Start incorporating the ARM integrated assembler.

src/IceAssembler.h

Index: src/IceAssembler.h
diff --git a/src/IceAssembler.h b/src/IceAssembler.h
index 8247e66cf1af8f27cdf3d686781cdbc5673dc821..68ff3f3f1a5e95d44a071c73988edb17caf91251 100644
--- a/src/IceAssembler.h
+++ b/src/IceAssembler.h
@@ -30,6 +30,8 @@
 
 namespace Ice {
 
+class GlobalContext;

[Jim Stichnoth, 2015/10/08 23:50:41]

GlobalContext is already forward-declared in IceDefs.h.

[John, 2015/10/09 12:12:23]

ah, the beauty of forward decs and cyclic includes!...

I tend to side with Karl here: if we need to forward declare something (i.e.,
Somehow there's a cycle between the #includes in IceGlobalcontext.h and
IceAssembler.h) then that should be done in the file that needs it.

[Jim Stichnoth, 2015/10/09 13:36:30]

Well, I tried out the patch and everything compiles without this forward
declaration.  Also note the preexisting declaration below:
  void emitIASBytes(GlobalContext *Ctx) const;
So this wasn't a problem that needed fixing.

[Karl, 2015/10/09 19:08:18]

Removed.

+
 /// A Label can be in one of three states:
 ///  - Unused.
 ///  - Linked, unplaced and tracking the position of branches to the label.
@@ -64,17 +66,20 @@ public:
     return Position - kWordSize;
   }
 
+  void setPosition(intptr_t NewValue) { Position = NewValue; }
+
   bool isBound() const { return Position < 0; }
   bool isLinked() const { return Position > 0; }
+
   virtual bool isUnused() const { return Position == 0; }
 
-protected:
   void bindTo(intptr_t position) {
     assert(!isBound());
     Position = -position - kWordSize;
     assert(isBound());
   }
 
+protected:
   void linkTo(intptr_t position) {
     assert(!isBound());
     Position = position + kWordSize;
@@ -83,7 +88,6 @@ protected:
 
   intptr_t Position = 0;
 
-private:
   // TODO(jvoung): why are labels offset by this?
   static constexpr uint32_t kWordSize = sizeof(uint32_t);
 };
@@ -272,6 +276,7 @@ public:
     return Buffer.createFixup(Kind, Value);
   }
 
+  // TODO(kschimpf): Remove Ctx argument, no longer needed.

[Jim Stichnoth, 2015/10/08 23:50:41]

Can you just do it now, and not leave a TODO?

   void emitIASBytes(GlobalContext *Ctx) const;

[John, 2015/10/09 12:12:23]

Just a personal preference, but I would not introduce this GlobalContext member.
I hate that (because LLVM Style Guide) parameters, locals, and class members all
look the same. I try to have local declarations (locals and/or parameters) as
much as I can. Feel free to ignore.

[Karl, 2015/10/09 19:08:17]

While I agree with this comment in general, the global context had to be added
to fix other issues (which would require major changes to many functions to pass
as parameters). Hence, having to pass something that is local appears to
introduce another set of issues. Changing to remove argument Ctx.

   bool getInternal() const { return IsInternal; }
   void setInternal(bool Internal) { IsInternal = Internal; }
@@ -286,8 +291,8 @@ public:
   AssemblerKind getKind() const { return Kind; }
 
 protected:
-  explicit Assembler(AssemblerKind Kind)
-      : Kind(Kind), Allocator(), Buffer(*this) {}
+  explicit Assembler(AssemblerKind Kind, GlobalContext *Ctx)
+      : Kind(Kind), Allocator(), Ctx(Ctx), Buffer(*this) {}
 
 private:
   const AssemblerKind Kind;
@@ -305,6 +310,7 @@ private:
   bool Preliminary = false;
 
 protected:
+  GlobalContext *Ctx;
   // Buffer's constructor uses the Allocator, so it needs to appear after it.
   // TODO(jpp): dependencies on construction order are a nice way of shooting
   // yourself in the foot. Fix this.