Merge branch 'dev'v6.70

10 files changed, 42 insertions, 22 deletions

diff --git a/CMakeLists.txt b/CMakeLists.txt
index 15ed204..12323ef 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -1,6 +1,6 @@
 cmake_minimum_required(VERSION 2.6...3.0.2)
 
-project(xbyak LANGUAGES CXX VERSION 6.69.2)
+project(xbyak LANGUAGES CXX VERSION 6.70)
 
 file(GLOB headers xbyak/*.h)
 
diff --git a/doc/changelog.md b/doc/changelog.md
index acdccd8..20e6e09 100644
--- a/doc/changelog.md
+++ b/doc/changelog.md
@@ -1,5 +1,6 @@
 # History
 
+* 2023/Jul/05 ver 6.70 add alias of vclmulqdq, correct alias of pclmulqdq
 * 2023/Jun/27 ver 6.69.2 add constexpr to `TypeT operator|` (thanks to Wunkolo)
 * 2023/Mar/23 ver 6.69.1 add detection of xsave (thanks to Wunkolo)
 * 2023/Feb/20 ver 6.69 util::Cpu supports AMD CPUs. support UINTR
diff --git a/gen/gen_code.cpp b/gen/gen_code.cpp
index a074db3..f638016 100644
--- a/gen/gen_code.cpp
+++ b/gen/gen_code.cpp
@@ -974,12 +974,13 @@ void put()
 		} tbl[] = {
 			{ "pclmullqlqdq", 0 },
 			{ "pclmulhqlqdq", 1 },
-			{ "pclmullqhdq", 0x10 },
-			{ "pclmulhqhdq", 0x11 },
+			{ "pclmullqhqdq", 0x10 },
+			{ "pclmulhqhqdq", 0x11 },
 		};
 		for (size_t i = 0; i < NUM_OF_ARRAY(tbl); i++) {
 			const Tbl *p = &tbl[i];
 			printf("void %s(const Xmm& xmm, const Operand& op) { pclmulqdq(xmm, op, 0x%02X); }\n", p->name, p->code);
+			printf("void v%s(const Xmm& x1, const Xmm& x2, const Operand& op) { vpclmulqdq(x1, x2, op, 0x%02X); }\n", p->name, p->code);
 		}
 	}
 	{
diff --git a/meson.build b/meson.build
index 53afa96..325cdc5 100644
--- a/meson.build
+++ b/meson.build
@@ -5,7 +5,7 @@
 project(
 	'xbyak',
 	'cpp',
-	version: '6.69.2',
+	version: '6.70',
 	license: 'BSD-3-Clause',
 	default_options: 'b_ndebug=if-release'
 )
diff --git a/readme.md b/readme.md
index f318608..843287a 100644
--- a/readme.md
+++ b/readme.md
@@ -1,5 +1,5 @@
 
-# Xbyak 6.69.2 [![Badge Build]][Build Status]
+# Xbyak 6.70 [![Badge Build]][Build Status]
 
 *A C++ JIT assembler for x86 (IA32), x64 (AMD64, x86-64)*
 
diff --git a/readme.txt b/readme.txt
index 48e9ab7..a165c23 100644
--- a/readme.txt
+++ b/readme.txt
@@ -1,5 +1,5 @@
 

-    C++用x86(IA-32), x64(AMD64, x86-64) JITアセンブラ Xbyak 6.69.2

+    C++用x86(IA-32), x64(AMD64, x86-64) JITアセンブラ Xbyak 6.70

 

 -----------------------------------------------------------------------------

 ◎概要

@@ -402,6 +402,7 @@ sample/{echo,hello}.bfは http://www.kmonos.net/alang/etc/brainfuck.php から
 -----------------------------------------------------------------------------

 ◎履歴

 

+2023/07/05 ver 6.70 vpclmulqdqのailas追加

 2023/06/27 ver 6.69.2 `TypeT operator|`にconstexpr追加(thanks to Wunkolo)

 2023/03/23 ver 6.69.1 xsave判定追加(thanks to Wunkolo)

