Set 'flat' interpolation qualifier for double precision input
variable as defined by the GL_ARB_gpu_shader_fp64 spec:

  "This extension does not support interpolation of double-precision
  values; doubles used as fragment shader inputs must be qualified
  as "flat"."

Also modified vertex output side, as the test uses GLSL 1.50 where
interpolation qualifier must match between stages.

(This test still fails on Nvidia binary driver (319.32) which would
additionally require GLSL >= 4.00 and GL_NV_gpu_shader5 for it to
work.)

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>