Commit Graph

171 Commits

Author SHA1 Message Date
Dan Winship
e0bb25c214 Remove an erroneous check in the non-sendmmsg() version of g_socket_send_messages()
The docs specify that *all* errors are ignored if we managed to send
any data successfully, not just timeout/wouldblock.

https://bugzilla.gnome.org/show_bug.cgi?id=768549
2016-07-12 09:14:09 -04:00
Christian Hergert
c16a0b53fe socket: set fd field to -1 after closing socket
This ensures that g_socket_get_fd() will return -1 after the socket has
been closed.

https://bugzilla.gnome.org/show_bug.cgi?id=765959
2016-05-04 16:47:54 +03:00
Rico Tzschichholz
d268d9f86a socket: Fix annotation of g_socket_receive_message
https://bugzilla.gnome.org/show_bug.cgi?id=761337
2016-02-02 10:15:04 +01:00
Philip Withnall
4e631d2e5f gio: Add GDatagramBased interface and rebase GSocket on it
GDatagramBased is an interface abstracting datagram-based communications
in the style of the Berkeley sockets API. It may be contrasted to (for
example) GIOStream, which supports only streaming I/O.

GDatagramBased allows socket-like communications to be done through any
object, not just a concrete GSocket (which wraps socket()).

This adds the GDatagramBased interface, and implements it in GSocket.

https://bugzilla.gnome.org/show_bug.cgi?id=697907
2015-10-13 15:33:48 +01:00
Philip Withnall
212b0c28cc gsocket: Fix g_socket_send_messages_with_timeout() on win32
Commit a0cefc2217 introduced an unresolved
symbol, g_socket_send_message_with_timeout(), on win32. Windows
unfortunately isn’t clever enough to fill in the gaps and magic up the
implementation of that function from nowhere, so we had better do it
ourselves.

Factor the blocking behaviour out of g_socket_send_message() into a new
internal g_socket_send_message_with_timeout().

https://bugzilla.gnome.org/show_bug.cgi?id=756054
2015-10-05 16:15:19 +01:00
Chun-wei Fan
e81d4ea988 gio/goscket.c: Fix build on Windows
5d68947 factored out resuable items, but some of these are only for
*NIX builds, which will break the build on Windows.  Fix this by
building these portions only when !G_OS_WIN32.

https://bugzilla.gnome.org/show_bug.cgi?id=756053
2015-10-05 17:15:12 +08:00
Philip Withnall
fc59c20e97 gsocket: Minor documentation clarifications
As suggested by Dan Winship on bug #697907.
2015-10-04 11:34:36 +01:00
Philip Withnall
237fec7e70 gsocket: Fix connected state if shutting down in two steps
The value of g_socket_is_connected() gets stuck high if the GSocket is
shut down in two steps:
   g_socket_shutdown (socket, TRUE, FALSE, NULL);
   g_socket_shutdown (socket, FALSE, TRUE, NULL);
rather than one:
   g_socket_shutdown (socket, TRUE, TRUE, NULL);

Fix that by tracking the connected status for the read half and the
write half of the connection separately.

https://bugzilla.gnome.org/show_bug.cgi?id=697907
2015-10-04 10:39:13 +01:00
Philip Withnall
1086507e75 gsocket: Fix error behaviour of g_socket_send_messages()
If an error in the underlying sendmmsg() syscall occurs after
successfully sending one or more messages, g_socket_send_messages()
should return the number of messages successfully sent, rather than an
error. This mirrors the documented sendmmsg() behaviour.

This is a slight behaviour change for g_socket_send_messages(), but as
it relaxes the error reporting (reporting errors in fewer situations
than before), it should not cause problems.

https://bugzilla.gnome.org/show_bug.cgi?id=751924
2015-10-01 14:10:10 +01:00
Philip Withnall
f62cbfc022 gsocket: Add g_socket_receive_messages()
Add support for receiving multiple messages with a single system call,
using recvmmsg() if available. Otherwise, fall back to looping over
g_socket_receive_message().

This adds new API, g_socket_receive_messages(), and corresponding unit
tests.

https://bugzilla.gnome.org/show_bug.cgi?id=751924
2015-10-01 14:10:10 +01:00
Philip Withnall
a0cefc2217 gsocket: Switch internal functions from blocking booleans to timeouts
In order to support per-operation timeouts on new API like
g_socket_receive_messages(), the internal GSocket API should use
timeouts rather than boolean blocking parameters.

