Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
/*
|
|
|
|
* Copyright (C) 2008 Red Hat. All rights reserved.
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or
|
|
|
|
* modify it under the terms of the GNU General Public
|
|
|
|
* License v2 as published by the Free Software Foundation.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
|
* General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public
|
|
|
|
* License along with this program; if not, write to the
|
|
|
|
* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
|
|
|
|
* Boston, MA 021110-1307, USA.
|
|
|
|
*/
|
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
#include <linux/pagemap.h>
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
#include <linux/sched.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 16:04:11 +08:00
|
|
|
#include <linux/slab.h>
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
#include <linux/math64.h>
|
2011-08-08 20:24:46 +08:00
|
|
|
#include <linux/ratelimit.h>
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
#include "ctree.h"
|
2009-04-03 21:47:43 +08:00
|
|
|
#include "free-space-cache.h"
|
|
|
|
#include "transaction.h"
|
2010-06-22 02:48:16 +08:00
|
|
|
#include "disk-io.h"
|
2011-04-01 22:55:00 +08:00
|
|
|
#include "extent_io.h"
|
Btrfs: Cache free inode numbers in memory
Currently btrfs stores the highest objectid of the fs tree, and it always
returns (highest+1) inode number when we create a file, so inode numbers
won't be reclaimed when we delete files, so we'll run out of inode numbers
as we keep create/delete files in 32bits machines.
This fixes it, and it works similarly to how we cache free space in block
cgroups.
We start a kernel thread to read the file tree. By scanning inode items,
we know which chunks of inode numbers are free, and we cache them in
an rb-tree.
Because we are searching the commit root, we have to carefully handle the
cross-transaction case.
The rb-tree is a hybrid extent+bitmap tree, so if we have too many small
chunks of inode numbers, we'll use bitmaps. Initially we allow 16K ram
of extents, and a bitmap will be used if we exceed this threshold. The
extents threshold is adjusted in runtime.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-04-20 10:06:11 +08:00
|
|
|
#include "inode-map.h"
|
Btrfs: fix race between fs trimming and block group remove/allocation
Our fs trim operation, which is completely transactionless (doesn't start
or joins an existing transaction) consists of visiting all block groups
and then for each one to iterate its free space entries and perform a
discard operation against the space range represented by the free space
entries. However before performing a discard, the corresponding free space
entry is removed from the free space rbtree, and when the discard completes
it is added back to the free space rbtree.
If a block group remove operation happens while the discard is ongoing (or
before it starts and after a free space entry is hidden), we end up not
waiting for the discard to complete, remove the extent map that maps
logical address to physical addresses and the corresponding chunk metadata
from the the chunk and device trees. After that and before the discard
completes, the current running transaction can finish and a new one start,
allowing for new block groups that map to the same physical addresses to
be allocated and written to.
So fix this by keeping the extent map in memory until the discard completes
so that the same physical addresses aren't reused before it completes.
If the physical locations that are under a discard operation end up being
used for a new metadata block group for example, and dirty metadata extents
are written before the discard finishes (the VM might call writepages() of
our btree inode's i_mapping for example, or an fsync log commit happens) we
end up overwriting metadata with zeroes, which leads to errors from fsck
like the following:
checking extents
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
read block failed check_tree_block
owner ref check failed [833912832 16384]
Errors found in extent allocation tree or chunk allocation
checking free space cache
checking fs roots
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
read block failed check_tree_block
root 5 root dir 256 error
root 5 inode 260 errors 2001, no inode item, link count wrong
unresolved ref dir 256 index 0 namelen 8 name foobar_3 filetype 1 errors 6, no dir index, no inode ref
root 5 inode 262 errors 2001, no inode item, link count wrong
unresolved ref dir 256 index 0 namelen 8 name foobar_5 filetype 1 errors 6, no dir index, no inode ref
root 5 inode 263 errors 2001, no inode item, link count wrong
(...)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-11-28 05:14:15 +08:00
|
|
|
#include "volumes.h"
|
2009-04-03 21:47:43 +08:00
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
#define BITS_PER_BITMAP (PAGE_CACHE_SIZE * 8)
|
2015-12-15 00:42:10 +08:00
|
|
|
#define MAX_CACHE_BYTES_PER_GIG SZ_32K
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
struct btrfs_trim_range {
|
|
|
|
u64 start;
|
|
|
|
u64 bytes;
|
|
|
|
struct list_head list;
|
|
|
|
};
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
static int link_free_space(struct btrfs_free_space_ctl *ctl,
|
2010-07-03 00:14:14 +08:00
|
|
|
struct btrfs_free_space *info);
|
2012-05-14 22:06:40 +08:00
|
|
|
static void unlink_free_space(struct btrfs_free_space_ctl *ctl,
|
|
|
|
struct btrfs_free_space *info);
|
2010-07-03 00:14:14 +08:00
|
|
|
|
2011-04-20 10:20:14 +08:00
|
|
|
static struct inode *__lookup_free_space_inode(struct btrfs_root *root,
|
|
|
|
struct btrfs_path *path,
|
|
|
|
u64 offset)
|
2010-06-22 02:48:16 +08:00
|
|
|
{
|
|
|
|
struct btrfs_key key;
|
|
|
|
struct btrfs_key location;
|
|
|
|
struct btrfs_disk_key disk_key;
|
|
|
|
struct btrfs_free_space_header *header;
|
|
|
|
struct extent_buffer *leaf;
|
|
|
|
struct inode *inode = NULL;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
key.objectid = BTRFS_FREE_SPACE_OBJECTID;
|
2011-04-20 10:20:14 +08:00
|
|
|
key.offset = offset;
|
2010-06-22 02:48:16 +08:00
|
|
|
key.type = 0;
|
|
|
|
|
|
|
|
ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
|
|
|
|
if (ret < 0)
|
|
|
|
return ERR_PTR(ret);
|
|
|
|
if (ret > 0) {
|
2011-04-21 07:20:15 +08:00
|
|
|
btrfs_release_path(path);
|
2010-06-22 02:48:16 +08:00
|
|
|
return ERR_PTR(-ENOENT);
|
|
|
|
}
|
|
|
|
|
|
|
|
leaf = path->nodes[0];
|
|
|
|
header = btrfs_item_ptr(leaf, path->slots[0],
|
|
|
|
struct btrfs_free_space_header);
|
|
|
|
btrfs_free_space_key(leaf, header, &disk_key);
|
|
|
|
btrfs_disk_key_to_cpu(&location, &disk_key);
|
2011-04-21 07:20:15 +08:00
|
|
|
btrfs_release_path(path);
|
2010-06-22 02:48:16 +08:00
|
|
|
|
|
|
|
inode = btrfs_iget(root->fs_info->sb, &location, root, NULL);
|
|
|
|
if (!inode)
|
|
|
|
return ERR_PTR(-ENOENT);
|
|
|
|
if (IS_ERR(inode))
|
|
|
|
return inode;
|
|
|
|
if (is_bad_inode(inode)) {
|
|
|
|
iput(inode);
|
|
|
|
return ERR_PTR(-ENOENT);
|
|
|
|
}
|
|
|
|
|
2012-04-13 23:03:55 +08:00
|
|
|
mapping_set_gfp_mask(inode->i_mapping,
|
2015-11-07 08:28:49 +08:00
|
|
|
mapping_gfp_constraint(inode->i_mapping,
|
|
|
|
~(__GFP_FS | __GFP_HIGHMEM)));
|
2011-03-31 17:43:23 +08:00
|
|
|
|
2011-04-20 10:20:14 +08:00
|
|
|
return inode;
|
|
|
|
}
|
|
|
|
|
|
|
|
struct inode *lookup_free_space_inode(struct btrfs_root *root,
|
|
|
|
struct btrfs_block_group_cache
|
|
|
|
*block_group, struct btrfs_path *path)
|
|
|
|
{
|
|
|
|
struct inode *inode = NULL;
|
2011-10-06 20:58:24 +08:00
|
|
|
u32 flags = BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW;
|
2011-04-20 10:20:14 +08:00
|
|
|
|
|
|
|
spin_lock(&block_group->lock);
|
|
|
|
if (block_group->inode)
|
|
|
|
inode = igrab(block_group->inode);
|
|
|
|
spin_unlock(&block_group->lock);
|
|
|
|
if (inode)
|
|
|
|
return inode;
|
|
|
|
|
|
|
|
inode = __lookup_free_space_inode(root, path,
|
|
|
|
block_group->key.objectid);
|
|
|
|
if (IS_ERR(inode))
|
|
|
|
return inode;
|
|
|
|
|
2010-06-22 02:48:16 +08:00
|
|
|
spin_lock(&block_group->lock);
|
2011-10-06 20:58:24 +08:00
|
|
|
if (!((BTRFS_I(inode)->flags & flags) == flags)) {
|
2013-03-20 06:41:23 +08:00
|
|
|
btrfs_info(root->fs_info,
|
|
|
|
"Old style space inode found, converting.");
|
2011-10-06 20:58:24 +08:00
|
|
|
BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM |
|
|
|
|
BTRFS_INODE_NODATACOW;
|
2011-06-11 03:31:13 +08:00
|
|
|
block_group->disk_cache_state = BTRFS_DC_CLEAR;
|
|
|
|
}
|
|
|
|
|
2011-08-30 02:06:00 +08:00
|
|
|
if (!block_group->iref) {
|
2010-06-22 02:48:16 +08:00
|
|
|
block_group->inode = igrab(inode);
|
|
|
|
block_group->iref = 1;
|
|
|
|
}
|
|
|
|
spin_unlock(&block_group->lock);
|
|
|
|
|
|
|
|
return inode;
|
|
|
|
}
|
|
|
|
|
2013-04-26 04:41:01 +08:00
|
|
|
static int __create_free_space_inode(struct btrfs_root *root,
|
|
|
|
struct btrfs_trans_handle *trans,
|
|
|
|
struct btrfs_path *path,
|
|
|
|
u64 ino, u64 offset)
|
2010-06-22 02:48:16 +08:00
|
|
|
{
|
|
|
|
struct btrfs_key key;
|
|
|
|
struct btrfs_disk_key disk_key;
|
|
|
|
struct btrfs_free_space_header *header;
|
|
|
|
struct btrfs_inode_item *inode_item;
|
|
|
|
struct extent_buffer *leaf;
|
2011-10-06 20:58:24 +08:00
|
|
|
u64 flags = BTRFS_INODE_NOCOMPRESS | BTRFS_INODE_PREALLOC;
|
2010-06-22 02:48:16 +08:00
|
|
|
int ret;
|
|
|
|
|
2011-04-20 10:20:14 +08:00
|
|
|
ret = btrfs_insert_empty_inode(trans, root, path, ino);
|
2010-06-22 02:48:16 +08:00
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
2011-10-06 20:58:24 +08:00
|
|
|
/* We inline crc's for the free disk space cache */
|
|
|
|
if (ino != BTRFS_FREE_INO_OBJECTID)
|
|
|
|
flags |= BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW;
|
|
|
|
|
2010-06-22 02:48:16 +08:00
|
|
|
leaf = path->nodes[0];
|
|
|
|
inode_item = btrfs_item_ptr(leaf, path->slots[0],
|
|
|
|
struct btrfs_inode_item);
|
|
|
|
btrfs_item_key(leaf, &disk_key, path->slots[0]);
|
|
|
|
memset_extent_buffer(leaf, 0, (unsigned long)inode_item,
|
|
|
|
sizeof(*inode_item));
|
|
|
|
btrfs_set_inode_generation(leaf, inode_item, trans->transid);
|
|
|
|
btrfs_set_inode_size(leaf, inode_item, 0);
|
|
|
|
btrfs_set_inode_nbytes(leaf, inode_item, 0);
|
|
|
|
btrfs_set_inode_uid(leaf, inode_item, 0);
|
|
|
|
btrfs_set_inode_gid(leaf, inode_item, 0);
|
|
|
|
btrfs_set_inode_mode(leaf, inode_item, S_IFREG | 0600);
|
2011-10-06 20:58:24 +08:00
|
|
|
btrfs_set_inode_flags(leaf, inode_item, flags);
|
2010-06-22 02:48:16 +08:00
|
|
|
btrfs_set_inode_nlink(leaf, inode_item, 1);
|
|
|
|
btrfs_set_inode_transid(leaf, inode_item, trans->transid);
|
2011-04-20 10:20:14 +08:00
|
|
|
btrfs_set_inode_block_group(leaf, inode_item, offset);
|
2010-06-22 02:48:16 +08:00
|
|
|
btrfs_mark_buffer_dirty(leaf);
|
2011-04-21 07:20:15 +08:00
|
|
|
btrfs_release_path(path);
|
2010-06-22 02:48:16 +08:00
|
|
|
|
|
|
|
key.objectid = BTRFS_FREE_SPACE_OBJECTID;
|
2011-04-20 10:20:14 +08:00
|
|
|
key.offset = offset;
|
2010-06-22 02:48:16 +08:00
|
|
|
key.type = 0;
|
|
|
|
ret = btrfs_insert_empty_item(trans, root, path, &key,
|
|
|
|
sizeof(struct btrfs_free_space_header));
|
|
|
|
if (ret < 0) {
|
2011-04-21 07:20:15 +08:00
|
|
|
btrfs_release_path(path);
|
2010-06-22 02:48:16 +08:00
|
|
|
return ret;
|
|
|
|
}
|
2015-04-05 08:14:42 +08:00
|
|
|
|
2010-06-22 02:48:16 +08:00
|
|
|
leaf = path->nodes[0];
|
|
|
|
header = btrfs_item_ptr(leaf, path->slots[0],
|
|
|
|
struct btrfs_free_space_header);
|
|
|
|
memset_extent_buffer(leaf, 0, (unsigned long)header, sizeof(*header));
|
|
|
|
btrfs_set_free_space_key(leaf, header, &disk_key);
|
|
|
|
btrfs_mark_buffer_dirty(leaf);
|
2011-04-21 07:20:15 +08:00
|
|
|
btrfs_release_path(path);
|
2010-06-22 02:48:16 +08:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-04-20 10:20:14 +08:00
|
|
|
int create_free_space_inode(struct btrfs_root *root,
|
|
|
|
struct btrfs_trans_handle *trans,
|
|
|
|
struct btrfs_block_group_cache *block_group,
|
|
|
|
struct btrfs_path *path)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
u64 ino;
|
|
|
|
|
|
|
|
ret = btrfs_find_free_objectid(root, &ino);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
return __create_free_space_inode(root, trans, path, ino,
|
|
|
|
block_group->key.objectid);
|
|
|
|
}
|
|
|
|
|
2013-05-13 21:55:09 +08:00
|
|
|
int btrfs_check_trunc_cache_free_space(struct btrfs_root *root,
|
|
|
|
struct btrfs_block_rsv *rsv)
|
2010-06-22 02:48:16 +08:00
|
|
|
{
|
2011-11-02 21:29:35 +08:00
|
|
|
u64 needed_bytes;
|
2013-05-13 21:55:09 +08:00
|
|
|
int ret;
|
2011-11-02 21:29:35 +08:00
|
|
|
|
|
|
|
/* 1 for slack space, 1 for updating the inode */
|
|
|
|
needed_bytes = btrfs_calc_trunc_metadata_size(root, 1) +
|
|
|
|
btrfs_calc_trans_metadata_size(root, 1);
|
|
|
|
|
2013-05-13 21:55:09 +08:00
|
|
|
spin_lock(&rsv->lock);
|
|
|
|
if (rsv->reserved < needed_bytes)
|
|
|
|
ret = -ENOSPC;
|
|
|
|
else
|
|
|
|
ret = 0;
|
|
|
|
spin_unlock(&rsv->lock);
|
2013-05-21 10:39:21 +08:00
|
|
|
return ret;
|
2013-05-13 21:55:09 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
int btrfs_truncate_free_space_cache(struct btrfs_root *root,
|
|
|
|
struct btrfs_trans_handle *trans,
|
2015-04-07 03:46:08 +08:00
|
|
|
struct btrfs_block_group_cache *block_group,
|
2013-05-13 21:55:09 +08:00
|
|
|
struct inode *inode)
|
|
|
|
{
|
|
|
|
int ret = 0;
|
2015-04-07 03:46:08 +08:00
|
|
|
struct btrfs_path *path = btrfs_alloc_path();
|
2015-05-01 00:47:05 +08:00
|
|
|
bool locked = false;
|
2015-04-07 03:46:08 +08:00
|
|
|
|
|
|
|
if (!path) {
|
|
|
|
ret = -ENOMEM;
|
|
|
|
goto fail;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (block_group) {
|
2015-05-01 00:47:05 +08:00
|
|
|
locked = true;
|
2015-04-07 03:46:08 +08:00
|
|
|
mutex_lock(&trans->transaction->cache_write_mutex);
|
|
|
|
if (!list_empty(&block_group->io_list)) {
|
|
|
|
list_del_init(&block_group->io_list);
|
|
|
|
|
|
|
|
btrfs_wait_cache_io(root, trans, block_group,
|
|
|
|
&block_group->io_ctl, path,
|
|
|
|
block_group->key.objectid);
|
|
|
|
btrfs_put_block_group(block_group);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* now that we've truncated the cache away, its no longer
|
|
|
|
* setup or written
|
|
|
|
*/
|
|
|
|
spin_lock(&block_group->lock);
|
|
|
|
block_group->disk_cache_state = BTRFS_DC_CLEAR;
|
|
|
|
spin_unlock(&block_group->lock);
|
|
|
|
}
|
|
|
|
btrfs_free_path(path);
|
2010-06-22 02:48:16 +08:00
|
|
|
|
|
|
|
btrfs_i_size_write(inode, 0);
|
2013-09-13 06:13:56 +08:00
|
|
|
truncate_pagecache(inode, 0);
|
2010-06-22 02:48:16 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* We don't need an orphan item because truncating the free space cache
|
|
|
|
* will never be split across transactions.
|
2014-12-18 01:41:04 +08:00
|
|
|
* We don't need to check for -EAGAIN because we're a free space
|
|
|
|
* cache inode
|
2010-06-22 02:48:16 +08:00
|
|
|
*/
|
|
|
|
ret = btrfs_truncate_inode_items(trans, root, inode,
|
|
|
|
0, BTRFS_EXTENT_DATA_KEY);
|
2015-05-01 00:47:05 +08:00
|
|
|
if (ret)
|
|
|
|
goto fail;
|
2010-06-22 02:48:16 +08:00
|
|
|
|
2011-04-20 10:33:24 +08:00
|
|
|
ret = btrfs_update_inode(trans, root, inode);
|
2015-04-07 03:46:08 +08:00
|
|
|
|
|
|
|
fail:
|
2015-05-01 00:47:05 +08:00
|
|
|
if (locked)
|
|
|
|
mutex_unlock(&trans->transaction->cache_write_mutex);
|
2012-03-12 23:03:00 +08:00
|
|
|
if (ret)
|
|
|
|
btrfs_abort_transaction(trans, root, ret);
|
2011-11-02 21:29:35 +08:00
|
|
|
|
2011-04-20 10:33:24 +08:00
|
|
|
return ret;
|
2010-06-22 02:48:16 +08:00
|
|
|
}
|
|
|
|
|
2010-08-26 04:54:15 +08:00
|
|
|
static int readahead_cache(struct inode *inode)
|
|
|
|
{
|
|
|
|
struct file_ra_state *ra;
|
|
|
|
unsigned long last_index;
|
|
|
|
|
|
|
|
ra = kzalloc(sizeof(*ra), GFP_NOFS);
|
|
|
|
if (!ra)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
file_ra_state_init(ra, inode->i_mapping);
|
|
|
|
last_index = (i_size_read(inode) - 1) >> PAGE_CACHE_SHIFT;
|
|
|
|
|
|
|
|
page_cache_sync_readahead(inode->i_mapping, ra, NULL, 0, last_index);
|
|
|
|
|
|
|
|
kfree(ra);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static int io_ctl_init(struct btrfs_io_ctl *io_ctl, struct inode *inode,
|
2014-06-19 10:42:49 +08:00
|
|
|
struct btrfs_root *root, int write)
|
2011-10-06 03:18:58 +08:00
|
|
|
{
|
2014-06-19 10:42:49 +08:00
|
|
|
int num_pages;
|
|
|
|
int check_crcs = 0;
|
|
|
|
|
2014-06-05 07:59:57 +08:00
|
|
|
num_pages = DIV_ROUND_UP(i_size_read(inode), PAGE_CACHE_SIZE);
|
2014-06-19 10:42:49 +08:00
|
|
|
|
|
|
|
if (btrfs_ino(inode) != BTRFS_FREE_INO_OBJECTID)
|
|
|
|
check_crcs = 1;
|
|
|
|
|
|
|
|
/* Make sure we can fit our crcs into the first page */
|
|
|
|
if (write && check_crcs &&
|
|
|
|
(num_pages * sizeof(u32)) >= PAGE_CACHE_SIZE)
|
|
|
|
return -ENOSPC;
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
memset(io_ctl, 0, sizeof(struct btrfs_io_ctl));
|
2014-06-19 10:42:49 +08:00
|
|
|
|
2015-02-21 01:00:26 +08:00
|
|
|
io_ctl->pages = kcalloc(num_pages, sizeof(struct page *), GFP_NOFS);
|
2011-10-06 03:18:58 +08:00
|
|
|
if (!io_ctl->pages)
|
|
|
|
return -ENOMEM;
|
2014-06-19 10:42:49 +08:00
|
|
|
|
|
|
|
io_ctl->num_pages = num_pages;
|
2011-10-06 03:18:58 +08:00
|
|
|
io_ctl->root = root;
|
2014-06-19 10:42:49 +08:00
|
|
|
io_ctl->check_crcs = check_crcs;
|
2015-04-05 08:14:42 +08:00
|
|
|
io_ctl->inode = inode;
|
2014-06-19 10:42:49 +08:00
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static void io_ctl_free(struct btrfs_io_ctl *io_ctl)
|
2011-10-06 03:18:58 +08:00
|
|
|
{
|
|
|
|
kfree(io_ctl->pages);
|
2015-04-05 08:14:42 +08:00
|
|
|
io_ctl->pages = NULL;
|
2011-10-06 03:18:58 +08:00
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static void io_ctl_unmap_page(struct btrfs_io_ctl *io_ctl)
|
2011-10-06 03:18:58 +08:00
|
|
|
{
|
|
|
|
if (io_ctl->cur) {
|
|
|
|
io_ctl->cur = NULL;
|
|
|
|
io_ctl->orig = NULL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static void io_ctl_map_page(struct btrfs_io_ctl *io_ctl, int clear)
|
2011-10-06 03:18:58 +08:00
|
|
|
{
|
2013-08-27 05:14:08 +08:00
|
|
|
ASSERT(io_ctl->index < io_ctl->num_pages);
|
2011-10-06 03:18:58 +08:00
|
|
|
io_ctl->page = io_ctl->pages[io_ctl->index++];
|
2015-04-06 22:48:20 +08:00
|
|
|
io_ctl->cur = page_address(io_ctl->page);
|
2011-10-06 03:18:58 +08:00
|
|
|
io_ctl->orig = io_ctl->cur;
|
|
|
|
io_ctl->size = PAGE_CACHE_SIZE;
|
|
|
|
if (clear)
|
|
|
|
memset(io_ctl->cur, 0, PAGE_CACHE_SIZE);
|
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static void io_ctl_drop_pages(struct btrfs_io_ctl *io_ctl)
|
2011-10-06 03:18:58 +08:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
io_ctl_unmap_page(io_ctl);
|
|
|
|
|
|
|
|
for (i = 0; i < io_ctl->num_pages; i++) {
|
2012-01-09 14:27:42 +08:00
|
|
|
if (io_ctl->pages[i]) {
|
|
|
|
ClearPageChecked(io_ctl->pages[i]);
|
|
|
|
unlock_page(io_ctl->pages[i]);
|
|
|
|
page_cache_release(io_ctl->pages[i]);
|
|
|
|
}
|
2011-10-06 03:18:58 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static int io_ctl_prepare_pages(struct btrfs_io_ctl *io_ctl, struct inode *inode,
|
2011-10-06 03:18:58 +08:00
|
|
|
int uptodate)
|
|
|
|
{
|
|
|
|
struct page *page;
|
|
|
|
gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping);
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < io_ctl->num_pages; i++) {
|
|
|
|
page = find_or_create_page(inode->i_mapping, i, mask);
|
|
|
|
if (!page) {
|
|
|
|
io_ctl_drop_pages(io_ctl);
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
io_ctl->pages[i] = page;
|
|
|
|
if (uptodate && !PageUptodate(page)) {
|
|
|
|
btrfs_readpage(NULL, page);
|
|
|
|
lock_page(page);
|
|
|
|
if (!PageUptodate(page)) {
|
2013-12-21 00:37:06 +08:00
|
|
|
btrfs_err(BTRFS_I(inode)->root->fs_info,
|
|
|
|
"error reading free space cache");
|
2011-10-06 03:18:58 +08:00
|
|
|
io_ctl_drop_pages(io_ctl);
|
|
|
|
return -EIO;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2011-11-15 22:31:24 +08:00
|
|
|
for (i = 0; i < io_ctl->num_pages; i++) {
|
|
|
|
clear_page_dirty_for_io(io_ctl->pages[i]);
|
|
|
|
set_page_extent_mapped(io_ctl->pages[i]);
|
|
|
|
}
|
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static void io_ctl_set_generation(struct btrfs_io_ctl *io_ctl, u64 generation)
|
2011-10-06 03:18:58 +08:00
|
|
|
{
|
2012-04-13 23:03:55 +08:00
|
|
|
__le64 *val;
|
2011-10-06 03:18:58 +08:00
|
|
|
|
|
|
|
io_ctl_map_page(io_ctl, 1);
|
|
|
|
|
|
|
|
/*
|
2011-10-06 20:58:24 +08:00
|
|
|
* Skip the csum areas. If we don't check crcs then we just have a
|
|
|
|
* 64bit chunk at the front of the first page.
|
2011-10-06 03:18:58 +08:00
|
|
|
*/
|
2011-10-06 20:58:24 +08:00
|
|
|
if (io_ctl->check_crcs) {
|
|
|
|
io_ctl->cur += (sizeof(u32) * io_ctl->num_pages);
|
|
|
|
io_ctl->size -= sizeof(u64) + (sizeof(u32) * io_ctl->num_pages);
|
|
|
|
} else {
|
|
|
|
io_ctl->cur += sizeof(u64);
|
|
|
|
io_ctl->size -= sizeof(u64) * 2;
|
|
|
|
}
|
2011-10-06 03:18:58 +08:00
|
|
|
|
|
|
|
val = io_ctl->cur;
|
|
|
|
*val = cpu_to_le64(generation);
|
|
|
|
io_ctl->cur += sizeof(u64);
|
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static int io_ctl_check_generation(struct btrfs_io_ctl *io_ctl, u64 generation)
|
2011-10-06 03:18:58 +08:00
|
|
|
{
|
2012-04-13 23:03:55 +08:00
|
|
|
__le64 *gen;
|
2011-10-06 03:18:58 +08:00
|
|
|
|
2011-10-06 20:58:24 +08:00
|
|
|
/*
|
|
|
|
* Skip the crc area. If we don't check crcs then we just have a 64bit
|
|
|
|
* chunk at the front of the first page.
|
|
|
|
*/
|
|
|
|
if (io_ctl->check_crcs) {
|
|
|
|
io_ctl->cur += sizeof(u32) * io_ctl->num_pages;
|
|
|
|
io_ctl->size -= sizeof(u64) +
|
|
|
|
(sizeof(u32) * io_ctl->num_pages);
|
|
|
|
} else {
|
|
|
|
io_ctl->cur += sizeof(u64);
|
|
|
|
io_ctl->size -= sizeof(u64) * 2;
|
|
|
|
}
|
2011-10-06 03:18:58 +08:00
|
|
|
|
|
|
|
gen = io_ctl->cur;
|
|
|
|
if (le64_to_cpu(*gen) != generation) {
|
2015-10-08 17:01:36 +08:00
|
|
|
btrfs_err_rl(io_ctl->root->fs_info,
|
|
|
|
"space cache generation (%llu) does not match inode (%llu)",
|
|
|
|
*gen, generation);
|
2011-10-06 03:18:58 +08:00
|
|
|
io_ctl_unmap_page(io_ctl);
|
|
|
|
return -EIO;
|
|
|
|
}
|
|
|
|
io_ctl->cur += sizeof(u64);
|
2011-10-06 20:58:24 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static void io_ctl_set_crc(struct btrfs_io_ctl *io_ctl, int index)
|
2011-10-06 20:58:24 +08:00
|
|
|
{
|
|
|
|
u32 *tmp;
|
|
|
|
u32 crc = ~(u32)0;
|
|
|
|
unsigned offset = 0;
|
|
|
|
|
|
|
|
if (!io_ctl->check_crcs) {
|
|
|
|
io_ctl_unmap_page(io_ctl);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (index == 0)
|
2011-11-22 00:43:28 +08:00
|
|
|
offset = sizeof(u32) * io_ctl->num_pages;
|
2011-10-06 20:58:24 +08:00
|
|
|
|
2013-03-14 22:57:45 +08:00
|
|
|
crc = btrfs_csum_data(io_ctl->orig + offset, crc,
|
2011-10-06 20:58:24 +08:00
|
|
|
PAGE_CACHE_SIZE - offset);
|
|
|
|
btrfs_csum_final(crc, (char *)&crc);
|
|
|
|
io_ctl_unmap_page(io_ctl);
|
2015-04-06 22:48:20 +08:00
|
|
|
tmp = page_address(io_ctl->pages[0]);
|
2011-10-06 20:58:24 +08:00
|
|
|
tmp += index;
|
|
|
|
*tmp = crc;
|
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static int io_ctl_check_crc(struct btrfs_io_ctl *io_ctl, int index)
|
2011-10-06 20:58:24 +08:00
|
|
|
{
|
|
|
|
u32 *tmp, val;
|
|
|
|
u32 crc = ~(u32)0;
|
|
|
|
unsigned offset = 0;
|
|
|
|
|
|
|
|
if (!io_ctl->check_crcs) {
|
|
|
|
io_ctl_map_page(io_ctl, 0);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (index == 0)
|
|
|
|
offset = sizeof(u32) * io_ctl->num_pages;
|
|
|
|
|
2015-04-06 22:48:20 +08:00
|
|
|
tmp = page_address(io_ctl->pages[0]);
|
2011-10-06 20:58:24 +08:00
|
|
|
tmp += index;
|
|
|
|
val = *tmp;
|
|
|
|
|
|
|
|
io_ctl_map_page(io_ctl, 0);
|
2013-03-14 22:57:45 +08:00
|
|
|
crc = btrfs_csum_data(io_ctl->orig + offset, crc,
|
2011-10-06 20:58:24 +08:00
|
|
|
PAGE_CACHE_SIZE - offset);
|
|
|
|
btrfs_csum_final(crc, (char *)&crc);
|
|
|
|
if (val != crc) {
|
2015-10-08 17:01:36 +08:00
|
|
|
btrfs_err_rl(io_ctl->root->fs_info,
|
|
|
|
"csum mismatch on free space cache");
|
2011-10-06 20:58:24 +08:00
|
|
|
io_ctl_unmap_page(io_ctl);
|
|
|
|
return -EIO;
|
|
|
|
}
|
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static int io_ctl_add_entry(struct btrfs_io_ctl *io_ctl, u64 offset, u64 bytes,
|
2011-10-06 03:18:58 +08:00
|
|
|
void *bitmap)
|
|
|
|
{
|
|
|
|
struct btrfs_free_space_entry *entry;
|
|
|
|
|
|
|
|
if (!io_ctl->cur)
|
|
|
|
return -ENOSPC;
|
|
|
|
|
|
|
|
entry = io_ctl->cur;
|
|
|
|
entry->offset = cpu_to_le64(offset);
|
|
|
|
entry->bytes = cpu_to_le64(bytes);
|
|
|
|
entry->type = (bitmap) ? BTRFS_FREE_SPACE_BITMAP :
|
|
|
|
BTRFS_FREE_SPACE_EXTENT;
|
|
|
|
io_ctl->cur += sizeof(struct btrfs_free_space_entry);
|
|
|
|
io_ctl->size -= sizeof(struct btrfs_free_space_entry);
|
|
|
|
|
|
|
|
if (io_ctl->size >= sizeof(struct btrfs_free_space_entry))
|
|
|
|
return 0;
|
|
|
|
|
2011-10-06 20:58:24 +08:00
|
|
|
io_ctl_set_crc(io_ctl, io_ctl->index - 1);
|
2011-10-06 03:18:58 +08:00
|
|
|
|
|
|
|
/* No more pages to map */
|
|
|
|
if (io_ctl->index >= io_ctl->num_pages)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/* map the next page */
|
|
|
|
io_ctl_map_page(io_ctl, 1);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static int io_ctl_add_bitmap(struct btrfs_io_ctl *io_ctl, void *bitmap)
|
2011-10-06 03:18:58 +08:00
|
|
|
{
|
|
|
|
if (!io_ctl->cur)
|
|
|
|
return -ENOSPC;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If we aren't at the start of the current page, unmap this one and
|
|
|
|
* map the next one if there is any left.
