net: Introduce recvmmsg socket syscall

RFC Superseded Arnaldo Carvalho de Melo 4 participants 10 messages Jun 11, 2009 – Jun 16, 2009

Subsystems (5)

Networking Network drivers x86 ABI/API +1 more subsystems

lore

Share & tools

lore

Arnaldo Carvalho de Melo

author

2009-06-11 03:42 UTC #0 lore raw

RFC v2 1/1 net: Introduce recvmmsg socket syscall

Meaning receive multiple messages, reducing the number of syscalls and net stack entry/exit operations.

Next patches will introduce mechanisms where protocols that want to optimize this operation will provide an unlocked_recvmsg operation.

This takes into account comments made by:

. Paul Moore: sock_recvmsg is called only for the first datagram,
sock_recvmsg_nosec is used for the rest.

. Caitlin Bestler: recvmmsg now has a struct timespec timeout, that
works in the same fashion as the ppoll one.

If the underlying protocol returns a datagram with MSG_OOB set, this will make recvmmsg return right away with as many datagrams (+ the OOB one) it has received so far.

. Rémi Denis-Courmont & Steven Whitehouse: If we receive N < vlen
datagrams and then recvmsg returns an error, recvmmsg will return the successfully received datagrams, store the error and return it in the next call.

I'll defer work on the subsequent optimization (sk_prot->unlocked_recvmmsg) till we get the syscall API sorted out.

One thing to think about is if programs that provide a timeout for the recvmmsg operation that is smaller than the one set (or the default) for SO_RCVTIMEO should get their neck broken by inproper rope usage.

I thought about checking that and doing the equivalent to sock_set_timeout(&sk->sk_rcvtimeo, recvmmsg_timeout), but felt like recvmmsg was getting too many lines of code, opinions?

Also details such as adding the syscall to all the other archs syscall tables and providing a socketcall interface for those that want it will be addressed before final submission.

Attached also goes the updated recvmmsg canonical usage tool.

Thanks for all the comments so far and keep them coming! :-)

- Arnaldo

arch/x86/ia32/ia32entry.S +1 -0

index e590261..2a188e5 100644
@@ -832,4 +832,5 @@ ia32_sys_call_table:
 	.quad compat_sys_pwritev
 	.quad compat_sys_rt_tgsigqueueinfo	/* 335 */
 	.quad sys_perf_counter_open
+	.quad compat_sys_recvmmsg
 ia32_syscall_end:

arch/x86/include/asm/unistd_32.h +1 -0

index 732a307..3e72cae 100644
@@ -342,6 +342,7 @@
 #define __NR_pwritev		334
 #define __NR_rt_tgsigqueueinfo	335
 #define __NR_perf_counter_open	336
+#define __NR_recvmmsg		337
 
 #ifdef __KERNEL__

arch/x86/include/asm/unistd_64.h +2 -0

index 900e161..713a32a 100644
@@ -661,6 +661,8 @@ __SYSCALL(__NR_pwritev, sys_pwritev)
 __SYSCALL(__NR_rt_tgsigqueueinfo, sys_rt_tgsigqueueinfo)
 #define __NR_perf_counter_open			298
 __SYSCALL(__NR_perf_counter_open, sys_perf_counter_open)
+#define __NR_recvmmsg				299
+__SYSCALL(__NR_recvmmsg, sys_recvmmsg)
 
 #ifndef __NO_STUBS
 #define __ARCH_WANT_OLD_READDIR

arch/x86/kernel/syscall_table_32.S +1 -0

index d51321d..4881b14 100644
@@ -336,3 +336,4 @@ ENTRY(sys_call_table)
 	.long sys_pwritev
 	.long sys_rt_tgsigqueueinfo	/* 335 */
 	.long sys_perf_counter_open
+	.long sys_recvmmsg

include/linux/socket.h +6 -0

index 421afb4..5aaa78a 100644
@@ -65,6 +65,12 @@ struct msghdr {
 	unsigned	msg_flags;
 };
 
+/* For recvmmsg/sendmmsg */
+struct mmsghdr {
+	struct msghdr   msg_hdr;
+	unsigned        msg_len;
+};
+
 /*
  *	POSIX 1003.1g - ancillary data object information
  *	Ancillary data consits of a sequence of pairs of

include/linux/syscalls.h +4 -0

index c6c84ad..afefa61 100644
@@ -25,6 +25,7 @@ struct linux_dirent64;
 struct list_head;
 struct msgbuf;
 struct msghdr;
+struct mmsghdr;
 struct msqid_ds;
 struct new_utsname;
 struct nfsctl_arg;

@@ -556,6 +557,9 @@ asmlinkage long sys_recv(int, void __user *, size_t, unsigned);
 asmlinkage long sys_recvfrom(int, void __user *, size_t, unsigned,
 				struct sockaddr __user *, int __user *);
 asmlinkage long sys_recvmsg(int fd, struct msghdr __user *msg, unsigned flags);
+asmlinkage long sys_recvmmsg(int fd, struct mmsghdr __user *msg,
+			     unsigned int vlen, unsigned flags,
+			     struct timespec __user *timeout);
 asmlinkage long sys_socket(int, int, int);
 asmlinkage long sys_socketpair(int, int, int, int __user *);
 asmlinkage long sys_socketcall(int call, unsigned long __user *args);

include/net/compat.h +8 -0

index 5bbf8bf..bcde814 100644
@@ -18,6 +18,11 @@ struct compat_msghdr {
 	compat_uint_t	msg_flags;
 };
 
+struct compat_mmsghdr {
+	struct compat_msghdr msg_hdr;
+	compat_uint_t        msg_len;
+};
+
 struct compat_cmsghdr {
 	compat_size_t	cmsg_len;
 	compat_int_t	cmsg_level;

