2007-06-12 21:07:21 +08:00
|
|
|
/*
|
|
|
|
|
* Copyright (C) 2007 Oracle. 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.
|
|
|
|
|
*/
|
|
|
|
|
|
2007-02-24 19:24:44 +08:00
|
|
|
#include "ctree.h"
|
|
|
|
|
#include "disk-io.h"
|
2007-05-03 03:53:43 +08:00
|
|
|
#include "print-tree.h"
|
2007-02-24 19:24:44 +08:00
|
|
|
|
2008-03-25 03:01:56 +08:00
|
|
|
static void print_chunk(struct extent_buffer *eb, struct btrfs_chunk *chunk)
|
|
|
|
|
{
|
|
|
|
|
int num_stripes = btrfs_chunk_num_stripes(eb, chunk);
|
|
|
|
|
int i;
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_INFO "\t\tchunk length %llu owner %llu type %llu "
|
|
|
|
|
"num_stripes %d\n",
|
2008-04-16 03:41:47 +08:00
|
|
|
(unsigned long long)btrfs_chunk_length(eb, chunk),
|
2008-03-25 03:01:56 +08:00
|
|
|
(unsigned long long)btrfs_chunk_owner(eb, chunk),
|
|
|
|
|
(unsigned long long)btrfs_chunk_type(eb, chunk),
|
|
|
|
|
num_stripes);
|
|
|
|
|
for (i = 0 ; i < num_stripes ; i++) {
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_INFO "\t\t\tstripe %d devid %llu offset %llu\n", i,
|
2008-03-25 03:01:56 +08:00
|
|
|
(unsigned long long)btrfs_stripe_devid_nr(eb, chunk, i),
|
|
|
|
|
(unsigned long long)btrfs_stripe_offset_nr(eb, chunk, i));
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
static void print_dev_item(struct extent_buffer *eb,
|
|
|
|
|
struct btrfs_dev_item *dev_item)
|
|
|
|
|
{
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_INFO "\t\tdev item devid %llu "
|
|
|
|
|
"total_bytes %llu bytes used %llu\n",
|
2008-03-25 03:01:56 +08:00
|
|
|
(unsigned long long)btrfs_device_id(eb, dev_item),
|
|
|
|
|
(unsigned long long)btrfs_device_total_bytes(eb, dev_item),
|
|
|
|
|
(unsigned long long)btrfs_device_bytes_used(eb, dev_item));
|
|
|
|
|
}
|
Btrfs: Mixed back reference (FORWARD ROLLING FORMAT CHANGE)
This commit introduces a new kind of back reference for btrfs metadata.
Once a filesystem has been mounted with this commit, IT WILL NO LONGER
BE MOUNTABLE BY OLDER KERNELS.
When a tree block in subvolume tree is cow'd, the reference counts of all
extents it points to are increased by one. At transaction commit time,
the old root of the subvolume is recorded in a "dead root" data structure,
and the btree it points to is later walked, dropping reference counts
and freeing any blocks where the reference count goes to 0.
The increments done during cow and decrements done after commit cancel out,
and the walk is a very expensive way to go about freeing the blocks that
are no longer referenced by the new btree root. This commit reduces the
transaction overhead by avoiding the need for dead root records.
When a non-shared tree block is cow'd, we free the old block at once, and the
new block inherits old block's references. When a tree block with reference
count > 1 is cow'd, we increase the reference counts of all extents
the new block points to by one, and decrease the old block's reference count by
one.
This dead tree avoidance code removes the need to modify the reference
counts of lower level extents when a non-shared tree block is cow'd.
But we still need to update back ref for all pointers in the block.
This is because the location of the block is recorded in the back ref
item.
We can solve this by introducing a new type of back ref. The new
back ref provides information about pointer's key, level and in which
tree the pointer lives. This information allow us to find the pointer
by searching the tree. The shortcoming of the new back ref is that it
only works for pointers in tree blocks referenced by their owner trees.
This is mostly a problem for snapshots, where resolving one of these
fuzzy back references would be O(number_of_snapshots) and quite slow.
The solution used here is to use the fuzzy back references in the common
case where a given tree block is only referenced by one root,
and use the full back references when multiple roots have a reference
on a given block.
This commit adds per subvolume red-black tree to keep trace of cached
inodes. The red-black tree helps the balancing code to find cached
inodes whose inode numbers within a given range.
This commit improves the balancing code by introducing several data
structures to keep the state of balancing. The most important one
is the back ref cache. It caches how the upper level tree blocks are
referenced. This greatly reduce the overhead of checking back ref.
The improved balancing code scales significantly better with a large
number of snapshots.
This is a very large commit and was written in a number of
pieces. But, they depend heavily on the disk format change and were
squashed together to make sure git bisect didn't end up in a
bad state wrt space balancing or the format change.
