lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Tue, 28 May 2019 11:47:06 -0700
From:   Fred Klassen <fklassen@...neta.com>
To:     "David S. Miller" <davem@...emloft.net>,
        Alexey Kuznetsov <kuznet@....inr.ac.ru>,
        Hideaki YOSHIFUJI <yoshfuji@...ux-ipv6.org>,
        Shuah Khan <shuah@...nel.org>, netdev@...r.kernel.org,
        linux-kernel@...r.kernel.org, linux-kselftest@...r.kernel.org,
        Willem de Bruijn <willemb@...gle.com>
Cc:     Fred Klassen <fklassen@...neta.com>
Subject: [PATCH net-next v2 1/3] net/udpgso_bench_tx: options to exercise TX CMSG

This enhancement adds options that facilitate load testing with
additional TX CMSG options, and to optionally print results of
various send CMSG operations.

These options are especially useful in isolating situations
where error-queue messages are lost when combined with other
CMSG operations (e.g. SO_ZEROCOPY).

New options:

    -a - count all CMSG messages and match to sent messages
    -T - add TX CMSG that requests TX software timestamps
    -H - similar to -T except request TX hardware timestamps
    -P - call poll() before reading error queue
    -v - print detailed results

v2: Enhancements as per Willem de Bruijn <willemb@...gle.com>
    - Updated control and buffer parameters for recvmsg
    - poll() parameter cleanup
    - fail on bad audit results
    - remove TOS options
    - improved reporting

Signed-off-by: Fred Klassen <fklassen@...neta.com>
---
 tools/testing/selftests/net/udpgso_bench_tx.c | 324 ++++++++++++++++++++++++--
 1 file changed, 307 insertions(+), 17 deletions(-)

diff --git a/tools/testing/selftests/net/udpgso_bench_tx.c b/tools/testing/selftests/net/udpgso_bench_tx.c
index 4074538b5df5..8506c7c2c354 100644
--- a/tools/testing/selftests/net/udpgso_bench_tx.c
+++ b/tools/testing/selftests/net/udpgso_bench_tx.c
@@ -5,6 +5,8 @@
 #include <arpa/inet.h>
 #include <errno.h>
 #include <error.h>
+#include <linux/errqueue.h>
+#include <linux/net_tstamp.h>
 #include <netinet/if_ether.h>
 #include <netinet/in.h>
 #include <netinet/ip.h>
@@ -19,6 +21,7 @@
 #include <string.h>
 #include <sys/socket.h>
 #include <sys/time.h>
+#include <sys/poll.h>
 #include <sys/types.h>
 #include <unistd.h>
 
@@ -34,6 +37,10 @@
 #define SO_ZEROCOPY	60
 #endif
 
+#ifndef SO_EE_ORIGIN_ZEROCOPY
+#define SO_EE_ORIGIN_ZEROCOPY 5
+#endif
+
 #ifndef MSG_ZEROCOPY
 #define MSG_ZEROCOPY	0x4000000
 #endif
@@ -48,12 +55,25 @@ static uint16_t	cfg_mss;
 static int	cfg_payload_len	= (1472 * 42);
 static int	cfg_port	= 8000;
 static int	cfg_runtime_ms	= -1;
+static bool	cfg_poll;
 static bool	cfg_segment;
 static bool	cfg_sendmmsg;
 static bool	cfg_tcp;
+static uint32_t	cfg_tx_ts = SOF_TIMESTAMPING_TX_SOFTWARE;
+static bool	cfg_tx_tstamp;
+static bool	cfg_audit;
+static bool	cfg_verbose;
 static bool	cfg_zerocopy;
 static int	cfg_msg_nr;
 static uint16_t	cfg_gso_size;
+static unsigned long total_num_msgs;
+static unsigned long total_num_sends;
+static unsigned long stat_tx_ts;
+static unsigned long stat_tx_ts_errors;
+static unsigned long tstart;
+static unsigned long tend;
+static unsigned long stat_zcopies;
+static unsigned long stat_zcopy_errors;
 
 static socklen_t cfg_alen;
 static struct sockaddr_storage cfg_dst_addr;
@@ -110,23 +130,149 @@ static void setup_sockaddr(int domain, const char *str_addr, void *sockaddr)
 	}
 }
 