@@ -35,6 +40,9 @@ extern int get_compat_msghdr(struct msghdr *, struct compat_msghdr __user *);
 extern int verify_compat_iovec(struct msghdr *, struct iovec *, struct sockaddr *, int);
 extern asmlinkage long compat_sys_sendmsg(int,struct compat_msghdr __user *,unsigned);
 extern asmlinkage long compat_sys_recvmsg(int,struct compat_msghdr __user *,unsigned);
+extern asmlinkage long compat_sys_recvmmsg(int, struct compat_mmsghdr __user *,
+					   unsigned, unsigned,
+					   struct compat_timespec __user *);
 extern asmlinkage long compat_sys_getsockopt(int, int, int, char __user *, int __user *);
 extern int put_cmsg_compat(struct msghdr*, int, int, int, void *);

kernel/sys_ni.c +2 -0

index 68320f6..f581fb0 100644
@@ -48,7 +48,9 @@ cond_syscall(sys_shutdown);
 cond_syscall(sys_sendmsg);
 cond_syscall(compat_sys_sendmsg);
 cond_syscall(sys_recvmsg);
+cond_syscall(sys_recvmmsg);
 cond_syscall(compat_sys_recvmsg);
+cond_syscall(compat_sys_recvmmsg);
 cond_syscall(sys_socketcall);
 cond_syscall(sys_futex);
 cond_syscall(compat_sys_futex);

net/compat.c +9 -0

index 8d73905..4fab14c 100644
@@ -743,6 +743,15 @@ asmlinkage long compat_sys_recvmsg(int fd, struct compat_msghdr __user *msg, uns
 	return sys_recvmsg(fd, (struct msghdr __user *)msg, flags | MSG_CMSG_COMPAT);
 }
 
+asmlinkage long compat_sys_recvmmsg(int fd, struct compat_mmsghdr __user *mmsg,
+				    unsigned vlen, unsigned int flags,
+				    struct compat_timespec __user *timeout)
+{
+	return sys_recvmmsg(fd, (struct mmsghdr __user *)mmsg, vlen,
+			    flags | MSG_CMSG_COMPAT,
+			    (struct timespec __user *)timeout);
+}
+
 asmlinkage long compat_sys_socketcall(int call, u32 __user *args)
 {
 	int ret;

net/socket.c +172 ~~-31~~

285 lines not on this page · raw message

#include <stdlib.h>
#include <syscall.h>
#include <stdio.h>
#include <sys/socket.h>
#include <unistd.h>
#include <netinet/in.h>
#include <netinet/tcp.h>
#include <arpa/inet.h>
#include <netdb.h>
#include <poll.h>
#include <string.h>

struct mmsghdr {
	struct msghdr	msg_hdr;
	unsigned	msg_len;
};

#if defined(__x86_64__) || defined(__i386__) #include "linux-2.6-tip/arch/x86/include/asm/unistd.h" #endif

#ifndef NSEC_PER_MSEC
#define NSEC_PER_MSEC 1000000UL
#endif

static inline int recvmmsg(int fd, struct mmsghdr *mmsg,
unsigned vlen, unsigned flags, struct timespec *timeout)

{
	return syscall(__NR_recvmmsg, fd, mmsg, vlen, flags, timeout);
}

static void print_stats_peer(struct mmsghdr *datagram, int count, int bytes)

{
	char peer[1024];
	int err = getnameinfo(datagram->msg_hdr.msg_name,
			      datagram->msg_hdr.msg_namelen,
			      peer, sizeof(peer), NULL, 0, 0);
	if (err != 0) {
		fprintf(stderr, "error using getnameinfo: %s\n",
			gai_strerror(err));
			return;
		}
	printf("    %d bytes received from %s in %d datagrams\n",
	       bytes, peer, count);
}

int main(int argc, char *argv[])

{
	struct addrinfo *host;
	struct addrinfo hints = {
		.ai_family   = AF_INET,
		.ai_socktype = SOCK_DGRAM,
		.ai_protocol = IPPROTO_UDP,
		.ai_flags    = AI_PASSIVE,
	};
	const char *port = "5001";
	int batch_size = 8;
	long timeout = 10 * NSEC_PER_MSEC;
	int err, fd;
	int i;

	if (argc > 1)
		port = argv[1];

	if (argc > 2)
		batch_size = atoi(argv[2]);

	if (argc > 3)
		timeout = atol(argv[3]) * NSEC_PER_MSEC;

	char buf[batch_size][256];
	struct iovec iovec[batch_size][1];
	struct sockaddr addr[batch_size];
	struct mmsghdr datagrams[batch_size];

	printf("usage: recvmmsg <port(def 5001)> <batch_size(def 8)> "
	       "<timeout(def 10ms>\n\nWaiting for datagrams...\n");

	err = getaddrinfo(NULL, port, &hints, &host);
	if (err != 0) {
		fprintf(stderr, "error using getaddrinfo: %s\n",
			gai_strerror(err));
		goto out;
	}
	
	fd = socket(host->ai_family, host->ai_socktype, host->ai_protocol);
	if (fd < 0) {
		perror("socket: ");
		goto out_freeaddrinfo;
	}

	if (bind(fd, host->ai_addr, host->ai_addrlen) < 0) {
		perror("bind: ");
		goto out_close_server;
	}

	for (i = 0; i < batch_size; ++i) {
		iovec[i][0].iov_base = buf[i];
		iovec[i][0].iov_len  = sizeof(buf[i]);
		datagrams[i].msg_hdr.msg_iov	 = iovec[i];
		datagrams[i].msg_hdr.msg_iovlen	 = 1;
		datagrams[i].msg_hdr.msg_name	 = &addr[i];
		datagrams[i].msg_hdr.msg_namelen = sizeof(addr[i]);
	}

