Locking

[folkertvanheusden]

The alsa-thread and main-thread can both access the queue concurrently. This may cause problems. This can be verified using helgrind:

==68196== Possible data race during write of size 4 at 0x4CA44B4 by thread #2
==68196== Locks held: none
==68196==    at 0x5851349: MidiInApi::MidiQueue::push(MidiInApi::MidiMessage const&) (RtMidi.cpp:696)
==68196==    by 0x5856333: alsaMidiHandler(void*) (RtMidi.cpp:1668)
==68196==    by 0x484D82A: ??? (in /usr/libexec/valgrind/vgpreload_helgrind-amd64-linux.so)
==68196==    by 0x4A2D44F: start_thread (pthread_create.c:473)
==68196==    by 0x4B5DD52: clone (clone.S:95)
==68196== 
==68196== This conflicts with a previous read of size 4 by thread #1
==68196== Locks held: none
==68196==    at 0x5851214: MidiInApi::MidiQueue::size(unsigned int*, unsigned int*) (RtMidi.cpp:671)
==68196==    by 0x585145F: MidiInApi::MidiQueue::pop(std::vector<unsigned char, std::allocator<unsigned char> >*, double*, RtMidiError::Type*, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >*) (RtMidi.cpp:709)
==68196==    by 0x58516B5: MidiInApi::getMessage(std::vector<unsigned char, std::allocator<unsigned char> >*) (RtMidi.cpp:657)
==68196==    by 0x584E54F: getMessage (RtMidi.h:623)
==68196==    by 0x584E54F: __pyx_pf_6rtmidi_7_rtmidi_6MidiIn_12get_message (_rtmidi.cpp:8167)
==68196==    by 0x584E54F: __pyx_pw_6rtmidi_7_rtmidi_6MidiIn_13get_message(_object*, _object*) (_rtmidi.cpp:8133)
==68196==    by 0x53578C: ??? (in /usr/bin/python3.9)
==68196==    by 0x516542: _PyEval_EvalFrameDefault (in /usr/bin/python3.9)
==68196==    by 0x514A74: ??? (in /usr/bin/python3.9)
==68196==    by 0x51480A: _PyEval_EvalCodeWithName (in /usr/bin/python3.9)
==68196==  Address 0x4ca44b4 is 84 bytes inside a block of size 232 alloc'd
==68196==    at 0x4845033: operator new(unsigned long) (in /usr/libexec/valgrind/vgpreload_helgrind-amd64-linux.so)
==68196==    by 0x58558C1: RtMidiIn::openMidiApi(RtMidi::Api, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, unsigned int) (RtMidi.cpp:416)
==68196==    by 0x5855A20: RtMidiIn::RtMidiIn(RtMidi::Api, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, unsigned int) (RtMidi.cpp:450)
==68196==    by 0x5848DA9: __pyx_pf_6rtmidi_7_rtmidi_6MidiIn___cinit__ (_rtmidi.cpp:7400)
==68196==    by 0x5848DA9: __pyx_pw_6rtmidi_7_rtmidi_6MidiIn_1__cinit__ (_rtmidi.cpp:7287)
==68196==    by 0x5848DA9: __pyx_tp_new_6rtmidi_7_rtmidi_MidiIn(_typeobject*, _object*, _object*) (_rtmidi.cpp:10705)
==68196==    by 0x521B55: _PyObject_MakeTpCall (in /usr/bin/python3.9)
==68196==    by 0x51B9F7: _PyEval_EvalFrameDefault (in /usr/bin/python3.9)
==68196==    by 0x514A74: ??? (in /usr/bin/python3.9)
==68196==    by 0x51480A: _PyEval_EvalCodeWithName (in /usr/bin/python3.9)
==68196==    by 0x5FB256: PyEval_EvalCode (in /usr/bin/python3.9)
==68196==    by 0x6205FA: ??? (in /usr/bin/python3.9)
==68196==    by 0x61B723: ??? (in /usr/bin/python3.9)
==68196==    by 0x61FB2C: ??? (in /usr/bin/python3.9)
==68196==  Block was alloc'd by thread #1

with something like:

#! /usr/bin/python3

import rtmidi
import time

midiin = rtmidi.MidiIn()

ports = midiin.get_ports()

nr = ports.index('CH345:CH345 MIDI 1 20:0')
midiin.open_port(nr)

while True:
    print('bla', midiin.get_message())
    time.sleep(0.1)

using:

valgrind --tool=helgrind python3 test3.py

[SpotlightKid]

Thanks for your PR and sorry for the slow response. I currently have only little time for FLOSS development but I'll look at your change as soon as I can.

[SpotlightKid]

It might be better, though, to file this PR upstream (https://github.com/thestk/rtmidi). Upstream is even slower at incorporating PRs, though, so including the change in my fork first is a possibility.