The free space cache has turned out to be a scalability bottleneck on
large, busy filesystems. When the cache for a lot of block groups needs
to be written out, we can get extremely long commit times; if this
happens in the critical section, things are especially bad because we
block new transactions from happening.
The main problem with the free space cache is that it has to be written
out in its entirety and is managed in an ad hoc fashion. Using a B-tree
to store free space fixes this: updates can be done as needed and we get
all of the benefits of using a B-tree: checksumming, RAID handling,
well-understood behavior.
With the free space tree, we get commit times that are about the same as
the no cache case with load times slower than the free space cache case
but still much faster than the no cache case. Free space is represented
with extents until it becomes more space-efficient to use bitmaps,
giving us similar space overhead to the free space cache.
The operations on the free space tree are: adding and removing free
space, handling the creation and deletion of block groups, and loading
the free space for a block group. We can also create the free space tree
by walking the extent tree and clear the free space tree.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Chris Mason <clm@fb.com>
This exercises the various parts of the new qgroup accounting code. We do some
basic stuff and do some things with the shared refs to make sure all that code
works. I had to add a bunch of infrastructure because I needed to be able to
insert items into a fake tree without having to do all the hard work myself,
hopefully this will be usefull in the future. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Chris Mason <clm@fb.com>
After the change titled "Btrfs: add support for inode properties", if
btrfs was built-in the kernel (i.e. not as a module), it would cause a
kernel panic, as reported recently by Fengguang:
[ 2.024722] BUG: unable to handle kernel NULL pointer dereference at (null)
[ 2.027814] IP: [<ffffffff81501594>] crc32c+0xc/0x6b
[ 2.028684] PGD 0
[ 2.028684] Oops: 0000 [#1] SMP
[ 2.028684] Modules linked in:
[ 2.028684] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.13.0-rc7-04795-ga7b57c2 #1
[ 2.028684] Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
[ 2.028684] task: ffff88000edba100 ti: ffff88000edd6000 task.ti: ffff88000edd6000
[ 2.028684] RIP: 0010:[<ffffffff81501594>] [<ffffffff81501594>] crc32c+0xc/0x6b
[ 2.028684] RSP: 0000:ffff88000edd7e58 EFLAGS: 00010246
[ 2.028684] RAX: 0000000000000000 RBX: ffffffff82295550 RCX: 0000000000000000
[ 2.028684] RDX: 0000000000000011 RSI: ffffffff81efe393 RDI: 00000000fffffffe
[ 2.028684] RBP: ffff88000edd7e60 R08: 0000000000000003 R09: 0000000000015d20
[ 2.028684] R10: ffffffff81ef225e R11: ffffffff811b0222 R12: ffffffffffffffff
[ 2.028684] R13: 0000000000000239 R14: 0000000000000000 R15: 0000000000000000
[ 2.028684] FS: 0000000000000000(0000) GS:ffff88000fa00000(0000) knlGS:0000000000000000
[ 2.028684] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[ 2.028684] CR2: 0000000000000000 CR3: 000000000220c000 CR4: 00000000000006f0
[ 2.028684] Stack:
[ 2.028684] ffffffff82295550 ffff88000edd7e80 ffffffff8238af62 ffffffff8238ac05
[ 2.028684] 0000000000000000 ffff88000edd7e98 ffffffff8238ac0f ffffffff8238ac05
[ 2.028684] ffff88000edd7f08 ffffffff810002ba ffff88000edd7f00 ffffffff810e2404
[ 2.028684] Call Trace:
[ 2.028684] [<ffffffff8238af62>] btrfs_props_init+0x4f/0x96
[ 2.028684] [<ffffffff8238ac05>] ? ftrace_define_fields_btrfs_space_reservation+0x145/0x145
[ 2.028684] [<ffffffff8238ac0f>] init_btrfs_fs+0xa/0xf0
[ 2.028684] [<ffffffff8238ac05>] ? ftrace_define_fields_btrfs_space_reservation+0x145/0x145
[ 2.028684] [<ffffffff810002ba>] do_one_initcall+0xa4/0x13a
[ 2.028684] [<ffffffff810e2404>] ? parse_args+0x25f/0x33d
[ 2.028684] [<ffffffff8234cf75>] kernel_init_freeable+0x1aa/0x230
[ 2.028684] [<ffffffff8234c785>] ? do_early_param+0x88/0x88
[ 2.028684] [<ffffffff819f61b5>] ? rest_init+0x89/0x89
[ 2.028684] [<ffffffff819f61c3>] kernel_init+0xe/0x109
The issue here is that the initialization function of btrfs (super.c:init_btrfs_fs)
started using crc32c (from lib/libcrc32c.c). But when it needs to call crc32c (as
part of the properties initialization routine), the libcrc32c is not yet initialized,
so crc32c derreferenced a NULL pointer (lib/libcrc32c.c:tfm), causing the kernel
panic on boot.
