ghidra/Ghidra/Features/Decompiler/src/decompile/cpp/address.cc
2020-01-23 14:36:34 -05:00

925 lines
27 KiB
C++

/* ###
* IP: GHIDRA
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "address.hh"
#include "translate.hh"
ostream &operator<<(ostream &s,const SeqNum &sq)
{
sq.pc.printRaw(s);
s << ':' << sq.uniq;
return s;
}
/// This allows an Address to be written to a stream using
/// the standard '<<' operator. This is a wrapper for the
/// printRaw method and is intended for debugging and console
/// mode uses.
/// \param s is the stream being written to
/// \param addr is the Address to write
/// \return the output stream
ostream &operator<<(ostream &s,const Address &addr)
{
addr.printRaw(s);
return s;
}
SeqNum::SeqNum(Address::mach_extreme ex) : pc(ex)
{
uniq = (ex == Address::m_minimal) ? 0 : ~((uintm)0);
}
void SeqNum::saveXml(ostream &s) const
{
s << "<seqnum";
pc.getSpace()->saveXmlAttributes(s,pc.getOffset());
a_v_u(s,"uniq",uniq);
s << "/>";
}
SeqNum SeqNum::restoreXml(const Element *el,const AddrSpaceManager *manage)
{
uintm uniq = ~((uintm)0);
Address pc = Address::restoreXml(el,manage); // Recover address
for(int4 i=0;i<el->getNumAttributes();++i)
if (el->getAttributeName(i) == "uniq") {
istringstream s2(el->getAttributeValue(i)); // Recover unique (if present)
s2.unsetf(ios::dec | ios::hex | ios::oct);
s2 >> uniq;
break;
}
return SeqNum(pc,uniq);
}
/// Some data structures sort on an Address, and it is convenient
/// to be able to create an Address that is either bigger than
/// or smaller than all other Addresses.
/// \param ex is either \e m_minimal or \e m_maximal
Address::Address(mach_extreme ex)
{
if (ex == m_minimal) {
base = (AddrSpace *)0;
offset = 0;
}
else {
base = (AddrSpace *) ~((uintp)0);
offset = ~((uintb)0);
}
}
/// \deprecated Convert this to the most basic physical address.
/// This routine is only present for backward compatibility
/// with SLED
void Address::toPhysical(void)
{ AddrSpace *phys = base->getContain();
if ((phys != (AddrSpace *)0)&&(base->getType()==IPTR_SPACEBASE))
base = phys;
}
/// Return \b true if the range starting at \b this extending the given number of bytes
/// is contained by the second given range.
/// \param sz is the given number of bytes in \b this range
/// \param op2 is the start of the second given range
/// \param sz2 is the number of bytes in the second given range
/// \return \b true if the second given range contains \b this range
bool Address::containedBy(int4 sz,const Address &op2,int4 sz2) const
{
if (base != op2.base) return false;
if (op2.offset > offset) return false;
uintb off1 = offset + (sz-1);
uintb off2 = op2.offset + (sz2-1);
return (off2 >= off1);
}
/// Return -1 if (\e op2,\e sz2) is not properly contained in (\e this,\e sz).
/// If it is contained, return the endian aware offset of (\e op2,\e sz2)
/// I.e. if the least significant byte of the \e op2 range falls on the least significant
/// byte of the \e this range, return 0. If it intersects the second least significant, return 1, etc.
/// The -forceleft- toggle causes the check to be made against the left (lowest address) side
/// of the container, regardless of the endianness. I.e. it forces a little endian interpretation.
/// \param sz is the size of \e this range
/// \param op2 is the address of the second range
/// \param sz2 is the size of the second range
/// \param forceleft is \b true if containments is forced to be on the left even for big endian
/// \return the endian aware offset, or -1
int4 Address::justifiedContain(int4 sz,const Address &op2,int4 sz2,bool forceleft) const
{ if (base != op2.base) return -1;
if (op2.offset < offset) return -1;
uintb off1 = offset + (sz-1);
uintb off2 = op2.offset + (sz2-1);
if (off2 > off1) return -1;
if (base->isBigEndian()&&(!forceleft)) {
return (int4)(off1 - off2);
}
return (int4)(op2.offset - offset);
}
/// If \e this + \e skip falls in the range
/// \e op to \e op + \e size, then a non-negative integer is
/// returned indicating where in the interval it falls. I.e.
