From 5984dc6cb5aa6cce342a44f01f984fde09ed05b1 Mon Sep 17 00:00:00 2001
From: Prashant Bhole <prashantbhole.linux@gmail.com>
Date: Mon, 16 Dec 2019 16:16:19 +0900
Subject: samples/bpf: Reintroduce missed build targets

Add xdp_redirect and per_socket_stats_example in build targets.
They got removed from build targets in Makefile reorganization.

Fixes: 1d97c6c2511f ("samples/bpf: Base target programs rules on Makefile.target")
Signed-off-by: Prashant Bhole <prashantbhole.linux@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20191216071619.25479-1-prashantbhole.linux@gmail.com
---
 samples/bpf/Makefile | 2 ++
 1 file changed, 2 insertions(+)

(limited to 'samples')

diff --git a/samples/bpf/Makefile b/samples/bpf/Makefile
index 1fc42ad8ff49..8003d2823fa8 100644
--- a/samples/bpf/Makefile
+++ b/samples/bpf/Makefile
@@ -38,6 +38,8 @@ tprogs-y += tc_l2_redirect
 tprogs-y += lwt_len_hist
 tprogs-y += xdp_tx_iptunnel
 tprogs-y += test_map_in_map
+tprogs-y += per_socket_stats_example
+tprogs-y += xdp_redirect
 tprogs-y += xdp_redirect_map
 tprogs-y += xdp_redirect_cpu
 tprogs-y += xdp_monitor
-- 
cgit v1.2.3


From 5615ed472dc04e38a1affca69059f17c1178f770 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Toke=20H=C3=B8iland-J=C3=B8rgensen?= <toke@redhat.com>
Date: Mon, 16 Dec 2019 11:24:05 +0100
Subject: samples/bpf: Add missing -lz to TPROGS_LDLIBS
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Since libbpf now links against zlib, this needs to be included in the
linker invocation for the userspace programs in samples/bpf that link
statically against libbpf.

Fixes: 166750bc1dd2 ("libbpf: Support libbpf-provided extern variables")
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Tested-by: Björn Töpel <bjorn.topel@intel.com>
Link: https://lore.kernel.org/bpf/20191216102405.353834-1-toke@redhat.com
---
 samples/bpf/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'samples')

diff --git a/samples/bpf/Makefile b/samples/bpf/Makefile
index 8003d2823fa8..227077ae66ed 100644
--- a/samples/bpf/Makefile
+++ b/samples/bpf/Makefile
@@ -198,7 +198,7 @@ endif
 
 TPROGCFLAGS_bpf_load.o += -Wno-unused-variable
 
-TPROGS_LDLIBS			+= $(LIBBPF) -lelf
+TPROGS_LDLIBS			+= $(LIBBPF) -lelf -lz
 TPROGLDLIBS_tracex4		+= -lrt
 TPROGLDLIBS_trace_output	+= -lrt
 TPROGLDLIBS_map_perf_test	+= -lrt
-- 
cgit v1.2.3


From 450278977acbf494a20367c22fbb38729772d1fc Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Toke=20H=C3=B8iland-J=C3=B8rgensen?= <toke@redhat.com>
Date: Mon, 16 Dec 2019 11:38:19 +0100
Subject: samples/bpf: Set -fno-stack-protector when building BPF programs
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

It seems Clang can in some cases turn on stack protection by default, which
doesn't work with BPF. This was reported once before[0], but it seems the
flag to explicitly turn off the stack protector wasn't added to the
Makefile, so do that now.

The symptom of this is compile errors like the following:

error: <unknown>:0:0: in function bpf_prog1 i32 (%struct.__sk_buff*): A call to built-in function '__stack_chk_fail' is not supported.

[0] https://www.spinics.net/lists/netdev/msg556400.html

Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20191216103819.359535-1-toke@redhat.com
---
 samples/bpf/Makefile | 1 +
 1 file changed, 1 insertion(+)

(limited to 'samples')

diff --git a/samples/bpf/Makefile b/samples/bpf/Makefile
index 227077ae66ed..5b89c0370f33 100644
--- a/samples/bpf/Makefile
+++ b/samples/bpf/Makefile
@@ -236,6 +236,7 @@ BTF_LLVM_PROBE := $(shell echo "int main() { return 0; }" | \
 			  readelf -S ./llvm_btf_verify.o | grep BTF; \
 			  /bin/rm -f ./llvm_btf_verify.o)
 
+BPF_EXTRA_CFLAGS += -fno-stack-protector
 ifneq ($(BTF_LLVM_PROBE),)
 	BPF_EXTRA_CFLAGS += -g
 else
-- 
cgit v1.2.3


From d50ecc46d18fa19ccf06e0c4d2ee8a050c665e3d Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Toke=20H=C3=B8iland-J=C3=B8rgensen?= <toke@redhat.com>
Date: Mon, 16 Dec 2019 12:07:42 +0100
Subject: samples/bpf: Attach XDP programs in driver mode by default
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

When attaching XDP programs, userspace can set flags to request the attach
mode (generic/SKB mode, driver mode or hw offloaded mode). If no such flags
are requested, the kernel will attempt to attach in driver mode, and then
silently fall back to SKB mode if this fails.

The silent fallback is a major source of user confusion, as users will try
to load a program on a device without XDP support, and instead of an error
they will get the silent fallback behaviour, not notice, and then wonder
why performance is not what they were expecting.

In an attempt to combat this, let's switch all the samples to default to
explicitly requesting driver-mode attach. As part of this, ensure that all
the userspace utilities have a switch to enable SKB mode. For those that
have a switch to request driver mode, keep it but turn it into a no-op.

Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Jesper Dangaard Brouer <brouer@redhat.com>
Acked-by: David Ahern <dsahern@gmail.com>
Link: https://lore.kernel.org/bpf/20191216110742.364456-1-toke@redhat.com
---
 samples/bpf/xdp1_user.c             |  5 ++++-
 samples/bpf/xdp_adjust_tail_user.c  |  5 ++++-
 samples/bpf/xdp_fwd_user.c          | 17 ++++++++++++++---
 samples/bpf/xdp_redirect_cpu_user.c |  4 ++++
 samples/bpf/xdp_redirect_map_user.c |  5 ++++-
 samples/bpf/xdp_redirect_user.c     |  5 ++++-
 samples/bpf/xdp_router_ipv4_user.c  |  3 +++
 samples/bpf/xdp_rxq_info_user.c     |  4 ++++
 samples/bpf/xdp_sample_pkts_user.c  | 12 +++++++++---
 samples/bpf/xdp_tx_iptunnel_user.c  |  5 ++++-
 samples/bpf/xdpsock_user.c          |  5 ++++-
 11 files changed, 58 insertions(+), 12 deletions(-)

(limited to 'samples')

diff --git a/samples/bpf/xdp1_user.c b/samples/bpf/xdp1_user.c
index 3e553eed95a7..38a8852cb57f 100644
--- a/samples/bpf/xdp1_user.c
+++ b/samples/bpf/xdp1_user.c
@@ -98,7 +98,7 @@ int main(int argc, char **argv)
 			xdp_flags |= XDP_FLAGS_SKB_MODE;
 			break;
 		case 'N':
-			xdp_flags |= XDP_FLAGS_DRV_MODE;
+			/* default, set below */
 			break;
 		case 'F':
 			xdp_flags &= ~XDP_FLAGS_UPDATE_IF_NOEXIST;
@@ -109,6 +109,9 @@ int main(int argc, char **argv)
 		}
 	}
 
+	if (!(xdp_flags & XDP_FLAGS_SKB_MODE))
+		xdp_flags |= XDP_FLAGS_DRV_MODE;
+
 	if (optind == argc) {
 		usage(basename(argv[0]));
 		return 1;
diff --git a/samples/bpf/xdp_adjust_tail_user.c b/samples/bpf/xdp_adjust_tail_user.c
index d86e9ad0356b..008789eb6ada 100644
--- a/samples/bpf/xdp_adjust_tail_user.c
+++ b/samples/bpf/xdp_adjust_tail_user.c
@@ -120,7 +120,7 @@ int main(int argc, char **argv)
 			xdp_flags |= XDP_FLAGS_SKB_MODE;
 			break;
 		case 'N':
-			xdp_flags |= XDP_FLAGS_DRV_MODE;
+			/* default, set below */
 			break;
 		case 'F':
 			xdp_flags &= ~XDP_FLAGS_UPDATE_IF_NOEXIST;
@@ -132,6 +132,9 @@ int main(int argc, char **argv)
 		opt_flags[opt] = 0;
 	}
 
