On heavy workloads, we're seeing soft lockup warnings on
root->inode_lock in __btrfs_release_delayed_node. The low hanging fruit
is to reduce the size of the critical section.
Signed-off-by: Jeff Mahoney <je...@suse.com>
---
fs/btrfs/delayed-inode.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c
index 33e561a..da775bf 100644
--- a/fs/btrfs/delayed-inode.c
+++ b/fs/btrfs/delayed-inode.c
@@ -149,8 +149,8 @@ again:
spin_lock(&root->inode_lock);
ret = radix_tree_insert(&root->delayed_nodes_tree, ino, node);
if (ret == -EEXIST) {
- kmem_cache_free(delayed_node_cache, node);
spin_unlock(&root->inode_lock);
+ kmem_cache_free(delayed_node_cache, node);
radix_tree_preload_end();
goto again;
}
@@ -267,14 +267,17 @@ static void __btrfs_release_delayed_node(
mutex_unlock(&delayed_node->mutex);
if (atomic_dec_and_test(&delayed_node->refs)) {
+ bool free = false;
struct btrfs_root *root = delayed_node->root;
spin_lock(&root->inode_lock);
if (atomic_read(&delayed_node->refs) == 0) {
radix_tree_delete(&root->delayed_nodes_tree,
delayed_node->inode_id);
- kmem_cache_free(delayed_node_cache, delayed_node);
+ free = true;
}
spin_unlock(&root->inode_lock);
+ if (free)
+ kmem_cache_free(delayed_node_cache, delayed_node);
}
}
--
Jeff Mahoney
SUSE Labs
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
