ffi 1.11.3 → 1.13.1

data/ext/ffi_c/libffi/src/powerpc/linux64.S

@@ -101,40 +101,70 @@ ffi_call_LINUX64:
 	ld	%r2, 8(%r29)
 # endif
 	/* Now do the call.  */
-	/* Set up cr1 with bits 4-7 of the flags.  */
-	mtcrf	0x40, %r31
+	/* Set up cr1 with bits 3-7 of the flags.  */
+	mtcrf	0xc0, %r31
 
 	/* Get the address to call into CTR.  */
 	mtctr	%r12
 	/* Load all those argument registers.  */
-	ld	%r3, -32-(8*8)(%r28)
-	ld	%r4, -32-(7*8)(%r28)
-	ld	%r5, -32-(6*8)(%r28)
-	ld	%r6, -32-(5*8)(%r28)
+	addi	%r29, %r28, -32-(8*8)
+	ld	%r3,  (0*8)(%r29)
+	ld	%r4,  (1*8)(%r29)
+	ld	%r5,  (2*8)(%r29)
+	ld	%r6,  (3*8)(%r29)
 	bf-	5, 1f
-	ld	%r7, -32-(4*8)(%r28)
-	ld	%r8, -32-(3*8)(%r28)
-	ld	%r9, -32-(2*8)(%r28)
-	ld	%r10, -32-(1*8)(%r28)
+	ld	%r7,  (4*8)(%r29)
+	ld	%r8,  (5*8)(%r29)
+	ld	%r9,  (6*8)(%r29)
+	ld	%r10, (7*8)(%r29)
 1:
 
 	/* Load all the FP registers.  */
 	bf-	6, 2f
-	lfd	%f1, -32-(21*8)(%r28)
-	lfd	%f2, -32-(20*8)(%r28)
-	lfd	%f3, -32-(19*8)(%r28)
-	lfd	%f4, -32-(18*8)(%r28)
-	lfd	%f5, -32-(17*8)(%r28)
-	lfd	%f6, -32-(16*8)(%r28)
-	lfd	%f7, -32-(15*8)(%r28)
-	lfd	%f8, -32-(14*8)(%r28)
-	lfd	%f9, -32-(13*8)(%r28)
-	lfd	%f10, -32-(12*8)(%r28)
-	lfd	%f11, -32-(11*8)(%r28)
-	lfd	%f12, -32-(10*8)(%r28)
-	lfd	%f13, -32-(9*8)(%r28)
+	addi	%r29, %r29, -(14*8)
+	lfd	%f1,  ( 1*8)(%r29)
+	lfd	%f2,  ( 2*8)(%r29)
+	lfd	%f3,  ( 3*8)(%r29)
+	lfd	%f4,  ( 4*8)(%r29)
+	lfd	%f5,  ( 5*8)(%r29)
+	lfd	%f6,  ( 6*8)(%r29)
+	lfd	%f7,  ( 7*8)(%r29)
+	lfd	%f8,  ( 8*8)(%r29)
+	lfd	%f9,  ( 9*8)(%r29)
+	lfd	%f10, (10*8)(%r29)
+	lfd	%f11, (11*8)(%r29)
+	lfd	%f12, (12*8)(%r29)
+	lfd	%f13, (13*8)(%r29)
 2:
 
+	/* Load all the vector registers.  */
+	bf-	3, 3f
+	addi	%r29, %r29, -16
+	lvx	%v13, 0, %r29
+	addi	%r29, %r29, -16
+	lvx	%v12, 0, %r29
+	addi	%r29, %r29, -16
+	lvx	%v11, 0, %r29
+	addi	%r29, %r29, -16
+	lvx	%v10, 0, %r29
+	addi	%r29, %r29, -16
+	lvx	%v9,  0, %r29
+	addi	%r29, %r29, -16
+	lvx	%v8,  0, %r29
+	addi	%r29, %r29, -16
+	lvx	%v7,  0, %r29
+	addi	%r29, %r29, -16
+	lvx	%v6,  0, %r29
+	addi	%r29, %r29, -16
+	lvx	%v5,  0, %r29
+	addi	%r29, %r29, -16
+	lvx	%v4,  0, %r29
+	addi	%r29, %r29, -16
+	lvx	%v3,  0, %r29
+	addi	%r29, %r29, -16
+	lvx	%v2,  0, %r29
+3:
+
 	/* Make the call.  */
 	ld	%r11, 8(%r28)
 	bctrl
@@ -152,6 +182,7 @@ ffi_call_LINUX64:
 	bt	31, .Lstruct_return_value
 	bt	30, .Ldone_return_value
 	bt	29, .Lfp_return_value
+	bt	28, .Lvec_return_value
 	std	%r3, 0(%r30)
 	/* Fall through...  */
 
