From 95a7be141a07a20b8465981b01116fcafb6b5f41 Mon Sep 17 00:00:00 2001 From: Rusty Wagner Date: Tue, 3 Oct 2017 23:07:48 -0400 Subject: Adding support for register stacks in IL (for x87) --- python/callingconvention.py | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'python/callingconvention.py') diff --git a/python/callingconvention.py b/python/callingconvention.py index e72475c9..18662fc5 100644 --- a/python/callingconvention.py +++ b/python/callingconvention.py @@ -308,6 +308,10 @@ class CallingConvention(object): return self.name def perform_get_incoming_reg_value(self, reg, func): + reg_stack = self.arch.get_reg_stack_for_reg(reg) + if reg_stack is not None: + if reg == self.arch.reg_stacks[reg_stack].stack_top_reg: + return function.RegisterValue.constant(0) return function.RegisterValue() def perform_get_incoming_flag_value(self, reg, func): -- cgit v1.3.1 From c24b1bd108ee2885d7164cecd15f75aa1715a8df Mon Sep 17 00:00:00 2001 From: Rusty Wagner Date: Mon, 27 Nov 2017 17:15:03 -0500 Subject: Register stack adjustments in calling conventions --- binaryninjaapi.h | 21 +++++- binaryninjacore.h | 29 +++++++- callingconvention.cpp | 44 ++++++++++++ function.cpp | 48 ++++++++++++- lowlevelilinstruction.cpp | 172 ++++++++++++++++++++++++++++++++++++++------ lowlevelilinstruction.h | 57 ++++++++++++--- python/architecture.py | 2 +- python/callingconvention.py | 85 ++++++++++++++++++++++ python/function.py | 67 ++++++++++++++--- python/lowlevelil.py | 24 ++++++- python/types.py | 15 ++++ 11 files changed, 512 insertions(+), 52 deletions(-) (limited to 'python/callingconvention.py') diff --git a/binaryninjaapi.h b/binaryninjaapi.h index 5c122966..af067ae9 100644 --- a/binaryninjaapi.h +++ b/binaryninjaapi.h @@ -2208,6 +2208,7 @@ namespace BinaryNinja Confidence> GetParameterVariables() const; Confidence HasVariableArguments() const; Confidence GetStackAdjustment() const; + std::map> GetRegisterStackAdjustments() const; Confidence> GetClobberedRegisters() const; void SetAutoType(Type* type); @@ -2217,6 +2218,7 @@ namespace BinaryNinja void SetAutoHasVariableArguments(const Confidence& varArgs); void SetAutoCanReturn(const Confidence& returns); void SetAutoStackAdjustment(const Confidence& stackAdjust); + void SetAutoRegisterStackAdjustments(const std::map>& regStackAdjust); void SetAutoClobberedRegisters(const Confidence>& clobbered); void SetUserType(Type* type); @@ -2226,6 +2228,7 @@ namespace BinaryNinja void SetHasVariableArguments(const Confidence& varArgs); void SetCanReturn(const Confidence& returns); void SetStackAdjustment(const Confidence& stackAdjust); + void SetRegisterStackAdjustments(const std::map>& regStackAdjust); void SetClobberedRegisters(const Confidence>& clobbered); void ApplyImportedTypes(Symbol* sym); @@ -2558,11 +2561,12 @@ namespace BinaryNinja ExprId JumpTo(ExprId dest, const std::vector& targets, const ILSourceLocation& loc = ILSourceLocation()); ExprId Call(ExprId dest, const ILSourceLocation& loc = ILSourceLocation()); - ExprId CallStackAdjust(ExprId dest, size_t adjust, const ILSourceLocation& loc = ILSourceLocation()); - ExprId CallSSA(const std::vector& output, ExprId dest, const std::vector& params, + ExprId CallStackAdjust(ExprId dest, size_t adjust, const std::map& regStackAdjust, + const ILSourceLocation& loc = ILSourceLocation()); + ExprId CallSSA(const std::vector& output, ExprId dest, const std::vector& params, const SSARegister& stack, size_t newMemoryVer, size_t prevMemoryVer, const ILSourceLocation& loc = ILSourceLocation()); - ExprId SystemCallSSA(const std::vector& output, const std::vector& params, + ExprId SystemCallSSA(const std::vector& output, const std::vector& params, const SSARegister& stack, size_t newMemoryVer, size_t prevMemoryVer, const ILSourceLocation& loc = ILSourceLocation()); ExprId Return(size_t dest, const ILSourceLocation& loc = ILSourceLocation()); @@ -3184,6 +3188,11 @@ namespace BinaryNinja static void GetIncomingRegisterValueCallback(void* ctxt, uint32_t reg, BNFunction* func, BNRegisterValue* result); static void GetIncomingFlagValueCallback(void* ctxt, uint32_t reg, BNFunction* func, BNRegisterValue* result); + static void GetIncomingVariableForParameterVariableCallback(void* ctxt, const BNVariable* var, + BNFunction* func, BNVariable* result); + static void GetParameterVariableForIncomingVariableCallback(void* ctxt, const BNVariable* var, + BNFunction* func, BNVariable* result); + public: Ref GetArchitecture() const; std::string GetName() const; @@ -3204,6 +3213,9 @@ namespace BinaryNinja virtual std::vector GetImplicitlyDefinedRegisters(); virtual RegisterValue GetIncomingRegisterValue(uint32_t reg, Function* func); virtual RegisterValue GetIncomingFlagValue(uint32_t flag, Function* func); + + virtual Variable GetIncomingVariableForParameterVariable(const Variable& var, Function* func); + virtual Variable GetParameterVariableForIncomingVariable(const Variable& var, Function* func); }; class CoreCallingConvention: public CallingConvention @@ -3227,6 +3239,9 @@ namespace BinaryNinja virtual std::vector GetImplicitlyDefinedRegisters() override; virtual RegisterValue GetIncomingRegisterValue(uint32_t reg, Function* func) override; virtual RegisterValue GetIncomingFlagValue(uint32_t flag, Function* func) override; + + virtual Variable GetIncomingVariableForParameterVariable(const Variable& var, Function* func) override; + virtual Variable GetParameterVariableForIncomingVariable(const Variable& var, Function* func) override; }; /*! diff --git a/binaryninjacore.h b/binaryninjacore.h index c32d42e3..278299bc 100644 --- a/binaryninjacore.h +++ b/binaryninjacore.h @@ -1341,6 +1341,11 @@ extern "C" uint32_t* (*getImplicitlyDefinedRegisters)(void* ctxt, size_t* count); void (*getIncomingRegisterValue)(void* ctxt, uint32_t reg, BNFunction* func, BNRegisterValue* result); void (*getIncomingFlagValue)(void* ctxt, uint32_t flag, BNFunction* func, BNRegisterValue* result); + + void (*getIncomingVariableForParameterVariable)(void* ctxt, const BNVariable* var, + BNFunction* func, BNVariable* result); + void (*getParameterVariableForIncomingVariable)(void* ctxt, const BNVariable* var, + BNFunction* func, BNVariable* result); }; struct BNVariableNameAndType @@ -1658,6 +1663,13 @@ extern "C" ArrayDataType }; + struct BNRegisterStackAdjustment + { + uint32_t regStack; + int32_t adjustment; + uint8_t confidence; + }; + BINARYNINJACOREAPI char* BNAllocString(const char* contents); BINARYNINJACOREAPI void BNFreeString(char* str); BINARYNINJACOREAPI char** BNAllocStringList(const char** contents, size_t size); @@ -2173,6 +2185,8 @@ extern "C" BINARYNINJACOREAPI void BNFreeParameterVariables(BNParameterVariablesWithConfidence* vars); BINARYNINJACOREAPI BNBoolWithConfidence BNFunctionHasVariableArguments(BNFunction* func); BINARYNINJACOREAPI BNSizeWithConfidence BNGetFunctionStackAdjustment(BNFunction* func); + BINARYNINJACOREAPI BNRegisterStackAdjustment* BNGetFunctionRegisterStackAdjustments(BNFunction* func, size_t* count); + BINARYNINJACOREAPI void BNFreeRegisterStackAdjustments(BNRegisterStackAdjustment* adjustments); BINARYNINJACOREAPI BNRegisterSetWithConfidence BNGetFunctionClobberedRegisters(BNFunction* func); BINARYNINJACOREAPI void BNFreeClobberedRegisters(BNRegisterSetWithConfidence* regs); @@ -2182,6 +2196,8 @@ extern "C" BINARYNINJACOREAPI void BNSetAutoFunctionHasVariableArguments(BNFunction* func, BNBoolWithConfidence* varArgs); BINARYNINJACOREAPI void BNSetAutoFunctionCanReturn(BNFunction* func, BNBoolWithConfidence* returns); BINARYNINJACOREAPI void BNSetAutoFunctionStackAdjustment(BNFunction* func, BNSizeWithConfidence* stackAdjust); + BINARYNINJACOREAPI void BNSetAutoFunctionRegisterStackAdjustments(BNFunction* func, + BNRegisterStackAdjustment* adjustments, size_t count); BINARYNINJACOREAPI void BNSetAutoFunctionClobberedRegisters(BNFunction* func, BNRegisterSetWithConfidence* regs); BINARYNINJACOREAPI void BNSetUserFunctionReturnType(BNFunction* func, BNTypeWithConfidence* type); @@ -2190,6 +2206,8 @@ extern "C" BINARYNINJACOREAPI void BNSetUserFunctionHasVariableArguments(BNFunction* func, BNBoolWithConfidence* varArgs); BINARYNINJACOREAPI void BNSetUserFunctionCanReturn(BNFunction* func, BNBoolWithConfidence* returns); BINARYNINJACOREAPI void BNSetUserFunctionStackAdjustment(BNFunction* func, BNSizeWithConfidence* stackAdjust); + BINARYNINJACOREAPI void BNSetUserFunctionRegisterStackAdjustments(BNFunction* func, + BNRegisterStackAdjustment* adjustments, size_t count); BINARYNINJACOREAPI void BNSetUserFunctionClobberedRegisters(BNFunction* func, BNRegisterSetWithConfidence* regs); BINARYNINJACOREAPI void BNApplyImportedTypes(BNFunction* func, BNSymbol* sym); @@ -2241,7 +2259,7 @@ extern "C" BINARYNINJACOREAPI void BNFreeStringReferenceList(BNStringReference* strings); BINARYNINJACOREAPI BNVariableNameAndType* BNGetStackLayout(BNFunction* func, size_t* count); - BINARYNINJACOREAPI void BNFreeVariableList(BNVariableNameAndType* vars, size_t count); + BINARYNINJACOREAPI void BNFreeVariableNameAndTypeList(BNVariableNameAndType* vars, size_t count); BINARYNINJACOREAPI void BNCreateAutoStackVariable(BNFunction* func, int64_t offset, BNTypeWithConfidence* type, const char* name); BINARYNINJACOREAPI void BNCreateUserStackVariable(BNFunction* func, int64_t offset, @@ -2917,6 +2935,15 @@ extern "C" BINARYNINJACOREAPI BNRegisterValue BNGetIncomingRegisterValue(BNCallingConvention* cc, uint32_t reg, BNFunction* func); BINARYNINJACOREAPI BNRegisterValue BNGetIncomingFlagValue(BNCallingConvention* cc, uint32_t reg, BNFunction* func); + BINARYNINJACOREAPI BNVariable BNGetIncomingVariableForParameterVariable(BNCallingConvention* cc, + const BNVariable* var, BNFunction* func); + BINARYNINJACOREAPI BNVariable BNGetParameterVariableForIncomingVariable(BNCallingConvention* cc, + const BNVariable* var, BNFunction* func); + BINARYNINJACOREAPI BNVariable BNGetDefaultIncomingVariableForParameterVariable(BNCallingConvention* cc, + const BNVariable* var); + BINARYNINJACOREAPI BNVariable BNGetDefaultParameterVariableForIncomingVariable(BNCallingConvention* cc, + const BNVariable* var); + BINARYNINJACOREAPI BNCallingConvention* BNGetArchitectureDefaultCallingConvention(BNArchitecture* arch); BINARYNINJACOREAPI BNCallingConvention* BNGetArchitectureCdeclCallingConvention(BNArchitecture* arch); BINARYNINJACOREAPI BNCallingConvention* BNGetArchitectureStdcallCallingConvention(BNArchitecture* arch); diff --git a/callingconvention.cpp b/callingconvention.cpp index b29d51e2..2eb57567 100644 --- a/callingconvention.cpp +++ b/callingconvention.cpp @@ -49,6 +49,8 @@ CallingConvention::CallingConvention(Architecture* arch, const string& name) cc.getImplicitlyDefinedRegisters = GetImplicitlyDefinedRegistersCallback; cc.getIncomingRegisterValue = GetIncomingRegisterValueCallback; cc.getIncomingFlagValue = GetIncomingFlagValueCallback; + cc.getIncomingVariableForParameterVariable = GetIncomingVariableForParameterVariableCallback; + cc.getParameterVariableForIncomingVariable = GetParameterVariableForIncomingVariableCallback; AddRefForRegistration(); m_object = BNCreateCallingConvention(arch->GetObject(), name.c_str(), &cc); @@ -189,6 +191,24 @@ void CallingConvention::GetIncomingFlagValueCallback(void* ctxt, uint32_t reg, B } +void CallingConvention::GetIncomingVariableForParameterVariableCallback(void* ctxt, const BNVariable* var, + BNFunction* func, BNVariable* result) +{ + CallingConvention* cc = (CallingConvention*)ctxt; + *result = cc->GetIncomingVariableForParameterVariable(*var, + func ? new Function(BNNewFunctionReference(func)) : nullptr); +} + + +void CallingConvention::GetParameterVariableForIncomingVariableCallback(void* ctxt, const BNVariable* var, + BNFunction* func, BNVariable* result) +{ + CallingConvention* cc = (CallingConvention*)ctxt; + *result = cc->GetParameterVariableForIncomingVariable(*var, + func ? new Function(BNNewFunctionReference(func)) : nullptr); +} + + Ref CallingConvention::GetArchitecture() const { return new CoreArchitecture(BNGetCallingConventionArchitecture(m_object)); @@ -284,6 +304,18 @@ RegisterValue CallingConvention::GetIncomingFlagValue(uint32_t, Function*) } +Variable CallingConvention::GetIncomingVariableForParameterVariable(const Variable& var, Function*) +{ + return BNGetDefaultIncomingVariableForParameterVariable(m_object, &var); +} + + +Variable CallingConvention::GetParameterVariableForIncomingVariable(const Variable& var, Function*) +{ + return BNGetDefaultParameterVariableForIncomingVariable(m_object, &var); +} + + CoreCallingConvention::CoreCallingConvention(BNCallingConvention* cc): CallingConvention(cc) { } @@ -385,3 +417,15 @@ RegisterValue CoreCallingConvention::GetIncomingFlagValue(uint32_t flag, Functio { return RegisterValue::FromAPIObject(BNGetIncomingFlagValue(m_object, flag, func ? func->GetObject() : nullptr)); } + + +Variable CoreCallingConvention::GetIncomingVariableForParameterVariable(const Variable& var, Function* func) +{ + return BNGetIncomingVariableForParameterVariable(m_object, &var, func ? func->GetObject() : nullptr); +} + + +Variable CoreCallingConvention::GetParameterVariableForIncomingVariable(const Variable& var, Function* func) +{ + return BNGetParameterVariableForIncomingVariable(m_object, &var, func ? func->GetObject() : nullptr); +} diff --git a/function.cpp b/function.cpp index 2d8db04c..879c7156 100644 --- a/function.cpp +++ b/function.cpp @@ -530,6 +530,18 @@ Confidence Function::GetStackAdjustment() const } +map> Function::GetRegisterStackAdjustments() const +{ + size_t count; + BNRegisterStackAdjustment* regStackAdjust = BNGetFunctionRegisterStackAdjustments(m_object, &count); + map> result; + for (size_t i = 0; i < count; i++) + result[regStackAdjust[i].regStack] = Confidence(regStackAdjust[i].adjustment, regStackAdjust[i].confidence); + BNFreeRegisterStackAdjustments(regStackAdjust); + return result; +} + + Confidence> Function::GetClobberedRegisters() const { BNRegisterSetWithConfidence regs = BNGetFunctionClobberedRegisters(m_object); @@ -611,6 +623,22 @@ void Function::SetAutoStackAdjustment(const Confidence& stackAdjust) } +void Function::SetAutoRegisterStackAdjustments(const map>& regStackAdjust) +{ + BNRegisterStackAdjustment* adjust = new BNRegisterStackAdjustment[regStackAdjust.size()]; + size_t i = 0; + for (auto& j : regStackAdjust) + { + adjust[i].regStack = j.first; + adjust[i].adjustment = j.second.GetValue(); + adjust[i].confidence = j.second.GetConfidence(); + i++; + } + BNSetAutoFunctionRegisterStackAdjustments(m_object, adjust, regStackAdjust.size()); + delete[] adjust; +} + + void Function::SetAutoClobberedRegisters(const Confidence>& clobbered) { BNRegisterSetWithConfidence regs; @@ -694,6 +722,22 @@ void Function::SetStackAdjustment(const Confidence& stackAdjust) } +void Function::SetRegisterStackAdjustments(const map>& regStackAdjust) +{ + BNRegisterStackAdjustment* adjust = new BNRegisterStackAdjustment[regStackAdjust.size()]; + size_t i = 0; + for (auto& j : regStackAdjust) + { + adjust[i].regStack = j.first; + adjust[i].adjustment = j.second.GetValue(); + adjust[i].confidence = j.second.GetConfidence(); + i++; + } + BNSetUserFunctionRegisterStackAdjustments(m_object, adjust, regStackAdjust.size()); + delete[] adjust; +} + + void Function::SetClobberedRegisters(const Confidence>& clobbered) { BNRegisterSetWithConfidence regs; @@ -743,7 +787,7 @@ map> Function::GetStackLayout() result[vars[i].var.storage].push_back(var); } - BNFreeVariableList(vars, count); + BNFreeVariableNameAndTypeList(vars, count); return result; } @@ -811,7 +855,7 @@ map Function::GetVariables() result[vars[i].var] = var; } - BNFreeVariableList(vars, count); + BNFreeVariableNameAndTypeList(vars, count); return result; } diff --git a/lowlevelilinstruction.cpp b/lowlevelilinstruction.cpp index 91af36b1..8259426a 100644 --- a/lowlevelilinstruction.cpp +++ b/lowlevelilinstruction.cpp @@ -74,12 +74,13 @@ unordered_map {FlagConditionLowLevelOperandUsage, FlagConditionLowLevelOperand}, {OutputSSARegistersLowLevelOperandUsage, SSARegisterListLowLevelOperand}, {OutputMemoryVersionLowLevelOperandUsage, IndexLowLevelOperand}, - {ParameterSSARegistersLowLevelOperandUsage, SSARegisterListLowLevelOperand}, + {ParameterExprsLowLevelOperandUsage, ExprListLowLevelOperand}, {SourceSSARegistersLowLevelOperandUsage, SSARegisterListLowLevelOperand}, {SourceSSARegisterStacksLowLevelOperandUsage, SSARegisterStackListLowLevelOperand}, {SourceSSAFlagsLowLevelOperandUsage, SSAFlagListLowLevelOperand}, {SourceMemoryVersionsLowLevelOperandUsage, IndexListLowLevelOperand}, - {TargetListLowLevelOperandUsage, IndexListLowLevelOperand} + {TargetListLowLevelOperandUsage, IndexListLowLevelOperand}, + {RegisterStackAdjustmentsLowLevelOperandUsage, RegisterStackAdjustmentsLowLevelOperand} }; @@ -133,7 +134,8 @@ unordered_map> {LLIL_JUMP, {DestExprLowLevelOperandUsage}}, {LLIL_JUMP_TO, {DestExprLowLevelOperandUsage, TargetListLowLevelOperandUsage}}, {LLIL_CALL, {DestExprLowLevelOperandUsage}}, - {LLIL_CALL_STACK_ADJUST, {DestExprLowLevelOperandUsage, StackAdjustmentLowLevelOperandUsage}}, + {LLIL_CALL_STACK_ADJUST, {DestExprLowLevelOperandUsage, StackAdjustmentLowLevelOperandUsage, + RegisterStackAdjustmentsLowLevelOperandUsage}}, {LLIL_RET, {DestExprLowLevelOperandUsage}}, {LLIL_IF, {ConditionExprLowLevelOperandUsage, TrueTargetLowLevelOperandUsage, FalseTargetLowLevelOperandUsage}}, @@ -142,10 +144,10 @@ unordered_map> {LLIL_TRAP, {VectorLowLevelOperandUsage}}, {LLIL_CALL_SSA, {OutputSSARegistersLowLevelOperandUsage, OutputMemoryVersionLowLevelOperandUsage, DestExprLowLevelOperandUsage, StackSSARegisterLowLevelOperandUsage, - StackMemoryVersionLowLevelOperandUsage, ParameterSSARegistersLowLevelOperandUsage}}, + StackMemoryVersionLowLevelOperandUsage, ParameterExprsLowLevelOperandUsage}}, {LLIL_SYSCALL_SSA, {OutputSSARegistersLowLevelOperandUsage, OutputMemoryVersionLowLevelOperandUsage, StackSSARegisterLowLevelOperandUsage, StackMemoryVersionLowLevelOperandUsage, - ParameterSSARegistersLowLevelOperandUsage}}, + ParameterExprsLowLevelOperandUsage}}, {LLIL_REG_PHI, {DestSSARegisterLowLevelOperandUsage, SourceSSARegistersLowLevelOperandUsage}}, {LLIL_REG_STACK_PHI, {DestSSARegisterStackLowLevelOperandUsage, SourceSSARegisterStacksLowLevelOperandUsage}}, {LLIL_FLAG_PHI, {DestSSAFlagLowLevelOperandUsage, SourceSSAFlagsLowLevelOperandUsage}}, @@ -239,7 +241,7 @@ static unordered_map() const } +const LowLevelILInstruction LowLevelILInstructionList::ListIterator::operator*() +{ + return LowLevelILInstruction(pos.GetFunction(), pos.GetFunction()->GetRawExpr((size_t)*pos), + (size_t)*pos, instructionIndex); +} + + +LowLevelILInstructionList::LowLevelILInstructionList(LowLevelILFunction* func, + const BNLowLevelILInstruction& instr, size_t count, size_t instrIndex): + m_list(func, instr, count), m_instructionIndex(instrIndex) +{ +} + + +LowLevelILInstructionList::const_iterator LowLevelILInstructionList::begin() const +{ + const_iterator result; + result.pos = m_list.begin(); + result.instructionIndex = m_instructionIndex; + return result; +} + + +LowLevelILInstructionList::const_iterator LowLevelILInstructionList::end() const +{ + const_iterator result; + result.pos = m_list.end(); + result.instructionIndex = m_instructionIndex; + return result; +} + + +size_t LowLevelILInstructionList::size() const +{ + return m_list.size(); +} + + +const LowLevelILInstruction LowLevelILInstructionList::operator[](size_t i) const +{ + if (i >= size()) + throw LowLevelILInstructionAccessException(); + auto iter = begin(); + for (size_t j = 0; j < i; j++) + ++iter; + return *iter; +} + + +LowLevelILInstructionList::operator vector() const +{ + vector result; + for (auto i : *this) + result.push_back(i); + return result; +} + + const SSARegister LowLevelILSSARegisterList::ListIterator::operator*() { LowLevelILIntegerList::const_iterator cur = pos; @@ -864,14 +925,20 @@ LowLevelILIndexList LowLevelILOperand::GetIndexList() const } +LowLevelILInstructionList LowLevelILOperand::GetExprList() const +{ + if (m_type != ExprListLowLevelOperand) + throw LowLevelILInstructionAccessException(); + return m_instr.GetRawOperandAsExpr(m_operandIndex).GetRawOperandAsExprList(0); +} + + LowLevelILSSARegisterList LowLevelILOperand::GetSSARegisterList() const { if (m_type != SSARegisterListLowLevelOperand) throw LowLevelILInstructionAccessException(); if (m_usage == OutputSSARegistersLowLevelOperandUsage) return m_instr.GetRawOperandAsExpr(m_operandIndex).GetRawOperandAsSSARegisterList(1); - if (m_usage == ParameterSSARegistersLowLevelOperandUsage) - return m_instr.GetRawOperandAsExpr(m_operandIndex).GetRawOperandAsSSARegisterList(0); return m_instr.GetRawOperandAsSSARegisterList(m_operandIndex); } @@ -892,6 +959,14 @@ LowLevelILSSAFlagList LowLevelILOperand::GetSSAFlagList() const } +map LowLevelILOperand::GetRegisterStackAdjustments() const +{ + if (m_type != RegisterStackAdjustmentsLowLevelOperand) + throw LowLevelILInstructionAccessException(); + return m_instr.GetRawOperandAsRegisterStackAdjustments(m_operandIndex); +} + + const LowLevelILOperand LowLevelILOperandList::ListIterator::operator*() { LowLevelILOperandUsage usage = *pos; @@ -1073,6 +1148,13 @@ LowLevelILIndexList LowLevelILInstructionBase::GetRawOperandAsIndexList(size_t o } +LowLevelILInstructionList LowLevelILInstructionBase::GetRawOperandAsExprList(size_t operand) const +{ + return LowLevelILInstructionList(function, function->GetRawExpr(operands[operand + 1]), operands[operand], + instructionIndex); +} + + LowLevelILSSARegisterList LowLevelILInstructionBase::GetRawOperandAsSSARegisterList(size_t operand) const { return LowLevelILSSARegisterList(function, function->GetRawExpr(operands[operand + 1]), operands[operand]); @@ -1091,6 +1173,24 @@ LowLevelILSSAFlagList LowLevelILInstructionBase::GetRawOperandAsSSAFlagList(size } +map LowLevelILInstructionBase::GetRawOperandAsRegisterStackAdjustments(size_t operand) const +{ + LowLevelILIntegerList list(function, function->GetRawExpr(operands[operand + 1]), operands[operand]); + map result; + for (auto i = list.begin(); i != list.end(); ) + { + uint32_t regStack = (uint32_t)*i; + ++i; + if (i == list.end()) + break; + int32_t adjust = (int32_t)*i; + ++i; + result[regStack] = adjust; + } + return result; +} + + void LowLevelILInstructionBase::UpdateRawOperand(size_t operandIndex, ExprId value) { operands[operandIndex] = value; @@ -1385,6 +1485,12 @@ void LowLevelILInstruction::VisitExprs(const std::function().VisitExprs(func); + for (auto& i : GetParameterExprs()) + i.VisitExprs(func); + break; + case LLIL_SYSCALL_SSA: + for (auto& i : GetParameterExprs()) + i.VisitExprs(func); break; case LLIL_RET: GetDestExpr().VisitExprs(func); @@ -1478,6 +1584,7 @@ ExprId LowLevelILInstruction::CopyTo(LowLevelILFunction* dest, const std::function& subExprHandler) const { vector labelList; + vector params; BNLowLevelILLabel* labelA; BNLowLevelILLabel* labelB; switch (operation) @@ -1576,7 +1683,7 @@ ExprId LowLevelILInstruction::CopyTo(LowLevelILFunction* dest, return dest->Call(subExprHandler(GetDestExpr()), *this); case LLIL_CALL_STACK_ADJUST: return dest->CallStackAdjust(subExprHandler(GetDestExpr()), - GetStackAdjustment(), *this); + GetStackAdjustment(), GetRegisterStackAdjustments(), *this); case LLIL_RET: return dest->Return(subExprHandler(GetDestExpr()), *this); case LLIL_JUMP_TO: @@ -1607,13 +1714,17 @@ ExprId LowLevelILInstruction::CopyTo(LowLevelILFunction* dest, case LLIL_TRAP: return dest->Trap(GetVector(), *this); case LLIL_CALL_SSA: + for (auto& i : GetParameterExprs()) + params.push_back(subExprHandler(i)); return dest->CallSSA(GetOutputSSARegisters(), subExprHandler(GetDestExpr()), - GetParameterSSARegisters(), GetStackSSARegister(), - GetDestMemoryVersion(), GetSourceMemoryVersion(), *this); + params, GetStackSSARegister(), GetDestMemoryVersion(), + GetSourceMemoryVersion(), *this); case LLIL_SYSCALL_SSA: + for (auto& i : GetParameterExprs()) + params.push_back(subExprHandler(i)); return dest->SystemCallSSA(GetOutputSSARegisters(), - GetParameterSSARegisters(), GetStackSSARegister(), - GetDestMemoryVersion(), GetSourceMemoryVersion(), *this); + params, GetStackSSARegister(), GetDestMemoryVersion(), + GetSourceMemoryVersion(), *this); case LLIL_REG_PHI: return dest->RegisterPhi(GetDestSSARegister(), GetSourceSSARegisters(), *this); case LLIL_REG_STACK_PHI: @@ -2055,11 +2166,11 @@ LowLevelILSSARegisterList LowLevelILInstruction::GetOutputSSARegisters() const } -LowLevelILSSARegisterList LowLevelILInstruction::GetParameterSSARegisters() const +LowLevelILInstructionList LowLevelILInstruction::GetParameterExprs() const { size_t operandIndex; - if (GetOperandIndexForUsage(ParameterSSARegistersLowLevelOperandUsage, operandIndex)) - return GetRawOperandAsExpr(operandIndex).GetRawOperandAsSSARegisterList(0); + if (GetOperandIndexForUsage(ParameterExprsLowLevelOperandUsage, operandIndex)) + return GetRawOperandAsExpr(operandIndex).GetRawOperandAsExprList(0); throw LowLevelILInstructionAccessException(); } @@ -2109,6 +2220,15 @@ LowLevelILIndexList LowLevelILInstruction::GetTargetList() const } +map LowLevelILInstruction::GetRegisterStackAdjustments() const +{ + size_t operandIndex; + if (GetOperandIndexForUsage(RegisterStackAdjustmentsLowLevelOperandUsage, operandIndex)) + return GetRawOperandAsRegisterStackAdjustments(operandIndex); + throw LowLevelILInstructionAccessException(); +} + + ExprId LowLevelILFunction::Nop(const ILSourceLocation& loc) { return AddExprWithLocation(LLIL_NOP, loc, 0, 0); @@ -2578,13 +2698,21 @@ ExprId LowLevelILFunction::Call(ExprId dest, const ILSourceLocation& loc) } -ExprId LowLevelILFunction::CallStackAdjust(ExprId dest, size_t adjust, const ILSourceLocation& loc) +ExprId LowLevelILFunction::CallStackAdjust(ExprId dest, size_t adjust, + const std::map& regStackAdjust, const ILSourceLocation& loc) { - return AddExprWithLocation(LLIL_CALL_STACK_ADJUST, loc, 0, 0, dest, adjust); + vector list; + for (auto& i : regStackAdjust) + { + list.push_back(i.first); + list.push_back(i.second); + } + return AddExprWithLocation(LLIL_CALL_STACK_ADJUST, loc, 0, 0, dest, adjust, list.size(), + AddIndexList(list)); } -ExprId LowLevelILFunction::CallSSA(const vector& output, ExprId dest, const vector& params, +ExprId LowLevelILFunction::CallSSA(const vector& output, ExprId dest, const vector& params, const SSARegister& stack, size_t newMemoryVer, size_t prevMemoryVer, const ILSourceLocation& loc) { return AddExprWithLocation(LLIL_CALL_SSA, loc, 0, 0, @@ -2592,11 +2720,11 @@ ExprId LowLevelILFunction::CallSSA(const vector& output, ExprId des output.size() * 2, AddSSARegisterList(output)), dest, AddExprWithLocation(LLIL_CALL_STACK_SSA, loc, 0, 0, stack.reg, stack.version, prevMemoryVer), AddExprWithLocation(LLIL_CALL_PARAM_SSA, loc, 0, 0, - params.size() * 2, AddSSARegisterList(params))); + params.size(), AddOperandList(params))); } -ExprId LowLevelILFunction::SystemCallSSA(const vector& output, const vector& params, +ExprId LowLevelILFunction::SystemCallSSA(const vector& output, const vector& params, const SSARegister& stack, size_t newMemoryVer, size_t prevMemoryVer, const ILSourceLocation& loc) { return AddExprWithLocation(LLIL_SYSCALL_SSA, loc, 0, 0, @@ -2604,7 +2732,7 @@ ExprId LowLevelILFunction::SystemCallSSA(const vector& output, cons output.size() * 2, AddSSARegisterList(output)), AddExprWithLocation(LLIL_CALL_STACK_SSA, loc, 0, 0, stack.reg, stack.version, prevMemoryVer), AddExprWithLocation(LLIL_CALL_PARAM_SSA, loc, 0, 0, - params.size() * 2, AddSSARegisterList(params))); + params.size(), AddOperandList(params))); } diff --git a/lowlevelilinstruction.h b/lowlevelilinstruction.h index 4e02632e..cfc7e2d7 100644 --- a/lowlevelilinstruction.h +++ b/lowlevelilinstruction.h @@ -112,9 +112,11 @@ namespace BinaryNinja SSARegisterStackLowLevelOperand, SSAFlagLowLevelOperand, IndexListLowLevelOperand, + ExprListLowLevelOperand, SSARegisterListLowLevelOperand, SSARegisterStackListLowLevelOperand, - SSAFlagListLowLevelOperand + SSAFlagListLowLevelOperand, + RegisterStackAdjustmentsLowLevelOperand }; enum LowLevelILOperandUsage @@ -158,12 +160,13 @@ namespace BinaryNinja FlagConditionLowLevelOperandUsage, OutputSSARegistersLowLevelOperandUsage, OutputMemoryVersionLowLevelOperandUsage, - ParameterSSARegistersLowLevelOperandUsage, + ParameterExprsLowLevelOperandUsage, SourceSSARegistersLowLevelOperandUsage, SourceSSARegisterStacksLowLevelOperandUsage, SourceSSAFlagsLowLevelOperandUsage, SourceMemoryVersionsLowLevelOperandUsage, - TargetListLowLevelOperandUsage + TargetListLowLevelOperandUsage, + RegisterStackAdjustmentsLowLevelOperandUsage }; } @@ -328,6 +331,36 @@ namespace BinaryNinja operator std::vector() const; }; + class LowLevelILInstructionList + { + struct ListIterator + { + LowLevelILIntegerList::const_iterator pos; + size_t instructionIndex; + bool operator==(const ListIterator& a) const { return pos == a.pos; } + bool operator!=(const ListIterator& a) const { return pos != a.pos; } + bool operator<(const ListIterator& a) const { return pos < a.pos; } + ListIterator& operator++() { ++pos; return *this; } + const LowLevelILInstruction operator*(); + }; + + LowLevelILIntegerList m_list; + size_t m_instructionIndex; + + public: + typedef ListIterator const_iterator; + + LowLevelILInstructionList(LowLevelILFunction* func, const BNLowLevelILInstruction& instr, + size_t count, size_t instrIndex); + + const_iterator begin() const; + const_iterator end() const; + size_t size() const; + const LowLevelILInstruction operator[](size_t i) const; + + operator std::vector() const; + }; + class LowLevelILSSARegisterList { struct ListIterator @@ -436,9 +469,11 @@ namespace BinaryNinja SSARegisterStack GetRawOperandAsPartialSSARegisterStackSource(size_t operand) const; SSAFlag GetRawOperandAsSSAFlag(size_t operand) const; LowLevelILIndexList GetRawOperandAsIndexList(size_t operand) const; + LowLevelILInstructionList GetRawOperandAsExprList(size_t operand) const; LowLevelILSSARegisterList GetRawOperandAsSSARegisterList(size_t operand) const; LowLevelILSSARegisterStackList GetRawOperandAsSSARegisterStackList(size_t operand) const; LowLevelILSSAFlagList GetRawOperandAsSSAFlagList(size_t operand) const; + std::map GetRawOperandAsRegisterStackAdjustments(size_t operand) const; void UpdateRawOperand(size_t operandIndex, ExprId value); void UpdateRawOperandAsSSARegisterList(size_t operandIndex, const std::vector& regs); @@ -574,12 +609,13 @@ namespace BinaryNinja template size_t GetDestMemoryVersion() const { return As().GetDestMemoryVersion(); } template BNLowLevelILFlagCondition GetFlagCondition() const { return As().GetFlagCondition(); } template LowLevelILSSARegisterList GetOutputSSARegisters() const { return As().GetOutputSSARegisters(); } - template LowLevelILSSARegisterList GetParameterSSARegisters() const { return As().GetParameterSSARegisters(); } + template LowLevelILInstructionList GetParameterExprs() const { return As().GetParameterExprs(); } template LowLevelILSSARegisterList GetSourceSSARegisters() const { return As().GetSourceSSARegisters(); } template LowLevelILSSARegisterStackList GetSourceSSARegisterStacks() const { return As().GetSourceSSARegisterStacks(); } template LowLevelILSSAFlagList GetSourceSSAFlags() const { return As().GetSourceSSAFlags(); } template LowLevelILIndexList GetSourceMemoryVersions() const { return As().GetSourceMemoryVersions(); } template LowLevelILIndexList GetTargetList() const { return As().GetTargetList(); } + template std::map GetRegisterStackAdjustments() const { return As().GetRegisterStackAdjustments(); } template void SetDestSSAVersion(size_t version) { As().SetDestSSAVersion(version); } template void SetSourceSSAVersion(size_t version) { As().SetSourceSSAVersion(version); } @@ -590,7 +626,6 @@ namespace BinaryNinja template void SetDestMemoryVersion(size_t version) { As().SetDestMemoryVersion(version); } template void SetSourceMemoryVersion(size_t version) { As().SetSourceMemoryVersion(version); } template void SetOutputSSARegisters(const std::vector& regs) { As().SetOutputSSARegisters(regs); } - template void SetParameterSSARegisters(const std::vector& regs) { As().SetParameterSSARegisters(regs); } bool GetOperandIndexForUsage(LowLevelILOperandUsage usage, size_t& operandIndex) const; @@ -632,12 +667,13 @@ namespace BinaryNinja size_t GetDestMemoryVersion() const; BNLowLevelILFlagCondition GetFlagCondition() const; LowLevelILSSARegisterList GetOutputSSARegisters() const; - LowLevelILSSARegisterList GetParameterSSARegisters() const; + LowLevelILInstructionList GetParameterExprs() const; LowLevelILSSARegisterList GetSourceSSARegisters() const; LowLevelILSSARegisterStackList GetSourceSSARegisterStacks() const; LowLevelILSSAFlagList GetSourceSSAFlags() const; LowLevelILIndexList GetSourceMemoryVersions() const; LowLevelILIndexList GetTargetList() const; + std::map GetRegisterStackAdjustments() const; }; class LowLevelILOperand @@ -665,9 +701,11 @@ namespace BinaryNinja SSARegisterStack GetSSARegisterStack() const; SSAFlag GetSSAFlag() const; LowLevelILIndexList GetIndexList() const; + LowLevelILInstructionList GetExprList() const; LowLevelILSSARegisterList GetSSARegisterList() const; LowLevelILSSARegisterStackList GetSSARegisterStackList() const; LowLevelILSSAFlagList GetSSAFlagList() const; + std::map GetRegisterStackAdjustments() const; }; class LowLevelILOperandList @@ -915,6 +953,7 @@ namespace BinaryNinja { LowLevelILInstruction GetDestExpr() const { return GetRawOperandAsExpr(0); } size_t GetStackAdjustment() const { return (size_t)GetRawOperandAsInteger(1); } + std::map GetRegisterStackAdjustments() const { return GetRawOperandAsRegisterStackAdjustments(2); } }; template <> struct LowLevelILInstructionAccessor: public LowLevelILInstructionBase { @@ -949,12 +988,11 @@ namespace BinaryNinja LowLevelILInstruction GetDestExpr() const { return GetRawOperandAsExpr(1); } SSARegister GetStackSSARegister() const { return GetRawOperandAsExpr(2).GetRawOperandAsSSARegister(0); } size_t GetSourceMemoryVersion() const { return GetRawOperandAsExpr(2).GetRawOperandAsIndex(2); } - LowLevelILSSARegisterList GetParameterSSARegisters() const { return GetRawOperandAsExpr(3).GetRawOperandAsSSARegisterList(0); } + LowLevelILInstructionList GetParameterExprs() const { return GetRawOperandAsExpr(3).GetRawOperandAsExprList(0); } void SetDestMemoryVersion(size_t version) { GetRawOperandAsExpr(0).UpdateRawOperand(0, version); } void SetSourceMemoryVersion(size_t version) { GetRawOperandAsExpr(2).UpdateRawOperand(2, version); } void SetStackSSAVersion(size_t version) { GetRawOperandAsExpr(2).UpdateRawOperand(1, version); } void SetOutputSSARegisters(const std::vector& regs) { GetRawOperandAsExpr(0).UpdateRawOperandAsSSARegisterList(1, regs); } - void SetParameterSSARegisters(const std::vector& regs) { GetRawOperandAsExpr(3).UpdateRawOperandAsSSARegisterList(0, regs); } }; template <> struct LowLevelILInstructionAccessor: public LowLevelILInstructionBase { @@ -962,12 +1000,11 @@ namespace BinaryNinja size_t GetDestMemoryVersion() const { return GetRawOperandAsExpr(0).GetRawOperandAsIndex(0); } SSARegister GetStackSSARegister() const { return GetRawOperandAsExpr(1).GetRawOperandAsSSARegister(0); } size_t GetSourceMemoryVersion() const { return GetRawOperandAsExpr(1).GetRawOperandAsIndex(2); } - LowLevelILSSARegisterList GetParameterSSARegisters() const { return GetRawOperandAsExpr(2).GetRawOperandAsSSARegisterList(0); } + LowLevelILInstructionList GetParameterExprs() const { return GetRawOperandAsExpr(2).GetRawOperandAsExprList(0); } void SetDestMemoryVersion(size_t version) { GetRawOperandAsExpr(0).UpdateRawOperand(0, version); } void SetSourceMemoryVersion(size_t version) { GetRawOperandAsExpr(1).UpdateRawOperand(2, version); } void SetStackSSAVersion(size_t version) { GetRawOperandAsExpr(1).UpdateRawOperand(1, version); } void SetOutputSSARegisters(const std::vector& regs) { GetRawOperandAsExpr(0).UpdateRawOperandAsSSARegisterList(1, regs); } - void SetParameterSSARegisters(const std::vector& regs) { GetRawOperandAsExpr(2).UpdateRawOperandAsSSARegisterList(0, regs); } }; template <> struct LowLevelILInstructionAccessor: public LowLevelILInstructionBase diff --git a/python/architecture.py b/python/architecture.py index 04179e7f..c4179b00 100644 --- a/python/architecture.py +++ b/python/architecture.py @@ -231,7 +231,7 @@ class Architecture(object): for j in xrange(0, info.topRelativeCount): top_rel.append(core.BNGetArchitectureRegisterName(self.handle, info.firstTopRelativeReg + j)) top = core.BNGetArchitectureRegisterName(self.handle, info.stackTopReg) - self.reg_stacks[name] = function.RegisterStackInfo(storage, top_rel, top) + self.reg_stacks[name] = function.RegisterStackInfo(storage, top_rel, top, regs[i]) core.BNFreeRegisterList(regs) else: startup._init_plugins() diff --git a/python/callingconvention.py b/python/callingconvention.py index 18662fc5..5aad3317 100644 --- a/python/callingconvention.py +++ b/python/callingconvention.py @@ -28,6 +28,7 @@ import log import types import function import binaryview +from enums import VariableSourceType class CallingConvention(object): @@ -68,6 +69,8 @@ class CallingConvention(object): self._cb.getImplicitlyDefinedRegisters = self._cb.getImplicitlyDefinedRegisters.__class__(self._get_implicitly_defined_regs) self._cb.getIncomingRegisterValue = self._cb.getIncomingRegisterValue.__class__(self._get_incoming_reg_value) self._cb.getIncomingFlagValue = self._cb.getIncomingFlagValue.__class__(self._get_incoming_flag_value) + self._cb.getIncomingVariableForParameterVariable = self._cb.getIncomingVariableForParameterVariable.__class__(self._get_incoming_var_for_parameter_var) + self._cb.getParameterVariableForIncomingVariable = self._cb.getParameterVariableForIncomingVariable.__class__(self._get_parameter_var_for_incoming_var) self.handle = core.BNCreateCallingConvention(arch.handle, name, self._cb) self.__class__._registered_calling_conventions.append(self) else: @@ -301,6 +304,42 @@ class CallingConvention(object): result[0].state = api_obj.state result[0].value = api_obj.value + def _get_incoming_var_for_parameter_var(self, ctxt, in_var, func, result): + try: + if func is None: + func_obj = None + else: + func_obj = function.Function(binaryview.BinaryView(handle = core.BNGetFunctionData(func)), + core.BNNewFunctionReference(func)) + in_var_obj = function.Variable(func_obj, in_var[0].type, in_var[0].index, in_var[0].storage) + out_var = self.perform_get_incoming_var_for_parameter_var(in_var_obj, func_obj) + result[0].type = out_var.source_type + result[0].index = out_var.index + result[0].storage = out_var.storage + except: + log.log_error(traceback.format_exc()) + result[0].type = in_var[0].type + result[0].index = in_var[0].index + result[0].storage = in_var[0].storage + + def _get_parameter_var_for_incoming_var(self, ctxt, in_var, func, result): + try: + if func is None: + func_obj = None + else: + func_obj = function.Function(binaryview.BinaryView(handle = core.BNGetFunctionData(func)), + core.BNNewFunctionReference(func)) + in_var_obj = function.Variable(func_obj, in_var[0].type, in_var[0].index, in_var[0].storage) + out_var = self.perform_get_parameter_var_for_incoming_var(in_var_obj, func_obj) + result[0].type = out_var.source_type + result[0].index = out_var.index + result[0].storage = out_var.storage + except: + log.log_error(traceback.format_exc()) + result[0].type = in_var[0].type + result[0].index = in_var[0].index + result[0].storage = in_var[0].storage + def __repr__(self): return "" % (self.arch.name, self.name) @@ -317,6 +356,25 @@ class CallingConvention(object): def perform_get_incoming_flag_value(self, reg, func): return function.RegisterValue() + def perform_get_incoming_var_for_parameter_var(self, in_var, func): + in_buf = core.BNVariable() + in_buf.type = in_var.source_type + in_buf.index = in_var.index + in_buf.storage = in_var.storage + out_var = core.BNGetDefaultIncomingVariableForParameterVariable(self.handle, in_buf) + name = None + if (func is not None) and (out_var.type == VariableSourceType.RegisterVariableSourceType): + name = func.arch.get_reg_name(out_var.storage) + return function.Variable(func, out_var.type, out_var.index, out_var.storage, name) + + def perform_get_parameter_var_for_incoming_var(self, in_var, func): + in_buf = core.BNVariable() + in_buf.type = in_var.source_type + in_buf.index = in_var.index + in_buf.storage = in_var.storage + out_var = core.BNGetDefaultParameterVariableForIncomingVariable(self.handle, in_buf) + return function.Variable(func, out_var.type, out_var.index, out_var.storage) + def with_confidence(self, confidence): return CallingConvention(self.arch, handle = core.BNNewCallingConventionReference(self.handle), confidence = confidence) @@ -334,3 +392,30 @@ class CallingConvention(object): if func is not None: func_handle = func.handle return function.RegisterValue(self.arch, core.BNGetIncomingFlagValue(self.handle, reg_num, func_handle)) + + def get_incoming_var_for_parameter_var(self, in_var, func): + in_buf = core.BNVariable() + in_buf.type = in_var.source_type + in_buf.index = in_var.index + in_buf.storage = in_var.storage + if func is None: + func_obj = None + else: + func_obj = func.handle + out_var = core.BNGetIncomingVariableForParameterVariable(self.handle, in_buf, func_obj) + name = None + if (func is not None) and (out_var.type == VariableSourceType.RegisterVariableSourceType): + name = func.arch.get_reg_name(out_var.storage) + return function.Variable(func, out_var.type, out_var.index, out_var.storage, name) + + def get_parameter_var_for_incoming_var(self, in_var, func): + in_buf = core.BNVariable() + in_buf.type = in_var.source_type + in_buf.index = in_var.index + in_buf.storage = in_var.storage + if func is None: + func_obj = None + else: + func_obj = func.handle + out_var = core.BNGetParameterVariableForIncomingVariable(self.handle, in_buf, func_obj) + return function.Variable(func, out_var.type, out_var.index, out_var.storage) diff --git a/python/function.py b/python/function.py index ba00147e..eabf86b3 100644 --- a/python/function.py +++ b/python/function.py @@ -263,14 +263,15 @@ class Variable(object): var.storage = storage self.identifier = core.BNToVariableIdentifier(var) - if name is None: - name = core.BNGetVariableName(func.handle, var) - if var_type is None: - var_type_conf = core.BNGetVariableType(func.handle, var) - if var_type_conf.type: - var_type = types.Type(var_type_conf.type, platform = func.platform, confidence = var_type_conf.confidence) - else: - var_type = None + if func is not None: + if name is None: + name = core.BNGetVariableName(func.handle, var) + if var_type is None: + var_type_conf = core.BNGetVariableType(func.handle, var) + if var_type_conf.type: + var_type = types.Type(var_type_conf.type, platform = func.platform, confidence = var_type_conf.confidence) + else: + var_type = None self.name = name self.type = var_type @@ -519,7 +520,7 @@ class Function(object): result.append(Variable(self, v[i].var.type, v[i].var.index, v[i].var.storage, v[i].name, types.Type(handle = core.BNNewTypeReference(v[i].type), platform = self.platform, confidence = v[i].typeConfidence))) result.sort(key = lambda x: x.identifier) - core.BNFreeVariableList(v, count.value) + core.BNFreeVariableNameAndTypeList(v, count.value) return result @property @@ -532,7 +533,7 @@ class Function(object): result.append(Variable(self, v[i].var.type, v[i].var.index, v[i].var.storage, v[i].name, types.Type(handle = core.BNNewTypeReference(v[i].type), platform = self.platform, confidence = v[i].typeConfidence))) result.sort(key = lambda x: x.identifier) - core.BNFreeVariableList(v, count.value) + core.BNFreeVariableNameAndTypeList(v, count.value) return result @property @@ -674,6 +675,35 @@ class Function(object): sc.confidence = types.max_confidence