mirror of
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git
synced 2024-09-03 17:40:04 +00:00
![Filipe Manana](/assets/img/avatar_default.png)
Currently, when logging a directory, we copy both dir items and dir index items from the fs/subvolume tree to the log tree. Both items have exactly the same data (same struct btrfs_dir_item), the difference lies in the key values, where a dir index key contains the index number of a directory entry while the dir item key does not, as it's used for doing fast lookups of an entry by name, while the former is used for sorting entries when listing a directory. We can exploit that and log only the dir index items, since they contain all the information needed to correctly add, replace and delete directory entries when replaying a log tree. Logging only the dir index items is also backward and forward compatible: an unpatched kernel (without this change) can correctly replay a log tree generated by a patched kernel (with this patch), and a patched kernel can correctly replay a log tree generated by an unpatched kernel. The backward compatibility is ensured because: 1) For inserting a new dentry: a dentry is only inserted when we find a new dir index key - we can only insert if we know the dir index offset, which is encoded in the dir index key's offset; 2) For deleting dentries: during log replay, before adding or replacing dentries, we first replay dentry deletions. Whenever we find a dir item key or a dir index key in the subvolume/fs tree that is not logged in a range for which the log tree is authoritative, we do the unlink of the dentry, which removes both the existing dir item key and the dir index key. Therefore logging just dir index keys is enough to ensure dentry deletions are correctly replayed; 3) For dentry replacements: they work when we log only dir index keys and this is mostly due to a combination of 1) and 2). If we replace a dentry with name "foobar" to point from inode A to inode B, then we know the dir index key for the new dentry is different from the old one, as it has an index number (key offset) larger than the old one. This results in replaying a deletion, through replay_dir_deletes(), that causes the old dentry to be removed, both the dir item key and the dir index key, as mentioned at 2). Then when processing the new dir index key, we add the new dentry, adding both a new dir item key and a new index key pointing to inode B, as stated in 1). The forward compatibility, the ability for a patched kernel to replay a log created by an older, unpatched kernel, comes from the changes required for making sure we are able to replay a log that only contains dir index keys - we simply ignore every dir item key we find. So modify directory logging to log only dir index items, and modify the log replay process to ignore dir item keys, from log trees created by an unpatched kernel, and process only with dir index keys. This reduces the amount of logged metadata by about half, and therefore the time spent logging or fsyncing large directories (less CPU time and less IO). The following test script was used to measure this change: #!/bin/bash DEV=/dev/nvme0n1 MNT=/mnt/nvme0n1 NUM_NEW_FILES=1000000 NUM_FILE_DELETES=10000 mkfs.btrfs -f $DEV mount -o ssd $DEV $MNT mkdir $MNT/testdir for ((i = 1; i <= $NUM_NEW_FILES; i++)); do echo -n > $MNT/testdir/file_$i done start=$(date +%s%N) xfs_io -c "fsync" $MNT/testdir end=$(date +%s%N) dur=$(( (end - start) / 1000000 )) echo "dir fsync took $dur ms after adding $NUM_NEW_FILES files" # sync to force transaction commit and wipeout the log. sync del_inc=$(( $NUM_NEW_FILES / $NUM_FILE_DELETES )) for ((i = 1; i <= $NUM_NEW_FILES; i += $del_inc)); do rm -f $MNT/testdir/file_$i done start=$(date +%s%N) xfs_io -c "fsync" $MNT/testdir end=$(date +%s%N) dur=$(( (end - start) / 1000000 )) echo "dir fsync took $dur ms after deleting $NUM_FILE_DELETES files" echo umount $MNT The tests were run on a physical machine, with a non-debug kernel (Debian's default kernel config), for different values of $NUM_NEW_FILES and $NUM_FILE_DELETES, and the results were the following: ** Before patch, NUM_NEW_FILES = 1 000 000, NUM_DELETE_FILES = 10 000 ** dir fsync took 8412 ms after adding 1000000 files dir fsync took 500 ms after deleting 10000 files ** After patch, NUM_NEW_FILES = 1 000 000, NUM_DELETE_FILES = 10 000 ** dir fsync took 4252 ms after adding 1000000 files (-49.5%) dir fsync took 269 ms after deleting 10000 files (-46.2%) ** Before patch, NUM_NEW_FILES = 100 000, NUM_DELETE_FILES = 1 000 ** dir fsync took 745 ms after adding 100000 files dir fsync took 59 ms after deleting 1000 files ** After patch, NUM_NEW_FILES = 100 000, NUM_DELETE_FILES = 1 000 ** dir fsync took 404 ms after adding 100000 files (-45.8%) dir fsync took 31 ms after deleting 1000 files (-47.5%) ** Before patch, NUM_NEW_FILES = 10 000, NUM_DELETE_FILES = 1 000 ** dir fsync took 67 ms after adding 10000 files dir fsync took 9 ms after deleting 1000 files ** After patch, NUM_NEW_FILES = 10 000, NUM_DELETE_FILES = 1 000 ** dir fsync took 36 ms after adding 10000 files (-46.3%) dir fsync took 5 ms after deleting 1000 files (-44.4%) ** Before patch, NUM_NEW_FILES = 1 000, NUM_DELETE_FILES = 100 ** dir fsync took 9 ms after adding 1000 files dir fsync took 4 ms after deleting 100 files ** After patch, NUM_NEW_FILES = 1 000, NUM_DELETE_FILES = 100 ** dir fsync took 7 ms after adding 1000 files (-22.2%) dir fsync took 3 ms after deleting 100 files (-25.0%) Reviewed-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: David Sterba <dsterba@suse.com>
419 lines
12 KiB
C
419 lines
12 KiB
C
/* SPDX-License-Identifier: GPL-2.0 */
|
|
/*
|
|
* Copyright (C) 2007 Oracle. All rights reserved.
