[PATCH] D9509: [cuda] Driver changes to build and stitch together host and device-side CUDA code.

[Douglas Katzman]

dougk added a subscriber: dougk.

================
Comment at: lib/Driver/Driver.cpp:822-823
@@ -820,3 +821,4 @@
 
 static unsigned PrintActions1(const Compilation &C, Action *A,
-                              std::map<Action*, unsigned> &Ids) {
+                              std::map<Action *, unsigned> &Ids);
+
----------------
tra wrote:
> echristo wrote:
> > Can you rearrange the code so that it doesn't need the forward declaration?
> Nope. PrintActions1 and PrintActionList call each other recursively. At least one of them must have forward declaration.
you could move PrintActionList back into PrintActions1 so that the "gpu binaries" clause and final clause each fall into some common code that performs the equivalent of what is now your PrintActionList function. (i.e. don't use mutually recursive functions)

If nothing else, consider renaming PrintActionList so not to imply that it performs the printing, perhaps "ActionListAsString". And use a range-based for loop. 

================
Comment at: lib/Driver/ToolChains.h:692
@@ -691,1 +691,3 @@
 
+class LLVM_LIBRARY_VISIBILITY Cuda : public Linux {
+public:
----------------
CudaToolChain would be are more readily understandable name.


http://reviews.llvm.org/D9509