|
|
|
|
*/
|
|
|
|
if (io_ctl->cur != io_ctl->orig) {
|
2011-10-06 20:58:24 +08:00
|
|
|
io_ctl_set_crc(io_ctl, io_ctl->index - 1);
|
2011-10-06 03:18:58 +08:00
|
|
|
if (io_ctl->index >= io_ctl->num_pages)
|
|
|
|
return -ENOSPC;
|
|
|
|
io_ctl_map_page(io_ctl, 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
memcpy(io_ctl->cur, bitmap, PAGE_CACHE_SIZE);
|
2011-10-06 20:58:24 +08:00
|
|
|
io_ctl_set_crc(io_ctl, io_ctl->index - 1);
|
2011-10-06 03:18:58 +08:00
|
|
|
if (io_ctl->index < io_ctl->num_pages)
|
|
|
|
io_ctl_map_page(io_ctl, 0);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static void io_ctl_zero_remaining_pages(struct btrfs_io_ctl *io_ctl)
|
2011-10-06 03:18:58 +08:00
|
|
|
{
|
2011-10-06 20:58:24 +08:00
|
|
|
/*
|
|
|
|
* If we're not on the boundary we know we've modified the page and we
|
|
|
|
* need to crc the page.
|
|
|
|
*/
|
|
|
|
if (io_ctl->cur != io_ctl->orig)
|
|
|
|
io_ctl_set_crc(io_ctl, io_ctl->index - 1);
|
|
|
|
else
|
|
|
|
io_ctl_unmap_page(io_ctl);
|
2011-10-06 03:18:58 +08:00
|
|
|
|
|
|
|
while (io_ctl->index < io_ctl->num_pages) {
|
|
|
|
io_ctl_map_page(io_ctl, 1);
|
2011-10-06 20:58:24 +08:00
|
|
|
io_ctl_set_crc(io_ctl, io_ctl->index - 1);
|
2011-10-06 03:18:58 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static int io_ctl_read_entry(struct btrfs_io_ctl *io_ctl,
|
2011-10-06 20:58:24 +08:00
|
|
|
struct btrfs_free_space *entry, u8 *type)
|
2011-10-06 03:18:58 +08:00
|
|
|
{
|
|
|
|
struct btrfs_free_space_entry *e;
|
2011-11-11 09:45:05 +08:00
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (!io_ctl->cur) {
|
|
|
|
ret = io_ctl_check_crc(io_ctl, io_ctl->index);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
}
|
2011-10-06 03:18:58 +08:00
|
|
|
|
|
|
|
e = io_ctl->cur;
|
|
|
|
entry->offset = le64_to_cpu(e->offset);
|
|
|
|
entry->bytes = le64_to_cpu(e->bytes);
|
2011-10-06 20:58:24 +08:00
|
|
|
*type = e->type;
|
2011-10-06 03:18:58 +08:00
|
|
|
io_ctl->cur += sizeof(struct btrfs_free_space_entry);
|
|
|
|
io_ctl->size -= sizeof(struct btrfs_free_space_entry);
|
|
|
|
|
|
|
|
if (io_ctl->size >= sizeof(struct btrfs_free_space_entry))
|
2011-10-06 20:58:24 +08:00
|
|
|
return 0;
|
2011-10-06 03:18:58 +08:00
|
|
|
|
|
|
|
io_ctl_unmap_page(io_ctl);
|
|
|
|
|
2011-11-11 09:45:05 +08:00
|
|
|
return 0;
|
2011-10-06 03:18:58 +08:00
|
|
|
}
|
|
|
|
|
2015-04-07 04:17:20 +08:00
|
|
|
static int io_ctl_read_bitmap(struct btrfs_io_ctl *io_ctl,
|
2011-10-06 20:58:24 +08:00
|
|
|
struct btrfs_free_space *entry)
|
2011-10-06 03:18:58 +08:00
|
|
|
{
|
2011-10-06 20:58:24 +08:00
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = io_ctl_check_crc(io_ctl, io_ctl->index);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
memcpy(entry->bitmap, io_ctl->cur, PAGE_CACHE_SIZE);
|
|
|
|
io_ctl_unmap_page(io_ctl);
|
2011-10-06 20:58:24 +08:00
|
|
|
|
|
|
|
return 0;
|
2011-10-06 03:18:58 +08:00
|
|
|
}
|
|
|
|
|
2012-05-14 22:06:40 +08:00
|
|
|
/*
|
|
|
|
* Since we attach pinned extents after the fact we can have contiguous sections
|
|
|
|
* of free space that are split up in entries. This poses a problem with the
|
|
|
|
* tree logging stuff since it could have allocated across what appears to be 2
|
|
|
|
* entries since we would have merged the entries when adding the pinned extents
|
|
|
|
* back to the free space cache. So run through the space cache that we just
|
|
|
|
* loaded and merge contiguous entries. This will make the log replay stuff not
|
|
|
|
* blow up and it will make for nicer allocator behavior.
|
|
|
|
*/
|
|
|
|
static void merge_space_tree(struct btrfs_free_space_ctl *ctl)
|
|
|
|
{
|
|
|
|
struct btrfs_free_space *e, *prev = NULL;
|
|
|
|
struct rb_node *n;
|
|
|
|
|
|
|
|
again:
|
|
|
|
spin_lock(&ctl->tree_lock);
|
|
|
|
for (n = rb_first(&ctl->free_space_offset); n; n = rb_next(n)) {
|
|
|
|
e = rb_entry(n, struct btrfs_free_space, offset_index);
|
|
|
|
if (!prev)
|
|
|
|
goto next;
|
|
|
|
if (e->bitmap || prev->bitmap)
|
|
|
|
goto next;
|
|
|
|
if (prev->offset + prev->bytes == e->offset) {
|
|
|
|
unlink_free_space(ctl, prev);
|
|
|
|
unlink_free_space(ctl, e);
|
|
|
|
prev->bytes += e->bytes;
|
|
|
|
kmem_cache_free(btrfs_free_space_cachep, e);
|
|
|
|
link_free_space(ctl, prev);
|
|
|
|
prev = NULL;
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
|
|
|
goto again;
|
|
|
|
}
|
|
|
|
next:
|
|
|
|
prev = e;
|
|
|
|
}
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
|
|
|
}
|
|
|
|
|
2013-04-26 04:41:01 +08:00
|
|
|
static int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
|
|
|
|
struct btrfs_free_space_ctl *ctl,
|
|
|
|
struct btrfs_path *path, u64 offset)
|
2010-08-26 04:54:15 +08:00
|
|
|
{
|
|
|
|
struct btrfs_free_space_header *header;
|
|
|
|
struct extent_buffer *leaf;
|
2015-04-07 04:17:20 +08:00
|
|
|
struct btrfs_io_ctl io_ctl;
|
2010-08-26 04:54:15 +08:00
|
|
|
struct btrfs_key key;
|
2011-10-06 03:18:58 +08:00
|
|
|
struct btrfs_free_space *e, *n;
|
2014-12-31 09:51:35 +08:00
|
|
|
LIST_HEAD(bitmaps);
|
2010-08-26 04:54:15 +08:00
|
|
|
u64 num_entries;
|
|
|
|
u64 num_bitmaps;
|
|
|
|
u64 generation;
|
2011-10-06 03:18:58 +08:00
|
|
|
u8 type;
|
2011-06-06 22:50:35 +08:00
|
|
|
int ret = 0;
|
2010-08-26 04:54:15 +08:00
|
|
|
|
|
|
|
/* Nothing in the space cache, goodbye */
|
2011-04-20 10:20:14 +08:00
|
|
|
if (!i_size_read(inode))
|
2011-10-06 03:18:58 +08:00
|
|
|
return 0;
|
2010-08-26 04:54:15 +08:00
|
|
|
|
|
|
|
key.objectid = BTRFS_FREE_SPACE_OBJECTID;
|
2011-04-20 10:20:14 +08:00
|
|
|
key.offset = offset;
|
2010-08-26 04:54:15 +08:00
|
|
|
key.type = 0;
|
|
|
|
|
|
|
|
ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
|
2011-04-20 10:20:14 +08:00
|
|
|
if (ret < 0)
|
2011-10-06 03:18:58 +08:00
|
|
|
return 0;
|
2011-04-20 10:20:14 +08:00
|
|
|
else if (ret > 0) {
|
2011-05-23 00:33:42 +08:00
|
|
|
btrfs_release_path(path);
|
2011-10-06 03:18:58 +08:00
|
|
|
return 0;
|
2010-08-26 04:54:15 +08:00
|
|
|
}
|
|
|
|
|
2011-04-20 10:20:14 +08:00
|
|
|
ret = -1;
|
|
|
|
|
2010-08-26 04:54:15 +08:00
|
|
|
leaf = path->nodes[0];
|
|
|
|
header = btrfs_item_ptr(leaf, path->slots[0],
|
|
|
|
struct btrfs_free_space_header);
|
|
|
|
num_entries = btrfs_free_space_entries(leaf, header);
|
|
|
|
num_bitmaps = btrfs_free_space_bitmaps(leaf, header);
|
|
|
|
generation = btrfs_free_space_generation(leaf, header);
|
2011-05-23 00:33:42 +08:00
|
|
|
btrfs_release_path(path);
|
2010-08-26 04:54:15 +08:00
|
|
|
|
Btrfs: fix broken free space cache after the system crashed
When we mounted the filesystem after the crash, we got the following
message:
BTRFS error (device xxx): block group xxxx has wrong amount of free space
BTRFS error (device xxx): failed to load free space cache for block group xxx
It is because we didn't update the metadata of the allocated space (in extent
tree) until the file data was written into the disk. During this time, there was
no information about the allocated spaces in either the extent tree nor the
free space cache. when we wrote out the free space cache at this time (commit
transaction), those spaces were lost. In fact, only the free space that is
used to store the file data had this problem, the others didn't because
the metadata of them is updated in the same transaction context.
There are many methods which can fix the above problem
- track the allocated space, and write it out when we write out the free
space cache
- account the size of the allocated space that is used to store the file
data, if the size is not zero, don't write out the free space cache.
The first one is complex and may make the performance drop down.
This patch chose the second method, we use a per-block-group variant to
account the size of that allocated space. Besides that, we also introduce
a per-block-group read-write semaphore to avoid the race between
the allocation and the free space cache write out.
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-06-19 10:42:50 +08:00
|
|
|
if (!BTRFS_I(inode)->generation) {
|
|
|
|
btrfs_info(root->fs_info,
|
|
|
|
"The free space cache file (%llu) is invalid. skip it\n",
|
|
|
|
offset);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2010-08-26 04:54:15 +08:00
|
|
|
if (BTRFS_I(inode)->generation != generation) {
|
2013-03-20 06:41:23 +08:00
|
|
|
btrfs_err(root->fs_info,
|
|
|
|
"free space inode generation (%llu) "
|
|
|
|
"did not match free space cache generation (%llu)",
|
2013-08-20 19:20:07 +08:00
|
|
|
BTRFS_I(inode)->generation, generation);
|
2011-10-06 03:18:58 +08:00
|
|
|
return 0;
|
2010-08-26 04:54:15 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
if (!num_entries)
|
2011-10-06 03:18:58 +08:00
|
|
|
return 0;
|
2010-08-26 04:54:15 +08:00
|
|
|
|
2014-06-19 10:42:49 +08:00
|
|
|
ret = io_ctl_init(&io_ctl, inode, root, 0);
|
2012-01-09 14:36:28 +08:00
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
2010-08-26 04:54:15 +08:00
|
|
|
ret = readahead_cache(inode);
|
2011-04-20 10:20:14 +08:00
|
|
|
if (ret)
|
2010-08-26 04:54:15 +08:00
|
|
|
goto out;
|
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
ret = io_ctl_prepare_pages(&io_ctl, inode, 1);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
2010-08-26 04:54:15 +08:00
|
|
|
|
2011-10-06 20:58:24 +08:00
|
|
|
ret = io_ctl_check_crc(&io_ctl, 0);
|
|
|
|
if (ret)
|
|
|
|
goto free_cache;
|
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
ret = io_ctl_check_generation(&io_ctl, generation);
|
|
|
|
if (ret)
|
|
|
|
goto free_cache;
|
2010-08-26 04:54:15 +08:00
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
while (num_entries) {
|
|
|
|
e = kmem_cache_zalloc(btrfs_free_space_cachep,
|
|
|
|
GFP_NOFS);
|
|
|
|
if (!e)
|
2010-08-26 04:54:15 +08:00
|
|
|
goto free_cache;
|
|
|
|
|
2011-10-06 20:58:24 +08:00
|
|
|
ret = io_ctl_read_entry(&io_ctl, e, &type);
|
|
|
|
if (ret) {
|
|
|
|
kmem_cache_free(btrfs_free_space_cachep, e);
|
|
|
|
goto free_cache;
|
|
|
|
}
|
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
if (!e->bytes) {
|
|
|
|
kmem_cache_free(btrfs_free_space_cachep, e);
|
|
|
|
goto free_cache;
|
2010-08-26 04:54:15 +08:00
|
|
|
}
|
2011-10-06 03:18:58 +08:00
|
|
|
|
|
|
|
if (type == BTRFS_FREE_SPACE_EXTENT) {
|
|
|
|
spin_lock(&ctl->tree_lock);
|
|
|
|
ret = link_free_space(ctl, e);
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
|
|
|
if (ret) {
|
2013-03-20 06:41:23 +08:00
|
|
|
btrfs_err(root->fs_info,
|
|
|
|
"Duplicate entries in free space cache, dumping");
|
2011-10-06 03:18:58 +08:00
|
|
|
kmem_cache_free(btrfs_free_space_cachep, e);
|
2010-08-26 04:54:15 +08:00
|
|
|
goto free_cache;
|
|
|
|
}
|
2011-10-06 03:18:58 +08:00
|
|
|
} else {
|
2013-08-27 05:14:08 +08:00
|
|
|
ASSERT(num_bitmaps);
|
2011-10-06 03:18:58 +08:00
|
|
|
num_bitmaps--;
|
|
|
|
e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
|
|
|
|
if (!e->bitmap) {
|
|
|
|
kmem_cache_free(
|
|
|
|
btrfs_free_space_cachep, e);
|
2010-08-26 04:54:15 +08:00
|
|
|
goto free_cache;
|
|
|
|
}
|
2011-10-06 03:18:58 +08:00
|
|
|
spin_lock(&ctl->tree_lock);
|
|
|
|
ret = link_free_space(ctl, e);
|
|
|
|
ctl->total_bitmaps++;
|
|
|
|
ctl->op->recalc_thresholds(ctl);
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
|
|
|
if (ret) {
|
2013-03-20 06:41:23 +08:00
|
|
|
btrfs_err(root->fs_info,
|
|
|
|
"Duplicate entries in free space cache, dumping");
|
2011-01-29 06:05:48 +08:00
|
|
|
kmem_cache_free(btrfs_free_space_cachep, e);
|
2010-08-26 04:54:15 +08:00
|
|
|
goto free_cache;
|
|
|
|
}
|
2011-10-06 03:18:58 +08:00
|
|
|
list_add_tail(&e->list, &bitmaps);
|
2010-08-26 04:54:15 +08:00
|
|
|
}
|
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
num_entries--;
|
|
|
|
}
|
2010-08-26 04:54:15 +08:00
|
|
|
|
2011-11-11 09:45:05 +08:00
|
|
|
io_ctl_unmap_page(&io_ctl);
|
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
/*
|
|
|
|
* We add the bitmaps at the end of the entries in order that
|
|
|
|
* the bitmap entries are added to the cache.
|
|
|
|
*/
|
|
|
|
list_for_each_entry_safe(e, n, &bitmaps, list) {
|
2010-08-26 04:54:15 +08:00
|
|
|
list_del_init(&e->list);
|
2011-10-06 20:58:24 +08:00
|
|
|
ret = io_ctl_read_bitmap(&io_ctl, e);
|
|
|
|
if (ret)
|
|
|
|
goto free_cache;
|
2010-08-26 04:54:15 +08:00
|
|
|
}
|
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
io_ctl_drop_pages(&io_ctl);
|
2012-05-14 22:06:40 +08:00
|
|
|
merge_space_tree(ctl);
|
2010-08-26 04:54:15 +08:00
|
|
|
ret = 1;
|
|
|
|
out:
|
2011-10-06 03:18:58 +08:00
|
|
|
io_ctl_free(&io_ctl);
|
2010-08-26 04:54:15 +08:00
|
|
|
return ret;
|
|
|
|
free_cache:
|
2011-10-06 03:18:58 +08:00
|
|
|
io_ctl_drop_pages(&io_ctl);
|
2011-04-20 10:20:14 +08:00
|
|
|
__btrfs_remove_free_space_cache(ctl);
|
2010-08-26 04:54:15 +08:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2011-04-20 10:20:14 +08:00
|
|
|
int load_free_space_cache(struct btrfs_fs_info *fs_info,
|
|
|
|
struct btrfs_block_group_cache *block_group)
|
2010-07-03 00:14:14 +08:00
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
2011-04-20 10:20:14 +08:00
|
|
|
struct btrfs_root *root = fs_info->tree_root;
|
|
|
|
struct inode *inode;
|
|
|
|
struct btrfs_path *path;
|
2011-10-06 20:58:24 +08:00
|
|
|
int ret = 0;
|
2011-04-20 10:20:14 +08:00
|
|
|
bool matched;
|
|
|
|
u64 used = btrfs_block_group_used(&block_group->item);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If this block group has been marked to be cleared for one reason or
|
|
|
|
* another then we can't trust the on disk cache, so just return.
|
|
|
|
*/
|
2010-08-26 04:54:15 +08:00
|
|
|
spin_lock(&block_group->lock);
|
2011-04-20 10:20:14 +08:00
|
|
|
if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) {
|
|
|
|
spin_unlock(&block_group->lock);
|
|
|
|
return 0;
|
|
|
|
}
|
2010-08-26 04:54:15 +08:00
|
|
|
spin_unlock(&block_group->lock);
|
2011-04-20 10:20:14 +08:00
|
|
|
|
|
|
|
path = btrfs_alloc_path();
|
|
|
|
if (!path)
|
|
|
|
return 0;
|
2012-04-13 04:03:57 +08:00
|
|
|
path->search_commit_root = 1;
|
|
|
|
path->skip_locking = 1;
|
2011-04-20 10:20:14 +08:00
|
|
|
|
|
|
|
inode = lookup_free_space_inode(root, block_group, path);
|
|
|
|
if (IS_ERR(inode)) {
|
|
|
|
btrfs_free_path(path);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-10-06 20:58:24 +08:00
|
|
|
/* We may have converted the inode and made the cache invalid. */
|
|
|
|
spin_lock(&block_group->lock);
|
|
|
|
if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) {
|
|
|
|
spin_unlock(&block_group->lock);
|
2012-02-14 16:12:23 +08:00
|
|
|
btrfs_free_path(path);
|
2011-10-06 20:58:24 +08:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
spin_unlock(&block_group->lock);
|
|
|
|
|
2011-04-20 10:20:14 +08:00
|
|
|
ret = __load_free_space_cache(fs_info->tree_root, inode, ctl,
|
|
|
|
path, block_group->key.objectid);
|
|
|
|
btrfs_free_path(path);
|
|
|
|
if (ret <= 0)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
spin_lock(&ctl->tree_lock);
|
|
|
|
matched = (ctl->free_space == (block_group->key.offset - used -
|
|
|
|
block_group->bytes_super));
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
|
|
|
|
|
|
|
if (!matched) {
|
|
|
|
__btrfs_remove_free_space_cache(ctl);
|
2014-04-24 13:31:55 +08:00
|
|
|
btrfs_warn(fs_info, "block group %llu has wrong amount of free space",
|
2013-03-20 06:41:23 +08:00
|
|
|
block_group->key.objectid);
|
2011-04-20 10:20:14 +08:00
|
|
|
ret = -1;
|
|
|
|
}
|
|
|
|
out:
|
|
|
|
if (ret < 0) {
|
|
|
|
/* This cache is bogus, make sure it gets cleared */
|
|
|
|
spin_lock(&block_group->lock);
|
|
|
|
block_group->disk_cache_state = BTRFS_DC_CLEAR;
|
|
|
|
spin_unlock(&block_group->lock);
|
2011-04-20 10:33:24 +08:00
|
|
|
ret = 0;
|
2011-04-20 10:20:14 +08:00
|
|
|
|
2015-11-28 00:32:04 +08:00
|
|
|
btrfs_warn(fs_info, "failed to load free space cache for block group %llu, rebuilding it now",
|
2013-03-20 06:41:23 +08:00
|
|
|
block_group->key.objectid);
|
2011-04-20 10:20:14 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
iput(inode);
|
|
|
|
return ret;
|
2010-08-26 04:54:15 +08:00
|
|
|
}
|
|
|
|
|
2014-05-20 11:47:56 +08:00
|
|
|
static noinline_for_stack
|
2015-04-07 04:17:20 +08:00
|
|
|
int write_cache_extent_entries(struct btrfs_io_ctl *io_ctl,
|
2014-05-20 11:47:56 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl,
|
|
|
|
struct btrfs_block_group_cache *block_group,
|
|
|
|
int *entries, int *bitmaps,
|
|
|
|
struct list_head *bitmap_list)
|
2010-07-03 00:14:14 +08:00
|
|
|
{
|
2011-08-30 22:19:10 +08:00
|
|
|
int ret;
|
2014-05-20 11:47:56 +08:00
|
|
|
struct btrfs_free_cluster *cluster = NULL;
|
2015-04-07 03:46:08 +08:00
|
|
|
struct btrfs_free_cluster *cluster_locked = NULL;
|
2014-05-20 11:47:56 +08:00
|
|
|
struct rb_node *node = rb_first(&ctl->free_space_offset);
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
struct btrfs_trim_range *trim_entry;
|
2011-04-07 01:05:22 +08:00
|
|
|
|
2011-04-01 22:55:00 +08:00
|
|
|
/* Get the cluster for this block_group if it exists */
|
2014-05-20 11:47:56 +08:00
|
|
|
if (block_group && !list_empty(&block_group->cluster_list)) {
|
2011-04-01 22:55:00 +08:00
|
|
|
cluster = list_entry(block_group->cluster_list.next,
|
|
|
|
struct btrfs_free_cluster,
|
|
|
|
block_group_list);
|
2014-05-20 11:47:56 +08:00
|
|
|
}
|
2011-04-01 22:55:00 +08:00
|
|
|
|
2011-10-05 22:00:18 +08:00
|
|
|
if (!node && cluster) {
|
2015-04-07 03:46:08 +08:00
|
|
|
cluster_locked = cluster;
|
|
|
|
spin_lock(&cluster_locked->lock);
|
2011-10-05 22:00:18 +08:00
|
|
|
node = rb_first(&cluster->root);
|
|
|
|
cluster = NULL;
|
|
|
|
}
|
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
/* Write out the extent entries */
|
|
|
|
while (node) {
|
|
|
|
struct btrfs_free_space *e;
|
2010-07-03 00:14:14 +08:00
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
e = rb_entry(node, struct btrfs_free_space, offset_index);
|
2014-05-20 11:47:56 +08:00
|
|
|
*entries += 1;
|
2010-07-03 00:14:14 +08:00
|
|
|
|
2014-05-20 11:47:56 +08:00
|
|
|
ret = io_ctl_add_entry(io_ctl, e->offset, e->bytes,
|
2011-10-06 03:18:58 +08:00
|
|
|
e->bitmap);
|
|
|
|
if (ret)
|
2014-05-20 11:47:56 +08:00
|
|
|
goto fail;
|
2011-06-11 03:31:13 +08:00
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
if (e->bitmap) {
|
2014-05-20 11:47:56 +08:00
|
|
|
list_add_tail(&e->list, bitmap_list);
|
|
|
|
*bitmaps += 1;
|
2011-06-11 03:31:13 +08:00
|
|
|
}
|
2011-10-06 03:18:58 +08:00
|
|
|
node = rb_next(node);
|
|
|
|
if (!node && cluster) {
|
|
|
|
node = rb_first(&cluster->root);
|
2015-04-07 03:46:08 +08:00
|
|
|
cluster_locked = cluster;
|
|
|
|
spin_lock(&cluster_locked->lock);
|
2011-10-06 03:18:58 +08:00
|
|
|
cluster = NULL;
|
2011-04-01 22:55:00 +08:00
|
|
|
}
|
2011-10-06 03:18:58 +08:00
|
|
|
}
|
2015-04-07 03:46:08 +08:00
|
|
|
if (cluster_locked) {
|
|
|
|
spin_unlock(&cluster_locked->lock);
|
|
|
|
cluster_locked = NULL;
|
|
|
|
}
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Make sure we don't miss any range that was removed from our rbtree
|
|
|
|
* because trimming is running. Otherwise after a umount+mount (or crash
|
|
|
|
* after committing the transaction) we would leak free space and get
|
|
|
|
* an inconsistent free space cache report from fsck.
|
|
|
|
*/
|
|
|
|
list_for_each_entry(trim_entry, &ctl->trimming_ranges, list) {
|
|
|
|
ret = io_ctl_add_entry(io_ctl, trim_entry->start,
|
|
|
|
trim_entry->bytes, NULL);
|
|
|
|
if (ret)
|
|
|
|
goto fail;
|
|
|
|
*entries += 1;
|
|
|
|
}
|
|
|
|
|
2014-05-20 11:47:56 +08:00
|
|
|
return 0;
|
|
|
|
fail:
|
2015-04-07 03:46:08 +08:00
|
|
|
if (cluster_locked)
|
|
|
|
spin_unlock(&cluster_locked->lock);
|
2014-05-20 11:47:56 +08:00
|
|
|
return -ENOSPC;
|
|
|
|
}
|
|
|
|
|
|
|
|
static noinline_for_stack int
|
|
|
|
update_cache_item(struct btrfs_trans_handle *trans,
|
|
|
|
struct btrfs_root *root,
|
|
|
|
struct inode *inode,
|
|
|
|
struct btrfs_path *path, u64 offset,
|
|
|
|
int entries, int bitmaps)
|
|
|
|
{
|
|
|
|
struct btrfs_key key;
|
|
|
|
struct btrfs_free_space_header *header;
|
|
|
|
struct extent_buffer *leaf;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
key.objectid = BTRFS_FREE_SPACE_OBJECTID;
|
|
|
|
key.offset = offset;
|
|
|
|
key.type = 0;
|
|
|
|
|
|
|
|
ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
|
|
|
|
if (ret < 0) {
|
|
|
|
clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, inode->i_size - 1,
|
|
|
|
EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, NULL,
|
|
|
|
GFP_NOFS);
|
|
|
|
goto fail;
|
|
|
|
}
|
|
|
|
leaf = path->nodes[0];
|
|
|
|
if (ret > 0) {
|
|
|
|
struct btrfs_key found_key;
|
|
|
|
ASSERT(path->slots[0]);
|
|
|
|
path->slots[0]--;
|
|
|
|
btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
|
|
|
|
if (found_key.objectid != BTRFS_FREE_SPACE_OBJECTID ||
|
|
|
|
found_key.offset != offset) {
|
|
|
|
clear_extent_bit(&BTRFS_I(inode)->io_tree, 0,
|
|
|
|
inode->i_size - 1,
|
|
|
|
EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0,
|
|
|
|
NULL, GFP_NOFS);
|
|
|
|
btrfs_release_path(path);
|
|
|
|
goto fail;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
BTRFS_I(inode)->generation = trans->transid;
|
|
|
|
header = btrfs_item_ptr(leaf, path->slots[0],
|
|
|
|
struct btrfs_free_space_header);
|
|
|
|
btrfs_set_free_space_entries(leaf, header, entries);
|
|
|
|
btrfs_set_free_space_bitmaps(leaf, header, bitmaps);
|
|
|
|
btrfs_set_free_space_generation(leaf, header, trans->transid);
|
|
|
|
btrfs_mark_buffer_dirty(leaf);
|
|
|
|
btrfs_release_path(path);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
fail:
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static noinline_for_stack int
|
2014-06-19 10:42:49 +08:00
|
|
|
write_pinned_extent_entries(struct btrfs_root *root,
|
|
|
|
struct btrfs_block_group_cache *block_group,
|
2015-04-07 04:17:20 +08:00
|
|
|
struct btrfs_io_ctl *io_ctl,
|
2014-06-19 10:42:49 +08:00
|
|
|
int *entries)
|
2014-05-20 11:47:56 +08:00
|
|
|
{
|
|
|
|
u64 start, extent_start, extent_end, len;
|
|
|
|
struct extent_io_tree *unpin = NULL;
|
|
|
|
int ret;
|
2011-04-01 22:55:00 +08:00
|
|
|
|
2014-06-19 10:42:49 +08:00
|
|
|
if (!block_group)
|
|
|
|
return 0;
|
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
/*
|
|
|
|
* We want to add any pinned extents to our free space cache
|
|
|
|
* so we don't leak the space
|
2014-05-20 11:47:56 +08:00
|
|
|
*
|
2012-01-10 16:41:01 +08:00
|
|
|
* We shouldn't have switched the pinned extents yet so this is the
|
|
|
|
* right one
|
|
|
|
*/
|
|
|
|
unpin = root->fs_info->pinned_extents;
|
|
|
|
|
2014-06-19 10:42:49 +08:00
|
|
|
start = block_group->key.objectid;
|
2012-01-10 16:41:01 +08:00
|
|
|
|
2014-06-19 10:42:49 +08:00
|
|
|
while (start < block_group->key.objectid + block_group->key.offset) {
|
2012-01-10 16:41:01 +08:00
|
|
|
ret = find_first_extent_bit(unpin, start,
|
|
|
|
&extent_start, &extent_end,
|
2012-09-28 05:07:30 +08:00
|
|
|
EXTENT_DIRTY, NULL);
|
2014-06-19 10:42:49 +08:00
|
|
|
if (ret)
|
|
|
|
return 0;
|
2010-07-03 00:14:14 +08:00
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
/* This pinned extent is out of our range */
|
2012-01-10 16:41:01 +08:00
|
|
|
if (extent_start >= block_group->key.objectid +
|
2011-10-06 03:18:58 +08:00
|
|
|
block_group->key.offset)
|
2014-06-19 10:42:49 +08:00
|
|
|
return 0;
|
2011-06-11 03:31:13 +08:00
|
|
|
|
2012-01-10 16:41:01 +08:00
|
|
|
extent_start = max(extent_start, start);
|
|
|
|
extent_end = min(block_group->key.objectid +
|
|
|
|
block_group->key.offset, extent_end + 1);
|
|
|
|
len = extent_end - extent_start;
|
2010-07-03 00:14:14 +08:00
|
|
|
|
2014-05-20 11:47:56 +08:00
|
|
|
*entries += 1;
|
|
|
|
ret = io_ctl_add_entry(io_ctl, extent_start, len, NULL);
|
2011-10-06 03:18:58 +08:00
|
|
|
if (ret)
|
2014-06-19 10:42:49 +08:00
|
|
|
return -ENOSPC;
|
2010-07-03 00:14:14 +08:00
|
|
|
|
2012-01-10 16:41:01 +08:00
|
|
|
start = extent_end;
|
2011-10-06 03:18:58 +08:00
|
|
|
}
|
2010-07-03 00:14:14 +08:00
|
|
|
|
2014-06-19 10:42:49 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static noinline_for_stack int
|
2015-04-07 04:17:20 +08:00
|
|
|
write_bitmap_entries(struct btrfs_io_ctl *io_ctl, struct list_head *bitmap_list)
|
2014-06-19 10:42:49 +08:00
|
|
|
{
|
2015-12-18 22:17:00 +08:00
|
|
|
struct btrfs_free_space *entry, *next;
|
2014-06-19 10:42:49 +08:00
|
|
|
int ret;
|
|
|
|
|
2010-07-03 00:14:14 +08:00
|
|
|
/* Write out the bitmaps */
|
2015-12-18 22:17:00 +08:00
|
|
|
list_for_each_entry_safe(entry, next, bitmap_list, list) {
|
2014-05-20 11:47:56 +08:00
|
|
|
ret = io_ctl_add_bitmap(io_ctl, entry->bitmap);
|
2011-10-06 03:18:58 +08:00
|
|
|
if (ret)
|
2014-06-19 10:42:49 +08:00
|
|
|
return -ENOSPC;
|
2010-07-03 00:14:14 +08:00
|
|
|
list_del_init(&entry->list);
|
2011-04-07 01:05:22 +08:00
|
|
|
}
|
|
|
|
|
2014-06-19 10:42:49 +08:00
|
|
|
return 0;
|
|
|
|
}
|
2010-07-03 00:14:14 +08:00
|
|
|
|
2014-06-19 10:42:49 +08:00
|
|
|
static int flush_dirty_cache(struct inode *inode)
|
|
|
|
{
|
|
|
|
int ret;
|
2011-04-07 01:05:22 +08:00
|
|
|
|
2013-10-26 04:13:35 +08:00
|
|
|
ret = btrfs_wait_ordered_range(inode, 0, (u64)-1);
|
2014-06-19 10:42:49 +08:00
|
|
|
if (ret)
|
2013-10-26 04:13:35 +08:00
|
|
|
clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, inode->i_size - 1,
|
|
|
|
EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, NULL,
|
|
|
|
GFP_NOFS);
|
2010-07-03 00:14:14 +08:00
|
|
|
|
2014-06-19 10:42:49 +08:00
|
|
|
return ret;
|
2014-05-20 11:47:56 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void noinline_for_stack
|
2015-04-25 02:00:00 +08:00
|
|
|
cleanup_bitmap_list(struct list_head *bitmap_list)
|
2014-05-20 11:47:56 +08:00
|
|
|
{
|
2015-12-18 22:17:00 +08:00
|
|
|
struct btrfs_free_space *entry, *next;
|
2014-06-19 10:42:49 +08:00
|
|
|
|
2015-12-18 22:17:00 +08:00
|
|
|
list_for_each_entry_safe(entry, next, bitmap_list, list)
|
2014-05-20 11:47:56 +08:00
|
|
|
list_del_init(&entry->list);
|
2015-04-25 02:00:00 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void noinline_for_stack
|
|
|
|
cleanup_write_cache_enospc(struct inode *inode,
|
|
|
|
struct btrfs_io_ctl *io_ctl,
|
|
|
|
struct extent_state **cached_state,
|
|
|
|
struct list_head *bitmap_list)
|
|
|
|
{
|
2014-05-20 11:47:56 +08:00
|
|
|
io_ctl_drop_pages(io_ctl);
|
|
|
|
unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
|
|
|
|
i_size_read(inode) - 1, cached_state,
|
|
|
|
GFP_NOFS);
|
|
|
|
}
|
2011-10-06 04:33:53 +08:00
|
|
|
|
2015-04-05 08:14:42 +08:00
|
|
|
int btrfs_wait_cache_io(struct btrfs_root *root,
|
|
|
|
struct btrfs_trans_handle *trans,
|
|
|
|
struct btrfs_block_group_cache *block_group,
|
|
|
|
struct btrfs_io_ctl *io_ctl,
|
|
|
|
struct btrfs_path *path, u64 offset)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
struct inode *inode = io_ctl->inode;
|
|
|
|
|
2015-04-07 03:46:08 +08:00
|
|
|
if (!inode)
|
|
|
|
return 0;
|
|
|
|
|
2015-04-23 23:02:49 +08:00
|
|
|
if (block_group)
|
|
|
|
root = root->fs_info->tree_root;
|
2015-04-05 08:14:42 +08:00
|
|
|
|
|
|
|
/* Flush the dirty pages in the cache file. */
|
|
|
|
ret = flush_dirty_cache(inode);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
/* Update the cache item to tell everyone this cache file is valid. */
|
|
|
|
ret = update_cache_item(trans, root, inode, path, offset,
|
|
|
|
io_ctl->entries, io_ctl->bitmaps);
|
|
|
|
out:
|
|
|
|
io_ctl_free(io_ctl);
|
|
|
|
if (ret) {
|
|
|
|
invalidate_inode_pages2(inode->i_mapping);
|
|
|
|
BTRFS_I(inode)->generation = 0;
|
|
|
|
if (block_group) {
|
|
|
|
#ifdef DEBUG
|
|
|
|
btrfs_err(root->fs_info,
|
|
|
|
"failed to write free space cache for block group %llu",
|
|
|
|
block_group->key.objectid);
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
}
|
|
|
|
btrfs_update_inode(trans, root, inode);
|
|
|
|
|
|
|
|
if (block_group) {
|
2015-04-07 03:46:08 +08:00
|
|
|
/* the dirty list is protected by the dirty_bgs_lock */
|
|
|
|
spin_lock(&trans->transaction->dirty_bgs_lock);
|
|
|
|
|
|
|
|
/* the disk_cache_state is protected by the block group lock */
|
2015-04-05 08:14:42 +08:00
|
|
|
spin_lock(&block_group->lock);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* only mark this as written if we didn't get put back on
|
2015-04-07 03:46:08 +08:00
|
|
|
* the dirty list while waiting for IO. Otherwise our
|
|
|
|
* cache state won't be right, and we won't get written again
|
2015-04-05 08:14:42 +08:00
|
|
|
*/
|
|
|
|
if (!ret && list_empty(&block_group->dirty_list))
|
|
|
|
block_group->disk_cache_state = BTRFS_DC_WRITTEN;
|
|
|
|
else if (ret)
|
|
|
|
block_group->disk_cache_state = BTRFS_DC_ERROR;
|
|
|
|
|
|
|
|
spin_unlock(&block_group->lock);
|
2015-04-07 03:46:08 +08:00
|
|
|
spin_unlock(&trans->transaction->dirty_bgs_lock);
|
2015-04-05 08:14:42 +08:00
|
|
|
io_ctl->inode = NULL;
|
|
|
|
iput(inode);
|
|
|
|
}
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
}
|
|
|
|
|
2014-05-20 11:47:56 +08:00
|
|
|
/**
|
|
|
|
* __btrfs_write_out_cache - write out cached info to an inode
|
|
|
|
* @root - the root the inode belongs to
|
|
|
|
* @ctl - the free space cache we are going to write out
|
|
|
|
* @block_group - the block_group for this cache if it belongs to a block_group
|
|
|
|
* @trans - the trans handle
|
|
|
|
* @path - the path to use
|
|
|
|
* @offset - the offset for the key we'll insert
|
|
|
|
*
|
|
|
|
* This function writes out a free space cache struct to disk for quick recovery
|
2015-10-04 17:05:32 +08:00
|
|
|
* on mount. This will return 0 if it was successful in writing the cache out,
|
2015-02-24 18:47:06 +08:00
|
|
|
* or an errno if it was not.
