address.hh File Reference

Classes for specifying addresses and other low-level constants. More...

#include "space.hh"

Classes
class	Address
	A low-level machine address for labelling bytes and data. More...
class	SeqNum
	A class for uniquely labelling and comparing PcodeOps. More...
class	Range
	A contiguous range of bytes in some address space. More...
class	RangeList
	A disjoint set of Ranges, possibly across multiple address spaces. More...

Functions
uintb	calc_mask (int4 size)
	Calculate a mask for a given byte size. More...
uintb	pcode_right (uintb val, int4 sa)
uintb	pcode_left (uintb val, int4 sa)
uintb	minimalmask (uintb val)
	Calculate smallest mask that covers the given value. More...
bool	signbit_negative (uintb val, int4 size)
	Return true if the sign-bit is set. More...
uintb	uintb_negate (uintb in, int4 size)
	Negate the sized value. More...
uintb	sign_extend (uintb in, int4 sizein, int4 sizeout)
	Sign-extend a value between two byte sizes. More...
void	sign_extend (intb &val, int4 bit)
	Sign extend above given bit. More...
void	zero_extend (intb &val, int4 bit)
	Clear all bits above given bit. More...
void	byte_swap (intb &val, int4 size)
	Swap bytes in the given value. More...
uintb	byte_swap (uintb val, int4 size)
	Return the given value with bytes swapped. More...
int4	leastsigbit_set (uintb val)
	Return index of least significant bit set in given value. More...
int4	mostsigbit_set (uintb val)
	Return index of most significant bit set in given value. More...
int4	popcount (uintb val)
	Return the number of one bits in the given value. More...
int4	count_leading_zeros (uintb val)
	Return the number of leading zero bits in the given value. More...
uintb	coveringmask (uintb val)
	Return a mask that covers the given value. More...
int4	bit_transitions (uintb val, int4 sz)
	Calculate the number of bit transitions in the sized value. More...
void	mult64to128 (uint8 *res, uint8 x, uint8 y)
	Multiply 2 unsigned 64-bit values, producing a 128-bit value. More...
void	unsignedSubtract128 (uint8 *a, uint8 *b)
	Subtract (in-place) a 128-bit value from a base 128-value. More...
int4	unsignedCompare128 (uint8 *a, uint8 *b)
	Compare two unsigned 128-bit values. More...
int4	power2Divide (int4 n, uint8 divisor, uint8 &q, uint8 &r)
	Unsigned division of a power of 2 (upto 2^127) by a 64-bit divisor. More...

Variables
uintb	uintbmasks []
	Precalculated masks indexed by size.

Detailed Description

Classes for specifying addresses and other low-level constants.

All addresses are absolute and there are are no registers in CPUI. However, all addresses are prefixed with an "immutable" pointer, which can specify a separate RAM space, a register space, an i/o space etc. Thus a translation from a real machine language will typically simulate registers by placing them in their own space, separate from RAM. Indirection (i.e. pointers) must be simulated through the LOAD and STORE ops.

Function Documentation

int4 bit_transitions	(	uintb	val,
		int4	sz
	)

Calculate the number of bit transitions in the sized value.

Treat val as a constant of size sz. Scanning across the bits of val return the number of transitions (from 0->1 or 1->0) If there are 2 or less transitions, this is an indication of a bit flag or a mask

Parameters

val	is the given value
sz	is the size to treat the value as

Returns

the number of transitions

Referenced by ActionConstantPtr::isPointer(), minimalmask(), and CircleRange::setNZMask().

void byte_swap	(	intb &	val,
		int4	size
	)

Swap bytes in the given value.

Swap the least significant size bytes in val

Parameters

val	is a reference to the value to swap
size	is the number of bytes to swap

Referenced by MemoryImage::find(), EmulatePcodeOp::getLoadImageValue(), EmulateSnippet::getLoadImageValue(), MemoryBank::getPage(), minimalmask(), and MemoryBank::setPage().

uintb byte_swap	(	uintb	val,
		int4	size
	)

Return the given value with bytes swapped.

Swap the least significant size bytes in val

Parameters

val	is the value to swap
size	is the number of bytes to swap

Returns

the swapped value

uintb calc_mask ( int4  size )

inline

Calculate a mask for a given byte size.

Parameters

size	is the desired size in bytes

Returns

a value appropriate for masking off the first size bytes

References uintbmasks.

