
+	stats.dirty	<<= 9;
+	stats.data	<<= 9;
+	stats.in_use	= (c->nbuckets - available) * 100 / c->nbuckets;
+	memcpy(&c->gc_stats, &stats, sizeof(struct gc_stat));
+	blktrace_msg_all(c, "Finished gc");
+
+	trace_bcache_gc_end(c->sb.set_uuid);
+	wake_up(&c->alloc_wait);
+
+	continue_at(cl, bch_moving_gc, bch_gc_wq);
+}
+
+void bch_queue_gc(struct cache_set *c)
+{
+	closure_trylock_call(&c->gc.cl, bch_btree_gc, bch_gc_wq, &c->cl);
+}
+
+/* Initial partial gc */
+
+static int bch_btree_check_recurse(struct btree *b, struct btree_op *op,
+				   unsigned long **seen)
+{
+	int ret;
+	unsigned i;
+	struct bkey *k;
+	struct bucket *g;
+	struct btree_iter iter;
+
+	for_each_key_filter(b, k, &iter, bch_ptr_invalid) {
+		for (i = 0; i < KEY_PTRS(k); i++) {
+			if (!ptr_available(b->c, k, i))
+				continue;
+
+			g = PTR_BUCKET(b->c, k, i);
+
+			if (!__test_and_set_bit(PTR_BUCKET_NR(b->c, k, i),
+						seen[PTR_DEV(k, i)]) ||
+			    !ptr_stale(b->c, k, i)) {
+				g->gen = PTR_GEN(k, i);
+
+				if (b->level)
+					g->prio = BTREE_PRIO;
+				else if (g->prio == BTREE_PRIO)
+					g->prio = INITIAL_PRIO;
+			}
+		}
+
+		btree_mark_key(b, k);
+	}
+
+	if (b->level) {
+		k = bch_next_recurse_key(b, &ZERO_KEY);
+
+		while (k) {
+			struct bkey *p = bch_next_recurse_key(b, k);
+			if (p)
+				btree_node_prefetch(b->c, p, b->level - 1);
+
+			ret = btree(check_recurse, k, b, op, seen);
+			if (ret)
+				return ret;
+
+			k = p;
+		}
+	}
+
+	return 0;
+}
+
+int bch_btree_check(struct cache_set *c, struct btree_op *op)
+{
+	int ret = -ENOMEM;
+	unsigned i;
+	unsigned long *seen[MAX_CACHES_PER_SET];
+
+	memset(seen, 0, sizeof(seen));
+
+	for (i = 0; c->cache[i]; i++) {
+		size_t n = DIV_ROUND_UP(c->cache[i]->sb.nbuckets, 8);
+		seen[i] = kmalloc(n, GFP_KERNEL);
+		if (!seen[i])
+			goto err;
+
+		/* Disables the seen array until prio_read() uses it too */
+		memset(seen[i], 0xFF, n);
+	}
+
+	ret = btree_root(check_recurse, c, op, seen);
+err:
+	for (i = 0; i < MAX_CACHES_PER_SET; i++)
+		kfree(seen[i]);
+	return ret;
+}
+
+/* Btree insertion */
+
+static void shift_keys(struct btree *b, struct bkey *where, struct bkey *insert)
+{
+	struct bset *i = b->sets[b->nsets].data;
+
+	memmove((uint64_t *) where + bkey_u64s(insert),
+		where,
+		(void *) end(i) - (void *) where);
+
+	i->keys += bkey_u64s(insert);
+	bkey_copy(where, insert);
+	bch_bset_fix_lookup_table(b, where);
+}
+
+static bool fix_overlapping_extents(struct btree *b,
+				    struct bkey *insert,
+				    struct btree_iter *iter,
+				    struct btree_op *op)
+{
+	void subtract_dirty(struct bkey *k, int sectors)
+	{
+		struct bcache_device *d = b->c->devices[KEY_INODE(k)];
+
+		if (KEY_DIRTY(k) && d)
+			atomic_long_sub(sectors, &d->sectors_dirty);
+	}
+
+	unsigned old_size, sectors_found = 0;
+
+	while (1) {
+		struct bkey *k = bch_btree_iter_next(iter);
+		if (!k ||
+		    bkey_cmp(&START_KEY(k), insert) >= 0)
+			break;
+
+		if (bkey_cmp(k, &START_KEY(insert)) <= 0)
+			continue;
+
+		old_size = KEY_SIZE(k);
+
+		/*
+		 * We might overlap with 0 size extents; we can't skip these
+		 * because if they're in the set we're inserting to we have to
+		 * adjust them so they don't overlap with the key we're
+		 * inserting. But we don't want to check them for BTREE_REPLACE
+		 * operations.
+		 */
+
+		if (op->type == BTREE_REPLACE &&
+		    KEY_SIZE(k)) {
+			/*
+			 * k might have been split since we inserted/found the
+			 * key we're replacing
+			 */
+			unsigned i;
+			uint64_t offset = KEY_START(k) -
+				KEY_START(&op->replace);
+
+			/* But it must be a subset of the replace key */
+			if (KEY_START(k) < KEY_START(&op->replace) ||
+			    KEY_OFFSET(k) > KEY_OFFSET(&op->replace))
+				goto check_failed;
+
+			/* We didn't find a key that we were supposed to */
+			if (KEY_START(k) > KEY_START(insert) + sectors_found)
+				goto check_failed;
+
+			if (KEY_PTRS(&op->replace) != KEY_PTRS(k))
+				goto check_failed;
+
+			/* skip past gen */
+			offset <<= 8;
+
+			BUG_ON(!KEY_PTRS(&op->replace));
+
+			for (i = 0; i < KEY_PTRS(&op->replace); i++)
+				if (k->ptr[i] != op->replace.ptr[i] + offset)
+					goto check_failed;
+
+			sectors_found = KEY_OFFSET(k) - KEY_START(insert);
+		}
+
+		if (bkey_cmp(insert, k) < 0 &&
+		    bkey_cmp(&START_KEY(insert), &START_KEY(k)) > 0) {
+			/*
+			 * We overlapped in the middle of an existing key: that
+			 * means we have to split the old key. But we have to do
+			 * slightly different things depending on whether the
+			 * old key has been written out yet.
+			 */
+
+			struct bkey *top;
+
+			subtract_dirty(k, KEY_SIZE(insert));
+
+			if (bkey_written(b, k)) {
+				/*
+				 * We insert a new key to cover the top of the
+				 * old key, and the old key is modified in place
+				 * to represent the bottom split.
+				 *
+				 * It's completely arbitrary whether the new key
+				 * is the top or the bottom, but it has to match
+				 * up with what btree_sort_fixup() does - it
+				 * doesn't check for this kind of overlap, it
+				 * depends on us inserting a new key for the top
+				 * here.
+				 */
+				top = bch_bset_search(b, &b->sets[b->nsets],
+						      insert);
+				shift_keys(b, top, k);
+			} else {
+				BKEY_PADDED(key) temp;
+				bkey_copy(&temp.key, k);
+				shift_keys(b, k, &temp.key);
+				top = bkey_next(k);
+			}
+
+			bch_cut_front(insert, top);
+			bch_cut_back(&START_KEY(insert), k);
+			bch_bset_fix_invalidated_key(b, k);
+			return false;
+		}
+
+		if (bkey_cmp(insert, k) < 0) {
+			bch_cut_front(insert, k);
+		} else {
+			if (bkey_written(b, k) &&
+			    bkey_cmp(&START_KEY(insert), &START_KEY(k)) <= 0) {
+				/*
+				 * Completely overwrote, so we don't have to
+				 * invalidate the binary search tree
+				 */
+				bch_cut_front(k, k);
+			} else {
+				__bch_cut_back(&START_KEY(insert), k);
+				bch_bset_fix_invalidated_key(b, k);
+			}
+		}
+
+		subtract_dirty(k, old_size - KEY_SIZE(k));
+	}
+
+check_failed:
+	if (op->type == BTREE_REPLACE) {
+		if (!sectors_found) {
+			op->insert_collision = true;
+			return true;
+		} else if (sectors_found < KEY_SIZE(insert)) {
+			SET_KEY_OFFSET(insert, KEY_OFFSET(insert) -
+				       (KEY_SIZE(insert) - sectors_found));
+			SET_KEY_SIZE(insert, sectors_found);
+		}
+	}
+
+	return false;
+}
+
+static bool btree_insert_key(struct btree *b, struct btree_op *op,
+			     struct bkey *k)
+{
+	struct bset *i = b->sets[b->nsets].data;
+	struct bkey *m, *prev;
+	const char *status = "insert";
+
+	BUG_ON(bkey_cmp(k, &b->key) > 0);
+	BUG_ON(b->level && !KEY_PTRS(k));
+	BUG_ON(!b->level && !KEY_OFFSET(k));
+
+	if (!b->level) {
+		struct btree_iter iter;
+		struct bkey search = KEY(KEY_INODE(k), KEY_START(k), 0);
+
+		/*
+		 * bset_search() returns the first key that is strictly greater
+		 * than the search key - but for back merging, we want to find
+		 * the first key that is greater than or equal to KEY_START(k) -
+		 * unless KEY_START(k) is 0.
+		 */
+		if (KEY_OFFSET(&search))
+			SET_KEY_OFFSET(&search, KEY_OFFSET(&search) - 1);
+
+		prev = NULL;
+		m = bch_btree_iter_init(b, &iter, &search);
+
+		if (fix_overlapping_extents(b, k, &iter, op))
+			return false;
+
+		while (m != end(i) &&
+		       bkey_cmp(k, &START_KEY(m)) > 0)
+			prev = m, m = bkey_next(m);
+
+		if (key_merging_disabled(b->c))
+			goto insert;
+
+		/* prev is in the tree, if we merge we're done */
+		status = "back merging";
+		if (prev &&
+		    bch_bkey_try_merge(b, prev, k))
+			goto merged;
+
+		status = "overwrote front";
+		if (m != end(i) &&
+		    KEY_PTRS(m) == KEY_PTRS(k) && !KEY_SIZE(m))
+			goto copy;
+
+		status = "front merge";
+		if (m != end(i) &&
+		    bch_bkey_try_merge(b, k, m))
+			goto copy;
+	} else
+		m = bch_bset_search(b, &b->sets[b->nsets], k);
+
+insert:	shift_keys(b, m, k);
+copy:	bkey_copy(m, k);
+merged:
+	bch_check_keys(b, "%s for %s at %s: %s", status,
+		       op_type(op), pbtree(b), pkey(k));
+	bch_check_key_order_msg(b, i, "%s for %s at %s: %s", status,
+				op_type(op), pbtree(b), pkey(k));
+
+	if (b->level && !KEY_OFFSET(k))
+		b->prio_blocked++;
+
+	pr_debug("%s for %s at %s: %s", status,
+		 op_type(op), pbtree(b), pkey(k));
+
+	return true;
+}
+
+bool bch_btree_insert_keys(struct btree *b, struct btree_op *op)
+{
+	bool ret = false;
+	struct bkey *k;
+	unsigned oldsize = bch_count_data(b);
+
+	while ((k = bch_keylist_pop(&op->keys))) {
+		bkey_put(b->c, k, b->level);
+		ret |= btree_insert_key(b, op, k);
+	}
+
+	BUG_ON(bch_count_data(b) < oldsize);
+	return ret;
+}
+
+bool bch_btree_insert_check_key(struct btree *b, struct btree_op *op,
+				   struct bio *bio)
+{
+	bool ret = false;
+	uint64_t btree_ptr = b->key.ptr[0];
+	unsigned long seq = b->seq;
+	BKEY_PADDED(k) tmp;
+
+	rw_unlock(false, b);
+	rw_lock(true, b, b->level);
+
+	if (b->key.ptr[0] != btree_ptr ||
+	    b->seq != seq + 1 ||
+	    should_split(b))
+		goto out;
+
+	op->replace = KEY(op->inode, bio_end(bio), bio_sectors(bio));
+
+	SET_KEY_PTRS(&op->replace, 1);
+	get_random_bytes(&op->replace.ptr[0], sizeof(uint64_t));
+
+	SET_PTR_DEV(&op->replace, 0, PTR_CHECK_DEV);
+
+	bkey_copy(&tmp.k, &op->replace);
+
+	BUG_ON(op->type != BTREE_INSERT);
+	BUG_ON(!btree_insert_key(b, op, &tmp.k));
+	bch_btree_write(b, false, NULL);
+	ret = true;
+out:
+	downgrade_write(&b->lock);
+	return ret;
+}
+
+static int btree_split(struct btree *b, struct btree_op *op)
+{
+	bool split, root = b == b->c->root;
+	struct btree *n1, *n2 = NULL, *n3 = NULL;
+	uint64_t start_time = local_clock();
+
+	if (b->level)
+		set_closure_blocking(&op->cl);
+
+	n1 = btree_node_alloc_replacement(b, &op->cl);
+	if (IS_ERR(n1))
+		goto err;
+
+	split = set_blocks(n1->sets[0].data, n1->c) > (btree_blocks(b) * 4) / 5;
+
+	pr_debug("%ssplitting at %s keys %i", split ? "" : "not ",
+		 pbtree(b), n1->sets[0].data->keys);
+
+	if (split) {
+		unsigned keys = 0;
+
+		n2 = bch_btree_node_alloc(b->c, b->level, &op->cl);
+		if (IS_ERR(n2))
+			goto err_free1;
+
+		if (root) {
+			n3 = bch_btree_node_alloc(b->c, b->level + 1, &op->cl);
+			if (IS_ERR(n3))
+				goto err_free2;
+		}
+
+		bch_btree_insert_keys(n1, op);
+
+		/* Has to be a linear search because we don't have an auxiliary
+		 * search tree yet
+		 */
+
+		while (keys < (n1->sets[0].data->keys * 3) / 5)
+			keys += bkey_u64s(node(n1->sets[0].data, keys));
+
+		bkey_copy_key(&n1->key, node(n1->sets[0].data, keys));
+		keys += bkey_u64s(node(n1->sets[0].data, keys));
+
+		n2->sets[0].data->keys = n1->sets[0].data->keys - keys;
+		n1->sets[0].data->keys = keys;
+
+		memcpy(n2->sets[0].data->start,
+		       end(n1->sets[0].data),
+		       n2->sets[0].data->keys * sizeof(uint64_t));
+
+		bkey_copy_key(&n2->key, &b->key);
+
+		bch_keylist_add(&op->keys, &n2->key);
+		bch_btree_write(n2, true, op);
+		rw_unlock(true, n2);
+	} else
+		bch_btree_insert_keys(n1, op);
+
+	bch_keylist_add(&op->keys, &n1->key);
+	bch_btree_write(n1, true, op);
+
+	if (n3) {
+		bkey_copy_key(&n3->key, &MAX_KEY);
+		bch_btree_insert_keys(n3, op);
+		bch_btree_write(n3, true, op);
+
+		closure_sync(&op->cl);
+		bch_btree_set_root(n3);
+		rw_unlock(true, n3);
+	} else if (root) {
+		op->keys.top = op->keys.bottom;
+		closure_sync(&op->cl);
+		bch_btree_set_root(n1);
+	} else {
+		unsigned i;
+
+		bkey_copy(op->keys.top, &b->key);
+		bkey_copy_key(op->keys.top, &ZERO_KEY);
+
+		for (i = 0; i < KEY_PTRS(&b->key); i++) {
+			uint8_t g = PTR_BUCKET(b->c, &b->key, i)->gen + 1;
+
+			SET_PTR_GEN(op->keys.top, i, g);
+		}
+
+		bch_keylist_push(&op->keys);
+		closure_sync(&op->cl);
+		atomic_inc(&b->c->prio_blocked);
+	}
+
+	rw_unlock(true, n1);
+	btree_node_free(b, op);
+
+	bch_time_stats_update(&b->c->btree_split_time, start_time);
+
+	return 0;
+err_free2:
+	__bkey_put(n2->c, &n2->key);
+	btree_node_free(n2, op);
+	rw_unlock(true, n2);
+err_free1:
+	__bkey_put(n1->c, &n1->key);
+	btree_node_free(n1, op);
+	rw_unlock(true, n1);
+err:
+	if (n3 == ERR_PTR(-EAGAIN) ||
+	    n2 == ERR_PTR(-EAGAIN) ||
+	    n1 == ERR_PTR(-EAGAIN))
+		return -EAGAIN;
+
+	pr_warn("couldn't split");
+	return -ENOMEM;
+}
+
+static int bch_btree_insert_recurse(struct btree *b, struct btree_op *op,
+				    struct keylist *stack_keys)
+{
+	if (b->level) {
+		int ret;
+		struct bkey *insert = op->keys.bottom;
+		struct bkey *k = bch_next_recurse_key(b, &START_KEY(insert));
+
+		if (!k) {
+			btree_bug(b, "no key to recurse on at level %i/%i",
+				  b->level, b->c->root->level);
+
+			op->keys.top = op->keys.bottom;
+			return -EIO;
+		}
+
+		if (bkey_cmp(insert, k) > 0) {
+			unsigned i;
+
+			if (op->type == BTREE_REPLACE) {
+				__bkey_put(b->c, insert);
+				op->keys.top = op->keys.bottom;
+				op->insert_collision = true;
+				return 0;
+			}
+
+			for (i = 0; i < KEY_PTRS(insert); i++)
+				atomic_inc(&PTR_BUCKET(b->c, insert, i)->pin);
+
+			bkey_copy(stack_keys->top, insert);
+
+			bch_cut_back(k, insert);
+			bch_cut_front(k, stack_keys->top);
+
+			bch_keylist_push(stack_keys);
+		}
+
+		ret = btree(insert_recurse, k, b, op, stack_keys);
+		if (ret)
+			return ret;
+	}
+
+	if (!bch_keylist_empty(&op->keys)) {
+		if (should_split(b)) {
+			if (op->lock <= b->c->root->level) {
+				BUG_ON(b->level);
+				op->lock = b->c->root->level + 1;
+				return -EINTR;
+			}
+			return btree_split(b, op);
+		}
+
+		BUG_ON(write_block(b) != b->sets[b->nsets].data);
+
+		if (bch_btree_insert_keys(b, op))
+			bch_btree_write(b, false, op);
+	}
+
+	return 0;
+}
+
+int bch_btree_insert(struct btree_op *op, struct cache_set *c)
+{
+	int ret = 0;
+	struct keylist stack_keys;
+
+	/*
+	 * Don't want to block with the btree locked unless we have to,
+	 * otherwise we get deadlocks with try_harder and between split/gc
+	 */
+	clear_closure_blocking(&op->cl);
+
+	BUG_ON(bch_keylist_empty(&op->keys));
+	bch_keylist_copy(&stack_keys, &op->keys);
+	bch_keylist_init(&op->keys);
+
+	while (!bch_keylist_empty(&stack_keys) ||
+	       !bch_keylist_empty(&op->keys)) {
+		if (bch_keylist_empty(&op->keys)) {
+			bch_keylist_add(&op->keys,
+					bch_keylist_pop(&stack_keys));
+			op->lock = 0;
+		}
+
+		ret = btree_root(insert_recurse, c, op, &stack_keys);
+
+		if (ret == -EAGAIN) {
+			ret = 0;
+			closure_sync(&op->cl);
+		} else if (ret) {
+			struct bkey *k;
+
+			pr_err("error %i trying to insert key for %s",
+			       ret, op_type(op));
+
+			while ((k = bch_keylist_pop(&stack_keys) ?:
+				    bch_keylist_pop(&op->keys)))
+				bkey_put(c, k, 0);
+		}
+	}
+
+	bch_keylist_free(&stack_keys);
+
+	if (op->journal)
+		atomic_dec_bug(op->journal);
+	op->journal = NULL;
+	return ret;
+}
+
+void bch_btree_set_root(struct btree *b)
+{
+	unsigned i;
+
+	BUG_ON(!b->written);
+
+	for (i = 0; i < KEY_PTRS(&b->key); i++)
+		BUG_ON(PTR_BUCKET(b->c, &b->key, i)->prio != BTREE_PRIO);
+
+	mutex_lock(&b->c->bucket_lock);
+	list_del_init(&b->list);
+	mutex_unlock(&b->c->bucket_lock);
+
+	b->c->root = b;
+	__bkey_put(b->c, &b->key);
+
+	bch_journal_meta(b->c, NULL);
+	pr_debug("%s for %pf", pbtree(b), __builtin_return_address(0));
+}
+
+/* Cache lookup */
+
+static int submit_partial_cache_miss(struct btree *b, struct btree_op *op,
+				     struct bkey *k)
+{
+	struct search *s = container_of(op, struct search, op);
+	struct bio *bio = &s->bio.bio;
+	int ret = 0;
+
+	while (!ret &&
+	       !op->lookup_done) {
+		unsigned sectors = INT_MAX;
+
+		if (KEY_INODE(k) == op->inode) {
+			if (KEY_START(k) <= bio->bi_sector)
+				break;
+
+			sectors = min_t(uint64_t, sectors,
+					KEY_START(k) - bio->bi_sector);
+		}
+
+		ret = s->d->cache_miss(b, s, bio, sectors);
+	}
+
+	return ret;
+}
+
+/*
+ * Read from a single key, handling the initial cache miss if the key starts in
+ * the middle of the bio
+ */
+static int submit_partial_cache_hit(struct btree *b, struct btree_op *op,
+				    struct bkey *k)
+{
+	struct search *s = container_of(op, struct search, op);
+	struct bio *bio = &s->bio.bio;
+	unsigned ptr;
+	struct bio *n;
+
+	int ret = submit_partial_cache_miss(b, op, k);
+	if (ret || op->lookup_done)
+		return ret;
+
+	/* XXX: figure out best pointer - for multiple cache devices */
+	ptr = 0;
+
+	PTR_BUCKET(b->c, k, ptr)->prio = INITIAL_PRIO;
+
+	while (!op->lookup_done &&
+	       KEY_INODE(k) == op->inode &&
+	       bio->bi_sector < KEY_OFFSET(k)) {
+		struct bkey *bio_key;
+		sector_t sector = PTR_OFFSET(k, ptr) +
+			(bio->bi_sector - KEY_START(k));
+		unsigned sectors = min_t(uint64_t, INT_MAX,
+					 KEY_OFFSET(k) - bio->bi_sector);
+
+		n = bch_bio_split(bio, sectors, GFP_NOIO, s->d->bio_split);
+		if (!n)
+			return -EAGAIN;
+
+		if (n == bio)
+			op->lookup_done = true;
+
+		bio_key = &container_of(n, struct bbio, bio)->key;
+
+		/*
+		 * The bucket we're reading from might be reused while our bio
+		 * is in flight, and we could then end up reading the wrong
+		 * data.
+		 *
+		 * We guard against this by checking (in cache_read_endio()) if
+		 * the pointer is stale again; if so, we treat it as an error
+		 * and reread from the backing device (but we don't pass that
+		 * error up anywhere).
+		 */
+
+		bch_bkey_copy_single_ptr(bio_key, k, ptr);
+		SET_PTR_OFFSET(bio_key, 0, sector);
+
+		n->bi_end_io	= bch_cache_read_endio;
+		n->bi_private	= &s->cl;
+
+		trace_bcache_cache_hit(n);
+		__bch_submit_bbio(n, b->c);
+	}
+
+	return 0;
+}
+
+int bch_btree_search_recurse(struct btree *b, struct btree_op *op)
+{
+	struct search *s = container_of(op, struct search, op);
+	struct bio *bio = &s->bio.bio;
+
+	int ret = 0;
+	struct bkey *k;
+	struct btree_iter iter;
+	bch_btree_iter_init(b, &iter, &KEY(op->inode, bio->bi_sector, 0));
+
+	pr_debug("at %s searching for %u:%llu", pbtree(b), op->inode,
+		 (uint64_t) bio->bi_sector);
+
+	do {
+		k = bch_btree_iter_next_filter(&iter, b, bch_ptr_bad);
+		if (!k) {
+			/*
+			 * b->key would be exactly what we want, except that
+			 * pointers to btree nodes have nonzero size - we
+			 * wouldn't go far enough
+			 */
+
+			ret = submit_partial_cache_miss(b, op,
+					&KEY(KEY_INODE(&b->key),
+					     KEY_OFFSET(&b->key), 0));
+			break;
+		}
+
+		ret = b->level
+			? btree(search_recurse, k, b, op)
+			: submit_partial_cache_hit(b, op, k);
+	} while (!ret &&
+		 !op->lookup_done);
+
+	return ret;
+}
+
+/* Keybuf code */
+
+static inline int keybuf_cmp(struct keybuf_key *l, struct keybuf_key *r)
+{
+	/* Overlapping keys compare equal */
+	if (bkey_cmp(&l->key, &START_KEY(&r->key)) <= 0)
+		return -1;
+	if (bkey_cmp(&START_KEY(&l->key), &r->key) >= 0)
+		return 1;
+	return 0;
+}
+
+static inline int keybuf_nonoverlapping_cmp(struct keybuf_key *l,
+					    struct keybuf_key *r)
+{
+	return clamp_t(int64_t, bkey_cmp(&l->key, &r->key), -1, 1);
+}
+
+static int bch_btree_refill_keybuf(struct btree *b, struct btree_op *op,
+				   struct keybuf *buf, struct bkey *end)
+{
+	struct btree_iter iter;
+	bch_btree_iter_init(b, &iter, &buf->last_scanned);
+
+	while (!array_freelist_empty(&buf->freelist)) {
+		struct bkey *k = bch_btree_iter_next_filter(&iter, b,
+							    bch_ptr_bad);
+
+		if (!b->level) {
+			if (!k) {
+				buf->last_scanned = b->key;
+				break;
+			}
+
+			buf->last_scanned = *k;
+			if (bkey_cmp(&buf->last_scanned, end) >= 0)
+				break;
+
+			if (buf->key_predicate(buf, k)) {
+				struct keybuf_key *w;
+
+				pr_debug("%s", pkey(k));
+
+				spin_lock(&buf->lock);
+
+				w = array_alloc(&buf->freelist);
+
+				w->private = NULL;
+				bkey_copy(&w->key, k);
+
+				if (RB_INSERT(&buf->keys, w, node, keybuf_cmp))
+					array_free(&buf->freelist, w);
+
+				spin_unlock(&buf->lock);
+			}
+		} else {
+			if (!k)
+				break;
+
+			btree(refill_keybuf, k, b, op, buf, end);
+			/*
+			 * Might get an error here, but can't really do anything
+			 * and it'll get logged elsewhere. Just read what we
+			 * can.
+			 */
+
+			if (bkey_cmp(&buf->last_scanned, end) >= 0)
+				break;
+
+			cond_resched();
+		}
+	}
+
+	return 0;
+}
+
+void bch_refill_keybuf(struct cache_set *c, struct keybuf *buf,
+			  struct bkey *end)
+{
+	struct bkey start = buf->last_scanned;
+	struct btree_op op;
+	bch_btree_op_init_stack(&op);
+
+	cond_resched();
+
+	btree_root(refill_keybuf, c, &op, buf, end);
+	closure_sync(&op.cl);
+
+	pr_debug("found %s keys from %llu:%llu to %llu:%llu",
+		 RB_EMPTY_ROOT(&buf->keys) ? "no" :
+		 array_freelist_empty(&buf->freelist) ? "some" : "a few",
+		 KEY_INODE(&start), KEY_OFFSET(&start),
+		 KEY_INODE(&buf->last_scanned), KEY_OFFSET(&buf->last_scanned));
+
+	spin_lock(&buf->lock);
+
+	if (!RB_EMPTY_ROOT(&buf->keys)) {
+		struct keybuf_key *w;
+		w = RB_FIRST(&buf->keys, struct keybuf_key, node);
+		buf->start	= START_KEY(&w->key);
+
+		w = RB_LAST(&buf->keys, struct keybuf_key, node);
+		buf->end	= w->key;
+	} else {
+		buf->start	= MAX_KEY;
+		buf->end	= MAX_KEY;
+	}
+
+	spin_unlock(&buf->lock);
+}
+
+static void __bch_keybuf_del(struct keybuf *buf, struct keybuf_key *w)
+{
+	rb_erase(&w->node, &buf->keys);
+	array_free(&buf->freelist, w);
+}
+
+void bch_keybuf_del(struct keybuf *buf, struct keybuf_key *w)
+{
+	spin_lock(&buf->lock);
+	__bch_keybuf_del(buf, w);
+	spin_unlock(&buf->lock);
+}
+
+bool bch_keybuf_check_overlapping(struct keybuf *buf, struct bkey *start,
+				  struct bkey *end)
+{
+	bool ret = false;
+	struct keybuf_key *p, *w, s;
+	s.key = *start;
+
+	if (bkey_cmp(end, &buf->start) <= 0 ||
+	    bkey_cmp(start, &buf->end) >= 0)
+		return false;
+
+	spin_lock(&buf->lock);
+	w = RB_GREATER(&buf->keys, s, node, keybuf_nonoverlapping_cmp);
+
+	while (w && bkey_cmp(&START_KEY(&w->key), end) < 0) {
+		p = w;
+		w = RB_NEXT(w, node);
+
+		if (p->private)
+			ret = true;
+		else
+			__bch_keybuf_del(buf, p);
+	}
+
+	spin_unlock(&buf->lock);
+	return ret;
+}
+
+struct keybuf_key *bch_keybuf_next(struct keybuf *buf)
+{
+	struct keybuf_key *w;
+	spin_lock(&buf->lock);
+
+	w = RB_FIRST(&buf->keys, struct keybuf_key, node);
+
+	while (w && w->private)
+		w = RB_NEXT(w, node);
+
+	if (w)
+		w->private = ERR_PTR(-EINTR);
+
+	spin_unlock(&buf->lock);
+	return w;
+}
+
+struct keybuf_key *bch_keybuf_next_rescan(struct cache_set *c,
+					     struct keybuf *buf,
+					     struct bkey *end)
+{
+	struct keybuf_key *ret;
+
+	while (1) {
+		ret = bch_keybuf_next(buf);
+		if (ret)
+			break;
+
+		if (bkey_cmp(&buf->last_scanned, end) >= 0) {
+			pr_debug("scan finished");
+			break;
+		}
+
+		bch_refill_keybuf(c, buf, end);
+	}
+
+	return ret;
+}
+
+void bch_keybuf_init(struct keybuf *buf, keybuf_pred_fn *fn)
+{
+	buf->key_predicate	= fn;
+	buf->last_scanned	= MAX_KEY;
+	buf->keys		= RB_ROOT;
+
+	spin_lock_init(&buf->lock);
+	array_allocator_init(&buf->freelist);
+}
+
+void bch_btree_exit(void)
+{
+	if (btree_io_wq)
+		destroy_workqueue(btree_io_wq);
+	if (bch_gc_wq)
+		destroy_workqueue(bch_gc_wq);
+}
+
+int __init bch_btree_init(void)
+{
+	if (!(bch_gc_wq = create_singlethread_workqueue("bch_btree_gc")) ||
+	    !(btree_io_wq = create_singlethread_workqueue("bch_btree_io")))
+		return -ENOMEM;
+
+	return 0;
+}

drivers/md/bcache/btree.h

···
+#ifndef _BCACHE_BTREE_H
+#define _BCACHE_BTREE_H
+
+/*
+ * THE BTREE:
+ *
+ * At a high level, bcache's btree is relatively standard b+ tree. All keys and
+ * pointers are in the leaves; interior nodes only have pointers to the child
+ * nodes.
+ *
+ * In the interior nodes, a struct bkey always points to a child btree node, and
+ * the key is the highest key in the child node - except that the highest key in
+ * an interior node is always MAX_KEY. The size field refers to the size on disk
+ * of the child node - this would allow us to have variable sized btree nodes
+ * (handy for keeping the depth of the btree 1 by expanding just the root).
+ *
+ * Btree nodes are themselves log structured, but this is hidden fairly
+ * thoroughly. Btree nodes on disk will in practice have extents that overlap
+ * (because they were written at different times), but in memory we never have
+ * overlapping extents - when we read in a btree node from disk, the first thing
+ * we do is resort all the sets of keys with a mergesort, and in the same pass
+ * we check for overlapping extents and adjust them appropriately.
+ *
+ * struct btree_op is a central interface to the btree code. It's used for
+ * specifying read vs. write locking, and the embedded closure is used for
+ * waiting on IO or reserve memory.
+ *
+ * BTREE CACHE:
+ *
+ * Btree nodes are cached in memory; traversing the btree might require reading
+ * in btree nodes which is handled mostly transparently.
+ *
+ * bch_btree_node_get() looks up a btree node in the cache and reads it in from
+ * disk if necessary. This function is almost never called directly though - the
+ * btree() macro is used to get a btree node, call some function on it, and
+ * unlock the node after the function returns.
+ *
+ * The root is special cased - it's taken out of the cache's lru (thus pinning
+ * it in memory), so we can find the root of the btree by just dereferencing a
+ * pointer instead of looking it up in the cache. This makes locking a bit
+ * tricky, since the root pointer is protected by the lock in the btree node it
+ * points to - the btree_root() macro handles this.
+ *
+ * In various places we must be able to allocate memory for multiple btree nodes
+ * in order to make forward progress. To do this we use the btree cache itself
+ * as a reserve; if __get_free_pages() fails, we'll find a node in the btree
+ * cache we can reuse. We can't allow more than one thread to be doing this at a
+ * time, so there's a lock, implemented by a pointer to the btree_op closure -
+ * this allows the btree_root() macro to implicitly release this lock.
+ *
+ * BTREE IO:
+ *
+ * Btree nodes never have to be explicitly read in; bch_btree_node_get() handles
+ * this.
+ *
+ * For writing, we have two btree_write structs embeddded in struct btree - one
+ * write in flight, and one being set up, and we toggle between them.
+ *
+ * Writing is done with a single function -  bch_btree_write() really serves two
+ * different purposes and should be broken up into two different functions. When
+ * passing now = false, it merely indicates that the node is now dirty - calling
+ * it ensures that the dirty keys will be written at some point in the future.
+ *
+ * When passing now = true, bch_btree_write() causes a write to happen
+ * "immediately" (if there was already a write in flight, it'll cause the write
+ * to happen as soon as the previous write completes). It returns immediately
+ * though - but it takes a refcount on the closure in struct btree_op you passed
+ * to it, so a closure_sync() later can be used to wait for the write to
+ * complete.
+ *
+ * This is handy because btree_split() and garbage collection can issue writes
+ * in parallel, reducing the amount of time they have to hold write locks.
+ *
+ * LOCKING:
+ *
+ * When traversing the btree, we may need write locks starting at some level -
+ * inserting a key into the btree will typically only require a write lock on
+ * the leaf node.
+ *
+ * This is specified with the lock field in struct btree_op; lock = 0 means we
+ * take write locks at level <= 0, i.e. only leaf nodes. bch_btree_node_get()
+ * checks this field and returns the node with the appropriate lock held.
+ *
+ * If, after traversing the btree, the insertion code discovers it has to split
+ * then it must restart from the root and take new locks - to do this it changes
+ * the lock field and returns -EINTR, which causes the btree_root() macro to
+ * loop.
+ *
+ * Handling cache misses require a different mechanism for upgrading to a write
+ * lock. We do cache lookups with only a read lock held, but if we get a cache
+ * miss and we wish to insert this data into the cache, we have to insert a
+ * placeholder key to detect races - otherwise, we could race with a write and
+ * overwrite the data that was just written to the cache with stale data from
+ * the backing device.
+ *
+ * For this we use a sequence number that write locks and unlocks increment - to
+ * insert the check key it unlocks the btree node and then takes a write lock,
+ * and fails if the sequence number doesn't match.
+ */
+
+#include "bset.h"
+#include "debug.h"
+
+struct btree_write {
+	struct closure		*owner;
+	atomic_t		*journal;
+
+	/* If btree_split() frees a btree node, it writes a new pointer to that
+	 * btree node indicating it was freed; it takes a refcount on
+	 * c->prio_blocked because we can't write the gens until the new
+	 * pointer is on disk. This allows btree_write_endio() to release the
+	 * refcount that btree_split() took.
+	 */
+	int			prio_blocked;
+};
+
+struct btree {
+	/* Hottest entries first */
+	struct hlist_node	hash;
+
+	/* Key/pointer for this btree node */
+	BKEY_PADDED(key);
+
+	/* Single bit - set when accessed, cleared by shrinker */
+	unsigned long		accessed;
+	unsigned long		seq;
+	struct rw_semaphore	lock;
+	struct cache_set	*c;
+
+	unsigned long		flags;
+	uint16_t		written;	/* would be nice to kill */
+	uint8_t			level;
+	uint8_t			nsets;
+	uint8_t			page_order;
+
+	/*
+	 * Set of sorted keys - the real btree node - plus a binary search tree
+	 *
+	 * sets[0] is special; set[0]->tree, set[0]->prev and set[0]->data point
+	 * to the memory we have allocated for this btree node. Additionally,
+	 * set[0]->data points to the entire btree node as it exists on disk.
+	 */
+	struct bset_tree	sets[MAX_BSETS];
+
+	/* Used to refcount bio splits, also protects b->bio */
+	struct closure_with_waitlist	io;
+
+	/* Gets transferred to w->prio_blocked - see the comment there */
+	int			prio_blocked;
+
+	struct list_head	list;
+	struct delayed_work	work;
+
+	uint64_t		io_start_time;
+	struct btree_write	writes[2];
+	struct bio		*bio;
+};
+
+#define BTREE_FLAG(flag)						\
+static inline bool btree_node_ ## flag(struct btree *b)			\
+{	return test_bit(BTREE_NODE_ ## flag, &b->flags); }		\
+									\
+static inline void set_btree_node_ ## flag(struct btree *b)		\
+{	set_bit(BTREE_NODE_ ## flag, &b->flags); }			\
+
+enum btree_flags {
+	BTREE_NODE_read_done,
+	BTREE_NODE_io_error,
+	BTREE_NODE_dirty,
+	BTREE_NODE_write_idx,
+};
+
+BTREE_FLAG(read_done);
+BTREE_FLAG(io_error);
+BTREE_FLAG(dirty);
+BTREE_FLAG(write_idx);
+
+static inline struct btree_write *btree_current_write(struct btree *b)
+{
+	return b->writes + btree_node_write_idx(b);
+}
+
+static inline struct btree_write *btree_prev_write(struct btree *b)
+{
+	return b->writes + (btree_node_write_idx(b) ^ 1);
+}
+
+static inline unsigned bset_offset(struct btree *b, struct bset *i)
+{
+	return (((size_t) i) - ((size_t) b->sets->data)) >> 9;
+}
+
+static inline struct bset *write_block(struct btree *b)
+{
+	return ((void *) b->sets[0].data) + b->written * block_bytes(b->c);
+}
+
+static inline bool bset_written(struct btree *b, struct bset_tree *t)
+{
+	return t->data < write_block(b);
+}
+
+static inline bool bkey_written(struct btree *b, struct bkey *k)
+{
+	return k < write_block(b)->start;
+}
+
+static inline void set_gc_sectors(struct cache_set *c)
+{
+	atomic_set(&c->sectors_to_gc, c->sb.bucket_size * c->nbuckets / 8);
+}
+
+static inline bool bch_ptr_invalid(struct btree *b, const struct bkey *k)
+{
+	return __bch_ptr_invalid(b->c, b->level, k);
+}
+
+static inline struct bkey *bch_btree_iter_init(struct btree *b,
+					       struct btree_iter *iter,
+					       struct bkey *search)
+{
+	return __bch_btree_iter_init(b, iter, search, b->sets);
+}
+
+/* Looping macros */
+
+#define for_each_cached_btree(b, c, iter)				\
+	for (iter = 0;							\
+	     iter < ARRAY_SIZE((c)->bucket_hash);			\
+	     iter++)							\
+		hlist_for_each_entry_rcu((b), (c)->bucket_hash + iter, hash)
+
+#define for_each_key_filter(b, k, iter, filter)				\
+	for (bch_btree_iter_init((b), (iter), NULL);			\
+	     ((k) = bch_btree_iter_next_filter((iter), b, filter));)
+
+#define for_each_key(b, k, iter)					\
+	for (bch_btree_iter_init((b), (iter), NULL);			\
+	     ((k) = bch_btree_iter_next(iter));)
+
+/* Recursing down the btree */
+
+struct btree_op {
+	struct closure		cl;
+	struct cache_set	*c;
+
+	/* Journal entry we have a refcount on */
+	atomic_t		*journal;
+
+	/* Bio to be inserted into the cache */
+	struct bio		*cache_bio;
+
+	unsigned		inode;
+
+	uint16_t		write_prio;
+
+	/* Btree level at which we start taking write locks */
+	short			lock;
+
+	/* Btree insertion type */
+	enum {
+		BTREE_INSERT,
+		BTREE_REPLACE
+	} type:8;
+
+	unsigned		csum:1;
+	unsigned		skip:1;
+	unsigned		flush_journal:1;
+
+	unsigned		insert_data_done:1;
+	unsigned		lookup_done:1;
+	unsigned		insert_collision:1;
+
+	/* Anything after this point won't get zeroed in do_bio_hook() */
+
+	/* Keys to be inserted */
+	struct keylist		keys;
+	BKEY_PADDED(replace);
+};
+
+void bch_btree_op_init_stack(struct btree_op *);
+
+static inline void rw_lock(bool w, struct btree *b, int level)
+{
+	w ? down_write_nested(&b->lock, level + 1)
+	  : down_read_nested(&b->lock, level + 1);
+	if (w)
+		b->seq++;
+}
+
+static inline void rw_unlock(bool w, struct btree *b)
+{
+#ifdef CONFIG_BCACHE_EDEBUG
+	unsigned i;
+
+	if (w &&
+	    b->key.ptr[0] &&
+	    btree_node_read_done(b))
+		for (i = 0; i <= b->nsets; i++)
+			bch_check_key_order(b, b->sets[i].data);
+#endif
+
+	if (w)
+		b->seq++;
+	(w ? up_write : up_read)(&b->lock);
+}
+
+#define insert_lock(s, b)	((b)->level <= (s)->lock)
+
+/*
+ * These macros are for recursing down the btree - they handle the details of
+ * locking and looking up nodes in the cache for you. They're best treated as
+ * mere syntax when reading code that uses them.
+ *
+ * op->lock determines whether we take a read or a write lock at a given depth.
+ * If you've got a read lock and find that you need a write lock (i.e. you're
+ * going to have to split), set op->lock and return -EINTR; btree_root() will
+ * call you again and you'll have the correct lock.
+ */
+
+/**
+ * btree - recurse down the btree on a specified key
+ * @fn:		function to call, which will be passed the child node
+ * @key:	key to recurse on
+ * @b:		parent btree node
+ * @op:		pointer to struct btree_op
+ */
+#define btree(fn, key, b, op, ...)					\
+({									\
+	int _r, l = (b)->level - 1;					\
+	bool _w = l <= (op)->lock;					\
+	struct btree *_b = bch_btree_node_get((b)->c, key, l, op);	\
+	if (!IS_ERR(_b)) {						\
+		_r = bch_btree_ ## fn(_b, op, ##__VA_ARGS__);		\
+		rw_unlock(_w, _b);					\
+	} else								\
+		_r = PTR_ERR(_b);					\
+	_r;								\
+})
+
+/**
+ * btree_root - call a function on the root of the btree
+ * @fn:		function to call, which will be passed the child node
+ * @c:		cache set
+ * @op:		pointer to struct btree_op
+ */
+#define btree_root(fn, c, op, ...)					\
+({									\
+	int _r = -EINTR;						\
+	do {								\
+		struct btree *_b = (c)->root;				\
+		bool _w = insert_lock(op, _b);				\
+		rw_lock(_w, _b, _b->level);				\
+		if (_b == (c)->root &&					\
+		    _w == insert_lock(op, _b))				\
+			_r = bch_btree_ ## fn(_b, op, ##__VA_ARGS__);	\
+		rw_unlock(_w, _b);					\
+		bch_cannibalize_unlock(c, &(op)->cl);		\
+	} while (_r == -EINTR);						\
+									\
+	_r;								\
+})
+
+static inline bool should_split(struct btree *b)
+{
+	struct bset *i = write_block(b);
+	return b->written >= btree_blocks(b) ||
+		(i->seq == b->sets[0].data->seq &&
+		 b->written + __set_blocks(i, i->keys + 15, b->c)
+		 > btree_blocks(b));
+}
+
+void bch_btree_read_done(struct closure *);
+void bch_btree_read(struct btree *);
+void bch_btree_write(struct btree *b, bool now, struct btree_op *op);
+
+void bch_cannibalize_unlock(struct cache_set *, struct closure *);
+void bch_btree_set_root(struct btree *);
+struct btree *bch_btree_node_alloc(struct cache_set *, int, struct closure *);
+struct btree *bch_btree_node_get(struct cache_set *, struct bkey *,
+				int, struct btree_op *);
+
+bool bch_btree_insert_keys(struct btree *, struct btree_op *);
+bool bch_btree_insert_check_key(struct btree *, struct btree_op *,
+				   struct bio *);
+int bch_btree_insert(struct btree_op *, struct cache_set *);
+
+int bch_btree_search_recurse(struct btree *, struct btree_op *);
+
+void bch_queue_gc(struct cache_set *);
+size_t bch_btree_gc_finish(struct cache_set *);
+void bch_moving_gc(struct closure *);
+int bch_btree_check(struct cache_set *, struct btree_op *);
+uint8_t __bch_btree_mark_key(struct cache_set *, int, struct bkey *);
+
+void bch_keybuf_init(struct keybuf *, keybuf_pred_fn *);
+void bch_refill_keybuf(struct cache_set *, struct keybuf *, struct bkey *);
+bool bch_keybuf_check_overlapping(struct keybuf *, struct bkey *,
+				  struct bkey *);
+void bch_keybuf_del(struct keybuf *, struct keybuf_key *);
+struct keybuf_key *bch_keybuf_next(struct keybuf *);
+struct keybuf_key *bch_keybuf_next_rescan(struct cache_set *,
+					  struct keybuf *, struct bkey *);
+
+#endif