core.BNSetUserFunctionStackAdjustment(self.handle, sc) + @property + def reg_stack_adjustments(self): + """Number of entries removed from each register stack after return""" + count = ctypes.c_ulonglong() + adjust = core.BNGetFunctionRegisterStackAdjustments(self.handle, count) + result = {} + for i in xrange(0, count.value): + name = self.arch.get_reg_stack_name(adjust[i].regStack) + value = types.RegisterStackAdjustmentWithConfidence(adjust[i].adjustment, + confidence = adjust[i].confidence) + result[name] = value + core.BNFreeRegisterStackAdjustments(adjust) + return result + + @reg_stack_adjustments.setter + def reg_stack_adjustments(self, value): + adjust = (core.BNRegisterStackAdjustment * len(value))() + i = 0 + for reg_stack in value.keys(): + adjust[i].regStack = self.arch.get_reg_stack_index(reg_stack) + if isinstance(value[reg_stack], types.RegisterStackAdjustmentWithConfidence): + adjust[i].adjustment = value[reg_stack].value + adjust[i].confidence = value[reg_stack].confidence + else: + adjust[i].adjustment = value[reg_stack] + adjust[i].confidence = types.max_confidence + i += 1 + core.BNSetUserFunctionRegisterStackAdjustments(self.handle, adjust, len(value)) + @property def clobbered_regs(self): """Registers that are modified by this function""" @@ -1099,6 +1129,20 @@ class Function(object): sc.confidence = types.max_confidence core.BNSetAutoFunctionStackAdjustment(self.handle, sc) + def set_auto_reg_stack_adjustments(self, value): + adjust = (core.BNRegisterStackAdjustment * len(value))() + i = 0 + for reg_stack in value.keys(): + adjust[i].regStack = self.arch.get_reg_stack_index(reg_stack) + if isinstance(value[reg_stack], types.RegisterStackAdjustmentWithConfidence): + adjust[i].adjustment = value[reg_stack].value + adjust[i].confidence = value[reg_stack].confidence + else: + adjust[i].adjustment = value[reg_stack] + adjust[i].confidence = types.max_confidence + i += 1 + core.BNSetAutoFunctionRegisterStackAdjustments(self.handle, adjust, len(value)) + def set_auto_clobbered_regs(self, value): regs = core.BNRegisterSetWithConfidence() regs.regs = (ctypes.c_uint * len(value))() @@ -1727,10 +1771,11 @@ class RegisterInfo(object): class RegisterStackInfo(object): - def __init__(self, storage_regs, top_relative_regs, stack_top_reg): + def __init__(self, storage_regs, top_relative_regs, stack_top_reg, index=None): self.storage_regs = storage_regs self.top_relative_regs = top_relative_regs self.stack_top_reg = stack_top_reg + self.index = index def __repr__(self): return "" % (len(self.storage_regs), self.stack_top_reg) diff --git a/python/lowlevelil.py b/python/lowlevelil.py index f5a39b7e..3810e742 100644 --- a/python/lowlevelil.py +++ b/python/lowlevelil.py @@ -197,7 +197,7 @@ class LowLevelILInstruction(object): LowLevelILOperation.LLIL_JUMP: [("dest", "expr")], LowLevelILOperation.LLIL_JUMP_TO: [("dest", "expr"), ("targets", "int_list")], LowLevelILOperation.LLIL_CALL: [("dest", "expr")], - LowLevelILOperation.LLIL_CALL_STACK_ADJUST: [("dest", "expr"), ("stack_adjustment", "int")], + LowLevelILOperation.LLIL_CALL_STACK_ADJUST: [("dest", "expr"), ("stack_adjustment", "int"), ("reg_stack_adjustments", "reg_stack_adjust")], LowLevelILOperation.LLIL_RET: [("dest", "expr")], LowLevelILOperation.LLIL_NORET: [], LowLevelILOperation.LLIL_IF: [("condition", "expr"), ("true", "int"), ("false", "int")], @@ -258,7 +258,7 @@ class LowLevelILInstruction(object): LowLevelILOperation.LLIL_SYSCALL_SSA: [("output", "expr"), ("stack", "expr"), ("param", "expr")], LowLevelILOperation.LLIL_CALL_OUTPUT_SSA: [("dest_memory", "int"), ("dest", "reg_ssa_list")], LowLevelILOperation.LLIL_CALL_STACK_SSA: [("src", "reg_ssa"), ("src_memory", "int")], - LowLevelILOperation.LLIL_CALL_PARAM_SSA: [("src", "reg_ssa_list")], + LowLevelILOperation.LLIL_CALL_PARAM_SSA: [("src", "expr_list")], LowLevelILOperation.LLIL_LOAD_SSA: [("src", "expr"), ("src_memory", "int")], LowLevelILOperation.LLIL_STORE_SSA: [("dest", "expr"), ("dest_memory", "int"), ("src_memory", "int"), ("src", "expr")], LowLevelILOperation.LLIL_REG_PHI: [("dest", "reg_ssa"), ("src", "reg_ssa_list")], @@ -334,6 +334,14 @@ class LowLevelILInstruction(object): for i in xrange(count.value): value.append(operand_list[i]) core.BNLowLevelILFreeOperandList(operand_list) + elif operand_type == "expr_list": + count = ctypes.c_ulonglong() + operand_list = core.BNLowLevelILGetOperandList(func.handle, self.expr_index, i, count) + i += 1 + value = [] + for i in xrange(count.value): + value.append(LowLevelILInstruction(func, operand_list[i])) + core.BNLowLevelILFreeOperandList(operand_list) elif operand_type == "reg_ssa_list": count = ctypes.c_ulonglong() operand_list = core.BNLowLevelILGetOperandList(func.handle, self.expr_index, i, count) @@ -364,6 +372,18 @@ class LowLevelILInstruction(object): flag_version = operand_list[(i * 2) + 1] value.append(SSAFlag(ILFlag(func.arch, flag), flag_version)) core.BNLowLevelILFreeOperandList(operand_list) + elif operand_type == "reg_stack_adjust": + count = ctypes.c_ulonglong() + operand_list = core.BNLowLevelILGetOperandList(func.handle, self.expr_index, i, count) + i += 1 + value = {} + for i in xrange(count.value / 2): + reg_stack = operand_list[i * 2] + adjust = operand_list[(i * 2) + 1] + if adjust & 0x80000000: + adjust |= ~0x80000000 + value[func.arch.get_reg_stack_name(reg_stack)] = adjust + core.BNLowLevelILFreeOperandList(operand_list) self.operands.append(value) self.__dict__[name] = value i += 1 diff --git a/python/types.py b/python/types.py index 2557db2c..d44cc736 100644 --- a/python/types.py +++ b/python/types.py @@ -680,6 +680,21 @@ class SizeWithConfidence(object): return self.value +class RegisterStackAdjustmentWithConfidence(object): + def __init__(self, value, confidence = max_confidence): + self.value = value + self.confidence = confidence + + def __str__(self): + return str(self.value) + + def __repr__(self): + return repr(self.value) + + def __int__(self): + return self.value + + class RegisterSet(object): def __init__(self, reg_list, confidence = max_confidence): self.regs = reg_list -- cgit v1.3.1 From 105fb2549bd8fc0e19907fefff1168322dee3bb3 Mon Sep 17 00:00:00 2001 From: Rusty Wagner Date: Mon, 26 Feb 2018 22:37:19 -0500 Subject: Architecture plugins no longer need to override the perform_* methods (you can now override get_instruction_info, not perform_get_instruction_info). The perform_* methods are now deprecated but will still function as expected. Added architecture hooks to Python API using this new style. --- python/architecture.py | 1763 +++++++++++++++++++++++++----------------- python/basicblock.py | 2 +- python/binaryview.py | 6 +- python/callingconvention.py | 2 +- python/examples/arch_hook.py | 16 + python/examples/nes.py | 28 +- python/function.py | 8 +- python/platform.py | 2 +- 8 files changed, 1091 insertions(+), 736 deletions(-) create mode 100644 python/examples/arch_hook.py (limited to 'python/callingconvention.py') diff --git a/python/architecture.py b/python/architecture.py index 2b52ea35..781edcfd 100644 --- a/python/architecture.py +++ b/python/architecture.py @@ -44,7 +44,7 @@ class _ArchitectureMetaClass(type): archs = core.BNGetArchitectureList(count) result = [] for i in xrange(0, count.value): - result.append(Architecture(archs[i])) + result.append(CoreArchitecture(archs[i])) core.BNFreeArchitectureList(archs) return result @@ -54,7 +54,7 @@ class _ArchitectureMetaClass(type): archs = core.BNGetArchitectureList(count) try: for i in xrange(0, count.value): - yield Architecture(archs[i]) + yield CoreArchitecture(archs[i]) finally: core.BNFreeArchitectureList(archs) @@ -63,7 +63,7 @@ class _ArchitectureMetaClass(type): arch = core.BNGetArchitectureByName(name) if arch is None: raise KeyError("'%s' is not a valid architecture" % str(name)) - return Architecture(arch) + return CoreArchitecture(arch) def register(cls): startup._init_plugins() @@ -133,438 +133,253 @@ class Architecture(object): __metaclass__ = _ArchitectureMetaClass next_address = 0 - def __init__(self, handle=None): - if handle is not None: - self.handle = core.handle_of_type(handle, core.BNArchitecture) - self.__dict__["name"] = core.BNGetArchitectureName(self.handle) - self.__dict__["endianness"] = Endianness(core.BNGetArchitectureEndianness(self.handle)) - self.__dict__["address_size"] = core.BNGetArchitectureAddressSize(self.handle) - self.__dict__["default_int_size"] = core.BNGetArchitectureDefaultIntegerSize(self.handle) - self.__dict__["instr_alignment"] = core.BNGetArchitectureInstructionAlignment(self.handle) - self.__dict__["max_instr_length"] = core.BNGetArchitectureMaxInstructionLength(self.handle) - self.__dict__["opcode_display_length"] = core.BNGetArchitectureOpcodeDisplayLength(self.handle) - self.__dict__["stack_pointer"] = core.BNGetArchitectureRegisterName(self.handle, - core.BNGetArchitectureStackPointerRegister(self.handle)) - self.__dict__["link_reg"] = core.BNGetArchitectureRegisterName(self.handle, - core.BNGetArchitectureLinkRegister(self.handle)) - - count = ctypes.c_ulonglong() - regs = core.BNGetAllArchitectureRegisters(self.handle, count) - self.__dict__["regs"] = {} - for i in xrange(0, count.value): - name = core.BNGetArchitectureRegisterName(self.handle, regs[i]) - info = core.BNGetArchitectureRegisterInfo(self.handle, regs[i]) - full_width_reg = core.BNGetArchitectureRegisterName(self.handle, info.fullWidthRegister) - self.regs[name] = function.RegisterInfo(full_width_reg, info.size, info.offset, - ImplicitRegisterExtend(info.extend), regs[i]) - core.BNFreeRegisterList(regs) - - count = ctypes.c_ulonglong() - flags = core.BNGetAllArchitectureFlags(self.handle, count) - self._flags = {} - self._flags_by_index = {} - self.__dict__["flags"] = [] - for i in xrange(0, count.value): - name = core.BNGetArchitectureFlagName(self.handle, flags[i]) - self._flags[name] = flags[i] - self._flags_by_index[flags[i]] = name - self.flags.append(name) - core.BNFreeRegisterList(flags) - - count = ctypes.c_ulonglong() - write_types = core.BNGetAllArchitectureFlagWriteTypes(self.handle, count) - self._flag_write_types = {} - self._flag_write_types_by_index = {} - self.__dict__["flag_write_types"] = [] - for i in xrange(0, count.value): - name = core.BNGetArchitectureFlagWriteTypeName(self.handle, write_types[i]) - self._flag_write_types[name] = write_types[i] - self._flag_write_types_by_index[write_types[i]] = name - self.flag_write_types.append(name) - core.BNFreeRegisterList(write_types) - - count = ctypes.c_ulonglong() - sem_classes = core.BNGetAllArchitectureSemanticFlagClasses(self.handle, count) - self._semantic_flag_classes = {} - self._semantic_flag_classes_by_index = {} - self.__dict__["semantic_flag_classes"] = [] - for i in xrange(0, count.value): - name = core.BNGetArchitectureSemanticFlagClassName(self.handle, sem_classes[i]) - self._semantic_flag_classes[name] = sem_classes[i] - self._semantic_flag_classes_by_index[sem_classes[i]] = name - self.semantic_flag_classes.append(name) - core.BNFreeRegisterList(sem_classes) + def __init__(self): + startup._init_plugins() - count = ctypes.c_ulonglong() - sem_groups = core.BNGetAllArchitectureSemanticFlagGroups(self.handle, count) - self._semantic_flag_groups = {} - self._semantic_flag_groups_by_index = {} - self.__dict__["semantic_flag_groups"] = [] - for i in xrange(0, count.value): - name = core.BNGetArchitectureSemanticFlagGroupName(self.handle, sem_groups[i]) - self._semantic_flag_groups[name] = sem_groups[i] - self._semantic_flag_groups_by_index[sem_groups[i]] = name - self.semantic_flag_groups.append(name) - core.BNFreeRegisterList(sem_groups) - - self._flag_roles = {} - self.__dict__["flag_roles"] = {} - for flag in self.__dict__["flags"]: - role = FlagRole(core.BNGetArchitectureFlagRole(self.handle, self._flags[flag], 0)) - self.__dict__["flag_roles"][flag] = role - self._flag_roles[self._flags[flag]] = role - - self.__dict__["flags_required_for_flag_condition"] = {} - for cond in LowLevelILFlagCondition: - count = ctypes.c_ulonglong() - flags = core.BNGetArchitectureFlagsRequiredForFlagCondition(self.handle, cond, 0, count) - flag_names = [] - for i in xrange(0, count.value): - flag_names.append(self._flags_by_index[flags[i]]) - core.BNFreeRegisterList(flags) - self.__dict__["flags_required_for_flag_condition"][cond] = flag_names - - self._flags_required_by_semantic_flag_group = {} - self.__dict__["flags_required_for_semantic_flag_group"] = {} - for group in self.semantic_flag_groups: - count = ctypes.c_ulonglong() - flags = core.BNGetArchitectureFlagsRequiredForSemanticFlagGroup(self.handle, - self._semantic_flag_groups[group], count) - flag_indexes = [] - flag_names = [] - for i in xrange(0, count.value): - flag_indexes.append(flags[i]) - flag_names.append(self._flags_by_index[flags[i]]) - core.BNFreeRegisterList(flags) - self._flags_required_by_semantic_flag_group[self._semantic_flag_groups[group]] = flag_indexes - self.__dict__["flags_required_for_semantic_flag_group"][cond] = flag_names - - self._flag_conditions_for_semantic_flag_group = {} - self.__dict__["flag_conditions_for_semantic_flag_group"] = {} - for group in self.semantic_flag_groups: - count = ctypes.c_ulonglong() - conditions = core.BNGetArchitectureFlagConditionsForSemanticFlagGroup(self.handle, - self._semantic_flag_groups[group], count) - class_index_cond = {} - class_cond = {} - for i in xrange(0, count.value): - class_index_cond[conditions[i].semanticClass] = conditions[i].condition - if conditions[i].semanticClass == 0: - class_cond[None] = conditions[i].condition - elif conditions[i].semanticClass in self._semantic_flag_classes_by_index: - class_cond[self._semantic_flag_classes_by_index[conditions[i].semanticClass]] = conditions[i].condition - core.BNFreeFlagConditionsForSemanticFlagGroup(conditions) - self._flag_conditions_for_semantic_flag_group[self._semantic_flag_groups[group]] = class_index_cond - self.__dict__["flag_conditions_for_semantic_flag_group"][group] = class_cond - - self._flags_written_by_flag_write_type = {} - self.__dict__["flags_written_by_flag_write_type"] = {} - for write_type in self.flag_write_types: - count = ctypes.c_ulonglong() - flags = core.BNGetArchitectureFlagsWrittenByFlagWriteType(self.handle, - self._flag_write_types[write_type], count) - flag_indexes = [] - flag_names = [] - for i in xrange(0, count.value): - flag_indexes.append(flags[i]) - flag_names.append(self._flags_by_index[flags[i]]) - core.BNFreeRegisterList(flags) - self._flags_written_by_flag_write_type[self._flag_write_types[write_type]] = flag_indexes - self.__dict__["flags_written_by_flag_write_type"][write_type] = flag_names - - self._semantic_class_for_flag_write_type = {} - self.__dict__["semantic_class_for_flag_write_type"] = {} - for write_type in self.flag_write_types: - sem_class = core.BNGetArchitectureSemanticClassForFlagWriteType(self.handle, - self._flag_write_types[write_type]) - if sem_class == 0: - sem_class_name = None + if self.__class__.opcode_display_length > self.__class__.max_instr_length: + self.__class__.opcode_display_length = self.__class__.max_instr_length + + self._cb = core.BNCustomArchitecture() + self._cb.context = 0 + self._cb.init = self._cb.init.__class__(self._init) + self._cb.getEndianness = self._cb.getEndianness.__class__(self._get_endianness) + self._cb.getAddressSize = self._cb.getAddressSize.__class__(self._get_address_size) + self._cb.getDefaultIntegerSize = self._cb.getDefaultIntegerSize.__class__(self._get_default_integer_size) + self._cb.getInstructionAlignment = self._cb.getInstructionAlignment.__class__(self._get_instruction_alignment) + self._cb.getMaxInstructionLength = self._cb.getMaxInstructionLength.__class__(self._get_max_instruction_length) + self._cb.getOpcodeDisplayLength = self._cb.getOpcodeDisplayLength.__class__(self._get_opcode_display_length) + self._cb.getAssociatedArchitectureByAddress = \ + self._cb.getAssociatedArchitectureByAddress.__class__(self._get_associated_arch_by_address) + self._cb.getInstructionInfo = self._cb.getInstructionInfo.__class__(self._get_instruction_info) + self._cb.getInstructionText = self._cb.getInstructionText.__class__(self._get_instruction_text) + self._cb.freeInstructionText = self._cb.freeInstructionText.__class__(self._free_instruction_text) + self._cb.getInstructionLowLevelIL = self._cb.getInstructionLowLevelIL.__class__( + self._get_instruction_low_level_il) + self._cb.getRegisterName = self._cb.getRegisterName.__class__(self._get_register_name) + self._cb.getFlagName = self._cb.getFlagName.__class__(self._get_flag_name) + self._cb.getFlagWriteTypeName = self._cb.getFlagWriteTypeName.__class__(self._get_flag_write_type_name) + self._cb.getSemanticFlagClassName = self._cb.getSemanticFlagClassName.__class__(self._get_semantic_flag_class_name) + self._cb.getSemanticFlagGroupName = self._cb.getSemanticFlagGroupName.__class__(self._get_semantic_flag_group_name) + self._cb.getFullWidthRegisters = self._cb.getFullWidthRegisters.__class__(self._get_full_width_registers) + self._cb.getAllRegisters = self._cb.getAllRegisters.__class__(self._get_all_registers) + self._cb.getAllFlags = self._cb.getAllRegisters.__class__(self._get_all_flags) + self._cb.getAllFlagWriteTypes = self._cb.getAllRegisters.__class__(self._get_all_flag_write_types) + self._cb.getAllSemanticFlagClasses = self._cb.getAllSemanticFlagClasses.__class__(self._get_all_semantic_flag_classes) + self._cb.getAllSemanticFlagGroups = self._cb.getAllSemanticFlagGroups.__class__(self._get_all_semantic_flag_groups) + self._cb.getFlagRole = self._cb.getFlagRole.__class__(self._get_flag_role) + self._cb.getFlagsRequiredForFlagCondition = self._cb.getFlagsRequiredForFlagCondition.__class__( + self._get_flags_required_for_flag_condition) + self._cb.getFlagsRequiredForSemanticFlagGroup = self._cb.getFlagsRequiredForSemanticFlagGroup.__class__( + self._get_flags_required_for_semantic_flag_group) + self._cb.getFlagConditionsForSemanticFlagGroup = self._cb.getFlagConditionsForSemanticFlagGroup.__class__( + self._get_flag_conditions_for_semantic_flag_group) + self._cb.freeFlagConditionsForSemanticFlagGroup = self._cb.freeFlagConditionsForSemanticFlagGroup.__class__( + self._free_flag_conditions_for_semantic_flag_group) + self._cb.getFlagsWrittenByFlagWriteType = self._cb.getFlagsWrittenByFlagWriteType.__class__( + self._get_flags_written_by_flag_write_type) + self._cb.getSemanticClassForFlagWriteType = self._cb.getSemanticClassForFlagWriteType.__class__( + self._get_semantic_class_for_flag_write_type) + self._cb.getFlagWriteLowLevelIL = self._cb.getFlagWriteLowLevelIL.__class__( + self._get_flag_write_low_level_il) + self._cb.getFlagConditionLowLevelIL = self._cb.getFlagConditionLowLevelIL.__class__( + self._get_flag_condition_low_level_il) + self._cb.getSemanticFlagGroupLowLevelIL = self._cb.getSemanticFlagGroupLowLevelIL.__class__( + self._get_semantic_flag_group_low_level_il) + self._cb.freeRegisterList = self._cb.freeRegisterList.__class__(self._free_register_list) + self._cb.getRegisterInfo = self._cb.getRegisterInfo.__class__(self._get_register_info) + self._cb.getStackPointerRegister = self._cb.getStackPointerRegister.__class__( + self._get_stack_pointer_register) + self._cb.getLinkRegister = self._cb.getLinkRegister.__class__(self._get_link_register) + self._cb.getGlobalRegisters = self._cb.getGlobalRegisters.__class__(self._get_global_registers) + self._cb.getRegisterStackName = self._cb.getRegisterStackName.__class__(self._get_register_stack_name) + self._cb.getAllRegisterStacks = self._cb.getAllRegisterStacks.__class__(self._get_all_register_stacks) + self._cb.getRegisterStackInfo = self._cb.getRegisterStackInfo.__class__(self._get_register_stack_info) + self._cb.getIntrinsicName = self._cb.getIntrinsicName.__class__(self._get_intrinsic_name) + self._cb.getAllIntrinsics = self._cb.getAllIntrinsics.__class__(self._get_all_intrinsics) + self._cb.getIntrinsicInputs = self._cb.getIntrinsicInputs.__class__(self._get_intrinsic_inputs) + self._cb.freeNameAndTypeList = self._cb.freeNameAndTypeList.__class__(self._free_name_and_type_list) + self._cb.getIntrinsicOutputs = self._cb.getIntrinsicOutputs.__class__(self._get_intrinsic_outputs) + self._cb.freeTypeList = self._cb.freeTypeList.__class__(self._free_type_list) + self._cb.assemble = self._cb.assemble.__class__(self._assemble) + self._cb.isNeverBranchPatchAvailable = self._cb.isNeverBranchPatchAvailable.__class__( + self._is_never_branch_patch_available) + self._cb.isAlwaysBranchPatchAvailable = self._cb.isAlwaysBranchPatchAvailable.__class__( + self._is_always_branch_patch_available) + self._cb.isInvertBranchPatchAvailable = self._cb.isInvertBranchPatchAvailable.__class__( + self._is_invert_branch_patch_available) + self._cb.isSkipAndReturnZeroPatchAvailable = self._cb.isSkipAndReturnZeroPatchAvailable.__class__( + self._is_skip_and_return_zero_patch_available) + self._cb.isSkipAndReturnValuePatchAvailable = self._cb.isSkipAndReturnValuePatchAvailable.__class__( + self._is_skip_and_return_value_patch_available) + self._cb.convertToNop = self._cb.convertToNop.__class__(self._convert_to_nop) + self._cb.alwaysBranch = self._cb.alwaysBranch.__class__(self._always_branch) + self._cb.invertBranch = self._cb.invertBranch.__class__(self._invert_branch) + self._cb.skipAndReturnValue = self._cb.skipAndReturnValue.__class__(self._skip_and_return_value) + + self.__dict__["endianness"] = self.__class__.endianness + self.__dict__["address_size"] = self.__class__.address_size + self.__dict__["default_int_size"] = self.__class__.default_int_size + self.__dict__["instr_alignment"] = self.__class__.instr_alignment + self.__dict__["max_instr_length"] = self.__class__.max_instr_length + self.__dict__["opcode_display_length"] = self.__class__.opcode_display_length + self.__dict__["stack_pointer"] = self.__class__.stack_pointer + self.__dict__["link_reg"] = self.__class__.link_reg + + self._all_regs = {} + self._full_width_regs = {} + self._regs_by_index = {} + self.