samples/bpf: start basic TC redirect benchmark

Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>

kernel/samples/bpf/Makefile

@@ -12,6 +12,11 @@ TARGETS += xdp_bench02_drop_pattern
 
 TARGETS += bpf_tail_calls01
 
+# TC bpf targets uses bpf-elf-loader included in tc/iproute2.  Thus,
+# it is unnecessary to link "user" binary with bpf_load.c.  TODO, if
+# somone cares, makefile should have separate target for TC.
+TARGETS += tc_bench01_redirect
+
 # Experimental target
 TARGETS += xdp_rxhash
 

kernel/samples/bpf/tc_bench01_redirect_kern.c

@@ -0,0 +1,110 @@
+/*  TC (Traffic Control) eBPF redirect benchmark
+ *
+ *  NOTICE: TC loading is different from XDP loading. TC bpf objects
+ *          use the 'tc' cmdline tool from iproute2 for loading and
+ *          attaching bpf programs.
+ *
+ *  Copyright(c) 2017 Jesper Dangaard Brouer, Red Hat Inc.
+ */
+#include <uapi/linux/bpf.h>
+#include <uapi/linux/if_ether.h>
+#include <uapi/linux/if_packet.h>
+#include <uapi/linux/if_vlan.h>
+#include <uapi/linux/ip.h>
+#include <uapi/linux/in.h>
+#include <uapi/linux/tcp.h>
+#include <uapi/linux/udp.h>
+
+#include <uapi/linux/pkt_cls.h>
+
+#include "bpf_helpers.h"
+
+/* Notice: TC and iproute2 bpf-loader uses another elf map layout */
+struct bpf_elf_map {
+	__u32 type;
+	__u32 size_key;
+	__u32 size_value;
+	__u32 max_elem;
+	__u32 flags;
+	__u32 id;
+	__u32 pinning;
+};
+
+/* TODO: Describe what this PIN_GLOBAL_NS value 2 means???
+ *
+ * A file is automatically created here:
+ *  /sys/fs/bpf/tc/globals/egress_ifindex
+ */
+#define PIN_GLOBAL_NS	2
+
+struct bpf_elf_map SEC("maps") egress_ifindex = {
+	.type = BPF_MAP_TYPE_ARRAY,
+	.size_key = sizeof(int),
+	.size_value = sizeof(int),
+	.pinning = PIN_GLOBAL_NS,
+	.max_elem = 1,
+};
+
+static void swap_src_dst_mac(void *data)
+{
+	unsigned short *p = data;
+	unsigned short dst[3];
+
+	dst[0] = p[0];
+	dst[1] = p[1];
+	dst[2] = p[2];
+	p[0] = p[3];
+	p[1] = p[4];
+	p[2] = p[5];
+	p[3] = dst[0];
+	p[4] = dst[1];
+	p[5] = dst[2];
+}
+
+/* Notice this section name is used when attaching TC filter
+ *
+ * Like:
+ *  $TC qdisc   add dev $DEV clsact
+ *  $TC filter  add dev $DEV ingress bpf da obj $BPF_OBJ sec ingress_redirect
+ *  $TC filter show dev $DEV ingress
+ *  $TC filter  del dev $DEV ingress
+ *
+ * Does TC redirect respect IP-forward settings?
+ *
+ */
+SEC("ingress_redirect")
+int _ingress_redirect(struct __sk_buff *skb)
+{
+	void *data     = (void *)(long)skb->data;
+	void *data_end = (void *)(long)skb->data_end;
+	struct ethhdr *eth = data;
+	int key = 0, *ifindex;
+
+	if (data + sizeof(*eth) > data_end)
+		return TC_ACT_OK;
+
+	/* Keep ARP resolution working */
+	if (eth->h_proto == htons(ETH_P_ARP))
+		return TC_ACT_OK;
+
+	/* Lookup what ifindex to redirect packets to */
+	ifindex = bpf_map_lookup_elem(&egress_ifindex, &key);
+	if (!ifindex)
+		return TC_ACT_OK;
+
+	if (*ifindex == 0)
+		return TC_ACT_OK; // or TC_ACT_SHOT ?
+
+	/* FIXME: with mlx5 we need to update MAC-addr else the HW
+	 * will drop the frames silently.
+	 */
+
+	/* Swap src and dst mac-addr if ingress==egress */
+	if (*ifindex == 5)
+		swap_src_dst_mac(data);
+
+	//return bpf_redirect(*ifindex, BPF_F_INGRESS); // __bpf_rx_skb
+	return bpf_redirect(*ifindex, 0); // __bpf_tx_skb / __dev_xmit_skb
+}
+
+char _license[] SEC("license") = "GPL";