Signed-off-by: Yan Zheng <zheng.yan@oracle.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-06-10 22:45:14 +08:00
|
|
|
static void print_extent_data_ref(struct extent_buffer *eb,
|
|
|
|
|
struct btrfs_extent_data_ref *ref)
|
|
|
|
|
{
|
|
|
|
|
printk(KERN_INFO "\t\textent data backref root %llu "
|
|
|
|
|
"objectid %llu offset %llu count %u\n",
|
|
|
|
|
(unsigned long long)btrfs_extent_data_ref_root(eb, ref),
|
|
|
|
|
(unsigned long long)btrfs_extent_data_ref_objectid(eb, ref),
|
|
|
|
|
(unsigned long long)btrfs_extent_data_ref_offset(eb, ref),
|
|
|
|
|
btrfs_extent_data_ref_count(eb, ref));
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void print_extent_item(struct extent_buffer *eb, int slot)
|
|
|
|
|
{
|
|
|
|
|
struct btrfs_extent_item *ei;
|
|
|
|
|
struct btrfs_extent_inline_ref *iref;
|
|
|
|
|
struct btrfs_extent_data_ref *dref;
|
|
|
|
|
struct btrfs_shared_data_ref *sref;
|
|
|
|
|
struct btrfs_disk_key key;
|
|
|
|
|
unsigned long end;
|
|
|
|
|
unsigned long ptr;
|
|
|
|
|
int type;
|
|
|
|
|
u32 item_size = btrfs_item_size_nr(eb, slot);
|
|
|
|
|
u64 flags;
|
|
|
|
|
u64 offset;
|
|
|
|
|
|
|
|
|
|
if (item_size < sizeof(*ei)) {
|
|
|
|
|
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
|
|
|
|
|
struct btrfs_extent_item_v0 *ei0;
|
|
|
|
|
BUG_ON(item_size != sizeof(*ei0));
|
|
|
|
|
ei0 = btrfs_item_ptr(eb, slot, struct btrfs_extent_item_v0);
|
|
|
|
|
printk(KERN_INFO "\t\textent refs %u\n",
|
|
|
|
|
btrfs_extent_refs_v0(eb, ei0));
|
|
|
|
|
return;
|
|
|
|
|
#else
|
|
|
|
|
BUG();
|
|
|
|
|
#endif
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
ei = btrfs_item_ptr(eb, slot, struct btrfs_extent_item);
|
|
|
|
|
flags = btrfs_extent_flags(eb, ei);
|
|
|
|
|
|
|
|
|
|
printk(KERN_INFO "\t\textent refs %llu gen %llu flags %llu\n",
|
|
|
|
|
(unsigned long long)btrfs_extent_refs(eb, ei),
|
|
|
|
|
(unsigned long long)btrfs_extent_generation(eb, ei),
|
|
|
|
|
(unsigned long long)flags);
|
|
|
|
|
|
|
|
|
|
if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
|
|
|
|
|
struct btrfs_tree_block_info *info;
|
|
|
|
|
info = (struct btrfs_tree_block_info *)(ei + 1);
|
|
|
|
|
btrfs_tree_block_key(eb, info, &key);
|
|
|
|
|
printk(KERN_INFO "\t\ttree block key (%llu %x %llu) "
|
|
|
|
|
"level %d\n",
|
|
|
|
|
(unsigned long long)btrfs_disk_key_objectid(&key),
|
|
|
|
|
key.type,
|
|
|
|
|
(unsigned long long)btrfs_disk_key_offset(&key),
|
|
|
|
|
btrfs_tree_block_level(eb, info));
|
|
|
|
|
iref = (struct btrfs_extent_inline_ref *)(info + 1);
|
|
|
|
|
} else {
|
|
|
|
|
iref = (struct btrfs_extent_inline_ref *)(ei + 1);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
ptr = (unsigned long)iref;
|
|
|
|
|
end = (unsigned long)ei + item_size;
|
|
|
|
|
while (ptr < end) {
|
|
|
|
|
iref = (struct btrfs_extent_inline_ref *)ptr;
|
|
|
|
|
type = btrfs_extent_inline_ref_type(eb, iref);
|
|
|
|
|
offset = btrfs_extent_inline_ref_offset(eb, iref);
|
|
|
|
|
switch (type) {
|
|
|
|
|
case BTRFS_TREE_BLOCK_REF_KEY:
|
|
|
|
|
printk(KERN_INFO "\t\ttree block backref "
|
|
|
|
|
"root %llu\n", (unsigned long long)offset);
|
|
|
|
|
break;
|
|
|
|
|
case BTRFS_SHARED_BLOCK_REF_KEY:
|
|
|
|
|
printk(KERN_INFO "\t\tshared block backref "
|
|
|
|
|
"parent %llu\n", (unsigned long long)offset);
|
|
|
|
|
break;
|
|
|
|
|
case BTRFS_EXTENT_DATA_REF_KEY:
|
|
|
|
|
dref = (struct btrfs_extent_data_ref *)(&iref->offset);
|
|
|
|
|
print_extent_data_ref(eb, dref);
|
|
|
|
|
break;
|
|
|
|
|
case BTRFS_SHARED_DATA_REF_KEY:
|
|
|
|
|
sref = (struct btrfs_shared_data_ref *)(iref + 1);
|
|
|
|
|
printk(KERN_INFO "\t\tshared data backref "
|
|
|
|
|
"parent %llu count %u\n",
|
|
|
|
|
(unsigned long long)offset,
|
|
|
|
|
btrfs_shared_data_ref_count(eb, sref));
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
BUG();
|
|
|
|
|
}
|
|
|
|
|
