1
0
MozillaThunderbird/libffi-ppc64le.patch

3751 lines
105 KiB
Diff

Index: mozilla/js/src/ctypes/libffi/src/powerpc/aix.S
===================================================================
--- mozilla.orig/js/src/ctypes/libffi/src/powerpc/aix.S
+++ mozilla/js/src/ctypes/libffi/src/powerpc/aix.S
@@ -1,5 +1,5 @@
/* -----------------------------------------------------------------------
- aix.S - Copyright (c) 2002,2009 Free Software Foundation, Inc.
+ aix.S - Copyright (c) 2002, 2009 Free Software Foundation, Inc.
based on darwin.S by John Hornkvist
PowerPC Assembly glue.
@@ -79,6 +79,8 @@
.set f20,20
.set f21,21
+ .extern .ffi_prep_args
+
#define LIBFFI_ASM
#include <fficonfig.h>
#include <ffi.h>
@@ -125,6 +127,7 @@ ffi_call_AIX:
/* Call ffi_prep_args. */
mr r4, r1
bl .ffi_prep_args
+ nop
/* Now do the call. */
ld r0, 0(r29)
@@ -226,6 +229,7 @@ L(float_return_value):
/* Call ffi_prep_args. */
mr r4, r1
bl .ffi_prep_args
+ nop
/* Now do the call. */
lwz r0, 0(r29)
Index: mozilla/js/src/ctypes/libffi/src/powerpc/ffi.c
===================================================================
--- mozilla.orig/js/src/ctypes/libffi/src/powerpc/ffi.c
+++ mozilla/js/src/ctypes/libffi/src/powerpc/ffi.c
@@ -1,7 +1,9 @@
/* -----------------------------------------------------------------------
- ffi.c - Copyright (c) 1998 Geoffrey Keating
- Copyright (C) 2007, 2008 Free Software Foundation, Inc
- Copyright (C) 2008 Red Hat, Inc
+ ffi.c - Copyright (C) 2011 Anthony Green
+ Copyright (C) 2011 Kyle Moffett
+ Copyright (C) 2008 Red Hat, Inc
+ Copyright (C) 2007, 2008 Free Software Foundation, Inc
+ Copyright (c) 1998 Geoffrey Keating
PowerPC Foreign Function Interface
@@ -39,32 +41,29 @@ enum {
/* The assembly depends on these exact flags. */
FLAG_RETURNS_SMST = 1 << (31-31), /* Used for FFI_SYSV small structs. */
FLAG_RETURNS_NOTHING = 1 << (31-30), /* These go in cr7 */
+#ifndef __NO_FPRS__
FLAG_RETURNS_FP = 1 << (31-29),
+#endif
FLAG_RETURNS_64BITS = 1 << (31-28),
FLAG_RETURNS_128BITS = 1 << (31-27), /* cr6 */
- FLAG_SYSV_SMST_R4 = 1 << (31-26), /* use r4 for FFI_SYSV 8 byte
- structs. */
- FLAG_SYSV_SMST_R3 = 1 << (31-25), /* use r3 for FFI_SYSV 4 byte
- structs. */
- /* Bits (31-24) through (31-19) store shift value for SMST */
FLAG_ARG_NEEDS_COPY = 1 << (31- 7),
+ FLAG_ARG_NEEDS_PSAVE = FLAG_ARG_NEEDS_COPY, /* Used by ELFv2 */
+#ifndef __NO_FPRS__
FLAG_FP_ARGUMENTS = 1 << (31- 6), /* cr1.eq; specified by ABI */
+#endif
FLAG_4_GPR_ARGUMENTS = 1 << (31- 5),
FLAG_RETVAL_REFERENCE = 1 << (31- 4)
};
/* About the SYSV ABI. */
-unsigned int NUM_GPR_ARG_REGISTERS = 8;
+#define ASM_NEEDS_REGISTERS 4
+#define NUM_GPR_ARG_REGISTERS 8
#ifndef __NO_FPRS__
-unsigned int NUM_FPR_ARG_REGISTERS = 8;
-#else
-unsigned int NUM_FPR_ARG_REGISTERS = 0;
+# define NUM_FPR_ARG_REGISTERS 8
#endif
-enum { ASM_NEEDS_REGISTERS = 4 };
-
/* ffi_prep_args_SYSV is called by the assembly routine once stack space
has been allocated for the function's arguments.
@@ -113,10 +112,12 @@ ffi_prep_args_SYSV (extended_cif *ecif,
valp gpr_base;
int intarg_count;
+#ifndef __NO_FPRS__
/* 'fpr_base' points at the space for fpr1, and grows upwards as
we use FPR registers. */
valp fpr_base;
int fparg_count;
+#endif
/* 'copy_space' grows down as we put structures in it. It should
stay 16-byte aligned. */
@@ -125,9 +126,11 @@ ffi_prep_args_SYSV (extended_cif *ecif,
/* 'next_arg' grows up as we put parameters in it. */
valp next_arg;
- int i, ii MAYBE_UNUSED;
+ int i;
ffi_type **ptr;
+#ifndef __NO_FPRS__
double double_tmp;
+#endif
union {
void **v;
char **c;
@@ -143,21 +146,22 @@ ffi_prep_args_SYSV (extended_cif *ecif,
size_t struct_copy_size;
unsigned gprvalue;
- if (ecif->cif->abi == FFI_LINUX_SOFT_FLOAT)
- NUM_FPR_ARG_REGISTERS = 0;
-
stacktop.c = (char *) stack + bytes;
gpr_base.u = stacktop.u - ASM_NEEDS_REGISTERS - NUM_GPR_ARG_REGISTERS;
intarg_count = 0;
+#ifndef __NO_FPRS__
fpr_base.d = gpr_base.d - NUM_FPR_ARG_REGISTERS;
fparg_count = 0;
copy_space.c = ((flags & FLAG_FP_ARGUMENTS) ? fpr_base.c : gpr_base.c);
+#else
+ copy_space.c = gpr_base.c;
+#endif
next_arg.u = stack + 2;
/* Check that everything starts aligned properly. */
- FFI_ASSERT (((unsigned) (char *) stack & 0xF) == 0);
- FFI_ASSERT (((unsigned) copy_space.c & 0xF) == 0);
- FFI_ASSERT (((unsigned) stacktop.c & 0xF) == 0);
+ FFI_ASSERT (((unsigned long) (char *) stack & 0xF) == 0);
+ FFI_ASSERT (((unsigned long) copy_space.c & 0xF) == 0);
+ FFI_ASSERT (((unsigned long) stacktop.c & 0xF) == 0);
FFI_ASSERT ((bytes & 0xF) == 0);
FFI_ASSERT (copy_space.c >= next_arg.c);
@@ -174,12 +178,28 @@ ffi_prep_args_SYSV (extended_cif *ecif,
i > 0;
i--, ptr++, p_argv.v++)
{
- switch ((*ptr)->type)
- {
+ unsigned short typenum = (*ptr)->type;
+
+ /* We may need to handle some values depending on ABI */
+ if (ecif->cif->abi == FFI_LINUX_SOFT_FLOAT) {
+ if (typenum == FFI_TYPE_FLOAT)
+ typenum = FFI_TYPE_UINT32;
+ if (typenum == FFI_TYPE_DOUBLE)
+ typenum = FFI_TYPE_UINT64;
+ if (typenum == FFI_TYPE_LONGDOUBLE)
+ typenum = FFI_TYPE_UINT128;
+ } else if (ecif->cif->abi != FFI_LINUX) {
+#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
+ if (typenum == FFI_TYPE_LONGDOUBLE)
+ typenum = FFI_TYPE_STRUCT;
+#endif
+ }
+
+ /* Now test the translated value */
+ switch (typenum) {
+#ifndef __NO_FPRS__
case FFI_TYPE_FLOAT:
/* With FFI_LINUX_SOFT_FLOAT floats are handled like UINT32. */
- if (ecif->cif->abi == FFI_LINUX_SOFT_FLOAT)
- goto soft_float_prep;
double_tmp = **p_argv.f;
if (fparg_count >= NUM_FPR_ARG_REGISTERS)
{
@@ -195,8 +215,6 @@ ffi_prep_args_SYSV (extended_cif *ecif,
case FFI_TYPE_DOUBLE:
/* With FFI_LINUX_SOFT_FLOAT doubles are handled like UINT64. */
- if (ecif->cif->abi == FFI_LINUX_SOFT_FLOAT)
- goto soft_double_prep;
double_tmp = **p_argv.d;
if (fparg_count >= NUM_FPR_ARG_REGISTERS)
@@ -218,43 +236,6 @@ ffi_prep_args_SYSV (extended_cif *ecif,
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
- if ((ecif->cif->abi != FFI_LINUX)
- && (ecif->cif->abi != FFI_LINUX_SOFT_FLOAT))
- goto do_struct;
- /* The soft float ABI for long doubles works like this,
- a long double is passed in four consecutive gprs if available.
- A maximum of 2 long doubles can be passed in gprs.
- If we do not have 4 gprs left, the long double is passed on the
- stack, 4-byte aligned. */
- if (ecif->cif->abi == FFI_LINUX_SOFT_FLOAT)
- {
- unsigned int int_tmp = (*p_argv.ui)[0];
- if (intarg_count >= NUM_GPR_ARG_REGISTERS - 3)
- {
- if (intarg_count < NUM_GPR_ARG_REGISTERS)
- intarg_count += NUM_GPR_ARG_REGISTERS - intarg_count;
- *next_arg.u = int_tmp;
- next_arg.u++;
- for (ii = 1; ii < 4; ii++)
- {
- int_tmp = (*p_argv.ui)[ii];
- *next_arg.u = int_tmp;
- next_arg.u++;
- }
- }
- else
- {
- *gpr_base.u++ = int_tmp;
- for (ii = 1; ii < 4; ii++)
- {
- int_tmp = (*p_argv.ui)[ii];
- *gpr_base.u++ = int_tmp;
- }
- }
- intarg_count +=4;
- }
- else
- {
double_tmp = (*p_argv.d)[0];
if (fparg_count >= NUM_FPR_ARG_REGISTERS - 1)
@@ -280,13 +261,40 @@ ffi_prep_args_SYSV (extended_cif *ecif,
fparg_count += 2;
FFI_ASSERT (flags & FLAG_FP_ARGUMENTS);
- }
break;
#endif
+#endif /* have FPRs */
+
+ /*
+ * The soft float ABI for long doubles works like this, a long double
+ * is passed in four consecutive GPRs if available. A maximum of 2
+ * long doubles can be passed in gprs. If we do not have 4 GPRs
+ * left, the long double is passed on the stack, 4-byte aligned.
+ */
+ case FFI_TYPE_UINT128: {
+ unsigned int int_tmp = (*p_argv.ui)[0];
+ unsigned int ii;
+ if (intarg_count >= NUM_GPR_ARG_REGISTERS - 3) {
+ if (intarg_count < NUM_GPR_ARG_REGISTERS)
+ intarg_count += NUM_GPR_ARG_REGISTERS - intarg_count;
+ *(next_arg.u++) = int_tmp;
+ for (ii = 1; ii < 4; ii++) {
+ int_tmp = (*p_argv.ui)[ii];
+ *(next_arg.u++) = int_tmp;
+ }
+ } else {
+ *(gpr_base.u++) = int_tmp;
+ for (ii = 1; ii < 4; ii++) {
+ int_tmp = (*p_argv.ui)[ii];
+ *(gpr_base.u++) = int_tmp;
+ }
+ }
+ intarg_count += 4;
+ break;
+ }
case FFI_TYPE_UINT64:
case FFI_TYPE_SINT64:
- soft_double_prep:
if (intarg_count == NUM_GPR_ARG_REGISTERS-1)
intarg_count++;
if (intarg_count >= NUM_GPR_ARG_REGISTERS)
@@ -319,9 +327,6 @@ ffi_prep_args_SYSV (extended_cif *ecif,
break;
case FFI_TYPE_STRUCT:
-#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
- do_struct:
-#endif
struct_copy_size = ((*ptr)->size + 15) & ~0xF;
copy_space.c -= struct_copy_size;
memcpy (copy_space.c, *p_argv.c, (*ptr)->size);
@@ -349,7 +354,6 @@ ffi_prep_args_SYSV (extended_cif *ecif,
case FFI_TYPE_UINT32:
case FFI_TYPE_SINT32:
case FFI_TYPE_POINTER:
- soft_float_prep:
gprvalue = **p_argv.ui;
@@ -366,8 +370,16 @@ ffi_prep_args_SYSV (extended_cif *ecif,
/* Check that we didn't overrun the stack... */
FFI_ASSERT (copy_space.c >= next_arg.c);
FFI_ASSERT (gpr_base.u <= stacktop.u - ASM_NEEDS_REGISTERS);
+ /* The assert below is testing that the number of integer arguments agrees
+ with the number found in ffi_prep_cif_machdep(). However, intarg_count
+ is incremented whenever we place an FP arg on the stack, so account for
+ that before our assert test. */
+#ifndef __NO_FPRS__
+ if (fparg_count > NUM_FPR_ARG_REGISTERS)
+ intarg_count -= fparg_count - NUM_FPR_ARG_REGISTERS;
FFI_ASSERT (fpr_base.u
<= stacktop.u - ASM_NEEDS_REGISTERS - NUM_GPR_ARG_REGISTERS);
+#endif
FFI_ASSERT (flags & FLAG_4_GPR_ARGUMENTS || intarg_count <= 4);
}
@@ -378,6 +390,45 @@ enum {
};
enum { ASM_NEEDS_REGISTERS64 = 4 };
+#if _CALL_ELF == 2
+static unsigned int
+discover_homogeneous_aggregate (const ffi_type *t, unsigned int *elnum)
+{
+ switch (t->type)
+ {
+ case FFI_TYPE_FLOAT:
+ case FFI_TYPE_DOUBLE:
+ *elnum = 1;
+ return (int) t->type;
+
+ case FFI_TYPE_STRUCT:;
+ {
+ unsigned int base_elt = 0, total_elnum = 0;
+ ffi_type **el = t->elements;
+ while (*el)
+ {
+ unsigned int el_elt, el_elnum = 0;
+ el_elt = discover_homogeneous_aggregate (*el, &el_elnum);
+ if (el_elt == 0
+ || (base_elt && base_elt != el_elt))
+ return 0;
+ base_elt = el_elt;
+ total_elnum += el_elnum;
+ if (total_elnum > 8)
+ return 0;
+ el++;
+ }
+ *elnum = total_elnum;
+ return base_elt;
+ }
+
+ default:
+ return 0;
+ }
+}
+#endif
+
+
/* ffi_prep_args64 is called by the assembly routine once stack space
has been allocated for the function's arguments.
