cudatoolkit: don't move $out/include to $out/usr_include

This effectively reverts 86c283824f
("If cuda headers are presented to nix [...]") and all the following
workarounds that was added due to that commit.

As far as I can tell[1] this hack isn't needed anymore. And moving
includes to $out/usr_include causes pain for cudatoolkit users, so
better get rid of it.

In patches that did more than the $out/usr_include workaround, I only
changed the line back to $out/include instead of re-generating the
patches and fully removing the changed line.

[1]: I build tested blender and caffe, and temporarily added
recurseIntoAttrs to rPackages and haskellPackages so that nox-review
could get proper coverage. However, many of the packages do not build
even before this patch. I also built CUDA samples with cudatoolkit7
that ran fine.

pkgs/development/r-modules/default.nix

@@ -1566,7 +1566,6 @@ let
     });
 
     iFes = old.iFes.overrideDerivation (attrs: {
-      patches = [ ./patches/iFes.patch ];
       CUDA_HOME = "${pkgs.cudatoolkit}";
     });
 
@@ -1661,7 +1660,7 @@ let
       patches = [ ./patches/gmatrix.patch ];
       CUDA_LIB_PATH = "${pkgs.cudatoolkit}/lib64";
       R_INC_PATH = "${pkgs.R}/lib/R/include";
-      CUDA_INC_PATH = "${pkgs.cudatoolkit}/usr_include";
+      CUDA_INC_PATH = "${pkgs.cudatoolkit}/include";
     });
 
     # It seems that we cannot override meta attributes with overrideDerivation.