netdev - Re: [PATCH v2.51 1/5] ofp-actions: Allow Consistency checking of OF1.3+ VLAN actions after mpls

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: <20131204212429.GA11354@nicira.com>
Date:	Wed, 4 Dec 2013 13:24:29 -0800
From:	Ben Pfaff <blp@...ira.com>
To:	Simon Horman <horms@...ge.net.au>
Cc:	dev@...nvswitch.org, netdev@...r.kernel.org,
	Jesse Gross <jesse@...ira.com>,
	Pravin B Shelar <pshelar@...ira.com>,
	Ravi K <rkerur@...il.com>, Joe Stringer <joe@...d.net.nz>
Subject: Re: [PATCH v2.51 1/5] ofp-actions: Allow Consistency checking of
 OF1.3+ VLAN actions after mpls_push

On Thu, Nov 21, 2013 at 12:46:42PM +0900, Simon Horman wrote:
> The aim of this patch is to support provide infrastructure for verification
> of VLAN actions after an mpls_push action for OpenFlow1.3. This supplements
> existing support for verifying these actions for pre-OpenFlow1.3.
> 
> In OpenFlow1.1 and 1.2 MPLS tags are pushed after any VLAN tags that
> immediately follow the ethernet header. This is pre-OpenFlow1.3 tag
> ordering. Open vSwitch also uses this ordering when supporting MPLS
> actions via Nicira extensions to OpenFlow1.0.
> 
> When using pre-OpenFlow1.3 tag ordering an MPLS push action does not
> affect the VLANs of a packet. If VLAN tags are present immediately after
> the ethernet header then they remain present there.
> 
> In of OpenFlow1.3+ MPLS LSEs are pushed before any VLAN tags that
> immediately follow the ethernet header. This is OpenFlow1.3+ tag
> ordering.
> 
> When using OpenFlow1.3+ tag ordering an MPLS push action affects the
> VLANs of a packet as any VLAN tags previously present after the ethernet
> header are moved to be immediately after the newly pushed MPLS LSE. Thus
> for the purpose of action consistency checking a packet may be changed
> from a VLAN packet to a non-VLAN packet.
> 
> In this way the effective value of the VLAN TCI of a packet may differ
> after an MPLS push depending on the OpenFlow version in use.
> 
> This patch does not enable the logic described above.
> Rather it is disabled in ofpacts_check__(). It should
> be enabled when support for OpenFlow1.3+ tag order is added
> and enabled.
> 
> Signed-off-by: Simon Horman <horms@...ge.net.au>

As far as I can tell this doesn't make sense, because where the MPLS
tag goes is a property of the action that we know at the time we parse
the push_mpls action.  So why isn't this patch just the following?

diff --git a/lib/ofp-actions.c b/lib/ofp-actions.c
index a02f842..f444374 100644
--- a/lib/ofp-actions.c
+++ b/lib/ofp-actions.c
@@ -2071,6 +2071,9 @@ ofpact_check__(enum ofputil_protocol *usable_protocols, struct ofpact *a,
          * Thus nothing can be assumed about the network protocol.
          * Temporarily mark that we have no nw_proto. */
         flow->nw_proto = 0;
+        if (ofpact_get_PUSH_MPLS(a)->position == OFPACT_MPLS_BEFORE_VLAN) {
+            flow->vlan_tci = 0;
+        }
         return 0;
 
     case OFPACT_POP_MPLS:
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html