@@ -423,6 +474,7 @@ ffi_prep_args64 (extended_cif *ecif, uns
unsigned long *ul;
float *f;
double *d;
+ size_t p;
} valp;
/* 'stacktop' points at the previous backchain pointer. */
@@ -438,9 +490,9 @@ ffi_prep_args64 (extended_cif *ecif, uns
/* 'fpr_base' points at the space for fpr3, and grows upwards as
we use FPR registers. */
valp fpr_base;
- int fparg_count;
+ unsigned int fparg_count;
- int i, words;
+ unsigned int i, words, nargs, nfixedargs;
ffi_type **ptr;
double double_tmp;
union {
@@ -457,11 +509,18 @@ ffi_prep_args64 (extended_cif *ecif, uns
double **d;
} p_argv;
unsigned long gprvalue;
+#ifdef __STRUCT_PARM_ALIGN__
+ unsigned long align;
+#endif
stacktop.c = (char *) stack + bytes;
gpr_base.ul = stacktop.ul - ASM_NEEDS_REGISTERS64 - NUM_GPR_ARG_REGISTERS64;
gpr_end.ul = gpr_base.ul + NUM_GPR_ARG_REGISTERS64;
+#if _CALL_ELF == 2
+ rest.ul = stack + 4 + NUM_GPR_ARG_REGISTERS64;
+#else
rest.ul = stack + 6 + NUM_GPR_ARG_REGISTERS64;
+#endif
fpr_base.d = gpr_base.d - NUM_FPR_ARG_REGISTERS64;
fparg_count = 0;
next_arg.ul = gpr_base.ul;
@@ -477,30 +536,36 @@ ffi_prep_args64 (extended_cif *ecif, uns
/* Now for the arguments. */
p_argv.v = ecif->avalue;
- for (ptr = ecif->cif->arg_types, i = ecif->cif->nargs;
- i > 0;
- i--, ptr++, p_argv.v++)
+ nargs = ecif->cif->nargs;
+ nfixedargs = ecif->cif->nfixedargs;
+ for (ptr = ecif->cif->arg_types, i = 0;
+ i < nargs;
+ i++, ptr++, p_argv.v++)
{
+ unsigned int elt, elnum;
+
switch ((*ptr)->type)
{
case FFI_TYPE_FLOAT:
double_tmp = **p_argv.f;
- *next_arg.f = (float) double_tmp;
+ if (fparg_count < NUM_FPR_ARG_REGISTERS64 && i < nfixedargs)
+ *fpr_base.d++ = double_tmp;
+ else
+ *next_arg.f = (float) double_tmp;
if (++next_arg.ul == gpr_end.ul)
next_arg.ul = rest.ul;
- if (fparg_count < NUM_FPR_ARG_REGISTERS64)
- *fpr_base.d++ = double_tmp;
fparg_count++;
FFI_ASSERT (flags & FLAG_FP_ARGUMENTS);
break;
case FFI_TYPE_DOUBLE:
double_tmp = **p_argv.d;
- *next_arg.d = double_tmp;
+ if (fparg_count < NUM_FPR_ARG_REGISTERS64 && i < nfixedargs)
+ *fpr_base.d++ = double_tmp;
+ else
+ *next_arg.d = double_tmp;
if (++next_arg.ul == gpr_end.ul)
next_arg.ul = rest.ul;
- if (fparg_count < NUM_FPR_ARG_REGISTERS64)
- *fpr_base.d++ = double_tmp;
fparg_count++;
FFI_ASSERT (flags & FLAG_FP_ARGUMENTS);
break;
@@ -508,18 +573,20 @@ ffi_prep_args64 (extended_cif *ecif, uns
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
double_tmp = (*p_argv.d)[0];
- *next_arg.d = double_tmp;
+ if (fparg_count < NUM_FPR_ARG_REGISTERS64 && i < nfixedargs)
+ *fpr_base.d++ = double_tmp;
+ else
+ *next_arg.d = double_tmp;
if (++next_arg.ul == gpr_end.ul)
next_arg.ul = rest.ul;
- if (fparg_count < NUM_FPR_ARG_REGISTERS64)
- *fpr_base.d++ = double_tmp;
fparg_count++;
double_tmp = (*p_argv.d)[1];
- *next_arg.d = double_tmp;
+ if (fparg_count < NUM_FPR_ARG_REGISTERS64 && i < nfixedargs)
+ *fpr_base.d++ = double_tmp;
+ else
+ *next_arg.d = double_tmp;
if (++next_arg.ul == gpr_end.ul)
next_arg.ul = rest.ul;
- if (fparg_count < NUM_FPR_ARG_REGISTERS64)
- *fpr_base.d++ = double_tmp;
fparg_count++;
FFI_ASSERT (__LDBL_MANT_DIG__ == 106);
FFI_ASSERT (flags & FLAG_FP_ARGUMENTS);
@@ -527,27 +594,86 @@ ffi_prep_args64 (extended_cif *ecif, uns
#endif
case FFI_TYPE_STRUCT:
- words = ((*ptr)->size + 7) / 8;
- if (next_arg.ul >= gpr_base.ul && next_arg.ul + words > gpr_end.ul)
+#ifdef __STRUCT_PARM_ALIGN__
+ align = (*ptr)->alignment;
+ if (align > __STRUCT_PARM_ALIGN__)
+ align = __STRUCT_PARM_ALIGN__;
+ if (align > 1)
+ next_arg.p = ALIGN (next_arg.p, align);
+#endif
+ elt = 0;
+#if _CALL_ELF == 2
+ elt = discover_homogeneous_aggregate (*ptr, &elnum);
+#endif
+ if (elt)
{
- size_t first = gpr_end.c - next_arg.c;
- memcpy (next_arg.c, *p_argv.c, first);
- memcpy (rest.c, *p_argv.c + first, (*ptr)->size - first);
- next_arg.c = rest.c + words * 8 - first;
+ union {
+ void *v;
+ float *f;
+ double *d;
+ } arg;
+
+ arg.v = *p_argv.v;
+ if (elt == FFI_TYPE_FLOAT)
+ {
+ do
+ {
+ double_tmp = *arg.f++;
+ if (fparg_count < NUM_FPR_ARG_REGISTERS64
+ && i < nfixedargs)
+ *fpr_base.d++ = double_tmp;
+ else
+ *next_arg.f = (float) double_tmp;
+ if (++next_arg.f == gpr_end.f)
+ next_arg.f = rest.f;
+ fparg_count++;
+ }
+ while (--elnum != 0);
+ if ((next_arg.p & 3) != 0)
+ {
+ if (++next_arg.f == gpr_end.f)
+ next_arg.f = rest.f;
+ }
+ }
+ else
+ do
+ {
+ double_tmp = *arg.d++;
+ if (fparg_count < NUM_FPR_ARG_REGISTERS64 && i < nfixedargs)
+ *fpr_base.d++ = double_tmp;
+ else
+ *next_arg.d = double_tmp;
+ if (++next_arg.d == gpr_end.d)
+ next_arg.d = rest.d;
+ fparg_count++;
+ }
+ while (--elnum != 0);
}
else
{
- char *where = next_arg.c;
-
- /* Structures with size less than eight bytes are passed
- left-padded. */
- if ((*ptr)->size < 8)
- where += 8 - (*ptr)->size;
+ words = ((*ptr)->size + 7) / 8;
+ if (next_arg.ul >= gpr_base.ul && next_arg.ul + words > gpr_end.ul)
+ {
+ size_t first = gpr_end.c - next_arg.c;
+ memcpy (next_arg.c, *p_argv.c, first);
+ memcpy (rest.c, *p_argv.c + first, (*ptr)->size - first);
+ next_arg.c = rest.c + words * 8 - first;
+ }
+ else
+ {
+ char *where = next_arg.c;
- memcpy (where, *p_argv.c, (*ptr)->size);
- next_arg.ul += words;
- if (next_arg.ul == gpr_end.ul)
- next_arg.ul = rest.ul;
+#ifndef __LITTLE_ENDIAN__
+ /* Structures with size less than eight bytes are passed
+ left-padded. */
+ if ((*ptr)->size < 8)
+ where += 8 - (*ptr)->size;
+#endif
+ memcpy (where, *p_argv.c, (*ptr)->size);
+ next_arg.ul += words;
+ if (next_arg.ul == gpr_end.ul)
+ next_arg.ul = rest.ul;
+ }
}
break;
@@ -591,27 +717,22 @@ ffi_prep_args64 (extended_cif *ecif, uns
/* Perform machine dependent cif processing */
-ffi_status
-ffi_prep_cif_machdep (ffi_cif *cif)
+static ffi_status
+ffi_prep_cif_machdep_core (ffi_cif *cif)
{
/* All this is for the SYSV and LINUX64 ABI. */
- int i;
ffi_type **ptr;
unsigned bytes;
- int fparg_count = 0, intarg_count = 0;
- unsigned flags = 0;
+ unsigned i, fparg_count = 0, intarg_count = 0;
+ unsigned flags = cif->flags;
unsigned struct_copy_size = 0;
unsigned type = cif->rtype->type;
unsigned size = cif->rtype->size;
- if (cif->abi == FFI_LINUX_SOFT_FLOAT)
- NUM_FPR_ARG_REGISTERS = 0;
-
+ /* The machine-independent calculation of cif->bytes doesn't work
+ for us. Redo the calculation. */
if (cif->abi != FFI_LINUX64)
{
- /* All the machine-independent calculation of cif->bytes will be wrong.
- Redo the calculation for SYSV. */
-
/* Space for the frame pointer, callee's LR, and the asm's temp regs. */
bytes = (2 + ASM_NEEDS_REGISTERS) * sizeof (int);
@@ -621,13 +742,20 @@ ffi_prep_cif_machdep (ffi_cif *cif)
else
{
/* 64-bit ABI. */
+#if _CALL_ELF == 2
+ /* Space for backchain, CR, LR, TOC and the asm's temp regs. */
+ bytes = (4 + ASM_NEEDS_REGISTERS64) * sizeof (long);
+ /* Space for the general registers. */
+ bytes += NUM_GPR_ARG_REGISTERS64 * sizeof (long);
+#else
/* Space for backchain, CR, LR, cc/ld doubleword, TOC and the asm's temp
regs. */
bytes = (6 + ASM_NEEDS_REGISTERS64) * sizeof (long);
/* Space for the mandatory parm save area and general registers. */
bytes += 2 * NUM_GPR_ARG_REGISTERS64 * sizeof (long);
+#endif
}
/* Return value handling. The rules for SYSV are as follows:
@@ -646,13 +774,30 @@ ffi_prep_cif_machdep (ffi_cif *cif)
- Single/double FP values in fpr1, long double in fpr1,fpr2.
- soft-float float/doubles are treated as UINT32/UINT64 respectivley.
- soft-float long doubles are returned in gpr3-gpr6. */
+ /* First translate for softfloat/nonlinux */
+ if (cif->abi == FFI_LINUX_SOFT_FLOAT)
+ {
+ if (type == FFI_TYPE_FLOAT)
+ type = FFI_TYPE_UINT32;
+ if (type == FFI_TYPE_DOUBLE)
+ type = FFI_TYPE_UINT64;
+ if (type == FFI_TYPE_LONGDOUBLE)
+ type = FFI_TYPE_UINT128;
+ }
+ else if (cif->abi != FFI_LINUX
+ && cif->abi != FFI_LINUX64)
+ {
+#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
+ if (type == FFI_TYPE_LONGDOUBLE)
+ type = FFI_TYPE_STRUCT;
+#endif
+ }
+
switch (type)
{
+#ifndef __NO_FPRS__
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
- if (cif->abi != FFI_LINUX && cif->abi != FFI_LINUX64
- && cif->abi != FFI_LINUX_SOFT_FLOAT)
- goto byref;
flags |= FLAG_RETURNS_128BITS;
/* Fall through. */
#endif
@@ -660,47 +805,52 @@ ffi_prep_cif_machdep (ffi_cif *cif)
flags |= FLAG_RETURNS_64BITS;
/* Fall through. */
case FFI_TYPE_FLOAT:
- /* With FFI_LINUX_SOFT_FLOAT no fp registers are used. */
- if (cif->abi != FFI_LINUX_SOFT_FLOAT)
- flags |= FLAG_RETURNS_FP;
+ flags |= FLAG_RETURNS_FP;
break;
+#endif
+ case FFI_TYPE_UINT128:
+ flags |= FLAG_RETURNS_128BITS;
+ /* Fall through. */
case FFI_TYPE_UINT64:
case FFI_TYPE_SINT64:
flags |= FLAG_RETURNS_64BITS;
break;
case FFI_TYPE_STRUCT:
- if (cif->abi == FFI_SYSV)
+ /*
+ * The final SYSV ABI says that structures smaller or equal 8 bytes
+ * are returned in r3/r4. The FFI_GCC_SYSV ABI instead returns them
+ * in memory.
+ *
+ * NOTE: The assembly code can safely assume that it just needs to
+ * store both r3 and r4 into a 8-byte word-aligned buffer, as
+ * we allocate a temporary buffer in ffi_call() if this flag is
+ * set.
+ */
+ if (cif->abi == FFI_SYSV && size <= 8)
{
- /* The final SYSV ABI says that structures smaller or equal 8 bytes
- are returned in r3/r4. The FFI_GCC_SYSV ABI instead returns them
- in memory. */
-
- /* Treat structs with size <= 8 bytes. */
- if (size <= 8)
+ flags |= FLAG_RETURNS_SMST;
+ break;
+ }
+#if _CALL_ELF == 2
+ if (cif->abi == FFI_LINUX64)
+ {
+ unsigned int elt, elnum;
+ elt = discover_homogeneous_aggregate (cif->rtype, &elnum);
+ if (elt)
+ {
+ if (elt == FFI_TYPE_DOUBLE)
+ flags |= FLAG_RETURNS_64BITS;
+ flags |= FLAG_RETURNS_FP | FLAG_RETURNS_SMST;
+ break;
+ }
+ if (size <= 16)
{
flags |= FLAG_RETURNS_SMST;
- /* These structs are returned in r3. We pack the type and the
- precalculated shift value (needed in the sysv.S) into flags.
- The same applies for the structs returned in r3/r4. */
- if (size <= 4)
- {
- flags |= FLAG_SYSV_SMST_R3;
- flags |= 8 * (4 - size) << 8;
- break;
- }
- /* These structs are returned in r3 and r4. See above. */
- if (size <= 8)
- {
- flags |= FLAG_SYSV_SMST_R3 | FLAG_SYSV_SMST_R4;
- flags |= 8 * (8 - size) << 8;
- break;
- }
+ break;
}
}
-#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
- byref:
#endif
intarg_count++;
flags |= FLAG_RETVAL_REFERENCE;
@@ -722,39 +872,36 @@ ffi_prep_cif_machdep (ffi_cif *cif)
Stuff on the stack needs to keep proper alignment. */
for (ptr = cif->arg_types, i = cif->nargs; i > 0; i--, ptr++)
{
- switch ((*ptr)->type)
- {
+ unsigned short typenum = (*ptr)->type;
+
+ /* We may need to handle some values depending on ABI */
+ if (cif->abi == FFI_LINUX_SOFT_FLOAT) {
+ if (typenum == FFI_TYPE_FLOAT)
+ typenum = FFI_TYPE_UINT32;
+ if (typenum == FFI_TYPE_DOUBLE)
+ typenum = FFI_TYPE_UINT64;
+ if (typenum == FFI_TYPE_LONGDOUBLE)
+ typenum = FFI_TYPE_UINT128;
+ } else if (cif->abi != FFI_LINUX && cif->abi != FFI_LINUX64) {
+#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
+ if (typenum == FFI_TYPE_LONGDOUBLE)
+ typenum = FFI_TYPE_STRUCT;
+#endif
+ }
+
+ switch (typenum) {
+#ifndef __NO_FPRS__
case FFI_TYPE_FLOAT:
- /* With FFI_LINUX_SOFT_FLOAT floats are handled like UINT32. */
- if (cif->abi == FFI_LINUX_SOFT_FLOAT)
- goto soft_float_cif;
fparg_count++;
/* floating singles are not 8-aligned on stack */
break;
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
- if (cif->abi != FFI_LINUX && cif->abi != FFI_LINUX_SOFT_FLOAT)
- goto do_struct;
- if (cif->abi == FFI_LINUX_SOFT_FLOAT)
- {
- if (intarg_count >= NUM_GPR_ARG_REGISTERS - 3
- || intarg_count < NUM_GPR_ARG_REGISTERS)
- /* A long double in FFI_LINUX_SOFT_FLOAT can use only
- a set of four consecutive gprs. If we have not enough,
- we have to adjust the intarg_count value. */
- intarg_count += NUM_GPR_ARG_REGISTERS - intarg_count;
- intarg_count += 4;
- break;
- }
- else
- fparg_count++;
+ fparg_count++;
/* Fall thru */
#endif
case FFI_TYPE_DOUBLE:
- /* With FFI_LINUX_SOFT_FLOAT doubles are handled like UINT64. */
- if (cif->abi == FFI_LINUX_SOFT_FLOAT)
- goto soft_double_cif;
fparg_count++;
/* If this FP arg is going on the stack, it must be
8-byte-aligned. */
@@ -763,10 +910,21 @@ ffi_prep_cif_machdep (ffi_cif *cif)
&& intarg_count % 2 != 0)
intarg_count++;
break;
+#endif
+ case FFI_TYPE_UINT128:
+ /*
+ * A long double in FFI_LINUX_SOFT_FLOAT can use only a set
+ * of four consecutive gprs. If we do not have enough, we
+ * have to adjust the intarg_count value.