__dict__["regs"] = self.__class__.regs + reg_index = 0 + + # Registers used for storage in register stacks must be sequential, so allocate these in order first + self._all_reg_stacks = {} + self._reg_stacks_by_index = {} + self.__dict__["reg_stacks"] = self.__class__.reg_stacks + reg_stack_index = 0 + for reg_stack in self.reg_stacks: + info = self.reg_stacks[reg_stack] + for reg in info.storage_regs: + self._all_regs[reg] = reg_index + self._regs_by_index[reg_index] = reg + self.regs[reg].index = reg_index + reg_index += 1 + for reg in info.top_relative_regs: + self._all_regs[reg] = reg_index + self._regs_by_index[reg_index] = reg + self.regs[reg].index = reg_index + reg_index += 1 + if reg_stack not in self._all_reg_stacks: + self._all_reg_stacks[reg_stack] = reg_stack_index + self._reg_stacks_by_index[reg_stack_index] = reg_stack + self.reg_stacks[reg_stack].index = reg_stack_index + reg_stack_index += 1 + + for reg in self.regs: + info = self.regs[reg] + if reg not in self._all_regs: + self._all_regs[reg] = reg_index + self._regs_by_index[reg_index] = reg + self.regs[reg].index = reg_index + reg_index += 1 + if info.full_width_reg not in self._all_regs: + self._all_regs[info.full_width_reg] = reg_index + self._regs_by_index[reg_index] = info.full_width_reg + self.regs[info.full_width_reg].index = reg_index + reg_index += 1 + if info.full_width_reg not in self._full_width_regs: + self._full_width_regs[info.full_width_reg] = self._all_regs[info.full_width_reg] + + self._flags = {} + self._flags_by_index = {} + self.__dict__["flags"] = self.__class__.flags + flag_index = 0 + for flag in self.__class__.flags: + if flag not in self._flags: + self._flags[flag] = flag_index + self._flags_by_index[flag_index] = flag + flag_index += 1 + + self._flag_write_types = {} + self._flag_write_types_by_index = {} + self.__dict__["flag_write_types"] = self.__class__.flag_write_types + write_type_index = 0 + for write_type in self.__class__.flag_write_types: + if write_type not in self._flag_write_types: + self._flag_write_types[write_type] = write_type_index + self._flag_write_types_by_index[write_type_index] = write_type + write_type_index += 1 + + self._semantic_flag_classes = {} + self._semantic_flag_classes_by_index = {} + self.__dict__["semantic_flag_classes"] = self.__class__.semantic_flag_classes + semantic_class_index = 1 + for sem_class in self.__class__.semantic_flag_classes: + if sem_class not in self._semantic_flag_classes: + self._semantic_flag_classes[sem_class] = semantic_class_index + self._semantic_flag_classes_by_index[semantic_class_index] = sem_class + semantic_class_index += 1 + + self._semantic_flag_groups = {} + self._semantic_flag_groups_by_index = {} + self.__dict__["semantic_flag_groups"] = self.__class__.semantic_flag_groups + semantic_group_index = 0 + for sem_group in self.__class__.semantic_flag_groups: + if sem_group not in self._semantic_flag_groups: + self._semantic_flag_groups[sem_group] = semantic_group_index + self._semantic_flag_groups_by_index[semantic_group_index] = sem_group + semantic_group_index += 1 + + self._flag_roles = {} + self.__dict__["flag_roles"] = self.__class__.flag_roles + for flag in self.__class__.flag_roles: + role = self.__class__.flag_roles[flag] + if isinstance(role, str): + role = FlagRole[role] + self._flag_roles[self._flags[flag]] = role + + self.__dict__["flags_required_for_flag_condition"] = self.__class__.flags_required_for_flag_condition + + self._flags_required_by_semantic_flag_group = {} + self.__dict__["flags_required_for_semantic_flag_group"] = self.__class__.flags_required_for_semantic_flag_group + for group in self.__class__.flags_required_for_semantic_flag_group: + flags = [] + for flag in self.__class__.flags_required_for_semantic_flag_group[group]: + flags.append(self._flags[flag]) + self._flags_required_by_semantic_flag_group[self._semantic_flag_groups[group]] = flags + + self._flag_conditions_for_semantic_flag_group = {} + self.__dict__["flag_conditions_for_semantic_flag_group"] = self.__class__.flag_conditions_for_semantic_flag_group + for group in self.__class__.flag_conditions_for_semantic_flag_group: + class_cond = {} + for sem_class in self.__class__.flag_conditions_for_semantic_flag_group[group]: + if sem_class is None: + class_cond[0] = self.__class__.flag_conditions_for_semantic_flag_group[group][sem_class] else: - sem_class_name = self._semantic_flag_classes_by_index[sem_class] - self._semantic_class_for_flag_write_type[self._flag_write_types[write_type]] = sem_class - self.__dict__["semantic_class_for_flag_write_type"][write_type] = sem_class_name - - count = ctypes.c_ulonglong() - regs = core.BNGetArchitectureGlobalRegisters(self.handle, count) - self.__dict__["global_regs"] = [] - for i in xrange(0, count.value): - self.global_regs.append(core.BNGetArchitectureRegisterName(self.handle, regs[i])) - core.BNFreeRegisterList(regs) - - count = ctypes.c_ulonglong() - regs = core.BNGetAllArchitectureRegisterStacks(self.handle, count) - self.__dict__["reg_stacks"] = {} - for i in xrange(0, count.value): - name = core.BNGetArchitectureRegisterStackName(self.handle, regs[i]) - info = core.BNGetArchitectureRegisterStackInfo(self.handle, regs[i]) - storage = [] - for j in xrange(0, info.storageCount): - storage.append(core.BNGetArchitectureRegisterName(self.handle, info.firstStorageReg + j)) - top_rel = [] - for j in xrange(0, info.topRelativeCount): - top_rel.append(core.BNGetArchitectureRegisterName(self.handle, info.firstTopRelativeReg + j)) - top = core.BNGetArchitectureRegisterName(self.handle, info.stackTopReg) - self.reg_stacks[name] = function.RegisterStackInfo(storage, top_rel, top, regs[i]) - core.BNFreeRegisterList(regs) + class_cond[self._semantic_flag_classes[sem_class]] = self.__class__.flag_conditions_for_semantic_flag_group[group][sem_class] + self._flag_conditions_for_semantic_flag_group[self._semantic_flag_groups[group]] = class_cond - count = ctypes.c_ulonglong() - intrinsics = core.BNGetAllArchitectureIntrinsics(self.handle, count) - self.__dict__["intrinsics"] = {} - for i in xrange(0, count.value): - name = core.BNGetArchitectureIntrinsicName(self.handle, intrinsics[i]) - input_count = ctypes.c_ulonglong() - inputs = core.BNGetArchitectureIntrinsicInputs(self.handle, intrinsics[i], input_count) - input_list = [] - for j in xrange(0, input_count.value): - input_name = inputs[j].name - type_obj = types.Type(core.BNNewTypeReference(inputs[j].type), confidence = inputs[j].typeConfidence) - input_list.append(function.IntrinsicInput(type_obj, input_name)) - core.BNFreeNameAndTypeList(inputs, input_count.value) - output_count = ctypes.c_ulonglong() - outputs = core.BNGetArchitectureIntrinsicOutputs(self.handle, intrinsics[i], output_count) - output_list = [] - for j in xrange(0, output_count.value): - output_list.append(types.Type(core.BNNewTypeReference(outputs[j].type), confidence = outputs[j].confidence)) - core.BNFreeOutputTypeList(outputs, output_count.value) - self.intrinsics[name] = function.IntrinsicInfo(input_list, output_list) - else: - startup._init_plugins() - - if self.__class__.opcode_display_length > self.__class__.max_instr_length: - self.__class__.opcode_display_length = self.__class__.max_instr_length - - self._cb = core.BNCustomArchitecture() - self._cb.context = 0 - self._cb.init = self._cb.init.__class__(self._init) - self._cb.getEndianness = self._cb.getEndianness.__class__(self._get_endianness) - self._cb.getAddressSize = self._cb.getAddressSize.__class__(self._get_address_size) - self._cb.getDefaultIntegerSize = self._cb.getDefaultIntegerSize.__class__(self._get_default_integer_size) - self._cb.getInstructionAlignment = self._cb.getInstructionAlignment.__class__(self._get_instruction_alignment) - self._cb.getMaxInstructionLength = self._cb.getMaxInstructionLength.__class__(self._get_max_instruction_length) - self._cb.getOpcodeDisplayLength = self._cb.getOpcodeDisplayLength.__class__(self._get_opcode_display_length) - self._cb.getAssociatedArchitectureByAddress = \ - self._cb.getAssociatedArchitectureByAddress.__class__(self._get_associated_arch_by_address) - self._cb.getInstructionInfo = self._cb.getInstructionInfo.__class__(self._get_instruction_info) - self._cb.getInstructionText = self._cb.getInstructionText.__class__(self._get_instruction_text) - self._cb.freeInstructionText = self._cb.freeInstructionText.__class__(self._free_instruction_text) - self._cb.getInstructionLowLevelIL = self._cb.getInstructionLowLevelIL.__class__( - self._get_instruction_low_level_il) - self._cb.getRegisterName = self._cb.getRegisterName.__class__(self._get_register_name) - self._cb.getFlagName = self._cb.getFlagName.__class__(self._get_flag_name) - self._cb.getFlagWriteTypeName = self._cb.getFlagWriteTypeName.__class__(self._get_flag_write_type_name) - self._cb.getSemanticFlagClassName = self._cb.getSemanticFlagClassName.__class__(self._get_semantic_flag_class_name) - self._cb.getSemanticFlagGroupName = self._cb.getSemanticFlagGroupName.__class__(self._get_semantic_flag_group_name) - self._cb.getFullWidthRegisters = self._cb.getFullWidthRegisters.__class__(self._get_full_width_registers) - self._cb.getAllRegisters = self._cb.getAllRegisters.__class__(self._get_all_registers) - self._cb.getAllFlags = self._cb.getAllRegisters.__class__(self._get_all_flags) - self._cb.getAllFlagWriteTypes = self._cb.getAllRegisters.__class__(self._get_all_flag_write_types) - self._cb.getAllSemanticFlagClasses = self._cb.getAllSemanticFlagClasses.__class__(self._get_all_semantic_flag_classes) - self._cb.getAllSemanticFlagGroups = self._cb.getAllSemanticFlagGroups.__class__(self._get_all_semantic_flag_groups) - self._cb.getFlagRole = self._cb.getFlagRole.__class__(self._get_flag_role) - self._cb.getFlagsRequiredForFlagCondition = self._cb.getFlagsRequiredForFlagCondition.__class__( - self._get_flags_required_for_flag_condition) - self._cb.getFlagsRequiredForSemanticFlagGroup = self._cb.getFlagsRequiredForSemanticFlagGroup.__class__( - self._get_flags_required_for_semantic_flag_group) - self._cb.getFlagConditionsForSemanticFlagGroup = self._cb.getFlagConditionsForSemanticFlagGroup.__class__( - self._get_flag_conditions_for_semantic_flag_group) - self._cb.freeFlagConditionsForSemanticFlagGroup = self._cb.freeFlagConditionsForSemanticFlagGroup.__class__( - self._free_flag_conditions_for_semantic_flag_group) - self._cb.getFlagsWrittenByFlagWriteType = self._cb.getFlagsWrittenByFlagWriteType.__class__( - self._get_flags_written_by_flag_write_type) - self._cb.getSemanticClassForFlagWriteType = self._cb.getSemanticClassForFlagWriteType.__class__( - self._get_semantic_class_for_flag_write_type) - self._cb.getFlagWriteLowLevelIL = self._cb.getFlagWriteLowLevelIL.__class__( - self._get_flag_write_low_level_il) - self._cb.getFlagConditionLowLevelIL = self._cb.getFlagConditionLowLevelIL.__class__( - self._get_flag_condition_low_level_il) - self._cb.getSemanticFlagGroupLowLevelIL = self._cb.getSemanticFlagGroupLowLevelIL.__class__( - self._get_semantic_flag_group_low_level_il) - self._cb.freeRegisterList = self._cb.freeRegisterList.__class__(self._free_register_list) - self._cb.getRegisterInfo = self._cb.getRegisterInfo.__class__(self._get_register_info) - self._cb.getStackPointerRegister = self._cb.getStackPointerRegister.__class__( - self._get_stack_pointer_register) - self._cb.getLinkRegister = self._cb.getLinkRegister.__class__(self._get_link_register) - self._cb.getGlobalRegisters = self._cb.getGlobalRegisters.__class__(self._get_global_registers) - self._cb.getRegisterStackName = self._cb.getRegisterStackName.__class__(self._get_register_stack_name) - self._cb.getAllRegisterStacks = self._cb.getAllRegisterStacks.__class__(self._get_all_register_stacks) - self._cb.getRegisterStackInfo = self._cb.getRegisterStackInfo.__class__(self._get_register_stack_info) - self._cb.getIntrinsicName = self._cb.getIntrinsicName.__class__(self._get_intrinsic_name) - self._cb.getAllIntrinsics = self._cb.getAllIntrinsics.__class__(self._get_all_intrinsics) - self._cb.getIntrinsicInputs = self._cb.getIntrinsicInputs.__class__(self._get_intrinsic_inputs) - self._cb.freeNameAndTypeList = self._cb.freeNameAndTypeList.__class__(self._free_name_and_type_list) - self._cb.getIntrinsicOutputs = self._cb.getIntrinsicOutputs.__class__(self._get_intrinsic_outputs) - self._cb.freeTypeList = self._cb.freeTypeList.__class__(self._free_type_list) - self._cb.assemble = self._cb.assemble.__class__(self._assemble) - self._cb.isNeverBranchPatchAvailable = self._cb.isNeverBranchPatchAvailable.__class__( - self._is_never_branch_patch_available) - self._cb.isAlwaysBranchPatchAvailable = self._cb.isAlwaysBranchPatchAvailable.__class__( - self._is_always_branch_patch_available) - self._cb.isInvertBranchPatchAvailable = self._cb.isInvertBranchPatchAvailable.__class__( - self._is_invert_branch_patch_available) - self._cb.isSkipAndReturnZeroPatchAvailable = self._cb.isSkipAndReturnZeroPatchAvailable.__class__( - self._is_skip_and_return_zero_patch_available) - self._cb.isSkipAndReturnValuePatchAvailable = self._cb.isSkipAndReturnValuePatchAvailable.__class__( - self._is_skip_and_return_value_patch_available) - self._cb.convertToNop = self._cb.convertToNop.__class__(self._convert_to_nop) - self._cb.alwaysBranch = self._cb.alwaysBranch.__class__(self._always_branch) - self._cb.invertBranch = self._cb.invertBranch.__class__(self._invert_branch) - self._cb.skipAndReturnValue = self._cb.skipAndReturnValue.__class__(self._skip_and_return_value) - - self._all_regs = {} - self._full_width_regs = {} - self._regs_by_index = {} - self.__dict__["regs"] = self.__class__.regs - reg_index = 0 - - # Registers used for storage in register stacks must be sequential, so allocate these in order first - self._all_reg_stacks = {} - self._reg_stacks_by_index = {} - self.__dict__["reg_stacks"] = self.__class__.reg_stacks - reg_stack_index = 0 - for reg_stack in self.reg_stacks: - info = self.reg_stacks[reg_stack] - for reg in info.storage_regs: - self._all_regs[reg] = reg_index - self._regs_by_index[reg_index] = reg - self.regs[reg].index = reg_index - reg_index += 1 - for reg in info.top_relative_regs: - self._all_regs[reg] = reg_index - self._regs_by_index[reg_index] = reg - self.regs[reg].index = reg_index - reg_index += 1 - if reg_stack not in self._all_reg_stacks: - self._all_reg_stacks[reg_stack] = reg_stack_index - self._reg_stacks_by_index[reg_stack_index] = reg_stack - self.reg_stacks[reg_stack].index = reg_stack_index - reg_stack_index += 1 - - for reg in self.regs: - info = self.regs[reg] - if reg not in self._all_regs: - self._all_regs[reg] = reg_index - self._regs_by_index[reg_index] = reg - self.regs[reg].index = reg_index - reg_index += 1 - if info.full_width_reg not in self._all_regs: - self._all_regs[info.full_width_reg] = reg_index - self._regs_by_index[reg_index] = info.full_width_reg - self.regs[info.full_width_reg].index = reg_index - reg_index += 1 - if info.full_width_reg not in self._full_width_regs: - self._full_width_regs[info.full_width_reg] = self._all_regs[info.full_width_reg] - - self._flags = {} - self._flags_by_index = {} - self.__dict__["flags"] = self.__class__.flags - flag_index = 0 - for flag in self.__class__.flags: - if flag not in self._flags: - self._flags[flag] = flag_index - self._flags_by_index[flag_index] = flag - flag_index += 1 - - self._flag_write_types = {} - self._flag_write_types_by_index = {} - self.__dict__["flag_write_types"] = self.__class__.flag_write_types - write_type_index = 0 - for write_type in self.__class__.flag_write_types: - if write_type not in self._flag_write_types: - self._flag_write_types[write_type] = write_type_index - self._flag_write_types_by_index[write_type_index] = write_type - write_type_index += 1 - - self._semantic_flag_classes = {} - self._semantic_flag_classes_by_index = {} - self.__dict__["semantic_flag_classes"] = self.__class__.semantic_flag_classes - semantic_class_index = 1 - for sem_class in self.__class__.semantic_flag_classes: - if sem_class not in self._semantic_flag_classes: - self._semantic_flag_classes[sem_class] = semantic_class_index - self._semantic_flag_classes_by_index[semantic_class_index] = sem_class - semantic_class_index += 1 - - self._semantic_flag_groups = {} - self._semantic_flag_groups_by_index = {} - self.__dict__["semantic_flag_groups"] = self.__class__.semantic_flag_groups - semantic_group_index = 0 - for sem_group in self.__class__.semantic_flag_groups: - if sem_group not in self._semantic_flag_groups: - self._semantic_flag_groups[sem_group] = semantic_group_index - self._semantic_flag_groups_by_index[semantic_group_index] = sem_group - semantic_group_index += 1 - - self._flag_roles = {} - self.__dict__["flag_roles"] = self.__class__.flag_roles - for flag in self.__class__.flag_roles: - role = self.__class__.flag_roles[flag] - if isinstance(role, str): - role = FlagRole[role] - self._flag_roles[self._flags[flag]] = role - - self.__dict__["flags_required_for_flag_condition"] = self.__class__.flags_required_for_flag_condition - - self._flags_required_by_semantic_flag_group = {} - self.__dict__["flags_required_for_semantic_flag_group"] = self.__class__.flags_required_for_semantic_flag_group - for group in self.__class__.flags_required_for_semantic_flag_group: - flags = [] - for flag in self.__class__.flags_required_for_semantic_flag_group[group]: - flags.append(self._flags[flag]) - self._flags_required_by_semantic_flag_group[self._semantic_flag_groups[group]] = flags - - self._flag_conditions_for_semantic_flag_group = {} - self.__dict__["flag_conditions_for_semantic_flag_group"] = self.__class__.flag_conditions_for_semantic_flag_group - for group in self.__class__.flag_conditions_for_semantic_flag_group: - class_cond = {} - for sem_class in self.__class__.flag_conditions_for_semantic_flag_group[group]: - if sem_class is None: - class_cond[0] = self.__class__.flag_conditions_for_semantic_flag_group[group][sem_class] - else: - class_cond[self._semantic_flag_classes[sem_class]] = self.__class__.flag_conditions_for_semantic_flag_group[group][sem_class] - self._flag_conditions_for_semantic_flag_group[self._semantic_flag_groups[group]] = class_cond - - self._flags_written_by_flag_write_type = {} - self.__dict__["flags_written_by_flag_write_type"] = self.__class__.flags_written_by_flag_write_type - for write_type in self.__class__.flags_written_by_flag_write_type: - flags = [] - for flag in self.__class__.flags_written_by_flag_write_type[write_type]: - flags.append(self._flags[flag]) - self._flags_written_by_flag_write_type[self._flag_write_types[write_type]] = flags - - self._semantic_class_for_flag_write_type = {} - self.__dict__["semantic_class_for_flag_write_type"] = self.__class__.semantic_class_for_flag_write_type - for write_type in self.__class__.semantic_class_for_flag_write_type: - sem_class = self.__class__.semantic_class_for_flag_write_type[write_type] - if sem_class in self._semantic_flag_classes: - sem_class_index = self._semantic_flag_classes[sem_class] - else: - sem_class_index = 0 - self._semantic_class_for_flag_write_type[self._flag_write_types[write_type]] = sem_class_index - - self.__dict__["global_regs"] = self.__class__.global_regs - - self._intrinsics = {} - self._intrinsics_by_index = {} - self.__dict__["intrinsics"] = self.__class__.intrinsics - intrinsic_index = 0 - for intrinsic in self.__class__.intrinsics.keys(): - if intrinsic not in self._intrinsics: - info = self.__class__.intrinsics[intrinsic] - for i in xrange(0, len(info.inputs)): - if isinstance(info.inputs[i], types.Type): - info.inputs[i] = function.IntrinsicInput(info.inputs[i]) - elif isinstance(info.inputs[i], tuple): - info.inputs[i] = function.IntrinsicInput(info.inputs[i][0], info.inputs[i][1]) - info.index = intrinsic_index - self._intrinsics[intrinsic] = intrinsic_index - self._intrinsics_by_index[intrinsic_index] = (intrinsic, info) - intrinsic_index += 1 - - self._pending_reg_lists = {} - self._pending_token_lists = {} - self._pending_condition_lists = {} - self._pending_name_and_type_lists = {} - self._pending_type_lists = {} + self._flags_written_by_flag_write_type = {} + self.__dict__["flags_written_by_flag_write_type"] = self.__class__.flags_written_by_flag_write_type + for write_type in self.__class__.flags_written_by_flag_write_type: + flags = [] + for flag in self.__class__.flags_written_by_flag_write_type[write_type]: + flags.append(self._flags[flag]) + self._flags_written_by_flag_write_type[self._flag_write_types[write_type]] = flags + + self._semantic_class_for_flag_write_type = {} + self.__dict__["semantic_class_for_flag_write_type"] = self.__class__.semantic_class_for_flag_write_type + for write_type in self.__class__.semantic_class_for_flag_write_type: + sem_class = self.__class__.semantic_class_for_flag_write_type[write_type] + if sem_class in self._semantic_flag_classes: + sem_class_index = self._semantic_flag_classes[sem_class] + else: + sem_class_index = 0 + self._semantic_class_for_flag_write_type[self._flag_write_types[write_type]] = sem_class_index + + self.__dict__["global_regs"] = self.__class__.global_regs + + self._intrinsics = {} + self._intrinsics_by_index = {} + self.__dict__["intrinsics"] = self.__class__.intrinsics + intrinsic_index = 0 + for intrinsic in self.__class__.intrinsics.keys(): + if intrinsic not in self._intrinsics: + info = self.__class__.intrinsics[intrinsic] + for i in xrange(0, len(info.inputs)): + if isinstance(info.inputs[i], types.Type): + info.inputs[i] = function.IntrinsicInput(info.inputs[i]) + elif isinstance(info.inputs[i], tuple): + info.inputs[i] = function.IntrinsicInput(info.inputs[i][0], info.inputs[i][1]) + info.index = intrinsic_index + self._intrinsics[intrinsic] = intrinsic_index + self._intrinsics_by_index[intrinsic_index] = (intrinsic, info) + intrinsic_index += 1 + + self._pending_reg_lists = {} + self._pending_token_lists = {} + self._pending_condition_lists = {} + self._pending_name_and_type_lists = {} + self._pending_type_lists = {} def __eq__(self, value): if not isinstance(value, Architecture): @@ -624,49 +439,49 @@ class Architecture(object): def _get_endianness(self, ctxt): try: - return self.