Merge branch 'dev'v7.08

21 files changed, 89 insertions, 21 deletions

diff --git a/CMakeLists.txt b/CMakeLists.txt
index 70c4314..6baf35e 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -1,6 +1,6 @@
 cmake_minimum_required(VERSION 3.5)
 
-project(xbyak LANGUAGES CXX VERSION 7.07.1)
+project(xbyak LANGUAGES CXX VERSION 7.08)
 
 file(GLOB headers xbyak/*.h)
 
diff --git a/doc/changelog.md b/doc/changelog.md
index 7b390a8..1f5c2c3 100644
--- a/doc/changelog.md
+++ b/doc/changelog.md
@@ -1,5 +1,6 @@
 # History
 
+* 2024/Oct/07 ver 7.08 support rdfsbase etc.
 * 2024/Aug/29 ver 7.07.1 adapt to NASM 2.16.03 output of xchg (The functionality stays the same.)
 * 2024/Jun/11 ver 7.07 support xresldtrk/xsusldtrk
 * 2024/Mar/07 ver 7.06 Xbyak::util::Cpu supports AMD processor
diff --git a/gen/gen_code.cpp b/gen/gen_code.cpp
index e43c9b2..ee8494f 100644
--- a/gen/gen_code.cpp
+++ b/gen/gen_code.cpp
@@ -2057,6 +2057,15 @@ void put64()
 	// encodekey
 	puts("void encodekey128(const Reg32& r1, const Reg32& r2) { opEncodeKey(r1, r2, 0xFA, 0xDA); }");
 	puts("void encodekey256(const Reg32& r1, const Reg32& r2) { opEncodeKey(r1, r2, 0xFB, 0xDB); }");
+	// read/write fs/gs
+	{
+		const char *tbl[] = {
+			"rdfs", "rdgs", "wrfs", "wrgs"
+		};
+		for (int i = 0; i < 4; i++) {
+			printf("void %sbase(const Reg32e& r) { opRR(%s, r, T_F3|T_0F, 0xAE); }\n", tbl[i], Reg(i, Operand::REG, 32).toString());
+		}
+	}
 }
 
 void putAMX_TILE()
diff --git a/meson.build b/meson.build
index f53bea2..b5a0ce7 100644
--- a/meson.build
+++ b/meson.build
@@ -5,7 +5,7 @@
 project(
 	'xbyak',
 	'cpp',
-	version: '7.07.1',
+	version: '7.08',
 	license: 'BSD-3-Clause',
 	default_options: 'b_ndebug=if-release'
 )
diff --git a/readme.md b/readme.md
index cede3ea..6cab2b8 100644
--- a/readme.md
+++ b/readme.md
@@ -1,5 +1,5 @@
 
-# Xbyak 7.07.1 [![Badge Build]][Build Status]
+# Xbyak 7.08 [![Badge Build]][Build Status]
 
 *A C++ JIT assembler for x86 (IA32), x64 (AMD64, x86-64)*
 
diff --git a/readme.txt b/readme.txt
index 1acca54..5bae6c4 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 7.07.1

+    C++用x86(IA-32), x64(AMD64, x86-64) JITアセンブラ Xbyak 7.08

 

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

 ◎概要

@@ -404,6 +404,7 @@ sample/{echo,hello}.bfは http://www.kmonos.net/alang/etc/brainfuck.php から
 -----------------------------------------------------------------------------

 ◎履歴

 

