Add HPPA(PA-RISC) architecture (capstone-engine#2265)

* Refactor HPPA

* Add full HPPA 1.1 instructions support

* Add HPPA 1.1 cs tests

* Fix HPPA dissassembler

* Add HPPA 2.0 instructions

* Add HPPA tests

* Fix HPPA disasm & printer

* Update HPPA tests

* Remove unused code

* Add implicit register access info & Refactor HPPA main files

* Add python bindings/tests and cstests

* Fix HPPA disasm wrong decoding

* Rewrite invalid test cases

* Update HPPA python constants

* Make HPPA python test executable

* Change HPPA python tests sequence to match c tests

* Refactor HPPA main files

* Write target instead of offset in details

* Add HPPA detail function support in cstest

* Rewrite targets in branch tests

* Make correct string modifier addition

* Add hppa test calls

* Add zero operands check

* Remove MCOperand array

* Change immediate values printing

* Add HPPA 2.0 wide support

* Fix invalid break instruction decode

Remove unused code

* Add HPPA to fuzzing tests

* Add HPPA to options

* Add HPPA to docs

* Refactor HPPA

* Fix invalid branch insn decoding

* Add HPPA to labeler

* clang-format hppa files

* Document internal structures and minor refactoring

* Add missing default statements

* Fix invalid default statement

.github/labeler.yml

@@ -57,6 +57,13 @@ EVM:
   - include/capstone/evm.h
   - tests/test_evm.c
 
+HPPA:
+  - arch/HPPA/*
+  - cstool/cstool_hppa.c
+  - include/capstone/hppa.h
+  - suite/MC/HPPA/*
+  - tests/test_hppa.c
+
 M680X:
   - arch/M680X/*
   - cstool/cstool_m680x.c

.gitignore

@@ -75,6 +75,8 @@ tests/test_bpf
 tests/test_sh
 tests/test_riscv
 tests/test_sh
+tests/test_alpha
+tests/test_hppa
 
 # regress binaries
 suite/regress/invalid_read_in_print_operand

CMakeLists.txt

@@ -64,8 +64,8 @@ if(APPLE AND NOT CAPSTONE_BUILD_MACOS_THIN)
     set(CMAKE_OSX_ARCHITECTURES "x86_64;arm64")
 endif()
 
-set(SUPPORTED_ARCHITECTURES ARM AARCH64 M68K MIPS PPC SPARC SYSZ XCORE X86 TMS320C64X M680X EVM MOS65XX WASM BPF RISCV SH TRICORE ALPHA)
-set(SUPPORTED_ARCHITECTURE_LABELS ARM AARCH64 M68K MIPS PowerPC Sparc SystemZ XCore x86 TMS320C64x M680x EVM MOS65XX WASM BPF RISCV SH TriCore Alpha)
+set(SUPPORTED_ARCHITECTURES ARM AARCH64 M68K MIPS PPC SPARC SYSZ XCORE X86 TMS320C64X M680X EVM MOS65XX WASM BPF RISCV SH TRICORE ALPHA HPPA)
+set(SUPPORTED_ARCHITECTURE_LABELS ARM AARCH64 M68K MIPS PowerPC Sparc SystemZ XCore x86 TMS320C64x M680x EVM MOS65XX WASM BPF RISCV SH TriCore Alpha HPPA)
 
 list(LENGTH SUPPORTED_ARCHITECTURES count)
 math(EXPR count "${count}-1")
@@ -169,6 +169,7 @@ set(HEADERS_COMMON
     include/capstone/platform.h
     include/capstone/sh.h
     include/capstone/alpha.h
+    include/capstone/hppa.h
 )
 
 set(TEST_SOURCES test_basic.c test_detail.c test_skipdata.c test_iter.c)
@@ -617,6 +618,24 @@ if (CAPSTONE_ALPHA_SUPPORT)
     set(TEST_SOURCES ${TEST_SOURCES} test_alpha.c)
 endif ()
 
