[PATCH] cipso: handle CIPSO options correctly when NetLabel is disabled

When NetLabel is not enabled, e.g. CONFIG_NETLABEL=n, and the system
receives a CIPSO tagged packet it is dropped (cipso_v4_validate()
returns non-zero).  In most cases this is the correct and desired
behavior, however, in the case where we are simply forwarding the
traffic, e.g. acting as a network bridge, this becomes a problem.

This patch fixes the forwarding problem by providing the basic CIPSO
validation code directly in ip_options_compile() without the need for
the NetLabel or CIPSO code.  The new validation code can not perform
any of the CIPSO option label/value verification that
cipso_v4_validate() does, but it can verify the basic CIPSO option
format.

The behavior when NetLabel is enabled is unchanged.

Signed-off-by: Paul Moore <redacted>
---
 net/ipv4/ip_options.c |   20 ++++++++++++++++++++
 1 files changed, 20 insertions(+), 0 deletions(-)

diff --git a/net/ipv4/ip_options.c b/net/ipv4/ip_options.c
index 708b994..ca2c919 100644
--- a/net/ipv4/ip_options.c
+++ b/net/ipv4/ip_options.c

@@ -439,10 +439,30 @@ int ip_options_compile(struct net *net,
 				goto error;
 			}
 			opt->cipso = optptr - iph;
+#ifndef CONFIG_NETLABEL
+			if (optlen < 8) {
+				pp_ptr = optptr + 1;
+				goto error;
+			}
+			if (get_unaligned_be32(&optptr[2]) != 0) {
+				unsigned int iter;
+				for (iter = 6; iter < optlen;) {
+					if (optptr[iter+1] > (optlen - iter)) {
+						pp_ptr = optptr + iter;
+						goto error;
+					}
+					iter += optptr[iter + 1];
+				}
+			} else {
+				pp_ptr = optptr + 2;
+				goto error;
+			}
+#else
 			if (cipso_v4_validate(skb, &optptr)) {
 				pp_ptr = optptr;
 				goto error;
 			}
+#endif /* CONFIG_NETLABEL */
 			break;
 		      case IPOPT_SEC:
 		      case IPOPT_SID:

From: Paul Moore <redacted>
Date: Thu, 31 May 2012 16:09:23 -0400

When NetLabel is not enabled, e.g. CONFIG_NETLABEL=n, and the system
receives a CIPSO tagged packet it is dropped (cipso_v4_validate()
returns non-zero).  In most cases this is the correct and desired
behavior, however, in the case where we are simply forwarding the
traffic, e.g. acting as a network bridge, this becomes a problem.

This patch fixes the forwarding problem by providing the basic CIPSO
validation code directly in ip_options_compile() without the need for
the NetLabel or CIPSO code.  The new validation code can not perform
any of the CIPSO option label/value verification that
cipso_v4_validate() does, but it can verify the basic CIPSO option
format.

The behavior when NetLabel is enabled is unchanged.

Signed-off-by: Paul Moore <redacted>

I don't like this at all.

The only conclusion I can come to is that cipso_v4_validate() is doing
the wrong thing when NETLABEL is disabled.

There is never a good reason to crap all over a function with ifdefs.
This is especially true when it's being done to paper over a function
with poor semantics.

The whole idea is to abstract and put all of this kind of logic into
cipso_v4_validate().

On Thursday, May 31, 2012 07:07:05 PM David Miller wrote:

From: Paul Moore <redacted>
Date: Thu, 31 May 2012 16:09:23 -0400

quoted

When NetLabel is not enabled, e.g. CONFIG_NETLABEL=n, and the system
receives a CIPSO tagged packet it is dropped (cipso_v4_validate()
returns non-zero).  In most cases this is the correct and desired
behavior, however, in the case where we are simply forwarding the
traffic, e.g. acting as a network bridge, this becomes a problem.

This patch fixes the forwarding problem by providing the basic CIPSO
validation code directly in ip_options_compile() without the need for
the NetLabel or CIPSO code.  The new validation code can not perform
any of the CIPSO option label/value verification that
cipso_v4_validate() does, but it can verify the basic CIPSO option
format.

The behavior when NetLabel is enabled is unchanged.

Signed-off-by: Paul Moore <redacted>

I don't like this at all.

The only conclusion I can come to is that cipso_v4_validate() is doing
the wrong thing when NETLABEL is disabled.

There is never a good reason to crap all over a function with ifdefs.
This is especially true when it's being done to paper over a function
with poor semantics.

The whole idea is to abstract and put all of this kind of logic into
cipso_v4_validate().

I originally had the #ifdef'd code in the non-CONFIG_NETLABEL 
cipso_v4_validate() in include/net/cipso_ipv4.h but thought it was too much 
code to put there.  No worries, I'll just move it back and resubmit.

-- 
paul moore
security and virtualization @ redhat