Does FreeBSD have sendmmsg or recvmmsg system calls?
Boris Astardzhiev
boris.astardzhiev at gmail.com
Fri Jan 22 09:15:21 UTC 2016
be>None of the above. Plain recvmsg() returns ssize_t and its len arg has
be>type size_t. That is excessively typedefed and excessively large with
be>64-bit ssize_t, but it is silly for the multiple-message variant to use
be>smaller types.
be>
be>Otherwise, all the integer types should be int.
It seems logical. I'll convert the ret easily to ssize_t and the vector
length
to size_t. Now it differs from the Linux prototype but I guess it's okay.
be>The errno method (and not checking ret at all) is best if for syscalls
that
be>return -1 for a non-error. It is not needed here.
Fixing it.
kb> I do not see any sense in making the functions with signature or
semantic
kb> different from Linux version. Right now, the goal of including the
patch
kb> is compatibility.
Regarding recvmmsg() -
I tried to implement MSG_WAITFORONE and the timeout stuff using pselect(2)
due to the timespec structure. I could have used ppoll and I'm not sure
which
of these two is more appropriate or maybe there's another approach?
Now it has timeout just as in the Linux prototype.
Comments are welcomed.
See patch.
Best regards,
Boris Astardzhiev
On Thu, Jan 21, 2016 at 3:27 PM, Bruce Evans <brde at optusnet.com.au> wrote:
> On Thu, 21 Jan 2016, Konstantin Belousov wrote:
>
> On Wed, Jan 20, 2016 at 10:29:47AM +0200, Boris Astardzhiev wrote:
>>
>
> kb>Shouldn't i and rcvd be unsigned as well ? Shouldn't return value
>>> kb>also be unsigned ?
>>> I think i and rcvd should be unsigned whereas ret should not - after all
>>> if an error occurred we get -1.
>>>
>> I looked at the real signatures and man pages for the Linux functions,
>> finally. There is indeed the timeout arg, the MSG_WAITFORONE flag for
>> recvmmsg(3), and Linux uses ints for what would be naturally size_t.
>>
>
> It knows better than POSIX, so has restored the v7 types :-). Probably
> not intentionally.
>
> FreeBSD-1 used the v7 types for recv* in FreeBSD-1, but this was changed
> in 4.4BSD-Lite1.
>
> kb>> + rcvd = 0;
>>> kb>> + for (i = 0; i < vlen; i++) {
>>> kb>> + errno = 0;
>>> kb>> + ret = __sys_recvmsg(s, &msgvec[i].msg_hdr, flags);
>>> kb>> + if (ret < 0 || errno != 0) {
>>> kb>I do not see why do you need to clear errno before, and then do this
>>> test.
>>> kb>Just check ret == -1, in which case errno was set from the immediate
>>> syscall.
>>> kb>
>>> kb>> + if (rcvd != 0) {
>>> kb>> + /* We've received messages. Let caller
>>> know. */
>>> kb>> + errno = 0;
>>> kb>This cleaning is not needed as well. For successfull functions
>>> returns,
>>> kb>errno value is undefined.
>>>
>>> Wouldn't I confuse apps if they check errno in the follow case - I want
>>> to
>>> receive two messages. The first __sys_recvmsg succeeds and then for the
>>> second __sys_recvmsg fails. Thus errno will be != 0 and I'm telling the
>>> app
>>> that I have received one message by returning 1 but errno will be != 0.
>>> Is this correct?
>>>
>>
>> errno value is only defined after the function explicitely returned error.
>> Apps which test for errno without testing for error are wrong.
>>
>
> Mostly such tests find wrong implementations. Only low quality
> implmentations set error for non-errors.
>
> I checked some details:
>
> - both C99 and POSIX.1-2001 forbid setting errno to 0 in library functions
>
> - C99 and draft C11 explicitly say that errno may be clobbered for
> non-errors by functions that are not documented to set errno. It
> intends to also say that errno may not be clobbered for non-errors by
> functions are documented to set errno, but actually says nothing
> about this case (except in a footnote, it says "Thus [the usual
> method of setting errno before the call and checking it after
> ``should'' be used to check for errors]." Footnotes are not part
> of the standard, so this also says nothing. But it gives the intent.
>
> - C99 doesn't say anything specific about this for strtol. But strtol
> needs errno to not be set on non-error for its API. This gives the
> intent of the documentation for errno in the standard itself.