+if(CAPSTONE_HPPA_SUPPORT)
+    add_definitions(-DCAPSTONE_HAS_HPPA)
+    set(SOURCES_HPPA
+        arch/HPPA/HPPADisassembler.c
+        arch/HPPA/HPPAInstPrinter.c
+        arch/HPPA/HPPAMapping.c
+        arch/HPPA/HPPAModule.c
+    )
+    set(HEADERS_HPPA
+        arch/HPPA/HPPAConstants.h
+        arch/HPPA/HPPADisassembler.h
+        arch/HPPA/HPPAInstPrinter.h
+        arch/HPPA/HPPAMapping.h
+        arch/HPPA/HPPAModule.h
+    )
+    set(TEST_SOURCES ${TEST_SOURCES} test_hppa.c)
+endif()
+
 if (CAPSTONE_OSXKERNEL_SUPPORT)
     add_definitions(-DCAPSTONE_HAS_OSXKERNEL)
 endif()
@@ -642,6 +661,7 @@ set(ALL_SOURCES
     ${SOURCES_SH}
     ${SOURCES_TRICORE}
     ${SOURCES_ALPHA}
+    ${SOURCES_HPPA}
 )
 
 set(ALL_HEADERS
@@ -666,6 +686,7 @@ set(ALL_HEADERS
     ${HEADERS_SH}
     ${HEADERS_TRICORE}
     ${HEADERS_ALPHA}
+    ${HEADERS_HPPA}
 )
 
 ## properties
@@ -730,6 +751,7 @@ source_group("Source\\RISCV" FILES ${SOURCES_RISCV})
 source_group("Source\\SH" FILES ${SOURCES_SH})
 source_group("Source\\TriCore" FILES ${SOURCES_TRICORE})
 source_group("Source\\Alpha" FILES ${SOURCES_ALPHA})
+source_group("Source\\HPPA" FILES ${SOURCES_HPPA})
 
 source_group("Include\\Common" FILES ${HEADERS_COMMON})
 source_group("Include\\Engine" FILES ${HEADERS_ENGINE})
@@ -752,6 +774,7 @@ source_group("Include\\RISCV" FILES ${HEADERS_RISCV})
 source_group("Include\\SH" FILES ${HEADERS_SH})
 source_group("Include\\TriCore" FILES ${HEADERS_TRICORE})
 source_group("Include\\Alpha" FILES ${HEADERS_ALPHA})
+source_group("Include\\HPPA" FILES ${HEADERS_HPPA})
 
 ## installation
 if(CAPSTONE_INSTALL)

COMPILE.TXT

@@ -89,6 +89,7 @@ Capstone requires no prerequisite packages, so it is easy to compile & install.
 	/usr/include/capstone/bpf.h
 	/usr/include/capstone/capstone.h
 	/usr/include/capstone/evm.h
+	/usr/include/capstone/hppa.h
 	/usr/include/capstone/m680x.h
 	/usr/include/capstone/m68k.h
 	/usr/include/capstone/mips.h

COMPILE_CMAKE.TXT

@@ -22,6 +22,7 @@ Get CMake for free from http://www.cmake.org.
   - CAPSTONE_ARM_SUPPORT: support ARM. Run cmake with -DCAPSTONE_ARM_SUPPORT=0 to remove ARM.
   - CAPSTONE_AARCH64_SUPPORT: support AARCH64. Run cmake with -DCAPSTONE_AARCH64_SUPPORT=0 to remove AARCH64.
   - CAPSTONE_ALPHA_SUPPORT: support Alpha. Run cmake with -DCAPSTONE_ALPHA_SUPPORT=0 to remove Alpha.
+  - CAPSTONE_HPPA_SUPPORT: support HPPA. Run cmake with -DCAPSTONE_HPPA_SUPPORT=0 to remove HPPA.
   - CAPSTONE_M680X_SUPPORT: support M680X. Run cmake with -DCAPSTONE_M680X_SUPPORT=0 to remove M680X.
   - CAPSTONE_M68K_SUPPORT: support M68K. Run cmake with -DCAPSTONE_M68K_SUPPORT=0 to remove M68K.
   - CAPSTONE_MIPS_SUPPORT: support Mips. Run cmake with -DCAPSTONE_MIPS_SUPPORT=0 to remove Mips.
@@ -117,7 +118,7 @@ Get CMake for free from http://www.cmake.org.
 
