linux-kernel - [PATCH] fs: add memory barrier in __fget

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 31 Oct 2022 18:13:07 +0100
From:   Jann Horn <jannh@...gle.com>
To:     Al Viro <viro@...iv.linux.org.uk>
Cc:     Linus Torvalds <torvalds@...ux-foundation.org>,
        Miklos Szeredi <mszeredi@...hat.com>,
        linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org,
        Will Deacon <will@...nel.org>
Subject: [PATCH] fs: add memory barrier in __fget_light()

We must prevent the CPU from reordering the files->count read with the
FD table access like this, on architectures where read-read reordering is
possible:

    files_lookup_fd_raw()
                                  close_fd()
                                  put_files_struct()
    atomic_read(&files->count)

I would like to mark this for stable, but the stable rules explicitly say
"no theoretical races", and given that the FD table pointer and
files->count are explicitly stored in the same cacheline, this sort of
reordering seems quite unlikely in practice...

If this is too expensive on platforms like arm64, I guess the more
performant alternative would be to add another flags field that tracks
whether the fs_struct was ever shared and check that instead of the
reference count in __fget_light().

Signed-off-by: Jann Horn <jannh@...gle.com>
---
 fs/file.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/fs/file.c b/fs/file.c
index 5f9c802a5d8d3..6144287ddc0fe 100644
--- a/fs/file.c
+++ b/fs/file.c
@@ -1004,6 +1004,18 @@ static unsigned long __fget_light(unsigned int fd, fmode_t mask)
 	struct file *file;
 
 	if (atomic_read(&files->count) == 1) {
+		/*
+		 * If another thread is concurrently calling close_fd() followed
+		 * by put_files_struct(), we must not observe the old table
+		 * entry combined with the new refcount - otherwise we could
+		 * return a file that is concurrently being freed.
+		 *
+		 * Pairs with atomic_dec_and_test() in put_files_struct().
+		 * An alternative to using a barrier here would be to use a
+		 * separate field in files_struct to track whether it was ever
+		 * shared.
+		 */
+		smp_rmb();
 		file = files_lookup_fd_raw(files, fd);
 		if (!file || unlikely(file->f_mode & mask))
 			return 0;

base-commit: 30a0b95b1335e12efef89dd78518ed3e4a71a763
-- 
2.38.1.273.g43a17bfeac-goog