+	if (!(xdp_flags & XDP_FLAGS_SKB_MODE))
+		xdp_flags |= XDP_FLAGS_DRV_MODE;
+
 	for (i = 0; i < strlen(optstr); i++) {
 		if (opt_flags[(unsigned int)optstr[i]]) {
 			fprintf(stderr, "Missing argument -%c\n", optstr[i]);
diff --git a/samples/bpf/xdp_fwd_user.c b/samples/bpf/xdp_fwd_user.c
index 97ff1dad7669..c30f9acfdb84 100644
--- a/samples/bpf/xdp_fwd_user.c
+++ b/samples/bpf/xdp_fwd_user.c
@@ -27,11 +27,13 @@
 #include "libbpf.h"
 #include <bpf/bpf.h>
 
+static __u32 xdp_flags = XDP_FLAGS_UPDATE_IF_NOEXIST;
+
 static int do_attach(int idx, int prog_fd, int map_fd, const char *name)
 {
 	int err;
 
-	err = bpf_set_link_xdp_fd(idx, prog_fd, 0);
+	err = bpf_set_link_xdp_fd(idx, prog_fd, xdp_flags);
 	if (err < 0) {
 		printf("ERROR: failed to attach program to %s\n", name);
 		return err;
@@ -49,7 +51,7 @@ static int do_detach(int idx, const char *name)
 {
 	int err;
 
-	err = bpf_set_link_xdp_fd(idx, -1, 0);
+	err = bpf_set_link_xdp_fd(idx, -1, xdp_flags);
 	if (err < 0)
 		printf("ERROR: failed to detach program from %s\n", name);
 
@@ -83,11 +85,17 @@ int main(int argc, char **argv)
 	int attach = 1;
 	int ret = 0;
 
-	while ((opt = getopt(argc, argv, ":dD")) != -1) {
+	while ((opt = getopt(argc, argv, ":dDSF")) != -1) {
 		switch (opt) {
 		case 'd':
 			attach = 0;
 			break;
+		case 'S':
+			xdp_flags |= XDP_FLAGS_SKB_MODE;
+			break;
+		case 'F':
+			xdp_flags &= ~XDP_FLAGS_UPDATE_IF_NOEXIST;
+			break;
 		case 'D':
 			prog_name = "xdp_fwd_direct";
 			break;
@@ -97,6 +105,9 @@ int main(int argc, char **argv)
 		}
 	}
 
+	if (!(xdp_flags & XDP_FLAGS_SKB_MODE))
+		xdp_flags |= XDP_FLAGS_DRV_MODE;
+
 	if (optind == argc) {
 		usage(basename(argv[0]));
 		return 1;
diff --git a/samples/bpf/xdp_redirect_cpu_user.c b/samples/bpf/xdp_redirect_cpu_user.c
index 0da6e9e7132e..72af628529b5 100644
--- a/samples/bpf/xdp_redirect_cpu_user.c
+++ b/samples/bpf/xdp_redirect_cpu_user.c
@@ -728,6 +728,10 @@ int main(int argc, char **argv)
 			return EXIT_FAIL_OPTION;
 		}
 	}
+
+	if (!(xdp_flags & XDP_FLAGS_SKB_MODE))
+		xdp_flags |= XDP_FLAGS_DRV_MODE;
+
 	/* Required option */
 	if (ifindex == -1) {
 		fprintf(stderr, "ERR: required option --dev missing\n");
diff --git a/samples/bpf/xdp_redirect_map_user.c b/samples/bpf/xdp_redirect_map_user.c
index f70ee33907fd..cc840661faab 100644
--- a/samples/bpf/xdp_redirect_map_user.c
+++ b/samples/bpf/xdp_redirect_map_user.c
@@ -116,7 +116,7 @@ int main(int argc, char **argv)
 			xdp_flags |= XDP_FLAGS_SKB_MODE;
 			break;
 		case 'N':
-			xdp_flags |= XDP_FLAGS_DRV_MODE;
+			/* default, set below */
 			break;
 		case 'F':
 			xdp_flags &= ~XDP_FLAGS_UPDATE_IF_NOEXIST;
@@ -127,6 +127,9 @@ int main(int argc, char **argv)
 		}
 	}
 
+	if (!(xdp_flags & XDP_FLAGS_SKB_MODE))
+		xdp_flags |= XDP_FLAGS_DRV_MODE;
+
 	if (optind == argc) {
 		printf("usage: %s <IFNAME|IFINDEX>_IN <IFNAME|IFINDEX>_OUT\n", argv[0]);
 		return 1;
diff --git a/samples/bpf/xdp_redirect_user.c b/samples/bpf/xdp_redirect_user.c
index 5440cd620607..71dff8e3382a 100644
--- a/samples/bpf/xdp_redirect_user.c
+++ b/samples/bpf/xdp_redirect_user.c
@@ -117,7 +117,7 @@ int main(int argc, char **argv)
 			xdp_flags |= XDP_FLAGS_SKB_MODE;
 			break;
 		case 'N':
-			xdp_flags |= XDP_FLAGS_DRV_MODE;
+			/* default, set below */
 			break;
 		case 'F':
 			xdp_flags &= ~XDP_FLAGS_UPDATE_IF_NOEXIST;
@@ -128,6 +128,9 @@ int main(int argc, char **argv)
 		}
 	}
 
+	if (!(xdp_flags & XDP_FLAGS_SKB_MODE))
+		xdp_flags |= XDP_FLAGS_DRV_MODE;
+
 	if (optind == argc) {
 		printf("usage: %s <IFNAME|IFINDEX>_IN <IFNAME|IFINDEX>_OUT\n", argv[0]);
 		return 1;
diff --git a/samples/bpf/xdp_router_ipv4_user.c b/samples/bpf/xdp_router_ipv4_user.c
index 1469b66ebad1..fef286c5add2 100644
--- a/samples/bpf/xdp_router_ipv4_user.c
+++ b/samples/bpf/xdp_router_ipv4_user.c
@@ -662,6 +662,9 @@ int main(int ac, char **argv)
 		}
 	}
 
+	if (!(flags & XDP_FLAGS_SKB_MODE))
+		flags |= XDP_FLAGS_DRV_MODE;
+
 	if (optind == ac) {
 		usage(basename(argv[0]));
 		return 1;
diff --git a/samples/bpf/xdp_rxq_info_user.c b/samples/bpf/xdp_rxq_info_user.c
index 8fc3ad01de72..fc4983fd6959 100644
--- a/samples/bpf/xdp_rxq_info_user.c
+++ b/samples/bpf/xdp_rxq_info_user.c
@@ -551,6 +551,10 @@ int main(int argc, char **argv)
 			return EXIT_FAIL_OPTION;
 		}
 	}
+
+	if (!(xdp_flags & XDP_FLAGS_SKB_MODE))
+		xdp_flags |= XDP_FLAGS_DRV_MODE;
+
 	/* Required option */
 	if (ifindex == -1) {
 		fprintf(stderr, "ERR: required option --dev missing\n");
diff --git a/samples/bpf/xdp_sample_pkts_user.c b/samples/bpf/xdp_sample_pkts_user.c
index a5760e8bf2c4..8c1af1b7372d 100644
--- a/samples/bpf/xdp_sample_pkts_user.c
+++ b/samples/bpf/xdp_sample_pkts_user.c
@@ -52,13 +52,13 @@ static int do_detach(int idx, const char *name)
 	__u32 curr_prog_id = 0;
 	int err = 0;
 
-	err = bpf_get_link_xdp_id(idx, &curr_prog_id, 0);
+	err = bpf_get_link_xdp_id(idx, &curr_prog_id, xdp_flags);
 	if (err) {
 		printf("bpf_get_link_xdp_id failed\n");
 		return err;
 	}
 	if (prog_id == curr_prog_id) {
-		err = bpf_set_link_xdp_fd(idx, -1, 0);
+		err = bpf_set_link_xdp_fd(idx, -1, xdp_flags);
 		if (err < 0)
 			printf("ERROR: failed to detach prog from %s\n", name);
 	} else if (!curr_prog_id) {
@@ -115,7 +115,7 @@ int main(int argc, char **argv)
 		.prog_type	= BPF_PROG_TYPE_XDP,
 	};
 	struct perf_buffer_opts pb_opts = {};
-	const char *optstr = "F";
+	const char *optstr = "FS";
 	int prog_fd, map_fd, opt;
 	struct bpf_object *obj;
 	struct bpf_map *map;
@@ -127,12 +127,18 @@ int main(int argc, char **argv)
 		case 'F':
 			xdp_flags &= ~XDP_FLAGS_UPDATE_IF_NOEXIST;
 			break;
+		case 'S':
+			xdp_flags |= XDP_FLAGS_SKB_MODE;
+			break;
 		default:
 			usage(basename(argv[0]));
 			return 1;
 		}
 	}
 
