Merge branch 'net-timestamp'

Willem de Bruijn says:

====================
net-timestamp: blinding

Changes
  (v2 -> v3)
  - rebase only: v2 did not make it to patchwork / netdev
  (v1 -> v2)
  - fix capability check in patch 2
      this could be moved into net/core/sock.c as sk_capable_nouser()
  (rfc -> v1)
  - dropped patch 4: timestamp batching
      due to complexity, as discussed
  - dropped patch 5: default mode
      because it does not really cover all use cases, as discussed
  - added documentation
  - minor fix, see patch 2

Two issues were raised during recent timestamping discussions:
1. looping full packets on the error queue exposes packet headers
2. TCP timestamping with retransmissions generates many timestamps

This RFC patchset is an attempt at addressing both without breaking
legacy behavior.

Patch 1 reintroduces the "no payload" timestamp option, which loops
timestamps onto an empty skb. This reduces the pressure on SO_RCVBUF
from looping many timestamps. It does not reduce the number of recv()
calls needed to process them. The timestamp cookie mechanism developed
in http://patchwork.ozlabs.org/patch/427213/ did, but this is
considerably simpler.

Patch 2 then gives administrators the power to block all timestamp
requests that contain data by unprivileged users. I proposed this
earlier as a backward compatible workaround in the discussion of

  net-timestamp: pull headers for SOCK_STREAM
  http://patchwork.ozlabs.org/patch/414810/

Patch 3 only updates the txtimestamp example to test this option.
Verified that with option '-n', length is zero in all cases and
option '-I' (PKTINFO) stops working.
====================

Acked-by: Richard Cochran <[email protected]>
Signed-off-by: David S. Miller <[email protected]>

Author: davem330

Documentation/networking/timestamping.txt

@@ -162,6 +162,27 @@ SOF_TIMESTAMPING_OPT_CMSG:
   option IP_PKTINFO simultaneously.
 
 
+SOF_TIMESTAMPING_OPT_TSONLY:
+
+  Applies to transmit timestamps only. Makes the kernel return the
+  timestamp as a cmsg alongside an empty packet, as opposed to
+  alongside the original packet. This reduces the amount of memory
+  charged to the socket's receive budget (SO_RCVBUF) and delivers
+  the timestamp even if sysctl net.core.tstamp_allow_data is 0.
+  This option disables SOF_TIMESTAMPING_OPT_CMSG.
+
+
+New applications are encouraged to pass SOF_TIMESTAMPING_OPT_ID to
+disambiguate timestamps and SOF_TIMESTAMPING_OPT_TSONLY to operate
+regardless of the setting of sysctl net.core.tstamp_allow_data.
+
+An exception is when a process needs additional cmsg data, for
+instance SOL_IP/IP_PKTINFO to detect the egress network interface.
+Then pass option SOF_TIMESTAMPING_OPT_CMSG. This option depends on
+having access to the contents of the original packet, so cannot be
+combined with SOF_TIMESTAMPING_OPT_TSONLY.
+
+
 1.4 Bytestream Timestamps
 
 The SO_TIMESTAMPING interface supports timestamping of bytes in a

Documentation/networking/timestamping/txtimestamp.c

@@ -70,6 +70,7 @@ static int do_ipv6 = 1;
 static int cfg_payload_len = 10;
 static bool cfg_show_payload;
 static bool cfg_do_pktinfo;
+static bool cfg_loop_nodata;
 static uint16_t dest_port = 9000;
 
 static struct sockaddr_in daddr;
@@ -141,6 +142,9 @@ static void print_payload(char *data, int len)
 {
 	int i;
 
+	if (!len)
+		return;
+
 	if (len > 70)
 		len = 70;
 
@@ -177,6 +181,7 @@ static void __recv_errmsg_cmsg(struct msghdr *msg, int payload_len)
 	struct sock_extended_err *serr = NULL;
 	struct scm_timestamping *tss = NULL;
 	struct cmsghdr *cm;
+	int batch = 0;
 
 	for (cm = CMSG_FIRSTHDR(msg);
 	     cm && cm->cmsg_len;
@@ -209,10 +214,18 @@ static void __recv_errmsg_cmsg(struct msghdr *msg, int payload_len)
 		} else
 			fprintf(stderr, "unknown cmsg %d,%d\n",
 					cm->cmsg_level, cm->cmsg_type);
+
+		if (serr && tss) {
+			print_timestamp(tss, serr->ee_info, serr->ee_data,
+					payload_len);
+			serr = NULL;
+			tss = NULL;
+			batch++;
+		}
 	}
 
-	if (serr && tss)
-		print_timestamp(tss, serr->ee_info, serr->ee_data, payload_len);
+	if (batch > 1)
+		fprintf(stderr, "batched %d timestamps\n", batch);
 }
 
 static int recv_errmsg(int fd)
@@ -244,7 +257,7 @@ static int recv_errmsg(int fd)
 	if (ret == -1 && errno != EAGAIN)
 		error(1, errno, "recvmsg");
 
