mirror of
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git
synced 2024-08-28 20:02:17 +00:00
![Filipe Manana](/assets/img/avatar_default.png)
Commit eafa4fd0ad
("btrfs: fix exhaustion of the system chunk array
due to concurrent allocations") fixed a problem that resulted in
exhausting the system chunk array in the superblock when there are many
tasks allocating chunks in parallel. Basically too many tasks enter the
first phase of chunk allocation without previous tasks having finished
their second phase of allocation, resulting in too many system chunks
being allocated. That was originally observed when running the fallocate
tests of stress-ng on a PowerPC machine, using a node size of 64K.
However that commit also introduced a deadlock where a task in phase 1 of
the chunk allocation waited for another task that had allocated a system
chunk to finish its phase 2, but that other task was waiting on an extent
buffer lock held by the first task, therefore resulting in both tasks not
making any progress. That change was later reverted by a patch with the
subject "btrfs: fix deadlock with concurrent chunk allocations involving
system chunks", since there is no simple and short solution to address it
and the deadlock is relatively easy to trigger on zoned filesystems, while
the system chunk array exhaustion is not so common.
This change reworks the chunk allocation to avoid the system chunk array
exhaustion. It accomplishes that by making the first phase of chunk
allocation do the updates of the device items in the chunk btree and the
insertion of the new chunk item in the chunk btree. This is done while
under the protection of the chunk mutex (fs_info->chunk_mutex), in the
same critical section that checks for available system space, allocates
a new system chunk if needed and reserves system chunk space. This way
we do not have chunk space reserved until the second phase completes.
The same logic is applied to chunk removal as well, since it keeps
reserved system space long after it is done updating the chunk btree.
For direct allocation of system chunks, the previous behaviour remains,
because otherwise we would deadlock on extent buffers of the chunk btree.
Changes to the chunk btree are by large done by chunk allocation and chunk
removal, which first reserve chunk system space and then later do changes
to the chunk btree. The other remaining cases are uncommon and correspond
to adding a device, removing a device and resizing a device. All these
other cases do not pre-reserve system space, they modify the chunk btree
right away, so they don't hold reserved space for a long period like chunk
allocation and chunk removal do.
The diff of this change is huge, but more than half of it is just addition
of comments describing both how things work regarding chunk allocation and
removal, including both the new behavior and the parts of the old behavior
that did not change.
CC: stable@vger.kernel.org # 5.12+
Tested-by: Shin'ichiro Kawasaki <shinichiro.kawasaki@wdc.com>
Tested-by: Naohiro Aota <naohiro.aota@wdc.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Tested-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
237 lines
7.6 KiB
C
237 lines
7.6 KiB
C
/* SPDX-License-Identifier: GPL-2.0 */
|
|
/*
|
|
* Copyright (C) 2007 Oracle. All rights reserved.
|
|
*/
|
|
|
|
#ifndef BTRFS_TRANSACTION_H
|
|
#define BTRFS_TRANSACTION_H
|
|
|
|
#include <linux/refcount.h>
|
|
#include "btrfs_inode.h"
|
|
#include "delayed-ref.h"
|
|
#include "ctree.h"
|
|
|
|
enum btrfs_trans_state {
|
|
TRANS_STATE_RUNNING,
|
|
TRANS_STATE_COMMIT_START,
|
|
TRANS_STATE_COMMIT_DOING,
|
|
TRANS_STATE_UNBLOCKED,
|
|
TRANS_STATE_SUPER_COMMITTED,
|
|
TRANS_STATE_COMPLETED,
|
|
TRANS_STATE_MAX,
|
|
};
|
|
|
|
#define BTRFS_TRANS_HAVE_FREE_BGS 0
|
|
#define BTRFS_TRANS_DIRTY_BG_RUN 1
|
|
#define BTRFS_TRANS_CACHE_ENOSPC 2
|
|
|
|
struct btrfs_transaction {
|
|
u64 transid;
|
|
/*
|
|
