summaryrefslogtreecommitdiff
path: root/docs/libcurl/libcurl-multi.3
blob: aa408242e4ccdab0ae1bf0939edd8a7ab437a2c0 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
.\" **************************************************************************
.\" *                                  _   _ ____  _
.\" *  Project                     ___| | | |  _ \| |
.\" *                             / __| | | | |_) | |
.\" *                            | (__| |_| |  _ <| |___
.\" *                             \___|\___/|_| \_\_____|
.\" *
.\" * Copyright (C) 1998 - 2021, Daniel Stenberg, <daniel@haxx.se>, et al.
.\" *
.\" * This software is licensed as described in the file COPYING, which
.\" * you should have received as part of this distribution. The terms
.\" * are also available at https://curl.se/docs/copyright.html.
.\" *
.\" * You may opt to use, copy, modify, merge, publish, distribute and/or sell
.\" * copies of the Software, and permit persons to whom the Software is
.\" * furnished to do so, under the terms of the COPYING file.
.\" *
.\" * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
.\" * KIND, either express or implied.
.\" *
.\" **************************************************************************
.\"
.TH libcurl-multi 3 "19 Sep 2014" "libcurl" "libcurl multi interface"
.SH NAME
libcurl-multi \- how to use the multi interface
.SH DESCRIPTION
This is an overview on how to use the libcurl multi interface in your C
programs. There are specific man pages for each function mentioned in
here. There's also the \fIlibcurl-tutorial(3)\fP man page for a complete
tutorial to programming with libcurl and the \fIlibcurl-easy(3)\fP man page
for an overview of the libcurl easy interface.

All functions in the multi interface are prefixed with curl_multi.
.SH "OBJECTIVES"
The multi interface offers several abilities that the easy interface doesn't.
They are mainly:

1. Enable a "pull" interface. The application that uses libcurl decides where
and when to ask libcurl to get/send data.

2. Enable multiple simultaneous transfers in the same thread without making it
complicated for the application.

3. Enable the application to wait for action on its own file descriptors and
curl's file descriptors simultaneously.

4. Enable event-based handling and scaling transfers up to and beyond
thousands of parallel connections.
.SH "ONE MULTI HANDLE MANY EASY HANDLES"
To use the multi interface, you must first create a 'multi handle' with
\fIcurl_multi_init(3)\fP. This handle is then used as input to all further
curl_multi_* functions.

With a multi handle and the multi interface you can do several simultaneous
transfers in parallel. Each single transfer is built up around an easy
handle. You create all the easy handles you need, and setup the appropriate
options for each easy handle using \fIcurl_easy_setopt(3)\fP.

There are two flavours of the multi interface, the select() oriented one and
the event based one we call multi_socket. You will benefit from reading
through the description of both versions to fully understand how they work and
differentiate. We start out with the select() oriented version.

When an easy handle is setup and ready for transfer, then instead of using
\fIcurl_easy_perform(3)\fP like when using the easy interface for transfers,
you should add the easy handle to the multi handle with
\fIcurl_multi_add_handle(3)\fP. You can add more easy handles to a multi
handle at any point, even if other transfers are already running.

Should you change your mind, the easy handle is again removed from the multi
stack using \fIcurl_multi_remove_handle(3)\fP. Once removed from the multi
handle, you can again use other easy interface functions like
\fIcurl_easy_perform(3)\fP on the handle or whatever you think is
necessary. You can remove handles at any point in time during transfers.

Adding the easy handle to the multi handle does not start the transfer.
Remember that one of the main ideas with this interface is to let your
application drive. You drive the transfers by invoking
\fIcurl_multi_perform(3)\fP. libcurl will then transfer data if there is
anything available to transfer. It'll use the callbacks and everything else
you have setup in the individual easy handles. It'll transfer data on all
current transfers in the multi stack that are ready to transfer anything. It
may be all, it may be none. When there's nothing more to do for now, it
returns back to the calling application.

Your application extracts info from libcurl about when it would like to get
invoked to transfer data or do other work. The most convenient way is to use
\fIcurl_multi_poll(3)\fP that will help you wait until the application should
call libcurl again. The older API to accomplish the same thing is
\fIcurl_multi_fdset(3)\fP that extracts fd_sets from libcurl to use in
select() or poll() calls in order to get to know when the transfers in the
multi stack might need attention. Both these APIs allow for your program to
wait for input on your own private file descriptors at the same time.
\fIcurl_multi_timeout(3)\fP also helps you with providing a suitable timeout
period for your select() calls.

