Implemented proper support for naked asm using llvm module level asm. Still not 100% complete, but already 1000 times better that what we had before. Don's BignumX86 implementation from Tango (when turned into a standalone unittest) seems to fully work with no changes, and great performance :)

Fixed align N; in asm blocks.

Fixed inreg parameter passing on x86 for ref/out params.

Removed support for lazy initialization of function local static variables, I have no idea why I ever implemented this, it's not in the D spec, and DMD doesn't support it :P

Some of the global variable related changes might cause minor regressions, but they should be easily fixable.

gen/asmstmt.cpp

@@ -15,7 +15,6 @@
 #include <cassert>
 #include <deque>
 #include <iostream>
-#include <sstream>
 #include <cstring>
 
 //#include "d-lang.h"
@@ -157,6 +156,8 @@ Statement *AsmStatement::semantic(Scope *sc)
     if (err)
         fatal();
 
+    //puts(toChars());
+
     sc->func->inlineAsm = 1;
     sc->func->inlineStatus = ILSno; // %% not sure
     // %% need to set DECL_UNINLINABLE too?
@@ -699,3 +700,35 @@ Statement *AsmBlockStatement::semantic(Scope *sc)
 
     return CompoundStatement::semantic(sc);
 }
+
+//////////////////////////////////////////////////////////////////////////////
+
+void AsmStatement::toNakedIR(IRState *p)
+{
+    Logger::println("AsmStatement::toNakedIR(): %s", loc.toChars());
+    LOG_SCOPE;
+
+    // is there code?
+    if (!asmcode)
+        return;
+    AsmCode * code = (AsmCode *) asmcode;
+
+    // build asm stmt
+    std::ostringstream& asmstr = p->nakedAsm;
+    asmstr << "\t";
+    asmstr.write(code->insnTemplate, code->insnTemplateLen);
+    asmstr << std::endl;
+}
+
+void AsmBlockStatement::toNakedIR(IRState *p)
+{
+    Logger::println("AsmBlockStatement::toNakedIR(): %s", loc.toChars());
+    LOG_SCOPE;
+
+    // do asm statements
+    for (unsigned i=0; i<statements->dim; i++)
+    {
+        Statement* s = (Statement*)statements->data[i];
+        if (s) s->toNakedIR(p);
+    }
+}