ip_tunnel: convert __be16 tunnel flags to bitmaps

Historically, tunnel flags like TUNNEL_CSUM or TUNNEL_ERSPAN_OPT
have been defined as __be16. Now all of those 16 bits are occupied
and there's no more free space for new flags.
It can't be simply switched to a bigger container with no
adjustments to the values, since it's an explicit Endian storage,
and on LE systems (__be16)0x0001 equals to
(__be64)0x0001000000000000.
We could probably define new 64-bit flags depending on the
Endianness, i.e. (__be64)0x0001 on BE and (__be64)0x00010000... on
LE, but that would introduce an Endianness dependency and spawn a
ton of Sparse warnings. To mitigate them, all of those places which
were adjusted with this change would be touched anyway, so why not
define stuff properly if there's no choice.

Define IP_TUNNEL_*_BIT counterparts as a bit number instead of the
value already coded and a fistful of <16 <-> bitmap> converters and
helpers. The two flags which have a different bit position are
SIT_ISATAP_BIT and VTI_ISVTI_BIT, as they were defined not as
__cpu_to_be16(), but as (__force __be16), i.e. had different
positions on LE and BE. Now they both have strongly defined places.
Change all __be16 fields which were used to store those flags, to
IP_TUNNEL_DECLARE_FLAGS() -> DECLARE_BITMAP(__IP_TUNNEL_FLAG_NUM) ->
unsigned long[1] for now, and replace all TUNNEL_* occurrences to
their bitmap counterparts. Use the converters in the places which talk
to the userspace, hardware (NFP) or other hosts (GRE header). The rest
must explicitly use the new flags only. This must be done at once,
otherwise there will be too many conversions throughout the code in
the intermediate commits.
Finally, disable the old __be16 flags for use in the kernel code
(except for the two 'irregular' flags mentioned above), to prevent
any accidental (mis)use of them. For the userspace, nothing is
changed, only additions were made.

Most noticeable bloat-o-meter difference (.text):

vmlinux:	307/-1 (306)
gre.ko:		62/0 (62)
ip_gre.ko:	941/-217 (724)	[*]
ip_tunnel.ko:	390/-900 (-510)	[**]
ip_vti.ko:	138/0 (138)
ip6_gre.ko:	534/-18 (516)	[*]
ip6_tunnel.ko:	118/-10 (108)

[*] gre_flags_to_tnl_flags() grew, but still is inlined
[**] ip_tunnel_find() got uninlined, hence such decrease

The average code size increase in non-extreme case is 100-200 bytes
per module, mostly due to sizeof(long) > sizeof(__be16), as
%__IP_TUNNEL_FLAG_NUM is less than %BITS_PER_LONG and the compilers
are able to expand the majority of bitmap_*() calls here into direct
operations on scalars.

Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

7 files changed, 177 insertions, 65 deletions

diff --git a/include/net/dst_metadata.h b/include/net/dst_metadata.h
index 1b7fae4c6b24..4160731dcb6e 100644
--- a/include/net/dst_metadata.h
+++ b/include/net/dst_metadata.h
@@ -198,7 +198,7 @@ static inline struct metadata_dst *__ip_tun_set_dst(__be32 saddr,
 						    __be32 daddr,
 						    __u8 tos, __u8 ttl,
 						    __be16 tp_dst,
-						    __be16 flags,
+						    const unsigned long *flags,
 						    __be64 tunnel_id,
 						    int md_size)
 {
@@ -215,7 +215,7 @@ static inline struct metadata_dst *__ip_tun_set_dst(__be32 saddr,
 }
 
 static inline struct metadata_dst *ip_tun_rx_dst(struct sk_buff *skb,
-						 __be16 flags,
+						 const unsigned long *flags,
 						 __be64 tunnel_id,
 						 int md_size)
 {
@@ -230,7 +230,7 @@ static inline struct metadata_dst *__ipv6_tun_set_dst(const struct in6_addr *sad
 						      __u8 tos, __u8 ttl,
 						      __be16 tp_dst,
 						      __be32 label,
-						      __be16 flags,
+						      const unsigned long *flags,
 						      __be64 tunnel_id,
 						      int md_size)
 {
@@ -243,7 +243,7 @@ static inline struct metadata_dst *__ipv6_tun_set_dst(const struct in6_addr *sad
 