-static void flush_zerocopy(int fd)
+static void flush_cmsg(struct cmsghdr *cmsg)
 {
-	struct msghdr msg = {0};	/* flush */
+	switch (cmsg->cmsg_level) {
+	case SOL_SOCKET:
+		if (cmsg->cmsg_type == SO_TIMESTAMPING) {
+			int i;
+
+			i = (cfg_tx_ts == SOF_TIMESTAMPING_TX_HARDWARE) ? 2 : 0;
+			struct scm_timestamping *tss;
+
+			tss = (struct scm_timestamping *)CMSG_DATA(cmsg);
+			if (tss->ts[i].tv_sec == 0)
+				stat_tx_ts_errors++;
+		} else {
+			error(1, 0,
+			      "unknown SOL_SOCKET cmsg type=%u level=%u\n",
+			      cmsg->cmsg_type, cmsg->cmsg_level);
+		}
+		break;
+	case SOL_IP:
+	case SOL_IPV6:
+		switch (cmsg->cmsg_type) {
+		case IP_RECVERR:
+		case IPV6_RECVERR:
+		{
+			struct sock_extended_err *err;
+
+			err = (struct sock_extended_err *)CMSG_DATA(cmsg);
+			switch (err->ee_origin) {
+			case SO_EE_ORIGIN_TIMESTAMPING:
+				// Got a TX timestamp from error queue
+				stat_tx_ts++;
+				break;
+			case SO_EE_ORIGIN_ICMP:
+			case SO_EE_ORIGIN_ICMP6:
+				if (cfg_verbose)
+					fprintf(stderr,
+						"received ICMP error: type=%u, code=%u\n",
+						err->ee_type, err->ee_code);
+				break;
+			case SO_EE_ORIGIN_ZEROCOPY:
+			{
+				__u32 lo = err->ee_info;
+				__u32 hi = err->ee_data;
+
+				if (hi == lo - 1) {
+					// TX was aborted
+					stat_zcopy_errors++;
+					if (cfg_verbose)
+						fprintf(stderr,
+							"Zerocopy TX aborted: lo=%u hi=%u\n",
+							lo, hi);
+				} else if (hi == lo) {
+					// single ID acknowledged
+					stat_zcopies++;
+				} else {
+					// range of IDs acknowledged
+					stat_zcopies += hi - lo + 1;
+				}
+				break;
+			}
+			case SO_EE_ORIGIN_LOCAL:
+				if (cfg_verbose)
+					fprintf(stderr,
+						"received packet with local origin: %u\n",
+						err->ee_origin);
+				break;
+			default:
+				error(0, 1,
+				      "received packet with origin: %u",
+				      err->ee_origin);
+			}
+
+			break;
+		}
+		default:
+			error(0, 1, "unknown IP msg type=%u level=%u\n",
+			      cmsg->cmsg_type, cmsg->cmsg_level);
+			break;
+		}
+		break;
+	default:
+		error(0, 1, "unknown cmsg type=%u level=%u\n",
+		      cmsg->cmsg_type, cmsg->cmsg_level);
+	}
+}
+
+static void flush_errqueue_recv(int fd)
+{
+	char control[CMSG_SPACE(sizeof(struct scm_timestamping)) +
+		     CMSG_SPACE(sizeof(struct sock_extended_err)) +
+		     CMSG_SPACE(sizeof(struct sockaddr_in6))] = {0};
+	char buf[cfg_payload_len];
+	struct msghdr msg = {0};
+	struct cmsghdr *cmsg;
+	struct iovec entry;
 	int ret;
 
+	/* add a buffer to the message to avoid MSG_TRUNC msg_flag */
+	entry.iov_base = buf;
+	entry.iov_len = sizeof(buf);
+	msg.msg_iovlen = 1;
+	msg.msg_iov = &entry;
+
 	while (1) {
+		msg.msg_control = control;
+		msg.msg_controllen = sizeof(control);
 		ret = recvmsg(fd, &msg, MSG_ERRQUEUE);
 		if (ret == -1 && errno == EAGAIN)
 			break;
 		if (ret == -1)
 			error(1, errno, "errqueue");
-		if (msg.msg_flags != (MSG_ERRQUEUE | MSG_CTRUNC))
+		if (msg.msg_flags != MSG_ERRQUEUE)
 			error(1, 0, "errqueue: flags 0x%x\n", msg.msg_flags);
+		if (cfg_audit) {
+			for (cmsg = CMSG_FIRSTHDR(&msg);
+					cmsg;
+					cmsg = CMSG_NXTHDR(&msg, cmsg))
+				flush_cmsg(cmsg);
+		}
 		msg.msg_flags = 0;
 	}
 }
 