-	if (ret > 0) {
+	if (ret >= 0) {
 		__recv_errmsg_cmsg(&msg, ret);
 		if (cfg_show_payload)
 			print_payload(data, cfg_payload_len);
@@ -309,6 +322,9 @@ static void do_test(int family, unsigned int opt)
 	opt |= SOF_TIMESTAMPING_SOFTWARE |
 	       SOF_TIMESTAMPING_OPT_CMSG |
 	       SOF_TIMESTAMPING_OPT_ID;
+	if (cfg_loop_nodata)
+		opt |= SOF_TIMESTAMPING_OPT_TSONLY;
+
 	if (setsockopt(fd, SOL_SOCKET, SO_TIMESTAMPING,
 		       (char *) &opt, sizeof(opt)))
 		error(1, 0, "setsockopt timestamping");
@@ -378,6 +394,7 @@ static void __attribute__((noreturn)) usage(const char *filepath)
 			"  -h:   show this message\n"
 			"  -I:   request PKTINFO\n"
 			"  -l N: send N bytes at a time\n"
+			"  -n:   set no-payload option\n"
 			"  -r:   use raw\n"
 			"  -R:   use raw (IP_HDRINCL)\n"
 			"  -p N: connect to port N\n"
@@ -392,7 +409,7 @@ static void parse_opt(int argc, char **argv)
 	int proto_count = 0;
 	char c;
 
-	while ((c = getopt(argc, argv, "46hIl:p:rRux")) != -1) {
+	while ((c = getopt(argc, argv, "46hIl:np:rRux")) != -1) {
 		switch (c) {
 		case '4':
 			do_ipv6 = 0;
@@ -403,6 +420,9 @@ static void parse_opt(int argc, char **argv)
 		case 'I':
 			cfg_do_pktinfo = true;
 			break;
+		case 'n':
+			cfg_loop_nodata = true;
+			break;
 		case 'r':
 			proto_count++;
 			cfg_proto = SOCK_RAW;

Documentation/sysctl/net.txt

@@ -97,6 +97,14 @@ rmem_max
 
 The maximum receive socket buffer size in bytes.
 
+tstamp_allow_data
+-----------------
+Allow processes to receive tx timestamps looped together with the original
+packet contents. If disabled, transmit timestamp requests from unprivileged
+processes are dropped unless socket option SOF_TIMESTAMPING_OPT_TSONLY is set.
+Default: 1 (on)
+
+
 wmem_default
 ------------
 

include/net/sock.h

@@ -2239,6 +2239,7 @@ bool sk_net_capable(const struct sock *sk, int cap);
 extern __u32 sysctl_wmem_max;
 extern __u32 sysctl_rmem_max;
 
+extern int sysctl_tstamp_allow_data;
 extern int sysctl_optmem_max;
 
 extern __u32 sysctl_wmem_default;

include/uapi/linux/net_tstamp.h

@@ -24,8 +24,9 @@ enum {
 	SOF_TIMESTAMPING_TX_SCHED = (1<<8),
 	SOF_TIMESTAMPING_TX_ACK = (1<<9),
 	SOF_TIMESTAMPING_OPT_CMSG = (1<<10),
+	SOF_TIMESTAMPING_OPT_TSONLY = (1<<11),
 
-	SOF_TIMESTAMPING_LAST = SOF_TIMESTAMPING_OPT_CMSG,
+	SOF_TIMESTAMPING_LAST = SOF_TIMESTAMPING_OPT_TSONLY,
 	SOF_TIMESTAMPING_MASK = (SOF_TIMESTAMPING_LAST - 1) |
 				 SOF_TIMESTAMPING_LAST
 };

net/core/skbuff.c

@@ -74,6 +74,8 @@
 #include <asm/uaccess.h>
 #include <trace/events/skb.h>
 #include <linux/highmem.h>
+#include <linux/capability.h>
+#include <linux/user_namespace.h>
 
 struct kmem_cache *skbuff_head_cache __read_mostly;
 static struct kmem_cache *skbuff_fclone_cache __read_mostly;
@@ -3690,11 +3692,28 @@ static void __skb_complete_tx_timestamp(struct sk_buff *skb,
 		kfree_skb(skb);
 }
 
+static bool skb_may_tx_timestamp(struct sock *sk, bool tsonly)
+{
+	bool ret;
+
+	if (likely(sysctl_tstamp_allow_data || tsonly))
+		return true;
+
+	read_lock_bh(&sk->sk_callback_lock);
+	ret = sk->sk_socket && sk->sk_socket->file &&
+	      file_ns_capable(sk->sk_socket->file, &init_user_ns, CAP_NET_RAW);
+	read_unlock_bh(&sk->sk_callback_lock);
+	return ret;
+}
+
 void skb_complete_tx_timestamp(struct sk_buff *skb,
 			       struct skb_shared_hwtstamps *hwtstamps)
 {
 	struct sock *sk = skb->sk;
 
+	if (!skb_may_tx_timestamp(sk, false))
+		return;
+
 	/* take a reference to prevent skb_orphan() from freeing the socket */
 	sock_hold(sk);
 