>
> - POSIX.1-2001 says that applications "should" only be examined when it is
> indicated to be valid by a function's return value. This is wrong for
> all functions where errno is part of the API (like strtol), and it
> would be inconsistent with C99 and C11 if these standards said what they
> intend to say, for many more functions. First, all of the C99 functions
> that are documented to set errno. Then, if POSIX follows the spirit of
> the intent of C99, then just about all POSIX functions must not clobber
> errno for non-errors, since (unlike in C99) just about all functions in
> POSIX are documented to set errno.
>
> - POSIX.1-2001 apparently agrees with me that C99 doesn't say what it
> intends to say. It explicitly says that strtol shall not change "t
> setting of" errno on success, and marks this as an extension of C99.
>
> - I couldn't find even a buggy generic clause in POSIX.1-2001 saying
> what C99 intends to say even for the C99 parts of POSIX. If there
> were such a clause, then strtol wouldn't need a special extension.
> This means that some other C99 functions need special clauses, and
> there are likely to be bugs in this since some don't really need
> them except to conform with C99's intentions.
>
> - not many C99 functions are specified to set errno. The main ones
> are the strtol family, math functions (most can set ERANGE or EDOM),
> and wide character functions (many can set EILSEQ). POSIX.1-2001
> seems to be missing a special clause for the first wide character
> function that I looked at -- fputwc(). For math functions, it has
> special statements ad nauseum (about 20 lines per function for
> duplicated for 20-50 functions).
>
> Bruce
>
-------------- next part --------------
diff --git a/lib/libc/include/namespace.h b/lib/libc/include/namespace.h
index 739d7b1..c95829e 100644
--- a/lib/libc/include/namespace.h
+++ b/lib/libc/include/namespace.h
@@ -208,6 +208,7 @@
#define readv _readv
#define recvfrom _recvfrom
#define recvmsg _recvmsg
+#define recvmmsg _recvmmsg
#define select _select
#define sem_close _sem_close
#define sem_destroy _sem_destroy
@@ -220,6 +221,7 @@
#define sem_unlink _sem_unlink
#define sem_wait _sem_wait
#define sendmsg _sendmsg
+#define sendmmsg _sendmmsg
#define sendto _sendto
#define setsockopt _setsockopt
/*#define sigaction _sigaction*/
diff --git a/lib/libc/include/un-namespace.h b/lib/libc/include/un-namespace.h
index f31fa7a..0233348 100644
--- a/lib/libc/include/un-namespace.h
+++ b/lib/libc/include/un-namespace.h
@@ -189,6 +189,7 @@
#undef readv
#undef recvfrom
#undef recvmsg
+#undef recvmmsg
#undef select
#undef sem_close
#undef sem_destroy
@@ -201,6 +202,7 @@
#undef sem_unlink
#undef sem_wait
#undef sendmsg
+#undef sendmmsg
#undef sendto
#undef setsockopt
#undef sigaction
diff --git a/lib/libc/sys/Makefile.inc b/lib/libc/sys/Makefile.inc
index e4fe1b2..5f8b699 100644
--- a/lib/libc/sys/Makefile.inc
+++ b/lib/libc/sys/Makefile.inc
@@ -28,6 +28,8 @@ SRCS+= futimens.c utimensat.c
NOASM+= futimens.o utimensat.o
PSEUDO+= _futimens.o _utimensat.o
+SRCS+= recvmmsg.c sendmmsg.c
+
INTERPOSED = \
accept \
accept4 \
diff --git a/lib/libc/sys/Symbol.map b/lib/libc/sys/Symbol.map
index 7b3257c..dc2ed0e 100644
--- a/lib/libc/sys/Symbol.map
+++ b/lib/libc/sys/Symbol.map
@@ -399,6 +399,8 @@ FBSD_1.4 {
utimensat;
numa_setaffinity;
numa_getaffinity;
+ sendmmsg;
+ recvmmsg;
};
FBSDprivate_1.0 {
diff --git a/lib/libc/sys/recv.2 b/lib/libc/sys/recv.2
index 326e7ff..fd2b2a1 100644
--- a/lib/libc/sys/recv.2
+++ b/lib/libc/sys/recv.2
@@ -34,8 +34,9 @@
.Sh NAME
.Nm recv ,
.Nm recvfrom ,
-.Nm recvmsg
-.Nd receive a message from a socket
+.Nm recvmsg ,
+.Nm recvmmsg
+.Nd receive message(s) from a socket
.Sh LIBRARY
.Lb libc
.Sh SYNOPSIS
@@ -47,11 +48,15 @@
.Fn recvfrom "int s" "void *buf" "size_t len" "int flags" "struct sockaddr * restrict from" "socklen_t * restrict fromlen"
.Ft ssize_t
.Fn recvmsg "int s" "struct msghdr *msg" "int flags"
+.Ft ssize_t
+.Fn recvmmsg "int s" "struct mmsghdr *msgvec" "size_t vlen" "int flags" "const struct timespec *timeout"
.Sh DESCRIPTION
The
.Fn recvfrom
and
.Fn recvmsg
+and
+.Fn recvmmsg
system calls
are used to receive messages from a socket,
and may be used to receive data on a socket whether or not
@@ -84,8 +89,30 @@ null pointer passed as its
.Fa from
argument.
