Files
gem5/src/cpu/reg_class.hh
Gabe Black e59f01a55e cpu: Make RegClass able to flatten RegIds.
This makes RegIds and the RegClass-es associated with them responsible
for their own flattening. If they don't need to be flattened (a common
case) then they just mark themselves as already flat and that step can
be skipped.

This will also make it possible to get rid of the (get|set)RegFlat APIs,
since if you want to use flattened registers, you'll either have or
create a flattened RegId and pass it into the same (get|set)Reg method.

By making flattening work on RegIds instead of RegIndexes, this will
also make it possible for registers to start out in one RegClass and
move into another one. This would be useful if, for instance, there were
multiple groups of integer registers which had different indexing
semantics, but which should all end up in the same pool for renaming.

For instance, on x86, there are three distinct classes of FP registers.
They are the MMX registers, the pairs of registers which back the XMM
registers, and the X87 registers. Only the last of these needs
flattening. These could all be treated as different RegClass-es
pre-flattening, and could converge on the underlying floating point
register file post-flattening.

Another example in x86 is that some registers can encode that they
should refer to either the first byte of one register, or the second
byte of another register. This only applies to some registers though,
and so only those would need to go through the flattening step.

Another major advantage is that this removes the need for flattening
functions on the ISA object. Having those, and treating the ISA object
as a TheISA::ISA instead of the more generic BaseISA, was done to make
the flattening functions inline, and to make them fold away in cases
where flattening is not necessary. This new scheme isn't *quite* as
streamline as that, since you'll actually need to check if something is
already flattened. You won't, however, need to check what type the
register is and then look up the right flattening function, so that will
likely compensate.

Change-Id: I3c648cc8c0776b0e1020504113445b7d033e665f
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/51227
Maintainer: Gabe Black <gabe.black@gmail.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
2022-08-09 09:17:53 +00:00