__class__.endianness + return self.endianness except: log.log_error(traceback.format_exc()) return Endianness.LittleEndian def _get_address_size(self, ctxt): try: - return self.__class__.address_size + return self.address_size except: log.log_error(traceback.format_exc()) return 8 def _get_default_integer_size(self, ctxt): try: - return self.__class__.default_int_size + return self.default_int_size except: log.log_error(traceback.format_exc()) return 4 def _get_instruction_alignment(self, ctxt): try: - return self.__class__.instr_alignment + return self.instr_alignment except: log.log_error(traceback.format_exc()) return 1 def _get_max_instruction_length(self, ctxt): try: - return self.__class__.max_instr_length + return self.max_instr_length except: log.log_error(traceback.format_exc()) return 16 def _get_opcode_display_length(self, ctxt): try: - return self.__class__.opcode_display_length + return self.opcode_display_length except: log.log_error(traceback.format_exc()) return 8 def _get_associated_arch_by_address(self, ctxt, addr): try: - result, new_addr = self.perform_get_associated_arch_by_address(addr[0]) + result, new_addr = self.get_associated_arch_by_address(addr[0]) addr[0] = new_addr return ctypes.cast(result.handle, ctypes.c_void_p).value except: @@ -677,7 +492,7 @@ class Architecture(object): try: buf = ctypes.create_string_buffer(max_len) ctypes.memmove(buf, data, max_len) - info = self.perform_get_instruction_info(buf.raw, addr) + info = self.get_instruction_info(buf.raw, addr) if info is None: return False result[0].length = info.length @@ -703,7 +518,7 @@ class Architecture(object): try: buf = ctypes.create_string_buffer(length[0]) ctypes.memmove(buf, data, length[0]) - info = self.perform_get_instruction_text(buf.raw, addr) + info = self.get_instruction_text(buf.raw, addr) if info is None: return False tokens = info[0] @@ -743,7 +558,7 @@ class Architecture(object): try: buf = ctypes.create_string_buffer(length[0]) ctypes.memmove(buf, data, length[0]) - result = self.perform_get_instruction_low_level_il(buf.raw, addr, + result = self.get_instruction_low_level_il(buf.raw, addr, lowlevelil.LowLevelILFunction(self, core.BNNewLowLevelILFunctionReference(il))) if result is None: return False @@ -782,8 +597,8 @@ class Architecture(object): def _get_semantic_flag_class_name(self, ctxt, sem_class): try: - if sem_class in self._semantic_flag_class_by_index: - return core.BNAllocString(self._semantic_flag_class_by_index[sem_class]) + if sem_class in self._semantic_flag_classes_by_index: + return core.BNAllocString(self._semantic_flag_classes_by_index[sem_class]) return core.BNAllocString("") except (KeyError, OSError): log.log_error(traceback.format_exc()) @@ -791,8 +606,8 @@ class Architecture(object): def _get_semantic_flag_group_name(self, ctxt, sem_group): try: - if sem_group in self._semantic_flag_group_by_index: - return core.BNAllocString(self._semantic_flag_group_by_index[sem_group]) + if sem_group in self._semantic_flag_groups_by_index: + return core.BNAllocString(self._semantic_flag_groups_by_index[sem_group]) return core.BNAllocString("") except (KeyError, OSError): log.log_error(traceback.format_exc()) @@ -894,23 +709,18 @@ class Architecture(object): sem_class = self._semantic_flag_classes_by_index[sem_class] else: sem_class = None - return self.perform_get_flag_role(flag, sem_class) + return self.get_flag_role(flag, sem_class) except KeyError: log.log_error(traceback.format_exc()) return FlagRole.SpecialFlagRole - def perform_get_flag_role(self, flag, sem_class): - if flag in self._flag_roles: - return self._flag_roles[flag] - return FlagRole.SpecialFlagRole - def _get_flags_required_for_flag_condition(self, ctxt, cond, sem_class, count): try: if sem_class in self._semantic_flag_classes_by_index: sem_class = self._semantic_flag_classes_by_index[sem_class] else: sem_class = None - flag_names = self.perform_get_flags_required_for_flag_condition(cond, sem_class) + flag_names = self.get_flags_required_for_flag_condition(cond, sem_class) flags = [] for name in flag_names: flags.append(self._flags[name]) @@ -926,11 +736,6 @@ class Architecture(object): count[0] = 0 return None - def perform_get_flags_required_for_flag_condition(self, cond, sem_class): - if cond in self.flags_required_for_flag_condition: - return self.flags_required_for_flag_condition[cond] - return [] - def _get_flags_required_for_semantic_flag_group(self, ctxt, sem_group, count): try: if sem_group in self._flags_required_by_semantic_flag_group: @@ -951,8 +756,8 @@ class Architecture(object): def _get_flag_conditions_for_semantic_flag_group(self, ctxt, sem_group, count): try: - if sem_group in self._flag_conditions_by_semantic_flag_group: - class_cond = self._flag_conditions_by_semantic_flag_group[sem_group] + if sem_group in self._flag_conditions_for_semantic_flag_group: + class_cond = self._flag_conditions_for_semantic_flag_group[sem_group] else: class_cond = {} count[0] = len(class_cond) @@ -963,7 +768,7 @@ class Architecture(object): cond_buf[i].condition = class_cond[class_index] i += 1 result = ctypes.cast(cond_buf, ctypes.c_void_p) - self._pending_conditions[result.value] = (result, cond_buf) + self._pending_condition_lists[result.value] = (result, cond_buf) return result.value except (KeyError, OSError): log.log_error(traceback.format_exc()) @@ -973,9 +778,9 @@ class Architecture(object): def _free_flag_conditions_for_semantic_flag_group(self, ctxt, conditions): try: buf = ctypes.cast(conditions, ctypes.c_void_p) - if buf.value not in self._pending_conditions: + if buf.value not in self._pending_condition_lists: raise ValueError("freeing condition list that wasn't allocated") - del self._pending_conditions[buf.value] + del self._pending_condition_lists[buf.value] except (ValueError, KeyError): log.log_error(traceback.format_exc()) @@ -1021,7 +826,7 @@ class Architecture(object): operand_list.append(lowlevelil.ILRegister(self, operands[i].reg)) else: operand_list.append(lowlevelil.ILRegister(self, operands[i].reg)) - return self.perform_get_flag_write_low_level_il(op, size, write_type_name, flag_name, operand_list, + return self.get_flag_write_low_level_il(op, size, write_type_name, flag_name, operand_list, lowlevelil.LowLevelILFunction(self, core.BNNewLowLevelILFunctionReference(il))).index except (KeyError, OSError): log.log_error(traceback.format_exc()) @@ -1033,7 +838,7 @@ class Architecture(object): sem_class_name = self._semantic_flag_classes_by_index[sem_class] else: sem_class_name = None - return self.perform_get_flag_condition_low_level_il(cond, sem_class_name, + return self.get_flag_condition_low_level_il(cond, sem_class_name, lowlevelil.LowLevelILFunction(self, core.BNNewLowLevelILFunctionReference(il))).index except OSError: log.log_error(traceback.format_exc()) @@ -1045,7 +850,7 @@ class Architecture(object): sem_group_name = self._semantic_flag_groups_by_index[sem_group] else: sem_group_name = None - return self.perform_get_semantic_flag_group_low_level_il(sem_group_name, + return self.get_semantic_flag_group_low_level_il(sem_group_name, lowlevelil.LowLevelILFunction(self, core.BNNewLowLevelILFunctionReference(il))).index except OSError: log.log_error(traceback.format_exc()) @@ -1068,7 +873,7 @@ class Architecture(object): result[0].size = 0 result[0].extend = ImplicitRegisterExtend.NoExtend return - info = self.__class__.regs[self._regs_by_index[reg]] + info = self.regs[self._regs_by_index[reg]] result[0].fullWidthRegister = self._all_regs[info.full_width_reg] result[0].offset = info.offset result[0].size = info.size @@ -1085,26 +890,26 @@ class Architecture(object): def _get_stack_pointer_register(self, ctxt): try: - return self._all_regs[self.__class__.stack_pointer] + return self._all_regs[self.stack_pointer] except KeyError: log.log_error(traceback.format_exc()) return 0 def _get_link_register(self, ctxt): try: - if self.__class__.link_reg is None: + if self.link_reg is None: return 0xffffffff - return self._all_regs[self.__class__.link_reg] + return self._all_regs[self.link_reg] except KeyError: log.log_error(traceback.format_exc()) return 0 def _get_global_registers(self, ctxt, count): try: - count[0] = len(self.__class__.global_regs) - reg_buf = (ctypes.c_uint * len(self.__class__.global_regs))() - for i in xrange(0, len(self.__class__.global_regs)): - reg_buf[i] = self._all_regs[self.__class__.global_regs[i]] + count[0] = len(self.global_regs) + reg_buf = (ctypes.c_uint * len(self.global_regs))() + for i in xrange(0, len(self.global_regs)): + reg_buf[i] = self._all_regs[self.global_regs[i]] result = ctypes.cast(reg_buf, ctypes.c_void_p) self._pending_reg_lists[result.value] = (result, reg_buf) return result.value @@ -1146,7 +951,7 @@ class Architecture(object): result[0].topRelativeCount = 0 result[0].stackTopReg = 0 return - info = self.__class__.regs[self._reg_stacks_by_index[reg_stack]] + info = self.reg_stacks[self._reg_stacks_by_index[reg_stack]] result[0].firstStorageReg = self._all_regs[info.storage_regs[0]] result[0].storageCount = len(info.storage_regs) if len(info.top_relative_regs) > 0: @@ -1208,7 +1013,7 @@ class Architecture(object): count[0] = 0 return None - def _free_name_and_type_list(self, ctxt, buf_raw): + def _free_name_and_type_list(self, ctxt, buf_raw, length): try: buf = ctypes.cast(buf_raw, ctypes.c_void_p) if buf.value not in self._pending_name_and_type_lists: @@ -1240,7 +1045,7 @@ class Architecture(object): count[0] = 0 return None - def _free_type_list(self, ctxt, buf_raw): + def _free_type_list(self, ctxt, buf_raw, length): try: buf = ctypes.cast(buf_raw, ctypes.c_void_p) if buf.value not in self._pending_type_lists: @@ -1255,7 +1060,7 @@ class Architecture(object): def _assemble(self, ctxt, code, addr, result, errors): try: - data, error_str = self.perform_assemble(code, addr) + data, error_str = self.assemble(code, addr) errors[0] = core.BNAllocString(str(error_str)) if data is None: return False @@ -1273,7 +1078,7 @@ class Architecture(object): try: buf = ctypes.create_string_buffer(length) ctypes.memmove(buf, data, length) - return self.perform_is_never_branch_patch_available(buf.raw, addr) + return self.is_never_branch_patch_available(buf.raw, addr) except: log.log_error(traceback.format_exc()) return False @@ -1282,7 +1087,7 @@ class Architecture(object): try: buf = ctypes.create_string_buffer(length) ctypes.memmove(buf, data, length) - return self.perform_is_always_branch_patch_available(buf.raw, addr) + return self.is_always_branch_patch_available(buf.raw, addr) except: log.log_error(traceback.format_exc()) return False @@ -1291,7 +1096,7 @@ class Architecture(object): try: buf = ctypes.create_string_buffer(length) ctypes.memmove(buf, data, length) - return self.perform_is_invert_branch_patch_available(buf.raw, addr) + return self.is_invert_branch_patch_available(buf.raw, addr) except: log.log_error(traceback.format_exc()) return False @@ -1300,7 +1105,7 @@ class Architecture(object): try: buf = ctypes.create_string_buffer(length) ctypes.memmove(buf, data, length) - return self.perform_is_skip_and_return_zero_patch_available(buf.raw, addr) + return self.is_skip_and_return_zero_patch_available(buf.raw, addr) except: log.log_error(traceback.format_exc()) return False @@ -1309,7 +1114,7 @@ class Architecture(object): try: buf = ctypes.create_string_buffer(length) ctypes.memmove(buf, data, length) - return self.perform_is_skip_and_return_value_patch_available(buf.raw, addr) + return self.is_skip_and_return_value_patch_available(buf.raw, addr) except: log.log_error(traceback.format_exc()) return False @@ -1318,7 +1123,7 @@ class Architecture(object): try: buf = ctypes.create_string_buffer(length) ctypes.memmove(buf, data, length) - result = self.perform_convert_to_nop(buf.raw, addr) + result = self.convert_to_nop(buf.raw, addr) if result is None: return False result = str(result) @@ -1334,7 +1139,7 @@ class Architecture(object): try: buf = ctypes.create_string_buffer(length) ctypes.memmove(buf, data, length) - result = self.perform_always_branch(buf.raw, addr) + result = self.always_branch(buf.raw, addr) if result is None: return False result = str(result) @@ -1350,7 +1155,7 @@ class Architecture(object): try: buf = ctypes.create_string_buffer(length) ctypes.memmove(buf, data, length) - result = self.perform_invert_branch(buf.raw, addr) + result = self.invert_branch(buf.raw, addr) if result is None: return False result = str(result) @@ -1366,7 +1171,7 @@ class Architecture(object): try: buf = ctypes.create_string_buffer(length) ctypes.memmove(buf, data, length) - result = self.perform_skip_and_return_value(buf.raw, addr, value) + result = self.skip_and_return_value(buf.raw, addr, value) if result is None: return False result = str(result) @@ -1379,27 +1184,15 @@ class Architecture(object): return False def perform_get_associated_arch_by_address(self, addr): + """ + Deprecated method provided for compatibility. Architecture plugins should override ``get_associated_arch_by_address``. + """ return self, addr @abc.abstractmethod def perform_get_instruction_info(self, data, addr): """ - ``perform_get_instruction_info`` implements a method which interpretes the bytes passed in ``data`` as an - :py:Class:`InstructionInfo` object. The InstructionInfo object should have the length of the current instruction. - If the instruction is a branch instruction the method should add a branch of the proper type: - - ===================== =================================================== - BranchType Description - ===================== =================================================== - UnconditionalBranch Branch will always be taken - FalseBranch False branch condition - TrueBranch True branch condition - CallDestination Branch is a call instruction (Branch with Link) - FunctionReturn Branch returns from a function - SystemCall System call instruction - IndirectBranch Branch destination is a memory address or register - UnresolvedBranch Call instruction that isn't - ===================== =================================================== + Deprecated method provided for compatibility. Architecture plugins should override ``get_instruction_info``. :param str data: bytes to decode :param int addr: virtual address of the byte to be decoded @@ -1411,8 +1204,7 @@ class Architecture(object): @abc.abstractmethod def perform_get_instruction_text(self, data, addr): """ - ``perform_get_instruction_text`` implements a method which interpretes the bytes passed in ``data`` as a - list of :py:class:`InstructionTextToken` objects. + Deprecated method provided for compatibility. Architecture plugins should override ``get_instruction_text``. :param str data: bytes to decode :param int addr: virtual address of the byte to be decoded @@ -1424,10 +1216,7 @@ class Architecture(object): @abc.abstractmethod def perform_get_instruction_low_level_il(self, data, addr, il): """ - ``perform_get_instruction_low_level_il`` implements a method to interpret the bytes passed in ``data`` to - low-level IL instructions. The il instructions must be appended to the :py:class:`LowLevelILFunction`. - - .. note:: Architecture subclasses should implement this method. + Deprecated method provided for compatibility. Architecture plugins should override ``get_instruction_low_level_il``. :param str data: bytes to be interpreted as low-level IL instructions :param int addr: virtual address of start of ``data`` @@ -1439,8 +1228,7 @@ class Architecture(object): @abc.abstractmethod def perform_get_flag_write_low_level_il(self, op, size, write_type, flag, operands, il): """ - .. note:: Architecture subclasses should implement this method. - .. warning:: This method should never be called directly. + Deprecated method provided for compatibility. Architecture plugins should override ``get_flag_write_low_level_il``. :param LowLevelILOperation op: :param int size: @@ -1458,8 +1246,7 @@ class Architecture(object): @abc.abstractmethod def perform_get_flag_condition_low_level_il(self, cond, sem_class, il): """ - .. note:: Architecture subclasses should implement this method. - .. warning:: This method should never be called directly. + Deprecated method provided for compatibility. Architecture plugins should override ``get_flag_condition_low_level_il``. :param LowLevelILFlagCondition cond: Flag condition to be computed :param str sem_class: Semantic class to be used (None for default semantics) @@ -1471,8 +1258,7 @@ class Architecture(object): @abc.abstractmethod def perform_get_semantic_flag_group_low_level_il(self, sem_group, il): """ - .. note:: Architecture subclasses should implement this method. - .. warning:: This method should never be called directly. + Deprecated method provided for compatibility. Architecture plugins should override ``get_semantic_flag_group_low_level_il``. :param str sem_group: Semantic group to be computed :param LowLevelILFunction il: LowLevelILFunction object to append LowLevelILExpr objects to @@ -1483,14 +1269,7 @@ class Architecture(object): @abc.abstractmethod def perform_assemble(self, code, addr): """ - ``perform_assemble`` implements a method to convert the string of assembly instructions ``code`` loaded at - virtual address ``addr`` to the byte representation of those instructions. This can be done by simply shelling - out to an assembler like yasm or llvm-mc, since this method isn't performance sensitive. - - .. note:: Architecture subclasses should implement this method. - .. note :: It is important that the assembler used accepts a syntax identical to the one emitted by the \ - disassembler. This will prevent confusing the user. - .. warning:: This method should never be called directly. + Deprecated method provided for compatibility. Architecture plugins should override ``assemble``. :param str code: string representation of the instructions to be assembled :param int addr: virtual address that the instructions will be loaded at @@ -1502,8 +1281,7 @@ class Architecture(object): @abc.abstractmethod def perform_is_never_branch_patch_available(self, data, addr): """ - ``perform_is_never_branch_patch_available`` implements a check to determine if the instruction represented by - the bytes contained in ``data`` at address addr is a branch instruction that can be made to never branch. + Deprecated method provided for compatibility. Architecture plugins should override ``is_never_branch_patch_available``. .. note:: Architecture subclasses should implement this method. .. warning:: This method should never be called directly. @@ -1518,11 +1296,7 @@ class Architecture(object): @abc.abstractmethod def perform_is_always_branch_patch_available(self, data, addr): """ - ``perform_is_always_branch_patch_available`` implements a check to determine if the instruction represented by - the bytes contained in ``data`` at address addr is a conditional branch that can be made unconditional. - - .. note:: Architecture subclasses should implement this method. - .. warning:: This method should never be called directly. + Deprecated method provided for compatibility. Architecture plugins should override ``is_always_branch_patch_available``. :param str data: bytes to be checked :param int addr: the virtual address of the instruction to be patched @@ -1534,11 +1308,7 @@ class Architecture(object): @abc.abstractmethod def perform_is_invert_branch_patch_available(self, data, addr): """ - ``perform_is_invert_branch_patch_available`` implements a check to determine if the instruction represented by - the bytes contained in ``data`` at address addr is a conditional branch which can be inverted. - - .. note:: Architecture subclasses should implement this method. - .. warning:: This method should never be called directly. + Deprecated method provided for compatibility. Architecture plugins should override ``is_invert_branch_patch_available``. :param int addr: the virtual address of the instruction to be patched :return: True if the instruction can be patched, False otherwise @@ -1549,13 +1319,7 @@ class Architecture(object): @abc.abstractmethod def perform_is_skip_and_return_zero_patch_available(self, data, addr): """ - ``perform_is_skip_and_return_zero_patch_available`` implements a check to determine if the instruction represented by - the bytes contained in ``data`` at address addr is a *call-like* instruction which can made into instructions - that are equivilent to "return 0". For example if ``data`` was the x86 instruction ``call eax`` which could be - converted into ``xor eax,eax`` thus this function would return True. - - .. note:: Architecture subclasses should implement this method. - .. warning:: This method should never be called directly. + Deprecated method provided for compatibility. Architecture plugins should override ``is_skip_and_return_zero_patch_available``. :param str data: bytes to be checked :param int addr: the virtual address of the instruction to be patched @@ -1567,13 +1331,7 @@ class Architecture(object): @abc.abstractmethod def perform_is_skip_and_return_value_patch_available(self, data, addr): """ - ``perform_is_skip_and_return_value_patch_available`` implements a check to determine if the instruction represented by - the bytes contained in ``data`` at address addr is a *call-like* instruction which can made into instructions - that are equivilent to "return 0". For example if ``data`` was the x86 instruction ``call 0xdeadbeef`` which could be - converted into ``mov eax, 42`` thus this function would return True. - - .. note:: Architecture subclasses should implement this method. - .. warning:: This method should never be called directly. + Deprecated method provided for compatibility. Architecture plugins should override ``is_skip_and_return_value_patch_available``. :param str data: bytes to be checked :param int addr: the virtual address of the instruction to be patched @@ -1585,10 +1343,7 @@ class Architecture(object): @abc.abstractmethod def perform_convert_to_nop(self, data, addr): """ - ``perform_convert_to_nop`` implements a method which returns a nop sequence of len(data) bytes long. - - .. note:: Architecture subclasses should implement this method. - .. warning:: This method should never be called directly. + Deprecated method provided for compatibility. Architecture plugins should override ``convert_to_nop``. :param str data: bytes at virtual address ``addr`` :param int addr: the virtual address of the instruction to be patched @@ -1600,11 +1355,7 @@ class Architecture(object): @abc.abstractmethod def perform_always_branch(self, data, addr): """ - ``perform_always_branch`` implements a method which converts the branch represented by the bytes in ``data`` to - at ``addr`` to an unconditional branch. - - .. note:: Architecture subclasses should implement this method. - .. warning:: This method should never be called directly. + Deprecated method provided for compatibility. Architecture plugins should override ``always_branch``. :param str data: bytes to be checked :param int addr: the virtual address of the instruction to be patched @@ -1616,11 +1367,7 @@ class Architecture(object): @abc.abstractmethod def perform_invert_branch(self, data, addr): """ - ``perform_invert_branch`` implements a method which inverts the branch represented by the bytes in ``data`` to - at ``addr``. - - .. note:: Architecture subclasses should implement this method. - .. warning:: This method should never be called directly. + Deprecated method provided for compatibility. Architecture plugins should override ``invert_branch``. :param str data: bytes to be checked :param int addr: the virtual address of the instruction to be patched @@ -1632,12 +1379,7 @@ class Architecture(object): @abc.abstractmethod def perform_skip_and_return_value(self, data, addr, value): """ - ``perform_skip_and_return_value`` implements a method which converts a *call-like* instruction represented by - the bytes in ``data`` at ``addr`` to one or more instructions that are equivilent to a function returning a - value. - - .. note:: Architecture subclasses should implement this method. - .. warning:: This method should never be called directly. + Deprecated method provided for compatibility. Architecture plugins should override ``skip_and_return_value``. :param str data: bytes to be checked :param int addr: the virtual address of the instruction to be patched @@ -1647,76 +1389,70 @@ class Architecture(object): """ return None + def perform_get_flag_role(self, flag, sem_class): + """ + Deprecated method provided for compatibility. Architecture plugins should override ``get_flag_role``. + """ + if flag in self._flag_roles: + return self._flag_roles[flag] + return FlagRole.SpecialFlagRole + + def perform_get_flags_required_for_flag_condition(self, cond, sem_class): + """ + Deprecated method provided for compatibility. Architecture plugins should override ``get_flags_required_for_flag_condition``. + """ + if cond in self.flags_required_for_flag_condition: + return self.flags_required_for_flag_condition[cond] + return [] + def get_associated_arch_by_address(self, addr): - new_addr = ctypes.c_ulonglong() - new_addr.value = addr - result = core.BNGetAssociatedArchitectureByAddress(self.handle, new_addr) - return Architecture(handle = result), new_addr.value + return self.perform_get_associated_arch_by_address(addr) def get_instruction_info(self, data, addr): """ ``get_instruction_info`` returns an InstructionInfo object for the instruction at the given virtual address ``addr`` with data ``data``. + .. note:: Architecture subclasses should implement this method. + .. note :: The instruction info object should always set the InstructionInfo.length to the instruction length, \ and the branches of the proper types shoulde be added if the instruction is a branch. + If the instruction is a branch instruction architecture plugins should add a branch of the proper type: + + ===================== =================================================== + BranchType Description + ===================== =================================================== + UnconditionalBranch Branch will always be taken + FalseBranch False branch condition + TrueBranch True branch condition + CallDestination Branch is a call instruction (Branch with Link) + FunctionReturn Branch returns from a function + SystemCall System call instruction + IndirectBranch Branch destination is a memory address or register + UnresolvedBranch Branch destination is an unknown address + ===================== =================================================== + :param str data: max_instruction_length bytes from the binary at virtual address ``addr`` :param int addr: virtual address of bytes in ``data`` :return: the InstructionInfo for the current instruction :rtype: InstructionInfo """ - info = core.BNInstructionInfo() - data = str(data) - buf = (ctypes.c_ubyte * len(data))() - ctypes.memmove(buf, data, len(data)) - if not core.BNGetInstructionInfo(self.handle, buf, addr, len(data), info): - return None - result = function.InstructionInfo() - result.length = info.length - result.arch_transition_by_target_addr = info.archTransitionByTargetAddr - result.branch_delay = info.branchDelay - for i in xrange(0, info.branchCount): - target = info.branchTarget[i] - if info.branchArch[i]: - arch = Architecture(info.branchArch[i]) - else: - arch = None - result.add_branch(BranchType(info.branchType[i]), target, arch) - return result + return self.perform_get_instruction_info(data, addr) def get_instruction_text(self, data, addr): """ ``get_instruction_text`` returns a list of InstructionTextToken objects for the instruction at the given virtual address ``addr`` with data ``data``. + .. note:: Architecture subclasses should implement this method. + :param str data: max_instruction_length bytes from the binary at virtual address ``addr`` :param int addr: virtual address of bytes in ``data`` :return: an InstructionTextToken list for the current instruction :rtype: list(InstructionTextToken) """ - data = str(data) - count = ctypes.c_ulonglong() - length = ctypes.c_ulonglong() - length.value = len(data) - buf = (ctypes.c_ubyte * len(data))() - ctypes.memmove(buf, data, len(data)) - tokens = ctypes.POINTER(core.BNInstructionTextToken)() - if not core.BNGetInstructionText(self.handle, buf, addr, length, tokens, count): - return None, 0 - result = [] - for i in xrange(0, count.value): - token_type = InstructionTextTokenType(tokens[i].type) - text = tokens[i].text - value = tokens[i].value - size = tokens[i].size - operand = tokens[i].operand - context = tokens[i].context - confidence = tokens[i].confidence - address = tokens[i].address - result.append(function.InstructionTextToken(token_type, text, value, size, operand, context, address, confidence)) - core.BNFreeInstructionText(tokens, count.value) - return result, length.value + return self.perform_get_instruction_text(data, addr) def get_instruction_low_level_il_instruction(self, bv, addr): il = lowlevelil.LowLevelILFunction(self) @@ -1732,19 +1468,15 @@ class Architecture(object): This is used to analyze arbitrary data at an address, if you are working with an existing binary, you likely want to be using ``Function.get_low_level_il_at``. + .. note:: Architecture subclasses should implement this method. + :param str data: max_instruction_length bytes from the binary at virtual address ``addr`` :param int addr: virtual address of bytes in ``data`` :param LowLevelILFunction il: The function the current instruction belongs to :return: the length of the current instruction :rtype: int """ - data = str(data) - length = ctypes.c_ulonglong() - length.value = len(data) - buf = (ctypes.c_ubyte * len(data))() - ctypes.memmove(buf, data, len(data)) - core.BNGetInstructionLowLevelIL(self.handle, buf, addr, length, il.handle) - return length.value + return self.perform_get_instruction_low_level_il(data, addr, il) def get_low_level_il_from_bytes(self, data, addr): """ @@ -1944,9 +1676,7 @@ class Architecture(object): :return: flag role :rtype: FlagRole """ - flag = self.get_flag_index(flag) - sem_class = self.get_semantic_flag_class_index(sem_class) - return FlagRole(core.BNGetArchitectureFlagRole(self.handle, flag, sem_class)) + return self.perform_get_flag_role(flag, sem_class) def get_flag_write_low_level_il(self, op, size, write_type, flag, operands, il): """ @@ -1958,20 +1688,7 @@ class Architecture(object): :param LowLevelILFunction il: :rtype: LowLevelILExpr """ - flag = self.get_flag_index(flag) - operand_list = (core.BNRegisterOrConstant * len(operands))() - for i in xrange(len(operands)): - if isinstance(operands[i], str): - operand_list[i].constant = False - operand_list[i].reg = self.regs[operands[i]].index - elif isinstance(operands[i], lowlevelil.ILRegister): - operand_list[i].constant = False - operand_list[i].reg = operands[i].index - else: - operand_list[i].constant = True - operand_list[i].value = operands[i] - return lowlevelil.LowLevelILExpr(core.BNGetArchitectureFlagWriteLowLevelIL(self.handle, op, size, - self._flag_write_types[write_type], flag, operand_list, len(operand_list), il.handle)) + return self.perform_get_flag_write_low_level_il(op, size, write_type, flag, operands, il) def get_default_flag_write_low_level_il(self, op, size, role, operands, il): """ @@ -1997,13 +1714,14 @@ class Architecture(object): return lowlevelil.LowLevelILExpr(core.BNGetDefaultArchitectureFlagWriteLowLevelIL(self.handle, op, size, role, operand_list, len(operand_list), il.handle)) - def get_flag_condition_low_level_il(self, cond, il): + def get_flag_condition_low_level_il(self, cond, sem_class, il): """ - :param LowLevelILFlagCondition cond: - :param LowLevelILFunction il: + :param LowLevelILFlagCondition cond: Flag condition to be computed + :param str sem_class: Semantic class to be used (None for default semantics) + :param LowLevelILFunction il: LowLevelILFunction object to append LowLevelILExpr objects to :rtype: LowLevelILExpr """ - return lowlevelil.LowLevelILExpr(core.BNGetArchitectureFlagConditionLowLevelIL(self.handle, cond, il.handle)) + return self.perform_get_flag_condition_low_level_il(cond, sem_class, il) def get_default_flag_condition_low_level_il(self, cond, sem_class, il): """ @@ -2021,18 +1739,10 @@ class Architecture(object): :param LowLevelILFunction il: :rtype: LowLevelILExpr """ - group_index = self.get_semantic_flag_group_index(sem_group) - return lowlevelil.LowLevelILExpr(core.BNGetArchitectureSemanticFlagGroupLowLevelIL(self.handle, group_index, il.handle)) + return self.perform_get_semantic_flag_group_low_level_il(sem_group, il) def get_flags_required_for_flag_condition(self, cond, sem_class = None): - sem_class = self.get_semantic_flag_class_index(sem_class) - count = ctypes.c_ulonglong() - flags = core.BNGetArchitectureFlagsRequiredForFlagCondition(self.handle, cond, sem_class, count) - flag_names = [] - for i in xrange(0, count.value): - flag_names.append(self._flags_by_index[flags[i]]) - core.BNFreeRegisterList(flags) - return flag_names + return self.perform_get_flags_required_for_flag_condition(cond, sem_class) def get_modified_regs_on_write(self, reg): """ @@ -2056,6 +1766,14 @@ class Architecture(object): ``assemble`` converts the string of assembly instructions ``code`` loaded at virtual address ``addr`` to the byte representation of those instructions. + .. note:: Architecture subclasses should implement this method. + + Architecture plugins can override this method to provide assembler functionality. This can be done by + simply shelling out to an assembler like yasm or llvm-mc, since this method isn't performance sensitive. + + .. note :: It is important that the assembler used accepts a syntax identical to the one emitted by the \ + disassembler. This will prevent confusing the user. + :param str code: string representation of the instructions to be assembled :param int addr: virtual address that the instructions will be loaded at :return: the bytes for the assembled instructions or error string @@ -2066,16 +1784,14 @@ class Architecture(object): ('\\x0f\\x84\\x04\\x00\\x00\\x00', '') >>> """ - result = databuffer.DataBuffer() - errors = ctypes.c_char_p() - if not core.BNAssemble(self.handle, code, addr, result.handle, errors): - return None, errors.value - return str(result), errors.value + return self.perform_assemble(code, addr) def is_never_branch_patch_available(self, data, addr): """ ``is_never_branch_patch_available`` determines if the instruction ``data`` at ``addr`` can be made to **never branch**. + .. note:: Architecture subclasses should implement this method. + :param str data: bytes for the instruction to be checked :param int addr: the virtual address of the instruction to be patched :return: True if the instruction can be patched, False otherwise @@ -2088,16 +1804,15 @@ class Architecture(object): False >>> """ - data = str(data) - buf = (ctypes.c_ubyte * len(data))() - ctypes.memmove(buf, data, len(data)) - return core.BNIsArchitectureNeverBranchPatchAvailable(self.handle, buf, addr, len(data)) + return self.perform_is_never_branch_patch_available(data, addr) def is_always_branch_patch_available(self, data, addr): """ ``is_always_branch_patch_available`` determines if the instruction ``data`` at ``addr`` can be made to **always branch**. + .. note:: Architecture subclasses should implement this method. + :param str data: bytes for the instruction to be checked :param int addr: the virtual address of the instruction to be patched :return: True if the instruction can be patched, False otherwise @@ -2110,15 +1825,14 @@ class Architecture(object): False >>> """ - data = str(data) - buf = (ctypes.c_ubyte * len(data))() - ctypes.memmove(buf, data, len(data)) - return core.BNIsArchitectureAlwaysBranchPatchAvailable(self.handle, buf, addr, len(data)) + return self.perform_is_always_branch_patch_available(data, addr) def is_invert_branch_patch_available(self, data, addr): """ ``is_always_branch_patch_available`` determines if the instruction ``data`` at ``addr`` can be inverted. + .. note:: Architecture subclasses should implement this method. + :param str data: bytes for the instruction to be checked :param int addr: the virtual address of the instruction to be patched :return: True if the instruction can be patched, False otherwise @@ -2131,16 +1845,15 @@ class Architecture(object): False >>> """ - data = str(data) - buf = (ctypes.c_ubyte * len(data))() - ctypes.memmove(buf, data, len(data)) - return core.BNIsArchitectureInvertBranchPatchAvailable(self.handle, buf, addr, len(data)) + return self.perform_is_invert_branch_patch_available(data, addr) def is_skip_and_return_zero_patch_available(self, data, addr): """ ``is_skip_and_return_zero_patch_available`` determines if the instruction ``data`` at ``addr`` is a *call-like* instruction that can be made into an instruction *returns zero*. + .. note:: Architecture subclasses should implement this method. + :param str data: bytes for the instruction to be checked :param int addr: the virtual address of the instruction to be patched :return: True if the instruction can be patched, False otherwise @@ -2155,39 +1868,37 @@ class Architecture(object): False >>> """ - data = str(data) - buf = (ctypes.c_ubyte * len(data))() - ctypes.memmove(buf, data, len(data)) - return core.BNIsArchitectureSkipAndReturnZeroPatchAvailable(self.handle, buf, addr, len(data)) + return self.perform_is_skip_and_return_zero_patch_available(data, addr) def is_skip_and_return_value_patch_available(self, data, addr): """ - ``is_skip_and_return_zero_patch_available`` determines if the instruction ``data`` at ``addr`` is a *call-like* + ``is_skip_and_return_value_patch_available`` determines if the instruction ``data`` at ``addr`` is a *call-like* instruction that can be made into an instruction *returns a value*. + .. note:: Architecture subclasses should implement this method. + :param str data: bytes for the instruction to be checked :param int addr: the virtual address of the instruction to be patched :return: True if the instruction can be patched, False otherwise :rtype: bool :Example: - >>> arch.is_skip_and_return_zero_patch_available(arch.assemble("call 0")[0], 0) + >>> arch.is_skip_and_return_value_patch_available(arch.assemble("call 0")[0], 0) True - >>> arch.is_skip_and_return_zero_patch_available(arch.assemble("jmp eax")[0], 0) + >>> arch.is_skip_and_return_value_patch_available(arch.assemble("jmp eax")[0], 0) False >>> """ - data = str(data) - buf = (ctypes.c_ubyte * len(data))() - ctypes.memmove(buf, data, len(data)) - return core.BNIsArchitectureSkipAndReturnValuePatchAvailable(self.handle, buf, addr, len(data)) + return self.perform_is_skip_and_return_value_patch_available(data, addr) def convert_to_nop(self, data, addr): """ ``convert_to_nop`` reads the instruction(s) in ``data`` at virtual address ``addr`` and returns a string of nop instructions of the same length as data. - :param str data: bytes for the instruction to be converted + .. note:: Architecture subclasses should implement this method. + + :param str data: bytes for the instruction to be converted :param int addr: the virtual address of the instruction to be patched :return: string containing len(data) worth of no-operation instructions :rtype: str @@ -2197,20 +1908,15 @@ class Architecture(object): '\\x90\\x90' >>> """ - data = str(data) - buf = (ctypes.c_ubyte * len(data))() - ctypes.memmove(buf, data, len(data)) - if not core.BNArchitectureConvertToNop(self.handle, buf, addr, len(data)): - return None - result = ctypes.create_string_buffer(len(data)) - ctypes.memmove(result, buf, len(data)) - return result.raw + return self.perform_convert_to_nop(data, addr) def always_branch(self, data, addr): """ ``always_branch`` reads the instruction(s) in ``data`` at virtual address ``addr`` and returns a string of bytes of the same length which always branches. + .. note:: Architecture subclasses should implement this method. + :param str data: bytes for the instruction to be converted :param int addr: the virtual address of the instruction to be patched :return: string containing len(data) which always branches to the same location as the provided instruction @@ -2224,20 +1930,15 @@ class Architecture(object): (['jmp ', '0x9'], 5L) >>> """ - data = str(data) - buf = (ctypes.c_ubyte * len(data))() - ctypes.memmove(buf, data, len(data)) - if not core.BNArchitectureAlwaysBranch(self.handle, buf, addr, len(data)): - return None - result = ctypes.create_string_buffer(len(data)) - ctypes.memmove(result, buf, len(data)) - return result.raw + return self.perform_always_branch(data, addr) def invert_branch(self, data, addr): """ ``invert_branch`` reads the instruction(s) in ``data`` at virtual address ``addr`` and returns a string of bytes of the same length which inverts the branch of provided instruction. + .. note:: Architecture subclasses should implement this method. + :param str data: bytes for the instruction to be converted :param int addr: the virtual address of the instruction to be patched :return: string containing len(data) which always branches to the same location as the provided instruction @@ -2252,20 +1953,15 @@ class Architecture(object): (['jl ', '0xa'], 6L) >>> """ - data = str(data) - buf = (ctypes.c_ubyte * len(data))() - ctypes.memmove(buf, data, len(data)) - if not core.BNArchitectureInvertBranch(self.handle, buf, addr, len(data)): - return None - result = ctypes.create_string_buffer(len(data)) - ctypes.memmove(result, buf, len(data)) - return result.raw + return self.perform_invert_branch(data, addr) def skip_and_return_value(self, data, addr, value): """ ``skip_and_return_value`` reads the instruction(s) in ``data`` at virtual address ``addr`` and returns a string of bytes of the same length which doesn't call and instead *return a value*. + .. note:: Architecture subclasses should implement this method. + :param str data: bytes for the instruction to be converted :param int addr: the virtual address of the instruction to be patched :return: string containing len(data) which always branches to the same location as the provided instruction @@ -2276,14 +1972,7 @@ class Architecture(object): (['mov ', 'eax', ', ', '0x0'], 5L) >>> """ - data = str(data) - buf = (ctypes.c_ubyte * len(data))() - ctypes.memmove(buf, data, len(data)) - if not core.BNArchitectureSkipAndReturnValue(self.handle, buf, addr, len(data), value): - return None - result = ctypes.create_string_buffer(len(data)) - ctypes.memmove(result, buf, len(data)) - return result.raw + return self.perform_skip_and_return_value(data, addr, value) def is_view_type_constant_defined(self, type_name, const_name): """ @@ -2348,6 +2037,656 @@ class Architecture(object): core.BNRegisterCallingConvention(self.handle, cc.handle) +class CoreArchitecture(Architecture): + def __init__(self, handle): + super(CoreArchitecture, self).__init__() + + self.handle = core.handle_of_type(handle, core.BNArchitecture) + self.__dict__["name"] = core.BNGetArchitectureName(self.handle) + self.__dict__["endianness"] = Endianness(core.BNGetArchitectureEndianness(self.handle)) + self.__dict__["address_size"] = core.BNGetArchitectureAddressSize(self.handle) + self.__dict__["default_int_size"] = core.BNGetArchitectureDefaultIntegerSize(self.handle) + self.__dict__["instr_alignment"] = core.BNGetArchitectureInstructionAlignment(self.handle) + self.__dict__["max_instr_length"] = core.BNGetArchitectureMaxInstructionLength(self.handle) + self.__dict__["opcode_display_length"] = core.BNGetArchitectureOpcodeDisplayLength(self.handle) + self.__dict__["stack_pointer"] = core.BNGetArchitectureRegisterName(self.handle, + core.BNGetArchitectureStackPointerRegister(self.handle)) + + link_reg = core.BNGetArchitectureLinkRegister(self.handle) + if link_reg == 0xffffffff: + self.