+2024/10/17 ver 7.08 rdfabaseなどサポート

 2024/08/29 ver 7.07.1 xchgの仕様をnasm 2.16.03の挙動に合わせる。

 2024/06/11 ver 7.07 xresldtrk/xsusldtrkサポート

 2024/03/07 ver 7.06 util::Cpuのキャッシュ判定周りがAMD CPU対応

diff --git a/sample/bf.cpp b/sample/bf.cpp
index 657c87c..06f1dac 100644
--- a/sample/bf.cpp
+++ b/sample/bf.cpp
@@ -1,4 +1,3 @@
-#define XBYAK_NO_OP_NAMES
 #include "xbyak/xbyak.h"
 #include <stdio.h>
 #include <stdlib.h>
@@ -189,18 +188,30 @@ int main(int argc, char *argv[])
 	fprintf(stderr, "64bit mode\n");
 #endif
 	if (argc == 1) {
-		fprintf(stderr, "bf filename.bf [0|1]\n");
+		fprintf(stderr, "bf filename.bf [0|1|2]\n");
 		return 1;
 	}
 	std::ifstream ifs(argv[1]);
 	int mode = argc == 3 ? atoi(argv[2]) : 0;
 	try {
 		Brainfuck bf(ifs);
-		if (mode == 0) {
+		switch (mode) {
+		case 0: {
 			static int stack[128 * 1024];
 			bf.getCode<void (*)(const void*, const void*, int *)>()(reinterpret_cast<const void*>(putchar), reinterpret_cast<const void*>(getchar), stack);
-		} else {
+			break;
+			}
+		case 1: {
 			dump(bf.getCode(), bf.getSize());
+			break;
+			}
+		default: {
+			const char *dumpName = "bf.dump";
+			printf("dump to %s\n", dumpName);
+			std::ofstream ofs(dumpName, std::ios::binary);
+			ofs.write((const char*)bf.getCode(), bf.getSize());
+			break;
+			}
 		}
 	} catch (std::exception& e) {
 		printf("ERR:%s\n", e.what());
diff --git a/sample/calc.cpp b/sample/calc.cpp
index 29ad9c3..d543d04 100644
--- a/sample/calc.cpp
+++ b/sample/calc.cpp
@@ -10,7 +10,6 @@
 #include <stdio.h>
 #include <sstream>
 #include <map>
-#define XBYAK_NO_OP_NAMES
 #include "xbyak/xbyak.h"
 #ifdef _MSC_VER
 	#pragma warning(disable : 4127) // for boost(constant condition)
diff --git a/sample/calc2.cpp b/sample/calc2.cpp
index a13d2cf..62fb900 100644
--- a/sample/calc2.cpp
+++ b/sample/calc2.cpp
@@ -22,7 +22,6 @@
 #include <assert.h>
 #include <string>
 #include <vector>
-#define XBYAK_NO_OP_NAMES
 #include "xbyak/xbyak.h"
 
 enum Operand {
diff --git a/sample/jmp_table.cpp b/sample/jmp_table.cpp
index fb47cac..4e0ca6e 100644
--- a/sample/jmp_table.cpp
+++ b/sample/jmp_table.cpp
@@ -2,7 +2,6 @@
 	sample of move(reg, LABEL);, L(LABEL), putL(LABEL);
 */
 #include <stdio.h>
-#define XBYAK_NO_OP_NAMES
 #include <xbyak/xbyak.h>
 
 const int expectTbl[] = {
diff --git a/sample/memfd.cpp b/sample/memfd.cpp
index 2b6d08b..312e7ca 100644
--- a/sample/memfd.cpp
+++ b/sample/memfd.cpp
@@ -25,6 +25,7 @@ public:
 
 int main()
 {
+	printf("pid=%d\n", getpid());
     Code c1("Xbyak::abc", 123);
     Code c2("Xbyak::xyz", 456);
     printf("c1 %d\n", c1.getCode<int (*)()>()());
@@ -36,4 +37,5 @@ int main()
 			printf("%s\n", line.c_str());
 		}
 	}
+	getchar();
 }
diff --git a/sample/memfunc.cpp b/sample/memfunc.cpp
index b97630b..65b0f04 100644
--- a/sample/memfunc.cpp
+++ b/sample/memfunc.cpp
@@ -1,7 +1,6 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include <memory.h>
-#define XBYAK_NO_OP_NAMES
 #include <xbyak/xbyak.h>
 
 struct A {
diff --git a/sample/protect-re.cpp b/sample/protect-re.cpp
index 6eaa863..a8f77cd 100644
--- a/sample/protect-re.cpp
+++ b/sample/protect-re.cpp
@@ -1,4 +1,3 @@
-#define XBYAK_NO_OP_NAMES
 #include <xbyak/xbyak.h>
 