.Pp
-All three routines return the length of the message on successful
-completion.
+The
+.Fn recvmmsg
+function is used to receive multiple
+messages at a call.
+Their number
+is supplied by
+.Fa vlen .
+The messages are placed in the
+.Fa msgvec
+vector after reception.
+The size of each received message is placed in the
+.Fa msg_len
+field of each element of the vector.
+If
+.Fa timeout
+is NULL the call will normally block. Otherwise it will wait for data
+for the specified amount of time. If the timeout expires and there is
+no data received a value of 0 is returned. pselect(2) is used for the
+implementation of the timeout mechanism.
+.Pp
+The first three routines return the length of the message on successful
+completion whereas
+.Fn recvmmsg
+returns the number of received messages.
If a message is too long to fit in the supplied buffer,
excess bytes may be discarded depending on the type of socket
the message is received from (see
@@ -100,7 +127,9 @@ in which case the value
.Va errno
is set to
.Er EAGAIN .
-The receive calls normally return any data available,
+The receive calls except
+.Fn recvmmsg
+normally return any data available,
up to the requested amount,
rather than waiting for receipt of the full amount requested;
this behavior is affected by the socket-level options
@@ -127,6 +156,9 @@ one or more of the values:
.It Dv MSG_WAITALL Ta wait for full request or error
.It Dv MSG_DONTWAIT Ta do not block
.It Dv MSG_CMSG_CLOEXEC Ta set received fds close-on-exec
+.It Dv MSG_WAITFORONE Ta do not block after receiving the first message
+(relevant only for
+.Fn recvmmsg )
.El
.Pp
The
@@ -158,6 +190,11 @@ is set to
This flag is not available in strict
.Tn ANSI
or C99 compilation mode.
+The
+.Dv MSG_WAITFORONE
+flag sets MSG_DONTWAIT after the first message has been received. This flag
+is only relevant for
+.Fn recvmmsg .
.Pp
The
.Fn recvmsg
@@ -290,9 +327,34 @@ control data were discarded due to lack of space in the buffer
for ancillary data.
.Dv MSG_OOB
is returned to indicate that expedited or out-of-band data were received.
+.Pp
+The
+.Fn recvmmsg
+system call uses the
+.Fa mmsghdr
+structure. Its form is as follows, as defined in
+.In sys/socket.h :
+.Bd -literal
+struct mmsghdr {
+ struct msghdr msg_hdr; /* message header */
+ unsigned int msg_len; /* message length */
+};
+.Ed
+.Pp
+For
+.Fa msg_hdr
+see above. On data reception the
+.Fa msg_len
+field is updated to the length of the received message. On
+data transmission it is updated to the number of
+characters sent.
.Sh RETURN VALUES
-These calls return the number of bytes received, or -1
-if an error occurred.
+These calls except
+.Fn recvmmsg
+return the number of bytes received.
+.Fn recvmmsg
+returns the number of messages received.
+A value of -1 is returned if an error occurred.