|
|
*/
|
|
|
|
#ifndef BTRFS_INODE_H
|
|
#define BTRFS_INODE_H
|
|
|
|
#include <linux/hash.h>
|
|
#include <linux/refcount.h>
|
|
#include "extent_map.h"
|
|
#include "extent_io.h"
|
|
#include "ordered-data.h"
|
|
#include "delayed-inode.h"
|
|
|
|
/*
|
|
* ordered_data_close is set by truncate when a file that used
|
|
* to have good data has been truncated to zero. When it is set
|
|
* the btrfs file release call will add this inode to the
|
|
* ordered operations list so that we make sure to flush out any
|
|
* new data the application may have written before commit.
|
|
*/
|
|
enum {
|
|
BTRFS_INODE_FLUSH_ON_CLOSE,
|
|
BTRFS_INODE_DUMMY,
|
|
BTRFS_INODE_IN_DEFRAG,
|
|
BTRFS_INODE_HAS_ASYNC_EXTENT,
|
|
/*
|
|
* Always set under the VFS' inode lock, otherwise it can cause races
|
|
* during fsync (we start as a fast fsync and then end up in a full
|
|
* fsync racing with ordered extent completion).
|
|
*/
|
|
BTRFS_INODE_NEEDS_FULL_SYNC,
|
|
BTRFS_INODE_COPY_EVERYTHING,
|
|
BTRFS_INODE_IN_DELALLOC_LIST,
|
|
BTRFS_INODE_HAS_PROPS,
|
|
BTRFS_INODE_SNAPSHOT_FLUSH,
|
|
/*
|
|
* Set and used when logging an inode and it serves to signal that an
|
|
* inode does not have xattrs, so subsequent fsyncs can avoid searching
|
|
* for xattrs to log. This bit must be cleared whenever a xattr is added
|
|
* to an inode.
|
|
*/
|
|
BTRFS_INODE_NO_XATTRS,
|
|
/*
|
|
* Set when we are in a context where we need to start a transaction and
|
|
* have dirty pages with the respective file range locked. This is to
|
|
* ensure that when reserving space for the transaction, if we are low
|
|
* on available space and need to flush delalloc, we will not flush
|
|
* delalloc for this inode, because that could result in a deadlock (on
|
|
* the file range, inode's io_tree).
|
|
*/
|
|
BTRFS_INODE_NO_DELALLOC_FLUSH,
|
|
/*
|
|
* Set when we are working on enabling verity for a file. Computing and
|
|
* writing the whole Merkle tree can take a while so we want to prevent
|
|
* races where two separate tasks attempt to simultaneously start verity
|
|
* on the same file.
|
|
*/
|
|
BTRFS_INODE_VERITY_IN_PROGRESS,
|
|
};
|
|
|
|
/* in memory btrfs inode */
|
|
struct btrfs_inode {
|
|
/* which subvolume this inode belongs to */
|
|
struct btrfs_root *root;
|
|
|
|
/* key used to find this inode on disk. This is used by the code
|
|
* to read in roots of subvolumes
|
|
*/
|
|
struct btrfs_key location;
|
|
|
|
/*
|
|
* Lock for counters and all fields used to determine if the inode is in
|
|
* the log or not (last_trans, last_sub_trans, last_log_commit,
|
|
* logged_trans), to access/update new_delalloc_bytes and to update the
|
|
* VFS' inode number of bytes used.