|
2014-05-20 11:47:56 +08:00
|
|
|
*/
|
|
|
|
static int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
|
|
|
|
struct btrfs_free_space_ctl *ctl,
|
|
|
|
struct btrfs_block_group_cache *block_group,
|
2015-04-05 08:14:42 +08:00
|
|
|
struct btrfs_io_ctl *io_ctl,
|
2014-05-20 11:47:56 +08:00
|
|
|
struct btrfs_trans_handle *trans,
|
|
|
|
struct btrfs_path *path, u64 offset)
|
|
|
|
{
|
|
|
|
struct extent_state *cached_state = NULL;
|
2014-06-19 10:42:49 +08:00
|
|
|
LIST_HEAD(bitmap_list);
|
2014-05-20 11:47:56 +08:00
|
|
|
int entries = 0;
|
|
|
|
int bitmaps = 0;
|
|
|
|
int ret;
|
2015-04-05 08:14:42 +08:00
|
|
|
int must_iput = 0;
|
2014-05-20 11:47:56 +08:00
|
|
|
|
|
|
|
if (!i_size_read(inode))
|
2015-02-24 18:47:06 +08:00
|
|
|
return -EIO;
|
2014-05-20 11:47:56 +08:00
|
|
|
|
2015-04-05 08:14:42 +08:00
|
|
|
WARN_ON(io_ctl->pages);
|
|
|
|
ret = io_ctl_init(io_ctl, inode, root, 1);
|
2014-05-20 11:47:56 +08:00
|
|
|
if (ret)
|
2015-02-24 18:47:06 +08:00
|
|
|
return ret;
|
2014-05-20 11:47:56 +08:00
|
|
|
|
Btrfs: fix broken free space cache after the system crashed
When we mounted the filesystem after the crash, we got the following
message:
BTRFS error (device xxx): block group xxxx has wrong amount of free space
BTRFS error (device xxx): failed to load free space cache for block group xxx
It is because we didn't update the metadata of the allocated space (in extent
tree) until the file data was written into the disk. During this time, there was
no information about the allocated spaces in either the extent tree nor the
free space cache. when we wrote out the free space cache at this time (commit
transaction), those spaces were lost. In fact, only the free space that is
used to store the file data had this problem, the others didn't because
the metadata of them is updated in the same transaction context.
There are many methods which can fix the above problem
- track the allocated space, and write it out when we write out the free
space cache
- account the size of the allocated space that is used to store the file
data, if the size is not zero, don't write out the free space cache.
The first one is complex and may make the performance drop down.
This patch chose the second method, we use a per-block-group variant to
account the size of that allocated space. Besides that, we also introduce
a per-block-group read-write semaphore to avoid the race between
the allocation and the free space cache write out.
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-06-19 10:42:50 +08:00
|
|
|
if (block_group && (block_group->flags & BTRFS_BLOCK_GROUP_DATA)) {
|
|
|
|
down_write(&block_group->data_rwsem);
|
|
|
|
spin_lock(&block_group->lock);
|
|
|
|
if (block_group->delalloc_bytes) {
|
|
|
|
block_group->disk_cache_state = BTRFS_DC_WRITTEN;
|
|
|
|
spin_unlock(&block_group->lock);
|
|
|
|
up_write(&block_group->data_rwsem);
|
|
|
|
BTRFS_I(inode)->generation = 0;
|
|
|
|
ret = 0;
|
2015-04-05 08:14:42 +08:00
|
|
|
must_iput = 1;
|
Btrfs: fix broken free space cache after the system crashed
When we mounted the filesystem after the crash, we got the following
message:
BTRFS error (device xxx): block group xxxx has wrong amount of free space
BTRFS error (device xxx): failed to load free space cache for block group xxx
It is because we didn't update the metadata of the allocated space (in extent
tree) until the file data was written into the disk. During this time, there was
no information about the allocated spaces in either the extent tree nor the
free space cache. when we wrote out the free space cache at this time (commit
transaction), those spaces were lost. In fact, only the free space that is
used to store the file data had this problem, the others didn't because
the metadata of them is updated in the same transaction context.
There are many methods which can fix the above problem
- track the allocated space, and write it out when we write out the free
space cache
- account the size of the allocated space that is used to store the file
data, if the size is not zero, don't write out the free space cache.
The first one is complex and may make the performance drop down.
This patch chose the second method, we use a per-block-group variant to
account the size of that allocated space. Besides that, we also introduce
a per-block-group read-write semaphore to avoid the race between
the allocation and the free space cache write out.
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-06-19 10:42:50 +08:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
spin_unlock(&block_group->lock);
|
|
|
|
}
|
|
|
|
|
2014-05-20 11:47:56 +08:00
|
|
|
/* Lock all pages first so we can lock the extent safely. */
|
2015-02-24 18:47:06 +08:00
|
|
|
ret = io_ctl_prepare_pages(io_ctl, inode, 0);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
2014-05-20 11:47:56 +08:00
|
|
|
|
|
|
|
lock_extent_bits(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1,
|
2015-12-03 21:30:40 +08:00
|
|
|
&cached_state);
|
2014-05-20 11:47:56 +08:00
|
|
|
|
2015-04-05 08:14:42 +08:00
|
|
|
io_ctl_set_generation(io_ctl, trans->transid);
|
2014-05-20 11:47:56 +08:00
|
|
|
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
mutex_lock(&ctl->cache_writeout_mutex);
|
2014-06-19 10:42:49 +08:00
|
|
|
/* Write out the extent entries in the free space cache */
|
2015-04-07 03:46:08 +08:00
|
|
|
spin_lock(&ctl->tree_lock);
|
2015-04-05 08:14:42 +08:00
|
|
|
ret = write_cache_extent_entries(io_ctl, ctl,
|
2014-05-20 11:47:56 +08:00
|
|
|
block_group, &entries, &bitmaps,
|
|
|
|
&bitmap_list);
|
2015-04-25 02:00:00 +08:00
|
|
|
if (ret)
|
|
|
|
goto out_nospc_locked;
|
2014-05-20 11:47:56 +08:00
|
|
|
|
2014-06-19 10:42:49 +08:00
|
|
|
/*
|
|
|
|
* Some spaces that are freed in the current transaction are pinned,
|
|
|
|
* they will be added into free space cache after the transaction is
|
|
|
|
* committed, we shouldn't lose them.
|
2015-04-07 03:46:08 +08:00
|
|
|
*
|
|
|
|
* If this changes while we are working we'll get added back to
|
|
|
|
* the dirty list and redo it. No locking needed
|
2014-06-19 10:42:49 +08:00
|
|
|
*/
|
2015-04-05 08:14:42 +08:00
|
|
|
ret = write_pinned_extent_entries(root, block_group, io_ctl, &entries);
|
2015-04-25 02:00:00 +08:00
|
|
|
if (ret)
|
|
|
|
goto out_nospc_locked;
|
2014-06-19 10:42:49 +08:00
|
|
|
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
/*
|
|
|
|
* At last, we write out all the bitmaps and keep cache_writeout_mutex
|
|
|
|
* locked while doing it because a concurrent trim can be manipulating
|
|
|
|
* or freeing the bitmap.
|
|
|
|
*/
|
2015-04-05 08:14:42 +08:00
|
|
|
ret = write_bitmap_entries(io_ctl, &bitmap_list);
|
2015-04-07 03:46:08 +08:00
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
mutex_unlock(&ctl->cache_writeout_mutex);
|
2014-06-19 10:42:49 +08:00
|
|
|
if (ret)
|
|
|
|
goto out_nospc;
|
|
|
|
|
|
|
|
/* Zero out the rest of the pages just to make sure */
|
2015-04-05 08:14:42 +08:00
|
|
|
io_ctl_zero_remaining_pages(io_ctl);
|
2014-05-20 11:47:56 +08:00
|
|
|
|
2014-06-19 10:42:49 +08:00
|
|
|
/* Everything is written out, now we dirty the pages in the file. */
|
2015-04-05 08:14:42 +08:00
|
|
|
ret = btrfs_dirty_pages(root, inode, io_ctl->pages, io_ctl->num_pages,
|
2014-06-19 10:42:49 +08:00
|
|
|
0, i_size_read(inode), &cached_state);
|
|
|
|
if (ret)
|
2014-05-20 11:47:56 +08:00
|
|
|
goto out_nospc;
|
2014-06-19 10:42:49 +08:00
|
|
|
|
Btrfs: fix broken free space cache after the system crashed
When we mounted the filesystem after the crash, we got the following
message:
BTRFS error (device xxx): block group xxxx has wrong amount of free space
BTRFS error (device xxx): failed to load free space cache for block group xxx
It is because we didn't update the metadata of the allocated space (in extent
tree) until the file data was written into the disk. During this time, there was
no information about the allocated spaces in either the extent tree nor the
free space cache. when we wrote out the free space cache at this time (commit
transaction), those spaces were lost. In fact, only the free space that is
used to store the file data had this problem, the others didn't because
the metadata of them is updated in the same transaction context.
There are many methods which can fix the above problem
- track the allocated space, and write it out when we write out the free
space cache
- account the size of the allocated space that is used to store the file
data, if the size is not zero, don't write out the free space cache.
The first one is complex and may make the performance drop down.
This patch chose the second method, we use a per-block-group variant to
account the size of that allocated space. Besides that, we also introduce
a per-block-group read-write semaphore to avoid the race between
the allocation and the free space cache write out.
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-06-19 10:42:50 +08:00
|
|
|
if (block_group && (block_group->flags & BTRFS_BLOCK_GROUP_DATA))
|
|
|
|
up_write(&block_group->data_rwsem);
|
2014-06-19 10:42:49 +08:00
|
|
|
/*
|
|
|
|
* Release the pages and unlock the extent, we will flush
|
|
|
|
* them out later
|
|
|
|
*/
|
2015-04-05 08:14:42 +08:00
|
|
|
io_ctl_drop_pages(io_ctl);
|
2014-06-19 10:42:49 +08:00
|
|
|
|
|
|
|
unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
|
|
|
|
i_size_read(inode) - 1, &cached_state, GFP_NOFS);
|
|
|
|
|
2015-04-05 08:14:42 +08:00
|
|
|
/*
|
|
|
|
* at this point the pages are under IO and we're happy,
|
|
|
|
* The caller is responsible for waiting on them and updating the
|
|
|
|
* the cache and the inode
|
|
|
|
*/
|
|
|
|
io_ctl->entries = entries;
|
|
|
|
io_ctl->bitmaps = bitmaps;
|
|
|
|
|
|
|
|
ret = btrfs_fdatawrite_range(inode, 0, (u64)-1);
|
2014-06-19 10:42:49 +08:00
|
|
|
if (ret)
|
2014-05-20 11:47:56 +08:00
|
|
|
goto out;
|
|
|
|
|
2015-04-05 08:14:42 +08:00
|
|
|
return 0;
|
|
|
|
|
2011-06-11 03:31:13 +08:00
|
|
|
out:
|
2015-04-05 08:14:42 +08:00
|
|
|
io_ctl->inode = NULL;
|
|
|
|
io_ctl_free(io_ctl);
|
2014-06-19 10:42:49 +08:00
|
|
|
if (ret) {
|
2011-10-06 03:18:58 +08:00
|
|
|
invalidate_inode_pages2(inode->i_mapping);
|
2010-07-03 00:14:14 +08:00
|
|
|
BTRFS_I(inode)->generation = 0;
|
|
|
|
}
|
|
|
|
btrfs_update_inode(trans, root, inode);
|
2015-04-05 08:14:42 +08:00
|
|
|
if (must_iput)
|
|
|
|
iput(inode);
|
2014-06-19 10:42:49 +08:00
|
|
|
return ret;
|
2011-10-06 03:18:58 +08:00
|
|
|
|
2015-04-25 02:00:00 +08:00
|
|
|
out_nospc_locked:
|
|
|
|
cleanup_bitmap_list(&bitmap_list);
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
|
|
|
mutex_unlock(&ctl->cache_writeout_mutex);
|
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
out_nospc:
|
2015-04-05 08:14:42 +08:00
|
|
|
cleanup_write_cache_enospc(inode, io_ctl, &cached_state, &bitmap_list);
|
Btrfs: fix broken free space cache after the system crashed
When we mounted the filesystem after the crash, we got the following
message:
BTRFS error (device xxx): block group xxxx has wrong amount of free space
BTRFS error (device xxx): failed to load free space cache for block group xxx
It is because we didn't update the metadata of the allocated space (in extent
tree) until the file data was written into the disk. During this time, there was
no information about the allocated spaces in either the extent tree nor the
free space cache. when we wrote out the free space cache at this time (commit
transaction), those spaces were lost. In fact, only the free space that is
used to store the file data had this problem, the others didn't because
the metadata of them is updated in the same transaction context.
There are many methods which can fix the above problem
- track the allocated space, and write it out when we write out the free
space cache
- account the size of the allocated space that is used to store the file
data, if the size is not zero, don't write out the free space cache.
The first one is complex and may make the performance drop down.
This patch chose the second method, we use a per-block-group variant to
account the size of that allocated space. Besides that, we also introduce
a per-block-group read-write semaphore to avoid the race between
the allocation and the free space cache write out.
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-06-19 10:42:50 +08:00
|
|
|
|
|
|
|
if (block_group && (block_group->flags & BTRFS_BLOCK_GROUP_DATA))
|
|
|
|
up_write(&block_group->data_rwsem);
|
|
|
|
|
2011-10-06 03:18:58 +08:00
|
|
|
goto out;
|
2011-04-20 10:20:14 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
int btrfs_write_out_cache(struct btrfs_root *root,
|
|
|
|
struct btrfs_trans_handle *trans,
|
|
|
|
struct btrfs_block_group_cache *block_group,
|
|
|
|
struct btrfs_path *path)
|
|
|
|
{
|
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
|
|
|
struct inode *inode;
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
root = root->fs_info->tree_root;
|
|
|
|
|
|
|
|
spin_lock(&block_group->lock);
|
|
|
|
if (block_group->disk_cache_state < BTRFS_DC_SETUP) {
|
|
|
|
spin_unlock(&block_group->lock);
|
Btrfs: fix broken free space cache after the system crashed
When we mounted the filesystem after the crash, we got the following
message:
BTRFS error (device xxx): block group xxxx has wrong amount of free space
BTRFS error (device xxx): failed to load free space cache for block group xxx
It is because we didn't update the metadata of the allocated space (in extent
tree) until the file data was written into the disk. During this time, there was
no information about the allocated spaces in either the extent tree nor the
free space cache. when we wrote out the free space cache at this time (commit
transaction), those spaces were lost. In fact, only the free space that is
used to store the file data had this problem, the others didn't because
the metadata of them is updated in the same transaction context.
There are many methods which can fix the above problem
- track the allocated space, and write it out when we write out the free
space cache
- account the size of the allocated space that is used to store the file
data, if the size is not zero, don't write out the free space cache.
The first one is complex and may make the performance drop down.
This patch chose the second method, we use a per-block-group variant to
account the size of that allocated space. Besides that, we also introduce
a per-block-group read-write semaphore to avoid the race between
the allocation and the free space cache write out.
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-06-19 10:42:50 +08:00
|
|
|
return 0;
|
|
|
|
}
|
2011-04-20 10:20:14 +08:00
|
|
|
spin_unlock(&block_group->lock);
|
|
|
|
|
|
|
|
inode = lookup_free_space_inode(root, block_group, path);
|
|
|
|
if (IS_ERR(inode))
|
|
|
|
return 0;
|
|
|
|
|
2015-04-05 08:14:42 +08:00
|
|
|
ret = __btrfs_write_out_cache(root, inode, ctl, block_group,
|
|
|
|
&block_group->io_ctl, trans,
|
2011-04-20 10:20:14 +08:00
|
|
|
path, block_group->key.objectid);
|
2011-08-30 22:19:10 +08:00
|
|
|
if (ret) {
|
|
|
|
#ifdef DEBUG
|
2013-03-20 06:41:23 +08:00
|
|
|
btrfs_err(root->fs_info,
|
|
|
|
"failed to write free space cache for block group %llu",
|
|
|
|
block_group->key.objectid);
|
2011-08-30 22:19:10 +08:00
|
|
|
#endif
|
2015-04-05 08:14:42 +08:00
|
|
|
spin_lock(&block_group->lock);
|
|
|
|
block_group->disk_cache_state = BTRFS_DC_ERROR;
|
|
|
|
spin_unlock(&block_group->lock);
|
|
|
|
|
|
|
|
block_group->io_ctl.inode = NULL;
|
|
|
|
iput(inode);
|
2011-04-20 10:20:14 +08:00
|
|
|
}
|
|
|
|
|
2015-04-05 08:14:42 +08:00
|
|
|
/*
|
|
|
|
* if ret == 0 the caller is expected to call btrfs_wait_cache_io
|
|
|
|
* to wait for IO and put the inode
|
|
|
|
*/
|
|
|
|
|
2010-07-03 00:14:14 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
static inline unsigned long offset_to_bit(u64 bitmap_start, u32 unit,
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
u64 offset)
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
{
|
2013-08-27 05:14:08 +08:00
|
|
|
ASSERT(offset >= bitmap_start);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
offset -= bitmap_start;
|
2011-03-29 13:46:06 +08:00
|
|
|
return (unsigned long)(div_u64(offset, unit));
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
}
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
static inline unsigned long bytes_to_bits(u64 bytes, u32 unit)
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
return (unsigned long)(div_u64(bytes, unit));
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
}
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
static inline u64 offset_to_bitmap(struct btrfs_free_space_ctl *ctl,
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
u64 offset)
|
|
|
|
{
|
|
|
|
u64 bitmap_start;
|
2015-01-17 00:26:13 +08:00
|
|
|
u32 bytes_per_bitmap;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
bytes_per_bitmap = BITS_PER_BITMAP * ctl->unit;
|
|
|
|
bitmap_start = offset - ctl->start;
|
2015-01-17 00:26:13 +08:00
|
|
|
bitmap_start = div_u64(bitmap_start, bytes_per_bitmap);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
bitmap_start *= bytes_per_bitmap;
|
2011-03-29 13:46:06 +08:00
|
|
|
bitmap_start += ctl->start;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
return bitmap_start;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
}
|
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
static int tree_insert_offset(struct rb_root *root, u64 offset,
|
|
|
|
struct rb_node *node, int bitmap)
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
{
|
|
|
|
struct rb_node **p = &root->rb_node;
|
|
|
|
struct rb_node *parent = NULL;
|
|
|
|
struct btrfs_free_space *info;
|
|
|
|
|
|
|
|
while (*p) {
|
|
|
|
parent = *p;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
info = rb_entry(parent, struct btrfs_free_space, offset_index);
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
if (offset < info->offset) {
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
p = &(*p)->rb_left;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
} else if (offset > info->offset) {
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
p = &(*p)->rb_right;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* we could have a bitmap entry and an extent entry
|
|
|
|
* share the same offset. If this is the case, we want
|
|
|
|
* the extent entry to always be found first if we do a
|
|
|
|
* linear search through the tree, since we want to have
|
|
|
|
* the quickest allocation time, and allocating from an
|
|
|
|
* extent is faster than allocating from a bitmap. So
|
|
|
|
* if we're inserting a bitmap and we find an entry at
|
|
|
|
* this offset, we want to go right, or after this entry
|
|
|
|
* logically. If we are inserting an extent and we've
|
|
|
|
* found a bitmap, we want to go left, or before
|
|
|
|
* logically.
|
|
|
|
*/
|
|
|
|
if (bitmap) {
|
2011-05-14 02:49:23 +08:00
|
|
|
if (info->bitmap) {
|
|
|
|
WARN_ON_ONCE(1);
|
|
|
|
return -EEXIST;
|
|
|
|
}
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
p = &(*p)->rb_right;
|
|
|
|
} else {
|
2011-05-14 02:49:23 +08:00
|
|
|
if (!info->bitmap) {
|
|
|
|
WARN_ON_ONCE(1);
|
|
|
|
return -EEXIST;
|
|
|
|
}
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
p = &(*p)->rb_left;
|
|
|
|
}
|
|
|
|
}
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
rb_link_node(node, parent, p);
|
|
|
|
rb_insert_color(node, root);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2009-04-03 22:14:19 +08:00
|
|
|
* searches the tree for the given offset.
|
|
|
|
*
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
* fuzzy - If this is set, then we are trying to make an allocation, and we just
|
|
|
|
* want a section that has at least bytes size and comes at or after the given
|
|
|
|
* offset.
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
*/
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
static struct btrfs_free_space *
|
2011-03-29 13:46:06 +08:00
|
|
|
tree_search_offset(struct btrfs_free_space_ctl *ctl,
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
u64 offset, int bitmap_only, int fuzzy)
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct rb_node *n = ctl->free_space_offset.rb_node;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
struct btrfs_free_space *entry, *prev = NULL;
|
|
|
|
|
|
|
|
/* find entry that is closest to the 'offset' */
|
|
|
|
while (1) {
|
|
|
|
if (!n) {
|
|
|
|
entry = NULL;
|
|
|
|
break;
|
|
|
|
}
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
|
|
|
entry = rb_entry(n, struct btrfs_free_space, offset_index);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
prev = entry;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
if (offset < entry->offset)
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
n = n->rb_left;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
else if (offset > entry->offset)
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
n = n->rb_right;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
else
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
if (bitmap_only) {
|
|
|
|
if (!entry)
|
|
|
|
return NULL;
|
|
|
|
if (entry->bitmap)
|
|
|
|
return entry;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
/*
|
|
|
|
* bitmap entry and extent entry may share same offset,
|
|
|
|
* in that case, bitmap entry comes after extent entry.
|
|
|
|
*/
|
|
|
|
n = rb_next(n);
|
|
|
|
if (!n)
|
|
|
|
return NULL;
|
|
|
|
entry = rb_entry(n, struct btrfs_free_space, offset_index);
|
|
|
|
if (entry->offset != offset)
|
|
|
|
return NULL;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
WARN_ON(!entry->bitmap);
|
|
|
|
return entry;
|
|
|
|
} else if (entry) {
|
|
|
|
if (entry->bitmap) {
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
/*
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
* if previous extent entry covers the offset,
|
|
|
|
* we should return it instead of the bitmap entry
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
*/
|
2012-10-18 16:18:01 +08:00
|
|
|
n = rb_prev(&entry->offset_index);
|
|
|
|
if (n) {
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
prev = rb_entry(n, struct btrfs_free_space,
|
|
|
|
offset_index);
|
2012-10-18 16:18:01 +08:00
|
|
|
if (!prev->bitmap &&
|
|
|
|
prev->offset + prev->bytes > offset)
|
|
|
|
entry = prev;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
}
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
}
|
|
|
|
return entry;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!prev)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
/* find last entry before the 'offset' */
|
|
|
|
entry = prev;
|
|
|
|
if (entry->offset > offset) {
|
|
|
|
n = rb_prev(&entry->offset_index);
|
|
|
|
if (n) {
|
|
|
|
entry = rb_entry(n, struct btrfs_free_space,
|
|
|
|
offset_index);
|
2013-08-27 05:14:08 +08:00
|
|
|
ASSERT(entry->offset <= offset);
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
} else {
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
if (fuzzy)
|
|
|
|
return entry;
|
|
|
|
else
|
|
|
|
return NULL;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
if (entry->bitmap) {
|
2012-10-18 16:18:01 +08:00
|
|
|
n = rb_prev(&entry->offset_index);
|
|
|
|
if (n) {
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
prev = rb_entry(n, struct btrfs_free_space,
|
|
|
|
offset_index);
|
2012-10-18 16:18:01 +08:00
|
|
|
if (!prev->bitmap &&
|
|
|
|
prev->offset + prev->bytes > offset)
|
|
|
|
return prev;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
}
|
2011-03-29 13:46:06 +08:00
|
|
|
if (entry->offset + BITS_PER_BITMAP * ctl->unit > offset)
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
return entry;
|
|
|
|
} else if (entry->offset + entry->bytes > offset)
|
|
|
|
return entry;
|
|
|
|
|
|
|
|
if (!fuzzy)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
while (1) {
|
|
|
|
if (entry->bitmap) {
|
|
|
|
if (entry->offset + BITS_PER_BITMAP *
|
2011-03-29 13:46:06 +08:00
|
|
|
ctl->unit > offset)
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
break;
|
|
|
|
} else {
|
|
|
|
if (entry->offset + entry->bytes > offset)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
n = rb_next(&entry->offset_index);
|
|
|
|
if (!n)
|
|
|
|
return NULL;
|
|
|
|
entry = rb_entry(n, struct btrfs_free_space, offset_index);
|
|
|
|
}
|
|
|
|
return entry;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
}
|
|
|
|
|
2010-11-09 14:57:39 +08:00
|
|
|
static inline void
|
2011-03-29 13:46:06 +08:00
|
|
|
__unlink_free_space(struct btrfs_free_space_ctl *ctl,
|
2010-11-09 14:57:39 +08:00
|
|
|
struct btrfs_free_space *info)
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
rb_erase(&info->offset_index, &ctl->free_space_offset);
|
|
|
|
ctl->free_extents--;
|
2010-11-09 14:57:39 +08:00
|
|
|
}
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
static void unlink_free_space(struct btrfs_free_space_ctl *ctl,
|
2010-11-09 14:57:39 +08:00
|
|
|
struct btrfs_free_space *info)
|
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
__unlink_free_space(ctl, info);
|
|
|
|
ctl->free_space -= info->bytes;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
}
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
static int link_free_space(struct btrfs_free_space_ctl *ctl,
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
struct btrfs_free_space *info)
|
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
|
2013-08-27 05:14:08 +08:00
|
|
|
ASSERT(info->bytes || info->bitmap);
|
2011-03-29 13:46:06 +08:00
|
|
|
ret = tree_insert_offset(&ctl->free_space_offset, info->offset,
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
&info->offset_index, (info->bitmap != NULL));
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
ctl->free_space += info->bytes;
|
|
|
|
ctl->free_extents++;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl)
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_block_group_cache *block_group = ctl->private;
|
2009-09-12 04:11:20 +08:00
|
|
|
u64 max_bytes;
|
|
|
|
u64 bitmap_bytes;
|
|
|
|
u64 extent_bytes;
|
2010-11-09 14:48:01 +08:00
|
|
|
u64 size = block_group->key.offset;
|
2015-01-17 00:26:13 +08:00
|
|
|
u32 bytes_per_bg = BITS_PER_BITMAP * ctl->unit;
|
|
|
|
u32 max_bitmaps = div_u64(size + bytes_per_bg - 1, bytes_per_bg);
|
2011-03-29 13:46:06 +08:00
|
|
|
|
2015-01-17 00:26:13 +08:00
|
|
|
max_bitmaps = max_t(u32, max_bitmaps, 1);
|
2013-02-13 03:07:51 +08:00
|
|
|
|
2013-08-27 05:14:08 +08:00
|
|
|
ASSERT(ctl->total_bitmaps <= max_bitmaps);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* The goal is to keep the total amount of memory used per 1gb of space
|
|
|
|
* at or below 32k, so we need to adjust how much memory we allow to be
|
|
|
|
* used by extent based free space tracking
|
|
|
|
*/
|
2015-12-15 00:42:10 +08:00
|
|
|
if (size < SZ_1G)
|
2010-11-09 14:48:01 +08:00
|
|
|
max_bytes = MAX_CACHE_BYTES_PER_GIG;
|
|
|
|
else
|
2015-12-15 00:42:10 +08:00
|
|
|
max_bytes = MAX_CACHE_BYTES_PER_GIG * div_u64(size, SZ_1G);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2009-09-12 04:11:20 +08:00
|
|
|
/*
|
|
|
|
* we want to account for 1 more bitmap than what we have so we can make
|
|
|
|
* sure we don't go over our overall goal of MAX_CACHE_BYTES_PER_GIG as
|
|
|
|
* we add more bitmaps.