	struct pollfd pfds[1] = {
		[0] = {
			.fd = fd,
			.events = POLLIN,
		},
	};

	while (1) {
		struct timespec timeout = { .tv_nsec = 10 * NSEC_PER_MSEC, };

		if (poll(pfds, 1, -1) < 0) {
			perror("poll: ");
			return EXIT_FAILURE;
		}

		int nr_datagrams = recvmmsg(fd, datagrams, batch_size,
					    0, &timeout);

		if (nr_datagrams == 0) {
			perror("recvmmsg: ");
			return EXIT_FAILURE;
		}

		printf("nr_datagrams received: %d, remaining: %luns\n",
		       nr_datagrams, timeout.tv_nsec);
		int peer_count = 1;
		int peer_bytes = datagrams[0].msg_len;
		for (i = 1; i < nr_datagrams; ++i) {
			if (memcmp(datagrams[i - 1].msg_hdr.msg_name,
				   datagrams[i].msg_hdr.msg_name,
				   datagrams[i].msg_hdr.msg_namelen) == 0) {
				++peer_count;
				peer_bytes += datagrams[i].msg_len;
				continue;
			}
			
			print_stats_peer(&datagrams[i - 1],
					 peer_count, peer_bytes);
			peer_bytes = datagrams[i].msg_len;
			peer_count = 1;
		}

		print_stats_peer(&datagrams[nr_datagrams - 1],
				 peer_count, peer_bytes);
	}
out_close_server:
	close(fd);
out_freeaddrinfo:
	freeaddrinfo(host);
out:
	return err;
}

David Miller

maintainer

2009-06-11 09:40 UTC #1 lore raw

From: Arnaldo Carvalho de Melo <[email protected]>
Date: Thu, 11 Jun 2009 00:40:22 -0300

net/socket.c SYSCALL_DEFINE3(recvmsg, int, fd, struct msghdr __user *, msg, · line 2118

Quoted 2 lines

net/socket.c +2 -0

+	struct compat_timespec *timeout_compat =
+		(struct compat_timespec *)timeout;

In order for this to build on non-compat platforms, or when CONFIG_COMPAT is disabled, you'll need to avoid trying to reference compat_timespec.

We get away with compat_msghdr because we define it as a NOP in net/compat.h when necessary.

Doing the same, and cleanly, with this non-networking compat knob is unlikely to be straightforward.

Arnaldo Carvalho de Melo

author

2009-06-11 15:35 UTC #2 lore raw

Em Thu, Jun 11, 2009 at 02:41:00AM -0700, David Miller escreveu:

Quoted 10 lines

From: Arnaldo Carvalho de Melo <[email protected]>
Date: Thu, 11 Jun 2009 00:40:22 -0300
net/socket.c SYSCALL_DEFINE3(recvmsg, int, fd, struct msghdr __user *, msg, · line 2118
Quoted 2 lines
net/socket.c +2 -0
+	struct compat_timespec *timeout_compat =
+		(struct compat_timespec *)timeout;
In order for this to build on non-compat platforms, or when CONFIG_COMPAT is disabled, you'll need to avoid trying to reference compat_timespec.

Quoted 5 lines

We get away with compat_msghdr because we define it as a NOP in net/compat.h when necessary.
Doing the same, and cleanly, with this non-networking compat knob is unlikely to be straightforward.

Done, not like msghdr, but should work, testing now.

Thanks for quickly pointing out this brainfart!

- Arnaldo

David Miller

maintainer

2009-06-12 00:00 UTC #3 lore raw

From: Arnaldo Carvalho de Melo <[email protected]>
Date: Thu, 11 Jun 2009 12:34:46 -0300

Quoted 1 line

Thanks for quickly pointing out this brainfart!

Couldn't everyone else smell it too?

Maybe I'm just super-sensitive.

Arnaldo Carvalho de Melo

author

2009-06-11 22:58 UTC #4 lore raw

RFC v3 1/1 net: Introduce recvmmsg socket syscall

In reply to David Miller

Em Thu, Jun 11, 2009 at 02:41:00AM -0700, David Miller escreveu:

Quoted 10 lines

From: Arnaldo Carvalho de Melo <[email protected]>
Date: Thu, 11 Jun 2009 00:40:22 -0300
Review
Quoted 2 lines
(quoted diff) +2 -0
+	struct compat_timespec *timeout_compat =
+		(struct compat_timespec *)timeout;
In order for this to build on non-compat platforms, or when CONFIG_COMPAT is disabled, you'll need to avoid trying to reference compat_timespec.

Quoted 5 lines

We get away with compat_msghdr because we define it as a NOP in net/compat.h when necessary.
Doing the same, and cleanly, with this non-networking compat knob is unlikely to be straightforward.

Better now?

Paul, this one also has the sock_recvmsg_nosec rewrite following your suggestion.

I'll write a proper changeset log once we get the more easy stuff out of the way and people are confortable with it.

