Merge pull request #197 from foxsen/mips_go_closure

Mips go closure support

src/mips/ffi.c

@@ -581,14 +581,15 @@ ffi_status ffi_prep_cif_machdep(ffi_cif *cif)
 /* Low level routine for calling O32 functions */
 extern int ffi_call_O32(void (*)(char *, extended_cif *, int, int), 
 			extended_cif *, unsigned, 
-			unsigned, unsigned *, void (*)(void));
+			unsigned, unsigned *, void (*)(void), void *closure);
 
 /* Low level routine for calling N32 functions */
 extern int ffi_call_N32(void (*)(char *, extended_cif *, int, int), 
 			extended_cif *, unsigned, 
-			unsigned, void *, void (*)(void));
+			unsigned, void *, void (*)(void), void *closure);
 
-void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue)
+void ffi_call_int(ffi_cif *cif, void (*fn)(void), void *rvalue, 
+	      void **avalue, void *closure)
 {
   extended_cif ecif;
 
@@ -610,7 +611,7 @@ void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue)
     case FFI_O32:
     case FFI_O32_SOFT_FLOAT:
       ffi_call_O32(ffi_prep_args, &ecif, cif->bytes, 
-		   cif->flags, ecif.rvalue, fn);
+		   cif->flags, ecif.rvalue, fn, closure);
       break;
 #endif
 
@@ -642,7 +643,7 @@ void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue)
 #endif
 	  }
         ffi_call_N32(ffi_prep_args, &ecif, cif->bytes,
-                     cif->flags, rvalue_copy, fn);
+                     cif->flags, rvalue_copy, fn, closure);
         if (copy_rvalue)
           memcpy(ecif.rvalue, rvalue_copy + copy_offset, cif->rtype->size);
       }
@@ -655,11 +656,27 @@ void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue)
     }
 }
 
+void
+ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue)
+{
+  ffi_call_int (cif, fn, rvalue, avalue, NULL);
+}
+
+void
+ffi_call_go (ffi_cif *cif, void (*fn)(void), void *rvalue,
+	     void **avalue, void *closure)
+{
+  ffi_call_int (cif, fn, rvalue, avalue, closure);
+}
+
+
 #if FFI_CLOSURES
 #if defined(FFI_MIPS_O32)
 extern void ffi_closure_O32(void);
+extern void ffi_go_closure_O32(void);
 #else
 extern void ffi_closure_N32(void);
+extern void ffi_go_closure_N32(void);
 #endif /* FFI_MIPS_O32 */
 
 ffi_status
@@ -762,17 +779,17 @@ ffi_prep_closure_loc (ffi_closure *closure,
  * Based on the similar routine for sparc.
  */
 int
-ffi_closure_mips_inner_O32 (ffi_closure *closure,
+ffi_closure_mips_inner_O32 (ffi_cif *cif,
+                            void (*fun)(ffi_cif*, void*, void**, void*),
+			    void *user_data,
 			    void *rvalue, ffi_arg *ar,
 			    double *fpr)
 {
-  ffi_cif *cif;
   void **avaluep;
   ffi_arg *avalue;
   ffi_type **arg_types;
   int i, avn, argn, seen_int;
 
-  cif = closure->cif;
   avalue = alloca (cif->nargs * sizeof (ffi_arg));
   avaluep = alloca (cif->nargs * sizeof (ffi_arg));
 
@@ -840,7 +857,7 @@ ffi_closure_mips_inner_O32 (ffi_closure *closure,
     }
 
   /* Invoke the closure. */
-  (closure->fun) (cif, rvalue, avaluep, closure->user_data);
+  fun(cif, rvalue, avaluep, user_data);
 
   if (cif->abi == FFI_O32_SOFT_FLOAT)
     {
@@ -916,11 +933,12 @@ copy_struct_N32(char *target, unsigned offset, ffi_abi abi, ffi_type *type,
  *
  */
 int
-ffi_closure_mips_inner_N32 (ffi_closure *closure,
+ffi_closure_mips_inner_N32 (ffi_cif *cif, 
+			    void (*fun)(ffi_cif*, void*, void**, void*),
+                            void *user_data,
 			    void *rvalue, ffi_arg *ar,
 			    ffi_arg *fpr)
 {
-  ffi_cif *cif;
   void **avaluep;
   ffi_arg *avalue;
   ffi_type **arg_types;
@@ -928,7 +946,6 @@ ffi_closure_mips_inner_N32 (ffi_closure *closure,
   int soft_float;
   ffi_arg *argp;
 
