forked from OSchip/llvm-project
2013 lines
75 KiB
ReStructuredText
2013 lines
75 KiB
ReStructuredText
===============================
|
|
TableGen Programmer's Reference
|
|
===============================
|
|
|
|
.. sectnum::
|
|
|
|
.. contents::
|
|
:local:
|
|
|
|
Introduction
|
|
============
|
|
|
|
The purpose of TableGen is to generate complex output files based on
|
|
information from source files that are significantly easier to code than the
|
|
output files would be, and also easier to maintain and modify over time. The
|
|
information is coded in a declarative style involving classes and records,
|
|
which are then processed by TableGen. The internalized records are passed on
|
|
to various *backends*, which extract information from a subset of the records
|
|
and generate one or more output files. These output files are typically
|
|
``.inc`` files for C++, but may be any type of file that the backend
|
|
developer needs.
|
|
|
|
This document describes the LLVM TableGen facility in detail. It is intended
|
|
for the programmer who is using TableGen to produce code for a project. If
|
|
you are looking for a simple overview, check out the :doc:`TableGen Overview
|
|
<./index>`. The various ``*-tblgen`` commands used to invoke TableGen are
|
|
described in :doc:`tblgen Family - Description to C++
|
|
Code<../CommandGuide/tblgen>`.
|
|
|
|
An example of a backend is ``RegisterInfo``, which generates the register
|
|
file information for a particular target machine, for use by the LLVM
|
|
target-independent code generator. See :doc:`TableGen Backends <./BackEnds>`
|
|
for a description of the LLVM TableGen backends, and :doc:`TableGen
|
|
Backend Developer's Guide <./BackGuide>` for a guide to writing a new
|
|
backend.
|
|
|
|
Here are a few of the things backends can do.
|
|
|
|
* Generate the register file information for a particular target machine.
|
|
|
|
* Generate the instruction definitions for a target.
|
|
|
|
* Generate the patterns that the code generator uses to match instructions
|
|
to intermediate representation (IR) nodes.
|
|
|
|
* Generate semantic attribute identifiers for Clang.
|
|
|
|
* Generate abstract syntax tree (AST) declaration node definitions for Clang.
|
|
|
|
* Generate AST statement node definitions for Clang.
|
|
|
|
|
|
Concepts
|
|
--------
|
|
|
|
TableGen source files contain two primary items: *abstract records* and
|
|
*concrete records*. In this and other TableGen documents, abstract records
|
|
are called *classes.* (These classes are different from C++ classes and do
|
|
not map onto them.) In addition, concrete records are usually just called
|
|
records, although sometimes the term *record* refers to both classes and
|
|
concrete records. The distinction should be clear in context.
|
|
|
|
Classes and concrete records have a unique *name*, either chosen by
|
|
the programmer or generated by TableGen. Associated with that name
|
|
is a list of *fields* with values and an optional list of *parent classes*
|
|
(sometimes called base or super classes). The fields are the primary data that
|
|
backends will process. Note that TableGen assigns no meanings to fields; the
|
|
meanings are entirely up to the backends and the programs that incorporate
|
|
the output of those backends.
|
|
|
|
.. note::
|
|
|
|
The term "parent class" can refer to a class that is a parent of another
|
|
class, and also to a class from which a concrete record inherits. This
|
|
nonstandard use of the term arises because TableGen treats classes and
|
|
concrete records similarly.
|
|
|
|
A backend processes some subset of the concrete records built by the
|
|
TableGen parser and emits the output files. These files are usually C++
|
|
``.inc`` files that are included by the programs that require the data in
|
|
those records. However, a backend can produce any type of output files. For
|
|
example, it could produce a data file containing messages tagged with
|
|
identifiers and substitution parameters. In a complex use case such as the
|
|
LLVM code generator, there can be many concrete records and some of them can
|
|
have an unexpectedly large number of fields, resulting in large output files.
|
|
|
|
In order to reduce the complexity of TableGen files, classes are used to
|
|
abstract out groups of record fields. For example, a few classes may
|
|
abstract the concept of a machine register file, while other classes may
|
|
abstract the instruction formats, and still others may abstract the
|
|
individual instructions. TableGen allows an arbitrary hierarchy of classes,
|
|
so that the abstract classes for two concepts can share a third superclass that
|
|
abstracts common "sub-concepts" from the two original concepts.
|
|
|
|
In order to make classes more useful, a concrete record (or another class)
|
|
can request a class as a parent class and pass *template arguments* to it.
|
|
These template arguments can be used in the fields of the parent class to
|
|
initialize them in a custom manner. That is, record or class ``A`` can
|
|
request parent class ``S`` with one set of template arguments, while record or class
|
|
``B`` can request ``S`` with a different set of arguments. Without template
|
|
arguments, many more classes would be required, one for each combination of
|
|
the template arguments.
|
|
|
|
Both classes and concrete records can include fields that are uninitialized.
|
|
The uninitialized "value" is represented by a question mark (``?``). Classes
|
|
often have uninitialized fields that are expected to be filled in when those
|
|
classes are inherited by concrete records. Even so, some fields of concrete
|
|
records may remain uninitialized.
|
|
|
|
TableGen provides *multiclasses* to collect a group of record definitions in
|
|
one place. A multiclass is a sort of macro that can be "invoked" to define
|
|
multiple concrete records all at once. A multiclass can inherit from other
|
|
multiclasses, which means that the multiclass inherits all the definitions
|
|
from its parent multiclasses.
|
|
|
|
`Appendix C: Sample Record`_ illustrates a complex record in the Intel X86
|
|
target and the simple way in which it is defined.
|
|
|
|
Source Files
|
|
============
|
|
|
|
TableGen source files are plain ASCII text files. The files can contain
|
|
statements, comments, and blank lines (see `Lexical Analysis`_). The standard file
|
|
extension for TableGen files is ``.td``.
|
|
|
|
TableGen files can grow quite large, so there is an include mechanism that
|
|
allows one file to include the content of another file (see `Include
|
|
Files`_). This allows large files to be broken up into smaller ones, and
|
|
also provides a simple library mechanism where multiple source files can
|
|
include the same library file.
|
|
|
|
TableGen supports a simple preprocessor that can be used to conditionalize
|
|
portions of ``.td`` files. See `Preprocessing Facilities`_ for more
|
|
information.
|
|
|
|
Lexical Analysis
|
|
================
|
|
|
|
The lexical and syntax notation used here is intended to imitate
|
|
`Python's`_ notation. In particular, for lexical definitions, the productions
|
|
operate at the character level and there is no implied whitespace between
|
|
elements. The syntax definitions operate at the token level, so there is
|
|
implied whitespace between tokens.
|
|
|
|
.. _`Python's`: http://docs.python.org/py3k/reference/introduction.html#notation
|
|
|
|
TableGen supports BCPL-style comments (``// ...``) and nestable C-style
|
|
comments (``/* ... */``).
|
|
TableGen also provides simple `Preprocessing Facilities`_.
|
|
|
|
Formfeed characters may be used freely in files to produce page breaks when
|
|
the file is printed for review.
|
|
|
|
The following are the basic punctuation tokens::
|
|
|
|
- + [ ] { } ( ) < > : ; . ... = ? #
|
|
|
|
Literals
|
|
--------
|
|
|
|
Numeric literals take one of the following forms:
|
|
|
|
.. productionlist::
|
|
TokInteger: `DecimalInteger` | `HexInteger` | `BinInteger`
|
|
DecimalInteger: ["+" | "-"] ("0"..."9")+
|
|
HexInteger: "0x" ("0"..."9" | "a"..."f" | "A"..."F")+
|
|
BinInteger: "0b" ("0" | "1")+
|
|
|
|
Observe that the :token:`DecimalInteger` token includes the optional ``+``
|
|
or ``-`` sign, unlike most languages where the sign would be treated as a
|
|
unary operator.
|
|
|
|
TableGen has two kinds of string literals:
|
|
|
|
.. productionlist::
|
|
TokString: '"' (non-'"' characters and escapes) '"'
|
|
TokCode: "[{" (shortest text not containing "}]") "}]"
|
|
|
|
A :token:`TokCode` is nothing more than a multi-line string literal
|
|
delimited by ``[{`` and ``}]``. It can break across lines and the
|
|
line breaks are retained in the string.
|
|
|
|
The current implementation accepts the following escape sequences::
|
|
|
|
\\ \' \" \t \n
|
|
|
|
Identifiers
|
|
-----------
|
|
|
|
TableGen has name- and identifier-like tokens, which are case-sensitive.
|
|
|
|
.. productionlist::
|
|
ualpha: "a"..."z" | "A"..."Z" | "_"
|
|
TokIdentifier: ("0"..."9")* `ualpha` (`ualpha` | "0"..."9")*
|
|
TokVarName: "$" `ualpha` (`ualpha` | "0"..."9")*
|
|
|
|
Note that, unlike most languages, TableGen allows :token:`TokIdentifier` to
|
|
begin with an integer. In case of ambiguity, a token is interpreted as a
|
|
numeric literal rather than an identifier.
|
|
|
|
TableGen has the following reserved keywords, which cannot be used as
|
|
identifiers::
|
|
|
|
assert bit bits class code
|
|
dag def else false foreach
|
|
defm defset defvar field if
|
|
in include int let list
|
|
multiclass string then true
|
|
|
|
.. warning::
|
|
The ``field`` reserved word is deprecated.
|
|
|
|
Bang operators
|
|
--------------
|
|
|
|
TableGen provides "bang operators" that have a wide variety of uses:
|
|
|
|
.. productionlist::
|
|
BangOperator: one of
|
|
: !add !and !cast !con !dag
|
|
: !empty !eq !filter !find !foldl
|
|
: !foreach !ge !getdagop !gt !head
|
|
: !if !interleave !isa !le !listconcat
|
|
: !listsplat !lt !mul !ne !not
|
|
: !or !setdagop !shl !size !sra
|
|
: !srl !strconcat !sub !subst !substr
|
|
: !tail !xor
|
|
|
|
The ``!cond`` operator has a slightly different
|
|
syntax compared to other bang operators, so it is defined separately:
|
|
|
|
.. productionlist::
|
|
CondOperator: !cond
|
|
|
|
See `Appendix A: Bang Operators`_ for a description of each bang operator.
|
|
|
|
Include files
|
|
-------------
|
|
|
|
TableGen has an include mechanism. The content of the included file
|
|
lexically replaces the ``include`` directive and is then parsed as if it was
|
|
originally in the main file.
|
|
|
|
.. productionlist::
|
|
IncludeDirective: "include" `TokString`
|
|
|
|
Portions of the main file and included files can be conditionalized using
|
|
preprocessor directives.
|
|
|
|
.. productionlist::
|
|
PreprocessorDirective: "#define" | "#ifdef" | "#ifndef"
|
|
|
|
Types
|
|
=====
|
|
|
|
The TableGen language is statically typed, using a simple but complete type
|
|
system. Types are used to check for errors, to perform implicit conversions,
|
|
and to help interface designers constrain the allowed input. Every value is
|
|
required to have an associated type.
|
|
|
|
TableGen supports a mixture of low-level types (e.g., ``bit``) and
|
|
high-level types (e.g., ``dag``). This flexibility allows you to describe a
|
|
wide range of records conveniently and compactly.