/// if \e this + \e skip == \e op, then 0 is returned. Otherwise
/// -1 is returned.
/// \param skip is an adjust to \e this address
/// \param op is the start of the range to check
/// \param size is the size of the range
/// \return an integer indicating how overlap occurs
int4 Address::overlap(int4 skip,const Address &op,int4 size) const
{
uintb dist;
if (base != op.base) return -1; // Must be in same address space to overlap
if (base->getType()==IPTR_CONSTANT) return -1; // Must not be constants
dist = base->wrapOffset(offset+skip-op.offset);
if (dist >= size) return -1; // but must fall before op+size
return (int4) dist;
}
/// Does the location \e this, \e sz form a contiguous region to \e loaddr, \e losz,
/// where \e this forms the most significant piece of the logical whole
/// \param sz is the size of \e this hi region
/// \param loaddr is the starting address of the low region
/// \param losz is the size of the low region
/// \return \b true if the pieces form a contiguous whole
bool Address::isContiguous(int4 sz,const Address &loaddr,int4 losz) const
{
if (base != loaddr.base) return false;
if (base->isBigEndian()) {
uintb nextoff = base->wrapOffset(offset+sz);
if (nextoff == loaddr.offset) return true;
}
else {
uintb nextoff = base->wrapOffset(loaddr.offset+losz);
if (nextoff == offset) return true;
}
return false;
}
/// This is usually used to build an address from an \b \<addr\>
/// tag, but it can be used to create an address from any tag
/// with the appropriate attributes
/// - \e space indicates the address space of the tag
/// - \e offset indicates the offset within the space
///
/// or a \e name attribute can be used to recover an address
/// based on a register name.
/// \param el is the parsed tag
/// \param manage is the address space manager for the program
/// \return the resulting Address
Address Address::restoreXml(const Element *el,const AddrSpaceManager *manage)
{
VarnodeData var;
var.restoreXml(el,manage);
return Address(var.space,var.offset);
}
/// This is usually used to build an address from an \b \<addr\>
/// tag, but it can be used to create an address from any tag
/// with the appropriate attributes
/// - \e space indicates the address space of the tag
/// - \e offset indicates the offset within the space
/// - \e size indicates the size of an address range
///
/// or a \e name attribute can be used to recover an address
/// and size based on a register name. If a size is recovered
/// it is stored in \e size reference.
/// \param el is the parsed tag
/// \param manage is the address space manager for the program
/// \param size is the reference to any recovered size
/// \return the resulting Address
Address Address::restoreXml(const Element *el,const AddrSpaceManager *manage,int4 &size)
{
VarnodeData var;
var.restoreXml(el,manage);
size = var.size;
return Address(var.space,var.offset);
}
/// Get the last address +1, updating the space, or returning
/// the extremal address if necessary
/// \param manage is used to fetch the next address space
Address Range::getLastAddrOpen(const AddrSpaceManager *manage) const
{
AddrSpace *curspc = spc;
uintb curlast = last;
if (curlast == curspc->getHighest()) {
curspc = manage->getNextSpaceInOrder(curspc);
curlast = 0;
}
else
curlast += 1;
if (curspc == (AddrSpace *)0)
return Address(Address::m_maximal);
return Address(curspc,curlast);
}
/// Output a description of this Range like: ram: 7f-9c
/// \param s is the output stream
void Range::printBounds(ostream &s) const
{
s << spc->getName() << ": ";
s << hex << first << '-' << last;
}
/// Write this object to a stream as a \<range> tag.