ptr += btrfs_extent_inline_ref_size(type);
|
|
|
|
|
}
|
|
|
|
|
WARN_ON(ptr > end);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
|
|
|
|
|
static void print_extent_ref_v0(struct extent_buffer *eb, int slot)
|
|
|
|
|
{
|
|
|
|
|
struct btrfs_extent_ref_v0 *ref0;
|
|
|
|
|
|
|
|
|
|
ref0 = btrfs_item_ptr(eb, slot, struct btrfs_extent_ref_v0);
|
|
|
|
|
printk("\t\textent back ref root %llu gen %llu "
|
|
|
|
|
"owner %llu num_refs %lu\n",
|
|
|
|
|
(unsigned long long)btrfs_ref_root_v0(eb, ref0),
|
|
|
|
|
(unsigned long long)btrfs_ref_generation_v0(eb, ref0),
|
|
|
|
|
(unsigned long long)btrfs_ref_objectid_v0(eb, ref0),
|
|
|
|
|
(unsigned long)btrfs_ref_count_v0(eb, ref0));
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
2007-10-16 04:14:19 +08:00
|
|
|
void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l)
|
2007-02-24 19:24:44 +08:00
|
|
|
{
|
|
|
|
|
int i;
|
Btrfs: Mixed back reference (FORWARD ROLLING FORMAT CHANGE)
This commit introduces a new kind of back reference for btrfs metadata.
Once a filesystem has been mounted with this commit, IT WILL NO LONGER
BE MOUNTABLE BY OLDER KERNELS.
When a tree block in subvolume tree is cow'd, the reference counts of all
extents it points to are increased by one. At transaction commit time,
the old root of the subvolume is recorded in a "dead root" data structure,
and the btree it points to is later walked, dropping reference counts
and freeing any blocks where the reference count goes to 0.
The increments done during cow and decrements done after commit cancel out,
and the walk is a very expensive way to go about freeing the blocks that
are no longer referenced by the new btree root. This commit reduces the
transaction overhead by avoiding the need for dead root records.
When a non-shared tree block is cow'd, we free the old block at once, and the
new block inherits old block's references. When a tree block with reference
count > 1 is cow'd, we increase the reference counts of all extents
the new block points to by one, and decrease the old block's reference count by
one.
This dead tree avoidance code removes the need to modify the reference
counts of lower level extents when a non-shared tree block is cow'd.
But we still need to update back ref for all pointers in the block.
This is because the location of the block is recorded in the back ref
item.
We can solve this by introducing a new type of back ref. The new
back ref provides information about pointer's key, level and in which
tree the pointer lives. This information allow us to find the pointer
by searching the tree. The shortcoming of the new back ref is that it
only works for pointers in tree blocks referenced by their owner trees.
This is mostly a problem for snapshots, where resolving one of these
fuzzy back references would be O(number_of_snapshots) and quite slow.
The solution used here is to use the fuzzy back references in the common
case where a given tree block is only referenced by one root,
and use the full back references when multiple roots have a reference
on a given block.
This commit adds per subvolume red-black tree to keep trace of cached
inodes. The red-black tree helps the balancing code to find cached
inodes whose inode numbers within a given range.
This commit improves the balancing code by introducing several data
structures to keep the state of balancing. The most important one
is the back ref cache. It caches how the upper level tree blocks are
referenced. This greatly reduce the overhead of checking back ref.
The improved balancing code scales significantly better with a large
number of snapshots.
This is a very large commit and was written in a number of
pieces. But, they depend heavily on the disk format change and were
squashed together to make sure git bisect didn't end up in a
bad state wrt space balancing or the format change.