@@ -3710,19 +3729,28 @@ void __skb_tstamp_tx(struct sk_buff *orig_skb,
 		     struct sock *sk, int tstype)
 {
 	struct sk_buff *skb;
+	bool tsonly = sk->sk_tsflags & SOF_TIMESTAMPING_OPT_TSONLY;
 
-	if (!sk)
+	if (!sk || !skb_may_tx_timestamp(sk, tsonly))
 		return;
 
-	if (hwtstamps)
-		*skb_hwtstamps(orig_skb) = *hwtstamps;
+	if (tsonly)
+		skb = alloc_skb(0, GFP_ATOMIC);
 	else
-		orig_skb->tstamp = ktime_get_real();
-
-	skb = skb_clone(orig_skb, GFP_ATOMIC);
+		skb = skb_clone(orig_skb, GFP_ATOMIC);
 	if (!skb)
 		return;
 
+	if (tsonly) {
+		skb_shinfo(skb)->tx_flags = skb_shinfo(orig_skb)->tx_flags;
+		skb_shinfo(skb)->tskey = skb_shinfo(orig_skb)->tskey;
+	}
+
+	if (hwtstamps)
+		*skb_hwtstamps(skb) = *hwtstamps;
+	else
+		skb->tstamp = ktime_get_real();
+
 	__skb_complete_tx_timestamp(skb, sk, tstype);
 }
 EXPORT_SYMBOL_GPL(__skb_tstamp_tx);

net/core/sock.c

@@ -325,6 +325,8 @@ __u32 sysctl_rmem_default __read_mostly = SK_RMEM_MAX;
 int sysctl_optmem_max __read_mostly = sizeof(unsigned long)*(2*UIO_MAXIOV+512);
 EXPORT_SYMBOL(sysctl_optmem_max);
 
+int sysctl_tstamp_allow_data __read_mostly = 1;
+
 struct static_key memalloc_socks = STATIC_KEY_INIT_FALSE;
 EXPORT_SYMBOL_GPL(memalloc_socks);
 
@@ -840,6 +842,7 @@ int sock_setsockopt(struct socket *sock, int level, int optname,
 			ret = -EINVAL;
 			break;
 		}
+
 		if (val & SOF_TIMESTAMPING_OPT_ID &&
 		    !(sk->sk_tsflags & SOF_TIMESTAMPING_OPT_ID)) {
 			if (sk->sk_protocol == IPPROTO_TCP) {

net/core/sysctl_net_core.c

@@ -321,6 +321,15 @@ static struct ctl_table net_core_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
+	{
+		.procname	= "tstamp_allow_data",
+		.data		= &sysctl_tstamp_allow_data,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &zero,
+		.extra2		= &one
+	},
 #ifdef CONFIG_RPS
 	{
 		.procname	= "rps_sock_flow_entries",

net/ipv4/ip_sockglue.c

@@ -483,7 +483,7 @@ int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len)
 
 	serr = SKB_EXT_ERR(skb);
 
-	if (sin) {
+	if (sin && skb->len) {
 		sin->sin_family = AF_INET;
 		sin->sin_addr.s_addr = *(__be32 *)(skb_network_header(skb) +
 						   serr->addr_offset);
@@ -496,8 +496,9 @@ int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len)
 	sin = &errhdr.offender;
 	memset(sin, 0, sizeof(*sin));
 
-	if (serr->ee.ee_origin == SO_EE_ORIGIN_ICMP ||
-	    ipv4_pktinfo_prepare_errqueue(sk, skb, serr->ee.ee_origin)) {
+	if (skb->len &&
+	    (serr->ee.ee_origin == SO_EE_ORIGIN_ICMP ||
+	     ipv4_pktinfo_prepare_errqueue(sk, skb, serr->ee.ee_origin))) {
 		sin->sin_family = AF_INET;
 		sin->sin_addr.s_addr = ip_hdr(skb)->saddr;
 		if (inet_sk(sk)->cmsg_flags)

net/ipv6/datagram.c

@@ -369,7 +369,7 @@ int ipv6_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len)
 
 	serr = SKB_EXT_ERR(skb);
 
-	if (sin) {
+	if (sin && skb->len) {
 		const unsigned char *nh = skb_network_header(skb);
 		sin->sin6_family = AF_INET6;
 		sin->sin6_flowinfo = 0;
@@ -394,8 +394,7 @@ int ipv6_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len)
 	memcpy(&errhdr.ee, &serr->ee, sizeof(struct sock_extended_err));
 	sin = &errhdr.offender;
 	memset(sin, 0, sizeof(*sin));
-
-	if (serr->ee.ee_origin != SO_EE_ORIGIN_LOCAL) {
+	if (serr->ee.ee_origin != SO_EE_ORIGIN_LOCAL && skb->len) {
 		sin->sin6_family = AF_INET6;
 		if (np->rxopt.all) {
 			if (serr->ee.ee_origin != SO_EE_ORIGIN_ICMP &&