ffi 1.10.0 → 1.15.5

data/ext/ffi_c/libffi/src/x86/sysv_intel.S

@@ -0,0 +1,995 @@
+/* -----------------------------------------------------------------------
+   sysv.S - Copyright (c) 2017  Anthony Green
+          - Copyright (c) 2013  The Written Word, Inc.
+          - Copyright (c) 1996,1998,2001-2003,2005,2008,2010  Red Hat, Inc.
+   
+   X86 Foreign Function Interface 
+
+   Permission is hereby granted, free of charge, to any person obtaining
+   a copy of this software and associated documentation files (the
+   ``Software''), to deal in the Software without restriction, including
+   without limitation the rights to use, copy, modify, merge, publish,
+   distribute, sublicense, and/or sell copies of the Software, and to
+   permit persons to whom the Software is furnished to do so, subject to
+   the following conditions:
+
+   The above copyright notice and this permission notice shall be included
+   in all copies or substantial portions of the Software.
+
+   THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND,
+   EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+   MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+   NONINFRINGEMENT.  IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
+   HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
+   WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+   OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+   DEALINGS IN THE SOFTWARE.
+   ----------------------------------------------------------------------- */
+
+#ifndef __x86_64__
+#ifdef _MSC_VER
+
+#define LIBFFI_ASM	
+#include <fficonfig.h>
+#include <ffi.h>
+#include <ffi_cfi.h>
+#include "internal.h" 
+
+#define C2(X, Y)  X ## Y
+#define C1(X, Y)  C2(X, Y)
+#define L(X)     C1(L, X)
+# define ENDF(X) X ENDP
+
+/* This macro allows the safe creation of jump tables without an
+   actual table.  The entry points into the table are all 8 bytes.
+   The use of ORG asserts that we're at the correct location.  */
+/* ??? The clang assembler doesn't handle .org with symbolic expressions.  */
+#if defined(__clang__) || defined(__APPLE__) || (defined (__sun__) && defined(__svr4__))
+# define E(BASE, X)	ALIGN 8
+#else
+# define E(BASE, X)	ALIGN 8; ORG BASE + X * 8
+#endif
+
+    .686P
+    .MODEL FLAT
+
+EXTRN	@ffi_closure_inner@8:PROC
+_TEXT SEGMENT
+
+/* This is declared as
+
+   void ffi_call_i386(struct call_frame *frame, char *argp)
+        __attribute__((fastcall));
+
+   Thus the arguments are present in
+
+        ecx: frame
+        edx: argp
+*/
+
+ALIGN 16
+PUBLIC @ffi_call_i386@8
+@ffi_call_i386@8 PROC
+L(UW0):
+	cfi_startproc
+ #if !HAVE_FASTCALL
+	mov	    ecx, [esp+4]
+	mov 	edx, [esp+8]
+ #endif
+	mov	    eax, [esp]		/* move the return address */
+	mov	    [ecx], ebp		/* store ebp into local frame */
+	mov 	[ecx+4], eax	/* store retaddr into local frame */
+
+	/* New stack frame based off ebp.  This is a itty bit of unwind
+	   trickery in that the CFA *has* changed.  There is no easy way
+	   to describe it correctly on entry to the function.  Fortunately,
+	   it doesn't matter too much since at all points we can correctly
+	   unwind back to ffi_call.  Note that the location to which we
+	   moved the return address is (the new) CFA-4, so from the
+	   perspective of the unwind info, it hasn't moved.  */
+	mov 	ebp, ecx
+L(UW1):
+	// cfi_def_cfa(%ebp, 8)
+	// cfi_rel_offset(%ebp, 0)
+
+	mov 	esp, edx		/* set outgoing argument stack */
+	mov 	eax, [20+R_EAX*4+ebp]	/* set register arguments */
+	mov 	edx, [20+R_EDX*4+ebp]
+	mov	    ecx, [20+R_ECX*4+ebp]
+
+	call	dword ptr [ebp+8]
+
+	mov	    ecx, [12+ebp]		/* load return type code */
+	mov 	[ebp+8], ebx		/* preserve %ebx */
+L(UW2):
+	// cfi_rel_offset(%ebx, 8)
+
+	and 	ecx, X86_RET_TYPE_MASK
+	lea 	ebx, [L(store_table) + ecx * 8]
+	mov 	ecx, [ebp+16]		/* load result address */
+	jmp	    ebx
+
+	ALIGN	8
+L(store_table):