Ah, here is how things are being distributed now in a test setup with two clients and running recvmmsg like this:

[acme@emilia git]$ perf record ./recvmmsg 5001 256 5000
usage: recvmmsg <port(def 5001)> <batch_size(def 8)> <timeout(def 10ms>

Waiting for datagrams...
nr_datagrams received: 256, remaining: 4416588ns
5376 bytes received from filo.ghostprotocols.net in 21 datagrams 768 bytes received from doppio.ghostprotocols.net in 3 datagrams 256 bytes received from filo.ghostprotocols.net in 1 datagrams 2304 bytes received from doppio.ghostprotocols.net in 9 datagrams 256 bytes received from filo.ghostprotocols.net in 1 datagrams 2560 bytes received from doppio.ghostprotocols.net in 10 datagrams 256 bytes received from filo.ghostprotocols.net in 1 datagrams
<SNIP>

[acme@emilia git]$ perf report --sort comm,dso,symbol | head -30

#
# (4532588 samples)
#

# Overhead            Command  Shared Object     Symbol
# ........ .........  .........................  ......
#
     4.85%  recvmmsg  [kernel]                   [k] copy_user_generic_string
     3.17%  recvmmsg  [kernel]                   [k] _spin_lock_irqsave
     2.35%  recvmmsg  [kernel]                   [k] _spin_lock
     2.18%  recvmmsg  [kernel]                   [k] udp_recvmsg
     2.04%  recvmmsg  /lib64/libnss_files-2.5.so [.] internal_getent
     1.52%  recvmmsg  [kernel]                   [k] n_tty_write
     1.49%  recvmmsg  [kernel]                   [k] _spin_lock_bh
     1.44%  recvmmsg  [kernel]                   [k] __rcu_read_lock
     1.30%  recvmmsg  [kernel]                   [k] __skb_recv_datagram
     1.23%  recvmmsg  [kernel]                   [k] __rcu_read_unlock
     1.14%  recvmmsg  [kernel]                   [k] __cache_free
     1.14%  recvmmsg  [kernel]                   [k] avc_has_perm_noaudit
     1.12%  recvmmsg  /lib64/libc-2.5.so         [.] _IO_vfprintf_internal
     1.10%  recvmmsg  [kernel]                   [k] __sys_recvmsg
     1.06%  recvmmsg  ./recvmmsg                 [.] main
     1.04%  recvmmsg  [kernel]                   [k] csum_partial
     1.02%  recvmmsg  [kernel]                   [k] _spin_unlock_irqrestore
     0.98%  recvmmsg  [kernel]                   [k] n_tty_receive_buf
     0.88%  recvmmsg  /lib64/libc-2.5.so         [.] __GI_inet_pton
     0.87%  recvmmsg  [kernel]                   [k] system_call
     0.86%  recvmmsg  [kernel]                   [k] ktime_get_ts
     0.83%  recvmmsg  /lib64/libc-2.5.so         [.] __memchr
     0.76%  recvmmsg  /lib64/libc-2.5.so         [.] _IO_file_xsputn_internal

[acme@emilia git]$

We'll see after sk_prot->unlocked_recvmsg gets reintroduced and I remove silly stuff such as fprintf from the test app... :-)

Thanks!

- Arnaldo

arch/x86/ia32/ia32entry.S +1 -0

index e590261..2a188e5 100644
@@ -832,4 +832,5 @@ ia32_sys_call_table:
 	.quad compat_sys_pwritev
 	.quad compat_sys_rt_tgsigqueueinfo	/* 335 */
 	.quad sys_perf_counter_open
+	.quad compat_sys_recvmmsg
 ia32_syscall_end:

arch/x86/include/asm/unistd_32.h +1 -0

index 732a307..3e72cae 100644
@@ -342,6 +342,7 @@
 #define __NR_pwritev		334
 #define __NR_rt_tgsigqueueinfo	335
 #define __NR_perf_counter_open	336
+#define __NR_recvmmsg		337
 
 #ifdef __KERNEL__

arch/x86/include/asm/unistd_64.h +2 -0

index 900e161..713a32a 100644
@@ -661,6 +661,8 @@ __SYSCALL(__NR_pwritev, sys_pwritev)
 __SYSCALL(__NR_rt_tgsigqueueinfo, sys_rt_tgsigqueueinfo)
 #define __NR_perf_counter_open			298
 __SYSCALL(__NR_perf_counter_open, sys_perf_counter_open)
+#define __NR_recvmmsg				299
+__SYSCALL(__NR_recvmmsg, sys_recvmmsg)
 
 #ifndef __NO_STUBS
 #define __ARCH_WANT_OLD_READDIR

arch/x86/kernel/syscall_table_32.S +1 -0

index d51321d..4881b14 100644
@@ -336,3 +336,4 @@ ENTRY(sys_call_table)
 	.long sys_pwritev
 	.long sys_rt_tgsigqueueinfo	/* 335 */
 	.long sys_perf_counter_open
+	.long sys_recvmmsg

include/linux/socket.h +10 -0

index 421afb4..5ed3644 100644
@@ -65,6 +65,12 @@ struct msghdr {
 	unsigned	msg_flags;
 };
 
+/* For recvmmsg/sendmmsg */
+struct mmsghdr {
+	struct msghdr   msg_hdr;
+	unsigned        msg_len;
+};
+
 /*
  *	POSIX 1003.1g - ancillary data object information
  *	Ancillary data consits of a sequence of pairs of

@@ -322,6 +328,10 @@ extern int move_addr_to_user(struct sockaddr *kaddr, int klen, void __user *uadd
 extern int move_addr_to_kernel(void __user *uaddr, int ulen, struct sockaddr *kaddr);
 extern int put_cmsg(struct msghdr*, int level, int type, int len, void *data);
 
+struct timespec;
+
+extern int __sys_recvmmsg(int fd, struct mmsghdr __user *mmsg, unsigned int vlen,
+			  unsigned int flags, struct timespec *timeout);
 #endif
 #endif /* not kernel and not glibc */
 #endif /* _LINUX_SOCKET_H */

include/linux/syscalls.h +4 -0

index c6c84ad..afefa61 100644
@@ -25,6 +25,7 @@ struct linux_dirent64;
 struct list_head;
 struct msgbuf;
 struct msghdr;
+struct mmsghdr;
 struct msqid_ds;
 struct new_utsname;
 struct nfsctl_arg;

