[PATCH] cachefiles: fix dentry leak in cachefiles_open_file()

libaokun at huaweicloud.com libaokun at huaweicloud.com
Mon Aug 26 14:00:18 AEST 2024


From: Baokun Li <libaokun1 at huawei.com>

In ondemand mode, dentry leaks may be caused when the mount has the
following concurrency with cull:

            P1             |             P2
-----------------------------------------------------------
cachefiles_lookup_cookie
  cachefiles_look_up_object
    lookup_one_positive_unlocked
     // get dentry
                            cachefiles_cull
                              inode->i_flags |= S_KERNEL_FILE;
    cachefiles_open_file
      cachefiles_mark_inode_in_use
        __cachefiles_mark_inode_in_use
          can_use = false
          if (!(inode->i_flags & S_KERNEL_FILE))
            can_use = true
	  return false
        return false
        // Returns an error but doesn't put dentry

After that the following WARNING will be triggered when the backend folder
is umounted:

==================================================================
BUG: Dentry 000000008ad87947{i=7a,n=Dx_1_1.img}  still in use (1) [unmount of ext4 sda]
WARNING: CPU: 4 PID: 359261 at fs/dcache.c:1767 umount_check+0x5d/0x70
CPU: 4 PID: 359261 Comm: umount Not tainted 6.6.0-dirty #25
RIP: 0010:umount_check+0x5d/0x70
Call Trace:
 <TASK>
 d_walk+0xda/0x2b0
 do_one_tree+0x20/0x40
 shrink_dcache_for_umount+0x2c/0x90
 generic_shutdown_super+0x20/0x160
 kill_block_super+0x1a/0x40
 ext4_kill_sb+0x22/0x40
 deactivate_locked_super+0x35/0x80
 cleanup_mnt+0x104/0x160
==================================================================

Add the missing dput() to cachefiles_open_file() for a quick fix.

Fixes: 1f08c925e7a3 ("cachefiles: Implement backing file wrangling")
Cc: stable at kernel.org
Signed-off-by: Baokun Li <libaokun1 at huawei.com>
---
 fs/cachefiles/namei.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/fs/cachefiles/namei.c b/fs/cachefiles/namei.c
index f53977169db4..0bd2f367c3ff 100644
--- a/fs/cachefiles/namei.c
+++ b/fs/cachefiles/namei.c
@@ -554,6 +554,7 @@ static bool cachefiles_open_file(struct cachefiles_object *object,
 	if (!cachefiles_mark_inode_in_use(object, d_inode(dentry))) {
 		pr_notice("cachefiles: Inode already in use: %pd (B=%lx)\n",
 			  dentry, d_inode(dentry)->i_ino);
+		dput(dentry);
 		return false;
 	}
 
-- 
2.39.2



More information about the Linux-erofs mailing list