Rebase from GCC

src/x86/win32.S

@@ -45,6 +45,7 @@ _TEXT SEGMENT
 ffi_call_win32 PROC NEAR,
     ffi_prep_args : NEAR PTR DWORD,
     ecif          : NEAR PTR DWORD,
+    cif_abi       : DWORD,
     cif_bytes     : DWORD,
     cif_flags     : DWORD,
     rvalue        : NEAR PTR DWORD,
@@ -64,6 +65,19 @@ ffi_call_win32 PROC NEAR,
         ;; Return stack to previous state and call the function
         add  esp, 8
 
+	;; Handle thiscall and fastcall
+	cmp cif_abi, 3 ;; FFI_THISCALL
+	jz do_thiscall
+	cmp cif_abi, 4 ;; FFI_FASTCALL
+	jnz do_stdcall
+	mov ecx, DWORD PTR [esp]
+	mov edx, DWORD PTR [esp+4]
+	add esp, 8
+	jmp do_stdcall
+do_thiscall:
+	mov ecx, DWORD PTR [esp]
+	add esp, 4
+do_stdcall:
         call fn
 
         ;; cdecl:   we restore esp in the epilogue, so there's no need to
@@ -405,7 +419,7 @@ _ffi_call_win32:
         movl  %esp,%ebp
 .LCFI1:
         # Make room for all of the new args.
-        movl  16(%ebp),%ecx                                                     
+        movl  20(%ebp),%ecx                                                     
         subl  %ecx,%esp
  
         movl  %esp,%eax
@@ -417,19 +431,34 @@ _ffi_call_win32:
  
         # Return stack to previous state and call the function
         addl  $8,%esp
- 
+
+	# Handle fastcall and thiscall
+	cmpl $3, 16(%ebp)  # FFI_THISCALL
+	jz .do_thiscall
+	cmpl $4, 16(%ebp) # FFI_FASTCALL
+	jnz .do_fncall
+	movl (%esp), %ecx
+	movl 4(%esp), %edx
+	addl $8, %esp
+	jmp .do_fncall
+.do_thiscall:
+	movl (%esp), %ecx
+	addl $4, %esp
+
+.do_fncall:
+	 
         # FIXME: Align the stack to a 128-bit boundary to avoid
         # potential performance hits.
 
-        call  *28(%ebp)
+        call  *32(%ebp)
  
         # stdcall functions pop arguments off the stack themselves
 
         # Load %ecx with the return type code
-        movl  20(%ebp),%ecx
+        movl  24(%ebp),%ecx
  
         # If the return value pointer is NULL, assume no return value.
-        cmpl  $0,24(%ebp)
+        cmpl  $0,28(%ebp)
         jne   0f
  
         # Even if there is no space for the return value, we are
@@ -488,50 +517,50 @@ _ffi_call_win32:
 
 .Lretint:
         # Load %ecx with the pointer to storage for the return value
-        movl  24(%ebp),%ecx
+        movl  28(%ebp),%ecx
         movl  %eax,0(%ecx)
         jmp   .Lepilogue
  
 .Lretfloat:
          # Load %ecx with the pointer to storage for the return value
-        movl  24(%ebp),%ecx
+        movl  28(%ebp),%ecx
         fstps (%ecx)
         jmp   .Lepilogue
  
 .Lretdouble:
         # Load %ecx with the pointer to storage for the return value
-        movl  24(%ebp),%ecx
+        movl  28(%ebp),%ecx
         fstpl (%ecx)
         jmp   .Lepilogue
  
 .Lretlongdouble:
         # Load %ecx with the pointer to storage for the return value
-        movl  24(%ebp),%ecx
+        movl  28(%ebp),%ecx
         fstpt (%ecx)
         jmp   .Lepilogue
  
 .Lretint64:
         # Load %ecx with the pointer to storage for the return value
-        movl  24(%ebp),%ecx
+        movl  28(%ebp),%ecx
         movl  %eax,0(%ecx)
         movl  %edx,4(%ecx)
 	jmp   .Lepilogue
 
 .Lretstruct1b:
         # Load %ecx with the pointer to storage for the return value
-        movl  24(%ebp),%ecx
+        movl  28(%ebp),%ecx
         movb  %al,0(%ecx)
         jmp   .Lepilogue
  
 .Lretstruct2b:
         # Load %ecx with the pointer to storage for the return value
-        movl  24(%ebp),%ecx
+        movl  28(%ebp),%ecx
         movw  %ax,0(%ecx)
         jmp   .Lepilogue
 
 .Lretstruct4b:
         # Load %ecx with the pointer to storage for the return value
-        movl  24(%ebp),%ecx
+        movl  28(%ebp),%ecx
         movl  %eax,0(%ecx)
         jmp   .Lepilogue