+E(L(store_table), X86_RET_FLOAT)
+	fstp	DWORD PTR [ecx]
+	jmp	L(e1)
+E(L(store_table), X86_RET_DOUBLE)
+	fstp	QWORD PTR [ecx]
+	jmp	L(e1)
+E(L(store_table), X86_RET_LDOUBLE)
+	fstp	QWORD PTR [ecx]
+	jmp	L(e1)
+E(L(store_table), X86_RET_SINT8)
+	movsx	eax, al
+	mov	[ecx], eax
+	jmp	L(e1)
+E(L(store_table), X86_RET_SINT16)
+	movsx	eax, ax
+	mov	[ecx], eax
+	jmp	L(e1)
+E(L(store_table), X86_RET_UINT8)
+	movzx	eax, al
+	mov	[ecx], eax
+	jmp	L(e1)
+E(L(store_table), X86_RET_UINT16)
+	movzx	eax, ax
+	mov	[ecx], eax
+	jmp	L(e1)
+E(L(store_table), X86_RET_INT64)
+	mov	[ecx+4], edx
+	/* fallthru */
+E(L(store_table), X86_RET_int 32)
+	mov	[ecx], eax
+	/* fallthru */
+E(L(store_table), X86_RET_VOID)
+L(e1):
+	mov	    ebx, [ebp+8]
+	mov	    esp, ebp
+	pop 	ebp
+L(UW3):
+	// cfi_remember_state
+	// cfi_def_cfa(%esp, 4)
+	// cfi_restore(%ebx)
+	// cfi_restore(%ebp)
+	ret
+L(UW4):
+	// cfi_restore_state
+
+E(L(store_table), X86_RET_STRUCTPOP)
+	jmp	    L(e1)
+E(L(store_table), X86_RET_STRUCTARG)
+	jmp	    L(e1)
+E(L(store_table), X86_RET_STRUCT_1B)
+	mov 	[ecx], al
+	jmp	    L(e1)
+E(L(store_table), X86_RET_STRUCT_2B)
+	mov 	[ecx], ax
+	jmp	    L(e1)
+
+	/* Fill out the table so that bad values are predictable.  */
+E(L(store_table), X86_RET_UNUSED14)
+	int 3
+E(L(store_table), X86_RET_UNUSED15)
+	int 3
+
+L(UW5):
+	// cfi_endproc
+ENDF(@ffi_call_i386@8)
+
+/* The inner helper is declared as
+
+   void ffi_closure_inner(struct closure_frame *frame, char *argp)
+	__attribute_((fastcall))
+
+   Thus the arguments are placed in
+
+	ecx:	frame
+	edx:	argp
+*/
+
+/* Macros to help setting up the closure_data structure.  */
+
+#if HAVE_FASTCALL
+# define closure_FS	(40 + 4)
+# define closure_CF	0
+#else
+# define closure_FS	(8 + 40 + 12)
+# define closure_CF	8
+#endif
+
+FFI_CLOSURE_SAVE_REGS MACRO
+	mov 	[esp + closure_CF+16+R_EAX*4], eax
+	mov 	[esp + closure_CF+16+R_EDX*4], edx
+	mov 	[esp + closure_CF+16+R_ECX*4], ecx
+ENDM
+
+FFI_CLOSURE_COPY_TRAMP_DATA MACRO
+	mov 	edx, [eax+FFI_TRAMPOLINE_SIZE]      /* copy cif */
+	mov 	ecx, [eax+FFI_TRAMPOLINE_SIZE+4]    /* copy fun */
+	mov 	eax, [eax+FFI_TRAMPOLINE_SIZE+8];   /* copy user_data */
+	mov 	[esp+closure_CF+28], edx
+	mov 	[esp+closure_CF+32], ecx
+	mov 	[esp+closure_CF+36], eax
+ENDM
+
+#if HAVE_FASTCALL
+FFI_CLOSURE_PREP_CALL MACRO
+	mov	    ecx, esp                    /* load closure_data */
+	lea 	edx, [esp+closure_FS+4]     /* load incoming stack */
+ENDM
+#else
+FFI_CLOSURE_PREP_CALL MACRO
+	lea 	ecx, [esp+closure_CF]       /* load closure_data */
+	lea 	edx, [esp+closure_FS+4]     /* load incoming stack */
+	mov 	[esp], ecx
+	mov 	[esp+4], edx
+ENDM
+#endif
+
+FFI_CLOSURE_CALL_INNER MACRO UWN
+	call	@ffi_closure_inner@8
+ENDM
+
+FFI_CLOSURE_MASK_AND_JUMP MACRO LABEL
+	and	    eax, X86_RET_TYPE_MASK
+	lea 	edx, [LABEL+eax*8]
+	mov 	eax, [esp+closure_CF]       /* optimiztic load */
+	jmp	    edx
+ENDM
+
+ALIGN 16
+PUBLIC ffi_go_closure_EAX
+ffi_go_closure_EAX PROC C
+L(UW6):
+	// cfi_startproc
+	sub	esp, closure_FS
+L(UW7):
+	// cfi_def_cfa_offset(closure_FS + 4)
+	FFI_CLOSURE_SAVE_REGS
+	mov     edx, [eax+4]			/* copy cif */
+	mov 	ecx, [eax +8]			/* copy fun */
+	mov 	[esp+closure_CF+28], edx
+	mov 	[esp+closure_CF+32], ecx
+	mov 	[esp+closure_CF+36], eax	/* closure is user_data */
+	jmp	L(do_closure_i386)
+L(UW8):
+	// cfi_endproc
+ENDF(ffi_go_closure_EAX)
+
+ALIGN 16
+PUBLIC ffi_go_closure_ECX
+ffi_go_closure_ECX PROC C
+L(UW9):
+	// cfi_startproc
+	sub 	esp, closure_FS
+L(UW10):
+	// cfi_def_cfa_offset(closure_FS + 4)
+	FFI_CLOSURE_SAVE_REGS
+	mov 	edx, [ecx+4]			/* copy cif */
+	mov 	eax, [ecx+8]			/* copy fun */
+	mov 	[esp+closure_CF+28], edx
+	mov 	[esp+closure_CF+32], eax
+	mov 	[esp+closure_CF+36], ecx	/* closure is user_data */
+	jmp	L(do_closure_i386)
+L(UW11):
+	// cfi_endproc
+ENDF(ffi_go_closure_ECX)
+
+/* The closure entry points are reached from the ffi_closure trampoline.