__dict__["link_reg"] = None + else: + self.__dict__["link_reg"] = core.BNGetArchitectureRegisterName(self.handle, link_reg) + + count = ctypes.c_ulonglong() + regs = core.BNGetAllArchitectureRegisters(self.handle, count) + self._all_regs = {} + self._regs_by_index = {} + self._full_width_regs = {} + self.__dict__["regs"] = {} + for i in xrange(0, count.value): + name = core.BNGetArchitectureRegisterName(self.handle, regs[i]) + info = core.BNGetArchitectureRegisterInfo(self.handle, regs[i]) + full_width_reg = core.BNGetArchitectureRegisterName(self.handle, info.fullWidthRegister) + self.regs[name] = function.RegisterInfo(full_width_reg, info.size, info.offset, + ImplicitRegisterExtend(info.extend), regs[i]) + self._all_regs[name] = regs[i] + self._regs_by_index[regs[i]] = name + for i in xrange(0, count.value): + info = core.BNGetArchitectureRegisterInfo(self.handle, regs[i]) + full_width_reg = core.BNGetArchitectureRegisterName(self.handle, info.fullWidthRegister) + if full_width_reg not in self._full_width_regs: + self._full_width_regs[full_width_reg] = self._all_regs[full_width_reg] + core.BNFreeRegisterList(regs) + + count = ctypes.c_ulonglong() + flags = core.BNGetAllArchitectureFlags(self.handle, count) + self._flags = {} + self._flags_by_index = {} + self.__dict__["flags"] = [] + for i in xrange(0, count.value): + name = core.BNGetArchitectureFlagName(self.handle, flags[i]) + self._flags[name] = flags[i] + self._flags_by_index[flags[i]] = name + self.flags.append(name) + core.BNFreeRegisterList(flags) + + count = ctypes.c_ulonglong() + write_types = core.BNGetAllArchitectureFlagWriteTypes(self.handle, count) + self._flag_write_types = {} + self._flag_write_types_by_index = {} + self.__dict__["flag_write_types"] = [] + for i in xrange(0, count.value): + name = core.BNGetArchitectureFlagWriteTypeName(self.handle, write_types[i]) + self._flag_write_types[name] = write_types[i] + self._flag_write_types_by_index[write_types[i]] = name + self.flag_write_types.append(name) + core.BNFreeRegisterList(write_types) + + count = ctypes.c_ulonglong() + sem_classes = core.BNGetAllArchitectureSemanticFlagClasses(self.handle, count) + self._semantic_flag_classes = {} + self._semantic_flag_classes_by_index = {} + self.__dict__["semantic_flag_classes"] = [] + for i in xrange(0, count.value): + name = core.BNGetArchitectureSemanticFlagClassName(self.handle, sem_classes[i]) + self._semantic_flag_classes[name] = sem_classes[i] + self._semantic_flag_classes_by_index[sem_classes[i]] = name + self.semantic_flag_classes.append(name) + core.BNFreeRegisterList(sem_classes) + + count = ctypes.c_ulonglong() + sem_groups = core.BNGetAllArchitectureSemanticFlagGroups(self.handle, count) + self._semantic_flag_groups = {} + self._semantic_flag_groups_by_index = {} + self.__dict__["semantic_flag_groups"] = [] + for i in xrange(0, count.value): + name = core.BNGetArchitectureSemanticFlagGroupName(self.handle, sem_groups[i]) + self._semantic_flag_groups[name] = sem_groups[i] + self._semantic_flag_groups_by_index[sem_groups[i]] = name + self.semantic_flag_groups.append(name) + core.BNFreeRegisterList(sem_groups) + + self._flag_roles = {} + self.__dict__["flag_roles"] = {} + for flag in self.__dict__["flags"]: + role = FlagRole(core.BNGetArchitectureFlagRole(self.handle, self._flags[flag], 0)) + self.__dict__["flag_roles"][flag] = role + self._flag_roles[self._flags[flag]] = role + + self.__dict__["flags_required_for_flag_condition"] = {} + for cond in LowLevelILFlagCondition: + count = ctypes.c_ulonglong() + flags = core.BNGetArchitectureFlagsRequiredForFlagCondition(self.handle, cond, 0, count) + flag_names = [] + for i in xrange(0, count.value): + flag_names.append(self._flags_by_index[flags[i]]) + core.BNFreeRegisterList(flags) + self.__dict__["flags_required_for_flag_condition"][cond] = flag_names + + self._flags_required_by_semantic_flag_group = {} + self.__dict__["flags_required_for_semantic_flag_group"] = {} + for group in self.semantic_flag_groups: + count = ctypes.c_ulonglong() + flags = core.BNGetArchitectureFlagsRequiredForSemanticFlagGroup(self.handle, + self._semantic_flag_groups[group], count) + flag_indexes = [] + flag_names = [] + for i in xrange(0, count.value): + flag_indexes.append(flags[i]) + flag_names.append(self._flags_by_index[flags[i]]) + core.BNFreeRegisterList(flags) + self._flags_required_by_semantic_flag_group[self._semantic_flag_groups[group]] = flag_indexes + self.__dict__["flags_required_for_semantic_flag_group"][cond] = flag_names + + self._flag_conditions_for_semantic_flag_group = {} + self.__dict__["flag_conditions_for_semantic_flag_group"] = {} + for group in self.semantic_flag_groups: + count = ctypes.c_ulonglong() + conditions = core.BNGetArchitectureFlagConditionsForSemanticFlagGroup(self.handle, + self._semantic_flag_groups[group], count) + class_index_cond = {} + class_cond = {} + for i in xrange(0, count.value): + class_index_cond[conditions[i].semanticClass] = conditions[i].condition + if conditions[i].semanticClass == 0: + class_cond[None] = conditions[i].condition + elif conditions[i].semanticClass in self._semantic_flag_classes_by_index: + class_cond[self._semantic_flag_classes_by_index[conditions[i].semanticClass]] = conditions[i].condition + core.BNFreeFlagConditionsForSemanticFlagGroup(conditions) + self._flag_conditions_for_semantic_flag_group[self._semantic_flag_groups[group]] = class_index_cond + self.__dict__["flag_conditions_for_semantic_flag_group"][group] = class_cond + + self._flags_written_by_flag_write_type = {} + self.__dict__["flags_written_by_flag_write_type"] = {} + for write_type in self.flag_write_types: + count = ctypes.c_ulonglong() + flags = core.BNGetArchitectureFlagsWrittenByFlagWriteType(self.handle, + self._flag_write_types[write_type], count) + flag_indexes = [] + flag_names = [] + for i in xrange(0, count.value): + flag_indexes.append(flags[i]) + flag_names.append(self._flags_by_index[flags[i]]) + core.BNFreeRegisterList(flags) + self._flags_written_by_flag_write_type[self._flag_write_types[write_type]] = flag_indexes + self.__dict__["flags_written_by_flag_write_type"][write_type] = flag_names + + self._semantic_class_for_flag_write_type = {} + self.__dict__["semantic_class_for_flag_write_type"] = {} + for write_type in self.flag_write_types: + sem_class = core.BNGetArchitectureSemanticClassForFlagWriteType(self.handle, + self._flag_write_types[write_type]) + if sem_class == 0: + sem_class_name = None + else: + sem_class_name = self._semantic_flag_classes_by_index[sem_class] + self._semantic_class_for_flag_write_type[self._flag_write_types[write_type]] = sem_class + self.__dict__["semantic_class_for_flag_write_type"][write_type] = sem_class_name + + count = ctypes.c_ulonglong() + regs = core.BNGetArchitectureGlobalRegisters(self.handle, count) + self.__dict__["global_regs"] = [] + for i in xrange(0, count.value): + self.global_regs.append(core.BNGetArchitectureRegisterName(self.handle, regs[i])) + core.BNFreeRegisterList(regs) + + count = ctypes.c_ulonglong() + regs = core.BNGetAllArchitectureRegisterStacks(self.handle, count) + self._all_reg_stacks = {} + self._reg_stacks_by_index = {} + self.__dict__["reg_stacks"] = {} + for i in xrange(0, count.value): + name = core.BNGetArchitectureRegisterStackName(self.handle, regs[i]) + info = core.BNGetArchitectureRegisterStackInfo(self.handle, regs[i]) + storage = [] + for j in xrange(0, info.storageCount): + storage.append(core.BNGetArchitectureRegisterName(self.handle, info.firstStorageReg + j)) + top_rel = [] + for j in xrange(0, info.topRelativeCount): + top_rel.append(core.BNGetArchitectureRegisterName(self.handle, info.firstTopRelativeReg + j)) + top = core.BNGetArchitectureRegisterName(self.handle, info.stackTopReg) + self.reg_stacks[name] = function.RegisterStackInfo(storage, top_rel, top, regs[i]) + self._all_reg_stacks[name] = regs[i] + self._reg_stacks_by_index[regs[i]] = name + core.BNFreeRegisterList(regs) + + count = ctypes.c_ulonglong() + intrinsics = core.BNGetAllArchitectureIntrinsics(self.handle, count) + self._intrinsics = {} + self._intrinsics_by_index = {} + self.__dict__["intrinsics"] = {} + for i in xrange(0, count.value): + name = core.BNGetArchitectureIntrinsicName(self.handle, intrinsics[i]) + input_count = ctypes.c_ulonglong() + inputs = core.BNGetArchitectureIntrinsicInputs(self.handle, intrinsics[i], input_count) + input_list = [] + for j in xrange(0, input_count.value): + input_name = inputs[j].name + type_obj = types.Type(core.BNNewTypeReference(inputs[j].type), confidence = inputs[j].typeConfidence) + input_list.append(function.IntrinsicInput(type_obj, input_name)) + core.BNFreeNameAndTypeList(inputs, input_count.value) + output_count = ctypes.c_ulonglong() + outputs = core.BNGetArchitectureIntrinsicOutputs(self.handle, intrinsics[i], output_count) + output_list = [] + for j in xrange(0, output_count.value): + output_list.append(types.Type(core.BNNewTypeReference(outputs[j].type), confidence = outputs[j].confidence)) + core.BNFreeOutputTypeList(outputs, output_count.value) + self.intrinsics[name] = function.IntrinsicInfo(input_list, output_list) + self._intrinsics[name] = intrinsics[i] + self._intrinsics_by_index[intrinsics[i]] = (name, self.intrinsics[name]) + core.BNFreeRegisterList(intrinsics) + + def get_associated_arch_by_address(self, addr): + new_addr = ctypes.c_ulonglong() + new_addr.value = addr + result = core.BNGetAssociatedArchitectureByAddress(self.handle, new_addr) + return CoreArchitecture(handle = result), new_addr.value + + def get_instruction_info(self, data, addr): + """ + ``get_instruction_info`` returns an InstructionInfo object for the instruction at the given virtual address + ``addr`` with data ``data``. + + .. note :: The instruction info object should always set the InstructionInfo.length to the instruction length, \ + and the branches of the proper types shoulde be added if the instruction is a branch. + + :param str data: max_instruction_length bytes from the binary at virtual address ``addr`` + :param int addr: virtual address of bytes in ``data`` + :return: the InstructionInfo for the current instruction + :rtype: InstructionInfo + """ + info = core.BNInstructionInfo() + data = str(data) + buf = (ctypes.c_ubyte * len(data))() + ctypes.memmove(buf, data, len(data)) + if not core.BNGetInstructionInfo(self.handle, buf, addr, len(data), info): + return None + result = function.InstructionInfo() + result.length = info.length + result.arch_transition_by_target_addr = info.archTransitionByTargetAddr + result.branch_delay = info.branchDelay + for i in xrange(0, info.branchCount): + target = info.branchTarget[i] + if info.branchArch[i]: + arch = CoreArchitecture(info.branchArch[i]) + else: + arch = None + result.add_branch(BranchType(info.branchType[i]), target, arch) + return result + + def get_instruction_text(self, data, addr): + """ + ``get_instruction_text`` returns a list of InstructionTextToken objects for the instruction at the given virtual + address ``addr`` with data ``data``. + + :param str data: max_instruction_length bytes from the binary at virtual address ``addr`` + :param int addr: virtual address of bytes in ``data`` + :return: an InstructionTextToken list for the current instruction + :rtype: list(InstructionTextToken) + """ + data = str(data) + count = ctypes.c_ulonglong() + length = ctypes.c_ulonglong() + length.value = len(data) + buf = (ctypes.c_ubyte * len(data))() + ctypes.memmove(buf, data, len(data)) + tokens = ctypes.POINTER(core.BNInstructionTextToken)() + if not core.BNGetInstructionText(self.handle, buf, addr, length, tokens, count): + return None, 0 + result = [] + for i in xrange(0, count.value): + token_type = InstructionTextTokenType(tokens[i].type) + text = tokens[i].text + value = tokens[i].value + size = tokens[i].size + operand = tokens[i].operand + context = tokens[i].context + confidence = tokens[i].confidence + address = tokens[i].address + result.append(function.InstructionTextToken(token_type, text, value, size, operand, context, address, confidence)) + core.BNFreeInstructionText(tokens, count.value) + return result, length.value + + def get_instruction_low_level_il(self, data, addr, il): + """ + ``get_instruction_low_level_il`` appends LowLevelILExpr objects to ``il`` for the instruction at the given + virtual address ``addr`` with data ``data``. + + This is used to analyze arbitrary data at an address, if you are working with an existing binary, you likely + want to be using ``Function.get_low_level_il_at``. + + :param str data: max_instruction_length bytes from the binary at virtual address ``addr`` + :param int addr: virtual address of bytes in ``data`` + :param LowLevelILFunction il: The function the current instruction belongs to + :return: the length of the current instruction + :rtype: int + """ + data = str(data) + length = ctypes.c_ulonglong() + length.value = len(data) + buf = (ctypes.c_ubyte * len(data))() + ctypes.memmove(buf, data, len(data)) + core.BNGetInstructionLowLevelIL(self.handle, buf, addr, length, il.handle) + return length.value + + def get_flag_write_low_level_il(self, op, size, write_type, flag, operands, il): + """ + :param LowLevelILOperation op: + :param int size: + :param str write_type: + :param list(str or int) operands: a list of either items that are either string register names or constant \ + integer values + :param LowLevelILFunction il: + :rtype: LowLevelILExpr + """ + flag = self.get_flag_index(flag) + operand_list = (core.BNRegisterOrConstant * len(operands))() + for i in xrange(len(operands)): + if isinstance(operands[i], str): + operand_list[i].constant = False + operand_list[i].reg = self.regs[operands[i]].index + elif isinstance(operands[i], lowlevelil.ILRegister): + operand_list[i].constant = False + operand_list[i].reg = operands[i].index + else: + operand_list[i].constant = True + operand_list[i].value = operands[i] + return lowlevelil.LowLevelILExpr(core.BNGetArchitectureFlagWriteLowLevelIL(self.handle, op, size, + self._flag_write_types[write_type], flag, operand_list, len(operand_list), il.handle)) + + def get_flag_condition_low_level_il(self, cond, sem_class, il): + """ + :param LowLevelILFlagCondition cond: Flag condition to be computed + :param str sem_class: Semantic class to be used (None for default semantics) + :param LowLevelILFunction il: LowLevelILFunction object to append LowLevelILExpr objects to + :rtype: LowLevelILExpr + """ + class_index = self.get_semantic_flag_class_index(sem_class) + return lowlevelil.LowLevelILExpr(core.BNGetArchitectureFlagConditionLowLevelIL(self.handle, cond, + class_index, il.handle)) + + def get_semantic_flag_group_low_level_il(self, sem_group, il): + """ + :param str sem_group: + :param LowLevelILFunction il: + :rtype: LowLevelILExpr + """ + group_index = self.get_semantic_flag_group_index(sem_group) + return lowlevelil.LowLevelILExpr(core.BNGetArchitectureSemanticFlagGroupLowLevelIL(self.handle, group_index, il.handle)) + + def assemble(self, code, addr=0): + """ + ``assemble`` converts the string of assembly instructions ``code`` loaded at virtual address ``addr`` to the + byte representation of those instructions. + + :param str code: string representation of the instructions to be assembled + :param int addr: virtual address that the instructions will be loaded at + :return: the bytes for the assembled instructions or error string + :rtype: (a tuple of instructions and empty string) or (or None and error string) + :Example: + + >>> arch.assemble("je 10") + ('\\x0f\\x84\\x04\\x00\\x00\\x00', '') + >>> + """ + result = databuffer.DataBuffer() + errors = ctypes.c_char_p() + if not core.BNAssemble(self.handle, code, addr, result.handle, errors): + return None, errors.value + return str(result), errors.value + + def is_never_branch_patch_available(self, data, addr): + """ + ``is_never_branch_patch_available`` determines if the instruction ``data`` at ``addr`` can be made to **never branch**. + + :param str data: bytes for the instruction to be checked + :param int addr: the virtual address of the instruction to be patched + :return: True if the instruction can be patched, False otherwise + :rtype: bool + :Example: + + >>> arch.is_never_branch_patch_available(arch.assemble("je 10")[0], 0) + True + >>> arch.is_never_branch_patch_available(arch.assemble("nop")[0], 0) + False + >>> + """ + data = str(data) + buf = (ctypes.c_ubyte * len(data))() + ctypes.memmove(buf, data, len(data)) + return core.BNIsArchitectureNeverBranchPatchAvailable(self.handle, buf, addr, len(data)) + + def is_always_branch_patch_available(self, data, addr): + """ + ``is_always_branch_patch_available`` determines if the instruction ``data`` at ``addr`` can be made to + **always branch**. + + :param str data: bytes for the instruction to be checked + :param int addr: the virtual address of the instruction to be patched + :return: True if the instruction can be patched, False otherwise + :rtype: bool + :Example: + + >>> arch.is_always_branch_patch_available(arch.assemble("je 10")[0], 0) + True + >>> arch.is_always_branch_patch_available(arch.assemble("nop")[0], 0) + False + >>> + """ + data = str(data) + buf = (ctypes.c_ubyte * len(data))() + ctypes.memmove(buf, data, len(data)) + return core.BNIsArchitectureAlwaysBranchPatchAvailable(self.handle, buf, addr, len(data)) + + def is_invert_branch_patch_available(self, data, addr): + """ + ``is_always_branch_patch_available`` determines if the instruction ``data`` at ``addr`` can be inverted. + + :param str data: bytes for the instruction to be checked + :param int addr: the virtual address of the instruction to be patched + :return: True if the instruction can be patched, False otherwise + :rtype: bool + :Example: + + >>> arch.is_invert_branch_patch_available(arch.assemble("je 10")[0], 0) + True + >>> arch.is_invert_branch_patch_available(arch.assemble("nop")[0], 0) + False + >>> + """ + data = str(data) + buf = (ctypes.c_ubyte * len(data))() + ctypes.memmove(buf, data, len(data)) + return core.BNIsArchitectureInvertBranchPatchAvailable(self.handle, buf, addr, len(data)) + + def is_skip_and_return_zero_patch_available(self, data, addr): + """ + ``is_skip_and_return_zero_patch_available`` determines if the instruction ``data`` at ``addr`` is a *call-like* + instruction that can be made into an instruction *returns zero*. + + :param str data: bytes for the instruction to be checked + :param int addr: the virtual address of the instruction to be patched + :return: True if the instruction can be patched, False otherwise + :rtype: bool + :Example: + + >>> arch.is_skip_and_return_zero_patch_available(arch.assemble("call 0")[0], 0) + True + >>> arch.is_skip_and_return_zero_patch_available(arch.assemble("call eax")[0], 0) + True + >>> arch.is_skip_and_return_zero_patch_available(arch.assemble("jmp eax")[0], 0) + False + >>> + """ + data = str(data) + buf = (ctypes.c_ubyte * len(data))() + ctypes.memmove(buf, data, len(data)) + return core.BNIsArchitectureSkipAndReturnZeroPatchAvailable(self.handle, buf, addr, len(data)) + + def is_skip_and_return_value_patch_available(self, data, addr): + """ + ``is_skip_and_return_value_patch_available`` determines if the instruction ``data`` at ``addr`` is a *call-like* + instruction that can be made into an instruction *returns a value*. + + :param str data: bytes for the instruction to be checked + :param int addr: the virtual address of the instruction to be patched + :return: True if the instruction can be patched, False otherwise + :rtype: bool + :Example: + + >>> arch.is_skip_and_return_value_patch_available(arch.assemble("call 0")[0], 0) + True + >>> arch.is_skip_and_return_value_patch_available(arch.assemble("jmp eax")[0], 0) + False + >>> + """ + data = str(data) + buf = (ctypes.c_ubyte * len(data))() + ctypes.memmove(buf, data, len(data)) + return core.BNIsArchitectureSkipAndReturnValuePatchAvailable(self.handle, buf, addr, len(data)) + + def convert_to_nop(self, data, addr): + """ + ``convert_to_nop`` reads the instruction(s) in ``data`` at virtual address ``addr`` and returns a string of nop + instructions of the same length as data. + + :param str data: bytes for the instruction to be converted + :param int addr: the virtual address of the instruction to be patched + :return: string containing len(data) worth of no-operation instructions + :rtype: str + :Example: + + >>> arch.convert_to_nop("\\x00\\x00", 0) + '\\x90\\x90' + >>> + """ + data = str(data) + buf = (ctypes.c_ubyte * len(data))() + ctypes.memmove(buf, data, len(data)) + if not core.BNArchitectureConvertToNop(self.handle, buf, addr, len(data)): + return None + result = ctypes.create_string_buffer(len(data)) + ctypes.memmove(result, buf, len(data)) + return result.raw + + def always_branch(self, data, addr): + """ + ``always_branch`` reads the instruction(s) in ``data`` at virtual address ``addr`` and returns a string of bytes + of the same length which always branches. + + :param str data: bytes for the instruction to be converted + :param int addr: the virtual address of the instruction to be patched + :return: string containing len(data) which always branches to the same location as the provided instruction + :rtype: str + :Example: + + >>> bytes = arch.always_branch(arch.assemble("je 10")[0], 0) + >>> arch.get_instruction_text(bytes, 0) + (['nop '], 1L) + >>> arch.get_instruction_text(bytes[1:], 0) + (['jmp ', '0x9'], 5L) + >>> + """ + data = str(data) + buf = (ctypes.c_ubyte * len(data))() + ctypes.memmove(buf, data, len(data)) + if not core.BNArchitectureAlwaysBranch(self.handle, buf, addr, len(data)): + return None + result = ctypes.create_string_buffer(len(data)) + ctypes.memmove(result, buf, len(data)) + return result.raw + + def invert_branch(self, data, addr): + """ + ``invert_branch`` reads the instruction(s) in ``data`` at virtual address ``addr`` and returns a string of bytes + of the same length which inverts the branch of provided instruction. + + :param str data: bytes for the instruction to be converted + :param int addr: the virtual address of the instruction to be patched + :return: string containing len(data) which always branches to the same location as the provided instruction + :rtype: str + :Example: + + >>> arch.get_instruction_text(arch.invert_branch(arch.assemble("je 10")[0], 0), 0) + (['jne ', '0xa'], 6L) + >>> arch.get_instruction_text(arch.invert_branch(arch.assemble("jo 10")[0], 0), 0) + (['jno ', '0xa'], 6L) + >>> arch.get_instruction_text(arch.invert_branch(arch.assemble("jge 10")[0], 0), 0) + (['jl ', '0xa'], 6L) + >>> + """ + data = str(data) + buf = (ctypes.c_ubyte * len(data))() + ctypes.memmove(buf, data, len(data)) + if not core.BNArchitectureInvertBranch(self.handle, buf, addr, len(data)): + return None + result = ctypes.create_string_buffer(len(data)) + ctypes.memmove(result, buf, len(data)) + return result.raw + + def skip_and_return_value(self, data, addr, value): + """ + ``skip_and_return_value`` reads the instruction(s) in ``data`` at virtual address ``addr`` and returns a string of + bytes of the same length which doesn't call and instead *return a value*. + + :param str data: bytes for the instruction to be converted + :param int addr: the virtual address of the instruction to be patched + :return: string containing len(data) which always branches to the same location as the provided instruction + :rtype: str + :Example: + + >>> arch.get_instruction_text(arch.skip_and_return_value(arch.assemble("call 10")[0], 0, 0), 0) + (['mov ', 'eax', ', ', '0x0'], 5L) + >>> + """ + data = str(data) + buf = (ctypes.c_ubyte * len(data))() + ctypes.memmove(buf, data, len(data)) + if not core.BNArchitectureSkipAndReturnValue(self.handle, buf, addr, len(data), value): + return None + result = ctypes.create_string_buffer(len(data)) + ctypes.memmove(result, buf, len(data)) + return result.raw + + def get_flag_role(self, flag, sem_class = None): + """ + ``get_flag_role`` gets the role of a given flag. + + :param int flag: flag + :param int sem_class: optional semantic flag class + :return: flag role + :rtype: FlagRole + """ + flag = self.get_flag_index(flag) + sem_class = self.get_semantic_flag_class_index(sem_class) + return FlagRole(core.BNGetArchitectureFlagRole(self.handle, flag, sem_class)) + + def get_flags_required_for_flag_condition(self, cond, sem_class = None): + sem_class = self.get_semantic_flag_class_index(sem_class) + count = ctypes.c_ulonglong() + flags = core.BNGetArchitectureFlagsRequiredForFlagCondition(self.handle, cond, sem_class, count) + flag_names = [] + for i in xrange(0, count.value): + flag_names.append(self._flags_by_index[flags[i]]) + core.BNFreeRegisterList(flags) + return flag_names + + +class ArchitectureHook(CoreArchitecture): + def __init__(self, base_arch): + self.base_arch = base_arch + super(ArchitectureHook, self).