[PATCH v4] flow_dissector: check device type before reading ETH_ADDRS

[Yun Zhou <yun.zhou@windriver.com>]

__skb_flow_dissect() unconditionally reads 12 bytes from eth_hdr(skb)
when FLOW_DISSECTOR_KEY_ETH_ADDRS is requested. This assumes the skb
has a valid Ethernet header at mac_header, which is not always the case.

The problem can be triggered by:
 1. Creating a TUN device in L3 mode (IFF_TUN, hard_header_len=0)
 2. Attaching a multiq qdisc with a flower filter matching on eth_src
 3. Sending a packet through AF_PACKET

Since TUN in L3 mode has no link-layer header, mac_header points to
the L3 data area. The flow dissector reads 12 bytes of uninitialized
skb memory, which then propagates through fl_set_masked_key() and is
used as a rhashtable lookup key in __fl_lookup(), as reported by KMSAN.

Rejecting the filter in the control path (at tc filter add time) is
not feasible because TC filter blocks can be shared between arbitrary
devices -- a filter installed on an Ethernet device may later classify
packets on a headerless device through a shared block. The device
association is not fixed at filter creation time.

Fix this by gating the memcpy on dev->type == ARPHRD_ETHER, which
ensures only true Ethernet-framed packets have their addresses read.
This is more precise than the previous hard_header_len >= 12 check,
which would incorrectly pass for non-Ethernet link types like IPoIB
(ARPHRD_INFINIBAND, hard_header_len=24) and FDDI (hard_header_len=21)
whose L2 headers are not in Ethernet format. Additionally check
skb_mac_header_was_set() to guard against the pathological case where
mac_header is the unset sentinel (~0U), which would cause eth_hdr() to
return a wild pointer.

For the act_mirred redirect case (Ethernet packet redirected to a
non-Ethernet device sharing a TC block), zeroing the key is the correct
behavior: the packet is now being classified on the target device, where
Ethernet address matching is not semantically meaningful.

Note: on non-Ethernet devices, the zeroed key will match a filter
configured with all-zero MAC addresses. This is an improvement over the
previous behavior where uninitialized memory could randomly match any
filter.

Reported-by: syzbot+fa2f5b1fb06147be5e16@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=fa2f5b1fb06147be5e16
Fixes: 67a900cc0436 ("flow_dissector: introduce support for Ethernet addresses")
Signed-off-by: Yun Zhou <yun.zhou@windriver.com>
---
v4:
 - Use dev->type == ARPHRD_ETHER instead of hard_header_len >= 12 to
   avoid false positives on non-Ethernet link types (IPoIB, FDDI)
 - Add skb_mac_header_was_set() guard against unset mac_header sentinel
 - Document act_mirred and all-zero key edge cases in commit message

v3:
 - Replace skb_tail_pointer() - skb_mac_header() length check with
    skb->dev->hard_header_len check.

v2:
 - Adjust commit message and comment.

 net/core/flow_dissector.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c
index 2a98f5fa74eb..8aa4f9b4df81 100644
--- a/net/core/flow_dissector.c
+++ b/net/core/flow_dissector.c
@@ -1173,13 +1173,21 @@ bool __skb_flow_dissect(const struct net *net,
 
 	if (dissector_uses_key(flow_dissector,
 			       FLOW_DISSECTOR_KEY_ETH_ADDRS)) {
-		struct ethhdr *eth = eth_hdr(skb);
 		struct flow_dissector_key_eth_addrs *key_eth_addrs;
 
 		key_eth_addrs = skb_flow_dissector_target(flow_dissector,
 							  FLOW_DISSECTOR_KEY_ETH_ADDRS,
 							  target_container);
-		memcpy(key_eth_addrs, eth, sizeof(*key_eth_addrs));
+		/* TC filter blocks can be shared across devices with
+		 * different link types, so we cannot validate this
+		 * when the filter is installed -- check at dissect time.
+		 */
+		if (skb && skb->dev &&
+		    skb->dev->type == ARPHRD_ETHER &&
+		    skb_mac_header_was_set(skb))
+			memcpy(key_eth_addrs, eth_hdr(skb), sizeof(*key_eth_addrs));
+		else
+			memset(key_eth_addrs, 0, sizeof(*key_eth_addrs));
 	}
 
 	if (dissector_uses_key(flow_dissector,
-- 
2.43.0