Merge branch 'bpf-misc'

davem330 · davem330 · commit 3004932c8042 · 2016-03-18T19:38:53.000-04:00
Daniel Borkmann says:

====================
Minor BPF follow-ups

Some minor last follow-ups I still had in my queue. The first one adds
readability support for __sk_buff's tc_classid member, the remaining
two are some minor cleanups. For details please see individual patches.
====================

Signed-off-by: David S. Miller &lt;davem@davemloft.net&gt;
diff --git a/include/net/dst.h b/include/net/dst.h
@@ -398,6 +398,18 @@ static inline void skb_tunnel_rx(struct sk_buff *skb, struct net_device *dev,
 	__skb_tunnel_rx(skb, dev, net);
 }
 
+static inline u32 dst_tclassid(const struct sk_buff *skb)
+{
+#ifdef CONFIG_IP_ROUTE_CLASSID
+	const struct dst_entry *dst;
+
+	dst = skb_dst(skb);
+	if (dst)
+		return dst->tclassid;
+#endif
+	return 0;
+}
+
 int dst_discard_out(struct net *net, struct sock *sk, struct sk_buff *skb);
 static inline int dst_discard(struct sk_buff *skb)
 {
diff --git a/include/net/ip_tunnels.h b/include/net/ip_tunnels.h
@@ -7,6 +7,8 @@
 #include <linux/socket.h>
 #include <linux/types.h>
 #include <linux/u64_stats_sync.h>
+#include <linux/bitops.h>
+
 #include <net/dsfield.h>
 #include <net/gro_cells.h>
 #include <net/inet_ecn.h>
@@ -57,6 +59,11 @@ struct ip_tunnel_key {
 #define IP_TUNNEL_INFO_TX	0x01	/* represents tx tunnel parameters */
 #define IP_TUNNEL_INFO_IPV6	0x02	/* key contains IPv6 addresses */
 
+/* Maximum tunnel options length. */
+#define IP_TUNNEL_OPTS_MAX					\
+	GENMASK((FIELD_SIZEOF(struct ip_tunnel_info,		\
+			      options_len) * BITS_PER_BYTE) - 1, 0)
+
 struct ip_tunnel_info {
 	struct ip_tunnel_key	key;
 #ifdef CONFIG_DST_CACHE
diff --git a/net/core/filter.c b/net/core/filter.c
@@ -1682,14 +1682,7 @@ static const struct bpf_func_proto bpf_get_cgroup_classid_proto = {
 
 static u64 bpf_get_route_realm(u64 r1, u64 r2, u64 r3, u64 r4, u64 r5)
 {
-#ifdef CONFIG_IP_ROUTE_CLASSID
-	const struct dst_entry *dst;
-
-	dst = skb_dst((struct sk_buff *) (unsigned long) r1);
-	if (dst)
-		return dst->tclassid;
-#endif
-	return 0;
+	return dst_tclassid((struct sk_buff *) (unsigned long) r1);
 }
 
 static const struct bpf_func_proto bpf_get_route_realm_proto = {
@@ -1911,8 +1904,6 @@ static const struct bpf_func_proto bpf_skb_set_tunnel_key_proto = {
 	.arg4_type	= ARG_ANYTHING,
 };
 
-#define BPF_TUNLEN_MAX	255
-
 static u64 bpf_skb_set_tunnel_opt(u64 r1, u64 r2, u64 size, u64 r4, u64 r5)
 {
 	struct sk_buff *skb = (struct sk_buff *) (long) r1;
@@ -1922,7 +1913,7 @@ static u64 bpf_skb_set_tunnel_opt(u64 r1, u64 r2, u64 size, u64 r4, u64 r5)
 
 	if (unlikely(info != &md->u.tun_info || (size & (sizeof(u32) - 1))))
 		return -EINVAL;
-	if (unlikely(size > BPF_TUNLEN_MAX))
+	if (unlikely(size > IP_TUNNEL_OPTS_MAX))
 		return -ENOMEM;
 