   Will just target the x86 architecture. The list of available architectures is: ARM,
   AARCH64, M68K, MIPS, PowerPC, Sparc, SystemZ, XCore, x86, TMS320C64x, M680x, EVM, MOS65XX,
-  WASM, BPF, RISCV, Alpha.
+  WASM, BPF, RISCV, Alpha, HPPA.
 
 (4) You can also create an installation image with cmake, by using the 'install' target.
   Use:

COMPILE_MSVC.TXT

@@ -33,6 +33,7 @@ or newer versions are required.
   - CAPSTONE_HAS_ARM: support ARM. Delete this to remove ARM support.
   - CAPSTONE_HAS_AARCH64: support AARCH64. Delete this to remove AARCH64 support.
   - CAPSTONE_HAS_ALPHA: support Alpha. Delete this to remove Alpha support.
+  - CAPSTONE_HAS_HPPA: support HPPA. Delete this to remove HPPA support.
   - CAPSTONE_HAS_M68K: support M68K. Delete this to remove M68K support.
   - CAPSTONE_HAS_MIPS: support Mips. Delete this to remove Mips support.
   - CAPSTONE_HAS_POWERPC: support PPC. Delete this to remove PPC support.

CREDITS.TXT

@@ -88,4 +88,4 @@ fanfuqiang & citypw & porto703 : RISCV architecture.
 Josh "blacktop" Maine: Arm64 architecture improvements.
 Finn Wilkinson: AArch64 update to Armv9.2-a (SME + SVE2 support)
 Billow & Sidneyp : TriCore architecture.
-Dmitry Sibirtsev: Alpha architecture.
+Dmitry Sibirtsev: Alpha & HPPA architecture.

HACK.TXT

@@ -10,6 +10,7 @@ Capstone source is organized as followings.
 │   ├── ARM         <- ARM engine
 │   ├── BPF         <- Berkeley Packet Filter engine
 │   ├── EVM         <- Ethereum engine
+│   ├── HPPA        <- HPPA engine
 │   ├── M680X       <- M680X engine
 │   ├── M68K        <- M68K engine
 │   ├── Mips        <- Mips engine

MCInst.c

@@ -36,6 +36,7 @@ void MCInst_Init(MCInst *inst)
 		inst->tied_op_idx[i] = -1;
 	inst->isAliasInstr = false;
 	inst->fillDetailOps = false;
+	memset(&inst->hppa_ext, 0, sizeof(inst->hppa_ext));
 }
 
 void MCInst_clear(MCInst *inst)

MCInst.h

@@ -131,6 +131,7 @@ struct MCInst {
 	uint8_t xAcquireRelease;   // X86 xacquire/xrelease
 	bool isAliasInstr; // Flag if this MCInst is an alias.
 	bool fillDetailOps; // If set, detail->operands gets filled.
+	hppa_ext hppa_ext;	///< for HPPA operand. Contains info about modifiers and their effect on the instruction
 };
 
 void MCInst_Init(MCInst *inst);

Makefile

@@ -335,11 +335,22 @@ ifneq (,$(findstring alpha,$(CAPSTONE_ARCHS)))
 	LIBOBJ_ALPHA += $(LIBSRC_ALPHA:%.c=$(OBJDIR)/%.o)
 endif
 