@@ -167,12 +198,16 @@ ffi_call_LINUX64:
 	ld	%r31, -8(%r1)
 	blr
 
+.Lvec_return_value:
+	stvx	%v2, 0, %r30
+	b	.Ldone_return_value
+
 .Lfp_return_value:
 	.cfi_def_cfa_register 28
-	bf	28, .Lfloat_return_value
-	stfd	%f1, 0(%r30)
 	mtcrf	0x02, %r31 /* cr6  */
-	bf	27, .Ldone_return_value
+	bf	27, .Lfloat_return_value
+	stfd	%f1, 0(%r30)
+	bf	26, .Ldone_return_value
 	stfd	%f2, 8(%r30)
 	b	.Ldone_return_value
 .Lfloat_return_value:
@@ -180,8 +215,9 @@ ffi_call_LINUX64:
 	b	.Ldone_return_value
 
 .Lstruct_return_value:
-	bf	29, .Lsmall_struct
-	bf	28, .Lfloat_homog_return_value
+	bf	29, .Lvec_homog_or_small_struct
+	mtcrf	0x02, %r31 /* cr6  */
+	bf	27, .Lfloat_homog_return_value
 	stfd	%f1, 0(%r30)
 	stfd	%f2, 8(%r30)
 	stfd	%f3, 16(%r30)
@@ -203,6 +239,25 @@ ffi_call_LINUX64:
 	stfs	%f8, 28(%r30)
 	b	.Ldone_return_value
 
+.Lvec_homog_or_small_struct:
+	bf	28, .Lsmall_struct
+	stvx	%v2, 0, %r30
+	addi	%r30, %r30, 16
+	stvx	%v3, 0, %r30
+	addi	%r30, %r30, 16
+	stvx	%v4, 0, %r30
+	addi	%r30, %r30, 16
+	stvx	%v5, 0, %r30
+	addi	%r30, %r30, 16
+	stvx	%v6, 0, %r30
+	addi	%r30, %r30, 16
+	stvx	%v7, 0, %r30
+	addi	%r30, %r30, 16
+	stvx	%v8, 0, %r30
+	addi	%r30, %r30, 16
+	stvx	%v9, 0, %r30
+	b	.Ldone_return_value
+
 .Lsmall_struct:
 	std	%r3, 0(%r30)
 	std	%r4, 8(%r30)

data/ext/ffi_c/libffi/src/powerpc/linux64_closure.S

@@ -61,9 +61,15 @@ ffi_closure_LINUX64:
 # endif
 
 # if _CALL_ELF == 2
-#  32 byte special reg save area + 64 byte parm save area
-#  + 64 byte retval area + 13*8 fpr save area + round to 16
-#  define STACKFRAME 272
+#  ifdef __VEC__
+#   32 byte special reg save area + 64 byte parm save area
+#   + 128 byte retval area + 13*8 fpr save area + 12*16 vec save area + round to 16
+#   define STACKFRAME 528
+#  else
+#   32 byte special reg save area + 64 byte parm save area
+#   + 64 byte retval area + 13*8 fpr save area + round to 16
+#   define STACKFRAME 272
+#  endif
 #  define PARMSAVE 32
 #  define RETVAL PARMSAVE+64
 # else
@@ -146,6 +152,35 @@ ffi_closure_LINUX64:
 	# load up the pointer to the saved fpr registers
 	addi	%r8, %r1, -104
 
+# ifdef __VEC__
+	# load up the pointer to the saved vector registers
+	# 8 bytes padding for 16-byte alignment at -112(%r1)
+	addi	%r9, %r8, -24
+	stvx	%v13, 0, %r9
+	addi	%r9, %r9, -16
+	stvx	%v12, 0, %r9
+	addi	%r9, %r9, -16
+	stvx	%v11, 0, %r9
+	addi	%r9, %r9, -16
+	stvx	%v10, 0, %r9
+	addi	%r9, %r9, -16
+	stvx	%v9, 0, %r9
+	addi	%r9, %r9, -16
+	stvx	%v8, 0, %r9
+	addi	%r9, %r9, -16
+	stvx	%v7, 0, %r9
+	addi	%r9, %r9, -16
+	stvx	%v6, 0, %r9
+	addi	%r9, %r9, -16
+	stvx	%v5, 0, %r9
+	addi	%r9, %r9, -16
+	stvx	%v4, 0, %r9
+	addi	%r9, %r9, -16
+	stvx	%v3, 0, %r9
+	addi	%r9, %r9, -16
+	stvx	%v2, 0, %r9
+# endif
+
 	# load up the pointer to the result storage
 	addi	%r6, %r1, -STACKFRAME+RETVAL
 