The approach to fix this is to use crypto component directly to use its crc32c (which
is basically what lib/libcrc32c.c is, a wrapper around crypto). This is what ext4 is
doing as well, it uses crypto directly to get crc32c functionality.
Verified this works both when btrfs is built-in and when it's loadable kernel module.
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com>
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Chris Mason <clm@fb.com>
This change adds infrastructure to allow for generic properties for
inodes. Properties are name/value pairs that can be associated with
inodes for different purposes. They are stored as xattrs with the
prefix "btrfs."
Properties can be inherited - this means when a directory inode has
inheritable properties set, these are added to new inodes created
under that directory. Further, subvolumes can also have properties
associated with them, and they can be inherited from their parent
subvolume. Naturally, directory properties have priority over subvolume
properties (in practice a subvolume property is just a regular
property associated with the root inode, objectid 256, of the
subvolume's fs tree).
This change also adds one specific property implementation, named
"compression", whose values can be "lzo" or "zlib" and it's an
inheritable property.
The corresponding changes to btrfs-progs were also implemented.
A patch with xfstests for this feature will follow once there's
agreement on this change/feature.
Further, the script at the bottom of this commit message was used to
do some benchmarks to measure any performance penalties of this feature.
Basically the tests correspond to:
Test 1 - create a filesystem and mount it with compress-force=lzo,
then sequentially create N files of 64Kb each, measure how long it took
to create the files, unmount the filesystem, mount the filesystem and
perform an 'ls -lha' against the test directory holding the N files, and
report the time the command took.
Test 2 - create a filesystem and don't use any compression option when
mounting it - instead set the compression property of the subvolume's
root to 'lzo'. Then create N files of 64Kb, and report the time it took.
The unmount the filesystem, mount it again and perform an 'ls -lha' like
in the former test. This means every single file ends up with a property
(xattr) associated to it.
Test 3 - same as test 2, but uses 4 properties - 3 are duplicates of the
compression property, have no real effect other than adding more work
when inheriting properties and taking more btree leaf space.
Test 4 - same as test 3 but with 10 properties per file.
Results (in seconds, and averages of 5 runs each), for different N
numbers of files follow.
* Without properties (test 1)
file creation time ls -lha time
10 000 files 3.49 0.76
100 000 files 47.19 8.37
1 000 000 files 518.51 107.06
* With 1 property (compression property set to lzo - test 2)
file creation time ls -lha time
10 000 files 3.63 0.93
100 000 files 48.56 9.74
1 000 000 files 537.72 125.11
* With 4 properties (test 3)
file creation time ls -lha time
10 000 files 3.94 1.20
100 000 files 52.14 11.48
1 000 000 files 572.70 142.13
* With 10 properties (test 4)
file creation time ls -lha time
10 000 files 4.61 1.35
100 000 files 58.86 13.83
1 000 000 files 656.01 177.61
The increased latencies with properties are essencialy because of:
*) When creating an inode, we now synchronously write 1 more item
(an xattr item) for each property inherited from the parent dir
(or subvolume). This could be done in an asynchronous way such
as we do for dir intex items (delayed-inode.c), which could help
reduce the file creation latency;
*) With properties, we now have larger fs trees. For this particular
test each xattr item uses 75 bytes of leaf space in the fs tree.
This could be less by using a new item for xattr items, instead of
the current btrfs_dir_item, since we could cut the 'location' and
'type' fields (saving 18 bytes) and maybe 'transid' too (saving a
total of 26 bytes per xattr item) from the btrfs_dir_item type.
Also tried batching the xattr insertions (ignoring proper hash
collision handling, since it didn't exist) when creating files that
inherit properties from their parent inode/subvolume, but the end
results were (surprisingly) essentially the same.