|
|
|
|
.. productionlist::
|
|
Type: "bit" | "int" | "string" | "dag"
|
|
:| "bits" "<" `TokInteger` ">"
|
|
:| "list" "<" `Type` ">"
|
|
:| `ClassID`
|
|
ClassID: `TokIdentifier`
|
|
|
|
``bit``
|
|
A ``bit`` is a boolean value that can be 0 or 1.
|
|
|
|
``int``
|
|
The ``int`` type represents a simple 64-bit integer value, such as 5 or
|
|
-42.
|
|
|
|
``string``
|
|
The ``string`` type represents an ordered sequence of characters of arbitrary
|
|
length.
|
|
|
|
``bits<``\ *n*\ ``>``
|
|
The ``bits`` type is a fixed-sized integer of arbitrary length *n* that
|
|
is treated as separate bits. These bits can be accessed individually.
|
|
A field of this type is useful for representing an instruction operation
|
|
code, register number, or address mode/register/displacement. The bits of
|
|
the field can be set individually or as subfields. For example, in an
|
|
instruction address, the addressing mode, base register number, and
|
|
displacement can be set separately.
|
|
|
|
``list<``\ *type*\ ``>``
|
|
This type represents a list whose elements are of the *type* specified in
|
|
angle brackets. The element type is arbitrary; it can even be another
|
|
list type. List elements are indexed from 0.
|
|
|
|
``dag``
|
|
This type represents a nestable directed acyclic graph (DAG) of nodes.
|
|
Each node has an *operator* and zero or more *arguments* (or *operands*).
|
|
An argument can be
|
|
another ``dag`` object, allowing an arbitrary tree of nodes and edges.
|
|
As an example, DAGs are used to represent code patterns for use by
|
|
the code generator instruction selection algorithms. See `Directed
|
|
acyclic graphs (DAGs)`_ for more details;
|
|
|
|
:token:`ClassID`
|
|
Specifying a class name in a type context indicates
|
|
that the type of the defined value must
|
|
be a subclass of the specified class. This is useful in conjunction with
|
|
the ``list`` type; for example, to constrain the elements of the list to a
|
|
common base class (e.g., a ``list<Register>`` can only contain definitions
|
|
derived from the ``Register`` class).
|
|
The :token:`ClassID` must name a class that has been previously
|
|
declared or defined.
|
|
|
|
|
|
Values and Expressions
|
|
======================
|
|
|
|
There are many contexts in TableGen statements where a value is required. A
|
|
common example is in the definition of a record, where each field is
|
|
specified by a name and an optional value. TableGen allows for a reasonable
|
|
number of different forms when building up value expressions. These forms
|
|
allow the TableGen file to be written in a syntax that is natural for the
|
|
application.
|
|
|
|
Note that all of the values have rules for converting them from one type to
|
|
another. For example, these rules allow you to assign a value like ``7``
|
|
to an entity of type ``bits<4>``.
|
|
|
|
.. productionlist::
|
|
Value: `SimpleValue` `ValueSuffix`*
|
|
:| `Value` "#" [`Value`]
|
|
ValueSuffix: "{" `RangeList` "}"
|
|
:| "[" `RangeList` "]"
|
|
:| "." `TokIdentifier`
|
|
RangeList: `RangePiece` ("," `RangePiece`)*
|
|
RangePiece: `TokInteger`
|
|
:| `TokInteger` "..." `TokInteger`
|
|
:| `TokInteger` "-" `TokInteger`
|
|
:| `TokInteger` `TokInteger`
|
|
|
|
.. warning::
|
|
The peculiar last form of :token:`RangePiece` is due to the fact that the
|
|
"``-``" is included in the :token:`TokInteger`, hence ``1-5`` gets lexed as
|
|
two consecutive tokens, with values ``1`` and ``-5``, instead of "1", "-",
|
|
and "5". The use of hyphen as the range punctuation is deprecated.
|
|
|
|
Simple values
|
|
-------------
|
|
|
|
The :token:`SimpleValue` has a number of forms.
|
|
|
|
.. productionlist::
|
|
SimpleValue: `TokInteger` | `TokString`+ | `TokCode`
|
|
|
|
A value can be an integer literal, a string literal, or a code literal.
|
|
Multiple adjacent string literals are concatenated as in C/C++; the simple
|
|
value is the concatenation of the strings. Code literals become strings and
|
|
are then indistinguishable from them.
|
|
|
|
.. productionlist::
|
|
SimpleValue2: "true" | "false"
|
|
|
|
The ``true`` and ``false`` literals are essentially syntactic sugar for the
|
|
integer values 1 and 0. They improve the readability of TableGen files when
|
|
boolean values are used in field initializations, bit sequences, ``if``
|
|
statements, etc. When parsed, these literals are converted to integers.
|
|
|
|
.. note::
|
|
|
|
Although ``true`` and ``false`` are literal names for 1 and 0, we
|
|
recommend as a stylistic rule that you use them for boolean
|
|
values only.
|
|
|
|
.. productionlist::
|
|
SimpleValue3: "?"
|
|
|
|
A question mark represents an uninitialized value.
|
|
|
|
.. productionlist::
|
|
SimpleValue4: "{" [`ValueList`] "}"
|
|
ValueList: `ValueListNE`
|
|
ValueListNE: `Value` ("," `Value`)*
|
|
|
|
This value represents a sequence of bits, which can be used to initialize a
|
|
``bits<``\ *n*\ ``>`` field (note the braces). When doing so, the values
|
|
must represent a total of *n* bits.
|
|
|
|
.. productionlist::
|
|
SimpleValue5: "[" `ValueList` "]" ["<" `Type` ">"]
|
|
|
|
This value is a list initializer (note the brackets). The values in brackets
|
|
are the elements of the list. The optional :token:`Type` can be used to
|
|
indicate a specific element type; otherwise the element type is inferred
|
|
from the given values. TableGen can usually infer the type, although
|
|
sometimes not when the value is the empty list (``[]``).
|
|
|
|
.. productionlist::
|
|
SimpleValue6: "(" `DagArg` [`DagArgList`] ")"
|
|
DagArgList: `DagArg` ("," `DagArg`)*
|
|
DagArg: `Value` [":" `TokVarName`] | `TokVarName`
|
|
|
|
This represents a DAG initializer (note the parentheses). The first
|
|
:token:`DagArg` is called the "operator" of the DAG and must be a record.
|
|
See `Directed acyclic graphs (DAGs)`_ for more details.
|
|
|
|
.. productionlist::
|
|
SimpleValue7: `TokIdentifier`
|
|
|
|
The resulting value is the value of the entity named by the identifier. The
|
|
possible identifiers are described here, but the descriptions will make more
|
|
sense after reading the remainder of this guide.
|
|
|
|
.. The code for this is exceptionally abstruse. These examples are a
|
|
best-effort attempt.
|
|
|
|
* A template argument of a ``class``, such as the use of ``Bar`` in::
|
|
|
|
class Foo <int Bar> {
|
|
int Baz = Bar;
|
|
}
|
|
|
|
* The implicit template argument ``NAME`` in a ``class`` or ``multiclass``
|
|
definition (see `NAME`_).
|
|
|
|
* A field local to a ``class``, such as the use of ``Bar`` in::
|
|
|
|
class Foo {
|
|
int Bar = 5;
|
|
int Baz = Bar;
|
|
}
|
|
|
|
* The name of a record definition, such as the use of ``Bar`` in the
|
|
definition of ``Foo``::
|
|
|
|
def Bar : SomeClass {
|
|
int X = 5;
|
|
}
|
|
|
|
def Foo {
|
|
SomeClass Baz = Bar;
|
|
}
|
|
|
|
* A field local to a record definition, such as the use of ``Bar`` in::
|
|
|
|
def Foo {
|
|
int Bar = 5;
|
|
int Baz = Bar;
|
|
}
|
|
|
|
Fields inherited from the record's parent classes can be accessed the same way.
|
|
|
|
* A template argument of a ``multiclass``, such as the use of ``Bar`` in::
|
|
|
|
multiclass Foo <int Bar> {
|
|
def : SomeClass<Bar>;
|
|
}
|
|
|
|
* A variable defined with the ``defvar`` or ``defset`` statements.
|
|
|
|
* The iteration variable of a ``foreach``, such as the use of ``i`` in::
|
|
|
|
foreach i = 0...5 in
|
|
def Foo#i;
|
|
|
|
.. productionlist::
|
|
SimpleValue8: `ClassID` "<" `ValueListNE` ">"
|
|
|
|
This form creates a new anonymous record definition (as would be created by an
|
|
unnamed ``def`` inheriting from the given class with the given template
|
|
arguments; see `def`_) and the value is that record. A field of the record can be
|
|
obtained using a suffix; see `Suffixed Values`_.
|
|
|
|
Invoking a class in this manner can provide a simple subroutine facility.
|
|
See `Using Classes as Subroutines`_ for more information.
|
|
|
|
.. productionlist::
|
|
SimpleValue9: `BangOperator` ["<" `Type` ">"] "(" `ValueListNE` ")"
|
|
:| `CondOperator` "(" `CondClause` ("," `CondClause`)* ")"
|
|
CondClause: `Value` ":" `Value`
|
|
|
|
The bang operators provide functions that are not available with the other
|
|
simple values. Except in the case of ``!cond``, a bang operator takes a list
|
|
of arguments enclosed in parentheses and performs some function on those
|
|
arguments, producing a value for that bang operator. The ``!cond`` operator
|
|
takes a list of pairs of arguments separated by colons. See `Appendix A:
|
|
Bang Operators`_ for a description of each bang operator.
|
|
|
|
|
|
Suffixed values
|
|
---------------
|
|
|
|
The :token:`SimpleValue` values described above can be specified with
|
|
certain suffixes. The purpose of a suffix is to obtain a subvalue of the
|
|
primary value. Here are the possible suffixes for some primary *value*.
|
|
|
|
*value*\ ``{17}``
|
|
The final value is bit 17 of the integer *value* (note the braces).
|
|
|
|
*value*\ ``{8...15}``
|
|
The final value is bits 8--15 of the integer *value*. The order of the
|
|
bits can be reversed by specifying ``{15...8}``.
|
|
|
|
*value*\ ``[4]``
|
|
The final value is element 4 of the list *value* (note the brackets).
|
|
In other words, the brackets act as a subscripting operator on the list.
|
|
This is the case only when a single element is specified.
|
|
|
|
*value*\ ``[4...7,17,2...3,4]``
|
|
The final value is a new list that is a slice of the list *value*.
|
|
The new list contains elements 4, 5, 6, 7, 17, 2, 3, and 4.
|
|
Elements may be included multiple times and in any order. This is the result
|
|
only when more than one element is specified.
|
|
|
|
*value*\ ``.``\ *field*
|
|
The final value is the value of the specified *field* in the specified
|
|
record *value*.
|
|
|
|
The paste operator
|
|
------------------
|
|
|
|
The paste operator (``#``) is the only infix operator available in TableGen
|
|
expressions. It allows you to concatenate strings or lists, but has a few
|
|
unusual features.
|
|
|
|
The paste operator can be used when specifying the record name in a
|
|
:token:`Def` or :token:`Defm` statement, in which case it must construct a
|
|
string. If an operand is an undefined name (:token:`TokIdentifier`) or the
|
|
name of a global :token:`Defvar` or :token:`Defset`, it is treated as a
|
|
verbatim string of characters. The value of a global name is not used.