|
|
|
|
*/
|
2011-03-29 13:46:06 +08:00
|
|
|
bitmap_bytes = (ctl->total_bitmaps + 1) * PAGE_CACHE_SIZE;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2009-09-12 04:11:20 +08:00
|
|
|
if (bitmap_bytes >= max_bytes) {
|
2011-03-29 13:46:06 +08:00
|
|
|
ctl->extents_thresh = 0;
|
2009-09-12 04:11:20 +08:00
|
|
|
return;
|
|
|
|
}
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2009-09-12 04:11:20 +08:00
|
|
|
/*
|
2015-01-17 00:21:12 +08:00
|
|
|
* we want the extent entry threshold to always be at most 1/2 the max
|
2009-09-12 04:11:20 +08:00
|
|
|
* bytes we can have, or whatever is less than that.
|
|
|
|
*/
|
|
|
|
extent_bytes = max_bytes - bitmap_bytes;
|
2015-01-17 00:21:12 +08:00
|
|
|
extent_bytes = min_t(u64, extent_bytes, max_bytes >> 1);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
ctl->extents_thresh =
|
2015-01-17 00:21:12 +08:00
|
|
|
div_u64(extent_bytes, sizeof(struct btrfs_free_space));
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
}
|
|
|
|
|
2011-08-05 17:32:35 +08:00
|
|
|
static inline void __bitmap_clear_bits(struct btrfs_free_space_ctl *ctl,
|
|
|
|
struct btrfs_free_space *info,
|
|
|
|
u64 offset, u64 bytes)
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
{
|
2011-03-14 13:40:51 +08:00
|
|
|
unsigned long start, count;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
start = offset_to_bit(info->offset, ctl->unit, offset);
|
|
|
|
count = bytes_to_bits(bytes, ctl->unit);
|
2013-08-27 05:14:08 +08:00
|
|
|
ASSERT(start + count <= BITS_PER_BITMAP);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2011-03-14 13:40:51 +08:00
|
|
|
bitmap_clear(info->bitmap, start, count);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
|
|
|
info->bytes -= bytes;
|
2011-08-05 17:32:35 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void bitmap_clear_bits(struct btrfs_free_space_ctl *ctl,
|
|
|
|
struct btrfs_free_space *info, u64 offset,
|
|
|
|
u64 bytes)
|
|
|
|
{
|
|
|
|
__bitmap_clear_bits(ctl, info, offset, bytes);
|
2011-03-29 13:46:06 +08:00
|
|
|
ctl->free_space -= bytes;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
}
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
static void bitmap_set_bits(struct btrfs_free_space_ctl *ctl,
|
Btrfs: async block group caching
This patch moves the caching of the block group off to a kthread in order to
allow people to allocate sooner. Instead of blocking up behind the caching
mutex, we instead kick of the caching kthread, and then attempt to make an
allocation. If we cannot, we wait on the block groups caching waitqueue, which
the caching kthread will wake the waiting threads up everytime it finds 2 meg
worth of space, and then again when its finished caching. This is how I tested
the speedup from this
mkfs the disk
mount the disk
fill the disk up with fs_mark
unmount the disk
mount the disk
time touch /mnt/foo
Without my changes this took 11 seconds on my box, with these changes it now
takes 1 second.
Another change thats been put in place is we lock the super mirror's in the
pinned extent map in order to keep us from adding that stuff as free space when
caching the block group. This doesn't really change anything else as far as the
pinned extent map is concerned, since for actual pinned extents we use
EXTENT_DIRTY, but it does mean that when we unmount we have to go in and unlock
those extents to keep from leaking memory.
I've also added a check where when we are reading block groups from disk, if the
amount of space used == the size of the block group, we go ahead and mark the
block group as cached. This drastically reduces the amount of time it takes to
cache the block groups. Using the same test as above, except doing a dd to a
file and then unmounting, it used to take 33 seconds to umount, now it takes 3
seconds.
This version uses the commit_root in the caching kthread, and then keeps track
of how many async caching threads are running at any given time so if one of the
async threads is still running as we cross transactions we can wait until its
finished before handling the pinned extents. Thank you,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
struct btrfs_free_space *info, u64 offset,
|
|
|
|
u64 bytes)
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
{
|
2011-03-14 13:40:51 +08:00
|
|
|
unsigned long start, count;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
start = offset_to_bit(info->offset, ctl->unit, offset);
|
|
|
|
count = bytes_to_bits(bytes, ctl->unit);
|
2013-08-27 05:14:08 +08:00
|
|
|
ASSERT(start + count <= BITS_PER_BITMAP);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2011-03-14 13:40:51 +08:00
|
|
|
bitmap_set(info->bitmap, start, count);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
|
|
|
info->bytes += bytes;
|
2011-03-29 13:46:06 +08:00
|
|
|
ctl->free_space += bytes;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
}
|
|
|
|
|
2013-09-09 13:19:42 +08:00
|
|
|
/*
|
|
|
|
* If we can not find suitable extent, we will use bytes to record
|
|
|
|
* the size of the max extent.
|
|
|
|
*/
|
2011-03-29 13:46:06 +08:00
|
|
|
static int search_bitmap(struct btrfs_free_space_ctl *ctl,
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
struct btrfs_free_space *bitmap_info, u64 *offset,
|
2015-10-03 04:12:23 +08:00
|
|
|
u64 *bytes, bool for_alloc)
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
{
|
|
|
|
unsigned long found_bits = 0;
|
2013-09-09 13:19:42 +08:00
|
|
|
unsigned long max_bits = 0;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
unsigned long bits, i;
|
|
|
|
unsigned long next_zero;
|
2013-09-09 13:19:42 +08:00
|
|
|
unsigned long extent_bits;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2015-10-03 04:09:42 +08:00
|
|
|
/*
|
|
|
|
* Skip searching the bitmap if we don't have a contiguous section that
|
|
|
|
* is large enough for this allocation.
|
|
|
|
*/
|
2015-10-03 04:12:23 +08:00
|
|
|
if (for_alloc &&
|
|
|
|
bitmap_info->max_extent_size &&
|
2015-10-03 04:09:42 +08:00
|
|
|
bitmap_info->max_extent_size < *bytes) {
|
|
|
|
*bytes = bitmap_info->max_extent_size;
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
i = offset_to_bit(bitmap_info->offset, ctl->unit,
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
max_t(u64, *offset, bitmap_info->offset));
|
2011-03-29 13:46:06 +08:00
|
|
|
bits = bytes_to_bits(*bytes, ctl->unit);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2012-09-14 10:29:02 +08:00
|
|
|
for_each_set_bit_from(i, bitmap_info->bitmap, BITS_PER_BITMAP) {
|
2015-10-03 04:12:23 +08:00
|
|
|
if (for_alloc && bits == 1) {
|
|
|
|
found_bits = 1;
|
|
|
|
break;
|
|
|
|
}
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
next_zero = find_next_zero_bit(bitmap_info->bitmap,
|
|
|
|
BITS_PER_BITMAP, i);
|
2013-09-09 13:19:42 +08:00
|
|
|
extent_bits = next_zero - i;
|
|
|
|
if (extent_bits >= bits) {
|
|
|
|
found_bits = extent_bits;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
break;
|
2013-09-09 13:19:42 +08:00
|
|
|
} else if (extent_bits > max_bits) {
|
|
|
|
max_bits = extent_bits;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
}
|
|
|
|
i = next_zero;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (found_bits) {
|
2011-03-29 13:46:06 +08:00
|
|
|
*offset = (u64)(i * ctl->unit) + bitmap_info->offset;
|
|
|
|
*bytes = (u64)(found_bits) * ctl->unit;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2013-09-09 13:19:42 +08:00
|
|
|
*bytes = (u64)(max_bits) * ctl->unit;
|
2015-10-03 04:09:42 +08:00
|
|
|
bitmap_info->max_extent_size = *bytes;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
2013-09-09 13:19:42 +08:00
|
|
|
/* Cache the size of the max extent in bytes */
|
2011-03-29 13:46:06 +08:00
|
|
|
static struct btrfs_free_space *
|
2013-01-30 07:40:14 +08:00
|
|
|
find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes,
|
2013-09-09 13:19:42 +08:00
|
|
|
unsigned long align, u64 *max_extent_size)
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
{
|
|
|
|
struct btrfs_free_space *entry;
|
|
|
|
struct rb_node *node;
|
2013-01-30 07:40:14 +08:00
|
|
|
u64 tmp;
|
|
|
|
u64 align_off;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
int ret;
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
if (!ctl->free_space_offset.rb_node)
|
2013-09-09 13:19:42 +08:00
|
|
|
goto out;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
entry = tree_search_offset(ctl, offset_to_bitmap(ctl, *offset), 0, 1);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
if (!entry)
|
2013-09-09 13:19:42 +08:00
|
|
|
goto out;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
|
|
|
for (node = &entry->offset_index; node; node = rb_next(node)) {
|
|
|
|
entry = rb_entry(node, struct btrfs_free_space, offset_index);
|
2013-09-09 13:19:42 +08:00
|
|
|
if (entry->bytes < *bytes) {
|
|
|
|
if (entry->bytes > *max_extent_size)
|
|
|
|
*max_extent_size = entry->bytes;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
continue;
|
2013-09-09 13:19:42 +08:00
|
|
|
}
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2013-01-30 07:40:14 +08:00
|
|
|
/* make sure the space returned is big enough
|
|
|
|
* to match our requested alignment
|
|
|
|
*/
|
|
|
|
if (*bytes >= align) {
|
2013-09-09 13:19:42 +08:00
|
|
|
tmp = entry->offset - ctl->start + align - 1;
|
2015-02-21 01:43:47 +08:00
|
|
|
tmp = div64_u64(tmp, align);
|
2013-01-30 07:40:14 +08:00
|
|
|
tmp = tmp * align + ctl->start;
|
|
|
|
align_off = tmp - entry->offset;
|
|
|
|
} else {
|
|
|
|
align_off = 0;
|
|
|
|
tmp = entry->offset;
|
|
|
|
}
|
|
|
|
|
2013-09-09 13:19:42 +08:00
|
|
|
if (entry->bytes < *bytes + align_off) {
|
|
|
|
if (entry->bytes > *max_extent_size)
|
|
|
|
*max_extent_size = entry->bytes;
|
2013-01-30 07:40:14 +08:00
|
|
|
continue;
|
2013-09-09 13:19:42 +08:00
|
|
|
}
|
2013-01-30 07:40:14 +08:00
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
if (entry->bitmap) {
|
2013-09-09 13:19:42 +08:00
|
|
|
u64 size = *bytes;
|
|
|
|
|
2015-10-03 04:12:23 +08:00
|
|
|
ret = search_bitmap(ctl, entry, &tmp, &size, true);
|
2013-01-30 07:40:14 +08:00
|
|
|
if (!ret) {
|
|
|
|
*offset = tmp;
|
2013-09-09 13:19:42 +08:00
|
|
|
*bytes = size;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
return entry;
|
2013-09-09 13:19:42 +08:00
|
|
|
} else if (size > *max_extent_size) {
|
|
|
|
*max_extent_size = size;
|
2013-01-30 07:40:14 +08:00
|
|
|
}
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
2013-01-30 07:40:14 +08:00
|
|
|
*offset = tmp;
|
|
|
|
*bytes = entry->bytes - align_off;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
return entry;
|
|
|
|
}
|
2013-09-09 13:19:42 +08:00
|
|
|
out:
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
static void add_new_bitmap(struct btrfs_free_space_ctl *ctl,
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
struct btrfs_free_space *info, u64 offset)
|
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
info->offset = offset_to_bitmap(ctl, offset);
|
2009-09-12 04:11:20 +08:00
|
|
|
info->bytes = 0;
|
Btrfs: initialize new bitmaps' list
We're failing to create clusters with bitmaps because
setup_cluster_no_bitmap checks that the list is empty before inserting
the bitmap entry in the list for setup_cluster_bitmap, but the list
field is only initialized when it is restored from the on-disk free
space cache, or when it is written out to disk.
Besides a potential race condition due to the multiple use of the list
field, filesystem performance severely degrades over time: as we use
up all non-bitmap free extents, the try-to-set-up-cluster dance is
done at every metadata block allocation. For every block group, we
fail to set up a cluster, and after failing on them all up to twice,
we fall back to the much slower unclustered allocation.
To make matters worse, before the unclustered allocation, we try to
create new block groups until we reach the 1% threshold, which
introduces additional bitmaps and thus block groups that we'll iterate
over at each metadata block request.
2011-11-28 22:04:43 +08:00
|
|
|
INIT_LIST_HEAD(&info->list);
|
2011-03-29 13:46:06 +08:00
|
|
|
link_free_space(ctl, info);
|
|
|
|
ctl->total_bitmaps++;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
ctl->op->recalc_thresholds(ctl);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
}
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
static void free_bitmap(struct btrfs_free_space_ctl *ctl,
|
2010-11-09 14:50:07 +08:00
|
|
|
struct btrfs_free_space *bitmap_info)
|
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
unlink_free_space(ctl, bitmap_info);
|
2010-11-09 14:50:07 +08:00
|
|
|
kfree(bitmap_info->bitmap);
|
2011-01-29 06:05:48 +08:00
|
|
|
kmem_cache_free(btrfs_free_space_cachep, bitmap_info);
|
2011-03-29 13:46:06 +08:00
|
|
|
ctl->total_bitmaps--;
|
|
|
|
ctl->op->recalc_thresholds(ctl);
|
2010-11-09 14:50:07 +08:00
|
|
|
}
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
static noinline int remove_from_bitmap(struct btrfs_free_space_ctl *ctl,
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
struct btrfs_free_space *bitmap_info,
|
|
|
|
u64 *offset, u64 *bytes)
|
|
|
|
{
|
|
|
|
u64 end;
|
2009-07-31 23:03:58 +08:00
|
|
|
u64 search_start, search_bytes;
|
|
|
|
int ret;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
|
|
|
again:
|
2011-03-29 13:46:06 +08:00
|
|
|
end = bitmap_info->offset + (u64)(BITS_PER_BITMAP * ctl->unit) - 1;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2009-07-31 23:03:58 +08:00
|
|
|
/*
|
2012-06-28 03:10:56 +08:00
|
|
|
* We need to search for bits in this bitmap. We could only cover some
|
|
|
|
* of the extent in this bitmap thanks to how we add space, so we need
|
|
|
|
* to search for as much as it as we can and clear that amount, and then
|
|
|
|
* go searching for the next bit.
|
2009-07-31 23:03:58 +08:00
|
|
|
*/
|
|
|
|
search_start = *offset;
|
2012-06-28 03:10:56 +08:00
|
|
|
search_bytes = ctl->unit;
|
2011-02-03 10:39:52 +08:00
|
|
|
search_bytes = min(search_bytes, end - search_start + 1);
|
2015-10-03 04:12:23 +08:00
|
|
|
ret = search_bitmap(ctl, bitmap_info, &search_start, &search_bytes,
|
|
|
|
false);
|
2013-04-26 03:55:30 +08:00
|
|
|
if (ret < 0 || search_start != *offset)
|
|
|
|
return -EINVAL;
|
2009-07-31 23:03:58 +08:00
|
|
|
|
2012-06-28 03:10:56 +08:00
|
|
|
/* We may have found more bits than what we need */
|
|
|
|
search_bytes = min(search_bytes, *bytes);
|
|
|
|
|
|
|
|
/* Cannot clear past the end of the bitmap */
|
|
|
|
search_bytes = min(search_bytes, end - search_start + 1);
|
|
|
|
|
|
|
|
bitmap_clear_bits(ctl, bitmap_info, search_start, search_bytes);
|
|
|
|
*offset += search_bytes;
|
|
|
|
*bytes -= search_bytes;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
|
|
|
if (*bytes) {
|
2009-07-31 23:03:58 +08:00
|
|
|
struct rb_node *next = rb_next(&bitmap_info->offset_index);
|
2010-11-09 14:50:07 +08:00
|
|
|
if (!bitmap_info->bytes)
|
2011-03-29 13:46:06 +08:00
|
|
|
free_bitmap(ctl, bitmap_info);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2009-07-31 23:03:58 +08:00
|
|
|
/*
|
|
|
|
* no entry after this bitmap, but we still have bytes to
|
|
|
|
* remove, so something has gone wrong.
|
|
|
|
*/
|
|
|
|
if (!next)
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
return -EINVAL;
|
|
|
|
|
2009-07-31 23:03:58 +08:00
|
|
|
bitmap_info = rb_entry(next, struct btrfs_free_space,
|
|
|
|
offset_index);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* if the next entry isn't a bitmap we need to return to let the
|
|
|
|
* extent stuff do its work.
|
|
|
|
*/
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
if (!bitmap_info->bitmap)
|
|
|
|
return -EAGAIN;
|
|
|
|
|
2009-07-31 23:03:58 +08:00
|
|
|
/*
|
|
|
|
* Ok the next item is a bitmap, but it may not actually hold
|
|
|
|
* the information for the rest of this free space stuff, so
|
|
|
|
* look for it, and if we don't find it return so we can try
|
|
|
|
* everything over again.
|
|
|
|
*/
|
|
|
|
search_start = *offset;
|
2012-06-28 03:10:56 +08:00
|
|
|
search_bytes = ctl->unit;
|
2011-03-29 13:46:06 +08:00
|
|
|
ret = search_bitmap(ctl, bitmap_info, &search_start,
|
2015-10-03 04:12:23 +08:00
|
|
|
&search_bytes, false);
|
2009-07-31 23:03:58 +08:00
|
|
|
if (ret < 0 || search_start != *offset)
|
|
|
|
return -EAGAIN;
|
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
goto again;
|
2010-11-09 14:50:07 +08:00
|
|
|
} else if (!bitmap_info->bytes)
|
2011-03-29 13:46:06 +08:00
|
|
|
free_bitmap(ctl, bitmap_info);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-05-28 02:07:49 +08:00
|
|
|
static u64 add_bytes_to_bitmap(struct btrfs_free_space_ctl *ctl,
|
|
|
|
struct btrfs_free_space *info, u64 offset,
|
|
|
|
u64 bytes)
|
|
|
|
{
|
|
|
|
u64 bytes_to_set = 0;
|
|
|
|
u64 end;
|
|
|
|
|
|
|
|
end = info->offset + (u64)(BITS_PER_BITMAP * ctl->unit);
|
|
|
|
|
|
|
|
bytes_to_set = min(end - offset, bytes);
|
|
|
|
|
|
|
|
bitmap_set_bits(ctl, info, offset, bytes_to_set);
|
|
|
|
|
2015-10-03 04:09:42 +08:00
|
|
|
/*
|
|
|
|
* We set some bytes, we have no idea what the max extent size is
|
|
|
|
* anymore.
|
|
|
|
*/
|
|
|
|
info->max_extent_size = 0;
|
|
|
|
|
2011-05-28 02:07:49 +08:00
|
|
|
return bytes_to_set;
|
|
|
|
|
|
|
|
}
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
static bool use_bitmap(struct btrfs_free_space_ctl *ctl,
|
|
|
|
struct btrfs_free_space *info)
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_block_group_cache *block_group = ctl->private;
|
2015-09-24 02:54:14 +08:00
|
|
|
bool forced = false;
|
|
|
|
|
|
|
|
#ifdef CONFIG_BTRFS_DEBUG
|
|
|
|
if (btrfs_should_fragment_free_space(block_group->fs_info->extent_root,
|
|
|
|
block_group))
|
|
|
|
forced = true;
|
|
|
|
#endif
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* If we are below the extents threshold then we can add this as an
|
|
|
|
* extent, and don't have to deal with the bitmap
|
|
|
|
*/
|
2015-09-24 02:54:14 +08:00
|
|
|
if (!forced && ctl->free_extents < ctl->extents_thresh) {
|
2011-03-19 04:16:21 +08:00
|
|
|
/*
|
|
|
|
* If this block group has some small extents we don't want to
|
|
|
|
* use up all of our free slots in the cache with them, we want
|
|
|
|
* to reserve them to larger extents, however if we have plent
|
|
|
|
* of cache left then go ahead an dadd them, no sense in adding
|
|
|
|
* the overhead of a bitmap if we don't have to.
|
|
|
|
*/
|
|
|
|
if (info->bytes <= block_group->sectorsize * 4) {
|
2011-03-29 13:46:06 +08:00
|
|
|
if (ctl->free_extents * 2 <= ctl->extents_thresh)
|
|
|
|
return false;
|
2011-03-19 04:16:21 +08:00
|
|
|
} else {
|
2011-03-29 13:46:06 +08:00
|
|
|
return false;
|
2011-03-19 04:16:21 +08:00
|
|
|
}
|
|
|
|
}
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
|
|
|
/*
|
2013-02-13 03:07:51 +08:00
|
|
|
* The original block groups from mkfs can be really small, like 8
|
|
|
|
* megabytes, so don't bother with a bitmap for those entries. However
|
|
|
|
* some block groups can be smaller than what a bitmap would cover but
|
|
|
|
* are still large enough that they could overflow the 32k memory limit,
|
|
|
|
* so allow those block groups to still be allowed to have a bitmap
|
|
|
|
* entry.
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
*/
|
2013-02-13 03:07:51 +08:00
|
|
|
if (((BITS_PER_BITMAP * ctl->unit) >> 1) > block_group->key.offset)
|
2011-03-29 13:46:06 +08:00
|
|
|
return false;
|
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
2015-11-19 18:42:28 +08:00
|
|
|
static const struct btrfs_free_space_op free_space_op = {
|
2011-05-28 02:07:49 +08:00
|
|
|
.recalc_thresholds = recalculate_thresholds,
|
|
|
|
.use_bitmap = use_bitmap,
|
|
|
|
};
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
static int insert_into_bitmap(struct btrfs_free_space_ctl *ctl,
|
|
|
|
struct btrfs_free_space *info)
|
|
|
|
{
|
|
|
|
struct btrfs_free_space *bitmap_info;
|
2011-05-28 02:07:49 +08:00
|
|
|
struct btrfs_block_group_cache *block_group = NULL;
|
2011-03-29 13:46:06 +08:00
|
|
|
int added = 0;
|
2011-05-28 02:07:49 +08:00
|
|
|
u64 bytes, offset, bytes_added;
|
2011-03-29 13:46:06 +08:00
|
|
|
int ret;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
|
|
|
bytes = info->bytes;
|
|
|
|
offset = info->offset;
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
if (!ctl->op->use_bitmap(ctl, info))
|
|
|
|
return 0;
|
|
|
|
|
2011-05-28 02:07:49 +08:00
|
|
|
if (ctl->op == &free_space_op)
|
|
|
|
block_group = ctl->private;
|
2011-06-11 04:36:57 +08:00
|
|
|
again:
|
2011-05-28 02:07:49 +08:00
|
|
|
/*
|
|
|
|
* Since we link bitmaps right into the cluster we need to see if we
|
|
|
|
* have a cluster here, and if so and it has our bitmap we need to add
|
|
|
|
* the free space to that bitmap.
|
|
|
|
*/
|
|
|
|
if (block_group && !list_empty(&block_group->cluster_list)) {
|
|
|
|
struct btrfs_free_cluster *cluster;
|
|
|
|
struct rb_node *node;
|
|
|
|
struct btrfs_free_space *entry;
|
|
|
|
|
|
|
|
cluster = list_entry(block_group->cluster_list.next,
|
|
|
|
struct btrfs_free_cluster,
|
|
|
|
block_group_list);
|
|
|
|
spin_lock(&cluster->lock);
|
|
|
|
node = rb_first(&cluster->root);
|
|
|
|
if (!node) {
|
|
|
|
spin_unlock(&cluster->lock);
|
2011-06-11 04:36:57 +08:00
|
|
|
goto no_cluster_bitmap;
|
2011-05-28 02:07:49 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
entry = rb_entry(node, struct btrfs_free_space, offset_index);
|
|
|
|
if (!entry->bitmap) {
|
|
|
|
spin_unlock(&cluster->lock);
|
2011-06-11 04:36:57 +08:00
|
|
|
goto no_cluster_bitmap;
|
2011-05-28 02:07:49 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
if (entry->offset == offset_to_bitmap(ctl, offset)) {
|
|
|
|
bytes_added = add_bytes_to_bitmap(ctl, entry,
|
|
|
|
offset, bytes);
|
|
|
|
bytes -= bytes_added;
|
|
|
|
offset += bytes_added;
|
|
|
|
}
|
|
|
|
spin_unlock(&cluster->lock);
|
|
|
|
if (!bytes) {
|
|
|
|
ret = 1;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
}
|
2011-06-11 04:36:57 +08:00
|
|
|
|
|
|
|
no_cluster_bitmap:
|
2011-03-29 13:46:06 +08:00
|
|
|
bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
1, 0);
|
|
|
|
if (!bitmap_info) {
|
2013-08-27 05:14:08 +08:00
|
|
|
ASSERT(added == 0);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
goto new_bitmap;
|
|
|
|
}
|
|
|
|
|
2011-05-28 02:07:49 +08:00
|
|
|
bytes_added = add_bytes_to_bitmap(ctl, bitmap_info, offset, bytes);
|
|
|
|
bytes -= bytes_added;
|
|
|
|
offset += bytes_added;
|
|
|
|
added = 0;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
|
|
|
if (!bytes) {
|
|
|
|
ret = 1;
|
|
|
|
goto out;
|
|
|
|
} else
|
|
|
|
goto again;
|
|
|
|
|
|
|
|
new_bitmap:
|
|
|
|
if (info && info->bitmap) {
|
2011-03-29 13:46:06 +08:00
|
|
|
add_new_bitmap(ctl, info, offset);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
added = 1;
|
|
|
|
info = NULL;
|
|
|
|
goto again;
|
|
|
|
} else {
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
|
|
|
/* no pre-allocated info, allocate a new one */
|
|
|
|
if (!info) {
|
2011-01-29 06:05:48 +08:00
|
|
|
info = kmem_cache_zalloc(btrfs_free_space_cachep,
|
|
|
|
GFP_NOFS);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
if (!info) {
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_lock(&ctl->tree_lock);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
ret = -ENOMEM;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* allocate the bitmap */
|
|
|
|
info->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_lock(&ctl->tree_lock);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
if (!info->bitmap) {
|
|
|
|
ret = -ENOMEM;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
goto again;
|
|
|
|
}
|
|
|
|
|
|
|
|
out:
|
|
|
|
if (info) {
|
|
|
|
if (info->bitmap)
|
|
|
|
kfree(info->bitmap);
|
2011-01-29 06:05:48 +08:00
|
|
|
kmem_cache_free(btrfs_free_space_cachep, info);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
}
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2011-05-23 00:33:42 +08:00
|
|
|
static bool try_merge_free_space(struct btrfs_free_space_ctl *ctl,
|
2010-11-09 14:57:39 +08:00
|
|
|
struct btrfs_free_space *info, bool update_stat)
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
{
|
2010-11-09 14:56:50 +08:00
|
|
|
struct btrfs_free_space *left_info;
|
|
|
|
struct btrfs_free_space *right_info;
|
|
|
|
bool merged = false;
|
|
|
|
u64 offset = info->offset;
|
|
|
|
u64 bytes = info->bytes;
|
2009-04-03 22:14:18 +08:00
|
|
|
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
/*
|
|
|
|
* first we want to see if there is free space adjacent to the range we
|
|
|
|
* are adding, if there is remove that struct and add a new one to
|
|
|
|
* cover the entire range
|
|
|
|
*/
|
2011-03-29 13:46:06 +08:00
|
|
|
right_info = tree_search_offset(ctl, offset + bytes, 0, 0);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
if (right_info && rb_prev(&right_info->offset_index))
|
|
|
|
left_info = rb_entry(rb_prev(&right_info->offset_index),
|
|
|
|
struct btrfs_free_space, offset_index);
|
|
|
|
else
|
2011-03-29 13:46:06 +08:00
|
|
|
left_info = tree_search_offset(ctl, offset - 1, 0, 0);
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
if (right_info && !right_info->bitmap) {
|
2010-11-09 14:57:39 +08:00
|
|
|
if (update_stat)
|
2011-03-29 13:46:06 +08:00
|
|
|
unlink_free_space(ctl, right_info);
|
2010-11-09 14:57:39 +08:00
|
|
|
else
|
2011-03-29 13:46:06 +08:00
|
|
|
__unlink_free_space(ctl, right_info);
|
2009-04-03 22:14:18 +08:00
|
|
|
info->bytes += right_info->bytes;
|
2011-01-29 06:05:48 +08:00
|
|
|
kmem_cache_free(btrfs_free_space_cachep, right_info);
|
2010-11-09 14:56:50 +08:00
|
|
|
merged = true;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
}
|
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
if (left_info && !left_info->bitmap &&
|
|
|
|
left_info->offset + left_info->bytes == offset) {
|
2010-11-09 14:57:39 +08:00
|
|
|
if (update_stat)
|
2011-03-29 13:46:06 +08:00
|
|
|
unlink_free_space(ctl, left_info);
|
2010-11-09 14:57:39 +08:00
|
|
|
else
|
2011-03-29 13:46:06 +08:00
|
|
|
__unlink_free_space(ctl, left_info);
|
2009-04-03 22:14:18 +08:00
|
|
|
info->offset = left_info->offset;
|
|
|
|
info->bytes += left_info->bytes;
|
2011-01-29 06:05:48 +08:00
|
|
|
kmem_cache_free(btrfs_free_space_cachep, left_info);
|
2010-11-09 14:56:50 +08:00
|
|
|
merged = true;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
}
|
|
|
|
|
2010-11-09 14:56:50 +08:00
|
|
|
return merged;
|
|
|
|
}
|
|
|
|
|
Btrfs: improve free space cache management and space allocation
While under random IO, a block group's free space cache eventually reaches
a state where it has a mix of extent entries and bitmap entries representing
free space regions.
As later free space regions are returned to the cache, some of them are merged
with existing extent entries if they are contiguous with them. But others are
not merged, because despite the existence of adjacent free space regions in
the cache, the merging doesn't happen because the existing free space regions
are represented in bitmap extents. Even when new free space regions are merged
with existing extent entries (enlarging the free space range they represent),
we create chances of having after an enlarged region that is contiguous with
some other region represented in a bitmap entry.
Both clustered and non-clustered space allocation work by iterating over our
extent and bitmap entries and skipping any that represents a region smaller
then the allocation request (and giving preference to extent entries before
bitmap entries). By having a contiguous free space region that is represented
by 2 (or more) entries (mix of extent and bitmap entries), we end up not
satisfying an allocation request with a size larger than the size of any of
the entries but no larger than the sum of their sizes. Making the caller assume
we're under a ENOSPC condition or force it to allocate multiple smaller space
regions (as we do for file data writes), which adds extra overhead and more
chances of causing fragmentation due to the smaller regions being all spread
apart from each other (more likely when under concurrency).