@@ -556,6 +557,9 @@ asmlinkage long sys_recv(int, void __user *, size_t, unsigned);
 asmlinkage long sys_recvfrom(int, void __user *, size_t, unsigned,
 				struct sockaddr __user *, int __user *);
 asmlinkage long sys_recvmsg(int fd, struct msghdr __user *msg, unsigned flags);
+asmlinkage long sys_recvmmsg(int fd, struct mmsghdr __user *msg,
+			     unsigned int vlen, unsigned flags,
+			     struct timespec __user *timeout);
 asmlinkage long sys_socket(int, int, int);
 asmlinkage long sys_socketpair(int, int, int, int __user *);
 asmlinkage long sys_socketcall(int call, unsigned long __user *args);

include/net/compat.h +8 -0

index 5bbf8bf..96c38d8 100644
@@ -18,6 +18,11 @@ struct compat_msghdr {
 	compat_uint_t	msg_flags;
 };
 
+struct compat_mmsghdr {
+	struct compat_msghdr msg_hdr;
+	compat_uint_t        msg_len;
+};
+
 struct compat_cmsghdr {
 	compat_size_t	cmsg_len;
 	compat_int_t	cmsg_level;

@@ -35,6 +40,9 @@ extern int get_compat_msghdr(struct msghdr *, struct compat_msghdr __user *);
 extern int verify_compat_iovec(struct msghdr *, struct iovec *, struct sockaddr *, int);
 extern asmlinkage long compat_sys_sendmsg(int,struct compat_msghdr __user *,unsigned);
 extern asmlinkage long compat_sys_recvmsg(int,struct compat_msghdr __user *,unsigned);
+extern asmlinkage long compat_sys_recvmmsg(int, struct compat_mmsghdr __user *,
+					   unsigned, unsigned,
+					   struct timespec __user *);
 extern asmlinkage long compat_sys_getsockopt(int, int, int, char __user *, int __user *);
 extern int put_cmsg_compat(struct msghdr*, int, int, int, void *);

kernel/sys_ni.c +2 -0

index 68320f6..f581fb0 100644
@@ -48,7 +48,9 @@ cond_syscall(sys_shutdown);
 cond_syscall(sys_sendmsg);
 cond_syscall(compat_sys_sendmsg);
 cond_syscall(sys_recvmsg);
+cond_syscall(sys_recvmmsg);
 cond_syscall(compat_sys_recvmsg);
+cond_syscall(compat_sys_recvmmsg);
 cond_syscall(sys_socketcall);
 cond_syscall(sys_futex);
 cond_syscall(compat_sys_futex);

net/compat.c +23 -0

index 8d73905..6f04f2d 100644
@@ -743,6 +743,29 @@ asmlinkage long compat_sys_recvmsg(int fd, struct compat_msghdr __user *msg, uns
 	return sys_recvmsg(fd, (struct msghdr __user *)msg, flags | MSG_CMSG_COMPAT);
 }
 
+asmlinkage long compat_sys_recvmmsg(int fd, struct compat_mmsghdr __user *mmsg,
+				    unsigned vlen, unsigned int flags,
+				    struct timespec __user *timeout)
+{
+	int datagrams;
+	struct timespec ktspec;
+	struct compat_timespec __user *utspec =
+			(struct compat_timespec __user *)timeout;
+
+	if (get_user(ktspec.tv_sec, &utspec->tv_sec) ||
+	    get_user(ktspec.tv_nsec, &utspec->tv_nsec))
+		return -EFAULT;
+
+	datagrams = __sys_recvmmsg(fd, (struct mmsghdr __user *)mmsg, vlen,
+				   flags | MSG_CMSG_COMPAT, &ktspec);
+	if (datagrams > 0 &&
+	    put_user(ktspec.tv_sec, &utspec->tv_sec) ||
+	    put_user(ktspec.tv_nsec, &utspec->tv_nsec))
+		datagrams = -EFAULT;
+
+	return datagrams;
+}
+
 asmlinkage long compat_sys_socketcall(int call, u32 __user *args)
 {
 	int ret;

net/socket.c +181 ~~-38~~

317 lines not on this page · raw message

Paul Moore

2009-06-11 18:09 UTC #5 lore raw

In reply to Arnaldo Carvalho de Melo

Quoted 19 lines · Arnaldo Carvalho de Melo

net/socket.c +11 -0

index 791d71a..f9f1e20 100644
@@ -702,6 +702,28 @@ int sock_recvmsg(struct socket *sock, struct msghdr
 *msg, return ret;
 }
 
+static int sock_recvmsg_nosec(struct socket *sock, struct msghdr *msg,
+			      size_t size, int flags)
+{
+	struct kiocb iocb;
+	struct sock_iocb siocb;
+	int ret;
+
+	init_sync_kiocb(&iocb, NULL);
+	iocb.private = &siocb;
+
+	siocb.sock = sock;

net/socket.c int sock_recvmsg(struct socket *sock, struct msghdr · line 725

Quoted 10 lines

net/socket.c +10 -0

+	siocb.scm = NULL;
+	siocb.msg = msg;
+	siocb.size = size;
+	siocb.flags = flags;
+
+	ret = sock->ops->recvmsg(&iocb, sock, msg, size, flags);
+	if (-EIOCBQUEUED == ret)
+		ret = wait_on_sync_kiocb(&iocb);
+	return ret;
+}

Hmmm, in an effort to reduce duplicated code how about updating __sock_recvmsg() to something like the following:

static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
struct msghdr *msg, size_t size, int flags)

{
        int err;

        err = security_socket_recvmsg(...);
        if (err)
                return err;

        return sock_recvmsg_nosec(...);
}

The only real difference is that now the *_kiocb() functions get called and I have no clue if that is good or bad but it is different :)