/// \param s is the output stream
void Range::saveXml(ostream &s) const
{
s << "<range";
a_v(s,"space",spc->getName());
a_v_u(s,"first",first);
a_v_u(s,"last",last);
s << "/>\n";
}
/// Reconstruct this object from an XML \<range> element
/// \param el is the XML element
/// \param manage is the space manage for recovering AddrSpace objects
void Range::restoreXml(const Element *el,const AddrSpaceManager *manage)
{
spc = (AddrSpace *)0;
first = 0;
last = ~((uintb)0);
for(int4 i=0;i<el->getNumAttributes();++i) {
if (el->getAttributeName(i) == "space") {
spc = manage->getSpaceByName(el->getAttributeValue(i));
if (spc == (AddrSpace *)0)
throw LowlevelError("Undefined space: "+el->getAttributeValue(i));
}
else if (el->getAttributeName(i) == "first") {
istringstream s(el->getAttributeValue(i));
s.unsetf(ios::dec | ios::hex | ios::oct);
s >> first;
}
else if (el->getAttributeName(i) == "last") {
istringstream s(el->getAttributeValue(i));
s.unsetf(ios::dec | ios::hex | ios::oct);
s >> last;
}
else if (el->getAttributeName(i) == "name") {
const Translate *trans = manage->getDefaultCodeSpace()->getTrans();
const VarnodeData &point(trans->getRegister(el->getAttributeValue(i)));
spc = point.space;
first = point.offset;
last = (first-1) + point.size;
break; // There should be no (space,first,last) attributes
}
}
if (spc == (AddrSpace *)0)
throw LowlevelError("No address space indicated in range tag");
last = spc->wrapOffset(last);
}
/// Insert a new Range merging as appropriate to maintain the disjoint cover
/// \param spc is the address space containing the new range
/// \param first is the offset of the first byte in the new range
/// \param last is the offset of the last byte in the new range
void RangeList::insertRange(AddrSpace *spc,uintb first,uintb last)
{
set<Range>::iterator iter1,iter2;
// we must have iter1.first > first
iter1 = tree.upper_bound(Range(spc,first,first));
// Set iter1 to first range with range.last >=first
// It is either current iter1 or the one before
if (iter1 != tree.begin()) {
--iter1;
if (((*iter1).spc!=spc)||((*iter1).last < first))
++iter1;
}
// Set iter2 to first range with range.first > last
iter2 = tree.upper_bound(Range(spc,last,last));
while(iter1!=iter2) {
if ((*iter1).first < first)
first = (*iter1).first;
if ((*iter1).last > last)
last = (*iter1).last;
tree.erase(iter1++);
}
tree.insert(Range(spc,first,last));
}
/// Remove/narrow/split existing Range objects to eliminate the indicated addresses
/// while still maintaining a disjoint cover.
/// \param spc is the address space of the address range to remove
/// \param first is the offset of the first byte of the range
/// \param last is the offset of the last byte of the range
void RangeList::removeRange(AddrSpace *spc,uintb first,uintb last)
{ // remove a range
set<Range>::iterator iter1,iter2;
if (tree.empty()) return; // Nothing to do
// we must have iter1.first > first
iter1 = tree.upper_bound(Range(spc,first,first));
// Set iter1 to first range with range.last >=first
// It is either current iter1 or the one before
if (iter1 != tree.begin()) {
--iter1;
if (((*iter1).spc!=spc)||((*iter1).last < first))
++iter1;
}
// Set iter2 to first range with range.first > last
iter2 = tree.upper_bound(Range(spc,last,last));
while(iter1!=iter2) {
uintb a,b;
a = (*iter1).first;
b = (*iter1).last;
tree.erase(iter1++);
if (a <first)
tree.insert(Range(spc,a,first-1));
if (b > last)
tree.insert(Range(spc,last+1,b));
}
}
void RangeList::merge(const RangeList &op2)
{ // Merge -op2- into this rangelist
set<Range>::const_iterator iter1,iter2;
iter1 = op2.tree.begin();
iter2 = op2.tree.end();
while(iter1 != iter2) {
const Range &range( *iter1 );
++iter1;
insertRange(range.spc, range.first, range.last);
}
}
/// Make sure indicated range of addresses is \e contained in \b this RangeList
/// \param addr is the first Address in the target range
/// \param size is the number of bytes in the target range
/// \return \b true is the range is fully contained by this RangeList
bool RangeList::inRange(const Address &addr,int4 size) const
{
set<Range>::const_iterator iter;
if (addr.isInvalid()) return true; // We don't really care
if (tree.empty()) return false;
// iter = first range with its first > addr
