1 files changed, 151 insertions, 58 deletions

diff --git a/net/ipv6/ioam6_iptunnel.c b/net/ipv6/ioam6_iptunnel.c
index 7563f8c6aa87..40df8bdfaacd 100644
--- a/net/ipv6/ioam6_iptunnel.c
+++ b/net/ipv6/ioam6_iptunnel.c
@@ -38,12 +38,15 @@ struct ioam6_lwt_freq {
 };
 
 struct ioam6_lwt {
+	struct dst_entry null_dst;
 	struct dst_cache cache;
 	struct ioam6_lwt_freq freq;
 	atomic_t pkt_cnt;
 	u8 mode;
+	bool has_tunsrc;
+	struct in6_addr tunsrc;
 	struct in6_addr tundst;
-	struct ioam6_lwt_encap	tuninfo;
+	struct ioam6_lwt_encap tuninfo;
 };
 
 static const struct netlink_range_validation freq_range = {
@@ -72,8 +75,10 @@ static const struct nla_policy ioam6_iptunnel_policy[IOAM6_IPTUNNEL_MAX + 1] = {
 	[IOAM6_IPTUNNEL_MODE]	= NLA_POLICY_RANGE(NLA_U8,
 						   IOAM6_IPTUNNEL_MODE_MIN,
 						   IOAM6_IPTUNNEL_MODE_MAX),
+	[IOAM6_IPTUNNEL_SRC]	= NLA_POLICY_EXACT_LEN(sizeof(struct in6_addr)),
 	[IOAM6_IPTUNNEL_DST]	= NLA_POLICY_EXACT_LEN(sizeof(struct in6_addr)),
-	[IOAM6_IPTUNNEL_TRACE]	= NLA_POLICY_EXACT_LEN(sizeof(struct ioam6_trace_hdr)),
+	[IOAM6_IPTUNNEL_TRACE]	= NLA_POLICY_EXACT_LEN(
+					sizeof(struct ioam6_trace_hdr)),
 };
 
 static bool ioam6_validate_trace_hdr(struct ioam6_trace_hdr *trace)
@@ -85,7 +90,7 @@ static bool ioam6_validate_trace_hdr(struct ioam6_trace_hdr *trace)
 	    trace->type.bit12 | trace->type.bit13 | trace->type.bit14 |
 	    trace->type.bit15 | trace->type.bit16 | trace->type.bit17 |
 	    trace->type.bit18 | trace->type.bit19 | trace->type.bit20 |
-	    trace->type.bit21)
+	    trace->type.bit21 | trace->type.bit23)
 		return false;
 
 	trace->nodelen = 0;
@@ -138,10 +143,13 @@ static int ioam6_build_state(struct net *net, struct nlattr *nla,
 		}
 	}
 
-	if (!tb[IOAM6_IPTUNNEL_MODE])
-		mode = IOAM6_IPTUNNEL_MODE_INLINE;
-	else
-		mode = nla_get_u8(tb[IOAM6_IPTUNNEL_MODE]);
+	mode = nla_get_u8_default(tb[IOAM6_IPTUNNEL_MODE],
+				  IOAM6_IPTUNNEL_MODE_INLINE);
+
+	if (tb[IOAM6_IPTUNNEL_SRC] && mode == IOAM6_IPTUNNEL_MODE_INLINE) {
+		NL_SET_ERR_MSG(extack, "no tunnel src expected with this mode");
+		return -EINVAL;
+	}
 
 	if (!tb[IOAM6_IPTUNNEL_DST] && mode != IOAM6_IPTUNNEL_MODE_INLINE) {
 		NL_SET_ERR_MSG(extack, "this mode needs a tunnel destination");
@@ -167,19 +175,48 @@ static int ioam6_build_state(struct net *net, struct nlattr *nla,
 
 	ilwt = ioam6_lwt_state(lwt);
 	err = dst_cache_init(&ilwt->cache, GFP_ATOMIC);
-	if (err) {
-		kfree(lwt);
-		return err;
-	}
+	if (err)
+		goto free_lwt;
+
+	/* This "fake" dst_entry will be stored in a dst_cache, which will call
+	 * dst_hold() and dst_release() on it. We must ensure that dst_destroy()
+	 * will never be called. For that, its initial refcount is 1 and +1 when
+	 * it is stored in the cache. Then, +1/-1 each time we read the cache
+	 * and release it. Long story short, we're fine.
+	 */
+	dst_init(&ilwt->null_dst, NULL, NULL, DST_OBSOLETE_NONE, DST_NOCOUNT);
 
 	atomic_set(&ilwt->pkt_cnt, 0);
 	ilwt->freq.k = freq_k;
 	ilwt->freq.n = freq_n;
 
