tcp: add tcp_v4_fill_cb()/tcp_v4_restore_cb()
authorEric Dumazet <edumazet@google.com>
Sun, 3 Dec 2017 17:32:59 +0000 (09:32 -0800)
committerDavid S. Miller <davem@davemloft.net>
Sun, 3 Dec 2017 17:39:14 +0000 (12:39 -0500)
James Morris reported kernel stack corruption bug [1] while
running the SELinux testsuite, and bisected to a recent
commit bffa72cf7f9d ("net: sk_buff rbnode reorg")

We believe this commit is fine, but exposes an older bug.

SELinux code runs from tcp_filter() and might send an ICMP,
expecting IP options to be found in skb->cb[] using regular IPCB placement.

We need to defer TCP mangling of skb->cb[] after tcp_filter() calls.

This patch adds tcp_v4_fill_cb()/tcp_v4_restore_cb() in a very
similar way we added them for IPv6.

[1]
[  339.806024] SELinux: failure in selinux_parse_skb(), unable to parse packet
[  339.822505] Kernel panic - not syncing: stack-protector: Kernel stack is corrupted in: ffffffff81745af5
[  339.822505]
[  339.852250] CPU: 4 PID: 3642 Comm: client Not tainted 4.15.0-rc1-test #15
[  339.868498] Hardware name: LENOVO 10FGS0VA1L/30BC, BIOS FWKT68A   01/19/2017
[  339.885060] Call Trace:
[  339.896875]  <IRQ>
[  339.908103]  dump_stack+0x63/0x87
[  339.920645]  panic+0xe8/0x248
[  339.932668]  ? ip_push_pending_frames+0x33/0x40
[  339.946328]  ? icmp_send+0x525/0x530
[  339.958861]  ? kfree_skbmem+0x60/0x70
[  339.971431]  __stack_chk_fail+0x1b/0x20
[  339.984049]  icmp_send+0x525/0x530
[  339.996205]  ? netlbl_skbuff_err+0x36/0x40
[  340.008997]  ? selinux_netlbl_err+0x11/0x20
[  340.021816]  ? selinux_socket_sock_rcv_skb+0x211/0x230
[  340.035529]  ? security_sock_rcv_skb+0x3b/0x50
[  340.048471]  ? sk_filter_trim_cap+0x44/0x1c0
[  340.061246]  ? tcp_v4_inbound_md5_hash+0x69/0x1b0
[  340.074562]  ? tcp_filter+0x2c/0x40
[  340.086400]  ? tcp_v4_rcv+0x820/0xa20
[  340.098329]  ? ip_local_deliver_finish+0x71/0x1a0
[  340.111279]  ? ip_local_deliver+0x6f/0xe0
[  340.123535]  ? ip_rcv_finish+0x3a0/0x3a0
[  340.135523]  ? ip_rcv_finish+0xdb/0x3a0
[  340.147442]  ? ip_rcv+0x27c/0x3c0
[  340.158668]  ? inet_del_offload+0x40/0x40
[  340.170580]  ? __netif_receive_skb_core+0x4ac/0x900
[  340.183285]  ? rcu_accelerate_cbs+0x5b/0x80
[  340.195282]  ? __netif_receive_skb+0x18/0x60
[  340.207288]  ? process_backlog+0x95/0x140
[  340.218948]  ? net_rx_action+0x26c/0x3b0
[  340.230416]  ? __do_softirq+0xc9/0x26a
[  340.241625]  ? do_softirq_own_stack+0x2a/0x40
[  340.253368]  </IRQ>
[  340.262673]  ? do_softirq+0x50/0x60
[  340.273450]  ? __local_bh_enable_ip+0x57/0x60
[  340.285045]  ? ip_finish_output2+0x175/0x350
[  340.296403]  ? ip_finish_output+0x127/0x1d0
[  340.307665]  ? nf_hook_slow+0x3c/0xb0
[  340.318230]  ? ip_output+0x72/0xe0
[  340.328524]  ? ip_fragment.constprop.54+0x80/0x80
[  340.340070]  ? ip_local_out+0x35/0x40
[  340.350497]  ? ip_queue_xmit+0x15c/0x3f0
[  340.361060]  ? __kmalloc_reserve.isra.40+0x31/0x90
[  340.372484]  ? __skb_clone+0x2e/0x130
[  340.382633]  ? tcp_transmit_skb+0x558/0xa10
[  340.393262]  ? tcp_connect+0x938/0xad0
[  340.403370]  ? ktime_get_with_offset+0x4c/0xb0
[  340.414206]  ? tcp_v4_connect+0x457/0x4e0
[  340.424471]  ? __inet_stream_connect+0xb3/0x300
[  340.435195]  ? inet_stream_connect+0x3b/0x60
[  340.445607]  ? SYSC_connect+0xd9/0x110
[  340.455455]  ? __audit_syscall_entry+0xaf/0x100
[  340.466112]  ? syscall_trace_enter+0x1d0/0x2b0
[  340.476636]  ? __audit_syscall_exit+0x209/0x290
[  340.487151]  ? SyS_connect+0xe/0x10
[  340.496453]  ? do_syscall_64+0x67/0x1b0
[  340.506078]  ? entry_SYSCALL64_slow_path+0x25/0x25