kernel/samples/bpf/tc_bench01_redirect_user.c

@@ -0,0 +1,108 @@
+/* Copyright(c) 2017 Jesper Dangaard Brouer, Red Hat, Inc.
+ */
+static const char *__doc__=
+ " TC redirect benchmark\n\n"
+ "  You must attach bpf object via TC cmdline tool\n"
+;
+
+#include <errno.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdbool.h>
+#include <string.h>
+#include <unistd.h>
+#include <locale.h>
+
+#include <getopt.h>
+#include <net/if.h>
+#include <time.h>
+
+#include "libbpf.h"
+
+static int verbose = 1;
+static const char *mapfile = "/sys/fs/bpf/tc/globals/egress_ifindex";
+
+static const struct option long_options[] = {
+	{"help",	no_argument,		NULL, 'h' },
+	{"egress",	required_argument,	NULL, 'o' },
+	{0, 0, NULL,  0 }
+};
+
+static void usage(char *argv[])
+{
+	int i;
+	printf("\nDOCUMENTATION:\n%s\n", __doc__);
+	printf("\n");
+	printf(" Usage: %s (options-see-below)\n",
+	       argv[0]);
+	printf(" Listing options:\n");
+	for (i = 0; long_options[i].name != 0; i++) {
+		printf(" --%-12s", long_options[i].name);
+		if (long_options[i].flag != NULL)
+			printf(" flag (internal value:%d)",
+			       *long_options[i].flag);
+		else
+			printf(" short-option: -%c",
+			       long_options[i].val);
+		printf("\n");
+	}
+	printf("\n");
+}
+
+int main(int argc, char **argv)
+{
+	int longindex = 0, opt, fd = -1;
+	int egress_ifindex = 0;
+	int ret = EXIT_SUCCESS;
+	int key = 0;
+
+	/* Parse commands line args */
+	while ((opt = getopt_long(argc, argv, "ho:",
+				  long_options, &longindex)) != -1) {
+		switch (opt) {
+		case 'o':
+			egress_ifindex = atoi(optarg);
+			break;
+		case 'h':
+		default:
+			usage(argv);
+			return EXIT_FAILURE;
+		}
+	}
+
+	fd = bpf_obj_get(mapfile);
+	if (fd < 0) {
+		fprintf(stderr, "ERROR: cannot open bpf_obj_get(%s): %s(%d)\n",
+			mapfile, strerror(errno), errno);
+		ret = EXIT_FAILURE;
+		goto out;
+	}
+
+	/* Only update/set egress port when set via cmdline */
+	if (egress_ifindex) {
+		ret = bpf_map_update_elem(fd, &key, &egress_ifindex, 0);
+		if (ret) {
+			perror("ERROR: bpf_map_update_elem");
+			ret = EXIT_FAILURE;
+			goto out;
+		}
+		if (verbose)
+			printf("Change egress redirect ifindex to: %d\n",
+			       egress_ifindex);
+	} else {
+		/* Read info from map */
+		ret = bpf_map_lookup_elem(fd, &key, &egress_ifindex);
+		if (ret) {
+			perror("ERROR: bpf_map_lookup_elem");
+			ret = EXIT_FAILURE;
+			goto out;
+		}
+		if (verbose)
+			printf("Current egress redirect ifindex: %d\n",
+			       egress_ifindex);
+	}
+out:
+	if (fd != -1)
+		close(fd);
+	return ret;
+}