It is way more complicated than it should be; and
locking it for thread-safety is too difficult.
[merged r7183 from branches/ew]
git-svn-id: https://svn.musicpd.org/mpd/trunk@7241 09075e82-0dd4-0310-85a5-a0d7c8717e4f
I initially started to do a heavy rewrite that changed the way processes
communicated, but that was too much to do at once. So this change only
focuses on replacing the player and decode processes with threads and
using condition variables instead of polling in loops; so the changeset
itself is quiet small.
* The shared output buffer variables will still need locking
to guard against race conditions. So in this effect, we're probably
just as buggy as before. The reduced context-switching overhead of
using threads instead of processes may even make bugs show up more or
less often...
* Basic functionality appears to be working for playing local (and NFS)
audio, including:
play, pause, stop, seek, previous, next, and main playlist editing
* I haven't tested HTTP streams yet, they should work.
* I've only tested ALSA and Icecast. ALSA works fine, Icecast
metadata seems to get screwy at times and breaks song
advancement in the playlist at times.
* state file loading works, too (after some last-minute hacks with
non-blocking wakeup functions)
* The non-blocking (*_nb) variants of the task management functions are
probably overused. They're more lenient and easier to use because
much of our code is still based on our previous polling-based system.
* It currently segfaults on exit. I haven't paid much attention
to the exit/signal-handling routines other than ensuring it
compiles. At least the state file seems to work. We don't
do any cleanups of the threads on exit, yet.
* Update is still done in a child process and not in a thread.
To do this in a thread, we'll need to ensure it does proper
locking and communication with the main thread; but should
require less memory in the end because we'll be updating
the database "in-place" rather than updating a copy and
then bulk-loading when done.
* We're more sensitive to bugs in 3rd party libraries now.
My plan is to eventually use a master process which forks()
and restarts the child when it dies:
locking and communication with the main thread; but should
require less memory in the end because we'll be updating
the database "in-place" rather than updating a copy and
then bulk-loading when done.
* We're more sensitive to bugs in 3rd party libraries now.
My plan is to eventually use a master process which forks()
and restarts the child when it dies:
master - just does waitpid() + fork() in a loop
\- main thread
\- decoder thread
\- player thread
At the beginning of every song, the main thread will set
a dirty flag and update the state file. This way, if we
encounter a song that triggers a segfault killing the
main thread, the master will start the replacement main
on the next song.
* The main thread still wakes up every second on select()
to check for signals; which affects power management.
[merged r7138 from branches/ew]
git-svn-id: https://svn.musicpd.org/mpd/trunk@7240 09075e82-0dd4-0310-85a5-a0d7c8717e4f
Do explicit casts before comparing signed with unsigned. The one in
log.c actually fixes another warning: in the expanded macro, there may
be a check "logLevel>=0", which is always true.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7230 09075e82-0dd4-0310-85a5-a0d7c8717e4f
Use unsigned integers in decoderParent() for chunk numbers which
cannot be negative.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7226 09075e82-0dd4-0310-85a5-a0d7c8717e4f
When the decoder receives SIGCONT during waitNotify(), the kernel
restarts the read() system call. This lets the decoder process block
indefinitely, while the player process waits for it to react. This
should probably be solved with a proper signal handler which aborts
the read() system call, but for now, we just write to the pipe to make
it wake up.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7216 09075e82-0dd4-0310-85a5-a0d7c8717e4f
When the decoder process is faster than the player process, all
decodedd buffers are full at some point in time. The decoder has to
wait for buffers to become free (finished playing). It used to do
this by polling the buffer status 100 times a second.
This generates a lot of unnecessary CPU wakeups. This patch adds a
way for the player process to notify the decoder process that it may
continue its work.
We could use pthread_cond for that, unfortunately inter-process
mutexes/conds are not supported by some kernels (Linux), so we cannot
use this light-weight method until mpd moves to using threads instead
of processes. The other method would be semaphores, which
historically are global resources with a unique name; this historic
API is cumbersome, and I wanted to avoid it.