+DEP_HPPA =
+DEP_HPPA += $(wildcard arch/HPPA/HPPA*.inc)
+
+LIBOBJ_HPPA =
+ifneq (,$(findstring hppa,$(CAPSTONE_ARCHS)))
+	CFLAGS += -DCAPSTONE_HAS_HPPA
+	LIBSRC_HPPA += $(wildcard arch/HPPA/HPPA*.c)
+	LIBOBJ_HPPA += $(LIBSRC_HPPA:%.c=$(OBJDIR)/%.o)
+endif
+
+
 LIBOBJ =
 LIBOBJ += $(OBJDIR)/cs.o $(OBJDIR)/utils.o $(OBJDIR)/SStream.o $(OBJDIR)/MCInstrDesc.o $(OBJDIR)/MCRegisterInfo.o $(OBJDIR)/MCInst.o $(OBJDIR)/MCInstPrinter.o $(OBJDIR)/Mapping.o
 LIBOBJ += $(LIBOBJ_ARM) $(LIBOBJ_AARCH64) $(LIBOBJ_M68K) $(LIBOBJ_MIPS) $(LIBOBJ_PPC) $(LIBOBJ_RISCV) $(LIBOBJ_SPARC) $(LIBOBJ_SYSZ) $(LIBOBJ_SH)
 LIBOBJ += $(LIBOBJ_X86) $(LIBOBJ_XCORE) $(LIBOBJ_TMS320C64X) $(LIBOBJ_M680X) $(LIBOBJ_EVM) $(LIBOBJ_MOS65XX) $(LIBOBJ_WASM) $(LIBOBJ_BPF)
-LIBOBJ += $(LIBOBJ_TRICORE) $(LIBOBJ_ALPHA)
+LIBOBJ += $(LIBOBJ_TRICORE) $(LIBOBJ_ALPHA) $(LIBOBJ_HPPA)
 
 
 ifeq ($(PKG_EXTRA),)
@@ -476,6 +487,7 @@ $(LIBOBJ_MOS65XX): $(DEP_MOS65XX)
 $(LIBOBJ_BPF): $(DEP_BPF)
 $(LIBOBJ_TRICORE): $(DEP_TRICORE)
 $(LIBOBJ_ALPHA): $(DEP_ALPHA)
+$(LIBOBJ_HPPA): $(DEP_HPPA)
 
 ifeq ($(CAPSTONE_STATIC),yes)
 $(ARCHIVE): $(LIBOBJ)
@@ -561,13 +573,13 @@ dist:
 	git archive --format=tar.gz --prefix=capstone-$(DIST_VERSION)/ $(TAG) > capstone-$(DIST_VERSION).tgz
 	git archive --format=zip --prefix=capstone-$(DIST_VERSION)/ $(TAG) > capstone-$(DIST_VERSION).zip
 
-TESTS  = test_basic test_detail test_arm test_aarch64 test_m68k test_mips test_ppc test_sparc test_tricore
+TESTS  = test_basic test_detail test_arm test_aarch64 test_m68k test_mips test_ppc test_sparc test_tricore test_hppa
 TESTS += test_systemz test_x86 test_xcore test_iter test_evm test_riscv test_mos65xx test_wasm test_bpf test_alpha
 TESTS += test_basic.static test_detail.static test_arm.static test_aarch64.static
 TESTS += test_m68k.static test_mips.static test_ppc.static test_sparc.static
 TESTS += test_systemz.static test_x86.static test_xcore.static test_m680x.static
 TESTS += test_skipdata test_skipdata.static test_iter.static test_evm.static test_riscv.static
-TESTS += test_mos65xx.static test_wasm.static test_bpf.static test_alpha.static
+TESTS += test_mos65xx.static test_wasm.static test_bpf.static test_alpha.static test_hppa.static
 
 check: $(TESTS)
 

Mapping.c

@@ -334,6 +334,7 @@ DEFINE_get_detail_op(ppc, PPC);
 DEFINE_get_detail_op(tricore, TriCore);
 DEFINE_get_detail_op(aarch64, AArch64);
 DEFINE_get_detail_op(alpha, Alpha);
+DEFINE_get_detail_op(hppa, HPPA);
 
 /// Returns true if for this architecture the
 /// alias operands should be filled.