iter = tree.upper_bound(Range(addr.getSpace(),addr.getOffset(),addr.getOffset()));
if (iter == tree.begin()) return false;
// Set iter to last range with range.first <= addr
--iter;
// if (iter == tree.end()) // iter can't be end if non-empty
// return false;
if ((*iter).spc != addr.getSpace()) return false;
if ((*iter).last >= addr.getOffset()+size-1)
return true;
return false;
}
/// If \b this RangeList contains the specific address (spaceid,offset), return it
/// \return the containing Range or NULL
const Range *RangeList::getRange(AddrSpace *spaceid,uintb offset) const
{
if (tree.empty()) return (const Range *)0;
// iter = first range with its first > offset
set<Range>::const_iterator iter = tree.upper_bound(Range(spaceid,offset,offset));
if (iter == tree.begin()) return (const Range *)0;
// Set iter to last range with range.first <= offset
--iter;
if ((*iter).spc != spaceid) return (const Range *)0;
if ((*iter).last >= offset)
return &(*iter);
return (const Range *)0;
}
/// Return the size of the biggest contiguous sequence of addresses in
/// \b this RangeList which contain the given address
/// \param addr is the given address
/// \param maxsize is the large range to consider before giving up
/// \return the size (in bytes) of the biggest range
uintb RangeList::longestFit(const Address &addr,uintb maxsize) const
{
set<Range>::const_iterator iter;
if (addr.isInvalid()) return 0;
if (tree.empty()) return 0;
// iter = first range with its first > addr
uintb offset = addr.getOffset();
iter = tree.upper_bound(Range(addr.getSpace(),offset,offset));
if (iter == tree.begin()) return 0;
// Set iter to last range with range.first <= addr
--iter;
uintb sizeres = 0;
if ((*iter).last < offset) return sizeres;
do {
if ((*iter).spc != addr.getSpace()) break;
if ((*iter).first > offset) break;
sizeres += ((*iter).last + 1 - offset); // Size extends to end of range
offset = (*iter).last + 1; // Try to chain on the next range
if (sizeres >= maxsize) break; // Don't bother if past maxsize
++iter; // Next range in the chain
} while(iter != tree.end());
return sizeres;
}
/// \return the first contiguous range of addresses or NULL if empty
const Range *RangeList::getFirstRange(void) const
{
if (tree.empty()) return (const Range *)0;
return &(*tree.begin());
}
/// \return the last contiguous range of addresses or NULL if empty
const Range *RangeList::getLastRange(void) const
{
if (tree.empty()) return (const Range *)0;
set<Range>::const_iterator iter = tree.end();
--iter;
return &(*iter);
}
/// Treating offsets with their high-bits set as coming \e before
/// offset where the high-bit is clear, return the last/latest contiguous
/// Range within the given address space
/// \param spaceid is the given address space
/// \return indicated Range or NULL if empty
const Range *RangeList::getLastSignedRange(AddrSpace *spaceid) const
{
uintb midway = spaceid->getHighest() / 2; // Maximal signed value
Range range(spaceid,midway,midway);
set<Range>::const_iterator iter = tree.upper_bound(range); // First element greater than -range- (should be MOST negative)
if (iter!=tree.begin()) {
--iter;
if ((*iter).getSpace() == spaceid)
return &(*iter);
}
// If there were no "positive" ranges, search for biggest negative range
range = Range(spaceid,spaceid->getHighest(),spaceid->getHighest());
iter = tree.upper_bound(range);
if (iter != tree.begin()) {
--iter;
if ((*iter).getSpace() == spaceid)
return &(*iter);
}
return (const Range *)0;
}
/// Print a one line description of each disjoint Range making up \b this RangeList
/// \param s is the output stream
void RangeList::printBounds(ostream &s) const
{
if (tree.empty())
s << "all" << endl;
else {
set<Range>::const_iterator iter;
for(iter=tree.begin();iter!=tree.end();++iter) {
(*iter).printBounds(s);
s << endl;
}
}
}
/// Serialize this object to an XML \<rangelist> tag
/// \param s is the output stream
void RangeList::saveXml(ostream &s) const
{
set<Range>::const_iterator iter;
s << "<rangelist>\n";
for(iter=tree.begin();iter!=tree.end();++iter) {
(*iter).saveXml(s);
}
s << "</rangelist>\n";
}
/// Recover each individual disjoint Range for \b this RangeList as encoded
/// in a \<rangelist> tag.