Test script:
$ cat test.pl
#!/usr/bin/perl -w
use strict;
use Time::HiRes qw(time);
use constant NUM_FILES => 10_000;
use constant FILE_SIZES => (64 * 1024);
use constant DEV => '/dev/sdb4';
use constant MNT_POINT => '/home/fdmanana/btrfs-tests/dev';
use constant TEST_DIR => (MNT_POINT . '/testdir');
system("mkfs.btrfs", "-l", "16384", "-f", DEV) == 0 or die "mkfs.btrfs failed!";
# following line for testing without properties
#system("mount", "-o", "compress-force=lzo", DEV, MNT_POINT) == 0 or die "mount failed!";
# following 2 lines for testing with properties
system("mount", DEV, MNT_POINT) == 0 or die "mount failed!";
system("btrfs", "prop", "set", MNT_POINT, "compression", "lzo") == 0 or die "set prop failed!";
system("mkdir", TEST_DIR) == 0 or die "mkdir failed!";
my ($t1, $t2);
$t1 = time();
for (my $i = 1; $i <= NUM_FILES; $i++) {
my $p = TEST_DIR . '/file_' . $i;
open(my $f, '>', $p) or die "Error opening file!";
$f->autoflush(1);
for (my $j = 0; $j < FILE_SIZES; $j += 4096) {
print $f ('A' x 4096) or die "Error writing to file!";
}
close($f);
}
$t2 = time();
print "Time to create " . NUM_FILES . ": " . ($t2 - $t1) . " seconds.\n";
system("umount", DEV) == 0 or die "umount failed!";
system("mount", DEV, MNT_POINT) == 0 or die "mount failed!";
$t1 = time();
system("bash -c 'ls -lha " . TEST_DIR . " > /dev/null'") == 0 or die "ls failed!";
$t2 = time();
print "Time to ls -lha all files: " . ($t2 - $t1) . " seconds.\n";
system("umount", DEV) == 0 or die "umount failed!";
Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com>
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Chris Mason <clm@fb.com>
I'm going to be removing hole extents in the near future so I wanted to make a
sanity test for btrfs_get_extent to make sure I don't break anything in the
meantime. This patch just puts btrfs_get_extent through its paces by giving it
a completely unreasonable mapping to look at and make sure it is giving us back
maps that make sense. Thanks,
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
So both Liu and I made huge messes of find_lock_delalloc_range trying to fix
stuff, me first by fixing extent size, then him by fixing something I broke and
then me again telling him to fix it a different way. So this is obviously a
candidate for some testing. This patch adds a pseudo fs so we can allocate fake
inodes for tests that need an inode or pages. Then it addes a bunch of tests to
make sure find_lock_delalloc_range is acting the way it is supposed to. With
this patch and all of our previous patches to find_lock_delalloc_range I am sure
it is working as expected now. Thanks,
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
While looking at somebodys corruption I became completely convinced that
btrfs_split_item was broken, so I wrote this test to verify that it was working
as it was supposed to. Thankfully it appears to be working as intended, so just
add this test to make sure nobody breaks it in the future. Thanks,
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
Mapping UUIDs to subvolume IDs is an operation with a high effort
today. Today, the algorithm even has quadratic effort (based on the
number of existing subvolumes), which means, that it takes minutes
to send/receive a single subvolume if 10,000 subvolumes exist. But
even linear effort would be too much since it is a waste. And these
data structures to allow mapping UUIDs to subvolume IDs are created
every time a btrfs send/receive instance is started.
It is much more efficient to maintain a searchable persistent data
structure in the filesystem, one that is updated whenever a
subvolume/snapshot is created and deleted, and when the received
subvolume UUID is set by the btrfs-receive tool.
Therefore kernel code is added with this commit that is able to
maintain data structures in the filesystem that allow to quickly
search for a given UUID and to retrieve data that is assigned to
this UUID, like which subvolume ID is related to this UUID.
This commit adds a new tree to hold UUID-to-data mapping items. The
key of the items is the full UUID plus the key type BTRFS_UUID_KEY.
Multiple data blocks can be stored for a given UUID, a type/length/
value scheme is used.
Now follows the lengthy justification, why a new tree was added
instead of using the existing root tree:
The first approach was to not create another tree that holds UUID
items. Instead, the items should just go into the top root tree.
Unfortunately this confused the algorithm to assign the objectid
of subvolumes and snapshots. The reason is that
btrfs_find_free_objectid() calls btrfs_find_highest_objectid() for
the first created subvol or snapshot after mounting a filesystem,
and this function simply searches for the largest used objectid in
the root tree keys to pick the next objectid to assign. Of course,
the UUID keys have always been the ones with the highest offset
value, and the next assigned subvol ID was wastefully huge.