+   On entry, %eax contains the address of the ffi_closure.  */
+
+ALIGN 16
+PUBLIC ffi_closure_i386
+ffi_closure_i386 PROC C
+L(UW12):
+	// cfi_startproc
+	sub	    esp, closure_FS
+L(UW13):
+	// cfi_def_cfa_offset(closure_FS + 4)
+
+	FFI_CLOSURE_SAVE_REGS
+	FFI_CLOSURE_COPY_TRAMP_DATA
+
+	/* Entry point from preceeding Go closures.  */
+L(do_closure_i386)::
+
+	FFI_CLOSURE_PREP_CALL
+	FFI_CLOSURE_CALL_INNER(14)
+	FFI_CLOSURE_MASK_AND_JUMP L(C1(load_table,2))
+
+    ALIGN 8
+L(load_table2):
+E(L(load_table2), X86_RET_FLOAT)
+	fld 	dword ptr [esp+closure_CF]
+	jmp	L(e2)
+E(L(load_table2), X86_RET_DOUBLE)
+	fld 	qword ptr [esp+closure_CF]
+	jmp	L(e2)
+E(L(load_table2), X86_RET_LDOUBLE)
+	fld 	qword ptr [esp+closure_CF]
+	jmp	L(e2)
+E(L(load_table2), X86_RET_SINT8)
+	movsx	eax, al
+	jmp	L(e2)
+E(L(load_table2), X86_RET_SINT16)
+	movsx	eax, ax
+	jmp	L(e2)
+E(L(load_table2), X86_RET_UINT8)
+	movzx	eax, al
+	jmp	L(e2)
+E(L(load_table2), X86_RET_UINT16)
+	movzx	eax, ax
+	jmp	L(e2)
+E(L(load_table2), X86_RET_INT64)
+	mov 	edx, [esp+closure_CF+4]
+	jmp	L(e2)
+E(L(load_table2), X86_RET_INT32)
+	nop
+	/* fallthru */
+E(L(load_table2), X86_RET_VOID)
+L(e2):
+	add 	esp, closure_FS
+L(UW16):
+	// cfi_adjust_cfa_offset(-closure_FS)
+	ret
+L(UW17):
+	// cfi_adjust_cfa_offset(closure_FS)
+E(L(load_table2), X86_RET_STRUCTPOP)
+	add 	esp, closure_FS
+L(UW18):
+	// cfi_adjust_cfa_offset(-closure_FS)
+	ret	4
+L(UW19):
+	// cfi_adjust_cfa_offset(closure_FS)
+E(L(load_table2), X86_RET_STRUCTARG)
+	jmp	L(e2)
+E(L(load_table2), X86_RET_STRUCT_1B)
+	movzx	eax, al
+	jmp	L(e2)
+E(L(load_table2), X86_RET_STRUCT_2B)
+	movzx	eax, ax
+	jmp	L(e2)
+
+	/* Fill out the table so that bad values are predictable.  */
+E(L(load_table2), X86_RET_UNUSED14)
+	int 3
+E(L(load_table2), X86_RET_UNUSED15)
+	int 3
+
+L(UW20):
+	// cfi_endproc
+ENDF(ffi_closure_i386)
+
+ALIGN 16
+PUBLIC	ffi_go_closure_STDCALL
+ffi_go_closure_STDCALL PROC C
+L(UW21):
+	// cfi_startproc
+	sub 	esp, closure_FS
+L(UW22):
+	// cfi_def_cfa_offset(closure_FS + 4)
+	FFI_CLOSURE_SAVE_REGS
+	mov 	edx, [ecx+4]			/* copy cif */
+	mov 	eax, [ecx+8]			/* copy fun */
+	mov 	[esp+closure_CF+28], edx
+	mov 	[esp+closure_CF+32], eax
+	mov 	[esp+closure_CF+36], ecx	/* closure is user_data */
+	jmp	L(do_closure_STDCALL)
+L(UW23):
+	// cfi_endproc
+ENDF(ffi_go_closure_STDCALL)
+
+/* For REGISTER, we have no available parameter registers, and so we
+   enter here having pushed the closure onto the stack.  */
+
+ALIGN 16
+PUBLIC ffi_closure_REGISTER
+ffi_closure_REGISTER PROC C
+L(UW24):
+	// cfi_startproc
+	// cfi_def_cfa(%esp, 8)
+	// cfi_offset(%eip, -8)
+	sub 	esp, closure_FS-4
+L(UW25):
+	// cfi_def_cfa_offset(closure_FS + 4)
+	FFI_CLOSURE_SAVE_REGS
+	mov	ecx, [esp+closure_FS-4] 	/* load retaddr */
+	mov	eax, [esp+closure_FS]		/* load closure */
+	mov	[esp+closure_FS], ecx		/* move retaddr */
+	jmp	L(do_closure_REGISTER)
+L(UW26):
+	// cfi_endproc
+ENDF(ffi_closure_REGISTER)
+
+/* For STDCALL (and others), we need to pop N bytes of arguments off
+   the stack following the closure.  The amount needing to be popped
+   is returned to us from ffi_closure_inner.  */
+
+ALIGN 16
+PUBLIC ffi_closure_STDCALL
+ffi_closure_STDCALL PROC C
+L(UW27):
+	// cfi_startproc
+	sub 	esp, closure_FS
+L(UW28):
+	// cfi_def_cfa_offset(closure_FS + 4)
+
+	FFI_CLOSURE_SAVE_REGS
+
+	/* Entry point from ffi_closure_REGISTER.  */
+L(do_closure_REGISTER)::
+
+	FFI_CLOSURE_COPY_TRAMP_DATA
+
+	/* Entry point from preceeding Go closure.  */
+L(do_closure_STDCALL)::
+
+	FFI_CLOSURE_PREP_CALL
+	FFI_CLOSURE_CALL_INNER(29)
+
+	mov 	ecx, eax
+	shr 	ecx, X86_RET_POP_SHIFT	    /* isolate pop count */
+	lea 	ecx, [esp+closure_FS+ecx]	/* compute popped esp */
+	mov 	edx, [esp+closure_FS]		/* move return address */
+	mov 	[ecx], edx
+
+	/* From this point on, the value of %esp upon return is %ecx+4,
+	   and we've copied the return address to %ecx to make return easy.