+	if (!(xdp_flags & XDP_FLAGS_SKB_MODE))
+		xdp_flags |= XDP_FLAGS_DRV_MODE;
+
 	if (optind == argc) {
 		usage(basename(argv[0]));
 		return 1;
diff --git a/samples/bpf/xdp_tx_iptunnel_user.c b/samples/bpf/xdp_tx_iptunnel_user.c
index 2fe4c7f5ffe5..5f33b5530032 100644
--- a/samples/bpf/xdp_tx_iptunnel_user.c
+++ b/samples/bpf/xdp_tx_iptunnel_user.c
@@ -231,7 +231,7 @@ int main(int argc, char **argv)
 			xdp_flags |= XDP_FLAGS_SKB_MODE;
 			break;
 		case 'N':
-			xdp_flags |= XDP_FLAGS_DRV_MODE;
+			/* default, set below */
 			break;
 		case 'F':
 			xdp_flags &= ~XDP_FLAGS_UPDATE_IF_NOEXIST;
@@ -243,6 +243,9 @@ int main(int argc, char **argv)
 		opt_flags[opt] = 0;
 	}
 
+	if (!(xdp_flags & XDP_FLAGS_SKB_MODE))
+		xdp_flags |= XDP_FLAGS_DRV_MODE;
+
 	for (i = 0; i < strlen(optstr); i++) {
 		if (opt_flags[(unsigned int)optstr[i]]) {
 			fprintf(stderr, "Missing argument -%c\n", optstr[i]);
diff --git a/samples/bpf/xdpsock_user.c b/samples/bpf/xdpsock_user.c
index a15480010828..e7829e5baaff 100644
--- a/samples/bpf/xdpsock_user.c
+++ b/samples/bpf/xdpsock_user.c
@@ -440,7 +440,7 @@ static void parse_command_line(int argc, char **argv)
 			opt_xdp_bind_flags |= XDP_COPY;
 			break;
 		case 'N':
-			opt_xdp_flags |= XDP_FLAGS_DRV_MODE;
+			/* default, set below */
 			break;
 		case 'n':
 			opt_interval = atoi(optarg);
@@ -474,6 +474,9 @@ static void parse_command_line(int argc, char **argv)
 		}
 	}
 
+	if (!(opt_xdp_flags & XDP_FLAGS_SKB_MODE))
+		opt_xdp_flags |= XDP_FLAGS_DRV_MODE;
+
 	opt_ifindex = if_nametoindex(opt_if);
 	if (!opt_ifindex) {
 		fprintf(stderr, "ERROR: interface \"%s\" does not exist\n",
-- 
cgit v1.2.3


From d3f11b018f6ce278e683008e9c225fe87afc532d Mon Sep 17 00:00:00 2001
From: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
Date: Fri, 20 Dec 2019 14:25:25 +0530
Subject: samples/bpf: xdpsock: Add duration option to specify how long to run

The application now supports '-d' or '--duration' option to specify number of
seconds to run. This is used in tx, rx and l2fwd features. If this option is
not provided, the application runs forever.

Signed-off-by: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20191220085530.4980-2-jay.jayatheerthan@intel.com
---
 samples/bpf/xdpsock_user.c | 37 +++++++++++++++++++++++++++++++++++--
 1 file changed, 35 insertions(+), 2 deletions(-)

(limited to 'samples')

diff --git a/samples/bpf/xdpsock_user.c b/samples/bpf/xdpsock_user.c
index e7829e5baaff..e188a79a9c31 100644
--- a/samples/bpf/xdpsock_user.c
+++ b/samples/bpf/xdpsock_user.c
@@ -65,6 +65,9 @@ static u32 opt_xdp_flags = XDP_FLAGS_UPDATE_IF_NOEXIST;
 static const char *opt_if = "";
 static int opt_ifindex;
 static int opt_queue;
+static unsigned long opt_duration;
+static unsigned long start_time;
+static bool benchmark_done;
 static int opt_poll;
 static int opt_interval = 1;
 static u32 opt_xdp_bind_flags = XDP_USE_NEED_WAKEUP;
@@ -167,10 +170,21 @@ static void dump_stats(void)
 	}
 }
 
+static bool is_benchmark_done(void)
+{
+	if (opt_duration > 0) {
+		unsigned long dt = (get_nsecs() - start_time);
+
+		if (dt >= opt_duration)
+			benchmark_done = true;
+	}
+	return benchmark_done;
+}
+
 static void *poller(void *arg)
 {
 	(void)arg;
-	for (;;) {
+	while (!is_benchmark_done()) {
 		sleep(opt_interval);
 		dump_stats();
 	}
@@ -375,6 +389,7 @@ static struct option long_options[] = {
 	{"unaligned", no_argument, 0, 'u'},
 	{"shared-umem", no_argument, 0, 'M'},
 	{"force", no_argument, 0, 'F'},
+	{"duration", required_argument, 0, 'd'},
 	{0, 0, 0, 0}
 };
 
@@ -399,6 +414,8 @@ static void usage(const char *prog)
 		"  -u, --unaligned	Enable unaligned chunk placement\n"
 		"  -M, --shared-umem	Enable XDP_SHARED_UMEM\n"
 		"  -F, --force		Force loading the XDP prog\n"
+		"  -d, --duration=n	Duration in secs to run command.\n"
+		"			Default: forever.\n"
 		"\n";
 	fprintf(stderr, str, prog, XSK_UMEM__DEFAULT_FRAME_SIZE);
 	exit(EXIT_FAILURE);
@@ -411,7 +428,7 @@ static void parse_command_line(int argc, char **argv)
 	opterr = 0;
 
 	for (;;) {
-		c = getopt_long(argc, argv, "Frtli:q:psSNn:czf:muM",
+		c = getopt_long(argc, argv, "Frtli:q:psSNn:czf:muMd:",
 				long_options, &option_index);
 		if (c == -1)
 			break;
@@ -469,6 +486,10 @@ static void parse_command_line(int argc, char **argv)
 		case 'M':
 			opt_num_xsks = MAX_SOCKS;
 			break;
+		case 'd':
+			opt_duration = atoi(optarg);
+			opt_duration *= 1000000000;
+			break;
 		default:
 			usage(basename(argv[0]));
 		}
@@ -622,6 +643,9 @@ static void rx_drop_all(void)
 
 		for (i = 0; i < num_socks; i++)
 			rx_drop(xsks[i], fds);
+
+		if (benchmark_done)
+			break;
 	}
 }
 
@@ -671,6 +695,9 @@ static void tx_only_all(void)
 
 		for (i = 0; i < num_socks; i++)
 			tx_only(xsks[i], frame_nb[i]);
+
+		if (benchmark_done)
+			break;
 	}
 }
 
@@ -739,6 +766,9 @@ static void l2fwd_all(void)
 
 		for (i = 0; i < num_socks; i++)
 			l2fwd(xsks[i], fds);
+
+		if (benchmark_done)
+			break;
 	}
 }
 
@@ -852,6 +882,7 @@ int main(int argc, char **argv)
 		exit_with_error(ret);
 
 	prev_time = get_nsecs();
+	start_time = prev_time;
 
 	if (opt_bench == BENCH_RXDROP)
 		rx_drop_all();
@@ -860,5 +891,7 @@ int main(int argc, char **argv)
 	else
 		l2fwd_all();
 
+	pthread_join(pt, NULL);
+
 	return 0;
 }
-- 
cgit v1.2.3


From 695255882bdf63da240db33d0f2aa9ccca1cbe67 Mon Sep 17 00:00:00 2001
From: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
Date: Fri, 20 Dec 2019 14:25:26 +0530
Subject: samples/bpf: xdpsock: Use common code to handle signal and main exit

Add code to do cleanup for signals and application completion in a unified
fashion. The signal handler sets benckmark_done flag terminating the
threads. The cleanup is called before returning from main() function.