522 lines
15 KiB
C++

/*
* Copyright (c) 2016-2019 ARM Limited
* All rights reserved
*
* The license below extends only to copyright in the software and shall
* not be construed as granting a license to any other intellectual
* property including but not limited to intellectual property relating
* to a hardware implementation of the functionality of the software
* licensed hereunder. You may use the software subject to the license
* terms below provided that you ensure that this notice is replicated
* unmodified and in its entirety in all distributions of the software,
* modified or unmodified, in source code or in binary form.
*
* Copyright (c) 2013 Advanced Micro Devices, Inc.
* All rights reserved
*.
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are
* met: redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer;
* redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution;
* neither the name of the copyright holders nor the names of its
* contributors may be used to endorse or promote products derived from
* this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef __CPU__REG_CLASS_HH__
#define __CPU__REG_CLASS_HH__
#include <cstddef>
#include <iterator>
#include <string>
#include "base/cprintf.hh"
#include "base/debug.hh"
#include "base/intmath.hh"
#include "base/types.hh"
#include "debug/InvalidReg.hh"
namespace gem5
{
/** Enumerate the classes of registers. */
enum RegClassType
{
IntRegClass, ///< Integer register
FloatRegClass, ///< Floating-point register
/** Vector Register. */
VecRegClass,
/** Vector Register Native Elem lane. */
VecElemClass,
VecPredRegClass,
CCRegClass, ///< Condition-code register
MiscRegClass, ///< Control (misc) register
InvalidRegClass = -1
};
// "Standard" register class names. Using these is encouraged but optional.
inline constexpr char IntRegClassName[] = "integer";
inline constexpr char FloatRegClassName[] = "floating_point";
inline constexpr char VecRegClassName[] = "vector";
inline constexpr char VecElemClassName[] = "vector_element";
inline constexpr char VecPredRegClassName[] = "vector_predicate";
inline constexpr char CCRegClassName[] = "condition_code";
inline constexpr char MiscRegClassName[] = "miscellaneous";
class RegClass;
class RegClassIterator;
class BaseISA;
/** Register ID: describe an architectural register with its class and index.
* This structure is used instead of just the register index to disambiguate
* between different classes of registers. For example, a integer register with
* index 3 is represented by Regid(IntRegClass, 3).
*/
class RegId
{
protected:
const RegClass *_regClass = nullptr;
RegIndex regIdx;
int numPinnedWrites;
friend struct std::hash<RegId>;
friend class RegClassIterator;
public:
inline constexpr RegId();
constexpr RegId(const RegClass &reg_class, RegIndex reg_idx)
: _regClass(&reg_class), regIdx(reg_idx), numPinnedWrites(0)
{}
constexpr operator RegIndex() const
{
return index();
}
constexpr bool
operator==(const RegId& that) const
{
return classValue() == that.classValue() && regIdx == that.index();
}
constexpr bool
operator!=(const RegId& that) const
{
return !(*this==that);
}
/** Order operator.
* The order is required to implement maps with key type RegId
*/
constexpr bool
operator<(const RegId& that) const
{
return classValue() < that.classValue() ||
(classValue() == that.classValue() && (regIdx < that.index()));
}
/**
* Return true if this register can be renamed
*/
constexpr bool
isRenameable() const
{
return classValue() != MiscRegClass && classValue() != InvalidRegClass;
}
/** @return true if it is of the specified class. */
inline constexpr bool is(RegClassType reg_class) const;
/** Index accessors */
/** @{ */
constexpr RegIndex index() const { return regIdx; }
/** Class accessor */
constexpr const RegClass &regClass() const { return *_regClass; }
inline constexpr RegClassType classValue() const;
/** Return a const char* with the register class name. */
inline constexpr const char* className() const;
inline constexpr bool isFlat() const;
inline RegId flatten(const BaseISA &isa) const;
int getNumPinnedWrites() const { return numPinnedWrites; }
void setNumPinnedWrites(int num_writes) { numPinnedWrites = num_writes; }
friend inline std::ostream& operator<<(std::ostream& os, const RegId& rid);
};
class RegClassOps
{
public:
/** Print the name of the register specified in id. */
virtual std::string regName(const RegId &id) const;
/** Print the value of a register pointed to by val of size size. */
virtual std::string valString(const void *val, size_t size) const;
/** Flatten register id id using information in the ISA object isa. */
virtual RegId
flatten(const BaseISA &isa, const RegId &id) const
{
return id;
}
};
class RegClassIterator;
class RegClass
{
private:
RegClassType _type;
const char *_name;
size_t _numRegs;
size_t _regBytes = sizeof(RegVal);
// This is how much to shift an index by to get an offset of a register in
// a register file from the register index, which would otherwise need to
// be calculated with a multiply.
size_t _regShift = ceilLog2(sizeof(RegVal));
static inline RegClassOps defaultOps;
const RegClassOps *_ops = &defaultOps;
const debug::Flag &debugFlag;
bool _flat = true;
public:
constexpr RegClass(RegClassType type, const char *new_name,
size_t num_regs, const debug::Flag &debug_flag) :
_type(type), _name(new_name), _numRegs(num_regs), debugFlag(debug_flag)
{}
constexpr RegClass
needsFlattening() const
{
RegClass reg_class = *this;
reg_class._flat = false;
return reg_class;
}
constexpr RegClass
ops(const RegClassOps &new_ops) const
{
RegClass reg_class = *this;
reg_class._ops = &new_ops;
return reg_class;
}
template <class RegType>
constexpr RegClass
regType() const
{
RegClass reg_class = *this;
reg_class._regBytes = sizeof(RegType);
reg_class._regShift = ceilLog2(reg_class._regBytes);
return reg_class;
}
constexpr RegClassType type() const { return _type; }
constexpr const char *name() const { return _name; }
constexpr size_t numRegs() const { return _numRegs; }
constexpr size_t regBytes() const { return _regBytes; }
constexpr size_t regShift() const { return _regShift; }
