Re: [PATCH 0/5] btrfs: avoid some block group rbtree lock contention

[Nikolay Borisov <nborisov@suse.com>]

On 13.04.22 г. 18:20 ч., fdmanana@kernel.org wrote:
> From: Filipe Manana <fdmanana@suse.com>
> 
> This patchset allows for better concurrency when accessing the red black
> tree of block groups, which is used very frequently and most accesses are
> read-only, as well as avoid some unnecessary searches in the tree during
> NOCOW writes. Details in the changelogs.
> 
> Filipe Manana (5):
>    btrfs: remove search start argument from first_logical_byte()
>    btrfs: use rbtree with leftmost node cached for tracking lowest block group
>    btrfs: use a read/write lock for protecting the block groups tree
>    btrfs: return block group directly at btrfs_next_block_group()
>    btrfs: avoid double search for block group during NOCOW writes
> 
>   fs/btrfs/block-group.c      | 130 ++++++++++++++++++++----------------
>   fs/btrfs/block-group.h      |   5 +-
>   fs/btrfs/ctree.h            |   5 +-
>   fs/btrfs/disk-io.c          |   5 +-
>   fs/btrfs/extent-tree.c      |  29 ++++----
>   fs/btrfs/free-space-cache.c |   2 +-
>   fs/btrfs/free-space-tree.c  |   2 +-
>   fs/btrfs/inode.c            |  26 +++++---
>   fs/btrfs/transaction.c      |   4 +-
>   9 files changed, 114 insertions(+), 94 deletions(-)
> 


Reviewed-by: Nikolay Borisov <nborisov@suse.com>