drivers/md/bcache/closure.c

···
+/*
+ * Asynchronous refcounty things
+ *
+ * Copyright 2010, 2011 Kent Overstreet <[e-mail]>
+ * Copyright 2012 Google, Inc.
+ */
+
+#include <linux/debugfs.h>
+#include <linux/module.h>
+#include <linux/seq_file.h>
+
+#include "closure.h"
+
+void closure_queue(struct closure *cl)
+{
+	struct workqueue_struct *wq = cl->wq;
+	if (wq) {
+		INIT_WORK(&cl->work, cl->work.func);
+		BUG_ON(!queue_work(wq, &cl->work));
+	} else
+		cl->fn(cl);
+}
+EXPORT_SYMBOL_GPL(closure_queue);
+
+#define CL_FIELD(type, field)					\
+	case TYPE_ ## type:					\
+	return &container_of(cl, struct type, cl)->field
+
+static struct closure_waitlist *closure_waitlist(struct closure *cl)
+{
+	switch (cl->type) {
+		CL_FIELD(closure_with_waitlist, wait);
+		CL_FIELD(closure_with_waitlist_and_timer, wait);
+	default:
+		return NULL;
+	}
+}
+
+static struct timer_list *closure_timer(struct closure *cl)
+{
+	switch (cl->type) {
+		CL_FIELD(closure_with_timer, timer);
+		CL_FIELD(closure_with_waitlist_and_timer, timer);
+	default:
+		return NULL;
+	}
+}
+
+static inline void closure_put_after_sub(struct closure *cl, int flags)
+{
+	int r = flags & CLOSURE_REMAINING_MASK;
+
+	BUG_ON(flags & CLOSURE_GUARD_MASK);
+	BUG_ON(!r && (flags & ~(CLOSURE_DESTRUCTOR|CLOSURE_BLOCKING)));
+
+	/* Must deliver precisely one wakeup */
+	if (r == 1 && (flags & CLOSURE_SLEEPING))
+		wake_up_process(cl->task);
+
+	if (!r) {
+		if (cl->fn && !(flags & CLOSURE_DESTRUCTOR)) {
+			/* CLOSURE_BLOCKING might be set - clear it */
+			atomic_set(&cl->remaining,
+				   CLOSURE_REMAINING_INITIALIZER);
+			closure_queue(cl);
+		} else {
+			struct closure *parent = cl->parent;
+			struct closure_waitlist *wait = closure_waitlist(cl);
+
+			closure_debug_destroy(cl);
+
+			atomic_set(&cl->remaining, -1);
+
+			if (wait)
+				closure_wake_up(wait);
+
+			if (cl->fn)
+				cl->fn(cl);
+
+			if (parent)
+				closure_put(parent);
+		}
+	}
+}
+
+/* For clearing flags with the same atomic op as a put */
+void closure_sub(struct closure *cl, int v)
+{
+	closure_put_after_sub(cl, atomic_sub_return(v, &cl->remaining));
+}
+EXPORT_SYMBOL_GPL(closure_sub);
+
+void closure_put(struct closure *cl)
+{
+	closure_put_after_sub(cl, atomic_dec_return(&cl->remaining));
+}
+EXPORT_SYMBOL_GPL(closure_put);
+
+static void set_waiting(struct closure *cl, unsigned long f)
+{
+#ifdef CONFIG_BCACHE_CLOSURES_DEBUG
+	cl->waiting_on = f;
+#endif
+}
+
+void __closure_wake_up(struct closure_waitlist *wait_list)
+{
+	struct llist_node *list;
+	struct closure *cl;
+	struct llist_node *reverse = NULL;
+
+	list = llist_del_all(&wait_list->list);
+
+	/* We first reverse the list to preserve FIFO ordering and fairness */
+
+	while (list) {
+		struct llist_node *t = list;
+		list = llist_next(list);
+
+		t->next = reverse;
+		reverse = t;
+	}
+
+	/* Then do the wakeups */
+
+	while (reverse) {
+		cl = container_of(reverse, struct closure, list);
+		reverse = llist_next(reverse);
+
+		set_waiting(cl, 0);
+		closure_sub(cl, CLOSURE_WAITING + 1);
+	}
+}
+EXPORT_SYMBOL_GPL(__closure_wake_up);
+
+bool closure_wait(struct closure_waitlist *list, struct closure *cl)
+{
+	if (atomic_read(&cl->remaining) & CLOSURE_WAITING)
+		return false;
+
+	set_waiting(cl, _RET_IP_);
+	atomic_add(CLOSURE_WAITING + 1, &cl->remaining);
+	llist_add(&cl->list, &list->list);
+
+	return true;
+}
+EXPORT_SYMBOL_GPL(closure_wait);
+
+/**
+ * closure_sync() - sleep until a closure a closure has nothing left to wait on
+ *
+ * Sleeps until the refcount hits 1 - the thread that's running the closure owns
+ * the last refcount.
+ */
+void closure_sync(struct closure *cl)
+{
+	while (1) {
+		__closure_start_sleep(cl);
+		closure_set_ret_ip(cl);
+
+		if ((atomic_read(&cl->remaining) &
+		     CLOSURE_REMAINING_MASK) == 1)
+			break;
+
+		schedule();
+	}
+
+	__closure_end_sleep(cl);
+}
+EXPORT_SYMBOL_GPL(closure_sync);
+
+/**
+ * closure_trylock() - try to acquire the closure, without waiting
+ * @cl:		closure to lock
+ *
+ * Returns true if the closure was succesfully locked.
+ */
+bool closure_trylock(struct closure *cl, struct closure *parent)
+{
+	if (atomic_cmpxchg(&cl->remaining, -1,
+			   CLOSURE_REMAINING_INITIALIZER) != -1)
+		return false;
+
+	closure_set_ret_ip(cl);
+
+	smp_mb();
+	cl->parent = parent;
+	if (parent)
+		closure_get(parent);
+
+	closure_debug_create(cl);
+	return true;
+}
+EXPORT_SYMBOL_GPL(closure_trylock);
+
+void __closure_lock(struct closure *cl, struct closure *parent,
+		    struct closure_waitlist *wait_list)
+{
+	struct closure wait;
+	closure_init_stack(&wait);
+
+	while (1) {
+		if (closure_trylock(cl, parent))
+			return;
+
+		closure_wait_event_sync(wait_list, &wait,
+					atomic_read(&cl->remaining) == -1);
+	}
+}
+EXPORT_SYMBOL_GPL(__closure_lock);
+
+static void closure_delay_timer_fn(unsigned long data)
+{
+	struct closure *cl = (struct closure *) data;
+	closure_sub(cl, CLOSURE_TIMER + 1);
+}
+
+void do_closure_timer_init(struct closure *cl)
+{
+	struct timer_list *timer = closure_timer(cl);
+
+	init_timer(timer);
+	timer->data	= (unsigned long) cl;
+	timer->function = closure_delay_timer_fn;
+}
+EXPORT_SYMBOL_GPL(do_closure_timer_init);
+
+bool __closure_delay(struct closure *cl, unsigned long delay,
+		     struct timer_list *timer)
+{
+	if (atomic_read(&cl->remaining) & CLOSURE_TIMER)
+		return false;
+
+	BUG_ON(timer_pending(timer));
+
+	timer->expires	= jiffies + delay;
+
+	atomic_add(CLOSURE_TIMER + 1, &cl->remaining);
+	add_timer(timer);
+	return true;
+}
+EXPORT_SYMBOL_GPL(__closure_delay);
+
+void __closure_flush(struct closure *cl, struct timer_list *timer)
+{
+	if (del_timer(timer))
+		closure_sub(cl, CLOSURE_TIMER + 1);
+}
+EXPORT_SYMBOL_GPL(__closure_flush);
+
+void __closure_flush_sync(struct closure *cl, struct timer_list *timer)
+{
+	if (del_timer_sync(timer))
+		closure_sub(cl, CLOSURE_TIMER + 1);
+}
+EXPORT_SYMBOL_GPL(__closure_flush_sync);
+
+#ifdef CONFIG_BCACHE_CLOSURES_DEBUG
+
+static LIST_HEAD(closure_list);
+static DEFINE_SPINLOCK(closure_list_lock);
+
+void closure_debug_create(struct closure *cl)
+{
+	unsigned long flags;
+
+	BUG_ON(cl->magic == CLOSURE_MAGIC_ALIVE);
+	cl->magic = CLOSURE_MAGIC_ALIVE;
+
+	spin_lock_irqsave(&closure_list_lock, flags);
+	list_add(&cl->all, &closure_list);
+	spin_unlock_irqrestore(&closure_list_lock, flags);
+}
+EXPORT_SYMBOL_GPL(closure_debug_create);
+
+void closure_debug_destroy(struct closure *cl)
+{
+	unsigned long flags;
+
+	BUG_ON(cl->magic != CLOSURE_MAGIC_ALIVE);
+	cl->magic = CLOSURE_MAGIC_DEAD;
+
+	spin_lock_irqsave(&closure_list_lock, flags);
+	list_del(&cl->all);
+	spin_unlock_irqrestore(&closure_list_lock, flags);
+}
+EXPORT_SYMBOL_GPL(closure_debug_destroy);
+
+static struct dentry *debug;
+
+#define work_data_bits(work) ((unsigned long *)(&(work)->data))
+
+static int debug_seq_show(struct seq_file *f, void *data)
+{
+	struct closure *cl;
+	spin_lock_irq(&closure_list_lock);
+
+	list_for_each_entry(cl, &closure_list, all) {
+		int r = atomic_read(&cl->remaining);
+
+		seq_printf(f, "%p: %pF -> %pf p %p r %i ",
+			   cl, (void *) cl->ip, cl->fn, cl->parent,
+			   r & CLOSURE_REMAINING_MASK);
+
+		seq_printf(f, "%s%s%s%s%s%s\n",
+			   test_bit(WORK_STRUCT_PENDING,
+				    work_data_bits(&cl->work)) ? "Q" : "",
+			   r & CLOSURE_RUNNING	? "R" : "",
+			   r & CLOSURE_BLOCKING	? "B" : "",
+			   r & CLOSURE_STACK	? "S" : "",
+			   r & CLOSURE_SLEEPING	? "Sl" : "",
+			   r & CLOSURE_TIMER	? "T" : "");
+
+		if (r & CLOSURE_WAITING)
+			seq_printf(f, " W %pF\n",
+				   (void *) cl->waiting_on);
+
+		seq_printf(f, "\n");
+	}
+
+	spin_unlock_irq(&closure_list_lock);
+	return 0;
+}
+
+static int debug_seq_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, debug_seq_show, NULL);
+}
+
+static const struct file_operations debug_ops = {
+	.owner		= THIS_MODULE,
+	.open		= debug_seq_open,
+	.read		= seq_read,
+	.release	= single_release
+};
+
+void __init closure_debug_init(void)
+{
+	debug = debugfs_create_file("closures", 0400, NULL, NULL, &debug_ops);
+}
+
+#endif
+
+MODULE_AUTHOR("Kent Overstreet <[e-mail]>");
+MODULE_LICENSE("GPL");

drivers/md/bcache/closure.h

···
+#ifndef _LINUX_CLOSURE_H
+#define _LINUX_CLOSURE_H
+
+#include <linux/llist.h>
+#include <linux/sched.h>
+#include <linux/workqueue.h>
+
+/*
+ * Closure is perhaps the most overused and abused term in computer science, but
+ * since I've been unable to come up with anything better you're stuck with it
+ * again.
+ *
+ * What are closures?
+ *
+ * They embed a refcount. The basic idea is they count "things that are in
+ * progress" - in flight bios, some other thread that's doing something else -
+ * anything you might want to wait on.
+ *
+ * The refcount may be manipulated with closure_get() and closure_put().
+ * closure_put() is where many of the interesting things happen, when it causes
+ * the refcount to go to 0.
+ *
+ * Closures can be used to wait on things both synchronously and asynchronously,
+ * and synchronous and asynchronous use can be mixed without restriction. To
+ * wait synchronously, use closure_sync() - you will sleep until your closure's
+ * refcount hits 1.
+ *
+ * To wait asynchronously, use
+ *   continue_at(cl, next_function, workqueue);
+ *
+ * passing it, as you might expect, the function to run when nothing is pending
+ * and the workqueue to run that function out of.
+ *
+ * continue_at() also, critically, is a macro that returns the calling function.
+ * There's good reason for this.
+ *
+ * To use safely closures asynchronously, they must always have a refcount while
+ * they are running owned by the thread that is running them. Otherwise, suppose
+ * you submit some bios and wish to have a function run when they all complete:
+ *
+ * foo_endio(struct bio *bio, int error)
+ * {
+ *	closure_put(cl);
+ * }
+ *
+ * closure_init(cl);
+ *
+ * do_stuff();
+ * closure_get(cl);
+ * bio1->bi_endio = foo_endio;
+ * bio_submit(bio1);
+ *
+ * do_more_stuff();
+ * closure_get(cl);
+ * bio2->bi_endio = foo_endio;
+ * bio_submit(bio2);
+ *
+ * continue_at(cl, complete_some_read, system_wq);
+ *
+ * If closure's refcount started at 0, complete_some_read() could run before the
+ * second bio was submitted - which is almost always not what you want! More
+ * importantly, it wouldn't be possible to say whether the original thread or
+ * complete_some_read()'s thread owned the closure - and whatever state it was
+ * associated with!
+ *
+ * So, closure_init() initializes a closure's refcount to 1 - and when a
+ * closure_fn is run, the refcount will be reset to 1 first.
+ *
+ * Then, the rule is - if you got the refcount with closure_get(), release it
+ * with closure_put() (i.e, in a bio->bi_endio function). If you have a refcount
+ * on a closure because you called closure_init() or you were run out of a
+ * closure - _always_ use continue_at(). Doing so consistently will help
+ * eliminate an entire class of particularly pernicious races.
+ *
+ * For a closure to wait on an arbitrary event, we need to introduce waitlists:
+ *
+ * struct closure_waitlist list;
+ * closure_wait_event(list, cl, condition);
+ * closure_wake_up(wait_list);
+ *
+ * These work analagously to wait_event() and wake_up() - except that instead of
+ * operating on the current thread (for wait_event()) and lists of threads, they
+ * operate on an explicit closure and lists of closures.
+ *
+ * Because it's a closure we can now wait either synchronously or
+ * asynchronously. closure_wait_event() returns the current value of the
+ * condition, and if it returned false continue_at() or closure_sync() can be
+ * used to wait for it to become true.
+ *
+ * It's useful for waiting on things when you can't sleep in the context in
+ * which you must check the condition (perhaps a spinlock held, or you might be
+ * beneath generic_make_request() - in which case you can't sleep on IO).
+ *
+ * closure_wait_event() will wait either synchronously or asynchronously,
+ * depending on whether the closure is in blocking mode or not. You can pick a
+ * mode explicitly with closure_wait_event_sync() and
+ * closure_wait_event_async(), which do just what you might expect.
+ *
+ * Lastly, you might have a wait list dedicated to a specific event, and have no
+ * need for specifying the condition - you just want to wait until someone runs
+ * closure_wake_up() on the appropriate wait list. In that case, just use
+ * closure_wait(). It will return either true or false, depending on whether the
+ * closure was already on a wait list or not - a closure can only be on one wait
+ * list at a time.
+ *
+ * Parents:
+ *
+ * closure_init() takes two arguments - it takes the closure to initialize, and
+ * a (possibly null) parent.
+ *
+ * If parent is non null, the new closure will have a refcount for its lifetime;
+ * a closure is considered to be "finished" when its refcount hits 0 and the
+ * function to run is null. Hence
+ *
+ * continue_at(cl, NULL, NULL);
+ *
+ * returns up the (spaghetti) stack of closures, precisely like normal return
+ * returns up the C stack. continue_at() with non null fn is better thought of
+ * as doing a tail call.
+ *
+ * All this implies that a closure should typically be embedded in a particular
+ * struct (which its refcount will normally control the lifetime of), and that
+ * struct can very much be thought of as a stack frame.
+ *
+ * Locking:
+ *
+ * Closures are based on work items but they can be thought of as more like
+ * threads - in that like threads and unlike work items they have a well
+ * defined lifetime; they are created (with closure_init()) and eventually
+ * complete after a continue_at(cl, NULL, NULL).
+ *
+ * Suppose you've got some larger structure with a closure embedded in it that's
+ * used for periodically doing garbage collection. You only want one garbage
+ * collection happening at a time, so the natural thing to do is protect it with
+ * a lock. However, it's difficult to use a lock protecting a closure correctly
+ * because the unlock should come after the last continue_to() (additionally, if
+ * you're using the closure asynchronously a mutex won't work since a mutex has
+ * to be unlocked by the same process that locked it).
+ *
+ * So to make it less error prone and more efficient, we also have the ability
+ * to use closures as locks:
+ *
+ * closure_init_unlocked();
+ * closure_trylock();
+ *
+ * That's all we need for trylock() - the last closure_put() implicitly unlocks
+ * it for you.  But for closure_lock(), we also need a wait list:
+ *
+ * struct closure_with_waitlist frobnicator_cl;
+ *
+ * closure_init_unlocked(&frobnicator_cl);
+ * closure_lock(&frobnicator_cl);
+ *
+ * A closure_with_waitlist embeds a closure and a wait list - much like struct
+ * delayed_work embeds a work item and a timer_list. The important thing is, use
+ * it exactly like you would a regular closure and closure_put() will magically
+ * handle everything for you.
+ *
+ * We've got closures that embed timers, too. They're called, appropriately
+ * enough:
+ * struct closure_with_timer;
+ *
+ * This gives you access to closure_delay(). It takes a refcount for a specified
+ * number of jiffies - you could then call closure_sync() (for a slightly
+ * convoluted version of msleep()) or continue_at() - which gives you the same
+ * effect as using a delayed work item, except you can reuse the work_struct
+ * already embedded in struct closure.
+ *
+ * Lastly, there's struct closure_with_waitlist_and_timer. It does what you
+ * probably expect, if you happen to need the features of both. (You don't
+ * really want to know how all this is implemented, but if I've done my job
+ * right you shouldn't have to care).
+ */
+
+struct closure;
+typedef void (closure_fn) (struct closure *);
+
+struct closure_waitlist {
+	struct llist_head	list;
+};
+
+enum closure_type {
+	TYPE_closure				= 0,
+	TYPE_closure_with_waitlist		= 1,
+	TYPE_closure_with_timer			= 2,
+	TYPE_closure_with_waitlist_and_timer	= 3,
+	MAX_CLOSURE_TYPE			= 3,
+};
+
+enum closure_state {
+	/*
+	 * CLOSURE_BLOCKING: Causes closure_wait_event() to block, instead of
+	 * waiting asynchronously
+	 *
+	 * CLOSURE_WAITING: Set iff the closure is on a waitlist. Must be set by
+	 * the thread that owns the closure, and cleared by the thread that's
+	 * waking up the closure.
+	 *
+	 * CLOSURE_SLEEPING: Must be set before a thread uses a closure to sleep
+	 * - indicates that cl->task is valid and closure_put() may wake it up.
+	 * Only set or cleared by the thread that owns the closure.
+	 *
+	 * CLOSURE_TIMER: Analagous to CLOSURE_WAITING, indicates that a closure
+	 * has an outstanding timer. Must be set by the thread that owns the
+	 * closure, and cleared by the timer function when the timer goes off.
+	 *
+	 * The rest are for debugging and don't affect behaviour:
+	 *
+	 * CLOSURE_RUNNING: Set when a closure is running (i.e. by
+	 * closure_init() and when closure_put() runs then next function), and
+	 * must be cleared before remaining hits 0. Primarily to help guard
+	 * against incorrect usage and accidentally transferring references.
+	 * continue_at() and closure_return() clear it for you, if you're doing
+	 * something unusual you can use closure_set_dead() which also helps
+	 * annotate where references are being transferred.
+	 *
+	 * CLOSURE_STACK: Sanity check - remaining should never hit 0 on a
+	 * closure with this flag set
+	 */
+
+	CLOSURE_BITS_START	= (1 << 19),
+	CLOSURE_DESTRUCTOR	= (1 << 19),
+	CLOSURE_BLOCKING	= (1 << 21),
+	CLOSURE_WAITING		= (1 << 23),
+	CLOSURE_SLEEPING	= (1 << 25),
+	CLOSURE_TIMER		= (1 << 27),
+	CLOSURE_RUNNING		= (1 << 29),
+	CLOSURE_STACK		= (1 << 31),
+};
+
+#define CLOSURE_GUARD_MASK					\
+	((CLOSURE_DESTRUCTOR|CLOSURE_BLOCKING|CLOSURE_WAITING|	\
+	  CLOSURE_SLEEPING|CLOSURE_TIMER|CLOSURE_RUNNING|CLOSURE_STACK) << 1)
+
+#define CLOSURE_REMAINING_MASK		(CLOSURE_BITS_START - 1)
+#define CLOSURE_REMAINING_INITIALIZER	(1|CLOSURE_RUNNING)
+
+struct closure {
+	union {
+		struct {
+			struct workqueue_struct *wq;
+			struct task_struct	*task;
+			struct llist_node	list;
+			closure_fn		*fn;
+		};
+		struct work_struct	work;
+	};
+
+	struct closure		*parent;
+
+	atomic_t		remaining;
+
+	enum closure_type	type;
+
+#ifdef CONFIG_BCACHE_CLOSURES_DEBUG
+#define CLOSURE_MAGIC_DEAD	0xc054dead
+#define CLOSURE_MAGIC_ALIVE	0xc054a11e
+
+	unsigned		magic;
+	struct list_head	all;
+	unsigned long		ip;
+	unsigned long		waiting_on;
+#endif
+};
+
+struct closure_with_waitlist {
+	struct closure		cl;
+	struct closure_waitlist	wait;
+};
+
+struct closure_with_timer {
+	struct closure		cl;
+	struct timer_list	timer;
+};
+
+struct closure_with_waitlist_and_timer {
+	struct closure		cl;
+	struct closure_waitlist	wait;
+	struct timer_list	timer;
+};
+
+extern unsigned invalid_closure_type(void);
+
+#define __CLOSURE_TYPE(cl, _t)						\
+	  __builtin_types_compatible_p(typeof(cl), struct _t)		\
+		? TYPE_ ## _t :						\
+
+#define __closure_type(cl)						\
+(									\
+	__CLOSURE_TYPE(cl, closure)					\
+	__CLOSURE_TYPE(cl, closure_with_waitlist)			\
+	__CLOSURE_TYPE(cl, closure_with_timer)				\
+	__CLOSURE_TYPE(cl, closure_with_waitlist_and_timer)		\
+	invalid_closure_type()						\
+)
+
+void closure_sub(struct closure *cl, int v);
+void closure_put(struct closure *cl);
+void closure_queue(struct closure *cl);
+void __closure_wake_up(struct closure_waitlist *list);
+bool closure_wait(struct closure_waitlist *list, struct closure *cl);
+void closure_sync(struct closure *cl);
+
+bool closure_trylock(struct closure *cl, struct closure *parent);
+void __closure_lock(struct closure *cl, struct closure *parent,
+		    struct closure_waitlist *wait_list);
+
+void do_closure_timer_init(struct closure *cl);
+bool __closure_delay(struct closure *cl, unsigned long delay,
+		     struct timer_list *timer);
+void __closure_flush(struct closure *cl, struct timer_list *timer);
+void __closure_flush_sync(struct closure *cl, struct timer_list *timer);
+
+#ifdef CONFIG_BCACHE_CLOSURES_DEBUG
+
+void closure_debug_init(void);
+void closure_debug_create(struct closure *cl);
+void closure_debug_destroy(struct closure *cl);
+
+#else
+
+static inline void closure_debug_init(void) {}
+static inline void closure_debug_create(struct closure *cl) {}
+static inline void closure_debug_destroy(struct closure *cl) {}
+
+#endif
+
+static inline void closure_set_ip(struct closure *cl)
+{
+#ifdef CONFIG_BCACHE_CLOSURES_DEBUG
+	cl->ip = _THIS_IP_;
+#endif
+}
+
+static inline void closure_set_ret_ip(struct closure *cl)
+{
+#ifdef CONFIG_BCACHE_CLOSURES_DEBUG
+	cl->ip = _RET_IP_;
+#endif
+}
+
+static inline void closure_get(struct closure *cl)
+{
+#ifdef CONFIG_BCACHE_CLOSURES_DEBUG
+	BUG_ON((atomic_inc_return(&cl->remaining) &
+		CLOSURE_REMAINING_MASK) <= 1);
+#else
+	atomic_inc(&cl->remaining);
+#endif
+}
+
+static inline void closure_set_stopped(struct closure *cl)
+{
+	atomic_sub(CLOSURE_RUNNING, &cl->remaining);
+}
+
+static inline bool closure_is_stopped(struct closure *cl)
+{
+	return !(atomic_read(&cl->remaining) & CLOSURE_RUNNING);
+}
+
+static inline bool closure_is_unlocked(struct closure *cl)
+{
+	return atomic_read(&cl->remaining) == -1;
+}
+
+static inline void do_closure_init(struct closure *cl, struct closure *parent,
+				   bool running)
+{
+	switch (cl->type) {
+	case TYPE_closure_with_timer:
+	case TYPE_closure_with_waitlist_and_timer:
+		do_closure_timer_init(cl);
+	default:
+		break;
+	}
+
+	cl->parent = parent;
+	if (parent)
+		closure_get(parent);
+
+	if (running) {
+		closure_debug_create(cl);
+		atomic_set(&cl->remaining, CLOSURE_REMAINING_INITIALIZER);
+	} else
+		atomic_set(&cl->remaining, -1);
+
+	closure_set_ip(cl);
+}
+
+/*
+ * Hack to get at the embedded closure if there is one, by doing an unsafe cast:
+ * the result of __closure_type() is thrown away, it's used merely for type
+ * checking.
+ */
+#define __to_internal_closure(cl)				\
+({								\
+	BUILD_BUG_ON(__closure_type(*cl) > MAX_CLOSURE_TYPE);	\
+	(struct closure *) cl;					\
+})
+
+#define closure_init_type(cl, parent, running)			\
+do {								\
+	struct closure *_cl = __to_internal_closure(cl);	\
+	_cl->type = __closure_type(*(cl));			\
+	do_closure_init(_cl, parent, running);			\
+} while (0)
+
+/**
+ * __closure_init() - Initialize a closure, skipping the memset()
+ *
+ * May be used instead of closure_init() when memory has already been zeroed.
+ */
+#define __closure_init(cl, parent)				\
+	closure_init_type(cl, parent, true)
+
+/**
+ * closure_init() - Initialize a closure, setting the refcount to 1
+ * @cl:		closure to initialize
+ * @parent:	parent of the new closure. cl will take a refcount on it for its
+ *		lifetime; may be NULL.
+ */
+#define closure_init(cl, parent)				\
+do {								\
+	memset((cl), 0, sizeof(*(cl)));				\
+	__closure_init(cl, parent);				\
+} while (0)
+
+static inline void closure_init_stack(struct closure *cl)
+{
+	memset(cl, 0, sizeof(struct closure));
+	atomic_set(&cl->remaining, CLOSURE_REMAINING_INITIALIZER|
+		   CLOSURE_BLOCKING|CLOSURE_STACK);
+}
+
+/**
+ * closure_init_unlocked() - Initialize a closure but leave it unlocked.
+ * @cl:		closure to initialize
+ *
+ * For when the closure will be used as a lock. The closure may not be used
+ * until after a closure_lock() or closure_trylock().
+ */
+#define closure_init_unlocked(cl)				\
+do {								\
+	memset((cl), 0, sizeof(*(cl)));				\
+	closure_init_type(cl, NULL, false);			\
+} while (0)
+
+/**
+ * closure_lock() - lock and initialize a closure.
+ * @cl:		the closure to lock
+ * @parent:	the new parent for this closure
+ *
+ * The closure must be of one of the types that has a waitlist (otherwise we
+ * wouldn't be able to sleep on contention).
+ *
+ * @parent has exactly the same meaning as in closure_init(); if non null, the
+ * closure will take a reference on @parent which will be released when it is
+ * unlocked.
+ */
+#define closure_lock(cl, parent)				\
+	__closure_lock(__to_internal_closure(cl), parent, &(cl)->wait)
+
+/**
+ * closure_delay() - delay some number of jiffies
+ * @cl:		the closure that will sleep
+ * @delay:	the delay in jiffies
+ *
+ * Takes a refcount on @cl which will be released after @delay jiffies; this may
+ * be used to have a function run after a delay with continue_at(), or
+ * closure_sync() may be used for a convoluted version of msleep().
+ */
+#define closure_delay(cl, delay)			\
+	__closure_delay(__to_internal_closure(cl), delay, &(cl)->timer)
+
+#define closure_flush(cl)				\
+	__closure_flush(__to_internal_closure(cl), &(cl)->timer)
+
+#define closure_flush_sync(cl)				\
+	__closure_flush_sync(__to_internal_closure(cl), &(cl)->timer)
+
+static inline void __closure_end_sleep(struct closure *cl)
+{
+	__set_current_state(TASK_RUNNING);
+
+	if (atomic_read(&cl->remaining) & CLOSURE_SLEEPING)
+		atomic_sub(CLOSURE_SLEEPING, &cl->remaining);
+}
+
+static inline void __closure_start_sleep(struct closure *cl)
+{
+	closure_set_ip(cl);
+	cl->task = current;
+	set_current_state(TASK_UNINTERRUPTIBLE);
+
+	if (!(atomic_read(&cl->remaining) & CLOSURE_SLEEPING))
+		atomic_add(CLOSURE_SLEEPING, &cl->remaining);
+}
+
+/**
+ * closure_blocking() - returns true if the closure is in blocking mode.
+ *
+ * If a closure is in blocking mode, closure_wait_event() will sleep until the
+ * condition is true instead of waiting asynchronously.
+ */
+static inline bool closure_blocking(struct closure *cl)
+{
+	return atomic_read(&cl->remaining) & CLOSURE_BLOCKING;
+}
+
+/**
+ * set_closure_blocking() - put a closure in blocking mode.
+ *
+ * If a closure is in blocking mode, closure_wait_event() will sleep until the
+ * condition is true instead of waiting asynchronously.
+ *
+ * Not thread safe - can only be called by the thread running the closure.
+ */
+static inline void set_closure_blocking(struct closure *cl)
+{
+	if (!closure_blocking(cl))
+		atomic_add(CLOSURE_BLOCKING, &cl->remaining);
+}
+
+/*
+ * Not thread safe - can only be called by the thread running the closure.
+ */
+static inline void clear_closure_blocking(struct closure *cl)
+{
+	if (closure_blocking(cl))
+		atomic_sub(CLOSURE_BLOCKING, &cl->remaining);
+}
+
+/**
+ * closure_wake_up() - wake up all closures on a wait list.
+ */
+static inline void closure_wake_up(struct closure_waitlist *list)
+{
+	smp_mb();
+	__closure_wake_up(list);
+}
+
+/*
+ * Wait on an event, synchronously or asynchronously - analogous to wait_event()
+ * but for closures.
+ *
+ * The loop is oddly structured so as to avoid a race; we must check the
+ * condition again after we've added ourself to the waitlist. We know if we were
+ * already on the waitlist because closure_wait() returns false; thus, we only
+ * schedule or break if closure_wait() returns false. If it returns true, we
+ * just loop again - rechecking the condition.
+ *
+ * The __closure_wake_up() is necessary because we may race with the event
+ * becoming true; i.e. we see event false -> wait -> recheck condition, but the
+ * thread that made the event true may have called closure_wake_up() before we
+ * added ourself to the wait list.
+ *
+ * We have to call closure_sync() at the end instead of just
+ * __closure_end_sleep() because a different thread might've called
+ * closure_wake_up() before us and gotten preempted before they dropped the
+ * refcount on our closure. If this was a stack allocated closure, that would be
+ * bad.
+ */
+#define __closure_wait_event(list, cl, condition, _block)		\
+({									\
+	bool block = _block;						\
+	typeof(condition) ret;						\
+									\
+	while (1) {							\
+		ret = (condition);					\
+		if (ret) {						\
+			__closure_wake_up(list);			\
+			if (block)					\
+				closure_sync(cl);			\
+									\
+			break;						\
+		}							\
+									\
+		if (block)						\
+			__closure_start_sleep(cl);			\
+									\
+		if (!closure_wait(list, cl)) {				\
+			if (!block)					\
+				break;					\
+									\
+			schedule();					\
+		}							\
+	}								\
+									\
+	ret;								\
+})
+
+/**
+ * closure_wait_event() - wait on a condition, synchronously or asynchronously.
+ * @list:	the wait list to wait on
+ * @cl:		the closure that is doing the waiting
+ * @condition:	a C expression for the event to wait for
+ *
+ * If the closure is in blocking mode, sleeps until the @condition evaluates to
+ * true - exactly like wait_event().
+ *
+ * If the closure is not in blocking mode, waits asynchronously; if the
+ * condition is currently false the @cl is put onto @list and returns. @list
+ * owns a refcount on @cl; closure_sync() or continue_at() may be used later to
+ * wait for another thread to wake up @list, which drops the refcount on @cl.
+ *
+ * Returns the value of @condition; @cl will be on @list iff @condition was
+ * false.
+ *
+ * closure_wake_up(@list) must be called after changing any variable that could
+ * cause @condition to become true.
+ */
+#define closure_wait_event(list, cl, condition)				\
+	__closure_wait_event(list, cl, condition, closure_blocking(cl))
+
+#define closure_wait_event_async(list, cl, condition)			\
+	__closure_wait_event(list, cl, condition, false)
+
+#define closure_wait_event_sync(list, cl, condition)			\
+	__closure_wait_event(list, cl, condition, true)
+
+static inline void set_closure_fn(struct closure *cl, closure_fn *fn,
+				  struct workqueue_struct *wq)
+{
+	BUG_ON(object_is_on_stack(cl));
+	closure_set_ip(cl);
+	cl->fn = fn;
+	cl->wq = wq;
+	/* between atomic_dec() in closure_put() */
+	smp_mb__before_atomic_dec();
+}
+
+#define continue_at(_cl, _fn, _wq)					\
+do {									\
+	set_closure_fn(_cl, _fn, _wq);					\
+	closure_sub(_cl, CLOSURE_RUNNING + 1);				\
+	return;								\
+} while (0)
+
+#define closure_return(_cl)	continue_at((_cl), NULL, NULL)
+
+#define continue_at_nobarrier(_cl, _fn, _wq)				\
+do {									\
+	set_closure_fn(_cl, _fn, _wq);					\
+	closure_queue(cl);						\
+	return;								\
+} while (0)
+
+#define closure_return_with_destructor(_cl, _destructor)		\
+do {									\
+	set_closure_fn(_cl, _destructor, NULL);				\
+	closure_sub(_cl, CLOSURE_RUNNING - CLOSURE_DESTRUCTOR + 1);	\
+	return;								\
+} while (0)
+
+static inline void closure_call(struct closure *cl, closure_fn fn,
+				struct workqueue_struct *wq,
+				struct closure *parent)
+{
+	closure_init(cl, parent);
+	continue_at_nobarrier(cl, fn, wq);
+}
+
+static inline void closure_trylock_call(struct closure *cl, closure_fn fn,
+					struct workqueue_struct *wq,
+					struct closure *parent)
+{
+	if (closure_trylock(cl, parent))
+		continue_at_nobarrier(cl, fn, wq);
+}
+
+#endif /* _LINUX_CLOSURE_H */

drivers/md/bcache/debug.c

···
+/*
+ * Assorted bcache debug code
+ *
+ * Copyright 2010, 2011 Kent Overstreet <[e-mail]>
+ * Copyright 2012 Google, Inc.
+ */
+
+#include "bcache.h"
+#include "btree.h"
+#include "debug.h"
+#include "request.h"
+
+#include <linux/console.h>
+#include <linux/debugfs.h>
+#include <linux/module.h>
+#include <linux/random.h>
+#include <linux/seq_file.h>
+
+static struct dentry *debug;
+
+const char *bch_ptr_status(struct cache_set *c, const struct bkey *k)
+{
+	unsigned i;
+
+	for (i = 0; i < KEY_PTRS(k); i++)
+		if (ptr_available(c, k, i)) {
+			struct cache *ca = PTR_CACHE(c, k, i);
+			size_t bucket = PTR_BUCKET_NR(c, k, i);
+			size_t r = bucket_remainder(c, PTR_OFFSET(k, i));
+
+			if (KEY_SIZE(k) + r > c->sb.bucket_size)
+				return "bad, length too big";
+			if (bucket <  ca->sb.first_bucket)
+				return "bad, short offset";
+			if (bucket >= ca->sb.nbuckets)
+				return "bad, offset past end of device";
+			if (ptr_stale(c, k, i))
+				return "stale";
+		}
+
+	if (!bkey_cmp(k, &ZERO_KEY))
+		return "bad, null key";
+	if (!KEY_PTRS(k))
+		return "bad, no pointers";
+	if (!KEY_SIZE(k))
+		return "zeroed key";
+	return "";
+}
+
+struct keyprint_hack bch_pkey(const struct bkey *k)
+{
+	unsigned i = 0;
+	struct keyprint_hack r;
+	char *out = r.s, *end = r.s + KEYHACK_SIZE;
+
+#define p(...)	(out += scnprintf(out, end - out, __VA_ARGS__))
+
+	p("%llu:%llu len %llu -> [", KEY_INODE(k), KEY_OFFSET(k), KEY_SIZE(k));
+
+	if (KEY_PTRS(k))
+		while (1) {
+			p("%llu:%llu gen %llu",
+			  PTR_DEV(k, i), PTR_OFFSET(k, i), PTR_GEN(k, i));
+
+			if (++i == KEY_PTRS(k))
+				break;
+
+			p(", ");
+		}
+
+	p("]");
+
+	if (KEY_DIRTY(k))
+		p(" dirty");
+	if (KEY_CSUM(k))
+		p(" cs%llu %llx", KEY_CSUM(k), k->ptr[1]);
+#undef p
+	return r;
+}
+
+struct keyprint_hack bch_pbtree(const struct btree *b)
+{
+	struct keyprint_hack r;
+
+	snprintf(r.s, 40, "%zu level %i/%i", PTR_BUCKET_NR(b->c, &b->key, 0),
+		 b->level, b->c->root ? b->c->root->level : -1);
+	return r;
+}
+
+#if defined(CONFIG_BCACHE_DEBUG) || defined(CONFIG_BCACHE_EDEBUG)
+
+static bool skipped_backwards(struct btree *b, struct bkey *k)
+{
+	return bkey_cmp(k, (!b->level)
+			? &START_KEY(bkey_next(k))
+			: bkey_next(k)) > 0;
+}
+
+static void dump_bset(struct btree *b, struct bset *i)
+{
+	struct bkey *k;
+	unsigned j;
+
+	for (k = i->start; k < end(i); k = bkey_next(k)) {
+		printk(KERN_ERR "block %zu key %zi/%u: %s", index(i, b),
+		       (uint64_t *) k - i->d, i->keys, pkey(k));
+
+		for (j = 0; j < KEY_PTRS(k); j++) {
+			size_t n = PTR_BUCKET_NR(b->c, k, j);
+			printk(" bucket %zu", n);
+
+			if (n >= b->c->sb.first_bucket && n < b->c->sb.nbuckets)
+				printk(" prio %i",
+				       PTR_BUCKET(b->c, k, j)->prio);
+		}
+
+		printk(" %s\n", bch_ptr_status(b->c, k));
+
+		if (bkey_next(k) < end(i) &&
+		    skipped_backwards(b, k))
+			printk(KERN_ERR "Key skipped backwards\n");
+	}
+}
+
+#endif
+
+#ifdef CONFIG_BCACHE_DEBUG
+
+void bch_btree_verify(struct btree *b, struct bset *new)
+{
+	struct btree *v = b->c->verify_data;
+	struct closure cl;
+	closure_init_stack(&cl);
+
+	if (!b->c->verify)
+		return;
+
+	closure_wait_event(&b->io.wait, &cl,
+			   atomic_read(&b->io.cl.remaining) == -1);
+
+	mutex_lock(&b->c->verify_lock);
+
+	bkey_copy(&v->key, &b->key);
+	v->written = 0;
+	v->level = b->level;
+
+	bch_btree_read(v);
+	closure_wait_event(&v->io.wait, &cl,
+			   atomic_read(&b->io.cl.remaining) == -1);
+
+	if (new->keys != v->sets[0].data->keys ||
+	    memcmp(new->start,
+		   v->sets[0].data->start,
+		   (void *) end(new) - (void *) new->start)) {
+		unsigned i, j;
+
+		console_lock();
+
+		printk(KERN_ERR "*** original memory node:\n");
+		for (i = 0; i <= b->nsets; i++)
+			dump_bset(b, b->sets[i].data);
+
+		printk(KERN_ERR "*** sorted memory node:\n");
+		dump_bset(b, new);
+
+		printk(KERN_ERR "*** on disk node:\n");
+		dump_bset(v, v->sets[0].data);
+
+		for (j = 0; j < new->keys; j++)
+			if (new->d[j] != v->sets[0].data->d[j])
+				break;
+
+		console_unlock();
+		panic("verify failed at %u\n", j);
+	}
+
+	mutex_unlock(&b->c->verify_lock);
+}
+
+static void data_verify_endio(struct bio *bio, int error)
+{
+	struct closure *cl = bio->bi_private;
+	closure_put(cl);
+}
+
+void bch_data_verify(struct search *s)
+{
+	char name[BDEVNAME_SIZE];
+	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
+	struct closure *cl = &s->cl;
+	struct bio *check;
+	struct bio_vec *bv;
+	int i;
+
+	if (!s->unaligned_bvec)
+		bio_for_each_segment(bv, s->orig_bio, i)
+			bv->bv_offset = 0, bv->bv_len = PAGE_SIZE;
+
+	check = bio_clone(s->orig_bio, GFP_NOIO);
+	if (!check)
+		return;
+
+	if (bch_bio_alloc_pages(check, GFP_NOIO))
+		goto out_put;
+
+	check->bi_rw		= READ_SYNC;
+	check->bi_private	= cl;
+	check->bi_end_io	= data_verify_endio;
+
+	closure_bio_submit(check, cl, &dc->disk);
+	closure_sync(cl);
+
+	bio_for_each_segment(bv, s->orig_bio, i) {
+		void *p1 = kmap(bv->bv_page);
+		void *p2 = kmap(check->bi_io_vec[i].bv_page);
+
+		if (memcmp(p1 + bv->bv_offset,
+			   p2 + bv->bv_offset,
+			   bv->bv_len))
+			printk(KERN_ERR
+			       "bcache (%s): verify failed at sector %llu\n",
+			       bdevname(dc->bdev, name),
+			       (uint64_t) s->orig_bio->bi_sector);
+
+		kunmap(bv->bv_page);
+		kunmap(check->bi_io_vec[i].bv_page);
+	}
+
+	__bio_for_each_segment(bv, check, i, 0)
+		__free_page(bv->bv_page);
+out_put:
+	bio_put(check);
+}
+
+#endif
+
+#ifdef CONFIG_BCACHE_EDEBUG
+
+unsigned bch_count_data(struct btree *b)
+{
+	unsigned ret = 0;
+	struct btree_iter iter;
+	struct bkey *k;
+
+	if (!b->level)
+		for_each_key(b, k, &iter)
+			ret += KEY_SIZE(k);
+	return ret;
+}
+
+static void vdump_bucket_and_panic(struct btree *b, const char *fmt,
+				   va_list args)
+{
+	unsigned i;
+
+	console_lock();
+
+	for (i = 0; i <= b->nsets; i++)
+		dump_bset(b, b->sets[i].data);
+
+	vprintk(fmt, args);
+
+	console_unlock();
+
+	panic("at %s\n", pbtree(b));
+}
+
+void bch_check_key_order_msg(struct btree *b, struct bset *i,
+			     const char *fmt, ...)
+{
+	struct bkey *k;
+
+	if (!i->keys)
+		return;
+
+	for (k = i->start; bkey_next(k) < end(i); k = bkey_next(k))
+		if (skipped_backwards(b, k)) {
+			va_list args;
+			va_start(args, fmt);
+
+			vdump_bucket_and_panic(b, fmt, args);
+			va_end(args);
+		}
+}
+
+void bch_check_keys(struct btree *b, const char *fmt, ...)
+{
+	va_list args;
+	struct bkey *k, *p = NULL;
+	struct btree_iter iter;
+
+	if (b->level)
+		return;
+
+	for_each_key(b, k, &iter) {
+		if (p && bkey_cmp(&START_KEY(p), &START_KEY(k)) > 0) {
+			printk(KERN_ERR "Keys out of order:\n");
+			goto bug;
+		}
+
+		if (bch_ptr_invalid(b, k))
+			continue;
+
+		if (p && bkey_cmp(p, &START_KEY(k)) > 0) {
+			printk(KERN_ERR "Overlapping keys:\n");
+			goto bug;
+		}
+		p = k;
+	}
+	return;
+bug:
+	va_start(args, fmt);
+	vdump_bucket_and_panic(b, fmt, args);
+	va_end(args);
+}
+
+#endif
+
+#ifdef CONFIG_DEBUG_FS
+
+/* XXX: cache set refcounting */
+
+struct dump_iterator {
+	char			buf[PAGE_SIZE];
+	size_t			bytes;
+	struct cache_set	*c;
+	struct keybuf		keys;
+};
+
+static bool dump_pred(struct keybuf *buf, struct bkey *k)
+{
+	return true;
+}
+
+static ssize_t bch_dump_read(struct file *file, char __user *buf,
+			     size_t size, loff_t *ppos)
+{
+	struct dump_iterator *i = file->private_data;
+	ssize_t ret = 0;
+
+	while (size) {
+		struct keybuf_key *w;
+		unsigned bytes = min(i->bytes, size);
+
+		int err = copy_to_user(buf, i->buf, bytes);
+		if (err)
+			return err;
+
+		ret	 += bytes;
+		buf	 += bytes;
+		size	 -= bytes;
+		i->bytes -= bytes;
+		memmove(i->buf, i->buf + bytes, i->bytes);
+
+		if (i->bytes)
+			break;
+
+		w = bch_keybuf_next_rescan(i->c, &i->keys, &MAX_KEY);
+		if (!w)
+			break;
+
+		i->bytes = snprintf(i->buf, PAGE_SIZE, "%s\n", pkey(&w->key));
+		bch_keybuf_del(&i->keys, w);
+	}
+
+	return ret;
+}
+
+static int bch_dump_open(struct inode *inode, struct file *file)
+{
+	struct cache_set *c = inode->i_private;
+	struct dump_iterator *i;
+
+	i = kzalloc(sizeof(struct dump_iterator), GFP_KERNEL);
+	if (!i)
+		return -ENOMEM;
+
+	file->private_data = i;
+	i->c = c;
+	bch_keybuf_init(&i->keys, dump_pred);
+	i->keys.last_scanned = KEY(0, 0, 0);
+
+	return 0;
+}
+
+static int bch_dump_release(struct inode *inode, struct file *file)
+{
+	kfree(file->private_data);
+	return 0;
+}
+
+static const struct file_operations cache_set_debug_ops = {
+	.owner		= THIS_MODULE,
+	.open		= bch_dump_open,
+	.read		= bch_dump_read,
+	.release	= bch_dump_release
+};
+
+void bch_debug_init_cache_set(struct cache_set *c)
+{
+	if (!IS_ERR_OR_NULL(debug)) {
+		char name[50];
+		snprintf(name, 50, "bcache-%pU", c->sb.set_uuid);
+
+		c->debug = debugfs_create_file(name, 0400, debug, c,
+					       &cache_set_debug_ops);
+	}
+}
+
+#endif
+
+/* Fuzz tester has rotted: */
+#if 0
+
+static ssize_t btree_fuzz(struct kobject *k, struct kobj_attribute *a,
+			  const char *buffer, size_t size)
+{
+	void dump(struct btree *b)
+	{
+		struct bset *i;
+
+		for (i = b->sets[0].data;
+		     index(i, b) < btree_blocks(b) &&
+		     i->seq == b->sets[0].data->seq;
+		     i = ((void *) i) + set_blocks(i, b->c) * block_bytes(b->c))
+			dump_bset(b, i);
+	}
+
+	struct cache_sb *sb;
+	struct cache_set *c;
+	struct btree *all[3], *b, *fill, *orig;
+	int j;
+
+	struct btree_op op;
+	bch_btree_op_init_stack(&op);
+
+	sb = kzalloc(sizeof(struct cache_sb), GFP_KERNEL);
+	if (!sb)
+		return -ENOMEM;
+
+	sb->bucket_size = 128;
+	sb->block_size = 4;
+
+	c = bch_cache_set_alloc(sb);
+	if (!c)
+		return -ENOMEM;
+
+	for (j = 0; j < 3; j++) {
+		BUG_ON(list_empty(&c->btree_cache));
+		all[j] = list_first_entry(&c->btree_cache, struct btree, list);
+		list_del_init(&all[j]->list);
+
+		all[j]->key = KEY(0, 0, c->sb.bucket_size);
+		bkey_copy_key(&all[j]->key, &MAX_KEY);
+	}
+
+	b = all[0];
+	fill = all[1];
+	orig = all[2];
+
+	while (1) {
+		for (j = 0; j < 3; j++)
+			all[j]->written = all[j]->nsets = 0;
+
+		bch_bset_init_next(b);
+
+		while (1) {
+			struct bset *i = write_block(b);
+			struct bkey *k = op.keys.top;
+			unsigned rand;
+
+			bkey_init(k);
+			rand = get_random_int();
+
+			op.type = rand & 1
+				? BTREE_INSERT
+				: BTREE_REPLACE;
+			rand >>= 1;
+
+			SET_KEY_SIZE(k, bucket_remainder(c, rand));
+			rand >>= c->bucket_bits;
+			rand &= 1024 * 512 - 1;
+			rand += c->sb.bucket_size;
+			SET_KEY_OFFSET(k, rand);
+#if 0
+			SET_KEY_PTRS(k, 1);
+#endif
+			bch_keylist_push(&op.keys);
+			bch_btree_insert_keys(b, &op);
+
+			if (should_split(b) ||
+			    set_blocks(i, b->c) !=
+			    __set_blocks(i, i->keys + 15, b->c)) {
+				i->csum = csum_set(i);
+
+				memcpy(write_block(fill),
+				       i, set_bytes(i));
+
+				b->written += set_blocks(i, b->c);
+				fill->written = b->written;
+				if (b->written == btree_blocks(b))
+					break;
+
+				bch_btree_sort_lazy(b);
+				bch_bset_init_next(b);
+			}
+		}
+
+		memcpy(orig->sets[0].data,
+		       fill->sets[0].data,
+		       btree_bytes(c));
+
+		bch_btree_sort(b);
+		fill->written = 0;
+		bch_btree_read_done(&fill->io.cl);
+
+		if (b->sets[0].data->keys != fill->sets[0].data->keys ||
+		    memcmp(b->sets[0].data->start,
+			   fill->sets[0].data->start,
+			   b->sets[0].data->keys * sizeof(uint64_t))) {
+			struct bset *i = b->sets[0].data;
+			struct bkey *k, *l;
+
+			for (k = i->start,
+			     l = fill->sets[0].data->start;
+			     k < end(i);
+			     k = bkey_next(k), l = bkey_next(l))
+				if (bkey_cmp(k, l) ||
+				    KEY_SIZE(k) != KEY_SIZE(l))
+					pr_err("key %zi differs: %s != %s",
+					       (uint64_t *) k - i->d,
+					       pkey(k), pkey(l));
+
+			for (j = 0; j < 3; j++) {
+				pr_err("**** Set %i ****", j);
+				dump(all[j]);
+			}
+			panic("\n");
+		}
+
+		pr_info("fuzz complete: %i keys", b->sets[0].data->keys);
+	}
+}
+
+kobj_attribute_write(fuzz, btree_fuzz);
+#endif
+
+void bch_debug_exit(void)
+{
+	if (!IS_ERR_OR_NULL(debug))
+		debugfs_remove_recursive(debug);
+}
+
+int __init bch_debug_init(struct kobject *kobj)
+{
+	int ret = 0;
+#if 0
+	ret = sysfs_create_file(kobj, &ksysfs_fuzz.attr);
+	if (ret)
+		return ret;
+#endif
+
+	debug = debugfs_create_dir("bcache", NULL);
+	return ret;
+}