+ */
+ if (intarg_count >= NUM_GPR_ARG_REGISTERS - 3
+ && intarg_count < NUM_GPR_ARG_REGISTERS)
+ intarg_count = NUM_GPR_ARG_REGISTERS;
+ intarg_count += 4;
+ break;
case FFI_TYPE_UINT64:
case FFI_TYPE_SINT64:
- soft_double_cif:
/* 'long long' arguments are passed as two words, but
either both words must fit in registers or both go
on the stack. If they go on the stack, they must
@@ -783,9 +941,6 @@ ffi_prep_cif_machdep (ffi_cif *cif)
break;
case FFI_TYPE_STRUCT:
-#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
- do_struct:
-#endif
/* We must allocate space for a copy of these to enforce
pass-by-value. Pad the space up to a multiple of 16
bytes (the maximum alignment required for anything under
@@ -793,50 +948,100 @@ ffi_prep_cif_machdep (ffi_cif *cif)
struct_copy_size += ((*ptr)->size + 15) & ~0xF;
/* Fall through (allocate space for the pointer). */
- default:
- soft_float_cif:
+ case FFI_TYPE_POINTER:
+ case FFI_TYPE_INT:
+ case FFI_TYPE_UINT32:
+ case FFI_TYPE_SINT32:
+ case FFI_TYPE_UINT16:
+ case FFI_TYPE_SINT16:
+ case FFI_TYPE_UINT8:
+ case FFI_TYPE_SINT8:
/* Everything else is passed as a 4-byte word in a GPR, either
the object itself or a pointer to it. */
intarg_count++;
break;
+ default:
+ FFI_ASSERT (0);
}
}
else
for (ptr = cif->arg_types, i = cif->nargs; i > 0; i--, ptr++)
{
+ unsigned int elt, elnum;
+#ifdef __STRUCT_PARM_ALIGN__
+ unsigned int align;
+#endif
+
switch ((*ptr)->type)
{
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
- if (cif->abi == FFI_LINUX_SOFT_FLOAT)
- intarg_count += 4;
- else
- {
- fparg_count += 2;
- intarg_count += 2;
- }
+ fparg_count += 2;
+ intarg_count += 2;
+ if (fparg_count > NUM_FPR_ARG_REGISTERS)
+ flags |= FLAG_ARG_NEEDS_PSAVE;
break;
#endif
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
fparg_count++;
intarg_count++;
+ if (fparg_count > NUM_FPR_ARG_REGISTERS)
+ flags |= FLAG_ARG_NEEDS_PSAVE;
break;
case FFI_TYPE_STRUCT:
+#ifdef __STRUCT_PARM_ALIGN__
+ align = (*ptr)->alignment;
+ if (align > __STRUCT_PARM_ALIGN__)
+ align = __STRUCT_PARM_ALIGN__;
+ align = align / 8;
+ if (align > 1)
+ intarg_count = ALIGN (intarg_count, align);
+#endif
intarg_count += ((*ptr)->size + 7) / 8;
+ elt = 0;
+#if _CALL_ELF == 2
+ elt = discover_homogeneous_aggregate (*ptr, &elnum);
+#endif
+ if (elt)
+ {
+ fparg_count += elnum;
+ if (fparg_count > NUM_FPR_ARG_REGISTERS)
+ flags |= FLAG_ARG_NEEDS_PSAVE;
+ }
+ else
+ {
+ if (intarg_count > NUM_GPR_ARG_REGISTERS)
+ flags |= FLAG_ARG_NEEDS_PSAVE;
+ }
break;
- default:
+ case FFI_TYPE_POINTER:
+ case FFI_TYPE_UINT64:
+ case FFI_TYPE_SINT64:
+ case FFI_TYPE_INT:
+ case FFI_TYPE_UINT32:
+ case FFI_TYPE_SINT32:
+ case FFI_TYPE_UINT16:
+ case FFI_TYPE_SINT16:
+ case FFI_TYPE_UINT8:
+ case FFI_TYPE_SINT8:
/* Everything else is passed as a 8-byte word in a GPR, either
the object itself or a pointer to it. */
intarg_count++;
+ if (intarg_count > NUM_GPR_ARG_REGISTERS)
+ flags |= FLAG_ARG_NEEDS_PSAVE;
break;
+ default:
+ FFI_ASSERT (0);
}
}
+#ifndef __NO_FPRS__
if (fparg_count != 0)
flags |= FLAG_FP_ARGUMENTS;
+#endif
if (intarg_count > 4)
flags |= FLAG_4_GPR_ARGUMENTS;
if (struct_copy_size != 0)
@@ -844,25 +1049,36 @@ ffi_prep_cif_machdep (ffi_cif *cif)
if (cif->abi != FFI_LINUX64)
{
+#ifndef __NO_FPRS__
/* Space for the FPR registers, if needed. */
if (fparg_count != 0)
bytes += NUM_FPR_ARG_REGISTERS * sizeof (double);
+#endif
/* Stack space. */
if (intarg_count > NUM_GPR_ARG_REGISTERS)
bytes += (intarg_count - NUM_GPR_ARG_REGISTERS) * sizeof (int);
+#ifndef __NO_FPRS__
if (fparg_count > NUM_FPR_ARG_REGISTERS)
bytes += (fparg_count - NUM_FPR_ARG_REGISTERS) * sizeof (double);
+#endif
}
else
{
+#ifndef __NO_FPRS__
/* Space for the FPR registers, if needed. */
if (fparg_count != 0)
bytes += NUM_FPR_ARG_REGISTERS64 * sizeof (double);
+#endif
/* Stack space. */
+#if _CALL_ELF == 2
+ if ((flags & FLAG_ARG_NEEDS_PSAVE) != 0)
+ bytes += intarg_count * sizeof (long);
+#else
if (intarg_count > NUM_GPR_ARG_REGISTERS64)
bytes += (intarg_count - NUM_GPR_ARG_REGISTERS64) * sizeof (long);
+#endif
}
/* The stack space allocated needs to be a multiple of 16 bytes. */
@@ -877,6 +1093,26 @@ ffi_prep_cif_machdep (ffi_cif *cif)
return FFI_OK;
}
+ffi_status
+ffi_prep_cif_machdep (ffi_cif *cif)
+{
+ cif->nfixedargs = cif->nargs;
+ return ffi_prep_cif_machdep_core (cif);
+}
+
+ffi_status
+ffi_prep_cif_machdep_var (ffi_cif *cif,
+ unsigned int nfixedargs,
+ unsigned int ntotalargs MAYBE_UNUSED)
+{
+ cif->nfixedargs = nfixedargs;
+#if _CALL_ELF == 2
+ if (cif->abi == FFI_LINUX64)
+ cif->flags |= FLAG_ARG_NEEDS_PSAVE;
+#endif
+ return ffi_prep_cif_machdep_core (cif);
+}
+
extern void ffi_call_SYSV(extended_cif *, unsigned, unsigned, unsigned *,
void (*fn)(void));
extern void FFI_HIDDEN ffi_call_LINUX64(extended_cif *, unsigned long,
@@ -886,28 +1122,39 @@ extern void FFI_HIDDEN ffi_call_LINUX64(
void
ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue)
{
+ /*
+ * The final SYSV ABI says that structures smaller or equal 8 bytes
+ * are returned in r3/r4. The FFI_GCC_SYSV ABI instead returns them
+ * in memory.
+ *
+ * We bounce-buffer SYSV small struct return values so that sysv.S
+ * can write r3 and r4 to memory without worrying about struct size.
+ *
+ * For ELFv2 ABI, use a bounce buffer for homogeneous structs too,
+ * for similar reasons.
+ */
+ unsigned long smst_buffer[8];
extended_cif ecif;
ecif.cif = cif;
ecif.avalue = avalue;
- /* If the return value is a struct and we don't have a return */
- /* value address then we need to make one */
-
- if ((rvalue == NULL) && (cif->rtype->type == FFI_TYPE_STRUCT))
- {
- ecif.rvalue = alloca(cif->rtype->size);
- }
- else
- ecif.rvalue = rvalue;
-
+ ecif.rvalue = rvalue;
+ if ((cif->flags & FLAG_RETURNS_SMST) != 0)
+ ecif.rvalue = smst_buffer;
+ /* Ensure that we have a valid struct return value.
+ FIXME: Isn't this just papering over a user problem? */
+ else if (!rvalue && cif->rtype->type == FFI_TYPE_STRUCT)
+ ecif.rvalue = alloca (cif->rtype->size);
switch (cif->abi)
{
#ifndef POWERPC64
+# ifndef __NO_FPRS__
case FFI_SYSV:
case FFI_GCC_SYSV:
case FFI_LINUX:
+# endif
case FFI_LINUX_SOFT_FLOAT:
ffi_call_SYSV (&ecif, -cif->bytes, cif->flags, ecif.rvalue, fn);
break;
@@ -920,10 +1167,29 @@ ffi_call(ffi_cif *cif, void (*fn)(void),
FFI_ASSERT (0);
break;
}
+
+ /* Check for a bounce-buffered return value */
+ if (rvalue && ecif.rvalue == smst_buffer)
+ {
+ unsigned int rsize = cif->rtype->size;
+#ifndef __LITTLE_ENDIAN__
+ /* The SYSV ABI returns a structure of up to 4 bytes in size
+ left-padded in r3. */
+ if (cif->abi == FFI_SYSV && rsize <= 4)
+ memcpy (rvalue, (char *) smst_buffer + 4 - rsize, rsize);
+ /* The SYSV ABI returns a structure of up to 8 bytes in size
+ left-padded in r3/r4, and the ELFv2 ABI similarly returns a
+ structure of up to 8 bytes in size left-padded in r3. */
+ else if (rsize <= 8)
+ memcpy (rvalue, (char *) smst_buffer + 8 - rsize, rsize);
+ else
+#endif
+ memcpy (rvalue, smst_buffer, rsize);
+ }
}
-#ifndef POWERPC64
+#if !defined POWERPC64 || _CALL_ELF == 2
#define MIN_CACHE_LINE_SIZE 8
static void
@@ -947,16 +1213,38 @@ ffi_prep_closure_loc (ffi_closure *closu
void *codeloc)
{
#ifdef POWERPC64
+# if _CALL_ELF == 2
+ unsigned int *tramp = (unsigned int *) &closure->tramp[0];
+
+ if (cif->abi != FFI_LINUX64)
+ return FFI_BAD_ABI;
+
+ tramp[0] = 0xe96c0018; /* 0: ld 11,2f-0b(12) */
+ tramp[1] = 0xe98c0010; /* ld 12,1f-0b(12) */
+ tramp[2] = 0x7d8903a6; /* mtctr 12 */
+ tramp[3] = 0x4e800420; /* bctr */
+ /* 1: .quad function_addr */
+ /* 2: .quad context */
+ *(void **) &tramp[4] = (void *) ffi_closure_LINUX64;
+ *(void **) &tramp[6] = codeloc;
+ flush_icache ((char *)tramp, (char *)codeloc, FFI_TRAMPOLINE_SIZE);
+# else
void **tramp = (void **) &closure->tramp[0];
- FFI_ASSERT (cif->abi == FFI_LINUX64);
+ if (cif->abi != FFI_LINUX64)
+ return FFI_BAD_ABI;
/* Copy function address and TOC from ffi_closure_LINUX64. */
memcpy (tramp, (char *) ffi_closure_LINUX64, 16);
tramp[2] = codeloc;
+# endif
#else
unsigned int *tramp;
- FFI_ASSERT (cif->abi == FFI_GCC_SYSV || cif->abi == FFI_SYSV);
+ if (! (cif->abi == FFI_GCC_SYSV
+ || cif->abi == FFI_SYSV
+ || cif->abi == FFI_LINUX
+ || cif->abi == FFI_LINUX_SOFT_FLOAT))
+ return FFI_BAD_ABI;
tramp = (unsigned int *) &closure->tramp[0];
tramp[0] = 0x7c0802a6; /* mflr r0 */
@@ -1011,32 +1299,38 @@ ffi_closure_helper_SYSV (ffi_closure *cl
void ** avalue;
ffi_type ** arg_types;
long i, avn;
- long nf; /* number of floating registers already used */
- long ng; /* number of general registers already used */
- ffi_cif * cif;
- double temp;
- unsigned size;
+#ifndef __NO_FPRS__
+ long nf = 0; /* number of floating registers already used */
+#endif
+ long ng = 0; /* number of general registers already used */
+
+ ffi_cif *cif = closure->cif;
+ unsigned size = cif->rtype->size;
+ unsigned short rtypenum = cif->rtype->type;
- cif = closure->cif;
avalue = alloca (cif->nargs * sizeof (void *));
- size = cif->rtype->size;
- nf = 0;
- ng = 0;
+ /* First translate for softfloat/nonlinux */
+ if (cif->abi == FFI_LINUX_SOFT_FLOAT) {
+ if (rtypenum == FFI_TYPE_FLOAT)
+ rtypenum = FFI_TYPE_UINT32;
+ if (rtypenum == FFI_TYPE_DOUBLE)
+ rtypenum = FFI_TYPE_UINT64;
+ if (rtypenum == FFI_TYPE_LONGDOUBLE)
+ rtypenum = FFI_TYPE_UINT128;
+ } else if (cif->abi != FFI_LINUX && cif->abi != FFI_LINUX64) {
+#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
+ if (rtypenum == FFI_TYPE_LONGDOUBLE)
+ rtypenum = FFI_TYPE_STRUCT;
+#endif
+ }
+
/* Copy the caller's structure return value address so that the closure
returns the data directly to the caller.
For FFI_SYSV the result is passed in r3/r4 if the struct size is less
or equal 8 bytes. */
-
- if ((cif->rtype->type == FFI_TYPE_STRUCT
- && !((cif->abi == FFI_SYSV) && (size <= 8)))
-#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
- || (cif->rtype->type == FFI_TYPE_LONGDOUBLE
- && cif->abi != FFI_LINUX && cif->abi != FFI_LINUX_SOFT_FLOAT)
-#endif
- )
- {
+ if (rtypenum == FFI_TYPE_STRUCT && ((cif->abi != FFI_SYSV) || (size > 8))) {
rvalue = (void *) *pgr;
ng++;
pgr++;
@@ -1047,12 +1341,112 @@ ffi_closure_helper_SYSV (ffi_closure *cl
arg_types = cif->arg_types;
/* Grab the addresses of the arguments from the stack frame. */
- while (i < avn)
- {
- switch (arg_types[i]->type)
- {
+ while (i < avn) {
+ unsigned short typenum = arg_types[i]->type;
+
+ /* We may need to handle some values depending on ABI */
+ if (cif->abi == FFI_LINUX_SOFT_FLOAT) {
+ if (typenum == FFI_TYPE_FLOAT)
+ typenum = FFI_TYPE_UINT32;
+ if (typenum == FFI_TYPE_DOUBLE)
+ typenum = FFI_TYPE_UINT64;
+ if (typenum == FFI_TYPE_LONGDOUBLE)
+ typenum = FFI_TYPE_UINT128;
+ } else if (cif->abi != FFI_LINUX && cif->abi != FFI_LINUX64) {
+#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
+ if (typenum == FFI_TYPE_LONGDOUBLE)
+ typenum = FFI_TYPE_STRUCT;
+#endif
+ }
+
+ switch (typenum) {
+#ifndef __NO_FPRS__
+ case FFI_TYPE_FLOAT:
+ /* unfortunately float values are stored as doubles
+ * in the ffi_closure_SYSV code (since we don't check
+ * the type in that routine).
+ */
+
+ /* there are 8 64bit floating point registers */
+
+ if (nf < 8)
+ {
+ double temp = pfr->d;
+ pfr->f = (float) temp;
+ avalue[i] = pfr;
+ nf++;
+ pfr++;
+ }
+ else
+ {
+ /* FIXME? here we are really changing the values
+ * stored in the original calling routines outgoing
+ * parameter stack. This is probably a really
+ * naughty thing to do but...
+ */
+ avalue[i] = pst;
+ pst += 1;
+ }
+ break;
+
+ case FFI_TYPE_DOUBLE:
+ /* On the outgoing stack all values are aligned to 8 */
+ /* there are 8 64bit floating point registers */
+
+ if (nf < 8)
+ {
+ avalue[i] = pfr;
+ nf++;
+ pfr++;
+ }
+ else
+ {
+ if (((long) pst) & 4)
+ pst++;
+ avalue[i] = pst;
+ pst += 2;
+ }
+ break;
+
+#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
+ case FFI_TYPE_LONGDOUBLE:
+ if (nf < 7)
+ {
+ avalue[i] = pfr;
+ pfr += 2;
+ nf += 2;
+ }
+ else
+ {
+ if (((long) pst) & 4)
+ pst++;
+ avalue[i] = pst;
+ pst += 4;
+ nf = 8;
+ }
+ break;
+#endif
+#endif /* have FPRS */
+
+ case FFI_TYPE_UINT128:
+ /*
+ * Test if for the whole long double, 4 gprs are available.
+ * otherwise the stuff ends up on the stack.
+ */
+ if (ng < 5) {
+ avalue[i] = pgr;
+ pgr += 4;
+ ng += 4;
+ } else {
+ avalue[i] = pst;
+ pst += 4;
+ ng = 8+4;
+ }
+ break;
+
case FFI_TYPE_SINT8:
case FFI_TYPE_UINT8:
+#ifndef __LITTLE_ENDIAN__
/* there are 8 gpr registers used to pass values */
if (ng < 8)
{
@@ -1066,9 +1460,11 @@ ffi_closure_helper_SYSV (ffi_closure *cl
pst++;
}
break;
+#endif
case FFI_TYPE_SINT16:
case FFI_TYPE_UINT16:
+#ifndef __LITTLE_ENDIAN__
/* there are 8 gpr registers used to pass values */
if (ng < 8)
{
@@ -1082,11 +1478,11 @@ ffi_closure_helper_SYSV (ffi_closure *cl
pst++;
}
break;
+#endif
case FFI_TYPE_SINT32:
case FFI_TYPE_UINT32:
case FFI_TYPE_POINTER:
- soft_float_closure:
/* there are 8 gpr registers used to pass values */
if (ng < 8)
{
@@ -1102,9 +1498,6 @@ ffi_closure_helper_SYSV (ffi_closure *cl
break;
case FFI_TYPE_STRUCT:
-#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
- do_struct:
-#endif
/* Structs are passed by reference. The address will appear in a
gpr if it is one of the first 8 arguments. */
if (ng < 8)
@@ -1122,7 +1515,6 @@ ffi_closure_helper_SYSV (ffi_closure *cl
case FFI_TYPE_SINT64:
case FFI_TYPE_UINT64:
- soft_double_closure:
/* passing long long ints are complex, they must
* be passed in suitable register pairs such as
* (r3,r4) or (r5,r6) or (r6,r7), or (r7,r8) or (r9,r10)
@@ -1154,99 +1546,8 @@ ffi_closure_helper_SYSV (ffi_closure *cl
}
break;
- case FFI_TYPE_FLOAT:
- /* With FFI_LINUX_SOFT_FLOAT floats are handled like UINT32. */
- if (cif->abi == FFI_LINUX_SOFT_FLOAT)
- goto soft_float_closure;
- /* unfortunately float values are stored as doubles
- * in the ffi_closure_SYSV code (since we don't check
- * the type in that routine).
- */
-
- /* there are 8 64bit floating point registers */
-
- if (nf < 8)
- {
- temp = pfr->d;
- pfr->f = (float) temp;
- avalue[i] = pfr;
- nf++;
- pfr++;
- }
- else
- {
- /* FIXME? here we are really changing the values
- * stored in the original calling routines outgoing
- * parameter stack. This is probably a really
- * naughty thing to do but...