Quoted 22 lines

net/socket.c +3 -3

/*

@@ -2018,46 +2029,47 @@ SYSCALL_DEFINE3(recvmsg, int, fd, struct msghdr
 __user *, msg, *      kernel msghdr to use the kernel address space)
 	 */
 
-	uaddr = (__force void __user *)msg_sys.msg_name;
+	uaddr = (__force void __user *)msg_sys->msg_name;
 	uaddr_len = COMPAT_NAMELEN(msg);
 	if (MSG_CMSG_COMPAT & flags) {
-		err = verify_compat_iovec(&msg_sys, iov,
+		err = verify_compat_iovec(msg_sys, iov,
 					  (struct sockaddr *)&addr,
 					  VERIFY_WRITE);
 	} else
-		err = verify_iovec(&msg_sys, iov,
+		err = verify_iovec(msg_sys, iov,
 				   (struct sockaddr *)&addr,
 				   VERIFY_WRITE);
 	if (err < 0)
 		goto out_freeiov;
 	total_len = err;

net/socket.c SYSCALL_DEFINE3(recvmsg, int, fd, struct msghdr · line 2052

Quoted 10 lines

net/socket.c +4 -3

-	cmsg_ptr = (unsigned long)msg_sys.msg_control;
-	msg_sys.msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
+	cmsg_ptr = (unsigned long)msg_sys->msg_control;
+	msg_sys->msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
 
 	if (sock->file->f_flags & O_NONBLOCK)
 		flags |= MSG_DONTWAIT;
-	err = sock_recvmsg(sock, &msg_sys, total_len, flags);
+	err = (nosec ? sock_recvmsg_nosec : sock_recvmsg)(sock, msg_sys,
+							  total_len, flags);

Perhaps I'm just being nit-picky here but why not this (it is much easier on my eyes at least <g>):

	if (nosec)
		err = sock_recvmsg_nosec(...);
	else
		err = sock_recvmsg(...);

Signature · 2 lines

paul moore
linux @ hp

Arnaldo Carvalho de Melo

author

2009-06-11 21:54 UTC #6 lore raw

Em Thu, Jun 11, 2009 at 02:09:22PM -0400, Paul Moore escreveu:

Quoted 46 lines · Arnaldo Carvalho de Melo

Quoted 19 lines

net/socket.c +11 -0

index 791d71a..f9f1e20 100644
@@ -702,6 +702,28 @@ int sock_recvmsg(struct socket *sock, struct msghdr
 *msg, return ret;
 }
 
+static int sock_recvmsg_nosec(struct socket *sock, struct msghdr *msg,
+			      size_t size, int flags)
+{
+	struct kiocb iocb;
+	struct sock_iocb siocb;
+	int ret;
+
+	init_sync_kiocb(&iocb, NULL);
+	iocb.private = &siocb;
+
+	siocb.sock = sock;

net/socket.c int sock_recvmsg(struct socket *sock, struct msghdr · line 725

Quoted 10 lines

net/socket.c +10 -0

+	siocb.scm = NULL;
+	siocb.msg = msg;
+	siocb.size = size;
+	siocb.flags = flags;
+
+	ret = sock->ops->recvmsg(&iocb, sock, msg, size, flags);
+	if (-EIOCBQUEUED == ret)
+		ret = wait_on_sync_kiocb(&iocb);
+	return ret;
+}

Hmmm, in an effort to reduce duplicated code how about updating __sock_recvmsg() to something like the following:

static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
struct msghdr *msg, size_t size, int flags)

{
        int err;

        err = security_socket_recvmsg(...);
        if (err)
                return err;

        return sock_recvmsg_nosec(...);
}

Quoted 2 lines

The only real difference is that now the *_kiocb() functions get called and I have no clue if that is good or bad but it is different :)

Yeah, gets clearer, like this:

static inline int __sock_recvmsg_nosec(struct kiocb *iocb, struct socket *sock,
struct msghdr *msg, size_t size, int flags)

{
	struct sock_iocb *si = kiocb_to_siocb(iocb);

	si->sock = sock;
	si->scm = NULL;
	si->msg = msg;
	si->size = size;
	si->flags = flags;

	return sock->ops->recvmsg(iocb, sock, msg, size, flags);
}

static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
struct msghdr *msg, size_t size, int flags)

{
	int err = security_socket_recvmsg(sock, msg, size, flags);

	return err ?: __sock_recvmsg_nosec(iocb, sock, msg, size, flags);
}

static int sock_recvmsg_nosec(struct socket *sock, struct msghdr *msg,
size_t size, int flags)

{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

	init_sync_kiocb(&iocb, NULL);
	iocb.private = &siocb;
	ret = __sock_recvmsg_nosec(&iocb, sock, msg, size, flags);
	if (-EIOCBQUEUED == ret)
		ret = wait_on_sync_kiocb(&iocb);
	return ret;
}

Better now? :-)

Quoted 36 lines

Quoted 22 lines

net/socket.c +3 -3

/*

@@ -2018,46 +2029,47 @@ SYSCALL_DEFINE3(recvmsg, int, fd, struct msghdr
 __user *, msg, *      kernel msghdr to use the kernel address space)
 	 */
 