drivers/md/bcache/debug.h

···
+#ifndef _BCACHE_DEBUG_H
+#define _BCACHE_DEBUG_H
+
+/* Btree/bkey debug printing */
+
+#define KEYHACK_SIZE 80
+struct keyprint_hack {
+	char s[KEYHACK_SIZE];
+};
+
+struct keyprint_hack bch_pkey(const struct bkey *k);
+struct keyprint_hack bch_pbtree(const struct btree *b);
+#define pkey(k)		(&bch_pkey(k).s[0])
+#define pbtree(b)	(&bch_pbtree(b).s[0])
+
+#ifdef CONFIG_BCACHE_EDEBUG
+
+unsigned bch_count_data(struct btree *);
+void bch_check_key_order_msg(struct btree *, struct bset *, const char *, ...);
+void bch_check_keys(struct btree *, const char *, ...);
+
+#define bch_check_key_order(b, i)			\
+	bch_check_key_order_msg(b, i, "keys out of order")
+#define EBUG_ON(cond)		BUG_ON(cond)
+
+#else /* EDEBUG */
+
+#define bch_count_data(b)				0
+#define bch_check_key_order(b, i)			do {} while (0)
+#define bch_check_key_order_msg(b, i, ...)		do {} while (0)
+#define bch_check_keys(b, ...)				do {} while (0)
+#define EBUG_ON(cond)					do {} while (0)
+
+#endif
+
+#ifdef CONFIG_BCACHE_DEBUG
+
+void bch_btree_verify(struct btree *, struct bset *);
+void bch_data_verify(struct search *);
+
+#else /* DEBUG */
+
+static inline void bch_btree_verify(struct btree *b, struct bset *i) {}
+static inline void bch_data_verify(struct search *s) {};
+
+#endif
+
+#ifdef CONFIG_DEBUG_FS
+void bch_debug_init_cache_set(struct cache_set *);
+#else
+static inline void bch_debug_init_cache_set(struct cache_set *c) {}
+#endif
+
+#endif

drivers/md/bcache/io.c

···
+/*
+ * Some low level IO code, and hacks for various block layer limitations
+ *
+ * Copyright 2010, 2011 Kent Overstreet <[e-mail]>
+ * Copyright 2012 Google, Inc.
+ */
+
+#include "bcache.h"
+#include "bset.h"
+#include "debug.h"
+
+static void bch_bi_idx_hack_endio(struct bio *bio, int error)
+{
+	struct bio *p = bio->bi_private;
+
+	bio_endio(p, error);
+	bio_put(bio);
+}
+
+static void bch_generic_make_request_hack(struct bio *bio)
+{
+	if (bio->bi_idx) {
+		struct bio *clone = bio_alloc(GFP_NOIO, bio_segments(bio));
+
+		memcpy(clone->bi_io_vec,
+		       bio_iovec(bio),
+		       bio_segments(bio) * sizeof(struct bio_vec));
+
+		clone->bi_sector	= bio->bi_sector;
+		clone->bi_bdev		= bio->bi_bdev;
+		clone->bi_rw		= bio->bi_rw;
+		clone->bi_vcnt		= bio_segments(bio);
+		clone->bi_size		= bio->bi_size;
+
+		clone->bi_private	= bio;
+		clone->bi_end_io	= bch_bi_idx_hack_endio;
+
+		bio = clone;
+	}
+
+	/*
+	 * Hack, since drivers that clone bios clone up to bi_max_vecs, but our
+	 * bios might have had more than that (before we split them per device
+	 * limitations).
+	 *
+	 * To be taken out once immutable bvec stuff is in.
+	 */
+	bio->bi_max_vecs = bio->bi_vcnt;
+
+	generic_make_request(bio);
+}
+
+/**
+ * bch_bio_split - split a bio
+ * @bio:	bio to split
+ * @sectors:	number of sectors to split from the front of @bio
+ * @gfp:	gfp mask
+ * @bs:		bio set to allocate from
+ *
+ * Allocates and returns a new bio which represents @sectors from the start of
+ * @bio, and updates @bio to represent the remaining sectors.
+ *
+ * If bio_sectors(@bio) was less than or equal to @sectors, returns @bio
+ * unchanged.
+ *
+ * The newly allocated bio will point to @bio's bi_io_vec, if the split was on a
+ * bvec boundry; it is the caller's responsibility to ensure that @bio is not
+ * freed before the split.
+ *
+ * If bch_bio_split() is running under generic_make_request(), it's not safe to
+ * allocate more than one bio from the same bio set. Therefore, if it is running
+ * under generic_make_request() it masks out __GFP_WAIT when doing the
+ * allocation. The caller must check for failure if there's any possibility of
+ * it being called from under generic_make_request(); it is then the caller's
+ * responsibility to retry from a safe context (by e.g. punting to workqueue).
+ */
+struct bio *bch_bio_split(struct bio *bio, int sectors,
+			  gfp_t gfp, struct bio_set *bs)
+{
+	unsigned idx = bio->bi_idx, vcnt = 0, nbytes = sectors << 9;
+	struct bio_vec *bv;
+	struct bio *ret = NULL;
+
+	BUG_ON(sectors <= 0);
+
+	/*
+	 * If we're being called from underneath generic_make_request() and we
+	 * already allocated any bios from this bio set, we risk deadlock if we
+	 * use the mempool. So instead, we possibly fail and let the caller punt
+	 * to workqueue or somesuch and retry in a safe context.
+	 */
+	if (current->bio_list)
+		gfp &= ~__GFP_WAIT;
+
+	if (sectors >= bio_sectors(bio))
+		return bio;
+
+	if (bio->bi_rw & REQ_DISCARD) {
+		ret = bio_alloc_bioset(gfp, 1, bs);
+		idx = 0;
+		goto out;
+	}
+
+	bio_for_each_segment(bv, bio, idx) {
+		vcnt = idx - bio->bi_idx;
+
+		if (!nbytes) {
+			ret = bio_alloc_bioset(gfp, vcnt, bs);
+			if (!ret)
+				return NULL;
+
+			memcpy(ret->bi_io_vec, bio_iovec(bio),
+			       sizeof(struct bio_vec) * vcnt);
+
+			break;
+		} else if (nbytes < bv->bv_len) {
+			ret = bio_alloc_bioset(gfp, ++vcnt, bs);
+			if (!ret)
+				return NULL;
+
+			memcpy(ret->bi_io_vec, bio_iovec(bio),
+			       sizeof(struct bio_vec) * vcnt);
+
+			ret->bi_io_vec[vcnt - 1].bv_len = nbytes;
+			bv->bv_offset	+= nbytes;
+			bv->bv_len	-= nbytes;
+			break;
+		}
+
+		nbytes -= bv->bv_len;
+	}
+out:
+	ret->bi_bdev	= bio->bi_bdev;
+	ret->bi_sector	= bio->bi_sector;
+	ret->bi_size	= sectors << 9;
+	ret->bi_rw	= bio->bi_rw;
+	ret->bi_vcnt	= vcnt;
+	ret->bi_max_vecs = vcnt;
+
+	bio->bi_sector	+= sectors;
+	bio->bi_size	-= sectors << 9;
+	bio->bi_idx	 = idx;
+
+	if (bio_integrity(bio)) {
+		if (bio_integrity_clone(ret, bio, gfp)) {
+			bio_put(ret);
+			return NULL;
+		}
+
+		bio_integrity_trim(ret, 0, bio_sectors(ret));
+		bio_integrity_trim(bio, bio_sectors(ret), bio_sectors(bio));
+	}
+
+	return ret;
+}
+
+static unsigned bch_bio_max_sectors(struct bio *bio)
+{
+	unsigned ret = bio_sectors(bio);
+	struct request_queue *q = bdev_get_queue(bio->bi_bdev);
+	unsigned max_segments = min_t(unsigned, BIO_MAX_PAGES,
+				      queue_max_segments(q));
+	struct bio_vec *bv, *end = bio_iovec(bio) +
+		min_t(int, bio_segments(bio), max_segments);
+
+	if (bio->bi_rw & REQ_DISCARD)
+		return min(ret, q->limits.max_discard_sectors);
+
+	if (bio_segments(bio) > max_segments ||
+	    q->merge_bvec_fn) {
+		ret = 0;
+
+		for (bv = bio_iovec(bio); bv < end; bv++) {
+			struct bvec_merge_data bvm = {
+				.bi_bdev	= bio->bi_bdev,
+				.bi_sector	= bio->bi_sector,
+				.bi_size	= ret << 9,
+				.bi_rw		= bio->bi_rw,
+			};
+
+			if (q->merge_bvec_fn &&
+			    q->merge_bvec_fn(q, &bvm, bv) < (int) bv->bv_len)
+				break;
+
+			ret += bv->bv_len >> 9;
+		}
+	}
+
+	ret = min(ret, queue_max_sectors(q));
+
+	WARN_ON(!ret);
+	ret = max_t(int, ret, bio_iovec(bio)->bv_len >> 9);
+
+	return ret;
+}
+
+static void bch_bio_submit_split_done(struct closure *cl)
+{
+	struct bio_split_hook *s = container_of(cl, struct bio_split_hook, cl);
+
+	s->bio->bi_end_io = s->bi_end_io;
+	s->bio->bi_private = s->bi_private;
+	bio_endio(s->bio, 0);
+
+	closure_debug_destroy(&s->cl);
+	mempool_free(s, s->p->bio_split_hook);
+}
+
+static void bch_bio_submit_split_endio(struct bio *bio, int error)
+{
+	struct closure *cl = bio->bi_private;
+	struct bio_split_hook *s = container_of(cl, struct bio_split_hook, cl);
+
+	if (error)
+		clear_bit(BIO_UPTODATE, &s->bio->bi_flags);
+
+	bio_put(bio);
+	closure_put(cl);
+}
+
+static void __bch_bio_submit_split(struct closure *cl)
+{
+	struct bio_split_hook *s = container_of(cl, struct bio_split_hook, cl);
+	struct bio *bio = s->bio, *n;
+
+	do {
+		n = bch_bio_split(bio, bch_bio_max_sectors(bio),
+				  GFP_NOIO, s->p->bio_split);
+		if (!n)
+			continue_at(cl, __bch_bio_submit_split, system_wq);
+
+		n->bi_end_io	= bch_bio_submit_split_endio;
+		n->bi_private	= cl;
+
+		closure_get(cl);
+		bch_generic_make_request_hack(n);
+	} while (n != bio);
+
+	continue_at(cl, bch_bio_submit_split_done, NULL);
+}
+
+void bch_generic_make_request(struct bio *bio, struct bio_split_pool *p)
+{
+	struct bio_split_hook *s;
+
+	if (!bio_has_data(bio) && !(bio->bi_rw & REQ_DISCARD))
+		goto submit;
+
+	if (bio_sectors(bio) <= bch_bio_max_sectors(bio))
+		goto submit;
+
+	s = mempool_alloc(p->bio_split_hook, GFP_NOIO);
+
+	s->bio		= bio;
+	s->p		= p;
+	s->bi_end_io	= bio->bi_end_io;
+	s->bi_private	= bio->bi_private;
+	bio_get(bio);
+
+	closure_call(&s->cl, __bch_bio_submit_split, NULL, NULL);
+	return;
+submit:
+	bch_generic_make_request_hack(bio);
+}
+
+/* Bios with headers */
+
+void bch_bbio_free(struct bio *bio, struct cache_set *c)
+{
+	struct bbio *b = container_of(bio, struct bbio, bio);
+	mempool_free(b, c->bio_meta);
+}
+
+struct bio *bch_bbio_alloc(struct cache_set *c)
+{
+	struct bbio *b = mempool_alloc(c->bio_meta, GFP_NOIO);
+	struct bio *bio = &b->bio;
+
+	bio_init(bio);
+	bio->bi_flags		|= BIO_POOL_NONE << BIO_POOL_OFFSET;
+	bio->bi_max_vecs	 = bucket_pages(c);
+	bio->bi_io_vec		 = bio->bi_inline_vecs;
+
+	return bio;
+}
+
+void __bch_submit_bbio(struct bio *bio, struct cache_set *c)
+{
+	struct bbio *b = container_of(bio, struct bbio, bio);
+
+	bio->bi_sector	= PTR_OFFSET(&b->key, 0);
+	bio->bi_bdev	= PTR_CACHE(c, &b->key, 0)->bdev;
+
+	b->submit_time_us = local_clock_us();
+	closure_bio_submit(bio, bio->bi_private, PTR_CACHE(c, &b->key, 0));
+}
+
+void bch_submit_bbio(struct bio *bio, struct cache_set *c,
+		     struct bkey *k, unsigned ptr)
+{
+	struct bbio *b = container_of(bio, struct bbio, bio);
+	bch_bkey_copy_single_ptr(&b->key, k, ptr);
+	__bch_submit_bbio(bio, c);
+}
+
+/* IO errors */
+
+void bch_count_io_errors(struct cache *ca, int error, const char *m)
+{
+	/*
+	 * The halflife of an error is:
+	 * log2(1/2)/log2(127/128) * refresh ~= 88 * refresh
+	 */
+
+	if (ca->set->error_decay) {
+		unsigned count = atomic_inc_return(&ca->io_count);
+
+		while (count > ca->set->error_decay) {
+			unsigned errors;
+			unsigned old = count;
+			unsigned new = count - ca->set->error_decay;
+
+			/*
+			 * First we subtract refresh from count; each time we
+			 * succesfully do so, we rescale the errors once:
+			 */
+
+			count = atomic_cmpxchg(&ca->io_count, old, new);
+
+			if (count == old) {
+				count = new;
+
+				errors = atomic_read(&ca->io_errors);
+				do {
+					old = errors;
+					new = ((uint64_t) errors * 127) / 128;
+					errors = atomic_cmpxchg(&ca->io_errors,
+								old, new);
+				} while (old != errors);
+			}
+		}
+	}
+
+	if (error) {
+		char buf[BDEVNAME_SIZE];
+		unsigned errors = atomic_add_return(1 << IO_ERROR_SHIFT,
+						    &ca->io_errors);
+		errors >>= IO_ERROR_SHIFT;
+
+		if (errors < ca->set->error_limit)
+			pr_err("%s: IO error on %s, recovering",
+			       bdevname(ca->bdev, buf), m);
+		else
+			bch_cache_set_error(ca->set,
+					    "%s: too many IO errors %s",
+					    bdevname(ca->bdev, buf), m);
+	}
+}
+
+void bch_bbio_count_io_errors(struct cache_set *c, struct bio *bio,
+			      int error, const char *m)
+{
+	struct bbio *b = container_of(bio, struct bbio, bio);
+	struct cache *ca = PTR_CACHE(c, &b->key, 0);
+
+	unsigned threshold = bio->bi_rw & REQ_WRITE
+		? c->congested_write_threshold_us
+		: c->congested_read_threshold_us;
+
+	if (threshold) {
+		unsigned t = local_clock_us();
+
+		int us = t - b->submit_time_us;
+		int congested = atomic_read(&c->congested);
+
+		if (us > (int) threshold) {
+			int ms = us / 1024;
+			c->congested_last_us = t;
+
+			ms = min(ms, CONGESTED_MAX + congested);
+			atomic_sub(ms, &c->congested);
+		} else if (congested < 0)
+			atomic_inc(&c->congested);
+	}
+
+	bch_count_io_errors(ca, error, m);
+}
+
+void bch_bbio_endio(struct cache_set *c, struct bio *bio,
+		    int error, const char *m)
+{
+	struct closure *cl = bio->bi_private;
+
+	bch_bbio_count_io_errors(c, bio, error, m);
+	bio_put(bio);
+	closure_put(cl);
+}

drivers/md/bcache/journal.c

···
+/*
+ * bcache journalling code, for btree insertions
+ *
+ * Copyright 2012 Google, Inc.
+ */
+
+#include "bcache.h"
+#include "btree.h"
+#include "debug.h"
+#include "request.h"
+
+/*
+ * Journal replay/recovery:
+ *
+ * This code is all driven from run_cache_set(); we first read the journal
+ * entries, do some other stuff, then we mark all the keys in the journal
+ * entries (same as garbage collection would), then we replay them - reinserting
+ * them into the cache in precisely the same order as they appear in the
+ * journal.
+ *
+ * We only journal keys that go in leaf nodes, which simplifies things quite a
+ * bit.
+ */
+
+static void journal_read_endio(struct bio *bio, int error)
+{
+	struct closure *cl = bio->bi_private;
+	closure_put(cl);
+}
+
+static int journal_read_bucket(struct cache *ca, struct list_head *list,
+			       struct btree_op *op, unsigned bucket_index)
+{
+	struct journal_device *ja = &ca->journal;
+	struct bio *bio = &ja->bio;
+
+	struct journal_replay *i;
+	struct jset *j, *data = ca->set->journal.w[0].data;
+	unsigned len, left, offset = 0;
+	int ret = 0;
+	sector_t bucket = bucket_to_sector(ca->set, ca->sb.d[bucket_index]);
+
+	pr_debug("reading %llu", (uint64_t) bucket);
+
+	while (offset < ca->sb.bucket_size) {
+reread:		left = ca->sb.bucket_size - offset;
+		len = min_t(unsigned, left, PAGE_SECTORS * 8);
+
+		bio_reset(bio);
+		bio->bi_sector	= bucket + offset;
+		bio->bi_bdev	= ca->bdev;
+		bio->bi_rw	= READ;
+		bio->bi_size	= len << 9;
+
+		bio->bi_end_io	= journal_read_endio;
+		bio->bi_private = &op->cl;
+		bch_bio_map(bio, data);
+
+		closure_bio_submit(bio, &op->cl, ca);
+		closure_sync(&op->cl);
+
+		/* This function could be simpler now since we no longer write
+		 * journal entries that overlap bucket boundaries; this means
+		 * the start of a bucket will always have a valid journal entry
+		 * if it has any journal entries at all.
+		 */
+
+		j = data;
+		while (len) {
+			struct list_head *where;
+			size_t blocks, bytes = set_bytes(j);
+
+			if (j->magic != jset_magic(ca->set))
+				return ret;
+
+			if (bytes > left << 9)
+				return ret;
+
+			if (bytes > len << 9)
+				goto reread;
+
+			if (j->csum != csum_set(j))
+				return ret;
+
+			blocks = set_blocks(j, ca->set);
+
+			while (!list_empty(list)) {
+				i = list_first_entry(list,
+					struct journal_replay, list);
+				if (i->j.seq >= j->last_seq)
+					break;
+				list_del(&i->list);
+				kfree(i);
+			}
+
+			list_for_each_entry_reverse(i, list, list) {
+				if (j->seq == i->j.seq)
+					goto next_set;
+
+				if (j->seq < i->j.last_seq)
+					goto next_set;
+
+				if (j->seq > i->j.seq) {
+					where = &i->list;
+					goto add;
+				}
+			}
+
+			where = list;
+add:
+			i = kmalloc(offsetof(struct journal_replay, j) +
+				    bytes, GFP_KERNEL);
+			if (!i)
+				return -ENOMEM;
+			memcpy(&i->j, j, bytes);
+			list_add(&i->list, where);
+			ret = 1;
+
+			ja->seq[bucket_index] = j->seq;
+next_set:
+			offset	+= blocks * ca->sb.block_size;
+			len	-= blocks * ca->sb.block_size;
+			j = ((void *) j) + blocks * block_bytes(ca);
+		}
+	}
+
+	return ret;
+}
+
+int bch_journal_read(struct cache_set *c, struct list_head *list,
+			struct btree_op *op)
+{
+#define read_bucket(b)							\
+	({								\
+		int ret = journal_read_bucket(ca, list, op, b);		\
+		__set_bit(b, bitmap);					\
+		if (ret < 0)						\
+			return ret;					\
+		ret;							\
+	})
+
+	struct cache *ca;
+	unsigned iter;
+
+	for_each_cache(ca, c, iter) {
+		struct journal_device *ja = &ca->journal;
+		unsigned long bitmap[SB_JOURNAL_BUCKETS / BITS_PER_LONG];
+		unsigned i, l, r, m;
+		uint64_t seq;
+
+		bitmap_zero(bitmap, SB_JOURNAL_BUCKETS);
+		pr_debug("%u journal buckets", ca->sb.njournal_buckets);
+
+		/* Read journal buckets ordered by golden ratio hash to quickly
+		 * find a sequence of buckets with valid journal entries
+		 */
+		for (i = 0; i < ca->sb.njournal_buckets; i++) {
+			l = (i * 2654435769U) % ca->sb.njournal_buckets;
+
+			if (test_bit(l, bitmap))
+				break;
+
+			if (read_bucket(l))
+				goto bsearch;
+		}
+
+		/* If that fails, check all the buckets we haven't checked
+		 * already
+		 */
+		pr_debug("falling back to linear search");
+
+		for (l = 0; l < ca->sb.njournal_buckets; l++) {
+			if (test_bit(l, bitmap))
+				continue;
+
+			if (read_bucket(l))
+				goto bsearch;
+		}
+bsearch:
+		/* Binary search */
+		m = r = find_next_bit(bitmap, ca->sb.njournal_buckets, l + 1);
+		pr_debug("starting binary search, l %u r %u", l, r);
+
+		while (l + 1 < r) {
+			m = (l + r) >> 1;
+
+			if (read_bucket(m))
+				l = m;
+			else
+				r = m;
+		}
+
+		/* Read buckets in reverse order until we stop finding more
+		 * journal entries
+		 */
+		pr_debug("finishing up");
+		l = m;
+
+		while (1) {
+			if (!l--)
+				l = ca->sb.njournal_buckets - 1;
+
+			if (l == m)
+				break;
+
+			if (test_bit(l, bitmap))
+				continue;
+
+			if (!read_bucket(l))
+				break;
+		}
+
+		seq = 0;
+
+		for (i = 0; i < ca->sb.njournal_buckets; i++)
+			if (ja->seq[i] > seq) {
+				seq = ja->seq[i];
+				ja->cur_idx = ja->discard_idx =
+					ja->last_idx = i;
+
+			}
+	}
+
+	c->journal.seq = list_entry(list->prev,
+				    struct journal_replay,
+				    list)->j.seq;
+
+	return 0;
+#undef read_bucket
+}
+
+void bch_journal_mark(struct cache_set *c, struct list_head *list)
+{
+	atomic_t p = { 0 };
+	struct bkey *k;
+	struct journal_replay *i;
+	struct journal *j = &c->journal;
+	uint64_t last = j->seq;
+
+	/*
+	 * journal.pin should never fill up - we never write a journal
+	 * entry when it would fill up. But if for some reason it does, we
+	 * iterate over the list in reverse order so that we can just skip that
+	 * refcount instead of bugging.
+	 */
+
+	list_for_each_entry_reverse(i, list, list) {
+		BUG_ON(last < i->j.seq);
+		i->pin = NULL;
+
+		while (last-- != i->j.seq)
+			if (fifo_free(&j->pin) > 1) {
+				fifo_push_front(&j->pin, p);
+				atomic_set(&fifo_front(&j->pin), 0);
+			}
+
+		if (fifo_free(&j->pin) > 1) {
+			fifo_push_front(&j->pin, p);
+			i->pin = &fifo_front(&j->pin);
+			atomic_set(i->pin, 1);
+		}
+
+		for (k = i->j.start;
+		     k < end(&i->j);
+		     k = bkey_next(k)) {
+			unsigned j;
+
+			for (j = 0; j < KEY_PTRS(k); j++) {
+				struct bucket *g = PTR_BUCKET(c, k, j);
+				atomic_inc(&g->pin);
+
+				if (g->prio == BTREE_PRIO &&
+				    !ptr_stale(c, k, j))
+					g->prio = INITIAL_PRIO;
+			}
+
+			__bch_btree_mark_key(c, 0, k);
+		}
+	}
+}
+
+int bch_journal_replay(struct cache_set *s, struct list_head *list,
+			  struct btree_op *op)
+{
+	int ret = 0, keys = 0, entries = 0;
+	struct bkey *k;
+	struct journal_replay *i =
+		list_entry(list->prev, struct journal_replay, list);
+
+	uint64_t start = i->j.last_seq, end = i->j.seq, n = start;
+
+	list_for_each_entry(i, list, list) {
+		BUG_ON(i->pin && atomic_read(i->pin) != 1);
+
+		if (n != i->j.seq)
+			pr_err(
+		"journal entries %llu-%llu missing! (replaying %llu-%llu)\n",
+		n, i->j.seq - 1, start, end);
+
+		for (k = i->j.start;
+		     k < end(&i->j);
+		     k = bkey_next(k)) {
+			pr_debug("%s", pkey(k));
+			bkey_copy(op->keys.top, k);
+			bch_keylist_push(&op->keys);
+
+			op->journal = i->pin;
+			atomic_inc(op->journal);
+
+			ret = bch_btree_insert(op, s);
+			if (ret)
+				goto err;
+
+			BUG_ON(!bch_keylist_empty(&op->keys));
+			keys++;
+
+			cond_resched();
+		}
+
+		if (i->pin)
+			atomic_dec(i->pin);
+		n = i->j.seq + 1;
+		entries++;
+	}
+
+	pr_info("journal replay done, %i keys in %i entries, seq %llu",
+		keys, entries, end);
+
+	while (!list_empty(list)) {
+		i = list_first_entry(list, struct journal_replay, list);
+		list_del(&i->list);
+		kfree(i);
+	}
+err:
+	closure_sync(&op->cl);
+	return ret;
+}
+
+/* Journalling */
+
+static void btree_flush_write(struct cache_set *c)
+{
+	/*
+	 * Try to find the btree node with that references the oldest journal
+	 * entry, best is our current candidate and is locked if non NULL:
+	 */
+	struct btree *b, *best = NULL;
+	unsigned iter;
+
+	for_each_cached_btree(b, c, iter) {
+		if (!down_write_trylock(&b->lock))
+			continue;
+
+		if (!btree_node_dirty(b) ||
+		    !btree_current_write(b)->journal) {
+			rw_unlock(true, b);
+			continue;
+		}
+
+		if (!best)
+			best = b;
+		else if (journal_pin_cmp(c,
+					 btree_current_write(best),
+					 btree_current_write(b))) {
+			rw_unlock(true, best);
+			best = b;
+		} else
+			rw_unlock(true, b);
+	}
+
+	if (best)
+		goto out;
+
+	/* We can't find the best btree node, just pick the first */
+	list_for_each_entry(b, &c->btree_cache, list)
+		if (!b->level && btree_node_dirty(b)) {
+			best = b;
+			rw_lock(true, best, best->level);
+			goto found;
+		}
+
+out:
+	if (!best)
+		return;
+found:
+	if (btree_node_dirty(best))
+		bch_btree_write(best, true, NULL);
+	rw_unlock(true, best);
+}
+
+#define last_seq(j)	((j)->seq - fifo_used(&(j)->pin) + 1)
+
+static void journal_discard_endio(struct bio *bio, int error)
+{
+	struct journal_device *ja =
+		container_of(bio, struct journal_device, discard_bio);
+	struct cache *ca = container_of(ja, struct cache, journal);
+
+	atomic_set(&ja->discard_in_flight, DISCARD_DONE);
+
+	closure_wake_up(&ca->set->journal.wait);
+	closure_put(&ca->set->cl);
+}
+
+static void journal_discard_work(struct work_struct *work)
+{
+	struct journal_device *ja =
+		container_of(work, struct journal_device, discard_work);
+
+	submit_bio(0, &ja->discard_bio);
+}
+
+static void do_journal_discard(struct cache *ca)
+{
+	struct journal_device *ja = &ca->journal;
+	struct bio *bio = &ja->discard_bio;
+
+	if (!ca->discard) {
+		ja->discard_idx = ja->last_idx;
+		return;
+	}
+
+	switch (atomic_read(&ja->discard_in_flight) == DISCARD_IN_FLIGHT) {
+	case DISCARD_IN_FLIGHT:
+		return;
+
+	case DISCARD_DONE:
+		ja->discard_idx = (ja->discard_idx + 1) %
+			ca->sb.njournal_buckets;
+
+		atomic_set(&ja->discard_in_flight, DISCARD_READY);
+		/* fallthrough */
+
+	case DISCARD_READY:
+		if (ja->discard_idx == ja->last_idx)
+			return;
+
+		atomic_set(&ja->discard_in_flight, DISCARD_IN_FLIGHT);
+
+		bio_init(bio);
+		bio->bi_sector		= bucket_to_sector(ca->set,
+						ca->sb.d[ja->discard_idx]);
+		bio->bi_bdev		= ca->bdev;
+		bio->bi_rw		= REQ_WRITE|REQ_DISCARD;
+		bio->bi_max_vecs	= 1;
+		bio->bi_io_vec		= bio->bi_inline_vecs;
+		bio->bi_size		= bucket_bytes(ca);
+		bio->bi_end_io		= journal_discard_endio;
+
+		closure_get(&ca->set->cl);
+		INIT_WORK(&ja->discard_work, journal_discard_work);
+		schedule_work(&ja->discard_work);
+	}
+}
+
+static void journal_reclaim(struct cache_set *c)
+{
+	struct bkey *k = &c->journal.key;
+	struct cache *ca;
+	uint64_t last_seq;
+	unsigned iter, n = 0;
+	atomic_t p;
+
+	while (!atomic_read(&fifo_front(&c->journal.pin)))
+		fifo_pop(&c->journal.pin, p);
+
+	last_seq = last_seq(&c->journal);
+
+	/* Update last_idx */
+
+	for_each_cache(ca, c, iter) {
+		struct journal_device *ja = &ca->journal;
+
+		while (ja->last_idx != ja->cur_idx &&
+		       ja->seq[ja->last_idx] < last_seq)
+			ja->last_idx = (ja->last_idx + 1) %
+				ca->sb.njournal_buckets;
+	}
+
+	for_each_cache(ca, c, iter)
+		do_journal_discard(ca);
+
+	if (c->journal.blocks_free)
+		return;
+
+	/*
+	 * Allocate:
+	 * XXX: Sort by free journal space
+	 */
+
+	for_each_cache(ca, c, iter) {
+		struct journal_device *ja = &ca->journal;
+		unsigned next = (ja->cur_idx + 1) % ca->sb.njournal_buckets;
+
+		/* No space available on this device */
+		if (next == ja->discard_idx)
+			continue;
+
+		ja->cur_idx = next;
+		k->ptr[n++] = PTR(0,
+				  bucket_to_sector(c, ca->sb.d[ja->cur_idx]),
+				  ca->sb.nr_this_dev);
+	}
+
+	bkey_init(k);
+	SET_KEY_PTRS(k, n);
+
+	if (n)
+		c->journal.blocks_free = c->sb.bucket_size >> c->block_bits;
+
+	if (!journal_full(&c->journal))
+		__closure_wake_up(&c->journal.wait);
+}
+
+void bch_journal_next(struct journal *j)
+{
+	atomic_t p = { 1 };
+
+	j->cur = (j->cur == j->w)
+		? &j->w[1]
+		: &j->w[0];
+
+	/*
+	 * The fifo_push() needs to happen at the same time as j->seq is
+	 * incremented for last_seq() to be calculated correctly
+	 */
+	BUG_ON(!fifo_push(&j->pin, p));
+	atomic_set(&fifo_back(&j->pin), 1);
+
+	j->cur->data->seq	= ++j->seq;
+	j->cur->need_write	= false;
+	j->cur->data->keys	= 0;
+
+	if (fifo_full(&j->pin))
+		pr_debug("journal_pin full (%zu)", fifo_used(&j->pin));
+}
+
+static void journal_write_endio(struct bio *bio, int error)
+{
+	struct journal_write *w = bio->bi_private;
+
+	cache_set_err_on(error, w->c, "journal io error");
+	closure_put(&w->c->journal.io.cl);
+}
+
+static void journal_write(struct closure *);
+
+static void journal_write_done(struct closure *cl)
+{
+	struct journal *j = container_of(cl, struct journal, io.cl);
+	struct cache_set *c = container_of(j, struct cache_set, journal);
+
+	struct journal_write *w = (j->cur == j->w)
+		? &j->w[1]
+		: &j->w[0];
+
+	__closure_wake_up(&w->wait);
+
+	if (c->journal_delay_ms)
+		closure_delay(&j->io, msecs_to_jiffies(c->journal_delay_ms));
+
+	continue_at(cl, journal_write, system_wq);
+}
+
+static void journal_write_unlocked(struct closure *cl)
+	__releases(c->journal.lock)
+{
+	struct cache_set *c = container_of(cl, struct cache_set, journal.io.cl);
+	struct cache *ca;
+	struct journal_write *w = c->journal.cur;
+	struct bkey *k = &c->journal.key;
+	unsigned i, sectors = set_blocks(w->data, c) * c->sb.block_size;
+
+	struct bio *bio;
+	struct bio_list list;
+	bio_list_init(&list);
+
+	if (!w->need_write) {
+		/*
+		 * XXX: have to unlock closure before we unlock journal lock,
+		 * else we race with bch_journal(). But this way we race
+		 * against cache set unregister. Doh.
+		 */
+		set_closure_fn(cl, NULL, NULL);
+		closure_sub(cl, CLOSURE_RUNNING + 1);
+		spin_unlock(&c->journal.lock);
+		return;
+	} else if (journal_full(&c->journal)) {
+		journal_reclaim(c);
+		spin_unlock(&c->journal.lock);
+
+		btree_flush_write(c);
+		continue_at(cl, journal_write, system_wq);
+	}
+
+	c->journal.blocks_free -= set_blocks(w->data, c);
+
+	w->data->btree_level = c->root->level;
+
+	bkey_copy(&w->data->btree_root, &c->root->key);
+	bkey_copy(&w->data->uuid_bucket, &c->uuid_bucket);
+
+	for_each_cache(ca, c, i)
+		w->data->prio_bucket[ca->sb.nr_this_dev] = ca->prio_buckets[0];
+
+	w->data->magic		= jset_magic(c);
+	w->data->version	= BCACHE_JSET_VERSION;
+	w->data->last_seq	= last_seq(&c->journal);
+	w->data->csum		= csum_set(w->data);
+
+	for (i = 0; i < KEY_PTRS(k); i++) {
+		ca = PTR_CACHE(c, k, i);
+		bio = &ca->journal.bio;
+
+		atomic_long_add(sectors, &ca->meta_sectors_written);
+
+		bio_reset(bio);
+		bio->bi_sector	= PTR_OFFSET(k, i);
+		bio->bi_bdev	= ca->bdev;
+		bio->bi_rw	= REQ_WRITE|REQ_SYNC|REQ_META|REQ_FLUSH;
+		bio->bi_size	= sectors << 9;
+
+		bio->bi_end_io	= journal_write_endio;
+		bio->bi_private = w;
+		bch_bio_map(bio, w->data);
+
+		trace_bcache_journal_write(bio);
+		bio_list_add(&list, bio);
+
+		SET_PTR_OFFSET(k, i, PTR_OFFSET(k, i) + sectors);
+
+		ca->journal.seq[ca->journal.cur_idx] = w->data->seq;
+	}
+
+	atomic_dec_bug(&fifo_back(&c->journal.pin));
+	bch_journal_next(&c->journal);
+	journal_reclaim(c);
+
+	spin_unlock(&c->journal.lock);
+
+	while ((bio = bio_list_pop(&list)))
+		closure_bio_submit(bio, cl, c->cache[0]);
+
+	continue_at(cl, journal_write_done, NULL);
+}
+
+static void journal_write(struct closure *cl)
+{
+	struct cache_set *c = container_of(cl, struct cache_set, journal.io.cl);
+
+	spin_lock(&c->journal.lock);
+	journal_write_unlocked(cl);
+}
+
+static void __journal_try_write(struct cache_set *c, bool noflush)
+	__releases(c->journal.lock)
+{
+	struct closure *cl = &c->journal.io.cl;
+
+	if (!closure_trylock(cl, &c->cl))
+		spin_unlock(&c->journal.lock);
+	else if (noflush && journal_full(&c->journal)) {
+		spin_unlock(&c->journal.lock);
+		continue_at(cl, journal_write, system_wq);
+	} else
+		journal_write_unlocked(cl);
+}
+
+#define journal_try_write(c)	__journal_try_write(c, false)
+
+void bch_journal_meta(struct cache_set *c, struct closure *cl)
+{
+	struct journal_write *w;
+
+	if (CACHE_SYNC(&c->sb)) {
+		spin_lock(&c->journal.lock);
+
+		w = c->journal.cur;
+		w->need_write = true;
+
+		if (cl)
+			BUG_ON(!closure_wait(&w->wait, cl));
+
+		__journal_try_write(c, true);
+	}
+}
+
+/*
+ * Entry point to the journalling code - bio_insert() and btree_invalidate()
+ * pass bch_journal() a list of keys to be journalled, and then
+ * bch_journal() hands those same keys off to btree_insert_async()
+ */
+
+void bch_journal(struct closure *cl)
+{
+	struct btree_op *op = container_of(cl, struct btree_op, cl);
+	struct cache_set *c = op->c;
+	struct journal_write *w;
+	size_t b, n = ((uint64_t *) op->keys.top) - op->keys.list;
+
+	if (op->type != BTREE_INSERT ||
+	    !CACHE_SYNC(&c->sb))
+		goto out;
+
+	/*
+	 * If we're looping because we errored, might already be waiting on
+	 * another journal write:
+	 */
+	while (atomic_read(&cl->parent->remaining) & CLOSURE_WAITING)
+		closure_sync(cl->parent);
+
+	spin_lock(&c->journal.lock);
+
+	if (journal_full(&c->journal)) {
+		/* XXX: tracepoint */
+		closure_wait(&c->journal.wait, cl);
+
+		journal_reclaim(c);
+		spin_unlock(&c->journal.lock);
+
+		btree_flush_write(c);
+		continue_at(cl, bch_journal, bcache_wq);
+	}
+
+	w = c->journal.cur;
+	w->need_write = true;
+	b = __set_blocks(w->data, w->data->keys + n, c);
+
+	if (b * c->sb.block_size > PAGE_SECTORS << JSET_BITS ||
+	    b > c->journal.blocks_free) {
+		/* XXX: If we were inserting so many keys that they won't fit in
+		 * an _empty_ journal write, we'll deadlock. For now, handle
+		 * this in bch_keylist_realloc() - but something to think about.
+		 */
+		BUG_ON(!w->data->keys);
+
+		/* XXX: tracepoint */
+		BUG_ON(!closure_wait(&w->wait, cl));
+
+		closure_flush(&c->journal.io);
+
+		journal_try_write(c);
+		continue_at(cl, bch_journal, bcache_wq);
+	}
+
+	memcpy(end(w->data), op->keys.list, n * sizeof(uint64_t));
+	w->data->keys += n;
+
+	op->journal = &fifo_back(&c->journal.pin);
+	atomic_inc(op->journal);
+
+	if (op->flush_journal) {
+		closure_flush(&c->journal.io);
+		closure_wait(&w->wait, cl->parent);
+	}
+
+	journal_try_write(c);
+out:
+	bch_btree_insert_async(cl);
+}
+
+void bch_journal_free(struct cache_set *c)
+{
+	free_pages((unsigned long) c->journal.w[1].data, JSET_BITS);
+	free_pages((unsigned long) c->journal.w[0].data, JSET_BITS);
+	free_fifo(&c->journal.pin);
+}
+
+int bch_journal_alloc(struct cache_set *c)
+{
+	struct journal *j = &c->journal;
+
+	closure_init_unlocked(&j->io);
+	spin_lock_init(&j->lock);
+
+	c->journal_delay_ms = 100;
+
+	j->w[0].c = c;
+	j->w[1].c = c;
+
+	if (!(init_fifo(&j->pin, JOURNAL_PIN, GFP_KERNEL)) ||
+	    !(j->w[0].data = (void *) __get_free_pages(GFP_KERNEL, JSET_BITS)) ||
+	    !(j->w[1].data = (void *) __get_free_pages(GFP_KERNEL, JSET_BITS)))
+		return -ENOMEM;
+
+	return 0;
+}

drivers/md/bcache/journal.h

···
+#ifndef _BCACHE_JOURNAL_H
+#define _BCACHE_JOURNAL_H
+
+/*
+ * THE JOURNAL:
+ *
+ * The journal is treated as a circular buffer of buckets - a journal entry
+ * never spans two buckets. This means (not implemented yet) we can resize the
+ * journal at runtime, and will be needed for bcache on raw flash support.
+ *
+ * Journal entries contain a list of keys, ordered by the time they were
+ * inserted; thus journal replay just has to reinsert the keys.
+ *
+ * We also keep some things in the journal header that are logically part of the
+ * superblock - all the things that are frequently updated. This is for future
+ * bcache on raw flash support; the superblock (which will become another
+ * journal) can't be moved or wear leveled, so it contains just enough
+ * information to find the main journal, and the superblock only has to be
+ * rewritten when we want to move/wear level the main journal.
+ *
+ * Currently, we don't journal BTREE_REPLACE operations - this will hopefully be
+ * fixed eventually. This isn't a bug - BTREE_REPLACE is used for insertions
+ * from cache misses, which don't have to be journaled, and for writeback and
+ * moving gc we work around it by flushing the btree to disk before updating the
+ * gc information. But it is a potential issue with incremental garbage
+ * collection, and it's fragile.
+ *
+ * OPEN JOURNAL ENTRIES:
+ *
+ * Each journal entry contains, in the header, the sequence number of the last
+ * journal entry still open - i.e. that has keys that haven't been flushed to
+ * disk in the btree.
+ *
+ * We track this by maintaining a refcount for every open journal entry, in a
+ * fifo; each entry in the fifo corresponds to a particular journal
+ * entry/sequence number. When the refcount at the tail of the fifo goes to
+ * zero, we pop it off - thus, the size of the fifo tells us the number of open
+ * journal entries
+ *
+ * We take a refcount on a journal entry when we add some keys to a journal
+ * entry that we're going to insert (held by struct btree_op), and then when we
+ * insert those keys into the btree the btree write we're setting up takes a
+ * copy of that refcount (held by struct btree_write). That refcount is dropped
+ * when the btree write completes.
+ *
+ * A struct btree_write can only hold a refcount on a single journal entry, but
+ * might contain keys for many journal entries - we handle this by making sure
+ * it always has a refcount on the _oldest_ journal entry of all the journal
+ * entries it has keys for.
+ *
+ * JOURNAL RECLAIM:
+ *
+ * As mentioned previously, our fifo of refcounts tells us the number of open
+ * journal entries; from that and the current journal sequence number we compute
+ * last_seq - the oldest journal entry we still need. We write last_seq in each
+ * journal entry, and we also have to keep track of where it exists on disk so
+ * we don't overwrite it when we loop around the journal.
+ *
+ * To do that we track, for each journal bucket, the sequence number of the
+ * newest journal entry it contains - if we don't need that journal entry we
+ * don't need anything in that bucket anymore. From that we track the last
+ * journal bucket we still need; all this is tracked in struct journal_device
+ * and updated by journal_reclaim().
+ *
+ * JOURNAL FILLING UP:
+ *
+ * There are two ways the journal could fill up; either we could run out of
+ * space to write to, or we could have too many open journal entries and run out
+ * of room in the fifo of refcounts. Since those refcounts are decremented
+ * without any locking we can't safely resize that fifo, so we handle it the
+ * same way.
+ *
+ * If the journal fills up, we start flushing dirty btree nodes until we can
+ * allocate space for a journal write again - preferentially flushing btree
+ * nodes that are pinning the oldest journal entries first.
+ */
+
+#define BCACHE_JSET_VERSION_UUIDv1	1
+/* Always latest UUID format */
+#define BCACHE_JSET_VERSION_UUID	1
+#define BCACHE_JSET_VERSION		1
+
+/*
+ * On disk format for a journal entry:
+ * seq is monotonically increasing; every journal entry has its own unique
+ * sequence number.
+ *
+ * last_seq is the oldest journal entry that still has keys the btree hasn't
+ * flushed to disk yet.
+ *
+ * version is for on disk format changes.
+ */
+struct jset {
+	uint64_t		csum;
+	uint64_t		magic;
+	uint64_t		seq;
+	uint32_t		version;
+	uint32_t		keys;
+
+	uint64_t		last_seq;
+
+	BKEY_PADDED(uuid_bucket);
+	BKEY_PADDED(btree_root);
+	uint16_t		btree_level;
+	uint16_t		pad[3];
+
+	uint64_t		prio_bucket[MAX_CACHES_PER_SET];
+
+	union {
+		struct bkey	start[0];
+		uint64_t	d[0];
+	};
+};
+
+/*
+ * Only used for holding the journal entries we read in btree_journal_read()
+ * during cache_registration
+ */
+struct journal_replay {
+	struct list_head	list;
+	atomic_t		*pin;
+	struct jset		j;
+};
+
+/*
+ * We put two of these in struct journal; we used them for writes to the
+ * journal that are being staged or in flight.
+ */
+struct journal_write {
+	struct jset		*data;
+#define JSET_BITS		3
+
+	struct cache_set	*c;
+	struct closure_waitlist	wait;
+	bool			need_write;
+};
+
+/* Embedded in struct cache_set */
+struct journal {
+	spinlock_t		lock;
+	/* used when waiting because the journal was full */
+	struct closure_waitlist	wait;
+	struct closure_with_timer io;
+
+	/* Number of blocks free in the bucket(s) we're currently writing to */
+	unsigned		blocks_free;
+	uint64_t		seq;
+	DECLARE_FIFO(atomic_t, pin);
+
+	BKEY_PADDED(key);
+
+	struct journal_write	w[2], *cur;
+};
+
+/*
+ * Embedded in struct cache. First three fields refer to the array of journal
+ * buckets, in cache_sb.
+ */
+struct journal_device {
+	/*
+	 * For each journal bucket, contains the max sequence number of the
+	 * journal writes it contains - so we know when a bucket can be reused.
+	 */
+	uint64_t		seq[SB_JOURNAL_BUCKETS];
+
+	/* Journal bucket we're currently writing to */
+	unsigned		cur_idx;
+
+	/* Last journal bucket that still contains an open journal entry */
+	unsigned		last_idx;
+
+	/* Next journal bucket to be discarded */
+	unsigned		discard_idx;
+
+#define DISCARD_READY		0
+#define DISCARD_IN_FLIGHT	1
+#define DISCARD_DONE		2
+	/* 1 - discard in flight, -1 - discard completed */
+	atomic_t		discard_in_flight;
+
+	struct work_struct	discard_work;
+	struct bio		discard_bio;
+	struct bio_vec		discard_bv;
+
+	/* Bio for journal reads/writes to this device */
+	struct bio		bio;
+	struct bio_vec		bv[8];
+};
+
+#define journal_pin_cmp(c, l, r)				\
+	(fifo_idx(&(c)->journal.pin, (l)->journal) >		\
+	 fifo_idx(&(c)->journal.pin, (r)->journal))
+
+#define JOURNAL_PIN	20000
+
+#define journal_full(j)						\
+	(!(j)->blocks_free || fifo_free(&(j)->pin) <= 1)
+
+struct closure;
+struct cache_set;
+struct btree_op;
+
+void bch_journal(struct closure *);
+void bch_journal_next(struct journal *);
+void bch_journal_mark(struct cache_set *, struct list_head *);
+void bch_journal_meta(struct cache_set *, struct closure *);
+int bch_journal_read(struct cache_set *, struct list_head *,
+			struct btree_op *);
+int bch_journal_replay(struct cache_set *, struct list_head *,
+			  struct btree_op *);
+
+void bch_journal_free(struct cache_set *);
+int bch_journal_alloc(struct cache_set *);
+
+#endif /* _BCACHE_JOURNAL_H */