@@ -313,6 +348,16 @@ ffi_closure_LINUX64:
 	.cfi_def_cfa_offset 0
 	blr
 	.cfi_def_cfa_offset STACKFRAME
+# case FFI_V2_TYPE_VECTOR
+	addi %r3, %r1, RETVAL
+	lvx %v2, 0, %r3
+	mtlr %r0
+	b .Lfinish
+# case FFI_V2_TYPE_VECTOR_HOMOG
+	addi %r3, %r1, RETVAL
+	lvx %v2, 0, %r3
+	addi %r3, %r3, 16
+	b .Lmorevector
 # case FFI_V2_TYPE_FLOAT_HOMOG
 	lfs %f1, RETVAL+0(%r1)
 	lfs %f2, RETVAL+4(%r1)
@@ -332,6 +377,25 @@ ffi_closure_LINUX64:
 	.cfi_def_cfa_offset 0
 	blr
 	.cfi_def_cfa_offset STACKFRAME
+.Lmorevector:
+	lvx %v3, 0, %r3
+	addi %r3, %r3, 16
+	lvx %v4, 0, %r3
+	addi %r3, %r3, 16
+	lvx %v5, 0, %r3
+	mtlr %r0
+	addi %r3, %r3, 16
+	lvx %v6, 0, %r3
+	addi %r3, %r3, 16
+	lvx %v7, 0, %r3
+	addi %r3, %r3, 16
+	lvx %v8, 0, %r3
+	addi %r3, %r3, 16
+	lvx %v9, 0, %r3
+	addi %r1, %r1, STACKFRAME
+	.cfi_def_cfa_offset 0
+	blr
+	.cfi_def_cfa_offset STACKFRAME
 .Lmorefloat:
 	lfs %f4, RETVAL+12(%r1)
 	mtlr %r0

data/ext/ffi_c/libffi/src/powerpc/sysv.S

@@ -104,17 +104,16 @@ ENTRY(ffi_call_SYSV)
 	bctrl
 
 	/* Now, deal with the return value.  */
-	mtcrf	0x01,%r31 /* cr7  */
+	mtcrf	0x03,%r31 /* cr6-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)
+	bf+	27,L(done_return_value)
 	stw	%r4,4(%r30)
-	mtcrf	0x02,%r31 /* cr6  */
-	bf	27,L(done_return_value)
+	bf	26,L(done_return_value)
 	stw     %r5,8(%r30)
 	stw	%r6,12(%r30)
 	/* Fall through...  */
@@ -145,10 +144,9 @@ L(done_return_value):
 #ifndef __NO_FPRS__
 L(fp_return_value):
 	.cfi_restore_state
-	bf	28,L(float_return_value)
+	bf	27,L(float_return_value)
 	stfd	%f1,0(%r30)
-	mtcrf   0x02,%r31 /* cr6  */
-	bf	27,L(done_return_value)
+	bf	26,L(done_return_value)
 	stfd	%f2,8(%r30)
 	b	L(done_return_value)
 L(float_return_value):

data/ext/ffi_c/libffi/src/x86/ffi.c

@@ -557,13 +557,16 @@ ffi_prep_closure_loc (ffi_closure* closure,
       return FFI_BAD_ABI;
     }
 
+  /* endbr32.  */
+  *(UINT32 *) tramp = 0xfb1e0ff3;
+
   /* movl or pushl immediate.  */
-  tramp[0] = op;
-  *(void **)(tramp + 1) = codeloc;
+  tramp[4] = op;
+  *(void **)(tramp + 5) = codeloc;
 
   /* jmp dest */
-  tramp[5] = 0xe9;
-  *(unsigned *)(tramp + 6) = (unsigned)dest - ((unsigned)codeloc + 10);
+  tramp[9] = 0xe9;
+  *(unsigned *)(tramp + 10) = (unsigned)dest - ((unsigned)codeloc + 14);
 
   closure->cif = cif;
   closure->fun = fun;