To use any other existing tree did not look proper. To apply a
workaround such as setting the objectid to zero in the UUID item
key and to implement collision handling would either add
limitations (in case of a btrfs_extend_item() approach to handle
the collisions) or a lot of complexity and source code (in case a
key would be looked up that is free of collisions). Adding new code
that introduces limitations is not good, and adding code that is
complex and lengthy for no good reason is also not good. That's the
justification why a completely new tree was introduced.
Signed-off-by: Stefan Behrens <sbehrens@giantdisaster.de>
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
The plan is to have a bunch of unit tests that run when btrfs is loaded when you
build with the appropriate config option. My ultimate goal is to have a test
for every non-static function we have, but at first I'm going to focus on the
things that cause us the most problems. To start out with this just adds a
tests/ directory and moves the existing free space cache tests into that
directory and sets up all of the infrastructure. Thanks,
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
This builds on David Woodhouse's original Btrfs raid5/6 implementation.
The code has changed quite a bit, blame Chris Mason for any bugs.
Read/modify/write is done after the higher levels of the filesystem have
prepared a given bio. This means the higher layers are not responsible
for building full stripes, and they don't need to query for the topology
of the extents that may get allocated during delayed allocation runs.
It also means different files can easily share the same stripe.
But, it does expose us to incorrect parity if we crash or lose power
while doing a read/modify/write cycle. This will be addressed in a
later commit.
Scrub is unable to repair crc errors on raid5/6 chunks.
Discard does not work on raid5/6 (yet)
The stripe size is fixed at 64KiB per disk. This will be tunable
in a later commit.
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
This adds a new file to the sources together with the header file
and the changes to ioctl.h and ctree.h that are required by the
new C source file. Additionally, 4 new functions are added to
volume.c that deal with device creation and destruction.
Signed-off-by: Stefan Behrens <sbehrens@giantdisaster.de>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
This is the kernel portion of btrfs send/receive
Conflicts:
fs/btrfs/Makefile
fs/btrfs/backref.h
fs/btrfs/ctree.c
fs/btrfs/ioctl.c
fs/btrfs/ioctl.h
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
This patch introduces the BTRFS_IOC_SEND ioctl that is
required for send. It allows btrfs-progs to implement
full and incremental sends. Patches for btrfs-progs will
follow.
Signed-off-by: Alexander Block <ablock84@googlemail.com>
Reviewed-by: David Sterba <dave@jikos.cz>
Reviewed-by: Arne Jansen <sensille@gmx.net>
Reviewed-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
Reviewed-by: Alex Lyakas <alex.bolshoy.btrfs@gmail.com>
ulist is a generic data structures to hold a collection of unique u64
values. The only operations it supports is adding to the list and
enumerating it.
It is possible to store an auxiliary value along with the key. The
implementation is preliminary and can probably be sped up significantly.
It is used by btrfs_find_all_roots() quota to translate recursions into
iterative loops.
Signed-off-by: Arne Jansen <sensille@gmx.net>
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
If the btrfs integrity check is enabled, the files required to
implement the checks are included in the build.
Signed-off-by: Stefan Behrens <sbehrens@giantdisaster.de>
This is the implementation for the generic read ahead framework.
To trigger a readahead, btrfs_reada_add must be called. It will start
a read ahead for the given range [start, end) on tree root. The returned
handle can either be used to wait on the readahead to finish
(btrfs_reada_wait), or to send it to the background (btrfs_reada_detach).
The read ahead works as follows:
On btrfs_reada_add, the root of the tree is inserted into a radix_tree.
reada_start_machine will then search for extents to prefetch and trigger
some reads. When a read finishes for a node, all contained node/leaf
pointers that lie in the given range will also be enqueued. The reads will
be triggered in sequential order, thus giving a big win over a naive
enumeration. It will also make use of multi-device layouts. Each disk
will have its on read pointer and all disks will by utilized in parallel.
Also will no two disks read both sides of a mirror simultaneously, as this
would waste seeking capacity. Instead both disks will read different parts
of the filesystem.
Any number of readaheads can be started in parallel. The read order will be
determined globally, i.e. 2 parallel readaheads will normally finish faster
than the 2 started one after another.
Changes v2:
- protect root->node by transaction instead of node_lock
- fix missed branches:
The readahead had a too simple check to determine if a branch from
a node should be checked or not. It now also records the upper bound
of each node to see if the requested RA range lies within.