Mapping.h

@@ -124,6 +124,7 @@ DECL_get_detail_op(ppc, PPC);
 DECL_get_detail_op(tricore, TriCore);
 DECL_get_detail_op(aarch64, AArch64);
 DECL_get_detail_op(alpha, Alpha);
+DECL_get_detail_op(hppa, HPPA);
 
 /// Increments the detail->arch.op_count by one.
 #define DEFINE_inc_detail_op_count(arch, ARCH) \
@@ -149,6 +150,8 @@ DEFINE_inc_detail_op_count(aarch64, AArch64);
 DEFINE_dec_detail_op_count(aarch64, AArch64);
 DEFINE_inc_detail_op_count(alpha, Alpha);
 DEFINE_dec_detail_op_count(alpha, Alpha);
+DEFINE_inc_detail_op_count(hppa, HPPA);
+DEFINE_dec_detail_op_count(hppa, HPPA);
 
 /// Returns true if a memory operand is currently edited.
 static inline bool doing_mem(const MCInst *MI)
@@ -175,6 +178,7 @@ DEFINE_get_arch_detail(ppc, PPC);
 DEFINE_get_arch_detail(tricore, TriCore);
 DEFINE_get_arch_detail(aarch64, AArch64);
 DEFINE_get_arch_detail(alpha, Alpha);
+DEFINE_get_arch_detail(hppa, HPPA);
 
 static inline bool detail_is_set(const MCInst *MI)
 {

MathExtras.h

@@ -430,6 +430,12 @@ static inline int64_t SignExtend64(uint64_t X, unsigned B) {
 	return (int64_t)(X << (64 - B)) >> (64 - B);
 }
 
+/// \brief Removes the rightmost bit of x and extends the field to the left with that
+/// bit to form a 64-bit quantity. The field is of size len
+static inline int64_t LowSignExtend64(uint64_t x, unsigned len) {
+    return (x >> 1) - ((x & 1) << (len - 1));
+}
+
 /// \brief One extend number X starting at bit B and returns it as int32_t.
 /// Requires 0 < B <= 32.
 static inline int32_t OneExtend32(uint32_t X, unsigned B) {
@@ -467,4 +473,16 @@ static inline unsigned int countLeadingZeros(int x)
 	return count;
 }
 
+/// \brief Get specified field from 32-bit instruction. Returns bits from the segment [from, to]
+static inline uint32_t get_insn_field(uint32_t insn, uint8_t from, uint8_t to) 
+{
+	return insn >> (31 - to) & ((1 << (to - from + 1)) - 1);
+}
+
+/// \brief Get specified bit from 32-bit instruction
+static inline uint32_t get_insn_bit(uint32_t insn, uint8_t bit) 
+{
+	return get_insn_field(insn, bit, bit);
+}
+
 #endif