-  cif = closure->cif;
   soft_float = cif->abi == FFI_N64_SOFT_FLOAT
     || cif->abi == FFI_N32_SOFT_FLOAT;
   avalue = alloca (cif->nargs * sizeof (ffi_arg));
@@ -1040,11 +1057,49 @@ ffi_closure_mips_inner_N32 (ffi_closure *closure,
     }
 
   /* Invoke the closure. */
-  (closure->fun) (cif, rvalue, avaluep, closure->user_data);
+  fun (cif, rvalue, avaluep, user_data);
 
   return cif->flags >> (FFI_FLAG_BITS * 8);
 }
 
 #endif /* FFI_MIPS_N32 */
 
+#if defined(FFI_MIPS_O32)
+extern void ffi_closure_O32(void);
+extern void ffi_go_closure_O32(void);
+#else
+extern void ffi_closure_N32(void);
+extern void ffi_go_closure_N32(void);
+#endif /* FFI_MIPS_O32 */
+
+ffi_status
+ffi_prep_go_closure (ffi_go_closure* closure, ffi_cif* cif,
+		     void (*fun)(ffi_cif*,void*,void**,void*))
+{
+  void * fn;
+
+#if defined(FFI_MIPS_O32)
+  if (cif->abi != FFI_O32 && cif->abi != FFI_O32_SOFT_FLOAT)
+    return FFI_BAD_ABI;
+  fn = ffi_go_closure_O32;
+#else
+#if _MIPS_SIM ==_ABIN32
+  if (cif->abi != FFI_N32
+      && cif->abi != FFI_N32_SOFT_FLOAT)
+    return FFI_BAD_ABI;
+#else
+  if (cif->abi != FFI_N64
+      && cif->abi != FFI_N64_SOFT_FLOAT)
+    return FFI_BAD_ABI;
+#endif
+  fn = ffi_go_closure_N32;
+#endif /* FFI_MIPS_O32 */
+
+  closure->tramp = (void *)fn;
+  closure->cif = cif;
+  closure->fun = fun;
+
+  return FFI_OK;
+}
+
 #endif /* FFI_CLOSURES */

src/mips/ffitarget.h

@@ -231,12 +231,14 @@ typedef enum ffi_abi {
 
 #if defined(FFI_MIPS_O32)
 #define FFI_CLOSURES 1
+#define FFI_GO_CLOSURES 1
 #define FFI_TRAMPOLINE_SIZE 20
 #else
 /* N32/N64. */
 # define FFI_CLOSURES 1
+#define FFI_GO_CLOSURES 1
 #if _MIPS_SIM==_ABI64
-#define FFI_TRAMPOLINE_SIZE 52
+#define FFI_TRAMPOLINE_SIZE 56
 #else
 #define FFI_TRAMPOLINE_SIZE 20
 #endif

src/mips/n32.S

@@ -37,8 +37,12 @@
 #define flags	 a3
 #define raddr    a4
 #define fn       a5
+#define closure  a6
 
-#define SIZEOF_FRAME	( 8 * FFI_SIZEOF_ARG )
+/* Note: to keep stack 16 byte aligned we need even number slots 
+   used 9 slots here
+*/
+#define SIZEOF_FRAME	( 10 * FFI_SIZEOF_ARG )
 
 #ifdef __GNUC__
 	.abicalls
@@ -49,24 +53,25 @@
 	.globl	ffi_call_N32
 	.ent	ffi_call_N32
 ffi_call_N32:	
-.LFB3:
+.LFB0:
 	.frame	$fp, SIZEOF_FRAME, ra
 	.mask	0xc0000000,-FFI_SIZEOF_ARG
 	.fmask	0x00000000,0
 