drivers/md/bcache/movinggc.c

···
+/*
+ * Moving/copying garbage collector
+ *
+ * Copyright 2012 Google, Inc.
+ */
+
+#include "bcache.h"
+#include "btree.h"
+#include "debug.h"
+#include "request.h"
+
+struct moving_io {
+	struct keybuf_key	*w;
+	struct search		s;
+	struct bbio		bio;
+};
+
+static bool moving_pred(struct keybuf *buf, struct bkey *k)
+{
+	struct cache_set *c = container_of(buf, struct cache_set,
+					   moving_gc_keys);
+	unsigned i;
+
+	for (i = 0; i < KEY_PTRS(k); i++) {
+		struct cache *ca = PTR_CACHE(c, k, i);
+		struct bucket *g = PTR_BUCKET(c, k, i);
+
+		if (GC_SECTORS_USED(g) < ca->gc_move_threshold)
+			return true;
+	}
+
+	return false;
+}
+
+/* Moving GC - IO loop */
+
+static void moving_io_destructor(struct closure *cl)
+{
+	struct moving_io *io = container_of(cl, struct moving_io, s.cl);
+	kfree(io);
+}
+
+static void write_moving_finish(struct closure *cl)
+{
+	struct moving_io *io = container_of(cl, struct moving_io, s.cl);
+	struct bio *bio = &io->bio.bio;
+	struct bio_vec *bv = bio_iovec_idx(bio, bio->bi_vcnt);
+
+	while (bv-- != bio->bi_io_vec)
+		__free_page(bv->bv_page);
+
+	pr_debug("%s %s", io->s.op.insert_collision
+		 ? "collision moving" : "moved",
+		 pkey(&io->w->key));
+
+	bch_keybuf_del(&io->s.op.c->moving_gc_keys, io->w);
+
+	atomic_dec_bug(&io->s.op.c->in_flight);
+	closure_wake_up(&io->s.op.c->moving_gc_wait);
+
+	closure_return_with_destructor(cl, moving_io_destructor);
+}
+
+static void read_moving_endio(struct bio *bio, int error)
+{
+	struct moving_io *io = container_of(bio->bi_private,
+					    struct moving_io, s.cl);
+
+	if (error)
+		io->s.error = error;
+
+	bch_bbio_endio(io->s.op.c, bio, error, "reading data to move");
+}
+
+static void moving_init(struct moving_io *io)
+{
+	struct bio *bio = &io->bio.bio;
+
+	bio_init(bio);
+	bio_get(bio);
+	bio_set_prio(bio, IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0));
+
+	bio->bi_size		= KEY_SIZE(&io->w->key) << 9;
+	bio->bi_max_vecs	= DIV_ROUND_UP(KEY_SIZE(&io->w->key),
+					       PAGE_SECTORS);
+	bio->bi_private		= &io->s.cl;
+	bio->bi_io_vec		= bio->bi_inline_vecs;
+	bch_bio_map(bio, NULL);
+}
+
+static void write_moving(struct closure *cl)
+{
+	struct search *s = container_of(cl, struct search, cl);
+	struct moving_io *io = container_of(s, struct moving_io, s);
+
+	if (!s->error) {
+		trace_bcache_write_moving(&io->bio.bio);
+
+		moving_init(io);
+
+		io->bio.bio.bi_sector	= KEY_START(&io->w->key);
+		s->op.lock		= -1;
+		s->op.write_prio	= 1;
+		s->op.cache_bio		= &io->bio.bio;
+
+		s->writeback		= KEY_DIRTY(&io->w->key);
+		s->op.csum		= KEY_CSUM(&io->w->key);
+
+		s->op.type = BTREE_REPLACE;
+		bkey_copy(&s->op.replace, &io->w->key);
+
+		closure_init(&s->op.cl, cl);
+		bch_insert_data(&s->op.cl);
+	}
+
+	continue_at(cl, write_moving_finish, NULL);
+}
+
+static void read_moving_submit(struct closure *cl)
+{
+	struct search *s = container_of(cl, struct search, cl);
+	struct moving_io *io = container_of(s, struct moving_io, s);
+	struct bio *bio = &io->bio.bio;
+
+	trace_bcache_read_moving(bio);
+	bch_submit_bbio(bio, s->op.c, &io->w->key, 0);
+
+	continue_at(cl, write_moving, bch_gc_wq);
+}
+
+static void read_moving(struct closure *cl)
+{
+	struct cache_set *c = container_of(cl, struct cache_set, moving_gc);
+	struct keybuf_key *w;
+	struct moving_io *io;
+	struct bio *bio;
+
+	/* XXX: if we error, background writeback could stall indefinitely */
+
+	while (!test_bit(CACHE_SET_STOPPING, &c->flags)) {
+		w = bch_keybuf_next_rescan(c, &c->moving_gc_keys, &MAX_KEY);
+		if (!w)
+			break;
+
+		io = kzalloc(sizeof(struct moving_io) + sizeof(struct bio_vec)
+			     * DIV_ROUND_UP(KEY_SIZE(&w->key), PAGE_SECTORS),
+			     GFP_KERNEL);
+		if (!io)
+			goto err;
+
+		w->private	= io;
+		io->w		= w;
+		io->s.op.inode	= KEY_INODE(&w->key);
+		io->s.op.c	= c;
+
+		moving_init(io);
+		bio = &io->bio.bio;
+
+		bio->bi_rw	= READ;
+		bio->bi_end_io	= read_moving_endio;
+
+		if (bch_bio_alloc_pages(bio, GFP_KERNEL))
+			goto err;
+
+		pr_debug("%s", pkey(&w->key));
+
+		closure_call(&io->s.cl, read_moving_submit, NULL, &c->gc.cl);
+
+		if (atomic_inc_return(&c->in_flight) >= 64) {
+			closure_wait_event(&c->moving_gc_wait, cl,
+					   atomic_read(&c->in_flight) < 64);
+			continue_at(cl, read_moving, bch_gc_wq);
+		}
+	}
+
+	if (0) {
+err:		if (!IS_ERR_OR_NULL(w->private))
+			kfree(w->private);
+
+		bch_keybuf_del(&c->moving_gc_keys, w);
+	}
+
+	closure_return(cl);
+}
+
+static bool bucket_cmp(struct bucket *l, struct bucket *r)
+{
+	return GC_SECTORS_USED(l) < GC_SECTORS_USED(r);
+}
+
+static unsigned bucket_heap_top(struct cache *ca)
+{
+	return GC_SECTORS_USED(heap_peek(&ca->heap));
+}
+
+void bch_moving_gc(struct closure *cl)
+{
+	struct cache_set *c = container_of(cl, struct cache_set, gc.cl);
+	struct cache *ca;
+	struct bucket *b;
+	unsigned i;
+
+	if (!c->copy_gc_enabled)
+		closure_return(cl);
+
+	mutex_lock(&c->bucket_lock);
+
+	for_each_cache(ca, c, i) {
+		unsigned sectors_to_move = 0;
+		unsigned reserve_sectors = ca->sb.bucket_size *
+			min(fifo_used(&ca->free), ca->free.size / 2);
+
+		ca->heap.used = 0;
+
+		for_each_bucket(b, ca) {
+			if (!GC_SECTORS_USED(b))
+				continue;
+
+			if (!heap_full(&ca->heap)) {
+				sectors_to_move += GC_SECTORS_USED(b);
+				heap_add(&ca->heap, b, bucket_cmp);
+			} else if (bucket_cmp(b, heap_peek(&ca->heap))) {
+				sectors_to_move -= bucket_heap_top(ca);
+				sectors_to_move += GC_SECTORS_USED(b);
+
+				ca->heap.data[0] = b;
+				heap_sift(&ca->heap, 0, bucket_cmp);
+			}
+		}
+
+		while (sectors_to_move > reserve_sectors) {
+			heap_pop(&ca->heap, b, bucket_cmp);
+			sectors_to_move -= GC_SECTORS_USED(b);
+		}
+
+		ca->gc_move_threshold = bucket_heap_top(ca);
+
+		pr_debug("threshold %u", ca->gc_move_threshold);
+	}
+
+	mutex_unlock(&c->bucket_lock);
+
+	c->moving_gc_keys.last_scanned = ZERO_KEY;
+
+	closure_init(&c->moving_gc, cl);
+	read_moving(&c->moving_gc);
+
+	closure_return(cl);
+}
+
+void bch_moving_init_cache_set(struct cache_set *c)
+{
+	bch_keybuf_init(&c->moving_gc_keys, moving_pred);
+}

drivers/md/bcache/request.c

···
+/*
+ * Main bcache entry point - handle a read or a write request and decide what to
+ * do with it; the make_request functions are called by the block layer.
+ *
+ * Copyright 2010, 2011 Kent Overstreet <[e-mail]>
+ * Copyright 2012 Google, Inc.
+ */
+
+#include "bcache.h"
+#include "btree.h"
+#include "debug.h"
+#include "request.h"
+
+#include <linux/cgroup.h>
+#include <linux/module.h>
+#include <linux/hash.h>
+#include <linux/random.h>
+#include "blk-cgroup.h"
+
+#include <trace/events/bcache.h>
+
+#define CUTOFF_CACHE_ADD	95
+#define CUTOFF_CACHE_READA	90
+#define CUTOFF_WRITEBACK	50
+#define CUTOFF_WRITEBACK_SYNC	75
+
+struct kmem_cache *bch_search_cache;
+
+static void check_should_skip(struct cached_dev *, struct search *);
+
+/* Cgroup interface */
+
+#ifdef CONFIG_CGROUP_BCACHE
+static struct bch_cgroup bcache_default_cgroup = { .cache_mode = -1 };
+
+static struct bch_cgroup *cgroup_to_bcache(struct cgroup *cgroup)
+{
+	struct cgroup_subsys_state *css;
+	return cgroup &&
+		(css = cgroup_subsys_state(cgroup, bcache_subsys_id))
+		? container_of(css, struct bch_cgroup, css)
+		: &bcache_default_cgroup;
+}
+
+struct bch_cgroup *bch_bio_to_cgroup(struct bio *bio)
+{
+	struct cgroup_subsys_state *css = bio->bi_css
+		? cgroup_subsys_state(bio->bi_css->cgroup, bcache_subsys_id)
+		: task_subsys_state(current, bcache_subsys_id);
+
+	return css
+		? container_of(css, struct bch_cgroup, css)
+		: &bcache_default_cgroup;
+}
+
+static ssize_t cache_mode_read(struct cgroup *cgrp, struct cftype *cft,
+			struct file *file,
+			char __user *buf, size_t nbytes, loff_t *ppos)
+{
+	char tmp[1024];
+	int len = bch_snprint_string_list(tmp, PAGE_SIZE, bch_cache_modes,
+					  cgroup_to_bcache(cgrp)->cache_mode + 1);
+
+	if (len < 0)
+		return len;
+
+	return simple_read_from_buffer(buf, nbytes, ppos, tmp, len);
+}
+
+static int cache_mode_write(struct cgroup *cgrp, struct cftype *cft,
+			    const char *buf)
+{
+	int v = bch_read_string_list(buf, bch_cache_modes);
+	if (v < 0)
+		return v;
+
+	cgroup_to_bcache(cgrp)->cache_mode = v - 1;
+	return 0;
+}
+
+static u64 bch_verify_read(struct cgroup *cgrp, struct cftype *cft)
+{
+	return cgroup_to_bcache(cgrp)->verify;
+}
+
+static int bch_verify_write(struct cgroup *cgrp, struct cftype *cft, u64 val)
+{
+	cgroup_to_bcache(cgrp)->verify = val;
+	return 0;
+}
+
+static u64 bch_cache_hits_read(struct cgroup *cgrp, struct cftype *cft)
+{
+	struct bch_cgroup *bcachecg = cgroup_to_bcache(cgrp);
+	return atomic_read(&bcachecg->stats.cache_hits);
+}
+
+static u64 bch_cache_misses_read(struct cgroup *cgrp, struct cftype *cft)
+{
+	struct bch_cgroup *bcachecg = cgroup_to_bcache(cgrp);
+	return atomic_read(&bcachecg->stats.cache_misses);
+}
+
+static u64 bch_cache_bypass_hits_read(struct cgroup *cgrp,
+					 struct cftype *cft)
+{
+	struct bch_cgroup *bcachecg = cgroup_to_bcache(cgrp);
+	return atomic_read(&bcachecg->stats.cache_bypass_hits);
+}
+
+static u64 bch_cache_bypass_misses_read(struct cgroup *cgrp,
+					   struct cftype *cft)
+{
+	struct bch_cgroup *bcachecg = cgroup_to_bcache(cgrp);
+	return atomic_read(&bcachecg->stats.cache_bypass_misses);
+}
+
+static struct cftype bch_files[] = {
+	{
+		.name		= "cache_mode",
+		.read		= cache_mode_read,
+		.write_string	= cache_mode_write,
+	},
+	{
+		.name		= "verify",
+		.read_u64	= bch_verify_read,
+		.write_u64	= bch_verify_write,
+	},
+	{
+		.name		= "cache_hits",
+		.read_u64	= bch_cache_hits_read,
+	},
+	{
+		.name		= "cache_misses",
+		.read_u64	= bch_cache_misses_read,
+	},
+	{
+		.name		= "cache_bypass_hits",
+		.read_u64	= bch_cache_bypass_hits_read,
+	},
+	{
+		.name		= "cache_bypass_misses",
+		.read_u64	= bch_cache_bypass_misses_read,
+	},
+	{ }	/* terminate */
+};
+
+static void init_bch_cgroup(struct bch_cgroup *cg)
+{
+	cg->cache_mode = -1;
+}
+
+static struct cgroup_subsys_state *bcachecg_create(struct cgroup *cgroup)
+{
+	struct bch_cgroup *cg;
+
+	cg = kzalloc(sizeof(*cg), GFP_KERNEL);
+	if (!cg)
+		return ERR_PTR(-ENOMEM);
+	init_bch_cgroup(cg);
+	return &cg->css;
+}
+
+static void bcachecg_destroy(struct cgroup *cgroup)
+{
+	struct bch_cgroup *cg = cgroup_to_bcache(cgroup);
+	free_css_id(&bcache_subsys, &cg->css);
+	kfree(cg);
+}
+
+struct cgroup_subsys bcache_subsys = {
+	.create		= bcachecg_create,
+	.destroy	= bcachecg_destroy,
+	.subsys_id	= bcache_subsys_id,
+	.name		= "bcache",
+	.module		= THIS_MODULE,
+};
+EXPORT_SYMBOL_GPL(bcache_subsys);
+#endif
+
+static unsigned cache_mode(struct cached_dev *dc, struct bio *bio)
+{
+#ifdef CONFIG_CGROUP_BCACHE
+	int r = bch_bio_to_cgroup(bio)->cache_mode;
+	if (r >= 0)
+		return r;
+#endif
+	return BDEV_CACHE_MODE(&dc->sb);
+}
+
+static bool verify(struct cached_dev *dc, struct bio *bio)
+{
+#ifdef CONFIG_CGROUP_BCACHE
+	if (bch_bio_to_cgroup(bio)->verify)
+		return true;
+#endif
+	return dc->verify;
+}
+
+static void bio_csum(struct bio *bio, struct bkey *k)
+{
+	struct bio_vec *bv;
+	uint64_t csum = 0;
+	int i;
+
+	bio_for_each_segment(bv, bio, i) {
+		void *d = kmap(bv->bv_page) + bv->bv_offset;
+		csum = bch_crc64_update(csum, d, bv->bv_len);
+		kunmap(bv->bv_page);
+	}
+
+	k->ptr[KEY_PTRS(k)] = csum & (~0ULL >> 1);
+}
+
+/* Insert data into cache */
+
+static void bio_invalidate(struct closure *cl)
+{
+	struct btree_op *op = container_of(cl, struct btree_op, cl);
+	struct bio *bio = op->cache_bio;
+
+	pr_debug("invalidating %i sectors from %llu",
+		 bio_sectors(bio), (uint64_t) bio->bi_sector);
+
+	while (bio_sectors(bio)) {
+		unsigned len = min(bio_sectors(bio), 1U << 14);
+
+		if (bch_keylist_realloc(&op->keys, 0, op->c))
+			goto out;
+
+		bio->bi_sector	+= len;
+		bio->bi_size	-= len << 9;
+
+		bch_keylist_add(&op->keys,
+				&KEY(op->inode, bio->bi_sector, len));
+	}
+
+	op->insert_data_done = true;
+	bio_put(bio);
+out:
+	continue_at(cl, bch_journal, bcache_wq);
+}
+
+struct open_bucket {
+	struct list_head	list;
+	struct task_struct	*last;
+	unsigned		sectors_free;
+	BKEY_PADDED(key);
+};
+
+void bch_open_buckets_free(struct cache_set *c)
+{
+	struct open_bucket *b;
+
+	while (!list_empty(&c->data_buckets)) {
+		b = list_first_entry(&c->data_buckets,
+				     struct open_bucket, list);
+		list_del(&b->list);
+		kfree(b);
+	}
+}
+
+int bch_open_buckets_alloc(struct cache_set *c)
+{
+	int i;
+
+	spin_lock_init(&c->data_bucket_lock);
+
+	for (i = 0; i < 6; i++) {
+		struct open_bucket *b = kzalloc(sizeof(*b), GFP_KERNEL);
+		if (!b)
+			return -ENOMEM;
+
+		list_add(&b->list, &c->data_buckets);
+	}
+
+	return 0;
+}
+
+/*
+ * We keep multiple buckets open for writes, and try to segregate different
+ * write streams for better cache utilization: first we look for a bucket where
+ * the last write to it was sequential with the current write, and failing that
+ * we look for a bucket that was last used by the same task.
+ *
+ * The ideas is if you've got multiple tasks pulling data into the cache at the
+ * same time, you'll get better cache utilization if you try to segregate their
+ * data and preserve locality.
+ *
+ * For example, say you've starting Firefox at the same time you're copying a
+ * bunch of files. Firefox will likely end up being fairly hot and stay in the
+ * cache awhile, but the data you copied might not be; if you wrote all that
+ * data to the same buckets it'd get invalidated at the same time.
+ *
+ * Both of those tasks will be doing fairly random IO so we can't rely on
+ * detecting sequential IO to segregate their data, but going off of the task
+ * should be a sane heuristic.
+ */
+static struct open_bucket *pick_data_bucket(struct cache_set *c,
+					    const struct bkey *search,
+					    struct task_struct *task,
+					    struct bkey *alloc)
+{
+	struct open_bucket *ret, *ret_task = NULL;
+
+	list_for_each_entry_reverse(ret, &c->data_buckets, list)
+		if (!bkey_cmp(&ret->key, search))
+			goto found;
+		else if (ret->last == task)
+			ret_task = ret;
+
+	ret = ret_task ?: list_first_entry(&c->data_buckets,
+					   struct open_bucket, list);
+found:
+	if (!ret->sectors_free && KEY_PTRS(alloc)) {
+		ret->sectors_free = c->sb.bucket_size;
+		bkey_copy(&ret->key, alloc);
+		bkey_init(alloc);
+	}
+
+	if (!ret->sectors_free)
+		ret = NULL;
+
+	return ret;
+}
+
+/*
+ * Allocates some space in the cache to write to, and k to point to the newly
+ * allocated space, and updates KEY_SIZE(k) and KEY_OFFSET(k) (to point to the
+ * end of the newly allocated space).
+ *
+ * May allocate fewer sectors than @sectors, KEY_SIZE(k) indicates how many
+ * sectors were actually allocated.
+ *
+ * If s->writeback is true, will not fail.
+ */
+static bool bch_alloc_sectors(struct bkey *k, unsigned sectors,
+			      struct search *s)
+{
+	struct cache_set *c = s->op.c;
+	struct open_bucket *b;
+	BKEY_PADDED(key) alloc;
+	struct closure cl, *w = NULL;
+	unsigned i;
+
+	if (s->writeback) {
+		closure_init_stack(&cl);
+		w = &cl;
+	}
+
+	/*
+	 * We might have to allocate a new bucket, which we can't do with a
+	 * spinlock held. So if we have to allocate, we drop the lock, allocate
+	 * and then retry. KEY_PTRS() indicates whether alloc points to
+	 * allocated bucket(s).
+	 */
+
+	bkey_init(&alloc.key);
+	spin_lock(&c->data_bucket_lock);
+
+	while (!(b = pick_data_bucket(c, k, s->task, &alloc.key))) {
+		unsigned watermark = s->op.write_prio
+			? WATERMARK_MOVINGGC
+			: WATERMARK_NONE;
+
+		spin_unlock(&c->data_bucket_lock);
+
+		if (bch_bucket_alloc_set(c, watermark, &alloc.key, 1, w))
+			return false;
+
+		spin_lock(&c->data_bucket_lock);
+	}
+
+	/*
+	 * If we had to allocate, we might race and not need to allocate the
+	 * second time we call find_data_bucket(). If we allocated a bucket but
+	 * didn't use it, drop the refcount bch_bucket_alloc_set() took:
+	 */
+	if (KEY_PTRS(&alloc.key))
+		__bkey_put(c, &alloc.key);
+
+	for (i = 0; i < KEY_PTRS(&b->key); i++)
+		EBUG_ON(ptr_stale(c, &b->key, i));
+
+	/* Set up the pointer to the space we're allocating: */
+
+	for (i = 0; i < KEY_PTRS(&b->key); i++)
+		k->ptr[i] = b->key.ptr[i];
+
+	sectors = min(sectors, b->sectors_free);
+
+	SET_KEY_OFFSET(k, KEY_OFFSET(k) + sectors);
+	SET_KEY_SIZE(k, sectors);
+	SET_KEY_PTRS(k, KEY_PTRS(&b->key));
+
+	/*
+	 * Move b to the end of the lru, and keep track of what this bucket was
+	 * last used for:
+	 */
+	list_move_tail(&b->list, &c->data_buckets);
+	bkey_copy_key(&b->key, k);
+	b->last = s->task;
+
+	b->sectors_free	-= sectors;
+
+	for (i = 0; i < KEY_PTRS(&b->key); i++) {
+		SET_PTR_OFFSET(&b->key, i, PTR_OFFSET(&b->key, i) + sectors);
+
+		atomic_long_add(sectors,
+				&PTR_CACHE(c, &b->key, i)->sectors_written);
+	}
+
+	if (b->sectors_free < c->sb.block_size)
+		b->sectors_free = 0;
+
+	/*
+	 * k takes refcounts on the buckets it points to until it's inserted
+	 * into the btree, but if we're done with this bucket we just transfer
+	 * get_data_bucket()'s refcount.
+	 */
+	if (b->sectors_free)
+		for (i = 0; i < KEY_PTRS(&b->key); i++)
+			atomic_inc(&PTR_BUCKET(c, &b->key, i)->pin);
+
+	spin_unlock(&c->data_bucket_lock);
+	return true;
+}
+
+static void bch_insert_data_error(struct closure *cl)
+{
+	struct btree_op *op = container_of(cl, struct btree_op, cl);
+
+	/*
+	 * Our data write just errored, which means we've got a bunch of keys to
+	 * insert that point to data that wasn't succesfully written.
+	 *
+	 * We don't have to insert those keys but we still have to invalidate
+	 * that region of the cache - so, if we just strip off all the pointers
+	 * from the keys we'll accomplish just that.
+	 */
+
+	struct bkey *src = op->keys.bottom, *dst = op->keys.bottom;
+
+	while (src != op->keys.top) {
+		struct bkey *n = bkey_next(src);
+
+		SET_KEY_PTRS(src, 0);
+		bkey_copy(dst, src);
+
+		dst = bkey_next(dst);
+		src = n;
+	}
+
+	op->keys.top = dst;
+
+	bch_journal(cl);
+}
+
+static void bch_insert_data_endio(struct bio *bio, int error)
+{
+	struct closure *cl = bio->bi_private;
+	struct btree_op *op = container_of(cl, struct btree_op, cl);
+	struct search *s = container_of(op, struct search, op);
+
+	if (error) {
+		/* TODO: We could try to recover from this. */
+		if (s->writeback)
+			s->error = error;
+		else if (s->write)
+			set_closure_fn(cl, bch_insert_data_error, bcache_wq);
+		else
+			set_closure_fn(cl, NULL, NULL);
+	}
+
+	bch_bbio_endio(op->c, bio, error, "writing data to cache");
+}
+
+static void bch_insert_data_loop(struct closure *cl)
+{
+	struct btree_op *op = container_of(cl, struct btree_op, cl);
+	struct search *s = container_of(op, struct search, op);
+	struct bio *bio = op->cache_bio, *n;
+
+	if (op->skip)
+		return bio_invalidate(cl);
+
+	if (atomic_sub_return(bio_sectors(bio), &op->c->sectors_to_gc) < 0) {
+		set_gc_sectors(op->c);
+		bch_queue_gc(op->c);
+	}
+
+	do {
+		unsigned i;
+		struct bkey *k;
+		struct bio_set *split = s->d
+			? s->d->bio_split : op->c->bio_split;
+
+		/* 1 for the device pointer and 1 for the chksum */
+		if (bch_keylist_realloc(&op->keys,
+					1 + (op->csum ? 1 : 0),
+					op->c))
+			continue_at(cl, bch_journal, bcache_wq);
+
+		k = op->keys.top;
+		bkey_init(k);
+		SET_KEY_INODE(k, op->inode);
+		SET_KEY_OFFSET(k, bio->bi_sector);
+
+		if (!bch_alloc_sectors(k, bio_sectors(bio), s))
+			goto err;
+
+		n = bch_bio_split(bio, KEY_SIZE(k), GFP_NOIO, split);
+		if (!n) {
+			__bkey_put(op->c, k);
+			continue_at(cl, bch_insert_data_loop, bcache_wq);
+		}
+
+		n->bi_end_io	= bch_insert_data_endio;
+		n->bi_private	= cl;
+
+		if (s->writeback) {
+			SET_KEY_DIRTY(k, true);
+
+			for (i = 0; i < KEY_PTRS(k); i++)
+				SET_GC_MARK(PTR_BUCKET(op->c, k, i),
+					    GC_MARK_DIRTY);
+		}
+
+		SET_KEY_CSUM(k, op->csum);
+		if (KEY_CSUM(k))
+			bio_csum(n, k);
+
+		pr_debug("%s", pkey(k));
+		bch_keylist_push(&op->keys);
+
+		trace_bcache_cache_insert(n, n->bi_sector, n->bi_bdev);
+		n->bi_rw |= REQ_WRITE;
+		bch_submit_bbio(n, op->c, k, 0);
+	} while (n != bio);
+
+	op->insert_data_done = true;
+	continue_at(cl, bch_journal, bcache_wq);
+err:
+	/* bch_alloc_sectors() blocks if s->writeback = true */
+	BUG_ON(s->writeback);
+
+	/*
+	 * But if it's not a writeback write we'd rather just bail out if
+	 * there aren't any buckets ready to write to - it might take awhile and
+	 * we might be starving btree writes for gc or something.
+	 */
+
+	if (s->write) {
+		/*
+		 * Writethrough write: We can't complete the write until we've
+		 * updated the index. But we don't want to delay the write while
+		 * we wait for buckets to be freed up, so just invalidate the
+		 * rest of the write.
+		 */
+		op->skip = true;
+		return bio_invalidate(cl);
+	} else {
+		/*
+		 * From a cache miss, we can just insert the keys for the data
+		 * we have written or bail out if we didn't do anything.
+		 */
+		op->insert_data_done = true;
+		bio_put(bio);
+
+		if (!bch_keylist_empty(&op->keys))
+			continue_at(cl, bch_journal, bcache_wq);
+		else
+			closure_return(cl);
+	}
+}
+
+/**
+ * bch_insert_data - stick some data in the cache
+ *
+ * This is the starting point for any data to end up in a cache device; it could
+ * be from a normal write, or a writeback write, or a write to a flash only
+ * volume - it's also used by the moving garbage collector to compact data in
+ * mostly empty buckets.
+ *
+ * It first writes the data to the cache, creating a list of keys to be inserted
+ * (if the data had to be fragmented there will be multiple keys); after the
+ * data is written it calls bch_journal, and after the keys have been added to
+ * the next journal write they're inserted into the btree.
+ *
+ * It inserts the data in op->cache_bio; bi_sector is used for the key offset,
+ * and op->inode is used for the key inode.
+ *
+ * If op->skip is true, instead of inserting the data it invalidates the region
+ * of the cache represented by op->cache_bio and op->inode.
+ */
+void bch_insert_data(struct closure *cl)
+{
+	struct btree_op *op = container_of(cl, struct btree_op, cl);
+
+	bch_keylist_init(&op->keys);
+	bio_get(op->cache_bio);
+	bch_insert_data_loop(cl);
+}
+
+void bch_btree_insert_async(struct closure *cl)
+{
+	struct btree_op *op = container_of(cl, struct btree_op, cl);
+	struct search *s = container_of(op, struct search, op);
+
+	if (bch_btree_insert(op, op->c)) {
+		s->error		= -ENOMEM;
+		op->insert_data_done	= true;
+	}
+
+	if (op->insert_data_done) {
+		bch_keylist_free(&op->keys);
+		closure_return(cl);
+	} else
+		continue_at(cl, bch_insert_data_loop, bcache_wq);
+}
+
+/* Common code for the make_request functions */
+
+static void request_endio(struct bio *bio, int error)
+{
+	struct closure *cl = bio->bi_private;
+
+	if (error) {
+		struct search *s = container_of(cl, struct search, cl);
+		s->error = error;
+		/* Only cache read errors are recoverable */
+		s->recoverable = false;
+	}
+
+	bio_put(bio);
+	closure_put(cl);
+}
+
+void bch_cache_read_endio(struct bio *bio, int error)
+{
+	struct bbio *b = container_of(bio, struct bbio, bio);
+	struct closure *cl = bio->bi_private;
+	struct search *s = container_of(cl, struct search, cl);
+
+	/*
+	 * If the bucket was reused while our bio was in flight, we might have
+	 * read the wrong data. Set s->error but not error so it doesn't get
+	 * counted against the cache device, but we'll still reread the data
+	 * from the backing device.
+	 */
+
+	if (error)
+		s->error = error;
+	else if (ptr_stale(s->op.c, &b->key, 0)) {
+		atomic_long_inc(&s->op.c->cache_read_races);
+		s->error = -EINTR;
+	}
+
+	bch_bbio_endio(s->op.c, bio, error, "reading from cache");
+}
+
+static void bio_complete(struct search *s)
+{
+	if (s->orig_bio) {
+		int cpu, rw = bio_data_dir(s->orig_bio);
+		unsigned long duration = jiffies - s->start_time;
+
+		cpu = part_stat_lock();
+		part_round_stats(cpu, &s->d->disk->part0);
+		part_stat_add(cpu, &s->d->disk->part0, ticks[rw], duration);
+		part_stat_unlock();
+
+		trace_bcache_request_end(s, s->orig_bio);
+		bio_endio(s->orig_bio, s->error);
+		s->orig_bio = NULL;
+	}
+}
+
+static void do_bio_hook(struct search *s)
+{
+	struct bio *bio = &s->bio.bio;
+	memcpy(bio, s->orig_bio, sizeof(struct bio));
+
+	bio->bi_end_io		= request_endio;
+	bio->bi_private		= &s->cl;
+	atomic_set(&bio->bi_cnt, 3);
+}
+
+static void search_free(struct closure *cl)
+{
+	struct search *s = container_of(cl, struct search, cl);
+	bio_complete(s);
+
+	if (s->op.cache_bio)
+		bio_put(s->op.cache_bio);
+
+	if (s->unaligned_bvec)
+		mempool_free(s->bio.bio.bi_io_vec, s->d->unaligned_bvec);
+
+	closure_debug_destroy(cl);
+	mempool_free(s, s->d->c->search);
+}
+
+static struct search *search_alloc(struct bio *bio, struct bcache_device *d)
+{
+	struct bio_vec *bv;
+	struct search *s = mempool_alloc(d->c->search, GFP_NOIO);
+	memset(s, 0, offsetof(struct search, op.keys));
+
+	__closure_init(&s->cl, NULL);
+
+	s->op.inode		= d->id;
+	s->op.c			= d->c;
+	s->d			= d;
+	s->op.lock		= -1;
+	s->task			= current;
+	s->orig_bio		= bio;
+	s->write		= (bio->bi_rw & REQ_WRITE) != 0;
+	s->op.flush_journal	= (bio->bi_rw & REQ_FLUSH) != 0;
+	s->op.skip		= (bio->bi_rw & REQ_DISCARD) != 0;
+	s->recoverable		= 1;
+	s->start_time		= jiffies;
+	do_bio_hook(s);
+
+	if (bio->bi_size != bio_segments(bio) * PAGE_SIZE) {
+		bv = mempool_alloc(d->unaligned_bvec, GFP_NOIO);
+		memcpy(bv, bio_iovec(bio),
+		       sizeof(struct bio_vec) * bio_segments(bio));
+
+		s->bio.bio.bi_io_vec	= bv;
+		s->unaligned_bvec	= 1;
+	}
+
+	return s;
+}
+
+static void btree_read_async(struct closure *cl)
+{
+	struct btree_op *op = container_of(cl, struct btree_op, cl);
+
+	int ret = btree_root(search_recurse, op->c, op);
+
+	if (ret == -EAGAIN)
+		continue_at(cl, btree_read_async, bcache_wq);
+
+	closure_return(cl);
+}
+
+/* Cached devices */
+
+static void cached_dev_bio_complete(struct closure *cl)
+{
+	struct search *s = container_of(cl, struct search, cl);
+	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
+
+	search_free(cl);
+	cached_dev_put(dc);
+}
+
+/* Process reads */
+
+static void cached_dev_read_complete(struct closure *cl)
+{
+	struct search *s = container_of(cl, struct search, cl);
+
+	if (s->op.insert_collision)
+		bch_mark_cache_miss_collision(s);
+
+	if (s->op.cache_bio) {
+		int i;
+		struct bio_vec *bv;
+
+		__bio_for_each_segment(bv, s->op.cache_bio, i, 0)
+			__free_page(bv->bv_page);
+	}
+
+	cached_dev_bio_complete(cl);
+}
+
+static void request_read_error(struct closure *cl)
+{
+	struct search *s = container_of(cl, struct search, cl);
+	struct bio_vec *bv;
+	int i;
+
+	if (s->recoverable) {
+		/* The cache read failed, but we can retry from the backing
+		 * device.
+		 */
+		pr_debug("recovering at sector %llu",
+			 (uint64_t) s->orig_bio->bi_sector);
+
+		s->error = 0;
+		bv = s->bio.bio.bi_io_vec;
+		do_bio_hook(s);
+		s->bio.bio.bi_io_vec = bv;
+
+		if (!s->unaligned_bvec)
+			bio_for_each_segment(bv, s->orig_bio, i)
+				bv->bv_offset = 0, bv->bv_len = PAGE_SIZE;
+		else
+			memcpy(s->bio.bio.bi_io_vec,
+			       bio_iovec(s->orig_bio),
+			       sizeof(struct bio_vec) *
+			       bio_segments(s->orig_bio));
+
+		/* XXX: invalidate cache */
+
+		trace_bcache_read_retry(&s->bio.bio);
+		closure_bio_submit(&s->bio.bio, &s->cl, s->d);
+	}
+
+	continue_at(cl, cached_dev_read_complete, NULL);
+}
+
+static void request_read_done(struct closure *cl)
+{
+	struct search *s = container_of(cl, struct search, cl);
+	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
+
+	/*
+	 * s->cache_bio != NULL implies that we had a cache miss; cache_bio now
+	 * contains data ready to be inserted into the cache.
+	 *
+	 * First, we copy the data we just read from cache_bio's bounce buffers
+	 * to the buffers the original bio pointed to:
+	 */
+
+	if (s->op.cache_bio) {
+		struct bio_vec *src, *dst;
+		unsigned src_offset, dst_offset, bytes;
+		void *dst_ptr;
+
+		bio_reset(s->op.cache_bio);
+		s->op.cache_bio->bi_sector	= s->cache_miss->bi_sector;
+		s->op.cache_bio->bi_bdev	= s->cache_miss->bi_bdev;
+		s->op.cache_bio->bi_size	= s->cache_bio_sectors << 9;
+		bch_bio_map(s->op.cache_bio, NULL);
+
+		src = bio_iovec(s->op.cache_bio);
+		dst = bio_iovec(s->cache_miss);
+		src_offset = src->bv_offset;
+		dst_offset = dst->bv_offset;
+		dst_ptr = kmap(dst->bv_page);
+
+		while (1) {
+			if (dst_offset == dst->bv_offset + dst->bv_len) {
+				kunmap(dst->bv_page);
+				dst++;
+				if (dst == bio_iovec_idx(s->cache_miss,
+						s->cache_miss->bi_vcnt))
+					break;
+
+				dst_offset = dst->bv_offset;
+				dst_ptr = kmap(dst->bv_page);
+			}
+
+			if (src_offset == src->bv_offset + src->bv_len) {
+				src++;
+				if (src == bio_iovec_idx(s->op.cache_bio,
+						 s->op.cache_bio->bi_vcnt))
+					BUG();
+
+				src_offset = src->bv_offset;
+			}
+
+			bytes = min(dst->bv_offset + dst->bv_len - dst_offset,
+				    src->bv_offset + src->bv_len - src_offset);
+
+			memcpy(dst_ptr + dst_offset,
+			       page_address(src->bv_page) + src_offset,
+			       bytes);
+
+			src_offset	+= bytes;
+			dst_offset	+= bytes;
+		}
+
+		bio_put(s->cache_miss);
+		s->cache_miss = NULL;
+	}
+
+	if (verify(dc, &s->bio.bio) && s->recoverable)
+		bch_data_verify(s);
+
+	bio_complete(s);
+
+	if (s->op.cache_bio &&
+	    !test_bit(CACHE_SET_STOPPING, &s->op.c->flags)) {
+		s->op.type = BTREE_REPLACE;
+		closure_call(&s->op.cl, bch_insert_data, NULL, cl);
+	}
+
+	continue_at(cl, cached_dev_read_complete, NULL);
+}
+
+static void request_read_done_bh(struct closure *cl)
+{
+	struct search *s = container_of(cl, struct search, cl);
+	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
+
+	bch_mark_cache_accounting(s, !s->cache_miss, s->op.skip);
+
+	if (s->error)
+		continue_at_nobarrier(cl, request_read_error, bcache_wq);
+	else if (s->op.cache_bio || verify(dc, &s->bio.bio))
+		continue_at_nobarrier(cl, request_read_done, bcache_wq);
+	else
+		continue_at_nobarrier(cl, cached_dev_read_complete, NULL);
+}
+
+static int cached_dev_cache_miss(struct btree *b, struct search *s,
+				 struct bio *bio, unsigned sectors)
+{
+	int ret = 0;
+	unsigned reada;
+	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
+	struct bio *miss;
+
+	miss = bch_bio_split(bio, sectors, GFP_NOIO, s->d->bio_split);
+	if (!miss)
+		return -EAGAIN;
+
+	if (miss == bio)
+		s->op.lookup_done = true;
+
+	miss->bi_end_io		= request_endio;
+	miss->bi_private	= &s->cl;
+
+	if (s->cache_miss || s->op.skip)
+		goto out_submit;
+
+	if (miss != bio ||
+	    (bio->bi_rw & REQ_RAHEAD) ||
+	    (bio->bi_rw & REQ_META) ||
+	    s->op.c->gc_stats.in_use >= CUTOFF_CACHE_READA)
+		reada = 0;
+	else {
+		reada = min(dc->readahead >> 9,
+			    sectors - bio_sectors(miss));
+
+		if (bio_end(miss) + reada > bdev_sectors(miss->bi_bdev))
+			reada = bdev_sectors(miss->bi_bdev) - bio_end(miss);
+	}
+
+	s->cache_bio_sectors = bio_sectors(miss) + reada;
+	s->op.cache_bio = bio_alloc_bioset(GFP_NOWAIT,
+			DIV_ROUND_UP(s->cache_bio_sectors, PAGE_SECTORS),
+			dc->disk.bio_split);
+
+	if (!s->op.cache_bio)
+		goto out_submit;
+
+	s->op.cache_bio->bi_sector	= miss->bi_sector;
+	s->op.cache_bio->bi_bdev	= miss->bi_bdev;
+	s->op.cache_bio->bi_size	= s->cache_bio_sectors << 9;
+
+	s->op.cache_bio->bi_end_io	= request_endio;
+	s->op.cache_bio->bi_private	= &s->cl;
+
+	/* btree_search_recurse()'s btree iterator is no good anymore */
+	ret = -EINTR;
+	if (!bch_btree_insert_check_key(b, &s->op, s->op.cache_bio))
+		goto out_put;
+
+	bch_bio_map(s->op.cache_bio, NULL);
+	if (bch_bio_alloc_pages(s->op.cache_bio, __GFP_NOWARN|GFP_NOIO))
+		goto out_put;
+
+	s->cache_miss = miss;
+	bio_get(s->op.cache_bio);
+
+	trace_bcache_cache_miss(s->orig_bio);
+	closure_bio_submit(s->op.cache_bio, &s->cl, s->d);
+
+	return ret;
+out_put:
+	bio_put(s->op.cache_bio);
+	s->op.cache_bio = NULL;
+out_submit:
+	closure_bio_submit(miss, &s->cl, s->d);
+	return ret;
+}
+
+static void request_read(struct cached_dev *dc, struct search *s)
+{
+	struct closure *cl = &s->cl;
+
+	check_should_skip(dc, s);
+	closure_call(&s->op.cl, btree_read_async, NULL, cl);
+
+	continue_at(cl, request_read_done_bh, NULL);
+}
+
+/* Process writes */
+
+static void cached_dev_write_complete(struct closure *cl)
+{
+	struct search *s = container_of(cl, struct search, cl);
+	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
+
+	up_read_non_owner(&dc->writeback_lock);
+	cached_dev_bio_complete(cl);
+}
+
+static bool should_writeback(struct cached_dev *dc, struct bio *bio)
+{
+	unsigned threshold = (bio->bi_rw & REQ_SYNC)
+		? CUTOFF_WRITEBACK_SYNC
+		: CUTOFF_WRITEBACK;
+
+	return !atomic_read(&dc->disk.detaching) &&
+		cache_mode(dc, bio) == CACHE_MODE_WRITEBACK &&
+		dc->disk.c->gc_stats.in_use < threshold;
+}
+
+static void request_write(struct cached_dev *dc, struct search *s)
+{
+	struct closure *cl = &s->cl;
+	struct bio *bio = &s->bio.bio;
+	struct bkey start, end;
+	start = KEY(dc->disk.id, bio->bi_sector, 0);
+	end = KEY(dc->disk.id, bio_end(bio), 0);
+
+	bch_keybuf_check_overlapping(&s->op.c->moving_gc_keys, &start, &end);
+
+	check_should_skip(dc, s);
+	down_read_non_owner(&dc->writeback_lock);
+
+	if (bch_keybuf_check_overlapping(&dc->writeback_keys, &start, &end)) {
+		s->op.skip	= false;
+		s->writeback	= true;
+	}
+
+	if (bio->bi_rw & REQ_DISCARD)
+		goto skip;
+
+	if (s->op.skip)
+		goto skip;
+
+	if (should_writeback(dc, s->orig_bio))
+		s->writeback = true;
+
+	if (!s->writeback) {
+		s->op.cache_bio = bio_clone_bioset(bio, GFP_NOIO,
+						   dc->disk.bio_split);
+
+		trace_bcache_writethrough(s->orig_bio);
+		closure_bio_submit(bio, cl, s->d);
+	} else {
+		s->op.cache_bio = bio;
+		trace_bcache_writeback(s->orig_bio);
+		bch_writeback_add(dc, bio_sectors(bio));
+	}
+out:
+	closure_call(&s->op.cl, bch_insert_data, NULL, cl);
+	continue_at(cl, cached_dev_write_complete, NULL);
+skip:
+	s->op.skip = true;
+	s->op.cache_bio = s->orig_bio;
+	bio_get(s->op.cache_bio);
+	trace_bcache_write_skip(s->orig_bio);
+
+	if ((bio->bi_rw & REQ_DISCARD) &&
+	    !blk_queue_discard(bdev_get_queue(dc->bdev)))
+		goto out;
+
+	closure_bio_submit(bio, cl, s->d);
+	goto out;
+}
+
+static void request_nodata(struct cached_dev *dc, struct search *s)
+{
+	struct closure *cl = &s->cl;
+	struct bio *bio = &s->bio.bio;
+
+	if (bio->bi_rw & REQ_DISCARD) {
+		request_write(dc, s);
+		return;
+	}
+
+	if (s->op.flush_journal)
+		bch_journal_meta(s->op.c, cl);
+
+	closure_bio_submit(bio, cl, s->d);
+
+	continue_at(cl, cached_dev_bio_complete, NULL);
+}
+
+/* Cached devices - read & write stuff */
+
+int bch_get_congested(struct cache_set *c)
+{
+	int i;
+
+	if (!c->congested_read_threshold_us &&
+	    !c->congested_write_threshold_us)
+		return 0;
+
+	i = (local_clock_us() - c->congested_last_us) / 1024;
+	if (i < 0)
+		return 0;
+
+	i += atomic_read(&c->congested);
+	if (i >= 0)
+		return 0;
+
+	i += CONGESTED_MAX;
+
+	return i <= 0 ? 1 : fract_exp_two(i, 6);
+}
+
+static void add_sequential(struct task_struct *t)
+{
+	ewma_add(t->sequential_io_avg,
+		 t->sequential_io, 8, 0);
+
+	t->sequential_io = 0;
+}
+
+static struct hlist_head *iohash(struct cached_dev *dc, uint64_t k)
+{
+	return &dc->io_hash[hash_64(k, RECENT_IO_BITS)];
+}
+
+static void check_should_skip(struct cached_dev *dc, struct search *s)
+{
+	struct cache_set *c = s->op.c;
+	struct bio *bio = &s->bio.bio;
+
+	long rand;
+	int cutoff = bch_get_congested(c);
+	unsigned mode = cache_mode(dc, bio);
+
+	if (atomic_read(&dc->disk.detaching) ||
+	    c->gc_stats.in_use > CUTOFF_CACHE_ADD ||
+	    (bio->bi_rw & REQ_DISCARD))
+		goto skip;
+
+	if (mode == CACHE_MODE_NONE ||
+	    (mode == CACHE_MODE_WRITEAROUND &&
+	     (bio->bi_rw & REQ_WRITE)))
+		goto skip;
+
+	if (bio->bi_sector   & (c->sb.block_size - 1) ||
+	    bio_sectors(bio) & (c->sb.block_size - 1)) {
+		pr_debug("skipping unaligned io");
+		goto skip;
+	}
+
+	if (!cutoff) {
+		cutoff = dc->sequential_cutoff >> 9;
+
+		if (!cutoff)
+			goto rescale;
+
+		if (mode == CACHE_MODE_WRITEBACK &&
+		    (bio->bi_rw & REQ_WRITE) &&
+		    (bio->bi_rw & REQ_SYNC))
+			goto rescale;
+	}
+
+	if (dc->sequential_merge) {
+		struct io *i;
+
+		spin_lock(&dc->io_lock);
+
+		hlist_for_each_entry(i, iohash(dc, bio->bi_sector), hash)
+			if (i->last == bio->bi_sector &&
+			    time_before(jiffies, i->jiffies))
+				goto found;
+
+		i = list_first_entry(&dc->io_lru, struct io, lru);
+
+		add_sequential(s->task);
+		i->sequential = 0;
+found:
+		if (i->sequential + bio->bi_size > i->sequential)
+			i->sequential	+= bio->bi_size;
+
+		i->last			 = bio_end(bio);
+		i->jiffies		 = jiffies + msecs_to_jiffies(5000);
+		s->task->sequential_io	 = i->sequential;
+
+		hlist_del(&i->hash);
+		hlist_add_head(&i->hash, iohash(dc, i->last));
+		list_move_tail(&i->lru, &dc->io_lru);
+
+		spin_unlock(&dc->io_lock);
+	} else {
+		s->task->sequential_io = bio->bi_size;
+
+		add_sequential(s->task);
+	}
+
+	rand = get_random_int();
+	cutoff -= bitmap_weight(&rand, BITS_PER_LONG);
+
+	if (cutoff <= (int) (max(s->task->sequential_io,
+				 s->task->sequential_io_avg) >> 9))
+		goto skip;
+
+rescale:
+	bch_rescale_priorities(c, bio_sectors(bio));
+	return;
+skip:
+	bch_mark_sectors_bypassed(s, bio_sectors(bio));
+	s->op.skip = true;
+}
+
+static void cached_dev_make_request(struct request_queue *q, struct bio *bio)
+{
+	struct search *s;
+	struct bcache_device *d = bio->bi_bdev->bd_disk->private_data;
+	struct cached_dev *dc = container_of(d, struct cached_dev, disk);
+	int cpu, rw = bio_data_dir(bio);
+
+	cpu = part_stat_lock();
+	part_stat_inc(cpu, &d->disk->part0, ios[rw]);
+	part_stat_add(cpu, &d->disk->part0, sectors[rw], bio_sectors(bio));
+	part_stat_unlock();
+
+	bio->bi_bdev = dc->bdev;
+	bio->bi_sector += dc->sb.data_offset;
+
+	if (cached_dev_get(dc)) {
+		s = search_alloc(bio, d);
+		trace_bcache_request_start(s, bio);
+
+		if (!bio_has_data(bio))
+			request_nodata(dc, s);
+		else if (rw)
+			request_write(dc, s);
+		else
+			request_read(dc, s);
+	} else {
+		if ((bio->bi_rw & REQ_DISCARD) &&
+		    !blk_queue_discard(bdev_get_queue(dc->bdev)))
+			bio_endio(bio, 0);
+		else
+			bch_generic_make_request(bio, &d->bio_split_hook);
+	}
+}
+
+static int cached_dev_ioctl(struct bcache_device *d, fmode_t mode,
+			    unsigned int cmd, unsigned long arg)
+{
+	struct cached_dev *dc = container_of(d, struct cached_dev, disk);
+	return __blkdev_driver_ioctl(dc->bdev, mode, cmd, arg);
+}
+
+static int cached_dev_congested(void *data, int bits)
+{
+	struct bcache_device *d = data;
+	struct cached_dev *dc = container_of(d, struct cached_dev, disk);
+	struct request_queue *q = bdev_get_queue(dc->bdev);
+	int ret = 0;
+
+	if (bdi_congested(&q->backing_dev_info, bits))
+		return 1;
+
+	if (cached_dev_get(dc)) {
+		unsigned i;
+		struct cache *ca;
+
+		for_each_cache(ca, d->c, i) {
+			q = bdev_get_queue(ca->bdev);
+			ret |= bdi_congested(&q->backing_dev_info, bits);
+		}
+
+		cached_dev_put(dc);
+	}
+
+	return ret;
+}
+
+void bch_cached_dev_request_init(struct cached_dev *dc)
+{
+	struct gendisk *g = dc->disk.disk;
+
+	g->queue->make_request_fn		= cached_dev_make_request;
+	g->queue->backing_dev_info.congested_fn = cached_dev_congested;
+	dc->disk.cache_miss			= cached_dev_cache_miss;
+	dc->disk.ioctl				= cached_dev_ioctl;
+}
+
+/* Flash backed devices */
+
+static int flash_dev_cache_miss(struct btree *b, struct search *s,
+				struct bio *bio, unsigned sectors)
+{
+	/* Zero fill bio */
+
+	while (bio->bi_idx != bio->bi_vcnt) {
+		struct bio_vec *bv = bio_iovec(bio);
+		unsigned j = min(bv->bv_len >> 9, sectors);
+
+		void *p = kmap(bv->bv_page);
+		memset(p + bv->bv_offset, 0, j << 9);
+		kunmap(bv->bv_page);
+
+		bv->bv_len	-= j << 9;
+		bv->bv_offset	+= j << 9;
+
+		if (bv->bv_len)
+			return 0;
+
+		bio->bi_sector	+= j;
+		bio->bi_size	-= j << 9;
+
+		bio->bi_idx++;
+		sectors		-= j;
+	}
+
+	s->op.lookup_done = true;
+
+	return 0;
+}
+
+static void flash_dev_make_request(struct request_queue *q, struct bio *bio)
+{
+	struct search *s;
+	struct closure *cl;
+	struct bcache_device *d = bio->bi_bdev->bd_disk->private_data;
+	int cpu, rw = bio_data_dir(bio);
+
+	cpu = part_stat_lock();
+	part_stat_inc(cpu, &d->disk->part0, ios[rw]);
+	part_stat_add(cpu, &d->disk->part0, sectors[rw], bio_sectors(bio));
+	part_stat_unlock();
+
+	s = search_alloc(bio, d);
+	cl = &s->cl;
+	bio = &s->bio.bio;
+
+	trace_bcache_request_start(s, bio);
+
+	if (bio_has_data(bio) && !rw) {
+		closure_call(&s->op.cl, btree_read_async, NULL, cl);
+	} else if (bio_has_data(bio) || s->op.skip) {
+		bch_keybuf_check_overlapping(&s->op.c->moving_gc_keys,
+					     &KEY(d->id, bio->bi_sector, 0),
+					     &KEY(d->id, bio_end(bio), 0));
+
+		s->writeback	= true;
+		s->op.cache_bio	= bio;
+
+		closure_call(&s->op.cl, bch_insert_data, NULL, cl);
+	} else {
+		/* No data - probably a cache flush */
+		if (s->op.flush_journal)
+			bch_journal_meta(s->op.c, cl);
+	}
+
+	continue_at(cl, search_free, NULL);
+}
+
+static int flash_dev_ioctl(struct bcache_device *d, fmode_t mode,
+			   unsigned int cmd, unsigned long arg)
+{
+	return -ENOTTY;
+}
+
+static int flash_dev_congested(void *data, int bits)
+{
+	struct bcache_device *d = data;
+	struct request_queue *q;
+	struct cache *ca;
+	unsigned i;
+	int ret = 0;
+
+	for_each_cache(ca, d->c, i) {
+		q = bdev_get_queue(ca->bdev);
+		ret |= bdi_congested(&q->backing_dev_info, bits);
+	}
+
+	return ret;
+}
+
+void bch_flash_dev_request_init(struct bcache_device *d)
+{
+	struct gendisk *g = d->disk;
+
+	g->queue->make_request_fn		= flash_dev_make_request;
+	g->queue->backing_dev_info.congested_fn = flash_dev_congested;
+	d->cache_miss				= flash_dev_cache_miss;
+	d->ioctl				= flash_dev_ioctl;
+}
+
+void bch_request_exit(void)
+{
+#ifdef CONFIG_CGROUP_BCACHE
+	cgroup_unload_subsys(&bcache_subsys);
+#endif
+	if (bch_search_cache)
+		kmem_cache_destroy(bch_search_cache);
+}
+
+int __init bch_request_init(void)
+{
+	bch_search_cache = KMEM_CACHE(search, 0);
+	if (!bch_search_cache)
+		return -ENOMEM;
+
+#ifdef CONFIG_CGROUP_BCACHE
+	cgroup_load_subsys(&bcache_subsys);
+	init_bch_cgroup(&bcache_default_cgroup);
+
+	cgroup_add_cftypes(&bcache_subsys, bch_files);
+#endif
+	return 0;
+}