- use KERN_CONT to debug output, to avoid line breaks
- defer reada_start_machine to worker to avoid deadlock
Changes v3:
- protect root->node by rcu
Changes v5:
- changed EIO-semantics of reada_tree_block_flagged
- remove spin_lock from reada_control and make elems an atomic_t
- remove unused read_total from reada_control
- kill reada_key_cmp, use btrfs_comp_cpu_keys instead
- use kref-style release functions where possible
- return struct reada_control * instead of void * from btrfs_reada_add
Signed-off-by: Arne Jansen <sensille@gmx.net>
These helper functions iterate back references and call a function for each
backref. There is also a function to resolve an inode to a path in the
file system.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
Changelog V5 -> V6:
- Fix oom when the memory load is high, by storing the delayed nodes into the
root's radix tree, and letting btrfs inodes go.
Changelog V4 -> V5:
- Fix the race on adding the delayed node to the inode, which is spotted by
Chris Mason.
- Merge Chris Mason's incremental patch into this patch.
- Fix deadlock between readdir() and memory fault, which is reported by
Itaru Kitayama.
Changelog V3 -> V4:
- Fix nested lock, which is reported by Itaru Kitayama, by updating space cache
inode in time.
Changelog V2 -> V3:
- Fix the race between the delayed worker and the task which does delayed items
balance, which is reported by Tsutomu Itoh.
- Modify the patch address David Sterba's comment.
- Fix the bug of the cpu recursion spinlock, reported by Chris Mason
Changelog V1 -> V2:
- break up the global rb-tree, use a list to manage the delayed nodes,
which is created for every directory and file, and used to manage the
delayed directory name index items and the delayed inode item.
- introduce a worker to deal with the delayed nodes.
Compare with Ext3/4, the performance of file creation and deletion on btrfs
is very poor. the reason is that btrfs must do a lot of b+ tree insertions,
such as inode item, directory name item, directory name index and so on.
If we can do some delayed b+ tree insertion or deletion, we can improve the
performance, so we made this patch which implemented delayed directory name
index insertion/deletion and delayed inode update.
Implementation:
- introduce a delayed root object into the filesystem, that use two lists to
manage the delayed nodes which are created for every file/directory.
One is used to manage all the delayed nodes that have delayed items. And the
other is used to manage the delayed nodes which is waiting to be dealt with
by the work thread.
- Every delayed node has two rb-tree, one is used to manage the directory name
index which is going to be inserted into b+ tree, and the other is used to
manage the directory name index which is going to be deleted from b+ tree.
- introduce a worker to deal with the delayed operation. This worker is used
to deal with the works of the delayed directory name index items insertion
and deletion and the delayed inode update.
When the delayed items is beyond the lower limit, we create works for some
delayed nodes and insert them into the work queue of the worker, and then
go back.
When the delayed items is beyond the upper bound, we create works for all
the delayed nodes that haven't been dealt with, and insert them into the work
queue of the worker, and then wait for that the untreated items is below some
threshold value.
- When we want to insert a directory name index into b+ tree, we just add the
information into the delayed inserting rb-tree.
And then we check the number of the delayed items and do delayed items
balance. (The balance policy is above.)
- When we want to delete a directory name index from the b+ tree, we search it
in the inserting rb-tree at first. If we look it up, just drop it. If not,
add the key of it into the delayed deleting rb-tree.
Similar to the delayed inserting rb-tree, we also check the number of the
delayed items and do delayed items balance.
(The same to inserting manipulation)
- When we want to update the metadata of some inode, we cached the data of the
inode into the delayed node. the worker will flush it into the b+ tree after
dealing with the delayed insertion and deletion.
- We will move the delayed node to the tail of the list after we access the
delayed node, By this way, we can cache more delayed items and merge more
inode updates.
- If we want to commit transaction, we will deal with all the delayed node.
- the delayed node will be freed when we free the btrfs inode.
- Before we log the inode items, we commit all the directory name index items
and the delayed inode update.
I did a quick test by the benchmark tool[1] and found we can improve the
performance of file creation by ~15%, and file deletion by ~20%.
Before applying this patch:
Create files:
Total files: 50000
Total time: 1.096108
Average time: 0.000022
Delete files:
Total files: 50000
Total time: 1.510403
Average time: 0.000030
After applying this patch:
Create files:
Total files: 50000
Total time: 0.932899
Average time: 0.000019
Delete files:
Total files: 50000
Total time: 1.215732
Average time: 0.000024
[1] http://marc.info/?l=linux-btrfs&m=128212635122920&q=p3
Many thanks for Kitayama-san's help!