|
|
|
|
The paste operator can be used in all other value expressions, in which case
|
|
it can construct a string or a list. Rather oddly, but consistent with the
|
|
previous case, if the *right-hand-side* operand is an undefined name or a
|
|
global name, it is treated as a verbatim string of characters. The
|
|
left-hand-side operand is treated normally.
|
|
|
|
Values can have a trailing paste operator, in which case the left-hand-side
|
|
operand is concatenated to an empty string.
|
|
|
|
`Appendix B: Paste Operator Examples`_ presents examples of the behavior of
|
|
the paste operator.
|
|
|
|
Statements
|
|
==========
|
|
|
|
The following statements may appear at the top level of TableGen source
|
|
files.
|
|
|
|
.. productionlist::
|
|
TableGenFile: (`Statement` | `IncludeDirective`
|
|
:| `PreprocessorDirective`)*
|
|
Statement: `Assert` | `Class` | `Def` | `Defm` | `Defset` | `Defvar`
|
|
:| `Foreach` | `If` | `Let` | `MultiClass`
|
|
|
|
The following sections describe each of these top-level statements.
|
|
|
|
|
|
``class`` --- define an abstract record class
|
|
---------------------------------------------
|
|
|
|
A ``class`` statement defines an abstract record class from which other
|
|
classes and records can inherit.
|
|
|
|
.. productionlist::
|
|
Class: "class" `ClassID` [`TemplateArgList`] `RecordBody`
|
|
TemplateArgList: "<" `TemplateArgDecl` ("," `TemplateArgDecl`)* ">"
|
|
TemplateArgDecl: `Type` `TokIdentifier` ["=" `Value`]
|
|
|
|
A class can be parameterized by a list of "template arguments," whose values
|
|
can be used in the class's record body. These template arguments are
|
|
specified each time the class is inherited by another class or record.
|
|
|
|
If a template argument is not assigned a default value with ``=``, it is
|
|
uninitialized (has the "value" ``?``) and must be specified in the template
|
|
argument list when the class is inherited (required argument). If an
|
|
argument is assigned a default value, then it need not be specified in the
|
|
argument list (optional argument). In the declaration, all required template
|
|
arguments must precede any optional arguments. The template argument default
|
|
values are evaluated from left to right.
|
|
|
|
The :token:`RecordBody` is defined below. It can include a list of
|
|
parent classes from which the current class inherits, along with field
|
|
definitions and other statements. When a class ``C`` inherits from another
|
|
class ``D``, the fields of ``D`` are effectively merged into the fields of
|
|
``C``.
|
|
|
|
A given class can only be defined once. A ``class`` statement is
|
|
considered to define the class if *any* of the following are true (the
|
|
:token:`RecordBody` elements are described below).
|
|
|
|
* The :token:`TemplateArgList` is present, or
|
|
* The :token:`ParentClassList` in the :token:`RecordBody` is present, or
|
|
* The :token:`Body` in the :token:`RecordBody` is present and not empty.
|
|
|
|
You can declare an empty class by specifying an empty :token:`TemplateArgList`
|
|
and an empty :token:`RecordBody`. This can serve as a restricted form of
|
|
forward declaration. Note that records derived from a forward-declared
|
|
class will inherit no fields from it, because those records are built when
|
|
their declarations are parsed, and thus before the class is finally defined.
|
|
|
|
.. _NAME:
|
|
|
|
Every class has an implicit template argument named ``NAME`` (uppercase),
|
|
which is bound to the name of the :token:`Def` or :token:`Defm` inheriting
|
|
from the class. If the class is inherited by an anonymous record, the name
|
|
is unspecified but globally unique.
|
|
|
|
See `Examples: classes and records`_ for examples.
|
|
|
|
Record Bodies
|
|
`````````````
|
|
|
|
Record bodies appear in both class and record definitions. A record body can
|
|
include a parent class list, which specifies the classes from which the
|
|
current class or record inherits fields. Such classes are called the
|
|
parent classes of the class or record. The record body also
|
|
includes the main body of the definition, which contains the specification
|
|
of the fields of the class or record.
|
|
|
|
.. productionlist::
|
|
RecordBody: `ParentClassList` `Body`
|
|
ParentClassList: [":" `ParentClassListNE`]
|
|
ParentClassListNE: `ClassRef` ("," `ClassRef`)*
|
|
ClassRef: (`ClassID` | `MultiClassID`) ["<" [`ValueList`] ">"]
|
|
|
|
A :token:`ParentClassList` containing a :token:`MultiClassID` is valid only
|
|
in the class list of a ``defm`` statement. In that case, the ID must be the
|
|
name of a multiclass.
|
|
|
|
.. productionlist::
|
|
Body: ";" | "{" `BodyItem`* "}"
|
|
BodyItem: (`Type` | "code") `TokIdentifier` ["=" `Value`] ";"
|
|
:| "let" `TokIdentifier` ["{" `RangeList` "}"] "=" `Value` ";"
|
|
:| "defvar" `TokIdentifier` "=" `Value` ";"
|
|
:| `Assert`
|
|
|
|
A field definition in the body specifies a field to be included in the class
|
|
or record. If no initial value is specified, then the field's value is
|
|
uninitialized. The type must be specified; TableGen will not infer it from
|
|
the value. The keyword ``code`` may be used to emphasize that the field
|
|
has a string value that is code.
|
|
|
|
The ``let`` form is used to reset a field to a new value. This can be done
|
|
for fields defined directly in the body or fields inherited from parent
|
|
classes. A :token:`RangeList` can be specified to reset certain bits in a
|
|
``bit<n>`` field.
|
|
|
|
The ``defvar`` form defines a variable whose value can be used in other
|
|
value expressions within the body. The variable is not a field: it does not
|
|
become a field of the class or record being defined. Variables are provided
|
|
to hold temporary values while processing the body. See `Defvar in a Record
|
|
Body`_ for more details.
|
|
|
|
When class ``C2`` inherits from class ``C1``, it acquires all the field
|
|
definitions of ``C1``. As those definitions are merged into class ``C2``, any
|
|
template arguments passed to ``C1`` by ``C2`` are substituted into the
|
|
definitions. In other words, the abstract record fields defined by ``C1`` are
|
|
expanded with the template arguments before being merged into ``C2``.
|
|
|
|
|
|
.. _def:
|
|
|
|
``def`` --- define a concrete record
|
|
------------------------------------
|
|
|
|
A ``def`` statement defines a new concrete record.
|
|
|
|
.. productionlist::
|
|
Def: "def" [`NameValue`] `RecordBody`
|
|
NameValue: `Value` (parsed in a special mode)
|
|
|
|
The name value is optional. If specified, it is parsed in a special mode
|
|
where undefined (unrecognized) identifiers are interpreted as literal
|
|
strings. In particular, global identifiers are considered unrecognized.
|
|
These include global variables defined by ``defvar`` and ``defset``. A
|
|
record name can be the null string.
|
|
|
|
If no name value is given, the record is *anonymous*. The final name of an
|
|
anonymous record is unspecified but globally unique.
|
|
|
|
Special handling occurs if a ``def`` appears inside a ``multiclass``
|
|
statement. See the ``multiclass`` section below for details.
|
|
|
|
A record can inherit from one or more classes by specifying the
|
|
:token:`ParentClassList` clause at the beginning of its record body. All of
|
|
the fields in the parent classes are added to the record. If two or more
|
|
parent classes provide the same field, the record ends up with the field value
|
|
of the last parent class.
|
|
|
|
As a special case, the name of a record can be passed as a template argument
|
|
to that record's parent classes. For example:
|
|
|
|
.. code-block:: text
|
|
|
|
class A <dag d> {
|
|
dag the_dag = d;
|
|
}
|
|
|
|
def rec1 : A<(ops rec1)>;
|
|
|
|
The DAG ``(ops rec1)`` is passed as a template argument to class ``A``. Notice
|
|
that the DAG includes ``rec1``, the record being defined.
|
|
|
|
The steps taken to create a new record are somewhat complex. See `How
|
|
records are built`_.
|
|
|
|
See `Examples: classes and records`_ for examples.
|
|
|
|
|
|
Examples: classes and records
|
|
-----------------------------
|
|
|
|
Here is a simple TableGen file with one class and two record definitions.
|
|
|
|
.. code-block:: text
|
|
|
|
class C {
|
|
bit V = true;
|
|
}
|
|
|
|
def X : C;
|
|
def Y : C {
|
|
let V = false;
|
|
string Greeting = "Hello!";
|
|
}
|
|
|
|
First, the abstract class ``C`` is defined. It has one field named ``V``
|
|
that is a bit initialized to true.
|
|
|
|
Next, two records are defined, derived from class ``C``; that is, with ``C``
|
|
as their parent class. Thus they both inherit the ``V`` field. Record ``Y``
|
|
also defines another string field, ``Greeting``, which is initialized to
|
|
``"Hello!"``. In addition, ``Y`` overrides the inherited ``V`` field,
|
|
setting it to false.
|
|
|
|
A class is useful for isolating the common features of multiple records in
|
|
one place. A class can initialize common fields to default values, but
|
|
records inheriting from that class can override the defaults.
|
|
|
|
TableGen supports the definition of parameterized classes as well as
|
|
nonparameterized ones. Parameterized classes specify a list of variable
|
|
declarations, which may optionally have defaults, that are bound when the
|
|
class is specified as a parent class of another class or record.
|
|
|
|
.. code-block:: text
|
|
|
|
class FPFormat <bits<3> val> {
|
|
bits<3> Value = val;
|
|
}
|
|
|
|
def NotFP : FPFormat<0>;
|
|
def ZeroArgFP : FPFormat<1>;
|
|
def OneArgFP : FPFormat<2>;
|
|
def OneArgFPRW : FPFormat<3>;
|
|
def TwoArgFP : FPFormat<4>;
|
|
def CompareFP : FPFormat<5>;
|
|
def CondMovFP : FPFormat<6>;
|
|
def SpecialFP : FPFormat<7>;
|
|
|
|
The purpose of the ``FPFormat`` class is to act as a sort of enumerated
|
|
type. It provides a single field, ``Value``, which holds a 3-bit number. Its
|
|
template argument, ``val``, is used to set the ``Value`` field. Each of the
|
|
eight records is defined with ``FPFormat`` as its parent class. The
|
|
enumeration value is passed in angle brackets as the template argument. Each
|
|
record will inherent the ``Value`` field with the appropriate enumeration
|
|
value.
|
|
|
|
Here is a more complex example of classes with template arguments. First, we
|
|
define a class similar to the ``FPFormat`` class above. It takes a template
|
|
argument and uses it to initialize a field named ``Value``. Then we define
|
|
four records that inherit the ``Value`` field with its four different
|
|
integer values.
|
|
|
|
.. code-block:: text
|
|
|
|
class ModRefVal <bits<2> val> {
|
|
bits<2> Value = val;
|
|
}
|
|
|
|
def None : ModRefVal<0>;
|
|
def Mod : ModRefVal<1>;
|
|
def Ref : ModRefVal<2>;
|
|
def ModRef : ModRefVal<3>;
|
|
|
|
This is somewhat contrived, but let's say we would like to examine the two
|
|
bits of the ``Value`` field independently. We can define a class that
|
|
accepts a ``ModRefVal`` record as a template argument and splits up its
|
|
value into two fields, one bit each. Then we can define records that inherit from
|
|
``ModRefBits`` and so acquire two fields from it, one for each bit in the
|
|
``ModRefVal`` record passed as the template argument.
