[PATCH net-next 1/3] net: gro: set {inner_,}network_header in receive phase

From: Richard Gobert
Date: Thu Feb 22 2024 - 16:05:40 EST

Next message: Conor Dooley: "Re: [RFC PATCH 1/1] riscv: dts: Allow BUILTIN_DTB for all socs"
Previous message: Bartosz Golaszewski: "[PATCH] arm64: gitignore: ignore relacheck"
In reply to: Richard Gobert: "[PATCH net-next 0/3] net: gro: encapsulation bug fix and flush checks improvements"
Next in thread: Eric Dumazet: "Re: [PATCH net-next 1/3] net: gro: set {inner_,}network_header in receive phase"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Commits a602456 ("udp: Add GRO functions to UDP socket") and 57c67ff ("udp:
additional GRO support") introduce incorrect usage of {ip,ipv6}_hdr in the
complete phase of gro. The functions always return skb->network_header,
which in the case of encapsulated packets at the gro complete phase, is
always set to the innermost L3 of the packet. That means that calling
{ip,ipv6}_hdr for skbs which completed the GRO receive phase (both in
gro_list and *_gro_complete) when parsing an encapsulated packet's _outer_
L3/L4 may return an unexpected value.

This incorrect usage leads to a bug in GRO's UDP socket lookup.
udp{4,6}_lib_lookup_skb functions use ip_hdr/ipv6_hdr respectively. These
*_hdr functions return network_header which will point to the innermost L3,
resulting in the wrong offset being used in __udp{4,6}_lib_lookup with
encapsulated packets.

Reproduction example:

Endpoint configuration example (fou + local address bind)

# ip fou add port 6666 ipproto 4
# ip link add name tun1 type ipip remote 2.2.2.1 local 2.2.2.2 encap fou encap-dport 5555 encap-sport 6666 mode ipip
# ip link set tun1 up
# ip a add 1.1.1.2/24 dev tun1

Netperf TCP_STREAM result on net-next before patch is applied:

net-next main, GRO enabled:
$ netperf -H 1.1.1.2 -t TCP_STREAM -l 5
Recv Send Send
Socket Socket Message Elapsed
Size Size Size Time Throughput
bytes bytes bytes secs. 10^6bits/sec

131072 16384 16384 5.28 2.37

net-next main, GRO disabled:
$ netperf -H 1.1.1.2 -t TCP_STREAM -l 5
Recv Send Send
Socket Socket Message Elapsed
Size Size Size Time Throughput
bytes bytes bytes secs. 10^6bits/sec

131072 16384 16384 5.01 2745.06

patch applied, GRO enabled:
$ netperf -H 1.1.1.2 -t TCP_STREAM -l 5
Recv Send Send
Socket Socket Message Elapsed
Size Size Size Time Throughput
bytes bytes bytes secs. 10^6bits/sec

131072 16384 16384 5.01 2877.38

This patch fixes this bug and prevents similar future misuse of
network_header by setting network_header and inner_network_header to their
respective values during the receive phase of GRO. This results in
more coherent {inner_,}network_header values for every skb in gro_list,
which also means there's no need to set/fix these values before passing
the packet forward.

network_header is already set in dev_gro_receive and under encapsulation we
set inner_network_header. *_gro_complete functions use a new helper
function - skb_gro_complete_network_header, which returns the
network_header/inner_network_header offset during the GRO complete phase,
depending on skb->encapsulation.

Fixes: 57c67ff4bd92 ("udp: additional GRO support")
Signed-off-by: Richard Gobert <richardbgobert@xxxxxxxxx>
---
include/net/gro.h | 14 +++++++++++++-
net/8021q/vlan_core.c | 3 +++
net/ipv4/af_inet.c | 8 ++++----
net/ipv4/tcp_offload.c | 2 +-
net/ipv4/udp_offload.c | 2 +-
net/ipv6/ip6_offload.c | 11 +++++------
net/ipv6/tcpv6_offload.c | 2 +-
net/ipv6/udp_offload.c | 2 +-
8 files changed, 29 insertions(+), 15 deletions(-)

diff --git a/include/net/gro.h b/include/net/gro.h
index b435f0ddbf64..89502a7e35ed 100644
--- a/include/net/gro.h
+++ b/include/net/gro.h
@@ -177,10 +177,22 @@ static inline void *skb_gro_header(struct sk_buff *skb,
return ptr;
}