Signed-off-by: Yan Zheng <zheng.yan@oracle.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-06-10 22:45:14 +08:00
|
|
|
u32 type;
|
2007-10-16 04:14:19 +08:00
|
|
|
u32 nr = btrfs_header_nritems(l);
|
2007-03-13 08:12:07 +08:00
|
|
|
struct btrfs_item *item;
|
2007-03-14 04:47:54 +08:00
|
|
|
struct btrfs_root_item *ri;
|
2007-03-16 03:18:43 +08:00
|
|
|
struct btrfs_dir_item *di;
|
2007-03-21 03:57:25 +08:00
|
|
|
struct btrfs_inode_item *ii;
|
2007-04-27 04:46:15 +08:00
|
|
|
struct btrfs_block_group_item *bi;
|
2007-06-18 21:57:58 +08:00
|
|
|
struct btrfs_file_extent_item *fi;
|
Btrfs: Mixed back reference (FORWARD ROLLING FORMAT CHANGE)
This commit introduces a new kind of back reference for btrfs metadata.
Once a filesystem has been mounted with this commit, IT WILL NO LONGER
BE MOUNTABLE BY OLDER KERNELS.
When a tree block in subvolume tree is cow'd, the reference counts of all
extents it points to are increased by one. At transaction commit time,
the old root of the subvolume is recorded in a "dead root" data structure,
and the btree it points to is later walked, dropping reference counts
and freeing any blocks where the reference count goes to 0.
The increments done during cow and decrements done after commit cancel out,
and the walk is a very expensive way to go about freeing the blocks that
are no longer referenced by the new btree root. This commit reduces the
transaction overhead by avoiding the need for dead root records.
When a non-shared tree block is cow'd, we free the old block at once, and the
new block inherits old block's references. When a tree block with reference
count > 1 is cow'd, we increase the reference counts of all extents
the new block points to by one, and decrease the old block's reference count by
one.
This dead tree avoidance code removes the need to modify the reference
counts of lower level extents when a non-shared tree block is cow'd.
But we still need to update back ref for all pointers in the block.
This is because the location of the block is recorded in the back ref
item.
We can solve this by introducing a new type of back ref. The new
back ref provides information about pointer's key, level and in which
tree the pointer lives. This information allow us to find the pointer
by searching the tree. The shortcoming of the new back ref is that it
only works for pointers in tree blocks referenced by their owner trees.
This is mostly a problem for snapshots, where resolving one of these
fuzzy back references would be O(number_of_snapshots) and quite slow.
The solution used here is to use the fuzzy back references in the common
case where a given tree block is only referenced by one root,
and use the full back references when multiple roots have a reference
on a given block.
This commit adds per subvolume red-black tree to keep trace of cached
inodes. The red-black tree helps the balancing code to find cached
inodes whose inode numbers within a given range.
This commit improves the balancing code by introducing several data
structures to keep the state of balancing. The most important one
is the back ref cache. It caches how the upper level tree blocks are
referenced. This greatly reduce the overhead of checking back ref.
The improved balancing code scales significantly better with a large
number of snapshots.
This is a very large commit and was written in a number of
pieces. But, they depend heavily on the disk format change and were
squashed together to make sure git bisect didn't end up in a
bad state wrt space balancing or the format change.
Signed-off-by: Yan Zheng <zheng.yan@oracle.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-06-10 22:45:14 +08:00
|
|
|
struct btrfs_extent_data_ref *dref;
|
|
|
|
|
struct btrfs_shared_data_ref *sref;
|
|
|
|
|
struct btrfs_dev_extent *dev_extent;
|
2007-10-16 04:14:19 +08:00
|
|
|
struct btrfs_key key;
|
|
|
|
|
struct btrfs_key found_key;
|
2007-03-16 03:18:43 +08:00
|
|
|
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_INFO "leaf %llu total ptrs %d free space %d\n",
|
2007-10-16 04:15:53 +08:00
|
|
|
(unsigned long long)btrfs_header_bytenr(l), nr,
|
2007-03-15 02:14:43 +08:00
|
|
|
btrfs_leaf_free_space(root, l));
|
2007-02-24 19:24:44 +08:00
|
|
|
for (i = 0 ; i < nr ; i++) {
|
2007-10-16 04:14:19 +08:00
|
|
|
item = btrfs_item_nr(l, i);
|
|
|
|
|
btrfs_item_key_to_cpu(l, &key, i);
|
|
|
|
|
type = btrfs_key_type(&key);
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_INFO "\titem %d key (%llu %x %llu) itemoff %d "
|
|
|
|
|
"itemsize %d\n",
|
2007-02-24 19:24:44 +08:00
|
|
|
i,
|
2007-10-16 04:14:19 +08:00
|
|
|
(unsigned long long)key.objectid, type,
|
|
|
|
|
(unsigned long long)key.offset,
|
|
|
|
|