Referenced by AddTreeState::AddTreeState(), ActionStackPtrFlow::analyzeExtraPop(), RuleCollectTerms::applyOp(), RuleOrMask::applyOp(), RuleNegateIdentity::applyOp(), RuleShiftBitops::applyOp(), RuleRightShiftAnd::applyOp(), RulePullsubMulti::applyOp(), RulePullsubIndirect::applyOp(), RuleHighOrderAnd::applyOp(), RuleAndDistribute::applyOp(), RuleAndCommute::applyOp(), RuleAndPiece::applyOp(), RuleAndCompare::applyOp(), RuleDoubleShift::applyOp(), RuleShiftCompare::applyOp(), RuleBoolZext::applyOp(), RuleSborrow::applyOp(), RuleSignShift::applyOp(), RuleTestSign::applyOp(), RuleTransformCpool::applyOp(), Rule2Comp2Mult::applyOp(), RuleCarryElim::applyOp(), RuleSub2Add::applyOp(), RuleSubCommute::applyOp(), RuleConcatCommute::applyOp(), RuleShiftAnd::applyOp(), RuleSubZext::applyOp(), RuleHumptyOr::applyOp(), RuleEmbed::applyOp(), RuleLess2Zero::applyOp(), RuleLessEqual2Zero::applyOp(), RuleSLess2Zero::applyOp(), RuleEqual2Zero::applyOp(), RuleEqual2Constant::applyOp(), RuleMultNegOne::applyOp(), RuleAddUnsigned::applyOp(), RuleDivTermAdd2::applyOp(), RuleSignDiv2::applyOp(), RuleSignNearMult::applyOp(), RuleModOpt::applyOp(), RuleSubvarAnd::applyOp(), RuleSubvarSubpiece::applyOp(), RuleSubvarCompZero::applyOp(), RuleSubvarZext::applyOp(), RuleSubvarSext::applyOp(), RuleFuncPtrEncoding::applyOp(), RuleThreeWayCompare::applyOp(), Funcdata::calcNZMask(), JumpBasic::calcRange(), AddrSpace::calcScaleMask(), ActionStackPtrFlow::checkClog(), ValueSetSolver::checkRelativeConstant(), CircleRange::CircleRange(), Funcdata::collapseIntMultMult(), RuleThreeWayCompare::detectThreeWay(), Funcdata::distributeIntMultAdd(), OpBehaviorIntAdd::evaluateBinary(), OpBehaviorIntSub::evaluateBinary(), OpBehaviorIntCarry::evaluateBinary(), OpBehaviorIntLeft::evaluateBinary(), OpBehaviorIntRight::evaluateBinary(), OpBehaviorIntSright::evaluateBinary(), OpBehaviorIntMult::evaluateBinary(), OpBehaviorSubpiece::evaluateBinary(), JumpTable::foldInNormalization(), ActionDeadCode::gatherConsumedReturn(), AliasChecker::gatherOffset(), RuleSLess2Zero::getHiBit(), EmulatePcodeOp::getLoadImageValue(), EmulateSnippet::getLoadImageValue(), PcodeOp::getNZMaskLocal(), ContextDatabase::getTrackedValue(), MemoryBank::getValue(), EquateSymbol::isValueClose(), ActionDeadCode::markConsumedParameters(), minimalmask(), TransformManager::newConstant(), TransformManager::newSplit(), FloatFormat::opTrunc(), Funcdata::opUndoPtradd(), ActionDeadCode::propagateConsumed(), CircleRange::pullBack(), CircleRange::pullBackBinary(), CircleRange::pullBackUnary(), PrintC::push_integer(), ActionDeadCode::pushConsumed(), PrintC::pushEquate(), CircleRange::pushForwardBinary(), CircleRange::pushForwardUnary(), OpBehaviorIntAdd::recoverInputBinary(), OpBehaviorIntSub::recoverInputBinary(), OpBehaviorIntLeft::recoverInputBinary(), OpBehaviorIntZext::recoverInputUnary(), OpBehaviorIntSext::recoverInputUnary(), Funcdata::replaceLessequal(), SegmentedResolver::resolve(), TypeFactory::setEnumValues(), CircleRange::setFull(), CircleRange::setNZMask(), CircleRange::setRange(), SubvariableFlow::setReplacement(), MemoryBank::setValue(), sign_extend(), SubvariableFlow::traceBackward(), SubvariableFlow::traceBackwardSext(), SubvariableFlow::traceForward(), SubvariableFlow::traceForwardSext(), ActionLikelyTrash::traceTrash(), Funcdata::transferVarnodeProperties(), and uintb_negate().

int4 count_leading_zeros

(

uintb

val

)

Return the number of leading zero bits in the given value.