.Sh ERRORS
The calls fail if:
.Bl -tag -width Er
diff --git a/lib/libc/sys/recvmmsg.c b/lib/libc/sys/recvmmsg.c
new file mode 100644
index 0000000..19a937b
--- /dev/null
+++ b/lib/libc/sys/recvmmsg.c
@@ -0,0 +1,96 @@
+/*
+ * Copyright (c) 2016 Boris Astardzhiev, Smartcom-Bulgaria AD
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice(s), this list of conditions and the following disclaimer as
+ * the first lines of this file unmodified other than the possible
+ * addition of one or more copyright notices.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice(s), this list of conditions and the following disclaimer in
+ * the documentation and/or other materials provided with the
+ * distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) ``AS IS'' AND ANY
+ * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER(S) BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
+ * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
+ * OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
+ * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <errno.h>
+#include <sys/types.h>
+#include <sys/syscall.h>
+#include <sys/socket.h>
+#include <sys/select.h>
+#include <pthread.h>
+#include "libc_private.h"
+
+#define CMTR(s, timeout) \
+ do { \
+ fd_set fds; \
+ int res; \
+ \
+ FD_ZERO(&fds); \
+ FD_SET((s), &fds); \
+ res = __sys_pselect((s)+1, &fds, NULL, NULL, (timeout), NULL);\
+ if (res == -1 || res == 0) \
+ return (res); \
+ if (!FD_ISSET((s), &fds)) \
+ return (-1); \
+ } while (0);
+
+ssize_t
+recvmmsg(int s, struct mmsghdr *msgvec, size_t vlen, int flags,
+ const struct timespec *timeout)
+{
+ size_t i, rcvd;
+ ssize_t ret;
+
+ if (timeout != NULL)
+ CMTR(s, timeout);
+
+ ret = __sys_recvmsg(s, &msgvec[0].msg_hdr, flags);
+ if (ret == -1)
+ return (ret);
+
+ /* Check initially for the presence of MSG_WAITFORONE.
+ * Turn on MSG_DONTWAIT if set. */
+ if (flags & MSG_WAITFORONE) {
+ flags |= MSG_DONTWAIT;
+ /* The kernel doesn't need to know about this flag. */
+ flags &= ~MSG_WAITFORONE;
+ }
+
+ rcvd = 1;
+ for (i = rcvd; i < vlen; i++) {
+ ret = __sys_recvmsg(s, &msgvec[i].msg_hdr, flags);
+ if (ret == -1) {
+ if (rcvd != 0) {
+ /* We've received messages. Let caller know. */
+ return (rcvd);
+ }
+ return (ret);
+ }
+
+ /* Save received bytes */
+ msgvec[i].msg_len = ret;
+ rcvd++;
+ }
+
+ return (rcvd);
+}
+
+#undef CMTR
diff --git a/lib/libc/sys/send.2 b/lib/libc/sys/send.2
index 8fa2c64..33fa58d 100644
--- a/lib/libc/sys/send.2
+++ b/lib/libc/sys/send.2
@@ -34,8 +34,9 @@
.Sh NAME
.Nm send ,
.Nm sendto ,
-.Nm sendmsg
-.Nd send a message from a socket
+.Nm sendmsg ,
+.Nm sendmmsg
+.Nd send message(s) from a socket
.Sh LIBRARY
.Lb libc
.Sh SYNOPSIS
@@ -47,6 +48,8 @@
.Fn sendto "int s" "const void *msg" "size_t len" "int flags" "const struct sockaddr *to" "socklen_t tolen"
.Ft ssize_t
.Fn sendmsg "int s" "const struct msghdr *msg" "int flags"
+.Ft ssize_t
+.Fn sendmmsg "int s" "struct mmsghdr *msgvec" "size_t vlen" "int flags"
.Sh DESCRIPTION
The
.Fn send
@@ -55,8 +58,10 @@ and
.Fn sendto
and
.Fn sendmsg
+and
+.Fn sendmmsg
system calls
-are used to transmit a message to another socket.
+are used to transmit one or multiple messages (with the latter call) to another socket.
The
.Fn send
function
@@ -66,6 +71,8 @@ state, while
.Fn sendto
and
.Fn sendmsg
+and
+.Fn sendmmsg
may be used at any time.
.Pp
The address of the target is given by
@@ -81,6 +88,18 @@ underlying protocol, the error
is returned, and
the message is not transmitted.
.Pp
+The
+.Fn sendmmsg
+function sends multiple messages at a call.
+They are given by the
+.Fa msgvec
+vector along with
+.Fa vlen
+specifying its size. The number of
+characters sent per each message is placed in the
+.Fa msg_len
+field of each element of the vector after transmission.
+.Pp
No indication of failure to deliver is implicit in a
.Fn send .
Locally detected errors are indicated by a return value of -1.
@@ -138,10 +157,16 @@ See
.Xr recv 2
for a description of the
.Fa msghdr
+structure and the
+.Fa mmsghdr
structure.