- */
- avalue[i] = pst;
- pst += 1;
- }
- break;
-
- case FFI_TYPE_DOUBLE:
- /* With FFI_LINUX_SOFT_FLOAT doubles are handled like UINT64. */
- if (cif->abi == FFI_LINUX_SOFT_FLOAT)
- goto soft_double_closure;
- /* On the outgoing stack all values are aligned to 8 */
- /* there are 8 64bit floating point registers */
-
- if (nf < 8)
- {
- avalue[i] = pfr;
- nf++;
- pfr++;
- }
- else
- {
- if (((long) pst) & 4)
- pst++;
- avalue[i] = pst;
- pst += 2;
- }
- break;
-
-#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
- case FFI_TYPE_LONGDOUBLE:
- if (cif->abi != FFI_LINUX && cif->abi != FFI_LINUX_SOFT_FLOAT)
- goto do_struct;
- if (cif->abi == FFI_LINUX_SOFT_FLOAT)
- { /* Test if for the whole long double, 4 gprs are available.
- otherwise the stuff ends up on the stack. */
- if (ng < 5)
- {
- avalue[i] = pgr;
- pgr += 4;
- ng += 4;
- }
- else
- {
- avalue[i] = pst;
- pst += 4;
- ng = 8;
- }
- break;
- }
- if (nf < 7)
- {
- avalue[i] = pfr;
- pfr += 2;
- nf += 2;
- }
- else
- {
- if (((long) pst) & 4)
- pst++;
- avalue[i] = pst;
- pst += 4;
- nf = 8;
- }
- break;
-#endif
-
default:
- FFI_ASSERT (0);
+ FFI_ASSERT (0);
}
i++;
@@ -1263,39 +1564,9 @@ ffi_closure_helper_SYSV (ffi_closure *cl
already used and we never have a struct with size zero. That is the reason
for the subtraction of 1. See the comment in ffitarget.h about ordering.
*/
- if (cif->abi == FFI_SYSV && cif->rtype->type == FFI_TYPE_STRUCT
- && size <= 8)
+ if (cif->abi == FFI_SYSV && rtypenum == FFI_TYPE_STRUCT && size <= 8)
return (FFI_SYSV_TYPE_SMALL_STRUCT - 1) + size;
-#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
- else if (cif->rtype->type == FFI_TYPE_LONGDOUBLE
- && cif->abi != FFI_LINUX && cif->abi != FFI_LINUX_SOFT_FLOAT)
- return FFI_TYPE_STRUCT;
-#endif
- /* With FFI_LINUX_SOFT_FLOAT floats and doubles are handled like UINT32
- respectivley UINT64. */
- if (cif->abi == FFI_LINUX_SOFT_FLOAT)
- {
- switch (cif->rtype->type)
- {
- case FFI_TYPE_FLOAT:
- return FFI_TYPE_UINT32;
- break;
- case FFI_TYPE_DOUBLE:
- return FFI_TYPE_UINT64;
- break;
-#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
- case FFI_TYPE_LONGDOUBLE:
- return FFI_TYPE_UINT128;
- break;
-#endif
- default:
- return cif->rtype->type;
- }
- }
- else
- {
- return cif->rtype->type;
- }
+ return rtypenum;
}
int FFI_HIDDEN ffi_closure_helper_LINUX64 (ffi_closure *, void *,
@@ -1312,16 +1583,20 @@ ffi_closure_helper_LINUX64 (ffi_closure
void **avalue;
ffi_type **arg_types;
- long i, avn;
+ unsigned long i, avn, nfixedargs;
ffi_cif *cif;
ffi_dblfl *end_pfr = pfr + NUM_FPR_ARG_REGISTERS64;
+#ifdef __STRUCT_PARM_ALIGN__
+ unsigned long align;
+#endif
cif = closure->cif;
avalue = alloca (cif->nargs * sizeof (void *));
- /* Copy the caller's structure return value address so that the closure
- returns the data directly to the caller. */
- if (cif->rtype->type == FFI_TYPE_STRUCT)
+ /* Copy the caller's structure return value address so that the
+ closure returns the data directly to the caller. */
+ if (cif->rtype->type == FFI_TYPE_STRUCT
+ && (cif->flags & FLAG_RETURNS_SMST) == 0)
{
rvalue = (void *) *pst;
pst++;
@@ -1329,30 +1604,39 @@ ffi_closure_helper_LINUX64 (ffi_closure
i = 0;
avn = cif->nargs;
+ nfixedargs = cif->nfixedargs;
arg_types = cif->arg_types;
/* Grab the addresses of the arguments from the stack frame. */
while (i < avn)
{
+ unsigned int elt, elnum;
+
switch (arg_types[i]->type)
{
case FFI_TYPE_SINT8:
case FFI_TYPE_UINT8:
+#ifndef __LITTLE_ENDIAN__
avalue[i] = (char *) pst + 7;
pst++;
break;
+#endif
case FFI_TYPE_SINT16:
case FFI_TYPE_UINT16:
+#ifndef __LITTLE_ENDIAN__
avalue[i] = (char *) pst + 6;
pst++;
break;
+#endif
case FFI_TYPE_SINT32:
case FFI_TYPE_UINT32:
+#ifndef __LITTLE_ENDIAN__
avalue[i] = (char *) pst + 4;
pst++;
break;
+#endif
case FFI_TYPE_SINT64:
case FFI_TYPE_UINT64:
@@ -1362,12 +1646,82 @@ ffi_closure_helper_LINUX64 (ffi_closure
break;
case FFI_TYPE_STRUCT:
- /* Structures with size less than eight bytes are passed
- left-padded. */
- if (arg_types[i]->size < 8)
- avalue[i] = (char *) pst + 8 - arg_types[i]->size;
+#ifdef __STRUCT_PARM_ALIGN__
+ align = arg_types[i]->alignment;
+ if (align > __STRUCT_PARM_ALIGN__)
+ align = __STRUCT_PARM_ALIGN__;
+ if (align > 1)
+ pst = (unsigned long *) ALIGN ((size_t) pst, align);
+#endif
+ elt = 0;
+#if _CALL_ELF == 2
+ elt = discover_homogeneous_aggregate (arg_types[i], &elnum);
+#endif
+ if (elt)
+ {
+ union {
+ void *v;
+ unsigned long *ul;
+ float *f;
+ double *d;
+ size_t p;
+ } to, from;
+
+ /* Repackage the aggregate from its parts. The
+ aggregate size is not greater than the space taken by
+ the registers so store back to the register/parameter
+ save arrays. */
+ if (pfr + elnum <= end_pfr)
+ to.v = pfr;
+ else
+ to.v = pst;
+
+ avalue[i] = to.v;
+ from.ul = pst;
+ if (elt == FFI_TYPE_FLOAT)
+ {
+ do
+ {
+ if (pfr < end_pfr && i < nfixedargs)
+ {
+ *to.f = (float) pfr->d;
+ pfr++;
+ }
+ else
+ *to.f = *from.f;
+ to.f++;
+ from.f++;
+ }
+ while (--elnum != 0);
+ }
+ else
+ {
+ do
+ {
+ if (pfr < end_pfr && i < nfixedargs)
+ {
+ *to.d = pfr->d;
+ pfr++;
+ }
+ else
+ *to.d = *from.d;
+ to.d++;
+ from.d++;
+ }
+ while (--elnum != 0);
+ }
+ }
else
- avalue[i] = pst;
+ {
+#ifndef __LITTLE_ENDIAN__
+ /* Structures with size less than eight bytes are passed
+ left-padded. */
+ if (arg_types[i]->size < 8)
+ avalue[i] = (char *) pst + 8 - arg_types[i]->size;
+ else
+#endif
+ avalue[i] = pst;
+ }
pst += (arg_types[i]->size + 7) / 8;
break;
@@ -1379,7 +1733,7 @@ ffi_closure_helper_LINUX64 (ffi_closure
/* there are 13 64bit floating point registers */
- if (pfr < end_pfr)
+ if (pfr < end_pfr && i < nfixedargs)
{
double temp = pfr->d;
pfr->f = (float) temp;
@@ -1395,7 +1749,7 @@ ffi_closure_helper_LINUX64 (ffi_closure
/* On the outgoing stack all values are aligned to 8 */
/* there are 13 64bit floating point registers */
- if (pfr < end_pfr)
+ if (pfr < end_pfr && i < nfixedargs)
{
avalue[i] = pfr;
pfr++;
@@ -1407,14 +1761,14 @@ ffi_closure_helper_LINUX64 (ffi_closure
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
- if (pfr + 1 < end_pfr)
+ if (pfr + 1 < end_pfr && i + 1 < nfixedargs)
{
avalue[i] = pfr;
pfr += 2;
}
else
{
- if (pfr < end_pfr)
+ if (pfr < end_pfr && i < nfixedargs)
{
/* Passed partly in f13 and partly on the stack.
Move it all to the stack. */
@@ -1438,5 +1792,14 @@ ffi_closure_helper_LINUX64 (ffi_closure
(closure->fun) (cif, rvalue, avalue, closure->user_data);
/* Tell ffi_closure_LINUX64 how to perform return type promotions. */
+ if ((cif->flags & FLAG_RETURNS_SMST) != 0)
+ {
+ if ((cif->flags & FLAG_RETURNS_FP) == 0)
+ return FFI_V2_TYPE_SMALL_STRUCT + cif->rtype->size - 1;
+ else if ((cif->flags & FLAG_RETURNS_64BITS) != 0)
+ return FFI_V2_TYPE_DOUBLE_HOMOG;
+ else
+ return FFI_V2_TYPE_FLOAT_HOMOG;
+ }
return cif->rtype->type;
}
Index: mozilla/js/src/ctypes/libffi/src/powerpc/ffi_darwin.c
===================================================================
--- mozilla.orig/js/src/ctypes/libffi/src/powerpc/ffi_darwin.c
+++ mozilla/js/src/ctypes/libffi/src/powerpc/ffi_darwin.c
@@ -3,7 +3,7 @@
Copyright (C) 1998 Geoffrey Keating
Copyright (C) 2001 John Hornkvist
- Copyright (C) 2002, 2006, 2007, 2009 Free Software Foundation, Inc.
+ Copyright (C) 2002, 2006, 2007, 2009, 2010 Free Software Foundation, Inc.
FFI support for Darwin and AIX.
@@ -35,11 +35,17 @@
extern void ffi_closure_ASM (void);
enum {
- /* The assembly depends on these exact flags. */
- FLAG_RETURNS_NOTHING = 1 << (31-30), /* These go in cr7 */
- FLAG_RETURNS_FP = 1 << (31-29),
- FLAG_RETURNS_64BITS = 1 << (31-28),
- FLAG_RETURNS_128BITS = 1 << (31-31),
+ /* The assembly depends on these exact flags.
+ For Darwin64 (when FLAG_RETURNS_STRUCT is set):
+ FLAG_RETURNS_FP indicates that the structure embeds FP data.
+ FLAG_RETURNS_128BITS signals a special struct size that is not
+ expanded for float content. */
+ FLAG_RETURNS_128BITS = 1 << (31-31), /* These go in cr7 */
+ FLAG_RETURNS_NOTHING = 1 << (31-30),
+ FLAG_RETURNS_FP = 1 << (31-29),
+ FLAG_RETURNS_64BITS = 1 << (31-28),
+
+ FLAG_RETURNS_STRUCT = 1 << (31-27), /* This goes in cr6 */
FLAG_ARG_NEEDS_COPY = 1 << (31- 7),
FLAG_FP_ARGUMENTS = 1 << (31- 6), /* cr1.eq; specified by ABI */
@@ -50,43 +56,61 @@ enum {
/* About the DARWIN ABI. */
enum {
NUM_GPR_ARG_REGISTERS = 8,
- NUM_FPR_ARG_REGISTERS = 13
+ NUM_FPR_ARG_REGISTERS = 13,
+ LINKAGE_AREA_GPRS = 6
};
-enum { ASM_NEEDS_REGISTERS = 4 };
+
+enum { ASM_NEEDS_REGISTERS = 4 }; /* r28-r31 */
/* ffi_prep_args is called by the assembly routine once stack space
has been allocated for the function's arguments.
+
+ m32/m64
The stack layout we want looks like this:
| Return address from ffi_call_DARWIN | higher addresses
|--------------------------------------------|
- | Previous backchain pointer 4 | stack pointer here
+ | Previous backchain pointer 4/8 | stack pointer here
|--------------------------------------------|<+ <<< on entry to
- | Saved r28-r31 4*4 | | ffi_call_DARWIN
+ | ASM_NEEDS_REGISTERS=r28-r31 4*(4/8) | | ffi_call_DARWIN
|--------------------------------------------| |
- | Parameters (at least 8*4=32) | |
+ | When we have any FP activity... the | |
+ | FPRs occupy NUM_FPR_ARG_REGISTERS slots | |
+ | here fp13 .. fp1 from high to low addr. | |
+ ~ ~ ~
+ | Parameters (at least 8*4/8=32/64) | | NUM_GPR_ARG_REGISTERS
|--------------------------------------------| |
- | Space for GPR2 4 | |
+ | TOC=R2 (AIX) Reserved (Darwin) 4/8 | |
|--------------------------------------------| | stack |
- | Reserved 2*4 | | grows |
+ | Reserved 2*4/8 | | grows |
|--------------------------------------------| | down V
- | Space for callee's LR 4 | |
+ | Space for callee's LR 4/8 | |
|--------------------------------------------| | lower addresses
- | Saved CR 4 | |
+ | Saved CR [low word for m64] 4/8 | |
|--------------------------------------------| | stack pointer here
- | Current backchain pointer 4 |-/ during
+ | Current backchain pointer 4/8 |-/ during
|--------------------------------------------| <<< ffi_call_DARWIN
*/
+#if defined(POWERPC_DARWIN64)
+static void
+darwin64_pass_struct_by_value
+ (ffi_type *, char *, unsigned, unsigned *, double **, unsigned long **);
+#endif
+
+/* This depends on GPR_SIZE = sizeof (unsigned long) */
+
void
ffi_prep_args (extended_cif *ecif, unsigned long *const stack)
{
const unsigned bytes = ecif->cif->bytes;
const unsigned flags = ecif->cif->flags;
const unsigned nargs = ecif->cif->nargs;
+#if !defined(POWERPC_DARWIN64)
const ffi_abi abi = ecif->cif->abi;
+#endif
/* 'stacktop' points at the previous backchain pointer. */
unsigned long *const stacktop = stack + (bytes / sizeof(unsigned long));
@@ -94,18 +118,19 @@ ffi_prep_args (extended_cif *ecif, unsig
/* 'fpr_base' points at the space for fpr1, and grows upwards as
we use FPR registers. */
double *fpr_base = (double *) (stacktop - ASM_NEEDS_REGISTERS) - NUM_FPR_ARG_REGISTERS;
- int fparg_count = 0;
-
+ int gp_count = 0, fparg_count = 0;
/* 'next_arg' grows up as we put parameters in it. */
- unsigned long *next_arg = stack + 6; /* 6 reserved positions. */
+ unsigned long *next_arg = stack + LINKAGE_AREA_GPRS; /* 6 reserved positions. */
int i;
double double_tmp;
void **p_argv = ecif->avalue;
unsigned long gprvalue;
ffi_type** ptr = ecif->cif->arg_types;
+#if !defined(POWERPC_DARWIN64)
char *dest_cpy;
+#endif
unsigned size_al = 0;
/* Check that everything starts aligned properly. */
@@ -130,25 +155,30 @@ ffi_prep_args (extended_cif *ecif, unsig
the size of the floating-point parameter are skipped. */
case FFI_TYPE_FLOAT:
double_tmp = *(float *) *p_argv;
- if (fparg_count >= NUM_FPR_ARG_REGISTERS)
- *(double *)next_arg = double_tmp;
- else
+ if (fparg_count < NUM_FPR_ARG_REGISTERS)
*fpr_base++ = double_tmp;
+#if defined(POWERPC_DARWIN)
+ *(float *)next_arg = *(float *) *p_argv;
+#else
+ *(double *)next_arg = double_tmp;
+#endif
next_arg++;
+ gp_count++;
fparg_count++;
FFI_ASSERT(flags & FLAG_FP_ARGUMENTS);
break;
case FFI_TYPE_DOUBLE:
double_tmp = *(double *) *p_argv;
- if (fparg_count >= NUM_FPR_ARG_REGISTERS)
- *(double *)next_arg = double_tmp;
- else
+ if (fparg_count < NUM_FPR_ARG_REGISTERS)
*fpr_base++ = double_tmp;
+ *(double *)next_arg = double_tmp;
#ifdef POWERPC64
next_arg++;
+ gp_count++;
#else
next_arg += 2;
+ gp_count += 2;
#endif
fparg_count++;
FFI_ASSERT(flags & FLAG_FP_ARGUMENTS);
@@ -157,30 +187,41 @@ ffi_prep_args (extended_cif *ecif, unsig
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
-#ifdef POWERPC64
+# if defined(POWERPC64) && !defined(POWERPC_DARWIN64)
+ /* ??? This will exceed the regs count when the value starts at fp13
+ and it will not put the extra bit on the stack. */
if (fparg_count < NUM_FPR_ARG_REGISTERS)
*(long double *) fpr_base++ = *(long double *) *p_argv;
else
*(long double *) next_arg = *(long double *) *p_argv;
next_arg += 2;
fparg_count += 2;
-#else
+# else
double_tmp = ((double *) *p_argv)[0];
if (fparg_count < NUM_FPR_ARG_REGISTERS)
*fpr_base++ = double_tmp;
- else
- *(double *) next_arg = double_tmp;
+ *(double *) next_arg = double_tmp;
+# if defined(POWERPC_DARWIN64)
+ next_arg++;
+ gp_count++;
+# else
next_arg += 2;
+ gp_count += 2;
+# endif
fparg_count++;
-
double_tmp = ((double *) *p_argv)[1];
if (fparg_count < NUM_FPR_ARG_REGISTERS)
*fpr_base++ = double_tmp;
- else
- *(double *) next_arg = double_tmp;
+ *(double *) next_arg = double_tmp;
+# if defined(POWERPC_DARWIN64)
+ next_arg++;
+ gp_count++;
+# else
next_arg += 2;
+ gp_count += 2;
+# endif
fparg_count++;
-#endif
+# endif
FFI_ASSERT(flags & FLAG_FP_ARGUMENTS);
break;
#endif
@@ -192,6 +233,7 @@ ffi_prep_args (extended_cif *ecif, unsig
#else
*(long long *) next_arg = *(long long *) *p_argv;
next_arg += 2;
+ gp_count += 2;
#endif
break;
case FFI_TYPE_POINTER:
@@ -211,32 +253,35 @@ ffi_prep_args (extended_cif *ecif, unsig
goto putgpr;
case FFI_TYPE_STRUCT:
-#ifdef POWERPC64
- dest_cpy = (char *) next_arg;
size_al = (*ptr)->size;
- if ((*ptr)->elements[0]->type == 3)
+#if defined(POWERPC_DARWIN64)
+ next_arg = (unsigned long *)ALIGN((char *)next_arg, (*ptr)->alignment);
+ darwin64_pass_struct_by_value (*ptr, (char *) *p_argv,
+ (unsigned) size_al,
+ (unsigned int *) &fparg_count,
+ &fpr_base, &next_arg);
+#else
+ dest_cpy = (char *) next_arg;
+
+ /* If the first member of the struct is a double, then include enough
+ padding in the struct size to align it to double-word. */
+ if ((*ptr)->elements[0]->type == FFI_TYPE_DOUBLE)
size_al = ALIGN((*ptr)->size, 8);
- if (size_al < 3 && abi == FFI_DARWIN)
- dest_cpy += 4 - size_al;
+# if defined(POWERPC64)
+ FFI_ASSERT (abi != FFI_DARWIN);
memcpy ((char *) dest_cpy, (char *) *p_argv, size_al);
next_arg += (size_al + 7) / 8;
-#else
- dest_cpy = (char *) next_arg;
-
+# else
/* Structures that match the basic modes (QI 1 byte, HI 2 bytes,
SI 4 bytes) are aligned as if they were those modes.
