summaryrefslogtreecommitdiff
path: root/net/ipv6/ioam6_iptunnel.c
diff options
context:
space:
mode:
Diffstat (limited to 'net/ipv6/ioam6_iptunnel.c')
-rw-r--r--net/ipv6/ioam6_iptunnel.c117
1 files changed, 77 insertions, 40 deletions
diff --git a/net/ipv6/ioam6_iptunnel.c b/net/ipv6/ioam6_iptunnel.c
index 9d8422e350f8..1fe7894f14dd 100644
--- a/net/ipv6/ioam6_iptunnel.c
+++ b/net/ipv6/ioam6_iptunnel.c
@@ -38,6 +38,7 @@ 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;
@@ -177,6 +178,14 @@ static int ioam6_build_state(struct net *net, struct nlattr *nla,
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;
@@ -253,14 +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;
@@ -291,7 +301,8 @@ static int ioam6_do_encap(struct net *net, struct sk_buff *skb,
struct ioam6_lwt_encap *tuninfo,
bool has_tunsrc,
struct in6_addr *tunsrc,
- struct in6_addr *tundst)
+ struct in6_addr *tundst,
+ struct dst_entry *cache_dst)
{
struct dst_entry *dst = skb_dst(skb);
struct ipv6hdr *hdr, *inner_hdr;
@@ -300,7 +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;
@@ -324,7 +335,7 @@ static int ioam6_do_encap(struct net *net, struct sk_buff *skb,
if (has_tunsrc)
memcpy(&hdr->saddr, tunsrc, sizeof(*tunsrc));
else
- ipv6_dev_get_saddr(net, dst->dev, &hdr->daddr,
+ ipv6_dev_get_saddr(net, dst_dev(dst), &hdr->daddr,
IPV6_PREFER_SRC_PUBLIC, &hdr->saddr);
skb_postpush_rcsum(skb, hdr, len);
@@ -334,8 +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;
@@ -343,14 +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:
@@ -359,7 +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;
@@ -369,7 +393,7 @@ do_encap:
/* Encapsulation (ip6ip6) */
err = ioam6_do_encap(net, skb, &ilwt->tuninfo,
ilwt->has_tunsrc, &ilwt->tunsrc,
- &ilwt->tundst);
+ &ilwt->tundst, dst);
if (unlikely(err))
goto drop;
@@ -387,52 +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;
+
+ 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;
- if (!ipv6_addr_equal(&orig_daddr, &ipv6_hdr(skb)->daddr)) {
+ 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();
- dst = dst_cache_get(&ilwt->cache);
+ 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();
- 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;
- }
-
- local_bh_disable();
- dst_cache_set_ip6(&ilwt->cache, dst, &fl6.saddr);
- local_bh_enable();
- }
+ err = skb_cow_head(skb, LL_RESERVED_SPACE(dst_dev(dst)));
+ 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);
}
generated by cgit v1.2.3 (git 2.39.0) at 2025-08-12 08:25:49 +0300