Signed-off-by: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20191220085530.4980-3-jay.jayatheerthan@intel.com
---
 samples/bpf/xdpsock_user.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

(limited to 'samples')

diff --git a/samples/bpf/xdpsock_user.c b/samples/bpf/xdpsock_user.c
index e188a79a9c31..7febc3d519a1 100644
--- a/samples/bpf/xdpsock_user.c
+++ b/samples/bpf/xdpsock_user.c
@@ -209,6 +209,11 @@ static void remove_xdp_program(void)
 }
 
 static void int_exit(int sig)
+{
+	benchmark_done = true;
+}
+
+static void xdpsock_cleanup(void)
 {
 	struct xsk_umem *umem = xsks[0]->umem->umem;
 	int i;
@@ -218,8 +223,6 @@ static void int_exit(int sig)
 		xsk_socket__delete(xsks[i]->xsk);
 	(void)xsk_umem__delete(umem);
 	remove_xdp_program();
-
-	exit(EXIT_SUCCESS);
 }
 
 static void __exit_with_error(int error, const char *file, const char *func,
@@ -893,5 +896,7 @@ int main(int argc, char **argv)
 
 	pthread_join(pt, NULL);
 
+	xdpsock_cleanup();
+
 	return 0;
 }
-- 
cgit v1.2.3


From cd9e72b6f21044b36a096833003811c2b2038455 Mon Sep 17 00:00:00 2001
From: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
Date: Fri, 20 Dec 2019 14:25:27 +0530
Subject: samples/bpf: xdpsock: Add option to specify batch size

New option to specify batch size for tx, rx and l2fwd has been added. This
allows fine tuning to maximize performance. It is specified using '-b' or
'--batch_size' options. When not specified default is 64.

Signed-off-by: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20191220085530.4980-4-jay.jayatheerthan@intel.com
---
 samples/bpf/xdpsock_user.c | 52 ++++++++++++++++++++++++++--------------------
 1 file changed, 29 insertions(+), 23 deletions(-)

(limited to 'samples')

diff --git a/samples/bpf/xdpsock_user.c b/samples/bpf/xdpsock_user.c
index 7febc3d519a1..1ba3e7142f39 100644
--- a/samples/bpf/xdpsock_user.c
+++ b/samples/bpf/xdpsock_user.c
@@ -45,7 +45,6 @@
 #endif
 
 #define NUM_FRAMES (4 * 1024)
-#define BATCH_SIZE 64
 
 #define DEBUG_HEXDUMP 0
 
@@ -68,6 +67,7 @@ static int opt_queue;
 static unsigned long opt_duration;
 static unsigned long start_time;
 static bool benchmark_done;
+static u32 opt_batch_size = 64;
 static int opt_poll;
 static int opt_interval = 1;
 static u32 opt_xdp_bind_flags = XDP_USE_NEED_WAKEUP;
@@ -237,7 +237,6 @@ static void __exit_with_error(int error, const char *file, const char *func,
 
 #define exit_with_error(error) __exit_with_error(error, __FILE__, __func__, \
 						 __LINE__)
-
 static const char pkt_data[] =
 	"\x3c\xfd\xfe\x9e\x7f\x71\xec\xb1\xd7\x98\x3a\xc0\x08\x00\x45\x00"
 	"\x00\x2e\x00\x00\x00\x00\x40\x11\x88\x97\x05\x08\x07\x08\xc8\x14"
@@ -291,11 +290,10 @@ static void hex_dump(void *pkt, size_t length, u64 addr)
 	printf("\n");
 }
 
-static size_t gen_eth_frame(struct xsk_umem_info *umem, u64 addr)
+static void gen_eth_frame(struct xsk_umem_info *umem, u64 addr)
 {
 	memcpy(xsk_umem__get_data(umem->buffer, addr), pkt_data,
 	       sizeof(pkt_data) - 1);
-	return sizeof(pkt_data) - 1;
 }
 
 static struct xsk_umem_info *xsk_configure_umem(void *buffer, u64 size)
@@ -393,6 +391,7 @@ static struct option long_options[] = {
 	{"shared-umem", no_argument, 0, 'M'},
 	{"force", no_argument, 0, 'F'},
 	{"duration", required_argument, 0, 'd'},
+	{"batch-size", required_argument, 0, 'b'},
 	{0, 0, 0, 0}
 };
 
@@ -419,8 +418,11 @@ static void usage(const char *prog)
 		"  -F, --force		Force loading the XDP prog\n"
 		"  -d, --duration=n	Duration in secs to run command.\n"
 		"			Default: forever.\n"
+		"  -b, --batch-size=n	Batch size for sending or receiving\n"
+		"			packets. Default: %d\n"
 		"\n";
-	fprintf(stderr, str, prog, XSK_UMEM__DEFAULT_FRAME_SIZE);
+	fprintf(stderr, str, prog, XSK_UMEM__DEFAULT_FRAME_SIZE,
+		opt_batch_size);
 	exit(EXIT_FAILURE);
 }
 
@@ -431,7 +433,7 @@ static void parse_command_line(int argc, char **argv)
 	opterr = 0;
 
 	for (;;) {
-		c = getopt_long(argc, argv, "Frtli:q:psSNn:czf:muMd:",
+		c = getopt_long(argc, argv, "Frtli:q:pSNn:czf:muMd:b:",
 				long_options, &option_index);
 		if (c == -1)
 			break;
@@ -493,6 +495,9 @@ static void parse_command_line(int argc, char **argv)
 			opt_duration = atoi(optarg);
 			opt_duration *= 1000000000;
 			break;
+		case 'b':
+			opt_batch_size = atoi(optarg);
+			break;
 		default:
 			usage(basename(argv[0]));
 		}
@@ -540,7 +545,7 @@ static inline void complete_tx_l2fwd(struct xsk_socket_info *xsk,
 	if (!opt_need_wakeup || xsk_ring_prod__needs_wakeup(&xsk->tx))
 		kick_tx(xsk);
 
-	ndescs = (xsk->outstanding_tx > BATCH_SIZE) ? BATCH_SIZE :
+	ndescs = (xsk->outstanding_tx > opt_batch_size) ? opt_batch_size :
 		xsk->outstanding_tx;
 
 	/* re-add completed Tx buffers */
@@ -580,7 +585,7 @@ static inline void complete_tx_only(struct xsk_socket_info *xsk)
 	if (!opt_need_wakeup || xsk_ring_prod__needs_wakeup(&xsk->tx))
 		kick_tx(xsk);
 
-	rcvd = xsk_ring_cons__peek(&xsk->umem->cq, BATCH_SIZE, &idx);
+	rcvd = xsk_ring_cons__peek(&xsk->umem->cq, opt_batch_size, &idx);
 	if (rcvd > 0) {
 		xsk_ring_cons__release(&xsk->umem->cq, rcvd);
 		xsk->outstanding_tx -= rcvd;
@@ -594,7 +599,7 @@ static void rx_drop(struct xsk_socket_info *xsk, struct pollfd *fds)
 	u32 idx_rx = 0, idx_fq = 0;
 	int ret;
 
-	rcvd = xsk_ring_cons__peek(&xsk->rx, BATCH_SIZE, &idx_rx);
+	rcvd = xsk_ring_cons__peek(&xsk->rx, opt_batch_size, &idx_rx);
 	if (!rcvd) {
 		if (xsk_ring_prod__needs_wakeup(&xsk->umem->fq))
 			ret = poll(fds, num_socks, opt_timeout);
@@ -655,23 +660,24 @@ static void rx_drop_all(void)
 static void tx_only(struct xsk_socket_info *xsk, u32 frame_nb)
 {
 	u32 idx;
+	unsigned int i;
 
-	if (xsk_ring_prod__reserve(&xsk->tx, BATCH_SIZE, &idx) == BATCH_SIZE) {
-		unsigned int i;
-
-		for (i = 0; i < BATCH_SIZE; i++) {
-			xsk_ring_prod__tx_desc(&xsk->tx, idx + i)->addr	=
-				(frame_nb + i) << XSK_UMEM__DEFAULT_FRAME_SHIFT;
-			xsk_ring_prod__tx_desc(&xsk->tx, idx + i)->len =
-				sizeof(pkt_data) - 1;
-		}
+	while (xsk_ring_prod__reserve(&xsk->tx, opt_batch_size, &idx) <
+				      opt_batch_size) {
+		complete_tx_only(xsk);
+	}
 
-		xsk_ring_prod__submit(&xsk->tx, BATCH_SIZE);
-		xsk->outstanding_tx += BATCH_SIZE;
-		frame_nb += BATCH_SIZE;
-		frame_nb %= NUM_FRAMES;
+	for (i = 0; i < opt_batch_size; i++) {
+		struct xdp_desc *tx_desc = xsk_ring_prod__tx_desc(&xsk->tx,
+								  idx + i);
+		tx_desc->addr = (frame_nb + i) << XSK_UMEM__DEFAULT_FRAME_SHIFT;
+		tx_desc->len = sizeof(pkt_data) - 1;
 	}
 