 	# Prologue
 	SUBU	$sp, SIZEOF_FRAME			# Frame size
-.LCFI0:
+.LCFI00:
 	REG_S	$fp, SIZEOF_FRAME - 2*FFI_SIZEOF_ARG($sp)	# Save frame pointer
 	REG_S	ra, SIZEOF_FRAME - 1*FFI_SIZEOF_ARG($sp)	# Save return address
-.LCFI1:
+.LCFI01:
 	move	$fp, $sp
-.LCFI3:
+.LCFI02:
 	move	t9, callback	# callback function pointer
 	REG_S	bytes, 2*FFI_SIZEOF_ARG($fp) # bytes
 	REG_S	flags, 3*FFI_SIZEOF_ARG($fp) # flags
 	REG_S	raddr, 4*FFI_SIZEOF_ARG($fp) # raddr
 	REG_S	fn,    5*FFI_SIZEOF_ARG($fp) # fn
+	REG_S	closure, 6*FFI_SIZEOF_ARG($fp) # closure
 
 	# Allocate at least 4 words in the argstack
 	move	v0, bytes
@@ -198,6 +203,9 @@ callit:
 	# Load the function pointer
 	REG_L	t9, 5*FFI_SIZEOF_ARG($fp)
 
+	# install the static chain(t7=$15)
+	REG_L	t7, 6*FFI_SIZEOF_ARG($fp)
+
 	# If the return value pointer is NULL, assume no return value.
 	REG_L	t5, 4*FFI_SIZEOF_ARG($fp)
 	beqz	t5, noretval
@@ -346,7 +354,7 @@ epilogue:
 	ADDU	$sp, SIZEOF_FRAME		      # Fix stack pointer
 	j	ra
 
-.LFE3:
+.LFE0:
 	.end	ffi_call_N32
 
 /* ffi_closure_N32. Expects address of the passed-in ffi_closure in t0
@@ -405,6 +413,41 @@ epilogue:
 #define RA_OFF2		(1  * FFI_SIZEOF_ARG)
 #define GP_OFF2		(0  * FFI_SIZEOF_ARG)
 
+	.align	2
+	.globl	ffi_go_closure_N32
+	.ent	ffi_go_closure_N32
+ffi_go_closure_N32:
+.LFB1:
+	.frame	$sp, SIZEOF_FRAME2, ra
+	.mask	0x90000000,-(SIZEOF_FRAME2 - RA_OFF2)
+	.fmask	0x00000000,0
+	SUBU	$sp, SIZEOF_FRAME2
+.LCFI10:
+	.cpsetup t9, GP_OFF2, ffi_go_closure_N32
+	REG_S	ra, RA_OFF2($sp)	# Save return address
+.LCFI11:
+
+	REG_S	a0, A0_OFF2($sp)
+	REG_S	a1, A1_OFF2($sp)
+	REG_S	a2, A2_OFF2($sp)
+	REG_S	a3, A3_OFF2($sp)
+	REG_S	a4, A4_OFF2($sp)
+	REG_S	a5, A5_OFF2($sp)
+
+	# Call ffi_closure_mips_inner_N32 to do the real work.
+	LA	t9, ffi_closure_mips_inner_N32
+	REG_L	a0, 8($15)   # cif
+	REG_L	a1, 16($15) # fun
+	move	a2, t7                     # userdata=closure
+	ADDU	a3, $sp, V0_OFF2           # rvalue
+	ADDU	a4, $sp, A0_OFF2           # ar
+	ADDU	a5, $sp, F12_OFF2          # fpr
+
+	b	$do_closure
+
+.LFE1:	
+	.end	ffi_go_closure_N32
+
 	.align	2
 	.globl	ffi_closure_N32
 	.ent	ffi_closure_N32
@@ -414,18 +457,29 @@ ffi_closure_N32:
 	.mask	0x90000000,-(SIZEOF_FRAME2 - RA_OFF2)
 	.fmask	0x00000000,0
 	SUBU	$sp, SIZEOF_FRAME2
-.LCFI5:
+.LCFI20:
 	.cpsetup t9, GP_OFF2, ffi_closure_N32
 	REG_S	ra, RA_OFF2($sp)	# Save return address
-.LCFI6:
+.LCFI21:
-	# Store all possible argument registers. If there are more than
-	# fit in registers, then they were stored on the stack.
 	REG_S	a0, A0_OFF2($sp)
 	REG_S	a1, A1_OFF2($sp)
 	REG_S	a2, A2_OFF2($sp)
 	REG_S	a3, A3_OFF2($sp)
 	REG_S	a4, A4_OFF2($sp)
 	REG_S	a5, A5_OFF2($sp)
+
+	# Call ffi_closure_mips_inner_N32 to do the real work.
+	LA	t9, ffi_closure_mips_inner_N32
+	REG_L	a0, 56($12)   # cif
+	REG_L	a1, 64($12)   # fun
+	REG_L	a2, 72($12) # user_data
+	ADDU	a3, $sp, V0_OFF2
+	ADDU	a4, $sp, A0_OFF2
+	ADDU	a5, $sp, F12_OFF2
+
+$do_closure:
+	# Store all possible argument registers. If there are more than
+	# fit in registers, then they were stored on the stack.
 	REG_S	a6, A6_OFF2($sp)
 	REG_S	a7, A7_OFF2($sp)
 