btrfs_item_offset(l, item), btrfs_item_size(l, item));
|
2007-03-16 00:56:47 +08:00
|
|
|
switch (type) {
|
|
|
|
|
case BTRFS_INODE_ITEM_KEY:
|
2007-03-21 03:57:25 +08:00
|
|
|
ii = btrfs_item_ptr(l, i, struct btrfs_inode_item);
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_INFO "\t\tinode generation %llu size %llu "
|
|
|
|
|
"mode %o\n",
|
|
|
|
|
(unsigned long long)
|
|
|
|
|
btrfs_inode_generation(l, ii),
|
2007-10-16 04:14:19 +08:00
|
|
|
(unsigned long long)btrfs_inode_size(l, ii),
|
|
|
|
|
btrfs_inode_mode(l, ii));
|
2007-03-16 00:56:47 +08:00
|
|
|
break;
|
|
|
|
|
case BTRFS_DIR_ITEM_KEY:
|
2007-03-16 03:18:43 +08:00
|
|
|
di = btrfs_item_ptr(l, i, struct btrfs_dir_item);
|
2007-10-16 04:14:19 +08:00
|
|
|
btrfs_dir_item_key_to_cpu(l, di, &found_key);
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_INFO "\t\tdir oid %llu type %u\n",
|
2007-10-16 04:14:19 +08:00
|
|
|
(unsigned long long)found_key.objectid,
|
|
|
|
|
btrfs_dir_type(l, di));
|
2007-03-16 00:56:47 +08:00
|
|
|
break;
|
|
|
|
|
case BTRFS_ROOT_ITEM_KEY:
|
|
|
|
|
ri = btrfs_item_ptr(l, i, struct btrfs_root_item);
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_INFO "\t\troot data bytenr %llu refs %u\n",
|
|
|
|
|
(unsigned long long)
|
|
|
|
|
btrfs_disk_root_bytenr(l, ri),
|
2007-10-16 04:14:19 +08:00
|
|
|
btrfs_disk_root_refs(l, ri));
|
2007-03-16 00:56:47 +08:00
|
|
|
break;
|
|
|
|
|
case BTRFS_EXTENT_ITEM_KEY:
|
Btrfs: Mixed back reference (FORWARD ROLLING FORMAT CHANGE)
This commit introduces a new kind of back reference for btrfs metadata.
Once a filesystem has been mounted with this commit, IT WILL NO LONGER
BE MOUNTABLE BY OLDER KERNELS.
When a tree block in subvolume tree is cow'd, the reference counts of all
extents it points to are increased by one. At transaction commit time,
the old root of the subvolume is recorded in a "dead root" data structure,
and the btree it points to is later walked, dropping reference counts
and freeing any blocks where the reference count goes to 0.
The increments done during cow and decrements done after commit cancel out,
and the walk is a very expensive way to go about freeing the blocks that
are no longer referenced by the new btree root. This commit reduces the
transaction overhead by avoiding the need for dead root records.
When a non-shared tree block is cow'd, we free the old block at once, and the
new block inherits old block's references. When a tree block with reference
count > 1 is cow'd, we increase the reference counts of all extents
the new block points to by one, and decrease the old block's reference count by
one.
This dead tree avoidance code removes the need to modify the reference
counts of lower level extents when a non-shared tree block is cow'd.
But we still need to update back ref for all pointers in the block.
This is because the location of the block is recorded in the back ref
item.
We can solve this by introducing a new type of back ref. The new
back ref provides information about pointer's key, level and in which
tree the pointer lives. This information allow us to find the pointer
by searching the tree. The shortcoming of the new back ref is that it
only works for pointers in tree blocks referenced by their owner trees.
This is mostly a problem for snapshots, where resolving one of these
fuzzy back references would be O(number_of_snapshots) and quite slow.
The solution used here is to use the fuzzy back references in the common
case where a given tree block is only referenced by one root,
and use the full back references when multiple roots have a reference
on a given block.
This commit adds per subvolume red-black tree to keep trace of cached
inodes. The red-black tree helps the balancing code to find cached
inodes whose inode numbers within a given range.
This commit improves the balancing code by introducing several data
structures to keep the state of balancing. The most important one
is the back ref cache. It caches how the upper level tree blocks are
referenced. This greatly reduce the overhead of checking back ref.
The improved balancing code scales significantly better with a large
number of snapshots.
This is a very large commit and was written in a number of
pieces. But, they depend heavily on the disk format change and were
squashed together to make sure git bisect didn't end up in a
bad state wrt space balancing or the format change.
Signed-off-by: Yan Zheng <zheng.yan@oracle.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-06-10 22:45:14 +08:00
|
|
|
print_extent_item(l, i);
|
2007-03-16 00:56:47 +08:00
|
|
|
break;
|
Btrfs: Mixed back reference (FORWARD ROLLING FORMAT CHANGE)
This commit introduces a new kind of back reference for btrfs metadata.