+	   There's no point in representing this in the unwind info, as
+	   there is always a window between the mov and the ret which
+	   will be wrong from one point of view or another.  */
+
+	FFI_CLOSURE_MASK_AND_JUMP  L(C1(load_table,3))
+
+    ALIGN 8
+L(load_table3):
+E(L(load_table3), X86_RET_FLOAT)
+	fld    DWORD PTR [esp+closure_CF]
+	mov     esp, ecx
+	ret
+E(L(load_table3), X86_RET_DOUBLE)
+	fld    QWORD PTR [esp+closure_CF]
+	mov     esp, ecx
+	ret
+E(L(load_table3), X86_RET_LDOUBLE)
+	fld    QWORD PTR [esp+closure_CF]
+	mov     esp, ecx
+	ret
+E(L(load_table3), X86_RET_SINT8)
+	movsx   eax, al
+	mov     esp, ecx
+	ret
+E(L(load_table3), X86_RET_SINT16)
+	movsx   eax, ax
+	mov     esp, ecx
+	ret
+E(L(load_table3), X86_RET_UINT8)
+	movzx   eax, al
+	mov     esp, ecx
+	ret
+E(L(load_table3), X86_RET_UINT16)
+	movzx   eax, ax
+	mov     esp, ecx
+	ret
+E(L(load_table3), X86_RET_INT64)
+	mov 	edx, [esp+closure_CF+4]
+	mov     esp, ecx
+	ret
+E(L(load_table3), X86_RET_int 32)
+	mov     esp, ecx
+	ret
+E(L(load_table3), X86_RET_VOID)
+	mov     esp, ecx
+	ret
+E(L(load_table3), X86_RET_STRUCTPOP)
+	mov     esp, ecx
+	ret
+E(L(load_table3), X86_RET_STRUCTARG)
+	mov 	esp, ecx
+	ret
+E(L(load_table3), X86_RET_STRUCT_1B)
+	movzx	eax, al
+	mov 	esp, ecx
+	ret
+E(L(load_table3), X86_RET_STRUCT_2B)
+	movzx	eax, ax
+	mov 	esp, ecx
+	ret
+
+	/* Fill out the table so that bad values are predictable.  */
+E(L(load_table3), X86_RET_UNUSED14)
+	int 3
+E(L(load_table3), X86_RET_UNUSED15)
+	int 3
+
+L(UW31):
+	// cfi_endproc
+ENDF(ffi_closure_STDCALL)
+
+#if !FFI_NO_RAW_API
+
+#define raw_closure_S_FS	(16+16+12)
+
+ALIGN 16
+PUBLIC ffi_closure_raw_SYSV
+ffi_closure_raw_SYSV PROC C
+L(UW32):
+	// cfi_startproc
+	sub 	esp, raw_closure_S_FS
+L(UW33):
+	// cfi_def_cfa_offset(raw_closure_S_FS + 4)
+	mov 	[esp+raw_closure_S_FS-4], ebx
+L(UW34):
+	// cfi_rel_offset(%ebx, raw_closure_S_FS-4)
+
+	mov 	edx, [eax+FFI_TRAMPOLINE_SIZE+8]	/* load cl->user_data */
+	mov 	[esp+12], edx
+	lea 	edx, [esp+raw_closure_S_FS+4]		/* load raw_args */
+	mov 	[esp+8], edx
+	lea 	edx, [esp+16]				/* load &res */
+	mov 	[esp+4], edx
+	mov 	ebx, [eax+FFI_TRAMPOLINE_SIZE]		/* load cl->cif */
+	mov 	[esp], ebx
+	call	DWORD PTR [eax+FFI_TRAMPOLINE_SIZE+4]		/* call cl->fun */
+
+	mov 	eax, [ebx+20]			/* load cif->flags */
+	and 	eax, X86_RET_TYPE_MASK
+// #ifdef __PIC__
+// 	call	__x86.get_pc_thunk.bx
+// L(pc4):
+// 	lea 	ecx, L(load_table4)-L(pc4)(%ebx, %eax, 8), %ecx
+// #else
+	lea 	ecx, [L(load_table4)+eax+8]
+// #endif
+	mov 	ebx, [esp+raw_closure_S_FS-4]
+L(UW35):
+	// cfi_restore(%ebx)
+	mov 	eax, [esp+16]				/* Optimistic load */
+	jmp	    dword ptr [ecx]
+
+	ALIGN 8
+L(load_table4):
+E(L(load_table4), X86_RET_FLOAT)