/// \param el is the XML element
/// \param manage is manager for retrieving address spaces
void RangeList::restoreXml(const Element *el,const AddrSpaceManager *manage)
{
const List &list(el->getChildren());
List::const_iterator iter;
for(iter=list.begin();iter!=list.end();++iter) {
const Element *subel = *iter;
Range range;
range.restoreXml(subel,manage);
tree.insert(range);
}
}
#ifdef UINTB4
uintb uintbmasks[9] = { 0, 0xff, 0xffff, 0xffffff, 0xffffffff, 0xffffffff, 0xffffffff, 0xffffffff, 0xffffffff };
#else
uintb uintbmasks[9] = { 0, 0xff, 0xffff, 0xffffff, 0xffffffff, 0xffffffffffLL,
0xffffffffffffLL, 0xffffffffffffffLL, 0xffffffffffffffffLL };
#endif
/// Treat the given \b val as a constant of \b size bytes
/// \param val is the given value
/// \param size is the size in bytes
/// \return \b true if the constant (as sized) has its sign bit set
bool signbit_negative(uintb val,int4 size)
{ // Return true if signbit is set (negative)
uintb mask = 0x80;
mask <<= 8*(size-1);
return ((val&mask) != 0);
}
/// Treat the given \b in as a constant of \b size bytes.
/// Negate this constant keeping the upper bytes zero.
/// \param in is the given value
/// \param size is the size in bytes
/// \return the negation of the sized constant
uintb uintb_negate(uintb in,int4 size)
{ // Invert bits
return ((~in)&calc_mask(size));
}
/// Take the first \b sizein bytes of the given \b in and sign-extend
/// this to \b sizeout bytes, keeping any more significant bytes zero
/// \param in is the given value
/// \param sizein is the size to treat that value as an input
/// \param sizeout is the size to sign-extend the value to
/// \return the sign-extended value
uintb sign_extend(uintb in,int4 sizein,int4 sizeout)
{
int4 signbit;
uintb mask;
signbit = sizein*8 - 1;
in &= calc_mask(sizein);
if (sizein >= sizeout) return in;
if ((in>>signbit) != 0) {
mask = calc_mask(sizeout);
uintb tmp = mask << signbit; // Split shift into two pieces
tmp = (tmp<<1) & mask; // In case, everything is shifted out
in |= tmp;
}
return in;
}
/// Sign extend \b val starting at \b bit
/// \param val is a reference to the value to be sign-extended
/// \param bit is the index of the bit to extend from (0=least significant bit)
void sign_extend(intb &val,int4 bit)
{
intb mask = 0;
mask = (~mask)<<bit;
if (((val>>bit)&1)!=0)
val |= mask;
else
val &= (~mask);
}
/// Zero extend \b val starting at \b bit
/// \param val is a reference to the value to be zero extended
/// \param bit is the index of the bit to extend from (0=least significant bit)
void zero_extend(intb &val,int4 bit)
{
intb mask = 0;
mask = (~mask)<<bit;
mask <<= 1;
val &= (~mask);
}
/// Swap the least significant \b size bytes in \b val
/// \param val is a reference to the value to swap
/// \param size is the number of bytes to swap
void byte_swap(intb &val,int4 size)
{
intb res = 0;
while(size>0) {
res <<= 8;
res |= (val&0xff);
val >>= 8;
size -= 1;
}
val = res;
}
/// Swap the least significant \b size bytes in \b val
/// \param val is the value to swap
/// \param size is the number of bytes to swap
/// \return the swapped value
uintb byte_swap(uintb val,int4 size)
{
uintb res=0;
while(size>0) {
res <<= 8;
res |= (val&0xff);
val >>= 8;
size -= 1;
}
return res;
}
/// The least significant bit is index 0.