@@ -439,12 +493,6 @@ ffi_closure_N32:
 	s.d	$f18, F18_OFF2($sp)
 	s.d	$f19, F19_OFF2($sp)
 
-	# Call ffi_closure_mips_inner_N32 to do the real work.
-	LA	t9, ffi_closure_mips_inner_N32
-	move	a0, $12	 # Pointer to the ffi_closure
-	ADDU	a1, $sp, V0_OFF2
-	ADDU	a2, $sp, A0_OFF2
-	ADDU	a3, $sp, F12_OFF2
 	jalr	t9
 
 	# Return flags are in v0
@@ -531,46 +579,66 @@ cls_epilogue:
         .align  EH_FRAME_ALIGN
 .LECIE1:
 
-.LSFDE1:
+.LSFDE0:
-        .4byte  .LEFDE1-.LASFDE1	# length.
+        .4byte  .LEFDE0-.LASFDE0	# length.
-.LASFDE1:
+.LASFDE0:
-        .4byte  .LASFDE1-.Lframe1	# CIE_pointer.
+        .4byte  .LASFDE0-.Lframe1	# CIE_pointer.
-        FDE_ADDR_BYTES  .LFB3		# initial_location.
+        FDE_ADDR_BYTES  .LFB0		# initial_location.
-        FDE_ADDR_BYTES  .LFE3-.LFB3	# address_range.
+        FDE_ADDR_BYTES  .LFE0-.LFB0	# address_range.
         .byte   0x4			# DW_CFA_advance_loc4
-        .4byte  .LCFI0-.LFB3		# to .LCFI0
+        .4byte  .LCFI00-.LFB0		# to .LCFI00
         .byte   0xe			# DW_CFA_def_cfa_offset
         .uleb128 SIZEOF_FRAME		# adjust stack.by SIZEOF_FRAME
         .byte   0x4			# DW_CFA_advance_loc4
-        .4byte  .LCFI1-.LCFI0		# to .LCFI1
+        .4byte  .LCFI01-.LCFI00		# to .LCFI01
         .byte   0x9e			# DW_CFA_offset of $fp
         .uleb128 2*FFI_SIZEOF_ARG/4	# 
         .byte   0x9f			# DW_CFA_offset of ra
         .uleb128 1*FFI_SIZEOF_ARG/4	# 
         .byte   0x4			# DW_CFA_advance_loc4
-        .4byte  .LCFI3-.LCFI1		# to .LCFI3
+        .4byte  .LCFI02-.LCFI01		# to .LCFI02
         .byte   0xd			# DW_CFA_def_cfa_register
         .uleb128 0x1e			# in $fp
         .align  EH_FRAME_ALIGN
-.LEFDE1:
+.LEFDE0:
-.LSFDE3:
+
-	.4byte	.LEFDE3-.LASFDE3	# length
+.LSFDE1:
-.LASFDE3:
+	.4byte	.LEFDE1-.LASFDE1	# length
-	.4byte	.LASFDE3-.Lframe1	# CIE_pointer.
+.LASFDE1:
-	FDE_ADDR_BYTES	.LFB2		# initial_location.
+	.4byte	.LASFDE1-.Lframe1	# CIE_pointer.
-	FDE_ADDR_BYTES	.LFE2-.LFB2	# address_range.