   (timeout == 0) === (blocking == FALSE)
   (timeout == -1) === (blocking == TRUE)
   (timeout > 0) === new behaviour

https://bugzilla.gnome.org/show_bug.cgi?id=751924
2015-10-01 13:59:08 +01:00
Philip Withnall
7f985b35ce gsocket: Factor out blocking parameter from g_socket_receive_message()
This will make future API additions easier. The factored version is
internal for the time being.

https://bugzilla.gnome.org/show_bug.cgi?id=751924
2015-10-01 13:58:44 +01:00
Philip Withnall
5d68947466 gsocket: Split out functions to convert to and from struct msghdr
As new methods are added to GSocket, we don’t want to duplicate this
code, so factor it out.

https://bugzilla.gnome.org/show_bug.cgi?id=751924
2015-10-01 13:58:27 +01:00
Philip Withnall
363fa18223 gsocket: Fix documentation for g_socket_send_message()
It is no longer the most fully featured version of this function —
g_socket_send_messages() stole that dubious honour with 2.44.

https://bugzilla.gnome.org/show_bug.cgi?id=751924
2015-09-28 12:47:09 +01:00
Philip Withnall
347e4a75ec gsocket: Clarify GSocket:blocking doesn’t apply to ops with a parameter
Operations which take an explicit blocking parameter are completely
unaffected by GSocket:blocking.

https://bugzilla.gnome.org/show_bug.cgi?id=751924
2015-09-28 12:47:09 +01:00
Philip Withnall
8fdc670188 gsocket: Clarify flags documentation for g_socket_receive_message()
The API design here is a bit awkward — the in/out flags argument should
actually have been an in flags argument and an out msg_flags argument.
Clarify that a bit in the documentation.

https://bugzilla.gnome.org/show_bug.cgi?id=751924
2015-09-28 12:47:07 +01:00
Dan Winship
1ab3e3ed3e gsocket: add a wrapper around g_set_error() to avoid extra work
If @error is NULL then we don't even need to evaluate the remaining
arguments. And if errno is EWOULDBLOCK, then no one should see the
error message anyway, so don't bother g_strdup_printf'ing up a pretty
one.

https://bugzilla.gnome.org/show_bug.cgi?id=752769
2015-08-29 08:46:25 -04:00
Philip Withnall
8520ae3ffa gsocket: Factor out blocking parameter from g_socket_send_messages()
This will make future API additions easier. The factored version is
internal for the time being.

https://bugzilla.gnome.org/show_bug.cgi?id=751924
2015-07-23 11:37:18 +01:00
Stef Walter
f405f42115 gsocket: Don't g_error() if file-descriptor is not a socket
This code was out of date with current coding practices.

Nowadays it's common to receive file descriptors over environment
variables from other processes like systemd. The unit files that
control these file descriptors are configurable by sysadmins.

It is not (necessarily) a programmer error when g_socket_details_from_fd()
is called with a file descriptor that is not a socket. It can also
be a system and/or configuration error.

https://bugzilla.gnome.org/show_bug.cgi?id=746339
2015-07-02 12:24:11 +02:00
Tim-Philipp Müller
fd789f1187 gsocket: avoid unnecessary select in _send_messages() and _receive_message()
For performance reasons we should always try to send or
receive our messages first and only wait for more space
or data to become available if we get an EAGAIN (and
are in blocking mode).

https://bugzilla.gnome.org/show_bug.cgi?id=751122
2015-06-21 10:28:14 +01:00
Philip Withnall
1b22df7822 gsocket: Document FD ownership with g_socket_new_from_fd()
https://bugzilla.gnome.org/show_bug.cgi?id=730188
2015-03-21 13:37:17 -04:00
Chun-wei Fan
6fe28eef3c Windows: Use Standard Networking Functions If Possible
Currently, the Windows code use Winsock2-specific APIs to try to emulate
calls such as inet_pton(), inet_ntop() and if_nametoindex(), which may not
do the job all the time.  On Vista and later, Winsock2 does provide a
proper implementation for these functions, so we can use them if they exist
on the system, by querying for them during g_networking_init().  Otherwise,
we continue to use the original code path for these, in the case of XP and
Server 2003.