 2023/02/20 ver 6.69 util::CpuがAMD対応 UINTR命令対応

diff --git a/test/make_nm.cpp b/test/make_nm.cpp
index e5939eb..ec627af 100644
--- a/test/make_nm.cpp
+++ b/test/make_nm.cpp
@@ -1493,18 +1493,6 @@ class Test {
 				put(p, XMM, XMM|MEM, IMM);
 			}
 		}
-		{
-			const char tbl[][16] = {
-				"pclmullqlqdq",
-				"pclmulhqlqdq",
-//				"pclmullqhdq", // QQQ : not supported by nasm/yasm
-//				"pclmulhqhdq",
-			};
-			for (size_t i = 0; i < NUM_OF_ARRAY(tbl); i++) {
-				const char *p = tbl[i];
-				put(p, XMM, XMM|MEM);
-			}
-		}
 		put("extractps", REG32e|MEM, XMM, IMM);
 		put("pextrw", REG32e|MEM, XMM, IMM); // pextrw for REG32 is for MMX2
 		put("pextrb", REG32e|MEM, XMM, IMM);
@@ -1522,6 +1510,23 @@ class Test {
 #endif
 
 	}
+	void putVpclmulqdq()
+	{
+		const char tbl[][16] = {
+			"vpclmullqlqdq",
+			"vpclmulhqlqdq",
+			"vpclmullqhqdq",
+			"vpclmulhqhqdq",
+		};
+		for (size_t i = 0; i < NUM_OF_ARRAY(tbl); i++) {
+			const char *p = tbl[i] + 1; // remove the top 'v'
+			put(p, XMM, XMM|MEM);
+			p = tbl[i]; // use the top 'v'
+			put(p, XMM, XMM, XMM|MEM);
+			put(p, YMM, YMM, YMM|MEM);
+			put(p, ZMM, ZMM, ZMM|MEM);
+		}
+	}
 	void putSHA() const
 	{
 		put("sha1rnds4", XMM, XMM|MEM, IMM);
@@ -2569,6 +2574,7 @@ public:
 		putPushPop8_16();
 #else
 		putSIMPLE();
+		putVpclmulqdq();
 		putReg1();
 		putBt();
 		putRorM();
diff --git a/xbyak/xbyak.h b/xbyak/xbyak.h
index 7660f3a..80627f1 100644
--- a/xbyak/xbyak.h
+++ b/xbyak/xbyak.h
@@ -155,7 +155,7 @@ namespace Xbyak {
 
 enum {
 	DEFAULT_MAX_CODE_SIZE = 4096,
-	VERSION = 0x6692 /* 0xABCD = A.BC(.D) */
+	VERSION = 0x6700 /* 0xABCD = A.BC(.D) */
 };
 