+	FDE_ADDR_BYTES	.LFB1		# initial_location.
+	FDE_ADDR_BYTES	.LFE1-.LFB1	# address_range.
 	.byte	0x4			# DW_CFA_advance_loc4
-	.4byte	.LCFI5-.LFB2		# to .LCFI5
+	.4byte	.LCFI10-.LFB1		# to .LCFI10
 	.byte	0xe			# DW_CFA_def_cfa_offset
 	.uleb128 SIZEOF_FRAME2		# adjust stack.by SIZEOF_FRAME
 	.byte	0x4			# DW_CFA_advance_loc4
-	.4byte	.LCFI6-.LCFI5		# to .LCFI6
+	.4byte	.LCFI11-.LCFI10		# to .LCFI11
 	.byte	0x9c			# DW_CFA_offset of $gp ($28)
 	.uleb128 (SIZEOF_FRAME2 - GP_OFF2)/4
 	.byte	0x9f			# DW_CFA_offset of ra ($31)
 	.uleb128 (SIZEOF_FRAME2 - RA_OFF2)/4
 	.align	EH_FRAME_ALIGN
-.LEFDE3:
+.LEFDE1:
+
+.LSFDE2:
+	.4byte	.LEFDE2-.LASFDE2	# length
+.LASFDE2:
+	.4byte	.LASFDE2-.Lframe1	# CIE_pointer.
+	FDE_ADDR_BYTES	.LFB2		# initial_location.
+	FDE_ADDR_BYTES	.LFE2-.LFB2	# address_range.
+	.byte	0x4			# DW_CFA_advance_loc4
+	.4byte	.LCFI20-.LFB2		# to .LCFI20
+	.byte	0xe			# DW_CFA_def_cfa_offset
+	.uleb128 SIZEOF_FRAME2		# adjust stack.by SIZEOF_FRAME
+	.byte	0x4			# DW_CFA_advance_loc4
+	.4byte	.LCFI21-.LCFI20		# to .LCFI21
+	.byte	0x9c			# DW_CFA_offset of $gp ($28)
+	.uleb128 (SIZEOF_FRAME2 - GP_OFF2)/4
+	.byte	0x9f			# DW_CFA_offset of ra ($31)
+	.uleb128 (SIZEOF_FRAME2 - RA_OFF2)/4
+	.align	EH_FRAME_ALIGN
+.LEFDE2:
 #endif /* __GNUC__ */	
 	
 #endif

src/mips/o32.S

@@ -50,14 +50,14 @@ ffi_call_O32:
 $LFB0:
 	# Prologue
 	SUBU	$sp, SIZEOF_FRAME	# Frame size
-$LCFI0:
+$LCFI00:
 	REG_S	$fp, FP_OFF($sp)	# Save frame pointer
-$LCFI1:
+$LCFI01:
 	REG_S	ra, RA_OFF($sp)		# Save return address
-$LCFI2:
+$LCFI02:
 	move	$fp, $sp
 
-$LCFI3:
+$LCFI03:
 	move	t9, callback		# callback function pointer
 	REG_S	flags, A3_OFF($fp)	# flags
 
@@ -132,6 +132,9 @@ pass_f_d:
 	l.d	$f14, 2*FFI_SIZEOF_ARG($sp)	# passing double and float
 
 call_it:	
+	# Load the static chain pointer
+	REG_L	t7, SIZEOF_FRAME + 6*FFI_SIZEOF_ARG($fp)
+
 	# Load the function pointer
 	REG_L	t9, SIZEOF_FRAME + 5*FFI_SIZEOF_ARG($fp)
 