 	info = &tun_dst->u.tun_info;
 	info->mode = IP_TUNNEL_INFO_IPV6;
-	info->key.tun_flags = flags;
+	ip_tunnel_flags_copy(info->key.tun_flags, flags);
 	info->key.tun_id = tunnel_id;
 	info->key.tp_src = 0;
 	info->key.tp_dst = tp_dst;
@@ -259,7 +259,7 @@ static inline struct metadata_dst *__ipv6_tun_set_dst(const struct in6_addr *sad
 }
 
 static inline struct metadata_dst *ipv6_tun_rx_dst(struct sk_buff *skb,
-						   __be16 flags,
+						   const unsigned long *flags,
 						   __be64 tunnel_id,
 						   int md_size)
 {
diff --git a/include/net/flow_dissector.h b/include/net/flow_dissector.h
index 1a7131d6cb0e..9ab376d1a677 100644
--- a/include/net/flow_dissector.h
+++ b/include/net/flow_dissector.h
@@ -97,7 +97,7 @@ struct flow_dissector_key_enc_opts {
 					 * here but seems difficult to #include
 					 */
 	u8 len;
-	__be16 dst_opt_type;
+	u32 dst_opt_type;
 };
 
 struct flow_dissector_key_keyid {
diff --git a/include/net/gre.h b/include/net/gre.h
index 4e209708b754..ccd293203284 100644
--- a/include/net/gre.h
+++ b/include/net/gre.h
@@ -49,67 +49,61 @@ static inline bool netif_is_ip6gretap(const struct net_device *dev)
 	       !strcmp(dev->rtnl_link_ops->kind, "ip6gretap");
 }
 
-static inline int gre_calc_hlen(__be16 o_flags)
+static inline int gre_calc_hlen(const unsigned long *o_flags)
 {
 	int addend = 4;
 
-	if (o_flags & TUNNEL_CSUM)
+	if (test_bit(IP_TUNNEL_CSUM_BIT, o_flags))
 		addend += 4;
-	if (o_flags & TUNNEL_KEY)
+	if (test_bit(IP_TUNNEL_KEY_BIT, o_flags))
 		addend += 4;
-	if (o_flags & TUNNEL_SEQ)
+	if (test_bit(IP_TUNNEL_SEQ_BIT, o_flags))
 		addend += 4;
 	return addend;
 }
 
-static inline __be16 gre_flags_to_tnl_flags(__be16 flags)
+static inline void gre_flags_to_tnl_flags(unsigned long *dst, __be16 flags)
 {
-	__be16 tflags = 0;
-
-	if (flags & GRE_CSUM)
-		tflags |= TUNNEL_CSUM;
-	if (flags & GRE_ROUTING)
-		tflags |= TUNNEL_ROUTING;
-	if (flags & GRE_KEY)
-		tflags |= TUNNEL_KEY;
-	if (flags & GRE_SEQ)
-		tflags |= TUNNEL_SEQ;
-	if (flags & GRE_STRICT)
-		tflags |= TUNNEL_STRICT;
-	if (flags & GRE_REC)
-		tflags |= TUNNEL_REC;
-	if (flags & GRE_VERSION)
-		tflags |= TUNNEL_VERSION;
-
-	return tflags;
+	IP_TUNNEL_DECLARE_FLAGS(res) = { };
+
+	__assign_bit(IP_TUNNEL_CSUM_BIT, res, flags & GRE_CSUM);
+	__assign_bit(IP_TUNNEL_ROUTING_BIT, res, flags & GRE_ROUTING);
+	__assign_bit(IP_TUNNEL_KEY_BIT, res, flags & GRE_KEY);
+	__assign_bit(IP_TUNNEL_SEQ_BIT, res, flags & GRE_SEQ);
+	__assign_bit(IP_TUNNEL_STRICT_BIT, res, flags & GRE_STRICT);
+	__assign_bit(IP_TUNNEL_REC_BIT, res, flags & GRE_REC);
+	__assign_bit(IP_TUNNEL_VERSION_BIT, res, flags & GRE_VERSION);
+
+	ip_tunnel_flags_copy(dst, res);
 }
 