Structures with 3 byte in size are padded upwards. */
- size_al = (*ptr)->size;
- /* If the first member of the struct is a double, then align
- the struct to double-word. */
- if ((*ptr)->elements[0]->type == FFI_TYPE_DOUBLE)
- size_al = ALIGN((*ptr)->size, 8);
if (size_al < 3 && abi == FFI_DARWIN)
dest_cpy += 4 - size_al;
memcpy((char *) dest_cpy, (char *) *p_argv, size_al);
next_arg += (size_al + 3) / 4;
+# endif
#endif
break;
@@ -249,6 +294,7 @@ ffi_prep_args (extended_cif *ecif, unsig
gprvalue = *(unsigned int *) *p_argv;
putgpr:
*next_arg++ = gprvalue;
+ gp_count++;
break;
default:
break;
@@ -262,8 +308,269 @@ ffi_prep_args (extended_cif *ecif, unsig
//FFI_ASSERT(flags & FLAG_4_GPR_ARGUMENTS || intarg_count <= 4);
}
+#if defined(POWERPC_DARWIN64)
+
+/* See if we can put some of the struct into fprs.
+ This should not be called for structures of size 16 bytes, since these are not
+ broken out this way. */
+static void
+darwin64_scan_struct_for_floats (ffi_type *s, unsigned *nfpr)
+{
+ int i;
+
+ FFI_ASSERT (s->type == FFI_TYPE_STRUCT)
+
+ for (i = 0; s->elements[i] != NULL; i++)
+ {
+ ffi_type *p = s->elements[i];
+ switch (p->type)
+ {
+ case FFI_TYPE_STRUCT:
+ darwin64_scan_struct_for_floats (p, nfpr);
+ break;
+ case FFI_TYPE_LONGDOUBLE:
+ (*nfpr) += 2;
+ break;
+ case FFI_TYPE_DOUBLE:
+ case FFI_TYPE_FLOAT:
+ (*nfpr) += 1;
+ break;
+ default:
+ break;
+ }
+ }
+}
+
+static int
+darwin64_struct_size_exceeds_gprs_p (ffi_type *s, char *src, unsigned *nfpr)
+{
+ unsigned struct_offset=0, i;
+
+ for (i = 0; s->elements[i] != NULL; i++)
+ {
+ char *item_base;
+ ffi_type *p = s->elements[i];
+ /* Find the start of this item (0 for the first one). */
+ if (i > 0)
+ struct_offset = ALIGN(struct_offset, p->alignment);
+
+ item_base = src + struct_offset;
+
+ switch (p->type)
+ {
+ case FFI_TYPE_STRUCT:
+ if (darwin64_struct_size_exceeds_gprs_p (p, item_base, nfpr))
+ return 1;
+ break;
+ case FFI_TYPE_LONGDOUBLE:
+ if (*nfpr >= NUM_FPR_ARG_REGISTERS)
+ return 1;
+ (*nfpr) += 1;
+ item_base += 8;
+ /* FALL THROUGH */
+ case FFI_TYPE_DOUBLE:
+ if (*nfpr >= NUM_FPR_ARG_REGISTERS)
+ return 1;
+ (*nfpr) += 1;
+ break;
+ case FFI_TYPE_FLOAT:
+ if (*nfpr >= NUM_FPR_ARG_REGISTERS)
+ return 1;
+ (*nfpr) += 1;
+ break;
+ default:
+ /* If we try and place any item, that is non-float, once we've
+ exceeded the 8 GPR mark, then we can't fit the struct. */
+ if ((unsigned long)item_base >= 8*8)
+ return 1;
+ break;
+ }
+ /* now count the size of what we just used. */
+ struct_offset += p->size;
+ }
+ return 0;
+}
+
+/* Can this struct be returned by value? */
+int
+darwin64_struct_ret_by_value_p (ffi_type *s)
+{
+ unsigned nfp = 0;
+
+ FFI_ASSERT (s && s->type == FFI_TYPE_STRUCT);
+
+ /* The largest structure we can return is 8long + 13 doubles. */
+ if (s->size > 168)
+ return 0;
+
+ /* We can't pass more than 13 floats. */
+ darwin64_scan_struct_for_floats (s, &nfp);
+ if (nfp > 13)
+ return 0;
+
+ /* If there are not too many floats, and the struct is
+ small enough to accommodate in the GPRs, then it must be OK. */
+ if (s->size <= 64)
+ return 1;
+
+ /* Well, we have to look harder. */
+ nfp = 0;
+ if (darwin64_struct_size_exceeds_gprs_p (s, NULL, &nfp))
+ return 0;
+
+ return 1;
+}
+
+void
+darwin64_pass_struct_floats (ffi_type *s, char *src,
+ unsigned *nfpr, double **fprs)
+{
+ int i;
+ double *fpr_base = *fprs;
+ unsigned struct_offset = 0;
+
+ /* We don't assume anything about the alignment of the source. */
+ for (i = 0; s->elements[i] != NULL; i++)
+ {
+ char *item_base;
+ ffi_type *p = s->elements[i];
+ /* Find the start of this item (0 for the first one). */
+ if (i > 0)
+ struct_offset = ALIGN(struct_offset, p->alignment);
+ item_base = src + struct_offset;
+
+ switch (p->type)
+ {
+ case FFI_TYPE_STRUCT:
+ darwin64_pass_struct_floats (p, item_base, nfpr,
+ &fpr_base);
+ break;
+ case FFI_TYPE_LONGDOUBLE:
+ if (*nfpr < NUM_FPR_ARG_REGISTERS)
+ *fpr_base++ = *(double *)item_base;
+ (*nfpr) += 1;
+ item_base += 8;
+ /* FALL THROUGH */
+ case FFI_TYPE_DOUBLE:
+ if (*nfpr < NUM_FPR_ARG_REGISTERS)
+ *fpr_base++ = *(double *)item_base;
+ (*nfpr) += 1;
+ break;
+ case FFI_TYPE_FLOAT:
+ if (*nfpr < NUM_FPR_ARG_REGISTERS)
+ *fpr_base++ = (double) *(float *)item_base;
+ (*nfpr) += 1;
+ break;
+ default:
+ break;
+ }
+ /* now count the size of what we just used. */
+ struct_offset += p->size;
+ }
+ /* Update the scores. */
+ *fprs = fpr_base;
+}
+
+/* Darwin64 special rules.
+ Break out a struct into params and float registers. */
+static void
+darwin64_pass_struct_by_value (ffi_type *s, char *src, unsigned size,
+ unsigned *nfpr, double **fprs, unsigned long **arg)
+{
+ unsigned long *next_arg = *arg;
+ char *dest_cpy = (char *)next_arg;
+
+ FFI_ASSERT (s->type == FFI_TYPE_STRUCT)
+
+ if (!size)
+ return;
+
+ /* First... special cases. */
+ if (size < 3
+ || (size == 4
+ && s->elements[0]
+ && s->elements[0]->type != FFI_TYPE_FLOAT))
+ {
+ /* Must be at least one GPR, padding is unspecified in value,
+ let's make it zero. */
+ *next_arg = 0UL;
+ dest_cpy += 8 - size;
+ memcpy ((char *) dest_cpy, src, size);
+ next_arg++;
+ }
+ else if (size == 16)
+ {
+ memcpy ((char *) dest_cpy, src, size);
+ next_arg += 2;
+ }
+ else
+ {
+ /* now the general case, we consider embedded floats. */
+ memcpy ((char *) dest_cpy, src, size);
+ darwin64_pass_struct_floats (s, src, nfpr, fprs);
+ next_arg += (size+7)/8;
+ }
+
+ *arg = next_arg;
+}
+
+double *
+darwin64_struct_floats_to_mem (ffi_type *s, char *dest, double *fprs, unsigned *nf)
+{
+ int i;
+ unsigned struct_offset = 0;
+
+ /* We don't assume anything about the alignment of the source. */
+ for (i = 0; s->elements[i] != NULL; i++)
+ {
+ char *item_base;
+ ffi_type *p = s->elements[i];
+ /* Find the start of this item (0 for the first one). */
+ if (i > 0)
+ struct_offset = ALIGN(struct_offset, p->alignment);
+ item_base = dest + struct_offset;
+
+ switch (p->type)
+ {
+ case FFI_TYPE_STRUCT:
+ fprs = darwin64_struct_floats_to_mem (p, item_base, fprs, nf);
+ break;
+ case FFI_TYPE_LONGDOUBLE:
+ if (*nf < NUM_FPR_ARG_REGISTERS)
+ {
+ *(double *)item_base = *fprs++ ;
+ (*nf) += 1;
+ }
+ item_base += 8;
+ /* FALL THROUGH */
+ case FFI_TYPE_DOUBLE:
+ if (*nf < NUM_FPR_ARG_REGISTERS)
+ {
+ *(double *)item_base = *fprs++ ;
+ (*nf) += 1;
+ }
+ break;
+ case FFI_TYPE_FLOAT:
+ if (*nf < NUM_FPR_ARG_REGISTERS)
+ {
+ *(float *)item_base = (float) *fprs++ ;
+ (*nf) += 1;
+ }
+ break;
+ default:
+ break;
+ }
+ /* now count the size of what we just used. */
+ struct_offset += p->size;
+ }
+ return fprs;
+}
+
+#endif
+
/* Adjust the size of S to be correct for Darwin.
- On Darwin, the first field of a structure has natural alignment. */
+ On Darwin m32, the first field of a structure has natural alignment.
+ On Darwin m64, all fields have natural alignment. */
static void
darwin_adjust_aggregate_sizes (ffi_type *s)
@@ -280,22 +587,29 @@ darwin_adjust_aggregate_sizes (ffi_type
int align;
p = s->elements[i];
- darwin_adjust_aggregate_sizes (p);
- if (i == 0
- && (p->type == FFI_TYPE_UINT64
- || p->type == FFI_TYPE_SINT64
- || p->type == FFI_TYPE_DOUBLE
- || p->alignment == 8))
- align = 8;
+ if (p->type == FFI_TYPE_STRUCT)
+ darwin_adjust_aggregate_sizes (p);
+#if defined(POWERPC_DARWIN64)
+ /* Natural alignment for all items. */
+ align = p->alignment;
+#else
+ /* Natrual alignment for the first item... */
+ if (i == 0)
+ align = p->alignment;
else if (p->alignment == 16 || p->alignment < 4)
+ /* .. subsequent items with vector or align < 4 have natural align. */
align = p->alignment;
else
+ /* .. or align is 4. */
align = 4;
+#endif
+ /* Pad, if necessary, before adding the current item. */
s->size = ALIGN(s->size, align) + p->size;
}
s->size = ALIGN(s->size, s->alignment);
+ /* This should not be necessary on m64, but harmless. */
if (s->elements[0]->type == FFI_TYPE_UINT64
|| s->elements[0]->type == FFI_TYPE_SINT64
|| s->elements[0]->type == FFI_TYPE_DOUBLE
@@ -347,7 +661,7 @@ ffi_prep_cif_machdep (ffi_cif *cif)
unsigned i;
ffi_type **ptr;
unsigned bytes;
- int fparg_count = 0, intarg_count = 0;
+ unsigned fparg_count = 0, intarg_count = 0;
unsigned flags = 0;
unsigned size_al = 0;
@@ -372,16 +686,25 @@ ffi_prep_cif_machdep (ffi_cif *cif)
/* Space for the frame pointer, callee's LR, CR, etc, and for
the asm's temp regs. */
- bytes = (6 + ASM_NEEDS_REGISTERS) * sizeof(long);
+ bytes = (LINKAGE_AREA_GPRS + ASM_NEEDS_REGISTERS) * sizeof(unsigned long);
- /* Return value handling. The rules are as follows:
+ /* Return value handling.
+ The rules m32 are as follows:
- 32-bit (or less) integer values are returned in gpr3;
- - Structures of size <= 4 bytes also returned in gpr3;
- - 64-bit integer values and structures between 5 and 8 bytes are returned
- in gpr3 and gpr4;
+ - structures of size <= 4 bytes also returned in gpr3;
+ - 64-bit integer values [??? and structures between 5 and 8 bytes] are
+ returned in gpr3 and gpr4;
- Single/double FP values are returned in fpr1;
- Long double FP (if not equivalent to double) values are returned in
fpr1 and fpr2;
+ m64:
+ - 64-bit or smaller integral values are returned in GPR3
+ - Single/double FP values are returned in fpr1;
+ - Long double FP values are returned in fpr1 and fpr2;
+ m64 Structures:
+ - If the structure could be accommodated in registers were it to be the
+ first argument to a routine, then it is returned in those registers.
+ m32/m64 structures otherwise:
- Larger structures values are allocated space and a pointer is passed
as the first argument. */
switch (cif->rtype->type)
@@ -410,9 +733,42 @@ ffi_prep_cif_machdep (ffi_cif *cif)
break;
case FFI_TYPE_STRUCT:
+#if defined(POWERPC_DARWIN64)
+ {
+ /* Can we fit the struct into regs? */
+ if (darwin64_struct_ret_by_value_p (cif->rtype))
+ {
+ unsigned nfpr = 0;
+ flags |= FLAG_RETURNS_STRUCT;
+ if (cif->rtype->size != 16)
+ darwin64_scan_struct_for_floats (cif->rtype, &nfpr) ;
+ else
+ flags |= FLAG_RETURNS_128BITS;
+ /* Will be 0 for 16byte struct. */
+ if (nfpr)
+ flags |= FLAG_RETURNS_FP;
+ }
+ else /* By ref. */
+ {
+ flags |= FLAG_RETVAL_REFERENCE;
+ flags |= FLAG_RETURNS_NOTHING;
+ intarg_count++;
+ }
+ }
+#elif defined(DARWIN_PPC)
+ if (cif->rtype->size <= 4)
+ flags |= FLAG_RETURNS_STRUCT;
+ else /* else by reference. */
+ {
+ flags |= FLAG_RETVAL_REFERENCE;
+ flags |= FLAG_RETURNS_NOTHING;
+ intarg_count++;
+ }
+#else /* assume we pass by ref. */
flags |= FLAG_RETVAL_REFERENCE;
flags |= FLAG_RETURNS_NOTHING;
intarg_count++;
+#endif
break;
case FFI_TYPE_VOID:
flags |= FLAG_RETURNS_NOTHING;
@@ -425,57 +781,83 @@ ffi_prep_cif_machdep (ffi_cif *cif)
/* The first NUM_GPR_ARG_REGISTERS words of integer arguments, and the
first NUM_FPR_ARG_REGISTERS fp arguments, go in registers; the rest
- goes on the stack. Structures are passed as a pointer to a copy of
- the structure. Stuff on the stack needs to keep proper alignment. */
+ goes on the stack.