 	ip_tunnel_info_opts_set(info, from, size);
@@ -1943,13 +1934,10 @@ static const struct bpf_func_proto *
 bpf_get_skb_set_tunnel_proto(enum bpf_func_id which)
 {
 	if (!md_dst) {
-		BUILD_BUG_ON(FIELD_SIZEOF(struct ip_tunnel_info,
-					  options_len) != 1);
-
 		/* Race is not possible, since it's called from verifier
 		 * that is holding verifier mutex.
 		 */
-		md_dst = metadata_dst_alloc_percpu(BPF_TUNLEN_MAX,
+		md_dst = metadata_dst_alloc_percpu(IP_TUNNEL_OPTS_MAX,
 						   GFP_KERNEL);
 		if (!md_dst)
 			return NULL;
@@ -2069,16 +2057,14 @@ static bool sk_filter_is_valid_access(int off, int size,
 static bool tc_cls_act_is_valid_access(int off, int size,
 				       enum bpf_access_type type)
 {
-	if (off == offsetof(struct __sk_buff, tc_classid))
-		return type == BPF_WRITE ? true : false;
-
 	if (type == BPF_WRITE) {
 		switch (off) {
 		case offsetof(struct __sk_buff, mark):
 		case offsetof(struct __sk_buff, tc_index):
 		case offsetof(struct __sk_buff, priority):
 		case offsetof(struct __sk_buff, cb[0]) ...
-			offsetof(struct __sk_buff, cb[4]):
+		     offsetof(struct __sk_buff, cb[4]):
+		case offsetof(struct __sk_buff, tc_classid):
 			break;
 		default:
 			return false;
@@ -2195,8 +2181,10 @@ static u32 bpf_net_convert_ctx_access(enum bpf_access_type type, int dst_reg,
 		ctx_off -= offsetof(struct __sk_buff, tc_classid);
 		ctx_off += offsetof(struct sk_buff, cb);
 		ctx_off += offsetof(struct qdisc_skb_cb, tc_classid);
-		WARN_ON(type != BPF_WRITE);
-		*insn++ = BPF_STX_MEM(BPF_H, dst_reg, src_reg, ctx_off);
+		if (type == BPF_WRITE)
+			*insn++ = BPF_STX_MEM(BPF_H, dst_reg, src_reg, ctx_off);
+		else
+			*insn++ = BPF_LDX_MEM(BPF_H, dst_reg, src_reg, ctx_off);
 		break;
 
 	case offsetof(struct __sk_buff, tc_index):
diff --git a/net/ipv4/ip_tunnel_core.c b/net/ipv4/ip_tunnel_core.c
@@ -398,6 +398,12 @@ static const struct lwtunnel_encap_ops ip6_tun_lwt_ops = {
 
 void __init ip_tunnel_core_init(void)
 {
+	/* If you land here, make sure whether increasing ip_tunnel_info's
+	 * options_len is a reasonable choice with its usage in front ends
+	 * (f.e., it's part of flow keys, etc).
+	 */
+	BUILD_BUG_ON(IP_TUNNEL_OPTS_MAX != 255);
+
 	lwtunnel_encap_add_ops(&ip_tun_lwt_ops, LWTUNNEL_ENCAP_IP);
 	lwtunnel_encap_add_ops(&ip6_tun_lwt_ops, LWTUNNEL_ENCAP_IP6);
 }
diff --git a/net/openvswitch/flow.h b/net/openvswitch/flow.h
@@ -55,7 +55,7 @@ struct ovs_tunnel_info {
 	FIELD_SIZEOF(struct sw_flow_key, recirc_id))
 
 struct sw_flow_key {
-	u8 tun_opts[255];
+	u8 tun_opts[IP_TUNNEL_OPTS_MAX];
 	u8 tun_opts_len;
 	struct ip_tunnel_key tun_key;	/* Encapsulating tunnel key. */
 	struct {

Original file line number	Diff line number	Diff line change
`@@ -398,6 +398,12 @@ static const struct lwtunnel_encap_ops ip6_tun_lwt_ops = {`
`398`	`398`
`399`	`399`	`void __init ip_tunnel_core_init(void)`
`400`	`400`	`{`
	`401`	`+ /* If you land here, make sure whether increasing ip_tunnel_info's`
	`402`	`+ * options_len is a reasonable choice with its usage in front ends`
	`403`	`+ * (f.e., it's part of flow keys, etc).`
	`404`	`+ */`
	`405`	`+ BUILD_BUG_ON(IP_TUNNEL_OPTS_MAX != 255);`
	`406`	`+`
`401`	`407`	`lwtunnel_encap_add_ops(&ip_tun_lwt_ops, LWTUNNEL_ENCAP_IP);`
`402`	`408`	`lwtunnel_encap_add_ops(&ip6_tun_lwt_ops, LWTUNNEL_ENCAP_IP6);`
`403`	`409`	`}`