-static inline __be16 gre_tnl_flags_to_gre_flags(__be16 tflags)
+static inline __be16 gre_tnl_flags_to_gre_flags(const unsigned long *tflags)
 {
 	__be16 flags = 0;
 
-	if (tflags & TUNNEL_CSUM)
+	if (test_bit(IP_TUNNEL_CSUM_BIT, tflags))
 		flags |= GRE_CSUM;
-	if (tflags & TUNNEL_ROUTING)
+	if (test_bit(IP_TUNNEL_ROUTING_BIT, tflags))
 		flags |= GRE_ROUTING;
-	if (tflags & TUNNEL_KEY)
+	if (test_bit(IP_TUNNEL_KEY_BIT, tflags))
 		flags |= GRE_KEY;
-	if (tflags & TUNNEL_SEQ)
+	if (test_bit(IP_TUNNEL_SEQ_BIT, tflags))
 		flags |= GRE_SEQ;
-	if (tflags & TUNNEL_STRICT)
+	if (test_bit(IP_TUNNEL_STRICT_BIT, tflags))
 		flags |= GRE_STRICT;
-	if (tflags & TUNNEL_REC)
+	if (test_bit(IP_TUNNEL_REC_BIT, tflags))
 		flags |= GRE_REC;
-	if (tflags & TUNNEL_VERSION)
+	if (test_bit(IP_TUNNEL_VERSION_BIT, tflags))
 		flags |= GRE_VERSION;
 
 	return flags;
 }
 
 static inline void gre_build_header(struct sk_buff *skb, int hdr_len,
-				    __be16 flags, __be16 proto,
+				    const unsigned long *flags, __be16 proto,
 				    __be32 key, __be32 seq)
 {
+	IP_TUNNEL_DECLARE_FLAGS(cond) = { };
 	struct gre_base_hdr *greh;
 
 	skb_push(skb, hdr_len);
@@ -120,18 +114,22 @@ static inline void gre_build_header(struct sk_buff *skb, int hdr_len,
 	greh->flags = gre_tnl_flags_to_gre_flags(flags);
 	greh->protocol = proto;
 
-	if (flags & (TUNNEL_KEY | TUNNEL_CSUM | TUNNEL_SEQ)) {
+	__set_bit(IP_TUNNEL_KEY_BIT, cond);
+	__set_bit(IP_TUNNEL_CSUM_BIT, cond);
+	__set_bit(IP_TUNNEL_SEQ_BIT, cond);
+
+	if (ip_tunnel_flags_intersect(flags, cond)) {
 		__be32 *ptr = (__be32 *)(((u8 *)greh) + hdr_len - 4);
 
-		if (flags & TUNNEL_SEQ) {
+		if (test_bit(IP_TUNNEL_SEQ_BIT, flags)) {
 			*ptr = seq;
 			ptr--;
 		}
-		if (flags & TUNNEL_KEY) {
+		if (test_bit(IP_TUNNEL_KEY_BIT, flags)) {
 			*ptr = key;
 			ptr--;
 		}
-		if (flags & TUNNEL_CSUM &&
+		if (test_bit(IP_TUNNEL_CSUM_BIT, flags) &&
 		    !(skb_shinfo(skb)->gso_type &
 		      (SKB_GSO_GRE | SKB_GSO_GRE_CSUM))) {
 			*ptr = 0;
diff --git a/include/net/ip6_tunnel.h b/include/net/ip6_tunnel.h
index 74b369bddf49..399592405c72 100644
--- a/include/net/ip6_tunnel.h
+++ b/include/net/ip6_tunnel.h
@@ -30,8 +30,8 @@ struct __ip6_tnl_parm {
 	struct in6_addr laddr;	/* local tunnel end-point address */
 	struct in6_addr raddr;	/* remote tunnel end-point address */
 
-	__be16			i_flags;
-	__be16			o_flags;
+	IP_TUNNEL_DECLARE_FLAGS(i_flags);
+	IP_TUNNEL_DECLARE_FLAGS(o_flags);
 	__be32			i_key;
 	__be32			o_key;
 
diff --git a/include/net/ip_tunnels.h b/include/net/ip_tunnels.h
index 20f0319ab149..ed8e48cc9054 100644
--- a/include/net/ip_tunnels.h
+++ b/include/net/ip_tunnels.h
@@ -36,6 +36,24 @@
 	(sizeof_field(struct ip_tunnel_key, u) -		\
 	 sizeof_field(struct ip_tunnel_key, u.ipv4))
 
