xdp-bench: Add support for the redirect action with --load-mode load-bytes

xdp-bench/README.org

@@ -78,13 +78,12 @@ flushing it back out if writing).
 The default for this option is =no-touch=.
 
 ** -l, --load-mode <MODE>
-Specify which mechanism xdp-bench should use to load the packet data when
-parsing the IP header (used with =-p parse-ip=). The following modes are
-available:
+Specify which mechanism xdp-bench should use to load (and store) the packet data.
+The following modes are available:
 
 #+begin_src sh
  dpa		- Use traditional Direct Packet Access from the XDP program
- load-bytes		- Use the xdp_load_bytes() helper function to load the data
+ load-bytes	- Use the xdp_load_bytes() and xdp_store_bytes() helper functions
 #+end_src
 
 This can be used to benchmark the various packet access modes supported by the
@@ -154,13 +153,12 @@ following actions are available:
 The default for this option is =no-touch=.
 
 ** -l, --load-mode <MODE>
-Specify which mechanism xdp-bench should use to load the packet data when
-parsing the IP header (used with =-p parse-ip=). The following modes are
-available:
+Specify which mechanism xdp-bench should use to load (and store) the packet data.
+The following modes are available:
 
 #+begin_src sh
  dpa		- Use traditional Direct Packet Access from the XDP program
- load-bytes		- Use the xdp_load_bytes() helper function to load the data
+ load-bytes	- Use the xdp_load_bytes() and xdp_store_bytes() helper functions
 #+end_src
 
 This can be used to benchmark the various packet access modes supported by the
@@ -234,13 +232,12 @@ allows this function to be turned off.
 The default for this option is =swap-macs=.
 
 ** -l, --load-mode <MODE>
-Specify which mechanism xdp-bench should use to load the packet data when
-parsing the IP header (used with =-p parse-ip=). The following modes are
-available:
+Specify which mechanism xdp-bench should use to load (and store) the packet data.
+The following modes are available:
 
 #+begin_src sh
  dpa		- Use traditional Direct Packet Access from the XDP program
- load-bytes		- Use the xdp_load_bytes() helper function to load the data
+ load-bytes	- Use the xdp_load_bytes() and xdp_store_bytes() helper functions
 #+end_src
 
 This can be used to benchmark the various packet access modes supported by the
@@ -293,6 +290,20 @@ be redirect to the output interface =<ifname_out>=.
 
 The supported options are:
 
+** -l, --load-mode <MODE>
+Specify which mechanism xdp-bench should use to load (and store) the packet data.
+The following modes are available:
+
+#+begin_src sh
+ dpa		- Use traditional Direct Packet Access from the XDP program
+ load-bytes	- Use the xdp_load_bytes() and xdp_store_bytes() helper functions
+#+end_src
+
+This can be used to benchmark the various packet access modes supported by the
+kernel.
+
+The default for this option is =dpa=.
+
 ** -i, --interval <SECONDS>
 Set the polling interval for collecting all statistics and displaying them to
 the output. The unit of interval is in seconds.

xdp-bench/tests/test-xdp-bench.sh

@@ -35,10 +35,20 @@ test_xdp_load_bytes()
     export XDP_SAMPLE_IMMEDIATE_EXIT=1
 
     for action in drop pass tx; do
-        check_run $XDP_BENCH $action $NS -p parse-ip -l load-bytes -vv
+        check_run $XDP_BENCH $action $NS -l load-bytes -vv
         check_run $XDP_BENCH $action $NS -p read-data -l load-bytes -vv
+        check_run $XDP_BENCH $action $NS -p parse-ip -l load-bytes -vv
         check_run $XDP_BENCH $action $NS -p swap-macs -l load-bytes -vv
+        check_run $XDP_BENCH $action $NS -m skb -l load-bytes -vv
+        check_run $XDP_BENCH $action $NS -e -l load-bytes -vv
     done
+
+    check_run ip link add dev btest0 type veth peer name btest1
+    check_run $XDP_BENCH redirect btest0 btest1 -l load-bytes -vv
+    check_run $XDP_BENCH redirect btest0 btest1 -s -l load-bytes -vv
+    check_run $XDP_BENCH redirect btest0 btest1 -m skb -l load-bytes -vv
+    check_run $XDP_BENCH redirect btest0 btest1 -e -l load-bytes -vv
+    ip link del dev btest0
 }
 
 test_rxq_stats()

xdp-bench/xdp-bench.8

@@ -92,14 +92,13 @@ The default for this option is \fIno\-touch\fP.
 