drivers/md/bcache/request.h

···
+#ifndef _BCACHE_REQUEST_H_
+#define _BCACHE_REQUEST_H_
+
+#include <linux/cgroup.h>
+
+struct search {
+	/* Stack frame for bio_complete */
+	struct closure		cl;
+
+	struct bcache_device	*d;
+	struct task_struct	*task;
+
+	struct bbio		bio;
+	struct bio		*orig_bio;
+	struct bio		*cache_miss;
+	unsigned		cache_bio_sectors;
+
+	unsigned		recoverable:1;
+	unsigned		unaligned_bvec:1;
+
+	unsigned		write:1;
+	unsigned		writeback:1;
+
+	/* IO error returned to s->bio */
+	short			error;
+	unsigned long		start_time;
+
+	/* Anything past op->keys won't get zeroed in do_bio_hook */
+	struct btree_op		op;
+};
+
+void bch_cache_read_endio(struct bio *, int);
+int bch_get_congested(struct cache_set *);
+void bch_insert_data(struct closure *cl);
+void bch_btree_insert_async(struct closure *);
+void bch_cache_read_endio(struct bio *, int);
+
+void bch_open_buckets_free(struct cache_set *);
+int bch_open_buckets_alloc(struct cache_set *);
+
+void bch_cached_dev_request_init(struct cached_dev *dc);
+void bch_flash_dev_request_init(struct bcache_device *d);
+
+extern struct kmem_cache *bch_search_cache, *bch_passthrough_cache;
+
+struct bch_cgroup {
+#ifdef CONFIG_CGROUP_BCACHE
+	struct cgroup_subsys_state	css;
+#endif
+	/*
+	 * We subtract one from the index into bch_cache_modes[], so that
+	 * default == -1; this makes it so the rest match up with d->cache_mode,
+	 * and we use d->cache_mode if cgrp->cache_mode < 0
+	 */
+	short				cache_mode;
+	bool				verify;
+	struct cache_stat_collector	stats;
+};
+
+struct bch_cgroup *bch_bio_to_cgroup(struct bio *bio);
+
+#endif /* _BCACHE_REQUEST_H_ */

drivers/md/bcache/stats.c

···
+/*
+ * bcache stats code
+ *
+ * Copyright 2012 Google, Inc.
+ */
+
+#include "bcache.h"
+#include "stats.h"
+#include "btree.h"
+#include "request.h"
+#include "sysfs.h"
+
+/*
+ * We keep absolute totals of various statistics, and addionally a set of three
+ * rolling averages.
+ *
+ * Every so often, a timer goes off and rescales the rolling averages.
+ * accounting_rescale[] is how many times the timer has to go off before we
+ * rescale each set of numbers; that gets us half lives of 5 minutes, one hour,
+ * and one day.
+ *
+ * accounting_delay is how often the timer goes off - 22 times in 5 minutes,
+ * and accounting_weight is what we use to rescale:
+ *
+ * pow(31 / 32, 22) ~= 1/2
+ *
+ * So that we don't have to increment each set of numbers every time we (say)
+ * get a cache hit, we increment a single atomic_t in acc->collector, and when
+ * the rescale function runs it resets the atomic counter to 0 and adds its
+ * old value to each of the exported numbers.
+ *
+ * To reduce rounding error, the numbers in struct cache_stats are all
+ * stored left shifted by 16, and scaled back in the sysfs show() function.
+ */
+
+static const unsigned DAY_RESCALE		= 288;
+static const unsigned HOUR_RESCALE		= 12;
+static const unsigned FIVE_MINUTE_RESCALE	= 1;
+static const unsigned accounting_delay		= (HZ * 300) / 22;
+static const unsigned accounting_weight		= 32;
+
+/* sysfs reading/writing */
+
+read_attribute(cache_hits);
+read_attribute(cache_misses);
+read_attribute(cache_bypass_hits);
+read_attribute(cache_bypass_misses);
+read_attribute(cache_hit_ratio);
+read_attribute(cache_readaheads);
+read_attribute(cache_miss_collisions);
+read_attribute(bypassed);
+
+SHOW(bch_stats)
+{
+	struct cache_stats *s =
+		container_of(kobj, struct cache_stats, kobj);
+#define var(stat)		(s->stat >> 16)
+	var_print(cache_hits);
+	var_print(cache_misses);
+	var_print(cache_bypass_hits);
+	var_print(cache_bypass_misses);
+
+	sysfs_print(cache_hit_ratio,
+		    DIV_SAFE(var(cache_hits) * 100,
+			     var(cache_hits) + var(cache_misses)));
+
+	var_print(cache_readaheads);
+	var_print(cache_miss_collisions);
+	sysfs_hprint(bypassed,	var(sectors_bypassed) << 9);
+#undef var
+	return 0;
+}
+
+STORE(bch_stats)
+{
+	return size;
+}
+
+static void bch_stats_release(struct kobject *k)
+{
+}
+
+static struct attribute *bch_stats_files[] = {
+	&sysfs_cache_hits,
+	&sysfs_cache_misses,
+	&sysfs_cache_bypass_hits,
+	&sysfs_cache_bypass_misses,
+	&sysfs_cache_hit_ratio,
+	&sysfs_cache_readaheads,
+	&sysfs_cache_miss_collisions,
+	&sysfs_bypassed,
+	NULL
+};
+static KTYPE(bch_stats);
+
+static void scale_accounting(unsigned long data);
+
+void bch_cache_accounting_init(struct cache_accounting *acc,
+			       struct closure *parent)
+{
+	kobject_init(&acc->total.kobj,		&bch_stats_ktype);
+	kobject_init(&acc->five_minute.kobj,	&bch_stats_ktype);
+	kobject_init(&acc->hour.kobj,		&bch_stats_ktype);
+	kobject_init(&acc->day.kobj,		&bch_stats_ktype);
+
+	closure_init(&acc->cl, parent);
+	init_timer(&acc->timer);
+	acc->timer.expires	= jiffies + accounting_delay;
+	acc->timer.data		= (unsigned long) acc;
+	acc->timer.function	= scale_accounting;
+	add_timer(&acc->timer);
+}
+
+int bch_cache_accounting_add_kobjs(struct cache_accounting *acc,
+				   struct kobject *parent)
+{
+	int ret = kobject_add(&acc->total.kobj, parent,
+			      "stats_total");
+	ret = ret ?: kobject_add(&acc->five_minute.kobj, parent,
+				 "stats_five_minute");
+	ret = ret ?: kobject_add(&acc->hour.kobj, parent,
+				 "stats_hour");
+	ret = ret ?: kobject_add(&acc->day.kobj, parent,
+				 "stats_day");
+	return ret;
+}
+
+void bch_cache_accounting_clear(struct cache_accounting *acc)
+{
+	memset(&acc->total.cache_hits,
+	       0,
+	       sizeof(unsigned long) * 7);
+}
+
+void bch_cache_accounting_destroy(struct cache_accounting *acc)
+{
+	kobject_put(&acc->total.kobj);
+	kobject_put(&acc->five_minute.kobj);
+	kobject_put(&acc->hour.kobj);
+	kobject_put(&acc->day.kobj);
+
+	atomic_set(&acc->closing, 1);
+	if (del_timer_sync(&acc->timer))
+		closure_return(&acc->cl);
+}
+
+/* EWMA scaling */
+
+static void scale_stat(unsigned long *stat)
+{
+	*stat =  ewma_add(*stat, 0, accounting_weight, 0);
+}
+
+static void scale_stats(struct cache_stats *stats, unsigned long rescale_at)
+{
+	if (++stats->rescale == rescale_at) {
+		stats->rescale = 0;
+		scale_stat(&stats->cache_hits);
+		scale_stat(&stats->cache_misses);
+		scale_stat(&stats->cache_bypass_hits);
+		scale_stat(&stats->cache_bypass_misses);
+		scale_stat(&stats->cache_readaheads);
+		scale_stat(&stats->cache_miss_collisions);
+		scale_stat(&stats->sectors_bypassed);
+	}
+}
+
+static void scale_accounting(unsigned long data)
+{
+	struct cache_accounting *acc = (struct cache_accounting *) data;
+
+#define move_stat(name) do {						\
+	unsigned t = atomic_xchg(&acc->collector.name, 0);		\
+	t <<= 16;							\
+	acc->five_minute.name += t;					\
+	acc->hour.name += t;						\
+	acc->day.name += t;						\
+	acc->total.name += t;						\
+} while (0)
+
+	move_stat(cache_hits);
+	move_stat(cache_misses);
+	move_stat(cache_bypass_hits);
+	move_stat(cache_bypass_misses);
+	move_stat(cache_readaheads);
+	move_stat(cache_miss_collisions);
+	move_stat(sectors_bypassed);
+
+	scale_stats(&acc->total, 0);
+	scale_stats(&acc->day, DAY_RESCALE);
+	scale_stats(&acc->hour, HOUR_RESCALE);
+	scale_stats(&acc->five_minute, FIVE_MINUTE_RESCALE);
+
+	acc->timer.expires += accounting_delay;
+
+	if (!atomic_read(&acc->closing))
+		add_timer(&acc->timer);
+	else
+		closure_return(&acc->cl);
+}
+
+static void mark_cache_stats(struct cache_stat_collector *stats,
+			     bool hit, bool bypass)
+{
+	if (!bypass)
+		if (hit)
+			atomic_inc(&stats->cache_hits);
+		else
+			atomic_inc(&stats->cache_misses);
+	else
+		if (hit)
+			atomic_inc(&stats->cache_bypass_hits);
+		else
+			atomic_inc(&stats->cache_bypass_misses);
+}
+
+void bch_mark_cache_accounting(struct search *s, bool hit, bool bypass)
+{
+	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
+	mark_cache_stats(&dc->accounting.collector, hit, bypass);
+	mark_cache_stats(&s->op.c->accounting.collector, hit, bypass);
+#ifdef CONFIG_CGROUP_BCACHE
+	mark_cache_stats(&(bch_bio_to_cgroup(s->orig_bio)->stats), hit, bypass);
+#endif
+}
+
+void bch_mark_cache_readahead(struct search *s)
+{
+	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
+	atomic_inc(&dc->accounting.collector.cache_readaheads);
+	atomic_inc(&s->op.c->accounting.collector.cache_readaheads);
+}
+
+void bch_mark_cache_miss_collision(struct search *s)
+{
+	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
+	atomic_inc(&dc->accounting.collector.cache_miss_collisions);
+	atomic_inc(&s->op.c->accounting.collector.cache_miss_collisions);
+}
+
+void bch_mark_sectors_bypassed(struct search *s, int sectors)
+{
+	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
+	atomic_add(sectors, &dc->accounting.collector.sectors_bypassed);
+	atomic_add(sectors, &s->op.c->accounting.collector.sectors_bypassed);
+}

drivers/md/bcache/stats.h

···
+#ifndef _BCACHE_STATS_H_
+#define _BCACHE_STATS_H_
+
+struct cache_stat_collector {
+	atomic_t cache_hits;
+	atomic_t cache_misses;
+	atomic_t cache_bypass_hits;
+	atomic_t cache_bypass_misses;
+	atomic_t cache_readaheads;
+	atomic_t cache_miss_collisions;
+	atomic_t sectors_bypassed;
+};
+
+struct cache_stats {
+	struct kobject		kobj;
+
+	unsigned long cache_hits;
+	unsigned long cache_misses;
+	unsigned long cache_bypass_hits;
+	unsigned long cache_bypass_misses;
+	unsigned long cache_readaheads;
+	unsigned long cache_miss_collisions;
+	unsigned long sectors_bypassed;
+
+	unsigned		rescale;
+};
+
+struct cache_accounting {
+	struct closure		cl;
+	struct timer_list	timer;
+	atomic_t		closing;
+
+	struct cache_stat_collector collector;
+
+	struct cache_stats total;
+	struct cache_stats five_minute;
+	struct cache_stats hour;
+	struct cache_stats day;
+};
+
+struct search;
+
+void bch_cache_accounting_init(struct cache_accounting *acc,
+			       struct closure *parent);
+
+int bch_cache_accounting_add_kobjs(struct cache_accounting *acc,
+				   struct kobject *parent);
+
+void bch_cache_accounting_clear(struct cache_accounting *acc);
+
+void bch_cache_accounting_destroy(struct cache_accounting *acc);
+
+void bch_mark_cache_accounting(struct search *s, bool hit, bool bypass);
+void bch_mark_cache_readahead(struct search *s);
+void bch_mark_cache_miss_collision(struct search *s);
+void bch_mark_sectors_bypassed(struct search *s, int sectors);
+
+#endif /* _BCACHE_STATS_H_ */