 struct Code1 : Xbyak::CodeGenerator {
diff --git a/sample/quantize.cpp b/sample/quantize.cpp
index ba0fd22..f45337c 100644
--- a/sample/quantize.cpp
+++ b/sample/quantize.cpp
@@ -42,7 +42,6 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include <time.h>
-#define XBYAK_NO_OP_NAMES
 #include "xbyak/xbyak.h"
 #ifdef _MSC_VER
 	#pragma warning(disable : 4996) // scanf
diff --git a/sample/stackframe.cpp b/sample/stackframe.cpp
index 7405b40..f324bde 100644
--- a/sample/stackframe.cpp
+++ b/sample/stackframe.cpp
@@ -1,4 +1,3 @@
-#define XBYAK_NO_OP_NAMES
 #include <xbyak/xbyak_util.h>
 
 #ifdef XBYAK32
diff --git a/sample/static_buf.cpp b/sample/static_buf.cpp
index 0a8ff57..0d87101 100644
--- a/sample/static_buf.cpp
+++ b/sample/static_buf.cpp
@@ -2,7 +2,6 @@
 	sample to use static memory
 */
 #include <stdio.h>
-#define XBYAK_NO_OP_NAMES
 #include "xbyak/xbyak.h"
 
 MIE_ALIGN(4096) char buf[4096];
diff --git a/sample/test0.cpp b/sample/test0.cpp
index afbaf83..924c816 100644
--- a/sample/test0.cpp
+++ b/sample/test0.cpp
@@ -7,7 +7,6 @@
 #endif
 #include <stdio.h>
 #include <stdlib.h>
-#define XBYAK_NO_OP_NAMES
 #include "xbyak/xbyak.h"
 
 class Sample : public Xbyak::CodeGenerator {
diff --git a/sample/toyvm.cpp b/sample/toyvm.cpp
index dff0cb7..b6549a1 100644
--- a/sample/toyvm.cpp
+++ b/sample/toyvm.cpp
@@ -27,7 +27,6 @@
 #include <stdlib.h>
 #include <memory.h>
 #include <vector>
-#define XBYAK_NO_OP_NAMES
 #include "xbyak/xbyak.h"
 #include "xbyak/xbyak_util.h"
 #define NUM_OF_ARRAY(x) (sizeof(x) / sizeof(x[0]))
diff --git a/test/misc.cpp b/test/misc.cpp
index 1457691..a15721f 100644
--- a/test/misc.cpp
+++ b/test/misc.cpp
@@ -2284,5 +2284,55 @@ CYBOZU_TEST_AUTO(avx_vnni_int)
 	CYBOZU_TEST_EQUAL_ARRAY(c.getCode(), tbl, n);
 }
 
-
+CYBOZU_TEST_AUTO(rdfsbase)
+{
+	struct Code : Xbyak::CodeGenerator {
+		Code()
+		{
+			rdfsbase(rax);
+			rdfsbase(r15d);
+			rdfsbase(r30d);
+			rdfsbase(r31);
+
+			rdgsbase(rax);
+			rdgsbase(r15d);
+			rdgsbase(r30d);
+			rdgsbase(r31);
+
+			wrfsbase(rax);
+			wrfsbase(r15d);
+			wrfsbase(r30d);
+			wrfsbase(r31);
+
+			wrgsbase(rax);
+			wrgsbase(r15d);
+			wrgsbase(r30d);
+			wrgsbase(r31);
+		}
+	} c;
+	const uint8_t tbl[] = {
+		0xf3, 0x48, 0x0f, 0xae, 0xc0,
+		0xf3, 0x41, 0x0f, 0xae, 0xc7,
+		0xf3, 0xd5, 0x91, 0xae, 0xc6,
+		0xf3, 0xd5, 0x99, 0xae, 0xc7,
+
+		0xf3, 0x48, 0x0f, 0xae, 0xc8,
+		0xf3, 0x41, 0x0f, 0xae, 0xcf,
+		0xf3, 0xd5, 0x91, 0xae, 0xce,
+		0xf3, 0xd5, 0x99, 0xae, 0xcf,
+
+		0xf3, 0x48, 0x0f, 0xae, 0xd0,
+		0xf3, 0x41, 0x0f, 0xae, 0xd7,
+		0xf3, 0xd5, 0x91, 0xae, 0xd6,
+		0xf3, 0xd5, 0x99, 0xae, 0xd7,
+
+		0xf3, 0x48, 0x0f, 0xae, 0xd8,
+		0xf3, 0x41, 0x0f, 0xae, 0xdf,
+		0xf3, 0xd5, 0x91, 0xae, 0xde,
+		0xf3, 0xd5, 0x99, 0xae, 0xdf,
+	};
+	const size_t n = sizeof(tbl) / sizeof(tbl[0]);
+	CYBOZU_TEST_EQUAL(c.getSize(), n);
+	CYBOZU_TEST_EQUAL_ARRAY(c.getCode(), tbl, n);
+}
 #endif
diff --git a/xbyak/xbyak.h b/xbyak/xbyak.h
index e67de82..292fcc7 100644
--- a/xbyak/xbyak.h
+++ b/xbyak/xbyak.h
@@ -155,7 +155,7 @@ namespace Xbyak {
 