 .SS "-l, --load-mode <MODE>"
 .PP
-Specify which mechanism xdp-bench should use to load the packet data when
-parsing the IP header (used with \fI\-p parse\-ip\fP). The following modes are
-available:
+Specify which mechanism xdp-bench should use to load (and store) the packet data.
+The following modes are available:
 
 .RS
 .nf
-\fCdpa		- Use traditional Direct Packet Access from the XDP program
-load-bytes		- Use the xdp_load_bytes() helper function to load the data
+\fCdpa			- Use traditional Direct Packet Access from the XDP program
+load-bytes	- Use the xdp_load_bytes() and xdp_store_bytes() helper functions
 \fP
 .fi
 .RE
@@ -191,14 +190,13 @@ The default for this option is \fIno\-touch\fP.
 
 .SS "-l, --load-mode <MODE>"
 .PP
-Specify which mechanism xdp-bench should use to load the packet data when
-parsing the IP header (used with \fI\-p parse\-ip\fP). The following modes are
-available:
+Specify which mechanism xdp-bench should use to load (and store) the packet data.
+The following modes are available:
 
 .RS
 .nf
-\fCdpa		- Use traditional Direct Packet Access from the XDP program
-load-bytes		- Use the xdp_load_bytes() helper function to load the data
+\fCdpa			- Use traditional Direct Packet Access from the XDP program
+load-bytes	- Use the xdp_load_bytes() and xdp_store_bytes() helper functions
 \fP
 .fi
 .RE
@@ -295,14 +293,13 @@ The default for this option is \fIswap\-macs\fP.
 
 .SS "-l, --load-mode <MODE>"
 .PP
-Specify which mechanism xdp-bench should use to load the packet data when
-parsing the IP header (used with \fI\-p parse\-ip\fP). The following modes are
-available:
+Specify which mechanism xdp-bench should use to load (and store) the packet data.
+The following modes are available:
 
 .RS
 .nf
-\fCdpa		- Use traditional Direct Packet Access from the XDP program
-load-bytes		- Use the xdp_load_bytes() helper function to load the data
+\fCdpa			- Use traditional Direct Packet Access from the XDP program
+load-bytes	- Use the xdp_load_bytes() and xdp_store_bytes() helper functions
 \fP
 .fi
 .RE
@@ -371,6 +368,26 @@ be redirect to the output interface \fI<ifname_out>\fP.
 .PP
 The supported options are:
 
+.SS "-l, --load-mode <MODE>"
+.PP
+Specify which mechanism xdp-bench should use to load (and store) the packet data.
+The following modes are available:
+
+.RS
+.nf
+\fCdpa			- Use traditional Direct Packet Access from the XDP program
+load-bytes	- Use the xdp_load_bytes() and xdp_store_bytes() helper functions
+\fP
+.fi
+.RE
+
+.PP
+This can be used to benchmark the various packet access modes supported by the
+kernel.
+
+.PP
+The default for this option is \fIdpa\fP.
+
 .SS "-i, --interval <SECONDS>"
 .PP
 Set the polling interval for collecting all statistics and displaying them to

xdp-bench/xdp-bench.c

@@ -82,7 +82,7 @@ struct prog_option basic_options[] = {
 		      .short_opt = 'l',
                       .metavar = "<mode>",
                       .typearg = basic_load_modes,
-		      .help = "How to load data when parsing IP header (with -p parse-ip; default dpa)"),
+		      .help = "How to load (and store) data; default dpa"),
 	DEFINE_OPTION("rxq-stats", OPT_BOOL, struct basic_opts, rxq_stats,
 		      .short_opt = 'r',
 		      .help = "Collect per-RXQ drop statistics"),
@@ -107,6 +107,11 @@ struct prog_option basic_options[] = {
 };
 