For example, if we have the following in the cache:
* extent entry representing free space range: [128Mb - 256Kb, 128Mb[
* bitmap entry covering the range [128Mb, 256Mb[, but only with the bits
representing the range [128Mb, 128Mb + 768Kb[ set - that is, only that
space in this 128Mb area is marked as free
An allocation request for 1Mb, starting at offset not greater than 128Mb - 256Kb,
would fail before, despite the existence of such contiguous free space area in the
cache. The caller could only allocate up to 768Kb of space at once and later another
256Kb (or vice-versa). In between each smaller allocation request, another task
working on a different file/inode might come in and take that space, preventing the
former task of getting a contiguous 1Mb region of free space.
Therefore this change implements the ability to move free space from bitmap
entries into existing and new free space regions represented with extent
entries. This is done when a space region is added to the cache.
A test was added to the sanity tests that explains in detail the issue too.
Some performance test results with compilebench on a 4 cores machine, with
32Gb of ram and using an HDD follow.
Test: compilebench -D /mnt -i 30 -r 1000 --makej
Before this change:
intial create total runs 30 avg 69.02 MB/s (user 0.28s sys 0.57s)
compile total runs 30 avg 314.96 MB/s (user 0.12s sys 0.25s)
read compiled tree total runs 3 avg 27.14 MB/s (user 1.52s sys 0.90s)
delete compiled tree total runs 30 avg 3.14 seconds (user 0.15s sys 0.66s)
After this change:
intial create total runs 30 avg 68.37 MB/s (user 0.29s sys 0.55s)
compile total runs 30 avg 382.83 MB/s (user 0.12s sys 0.24s)
read compiled tree total runs 3 avg 27.82 MB/s (user 1.45s sys 0.97s)
delete compiled tree total runs 30 avg 3.18 seconds (user 0.17s sys 0.65s)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-08-29 20:35:13 +08:00
|
|
|
static bool steal_from_bitmap_to_end(struct btrfs_free_space_ctl *ctl,
|
|
|
|
struct btrfs_free_space *info,
|
|
|
|
bool update_stat)
|
|
|
|
{
|
|
|
|
struct btrfs_free_space *bitmap;
|
|
|
|
unsigned long i;
|
|
|
|
unsigned long j;
|
|
|
|
const u64 end = info->offset + info->bytes;
|
|
|
|
const u64 bitmap_offset = offset_to_bitmap(ctl, end);
|
|
|
|
u64 bytes;
|
|
|
|
|
|
|
|
bitmap = tree_search_offset(ctl, bitmap_offset, 1, 0);
|
|
|
|
if (!bitmap)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
i = offset_to_bit(bitmap->offset, ctl->unit, end);
|
|
|
|
j = find_next_zero_bit(bitmap->bitmap, BITS_PER_BITMAP, i);
|
|
|
|
if (j == i)
|
|
|
|
return false;
|
|
|
|
bytes = (j - i) * ctl->unit;
|
|
|
|
info->bytes += bytes;
|
|
|
|
|
|
|
|
if (update_stat)
|
|
|
|
bitmap_clear_bits(ctl, bitmap, end, bytes);
|
|
|
|
else
|
|
|
|
__bitmap_clear_bits(ctl, bitmap, end, bytes);
|
|
|
|
|
|
|
|
if (!bitmap->bytes)
|
|
|
|
free_bitmap(ctl, bitmap);
|
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
static bool steal_from_bitmap_to_front(struct btrfs_free_space_ctl *ctl,
|
|
|
|
struct btrfs_free_space *info,
|
|
|
|
bool update_stat)
|
|
|
|
{
|
|
|
|
struct btrfs_free_space *bitmap;
|
|
|
|
u64 bitmap_offset;
|
|
|
|
unsigned long i;
|
|
|
|
unsigned long j;
|
|
|
|
unsigned long prev_j;
|
|
|
|
u64 bytes;
|
|
|
|
|
|
|
|
bitmap_offset = offset_to_bitmap(ctl, info->offset);
|
|
|
|
/* If we're on a boundary, try the previous logical bitmap. */
|
|
|
|
if (bitmap_offset == info->offset) {
|
|
|
|
if (info->offset == 0)
|
|
|
|
return false;
|
|
|
|
bitmap_offset = offset_to_bitmap(ctl, info->offset - 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
bitmap = tree_search_offset(ctl, bitmap_offset, 1, 0);
|
|
|
|
if (!bitmap)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
i = offset_to_bit(bitmap->offset, ctl->unit, info->offset) - 1;
|
|
|
|
j = 0;
|
|
|
|
prev_j = (unsigned long)-1;
|
|
|
|
for_each_clear_bit_from(j, bitmap->bitmap, BITS_PER_BITMAP) {
|
|
|
|
if (j > i)
|
|
|
|
break;
|
|
|
|
prev_j = j;
|
|
|
|
}
|
|
|
|
if (prev_j == i)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
if (prev_j == (unsigned long)-1)
|
|
|
|
bytes = (i + 1) * ctl->unit;
|
|
|
|
else
|
|
|
|
bytes = (i - prev_j) * ctl->unit;
|
|
|
|
|
|
|
|
info->offset -= bytes;
|
|
|
|
info->bytes += bytes;
|
|
|
|
|
|
|
|
if (update_stat)
|
|
|
|
bitmap_clear_bits(ctl, bitmap, info->offset, bytes);
|
|
|
|
else
|
|
|
|
__bitmap_clear_bits(ctl, bitmap, info->offset, bytes);
|
|
|
|
|
|
|
|
if (!bitmap->bytes)
|
|
|
|
free_bitmap(ctl, bitmap);
|
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We prefer always to allocate from extent entries, both for clustered and
|
|
|
|
* non-clustered allocation requests. So when attempting to add a new extent
|
|
|
|
* entry, try to see if there's adjacent free space in bitmap entries, and if
|
|
|
|
* there is, migrate that space from the bitmaps to the extent.
|
|
|
|
* Like this we get better chances of satisfying space allocation requests
|
|
|
|
* because we attempt to satisfy them based on a single cache entry, and never
|
|
|
|
* on 2 or more entries - even if the entries represent a contiguous free space
|
|
|
|
* region (e.g. 1 extent entry + 1 bitmap entry starting where the extent entry
|
|
|
|
* ends).
|
|
|
|
*/
|
|
|
|
static void steal_from_bitmap(struct btrfs_free_space_ctl *ctl,
|
|
|
|
struct btrfs_free_space *info,
|
|
|
|
bool update_stat)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* Only work with disconnected entries, as we can change their offset,
|
|
|
|
* and must be extent entries.
|
|
|
|
*/
|
|
|
|
ASSERT(!info->bitmap);
|
|
|
|
ASSERT(RB_EMPTY_NODE(&info->offset_index));
|
|
|
|
|
|
|
|
if (ctl->total_bitmaps > 0) {
|
|
|
|
bool stole_end;
|
|
|
|
bool stole_front = false;
|
|
|
|
|
|
|
|
stole_end = steal_from_bitmap_to_end(ctl, info, update_stat);
|
|
|
|
if (ctl->total_bitmaps > 0)
|
|
|
|
stole_front = steal_from_bitmap_to_front(ctl, info,
|
|
|
|
update_stat);
|
|
|
|
|
|
|
|
if (stole_end || stole_front)
|
|
|
|
try_merge_free_space(ctl, info, update_stat);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
Btrfs: Cache free inode numbers in memory
Currently btrfs stores the highest objectid of the fs tree, and it always
returns (highest+1) inode number when we create a file, so inode numbers
won't be reclaimed when we delete files, so we'll run out of inode numbers
as we keep create/delete files in 32bits machines.
This fixes it, and it works similarly to how we cache free space in block
cgroups.
We start a kernel thread to read the file tree. By scanning inode items,
we know which chunks of inode numbers are free, and we cache them in
an rb-tree.
Because we are searching the commit root, we have to carefully handle the
cross-transaction case.
The rb-tree is a hybrid extent+bitmap tree, so if we have too many small
chunks of inode numbers, we'll use bitmaps. Initially we allow 16K ram
of extents, and a bitmap will be used if we exceed this threshold. The
extents threshold is adjusted in runtime.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-04-20 10:06:11 +08:00
|
|
|
int __btrfs_add_free_space(struct btrfs_free_space_ctl *ctl,
|
|
|
|
u64 offset, u64 bytes)
|
2010-11-09 14:56:50 +08:00
|
|
|
{
|
|
|
|
struct btrfs_free_space *info;
|
|
|
|
int ret = 0;
|
|
|
|
|
2011-01-29 06:05:48 +08:00
|
|
|
info = kmem_cache_zalloc(btrfs_free_space_cachep, GFP_NOFS);
|
2010-11-09 14:56:50 +08:00
|
|
|
if (!info)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
info->offset = offset;
|
|
|
|
info->bytes = bytes;
|
Btrfs: improve free space cache management and space allocation
While under random IO, a block group's free space cache eventually reaches
a state where it has a mix of extent entries and bitmap entries representing
free space regions.
As later free space regions are returned to the cache, some of them are merged
with existing extent entries if they are contiguous with them. But others are
not merged, because despite the existence of adjacent free space regions in
the cache, the merging doesn't happen because the existing free space regions
are represented in bitmap extents. Even when new free space regions are merged
with existing extent entries (enlarging the free space range they represent),
we create chances of having after an enlarged region that is contiguous with
some other region represented in a bitmap entry.
Both clustered and non-clustered space allocation work by iterating over our
extent and bitmap entries and skipping any that represents a region smaller
then the allocation request (and giving preference to extent entries before
bitmap entries). By having a contiguous free space region that is represented
by 2 (or more) entries (mix of extent and bitmap entries), we end up not
satisfying an allocation request with a size larger than the size of any of
the entries but no larger than the sum of their sizes. Making the caller assume
we're under a ENOSPC condition or force it to allocate multiple smaller space
regions (as we do for file data writes), which adds extra overhead and more
chances of causing fragmentation due to the smaller regions being all spread
apart from each other (more likely when under concurrency).
For example, if we have the following in the cache:
* extent entry representing free space range: [128Mb - 256Kb, 128Mb[
* bitmap entry covering the range [128Mb, 256Mb[, but only with the bits
representing the range [128Mb, 128Mb + 768Kb[ set - that is, only that
space in this 128Mb area is marked as free
An allocation request for 1Mb, starting at offset not greater than 128Mb - 256Kb,
would fail before, despite the existence of such contiguous free space area in the
cache. The caller could only allocate up to 768Kb of space at once and later another
256Kb (or vice-versa). In between each smaller allocation request, another task
working on a different file/inode might come in and take that space, preventing the
former task of getting a contiguous 1Mb region of free space.
Therefore this change implements the ability to move free space from bitmap
entries into existing and new free space regions represented with extent
entries. This is done when a space region is added to the cache.
A test was added to the sanity tests that explains in detail the issue too.
Some performance test results with compilebench on a 4 cores machine, with
32Gb of ram and using an HDD follow.
Test: compilebench -D /mnt -i 30 -r 1000 --makej
Before this change:
intial create total runs 30 avg 69.02 MB/s (user 0.28s sys 0.57s)
compile total runs 30 avg 314.96 MB/s (user 0.12s sys 0.25s)
read compiled tree total runs 3 avg 27.14 MB/s (user 1.52s sys 0.90s)
delete compiled tree total runs 30 avg 3.14 seconds (user 0.15s sys 0.66s)
After this change:
intial create total runs 30 avg 68.37 MB/s (user 0.29s sys 0.55s)
compile total runs 30 avg 382.83 MB/s (user 0.12s sys 0.24s)
read compiled tree total runs 3 avg 27.82 MB/s (user 1.45s sys 0.97s)
delete compiled tree total runs 30 avg 3.18 seconds (user 0.17s sys 0.65s)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-08-29 20:35:13 +08:00
|
|
|
RB_CLEAR_NODE(&info->offset_index);
|
2010-11-09 14:56:50 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_lock(&ctl->tree_lock);
|
2010-11-09 14:56:50 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
if (try_merge_free_space(ctl, info, true))
|
2010-11-09 14:56:50 +08:00
|
|
|
goto link;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* There was no extent directly to the left or right of this new
|
|
|
|
* extent then we know we're going to have to allocate a new extent, so
|
|
|
|
* before we do that see if we need to drop this into a bitmap
|
|
|
|
*/
|
2011-03-29 13:46:06 +08:00
|
|
|
ret = insert_into_bitmap(ctl, info);
|
2010-11-09 14:56:50 +08:00
|
|
|
if (ret < 0) {
|
|
|
|
goto out;
|
|
|
|
} else if (ret) {
|
|
|
|
ret = 0;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
link:
|
Btrfs: improve free space cache management and space allocation
While under random IO, a block group's free space cache eventually reaches
a state where it has a mix of extent entries and bitmap entries representing
free space regions.
As later free space regions are returned to the cache, some of them are merged
with existing extent entries if they are contiguous with them. But others are
not merged, because despite the existence of adjacent free space regions in
the cache, the merging doesn't happen because the existing free space regions
are represented in bitmap extents. Even when new free space regions are merged
with existing extent entries (enlarging the free space range they represent),
we create chances of having after an enlarged region that is contiguous with
some other region represented in a bitmap entry.
Both clustered and non-clustered space allocation work by iterating over our
extent and bitmap entries and skipping any that represents a region smaller
then the allocation request (and giving preference to extent entries before
bitmap entries). By having a contiguous free space region that is represented
by 2 (or more) entries (mix of extent and bitmap entries), we end up not
satisfying an allocation request with a size larger than the size of any of
the entries but no larger than the sum of their sizes. Making the caller assume
we're under a ENOSPC condition or force it to allocate multiple smaller space
regions (as we do for file data writes), which adds extra overhead and more
chances of causing fragmentation due to the smaller regions being all spread
apart from each other (more likely when under concurrency).
For example, if we have the following in the cache:
* extent entry representing free space range: [128Mb - 256Kb, 128Mb[
* bitmap entry covering the range [128Mb, 256Mb[, but only with the bits
representing the range [128Mb, 128Mb + 768Kb[ set - that is, only that
space in this 128Mb area is marked as free
An allocation request for 1Mb, starting at offset not greater than 128Mb - 256Kb,
would fail before, despite the existence of such contiguous free space area in the
cache. The caller could only allocate up to 768Kb of space at once and later another
256Kb (or vice-versa). In between each smaller allocation request, another task
working on a different file/inode might come in and take that space, preventing the
former task of getting a contiguous 1Mb region of free space.
Therefore this change implements the ability to move free space from bitmap
entries into existing and new free space regions represented with extent
entries. This is done when a space region is added to the cache.
A test was added to the sanity tests that explains in detail the issue too.
Some performance test results with compilebench on a 4 cores machine, with
32Gb of ram and using an HDD follow.
Test: compilebench -D /mnt -i 30 -r 1000 --makej
Before this change:
intial create total runs 30 avg 69.02 MB/s (user 0.28s sys 0.57s)
compile total runs 30 avg 314.96 MB/s (user 0.12s sys 0.25s)
read compiled tree total runs 3 avg 27.14 MB/s (user 1.52s sys 0.90s)
delete compiled tree total runs 30 avg 3.14 seconds (user 0.15s sys 0.66s)
After this change:
intial create total runs 30 avg 68.37 MB/s (user 0.29s sys 0.55s)
compile total runs 30 avg 382.83 MB/s (user 0.12s sys 0.24s)
read compiled tree total runs 3 avg 27.82 MB/s (user 1.45s sys 0.97s)
delete compiled tree total runs 30 avg 3.18 seconds (user 0.17s sys 0.65s)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-08-29 20:35:13 +08:00
|
|
|
/*
|
|
|
|
* Only steal free space from adjacent bitmaps if we're sure we're not
|
|
|
|
* going to add the new free space to existing bitmap entries - because
|
|
|
|
* that would mean unnecessary work that would be reverted. Therefore
|
|
|
|
* attempt to steal space from bitmaps if we're adding an extent entry.
|
|
|
|
*/
|
|
|
|
steal_from_bitmap(ctl, info, true);
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
ret = link_free_space(ctl, info);
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
if (ret)
|
2011-01-29 06:05:48 +08:00
|
|
|
kmem_cache_free(btrfs_free_space_cachep, info);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
out:
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_unlock(&ctl->tree_lock);
|
2009-04-03 22:14:18 +08:00
|
|
|
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
if (ret) {
|
2013-12-21 00:37:06 +08:00
|
|
|
printk(KERN_CRIT "BTRFS: unable to add free space :%d\n", ret);
|
2013-08-27 05:14:08 +08:00
|
|
|
ASSERT(ret != -EEXIST);
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2009-04-03 22:14:18 +08:00
|
|
|
int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
|
|
|
|
u64 offset, u64 bytes)
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
struct btrfs_free_space *info;
|
2012-12-19 00:39:19 +08:00
|
|
|
int ret;
|
|
|
|
bool re_search = false;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_lock(&ctl->tree_lock);
|
2009-04-03 22:14:18 +08:00
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
again:
|
2012-12-19 00:39:19 +08:00
|
|
|
ret = 0;
|
2012-06-28 03:10:56 +08:00
|
|
|
if (!bytes)
|
|
|
|
goto out_lock;
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
info = tree_search_offset(ctl, offset, 0, 0);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
if (!info) {
|
2009-07-31 23:03:58 +08:00
|
|
|
/*
|
|
|
|
* oops didn't find an extent that matched the space we wanted
|
|
|
|
* to remove, look for a bitmap instead
|
|
|
|
*/
|
2011-03-29 13:46:06 +08:00
|
|
|
info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
|
2009-07-31 23:03:58 +08:00
|
|
|
1, 0);
|
|
|
|
if (!info) {
|
2012-12-19 00:39:19 +08:00
|
|
|
/*
|
|
|
|
* If we found a partial bit of our free space in a
|
|
|
|
* bitmap but then couldn't find the other part this may
|
|
|
|
* be a problem, so WARN about it.
|
2011-11-21 22:39:11 +08:00
|
|
|
*/
|
2012-12-19 00:39:19 +08:00
|
|
|
WARN_ON(re_search);
|
2009-07-31 23:03:58 +08:00
|
|
|
goto out_lock;
|
|
|
|
}
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
}
|
|
|
|
|
2012-12-19 00:39:19 +08:00
|
|
|
re_search = false;
|
2012-06-28 03:10:56 +08:00
|
|
|
if (!info->bitmap) {
|
2011-03-29 13:46:06 +08:00
|
|
|
unlink_free_space(ctl, info);
|
2012-06-28 03:10:56 +08:00
|
|
|
if (offset == info->offset) {
|
|
|
|
u64 to_free = min(bytes, info->bytes);
|
|
|
|
|
|
|
|
info->bytes -= to_free;
|
|
|
|
info->offset += to_free;
|
|
|
|
if (info->bytes) {
|
|
|
|
ret = link_free_space(ctl, info);
|
|
|
|
WARN_ON(ret);
|
|
|
|
} else {
|
|
|
|
kmem_cache_free(btrfs_free_space_cachep, info);
|
|
|
|
}
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
2012-06-28 03:10:56 +08:00
|
|
|
offset += to_free;
|
|
|
|
bytes -= to_free;
|
|
|
|
goto again;
|
|
|
|
} else {
|
|
|
|
u64 old_end = info->bytes + info->offset;
|
2008-09-24 23:23:25 +08:00
|
|
|
|
2012-06-28 03:10:56 +08:00
|
|
|
info->bytes = offset - info->offset;
|
2011-03-29 13:46:06 +08:00
|
|
|
ret = link_free_space(ctl, info);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
WARN_ON(ret);
|
|
|
|
if (ret)
|
|
|
|
goto out_lock;
|
|
|
|
|
2012-06-28 03:10:56 +08:00
|
|
|
/* Not enough bytes in this entry to satisfy us */
|
|
|
|
if (old_end < offset + bytes) {
|
|
|
|
bytes -= old_end - offset;
|
|
|
|
offset = old_end;
|
|
|
|
goto again;
|
|
|
|
} else if (old_end == offset + bytes) {
|
|
|
|
/* all done */
|
|
|
|
goto out_lock;
|
|
|
|
}
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
|
|
|
|
|
|
|
ret = btrfs_add_free_space(block_group, offset + bytes,
|
|
|
|
old_end - (offset + bytes));
|
|
|
|
WARN_ON(ret);
|
|
|
|
goto out;
|
|
|
|
}
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
}
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
ret = remove_from_bitmap(ctl, info, &offset, &bytes);
|
2012-12-19 00:39:19 +08:00
|
|
|
if (ret == -EAGAIN) {
|
|
|
|
re_search = true;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
goto again;
|
2012-12-19 00:39:19 +08:00
|
|
|
}
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
out_lock:
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
out:
|
Btrfs: nuke fs wide allocation mutex V2
This patch removes the giant fs_info->alloc_mutex and replaces it with a bunch
of little locks.
There is now a pinned_mutex, which is used when messing with the pinned_extents
extent io tree, and the extent_ins_mutex which is used with the pending_del and
extent_ins extent io trees.
The locking for the extent tree stuff was inspired by a patch that Yan Zheng
wrote to fix a race condition, I cleaned it up some and changed the locking
around a little bit, but the idea remains the same. Basically instead of
holding the extent_ins_mutex throughout the processing of an extent on the
extent_ins or pending_del trees, we just hold it while we're searching and when
we clear the bits on those trees, and lock the extent for the duration of the
operations on the extent.
Also to keep from getting hung up waiting to lock an extent, I've added a
try_lock_extent so if we cannot lock the extent, move on to the next one in the
tree and we'll come back to that one. I have tested this heavily and it does
not appear to break anything. This has to be applied on top of my
find_free_extent redo patch.
I tested this patch on top of Yan's space reblancing code and it worked fine.
The only thing that has changed since the last version is I pulled out all my
debugging stuff, apparently I forgot to run guilt refresh before I sent the
last patch out. Thank you,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
2008-10-30 02:49:05 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
|
|
|
|
u64 bytes)
|
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
struct btrfs_free_space *info;
|
|
|
|
struct rb_node *n;
|
|
|
|
int count = 0;
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
for (n = rb_first(&ctl->free_space_offset); n; n = rb_next(n)) {
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
info = rb_entry(n, struct btrfs_free_space, offset_index);
|
2012-07-06 17:31:36 +08:00
|
|
|
if (info->bytes >= bytes && !block_group->ro)
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
count++;
|
2013-12-21 00:37:06 +08:00
|
|
|
btrfs_crit(block_group->fs_info,
|
|
|
|
"entry offset %llu, bytes %llu, bitmap %s",
|
|
|
|
info->offset, info->bytes,
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
(info->bitmap) ? "yes" : "no");
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
}
|
2013-12-21 00:37:06 +08:00
|
|
|
btrfs_info(block_group->fs_info, "block group has cluster?: %s",
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
list_empty(&block_group->cluster_list) ? "no" : "yes");
|
2013-12-21 00:37:06 +08:00
|
|
|
btrfs_info(block_group->fs_info,
|
|
|
|
"%d blocks of free space at or bigger than bytes is", count);
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
}
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
void btrfs_init_free_space_ctl(struct btrfs_block_group_cache *block_group)
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_lock_init(&ctl->tree_lock);
|
|
|
|
ctl->unit = block_group->sectorsize;
|
|
|
|
ctl->start = block_group->key.objectid;
|
|
|
|
ctl->private = block_group;
|
|
|
|
ctl->op = &free_space_op;
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
INIT_LIST_HEAD(&ctl->trimming_ranges);
|
|
|
|
mutex_init(&ctl->cache_writeout_mutex);
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
/*
|
|
|
|
* we only want to have 32k of ram per block group for keeping
|
|
|
|
* track of free space, and if we pass 1/2 of that we want to
|
|
|
|
* start converting things over to using bitmaps
|
|
|
|
*/
|
2015-12-15 00:42:10 +08:00
|
|
|
ctl->extents_thresh = (SZ_32K / 2) / sizeof(struct btrfs_free_space);
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
}
|
|
|
|
|
2009-04-03 21:47:43 +08:00
|
|
|
/*
|
|
|
|
* for a given cluster, put all of its extents back into the free
|
|
|
|
* space cache. If the block group passed doesn't match the block group
|
|
|
|
* pointed to by the cluster, someone else raced in and freed the
|
|
|
|
* cluster already. In that case, we just return without changing anything
|
|
|
|
*/
|
|
|
|
static int
|
|
|
|
__btrfs_return_cluster_to_free_space(
|
|
|
|
struct btrfs_block_group_cache *block_group,
|
|
|
|
struct btrfs_free_cluster *cluster)
|
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
2009-04-03 21:47:43 +08:00
|
|
|
struct btrfs_free_space *entry;
|
|
|
|
struct rb_node *node;
|
|
|
|
|
|
|
|
spin_lock(&cluster->lock);
|
|
|
|
if (cluster->block_group != block_group)
|
|
|
|
goto out;
|
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
cluster->block_group = NULL;
|
2009-04-03 21:47:43 +08:00
|
|
|
cluster->window_start = 0;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
list_del_init(&cluster->block_group_list);
|
|
|
|
|
2009-04-03 21:47:43 +08:00
|
|
|
node = rb_first(&cluster->root);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
while (node) {
|
2011-03-21 22:11:24 +08:00
|
|
|
bool bitmap;
|
|
|
|
|
2009-04-03 21:47:43 +08:00
|
|
|
entry = rb_entry(node, struct btrfs_free_space, offset_index);
|
|
|
|
node = rb_next(&entry->offset_index);
|
|
|
|
rb_erase(&entry->offset_index, &cluster->root);
|
Btrfs: improve free space cache management and space allocation
While under random IO, a block group's free space cache eventually reaches
a state where it has a mix of extent entries and bitmap entries representing
free space regions.
As later free space regions are returned to the cache, some of them are merged
with existing extent entries if they are contiguous with them. But others are
not merged, because despite the existence of adjacent free space regions in
the cache, the merging doesn't happen because the existing free space regions
are represented in bitmap extents. Even when new free space regions are merged
with existing extent entries (enlarging the free space range they represent),
we create chances of having after an enlarged region that is contiguous with
some other region represented in a bitmap entry.
Both clustered and non-clustered space allocation work by iterating over our
extent and bitmap entries and skipping any that represents a region smaller
then the allocation request (and giving preference to extent entries before
bitmap entries). By having a contiguous free space region that is represented
by 2 (or more) entries (mix of extent and bitmap entries), we end up not
satisfying an allocation request with a size larger than the size of any of
the entries but no larger than the sum of their sizes. Making the caller assume
we're under a ENOSPC condition or force it to allocate multiple smaller space
regions (as we do for file data writes), which adds extra overhead and more
chances of causing fragmentation due to the smaller regions being all spread
apart from each other (more likely when under concurrency).
For example, if we have the following in the cache:
* extent entry representing free space range: [128Mb - 256Kb, 128Mb[
* bitmap entry covering the range [128Mb, 256Mb[, but only with the bits
representing the range [128Mb, 128Mb + 768Kb[ set - that is, only that
space in this 128Mb area is marked as free
An allocation request for 1Mb, starting at offset not greater than 128Mb - 256Kb,
would fail before, despite the existence of such contiguous free space area in the
cache. The caller could only allocate up to 768Kb of space at once and later another
256Kb (or vice-versa). In between each smaller allocation request, another task
working on a different file/inode might come in and take that space, preventing the
former task of getting a contiguous 1Mb region of free space.
Therefore this change implements the ability to move free space from bitmap
entries into existing and new free space regions represented with extent
entries. This is done when a space region is added to the cache.
A test was added to the sanity tests that explains in detail the issue too.
Some performance test results with compilebench on a 4 cores machine, with
32Gb of ram and using an HDD follow.
Test: compilebench -D /mnt -i 30 -r 1000 --makej
Before this change:
intial create total runs 30 avg 69.02 MB/s (user 0.28s sys 0.57s)
compile total runs 30 avg 314.96 MB/s (user 0.12s sys 0.25s)
read compiled tree total runs 3 avg 27.14 MB/s (user 1.52s sys 0.90s)
delete compiled tree total runs 30 avg 3.14 seconds (user 0.15s sys 0.66s)
After this change:
intial create total runs 30 avg 68.37 MB/s (user 0.29s sys 0.55s)
compile total runs 30 avg 382.83 MB/s (user 0.12s sys 0.24s)
read compiled tree total runs 3 avg 27.82 MB/s (user 1.45s sys 0.97s)
delete compiled tree total runs 30 avg 3.18 seconds (user 0.17s sys 0.65s)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-08-29 20:35:13 +08:00
|
|
|
RB_CLEAR_NODE(&entry->offset_index);
|
2011-03-21 22:11:24 +08:00
|
|
|
|
|
|
|
bitmap = (entry->bitmap != NULL);
|
Btrfs: improve free space cache management and space allocation
While under random IO, a block group's free space cache eventually reaches
a state where it has a mix of extent entries and bitmap entries representing
free space regions.
As later free space regions are returned to the cache, some of them are merged
with existing extent entries if they are contiguous with them. But others are
not merged, because despite the existence of adjacent free space regions in
the cache, the merging doesn't happen because the existing free space regions
are represented in bitmap extents. Even when new free space regions are merged
with existing extent entries (enlarging the free space range they represent),
we create chances of having after an enlarged region that is contiguous with
some other region represented in a bitmap entry.
Both clustered and non-clustered space allocation work by iterating over our
extent and bitmap entries and skipping any that represents a region smaller
then the allocation request (and giving preference to extent entries before
bitmap entries). By having a contiguous free space region that is represented
by 2 (or more) entries (mix of extent and bitmap entries), we end up not
satisfying an allocation request with a size larger than the size of any of
the entries but no larger than the sum of their sizes. Making the caller assume
we're under a ENOSPC condition or force it to allocate multiple smaller space
regions (as we do for file data writes), which adds extra overhead and more
chances of causing fragmentation due to the smaller regions being all spread
apart from each other (more likely when under concurrency).
For example, if we have the following in the cache:
* extent entry representing free space range: [128Mb - 256Kb, 128Mb[
* bitmap entry covering the range [128Mb, 256Mb[, but only with the bits
representing the range [128Mb, 128Mb + 768Kb[ set - that is, only that
space in this 128Mb area is marked as free
An allocation request for 1Mb, starting at offset not greater than 128Mb - 256Kb,
would fail before, despite the existence of such contiguous free space area in the
cache. The caller could only allocate up to 768Kb of space at once and later another
256Kb (or vice-versa). In between each smaller allocation request, another task
working on a different file/inode might come in and take that space, preventing the
former task of getting a contiguous 1Mb region of free space.
Therefore this change implements the ability to move free space from bitmap
entries into existing and new free space regions represented with extent
entries. This is done when a space region is added to the cache.
A test was added to the sanity tests that explains in detail the issue too.
Some performance test results with compilebench on a 4 cores machine, with
32Gb of ram and using an HDD follow.
Test: compilebench -D /mnt -i 30 -r 1000 --makej
Before this change:
intial create total runs 30 avg 69.02 MB/s (user 0.28s sys 0.57s)
compile total runs 30 avg 314.96 MB/s (user 0.12s sys 0.25s)
read compiled tree total runs 3 avg 27.14 MB/s (user 1.52s sys 0.90s)
delete compiled tree total runs 30 avg 3.14 seconds (user 0.15s sys 0.66s)
After this change:
intial create total runs 30 avg 68.37 MB/s (user 0.29s sys 0.55s)
compile total runs 30 avg 382.83 MB/s (user 0.12s sys 0.24s)
read compiled tree total runs 3 avg 27.82 MB/s (user 1.45s sys 0.97s)
delete compiled tree total runs 30 avg 3.18 seconds (user 0.17s sys 0.65s)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-08-29 20:35:13 +08:00
|
|
|
if (!bitmap) {
|
2011-03-29 13:46:06 +08:00
|
|
|
try_merge_free_space(ctl, entry, false);
|
Btrfs: improve free space cache management and space allocation
While under random IO, a block group's free space cache eventually reaches
a state where it has a mix of extent entries and bitmap entries representing
free space regions.
As later free space regions are returned to the cache, some of them are merged
with existing extent entries if they are contiguous with them. But others are
not merged, because despite the existence of adjacent free space regions in
the cache, the merging doesn't happen because the existing free space regions
are represented in bitmap extents. Even when new free space regions are merged
with existing extent entries (enlarging the free space range they represent),
we create chances of having after an enlarged region that is contiguous with
some other region represented in a bitmap entry.
