[tls] Guard against a premature server Finished

A malicious server that immediately sends a Finished record (without
ever having sent a ServerHello) will currently cause tls_prf() to get
stuck in an infinite loop attempting to generate pseudorandom data
using the null digest algorithm.

Fix by checking that the key schedule digest size is non-zero
(i.e. that the digest is not the null digest) before attempting to
process the Finished record.

Signed-off-by: Michael Brown <mcb30@ipxe.org>

src/net/tls.c

diff --git a/src/net/tls.c b/src/net/tls.c
index a53f8c8..c5cf92f 100644
--- a/src/net/tls.c
+++ b/src/net/tls.c
@@ -2888,10 +2888,15 @@
 	} __attribute__ (( packed )) *finished = data;
 	uint8_t digest_out[ digest->digestsize ];
 
-	/* Sanity check */
+	/* Sanity checks */
+	if ( ! digest->digestsize ) {
+		DBGC ( tls, "TLS %p received premature Finished\n", tls );
+		DBGC_HDA ( tls, 0, data, len );
+		return -EINVAL_FINISHED;
+	}
 	if ( sizeof ( *finished ) != len ) {
 		DBGC ( tls, "TLS %p received overlength Finished\n", tls );
-		DBGC_HD ( tls, data, len );
+		DBGC_HDA ( tls, 0, data, len );
 		return -EINVAL_FINISHED;
 	}