 #ifndef MIE_INTEGER_TYPE_DEFINED
diff --git a/xbyak/xbyak_mnemonic.h b/xbyak/xbyak_mnemonic.h
index 1a2ec1c..7b38aa1 100644
--- a/xbyak/xbyak_mnemonic.h
+++ b/xbyak/xbyak_mnemonic.h
@@ -1,4 +1,4 @@
-const char *getVersionString() const { return "6.69.2"; }
+const char *getVersionString() const { return "6.70"; }
 void aadd(const Address& addr, const Reg32e &reg) { opModM(addr, reg, 0x0F, 0x38, 0x0FC); }
 void aand(const Address& addr, const Reg32e &reg) { db(0x66); opModM(addr, reg, 0x0F, 0x38, 0x0FC); }
 void adc(const Operand& op, uint32_t imm) { opRM_I(op, imm, 0x10, 2); }
@@ -589,9 +589,9 @@ void pavgb(const Mmx& mmx, const Operand& op) { opMMX(mmx, op, 0xE0); }
 void pavgw(const Mmx& mmx, const Operand& op) { opMMX(mmx, op, 0xE3); }
 void pblendvb(const Xmm& xmm, const Operand& op) { opGen(xmm, op, 0x10, 0x66, isXMM_XMMorMEM, NONE, 0x38); }
 void pblendw(const Xmm& xmm, const Operand& op, int imm) { opGen(xmm, op, 0x0E, 0x66, isXMM_XMMorMEM, static_cast<uint8_t>(imm), 0x3A); }
-void pclmulhqhdq(const Xmm& xmm, const Operand& op) { pclmulqdq(xmm, op, 0x11); }
+void pclmulhqhqdq(const Xmm& xmm, const Operand& op) { pclmulqdq(xmm, op, 0x11); }
 void pclmulhqlqdq(const Xmm& xmm, const Operand& op) { pclmulqdq(xmm, op, 0x01); }
-void pclmullqhdq(const Xmm& xmm, const Operand& op) { pclmulqdq(xmm, op, 0x10); }
+void pclmullqhqdq(const Xmm& xmm, const Operand& op) { pclmulqdq(xmm, op, 0x10); }
 void pclmullqlqdq(const Xmm& xmm, const Operand& op) { pclmulqdq(xmm, op, 0x00); }
 void pclmulqdq(const Xmm& xmm, const Operand& op, int imm) { opGen(xmm, op, 0x44, 0x66, isXMM_XMMorMEM, static_cast<uint8_t>(imm), 0x3A); }
 void pcmpeqb(const Mmx& mmx, const Operand& op) { opMMX(mmx, op, 0x74); }
@@ -1192,6 +1192,10 @@ void vpbroadcastb(const Xmm& x, const Operand& op) { if (!(op.isXMM() || op.isME
 void vpbroadcastd(const Xmm& x, const Operand& op) { if (!(op.isXMM() || op.isMEM())) XBYAK_THROW(ERR_BAD_COMBINATION) opAVX_X_XM_IMM(x, op, T_N4 | T_66 | T_0F38 | T_W0 | T_YMM | T_EVEX, 0x58); }
 void vpbroadcastq(const Xmm& x, const Operand& op) { if (!(op.isXMM() || op.isMEM())) XBYAK_THROW(ERR_BAD_COMBINATION) opAVX_X_XM_IMM(x, op, T_N8 | T_66 | T_0F38 | T_W0 | T_EW1 | T_YMM | T_EVEX, 0x59); }
 void vpbroadcastw(const Xmm& x, const Operand& op) { if (!(op.isXMM() || op.isMEM())) XBYAK_THROW(ERR_BAD_COMBINATION) opAVX_X_XM_IMM(x, op, T_N2 | T_66 | T_0F38 | T_W0 | T_YMM | T_EVEX, 0x79); }
+void vpclmulhqhqdq(const Xmm& x1, const Xmm& x2, const Operand& op) { vpclmulqdq(x1, x2, op, 0x11); }
+void vpclmulhqlqdq(const Xmm& x1, const Xmm& x2, const Operand& op) { vpclmulqdq(x1, x2, op, 0x01); }
+void vpclmullqhqdq(const Xmm& x1, const Xmm& x2, const Operand& op) { vpclmulqdq(x1, x2, op, 0x10); }
+void vpclmullqlqdq(const Xmm& x1, const Xmm& x2, const Operand& op) { vpclmulqdq(x1, x2, op, 0x00); }
 void vpclmulqdq(const Xmm& x1, const Xmm& x2, const Operand& op, uint8_t imm) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F3A | T_W0 | T_YMM | T_EVEX, 0x44, imm); }
 void vpcmpeqb(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F | T_YMM, 0x74); }
 void vpcmpeqd(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F | T_YMM, 0x76); }
diff --git a/xbyak/xbyak_util.h b/xbyak/xbyak_util.h
index d56b105..12c8b36 100644
--- a/xbyak/xbyak_util.h
+++ b/xbyak/xbyak_util.h
@@ -9,6 +9,13 @@
 	#define XBYAK_THROW(x) ;
 	#define XBYAK_THROW_RET(x, y) return y;
 #endif
+#ifndef XBYAK_CONSTEXPR
+#if ((__cplusplus >= 201402L) && !(!defined(__clang__) && defined(__GNUC__) && (__GNUC__ <= 5))) || (defined(_MSC_VER) && _MSC_VER >= 1910)
+	#define XBYAK_CONSTEXPR constexpr
+#else
+	#define XBYAK_CONSTEXPR
+#endif
+#endif
 #else
 #include <string.h>