drivers/md/bcache/super.c

···
+/*
+ * bcache setup/teardown code, and some metadata io - read a superblock and
+ * figure out what to do with it.
+ *
+ * Copyright 2010, 2011 Kent Overstreet <[e-mail]>
+ * Copyright 2012 Google, Inc.
+ */
+
+#include "bcache.h"
+#include "btree.h"
+#include "debug.h"
+#include "request.h"
+
+#include <linux/buffer_head.h>
+#include <linux/debugfs.h>
+#include <linux/genhd.h>
+#include <linux/module.h>
+#include <linux/random.h>
+#include <linux/reboot.h>
+#include <linux/sysfs.h>
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Kent Overstreet <[e-mail]>");
+
+static const char bcache_magic[] = {
+	0xc6, 0x85, 0x73, 0xf6, 0x4e, 0x1a, 0x45, 0xca,
+	0x82, 0x65, 0xf5, 0x7f, 0x48, 0xba, 0x6d, 0x81
+};
+
+static const char invalid_uuid[] = {
+	0xa0, 0x3e, 0xf8, 0xed, 0x3e, 0xe1, 0xb8, 0x78,
+	0xc8, 0x50, 0xfc, 0x5e, 0xcb, 0x16, 0xcd, 0x99
+};
+
+/* Default is -1; we skip past it for struct cached_dev's cache mode */
+const char * const bch_cache_modes[] = {
+	"default",
+	"writethrough",
+	"writeback",
+	"writearound",
+	"none",
+	NULL
+};
+
+struct uuid_entry_v0 {
+	uint8_t		uuid[16];
+	uint8_t		label[32];
+	uint32_t	first_reg;
+	uint32_t	last_reg;
+	uint32_t	invalidated;
+	uint32_t	pad;
+};
+
+static struct kobject *bcache_kobj;
+struct mutex bch_register_lock;
+LIST_HEAD(bch_cache_sets);
+static LIST_HEAD(uncached_devices);
+
+static int bcache_major, bcache_minor;
+static wait_queue_head_t unregister_wait;
+struct workqueue_struct *bcache_wq;
+
+#define BTREE_MAX_PAGES		(256 * 1024 / PAGE_SIZE)
+
+static void bio_split_pool_free(struct bio_split_pool *p)
+{
+	if (p->bio_split_hook)
+		mempool_destroy(p->bio_split_hook);
+
+	if (p->bio_split)
+		bioset_free(p->bio_split);
+}
+
+static int bio_split_pool_init(struct bio_split_pool *p)
+{
+	p->bio_split = bioset_create(4, 0);
+	if (!p->bio_split)
+		return -ENOMEM;
+
+	p->bio_split_hook = mempool_create_kmalloc_pool(4,
+				sizeof(struct bio_split_hook));
+	if (!p->bio_split_hook)
+		return -ENOMEM;
+
+	return 0;
+}
+
+/* Superblock */
+
+static const char *read_super(struct cache_sb *sb, struct block_device *bdev,
+			      struct page **res)
+{
+	const char *err;
+	struct cache_sb *s;
+	struct buffer_head *bh = __bread(bdev, 1, SB_SIZE);
+	unsigned i;
+
+	if (!bh)
+		return "IO error";
+
+	s = (struct cache_sb *) bh->b_data;
+
+	sb->offset		= le64_to_cpu(s->offset);
+	sb->version		= le64_to_cpu(s->version);
+
+	memcpy(sb->magic,	s->magic, 16);
+	memcpy(sb->uuid,	s->uuid, 16);
+	memcpy(sb->set_uuid,	s->set_uuid, 16);
+	memcpy(sb->label,	s->label, SB_LABEL_SIZE);
+
+	sb->flags		= le64_to_cpu(s->flags);
+	sb->seq			= le64_to_cpu(s->seq);
+	sb->last_mount		= le32_to_cpu(s->last_mount);
+	sb->first_bucket	= le16_to_cpu(s->first_bucket);
+	sb->keys		= le16_to_cpu(s->keys);
+
+	for (i = 0; i < SB_JOURNAL_BUCKETS; i++)
+		sb->d[i] = le64_to_cpu(s->d[i]);
+
+	pr_debug("read sb version %llu, flags %llu, seq %llu, journal size %u",
+		 sb->version, sb->flags, sb->seq, sb->keys);
+
+	err = "Not a bcache superblock";
+	if (sb->offset != SB_SECTOR)
+		goto err;
+
+	if (memcmp(sb->magic, bcache_magic, 16))
+		goto err;
+
+	err = "Too many journal buckets";
+	if (sb->keys > SB_JOURNAL_BUCKETS)
+		goto err;
+
+	err = "Bad checksum";
+	if (s->csum != csum_set(s))
+		goto err;
+
+	err = "Bad UUID";
+	if (bch_is_zero(sb->uuid, 16))
+		goto err;
+
+	sb->block_size	= le16_to_cpu(s->block_size);
+
+	err = "Superblock block size smaller than device block size";
+	if (sb->block_size << 9 < bdev_logical_block_size(bdev))
+		goto err;
+
+	switch (sb->version) {
+	case BCACHE_SB_VERSION_BDEV:
+		sb->data_offset	= BDEV_DATA_START_DEFAULT;
+		break;
+	case BCACHE_SB_VERSION_BDEV_WITH_OFFSET:
+		sb->data_offset	= le64_to_cpu(s->data_offset);
+
+		err = "Bad data offset";
+		if (sb->data_offset < BDEV_DATA_START_DEFAULT)
+			goto err;
+
+		break;
+	case BCACHE_SB_VERSION_CDEV:
+	case BCACHE_SB_VERSION_CDEV_WITH_UUID:
+		sb->nbuckets	= le64_to_cpu(s->nbuckets);
+		sb->block_size	= le16_to_cpu(s->block_size);
+		sb->bucket_size	= le16_to_cpu(s->bucket_size);
+
+		sb->nr_in_set	= le16_to_cpu(s->nr_in_set);
+		sb->nr_this_dev	= le16_to_cpu(s->nr_this_dev);
+
+		err = "Too many buckets";
+		if (sb->nbuckets > LONG_MAX)
+			goto err;
+
+		err = "Not enough buckets";
+		if (sb->nbuckets < 1 << 7)
+			goto err;
+
+		err = "Bad block/bucket size";
+		if (!is_power_of_2(sb->block_size) ||
+		    sb->block_size > PAGE_SECTORS ||
+		    !is_power_of_2(sb->bucket_size) ||
+		    sb->bucket_size < PAGE_SECTORS)
+			goto err;
+
+		err = "Invalid superblock: device too small";
+		if (get_capacity(bdev->bd_disk) < sb->bucket_size * sb->nbuckets)
+			goto err;
+
+		err = "Bad UUID";
+		if (bch_is_zero(sb->set_uuid, 16))
+			goto err;
+
+		err = "Bad cache device number in set";
+		if (!sb->nr_in_set ||
+		    sb->nr_in_set <= sb->nr_this_dev ||
+		    sb->nr_in_set > MAX_CACHES_PER_SET)
+			goto err;
+
+		err = "Journal buckets not sequential";
+		for (i = 0; i < sb->keys; i++)
+			if (sb->d[i] != sb->first_bucket + i)
+				goto err;
+
+		err = "Too many journal buckets";
+		if (sb->first_bucket + sb->keys > sb->nbuckets)
+			goto err;
+
+		err = "Invalid superblock: first bucket comes before end of super";
+		if (sb->first_bucket * sb->bucket_size < 16)
+			goto err;
+
+		break;
+	default:
+		err = "Unsupported superblock version";
+		goto err;
+	}
+
+	sb->last_mount = get_seconds();
+	err = NULL;
+
+	get_page(bh->b_page);
+	*res = bh->b_page;
+err:
+	put_bh(bh);
+	return err;
+}
+
+static void write_bdev_super_endio(struct bio *bio, int error)
+{
+	struct cached_dev *dc = bio->bi_private;
+	/* XXX: error checking */
+
+	closure_put(&dc->sb_write.cl);
+}
+
+static void __write_super(struct cache_sb *sb, struct bio *bio)
+{
+	struct cache_sb *out = page_address(bio->bi_io_vec[0].bv_page);
+	unsigned i;
+
+	bio->bi_sector	= SB_SECTOR;
+	bio->bi_rw	= REQ_SYNC|REQ_META;
+	bio->bi_size	= SB_SIZE;
+	bch_bio_map(bio, NULL);
+
+	out->offset		= cpu_to_le64(sb->offset);
+	out->version		= cpu_to_le64(sb->version);
+
+	memcpy(out->uuid,	sb->uuid, 16);
+	memcpy(out->set_uuid,	sb->set_uuid, 16);
+	memcpy(out->label,	sb->label, SB_LABEL_SIZE);
+
+	out->flags		= cpu_to_le64(sb->flags);
+	out->seq		= cpu_to_le64(sb->seq);
+
+	out->last_mount		= cpu_to_le32(sb->last_mount);
+	out->first_bucket	= cpu_to_le16(sb->first_bucket);
+	out->keys		= cpu_to_le16(sb->keys);
+
+	for (i = 0; i < sb->keys; i++)
+		out->d[i] = cpu_to_le64(sb->d[i]);
+
+	out->csum = csum_set(out);
+
+	pr_debug("ver %llu, flags %llu, seq %llu",
+		 sb->version, sb->flags, sb->seq);
+
+	submit_bio(REQ_WRITE, bio);
+}
+
+void bch_write_bdev_super(struct cached_dev *dc, struct closure *parent)
+{
+	struct closure *cl = &dc->sb_write.cl;
+	struct bio *bio = &dc->sb_bio;
+
+	closure_lock(&dc->sb_write, parent);
+
+	bio_reset(bio);
+	bio->bi_bdev	= dc->bdev;
+	bio->bi_end_io	= write_bdev_super_endio;
+	bio->bi_private = dc;
+
+	closure_get(cl);
+	__write_super(&dc->sb, bio);
+
+	closure_return(cl);
+}
+
+static void write_super_endio(struct bio *bio, int error)
+{
+	struct cache *ca = bio->bi_private;
+
+	bch_count_io_errors(ca, error, "writing superblock");
+	closure_put(&ca->set->sb_write.cl);
+}
+
+void bcache_write_super(struct cache_set *c)
+{
+	struct closure *cl = &c->sb_write.cl;
+	struct cache *ca;
+	unsigned i;
+
+	closure_lock(&c->sb_write, &c->cl);
+
+	c->sb.seq++;
+
+	for_each_cache(ca, c, i) {
+		struct bio *bio = &ca->sb_bio;
+
+		ca->sb.version		= BCACHE_SB_VERSION_CDEV_WITH_UUID;
+		ca->sb.seq		= c->sb.seq;
+		ca->sb.last_mount	= c->sb.last_mount;
+
+		SET_CACHE_SYNC(&ca->sb, CACHE_SYNC(&c->sb));
+
+		bio_reset(bio);
+		bio->bi_bdev	= ca->bdev;
+		bio->bi_end_io	= write_super_endio;
+		bio->bi_private = ca;
+
+		closure_get(cl);
+		__write_super(&ca->sb, bio);
+	}
+
+	closure_return(cl);
+}
+
+/* UUID io */
+
+static void uuid_endio(struct bio *bio, int error)
+{
+	struct closure *cl = bio->bi_private;
+	struct cache_set *c = container_of(cl, struct cache_set, uuid_write.cl);
+
+	cache_set_err_on(error, c, "accessing uuids");
+	bch_bbio_free(bio, c);
+	closure_put(cl);
+}
+
+static void uuid_io(struct cache_set *c, unsigned long rw,
+		    struct bkey *k, struct closure *parent)
+{
+	struct closure *cl = &c->uuid_write.cl;
+	struct uuid_entry *u;
+	unsigned i;
+
+	BUG_ON(!parent);
+	closure_lock(&c->uuid_write, parent);
+
+	for (i = 0; i < KEY_PTRS(k); i++) {
+		struct bio *bio = bch_bbio_alloc(c);
+
+		bio->bi_rw	= REQ_SYNC|REQ_META|rw;
+		bio->bi_size	= KEY_SIZE(k) << 9;
+
+		bio->bi_end_io	= uuid_endio;
+		bio->bi_private = cl;
+		bch_bio_map(bio, c->uuids);
+
+		bch_submit_bbio(bio, c, k, i);
+
+		if (!(rw & WRITE))
+			break;
+	}
+
+	pr_debug("%s UUIDs at %s", rw & REQ_WRITE ? "wrote" : "read",
+		 pkey(&c->uuid_bucket));
+
+	for (u = c->uuids; u < c->uuids + c->nr_uuids; u++)
+		if (!bch_is_zero(u->uuid, 16))
+			pr_debug("Slot %zi: %pU: %s: 1st: %u last: %u inv: %u",
+				 u - c->uuids, u->uuid, u->label,
+				 u->first_reg, u->last_reg, u->invalidated);
+
+	closure_return(cl);
+}
+
+static char *uuid_read(struct cache_set *c, struct jset *j, struct closure *cl)
+{
+	struct bkey *k = &j->uuid_bucket;
+
+	if (__bch_ptr_invalid(c, 1, k))
+		return "bad uuid pointer";
+
+	bkey_copy(&c->uuid_bucket, k);
+	uuid_io(c, READ_SYNC, k, cl);
+
+	if (j->version < BCACHE_JSET_VERSION_UUIDv1) {
+		struct uuid_entry_v0	*u0 = (void *) c->uuids;
+		struct uuid_entry	*u1 = (void *) c->uuids;
+		int i;
+
+		closure_sync(cl);
+
+		/*
+		 * Since the new uuid entry is bigger than the old, we have to
+		 * convert starting at the highest memory address and work down
+		 * in order to do it in place
+		 */
+
+		for (i = c->nr_uuids - 1;
+		     i >= 0;
+		     --i) {
+			memcpy(u1[i].uuid,	u0[i].uuid, 16);
+			memcpy(u1[i].label,	u0[i].label, 32);
+
+			u1[i].first_reg		= u0[i].first_reg;
+			u1[i].last_reg		= u0[i].last_reg;
+			u1[i].invalidated	= u0[i].invalidated;
+
+			u1[i].flags	= 0;
+			u1[i].sectors	= 0;
+		}
+	}
+
+	return NULL;
+}
+
+static int __uuid_write(struct cache_set *c)
+{
+	BKEY_PADDED(key) k;
+	struct closure cl;
+	closure_init_stack(&cl);
+
+	lockdep_assert_held(&bch_register_lock);
+
+	if (bch_bucket_alloc_set(c, WATERMARK_METADATA, &k.key, 1, &cl))
+		return 1;
+
+	SET_KEY_SIZE(&k.key, c->sb.bucket_size);
+	uuid_io(c, REQ_WRITE, &k.key, &cl);
+	closure_sync(&cl);
+
+	bkey_copy(&c->uuid_bucket, &k.key);
+	__bkey_put(c, &k.key);
+	return 0;
+}
+
+int bch_uuid_write(struct cache_set *c)
+{
+	int ret = __uuid_write(c);
+
+	if (!ret)
+		bch_journal_meta(c, NULL);
+
+	return ret;
+}
+
+static struct uuid_entry *uuid_find(struct cache_set *c, const char *uuid)
+{
+	struct uuid_entry *u;
+
+	for (u = c->uuids;
+	     u < c->uuids + c->nr_uuids; u++)
+		if (!memcmp(u->uuid, uuid, 16))
+			return u;
+
+	return NULL;
+}
+
+static struct uuid_entry *uuid_find_empty(struct cache_set *c)
+{
+	static const char zero_uuid[16] = "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0";
+	return uuid_find(c, zero_uuid);
+}
+
+/*
+ * Bucket priorities/gens:
+ *
+ * For each bucket, we store on disk its
+   * 8 bit gen
+   * 16 bit priority
+ *
+ * See alloc.c for an explanation of the gen. The priority is used to implement
+ * lru (and in the future other) cache replacement policies; for most purposes
+ * it's just an opaque integer.
+ *
+ * The gens and the priorities don't have a whole lot to do with each other, and
+ * it's actually the gens that must be written out at specific times - it's no
+ * big deal if the priorities don't get written, if we lose them we just reuse
+ * buckets in suboptimal order.
+ *
+ * On disk they're stored in a packed array, and in as many buckets are required
+ * to fit them all. The buckets we use to store them form a list; the journal
+ * header points to the first bucket, the first bucket points to the second
+ * bucket, et cetera.
+ *
+ * This code is used by the allocation code; periodically (whenever it runs out
+ * of buckets to allocate from) the allocation code will invalidate some
+ * buckets, but it can't use those buckets until their new gens are safely on
+ * disk.
+ */
+
+static void prio_endio(struct bio *bio, int error)
+{
+	struct cache *ca = bio->bi_private;
+
+	cache_set_err_on(error, ca->set, "accessing priorities");
+	bch_bbio_free(bio, ca->set);
+	closure_put(&ca->prio);
+}
+
+static void prio_io(struct cache *ca, uint64_t bucket, unsigned long rw)
+{
+	struct closure *cl = &ca->prio;
+	struct bio *bio = bch_bbio_alloc(ca->set);
+
+	closure_init_stack(cl);
+
+	bio->bi_sector	= bucket * ca->sb.bucket_size;
+	bio->bi_bdev	= ca->bdev;
+	bio->bi_rw	= REQ_SYNC|REQ_META|rw;
+	bio->bi_size	= bucket_bytes(ca);
+
+	bio->bi_end_io	= prio_endio;
+	bio->bi_private = ca;
+	bch_bio_map(bio, ca->disk_buckets);
+
+	closure_bio_submit(bio, &ca->prio, ca);
+	closure_sync(cl);
+}
+
+#define buckets_free(c)	"free %zu, free_inc %zu, unused %zu",		\
+	fifo_used(&c->free), fifo_used(&c->free_inc), fifo_used(&c->unused)
+
+void bch_prio_write(struct cache *ca)
+{
+	int i;
+	struct bucket *b;
+	struct closure cl;
+
+	closure_init_stack(&cl);
+
+	lockdep_assert_held(&ca->set->bucket_lock);
+
+	for (b = ca->buckets;
+	     b < ca->buckets + ca->sb.nbuckets; b++)
+		b->disk_gen = b->gen;
+
+	ca->disk_buckets->seq++;
+
+	atomic_long_add(ca->sb.bucket_size * prio_buckets(ca),
+			&ca->meta_sectors_written);
+
+	pr_debug("free %zu, free_inc %zu, unused %zu", fifo_used(&ca->free),
+		 fifo_used(&ca->free_inc), fifo_used(&ca->unused));
+	blktrace_msg(ca, "Starting priorities: " buckets_free(ca));
+
+	for (i = prio_buckets(ca) - 1; i >= 0; --i) {
+		long bucket;
+		struct prio_set *p = ca->disk_buckets;
+		struct bucket_disk *d = p->data;
+		struct bucket_disk *end = d + prios_per_bucket(ca);
+
+		for (b = ca->buckets + i * prios_per_bucket(ca);
+		     b < ca->buckets + ca->sb.nbuckets && d < end;
+		     b++, d++) {
+			d->prio = cpu_to_le16(b->prio);
+			d->gen = b->gen;
+		}
+
+		p->next_bucket	= ca->prio_buckets[i + 1];
+		p->magic	= pset_magic(ca);
+		p->csum		= bch_crc64(&p->magic, bucket_bytes(ca) - 8);
+
+		bucket = bch_bucket_alloc(ca, WATERMARK_PRIO, &cl);
+		BUG_ON(bucket == -1);
+
+		mutex_unlock(&ca->set->bucket_lock);
+		prio_io(ca, bucket, REQ_WRITE);
+		mutex_lock(&ca->set->bucket_lock);
+
+		ca->prio_buckets[i] = bucket;
+		atomic_dec_bug(&ca->buckets[bucket].pin);
+	}
+
+	mutex_unlock(&ca->set->bucket_lock);
+
+	bch_journal_meta(ca->set, &cl);
+	closure_sync(&cl);
+
+	mutex_lock(&ca->set->bucket_lock);
+
+	ca->need_save_prio = 0;
+
+	/*
+	 * Don't want the old priorities to get garbage collected until after we
+	 * finish writing the new ones, and they're journalled
+	 */
+	for (i = 0; i < prio_buckets(ca); i++)
+		ca->prio_last_buckets[i] = ca->prio_buckets[i];
+}
+
+static void prio_read(struct cache *ca, uint64_t bucket)
+{
+	struct prio_set *p = ca->disk_buckets;
+	struct bucket_disk *d = p->data + prios_per_bucket(ca), *end = d;
+	struct bucket *b;
+	unsigned bucket_nr = 0;
+
+	for (b = ca->buckets;
+	     b < ca->buckets + ca->sb.nbuckets;
+	     b++, d++) {
+		if (d == end) {
+			ca->prio_buckets[bucket_nr] = bucket;
+			ca->prio_last_buckets[bucket_nr] = bucket;
+			bucket_nr++;
+
+			prio_io(ca, bucket, READ_SYNC);
+
+			if (p->csum != bch_crc64(&p->magic, bucket_bytes(ca) - 8))
+				pr_warn("bad csum reading priorities");
+
+			if (p->magic != pset_magic(ca))
+				pr_warn("bad magic reading priorities");
+
+			bucket = p->next_bucket;
+			d = p->data;
+		}
+
+		b->prio = le16_to_cpu(d->prio);
+		b->gen = b->disk_gen = b->last_gc = b->gc_gen = d->gen;
+	}
+}
+
+/* Bcache device */
+
+static int open_dev(struct block_device *b, fmode_t mode)
+{
+	struct bcache_device *d = b->bd_disk->private_data;
+	if (atomic_read(&d->closing))
+		return -ENXIO;
+
+	closure_get(&d->cl);
+	return 0;
+}
+
+static int release_dev(struct gendisk *b, fmode_t mode)
+{
+	struct bcache_device *d = b->private_data;
+	closure_put(&d->cl);
+	return 0;
+}
+
+static int ioctl_dev(struct block_device *b, fmode_t mode,
+		     unsigned int cmd, unsigned long arg)
+{
+	struct bcache_device *d = b->bd_disk->private_data;
+	return d->ioctl(d, mode, cmd, arg);
+}
+
+static const struct block_device_operations bcache_ops = {
+	.open		= open_dev,
+	.release	= release_dev,
+	.ioctl		= ioctl_dev,
+	.owner		= THIS_MODULE,
+};
+
+void bcache_device_stop(struct bcache_device *d)
+{
+	if (!atomic_xchg(&d->closing, 1))
+		closure_queue(&d->cl);
+}
+
+static void bcache_device_unlink(struct bcache_device *d)
+{
+	unsigned i;
+	struct cache *ca;
+
+	sysfs_remove_link(&d->c->kobj, d->name);
+	sysfs_remove_link(&d->kobj, "cache");
+
+	for_each_cache(ca, d->c, i)
+		bd_unlink_disk_holder(ca->bdev, d->disk);
+}
+
+static void bcache_device_link(struct bcache_device *d, struct cache_set *c,
+			       const char *name)
+{
+	unsigned i;
+	struct cache *ca;
+
+	for_each_cache(ca, d->c, i)
+		bd_link_disk_holder(ca->bdev, d->disk);
+
+	snprintf(d->name, BCACHEDEVNAME_SIZE,
+		 "%s%u", name, d->id);
+
+	WARN(sysfs_create_link(&d->kobj, &c->kobj, "cache") ||
+	     sysfs_create_link(&c->kobj, &d->kobj, d->name),
+	     "Couldn't create device <-> cache set symlinks");
+}
+
+static void bcache_device_detach(struct bcache_device *d)
+{
+	lockdep_assert_held(&bch_register_lock);
+
+	if (atomic_read(&d->detaching)) {
+		struct uuid_entry *u = d->c->uuids + d->id;
+
+		SET_UUID_FLASH_ONLY(u, 0);
+		memcpy(u->uuid, invalid_uuid, 16);
+		u->invalidated = cpu_to_le32(get_seconds());
+		bch_uuid_write(d->c);
+
+		atomic_set(&d->detaching, 0);
+	}
+
+	bcache_device_unlink(d);
+
+	d->c->devices[d->id] = NULL;
+	closure_put(&d->c->caching);
+	d->c = NULL;
+}
+
+static void bcache_device_attach(struct bcache_device *d, struct cache_set *c,
+				 unsigned id)
+{
+	BUG_ON(test_bit(CACHE_SET_STOPPING, &c->flags));
+
+	d->id = id;
+	d->c = c;
+	c->devices[id] = d;
+
+	closure_get(&c->caching);
+}
+
+static void bcache_device_free(struct bcache_device *d)
+{
+	lockdep_assert_held(&bch_register_lock);
+
+	pr_info("%s stopped", d->disk->disk_name);
+
+	if (d->c)
+		bcache_device_detach(d);
+
+	if (d->disk)
+		del_gendisk(d->disk);
+	if (d->disk && d->disk->queue)
+		blk_cleanup_queue(d->disk->queue);
+	if (d->disk)
+		put_disk(d->disk);
+
+	bio_split_pool_free(&d->bio_split_hook);
+	if (d->unaligned_bvec)
+		mempool_destroy(d->unaligned_bvec);
+	if (d->bio_split)
+		bioset_free(d->bio_split);
+
+	closure_debug_destroy(&d->cl);
+}
+
+static int bcache_device_init(struct bcache_device *d, unsigned block_size)
+{
+	struct request_queue *q;
+
+	if (!(d->bio_split = bioset_create(4, offsetof(struct bbio, bio))) ||
+	    !(d->unaligned_bvec = mempool_create_kmalloc_pool(1,
+				sizeof(struct bio_vec) * BIO_MAX_PAGES)) ||
+	    bio_split_pool_init(&d->bio_split_hook))
+
+		return -ENOMEM;
+
+	d->disk = alloc_disk(1);
+	if (!d->disk)
+		return -ENOMEM;
+
+	snprintf(d->disk->disk_name, DISK_NAME_LEN, "bcache%i", bcache_minor);
+
+	d->disk->major		= bcache_major;
+	d->disk->first_minor	= bcache_minor++;
+	d->disk->fops		= &bcache_ops;
+	d->disk->private_data	= d;
+
+	q = blk_alloc_queue(GFP_KERNEL);
+	if (!q)
+		return -ENOMEM;
+
+	blk_queue_make_request(q, NULL);
+	d->disk->queue			= q;
+	q->queuedata			= d;
+	q->backing_dev_info.congested_data = d;
+	q->limits.max_hw_sectors	= UINT_MAX;
+	q->limits.max_sectors		= UINT_MAX;
+	q->limits.max_segment_size	= UINT_MAX;
+	q->limits.max_segments		= BIO_MAX_PAGES;
+	q->limits.max_discard_sectors	= UINT_MAX;
+	q->limits.io_min		= block_size;
+	q->limits.logical_block_size	= block_size;
+	q->limits.physical_block_size	= block_size;
+	set_bit(QUEUE_FLAG_NONROT,	&d->disk->queue->queue_flags);
+	set_bit(QUEUE_FLAG_DISCARD,	&d->disk->queue->queue_flags);
+
+	return 0;
+}
+
+/* Cached device */
+
+static void calc_cached_dev_sectors(struct cache_set *c)
+{
+	uint64_t sectors = 0;
+	struct cached_dev *dc;
+
+	list_for_each_entry(dc, &c->cached_devs, list)
+		sectors += bdev_sectors(dc->bdev);
+
+	c->cached_dev_sectors = sectors;
+}
+
+void bch_cached_dev_run(struct cached_dev *dc)
+{
+	struct bcache_device *d = &dc->disk;
+
+	if (atomic_xchg(&dc->running, 1))
+		return;
+
+	if (!d->c &&
+	    BDEV_STATE(&dc->sb) != BDEV_STATE_NONE) {
+		struct closure cl;
+		closure_init_stack(&cl);
+
+		SET_BDEV_STATE(&dc->sb, BDEV_STATE_STALE);
+		bch_write_bdev_super(dc, &cl);
+		closure_sync(&cl);
+	}
+
+	add_disk(d->disk);
+	bd_link_disk_holder(dc->bdev, dc->disk.disk);
+#if 0
+	char *env[] = { "SYMLINK=label" , NULL };
+	kobject_uevent_env(&disk_to_dev(d->disk)->kobj, KOBJ_CHANGE, env);
+#endif
+	if (sysfs_create_link(&d->kobj, &disk_to_dev(d->disk)->kobj, "dev") ||
+	    sysfs_create_link(&disk_to_dev(d->disk)->kobj, &d->kobj, "bcache"))
+		pr_debug("error creating sysfs link");
+}
+
+static void cached_dev_detach_finish(struct work_struct *w)
+{
+	struct cached_dev *dc = container_of(w, struct cached_dev, detach);
+	char buf[BDEVNAME_SIZE];
+	struct closure cl;
+	closure_init_stack(&cl);
+
+	BUG_ON(!atomic_read(&dc->disk.detaching));
+	BUG_ON(atomic_read(&dc->count));
+
+	mutex_lock(&bch_register_lock);
+
+	memset(&dc->sb.set_uuid, 0, 16);
+	SET_BDEV_STATE(&dc->sb, BDEV_STATE_NONE);
+
+	bch_write_bdev_super(dc, &cl);
+	closure_sync(&cl);
+
+	bcache_device_detach(&dc->disk);
+	list_move(&dc->list, &uncached_devices);
+
+	mutex_unlock(&bch_register_lock);
+
+	pr_info("Caching disabled for %s", bdevname(dc->bdev, buf));
+
+	/* Drop ref we took in cached_dev_detach() */
+	closure_put(&dc->disk.cl);
+}
+
+void bch_cached_dev_detach(struct cached_dev *dc)
+{
+	lockdep_assert_held(&bch_register_lock);
+
+	if (atomic_read(&dc->disk.closing))
+		return;
+
+	if (atomic_xchg(&dc->disk.detaching, 1))
+		return;
+
+	/*
+	 * Block the device from being closed and freed until we're finished
+	 * detaching
+	 */
+	closure_get(&dc->disk.cl);
+
+	bch_writeback_queue(dc);
+	cached_dev_put(dc);
+}
+
+int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c)
+{
+	uint32_t rtime = cpu_to_le32(get_seconds());
+	struct uuid_entry *u;
+	char buf[BDEVNAME_SIZE];
+
+	bdevname(dc->bdev, buf);
+
+	if (memcmp(dc->sb.set_uuid, c->sb.set_uuid, 16))
+		return -ENOENT;
+
+	if (dc->disk.c) {
+		pr_err("Can't attach %s: already attached", buf);
+		return -EINVAL;
+	}
+
+	if (test_bit(CACHE_SET_STOPPING, &c->flags)) {
+		pr_err("Can't attach %s: shutting down", buf);
+		return -EINVAL;
+	}
+
+	if (dc->sb.block_size < c->sb.block_size) {
+		/* Will die */
+		pr_err("Couldn't attach %s: block size less than set's block size",
+		       buf);
+		return -EINVAL;
+	}
+
+	u = uuid_find(c, dc->sb.uuid);
+
+	if (u &&
+	    (BDEV_STATE(&dc->sb) == BDEV_STATE_STALE ||
+	     BDEV_STATE(&dc->sb) == BDEV_STATE_NONE)) {
+		memcpy(u->uuid, invalid_uuid, 16);
+		u->invalidated = cpu_to_le32(get_seconds());
+		u = NULL;
+	}
+
+	if (!u) {
+		if (BDEV_STATE(&dc->sb) == BDEV_STATE_DIRTY) {
+			pr_err("Couldn't find uuid for %s in set", buf);
+			return -ENOENT;
+		}
+
+		u = uuid_find_empty(c);
+		if (!u) {
+			pr_err("Not caching %s, no room for UUID", buf);
+			return -EINVAL;
+		}
+	}
+
+	/* Deadlocks since we're called via sysfs...
+	sysfs_remove_file(&dc->kobj, &sysfs_attach);
+	 */
+
+	if (bch_is_zero(u->uuid, 16)) {
+		struct closure cl;
+		closure_init_stack(&cl);
+
+		memcpy(u->uuid, dc->sb.uuid, 16);
+		memcpy(u->label, dc->sb.label, SB_LABEL_SIZE);
+		u->first_reg = u->last_reg = rtime;
+		bch_uuid_write(c);
+
+		memcpy(dc->sb.set_uuid, c->sb.set_uuid, 16);
+		SET_BDEV_STATE(&dc->sb, BDEV_STATE_CLEAN);
+
+		bch_write_bdev_super(dc, &cl);
+		closure_sync(&cl);
+	} else {
+		u->last_reg = rtime;
+		bch_uuid_write(c);
+	}
+
+	bcache_device_attach(&dc->disk, c, u - c->uuids);
+	list_move(&dc->list, &c->cached_devs);
+	calc_cached_dev_sectors(c);
+
+	smp_wmb();
+	/*
+	 * dc->c must be set before dc->count != 0 - paired with the mb in
+	 * cached_dev_get()
+	 */
+	atomic_set(&dc->count, 1);
+
+	if (BDEV_STATE(&dc->sb) == BDEV_STATE_DIRTY) {
+		atomic_set(&dc->has_dirty, 1);
+		atomic_inc(&dc->count);
+		bch_writeback_queue(dc);
+	}
+
+	bch_cached_dev_run(dc);
+	bcache_device_link(&dc->disk, c, "bdev");
+
+	pr_info("Caching %s as %s on set %pU",
+		bdevname(dc->bdev, buf), dc->disk.disk->disk_name,
+		dc->disk.c->sb.set_uuid);
+	return 0;
+}
+
+void bch_cached_dev_release(struct kobject *kobj)
+{
+	struct cached_dev *dc = container_of(kobj, struct cached_dev,
+					     disk.kobj);
+	kfree(dc);
+	module_put(THIS_MODULE);
+}
+
+static void cached_dev_free(struct closure *cl)
+{
+	struct cached_dev *dc = container_of(cl, struct cached_dev, disk.cl);
+
+	cancel_delayed_work_sync(&dc->writeback_rate_update);
+
+	mutex_lock(&bch_register_lock);
+
+	bd_unlink_disk_holder(dc->bdev, dc->disk.disk);
+	bcache_device_free(&dc->disk);
+	list_del(&dc->list);
+
+	mutex_unlock(&bch_register_lock);
+
+	if (!IS_ERR_OR_NULL(dc->bdev)) {
+		blk_sync_queue(bdev_get_queue(dc->bdev));
+		blkdev_put(dc->bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
+	}
+
+	wake_up(&unregister_wait);
+
+	kobject_put(&dc->disk.kobj);
+}
+
+static void cached_dev_flush(struct closure *cl)
+{
+	struct cached_dev *dc = container_of(cl, struct cached_dev, disk.cl);
+	struct bcache_device *d = &dc->disk;
+
+	bch_cache_accounting_destroy(&dc->accounting);
+	kobject_del(&d->kobj);
+
+	continue_at(cl, cached_dev_free, system_wq);
+}
+
+static int cached_dev_init(struct cached_dev *dc, unsigned block_size)
+{
+	int err;
+	struct io *io;
+
+	closure_init(&dc->disk.cl, NULL);
+	set_closure_fn(&dc->disk.cl, cached_dev_flush, system_wq);
+
+	__module_get(THIS_MODULE);
+	INIT_LIST_HEAD(&dc->list);
+	kobject_init(&dc->disk.kobj, &bch_cached_dev_ktype);
+
+	bch_cache_accounting_init(&dc->accounting, &dc->disk.cl);
+
+	err = bcache_device_init(&dc->disk, block_size);
+	if (err)
+		goto err;
+
+	spin_lock_init(&dc->io_lock);
+	closure_init_unlocked(&dc->sb_write);
+	INIT_WORK(&dc->detach, cached_dev_detach_finish);
+
+	dc->sequential_merge		= true;
+	dc->sequential_cutoff		= 4 << 20;
+
+	INIT_LIST_HEAD(&dc->io_lru);
+	dc->sb_bio.bi_max_vecs	= 1;
+	dc->sb_bio.bi_io_vec	= dc->sb_bio.bi_inline_vecs;
+
+	for (io = dc->io; io < dc->io + RECENT_IO; io++) {
+		list_add(&io->lru, &dc->io_lru);
+		hlist_add_head(&io->hash, dc->io_hash + RECENT_IO);
+	}
+
+	bch_writeback_init_cached_dev(dc);
+	return 0;
+err:
+	bcache_device_stop(&dc->disk);
+	return err;
+}
+
+/* Cached device - bcache superblock */
+
+static const char *register_bdev(struct cache_sb *sb, struct page *sb_page,
+				 struct block_device *bdev,
+				 struct cached_dev *dc)
+{
+	char name[BDEVNAME_SIZE];
+	const char *err = "cannot allocate memory";
+	struct gendisk *g;
+	struct cache_set *c;
+
+	if (!dc || cached_dev_init(dc, sb->block_size << 9) != 0)
+		return err;
+
+	memcpy(&dc->sb, sb, sizeof(struct cache_sb));
+	dc->sb_bio.bi_io_vec[0].bv_page = sb_page;
+	dc->bdev = bdev;
+	dc->bdev->bd_holder = dc;
+
+	g = dc->disk.disk;
+
+	set_capacity(g, dc->bdev->bd_part->nr_sects - dc->sb.data_offset);
+
+	g->queue->backing_dev_info.ra_pages =
+		max(g->queue->backing_dev_info.ra_pages,
+		    bdev->bd_queue->backing_dev_info.ra_pages);
+
+	bch_cached_dev_request_init(dc);
+
+	err = "error creating kobject";
+	if (kobject_add(&dc->disk.kobj, &part_to_dev(bdev->bd_part)->kobj,
+			"bcache"))
+		goto err;
+	if (bch_cache_accounting_add_kobjs(&dc->accounting, &dc->disk.kobj))
+		goto err;
+
+	list_add(&dc->list, &uncached_devices);
+	list_for_each_entry(c, &bch_cache_sets, list)
+		bch_cached_dev_attach(dc, c);
+
+	if (BDEV_STATE(&dc->sb) == BDEV_STATE_NONE ||
+	    BDEV_STATE(&dc->sb) == BDEV_STATE_STALE)
+		bch_cached_dev_run(dc);
+
+	return NULL;
+err:
+	kobject_put(&dc->disk.kobj);
+	pr_notice("error opening %s: %s", bdevname(bdev, name), err);
+	/*
+	 * Return NULL instead of an error because kobject_put() cleans
+	 * everything up
+	 */
+	return NULL;
+}
+
+/* Flash only volumes */
+
+void bch_flash_dev_release(struct kobject *kobj)
+{
+	struct bcache_device *d = container_of(kobj, struct bcache_device,
+					       kobj);
+	kfree(d);
+}
+
+static void flash_dev_free(struct closure *cl)
+{
+	struct bcache_device *d = container_of(cl, struct bcache_device, cl);
+	bcache_device_free(d);
+	kobject_put(&d->kobj);
+}
+
+static void flash_dev_flush(struct closure *cl)
+{
+	struct bcache_device *d = container_of(cl, struct bcache_device, cl);
+
+	bcache_device_unlink(d);
+	kobject_del(&d->kobj);
+	continue_at(cl, flash_dev_free, system_wq);
+}
+
+static int flash_dev_run(struct cache_set *c, struct uuid_entry *u)
+{
+	struct bcache_device *d = kzalloc(sizeof(struct bcache_device),
+					  GFP_KERNEL);
+	if (!d)
+		return -ENOMEM;
+
+	closure_init(&d->cl, NULL);
+	set_closure_fn(&d->cl, flash_dev_flush, system_wq);
+
+	kobject_init(&d->kobj, &bch_flash_dev_ktype);
+
+	if (bcache_device_init(d, block_bytes(c)))
+		goto err;
+
+	bcache_device_attach(d, c, u - c->uuids);
+	set_capacity(d->disk, u->sectors);
+	bch_flash_dev_request_init(d);
+	add_disk(d->disk);
+
+	if (kobject_add(&d->kobj, &disk_to_dev(d->disk)->kobj, "bcache"))
+		goto err;
+
+	bcache_device_link(d, c, "volume");
+
+	return 0;
+err:
+	kobject_put(&d->kobj);
+	return -ENOMEM;
+}
+
+static int flash_devs_run(struct cache_set *c)
+{
+	int ret = 0;
+	struct uuid_entry *u;
+
+	for (u = c->uuids;
+	     u < c->uuids + c->nr_uuids && !ret;
+	     u++)
+		if (UUID_FLASH_ONLY(u))
+			ret = flash_dev_run(c, u);
+
+	return ret;
+}
+
+int bch_flash_dev_create(struct cache_set *c, uint64_t size)
+{
+	struct uuid_entry *u;
+
+	if (test_bit(CACHE_SET_STOPPING, &c->flags))
+		return -EINTR;
+
+	u = uuid_find_empty(c);
+	if (!u) {
+		pr_err("Can't create volume, no room for UUID");
+		return -EINVAL;
+	}
+
+	get_random_bytes(u->uuid, 16);
+	memset(u->label, 0, 32);
+	u->first_reg = u->last_reg = cpu_to_le32(get_seconds());
+
+	SET_UUID_FLASH_ONLY(u, 1);
+	u->sectors = size >> 9;
+
+	bch_uuid_write(c);
+
+	return flash_dev_run(c, u);
+}
+
+/* Cache set */
+
+__printf(2, 3)
+bool bch_cache_set_error(struct cache_set *c, const char *fmt, ...)
+{
+	va_list args;
+
+	if (test_bit(CACHE_SET_STOPPING, &c->flags))
+		return false;
+
+	/* XXX: we can be called from atomic context
+	acquire_console_sem();
+	*/
+
+	printk(KERN_ERR "bcache: error on %pU: ", c->sb.set_uuid);
+
+	va_start(args, fmt);
+	vprintk(fmt, args);
+	va_end(args);
+
+	printk(", disabling caching\n");
+
+	bch_cache_set_unregister(c);
+	return true;
+}
+
+void bch_cache_set_release(struct kobject *kobj)
+{
+	struct cache_set *c = container_of(kobj, struct cache_set, kobj);
+	kfree(c);
+	module_put(THIS_MODULE);
+}
+
+static void cache_set_free(struct closure *cl)
+{
+	struct cache_set *c = container_of(cl, struct cache_set, cl);
+	struct cache *ca;
+	unsigned i;
+
+	if (!IS_ERR_OR_NULL(c->debug))
+		debugfs_remove(c->debug);
+
+	bch_open_buckets_free(c);
+	bch_btree_cache_free(c);
+	bch_journal_free(c);
+
+	for_each_cache(ca, c, i)
+		if (ca)
+			kobject_put(&ca->kobj);
+
+	free_pages((unsigned long) c->uuids, ilog2(bucket_pages(c)));
+	free_pages((unsigned long) c->sort, ilog2(bucket_pages(c)));
+
+	kfree(c->fill_iter);
+	if (c->bio_split)
+		bioset_free(c->bio_split);
+	if (c->bio_meta)
+		mempool_destroy(c->bio_meta);
+	if (c->search)
+		mempool_destroy(c->search);
+	kfree(c->devices);
+
+	mutex_lock(&bch_register_lock);
+	list_del(&c->list);
+	mutex_unlock(&bch_register_lock);
+
+	pr_info("Cache set %pU unregistered", c->sb.set_uuid);
+	wake_up(&unregister_wait);
+
+	closure_debug_destroy(&c->cl);
+	kobject_put(&c->kobj);
+}
+
+static void cache_set_flush(struct closure *cl)
+{
+	struct cache_set *c = container_of(cl, struct cache_set, caching);
+	struct btree *b;
+
+	/* Shut down allocator threads */
+	set_bit(CACHE_SET_STOPPING_2, &c->flags);
+	wake_up(&c->alloc_wait);
+
+	bch_cache_accounting_destroy(&c->accounting);
+
+	kobject_put(&c->internal);
+	kobject_del(&c->kobj);
+
+	if (!IS_ERR_OR_NULL(c->root))
+		list_add(&c->root->list, &c->btree_cache);
+
+	/* Should skip this if we're unregistering because of an error */
+	list_for_each_entry(b, &c->btree_cache, list)
+		if (btree_node_dirty(b))
+			bch_btree_write(b, true, NULL);
+
+	closure_return(cl);
+}
+
+static void __cache_set_unregister(struct closure *cl)
+{
+	struct cache_set *c = container_of(cl, struct cache_set, caching);
+	struct cached_dev *dc, *t;
+	size_t i;
+
+	mutex_lock(&bch_register_lock);
+
+	if (test_bit(CACHE_SET_UNREGISTERING, &c->flags))
+		list_for_each_entry_safe(dc, t, &c->cached_devs, list)
+			bch_cached_dev_detach(dc);
+
+	for (i = 0; i < c->nr_uuids; i++)
+		if (c->devices[i] && UUID_FLASH_ONLY(&c->uuids[i]))
+			bcache_device_stop(c->devices[i]);
+
+	mutex_unlock(&bch_register_lock);
+
+	continue_at(cl, cache_set_flush, system_wq);
+}
+
+void bch_cache_set_stop(struct cache_set *c)
+{
+	if (!test_and_set_bit(CACHE_SET_STOPPING, &c->flags))
+		closure_queue(&c->caching);
+}
+
+void bch_cache_set_unregister(struct cache_set *c)
+{
+	set_bit(CACHE_SET_UNREGISTERING, &c->flags);
+	bch_cache_set_stop(c);
+}
+
+#define alloc_bucket_pages(gfp, c)			\
+	((void *) __get_free_pages(__GFP_ZERO|gfp, ilog2(bucket_pages(c))))
+
+struct cache_set *bch_cache_set_alloc(struct cache_sb *sb)
+{
+	int iter_size;
+	struct cache_set *c = kzalloc(sizeof(struct cache_set), GFP_KERNEL);
+	if (!c)
+		return NULL;
+
+	__module_get(THIS_MODULE);
+	closure_init(&c->cl, NULL);
+	set_closure_fn(&c->cl, cache_set_free, system_wq);
+
+	closure_init(&c->caching, &c->cl);
+	set_closure_fn(&c->caching, __cache_set_unregister, system_wq);
+
+	/* Maybe create continue_at_noreturn() and use it here? */
+	closure_set_stopped(&c->cl);
+	closure_put(&c->cl);
+
+	kobject_init(&c->kobj, &bch_cache_set_ktype);
+	kobject_init(&c->internal, &bch_cache_set_internal_ktype);
+
+	bch_cache_accounting_init(&c->accounting, &c->cl);
+
+	memcpy(c->sb.set_uuid, sb->set_uuid, 16);
+	c->sb.block_size	= sb->block_size;
+	c->sb.bucket_size	= sb->bucket_size;
+	c->sb.nr_in_set		= sb->nr_in_set;
+	c->sb.last_mount	= sb->last_mount;
+	c->bucket_bits		= ilog2(sb->bucket_size);
+	c->block_bits		= ilog2(sb->block_size);
+	c->nr_uuids		= bucket_bytes(c) / sizeof(struct uuid_entry);
+
+	c->btree_pages		= c->sb.bucket_size / PAGE_SECTORS;
+	if (c->btree_pages > BTREE_MAX_PAGES)
+		c->btree_pages = max_t(int, c->btree_pages / 4,
+				       BTREE_MAX_PAGES);
+
+	init_waitqueue_head(&c->alloc_wait);
+	mutex_init(&c->bucket_lock);
+	mutex_init(&c->fill_lock);
+	mutex_init(&c->sort_lock);
+	spin_lock_init(&c->sort_time_lock);
+	closure_init_unlocked(&c->sb_write);
+	closure_init_unlocked(&c->uuid_write);
+	spin_lock_init(&c->btree_read_time_lock);
+	bch_moving_init_cache_set(c);
+
+	INIT_LIST_HEAD(&c->list);
+	INIT_LIST_HEAD(&c->cached_devs);
+	INIT_LIST_HEAD(&c->btree_cache);
+	INIT_LIST_HEAD(&c->btree_cache_freeable);
+	INIT_LIST_HEAD(&c->btree_cache_freed);
+	INIT_LIST_HEAD(&c->data_buckets);
+
+	c->search = mempool_create_slab_pool(32, bch_search_cache);
+	if (!c->search)
+		goto err;
+
+	iter_size = (sb->bucket_size / sb->block_size + 1) *
+		sizeof(struct btree_iter_set);
+
+	if (!(c->devices = kzalloc(c->nr_uuids * sizeof(void *), GFP_KERNEL)) ||
+	    !(c->bio_meta = mempool_create_kmalloc_pool(2,
+				sizeof(struct bbio) + sizeof(struct bio_vec) *
+				bucket_pages(c))) ||
+	    !(c->bio_split = bioset_create(4, offsetof(struct bbio, bio))) ||
+	    !(c->fill_iter = kmalloc(iter_size, GFP_KERNEL)) ||
+	    !(c->sort = alloc_bucket_pages(GFP_KERNEL, c)) ||
+	    !(c->uuids = alloc_bucket_pages(GFP_KERNEL, c)) ||
+	    bch_journal_alloc(c) ||
+	    bch_btree_cache_alloc(c) ||
+	    bch_open_buckets_alloc(c))
+		goto err;
+
+	c->fill_iter->size = sb->bucket_size / sb->block_size;
+
+	c->congested_read_threshold_us	= 2000;
+	c->congested_write_threshold_us	= 20000;
+	c->error_limit	= 8 << IO_ERROR_SHIFT;
+
+	return c;
+err:
+	bch_cache_set_unregister(c);
+	return NULL;
+}
+
+static void run_cache_set(struct cache_set *c)
+{
+	const char *err = "cannot allocate memory";
+	struct cached_dev *dc, *t;
+	struct cache *ca;
+	unsigned i;
+
+	struct btree_op op;
+	bch_btree_op_init_stack(&op);
+	op.lock = SHRT_MAX;
+
+	for_each_cache(ca, c, i)
+		c->nbuckets += ca->sb.nbuckets;
+
+	if (CACHE_SYNC(&c->sb)) {
+		LIST_HEAD(journal);
+		struct bkey *k;
+		struct jset *j;
+
+		err = "cannot allocate memory for journal";
+		if (bch_journal_read(c, &journal, &op))
+			goto err;
+
+		pr_debug("btree_journal_read() done");
+
+		err = "no journal entries found";
+		if (list_empty(&journal))
+			goto err;
+
+		j = &list_entry(journal.prev, struct journal_replay, list)->j;
+
+		err = "IO error reading priorities";
+		for_each_cache(ca, c, i)
+			prio_read(ca, j->prio_bucket[ca->sb.nr_this_dev]);
+
+		/*
+		 * If prio_read() fails it'll call cache_set_error and we'll
+		 * tear everything down right away, but if we perhaps checked
+		 * sooner we could avoid journal replay.
+		 */
+
+		k = &j->btree_root;
+
+		err = "bad btree root";
+		if (__bch_ptr_invalid(c, j->btree_level + 1, k))
+			goto err;
+
+		err = "error reading btree root";
+		c->root = bch_btree_node_get(c, k, j->btree_level, &op);
+		if (IS_ERR_OR_NULL(c->root))
+			goto err;
+
+		list_del_init(&c->root->list);
+		rw_unlock(true, c->root);
+
+		err = uuid_read(c, j, &op.cl);
+		if (err)
+			goto err;
+
+		err = "error in recovery";
+		if (bch_btree_check(c, &op))
+			goto err;
+
+		bch_journal_mark(c, &journal);
+		bch_btree_gc_finish(c);
+		pr_debug("btree_check() done");
+
+		/*
+		 * bcache_journal_next() can't happen sooner, or
+		 * btree_gc_finish() will give spurious errors about last_gc >
+		 * gc_gen - this is a hack but oh well.
+		 */
+		bch_journal_next(&c->journal);
+
+		for_each_cache(ca, c, i)
+			closure_call(&ca->alloc, bch_allocator_thread,
+				     system_wq, &c->cl);
+
+		/*
+		 * First place it's safe to allocate: btree_check() and
+		 * btree_gc_finish() have to run before we have buckets to
+		 * allocate, and bch_bucket_alloc_set() might cause a journal
+		 * entry to be written so bcache_journal_next() has to be called
+		 * first.
+		 *
+		 * If the uuids were in the old format we have to rewrite them
+		 * before the next journal entry is written:
+		 */
+		if (j->version < BCACHE_JSET_VERSION_UUID)
+			__uuid_write(c);
+
+		bch_journal_replay(c, &journal, &op);
+	} else {
+		pr_notice("invalidating existing data");
+		/* Don't want invalidate_buckets() to queue a gc yet */
+		closure_lock(&c->gc, NULL);
+
+		for_each_cache(ca, c, i) {
+			unsigned j;
+
+			ca->sb.keys = clamp_t(int, ca->sb.nbuckets >> 7,
+					      2, SB_JOURNAL_BUCKETS);
+
+			for (j = 0; j < ca->sb.keys; j++)
+				ca->sb.d[j] = ca->sb.first_bucket + j;
+		}
+
+		bch_btree_gc_finish(c);
+
+		for_each_cache(ca, c, i)
+			closure_call(&ca->alloc, bch_allocator_thread,
+				     ca->alloc_workqueue, &c->cl);
+
+		mutex_lock(&c->bucket_lock);
+		for_each_cache(ca, c, i)
+			bch_prio_write(ca);
+		mutex_unlock(&c->bucket_lock);
+
+		wake_up(&c->alloc_wait);
+
+		err = "cannot allocate new UUID bucket";
+		if (__uuid_write(c))
+			goto err_unlock_gc;
+
+		err = "cannot allocate new btree root";
+		c->root = bch_btree_node_alloc(c, 0, &op.cl);
+		if (IS_ERR_OR_NULL(c->root))
+			goto err_unlock_gc;
+
+		bkey_copy_key(&c->root->key, &MAX_KEY);
+		bch_btree_write(c->root, true, &op);
+
+		bch_btree_set_root(c->root);
+		rw_unlock(true, c->root);
+
+		/*
+		 * We don't want to write the first journal entry until
+		 * everything is set up - fortunately journal entries won't be
+		 * written until the SET_CACHE_SYNC() here:
+		 */
+		SET_CACHE_SYNC(&c->sb, true);
+
+		bch_journal_next(&c->journal);
+		bch_journal_meta(c, &op.cl);
+
+		/* Unlock */
+		closure_set_stopped(&c->gc.cl);
+		closure_put(&c->gc.cl);
+	}
+
+	closure_sync(&op.cl);
+	c->sb.last_mount = get_seconds();
+	bcache_write_super(c);
+
+	list_for_each_entry_safe(dc, t, &uncached_devices, list)
+		bch_cached_dev_attach(dc, c);
+
+	flash_devs_run(c);
+
+	return;
+err_unlock_gc:
+	closure_set_stopped(&c->gc.cl);
+	closure_put(&c->gc.cl);
+err:
+	closure_sync(&op.cl);
+	/* XXX: test this, it's broken */
+	bch_cache_set_error(c, err);
+}
+
+static bool can_attach_cache(struct cache *ca, struct cache_set *c)
+{
+	return ca->sb.block_size	== c->sb.block_size &&
+		ca->sb.bucket_size	== c->sb.block_size &&
+		ca->sb.nr_in_set	== c->sb.nr_in_set;
+}
+
+static const char *register_cache_set(struct cache *ca)
+{
+	char buf[12];
+	const char *err = "cannot allocate memory";
+	struct cache_set *c;
+
+	list_for_each_entry(c, &bch_cache_sets, list)
+		if (!memcmp(c->sb.set_uuid, ca->sb.set_uuid, 16)) {
+			if (c->cache[ca->sb.nr_this_dev])
+				return "duplicate cache set member";
+
+			if (!can_attach_cache(ca, c))
+				return "cache sb does not match set";
+
+			if (!CACHE_SYNC(&ca->sb))
+				SET_CACHE_SYNC(&c->sb, false);
+
+			goto found;
+		}
+
+	c = bch_cache_set_alloc(&ca->sb);
+	if (!c)
+		return err;
+
+	err = "error creating kobject";
+	if (kobject_add(&c->kobj, bcache_kobj, "%pU", c->sb.set_uuid) ||
+	    kobject_add(&c->internal, &c->kobj, "internal"))
+		goto err;
+
+	if (bch_cache_accounting_add_kobjs(&c->accounting, &c->kobj))
+		goto err;
+
+	bch_debug_init_cache_set(c);
+
+	list_add(&c->list, &bch_cache_sets);
+found:
+	sprintf(buf, "cache%i", ca->sb.nr_this_dev);
+	if (sysfs_create_link(&ca->kobj, &c->kobj, "set") ||
+	    sysfs_create_link(&c->kobj, &ca->kobj, buf))
+		goto err;
+
+	if (ca->sb.seq > c->sb.seq) {
+		c->sb.version		= ca->sb.version;
+		memcpy(c->sb.set_uuid, ca->sb.set_uuid, 16);
+		c->sb.flags             = ca->sb.flags;
+		c->sb.seq		= ca->sb.seq;
+		pr_debug("set version = %llu", c->sb.version);
+	}
+
+	ca->set = c;
+	ca->set->cache[ca->sb.nr_this_dev] = ca;
+	c->cache_by_alloc[c->caches_loaded++] = ca;
+
+	if (c->caches_loaded == c->sb.nr_in_set)
+		run_cache_set(c);
+
+	return NULL;
+err:
+	bch_cache_set_unregister(c);
+	return err;
+}
+
+/* Cache device */
+
+void bch_cache_release(struct kobject *kobj)
+{
+	struct cache *ca = container_of(kobj, struct cache, kobj);
+
+	if (ca->set)
+		ca->set->cache[ca->sb.nr_this_dev] = NULL;
+
+	bch_cache_allocator_exit(ca);
+
+	bio_split_pool_free(&ca->bio_split_hook);
+
+	if (ca->alloc_workqueue)
+		destroy_workqueue(ca->alloc_workqueue);
+
+	free_pages((unsigned long) ca->disk_buckets, ilog2(bucket_pages(ca)));
+	kfree(ca->prio_buckets);
+	vfree(ca->buckets);
+
+	free_heap(&ca->heap);
+	free_fifo(&ca->unused);
+	free_fifo(&ca->free_inc);
+	free_fifo(&ca->free);
+
+	if (ca->sb_bio.bi_inline_vecs[0].bv_page)
+		put_page(ca->sb_bio.bi_io_vec[0].bv_page);
+
+	if (!IS_ERR_OR_NULL(ca->bdev)) {
+		blk_sync_queue(bdev_get_queue(ca->bdev));
+		blkdev_put(ca->bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
+	}
+
+	kfree(ca);
+	module_put(THIS_MODULE);
+}
+
+static int cache_alloc(struct cache_sb *sb, struct cache *ca)
+{
+	size_t free;
+	struct bucket *b;
+
+	if (!ca)
+		return -ENOMEM;
+
+	__module_get(THIS_MODULE);
+	kobject_init(&ca->kobj, &bch_cache_ktype);
+
+	memcpy(&ca->sb, sb, sizeof(struct cache_sb));
+
+	INIT_LIST_HEAD(&ca->discards);
+
+	bio_init(&ca->sb_bio);
+	ca->sb_bio.bi_max_vecs	= 1;
+	ca->sb_bio.bi_io_vec	= ca->sb_bio.bi_inline_vecs;
+
+	bio_init(&ca->journal.bio);
+	ca->journal.bio.bi_max_vecs = 8;
+	ca->journal.bio.bi_io_vec = ca->journal.bio.bi_inline_vecs;
+
+	free = roundup_pow_of_two(ca->sb.nbuckets) >> 9;
+	free = max_t(size_t, free, (prio_buckets(ca) + 8) * 2);
+
+	if (!init_fifo(&ca->free,	free, GFP_KERNEL) ||
+	    !init_fifo(&ca->free_inc,	free << 2, GFP_KERNEL) ||
+	    !init_fifo(&ca->unused,	free << 2, GFP_KERNEL) ||
+	    !init_heap(&ca->heap,	free << 3, GFP_KERNEL) ||
+	    !(ca->buckets	= vmalloc(sizeof(struct bucket) *
+					  ca->sb.nbuckets)) ||
+	    !(ca->prio_buckets	= kzalloc(sizeof(uint64_t) * prio_buckets(ca) *
+					  2, GFP_KERNEL)) ||
+	    !(ca->disk_buckets	= alloc_bucket_pages(GFP_KERNEL, ca)) ||
+	    !(ca->alloc_workqueue = alloc_workqueue("bch_allocator", 0, 1)) ||
+	    bio_split_pool_init(&ca->bio_split_hook))
+		goto err;
+
+	ca->prio_last_buckets = ca->prio_buckets + prio_buckets(ca);
+
+	memset(ca->buckets, 0, ca->sb.nbuckets * sizeof(struct bucket));
+	for_each_bucket(b, ca)
+		atomic_set(&b->pin, 0);
+
+	if (bch_cache_allocator_init(ca))
+		goto err;
+
+	return 0;
+err:
+	kobject_put(&ca->kobj);
+	return -ENOMEM;
+}
+
+static const char *register_cache(struct cache_sb *sb, struct page *sb_page,
+				  struct block_device *bdev, struct cache *ca)
+{
+	char name[BDEVNAME_SIZE];
+	const char *err = "cannot allocate memory";
+
+	if (cache_alloc(sb, ca) != 0)
+		return err;
+
+	ca->sb_bio.bi_io_vec[0].bv_page = sb_page;
+	ca->bdev = bdev;
+	ca->bdev->bd_holder = ca;
+
+	if (blk_queue_discard(bdev_get_queue(ca->bdev)))
+		ca->discard = CACHE_DISCARD(&ca->sb);
+
+	err = "error creating kobject";
+	if (kobject_add(&ca->kobj, &part_to_dev(bdev->bd_part)->kobj, "bcache"))
+		goto err;
+
+	err = register_cache_set(ca);
+	if (err)
+		goto err;
+
+	pr_info("registered cache device %s", bdevname(bdev, name));
+
+	return NULL;
+err:
+	kobject_put(&ca->kobj);
+	pr_info("error opening %s: %s", bdevname(bdev, name), err);
+	/* Return NULL instead of an error because kobject_put() cleans
+	 * everything up
+	 */
+	return NULL;
+}
+
+/* Global interfaces/init */
+
+static ssize_t register_bcache(struct kobject *, struct kobj_attribute *,
+			       const char *, size_t);
+
+kobj_attribute_write(register,		register_bcache);
+kobj_attribute_write(register_quiet,	register_bcache);
+
+static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
+			       const char *buffer, size_t size)
+{
+	ssize_t ret = size;
+	const char *err = "cannot allocate memory";
+	char *path = NULL;
+	struct cache_sb *sb = NULL;
+	struct block_device *bdev = NULL;
+	struct page *sb_page = NULL;
+
+	if (!try_module_get(THIS_MODULE))
+		return -EBUSY;
+
+	mutex_lock(&bch_register_lock);
+
+	if (!(path = kstrndup(buffer, size, GFP_KERNEL)) ||
+	    !(sb = kmalloc(sizeof(struct cache_sb), GFP_KERNEL)))
+		goto err;
+
+	err = "failed to open device";
+	bdev = blkdev_get_by_path(strim(path),
+				  FMODE_READ|FMODE_WRITE|FMODE_EXCL,
+				  sb);
+	if (bdev == ERR_PTR(-EBUSY))
+		err = "device busy";
+
+	if (IS_ERR(bdev) ||
+	    set_blocksize(bdev, 4096))
+		goto err;
+
+	err = read_super(sb, bdev, &sb_page);
+	if (err)
+		goto err_close;
+
+	if (SB_IS_BDEV(sb)) {
+		struct cached_dev *dc = kzalloc(sizeof(*dc), GFP_KERNEL);
+
+		err = register_bdev(sb, sb_page, bdev, dc);
+	} else {
+		struct cache *ca = kzalloc(sizeof(*ca), GFP_KERNEL);
+
+		err = register_cache(sb, sb_page, bdev, ca);
+	}
+
+	if (err) {
+		/* register_(bdev|cache) will only return an error if they
+		 * didn't get far enough to create the kobject - if they did,
+		 * the kobject destructor will do this cleanup.
+		 */
+		put_page(sb_page);
+err_close:
+		blkdev_put(bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
+err:
+		if (attr != &ksysfs_register_quiet)
+			pr_info("error opening %s: %s", path, err);
+		ret = -EINVAL;
+	}
+
+	kfree(sb);
+	kfree(path);
+	mutex_unlock(&bch_register_lock);
+	module_put(THIS_MODULE);
+	return ret;
+}
+
+static int bcache_reboot(struct notifier_block *n, unsigned long code, void *x)
+{
+	if (code == SYS_DOWN ||
+	    code == SYS_HALT ||
+	    code == SYS_POWER_OFF) {
+		DEFINE_WAIT(wait);
+		unsigned long start = jiffies;
+		bool stopped = false;
+
+		struct cache_set *c, *tc;
+		struct cached_dev *dc, *tdc;
+
+		mutex_lock(&bch_register_lock);
+
+		if (list_empty(&bch_cache_sets) &&
+		    list_empty(&uncached_devices))
+			goto out;
+
+		pr_info("Stopping all devices:");
+
+		list_for_each_entry_safe(c, tc, &bch_cache_sets, list)
+			bch_cache_set_stop(c);
+
+		list_for_each_entry_safe(dc, tdc, &uncached_devices, list)
+			bcache_device_stop(&dc->disk);
+
+		/* What's a condition variable? */
+		while (1) {
+			long timeout = start + 2 * HZ - jiffies;
+
+			stopped = list_empty(&bch_cache_sets) &&
+				list_empty(&uncached_devices);
+
+			if (timeout < 0 || stopped)
+				break;
+
+			prepare_to_wait(&unregister_wait, &wait,
+					TASK_UNINTERRUPTIBLE);
+
+			mutex_unlock(&bch_register_lock);
+			schedule_timeout(timeout);
+			mutex_lock(&bch_register_lock);
+		}
+
+		finish_wait(&unregister_wait, &wait);
+
+		if (stopped)
+			pr_info("All devices stopped");
+		else
+			pr_notice("Timeout waiting for devices to be closed");
+out:
+		mutex_unlock(&bch_register_lock);
+	}
+
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block reboot = {
+	.notifier_call	= bcache_reboot,
+	.priority	= INT_MAX, /* before any real devices */
+};
+
+static void bcache_exit(void)
+{
+	bch_debug_exit();
+	bch_writeback_exit();
+	bch_request_exit();
+	bch_btree_exit();
+	if (bcache_kobj)
+		kobject_put(bcache_kobj);
+	if (bcache_wq)
+		destroy_workqueue(bcache_wq);
+	unregister_blkdev(bcache_major, "bcache");
+	unregister_reboot_notifier(&reboot);
+}
+
+static int __init bcache_init(void)
+{
+	static const struct attribute *files[] = {
+		&ksysfs_register.attr,
+		&ksysfs_register_quiet.attr,
+		NULL
+	};
+
+	mutex_init(&bch_register_lock);
+	init_waitqueue_head(&unregister_wait);
+	register_reboot_notifier(&reboot);
+	closure_debug_init();
+
+	bcache_major = register_blkdev(0, "bcache");
+	if (bcache_major < 0)
+		return bcache_major;
+
+	if (!(bcache_wq = create_workqueue("bcache")) ||
+	    !(bcache_kobj = kobject_create_and_add("bcache", fs_kobj)) ||
+	    sysfs_create_files(bcache_kobj, files) ||
+	    bch_btree_init() ||
+	    bch_request_init() ||
+	    bch_writeback_init() ||
+	    bch_debug_init(bcache_kobj))
+		goto err;
+
+	return 0;
+err:
+	bcache_exit();
+	return -ENOMEM;
+}
+
+module_exit(bcache_exit);
+module_init(bcache_init);

