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>

3 files changed, 40 insertions, 30 deletions

diff --git a/net/netfilter/ipvs/ip_vs_core.c b/net/netfilter/ipvs/ip_vs_core.c
index a2c16b501087..c7a8a08b7308 100644
--- a/net/netfilter/ipvs/ip_vs_core.c
+++ b/net/netfilter/ipvs/ip_vs_core.c
@@ -1550,6 +1550,7 @@ static int ipvs_gre_decap(struct netns_ipvs *ipvs, struct sk_buff *skb,
 	if (!dest)
 		goto unk;
 	if (dest->tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GRE) {
+		IP_TUNNEL_DECLARE_FLAGS(flags);
 		__be16 type;
 
 		/* Only support version 0 and C (csum) */
@@ -1560,7 +1561,10 @@ static int ipvs_gre_decap(struct netns_ipvs *ipvs, struct sk_buff *skb,
 		if (type != htons(ETH_P_IP))
 			goto unk;
 		*proto = IPPROTO_IPIP;
-		return gre_calc_hlen(gre_flags_to_tnl_flags(greh->flags));
+
+		gre_flags_to_tnl_flags(flags, greh->flags);
+
+		return gre_calc_hlen(flags);
 	}
 
 unk:
diff --git a/net/netfilter/ipvs/ip_vs_xmit.c b/net/netfilter/ipvs/ip_vs_xmit.c
index 65e0259178da..39b5fd6bbf65 100644
--- a/net/netfilter/ipvs/ip_vs_xmit.c
+++ b/net/netfilter/ipvs/ip_vs_xmit.c
@@ -390,10 +390,10 @@ __ip_vs_get_out_rt(struct netns_ipvs *ipvs, int skb_af, struct sk_buff *skb,
 			    skb->ip_summed == CHECKSUM_PARTIAL)
 				mtu -= GUE_PLEN_REMCSUM + GUE_LEN_PRIV;
 		} else if (dest->tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GRE) {
-			__be16 tflags = 0;
+			IP_TUNNEL_DECLARE_FLAGS(tflags) = { };
 
 			if (dest->tun_flags & IP_VS_TUNNEL_ENCAP_FLAG_CSUM)
-				tflags |= TUNNEL_CSUM;
+				__set_bit(IP_TUNNEL_CSUM_BIT, tflags);
 			mtu -= gre_calc_hlen(tflags);
 		}
 		if (mtu < 68) {
@@ -553,10 +553,10 @@ __ip_vs_get_out_rt_v6(struct netns_ipvs *ipvs, int skb_af, struct sk_buff *skb,
 			    skb->ip_summed == CHECKSUM_PARTIAL)
 				mtu -= GUE_PLEN_REMCSUM + GUE_LEN_PRIV;
 		} else if (dest->tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GRE) {
-			__be16 tflags = 0;
+			IP_TUNNEL_DECLARE_FLAGS(tflags) = { };
 
 			if (dest->tun_flags & IP_VS_TUNNEL_ENCAP_FLAG_CSUM)
-				tflags |= TUNNEL_CSUM;
+				__set_bit(IP_TUNNEL_CSUM_BIT, tflags);
 			mtu -= gre_calc_hlen(tflags);
 		}
 		if (mtu < IPV6_MIN_MTU) {
@@ -1082,11 +1082,11 @@ ipvs_gre_encap(struct net *net, struct sk_buff *skb,
 {
 	__be16 proto = *next_protocol == IPPROTO_IPIP ?
 				htons(ETH_P_IP) : htons(ETH_P_IPV6);
-	__be16 tflags = 0;
+	IP_TUNNEL_DECLARE_FLAGS(tflags) = { };
 	size_t hdrlen;
 
 	if (cp->dest->tun_flags & IP_VS_TUNNEL_ENCAP_FLAG_CSUM)
-		tflags |= TUNNEL_CSUM;
+		__set_bit(IP_TUNNEL_CSUM_BIT, tflags);
 
 	hdrlen = gre_calc_hlen(tflags);
 	gre_build_header(skb, hdrlen, tflags, proto, 0, 0);
@@ -1165,11 +1165,11 @@ ip_vs_tunnel_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
 
 		max_headroom += sizeof(struct udphdr) + gue_hdrlen;
 	} else if (tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GRE) {
+		IP_TUNNEL_DECLARE_FLAGS(tflags) = { };
 		size_t gre_hdrlen;
-		__be16 tflags = 0;
 
 		if (tun_flags & IP_VS_TUNNEL_ENCAP_FLAG_CSUM)
-			tflags |= TUNNEL_CSUM;
+			__set_bit(IP_TUNNEL_CSUM_BIT, tflags);
 		gre_hdrlen = gre_calc_hlen(tflags);
 
 		max_headroom += gre_hdrlen;
@@ -1310,11 +1310,11 @@ ip_vs_tunnel_xmit_v6(struct sk_buff *skb, struct ip_vs_conn *cp,
 