@@ -204,13 +207,15 @@ $LFE0:
 	-8 - f14 (le low, be high)
 	-9 - f12 (le high, be low)
        -10 - f12 (le low, be high)
-       -11 - Called function a3 save
+       -11 - Called function a5 save
-       -12 - Called function a2 save
+       -12 - Called function a4 save
-       -13 - Called function a1 save
+       -13 - Called function a3 save
-       -14 - Called function a0 save, our sp and fp point here
+       -14 - Called function a2 save
+       -15 - Called function a1 save
+       -16 - Called function a0 save, our sp and fp point here
 	 */
 	
-#define SIZEOF_FRAME2	(14 * FFI_SIZEOF_ARG)
+#define SIZEOF_FRAME2	(16 * FFI_SIZEOF_ARG)
 #define A3_OFF2		(SIZEOF_FRAME2 + 3 * FFI_SIZEOF_ARG)
 #define A2_OFF2		(SIZEOF_FRAME2 + 2 * FFI_SIZEOF_ARG)
 #define A1_OFF2		(SIZEOF_FRAME2 + 1 * FFI_SIZEOF_ARG)
@@ -225,12 +230,15 @@ $LFE0:
 #define FA_1_0_OFF2	(SIZEOF_FRAME2 - 8 * FFI_SIZEOF_ARG)
 #define FA_0_1_OFF2	(SIZEOF_FRAME2 - 9 * FFI_SIZEOF_ARG)
 #define FA_0_0_OFF2	(SIZEOF_FRAME2 - 10 * FFI_SIZEOF_ARG)
+#define CALLED_A5_OFF2  (SIZEOF_FRAME2 - 11 * FFI_SIZEOF_ARG)
+#define CALLED_A4_OFF2  (SIZEOF_FRAME2 - 12 * FFI_SIZEOF_ARG)
 
 	.text
+
 	.align	2
-	.globl	ffi_closure_O32
+	.globl	ffi_go_closure_O32
-	.ent	ffi_closure_O32
+	.ent	ffi_go_closure_O32
-ffi_closure_O32:
+ffi_go_closure_O32:
 $LFB1:
 	# Prologue
 	.frame	$fp, SIZEOF_FRAME2, ra
@@ -239,14 +247,69 @@ $LFB1:
 	.set	reorder
 	SUBU	$sp, SIZEOF_FRAME2
 	.cprestore GP_OFF2
-$LCFI4:
+$LCFI10:
+
 	REG_S	$16, S0_OFF2($sp)	 # Save s0
 	REG_S	$fp, FP_OFF2($sp)	 # Save frame pointer
 	REG_S	ra, RA_OFF2($sp)	 # Save return address
-$LCFI6:
+$LCFI11:
+
+	move	$fp, $sp
+$LCFI12:
+
+	REG_S	a0, A0_OFF2($fp)
+	REG_S	a1, A1_OFF2($fp)
+	REG_S	a2, A2_OFF2($fp)
+	REG_S	a3, A3_OFF2($fp)
+
+	# Load ABI enum to s0
+	REG_L	$16, 4($15)	# cif 
+	REG_L	$16, 0($16)	# abi is first member.
+
+	li	$13, 1		# FFI_O32
+	bne	$16, $13, 1f	# Skip fp save if FFI_O32_SOFT_FLOAT
+	
+	# Store all possible float/double registers.
+	s.d	$f12, FA_0_0_OFF2($fp)
+	s.d	$f14, FA_1_0_OFF2($fp)
+1:
+	# prepare arguments for ffi_closure_mips_inner_O32
+	REG_L	a0, 4($15)	 # cif 
+	REG_L	a1, 8($15)	 # fun
+	move	a2, $15		 # user_data = go closure
+	addu	a3, $fp, V0_OFF2 # rvalue
+
+	addu	t9, $fp, A0_OFF2 # ar
+	REG_S   t9, CALLED_A4_OFF2($fp)
+
+	addu	t9, $fp, FA_0_0_OFF2 #fpr
+	REG_S   t9, CALLED_A5_OFF2($fp)
+
+	b $do_closure
+
+$LFE1:
+	.end ffi_go_closure_O32
+
+	.align	2
+	.globl	ffi_closure_O32
+	.ent	ffi_closure_O32
+ffi_closure_O32:
+$LFB2:
+	# Prologue
+	.frame	$fp, SIZEOF_FRAME2, ra
+	.set	noreorder
+	.cpload	t9
+	.set	reorder
+	SUBU	$sp, SIZEOF_FRAME2
+	.cprestore GP_OFF2
+$LCFI20:
+	REG_S	$16, S0_OFF2($sp)	 # Save s0
+	REG_S	$fp, FP_OFF2($sp)	 # Save frame pointer
+	REG_S	ra, RA_OFF2($sp)	 # Save return address
+$LCFI21:
 	move	$fp, $sp
 