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Reviewed-by: David Sterba <dave@jikos.cz>
Tested-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Tested-by: Itaru Kitayama <kitayama@cl.bb4u.ne.jp>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
This adds an initial implementation for scrub. It works quite
straightforward. The usermode issues an ioctl for each device in the
fs. For each device, it enumerates the allocated device chunks. For
each chunk, the contained extents are enumerated and the data checksums
fetched. The extents are read sequentially and the checksums verified.
If an error occurs (checksum or EIO), a good copy is searched for. If
one is found, the bad copy will be rewritten.
All enumerations happen from the commit roots. During a transaction
commit, the scrubs get paused and afterwards continue from the new
roots.
This commit is based on the series originally posted to linux-btrfs
with some improvements that resulted from comments from David Sterba,
Ilya Dryomov and Jan Schmidt.
Signed-off-by: Arne Jansen <sensille@gmx.net>
Lzo is a much faster compression algorithm than gzib, so would allow
more users to enable transparent compression, and some users can
choose from compression ratio and speed for different applications
Usage:
# mount -t btrfs -o compress[=<zlib,lzo>] dev /mnt
or
# mount -t btrfs -o compress-force[=<zlib,lzo>] dev /mnt
"-o compress" without argument is still allowed for compatability.
Compatibility:
If we mount a filesystem with lzo compression, it will not be able be
mounted in old kernels. One reason is, otherwise btrfs will directly
dump compressed data, which sits in inline extent, to user.
Performance:
The test copied a linux source tarball (~400M) from an ext4 partition
to the btrfs partition, and then extracted it.
(time in second)
lzo zlib nocompress
copy: 10.6 21.7 14.9
extract: 70.1 94.4 66.6
(data size in MB)
lzo zlib nocompress
copy: 185.87 108.69 394.49
extract: 193.80 132.36 381.21
Changelog:
v1 -> v2:
- Select LZO_COMPRESS and LZO_DECOMPRESS in btrfs Kconfig.
- Add incompability flag.
- Fix error handling in compress code.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
This commit introduces a new kind of back reference for btrfs metadata.
Once a filesystem has been mounted with this commit, IT WILL NO LONGER
BE MOUNTABLE BY OLDER KERNELS.
When a tree block in subvolume tree is cow'd, the reference counts of all
extents it points to are increased by one. At transaction commit time,
the old root of the subvolume is recorded in a "dead root" data structure,
and the btree it points to is later walked, dropping reference counts
and freeing any blocks where the reference count goes to 0.
The increments done during cow and decrements done after commit cancel out,
and the walk is a very expensive way to go about freeing the blocks that
are no longer referenced by the new btree root. This commit reduces the
transaction overhead by avoiding the need for dead root records.
When a non-shared tree block is cow'd, we free the old block at once, and the
new block inherits old block's references. When a tree block with reference
count > 1 is cow'd, we increase the reference counts of all extents
the new block points to by one, and decrease the old block's reference count by
one.
This dead tree avoidance code removes the need to modify the reference
counts of lower level extents when a non-shared tree block is cow'd.
But we still need to update back ref for all pointers in the block.
This is because the location of the block is recorded in the back ref
item.
We can solve this by introducing a new type of back ref. The new
back ref provides information about pointer's key, level and in which
tree the pointer lives. This information allow us to find the pointer
by searching the tree. The shortcoming of the new back ref is that it
only works for pointers in tree blocks referenced by their owner trees.
This is mostly a problem for snapshots, where resolving one of these
fuzzy back references would be O(number_of_snapshots) and quite slow.
The solution used here is to use the fuzzy back references in the common
case where a given tree block is only referenced by one root,
and use the full back references when multiple roots have a reference
on a given block.
This commit adds per subvolume red-black tree to keep trace of cached
inodes. The red-black tree helps the balancing code to find cached
inodes whose inode numbers within a given range.
This commit improves the balancing code by introducing several data
structures to keep the state of balancing. The most important one
is the back ref cache. It caches how the upper level tree blocks are
referenced. This greatly reduce the overhead of checking back ref.
The improved balancing code scales significantly better with a large
number of snapshots.
This is a very large commit and was written in a number of
pieces. But, they depend heavily on the disk format change and were
squashed together to make sure git bisect didn't end up in a
bad state wrt space balancing or the format change.