Count the number of more significant zero bits before the most significant one bit in the representation of the given value;

Parameters

val	is the given value

Returns

the number of zero bits

Referenced by RuleDivOpt::findForm(), CircleRange::getMaxInfo(), minimalmask(), CircleRange::pushForwardBinary(), and JumpValuesRange::truncate().

uintb coveringmask

(

uintb

val

)

Return a mask that covers the given value.

Return smallest number of form 2^n-1, bigger or equal to the given value

Parameters

val	is the given value

Returns

the mask

Referenced by JumpBasic::calcRange(), PcodeOp::getNZMaskLocal(), minimalmask(), and ActionDeadCode::propagateConsumed().

int4 leastsigbit_set

(

uintb

val

)

Return index of least significant bit set in given value.

The least significant bit is index 0.

Parameters

val	is the given value

Returns

the index of the least significant set bit, or -1 if none are set

Referenced by SubvariableFlow::addConstant(), SubvariableFlow::addSuggestedPatch(), RuleShiftBitops::applyOp(), RuleDoubleShift::applyOp(), RuleShiftCompare::applyOp(), RuleShiftAnd::applyOp(), RuleSubvarCompZero::applyOp(), RulePopcountBoolXor::applyOp(), PcodeOp::getNZMaskLocal(), SubvariableFlow::getReplacementAddress(), CircleRange::minimalContainer(), minimalmask(), ActionDeadCode::propagateConsumed(), CircleRange::setNZMask(), SubvariableFlow::SubvariableFlow(), SubvariableFlow::traceBackward(), and SubvariableFlow::traceForward().

uintb minimalmask ( uintb  val )

inline

Calculate smallest mask that covers the given value.

Calculcate a mask that covers either the least significant byte, uint2, uint4, or uint8, whatever is smallest.

Parameters

val	is the given value

Returns

the minimal mask

References bit_transitions(), byte_swap(), calc_mask(), count_leading_zeros(), coveringmask(), leastsigbit_set(), mostsigbit_set(), mult64to128(), popcount(), power2Divide(), sign_extend(), signbit_negative(), uintb_negate(), unsignedCompare128(), unsignedSubtract128(), and zero_extend().

Referenced by JumpTable::foldInNormalization(), ActionDeadCode::gatherConsumedReturn(), and ActionDeadCode::markConsumedParameters().

int4 mostsigbit_set

(

uintb

val

)

Return index of most significant bit set in given value.

The least significant bit is index 0.

Parameters

val	is the given value

Returns

the index of the most significant set bit, or -1 if none are set

Referenced by RulePopcountBoolXor::applyOp(), PrintLanguage::formatBinary(), PcodeOp::getNZMaskLocal(), CircleRange::minimalContainer(), minimalmask(), CircleRange::pullBack(), GuardRecord::quasiCopy(), SubvariableFlow::SubvariableFlow(), and StringManager::writeUtf8().

void mult64to128	(	uint8 *	res,
		uint8	x,
		uint8	y
	)

Multiply 2 unsigned 64-bit values, producing a 128-bit value.

TODO: Remove once we import a full multiprecision library.

Parameters

res	points to the result array (2 uint8 pieces)
x	is the first 64-bit value
y	is the second 64-bit value

Referenced by minimalmask(), and power2Divide().

uintb pcode_left ( uintb  val, int4  sa  )

inline

Perform a CPUI_INT_LEFT on the given val

Parameters

val	is the value to shift
sa	is the number of bits to shift

Returns

the shifted value

Referenced by RuleShiftBitops::applyOp(), and PcodeOp::getNZMaskLocal().

uintb pcode_right ( uintb  val, int4  sa  )

inline

Perform a CPUI_INT_RIGHT on the given val

Parameters

val	is the value to shift
sa	is the number of bits to shift

Returns

the shifted value

Referenced by RuleShiftBitops::applyOp(), and PcodeOp::getNZMaskLocal().

int4 popcount

(

uintb

val

)

Return the number of one bits in the given value.

Count the number (population) bits set.

Parameters

val	is the given value

Returns

the number of one bits

Referenced by RulePopcountBoolXor::applyOp(), OpBehaviorPopcount::evaluateUnary(), PcodeOp::getNZMaskLocal(), and minimalmask().

int4 power2Divide	(	int4	n,
		uint8	divisor,
		uint8 &	q,
		uint8 &	r
	)

Unsigned division of a power of 2 (upto 2^127) by a 64-bit divisor.

