add (new) "url" protocol action, this can be used to match/filter URL

suffix/prefix expressions like "example.com/index.html?args". a digest
mode allows to match against anonymized SHA1/MD5 digests of
suffix/prefix expressions.

4 files changed, 162 insertions, 10 deletions

diff --git a/usr.sbin/relayd/parse.y b/usr.sbin/relayd/parse.y
index f8e0829996f..86ebc48c691 100644
--- a/usr.sbin/relayd/parse.y
+++ b/usr.sbin/relayd/parse.y
@@ -1,4 +1,4 @@
-/*	$OpenBSD: parse.y,v 1.92 2007/11/21 20:31:03 reyk Exp $	*/
+/*	$OpenBSD: parse.y,v 1.93 2007/11/22 10:09:53 reyk Exp $	*/
 
 /*
  * Copyright (c) 2006 Pierre-Yves Ritschard <pyr@openbsd.org>
@@ -123,7 +123,7 @@ typedef struct {
 %token  TIMEOUT CODE DIGEST PORT TAG INTERFACE STYLE RETURN
 %token	VIRTUAL INTERVAL DISABLE STICKYADDR BACKLOG PATH SCRIPT
 %token	SEND EXPECT NOTHING SSL LOADBALANCE ROUNDROBIN CIPHERS COOKIE
-%token	RELAY LISTEN ON FORWARD TO NAT LOOKUP PREFORK NO MARK MARKED
+%token	RELAY LISTEN ON FORWARD TO NAT LOOKUP PREFORK NO MARK MARKED URL
 %token	PROTO SESSION CACHE APPEND CHANGE REMOVE FROM FILTER HASH HEADER
 %token	LOG UPDATES ALL DEMOTE NODELAY SACK SOCKET BUFFER QUERYSTR RETRY IP
 %token	ERROR
@@ -728,6 +728,8 @@ protoptsl	: SSL sslflags
 
 			if (node.type == NODE_TYPE_COOKIE)
 				pk.key = "Cookie";
+			else if (node.type == NODE_TYPE_URL)
+				pk.key = "Host";
 			else
 				pk.key = "GET";
 			if (node.type != NODE_TYPE_HEADER) {
@@ -763,6 +765,15 @@ protoptsl	: SSL sslflags
 				case NODE_TYPE_COOKIE:
 					pn->flags |= PNFLAG_LOOKUP_COOKIE;
 					break;
+				case NODE_TYPE_URL:
+					if (node.flags &
+					    PNFLAG_LOOKUP_URL_DIGEST)
+						pn->flags |= node.flags &
+						    PNFLAG_LOOKUP_URL_DIGEST;
+					else
+						pn->flags |=
+						    PNFLAG_LOOKUP_DIGEST(0);
+					break;
 				default:
 					break;
 				}
@@ -905,6 +916,21 @@ protonode	: nodetype APPEND STRING TO STRING marked	{
 			free($3);
 			proto->lateconnect++;
 		}
+		| nodetype EXPECT digest mark {
+			if (node.type != NODE_TYPE_URL) {
+				yyerror("digest not supported for this type");
+				free($3.digest);
+				YYERROR;
+			}
+			node.action = NODE_ACTION_EXPECT;
+			node.key = strdup($3.digest);
+			node.flags |= PNFLAG_LOOKUP_DIGEST($3.type);
+			node.value = strdup("*");
+			if (node.key == NULL || node.value == NULL)
+				fatal("out of memory");
+			free($3.digest);
+			proto->lateconnect++;
+		}
 		| nodetype FILTER STRING FROM STRING mark	{
 			node.action = NODE_ACTION_FILTER;
 			node.key = strdup($5);
@@ -924,6 +950,21 @@ protonode	: nodetype APPEND STRING TO STRING marked	{
 			free($3);
 			proto->lateconnect++;
 		}
+		| nodetype FILTER digest mark {
+			if (node.type != NODE_TYPE_URL) {
+				yyerror("digest not supported for this type");
+				free($3.digest);
+				YYERROR;
+			}
+			node.action = NODE_ACTION_FILTER;
+			node.key = strdup($3.digest);
+			node.flags |= PNFLAG_LOOKUP_DIGEST($3.type);
+			node.value = strdup("*");
+			if (node.key == NULL || node.value == NULL)
+				fatal("out of memory");
+			free($3.digest);
+			proto->lateconnect++;
+		}
 		| nodetype HASH STRING marked			{
 			node.action = NODE_ACTION_HASH;
 			node.key = strdup($3);
@@ -963,6 +1004,7 @@ nodetype	: HEADER			{
 			proto->flags |= F_LOOKUP_PATH;
 			node.type = NODE_TYPE_PATH;
 		}
+		| URL				{ node.type = NODE_TYPE_URL; }
 		;
 
