There is a coarse description on the "features" page of how blocking operations in general are handled. The same principle applies to all kinds of blocking operations (waiting for a mutex lock, sleeping, waiting for data on a pipe etc.).
Looking at the commit history, GDC has been in steady development since about 4 years now after a pause and LDC is being steadily developed since its beginnings (with changing main authors). Both have very active maintainers these days.
My feeling is that you really have to try it out for some time to see why it really makes a difference* . And, although C++11 indeed has caught up in some areas, there are in fact many smaller and some larger things that still do make a big difference in practice. But in general the fact that it's not just one or maybe a few primary features indeed probably makes it more difficult to judge the actual benefits from the outside.
The drawbacks today are in my eyes mostly the relatively minimal support for non-PC platforms and that a number of language features are not yet completely finished. But if the main target are x86 based systems and the use of cutting edge features is not the primary concern, D2 is a great development tool today.
* I can only really talk about my experiences going from C++ -> D, but it was a long process where my way of designing systems and solving problems was gradually altered in many areas - much to the better. And the immediate reward of course is the extreme gain in productivity due to concise syntax, fast compilation times and, last but not least, the lack of a need for header files.