[Bug 30118] Sound occasionally stalls past underrun, bypassing the ALSA lead-in writer

http://bugs.winehq.org/show_bug.cgi?id=30118

Jörg Höhle  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #2 from Jörg Höhle  2012-03-11 11:50:45 CDT ---
There's something I don't understand.  James' Civ5 log at bug #29294, comment
#23 look like ALSA stalled after the broken pipe, despite 5120 frames in
buffer.  However no stalling is seen in the render test in bug #29294, comment
#35, where another underrun soon follows the broken pipe.  How to explain the
difference?

Civ5:
 23.054:dump_fmt nSamplesPerSec: 44100 16bit stereo
 23.054:AudioClient_Initialize ALSA period: 5512 frames
 23.055:AudioClient_Initialize ALSA buffer: 16537 frames
 23.055:AudioClient_Initialize MMDevice period: 441 frames
 23.055:AudioClient_Initialize MMDevice buffer: 5120 frames
mmdevapi render:
1838.608:dump_fmt nSamplesPerSec: 48000 float stereo
1838.609:AudioClient_Initialize ALSA period: 6000 frames
1838.609:AudioClient_Initialize ALSA buffer: 18000 frames
1838.609:AudioClient_Initialize MMDevice period: 480 frames
1838.609:AudioClient_Initialize MMDevice buffer: 2400 frames

James, please use the Civ5 parameters with the render test. After GetMixFormat
in mmdevapi/tests/render.c:test_worse_case, add:
    fmt->wFormatTag = WAVE_FORMAT_PCM;
    fmt->cbSize = 0;
    fmt->nSamplesPerSec = 44100;
    fmt->nAvgBytesPerSec= fmt.nBlockAlign * fmt.nSamplesPerSec;