|
|
*/
|
|
spinlock_t lock;
|
|
|
|
/* the extent_tree has caches of all the extent mappings to disk */
|
|
struct extent_map_tree extent_tree;
|
|
|
|
/* the io_tree does range state (DIRTY, LOCKED etc) */
|
|
struct extent_io_tree io_tree;
|
|
|
|
/* special utility tree used to record which mirrors have already been
|
|
* tried when checksums fail for a given block
|
|
*/
|
|
struct extent_io_tree io_failure_tree;
|
|
|
|
/*
|
|
* Keep track of where the inode has extent items mapped in order to
|
|
* make sure the i_size adjustments are accurate
|
|
*/
|
|
struct extent_io_tree file_extent_tree;
|
|
|
|
/* held while logging the inode in tree-log.c */
|
|
struct mutex log_mutex;
|
|
|
|
/* used to order data wrt metadata */
|
|
struct btrfs_ordered_inode_tree ordered_tree;
|
|
|
|
/* list of all the delalloc inodes in the FS. There are times we need
|
|
* to write all the delalloc pages to disk, and this list is used
|
|
* to walk them all.
|
|
*/
|
|
struct list_head delalloc_inodes;
|
|
|
|
/* node for the red-black tree that links inodes in subvolume root */
|
|
struct rb_node rb_node;
|
|
|
|
unsigned long runtime_flags;
|
|
|
|
/* Keep track of who's O_SYNC/fsyncing currently */
|
|
atomic_t sync_writers;
|
|
|
|
/* full 64 bit generation number, struct vfs_inode doesn't have a big
|
|
* enough field for this.
|
|
*/
|
|
u64 generation;
|
|
|
|
/*
|
|
* transid of the trans_handle that last modified this inode
|
|
*/
|
|
u64 last_trans;
|
|
|
|
/*
|
|
* transid that last logged this inode
|
|
*/
|
|
u64 logged_trans;
|
|
|
|
/*
|
|
* log transid when this inode was last modified
|
|
*/
|
|
int last_sub_trans;
|
|
|
|
/* a local copy of root's last_log_commit */
|
|
int last_log_commit;
|
|
|
|
/*
|
|
* Total number of bytes pending delalloc, used by stat to calculate the
|
|
* real block usage of the file. This is used only for files.
|
|
*/
|
|
u64 delalloc_bytes;
|
|
|
|
union {
|
|
/*
|
|
* Total number of bytes pending delalloc that fall within a file
|
|
* range that is either a hole or beyond EOF (and no prealloc extent
|
|
* exists in the range). This is always <= delalloc_bytes and this
|
|
* is used only for files.
|
|
*/
|
|
u64 new_delalloc_bytes;
|
|
/*
|
|
* The offset of the last dir index key that was logged.
|
|
* This is used only for directories.
|
|
*/
|
|
u64 last_dir_index_offset;
|
|
};
|
|
|
|
/*
|
|
* total number of bytes pending defrag, used by stat to check whether
|
|
* it needs COW.
|
|
*/
|
|
u64 defrag_bytes;
|
|
|
|
/*
|
|
* the size of the file stored in the metadata on disk. data=ordered
|
|
* means the in-memory i_size might be larger than the size on disk
|
|
* because not all the blocks are written yet.
|
|
*/
|
|
u64 disk_i_size;
|
|
|
|
/*
|
|
* if this is a directory then index_cnt is the counter for the index
|
|
* number for new files that are created
|
|
*/
|
|
u64 index_cnt;
|
|
|
|
/* Cache the directory index number to speed the dir/file remove */
|
|
u64 dir_index;
|
|
|
|
/* the fsync log has some corner cases that mean we have to check
|
|
* directories to see if any unlinks have been done before
|
|
* the directory was logged. See tree-log.c for all the
|
|
* details
|
|
*/
|
|
u64 last_unlink_trans;
|
|
|
|
/*
|
|
* The id/generation of the last transaction where this inode was
|
|
* either the source or the destination of a clone/dedupe operation.
|
|
* Used when logging an inode to know if there are shared extents that
|
|
* need special care when logging checksum items, to avoid duplicate
|
|
* checksum items in a log (which can lead to a corruption where we end
|
|
* up with missing checksum ranges after log replay).
|
|
* Protected by the vfs inode lock.
