rbdc 0.8.7 → 0.9.0

data/dyncall/dyncall/dyncall_call_arm32_arm.S

@@ -6,7 +6,7 @@
  Description: Call Kernel for ARM 32-bit ARM Architecture
  License:
 
-   Copyright (c) 2007-2011 Daniel Adler <dadler@uni-goettingen.de>, 
+   Copyright (c) 2007-2015 Daniel Adler <dadler@uni-goettingen.de>, 
                            Tassilo Philipp <tphilipp@potion-studios.com>
 
    Permission to use, copy, modify, and distribute this software for any
@@ -24,6 +24,7 @@
 */
 
 
+
 #include "../portasm/portasm-arm.S"
 
 /* ============================================================================

data/dyncall/dyncall/dyncall_call_arm32_arm.h

@@ -6,7 +6,7 @@
  Description: 
  License:
 
-   Copyright (c) 2007-2011 Daniel Adler <dadler@uni-goettingen.de>, 
+   Copyright (c) 2007-2015 Daniel Adler <dadler@uni-goettingen.de>, 
                            Tassilo Philipp <tphilipp@potion-studios.com>
 
    Permission to use, copy, modify, and distribute this software for any
@@ -24,6 +24,7 @@
 */
 
 
+
 /*
 
   dyncall 32bit ARM32 family interface (ARM mode)
@@ -49,9 +50,14 @@ extern "C" {
 **
 ** - hybrid return-type call (bool ... pointer)
 **
+** Note the return type of this declaration is intentially of double-word size.
+** On some platforms (FreeBSD/arm, Nintendo DS, ...) the compiler generates cleanup code
+** in the caller (dc_callvm_call_arm32_arm) that reuses, thus overwrites r0 and r1.
+** With this "hint", we preserve those registers by letting the compiler assume both
+** registers are used for the return type.
 */
 
-void dcCall_arm32_arm(DCpointer target, DCpointer stackdata, DCsize size);
+DClonglong dcCall_arm32_arm(DCpointer target, DCpointer stackdata, DCsize size);
 
 #ifdef __cplusplus
 }

data/dyncall/dyncall/dyncall_call_arm32_arm_armhf.S

@@ -6,7 +6,7 @@
  Description: Call Kernel for ARM 32-bit ARM Architecture - Hard Float
  License:
 
-   Copyright (c) 2007-2012 Daniel Adler <dadler@uni-goettingen.de>, 
+   Copyright (c) 2007-2015 Daniel Adler <dadler@uni-goettingen.de>, 
                            Tassilo Philipp <tphilipp@potion-studios.com>
 
    Permission to use, copy, modify, and distribute this software for any
@@ -24,6 +24,7 @@
 */
 
 
+
 #include "../portasm/portasm-arm.S"
 
 /* ============================================================================
@@ -40,6 +41,12 @@
 .arch armv6
 .fpu  vfp
 
+/*
+  1st arg / r0 = funptr
+  2st arg / r1 = ptr to int args
+  3st arg / r2 = size
+  4st arg / r3 = ptr to float args
+*/
 GLOBAL_C(dcCall_arm32_armhf)
 ENTRY_C(dcCall_arm32_armhf)
 
@@ -52,24 +59,9 @@ ENTRY_C(dcCall_arm32_armhf)
 	mov	r5  , r1	 /* r5 = 'args' (2nd argument is passed in r1). */
 	
 	/* Load 16 single-precision registers (= 8 double-precision registers). */
-	flds	s0,  [r3,#0 ]
-	flds	s1,  [r3,#4 ]
-	flds	s2,  [r3,#8 ]
-	flds	s3,  [r3,#12]
-	flds	s4,  [r3,#16]
-	flds	s5,  [r3,#20]
-	flds	s6,  [r3,#24]
-	flds	s7,  [r3,#28]
-	flds	s8,  [r3,#32]
-	flds	s9,  [r3,#36]
-	flds	s10, [r3,#40]
-	flds	s11, [r3,#44]
-	flds	s12, [r3,#48]
-	flds	s13, [r3,#52]
-	flds	s14, [r3,#56]
-	flds	s15, [r3,#60]
-
-	sub	r2 , r2 , #16	
+	fldmiad	r3, {d0-d7}
+
+	sub	r2 , r2 , #16	/* skip spill area */
 	cmp     r2, #0
 	ble	armhf_call
 

data/dyncall/dyncall/dyncall_call_arm32_arm_armhf.h

@@ -6,7 +6,7 @@
  Description: 
  License:
 