+static void flush_errqueue(int fd, const bool do_poll)
+{
+	if (do_poll) {
+		struct pollfd fds = {0};
+		int ret;
+
+		fds.fd = fd;
+		ret = poll(&fds, 1, 500);
+		if (ret == 0) {
+			if (cfg_verbose)
+				fprintf(stderr, "poll timeout\n");
+		} else if (ret < 0) {
+			error(1, errno, "poll");
+		}
+	}
+
+	flush_errqueue_recv(fd);
+}
+
 static int send_tcp(int fd, char *data)
 {
 	int ret, done = 0, count = 0;
@@ -168,16 +314,40 @@ static int send_udp(int fd, char *data)
 	return count;
 }
 
+static void send_ts_cmsg(struct cmsghdr *cm)
+{
+	uint32_t *valp;
+
+	cm->cmsg_level = SOL_SOCKET;
+	cm->cmsg_type = SO_TIMESTAMPING;
+	cm->cmsg_len = CMSG_LEN(sizeof(cfg_tx_ts));
+	valp = (void *)CMSG_DATA(cm);
+	*valp = cfg_tx_ts;
+}
+
 static int send_udp_sendmmsg(int fd, char *data)
 {
+	char control[CMSG_SPACE(sizeof(cfg_tx_ts))] = {0};
 	const int max_nr_msg = ETH_MAX_MTU / ETH_DATA_LEN;
 	struct mmsghdr mmsgs[max_nr_msg];
 	struct iovec iov[max_nr_msg];
 	unsigned int off = 0, left;
+	size_t msg_controllen = 0;
 	int i = 0, ret;
 
 	memset(mmsgs, 0, sizeof(mmsgs));
 
+	if (cfg_tx_tstamp) {
+		struct msghdr msg = {0};
+		struct cmsghdr *cmsg;
+
+		msg.msg_control = control;
+		msg.msg_controllen = sizeof(control);
+		cmsg = CMSG_FIRSTHDR(&msg);
+		send_ts_cmsg(cmsg);
+		msg_controllen += CMSG_SPACE(sizeof(cfg_tx_ts));
+	}
+
 	left = cfg_payload_len;
 	while (left) {
 		if (i == max_nr_msg)
@@ -188,6 +358,12 @@ static int send_udp_sendmmsg(int fd, char *data)
 
 		mmsgs[i].msg_hdr.msg_iov = iov + i;
 		mmsgs[i].msg_hdr.msg_iovlen = 1;
+		mmsgs[i].msg_hdr.msg_name = (void *)&cfg_dst_addr;
+		mmsgs[i].msg_hdr.msg_namelen = cfg_alen;
+		if (msg_controllen) {
+			mmsgs[i].msg_hdr.msg_control = control;
+			mmsgs[i].msg_hdr.msg_controllen = msg_controllen;
+		}
 
 		off += iov[i].iov_len;
 		left -= iov[i].iov_len;
@@ -214,9 +390,12 @@ static void send_udp_segment_cmsg(struct cmsghdr *cm)
 
 static int send_udp_segment(int fd, char *data)
 {
-	char control[CMSG_SPACE(sizeof(cfg_gso_size))] = {0};
+	char control[CMSG_SPACE(sizeof(cfg_gso_size)) +
+		     CMSG_SPACE(sizeof(cfg_tx_ts))] = {0};
 	struct msghdr msg = {0};
 	struct iovec iov = {0};
+	size_t msg_controllen;
+	struct cmsghdr *cmsg;
 	int ret;
 
 	iov.iov_base = data;
@@ -227,8 +406,16 @@ static int send_udp_segment(int fd, char *data)
 
 	msg.msg_control = control;
 	msg.msg_controllen = sizeof(control);
-	send_udp_segment_cmsg(CMSG_FIRSTHDR(&msg));
+	cmsg = CMSG_FIRSTHDR(&msg);
+	send_udp_segment_cmsg(cmsg);
+	msg_controllen = CMSG_SPACE(sizeof(cfg_mss));
+	if (cfg_tx_tstamp) {
+		cmsg = CMSG_NXTHDR(&msg, cmsg);
+		send_ts_cmsg(cmsg);
+		msg_controllen += CMSG_SPACE(sizeof(cfg_tx_ts));
+	}
 
+	msg.msg_controllen = msg_controllen;
 	msg.msg_name = (void *)&cfg_dst_addr;
 	msg.msg_namelen = cfg_alen;
 