I came up with a quite naive solution for now: I create an anonymous
pipe with pipe(), and the decoder process reads on that pipe. Until
the player process sends data on it as a signal, the decoder process
blocks.
This can be optimized in a number of ways:
- if the decoder process is still working (instead of waiting for
buffers), we could save the write() system call, since there is
nobody waiting for the notification.
[ew: I tried this using a counter in shared memory, didn't help]
- the pipe buffer will be full at some point, when the decoder thread
is too slow. For this reason, the writer side of the pipe is
non-blocking, and mpd can ignore the resulting EWOULDBLOCK.
- since we have shared memory, we could check whether somebody is
actually waiting without a context switch, and we could just not
write the notification byte.
[ew: tried same method/result as first point above]
- if there is already a notification in the pipe, we could also not
write another one.
[ew: tried same method/result as first/third points above]
- the decoder will only consume 64 bytes at a time. If the pipe
buffer is full, this will result in a lot of read() invocations.
This does not hurt badly, but on a heavily loaded system, this might
add a little bit more load. The preceding optimizations however
are able eliminate the this.
- finally, we should use another method for inter process
notifications - maybe kill() or just make mpd use threads, finally.
In spite of all these possibilities to optimize this code further,
this pipe notification trick is faster than the 100 Hz poll. On my
machine, it reduced the number of wakeups to less than 30%.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7215 09075e82-0dd4-0310-85a5-a0d7c8717e4f
[ew: cleaned up the dirty union hack a bit]
Signed-off-by: Eric Wong <normalperson@yhbt.net>
git-svn-id: https://svn.musicpd.org/mpd/trunk@7180 09075e82-0dd4-0310-85a5-a0d7c8717e4f
This will make refactoring features easier, especially now that
pthreads support and larger refactorings are on the horizon.
Hopefully, this will make porting to other platforms (even
non-UNIX-like ones for masochists) easier, too.
os_compat.h will house all the #includes for system headers
considered to be the "core" of MPD. Headers for optional
features will be left to individual source files.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7130 09075e82-0dd4-0310-85a5-a0d7c8717e4f
DECODE_STATE_STOP is always set as dc->state, and dc->stop
is always cleared. So handle it in decodeStart once rather
than doing it in every plugin.
While we're at it, fix a long-standing (but difficult to
trigger) bug in mpc_decode where we failed to return
if mpc_decoder_initialize() fails.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7122 09075e82-0dd4-0310-85a5-a0d7c8717e4f
thread-safety work in preparation for rewrite to use pthreads
Expect no regressions against trunk (r7078), possibly minor
performance improvements in update (due to fewer heap
allocations), but increased stack usage.
Applied the following patches:
* maxpath_str for reentrancy (temporary fix, reverted)
* path: start working on thread-safe variants of these methods
* Re-entrancy work on path/character-set conversions
* directory.c: exploreDirectory() use reentrant functions here
* directory/update: more use of reentrant functions + cleanups
* string_toupper: a strdup-less version of strDupToUpper
* get_song_url: a static-variable-free version of getSongUrl()
* Use reentrant/thread-safe get_song_url everywhere
* replace rmp2amp with the reentrant version, rmp2amp_r
* Get rid of the non-reentrant/non-thread-safe rpp2app, too.
* buffer2array: assert strdup() returns a usable value in unit tests
* replace utf8ToFsCharset and fsCharsetToUtf8 with thread-safe variants
* fix storing playlists w/o absolute paths
* parent_path(), a reentrant version of parentPath()
* parentPath => parent_path for reentrancy and thread-safety
* allow "make test" to automatically run embedded unit tests
* remove convStrDup() and maxpath_str()
* use MPD_PATH_MAX everywhere instead of MAXPATHLEN
* path: get rid of appendSlash, pfx_path and just use pfx_dir
* get_song_url: fix the ability to play songs in the top-level music_directory
git-svn-id: https://svn.musicpd.org/mpd/trunk@7106 09075e82-0dd4-0310-85a5-a0d7c8717e4f
We need to SIGCONT the decoder process to allow for seeking
while paused.
git-svn-id: https://svn.musicpd.org/mpd/trunk@6864 09075e82-0dd4-0310-85a5-a0d7c8717e4f
The problems I had were related to the OSS driver and USB
device I was using. The problems existed even with the old
busy-waiting scheme enabled.