Both clustered and non-clustered space allocation work by iterating over our
extent and bitmap entries and skipping any that represents a region smaller
then the allocation request (and giving preference to extent entries before
bitmap entries). By having a contiguous free space region that is represented
by 2 (or more) entries (mix of extent and bitmap entries), we end up not
satisfying an allocation request with a size larger than the size of any of
the entries but no larger than the sum of their sizes. Making the caller assume
we're under a ENOSPC condition or force it to allocate multiple smaller space
regions (as we do for file data writes), which adds extra overhead and more
chances of causing fragmentation due to the smaller regions being all spread
apart from each other (more likely when under concurrency).
For example, if we have the following in the cache:
* extent entry representing free space range: [128Mb - 256Kb, 128Mb[
* bitmap entry covering the range [128Mb, 256Mb[, but only with the bits
representing the range [128Mb, 128Mb + 768Kb[ set - that is, only that
space in this 128Mb area is marked as free
An allocation request for 1Mb, starting at offset not greater than 128Mb - 256Kb,
would fail before, despite the existence of such contiguous free space area in the
cache. The caller could only allocate up to 768Kb of space at once and later another
256Kb (or vice-versa). In between each smaller allocation request, another task
working on a different file/inode might come in and take that space, preventing the
former task of getting a contiguous 1Mb region of free space.
Therefore this change implements the ability to move free space from bitmap
entries into existing and new free space regions represented with extent
entries. This is done when a space region is added to the cache.
A test was added to the sanity tests that explains in detail the issue too.
Some performance test results with compilebench on a 4 cores machine, with
32Gb of ram and using an HDD follow.
Test: compilebench -D /mnt -i 30 -r 1000 --makej
Before this change:
intial create total runs 30 avg 69.02 MB/s (user 0.28s sys 0.57s)
compile total runs 30 avg 314.96 MB/s (user 0.12s sys 0.25s)
read compiled tree total runs 3 avg 27.14 MB/s (user 1.52s sys 0.90s)
delete compiled tree total runs 30 avg 3.14 seconds (user 0.15s sys 0.66s)
After this change:
intial create total runs 30 avg 68.37 MB/s (user 0.29s sys 0.55s)
compile total runs 30 avg 382.83 MB/s (user 0.12s sys 0.24s)
read compiled tree total runs 3 avg 27.82 MB/s (user 1.45s sys 0.97s)
delete compiled tree total runs 30 avg 3.18 seconds (user 0.17s sys 0.65s)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-08-29 20:35:13 +08:00
|
|
|
steal_from_bitmap(ctl, entry, false);
|
|
|
|
}
|
2011-03-29 13:46:06 +08:00
|
|
|
tree_insert_offset(&ctl->free_space_offset,
|
2011-03-21 22:11:24 +08:00
|
|
|
entry->offset, &entry->offset_index, bitmap);
|
2009-04-03 21:47:43 +08:00
|
|
|
}
|
2010-02-24 03:43:04 +08:00
|
|
|
cluster->root = RB_ROOT;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2009-04-03 21:47:43 +08:00
|
|
|
out:
|
|
|
|
spin_unlock(&cluster->lock);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
btrfs_put_block_group(block_group);
|
2009-04-03 21:47:43 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2013-04-26 04:41:01 +08:00
|
|
|
static void __btrfs_remove_free_space_cache_locked(
|
|
|
|
struct btrfs_free_space_ctl *ctl)
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
{
|
|
|
|
struct btrfs_free_space *info;
|
|
|
|
struct rb_node *node;
|
Btrfs: Cache free inode numbers in memory
Currently btrfs stores the highest objectid of the fs tree, and it always
returns (highest+1) inode number when we create a file, so inode numbers
won't be reclaimed when we delete files, so we'll run out of inode numbers
as we keep create/delete files in 32bits machines.
This fixes it, and it works similarly to how we cache free space in block
cgroups.
We start a kernel thread to read the file tree. By scanning inode items,
we know which chunks of inode numbers are free, and we cache them in
an rb-tree.
Because we are searching the commit root, we have to carefully handle the
cross-transaction case.
The rb-tree is a hybrid extent+bitmap tree, so if we have too many small
chunks of inode numbers, we'll use bitmaps. Initially we allow 16K ram
of extents, and a bitmap will be used if we exceed this threshold. The
extents threshold is adjusted in runtime.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-04-20 10:06:11 +08:00
|
|
|
|
|
|
|
while ((node = rb_last(&ctl->free_space_offset)) != NULL) {
|
|
|
|
info = rb_entry(node, struct btrfs_free_space, offset_index);
|
2011-06-25 00:02:51 +08:00
|
|
|
if (!info->bitmap) {
|
|
|
|
unlink_free_space(ctl, info);
|
|
|
|
kmem_cache_free(btrfs_free_space_cachep, info);
|
|
|
|
} else {
|
|
|
|
free_bitmap(ctl, info);
|
|
|
|
}
|
2015-01-08 22:20:54 +08:00
|
|
|
|
|
|
|
cond_resched_lock(&ctl->tree_lock);
|
Btrfs: Cache free inode numbers in memory
Currently btrfs stores the highest objectid of the fs tree, and it always
returns (highest+1) inode number when we create a file, so inode numbers
won't be reclaimed when we delete files, so we'll run out of inode numbers
as we keep create/delete files in 32bits machines.
This fixes it, and it works similarly to how we cache free space in block
cgroups.
We start a kernel thread to read the file tree. By scanning inode items,
we know which chunks of inode numbers are free, and we cache them in
an rb-tree.
Because we are searching the commit root, we have to carefully handle the
cross-transaction case.
The rb-tree is a hybrid extent+bitmap tree, so if we have too many small
chunks of inode numbers, we'll use bitmaps. Initially we allow 16K ram
of extents, and a bitmap will be used if we exceed this threshold. The
extents threshold is adjusted in runtime.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-04-20 10:06:11 +08:00
|
|
|
}
|
2011-05-21 21:27:38 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
void __btrfs_remove_free_space_cache(struct btrfs_free_space_ctl *ctl)
|
|
|
|
{
|
|
|
|
spin_lock(&ctl->tree_lock);
|
|
|
|
__btrfs_remove_free_space_cache_locked(ctl);
|
Btrfs: Cache free inode numbers in memory
Currently btrfs stores the highest objectid of the fs tree, and it always
returns (highest+1) inode number when we create a file, so inode numbers
won't be reclaimed when we delete files, so we'll run out of inode numbers
as we keep create/delete files in 32bits machines.
This fixes it, and it works similarly to how we cache free space in block
cgroups.
We start a kernel thread to read the file tree. By scanning inode items,
we know which chunks of inode numbers are free, and we cache them in
an rb-tree.
Because we are searching the commit root, we have to carefully handle the
cross-transaction case.
The rb-tree is a hybrid extent+bitmap tree, so if we have too many small
chunks of inode numbers, we'll use bitmaps. Initially we allow 16K ram
of extents, and a bitmap will be used if we exceed this threshold. The
extents threshold is adjusted in runtime.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-04-20 10:06:11 +08:00
|
|
|
spin_unlock(&ctl->tree_lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group)
|
|
|
|
{
|
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
2009-04-03 21:47:43 +08:00
|
|
|
struct btrfs_free_cluster *cluster;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
struct list_head *head;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_lock(&ctl->tree_lock);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
while ((head = block_group->cluster_list.next) !=
|
|
|
|
&block_group->cluster_list) {
|
|
|
|
cluster = list_entry(head, struct btrfs_free_cluster,
|
|
|
|
block_group_list);
|
2009-04-03 21:47:43 +08:00
|
|
|
|
|
|
|
WARN_ON(cluster->block_group != block_group);
|
|
|
|
__btrfs_return_cluster_to_free_space(block_group, cluster);
|
2015-01-08 22:20:54 +08:00
|
|
|
|
|
|
|
cond_resched_lock(&ctl->tree_lock);
|
2009-04-03 21:47:43 +08:00
|
|
|
}
|
2011-05-21 21:27:38 +08:00
|
|
|
__btrfs_remove_free_space_cache_locked(ctl);
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_unlock(&ctl->tree_lock);
|
2009-04-03 21:47:43 +08:00
|
|
|
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
}
|
|
|
|
|
2009-04-03 22:14:18 +08:00
|
|
|
u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
|
2013-09-09 13:19:42 +08:00
|
|
|
u64 offset, u64 bytes, u64 empty_size,
|
|
|
|
u64 *max_extent_size)
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
2009-04-03 22:14:18 +08:00
|
|
|
struct btrfs_free_space *entry = NULL;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
u64 bytes_search = bytes + empty_size;
|
2009-04-03 22:14:18 +08:00
|
|
|
u64 ret = 0;
|
2013-01-30 07:40:14 +08:00
|
|
|
u64 align_gap = 0;
|
|
|
|
u64 align_gap_len = 0;
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_lock(&ctl->tree_lock);
|
2013-01-30 07:40:14 +08:00
|
|
|
entry = find_free_space(ctl, &offset, &bytes_search,
|
2013-09-09 13:19:42 +08:00
|
|
|
block_group->full_stripe_len, max_extent_size);
|
2009-04-03 22:14:18 +08:00
|
|
|
if (!entry)
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
goto out;
|
|
|
|
|
|
|
|
ret = offset;
|
|
|
|
if (entry->bitmap) {
|
2011-03-29 13:46:06 +08:00
|
|
|
bitmap_clear_bits(ctl, entry, offset, bytes);
|
2010-11-09 14:50:07 +08:00
|
|
|
if (!entry->bytes)
|
2011-03-29 13:46:06 +08:00
|
|
|
free_bitmap(ctl, entry);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
} else {
|
2011-03-29 13:46:06 +08:00
|
|
|
unlink_free_space(ctl, entry);
|
2013-01-30 07:40:14 +08:00
|
|
|
align_gap_len = offset - entry->offset;
|
|
|
|
align_gap = entry->offset;
|
|
|
|
|
|
|
|
entry->offset = offset + bytes;
|
|
|
|
WARN_ON(entry->bytes < bytes + align_gap_len);
|
|
|
|
|
|
|
|
entry->bytes -= bytes + align_gap_len;
|
2009-04-03 22:14:18 +08:00
|
|
|
if (!entry->bytes)
|
2011-01-29 06:05:48 +08:00
|
|
|
kmem_cache_free(btrfs_free_space_cachep, entry);
|
2009-04-03 22:14:18 +08:00
|
|
|
else
|
2011-03-29 13:46:06 +08:00
|
|
|
link_free_space(ctl, entry);
|
2009-04-03 22:14:18 +08:00
|
|
|
}
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
out:
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: async block group caching
This patch moves the caching of the block group off to a kthread in order to
allow people to allocate sooner. Instead of blocking up behind the caching
mutex, we instead kick of the caching kthread, and then attempt to make an
allocation. If we cannot, we wait on the block groups caching waitqueue, which
the caching kthread will wake the waiting threads up everytime it finds 2 meg
worth of space, and then again when its finished caching. This is how I tested
the speedup from this
mkfs the disk
mount the disk
fill the disk up with fs_mark
unmount the disk
mount the disk
time touch /mnt/foo
Without my changes this took 11 seconds on my box, with these changes it now
takes 1 second.
Another change thats been put in place is we lock the super mirror's in the
pinned extent map in order to keep us from adding that stuff as free space when
caching the block group. This doesn't really change anything else as far as the
pinned extent map is concerned, since for actual pinned extents we use
EXTENT_DIRTY, but it does mean that when we unmount we have to go in and unlock
those extents to keep from leaking memory.
I've also added a check where when we are reading block groups from disk, if the
amount of space used == the size of the block group, we go ahead and mark the
block group as cached. This drastically reduces the amount of time it takes to
cache the block groups. Using the same test as above, except doing a dd to a
file and then unmounting, it used to take 33 seconds to umount, now it takes 3
seconds.
This version uses the commit_root in the caching kthread, and then keeps track
of how many async caching threads are running at any given time so if one of the
async threads is still running as we cross transactions we can wait until its
finished before handling the pinned extents. Thank you,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2013-01-30 07:40:14 +08:00
|
|
|
if (align_gap_len)
|
|
|
|
__btrfs_add_free_space(ctl, align_gap, align_gap_len);
|
Btrfs: free space accounting redo
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>
2008-09-24 01:14:11 +08:00
|
|
|
return ret;
|
|
|
|
}
|
2009-04-03 21:47:43 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* given a cluster, put all of its extents back into the free space
|
|
|
|
* cache. If a block group is passed, this function will only free
|
|
|
|
* a cluster that belongs to the passed block group.
|
|
|
|
*
|
|
|
|
* Otherwise, it'll get a reference on the block group pointed to by the
|
|
|
|
* cluster and remove the cluster from it.
|
|
|
|
*/
|
|
|
|
int btrfs_return_cluster_to_free_space(
|
|
|
|
struct btrfs_block_group_cache *block_group,
|
|
|
|
struct btrfs_free_cluster *cluster)
|
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl;
|
2009-04-03 21:47:43 +08:00
|
|
|
int ret;
|
|
|
|
|
|
|
|
/* first, get a safe pointer to the block group */
|
|
|
|
spin_lock(&cluster->lock);
|
|
|
|
if (!block_group) {
|
|
|
|
block_group = cluster->block_group;
|
|
|
|
if (!block_group) {
|
|
|
|
spin_unlock(&cluster->lock);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
} else if (cluster->block_group != block_group) {
|
|
|
|
/* someone else has already freed it don't redo their work */
|
|
|
|
spin_unlock(&cluster->lock);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
atomic_inc(&block_group->count);
|
|
|
|
spin_unlock(&cluster->lock);
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
ctl = block_group->free_space_ctl;
|
|
|
|
|
2009-04-03 21:47:43 +08:00
|
|
|
/* now return any extents the cluster had on it */
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_lock(&ctl->tree_lock);
|
2009-04-03 21:47:43 +08:00
|
|
|
ret = __btrfs_return_cluster_to_free_space(block_group, cluster);
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_unlock(&ctl->tree_lock);
|
2009-04-03 21:47:43 +08:00
|
|
|
|
|
|
|
/* finally drop our ref */
|
|
|
|
btrfs_put_block_group(block_group);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group,
|
|
|
|
struct btrfs_free_cluster *cluster,
|
2011-03-21 22:11:24 +08:00
|
|
|
struct btrfs_free_space *entry,
|
2013-09-09 13:19:42 +08:00
|
|
|
u64 bytes, u64 min_start,
|
|
|
|
u64 *max_extent_size)
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
int err;
|
|
|
|
u64 search_start = cluster->window_start;
|
|
|
|
u64 search_bytes = bytes;
|
|
|
|
u64 ret = 0;
|
|
|
|
|
|
|
|
search_start = min_start;
|
|
|
|
search_bytes = bytes;
|
|
|
|
|
2015-10-03 04:12:23 +08:00
|
|
|
err = search_bitmap(ctl, entry, &search_start, &search_bytes, true);
|
2013-09-09 13:19:42 +08:00
|
|
|
if (err) {
|
|
|
|
if (search_bytes > *max_extent_size)
|
|
|
|
*max_extent_size = search_bytes;
|
2011-03-21 22:11:24 +08:00
|
|
|
return 0;
|
2013-09-09 13:19:42 +08:00
|
|
|
}
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
|
|
|
ret = search_start;
|
2011-08-05 17:32:35 +08:00
|
|
|
__bitmap_clear_bits(ctl, entry, ret, bytes);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2009-04-03 21:47:43 +08:00
|
|
|
/*
|
|
|
|
* given a cluster, try to allocate 'bytes' from it, returns 0
|
|
|
|
* if it couldn't find anything suitably large, or a logical disk offset
|
|
|
|
* if things worked out
|
|
|
|
*/
|
|
|
|
u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group,
|
|
|
|
struct btrfs_free_cluster *cluster, u64 bytes,
|
2013-09-09 13:19:42 +08:00
|
|
|
u64 min_start, u64 *max_extent_size)
|
2009-04-03 21:47:43 +08:00
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
2009-04-03 21:47:43 +08:00
|
|
|
struct btrfs_free_space *entry = NULL;
|
|
|
|
struct rb_node *node;
|
|
|
|
u64 ret = 0;
|
|
|
|
|
|
|
|
spin_lock(&cluster->lock);
|
|
|
|
if (bytes > cluster->max_size)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
if (cluster->block_group != block_group)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
node = rb_first(&cluster->root);
|
|
|
|
if (!node)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
entry = rb_entry(node, struct btrfs_free_space, offset_index);
|
2013-10-31 13:03:04 +08:00
|
|
|
while (1) {
|
2013-09-09 13:19:42 +08:00
|
|
|
if (entry->bytes < bytes && entry->bytes > *max_extent_size)
|
|
|
|
*max_extent_size = entry->bytes;
|
|
|
|
|
2011-03-21 22:11:24 +08:00
|
|
|
if (entry->bytes < bytes ||
|
|
|
|
(!entry->bitmap && entry->offset < min_start)) {
|
2009-04-03 21:47:43 +08:00
|
|
|
node = rb_next(&entry->offset_index);
|
|
|
|
if (!node)
|
|
|
|
break;
|
|
|
|
entry = rb_entry(node, struct btrfs_free_space,
|
|
|
|
offset_index);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
2011-03-21 22:11:24 +08:00
|
|
|
if (entry->bitmap) {
|
|
|
|
ret = btrfs_alloc_from_bitmap(block_group,
|
|
|
|
cluster, entry, bytes,
|
2013-09-09 13:19:42 +08:00
|
|
|
cluster->window_start,
|
|
|
|
max_extent_size);
|
2011-03-21 22:11:24 +08:00
|
|
|
if (ret == 0) {
|
|
|
|
node = rb_next(&entry->offset_index);
|
|
|
|
if (!node)
|
|
|
|
break;
|
|
|
|
entry = rb_entry(node, struct btrfs_free_space,
|
|
|
|
offset_index);
|
|
|
|
continue;
|
|
|
|
}
|
2012-01-27 04:01:12 +08:00
|
|
|
cluster->window_start += bytes;
|
2011-03-21 22:11:24 +08:00
|
|
|
} else {
|
|
|
|
ret = entry->offset;
|
|
|
|
|
|
|
|
entry->offset += bytes;
|
|
|
|
entry->bytes -= bytes;
|
|
|
|
}
|
2009-04-03 21:47:43 +08:00
|
|
|
|
2010-11-09 14:55:34 +08:00
|
|
|
if (entry->bytes == 0)
|
2009-04-03 21:47:43 +08:00
|
|
|
rb_erase(&entry->offset_index, &cluster->root);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
out:
|
|
|
|
spin_unlock(&cluster->lock);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2010-11-09 14:55:34 +08:00
|
|
|
if (!ret)
|
|
|
|
return 0;
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_lock(&ctl->tree_lock);
|
2010-11-09 14:55:34 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
ctl->free_space -= bytes;
|
2010-11-09 14:55:34 +08:00
|
|
|
if (entry->bytes == 0) {
|
2011-03-29 13:46:06 +08:00
|
|
|
ctl->free_extents--;
|
2011-03-21 22:11:24 +08:00
|
|
|
if (entry->bitmap) {
|
|
|
|
kfree(entry->bitmap);
|
2011-03-29 13:46:06 +08:00
|
|
|
ctl->total_bitmaps--;
|
|
|
|
ctl->op->recalc_thresholds(ctl);
|
2011-03-21 22:11:24 +08:00
|
|
|
}
|
2011-01-29 06:05:48 +08:00
|
|
|
kmem_cache_free(btrfs_free_space_cachep, entry);
|
2010-11-09 14:55:34 +08:00
|
|
|
}
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_unlock(&ctl->tree_lock);
|
2010-11-09 14:55:34 +08:00
|
|
|
|
2009-04-03 21:47:43 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group,
|
|
|
|
struct btrfs_free_space *entry,
|
|
|
|
struct btrfs_free_cluster *cluster,
|
2011-10-14 23:10:36 +08:00
|
|
|
u64 offset, u64 bytes,
|
|
|
|
u64 cont1_bytes, u64 min_bytes)
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
unsigned long next_zero;
|
|
|
|
unsigned long i;
|
2011-10-14 23:10:36 +08:00
|
|
|
unsigned long want_bits;
|
|
|
|
unsigned long min_bits;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
unsigned long found_bits;
|
2015-10-03 04:09:42 +08:00
|
|
|
unsigned long max_bits = 0;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
unsigned long start = 0;
|
|
|
|
unsigned long total_found = 0;
|
2011-03-21 22:11:24 +08:00
|
|
|
int ret;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2012-11-30 14:30:14 +08:00
|
|
|
i = offset_to_bit(entry->offset, ctl->unit,
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
max_t(u64, offset, entry->offset));
|
2012-11-30 14:30:14 +08:00
|
|
|
want_bits = bytes_to_bits(bytes, ctl->unit);
|
|
|
|
min_bits = bytes_to_bits(min_bytes, ctl->unit);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2015-10-03 04:09:42 +08:00
|
|
|
/*
|
|
|
|
* Don't bother looking for a cluster in this bitmap if it's heavily
|
|
|
|
* fragmented.
|
|
|
|
*/
|
|
|
|
if (entry->max_extent_size &&
|
|
|
|
entry->max_extent_size < cont1_bytes)
|
|
|
|
return -ENOSPC;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
again:
|
|
|
|
found_bits = 0;
|
2012-09-14 10:29:02 +08:00
|
|
|
for_each_set_bit_from(i, entry->bitmap, BITS_PER_BITMAP) {
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
next_zero = find_next_zero_bit(entry->bitmap,
|
|
|
|
BITS_PER_BITMAP, i);
|
2011-10-14 23:10:36 +08:00
|
|
|
if (next_zero - i >= min_bits) {
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
found_bits = next_zero - i;
|
2015-10-03 04:09:42 +08:00
|
|
|
if (found_bits > max_bits)
|
|
|
|
max_bits = found_bits;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
break;
|
|
|
|
}
|
2015-10-03 04:09:42 +08:00
|
|
|
if (next_zero - i > max_bits)
|
|
|
|
max_bits = next_zero - i;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
i = next_zero;
|
|
|
|
}
|
|
|
|
|
2015-10-03 04:09:42 +08:00
|
|
|
if (!found_bits) {
|
|
|
|
entry->max_extent_size = (u64)max_bits * ctl->unit;
|
2011-03-21 22:11:24 +08:00
|
|
|
return -ENOSPC;
|
2015-10-03 04:09:42 +08:00
|
|
|
}
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2011-10-14 23:10:36 +08:00
|
|
|
if (!total_found) {
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
start = i;
|
2011-12-01 02:43:00 +08:00
|
|
|
cluster->max_size = 0;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
total_found += found_bits;
|
|
|
|
|
2012-11-30 14:30:14 +08:00
|
|
|
if (cluster->max_size < found_bits * ctl->unit)
|
|
|
|
cluster->max_size = found_bits * ctl->unit;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2011-10-14 23:10:36 +08:00
|
|
|
if (total_found < want_bits || cluster->max_size < cont1_bytes) {
|
|
|
|
i = next_zero + 1;
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
goto again;
|
|
|
|
}
|
|
|
|
|
2012-11-30 14:30:14 +08:00
|
|
|
cluster->window_start = start * ctl->unit + entry->offset;
|
2011-03-29 13:46:06 +08:00
|
|
|
rb_erase(&entry->offset_index, &ctl->free_space_offset);
|
2011-03-21 22:11:24 +08:00
|
|
|
ret = tree_insert_offset(&cluster->root, entry->offset,
|
|
|
|
&entry->offset_index, 1);
|
2013-08-27 05:14:08 +08:00
|
|
|
ASSERT(!ret); /* -EEXIST; Logic error */
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
|
2011-11-10 21:29:20 +08:00
|
|
|
trace_btrfs_setup_cluster(block_group, cluster,
|
2012-11-30 14:30:14 +08:00
|
|
|
total_found * ctl->unit, 1);
|
Btrfs: use hybrid extents+bitmap rb tree for free space
Currently btrfs has a problem where it can use a ridiculous amount of RAM simply
tracking free space. As free space gets fragmented, we end up with thousands of
entries on an rb-tree per block group, which usually spans 1 gig of area. Since
we currently don't ever flush free space cache back to disk this gets to be a
bit unweildly on large fs's with lots of fragmentation.
This patch solves this problem by using PAGE_SIZE bitmaps for parts of the free
space cache. Initially we calculate a threshold of extent entries we can
handle, which is however many extent entries we can cram into 16k of ram. The
maximum amount of RAM that should ever be used to track 1 gigabyte of diskspace
will be 32k of RAM, which scales much better than we did before.
Once we pass the extent threshold, we start adding bitmaps and using those
instead for tracking the free space. This patch also makes it so that any free
space thats less than 4 * sectorsize we go ahead and put into a bitmap. This is
nice since we try and allocate out of the front of a block group, so if the
front of a block group is heavily fragmented and then has a huge chunk of free
space at the end, we go ahead and add the fragmented areas to bitmaps and use a
normal extent entry to track the big chunk at the back of the block group.
I've also taken the opportunity to revamp how we search for free space.
Previously we indexed free space via an offset indexed rb tree and a bytes
indexed rb tree. I've dropped the bytes indexed rb tree and use only the offset
indexed rb tree. This cuts the number of tree operations we were doing
previously down by half, and gives us a little bit of a better allocation
pattern since we will always start from a specific offset and search forward
from there, instead of searching for the size we need and try and get it as
close as possible to the offset we want.
I've given this a healthy amount of testing pre-new format stuff, as well as
post-new format stuff. I've booted up my fedora box which is installed on btrfs
with this patch and ran with it for a few days without issues. I've not seen
any performance regressions in any of my tests.
Since the last patch Yan Zheng fixed a problem where we could have overlapping
entries, so updating their offset inline would cause problems. Thanks,
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-07-14 09:29:25 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-03-21 22:11:24 +08:00
|
|
|
/*
|
|
|
|
* This searches the block group for just extents to fill the cluster with.
|
2011-10-14 23:10:36 +08:00
|
|
|
* Try to find a cluster with at least bytes total bytes, at least one
|
|
|
|
* extent of cont1_bytes, and other clusters of at least min_bytes.
|
2011-03-21 22:11:24 +08:00
|
|
|
*/
|
2011-05-26 01:07:37 +08:00
|
|
|
static noinline int
|
|
|
|
setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group,
|
|
|
|
struct btrfs_free_cluster *cluster,
|
|
|
|
struct list_head *bitmaps, u64 offset, u64 bytes,
|
2011-10-14 23:10:36 +08:00
|
|
|
u64 cont1_bytes, u64 min_bytes)
|
2011-03-21 22:11:24 +08:00
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
2011-03-21 22:11:24 +08:00
|
|
|
struct btrfs_free_space *first = NULL;
|
|
|
|
struct btrfs_free_space *entry = NULL;
|
|
|
|
struct btrfs_free_space *last;
|
|
|
|
struct rb_node *node;
|
|
|
|
u64 window_free;
|
|
|
|
u64 max_extent;
|
2011-11-10 21:29:20 +08:00
|
|
|
u64 total_size = 0;
|
2011-03-21 22:11:24 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
entry = tree_search_offset(ctl, offset, 0, 1);
|
2011-03-21 22:11:24 +08:00
|
|
|
if (!entry)
|
|
|
|
return -ENOSPC;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We don't want bitmaps, so just move along until we find a normal
|
|
|
|
* extent entry.
|
|
|
|
*/
|
2011-10-14 23:10:36 +08:00
|
|
|
while (entry->bitmap || entry->bytes < min_bytes) {
|
|
|
|
if (entry->bitmap && list_empty(&entry->list))
|
2011-05-26 01:03:16 +08:00
|
|
|
list_add_tail(&entry->list, bitmaps);
|
2011-03-21 22:11:24 +08:00
|
|
|
node = rb_next(&entry->offset_index);
|
|
|
|
if (!node)
|
|
|
|
return -ENOSPC;
|
|
|
|
entry = rb_entry(node, struct btrfs_free_space, offset_index);
|
|
|
|
}
|
|
|
|
|
|
|
|
window_free = entry->bytes;
|
|
|
|
max_extent = entry->bytes;
|
|
|
|
first = entry;
|
|
|
|
last = entry;
|
|
|
|
|
2011-10-14 23:10:36 +08:00
|
|
|
for (node = rb_next(&entry->offset_index); node;
|
|
|
|
node = rb_next(&entry->offset_index)) {
|
2011-03-21 22:11:24 +08:00
|
|
|
entry = rb_entry(node, struct btrfs_free_space, offset_index);
|
|
|
|
|
2011-05-26 01:03:16 +08:00
|
|
|
if (entry->bitmap) {
|
|
|
|
if (list_empty(&entry->list))
|
|
|
|
list_add_tail(&entry->list, bitmaps);
|
2011-03-21 22:11:24 +08:00
|
|
|
continue;
|
2011-05-26 01:03:16 +08:00
|
|
|
}
|
|
|
|
|
2011-10-14 23:10:36 +08:00
|
|
|
if (entry->bytes < min_bytes)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
last = entry;
|
|
|
|
window_free += entry->bytes;
|
|
|
|
if (entry->bytes > max_extent)
|
2011-03-21 22:11:24 +08:00
|
|
|
max_extent = entry->bytes;
|
|
|
|
}
|
|
|
|
|
2011-10-14 23:10:36 +08:00
|
|
|
if (window_free < bytes || max_extent < cont1_bytes)
|
|
|
|
return -ENOSPC;
|
|
|
|
|
2011-03-21 22:11:24 +08:00
|
|
|
cluster->window_start = first->offset;
|
|
|
|
|
|
|
|
node = &first->offset_index;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* now we've found our entries, pull them out of the free space
|
|
|
|
* cache and put them into the cluster rbtree
|
|
|
|
*/
|
|
|
|
do {
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
entry = rb_entry(node, struct btrfs_free_space, offset_index);
|
|
|
|
node = rb_next(&entry->offset_index);
|
2011-10-14 23:10:36 +08:00
|
|
|
if (entry->bitmap || entry->bytes < min_bytes)
|
2011-03-21 22:11:24 +08:00
|
|
|
continue;
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
rb_erase(&entry->offset_index, &ctl->free_space_offset);
|
2011-03-21 22:11:24 +08:00
|
|
|
ret = tree_insert_offset(&cluster->root, entry->offset,
|
|
|
|
&entry->offset_index, 0);
|
2011-11-10 21:29:20 +08:00
|
|
|
total_size += entry->bytes;
|
2013-08-27 05:14:08 +08:00
|
|
|
ASSERT(!ret); /* -EEXIST; Logic error */
|
2011-03-21 22:11:24 +08:00
|
|
|
} while (node && entry != last);
|
|
|
|
|
|
|
|
cluster->max_size = max_extent;
|
2011-11-10 21:29:20 +08:00
|
|
|
trace_btrfs_setup_cluster(block_group, cluster, total_size, 0);
|
2011-03-21 22:11:24 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This specifically looks for bitmaps that may work in the cluster, we assume
|
|
|
|
* that we have already failed to find extents that will work.
|
|
|
|
*/
|
2011-05-26 01:07:37 +08:00
|
|
|
static noinline int
|
|
|
|
setup_cluster_bitmap(struct btrfs_block_group_cache *block_group,
|
|
|
|
struct btrfs_free_cluster *cluster,
|
|
|
|
struct list_head *bitmaps, u64 offset, u64 bytes,
|
2011-10-14 23:10:36 +08:00
|
|
|
u64 cont1_bytes, u64 min_bytes)
|
2011-03-21 22:11:24 +08:00
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
2015-12-15 23:15:32 +08:00
|
|
|
struct btrfs_free_space *entry = NULL;
|
2011-03-21 22:11:24 +08:00
|
|
|
int ret = -ENOSPC;
|
Btrfs: fix to search one more bitmap for cluster setup
Suppose there are two bitmaps [0, 256], [256, 512] and one extent
[100, 120] in the free space cache, and we want to setup a cluster
with offset=100, bytes=50.
In this case, there will be only one bitmap [256, 512] in the temporary
bitmaps list, and then setup_cluster_bitmap() won't search bitmap [0, 256].
The cause is, the list is constructed in setup_cluster_no_bitmap(),
and only bitmaps with bitmap_entry->offset >= offset will be added
into the list, and the very bitmap that convers offset has
bitmap_entry->offset <= offset.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2011-11-20 20:33:38 +08:00
|
|
|
u64 bitmap_offset = offset_to_bitmap(ctl, offset);
|
2011-03-21 22:11:24 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
if (ctl->total_bitmaps == 0)
|
2011-03-21 22:11:24 +08:00
|
|
|
return -ENOSPC;
|
|
|
|
|
Btrfs: fix to search one more bitmap for cluster setup
Suppose there are two bitmaps [0, 256], [256, 512] and one extent
[100, 120] in the free space cache, and we want to setup a cluster
with offset=100, bytes=50.