+	xsk_ring_prod__submit(&xsk->tx, opt_batch_size);
+	xsk->outstanding_tx += opt_batch_size;
+	frame_nb += opt_batch_size;
+	frame_nb %= NUM_FRAMES;
 	complete_tx_only(xsk);
 }
 
@@ -712,7 +718,7 @@ static void l2fwd(struct xsk_socket_info *xsk, struct pollfd *fds)
 
 	complete_tx_l2fwd(xsk, fds);
 
-	rcvd = xsk_ring_cons__peek(&xsk->rx, BATCH_SIZE, &idx_rx);
+	rcvd = xsk_ring_cons__peek(&xsk->rx, opt_batch_size, &idx_rx);
 	if (!rcvd) {
 		if (xsk_ring_prod__needs_wakeup(&xsk->umem->fq))
 			ret = poll(fds, num_socks, opt_timeout);
-- 
cgit v1.2.3


From ece6e9694751a4f0b99372724a0705a0217132b3 Mon Sep 17 00:00:00 2001
From: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
Date: Fri, 20 Dec 2019 14:25:28 +0530
Subject: samples/bpf: xdpsock: Add option to specify number of packets to send

Use '-C' or '--tx-pkt-count' to specify number of packets to send.
If it is not specified, the application sends packets forever. If packet
count is not a multiple of batch size, last batch sent is less than the
batch size.

Signed-off-by: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20191220085530.4980-5-jay.jayatheerthan@intel.com
---
 samples/bpf/xdpsock_user.c | 73 +++++++++++++++++++++++++++++++++++++---------
 1 file changed, 59 insertions(+), 14 deletions(-)

(limited to 'samples')

diff --git a/samples/bpf/xdpsock_user.c b/samples/bpf/xdpsock_user.c
index 1ba3e7142f39..f96ce3055d46 100644
--- a/samples/bpf/xdpsock_user.c
+++ b/samples/bpf/xdpsock_user.c
@@ -68,6 +68,7 @@ static unsigned long opt_duration;
 static unsigned long start_time;
 static bool benchmark_done;
 static u32 opt_batch_size = 64;
+static int opt_pkt_count;
 static int opt_poll;
 static int opt_interval = 1;
 static u32 opt_xdp_bind_flags = XDP_USE_NEED_WAKEUP;
@@ -392,6 +393,7 @@ static struct option long_options[] = {
 	{"force", no_argument, 0, 'F'},
 	{"duration", required_argument, 0, 'd'},
 	{"batch-size", required_argument, 0, 'b'},
+	{"tx-pkt-count", required_argument, 0, 'C'},
 	{0, 0, 0, 0}
 };
 
@@ -420,6 +422,8 @@ static void usage(const char *prog)
 		"			Default: forever.\n"
 		"  -b, --batch-size=n	Batch size for sending or receiving\n"
 		"			packets. Default: %d\n"
+		"  -C, --tx-pkt-count=n	Number of packets to send.\n"
+		"			Default: Continuous packets.\n"
 		"\n";
 	fprintf(stderr, str, prog, XSK_UMEM__DEFAULT_FRAME_SIZE,
 		opt_batch_size);
@@ -433,7 +437,7 @@ static void parse_command_line(int argc, char **argv)
 	opterr = 0;
 
 	for (;;) {
-		c = getopt_long(argc, argv, "Frtli:q:pSNn:czf:muMd:b:",
+		c = getopt_long(argc, argv, "Frtli:q:pSNn:czf:muMd:b:C:",
 				long_options, &option_index);
 		if (c == -1)
 			break;
@@ -498,6 +502,9 @@ static void parse_command_line(int argc, char **argv)
 		case 'b':
 			opt_batch_size = atoi(optarg);
 			break;
+		case 'C':
+			opt_pkt_count = atoi(optarg);
+			break;
 		default:
 			usage(basename(argv[0]));
 		}
@@ -574,7 +581,8 @@ static inline void complete_tx_l2fwd(struct xsk_socket_info *xsk,
 	}
 }
 
-static inline void complete_tx_only(struct xsk_socket_info *xsk)
+static inline void complete_tx_only(struct xsk_socket_info *xsk,
+				    int batch_size)
 {
 	unsigned int rcvd;
 	u32 idx;
@@ -585,7 +593,7 @@ static inline void complete_tx_only(struct xsk_socket_info *xsk)
 	if (!opt_need_wakeup || xsk_ring_prod__needs_wakeup(&xsk->tx))
 		kick_tx(xsk);
 
-	rcvd = xsk_ring_cons__peek(&xsk->umem->cq, opt_batch_size, &idx);
+	rcvd = xsk_ring_cons__peek(&xsk->umem->cq, batch_size, &idx);
 	if (rcvd > 0) {
 		xsk_ring_cons__release(&xsk->umem->cq, rcvd);
 		xsk->outstanding_tx -= rcvd;
@@ -657,34 +665,62 @@ static void rx_drop_all(void)
 	}
 }
 
-static void tx_only(struct xsk_socket_info *xsk, u32 frame_nb)
+static void tx_only(struct xsk_socket_info *xsk, u32 frame_nb, int batch_size)
 {
 	u32 idx;
 	unsigned int i;
 
-	while (xsk_ring_prod__reserve(&xsk->tx, opt_batch_size, &idx) <
-				      opt_batch_size) {
-		complete_tx_only(xsk);
+	while (xsk_ring_prod__reserve(&xsk->tx, batch_size, &idx) <
+				      batch_size) {
+		complete_tx_only(xsk, batch_size);
 	}
 
-	for (i = 0; i < opt_batch_size; i++) {
+	for (i = 0; i < batch_size; i++) {
 		struct xdp_desc *tx_desc = xsk_ring_prod__tx_desc(&xsk->tx,
 								  idx + i);
 		tx_desc->addr = (frame_nb + i) << XSK_UMEM__DEFAULT_FRAME_SHIFT;
 		tx_desc->len = sizeof(pkt_data) - 1;
 	}
 
-	xsk_ring_prod__submit(&xsk->tx, opt_batch_size);
-	xsk->outstanding_tx += opt_batch_size;
-	frame_nb += opt_batch_size;
+	xsk_ring_prod__submit(&xsk->tx, batch_size);
+	xsk->outstanding_tx += batch_size;
+	frame_nb += batch_size;
 	frame_nb %= NUM_FRAMES;
-	complete_tx_only(xsk);
+	complete_tx_only(xsk, batch_size);
+}
+
+static inline int get_batch_size(int pkt_cnt)
+{
+	if (!opt_pkt_count)
+		return opt_batch_size;
+
+	if (pkt_cnt + opt_batch_size <= opt_pkt_count)
+		return opt_batch_size;
+
+	return opt_pkt_count - pkt_cnt;
+}
+
+static void complete_tx_only_all(void)
+{
+	bool pending;
+	int i;
+
+	do {
+		pending = false;
+		for (i = 0; i < num_socks; i++) {
+			if (xsks[i]->outstanding_tx) {
+				complete_tx_only(xsks[i], opt_batch_size);
+				pending = !!xsks[i]->outstanding_tx;
+			}
+		}
+	} while (pending);
 }
 
 static void tx_only_all(void)
 {
 	struct pollfd fds[MAX_SOCKS] = {};
 	u32 frame_nb[MAX_SOCKS] = {};
+	int pkt_cnt = 0;
 	int i, ret;
 
 	for (i = 0; i < num_socks; i++) {
@@ -692,7 +728,9 @@ static void tx_only_all(void)
 		fds[0].events = POLLOUT;
 	}
 
-	for (;;) {
+	while ((opt_pkt_count && pkt_cnt < opt_pkt_count) || !opt_pkt_count) {
+		int batch_size = get_batch_size(pkt_cnt);
+
 		if (opt_poll) {
 			ret = poll(fds, num_socks, opt_timeout);
 			if (ret <= 0)
@@ -703,11 +741,16 @@ static void tx_only_all(void)
 		}
 
 		for (i = 0; i < num_socks; i++)
-			tx_only(xsks[i], frame_nb[i]);
+			tx_only(xsks[i], frame_nb[i], batch_size);
+
+		pkt_cnt += batch_size;
 
 		if (benchmark_done)
 			break;
 	}
+
+	if (opt_pkt_count)
+		complete_tx_only_all();
 }
 
 static void l2fwd(struct xsk_socket_info *xsk, struct pollfd *fds)
@@ -900,6 +943,8 @@ int main(int argc, char **argv)
 	else
 		l2fwd_all();
 
+	benchmark_done = true;
+
 	pthread_join(pt, NULL);
 
 	xdpsock_cleanup();
-- 
cgit v1.2.3


From 4a3c23ae3acc6185a9d418830f22672a52ff986a Mon Sep 17 00:00:00 2001
From: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
Date: Fri, 20 Dec 2019 14:25:29 +0530
Subject: samples/bpf: xdpsock: Add option to specify tx packet size

New option '-s' or '--tx-pkt-size' has been added to specify the transmit
packet size. The packet size ranges from 47 to 4096 bytes. When this
option is not provided, it defaults to 64 byte packet.

