to replace all-zeros hash, and ease implementation for
separate LS and RI databases, as documented in i2np spec.

   types Standard, IRC, and Socks IRC, if non-shared.

   - Don't rate peers in some countries as high capacity
   - Don't enable auto-floodfill in some countries
   - Don't prefer floodfills in some countries

  - code cleanup

 - Remove duplicate load3 and load4 methods
 - Change load3 return type to int
 - Comment out dead stores
 - Re-roll the add, subtract, and negate loops;
   there's no speed benefit or timing reason to unroll them
 - Check for field already set
 - Remove shifts by 0

 - code cleanup in direction handling

 - More synchronization

  - Fix main() model and family calculation
  - Add model string fetch from processor
  - AMD model string tweaks

  - Fix model and family calculation
  - Fix most AMD family 15 IDs
  - Add AMD Llano, Jaguar, Bulldozer 2
  - Add Intel Ivy Bridge, Haswell, Broadwell, Penryn, Pineview, Cedarview, Bay Trail, Avoton, and others
  - Set best-guess capabilities for most Intel processors
  - Supply best-guess model string in most cases
  - Processors listed above, and some others, may see crypto speedups as a result
  - Code cleanup, reduce number of JNI calls
  - Merge dup cases
  - Tab removal
  - Javadocs

SAM:
 - Don't publish LS for DIRECTION=RECEIVE
 - Set default tunnel name

   - Increase default to 3 hops (ticket #966)
   - Reduce upload threshold for auto-stop
   - Revert addition of js mime type, it was already in the default

 - Show info hash on details page
 - Null storage check just in case