 struct prog_option redirect_basic_options[] = {
+	DEFINE_OPTION("load-mode", OPT_ENUM, struct redirect_opts, load_mode,
+		      .short_opt = 'l',
+                      .metavar = "<mode>",
+                      .typearg = basic_load_modes,
+		      .help = "How to load (and store) data; default dpa"),
 	DEFINE_OPTION("interval", OPT_U32, struct redirect_opts, interval,
 		      .short_opt = 'i',
 		      .metavar = "<seconds>",

xdp-bench/xdp-bench.h

@@ -43,6 +43,7 @@ struct redirect_opts {
 	bool extended;
 	__u32 interval;
 	enum xdp_attach_mode mode;
+	enum basic_load_mode load_mode;
 	struct iface iface_in;
 	struct iface iface_out;
 };

xdp-bench/xdp_redirect_basic.bpf.c

@@ -16,10 +16,15 @@
 #include <xdp/xdp_sample_common.bpf.h>
 #include <linux/if_ether.h>
 
+#ifndef HAVE_LIBBPF_BPF_PROGRAM__TYPE
+static long (*bpf_xdp_load_bytes)(struct xdp_md *xdp_md, __u32 offset, void *buf, __u32 len) = (void *) 189;
+static long (*bpf_xdp_store_bytes)(struct xdp_md *xdp_md, __u32 offset, void *buf, __u32 len) = (void *) 190;
+#endif
+
 const volatile int ifindex_out;
 
 SEC("xdp")
-int xdp_redirect_basic_prog(struct xdp_md *ctx)
+int xdp_redirect_prog(struct xdp_md *ctx)
 {
 	void *data_end = (void *)(long)ctx->data_end;
 	void *data = (void *)(long)ctx->data;
@@ -41,4 +46,30 @@ int xdp_redirect_basic_prog(struct xdp_md *ctx)
 	return bpf_redirect(ifindex_out, 0);
 }
 
+SEC("xdp")
+int xdp_redirect_load_bytes_prog(struct xdp_md *ctx)
+{
+	__u32 key = bpf_get_smp_processor_id();
+	int err, offset = 0;
+	struct datarec *rec;
+	struct ethhdr eth;
+
+	err = bpf_xdp_load_bytes(ctx, offset, &eth, sizeof(eth));
+	if (err)
+		return err;
+
+	rec = bpf_map_lookup_elem(&rx_cnt, &key);
+	if (!rec)
+		return XDP_PASS;
+	NO_TEAR_INC(rec->processed);
+
+	swap_src_dst_mac(&eth);
+
+	err = bpf_xdp_store_bytes(ctx, offset, &eth, sizeof(eth));
+	if (err)
+		return err;
+
+	return bpf_redirect(ifindex_out, 0);
+}
+
 char _license[] SEC("license") = "GPL";

xdp-bench/xdp_redirect_basic.c

@@ -39,6 +39,7 @@ int do_redirect_basic(const void *cfg, __unused const char *pin_root_path)
 
 	struct xdp_program *xdp_prog = NULL, *dummy_prog = NULL;
 	DECLARE_LIBBPF_OPTS(xdp_program_opts, opts);
+	struct bpf_program *prog = NULL;
 	struct xdp_redirect_basic *skel;
 	char str[2 * IF_NAMESIZE + 1];
 	int ret = EXIT_FAIL_OPTION;
@@ -73,8 +74,14 @@ int do_redirect_basic(const void *cfg, __unused const char *pin_root_path)
 	skel->rodata->to_match[0] = opt->iface_out.ifindex;
 	skel->rodata->ifindex_out = opt->iface_out.ifindex;
 
+	/* Make sure we only load the one XDP program we are interested in */
+	while ((prog = bpf_object__next_program(skel->obj, prog)) != NULL)
+		if (bpf_program__type(prog) == BPF_PROG_TYPE_XDP &&
+		    bpf_program__expected_attach_type(prog) == BPF_XDP)
+			bpf_program__set_autoload(prog, false);
+
 	opts.obj = skel->obj;
-	opts.prog_name = bpf_program__name(skel->progs.xdp_redirect_basic_prog);
+	opts.prog_name = (opt->load_mode == BASIC_LOAD_BYTES) ? "xdp_redirect_load_bytes_prog" : "xdp_redirect_prog";
 	xdp_prog = xdp_program__create(&opts);
 	if (!xdp_prog) {
 		ret = -errno;