* total external writers(USERSPACE/START/ATTACH) in this
|
|
* transaction, it must be zero before the transaction is
|
|
* being committed
|
|
*/
|
|
atomic_t num_extwriters;
|
|
/*
|
|
* total writers in this transaction, it must be zero before the
|
|
* transaction can end
|
|
*/
|
|
atomic_t num_writers;
|
|
refcount_t use_count;
|
|
|
|
unsigned long flags;
|
|
|
|
/* Be protected by fs_info->trans_lock when we want to change it. */
|
|
enum btrfs_trans_state state;
|
|
int aborted;
|
|
struct list_head list;
|
|
struct extent_io_tree dirty_pages;
|
|
time64_t start_time;
|
|
wait_queue_head_t writer_wait;
|
|
wait_queue_head_t commit_wait;
|
|
struct list_head pending_snapshots;
|
|
struct list_head dev_update_list;
|
|
struct list_head switch_commits;
|
|
struct list_head dirty_bgs;
|
|
|
|
/*
|
|
* There is no explicit lock which protects io_bgs, rather its
|
|
* consistency is implied by the fact that all the sites which modify
|
|
* it do so under some form of transaction critical section, namely:
|
|
*
|
|
* - btrfs_start_dirty_block_groups - This function can only ever be
|
|
* run by one of the transaction committers. Refer to
|
|
* BTRFS_TRANS_DIRTY_BG_RUN usage in btrfs_commit_transaction
|
|
*
|
|
* - btrfs_write_dirty_blockgroups - this is called by
|
|
* commit_cowonly_roots from transaction critical section
|
|
* (TRANS_STATE_COMMIT_DOING)
|
|
*
|
|
* - btrfs_cleanup_dirty_bgs - called on transaction abort
|
|
*/
|
|
struct list_head io_bgs;
|
|
struct list_head dropped_roots;
|
|
struct extent_io_tree pinned_extents;
|
|
|
|
/*
|
|
* we need to make sure block group deletion doesn't race with
|
|
* free space cache writeout. This mutex keeps them from stomping
|
|
* on each other
|
|
*/
|
|
struct mutex cache_write_mutex;
|
|
spinlock_t dirty_bgs_lock;
|
|
/* Protected by spin lock fs_info->unused_bgs_lock. */
|
|
struct list_head deleted_bgs;
|
|
spinlock_t dropped_roots_lock;
|
|
struct btrfs_delayed_ref_root delayed_refs;
|
|
struct btrfs_fs_info *fs_info;
|
|
|
|
/*
|
|
* Number of ordered extents the transaction must wait for before
|
|
* committing. These are ordered extents started by a fast fsync.
|
|
*/
|
|
atomic_t pending_ordered;
|
|
wait_queue_head_t pending_wait;
|
|
|
|
spinlock_t releasing_ebs_lock;
|
|
struct list_head releasing_ebs;
|
|
};
|
|
|
|
#define __TRANS_FREEZABLE (1U << 0)
|
|
|
|
#define __TRANS_START (1U << 9)
|
|
#define __TRANS_ATTACH (1U << 10)
|
|
#define __TRANS_JOIN (1U << 11)
|
|
#define __TRANS_JOIN_NOLOCK (1U << 12)
|
|
#define __TRANS_DUMMY (1U << 13)
|
|
#define __TRANS_JOIN_NOSTART (1U << 14)
|
|
|
|
#define TRANS_START (__TRANS_START | __TRANS_FREEZABLE)
|
|
#define TRANS_ATTACH (__TRANS_ATTACH)
|
|
#define TRANS_JOIN (__TRANS_JOIN | __TRANS_FREEZABLE)
|
|
#define TRANS_JOIN_NOLOCK (__TRANS_JOIN_NOLOCK)
|
|
#define TRANS_JOIN_NOSTART (__TRANS_JOIN_NOSTART)
|
|
|
|
#define TRANS_EXTWRITERS (__TRANS_START | __TRANS_ATTACH)
|
|
|
|
struct btrfs_trans_handle {
|
|
u64 transid;
|
|
u64 bytes_reserved;
|
|
u64 chunk_bytes_reserved;
|
|
unsigned long delayed_ref_updates;
|
|
struct btrfs_transaction *transaction;
|
|
struct btrfs_block_rsv *block_rsv;
|
|
struct btrfs_block_rsv *orig_rsv;
|
|
refcount_t use_count;
|
|
unsigned int type;
|
|
/*
|
|
* Error code of transaction abort, set outside of locks and must use
|
|
* the READ_ONCE/WRITE_ONCE access
|
|
*/
|
|
short aborted;
|
|
bool adding_csums;
|
|
bool allocating_chunk;
|
|
bool removing_chunk;
|
|
bool reloc_reserved;
|
|
bool in_fsync;
|
|
struct btrfs_root *root;
|
|
struct btrfs_fs_info *fs_info;
|
|
struct list_head new_bgs;
|
|
};
|
|
|
|
/*
|
|
* The abort status can be changed between calls and is not protected by locks.
|
|
* This accepts btrfs_transaction and btrfs_trans_handle as types. Once it's
|
|
* set to a non-zero value it does not change, so the macro should be in checks
|
|
* but is not necessary for further reads of the value.