drivers/md/bcache/sysfs.c

···
+/*
+ * bcache sysfs interfaces
+ *
+ * Copyright 2010, 2011 Kent Overstreet <[e-mail]>
+ * Copyright 2012 Google, Inc.
+ */
+
+#include "bcache.h"
+#include "sysfs.h"
+#include "btree.h"
+#include "request.h"
+
+#include <linux/sort.h>
+
+static const char * const cache_replacement_policies[] = {
+	"lru",
+	"fifo",
+	"random",
+	NULL
+};
+
+write_attribute(attach);
+write_attribute(detach);
+write_attribute(unregister);
+write_attribute(stop);
+write_attribute(clear_stats);
+write_attribute(trigger_gc);
+write_attribute(prune_cache);
+write_attribute(flash_vol_create);
+
+read_attribute(bucket_size);
+read_attribute(block_size);
+read_attribute(nbuckets);
+read_attribute(tree_depth);
+read_attribute(root_usage_percent);
+read_attribute(priority_stats);
+read_attribute(btree_cache_size);
+read_attribute(btree_cache_max_chain);
+read_attribute(cache_available_percent);
+read_attribute(written);
+read_attribute(btree_written);
+read_attribute(metadata_written);
+read_attribute(active_journal_entries);
+
+sysfs_time_stats_attribute(btree_gc,	sec, ms);
+sysfs_time_stats_attribute(btree_split, sec, us);
+sysfs_time_stats_attribute(btree_sort,	ms,  us);
+sysfs_time_stats_attribute(btree_read,	ms,  us);
+sysfs_time_stats_attribute(try_harder,	ms,  us);
+
+read_attribute(btree_nodes);
+read_attribute(btree_used_percent);
+read_attribute(average_key_size);
+read_attribute(dirty_data);
+read_attribute(bset_tree_stats);
+
+read_attribute(state);
+read_attribute(cache_read_races);
+read_attribute(writeback_keys_done);
+read_attribute(writeback_keys_failed);
+read_attribute(io_errors);
+read_attribute(congested);
+rw_attribute(congested_read_threshold_us);
+rw_attribute(congested_write_threshold_us);
+
+rw_attribute(sequential_cutoff);
+rw_attribute(sequential_merge);
+rw_attribute(data_csum);
+rw_attribute(cache_mode);
+rw_attribute(writeback_metadata);
+rw_attribute(writeback_running);
+rw_attribute(writeback_percent);
+rw_attribute(writeback_delay);
+rw_attribute(writeback_rate);
+
+rw_attribute(writeback_rate_update_seconds);
+rw_attribute(writeback_rate_d_term);
+rw_attribute(writeback_rate_p_term_inverse);
+rw_attribute(writeback_rate_d_smooth);
+read_attribute(writeback_rate_debug);
+
+rw_attribute(synchronous);
+rw_attribute(journal_delay_ms);
+rw_attribute(discard);
+rw_attribute(running);
+rw_attribute(label);
+rw_attribute(readahead);
+rw_attribute(io_error_limit);
+rw_attribute(io_error_halflife);
+rw_attribute(verify);
+rw_attribute(key_merging_disabled);
+rw_attribute(gc_always_rewrite);
+rw_attribute(freelist_percent);
+rw_attribute(cache_replacement_policy);
+rw_attribute(btree_shrinker_disabled);
+rw_attribute(copy_gc_enabled);
+rw_attribute(size);
+
+SHOW(__bch_cached_dev)
+{
+	struct cached_dev *dc = container_of(kobj, struct cached_dev,
+					     disk.kobj);
+	const char *states[] = { "no cache", "clean", "dirty", "inconsistent" };
+
+#define var(stat)		(dc->stat)
+
+	if (attr == &sysfs_cache_mode)
+		return bch_snprint_string_list(buf, PAGE_SIZE,
+					       bch_cache_modes + 1,
+					       BDEV_CACHE_MODE(&dc->sb));
+
+	sysfs_printf(data_csum,		"%i", dc->disk.data_csum);
+	var_printf(verify,		"%i");
+	var_printf(writeback_metadata,	"%i");
+	var_printf(writeback_running,	"%i");
+	var_print(writeback_delay);
+	var_print(writeback_percent);
+	sysfs_print(writeback_rate,	dc->writeback_rate.rate);
+
+	var_print(writeback_rate_update_seconds);
+	var_print(writeback_rate_d_term);
+	var_print(writeback_rate_p_term_inverse);
+	var_print(writeback_rate_d_smooth);
+
+	if (attr == &sysfs_writeback_rate_debug) {
+		char dirty[20];
+		char derivative[20];
+		char target[20];
+		bch_hprint(dirty,
+		       atomic_long_read(&dc->disk.sectors_dirty) << 9);
+		bch_hprint(derivative,	dc->writeback_rate_derivative << 9);
+		bch_hprint(target,	dc->writeback_rate_target << 9);
+
+		return sprintf(buf,
+			       "rate:\t\t%u\n"
+			       "change:\t\t%i\n"
+			       "dirty:\t\t%s\n"
+			       "derivative:\t%s\n"
+			       "target:\t\t%s\n",
+			       dc->writeback_rate.rate,
+			       dc->writeback_rate_change,
+			       dirty, derivative, target);
+	}
+
+	sysfs_hprint(dirty_data,
+		     atomic_long_read(&dc->disk.sectors_dirty) << 9);
+
+	var_printf(sequential_merge,	"%i");
+	var_hprint(sequential_cutoff);
+	var_hprint(readahead);
+
+	sysfs_print(running,		atomic_read(&dc->running));
+	sysfs_print(state,		states[BDEV_STATE(&dc->sb)]);
+
+	if (attr == &sysfs_label) {
+		memcpy(buf, dc->sb.label, SB_LABEL_SIZE);
+		buf[SB_LABEL_SIZE + 1] = '\0';
+		strcat(buf, "\n");
+		return strlen(buf);
+	}
+
+#undef var
+	return 0;
+}
+SHOW_LOCKED(bch_cached_dev)
+
+STORE(__cached_dev)
+{
+	struct cached_dev *dc = container_of(kobj, struct cached_dev,
+					     disk.kobj);
+	unsigned v = size;
+	struct cache_set *c;
+
+#define d_strtoul(var)		sysfs_strtoul(var, dc->var)
+#define d_strtoi_h(var)		sysfs_hatoi(var, dc->var)
+
+	sysfs_strtoul(data_csum,	dc->disk.data_csum);
+	d_strtoul(verify);
+	d_strtoul(writeback_metadata);
+	d_strtoul(writeback_running);
+	d_strtoul(writeback_delay);
+	sysfs_strtoul_clamp(writeback_rate,
+			    dc->writeback_rate.rate, 1, 1000000);
+	sysfs_strtoul_clamp(writeback_percent, dc->writeback_percent, 0, 40);
+
+	d_strtoul(writeback_rate_update_seconds);
+	d_strtoul(writeback_rate_d_term);
+	d_strtoul(writeback_rate_p_term_inverse);
+	sysfs_strtoul_clamp(writeback_rate_p_term_inverse,
+			    dc->writeback_rate_p_term_inverse, 1, INT_MAX);
+	d_strtoul(writeback_rate_d_smooth);
+
+	d_strtoul(sequential_merge);
+	d_strtoi_h(sequential_cutoff);
+	d_strtoi_h(readahead);
+
+	if (attr == &sysfs_clear_stats)
+		bch_cache_accounting_clear(&dc->accounting);
+
+	if (attr == &sysfs_running &&
+	    strtoul_or_return(buf))
+		bch_cached_dev_run(dc);
+
+	if (attr == &sysfs_cache_mode) {
+		ssize_t v = bch_read_string_list(buf, bch_cache_modes + 1);
+
+		if (v < 0)
+			return v;
+
+		if ((unsigned) v != BDEV_CACHE_MODE(&dc->sb)) {
+			SET_BDEV_CACHE_MODE(&dc->sb, v);
+			bch_write_bdev_super(dc, NULL);
+		}
+	}
+
+	if (attr == &sysfs_label) {
+		memcpy(dc->sb.label, buf, SB_LABEL_SIZE);
+		bch_write_bdev_super(dc, NULL);
+		if (dc->disk.c) {
+			memcpy(dc->disk.c->uuids[dc->disk.id].label,
+			       buf, SB_LABEL_SIZE);
+			bch_uuid_write(dc->disk.c);
+		}
+	}
+
+	if (attr == &sysfs_attach) {
+		if (bch_parse_uuid(buf, dc->sb.set_uuid) < 16)
+			return -EINVAL;
+
+		list_for_each_entry(c, &bch_cache_sets, list) {
+			v = bch_cached_dev_attach(dc, c);
+			if (!v)
+				return size;
+		}
+
+		pr_err("Can't attach %s: cache set not found", buf);
+		size = v;
+	}
+
+	if (attr == &sysfs_detach && dc->disk.c)
+		bch_cached_dev_detach(dc);
+
+	if (attr == &sysfs_stop)
+		bcache_device_stop(&dc->disk);
+
+	return size;
+}
+
+STORE(bch_cached_dev)
+{
+	struct cached_dev *dc = container_of(kobj, struct cached_dev,
+					     disk.kobj);
+
+	mutex_lock(&bch_register_lock);
+	size = __cached_dev_store(kobj, attr, buf, size);
+
+	if (attr == &sysfs_writeback_running)
+		bch_writeback_queue(dc);
+
+	if (attr == &sysfs_writeback_percent)
+		schedule_delayed_work(&dc->writeback_rate_update,
+				      dc->writeback_rate_update_seconds * HZ);
+
+	mutex_unlock(&bch_register_lock);
+	return size;
+}
+
+static struct attribute *bch_cached_dev_files[] = {
+	&sysfs_attach,
+	&sysfs_detach,
+	&sysfs_stop,
+#if 0
+	&sysfs_data_csum,
+#endif
+	&sysfs_cache_mode,
+	&sysfs_writeback_metadata,
+	&sysfs_writeback_running,
+	&sysfs_writeback_delay,
+	&sysfs_writeback_percent,
+	&sysfs_writeback_rate,
+	&sysfs_writeback_rate_update_seconds,
+	&sysfs_writeback_rate_d_term,
+	&sysfs_writeback_rate_p_term_inverse,
+	&sysfs_writeback_rate_d_smooth,
+	&sysfs_writeback_rate_debug,
+	&sysfs_dirty_data,
+	&sysfs_sequential_cutoff,
+	&sysfs_sequential_merge,
+	&sysfs_clear_stats,
+	&sysfs_running,
+	&sysfs_state,
+	&sysfs_label,
+	&sysfs_readahead,
+#ifdef CONFIG_BCACHE_DEBUG
+	&sysfs_verify,
+#endif
+	NULL
+};
+KTYPE(bch_cached_dev);
+
+SHOW(bch_flash_dev)
+{
+	struct bcache_device *d = container_of(kobj, struct bcache_device,
+					       kobj);
+	struct uuid_entry *u = &d->c->uuids[d->id];
+
+	sysfs_printf(data_csum,	"%i", d->data_csum);
+	sysfs_hprint(size,	u->sectors << 9);
+
+	if (attr == &sysfs_label) {
+		memcpy(buf, u->label, SB_LABEL_SIZE);
+		buf[SB_LABEL_SIZE + 1] = '\0';
+		strcat(buf, "\n");
+		return strlen(buf);
+	}
+
+	return 0;
+}
+
+STORE(__bch_flash_dev)
+{
+	struct bcache_device *d = container_of(kobj, struct bcache_device,
+					       kobj);
+	struct uuid_entry *u = &d->c->uuids[d->id];
+
+	sysfs_strtoul(data_csum,	d->data_csum);
+
+	if (attr == &sysfs_size) {
+		uint64_t v;
+		strtoi_h_or_return(buf, v);
+
+		u->sectors = v >> 9;
+		bch_uuid_write(d->c);
+		set_capacity(d->disk, u->sectors);
+	}
+
+	if (attr == &sysfs_label) {
+		memcpy(u->label, buf, SB_LABEL_SIZE);
+		bch_uuid_write(d->c);
+	}
+
+	if (attr == &sysfs_unregister) {
+		atomic_set(&d->detaching, 1);
+		bcache_device_stop(d);
+	}
+
+	return size;
+}
+STORE_LOCKED(bch_flash_dev)
+
+static struct attribute *bch_flash_dev_files[] = {
+	&sysfs_unregister,
+#if 0
+	&sysfs_data_csum,
+#endif
+	&sysfs_label,
+	&sysfs_size,
+	NULL
+};
+KTYPE(bch_flash_dev);
+
+SHOW(__bch_cache_set)
+{
+	unsigned root_usage(struct cache_set *c)
+	{
+		unsigned bytes = 0;
+		struct bkey *k;
+		struct btree *b;
+		struct btree_iter iter;
+
+		goto lock_root;
+
+		do {
+			rw_unlock(false, b);
+lock_root:
+			b = c->root;
+			rw_lock(false, b, b->level);
+		} while (b != c->root);
+
+		for_each_key_filter(b, k, &iter, bch_ptr_bad)
+			bytes += bkey_bytes(k);
+
+		rw_unlock(false, b);
+
+		return (bytes * 100) / btree_bytes(c);
+	}
+
+	size_t cache_size(struct cache_set *c)
+	{
+		size_t ret = 0;
+		struct btree *b;
+
+		mutex_lock(&c->bucket_lock);
+		list_for_each_entry(b, &c->btree_cache, list)
+			ret += 1 << (b->page_order + PAGE_SHIFT);
+
+		mutex_unlock(&c->bucket_lock);
+		return ret;
+	}
+
+	unsigned cache_max_chain(struct cache_set *c)
+	{
+		unsigned ret = 0;
+		struct hlist_head *h;
+
+		mutex_lock(&c->bucket_lock);
+
+		for (h = c->bucket_hash;
+		     h < c->bucket_hash + (1 << BUCKET_HASH_BITS);
+		     h++) {
+			unsigned i = 0;
+			struct hlist_node *p;
+
+			hlist_for_each(p, h)
+				i++;
+
+			ret = max(ret, i);
+		}
+
+		mutex_unlock(&c->bucket_lock);
+		return ret;
+	}
+
+	unsigned btree_used(struct cache_set *c)
+	{
+		return div64_u64(c->gc_stats.key_bytes * 100,
+				 (c->gc_stats.nodes ?: 1) * btree_bytes(c));
+	}
+
+	unsigned average_key_size(struct cache_set *c)
+	{
+		return c->gc_stats.nkeys
+			? div64_u64(c->gc_stats.data, c->gc_stats.nkeys)
+			: 0;
+	}
+
+	struct cache_set *c = container_of(kobj, struct cache_set, kobj);
+
+	sysfs_print(synchronous,		CACHE_SYNC(&c->sb));
+	sysfs_print(journal_delay_ms,		c->journal_delay_ms);
+	sysfs_hprint(bucket_size,		bucket_bytes(c));
+	sysfs_hprint(block_size,		block_bytes(c));
+	sysfs_print(tree_depth,			c->root->level);
+	sysfs_print(root_usage_percent,		root_usage(c));
+
+	sysfs_hprint(btree_cache_size,		cache_size(c));
+	sysfs_print(btree_cache_max_chain,	cache_max_chain(c));
+	sysfs_print(cache_available_percent,	100 - c->gc_stats.in_use);
+
+	sysfs_print_time_stats(&c->btree_gc_time,	btree_gc, sec, ms);
+	sysfs_print_time_stats(&c->btree_split_time,	btree_split, sec, us);
+	sysfs_print_time_stats(&c->sort_time,		btree_sort, ms, us);
+	sysfs_print_time_stats(&c->btree_read_time,	btree_read, ms, us);
+	sysfs_print_time_stats(&c->try_harder_time,	try_harder, ms, us);
+
+	sysfs_print(btree_used_percent,	btree_used(c));
+	sysfs_print(btree_nodes,	c->gc_stats.nodes);
+	sysfs_hprint(dirty_data,	c->gc_stats.dirty);
+	sysfs_hprint(average_key_size,	average_key_size(c));
+
+	sysfs_print(cache_read_races,
+		    atomic_long_read(&c->cache_read_races));
+
+	sysfs_print(writeback_keys_done,
+		    atomic_long_read(&c->writeback_keys_done));
+	sysfs_print(writeback_keys_failed,
+		    atomic_long_read(&c->writeback_keys_failed));
+
+	/* See count_io_errors for why 88 */
+	sysfs_print(io_error_halflife,	c->error_decay * 88);
+	sysfs_print(io_error_limit,	c->error_limit >> IO_ERROR_SHIFT);
+
+	sysfs_hprint(congested,
+		     ((uint64_t) bch_get_congested(c)) << 9);
+	sysfs_print(congested_read_threshold_us,
+		    c->congested_read_threshold_us);
+	sysfs_print(congested_write_threshold_us,
+		    c->congested_write_threshold_us);
+
+	sysfs_print(active_journal_entries,	fifo_used(&c->journal.pin));
+	sysfs_printf(verify,			"%i", c->verify);
+	sysfs_printf(key_merging_disabled,	"%i", c->key_merging_disabled);
+	sysfs_printf(gc_always_rewrite,		"%i", c->gc_always_rewrite);
+	sysfs_printf(btree_shrinker_disabled,	"%i", c->shrinker_disabled);
+	sysfs_printf(copy_gc_enabled,		"%i", c->copy_gc_enabled);
+
+	if (attr == &sysfs_bset_tree_stats)
+		return bch_bset_print_stats(c, buf);
+
+	return 0;
+}
+SHOW_LOCKED(bch_cache_set)
+
+STORE(__bch_cache_set)
+{
+	struct cache_set *c = container_of(kobj, struct cache_set, kobj);
+
+	if (attr == &sysfs_unregister)
+		bch_cache_set_unregister(c);
+
+	if (attr == &sysfs_stop)
+		bch_cache_set_stop(c);
+
+	if (attr == &sysfs_synchronous) {
+		bool sync = strtoul_or_return(buf);
+
+		if (sync != CACHE_SYNC(&c->sb)) {
+			SET_CACHE_SYNC(&c->sb, sync);
+			bcache_write_super(c);
+		}
+	}
+
+	if (attr == &sysfs_flash_vol_create) {
+		int r;
+		uint64_t v;
+		strtoi_h_or_return(buf, v);
+
+		r = bch_flash_dev_create(c, v);
+		if (r)
+			return r;
+	}
+
+	if (attr == &sysfs_clear_stats) {
+		atomic_long_set(&c->writeback_keys_done,	0);
+		atomic_long_set(&c->writeback_keys_failed,	0);
+
+		memset(&c->gc_stats, 0, sizeof(struct gc_stat));
+		bch_cache_accounting_clear(&c->accounting);
+	}
+
+	if (attr == &sysfs_trigger_gc)
+		bch_queue_gc(c);
+
+	if (attr == &sysfs_prune_cache) {
+		struct shrink_control sc;
+		sc.gfp_mask = GFP_KERNEL;
+		sc.nr_to_scan = strtoul_or_return(buf);
+		c->shrink.shrink(&c->shrink, &sc);
+	}
+
+	sysfs_strtoul(congested_read_threshold_us,
+		      c->congested_read_threshold_us);
+	sysfs_strtoul(congested_write_threshold_us,
+		      c->congested_write_threshold_us);
+
+	if (attr == &sysfs_io_error_limit)
+		c->error_limit = strtoul_or_return(buf) << IO_ERROR_SHIFT;
+
+	/* See count_io_errors() for why 88 */
+	if (attr == &sysfs_io_error_halflife)
+		c->error_decay = strtoul_or_return(buf) / 88;
+
+	sysfs_strtoul(journal_delay_ms,		c->journal_delay_ms);
+	sysfs_strtoul(verify,			c->verify);
+	sysfs_strtoul(key_merging_disabled,	c->key_merging_disabled);
+	sysfs_strtoul(gc_always_rewrite,	c->gc_always_rewrite);
+	sysfs_strtoul(btree_shrinker_disabled,	c->shrinker_disabled);
+	sysfs_strtoul(copy_gc_enabled,		c->copy_gc_enabled);
+
+	return size;
+}
+STORE_LOCKED(bch_cache_set)
+
+SHOW(bch_cache_set_internal)
+{
+	struct cache_set *c = container_of(kobj, struct cache_set, internal);
+	return bch_cache_set_show(&c->kobj, attr, buf);
+}
+
+STORE(bch_cache_set_internal)
+{
+	struct cache_set *c = container_of(kobj, struct cache_set, internal);
+	return bch_cache_set_store(&c->kobj, attr, buf, size);
+}
+
+static void bch_cache_set_internal_release(struct kobject *k)
+{
+}
+
+static struct attribute *bch_cache_set_files[] = {
+	&sysfs_unregister,
+	&sysfs_stop,
+	&sysfs_synchronous,
+	&sysfs_journal_delay_ms,
+	&sysfs_flash_vol_create,
+
+	&sysfs_bucket_size,
+	&sysfs_block_size,
+	&sysfs_tree_depth,
+	&sysfs_root_usage_percent,
+	&sysfs_btree_cache_size,
+	&sysfs_cache_available_percent,
+
+	&sysfs_average_key_size,
+	&sysfs_dirty_data,
+
+	&sysfs_io_error_limit,
+	&sysfs_io_error_halflife,
+	&sysfs_congested,
+	&sysfs_congested_read_threshold_us,
+	&sysfs_congested_write_threshold_us,
+	&sysfs_clear_stats,
+	NULL
+};
+KTYPE(bch_cache_set);
+
+static struct attribute *bch_cache_set_internal_files[] = {
+	&sysfs_active_journal_entries,
+
+	sysfs_time_stats_attribute_list(btree_gc, sec, ms)
+	sysfs_time_stats_attribute_list(btree_split, sec, us)
+	sysfs_time_stats_attribute_list(btree_sort, ms, us)
+	sysfs_time_stats_attribute_list(btree_read, ms, us)
+	sysfs_time_stats_attribute_list(try_harder, ms, us)
+
+	&sysfs_btree_nodes,
+	&sysfs_btree_used_percent,
+	&sysfs_btree_cache_max_chain,
+
+	&sysfs_bset_tree_stats,
+	&sysfs_cache_read_races,
+	&sysfs_writeback_keys_done,
+	&sysfs_writeback_keys_failed,
+
+	&sysfs_trigger_gc,
+	&sysfs_prune_cache,
+#ifdef CONFIG_BCACHE_DEBUG
+	&sysfs_verify,
+	&sysfs_key_merging_disabled,
+#endif
+	&sysfs_gc_always_rewrite,
+	&sysfs_btree_shrinker_disabled,
+	&sysfs_copy_gc_enabled,
+	NULL
+};
+KTYPE(bch_cache_set_internal);
+
+SHOW(__bch_cache)
+{
+	struct cache *ca = container_of(kobj, struct cache, kobj);
+
+	sysfs_hprint(bucket_size,	bucket_bytes(ca));
+	sysfs_hprint(block_size,	block_bytes(ca));
+	sysfs_print(nbuckets,		ca->sb.nbuckets);
+	sysfs_print(discard,		ca->discard);
+	sysfs_hprint(written, atomic_long_read(&ca->sectors_written) << 9);
+	sysfs_hprint(btree_written,
+		     atomic_long_read(&ca->btree_sectors_written) << 9);
+	sysfs_hprint(metadata_written,
+		     (atomic_long_read(&ca->meta_sectors_written) +
+		      atomic_long_read(&ca->btree_sectors_written)) << 9);
+
+	sysfs_print(io_errors,
+		    atomic_read(&ca->io_errors) >> IO_ERROR_SHIFT);
+
+	sysfs_print(freelist_percent, ca->free.size * 100 /
+		    ((size_t) ca->sb.nbuckets));
+
+	if (attr == &sysfs_cache_replacement_policy)
+		return bch_snprint_string_list(buf, PAGE_SIZE,
+					       cache_replacement_policies,
+					       CACHE_REPLACEMENT(&ca->sb));
+
+	if (attr == &sysfs_priority_stats) {
+		int cmp(const void *l, const void *r)
+		{	return *((uint16_t *) r) - *((uint16_t *) l); }
+
+		/* Number of quantiles we compute */
+		const unsigned nq = 31;
+
+		size_t n = ca->sb.nbuckets, i, unused, btree;
+		uint64_t sum = 0;
+		uint16_t q[nq], *p, *cached;
+		ssize_t ret;
+
+		cached = p = vmalloc(ca->sb.nbuckets * sizeof(uint16_t));
+		if (!p)
+			return -ENOMEM;
+
+		mutex_lock(&ca->set->bucket_lock);
+		for (i = ca->sb.first_bucket; i < n; i++)
+			p[i] = ca->buckets[i].prio;
+		mutex_unlock(&ca->set->bucket_lock);
+
+		sort(p, n, sizeof(uint16_t), cmp, NULL);
+
+		while (n &&
+		       !cached[n - 1])
+			--n;
+
+		unused = ca->sb.nbuckets - n;
+
+		while (cached < p + n &&
+		       *cached == BTREE_PRIO)
+			cached++;
+
+		btree = cached - p;
+		n -= btree;
+
+		for (i = 0; i < n; i++)
+			sum += INITIAL_PRIO - cached[i];
+
+		if (n)
+			do_div(sum, n);
+
+		for (i = 0; i < nq; i++)
+			q[i] = INITIAL_PRIO - cached[n * (i + 1) / (nq + 1)];
+
+		vfree(p);
+
+		ret = snprintf(buf, PAGE_SIZE,
+			       "Unused:		%zu%%\n"
+			       "Metadata:	%zu%%\n"
+			       "Average:	%llu\n"
+			       "Sectors per Q:	%zu\n"
+			       "Quantiles:	[",
+			       unused * 100 / (size_t) ca->sb.nbuckets,
+			       btree * 100 / (size_t) ca->sb.nbuckets, sum,
+			       n * ca->sb.bucket_size / (nq + 1));
+
+		for (i = 0; i < nq && ret < (ssize_t) PAGE_SIZE; i++)
+			ret += snprintf(buf + ret, PAGE_SIZE - ret,
+					i < nq - 1 ? "%u " : "%u]\n", q[i]);
+
+		buf[PAGE_SIZE - 1] = '\0';
+		return ret;
+	}
+
+	return 0;
+}
+SHOW_LOCKED(bch_cache)
+
+STORE(__bch_cache)
+{
+	struct cache *ca = container_of(kobj, struct cache, kobj);
+
+	if (attr == &sysfs_discard) {
+		bool v = strtoul_or_return(buf);
+
+		if (blk_queue_discard(bdev_get_queue(ca->bdev)))
+			ca->discard = v;
+
+		if (v != CACHE_DISCARD(&ca->sb)) {
+			SET_CACHE_DISCARD(&ca->sb, v);
+			bcache_write_super(ca->set);
+		}
+	}
+
+	if (attr == &sysfs_cache_replacement_policy) {
+		ssize_t v = bch_read_string_list(buf, cache_replacement_policies);
+
+		if (v < 0)
+			return v;
+
+		if ((unsigned) v != CACHE_REPLACEMENT(&ca->sb)) {
+			mutex_lock(&ca->set->bucket_lock);
+			SET_CACHE_REPLACEMENT(&ca->sb, v);
+			mutex_unlock(&ca->set->bucket_lock);
+
+			bcache_write_super(ca->set);
+		}
+	}
+
+	if (attr == &sysfs_freelist_percent) {
+		DECLARE_FIFO(long, free);
+		long i;
+		size_t p = strtoul_or_return(buf);
+
+		p = clamp_t(size_t,
+			    ((size_t) ca->sb.nbuckets * p) / 100,
+			    roundup_pow_of_two(ca->sb.nbuckets) >> 9,
+			    ca->sb.nbuckets / 2);
+
+		if (!init_fifo_exact(&free, p, GFP_KERNEL))
+			return -ENOMEM;
+
+		mutex_lock(&ca->set->bucket_lock);
+
+		fifo_move(&free, &ca->free);
+		fifo_swap(&free, &ca->free);
+
+		mutex_unlock(&ca->set->bucket_lock);
+
+		while (fifo_pop(&free, i))
+			atomic_dec(&ca->buckets[i].pin);
+
+		free_fifo(&free);
+	}
+
+	if (attr == &sysfs_clear_stats) {
+		atomic_long_set(&ca->sectors_written, 0);
+		atomic_long_set(&ca->btree_sectors_written, 0);
+		atomic_long_set(&ca->meta_sectors_written, 0);
+		atomic_set(&ca->io_count, 0);
+		atomic_set(&ca->io_errors, 0);
+	}
+
+	return size;
+}
+STORE_LOCKED(bch_cache)
+
+static struct attribute *bch_cache_files[] = {
+	&sysfs_bucket_size,
+	&sysfs_block_size,
+	&sysfs_nbuckets,
+	&sysfs_priority_stats,
+	&sysfs_discard,
+	&sysfs_written,
+	&sysfs_btree_written,
+	&sysfs_metadata_written,
+	&sysfs_io_errors,
+	&sysfs_clear_stats,
+	&sysfs_freelist_percent,
+	&sysfs_cache_replacement_policy,
+	NULL
+};
+KTYPE(bch_cache);

drivers/md/bcache/sysfs.h

···
+#ifndef _BCACHE_SYSFS_H_
+#define _BCACHE_SYSFS_H_
+
+#define KTYPE(type)							\
+struct kobj_type type ## _ktype = {					\
+	.release	= type ## _release,				\
+	.sysfs_ops	= &((const struct sysfs_ops) {			\
+		.show	= type ## _show,				\
+		.store	= type ## _store				\
+	}),								\
+	.default_attrs	= type ## _files				\
+}
+
+#define SHOW(fn)							\
+static ssize_t fn ## _show(struct kobject *kobj, struct attribute *attr,\
+			   char *buf)					\
+
+#define STORE(fn)							\
+static ssize_t fn ## _store(struct kobject *kobj, struct attribute *attr,\
+			    const char *buf, size_t size)		\
+
+#define SHOW_LOCKED(fn)							\
+SHOW(fn)								\
+{									\
+	ssize_t ret;							\
+	mutex_lock(&bch_register_lock);					\
+	ret = __ ## fn ## _show(kobj, attr, buf);			\
+	mutex_unlock(&bch_register_lock);				\
+	return ret;							\
+}
+
+#define STORE_LOCKED(fn)						\
+STORE(fn)								\
+{									\
+	ssize_t ret;							\
+	mutex_lock(&bch_register_lock);					\
+	ret = __ ## fn ## _store(kobj, attr, buf, size);		\
+	mutex_unlock(&bch_register_lock);				\
+	return ret;							\
+}
+
+#define __sysfs_attribute(_name, _mode)					\
+	static struct attribute sysfs_##_name =				\
+		{ .name = #_name, .mode = _mode }
+
+#define write_attribute(n)	__sysfs_attribute(n, S_IWUSR)
+#define read_attribute(n)	__sysfs_attribute(n, S_IRUGO)
+#define rw_attribute(n)		__sysfs_attribute(n, S_IRUGO|S_IWUSR)
+
+#define sysfs_printf(file, fmt, ...)					\
+do {									\
+	if (attr == &sysfs_ ## file)					\
+		return snprintf(buf, PAGE_SIZE, fmt "\n", __VA_ARGS__);	\
+} while (0)
+
+#define sysfs_print(file, var)						\
+do {									\
+	if (attr == &sysfs_ ## file)					\
+		return snprint(buf, PAGE_SIZE, var);			\
+} while (0)
+
+#define sysfs_hprint(file, val)						\
+do {									\
+	if (attr == &sysfs_ ## file) {					\
+		ssize_t ret = bch_hprint(buf, val);			\
+		strcat(buf, "\n");					\
+		return ret + 1;						\
+	}								\
+} while (0)
+
+#define var_printf(_var, fmt)	sysfs_printf(_var, fmt, var(_var))
+#define var_print(_var)		sysfs_print(_var, var(_var))
+#define var_hprint(_var)	sysfs_hprint(_var, var(_var))
+
+#define sysfs_strtoul(file, var)					\
+do {									\
+	if (attr == &sysfs_ ## file)					\
+		return strtoul_safe(buf, var) ?: (ssize_t) size;	\
+} while (0)
+
+#define sysfs_strtoul_clamp(file, var, min, max)			\
+do {									\
+	if (attr == &sysfs_ ## file)					\
+		return strtoul_safe_clamp(buf, var, min, max)		\
+			?: (ssize_t) size;				\
+} while (0)
+
+#define strtoul_or_return(cp)						\
+({									\
+	unsigned long _v;						\
+	int _r = kstrtoul(cp, 10, &_v);					\
+	if (_r)								\
+		return _r;						\
+	_v;								\
+})
+
+#define strtoi_h_or_return(cp, v)					\
+do {									\
+	int _r = strtoi_h(cp, &v);					\
+	if (_r)								\
+		return _r;						\
+} while (0)
+
+#define sysfs_hatoi(file, var)						\
+do {									\
+	if (attr == &sysfs_ ## file)					\
+		return strtoi_h(buf, &var) ?: (ssize_t) size;		\
+} while (0)
+
+#endif  /* _BCACHE_SYSFS_H_ */

drivers/md/bcache/trace.c

···
+#include "bcache.h"
+#include "btree.h"
+#include "request.h"
+
+#include <linux/module.h>
+
+#define CREATE_TRACE_POINTS
+#include <trace/events/bcache.h>
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_request_start);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_request_end);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_passthrough);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_cache_hit);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_cache_miss);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_read_retry);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_writethrough);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_writeback);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_write_skip);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_read);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_write);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_write_dirty);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_read_dirty);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_journal_write);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_cache_insert);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_gc_start);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_gc_end);

drivers/md/bcache/util.c

···
+/*
+ * random utiility code, for bcache but in theory not specific to bcache
+ *
+ * Copyright 2010, 2011 Kent Overstreet <[e-mail]>
+ * Copyright 2012 Google, Inc.
+ */
+
+#include <linux/bio.h>
+#include <linux/blkdev.h>
+#include <linux/ctype.h>
+#include <linux/debugfs.h>
+#include <linux/module.h>
+#include <linux/seq_file.h>
+#include <linux/types.h>
+
+#include "util.h"
+
+#define simple_strtoint(c, end, base)	simple_strtol(c, end, base)
+#define simple_strtouint(c, end, base)	simple_strtoul(c, end, base)
+
+#define STRTO_H(name, type)					\
+int bch_ ## name ## _h(const char *cp, type *res)		\
+{								\
+	int u = 0;						\
+	char *e;						\
+	type i = simple_ ## name(cp, &e, 10);			\
+								\
+	switch (tolower(*e)) {					\
+	default:						\
+		return -EINVAL;					\
+	case 'y':						\
+	case 'z':						\
+		u++;						\
+	case 'e':						\
+		u++;						\
+	case 'p':						\
+		u++;						\
+	case 't':						\
+		u++;						\
+	case 'g':						\
+		u++;						\
+	case 'm':						\
+		u++;						\
+	case 'k':						\
+		u++;						\
+		if (e++ == cp)					\
+			return -EINVAL;				\
+	case '\n':						\
+	case '\0':						\
+		if (*e == '\n')					\
+			e++;					\
+	}							\
+								\
+	if (*e)							\
+		return -EINVAL;					\
+								\
+	while (u--) {						\
+		if ((type) ~0 > 0 &&				\
+		    (type) ~0 / 1024 <= i)			\
+			return -EINVAL;				\
+		if ((i > 0 && ANYSINT_MAX(type) / 1024 < i) ||	\
+		    (i < 0 && -ANYSINT_MAX(type) / 1024 > i))	\
+			return -EINVAL;				\
+		i *= 1024;					\
+	}							\
+								\
+	*res = i;						\
+	return 0;						\
+}								\
+
+STRTO_H(strtoint, int)
+STRTO_H(strtouint, unsigned int)
+STRTO_H(strtoll, long long)
+STRTO_H(strtoull, unsigned long long)
+
+ssize_t bch_hprint(char *buf, int64_t v)
+{
+	static const char units[] = "?kMGTPEZY";
+	char dec[4] = "";
+	int u, t = 0;
+
+	for (u = 0; v >= 1024 || v <= -1024; u++) {
+		t = v & ~(~0 << 10);
+		v >>= 10;
+	}
+
+	if (!u)
+		return sprintf(buf, "%llu", v);
+
+	if (v < 100 && v > -100)
+		snprintf(dec, sizeof(dec), ".%i", t / 100);
+
+	return sprintf(buf, "%lli%s%c", v, dec, units[u]);
+}
+
+ssize_t bch_snprint_string_list(char *buf, size_t size, const char * const list[],
+			    size_t selected)
+{
+	char *out = buf;
+	size_t i;
+
+	for (i = 0; list[i]; i++)
+		out += snprintf(out, buf + size - out,
+				i == selected ? "[%s] " : "%s ", list[i]);
+
+	out[-1] = '\n';
+	return out - buf;
+}
+
+ssize_t bch_read_string_list(const char *buf, const char * const list[])
+{
+	size_t i;
+	char *s, *d = kstrndup(buf, PAGE_SIZE - 1, GFP_KERNEL);
+	if (!d)
+		return -ENOMEM;
+
+	s = strim(d);
+
+	for (i = 0; list[i]; i++)
+		if (!strcmp(list[i], s))
+			break;
+
+	kfree(d);
+
+	if (!list[i])
+		return -EINVAL;
+
+	return i;
+}
+
+bool bch_is_zero(const char *p, size_t n)
+{
+	size_t i;
+
+	for (i = 0; i < n; i++)
+		if (p[i])
+			return false;
+	return true;
+}
+
+int bch_parse_uuid(const char *s, char *uuid)
+{
+	size_t i, j, x;
+	memset(uuid, 0, 16);
+
+	for (i = 0, j = 0;
+	     i < strspn(s, "-0123456789:ABCDEFabcdef") && j < 32;
+	     i++) {
+		x = s[i] | 32;
+
+		switch (x) {
+		case '0'...'9':
+			x -= '0';
+			break;
+		case 'a'...'f':
+			x -= 'a' - 10;
+			break;
+		default:
+			continue;
+		}
+
+		if (!(j & 1))
+			x <<= 4;
+		uuid[j++ >> 1] |= x;
+	}
+	return i;
+}
+
+void bch_time_stats_update(struct time_stats *stats, uint64_t start_time)
+{
+	uint64_t now		= local_clock();
+	uint64_t duration	= time_after64(now, start_time)
+		? now - start_time : 0;
+	uint64_t last		= time_after64(now, stats->last)
+		? now - stats->last : 0;
+
+	stats->max_duration = max(stats->max_duration, duration);
+
+	if (stats->last) {
+		ewma_add(stats->average_duration, duration, 8, 8);
+
+		if (stats->average_frequency)
+			ewma_add(stats->average_frequency, last, 8, 8);
+		else
+			stats->average_frequency  = last << 8;
+	} else {
+		stats->average_duration  = duration << 8;
+	}
+
+	stats->last = now ?: 1;
+}
+
+unsigned bch_next_delay(struct ratelimit *d, uint64_t done)
+{
+	uint64_t now = local_clock();
+
+	d->next += div_u64(done, d->rate);
+
+	return time_after64(d->next, now)
+		? div_u64(d->next - now, NSEC_PER_SEC / HZ)
+		: 0;
+}
+
+void bch_bio_map(struct bio *bio, void *base)
+{
+	size_t size = bio->bi_size;
+	struct bio_vec *bv = bio->bi_io_vec;
+
+	BUG_ON(!bio->bi_size);
+	BUG_ON(bio->bi_vcnt);
+
+	bv->bv_offset = base ? ((unsigned long) base) % PAGE_SIZE : 0;
+	goto start;
+
+	for (; size; bio->bi_vcnt++, bv++) {
+		bv->bv_offset	= 0;
+start:		bv->bv_len	= min_t(size_t, PAGE_SIZE - bv->bv_offset,
+					size);
+		if (base) {
+			bv->bv_page = is_vmalloc_addr(base)
+				? vmalloc_to_page(base)
+				: virt_to_page(base);
+
+			base += bv->bv_len;
+		}
+
+		size -= bv->bv_len;
+	}
+}
+
+int bch_bio_alloc_pages(struct bio *bio, gfp_t gfp)
+{
+	int i;
+	struct bio_vec *bv;
+
+	bio_for_each_segment(bv, bio, i) {
+		bv->bv_page = alloc_page(gfp);
+		if (!bv->bv_page) {
+			while (bv-- != bio->bi_io_vec + bio->bi_idx)
+				__free_page(bv->bv_page);
+			return -ENOMEM;
+		}
+	}
+
+	return 0;
+}
+
+/*
+ * Portions Copyright (c) 1996-2001, PostgreSQL Global Development Group (Any
+ * use permitted, subject to terms of PostgreSQL license; see.)
+
+ * If we have a 64-bit integer type, then a 64-bit CRC looks just like the
+ * usual sort of implementation. (See Ross Williams' excellent introduction
+ * A PAINLESS GUIDE TO CRC ERROR DETECTION ALGORITHMS, available from
+ * ftp://ftp.rocksoft.com/papers/crc_v3.txt or several other net sites.)
+ * If we have no working 64-bit type, then fake it with two 32-bit registers.
+ *
+ * The present implementation is a normal (not "reflected", in Williams'
+ * terms) 64-bit CRC, using initial all-ones register contents and a final
+ * bit inversion. The chosen polynomial is borrowed from the DLT1 spec
+ * (ECMA-182, available from http://www.ecma.ch/ecma1/STAND/ECMA-182.HTM):
+ *
+ * x^64 + x^62 + x^57 + x^55 + x^54 + x^53 + x^52 + x^47 + x^46 + x^45 +
+ * x^40 + x^39 + x^38 + x^37 + x^35 + x^33 + x^32 + x^31 + x^29 + x^27 +
+ * x^24 + x^23 + x^22 + x^21 + x^19 + x^17 + x^13 + x^12 + x^10 + x^9 +
+ * x^7 + x^4 + x + 1
+*/
+
+static const uint64_t crc_table[256] = {
+	0x0000000000000000ULL, 0x42F0E1EBA9EA3693ULL, 0x85E1C3D753D46D26ULL,
+	0xC711223CFA3E5BB5ULL, 0x493366450E42ECDFULL, 0x0BC387AEA7A8DA4CULL,
+	0xCCD2A5925D9681F9ULL, 0x8E224479F47CB76AULL, 0x9266CC8A1C85D9BEULL,
+	0xD0962D61B56FEF2DULL, 0x17870F5D4F51B498ULL, 0x5577EEB6E6BB820BULL,
+	0xDB55AACF12C73561ULL, 0x99A54B24BB2D03F2ULL, 0x5EB4691841135847ULL,
+	0x1C4488F3E8F96ED4ULL, 0x663D78FF90E185EFULL, 0x24CD9914390BB37CULL,
+	0xE3DCBB28C335E8C9ULL, 0xA12C5AC36ADFDE5AULL, 0x2F0E1EBA9EA36930ULL,
+	0x6DFEFF5137495FA3ULL, 0xAAEFDD6DCD770416ULL, 0xE81F3C86649D3285ULL,
+	0xF45BB4758C645C51ULL, 0xB6AB559E258E6AC2ULL, 0x71BA77A2DFB03177ULL,
+	0x334A9649765A07E4ULL, 0xBD68D2308226B08EULL, 0xFF9833DB2BCC861DULL,
+	0x388911E7D1F2DDA8ULL, 0x7A79F00C7818EB3BULL, 0xCC7AF1FF21C30BDEULL,
+	0x8E8A101488293D4DULL, 0x499B3228721766F8ULL, 0x0B6BD3C3DBFD506BULL,
+	0x854997BA2F81E701ULL, 0xC7B97651866BD192ULL, 0x00A8546D7C558A27ULL,
+	0x4258B586D5BFBCB4ULL, 0x5E1C3D753D46D260ULL, 0x1CECDC9E94ACE4F3ULL,
+	0xDBFDFEA26E92BF46ULL, 0x990D1F49C77889D5ULL, 0x172F5B3033043EBFULL,
+	0x55DFBADB9AEE082CULL, 0x92CE98E760D05399ULL, 0xD03E790CC93A650AULL,
+	0xAA478900B1228E31ULL, 0xE8B768EB18C8B8A2ULL, 0x2FA64AD7E2F6E317ULL,
+	0x6D56AB3C4B1CD584ULL, 0xE374EF45BF6062EEULL, 0xA1840EAE168A547DULL,
+	0x66952C92ECB40FC8ULL, 0x2465CD79455E395BULL, 0x3821458AADA7578FULL,
+	0x7AD1A461044D611CULL, 0xBDC0865DFE733AA9ULL, 0xFF3067B657990C3AULL,
+	0x711223CFA3E5BB50ULL, 0x33E2C2240A0F8DC3ULL, 0xF4F3E018F031D676ULL,
+	0xB60301F359DBE0E5ULL, 0xDA050215EA6C212FULL, 0x98F5E3FE438617BCULL,
+	0x5FE4C1C2B9B84C09ULL, 0x1D14202910527A9AULL, 0x93366450E42ECDF0ULL,
+	0xD1C685BB4DC4FB63ULL, 0x16D7A787B7FAA0D6ULL, 0x5427466C1E109645ULL,
+	0x4863CE9FF6E9F891ULL, 0x0A932F745F03CE02ULL, 0xCD820D48A53D95B7ULL,
+	0x8F72ECA30CD7A324ULL, 0x0150A8DAF8AB144EULL, 0x43A04931514122DDULL,
+	0x84B16B0DAB7F7968ULL, 0xC6418AE602954FFBULL, 0xBC387AEA7A8DA4C0ULL,
+	0xFEC89B01D3679253ULL, 0x39D9B93D2959C9E6ULL, 0x7B2958D680B3FF75ULL,
+	0xF50B1CAF74CF481FULL, 0xB7FBFD44DD257E8CULL, 0x70EADF78271B2539ULL,
+	0x321A3E938EF113AAULL, 0x2E5EB66066087D7EULL, 0x6CAE578BCFE24BEDULL,
+	0xABBF75B735DC1058ULL, 0xE94F945C9C3626CBULL, 0x676DD025684A91A1ULL,
+	0x259D31CEC1A0A732ULL, 0xE28C13F23B9EFC87ULL, 0xA07CF2199274CA14ULL,
+	0x167FF3EACBAF2AF1ULL, 0x548F120162451C62ULL, 0x939E303D987B47D7ULL,
+	0xD16ED1D631917144ULL, 0x5F4C95AFC5EDC62EULL, 0x1DBC74446C07F0BDULL,
+	0xDAAD56789639AB08ULL, 0x985DB7933FD39D9BULL, 0x84193F60D72AF34FULL,
+	0xC6E9DE8B7EC0C5DCULL, 0x01F8FCB784FE9E69ULL, 0x43081D5C2D14A8FAULL,
+	0xCD2A5925D9681F90ULL, 0x8FDAB8CE70822903ULL, 0x48CB9AF28ABC72B6ULL,
+	0x0A3B7B1923564425ULL, 0x70428B155B4EAF1EULL, 0x32B26AFEF2A4998DULL,
+	0xF5A348C2089AC238ULL, 0xB753A929A170F4ABULL, 0x3971ED50550C43C1ULL,
+	0x7B810CBBFCE67552ULL, 0xBC902E8706D82EE7ULL, 0xFE60CF6CAF321874ULL,
+	0xE224479F47CB76A0ULL, 0xA0D4A674EE214033ULL, 0x67C58448141F1B86ULL,
+	0x253565A3BDF52D15ULL, 0xAB1721DA49899A7FULL, 0xE9E7C031E063ACECULL,
+	0x2EF6E20D1A5DF759ULL, 0x6C0603E6B3B7C1CAULL, 0xF6FAE5C07D3274CDULL,
+	0xB40A042BD4D8425EULL, 0x731B26172EE619EBULL, 0x31EBC7FC870C2F78ULL,
+	0xBFC9838573709812ULL, 0xFD39626EDA9AAE81ULL, 0x3A28405220A4F534ULL,
+	0x78D8A1B9894EC3A7ULL, 0x649C294A61B7AD73ULL, 0x266CC8A1C85D9BE0ULL,
+	0xE17DEA9D3263C055ULL, 0xA38D0B769B89F6C6ULL, 0x2DAF4F0F6FF541ACULL,
+	0x6F5FAEE4C61F773FULL, 0xA84E8CD83C212C8AULL, 0xEABE6D3395CB1A19ULL,
+	0x90C79D3FEDD3F122ULL, 0xD2377CD44439C7B1ULL, 0x15265EE8BE079C04ULL,
+	0x57D6BF0317EDAA97ULL, 0xD9F4FB7AE3911DFDULL, 0x9B041A914A7B2B6EULL,
+	0x5C1538ADB04570DBULL, 0x1EE5D94619AF4648ULL, 0x02A151B5F156289CULL,
+	0x4051B05E58BC1E0FULL, 0x87409262A28245BAULL, 0xC5B073890B687329ULL,
+	0x4B9237F0FF14C443ULL, 0x0962D61B56FEF2D0ULL, 0xCE73F427ACC0A965ULL,
+	0x8C8315CC052A9FF6ULL, 0x3A80143F5CF17F13ULL, 0x7870F5D4F51B4980ULL,
+	0xBF61D7E80F251235ULL, 0xFD913603A6CF24A6ULL, 0x73B3727A52B393CCULL,
+	0x31439391FB59A55FULL, 0xF652B1AD0167FEEAULL, 0xB4A25046A88DC879ULL,
+	0xA8E6D8B54074A6ADULL, 0xEA16395EE99E903EULL, 0x2D071B6213A0CB8BULL,
+	0x6FF7FA89BA4AFD18ULL, 0xE1D5BEF04E364A72ULL, 0xA3255F1BE7DC7CE1ULL,
+	0x64347D271DE22754ULL, 0x26C49CCCB40811C7ULL, 0x5CBD6CC0CC10FAFCULL,
+	0x1E4D8D2B65FACC6FULL, 0xD95CAF179FC497DAULL, 0x9BAC4EFC362EA149ULL,
+	0x158E0A85C2521623ULL, 0x577EEB6E6BB820B0ULL, 0x906FC95291867B05ULL,
+	0xD29F28B9386C4D96ULL, 0xCEDBA04AD0952342ULL, 0x8C2B41A1797F15D1ULL,
+	0x4B3A639D83414E64ULL, 0x09CA82762AAB78F7ULL, 0x87E8C60FDED7CF9DULL,
+	0xC51827E4773DF90EULL, 0x020905D88D03A2BBULL, 0x40F9E43324E99428ULL,
+	0x2CFFE7D5975E55E2ULL, 0x6E0F063E3EB46371ULL, 0xA91E2402C48A38C4ULL,
+	0xEBEEC5E96D600E57ULL, 0x65CC8190991CB93DULL, 0x273C607B30F68FAEULL,
+	0xE02D4247CAC8D41BULL, 0xA2DDA3AC6322E288ULL, 0xBE992B5F8BDB8C5CULL,
+	0xFC69CAB42231BACFULL, 0x3B78E888D80FE17AULL, 0x7988096371E5D7E9ULL,
+	0xF7AA4D1A85996083ULL, 0xB55AACF12C735610ULL, 0x724B8ECDD64D0DA5ULL,
+	0x30BB6F267FA73B36ULL, 0x4AC29F2A07BFD00DULL, 0x08327EC1AE55E69EULL,
+	0xCF235CFD546BBD2BULL, 0x8DD3BD16FD818BB8ULL, 0x03F1F96F09FD3CD2ULL,
+	0x41011884A0170A41ULL, 0x86103AB85A2951F4ULL, 0xC4E0DB53F3C36767ULL,
+	0xD8A453A01B3A09B3ULL, 0x9A54B24BB2D03F20ULL, 0x5D45907748EE6495ULL,
+	0x1FB5719CE1045206ULL, 0x919735E51578E56CULL, 0xD367D40EBC92D3FFULL,
+	0x1476F63246AC884AULL, 0x568617D9EF46BED9ULL, 0xE085162AB69D5E3CULL,
+	0xA275F7C11F7768AFULL, 0x6564D5FDE549331AULL, 0x279434164CA30589ULL,
+	0xA9B6706FB8DFB2E3ULL, 0xEB46918411358470ULL, 0x2C57B3B8EB0BDFC5ULL,
+	0x6EA7525342E1E956ULL, 0x72E3DAA0AA188782ULL, 0x30133B4B03F2B111ULL,
+	0xF7021977F9CCEAA4ULL, 0xB5F2F89C5026DC37ULL, 0x3BD0BCE5A45A6B5DULL,
+	0x79205D0E0DB05DCEULL, 0xBE317F32F78E067BULL, 0xFCC19ED95E6430E8ULL,
+	0x86B86ED5267CDBD3ULL, 0xC4488F3E8F96ED40ULL, 0x0359AD0275A8B6F5ULL,
+	0x41A94CE9DC428066ULL, 0xCF8B0890283E370CULL, 0x8D7BE97B81D4019FULL,
+	0x4A6ACB477BEA5A2AULL, 0x089A2AACD2006CB9ULL, 0x14DEA25F3AF9026DULL,
+	0x562E43B4931334FEULL, 0x913F6188692D6F4BULL, 0xD3CF8063C0C759D8ULL,
+	0x5DEDC41A34BBEEB2ULL, 0x1F1D25F19D51D821ULL, 0xD80C07CD676F8394ULL,
+	0x9AFCE626CE85B507ULL,
+};
+
+uint64_t bch_crc64_update(uint64_t crc, const void *_data, size_t len)
+{
+	const unsigned char *data = _data;
+
+	while (len--) {
+		int i = ((int) (crc >> 56) ^ *data++) & 0xFF;
+		crc = crc_table[i] ^ (crc << 8);
+	}
+
+	return crc;
+}
+
+uint64_t bch_crc64(const void *data, size_t len)
+{
+	uint64_t crc = 0xffffffffffffffffULL;
+
+	crc = bch_crc64_update(crc, data, len);
+
+	return crc ^ 0xffffffffffffffffULL;
+}

