http://bugs.winehq.org/show_bug.cgi?id=14717
--- Comment #152 from Alexander E. Patrakov patrakov@gmail.com 2011-02-22 18:47:27 CST --- I wrote a simple directsound program that creates primary & secondary buffers at different rates and locks/unlocks each portion of the secondary buffer twice: once writing a portion of a sine wave there, and once writing silence. The end result is expected to be silence, and is actually silence. The patch doesn't break this test with both relevant values of MaxShadowSize, so my concerns about this were invalid.