a3f7947ad2
server_socket: use resolve_host_port() instead of getaddrinfo()
...
Allow port specification in "bind_to_address" settings.
2011-09-20 21:15:22 +02:00
7d9d459ac2
resolver: add function resolve_host_port()
2011-09-20 21:15:05 +02:00
3ea1073809
socket_util: move sockaddr_to_string() to resolver.c
2011-09-20 08:49:36 +02:00
76a959a578
rtsp_client, output/raop: use close_socket() instead of close()
2011-09-20 08:48:48 +02:00
f43873dc5f
tcp_socket, udp_server: use g_io_channel_new_socket()
2011-09-20 08:47:42 +02:00
921553d7bb
Merge branch 'v0.16.x'
2011-09-20 08:47:06 +02:00
c476819cb1
fd_util: add function close_socket()
...
Wrap close(), use closesocket() on WIN32/WinSock.
2011-09-20 08:38:58 +02:00
77a56c7c5a
fd_util: check HAVE_INOTIFY_INIT in header
...
Don't provide the prototype if inotify_init() was not detected.
2011-09-20 08:38:17 +02:00
46deb7ca82
fd_util: include check.h, verify config.h was included
2011-09-20 08:38:08 +02:00
b03f9ece05
glib_socket.h: wrap g_io_channel_*_new() calls portably
...
The server_socket library (used by the httpd output plugin) didn't
check for WIN32, that's fixed now.
2011-09-20 08:35:25 +02:00
27946a981f
Set socket TCP keepalive option on incoming connections
...
If a connected host disappears without our knowledge, as can happen over
wireless or a hibernating machine, we continue to hold the port open waiting
for messages. Because we never try to send anything down this now-broken
pipe, the connection will sit idle taking up a slot in our allowed incoming
connections list.
If enough of these happen, an unintended Denial of Service takes place,
where all connection slots are filled with now-broken, never ending
connections. Setting the TCP keepalive option at least allows these to time
out after the default two hours, which is sufficient in the non-malicious
case.
Signed-off-by: Dan McGee <dan@archlinux.org>
2011-09-19 17:11:09 +02:00
74617389c8
output_plugin: the plugin allocates the audio_output object
...
Pass audio_output objects around instead of void pointers. This will
give some more control to the plugin, and prepares for non-blocking
audio outputs.
2011-09-19 09:41:21 +02:00
1d8840412f
configure.ac: add option --enable-solaris-output
...
Allow enabling the plugin explicitly without running Solaris, to test
the build.
2011-09-19 09:39:35 +02:00
ffa9f4b47f
output/roar: check return values
...
Fixes gcc warnings.
2011-09-18 12:07:06 +02:00
667e22bbc7
output/roar: unknown volume is -1, not 0
2011-09-18 12:06:46 +02:00
d4d838c8bb
output/roar: move code to roar_cancel_locked() and others
...
Wrap the locking code.
2011-09-18 11:51:18 +02:00
c61a889c86
output/roar: use GLib memory allocation functions
...
GLib has a built-in OOM handler.
2011-09-17 20:12:45 +02:00
567fe89a77
output/roar: use config_get_block_string() for temporary
2011-09-17 20:11:51 +02:00
950772ab8a
output/roar: reorder methods
2011-09-17 20:10:26 +02:00
0bd0e2a3ec
output: rename plugin variables
...
Consistent naming.
2011-09-17 20:04:18 +02:00
47774ce882
Merge branch 'v0.16.x'
2011-09-17 19:59:52 +02:00
3c9bcdd347
output: per-plugin header
...
Move the "extern" declarations from output_list.c, for more type
safety.
2011-09-17 19:51:41 +02:00
3934d2d30c
output/pulse: don't expose internal struct in public header
...
Provide _lock() and _unlock() to wrap all accesses from the mixer
plugin.
2011-09-17 19:50:36 +02:00
423ce9557a
output/raop: don't expose internal structs in public header
2011-09-17 19:44:30 +02:00
947848ebf6
output/roar: export volume methods
...
Use these instead of exposing the internal roar_t struct.