-$LCFI7:
+$LCFI22:
 	# Store all possible argument registers. If there are more than
 	# four arguments, then they are stored above where we put a3.
 	REG_S	a0, A0_OFF2($fp)
@@ -265,12 +328,21 @@ $LCFI7:
 	s.d	$f12, FA_0_0_OFF2($fp)
 	s.d	$f14, FA_1_0_OFF2($fp)
 1:	
-	# Call ffi_closure_mips_inner_O32 to do the work.
+	# prepare arguments for ffi_closure_mips_inner_O32
+	REG_L	a0, 20($12)	 # cif pointer follows tramp.
+	REG_L	a1, 24($12)	 # fun
+	REG_L	a2, 28($12)	 # user_data
+	addu	a3, $fp, V0_OFF2 # rvalue
+
+	addu	t9, $fp, A0_OFF2 # ar
+	REG_S   t9, CALLED_A4_OFF2($fp)
+
+	addu	t9, $fp, FA_0_0_OFF2 #fpr
+	REG_S   t9, CALLED_A5_OFF2($fp)
+
+$do_closure:
 	la	t9, ffi_closure_mips_inner_O32
-	move	a0, $12	 # Pointer to the ffi_closure
+	# Call ffi_closure_mips_inner_O32 to do the work.
-	addu	a1, $fp, V0_OFF2
-	addu	a2, $fp, A0_OFF2
-	addu	a3, $fp, FA_0_0_OFF2
 	jalr	t9
 
 	# Load the return value into the appropriate register.
@@ -300,7 +372,7 @@ closure_done:
 	REG_L	ra,  RA_OFF2($sp)	 # Restore return address
 	ADDU	$sp, SIZEOF_FRAME2
 	j	ra
-$LFE1:
+$LFE2:
 	.end	ffi_closure_O32
 