The code uses struct ethhdr, struct iphdr and struct udphdr to form the
transmit packet. The MAC address, IP address and UDP ports are set to default
values.

The code calculates IP and UDP checksums before sending the packet.
Checksum calculation code in Linux kernel is used for this purpose.
The Ethernet FCS is not filled by the code.

Signed-off-by: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20191220085530.4980-6-jay.jayatheerthan@intel.com
---
 samples/bpf/xdpsock_user.c | 276 +++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 265 insertions(+), 11 deletions(-)

(limited to 'samples')

diff --git a/samples/bpf/xdpsock_user.c b/samples/bpf/xdpsock_user.c
index f96ce3055d46..2297158a32bd 100644
--- a/samples/bpf/xdpsock_user.c
+++ b/samples/bpf/xdpsock_user.c
@@ -10,6 +10,9 @@
 #include <linux/if_link.h>
 #include <linux/if_xdp.h>
 #include <linux/if_ether.h>
+#include <linux/ip.h>
+#include <linux/udp.h>
+#include <arpa/inet.h>
 #include <locale.h>
 #include <net/ethernet.h>
 #include <net/if.h>
@@ -45,11 +48,14 @@
 #endif
 
 #define NUM_FRAMES (4 * 1024)
+#define MIN_PKT_SIZE 64
 
 #define DEBUG_HEXDUMP 0
 
 typedef __u64 u64;
 typedef __u32 u32;
+typedef __u16 u16;
+typedef __u8  u8;
 
 static unsigned long prev_time;
 
@@ -69,6 +75,8 @@ static unsigned long start_time;
 static bool benchmark_done;
 static u32 opt_batch_size = 64;
 static int opt_pkt_count;
+static u16 opt_pkt_size = MIN_PKT_SIZE;
+static u32 pkt_fill_pattern = 0x12345678;
 static int opt_poll;
 static int opt_interval = 1;
 static u32 opt_xdp_bind_flags = XDP_USE_NEED_WAKEUP;
@@ -238,12 +246,6 @@ static void __exit_with_error(int error, const char *file, const char *func,
 
 #define exit_with_error(error) __exit_with_error(error, __FILE__, __func__, \
 						 __LINE__)
-static const char pkt_data[] =
-	"\x3c\xfd\xfe\x9e\x7f\x71\xec\xb1\xd7\x98\x3a\xc0\x08\x00\x45\x00"
-	"\x00\x2e\x00\x00\x00\x00\x40\x11\x88\x97\x05\x08\x07\x08\xc8\x14"
-	"\x1e\x04\x10\x92\x10\x92\x00\x1a\x6d\xa3\x34\x33\x1f\x69\x40\x6b"
-	"\x54\x59\xb6\x14\x2d\x11\x44\xbf\xaf\xd9\xbe\xaa";
-
 static void swap_mac_addresses(void *data)
 {
 	struct ether_header *eth = (struct ether_header *)data;
@@ -291,10 +293,243 @@ static void hex_dump(void *pkt, size_t length, u64 addr)
 	printf("\n");
 }
 
+static void *memset32_htonl(void *dest, u32 val, u32 size)
+{
+	u32 *ptr = (u32 *)dest;
+	int i;
+
+	val = htonl(val);
+
+	for (i = 0; i < (size & (~0x3)); i += 4)
+		ptr[i >> 2] = val;
+
+	for (; i < size; i++)
+		((char *)dest)[i] = ((char *)&val)[i & 3];
+
+	return dest;
+}
+
+/*
+ * This function code has been taken from
+ * Linux kernel lib/checksum.c
+ */
+static inline unsigned short from32to16(unsigned int x)
+{
+	/* add up 16-bit and 16-bit for 16+c bit */
+	x = (x & 0xffff) + (x >> 16);
+	/* add up carry.. */
+	x = (x & 0xffff) + (x >> 16);
+	return x;
+}
+
+/*
+ * This function code has been taken from
+ * Linux kernel lib/checksum.c
+ */
+static unsigned int do_csum(const unsigned char *buff, int len)
+{
+	unsigned int result = 0;
+	int odd;
+
+	if (len <= 0)
+		goto out;
+	odd = 1 & (unsigned long)buff;
+	if (odd) {
+#ifdef __LITTLE_ENDIAN
+		result += (*buff << 8);
+#else
+		result = *buff;
+#endif
+		len--;
+		buff++;
+	}
+	if (len >= 2) {
+		if (2 & (unsigned long)buff) {
+			result += *(unsigned short *)buff;
+			len -= 2;
+			buff += 2;
+		}
+		if (len >= 4) {
+			const unsigned char *end = buff +
+						   ((unsigned int)len & ~3);
+			unsigned int carry = 0;
+
+			do {
+				unsigned int w = *(unsigned int *)buff;
+
+				buff += 4;
+				result += carry;
+				result += w;
+				carry = (w > result);
+			} while (buff < end);
+			result += carry;
+			result = (result & 0xffff) + (result >> 16);
+		}
+		if (len & 2) {
+			result += *(unsigned short *)buff;
+			buff += 2;
+		}
+	}
+	if (len & 1)
+#ifdef __LITTLE_ENDIAN
+		result += *buff;
+#else
+		result += (*buff << 8);
+#endif
+	result = from32to16(result);
+	if (odd)
+		result = ((result >> 8) & 0xff) | ((result & 0xff) << 8);
+out:
+	return result;
+}
+
+__sum16 ip_fast_csum(const void *iph, unsigned int ihl);
+
+/*
+ *	This is a version of ip_compute_csum() optimized for IP headers,
+ *	which always checksum on 4 octet boundaries.
+ *	This function code has been taken from
+ *	Linux kernel lib/checksum.c
+ */
+__sum16 ip_fast_csum(const void *iph, unsigned int ihl)
+{
+	return (__force __sum16)~do_csum(iph, ihl * 4);
+}
+
+/*
+ * Fold a partial checksum
+ * This function code has been taken from
+ * Linux kernel include/asm-generic/checksum.h
+ */
+static inline __sum16 csum_fold(__wsum csum)
+{
+	u32 sum = (__force u32)csum;
+
+	sum = (sum & 0xffff) + (sum >> 16);
+	sum = (sum & 0xffff) + (sum >> 16);
+	return (__force __sum16)~sum;
+}
+
+/*
+ * This function code has been taken from
+ * Linux kernel lib/checksum.c
+ */
+static inline u32 from64to32(u64 x)
+{
+	/* add up 32-bit and 32-bit for 32+c bit */
+	x = (x & 0xffffffff) + (x >> 32);
+	/* add up carry.. */
+	x = (x & 0xffffffff) + (x >> 32);
+	return (u32)x;
+}
+
+__wsum csum_tcpudp_nofold(__be32 saddr, __be32 daddr,
+			  __u32 len, __u8 proto, __wsum sum);
+
+/*
+ * This function code has been taken from
+ * Linux kernel lib/checksum.c
+ */
+__wsum csum_tcpudp_nofold(__be32 saddr, __be32 daddr,
+			  __u32 len, __u8 proto, __wsum sum)
+{
+	unsigned long long s = (__force u32)sum;
+
+	s += (__force u32)saddr;
+	s += (__force u32)daddr;
+#ifdef __BIG_ENDIAN__
+	s += proto + len;
+#else
+	s += (proto + len) << 8;
+#endif
+	return (__force __wsum)from64to32(s);
+}
+
+/*
+ * This function has been taken from
+ * Linux kernel include/asm-generic/checksum.h
+ */
+static inline __sum16
+csum_tcpudp_magic(__be32 saddr, __be32 daddr, __u32 len,
+		  __u8 proto, __wsum sum)
+{
+	return csum_fold(csum_tcpudp_nofold(saddr, daddr, len, proto, sum));
+}
+
+static inline u16 udp_csum(u32 saddr, u32 daddr, u32 len,
+			   u8 proto, u16 *udp_pkt)
+{
+	u32 csum = 0;
+	u32 cnt = 0;
+
+	/* udp hdr and data */
+	for (; cnt < len; cnt += 2)
+		csum += udp_pkt[cnt >> 1];
+
+	return csum_tcpudp_magic(saddr, daddr, len, proto, csum);
+}
+
+#define ETH_FCS_SIZE 4
+
+#define PKT_HDR_SIZE (sizeof(struct ethhdr) + sizeof(struct iphdr) + \
+		      sizeof(struct udphdr))
+
+#define PKT_SIZE		(opt_pkt_size - ETH_FCS_SIZE)
+#define IP_PKT_SIZE		(PKT_SIZE - sizeof(struct ethhdr))
+#define UDP_PKT_SIZE		(IP_PKT_SIZE - sizeof(struct iphdr))
+#define UDP_PKT_DATA_SIZE	(UDP_PKT_SIZE - sizeof(struct udphdr))
+
+static u8 pkt_data[XSK_UMEM__DEFAULT_FRAME_SIZE];
+
+static void gen_eth_hdr_data(void)
+{
+	struct udphdr *udp_hdr = (struct udphdr *)(pkt_data +
+						   sizeof(struct ethhdr) +
+						   sizeof(struct iphdr));
+	struct iphdr *ip_hdr = (struct iphdr *)(pkt_data +
+						sizeof(struct ethhdr));
+	struct ethhdr *eth_hdr = (struct ethhdr *)pkt_data;
+
+	/* ethernet header */
+	memcpy(eth_hdr->h_dest, "\x3c\xfd\xfe\x9e\x7f\x71", ETH_ALEN);
+	memcpy(eth_hdr->h_source, "\xec\xb1\xd7\x98\x3a\xc0", ETH_ALEN);
+	eth_hdr->h_proto = htons(ETH_P_IP);
+
+	/* IP header */
+	ip_hdr->version = IPVERSION;
+	ip_hdr->ihl = 0x5; /* 20 byte header */
+	ip_hdr->tos = 0x0;
+	ip_hdr->tot_len = htons(IP_PKT_SIZE);
+	ip_hdr->id = 0;
+	ip_hdr->frag_off = 0;
+	ip_hdr->ttl = IPDEFTTL;
+	ip_hdr->protocol = IPPROTO_UDP;
+	ip_hdr->saddr = htonl(0x0a0a0a10);
+	ip_hdr->daddr = htonl(0x0a0a0a20);
+
+	/* IP header checksum */
+	ip_hdr->check = 0;
+	ip_hdr->check = ip_fast_csum((const void *)ip_hdr, ip_hdr->ihl);
+
+	/* UDP header */
+	udp_hdr->source = htons(0x1000);
+	udp_hdr->dest = htons(0x1000);
+	udp_hdr->len = htons(UDP_PKT_SIZE);
+
+	/* UDP data */
+	memset32_htonl(pkt_data + PKT_HDR_SIZE, pkt_fill_pattern,
+		       UDP_PKT_DATA_SIZE);
+
+	/* UDP header checksum */
+	udp_hdr->check = 0;
+	udp_hdr->check = udp_csum(ip_hdr->saddr, ip_hdr->daddr, UDP_PKT_SIZE,
+				  IPPROTO_UDP, (u16 *)udp_hdr);
+}
+
 static void gen_eth_frame(struct xsk_umem_info *umem, u64 addr)
 {
 	memcpy(xsk_umem__get_data(umem->buffer, addr), pkt_data,
-	       sizeof(pkt_data) - 1);
+	       PKT_SIZE);
 }
 