Once a filesystem has been mounted with this commit, IT WILL NO LONGER
BE MOUNTABLE BY OLDER KERNELS.
When a tree block in subvolume tree is cow'd, the reference counts of all
extents it points to are increased by one. At transaction commit time,
the old root of the subvolume is recorded in a "dead root" data structure,
and the btree it points to is later walked, dropping reference counts
and freeing any blocks where the reference count goes to 0.
The increments done during cow and decrements done after commit cancel out,
and the walk is a very expensive way to go about freeing the blocks that
are no longer referenced by the new btree root. This commit reduces the
transaction overhead by avoiding the need for dead root records.
When a non-shared tree block is cow'd, we free the old block at once, and the
new block inherits old block's references. When a tree block with reference
count > 1 is cow'd, we increase the reference counts of all extents
the new block points to by one, and decrease the old block's reference count by
one.
This dead tree avoidance code removes the need to modify the reference
counts of lower level extents when a non-shared tree block is cow'd.
But we still need to update back ref for all pointers in the block.
This is because the location of the block is recorded in the back ref
item.
We can solve this by introducing a new type of back ref. The new
back ref provides information about pointer's key, level and in which
tree the pointer lives. This information allow us to find the pointer
by searching the tree. The shortcoming of the new back ref is that it
only works for pointers in tree blocks referenced by their owner trees.
This is mostly a problem for snapshots, where resolving one of these
fuzzy back references would be O(number_of_snapshots) and quite slow.
The solution used here is to use the fuzzy back references in the common
case where a given tree block is only referenced by one root,
and use the full back references when multiple roots have a reference
on a given block.
This commit adds per subvolume red-black tree to keep trace of cached
inodes. The red-black tree helps the balancing code to find cached
inodes whose inode numbers within a given range.
This commit improves the balancing code by introducing several data
structures to keep the state of balancing. The most important one
is the back ref cache. It caches how the upper level tree blocks are
referenced. This greatly reduce the overhead of checking back ref.
The improved balancing code scales significantly better with a large
number of snapshots.
This is a very large commit and was written in a number of
pieces. But, they depend heavily on the disk format change and were
squashed together to make sure git bisect didn't end up in a
bad state wrt space balancing or the format change.
Signed-off-by: Yan Zheng <zheng.yan@oracle.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-06-10 22:45:14 +08:00
|
|
|
case BTRFS_TREE_BLOCK_REF_KEY:
|
|
|
|
|
printk(KERN_INFO "\t\ttree block backref\n");
|
|
|
|
|
break;
|
|
|
|
|
case BTRFS_SHARED_BLOCK_REF_KEY:
|
|
|
|
|
printk(KERN_INFO "\t\tshared block backref\n");
|
|
|
|
|
break;
|
|
|
|
|
case BTRFS_EXTENT_DATA_REF_KEY:
|
|
|
|
|
dref = btrfs_item_ptr(l, i,
|
|
|
|
|
struct btrfs_extent_data_ref);
|
|
|
|
|
print_extent_data_ref(l, dref);
|
|
|
|
|
break;
|
|
|
|
|
case BTRFS_SHARED_DATA_REF_KEY:
|
|
|
|
|
sref = btrfs_item_ptr(l, i,
|
|
|
|
|
struct btrfs_shared_data_ref);
|
|
|
|
|
printk(KERN_INFO "\t\tshared data backref count %u\n",
|
|
|
|
|
btrfs_shared_data_ref_count(l, sref));
|
2007-12-11 22:25:06 +08:00
|
|
|
break;
|
2007-06-18 21:57:58 +08:00
|
|
|
case BTRFS_EXTENT_DATA_KEY:
|
|
|
|
|
fi = btrfs_item_ptr(l, i,
|
|
|
|
|
struct btrfs_file_extent_item);
|
2007-10-16 04:14:19 +08:00
|
|
|
if (btrfs_file_extent_type(l, fi) ==
|
2007-06-18 21:57:58 +08:00
|
|
|
BTRFS_FILE_EXTENT_INLINE) {
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_INFO "\t\tinline extent data "
|
|
|
|
|
"size %u\n",
|
|
|
|
|
btrfs_file_extent_inline_len(l, fi));
|
2007-06-18 21:57:58 +08:00
|
|
|
break;
|
|
|
|
|
}
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_INFO "\t\textent data disk bytenr %llu "
|
|
|
|
|
"nr %llu\n",
|
|
|
|
|
(unsigned long long)
|
|
|
|
|
btrfs_file_extent_disk_bytenr(l, fi),
|
|
|
|
|
(unsigned long long)
|
|
|
|
|
btrfs_file_extent_disk_num_bytes(l, fi));
|
|
|
|
|
printk(KERN_INFO "\t\textent data offset %llu "
|
|
|
|
|
"nr %llu ram %llu\n",
|
|
|
|
|
(unsigned long long)
|
|
|
|
|
btrfs_file_extent_offset(l, fi),
|
|
|
|
|
(unsigned long long)
|
|
|
|
|
btrfs_file_extent_num_bytes(l, fi),
|
|
|
|
|
(unsigned long long)
|
|
|
|
|
btrfs_file_extent_ram_bytes(l, fi));
|
2007-06-18 21:57:58 +08:00
|
|
|
break;
|
Btrfs: Mixed back reference (FORWARD ROLLING FORMAT CHANGE)
This commit introduces a new kind of back reference for btrfs metadata.
