Existing bpf_clone_redirect() helper clones skb before redirecting
it to RX or TX of destination netdev.
Introduce bpf_redirect() helper that does that without cloning.
Benchmarked with two hosts using 10G ixgbe NICs.
One host is doing line rate pktgen.
Another host is configured as:
$ tc qdisc add dev $dev ingress
$ tc filter add dev $dev root pref 10 u32 match u32 0 0 flowid 1:2 \
action bpf run object-file tcbpf1_kern.o section clone_redirect_xmit drop
so it receives the packet on $dev and immediately xmits it on $dev + 1
The section 'clone_redirect_xmit' in tcbpf1_kern.o file has the program
that does bpf_clone_redirect() and performance is 2.0 Mpps
$ tc filter add dev $dev root pref 10 u32 match u32 0 0 flowid 1:2 \
action bpf run object-file tcbpf1_kern.o section redirect_xmit drop
which is using bpf_redirect() - 2.4 Mpps
and using cls_bpf with integrated actions as:
$ tc filter add dev $dev root pref 10 \
bpf run object-file tcbpf1_kern.o section redirect_xmit integ_act classid 1
performance is 2.5 Mpps
To summarize:
u32+act_bpf using clone_redirect - 2.0 Mpps
u32+act_bpf using redirect - 2.4 Mpps
cls_bpf using redirect - 2.5 Mpps
For comparison linux bridge in this setup is doing 2.1 Mpps
and ixgbe rx + drop in ip_rcv - 7.8 Mpps
Signed-off-by: Alexei Starovoitov <a...@plumgrid.com>
Acked-by: Daniel Borkmann <dan...@iogearbox.net>
---
This approach is using per_cpu scratch area to store ifindex and flags.
The other alternatives discussed at plumbers are slower and more intrusive.
include/net/sch_generic.h | 1 +
include/uapi/linux/bpf.h | 8 +++++++
include/uapi/linux/pkt_cls.h | 1 +
net/core/dev.c | 8 +++++++
net/core/filter.c | 49 ++++++++++++++++++++++++++++++++++++++++++
net/sched/act_bpf.c | 1 +
net/sched/cls_bpf.c | 1 +
samples/bpf/bpf_helpers.h | 4 ++++
samples/bpf/tcbpf1_kern.c | 24 ++++++++++++++++++++-
9 files changed, 96 insertions(+), 1 deletion(-)
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index da61febb9091..4c79ce8c1f92 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -402,6 +402,7 @@ void __qdisc_calculate_pkt_len(struct sk_buff *skb,
const struct qdisc_size_table *stab);
bool tcf_destroy(struct tcf_proto *tp, bool force);
void tcf_destroy_chain(struct tcf_proto __rcu **fl);
+int skb_do_redirect(struct sk_buff *);
/* Reset all TX qdiscs greater then index of a device. */
static inline void qdisc_reset_all_tx_gt(struct net_device *dev, unsigned int
i)
diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
index 2fbd1c71fa3b..4ec0b5488294 100644
--- a/include/uapi/linux/bpf.h
+++ b/include/uapi/linux/bpf.h
@@ -272,6 +272,14 @@ enum bpf_func_id {
BPF_FUNC_skb_get_tunnel_key,
BPF_FUNC_skb_set_tunnel_key,
BPF_FUNC_perf_event_read, /* u64 bpf_perf_event_read(&map, index)
*/
+ /**
+ * bpf_redirect(ifindex, flags) - redirect to another netdev
+ * @ifindex: ifindex of the net device
+ * @flags: bit 0 - if set, redirect to ingress instead of egress
+ * other bits - reserved
+ * Return: TC_ACT_REDIRECT
+ */
+ BPF_FUNC_redirect,
__BPF_FUNC_MAX_ID,
};
diff --git a/include/uapi/linux/pkt_cls.h b/include/uapi/linux/pkt_cls.h
index 0a262a83f9d4..439873775d49 100644
--- a/include/uapi/linux/pkt_cls.h
+++ b/include/uapi/linux/pkt_cls.h
@@ -87,6 +87,7 @@ enum {
#define TC_ACT_STOLEN 4
#define TC_ACT_QUEUED 5
#define TC_ACT_REPEAT 6
+#define TC_ACT_REDIRECT 7
#define TC_ACT_JUMP 0x10000000
/* Action type identifiers*/
diff --git a/net/core/dev.c b/net/core/dev.c
index 877c84834d81..d6a492e57874 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -3668,6 +3668,14 @@ static inline struct sk_buff *handle_ing(struct sk_buff
*skb,
case TC_ACT_QUEUED:
kfree_skb(skb);
return NULL;
+ case TC_ACT_REDIRECT:
+ /* skb_mac_header check was done by cls/act_bpf, so
+ * we can safely push the L2 header back before
+ * redirecting to another netdev
+ */
+ __skb_push(skb, skb->mac_len);
+ skb_do_redirect(skb);
+ return NULL;
default:
break;
}
diff --git a/net/core/filter.c b/net/core/filter.c