 sslcache	: NUMBER			{
@@ -1349,6 +1391,7 @@ lookup(char *s)
 		{ "timeout",		TIMEOUT },
 		{ "to",			TO },
 		{ "updates",		UPDATES },
+		{ "url",		URL },
 		{ "virtual",		VIRTUAL }
 	};
 	const struct keywords	*p;
diff --git a/usr.sbin/relayd/relay.c b/usr.sbin/relayd/relay.c
index 10c2fbc75f6..466e17323f6 100644
--- a/usr.sbin/relayd/relay.c
+++ b/usr.sbin/relayd/relay.c
@@ -1,4 +1,4 @@
-/*	$OpenBSD: relay.c,v 1.64 2007/11/21 20:41:40 reyk Exp $	*/
+/*	$OpenBSD: relay.c,v 1.65 2007/11/22 10:09:53 reyk Exp $	*/
 
 /*
  * Copyright (c) 2006, 2007 Reyk Floeter <reyk@openbsd.org>
@@ -91,6 +91,8 @@ int		 relay_handle_http(struct ctl_relay_event *,
 		    struct protonode *, struct protonode *,
 		    struct protonode *, int);
 void		 relay_read_http(struct bufferevent *, void *);
+int		 relay_lookup_url(struct ctl_relay_event *,
+		    const char *, enum digest_type);
 int		 relay_lookup_query(struct ctl_relay_event *);
 int		 relay_lookup_cookie(struct ctl_relay_event *, const char *);
 void		 relay_read_httpcontent(struct bufferevent *, void *);
@@ -284,6 +286,7 @@ void
 relay_nodedebug(const char *name, struct protonode *pn)
 {
 	const char	*s;
+	int		 digest;
 
 	if (pn->action == NODE_ACTION_NONE)
 		return;
@@ -303,6 +306,9 @@ relay_nodedebug(const char *name, struct protonode *pn)
 	case NODE_TYPE_PATH:
 		fprintf(stderr, "path ");
 		break;
+	case NODE_TYPE_URL:
+		fprintf(stderr, "url ");
+		break;
 	}
 
 	switch (pn->action) {
@@ -321,8 +327,10 @@ relay_nodedebug(const char *name, struct protonode *pn)
 	case NODE_ACTION_EXPECT:
 	case NODE_ACTION_FILTER:
 		s = pn->action == NODE_ACTION_EXPECT ? "expect" : "filter";
+		digest = pn->flags & PNFLAG_LOOKUP_URL_DIGEST;
 		if (strcmp(pn->value, "*") == 0)
-			fprintf(stderr, "%s \"%s\"", s, pn->key);
+			fprintf(stderr, "%s %s\"%s\"", s,
+			    digest ? "digest " : "", pn->key);
 		else
 			fprintf(stderr, "%s \"%s\" from \"%s\"", s,
 			    pn->value, pn->key);
@@ -1378,7 +1386,24 @@ relay_read_http(struct bufferevent *bev, void *arg)
 		if (cre->dir == RELAY_DIR_RESPONSE)
 			goto handle;
 