@@ -236,15 +423,16 @@ static int send_udp_segment(int fd, char *data)
 	if (ret == -1)
 		error(1, errno, "sendmsg");
 	if (ret != iov.iov_len)
-		error(1, 0, "sendmsg: %u != %lu\n", ret, iov.iov_len);
+		error(1, 0, "sendmsg: %u != %lu", ret, iov.iov_len);
 
 	return 1;
 }
 
 static void usage(const char *filepath)
 {
-	error(1, 0, "Usage: %s [-46cmtuz] [-C cpu] [-D dst ip] [-l secs] [-m messagenr] [-p port] [-s sendsize] [-S gsosize]",
-		    filepath);
+	error(1, 0,
+	      "Usage: %s [-46acmHPtTuvz] [-C cpu] [-D dst ip] [-l secs] [-M messagenr] [-p port] [-s sendsize] [-S gsosize]",
+	      filepath);
 }
 
 static void parse_opts(int argc, char **argv)
@@ -252,7 +440,7 @@ static void parse_opts(int argc, char **argv)
 	int max_len, hdrlen;
 	int c;
 
-	while ((c = getopt(argc, argv, "46cC:D:l:mM:p:s:S:tuz")) != -1) {
+	while ((c = getopt(argc, argv, "46acC:D:Hl:mM:p:s:PS:tTuvz")) != -1) {
 		switch (c) {
 		case '4':
 			if (cfg_family != PF_UNSPEC)
@@ -266,6 +454,9 @@ static void parse_opts(int argc, char **argv)
 			cfg_family = PF_INET6;
 			cfg_alen = sizeof(struct sockaddr_in6);
 			break;
+		case 'a':
+			cfg_audit = true;
+			break;
 		case 'c':
 			cfg_cache_trash = true;
 			break;
@@ -287,6 +478,9 @@ static void parse_opts(int argc, char **argv)
 		case 'p':
 			cfg_port = strtoul(optarg, NULL, 0);
 			break;
+		case 'P':
+			cfg_poll = true;
+			break;
 		case 's':
 			cfg_payload_len = strtoul(optarg, NULL, 0);
 			break;
@@ -294,12 +488,22 @@ static void parse_opts(int argc, char **argv)
 			cfg_gso_size = strtoul(optarg, NULL, 0);
 			cfg_segment = true;
 			break;
+		case 'H':
+			cfg_tx_ts = SOF_TIMESTAMPING_TX_HARDWARE;
+			cfg_tx_tstamp = true;
+			break;
 		case 't':
 			cfg_tcp = true;
 			break;
+		case 'T':
+			cfg_tx_tstamp = true;
+			break;
 		case 'u':
 			cfg_connected = false;
 			break;
+		case 'v':
+			cfg_verbose = true;
+			break;
 		case 'z':
 			cfg_zerocopy = true;
 			break;
@@ -315,6 +519,8 @@ static void parse_opts(int argc, char **argv)
 		error(1, 0, "connectionless tcp makes no sense");
 	if (cfg_segment && cfg_sendmmsg)
 		error(1, 0, "cannot combine segment offload and sendmmsg");
+	if (cfg_tx_tstamp && !(cfg_segment || cfg_sendmmsg))
+		error(1, 0, "Options -T and -H require either -S or -m option");
 
 	if (cfg_family == PF_INET)
 		hdrlen = sizeof(struct iphdr) + sizeof(struct udphdr);
@@ -349,6 +555,79 @@ static void set_pmtu_discover(int fd, bool is_ipv4)
 		error(1, errno, "setsockopt path mtu");
 }
 