 enum {
 	DEFAULT_MAX_CODE_SIZE = 4096,
-	VERSION = 0x7071 /* 0xABCD = A.BC(.D) */
+	VERSION = 0x7080 /* 0xABCD = A.BC(.D) */
 };
 
 #ifndef MIE_INTEGER_TYPE_DEFINED
diff --git a/xbyak/xbyak_mnemonic.h b/xbyak/xbyak_mnemonic.h
index 6eb3d63..06eb6d0 100644
--- a/xbyak/xbyak_mnemonic.h
+++ b/xbyak/xbyak_mnemonic.h
@@ -1,4 +1,4 @@
-const char *getVersionString() const { return "7.07.1"; }
+const char *getVersionString() const { return "7.08"; }
 void aadd(const Address& addr, const Reg32e &reg) { opMR(addr, reg, T_0F38, 0x0FC, T_APX); }
 void aand(const Address& addr, const Reg32e &reg) { opMR(addr, reg, T_0F38|T_66, 0x0FC, T_APX|T_66); }
 void adc(const Operand& op, uint32_t imm) { opOI(op, imm, 0x10, 2); }
@@ -1937,6 +1937,10 @@ void aesencwide128kl(const Address& addr) { opSSE_APX(xmm0, addr, T_F3|T_0F38, 0
 void aesencwide256kl(const Address& addr) { opSSE_APX(xmm2, addr, T_F3|T_0F38, 0xD8, T_F3|T_MUST_EVEX, 0xD8); }
 void encodekey128(const Reg32& r1, const Reg32& r2) { opEncodeKey(r1, r2, 0xFA, 0xDA); }
 void encodekey256(const Reg32& r1, const Reg32& r2) { opEncodeKey(r1, r2, 0xFB, 0xDB); }
+void rdfsbase(const Reg32e& r) { opRR(eax, r, T_F3|T_0F, 0xAE); }
+void rdgsbase(const Reg32e& r) { opRR(ecx, r, T_F3|T_0F, 0xAE); }
+void wrfsbase(const Reg32e& r) { opRR(edx, r, T_F3|T_0F, 0xAE); }
+void wrgsbase(const Reg32e& r) { opRR(ebx, r, T_F3|T_0F, 0xAE); }
 void ldtilecfg(const Address& addr) { if (opROO(Reg(), addr, tmm0, T_APX|T_0F38|T_W0, 0x49)) return; opVex(tmm0, &tmm0, addr, T_0F38|T_W0, 0x49); }
 void sttilecfg(const Address& addr) { if (opROO(Reg(), addr, tmm0, T_APX|T_66|T_0F38|T_W0, 0x49)) return; opVex(tmm0, &tmm0, addr, T_66|T_0F38 | T_W0, 0x49); }
 void tileloadd(const Tmm& tm, const Address& addr) { opAMX(tm, addr, T_F2|T_0F38|T_W0, 0x4B); }