Signed-off-by: Yan Zheng <zheng.yan@oracle.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Get rid of the hacks for building out of tree, and always use += for
assigning to the object lists.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
The extent allocation tree maintains a reference count and full
back reference information for every extent allocated in the
filesystem. For subvolume and snapshot trees, every time
a block goes through COW, the new copy of the block adds a reference
on every block it points to.
If a btree node points to 150 leaves, then the COW code needs to go
and add backrefs on 150 different extents, which might be spread all
over the extent allocation tree.
These updates currently happen during btrfs_cow_block, and most COWs
happen during btrfs_search_slot. btrfs_search_slot has locks held
on both the parent and the node we are COWing, and so we really want
to avoid IO during the COW if we can.
This commit adds an rbtree of pending reference count updates and extent
allocations. The tree is ordered by byte number of the extent and byte number
of the parent for the back reference. The tree allows us to:
1) Modify back references in something close to disk order, reducing seeks
2) Significantly reduce the number of modifications made as block pointers
are balanced around
3) Do all of the extent insertion and back reference modifications outside
of the performance critical btrfs_search_slot code.
#3 has the added benefit of greatly reducing the btrfs stack footprint.
The extent allocation tree modifications are done without the deep
(and somewhat recursive) call chains used in the past.
These delayed back reference updates must be done before the transaction
commits, and so the rbtree is tied to the transaction. Throttling is
implemented to help keep the queue of backrefs at a reasonable size.
Since there was a similar mechanism in place for the extent tree
extents, that is removed and replaced by the delayed reference tree.
Yan Zheng <yan.zheng@oracle.com> helped review and fixup this code.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
This is a large change for adding compression on reading and writing,
both for inline and regular extents. It does some fairly large
surgery to the writeback paths.
Compression is off by default and enabled by mount -o compress. Even
when the -o compress mount option is not used, it is possible to read
compressed extents off the disk.
If compression for a given set of pages fails to make them smaller, the
file is flagged to avoid future compression attempts later.
* While finding delalloc extents, the pages are locked before being sent down
to the delalloc handler. This allows the delalloc handler to do complex things
such as cleaning the pages, marking them writeback and starting IO on their
behalf.
* Inline extents are inserted at delalloc time now. This allows us to compress
the data before inserting the inline extent, and it allows us to insert
an inline extent that spans multiple pages.
* All of the in-memory extent representations (extent_map.c, ordered-data.c etc)
are changed to record both an in-memory size and an on disk size, as well
as a flag for compression.
From a disk format point of view, the extent pointers in the file are changed
to record the on disk size of a given extent and some encoding flags.
Space in the disk format is allocated for compression encoding, as well
as encryption and a generic 'other' field. Neither the encryption or the
'other' field are currently used.
In order to limit the amount of data read for a single random read in the
file, the size of a compressed extent is limited to 128k. This is a
software only limit, the disk format supports u64 sized compressed extents.
In order to limit the ram consumed while processing extents, the uncompressed
size of a compressed extent is limited to 256k. This is a software only limit
and will be subject to tuning later.
Checksumming is still done on compressed extents, and it is done on the
uncompressed version of the data. This way additional encodings can be
layered on without having to figure out which encoding to checksum.
Compression happens at delalloc time, which is basically singled threaded because
it is usually done by a single pdflush thread. This makes it tricky to
spread the compression load across all the cpus on the box. We'll have to
look at parallel pdflush walks of dirty inodes at a later time.
Decompression is hooked into readpages and it does spread across CPUs nicely.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
This fixes the btrfs makefile for building in the tree and out of the tree
both as a module and static.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
This improves the comments at the top of many functions. It didn't
dive into the guts of functions because I was trying to
avoid merging problems with the new allocator and back reference work.
extent-tree.c and volumes.c were both skipped, and there is definitely
more work todo in cleaning and commenting the code.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
btrfs had magic to put the chagneset id into a printk on module load.
This removes that from the Makefile and hardcodes the printk to print
"Btrfs"
Signed-off-by: Chris Mason <chris.mason@oracle.com>
1) replace the per fs_info extent_io_tree that tracked free space with two
rb-trees per block group to track free space areas via offset and size. The
reason to do this is because most allocations come with a hint byte where to
start, so we can usually find a chunk of free space at that hint byte to satisfy
the allocation and get good space packing. If we cannot find free space at or
after the given offset we fall back on looking for a chunk of the given size as
close to that given offset as possible. When we fall back on the size search we
also try to find a slot as close to the size we want as possible, to avoid
breaking small chunks off of huge areas if possible.