Fixes: 971f10eca186 ("tcp: better TCP_SKB_CB layout to reduce cache line misses")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: James Morris <james.l.morris@oracle.com>
Tested-by: James Morris <james.l.morris@oracle.com>
Tested-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv4/tcp_ipv4.c
net/ipv6/tcp_ipv6.c

index c6bc0c4d19c624888b0d0b5a4246c7183edf63f5..77ea45da0fe9c746907a312989658af3ad3b198d 100644 (file)
@@ -1591,6 +1591,34 @@ int tcp_filter(struct sock *sk, struct sk_buff *skb)
 }
 EXPORT_SYMBOL(tcp_filter);
 
+static void tcp_v4_restore_cb(struct sk_buff *skb)
+{
+       memmove(IPCB(skb), &TCP_SKB_CB(skb)->header.h4,
+               sizeof(struct inet_skb_parm));
+}
+
+static void tcp_v4_fill_cb(struct sk_buff *skb, const struct iphdr *iph,
+                          const struct tcphdr *th)
+{
+       /* This is tricky : We move IPCB at its correct location into TCP_SKB_CB()
+        * barrier() makes sure compiler wont play fool^Waliasing games.
+        */
+       memmove(&TCP_SKB_CB(skb)->header.h4, IPCB(skb),
+               sizeof(struct inet_skb_parm));
+       barrier();
+
+       TCP_SKB_CB(skb)->seq = ntohl(th->seq);
+       TCP_SKB_CB(skb)->end_seq = (TCP_SKB_CB(skb)->seq + th->syn + th->fin +
+                                   skb->len - th->doff * 4);
+       TCP_SKB_CB(skb)->ack_seq = ntohl(th->ack_seq);
+       TCP_SKB_CB(skb)->tcp_flags = tcp_flag_byte(th);
+       TCP_SKB_CB(skb)->tcp_tw_isn = 0;
+       TCP_SKB_CB(skb)->ip_dsfield = ipv4_get_dsfield(iph);
+       TCP_SKB_CB(skb)->sacked  = 0;
+       TCP_SKB_CB(skb)->has_rxtstamp =
+                       skb->tstamp || skb_hwtstamps(skb)->hwtstamp;
+}
+
 /*
  *     From tcp_input.c
  */
@@ -1631,24 +1659,6 @@ int tcp_v4_rcv(struct sk_buff *skb)
 