+	fld 	DWORD PTR [esp +16]
+	jmp	L(e4)
+E(L(load_table4), X86_RET_DOUBLE)
+	fld 	QWORD PTR [esp +16]
+	jmp	L(e4)
+E(L(load_table4), X86_RET_LDOUBLE)
+	fld 	QWORD PTR [esp +16]
+	jmp	L(e4)
+E(L(load_table4), X86_RET_SINT8)
+	movsx	eax, al
+	jmp	L(e4)
+E(L(load_table4), X86_RET_SINT16)
+	movsx	eax, ax
+	jmp	L(e4)
+E(L(load_table4), X86_RET_UINT8)
+	movzx	eax, al
+	jmp	L(e4)
+E(L(load_table4), X86_RET_UINT16)
+	movzx	eax, ax
+	jmp	L(e4)
+E(L(load_table4), X86_RET_INT64)
+	mov 	edx, [esp+16+4]
+	jmp	L(e4)
+E(L(load_table4), X86_RET_int 32)
+	nop
+	/* fallthru */
+E(L(load_table4), X86_RET_VOID)
+L(e4):
+	add 	esp, raw_closure_S_FS
+L(UW36):
+	// cfi_adjust_cfa_offset(-raw_closure_S_FS)
+	ret
+L(UW37):
+	// cfi_adjust_cfa_offset(raw_closure_S_FS)
+E(L(load_table4), X86_RET_STRUCTPOP)
+	add 	esp, raw_closure_S_FS
+L(UW38):
+	// cfi_adjust_cfa_offset(-raw_closure_S_FS)
+	ret	4
+L(UW39):
+	// cfi_adjust_cfa_offset(raw_closure_S_FS)
+E(L(load_table4), X86_RET_STRUCTARG)
+	jmp	L(e4)
+E(L(load_table4), X86_RET_STRUCT_1B)
+	movzx	eax, al
+	jmp	L(e4)
+E(L(load_table4), X86_RET_STRUCT_2B)
+	movzx	eax, ax
+	jmp	L(e4)
+
+	/* Fill out the table so that bad values are predictable.  */
+E(L(load_table4), X86_RET_UNUSED14)
+	int 3
+E(L(load_table4), X86_RET_UNUSED15)
+	int 3
+
+L(UW40):
+	// cfi_endproc
+ENDF(ffi_closure_raw_SYSV)
+
+#define raw_closure_T_FS	(16+16+8)
+
+ALIGN 16
+PUBLIC ffi_closure_raw_THISCALL
+ffi_closure_raw_THISCALL PROC C
+L(UW41):
+	// cfi_startproc
+	/* Rearrange the stack such that %ecx is the first argument.
+	   This means moving the return address.  */
+	pop 	edx
+L(UW42):
+	// cfi_def_cfa_offset(0)
+	// cfi_register(%eip, %edx)
+	push	ecx
+L(UW43):
+	// cfi_adjust_cfa_offset(4)
+	push 	edx
+L(UW44):
+	// cfi_adjust_cfa_offset(4)
+	// cfi_rel_offset(%eip, 0)
+	sub 	esp, raw_closure_T_FS
+L(UW45):
+	// cfi_adjust_cfa_offset(raw_closure_T_FS)
+	mov 	[esp+raw_closure_T_FS-4], ebx
+L(UW46):
+	// cfi_rel_offset(%ebx, raw_closure_T_FS-4)
+
+	mov 	edx, [eax+FFI_TRAMPOLINE_SIZE+8]	/* load cl->user_data */
+	mov 	[esp+12], edx
+	lea 	edx, [esp+raw_closure_T_FS+4]		/* load raw_args */
+	mov 	[esp+8], edx
+	lea 	edx, [esp+16]				/* load &res */
+	mov 	[esp+4], edx
+	mov 	ebx, [eax+FFI_TRAMPOLINE_SIZE]		/* load cl->cif */
+	mov 	[esp], ebx
+	call	DWORD PTR [eax+FFI_TRAMPOLINE_SIZE+4]		/* call cl->fun */
+
+	mov 	eax, [ebx+20]				/* load cif->flags */
+	and 	eax, X86_RET_TYPE_MASK
+// #ifdef __PIC__
+// 	call	__x86.get_pc_thunk.bx
+// L(pc5):
+// 	leal	L(load_table5)-L(pc5)(%ebx, %eax, 8), %ecx
+// #else
+	lea 	ecx, [L(load_table5)+eax*8]
+//#endif
+	mov 	ebx, [esp+raw_closure_T_FS-4]
+L(UW47):
+	// cfi_restore(%ebx)
+	mov 	eax, [esp+16]				/* Optimistic load */
+	jmp	    DWORD PTR [ecx]
+
+	AlIGN 4