|
|
|
|
.. code-block:: text
|
|
|
|
class ModRefBits <ModRefVal mrv> {
|
|
// Break the value up into its bits, which can provide a nice
|
|
// interface to the ModRefVal values.
|
|
bit isMod = mrv.Value{0};
|
|
bit isRef = mrv.Value{1};
|
|
}
|
|
|
|
// Example uses.
|
|
def foo : ModRefBits<Mod>;
|
|
def bar : ModRefBits<Ref>;
|
|
def snork : ModRefBits<ModRef>;
|
|
|
|
This illustrates how one class can be defined to reorganize the
|
|
fields in another class, thus hiding the internal representation of that
|
|
other class.
|
|
|
|
Running ``llvm-tblgen`` on the example prints the following definitions:
|
|
|
|
.. code-block:: text
|
|
|
|
def bar { // Value
|
|
bit isMod = 0;
|
|
bit isRef = 1;
|
|
}
|
|
def foo { // Value
|
|
bit isMod = 1;
|
|
bit isRef = 0;
|
|
}
|
|
def snork { // Value
|
|
bit isMod = 1;
|
|
bit isRef = 1;
|
|
}
|
|
|
|
``let`` --- override fields in classes or records
|
|
-------------------------------------------------
|
|
|
|
A ``let`` statement collects a set of field values (sometimes called
|
|
*bindings*) and applies them to all the classes and records defined by
|
|
statements within the scope of the ``let``.
|
|
|
|
.. productionlist::
|
|
Let: "let" `LetList` "in" "{" `Statement`* "}"
|
|
:| "let" `LetList` "in" `Statement`
|
|
LetList: `LetItem` ("," `LetItem`)*
|
|
LetItem: `TokIdentifier` ["<" `RangeList` ">"] "=" `Value`
|
|
|
|
The ``let`` statement establishes a scope, which is a sequence of statements
|
|
in braces or a single statement with no braces. The bindings in the
|
|
:token:`LetList` apply to the statements in that scope.
|
|
|
|
The field names in the :token:`LetList` must name fields in classes inherited by
|
|
the classes and records defined in the statements. The field values are
|
|
applied to the classes and records *after* the records inherit all the fields from
|
|
their parent classes. So the ``let`` acts to override inherited field
|
|
values. A ``let`` cannot override the value of a template argument.
|
|
|
|
Top-level ``let`` statements are often useful when a few fields need to be
|
|
overridden in several records. Here are two examples. Note that ``let``
|
|
statements can be nested.
|
|
|
|
.. code-block:: text
|
|
|
|
let isTerminator = true, isReturn = true, isBarrier = true, hasCtrlDep = true in
|
|
def RET : I<0xC3, RawFrm, (outs), (ins), "ret", [(X86retflag 0)]>;
|
|
|
|
let isCall = true in
|
|
// All calls clobber the non-callee saved registers...
|
|
let Defs = [EAX, ECX, EDX, FP0, FP1, FP2, FP3, FP4, FP5, FP6, ST0,
|
|
MM0, MM1, MM2, MM3, MM4, MM5, MM6, MM7, XMM0, XMM1, XMM2,
|
|
XMM3, XMM4, XMM5, XMM6, XMM7, EFLAGS] in {
|
|
def CALLpcrel32 : Ii32<0xE8, RawFrm, (outs), (ins i32imm:$dst, variable_ops),
|
|
"call\t${dst:call}", []>;
|
|
def CALL32r : I<0xFF, MRM2r, (outs), (ins GR32:$dst, variable_ops),
|
|
"call\t{*}$dst", [(X86call GR32:$dst)]>;
|
|
def CALL32m : I<0xFF, MRM2m, (outs), (ins i32mem:$dst, variable_ops),
|
|
"call\t{*}$dst", []>;
|
|
}
|
|
|
|
Note that a top-level ``let`` will not override fields defined in the classes or records
|
|
themselves.
|
|
|
|
|
|
``multiclass`` --- define multiple records
|
|
------------------------------------------
|
|
|
|
While classes with template arguments are a good way to factor out commonality
|
|
between multiple records, multiclasses allow a convenient method for
|
|
defining many records at once. For example, consider a 3-address
|
|
instruction architecture whose instructions come in two formats: ``reg = reg
|
|
op reg`` and ``reg = reg op imm`` (e.g., SPARC). We would like to specify in
|
|
one place that these two common formats exist, then in a separate place
|
|
specify what all the operations are. The ``multiclass`` and ``defm``
|
|
statements accomplish this goal. You can think of a multiclass as a macro or
|
|
template that expands into multiple records.
|
|
|
|
.. productionlist::
|
|
MultiClass: "multiclass" `TokIdentifier` [`TemplateArgList`]
|
|
: `ParentClassList`
|
|
: "{" `MultiClassStatement`+ "}"
|
|
MultiClassID: `TokIdentifier`
|
|
MultiClassStatement: `Assert` | `Def` | `Defm` | `Defvar` | `Foreach` | `If` | `Let`
|
|
|
|
As with regular classes, the multiclass has a name and can accept template
|
|
arguments. A multiclass can inherit from other multiclasses, which causes
|
|
the other multiclasses to be expanded and contribute to the record
|
|
definitions in the inheriting multiclass. The body of the multiclass
|
|
contains a series of statements that define records, using :token:`Def` and
|
|
:token:`Defm`. In addition, :token:`Defvar`, :token:`Foreach`, and
|
|
:token:`Let` statements can be used to factor out even more common elements.
|
|
The :token:`If` and :token:`Assert` statements can also be used.
|
|
|
|
Also as with regular classes, the multiclass has the implicit template
|
|
argument ``NAME`` (see NAME_). When a named (non-anonymous) record is
|
|
defined in a multiclass and the record's name does not include a use of the
|
|
template argument ``NAME``, such a use is automatically *prepended*
|
|
to the name. That is, the following are equivalent inside a multiclass::
|
|
|
|
def Foo ...
|
|
def NAME # Foo ...
|
|
|
|
The records defined in a multiclass are created when the multiclass is
|
|
"instantiated" or "invoked" by a ``defm`` statement outside the multiclass
|
|
definition. Each ``def`` statement in the multiclass produces a record. As
|
|
with top-level ``def`` statements, these definitions can inherit from
|
|
multiple parent classes.
|
|
|
|
See `Examples: multiclasses and defms`_ for examples.
|
|
|
|
|
|
``defm`` --- invoke multiclasses to define multiple records
|
|
-----------------------------------------------------------
|
|
|
|
Once multiclasses have been defined, you use the ``defm`` statement to
|
|
"invoke" them and process the multiple record definitions in those
|
|
multiclasses. Those record definitions are specified by ``def``
|
|
statements in the multiclasses, and indirectly by ``defm`` statements.
|
|
|
|
.. productionlist::
|
|
Defm: "defm" [`NameValue`] `ParentClassList` ";"
|
|
|
|
The optional :token:`NameValue` is formed in the same way as the name of a
|
|
``def``. The :token:`ParentClassList` is a colon followed by a list of at
|
|
least one multiclass and any number of regular classes. The multiclasses
|
|
must precede the regular classes. Note that the ``defm`` does not have a
|
|
body.
|
|
|
|
This statement instantiates all the records defined in all the specified
|
|
multiclasses, either directly by ``def`` statements or indirectly by
|
|
``defm`` statements. These records also receive the fields defined in any
|
|
regular classes included in the parent class list. This is useful for adding
|
|
a common set of fields to all the records created by the ``defm``.
|
|
|
|
The name is parsed in the same special mode used by ``def``. If the name is
|
|
not included, an unspecified but globally unique name is provided. That is,
|
|
the following examples end up with different names::
|
|
|
|
defm : SomeMultiClass<...>; // A globally unique name.
|
|
defm "" : SomeMultiClass<...>; // An empty name.
|
|
|
|
The ``defm`` statement can be used in a multiclass body. When this occurs,
|
|
the second variant is equivalent to::
|
|
|
|
defm NAME : SomeMultiClass<...>;
|
|
|
|
More generally, when ``defm`` occurs in a multiclass and its name does not
|
|
include a use of the implicit template argument ``NAME``, then ``NAME`` will
|
|
be prepended automatically. That is, the following are equivalent inside a
|
|
multiclass::
|
|
|
|
defm Foo : SomeMultiClass<...>;
|
|
defm NAME # Foo : SomeMultiClass<...>;
|
|
|
|
See `Examples: multiclasses and defms`_ for examples.
|
|
|
|
Examples: multiclasses and defms
|
|
--------------------------------
|
|
|
|
Here is a simple example using ``multiclass`` and ``defm``. Consider a
|
|
3-address instruction architecture whose instructions come in two formats:
|
|
``reg = reg op reg`` and ``reg = reg op imm`` (immediate). The SPARC is an
|
|
example of such an architecture.
|
|
|
|
.. code-block:: text
|
|
|
|
def ops;
|
|
def GPR;
|
|
def Imm;
|
|
class inst <int opc, string asmstr, dag operandlist>;
|
|
|
|
multiclass ri_inst <int opc, string asmstr> {
|
|
def _rr : inst<opc, !strconcat(asmstr, " $dst, $src1, $src2"),
|
|
(ops GPR:$dst, GPR:$src1, GPR:$src2)>;
|
|
def _ri : inst<opc, !strconcat(asmstr, " $dst, $src1, $src2"),
|
|
(ops GPR:$dst, GPR:$src1, Imm:$src2)>;
|
|
}
|
|
|
|
// Define records for each instruction in the RR and RI formats.
|
|
defm ADD : ri_inst<0b111, "add">;
|
|
defm SUB : ri_inst<0b101, "sub">;
|
|
defm MUL : ri_inst<0b100, "mul">;
|
|
|
|
Each use of the ``ri_inst`` multiclass defines two records, one with the
|
|
``_rr`` suffix and one with ``_ri``. Recall that the name of the ``defm``
|
|
that uses a multiclass is prepended to the names of the records defined in
|
|
that multiclass. So the resulting definitions are named::
|
|
|
|
ADD_rr, ADD_ri
|
|
SUB_rr, SUB_ri
|
|
MUL_rr, MUL_ri
|
|
|
|
Without the ``multiclass`` feature, the instructions would have to be
|
|
defined as follows.
|
|
|
|
.. code-block:: text
|
|
|
|
def ops;
|
|
def GPR;
|
|
def Imm;
|
|
class inst <int opc, string asmstr, dag operandlist>;
|
|
|
|
class rrinst <int opc, string asmstr>
|
|
: inst<opc, !strconcat(asmstr, " $dst, $src1, $src2"),
|
|
(ops GPR:$dst, GPR:$src1, GPR:$src2)>;
|
|
|
|
class riinst <int opc, string asmstr>
|
|
: inst<opc, !strconcat(asmstr, " $dst, $src1, $src2"),
|
|
(ops GPR:$dst, GPR:$src1, Imm:$src2)>;
|
|
|
|
// Define records for each instruction in the RR and RI formats.
