From: Kent Overstreet <kent.overstreet@linux.dev>
Date: Thu, 22 May 2025 22:12:54 +0000 (-0400)
Subject: bcachefs: bch2_path_get() reuses paths if upgrade_fails & !should_be_locked
X-Git-Tag: v6.16-rc1~211^2~12
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=be9fecdcdaf730dbf2ca70dbe5b6d42922df50d6;p=thirdparty%2Flinux.git

bcachefs: bch2_path_get() reuses paths if upgrade_fails & !should_be_locked

Small additional optimization over the previous patch, bringing us
closer to the original behaviour, except when we need to clone to avoid
a transaction restart.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
---

diff --git a/fs/bcachefs/btree_locking.c b/fs/bcachefs/btree_locking.c
index 6e43269a9c47a..78f485ed17469 100644
--- a/fs/bcachefs/btree_locking.c
+++ b/fs/bcachefs/btree_locking.c
@@ -628,8 +628,13 @@ bool __bch2_btree_path_upgrade_norestart(struct btree_trans *trans,
 {
 	path->locks_want = new_locks_want;
 
-	struct get_locks_fail f = {};
-	bool ret = !btree_path_get_locks(trans, path, true, &f, 0);
+	/*
+	 * If we need it locked, we can't touch it. Otherwise, we can return
+	 * success - bch2_path_get() will use this path, and it'll just be
+	 * retraversed:
+	 */
+	bool ret = !btree_path_get_locks(trans, path, true, NULL, 0) ||
+		!path->should_be_locked;
 
 	bch2_btree_path_verify_locks(path);
 	return ret;