/// \param val is the given value
/// \return the index of the least significant set bit, or -1 if none are set
int4 leastsigbit_set(uintb val)
{
if (val==0) return -1;
int4 res = 0;
int4 sz = 4*sizeof(uintb);
uintb mask = ~((uintb)0);
do {
mask >>= sz;
if ((mask&val)==0) {
res += sz;
val >>= sz;
}
sz >>= 1;
} while(sz!=0);
return res;
}
/// The least significant bit is index 0.
/// \param val is the given value
/// \return the index of the most significant set bit, or -1 if none are set
int4 mostsigbit_set(uintb val)
{
if (val==0) return -1;
int4 res = 8*sizeof(uintb)-1;
int4 sz = 4*sizeof(uintb);
uintb mask = ~((uintb)0);
do {
mask <<= sz;
if ((mask&val)==0) {
res -= sz;
val <<= sz;
}
sz >>= 1;
} while(sz != 0);
return res;
}
/// Count the number (population) bits set.
/// \param val is the given value
/// \return the number of one bits
int4 popcount(uintb val)
{
val = (val & 0x5555555555555555L) + ((val >> 1) & 0x5555555555555555L);
val = (val & 0x3333333333333333L) + ((val >> 2) & 0x3333333333333333L);
val = (val & 0x0f0f0f0f0f0f0f0fL) + ((val >> 4) & 0x0f0f0f0f0f0f0f0fL);
val = (val & 0x00ff00ff00ff00ffL) + ((val >> 8) & 0x00ff00ff00ff00ffL);
val = (val & 0x0000ffff0000ffffL) + ((val >> 16) & 0x0000ffff0000ffffL);
int4 res = (int4)(val & 0xff);
res += (int4)((val >> 32) & 0xff);
return res;
}
/// Count the number of more significant zero bits before the most significant
/// one bit in the representation of the given value;
/// \param val is the given value
/// \return the number of zero bits
int4 count_leading_zeros(uintb val)
{
if (val == 0)
return 8*sizeof(uintb);
uintb mask = ~((uintb)0);
int4 maskSize = 4*sizeof(uintb);
mask &= (mask << maskSize);
int4 bit = 0;
do {
if ((mask & val)==0) {
bit += maskSize;
maskSize >>= 1;
mask |= (mask >> maskSize);
}
else {
maskSize >>= 1;
mask &= (mask << maskSize);
}
} while(maskSize != 0);
return bit;
}
/// Return smallest number of form 2^n-1, bigger or equal to the given value
/// \param val is the given value
/// \return the mask
uintb coveringmask(uintb val)
{
uintb res = val;
int4 sz = 1;
while(sz < 8*sizeof(uintb)) {
res = res | (res>>sz);
sz <<= 1;
}
return res;
}
/// Treat \b val as a constant of size \b sz.