|
|
def ADD_rr : rrinst<0b111, "add">;
|
|
def ADD_ri : riinst<0b111, "add">;
|
|
def SUB_rr : rrinst<0b101, "sub">;
|
|
def SUB_ri : riinst<0b101, "sub">;
|
|
def MUL_rr : rrinst<0b100, "mul">;
|
|
def MUL_ri : riinst<0b100, "mul">;
|
|
|
|
A ``defm`` can be used in a multiclass to "invoke" other multiclasses and
|
|
create the records defined in those multiclasses in addition to the records
|
|
defined in the current multiclass. In the following example, the ``basic_s``
|
|
and ``basic_p`` multiclasses contain ``defm`` statements that refer to the
|
|
``basic_r`` multiclass. The ``basic_r`` multiclass contains only ``def``
|
|
statements.
|
|
|
|
.. code-block:: text
|
|
|
|
class Instruction <bits<4> opc, string Name> {
|
|
bits<4> opcode = opc;
|
|
string name = Name;
|
|
}
|
|
|
|
multiclass basic_r <bits<4> opc> {
|
|
def rr : Instruction<opc, "rr">;
|
|
def rm : Instruction<opc, "rm">;
|
|
}
|
|
|
|
multiclass basic_s <bits<4> opc> {
|
|
defm SS : basic_r<opc>;
|
|
defm SD : basic_r<opc>;
|
|
def X : Instruction<opc, "x">;
|
|
}
|
|
|
|
multiclass basic_p <bits<4> opc> {
|
|
defm PS : basic_r<opc>;
|
|
defm PD : basic_r<opc>;
|
|
def Y : Instruction<opc, "y">;
|
|
}
|
|
|
|
defm ADD : basic_s<0xf>, basic_p<0xf>;
|
|
|
|
The final ``defm`` creates the following records, five from the ``basic_s``
|
|
multiclass and five from the ``basic_p`` multiclass::
|
|
|
|
ADDSSrr, ADDSSrm
|
|
ADDSDrr, ADDSDrm
|
|
ADDX
|
|
ADDPSrr, ADDPSrm
|
|
ADDPDrr, ADDPDrm
|
|
ADDY
|
|
|
|
A ``defm`` statement, both at top level and in a multiclass, can inherit
|
|
from regular classes in addition to multiclasses. The rule is that the
|
|
regular classes must be listed after the multiclasses, and there must be at least
|
|
one multiclass.
|
|
|
|
.. code-block:: text
|
|
|
|
class XD {
|
|
bits<4> Prefix = 11;
|
|
}
|
|
class XS {
|
|
bits<4> Prefix = 12;
|
|
}
|
|
class I <bits<4> op> {
|
|
bits<4> opcode = op;
|
|
}
|
|
|
|
multiclass R {
|
|
def rr : I<4>;
|
|
def rm : I<2>;
|
|
}
|
|
|
|
multiclass Y {
|
|
defm SS : R, XD; // First multiclass R, then regular class XD.
|
|
defm SD : R, XS;
|
|
}
|
|
|
|
defm Instr : Y;
|
|
|
|
This example will create four records, shown here in alphabetical order with
|
|
their fields.
|
|
|
|
.. code-block:: text
|
|
|
|
def InstrSDrm {
|
|
bits<4> opcode = { 0, 0, 1, 0 };
|
|
bits<4> Prefix = { 1, 1, 0, 0 };
|
|
}
|
|
|
|
def InstrSDrr {
|
|
bits<4> opcode = { 0, 1, 0, 0 };
|
|
bits<4> Prefix = { 1, 1, 0, 0 };
|
|
}
|
|
|
|
def InstrSSrm {
|
|
bits<4> opcode = { 0, 0, 1, 0 };
|
|
bits<4> Prefix = { 1, 0, 1, 1 };
|
|
}
|
|
|
|
def InstrSSrr {
|
|
bits<4> opcode = { 0, 1, 0, 0 };
|
|
bits<4> Prefix = { 1, 0, 1, 1 };
|
|
}
|
|
|
|
It's also possible to use ``let`` statements inside multiclasses, providing
|
|
another way to factor out commonality from the records, especially when
|
|
using several levels of multiclass instantiations.
|
|
|
|
.. code-block:: text
|
|
|
|
multiclass basic_r <bits<4> opc> {
|
|
let Predicates = [HasSSE2] in {
|
|
def rr : Instruction<opc, "rr">;
|
|
def rm : Instruction<opc, "rm">;
|
|
}
|
|
let Predicates = [HasSSE3] in
|
|
def rx : Instruction<opc, "rx">;
|
|
}
|
|
|
|
multiclass basic_ss <bits<4> opc> {
|
|
let IsDouble = false in
|
|
defm SS : basic_r<opc>;
|
|
|
|
let IsDouble = true in
|
|
defm SD : basic_r<opc>;
|
|
}
|
|
|
|
defm ADD : basic_ss<0xf>;
|
|
|
|
|
|
``defset`` --- create a definition set
|
|
--------------------------------------
|
|
|
|
The ``defset`` statement is used to collect a set of records into a global
|
|
list of records.
|
|
|
|
.. productionlist::
|
|
Defset: "defset" `Type` `TokIdentifier` "=" "{" `Statement`* "}"
|
|
|
|
All records defined inside the braces via ``def`` and ``defm`` are defined
|
|
as usual, and they are also collected in a global list of the given name
|
|
(:token:`TokIdentifier`).
|
|
|
|
The specified type must be ``list<``\ *class*\ ``>``, where *class* is some
|
|
record class. The ``defset`` statement establishes a scope for its
|
|
statements. It is an error to define a record in the scope of the
|
|
``defset`` that is not of type *class*.
|
|
|
|
The ``defset`` statement can be nested. The inner ``defset`` adds the
|
|
records to its own set, and all those records are also added to the outer
|
|
set.
|
|
|
|
Anonymous records created inside initialization expressions using the
|
|
``ClassID<...>`` syntax are not collected in the set.
|
|
|
|
|
|
``defvar`` --- define a variable
|
|
--------------------------------
|
|
|
|
A ``defvar`` statement defines a global variable. Its value can be used
|
|
throughout the statements that follow the definition.
|
|
|
|
.. productionlist::
|
|
Defvar: "defvar" `TokIdentifier` "=" `Value` ";"
|
|
|
|
The identifier on the left of the ``=`` is defined to be a global variable
|
|
whose value is given by the value expression on the right of the ``=``. The
|
|
type of the variable is automatically inferred.
|
|
|
|
Once a variable has been defined, it cannot be set to another value.
|
|
|
|
Variables defined in a top-level ``foreach`` go out of scope at the end of
|
|
each loop iteration, so their value in one iteration is not available in
|
|
the next iteration. The following ``defvar`` will not work::
|
|
|
|
defvar i = !add(i, 1);
|
|
|
|
Variables can also be defined with ``defvar`` in a record body. See
|
|
`Defvar in a Record Body`_ for more details.
|
|
|
|
``foreach`` --- iterate over a sequence of statements
|
|
-----------------------------------------------------
|
|
|
|
The ``foreach`` statement iterates over a series of statements, varying a
|
|
variable over a sequence of values.
|
|
|
|
.. productionlist::
|
|
Foreach: "foreach" `ForeachIterator` "in" "{" `Statement`* "}"
|
|
:| "foreach" `ForeachIterator` "in" `Statement`
|
|
ForeachIterator: `TokIdentifier` "=" ("{" `RangeList` "}" | `RangePiece` | `Value`)
|
|
|
|
The body of the ``foreach`` is a series of statements in braces or a
|
|
single statement with no braces. The statements are re-evaluated once for
|
|
each value in the range list, range piece, or single value. On each
|
|
iteration, the :token:`TokIdentifier` variable is set to the value and can
|
|
be used in the statements.
|
|
|
|
The statement list establishes an inner scope. Variables local to a
|
|
``foreach`` go out of scope at the end of each loop iteration, so their
|
|
values do not carry over from one iteration to the next. Foreach loops may
|
|
be nested.
|
|
|
|
The ``foreach`` statement can also be used in a record :token:`Body`.
|
|
|
|
.. Note that the productions involving RangeList and RangePiece have precedence
|
|
over the more generic value parsing based on the first token.
|
|
|
|
.. code-block:: text
|
|
|
|
foreach i = [0, 1, 2, 3] in {
|
|
def R#i : Register<...>;
|
|
def F#i : Register<...>;
|
|
}
|
|
|
|
This loop defines records named ``R0``, ``R1``, ``R2``, and ``R3``, along
|
|
with ``F0``, ``F1``, ``F2``, and ``F3``.
|
|
|
|
|
|
``if`` --- select statements based on a test
|
|
--------------------------------------------
|
|
|
|
The ``if`` statement allows one of two statement groups to be selected based
|
|
on the value of an expression.
|
|
|
|
.. productionlist::
|
|
If: "if" `Value` "then" `IfBody`
|
|
:| "if" `Value` "then" `IfBody` "else" `IfBody`
|
|
IfBody: "{" `Statement`* "}" | `Statement`
|
|
|
|
The value expression is evaluated. If it evaluates to true (in the same
|
|
sense used by the bang operators), then the statements following the
|
|
``then`` reserved word are processed. Otherwise, if there is an ``else``
|
|
reserved word, the statements following the ``else`` are processed. If the
|
|
value is false and there is no ``else`` arm, no statements are processed.
|
|
|
|
Because the braces around the ``then`` statements are optional, this grammar rule
|
|
has the usual ambiguity with "dangling else" clauses, and it is resolved in
|
|
the usual way: in a case like ``if v1 then if v2 then {...} else {...}``, the
|
|
``else`` associates with the inner ``if`` rather than the outer one.
|
|
|
|
The :token:`IfBody` of the then and else arms of the ``if`` establish an
|
|
inner scope. Any ``defvar`` variables defined in the bodies go out of scope
|
|
when the bodies are finished (see `Defvar in a Record Body`_ for more details).
|
|
|
|
The ``if`` statement can also be used in a record :token:`Body`.
|
|
|
|
|
|
``assert`` --- check that a condition is true
|
|
---------------------------------------------
|
|
|
|
The ``assert`` statement checks a boolean condition to be sure that it is true
|
|
and prints an error message if it is not.
|
|
|
|
.. productionlist::
|
|
Assert: "assert" `condition` "," `message` ";"
|
|
|
|
If the boolean condition is true, the statement does nothing. If the
|
|
condition is false, it prints a nonfatal error message. The **message**, which
|
|
can be an arbitrary string expression, is included in the error message as a
|
|
note. The exact behavior of the ``assert`` statement depends on its
|
|
placement.
|
|
|
|
* At top level, the assertion is checked immediately.
|
|
|
|
* In a record definition, the statement is saved and all assertions are
|
|
checked after the record is completely built.
|
|
|
|
* In a class definition, the assertions are saved and inherited by all
|
|
the subclasses and records that inherit from the class. The assertions are
|
|
then checked when the records are completely built.
|
|
|
|
* In a multiclass definition, the assertions are saved with the other
|
|
components of the multiclass and then checked each time the multiclass
|
|
is instantiated with ``defm``.
|
|
|
|
Using assertions in TableGen files can simplify record checking in TableGen
|
|
backends. Here is an example of an ``assert`` in two class definitions.