drivers/md/bcache/util.h

···
+
+#ifndef _BCACHE_UTIL_H
+#define _BCACHE_UTIL_H
+
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/llist.h>
+#include <linux/ratelimit.h>
+#include <linux/vmalloc.h>
+#include <linux/workqueue.h>
+
+#include "closure.h"
+
+#define PAGE_SECTORS		(PAGE_SIZE / 512)
+
+struct closure;
+
+#include <trace/events/bcache.h>
+
+#ifdef CONFIG_BCACHE_EDEBUG
+
+#define atomic_dec_bug(v)	BUG_ON(atomic_dec_return(v) < 0)
+#define atomic_inc_bug(v, i)	BUG_ON(atomic_inc_return(v) <= i)
+
+#else /* EDEBUG */
+
+#define atomic_dec_bug(v)	atomic_dec(v)
+#define atomic_inc_bug(v, i)	atomic_inc(v)
+
+#endif
+
+#define BITMASK(name, type, field, offset, size)		\
+static inline uint64_t name(const type *k)			\
+{ return (k->field >> offset) & ~(((uint64_t) ~0) << size); }	\
+								\
+static inline void SET_##name(type *k, uint64_t v)		\
+{								\
+	k->field &= ~(~((uint64_t) ~0 << size) << offset);	\
+	k->field |= v << offset;				\
+}
+
+#define DECLARE_HEAP(type, name)					\
+	struct {							\
+		size_t size, used;					\
+		type *data;						\
+	} name
+
+#define init_heap(heap, _size, gfp)					\
+({									\
+	size_t _bytes;							\
+	(heap)->used = 0;						\
+	(heap)->size = (_size);						\
+	_bytes = (heap)->size * sizeof(*(heap)->data);			\
+	(heap)->data = NULL;						\
+	if (_bytes < KMALLOC_MAX_SIZE)					\
+		(heap)->data = kmalloc(_bytes, (gfp));			\
+	if ((!(heap)->data) && ((gfp) & GFP_KERNEL))			\
+		(heap)->data = vmalloc(_bytes);				\
+	(heap)->data;							\
+})
+
+#define free_heap(heap)							\
+do {									\
+	if (is_vmalloc_addr((heap)->data))				\
+		vfree((heap)->data);					\
+	else								\
+		kfree((heap)->data);					\
+	(heap)->data = NULL;						\
+} while (0)
+
+#define heap_swap(h, i, j)	swap((h)->data[i], (h)->data[j])
+
+#define heap_sift(h, i, cmp)						\
+do {									\
+	size_t _r, _j = i;						\
+									\
+	for (; _j * 2 + 1 < (h)->used; _j = _r) {			\
+		_r = _j * 2 + 1;					\
+		if (_r + 1 < (h)->used &&				\
+		    cmp((h)->data[_r], (h)->data[_r + 1]))		\
+			_r++;						\
+									\
+		if (cmp((h)->data[_r], (h)->data[_j]))			\
+			break;						\
+		heap_swap(h, _r, _j);					\
+	}								\
+} while (0)
+
+#define heap_sift_down(h, i, cmp)					\
+do {									\
+	while (i) {							\
+		size_t p = (i - 1) / 2;					\
+		if (cmp((h)->data[i], (h)->data[p]))			\
+			break;						\
+		heap_swap(h, i, p);					\
+		i = p;							\
+	}								\
+} while (0)
+
+#define heap_add(h, d, cmp)						\
+({									\
+	bool _r = !heap_full(h);					\
+	if (_r) {							\
+		size_t _i = (h)->used++;				\
+		(h)->data[_i] = d;					\
+									\
+		heap_sift_down(h, _i, cmp);				\
+		heap_sift(h, _i, cmp);					\
+	}								\
+	_r;								\
+})
+
+#define heap_pop(h, d, cmp)						\
+({									\
+	bool _r = (h)->used;						\
+	if (_r) {							\
+		(d) = (h)->data[0];					\
+		(h)->used--;						\
+		heap_swap(h, 0, (h)->used);				\
+		heap_sift(h, 0, cmp);					\
+	}								\
+	_r;								\
+})
+
+#define heap_peek(h)	((h)->size ? (h)->data[0] : NULL)
+
+#define heap_full(h)	((h)->used == (h)->size)
+
+#define DECLARE_FIFO(type, name)					\
+	struct {							\
+		size_t front, back, size, mask;				\
+		type *data;						\
+	} name
+
+#define fifo_for_each(c, fifo, iter)					\
+	for (iter = (fifo)->front;					\
+	     c = (fifo)->data[iter], iter != (fifo)->back;		\
+	     iter = (iter + 1) & (fifo)->mask)
+
+#define __init_fifo(fifo, gfp)						\
+({									\
+	size_t _allocated_size, _bytes;					\
+	BUG_ON(!(fifo)->size);						\
+									\
+	_allocated_size = roundup_pow_of_two((fifo)->size + 1);		\
+	_bytes = _allocated_size * sizeof(*(fifo)->data);		\
+									\
+	(fifo)->mask = _allocated_size - 1;				\
+	(fifo)->front = (fifo)->back = 0;				\
+	(fifo)->data = NULL;						\
+									\
+	if (_bytes < KMALLOC_MAX_SIZE)					\
+		(fifo)->data = kmalloc(_bytes, (gfp));			\
+	if ((!(fifo)->data) && ((gfp) & GFP_KERNEL))			\
+		(fifo)->data = vmalloc(_bytes);				\
+	(fifo)->data;							\
+})
+
+#define init_fifo_exact(fifo, _size, gfp)				\
+({									\
+	(fifo)->size = (_size);						\
+	__init_fifo(fifo, gfp);						\
+})
+
+#define init_fifo(fifo, _size, gfp)					\
+({									\
+	(fifo)->size = (_size);						\
+	if ((fifo)->size > 4)						\
+		(fifo)->size = roundup_pow_of_two((fifo)->size) - 1;	\
+	__init_fifo(fifo, gfp);						\
+})
+
+#define free_fifo(fifo)							\
+do {									\
+	if (is_vmalloc_addr((fifo)->data))				\
+		vfree((fifo)->data);					\
+	else								\
+		kfree((fifo)->data);					\
+	(fifo)->data = NULL;						\
+} while (0)
+
+#define fifo_used(fifo)		(((fifo)->back - (fifo)->front) & (fifo)->mask)
+#define fifo_free(fifo)		((fifo)->size - fifo_used(fifo))
+
+#define fifo_empty(fifo)	(!fifo_used(fifo))
+#define fifo_full(fifo)		(!fifo_free(fifo))
+
+#define fifo_front(fifo)	((fifo)->data[(fifo)->front])
+#define fifo_back(fifo)							\
+	((fifo)->data[((fifo)->back - 1) & (fifo)->mask])
+
+#define fifo_idx(fifo, p)	(((p) - &fifo_front(fifo)) & (fifo)->mask)
+
+#define fifo_push_back(fifo, i)						\
+({									\
+	bool _r = !fifo_full((fifo));					\
+	if (_r) {							\
+		(fifo)->data[(fifo)->back++] = (i);			\
+		(fifo)->back &= (fifo)->mask;				\
+	}								\
+	_r;								\
+})
+
+#define fifo_pop_front(fifo, i)						\
+({									\
+	bool _r = !fifo_empty((fifo));					\
+	if (_r) {							\
+		(i) = (fifo)->data[(fifo)->front++];			\
+		(fifo)->front &= (fifo)->mask;				\
+	}								\
+	_r;								\
+})
+
+#define fifo_push_front(fifo, i)					\
+({									\
+	bool _r = !fifo_full((fifo));					\
+	if (_r) {							\
+		--(fifo)->front;					\
+		(fifo)->front &= (fifo)->mask;				\
+		(fifo)->data[(fifo)->front] = (i);			\
+	}								\
+	_r;								\
+})
+
+#define fifo_pop_back(fifo, i)						\
+({									\
+	bool _r = !fifo_empty((fifo));					\
+	if (_r) {							\
+		--(fifo)->back;						\
+		(fifo)->back &= (fifo)->mask;				\
+		(i) = (fifo)->data[(fifo)->back]			\
+	}								\
+	_r;								\
+})
+
+#define fifo_push(fifo, i)	fifo_push_back(fifo, (i))
+#define fifo_pop(fifo, i)	fifo_pop_front(fifo, (i))
+
+#define fifo_swap(l, r)							\
+do {									\
+	swap((l)->front, (r)->front);					\
+	swap((l)->back, (r)->back);					\
+	swap((l)->size, (r)->size);					\
+	swap((l)->mask, (r)->mask);					\
+	swap((l)->data, (r)->data);					\
+} while (0)
+
+#define fifo_move(dest, src)						\
+do {									\
+	typeof(*((dest)->data)) _t;					\
+	while (!fifo_full(dest) &&					\
+	       fifo_pop(src, _t))					\
+		fifo_push(dest, _t);					\
+} while (0)
+
+/*
+ * Simple array based allocator - preallocates a number of elements and you can
+ * never allocate more than that, also has no locking.
+ *
+ * Handy because if you know you only need a fixed number of elements you don't
+ * have to worry about memory allocation failure, and sometimes a mempool isn't
+ * what you want.
+ *
+ * We treat the free elements as entries in a singly linked list, and the
+ * freelist as a stack - allocating and freeing push and pop off the freelist.
+ */
+
+#define DECLARE_ARRAY_ALLOCATOR(type, name, size)			\
+	struct {							\
+		type	*freelist;					\
+		type	data[size];					\
+	} name
+
+#define array_alloc(array)						\
+({									\
+	typeof((array)->freelist) _ret = (array)->freelist;		\
+									\
+	if (_ret)							\
+		(array)->freelist = *((typeof((array)->freelist) *) _ret);\
+									\
+	_ret;								\
+})
+
+#define array_free(array, ptr)						\
+do {									\
+	typeof((array)->freelist) _ptr = ptr;				\
+									\
+	*((typeof((array)->freelist) *) _ptr) = (array)->freelist;	\
+	(array)->freelist = _ptr;					\
+} while (0)
+
+#define array_allocator_init(array)					\
+do {									\
+	typeof((array)->freelist) _i;					\
+									\
+	BUILD_BUG_ON(sizeof((array)->data[0]) < sizeof(void *));	\
+	(array)->freelist = NULL;					\
+									\
+	for (_i = (array)->data;					\
+	     _i < (array)->data + ARRAY_SIZE((array)->data);		\
+	     _i++)							\
+		array_free(array, _i);					\
+} while (0)
+
+#define array_freelist_empty(array)	((array)->freelist == NULL)
+
+#define ANYSINT_MAX(t)							\
+	((((t) 1 << (sizeof(t) * 8 - 2)) - (t) 1) * (t) 2 + (t) 1)
+
+int bch_strtoint_h(const char *, int *);
+int bch_strtouint_h(const char *, unsigned int *);
+int bch_strtoll_h(const char *, long long *);
+int bch_strtoull_h(const char *, unsigned long long *);
+
+static inline int bch_strtol_h(const char *cp, long *res)
+{
+#if BITS_PER_LONG == 32
+	return bch_strtoint_h(cp, (int *) res);
+#else
+	return bch_strtoll_h(cp, (long long *) res);
+#endif
+}
+
+static inline int bch_strtoul_h(const char *cp, long *res)
+{
+#if BITS_PER_LONG == 32
+	return bch_strtouint_h(cp, (unsigned int *) res);
+#else
+	return bch_strtoull_h(cp, (unsigned long long *) res);
+#endif
+}
+
+#define strtoi_h(cp, res)						\
+	(__builtin_types_compatible_p(typeof(*res), int)		\
+	? bch_strtoint_h(cp, (void *) res)				\
+	: __builtin_types_compatible_p(typeof(*res), long)		\
+	? bch_strtol_h(cp, (void *) res)				\
+	: __builtin_types_compatible_p(typeof(*res), long long)		\
+	? bch_strtoll_h(cp, (void *) res)				\
+	: __builtin_types_compatible_p(typeof(*res), unsigned int)	\
+	? bch_strtouint_h(cp, (void *) res)				\
+	: __builtin_types_compatible_p(typeof(*res), unsigned long)	\
+	? bch_strtoul_h(cp, (void *) res)				\
+	: __builtin_types_compatible_p(typeof(*res), unsigned long long)\
+	? bch_strtoull_h(cp, (void *) res) : -EINVAL)
+
+#define strtoul_safe(cp, var)						\
+({									\
+	unsigned long _v;						\
+	int _r = kstrtoul(cp, 10, &_v);					\
+	if (!_r)							\
+		var = _v;						\
+	_r;								\
+})
+
+#define strtoul_safe_clamp(cp, var, min, max)				\
+({									\
+	unsigned long _v;						\
+	int _r = kstrtoul(cp, 10, &_v);					\
+	if (!_r)							\
+		var = clamp_t(typeof(var), _v, min, max);		\
+	_r;								\
+})
+
+#define snprint(buf, size, var)						\
+	snprintf(buf, size,						\
+		__builtin_types_compatible_p(typeof(var), int)		\
+		     ? "%i\n" :						\
+		__builtin_types_compatible_p(typeof(var), unsigned)	\
+		     ? "%u\n" :						\
+		__builtin_types_compatible_p(typeof(var), long)		\
+		     ? "%li\n" :					\
+		__builtin_types_compatible_p(typeof(var), unsigned long)\
+		     ? "%lu\n" :					\
+		__builtin_types_compatible_p(typeof(var), int64_t)	\
+		     ? "%lli\n" :					\
+		__builtin_types_compatible_p(typeof(var), uint64_t)	\
+		     ? "%llu\n" :					\
+		__builtin_types_compatible_p(typeof(var), const char *)	\
+		     ? "%s\n" : "%i\n", var)
+
+ssize_t bch_hprint(char *buf, int64_t v);
+
+bool bch_is_zero(const char *p, size_t n);
+int bch_parse_uuid(const char *s, char *uuid);
+
+ssize_t bch_snprint_string_list(char *buf, size_t size, const char * const list[],
+			    size_t selected);
+
+ssize_t bch_read_string_list(const char *buf, const char * const list[]);
+
+struct time_stats {
+	/*
+	 * all fields are in nanoseconds, averages are ewmas stored left shifted
+	 * by 8
+	 */
+	uint64_t	max_duration;
+	uint64_t	average_duration;
+	uint64_t	average_frequency;
+	uint64_t	last;
+};
+
+void bch_time_stats_update(struct time_stats *stats, uint64_t time);
+
+#define NSEC_PER_ns			1L
+#define NSEC_PER_us			NSEC_PER_USEC
+#define NSEC_PER_ms			NSEC_PER_MSEC
+#define NSEC_PER_sec			NSEC_PER_SEC
+
+#define __print_time_stat(stats, name, stat, units)			\
+	sysfs_print(name ## _ ## stat ## _ ## units,			\
+		    div_u64((stats)->stat >> 8, NSEC_PER_ ## units))
+
+#define sysfs_print_time_stats(stats, name,				\
+			       frequency_units,				\
+			       duration_units)				\
+do {									\
+	__print_time_stat(stats, name,					\
+			  average_frequency,	frequency_units);	\
+	__print_time_stat(stats, name,					\
+			  average_duration,	duration_units);	\
+	__print_time_stat(stats, name,					\
+			  max_duration,		duration_units);	\
+									\
+	sysfs_print(name ## _last_ ## frequency_units, (stats)->last	\
+		    ? div_s64(local_clock() - (stats)->last,		\
+			      NSEC_PER_ ## frequency_units)		\
+		    : -1LL);						\
+} while (0)
+
+#define sysfs_time_stats_attribute(name,				\
+				   frequency_units,			\
+				   duration_units)			\
+read_attribute(name ## _average_frequency_ ## frequency_units);		\
+read_attribute(name ## _average_duration_ ## duration_units);		\
+read_attribute(name ## _max_duration_ ## duration_units);		\
+read_attribute(name ## _last_ ## frequency_units)
+
+#define sysfs_time_stats_attribute_list(name,				\
+					frequency_units,		\
+					duration_units)			\
+&sysfs_ ## name ## _average_frequency_ ## frequency_units,		\
+&sysfs_ ## name ## _average_duration_ ## duration_units,		\
+&sysfs_ ## name ## _max_duration_ ## duration_units,			\
+&sysfs_ ## name ## _last_ ## frequency_units,
+
+#define ewma_add(ewma, val, weight, factor)				\
+({									\
+	(ewma) *= (weight) - 1;						\
+	(ewma) += (val) << factor;					\
+	(ewma) /= (weight);						\
+	(ewma) >> factor;						\
+})
+
+struct ratelimit {
+	uint64_t		next;
+	unsigned		rate;
+};
+
+static inline void ratelimit_reset(struct ratelimit *d)
+{
+	d->next = local_clock();
+}
+
+unsigned bch_next_delay(struct ratelimit *d, uint64_t done);
+
+#define __DIV_SAFE(n, d, zero)						\
+({									\
+	typeof(n) _n = (n);						\
+	typeof(d) _d = (d);						\
+	_d ? _n / _d : zero;						\
+})
+
+#define DIV_SAFE(n, d)	__DIV_SAFE(n, d, 0)
+
+#define container_of_or_null(ptr, type, member)				\
+({									\
+	typeof(ptr) _ptr = ptr;						\
+	_ptr ? container_of(_ptr, type, member) : NULL;			\
+})
+
+#define RB_INSERT(root, new, member, cmp)				\
+({									\
+	__label__ dup;							\
+	struct rb_node **n = &(root)->rb_node, *parent = NULL;		\
+	typeof(new) this;						\
+	int res, ret = -1;						\
+									\
+	while (*n) {							\
+		parent = *n;						\
+		this = container_of(*n, typeof(*(new)), member);	\
+		res = cmp(new, this);					\
+		if (!res)						\
+			goto dup;					\
+		n = res < 0						\
+			? &(*n)->rb_left				\
+			: &(*n)->rb_right;				\
+	}								\
+									\
+	rb_link_node(&(new)->member, parent, n);			\
+	rb_insert_color(&(new)->member, root);				\
+	ret = 0;							\
+dup:									\
+	ret;								\
+})
+
+#define RB_SEARCH(root, search, member, cmp)				\
+({									\
+	struct rb_node *n = (root)->rb_node;				\
+	typeof(&(search)) this, ret = NULL;				\
+	int res;							\
+									\
+	while (n) {							\
+		this = container_of(n, typeof(search), member);		\
+		res = cmp(&(search), this);				\
+		if (!res) {						\
+			ret = this;					\
+			break;						\
+		}							\
+		n = res < 0						\
+			? n->rb_left					\
+			: n->rb_right;					\
+	}								\
+	ret;								\
+})
+
+#define RB_GREATER(root, search, member, cmp)				\
+({									\
+	struct rb_node *n = (root)->rb_node;				\
+	typeof(&(search)) this, ret = NULL;				\
+	int res;							\
+									\
+	while (n) {							\
+		this = container_of(n, typeof(search), member);		\
+		res = cmp(&(search), this);				\
+		if (res < 0) {						\
+			ret = this;					\
+			n = n->rb_left;					\
+		} else							\
+			n = n->rb_right;				\
+	}								\
+	ret;								\
+})
+
+#define RB_FIRST(root, type, member)					\
+	container_of_or_null(rb_first(root), type, member)
+
+#define RB_LAST(root, type, member)					\
+	container_of_or_null(rb_last(root), type, member)
+
+#define RB_NEXT(ptr, member)						\
+	container_of_or_null(rb_next(&(ptr)->member), typeof(*ptr), member)
+
+#define RB_PREV(ptr, member)						\
+	container_of_or_null(rb_prev(&(ptr)->member), typeof(*ptr), member)
+
+/* Does linear interpolation between powers of two */
+static inline unsigned fract_exp_two(unsigned x, unsigned fract_bits)
+{
+	unsigned fract = x & ~(~0 << fract_bits);
+
+	x >>= fract_bits;
+	x   = 1 << x;
+	x  += (x * fract) >> fract_bits;
+
+	return x;
+}
+
+#define bio_end(bio)	((bio)->bi_sector + bio_sectors(bio))
+
+void bch_bio_map(struct bio *bio, void *base);
+
+int bch_bio_alloc_pages(struct bio *bio, gfp_t gfp);
+
+static inline sector_t bdev_sectors(struct block_device *bdev)
+{
+	return bdev->bd_inode->i_size >> 9;
+}
+
+#define closure_bio_submit(bio, cl, dev)				\
+do {									\
+	closure_get(cl);						\
+	bch_generic_make_request(bio, &(dev)->bio_split_hook);		\
+} while (0)
+
+uint64_t bch_crc64_update(uint64_t, const void *, size_t);
+uint64_t bch_crc64(const void *, size_t);
+
+#endif /* _BCACHE_UTIL_H */

drivers/md/bcache/writeback.c

···
+/*
+ * background writeback - scan btree for dirty data and write it to the backing
+ * device
+ *
+ * Copyright 2010, 2011 Kent Overstreet <[e-mail]>
+ * Copyright 2012 Google, Inc.
+ */
+
+#include "bcache.h"
+#include "btree.h"
+#include "debug.h"
+
+static struct workqueue_struct *dirty_wq;
+
+static void read_dirty(struct closure *);
+
+struct dirty_io {
+	struct closure		cl;
+	struct cached_dev	*dc;
+	struct bio		bio;
+};
+
+/* Rate limiting */
+
+static void __update_writeback_rate(struct cached_dev *dc)
+{
+	struct cache_set *c = dc->disk.c;
+	uint64_t cache_sectors = c->nbuckets * c->sb.bucket_size;
+	uint64_t cache_dirty_target =
+		div_u64(cache_sectors * dc->writeback_percent, 100);
+
+	int64_t target = div64_u64(cache_dirty_target * bdev_sectors(dc->bdev),
+				   c->cached_dev_sectors);
+
+	/* PD controller */
+
+	int change = 0;
+	int64_t error;
+	int64_t dirty = atomic_long_read(&dc->disk.sectors_dirty);
+	int64_t derivative = dirty - dc->disk.sectors_dirty_last;
+
+	dc->disk.sectors_dirty_last = dirty;
+
+	derivative *= dc->writeback_rate_d_term;
+	derivative = clamp(derivative, -dirty, dirty);
+
+	derivative = ewma_add(dc->disk.sectors_dirty_derivative, derivative,
+			      dc->writeback_rate_d_smooth, 0);
+
+	/* Avoid divide by zero */
+	if (!target)
+		goto out;
+
+	error = div64_s64((dirty + derivative - target) << 8, target);
+
+	change = div_s64((dc->writeback_rate.rate * error) >> 8,
+			 dc->writeback_rate_p_term_inverse);
+
+	/* Don't increase writeback rate if the device isn't keeping up */
+	if (change > 0 &&
+	    time_after64(local_clock(),
+			 dc->writeback_rate.next + 10 * NSEC_PER_MSEC))
+		change = 0;
+
+	dc->writeback_rate.rate =
+		clamp_t(int64_t, dc->writeback_rate.rate + change,
+			1, NSEC_PER_MSEC);
+out:
+	dc->writeback_rate_derivative = derivative;
+	dc->writeback_rate_change = change;
+	dc->writeback_rate_target = target;
+
+	schedule_delayed_work(&dc->writeback_rate_update,
+			      dc->writeback_rate_update_seconds * HZ);
+}
+
+static void update_writeback_rate(struct work_struct *work)
+{
+	struct cached_dev *dc = container_of(to_delayed_work(work),
+					     struct cached_dev,
+					     writeback_rate_update);
+
+	down_read(&dc->writeback_lock);
+
+	if (atomic_read(&dc->has_dirty) &&
+	    dc->writeback_percent)
+		__update_writeback_rate(dc);
+
+	up_read(&dc->writeback_lock);
+}
+
+static unsigned writeback_delay(struct cached_dev *dc, unsigned sectors)
+{
+	if (atomic_read(&dc->disk.detaching) ||
+	    !dc->writeback_percent)
+		return 0;
+
+	return bch_next_delay(&dc->writeback_rate, sectors * 10000000ULL);
+}
+
+/* Background writeback */
+
+static bool dirty_pred(struct keybuf *buf, struct bkey *k)
+{
+	return KEY_DIRTY(k);
+}
+
+static void dirty_init(struct keybuf_key *w)
+{
+	struct dirty_io *io = w->private;
+	struct bio *bio = &io->bio;
+
+	bio_init(bio);
+	if (!io->dc->writeback_percent)
+		bio_set_prio(bio, IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0));
+
+	bio->bi_size		= KEY_SIZE(&w->key) << 9;
+	bio->bi_max_vecs	= DIV_ROUND_UP(KEY_SIZE(&w->key), PAGE_SECTORS);
+	bio->bi_private		= w;
+	bio->bi_io_vec		= bio->bi_inline_vecs;
+	bch_bio_map(bio, NULL);
+}
+
+static void refill_dirty(struct closure *cl)
+{
+	struct cached_dev *dc = container_of(cl, struct cached_dev,
+					     writeback.cl);
+	struct keybuf *buf = &dc->writeback_keys;
+	bool searched_from_start = false;
+	struct bkey end = MAX_KEY;
+	SET_KEY_INODE(&end, dc->disk.id);
+
+	if (!atomic_read(&dc->disk.detaching) &&
+	    !dc->writeback_running)
+		closure_return(cl);
+
+	down_write(&dc->writeback_lock);
+
+	if (!atomic_read(&dc->has_dirty)) {
+		SET_BDEV_STATE(&dc->sb, BDEV_STATE_CLEAN);
+		bch_write_bdev_super(dc, NULL);
+
+		up_write(&dc->writeback_lock);
+		closure_return(cl);
+	}
+
+	if (bkey_cmp(&buf->last_scanned, &end) >= 0) {
+		buf->last_scanned = KEY(dc->disk.id, 0, 0);
+		searched_from_start = true;
+	}
+
+	bch_refill_keybuf(dc->disk.c, buf, &end);
+
+	if (bkey_cmp(&buf->last_scanned, &end) >= 0 && searched_from_start) {
+		/* Searched the entire btree  - delay awhile */
+
+		if (RB_EMPTY_ROOT(&buf->keys)) {
+			atomic_set(&dc->has_dirty, 0);
+			cached_dev_put(dc);
+		}
+
+		if (!atomic_read(&dc->disk.detaching))
+			closure_delay(&dc->writeback, dc->writeback_delay * HZ);
+	}
+
+	up_write(&dc->writeback_lock);
+
+	ratelimit_reset(&dc->writeback_rate);
+
+	/* Punt to workqueue only so we don't recurse and blow the stack */
+	continue_at(cl, read_dirty, dirty_wq);
+}
+
+void bch_writeback_queue(struct cached_dev *dc)
+{
+	if (closure_trylock(&dc->writeback.cl, &dc->disk.cl)) {
+		if (!atomic_read(&dc->disk.detaching))
+			closure_delay(&dc->writeback, dc->writeback_delay * HZ);
+
+		continue_at(&dc->writeback.cl, refill_dirty, dirty_wq);
+	}
+}
+
+void bch_writeback_add(struct cached_dev *dc, unsigned sectors)
+{
+	atomic_long_add(sectors, &dc->disk.sectors_dirty);
+
+	if (!atomic_read(&dc->has_dirty) &&
+	    !atomic_xchg(&dc->has_dirty, 1)) {
+		atomic_inc(&dc->count);
+
+		if (BDEV_STATE(&dc->sb) != BDEV_STATE_DIRTY) {
+			SET_BDEV_STATE(&dc->sb, BDEV_STATE_DIRTY);
+			/* XXX: should do this synchronously */
+			bch_write_bdev_super(dc, NULL);
+		}
+
+		bch_writeback_queue(dc);
+
+		if (dc->writeback_percent)
+			schedule_delayed_work(&dc->writeback_rate_update,
+				      dc->writeback_rate_update_seconds * HZ);
+	}
+}
+
+/* Background writeback - IO loop */
+
+static void dirty_io_destructor(struct closure *cl)
+{
+	struct dirty_io *io = container_of(cl, struct dirty_io, cl);
+	kfree(io);
+}
+
+static void write_dirty_finish(struct closure *cl)
+{
+	struct dirty_io *io = container_of(cl, struct dirty_io, cl);
+	struct keybuf_key *w = io->bio.bi_private;
+	struct cached_dev *dc = io->dc;
+	struct bio_vec *bv = bio_iovec_idx(&io->bio, io->bio.bi_vcnt);
+
+	while (bv-- != io->bio.bi_io_vec)
+		__free_page(bv->bv_page);
+
+	/* This is kind of a dumb way of signalling errors. */
+	if (KEY_DIRTY(&w->key)) {
+		unsigned i;
+		struct btree_op op;
+		bch_btree_op_init_stack(&op);
+
+		op.type = BTREE_REPLACE;
+		bkey_copy(&op.replace, &w->key);
+
+		SET_KEY_DIRTY(&w->key, false);
+		bch_keylist_add(&op.keys, &w->key);
+
+		for (i = 0; i < KEY_PTRS(&w->key); i++)
+			atomic_inc(&PTR_BUCKET(dc->disk.c, &w->key, i)->pin);
+
+		pr_debug("clearing %s", pkey(&w->key));
+		bch_btree_insert(&op, dc->disk.c);
+		closure_sync(&op.cl);
+
+		atomic_long_inc(op.insert_collision
+				? &dc->disk.c->writeback_keys_failed
+				: &dc->disk.c->writeback_keys_done);
+	}
+
+	bch_keybuf_del(&dc->writeback_keys, w);
+	atomic_dec_bug(&dc->in_flight);
+
+	closure_wake_up(&dc->writeback_wait);
+
+	closure_return_with_destructor(cl, dirty_io_destructor);
+}
+
+static void dirty_endio(struct bio *bio, int error)
+{
+	struct keybuf_key *w = bio->bi_private;
+	struct dirty_io *io = w->private;
+
+	if (error)
+		SET_KEY_DIRTY(&w->key, false);
+
+	closure_put(&io->cl);
+}
+
+static void write_dirty(struct closure *cl)
+{
+	struct dirty_io *io = container_of(cl, struct dirty_io, cl);
+	struct keybuf_key *w = io->bio.bi_private;
+
+	dirty_init(w);
+	io->bio.bi_rw		= WRITE;
+	io->bio.bi_sector	= KEY_START(&w->key);
+	io->bio.bi_bdev		= io->dc->bdev;
+	io->bio.bi_end_io	= dirty_endio;
+
+	trace_bcache_write_dirty(&io->bio);
+	closure_bio_submit(&io->bio, cl, &io->dc->disk);
+
+	continue_at(cl, write_dirty_finish, dirty_wq);
+}
+
+static void read_dirty_endio(struct bio *bio, int error)
+{
+	struct keybuf_key *w = bio->bi_private;
+	struct dirty_io *io = w->private;
+
+	bch_count_io_errors(PTR_CACHE(io->dc->disk.c, &w->key, 0),
+			    error, "reading dirty data from cache");
+
+	dirty_endio(bio, error);
+}
+
+static void read_dirty_submit(struct closure *cl)
+{
+	struct dirty_io *io = container_of(cl, struct dirty_io, cl);
+
+	trace_bcache_read_dirty(&io->bio);
+	closure_bio_submit(&io->bio, cl, &io->dc->disk);
+
+	continue_at(cl, write_dirty, dirty_wq);
+}
+
+static void read_dirty(struct closure *cl)
+{
+	struct cached_dev *dc = container_of(cl, struct cached_dev,
+					     writeback.cl);
+	unsigned delay = writeback_delay(dc, 0);
+	struct keybuf_key *w;
+	struct dirty_io *io;
+
+	/*
+	 * XXX: if we error, background writeback just spins. Should use some
+	 * mempools.
+	 */
+
+	while (1) {
+		w = bch_keybuf_next(&dc->writeback_keys);
+		if (!w)
+			break;
+
+		BUG_ON(ptr_stale(dc->disk.c, &w->key, 0));
+
+		if (delay > 0 &&
+		    (KEY_START(&w->key) != dc->last_read ||
+		     jiffies_to_msecs(delay) > 50)) {
+			w->private = NULL;
+
+			closure_delay(&dc->writeback, delay);
+			continue_at(cl, read_dirty, dirty_wq);
+		}
+
+		dc->last_read	= KEY_OFFSET(&w->key);
+
+		io = kzalloc(sizeof(struct dirty_io) + sizeof(struct bio_vec)
+			     * DIV_ROUND_UP(KEY_SIZE(&w->key), PAGE_SECTORS),
+			     GFP_KERNEL);
+		if (!io)
+			goto err;
+
+		w->private	= io;
+		io->dc		= dc;
+
+		dirty_init(w);
+		io->bio.bi_sector	= PTR_OFFSET(&w->key, 0);
+		io->bio.bi_bdev		= PTR_CACHE(dc->disk.c,
+						    &w->key, 0)->bdev;
+		io->bio.bi_rw		= READ;
+		io->bio.bi_end_io	= read_dirty_endio;
+
+		if (bch_bio_alloc_pages(&io->bio, GFP_KERNEL))
+			goto err_free;
+
+		pr_debug("%s", pkey(&w->key));
+
+		closure_call(&io->cl, read_dirty_submit, NULL, &dc->disk.cl);
+
+		delay = writeback_delay(dc, KEY_SIZE(&w->key));
+
+		atomic_inc(&dc->in_flight);
+
+		if (!closure_wait_event(&dc->writeback_wait, cl,
+					atomic_read(&dc->in_flight) < 64))
+			continue_at(cl, read_dirty, dirty_wq);
+	}
+
+	if (0) {
+err_free:
+		kfree(w->private);
+err:
+		bch_keybuf_del(&dc->writeback_keys, w);
+	}
+
+	refill_dirty(cl);
+}
+
+void bch_writeback_init_cached_dev(struct cached_dev *dc)
+{
+	closure_init_unlocked(&dc->writeback);
+	init_rwsem(&dc->writeback_lock);
+
+	bch_keybuf_init(&dc->writeback_keys, dirty_pred);
+
+	dc->writeback_metadata		= true;
+	dc->writeback_running		= true;
+	dc->writeback_percent		= 10;
+	dc->writeback_delay		= 30;
+	dc->writeback_rate.rate		= 1024;
+
+	dc->writeback_rate_update_seconds = 30;
+	dc->writeback_rate_d_term	= 16;
+	dc->writeback_rate_p_term_inverse = 64;
+	dc->writeback_rate_d_smooth	= 8;
+
+	INIT_DELAYED_WORK(&dc->writeback_rate_update, update_writeback_rate);
+	schedule_delayed_work(&dc->writeback_rate_update,
+			      dc->writeback_rate_update_seconds * HZ);
+}
+
+void bch_writeback_exit(void)
+{
+	if (dirty_wq)
+		destroy_workqueue(dirty_wq);
+}
+
+int __init bch_writeback_init(void)
+{
+	dirty_wq = create_singlethread_workqueue("bcache_writeback");
+	if (!dirty_wq)
+		return -ENOMEM;
+
+	return 0;
+}

include/linux/cgroup_subsys.h

···
 #endif
 
 /* */
+
+#ifdef CONFIG_CGROUP_BCACHE
+SUBSYS(bcache)
+#endif
+
+/* */

include/linux/drbd.h

···
 #endif
 
 extern const char *drbd_buildtag(void);
-#define REL_VERSION "8.4.2"
+#define REL_VERSION "8.4.3"
 #define API_VERSION 1
 #define PRO_VERSION_MIN 86
 #define PRO_VERSION_MAX 101
···
 	SS_IN_TRANSIENT_STATE = -18,  /* Retry after the next state change */
 	SS_CONCURRENT_ST_CHG = -19,   /* Concurrent cluster side state change! */
 	SS_O_VOL_PEER_PRI = -20,
-	SS_AFTER_LAST_ERROR = -21,    /* Keep this at bottom */
+	SS_OUTDATE_WO_CONN = -21,
+	SS_AFTER_LAST_ERROR = -22,    /* Keep this at bottom */
 };
 
 /* from drbd_strings.c */

include/linux/drbd_limits.h

···
 #define DRBD_RESYNC_RATE_DEF 250
 #define DRBD_RESYNC_RATE_SCALE 'k'  /* kilobytes */
 
-  /* less than 7 would hit performance unnecessarily.
-   * 919 slots context information per transaction,
-   * 32k activity log, 4k transaction size,
-   * one transaction in flight:
-   * 919 * 7 = 6433 */
+  /* less than 7 would hit performance unnecessarily. */
 #define DRBD_AL_EXTENTS_MIN  7
-#define DRBD_AL_EXTENTS_MAX  6433
+  /* we use u16 as "slot number", (u16)~0 is "FREE".
+   * If you use >= 292 kB on-disk ring buffer,
+   * this is the maximum you can use: */
+#define DRBD_AL_EXTENTS_MAX  0xfffe
 #define DRBD_AL_EXTENTS_DEF  1237
 #define DRBD_AL_EXTENTS_SCALE '1'
 

include/linux/idr.h

···
  * @idp:     idr handle
  * @entry:   the type * to use as cursor
  * @id:      id entry's key
+ *
+ * @entry and @id do not need to be initialized before the loop, and
+ * after normal terminatinon @entry is left with the value NULL.  This
+ * is convenient for a "not found" value.
  */
-#define idr_for_each_entry(idp, entry, id)				\
-	for (id = 0, entry = (typeof(entry))idr_get_next((idp), &(id)); \
-	     entry != NULL;                                             \
-	     ++id, entry = (typeof(entry))idr_get_next((idp), &(id)))
+#define idr_for_each_entry(idp, entry, id)			\
+	for (id = 0; ((entry) = idr_get_next(idp, &(id))) != NULL; ++id)
 
 /*
  * Don't use the following functions.  These exist only to suppress

include/linux/lru_cache.h

···
 extern void lc_set(struct lru_cache *lc, unsigned int enr, int index);
 extern void lc_del(struct lru_cache *lc, struct lc_element *element);
 
+extern struct lc_element *lc_get_cumulative(struct lru_cache *lc, unsigned int enr);
 extern struct lc_element *lc_try_get(struct lru_cache *lc, unsigned int enr);
 extern struct lc_element *lc_find(struct lru_cache *lc, unsigned int enr);
 extern struct lc_element *lc_get(struct lru_cache *lc, unsigned int enr);

include/linux/rwsem.h

···
 	_down_write_nest_lock(sem, &(nest_lock)->dep_map);	\
 } while (0);
 
+/*
+ * Take/release a lock when not the owner will release it.
+ *
+ * [ This API should be avoided as much as possible - the
+ *   proper abstraction for this case is completions. ]
+ */
+extern void down_read_non_owner(struct rw_semaphore *sem);
+extern void up_read_non_owner(struct rw_semaphore *sem);
 #else
 # define down_read_nested(sem, subclass)		down_read(sem)
 # define down_write_nest_lock(sem, nest_lock)	down_write(sem)
 # define down_write_nested(sem, subclass)	down_write(sem)
+# define down_read_non_owner(sem)		down_read(sem)
+# define up_read_non_owner(sem)			up_read(sem)
 #endif
 
 #endif /* _LINUX_RWSEM_H */

include/linux/sched.h

···
 #ifdef CONFIG_UPROBES
 	struct uprobe_task *utask;
 #endif
+#if defined(CONFIG_BCACHE) || defined(CONFIG_BCACHE_MODULE)
+	unsigned int	sequential_io;
+	unsigned int	sequential_io_avg;
+#endif
 };
 
 /* Future-safe accessor for struct task_struct's cpus_allowed. */

include/trace/events/bcache.h

···
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM bcache
+
+#if !defined(_TRACE_BCACHE_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_BCACHE_H
+
+#include <linux/tracepoint.h>
+
+struct search;
+
+DECLARE_EVENT_CLASS(bcache_request,
+
+	TP_PROTO(struct search *s, struct bio *bio),
+
+	TP_ARGS(s, bio),
+
+	TP_STRUCT__entry(
+		__field(dev_t,		dev			)
+		__field(unsigned int,	orig_major		)
+		__field(unsigned int,	orig_minor		)
+		__field(sector_t,	sector			)
+		__field(dev_t,		orig_sector		)
+		__field(unsigned int,	nr_sector		)
+		__array(char,		rwbs,	6		)
+		__array(char,		comm,	TASK_COMM_LEN	)
+	),
+
+	TP_fast_assign(
+		__entry->dev		= bio->bi_bdev->bd_dev;
+		__entry->orig_major	= s->d->disk->major;
+		__entry->orig_minor	= s->d->disk->first_minor;
+		__entry->sector		= bio->bi_sector;
+		__entry->orig_sector	= bio->bi_sector - 16;
+		__entry->nr_sector	= bio->bi_size >> 9;
+		blk_fill_rwbs(__entry->rwbs, bio->bi_rw, bio->bi_size);
+		memcpy(__entry->comm, current->comm, TASK_COMM_LEN);
+	),
+
+	TP_printk("%d,%d %s %llu + %u [%s] (from %d,%d @ %llu)",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->rwbs,
+		  (unsigned long long)__entry->sector,
+		  __entry->nr_sector, __entry->comm,
+		  __entry->orig_major, __entry->orig_minor,
+		  (unsigned long long)__entry->orig_sector)
+);
+
+DEFINE_EVENT(bcache_request, bcache_request_start,
+
+	TP_PROTO(struct search *s, struct bio *bio),
+
+	TP_ARGS(s, bio)
+);
+
+DEFINE_EVENT(bcache_request, bcache_request_end,
+
+	TP_PROTO(struct search *s, struct bio *bio),
+
+	TP_ARGS(s, bio)
+);
+
+DECLARE_EVENT_CLASS(bcache_bio,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio),
+
+	TP_STRUCT__entry(
+		__field(dev_t,		dev			)
+		__field(sector_t,	sector			)
+		__field(unsigned int,	nr_sector		)
+		__array(char,		rwbs,	6		)
+		__array(char,		comm,	TASK_COMM_LEN	)
+	),
+
+	TP_fast_assign(
+		__entry->dev		= bio->bi_bdev->bd_dev;
+		__entry->sector		= bio->bi_sector;
+		__entry->nr_sector	= bio->bi_size >> 9;
+		blk_fill_rwbs(__entry->rwbs, bio->bi_rw, bio->bi_size);
+		memcpy(__entry->comm, current->comm, TASK_COMM_LEN);
+	),
+
+	TP_printk("%d,%d  %s %llu + %u [%s]",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->rwbs,
+		  (unsigned long long)__entry->sector,
+		  __entry->nr_sector, __entry->comm)
+);
+
+
+DEFINE_EVENT(bcache_bio, bcache_passthrough,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio)
+);
+
+DEFINE_EVENT(bcache_bio, bcache_cache_hit,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio)
+);
+
+DEFINE_EVENT(bcache_bio, bcache_cache_miss,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio)
+);
+
+DEFINE_EVENT(bcache_bio, bcache_read_retry,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio)
+);
+
+DEFINE_EVENT(bcache_bio, bcache_writethrough,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio)
+);
+
+DEFINE_EVENT(bcache_bio, bcache_writeback,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio)
+);
+
+DEFINE_EVENT(bcache_bio, bcache_write_skip,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio)
+);
+
+DEFINE_EVENT(bcache_bio, bcache_btree_read,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio)
+);
+
+DEFINE_EVENT(bcache_bio, bcache_btree_write,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio)
+);
+
+DEFINE_EVENT(bcache_bio, bcache_write_dirty,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio)
+);
+
+DEFINE_EVENT(bcache_bio, bcache_read_dirty,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio)
+);
+
+DEFINE_EVENT(bcache_bio, bcache_write_moving,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio)
+);
+
+DEFINE_EVENT(bcache_bio, bcache_read_moving,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio)
+);
+
+DEFINE_EVENT(bcache_bio, bcache_journal_write,
+
+	TP_PROTO(struct bio *bio),
+
+	TP_ARGS(bio)
+);
+
+DECLARE_EVENT_CLASS(bcache_cache_bio,
+
+	TP_PROTO(struct bio *bio,
+		 sector_t orig_sector,
+		 struct block_device* orig_bdev),
+
+	TP_ARGS(bio, orig_sector, orig_bdev),
+
+	TP_STRUCT__entry(
+		__field(dev_t,		dev			)
+		__field(dev_t,		orig_dev		)
+		__field(sector_t,	sector			)
+		__field(sector_t,	orig_sector		)
+		__field(unsigned int,	nr_sector		)
+		__array(char,		rwbs,	6		)
+		__array(char,		comm,	TASK_COMM_LEN	)
+	),
+
+	TP_fast_assign(
+		__entry->dev		= bio->bi_bdev->bd_dev;
+		__entry->orig_dev	= orig_bdev->bd_dev;
+		__entry->sector		= bio->bi_sector;
+		__entry->orig_sector	= orig_sector;
+		__entry->nr_sector	= bio->bi_size >> 9;
+		blk_fill_rwbs(__entry->rwbs, bio->bi_rw, bio->bi_size);
+		memcpy(__entry->comm, current->comm, TASK_COMM_LEN);
+	),
+
+	TP_printk("%d,%d  %s %llu + %u [%s] (from %d,%d %llu)",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->rwbs,
+		  (unsigned long long)__entry->sector,
+		  __entry->nr_sector, __entry->comm,
+		  MAJOR(__entry->orig_dev), MINOR(__entry->orig_dev),
+		  (unsigned long long)__entry->orig_sector)
+);
+
+DEFINE_EVENT(bcache_cache_bio, bcache_cache_insert,
+
+	TP_PROTO(struct bio *bio,
+		 sector_t orig_sector,
+		 struct block_device *orig_bdev),
+
+	TP_ARGS(bio, orig_sector, orig_bdev)
+);
+
+DECLARE_EVENT_CLASS(bcache_gc,
+
+	TP_PROTO(uint8_t *uuid),
+
+	TP_ARGS(uuid),
+
+	TP_STRUCT__entry(
+		__field(uint8_t *,	uuid)
+	),
+
+	TP_fast_assign(
+		__entry->uuid		= uuid;
+	),
+
+	TP_printk("%pU", __entry->uuid)
+);
+
+
+DEFINE_EVENT(bcache_gc, bcache_gc_start,
+
+	     TP_PROTO(uint8_t *uuid),
+
+	     TP_ARGS(uuid)
+);
+
+DEFINE_EVENT(bcache_gc, bcache_gc_end,
+
+	     TP_PROTO(uint8_t *uuid),
+
+	     TP_ARGS(uuid)
+);
+
+#endif /* _TRACE_BCACHE_H */
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>

kernel/fork.c

···
 	p->memcg_batch.do_batch = 0;
 	p->memcg_batch.memcg = NULL;
 #endif
+#ifdef CONFIG_BCACHE
+	p->sequential_io	= 0;
+	p->sequential_io_avg	= 0;
+#endif
 
 	/* Perform scheduler related setup. Assign this task to a CPU. */
 	sched_fork(p);

kernel/lockdep.c

···
 EXPORT_SYMBOL_GPL(lockdep_init_map);
 
 struct lock_class_key __lockdep_no_validate__;
+EXPORT_SYMBOL_GPL(__lockdep_no_validate__);
 
 static int
 print_lock_nested_lock_not_held(struct task_struct *curr,

kernel/rwsem.c

···
 
 EXPORT_SYMBOL(_down_write_nest_lock);
 
+void down_read_non_owner(struct rw_semaphore *sem)
+{
+	might_sleep();
+
+	__down_read(sem);
+}
+
+EXPORT_SYMBOL(down_read_non_owner);
+
 void down_write_nested(struct rw_semaphore *sem, int subclass)
 {
 	might_sleep();
···
 }
 
 EXPORT_SYMBOL(down_write_nested);
+
+void up_read_non_owner(struct rw_semaphore *sem)
+{
+	__up_read(sem);
+}
+
+EXPORT_SYMBOL(up_read_non_owner);
 
 #endif
 

kernel/trace/blktrace.c

···
 
 	rwbs[i] = '\0';
 }
+EXPORT_SYMBOL_GPL(blk_fill_rwbs);
 
 #endif /* CONFIG_EVENT_TRACING */
 

lib/lru_cache.c

···
 	return 0;
 }
 
-static struct lc_element *__lc_get(struct lru_cache *lc, unsigned int enr, bool may_change)
+/* used as internal flags to __lc_get */
+enum {
+	LC_GET_MAY_CHANGE = 1,
+	LC_GET_MAY_USE_UNCOMMITTED = 2,
+};
+
+static struct lc_element *__lc_get(struct lru_cache *lc, unsigned int enr, unsigned int flags)
 {
 	struct lc_element *e;
 
···
 	 * this enr is currently being pulled in already,
 	 * and will be available once the pending transaction
 	 * has been committed. */
-	if (e && e->lc_new_number == e->lc_number) {
+	if (e) {
+		if (e->lc_new_number != e->lc_number) {
+			/* It has been found above, but on the "to_be_changed"
+			 * list, not yet committed.  Don't pull it in twice,
+			 * wait for the transaction, then try again...
+			 */
+			if (!(flags & LC_GET_MAY_USE_UNCOMMITTED))
+				RETURN(NULL);
+			/* ... unless the caller is aware of the implications,
+			 * probably preparing a cumulative transaction. */
+			++e->refcnt;
+			++lc->hits;
+			RETURN(e);
+		}
+		/* else: lc_new_number == lc_number; a real hit. */
 		++lc->hits;
 		if (e->refcnt++ == 0)
 			lc->used++;
 		list_move(&e->list, &lc->in_use); /* Not evictable... */
 		RETURN(e);
 	}
+	/* e == NULL */
 
 	++lc->misses;
-	if (!may_change)
-		RETURN(NULL);
-
-	/* It has been found above, but on the "to_be_changed" list, not yet
-	 * committed.  Don't pull it in twice, wait for the transaction, then
-	 * try again */
-	if (e)
+	if (!(flags & LC_GET_MAY_CHANGE))
 		RETURN(NULL);
 
 	/* To avoid races with lc_try_lock(), first, mark us dirty
···
  */
 struct lc_element *lc_get(struct lru_cache *lc, unsigned int enr)
 {
-	return __lc_get(lc, enr, 1);
+	return __lc_get(lc, enr, LC_GET_MAY_CHANGE);
+}
+
+/**
+ * lc_get_cumulative - like lc_get; also finds to-be-changed elements
+ * @lc: the lru cache to operate on
+ * @enr: the label to look up
+ *
+ * Unlike lc_get this also returns the element for @enr, if it is belonging to
+ * a pending transaction, so the return values are like for lc_get(),
+ * plus:
+ *
+ * pointer to an element already on the "to_be_changed" list.
+ * 	In this case, the cache was already marked %LC_DIRTY.
+ *
+ * Caller needs to make sure that the pending transaction is completed,
+ * before proceeding to actually use this element.
+ */
+struct lc_element *lc_get_cumulative(struct lru_cache *lc, unsigned int enr)
+{
+	return __lc_get(lc, enr, LC_GET_MAY_CHANGE|LC_GET_MAY_USE_UNCOMMITTED);
 }
 
 /**
···
 EXPORT_SYMBOL(lc_seq_dump_details);
 EXPORT_SYMBOL(lc_try_lock);
 EXPORT_SYMBOL(lc_is_used);
+EXPORT_SYMBOL(lc_get_cumulative);