Once a filesystem has been mounted with this commit, IT WILL NO LONGER
BE MOUNTABLE BY OLDER KERNELS.
When a tree block in subvolume tree is cow'd, the reference counts of all
extents it points to are increased by one. At transaction commit time,
the old root of the subvolume is recorded in a "dead root" data structure,
and the btree it points to is later walked, dropping reference counts
and freeing any blocks where the reference count goes to 0.
The increments done during cow and decrements done after commit cancel out,
and the walk is a very expensive way to go about freeing the blocks that
are no longer referenced by the new btree root. This commit reduces the
transaction overhead by avoiding the need for dead root records.
When a non-shared tree block is cow'd, we free the old block at once, and the
new block inherits old block's references. When a tree block with reference
count > 1 is cow'd, we increase the reference counts of all extents
the new block points to by one, and decrease the old block's reference count by
one.
This dead tree avoidance code removes the need to modify the reference
counts of lower level extents when a non-shared tree block is cow'd.
But we still need to update back ref for all pointers in the block.
This is because the location of the block is recorded in the back ref
item.
We can solve this by introducing a new type of back ref. The new
back ref provides information about pointer's key, level and in which
tree the pointer lives. This information allow us to find the pointer
by searching the tree. The shortcoming of the new back ref is that it
only works for pointers in tree blocks referenced by their owner trees.
This is mostly a problem for snapshots, where resolving one of these
fuzzy back references would be O(number_of_snapshots) and quite slow.
The solution used here is to use the fuzzy back references in the common
case where a given tree block is only referenced by one root,
and use the full back references when multiple roots have a reference
on a given block.
This commit adds per subvolume red-black tree to keep trace of cached
inodes. The red-black tree helps the balancing code to find cached
inodes whose inode numbers within a given range.
This commit improves the balancing code by introducing several data
structures to keep the state of balancing. The most important one
is the back ref cache. It caches how the upper level tree blocks are
referenced. This greatly reduce the overhead of checking back ref.
The improved balancing code scales significantly better with a large
number of snapshots.
This is a very large commit and was written in a number of
pieces. But, they depend heavily on the disk format change and were
squashed together to make sure git bisect didn't end up in a
bad state wrt space balancing or the format change.
Signed-off-by: Yan Zheng <zheng.yan@oracle.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2009-06-10 22:45:14 +08:00
|
|
|
case BTRFS_EXTENT_REF_V0_KEY:
|
|
|
|
|
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
|
|
|
|
|
print_extent_ref_v0(l, i);
|
|
|
|
|
#else
|
|
|
|
|
BUG();
|
|
|
|
|
#endif
|
2007-04-27 04:46:15 +08:00
|
|
|
case BTRFS_BLOCK_GROUP_ITEM_KEY:
|
|
|
|
|
bi = btrfs_item_ptr(l, i,
|
|
|
|
|
struct btrfs_block_group_item);
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_INFO "\t\tblock group used %llu\n",
|
|
|
|
|
(unsigned long long)
|
|
|
|
|
btrfs_disk_block_group_used(l, bi));
|