-   Copyright (c) 2007-2012 Daniel Adler <dadler@uni-goettingen.de>, 
+   Copyright (c) 2007-2015 Daniel Adler <dadler@uni-goettingen.de>, 
                            Tassilo Philipp <tphilipp@potion-studios.com>
 
    Permission to use, copy, modify, and distribute this software for any
@@ -24,6 +24,7 @@
 */
 
 
+
 #ifndef DYNCALL_CALL_ARM32_ARM_ARMHF_H
 #define DYNCALL_CALL_ARM32_ARM_ARMHF_H
 
@@ -34,7 +35,19 @@
 extern "C" {
 #endif
 
-void dcCall_arm32_armhf(DCpointer target, DCpointer stackdata, DCsize size, DCfloat* p_s16);
+/* 
+** arm32 armhf mode calling convention calls 
+**
+** - hybrid return-type call (bool ... pointer)
+**
+** Note the return type of this declaration is intentially of double-word size.
+** On some platforms the compiler generates cleanup code in the caller (dyncall_callvm_arm32_arm_armhf.c's
+** call()) that reuses,thus overwrites r0 and r1.
+** With this "hint", we preserve those registers by letting the compiler assume both
+** registers are used for the return type.
+*/
+
+DClonglong dcCall_arm32_armhf(DCpointer target, DCpointer stackdata, DCsize size, DCfloat* p_s16);
 
 #ifdef __cplusplus
 }

data/dyncall/dyncall/dyncall_call_arm32_thumb.h

@@ -6,7 +6,7 @@
  Description: 
  License:
 
-   Copyright (c) 2007-2011 Daniel Adler <dadler@uni-goettingen.de>, 
+   Copyright (c) 2007-2015 Daniel Adler <dadler@uni-goettingen.de>, 
                            Tassilo Philipp <tphilipp@potion-studios.com>
 
    Permission to use, copy, modify, and distribute this software for any
@@ -24,6 +24,7 @@
 */
 
 
+
 /*
 
   dyncall 32bit ARM32 family interface (THUMB mode)
@@ -49,13 +50,14 @@ extern "C" {
 **
 ** - hybrid return-type call (bool ... pointer)
 **
+** Note the return type of this declaration is intentially of double-word size.
+** On some platforms (FreeBSD/arm, Nintendo DS, ...) the compiler generates cleanup code
+** in the caller (dc_callvm_call_arm32_thumb) that reuses, thus overwrites r0 and r1.
+** With this "hint", we preserve those registers by letting the compiler assume both
+** registers are used for the return type.
 */
 
-void dcCall_arm32_thumb(DCpointer target, DCpointer stackdata, DCsize size);
-
-/* Internally used to avoid compiler overwriting r0 and r1 in call stub */
-DClong     dcCall_arm32_thumb_word (DCpointer target, DCpointer stackdata, DCsize size);
-DClonglong dcCall_arm32_thumb_dword(DCpointer target, DCpointer stackdata, DCsize size);
+DClonglong dcCall_arm32_thumb(DCpointer target, DCpointer stackdata, DCsize size);
 
 #ifdef __cplusplus
 }

data/dyncall/dyncall/dyncall_call_arm32_thumb_armhf.S

@@ -3,10 +3,10 @@
  Package: dyncall
  Library: dyncall
  File: dyncall/dyncall_call_arm32_thumb_armhf.S
- Description: Call Kernel for ARM 32-bit ARM Architecture - Hard Float in Thumb code 
+ Description: Call Kernel for ARM 32-bit ARM Architecture - Hard Float in Thumb code
  License:
 
-   Copyright (c) 2007-2012 Daniel Adler <dadler@uni-goettingen.de>, 
+   Copyright (c) 2007-2015 Daniel Adler <dadler@uni-goettingen.de>, 
                            Tassilo Philipp <tphilipp@potion-studios.com>
 
    Permission to use, copy, modify, and distribute this software for any
@@ -24,6 +24,7 @@
 */
 
 
+
 #include "../portasm/portasm-arm.S"
 
 /* ============================================================================
@@ -43,6 +44,12 @@
 // .arch armv6
 // .fpu  vfp
 