 static struct xsk_umem_info *xsk_configure_umem(void *buffer, u64 size)
@@ -394,6 +629,7 @@ static struct option long_options[] = {
 	{"duration", required_argument, 0, 'd'},
 	{"batch-size", required_argument, 0, 'b'},
 	{"tx-pkt-count", required_argument, 0, 'C'},
+	{"tx-pkt-size", required_argument, 0, 's'},
 	{0, 0, 0, 0}
 };
 
@@ -424,9 +660,14 @@ static void usage(const char *prog)
 		"			packets. Default: %d\n"
 		"  -C, --tx-pkt-count=n	Number of packets to send.\n"
 		"			Default: Continuous packets.\n"
+		"  -s, --tx-pkt-size=n	Transmit packet size.\n"
+		"			(Default: %d bytes)\n"
+		"			Min size: %d, Max size %d.\n"
 		"\n";
 	fprintf(stderr, str, prog, XSK_UMEM__DEFAULT_FRAME_SIZE,
-		opt_batch_size);
+		opt_batch_size, MIN_PKT_SIZE, MIN_PKT_SIZE,
+		XSK_UMEM__DEFAULT_FRAME_SIZE);
+
 	exit(EXIT_FAILURE);
 }
 
@@ -437,7 +678,7 @@ static void parse_command_line(int argc, char **argv)
 	opterr = 0;
 
 	for (;;) {
-		c = getopt_long(argc, argv, "Frtli:q:pSNn:czf:muMd:b:C:",
+		c = getopt_long(argc, argv, "Frtli:q:pSNn:czf:muMd:b:C:s:",
 				long_options, &option_index);
 		if (c == -1)
 			break;
@@ -505,6 +746,16 @@ static void parse_command_line(int argc, char **argv)
 		case 'C':
 			opt_pkt_count = atoi(optarg);
 			break;
+		case 's':
+			opt_pkt_size = atoi(optarg);
+			if (opt_pkt_size > (XSK_UMEM__DEFAULT_FRAME_SIZE) ||
+			    opt_pkt_size < MIN_PKT_SIZE) {
+				fprintf(stderr,
+					"ERROR: Invalid frame size %d\n",
+					opt_pkt_size);
+				usage(basename(argv[0]));
+			}
+			break;
 		default:
 			usage(basename(argv[0]));
 		}
@@ -679,7 +930,7 @@ static void tx_only(struct xsk_socket_info *xsk, u32 frame_nb, int batch_size)
 		struct xdp_desc *tx_desc = xsk_ring_prod__tx_desc(&xsk->tx,
 								  idx + i);
 		tx_desc->addr = (frame_nb + i) << XSK_UMEM__DEFAULT_FRAME_SHIFT;
-		tx_desc->len = sizeof(pkt_data) - 1;
+		tx_desc->len = PKT_SIZE;
 	}
 
 	xsk_ring_prod__submit(&xsk->tx, batch_size);
@@ -916,9 +1167,12 @@ int main(int argc, char **argv)
 	for (i = 0; i < opt_num_xsks; i++)
 		xsks[num_socks++] = xsk_configure_socket(umem, rx, tx);
 
-	if (opt_bench == BENCH_TXONLY)
+	if (opt_bench == BENCH_TXONLY) {
+		gen_eth_hdr_data();
+
 		for (i = 0; i < NUM_FRAMES; i++)
 			gen_eth_frame(umem, i * opt_xsk_frame_size);
+	}
 
 	if (opt_num_xsks > 1 && opt_bench != BENCH_TXONLY)
 		enter_xsks_into_map(obj);
-- 
cgit v1.2.3


From 46e3268eaaca9f8a0f145872b96fe6d54a232890 Mon Sep 17 00:00:00 2001
From: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
Date: Fri, 20 Dec 2019 14:25:30 +0530
Subject: samples/bpf: xdpsock: Add option to specify transmit fill pattern

The UDP payload fill pattern can be specified using '-P' or '--tx-pkt-pattern'
option. It is an unsigned 32 bit field and defaulted to 0x12345678.

The IP and UDP checksum is calculated by the code as per the content of
the packet before transmission.