 	ilwt->mode = mode;
-	if (tb[IOAM6_IPTUNNEL_DST])
+
+	if (!tb[IOAM6_IPTUNNEL_SRC]) {
+		ilwt->has_tunsrc = false;
+	} else {
+		ilwt->has_tunsrc = true;
+		ilwt->tunsrc = nla_get_in6_addr(tb[IOAM6_IPTUNNEL_SRC]);
+
+		if (ipv6_addr_any(&ilwt->tunsrc)) {
+			NL_SET_ERR_MSG_ATTR(extack, tb[IOAM6_IPTUNNEL_SRC],
+					    "invalid tunnel source address");
+			err = -EINVAL;
+			goto free_cache;
+		}
+	}
+
+	if (tb[IOAM6_IPTUNNEL_DST]) {
 		ilwt->tundst = nla_get_in6_addr(tb[IOAM6_IPTUNNEL_DST]);
 
+		if (ipv6_addr_any(&ilwt->tundst)) {
+			NL_SET_ERR_MSG_ATTR(extack, tb[IOAM6_IPTUNNEL_DST],
+					    "invalid tunnel dest address");
+			err = -EINVAL;
+			goto free_cache;
+		}
+	}
+
 	tuninfo = ioam6_lwt_info(lwt);
 	tuninfo->eh.hdrlen = ((sizeof(*tuninfo) + len_aligned) >> 3) - 1;
 	tuninfo->pad[0] = IPV6_TLV_PADN;
@@ -201,6 +238,11 @@ static int ioam6_build_state(struct net *net, struct nlattr *nla,
 	*ts = lwt;
 
 	return 0;
+free_cache:
+	dst_cache_destroy(&ilwt->cache);
+free_lwt:
+	kfree(lwt);
+	return err;
 }
 
 static int ioam6_do_fill(struct net *net, struct sk_buff *skb)
@@ -220,14 +262,15 @@ static int ioam6_do_fill(struct net *net, struct sk_buff *skb)
 }
 
 static int ioam6_do_inline(struct net *net, struct sk_buff *skb,
-			   struct ioam6_lwt_encap *tuninfo)
+			   struct ioam6_lwt_encap *tuninfo,
+			   struct dst_entry *cache_dst)
 {
 	struct ipv6hdr *oldhdr, *hdr;
 	int hdrlen, err;
 
 	hdrlen = (tuninfo->eh.hdrlen + 1) << 3;
 
-	err = skb_cow_head(skb, hdrlen + skb->mac_len);
+	err = skb_cow_head(skb, hdrlen + dst_dev_overhead(cache_dst, skb));
 	if (unlikely(err))
 		return err;
 
@@ -256,7 +299,10 @@ static int ioam6_do_inline(struct net *net, struct sk_buff *skb,
 
 static int ioam6_do_encap(struct net *net, struct sk_buff *skb,
 			  struct ioam6_lwt_encap *tuninfo,
-			  struct in6_addr *tundst)
+			  bool has_tunsrc,
+			  struct in6_addr *tunsrc,
+			  struct in6_addr *tundst,
+			  struct dst_entry *cache_dst)
 {
 	struct dst_entry *dst = skb_dst(skb);
 	struct ipv6hdr *hdr, *inner_hdr;
@@ -265,7 +311,7 @@ static int ioam6_do_encap(struct net *net, struct sk_buff *skb,
 	hdrlen = (tuninfo->eh.hdrlen + 1) << 3;
 	len = sizeof(*hdr) + hdrlen;
 
-	err = skb_cow_head(skb, len + skb->mac_len);
+	err = skb_cow_head(skb, len + dst_dev_overhead(cache_dst, skb));
 	if (unlikely(err))
 		return err;
 
@@ -285,8 +331,12 @@ static int ioam6_do_encap(struct net *net, struct sk_buff *skb,
 	hdr->nexthdr = NEXTHDR_HOP;
 	hdr->payload_len = cpu_to_be16(skb->len - sizeof(*hdr));
 	hdr->daddr = *tundst;
-	ipv6_dev_get_saddr(net, dst->dev, &hdr->daddr,
-			   IPV6_PREFER_SRC_PUBLIC, &hdr->saddr);
+
+	if (has_tunsrc)
+		memcpy(&hdr->saddr, tunsrc, sizeof(*tunsrc));
+	else
+		ipv6_dev_get_saddr(net, dst->dev, &hdr->daddr,
+				   IPV6_PREFER_SRC_PUBLIC, &hdr->saddr);
 
 	skb_postpush_rcsum(skb, hdr, len);
 