|
|
*/
|
|
u64 last_reflink_trans;
|
|
|
|
/*
|
|
* Number of bytes outstanding that are going to need csums. This is
|
|
* used in ENOSPC accounting.
|
|
*/
|
|
u64 csum_bytes;
|
|
|
|
/* Backwards incompatible flags, lower half of inode_item::flags */
|
|
u32 flags;
|
|
/* Read-only compatibility flags, upper half of inode_item::flags */
|
|
u32 ro_flags;
|
|
|
|
/*
|
|
* Counters to keep track of the number of extent item's we may use due
|
|
* to delalloc and such. outstanding_extents is the number of extent
|
|
* items we think we'll end up using, and reserved_extents is the number
|
|
* of extent items we've reserved metadata for.
|
|
*/
|
|
unsigned outstanding_extents;
|
|
|
|
struct btrfs_block_rsv block_rsv;
|
|
|
|
/*
|
|
* Cached values of inode properties
|
|
*/
|
|
unsigned prop_compress; /* per-file compression algorithm */
|
|
/*
|
|
* Force compression on the file using the defrag ioctl, could be
|
|
* different from prop_compress and takes precedence if set
|
|
*/
|
|
unsigned defrag_compress;
|
|
|
|
struct btrfs_delayed_node *delayed_node;
|
|
|
|
/* File creation time. */
|
|
struct timespec64 i_otime;
|
|
|
|
/* Hook into fs_info->delayed_iputs */
|
|
struct list_head delayed_iput;
|
|
|
|
struct rw_semaphore i_mmap_lock;
|
|
struct inode vfs_inode;
|
|
};
|
|
|
|
static inline u32 btrfs_inode_sectorsize(const struct btrfs_inode *inode)
|
|
{
|
|
return inode->root->fs_info->sectorsize;
|
|
}
|
|
|
|
static inline struct btrfs_inode *BTRFS_I(const struct inode *inode)
|
|
{
|
|
return container_of(inode, struct btrfs_inode, vfs_inode);
|
|
}
|
|
|
|
static inline unsigned long btrfs_inode_hash(u64 objectid,
|
|
const struct btrfs_root *root)
|
|
{
|
|
u64 h = objectid ^ (root->root_key.objectid * GOLDEN_RATIO_PRIME);
|
|
|
|
#if BITS_PER_LONG == 32
|
|
h = (h >> 32) ^ (h & 0xffffffff);
|
|
#endif
|
|
|
|
return (unsigned long)h;
|
|
}
|
|
|
|
static inline void btrfs_insert_inode_hash(struct inode *inode)
|
|
{
|
|
unsigned long h = btrfs_inode_hash(inode->i_ino, BTRFS_I(inode)->root);
|
|
|
|
__insert_inode_hash(inode, h);
|
|
}
|
|
|
|
static inline u64 btrfs_ino(const struct btrfs_inode *inode)
|
|
{
|
|
u64 ino = inode->location.objectid;
|
|
|
|
/*
|
|
* !ino: btree_inode
|
|
* type == BTRFS_ROOT_ITEM_KEY: subvol dir
|
|
*/
|
|
if (!ino || inode->location.type == BTRFS_ROOT_ITEM_KEY)
|
|
ino = inode->vfs_inode.i_ino;
|
|
return ino;
|
|
}
|
|
|
|
static inline void btrfs_i_size_write(struct btrfs_inode *inode, u64 size)
|
|
{
|
|
i_size_write(&inode->vfs_inode, size);
|
|
inode->disk_i_size = size;
|
|
}
|
|
|
|
static inline bool btrfs_is_free_space_inode(struct btrfs_inode *inode)
|
|
{
|
|
struct btrfs_root *root = inode->root;
|
|
|
|
if (root == root->fs_info->tree_root &&
|
|
btrfs_ino(inode) != BTRFS_BTREE_INODE_OBJECTID)
|
|
return true;
|
|
if (inode->location.objectid == BTRFS_FREE_INO_OBJECTID)
|
|
return true;
|
|
return false;
|
|
}
|
|
|
|
static inline bool is_data_inode(struct inode *inode)
|
|
{
|
|
return btrfs_ino(BTRFS_I(inode)) != BTRFS_BTREE_INODE_OBJECTID;
|
|
}
|
|
|
|
static inline void btrfs_mod_outstanding_extents(struct btrfs_inode *inode,
|
|
int mod)
|
|
{
|
|
lockdep_assert_held(&inode->lock);
|
|
inode->outstanding_extents += mod;
|
|
if (btrfs_is_free_space_inode(inode))
|
|
return;
|
|
trace_btrfs_inode_mod_outstanding_extents(inode->root, btrfs_ino(inode),
|
|
mod);
|
|
}
|
|
|
|
/*
|
|
* Called every time after doing a buffered, direct IO or memory mapped write.