+ ??? Structures are passed as a pointer to a copy of the structure.
+ Stuff on the stack needs to keep proper alignment.
+ For m64 the count is effectively of half-GPRs. */
for (ptr = cif->arg_types, i = cif->nargs; i > 0; i--, ptr++)
{
+ unsigned align_words;
switch ((*ptr)->type)
{
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
fparg_count++;
+#if !defined(POWERPC_DARWIN64)
/* If this FP arg is going on the stack, it must be
8-byte-aligned. */
if (fparg_count > NUM_FPR_ARG_REGISTERS
- && intarg_count%2 != 0)
+ && (intarg_count & 0x01) != 0)
intarg_count++;
+#endif
break;
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
-
case FFI_TYPE_LONGDOUBLE:
fparg_count += 2;
/* If this FP arg is going on the stack, it must be
- 8-byte-aligned. */
- if (fparg_count > NUM_FPR_ARG_REGISTERS
- && intarg_count%2 != 0)
- intarg_count++;
- intarg_count +=2;
+ 16-byte-aligned. */
+ if (fparg_count >= NUM_FPR_ARG_REGISTERS)
+#if defined (POWERPC64)
+ intarg_count = ALIGN(intarg_count, 2);
+#else
+ intarg_count = ALIGN(intarg_count, 4);
+#endif
break;
#endif
case FFI_TYPE_UINT64:
case FFI_TYPE_SINT64:
+#if defined(POWERPC64)
+ intarg_count++;
+#else
/* 'long long' arguments are passed as two words, but
either both words must fit in registers or both go
on the stack. If they go on the stack, they must
be 8-byte-aligned. */
if (intarg_count == NUM_GPR_ARG_REGISTERS-1
- || (intarg_count >= NUM_GPR_ARG_REGISTERS && intarg_count%2 != 0))
+ || (intarg_count >= NUM_GPR_ARG_REGISTERS
+ && (intarg_count & 0x01) != 0))
intarg_count++;
intarg_count += 2;
+#endif
break;
case FFI_TYPE_STRUCT:
size_al = (*ptr)->size;
+#if defined(POWERPC_DARWIN64)
+ align_words = (*ptr)->alignment >> 3;
+ if (align_words)
+ intarg_count = ALIGN(intarg_count, align_words);
+ /* Base size of the struct. */
+ intarg_count += (size_al + 7) / 8;
+ /* If 16 bytes then don't worry about floats. */
+ if (size_al != 16)
+ /* Scan through for floats to be placed in regs. */
+ darwin64_scan_struct_for_floats (*ptr, &fparg_count) ;
+#else
+ align_words = (*ptr)->alignment >> 2;
+ if (align_words)
+ intarg_count = ALIGN(intarg_count, align_words);
/* If the first member of the struct is a double, then align
- the struct to double-word. */
+ the struct to double-word.
if ((*ptr)->elements[0]->type == FFI_TYPE_DOUBLE)
- size_al = ALIGN((*ptr)->size, 8);
-#ifdef POWERPC64
+ size_al = ALIGN((*ptr)->size, 8); */
+# ifdef POWERPC64
intarg_count += (size_al + 7) / 8;
-#else
+# else
intarg_count += (size_al + 3) / 4;
+# endif
#endif
break;
@@ -490,9 +872,18 @@ ffi_prep_cif_machdep (ffi_cif *cif)
if (fparg_count != 0)
flags |= FLAG_FP_ARGUMENTS;
+#if defined(POWERPC_DARWIN64)
+ /* Space to image the FPR registers, if needed - which includes when they might be
+ used in a struct return. */
+ if (fparg_count != 0
+ || ((flags & FLAG_RETURNS_STRUCT)
+ && (flags & FLAG_RETURNS_FP)))
+ bytes += NUM_FPR_ARG_REGISTERS * sizeof(double);
+#else
/* Space for the FPR registers, if needed. */
if (fparg_count != 0)
bytes += NUM_FPR_ARG_REGISTERS * sizeof(double);
+#endif
/* Stack space. */
#ifdef POWERPC64
@@ -506,7 +897,7 @@ ffi_prep_cif_machdep (ffi_cif *cif)
bytes += NUM_GPR_ARG_REGISTERS * sizeof(long);
/* The stack space allocated needs to be a multiple of 16 bytes. */
- bytes = (bytes + 15) & ~0xF;
+ bytes = ALIGN(bytes, 16) ;
cif->flags = flags;
cif->bytes = bytes;
@@ -516,8 +907,9 @@ ffi_prep_cif_machdep (ffi_cif *cif)
extern void ffi_call_AIX(extended_cif *, long, unsigned, unsigned *,
void (*fn)(void), void (*fn2)(void));
+
extern void ffi_call_DARWIN(extended_cif *, long, unsigned, unsigned *,
- void (*fn)(void), void (*fn2)(void));
+ void (*fn)(void), void (*fn2)(void), ffi_type*);
void
ffi_call (ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue)
@@ -546,7 +938,7 @@ ffi_call (ffi_cif *cif, void (*fn)(void)
break;
case FFI_DARWIN:
ffi_call_DARWIN(&ecif, -(long)cif->bytes, cif->flags, ecif.rvalue, fn,
- FFI_FN(ffi_prep_args));
+ FFI_FN(ffi_prep_args), cif->rtype);
break;
default:
FFI_ASSERT(0);
@@ -566,58 +958,48 @@ typedef struct aix_fd_struct {
} aix_fd;
/* here I'd like to add the stack frame layout we use in darwin_closure.S
- and aix_clsoure.S
+ and aix_closure.S
+
+ m32/m64
- SP previous -> +---------------------------------------+ <--- child frame
- | back chain to caller 4 |
- +---------------------------------------+ 4
- | saved CR 4 |
- +---------------------------------------+ 8
- | saved LR 4 |
- +---------------------------------------+ 12
- | reserved for compilers 4 |
- +---------------------------------------+ 16
- | reserved for binders 4 |
- +---------------------------------------+ 20
- | saved TOC pointer 4 |
- +---------------------------------------+ 24
- | always reserved 8*4=32 (previous GPRs)|
- | according to the linkage convention |
- | from AIX |
- +---------------------------------------+ 56
- | our FPR area 13*8=104 |
- | f1 |
- | . |
- | f13 |
- +---------------------------------------+ 160
- | result area 8 |
- +---------------------------------------+ 168
- | alignement to the next multiple of 16 |
-SP current --> +---------------------------------------+ 176 <- parent frame
- | back chain to caller 4 |
- +---------------------------------------+ 180
- | saved CR 4 |
- +---------------------------------------+ 184
- | saved LR 4 |
- +---------------------------------------+ 188
- | reserved for compilers 4 |
- +---------------------------------------+ 192
- | reserved for binders 4 |
- +---------------------------------------+ 196
- | saved TOC pointer 4 |
- +---------------------------------------+ 200
- | always reserved 8*4=32 we store our |
- | GPRs here |
- | r3 |
- | . |
- | r10 |
- +---------------------------------------+ 232
- | overflow part |
- +---------------------------------------+ xxx
- | ???? |
- +---------------------------------------+ xxx
+ The stack layout looks like this:
+
+ | Additional params... | | Higher address
+ ~ ~ ~
+ | Parameters (at least 8*4/8=32/64) | | NUM_GPR_ARG_REGISTERS
+ |--------------------------------------------| |
+ | TOC=R2 (AIX) Reserved (Darwin) 4/8 | |
+ |--------------------------------------------| |
+ | Reserved 2*4/8 | |
+ |--------------------------------------------| |
+ | Space for callee's LR 4/8 | |
+ |--------------------------------------------| |
+ | Saved CR [low word for m64] 4/8 | |
+ |--------------------------------------------| |
+ | Current backchain pointer 4/8 |-/ Parent's frame.
+ |--------------------------------------------| <+ <<< on entry to ffi_closure_ASM
+ | Result Bytes 16 | |
+ |--------------------------------------------| |
+ ~ padding to 16-byte alignment ~ ~
+ |--------------------------------------------| |
+ | NUM_FPR_ARG_REGISTERS slots | |
+ | here fp13 .. fp1 13*8 | |
+ |--------------------------------------------| |
+ | R3..R10 8*4/8=32/64 | | NUM_GPR_ARG_REGISTERS
+ |--------------------------------------------| |
+ | TOC=R2 (AIX) Reserved (Darwin) 4/8 | |
+ |--------------------------------------------| | stack |
+ | Reserved [compiler,binder] 2*4/8 | | grows |
+ |--------------------------------------------| | down V
+ | Space for callee's LR 4/8 | |
+ |--------------------------------------------| | lower addresses
+ | Saved CR [low word for m64] 4/8 | |
+ |--------------------------------------------| | stack pointer here
+ | Current backchain pointer 4/8 |-/ during
+ |--------------------------------------------| <<< ffi_closure_ASM.
*/
+
ffi_status
ffi_prep_closure_loc (ffi_closure* closure,
ffi_cif* cif,
@@ -631,30 +1013,44 @@ ffi_prep_closure_loc (ffi_closure* closu
switch (cif->abi)
{
- case FFI_DARWIN:
+ case FFI_DARWIN:
- FFI_ASSERT (cif->abi == FFI_DARWIN);
+ FFI_ASSERT (cif->abi == FFI_DARWIN);
- tramp = (unsigned int *) &closure->tramp[0];
- tramp[0] = 0x7c0802a6; /* mflr r0 */
- tramp[1] = 0x429f000d; /* bcl- 20,4*cr7+so,0x10 */
- tramp[4] = 0x7d6802a6; /* mflr r11 */
- tramp[5] = 0x818b0000; /* lwz r12,0(r11) function address */
- tramp[6] = 0x7c0803a6; /* mtlr r0 */
- tramp[7] = 0x7d8903a6; /* mtctr r12 */
- tramp[8] = 0x816b0004; /* lwz r11,4(r11) static chain */
- tramp[9] = 0x4e800420; /* bctr */
- tramp[2] = (unsigned long) ffi_closure_ASM; /* function */
- tramp[3] = (unsigned long) codeloc; /* context */
+ tramp = (unsigned int *) &closure->tramp[0];
+#if defined(POWERPC_DARWIN64)
+ tramp[0] = 0x7c0802a6; /* mflr r0 */
+ tramp[1] = 0x429f0015; /* bcl- 20,4*cr7+so, +0x18 (L1) */
+ /* We put the addresses here. */
+ tramp[6] = 0x7d6802a6; /*L1: mflr r11 */
+ tramp[7] = 0xe98b0000; /* ld r12,0(r11) function address */
+ tramp[8] = 0x7c0803a6; /* mtlr r0 */
+ tramp[9] = 0x7d8903a6; /* mtctr r12 */
+ tramp[10] = 0xe96b0008; /* lwz r11,8(r11) static chain */
+ tramp[11] = 0x4e800420; /* bctr */
- closure->cif = cif;
- closure->fun = fun;
- closure->user_data = user_data;
+ *((unsigned long *)&tramp[2]) = (unsigned long) ffi_closure_ASM; /* function */
+ *((unsigned long *)&tramp[4]) = (unsigned long) codeloc; /* context */
+#else
+ tramp[0] = 0x7c0802a6; /* mflr r0 */
+ tramp[1] = 0x429f000d; /* bcl- 20,4*cr7+so,0x10 */
+ tramp[4] = 0x7d6802a6; /* mflr r11 */
+ tramp[5] = 0x818b0000; /* lwz r12,0(r11) function address */
+ tramp[6] = 0x7c0803a6; /* mtlr r0 */
+ tramp[7] = 0x7d8903a6; /* mtctr r12 */
+ tramp[8] = 0x816b0004; /* lwz r11,4(r11) static chain */
+ tramp[9] = 0x4e800420; /* bctr */
+ tramp[2] = (unsigned long) ffi_closure_ASM; /* function */
+ tramp[3] = (unsigned long) codeloc; /* context */
+#endif
+ closure->cif = cif;
+ closure->fun = fun;
+ closure->user_data = user_data;
- /* Flush the icache. Only necessary on Darwin. */
- flush_range(codeloc, FFI_TRAMPOLINE_SIZE);
+ /* Flush the icache. Only necessary on Darwin. */
+ flush_range(codeloc, FFI_TRAMPOLINE_SIZE);
- break;
+ break;
case FFI_AIX:
@@ -669,10 +1065,10 @@ ffi_prep_closure_loc (ffi_closure* closu
closure->cif = cif;
closure->fun = fun;
closure->user_data = user_data;
+ break;
default:
-
- FFI_ASSERT(0);
+ return FFI_BAD_ABI;
break;
}
return FFI_OK;
@@ -708,7 +1104,7 @@ typedef union
double d;
} ffi_dblfl;
-int
+ffi_type *
ffi_closure_helper_DARWIN (ffi_closure *, void *,
unsigned long *, ffi_dblfl *);
@@ -719,7 +1115,7 @@ ffi_closure_helper_DARWIN (ffi_closure *
up space for a return value, ffi_closure_ASM invokes the
following helper function to do most of the work. */
-int
+ffi_type *
ffi_closure_helper_DARWIN (ffi_closure *closure, void *rvalue,
unsigned long *pgr, ffi_dblfl *pfr)
{
@@ -741,16 +1137,32 @@ ffi_closure_helper_DARWIN (ffi_closure *
ffi_cif * cif;
ffi_dblfl * end_pfr = pfr + NUM_FPR_ARG_REGISTERS;
unsigned size_al;
+#if defined(POWERPC_DARWIN64)
+ unsigned fpsused = 0;
+#endif
cif = closure->cif;
avalue = alloca (cif->nargs * sizeof(void *));
- /* Copy the caller's structure return value address so that the closure
- returns the data directly to the caller. */
if (cif->rtype->type == FFI_TYPE_STRUCT)
{
+#if defined(POWERPC_DARWIN64)
+ if (!darwin64_struct_ret_by_value_p (cif->rtype))
+ {
+ /* Won't fit into the regs - return by ref. */
+ rvalue = (void *) *pgr;
+ pgr++;
+ }
+#elif defined(DARWIN_PPC)
+ if (cif->rtype->size > 4)
+ {
+ rvalue = (void *) *pgr;
+ pgr++;
+ }
+#else /* assume we return by ref. */
rvalue = (void *) *pgr;
pgr++;
+#endif
}
i = 0;
@@ -764,7 +1176,7 @@ ffi_closure_helper_DARWIN (ffi_closure *
{
case FFI_TYPE_SINT8:
case FFI_TYPE_UINT8:
-#ifdef POWERPC64
+#if defined(POWERPC64)
avalue[i] = (char *) pgr + 7;
#else
avalue[i] = (char *) pgr + 3;
@@ -774,7 +1186,7 @@ ffi_closure_helper_DARWIN (ffi_closure *
case FFI_TYPE_SINT16:
case FFI_TYPE_UINT16:
-#ifdef POWERPC64
+#if defined(POWERPC64)
avalue[i] = (char *) pgr + 6;
#else
avalue[i] = (char *) pgr + 2;
@@ -784,7 +1196,7 @@ ffi_closure_helper_DARWIN (ffi_closure *
case FFI_TYPE_SINT32:
case FFI_TYPE_UINT32:
-#ifdef POWERPC64
+#if defined(POWERPC64)
avalue[i] = (char *) pgr + 4;
#else
case FFI_TYPE_POINTER:
@@ -794,34 +1206,53 @@ ffi_closure_helper_DARWIN (ffi_closure *
break;
case FFI_TYPE_STRUCT:
-#ifdef POWERPC64
size_al = arg_types[i]->size;
- if (arg_types[i]->elements[0]->type == FFI_TYPE_DOUBLE)
- size_al = ALIGN (arg_types[i]->size, 8);
- if (size_al < 3 && cif->abi == FFI_DARWIN)
- avalue[i] = (char *) pgr + 8 - size_al;
- else
- avalue[i] = pgr;
+#if defined(POWERPC_DARWIN64)
+ pgr = (unsigned long *)ALIGN((char *)pgr, arg_types[i]->alignment);
+ if (size_al < 3 || size_al == 4)
+ {
+ avalue[i] = ((char *)pgr)+8-size_al;
+ if (arg_types[i]->elements[0]->type == FFI_TYPE_FLOAT
+ && fpsused < NUM_FPR_ARG_REGISTERS)
+ {
+ *(float *)pgr = (float) *(double *)pfr;
+ pfr++;
+ fpsused++;
+ }
+ }
+ else
+ {
+ if (size_al != 16)
+ pfr = (ffi_dblfl *)
+ darwin64_struct_floats_to_mem (arg_types[i], (char *)pgr,
+ (double *)pfr, &fpsused);
+ avalue[i] = pgr;
+ }
pgr += (size_al + 7) / 8;
#else
- /* Structures that match the basic modes (QI 1 byte, HI 2 bytes,
- SI 4 bytes) are aligned as if they were those modes. */
- size_al = arg_types[i]->size;
/* If the first member of the struct is a double, then align
the struct to double-word. */
if (arg_types[i]->elements[0]->type == FFI_TYPE_DOUBLE)
size_al = ALIGN(arg_types[i]->size, 8);
+# if defined(POWERPC64)
+ FFI_ASSERT (cif->abi != FFI_DARWIN);
+ avalue[i] = pgr;
+ pgr += (size_al + 7) / 8;
+# else
+ /* Structures that match the basic modes (QI 1 byte, HI 2 bytes,
+ SI 4 bytes) are aligned as if they were those modes. */
if (size_al < 3 && cif->abi == FFI_DARWIN)
avalue[i] = (char*) pgr + 4 - size_al;
else
avalue[i] = pgr;
pgr += (size_al + 3) / 4;
+# endif
#endif
break;
case FFI_TYPE_SINT64:
case FFI_TYPE_UINT64:
-#ifdef POWERPC64
+#if defined(POWERPC64)
case FFI_TYPE_POINTER:
avalue[i] = pgr;
pgr++;
@@ -924,5 +1355,5 @@ ffi_closure_helper_DARWIN (ffi_closure *
(closure->fun) (cif, rvalue, avalue, closure->user_data);
/* Tell ffi_closure_ASM to perform return type promotions. */
- return cif->rtype->type;
+ return cif->rtype;
}
Index: mozilla/js/src/ctypes/libffi/src/powerpc/ffitarget.h
===================================================================
--- mozilla.orig/js/src/ctypes/libffi/src/powerpc/ffitarget.h
+++ mozilla/js/src/ctypes/libffi/src/powerpc/ffitarget.h
@@ -1,6 +1,8 @@
/* -----------------------------------------------------------------*-C-*-
- ffitarget.h - Copyright (c) 1996-2003 Red Hat, Inc.