data/ext/ffi_c/libffi/src/x86/ffi64.c

@@ -728,13 +728,15 @@ ffi_prep_closure_loc (ffi_closure* closure,
 		      void *user_data,
 		      void *codeloc)
 {
-  static const unsigned char trampoline[16] = {
-    /* leaq  -0x7(%rip),%r10   # 0x0  */
-    0x4c, 0x8d, 0x15, 0xf9, 0xff, 0xff, 0xff,
-    /* jmpq  *0x3(%rip)        # 0x10 */
-    0xff, 0x25, 0x03, 0x00, 0x00, 0x00,
-    /* nopl  (%rax) */
-    0x0f, 0x1f, 0x00
+  static const unsigned char trampoline[24] = {
+    /* endbr64 */
+    0xf3, 0x0f, 0x1e, 0xfa,
+    /* leaq  -0xb(%rip),%r10   # 0x0  */
+    0x4c, 0x8d, 0x15, 0xf5, 0xff, 0xff, 0xff,
+    /* jmpq  *0x7(%rip)        # 0x18 */
+    0xff, 0x25, 0x07, 0x00, 0x00, 0x00,
+    /* nopl  0(%rax) */
+    0x0f, 0x1f, 0x80, 0x00, 0x00, 0x00, 0x00
   };
   void (*dest)(void);
   char *tramp = closure->tramp;
@@ -752,7 +754,7 @@ ffi_prep_closure_loc (ffi_closure* closure,
     dest = ffi_closure_unix64;
 
   memcpy (tramp, trampoline, sizeof(trampoline));
-  *(UINT64 *)(tramp + 16) = (uintptr_t)dest;
+  *(UINT64 *)(tramp + sizeof (trampoline)) = (uintptr_t)dest;
 
   closure->cif = cif;
   closure->fun = fun;

data/ext/ffi_c/libffi/src/x86/ffitarget.h

@@ -136,12 +136,25 @@ typedef enum ffi_abi {
 
 #if defined (X86_64) || defined(X86_WIN64) \
     || (defined (__x86_64__) && defined (X86_DARWIN))
-# define FFI_TRAMPOLINE_SIZE 24
+/* 4 bytes of ENDBR64 + 7 bytes of LEA + 6 bytes of JMP + 7 bytes of NOP
+   + 8 bytes of pointer.  */
+# define FFI_TRAMPOLINE_SIZE 32
 # define FFI_NATIVE_RAW_API 0
 #else
-# define FFI_TRAMPOLINE_SIZE 12
+/* 4 bytes of ENDBR32 + 5 bytes of MOV + 5 bytes of JMP + 2 unused
+   bytes.  */
+# define FFI_TRAMPOLINE_SIZE 16
 # define FFI_NATIVE_RAW_API 1  /* x86 has native raw api support */
 #endif
 
+#if !defined(GENERATE_LIBFFI_MAP) && defined(__ASSEMBLER__) \
+    && defined(__CET__)
+# include <cet.h>
+# define _CET_NOTRACK notrack
+#else
+# define _CET_ENDBR
+# define _CET_NOTRACK
+#endif
+
 #endif
 

data/ext/ffi_c/libffi/src/x86/ffiw64.c

@@ -196,13 +196,15 @@ EFI64(ffi_prep_closure_loc)(ffi_closure* closure,
 		      void *user_data,
 		      void *codeloc)
 {
-  static const unsigned char trampoline[16] = {
-    /* leaq  -0x7(%rip),%r10   # 0x0  */
-    0x4c, 0x8d, 0x15, 0xf9, 0xff, 0xff, 0xff,
-    /* jmpq  *0x3(%rip)        # 0x10 */
-    0xff, 0x25, 0x03, 0x00, 0x00, 0x00,
-    /* nopl  (%rax) */
-    0x0f, 0x1f, 0x00
+  static const unsigned char trampoline[FFI_TRAMPOLINE_SIZE - 8] = {
+    /* endbr64 */
+    0xf3, 0x0f, 0x1e, 0xfa,
+    /* leaq  -0xb(%rip),%r10   # 0x0  */
+    0x4c, 0x8d, 0x15, 0xf5, 0xff, 0xff, 0xff,
+    /* jmpq  *0x7(%rip)        # 0x18 */
+    0xff, 0x25, 0x07, 0x00, 0x00, 0x00,
+    /* nopl  0(%rax) */
+    0x0f, 0x1f, 0x80, 0x00, 0x00, 0x00, 0x00
   };
   char *tramp = closure->tramp;
 
@@ -216,7 +218,7 @@ EFI64(ffi_prep_closure_loc)(ffi_closure* closure,
     }
 
   memcpy (tramp, trampoline, sizeof(trampoline));
-  *(UINT64 *)(tramp + 16) = (uintptr_t)ffi_closure_win64;
+  *(UINT64 *)(tramp + sizeof (trampoline)) = (uintptr_t)ffi_closure_win64;
 
   closure->cif = cif;
   closure->fun = fun;