-	uaddr = (__force void __user *)msg_sys.msg_name;
+	uaddr = (__force void __user *)msg_sys->msg_name;
 	uaddr_len = COMPAT_NAMELEN(msg);
 	if (MSG_CMSG_COMPAT & flags) {
-		err = verify_compat_iovec(&msg_sys, iov,
+		err = verify_compat_iovec(msg_sys, iov,
 					  (struct sockaddr *)&addr,
 					  VERIFY_WRITE);
 	} else
-		err = verify_iovec(&msg_sys, iov,
+		err = verify_iovec(msg_sys, iov,
 				   (struct sockaddr *)&addr,
 				   VERIFY_WRITE);
 	if (err < 0)
 		goto out_freeiov;
 	total_len = err;

net/socket.c SYSCALL_DEFINE3(recvmsg, int, fd, struct msghdr · line 2052

Quoted 10 lines

net/socket.c +4 -3

-	cmsg_ptr = (unsigned long)msg_sys.msg_control;
-	msg_sys.msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
+	cmsg_ptr = (unsigned long)msg_sys->msg_control;
+	msg_sys->msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
 
 	if (sock->file->f_flags & O_NONBLOCK)
 		flags |= MSG_DONTWAIT;
-	err = sock_recvmsg(sock, &msg_sys, total_len, flags);
+	err = (nosec ? sock_recvmsg_nosec : sock_recvmsg)(sock, msg_sys,
+							  total_len, flags);

Perhaps I'm just being nit-picky here but why not this (it is much easier on my eyes at least <g>):

Quoted 4 lines

	if (nosec)
		err = sock_recvmsg_nosec(...);
	else
		err = sock_recvmsg(...);

Well, its like "if (foo)" versus "if (foo != NULL)", I prefer to reduce the number of source code lines and stress that the parameter list is the same, anybody else feels confused by this?

- Arnaldo

Paul Moore

2009-06-16 20:36 UTC #7 lore raw

Quoted 89 lines · Arnaldo Carvalho de Melo

Em Thu, Jun 11, 2009 at 02:09:22PM -0400, Paul Moore escreveu:
Quoted 46 lines · Arnaldo Carvalho de Melo
Quoted 19 lines
net/socket.c +11 -0
index 791d71a..f9f1e20 100644
@@ -702,6 +702,28 @@ int sock_recvmsg(struct socket *sock, struct
 msghdr *msg, return ret;
 }
 
+static int sock_recvmsg_nosec(struct socket *sock, struct msghdr *msg,
+			      size_t size, int flags)
+{
+	struct kiocb iocb;
+	struct sock_iocb siocb;
+	int ret;
+
+	init_sync_kiocb(&iocb, NULL);
+	iocb.private = &siocb;
+
+	siocb.sock = sock;
net/socket.c int sock_recvmsg(struct socket *sock, struct · line 725
Quoted 10 lines
net/socket.c +10 -0
+	siocb.scm = NULL;
+	siocb.msg = msg;
+	siocb.size = size;
+	siocb.flags = flags;
+
+	ret = sock->ops->recvmsg(&iocb, sock, msg, size, flags);
+	if (-EIOCBQUEUED == ret)
+		ret = wait_on_sync_kiocb(&iocb);
+	return ret;
+}
Hmmm, in an effort to reduce duplicated code how about updating __sock_recvmsg() to something like the following:
static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
struct msghdr *msg, size_t size, int
flags) {
        int err;

        err = security_socket_recvmsg(...);
        if (err)
                return err;

        return sock_recvmsg_nosec(...);
}
Quoted 2 lines
The only real difference is that now the *_kiocb() functions get called and I have no clue if that is good or bad but it is different :)
Yeah, gets clearer, like this:
static inline int __sock_recvmsg_nosec(struct kiocb *iocb, struct socket *sock, struct msghdr *msg, size_t size, int flags)
{
	struct sock_iocb *si = kiocb_to_siocb(iocb);

	si->sock = sock;
	si->scm = NULL;
	si->msg = msg;
	si->size = size;
	si->flags = flags;

	return sock->ops->recvmsg(iocb, sock, msg, size, flags);
}
static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
struct msghdr *msg, size_t size, int flags)
{
	int err = security_socket_recvmsg(sock, msg, size, flags);

	return err ?: __sock_recvmsg_nosec(iocb, sock, msg, size, flags);
}
static int sock_recvmsg_nosec(struct socket *sock, struct msghdr *msg,
size_t size, int flags)
{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

	init_sync_kiocb(&iocb, NULL);
	iocb.private = &siocb;
	ret = __sock_recvmsg_nosec(&iocb, sock, msg, size, flags);
	if (-EIOCBQUEUED == ret)
		ret = wait_on_sync_kiocb(&iocb);
	return ret;
}

Quoted 1 line

Better now? :-)

Sorry for the delay, I was away for a few days and just got back to email this morning ... but yes, much better now, thanks :)

Signature · 2 lines

paul moore
linux @ hp

Rémi Denis-Courmont

2009-06-12 07:26 UTC #8 lore raw

In reply to Arnaldo Carvalho de Melo

arch/x86/include/asm/unistd_64.h __SYSCALL(__NR_pwritev, sys_pwritev) · line 665

Quoted 10 lines · On Thursday 11 June 2009 06:40:22 ext Arnaldo Carvalho de Melo wrote:

(quoted diff) +0 -0

b/arch/x86/include/asm/unistd_64.h index 900e161..713a32a 100644

arch/x86/include/asm/unistd_64.h +2 -0

@@ -661,6 +661,8 @@ __SYSCALL(__NR_pwritev, sys_pwritev)
 __SYSCALL(__NR_rt_tgsigqueueinfo, sys_rt_tgsigqueueinfo)
 #define __NR_perf_counter_open                 298
 __SYSCALL(__NR_perf_counter_open, sys_perf_counter_open)
+#define __NR_recvmmsg                          299
+__SYSCALL(__NR_recvmmsg, sys_recvmmsg)

I guess socketcall is deprecated in favor of full syscalls, then? (sorry if this is a stupid question)

arch/x86/include/asm/unistd_64.h __SYSCALL(__NR_pwritev, sys_pwritev)

Quoted 4 lines

arch/x86/include/asm/unistd_64.h +4 -0

+       if (timeout) {
+               /* Doesn't make much sense */
+               if (flags & MSG_DONTWAIT)
+                       return -EINVAL;

An application could possibly hit this degenerated case at the end of a loop or whatever, and EINVAL makes it look like this is a bug. Why not EAGAIN?

