From 2fe209d0ad2e2729f7e22b9b31a86cc3ff0db550 Mon Sep 17 00:00:00 2001 From: Casey Schaufler Date: Wed, 5 Jun 2024 15:41:50 -0700 Subject: smack: tcp: ipv4, fix incorrect labeling Currently, Smack mirrors the label of incoming tcp/ipv4 connections: when a label 'foo' connects to a label 'bar' with tcp/ipv4, 'foo' always gets 'foo' in returned ipv4 packets. So, 1) returned packets are incorrectly labeled ('foo' instead of 'bar') 2) 'bar' can write to 'foo' without being authorized to write. Here is a scenario how to see this: * Take two machines, let's call them C and S, with active Smack in the default state (no settings, no rules, no labeled hosts, only builtin labels) * At S, add Smack rule 'foo bar w' (labels 'foo' and 'bar' are instantiated at S at this moment) * At S, at label 'bar', launch a program that listens for incoming tcp/ipv4 connections * From C, at label 'foo', connect to the listener at S. (label 'foo' is instantiated at C at this moment) Connection succeedes and works. * Send some data in both directions. * Collect network traffic of this connection. All packets in both directions are labeled with the CIPSO of the label 'foo'. Hence, label 'bar' writes to 'foo' without being authorized, and even without ever being known at C. If anybody cares: exactly the same happens with DCCP. This behavior 1st manifested in release 2.6.29.4 (see Fixes below) and it looks unintentional. At least, no explanation was provided. I changed returned packes label into the 'bar', to bring it into line with the Smack documentation claims. Signed-off-by: Konstantin Andreev Signed-off-by: Casey Schaufler --- security/smack/smack_lsm.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/security/smack/smack_lsm.c b/security/smack/smack_lsm.c index 70ba2841e181..56e02cc5c44d 100644 --- a/security/smack/smack_lsm.c +++ b/security/smack/smack_lsm.c @@ -4431,7 +4431,7 @@ static int smack_inet_conn_request(const struct sock *sk, struct sk_buff *skb, rcu_read_unlock(); if (hskp == NULL) - rc = netlbl_req_setattr(req, &skp->smk_netlabel); + rc = netlbl_req_setattr(req, &ssp->smk_out->smk_netlabel); else netlbl_req_delattr(req); -- cgit v1.2.3-58-ga151 From e86cac0acdb1a74f608bacefe702f2034133a047 Mon Sep 17 00:00:00 2001 From: Konstantin Andreev Date: Mon, 17 Jun 2024 01:44:30 +0300 Subject: smack: unix sockets: fix accept()ed socket label When a process accept()s connection from a unix socket (either stream or seqpacket) it gets the socket with the label of the connecting process. For example, if a connecting process has a label 'foo', the accept()ed socket will also have 'in' and 'out' labels 'foo', regardless of the label of the listener process. This is because kernel creates unix child sockets in the context of the connecting process. I do not see any obvious way for the listener to abuse alien labels coming with the new socket, but, to be on the safe side, it's better fix new socket labels. Signed-off-by: Konstantin Andreev Signed-off-by: Casey Schaufler --- security/smack/smack_lsm.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) diff --git a/security/smack/smack_lsm.c b/security/smack/smack_lsm.c index 56e02cc5c44d..d0d484c1599a 100644 --- a/security/smack/smack_lsm.c +++ b/security/smack/smack_lsm.c @@ -3846,12 +3846,18 @@ static int smack_unix_stream_connect(struct sock *sock, } } - /* - * Cross reference the peer labels for SO_PEERSEC. - */ if (rc == 0) { + /* + * Cross reference the peer labels for SO_PEERSEC. + */ nsp->smk_packet = ssp->smk_out; ssp->smk_packet = osp->smk_out; + + /* + * new/child/established socket must inherit listening socket labels + */ + nsp->smk_out = osp->smk_out; + nsp->smk_in = osp->smk_in; } return rc; -- cgit v1.2.3-58-ga151