+#define __ipt_flag_op(op, ...)					\
+	op(__VA_ARGS__, __IP_TUNNEL_FLAG_NUM)
+
+#define IP_TUNNEL_DECLARE_FLAGS(...)				\
+	__ipt_flag_op(DECLARE_BITMAP, __VA_ARGS__)
+
+#define ip_tunnel_flags_zero(...)	__ipt_flag_op(bitmap_zero, __VA_ARGS__)
+#define ip_tunnel_flags_copy(...)	__ipt_flag_op(bitmap_copy, __VA_ARGS__)
+#define ip_tunnel_flags_and(...)	__ipt_flag_op(bitmap_and, __VA_ARGS__)
+#define ip_tunnel_flags_or(...)		__ipt_flag_op(bitmap_or, __VA_ARGS__)
+
+#define ip_tunnel_flags_empty(...)				\
+	__ipt_flag_op(bitmap_empty, __VA_ARGS__)
+#define ip_tunnel_flags_intersect(...)				\
+	__ipt_flag_op(bitmap_intersects, __VA_ARGS__)
+#define ip_tunnel_flags_subset(...)				\
+	__ipt_flag_op(bitmap_subset, __VA_ARGS__)
+
 struct ip_tunnel_key {
 	__be64			tun_id;
 	union {
@@ -48,11 +66,11 @@ struct ip_tunnel_key {
 			struct in6_addr dst;
 		} ipv6;
 	} u;
-	__be16			tun_flags;
-	u8			tos;		/* TOS for IPv4, TC for IPv6 */
-	u8			ttl;		/* TTL for IPv4, HL for IPv6 */
+	IP_TUNNEL_DECLARE_FLAGS(tun_flags);
 	__be32			label;		/* Flow Label for IPv6 */
 	u32			nhid;
+	u8			tos;		/* TOS for IPv4, TC for IPv6 */
+	u8			ttl;		/* TTL for IPv4, HL for IPv6 */
 	__be16			tp_src;
 	__be16			tp_dst;
 	__u8			flow_flags;
@@ -110,14 +128,14 @@ struct ip_tunnel_prl_entry {
 
 struct metadata_dst;
 
-/* Kernel-side copy of ip_tunnel_parm */
+/* Kernel-side variant of ip_tunnel_parm */
 struct ip_tunnel_parm_kern {
 	char			name[IFNAMSIZ];
-	int			link;
-	__be16			i_flags;
-	__be16			o_flags;
+	IP_TUNNEL_DECLARE_FLAGS(i_flags);
+	IP_TUNNEL_DECLARE_FLAGS(o_flags);
 	__be32			i_key;
 	__be32			o_key;
+	int			link;
 	struct iphdr		iph;
 };
 
@@ -168,7 +186,7 @@ struct ip_tunnel {
 };
 
 struct tnl_ptk_info {
-	__be16 flags;
+	IP_TUNNEL_DECLARE_FLAGS(flags);
 	__be16 proto;
 	__be32 key;
 	__be32 seq;
@@ -190,11 +208,77 @@ struct ip_tunnel_net {
 	int type;
 };
 