+static inline int skb_gro_network_offset(struct sk_buff *skb)
+{
+ return NAPI_GRO_CB(skb)->encap_mark ? skb_inner_network_offset(skb) :
+ skb_network_offset(skb);
+}
+
static inline void *skb_gro_network_header(struct sk_buff *skb)
{
return (NAPI_GRO_CB(skb)->frag0 ?: skb->data) +
- skb_network_offset(skb);
+ skb_gro_network_offset(skb);
+}
+
+static inline void *skb_gro_complete_network_header(struct sk_buff *skb)
+{
+ return skb->encapsulation ? skb_inner_network_header(skb) :
+ skb_network_header(skb);
}

static inline __wsum inet_gro_compute_pseudo(struct sk_buff *skb, int proto)
diff --git a/net/8021q/vlan_core.c b/net/8021q/vlan_core.c
index f00158234505..8bc871397e47 100644
--- a/net/8021q/vlan_core.c
+++ b/net/8021q/vlan_core.c
@@ -478,6 +478,9 @@ static struct sk_buff *vlan_gro_receive(struct list_head *head,
if (unlikely(!vhdr))
goto out;

+ if (!NAPI_GRO_CB(skb)->encap_mark)
+ skb_set_network_header(skb, hlen);
+
type = vhdr->h_vlan_encapsulated_proto;

ptype = gro_find_receive_by_type(type);
diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 835f4f9d98d2..c0f3c162bf73 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -1564,7 +1564,9 @@ struct sk_buff *inet_gro_receive(struct list_head *head, struct sk_buff *skb)

NAPI_GRO_CB(skb)->is_atomic = !!(iph->frag_off & htons(IP_DF));
NAPI_GRO_CB(skb)->flush |= flush;
- skb_set_network_header(skb, off);
+ if (NAPI_GRO_CB(skb)->encap_mark)
+ skb_set_inner_network_header(skb, off);
+
/* The above will be needed by the transport layer if there is one
* immediately following this IP hdr.
*/
@@ -1643,10 +1645,8 @@ int inet_gro_complete(struct sk_buff *skb, int nhoff)
int proto = iph->protocol;
int err = -ENOSYS;

- if (skb->encapsulation) {
+ if (skb->encapsulation)
skb_set_inner_protocol(skb, cpu_to_be16(ETH_P_IP));
- skb_set_inner_network_header(skb, nhoff);
- }

iph_set_totlen(iph, skb->len - nhoff);
csum_replace2(&iph->check, totlen, iph->tot_len);
diff --git a/net/ipv4/tcp_offload.c b/net/ipv4/tcp_offload.c
index 8311c38267b5..8bbcd3f502ac 100644
--- a/net/ipv4/tcp_offload.c
+++ b/net/ipv4/tcp_offload.c
@@ -330,7 +330,7 @@ struct sk_buff *tcp4_gro_receive(struct list_head *head, struct sk_buff *skb)

INDIRECT_CALLABLE_SCOPE int tcp4_gro_complete(struct sk_buff *skb, int thoff)
{
- const struct iphdr *iph = ip_hdr(skb);
+ const struct iphdr *iph = skb_gro_complete_network_header(skb);
struct tcphdr *th = tcp_hdr(skb);

th->check = ~tcp_v4_check(skb->len - thoff, iph->saddr,
diff --git a/net/ipv4/udp_offload.c b/net/ipv4/udp_offload.c
index 6c95d28d0c4a..7f59cede67f5 100644
--- a/net/ipv4/udp_offload.c
+++ b/net/ipv4/udp_offload.c
@@ -709,7 +709,7 @@ EXPORT_SYMBOL(udp_gro_complete);

INDIRECT_CALLABLE_SCOPE int udp4_gro_complete(struct sk_buff *skb, int nhoff)
{
- const struct iphdr *iph = ip_hdr(skb);
+ const struct iphdr *iph = skb_gro_complete_network_header(skb);
struct udphdr *uh = (struct udphdr *)(skb->data + nhoff);