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

@@ -92,6 +92,7 @@
 ffi_call_i386:
 L(UW0):
 	# cfi_startproc
+	_CET_ENDBR
 #if !HAVE_FASTCALL
 	movl	4(%esp), %ecx
 	movl	8(%esp), %edx
@@ -133,7 +134,7 @@ L(pc1):
 	leal	L(store_table)(,%ecx, 8), %ebx
 #endif
 	movl	16(%ebp), %ecx		/* load result address */
-	jmp	*%ebx
+	_CET_NOTRACK jmp *%ebx
 
 	.balign	8
 L(store_table):
@@ -256,7 +257,7 @@ ENDF(ffi_call_i386)
 	andl	$X86_RET_TYPE_MASK, %eax;				\
 	leal	L(C1(load_table,N))(, %eax, 8), %edx;			\
 	movl	closure_CF(%esp), %eax;		/* optimiztic load */	\
-	jmp	*%edx
+	_CET_NOTRACK jmp *%edx
 
 #ifdef __PIC__
 # if defined X86_DARWIN || defined HAVE_HIDDEN_VISIBILITY_ATTRIBUTE
@@ -267,7 +268,7 @@ ENDF(ffi_call_i386)
 L(C1(pc,N)):								\
 	leal	L(C1(load_table,N))-L(C1(pc,N))(%edx, %eax, 8), %edx;	\
 	movl	closure_CF(%esp), %eax;		/* optimiztic load */	\
-	jmp	*%edx
+	_CET_NOTRACK jmp *%edx
 # else
 #  define FFI_CLOSURE_CALL_INNER_SAVE_EBX
 #  undef FFI_CLOSURE_CALL_INNER
@@ -286,7 +287,7 @@ L(C1(UW,UWN)):								\
 L(C1(UW,UWN)):								\
 	/* cfi_restore(%ebx); */					\
 	movl	closure_CF(%esp), %eax;		/* optimiztic load */	\
-	jmp	*%edx
+	_CET_NOTRACK jmp *%edx
 # endif /* DARWIN || HIDDEN */
 #endif /* __PIC__ */
 
@@ -296,6 +297,7 @@ L(C1(UW,UWN)):								\
 C(ffi_go_closure_EAX):
 L(UW6):
 	# cfi_startproc
+	_CET_ENDBR
 	subl	$closure_FS, %esp
 L(UW7):
 	# cfi_def_cfa_offset(closure_FS + 4)
@@ -316,6 +318,7 @@ ENDF(C(ffi_go_closure_EAX))
 C(ffi_go_closure_ECX):
 L(UW9):
 	# cfi_startproc
+	_CET_ENDBR
 	subl	$closure_FS, %esp
 L(UW10):
 	# cfi_def_cfa_offset(closure_FS + 4)
@@ -340,6 +343,7 @@ ENDF(C(ffi_go_closure_ECX))
 C(ffi_closure_i386):
 L(UW12):
 	# cfi_startproc
+	_CET_ENDBR
 	subl	$closure_FS, %esp
 L(UW13):
 	# cfi_def_cfa_offset(closure_FS + 4)
@@ -423,6 +427,7 @@ ENDF(C(ffi_closure_i386))
 C(ffi_go_closure_STDCALL):
 L(UW21):
 	# cfi_startproc
+	_CET_ENDBR
 	subl	$closure_FS, %esp
 L(UW22):
 	# cfi_def_cfa_offset(closure_FS + 4)
@@ -448,6 +453,7 @@ L(UW24):
 	# cfi_startproc
 	# cfi_def_cfa(%esp, 8)
 	# cfi_offset(%eip, -8)
+	_CET_ENDBR
 	subl	$closure_FS-4, %esp
 L(UW25):
 	# cfi_def_cfa_offset(closure_FS + 4)
@@ -470,6 +476,7 @@ ENDF(C(ffi_closure_REGISTER))
 C(ffi_closure_STDCALL):
 L(UW27):
 	# cfi_startproc
+	_CET_ENDBR
 	subl	$closure_FS, %esp
 L(UW28):
 	# cfi_def_cfa_offset(closure_FS + 4)
@@ -576,6 +583,7 @@ ENDF(C(ffi_closure_STDCALL))
 C(ffi_closure_raw_SYSV):
 L(UW32):
 	# cfi_startproc
+	_CET_ENDBR
 	subl	$raw_closure_S_FS, %esp
 L(UW33):
 	# cfi_def_cfa_offset(raw_closure_S_FS + 4)
@@ -679,6 +687,7 @@ ENDF(C(ffi_closure_raw_SYSV))
 C(ffi_closure_raw_THISCALL):
 L(UW41):
 	# cfi_startproc
+	_CET_ENDBR
 	/* Rearrange the stack such that %ecx is the first argument.
 	   This means moving the return address.  */
 	popl	%edx