Signed-off-by: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20191220085530.4980-7-jay.jayatheerthan@intel.com
---
 samples/bpf/xdpsock_user.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

(limited to 'samples')

diff --git a/samples/bpf/xdpsock_user.c b/samples/bpf/xdpsock_user.c
index 2297158a32bd..d74c4c83fc93 100644
--- a/samples/bpf/xdpsock_user.c
+++ b/samples/bpf/xdpsock_user.c
@@ -76,7 +76,7 @@ static bool benchmark_done;
 static u32 opt_batch_size = 64;
 static int opt_pkt_count;
 static u16 opt_pkt_size = MIN_PKT_SIZE;
-static u32 pkt_fill_pattern = 0x12345678;
+static u32 opt_pkt_fill_pattern = 0x12345678;
 static int opt_poll;
 static int opt_interval = 1;
 static u32 opt_xdp_bind_flags = XDP_USE_NEED_WAKEUP;
@@ -517,7 +517,7 @@ static void gen_eth_hdr_data(void)
 	udp_hdr->len = htons(UDP_PKT_SIZE);
 
 	/* UDP data */
-	memset32_htonl(pkt_data + PKT_HDR_SIZE, pkt_fill_pattern,
+	memset32_htonl(pkt_data + PKT_HDR_SIZE, opt_pkt_fill_pattern,
 		       UDP_PKT_DATA_SIZE);
 
 	/* UDP header checksum */
@@ -630,6 +630,7 @@ static struct option long_options[] = {
 	{"batch-size", required_argument, 0, 'b'},
 	{"tx-pkt-count", required_argument, 0, 'C'},
 	{"tx-pkt-size", required_argument, 0, 's'},
+	{"tx-pkt-pattern", required_argument, 0, 'P'},
 	{0, 0, 0, 0}
 };
 
@@ -663,10 +664,11 @@ static void usage(const char *prog)
 		"  -s, --tx-pkt-size=n	Transmit packet size.\n"
 		"			(Default: %d bytes)\n"
 		"			Min size: %d, Max size %d.\n"
+		"  -P, --tx-pkt-pattern=nPacket fill pattern. Default: 0x%x\n"
 		"\n";
 	fprintf(stderr, str, prog, XSK_UMEM__DEFAULT_FRAME_SIZE,
 		opt_batch_size, MIN_PKT_SIZE, MIN_PKT_SIZE,
-		XSK_UMEM__DEFAULT_FRAME_SIZE);
+		XSK_UMEM__DEFAULT_FRAME_SIZE, opt_pkt_fill_pattern);
 
 	exit(EXIT_FAILURE);
 }
@@ -678,7 +680,7 @@ static void parse_command_line(int argc, char **argv)
 	opterr = 0;
 
 	for (;;) {
-		c = getopt_long(argc, argv, "Frtli:q:pSNn:czf:muMd:b:C:s:",
+		c = getopt_long(argc, argv, "Frtli:q:pSNn:czf:muMd:b:C:s:P:",
 				long_options, &option_index);
 		if (c == -1)
 			break;
@@ -756,6 +758,9 @@ static void parse_command_line(int argc, char **argv)
 				usage(basename(argv[0]));
 			}
 			break;
+		case 'P':
+			opt_pkt_fill_pattern = strtol(optarg, NULL, 16);
+			break;
 		default:
 			usage(basename(argv[0]));
 		}
-- 
cgit v1.2.3


From f9e6bfdbaf0cf304d72c70a05d81acac01a04f48 Mon Sep 17 00:00:00 2001
From: Jesper Dangaard Brouer <brouer@redhat.com>
Date: Fri, 20 Dec 2019 17:19:36 +0100
Subject: samples/bpf: Xdp_redirect_cpu fix missing tracepoint attach

When sample xdp_redirect_cpu was converted to use libbpf, the
tracepoints used by this sample were not getting attached automatically
like with bpf_load.c. The BPF-maps was still getting loaded, thus
nobody notice that the tracepoints were not updating these maps.

This fix doesn't use the new skeleton code, as this bug was introduced
in v5.1 and stable might want to backport this. E.g. Red Hat QA uses
this sample as part of their testing.

Fixes: bbaf6029c49c ("samples/bpf: Convert XDP samples to libbpf usage")
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/157685877642.26195.2798780195186786841.stgit@firesoul
---
 samples/bpf/xdp_redirect_cpu_user.c | 59 ++++++++++++++++++++++++++++++++++---
 1 file changed, 55 insertions(+), 4 deletions(-)

(limited to 'samples')

diff --git a/samples/bpf/xdp_redirect_cpu_user.c b/samples/bpf/xdp_redirect_cpu_user.c
index 72af628529b5..79a2fb7d16cb 100644
--- a/samples/bpf/xdp_redirect_cpu_user.c
+++ b/samples/bpf/xdp_redirect_cpu_user.c
@@ -16,6 +16,10 @@ static const char *__doc__ =
 #include <getopt.h>
 #include <net/if.h>
 #include <time.h>
+#include <linux/limits.h>
+
+#define __must_check
+#include <linux/err.h>
 
 #include <arpa/inet.h>
 #include <linux/if_link.h>
@@ -46,6 +50,10 @@ static int cpus_count_map_fd;
 static int cpus_iterator_map_fd;
 static int exception_cnt_map_fd;
 
+#define NUM_TP 5
+struct bpf_link *tp_links[NUM_TP] = { 0 };
+static int tp_cnt = 0;
+
 /* Exit return codes */
 #define EXIT_OK		0
 #define EXIT_FAIL		1
@@ -88,6 +96,10 @@ static void int_exit(int sig)
 			printf("program on interface changed, not removing\n");
 		}
 	}
+	/* Detach tracepoints */
+	while (tp_cnt)
+		bpf_link__destroy(tp_links[--tp_cnt]);
+
 	exit(EXIT_OK);
 }
 
@@ -588,23 +600,61 @@ static void stats_poll(int interval, bool use_separators, char *prog_name,
 	free_stats_record(prev);
 }
 
+static struct bpf_link * attach_tp(struct bpf_object *obj,
+				   const char *tp_category,
+				   const char* tp_name)
+{
+	struct bpf_program *prog;
+	struct bpf_link *link;
+	char sec_name[PATH_MAX];
+	int len;
+
+	len = snprintf(sec_name, PATH_MAX, "tracepoint/%s/%s",
+		       tp_category, tp_name);
+	if (len < 0)
+		exit(EXIT_FAIL);
+
+	prog = bpf_object__find_program_by_title(obj, sec_name);
+	if (!prog) {
+		fprintf(stderr, "ERR: finding progsec: %s\n", sec_name);
+		exit(EXIT_FAIL_BPF);
+	}
+
+	link = bpf_program__attach_tracepoint(prog, tp_category, tp_name);
+	if (IS_ERR(link))
+		exit(EXIT_FAIL_BPF);
+
+	return link;
+}
+
+static void init_tracepoints(struct bpf_object *obj) {
+	tp_links[tp_cnt++] = attach_tp(obj, "xdp", "xdp_redirect_err");
+	tp_links[tp_cnt++] = attach_tp(obj, "xdp", "xdp_redirect_map_err");
+	tp_links[tp_cnt++] = attach_tp(obj, "xdp", "xdp_exception");
+	tp_links[tp_cnt++] = attach_tp(obj, "xdp", "xdp_cpumap_enqueue");
+	tp_links[tp_cnt++] = attach_tp(obj, "xdp", "xdp_cpumap_kthread");
+}
+
 static int init_map_fds(struct bpf_object *obj)
 {
-	cpu_map_fd = bpf_object__find_map_fd_by_name(obj, "cpu_map");
-	rx_cnt_map_fd = bpf_object__find_map_fd_by_name(obj, "rx_cnt");
+	/* Maps updated by tracepoints */
 	redirect_err_cnt_map_fd =
 		bpf_object__find_map_fd_by_name(obj, "redirect_err_cnt");
+	exception_cnt_map_fd =
+		bpf_object__find_map_fd_by_name(obj, "exception_cnt");
 	cpumap_enqueue_cnt_map_fd =
 		bpf_object__find_map_fd_by_name(obj, "cpumap_enqueue_cnt");
 	cpumap_kthread_cnt_map_fd =
 		bpf_object__find_map_fd_by_name(obj, "cpumap_kthread_cnt");
+
+	/* Maps used by XDP */
+	rx_cnt_map_fd = bpf_object__find_map_fd_by_name(obj, "rx_cnt");
+	cpu_map_fd = bpf_object__find_map_fd_by_name(obj, "cpu_map");
 	cpus_available_map_fd =
 		bpf_object__find_map_fd_by_name(obj, "cpus_available");
 	cpus_count_map_fd = bpf_object__find_map_fd_by_name(obj, "cpus_count");
 	cpus_iterator_map_fd =
 		bpf_object__find_map_fd_by_name(obj, "cpus_iterator");
-	exception_cnt_map_fd =
-		bpf_object__find_map_fd_by_name(obj, "exception_cnt");
 
 	if (cpu_map_fd < 0 || rx_cnt_map_fd < 0 ||
 	    redirect_err_cnt_map_fd < 0 || cpumap_enqueue_cnt_map_fd < 0 ||
@@ -662,6 +712,7 @@ int main(int argc, char **argv)
 			strerror(errno));
 		return EXIT_FAIL;
 	}
+	init_tracepoints(obj);
 	if (init_map_fds(obj) < 0) {
 		fprintf(stderr, "bpf_object__find_map_fd_by_name failed\n");
 		return EXIT_FAIL;
-- 
cgit v1.2.3