 		max_headroom += sizeof(struct udphdr) + gue_hdrlen;
 	} else if (tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GRE) {
+		IP_TUNNEL_DECLARE_FLAGS(tflags) = { };
 		size_t gre_hdrlen;
-		__be16 tflags = 0;
 
 		if (tun_flags & IP_VS_TUNNEL_ENCAP_FLAG_CSUM)
-			tflags |= TUNNEL_CSUM;
+			__set_bit(IP_TUNNEL_CSUM_BIT, tflags);
 		gre_hdrlen = gre_calc_hlen(tflags);
 
 		max_headroom += gre_hdrlen;
diff --git a/net/netfilter/nft_tunnel.c b/net/netfilter/nft_tunnel.c
index f735d79d8be5..60a76e6e348e 100644
--- a/net/netfilter/nft_tunnel.c
+++ b/net/netfilter/nft_tunnel.c
@@ -174,8 +174,8 @@ struct nft_tunnel_opts {
 		struct erspan_metadata	erspan;
 		u8	data[IP_TUNNEL_OPTS_MAX];
 	} u;
+	IP_TUNNEL_DECLARE_FLAGS(flags);
 	u32	len;
-	__be16	flags;
 };
 
 struct nft_tunnel_obj {
@@ -271,7 +271,8 @@ static int nft_tunnel_obj_vxlan_init(const struct nlattr *attr,
 	opts->u.vxlan.gbp = ntohl(nla_get_be32(tb[NFTA_TUNNEL_KEY_VXLAN_GBP]));
 
 	opts->len	= sizeof(struct vxlan_metadata);
-	opts->flags	= TUNNEL_VXLAN_OPT;
+	ip_tunnel_flags_zero(opts->flags);
+	__set_bit(IP_TUNNEL_VXLAN_OPT_BIT, opts->flags);
 
 	return 0;
 }
@@ -325,7 +326,8 @@ static int nft_tunnel_obj_erspan_init(const struct nlattr *attr,
 	opts->u.erspan.version = version;
 
 	opts->len	= sizeof(struct erspan_metadata);
-	opts->flags	= TUNNEL_ERSPAN_OPT;
+	ip_tunnel_flags_zero(opts->flags);
+	__set_bit(IP_TUNNEL_ERSPAN_OPT_BIT, opts->flags);
 
 	return 0;
 }
@@ -366,7 +368,8 @@ static int nft_tunnel_obj_geneve_init(const struct nlattr *attr,
 	opt->length = data_len / 4;
 	opt->opt_class = nla_get_be16(tb[NFTA_TUNNEL_KEY_GENEVE_CLASS]);
 	opt->type = nla_get_u8(tb[NFTA_TUNNEL_KEY_GENEVE_TYPE]);
-	opts->flags = TUNNEL_GENEVE_OPT;
+	ip_tunnel_flags_zero(opts->flags);
+	__set_bit(IP_TUNNEL_GENEVE_OPT_BIT, opts->flags);
 
 	return 0;
 }
@@ -385,8 +388,8 @@ static int nft_tunnel_obj_opts_init(const struct nft_ctx *ctx,
 				    struct nft_tunnel_opts *opts)
 {
 	struct nlattr *nla;
-	__be16 type = 0;
 	int err, rem;
+	u32 type = 0;
 
 	err = nla_validate_nested_deprecated(attr, NFTA_TUNNEL_KEY_OPTS_MAX,
 					     nft_tunnel_opts_policy, NULL);
@@ -401,7 +404,7 @@ static int nft_tunnel_obj_opts_init(const struct nft_ctx *ctx,
 			err = nft_tunnel_obj_vxlan_init(nla, opts);
 			if (err)
 				return err;
-			type = TUNNEL_VXLAN_OPT;
+			type = IP_TUNNEL_VXLAN_OPT_BIT;
 			break;
 		case NFTA_TUNNEL_KEY_OPTS_ERSPAN:
 			if (type)
@@ -409,15 +412,15 @@ static int nft_tunnel_obj_opts_init(const struct nft_ctx *ctx,
 			err = nft_tunnel_obj_erspan_init(nla, opts);
 			if (err)
 				return err;
-			type = TUNNEL_ERSPAN_OPT;
+			type = IP_TUNNEL_ERSPAN_OPT_BIT;
 			break;
 		case NFTA_TUNNEL_KEY_OPTS_GENEVE:
-			if (type && type != TUNNEL_GENEVE_OPT)
+			if (type && type != IP_TUNNEL_GENEVE_OPT_BIT)
 				return -EINVAL;
 			err = nft_tunnel_obj_geneve_init(nla, opts);
 			if (err)
 				return err;
-			type = TUNNEL_GENEVE_OPT;
+			type = IP_TUNNEL_GENEVE_OPT_BIT;
 			break;
 		default:
 			return -EOPNOTSUPP;
@@ -454,7 +457,9 @@ static int nft_tunnel_obj_init(const struct nft_ctx *ctx,
 	memset(&info, 0, sizeof(info));
 	info.mode		= IP_TUNNEL_INFO_TX;
 	info.key.tun_id		= key32_to_tunnel_id(nla_get_be32(tb[NFTA_TUNNEL_KEY_ID]));
-	info.key.tun_flags	= TUNNEL_KEY | TUNNEL_CSUM | TUNNEL_NOCACHE;
+	__set_bit(IP_TUNNEL_KEY_BIT, info.key.tun_flags);
+	__set_bit(IP_TUNNEL_CSUM_BIT, info.key.tun_flags);
+	__set_bit(IP_TUNNEL_NOCACHE_BIT, info.key.tun_flags);
 