OSS - Bithead USB => bad
ALSA - Bithead USB => OK
OSS - Onboard i8x0 => OK
ALSA - Onboard i8x0 => OK
bad - slow shutdown, pauses, dropped audio after pause/resume
git-svn-id: https://svn.musicpd.org/mpd/trunk@6861 09075e82-0dd4-0310-85a5-a0d7c8717e4f
Until we can fix it properly (or replace it with a cleaner event
system), I don't want this in trunk. Currently there are
strange pauses when queueing and during shutdown that I can't
seem to figure out right away.
git-svn-id: https://svn.musicpd.org/mpd/trunk@6860 09075e82-0dd4-0310-85a5-a0d7c8717e4f
As unfortunate as it is to remove such useful debugging messages, it's
necessary to fix a potential deadlock with signal handling. A bunch of
functions the debug functions call aren't safe to call from a signal
handler. There are some alternate solutions, but they're neither pretty
nor simple. So just remove them entirely for now.
git-svn-id: https://svn.musicpd.org/mpd/trunk@6828 09075e82-0dd4-0310-85a5-a0d7c8717e4f
as with the stop command, this will cause the player and decoder
to suspend and not wake up hundreds of times a second to poll
a variable for wakeup. This will reduce power consumption
on some CPUs while mpd is paused and not playing.
tests:
pause && unpause => OK
pause && stop && play => OK
pause && exit && restart w/statefile && unpause => OK
pause && block sound device && \
unpause => failed to open sound device \
=> still paused and suspended => unblock sound device &&
unpause => OK (playing)
In all cases, the player process releases the audio device
when paused before going into the suspended state.
git-svn-id: https://svn.musicpd.org/mpd/trunk@6822 09075e82-0dd4-0310-85a5-a0d7c8717e4f
Currently, if we start decoding while the pause flag is set, we open the
audio device and leave it opened, blocking other apps from using it. The
obvious thing to do is to not open the audio device if the pause flag is
set, but the open call also sets the audio format. Therefore I'm leaving
the open call in, and just closing it immediately afterwards if the pause
flag is set.
git-svn-id: https://svn.musicpd.org/mpd/trunk@6745 09075e82-0dd4-0310-85a5-a0d7c8717e4f
Only wavpack implements both fileDecodeFunc and streamDecodeFunc, and it's
fileDecodeFunc provides more functionality. So try using that first.
This commit also fixes a bug where the plugin test loop wouldn't break once
a suitable plugin was found if it used fileDecodeFunc.
git-svn-id: https://svn.musicpd.org/mpd/trunk@6655 09075e82-0dd4-0310-85a5-a0d7c8717e4f
top depending on !quit, which doesn't set it anywhere before the if (quit)
block is reached, and the inner one which doesn't set quit at all. Since
it's a local variable and can't be modified externally, it'll never be hit.
git-svn-id: https://svn.musicpd.org/mpd/trunk@6524 09075e82-0dd4-0310-85a5-a0d7c8717e4f
completely stopped. Instead, send them SIGSTOP to pause the process until
they're needed again. Then send them SIGCONT instead of re-spawning them.
git-svn-id: https://svn.musicpd.org/mpd/trunk@6485 09075e82-0dd4-0310-85a5-a0d7c8717e4f
#2) fix a deadlock condition when attempting to seek if the decoder quit and returned to playerInit()
git-svn-id: https://svn.musicpd.org/mpd/trunk@5325 09075e82-0dd4-0310-85a5-a0d7c8717e4f
Some compilers and linkers aren't smart enough to optimize this,
as global variables are implictly initialized to zero. As a
result, binaries are a bit smaller as more goes in the .bss and
less in the text section.
git-svn-id: https://svn.musicpd.org/mpd/trunk@5254 09075e82-0dd4-0310-85a5-a0d7c8717e4f
This fixes a bug where streams that won't play somehow appear with the
metadata of a previously played stream. As far as I can tell, the only
reason this is done is to sync any buffered metadata with the displayed
metadata when decoding stops, so there should be no other adverse effects.
git-svn-id: https://svn.musicpd.org/mpd/trunk@5161 09075e82-0dd4-0310-85a5-a0d7c8717e4f
I'm checking for zero-size allocations and assert()-ing them,
so we can more easily get backtraces and debug problems, but we'll
also allow -DNDEBUG people to live on the edge if they wish.