+static inline void ip_tunnel_set_options_present(unsigned long *flags)
+{
+	IP_TUNNEL_DECLARE_FLAGS(present) = { };
+
+	__set_bit(IP_TUNNEL_GENEVE_OPT_BIT, present);
+	__set_bit(IP_TUNNEL_VXLAN_OPT_BIT, present);
+	__set_bit(IP_TUNNEL_ERSPAN_OPT_BIT, present);
+	__set_bit(IP_TUNNEL_GTP_OPT_BIT, present);
+
+	ip_tunnel_flags_or(flags, flags, present);
+}
+
+static inline void ip_tunnel_clear_options_present(unsigned long *flags)
+{
+	IP_TUNNEL_DECLARE_FLAGS(present) = { };
+
+	__set_bit(IP_TUNNEL_GENEVE_OPT_BIT, present);
+	__set_bit(IP_TUNNEL_VXLAN_OPT_BIT, present);
+	__set_bit(IP_TUNNEL_ERSPAN_OPT_BIT, present);
+	__set_bit(IP_TUNNEL_GTP_OPT_BIT, present);
+
+	__ipt_flag_op(bitmap_andnot, flags, flags, present);
+}
+
+static inline bool ip_tunnel_is_options_present(const unsigned long *flags)
+{
+	IP_TUNNEL_DECLARE_FLAGS(present) = { };
+
+	__set_bit(IP_TUNNEL_GENEVE_OPT_BIT, present);
+	__set_bit(IP_TUNNEL_VXLAN_OPT_BIT, present);
+	__set_bit(IP_TUNNEL_ERSPAN_OPT_BIT, present);
+	__set_bit(IP_TUNNEL_GTP_OPT_BIT, present);
+
+	return ip_tunnel_flags_intersect(flags, present);
+}
+
+static inline bool ip_tunnel_flags_is_be16_compat(const unsigned long *flags)
+{
+	IP_TUNNEL_DECLARE_FLAGS(supp) = { };
+
+	bitmap_set(supp, 0, BITS_PER_TYPE(__be16));
+	__set_bit(IP_TUNNEL_VTI_BIT, supp);
+
+	return ip_tunnel_flags_subset(flags, supp);
+}
+
+static inline void ip_tunnel_flags_from_be16(unsigned long *dst, __be16 flags)
+{
+	ip_tunnel_flags_zero(dst);
+
+	bitmap_write(dst, be16_to_cpu(flags), 0, BITS_PER_TYPE(__be16));
+	__assign_bit(IP_TUNNEL_VTI_BIT, dst, flags & VTI_ISVTI);
+}
+
+static inline __be16 ip_tunnel_flags_to_be16(const unsigned long *flags)
+{
+	__be16 ret;
+
+	ret = cpu_to_be16(bitmap_read(flags, 0, BITS_PER_TYPE(__be16)));
+	if (test_bit(IP_TUNNEL_VTI_BIT, flags))
+		ret |= VTI_ISVTI;
+
+	return ret;
+}
+
 static inline void ip_tunnel_key_init(struct ip_tunnel_key *key,
 				      __be32 saddr, __be32 daddr,
 				      u8 tos, u8 ttl, __be32 label,
 				      __be16 tp_src, __be16 tp_dst,
-				      __be64 tun_id, __be16 tun_flags)
+				      __be64 tun_id,
+				      const unsigned long *tun_flags)
 {
 	key->tun_id = tun_id;
 	key->u.ipv4.src = saddr;
@@ -204,7 +288,7 @@ static inline void ip_tunnel_key_init(struct ip_tunnel_key *key,
 	key->tos = tos;
 	key->ttl = ttl;
 	key->label = label;
-	key->tun_flags = tun_flags;
+	ip_tunnel_flags_copy(key->tun_flags, tun_flags);
 
 	/* For the tunnel types on the top of IPsec, the tp_src and tp_dst of
 	 * the upper tunnel are used.
@@ -225,12 +309,8 @@ ip_tunnel_dst_cache_usable(const struct sk_buff *skb,
 {
 	if (skb->mark)
 		return false;
-	if (!info)
-		return true;
-	if (info->key.tun_flags & TUNNEL_NOCACHE)
-		return false;
 
-	return true;
+	return !info || !test_bit(IP_TUNNEL_NOCACHE_BIT, info->key.tun_flags);
 }
 
 static inline unsigned short ip_tunnel_info_af(const struct ip_tunnel_info
@@ -313,7 +393,7 @@ int __ip_tunnel_change_mtu(struct net_device *dev, int new_mtu, bool strict);
 int ip_tunnel_change_mtu(struct net_device *dev, int new_mtu);
 
 struct ip_tunnel *ip_tunnel_lookup(struct ip_tunnel_net *itn,
-				   int link, __be16 flags,
+				   int link, const unsigned long *flags,
 				   __be32 remote, __be32 local,
 				   __be32 key);
 
