http://bugs.winehq.org/show_bug.cgi?id=15074
--- Comment #3 from Alexander Dorofeyev alexd4@inbox.lv 2008-08-31 06:59:20 --- The actual debug channel for that file seems to be dsalsa. I'll try to get a log. I did notice a suspicious detail when messing with debug outputs. During normal operation, "writepos" appears to always be in front of "playpos", which makes sense as buffer should be filled in front of play position. E.g.:
err:dsalsa:IDsDriverBufferImpl_GetPosition hw_pptr=0x000002f0, hw_wptr=0x00001510 playpos=3008, writepos=21568 err:dsalsa:IDsDriverBufferImpl_GetPosition hw_pptr=0x000004cf, hw_wptr=0x000016f0 playpos=4924, writepos=23488 err:dsalsa:IDsDriverBufferImpl_GetPosition hw_pptr=0x000006af, hw_wptr=0x000018cf playpos=6844, writepos=25404
But after the sound disappears, writepos is trailing playpos, e.g.:
err:dsalsa:IDsDriverBufferImpl_GetPosition hw_pptr=0x0000031f, hw_wptr=0x00000140 playpos=3196, writepos=1280 err:dsalsa:IDsDriverBufferImpl_GetPosition hw_pptr=0x000004ff, hw_wptr=0x0000031f playpos=5116, writepos=3196 err:dsalsa:IDsDriverBufferImpl_GetPosition hw_pptr=0x000006df, hw_wptr=0x000004ff playpos=7036, writepos=5116
Maybe that's the problem.