|
|
*
|
|
* This is to ensure that if we write to a file that was previously fsynced in
|
|
* the current transaction, then try to fsync it again in the same transaction,
|
|
* we will know that there were changes in the file and that it needs to be
|
|
* logged.
|
|
*/
|
|
static inline void btrfs_set_inode_last_sub_trans(struct btrfs_inode *inode)
|
|
{
|
|
spin_lock(&inode->lock);
|
|
inode->last_sub_trans = inode->root->log_transid;
|
|
spin_unlock(&inode->lock);
|
|
}
|
|
|
|
static inline bool btrfs_inode_in_log(struct btrfs_inode *inode, u64 generation)
|
|
{
|
|
bool ret = false;
|
|
|
|
spin_lock(&inode->lock);
|
|
if (inode->logged_trans == generation &&
|
|
inode->last_sub_trans <= inode->last_log_commit &&
|
|
inode->last_sub_trans <= inode->root->last_log_commit)
|
|
ret = true;
|
|
spin_unlock(&inode->lock);
|
|
return ret;
|
|
}
|
|
|
|
struct btrfs_dio_private {
|
|
struct inode *inode;
|
|
|
|
/*
|
|
* Since DIO can use anonymous page, we cannot use page_offset() to
|
|
* grab the file offset, thus need a dedicated member for file offset.
|
|
*/
|
|
u64 file_offset;
|
|
u64 disk_bytenr;
|
|
/* Used for bio::bi_size */
|
|
u32 bytes;
|
|
|
|
/*
|
|
* References to this structure. There is one reference per in-flight
|
|
* bio plus one while we're still setting up.
|
|
*/
|
|
refcount_t refs;
|
|
|
|
/* dio_bio came from fs/direct-io.c */
|
|
struct bio *dio_bio;
|
|
|
|
/* Array of checksums */
|
|
u8 csums[];
|
|
};
|
|
|
|
/*
|
|
* btrfs_inode_item stores flags in a u64, btrfs_inode stores them in two
|
|
* separate u32s. These two functions convert between the two representations.
|
|
*/
|
|
static inline u64 btrfs_inode_combine_flags(u32 flags, u32 ro_flags)
|
|
{
|
|
return (flags | ((u64)ro_flags << 32));
|
|
}
|
|
|
|
static inline void btrfs_inode_split_flags(u64 inode_item_flags,
|
|
u32 *flags, u32 *ro_flags)
|
|
{
|
|
*flags = (u32)inode_item_flags;
|
|
*ro_flags = (u32)(inode_item_flags >> 32);
|
|
}
|
|
|
|
/* Array of bytes with variable length, hexadecimal format 0x1234 */
|
|
#define CSUM_FMT "0x%*phN"
|
|
#define CSUM_FMT_VALUE(size, bytes) size, bytes
|
|
|
|
static inline void btrfs_print_data_csum_error(struct btrfs_inode *inode,
|
|
u64 logical_start, u8 *csum, u8 *csum_expected, int mirror_num)
|
|
{
|
|
struct btrfs_root *root = inode->root;
|
|
const u32 csum_size = root->fs_info->csum_size;
|
|
|
|
/* Output minus objectid, which is more meaningful */
|
|
if (root->root_key.objectid >= BTRFS_LAST_FREE_OBJECTID)
|
|
btrfs_warn_rl(root->fs_info,
|
|
"csum failed root %lld ino %lld off %llu csum " CSUM_FMT " expected csum " CSUM_FMT " mirror %d",
|
|
root->root_key.objectid, btrfs_ino(inode),
|
|
logical_start,
|
|
CSUM_FMT_VALUE(csum_size, csum),
|
|
CSUM_FMT_VALUE(csum_size, csum_expected),
|
|
mirror_num);
|
|
else
|
|
btrfs_warn_rl(root->fs_info,
|
|
"csum failed root %llu ino %llu off %llu csum " CSUM_FMT " expected csum " CSUM_FMT " mirror %d",
|
|
root->root_key.objectid, btrfs_ino(inode),
|
|
logical_start,
|
|
CSUM_FMT_VALUE(csum_size, csum),
|
|
CSUM_FMT_VALUE(csum_size, csum_expected),
|
|
mirror_num);
|
|
}
|
|
|
|
#endif
|