netdev - [PATCH net-next 2/3] net: mscc: ocelot: refine the ocelot_ace_is_problematic_mac

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

Message-Id: <20200420162743.15847-3-olteanv@gmail.com>
Date:   Mon, 20 Apr 2020 19:27:42 +0300
From:   Vladimir Oltean <olteanv@...il.com>
To:     davem@...emloft.net, netdev@...r.kernel.org
Cc:     idosch@...sch.org, allan.nielsen@...rochip.com,
        horatiu.vultur@...rochip.com, alexandre.belloni@...tlin.com,
        antoine.tenart@...tlin.com, andrew@...n.ch, f.fainelli@...il.com,
        vivien.didelot@...il.com, joergen.andreasen@...rochip.com,
        claudiu.manoil@....com, UNGLinuxDriver@...rochip.com,
        alexandru.marginean@....com, xiaoliang.yang_1@....com,
        yangbo.lu@....com, po.liu@....com, jiri@...lanox.com,
        kuba@...nel.org
Subject: [PATCH net-next 2/3] net: mscc: ocelot: refine the ocelot_ace_is_problematic_mac_etype function

From: Vladimir Oltean <vladimir.oltean@....com>

The commit mentioned below was a bit too harsh, and while it restricted
the invalid key combinations which are known to not work, such as:

tc filter add dev swp0 ingress proto ip \
      flower src_ip 192.0.2.1 action drop
tc filter add dev swp0 ingress proto all \
      flower src_mac 00:11:22:33:44:55 action drop

it also restricted some which still should work, such as:

tc filter add dev swp0 ingress proto ip \
      flower src_ip 192.0.2.1 action drop
tc filter add dev swp0 ingress proto 0x22f0 \
      flower src_mac 00:11:22:33:44:55 action drop

What actually does not match "sanely" is a MAC_ETYPE rule on frames
having an EtherType of ARP, IPv4, IPv6, in addition to SNAP and OAM
frames (which the ocelot tc-flower implementation does not parse yet, so
the function might need to be revisited again in the future).

So just make the function recognize the problematic MAC_ETYPE rules by
EtherType - thus the VCAP IS2 can be forced to match even on those
packets.

This patch makes it possible for IP rules to live on a port together
with MAC_ETYPE rules that are non-all, non-arp, non-ip and non-ipv6.

Fixes: d4d0cb741d7b ("net: mscc: ocelot: deal with problematic MAC_ETYPE VCAP IS2 rules")
Reported-by: Allan W. Nielsen <allan.nielsen@...rochip.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@....com>
---
 drivers/net/ethernet/mscc/ocelot_ace.c | 18 ++++++++++++++----
 1 file changed, 14 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/mscc/ocelot_ace.c b/drivers/net/ethernet/mscc/ocelot_ace.c
index 8a2f7d13ef6d..dfd82a3baab2 100644
--- a/drivers/net/ethernet/mscc/ocelot_ace.c
+++ b/drivers/net/ethernet/mscc/ocelot_ace.c
@@ -739,14 +739,24 @@ static void ocelot_match_all_as_mac_etype(struct ocelot *ocelot, int port,
 
 static bool ocelot_ace_is_problematic_mac_etype(struct ocelot_ace_rule *ace)
 {
+	u16 proto, mask;
+
 	if (ace->type != OCELOT_ACE_TYPE_ETYPE)
 		return false;
-	if (ether_addr_to_u64(ace->frame.etype.dmac.value) &
-	    ether_addr_to_u64(ace->frame.etype.dmac.mask))
+
+	proto = ntohs(*(u16 *)ace->frame.etype.etype.value);
+	mask = ntohs(*(u16 *)ace->frame.etype.etype.mask);
+
+	/* ETH_P_ALL match, so all protocols below are included */
+	if (mask == 0)
 		return true;
-	if (ether_addr_to_u64(ace->frame.etype.smac.value) &
-	    ether_addr_to_u64(ace->frame.etype.smac.mask))
+	if (proto == ETH_P_ARP)
 		return true;
+	if (proto == ETH_P_IP)
+		return true;
+	if (proto == ETH_P_IPV6)
+		return true;
+
 	return false;
 }
 
-- 
2.17.1