 /* DWARF-2 unwind info. */
@@ -322,6 +394,7 @@ $LSCIE0:
 	.uleb128 0x0
 	.align	2
 $LECIE0:
+
 $LSFDE0:
 	.4byte	$LEFDE0-$LASFDE0	 # FDE Length
 $LASFDE0:
@@ -330,11 +403,11 @@ $LASFDE0:
 	.4byte	$LFE0-$LFB0	 # FDE address range
 	.uleb128 0x0	 # Augmentation size
 	.byte	0x4	 # DW_CFA_advance_loc4
-	.4byte	$LCFI0-$LFB0
+	.4byte	$LCFI00-$LFB0
 	.byte	0xe	 # DW_CFA_def_cfa_offset
 	.uleb128 0x18
 	.byte	0x4	 # DW_CFA_advance_loc4
-	.4byte	$LCFI2-$LCFI0
+	.4byte	$LCFI01-$LCFI00
 	.byte	0x11	 # DW_CFA_offset_extended_sf
 	.uleb128 0x1e	 # $fp
 	.sleb128 -2	 # SIZEOF_FRAME2 - 2*FFI_SIZEOF_ARG($sp)
@@ -342,12 +415,13 @@ $LASFDE0:
 	.uleb128 0x1f	 # $ra
 	.sleb128 -1	 # SIZEOF_FRAME2 - 1*FFI_SIZEOF_ARG($sp)
 	.byte	0x4	 # DW_CFA_advance_loc4
-	.4byte	$LCFI3-$LCFI2
+	.4byte	$LCFI02-$LCFI01
 	.byte	0xc	 # DW_CFA_def_cfa
 	.uleb128 0x1e
 	.uleb128 0x18
 	.align	2
 $LEFDE0:
+
 $LSFDE1:
 	.4byte	$LEFDE1-$LASFDE1	 # FDE Length
 $LASFDE1:
@@ -356,11 +430,11 @@ $LASFDE1:
 	.4byte	$LFE1-$LFB1	 # FDE address range
 	.uleb128 0x0	 # Augmentation size
 	.byte	0x4	 # DW_CFA_advance_loc4
-	.4byte	$LCFI4-$LFB1
+	.4byte	$LCFI10-$LFB1
 	.byte	0xe	 # DW_CFA_def_cfa_offset
-	.uleb128 0x38
+	.uleb128 SIZEOF_FRAME2
 	.byte	0x4	 # DW_CFA_advance_loc4
-	.4byte	$LCFI6-$LCFI4
+	.4byte	$LCFI11-$LCFI10
 	.byte	0x11	 # DW_CFA_offset_extended_sf
 	.uleb128 0x10	 # $16
 	.sleb128 -3	 # SIZEOF_FRAME2 - 3*FFI_SIZEOF_ARG($sp)
@@ -371,11 +445,41 @@ $LASFDE1:
 	.uleb128 0x1f	 # $ra
 	.sleb128 -1	 # SIZEOF_FRAME2 - 1*FFI_SIZEOF_ARG($sp)
 	.byte	0x4	 # DW_CFA_advance_loc4
-	.4byte	$LCFI7-$LCFI6
+	.4byte	$LCFI12-$LCFI11
 	.byte	0xc	 # DW_CFA_def_cfa
 	.uleb128 0x1e
-	.uleb128 0x38
+	.uleb128 SIZEOF_FRAME2
 	.align	2
 $LEFDE1:
 
+$LSFDE2:
+	.4byte	$LEFDE2-$LASFDE2	 # FDE Length
+$LASFDE2:
+	.4byte	$LASFDE2-$Lframe0	 # FDE CIE offset
+	.4byte	$LFB2	 # FDE initial location
+	.4byte	$LFE2-$LFB2	 # FDE address range
+	.uleb128 0x0	 # Augmentation size
+	.byte	0x4	 # DW_CFA_advance_loc4
+	.4byte	$LCFI20-$LFB2
+	.byte	0xe	 # DW_CFA_def_cfa_offset
+	.uleb128 SIZEOF_FRAME2
+	.byte	0x4	 # DW_CFA_advance_loc4
+	.4byte	$LCFI21-$LCFI20
+	.byte	0x11	 # DW_CFA_offset_extended_sf
+	.uleb128 0x10	 # $16
+	.sleb128 -3	 # SIZEOF_FRAME2 - 3*FFI_SIZEOF_ARG($sp)
+	.byte	0x11	 # DW_CFA_offset_extended_sf
+	.uleb128 0x1e	 # $fp
+	.sleb128 -2	 # SIZEOF_FRAME2 - 2*FFI_SIZEOF_ARG($sp)
+	.byte	0x11	 # DW_CFA_offset_extended_sf
+	.uleb128 0x1f	 # $ra
+	.sleb128 -1	 # SIZEOF_FRAME2 - 1*FFI_SIZEOF_ARG($sp)
+	.byte	0x4	 # DW_CFA_advance_loc4
+	.4byte	$LCFI22-$LCFI21
+	.byte	0xc	 # DW_CFA_def_cfa
+	.uleb128 0x1e
+	.uleb128 SIZEOF_FRAME2
+	.align	2
+$LEFDE2:
+
 #endif