2007-03-16 00:56:47 +08:00
|
|
|
break;
|
2008-03-25 03:01:56 +08:00
|
|
|
case BTRFS_CHUNK_ITEM_KEY:
|
2009-01-06 10:25:51 +08:00
|
|
|
print_chunk(l, btrfs_item_ptr(l, i,
|
|
|
|
|
struct btrfs_chunk));
|
2008-03-25 03:01:56 +08:00
|
|
|
break;
|
|
|
|
|
case BTRFS_DEV_ITEM_KEY:
|
|
|
|
|
print_dev_item(l, btrfs_item_ptr(l, i,
|
|
|
|
|
struct btrfs_dev_item));
|
|
|
|
|
break;
|
|
|
|
|
case BTRFS_DEV_EXTENT_KEY:
|
|
|
|
|
dev_extent = btrfs_item_ptr(l, i,
|
|
|
|
|
struct btrfs_dev_extent);
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_INFO "\t\tdev extent chunk_tree %llu\n"
|
2008-04-16 03:41:47 +08:00
|
|
|
"\t\tchunk objectid %llu chunk offset %llu "
|
|
|
|
|
"length %llu\n",
|
|
|
|
|
(unsigned long long)
|
|
|
|
|
btrfs_dev_extent_chunk_tree(l, dev_extent),
|
|
|
|
|
(unsigned long long)
|
|
|
|
|
btrfs_dev_extent_chunk_objectid(l, dev_extent),
|
|
|
|
|
(unsigned long long)
|
|
|
|
|
btrfs_dev_extent_chunk_offset(l, dev_extent),
|
|
|
|
|
(unsigned long long)
|
|
|
|
|
btrfs_dev_extent_length(l, dev_extent));
|
2007-03-16 00:56:47 +08:00
|
|
|
};
|
2007-02-24 19:24:44 +08:00
|
|
|
}
|
|
|
|
|
}
|
2007-03-23 00:13:20 +08:00
|
|
|
|
2007-10-16 04:14:19 +08:00
|
|
|
void btrfs_print_tree(struct btrfs_root *root, struct extent_buffer *c)
|
2007-02-24 19:24:44 +08:00
|
|
|
{
|
2008-04-16 03:41:47 +08:00
|
|
|
int i; u32 nr;
|
2007-10-16 04:14:19 +08:00
|
|
|
struct btrfs_key key;
|
2007-10-16 04:15:53 +08:00
|
|
|
int level;
|
2007-02-24 19:24:44 +08:00
|
|
|
|
2007-10-16 04:14:19 +08:00
|
|
|
if (!c)
|
2007-02-24 19:24:44 +08:00
|
|
|
return;
|
2007-10-16 04:14:19 +08:00
|
|
|
nr = btrfs_header_nritems(c);
|
2007-10-16 04:15:53 +08:00
|
|
|
level = btrfs_header_level(c);
|
|
|
|
|
if (level == 0) {
|
2007-10-16 04:14:19 +08:00
|
|
|
btrfs_print_leaf(root, c);
|
2007-02-24 19:24:44 +08:00
|
|
|
return;
|
|
|
|
|
}
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_INFO "node %llu level %d total ptrs %d free spc %u\n",
|
2007-10-16 04:15:53 +08:00
|
|
|
(unsigned long long)btrfs_header_bytenr(c),
|
2007-10-16 04:14:19 +08:00
|
|
|
btrfs_header_level(c), nr,
|
2007-03-23 00:13:20 +08:00
|
|
|
(u32)BTRFS_NODEPTRS_PER_BLOCK(root) - nr);
|
2007-02-24 19:24:44 +08:00
|
|
|
for (i = 0; i < nr; i++) {
|
2007-10-16 04:14:19 +08:00
|
|
|
btrfs_node_key_to_cpu(c, &key, i);
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_INFO "\tkey %d (%llu %u %llu) block %llu\n",
|
2007-02-24 19:24:44 +08:00
|
|
|
i,
|
2007-10-16 04:14:19 +08:00
|
|
|
(unsigned long long)key.objectid,
|
|
|
|
|
key.type,
|
|
|
|
|
(unsigned long long)key.offset,
|
2007-06-12 19:50:13 +08:00
|
|
|
(unsigned long long)btrfs_node_blockptr(c, i));
|
2007-02-24 19:24:44 +08:00
|
|
|
}
|
|
|
|
|
for (i = 0; i < nr; i++) {
|
2007-10-16 04:14:19 +08:00
|
|
|
struct extent_buffer *next = read_tree_block(root,
|
2007-10-16 04:15:53 +08:00
|
|
|
btrfs_node_blockptr(c, i),
|
2008-05-13 00:59:19 +08:00
|
|
|
btrfs_level_size(root, level - 1),
|
|
|
|
|
btrfs_node_ptr_generation(c, i));
|
2007-03-13 00:01:18 +08:00
|
|
|
if (btrfs_is_leaf(next) &&
|
2007-10-16 04:14:19 +08:00
|
|
|
btrfs_header_level(c) != 1)
|
2007-02-24 19:24:44 +08:00
|
|
|
BUG();
|
2007-10-16 04:14:19 +08:00
|
|
|
if (btrfs_header_level(next) !=
|
|
|
|
|
btrfs_header_level(c) - 1)
|
2007-02-24 19:24:44 +08:00
|
|
|
BUG();
|
2007-10-16 04:14:19 +08:00
|
|
|
btrfs_print_tree(root, next);
|
|
|
|
|
free_extent_buffer(next);
|
2007-02-24 19:24:44 +08:00
|
|
|
}
|
|
|
|
|
}
|