2011-09-17 19:33:51 +02:00
5e22fe488e
output: rename plugin source files
2011-09-17 08:54:50 +02:00
c666cf1c44
output/pulse: move code to _setup_stream()
2011-09-17 08:53:53 +02:00
79435dbdec
decoder/audiofile: include stdio header for SEEK_* defines
...
The stdio header is no longer pulled in by af_vfs.h in audiofile-0.3.0.
2011-09-17 07:33:42 +02:00
27206368da
output/pulse: improve locking
...
Always lock the main loop when operating on PULSE objects. Document
this.
2011-09-17 07:30:59 +02:00
2be245b75e
input/soup: wait some more before resuming the stream
2011-09-16 21:50:05 +02:00
17927f5c26
input/soup: improve error messages
...
Copy the reason_phrase from the SoupMessage, and check for transport
errors in the session callback.
2011-09-16 21:50:05 +02:00
754f26a97c
input_stream: non-blocking I/O
...
Add GMutex, GCond attributes which will be used by callers to
conditionally wait on the stream.
Remove the (now-useless) plugin method buffer(), wait on GCond
instead. Lock the input_stream before each method call. Do the same
with the playlist plugins.
2011-09-16 21:22:13 +02:00
29241c4f83
input_plugin: add method check()
...
To check for errors without reading. The decoder thread wants to do
that, before it passes the input stream to the plugin.
2011-09-16 21:22:02 +02:00
6f655eb9b9
input/{soup,curl}: free unused postponed_error
...
Fix memory leak.
2011-09-16 21:13:25 +02:00
65dfd90141
input/soup: move libsoup calls to the I/O thread
...
libsoup's asynchronous API is not thread safe. By moving the calls
into the I/O thread, several crash bugs will be fixed.
2011-09-16 20:34:47 +02:00
59abdbd2dd
input/soup: wait for _cancel_message() to complete
...
Add the "complete" attribute, and set it in _session_callback().
_close() waits for it to become true.
2011-09-16 20:34:47 +02:00
a6d6873856
input/soup: return real GErrors to the caller
...
Add attribute postponed_error, pass this GError to the caller.
2011-09-16 11:36:09 +02:00
e99df3a3be
input/soup: set the "ready" flag on failure
...
Fix deadlock.
2011-09-16 11:35:02 +02:00
8b657255cc
input/soup: remove redundant flag "ready"
...
We'll add proper locking now, and there's no need to carry this
duplicate flag.
2011-09-16 11:34:50 +02:00
0e1dc79321
input/soup: report failure in _read()
...
To report failures to callers who don't use _buffer().
2011-09-16 11:30:56 +02:00
0ef15e5fac
input/soup: unlock before cancelling the soup message
...
Avoid recursive deadlock in _session_callback().
2011-09-16 11:30:30 +02:00
9ea8a840a9
decoder_api: remove the loop from decoder_read()
...
This was useless, because input_stream_read() blocks until data is
available. Calling it in a loop doesn't make sense.
2011-09-16 09:15:10 +02:00
7b88089593
decoder_api: move code to decoder_check_cancel_read()
2011-09-16 09:15:10 +02:00
b513943893
input/curl: wait some more before resuming the stream
...
Pausing and resuming after every little chunk adds lots of overhead.
Add a lower level for resuming the stream.
2011-09-16 09:14:29 +02:00
edffc56600
input/curl: clear the "paused" flag before resuming
...
curl_easy_pause(CURLPAUSE_CONT) may have added enough data to go into
pause again.
2011-09-16 09:11:15 +02:00
856b0e6886
input/curl: remove _schedule_update()
...
Call _update_fds() directly. This is possible because it's only
called from within the I/O thread.
2011-09-16 09:11:09 +02:00
bf1eb46b87
input/curl: unlock the mutex for io_thread_call()
...
Fix deadlock.
2011-09-16 09:10:39 +02:00
53174ea45f
input/soup: fix "unused local variable" warnings
2011-09-16 09:10:33 +02:00
87b624f5d5
Merge branch 'v0.16.x'
2011-09-16 07:52:39 +02:00
1cbba4fc59
input/curl, output/pulse: fix "unused local variable" warnings
2011-09-16 07:41:41 +02:00
Max Kellermann