-		if (pn->flags & PNFLAG_LOOKUP_QUERY) {
+		if (pn->flags & PNFLAG_LOOKUP_URL) {
+			/*
+			 * Lookup the URL of type example.com/path?args.
+			 * Either as a plain string or SHA1/MD5 digest.
+			 */
+			if ((pn->flags & PNFLAG_LOOKUP_DIGEST(0)) &&
+			    relay_lookup_url(cre, pk.value,
+			    DIGEST_NONE) == PN_FAIL)
+				goto abort;
+			if ((pn->flags & PNFLAG_LOOKUP_DIGEST(DIGEST_SHA1)) &&
+			    relay_lookup_url(cre, pk.value,
+			    DIGEST_SHA1) == PN_FAIL)
+				goto abort;
+			if ((pn->flags & PNFLAG_LOOKUP_DIGEST(DIGEST_MD5)) &&
+			    relay_lookup_url(cre, pk.value,
+			    DIGEST_MD5) == PN_FAIL)
+				goto abort;
+		} else if (pn->flags & PNFLAG_LOOKUP_QUERY) {
 			/* Lookup the HTTP query arguments */
 			if (relay_lookup_query(cre) == PN_FAIL)
 				goto abort;
@@ -1476,6 +1501,60 @@ relay_read_http(struct bufferevent *bev, void *arg)
 }
 