- Copyright (C) 2007, 2008 Free Software Foundation, Inc
+ ffitarget.h - Copyright (c) 2012 Anthony Green
+ Copyright (C) 2007, 2008, 2010 Free Software Foundation, Inc
+ Copyright (c) 1996-2003 Red Hat, Inc.
+
Target configuration macros for PowerPC.
Permission is hereby granted, free of charge, to any person obtaining
@@ -28,16 +30,23 @@
#ifndef LIBFFI_TARGET_H
#define LIBFFI_TARGET_H
+#ifndef LIBFFI_H
+#error "Please do not include ffitarget.h directly into your source. Use ffi.h instead."
+#endif
+
/* ---- System specific configurations ----------------------------------- */
#if defined (POWERPC) && defined (__powerpc64__) /* linux64 */
#ifndef POWERPC64
#define POWERPC64
#endif
-#elif defined (POWERPC_DARWIN) && defined (__ppc64__) /* Darwin */
+#elif defined (POWERPC_DARWIN) && defined (__ppc64__) /* Darwin64 */
#ifndef POWERPC64
#define POWERPC64
#endif
+#ifndef POWERPC_DARWIN64
+#define POWERPC_DARWIN64
+#endif
#elif defined (POWERPC_AIX) && defined (__64BIT__) /* AIX64 */
#ifndef POWERPC64
#define POWERPC64
@@ -57,18 +66,14 @@ typedef enum ffi_abi {
FFI_LINUX64,
FFI_LINUX,
FFI_LINUX_SOFT_FLOAT,
-# ifdef POWERPC64
+# if defined(POWERPC64)
FFI_DEFAULT_ABI = FFI_LINUX64,
-# else
-# if (!defined(__NO_FPRS__) && (__LDBL_MANT_DIG__ == 106))
- FFI_DEFAULT_ABI = FFI_LINUX,
-# else
-# ifdef __NO_FPRS__
+# elif defined(__NO_FPRS__)
FFI_DEFAULT_ABI = FFI_LINUX_SOFT_FLOAT,
-# else
+# elif (__LDBL_MANT_DIG__ == 106)
+ FFI_DEFAULT_ABI = FFI_LINUX,
+# else
FFI_DEFAULT_ABI = FFI_GCC_SYSV,
-# endif
-# endif
# endif
#endif
@@ -101,6 +106,10 @@ typedef enum ffi_abi {
#define FFI_CLOSURES 1
#define FFI_NATIVE_RAW_API 0
+#if defined (POWERPC) || defined (POWERPC_FREEBSD)
+# define FFI_TARGET_SPECIFIC_VARIADIC 1
+# define FFI_EXTRA_CIF_FIELDS unsigned nfixedargs
+#endif
/* For additional types like the below, take care about the order in
ppc_closures.S. They must follow after the FFI_TYPE_LAST. */
@@ -113,10 +122,23 @@ typedef enum ffi_abi {
defined in ffi.c, to determine the exact return type and its size. */
#define FFI_SYSV_TYPE_SMALL_STRUCT (FFI_TYPE_LAST + 2)
-#if defined(POWERPC64) || defined(POWERPC_AIX)
-#define FFI_TRAMPOLINE_SIZE 24
-#else /* POWERPC || POWERPC_AIX */
-#define FFI_TRAMPOLINE_SIZE 40
+/* Used by ELFv2 for homogenous structure returns. */
+#define FFI_V2_TYPE_FLOAT_HOMOG (FFI_TYPE_LAST + 1)
+#define FFI_V2_TYPE_DOUBLE_HOMOG (FFI_TYPE_LAST + 2)
+#define FFI_V2_TYPE_SMALL_STRUCT (FFI_TYPE_LAST + 3)
+
+#if _CALL_ELF == 2
+# define FFI_TRAMPOLINE_SIZE 32
+#else
+# if defined(POWERPC64) || defined(POWERPC_AIX)
+# if defined(POWERPC_DARWIN64)
+# define FFI_TRAMPOLINE_SIZE 48
+# else
+# define FFI_TRAMPOLINE_SIZE 24
+# endif
+# else /* POWERPC || POWERPC_AIX */
+# define FFI_TRAMPOLINE_SIZE 40
+# endif
#endif
#ifndef LIBFFI_ASM
Index: mozilla/js/src/ctypes/libffi/src/powerpc/linux64_closure.S
===================================================================
--- mozilla.orig/js/src/ctypes/libffi/src/powerpc/linux64_closure.S
+++ mozilla/js/src/ctypes/libffi/src/powerpc/linux64_closure.S
@@ -32,66 +32,127 @@
#ifdef __powerpc64__
FFI_HIDDEN (ffi_closure_LINUX64)
- FFI_HIDDEN (.ffi_closure_LINUX64)
- .globl ffi_closure_LINUX64, .ffi_closure_LINUX64
+ .globl ffi_closure_LINUX64
+# if _CALL_ELF == 2
+ .text
+ffi_closure_LINUX64:
+ addis %r2, %r12, .TOC.-ffi_closure_LINUX64@ha
+ addi %r2, %r2, .TOC.-ffi_closure_LINUX64@l
+ .localentry ffi_closure_LINUX64, . - ffi_closure_LINUX64
+# else
.section ".opd","aw"
.align 3
ffi_closure_LINUX64:
+# ifdef _CALL_LINUX
+ .quad .L.ffi_closure_LINUX64,.TOC.@tocbase,0
+ .type ffi_closure_LINUX64,@function
+ .text
+.L.ffi_closure_LINUX64:
+# else
+ FFI_HIDDEN (.ffi_closure_LINUX64)
+ .globl .ffi_closure_LINUX64
.quad .ffi_closure_LINUX64,.TOC.@tocbase,0
.size ffi_closure_LINUX64,24
.type .ffi_closure_LINUX64,@function
.text
.ffi_closure_LINUX64:
+# endif
+# endif
+
+# if _CALL_ELF == 2
+# 32 byte special reg save area + 64 byte parm save area and retval
+# + 13*8 fpr save area + round to 16
+# define STACKFRAME 208
+# define PARMSAVE 32
+# No parameter save area is needed for the call to ffi_closure_helper_LINUX64,
+# so return value can start there.
+# define RETVAL PARMSAVE
+# else
+# 48 bytes special reg save area + 64 bytes parm save area
+# + 16 bytes retval area + 13*8 bytes fpr save area + round to 16
+# define STACKFRAME 240
+# define PARMSAVE 48
+# define RETVAL PARMSAVE+64
+# endif
+
.LFB1:
- # save general regs into parm save area
- std %r3, 48(%r1)
- std %r4, 56(%r1)
- std %r5, 64(%r1)
- std %r6, 72(%r1)
+# if _CALL_ELF == 2
+ ld %r12, FFI_TRAMPOLINE_SIZE(%r11) # closure->cif
+ mflr %r0
+ lwz %r12, 28(%r12) # cif->flags
+ mtcrf 0x40, %r12
+ addi %r12, %r1, PARMSAVE
+ bt 7, .Lparmsave
+ # Our caller has not allocated a parameter save area.
+ # We need to allocate one here and use it to pass gprs to
+ # ffi_closure_helper_LINUX64. The return value area will do.
+ addi %r12, %r1, -STACKFRAME+RETVAL
+.Lparmsave:
+ std %r0, 16(%r1)
+ # Save general regs into parm save area
+ std %r3, 0(%r12)
+ std %r4, 8(%r12)
+ std %r5, 16(%r12)
+ std %r6, 24(%r12)
+ std %r7, 32(%r12)
+ std %r8, 40(%r12)
+ std %r9, 48(%r12)
+ std %r10, 56(%r12)
+
+ # load up the pointer to the parm save area
+ mr %r5, %r12
+# else
mflr %r0
+ # Save general regs into parm save area
+ # This is the parameter save area set up by our caller.
+ std %r3, PARMSAVE+0(%r1)
+ std %r4, PARMSAVE+8(%r1)
+ std %r5, PARMSAVE+16(%r1)
+ std %r6, PARMSAVE+24(%r1)
+ std %r7, PARMSAVE+32(%r1)
+ std %r8, PARMSAVE+40(%r1)
+ std %r9, PARMSAVE+48(%r1)
+ std %r10, PARMSAVE+56(%r1)
- std %r7, 80(%r1)
- std %r8, 88(%r1)
- std %r9, 96(%r1)
- std %r10, 104(%r1)
std %r0, 16(%r1)
- # mandatory 48 bytes special reg save area + 64 bytes parm save area
- # + 16 bytes retval area + 13*8 bytes fpr save area + round to 16
- stdu %r1, -240(%r1)
-.LCFI0:
+ # load up the pointer to the parm save area
+ addi %r5, %r1, PARMSAVE
+# endif
# next save fpr 1 to fpr 13
- stfd %f1, 128+(0*8)(%r1)
- stfd %f2, 128+(1*8)(%r1)
- stfd %f3, 128+(2*8)(%r1)
- stfd %f4, 128+(3*8)(%r1)
- stfd %f5, 128+(4*8)(%r1)
- stfd %f6, 128+(5*8)(%r1)
- stfd %f7, 128+(6*8)(%r1)
- stfd %f8, 128+(7*8)(%r1)
- stfd %f9, 128+(8*8)(%r1)
- stfd %f10, 128+(9*8)(%r1)
- stfd %f11, 128+(10*8)(%r1)
- stfd %f12, 128+(11*8)(%r1)
- stfd %f13, 128+(12*8)(%r1)
+ stfd %f1, -104+(0*8)(%r1)
+ stfd %f2, -104+(1*8)(%r1)
+ stfd %f3, -104+(2*8)(%r1)
+ stfd %f4, -104+(3*8)(%r1)
+ stfd %f5, -104+(4*8)(%r1)
+ stfd %f6, -104+(5*8)(%r1)
+ stfd %f7, -104+(6*8)(%r1)
+ stfd %f8, -104+(7*8)(%r1)
+ stfd %f9, -104+(8*8)(%r1)
+ stfd %f10, -104+(9*8)(%r1)
+ stfd %f11, -104+(10*8)(%r1)
+ stfd %f12, -104+(11*8)(%r1)
+ stfd %f13, -104+(12*8)(%r1)
- # set up registers for the routine that actually does the work
- # get the context pointer from the trampoline
- mr %r3, %r11
+ # load up the pointer to the saved fpr registers */
+ addi %r6, %r1, -104
- # now load up the pointer to the result storage
- addi %r4, %r1, 112
+ # load up the pointer to the result storage
+ addi %r4, %r1, -STACKFRAME+RETVAL
- # now load up the pointer to the parameter save area
- # in the previous frame
- addi %r5, %r1, 240 + 48
+ stdu %r1, -STACKFRAME(%r1)
+.LCFI0:
- # now load up the pointer to the saved fpr registers */
- addi %r6, %r1, 128
+ # get the context pointer from the trampoline
+ mr %r3, %r11
# make the call
+# if defined _CALL_LINUX || _CALL_ELF == 2
+ bl ffi_closure_helper_LINUX64
+# else
bl .ffi_closure_helper_LINUX64
+# endif
.Lret:
# now r3 contains the return type
@@ -100,10 +161,12 @@ ffi_closure_LINUX64:
# look up the proper starting point in table
# by using return type as offset
+ ld %r0, STACKFRAME+16(%r1)
+ cmpldi %r3, FFI_V2_TYPE_SMALL_STRUCT
+ bge .Lsmall
mflr %r4 # move address of .Lret to r4
sldi %r3, %r3, 4 # now multiply return type by 16
addi %r4, %r4, .Lret_type0 - .Lret
- ld %r0, 240+16(%r1)
add %r3, %r3, %r4 # add contents of table to table address
mtctr %r3
bctr # jump to it
@@ -116,85 +179,175 @@ ffi_closure_LINUX64:
.Lret_type0:
# case FFI_TYPE_VOID
mtlr %r0
- addi %r1, %r1, 240
+ addi %r1, %r1, STACKFRAME
blr
nop
# case FFI_TYPE_INT
- lwa %r3, 112+4(%r1)
+# ifdef __LITTLE_ENDIAN__
+ lwa %r3, RETVAL+0(%r1)
+# else
+ lwa %r3, RETVAL+4(%r1)
+# endif
mtlr %r0
- addi %r1, %r1, 240
+ addi %r1, %r1, STACKFRAME
blr
# case FFI_TYPE_FLOAT
- lfs %f1, 112+0(%r1)
+ lfs %f1, RETVAL+0(%r1)
mtlr %r0
- addi %r1, %r1, 240
+ addi %r1, %r1, STACKFRAME
blr
# case FFI_TYPE_DOUBLE
- lfd %f1, 112+0(%r1)
+ lfd %f1, RETVAL+0(%r1)
mtlr %r0
- addi %r1, %r1, 240
+ addi %r1, %r1, STACKFRAME
blr
# case FFI_TYPE_LONGDOUBLE
- lfd %f1, 112+0(%r1)
+ lfd %f1, RETVAL+0(%r1)
mtlr %r0
- lfd %f2, 112+8(%r1)
+ lfd %f2, RETVAL+8(%r1)
b .Lfinish
# case FFI_TYPE_UINT8
- lbz %r3, 112+7(%r1)
+# ifdef __LITTLE_ENDIAN__
+ lbz %r3, RETVAL+0(%r1)
+# else
+ lbz %r3, RETVAL+7(%r1)
+# endif
mtlr %r0
- addi %r1, %r1, 240
+ addi %r1, %r1, STACKFRAME
blr
# case FFI_TYPE_SINT8
- lbz %r3, 112+7(%r1)
+# ifdef __LITTLE_ENDIAN__
+ lbz %r3, RETVAL+0(%r1)
+# else
+ lbz %r3, RETVAL+7(%r1)
+# endif
extsb %r3,%r3
mtlr %r0
b .Lfinish
# case FFI_TYPE_UINT16
- lhz %r3, 112+6(%r1)
+# ifdef __LITTLE_ENDIAN__
+ lhz %r3, RETVAL+0(%r1)
+# else
+ lhz %r3, RETVAL+6(%r1)
+# endif
mtlr %r0
.Lfinish:
- addi %r1, %r1, 240
+ addi %r1, %r1, STACKFRAME
blr
# case FFI_TYPE_SINT16
- lha %r3, 112+6(%r1)
+# ifdef __LITTLE_ENDIAN__
+ lha %r3, RETVAL+0(%r1)
+# else
+ lha %r3, RETVAL+6(%r1)
+# endif
mtlr %r0
- addi %r1, %r1, 240
+ addi %r1, %r1, STACKFRAME
blr
# case FFI_TYPE_UINT32
- lwz %r3, 112+4(%r1)
+# ifdef __LITTLE_ENDIAN__
+ lwz %r3, RETVAL+0(%r1)
+# else
+ lwz %r3, RETVAL+4(%r1)
+# endif
mtlr %r0
- addi %r1, %r1, 240
+ addi %r1, %r1, STACKFRAME
blr
# case FFI_TYPE_SINT32
- lwa %r3, 112+4(%r1)
+# ifdef __LITTLE_ENDIAN__
+ lwa %r3, RETVAL+0(%r1)
+# else
+ lwa %r3, RETVAL+4(%r1)
+# endif
mtlr %r0
- addi %r1, %r1, 240
+ addi %r1, %r1, STACKFRAME
blr
# case FFI_TYPE_UINT64
- ld %r3, 112+0(%r1)
+ ld %r3, RETVAL+0(%r1)
mtlr %r0
- addi %r1, %r1, 240
+ addi %r1, %r1, STACKFRAME
blr
# case FFI_TYPE_SINT64
- ld %r3, 112+0(%r1)
+ ld %r3, RETVAL+0(%r1)
mtlr %r0
- addi %r1, %r1, 240
+ addi %r1, %r1, STACKFRAME
blr
# case FFI_TYPE_STRUCT
mtlr %r0
- addi %r1, %r1, 240
+ addi %r1, %r1, STACKFRAME
blr
nop
# case FFI_TYPE_POINTER
- ld %r3, 112+0(%r1)
+ ld %r3, RETVAL+0(%r1)
mtlr %r0
- addi %r1, %r1, 240
+ addi %r1, %r1, STACKFRAME
+ blr
+# case FFI_V2_TYPE_FLOAT_HOMOG
+ lfs %f1, RETVAL+0(%r1)
+ lfs %f2, RETVAL+4(%r1)
+ lfs %f3, RETVAL+8(%r1)
+ b .Lmorefloat
+# case FFI_V2_TYPE_DOUBLE_HOMOG
+ lfd %f1, RETVAL+0(%r1)
+ lfd %f2, RETVAL+8(%r1)
+ lfd %f3, RETVAL+16(%r1)
+ lfd %f4, RETVAL+24(%r1)
+ mtlr %r0
+ lfd %f5, RETVAL+32(%r1)
+ lfd %f6, RETVAL+40(%r1)
+ lfd %f7, RETVAL+48(%r1)
+ lfd %f8, RETVAL+56(%r1)
+ addi %r1, %r1, STACKFRAME
+ blr
+.Lmorefloat:
+ lfs %f4, RETVAL+12(%r1)
+ mtlr %r0
+ lfs %f5, RETVAL+16(%r1)
+ lfs %f6, RETVAL+20(%r1)
+ lfs %f7, RETVAL+24(%r1)
+ lfs %f8, RETVAL+28(%r1)
+ addi %r1, %r1, STACKFRAME
+ blr
+.Lsmall:
+# ifdef __LITTLE_ENDIAN__
+ ld %r3,RETVAL+0(%r1)
+ mtlr %r0
+ ld %r4,RETVAL+8(%r1)
+ addi %r1, %r1, STACKFRAME
+ blr
+# else
+ # A struct smaller than a dword is returned in the low bits of r3
+ # ie. right justified. Larger structs are passed left justified
+ # in r3 and r4. The return value area on the stack will have
+ # the structs as they are usually stored in memory.