index 971d6ba89758..5bf273bab781 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -1427,6 +1427,53 @@ const struct bpf_func_proto bpf_clone_redirect_proto = {
.arg3_type = ARG_ANYTHING,
};
+struct redirect_info {
+ u32 ifindex;
+ u32 flags;
+};
+
+static DEFINE_PER_CPU(struct redirect_info, redirect_info);
+static u64 bpf_redirect(u64 ifindex, u64 flags, u64 r3, u64 r4, u64 r5)
+{
+ struct redirect_info *ri = this_cpu_ptr(&redirect_info);
+
+ ri->ifindex = ifindex;
+ ri->flags = flags;
+ return TC_ACT_REDIRECT;
+}
+
+int skb_do_redirect(struct sk_buff *skb)
+{
+ struct redirect_info *ri = this_cpu_ptr(&redirect_info);
+ struct net_device *dev;
+
+ dev = dev_get_by_index_rcu(dev_net(skb->dev), ri->ifindex);
+ ri->ifindex = 0;
+ if (unlikely(!dev)) {
+ kfree_skb(skb);
+ return -EINVAL;
+ }
+
+ if (unlikely(!(dev->flags & IFF_UP))) {
+ kfree_skb(skb);
+ return -EINVAL;
+ }
+
+ if (BPF_IS_REDIRECT_INGRESS(ri->flags))
+ return dev_forward_skb(dev, skb);
+
+ skb->dev = dev;
+ return dev_queue_xmit(skb);
+}
+
+const struct bpf_func_proto bpf_redirect_proto = {
+ .func = bpf_redirect,
+ .gpl_only = false,
+ .ret_type = RET_INTEGER,
+ .arg1_type = ARG_ANYTHING,
+ .arg2_type = ARG_ANYTHING,
+};
+
static u64 bpf_get_cgroup_classid(u64 r1, u64 r2, u64 r3, u64 r4, u64 r5)
{
return task_get_classid((struct sk_buff *) (unsigned long) r1);
@@ -1607,6 +1654,8 @@ tc_cls_act_func_proto(enum bpf_func_id func_id)
return &bpf_skb_get_tunnel_key_proto;
case BPF_FUNC_skb_set_tunnel_key:
return bpf_get_skb_set_tunnel_key_proto();
+ case BPF_FUNC_redirect:
+ return &bpf_redirect_proto;
default:
return sk_filter_func_proto(func_id);
}
diff --git a/net/sched/act_bpf.c b/net/sched/act_bpf.c
index 559bfa011bda..0bc6f912f870 100644
--- a/net/sched/act_bpf.c
+++ b/net/sched/act_bpf.c
@@ -72,6 +72,7 @@ static int tcf_bpf(struct sk_buff *skb, const struct
tc_action *act,
case TC_ACT_PIPE:
case TC_ACT_RECLASSIFY:
case TC_ACT_OK:
+ case TC_ACT_REDIRECT:
action = filter_res;
break;
case TC_ACT_SHOT:
diff --git a/net/sched/cls_bpf.c b/net/sched/cls_bpf.c
index 77b0ef148256..0590816ab7b0 100644
--- a/net/sched/cls_bpf.c
+++ b/net/sched/cls_bpf.c
@@ -70,6 +70,7 @@ static int cls_bpf_exec_opcode(int code)
case TC_ACT_PIPE:
case TC_ACT_STOLEN:
case TC_ACT_QUEUED:
+ case TC_ACT_REDIRECT:
case TC_ACT_UNSPEC:
return code;
default:
diff --git a/samples/bpf/bpf_helpers.h b/samples/bpf/bpf_helpers.h
index 3a44d3a272af..21aa1b44c30c 100644
--- a/samples/bpf/bpf_helpers.h
+++ b/samples/bpf/bpf_helpers.h
@@ -33,6 +33,10 @@ static int (*bpf_get_current_comm)(void *buf, int buf_size) =
(void *) BPF_FUNC_get_current_comm;
static int (*bpf_perf_event_read)(void *map, int index) =
(void *) BPF_FUNC_perf_event_read;
+static int (*bpf_clone_redirect)(void *ctx, int ifindex, int flags) =
+ (void *) BPF_FUNC_clone_redirect;
+static int (*bpf_redirect)(int ifindex, int flags) =
+ (void *) BPF_FUNC_redirect;
/* llvm builtin functions that eBPF C program may use to
* emit BPF_LD_ABS and BPF_LD_IND instructions
diff --git a/samples/bpf/tcbpf1_kern.c b/samples/bpf/tcbpf1_kern.c
index 9bfb2eb34563..fa051b3d53ee 100644
--- a/samples/bpf/tcbpf1_kern.c
+++ b/samples/bpf/tcbpf1_kern.c
@@ -5,7 +5,7 @@
#include <uapi/linux/in.h>
#include <uapi/linux/tcp.h>
#include <uapi/linux/filter.h>
-
+#include <uapi/linux/pkt_cls.h>
#include "bpf_helpers.h"
/* compiler workaround */
@@ -64,4 +64,26 @@ int bpf_prog1(struct __sk_buff *skb)
return 0;
}
+SEC("redirect_xmit")
+int _redirect_xmit(struct __sk_buff *skb)
+{
+ return bpf_redirect(skb->ifindex + 1, 0);
+}
+SEC("redirect_recv")
+int _redirect_recv(struct __sk_buff *skb)
+{
+ return bpf_redirect(skb->ifindex + 1, 1);
+}
+SEC("clone_redirect_xmit")
+int _clone_redirect_xmit(struct __sk_buff *skb)
+{
+ bpf_clone_redirect(skb, skb->ifindex + 1, 0);
+ return TC_ACT_SHOT;
+}
+SEC("clone_redirect_recv")
+int _clone_redirect_recv(struct __sk_buff *skb)
+{
+ bpf_clone_redirect(skb, skb->ifindex + 1, 1);
+ return TC_ACT_SHOT;
+}
char _license[] SEC("license") = "GPL";
--
1.7.9.5
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