\fIcurl_multi_perform(3)\fP stores the number of still running transfers in
one of its input arguments, and by reading that you can figure out when all
the transfers in the multi handles are done. 'done' does not mean
successful. One or more of the transfers may have failed.

To get information about completed transfers, to figure out success or not and
similar, \fIcurl_multi_info_read(3)\fP should be called. It can return a
message about a current or previous transfer. Repeated invokes of the function
get more messages until the message queue is empty. The information you
receive there includes an easy handle pointer which you may use to identify
which easy handle the information regards.

When a single transfer is completed, the easy handle is still left added to
the multi stack. You need to first remove the easy handle with
\fIcurl_multi_remove_handle(3)\fP and then close it with
\fIcurl_easy_cleanup(3)\fP, or possibly set new options to it and add it again
with \fIcurl_multi_add_handle(3)\fP to start another transfer.

When all transfers in the multi stack are done, close the multi handle with
\fIcurl_multi_cleanup(3)\fP. Be careful and please note that you \fBMUST\fP
invoke separate \fIcurl_easy_cleanup(3)\fP calls for every single easy handle
to clean them up properly.

If you want to re-use an easy handle that was added to the multi handle for
transfer, you must first remove it from the multi stack and then re-add it
again (possibly after having altered some options at your own choice).
.SH "MULTI_SOCKET"
\fIcurl_multi_socket_action(3)\fP function offers a way for applications to
not only avoid being forced to use select(), but it also offers a much more
high-performance API that will make a significant difference for applications
using large numbers of simultaneous connections.

\fIcurl_multi_socket_action(3)\fP is then used instead of
\fIcurl_multi_perform(3)\fP.

When using this API, you add easy handles to the multi handle just as with the
normal multi interface. Then you also set two callbacks with the
\fICURLMOPT_SOCKETFUNCTION(3)\fP and \fICURLMOPT_TIMERFUNCTION(3)\fP options
to \fIcurl_multi_setopt(3)\fP. They are two callback functions that libcurl
will call with information about what sockets to wait for, and for what
activity, and what the current timeout time is - if that expires libcurl
should be notified.

The multi_socket API is designed to inform your application about which
sockets libcurl is currently using and for what activities (read and/or write)
on those sockets your application is expected to wait for.

Your application must make sure to receive all sockets informed about in the
\fICURLMOPT_SOCKETFUNCTION(3)\fP callback and make sure it reacts on the given
activity on them. When a socket has the given activity, you call
\fIcurl_multi_socket_action(3)\fP specifying which socket and action there
are.

The \fICURLMOPT_TIMERFUNCTION(3)\fP callback is called to set a timeout. When
that timeout expires, your application should call the
\fIcurl_multi_socket_action(3)\fP function saying it was due to a timeout.

This API is typically used with an event-driven underlying functionality (like
libevent, libev, kqueue, epoll or similar) with which the application
"subscribes" on socket changes. This allows applications and libcurl to much
better scale upward and beyond thousands of simultaneous transfers without
losing performance.

When you've added your initial set of handles, you call
\fIcurl_multi_socket_action(3)\fP with CURL_SOCKET_TIMEOUT set in the sockfd
argument, and you'll get callbacks call that sets you up and you then continue
to call \fIcurl_multi_socket_action(3)\fP accordingly when you get activity on
the sockets you've been asked to wait on, or if the timeout timer expires.

You can poll \fIcurl_multi_info_read(3)\fP to see if any transfer has
completed, as it then has a message saying so.
.SH "BLOCKING"
A few areas in the code are still using blocking code, even when used from the
multi interface. While we certainly want and intend for these to get fixed in
the future, you should be aware of the following current restrictions:

.nf
 - Name resolves unless the c-ares or threaded-resolver backends are used
 - SOCKS proxy handshakes
 - file:// transfers
 - TELNET transfers
.fi
.SH "SEE ALSO"
.BR libcurl-errors "(3), " libcurl-easy "(3), " libcurl "(3) "