s4-dcerpc: fixed auth padding to be relative to the stub, not packet

The recent dcerpc padding changes made our padding relative to the
packet header, instead of the start of the stub. Surprisingly, this
broke w2k8r2 doing a dcpromo join to a s4 server. It seems that w2k8r2
is very fussy about the padding it gets in some circumstances.

1 files changed, 7 insertions, 5 deletions

diff --git a/source4/librpc/rpc/dcerpc.c b/source4/librpc/rpc/dcerpc.c
index ccafe070ab..63ff36768b 100644
--- a/source4/librpc/rpc/dcerpc.c
+++ b/source4/librpc/rpc/dcerpc.c
@@ -318,7 +318,6 @@ static NTSTATUS ncacn_push_request_sign(struct dcerpc_connection *c,
 	size_t payload_length;
 	enum ndr_err_code ndr_err;
 	size_t hdr_size = DCERPC_REQUEST_LENGTH;
-	uint32_t offset;
 
 	/* non-signed packets are simpler */
 	if (sig_size == 0) {
@@ -365,13 +364,16 @@ static NTSTATUS ncacn_push_request_sign(struct dcerpc_connection *c,
 	}
 
 	/* pad to 16 byte multiple in the payload portion of the
-	   packet. This matches what w2k3 does */
-	offset = ndr->offset;
-	ndr_err = ndr_push_align(ndr, 16);
+	   packet. This matches what w2k3 does. Note that we can't use
+	   ndr_push_align() as that is relative to the start of the
+	   whole packet, whereas w2k8 wants it relative to the start
+	   of the stub */
+	c->security_state.auth_info->auth_pad_length =
+		(16 - (pkt->u.request.stub_and_verifier.length & 15)) & 15;
+	ndr_err = ndr_push_zero(ndr, c->security_state.auth_info->auth_pad_length);
 	if (!NDR_ERR_CODE_IS_SUCCESS(ndr_err)) {
 		return ndr_map_error2ntstatus(ndr_err);
 	}
-	c->security_state.auth_info->auth_pad_length = ndr->offset - offset;
 
 	payload_length = pkt->u.request.stub_and_verifier.length + 
 		c->security_state.auth_info->auth_pad_length;