A few fixes for generalized intrinsics

[Lupusoru, Razvan A]

Hi all,

I have fixed the following issues following the generalized intrinsics implementation. These issues were all reproducible with intrinsification turned off - aka they were java implementation bugs:

-          trueMask method had implementation in both species and mask class. The mask class implementation was calling itself leading to StackOverflow. The condition leading to the overflow was calling of "allTrue".

-          Mask operations had incorrect default implementation for "and" and "or"

-          mulAll was using 0 as starting value leading to all results to be zero.

The patch is available for review here:
http://cr.openjdk.java.net/~rlupusoru/panama/webrev_genintrinfix_00/

Thanks,
Razvan

PS: The webrev mentions two revisions including "49068". This seems to have been generated due to a limitation in webrev tool dealing with branch split. No changes from 49068 are included in this patch since it is a local change not yet merged into panama.