netdev - [Linux-kernel-mentees] [PATCH net] rds: Prevent kernel-infoleak in rds_notify_queue

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]

Message-Id: <20200730192026.110246-1-yepeilin.cs@gmail.com>
Date:   Thu, 30 Jul 2020 15:20:26 -0400
From:   Peilin Ye <yepeilin.cs@...il.com>
To:     Santosh Shilimkar <santosh.shilimkar@...cle.com>
Cc:     Peilin Ye <yepeilin.cs@...il.com>,
        "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Dan Carpenter <dan.carpenter@...cle.com>,
        Arnd Bergmann <arnd@...db.de>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        linux-kernel-mentees@...ts.linuxfoundation.org,
        netdev@...r.kernel.org, linux-rdma@...r.kernel.org,
        rds-devel@....oracle.com, linux-kernel@...r.kernel.org
Subject: [Linux-kernel-mentees] [PATCH net] rds: Prevent kernel-infoleak in rds_notify_queue_get()

rds_notify_queue_get() is potentially copying uninitialized kernel stack
memory to userspace since the compiler may leave a 4-byte hole at the end
of `cmsg`.

In 2016 we tried to fix this issue by doing `= { 0 };` on `cmsg`, which
unfortunately does not always initialize that 4-byte hole. Fix it by using
memset() instead.

Cc: stable@...r.kernel.org
Fixes: f037590fff30 ("rds: fix a leak of kernel memory")
Fixes: bdbe6fbc6a2f ("RDS: recv.c")
Suggested-by: Dan Carpenter <dan.carpenter@...cle.com>
Signed-off-by: Peilin Ye <yepeilin.cs@...il.com>
---
Note: the "real" copy_to_user() happens in put_cmsg(), where
`cmlen - sizeof(*cm)` equals to `sizeof(cmsg)`.

Reference: https://lwn.net/Articles/417989/

$ pahole -C "rds_rdma_notify" net/rds/recv.o
struct rds_rdma_notify {
	__u64                      user_token;           /*     0     8 */
	__s32                      status;               /*     8     4 */

	/* size: 16, cachelines: 1, members: 2 */
	/* padding: 4 */
	/* last cacheline: 16 bytes */
};

 net/rds/recv.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/net/rds/recv.c b/net/rds/recv.c
index c8404971d5ab..aba4afe4dfed 100644
--- a/net/rds/recv.c
+++ b/net/rds/recv.c
@@ -450,12 +450,13 @@ static int rds_still_queued(struct rds_sock *rs, struct rds_incoming *inc,
 int rds_notify_queue_get(struct rds_sock *rs, struct msghdr *msghdr)
 {
 	struct rds_notifier *notifier;
-	struct rds_rdma_notify cmsg = { 0 }; /* fill holes with zero */
+	struct rds_rdma_notify cmsg;
 	unsigned int count = 0, max_messages = ~0U;
 	unsigned long flags;
 	LIST_HEAD(copy);
 	int err = 0;
 
+	memset(&cmsg, 0, sizeof(cmsg));	/* fill holes with zero */
 
 	/* put_cmsg copies to user space and thus may sleep. We can't do this
 	 * with rs_lock held, so first grab as many notifications as we can stuff
-- 
2.25.1