|
|
|
|
.. code-block:: text
|
|
|
|
class PersonName<string name> {
|
|
assert !le(!size(name), 32), "person name is too long: " # name;
|
|
string Name = name;
|
|
}
|
|
|
|
class Person<string name, int age> : PersonName<name> {
|
|
assert !and(!ge(age, 1), !le(age, 120)), "person age is invalid: " # age;
|
|
int Age = age;
|
|
}
|
|
|
|
def Rec20 : Person<"Donald Knuth", 60> {
|
|
...
|
|
}
|
|
|
|
|
|
Additional Details
|
|
==================
|
|
|
|
Directed acyclic graphs (DAGs)
|
|
------------------------------
|
|
|
|
A directed acyclic graph can be represented directly in TableGen using the
|
|
``dag`` datatype. A DAG node consists of an operator and zero or more
|
|
arguments (or operands). Each argument can be of any desired type. By using
|
|
another DAG node as an argument, an arbitrary graph of DAG nodes can be
|
|
built.
|
|
|
|
The syntax of a ``dag`` instance is:
|
|
|
|
``(`` *operator* *argument1*\ ``,`` *argument2*\ ``,`` ... ``)``
|
|
|
|
The operator must be present and must be a record. There can be zero or more
|
|
arguments, separated by commas. The operator and arguments can have three
|
|
formats.
|
|
|
|
====================== =============================================
|
|
Format Meaning
|
|
====================== =============================================
|
|
*value* argument value
|
|
*value*\ ``:``\ *name* argument value and associated name
|
|
*name* argument name with unset (uninitialized) value
|
|
====================== =============================================
|
|
|
|
The *value* can be any TableGen value. The *name*, if present, must be a
|
|
:token:`TokVarName`, which starts with a dollar sign (``$``). The purpose of
|
|
a name is to tag an operator or argument in a DAG with a particular meaning,
|
|
or to associate an argument in one DAG with a like-named argument in another
|
|
DAG.
|
|
|
|
The following bang operators are useful for working with DAGs:
|
|
``!con``, ``!dag``, ``!empty``, ``!foreach``, ``!getdagop``, ``!setdagop``, ``!size``.
|
|
|
|
Defvar in a record body
|
|
-----------------------
|
|
|
|
In addition to defining global variables, the ``defvar`` statement can
|
|
be used inside the :token:`Body` of a class or record definition to define
|
|
local variables. The scope of the variable extends from the ``defvar``
|
|
statement to the end of the body. It cannot be set to a different value
|
|
within its scope. The ``defvar`` statement can also be used in the statement
|
|
list of a ``foreach``, which establishes a scope.
|
|
|
|
A variable named ``V`` in an inner scope shadows (hides) any variables ``V``
|
|
in outer scopes. In particular, ``V`` in a record body shadows a global
|
|
``V``, and ``V`` in a ``foreach`` statement list shadows any ``V`` in
|
|
surrounding record or global scopes.
|
|
|
|
Variables defined in a ``foreach`` go out of scope at the end of
|
|
each loop iteration, so their value in one iteration is not available in
|
|
the next iteration. The following ``defvar`` will not work::
|
|
|
|
defvar i = !add(i, 1)
|
|
|
|
How records are built
|
|
---------------------
|
|
|
|
The following steps are taken by TableGen when a record is built. Classes are simply
|
|
abstract records and so go through the same steps.
|
|
|
|
1. Build the record name (:token:`NameValue`) and create an empty record.
|
|
|
|
2. Parse the parent classes in the :token:`ParentClassList` from left to
|
|
right, visiting each parent class's ancestor classes from top to bottom.
|
|
|
|
a. Add the fields from the parent class to the record.
|
|
b. Substitute the template arguments into those fields.
|
|
c. Add the parent class to the record's list of inherited classes.
|
|
|
|
3. Apply any top-level ``let`` bindings to the record. Recall that top-level
|
|
bindings only apply to inherited fields.
|
|
|
|
4. Parse the body of the record.
|
|
|
|
* Add any fields to the record.
|
|
* Modify the values of fields according to local ``let`` statements.
|
|
* Define any ``defvar`` variables.
|
|
|
|
5. Make a pass over all the fields to resolve any inter-field references.
|
|
|
|
6. Add the record to the final record list.
|
|
|
|
Because references between fields are resolved (step 5) after ``let`` bindings are
|
|
applied (step 3), the ``let`` statement has unusual power. For example:
|
|
|
|
.. code-block:: text
|
|
|
|
class C <int x> {
|
|
int Y = x;
|
|
int Yplus1 = !add(Y, 1);
|
|
int xplus1 = !add(x, 1);
|
|
}
|
|
|
|
let Y = 10 in {
|
|
def rec1 : C<5> {
|
|
}
|
|
}
|
|
|
|
def rec2 : C<5> {
|
|
let Y = 10;
|
|
}
|
|
|
|
In both cases, one where a top-level ``let`` is used to bind ``Y`` and one
|
|
where a local ``let`` does the same thing, the results are:
|
|
|
|
.. code-block:: text
|
|
|
|
def rec1 { // C
|
|
int Y = 10;
|
|
int Yplus1 = 11;
|
|
int xplus1 = 6;
|
|
}
|
|
def rec2 { // C
|
|
int Y = 10;
|
|
int Yplus1 = 11;
|
|
int xplus1 = 6;
|
|
}
|
|
|
|
``Yplus1`` is 11 because the ``let Y`` is performed before the ``!add(Y,
|
|
1)`` is resolved. Use this power wisely.
|
|
|
|
|
|
Using Classes as Subroutines
|
|
============================
|
|
|
|
As described in `Simple values`_, a class can be invoked in an expression
|
|
and passed template arguments. This causes TableGen to create a new anonymous
|
|
record inheriting from that class. As usual, the record receives all the
|
|
fields defined in the class.
|
|
|
|
This feature can be employed as a simple subroutine facility. The class can
|
|
use the template arguments to define various variables and fields, which end
|
|
up in the anonymous record. Those fields can then be retrieved in the
|
|
expression invoking the class as follows. Assume that the field ``ret``
|
|
contains the final value of the subroutine.
|
|
|
|
.. code-block:: text
|
|
|
|
int Result = ... CalcValue<arg>.ret ...;
|
|
|
|
The ``CalcValue`` class is invoked with the template argument ``arg``. It
|
|
calculates a value for the ``ret`` field, which is then retrieved at the
|
|
"point of call" in the initialization for the Result field. The anonymous
|
|
record created in this example serves no other purpose than to carry the
|
|
result value.
|
|
|
|
Here is a practical example. The class ``isValidSize`` determines whether a
|
|
specified number of bytes represents a valid data size. The bit ``ret`` is
|
|
set appropriately. The field ``ValidSize`` obtains its initial value by
|
|
invoking ``isValidSize`` with the data size and retrieving the ``ret`` field
|
|
from the resulting anonymous record.
|
|
|
|
.. code-block:: text
|
|
|
|
class isValidSize<int size> {
|
|
bit ret = !cond(!eq(size, 1): 1,
|
|
!eq(size, 2): 1,
|
|
!eq(size, 4): 1,
|
|
!eq(size, 8): 1,
|
|
!eq(size, 16): 1,
|
|
true: 0);
|
|
}
|
|
|
|
def Data1 {
|
|
int Size = ...;
|
|
bit ValidSize = isValidSize<Size>.ret;
|
|
}
|
|
|
|
Preprocessing Facilities
|
|
========================
|
|
|
|
The preprocessor embedded in TableGen is intended only for simple
|
|
conditional compilation. It supports the following directives, which are
|
|
specified somewhat informally.
|
|
|
|
.. productionlist::
|
|
LineBegin: beginning of line
|
|
LineEnd: newline | return | EOF
|
|
WhiteSpace: space | tab
|
|
CComment: "/*" ... "*/"
|
|
BCPLComment: "//" ... `LineEnd`
|
|
WhiteSpaceOrCComment: `WhiteSpace` | `CComment`
|
|
WhiteSpaceOrAnyComment: `WhiteSpace` | `CComment` | `BCPLComment`
|
|
MacroName: `ualpha` (`ualpha` | "0"..."9")*
|
|
PreDefine: `LineBegin` (`WhiteSpaceOrCComment`)*
|
|
: "#define" (`WhiteSpace`)+ `MacroName`
|
|
: (`WhiteSpaceOrAnyComment`)* `LineEnd`
|
|
PreIfdef: `LineBegin` (`WhiteSpaceOrCComment`)*
|
|
: ("#ifdef" | "#ifndef") (`WhiteSpace`)+ `MacroName`
|
|
: (`WhiteSpaceOrAnyComment`)* `LineEnd`
|
|
PreElse: `LineBegin` (`WhiteSpaceOrCComment`)*
|
|
: "#else" (`WhiteSpaceOrAnyComment`)* `LineEnd`
|
|
PreEndif: `LineBegin` (`WhiteSpaceOrCComment`)*
|
|
: "#endif" (`WhiteSpaceOrAnyComment`)* `LineEnd`
|
|
|
|
..
|
|
PreRegContentException: `PreIfdef` | `PreElse` | `PreEndif` | EOF
|
|
PreRegion: .* - `PreRegContentException`
|
|
:| `PreIfdef`
|
|
: (`PreRegion`)*
|
|
: [`PreElse`]
|
|
: (`PreRegion`)*
|
|
: `PreEndif`
|
|
|
|
A :token:`MacroName` can be defined anywhere in a TableGen file. The name has
|
|
no value; it can only be tested to see whether it is defined.
|
|
|
|
A macro test region begins with an ``#ifdef`` or ``#ifndef`` directive. If
|
|
the macro name is defined (``#ifdef``) or undefined (``#ifndef``), then the
|
|
source code between the directive and the corresponding ``#else`` or
|
|
``#endif`` is processed. If the test fails but there is an ``#else``
|
|
clause, the source code between the ``#else`` and the ``#endif`` is
|
|
processed. If the test fails and there is no ``#else`` clause, then no
|
|
source code in the test region is processed.
|
|
|
|
Test regions may be nested, but they must be properly nested. A region
|
|
started in a file must end in that file; that is, must have its
|
|
``#endif`` in the same file.
|
|
|
|
A :token:`MacroName` may be defined externally using the ``-D`` option on the
|
|
``*-tblgen`` command line::
|
|
|
|
llvm-tblgen self-reference.td -Dmacro1 -Dmacro3
|
|
|
|
Appendix A: Bang Operators
|
|
==========================
|
|
|
|
Bang operators act as functions in value expressions. A bang operator takes
|
|
one or more arguments, operates on them, and produces a result. If the
|
|
operator produces a boolean result, the result value will be 1 for true or 0
|
|
for false. When an operator tests a boolean argument, it interprets 0 as false
|
|
and non-0 as true.
|
|
|
|
.. warning::
|
|
The ``!getop`` and ``!setop`` bang operators are deprecated in favor of
|
|
``!getdagop`` and ``!setdagop``.
|
|
|
|
``!add(``\ *a*\ ``,`` *b*\ ``, ...)``
|
|
This operator adds *a*, *b*, etc., and produces the sum.
|
|
|
|
``!and(``\ *a*\ ``,`` *b*\ ``, ...)``
|
|
This operator does a bitwise AND on *a*, *b*, etc., and produces the
|
|
result. A logical AND can be performed if all the arguments are either
|
|
0 or 1.