@@ -529,12 +609,13 @@ static inline void ip_tunnel_info_opts_get(void *to,
 
 static inline void ip_tunnel_info_opts_set(struct ip_tunnel_info *info,
 					   const void *from, int len,
-					   __be16 flags)
+					   const unsigned long *flags)
 {
 	info->options_len = len;
 	if (len > 0) {
 		memcpy(ip_tunnel_info_opts(info), from, len);
-		info->key.tun_flags |= flags;
+		ip_tunnel_flags_or(info->key.tun_flags, info->key.tun_flags,
+				   flags);
 	}
 }
 
@@ -578,7 +659,7 @@ static inline void ip_tunnel_info_opts_get(void *to,
 
 static inline void ip_tunnel_info_opts_set(struct ip_tunnel_info *info,
 					   const void *from, int len,
-					   __be16 flags)
+					   const unsigned long *flags)
 {
 	info->options_len = 0;
 }
diff --git a/include/net/udp_tunnel.h b/include/net/udp_tunnel.h
index d716214fe03d..a93dc51f6323 100644
--- a/include/net/udp_tunnel.h
+++ b/include/net/udp_tunnel.h
@@ -179,8 +179,8 @@ struct dst_entry *udp_tunnel6_dst_lookup(struct sk_buff *skb,
 					 struct dst_cache *dst_cache);
 
 struct metadata_dst *udp_tun_rx_dst(struct sk_buff *skb, unsigned short family,
-				    __be16 flags, __be64 tunnel_id,
-				    int md_size);
+				    const unsigned long *flags,
+				    __be64 tunnel_id, int md_size);
 
 #ifdef CONFIG_INET
 static inline int udp_tunnel_handle_offloads(struct sk_buff *skb, bool udp_csum)
diff --git a/include/uapi/linux/if_tunnel.h b/include/uapi/linux/if_tunnel.h
index 102119628ff5..838927dd73a4 100644
--- a/include/uapi/linux/if_tunnel.h
+++ b/include/uapi/linux/if_tunnel.h
@@ -161,6 +161,14 @@ enum {
 
 #define IFLA_VTI_MAX	(__IFLA_VTI_MAX - 1)
 
+#ifndef __KERNEL__
+/* Historically, tunnel flags have been defined as __be16 and now there are
+ * no free bits left. It is strongly advised to switch the already existing
+ * userspace code to the new *_BIT definitions from down below, as __be16
+ * can't be simply cast to a wider type on LE systems. All new flags and
+ * code must use *_BIT only.
+ */
+
 #define TUNNEL_CSUM		__cpu_to_be16(0x01)
 #define TUNNEL_ROUTING		__cpu_to_be16(0x02)
 #define TUNNEL_KEY		__cpu_to_be16(0x04)
@@ -181,5 +189,30 @@ enum {
 #define TUNNEL_OPTIONS_PRESENT \
 		(TUNNEL_GENEVE_OPT | TUNNEL_VXLAN_OPT | TUNNEL_ERSPAN_OPT | \
 		TUNNEL_GTP_OPT)
+#endif
+
+enum {
+	IP_TUNNEL_CSUM_BIT		= 0U,
+	IP_TUNNEL_ROUTING_BIT,
+	IP_TUNNEL_KEY_BIT,
+	IP_TUNNEL_SEQ_BIT,
+	IP_TUNNEL_STRICT_BIT,
+	IP_TUNNEL_REC_BIT,
+	IP_TUNNEL_VERSION_BIT,
+	IP_TUNNEL_NO_KEY_BIT,
+	IP_TUNNEL_DONT_FRAGMENT_BIT,
+	IP_TUNNEL_OAM_BIT,
+	IP_TUNNEL_CRIT_OPT_BIT,
+	IP_TUNNEL_GENEVE_OPT_BIT,	/* OPTIONS_PRESENT */
+	IP_TUNNEL_VXLAN_OPT_BIT,	/* OPTIONS_PRESENT */
+	IP_TUNNEL_NOCACHE_BIT,
+	IP_TUNNEL_ERSPAN_OPT_BIT,	/* OPTIONS_PRESENT */
+	IP_TUNNEL_GTP_OPT_BIT,		/* OPTIONS_PRESENT */
+
+	IP_TUNNEL_VTI_BIT,
+	IP_TUNNEL_SIT_ISATAP_BIT	= IP_TUNNEL_VTI_BIT,
+
+	__IP_TUNNEL_FLAG_NUM,
+};
 
 #endif /* _UAPI_IF_TUNNEL_H_ */