In this case, there will be only one bitmap [256, 512] in the temporary
bitmaps list, and then setup_cluster_bitmap() won't search bitmap [0, 256].
The cause is, the list is constructed in setup_cluster_no_bitmap(),
and only bitmaps with bitmap_entry->offset >= offset will be added
into the list, and the very bitmap that convers offset has
bitmap_entry->offset <= offset.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2011-11-20 20:33:38 +08:00
|
|
|
/*
|
|
|
|
* The bitmap that covers offset won't be in the list unless offset
|
|
|
|
* is just its start offset.
|
|
|
|
*/
|
2015-12-15 23:15:32 +08:00
|
|
|
if (!list_empty(bitmaps))
|
|
|
|
entry = list_first_entry(bitmaps, struct btrfs_free_space, list);
|
|
|
|
|
|
|
|
if (!entry || entry->offset != bitmap_offset) {
|
Btrfs: fix to search one more bitmap for cluster setup
Suppose there are two bitmaps [0, 256], [256, 512] and one extent
[100, 120] in the free space cache, and we want to setup a cluster
with offset=100, bytes=50.
In this case, there will be only one bitmap [256, 512] in the temporary
bitmaps list, and then setup_cluster_bitmap() won't search bitmap [0, 256].
The cause is, the list is constructed in setup_cluster_no_bitmap(),
and only bitmaps with bitmap_entry->offset >= offset will be added
into the list, and the very bitmap that convers offset has
bitmap_entry->offset <= offset.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2011-11-20 20:33:38 +08:00
|
|
|
entry = tree_search_offset(ctl, bitmap_offset, 1, 0);
|
|
|
|
if (entry && list_empty(&entry->list))
|
|
|
|
list_add(&entry->list, bitmaps);
|
|
|
|
}
|
|
|
|
|
2011-05-26 01:03:16 +08:00
|
|
|
list_for_each_entry(entry, bitmaps, list) {
|
2012-01-27 04:01:11 +08:00
|
|
|
if (entry->bytes < bytes)
|
2011-05-26 01:03:16 +08:00
|
|
|
continue;
|
|
|
|
ret = btrfs_bitmap_cluster(block_group, entry, cluster, offset,
|
2011-10-14 23:10:36 +08:00
|
|
|
bytes, cont1_bytes, min_bytes);
|
2011-05-26 01:03:16 +08:00
|
|
|
if (!ret)
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2011-11-20 20:33:38 +08:00
|
|
|
* The bitmaps list has all the bitmaps that record free space
|
|
|
|
* starting after offset, so no more search is required.
|
2011-05-26 01:03:16 +08:00
|
|
|
*/
|
2011-11-20 20:33:38 +08:00
|
|
|
return -ENOSPC;
|
2011-03-21 22:11:24 +08:00
|
|
|
}
|
|
|
|
|
2009-04-03 21:47:43 +08:00
|
|
|
/*
|
|
|
|
* here we try to find a cluster of blocks in a block group. The goal
|
2011-10-14 23:10:36 +08:00
|
|
|
* is to find at least bytes+empty_size.
|
2009-04-03 21:47:43 +08:00
|
|
|
* We might not find them all in one contiguous area.
|
|
|
|
*
|
|
|
|
* returns zero and sets up cluster if things worked out, otherwise
|
|
|
|
* it returns -enospc
|
|
|
|
*/
|
2013-08-15 02:02:47 +08:00
|
|
|
int btrfs_find_space_cluster(struct btrfs_root *root,
|
2009-04-03 21:47:43 +08:00
|
|
|
struct btrfs_block_group_cache *block_group,
|
|
|
|
struct btrfs_free_cluster *cluster,
|
|
|
|
u64 offset, u64 bytes, u64 empty_size)
|
|
|
|
{
|
2011-03-29 13:46:06 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
2011-05-26 01:03:16 +08:00
|
|
|
struct btrfs_free_space *entry, *tmp;
|
2011-11-20 20:33:38 +08:00
|
|
|
LIST_HEAD(bitmaps);
|
2009-04-03 21:47:43 +08:00
|
|
|
u64 min_bytes;
|
2011-10-14 23:10:36 +08:00
|
|
|
u64 cont1_bytes;
|
2009-04-03 21:47:43 +08:00
|
|
|
int ret;
|
|
|
|
|
2011-10-14 23:10:36 +08:00
|
|
|
/*
|
|
|
|
* Choose the minimum extent size we'll require for this
|
|
|
|
* cluster. For SSD_SPREAD, don't allow any fragmentation.
|
|
|
|
* For metadata, allow allocates with smaller extents. For
|
|
|
|
* data, keep it dense.
|
|
|
|
*/
|
2009-06-10 08:28:34 +08:00
|
|
|
if (btrfs_test_opt(root, SSD_SPREAD)) {
|
2011-10-14 23:10:36 +08:00
|
|
|
cont1_bytes = min_bytes = bytes + empty_size;
|
2009-06-10 08:28:34 +08:00
|
|
|
} else if (block_group->flags & BTRFS_BLOCK_GROUP_METADATA) {
|
2011-10-14 23:10:36 +08:00
|
|
|
cont1_bytes = bytes;
|
|
|
|
min_bytes = block_group->sectorsize;
|
|
|
|
} else {
|
|
|
|
cont1_bytes = max(bytes, (bytes + empty_size) >> 2);
|
|
|
|
min_bytes = block_group->sectorsize;
|
|
|
|
}
|
2009-04-03 21:47:43 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_lock(&ctl->tree_lock);
|
2011-03-19 03:13:42 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* If we know we don't have enough space to make a cluster don't even
|
|
|
|
* bother doing all the work to try and find one.
|
|
|
|
*/
|
2011-10-14 23:10:36 +08:00
|
|
|
if (ctl->free_space < bytes) {
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_unlock(&ctl->tree_lock);
|
2011-03-19 03:13:42 +08:00
|
|
|
return -ENOSPC;
|
|
|
|
}
|
|
|
|
|
2009-04-03 21:47:43 +08:00
|
|
|
spin_lock(&cluster->lock);
|
|
|
|
|
|
|
|
/* someone already found a cluster, hooray */
|
|
|
|
if (cluster->block_group) {
|
|
|
|
ret = 0;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2011-11-10 21:29:20 +08:00
|
|
|
trace_btrfs_find_cluster(block_group, offset, bytes, empty_size,
|
|
|
|
min_bytes);
|
|
|
|
|
2011-05-26 01:03:16 +08:00
|
|
|
ret = setup_cluster_no_bitmap(block_group, cluster, &bitmaps, offset,
|
2011-10-14 23:10:36 +08:00
|
|
|
bytes + empty_size,
|
|
|
|
cont1_bytes, min_bytes);
|
2011-03-21 22:11:24 +08:00
|
|
|
if (ret)
|
2011-05-26 01:03:16 +08:00
|
|
|
ret = setup_cluster_bitmap(block_group, cluster, &bitmaps,
|
2011-10-14 23:10:36 +08:00
|
|
|
offset, bytes + empty_size,
|
|
|
|
cont1_bytes, min_bytes);
|
2011-05-26 01:03:16 +08:00
|
|
|
|
|
|
|
/* Clear our temporary list */
|
|
|
|
list_for_each_entry_safe(entry, tmp, &bitmaps, list)
|
|
|
|
list_del_init(&entry->list);
|
2009-04-03 21:47:43 +08:00
|
|
|
|
2011-03-21 22:11:24 +08:00
|
|
|
if (!ret) {
|
|
|
|
atomic_inc(&block_group->count);
|
|
|
|
list_add_tail(&cluster->block_group_list,
|
|
|
|
&block_group->cluster_list);
|
|
|
|
cluster->block_group = block_group;
|
2011-11-10 21:29:20 +08:00
|
|
|
} else {
|
|
|
|
trace_btrfs_failed_cluster_setup(block_group);
|
2009-04-03 21:47:43 +08:00
|
|
|
}
|
|
|
|
out:
|
|
|
|
spin_unlock(&cluster->lock);
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_unlock(&ctl->tree_lock);
|
2009-04-03 21:47:43 +08:00
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* simple code to zero out a cluster
|
|
|
|
*/
|
|
|
|
void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster)
|
|
|
|
{
|
|
|
|
spin_lock_init(&cluster->lock);
|
|
|
|
spin_lock_init(&cluster->refill_lock);
|
2010-02-24 03:43:04 +08:00
|
|
|
cluster->root = RB_ROOT;
|
2009-04-03 21:47:43 +08:00
|
|
|
cluster->max_size = 0;
|
2015-10-03 03:25:10 +08:00
|
|
|
cluster->fragmented = false;
|
2009-04-03 21:47:43 +08:00
|
|
|
INIT_LIST_HEAD(&cluster->block_group_list);
|
|
|
|
cluster->block_group = NULL;
|
|
|
|
}
|
|
|
|
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
static int do_trimming(struct btrfs_block_group_cache *block_group,
|
|
|
|
u64 *total_trimmed, u64 start, u64 bytes,
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
u64 reserved_start, u64 reserved_bytes,
|
|
|
|
struct btrfs_trim_range *trim_entry)
|
2011-03-24 18:24:28 +08:00
|
|
|
{
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
struct btrfs_space_info *space_info = block_group->space_info;
|
2011-03-24 18:24:28 +08:00
|
|
|
struct btrfs_fs_info *fs_info = block_group->fs_info;
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
int ret;
|
|
|
|
int update = 0;
|
|
|
|
u64 trimmed = 0;
|
2011-03-24 18:24:28 +08:00
|
|
|
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
spin_lock(&space_info->lock);
|
|
|
|
spin_lock(&block_group->lock);
|
|
|
|
if (!block_group->ro) {
|
|
|
|
block_group->reserved += reserved_bytes;
|
|
|
|
space_info->bytes_reserved += reserved_bytes;
|
|
|
|
update = 1;
|
|
|
|
}
|
|
|
|
spin_unlock(&block_group->lock);
|
|
|
|
spin_unlock(&space_info->lock);
|
|
|
|
|
2014-12-08 22:01:12 +08:00
|
|
|
ret = btrfs_discard_extent(fs_info->extent_root,
|
|
|
|
start, bytes, &trimmed);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
if (!ret)
|
|
|
|
*total_trimmed += trimmed;
|
|
|
|
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
mutex_lock(&ctl->cache_writeout_mutex);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
btrfs_add_free_space(block_group, reserved_start, reserved_bytes);
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
list_del(&trim_entry->list);
|
|
|
|
mutex_unlock(&ctl->cache_writeout_mutex);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
|
|
|
|
if (update) {
|
|
|
|
spin_lock(&space_info->lock);
|
|
|
|
spin_lock(&block_group->lock);
|
|
|
|
if (block_group->ro)
|
|
|
|
space_info->bytes_readonly += reserved_bytes;
|
|
|
|
block_group->reserved -= reserved_bytes;
|
|
|
|
space_info->bytes_reserved -= reserved_bytes;
|
|
|
|
spin_unlock(&space_info->lock);
|
|
|
|
spin_unlock(&block_group->lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int trim_no_bitmap(struct btrfs_block_group_cache *block_group,
|
|
|
|
u64 *total_trimmed, u64 start, u64 end, u64 minlen)
|
|
|
|
{
|
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
|
|
|
struct btrfs_free_space *entry;
|
|
|
|
struct rb_node *node;
|
|
|
|
int ret = 0;
|
|
|
|
u64 extent_start;
|
|
|
|
u64 extent_bytes;
|
|
|
|
u64 bytes;
|
2011-03-24 18:24:28 +08:00
|
|
|
|
|
|
|
while (start < end) {
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
struct btrfs_trim_range trim_entry;
|
|
|
|
|
|
|
|
mutex_lock(&ctl->cache_writeout_mutex);
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_lock(&ctl->tree_lock);
|
2011-03-24 18:24:28 +08:00
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
if (ctl->free_space < minlen) {
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
mutex_unlock(&ctl->cache_writeout_mutex);
|
2011-03-24 18:24:28 +08:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
entry = tree_search_offset(ctl, start, 0, 1);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
if (!entry) {
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
mutex_unlock(&ctl->cache_writeout_mutex);
|
2011-03-24 18:24:28 +08:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
/* skip bitmaps */
|
|
|
|
while (entry->bitmap) {
|
|
|
|
node = rb_next(&entry->offset_index);
|
|
|
|
if (!node) {
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
mutex_unlock(&ctl->cache_writeout_mutex);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
goto out;
|
2011-03-24 18:24:28 +08:00
|
|
|
}
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
entry = rb_entry(node, struct btrfs_free_space,
|
|
|
|
offset_index);
|
2011-03-24 18:24:28 +08:00
|
|
|
}
|
|
|
|
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
if (entry->offset >= end) {
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
mutex_unlock(&ctl->cache_writeout_mutex);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
break;
|
2011-03-24 18:24:28 +08:00
|
|
|
}
|
|
|
|
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
extent_start = entry->offset;
|
|
|
|
extent_bytes = entry->bytes;
|
|
|
|
start = max(start, extent_start);
|
|
|
|
bytes = min(extent_start + extent_bytes, end) - start;
|
|
|
|
if (bytes < minlen) {
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
mutex_unlock(&ctl->cache_writeout_mutex);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
goto next;
|
2011-03-24 18:24:28 +08:00
|
|
|
}
|
|
|
|
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
unlink_free_space(ctl, entry);
|
|
|
|
kmem_cache_free(btrfs_free_space_cachep, entry);
|
|
|
|
|
2011-03-29 13:46:06 +08:00
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
trim_entry.start = extent_start;
|
|
|
|
trim_entry.bytes = extent_bytes;
|
|
|
|
list_add_tail(&trim_entry.list, &ctl->trimming_ranges);
|
|
|
|
mutex_unlock(&ctl->cache_writeout_mutex);
|
2011-03-24 18:24:28 +08:00
|
|
|
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
ret = do_trimming(block_group, total_trimmed, start, bytes,
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
extent_start, extent_bytes, &trim_entry);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
if (ret)
|
|
|
|
break;
|
|
|
|
next:
|
|
|
|
start += bytes;
|
2011-03-24 18:24:28 +08:00
|
|
|
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
if (fatal_signal_pending(current)) {
|
|
|
|
ret = -ERESTARTSYS;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
cond_resched();
|
|
|
|
}
|
|
|
|
out:
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int trim_bitmaps(struct btrfs_block_group_cache *block_group,
|
|
|
|
u64 *total_trimmed, u64 start, u64 end, u64 minlen)
|
|
|
|
{
|
|
|
|
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
|
|
|
|
struct btrfs_free_space *entry;
|
|
|
|
int ret = 0;
|
|
|
|
int ret2;
|
|
|
|
u64 bytes;
|
|
|
|
u64 offset = offset_to_bitmap(ctl, start);
|
|
|
|
|
|
|
|
while (offset < end) {
|
|
|
|
bool next_bitmap = false;
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
struct btrfs_trim_range trim_entry;
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
mutex_lock(&ctl->cache_writeout_mutex);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
spin_lock(&ctl->tree_lock);
|
|
|
|
|
|
|
|
if (ctl->free_space < minlen) {
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
mutex_unlock(&ctl->cache_writeout_mutex);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
entry = tree_search_offset(ctl, offset, 1, 0);
|
|
|
|
if (!entry) {
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
mutex_unlock(&ctl->cache_writeout_mutex);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
next_bitmap = true;
|
|
|
|
goto next;
|
|
|
|
}
|
|
|
|
|
|
|
|
bytes = minlen;
|
2015-10-03 04:12:23 +08:00
|
|
|
ret2 = search_bitmap(ctl, entry, &start, &bytes, false);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
if (ret2 || start >= end) {
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
mutex_unlock(&ctl->cache_writeout_mutex);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
next_bitmap = true;
|
|
|
|
goto next;
|
|
|
|
}
|
|
|
|
|
|
|
|
bytes = min(bytes, end - start);
|
|
|
|
if (bytes < minlen) {
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
mutex_unlock(&ctl->cache_writeout_mutex);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
goto next;
|
|
|
|
}
|
|
|
|
|
|
|
|
bitmap_clear_bits(ctl, entry, start, bytes);
|
|
|
|
if (entry->bytes == 0)
|
|
|
|
free_bitmap(ctl, entry);
|
|
|
|
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
trim_entry.start = start;
|
|
|
|
trim_entry.bytes = bytes;
|
|
|
|
list_add_tail(&trim_entry.list, &ctl->trimming_ranges);
|
|
|
|
mutex_unlock(&ctl->cache_writeout_mutex);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
|
|
|
|
ret = do_trimming(block_group, total_trimmed, start, bytes,
|
Btrfs: fix race between writing free space cache and trimming
Trimming is completely transactionless, and the way it operates consists
of hiding free space entries from a block group, perform the trim/discard
and then make the free space entries visible again.
Therefore while a free space entry is being trimmed, we can have free space
cache writing running in parallel (as part of a transaction commit) which
will miss the free space entry. This means that an unmount (or crash/reboot)
after that transaction commit and mount again before another transaction
starts/commits after the discard finishes, we will have some free space
that won't be used again unless the free space cache is rebuilt. After the
unmount, fsck (btrfsck, btrfs check) reports the issue like the following
example:
*** fsck.btrfs output ***
checking extents
checking free space cache
There is no free space entry for 521764864-521781248
There is no free space entry for 521764864-1103101952
cache appears valid but isnt 29360128
Checking filesystem on /dev/sdc
UUID: b4789e27-4774-4626-98e9-ae8dfbfb0fb5
found 1235681286 bytes used err is -22
(...)
Another issue caused by this race is a crash while writing bitmap entries
to the cache, because while the cache writeout task accesses the bitmaps,
the trim task can be concurrently modifying the bitmap or worse might
be freeing the bitmap. The later case results in the following crash:
[55650.804460] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC
[55650.804835] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc loop parport_pc parport i2c_piix4 psmouse evdev pcspkr microcode processor i2ccore serio_raw thermal_sys button ext4 crc16 jbd2 mbcache sg sd_mod crc_t10dif sr_mod cdrom crct10dif_generic crct10dif_common ata_generic virtio_scsi floppy ata_piix libata virtio_pci virtio_ring virtio scsi_mod e1000 [last unloaded: btrfs]
[55650.806169] CPU: 1 PID: 31002 Comm: btrfs-transacti Tainted: G W 3.17.0-rc5-btrfs-next-1+ #1
[55650.806493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[55650.806867] task: ffff8800b12f6410 ti: ffff880071538000 task.ti: ffff880071538000
[55650.807166] RIP: 0010:[<ffffffffa037cf45>] [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.807514] RSP: 0018:ffff88007153bc30 EFLAGS: 00010246
[55650.807687] RAX: 000000005d1ec000 RBX: ffff8800a665df08 RCX: 0000000000000400
[55650.807885] RDX: ffff88005d1ec000 RSI: 6b6b6b6b6b6b6b6b RDI: ffff88005d1ec000
[55650.808017] RBP: ffff88007153bc58 R08: 00000000ddd51536 R09: 00000000000001e0
[55650.808017] R10: 0000000000000000 R11: 0000000000000037 R12: 6b6b6b6b6b6b6b6b
[55650.808017] R13: ffff88007153bca8 R14: 6b6b6b6b6b6b6b6b R15: ffff88007153bc98
[55650.808017] FS: 0000000000000000(0000) GS:ffff88023ec80000(0000) knlGS:0000000000000000
[55650.808017] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[55650.808017] CR2: 0000000002273b88 CR3: 00000000b18f6000 CR4: 00000000000006e0
[55650.808017] Stack:
[55650.808017] ffff88020e834e00 ffff880172d68db0 0000000000000000 ffff88019257c800
[55650.808017] ffff8801d42ea720 ffff88007153bd10 ffffffffa037d2fa ffff880224e99180
[55650.808017] ffff8801469a6188 ffff880224e99140 ffff880172d68c50 00000003000000b7
[55650.808017] Call Trace:
[55650.808017] [<ffffffffa037d2fa>] __btrfs_write_out_cache+0x1ea/0x37f [btrfs]
[55650.808017] [<ffffffffa037d959>] btrfs_write_out_cache+0xa1/0xd8 [btrfs]
[55650.808017] [<ffffffffa033936b>] btrfs_write_dirty_block_groups+0x4b5/0x505 [btrfs]
[55650.808017] [<ffffffffa03aa98e>] commit_cowonly_roots+0x15e/0x1f7 [btrfs]
[55650.808017] [<ffffffff813eb9c7>] ? _raw_spin_lock+0xe/0x10
[55650.808017] [<ffffffffa0346e46>] btrfs_commit_transaction+0x411/0x882 [btrfs]
[55650.808017] [<ffffffffa03432a4>] transaction_kthread+0xf2/0x1a4 [btrfs]
[55650.808017] [<ffffffffa03431b2>] ? btrfs_cleanup_transaction+0x3d8/0x3d8 [btrfs]
[55650.808017] [<ffffffff8105966b>] kthread+0xb7/0xbf
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] [<ffffffff813ebeac>] ret_from_fork+0x7c/0xb0
[55650.808017] [<ffffffff810595b4>] ? __kthread_parkme+0x67/0x67
[55650.808017] Code: 4c 89 ef 8d 70 ff e8 d4 fc ff ff 41 8b 45 34 41 39 45 30 7d 5c 31 f6 4c 89 ef e8 80 f6 ff ff 49 8b 7d 00 4c 89 f6 b9 00 04 00 00 <f3> a5 4c 89 ef 41 8b 45 30 8d 70 ff e8 a3 fc ff ff 41 8b 45 34
[55650.808017] RIP [<ffffffffa037cf45>] write_bitmap_entries+0x65/0xbb [btrfs]
[55650.808017] RSP <ffff88007153bc30>
[55650.815725] ---[ end trace 1c032e96b149ff86 ]---
Fix this by serializing both tasks in such a way that cache writeout
doesn't wait for the trim/discard of free space entries to finish and
doesn't miss any free space entry.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-12-02 01:04:09 +08:00
|
|
|
start, bytes, &trim_entry);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
if (ret)
|
|
|
|
break;
|
|
|
|
next:
|
|
|
|
if (next_bitmap) {
|
|
|
|
offset += BITS_PER_BITMAP * ctl->unit;
|
|
|
|
} else {
|
|
|
|
start += bytes;
|
|
|
|
if (start >= offset + BITS_PER_BITMAP * ctl->unit)
|
|
|
|
offset += BITS_PER_BITMAP * ctl->unit;
|
2011-03-24 18:24:28 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
if (fatal_signal_pending(current)) {
|
|
|
|
ret = -ERESTARTSYS;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
cond_resched();
|
|
|
|
}
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
Btrfs: Cache free inode numbers in memory
Currently btrfs stores the highest objectid of the fs tree, and it always
returns (highest+1) inode number when we create a file, so inode numbers
won't be reclaimed when we delete files, so we'll run out of inode numbers
as we keep create/delete files in 32bits machines.
This fixes it, and it works similarly to how we cache free space in block
cgroups.
We start a kernel thread to read the file tree. By scanning inode items,
we know which chunks of inode numbers are free, and we cache them in
an rb-tree.
Because we are searching the commit root, we have to carefully handle the
cross-transaction case.
The rb-tree is a hybrid extent+bitmap tree, so if we have too many small
chunks of inode numbers, we'll use bitmaps. Initially we allow 16K ram
of extents, and a bitmap will be used if we exceed this threshold. The
extents threshold is adjusted in runtime.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-04-20 10:06:11 +08:00
|
|
|
|
2015-06-15 21:41:19 +08:00
|
|
|
void btrfs_get_block_group_trimming(struct btrfs_block_group_cache *cache)
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
{
|
2015-06-15 21:41:19 +08:00
|
|
|
atomic_inc(&cache->trimming);
|
|
|
|
}
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
|
2015-06-15 21:41:19 +08:00
|
|
|
void btrfs_put_block_group_trimming(struct btrfs_block_group_cache *block_group)
|
|
|
|
{
|
|
|
|
struct extent_map_tree *em_tree;
|
|
|
|
struct extent_map *em;
|
|
|
|
bool cleanup;
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
|
Btrfs: fix race between fs trimming and block group remove/allocation
Our fs trim operation, which is completely transactionless (doesn't start
or joins an existing transaction) consists of visiting all block groups
and then for each one to iterate its free space entries and perform a
discard operation against the space range represented by the free space
entries. However before performing a discard, the corresponding free space
entry is removed from the free space rbtree, and when the discard completes
it is added back to the free space rbtree.
If a block group remove operation happens while the discard is ongoing (or
before it starts and after a free space entry is hidden), we end up not
waiting for the discard to complete, remove the extent map that maps
logical address to physical addresses and the corresponding chunk metadata
from the the chunk and device trees. After that and before the discard
completes, the current running transaction can finish and a new one start,
allowing for new block groups that map to the same physical addresses to
be allocated and written to.
So fix this by keeping the extent map in memory until the discard completes
so that the same physical addresses aren't reused before it completes.
If the physical locations that are under a discard operation end up being
used for a new metadata block group for example, and dirty metadata extents
are written before the discard finishes (the VM might call writepages() of
our btree inode's i_mapping for example, or an fsync log commit happens) we
end up overwriting metadata with zeroes, which leads to errors from fsck
like the following:
checking extents
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
read block failed check_tree_block
owner ref check failed [833912832 16384]
Errors found in extent allocation tree or chunk allocation
checking free space cache
checking fs roots
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
read block failed check_tree_block
root 5 root dir 256 error
root 5 inode 260 errors 2001, no inode item, link count wrong
unresolved ref dir 256 index 0 namelen 8 name foobar_3 filetype 1 errors 6, no dir index, no inode ref
root 5 inode 262 errors 2001, no inode item, link count wrong
unresolved ref dir 256 index 0 namelen 8 name foobar_5 filetype 1 errors 6, no dir index, no inode ref
root 5 inode 263 errors 2001, no inode item, link count wrong
(...)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-11-28 05:14:15 +08:00
|
|
|
spin_lock(&block_group->lock);
|
2015-06-15 21:41:19 +08:00
|
|
|
cleanup = (atomic_dec_and_test(&block_group->trimming) &&
|
|
|
|
block_group->removed);
|
Btrfs: fix race between fs trimming and block group remove/allocation
Our fs trim operation, which is completely transactionless (doesn't start
or joins an existing transaction) consists of visiting all block groups
and then for each one to iterate its free space entries and perform a
discard operation against the space range represented by the free space
entries. However before performing a discard, the corresponding free space
entry is removed from the free space rbtree, and when the discard completes
it is added back to the free space rbtree.
If a block group remove operation happens while the discard is ongoing (or
before it starts and after a free space entry is hidden), we end up not
waiting for the discard to complete, remove the extent map that maps
logical address to physical addresses and the corresponding chunk metadata
from the the chunk and device trees. After that and before the discard
completes, the current running transaction can finish and a new one start,
allowing for new block groups that map to the same physical addresses to
be allocated and written to.
So fix this by keeping the extent map in memory until the discard completes
so that the same physical addresses aren't reused before it completes.
If the physical locations that are under a discard operation end up being
used for a new metadata block group for example, and dirty metadata extents
are written before the discard finishes (the VM might call writepages() of
our btree inode's i_mapping for example, or an fsync log commit happens) we
end up overwriting metadata with zeroes, which leads to errors from fsck
like the following:
checking extents
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
read block failed check_tree_block
owner ref check failed [833912832 16384]
Errors found in extent allocation tree or chunk allocation
checking free space cache
checking fs roots
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
read block failed check_tree_block
root 5 root dir 256 error
root 5 inode 260 errors 2001, no inode item, link count wrong
unresolved ref dir 256 index 0 namelen 8 name foobar_3 filetype 1 errors 6, no dir index, no inode ref
root 5 inode 262 errors 2001, no inode item, link count wrong
unresolved ref dir 256 index 0 namelen 8 name foobar_5 filetype 1 errors 6, no dir index, no inode ref
root 5 inode 263 errors 2001, no inode item, link count wrong
(...)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-11-28 05:14:15 +08:00
|
|
|
spin_unlock(&block_group->lock);
|
|
|
|
|
2015-06-15 21:41:19 +08:00
|
|
|
if (cleanup) {
|
2014-12-04 23:31:01 +08:00
|
|
|
lock_chunks(block_group->fs_info->chunk_root);
|
Btrfs: fix race between fs trimming and block group remove/allocation
Our fs trim operation, which is completely transactionless (doesn't start
or joins an existing transaction) consists of visiting all block groups
and then for each one to iterate its free space entries and perform a
discard operation against the space range represented by the free space
entries. However before performing a discard, the corresponding free space
entry is removed from the free space rbtree, and when the discard completes
it is added back to the free space rbtree.
If a block group remove operation happens while the discard is ongoing (or
before it starts and after a free space entry is hidden), we end up not
waiting for the discard to complete, remove the extent map that maps
logical address to physical addresses and the corresponding chunk metadata
from the the chunk and device trees. After that and before the discard
completes, the current running transaction can finish and a new one start,
allowing for new block groups that map to the same physical addresses to
be allocated and written to.
So fix this by keeping the extent map in memory until the discard completes
so that the same physical addresses aren't reused before it completes.
If the physical locations that are under a discard operation end up being
used for a new metadata block group for example, and dirty metadata extents
are written before the discard finishes (the VM might call writepages() of
our btree inode's i_mapping for example, or an fsync log commit happens) we
end up overwriting metadata with zeroes, which leads to errors from fsck
like the following:
checking extents
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
read block failed check_tree_block
owner ref check failed [833912832 16384]
Errors found in extent allocation tree or chunk allocation
checking free space cache
checking fs roots
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
read block failed check_tree_block
root 5 root dir 256 error
root 5 inode 260 errors 2001, no inode item, link count wrong
unresolved ref dir 256 index 0 namelen 8 name foobar_3 filetype 1 errors 6, no dir index, no inode ref
root 5 inode 262 errors 2001, no inode item, link count wrong
unresolved ref dir 256 index 0 namelen 8 name foobar_5 filetype 1 errors 6, no dir index, no inode ref
root 5 inode 263 errors 2001, no inode item, link count wrong
(...)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-11-28 05:14:15 +08:00
|
|
|
em_tree = &block_group->fs_info->mapping_tree.map_tree;
|
|
|
|
write_lock(&em_tree->lock);
|
|
|
|
em = lookup_extent_mapping(em_tree, block_group->key.objectid,
|
|
|
|
1);
|
|
|
|
BUG_ON(!em); /* logic error, can't happen */
|
2014-12-04 23:31:01 +08:00
|
|
|
/*
|
|
|
|
* remove_extent_mapping() will delete us from the pinned_chunks
|
|
|
|
* list, which is protected by the chunk mutex.
|
|
|
|
*/
|
Btrfs: fix race between fs trimming and block group remove/allocation
Our fs trim operation, which is completely transactionless (doesn't start
or joins an existing transaction) consists of visiting all block groups
and then for each one to iterate its free space entries and perform a
discard operation against the space range represented by the free space
entries. However before performing a discard, the corresponding free space
entry is removed from the free space rbtree, and when the discard completes
it is added back to the free space rbtree.
If a block group remove operation happens while the discard is ongoing (or
before it starts and after a free space entry is hidden), we end up not
waiting for the discard to complete, remove the extent map that maps
logical address to physical addresses and the corresponding chunk metadata
from the the chunk and device trees. After that and before the discard
completes, the current running transaction can finish and a new one start,
allowing for new block groups that map to the same physical addresses to
be allocated and written to.
So fix this by keeping the extent map in memory until the discard completes
so that the same physical addresses aren't reused before it completes.
