aix.S (ffi_call_AIX): Convert to more standard register usage.

        * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
        register usage.  Call ffi_prep_args directly.  Add long double
        return value support.
        * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
        applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
        Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
        (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
        mode.
        (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
        into case.
        * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
        Allocate result area between params and FPRs.

From-SVN: r154892

1 files changed, 32 insertions, 45 deletions

diff --git a/libffi/src/powerpc/aix.S b/libffi/src/powerpc/aix.S
index 7b73999..5b477ca 100644
--- a/libffi/src/powerpc/aix.S
+++ b/libffi/src/powerpc/aix.S
@@ -104,47 +104,34 @@ ffi_call_AIX:
 	.llong .ffi_call_AIX, TOC[tc0], 0
 	.csect .text[PR]
 .ffi_call_AIX:
-	mr	r12,r8 // We only need r12 until the call, so it doesn't have to be saved...
-	/* Save the old stack pointer as AP.  */
-	mr	r8,r1
-
-	/* Allocate the stack space we need.  */
-	stdux	r1,r1,r4
-
 	/* Save registers we use.  */
-	mflr	r9
+	mflr	r0
 
-	std	r28,-32(r8)
-	std	r29,-24(r8)
-	std	r30,-16(r8)
-	std	r31, -8(r8)
+	std	r28,-32(r1)
+	std	r29,-24(r1)
+	std	r30,-16(r1)
+	std	r31, -8(r1)
 
-	std	r9, 16(r8)
-	std	r2, 40(r1)
+	std	r0, 16(r1)
+	mr	r28, r1		/* our AP.  */
+	stdux	r1, r1, r4
 
 	/* Save arguments over call...  */
-	mr	r31,r5	/* flags, */
-	mr	r30,r6	/* rvalue, */
-	mr	r29,r7	/* function address, */
-	mr	r28,r8	/* our AP. */
+	mr	r31, r5	/* flags, */
+	mr	r30, r6	/* rvalue, */
+	mr	r29, r7	/* function address.  */
+	std	r2, 40(r1)
 
 	/* Call ffi_prep_args.  */
-	mr	r4,r1
-	li	r9,0
-
-	ld	r2,8(r12)
-	ld	r12,0(r12)
-	mtctr	r12 // r12 holds address of _ffi_prep_args
-	bctrl
-	ld	r2,40(r1)
+	mr	r4, r1
+	bl .ffi_prep_args
 
 	/* Now do the call.  */
-	ld	r12,0(r29)
+	ld	r0, 0(r29)
+	ld	r2, 8(r29)
 	/* Set up cr1 with bits 4-7 of the flags.  */
-	mtcrf	0x40,r31
-	std	r2,40(r1)
-	mtctr	r12
-	ld	r2,8(r29)
+	mtcrf	0x40, r31
+	mtctr	r0
 	/* Load all those argument registers.  */
 	// We have set up a nice stack frame, just load it into registers.
 	ld	r3, 40+(1*8)(r1)
@@ -180,33 +167,33 @@ L1:
 L2:
 	/* Make the call.  */
 	bctrl
-	ld	r2,40(r1)
+	ld	r2, 40(r1)
 
 	/* Now, deal with the return value.  */
-	mtcrf	0x01,r31
+	mtcrf	0x01, r31
 
-	bt	30,L(done_return_value)
-	bt	29,L(fp_return_value)
-	std	r3,0(r30)
-	bf	28,L(done_return_value)
-	std	r4,4(r30)
+	bt	30, L(done_return_value)
+	bt	29, L(fp_return_value)
+	std	r3, 0(r30)
 
 	/* Fall through...  */
 
 L(done_return_value):
 	/* Restore the registers we used and return.  */
-	ld	r9,16(r28)
-	ld	r31,-8(r28)
-	mtlr	r9
-	ld	r30,-16(r28)
-	ld	r29,-24(r28)
-	ld	r28,-32(r28)
-	ld	r1,0(r1)
+	mr	r1, r28
+	ld	r0, 16(r28)
+	ld	r28,-32(r1)
+	mtlr	r0
+	ld	r29,-24(r1)
+	ld	r30,-16(r1)
+	ld	r31,-8(r1)
 	blr
 
 L(fp_return_value):
 	bf	28,L(float_return_value)
 	stfd	f1,0(r30)
+	bf	31,L(done_return_value)
+	stfd	f2,8(r30)
 	b	L(done_return_value)
 L(float_return_value):
 	stfs	f1,0(r30)