constexpr const debug::Flag &debug() const { return debugFlag; }
constexpr bool isFlat() const { return _flat; }
std::string regName(const RegId &id) const { return _ops->regName(id); }
std::string
valString(const void *val) const
{
return _ops->valString(val, regBytes());
}
RegId
flatten(const BaseISA &isa, const RegId &id) const
{
return isFlat() ? id : _ops->flatten(isa, id);
}
using iterator = RegClassIterator;
inline iterator begin() const;
inline iterator end() const;
inline constexpr RegId operator[](RegIndex idx) const;
};
inline constexpr RegClass
invalidRegClass(InvalidRegClass, "invalid", 0, debug::InvalidReg);
constexpr RegId::RegId() : RegId(invalidRegClass, 0) {}
constexpr bool
RegId::is(RegClassType reg_class) const
{
return _regClass->type() == reg_class;
}
constexpr RegClassType RegId::classValue() const { return _regClass->type(); }
constexpr const char* RegId::className() const { return _regClass->name(); }
constexpr bool RegId::isFlat() const { return _regClass->isFlat(); }
RegId
RegId::flatten(const BaseISA &isa) const
{
return _regClass->flatten(isa, *this);
}
std::ostream&
operator<<(std::ostream& os, const RegId& rid)
{
return os << rid.regClass().regName(rid);
}
class RegClassIterator
{
private:
RegId id;
RegClassIterator(const RegClass &reg_class, RegIndex idx) :
id(reg_class, idx)
{}
friend class RegClass;
public:
using iterator_category = std::forward_iterator_tag;
using difference_type = std::size_t;
using value_type = const RegId;
using pointer = value_type *;
using reference = value_type &;
reference operator*() const { return id; }
pointer operator->() { return &id; }
RegClassIterator &
operator++()
{
id.regIdx++;
return *this;
}
RegClassIterator
operator++(int)
{
auto tmp = *this;
++(*this);
return tmp;
}
bool
operator==(const RegClassIterator &other) const
{
return id == other.id;
}
bool
operator!=(const RegClassIterator &other) const
{
return id != other.id;
}
};
RegClassIterator
RegClass::begin() const
{
return RegClassIterator(*this, 0);
}
RegClassIterator
RegClass::end() const
{
return RegClassIterator(*this, numRegs());
}
constexpr RegId
RegClass::operator[](RegIndex idx) const
{
return RegId(*this, idx);
}
template <typename ValueType>
class TypedRegClassOps : public RegClassOps
{
public:
std::string
valString(const void *val, size_t size) const override
{
assert(size == sizeof(ValueType));
return csprintf("%s", *(const ValueType *)val);
}
};
template <typename ValueType>
class VecElemRegClassOps : public TypedRegClassOps<ValueType>
{
protected:
size_t elemsPerVec;
public:
explicit VecElemRegClassOps(size_t elems_per_vec) :
elemsPerVec(elems_per_vec)
{}
std::string
regName(const RegId &id) const override
{
RegIndex reg_idx = id.index() / elemsPerVec;
RegIndex elem_idx = id.index() % elemsPerVec;
return csprintf("v%d[%d]", reg_idx, elem_idx);
}
};
/** Physical register ID.
* Like a register ID but physical. The inheritance is private because the
* only relationship between this types is functional, and it is done to
* prevent code replication. */
class PhysRegId : private RegId
{
private:
RegIndex flatIdx;
int numPinnedWritesToComplete;
bool pinned;
public:
explicit PhysRegId() : RegId(invalidRegClass, -1), flatIdx(-1),
numPinnedWritesToComplete(0)
{}
/** Scalar PhysRegId constructor. */
explicit PhysRegId(const RegClass &reg_class, RegIndex _regIdx,
RegIndex _flatIdx)
: RegId(reg_class, _regIdx), flatIdx(_flatIdx),
numPinnedWritesToComplete(0), pinned(false)
{}
/** Visible RegId methods */
/** @{ */
using RegId::index;
using RegId::regClass;
using RegId::classValue;
using RegId::className;
using RegId::is;
/** @} */
/**
* Explicit forward methods, to prevent comparisons of PhysRegId with
* RegIds.
*/
/** @{ */
bool
operator<(const PhysRegId& that) const
{
return RegId::operator<(that);
}
bool
operator==(const PhysRegId& that) const
{
return RegId::operator==(that);
}
bool
operator!=(const PhysRegId& that) const
{
return RegId::operator!=(that);
}
/** @} */
/**
* Returns true if this register is always associated to the same
* architectural register.
*/
bool isFixedMapping() const { return !isRenameable(); }
/** Flat index accessor */
const RegIndex& flatIndex() const { return flatIdx; }
int getNumPinnedWrites() const { return numPinnedWrites; }
void
setNumPinnedWrites(int numWrites)
{
// An instruction with a pinned destination reg can get
// squashed. The numPinnedWrites counter may be zero when
// the squash happens but we need to know if the dest reg
// was pinned originally in order to reset counters properly
// for a possible re-rename using the same physical reg (which
// may be required in case of a mem access order violation).
pinned = (numWrites != 0);
numPinnedWrites = numWrites;
}
void decrNumPinnedWrites() { --numPinnedWrites; }
void incrNumPinnedWrites() { ++numPinnedWrites; }
bool isPinned() const { return pinned; }
int
getNumPinnedWritesToComplete() const
{
return numPinnedWritesToComplete;
}
void
setNumPinnedWritesToComplete(int numWrites)
{
numPinnedWritesToComplete = numWrites;
}
void decrNumPinnedWritesToComplete() { --numPinnedWritesToComplete; }
void incrNumPinnedWritesToComplete() { ++numPinnedWritesToComplete; }
};
using PhysRegIdPtr = PhysRegId*;
} // namespace gem5
namespace std
{
template<>
struct hash<gem5::RegId>
{
size_t
operator()(const gem5::RegId& reg_id) const
{
// Extract unique integral values for the effective fields of a RegId.
const size_t index = static_cast<size_t>(reg_id.index());
const size_t class_num = static_cast<size_t>(reg_id.classValue());
const size_t shifted_class_num =
class_num << (sizeof(gem5::RegIndex) << 3);
// Concatenate the class_num to the end of the flat_index, in order to
// maximize information retained.
const size_t concatenated_hash = index | shifted_class_num;
// If RegIndex is larger than size_t, then class_num will not be
// considered by this hash function, so we may wish to perform a
// different operation to include that information in the hash.
static_assert(sizeof(gem5::RegIndex) < sizeof(size_t),
"sizeof(RegIndex) should be less than sizeof(size_t)");
return concatenated_hash;
}
};
} // namespace std
#endif // __CPU__REG_CLASS_HH__