__init__(base_arch.handle) + + # To improve performance of simpler hooks, use null callback for functions that are not being overridden + if self.get_associated_arch_by_address.__code__ == CoreArchitecture.get_associated_arch_by_address.__code__: + self._cb.getAssociatedArchitectureByAddress = self._cb.getAssociatedArchitectureByAddress.__class__() + if self.get_instruction_info.__code__ == CoreArchitecture.get_instruction_info.__code__: + self._cb.getInstructionInfo = self._cb.getInstructionInfo.__class__() + if self.get_instruction_text.__code__ == CoreArchitecture.get_instruction_text.__code__: + self._cb.getInstructionText = self._cb.getInstructionText.__class__() + if self.__class__.stack_pointer is None: + self._cb.getStackPointerRegister = self._cb.getStackPointerRegister.__class__() + if self.__class__.link_reg is None: + self._cb.getLinkRegister = self._cb.getLinkRegister.__class__() + if len(self.__class__.regs) == 0: + self._cb.getRegisterInfo = self._cb.getRegisterInfo.__class__() + self._cb.getRegisterName = self._cb.getRegisterName.__class__() + if len(self.__class__.reg_stacks) == 0: + self._cb.getRegisterStackName = self._cb.getRegisterStackName.__class__() + self._cb.getRegisterStackInfo = self._cb.getRegisterStackInfo.__class__() + if len(self.__class__.intrinsics) == 0: + self._cb.getIntrinsicName = self._cb.getIntrinsicName.__class__() + self._cb.getIntrinsicInputs = self._cb.getIntrinsicInputs.__class__() + self._cb.freeNameAndTypeList = self._cb.freeNameAndTypeList.__class__() + self._cb.getIntrinsicOutputs = self._cb.getIntrinsicOutputs.__class__() + self._cb.freeTypeList = self._cb.freeTypeList.__class__() + + def register(self): + self.__class__._registered_cb = self._cb + self.handle = core.BNRegisterArchitectureHook(self.base_arch.handle, self._cb) + + class ReferenceSource(object): def __init__(self, func, arch, addr): self.function = func diff --git a/python/basicblock.py b/python/basicblock.py index 8e64c1c1..c93d0b2c 100644 --- a/python/basicblock.py +++ b/python/basicblock.py @@ -100,7 +100,7 @@ class BasicBlock(object): arch = core.BNGetBasicBlockArchitecture(self.handle) if arch is None: return None - self._arch = architecture.Architecture(arch) + self._arch = architecture.CoreArchitecture(arch) return self._arch @property diff --git a/python/binaryview.py b/python/binaryview.py index fae98c14..7a6bc875 100644 --- a/python/binaryview.py +++ b/python/binaryview.py @@ -400,7 +400,7 @@ class BinaryViewType(object): arch = core.BNGetArchitectureForViewType(self.handle, ident, endian) if arch is None: return None - return architecture.Architecture(arch) + return architecture.CoreArchitecture(arch) def register_platform(self, ident, arch, plat): core.BNRegisterPlatformForViewType(self.handle, ident, arch.handle, plat.handle) @@ -812,7 +812,7 @@ class BinaryView(object): arch = core.BNGetDefaultArchitecture(self.handle) if arch is None: return None - return architecture.Architecture(handle=arch) + return architecture.CoreArchitecture(handle=arch) @arch.setter def arch(self, value): @@ -2191,7 +2191,7 @@ class BinaryView(object): else: func = None if refs[i].arch: - arch = architecture.Architecture(refs[i].arch) + arch = architecture.CoreArchitecture(refs[i].arch) else: arch = None addr = refs[i].addr diff --git a/python/callingconvention.py b/python/callingconvention.py index 5aad3317..49ef7666 100644 --- a/python/callingconvention.py +++ b/python/callingconvention.py @@ -75,7 +75,7 @@ class CallingConvention(object): self.__class__._registered_calling_conventions.append(self) else: self.handle = handle - self.arch = architecture.Architecture(core.BNGetCallingConventionArchitecture(self.handle)) + self.arch = architecture.CoreArchitecture(core.BNGetCallingConventionArchitecture(self.handle)) self.__dict__["name"] = core.BNGetCallingConventionName(self.handle) self.__dict__["arg_regs_share_index"] = core.BNAreArgumentRegistersSharedIndex(self.handle) self.__dict__["stack_reserved_for_arg_regs"] = core.BNIsStackReservedForArgumentRegisters(self.handle) diff --git a/python/examples/arch_hook.py b/python/examples/arch_hook.py new file mode 100644 index 00000000..452bd2b6 --- /dev/null +++ b/python/examples/arch_hook.py @@ -0,0 +1,16 @@ +from binaryninja.architecture import Architecture, ArchitectureHook + +class X86ReturnHook(ArchitectureHook): + def get_instruction_text(self, data, addr): + # Call the original implementation's method by calling the superclass + result, length = super(X86ReturnHook, self).get_instruction_text(data, addr) + + # Patch the name of the 'retn' instruction to 'ret' + if len(result) > 0 and result[0].text == 'retn': + result[0].text = 'ret' + + return result, length + +# Install the hook by constructing it with the desired architecture to hook, then registering it +X86ReturnHook(Architecture['x86']).register() + diff --git a/python/examples/nes.py b/python/examples/nes.py index 39544c9f..00451abd 100644 --- a/python/examples/nes.py +++ b/python/examples/nes.py @@ -423,7 +423,7 @@ class M6502(Architecture): return instr, operand, length, value - def perform_get_instruction_info(self, data, addr): + def get_instruction_info(self, data, addr): instr, operand, length, value = self.decode_instruction(data, addr) if instr is None: return None @@ -445,7 +445,7 @@ class M6502(Architecture): result.add_branch(BranchType.FalseBranch, addr + 2) return result - def perform_get_instruction_text(self, data, addr): + def get_instruction_text(self, data, addr): instr, operand, length, value = self.decode_instruction(data, addr) if instr is None: return None @@ -455,7 +455,7 @@ class M6502(Architecture): tokens += OperandTokens[operand](value) return tokens, length - def perform_get_instruction_low_level_il(self, data, addr, il): + def get_instruction_low_level_il(self, data, addr, il): instr, operand, length, value = self.decode_instruction(data, addr) if instr is None: return None @@ -470,48 +470,48 @@ class M6502(Architecture): return length - def perform_get_flag_write_low_level_il(self, op, size, write_type, flag, operands, il): + def get_flag_write_low_level_il(self, op, size, write_type, flag, operands, il): if flag == 'c': if (op == LowLevelILOperation.LLIL_SUB) or (op == LowLevelILOperation.LLIL_SBB): # Subtraction carry flag is inverted from the commom implementation return il.not_expr(0, self.get_default_flag_write_low_level_il(op, size, FlagRole.CarryFlagRole, operands, il)) # Other operations use a normal carry flag return self.get_default_flag_write_low_level_il(op, size, FlagRole.CarryFlagRole, operands, il) - return Architecture.perform_get_flag_write_low_level_il(self, op, size, write_type, flag, operands, il) + return Architecture.get_flag_write_low_level_il(self, op, size, write_type, flag, operands, il) - def perform_is_never_branch_patch_available(self, data, addr): + def is_never_branch_patch_available(self, data, addr): if (data[0] == "\x10") or (data[0] == "\x30") or (data[0] == "\x50") or (data[0] == "\x70") or (data[0] == "\x90") or (data[0] == "\xb0") or (data[0] == "\xd0") or (data[0] == "\xf0"): return True return False - def perform_is_invert_branch_patch_available(self, data, addr): + def is_invert_branch_patch_available(self, data, addr): if (data[0] == "\x10") or (data[0] == "\x30") or (data[0] == "\x50") or (data[0] == "\x70") or (data[0] == "\x90") or (data[0] == "\xb0") or (data[0] == "\xd0") or (data[0] == "\xf0"): return True return False - def perform_is_always_branch_patch_available(self, data, addr): + def is_always_branch_patch_available(self, data, addr): return False - def perform_is_skip_and_return_zero_patch_available(self, data, addr): + def is_skip_and_return_zero_patch_available(self, data, addr): return (data[0] == "\x20") and (len(data) == 3) - def perform_is_skip_and_return_value_patch_available(self, data, addr): + def is_skip_and_return_value_patch_available(self, data, addr): return (data[0] == "\x20") and (len(data) == 3) - def perform_convert_to_nop(self, data, addr): + def convert_to_nop(self, data, addr): return "\xea" * len(data) - def perform_never_branch(self, data, addr): + def never_branch(self, data, addr): if (data[0] == "\x10") or (data[0] == "\x30") or (data[0] == "\x50") or (data[0] == "\x70") or (data[0] == "\x90") or (data[0] == "\xb0") or (data[0] == "\xd0") or (data[0] == "\xf0"): return "\xea" * len(data) return None - def perform_invert_branch(self, data, addr): + def invert_branch(self, data, addr): if (data[0] == "\x10") or (data[0] == "\x30") or (data[0] == "\x50") or (data[0] == "\x70") or (data[0] == "\x90") or (data[0] == "\xb0") or (data[0] == "\xd0") or (data[0] == "\xf0"): return chr(ord(data[0]) ^ 0x20) + data[1:] return None - def perform_skip_and_return_value(self, data, addr, value): + def skip_and_return_value(self, data, addr, value): if (data[0] != "\x20") or (len(data) != 3): return None return "\xa9" + chr(value & 0xff) + "\xea" diff --git a/python/function.py b/python/function.py index 39ace7c6..f68aa76f 100644 --- a/python/function.py +++ b/python/function.py @@ -419,7 +419,7 @@ class Function(object): arch = core.BNGetFunctionArchitecture(self.handle) if arch is None: return None - self._arch = architecture.Architecture(arch) + self._arch = architecture.CoreArchitecture(arch) return self._arch @property @@ -557,7 +557,7 @@ class Function(object): branches = core.BNGetIndirectBranches(self.handle, count) result = [] for i in xrange(0, count.value): - result.append(IndirectBranchInfo(architecture.Architecture(branches[i].sourceArch), branches[i].sourceAddr, architecture.Architecture(branches[i].destArch), branches[i].destAddr, branches[i].autoDefined)) + result.append(IndirectBranchInfo(architecture.CoreArchitecture(branches[i].sourceArch), branches[i].sourceAddr, architecture.Architecture(branches[i].destArch), branches[i].destAddr, branches[i].autoDefined)) core.BNFreeIndirectBranchList(branches) return result @@ -1115,7 +1115,7 @@ class Function(object): branches = core.BNGetIndirectBranchesAt(self.handle, arch.handle, addr, count) result = [] for i in xrange(0, count.value): - result.append(IndirectBranchInfo(architecture.Architecture(branches[i].sourceArch), branches[i].sourceAddr, architecture.Architecture(branches[i].destArch), branches[i].destAddr, branches[i].autoDefined)) + result.append(IndirectBranchInfo(architecture.CoreArchitecture(branches[i].sourceArch), branches[i].sourceAddr, architecture.Architecture(branches[i].destArch), branches[i].destAddr, branches[i].autoDefined)) core.BNFreeIndirectBranchList(branches) return result @@ -1633,7 +1633,7 @@ class FunctionGraphBlock(object): arch = core.BNGetFunctionGraphBlockArchitecture(self.handle) if arch is None: return None - return architecture.Architecture(arch) + return architecture.CoreArchitecture(arch) @property def start(self): diff --git a/python/platform.py b/python/platform.py index 5e63d836..09670bac 100644 --- a/python/platform.py +++ b/python/platform.py @@ -105,7 +105,7 @@ class Platform(object): else: self.handle = handle self.__dict__["name"] = core.BNGetPlatformName(self.handle) - self.arch = architecture.Architecture(core.BNGetPlatformArchitecture(self.handle)) + self.arch = architecture.CoreArchitecture(core.BNGetPlatformArchitecture(self.handle)) def __del__(self): core.BNFreePlatform(self.handle) -- cgit v1.3.1 From 5f07bf18df013860c9d9870a4ad0fba78379b0f1 Mon Sep 17 00:00:00 2001 From: Ryan Snyder Date: Tue, 20 Mar 2018 09:53:59 -0400 Subject: Cache all created CoreArchitecture objects --- python/architecture.py | 18 +++++++++++++----- python/basicblock.py | 2 +- python/binaryview.py | 6 +++--- python/callingconvention.py | 2 +- python/function.py | 8 ++++---- python/platform.py | 2 +- 6 files changed, 23 insertions(+), 15 deletions(-) (limited to 'python/callingconvention.py') diff --git a/python/architecture.py b/python/architecture.py index 781edcfd..c5f87ec6 100644 --- a/python/architecture.py +++ b/python/architecture.py @@ -44,7 +44,7 @@ class _ArchitectureMetaClass(type): archs = core.BNGetArchitectureList(count) result = [] for i in xrange(0, count.value): - result.append(CoreArchitecture(archs[i])) + result.append(CoreArchitecture._from_cache(archs[i])) core.BNFreeArchitectureList(archs) return result @@ -54,7 +54,7 @@ class _ArchitectureMetaClass(type): archs = core.BNGetArchitectureList(count) try: for i in xrange(0, count.value): - yield CoreArchitecture(archs[i]) + yield CoreArchitecture._from_cache(archs[i]) finally: core.BNFreeArchitectureList(archs) @@ -63,7 +63,7 @@ class _ArchitectureMetaClass(type): arch = core.BNGetArchitectureByName(name) if arch is None: raise KeyError("'%s' is not a valid architecture" % str(name)) - return CoreArchitecture(arch) + return CoreArchitecture._from_cache(arch) def register(cls): startup._init_plugins() @@ -2037,6 +2037,7 @@ class Architecture(object): core.BNRegisterCallingConvention(self.handle, cc.handle) +_architecture_cache = {} class CoreArchitecture(Architecture): def __init__(self, handle): super(CoreArchitecture, self).__init__() @@ -2256,12 +2257,19 @@ class CoreArchitecture(Architecture): self._intrinsics[name] = intrinsics[i] self._intrinsics_by_index[intrinsics[i]] = (name, self.intrinsics[name]) core.BNFreeRegisterList(intrinsics) + global _architecture_cache + _architecture_cache[ctypes.addressof(handle.contents)] = self + + @classmethod + def _from_cache(cls, handle): + global _architecture_cache + return _architecture_cache.get(ctypes.addressof(handle.contents)) or cls(handle) def get_associated_arch_by_address(self, addr): new_addr = ctypes.c_ulonglong() new_addr.value = addr result = core.BNGetAssociatedArchitectureByAddress(self.handle, new_addr) - return CoreArchitecture(handle = result), new_addr.value + return CoreArchitecture._from_cache(handle = result), new_addr.value def get_instruction_info(self, data, addr): """ @@ -2289,7 +2297,7 @@ class CoreArchitecture(Architecture): for i in xrange(0, info.branchCount): target = info.branchTarget[i] if info.branchArch[i]: - arch = CoreArchitecture(info.branchArch[i]) + arch = CoreArchitecture._from_cache(info.branchArch[i]) else: arch = None result.add_branch(BranchType(info.branchType[i]), target, arch) diff --git a/python/basicblock.py b/python/basicblock.py index c93d0b2c..1d932c5e 100644 --- a/python/basicblock.py +++ b/python/basicblock.py @@ -100,7 +100,7 @@ class BasicBlock(object): arch = core.BNGetBasicBlockArchitecture(self.handle) if arch is None: return None - self._arch = architecture.CoreArchitecture(arch) + self._arch = architecture.CoreArchitecture._from_cache(arch) return self._arch @property diff --git a/python/binaryview.py b/python/binaryview.py index 3af0ee50..0cf25e70 100644 --- a/python/binaryview.py +++ b/python/binaryview.py @@ -411,7 +411,7 @@ class BinaryViewType(object): arch = core.BNGetArchitectureForViewType(self.handle, ident, endian) if arch is None: return None - return architecture.CoreArchitecture(arch) + return architecture.CoreArchitecture._from_cache(arch) def register_platform(self, ident, arch, plat): core.BNRegisterPlatformForViewType(self.handle, ident, arch.handle, plat.handle) @@ -823,7 +823,7 @@ class BinaryView(object): arch = core.BNGetDefaultArchitecture(self.handle) if arch is None: return None - return architecture.CoreArchitecture(handle=arch) + return architecture.CoreArchitecture._from_cache(handle=arch) @arch.setter def arch(self, value): @@ -2211,7 +2211,7 @@ class BinaryView(object): else: func = None if refs[i].arch: - arch = architecture.CoreArchitecture(refs[i].arch) + arch = architecture.CoreArchitecture._from_cache(refs[i].arch) else: arch = None addr = refs[i].addr diff --git a/python/callingconvention.py b/python/callingconvention.py index 49ef7666..e3ec7261 100644 --- a/python/callingconvention.py +++ b/python/callingconvention.py @@ -75,7 +75,7 @@ class CallingConvention(object): self.__class__._registered_calling_conventions.append(self) else: self.handle = handle - self.arch = architecture.CoreArchitecture(core.BNGetCallingConventionArchitecture(self.handle)) + self.arch = architecture.CoreArchitecture._from_cache(core.BNGetCallingConventionArchitecture(self.handle)) self.__dict__["name"] = core.BNGetCallingConventionName(self.handle) self.__dict__["arg_regs_share_index"] = core.BNAreArgumentRegistersSharedIndex(self.handle) self.__dict__["stack_reserved_for_arg_regs"] = core.BNIsStackReservedForArgumentRegisters(self.handle) diff --git a/python/function.py b/python/function.py index 0d796d98..8c4b7765 100644 --- a/python/function.py +++ b/python/function.py @@ -420,7 +420,7 @@ class Function(object): arch = core.BNGetFunctionArchitecture(self.handle) if arch is None: return None - self._arch = architecture.CoreArchitecture(arch) + self._arch = architecture.CoreArchitecture._from_cache(arch) return self._arch @property @@ -558,7 +558,7 @@ class Function(object): branches = core.BNGetIndirectBranches(self.handle, count) result = [] for i in xrange(0, count.value): - result.append(IndirectBranchInfo(architecture.CoreArchitecture(branches[i].sourceArch), branches[i].sourceAddr, architecture.Architecture(branches[i].destArch), branches[i].destAddr, branches[i].autoDefined)) + result.append(IndirectBranchInfo(architecture.CoreArchitecture._from_cache(branches[i].sourceArch), branches[i].sourceAddr, architecture.Architecture(branches[i].destArch), branches[i].destAddr, branches[i].autoDefined)) core.BNFreeIndirectBranchList(branches) return result @@ -1142,7 +1142,7 @@ class Function(object): branches = core.BNGetIndirectBranchesAt(self.handle, arch.handle, addr, count) result = [] for i in xrange(0, count.value): - result.append(IndirectBranchInfo(architecture.CoreArchitecture(branches[i].sourceArch), branches[i].sourceAddr, architecture.Architecture(branches[i].destArch), branches[i].destAddr, branches[i].autoDefined)) + result.append(IndirectBranchInfo(architecture.CoreArchitecture._from_cache(branches[i].sourceArch), branches[i].sourceAddr, architecture.Architecture(branches[i].destArch), branches[i].destAddr, branches[i].autoDefined)) core.BNFreeIndirectBranchList(branches) return result @@ -1660,7 +1660,7 @@ class FunctionGraphBlock(object): arch = core.BNGetFunctionGraphBlockArchitecture(self.handle) if arch is None: return None - return architecture.CoreArchitecture(arch) + return architecture.CoreArchitecture._from_cache(arch) @property def start(self): diff --git a/python/platform.py b/python/platform.py index 09670bac..a79e3b9a 100644 --- a/python/platform.py +++ b/python/platform.py @@ -105,7 +105,7 @@ class Platform(object): else: self.handle = handle self.__dict__["name"] = core.BNGetPlatformName(self.handle) - self.arch = architecture.CoreArchitecture(core.BNGetPlatformArchitecture(self.handle)) + self.arch = architecture.CoreArchitecture._from_cache(core.BNGetPlatformArchitecture(self.handle)) def __del__(self): core.BNFreePlatform(self.handle) -- cgit v1.3.1