        th = (const struct tcphdr *)skb->data;
        iph = ip_hdr(skb);
-       /* This is tricky : We move IPCB at its correct location into TCP_SKB_CB()
-        * barrier() makes sure compiler wont play fool^Waliasing games.
-        */
-       memmove(&TCP_SKB_CB(skb)->header.h4, IPCB(skb),
-               sizeof(struct inet_skb_parm));
-       barrier();
-
-       TCP_SKB_CB(skb)->seq = ntohl(th->seq);
-       TCP_SKB_CB(skb)->end_seq = (TCP_SKB_CB(skb)->seq + th->syn + th->fin +
-                                   skb->len - th->doff * 4);
-       TCP_SKB_CB(skb)->ack_seq = ntohl(th->ack_seq);
-       TCP_SKB_CB(skb)->tcp_flags = tcp_flag_byte(th);
-       TCP_SKB_CB(skb)->tcp_tw_isn = 0;
-       TCP_SKB_CB(skb)->ip_dsfield = ipv4_get_dsfield(iph);
-       TCP_SKB_CB(skb)->sacked  = 0;
-       TCP_SKB_CB(skb)->has_rxtstamp =
-                       skb->tstamp || skb_hwtstamps(skb)->hwtstamp;
-
 lookup:
        sk = __inet_lookup_skb(&tcp_hashinfo, skb, __tcp_hdrlen(th), th->source,
                               th->dest, sdif, &refcounted);
@@ -1679,14 +1689,19 @@ process:
                sock_hold(sk);
                refcounted = true;
                nsk = NULL;
-               if (!tcp_filter(sk, skb))
+               if (!tcp_filter(sk, skb)) {
+                       th = (const struct tcphdr *)skb->data;
+                       iph = ip_hdr(skb);
+                       tcp_v4_fill_cb(skb, iph, th);
                        nsk = tcp_check_req(sk, skb, req, false);
+               }
                if (!nsk) {
                        reqsk_put(req);
                        goto discard_and_relse;
                }
                if (nsk == sk) {
                        reqsk_put(req);
+                       tcp_v4_restore_cb(skb);
                } else if (tcp_child_process(sk, nsk, skb)) {
                        tcp_v4_send_reset(nsk, skb);
                        goto discard_and_relse;
@@ -1712,6 +1727,7 @@ process:
                goto discard_and_relse;
        th = (const struct tcphdr *)skb->data;
        iph = ip_hdr(skb);
+       tcp_v4_fill_cb(skb, iph, th);
 
        skb->dev = NULL;
 
@@ -1742,6 +1758,8 @@ no_tcp_socket:
        if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb))
                goto discard_it;
 
+       tcp_v4_fill_cb(skb, iph, th);
+
        if (tcp_checksum_complete(skb)) {
 csum_error:
                __TCP_INC_STATS(net, TCP_MIB_CSUMERRORS);
@@ -1768,6 +1786,8 @@ do_time_wait:
                goto discard_it;
        }
 
+       tcp_v4_fill_cb(skb, iph, th);
+
        if (tcp_checksum_complete(skb)) {
                inet_twsk_put(inet_twsk(sk));
                goto csum_error;
@@ -1784,6 +1804,7 @@ do_time_wait:
                if (sk2) {
                        inet_twsk_deschedule_put(inet_twsk(sk));
                        sk = sk2;
+                       tcp_v4_restore_cb(skb);
                        refcounted = false;
                        goto process;
                }
index be11dc13aa705145a83177e17d23594e9416e11a..1f04ec0e4a7aa2c11b8ee27cbdd4067b5bcf32e5 100644 (file)
@@ -1454,7 +1454,6 @@ process:
                struct sock *nsk;
 
                sk = req->rsk_listener;
-               tcp_v6_fill_cb(skb, hdr, th);
                if (tcp_v6_inbound_md5_hash(sk, skb)) {
                        sk_drops_add(sk, skb);
                        reqsk_put(req);
@@ -1467,8 +1466,12 @@ process:
                sock_hold(sk);
                refcounted = true;
                nsk = NULL;
-               if (!tcp_filter(sk, skb))
+               if (!tcp_filter(sk, skb)) {
+                       th = (const struct tcphdr *)skb->data;
+                       hdr = ipv6_hdr(skb);
+                       tcp_v6_fill_cb(skb, hdr, th);
                        nsk = tcp_check_req(sk, skb, req, false);
+               }
                if (!nsk) {
                        reqsk_put(req);
                        goto discard_and_relse;
@@ -1492,8 +1495,6 @@ process:
        if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb))
                goto discard_and_relse;
 
-       tcp_v6_fill_cb(skb, hdr, th);
-
        if (tcp_v6_inbound_md5_hash(sk, skb))
                goto discard_and_relse;
 
@@ -1501,6 +1502,7 @@ process:
                goto discard_and_relse;
        th = (const struct tcphdr *)skb->data;
        hdr = ipv6_hdr(skb);
+       tcp_v6_fill_cb(skb, hdr, th);
 
        skb->dev = NULL;