This enables many of the network-address tests to pass on Windows as a
result, when the native Winsock2 implementations can be used.

https://bugzilla.gnome.org/show_bug.cgi?id=730352
2015-03-06 23:40:03 +08:00
Ignacio Casal Quinteiro
cf03e82478 gsocket: always try before waiting for condition
When implementing blocking operations on top of
nonblocking sockets we should always first try to
perform the operation and then if needed handle
EAGAIN and wait with g_socket_wait_condition.
This is an optimization since we avoid calling
wait condition when it is not needed, but most
importantly this fixes hangs on win32 where some
events (in particular FD_WRITE) are only emitted
after the operation fails with EWOULDBLOCK.

https://bugzilla.gnome.org/show_bug.cgi?id=732439
https://bugzilla.gnome.org/show_bug.cgi?id=741707
2015-01-17 15:04:25 +01:00
Tim-Philipp Müller
fff5c7cd63 gsocket: add g_socket_send_messages()
Allows sending of multiple messages (packets, datagrams)
in one go using sendmmsg(), thus drastically reducing the
number of syscalls when sending out a lot of data, or when
sending out the same data to multiple recipients.

https://bugzilla.gnome.org/show_bug.cgi?id=719646
2014-12-11 15:10:44 +00:00
Dan Winship
634b692199 gsocket: Set SO_NOSIGPIPE on sockets on Darwin
This is a best-effort approach to preventing SIGPIPE emissions on Darwin
and iOS, where they continue to be intercepted by the Xcode debugger
even if SIG_IGN prevents them crashing the program.

This is similar to the existing code which sets MSG_NOSIGNAL on all
send() calls. MSG_NOSIGNAL doesn't exist on Darwin though.

Based on a patch from Philip Withnall.

https://bugzilla.gnome.org/show_bug.cgi?id=728730
2014-07-21 11:31:27 -04:00
Dan Winship
7b34e07ac4 gsocket: move set_fd_nonblocking() into g_socket_constructed() 2014-07-21 11:31:27 -04:00
Patrick Welche
afce39c228 gcredentials: add NetBSD support
https://bugzilla.gnome.org/show_bug.cgi?id=728256
2014-06-28 14:06:36 -04:00
Philip Withnall
00e733315d gsocket: Document that g_socket_create_source() holds a socket ref
This is quite important, as it means you can safely let the GSocket drop
out of scope while maintaining a reference to the GSource, and the
socket will remain open. That means fewer closure structures, simpler
code, and fewer allocations.

https://bugzilla.gnome.org/show_bug.cgi?id=732081
2014-06-23 16:24:47 +01:00
Philip Withnall
7f5cc2c3c5 gsocket: Add missing preconditions to g_socket_send_message()
https://bugzilla.gnome.org/show_bug.cgi?id=730190
2014-05-15 14:35:19 +01:00
Philip Withnall
2b247e1589 gio: Document that GSocket is not thread safe
https://bugzilla.gnome.org/show_bug.cgi?id=726318
2014-04-16 19:04:25 +01:00
Dan Winship
3da5d59078 gio: move Winsock error mapping to g_io_error_from_win32_error()
Rather than having special code in gsocket.c, handle Winsock errors
along with other Win32 errors in gioerror.c

Also, reference g_win32_error_message() from the
g_io_error_from_win32_error() docs, and update the
g_win32_error_message() docs to clarify that it works with Winsock
error codes too.
2014-03-30 11:55:47 -04:00
Olivier Crête
0c65f7e45a GSocket: ignore timed out state when not relevant
Do this by separating the timeout check from the other socket checks.

https://bugzilla.gnome.org/show_bug.cgi?id=726375
2014-03-23 16:17:09 -04:00
Ryan Lortie
8da795de4f gsocket: use check/prepare only on win32
There is no longer any code left in the check/prepare functions on UNIX,
so put %NULL in the GSourceFuncs vtable.

This also allows us to simplify some logic.

https://bugzilla.gnome.org/show_bug.cgi?id=724707
2014-02-22 10:25:42 -05:00
Ryan Lortie
e8f26efe0d gsocket: use _add_unix_fd() instead of _add_poll()
Use g_source_add_unix_fd() on UNIX instead of using a GPollFD.

https://bugzilla.gnome.org/show_bug.cgi?id=724707
2014-02-22 10:25:42 -05:00
Ryan Lortie
04aee2d920 gsocket: make use of g_source_set_ready_time()
Drop our own hand-rolled version of the same functionality.

https://bugzilla.gnome.org/show_bug.cgi?id=724707
2014-02-22 10:25:42 -05:00
Ryan Lortie
1f7100506a gsocket: don't abuse GPollFD.revents field
We are reusing the GPollFD.revents field of the source to store a
temporary value.  Use a local variable for that instead.