+L(load_table5):
+E(L(load_table5), X86_RET_FLOAT)
+	fld	DWORD PTR [esp +16]
+	jmp	L(e5)
+E(L(load_table5), X86_RET_DOUBLE)
+	fld	QWORD PTR [esp +16]
+	jmp	L(e5)
+E(L(load_table5), X86_RET_LDOUBLE)
+	fld	QWORD PTR [esp+16]
+	jmp	L(e5)
+E(L(load_table5), X86_RET_SINT8)
+	movsx	eax, al
+	jmp	L(e5)
+E(L(load_table5), X86_RET_SINT16)
+	movsx	eax, ax
+	jmp	L(e5)
+E(L(load_table5), X86_RET_UINT8)
+	movzx	eax, al
+	jmp	L(e5)
+E(L(load_table5), X86_RET_UINT16)
+	movzx	eax, ax
+	jmp	L(e5)
+E(L(load_table5), X86_RET_INT64)
+	mov 	edx, [esp+16+4]
+	jmp	L(e5)
+E(L(load_table5), X86_RET_int 32)
+	nop
+	/* fallthru */
+E(L(load_table5), X86_RET_VOID)
+L(e5):
+	add 	esp, raw_closure_T_FS
+L(UW48):
+	// cfi_adjust_cfa_offset(-raw_closure_T_FS)
+	/* Remove the extra %ecx argument we pushed.  */
+	ret	4
+L(UW49):
+	// cfi_adjust_cfa_offset(raw_closure_T_FS)
+E(L(load_table5), X86_RET_STRUCTPOP)
+	add 	esp, raw_closure_T_FS
+L(UW50):
+	// cfi_adjust_cfa_offset(-raw_closure_T_FS)
+	ret	8
+L(UW51):
+	// cfi_adjust_cfa_offset(raw_closure_T_FS)
+E(L(load_table5), X86_RET_STRUCTARG)
+	jmp	L(e5)
+E(L(load_table5), X86_RET_STRUCT_1B)
+	movzx	eax, al
+	jmp	L(e5)
+E(L(load_table5), X86_RET_STRUCT_2B)
+	movzx	eax, ax
+	jmp	L(e5)
+
+	/* Fill out the table so that bad values are predictable.  */
+E(L(load_table5), X86_RET_UNUSED14)
+	int 3
+E(L(load_table5), X86_RET_UNUSED15)
+	int 3
+
+L(UW52):
+	// cfi_endproc
+ENDF(ffi_closure_raw_THISCALL)
+
+#endif /* !FFI_NO_RAW_API */
+
+#ifdef X86_DARWIN
+# define COMDAT(X)							\
+        .section __TEXT,__text,coalesced,pure_instructions;		\
+        .weak_definition X;						\
+        FFI_HIDDEN(X)
+#elif defined __ELF__ && !(defined(__sun__) && defined(__svr4__))
+# define COMDAT(X)							\
+	.section .text.X,"axG",@progbits,X,comdat;			\
+	PUBLIC	X;							\
+	FFI_HIDDEN(X)
+#else
+# define COMDAT(X)
+#endif
+
+// #if defined(__PIC__)
+// 	COMDAT(C(__x86.get_pc_thunk.bx))
+// C(__x86.get_pc_thunk.bx):
+// 	movl	(%esp), %ebx
+// 	ret
+// ENDF(C(__x86.get_pc_thunk.bx))
+// # if defined X86_DARWIN || defined HAVE_HIDDEN_VISIBILITY_ATTRIBUTE
+// 	COMDAT(C(__x86.get_pc_thunk.dx))
+// C(__x86.get_pc_thunk.dx):
+// 	movl	(%esp), %edx
+// 	ret
+// ENDF(C(__x86.get_pc_thunk.dx))
+// #endif /* DARWIN || HIDDEN */
+// #endif /* __PIC__ */
+
+#if 0
+/* Sadly, OSX cctools-as doesn't understand .cfi directives at all.  */
+
+#ifdef __APPLE__
+.section __TEXT,__eh_frame,coalesced,no_toc+strip_static_syms+live_support
+EHFrame0:
+#elif defined(X86_WIN32)
+.section .eh_frame,"r"
+#elif defined(HAVE_AS_X86_64_UNWIND_SECTION_TYPE)
+.section .eh_frame,EH_FRAME_FLAGS,@unwind
+#else
+.section .eh_frame,EH_FRAME_FLAGS,@progbits
+#endif
+
+#ifdef HAVE_AS_X86_PCREL
+# define PCREL(X)	X - .