+/*
+  1st arg / r0 = funptr
+  2st arg / r1 = ptr to int args
+  3st arg / r2 = size
+  4st arg / r3 = ptr to float args
+*/
 GLOBAL_C(dcCall_arm32_armhf)
 .thumb_func
 ENTRY_C(dcCall_arm32_armhf)
@@ -61,22 +68,7 @@ ENTRY_C(dcCall_arm32_armhf)
 	mov	r5  , r1	 /* r5 = 'args' (2nd argument is passed in r1). */
 	
 	/* Load 16 single-precision registers (= 8 double-precision registers). */
-	flds	s0,  [r3,#0 ]
-	flds	s1,  [r3,#4 ]
-	flds	s2,  [r3,#8 ]
-	flds	s3,  [r3,#12]
-	flds	s4,  [r3,#16]
-	flds	s5,  [r3,#20]
-	flds	s6,  [r3,#24]
-	flds	s7,  [r3,#28]
-	flds	s8,  [r3,#32]
-	flds	s9,  [r3,#36]
-	flds	s10, [r3,#40]
-	flds	s11, [r3,#44]
-	flds	s12, [r3,#48]
-	flds	s13, [r3,#52]
-	flds	s14, [r3,#56]
-	flds	s15, [r3,#60]
+	fldmiad	r3, {d0-d7}
 
 	sub	r2 , #16	
 	cmp     r2, #0

data/dyncall/dyncall/dyncall_call_arm32_thumb_gas.s

@@ -83,19 +83,3 @@ call:
 	mov		%r13, %r7	/* Reset stack ptr. */
 	pop		{%r4-%r7, %r15}	/* Restore permanent registers and program counter. (Force a stay in THUMB in ARMv4, whether ARMv5 can return in ARM or THUMB depending on the bit 0. */
 
-
-
-/* Internally used to avoid compiler overwriting r0 and r1 in call stub */
-.globl dcCall_arm32_thumb_word
-
-.thumb_func
-dcCall_arm32_thumb_word:
-	b	dcCall_arm32_thumb
-
-
-.globl dcCall_arm32_thumb_dword
-
-.thumb_func
-dcCall_arm32_thumb_dword:
-	b	dcCall_arm32_thumb
-

