Btrfs: grab write lock directly if write_lock_level is the max level
authorLiu Bo <bo.liu@linux.alibaba.com>
Fri, 18 May 2018 03:00:23 +0000 (11:00 +0800)
committerDavid Sterba <dsterba@suse.com>
Wed, 30 May 2018 14:46:51 +0000 (16:46 +0200)
commit662c653bfda58698cf48d7143a39bd3a063fd9c6
treec4996ca4d8c9fd5410c400dfdffaa1dadddb2cc6
parent1fc28d8e2e9bf22044f1bacd17fe941cd0df5ba6
Btrfs: grab write lock directly if write_lock_level is the max level

Typically, when acquiring root node's lock, btrfs tries its best to get
read lock and trade for write lock if @write_lock_level implies to do so.

In case of (cow && (p->keep_locks || p->lowest_level)), write_lock_level
is set to BTRFS_MAX_LEVEL, which means we need to acquire root node's
write lock directly.

In this particular case, the dance of acquiring read lock and then trading
for write lock can be saved.

Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
fs/btrfs/ctree.c