.Sh RETURN VALUES
-The call returns the number of characters sent, or -1
-if an error occurred.
+All calls except
+.Fn sendmmsg
+return the number of characters sent. The
+.Fn sendmmsg
+call returns the number of messages sent.
+If an error occurred a value of -1 is returned.
.Sh ERRORS
The
.Fn send
@@ -149,6 +174,8 @@ function and
.Fn sendto
and
.Fn sendmsg
+and
+.Fn sendmmsg
system calls
fail if:
.Bl -tag -width Er
diff --git a/lib/libc/sys/sendmmsg.c b/lib/libc/sys/sendmmsg.c
new file mode 100644
index 0000000..cef35a3
--- /dev/null
+++ b/lib/libc/sys/sendmmsg.c
@@ -0,0 +1,63 @@
+/*
+ * Copyright (c) 2016 Boris Astardzhiev, Smartcom-Bulgaria AD
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice(s), this list of conditions and the following disclaimer as
+ * the first lines of this file unmodified other than the possible
+ * addition of one or more copyright notices.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice(s), this list of conditions and the following disclaimer in
+ * the documentation and/or other materials provided with the
+ * distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) ``AS IS'' AND ANY
+ * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER(S) BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
+ * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
+ * OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
+ * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <errno.h>
+#include <sys/types.h>
+#include <sys/syscall.h>
+#include <sys/socket.h>
+#include <pthread.h>
+#include "libc_private.h"
+
+ssize_t
+sendmmsg(int s, struct mmsghdr *msgvec, size_t vlen, int flags)
+{
+ size_t i, sent;
+ ssize_t ret;
+
+ sent = 0;
+ for (i = 0; i < vlen; i++) {
+ ret = __sys_sendmsg(s, &msgvec[i].msg_hdr, flags);
+ if (ret == -1) {
+ if (sent != 0) {
+ /* We have sent messages. Let caller know. */
+ return (sent);
+ }
+ return (ret);
+ }
+
+ /* Save sent bytes */
+ msgvec[i].msg_len = ret;
+ sent++;
+ }
+
+ return (sent);
+}
diff --git a/sys/sys/socket.h b/sys/sys/socket.h
index 18e2de1..d95f29e 100644
--- a/sys/sys/socket.h
+++ b/sys/sys/socket.h
@@ -435,6 +435,11 @@ struct msghdr {
#ifdef _KERNEL
#define MSG_SOCALLBCK 0x10000 /* for use by socket callbacks - soreceive (TCP) */
#endif
+#ifndef _KERNEL
+#ifdef __BSD_VISIBLE
+#define MSG_WAITFORONE 0x100000 /* used in recvmmsg() */
+#endif /* __BSD_VISIBLE */
+#endif /* !_KERNEL */
/*
* Header for ancillary data objects in msg_control buffer.
@@ -595,6 +600,18 @@ struct sf_hdtr {
#endif /* _KERNEL */
#endif /* __BSD_VISIBLE */
+#ifndef _KERNEL
+#ifdef __BSD_VISIBLE
+/*
+ * Send/recvmmsg specific structure(s)
+ */
+struct mmsghdr {
+ struct msghdr msg_hdr; /* message header */
+ unsigned int msg_len; /* message length */
+};
+#endif /* __BSD_VISIBLE */
+#endif /* !_KERNEL */
+
#ifndef _KERNEL
#include <sys/cdefs.h>
@@ -615,11 +632,19 @@ int listen(int, int);
ssize_t recv(int, void *, size_t, int);
ssize_t recvfrom(int, void *, size_t, int, struct sockaddr * __restrict, socklen_t * __restrict);
ssize_t recvmsg(int, struct msghdr *, int);
+#if __BSD_VISIBLE
+struct timespec;
+ssize_t recvmmsg(int, struct mmsghdr *, size_t, int,
+ const struct timespec *);
+#endif
ssize_t send(int, const void *, size_t, int);
ssize_t sendto(int, const void *,
size_t, int, const struct sockaddr *, socklen_t);
ssize_t sendmsg(int, const struct msghdr *, int);
#if __BSD_VISIBLE
+ssize_t sendmmsg(int, struct mmsghdr *, size_t, int);
+#endif
+#if __BSD_VISIBLE
int sendfile(int, int, off_t, size_t, struct sf_hdtr *, off_t *, int);
int setfib(int);
#endif
More information about the freebsd-net
mailing list