+#else
+# define PCREL(X)	X@rel
+#endif
+
+/* Simplify advancing between labels.  Assume DW_CFA_advance_loc1 fits.  */
+#define ADV(N, P)	.byte 2, L(N)-L(P)
+
+	.balign 4
+L(CIE):
+	.set	L(set0),L(ECIE)-L(SCIE)
+	.long	L(set0)			/* CIE Length */
+L(SCIE):
+	.long	0			/* CIE Identifier Tag */
+	.byte	1			/* CIE Version */
+	.ascii	"zR\0"			/* CIE Augmentation */
+	.byte	1			/* CIE Code Alignment Factor */
+	.byte	0x7c			/* CIE Data Alignment Factor */
+	.byte	0x8			/* CIE RA Column */
+	.byte	1			/* Augmentation size */
+	.byte	0x1b			/* FDE Encoding (pcrel sdata4) */
+	.byte	0xc, 4, 4		/* DW_CFA_def_cfa, %esp offset 4 */
+	.byte	0x80+8, 1		/* DW_CFA_offset, %eip offset 1*-4 */
+	.balign 4
+L(ECIE):
+
+	.set	L(set1),L(EFDE1)-L(SFDE1)
+	.long	L(set1)			/* FDE Length */
+L(SFDE1):
+	.long	L(SFDE1)-L(CIE)		/* FDE CIE offset */
+	.long	PCREL(L(UW0))		/* Initial location */
+	.long	L(UW5)-L(UW0)		/* Address range */
+	.byte	0			/* Augmentation size */
+	ADV(UW1, UW0)
+	.byte	0xc, 5, 8		/* DW_CFA_def_cfa, %ebp 8 */
+	.byte	0x80+5, 2		/* DW_CFA_offset, %ebp 2*-4 */
+	ADV(UW2, UW1)
+	.byte	0x80+3, 0		/* DW_CFA_offset, %ebx 0*-4 */
+	ADV(UW3, UW2)
+	.byte	0xa			/* DW_CFA_remember_state */
+	.byte	0xc, 4, 4		/* DW_CFA_def_cfa, %esp 4 */
+	.byte	0xc0+3			/* DW_CFA_restore, %ebx */
+	.byte	0xc0+5			/* DW_CFA_restore, %ebp */
+	ADV(UW4, UW3)
+	.byte	0xb			/* DW_CFA_restore_state */
+	.balign	4
+L(EFDE1):
+
+	.set	L(set2),L(EFDE2)-L(SFDE2)
+	.long	L(set2)			/* FDE Length */
+L(SFDE2):
+	.long	L(SFDE2)-L(CIE)		/* FDE CIE offset */
+	.long	PCREL(L(UW6))		/* Initial location */
+	.long	L(UW8)-L(UW6)		/* Address range */
+	.byte	0			/* Augmentation size */
+	ADV(UW7, UW6)
+	.byte	0xe, closure_FS+4	/* DW_CFA_def_cfa_offset */
+	.balign	4
+L(EFDE2):
+
+	.set	L(set3),L(EFDE3)-L(SFDE3)
+	.long	L(set3)			/* FDE Length */
+L(SFDE3):
+	.long	L(SFDE3)-L(CIE)		/* FDE CIE offset */
+	.long	PCREL(L(UW9))		/* Initial location */
+	.long	L(UW11)-L(UW9)		/* Address range */
+	.byte	0			/* Augmentation size */
+	ADV(UW10, UW9)
+	.byte	0xe, closure_FS+4	/* DW_CFA_def_cfa_offset */
+	.balign	4
+L(EFDE3):
+
+	.set	L(set4),L(EFDE4)-L(SFDE4)
+	.long	L(set4)			/* FDE Length */
+L(SFDE4):
+	.long	L(SFDE4)-L(CIE)		/* FDE CIE offset */
+	.long	PCREL(L(UW12))		/* Initial location */
+	.long	L(UW20)-L(UW12)		/* Address range */
+	.byte	0			/* Augmentation size */
+	ADV(UW13, UW12)
+	.byte	0xe, closure_FS+4	/* DW_CFA_def_cfa_offset */
+#ifdef FFI_CLOSURE_CALL_INNER_SAVE_EBX
+	ADV(UW14, UW13)
+	.byte	0x80+3, (40-(closure_FS+4))/-4  /* DW_CFA_offset %ebx */
+	ADV(UW15, UW14)
+	.byte	0xc0+3			/* DW_CFA_restore %ebx */
+	ADV(UW16, UW15)
+#else
+	ADV(UW16, UW13)
+#endif
+	.byte	0xe, 4			/* DW_CFA_def_cfa_offset */
+	ADV(UW17, UW16)
+	.byte	0xe, closure_FS+4	/* DW_CFA_def_cfa_offset */
+	ADV(UW18, UW17)
+	.byte	0xe, 4			/* DW_CFA_def_cfa_offset */
+	ADV(UW19, UW18)
+	.byte	0xe, closure_FS+4	/* DW_CFA_def_cfa_offset */
+	.balign	4
+L(EFDE4):
+
+	.set	L(set5),L(EFDE5)-L(SFDE5)
+	.long	L(set5)			/* FDE Length */
+L(SFDE5):
+	.long	L(SFDE5)-L(CIE)		/* FDE CIE offset */
+	.long	PCREL(L(UW21))		/* Initial location */
+	.long	L(UW23)-L(UW21)		/* Address range */
+	.byte	0			/* Augmentation size */
+	ADV(UW22, UW21)