We do not rely on errno when checking for OOM errors because
some implementations of malloc do not set it, and malloc
is commonly overridden by userspace wrappers.
I've spent some time looking through the source and didn't find any
obvious places where we would explicitly allocate 0 bytes, so we
shouldn't trip any of those assertions.
We also avoid allocating zero bytes because C libraries don't
handle this consistently (some return NULL, some not); and it's
dangerous either way.
git-svn-id: https://svn.musicpd.org/mpd/trunk@4690 09075e82-0dd4-0310-85a5-a0d7c8717e4f
Add -Wmissing-prototypes if compiling with gcc
Static where possible
git-svn-id: https://svn.musicpd.org/mpd/trunk@4657 09075e82-0dd4-0310-85a5-a0d7c8717e4f
utf8ToFsCharset() and fsCharsetToUtf8() got very broken in r4311, and
resulted in several commits to fix those leaks. Unfortunately, not all
of those newly introduced leaks were fixed, nor was the result pretty.
Also, fixed a double-free in lsPlaylists(). This is very hard
to trigger (and therefore exploit) at the moment because we
check printDirectoryInfo() beforehand.
Intended behavior for utf8ToFsCharset() and fsCharsetToUtf8() as
God^H^H^Hshank originally intended is now documented in path.h
to prevent future errors like this.
mpd could still use some good valgrind testing before the 0.12.0
release.
<plug>In addition to reducing heap fragmentation, malloc
reductions from mpd-ke greatly reduces the chance of leaks from
happening due to programming errors.</plug>
git-svn-id: https://svn.musicpd.org/mpd/trunk@4639 09075e82-0dd4-0310-85a5-a0d7c8717e4f
* less-commonly compiled things like ao/mvp outputs
* Adding -Wno-transparent-union to SPARSE_FLAGS makes it check
inside decode.c, directory.c, player.c, and sig_handlers.c
* remove unused variables leftover from the master process
in sig_handlers.c
git-svn-id: https://svn.musicpd.org/mpd/trunk@4598 09075e82-0dd4-0310-85a5-a0d7c8717e4f
* Moved all logging-related stuff into log.c
(and not myfprintf.c)
* ISO C90-compliant strftime usage:
%e and %R replaced with %d and %H:%M respectively
* Got rid of variadic macros since some old-school compilers
don't like them
* compiling with -DNDEBUG disables the DEBUG() macro
git-svn-id: https://svn.musicpd.org/mpd/trunk@4512 09075e82-0dd4-0310-85a5-a0d7c8717e4f
strncpy isn't really safe because it doesn't guarantee null termination,
and we have had to work around it in several places.
strlcpy (from OpenBSD) isn't great, either because it often leaves
errors going unchecked (by truncating strings).
So we'll add the pathcpy_trunc() function with is basically strlcpy
with a hardcoded MAXPATHLEN as the limit, and we'll acknowledge
truncation since we only work on paths and MAXPATHLEN should be
set correctly by the system headers[1].
file-specific notes:
inputStream_http:
eyeballing the changes here, it seems to look alright but I
haven't actually tested it myself.
ls:
don't even bother printing a file if the filename is too long
(and when is it ever?) since we won't be able to read it anyways.
metadataChunk:
it's only metadata, and it's only for showin the user, so truncating
it here souldn't be a big issue.
memset to zero in init is unecessary, so lets not waste cycles
[1] - If the system headers are screwed up, then we're majorly
screwed regardless of what we do :x
git-svn-id: https://svn.musicpd.org/mpd/trunk@4491 09075e82-0dd4-0310-85a5-a0d7c8717e4f