|
|
|
|
``!cast<``\ *type*\ ``>(``\ *a*\ ``)``
|
|
This operator performs a cast on *a* and produces the result.
|
|
If *a* is not a string, then a straightforward cast is performed, say
|
|
between an ``int`` and a ``bit``, or between record types. This allows
|
|
casting a record to a class. If a record is cast to ``string``, the
|
|
record's name is produced.
|
|
|
|
If *a* is a string, then it is treated as a record name and looked up in
|
|
the list of all defined records. The resulting record is expected to be of
|
|
the specified *type*.
|
|
|
|
For example, if ``!cast<``\ *type*\ ``>(``\ *name*\ ``)``
|
|
appears in a multiclass definition, or in a
|
|
class instantiated inside a multiclass definition, and the *name* does not
|
|
reference any template arguments of the multiclass, then a record by
|
|
that name must have been instantiated earlier
|
|
in the source file. If *name* does reference
|
|
a template argument, then the lookup is delayed until ``defm`` statements
|
|
instantiating the multiclass (or later, if the defm occurs in another
|
|
multiclass and template arguments of the inner multiclass that are
|
|
referenced by *name* are substituted by values that themselves contain
|
|
references to template arguments of the outer multiclass).
|
|
|
|
If the type of *a* does not match *type*, TableGen raises an error.
|
|
|
|
``!con(``\ *a*\ ``,`` *b*\ ``, ...)``
|
|
This operator concatenates the DAG nodes *a*, *b*, etc. Their operations
|
|
must equal.
|
|
|
|
``!con((op a1:$name1, a2:$name2), (op b1:$name3))``
|
|
|
|
results in the DAG node ``(op a1:$name1, a2:$name2, b1:$name3)``.
|
|
|
|
``!cond(``\ *cond1* ``:`` *val1*\ ``,`` *cond2* ``:`` *val2*\ ``, ...,`` *condn* ``:`` *valn*\ ``)``
|
|
This operator tests *cond1* and returns *val1* if the result is true.
|
|
If false, the operator tests *cond2* and returns *val2* if the result is
|
|
true. And so forth. An error is reported if no conditions are true.
|
|
|
|
This example produces the sign word for an integer::
|
|
|
|
!cond(!lt(x, 0) : "negative", !eq(x, 0) : "zero", true : "positive")
|
|
|
|
``!dag(``\ *op*\ ``,`` *arguments*\ ``,`` *names*\ ``)``
|
|
This operator creates a DAG node with the given operator and
|
|
arguments. The *arguments* and *names* arguments must be lists
|
|
of equal length or uninitialized (``?``). The *names* argument
|
|
must be of type ``list<string>``.
|
|
|
|
Due to limitations of the type system, *arguments* must be a list of items
|
|
of a common type. In practice, this means that they should either have the
|
|
same type or be records with a common parent class. Mixing ``dag`` and
|
|
non-``dag`` items is not possible. However, ``?`` can be used.
|
|
|
|
Example: ``!dag(op, [a1, a2, ?], ["name1", "name2", "name3"])`` results in
|
|
``(op a1-value:$name1, a2-value:$name2, ?:$name3)``.
|
|
|
|
``!empty(``\ *a*\ ``)``
|
|
This operator produces 1 if the string, list, or DAG *a* is empty; 0 otherwise.
|
|
A dag is empty if it has no arguments; the operator does not count.
|
|
|
|
``!eq(`` *a*\ `,` *b*\ ``)``
|
|
This operator produces 1 if *a* is equal to *b*; 0 otherwise.
|
|
The arguments must be ``bit``, ``bits``, ``int``, ``string``, or
|
|
record values. Use ``!cast<string>`` to compare other types of objects.
|
|
|
|
``!filter(``\ *var*\ ``,`` *list*\ ``,`` *predicate*\ ``)``
|
|
|
|
This operator creates a new ``list`` by filtering the elements in
|
|
*list*. To perform the filtering, TableGen binds the variable *var* to each
|
|
element and then evaluates the *predicate* expression, which presumably
|
|
refers to *var*. The predicate must
|
|
produce a boolean value (``bit``, ``bits``, or ``int``). The value is
|
|
interpreted as with ``!if``:
|
|
if the value is 0, the element is not included in the new list. If the value
|
|
is anything else, the element is included.
|
|
|
|
``!find(``\ *string1*\ ``,`` *string2*\ [``,`` *start*]\ ``)``
|
|
This operator searches for *string2* in *string1* and produces its
|
|
position. The starting position of the search may be specified by *start*,
|
|
which can range between 0 and the length of *string1*; the default is 0.
|
|
If the string is not found, the result is -1.
|
|
|
|
``!foldl(``\ *init*\ ``,`` *list*\ ``,`` *acc*\ ``,`` *var*\ ``,`` *expr*\ ``)``
|
|
This operator performs a left-fold over the items in *list*. The
|
|
variable *acc* acts as the accumulator and is initialized to *init*.
|
|
The variable *var* is bound to each element in the *list*. The
|
|
expression is evaluated for each element and presumably uses *acc* and
|
|
*var* to calculate the accumulated value, which ``!foldl`` stores back in
|
|
*acc*. The type of *acc* is the same as *init*; the type of *var* is the
|
|
same as the elements of *list*; *expr* must have the same type as *init*.
|
|
|
|
The following example computes the total of the ``Number`` field in the
|
|
list of records in ``RecList``::
|
|
|
|
int x = !foldl(0, RecList, total, rec, !add(total, rec.Number));
|
|
|
|
If your goal is to filter the list and produce a new list that includes only
|
|
some of the elements, see ``!filter``.
|
|
|
|
``!foreach(``\ *var*\ ``,`` *sequence*\ ``,`` *expr*\ ``)``
|
|
This operator creates a new ``list``/``dag`` in which each element is a
|
|
function of the corresponding element in the *sequence* ``list``/``dag``.
|
|
To perform the function, TableGen binds the variable *var* to an element
|
|
and then evaluates the expression. The expression presumably refers
|
|
to the variable *var* and calculates the result value.
|
|
|
|
If you simply want to create a list of a certain length containing
|
|
the same value repeated multiple times, see ``!listsplat``.
|
|
|
|
``!ge(``\ *a*\ `,` *b*\ ``)``
|
|
This operator produces 1 if *a* is greater than or equal to *b*; 0 otherwise.
|
|
The arguments must be ``bit``, ``bits``, ``int``, or ``string`` values.
|
|
|
|
``!getdagop(``\ *dag*\ ``)`` --or-- ``!getdagop<``\ *type*\ ``>(``\ *dag*\ ``)``
|
|
This operator produces the operator of the given *dag* node.
|
|
Example: ``!getdagop((foo 1, 2))`` results in ``foo``. Recall that
|
|
DAG operators are always records.
|
|
|
|
The result of ``!getdagop`` can be used directly in a context where
|
|
any record class at all is acceptable (typically placing it into
|
|
another dag value). But in other contexts, it must be explicitly
|
|
cast to a particular class. The ``<``\ *type*\ ``>`` syntax is
|
|
provided to make this easy.
|
|
|
|
For example, to assign the result to a value of type ``BaseClass``, you
|
|
could write either of these::
|
|
|
|
BaseClass b = !getdagop<BaseClass>(someDag);
|
|
BaseClass b = !cast<BaseClass>(!getdagop(someDag));
|
|
|
|
But to create a new DAG node that reuses the operator from another, no
|
|
cast is necessary::
|
|
|
|
dag d = !dag(!getdagop(someDag), args, names);
|
|
|
|
``!gt(``\ *a*\ `,` *b*\ ``)``
|
|
This operator produces 1 if *a* is greater than *b*; 0 otherwise.
|
|
The arguments must be ``bit``, ``bits``, ``int``, or ``string`` values.
|
|
|
|
``!head(``\ *a*\ ``)``
|
|
This operator produces the zeroth element of the list *a*.
|
|
(See also ``!tail``.)
|
|
|
|
``!if(``\ *test*\ ``,`` *then*\ ``,`` *else*\ ``)``
|
|
This operator evaluates the *test*, which must produce a ``bit`` or
|
|
``int``. If the result is not 0, the *then* expression is produced; otherwise
|
|
the *else* expression is produced.
|
|
|
|
``!interleave(``\ *list*\ ``,`` *delim*\ ``)``
|
|
This operator concatenates the items in the *list*, interleaving the
|
|
*delim* string between each pair, and produces the resulting string.
|
|
The list can be a list of string, int, bits, or bit. An empty list
|
|
results in an empty string. The delimiter can be the empty string.
|
|
|
|
``!isa<``\ *type*\ ``>(``\ *a*\ ``)``
|
|
This operator produces 1 if the type of *a* is a subtype of the given *type*; 0
|
|
otherwise.
|
|
|
|
``!le(``\ *a*\ ``,`` *b*\ ``)``
|
|
This operator produces 1 if *a* is less than or equal to *b*; 0 otherwise.
|
|
The arguments must be ``bit``, ``bits``, ``int``, or ``string`` values.
|
|
|
|
``!listconcat(``\ *list1*\ ``,`` *list2*\ ``, ...)``
|
|
This operator concatenates the list arguments *list1*, *list2*, etc., and
|
|
produces the resulting list. The lists must have the same element type.
|
|
|
|
``!listsplat(``\ *value*\ ``,`` *count*\ ``)``
|
|
This operator produces a list of length *count* whose elements are all
|
|
equal to the *value*. For example, ``!listsplat(42, 3)`` results in
|
|
``[42, 42, 42]``.
|
|
|
|
``!lt(``\ *a*\ `,` *b*\ ``)``
|
|
This operator produces 1 if *a* is less than *b*; 0 otherwise.
|
|
The arguments must be ``bit``, ``bits``, ``int``, or ``string`` values.
|
|
|
|
``!mul(``\ *a*\ ``,`` *b*\ ``, ...)``
|
|
This operator multiplies *a*, *b*, etc., and produces the product.
|
|
|
|
``!ne(``\ *a*\ `,` *b*\ ``)``
|
|
This operator produces 1 if *a* is not equal to *b*; 0 otherwise.
|
|
The arguments must be ``bit``, ``bits``, ``int``, ``string``,
|
|
or record values. Use ``!cast<string>`` to compare other types of objects.
|
|
|
|
``!not(``\ *a*\ ``)``
|
|
This operator performs a logical NOT on *a*, which must be
|
|
an integer. The argument 0 results in 1 (true); any other
|
|
argument results in 0 (false).
|
|
|
|
``!or(``\ *a*\ ``,`` *b*\ ``, ...)``
|
|
This operator does a bitwise OR on *a*, *b*, etc., and produces the
|
|
result. A logical OR can be performed if all the arguments are either
|
|
0 or 1.
|
|
|
|
``!setdagop(``\ *dag*\ ``,`` *op*\ ``)``
|
|
This operator produces a DAG node with the same arguments as *dag*, but with its
|
|
operator replaced with *op*.
|
|
|
|
Example: ``!setdagop((foo 1, 2), bar)`` results in ``(bar 1, 2)``.
|
|
|
|
``!shl(``\ *a*\ ``,`` *count*\ ``)``
|
|
This operator shifts *a* left logically by *count* bits and produces the resulting
|
|
value. The operation is performed on a 64-bit integer; the result
|
|
is undefined for shift counts outside 0...63.
