More correct fix for dsp empty case

[Jeremy White]

It turns out that it's only my OSS implementation that properly
respects the SND_POST ioctl; others are not so nice.  This makes
the SND_GETOSPACE call inaccurate when we're down to the last
fragment or two.  This patch uses the clock to predict when
that dsp is clear (rather than just assuming it's autmoatically
clear), and allows 'good' OSS implementations to respond to
a SND_POST ioctl.

This patch almost eliminates all of the test failures from
winmm, so now I can start thinking about timing problems...

Cheers,

Jer
-------------- next part --------------
A non-text attachment was scrubbed...
Name: fragfix2.diff
Type: text/x-patch
Size: 4342 bytes
Desc: not available
Url : http://www.winehq.org/pipermail/wine-patches/attachments/20050223/cd717c6c/fragfix2.bin