The result must be less than 2^64. The remainder is calculated.

Parameters

n	is the power of 2 for the numerand
divisor	is the 64-bit divisor
q	is the passed back 64-bit quotient
r	is the passed back 64-bit remainder

Returns

0 if successful, 1 if result is too big, 2 if divide by 0

References mult64to128(), unsignedCompare128(), and unsignedSubtract128().

Referenced by RuleDivOpt::calcDivisor(), and minimalmask().

uintb sign_extend	(	uintb	in,
		int4	sizein,
		int4	sizeout
	)

Sign-extend a value between two byte sizes.

Take the first sizein bytes of the given in and sign-extend this to sizeout bytes, keeping any more significant bytes zero

Parameters

in	is the given value
sizein	is the size to treat that value as an input
sizeout	is the size to sign-extend the value to

Returns

the sign-extended value

References calc_mask().

Referenced by MapState::addRange(), RuleSubCommute::applyOp(), AddTreeState::buildMultiples(), ScopeLocal::buildVariableName(), AddTreeState::calcSubtype(), AddTreeState::checkMultTerm(), AddTreeState::checkTerm(), OpBehaviorIntSdiv::evaluateBinary(), OpBehaviorIntSrem::evaluateBinary(), OpBehaviorIntSext::evaluateUnary(), AddTreeState::findArrayHint(), PcodeOp::getNZMaskLocal(), MapState::initialize(), EquateSymbol::isValueClose(), minimalmask(), FloatFormat::opInt2Float(), CircleRange::pullBackUnary(), PrintC::pushEquate(), CircleRange::pushForwardBinary(), CircleRange::pushForwardUnary(), Funcdata::replaceLessequal(), and SubvariableFlow::setReplacement().

void sign_extend	(	intb &	val,
		int4	bit
	)

Sign extend above given bit.

Sign extend val starting at bit

Parameters

val	is a reference to the value to be sign-extended
bit	is the index of the bit to extend from (0=least significant bit)

bool signbit_negative	(	uintb	val,
		int4	size
	)

Return true if the sign-bit is set.

Treat the given val as a constant of size bytes

Parameters

val	is the given value
size	is the size in bytes

Returns

true if the constant (as sized) has its sign bit set

Referenced by RuleSlessToLess::applyOp(), RuleDivTermAdd::applyOp(), OpBehaviorIntSright::evaluateBinary(), CastStrategyC::localExtensionType(), minimalmask(), and ConditionMarker::sameOpComplement().

uintb uintb_negate	(	uintb	in,
		int4	size
	)

Negate the sized value.

Treat the given in as a constant of size bytes. Negate this constant keeping the upper bytes zero.

Parameters

in	is the given value
size	is the size in bytes

Returns

the negation of the sized constant

References calc_mask().

Referenced by RuleSborrow::applyOp(), RuleEqual2Zero::applyOp(), AddTreeState::buildExtra(), OpBehaviorInt2Comp::evaluateUnary(), OpBehaviorIntNegate::evaluateUnary(), and minimalmask().

int4 unsignedCompare128	(	uint8 *	a,
		uint8 *	b
	)

Compare two unsigned 128-bit values.

TODO: Remove once we import a full multiprecision library. Given a first and second value, return -1, 0, or 1 depending on whether the first value is less, equal, or greater than the second value.

Parameters

a	is the first 128-bit value (as an array of 2 uint8 elements)
b	is the second 128-bit value

Returns

the comparison code

Referenced by minimalmask(), and power2Divide().

void unsignedSubtract128	(	uint8 *	a,
		uint8 *	b
	)

Subtract (in-place) a 128-bit value from a base 128-value.

The base value is altered in place. TODO: Remove once we import a full multiprecision library.

Parameters

a	is the base 128-bit value being subtracted from in-place
b	is the other 128-bit value being subtracted

Referenced by minimalmask(), and power2Divide().

void zero_extend	(	intb &	val,
		int4	bit
	)

Clear all bits above given bit.

Zero extend val starting at bit

Parameters

val	is a reference to the value to be zero extended
bit	is the index of the bit to extend from (0=least significant bit)

Referenced by OpBehaviorIntSdiv::evaluateBinary(), OpBehaviorIntSrem::evaluateBinary(), and minimalmask().