|
|
|
|
``!size(``\ *a*\ ``)``
|
|
This operator produces the size of the string, list, or dag *a*.
|
|
The size of a DAG is the number of arguments; the operator does not count.
|
|
|
|
``!sra(``\ *a*\ ``,`` *count*\ ``)``
|
|
This operator shifts *a* right arithmetically by *count* bits and produces the resulting
|
|
value. The operation is performed on a 64-bit integer; the result
|
|
is undefined for shift counts outside 0...63.
|
|
|
|
``!srl(``\ *a*\ ``,`` *count*\ ``)``
|
|
This operator shifts *a* right logically by *count* bits and produces the resulting
|
|
value. The operation is performed on a 64-bit integer; the result
|
|
is undefined for shift counts outside 0...63.
|
|
|
|
``!strconcat(``\ *str1*\ ``,`` *str2*\ ``, ...)``
|
|
This operator concatenates the string arguments *str1*, *str2*, etc., and
|
|
produces the resulting string.
|
|
|
|
``!sub(``\ *a*\ ``,`` *b*\ ``)``
|
|
This operator subtracts *b* from *a* and produces the arithmetic difference.
|
|
|
|
``!subst(``\ *target*\ ``,`` *repl*\ ``,`` *value*\ ``)``
|
|
This operator replaces all occurrences of the *target* in the *value* with
|
|
the *repl* and produces the resulting value. The *value* can
|
|
be a string, in which case substring substitution is performed.
|
|
|
|
The *value* can be a record name, in which case the operator produces the *repl*
|
|
record if the *target* record name equals the *value* record name; otherwise it
|
|
produces the *value*.
|
|
|
|
``!substr(``\ *string*\ ``,`` *start*\ [``,`` *length*]\ ``)``
|
|
This operator extracts a substring of the given *string*. The starting
|
|
position of the substring is specified by *start*, which can range
|
|
between 0 and the length of the string. The length of the substring
|
|
is specified by *length*; if not specified, the rest of the string is
|
|
extracted. The *start* and *length* arguments must be integers.
|
|
|
|
``!tail(``\ *a*\ ``)``
|
|
This operator produces a new list with all the elements
|
|
of the list *a* except for the zeroth one. (See also ``!head``.)
|
|
|
|
``!xor(``\ *a*\ ``,`` *b*\ ``, ...)``
|
|
This operator does a bitwise EXCLUSIVE OR on *a*, *b*, etc., and produces
|
|
the result. A logical XOR can be performed if all the arguments are either
|
|
0 or 1.
|
|
|
|
Appendix B: Paste Operator Examples
|
|
===================================
|
|
|
|
Here is an example illustrating the use of the paste operator in record names.
|
|
|
|
.. code-block:: text
|
|
|
|
defvar suffix = "_suffstring";
|
|
defvar some_ints = [0, 1, 2, 3];
|
|
|
|
def name # suffix {
|
|
}
|
|
|
|
foreach i = [1, 2] in {
|
|
def rec # i {
|
|
}
|
|
}
|
|
|
|
The first ``def`` does not use the value of the ``suffix`` variable. The
|
|
second def does use the value of the ``i`` iterator variable, because it is not a
|
|
global name. The following records are produced.
|
|
|
|
.. code-block:: text
|
|
|
|
def namesuffix {
|
|
}
|
|
def rec1 {
|
|
}
|
|
def rec2 {
|
|
}
|
|
|
|
Here is a second example illustrating the paste operator in field value expressions.
|
|
|
|
.. code-block:: text
|
|
|
|
def test {
|
|
string strings = suffix # suffix;
|
|
list<int> integers = some_ints # [4, 5, 6];
|
|
}
|
|
|
|
The ``strings`` field expression uses ``suffix`` on both sides of the paste
|
|
operator. It is evaluated normally on the left hand side, but taken verbatim
|
|
on the right hand side. The ``integers`` field expression uses the value of
|
|
the ``some_ints`` variable and a literal list. The following record is
|
|
produced.
|
|
|
|
.. code-block:: text
|
|
|
|
def test {
|
|
string strings = "_suffstringsuffix";
|
|
list<int> ints = [0, 1, 2, 3, 4, 5, 6];
|
|
}
|
|
|
|
|
|
Appendix C: Sample Record
|
|
=========================
|
|
|
|
One target machine supported by LLVM is the Intel x86. The following output
|
|
from TableGen shows the record that is created to represent the 32-bit
|
|
register-to-register ADD instruction.
|
|
|
|
.. code-block:: text
|
|
|
|
def ADD32rr { // InstructionEncoding Instruction X86Inst I ITy Sched BinOpRR BinOpRR_RF
|
|
int Size = 0;
|
|
string DecoderNamespace = "";
|
|
list<Predicate> Predicates = [];
|
|
string DecoderMethod = "";
|
|
bit hasCompleteDecoder = 1;
|
|
string Namespace = "X86";
|
|
dag OutOperandList = (outs GR32:$dst);
|
|
dag InOperandList = (ins GR32:$src1, GR32:$src2);
|
|
string AsmString = "add{l} {$src2, $src1|$src1, $src2}";
|
|
EncodingByHwMode EncodingInfos = ?;
|
|
list<dag> Pattern = [(set GR32:$dst, EFLAGS, (X86add_flag GR32:$src1, GR32:$src2))];
|
|
list<Register> Uses = [];
|
|
list<Register> Defs = [EFLAGS];
|
|
int CodeSize = 3;
|
|
int AddedComplexity = 0;
|
|
bit isPreISelOpcode = 0;
|
|
bit isReturn = 0;
|
|
bit isBranch = 0;
|
|
bit isEHScopeReturn = 0;
|
|
bit isIndirectBranch = 0;
|
|
bit isCompare = 0;
|
|
bit isMoveImm = 0;
|
|
bit isMoveReg = 0;
|
|
bit isBitcast = 0;
|
|
bit isSelect = 0;
|
|
bit isBarrier = 0;
|
|
bit isCall = 0;
|
|
bit isAdd = 0;
|
|
bit isTrap = 0;
|
|
bit canFoldAsLoad = 0;
|
|
bit mayLoad = ?;
|
|
bit mayStore = ?;
|
|
bit mayRaiseFPException = 0;
|
|
bit isConvertibleToThreeAddress = 1;
|
|
bit isCommutable = 1;
|
|
bit isTerminator = 0;
|
|
bit isReMaterializable = 0;
|
|
bit isPredicable = 0;
|
|
bit isUnpredicable = 0;
|
|
bit hasDelaySlot = 0;
|
|
bit usesCustomInserter = 0;
|
|
bit hasPostISelHook = 0;
|
|
bit hasCtrlDep = 0;
|
|
bit isNotDuplicable = 0;
|
|
bit isConvergent = 0;
|
|
bit isAuthenticated = 0;
|
|
bit isAsCheapAsAMove = 0;
|
|
bit hasExtraSrcRegAllocReq = 0;
|
|
bit hasExtraDefRegAllocReq = 0;
|
|
bit isRegSequence = 0;
|
|
bit isPseudo = 0;
|
|
bit isExtractSubreg = 0;
|
|
bit isInsertSubreg = 0;
|
|
bit variadicOpsAreDefs = 0;
|
|
bit hasSideEffects = ?;
|
|
bit isCodeGenOnly = 0;
|
|
bit isAsmParserOnly = 0;
|
|
bit hasNoSchedulingInfo = 0;
|
|
InstrItinClass Itinerary = NoItinerary;
|
|
list<SchedReadWrite> SchedRW = [WriteALU];
|
|
string Constraints = "$src1 = $dst";
|
|
string DisableEncoding = "";
|
|
string PostEncoderMethod = "";
|
|
bits<64> TSFlags = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 0 };
|
|
string AsmMatchConverter = "";
|
|
string TwoOperandAliasConstraint = "";
|
|
string AsmVariantName = "";
|
|
bit UseNamedOperandTable = 0;
|
|
bit FastISelShouldIgnore = 0;
|
|
bits<8> Opcode = { 0, 0, 0, 0, 0, 0, 0, 1 };
|
|
Format Form = MRMDestReg;
|
|
bits<7> FormBits = { 0, 1, 0, 1, 0, 0, 0 };
|
|
ImmType ImmT = NoImm;
|
|
bit ForceDisassemble = 0;
|
|
OperandSize OpSize = OpSize32;
|
|
bits<2> OpSizeBits = { 1, 0 };
|
|
AddressSize AdSize = AdSizeX;
|
|
bits<2> AdSizeBits = { 0, 0 };
|
|
Prefix OpPrefix = NoPrfx;
|
|
bits<3> OpPrefixBits = { 0, 0, 0 };
|
|
Map OpMap = OB;
|
|
bits<3> OpMapBits = { 0, 0, 0 };
|
|
bit hasREX_WPrefix = 0;
|
|
FPFormat FPForm = NotFP;
|
|
bit hasLockPrefix = 0;
|
|
Domain ExeDomain = GenericDomain;
|
|
bit hasREPPrefix = 0;
|
|
Encoding OpEnc = EncNormal;
|
|
bits<2> OpEncBits = { 0, 0 };
|
|
bit HasVEX_W = 0;
|
|
bit IgnoresVEX_W = 0;
|
|
bit EVEX_W1_VEX_W0 = 0;
|
|
bit hasVEX_4V = 0;
|
|
bit hasVEX_L = 0;
|
|
bit ignoresVEX_L = 0;
|
|
bit hasEVEX_K = 0;
|
|
bit hasEVEX_Z = 0;
|
|
bit hasEVEX_L2 = 0;
|
|
bit hasEVEX_B = 0;
|
|
bits<3> CD8_Form = { 0, 0, 0 };
|
|
int CD8_EltSize = 0;
|
|
bit hasEVEX_RC = 0;
|
|
bit hasNoTrackPrefix = 0;
|
|
bits<7> VectSize = { 0, 0, 1, 0, 0, 0, 0 };
|
|
bits<7> CD8_Scale = { 0, 0, 0, 0, 0, 0, 0 };
|
|
string FoldGenRegForm = ?;
|
|
string EVEX2VEXOverride = ?;
|
|
bit isMemoryFoldable = 1;
|
|
bit notEVEX2VEXConvertible = 0;
|
|
}
|
|
|
|
On the first line of the record, you can see that the ``ADD32rr`` record
|
|
inherited from eight classes. Although the inheritance hierarchy is complex,
|
|
using parent classes is much simpler than specifying the 109 individual
|
|
fields for each instruction.
|
|
|
|
Here is the code fragment used to define ``ADD32rr`` and multiple other
|
|
``ADD`` instructions:
|
|
|
|
.. code-block:: text
|
|
|
|
defm ADD : ArithBinOp_RF<0x00, 0x02, 0x04, "add", MRM0r, MRM0m,
|
|
X86add_flag, add, 1, 1, 1>;
|
|
|
|
The ``defm`` statement tells TableGen that ``ArithBinOp_RF`` is a
|
|
multiclass, which contains multiple concrete record definitions that inherit
|
|
from ``BinOpRR_RF``. That class, in turn, inherits from ``BinOpRR``, which
|
|
inherits from ``ITy`` and ``Sched``, and so forth. The fields are inherited
|
|
from all the parent classes; for example, ``IsIndirectBranch`` is inherited
|
|
from the ``Instruction`` class.
|