/// Scanning across the bits of \b 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
/// \param val is the given value
/// \param sz is the size to treat the value as
/// \return the number of transitions
int4 bit_transitions(uintb val,int4 sz)
{
int4 res = 0;
int4 last = val & 1;
int4 cur;
for(int4 i=1;i<8*sz;++i) {
val >>= 1;
cur = val & 1;
if (cur != last) {
res += 1;
last = cur;
}
if (val==0) break;
}
return res;
}
/// \brief Multiply 2 unsigned 64-bit values, producing a 128-bit value
///
/// TODO: Remove once we import a full multiprecision library.
/// \param res points to the result array (2 uint8 pieces)
/// \param x is the first 64-bit value
/// \param y is the second 64-bit value
void mult64to128(uint8 *res,uint8 x,uint8 y)
{
uint8 f = x & 0xffffffff;
uint8 e = x >> 32;
uint8 d = y & 0xffffffff;
uint8 c = y >> 32;
uint8 fd = f * d;
uint8 fc = f * c;
uint8 ed = e * d;
uint8 ec = e * c;
uint8 tmp = (fd >> 32) + (fc & 0xffffffff) + (ed & 0xffffffff);
res[1] = (tmp>>32) + (fc>>32) + (ed>>32) + ec;
res[0] = (tmp<<32) + (fd & 0xffffffff);
}
/// \brief 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.
/// \param a is the base 128-bit value being subtracted from in-place
/// \param b is the other 128-bit value being subtracted
void unsignedSubtract128(uint8 *a,uint8 *b)
{
bool borrow = (a[0] < b[0]);
a[0] -= b[0];
a[1] -= b[1];
if (borrow)
a[1] -= 1;
}
/// \brief 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 \e less, \e equal, or \e greater than the second value.
/// \param a is the first 128-bit value (as an array of 2 uint8 elements)
/// \param b is the second 128-bit value
/// \return the comparison code
int4 unsignedCompare128(uint8 *a,uint8 *b)
{
if (a[1] != b[1])
return (a[1] < b[1]) ? -1 : 1;
if (a[0] != b[0])
return (a[0] < b[0]) ? -1 : 1;
return 0;
}
/// \brief 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.
/// \param n is the power of 2 for the numerand
/// \param divisor is the 64-bit divisor
/// \param q is the passed back 64-bit quotient
/// \param r is the passed back 64-bit remainder
/// \return 0 if successful, 1 if result is too big, 2 if divide by 0
int4 power2Divide(int4 n,uint8 divisor,uint8 &q,uint8 &r)
{
if (divisor == 0) return 2;
uint8 power = 1;
if (n < 64) {
power <<= n;
q = power / divisor;
r = power % divisor;
return 0;
}
// Divide numerand and divisor by 2^(n-63) to get approximation of result
uint8 y = divisor >> (n-64); // Most of the way on divisor
if (y == 0) return 1; // Check if result will be too big
y >>= 1; // Divide divisor by final bit
power <<= 63;
uint8 max;
if (y == 0) {
max = 0;
max -= 1; // Could be maximal
// Check if divisor is a power of 2
if ((((uint8)1) << (n-64)) == divisor)
return 1;
}
else
max = power / y + 1;
uint8 min = power / (y+1);
if (min != 0)
min -= 1;
uint8 fullpower[2];
fullpower[1] = ((uint8)1)<<(n-64);
fullpower[0] = 0;
uint8 mult[2];
mult[0] = 0;
mult[1] = 0;
uint8 tmpq = 0;
while(max > min+1) {
tmpq = max + min;
if (tmpq < min) {
tmpq = (tmpq>>1) + 0x8000000000000000L;
}
else
tmpq >>= 1;
mult64to128(mult,divisor,tmpq);
if (unsignedCompare128(fullpower,mult) < 0)
max = tmpq-1;
else
min = tmpq;
}
// min is now our putative quotient
if (tmpq != min)
mult64to128(mult,divisor,min);
unsignedSubtract128(fullpower,mult); // Calculate remainder
// min might be 1 too small
if (fullpower[1] != 0 || fullpower[0] >= divisor) {
q = min + 1;
r = fullpower[0] - divisor;
}
else {
q = min;
r = fullpower[0];
}
return 0;
}