@@ -295,8 +345,8 @@ static int ioam6_do_encap(struct net *net, struct sk_buff *skb,
 
 static int ioam6_output(struct net *net, struct sock *sk, struct sk_buff *skb)
 {
-	struct dst_entry *dst = skb_dst(skb);
-	struct in6_addr orig_daddr;
+	struct dst_entry *orig_dst = skb_dst(skb);
+	struct dst_entry *dst = NULL;
 	struct ioam6_lwt *ilwt;
 	int err = -EINVAL;
 	u32 pkt_cnt;
@@ -304,14 +354,27 @@ static int ioam6_output(struct net *net, struct sock *sk, struct sk_buff *skb)
 	if (skb->protocol != htons(ETH_P_IPV6))
 		goto drop;
 
-	ilwt = ioam6_lwt_state(dst->lwtstate);
+	ilwt = ioam6_lwt_state(orig_dst->lwtstate);
 
 	/* Check for insertion frequency (i.e., "k over n" insertions) */
 	pkt_cnt = atomic_fetch_inc(&ilwt->pkt_cnt);
 	if (pkt_cnt % ilwt->freq.n >= ilwt->freq.k)
 		goto out;
 
-	orig_daddr = ipv6_hdr(skb)->daddr;
+	local_bh_disable();
+	dst = dst_cache_get(&ilwt->cache);
+	local_bh_enable();
+
+	/* This is how we notify that the destination does not change after
+	 * transformation and that we need to use orig_dst instead of the cache
+	 */
+	if (dst == &ilwt->null_dst) {
+		dst_release(dst);
+
+		dst = orig_dst;
+		/* keep refcount balance: dst_release() is called at the end */
+		dst_hold(dst);
+	}
 
 	switch (ilwt->mode) {
 	case IOAM6_IPTUNNEL_MODE_INLINE:
@@ -320,7 +383,7 @@ do_inline:
 		if (ipv6_hdr(skb)->nexthdr == NEXTHDR_HOP)
 			goto out;
 
-		err = ioam6_do_inline(net, skb, &ilwt->tuninfo);
+		err = ioam6_do_inline(net, skb, &ilwt->tuninfo, dst);
 		if (unlikely(err))
 			goto drop;
 
@@ -328,7 +391,9 @@ do_inline:
 	case IOAM6_IPTUNNEL_MODE_ENCAP:
 do_encap:
 		/* Encapsulation (ip6ip6) */
-		err = ioam6_do_encap(net, skb, &ilwt->tuninfo, &ilwt->tundst);
+		err = ioam6_do_encap(net, skb, &ilwt->tuninfo,
+				     ilwt->has_tunsrc, &ilwt->tunsrc,
+				     &ilwt->tundst, dst);
 		if (unlikely(err))
 			goto drop;
 
@@ -346,52 +411,65 @@ do_encap:
 		goto drop;
 	}
 
-	err = skb_cow_head(skb, LL_RESERVED_SPACE(dst->dev));
-	if (unlikely(err))
-		goto drop;
+	if (unlikely(!dst)) {
+		struct ipv6hdr *hdr = ipv6_hdr(skb);
+		struct flowi6 fl6;
 
-	if (!ipv6_addr_equal(&orig_daddr, &ipv6_hdr(skb)->daddr)) {
-		preempt_disable();
-		dst = dst_cache_get(&ilwt->cache);
-		preempt_enable();
-
-		if (unlikely(!dst)) {
-			struct ipv6hdr *hdr = ipv6_hdr(skb);
-			struct flowi6 fl6;
-
-			memset(&fl6, 0, sizeof(fl6));
-			fl6.daddr = hdr->daddr;
-			fl6.saddr = hdr->saddr;
-			fl6.flowlabel = ip6_flowinfo(hdr);
-			fl6.flowi6_mark = skb->mark;
-			fl6.flowi6_proto = hdr->nexthdr;
-
-			dst = ip6_route_output(net, NULL, &fl6);
-			if (dst->error) {
-				err = dst->error;
-				dst_release(dst);
-				goto drop;
-			}
-
-			preempt_disable();
-			dst_cache_set_ip6(&ilwt->cache, dst, &fl6.saddr);
-			preempt_enable();
+		memset(&fl6, 0, sizeof(fl6));
+		fl6.daddr = hdr->daddr;
+		fl6.saddr = hdr->saddr;
+		fl6.flowlabel = ip6_flowinfo(hdr);
+		fl6.flowi6_mark = skb->mark;
+		fl6.flowi6_proto = hdr->nexthdr;
+
+		dst = ip6_route_output(net, NULL, &fl6);
+		if (dst->error) {
+			err = dst->error;
+			goto drop;
 		}
 