If the physical locations that are under a discard operation end up being
used for a new metadata block group for example, and dirty metadata extents
are written before the discard finishes (the VM might call writepages() of
our btree inode's i_mapping for example, or an fsync log commit happens) we
end up overwriting metadata with zeroes, which leads to errors from fsck
like the following:
checking extents
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
read block failed check_tree_block
owner ref check failed [833912832 16384]
Errors found in extent allocation tree or chunk allocation
checking free space cache
checking fs roots
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
read block failed check_tree_block
root 5 root dir 256 error
root 5 inode 260 errors 2001, no inode item, link count wrong
unresolved ref dir 256 index 0 namelen 8 name foobar_3 filetype 1 errors 6, no dir index, no inode ref
root 5 inode 262 errors 2001, no inode item, link count wrong
unresolved ref dir 256 index 0 namelen 8 name foobar_5 filetype 1 errors 6, no dir index, no inode ref
root 5 inode 263 errors 2001, no inode item, link count wrong
(...)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-11-28 05:14:15 +08:00
|
|
|
remove_extent_mapping(em_tree, em);
|
|
|
|
write_unlock(&em_tree->lock);
|
|
|
|
unlock_chunks(block_group->fs_info->chunk_root);
|
|
|
|
|
|
|
|
/* once for us and once for the tree */
|
|
|
|
free_extent_map(em);
|
|
|
|
free_extent_map(em);
|
2014-12-02 01:04:40 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* We've left one free space entry and other tasks trimming
|
|
|
|
* this block group have left 1 entry each one. Free them.
|
|
|
|
*/
|
|
|
|
__btrfs_remove_free_space_cache(block_group->free_space_ctl);
|
2015-06-15 21:41:19 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group,
|
|
|
|
u64 *trimmed, u64 start, u64 end, u64 minlen)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
*trimmed = 0;
|
|
|
|
|
|
|
|
spin_lock(&block_group->lock);
|
|
|
|
if (block_group->removed) {
|
Btrfs: fix race between fs trimming and block group remove/allocation
Our fs trim operation, which is completely transactionless (doesn't start
or joins an existing transaction) consists of visiting all block groups
and then for each one to iterate its free space entries and perform a
discard operation against the space range represented by the free space
entries. However before performing a discard, the corresponding free space
entry is removed from the free space rbtree, and when the discard completes
it is added back to the free space rbtree.
If a block group remove operation happens while the discard is ongoing (or
before it starts and after a free space entry is hidden), we end up not
waiting for the discard to complete, remove the extent map that maps
logical address to physical addresses and the corresponding chunk metadata
from the the chunk and device trees. After that and before the discard
completes, the current running transaction can finish and a new one start,
allowing for new block groups that map to the same physical addresses to
be allocated and written to.
So fix this by keeping the extent map in memory until the discard completes
so that the same physical addresses aren't reused before it completes.
If the physical locations that are under a discard operation end up being
used for a new metadata block group for example, and dirty metadata extents
are written before the discard finishes (the VM might call writepages() of
our btree inode's i_mapping for example, or an fsync log commit happens) we
end up overwriting metadata with zeroes, which leads to errors from fsck
like the following:
checking extents
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
read block failed check_tree_block
owner ref check failed [833912832 16384]
Errors found in extent allocation tree or chunk allocation
checking free space cache
checking fs roots
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
read block failed check_tree_block
root 5 root dir 256 error
root 5 inode 260 errors 2001, no inode item, link count wrong
unresolved ref dir 256 index 0 namelen 8 name foobar_3 filetype 1 errors 6, no dir index, no inode ref
root 5 inode 262 errors 2001, no inode item, link count wrong
unresolved ref dir 256 index 0 namelen 8 name foobar_5 filetype 1 errors 6, no dir index, no inode ref
root 5 inode 263 errors 2001, no inode item, link count wrong
(...)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-11-28 05:14:15 +08:00
|
|
|
spin_unlock(&block_group->lock);
|
2015-06-15 21:41:19 +08:00
|
|
|
return 0;
|
Btrfs: fix race between fs trimming and block group remove/allocation
Our fs trim operation, which is completely transactionless (doesn't start
or joins an existing transaction) consists of visiting all block groups
and then for each one to iterate its free space entries and perform a
discard operation against the space range represented by the free space
entries. However before performing a discard, the corresponding free space
entry is removed from the free space rbtree, and when the discard completes
it is added back to the free space rbtree.
If a block group remove operation happens while the discard is ongoing (or
before it starts and after a free space entry is hidden), we end up not
waiting for the discard to complete, remove the extent map that maps
logical address to physical addresses and the corresponding chunk metadata
from the the chunk and device trees. After that and before the discard
completes, the current running transaction can finish and a new one start,
allowing for new block groups that map to the same physical addresses to
be allocated and written to.
So fix this by keeping the extent map in memory until the discard completes
so that the same physical addresses aren't reused before it completes.
If the physical locations that are under a discard operation end up being
used for a new metadata block group for example, and dirty metadata extents
are written before the discard finishes (the VM might call writepages() of
our btree inode's i_mapping for example, or an fsync log commit happens) we
end up overwriting metadata with zeroes, which leads to errors from fsck
like the following:
checking extents
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
read block failed check_tree_block
owner ref check failed [833912832 16384]
Errors found in extent allocation tree or chunk allocation
checking free space cache
checking fs roots
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
Check tree block failed, want=833912832, have=0
read block failed check_tree_block
root 5 root dir 256 error
root 5 inode 260 errors 2001, no inode item, link count wrong
unresolved ref dir 256 index 0 namelen 8 name foobar_3 filetype 1 errors 6, no dir index, no inode ref
root 5 inode 262 errors 2001, no inode item, link count wrong
unresolved ref dir 256 index 0 namelen 8 name foobar_5 filetype 1 errors 6, no dir index, no inode ref
root 5 inode 263 errors 2001, no inode item, link count wrong
(...)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-11-28 05:14:15 +08:00
|
|
|
}
|
2015-06-15 21:41:19 +08:00
|
|
|
btrfs_get_block_group_trimming(block_group);
|
|
|
|
spin_unlock(&block_group->lock);
|
|
|
|
|
|
|
|
ret = trim_no_bitmap(block_group, trimmed, start, end, minlen);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
|
2015-06-15 21:41:19 +08:00
|
|
|
ret = trim_bitmaps(block_group, trimmed, start, end, minlen);
|
|
|
|
out:
|
|
|
|
btrfs_put_block_group_trimming(block_group);
|
Btrfs: rewrite btrfs_trim_block_group()
There are various bugs in block group trimming:
- It may trim from offset smaller than user-specified offset.
- It may trim beyond user-specified range.
- It may leak free space for extents smaller than specified minlen.
- It may truncate the last trimmed extent thus leak free space.
- With mixed extents+bitmaps, some extents may not be trimmed.
- With mixed extents+bitmaps, some bitmaps may not be trimmed (even
none will be trimmed). Even for those trimmed, not all the free space
in the bitmaps will be trimmed.
I rewrite btrfs_trim_block_group() and break it into two functions.
One is to trim extents only, and the other is to trim bitmaps only.
Before patching:
# fstrim -v /mnt/
/mnt/: 1496465408 bytes were trimmed
After patching:
# fstrim -v /mnt/
/mnt/: 2193768448 bytes were trimmed
And this matches the total free space:
# btrfs fi df /mnt
Data: total=3.58GB, used=1.79GB
System, DUP: total=8.00MB, used=4.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=205.12MB, used=97.14MB
Metadata: total=8.00MB, used=0.00
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-12-29 14:47:27 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
Btrfs: Cache free inode numbers in memory
Currently btrfs stores the highest objectid of the fs tree, and it always
returns (highest+1) inode number when we create a file, so inode numbers
won't be reclaimed when we delete files, so we'll run out of inode numbers
as we keep create/delete files in 32bits machines.
This fixes it, and it works similarly to how we cache free space in block
cgroups.
We start a kernel thread to read the file tree. By scanning inode items,
we know which chunks of inode numbers are free, and we cache them in
an rb-tree.
Because we are searching the commit root, we have to carefully handle the
cross-transaction case.
The rb-tree is a hybrid extent+bitmap tree, so if we have too many small
chunks of inode numbers, we'll use bitmaps. Initially we allow 16K ram
of extents, and a bitmap will be used if we exceed this threshold. The
extents threshold is adjusted in runtime.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-04-20 10:06:11 +08:00
|
|
|
/*
|
|
|
|
* Find the left-most item in the cache tree, and then return the
|
|
|
|
* smallest inode number in the item.
|
|
|
|
*
|
|
|
|
* Note: the returned inode number may not be the smallest one in
|
|
|
|
* the tree, if the left-most item is a bitmap.
|
|
|
|
*/
|
|
|
|
u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root)
|
|
|
|
{
|
|
|
|
struct btrfs_free_space_ctl *ctl = fs_root->free_ino_ctl;
|
|
|
|
struct btrfs_free_space *entry = NULL;
|
|
|
|
u64 ino = 0;
|
|
|
|
|
|
|
|
spin_lock(&ctl->tree_lock);
|
|
|
|
|
|
|
|
if (RB_EMPTY_ROOT(&ctl->free_space_offset))
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
entry = rb_entry(rb_first(&ctl->free_space_offset),
|
|
|
|
struct btrfs_free_space, offset_index);
|
|
|
|
|
|
|
|
if (!entry->bitmap) {
|
|
|
|
ino = entry->offset;
|
|
|
|
|
|
|
|
unlink_free_space(ctl, entry);
|
|
|
|
entry->offset++;
|
|
|
|
entry->bytes--;
|
|
|
|
if (!entry->bytes)
|
|
|
|
kmem_cache_free(btrfs_free_space_cachep, entry);
|
|
|
|
else
|
|
|
|
link_free_space(ctl, entry);
|
|
|
|
} else {
|
|
|
|
u64 offset = 0;
|
|
|
|
u64 count = 1;
|
|
|
|
int ret;
|
|
|
|
|
2015-10-03 04:12:23 +08:00
|
|
|
ret = search_bitmap(ctl, entry, &offset, &count, true);
|
2012-03-12 23:03:00 +08:00
|
|
|
/* Logic error; Should be empty if it can't find anything */
|
2013-08-27 05:14:08 +08:00
|
|
|
ASSERT(!ret);
|
Btrfs: Cache free inode numbers in memory
Currently btrfs stores the highest objectid of the fs tree, and it always
returns (highest+1) inode number when we create a file, so inode numbers
won't be reclaimed when we delete files, so we'll run out of inode numbers
as we keep create/delete files in 32bits machines.
This fixes it, and it works similarly to how we cache free space in block
cgroups.
We start a kernel thread to read the file tree. By scanning inode items,
we know which chunks of inode numbers are free, and we cache them in
an rb-tree.
Because we are searching the commit root, we have to carefully handle the
cross-transaction case.
The rb-tree is a hybrid extent+bitmap tree, so if we have too many small
chunks of inode numbers, we'll use bitmaps. Initially we allow 16K ram
of extents, and a bitmap will be used if we exceed this threshold. The
extents threshold is adjusted in runtime.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
2011-04-20 10:06:11 +08:00
|
|
|
|
|
|
|
ino = offset;
|
|
|
|
bitmap_clear_bits(ctl, entry, offset, 1);
|
|
|
|
if (entry->bytes == 0)
|
|
|
|
free_bitmap(ctl, entry);
|
|
|
|
}
|
|
|
|
out:
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
|
|
|
|
|
|
|
return ino;
|
|
|
|
}
|
2011-04-20 10:33:24 +08:00
|
|
|
|
|
|
|
struct inode *lookup_free_ino_inode(struct btrfs_root *root,
|
|
|
|
struct btrfs_path *path)
|
|
|
|
{
|
|
|
|
struct inode *inode = NULL;
|
|
|
|
|
2014-02-05 09:37:48 +08:00
|
|
|
spin_lock(&root->ino_cache_lock);
|
|
|
|
if (root->ino_cache_inode)
|
|
|
|
inode = igrab(root->ino_cache_inode);
|
|
|
|
spin_unlock(&root->ino_cache_lock);
|
2011-04-20 10:33:24 +08:00
|
|
|
if (inode)
|
|
|
|
return inode;
|
|
|
|
|
|
|
|
inode = __lookup_free_space_inode(root, path, 0);
|
|
|
|
if (IS_ERR(inode))
|
|
|
|
return inode;
|
|
|
|
|
2014-02-05 09:37:48 +08:00
|
|
|
spin_lock(&root->ino_cache_lock);
|
2011-06-01 00:07:27 +08:00
|
|
|
if (!btrfs_fs_closing(root->fs_info))
|
2014-02-05 09:37:48 +08:00
|
|
|
root->ino_cache_inode = igrab(inode);
|
|
|
|
spin_unlock(&root->ino_cache_lock);
|
2011-04-20 10:33:24 +08:00
|
|
|
|
|
|
|
return inode;
|
|
|
|
}
|
|
|
|
|
|
|
|
int create_free_ino_inode(struct btrfs_root *root,
|
|
|
|
struct btrfs_trans_handle *trans,
|
|
|
|
struct btrfs_path *path)
|
|
|
|
{
|
|
|
|
return __create_free_space_inode(root, trans, path,
|
|
|
|
BTRFS_FREE_INO_OBJECTID, 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
int load_free_ino_cache(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
|
|
|
|
{
|
|
|
|
struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
|
|
|
|
struct btrfs_path *path;
|
|
|
|
struct inode *inode;
|
|
|
|
int ret = 0;
|
|
|
|
u64 root_gen = btrfs_root_generation(&root->root_item);
|
|
|
|
|
2011-06-03 21:36:29 +08:00
|
|
|
if (!btrfs_test_opt(root, INODE_MAP_CACHE))
|
|
|
|
return 0;
|
|
|
|
|
2011-04-20 10:33:24 +08:00
|
|
|
/*
|
|
|
|
* If we're unmounting then just return, since this does a search on the
|
|
|
|
* normal root and not the commit root and we could deadlock.
|
|
|
|
*/
|
2011-06-01 00:07:27 +08:00
|
|
|
if (btrfs_fs_closing(fs_info))
|
2011-04-20 10:33:24 +08:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
path = btrfs_alloc_path();
|
|
|
|
if (!path)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
inode = lookup_free_ino_inode(root, path);
|
|
|
|
if (IS_ERR(inode))
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
if (root_gen != BTRFS_I(inode)->generation)
|
|
|
|
goto out_put;
|
|
|
|
|
|
|
|
ret = __load_free_space_cache(root, inode, ctl, path, 0);
|
|
|
|
|
|
|
|
if (ret < 0)
|
2013-03-20 06:41:23 +08:00
|
|
|
btrfs_err(fs_info,
|
|
|
|
"failed to load free ino cache for root %llu",
|
|
|
|
root->root_key.objectid);
|
2011-04-20 10:33:24 +08:00
|
|
|
out_put:
|
|
|
|
iput(inode);
|
|
|
|
out:
|
|
|
|
btrfs_free_path(path);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
int btrfs_write_out_ino_cache(struct btrfs_root *root,
|
|
|
|
struct btrfs_trans_handle *trans,
|
2013-09-20 21:43:28 +08:00
|
|
|
struct btrfs_path *path,
|
|
|
|
struct inode *inode)
|
2011-04-20 10:33:24 +08:00
|
|
|
{
|
|
|
|
struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
|
|
|
|
int ret;
|
2015-04-05 08:14:42 +08:00
|
|
|
struct btrfs_io_ctl io_ctl;
|
2015-05-05 22:21:27 +08:00
|
|
|
bool release_metadata = true;
|
2011-04-20 10:33:24 +08:00
|
|
|
|
2011-06-03 21:36:29 +08:00
|
|
|
if (!btrfs_test_opt(root, INODE_MAP_CACHE))
|
|
|
|
return 0;
|
|
|
|
|
2015-04-23 23:02:49 +08:00
|
|
|
memset(&io_ctl, 0, sizeof(io_ctl));
|
2015-04-05 08:14:42 +08:00
|
|
|
ret = __btrfs_write_out_cache(root, inode, ctl, NULL, &io_ctl,
|
2015-04-23 23:02:49 +08:00
|
|
|
trans, path, 0);
|
2015-05-05 22:21:27 +08:00
|
|
|
if (!ret) {
|
|
|
|
/*
|
|
|
|
* At this point writepages() didn't error out, so our metadata
|
|
|
|
* reservation is released when the writeback finishes, at
|
|
|
|
* inode.c:btrfs_finish_ordered_io(), regardless of it finishing
|
|
|
|
* with or without an error.
|
|
|
|
*/
|
|
|
|
release_metadata = false;
|
2015-04-23 23:02:49 +08:00
|
|
|
ret = btrfs_wait_cache_io(root, trans, NULL, &io_ctl, path, 0);
|
2015-05-05 22:21:27 +08:00
|
|
|
}
|
2015-04-23 23:02:49 +08:00
|
|
|
|
2011-08-30 22:19:10 +08:00
|
|
|
if (ret) {
|
2015-05-05 22:21:27 +08:00
|
|
|
if (release_metadata)
|
|
|
|
btrfs_delalloc_release_metadata(inode, inode->i_size);
|
2011-08-30 22:19:10 +08:00
|
|
|
#ifdef DEBUG
|
2013-03-20 06:41:23 +08:00
|
|
|
btrfs_err(root->fs_info,
|
|
|
|
"failed to write free ino cache for root %llu",
|
|
|
|
root->root_key.objectid);
|
2011-08-30 22:19:10 +08:00
|
|
|
#endif
|
|
|
|
}
|
2011-04-20 10:33:24 +08:00
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
2013-03-15 21:47:08 +08:00
|
|
|
|
|
|
|
#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
|
2013-08-15 03:05:12 +08:00
|
|
|
/*
|
|
|
|
* Use this if you need to make a bitmap or extent entry specifically, it
|
|
|
|
* doesn't do any of the merging that add_free_space does, this acts a lot like
|
|
|
|
* how the free space cache loading stuff works, so you can get really weird
|
|
|
|
* configurations.
|
|
|
|
*/
|
|
|
|
int test_add_free_space_entry(struct btrfs_block_group_cache *cache,
|
|
|
|
u64 offset, u64 bytes, bool bitmap)
|
2013-03-15 21:47:08 +08:00
|
|
|
{
|
2013-08-15 03:05:12 +08:00
|
|
|
struct btrfs_free_space_ctl *ctl = cache->free_space_ctl;
|
|
|
|
struct btrfs_free_space *info = NULL, *bitmap_info;
|
|
|
|
void *map = NULL;
|
|
|
|
u64 bytes_added;
|
|
|
|
int ret;
|
2013-03-15 21:47:08 +08:00
|
|
|
|
2013-08-15 03:05:12 +08:00
|
|
|
again:
|
|
|
|
if (!info) {
|
|
|
|
info = kmem_cache_zalloc(btrfs_free_space_cachep, GFP_NOFS);
|
|
|
|
if (!info)
|
|
|
|
return -ENOMEM;
|
2013-03-15 21:47:08 +08:00
|
|
|
}
|
|
|
|
|
2013-08-15 03:05:12 +08:00
|
|
|
if (!bitmap) {
|
|
|
|
spin_lock(&ctl->tree_lock);
|
|
|
|
info->offset = offset;
|
|
|
|
info->bytes = bytes;
|
2015-10-03 04:09:42 +08:00
|
|
|
info->max_extent_size = 0;
|
2013-08-15 03:05:12 +08:00
|
|
|
ret = link_free_space(ctl, info);
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
|
|
|
if (ret)
|
|
|
|
kmem_cache_free(btrfs_free_space_cachep, info);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!map) {
|
|
|
|
map = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
|
|
|
|
if (!map) {
|
|
|
|
kmem_cache_free(btrfs_free_space_cachep, info);
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
spin_lock(&ctl->tree_lock);
|
|
|
|
bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
|
|
|
|
1, 0);
|
|
|
|
if (!bitmap_info) {
|
|
|
|
info->bitmap = map;
|
|
|
|
map = NULL;
|
|
|
|
add_new_bitmap(ctl, info, offset);
|
|
|
|
bitmap_info = info;
|
Btrfs: improve free space cache management and space allocation
While under random IO, a block group's free space cache eventually reaches
a state where it has a mix of extent entries and bitmap entries representing
free space regions.
As later free space regions are returned to the cache, some of them are merged
with existing extent entries if they are contiguous with them. But others are
not merged, because despite the existence of adjacent free space regions in
the cache, the merging doesn't happen because the existing free space regions
are represented in bitmap extents. Even when new free space regions are merged
with existing extent entries (enlarging the free space range they represent),
we create chances of having after an enlarged region that is contiguous with
some other region represented in a bitmap entry.
Both clustered and non-clustered space allocation work by iterating over our
extent and bitmap entries and skipping any that represents a region smaller
then the allocation request (and giving preference to extent entries before
bitmap entries). By having a contiguous free space region that is represented
by 2 (or more) entries (mix of extent and bitmap entries), we end up not
satisfying an allocation request with a size larger than the size of any of
the entries but no larger than the sum of their sizes. Making the caller assume
we're under a ENOSPC condition or force it to allocate multiple smaller space
regions (as we do for file data writes), which adds extra overhead and more
chances of causing fragmentation due to the smaller regions being all spread
apart from each other (more likely when under concurrency).
For example, if we have the following in the cache:
* extent entry representing free space range: [128Mb - 256Kb, 128Mb[
* bitmap entry covering the range [128Mb, 256Mb[, but only with the bits
representing the range [128Mb, 128Mb + 768Kb[ set - that is, only that
space in this 128Mb area is marked as free
An allocation request for 1Mb, starting at offset not greater than 128Mb - 256Kb,
would fail before, despite the existence of such contiguous free space area in the
cache. The caller could only allocate up to 768Kb of space at once and later another
256Kb (or vice-versa). In between each smaller allocation request, another task
working on a different file/inode might come in and take that space, preventing the
former task of getting a contiguous 1Mb region of free space.
Therefore this change implements the ability to move free space from bitmap
entries into existing and new free space regions represented with extent
entries. This is done when a space region is added to the cache.
A test was added to the sanity tests that explains in detail the issue too.
Some performance test results with compilebench on a 4 cores machine, with
32Gb of ram and using an HDD follow.
Test: compilebench -D /mnt -i 30 -r 1000 --makej
Before this change:
intial create total runs 30 avg 69.02 MB/s (user 0.28s sys 0.57s)
compile total runs 30 avg 314.96 MB/s (user 0.12s sys 0.25s)
read compiled tree total runs 3 avg 27.14 MB/s (user 1.52s sys 0.90s)
delete compiled tree total runs 30 avg 3.14 seconds (user 0.15s sys 0.66s)
After this change:
intial create total runs 30 avg 68.37 MB/s (user 0.29s sys 0.55s)
compile total runs 30 avg 382.83 MB/s (user 0.12s sys 0.24s)
read compiled tree total runs 3 avg 27.82 MB/s (user 1.45s sys 0.97s)
delete compiled tree total runs 30 avg 3.18 seconds (user 0.17s sys 0.65s)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-08-29 20:35:13 +08:00
|
|
|
info = NULL;
|
2013-08-15 03:05:12 +08:00
|
|
|
}
|
2013-03-15 21:47:08 +08:00
|
|
|
|
2013-08-15 03:05:12 +08:00
|
|
|
bytes_added = add_bytes_to_bitmap(ctl, bitmap_info, offset, bytes);
|
2015-10-03 04:09:42 +08:00
|
|
|
|
2013-08-15 03:05:12 +08:00
|
|
|
bytes -= bytes_added;
|
|
|
|
offset += bytes_added;
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
2013-03-15 21:47:08 +08:00
|
|
|
|
2013-08-15 03:05:12 +08:00
|
|
|
if (bytes)
|
|
|
|
goto again;
|
2013-03-15 21:47:08 +08:00
|
|
|
|
Btrfs: improve free space cache management and space allocation
While under random IO, a block group's free space cache eventually reaches
a state where it has a mix of extent entries and bitmap entries representing
free space regions.
As later free space regions are returned to the cache, some of them are merged
with existing extent entries if they are contiguous with them. But others are
not merged, because despite the existence of adjacent free space regions in
the cache, the merging doesn't happen because the existing free space regions
are represented in bitmap extents. Even when new free space regions are merged
with existing extent entries (enlarging the free space range they represent),
we create chances of having after an enlarged region that is contiguous with
some other region represented in a bitmap entry.
Both clustered and non-clustered space allocation work by iterating over our
extent and bitmap entries and skipping any that represents a region smaller
then the allocation request (and giving preference to extent entries before
bitmap entries). By having a contiguous free space region that is represented
by 2 (or more) entries (mix of extent and bitmap entries), we end up not
satisfying an allocation request with a size larger than the size of any of
the entries but no larger than the sum of their sizes. Making the caller assume
we're under a ENOSPC condition or force it to allocate multiple smaller space
regions (as we do for file data writes), which adds extra overhead and more
chances of causing fragmentation due to the smaller regions being all spread
apart from each other (more likely when under concurrency).
For example, if we have the following in the cache:
* extent entry representing free space range: [128Mb - 256Kb, 128Mb[
* bitmap entry covering the range [128Mb, 256Mb[, but only with the bits
representing the range [128Mb, 128Mb + 768Kb[ set - that is, only that
space in this 128Mb area is marked as free
An allocation request for 1Mb, starting at offset not greater than 128Mb - 256Kb,
would fail before, despite the existence of such contiguous free space area in the
cache. The caller could only allocate up to 768Kb of space at once and later another
256Kb (or vice-versa). In between each smaller allocation request, another task
working on a different file/inode might come in and take that space, preventing the
former task of getting a contiguous 1Mb region of free space.
Therefore this change implements the ability to move free space from bitmap
entries into existing and new free space regions represented with extent
entries. This is done when a space region is added to the cache.
A test was added to the sanity tests that explains in detail the issue too.
Some performance test results with compilebench on a 4 cores machine, with
32Gb of ram and using an HDD follow.
Test: compilebench -D /mnt -i 30 -r 1000 --makej
Before this change:
intial create total runs 30 avg 69.02 MB/s (user 0.28s sys 0.57s)
compile total runs 30 avg 314.96 MB/s (user 0.12s sys 0.25s)
read compiled tree total runs 3 avg 27.14 MB/s (user 1.52s sys 0.90s)
delete compiled tree total runs 30 avg 3.14 seconds (user 0.15s sys 0.66s)
After this change:
intial create total runs 30 avg 68.37 MB/s (user 0.29s sys 0.55s)
compile total runs 30 avg 382.83 MB/s (user 0.12s sys 0.24s)
read compiled tree total runs 3 avg 27.82 MB/s (user 1.45s sys 0.97s)
delete compiled tree total runs 30 avg 3.18 seconds (user 0.17s sys 0.65s)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-08-29 20:35:13 +08:00
|
|
|
if (info)
|
|
|
|
kmem_cache_free(btrfs_free_space_cachep, info);
|
2013-08-15 03:05:12 +08:00
|
|
|
if (map)
|
|
|
|
kfree(map);
|
|
|
|
return 0;
|
2013-03-15 21:47:08 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Checks to see if the given range is in the free space cache. This is really
|
|
|
|
* just used to check the absence of space, so if there is free space in the
|
|
|
|
* range at all we will return 1.
|
|
|
|
*/
|
2013-08-15 03:05:12 +08:00
|
|
|
int test_check_exists(struct btrfs_block_group_cache *cache,
|
|
|
|
u64 offset, u64 bytes)
|
2013-03-15 21:47:08 +08:00
|
|
|
{
|
|
|
|
struct btrfs_free_space_ctl *ctl = cache->free_space_ctl;
|
|
|
|
struct btrfs_free_space *info;
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
spin_lock(&ctl->tree_lock);
|
|
|
|
info = tree_search_offset(ctl, offset, 0, 0);
|
|
|
|
if (!info) {
|
|
|
|
info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
|
|
|
|
1, 0);
|
|
|
|
if (!info)
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
have_info:
|
|
|
|
if (info->bitmap) {
|
|
|
|
u64 bit_off, bit_bytes;
|
|
|
|
struct rb_node *n;
|
|
|
|
struct btrfs_free_space *tmp;
|
|
|
|
|
|
|
|
bit_off = offset;
|
|
|
|
bit_bytes = ctl->unit;
|
2015-10-03 04:12:23 +08:00
|
|
|
ret = search_bitmap(ctl, info, &bit_off, &bit_bytes, false);
|
2013-03-15 21:47:08 +08:00
|
|
|
if (!ret) {
|
|
|
|
if (bit_off == offset) {
|
|
|
|
ret = 1;
|
|
|
|
goto out;
|
|
|
|
} else if (bit_off > offset &&
|
|
|
|
offset + bytes > bit_off) {
|
|
|
|
ret = 1;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
n = rb_prev(&info->offset_index);
|
|
|
|
while (n) {
|
|
|
|
tmp = rb_entry(n, struct btrfs_free_space,
|
|
|
|
offset_index);
|
|
|
|
if (tmp->offset + tmp->bytes < offset)
|
|
|
|
break;
|
|
|
|
if (offset + bytes < tmp->offset) {
|
|
|
|
n = rb_prev(&info->offset_index);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
info = tmp;
|
|
|
|
goto have_info;
|
|
|
|
}
|
|
|
|
|
|
|
|
n = rb_next(&info->offset_index);
|
|
|
|
while (n) {
|
|
|
|
tmp = rb_entry(n, struct btrfs_free_space,
|
|
|
|
offset_index);
|
|
|
|
if (offset + bytes < tmp->offset)
|
|
|
|
break;
|
|
|
|
if (tmp->offset + tmp->bytes < offset) {
|
|
|
|
n = rb_next(&info->offset_index);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
info = tmp;
|
|
|
|
goto have_info;
|
|
|
|
}
|
|
|
|
|
Btrfs: improve free space cache management and space allocation
While under random IO, a block group's free space cache eventually reaches
a state where it has a mix of extent entries and bitmap entries representing
free space regions.
As later free space regions are returned to the cache, some of them are merged
with existing extent entries if they are contiguous with them. But others are
not merged, because despite the existence of adjacent free space regions in
the cache, the merging doesn't happen because the existing free space regions
are represented in bitmap extents. Even when new free space regions are merged
with existing extent entries (enlarging the free space range they represent),
we create chances of having after an enlarged region that is contiguous with
some other region represented in a bitmap entry.
Both clustered and non-clustered space allocation work by iterating over our
extent and bitmap entries and skipping any that represents a region smaller
then the allocation request (and giving preference to extent entries before
bitmap entries). By having a contiguous free space region that is represented
by 2 (or more) entries (mix of extent and bitmap entries), we end up not
satisfying an allocation request with a size larger than the size of any of
the entries but no larger than the sum of their sizes. Making the caller assume
we're under a ENOSPC condition or force it to allocate multiple smaller space
regions (as we do for file data writes), which adds extra overhead and more
chances of causing fragmentation due to the smaller regions being all spread
apart from each other (more likely when under concurrency).
For example, if we have the following in the cache:
* extent entry representing free space range: [128Mb - 256Kb, 128Mb[
* bitmap entry covering the range [128Mb, 256Mb[, but only with the bits
representing the range [128Mb, 128Mb + 768Kb[ set - that is, only that
space in this 128Mb area is marked as free
An allocation request for 1Mb, starting at offset not greater than 128Mb - 256Kb,
would fail before, despite the existence of such contiguous free space area in the
cache. The caller could only allocate up to 768Kb of space at once and later another
256Kb (or vice-versa). In between each smaller allocation request, another task
working on a different file/inode might come in and take that space, preventing the
former task of getting a contiguous 1Mb region of free space.
Therefore this change implements the ability to move free space from bitmap
entries into existing and new free space regions represented with extent
entries. This is done when a space region is added to the cache.
A test was added to the sanity tests that explains in detail the issue too.
Some performance test results with compilebench on a 4 cores machine, with
32Gb of ram and using an HDD follow.
Test: compilebench -D /mnt -i 30 -r 1000 --makej
Before this change:
intial create total runs 30 avg 69.02 MB/s (user 0.28s sys 0.57s)
compile total runs 30 avg 314.96 MB/s (user 0.12s sys 0.25s)
read compiled tree total runs 3 avg 27.14 MB/s (user 1.52s sys 0.90s)
delete compiled tree total runs 30 avg 3.14 seconds (user 0.15s sys 0.66s)
After this change:
intial create total runs 30 avg 68.37 MB/s (user 0.29s sys 0.55s)
compile total runs 30 avg 382.83 MB/s (user 0.12s sys 0.24s)
read compiled tree total runs 3 avg 27.82 MB/s (user 1.45s sys 0.97s)
delete compiled tree total runs 30 avg 3.18 seconds (user 0.17s sys 0.65s)
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
2014-08-29 20:35:13 +08:00
|
|
|
ret = 0;
|
2013-03-15 21:47:08 +08:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (info->offset == offset) {
|
|
|
|
ret = 1;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (offset > info->offset && offset < info->offset + info->bytes)
|
|
|
|
ret = 1;
|
|
|
|
out:
|
|
|
|
spin_unlock(&ctl->tree_lock);
|
|
|
|
return ret;
|
|
|
|
}
|
2013-08-15 03:05:12 +08:00
|
|
|
#endif /* CONFIG_BTRFS_FS_RUN_SANITY_TESTS */
|