2) remove the extent_io_tree that tracked the block group cache from fs_info and
replaced it with an rb-tree thats tracks block group cache via offset. also
added a per space_info list that tracks the block group cache for the particular
space so we can lookup related block groups easily.
3) cleaned up the allocation code to make it a little easier to read and a
little less complicated. Basically there are 3 steps, first look from our
provided hint. If we couldn't find from that given hint, start back at our
original search start and look for space from there. If that fails try to
allocate space if we can and start looking again. If not we're screwed and need
to start over again.
4) small fixes. there were some issues in volumes.c where we wouldn't allocate
the rest of the disk. fixed cow_file_range to actually pass the alloc_hint,
which has helped a good bit in making the fs_mark test I run have semi-normal
results as we run out of space. Generally with data allocations we don't track
where we last allocated from, so everytime we did a data allocation we'd search
through every block group that we have looking for free space. Now searching a
block group with no free space isn't terribly time consuming, it was causing a
slight degradation as we got more data block groups. The alloc_hint has fixed
this slight degredation and made things semi-normal.
There is still one nagging problem I'm working on where we will get ENOSPC when
there is definitely plenty of space. This only happens with metadata
allocations, and only when we are almost full. So you generally hit the 85%
mark first, but sometimes you'll hit the BUG before you hit the 85% wall. I'm
still tracking it down, but until then this seems to be pretty stable and make a
significant performance gain.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
File syncs and directory syncs are optimized by copying their
items into a special (copy-on-write) log tree. There is one log tree per
subvolume and the btrfs super block points to a tree of log tree roots.
After a crash, items are copied out of the log tree and back into the
subvolume. See tree-log.c for all the details.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
This patch makes btrfs so it will compile properly when acls are disabled. I
tested this and it worked with CONFIG_FS_POSIX_ACL off and on.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Date: Tue, 19 Aug 2008 19:21:57 +0100
Using a 64-bit hash as the readdir cookie is just asking for trouble.
And gets it, when we try to export the file system by NFS.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Date: Mon, 21 Jul 2008 02:01:56 +0530
Here's an implementation of NFS support for btrfs. It relies on the
fixes which are going in to 2.6.28 for the NFS readdir/lookup deadlock.
This uses the btrfs_iget helper introduced previously.
[dwmw2: Tidy up a little, switch to d_obtain_alias() w/compat routine,
change fh_type, store parent's root object ID where needed,
fix some get_parent() and fs_to_dentry() bugs]
Signed-off-by: Balaji Rao <balajirrao@gmail.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Much of the IO done while dropping snapshots is done looking up
leaves in the filesystem trees to see if they point to any extents and
to drop the references on any extents found.
This creates a cache so that IO isn't required.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
The allocation trees and the chunk trees are serialized via their own
dedicated mutexes. This means allocation location is still not very
fine grained.
The main FS btree is protected by locks on each block in the btree. Locks
are taken top / down, and as processing finishes on a given level of the
tree, the lock is released after locking the lower level.
The end result of a search is now a path where only the lowest level
is locked. Releasing or freeing the path drops any locks held.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Split the ioctl handling out of inode.c into a file of it's own.
Also fix up checkpatch.pl warnings for the moved code.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Btrfs has been using workqueues to spread the checksumming load across
other CPUs in the system. But, workqueues only schedule work on the
same CPU that queued the work, giving them a limited benefit for systems with
higher CPU counts.
This code adds a generic facility to schedule work with pools of kthreads,
and changes the bio submission code to queue bios up. The queueing is
important to make sure large numbers of procs on the system don't
turn streaming workloads into random workloads by sending IO down
concurrently.
The end result of all of this is much higher performance (and CPU usage) when
doing checksumming on large machines. Two worker pools are created,
one for writes and one for endio processing. The two could deadlock if
we tried to service both from a single pool.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
use normal kbuild syntax to build acl.o conditinally and remove comment
out lines.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
There is now extent_map for mapping offsets in the file to disk and
extent_io for state tracking, IO submission and extent_bufers.
The new extent_map code shifts from [start,end] pairs to [start,len], and
pushes the locking out into the caller. This allows a few performance
optimizations and is easier to use.
A number of extent_map usage bugs were fixed, mostly with failing
to remove extent_map entries when changing the file.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
This forces file data extents down the disk along with the metadata that
references them. The current implementation is fairly simple, and just
writes out all of the dirty pages in an inode before the commit.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Omar Sandoval