+ cmpldi %r3, FFI_V2_TYPE_SMALL_STRUCT + 7 # size 8 bytes?
+ neg %r5, %r3
+ ld %r3,RETVAL+0(%r1)
+ blt .Lsmalldown
+ mtlr %r0
+ ld %r4,RETVAL+8(%r1)
+ addi %r1, %r1, STACKFRAME
+ blr
+.Lsmalldown:
+ addi %r5, %r5, FFI_V2_TYPE_SMALL_STRUCT + 7
+ mtlr %r0
+ sldi %r5, %r5, 3
+ addi %r1, %r1, STACKFRAME
+ srd %r3, %r3, %r5
blr
-# esac
+# endif
+
.LFE1:
.long 0
.byte 0,12,0,1,128,0,0,0
+# if _CALL_ELF == 2
+ .size ffi_closure_LINUX64,.-ffi_closure_LINUX64
+# else
+# ifdef _CALL_LINUX
+ .size ffi_closure_LINUX64,.-.L.ffi_closure_LINUX64
+# else
.size .ffi_closure_LINUX64,.-.ffi_closure_LINUX64
+# endif
+# endif
.section .eh_frame,EH_FRAME_FLAGS,@progbits
.Lframe1:
@@ -223,14 +376,14 @@ ffi_closure_LINUX64:
.byte 0x2 # DW_CFA_advance_loc1
.byte .LCFI0-.LFB1
.byte 0xe # DW_CFA_def_cfa_offset
- .uleb128 240
+ .uleb128 STACKFRAME
.byte 0x11 # DW_CFA_offset_extended_sf
.uleb128 0x41
.sleb128 -2
.align 3
.LEFDE1:
-#endif
-#if defined __ELF__ && defined __linux__
+# if defined __ELF__ && defined __linux__
.section .note.GNU-stack,"",@progbits
+# endif
#endif
Index: mozilla/js/src/ctypes/libffi/src/powerpc/linux64.S
===================================================================
--- mozilla.orig/js/src/ctypes/libffi/src/powerpc/linux64.S
+++ mozilla/js/src/ctypes/libffi/src/powerpc/linux64.S
@@ -30,16 +30,33 @@
#include <ffi.h>
#ifdef __powerpc64__
- .hidden ffi_call_LINUX64, .ffi_call_LINUX64
- .globl ffi_call_LINUX64, .ffi_call_LINUX64
+ .hidden ffi_call_LINUX64
+ .globl ffi_call_LINUX64
+# if _CALL_ELF == 2
+ .text
+ffi_call_LINUX64:
+ addis %r2, %r12, .TOC.-ffi_call_LINUX64@ha
+ addi %r2, %r2, .TOC.-ffi_call_LINUX64@l
+ .localentry ffi_call_LINUX64, . - ffi_call_LINUX64
+# else
.section ".opd","aw"
.align 3
ffi_call_LINUX64:
+# ifdef _CALL_LINUX
+ .quad .L.ffi_call_LINUX64,.TOC.@tocbase,0
+ .type ffi_call_LINUX64,@function
+ .text
+.L.ffi_call_LINUX64:
+# else
+ .hidden .ffi_call_LINUX64
+ .globl .ffi_call_LINUX64
.quad .ffi_call_LINUX64,.TOC.@tocbase,0
.size ffi_call_LINUX64,24
.type .ffi_call_LINUX64,@function
.text
.ffi_call_LINUX64:
+# endif
+# endif
.LFB1:
mflr %r0
std %r28, -32(%r1)
@@ -54,22 +71,35 @@ ffi_call_LINUX64:
mr %r31, %r5 /* flags, */
mr %r30, %r6 /* rvalue, */
mr %r29, %r7 /* function address. */
+/* Save toc pointer, not for the ffi_prep_args64 call, but for the later
+ bctrl function call. */
+# if _CALL_ELF == 2
+ std %r2, 24(%r1)
+# else
std %r2, 40(%r1)
+# endif
/* Call ffi_prep_args64. */
mr %r4, %r1
+# if defined _CALL_LINUX || _CALL_ELF == 2
+ bl ffi_prep_args64
+# else
bl .ffi_prep_args64
+# endif
- ld %r0, 0(%r29)
+# if _CALL_ELF == 2
+ mr %r12, %r29
+# else
+ ld %r12, 0(%r29)
ld %r2, 8(%r29)
ld %r11, 16(%r29)
-
+# endif
/* Now do the call. */
/* Set up cr1 with bits 4-7 of the flags. */
mtcrf 0x40, %r31
/* Get the address to call into CTR. */
- mtctr %r0
+ mtctr %r12
/* Load all those argument registers. */
ld %r3, -32-(8*8)(%r28)
ld %r4, -32-(7*8)(%r28)
@@ -104,12 +134,17 @@ ffi_call_LINUX64:
/* This must follow the call immediately, the unwinder
uses this to find out if r2 has been saved or not. */
+# if _CALL_ELF == 2
+ ld %r2, 24(%r1)
+# else
ld %r2, 40(%r1)
+# endif
/* Now, deal with the return value. */
mtcrf 0x01, %r31
- bt- 30, .Ldone_return_value
- bt- 29, .Lfp_return_value
+ bt 31, .Lstruct_return_value
+ bt 30, .Ldone_return_value
+ bt 29, .Lfp_return_value
std %r3, 0(%r30)
/* Fall through... */
@@ -117,7 +152,7 @@ ffi_call_LINUX64:
/* Restore the registers we used and return. */
mr %r1, %r28
ld %r0, 16(%r28)
- ld %r28, -32(%r1)
+ ld %r28, -32(%r28)
mtlr %r0
ld %r29, -24(%r1)
ld %r30, -16(%r1)
@@ -134,10 +169,48 @@ ffi_call_LINUX64:
.Lfloat_return_value:
stfs %f1, 0(%r30)
b .Ldone_return_value
+
+.Lstruct_return_value:
+ bf 29, .Lsmall_struct
+ bf 28, .Lfloat_homog_return_value
+ stfd %f1, 0(%r30)
+ stfd %f2, 8(%r30)
+ stfd %f3, 16(%r30)
+ stfd %f4, 24(%r30)
+ stfd %f5, 32(%r30)
+ stfd %f6, 40(%r30)
+ stfd %f7, 48(%r30)
+ stfd %f8, 56(%r30)
+ b .Ldone_return_value
+
+.Lfloat_homog_return_value:
+ stfs %f1, 0(%r30)
+ stfs %f2, 4(%r30)
+ stfs %f3, 8(%r30)
+ stfs %f4, 12(%r30)
+ stfs %f5, 16(%r30)
+ stfs %f6, 20(%r30)
+ stfs %f7, 24(%r30)
+ stfs %f8, 28(%r30)
+ b .Ldone_return_value
+
+.Lsmall_struct:
+ std %r3, 0(%r30)
+ std %r4, 8(%r30)
+ b .Ldone_return_value
+
.LFE1:
.long 0
.byte 0,12,0,1,128,4,0,0
+# if _CALL_ELF == 2
+ .size ffi_call_LINUX64,.-ffi_call_LINUX64
+# else
+# ifdef _CALL_LINUX
+ .size ffi_call_LINUX64,.-.L.ffi_call_LINUX64
+# else
.size .ffi_call_LINUX64,.-.ffi_call_LINUX64
+# endif
+# endif
.section .eh_frame,EH_FRAME_FLAGS,@progbits
.Lframe1:
@@ -180,8 +253,8 @@ ffi_call_LINUX64:
.uleb128 0x4
.align 3
.LEFDE1:
-#endif
-#if defined __ELF__ && defined __linux__
+# if (defined __ELF__ && defined __linux__) || _CALL_ELF == 2
.section .note.GNU-stack,"",@progbits
+# endif
#endif
Index: mozilla/js/src/ctypes/libffi/src/powerpc/ppc_closure.S
===================================================================
--- mozilla.orig/js/src/ctypes/libffi/src/powerpc/ppc_closure.S
+++ mozilla/js/src/ctypes/libffi/src/powerpc/ppc_closure.S
@@ -122,43 +122,78 @@ ENTRY(ffi_closure_SYSV)
blr
# case FFI_TYPE_FLOAT
+#ifndef __NO_FPRS__
lfs %f1,112+0(%r1)
mtlr %r0
addi %r1,%r1,144
+#else
+ nop
+ nop
+ nop
+#endif
blr
# case FFI_TYPE_DOUBLE
+#ifndef __NO_FPRS__
lfd %f1,112+0(%r1)
mtlr %r0
addi %r1,%r1,144
+#else
+ nop
+ nop
+ nop
+#endif
blr
# case FFI_TYPE_LONGDOUBLE
+#ifndef __NO_FPRS__
lfd %f1,112+0(%r1)
lfd %f2,112+8(%r1)
mtlr %r0
b .Lfinish
+#else
+ nop
+ nop
+ nop
+ blr
+#endif
# case FFI_TYPE_UINT8
+#ifdef __LITTLE_ENDIAN__
+ lbz %r3,112+0(%r1)
+#else
lbz %r3,112+3(%r1)
+#endif
mtlr %r0
addi %r1,%r1,144
blr
# case FFI_TYPE_SINT8
+#ifdef __LITTLE_ENDIAN__
+ lbz %r3,112+0(%r1)
+#else
lbz %r3,112+3(%r1)
+#endif
extsb %r3,%r3
mtlr %r0
b .Lfinish
# case FFI_TYPE_UINT16
+#ifdef __LITTLE_ENDIAN__
+ lhz %r3,112+0(%r1)
+#else
lhz %r3,112+2(%r1)
+#endif
mtlr %r0
addi %r1,%r1,144
blr
# case FFI_TYPE_SINT16
+#ifdef __LITTLE_ENDIAN__
+ lha %r3,112+0(%r1)
+#else
lha %r3,112+2(%r1)
+#endif
mtlr %r0
addi %r1,%r1,144
blr
@@ -203,7 +238,7 @@ ENTRY(ffi_closure_SYSV)
lwz %r3,112+0(%r1)
lwz %r4,112+4(%r1)
lwz %r5,112+8(%r1)
- bl .Luint128
+ b .Luint128
# The return types below are only used when the ABI type is FFI_SYSV.
# case FFI_SYSV_TYPE_SMALL_STRUCT + 1. One byte struct.
@@ -220,9 +255,15 @@ ENTRY(ffi_closure_SYSV)
# case FFI_SYSV_TYPE_SMALL_STRUCT + 3. Three byte struct.
lwz %r3,112+0(%r1)
+#ifdef __LITTLE_ENDIAN__
+ mtlr %r0
+ addi %r1,%r1,144
+ blr
+#else
srwi %r3,%r3,8
mtlr %r0
b .Lfinish
+#endif
# case FFI_SYSV_TYPE_SMALL_STRUCT + 4. Four byte struct.
lwz %r3,112+0(%r1)
@@ -233,20 +274,35 @@ ENTRY(ffi_closure_SYSV)
# case FFI_SYSV_TYPE_SMALL_STRUCT + 5. Five byte struct.
lwz %r3,112+0(%r1)
lwz %r4,112+4(%r1)
+#ifdef __LITTLE_ENDIAN__
+ mtlr %r0
+ b .Lfinish
+#else
li %r5,24
b .Lstruct567
+#endif
# case FFI_SYSV_TYPE_SMALL_STRUCT + 6. Six byte struct.
lwz %r3,112+0(%r1)
lwz %r4,112+4(%r1)
+#ifdef __LITTLE_ENDIAN__
+ mtlr %r0
+ b .Lfinish
+#else
li %r5,16
b .Lstruct567
+#endif
# case FFI_SYSV_TYPE_SMALL_STRUCT + 7. Seven byte struct.
lwz %r3,112+0(%r1)
lwz %r4,112+4(%r1)
+#ifdef __LITTLE_ENDIAN__
+ mtlr %r0
+ b .Lfinish
+#else
li %r5,8
b .Lstruct567
+#endif
# case FFI_SYSV_TYPE_SMALL_STRUCT + 8. Eight byte struct.
lwz %r3,112+0(%r1)
@@ -254,6 +310,7 @@ ENTRY(ffi_closure_SYSV)
mtlr %r0
b .Lfinish
+#ifndef __LITTLE_ENDIAN__
.Lstruct567:
subfic %r6,%r5,32
srw %r4,%r4,%r5
@@ -263,6 +320,7 @@ ENTRY(ffi_closure_SYSV)
mtlr %r0
addi %r1,%r1,144
blr
+#endif
.Luint128:
lwz %r6,112+12(%r1)
Index: mozilla/js/src/ctypes/libffi/src/powerpc/sysv.S
===================================================================
--- mozilla.orig/js/src/ctypes/libffi/src/powerpc/sysv.S
+++ mozilla/js/src/ctypes/libffi/src/powerpc/sysv.S
@@ -83,6 +83,7 @@ ENTRY(ffi_call_SYSV)
nop
1:
+#ifndef __NO_FPRS__
/* Load all the FP registers. */
bf- 6,2f
lfd %f1,-16-(8*4)-(8*8)(%r28)
@@ -94,6 +95,7 @@ ENTRY(ffi_call_SYSV)
lfd %f6,-16-(8*4)-(3*8)(%r28)
lfd %f7,-16-(8*4)-(2*8)(%r28)
lfd %f8,-16-(8*4)-(1*8)(%r28)
+#endif
2:
/* Make the call. */
@@ -103,7 +105,9 @@ ENTRY(ffi_call_SYSV)
mtcrf 0x01,%r31 /* cr7 */
bt- 31,L(small_struct_return_value)
bt- 30,L(done_return_value)
+#ifndef __NO_FPRS__
bt- 29,L(fp_return_value)
+#endif
stw %r3,0(%r30)
bf+ 28,L(done_return_value)
stw %r4,4(%r30)
@@ -124,6 +128,7 @@ L(done_return_value):
lwz %r1,0(%r1)
blr
+#ifndef __NO_FPRS__
L(fp_return_value):
bf 28,L(float_return_value)
stfd %f1,0(%r30)
@@ -134,21 +139,17 @@ L(fp_return_value):
L(float_return_value):
stfs %f1,0(%r30)
b L(done_return_value)
+#endif
L(small_struct_return_value):
- extrwi %r6,%r31,2,19 /* number of bytes padding = shift/8 */
- mtcrf 0x02,%r31 /* copy flags to cr[24:27] (cr6) */
- extrwi %r5,%r31,5,19 /* r5 <- number of bits of padding */
- subfic %r6,%r6,4 /* r6 <- number of useful bytes in r3 */
- bf- 25,L(done_return_value) /* struct in r3 ? if not, done. */
-/* smst_one_register: */
- slw %r3,%r3,%r5 /* Left-justify value in r3 */
- mtxer %r6 /* move byte count to XER ... */
- stswx %r3,0,%r30 /* ... and store that many bytes */
- bf+ 26,L(done_return_value) /* struct in r3:r4 ? */
- add %r6,%r6,%r30 /* adjust pointer */
- stswi %r4,%r6,4 /* store last four bytes */
- b L(done_return_value)
+ /*
+ * The C code always allocates a properly-aligned 8-byte bounce
+ * buffer to make this assembly code very simple. Just write out
+ * r3 and r4 to the buffer to allow the C code to handle the rest.
+ */
+ stw %r3, 0(%r30)
+ stw %r4, 4(%r30)
+ b L(done_return_value)
.LFE1:
END(ffi_call_SYSV)