+		/* If the destination is the same after transformation (which is
+		 * a valid use case for IOAM), then we don't want to add it to
+		 * the cache in order to avoid a reference loop. Instead, we add
+		 * our fake dst_entry to the cache as a way to detect this case.
+		 * Otherwise, we add the resolved destination to the cache.
+		 */
+		local_bh_disable();
+		if (orig_dst->lwtstate == dst->lwtstate)
+			dst_cache_set_ip6(&ilwt->cache,
+					  &ilwt->null_dst, &fl6.saddr);
+		else
+			dst_cache_set_ip6(&ilwt->cache, dst, &fl6.saddr);
+		local_bh_enable();
+
+		err = skb_cow_head(skb, LL_RESERVED_SPACE(dst->dev));
+		if (unlikely(err))
+			goto drop;
+	}
+
+	/* avoid lwtunnel_output() reentry loop when destination is the same
+	 * after transformation (e.g., with the inline mode)
+	 */
+	if (orig_dst->lwtstate != dst->lwtstate) {
 		skb_dst_drop(skb);
 		skb_dst_set(skb, dst);
-
 		return dst_output(net, sk, skb);
 	}
 out:
-	return dst->lwtstate->orig_output(net, sk, skb);
+	dst_release(dst);
+	return orig_dst->lwtstate->orig_output(net, sk, skb);
 drop:
+	dst_release(dst);
 	kfree_skb(skb);
 	return err;
 }
 
 static void ioam6_destroy_state(struct lwtunnel_state *lwt)
 {
+	/* Since the refcount of per-cpu dst_entry caches will never be 0 (see
+	 * why above) when our "fake" dst_entry is used, it is not necessary to
+	 * remove them before calling dst_cache_destroy()
+	 */
 	dst_cache_destroy(&ioam6_lwt_state(lwt)->cache);
 }
 
@@ -414,6 +492,13 @@ static int ioam6_fill_encap_info(struct sk_buff *skb,
 		goto ret;
 
 	if (ilwt->mode != IOAM6_IPTUNNEL_MODE_INLINE) {
+		if (ilwt->has_tunsrc) {
+			err = nla_put_in6_addr(skb, IOAM6_IPTUNNEL_SRC,
+					       &ilwt->tunsrc);
+			if (err)
+				goto ret;
+		}
+
 		err = nla_put_in6_addr(skb, IOAM6_IPTUNNEL_DST, &ilwt->tundst);
 		if (err)
 			goto ret;
@@ -435,8 +520,12 @@ static int ioam6_encap_nlsize(struct lwtunnel_state *lwtstate)
 		  nla_total_size(sizeof(ilwt->mode)) +
 		  nla_total_size(sizeof(ilwt->tuninfo.traceh));
 
-	if (ilwt->mode != IOAM6_IPTUNNEL_MODE_INLINE)
+	if (ilwt->mode != IOAM6_IPTUNNEL_MODE_INLINE) {
+		if (ilwt->has_tunsrc)
+			nlsize += nla_total_size(sizeof(ilwt->tunsrc));
+
 		nlsize += nla_total_size(sizeof(ilwt->tundst));
+	}
 
 	return nlsize;
 }
@@ -451,17 +540,21 @@ static int ioam6_encap_cmp(struct lwtunnel_state *a, struct lwtunnel_state *b)
 	return (ilwt_a->freq.k != ilwt_b->freq.k ||
 		ilwt_a->freq.n != ilwt_b->freq.n ||
 		ilwt_a->mode != ilwt_b->mode ||
+		ilwt_a->has_tunsrc != ilwt_b->has_tunsrc ||
 		(ilwt_a->mode != IOAM6_IPTUNNEL_MODE_INLINE &&
 		 !ipv6_addr_equal(&ilwt_a->tundst, &ilwt_b->tundst)) ||
+		(ilwt_a->mode != IOAM6_IPTUNNEL_MODE_INLINE &&
+		 ilwt_a->has_tunsrc &&
+		 !ipv6_addr_equal(&ilwt_a->tunsrc, &ilwt_b->tunsrc)) ||
 		trace_a->namespace_id != trace_b->namespace_id);
 }
 
 static const struct lwtunnel_encap_ops ioam6_iptun_ops = {
 	.build_state		= ioam6_build_state,
 	.destroy_state		= ioam6_destroy_state,
-	.output		= ioam6_output,
+	.output			= ioam6_output,
 	.fill_encap		= ioam6_fill_encap_info,
-	.get_encap_size	= ioam6_encap_nlsize,
+	.get_encap_size		= ioam6_encap_nlsize,
 	.cmp_encap		= ioam6_encap_cmp,
 	.owner			= THIS_MODULE,
 };