On Tue, Sep 7, 2021 at 9:33 AM Murphy Zhou <jencce.kernel@xxxxxxxxx> wrote:
>
> Hi,
>
> Since this commit:
>
> commit ec44610fe2b86daef70f3f53f47d2a2542d7094f
> Author: Amir Goldstein <amir73il@xxxxxxxxx>
> Date: Tue Aug 10 18:12:19 2021 +0300
>
> fsnotify: count all objects with attached connectors
>
>
>
>
> Kernel fsnotify can't finish a stress testcase that used to pass quickly.
>
> Kernel hung at umount. Can not be killed but restarting the server.
>
> Reproducer text is attached.
>
Hi Murphy,
Thank you for the detailed report.
I was able to reproduce the hang and the attached patch fixes it for me.
Cloud you please verify the fix yourself as well?
This is a good regression test.
Did you consider contributing it to LTP?
I think the LTP team could also help converting your reproducer to
an LTP test (CC: Petr).
Thanks,
Amir.
From 14d3c313062dfbc86b3d2c4d7deec56a096432f7 Mon Sep 17 00:00:00 2001
From: Amir Goldstein <amir73il@xxxxxxxxx>
Date: Thu, 9 Sep 2021 13:46:34 +0300
Subject: [PATCH] fsnotify: fix sb_connectors leak
Fix a leak in s_fsnotify_connectors counter in case of a race between
concurrent add of new fsnotify mark to an object.
The task that lost the race fails to drop the counter before freeing
the unused connector.
Fixes: ec44610fe2b8 ("fsnotify: count all objects with attached connectors")
Reported-by: Murphy Zhou <jencce.kernel@xxxxxxxxx>
Link: https://lore.kernel.org/linux-fsdevel/20210907063338.ycaw6wvhzrfsfdlp@xxxxxxxxxxxxxxxxxxxxxxxxx/
Signed-off-by: Amir Goldstein <amir73il@xxxxxxxxx>
---
fs/notify/mark.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/fs/notify/mark.c b/fs/notify/mark.c
index 95006d1d29ab..fa1d99101f89 100644
--- a/fs/notify/mark.c
+++ b/fs/notify/mark.c
@@ -531,6 +531,7 @@ static int fsnotify_attach_connector_to_object(fsnotify_connp_t *connp,
/* Someone else created list structure for us */
if (inode)
fsnotify_put_inode_ref(inode);
+ fsnotify_put_sb_connectors(conn);
kmem_cache_free(fsnotify_mark_connector_cachep, conn);
}
--
2.25.1
