[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <10CFECF7-B9FC-4562-A445-4811F4C27655@gmail.com>
Date: Thu, 10 Aug 2023 20:59:26 -0400
From: Sishuai Gong <sishuai.system@...il.com>
To: rostedt@...dmis.org, mhiramat@...nel.org
Cc: linux-kernel@...r.kernel.org, linux-trace-kernel@...r.kernel.org
Subject: [PATCH] tracefs: avoid setting i_mode to a temp value
Right now inode->i_mode is updated twice to reach the desired value
in tracefs_apply_options(). Because there is no lock protecting the two
writes, other threads might read the intermediate value of inode->i_mode.
Thread-1 Thread-2
// tracefs_apply_options() //e.g., acl_permission_check
inode->i_mode &= ~S_IALLUGO;
unsigned int mode = inode->i_mode;
inode->i_mode |= opts->mode;
I think there is no need to introduce a lock but it is better to
only update inode->i_mode ONCE, so the readers will either see the old
or latest value, rather than an intermediate/temporary value.
Signed-off-by: Sishuai Gong <sishuai.system@...il.com>
---
fs/tracefs/inode.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/fs/tracefs/inode.c b/fs/tracefs/inode.c
index 57ac8aa4a724..dca84ebb62fa 100644
--- a/fs/tracefs/inode.c
+++ b/fs/tracefs/inode.c
@@ -297,8 +297,7 @@ static int tracefs_apply_options(struct super_block *sb, bool remount)
*/
if (!remount || opts->opts & BIT(Opt_mode)) {
- inode->i_mode &= ~S_IALLUGO;
- inode->i_mode |= opts->mode;
+ inode->i_mode = (inode->i_mode & ~S_IALLUGO) | opts->mode;
}
if (!remount || opts->opts & BIT(Opt_uid))
--
2.39.2 (Apple Git-143)
Powered by blists - more mailing lists