 int
+relay_lookup_url(struct ctl_relay_event *cre, const char *str,
+    enum digest_type type)
+{
+	struct session		*con = (struct session *)cre->con;
+	struct protonode	*proot, *pnv, pkv;
+	char			*val, *md = NULL;
+	int			 ret = PN_FAIL;
+
+	if (cre->path == NULL)
+		return (PN_PASS);
+
+	if (asprintf(&val, "%s%s%s%s",
+	    str, cre->path,
+	    cre->args == NULL ? "" : "?",
+	    cre->args == NULL ? "" : cre->args) == -1) {
+		relay_close_http(con, 500, "failed to allocate URL");
+		return (PN_FAIL);
+	}
+
+	switch (type) {
+	case DIGEST_SHA1:
+	case DIGEST_MD5:
+		if ((md = digeststr(type, val, strlen(val), NULL)) == NULL) {
+			relay_close_http(con, 500, "failed to allocate digest");
+			goto fail;
+		}
+		pkv.key = md;
+		break;
+	case DIGEST_NONE:
+		pkv.key = val;
+		break;
+	}
+	pkv.type = NODE_TYPE_URL;
+	pkv.value = "";
+
+	if ((proot = RB_FIND(proto_tree, cre->tree, &pkv)) == NULL)
+		goto done;
+
+	PROTONODE_FOREACH(pnv, proot, entry) {
+		ret = relay_handle_http(cre, proot, pnv, &pkv, 0);
+		if (ret == PN_FAIL)
+			goto done;
+	}
+
+ done:
+	ret = PN_PASS;
+ fail:
+	if (md != NULL)
+		free(md);
+	free(val);
+	return (ret);
+}
+
+int
 relay_lookup_query(struct ctl_relay_event *cre)
 {
 	struct session		*con = (struct session *)cre->con;
diff --git a/usr.sbin/relayd/relayd.conf.5 b/usr.sbin/relayd/relayd.conf.5
index 844e92c0fd9..23cf9c0f026 100644
--- a/usr.sbin/relayd/relayd.conf.5
+++ b/usr.sbin/relayd/relayd.conf.5
@@ -1,4 +1,4 @@
-.\"	$OpenBSD: relayd.conf.5,v 1.62 2007/11/21 20:24:28 reyk Exp $
+.\"	$OpenBSD: relayd.conf.5,v 1.63 2007/11/22 10:09:53 reyk Exp $
 .\"
 .\" Copyright (c) 2006, 2007 Reyk Floeter <reyk@openbsd.org>
 .\" Copyright (c) 2006, 2007 Pierre-Yves Ritschard <pyr@openbsd.org>
@@ -15,7 +15,7 @@
 .\" ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
 .\" OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 .\"
-.Dd $Mdocdate: November 21 2007 $
+.Dd $Mdocdate: November 22 2007 $
 .Dt HOSTSTATED.CONF 5
 .Os
 .Sh NAME
@@ -553,7 +553,7 @@ will match the path of the requested URL without the hostname
 and query and the value will match the complete query,
 for example:
 .Bd -literal -offset indent
-request path filter "*" from "/index.html"
+request path filter "/index.html"
 request path filter "foo=bar*" from "/cgi-bin/t.cgi"
 .Ed
 .It Ic query
@@ -567,6 +567,18 @@ for example:
 # Will match /cgi-bin/example.pl?foo=bar&ok=yes
 request query expect "bar" from "foo"
 .Ed
+.It Ic url
+Look up the entity as an URL suffix/prefix expression consisting of a
+host name or suffix and a path name or prefix when using the
+.Ic http
+protocol.
+This type is only available with the direction
+.Ic request ,
+for example:
+.Bd -literal -offset indent
+request url filter "example.com/index.html"
+request url filter "example.com/test.cgi?val=1"
+.Ed
 .El
 .Pp
 The following actions are available:
@@ -604,6 +616,7 @@ If the entity is not present or the value doesn't match, the connection
 will be dropped.
 .It Xo
 .Ic expect
+.Op Ic digest
 .Ar key
 .Xc
 Expect an entity
@@ -611,6 +624,18 @@ Expect an entity
 with any possible value.
 This is the short form of
 .Ic expect Ar * Ic from Ar key .
+.Pp
+If the
+.Ic digest
+keyword is specified,
+compare the message digest of the entity against the defined string.
+The algorithm used is determined by the string length of the
+.Ar key
+argument, either SHA1 (40 characters) or MD5 (32 characters).
+To compute the digest, use this simple command:
+.Bd -literal -offset indent
+echo -n "example.com/path/?args" | sha1
+.Ed
 .It Ic filter Ar value Ic from Ar key
 Like the
 .Ic expect Ar .. Ic from
@@ -620,6 +645,7 @@ and a matching
 .Ar value .
 .It Xo
 .Ic filter
+.Op Ic digest
 .Ar key
 .Xc
 Like the
diff --git a/usr.sbin/relayd/relayd.h b/usr.sbin/relayd/relayd.h
index f3642a3d645..b016a3a5283 100644
--- a/usr.sbin/relayd/relayd.h
+++ b/usr.sbin/relayd/relayd.h
@@ -1,4 +1,4 @@
-/*	$OpenBSD: relayd.h,v 1.80 2007/11/21 20:28:38 reyk Exp $	*/
+/*	$OpenBSD: relayd.h,v 1.81 2007/11/22 10:09:53 reyk Exp $	*/
 
 /*
  * Copyright (c) 2006, 2007 Pierre-Yves Ritschard <pyr@openbsd.org>
@@ -437,7 +437,8 @@ enum nodetype {
 	NODE_TYPE_HEADER	= 0,
 	NODE_TYPE_QUERY		= 1,
 	NODE_TYPE_COOKIE	= 2,
-	NODE_TYPE_PATH		= 3
+	NODE_TYPE_PATH		= 3,
+	NODE_TYPE_URL		= 4
 };
 
 #define PNFLAG_MACRO			0x01
@@ -445,6 +446,9 @@ enum nodetype {
 #define PNFLAG_LOG			0x04
 #define PNFLAG_LOOKUP_QUERY		0x08
 #define PNFLAG_LOOKUP_COOKIE		0x10
+#define PNFLAG_LOOKUP_URL		0xe0
+#define PNFLAG_LOOKUP_URL_DIGEST	0xc0
+#define PNFLAG_LOOKUP_DIGEST(x)		(0x20 << x)
 
 enum noderesult {
 	PN_DROP			= 0,