[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20240304163543.6700-3-shikemeng@huaweicloud.com>
Date: Tue, 5 Mar 2024 00:35:42 +0800
From: Kemeng Shi <shikemeng@...weicloud.com>
To: tytso@....edu,
adilger.kernel@...ger.ca
Cc: linux-ext4@...r.kernel.org,
linux-kernel@...r.kernel.org,
naresh.kamboju@...aro.org,
daniel.diaz@...aro.org,
linux@...ck-us.net,
brauner@...nel.org
Subject: [PATCH v4 2/3] ext4: hold group lock in ext4 kunit test
Although there is no concurrent block allocation/free in unit test,
internal functions mb_mark_used and mb_free_blocks assert group
lock is always held. Acquire group before calling mb_mark_used and
mb_free_blocks in unit test to avoid the assertion.
Signed-off-by: Kemeng Shi <shikemeng@...weicloud.com>
Reported-by: Guenter Roeck <linux@...ck-us.net>
---
fs/ext4/mballoc-test.c | 9 +++++++++
1 file changed, 9 insertions(+)
diff --git a/fs/ext4/mballoc-test.c b/fs/ext4/mballoc-test.c
index 9d061f80f150..4d9eb4807c98 100644
--- a/fs/ext4/mballoc-test.c
+++ b/fs/ext4/mballoc-test.c
@@ -701,7 +701,10 @@ test_mb_mark_used_range(struct kunit *test, struct ext4_buddy *e4b,
ex.fe_start = start;
ex.fe_len = len;
ex.fe_group = TEST_GOAL_GROUP;
+
+ ext4_lock_group(sb, TEST_GOAL_GROUP);
mb_mark_used(e4b, &ex);
+ ext4_unlock_group(sb, TEST_GOAL_GROUP);
mb_set_bits(bitmap, start, len);
/* bypass bb_free validatoin in ext4_mb_generate_buddy */
@@ -761,7 +764,9 @@ test_mb_free_blocks_range(struct kunit *test, struct ext4_buddy *e4b,
if (len == 0)
return;
+ ext4_lock_group(sb, e4b->bd_group);
mb_free_blocks(NULL, e4b, start, len);
+ ext4_unlock_group(sb, e4b->bd_group);
mb_clear_bits(bitmap, start, len);
/* bypass bb_free validatoin in ext4_mb_generate_buddy */
@@ -805,7 +810,11 @@ static void test_mb_free_blocks(struct kunit *test)
ex.fe_start = 0;
ex.fe_len = EXT4_CLUSTERS_PER_GROUP(sb);
ex.fe_group = TEST_GOAL_GROUP;
+
+ ext4_lock_group(sb, TEST_GOAL_GROUP);
mb_mark_used(&e4b, &ex);
+ ext4_unlock_group(sb, TEST_GOAL_GROUP);
+
grp->bb_free = 0;
memset(bitmap, 0xff, sb->s_blocksize);
--
2.30.0
Powered by blists - more mailing lists