Signature · 2 lines

Rémi Denis-Courmont
Nokia Devices R&D, Maemo Software, Helsinki

Arnaldo Carvalho de Melo

author

2009-06-12 14:19 UTC #9 lore raw

Em Fri, Jun 12, 2009 at 10:26:25AM +0300, Rémi Denis-Courmont escreveu:

Quoted 6 lines · On Thursday 11 June 2009 06:40:22 ext Arnaldo Carvalho de Melo wrote:

Quoted 5 lines
(quoted diff) +0 -0
b/arch/x86/include/asm/unistd_64.h index 900e161..713a32a 100644
arch/x86/include/asm/unistd_64.h +0 -0
@@ -661,6 +661,8 @@ __SYSCALL(__NR_pwritev, sys_pwritev)

Quoted 8 lines

arch/x86/include/asm/unistd_64.h __SYSCALL(__NR_pwritev, sys_pwritev) · line 665
Quoted 5 lines
arch/x86/include/asm/unistd_64.h +2 -0
 __SYSCALL(__NR_rt_tgsigqueueinfo, sys_rt_tgsigqueueinfo)
 #define __NR_perf_counter_open                 298
 __SYSCALL(__NR_perf_counter_open, sys_perf_counter_open)
+#define __NR_recvmmsg                          299
+__SYSCALL(__NR_recvmmsg, sys_recvmmsg)
I guess socketcall is deprecated in favor of full syscalls, then? (sorry if this is a stupid question)

In some architectures, like x86_64, yes, but some still need it, the ones that set __ARCH_WANT_SYS_SOCKETCALL:

[acme@doppio linux-2.6-tip]$ echo $(find . -type f | xargs egrep "define.+__ARCH_WANT_SYS_SOCKETCALL" | sed -r 's#./arch/(\w+)/.*#\1#g' | sort)
arm cris frv h8300 m32r m68k microblaze mips mn10300 parisc powerpc s390 sh sh sparc x86 x86

So the alpha and ia64, for instance, doesn't want sys_socketcall:

[acme@doppio linux-2.6-tip]$ egrep 'sys_((recv|send)msg|[gs]etsockopt)' arch/alpha/kernel/systbls.S

	.quad sys_setsockopt			/* 105 */
	.quad sys_recvmsg
	.quad sys_sendmsg
	.quad sys_getsockopt

[acme@doppio linux-2.6-tip]$ egrep sys_socketcall arch/alpha/kernel/systbls.S [acme@doppio linux-2.6-tip]$

[acme@doppio linux-2.6-tip]$ egrep 'sys_((recv|send)msg|[gs]etsockopt)' arch/ia64/kernel/entry.S

	data8 sys_setsockopt
	data8 sys_getsockopt
	data8 sys_sendmsg			// 1205
	data8 sys_recvmsg

[acme@doppio linux-2.6-tip]$ egrep sys_socketcall
arch/ia64/kernel/entry.S
[acme@doppio linux-2.6-tip]$

But some define __ARCH_WANT_SYS_SOCKETCALL conditionally, like x86_64 and um:

[acme@doppio linux-2.6-tip]$ find arch -type f | xargs egrep 'define.+__NO_STUBS' arch/x86/kernel/syscall_64.c:#define __NO_STUBS arch/x86/kernel/asm-offsets_64.c:#define __NO_STUBS 1 arch/um/sys-x86_64/syscall_table.c:#define __NO_STUBS
[acme@doppio linux-2.6-tip]$

And others want socketcall if preserving old ABIs, like ARM:

arch/arm/include/asm/unistd.h

#if !defined(CONFIG_AEABI) || defined(CONFIG_OABI_COMPAT) #define __ARCH_WANT_SYS_TIME
#define __ARCH_WANT_SYS_OLDUMOUNT
#define __ARCH_WANT_SYS_ALARM
#define __ARCH_WANT_SYS_UTIME
#define __ARCH_WANT_SYS_OLD_GETRLIMIT
#define __ARCH_WANT_OLD_READDIR
#define __ARCH_WANT_SYS_SOCKETCALL
#endif

So for the final revision indeed I need to provide a sys_socketcall interface to sys_recvmmsg.

Quoted 7 lines

arch/x86/include/asm/unistd_64.h __SYSCALL(__NR_pwritev, sys_pwritev)
Quoted 4 lines
arch/x86/include/asm/unistd_64.h +4 -0
+       if (timeout) {
+               /* Doesn't make much sense */
+               if (flags & MSG_DONTWAIT)
+                       return -EINVAL;
An application could possibly hit this degenerated case at the end of a loop or whatever, and EINVAL makes it look like this is a bug. Why not EAGAIN?

EAGAIN for me looks like something that when repeated would possibly produce a valid result, but in this case it will never be that way.

But lemme look around, perhaps this assumption of mine is invalidated by some strange standard... EAGAIN = EWOULDBLOCK, and in the current implementation it wouldn't block at all, because recvmsg would return right away...

The only semantic we could associate to this would be to make recvmmsg call, in a non blocking way recvmsg (as specified in the flags) and if it returns EAGAIN go to sleep waiting for more packets, i.e. a mixed blocking recvmmsg with a noblocking recvmsg, sounds of creeping featuritis...

Or perhaps just remove this check and let recvmmsg return imediately, the app could then use this degenerated case to measure how long it took for the packets to be received, looks like the most liberal and removes 3 lines of source code and a branch :-\

Keep it simple?

- Arnaldo