+static void set_tx_timestamping(int fd)
+{
+	int val = SOF_TIMESTAMPING_OPT_CMSG | SOF_TIMESTAMPING_OPT_ID;
+
+	if (cfg_tx_ts == SOF_TIMESTAMPING_TX_SOFTWARE)
+		val |= SOF_TIMESTAMPING_SOFTWARE;
+	else
+		val |= SOF_TIMESTAMPING_RAW_HARDWARE;
+
+	if (setsockopt(fd, SOL_SOCKET, SO_TIMESTAMPING, &val, sizeof(val)))
+		error(1, errno, "setsockopt tx timestamping");
+}
+
+static void print_audit_report(unsigned long num_msgs, unsigned long num_sends)
+{
+	unsigned long tdelta;
+
+	tdelta = tend - tstart;
+	if (!tdelta)
+		return;
+
+	fprintf(stderr, "Summary over %lu.%03lu seconds...\n",
+			tdelta / 1000, tdelta % 1000);
+	fprintf(stderr,
+		"sum %s tx: %6lu MB/s %10lu calls (%lu/s) %10lu msgs (%lu/s)\n",
+		cfg_tcp ? "tcp" : "udp",
+		((num_msgs * cfg_payload_len) >> 10) / tdelta,
+		num_sends, num_sends * 1000 / tdelta,
+		num_msgs, num_msgs * 1000 / tdelta);
+
+	if (cfg_tx_tstamp) {
+		if (stat_tx_ts_errors)
+			error(1, 0,
+			      "Expected clean TX Timestamps: %9lu msgs received %6lu errors",
+			      stat_tx_ts, stat_tx_ts_errors);
+		if (stat_tx_ts != num_sends)
+			error(1, 0,
+			      "Unexpected number of TX Timestamps: %9lu expected %9lu received",
+			      num_sends, stat_tx_ts);
+		fprintf(stderr,
+			"Tx Timestamps: %19lu received %17lu errors\n",
+			stat_tx_ts, stat_tx_ts_errors);
+	}
+
+	if (cfg_zerocopy) {
+		if (stat_zcopy_errors)
+			error(1, 0,
+			      "Expected clean Zerocopies: %9lu msgs received %6lu errors",
+			      stat_zcopies, stat_zcopy_errors);
+		if (stat_zcopies != num_sends)
+			error(1, 0,
+			      "Unexpected number of Zerocopy completions: %9lu expected %9lu received",
+			      num_sends, stat_zcopies);
+		fprintf(stderr,
+			"Zerocopy acks: %19lu received %17lu errors\n",
+			stat_zcopies, stat_zcopy_errors);
+	}
+}
+
+static void print_report(unsigned long num_msgs, unsigned long num_sends)
+{
+	fprintf(stderr,
+		"%s tx: %6lu MB/s %8lu calls/s %6lu msg/s\n",
+		cfg_tcp ? "tcp" : "udp",
+		(num_msgs * cfg_payload_len) >> 20,
+		num_sends, num_msgs);
+
+	if (cfg_audit) {
+		total_num_msgs += num_msgs;
+		total_num_sends += num_sends;
+	}
+}
+
 int main(int argc, char **argv)
 {
 	unsigned long num_msgs, num_sends;
@@ -384,8 +663,13 @@ int main(int argc, char **argv)
 	if (cfg_segment)
 		set_pmtu_discover(fd, cfg_family == PF_INET);
 
+	if (cfg_tx_tstamp)
+		set_tx_timestamping(fd);
+
 	num_msgs = num_sends = 0;
 	tnow = gettimeofday_ms();
+	tstart = tnow;
+	tend = tnow;
 	tstop = tnow + cfg_runtime_ms;
 	treport = tnow + 1000;
 
@@ -400,19 +684,15 @@ int main(int argc, char **argv)
 		else
 			num_sends += send_udp(fd, buf[i]);
 		num_msgs++;
-		if (cfg_zerocopy && ((num_msgs & 0xF) == 0))
-			flush_zerocopy(fd);
+		if ((cfg_zerocopy && ((num_msgs & 0xF) == 0)) || cfg_tx_tstamp)
+			flush_errqueue(fd, cfg_poll);
 
 		if (cfg_msg_nr && num_msgs >= cfg_msg_nr)
 			break;
 
 		tnow = gettimeofday_ms();
-		if (tnow > treport) {
-			fprintf(stderr,
-				"%s tx: %6lu MB/s %8lu calls/s %6lu msg/s\n",
-				cfg_tcp ? "tcp" : "udp",
-				(num_msgs * cfg_payload_len) >> 20,
-				num_sends, num_msgs);
+		if (tnow >= treport) {
+			print_report(num_msgs, num_sends);
 			num_msgs = num_sends = 0;
 			treport = tnow + 1000;
 		}
@@ -423,8 +703,18 @@ int main(int argc, char **argv)
 
 	} while (!interrupted && (cfg_runtime_ms == -1 || tnow < tstop));
 
+	if (cfg_zerocopy || cfg_tx_tstamp)
+		flush_errqueue(fd, true);
+
 	if (close(fd))
 		error(1, errno, "close");
 
+	if (cfg_audit) {
+		tend = tnow;
+		total_num_msgs += num_msgs;
+		total_num_sends += num_sends;
+		print_audit_report(total_num_msgs, total_num_sends);
+	}
+
 	return 0;
 }
-- 
2.11.0

Powered by blists - more mailing lists