 	if (tb[NFTA_TUNNEL_KEY_IP]) {
 		err = nft_tunnel_obj_ip_init(ctx, tb[NFTA_TUNNEL_KEY_IP], &info);
@@ -483,11 +488,12 @@ static int nft_tunnel_obj_init(const struct nft_ctx *ctx,
 			return -EOPNOTSUPP;
 
 		if (tun_flags & NFT_TUNNEL_F_ZERO_CSUM_TX)
-			info.key.tun_flags &= ~TUNNEL_CSUM;
+			__clear_bit(IP_TUNNEL_CSUM_BIT, info.key.tun_flags);
 		if (tun_flags & NFT_TUNNEL_F_DONT_FRAGMENT)
-			info.key.tun_flags |= TUNNEL_DONT_FRAGMENT;
+			__set_bit(IP_TUNNEL_DONT_FRAGMENT_BIT,
+				  info.key.tun_flags);
 		if (tun_flags & NFT_TUNNEL_F_SEQ_NUMBER)
-			info.key.tun_flags |= TUNNEL_SEQ;
+			__set_bit(IP_TUNNEL_SEQ_BIT, info.key.tun_flags);
 	}
 	if (tb[NFTA_TUNNEL_KEY_TOS])
 		info.key.tos = nla_get_u8(tb[NFTA_TUNNEL_KEY_TOS]);
@@ -583,7 +589,7 @@ static int nft_tunnel_opts_dump(struct sk_buff *skb,
 	if (!nest)
 		return -1;
 
-	if (opts->flags & TUNNEL_VXLAN_OPT) {
+	if (test_bit(IP_TUNNEL_VXLAN_OPT_BIT, opts->flags)) {
 		inner = nla_nest_start_noflag(skb, NFTA_TUNNEL_KEY_OPTS_VXLAN);
 		if (!inner)
 			goto failure;
@@ -591,7 +597,7 @@ static int nft_tunnel_opts_dump(struct sk_buff *skb,
 				 htonl(opts->u.vxlan.gbp)))
 			goto inner_failure;
 		nla_nest_end(skb, inner);
-	} else if (opts->flags & TUNNEL_ERSPAN_OPT) {
+	} else if (test_bit(IP_TUNNEL_ERSPAN_OPT_BIT, opts->flags)) {
 		inner = nla_nest_start_noflag(skb, NFTA_TUNNEL_KEY_OPTS_ERSPAN);
 		if (!inner)
 			goto failure;
@@ -613,7 +619,7 @@ static int nft_tunnel_opts_dump(struct sk_buff *skb,
 			break;
 		}
 		nla_nest_end(skb, inner);
-	} else if (opts->flags & TUNNEL_GENEVE_OPT) {
+	} else if (test_bit(IP_TUNNEL_GENEVE_OPT_BIT, opts->flags)) {
 		struct geneve_opt *opt;
 		int offset = 0;
 
@@ -658,11 +664,11 @@ static int nft_tunnel_flags_dump(struct sk_buff *skb,
 {
 	u32 flags = 0;
 
-	if (info->key.tun_flags & TUNNEL_DONT_FRAGMENT)
+	if (test_bit(IP_TUNNEL_DONT_FRAGMENT_BIT, info->key.tun_flags))
 		flags |= NFT_TUNNEL_F_DONT_FRAGMENT;
-	if (!(info->key.tun_flags & TUNNEL_CSUM))
+	if (!test_bit(IP_TUNNEL_CSUM_BIT, info->key.tun_flags))
 		flags |= NFT_TUNNEL_F_ZERO_CSUM_TX;
-	if (info->key.tun_flags & TUNNEL_SEQ)
+	if (test_bit(IP_TUNNEL_SEQ_BIT, info->key.tun_flags))
 		flags |= NFT_TUNNEL_F_SEQ_NUMBER;
 
 	if (nla_put_be32(skb, NFTA_TUNNEL_KEY_FLAGS, htonl(flags)) < 0)