|
|
*/
|
|
#define TRANS_ABORTED(trans) (unlikely(READ_ONCE((trans)->aborted)))
|
|
|
|
struct btrfs_pending_snapshot {
|
|
struct dentry *dentry;
|
|
struct inode *dir;
|
|
struct btrfs_root *root;
|
|
struct btrfs_root_item *root_item;
|
|
struct btrfs_root *snap;
|
|
struct btrfs_qgroup_inherit *inherit;
|
|
struct btrfs_path *path;
|
|
/* block reservation for the operation */
|
|
struct btrfs_block_rsv block_rsv;
|
|
/* extra metadata reservation for relocation */
|
|
int error;
|
|
/* Preallocated anonymous block device number */
|
|
dev_t anon_dev;
|
|
bool readonly;
|
|
struct list_head list;
|
|
};
|
|
|
|
static inline void btrfs_set_inode_last_trans(struct btrfs_trans_handle *trans,
|
|
struct btrfs_inode *inode)
|
|
{
|
|
spin_lock(&inode->lock);
|
|
inode->last_trans = trans->transaction->transid;
|
|
inode->last_sub_trans = inode->root->log_transid;
|
|
inode->last_log_commit = inode->last_sub_trans - 1;
|
|
spin_unlock(&inode->lock);
|
|
}
|
|
|
|
/*
|
|
* Make qgroup codes to skip given qgroupid, means the old/new_roots for
|
|
* qgroup won't contain the qgroupid in it.
|
|
*/
|
|
static inline void btrfs_set_skip_qgroup(struct btrfs_trans_handle *trans,
|
|
u64 qgroupid)
|
|
{
|
|
struct btrfs_delayed_ref_root *delayed_refs;
|
|
|
|
delayed_refs = &trans->transaction->delayed_refs;
|
|
WARN_ON(delayed_refs->qgroup_to_skip);
|
|
delayed_refs->qgroup_to_skip = qgroupid;
|
|
}
|
|
|
|
static inline void btrfs_clear_skip_qgroup(struct btrfs_trans_handle *trans)
|
|
{
|
|
struct btrfs_delayed_ref_root *delayed_refs;
|
|
|
|
delayed_refs = &trans->transaction->delayed_refs;
|
|
WARN_ON(!delayed_refs->qgroup_to_skip);
|
|
delayed_refs->qgroup_to_skip = 0;
|
|
}
|
|
|
|
int btrfs_end_transaction(struct btrfs_trans_handle *trans);
|
|
struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root,
|
|
unsigned int num_items);
|
|
struct btrfs_trans_handle *btrfs_start_transaction_fallback_global_rsv(
|
|
struct btrfs_root *root,
|
|
unsigned int num_items);
|
|
struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root);
|
|
struct btrfs_trans_handle *btrfs_join_transaction_spacecache(struct btrfs_root *root);
|
|
struct btrfs_trans_handle *btrfs_join_transaction_nostart(struct btrfs_root *root);
|
|
struct btrfs_trans_handle *btrfs_attach_transaction(struct btrfs_root *root);
|
|
struct btrfs_trans_handle *btrfs_attach_transaction_barrier(
|
|
struct btrfs_root *root);
|
|
int btrfs_wait_for_commit(struct btrfs_fs_info *fs_info, u64 transid);
|
|
|
|
void btrfs_add_dead_root(struct btrfs_root *root);
|
|
int btrfs_defrag_root(struct btrfs_root *root);
|
|
int btrfs_clean_one_deleted_snapshot(struct btrfs_root *root);
|
|
int btrfs_commit_transaction(struct btrfs_trans_handle *trans);
|
|
int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans);
|
|
int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans);
|
|
bool btrfs_should_end_transaction(struct btrfs_trans_handle *trans);
|
|
void btrfs_throttle(struct btrfs_fs_info *fs_info);
|
|
int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans,
|
|
struct btrfs_root *root);
|
|
int btrfs_write_marked_extents(struct btrfs_fs_info *fs_info,
|
|
struct extent_io_tree *dirty_pages, int mark);
|
|
int btrfs_wait_tree_log_extents(struct btrfs_root *root, int mark);
|
|
int btrfs_transaction_blocked(struct btrfs_fs_info *info);
|
|
int btrfs_transaction_in_commit(struct btrfs_fs_info *info);
|
|
void btrfs_put_transaction(struct btrfs_transaction *transaction);
|
|
void btrfs_apply_pending_changes(struct btrfs_fs_info *fs_info);
|
|
void btrfs_add_dropped_root(struct btrfs_trans_handle *trans,
|
|
struct btrfs_root *root);
|
|
void btrfs_trans_release_chunk_metadata(struct btrfs_trans_handle *trans);
|
|
|
|
#endif
|