arch/HPPA/HPPAConstants.h

@@ -0,0 +1,74 @@
+/* Capstone Disassembly Engine */
+/* By Dmitry Sibirtsev  <[email protected]>, 2023 */
+
+/* This file defines constants and macros used for parsing a HPPA instruction */
+
+#ifndef CS_HPPA_CONSTANTS_H
+#define CS_HPPA_CONSTANTS_H
+
+#define HPPA_OP_TYPE(byte) (byte) >> 2
+#define MODE_IS_HPPA_20(mode) (((mode)&CS_MODE_HPPA_20) != 0)
+#define MODE_IS_HPPA_20W(mode) (((mode) & (1 << 3)) != 0)
+
+///> HPPA opcode types
+#define HPPA_OP_TYPE_SYSOP 0x00
+#define HPPA_OP_TYPE_MEMMGMT 0x01
+#define HPPA_OP_TYPE_ALU 0x02
+#define HPPA_OP_TYPE_IDXMEM 0x03
+#define HPPA_OP_TYPE_SPOP 0x04
+#define HPPA_OP_TYPE_DIAG 0x05
+#define HPPA_OP_TYPE_FMPYADD 0x06
+#define HPPA_OP_TYPE_LDIL 0x08
+#define HPPA_OP_TYPE_COPRW 0x09
+#define HPPA_OP_TYPE_ADDIL 0x0a
+#define HPPA_OP_TYPE_COPRDW 0x0b
+#define HPPA_OP_TYPE_COPR 0x0c
+#define HPPA_OP_TYPE_LDO 0x0d
+#define HPPA_OP_TYPE_FLOAT 0x0e
+#define HPPA_OP_TYPE_PRDSPEC 0x0f
+#define HPPA_OP_TYPE_LDB 0x10
+#define HPPA_OP_TYPE_LDH 0x11
+#define HPPA_OP_TYPE_LDW 0x12
+#define HPPA_OP_TYPE_LDWM 0x13
+#define HPPA_OP_TYPE_LOADDW 0x14
+#define HPPA_OP_TYPE_FLDW 0x16
+#define HPPA_OP_TYPE_LOADW 0x17
+#define HPPA_OP_TYPE_STB 0x18
+#define HPPA_OP_TYPE_STH 0x19
+#define HPPA_OP_TYPE_STW 0x1a
+#define HPPA_OP_TYPE_STWM 0x1b
+#define HPPA_OP_TYPE_STOREDW 0x1c
+#define HPPA_OP_TYPE_FSTW 0x1e
+#define HPPA_OP_TYPE_STOREW 0x1f
+#define HPPA_OP_TYPE_CMPBT 0x20
+#define HPPA_OP_TYPE_CMPIBT 0x21
+#define HPPA_OP_TYPE_CMPBF 0x22
+#define HPPA_OP_TYPE_CMPIBF 0x23
+#define HPPA_OP_TYPE_CMPICLR 0x24
+#define HPPA_OP_TYPE_SUBI 0x25
+#define HPPA_OP_TYPE_FMPYSUB 0x26
+#define HPPA_OP_TYPE_CMPBDWT 0x27
+#define HPPA_OP_TYPE_ADDBT 0x28
+#define HPPA_OP_TYPE_ADDIBT 0x29
+#define HPPA_OP_TYPE_ADDBF 0x2a
+#define HPPA_OP_TYPE_ADDIBF 0x2b
+#define HPPA_OP_TYPE_ADDIT 0x2c
+#define HPPA_OP_TYPE_ADDI 0x2d
+#define HPPA_OP_TYPE_FPFUSED 0x2e
+#define HPPA_OP_TYPE_CMPBDWF 0x2f
+#define HPPA_OP_TYPE_BBS 0x30
+#define HPPA_OP_TYPE_BB 0x31
+#define HPPA_OP_TYPE_MOVB 0x32
+#define HPPA_OP_TYPE_MOVIB 0x33
+#define HPPA_OP_TYPE_SHEXDEP0 0x34
+#define HPPA_OP_TYPE_SHEXDEP1 0x35
+#define HPPA_OP_TYPE_SHEXDEP2 0x36
+#define HPPA_OP_TYPE_BE 0x38
+#define HPPA_OP_TYPE_BLE 0x39
+#define HPPA_OP_TYPE_BRANCH 0x3a
+#define HPPA_OP_TYPE_CMPIBDW 0x3b
+#define HPPA_OP_TYPE_SHEXDEP3 0x3c
+#define HPPA_OP_TYPE_SHEXDEP4 0x3d
+#define HPPA_OP_TYPE_MULTMED 0x3e
+
+#endif // CS_HPPA_CONSTANTS_H