This is a refactor to make the next commit easier to understand.

https://bugzilla.gnome.org/show_bug.cgi?id=724707
2014-02-22 10:24:19 -05:00
Ryan Lortie
ff96f88e17 gsocket: use cancellable child source
Now that GCancellable's GSource is based on _set_ready_time() instead of
an fd, we should use it as a child source, instead of forcing the
creation of the fd and adding it as a poll.

https://bugzilla.gnome.org/show_bug.cgi?id=724707
2014-02-22 10:24:19 -05:00
Ryan Lortie
d8263dd793 gsocket: trivial typo fix
https://bugzilla.gnome.org/show_bug.cgi?id=724707
2014-02-22 10:24:19 -05:00
Dan Winship
4139b26f3e gsocket: fix g_socket_condition_timed_wait() recovery after EINTR
After getting an EINTR, g_socket_condition_timed_wait() has to adjust
its timeout, but it was trying to convert from nanoseconds to
microseconds by multiplying by 1000 rather than dividing... Oops.

https://bugzilla.gnome.org/show_bug.cgi?id=724239
2014-02-17 11:38:11 -05:00
Dan Winship
074df39681 Fix g_socket_get_available() with TCP on Windows
Windows needs a special inefficient hack to implement
g_socket_get_available() correctly for UDP sockets, but that hack
isn't needed for TCP, and in fact, might give the wrong answer in that
case. Fix it to only use the hack with UDP.

Also, fix that case to handle non-blocking sockets as well.

And add a test case for g_socket_get_available() with TCP.

https://bugzilla.gnome.org/show_bug.cgi?id=723422
2014-02-15 10:06:04 -05:00
Matthias Clasen
35066ed6c6 Docs: Drop entities, switch away from sgml mode
Since all element markup is now gone from the doc comments,
we can turn off the gtk-doc sgml mode, which means that from
now on, docbook markup is no longer allowed in doc comments.

To make this possible, we have to replace all remaining
entities in doc comments by their replacement text, & -> &
and so on.
2014-02-09 02:07:26 -05:00
Matthias Clasen
e7fd3de86d Eradicate links and xrefs
These are all replaced by markdown ref links.
2014-02-08 12:26:56 -05:00
Matthias Clasen
3232425785 Docs: replace <literal> by ` 2014-02-06 08:07:16 -05:00
Matthias Clasen
42cf80780b Docs: Big entity cleanup
Strip lots of entity use from |[ ]| examples (which are now
implicit CDATA). Also remove many redundant uses of <!-- -->.
2014-02-01 12:00:30 -05:00
Daniel Mustieles
078dbda148 Updated FSF's address 2014-01-31 14:31:55 +01:00
Dan Winship
3981cddbf8 Require POSIX.1 (1990) compliance on unix
Assume unix platforms support the original POSIX.1 standard.
Specifically, assume that if G_OS_UNIX, then we have chown(),
getcwd(), getgrgid(), getpwuid(), link(), <grp.h>, <pwd.h>,
<sys/types.h>, <sys/uio.h>, <sys/wait.h>, and <unistd.h>.

Additionally, since all versions of Windows that we care about also
have <sys/types.h>, we can remove HAVE_SYS_TYPES_H checks everywhere.

Also remove one include of <sys/times.h>, and the corresponding
configure check, since the include is not currently needed (and may
always have just been a typo for <sys/time.h>).

https://bugzilla.gnome.org/show_bug.cgi?id=710519
2013-11-20 09:17:42 -05:00
Dan Winship
5a269e5a90 gcredentials: add Solaris support
Based on patches from Igor Pashev.

https://bugzilla.gnome.org/show_bug.cgi?id=705029
2013-10-04 09:51:31 -04:00
Dan Winship
cedd697bf8 gcredentials: add internal macros to simplify all the #ifdefs
Rather than having lots of obscure platform-based #ifdefs all over
gio, define some macros in gcredentialsprivate.h, and use those to
simplify the rest of the code.

https://bugzilla.gnome.org/show_bug.cgi?id=701482
2013-10-04 09:51:31 -04:00
Svante Signell
1d4bb3f5d0 gio/gsocket.c: Fix error code checks when SOCK_CLOEXEC is defined but
not supported on GNU/Hurd.

https://bugzilla.gnome.org/show_bug.cgi?id=708266
2013-09-24 08:59:38 -04:00
Dan Winship
9fd0927e58 gsocket: Fix g_socket_get_available_bytes() on Windows and OS X
On Windows and OS X, FIONREAD on a UDP socket gets the total number of
bytes available, not the number of bytes available in the next packet,
which is the more useful number (and how the function always behaved
on Linux).

On OS X, fix this by using SO_NREAD. On Windows, fix this by doing a
MSG_PEEK recv() into a giant buffer, since there is apparently no
other way to get the information.

https://bugzilla.gnome.org/show_bug.cgi?id=686786
2013-08-31 11:40:44 -04:00