+	.byte	0xe, closure_FS+4	/* DW_CFA_def_cfa_offset */
+	.balign	4
+L(EFDE5):
+
+	.set	L(set6),L(EFDE6)-L(SFDE6)
+	.long	L(set6)			/* FDE Length */
+L(SFDE6):
+	.long	L(SFDE6)-L(CIE)		/* FDE CIE offset */
+	.long	PCREL(L(UW24))		/* Initial location */
+	.long	L(UW26)-L(UW24)		/* Address range */
+	.byte	0			/* Augmentation size */
+	.byte	0xe, 8			/* DW_CFA_def_cfa_offset */
+	.byte	0x80+8, 2		/* DW_CFA_offset %eip, 2*-4 */
+	ADV(UW25, UW24)
+	.byte	0xe, closure_FS+4	/* DW_CFA_def_cfa_offset */
+	.balign	4
+L(EFDE6):
+
+	.set	L(set7),L(EFDE7)-L(SFDE7)
+	.long	L(set7)			/* FDE Length */
+L(SFDE7):
+	.long	L(SFDE7)-L(CIE)		/* FDE CIE offset */
+	.long	PCREL(L(UW27))		/* Initial location */
+	.long	L(UW31)-L(UW27)		/* Address range */
+	.byte	0			/* Augmentation size */
+	ADV(UW28, UW27)
+	.byte	0xe, closure_FS+4	/* DW_CFA_def_cfa_offset */
+#ifdef FFI_CLOSURE_CALL_INNER_SAVE_EBX
+	ADV(UW29, UW28)
+	.byte	0x80+3, (40-(closure_FS+4))/-4  /* DW_CFA_offset %ebx */
+	ADV(UW30, UW29)
+	.byte	0xc0+3			/* DW_CFA_restore %ebx */
+#endif
+	.balign	4
+L(EFDE7):
+
+#if !FFI_NO_RAW_API
+	.set	L(set8),L(EFDE8)-L(SFDE8)
+	.long	L(set8)			/* FDE Length */
+L(SFDE8):
+	.long	L(SFDE8)-L(CIE)		/* FDE CIE offset */
+	.long	PCREL(L(UW32))		/* Initial location */
+	.long	L(UW40)-L(UW32)		/* Address range */
+	.byte	0			/* Augmentation size */
+	ADV(UW33, UW32)
+	.byte	0xe, raw_closure_S_FS+4	/* DW_CFA_def_cfa_offset */
+	ADV(UW34, UW33)
+	.byte	0x80+3, 2		/* DW_CFA_offset %ebx 2*-4 */
+	ADV(UW35, UW34)
+	.byte	0xc0+3			/* DW_CFA_restore %ebx */
+	ADV(UW36, UW35)
+	.byte	0xe, 4			/* DW_CFA_def_cfa_offset */
+	ADV(UW37, UW36)
+	.byte	0xe, raw_closure_S_FS+4	/* DW_CFA_def_cfa_offset */
+	ADV(UW38, UW37)
+	.byte	0xe, 4			/* DW_CFA_def_cfa_offset */
+	ADV(UW39, UW38)
+	.byte	0xe, raw_closure_S_FS+4	/* DW_CFA_def_cfa_offset */
+	.balign	4
+L(EFDE8):
+
+	.set	L(set9),L(EFDE9)-L(SFDE9)
+	.long	L(set9)			/* FDE Length */
+L(SFDE9):
+	.long	L(SFDE9)-L(CIE)		/* FDE CIE offset */
+	.long	PCREL(L(UW41))		/* Initial location */
+	.long	L(UW52)-L(UW41)		/* Address range */
+	.byte	0			/* Augmentation size */
+	ADV(UW42, UW41)
+	.byte	0xe, 0			/* DW_CFA_def_cfa_offset */
+	.byte	0x9, 8, 2		/* DW_CFA_register %eip, %edx */
+	ADV(UW43, UW42)
+	.byte	0xe, 4			/* DW_CFA_def_cfa_offset */
+	ADV(UW44, UW43)
+	.byte	0xe, 8			/* DW_CFA_def_cfa_offset */
+	.byte	0x80+8, 2		/* DW_CFA_offset %eip 2*-4 */
+	ADV(UW45, UW44)
+	.byte	0xe, raw_closure_T_FS+8	/* DW_CFA_def_cfa_offset */
+	ADV(UW46, UW45)
+	.byte	0x80+3, 3		/* DW_CFA_offset %ebx 3*-4 */
+	ADV(UW47, UW46)
+	.byte	0xc0+3			/* DW_CFA_restore %ebx */
+	ADV(UW48, UW47)
+	.byte	0xe, 8			/* DW_CFA_def_cfa_offset */
+	ADV(UW49, UW48)
+	.byte	0xe, raw_closure_T_FS+8	/* DW_CFA_def_cfa_offset */
+	ADV(UW50, UW49)
+	.byte	0xe, 8			/* DW_CFA_def_cfa_offset */
+	ADV(UW51, UW50)
+	.byte	0xe, raw_closure_T_FS+8	/* DW_CFA_def_cfa_offset */
+	.balign	4
+L(EFDE9):
+#endif /* !FFI_NO_RAW_API */
+
+#ifdef _WIN32
+	.def	 @feat.00;
+	.scl	3;
+	.type	0;
+	.endef
+	PUBLIC	@feat.00
+@feat.00 = 1
+#endif
+
+#endif /* ifndef _MSC_VER */
+#endif /* ifndef __x86_64__ */
+
+#if defined __ELF__ && defined __linux__
+	.section	.note.GNU-stack,"",@progbits
+#endif
+#endif
+
+END