data/dyncall/dyncall/dyncall_call_arm64.S

@@ -0,0 +1,226 @@
+/*
+
+ Package: dyncall
+ Library: dyncall
+ File: dyncall/dyncall_call_arm64.S
+ Description: Call Kernel for ARM 64-bit Architecture (aka ARM64, AArch64)
+ License:
+
+   Copyright (c) 2015 Daniel Adler <dadler@uni-goettingen.de>, 
+                      Tassilo Philipp <tphilipp@potion-studios.com>
+
+   Permission to use, copy, modify, and distribute this software for any
+   purpose with or without fee is hereby granted, provided that the above
+   copyright notice and this permission notice appear in all copies.
+
+   THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+   WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+   MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+   ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+   WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+   ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+   OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+
+*/
+
+
+
+#include "../portasm/portasm-arm.S"
+
+/* ============================================================================
+   DynCall Call Kernel for ARM 64-bit ARM Architecture 
+   ----------------------------------------------------------------------------
+   C Interface:
+     dcCall_arm64 (DCpointer target, DCpointer data, DCsize size, DCfloat* regdata);
+
+   This Call Kernel was tested on Debian/qemu-debootstrap arm64 jessie.
+*/
+
+.text
+//
+// DynCall Back-End arm64 
+// 
+// Supported ABIs:
+// - 'ARM 64-bit AArch64 PCS' (@dadler: work in progress)
+// 
+// Useful Links:
+// - http://lxr.free-electrons.com/source/arch/arm64/kernel/stacktrace.c
+
+GLOBAL_C(dcCall_arm64)
+ENTRY_C(dcCall_arm64)
+.align 2
+
+// input:
+//   x0: target   (address of target)
+//   x1: data     (address of stack copy data)
+//   x2: size     (number of 'pair' 16-byte units)
+//   x3: regdata  (address of register data)
+
+// prolog:
+	
+	stp  x29, x30, [sp, #-16]!	// allocate frame
+	mov  x29,  sp
+
+// load 64-bit floating-point registers
+        
+	ldr  d0,  [x3,#0 ]
+	ldr  d1,  [x3,#8 ]
+	ldr  d2,  [x3,#16]
+	ldr  d3,  [x3,#24]
+	ldr  d4,  [x3,#32]
+	ldr  d5,  [x3,#40]
+	ldr  d6,  [x3,#48]
+	ldr  d7,  [x3,#56]
+
+// copy to stack
+	
+	sub  sp, sp, x2		// create call-frame
+
+	eor  x4, x4, x4		// x4: cnt = 0
+
+	mov  x5, x1		// x5: read pointer = data
+	mov  x6, sp		// x6: write pointer = sp
+
+.next:
+	cmp  x4, x2
+	b.ge  .done
+	
+	ldp  x7, x9, [x5], #16	// get pair from data
+	stp  x7, x9, [x6], #16	// put to stack
+	add  x4, x4, 16	        // advance 16 bytes
+
+	b   .next
+
+.done:
+	
+// rescue temp int registers
+
+	mov  x9 , x0			// x9: target
+	add  x10, x3, 64                // x3: integer reg buffer
+
+// load 64-bit integer registers ( 8 x 64-bit )
+	
+	// load register set
+
+	ldr  x0, [x10, #0]
+	ldr  x1, [x10, #8]
+	ldr  x2, [x10, #16]
+	ldr  x3, [x10, #24]
+	ldr  x4, [x10, #32]
+	ldr  x5, [x10, #40]
+	ldr  x6, [x10, #48]
+	ldr  x7, [x10, #56]
+	
+// call target:
+	
+	blr  x9
+
+// epilog:
+
+	mov  sp,  x29
+	ldp  x29, x30, [sp], 16
+
+	ret
+
+#if 0
+
+	
+
+// epilog:
+
+	add  sp, x28, 0		// remove call record
+	
+	ret
+
+	// -- OLD:
+
+	str  x27,      [sp, 16]			// use 1 local var (size)
+	
+
+        ldr  q0,  [x3,#0 ]
+	ldr  q1,  [x3,#8 ]
+	ldr  q2,  [x3,#16]
+	ldr  q3,  [x3,#24]
+	ldr  q4,  [x3,#32]
+	ldr  q5,  [x3,#40]
+	ldr  q6,  [x3,#48]
+	ldr  q7,  [x3,#56]
+
+	ldr  d8,  [x3,#32]
+	ldr  d9,  [x3,#36]
+	ldr  d10, [x3,#40]
+	ldr  d11, [x3,#44]
+	ldr  d12, [x3,#48]
+	ldr  d13, [x3,#52]
+	ldr  d14, [x3,#56]
+	ldr  d15, [x3,#60]
+
+
+	// load float ( 16 x 32-bit ) 
+
+        ldr  s0,  [x3,#0 ]
+	ldr  s1,  [x3,#4 ]
+	ldr  s2,  [x3,#8 ]
+	ldr  s3,  [x3,#12]
+	ldr  s4,  [x3,#16]
+	ldr  s5,  [x3,#20]
+	ldr  s6,  [x3,#24]
+	ldr  s7,  [x3,#28]
+	ldr  s8,  [x3,#32]
+	ldr  s9,  [x3,#36]
+	ldr  s10, [x3,#40]
+	ldr  s11, [x3,#44]
+	ldr  s12, [x3,#48]
+	ldr  s13, [x3,#52]
+	ldr  s14, [x3,#56]
+	ldr  s15, [x3,#60]
+ 
+	// call
+	
+	blr  x0
+
+	// epilog
+
+	ldp  x29, x30, [sp], 32
+	ret
+
+	// stack copy 
+
+	sub  sp, sp, x2	// decrement stack by size (x2) 
+	eor  x3, x3, x3	// x3 = counter, set to zero
+
+// .next:
+	ldr  x4, [x1, x3]  // x4 = 64-bit stack data
+	str  x4, [sp, x3]  // store to stack
+	add  x3, x3, #8
+	cmp  x3, x2
+	blt .next
+
+
+	// rescue int registers
+
+	mov  x9 , x0			// x9 = code ptr
+	mov  x10, x2
+
+	// load int ( 8 x 64-bit )
+
+	ldr  x0, [sp, #0]
+	ldr  x1, [sp, #8]
+	ldr  x2, [sp, #16]
+	ldr  x3, [sp, #24]
+	ldr  x4, [sp, #32]
+	ldr  x5, [sp, #40]
+	ldr  x6, [sp, #48]
+	ldr  x7, [sp, #56]
+
+	// call
+	
+	blr  x9
+
+	// epilog
+
+	ldp  x29, x30, [sp], 32
+	ret
+#endif
+ 
+