/* do fraglist only if there is no outer UDP encap (or we already processed it) */
diff --git a/net/ipv6/ip6_offload.c b/net/ipv6/ip6_offload.c
index cca64c7809be..db7e3db587b9 100644
--- a/net/ipv6/ip6_offload.c
+++ b/net/ipv6/ip6_offload.c
@@ -67,7 +67,7 @@ static int ipv6_gro_pull_exthdrs(struct sk_buff *skb, int off, int proto)
off += len;
}

- skb_gro_pull(skb, off - skb_network_offset(skb));
+ skb_gro_pull(skb, off - skb_gro_network_offset(skb));
return proto;
}

@@ -236,7 +236,8 @@ INDIRECT_CALLABLE_SCOPE struct sk_buff *ipv6_gro_receive(struct list_head *head,
if (unlikely(!iph))
goto out;

- skb_set_network_header(skb, off);
+ if (NAPI_GRO_CB(skb)->encap_mark)
+ skb_set_inner_network_header(skb, off);

flush += ntohs(iph->payload_len) != skb->len - hlen;

@@ -259,7 +260,7 @@ INDIRECT_CALLABLE_SCOPE struct sk_buff *ipv6_gro_receive(struct list_head *head,
NAPI_GRO_CB(skb)->proto = proto;

flush--;
- nlen = skb_network_header_len(skb);
+ nlen = skb_gro_offset(skb) - off;

list_for_each_entry(p, head, list) {
const struct ipv6hdr *iph2;
@@ -353,10 +354,8 @@ INDIRECT_CALLABLE_SCOPE int ipv6_gro_complete(struct sk_buff *skb, int nhoff)
int err = -ENOSYS;
u32 payload_len;

- if (skb->encapsulation) {
+ if (skb->encapsulation)
skb_set_inner_protocol(skb, cpu_to_be16(ETH_P_IPV6));
- skb_set_inner_network_header(skb, nhoff);
- }

payload_len = skb->len - nhoff - sizeof(*iph);
if (unlikely(payload_len > IPV6_MAXPLEN)) {
diff --git a/net/ipv6/tcpv6_offload.c b/net/ipv6/tcpv6_offload.c
index bf0c957e4b5e..79eeaced2834 100644
--- a/net/ipv6/tcpv6_offload.c
+++ b/net/ipv6/tcpv6_offload.c
@@ -29,7 +29,7 @@ struct sk_buff *tcp6_gro_receive(struct list_head *head, struct sk_buff *skb)

INDIRECT_CALLABLE_SCOPE int tcp6_gro_complete(struct sk_buff *skb, int thoff)
{
- const struct ipv6hdr *iph = ipv6_hdr(skb);
+ const struct ipv6hdr *iph = skb_gro_complete_network_header(skb);
struct tcphdr *th = tcp_hdr(skb);

th->check = ~tcp_v6_check(skb->len - thoff, &iph->saddr,
diff --git a/net/ipv6/udp_offload.c b/net/ipv6/udp_offload.c
index 6b95ba241ebe..897caa2e39fb 100644
--- a/net/ipv6/udp_offload.c
+++ b/net/ipv6/udp_offload.c
@@ -164,7 +164,7 @@ struct sk_buff *udp6_gro_receive(struct list_head *head, struct sk_buff *skb)

INDIRECT_CALLABLE_SCOPE int udp6_gro_complete(struct sk_buff *skb, int nhoff)
{
- const struct ipv6hdr *ipv6h = ipv6_hdr(skb);
+ const struct ipv6hdr *ipv6h = skb_gro_complete_network_header(skb);
struct udphdr *uh = (struct udphdr *)(skb->data + nhoff);

/* do fraglist only if there is no outer UDP encap (or we already processed it) */
--
2.36.1

Next message: Conor Dooley: "Re: [RFC PATCH 1/1] riscv: dts: Allow BUILTIN_DTB for all socs"
Previous message: Bartosz Golaszewski: "[PATCH] arm64: gitignore: ignore relacheck"
In reply to: Richard Gobert: "[PATCH net-next 0/3] net: gro: encapsulation bug fix and flush checks improvements"
Next in thread: Eric Dumazet: "Re: [PATCH net-next 1/3] net: gro: set {inner_,}network_header in receive phase"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]